&EPA
          United States
          Environmental Protection
          Agency
          Office of Research and
          DeveJopment
          Washington DC 20460
EPA/600/9-90037
August 1090
          Research and Development
Remedial Action,
Treatment, and Disposal
of Hazardous Waste

Proceedings of the
Sixteenth Annual
RREL Hazardous
Waste Research
Symposium

-------
                                                     EPA/600/9-90/037
                                                     Aug.  1990
               REMEDIAL ACTION, TREATMENT, AND DISPOSAL
                          OF HAZARDOUS WASTE
Proceedings of the Sixteenth Annual Hazardous Waste Research Symposium
                   Cincinnati, Ohio, April 3-5, 1990
      Sponsored by the U.S. EPA, Office of Research & Development
                 Risk Reduction Engineering Laboratory
                         Cincinnati, OH  45268
                            Coordinated by:

                              JACA Corp.
                      Fort Washington, PA  19034

                         under subcontract to:

                            PEI Associates
                         Cincinnati, OH  45246
                           Project Officers:
                          Marta K. Richards
                            Eugene F. Harris
                             Dave Ferguson
                             Teri Shearer
                  RISK  REDUCTION  ENGINEERING  LABORATORY
                   OFFICE  OF  RESEARCH AND  DEVELOPMENT
                  U.S.  ENVIRONMENTAL PROTECTION AGENCY
                          CINCINNATI, OH   45268

-------
                                  NOTICE


        These  Proceedings  have been  reviewed  1n accordance with  the  U.S.
Environmental  Protection Agency's  peer and administrative  review policies
and approved  for  presentation and  publication.  Mention of  trade names or
commercial products  does  not constitute endorsement or  recommendation for
use*

-------
                                 FOREWORD


      Today's  rapidly  developing  technologies  and  Industrial  practices
frequently carry with  them the increased generation of materials,  that  1f
improperly dealt with, can threaten both public health and the environment.
The  U.S.   Environmental   Protection  Agency  Is  charged   by  Congress  with
protecting the Nation's land, air,  and water resources.  Under a mandate  of
national environmental  laws, the agency strives  to  formulate and implement
actions leading  to a  compatible balance between human activities  and the
ability of natural  systems  to  support  and  nurture  life.   These laws direct
the EPA to  perform research to define  our  environmental  problems,  measure
the impacts, and search for solutions.

      The  Risk   Reduction  Engineering  Laboratory   1s   responsible  for
planning,    implementing,   and   managing   research,   development,   and
demonstration  programs.     These   provide  an  authoritative,  defensible
engineering  basis  in  support of the policies,  programs,  and regulations  of
the  EPA  with  respect to  drinking  water,  wastewater,   pesticides,  toxic
substances,  solid  and hazardous wastes,  and Superfund-related activities.
This  publication 1s one  of the products  of that research  and provides a
vital communication link  between researchers and users.

      These  Proceedings  from  the  1990 Symposium provide  the  results  of
projects  recently  completed  by RREL  and  current information  on projects
presently underway.  Those  wishing additional  information on these projects
are urged to contact the  author or the  EPA  Project Officer,

      RREL  sponsors a symposium  each year  in  order to assure  that the
results  of  its  research  efforts  are rapidly  transmitted  to the  user
community.   The  1990  symposium attracted over 800 attendees from industry,
Federal and  State agencies, consulting  firms, and  universities.   The 1991
symposium  is planned for  April 9, 10,  and  11 in Cincinnati, Ohio.

                        E.  Timothy Oppelt,  Director
                    Risk Reduction Engineering  Laboratory

-------
                                 ABSTRACT


      The   Sixteenth   Annual  Research   Symposium  on  Remedial   Action,
Treatment,  and  Disposal  of  Hazardous  Waste  was held  in Cincinnati,  Ohio,
April 3-5,  1990.   The purpose of this Symposium was  to  present the latest
significant research  findings  from  ongoing and  recently  completed projects
funded by the Risk Reduction Engineering Laboratory (RREL).

      These Proceedings  are  organized in three sections:  Sessions  A  and B
consist of  paper  presentations.   Session C  contains  the poster abstracts.
Subjects  include   remedial  action  treatment  and  control  technologies  for
waste  disposal,  landfill  liner  and  cover  systems,  underground  storage
tanks,   and  demonstration   and    development  of  innovative/alternative
treatment   technologies  for  hazardous   waste.    Alternative  technology
subjects   include   thermal    destruction   of  hazardous   wastes,   field
evaluations,  existing  treatment  options,   emerging   treatment  processes,
waste minimization, and  biosystems  for hazardous waste destruction.
                                       IV

-------
                                 CONTENTS


                                 SESSION A
                                                                     Page
PIC Formation - Research Status and Control  Implications
  Barry Del linger, University of Dayton Research Institute	   1-11

Parametric Evaluation of Metal Partitioning  at the U.S. EPA
Incineration Research Facility
  Gregory J. Carroll, U.S. Environmental Protection Agency	   12-22

Sorption and Desorption of POHCs and PICs in a Full-Scale Boiler
Under Sooting Conditions
  Gary D. Hinshaw, Midwest Research Institute	   23-38

State of the Art Assessment of Medical Waste Incineration
Technology
  R.G. Barton, Energy and Environmental Research Corporation.,	   39-49

The Incineration of Arsenic-Contaminated Soils Related to the
Comprehensive Environmental Response, Compensation and Liability
Act (CERCLA)
  Howard 0. Wall, U.S. Environmental Protection Agency	.,	   50-58

Fundamental Studies on Particulate Emissions from Hazardous Waste
Incinerators
  Virendra Sethi, University  of Cincinnati	   59-67

Fate of Volatiles in Wastewater Treatment
  Richard A. Dobbs, U.S.  Environmental  Protection Agency	   68-86

Field Assessment  of Air Emissions from  Hazardous Waste Dewatering
Operations
  Thomas C. Ponder, Jr.,  PEI  Associates, Inc	   87-96

Bench-Scale Wet Air Oxidation of Complexed Metal Cyanide Sludges
  Ronald J. Turner, U.S.  Environmental  Protection Agency	   97-103

Passive Treatment of Metals Mine Drainage Through Use  of a
Constructed Wetland
  S.D. Machemer,  Colorado School of Mines	   104-114

Soil Vapor  Extraction Technology Assessment
  James T.  Curtis, Camp Dresser & McKee, Inc	   115-129

Algasorb®:  A New Technology  for Removal and Recovery  of Metal  Ions
from Groundwaters
  Dennis W. Darnall, Bio-Recovery Systems, Inc	   130-138

-------
                           SESSION A (Continued)

                                                                     Page

A SITE Demonstration of Ultrox UV/Ox1dat1on Technology
  Norma Lewis, U.S. Environmental Protection Agency	  139-148

SITE Demonstration of Biological Treatment of Groundwater by
BloTrol,  Inc. at a Wood Preserving Site in New Brighton, MN
  Mary K. Stinson, U.S. Environmental Protection Agency	  149-158

Factors Affecting the Reliability of Operations of the EPA Mobile
Incineration System
  James P. Stumbar, Foster Wheeler Env1 response, Inc	»..  159-177

Characterization of Internal Inspection Procedures & Equipment for
Underground Storage Tanks
  Susan E. Rohland, PEI Associates, Inc....	.	  178-191

Evaluation of Internal Leak Detection Technology for Large
Underground Storage Tanks
  Joseph  W. Maresca, Jr., Vista Research, Inc	  192-200

Source of Contamination Associated with Closure of Underground
.Storage Tanks
  Warren  J. Lyman, Camp Dresser & McKee Inc..	  201-206

Removal of Soluble Toxic Metals from Water
  L.P. Buckley, Atomic Energy of Canada Limited	  207-216

Laboratory Studies of the Thermal Destruction of Toxic Organ!cs in
Sewage Sludge
  Barry DelUnger, University of Dayton Research Institute	  217-227

Predicted and Observed Organic Emissions  from Sewage Sludge
Incineration
  Debra A. T1rey, University of Dayton Research Institute..........  228-237

Characterization of Municipal Waste Combustion Ashes and Leachate:
Field Study Results
  Ha1a K. Roffman, AWD Technologies, Inc	  238-250

Mobility  of Dioxlns and Furans Associated with  Stabilized
Incineration Residues  in the Marine Environment
  Frank J. Roethel, State University of New  York	  251-262

Water Movement Through an Experimental .Soil  Liner
  I.G. Krapac,  Illinois State Geological  Survey.	  263-273
                                       VI

-------
                               SESSION B

                                                                     Page

Hazardous and Toxic Wastes Associated with Urban Stormwater Runoff
  Robert E. P1tt, University of Alabama at Birmingham	  274-289

Degradation of Chlorinated Blphenyls by Escher1ch1a coll Containing
Cloned Genes of the Pseudomonas putlda Fl Toluene Catabollc Pathway
  Gerben J. Zylstra, University of Iowa College of Medicine..,,	  290-302

Use of DNA Hybridization Probes 1n the Monitoring of Blotreatment
Systems
  Richard A. Haugland, University of Cincinnati	  303-312

B1odegradab1l1ty Studies of the Alaskan Weathered Crude 011
Constituents with the Use of Electrolytic Resplrometry and Shaker
Flask Systems
  Henry H. Tabak, U.S. Environmental Protection Agency	  313-339

Biological Treatment of Wastewater Containing Hazardous Organic
Compounds:  2-Ethoxyethanol
  Margaret J. Kupferle, University of Cincinnati	  340-349

Bench-Scale B1odegradat1on Studies with Organic Pollutants Using a
White Rot Fungus
  John A. Glaser, U.S. Environmental Protection Agency	  350-361

Evaluation of EPA Soil Washing Technology for Remediation at IJST
Sites
  Mary E. Tabak, Camp Dresser & McKee Inc	  362-376

Vacuum-Assisted Steam Stripping to Remove Pollutants from
Contaminated Soil:  A Laboratory Study
  Arthur E. Lord, Jr., Drexel University	  377-395

Evaluation of Limited-Use Chemical Protective Clothing Used for EPA
Superfund Activities
  Jack C. Saw1ck1,  Arthur D. Little,  Inc	  396-407

Methods  to Evaluate the Stress Crack  Resistance of High Density
Polyethylene FML Sheets and Seams
  Y1ck H. Halse, Drexel University	  408-424

Landfill Test Cells
  N.C. vasukl, Delaware Solid Waste Authority	  425-438

Revision of  the  HELP Model
  Paul R.  Schroeder, USAE Waterways  Experiment  Station	  439-449

Waste Minimization  Assessments at  Selected  DOD  Facilities
  James  S. Bridges, U.S.  Environmental  Protection Agency	  450-460

Chemical Substitution  for  1,1,1-TMchloroethane and  Methanol  in
Manufacturing Operations
  Lisa M.  Brown, U.S.  Environmental  Protection  Agency	  461-471
                                       vii

-------
                           SESSION B (Continued)

                                                                     Page

Waste Minimization Assessment Centers
  F. William Klrsch, University City Science Center	  472-482

Veterans Affairs Hospital and Hospital Waste Minimization Case
Studies
  Kenneth R. Stone, U.S. Environmental Protection Agency	  483-500

A Ten-Year Review of Plastics Recycling
  S. Garry Howell, U.S. Environmental Protection Agency	  501-526

Evaluation of EPA Waste Minimization Assessment
  Mary Ann Curran, U.S. Environmental Protection Agency	  527-535

Slurry-Based Blotreatment of Still Bottoms Sorbed Onto Soil Fines
  Robert C. Ahlert, Rutgers University	  536-547

Anaerobic Treatment of Leachate
  Sanjoy K. Bhattacharya, Tulane University	  548-560

The Effects of Temperature, Concentration of Substrate and
Hicrobial Inoculum and the Source of Sludge Biomass on the Kinetics
of B1odegradat1on
  Henry H. Tabak, U.S. Environmental Protection Agency	  561-585

Chemical Compatibility of Geotextiles, Geonets and Pipes with
Hazardous Waste
  P. Cassldy, Southwest Texas State University	  586-595

Stabilization/Solidification for Treatment of Superfund Soils
  Franklin R. Alvarez, U.S. Environmental Protection Agency	  596-614

Treatabmty of Tr1 butyl tin 1n POTWs
  Richard A. Dobbs, U.S. Environmental Protection Agency	  615-634
                                      VIII

-------
                                 SESSION C

Superfund Technical Support Activities
  Benjamin L. Blaney, U.S. Environmental Protection Agency...	   635

Guide for Conducting TreatablHty Studies Under CERCLA
  Jonathan G. Herrmann, U.S. Environmental Protection Agency	   636

Implementing a Waste Minimization Program at the A.W. Breldenbach
Environmental Research Center
  Robert L. Gould, Science Applications International Corporation..   637

Biological and Physi co-Chemical Remediation of a Mercury
Contaminated Hazardous Waste
  Conly L. Hansen, Utah State University..	   638

Steam Stripping and Batch Distillation  for Removal and/or Recovery
of Volatile Organic Compounds
  Sardar Hassan, University of Cincinnati....	   639

Ozone/Ultraviolet  Light Treatment of  Ferri-Cyanide Complexes
  Mark J.  Briggs,  PEI Associates, Inc.	   640

The  USEPA/WRITE Program
  Paul M.  Randall, U.S. Environmental Protection Agency	   641

American  Institute for Pollution Prevention
  Thomas  R.  Hauser,  University of Cincinnati	   642

U.S. EPA  Incineration Research Facility Update
  Johannes W.  Lee, Acurex Corp	  643

Development  and Demonstration of a  Pilot-Scale Debris Washing
System
  Michael  Taylor,  PEI Associates, Inc..........	  644

Engineering  Handbook for  Hazardous  Waste Incineration
  Leo Weitzman, LVW  Associates,  Inc	  645

An Assessment  of  the Current State  of the Art of
Blo/Photodegradable  Plastics
  R.B. Sukol,  PEI  Associates,  Inc	  646

Field Tests  of Hydraulic  Fracturing to  Increase Fluid Flow  In Soils
  L.C. Murdoch, University  of Cincinnati	  647

Barrier  Equivalency  of Liner and Cap Materials
  David  E. Daniel, The University of Texas at Austin...	  648

An Upcoming  S.I.T.E. Demonstration:  AWD Technologies, Inc.,
 "AquaDetox/SVE System"
  Gordon  M.  Evans, U.S.  Environmental Protection Agency	  649

                                        ix

-------
                           SESSION C (Continued)

                                                                     Page

Guidance on Metals and PIC Emission Controls from Hazardous Waste
Incineration
  Shiva Garg, U.S. Environmental Protection Agency	   650

Reduction of Transient Puffs from a Rotary K1ln Incinerator:  The
Role of the Secondary Combustion Chamber
  P.M. Lemleux, U.S. Environmental Protection Agency	   651

The Use of Electrokinetics for Hazardous Waste Site Remediation
  Joseph T. Swartzbaugh, PEER Consultants, P.C	   652

Ground Penetrating Radar Equipped Field Robotic Vehicle for
Surveying CERCLA Sites
  James F. Osborn, Carnegie Mellon University	   653

Detoxification of Released Vapors/Particulates by Entrapment 1n
Chemically Active Foams
  Patricia M. Brown, Foster Wheeler Envlresponse, Inc	   654

Assessment of Materials Handling Technologies
  MaJ1d A. Dosanl, PEI Associates, Inc.....	   655

Metal Treatment by Adsorptlve Filtration
  Mark M. Benjamin, University of Washington	   656

-------
                      PIC Formation - Research Status and
                              Control Implications

                       Barry Bellinger, Philip H. Taylor,
                    Debra A. Tirey, and Richard C. Striebich
                          Environmental Sciences Group
                    University of Dayton Research Institute
                                300 College Park
                               Dayton, Ohio 45469

                                      and

                                   C. C. Lee
                                    U.S. EPA
                     Risk Reduction Engineering Laboratory
                         26 W. Martin Luther King Drive
                              Cincinnati, OH 45268

                                    ABSTRACT

     Control of emissions of toxic organic and organometallic compounds is one
of the major technical and sociological issues surrounding the further implemen-
tation of incineration as a waste disposal alternative.  In this paper, we
briefly review the status of research concerning the emission of organic
products of incomplete combustion (PICs).  Emphasis is placed on the complete-
ness of full-scale emissions test data and any identifiable deficiencies in our
fundamental knowledge of PIC formation.  From the laboratory perspective, the
urgent need for detailed studies of PIC characterization, yields, and emission
mechanisms at elevated temperatures is discussed.  The benefits of interactions
between laboratory-scale PIC studies, PIC lexicological studies, and sampling
and analysis methods development for full-scale systems is also discussed.

INTRODUCTION

     Of all the issues surrounding toxic waste disposal, the potential formation
and emission of so-called "Products of Incomplete Combustion" (PICs) from the
thermal disposal of toxic wastes is the most controversial.  Much of the con-
troversy is due to the lack of understanding of PICs, the current regulations
only addressing destruction and removal efficiency of so-called "Principal
Organic Hazardous Constituents" (POHCs) contained in the waste stream.  This has
led to the unfortunate public misconception of incineration as a "landfill in
the sky."
     It would seem that this important subject and controversy which has raged
all through the 80's would be well researched and significant strides made
towards answering key issues.  However, we in reality know very little about PIC
emissions.  The purpose of this discussion is to define what we real!y know
about PICs, what we think we know, and what we don't know.  Although every
effort has been made to approach this subject scientifically and arrive at
logical, factually-based conclusions, there will undoubtedly continue to arise
non-factual or even clearly incorrect statements concerning PICs.  It is hoped
that this paper will at least serve to establish a framework for rational dis-
cussion of the issues surrounding PICs.
                                      1

-------
     WHAT WE KNOW

           There are only a few issues for which we have sufficient data that agrees
     with scientific theory that will allow us to say we know this about PICs.
     Although we do pot. have enough data to advance a series of "Laws of PIC
     Formation," we can advance several  theories.  These theories are discussed below.

     Theory  0 - PICs are a natural  consequence of thermal degradation of toxic wastes

           This theory is so fundamental  and obvious that it is designated as the
     Zeroth  Theory of PIC Formation.  The formation of stable PICs has been
     repeatedly demonstrated in numerous lab and full-scale studies.   Possibly, one
     source  of confusion is the concentration or toxicity a combustion byproduct must
     have to be designated a PIC.   From a scientific point of view any product of
     the  oxidative or pyrolytic degradation of a compound, other than the ther-
     modynamlcally most stable end-products, should be Included in the definition of
     PIC.   If we want to designate  PICs  of environmental concern where toxicity and
     concentration are factors, then we  can define a new term; Toxic, Principal,
     Product of Incomplete Combustion (ToP-PIC),  The question is not if PICs are
     emitted from hazardous waste incinerators, because they will undoubtedly be
     formed  and emitted from any combustion source.  The true issue is whether they
     are  of  environmental consequence, which is a very complex issue indeed.

     Theory  1 - PIC emission rates  are kineticallv. not thermodynamicallv. controlled

           If an incinerator achieves thermodynamic equilibrium, then one only needs
     to know the elemental composition of the waste/fuel/oxidizer (air) feed system
     and  the temperature to calculate the emissions from the system.  For example,
     thermodynamically complete combustion or pyrolysis of a chlorinated hydrocarbon
     at 1800*F results in the formation  of a few stable products, e.g., C02, H«0,
     HC1,  H?, C12, etc., and3trace  quantities of carbon monoxide (CO) and organic
     byproducts \cf. Eqn. 1) :

               CXH Clf * (* * *J£) 9t • x C8- * t MCI + ££ tLC * trace orsinici and CO   for y>i
oxidation:                   *                    4   *
               CiHyC1* + x Oj » x C02 * y HC1 * t-y C1f * tract orfantcs and Cfl          for y
-------
     Modern,  well-designed and managed Incinerators operate  at temperatures in
excess of  1800'F,  residence times  greater than 2.0 s, and  at least 50% excess
air.   Detailed elementary and global  oxidation kinetic models consistently show
that destruction efficiencies (DE's)  for POHCs under thesi conditions are
greater than  99.9999% for even the most stable compounds.    Similar calculations
for PICs result in yields < 0.000001  (cf. columns 3, 4, and  6, Table 1).
     Since the nominal operating range of incinerators preclude measurable POHC
or PIC emissions,  we are left with the conclusion that temporal or spatial
excursions from the measured conditions are responsible for  the observed
emissions. These "failure modes"  may be due to excursions in temperature
(thermal), residence time (temporal),  or oxygen concentration (mixing).
                                     TABLE  1

               Kinetic Calculations of the Destruction Efficiency for PICs1
                for Various Flame and Post-Rame Reaction Conditions
  POHQ
  1,1-Dtehloroetharw  Chloroethene

  PentacWoroethane  TetraeMoroctrMme

  Chlorobenzene     Benzene

  Notes:
Optimal
 Flame
 DE2
                                                                  (Optimal)
                                                                   D£8
          Optima]   ThejmfJ
          Thermai    fjftffB
           IK9      d4

99.9999+ 99.9999+  99.9997   99.9999+

99.9999+ 99.9999+  27.0184   99.9999+

99.9999+ 99.9999+  92.2267   99.9999+
 Qyera|
(Fallum)

  J2E8

99,9999+

99.2702

99.9223
  1. PIC destruction calculation* based on analytical solution of th* following reaction scheme (cJ.
    reference 10 for Wnatte parameters):
                    M      fe
              POHC —> PIC —> Combustion Endproducts.

  2. DEca(aHatedforoptimaJfiam»ooodRton«(T(-2780aF,V-0.1 »).

  3. OE calculated for optima) post-flame commons 0> • 2dOO°F, tpf - 2.0 s).

  4. OE calculated tor sub-optimal post-flam* conditions (Tpf»1340°F, tpf - 2.0 s).

  S. Overall DE based on 99% destruction of the waste experiencing flame conditions and 1% of the waste
    bypassing the flame but experiencing optimal post-flame conditions.

  ft. Overa*DEt>ased on 99% destruction of the waste experiem^
    experiencing failure post-flame condWone.


WHAT WE THINK WE KNOW

      In addition to  these theories which the preponderance of the data support,
there is a body of qualitative and semi-quantitative lab and field data  that
have led to certain  observations concerning  PIC  emissions.  We refer to  these as
partially verified hypothesis.
                                        3

-------
Hypothesis 1  - Flamereactions control bulk or absolute emissions of PICs while
post-flame (thermal) reactions control relative emission rates

     This hypothesis is based on the reasonable assumption (Theory 2) that flame
temperatures are high enough that any material experiencing nominal oxidizing
f 1ame conditions will be totally destroyed (POHCs and PICs) (cf., column 3,
Table 1).  The small fraction of the waste that somehow escapes the flame is
responsible for observed POHC and PIC emissions.  Since compounds fed as mix-
tures will now experience very similar post-flame destruction conditions, the
relative thermal stability (oxidative or pyrolytic) of the escaping POHCs will
determine their relative emission rates.  In a similar manner, the post-flame
propensity for formation and stability of PICs will control their relative
emission rates (cf. column 7, Table 1),
     However, there is insufficient data on the range of temperature, residence
times, and oxygen concentrations that can potentially exist in an incinerator
flame to absolutely state that no PICs are emitted from the flame.  Although a
two zone system seems to be conceptually useful, there is also a transition zone
between the flame and post-flame zones where the distinction is blurred.  In
addition, differing vaporization rates of compounds can, in principle  result in
not all species experiencing the exact same combustion conditions.   However,
preliminary calculations suggest that droplet vaporization does not sig-
nificantly affect relative DEs.

Hypothesis 2 - Host PIC emissions are the result of pyrolvsis pathways

     This is essentially a narrowing of the focus of the failure mode concept
presented as the Second Theory of PIC Formation.  Theoretical calculations
concerning the oxidation of the most stable organic PICs indicate that tempera-
tures below 1450*F for post-flame mean residence times of 2.0 s (cf. Columns 5
and 7, Table 1), or residence times less than 0.1 s for post-flame mean tempera-
tures of 2300*F, are necessary for measurable PIC breakthrough.  Since these
temperatures are so low, it seems reasonable that pyrolysis pathways (which are
much slower than oxidation reactions) may be responsible for a large fraction of
POHC and PIC emissions.    This hypothesis has been,parti ally supported by
comparison of actual versus predicted CO emissions.    Since CO is only
destroyed under oxidative conditions (primarily by OH), its stack concentration
may be considered an indicator of the fraction of the incinerator which is
acting as an oxidlzer.  Theoretical calculations indicate that CO levels should
be on the order of a few ppb instead of the observed levels of 1 ppm - 100 ppm,
suggesting a significant fraction of the CO does not experience an oxidative
environment.  Full-scale emissions of undestroyed POHCs also agrees best with a
theoretical prediction of POHC stability based on pyrolysis kinetics.
     Recent bench-scale experiments simulating the burning of plastics in a
rotary kiln suggest that rapid thermal degradation results in a "transient puff"
of hydrocarbons and chlorocarbons that can overload the system such that
intimate waste/air mixing is not completed.    Consequently, high concentrations
of various PICs and soot are observed, apparently due to a transient pyrolysis
condition.
     Some researchers have argued that upset pyrolysis conditions in full-scale
incineration facilities actually result in lower PIC emissions.   However, other
researchers have shown that the high levels o|5soot that are formed may act as a
reservoir for the PICs by surface adsorption    Although air emissions may be
mitigated by a high efficiency particulate APCD, the total PIC yield and en-
vironmental burden is increased.  The impact is simply shifted from the air to
the land.
                                      4

-------
     There also exists a measurabl^number of cases where emissions do not
correlate with pyrolysls kinetics.    In these instances, relative POHC and PIC
emission rates may be due to other "physical" kinetic conditions which are
facility specific.

Hypothesis3- A significantfraction of the observed PICs is. formed outside
the high temperaturezones

     A detailed study of the types of PICs emitted from full-scale hazardous
waste incinerators, boilers, and kilns suggests that as high as 50% of observed
PICs are likely formed by radical-radical and radical-02 association reactions
(cf. Table 2).  These types of reactions are important as the system temperature
1s lowered because high activation energy unimolecular and radical-molecule
reactions are not favorable.  This may explain why thermally fragile PICs such
as chloroform, I,l,l-trichloroethane6 and carbon tetrachloride are observed in
the effluent of full-scale systems.    In addition, it has been proposed and
limited full-scale data suggest  that formation of polychlorinattd dibenzo-p-
dioxins (PCDDs) is due to surface catalyzed reactions in the cool,zones of the
incinerator (i.e., transfer ducts air pollution control devices).  *    Both lab
and full-scale research on the mechanism of PCDD formation is incomplete, and no
quantitative study of gas-phase formation of PCDD from possible precursors has
been performed.

Hypfttb.§s1s 4 - A significant fraction of observed PICs is  due to fuel/waste
interactions

     Analysis of full-scale data reveals that many of the observed PIC emissions
are aromatic and polynuclear aromatic (PNAs) species, i.e., benzene, toluene,
naphthalene, etc.   These compounds are commonly observed in the effluent from
most combustion devices.  However, further analysis indicates that the presence
of chlorine containing wastes Increases the emission of these species as well as
partially chlorinated hydrocarbons.AAlthough chlorination of a stable
hydrocarbon is kinetically unfavorable,   already chlorinated radicals may
participate in condensation-type molecular growth reactions resulting 1n forma-
tion of chlorinated aromatic species and PNAs.  Chlorine can also Increase
overall PlC-emissions through the traditionally proposed effect of flame
inhibition/"
     However, not enough parametric full-scale studies have been performed to
confirm the hypothesis.  Although preliminary lab studies support this concept,
the chemistry of chlorinated hydrocarbons and other hazardous materials 1s only
beginning to be addressed.

Hypothesis 5- Carbon monoxide and total unburned hydrocarbons are surrogates
for PIC emissions

     From a scientific viewpoint, the emission rate of a surrogate should corre-
late (viz. have a statistically significant relationship through a continuous
function) with the emission rate of a specific PIC or total PICs.  From a
regulatory viewpoint, it appears that a reasonable surrogate must only have a
concentration greater than the PIC of interest such that 1f the CO concentration
1s below a given ifvel, then the  PIC will be below a given environmentally
acceptable level.
     The concept of CO as a PIC surrogate arises from the fact that CO is more
stable than any organic PIC.  Total unburned hydrocarbons (TUHC) has been
                                      5

-------
                                         TABLE 2
       Prevalent, High-Yield PICsa-b and their Postulated Formation Mechanisms
                     from Full-Scale Incinerators, Boilers, and Kilns
Compound
Benzene
Chloroform
Mothylene Chloride
Toluene
Tetrachloroethylene
Carbon Tetrachtoride
Chloromethane
1,1,1 -Trichloroethane
1 ,2-DlchIoroethane
Naphthalene
Footnotes:
Geometric Mean0
Emission Rate (fio/min)
38,9SQ*
10,047®
6,757*
31,391®
2,250*
1,234^
96,408®
2,100<*
1,623*
21,053®
1,982f
96,408®
10,308®
9,593®
8,628®
2,977®
1,461*

Formation
Mechanism
Radical-
molecule
Association
Radical-
molecule
Radical-
molecule
Radical-
molecule
Association
Radical-
molecule
Association
Association
Radical-
molecule

• Prevalent PICs defined as PICs observed In at least 75% of all tests; this excludes tests at sKes where
PICs were designated as a POHC in the waste feed.
b High-yield PICs defined as PICs with geometric mean emission rates 2 1000 ng/mln for the Incinerator
and kflo tests and PICs wtth geometric emission mean rates * 7500 ng/mln for the boiler tests.
0 PIC emission rate calculated as the geometric mean for all sites; geometric mean of data for Individual
runs based on the number of measurements above quoted detection limits.
d PIC emission rate measured In kin tests.
* PIC emission rate measured hi boiler tests.
f PK) emission rale measured In indnerator tests.   -,

-------
proposed as a PIC surrogate because TUHC, in principal, is the total of un-
destroyed POHC and organic PICs.  Results of some full-scale tests indicate that
CO and TUHC2fett the regulatory definition of a surrogate for some PICs at some
facilities.  '    Furthermore, statistically significant rank/order correlations
between CO, TUHC andjvarious measures of PIC emissions were observed at a few
sites (cf. Table 3).   This is encouraging considering the limited range of
conditions tested and the number of data points available.
     However, the simplistic concept that the extreme stability (slow destruc-
tion kinetics) of CO makes it a suitable PIC surrogate is complicated by the
fact that it does not recognize the formation kinetics of CO versus organic
PICs.  Under some failure conditions, PIC yields may be high while CO formation
has yet to reach its maximum.

Hypothesis 6 - Surrogates that are indicative of poor waste/air mixing, viz.
benzene emissions rate, stack 0- concentration^ and waste/heat jpad,are
indicators of PIC emissions

     Rank-order statistical analyses of full-scale results have revealed that
some facilities show a significant positive correlation between PIC emission
rate and the following independent variables: normalized benzene emission rate,
normalized toluene emission rate, waste or heat load, and the stack oxygen
concentration (cf. Table 3).   Each of these parameters is indicative of poor
mixing of waste and air, resulting in poor combustion.
     Benzene and toluene are PICs expected to be formed under oxygen-deficient
conditions.  High waste or heat loads are indicative of system overloads such
that there is not sufficient air or time for complete mixing on the molecular
level.  The direct correlation between stack 02 concentration and PIC emissions
was unexpected, but is apparently an indicator of unburned 0« due to poor
waste/air mixing.  Although these correlations were not observed for every
facility and were not always highly significant (i.e., 80 and 90% confidence
intervals) they were the strongest observed in these comparisons.

WHAT WE DON'T KNOW

     There are clearly a number of issues concerning PICs that remain to be
resolved.

Issue 1 - Lack of carbon and chlorine mass balance

     Even the most detailed, complete full-scale studies have only accounted,fog
about 70% of the unburned waste/fuel feed including all facility effluents.  *
It is not clear whether this unaccounted material resides  in the bottom ash,
APCD ash, stack effluent, or other possible effluent streams.  Although it is
tempting to attribute this deficiency to analytical uncertainties, it may also
be true that our sampling and analytical methodologies are not j;ufficient to
detect all PICs being emitted from the facility.  Closer scrutiny must be placed
on determining total yields of a potential multitude of trace species that may
account for the mass balance deficiency.

Issue 2 - Are high and low molecular weight..PICs being full characterized?

     These species represent special sampling and analytical difficulties.  High
resolution capillary chromatography and liquid chromatography may be needed for
analysis of high molecular weight species, while gas-absorption chromatography
is necessary for very low molecular weight PICs.  Research has shown that high
                                      7

-------
                                  TABLE 3

Selected  Statistical  Rank-Order Correlations for  Pull-Scale
                  Hazardous Waste Incinerators
KEY:
btai* - no! «0«npto4
O
2
C
O
       ' comiuton
** dmto Ntf UNA wm '+' ind
   '-' cottvMfon co»ffidtnt»
 s
 T
 K

*v,%
                                               S
                                               T
                                               K
T
U
H
C
             C
             O
             L
             D
                     e
                     z

                     E
                     M
                     I
             C
             O
             N
             C
BZ+TOL

  1
  M
  I
  8
  S
             C
             O
             N
             C
                                                   txxn.3%O2  uaftacoL
TOTAL PIC EMISSION RATE (utfrrtn)
TOTAL PIC EMISSION CONCENTRATION (utffcorQ
TOTAL PIC EMISSION CONCENTRATION (10% OS}
% PIC EMISSION/ POHC MPUT
% PIC EMISSION /WASTE MPUT
TOTAL CHLORINATED PIC EMISSION RATE (ugJMn)
TOTAL CHLORINATED PIC EMISSION CONCENTRATION (ugAfecm)
TOTAL CHLORINATED PIC EMISSION CONCENTRATION (10% O2)
% CHLORINATED PIC EMISSION/ CHLORINATED POHC MPUT
% CHLORINATEOPIC EMISSION/ CHLORMATED WASTE MPUT
PIC: POHC RATIO
BENZENE EMISSION RATE (ugMn)
BENZENE EMISSION CONCENTRATION (ueUMtn)
BENZENE EMISSION CONCENTRATION (10% O2)
% BENZENE EMISSION/ POHC MPUT
% BENZENE EMISSION/ WASTE MPUT
TOLUENE EMISSION RATE (utfrnin)
TOLUENE EMISSION CONCENTRATION (ugfttocm)
TOLUENE EMISSION CONCENTRATION (10% OS)
% TOLUENE EMISSION / POHC MPUT
% TOLUENE EMISSION/ WASTE MPUT
TTOCHLOflOMETHANE EMISSION RATE (ugtoin)
TRICHLOROMETHANE EMISSION CONCENTRATION (ugMMfflt
TRICHLOROMETHANE EMISSION CONCENTRATION (10% O2)
% TRICHLOROMETHANE EMISSION / POHC MPUT
% TRICHLOROMETHANE EMISSION / WASTE MPUT
CHLOROMETHANE EMISSION RATE (ugftrtn)
CHLOOOMETHANE EMISSION CONCENTRATION (ugUtcn*
CHLOROU ETHANE EMISSION CONCENTRATION (10% OS)
% CHLOROMETHANE EMISSION / POHC MPUT
% CHLOROMETHANE EMISSION / WASTE MPUT
DCHLOROMETHANE EMISSION RATE (inMn)
DtCHLOROMETHANE EMISSION CONCENTRATION (UQAfcCffl)
DICHLOROMETKANE EMISSION CONCENTRATION (10% O2)
% WCHLOROMETHANE EMISSION / POHC MPUT
% DICHLOnOMETHANE EMISSION/ WASTE MPUT
1,3-QiCHLOROBENZENE EMISSION RATE (uaftrtnj
1>OICHLOROBENZENE EMISSION CONCENTRATION (ugAfccnt
1,3-DfCHLOROBENZENE EMISSION CONCENTRATION (10% 09)
% 1,3-DKHLOROBENZENE EMISSION / POHC MPUT
% 1,34»CHLQROBENZENE EMISSION/ WASTE MPUT
CHLOROBENZENE EMISSION RATE (u£Mn)
CHLOROBENZENE EMISSION CONCENTRATION (ugttiem)
CHLOROBENZENE EMISSION CONCENTRATION (10% O2)
% CHLOROBENZENE EMISSION / POHC MPUT
% CHLOROBENZENE EMISSION / WASTE MPUT
TRJCHLOROETHENE EMISSION RATE (ugMn)
TRJCHLOROETHENE EMISSION CONCENTRATION fugftfecm)
TWCHLOROETHENE EMISSION CONCENTRATION (10% OS)
% TR1CHLORQETHENE EMISSION / POHC MPUT
% TRICHLOROETHENE EMISSION / WASTE MPUT
PENTACHLOROETHANE EMISSION RATE (u»Mn)
PENTACHLOROETHANE EMISSION CONCENTRATION (u0ttKm)
PENTACHLOROETHANE EMISSION CONCENTRATION (10% 02)
% PENTACHLOROETHANE EMISSION / POHC MPUT
% PENTACHLOROETHANE EMISSION / WASTE MPUT
1,1,1-TRJCHLOROETHANE EMISSION RATE (uoMn)
1.1.1-TRtCHLOROeTHANE EMISSION CONCENTRATION (utfkcnf
1.1.1-TRBHLOROETHANE EMISSION CONCENTRATION (10% O2)
% 1,1.1-THCHLOROeTHANe EMISSION / POHC NVT
% 1.1.1-TRICHLOROETHANE EMISSION / WASTE MPUT
1.14MCHLOROETHENE EMISSION RATE (uflftrfnl
1.1-DICHLOROETHENE EMISSION CONCENTRATION (ugAtefflO
1,1-DICHLOROETHENE EMISSION CONCENTRATION (10% 00}
% 1,1-DICHLOROETHENE EMISSION / POHC INPUT
% 1,1-OICHLOROETHENE EMISSION / WASTE MPUT
VNYt CHLORIDE EMISSION RATE (ugto*»
VtM. CHLORIDE EMISSION CONCENTRATION (ugftfeon)
Vtm. CHLOROE EMISSION CONCENTRATION (10% O2)
% VMYL CHLORCE EMISSION / POHC MPUT
% VMYL CHLCfUDE EMISSION / WASTE MPUT
I-




2/4
?''
21'
1 /<
2/'
O/l
O/!






1/3

O/
1 /*
1 1'
i 1
( /
1/!
5'
1 14
• t'l
„ 	 ,('/
ll
II,
' r
• "T-











•

•
*
1/1
1/1*
















0/1
§/
/

-------
molecular weighlgPICs are numerous but are usually present in very low
concentrations.    This further exacerbates the analytic problems.  Quantitative
trapping and recovery of low molecular weight PICs (e.g., ehloromethane,
phosgene) is the major problem that must be addressed to achieve a complete
emissions characterization.

Issue 3 - What are environmentally acceptable PIC emission rates?

     This is of course a key issue and probably the most complex.  Because many
PICs may be formed in ultra-trace concentrations, we may never identify all
possible PIC emissions and thus never be able to ascertain the biological ac-
tivity of hazardous waste incinerator emissions.  Since some species may act as
activators, promoters, or inhibitors of biological activity of other species, a
precise knowledge of all PICs would be necessary to accurately evaluate the
influence of PICs on human health.

Issue 4 - What is the relative health risk associated with PICs from different
sources?

     Since the evaluation of the biological activity of PICs from hazardous
waste incinerators may be prohibitively time-consuming and expensive, it would
be wise to be sure we are focusing on the most serious offender.  Sufficient PIC
data has not been obtained on a wide range of stationary combustion sources
(e.g., municipal incinerators, sewage sludge incinerators, medical waste in-
cinerators, coal and oil fired power plants, or other industrial processes) to
determine which source results in the greatest environmental insult.  It will be
necessary to retest these sources using consistent sampling and analytical
procedures, searching for the same PICs before we will truly know their relative
environmental impact.  Only then can we focus our energies on the sources of the
most serious environmental problems.

CONTROL IMPLICATIONS

     Based on a review of theoretical, laboratory, and full-scale results in
combination with the failure mode theories of PIC emissions, several concepts
have been identified for consideration for development into PIC control
strategies.  Although many of these ideas have been previously proposed, suffi-
cient research has now been completed such that these concepts can be
implemented with a firmer scientific basis.  We have divided these potential
strategies into design, operation, and regulatory control implications as follows:

Design Implications
•  Enhanced waste/air micro-mixing
•  Plug flow conditions in post flame zones
•  Positive temperature gradient in high-temperature zones
t  Slow quenching in cool zones

Operation Implications
•  Low concentration of direct chlorinated PIC precursors
•  Low C1:H ratios in the waste or waste/fuel mixture
•  Waste homogenization to remove high chlorine pockets

Regulatory Implications
•  Comprehensive emission modeling
•  Test burn sampling for key ToP-PICs

-------
*  Continuous monitoring of PICs

     In conclusion, we briefly discuss the regulatory PIC control implications.
It now seems feasible to establish a reaction kinetic data base that can be used
to model and predict the relative emission rates of residual POHCs and PICs from
full-scale systems.  Limited trial burn emissions testing on a few "calibration"
PICs can then be used to convert these relative emission rates into quantitative
emissions estimates for "all" PICs and residual POHCs.  These emissions es-
timates can be used in conjunction with available potency data to conduct risk
assessments to ensure regulatory compliances.  Even when the appropriate
toxicological data is not available, this approach can still be used to establish
a list of target PICs for trial burn testing.  Such data is also invaluable to
the toxicological community who must ultimately determine the potency of
specific pollutants.  This approach may also drive the development of reliable
and appropriate PIC continuous monitors.

Credit

     This work was conducted under the partial sponsorship by the U.S. EPA under
Co-operative Agreement CR-813938.

REFERENCES

1.  Dellinger, B, Taylor, P.H., and Tirey, D.A., "Minimization and Control of
Hazardous Combustion Byproducts," Final Report Prepared for EPA Co-operative
Agreement CR-813938, February 1990.

2.  Lee, K.C., (1987), JAPCA. 31, 1011.

3.  Senkan, S.M., (1982), In Detoxification of Hazardous Waste. Exner, J.H.,
ed., Ann Arbor Science Publishers, Ann Arbor, MI, p. 61.

4.  Chang, D.P.Y., Hournighan, R.E, and Huffman, G.L., (1987), Thirteenth Annual
EPA Research Symp., EPA/600/9-87/015, p. 235.

5.  Hart, O.H., (1989), "The Use of Combustion Modeling in Permitting Hazardous
Waste Incinerators," 3rd Annual National Symposium on the Incineration of
Industrial Wastes, San Diego, CA, paper No. 15.

6.  Sorbo, N.W., Steeper, R.R., Law, C.K., and Chang, D.P.Y., (1990), "An
Experimental Investigation of the Incineration and Incinerability of Chlorinated
Alkane Droplets," Twenty-Second Symp. (Int.) on Combustion, The Combustion
Institute, in press.

7.  Graham, M.D., Taylor, P.H., and Dellinger, B., (1986), "A Model Study of
Vaporization and Thermal Oxidation of a Multi-component Hazardous Waste
Droplet," Proceedings of the Eastern States Section - Combustion Institute
Meeting, San Juan, Puerto Rico.

8.  Dellinger, B., Graham, M.D., and Tirey, D.A., (1986), Haz. Waste Haz. Mat..
1, 293.

9.  Kramlich, J.C., Heop, H.P., Seeker, W.R., and Samuel son, G.S., (1984),
Twentieth Svmp. (Int.) on Combustion, The Combustion Institute, p. 1991.
                                      10

-------
10.  (a) Dellinger, 8., Torres, J.L., Rubey, W.A., Hall, D.L., and Graham, J.L.,
(1984), Haz. Waste. I, 137. (b) Lee, K.C., Morgan, H., Hansen, J.L., Whipple,
G.M., (1982), 75th Annual Air Pollution Control Assoc. Meeting, New Orleans, LA,
paper 82-5.3.

11.  Graham, J.L., Hall, D.L., and Dellinger, B., (1986), Environ. Sci.
Techno!..  20> 703.

12.  Tsang, W. (1986), "Fundamental Aspects of Key Issues in Hazardous Waste
Incineration," ASME Publication 86 - WA/HT. 27.

13.  Dellinger, B., Rubey, W.A., Hall, D.L., and Graham, J.L., (1986), Haz.
Waste Haz. Mater., 3, 139.

14.  (a) Linak, W.P. Kilgroe, J.D., McSorley, J.A., Wendt, J.D.L., and Dunn,
J.E., (1987), JAPCA. 37, 54.  (b) Linak, W. P., McSorley, J. A., Wendt, J. D.
L., and Dunn, J. E., (1987), JAPCA. 37, 934.

15.  VanDell, R.D. and Mahle, H.H. (1988), "The Rate of Carbon Particulate
Surface Area on PIC Emission,", 195th ACS National Meeting,Preprint Extended
Abstract.  28, 76.

16.  Trenholm, A., Gorman, P., and Jungclaus, G., (1985), "Performance
Evaluation of Full-Scale Hazardous Waste Incinerators,", Vol.  I-V, EPA-600/2-84-
181 a-e, U.S. EPA, Cincinnati, OH.

17.  Hagenmaier, H., Kraft, M., Brunner, H., and Haag, R., 1987,, Environ.
Sci. Techno!.. 21, 1080.

18.  Municipal Waste Combustion Assessment: Technical Basis for Good Combustion
Practice,  EPA-600/8-89-063, prepared by Air and Energy Engineering Research
Laboratory, RTP, NC, August 1989.

19.  Tsang, W., (1990),  "Mechanisms for the Formation and Destruction of
Chlorinated Organic Products of Incomplete Combustion," Submitted to
Combust. Sci. Technol.

20.  Westbrook, C.K.,  (1982), Nineteenth Svmp. (Int.) on Combustion. The
Combustion  Institute, p. 127.

21.  Report of the Products of Incomplete Combustion Subcommittee of the Science
Advisory Board, "Review  of OSW Proposed Controls for Hazardous Waste
Incineration Products of Incomplete Combustion," EPA Science Advisory Board, US-
EPA, EPA-SAB-EC-90-004,  January 1990.

22.  Guidance on PIC Controls for Hazardous Waste Incinerators, Vol. 5 of the
Hazardous Waste Incineration Guidance Series, Draft Final Report prepared for
US-EPA/OSW, April, 1989.

23.  Dellinger, B. and Hall, D.L., (1986), JAPCA. 36, 179.

24.  Trenholm, A.  (1988), Measurements of Particulates, Metals, and Organics at
a Hazardous Waste  Incinerator, Draft final report, EPA Prime Contract No. 69-01-
7287, Work Assignment No. 36.


25.  Dellinger,  B.,  (1987), JAPCA, 37,  1019.

                                      11

-------
                PARAMETRIC EVALUATION OF METAL PARTITIONING AT
                  THE  U.S.  EPA  INCINERATION  RESEARCH FACILITY

                  by:    Gregory J. Carroll
                        U.S. Environmental Protection  Agency
                        Risk Reduction Engineering Laboratory
                        Cincinnati, OH 45268

                        Larry R. Water!and
                        Donald J. Fournier,  Jr.
                        Acurex Corporation
                        Mountain View, CA 94039


                                   ABSTRACT


      Two series of tests were performed at the U.S.  EPA Incineration  Research
Facility (IRF) to evaluate the fate of trace metals fed to a pilot-scale
rotary kiln incinerator, as a function of incinerator  operating temperatures
and feed chlorine content.  The two series were similar in all  respects with
the exception of the device used for particulate and  acid gas control.  In the
first series, the primary air pollution control device (APCD) consisted of a
venturi scrubber followed by a packed column scrubber.  In the second  series,
an ionizing wet scrubber was evaluated as the primary  APCD.

      Synthetic waste formulations contained five hazardous trace metals
(arsenic, barium, cadmium, chromium, and lead) and four nonhazardous  trace
metals (bismuth, copper, magnesium, and strontium) spiked onto a clay
absorbent material.  Test variables were waste feed chlorine content  (0,  4 and
8 percent), kiln temperature (816°, 871° and 927°C),  and afterburner  temper-
ature (98E°, 1093° and 1204°C).

      A number of conclusions were made from the results of the first test
series.  Cadmium, lead and bismuth are relatively volatile, based on  normal-
ized discharge distribution data; less than 32 percent of their discharge was
accounted for by kiln ash.  Barium, copper, strontium, chromium and magnesium
are relatively nonvolatile; more than 75 percent of their discharge was in the
kiln ash.  Average apparent scrubber efficiency for the individual  metals
ranged from 32 to 88 percent; scrubber efficiency for  the three volatile
metals was lower than that for five of the six nonvolatile metals.  Both  kiln
ash partitioning and scrubber efficiency appear to be  impacted negatively by
increases 1n feed chlorine content and, to a lesser extent, increases in  kiln
temperature.  With the exception of arsenic, discharge distributions  of the
metals correlate strongly with volatility temperatures.


                                      12

-------
      Data from the second test series are currently being evaluated.
Preliminary review indicates that the results generally support the
conclusions of the first test series.

      This paper has been reviewed in accordance with the U.S.  Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.

                                 INTRODUCTION


      Metals, such as arsenic, barium, beryllium, cadmium, chromium,  lead,
mercury, nickel, and zinc are of concern in waste incineration  because  of
their presence in many hazardous wastes and because of possible adverse health
effects from human exposure to emissions.  Incineration will  change the form
of metal fractions in waste streams, but it will not destroy the metals. As  a
result, metals are expected to emerge from the combustion zone  essentially  in
the same total quantity as the input.  The principal environmental  concern
centers around where and in what physical or chemical  form the  metals end up
in the combustion system, i.e., bottom ash, air pollution control  device
(APCD) residues, or stack emissions.

      Most interest has traditionally focused on stack emissions of metals.
However, increasing attention is being given to the quality of  residuals from
incineration of metal-bearing wastes since disposal of these materials  is
becoming subject to restrictions on land disposal under the Hazardous and
Solid Waste Act of 1984 (HSWA) (1).

      A major role of the U.S. EPA Incineration Research Facility (IRF) in
Jefferson, Arkansas is to perform research in support of regulatory develop-
ment by the U.S. EPA Office of Solid Waste (OSW).  Accordingly, the IRF has
conducted research to study the partitioning of trace metals during the
incineration of a metal-bearing surrogate waste (2).  The testing described
herein took place in the Rotary Kiln System (RKS) of the IRF.

                                    PURPOSE


      Field studies to date indicate that emissions of metals from  incin-
erators burning wastes of relatively low metal  content probably do  not  pose an
unacceptable level of risk.  However, as the level  of metals in wastes  which
are incinerated rises, there is concern that some incinerators  could  create
unacceptable risk.  Current regulatory thinking suggests implementation of
metal emission limits and metal feedrate limits, calculated using health risk
levels and dispersion modeling.

      Information gathered from the subject research will be used to  further
examine the impact of metals upon environmental emissions and to assist in
developing and refining regulatory strategies for dealing with  such an
impact (3).
                                      13

-------
      The research was designed to identify:

-  the distribution of metal emissions among kiln ash, scrubber blowdown
   water, flue gas particulate and flue gas vapor

-  the sensitivity of metal fate to RKS operating conditions

-  the dependence of metal emissions on chlorine content of the incinerated
   organic waste

-  the collection efficiency of several air pollution control  devices (APCDs)
   for the metals.

      This testing was conducted in conjunction with the development of a
metal partitioning model  under an independent EPA contract; data from these
tests will be used in part to evaluate the predictive capability of the model.

                                  APPROACH


      As indicated in Figure 1, the RKS consists of a rotary kiln primary
combustion chamber followed by an afterburner.  Combustion gases exiting the
afterburner are quenched after which they enter a primary air  pollution
control  system.  In the first of two test series, this system  consisted of a
venturi  scrubber and packed column scrubber.  In the second test series, an
ionizing wet scrubber (IWS) was substituted as the primary APCD.  Hereafter,
the first and second test series will be referred to as the "venturi" and
"IWS" test series, respectively.  During both series, primary  air pollution
control  was followed by a secondary system consisting of a carbon bed adsorber
and a high efficiency particulate filter.

      The feed material consisted of synthetic waste formulations mixed
into a clay absorbent material.  In each of the four-week series of parametric
tests, metals in an aqueous solution were spiked onto the solid material,
which was then screw-fed  to the RKS.

      The tests examined the fate of the hazardous constituent trace elements
arsenic  (As), lead (Pb),  cadmium (Cd), chromium (Cr), and barium (Ba).   In
addition, the nonhazardous elements copper (Cu), magnesium (Mg), bismuth (Bi),
and strontium (Sr) were spiked onto the test feed to establish whether  their
discharge distributions were comparable to those for the hazardous elements.
In addition to the metals, toluene, chlorobenzene, and tetrachloroethylene
were added to the solid material in predetermined amounts to achieve the
necessary chlorine levels and heat content in the feed.

      Kiln temperature, afterburner temperature and the chlorine content of
the organic liquid were varied based on a factorial experimental design for
three variables over three levels.  The target values for the  three variables
were: feed liquid matrix chlorine content (0, 4 and 8 percent); kiln temp-
erature  (816°, 871° and 927°C [1500°, 1600° and 1700°F]); and  afterburner
temperature (982°, 1093°  and 1204°C  [1800°, 2000° and 2200°F]).


                                      14

-------
           ROTARY
           KILN
          ROTARY KILN
          INCINERATOR
MODULAR PRIMARY AIR
 POLLUTION CONTROL
     DEVICES
.   REDUNDANT AIR
 POLLUTION CONTROL
1     SYSTEM
               Figure  1.   Schematic  of  the rotary kiln system
      Table 1 gives the values for each of the test variables for seven test
points specified by the design algorithm.  The eighth test point represents a
duplicate of Test Point 4.  This duplicate test condition was added to provide
information on test measurement precision in both test series, and was a
component of a trial burn undertaken during the IWS test series.  Test Point 9
provided the final component of the triplicate measurements for the trial
burn.

      Table 2 lists the feed concentrations of each of the metals.  These
values represent the sum of metals in the spike solution and background levels
of metals in the clay.

      All tests were conducted at the same nominal kiln exit flue gas oxygen
(12%), afterburner exit flue gas oxygen (7.5%), and synthetic waste feedrate
(63 kg/hr [140 lb/ hr], of which 18 kg/hr (40 Ib/hr) was the organic liquid
matrix).  For all tests, the kiln rotational speed was held constant at 0.2
rpm, and the primary APCDs were operated at their nominal design settings.

      Sampling for each test included:

- a composite sample of all feed material (clay/organic liquid mixture
  just before kiln introduction, and aqueous metal spike solution)
                                      IS

-------
- samples of scrubber blowdown water over time
- composite samples of kiln ash from each test
- samples of the flue gas at the afterburner and scrubber exits  for metal
  capture in the particulate and backup impinger train
- samples of the flue gas at the afterburner and scrubber exits  for volatile
  organic hazardous constituents (2).
Table 1.
Test
1
2
3
4
5
6
7
8a
9b
aTest
Point
Parametric test design conditions.
Feed Cl
Content
(wt %)
0
4
4
4
4
4
8
4
4
Point 8 Is
4.
bTest Point 9 is
Point 4 and was
purposes in the
Kiln Exit Afterburner
Temp, Exit Temp,
QC (°F) °c (°F)
871
(1,600)
816
(1,500)
927
(1,700)
871
(1,600)
871
(1,600)
871
(1,600)
871
(1,600)
871
(1,600)
871
(1,600)
a duplicate
1,093
(2,000)
1,093
(2,000)
1,093
(2,000)
1,093
(2,000)
1,204
(2,200)
982
(1,800)
1,093
(2,000)
1,093
(2,000)
1,093
(2,000)
of Test
a triplicate of Test
added for trial burn
IMS test series only.
                                                Table 2.   Nominal  feed metal
                                                          concentrations.3
Metal
Arsenic
Barium
Bismuth
Cadmium
Chromium
Copper
Lead
Magnesium
Strontium
                                                             Concentration
                                                                (ppm)
                                                                  50
                                                                  50
                                                                 180
                                                                  10
                                                                  90
                                                                 500
                                                                  50
                                                               17,000
                                                                 300
                                                aBased  on  average  clay
                                                 matrix metals  concen-
                                                 trations  of: Bi  (12  ppm);
                                                 Cr (53 ppm); Pb  (3 ppm);
                                                 Mg (2.2 percent); Sr (34
                                                 ppm).
                                    16

-------
                                   RESULTS
      Evaluation of the results of the venturi  test series  has  been  completed.
Table 3 represents a summary of metal  discharge distribution  data  collected
during the eight-test series.  Analysis of data from the IWS  test  series  is
not yet complete.  Results from that series are therefore not presented
herein; this discussion focusses only on the results of the venturi  test
series.

TEST CONDITIONS

      Actual test conditions achieved were very close to target test
conditions for all test points except Point 2.   The average afterburner exit
temperature was approximately 22°C lower than the target temperature of 1093°C
for this test.  For all other parametric tests, average actual  operating
temperatures were within 10°C of target conditions.

      Actual feed chlorine content was 0 percent for Test 1;  8.3 percent  for
Test 7; and ranged from 3.4 to 4.6 percent for Tests 2 through  6,  and Test 8.

METAL DISCHARGE DISTRIBUTIONS

      Metals exiting the kiln followed one of three pathways; they either left
in the kiln ash, or traveled to the afterburner with fly ash  or as volatilized
compounds.  In this discussion, a distinction is made between volatile and
nonvolatile (or refractory) metals.  These are relative terms based  on the
amount of each metal found in the kiln ash, as  a percentage of the total  metal
measured among the process streams (ash, scrubber blowdown, and flue gas).  An
assumption in making these distinctions is that metals traveling from the kiln
to the afterburner are predominantly the result of volatilization, rather than
entrainment in fly ash.

      Cadmium, lead, and bismuth appeared to be the most volatile  of the  nine
metals over the range of kiln temperatures tested, with average partitioning
to the kiln ash of <15, 20, and 32 percent of the measured  amounts for each of
the three metals, respectively.

      Barium, copper, strontium, arsenic, chromium, and magnesium  were
relatively nonvolatile, with average partitioning of 77, 79,  89, 91, 93,  and
99 percent to the kiln ash for each of the metals, respectively.

      In Figure 2, the discharge distribution data of each  of the  metals  is
plotted versus their volatility temperatures (the temperatures  at  which the
vapor pressure of the most volatile principal species of each metal  under
oxidizing conditions is 10 ~6 atm).   With the exception of arsenic, the
distribution of the metals correlates strongly with that which  would be
predicted by volatility temperatures.  The fact that arsenic  is significantly
less volatile than expected suggests that either a refractory compound is
preferred ever the predominant arsenic species (AS203), or  that some chemical
interaction, such as strong adsorption to the clay, occurred  (4).
                                      17

-------
Table 3.  Normalized metal  discharge distributions  (% of total measured).
Test Huober: 2
Primary
Variable: Kiln
Target Value: 816
Test Average: 825
4

8

3

Exit Temperature (°C)
871
878
Constants 1;2: Afterburner'
Constant 1 Avg.: 1071
Constant 2 Avg.: 3.7
Arsenic
Kiln Ash 94.4
Scrubber Exit Gas 2.1-2.9
Scrubber Water 2.7
BarlMi
Kiln Ash 74.3
Scrubber Exit Gas 3.8
Scrubber Hater 21.9
Blswth
Kiln Ash 25.8
Scrubber Exit Gas 41.5
Scrubber Hater 32.6
Cadaiuo
Kiln Ash <15.2
Scrubber Exit Gas 43-49
Scrubber Water 42-51
Chroalw
Kiln Ash 94.7
Scrubber Exit Gas 3.0
Scrubber Water 2.3
Copper
Kiln Ash 84.2
Scrubber Exit Gas 12.9
Scrubber Water 3.0
Lead
Kiln Ash 12.6
Scrubber Exit 6as 50.4
Scrubber Water 37.0
Nagneslwi
Kiln Ash 99.4
Scrubber Exit Gas 0.2
Scrubber Hater 0.4
Strontiua
Kiln Ash 82.9
Scrubber Exit Gas l.l
Scrubber Water 16.0
1088
3.8

86.1
3.8-5.8
8.2

79.6
2.2
18.2

22.2
41.1
36.7

<10.3
56-61
34-39

94.1
2.0
3.9

75.8
15.1
9.1

15.0
48.9
36.1

99.3
0.1
0.6

93.0
1.7
5.3
871
873
1093°C; C1
1094
4.6

92.3
2.3-4.1
3.6

69.9
5.5
24.7

30.0
35.2
34.7

<12.9
42-45
45-55

85.9
1.9
12.2

76.2
17.8
5.9

13.7
50.2
36.0

99.3
0.2
0.5

89.8
3.5
6.6
927
928
>4X
1092
4.2

84.0
6.8-8.4
7.6

69.6
1.6
28.8

22.9
50.7
26.3

<10.7
62-69
27-31

95.3
2.1
2.6

82.3
14.1
3.6

10.4
67.2
22.4

99.5
0.1
0.4

90.1
1.6
8.3
6

4

Afterburner
982
983

875
3.4

93.6
2.6-3.8
2.6

85.2
2.2
12.5

20.9
47.4
31.6

<13.9
61-69
25-31

95.5
1.1
3.4

79.2
15.2
5.6

5.8
60.6
33.6

99.3
0.1
0.6

94.3
1.3
4.4
1093
1088
Kiln-
878
3.8

86.1
3.8-5.
8.2

79.6
2.2
18.2

22.2
41.1
36.7

<10.3
56-61
34-39

94.1
2.0
3.9

75.8
15.1
9.1

15.0
48.9
36.1

99.3
0.1
0.6

93.0
1.7
5.3
8

Teaperature
1093
1094
5

(°C)
1204
1196
871°C; Cl«4*
873
4.6

92.3
8 2.3-4.1
3.6

69.9
5.5
24.7

30.0
35.2
34.7

<12.9
42-45
45-55

85.9
1.9
12.2

76.2
17.8
5.9

13.7
50.2
36.0

99.3
0.2
0.5

89.8
3.5
6.6
871
3.6

91.2
3.0-4.3
4.6

86.9
1.6
11.5

30.1
37.1
32.8

<14.5
55-62
31-38

89.3
4.2
6.6

75.1
16.0
8.9

13.8
45.0
41.1

99.2
0.1
0.7

81.9
3.7
14.4
1

Feed
0
0
Kiln
874
1093

93.9
1.7-2.2
3.9

68.8
2.0
28.8

64.8
15.7
19.5

<29.3
42-54
29-46

95.7
1.4
2.8

97.6
0.8
1.6

83.7
11.6
4.7

99.6
0.03
0.36

91.8
2.5
5.7
4

Chlorine
4
3.8
- 871°C;
878
1088

86.1
3.8-5.8
8.2

79.6
2.2
18.2

22.2
41.1
36.7

<10.3
56-61
34-39

94.1
2.0
3.9

75.8
15.1
9.1

15.0
48.9
36.1

99.3
0.1
0.6

93.0
1.7
5.3
C

Content
4
4.6
7

(wt J)
8
8.3
AC • 1093°C
873
1094

92.3
2.3-4.1
3.6

69.9
5.5
24.7

30.0
35.2
34.7

<12.9
42-45
45-55

85.9
1.9
12.2

76.2
17.8
5.9

13.7
50.2
36.0

99.3
0.2
0.5

89.8
3.5
6.6
870
1092

92.4
4.0-4.8
2.7

78.6
2.4
19.0

36.3
38.4
25.4

<9.3
68-74
68-74

92.1
2.8
5.1

58.0
33.2
8.8

6.0
73.6
20.3

99.4
0.1
0.5

90.7
1.6
7.7

-------
                             KILN ASH
1UU
Los 80
Z w 60
— LU
|fe 4°
£ 20
0
c
JAS _ 7 sr?"9!0'
—
-
Bi
-
r
• Cd


•«
1
1 Ba - - Cu •*•
4-
Pb

Ll 1 1 1 1 1
) 200 400 600 800 1000 1200 1400 1600
s
g
c/o
LU




VOLATILITY TEMPERATURE CC)
45
tr 40
gOQ 35
5: go: 30
111 2S
^ > LLJ
Z§3 20
P1;:
5
0
C
SCRUBBER FLUE GAS
—
~
-

- Cd Bi

-
I*




Pb

-- Ba
tcu Sr"Mgicr
i i i i i •*• i i * r
200 400 600 800 1000 1200 1400 1600






VOLATILITY TEMPERATURE CC)
80
| 70
5 Q 60
ffl LU
|gi so
CO LU LU 40-
1^5 30
| & 20
£ 10
0
SCRUBBER SLOWDOWN
—
-
—
—
Cd

er
—
—
•^



S*8 1 1



Pb


.•_ Cu Sr u_ Cr
1 Ba^l | ..| 1*??*







         200    400    600    800    1000   1200   1400
                      VOLATILITY TEMPERATURE CC)
1600
Figure  2.  Distribution  of metals  in discharge stream.
                             19

-------
PARAMETRIC EFFECTS

Kiln Ash Part 111 oning

      The only metals for which kiln ash partitioning appeared to be affected
by changes in kiln temperature were the most volatile metals (cadmium and
lead) and arsenic.  For these three metals, slight decreases in kiln ash
fractions were seen with increasing kiln temperature.  Kiln temperature
effects for the other individual metals, as well as for total  metals, were
not significant within data variability.

      The effect of increasing feed chlorine content on the partitioning of
metals appeared to be much more significant than that of kiln  temperature.
As kiln and afterburner temperatures were held constant, and feed chlorine
content increased from 0.0 to 8.3 percent, the overall fraction of metals
partitioning to the kiln ash dropped from 81 to 63 percent.

      Kiln ash partitioning of the three volatile metals decreased measurably
as chlorine content rose.  In the case of bismuth, the fraction of the
measured metal found in the kiln ash dropped from 65 to 36 percent as feed
chlorine concentration increased from 0.0 to 8.3 percent.  Kiln ash fractions
of cadmium and lead dropped from 29 to 9 percent, and from 84  to 6 percent,
respectively, over the same chlorine range.

      Only one of the refractory metals showed an increase in  volatility with
increasing feed chlorine content.  The fraction of measured copper found in
the kiln ash dropped from 98 to 58 percent as feed chlorine increased.  The
volatility of the other refractory metals appeared relatively  stable with
increasing feed chlorine.

Scrubber Efficiency

      The split of metals between scrubber exit gas and scrubber water
determines the apparent scrubber efficiency, defined as the ratio of metals
in the water to the sum of the metals in the two splits.

      Scrubber efficiency for total metals ranged from 33 to 56 percent, and
to a slight extent was negatively impacted by increases in kiln temperature
and feed chlorine content.

      Average scrubber efficiencies for the individual metals  ranged from a
low of 32 percent for copper to a high of 88 percent for barium.  Average
scrubber efficiency for the three volatile metals was lower than that for five
of the six nonvolatile metals.

      Kiln temperature effects on scrubber efficiency were measurable in the
cases of two of the volatile metals.  Decreases in scrubber efficiency, from
51 to 31 percent for cadmium and from 42 to 25 percent for lead, were seen
with increases in kiln temperature from 825°C to 927°C.  This  is consistent
with the changes in volatilization of these metals experienced with changes in
kiln temperature.


                                      20

-------
      The effect of chlorine on scrubber efficiency for individual  metals  was
significant for copper, arsenic, bismuth, and cadmium.   In each case,  the
effect of increasing feed chlorine was to decrease scrubber efficiency.
Decreases were as follows: copper (from 67 to 21 percent); arsenic  (67 to  38
percent); cadmium (46 to 26 percent); and bismuth (55 to 40 percent).

      The effect of afterburner temperature on scrubber efficiency  for
Individual metals was less substantial than that of kiln temperature and feed
chlorine content.  Modest increases in scrubber efficiency were seen along the
increasing afterburner temperature range for bismuth and lead.

Flue Ga s Phase Distribution

      Standard Method 5 trains were used to collect samples of flue gas
particulate- and vapor-phase metals.  In this discussion, two assumptions
regarding this sampling technique are made: (1) all particulate-phase metals
are collected on the filters, and (2) all vapor-phase metals are collected in
the impingers; vapor-phase metals do not condense on the filters, nor do they
pass through the impingers. (In actuality, the "vapor-phase" likely includes
some water-soluble metal that has wept through the filters and collected in
the impingers.)

      The distribution of metals in the flue gas favored the particulate phase
over vapor.  Average particulate phase metals as a fraction of total flue  gas
metals were highest for the three volatile metals, with lead at 96  percent,
cadmium at 90 percent and bismuth at 84 percent.  For the nonvolatile metals,
the fraction of flue gas metals exiting as particulate ranged from  82 percent
for copper to 31 percent for barium.

     None of the test variables had a clear effect on the split of  total flue
gas metals between the two phases.  Likewise, effects of test variables  on the
flue gas phase distribution for the individual metals were not significant
within data variability (4).

MASS BALANCE CLOSURE

      Average mass balance closures achieved around the kiln ash and scrubber
discharges ranged from 48 percent for strontium to 96 percent for cadmium.
Overall average closure was 71 percent.  Typical trace metal mass balance
closure results from past experience with combustion sources are in the 30 to
200 percent range.

                                  CONCLUSIONS


      In the subject test series, the trace metals cadmium, lead and bismuth
were found to be relatively volatile, while barium, copper, strontium,
arsenic, chromium and magnesium were  relatively nonvolatile.  With  the
exception of arsenic, their observed volatilities generally agree with
theoretical predictions based on volatility temperature.  A correlation also
                                      21

-------
existed between venturi/packed column scrubber efficiencies  and metal
volatility; average apparent scrubber efficiency for the metals decreased  with
increasing metal volatility.

      Both kiln ash partitioning and scrubber efficiency appeared  to  be
impacted negatively by increases in feed chlorine content and, to  a lesser
extent, increases in kiln temperature.  Neither was  significantly  affected by
changes in afterburner temperature.

      In the flue gas, particulate-phase metals predominated over  those  in the
vapor phase.  Average particulate-phase metals, as a fraction of total flue
gas metals, were highest for the volatile metals.  No clear  effects of the
three test variables on flue gas phase distributions were apparent.

                                  REFERENCES


1.    Oppelt, E.T.,  "Incineration of Hazardous Waste: A Critical  Review",
      JAPCA 37: 558, 1987.

2.    Acurex  Corp.,  "Test Plan for Evaluating the  Fate of  Trace  Metals  in
      Rotary Kiln Incineration with Venturi  Scrubber/Packed  Tower  Scrubber
      Particulate/Acid Gas Control"; U.S. EPA Contract No. 68-03-3267;
      Cincinnati; July, 1988.

3.    Skinner, J.H. and  G.J. Carroll, "Hazardous Waste Incineration: Status
      and Direction", Int'l. Conference on Incineration of Hazardous,
      Radioactive and Mixed Wastes; San Francisco; May, 1988.

4.    Fournier, D.J. Jr. and L.R. Waterland, "Pilot-scale Evaluation  of  the
      Fate of Trace Metals in a Rotary Kiln  Incinerator with a Venturi
      Scrubber/ Packed Column Scrubber - Draft", U.S. EPA Contract 68-C9-0038,
      Cincinnati; October, 1989.
                                      22

-------
       SORPTION AND DESORPTION OF POHCs AND PICs IN A FULL-SCALE BOILER
                                 UNDER SOOTING CONDITIONS

                     by:  Gary D. Hinsbaw and Scott W. Klamm
                          Midwest Research Institute
                          Kansas City, Missouri 64110

                          George L. Huffman and Philip C. L. Lin
                          U.S. EPA Risk Reduction Engineering Laboratory
                          Cincinnati, Ohio 45268

                                            ABSTRACT

     Recent EPA test results from boilers cofiring hazardous waste have indicated that unbumed organics can
continue to be released in stack gases after cofiring ceases, an effect which has been termed hysteresis. In those
tests it appeared that both principal organic hazardous constituents (POHCs) and products of incomplete combus-
tion (PICs) can be sorbed on soot deposits during cofiring and then desorbed after cofiring ends, but the evidence
for this hysteresis effect was limited and the impact on determination of destruction and removal efficiency (DRE)
was not known. The current study was designed to evaluate sorption and desorption of organics on combustion-
generated soot in a full-scale boiler during cofiring,  to determine unequivocally if hysteresis exists and, if so, to
evaluate its effect on DRE measurement The results will provide guidance on the RCRA trial burn protocol for
boilers.  A two-week test involved a series  of boiler operating conditions including background measurements,
cofiring of a simulated waste, soot generation, and subsequent waste-free operation to look for organics desorp-
tion. Sampling included stack testing for volatile and semivolatile organics and continuous emission monitoring
of stack gases, as well as collection of soot at several locations. Preliminary conclusions are:  (1) the hysteresis
effect is real and is measurable under certain conditions; (2) hysteresis can be observed both in the gas phase
(stack emissions) and in the solid, sorbent phase (soot); and (3) the extent  of hysteresis is compound-specific,
apparently relating to physical and chemical properties such as volatility or molecular size.  Assessing the full
impact of hysteresis upon DRE quantitation will require further evaluation. It appears possible that a DRE could
be measurably affected under some conditions, but it is also likely that the effect would be much less than an order
of magnitude. The duration of the trial bum and the duration of waste cofiring which precedes it are important
factors in DRE measurement where soot is generated.

     This paper has been reviewed in accordance with the U.S. Environmental Protection Agency's peer and
administrative review policies and approved for presentation and publication.

                                         INTRODUCTION

     EPA test results have demonstrated that industrial boilers can provide adequate thermal environments for
hazardous waste destruction, achieving an average destruction and removal efficiency (DRE) of 99.998% for
RCRA-listed toxic organics, i.e., principal organic hazardous constituents  (POHCs) (1). Results from recent test
programs, however, have shown that unbumed POHCs and products of incomplete combustion (PICs) can con-
tinue to be released in the stack gases after waste cofiring ceases (2), an effect which has been termed hysteresis.
In some cases, the levels of organics peaked and then tailed off for a period.  Soot blowing e?:periments indicated
                                               23

-------
that with cleaner boiler tube surfaces, such hysteresis effect was greatly reduced. Thus it appeared that some of
the POHCs and PICs can be adsorbed on soot deposited on boiler surfaces during coining and desorbed back into
the combustion gases after waste cofiring ceases. There was some limited evidence supporting this hysteresis
effect, but the impact on the determination of DRE for POHCs was not known, A full-scale test to determine
unequivocally if hysteresis exists and, if so, to evaluate its effect on DRE measurements was therefore warranted.
Such a test was performed under EPA Office of Toxic Substances (OTS) Contract No. 68-02-4252, Work
Assignment 14.

     The objective of this study was to evaluate the sorption and desorption of organic compounds on
combustion-generated soot during the cofiring of hazardous organics with fuel oil in a full-scale boiler. Testing
was performed by firing a watertube package boiler at very low load, under a "smoking" condition, while spiking
moderate levels of hazardous organic compounds into the fuel

     The test was not designed to enable a quantitative redefinition of DRE for baiters, but was instead intended
as a "range-finding" experiment to determine if soot sorption and desorption could be a major factor in organic
emissions relative to trial bums under certain conditions. Test results will primarily be used to provide guidance
in the design ofRCRA trial burns for boilers. Evaluation of the test results should allow development of recom-
mended guidelines on:

      *  Duration of waste firing at test conditions  (including waste firing and POHC input) prior to initiating a
        trial bum, to ensure "steady-state" POHC and PIC emissions,
      «  Duration of trial bum necessary to accurately measure long-term DRE performance.
      *  Sampling and process monitoring procedures that account for soot sorption.

                                        TEST PROTOCOL

FACILITY DESCRIPTION

     Testing was conducted during September 1989 at the Tampela/Keeler (formerly Keeler/Dorr-Oliver) boiler
manufacturing plant in WUIiamiport, Pennsylvania. A plant boiler was selected on the basis of its ability to five
either nitural gas or fuel oil, its seasonal availability, and because of the cooperation and support available from
the Tumpcla/Keelcr personnel.

     The test unit was t Model DS-17,5 boiler used for plant beating and test purposes, with a capacity of
17,500 Ib/h (2,21 kg/i) of steam and a furnace volume of B 415 ft3 (11,8 m1) (see Figure 1), The unit is fired by
to lir-itomlzing oil burner (8:1 turndown ratio) which can accommodate several fuel types, including natural gas,
No, 2 fuel oil, tod No, 6 fuel oil. Stack gases are monitored by oxygen (Oa) and nitrogen oxide (NO,,) analyzers
and in opacity meter, The facility is not equipped with any air pollution control devices, and none t& required,

     Prior to tenting, Use boiler was cleaned and ns much residual soot as possible was removed from all acces-
sible Interior surf ices, An estimated 60% of the boiler tube surface was cleaned, the remainder being physically
unrcichable,

EXPERIMENTAL DESIGN

Test Matrix

     A variety of parameters can affect the sorption and desorption of POHCs and PICs, including boiler operat-
ing conditions (e.g., load, fuel type and firing rate, waste cofiring rate, burner turndown, excess air, combustion
chamber temperature, and flue gas composition), waste characteristics (e.g., chlorine concentration and chemical


                                              14

-------
                                                                      Stack
                                                                        Boiler Outlet
                                                                        Soot Sampling
                                                                        Pens
                        l"b o o o o o o~"o""o"o""d~6""<£"6""o""o"'o o
                           -
                                                                              Burner
                           Rgurel. Keeler watertube package boiler: top view.

composition), and POHC characteristics (e,g,, sorptive characteristics, thermal stability, and behavior relative to
PICs).  Although it was not possible to address every parameter individually, this test was designed to create
conditions that were expected to allow the greatest sorption effects and then to measure those effects. During the
test, operation of the boiler was deliberately altered to produce a poor combustion mode (i.e., low load and low
excess air), and also nonoptimum air distribution within the burner, in order to deliberately produce soot.

     Both monochlorobenzene (MCB) and trichloroethylene (TCE) were selected as POHCs on the basis of their
chemical structure, thermal stability, adsorption characteristics, analytical detectability, and availability. A
separate pump and metering system was set up to spike each POHC directly into the fuel oil i'eed line. The two
POHCs and the fuel oil, which simulated a hazardous waste, were then pumped through an in-line mixer and
injected into the burner.

     The test program covered six different boiler operating conditions. These conditions and the associated
sampling and analysis activities are summarized in Table 1. The first two conditions consisted of short back-
ground tests in which the boiler was fired with fuel  only, either (1) natural gas or (2) No. 6 fuel oil.

     Test condition 3 used POHCs cofired with No. 6 fiiel oil under nonsooting combustion conditions. POHCs
for the test series were each to be fired at 5% of the fuel oil feed rate.  The test period involved one day (24 h) of
continuous boiler operation with periodic sampling activities.

     Test condition 4 involved cofiring the two POHCs with fuel oil under soot-producing conditions which were
determined on-site. Excess air was lowered to induce soot generation. Both stack opacity readings and visual
observations through boiler sight glasses confirmed the generation of soot.  The test period covered two contin-
uous days of boiler operation with periodic sampling activities.

     To initiate test condition 5, the fuel was switched from fuel oil to natural gas, and POHC cofiring was
ceased. The combustion furnace exit temperature was maintained at die same level as in test condition 4. Under
this condition the hysteresis effects of the boiler were measured for three continuous days. The test duration was
based partly upon organic levels indicated by field GC (i.e., using near-"real time" feedback).

     Test condition 6 involved a final collection of soot samples after the boiler was allowed to cool. Soot
sampling shields and soot scrapings taken directly from the boiler tubes were collected at this time.
                                                25

-------
                                        TABLE 1. EXPERIMENTAL DESIGN AND SAMPLING MATOIX
Run No/tot condition
                        Run length
                                                    VOST
                                                                      Sampling methods* and frequencies
                                                                         MM5
                                                                                             Soot
                                                                                                                  Peals
                                                                                                    CSM/QC
 I.    Natural gas
      (background)

2.    No. <5 fuel oil
      (background)
4.
No. 6 fuel oil
wltfiPOHCs
co filing
(nonsootlng)

No. 6 fuel oil
whhPOHCs
eofirtag (sowing)
5.
     Niton] gas, no
     POHCs (hysteresis)
6.   Cold ftimace
                             4h
                       6h
                             24 h
                             (continuous)
51 h
(continuous)
(3 hut Ugh
eofltlng
rate. 4 h
with no
corning,
and 44 hat
reduced
eoflring
rate)

73 h
(continuous)
                             After fur-
                             nace- cooling
                 One 40-mln sample
                 coltoctod hourly

                 One 40-mhi sample
                 collected hourly
                 One 40-mln sample
                 collected every
                 other hour
One 40-mln sample
collected hourly
                                        One 40-mln sample
                                        collected hourly
                                        until field GC
                                        leadings indi-
                                        cated "Wl off';
                                        then collected
                                        with diminishing
                                        frequency to end
                                        of run
                                                              One 3-h samp le
                                                              staitingZh
                                                              Into run
One 8-h sample
starting at
beginning of
run

One 3-h sample
starting at
beginning of
run (termi-
nated when
coflring
interrupted)
and one 8-h
sample start-
ing 23 h Into
ran

Three 8-h
samples
starting at
beginning.
20h,and44h
into run,
respectively
                                         3seU(l set each
                                         from furnace nod
                                         both duo loca-
                                         tion*) collector!
                                         removed from
                                         probes 3 h into
                                         run; probe*
                                         returned to
                                         sampling position
3 sets collectors
removed 28 h Into
run; 1 set col-
lectors removed
from boiler Inlet
probe 48 h Into
run (following
probe overheating)
                                         2 groups of 3 sets
                                         collector* removed
                                         4 hand 28 h Into
                                         run, respectively
                                                                                ISsootihklds
                                                                                removed front
                                                                                boiler tubes and
                                                                                9 scrap Ings
                                                                                collected
                                                                                                                                 Continuous
                                          One grab sample     Continuous
                                               l oil
                                                                One grab sample
                                                                of fuel oil and
                                                                enchPOHC
                                                                                                             One grab sample
                                                                                                             of rcieloll and
                                                                                                             each POHC for
                                                                                                             each day of
                                                                                                             sampling
                                                                                                                                Continuous
                                                                                                                                Continuous
                                                                                                                                Continuous
•CEM m Continuous emission monitoring for CO, COr O2, and THC (Methods 10,3A. 3A, and 25A, respectively).
 GC «• Held cryofentc trap/OC/HD for POHCj (monochlocobenzene and trichloroethytene) and selected PIC*.
 VOST <• VotstJlo organic sampling train for volatile POHCs end PICs (Method 0030). Two-thiids of sample pairs were combined for
   analysis, while one-third underwent separate trap analyses (Methods 5030,5040, and 8240).
 MMJ - Modified Method 5 for scmivolatilo PICs (Method 0010), analyzed by Method 8270.
 Soot« Thrco types of samples:
    t. Soot sampling water-cooled probes with clsmp-on collectors. Three probes simultaneously collected soot at two sampling locations (fttrnace and flue).
       Each "set" of samples was three separate collectors from a probe. Collectors were quickly removed from the withdrawn prob: and carefully stored until
       analysts.
    2. Shields were clamped onto existing boiler tubes prior to die test for collection of soot throughout the entire test series.
    3. Scrapings of soot from the existing boiler tubes were collected following the hysteresis testing.
3pDeds m One fuel oil sample is to be analyzed, and additional samples will be archived. POHC samples will be archived.
                                                                  26

-------
     Practical constraints limited the amount of soot formed. Since the test boiler was not equipped with
pollution control devices, state air quality standards limited the extent of soot allowable due to :i 20% stack
opacity limit.

Specialized Sampling and Analysis Techniques

     This project used several specialized, nonstandard sampling and analysis techniques. Dismissed below are
the techniques for soot sampling and analysis and the field GC used for on-site POHC monitoring.

     Soot samples were collected by three methods: (1) soot collection probes designed to collect soot in the
furnace and at the boiler outlet, (2) soot collection shields attached to boiler tubes, and (3) scrapings of soot from
boiler tubes. Soot collection probes and shields were placed in position before the test began. The probes were
withdrawn at specified times for removal of soot samples (during Runs 2,4, and 5). Following all test activities,
soot shields were removed for sample collection, and boiler tube scrapings from  the same locations as the shields
were also collected.

     Three soot collection probes were used to collect soot and to withdraw cumulative samples at various times
(see Figure 2). The probes were built with a double shell to allow ethylene glycol coolant to flow between.  In this
fashion, the outer wall of the probe was kept relatively cool (S 300°F) in the hot boiler environment to thus
simulate actual boiler tube surfaces.
       mm vm mm mmWM \\mm& &*mm K^\ km ww

                                                                                               VaJw
                                                                                            I Coolant
                                               PROBE A       Kwesa                           i-
                                            (Furnace Location)
                                     Figure 2. Soot probe schematic.

     Each probe was fitted with a series of carbon steel sample collection rings ("collectors") which were clamped
directly onto the probe outer shell. Volatile as well as semivolatile organic compounds adsorbed onto the soot
were thus "trapped" by the relatively cool surface arid were thermally desorbed during subsequent analysis.
Thermocouples monitored probe surface and combustion gas temperatures. When a probe was withdrawn for
taking samples, a predetermined set of collectors was retrieved as quickly as possible. Storage procedures for the
soot samples were geared toward reducing loss of organics and minimizing outside contamination.

     Eighteen soot collection shields,  similar in size and material to the probe collectors, were mounted to
selected boiler tubes in three locations: (1) along the furnace wall opposite the flame, (2) near the furnace exit,
and (3) near the boiler exit.  The shields collected soot throughout the entire test sequence. After the boiler
cooled, the shields were retrieved, and in addition, soot was scraped from the boiler tubes in the: same locations as
the shields.  Storage procedures for these two sample types paralleled those of the probe samples.

     Three methods for the analysis of organic compounds in soot were investigated, including thermal
desorption-gas chromatography/mass spectrometry (GC/MS), solvent extraction-GC/MS, and supercritical fluid
extraction-GC. Thermal desorption-GC/MS was selected for test sample analysis on the basis of spike recoveries,
detection limit, reproducibility, and sample size requirements. Using this technique, the soot samples were heated


                                                 27

-------
baUistieally in the interface chamber at the Met of a GC/MS, Any volatile organic species thereby released were
subsequently analyzed by conventional GC/MS methods. The effect of desorption temperature on POHC yield
from soot was evaluated. The yield of POHCs and several hydrocarbon species observed during the soot analysis
was highly temperature-dependent, and a desorption temperature of900°C was determined to be the optimum for
desorption of POHCs. These results indicate that POHCs (and other species) incorporated into the soot during the
actual soot formation are tightly bound to the soot matrix.  Despite the high desorption temperatures used, no
POHC decomposition was observed. Method detection limits for the thermal desorption-GC/MS method were
shown to be 500 pg/mg (500 ppb) and 20 pg/mg (20 ppb) of POHCs in soot for full-scan and selected-ion-
monitoring modes, respectively. Method precision was determined during the analysis of the calibration curves
for each POHC by each GC/MS method and found to be better than ±20% (relative standard deviation) in all
cases. Additional details on the analysis method and its development are available in a separate method
development report (3).

     The field gas chromatograph (GC) used to monitor organic compounds covers a boiling point range roughly
equivalent to those of Cj to C17 hydrocarbons (-160°C to 300°C). It was used as an indicator to monitor boiler
outlet organic levels, especially during the hysteresis "tail-off* phase, and to assist in coordinating the other sam-
pling methods. The system allowed semicontinuous monitoring of selected organic compounds collected continu-
ously, via their concentration in a cryogenic trap and subsequent analysis by GC/flame ionization detection
(GC/HD).

     The sample to the field GC was split off from a Teflon® heated line and maintained at 250°C, passed
through a Perma Pure® dryer, and then injected into a liquid CO2-based cryogenic cooling loop. The cooling
loop acted as a concentrator to enhance sensitivity, resulting in an integrated sample which could be injected peri-
odically (roughly hourly) into the GC/MD, The GC was fitted with a megabore column and temperature-
programmed to improve separation of target organic compounds. The HD was adjusted to scan for the two test
POHCs and selected PICs.

Technical Considerations

     Major practical limitations affected the establishment of priorities and data end use. In particular, known,
validated techniques did not exist for the collection, preservation, and analysis of soot for this study, although the
best procedures possible and immediately available were used to increase the likelihood of generating useful data.

     Key limitations relating to the usefulness of the soot results are:

     *  It was not possible to quantitate the rate of soot generation in the test boiler because the probe-based
        sampling technique did not necessarily provide for representative sampling of the entire gaseous stream
        within the combustion chamber or boiler exit Accordingly, the amount of total POHCs or PICs in the
        combustion gases that were trapped in the soot could not be quantitated.  Therefore, a mass "balance" of
        the partitioning of organics between the solid phase (soot) and the gaseous phase (stack gases) was not
        possible.
     «  Some of the organics in the soot may have escaped prior to analysis. It was not possible to determine the
        degree of sample integrity, although it was controlled as well as possible. Furthermore, the efficiency of
        desorption of organics from soot for subsequent analysis was not known, and measurement of true recov-
        eries was not possible. Nevertheless, the results are quite useful on a relative basis.

     Considering these and other limitations, the soot testing portion was intended to provide more qualitative
data than quantitative. An objective was to look at trends over time (sorption and desorptiou),  both in the soot
samples and in the stack samples. Comparison of results for each phase (i.e., solid and gas) assisted in evaluating
the role that soot on boiler tubes plays in the hysteresis effect.
                                               28

-------
                                              RESULTS
PROCESS CONDITIONS
     The process operating conditions and stack gas levels for each test period are summarised in Tables 2 and 3,
respectively. The combustion temperature (furnace exit temperature) was low, only 5 900° to 1100°F, compared
with typical trial bum temperatures of 1600°F or more. Although the boiler operation is not normally controlled
by furnace temperature, the temperature was maintained as constant as possible.

     POHC cofiring began in Run 3 at the planned rate, about 5% by volume for each POHC. The rates were
reduced during Run 4 due to extremely high levels of acid gases and organics believed to be present, and because
of the potential for boiler corrosion and possible YOST analysis problems, respectively. The changes in cofiring
rates complicate data interpretation for organics results.

     The stack flow rate ranged from 1460 to 1740 acf/min (840 to 1080 dscf/min), as measured during MM5
sampling. The stack opacity levels were generally maintained under 20% by carefully controlling excess air.
Table 3 summarizes the stack levels of gases as measured by CEMs.
                                   TABLE 2. KEY PROCESS CONDITIONS

Appro*.
Run No./ duration
Ten condition (h)
1. Natural sa» 4
(background)
2. Fuel oil 6
(background)
3, POHCeoflrini 24
(neiuooting)
4, POHCeoflrini 31
(looting)
S, Niturdgu 73
(hyiMreili)
Nominal POHC codling
Fumaco item Fuel POHC rate-volumMrlc
temp.* load* flow feed rate (Ib/h) baala (% total) Opacity
(°F) (Ib/h) rale1 TCB
990.1130 3000 39 oft 0
(N,a.)
963.1080 3000 38gph 0

933-1007 3000 28gph 16.7

936-lOfll 2700 Mgph 1&S/B.I*

910.1070 2500 40ofh 0
(N.O.)
MCB TCB MCB Total (%}
0 0000*

0 0000*

124 4,3 4,4 8.1 6-1

I3.2/M* 4J/2J* 4,W»* 9.W.2* 11-22

0 0 0 0 0-10


   'Monitored it mot probe optlttomd near bolter Intel. Notoi Hi* duo gai temperature, monitored at location of bolter outlet toot
    probei, ranged from 3SO  lo 376°P.
   ^Nominal itoim pmiur* wu 100 pilg,
   •No. 6 fuel oil, union noted "N.Q," for natural gii,
   *Sllght bMolInc drift oblorvad,
   'initially, cofiring rotes for Run 4 continued el the nme rated at during Run 3, After about 3 h, eoflrlng wi* ilopped for about 4 h,
    then returned it about 30% of the prevloui rota, which continued for 44 h,

-------
                          TABLE 3. CONTINUOUS EMISSION MONITORING RESULTS
                                         Average stock concentration (range shown to parentheses)*'
Run No.
t
2
3
4
5
02(%)
5,0(4.1-5.8)
5.6(5.1-8.2)
4.8(3.1-7.8)
2.2(1.7-8.6}
2.8(0.1-11.9)
C02(%)
8.6 (6.5-8.8)
11 3 (9.3-12J)
12.4(11.0-12.8)
14,3 (8.9-15.1)
10.1 (1.4-14.9)
CO (ppm)
10(0-23)
64(38-127)
97 (52-180)
626(91-8071)
23(2-828)
THC(ppm)
2(0-3)
2(1-4)
2(0-3)
13 (1-293)
1(0-26)

*AH concentrations reported on a dry basis, ss measured.
     The oxygen (O-j) level, ie., excess air, was reduced in Run 4 to deliberately generate soot Since the oxygen
level was manually controlled to try to maintain relatively constant stack opacity, more variation is seen in Run 4.

     Carbon monoxide (CO) levels were reasonably low in Runs 1,2, and 3, but were very high in Run 4, with a
maximum of over 8000 ppm. CO was initially high in Run 5, but quickly stabilized at less than 25 ppm. The total
hydrocarbon (THC) trends are somewhat similar to those of CO. THC levels were very low and varied little,
except during Run 4.

STACK EMISSIONS

POHC Results

     Stack gas levels of the two POHCs are shown in Figure 3 for all of the test conditions. (The inset is a
"blow-up" of the hysteresis portion of the graph.) The TCE concentration was high in the initial natural gas
baseline condition (Run 1), but it decreased substantially by the end of Run 2. It is possible that the boiler was
initially contaminated with TCE but that the first few hours of operation essentially "burned out" the contamina-
tion. The background levels of MCB were relatively  low.

     'When cofiring began, the POHC levels generally continued to climb during Run 3 and did not level out The
sooting condition (Run 4) began with a marked increase in stack gas POHC levels, but the levels decreased with
the reduction in cofiring rate. 'With time, the levels generally dropped, climbed, and dropped again before the
boiler feed was switched to natural gas for the hysteresis period. Some of this variation may be due to the gradual
buildup of a soot layer which sorbed more and more organics from the gas stream. However, some of the more
abrupt changes in concentration (Le., "sawtooth" patterns) may be an artifact of the analysis procedures.

     Throughout Runs 3 and 4, the higher level POHC switched back and forth between TCE and MCB,
indicating similar but varying levels of destruction and removal.
                                               30

-------
 CD


 O
 o
 O
 Baseline
Conditions
(Runs 1-2)
                             10      30       50      70      90      110     130     150
                                   Time After Start of Cofiring (h)
      	  Trichloroethylene                        	  Monochiorobenzene

                         Figure 3. POHC stack gas concentrations from VOST,

     The POHC levels rapidly dropped upon initiation of the hysteresis condition (Run 5), Levels generally con-
tinued to drop, then stabilized somewhat after roughly 10 hours. (The initial two high TCE peaks may be analysis
artifacts. Further planned evaluation of the VOST and field GC results may help to clarify.) Later in the run, the
TCE levels increased, reaching a maximum roughly 33 hours after the start of "hysteresis." Interestingly, the
MCB levels increased and maximized even later, but since this increase accompanied another "peak" of TCE, this
may be due to some Mnd of perturbation in the boiler.

     At the end of the hysteresis period, soot was deliberately "blown" from the boiler batik by manually injecting
steam in several pulses, to learn  if the POHC and PIC levels would increase. This soot blowing episode, which
lasted about 10 minutes, resulted in a minor rise of TCE, but no MCB was detected. Thus, there was not a
dramatic rise in POHCs, although much of the sorbed POHCs may have already "bled off' during the 3 days alter
cofiring ended.

     Destruction and removal efficiencies (DREs) were determined to further evaluate combustion conditions and
to better relate other test results to trial bum practices. As summarized in Table 4, DREs were generally low, only
three to four "nines." Interestingly, the averages were about the same for either POHC, indicating that the actual
"incinerabilities" ranked about equally under these operating conditions. DREs were higher in Run 3, as expected,
since oxygen had not yet been lowered. DREs were higher in the first part of Run 4, compared to the second part.
This was expected, since the cofiring rate was higher, considering the generally observed trend of increasing DRE
with increasing POHC concentration (4).
                                             31

-------
                                  TABLE 4. DMB» DURING POHCCOF1UNG

Run
No.
3
4*
4b
No. of
sample Destruction/removal efficiencies (DREs): average and range (%)
Test condition pain Trichloroethylene (TCE)
POHC 11 99.9947(99.9923-99.9975)
coflrlng/nonsooting
POHC 3 99.9904(99.987-99.9945)
cofiring/ioctlng"
POHC 27 99.986 (99.951-99.9972)
eoflring/Booting*
Monochlorobenzene (MCB)
99.9943 $9.9913-99.9974)
99.9913 (99.989-99.9929)
99.987 (99.970-99.9972)

Note: The number of DRB determinations was the same as the number of VOST sample pain for each teat condition. Bach DRE w«a based
upon the POHC output rate from one VOST pair and the POHC input rate from the closest matching time period with feed data.

'initial coflring rate nominally the same as Run 3.
'Adjuilcd coflring rate to approximately 50% that of Runs 3 and 4a.

     These results show that the test was conducted on the "tagged edge" of four cones DRE, which is where any
liysteresis effect would be expected to matter the most from a compliance viewpoint.

     Hysteresis effect quantitation much depends upon the conditions in which the measurements were made as
well as other assumptions. The test design and execution may have biased the rate of organics soiption and
desorption and also the approach for measuring hysteresis. Table 5 shows the hysteresis effect, both in terms of
the total mass emitted and also on a rate basis, which takes into account sampling duration. The mass values ate
an integration of the "area under the curve" for mass emission rates versus time. Calculating the masses of
POHCs for each test condition served as a means to assess the hysteresis effect.  The background amounts in Runs
1 and 2 can be compared against those in Runs 3,4, and 5. As discussed above, the TCE levels were high in Run
1, but die TCE appeared to be largely "purged out of the system" before Run 2 began.

     Assuming no sorbents for the POHCs were present prior to the deliberate soot generation in Run 4, then
hysteresis involved only what was emitted in Run S as compared to the amount emitted in Run 4, The ratio of die
mass emitted during hysteresis to  the mass emitted during soot formation was calculated as 10.5% for TCE and
5.5% for MCB; however, it must be emphasized that this test was conducted as a "worst case" type experiment
The real interest is in what would happen in a typical trial burn period of about 3 to 6 hours. Continued data
evaluation will focus on extrapolating results to trial bum situations.

     Several key questions that relate to an assessment of the hysteresis effect will be further evaluated:

     *  Had the soot reached a terminal thickness? (As soot builds up on a "cool" boiler tube surface, it
        gradually will decrease the heat transfer between the hot gases and the much cooler boiler tube. It is
        expected that the outermost layer of soot particles will eventually become so hot that no more soot will
        tend to collect, hence, a "terminal" thickness will be reached.)
     •  Had the sorption and desorption processes reached "steady-state" by the end of Run 4 (i.e., was there still
        a net decrease of POHCs due to sorption, or had the soot reached saturation capacity)?
     *  Whit  effect did varying chemical and physical characteristics of the POHCs and PICs have on their
        sorption and desorption?


                                                 32

-------
                            TABLE 5. POHC MASS EMISSIONS BY TEST CONDITION
—
Run
No.
1
2
3
4aM
4b (
5
5
5:4«
Test condition
Natural gas (background)
Fuel oil (background)
POHC cofiring/nonsooting

POHC cofiring/sooting
Natural gas (hysteresis)
Soot blowing (hysteresis)
Mass during hysteresis
versus mass during sooting
Duration

-------
           900
     3-   800
     £.   700
      o
     I
      I
      o
     O
      0>
      D)
      S
     I
600
500
400
300
200
100
                     Run 2    Run 3   Run 4A   Run 4B  Run 5A   Run 5B  Run 5C
                 Total POHCsESsa Total Vol. PICsE^J Benzene cxa Total SemiVol. PICs
                     Figure 4. Average stack gas concentrations of POHCs and PICs.
    The average levels of the major PICs (other than benzene) are shown in Figure 5. The highest level volatile
PICs included benzene, 1,1-dichloroethylene (DCE), methylene chloride (MeCy, dichlorobenzenes (DCBz's, as
the total of 3 isomers), and toluene.  The major semivolatile PICs included naphthalene and phthalic anhydride.
As seen in both Figures 4 and 5, the semivolatile PICs had a lower ratio of Run 5 concentrations to Run 4 con-
centrations than did the volatile PICs.  This could mean either (a) the amount of semivolatiles "held up" was less,
or (b) assuming similar levels were sorted, they were released less completely in the 3-day sampling period.
IOU
120
5" 110
FJJ
Ct rtrt
1 UU
¥ 9°
|j 80
ir 70
Q>
« 60
c

X
X
x
X
X
X
X
X
X
X






~,
/
7-,



1
1
1
1

::::







S
s
\







^
%
X,\,
K X
>
-------
     Levels of some of the other PICs of significance are shown in Figure 6, including the volatile compounds
1,1,1-triehloroethane (1,1,1-TCA), chloroform (CHC13), and caibon tetrachloride (CCL^, and the semivolatfle
compounds phenantlirene, dibenzofuran, and benzaldehyde. Again, volatiles tended to predominate during Run 5.
1,1,1-Trichloroethane was found at high levels during the background runs and may have been a contaminant in
the boiler.
                    57
       CD


       O

       I
       *«*
       Q>
       O
       OJ
       CD
£X>
24
22
20
18
16
14
12
10
8
6
w
4
2
n

-
-
-
-
-
-
-
-
-


-























n
/
/












I




_
S
S
\
s
X
s
s
\
s
s
s
N
s
N























x
x
X
*
X






I
I
1
1
!
,

i
i
I?
i/
?/
!/












1










s
V
S
s
s

















X
X
X
X
X
X
X
X
X
x
X
x
X
X
X
X








7
/
/

/•
/
/
/
/•





1
1
:f
i
I
1


|












N
S
S




















_
x
X
X
y
X I
X ^
^ 1
xl
X ;
X |
X '•



7
/
",
',
'
/
/
/

,
/





I
\
I
\
ji
|
%
|







-<;
S
^
S
S
S
\
s

























^
-"'" i»B(#*»(Ma3W












                     Run 2    Run 3    Run4A   Run4B  RunSA  Run 5B  Run 5C
               1,1,1-TCA
                             CHCI3       CCI4
                                                 Ptiananthrene
                                                                  Dibenzofuran
                                                                                  Benzaldahyda
                         Figure 6.  Average stack gas concentrations of other PICs.

     Of the primary PICs discussed above, those displaying more apparent hysteresis effects (i.e., higher Run 5
concentrations as compared with Run 4 concentrations) included 1,1,1-triehloroethane, methylene chloride, carbon
tetrachloride, and benzaldehyde. Benzene, toluene, chloroform, naphthalene, phenantlirene, and dibenzofuran
displayed less effect, and 1,1-dichloroethylene, dichloroben/enes, and phthalic anhydride displayed the least
effect. In general, the POHCs showed more hysteresis effect than the PICs, and the volatile PICs showed more
effect than did the semivolatile PICs.

     The highest level PICs seen during soot blowing tended to be the more volatile compounds sampled by
VOST.
SOOT RESULTS

     Figure 7 shows the levels of POHCs found in the soot probe samples. Levels of POHCs in the background
samples from Run 2 were very low. (Only samples from the boiler inlet probe were analyzed since no soot was
expected to be present. It is likely that the small amount collected was a different type of material than was col-
lected during Runs 4 and 5.) TCB and MCB were found at similar levels in Run 4 samples, the sooting condition.

     During Run 4, the boiler inlet probe overheated (from nominally 300°F to over 7QO°F) between the 28-hour
and the 48-hour samples. The latter sample was collected to determine whether soot on that probe had built back
up organic levels, assuming that the overheating had caused a loss. It can be seen that the levels at the end of Run
4 did essentially match those of the first sample set collected in Run 4.

-------
   ~     40
   *-»"
"1-03
           TCE @ Inlet
               	>NotSmpl.
                INU3" ' OUTUfcl
                 4-48
   .._.i No. -ElapseaTime (h,
MCB ©Inlet   IBB  TCE ©Outlet
                                      28
INLET  - -
  5-04
MUJJ
  tD™,
           •28
CZD   MCB@ Outlet
                            Figure 7. POHC results from soot probe samples.

     Hie hysteresis effect can be evaluated from the soot samples as follows: for the boiler inlet, compare the
"4-48" sample results with the Run 5 sample results; whereas for the boiler outlet, compare the "4-28" sample
results with the Run 5 results. The Run 2 results can be misleading since a different type of material was likely
collected and much less of the material was present than in the Runs 4 and 5 samples. This applies also to the PIC
data, which are discussed below.

     Stack gas results indicated that more TCE was emitted during the hysteresis period than was MCB; in fact,
roughly double the amount was emitted (refer to Figure 3 and Table 5). The soot results show some trends con-
sistent with the stack gas results. Comparing Runs 4 and 5, the TCE levels dropped much more at the boiler inlet
than did the MCB levels.  Actually, MCB appears to be "held up" more in the soot, as further evidenced by the
marked "delayed" increase in MCB levels in the boiler outlet soot during the time period of Runs 4 and 5. TCE
also rose in the outlet soot, but not as much.  All of this behavior is consistent with the physical characteristics of
the two POHCs (e.g., volatility and molecular weight).

     The soot PIC data show similar trends, but actual concentrations cannot be compared because only relative
units were appropriate according to the analysis procedure.* (Refer to Figures 8 and 9; units are in mg"1.) Ana-
lytes included only a limited number of preselected PICs. For those PICs the following order was observed: in
order of decreasing relative concentration, toluene > benzene > naphthalene > xylenes (total of 3 isomers) >
dlchlorobenzenes (total of 3 isomers) > 1,1-diehloroethylene (DCE). With the exception of DCE, in comparing
*Bach PIC was qunntitated by computing the relative peak area for the PIC with respect to the peak area for the
 internal standard (IS), chlorobenzene-dj, and to the total soot mass. Iqother words, a unitless ratio of peak areas
 was divided by the mass of the soot aliquot which was analyzed. The units are mg"1.  This method of quantita-
 tion was chosen for three reasons.  First, the variability in the results precludes accurate absolute quanu'tation.
 Second, no standards for the PICs were used so no relative response information was available. Finally, the mass
 spectrometry data were collected in the selected-ion-monitoring mode (SIM), so that an assumption of a response
 factor of 1.0 relative  to the IS would likely be in error by as much as a factor of 20.
                                               36

-------
O)
I,  3

If3
•J3   3
to   «

C   d,
S  n o
C   P ^
o  J:
O  ^5
.1   o

g   1

g
Q.
                    Not And.
 INLET ' OUTLET
   2-03

Toluene i
Toluene <
                              INLET ' OUTLET
               INLET ' OUTLET
                        ! Inlet
                        (Outlet
   4-28         "4-48
  Run No. - Elapsed Time (h)
            Benzene @ Inlet
            Benzene @ Outlet
INLET ' OUTLET
                                                              5-04
INLET ' OUTLET
  5-28
         Naphthalene @ Inlet
         Naphthalene @ Outlet
                    Figure 8. Major PIC results from soot probe samples.
            .NO NalAnil
          INLET ' OUTU5T
            2-03

        Xylenes @ Inlet
        Xylenes @ Outlet
                           INLET ' OUTLET
                                            INLET ' OUTLET
4-28            4-48
  Run No. - Elapsed Time (h)
             DCBs® Inlet
             DCBa @ Outlet
                                INLET' OUTLET
    5-04
                                                  INLET' OUTLET
                                                                  5-28
               DCE @ Inlet
               DCE © Outlet
                     Figure 9, Other PIC results from soot probe samples.
                                          37

-------
 the levels found in Run 4 and Run S, all of the PICs generally displayed a time-dependent behavior of falling
 levels at the boiler inlet and rising levels at the boiler outlet. This is similar to what was seen in the POHCs,
 especially MCB. Of the PICs, toluene showed this hysteresis-type behavior very dearly, whereas the xylenes and
 naphthalene displayed the effect less clearly than the other PICS.  Interestingly, the dichlorobenzenes, which were
 present at much lower concentrations, also clearly showed the effect

     DCB, which was found only at trace levels, was selected for analysis because it was seen at high stack gas
 levels. The soot levels are probably low due to the compound's extremely high volatility and related properties.

                                          CONCLUSIONS

     At this point, only a limited set of conclusions are offered. Further data evaluation is ongoing, and additional
 conclusions may be possible in the final project report

     The test successfully accomplished the major objective to evaluate the sorption and desorption of organic
 compounds on combustion-generated soot during POHC cofiring.

     The primary conclusions are:

      *  The hysteresis effect is real and is measurable under certain conditions.
      »  Hysteresis can be observed both in the gas phase (stack emissions) and in the solid, sorbent phase (soot).
      *  The extent of hysteresis is compound-specific, apparently relating to physical and chemical properties
        such as volatility or molecular size.

     The exact impact of the hysteresis effect upon DRE quantisation is not yet known.  It appears possible that a
 DRE could be measurably affected under some conditions, but it also seems likely that the effect would be much
 less than an order of magnitude. To  some extent the degree to which soot is saturated with organics as well as the
 rates of POHC sorption and desorption will affect DRE measurement Also, the duration of the trial burn and the
 duration of waste cofiring which precedes it will be major factors.

                                           REFERENCES

 1.   Wool, M, Castaldini, C, and Lips, H.  Engineering assessment report: hazardous waste cofiring in
     industrial boilers under nonsteady operating conditions. Acurex Summary Report TR-86-103/ESD, U.S.
     Environmental Protection Agency, Risk Reduction Engineering Laboratory, Cincinnati, Ohio, My 19S9.

 2,   Mason H,B,,eiaJ.  Pilot-scale  boiler cofiring tests to investigate nonsteady effects. In: Proceedings of the
      14th Annual EPA Research Symposium on Land Disposal, Remedial Action, Incineration and Treatment of
     Hazardous Waste. Oncinnati, Ohio, May 9-11,1988, EPA/600/9-88/Q21, My 1988. pp. 332-345.

3.   Method development for the determination of principal organic hazardous constituents in soot.  Midwest
     Research Institute Summary Report, EPA Contract No. 68-02-4252, U.S. Environmental Protection Agency,
     Risk Reduction Engineering Laboratory, Cincinnati, Ohio, November 22, 1989.

4.   Trenholm, A., Gorman, P., and Jungclaus, G. Performance evaluation of full-scale hazardous waste
     incinerators, Vol. H: Incinerator performance results. EPA-600/2-84-181b, PB85-129518. U.S.
     Environmental Protection Agency, Cincinnati, Ohio, November 1984.
                                               38

-------
                        STATE OF THE ART ASSESSMENT
                OF MEDICAL WASTE INCINERATION TECHNOLOGY

            by:    R. G. Barton, G. R. Hassel, W. S. Lanier, W. R. Seeker
                  Energy and Environmental Research Corporation
                  18 Mason
                  Irvine, California 92718

                  C. C. Lee
                  Risk Reduction Engineering Laboratory
                  U.S. Environmental Protection Agency
                  Cincinnati, Ohio 45268

                                    ABSTRACT


    Incineration is being increasingly utilized as a method for disposing of medical waste since it
reduces both the waste's volume and potential hazard.  This paper is based upon an ongoing
study examining current practices in design and operation of medical waste incinerators with
emphasis on identifying incineration process parameters controlling potential toxic emissions.
Recent field test results have indicated that medical waste incinerators may be prone to emitting
high concentrations of acid gases, toxic organic compounds and other hazardous substances. A
variety of incinerator design and operating parameters including chamber temperatures, gas phase
mixing, waste feed rate and inclusion of air pollution control devices were  found  to have
important impacts on emissions. It was determined that many of the same mechanisms that are
responsible for toxic emissions from municipal and hazardous  waste incineration sys.tens .are also
responsible for emissions from medical waste incinerators.  However, medical waste incinerators
present unique challenges due to their size and the heterogeneity of the waste.

    This paper has been reviewed in accordance with U.S. Envkonmental Protection Agency's
peer and administrative review policies and approved for presentation and publication.
                                         39

-------
                                  INTRODUCTION


    Medical waste treatment and disposal is becoming an increasingly important issue.  The
EPA estimates as much as 14,000 ton/day of medical waste are produced in hospitals in the
United States (1). Additional waste is produced in research and diagnostic laboratories, nursing
homes, doctors offices and veterinary clinics.  Not only is this waste extremely voluminous, but
its infectious and sometimes radioactive nature waste requires careful management.  The
technologies most commonly used to treat medical waste in the United States include steam
sterilization, shredding/chemical disinfection, and incineration. Approximately 67 percent of all
hospital waste is incinerated on-site. Incineration can provide up to 95 percent waste volume
reduction, as well as effectively destroying infectious organisms in the waste.  With landfill
capacity rapidly diminishing and medical waste generation increasing, incineration is being
increasingly viewed as a viable medical waste management technique.

    While effective for volume/hazard reduction, incineration of medical wastes presents a unique
set of problems.  Destruction of hospital wastes by combustion results in formation .of air
pollutants in solid, liquid, condensible, and gaseous forms. This is one of the most sensitive of
air pollution problems, because not only is the general public exposed to these emissions, but the
greatest exposure is potentially sustained by that segment of the population which are the least
capable of withstanding any further stress to their health:  hospital patients.

                                    OBJECTIVES


     The objectives of the current study were to define specifically what  the most  critical
emissions problems are and to define the methods currently being used to deal with the problems.

                              INCINERATION SYSTEMS


    There are a wide range of medical waste incineration system types currently in use in the
United States. The three most common types are: retort incinerators, modular controlled air
systems, said rotary kilns. This paper will focus on these incinerator types.  Systems are generally
available in either batch or semi-continuous models.  Modular controlled air systems are the most
commonly manufactured systems in the United States. Figure 1 is a schematic illustration of a
modular controlled air incinerator. The primary zone  is operated under fuel rich conditions
(oxygen deficient) with additional air added at or near the entrance to the secondary chamber.
Waste is fed to the primary chamber, usually with a hydraulic ram, allowing the waste to burn on
one or more fixed hearths. Small systems typically have only one hearth while larger systems
may have 2 or 3 hearths. Small systems, even if they have continuous feed, typically do not
provide for continuous ash removal.  Such systems are generally designed to operate 8 to 10
hours per day. Large systems typically provide for continuous waste feed and ash removal.

    The principal control variables for modern controlled air incinerators are temperatures in the
primary and secondary chambers and the volume of those two chambers. Combustion air flow is
modulated to maintain primary temperature in the 1400-1600°F range while secondary chamber
temperature is controlled typically in the 1800-2000°F range. Auxiliary fuel fired burners are
provided to preheat the systems and to assure that secondary chamber temperatures are maintained
at regulatory dictated levels. The combination of substoichiometric air addition and large chamber
volume in the primary zone create low gas velocities in that region which permit these systems to
                                         40

-------
achieve low particulate emissions (typically less than 0.1 gr/dSCF) without air pollution control
zones.  The size of the secondary chambers is established by requirement of local regulations but
typically 1 or 2 second residence rime is provided in that zone.

                             POLLUTANT FORMATION


    The pollutants of concern are: radioactive materials, pathogens, cytotoxic compounds, toxic
metals, trace organics (e.g. PCDD/PCDF) and criteria pollutants (CO, NOS, SO2, HC1,  PM,
PM10).


RADIOACTIVE AND CYTOTOXIC MATERIALS

    Radioactive  and cytotoxic materials are only a problem when present in the waste. In
general, these materials are segregated and disposed of separately.

PATHOGENS

    The destruction of pathogens is one of the primary goals of medical waste incineration.
There exist some data that indicate that a properly designed and operated incinerator is capable of
completely destroying pathogens (2) (3),  These data were obtained by spiking the feed with a
particular bacteria spore and men testing the residuals and exhaust gases for spore activity. Only
in an incinerator operated at very  low temperatures, 1100°F,  was there found any residual
activity. While pathogen destruction is generally complete, an incinerator can be made to operate
with poor destruction under abnormal situations.  The phenomena likely to  affect pathogen
destruction include: waste moisture content,  uniformity of  combustion zone  conditions,
combustion zone temperature, residence times of solids in the combustion zone including both
solids on the bed  and particles entrained into the combustion gases, and excess air levels.

TOXIC METALS

    Metals are present in large amounts in hospital wastes.  Contaminated needles are the most
obvious source of metals. However, there are a number of less  obvious sources, ranging from
the dyes and inks used in printed matter to some pharmaceutical preparations.  It is these trace
sources of metals which represent the greatest potential danger to human health, for they generally
contain the most toxic metals.

    In Table 1 are data from a hospital waste incinerator in California (4).  This study indicated
that the fly ash was highly enriched relative to the bottom ash with lead, cadmium, chromium and
arsenic. Even  when equipped with a baghouse, emission of cadmium and lead were found to be
relatively high.  Because medical waste incinerators are small, these emissions may not
significantly increase the metals concentrations in the atmosphere. However, multipathway risk
analyses performed by California Air Resource Board indicate that the incremental lifetime cancer
risk associated with the cadmium emissions from medical waste incinerators can be as high as
500 chances in a million (5).

    A number of mechanisms control the behavior of metals during the incineration of hospital
wastes. These mechanisms can be identified by examining data available from a variety of
incineration systems and are discussed in detail by Barton et  al. (6).  The key phenomena,
illustrated in  Figure 2, are vaporization and subsequent condensation of volatile metals,
                                         41

-------
entrauunent of metals bearing solids, and reactions between metals and other elements (esp. Cl).
New insight into the importance of these mechanisms has been obtained from recent tests of
hazardous waste incinerators (7) (8).

    The important hospital waste incinerator operating parameters that influence metals emissions
can be identified using field data and an understanding of the controlling mechanisms. The most
important parameter is, of course, the amount and type of metals in the waste. As previously
discussed, the metals that are volatile at combustion conditions or are converted to volatile form
such as As,  Cd, Pb and Hg are particularly difficult to control. Other system parameters expected
to influence metal emissions include the following: the maximum temperatures of solids in the
primary zone, chlorine content of waste since chlorides of metals such as lead tetraehloride can
form salts which are extremely volatile, primary zone gas velocity which dictates particle
entrainment, the temperature of the particle control device since this determines how much of the
volatile metals have recondensed and can therefore be captured, and the fine particle control of the
APCD and parameters such as pressure drop which influence the control levels.

CHLORINATED DIOXM/FURANS

    One of the greatest challenges that remains for the disposal of medical waste by incineration
is the control of emissions of trace organics such as polychlorinated dibenzo(p)dioxin and furans
(PCDD/PCDF). There have now been several studies which have indicated that medical waste
incinerators can generate relatively Mgh levels of PCDD/PCDF. In Table  2 are provided data
from a wide variety of combustion sources, including three medical waste incinerators. These
data have an average total PCDD/PCDF level of approximately 700 ng/Nm3.  This compares to
large modern mass burn municipal waste incinerators such as Marion, Tulsa, and Wurzburg
which have average total PCDD/PCDF levels of less than 30 ng/Nm3 and   hazarous waste
incinerators which can range from 25 to 3000 ng/Nm3.

    Figure 3 illustrates the various potential mechanisms that have been proposed for formation
of PCDD/PCDF in incinerators.  One of the simplest mechanisms is that the emitted dioxins were
originally present in the waste and were not destroyed in the incinerator.  While this mechanism
may be responsible for a very small fraction of the emitted PCDD/PCDF  most wastes do not
contain sufficient quantities of PCDD/PCDF to account for the observed emission levels.

    In the second potential mechanism, dioxins are formed as intermediates in the oxidation of
more complex hydrocarbons. The hydrocarbons may be chlorinated (PVC  for example) or not
(cellulose).  If the dioxins are originally unchlorinated the chlorination must take place as a second
step.

    The third potential dioxin formation mechanism involves reactions between relatively simple
gas phase precursors such as phenols and chlorobenzenes. Shaub and  Tsang (9) developed a
kinetic model to study the characteristics of the proposed reactions.  Ballschmitter et al. (10) and
Benenfinati et al. (11) examined the  emissions from full  scale incinerators and found a close
relationship between the dioxin  emissions and the  quantity of  polychlprobiphenol and
polychlorpphenol in the exhaust. They suggested that this indicates that dioxins are formed  by
reactions involving PCBs and PCPs.

    The final mechanism that has been proposed calls for fly ash catalyzed formation of dioxins
in relatively cool regions of the incinerator.  This last mechanism was originally proposed  by
Vogg and Stieglitz (12)and is supported by laboratory experiments. Extensive experiments are
under way in Germany, Canada and the U.S. to extend Vogg and Stieglitz's work on full-scale
                                         42

-------
systems. Limited results available to date indicate that PCDD/PCDF emissions are closely related
to the quantity of entrained particles and the amount of gas phase hydrocarbons (13).

    Based on the mechanisms described above, it can be hypothesized that PCDD/PCDF
formation can be minimized by controlling particle emission levels within the incinerator,
minimizing the time particles are held at key temperatures (between 480°F and 666°F) and by
maximizing the destruction of precursors both vapor and particle bound within the incinerator.
Also dioxins can ultimately be removed from the flue gas through the use of fine particle control
since PCDD/PCDF will condense on particles at low temperatures.

CRITERIA POLLUTANTS

    The incineration of medical wastes can generate a variety of acid gases such as SC«2, SOs,
NOX, HC1 and HF. NOX can be formed by oxidation of the nitrogen in air and in the wastes. The
other gases are typically formed by the chemical reaction of sulfur, chlorine and other elements in
the waste. The most common occurrence is the formation of HC1, and thus in most hospital
waste incinerators, HC1 is the principal acid gas of interest. During  combustion, the organic
chloride in the waste reacts with hydrogen to form hydrogen chloride, HC1. The organic chloride
can be found in many substances commonly burned  in hospital waste incinerators including
plastic bags, disposable syringes and plastic tubing. It is possible for free chlorine to form when
the combustion chamber contains  an insufficient quantity of hydrogen to convert all of the organic
chlorine to HC1, but this is very uncommon.

AIR POLLUTION CONTROL DEVICES

    Until very recently medical waste incinerators were typically not equipped with air pollution
control devices. Prevailing particulate emissions standards could be achieved without APCDs;
CO emission levels were generally below 25 ppm and acid gas from small sources were not a
major concern for local regulators.  However, that situation is changing drastically. New York,
for example, is requiring that APCDs be retrofit to all existing medical waste incinerators. While
many types of advanced APCDs are available (e.g. wet electrostatic precipitatots and ionizing wet
scrubbers) two types of add-on APCD are receiving major attention.  These are illustrated in
Figures 4 and 5 and include (1) a venturi scrubber followed by a packed tower and (2) duct
injection of a calcium based sorbent followed by a fabric filter. The venturi scrubber achieves
particulate capture largely by  impacting the particles with water droplets.  The venturi
configuration is used to provide a high relative velocity between  the particles and water.
Increased relative velocity increases impaction  and particulate capture but is achieved at the
expense of high pressure drop.  Modern, venturi scrubbers provide 30 to 40 inches W.G. to
achieve exit particulate loadings  on the order of 0.03 gr/dSCF. Since HC1 and SOa are highly
soluble in water the combination of water sprays in the venturi and in the packed tower are
extremely effective in removing acid gases.  Typically, caustic is added to keep the scrubber
liquid pH in the 6.5 to 7.0 range. Total acid gas removal efficiency on the order of 98 to 99
percent is routinely achieved.

    Duct injection followed by fabric filtration is currently  finding increased  popularity—
particularly among regulators who have based rules for medical  waste incinerators on municipal
waste combustion experience. As with municipal waste combustion applications the fabric filter
is able to achieve particulate loading on the order of 0.015 gr/dSCF with pressure drops on the
order of 10 inches W.G. This system does not drop the flue gas temperature below the dew point
and must achieve acid gas capture with injected sorbent. Using about 2:1 stoichiometric calcium
to acid gas molar ratios, about 90 percent acid gas control is achievable.
                                         43

-------
                                     SUMMARY


    The pollutants of principal concern from medical waste incinerators include: volatile metals
such as As, Pb, Cd, Hg; highly toxic materials such as: PCDD/PCDF and hexavalent chromium;
HC1; and fine particles (PM10).  In the case of HC1, the waste chlorine level and the performance
of any air pollution control device scrubbers are dominant. Waste chlorine levels are highly
variable and not easily quantified. Thus the focus comes back to the proving that the scrubber
efficiency is sufficient to control any levels of HC1 emissions.  Since even short-term exposures
to HC1 are a measurable risk, the short-term control efficiency must be determined.

    For toxic metals,  the volatile species such as arsenic, lead and cadmium are of principal
concern because they escape the incinerator as a vapor and condense into an ultrafine fume that is
not easily  captured and is highly respirable. However, some metals are carcinogenic and even
low emission levels are important.  Again, the amount of these metals present in the hospital
waste is, of course, the dominant variable for control of their emission. However, it is virtually
impossible to either quantify the metals in the input waste stream or to control the levels due to the
trace quantities of concern and the ubiquitous nature of many of the metals such as chromium and
lead.  The key operating parameters are those that control the volatilization of metals (such as
temperature, excess air and chlorine levels) and those that control fine paniculate capture.

    Finally, the emissions of PCDD/PCDF are expected to  be the result of a complicated
interrelationship between waste properties, combustion conditions, and scrubber/fine particulate
control.  The primary control parameters were identified as - combustion uniformity, combustion
zone mean temperature, fine particle control efficiency, APCD temperature, and particle loading
exiting furnace (determined by incinerator load, velocities and waste properties).

                                   REFERENCES


 1.  Infectious Waste News, "EPA releases estimates on infectious wastes generation for  this
    week's meeting,"  November 17,1988.

 2.  Barba, P.,"Test results from bacterial sample burns from nine infectious waste incinerators,"
    APCA Mid-Atlantic States Section, Nov. 1987.

 3.  Allen, R.J. et al.,  "Bacterial emissions from incineration of hospital waste, final report,"
    ILENR/RE-AQ-88/17, Illinois Department of Energy and Natural Resources, July, 1988.

 4.   Jenkins, A., et al., "Evaluation test on a hospital refuse incinerator at Cedars-Sinai Medical
     Center, Los Angeles California," California Air Resources Board, ARB/SS-87-11, April,
     1987.

 5.   California Air Resources Board (GARB) "Draft dioxins and cadmium control measure for
     medical waste incinerators," Feb. 22,1990.

 6.   Barton, R.G., et al.,  "Prediction of the fate of toxic metals in hazardous waste incinerators,"
     Final Report for EPA Contract 68-03-3365,1988.
                                        44

-------
7.  Carroll, G. J. et al, "The partitioning of metals in rotary Mln incineration," In: Proceedings
    of the Third International Conference on New Frontiers of Hazardous Waste Management,
    Pittsburgh, Sept., 1989.
8.  Radian Corporation, "Draft test report: A performance test on a spray dryer, fabric filter,
    and wet scrubber system," Oct., 1989.
9.  Shaub, W. M. and W. Tsang, Environ. Sci. and Tech.. 17:721,1983.
10. BaUschmiter, K.» W. Zoller, C. Scholtz, A. Nottrodt, Chemosphere. 12:585,1983.
11. Benefenanti, E., F.  Gizzi, R. Reginato,  R. Fanelli, M. Lodi,  and R. Tagliaferri,
    Chemosphere. 12:1151, 1983.
12. Vogg, H. and L. Stieglitz, Chemosphere. 15:1373,1986.
13. Barton, R.G., et al., "Draft topical report: analysis of Quebec City incineration tests," EPA
    Contract 68-03-3365, April 1988.
                                       45

-------
     Waste
     Feed
                      Secondary
                      Chamber
Primary Chamber
                      Secondary
                      Burner
                 Primary
                 Burner

             Figure 1. Typical modular starved incinerator.
TABLE 1. METALS EMISSIONS FROM CEDARS-SINAI MEDICAL CENTER
        INCINERATOR
METAL
LEAD
IRON
MANGANESE
NICKEL
CADMIUM
CHROMIUM
ARSENIC
BAGHOUSE
(wj/gm)
18051
2401
175
30.7
1792
78
86
BAGHOUSE
CATCH
(ng/gm)
89
3100
240
33
1700
1
4.7
BOTTOM ASH
(na/gm)
1700
17400
230
33
2.8
0.7
0.7

-------
        REDUCMG
        ENWKWMENT
rCXVDUAL
PARTICLE
OR
DROPLET

DURNQ
COMBUSTION
                        SPRAY
                   Of LIQUD
                    WASTE
                                        VAPOR
                                                              HOMOGENEOUS
                                                              CONDENSATION
                   BURNING BED
                     OF SOLD
                      WASTE
CHLOROE8
 8ULF10CS
>XDE8,ET
                                                              FLY ASH
                                                                RESCUALS
           Figure 2. Phenomena affecting metals behavior during incineration.
         Formation From
         Precursors
   Incomplete  Destruction
   of Long Chain Organics
                                    Low  Temperature
                                    Catalyzed  Reactions
       PCDD/PCDF in Waste
                Figure 3.  Potential PCDD/PCDF formation mechanisms.
                                        47

-------
             TABLE 2. PCDD/PCDF EMISSIONS SUMMARY
Hospital Waste

Cedar Sinai
Saint Agnes
Royal Jubilee

hazardous Waste

Hazardous Waste Incinerators
                                    PCDD
                            ug/Mg         ng/Nm3
    1988
    6272
1625-2680
Hampton
North Andover
Marion Co.
Prince Edward Island
Tulsa
Wurzburg
Akron

Tier IV

Black Liquor
Wood Fired Boiler
Carbon Reg Furnace
Sewage Sludge
Drum & Barrel Reclamation
                                    PCDF
                            ug/Mg       ng/Nm3
160-260
290-450
117-197
                20-600
   5384
   10961
715-1115
                    2-17
                     102
                    28.8
                     114
                     687
386-700
700-785
  52-84
                                         5-2400
1000-27000

5 2
200-500
74.5
62.7
636
243-10700
225
1.13
60-125
18.9
22.1
258
1770-41200

1.9
300-500
61.1
79.2
1680
400-37500
«3/%*2
323

100-160
15,5
27.9
679
                           1.5-45
                             154
                            70.1
                             507
                            2107
                                  1000
                                                                         STACK
 Figure 4. Schematic of wet yenturi scrubber and packed tower for application
           to medical waste incinerators.
                                   41

-------
   ECONOMIZER OR
      GAS-TO-AIR
   HEAT EXCHANGER
         LIME INJECTION
                       FABRIC FILTER
                                            STACK
/     \
      - Injection of dry reagent
      , most often hydrated Nme
                                                                 Periodic bag pulsing  -
                                                                 removes collected
                                                                 paniculate; the system
                                                                 remains on-line
Reagent converts HCI
gas to dry, collectible
particulate
Recovers heat, reduces gas
temperature tor improved
reaction, and allows use of
lower-cost bags
            -Adjustable venturi throat
             accommodates changing
             gas flows, maintain even
             distribution of reagent

Counter-current injection provides
good distribution and entrapment
of reagent
RMOIonZon*
                                                                   Flyash. reacted gas,
                                                                   and unused reagent
                                                                   are removed continually
                                                                   in a dry state
•Bags: typical system contains
 40- 180 bags. Fabric selection
 Is based on cost and material's
 ability to withstand heat, acid,
 and abrasion
•Gas passes from outside of
 bag to inside

•Ash and reacted acid gas
 collect on outside of bag
                                                                                                        Internal system recirculates
                                                                                                        reagent for optimum use
                                                                                                        and effectiveness without
                                                                                                        material handling equipment
No visible
steam plume
                Figure 5. Schematic of duct injection fabric filter APCD for application to medical waste
                            incinerators.

-------
                THE INCINERATION OF ARSENIC-CONTAMINATED SOILS
             RELATED TO THE COMPREHENSIVE ENVIRONMENTAL RESPONSE.
                    COMPENSATION AND LIABILITY  ACT (CERCLA)
                  by:  Howard 0. Wall and Marta K. Richards
                       Risk Reduction Engineering Laboratory
                       U.S. Environmental Protection Agency
                       Cincinnati, OH  45268
                                   ABSTRACT


      The thermal fate of arsenic in CERCLA soils has been evaluated at the
USEPA Incineration Research Facility at Jefferson, Arkansas.  This facility
houses a pilot-scale rotary kiln incinerator which was fed an arsenic-
contaminated soil for the reported study.  The analytical results indicate
that as much as 75 percent of the measured arsenic fed to the kiln was
contained in the kiln ash; about 20 percent was removed by the air pollution
abatement system; and the other 5 percent was found in the particulate
emissions.  Preliminary muffle furnace tests conpleted on the same soils
indicated that both the volatility and Teachability of arsenic were reduced
when lime was added to the arsenic-containing soil before thermal treatment.

                                 INTRODUCTION


      The Incineration Research Facility (IRF) at Jefferson, Arkansas, has, as
a primary mission, the support of the Environmental Protection Agency (EPA)
Regional Offices and the Office of Solid Waste and Emergency Response in
evaluation of incineration as a disposal option for wastes which are
undergoing remedial actions at Superfund sites.  A priority site in Region I
is the Baird & McGuire location in Hoi brook, Massachusetts.  EPA Region I
requested test burns of contaminated soil from this location to support the
evaluation of the suitability of incineration as a treatment technology for
remediation of this Superfund site.  The soil contained the principal  metals
arsenic and lead and a wide variety of pesticides (pp'-DDT, pp'-DDD and
pp'-DDE and Methojo?chlor).  Incineration was proposed to destroy the organics,
but the fate of arsenic (or metals) was of concern.  The principal  mission of
the test was to evaluate the incinerator operating conditions, and  how they
would impact the partitioning of the arsenic to the various residual  streams.
The purpose of this presentation is to identify the arsenic content of the
discharge streams:  ash, scrubber water and atmospheric emissions.
                                      SO

-------
                                   EQUIPMENT
      The experimental work was conducted at the USEPA Incineration Research
Facility (1). The rotary kiln incineration system (RKS) was used for this
research.  The design characteristics of the kiln system are:
Length, outside
Diameter, outside
Length, inside
Diameter, inside
Chamber volume
Construction
Refractory
Rotation
Solids retention time
Burner

Primary fuel
Feed System
  Liquids
  Sludges
  Solids

Temperature (max)
Length, outside
Diameter, outside
Length, inside
Diameter, inside
Chamber volume
Construction
Refractory
Gas residence time
Burner
Primary fuel
Temperature
               Main Chamber

      2.61 m (8 ft - 7 in.)
      1.22 m (4 ft)
      2.44 m (8 ft)
      0.95 m (3 ft - 1-1/2 in.)
      1.74 m3 (61.4 ft3)
      0.63 cm (0,25 in.)  thick cold-rolled steel.
      12.7 cm (5 in.) thick high alumina castable
      refractory, variable depth to produce a
      frustroconical effect for  moving solids.
      Clockwise or counterclockwise 0.2 to 1.5  rpm
      1 hr (at 0.2 rpm)
      North American Burner, rated at 770 kW (2.6  MMBtu/hr)
      with liquid feed capability.
      Natural Gas

      Positive displacement pump via water-cooled  lance.
      Moyno pump via front face, water-cooled lance.
      Metered twin-auger screw feeder or fiber  pack ram
      feeder.
      1010°C (1850°F)

Characteristics of the Afterburner Chamber

      3.05 m (10 ft)
      1.22 m (4 ft)
      2.74 m (9 ft)
      0.91 m (3 ft)
      1.80 m3 (63.6 ft3)
      0.63 cm (0.25 in.)  thick cold rolled steel
      15.24 cm (6 in.) thick alumina castable refractory
      1.2 to 2.5 sec depending on temperature and  excess air
      North American burner rated at 590 kW (2.0 MMBtu/hr)
      with liquid feed capability
      Natural gas
      1200°C (2200°F)
               Characteristics  of the Ionizing  Wet  Scrubber  APCD
System capacity
  inlet gas flow
Pressure drop
Liquid flow
pH control
      85 m3/min (3000 acfm) at 78°C (172°F) and 101 kPa
      (14.7 psia)
      1.5 kPa (6 in. we)
      15.1 L/min (4 gpm) at 345 kPa (50 psig)
      Feedback control by NaOH solution addition
                                      SI

-------
                           EXPERIMENTAL TEST  PROGRAM
MUFFLE FURNACE EXPERIMENTATION

      A series of waste composition/Teachability tests was conducted in  a
muffle furnace prior to incinerator testing to determine the optimum
experimental conditions.  The muffle furnace tests consisted of 9 experiments
during which various concentrations of arsenic in the contaminated soil  were
heated at 982°C (1800°F) for 1 hour.  The weight losses (moisture and
volatiles) were determined for each sample, and the resulting ash was analyzed
for total arsenic (As).  Clay was mixed with the soil to get variable arsenic
concentrations in the feed.  Toxicity Characteristic Leaching Procedure  (TCLP)
leachates of the feed as well as the ash were also taken and analyzed for
arsenic and compared with the total waste/ash analysis.  Two percent lime  by
weight was added to one of the samples and two percent alum (ferric ammonium
sulfate, FeNH^SO^) was added to another of the samples.

INCINERATOR TESTING

      A series of five (5) tests was performed using the RKS at the IRF  to
determine the relative partitioning of arsenic to the different waste streams.
These were a part of the testing series to establish that incineration
effectively destroyed the organic contaminants in the soil and to evaluate the
fate of arsenic and lead as a function of incineration conditions.

      The incineration condition test variables were kiln temperature and  kiln
air flow.  Kiln temperature was targeted for 816 and 982°C (1500 and 1800°F),
and the air flow from the kiln exit flue was targeted at 6 and 10 percent
oxygen.

      All the soil was fed to the kiln in fiber-pack drums via the RKS ram
feeder system.  The fiber-packs, each of which contained about 5.0 kg (11  Ib)
of soil, were fed at the rate of one every 5 minutes.  The kiln rotation was
set to give a nominal solids residence time of 0.5 hr.

SAMPLES

      The samples taken during each test were:

1.  Scrubber blowdown water for arsenic,

2.  Feed and kiln ash for total arsenic analysis,

3.  Og concentrations at kiln outlet,

4.  Stack gas for arsenic incorporated into the particulate matter and
    POHCs, and

5.  Feed and ash for TCLP determinations.
                                      52

-------
                         TEST RESULTS AND DISCUSSION
MUFFLE FURNACE TEST RESULTS

      Soils with varying concentrations of arsenic were heated in  the muffle
furnace to determine the effect of arsenic feed concentration on the resultant
arsenic ash concentration and Teachability.  TCLP leachates of the feeds  and
ashes were analyzed for arsenic to observe the effects on arsenic  mobility.

      The test results are summarized in Table 1.  These data show the various
concentrations of arsenic which were derived by mixing the contaminated soil,
which contained 655 mg/kg of arsenic, with uncontaminated clay.

      These data from muffle furnace tests suggest that an initial arsenic
concentration of about 200 mg/kg would be the maximum feed contamination  level
which can be incinerated if the TCLP guidance level of 5 mg/L for the ash is
to be maintained.  It may be possible to increase the partitioning of arsenic
to the ash and maintain low TCLP values if lime was added to the material
before it was incinerated.

      The addition of 2 percent lime by weight to 100 percent of the
contaminated soil increased the arsenic retention 1n the ash to 83 percent.
The addition of the lime reduced the TCLP value by 53 percent when compared to
an untreated sample.

      The addition of 2 percent alum had the opposite effect on the arsenic
retention 1n the ash with the retention being reduced to 38 percent.  The TCLP
was 9.8 mg/L as compared with 6.5 mg/L when lime was added Indicating that the
retaining properties of alum were less than lime,  Figure 1 shows the TCLP
values datermlntd for each soil concentration and the resulting ash TCLP  for
the data shown 1n Table 1.  The soil that was processed had a consistently
higher TCLP value than Its corresponding untreated soil.

INCINERATION TEST RESULTS

      Five Incineration tests were conducted at the IRF on a contaminated soil
(at the arsenic concentration as received) from the Balrd & Mcfiulre site.
Table 2 summarizes the resulting distribution of arsenic 1n the ash, kiln exit
gases and scrubber water.  The ash generation rate was determined by using the
feed rate, less the volatlles and moisture.

      The test results are contained 1n Table 2.  Table 2 1s the arsenic
material balance and a general summary.  This table has been arranged by
temperatures of operation, so the Tests 1, 2 and 5 are the 816°C (liOO°F)
condition, and Tests 3 and 4 are the 982°C (18QQ°F) operating condition.   Test
5 was a duplicate of Test 2 to check the data precision.

      Table 2 shows the conditions of operation.  Tests 1, 2 and 5 had target
temperatures of 816°C (1500°F).  The mean experimental temperatures achieved
were 832°C (1530°F). 839°C {154QOF) and 844° (1561°F).  Tests 3 and 4 had a
                                      S3

-------
target temperature of 982°C (1800°F) and operated at 994°C (18210F) for both
tests.  Temperatures for all tests were close to the target conditions and
judged successful.  For ease of presentation, the temperatures 816°C (1500°F)
will be used for Tests 1, 2 and 5, and 982°C (1800°F) will be used for Tests 3
and 4.  The actual temperatures achieved are used in the table.

      For the same reasons discussed above, the oxygen concentrations of
7 percent and 11 percent are used in the text presentation.  Tests 1, 3 and 5
were at the 11 percent concentration and Tests 2 and 4 were at the 7 percent
concentration.  Operating the kiln at 816°C (1500°F) resulted in the lower
arsenic emissions at the stack.  At the 11 percent kiln oxygen rate (high air
flow rate), the arsenic emissions at the stack were 4.3 and 3.6  percent of the
feed for Tests 1 and 5.  For Test 3, at the 11 percent oxygen and 982°C
(1900°F) kiln operating temperature, the emissions of arsenic at the stack
were 5.5 percent of the feed.  Using the lower temperature reduced the arsenic
emissions at the stack over 35 percent.

      The lower oxygen concentration (7 percent) (lower air flow rate) at the
kiln outlet further reduced arsenic emissions to the atmosphere.  At 816°C
(1500°F) and 7 percent oxygen at the kiln, the emissions were 2.1 percent of
the feed compared to 4.7 percent at 982°C (1800°F).  The 982°C (1800°F) had an
emission rate of arsenic over 100 percent higher than the 816°C  (15QO°F)
operating at the lower air flow.

      Incinerating the soil at 816°C (1500°F) and 7 percent oxygen at the kiln
outlet (low air flow) was the best operating condition of the five tests for
limiting arsenic stack emissions.  The retention of ash in the feed was 85
percent of the arsenic feed and the emissions to the atmosphere  were 2.1
percent of the arsenic feed.  Operating at the lower kiln temperature of 816°C
(1500°F) and the lower oxygen concentration, lower air flow increased the
retention of arsenic in the ash and reduced the emissions to the atmosphere.
The retention of arsenic in the ash was 84.6 percent at 816°C (1500°F) and 6.8
percent oxygen.

      The lowest retention of arsenic in the ash was 29.8 percent at 982°C
(1800°F) operating temperature, and elevated oxygen concentrations.  Ash
retention data reinforces the emissions data at the lower air flow and lower
temperature of operation.

      The ash from the RKS tests had a greater TCLP value than the feed in all
cases (Table 3).  It was suggested that the incineration of the  contaminated
soil destroyed the part of the soil matrix that was holding the  arsenic.  Once
incinerated, the arsenic became much more mobile and produced higher TCLP
values.  Based on drinking water health effects data, the guidance TCLP for
arsenic is 5 mg/L, and the highest ash TCLP value from these tests was 1.24
mg/L.  Therefore the TCLP of the ash was well below guidance values for all
tests.

      The data reported in this paper are in line with full scale emissions
data previously obtained on arsenic.  Gerstle (2), in analysis of metals data
from sewage sludge incinerators found that an increase in temperature from
1300°F to 1500°F would double the emissions of arsenic.  An additional

                                      54

-------
increase in temperature to 1700°F would quadruple those experienced at 1300°F.
Higher temperatures for incineration require more oxygen (air flow), thus  more
air flow resulting in a more turbulent combustion zone.  The turbulent
conditions will entrain more particulate.  A reduction of (^ at the kiln
outlet for the RKS may be equivalent to Gerstle's reduction in temperature.  If
the RKS behaved thermally in the same manner as the sludge incinerator,  the
average expected arsenic emission would be 0.8 to 2 percent of the feed  at
1500°F and would be 20 to 100 percent at 1800°F.  The question of
comparability not withstanding, the RKS appears to have done much better in
partitioning the arsenic to the ash.

                                  CONCLUSIONS


1.    Of the operating conditions tested, the largest percentage (85 percent)
      of arsenic retained in the ash was at 816°C (1500°F) and 7.5 percent
      oxygen.

2.    Operating conditions at the lower temperature 816°C vs. 982°C (1500°F
      vs. 1800°F), increased arsenic retention in the ash.

3.    Reduced oxygen rates (air flow rates) gave increased retention of the
      arsenic regardless of the operating temperature.

4.    Reduced oxygen rates (air flow rates) decreased emissions at the stack
      outlet.

5.    The TCLP of the ash in these tests was always lower than the 5 mg/L
      guidance 1imits.

                                  REFERENCES
1.    Acurex Corporation, "Test Plan for Evaluating the Incinerability of
      Wastes from Baird & McGuire Superfund Site," USEPA Contract No.  68-03-
      3267, Work Assignment 3-5, Cincinnati, Ohio, August 1989.

2.    Gerstle, Richard W. and Albrinck, Diana, Atmospheric Emission  of Metals
      from Sewage Sludge Incineration, Journal of the Air Pollution  Control
      Association. Volume 32, No. 11, p. 1119-1123, November 1982.
                                      55

-------
 O
 LU
 o

 I
 LLJ




 1
 LU
1
     14
     12
     10
                                                           D
TCLP LIMIT
                      200             400
                   600
                              CONCENTRATION (ing/kg)
800
Figure 1.  TCLP  leachate,  mg/L,vs. arsenic concentration in soil  and
           in resulting  ash  from muffle furnace,
                                   .5!

-------
                        Table  1.   Muffle Furnace Results

Concentration Expected ash Measured ash
of arsenic arsenic concentration
in soil concentration arsenic
mg/Kg (dry basis)* mg/Kg mg/Kg
655
524
339
328
262
131
<5
642*
642*
863 468
701 351
528 310
437 299
351 211
174 130
<7 5.1
862 716
879 330
Fraction
remaining TCLP
in ash ash
(percent) mg/L
54
50
59
68
60
75
-
83
38
13.7
13.0
9.4
9.8
7.6
4.5
<0.07
6.5
9.8
*dry basis calculated using moisture and volatile removal
+2% lime added
#2% alum added

Table 2. Incinerator Test Results



Incinerator
Test kiln
No. tenp.°
1 832
5 839
2 844
3 994
4 994
Arsenic Arsenic
% Oxygen content retention in scrubber Arsenic
of flue gas at in ash, % water* in the
kiln exit of feed mass % of feed stack gas
11 71.9 19
11 57.2 21
6.8 84.6 22
10.4 29.8 34
7.5 41.9 30
4.3
3.6
2.1
5.5
4.7
Metal balance
% of feed
95
82
109
69
76
*scrubber water and stack emissions

-------
                       Table 3.  Summary of Test Results
                           Arsenic
             Arsenic in  retained in
Test  Temp   feed rate    ash % of
No.    °C      mg/kg        feed
                        Arsenic emitted
                        at stack ratio
                         to feed mass
                             Arsenic
                             TCLP of
                               ash
                               mg/L
                        Arsenic
                        TCLP  of
                         feed
                         mg/L
 1     832

 5     839

 2     844

 3     994

 4     994
82.2

82.8

92.6

80.5

83.5
71.9

57.2

84.6

29.8

41.9
.0035

.0031

.0081

.0051

.0038
0.3

0.2

1.2

0 2

1.2
0.1

0.1

0.1

0.1

0.1
                                       58

-------
              FUNDAMENTAL STUDIES ON PARTICULATE EMISSIONS
                    FROM HAZARDOUS WASTE  INCINERATORS
                    Virendra Sethi and Pratim Biswas

            Department  of Civil  and  Environmental Engineering
                        University of Cincinnati
                      Cincinnati, Ohio 45221-0071.
                                ABSTRACT

   A fundamental study has been carried out to understand the formation
and growth of metallic particles in incinerators.   A bench scale flame
incinerator was used to perform controlled experiments with lead acetate
and silicon tetrachloride as the test compounds.   A dilution probe and
real time aerosol instruments were used to measure the particle size
distributions at different locations in the flame region.  A lognormal
aerosol model accounting for particle formation by nucleation and
chemical reaction, and growth by condensation and coagulation was used to
predict the evolution of the particle size distribution.  Reasonable
agreement was obtained between the model predictions and experimental
results in view of the limitation of the aerosol measuring instruments.
Condensation was the dominant growth mechanism for the lead aerosol.
With silicon tetrachloride as the test compound,  non-volatil« silica
particles were formed, and coagulation was the dominant mechanism of
particle growth.


   This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved
for presentation and publication.  The conclusions represent the views of
the authors and do not necessarily represent the opinions, policies, or
recommendations of the U.S. Environmental Protection Agency.
                                   59

-------
                              INTRODUCTION
   Emissions from hazardous waste incinerators are currently regulated by
the Resource Conservation and Recovery Act (RCRA) incinerator performance
standards (1).  The requirements of 99,99 % destruction and removal
efficiency for each principal organic hazardous constituent (POHC), and
99 % removal of the HC1 gas have been met in most cases of hazardous
waste incinerator operation. However, the particulate emission standard
of 180 mg/nr* has often been violated (2).  Combustion processes are the
major sources of toxic metal loadings in the atmosphere, with waste
incineration and coal combustion being the two major contributors (3),
Waste incinerators are becoming popular as they lead to significant
reduction in volume of the original waste (90 to 95 % by volume),
However, stack gas emissions from these units, especially those of toxic
volatile metals, are of great concern.  Vogg et al. (4) report enrichment
factors of 20 for Pb, 60 for Sb, 100 for Cd, and 600 for Hg during waste
incineration.  It has been well recognized that the submicron particles
are enriched with volatile metals such as Hg, Pb, Cd, As, and Zn.

   Many studies have been conducted to examine particulate emissions from
incinerators in operation.  Very few have examined detailed size
distributions of these particulate emissions, especially in the submicron
range.  These submicron sized particles are collected with lowest
efficiency in the control devices, and are the most harmful to human
health (4, 5).  A notable exception is the work by Kauppinen and Pakkanen
(5) who reported aerosol size distributions from hospital refuse
incineration.  The submicron sized particles were significantly enriched
with toxic elements like Pb and Cd.

   If incineration is to be an effective alternative to conventional
hazardous waste disposal methods, there is need for a better
understanding of toxic particle formation processes in such systems.  In
order to effectively design and choose particle control devices for such
incinerators, it is imperative that the particle size distributions of
the emissions be known,  A model to predict formation and growth of
volatile particles in a. flame incinerator system has been developed by
Sethi and Biswas (6).  This paper describes a bench scale experimental
system to perform studies to verify the predictions of the model.


                           EXPERIMENTAL SYSTEM
   The schematic of the experimental system to study the evolution of the
                                    60

-------
       Dilution Air fParticle Free*
       Dilution Air (Particle Free)
       Primary Dilution
      Air
                            Sampling
                               Probe
      Methane
              Diffusion Dryer
             Atomizer
                                                                    Exhaust
    Secondary
    Dilution
Burner
                   _  Diluted
                      Sample
                                                        a
                                                        o o
                                                         0

                                                       OMPS
                                                                ED
                                                                   Exhaust.
        Figure 1.   Schematic diagram of the experimental set-up.
particle size  distribution is shown in Figure 1,  It consist:: of  a
methane/air  laminar flow burner.  The test compound is introduced into
the flame region through one of the ports of the burner.  Experiments
were performed using lead acetate and silicon tetrachloride.  In  the case
of Pb, an aqueous solution of lead acetate was atomized, diffusion-dried,
and introduced into the flame as an aerosol.  Silicon tetrachloride, for
the other case,  was introduced into the flame as a vapor entrained in Ar
as the carrier gas passed through a bubbler. A dilution probia was used  to
draw out samples from the flame for size distribution measurements in
real time (7).   Particle free air was used as a dilution gas to freeze
all chemical reactions and aerosol dynamic effects once the aerosol
enters the probe.   A secondary dilution system using particLs free air
was used to  ensure that the concentrations were lower than 10*
particles/cm3  to be within the instrument measuring range.
A Differential Mobility Particle Sizer (DMPS; manufactured by TSI, Inc.)
was used to  measure the particle size distributions for the silica
tetrachloride  feed tests and an Optical Particle Counter (OPC;
manufactured by PMS, Inc.) was used for the Pb aerosol feed tests.  The gas

                                    II

-------
was passed through a. filter before being exhausted into the hood.  The
experimental conditions are listed in Table I.
                    TABLE I. EXPERIMENTAL CONDITIONS
FLAME CHARACTERISTICS:

Methane Flow Rate, 1pm                                1.0
Air Flow Rate, 1pm                                    2,5
Burner Diameter, cm                                   0.2
LEAD COMPOUND:

Solution Concentration, g/cm^ lead acetate            0.159
in Atomizer
Feed Aerosol Concentration, 10^ #/cm^                 5.0

Feed Aerosol Volume Average Diameter, urn              0.5

Feed Aerosol Geometric  Standard Deviation             1.4


SILICON COMPOUND:

SiCl4 Feed Rate, mg/min                              20
Carrier Gas                                           Ar




                          RESULTS AND DISCUSSION
   The temperature history  in  the flame is important as it determines the
rate of different physico-chemical phenomena.  The experimentally
determined temperature profile as a function of distance from the burner
inlet as measured by a very fine wire  fR' type thermocouple is shown in
Figure 2.  The solid line in Figure 2  is a prediction of the temperature
profile (6). Reasonable agreement is obtained between model predictions
and measured data; the trend and peak values are in good agreement.  The
differences close to the burner inlet  are because the model assumes the
flame to be a pure diffusion flame and neglects radiative heat transfer
with the surroundings.

                                   62

-------
   2000
    1500 -
V.
3
1000 -
     500
                                       THEORETICAL PREDICTION
                                 DOOOD EXPERIMENTAL  MEASUREMENTS
                         10       15      20       25
                       Distance fron Burner  Inlet  (cm)

        Figure 2.  Comparison of  the theoretical and experimental flame
        temperature profiles.
   Experiments were conducted with lead acetate aerosol and the results
are plotted in Figure 3.  The experiments were repeated three times for
the same inlet conditions and the evolution of the average volume
concentration, number concentration, volumetric mean particle size, and
geometric standard deviation is plotted in Figures 3(a), (b) , .(c).'and
(d), respectively.  The vertical bars in the figures represent
variability in the data while repeating the experiments.  As the Pb
compound enters the flame, the particles tend to evaporate in the high
temperature region of the flame, and there is a steep decrease in both
the aerosol number and volume concentration (0 to 10 cm).   Predictions
from the theoretical model indicate that the Pb will be completely in the
vapor state in the flame region (6).  On proceeding downstream of the
flame, the temperature decreases (Figure 2); therefore, the saturation
ratio of the Pb vapor increases.  This vapor then nucleates into fine
particles, and there is an increase in both number and volume
concentration, as the vapor is now converted to the particle phase where
further growth takes place by condensation and coagulation.  The
experimental data reflect a similar trend.  The measured number
concentrations are low in the high temperature regions of  the flame zone
(3 to 10 cm) and essentially reflect complete evaporation  of the Pb
                                   S3

-------
    io-7

 c  10

«"  10"
 a

 1 10-tc|
 J
 1 ,0-n

 f 10'"
s  10-"
o
10
                                              (a)
          \
          to
                               1
                                                              10
                                                         3 io-s
                                                              10
                                                                              _l	I	I	t
  tic8
    10 3
    10'
     10
                                                          "S
                                                          I-
                                                          u
                                                          w
                                                           o
                                                           s
                 10     15    20    25     30
                Distance Along Plane Axlt ( en )
                                                   35
                                                         40
                                                               0.0
                                                                                                   (d)
                                                                       *•*.  Experimental Results
                                                                       - Theoretical Prediction
                                                                     +
                                                                                                -*-
 10     IS    20    25     30
Distance Along Tltat Axis { em )
                                                                                                             35
                                                                                                                    40
                    Figure 3.  Lead aerosol characteristics as a function of dlst.ince  along
                    Clamv axis for an aerosol volume feed rate of 5.4E-8 cm^/cra^,  georautrlc
                    nean particle size of O.S pm, and a geometric standard deviation of  1.4
                    (a) aerosol voluae concentration, (b) number concentration,  (c)  volume
                    average particle dianeter, and (d) geometric standard davlatlon.

-------
aerosol.  Increases in number concentration and volume concentration in
the post-flame region are observed to be close to the theoretically
predicted values.  The differences between predicted and measured values
at locations closer to the inlet arise due to the limitations of the OPC
in measuring particles smaller than 0.1 fim.  Better agreement between
aodel predictions and experimental data is obtained later f,s the
particles grow larger by vapor condensation and coagulation.

   The results for the silicon tetrachloride  experiments are plotted in
Figures 4 (a) through (d).  Silicon tetrachloride is oxidized to form
silica, and the molecule itself is a chermodynamically stable particle
(7).  Hence, the reaction rate governs particle formation rates, and
coagulation is the mode by which particles grow in the flame region.
Significant differences are obtained between the theoretical and the
predicted values of the aerosol volume and the number concentrations at
initial locations.  This again arises due to the limitation of the DMPS
in measuring particles smaller than 0.05 /im, and the silica particle when
first formed is about 4 A. in diameter. Subsequently, the decrease in
theoretically predicted number concentrations by coagulation agrees well
with the experimental results as the particles grow to sizes that are
readily detected by the DMPS.

   In full scale incinerator systems, multiple elements are present.  The
vapors of the volatile metals (such as Pb) may condense onto the non-
volatile particles (such as silica).  Controlled experimental studies
need to be performed with both compounds present simultaneously.  The
model has been used to predict Pb aerosol size distributions in the
presence of seed particles, and nucleation of the volatile element was
found to be completely quenched (6).  This has important ramifications
with regards to effective incinerator operation.  By properly choosing
design and operating conditions, one may effectively condense these
volatile toxic species either homogeneously or on existing seed materials
to grow large enough so that they can be efficiently captured in
particulate control devices.
                               CONCLUSIONS
   An experimental  system to  study  particle  formation and the  evolution
of the aerosol size distribution was  developed.  Experiments were
performed with single  components, and reasonable agreement was  obtained
with the predictions of  a comprehensive model.  Volatile elements
nucleated to form tiny submicron particles in the downstream regions of
the flame.  Further growth of these particles occurred by rapid
condensation of  Pb  vapor onto existing seed  particles.  When silicon .
tetrachloride was used as the test  compound, stable silica particles were
formed which grew by coagulation.   In full scale incinerator systems,
volatile and non-volatile elements  are expected to be present,  and hence
condensation onto particles is expected to be a dominant mechanism.  The
theoretical model may  be further developed for application in  full scale
                                   85

-------
  10
  10
  10
  10
  10
3 10
   10*


   10'


   10
                                              (b)
-«- • -*-
                "5         10        15        20
                   Distance Along flame Axis (  c« )
                                                               3.0
                                   25
                                                             5. 2.0
                                                             "
                                        .21.0
                                         0.0
                                                                                   (d)
                                                            •-••- Experimental Results
                                                            	Theoretical Prediction
                                                                10        15        20
                                                           Distance Along FUa« Axis ( ca )
                                                                                               25
                     Figure 4.   Siller  aerosol characteristics as a function of distance along
                     the flame  axis  for a  vapor  ?eed rate of 20 mg/niln  : (a) aerosol volume
                     concentration,  (b) number concentration, (c) volume average particle
                     diameter,  and (d)  geometric standard deviation.

-------
incinerators and used to determine operating conditions to achieve
efficient capture of toxic elements.
                             ACKNOWLEDGEMENT
   We acknowledge the support of the U. S. Environmental Protection
Agency under contract 68-03-4038, Work Assignment 2-RTP42.0 .
                               REFERENCES
 1.   Federal Register 47:27516,  1982.

 2.   Oppelt, E. T. Hazardous waste destruction. Environ.  Sci. 	Tech.
     20:312, 1986.

 3.   Kowalcyzk, G. S.,  Choquette, C,  E. and Gordon G. E. Chemical element
     balances  and identification of air pollution sources in Washington,
     D. C. Atmos.  Environ.  12:1143, 1978.

 4.   Vogg, H. , Braun, H. , Metzger, M.  and Schneider, J. The specific role
     of cadmium and  mercury in municipal solid waste incineration.
     Waste Manag.  Res.  4:65, 1986.

 5.   Kauppinen, E. I. and Pakkanen, T. A.  Mass and trace element size
     distributions of aerosols emitted by a hospital refuse incinerator.
     Atmos. Environ. In press, 1990.

 6.   Sethi, V. and Biswas,  P.  Modeling of particle formation and dynamics
       in  a flame  incinerator.  JAWMA  40:42, 1990.

 7.   Biswas, P.,  Li, X.  and Pratsinis, S. E. Optical waveguide preform
     fabrication:  silica formation and growth in a high temperature
     aerosol reactor.   J. Appl.  Phys.  65:2445, 1989.
                                    I?

-------
                   FATE OF VOLATILES IN WASTEWATER TREATMENT

           by; Richard A. Dobbs and 2Sanjoy K.  Bhattacharya

               lRisk Reduction Engineering Laboratory, United States
                Environmental Protection Agency, Cincinnati, OH  45268

               2Civil Engineering Department, Tulane University,
                New Orleans, LA  70118
                                   ABSTRACT

      Factors which affect the transport of volatile compounds from the
aqueous phase to the gas phase by surface desorption (volatilization) and air
stripping in conventional primary/activated sludge wastewater treatment plants
are reviewed.

      Fate of eleven volatile organic compounds spiked to pilot-scale
activated sludge system was determined.  Removals by sorption, volatilization
and stripping, and biodegradation were assessed by measuring influent,
effluent, primary and secondary sludges, and primary and aeration basin off-
gas samples for the compounds studied.
                                    68

-------
                                 INTRODUCTION

      Tht fate of volatile organles 1n conventional primary/activated sludge
treatment systems 1s controlled by several Integrated removal mechanisms which
operate simultaneously.  The major mechanisms Include;  (a) sorption on sol Ids
and biomass, (b) volatilization to the atmosphere through surface desorption
and air stripping, and (c) blodegradatlon or transformation by aerobic or
anaerobic processes.  Direct oxidation, hydrolysis, and other mechanisms may
play a minor role In the removal of certain organic compounds.  Sorption on
wastewater solids has been correlated with the octanol/water partition
coefficient and modified Randic* Indexes (1).  The correlations developed are
useful for modeling the role of sorption in a wastewater treatment plant.
Volatilization and air stripping and the biological processes are more
complicated and are controlled by compound specific properties, wastewater
characteristics, and plant design and operational parameters.  This paper will
review some of the more important factors which control the fate of volatile
compounds in wastewater plants with emphasis on air emissions from primary and
activated sludge treatment systems.

FACTORS AFFECTING VOLATILE EMISSIONS

      Several research studies which relate to the problem of air emissions
from wastewater treatment plants have been reported (2-32),  These studies
have attempted to develop an understanding of the principles which affect the'
transport of volatile compounds from the aqueous phase to the gas phase by
surface desorption (volatilization) and air stripping.  Table 1 lists several
factors which are important in air emissions from wastewater treatment plants.
This section will discuss these factors in some detail.

                   TABLE  1.   FACTORS AFFECTING  AIR EMISSIONS


                                 Aeration  Rate

                             Henry's Law Constant

                                  Acclimation

                         Blodegradatlon Rate Constant

                                      Psi

                                     Alpha

                                Aeration Device

                                 Diffuser Type

                          Surfactants/Other Organles

                                  Metabolism

                                  Temperature

                                     IS

-------
AERATION RATE

      The effect of aeration rate on the stripping of volatlles  from water and
wastewater has been studied extensively.  The effect 1s Illustrated  In  Figure
1 where log concentration 1s plotted as a function of time for four  different
aeration rates from a 10-11ter bioreactor without biomass  (31).

             Concentration
                (ug/l)
                    100
                                                1.0 L/min
                   10.0
                    1.0
3.0 L/min



 4.0  L/min
                                                  5.0 L/min
                            10    20    30    40
                                  Time(Minutes)
      50
60
             Figure  1.   Effect of  aeration rate on stripping

                         of benzene  without  biomass.


      Stripping rates were reproducible and not affected  by initial  concen-
trations.  When comparing air emission data from different  sources,  the gas to
liquid ratio will be an Important factor.

HENRY'S LAW CONSTANT

      Henry's Law Constant (H.)  1s a critical  parameter in  overall stripping
models for volatiles from wastewater treatment plants. Figure 2  (31)
Illustrates the observed correlation between Hc  and stripping  rates  for
selected volatile compounds.
                                    70

-------
      Concentration
         (ug/0
         100
        10:0
         1.0
                          \
       1,2,4-Trichlorobenzene  (2.32)*

       1,2-Dichlorobenzene (2.96)*

 Chlorobenzene  (3.71)*

 0-Xyiene (5.27)*

Benzene (5.49)*

 Toluene (6.66)*



 Ethylbenzene  (8.73)*
                                           * Hc ,(atm-m^/mol)  x 10
                                   ,-3
            0   10   20   30  40    50   60
                   Time(Minutes)
       Figure  2,  Effect of Hc on  stripping of  volatile  organics.


      In the absence of other factors, stripping of volatile compounds was
directly related to Henry's Law Constant.

ACCLIMATION

      When biota are exposed to compounds which are not readily degraded,
adaptive changes can occur within the population which enable  certain
organisms to metabolize the substances.  This process, known as acclimation,
may take several days to  several weeks or even longer to  occur.  A1r emissions
of volatiles are substantially reduced when the biomass 1s acclimated.   The
magnitude of this effect  1s shown 1n Table 2.
                                     II

-------
               TABLE 2.   EFFECT OF ACCLIMATION ON AIR EMISSIONS*
                                             % Stripped
Compound
Benzene
To! uene
Ethyl benzene
o-Xyl ene
Chlorobenzene
UjaccJ'Jiated
45
40
88
90
50
Acclimated
16
17
22
25
20
*Reference 31

      Under acclimated conditions, blodegradation can compete effectively with
volatilization and stripping as a removal mechanism and can significantly
reduce the air emissions from wastewater treatment plants.  An acclimated
blomass can maintain the ability to degrade specific compounds 1f they are
usually present 1n the Influent wastewater.  In the case of Intermittent
discharges, specific compounds may not be present in the Influent wastewater
for a period of time.  Under these conditions, an acclimated blomass may lose
the ability to degrade a given compound.  However, 1f the component 1s again
discharged within a reasonable period of time (I.e., less than 3 sludge
retention times) the reaccl1mat1on process 1s substantially shortened as
Illustrated by the results summarized 1n Table 3.

         TABLE  3.   COMPARISON OF  INITIAL ACCLIMATION AND REACCLIMATION
                AFTER 14  DAYS WITHOUT TOXICS ADDED TO  INFLUENT*


                      Overal1 Removal (%1            Acclimation Period (Pavs)
Compound
Benzene
Toluene
Ethyl benzene
o-Xylene
Chlorobenzene
Initial
86
81
79
78
81
Reaccllmatlon
90
88
90
81
86
Initial Reaccllmatlon
16-20
16-20
16-20
16-20
16-20
3-4
3-4
3-4
3-4
3-4
*Referance 31 (Sludge Retention Time 6 Days)

      Approximately 2-3 weeks were required for Initial acclimation compared

                                    72

-------
to 3-4 days for reacclimation for the compounds listed in Table 3.

BIODEGRADATION RATE CONSTANT

      Volatilization and stripping of volatiles from a wastewater treatment
plant are maximized if no competing mechanism contributes to the removal.
However, for many compounds the biodegradation rate constant (KJ  is large
enough so that this mechanism competes effectively with volatilization and
stripping in the overall removal process.  Under this condition, air emissions
are substantially reduced.  The effect of biodegradation rate constants on the
fate of volatiles in two large wastewater treatment plants has been reported
(28).  Results for one of the plants are presented graphically in Figure 3.
               voc
             Removal
              100

              90

              80

              70

              60


              50

              40


              30

              20


              10
                                                 Total
                                                        Sorption
                 o.oooi
                               o.ooi
                                           o.oi
                                                     0.1
                                                              10
           Figure 3.  Effect of biodegradation rate  constant
                       on removal of  volatiles.
      As the biodegradation rate constant Increases, the total removal of a
given volatile compound Increases, removal by stripping decreases, and
sorption remains essentially constant.

PS I

      Stripping can be defined as the transfer of organics at dispersed
gas/water interfaces such as the surface of air bubbles produced by subsurface
aerators.  In this case, the rate of stripping can be defined as follows:
                              rs -  KLsa(C*Li - CL!)
                                                          (1)
where K,sa - overall  mass transfer coefficient from the liquid phase
  LI
liquid phase concentration of compound i in equilibrium with the gas

                              73

-------
phase CLi » effluent concentration of compound i.

      Since oxygen  transfer  for a given system can be measured, oxygen is used
as a reference for  expressing the mass transfer of other volatile organic
compounds  as shown  in the following  equation:

                                KLsa, = ^KL a02                   (2)

where KLsat - overall mass transfer coefficient of compound  i
0, » transfer rate proportionality coefficient.   The value of $f  can  be
measured or estimated from the relative diffusivity of the compound and oxygen
using the  following equation:

                                                                  (3)
                                        D02j

Measured values of n for selected volatile compounds ranged from 0.61 to 0.66
(17, 23).   In Barton's model for stripping, n was assigned a value of 0.6
(29).

ALPHA

      Alpha is an oxygen transfer correction factor.  It is defined by the
simple ratio of KLa in wastewater/KLa in clean water.  Standard procedures  for
oxygen transfer determinations have been described for both clean water (33)
and wastewater (34).

AERATION DEVICES

      Although most wastewater treatment plants use bubble or dispersed
aeration systems, some do use surface aeration.  Both systems have been
modeled by several researchers cited in the reference section.  In general,
surface aeration achieves greater transfer of organic solutes to the
atmosphere than does bubble aeration for a given oxygen transfer rate (24).
The same authors reported that in bubble aeration the transfer efficiency of
organic solutes decreases with increasing oxygen transfer efficiency when the
overall transfer requirement is held constant.

DIFFUSER TYPE

      Three levels of oxygen transfer efficiency, viz.,  6%, 10%, and 15%,
correspond to values for coarse, medium, and fine bubble diffused-air
aeration, respectively, with a submergence depth of 4 meters (35).  Since
coarse bubble diffusers have the smallest percent oxygen transfer efficiency
(6%), a higher gas/liquid ratio 1s required to maintain dissolved oxygen at
the desired level.  Under these conditions, stripping of volatiles is
increased.  As a consequence, the use of medium or fine bubble diffusers would.
decrease the emissions from a wastewater treatment plant.

SURFACTANTS/OTHER ORGANICS

      The rate of mass transfer in aeration systems is strongly dependent on

                                    74

-------
wastewater composition.  Low concentrations of surface active compounds in
water have been shown to reduce the overall mass transfer coefficient, KLa
(36-39).  The effectiveness of porous diffusers has been shown to be increased
significantly by organic constituents of wastewater such as phenols, alcohols,
and acids.  Alpha values approaching 3.0 have been observed in 0.1 millimolar
acetic acid (40).

METABOLISM

      Microorganisms can convert many toxic synthetic organic chemicals to
inorganic products.  These microbial processes can lead to environmental
detoxification, the formation of new toxicants, or the biosynthesis of
persistent products (41).  Research has shown that the source of the highly
volatile chloroethene compounds, vinyl chloride, vinylidene chloride, cis and
trans 1,2-dichloroethene in a raw groundwater was the result of biodegradation
of tricholorethylene and/or tetrachloroethylene which are found widely spread
in the environment as a result of extensive use of these compounds (42).  Both
aerobic and anaerobic organisms were capable of degrading chlorinated
solvents.  In complex mixtures where several chlorinated or aromatic solvents
are present, metabolism can result in the transformation of on« compound into
another.  Under these conditions one can measure stripping efficiencies in
excess of 100 percent.  This effect will be discussed in more detail in a
later section.

TEMPERATURE

      The general expression for the volatilization rate constant Kv can be
expressed in terms of the mass transfer rates of the compound across liquid
and gas phase boundary layers.  The general expression can be written as
fol1ows:
kv  =  in   + .BIT1                 (4)
             HkgJ
                              v =              1
                                           cg
where L is the depth which equals the interfacial area, A, divided by the
liquid volume, V; k, is the liquid film mass transfer coefficient; R is the
gas constant; T is the temperature; H  is the Henry's law constant; and kg  is
the gas-film mass transfer coefficient (17).  The two-film mas:; transfer model
for volatilization has been reviewed and applied to environmental problems
(43, 44).  As shown by Equation 4, the rate constant for volatilization or
stripping is directly proportional to temperature.

      An understanding of the factors described in this section is essential
for accurate estimates of volatile emission rates from wastewater treatment
plants.

                      FATE OF VOLATILES IN CONVENTIONAL
                      PRIMARY/ACTIVATED SLUDGE TREATMENT

      The removal of volatiles by six wastewater treatment processes was
studied by Hannah et al. (45, 46).  The processes included conventional
primary/activated sludge treatment.  Weber and Jones (31) conducted bench-

                                     75

-------
full scale  treatment plant data.   None  of the studies cited  in this section
measured loss of volatiles in primary treatment.  The first attempt to develop
fate-1n-treatment data  for a  conventional  primary/activated sludge system was
reported by Bhattacharya et al.  (47) for selected volatiles and semivolatiles.

      In the present study, an attempt was  made to obtain improved mass balance
data for  the fate  of selected  volatiles  from the Resource  Conservation and
Recovery Act (RCRA) list of pollutants for  both primary and  secondary treatment
processes.  Two pilot-scale activated sludge systems were operated in parallel
at  the  United  States Environmental  Protection  Agency's  Test  and Evaluation
Facility in Cincinnati,  Ohio.  One pilot plant was used as the spiked test system
while the other served  as  an  unspiked control.  The eleven compounds selected
for this study from the RCRA  list of pollutants are  shown in Table 4 along with
physical-chemical properties.

                   TABLE 4.   LIST OF SELECTED  RCRA  COMPOUNDS
 Compound
CAS No.   TTVH_
         Ippm) (atm.m /mole}
Acetone
Cyclohexanone
Methyl ethyl ketone
Tetrahydrofuran
Ethyl benzene
Tr i chl oroethyl ene
Carbon tetrachloride
1,1, 1-Tr i chl oroethane
Chlorobenzene
1 , 1 , 2-Tri chl oroethane
Tetrachl oroethyl ene
000-067-641
000-108-941
000-078-933
0109199-9
000-100-414
000-079-016
000-056-235
000-071-556
000-108-907
000-079-005
000-079-016
750
25
200
200
100
100
10
450
75
20
50
6.8 X 10"6
2.5 X 10"5
5.8 X 10"5
1.1 X 10'4
6.4 X 10"3
9.1 X ID"3
2.3 X 10"2
3.0 X 10"2
3.5 X 10"3
7.4 X 10"4
1.5 X 10"2
0.57
6.46
N.A,
5.4
1412
263
436
310
690
117
759
  H_ - Henry's Law Constant
 N.A. - Not available

MATERIALS AND METHODS
        octanol/water partition  coefficient
      Two pilot-scale activated sludge systems were operated at a flow rate of
35 gpm  (2.2 L/s)  and  a  hydraulic  retention time of 7.5 hours.   An operational
sludge retention  time (SRT)  of  4  days  was  used.   The organics  were added in a
single spike mixture to produce concentrations of approximately 0.25 mg/L of each
compound in the influent to the treatment system.   The operating conditions and
design characteristics for the two systems  used in this study are shown  in Table
                                     76

-------
35 gpm (2.2 L/s) and a hydraulic retention time of 7.5 hours.  An operational
sludge retention time (SRT) of 4 days was used.  The organlcs were added in a
single spike mixture to produce concentrations of approximately 0.25 mg/L of
each compound in the Influent to the treatment system.  The operating
conditions and design characteristics for the two systems used in this study
are shown 1n Table 5.


 TABLE 5.  OPERATING CONDITIONS AND DESIGN CHARACTERISTICS OF THE PILOT PLANT


  I.  Design Flow          = 35 gpm
                           = 50,400 gpd

 II.  Primary Clar1f1ers  -  Diameter              • 9'-8"
                             Weir Diameter         = 9'-l"
                             Surface Area          =73.4 ftz
                             Surface Overflow Rate = 687 gpd/ft2

III.  Aeration Basins - L:W:D                      - 17'-7":10'-0":12'-0"
                        Surface Area               - 175.8 ft2
                        Volume                     = 15,780 gal.
                        Residence Time             = 7.5 hrs.
                        Gas to Liquid Ratio        = 34.2
                        Sanitaire Course Air Diffusers, Model D24

 IV.  Secondary Clariflers - Diameter              * 11'-11"
                             Surface Area          - 111.5 ft2
                             Surface Overflow Rate = 452 gpd/ft2


      In order to sample the air space above the primary clarlfiers, loosely
fitted covers with a 1-inch (2.5-cm) gap at the edge were fabricated for the
units.  The covers were vented through a duct to the roof.  An air sweep
equivalent to a 4 kilometer per hour wind was maintained over the surface of
each primary clarlfler by exhausting air at 8,500 liters/minute.  The aeration
basins were fitted with airtight covers and the off gas was vented to the
roof.  The average air flow 1n each aeration basin was 4,250 liters/minute.
A1r samples were collected 1n stainless steel canisters and were analyzed by
GC/MS by a contract laboratory (PEI Associates, Inc., Cincinnati, Ohio) using
Method TO-14 (48).  Sludge and liquid samples for RCRA analytes were measured
using Method 1624 (49).

      Automated analytical procedures were used for the conventional
pollutants.  These analyses (soluble chemical oxygen demand and ammonia
nitrogen) were performed three times per week.  Total and volatile suspended
solids 1n the Influents, effluents, and mixed liquors were measured dally
using composite samples collected three times per day.  All these parameters
were measured following Standard Methods (50).

      The test system was spiked for ten weeks.  Following an acclimation
period of three weeks, composite aqueous and sludge samples were collected
                                    77

-------
during a 24-hr, period once every week for seven weeks.  Composite air samples
were also collected during this 24-hr, test period.  The average air flow
rates were also determined for each sampling event.  The wastewater flows,
dissolved oxygen concentrations, and pH were checked by the operators three
times per day.
                            RESULTS AND DISCUSSION
      The average percent removals and standard deviations for the
conventional pollutants during the test period are summarized in Table 6.

         TABLE  6.   AVERAGE PERCENT REMOVALS OF CONVENTIONAL  PARAMETERS
PARAMETER
  CONTROL SYSTEM
  PERCENT REMOVAL
            STD
AVERAGE   DEVIATION
   SPIKED SYSTEM
  PERCENT REMOVAL
            STD
AVERAGE   DEVIATION
Total Suspended Solids       96       2

Volatile Suspended Solids    96       2

Soluble Chemical             69      11
Oxygen Demand

Ammonia-Nitrogen             92      14
                           94

                           96

                           69


                           93
              4

              3

             14


             12
      The spiked toxic organic compounds in the wastewater produced no
significant effects on the treatment of conventional pollutants.  Ammonia
removals higher than 90% were achieved at an SRT as low as 4 days.  Since this
study was performed in the summer, high temperature probably enhanced
nitrification.  Nitrifiers are expected to be more adversely affected by
toxicants than activated sludge organisms.  High nitrification showed that the
added organic compounds were not toxic to even the nitrifiers.

      The fates of the volatile organic compounds were studied by measuring
the concentrations of each spiked compound in the primary influent, primary
effluent, secondary effluent, primary and secondary sludges, and primary and
secondary air samples.  Total removals for both primary and secondary
treatment processes are summarized in Table 7 for the volatiles studied.
                                     78

-------
            TABLE  7.   SUMMARY TOTAL REMOVALS (PRIMARY +  SECONDARY)
PERCENT TOTAL REMOVAL
NAME OF COMPOUND

Acetone
Methyl ethyl ketone
Cyclohexanone
Tetrahydrofuran
Carbon tetrachloride
1.1, 1-Trichloroethane
1,1.2-Trlchloroethane
TMchloroethylene
Tetrachloroethylene
Chlorobenzene
Ethylbenzene
EVENT
1
95.7
94.8
95.0
54.8
98.2
97.4
59.3
99.3
95.7
99.0
99.1
EVENT
2
94.
95.
-
-
98.
98.
66.
-
96.
98.
94.

0
0


8
1
2

0
3
6
EVENT
3
96.7
97.2
-
91.3
99.2
98.7
69.3
99.8
97.8
99.4
97.7
EVENT
4
96.8
96.7
92.4
92.0
98.4
98.6
68.9
99.9
97.7
99.9
99.2
EVENT
5
95.3
98.0
77.0
96.5
99.6
99.2
78.3
99.8
98.8
99.7
100.0
EVENT
6
99.3
98.8
-
100.0
99.3
98.9
79.4
100.0
98.3
100.0
99.8
EVENT
7
80.8
-
66.8
100.0
98.1
99.2
74.1
99.3
97.0
97.3
97.5
AVERAGE

94
96
82
76
98
98
70
99
97
99
98

.1
.7
.8
.4
.8
.6
.8
.7
.3
.1
.3
STANDARD
DEVIATION
5.64
1.45
11.52
34.41
0.52
0.60
6.55
0.27
1.07
0.91
1.75
(-) UNAVAILABLE DATA (Decomposed In sampling canister)
      Total removals  generally exceeded'90 percent  (1,1,2-trichloroethane and
tetrahydrofuran were  the exceptions).

      Removals by  volatilization and stripping are  presented  in Table 8 for
the combined primary  and secondary processes.

               TABLE  8.   SUMMARY OF REMOVALS BY VOLATILIZATION
                          AND STRIPPING (PRIMARY + SECONDARY)
PERCENT STRIPPED
NAME OF COMPOUND EVENT
1
Acetone 1 . 1
Methyl ethyl ketone 1.6
Cyclohexanone
Tetrahydrofuran 15.6
Carbon tetrachloride 117
1.1. 1-Trichloroethane 122
1,1,2-Trlchloroethane 16.2
Trichloroethylene 62.4
Tetrachloroethylene 122
Chlorobenzene 14.7
Ethylbenzene 16.9
(-) UNAVAILABLE DATA (Decomposed
EVENT
2
1.1
1.5
-
11.0
129
137
29.0
-
158
14.7
23.0
EVENT
3
2.6
1.5
-
8.1
61.7
94.3
23.9
51.2
81.0
11.1
16.5
EVENT
4
2.4
2.7
-
5.9
125
111
38.0
57.0
108
9.9
13.3
EVENT EVENT
5
1.8
1.8
-
10.2
107
106
51.5
53.7
108
10.4
12.0
6
* 1.
* 1.
-
* 10.
95.
105
52.
51.
120
15.
16.

8*
8*

2*
6

0
5

9
0
EVENT
7
1.8*
1.8*
-
10.2*
83.3
85.5
48.6
38.5
78.7
11.4
13.9
AVERAGE
STANDARD
DEVIATION
!
ii

10
103
10!)
37
5!!
Ill
12
115
.8
.8

.2


.0
.4

.6
.9
0
0
-
4
22
15
13
7
25
2
3
.81*
.59*

.19*
.5
.9
.3
.29
.0
.24
.32
In sampling canister)
* AVERAGE VALUES AND STANDARD DEVIATIONS
CALCULATED FROM
FIRST
FOUR EVENTS
      Since both  primary and aeration basin air samples were  analyzed, the

                                     79

-------
data presented  in Table 9 can be factored to describe  air  emissions  for the
separate  processes.   Percent volatized in primary treatment  and  stripped in
the aeration  basin 1s shown in Table 9.


          TABLE  9.  AIR EMISSIONS FROM PRIMARY AND SECONDARY  TREATMENT
HAME OF COMPOUND

Acetona
Ha thy 1 ethyl katone
Cyclohexanona
Tatrahydrofuran
Carbon tetrachlorlda
1,1, 1-Tr Ichloroothana
1 , 1 ,2-Tr ichloroathana
Trlchlorotthyleni
Tatrachlorotthyltna
Chlorobanzana
Ethylbenzena
PERCENT VOLATILIZED
PRIMARY TREATMENT
1.5
0.46
-
3,3
12.4
12.4
5.0
10.2
11.6
8.8
8.2
STANDARD
DEVIATION
0.6
0.6
.
0.1
1.7
1.2
1.3
0.7
0.8
0.7
0.7
PERCENT STRIPPED
AERATION BASIN
0.3
N.D.
.
6.9
90.6
96.6
32.0
42.2
99.4
3.8
7.6
STANDARD
DEVIATION
0.06
N.A.
N,A.
3.6
20. 6
14.3
11. 8
8.2
23.2
2.1
3.6
(-) UNAVAILABLE DATA (Decomposed In sampling canlstar)
H.D. « NOT DETECTED         N.A. • NOT AVAILABLE
      Air  emissions from primary treatment were significant.   The  dominant
factors which  controlled the process were volatile compound properties
(Henry's law constant and octanol/water partition coefficient)  and the nature
of the air water Interface which 1s a function of process design parameters
and wind speed.   In the aeration basin, air emissions were largely controlled
by the gas to  liquid flow ratio and the magnitude of the biodegradation rate
constant under the acclimated conditions used in this study.

    •  The  removal  of volatiles by sorption and biodegradation  were also
assessed.  A summary of the removals obtained by sorption is shown 1n Table
10.

-------
        TABLE 10.  SUMMARY  OF  REMOVALS BY SORPTION (PRIMARY+SECONDARY)
                                  PERCENT SORBED
NAME OF COMPOUND
Acetone
Methyl ethyl ketone
Cyclohexanone
Tetrahydrofuran
Carbon tetrachlortde
1 , 1 , 1-Trichloroethane
1,1, 2-Tr iohloroethane
THchloroethylene
Tetrach loroethy lene
Chlorobenzene
Ethyl benzene
EVENT
1
0
1
2
2
1
1
1
0
1
1
2
.4
.0
.2
.0
.1
.0
.8
,9
.2
.4
.3
EVENT
2
0.9
0.6
-
1.9
0,9
2.3
2.5
-
4.4
3.4
5.2
EVENT
3
0.9
O.i
-
l.l
0.6
1.8
2.0
2.6
2.4
2.4
5.1
EVENT
4
1.0
0.8
-
1.1
0.8
1.4
1.8
1.9
2.4
2.0
2.9
EVENT
5
0.6
0.5
0.2
1.2
0.1
1.2
1.7
1.8
1.0
1.6
2.4
EVENT
6
0.1
0.3
-
0.5
0.4
1.8
2.3
1.1
2.3
2.7
4.7
EVENT
7
0.2
-
0.6
0.9
0,1
2.5
2.3
2.8
3.8
6.2
13.1
AVERAGE
0.
0,
I.
1,
0.
1.
2.
1.
2.
2.
5.
6
6
0
2
6
7
1
8
5
8
1
STANDARD
DEVIATION
0.32
0,23
0.88
0.49
0.38
0.52
0.30
0.71
1.17
1.53
3.46
(-) UNAVAILABLE DATA (Decomposed in sampling canister)
      Sorptlon of  volatiles by primary sludge and mixed-liquor solids
accounted for 0.5  -  5  percent of'the total removal in primary/activated  sludge
treatment.  The  sorption capacity of wastewater solids and sludges  for organic
compounds has been correlated with octanol-water partition coefficient.
Capacities must  be based on organic matter in the wastewater solids  or sludges
as measured by volatile solids (loss on ignition at 550*C) for predictive
purposes (1).

      The percent  biodegraded can be estimated from mass balance considera-
tions (% biodegraded - % total removal -  % sorbed - % volatilized  and
stripped).  Estimated  removals by the biodegradation mechanism are  presented
in Table 11.
                                     81

-------
                 TABLE 11.   SUMMARY OF ESTIMATED BIODEGRADATION
                            PERCENT 8IOOEGRAOEO (ESTIMATED)
NAME OF COMPOUND
Acetone
Methyl ethyl kctone
Cyclohexanone
Tetrahydrofuran
Carbon tetrachloHde
1.1.1-Trlchloroethane
1,1,2-Trlchloroethane
Trlchloroethylene
Tctrachloroethylene
Chlorobenzene
Ethylbenzene
EVENT
1
94.1
92.2
92.8
37.1
-19.6
-25.2
41.3
36.0
-28.0
82.8
79.9
EVENT
2
92.0
92.9
-
78.4
-31.4
-41.7
34.7
-
-66.3
80.2
66.4
EVENT
3
93.3
95.1
-
82.1
37.0
2.6
43.3
46.1
14.4
85.9
76.1
EVENT
4
93.4
93.2
92.4
85.0
-27.7
-13.7
29.1
41.0
-12.8
87.9
83.0
EVENT
5
92.8
96.7
76.8
86.6
-7.9
-8.3
25.2
44.3
-10.0
87.7
85.6
EVENT
6
97.6
97.5
-
86.8
3.4
-7.9
25.1
47.4
-24.1
81.4
79.1
EVENT
7
76.8
-
66.3
93.0
14.8
11.2
23.3
58.0
14.4
79.6
70.5
AVERAGE
91.
94.
82.
78.
-4.
-11.
31.
45.
-16.
83.
77.
4
6
1
4
5
9
7
5
0
6
2
STANDARD
DEVIATION
6.
2.
11.
18.
22.
16.
7.
6.
25.
3.
6.
2
0
2
8
9
2
6
7
7
2
3
(•) X BIODEGRADEO - X TOTAL REMOVAL - X SOR8ED - X STRIPPED
(-) UNAVAILABLE DATA (Decomposed in sampling canister)
      Polar  solvents (acetone, methyl ethyl ketone, cyclohexanone,  and
tetrahydrofuran)  were removed primarily by biodegradation.  Chlorinated
aliphatic  solvents  showed little or no biodegradation (except  for
trichloroethylene).   Volatilization and stripping to the atmosphere was  the
dominant removal  mechanism.  Aromatic solvents were removed primarily by
biodegradation,  however,  volatilization and stripping accounted  for
approximately  15 percent  of the total removal.  Negative values  for %
biodegradation can  be the result of analytical error in sample analyses  or can
be due to  metabolism which produces the compound as a product  of degradation
of another component in the spike mixture or in the wastewater feed.


                            SUMMARY AND  CONCLUSIONS

      The  eleven compounds spiked Into the activated sludge system  did not
cause any  adverse effect  on plant performance including nitrification.
Sorption accounted  for 0.5 - 5% of the removal of volatiles.   Total removal
was greater  than 94% for  eight of the eleven compounds studied.  The minimum
removal was  70%  for 1,1,2-trichloroethane.

      Volatilization and  stripping was the primary removal mechanism for
carbon tetrachloride, 1,1,1-trichloroethane, and tetrachloroethylene.  Polar
solvents were  biodegraded while the two aromatic solvents  showed intermediate
biodegradation with substantial (15%) emissions to the atmosphere.
                                    82

-------
                                  REFERENCES


1.  Dobbs, R.A., Wang, L. and Govind, R.  Sorption of Toxic Organic Compounds
    on Wastewater Solids: Correlation with Fundamental Properties.  Environ.
    Sci. Technol. 23 (9): 1092, (1989).

2.  Engelbrecht, R.S., Gaudy, A.F., and Cederstrand, J.M.  Diffused air
    stripping of volatile components of petrochemical wastes.  J. Water Poll.
    Control Fed. 33: 127, 1961.

3.  Gaudy, A.F., and Engelbrecht, R.S.  Stripping kinetics of volatile
    compounds of petrochemical wastes.  J. Water Poll. Control Fed. 33: 382,
    1961.

4.  Gaudy, A.F., Turner, B.G., and Presztoszeri, S.  Biological treatment of
    volatile waste components.  J. Water Poll. Control Fed. 35: 75, 1963.

5.  Thibodeaux, L.J., and Parker, D.G.  Desorption limits of selected
    industrial gases and liquids from aerated basins.  AICHE Symp. Series 156,
    72: 424, 1976.

6.  Dilllng, W.L.  Interphase transfer processes II.  Evaporation of chloro-
    methanes, ethanes, ethylenes, propanes, and propylenes from dilute aqueous
    solutions.  Comparisons with theoretical predictions.  Environ. Sci.
    Technol. 11: 405, 1977.

7.  Thibodeaux, L.J., and Millican, J.D.  Quantity and relative desorption
    rates of air strippable organics in industrial wastewater.  Environ. Sci.
    Technol. 11: 879, 1977.

8.  Mackay, D., and Wolkoff, A.W.  Rate of evaporation of low-solubility
    contaminants from water bodies to atmosphere.  Environ. Sci. Technol. 7:
    611, 1973.

9.  Mackay, D., Shiu, W.Y., and Sutherland, R.P.  Determination of air-water
    Henry's Law Constants for hydrophobic pollutants.  Environ. Sci. Technol.
    13: 133, 1979.

10. Matter-Muller, C., Gujer, W., and Stumm, W.  Non-biological elimination
    mechanisms in a biological sewage treatment plant.  Program Water Technol.
    11: 299, 1980.

11. Matter-Muller, C., Gujer, C., and Giger, W.  Transfer of volatile
    substances from water to the atmosphere.  Water Res. 15: 1271, 1981.

12. Smith, J.H., and Bomberger, D.C.  Prediction of volatilization rates of
    chemicals in water,  in.:  B.K. Afghan and D. Mackay (ed.), Hydrocarbons
    and Halogenated Hydrocarbons in Aquatic Environment.  Plenum Press, New
    York, 1980.  p. 445-451.


                                     83

-------
13. Freeman. R.A., Schroy J.M., Klieve, J.R,, and Archer, M.C.  Experimental
    studies on the rate of air stripping of hazardous chemicals from waste
    treatment systems.  73rd Annual APCA Meeting, Montreal, Canada, June 22-
    27, 1980.  Reprint Number 80 - 16.7.

14. Freeman, R.A.  Secondary emissions from subsurface aerated treatment
    systems.  Environ. Progress. 1(2): 118, 1982.

IS. Stover, E.L., and Kincannon, D.F.  Biological treatability of specific
    organic compounds found in chemical industry wastewaters.  In:
    Proceedings of the Thirty-Sixth Purdue Industrial Waste Conference.  West
    Lafayette, Indiana, 1981.

16. Kincannon, D.F., Stover, E.L., Nichols, V., and Medley, D.  Removal
    mechanisms for toxic priority pollutants.  J. Water Poll. Control Fed.
    55(2): 157, 1983.

17. Smith, J.H., Bomberger, D.C., Jr., and Haynes, D.L.  Prediction of the
    volatilization rates of high-volatility chemicals from natural water
    bodies.  Environ. Sci. Techno!. 14: 1332, 1980.

18. Smith, J.H., Bomberger, D.C., Jr., and Haynes, D.L.  Volatilization rates
    of intermediate and low volatility chemicals in water.  Chemosphere 10:
    281, 1981.

19. Smith, J.H., and Bomberger, D.C., Jr.,   Prediction of volatilization
    rates of chemicals in water.  AICHE J. 75: 375, 1979.

20. Lurker, P.A., Clark, S., and Elia, V.J.  Atmospheric release of
    chlorinated organic compounds from the activated sludge process.  J. Water
    Poll. Control. Fed. 54: 1566, 1982.

21. Kobayashi, H., and Rittman, B.E.  Microbial removal of hazardous organic
    compounds.  Environ. Sci. Techno!. 16: 170A, 1982.

22. Thibodeaux, L.G., Parker, D.6., and Heck, H.H.  Measurement of volatile
    emissions from wastewater basins.  EPA 600/2-82-095, U.S. Environmental
    Protection Agency, 1982.

23. Roberts, P.V., and Dandliker, P.6.  Mass transfer of volatile organic
    contaminants from aqueous solution to the atmosphere during surface
    aeration.  Environ. Sci. Techno!. 17: 484, 1983.

24. Roberts, P.V., Munz, C., and Dandliker, P.G.  Modeling volatile organic
    solute removal by surface and bubble aeration.  J. Water Poll. Control
    Fed. 56(2): 157, 1984.

25. Mackay, D., and Yeun, A.T.K.  Mass transfer coefficient correlations for
    volatilization of organic solutes from water.  Environ. Sci. Techno!.
    17(4): 211, 1983.

-------
26. Truong, K.N., and Blackburn, J.W.  The stripping of organic chemicals In
    biological treatment proctssts.  Environ. Progress 3(3): 143, 1984.

27. Lurkir, P.A., Clark, C.S., Ella, V.J., Gartslde, P.S., and Klnman, R.N.
    Aerial organic chemical release from activated sludge.  Water Res. 18;
    489, 1984.

28. Namkung, £., and Rittman, B.E.  Estimating volatile organic compound
    emissions from publicly owned treatment works.  J. Water foil. Control
    Fed. 59(7): 670, 1987.

29. Barton, D.A.  Intermedia transport of organic compounds 1m biological
    wastewater treatment processes.  Environ. Progress 6(4): 246, 1987.

30. Dunovant, V.S., Clark, C.S., QueHee, S.S., Hertiberg, V.S., and Trapp,
    J.H.  Volatile organlcs 1n the wastewater and airspaces of three
    wastewater treatment plants.  J. Water Poll. Control Fed. 58(9): 886,
    1986.

31. Weber, W.J., and Jones, B.E.  Toxic substance removal in activated sludge
    and PAC treatment systems.  EPA 600/2-86-Q4S, U.S. Environmental
    Protection Agency, Cincinnati, Ohio, 1986.

32. Mtlcer, H,, Thompson, D., Bell, «3., and Montelth, H.  Stripping of
    volatile organic compounds at municipal wastewater treatment plants.
    Presented at the AWMA/EPA International Symposium on Hazardous Waste
    Treatment: Blosystems for Pollution Control, Cincinnati, Ohio, February
    20-23, 1989.

33. Boyle, W.C., Chairman, American Society of C1v1l Engineers;, Oxygen
    Transfer Standards Subcommittee,  Development of Standard Procedures for
    Evaluating Oxygen Transfer Devices.  EPA 600/2-83-102, U.S. Environmental
    Protection Agency, Cincinnati, Ohio, 1983.

34. Redmon, D., Boyle, W.C., and Ewlng, L.  Oxygen transfer efficiency
    measurements 1n mixed liquor using off-gas techniques.  J, Water Poll.
    Control Fed. 55(11): 1338, 1983.

35. Metcalf & Eddy, Inc.  Wastewater engineering treatment, disposal and
    reuse.  G. Tchobanoglous (ed.), 2nd edition.  McGraw-Hill Book Co., New
    York, 1979.  p.274.

36. Nancy, K.H., and Okun, D.A.  Effects of surface active agents on bubble
    aeration.  J. Water Poll. Control Fed. 32: 351, 1960.

37. McKeown, J.J., and Okun, D.A.  Effects of surface active agents on oxygen
    bubble characteristics,  la: Proceedings Third Biological Waste Treatment
    Conference. Manhattan  College, 1960.

38, Kolde, K., HayasM, T., Sumlno, K., and Iwamoto, S.  Mass transfer from
   'single bubbles 1n aqueous solutions of surfactants.  Chem, Eng. Sc1. 31:
    963, 1976.
                                    IS

-------
39. Eckenfelder, W.W.* Jr., and Barnhart, E.L.  The effect of organic
    substances on the transfer of oxygen from air bubbles in water.  Am.  Ind.
    Chem. Eng. J. 7: 7631, 1961.

40. Gurol, M.D., and Nekovinaini, S.  Effect of organic substances on mass
    transfer in bubble aeration.  J. Water Poll. Control Fed. 57(3): 235,
    1985.

41. Alexander M.  Biodegradation of chemicals of environmental concern.
    Science 211: 132,1981.

42. Wood, P.R., Parsons, F.Z., Demarco, J., Harveen, H.J'., Lang, R.F., Payan,
    I.L., Meyer, L.M., Ruiz, M.C., and Ravelo, E.D.  Introductory study of the
    blodegradation of the chlorinated methane, ethane, and ethene compounds.
    Cooperative Agreement Number CR806890-01, U.S. Environmental Protection
    Agency, Cincinnati, Ohio, 1980.

43. Liss, P.S., and Slater, P.G.  Flux of gases across the air-sea interface.
    Nature 247: 181, 1974.

44. Mackay, D., and Leinonen, P.J.  Rate of evaporation of low-solubility
    contaminants from water bodies to atmosphere.  Environ. Sci. Technol. 9:
    1178, 1975.

45. Hannah, S.A., Austern, B.M., "Eralp, A.E., and Wise, R.H.  Comparative
    removal of toxic pollutants by six wastewater treatment processes.  J.
    Water Poll. Control Fed. 58(1): 27 (1986).

46. Hannah, S.A., Austern, B.M., Eralp, A.E., and Dobbs, R.A.  Removal of
    organic toxic pollutants by trickling filter and activated sludge.  J.
    Water-Poll. Control Fed. 60(7): 1281 (1988).

47. Bhattacharya, S.K., Angara, R.V.R., Hannah, S.A., Bishop, D.F., Jr.,
    Dobbs, R.A., and Austern, B.M.  Fate and effects of selected RCRA and
    CERCLA compounds in activated sludge systems.  Presented at EPA Fifteenth
    Annual Research Symposium on Remedial Action, Treatment and Disposal of
    Hazardous Wastes, Cincinnati, Ohio.  April 10-12, 1989.

48. Compendium Method TO-14,  The Determination  of Volatile  Organic  Compounds
    (VOCs) in Ambient Air Using SUMMA Canister Sampling  and  Gas  Chromatograph
    Analysis.  Quality Assurance Division,  Environmental Monitoring Systems
    Laboratory.  Environmental  Protection Agency,  Research  Triangle Park,  NC
    27711.  May 1988.

49. Federal Register. Guidelines establishing test procedures for the analysis
    of pollutants under the clean water act; 40 CFR, 136, 49_, No. 209, 1984.

50. Standard Methods for the Examination of Water and Wastewater, 16th
    Edition, APHA-AWWA-WPCF (1985).
                                     86

-------
                    FIELD ASSESSMENT OF AIR EMISSIONS FROM
                    HAZARDOUSWASTE PEWATERING  OPERATIONS

                     by:  Thomas C.  Ponder, Jr., PE, CCE
                          Cynthia J. Bishop
                          PEI Associates, Inc.
                          Arlington, Texas 76012
                                   ABSTRACT
    Petroleum companies generate millions of tons of sludge every year from
American Petroleum Institute (API) separators, Dissolved Air Floatation (DAF)
float, and biological waste.  This sludge is often dewatered to reduce the
volume of solid waste that will require disposal.  Since the sludge is derived
from petroleum products, a variety of volatile and semi-volatile organics can
be contained in the sludge mixture.  Sludges are often heated and dewatering
equipment is open to the atmosphere, therefore increasing the potential for
the release of organics to the air from the dewatering process.  Organic con-
centrations as high as 5,000 ppm have been found at the vents from the dewater-
ing  process.  This paper presents the results of a comprehensive study that
was performed for the EPA to quantify the release of volatile and semi-vola-
tile organics from sludge dewatering operations.  The study included field
tests of three dewatering operations in which material balance was used to
calculate air releases.

    This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
                                     87

-------
                                  INTRODUCTION


    Sludge dewatering  is  a process used by  industry to reduce the volume of
waste before disposal.  The sludge dewatering process involves extracting the
liquid out of the sludge.  After  an exhaustive study of dewatering operations,
it was found that only sludge dewatering operations at petroleum refineries
contained concentrations  of organic material over 500 ppm  (1).  These petro-
leum refinery waste sludges contain high concentrations of oil and organic
chemicals and include  API separator sludge, DAF float, and biological sludge  .
Both DAF and API sludge are considered hazardous waste (K048 and K051, respec-
tively) .  Dewatering of refinery  sludge may be a significant source of vola-
tile organic emissions since few  refineries utilize volatile organic emissions
collection and/or control devices.

    The U.S. Environmental Protection Agency (EPA) is currently collecting
information to develop standards  as necessary for the control of air emissions
from hazardous waste treatment, storage, and disposal facilities (TSDFs).
These field tests were conducted  to aid in the development of standards for
hazardous waste dewatering operations.  The main objectives of this study were
to:

    * Determine the percent of volatile and semi-volatile  organic
      compounds removed from the  sludge during dewatering, and

    * Determine the air emissions of volatile and semi-volatile
      organic compounds from the  dewatering process.

    This paper provides general information on the dewatering process and the
types of dewatering methods currently in use, discusses the dewatering opera-
tions at the three refineries tested, and discusses the procedures and methods
used to determine air  emissions from the dewatering process.

                              DEWATERING METHODS
    As mentioned earlier, the objective of sludge dewatering is to reduce the
amount of liquid in the sludge, therefore, reducing the volume of waste that
must be disposed.  The input or sludge to the dewatering device Is called the
feed.  The liquid removed from the feed (or sludge) is called the filtrate.
The final product of dewatering is called the cake.  There are four generally
accepted ways to dewater sludge;
    * centrifuge;
    * vacuum filtration;
    * plate and frame filtration; and
    * belt press filtration.
    A centrifuge accelerates sedimentation through the use of centrifugal
force.  In a centrifuge, a rotating bowl acts as a highly effective settling
tank.  The most common type of centrifuge used in sludge dewatering is the
imperforate basket centrifuge which handles sludges with high solids content.
Centrifuges are not used as readily as other dewatering devices due to high
maintenance costs and problems with separating suspended solids,
                                     08

-------
    Vacuum filters use atmospheric pressure as the driving force which causes
the liquid phase to move through a porous media and separate from the solids.
Vacuum filtration consists of a large cylindrical drum which rotates through a
vat containing sludge.  The drum rotates through three zones.  In the cake
forming zone, a vacuum is applied to the submerged section of the drum which
causes the filtrate to pass through the porous surface media and cake to form
on the surface of the drum.  As the drum rotates, it is carried to the cake
drying zone.  This zone is also under vacuum and further dries the cake.  As
the drum rotates further, it is carried into the cake discharge zone where the
vacuum is removed and the cake is scraped off the drum.  The last few years
have seen a decline in the use of the vacuum filter due to improvements in
other dewatering methods which make them more economical and technically fea-
sible.

    Plate and frame filters use fluid pressure generated by pumping sludge
into the unit as the driving force to separate solids from liquids.  Plate and
frame filters (or recessed plate pressure filters) are constructed from a
series of recessed plates.  As more sludge is pumped into the filter, the
pressure increases causing the filtrate to pass through the filter cloth leav-
ing the cake.  Plate and frame filters are used for sludges of poor dewater-
ability or for cases where a high solids content cake is required.   Plate and
frame filters are operated on a batch process and do not run continuously.

    Belt filter presses employ single and double moving belts to dewater slu-
dges continuously.  Any belt filtration process includes three basic opera-
tional stages:  chemical conditioning of the feed, gravity drainage to a non-
fluid consistency, and compaction of the pre-dewatered sludge.  A majority of
refineries dewatering waste use the belt filter press because of its ability
to handle high sludge throughput capacities.

              DEWATERING PROCESSES FOR THREE SELECTED FACILITIES
    Five facilities across the country were visited to observe and review
their sludge dewatering processes.  The following three sites were selected
for testing because the sludge feed contained a significant concentration of
organics:

    0 Sun Refining and Marketing Company - Toledo, Ohio
    0 British Petroleum Refinery - Lima, Ohio
    ° Sun Refining and Marketing Company - Tulsa, Oklahoma

SUN REFINING AND MARKETING COMPANY - TOLEDO, OHIO

    Sun Refining and Marketing Company in Toledo, Ohio (Sun-Toledo) has two
processing units with a total throughput of 120,000 barrels per day (1).  As a
result of their production facilities, they produce API bottoms from the API
separator, DAF float from the air flotation unit, and biological sludge.  The
filter cake is landfilled and the filtrate is sent to the wastewater treatment
plant.

    Sun-Toledo has one belt filter press which processes approximately 20,000
gallon per day (gpd) of sludge and generates approximately 18 tons per operat-
ing day of cake.  The press operates 4 days per week.   API and DAF sludges are
mixed and dewatered in the press.  Biological sludges are also clewatered in
this press, but are kept separate from the API and DAF wastes.
                                      89

-------
    The sludges  go  through a series of mixing and. treatment processes before
being dewatered.  The  API and DAF sludges  are mixed in two  large  1-millon gal-
lon tanks where  the sludge is steam heated to 100°F.   Large capacity pumps
agitate the  sludge  and transfer it to  the  oily waste  feed tank.   Once in the
oily waste feed  tank,  the sludge is heated to 140°F and pumped to the belt
filter press.  A conditioning polymer  is added to the sludge before  it is
pressed.  The belt  filter press is fed at  approximately 35  gallons per minute.
During filtration,  the belt is backwashed  with water.   The  filter cake from
the belt press filter  is  then conveyed to  the cake storage  bins,  after which
it is landfilled.   The filtrate is collected and sent to the waste water
treatment plant.  A side  draft hood, located on the wall, ducts air  to a car-
bon filter and exhaust fan on the roof.  Figure 1 illustrates the dewatering
process.
                                               WASHWATIR
                                                POLYMEfl
                  >• SAMPLE FONT
         Figure 1.  Dewatering Process at Sun Refining - Toledo, Ohio,
BRITISH PETROLEUM REFINERY - LIMA, OHIO

    British Petroleum Refinery  in Lima, Ohio  (BP-Lima) produces  gasoline,
diesel fuel, jet fuel, kerosene,  lube oil, and benzene (2).  As  a result of
their production facilities, they produce API bottoms from the API separator,
DAF float from the air flotation  unit, and biological sludge which must be
dewatered to reduce its volume  before disposal.  The filter cake is disposed
of on a land farm, and filtrate is sent to the waste water treatment plant.

    BP-Lima uses a. belt filter  press for its dewatering operations.  The
press operates 24 hours per day,  7 days per week.  About 40 continuous hours
of this is spent dewatering biological sludge from the clarifier bottoms.
During the balance of the  time, oily sludge from the API separators and DAF
float is dewatered.  This  sludge  contains about 1-3 percent solids and has a
pH of 6.5.  The sludge is  preheated to approximately 130°F and a conditioner
polymer is added before being dewatered.  As the sludge is dewatered, the
filtrate drains from the press  into a drainage hole below the filter.  The
filter cake is conveyed to a cake bin.  Trucks transport the cake to the land-
farm.  Wash water is sprayed on the filter belts as they return  to the front
of the press.  Air emissions from the process are vented through a stack in
the roof.  Figure 2 illustrates the general process flow for the dewatering
unit.
                                      90

-------
                                              WASHWATER

                                               POLYMER
          Figure 2.  Dewatering Process at BP Refinery - Lima, Ohio.

SUN REFINING AND MARKETING COMPANY - TULSA, OKLAHOMA

    Sun Refining and Marketing Company (Sun-Tulsa) in addition to producing
typical fuel products produces benzene, toluene, propylene, and cyclohexane
(3).  As a result of their production facilities, they produce API bottoms
from the API separator, DAF float from the air flotation unit, and biological
sludge which must be dewatered to reduce its volume before disposal.  The
filter cake is disposed of to an onsite land farm and the filtrate is sent to
an oil-water separator to recover the oil.

    Sun-Tulsa has two belt presses housed in the refinery's belt press build-
ing.  Each belt press is 2.2 meters wide and operates on an as needed basis
which is approximately one shift per day.  These presses were installed in
1986 to replace centrifuges.  The centrifuges were replaced because solids in
the sludges were reduced to fine particles, leading to Total Suspended Solids
problems.  In addition, down time was excessive.

    The sludge dewatering process is shown in Figure 3.  The sludge charge is
stored in one of two 2,015-barrel charge tanks where it is heated before being
pumped to the belt press.  A polymer solution, Amerfloc Plus 5485, is injected
from a 4,000 gallon tank into the feed line.  At the belt press,  the sludge is
sprayed with wash water to aid in filtration.  The filtrate exits the bottom
of the belt press through a drain and flows to an oil-water separator.  The
filter cake is then conveyed to a cake bin and transported to the landfarm by
trucks.  Air emissions from each press are vented to a carbon filter and ex-
haust fan on the roof.
                          X - SAMPtC TOWT
             Figure  3.   Dewatering Process  at  Sun-Tulsa, Oklahoma.
                                     91

-------
                          TESTING METHODS AND RESULTS
     As discussed previously,  the objectives of the field tests  were  to  deter-
mine the  percent of volatile  and semi-volatile organics  removed from the  slud-
ges  during dewatering and to  estimate the  air emissions  during  the dewatering
process.   To meet these objectives,  samples of the feed,  filtrate, wash water,
polymer,  and cake were collected and analyzed for  volatile  and  semi-volatile
organic compound concentrations.   Flow rates for all  inlet  and  outlet streams
were measured "by reading flow meters or collecting a  known  quantity  of  materi-
al for a  fixed  time period.   Biological sludge was not sampled  because  of its
typically low concentration of organic compounds.   Figures  1  through 3  show
the  sample locations at each  refinery.

     At the BP-Lima and Sun-Tulsa facilities, three 3-hour runs  were  conduct-
ed.   At the Sun-Toledo facility,  two 6-hour runs were conducted:  one with API
feed and  one with DAF feed.   Samples were  collected and  flow  rates were mea-
sured at  half-hour intervals  during  each run.   Table  1 shows  the measured feed
rates for each  of the field tests.

       TABLE 1.   MASS FLOW BATES FOR  INFLUENT AND EFFLUENT STREAMS (Ib/h)

SUN-TOLEDO
API test
DAF test
BP-LIMA
Run 1
Eun 2
Run 3
Average
Std. Dev.
SUN-TULSA
Run 1
Run 2
Run 3
Average
Std. Dev.
Feed

20,820
8,110

41,718
40,410
43 , 104
41,744
1,100

18,402
25,716
20,466
21,528
3,079
Polymer

823
971

442
1,529
332
768
540

4,252
3,345
3,204
3,600
464
Wash
water

16,127
15,938

13,032
12,726
12,960
12,906
131

49,758
49,818
50,676
50,084
419
Cake

2,617
1,135

1,188
576
852
872
250

1,860
300
1,800
1,320
722
Filtrate

..
..

68,914
34,358
44,138
49,137
14,543

27,571
39,127
32,327
33,008
4,742
Samples of the feed and wash water streams were collected from valve openings.
The polymer was sampled by scooping the polymer solution from its feed tank.
The filter cake was collected from the conveyor at the outlet of the filter
press.  The filtrate was collected by holding a large sample jar under the
belt press where the filtrate drains.  Homogeneous filtrate samples were dif-
ficult to sample and analyze because the oily mixture would immediately begin
to separate.  At the Sun-Toledo facility, the filtrate could not be measured
                                     92

-------
because the flow was too widespread across the belt to collect and measure.
The filtrate at the BP-Lima facility was also difficult to measure.  As a
result lithium chloride was pumped into the filtrate and sampled at  the inlet
and outlet points.  The ratio of the concentrations determined the filtrate
flow rate.

    Stack tests were conducted on the exhaust from the dewatering process.
Volatile organics were collected from the exhaust stream with charcoal sorb-
ents.  Semi-volatile organics were collected using the modified Method 5 train
with an XAD-2 sorbent.  Total hydrocarbons were measured with a flame ioniza-
tion analyzer.  All analytical methods were taken from the third edition of
SW-846.  Method 8240 was followed for all volatile organic analyses, and Meth-
od 8270 was used for all semi-volatile organic analyses.  Both methods employ
gas chromatography and mass spectroscopy (GC/MS) techniques.

    Tables 2 and 3 show the average mass flow rates of the volatile  and semi-
volatile organics, respectively, in the influent stream (feed, wash  water, and
polymer) and out the stack for each facility.  At the Sun-Toledo facility,
only one run was conducted on each sludge type, DAF and API, therefore stan-
dard deviation calculations are not applicable.

             TABLE 2.   MASS FLOW RATES OF VOLATILE ORGANICS (Ib/h)

Xylenes
1,2,3-
Trimethyl
benzene
1,3,5-
Trimethyl
hexanc
Ethylbenzene
Methyl
eyclohexane
n- Heptane
n-Hexane
Toluene
Benzene
Ethyl toluene
Sun-Toledo
Feed
API
17
NT
NT
2.3
10
12
NT
15
2.1
2.8
DAF
7.5
NT
NT
0.99
4.8
5.2
NT
4.8
0.68
1.7
Stack
API
ND
NT
NT
0.40
2.8
3.8
NT
3.7
0.80
0.17
DAF
ND
NT
NT
0.14
1.1
1.4
NT
0.81
ND
0.10
BP-Lima
Feed
Avg.
17
4.01
NT
3.3
4.5
4.1
NT
12
1.4
10
Std.
Dev.
8.8
1.8
NT
1.9
2.2
2.0
NT
6.5
0.80
4.9
Stack
Avg.
0.80
0.06
NT
0.20
0.94
1.3
NT
1.4
0.29
0.24
Std.
Dev.
0.14
0.01
NT
0.04
0.33
0.47
NT
0.43
0.09
0.04
Sun-Tulsa
Feed
Avg.
9.6
3.5
2.7
NT
1.8
3.8
3.5
9.2
1.6
7.5
Std.
Dev.
2.7
0.28
2.0
NT
1.8
3.0
3.8
6.8
1.6
0.43
Stack
Avg.
1.6
0.20
ND
NT
2.0
3.3
4.3
3.3
0.35
0.45
Std.
Dev.
0.43
0.00
*™
NT
1.3
2.8
5.1
2.5
0.35
0.07
 NO * Not detected.
 Ill - Not tested.
                                      13

-------
          TABLE 3.   MASS FLOW RATES FOR SEMI-VOLATILE ORGANICS  (Ib/h)

Naphthalene
2-Hethyt
ntphth*ltne
Fluorcrw
Minwnthrene
fyrfne
n-Hltrojodl-
phenylMlne
Dlberuofuran
Chrystne
1.3,5*Trl-
Btthylbcnxent
Sun-Toledo
Feed
API
0.23
NT
HT
HT
HT
NT
NT
0.040
0.75
DAF
O.EO
HT
KT
HT
HT
HT
HT
0.030
0.57
Stack
API
o.oto
HT
HT
HT
HT
HT
NT
NO
0.090
DAF
0.10
HT
HT
HT
HI
NT
HT
NO
0.090
BP-UiM
Feed
Avg.
4.6
14
1.8
6.1
3.7
NO
HT
NT
NT
Std.
Dev.
H.O
6.6
0.90
2.7
1.5
•-
NT
NT
NT
Stack
Avg.
0,011
O.OOS5
9.0E-5
7.0i-5
1.0E-5
NO
NT
HT
NT
Std.
Dev.
0.0022
0.0018
1,06-5
1.06-5
1.06-5
--
NT
NT
NT
Sun-Tulsa
feed
Avg.
6.0
23
1.7
3.7
HT
7.9
ND
NT
NT
Std.
Dev.
0.42
3.54
0.35
0.57
NT
1.6
--
NT
NT
Stick
Avg.
0.09
0.17
<0.01
<0.01
HT
NO
<0.01
HT
NT
Std.
Dev.
0.014
0.042
--
-
NT
—
-
NT
NT
  HT • Not tested.
  MO - Not detected.
    Volatile and semi-volatile  organic emission factors,  calculated from the
data in Tables 2 and 3,  are  shown in Tables  4 and 5.   The emission factor for
each chemical is the ratio of the flow rate  out the stack to the flow rate in
the feed.  This represents the  fraction of each organic that would be released
to the air.  For example, the emission factor for a^ylenes at the BP-Lima
facility was calculated  as 0.046  Ib/lb feed.  Therefore,  4.6 percent of the
xylene that was fed to the belt press was  released out the stack.
       TABLE 4.   AIR EMISSION FACTORS FOR VOLATILE ORGANICS  (Ib/lb feed)

Xylenes
1,2,3-Tri-
methylbenzene
1,3,5-Tri-
me thy Ihexane
Ethylbenzene
Methyleyelo-
hexane
n-Heptane
n-Hexane
Toluene
Benzene
Ethyl toluene
Sun-Toledo
API
ND
NT
NT
0.17
0.28
0.32
NT
0.25
0.38
0.061
DAF
ND
NT
NT
0.14
0.23
0.28
NT
0.17
ND
0.059
BP-Lima
0.046
0.015
NT
0.061
0.21
0.31
NT
0.12
0.21
0.024
Sun-Tulsa
0.19
0.071
ND
NT
0.78
0.92
1.3
0.39
0.26
0.073
           detected.
  NT - Not tested.
94

-------
  TABLE 5.  AIR EMISSION FACTORS FOR SEMI-VOLATILE ORGANICS (lb/lb feed)

Naphthalene
2 - Me thy 1-
naphthalene
Fluorene
Phenanthrene
Pyrene
n-Nitrosodl-
phenylamine
Dibenzofuran
Chrysene
1,3,5-Tri-
methylbenzene
Sun-Toledo
API
0.044
NT
NT
NT
NT
NT

NT
ND
0.12
DAF
0.50
NT
NT
NT
NT
NT

NT
ND
0.16
BP-Lima
0.0024
0.00057
0,000050
ND
ND
ND

NT
NT
NT
Sun-Tulsa
0.018
0.0087
0,00057
0.00017
ND
ND

NT
NT
NT
ND = Not detected,
NT = Not tested.

-------
                                    SUMMARY
    Dewatering is used by industry to reduce the volume of hazardous waste
sludges before disposal.  Common dewatering equipment include the centrifuge,
the vacuum filter, the plate and frame filter, and the belt press filter.
Based on an industry survey, the belt press filter is the most common
dewatering device.

    During hazardous waste dewatering, organics may be released to the air.
In order to quantify these emissions, PEI conducted field tests at three
petroleum refineries that use belt presses to dewater API and DAF sludges.
Samples of the influent and effluent streams were taken and analyzed for
volatile and semi-volatile compounds.  Stack tests were also conducted on the
exhaust gas.

    The results of the field tests were used to determine the organics mass
flows in the influent and effluent streams.  Some major points from the
results are as follows:

    The highest influent mass flow of volatile and semi-volatile organics in
    the feed streams tested were 17 Ib/h for xylenes and 15 Ib/h for 2-
    methylnaphthalene.

    The highest air emission rates for the volatile organics tested were 3.8
    Ib/h for n-heptane and 3.7 Ib/h for toluene.

    The highest emission rate for the semi-volatiles tested was 0.13 Ib/h for
    2-methylnaphthalene.

    The tested organics with the highest air emission factors (Ib released to
    the air/lb feed) were n-hexane, n-heptane, and naphthalene.
                                   REFERENCES


 1,   PEI  Associates,  Inc.   Field Test and Evaluation  of a  Sludge Dewatering
     Operation at A Petroleum Refinery.   Prepared  for the  U.S. Environmental
     Protection Agency Office of Research and Development  Hazardous Waste
     Engineering  Laboratory,  Contract No.  68-02-3995,  May  1988.

 2.   PEI  Associates,  Inc.   Field Evaluation of a Sludge Dewatering Unit at BP
     Refinery,  Lima,  Ohio.   Prepared for  the U.S.  Environmental Protection
     Agency Office of Research and Development Risk Reduction Engineering
     Laboratory,  Contract No.  68-02-4284,  February 1990.

 3,   PEI  Associates,  Inc.   Field Evaluation of a Sludge Dewatering Unit at Sun
     Oil  Refinery,  Tulsa, Oklahoma.   Prepared for  the U.S. Environmental
     Protection Agency Office of Research and Development  Risk Reduction
     Engineering  Laboratory,  Contract No.  68-02-4284,  February 1990.

                                     96

-------
                       BENCH-SCALE WET AIR OXIDATION OF
                       COMPLEXED  METAL CYANIDE SLUDGES
                               RONALD J. TURNER
                     U.S. ENVIRONHENTAL PROTECTION AGENCY
                     RISK REDUCTION  ENGINEERING  LABORATORY
                               CINCINNATI, OHIO
                                   ABSTRACT
      Certain aqueous and solid hazardous wastes from metal finishing and
electroplating operations contain metal-cyanide complexes which must receive
treatment to significantly reduce the cyanide concentrations prior to final
disposal.  The applicable processes for treating the cyanide wastes include
alkaline chlorination, electrolytic oxidation, ultraviolet light/ozone,
hydrolysis, and wet-air oxidation (WAO),  Resource Conservation and Recovery
Act (RCRA) wastes F006, F007, and F019 were treated by wet-air oxidation to
determine the degree of cyanide destruction as measured by the total and
amenable cyanide concentrations in the oxidized effluent and filtered  solids.

INTRODUCTION

      Wet-air oxidation (WAO) is the liquid phase oxidation of organics or
oxidizable inorganic compounds at elevated temperatures and pressures.  The
oxidation is brought about by combining a wastewater or dilute sludge with a
gaseous source of oxygen (air) at temperatures of about 175 to 327 degrees
Centigrade (347* to 620" F) and pressures of about 2069 to 20,690 kPa (300 to
3000 psig).

      Wet-air oxidation has been demonstrated at bench-scale, pilot-scale, and
full-scale as a technology capable of breaking down hazardous compounds to
carbon dioxide and other less toxic end products.  The cyanides in
electroplating wastes are converted to carbonate and ammonium ions when
oxidized as shown by the following reaction:

                   2NaCN + 02  +  4H20  -*  Na2C03  + (NH4)2C03

      High cyanide wastestreams (greater than 1%) have been successfully
treated with WAO and destructions of more than 99% are typical. (1) This paper
presents the results obtained from bench-scale oxidation testing to provide
data for three RCRA wastes;
                                      9?

-------
      F006   -    wastewater treatment sludges from electroplating operations

      F007   -    spent cyanide plating bath solutions

      F019   -    wastewater treatment sludges from the chemical conversion
                  coating of aluminum

CYANIDE TERMS

      The characterization of cyanide species by ion structure include simple
and complex cyanides.  In solution, the simple cyanide compounds dissociate
into the cyanide anion, and the complex cyanides dissociate into a cyanide-
metal anion.  Another category of cyanide is called "free cyanide", which is
the sum of the cyanide anions and hydrogen cyanide in solution.

      Amenable cyanide refers to the cyanide species removable by alkaline
chlorination.  This procedure is described in SW-846 Methods 9010 and 9012.
(2) The total cyanide generally includes all cyanide groups, free, simple and
cyanide-metal complexes, except cobalt, gold and some of the platinum group
metals.

WASTE CHARACTERIZATIONS OF WAO TEST MATERIALS

      The F006 filter cake was generated by treatment of spent solutions from
barrel and rack plating lines dedicated to nickel-chromium, nickel-copper,
cadmium, and chrome plating.  The treatment system consists of chromium
reduction, single-stage alkaline chlorination, neutralization, flocculation,
sedimentation, sand filtration, sludge thickening, and filter press
dewatering.  The F006 contained a chemical oxygen demand (COD) of 37,000 mg/L,
400 mg/L total cyanide and total solids of 45 percent.

      The F007 plating bath solution for wet-air oxidation tests was obtained
from a manufacturer of decorative hardware, with brass, copper, zinc, nickel,
and chrome plated products.  The spent plating bath had a COD of approximately
33,000 mg/L, 24,000 mg/L total cyanide, and suspended solids of 15,850 mg/L
(about 16 percent).

      The F019 filter cake samples were obtained from a manufacturer of
automotive refrigeration systems.  The process wastewaters received treatment
to reduce hexavalent chromium to trivalent, adjust the ph, precipitate metals
and dewater the sludge.  The samples contained 43,000 mg/L COD, about 5,000
mg/L total cyanide (as ferricyanide, Fe(CN)6"3), and about 40 percent total
solids.

WET AIR OXIDATION EXPERIMENTAL PROCEDURE

      The tests were conducted at Zimpro/Passavant in their 0.5-liter capacity
titanium shaking autoclave.  Oxidations were performed for each waste at
temperatures of 200*, 240°, and 280° C and a residence time of about 1 hour.
The F006 was diluted 1:10 (volume) and potassium hexacyanoferrate added to
restore the sample to approximately the original cyanide concentration.  The
F007 test was conducted on a 1:1 dilution, while the F019 was diluted 1:4.

                                     98

-------
Pressures were determined by the amount of air charged to the autoclave.
After charging with a dilute waste sample and enough air to provide 120% of
the F006 waste's oxygen demand (150-200 psig air charged), the autoclave was
placed in a heater/shaker for the duration of the test.  The autoclave was
then cooled, the contents filtered, and both fractions analyzed.

TEST RESULTS

      The results of the autoclave oxidations of F006, F007 and F019 are
presented in the following tables.  Compared to the dilute wastes, the COD
reductions were in the range of 92-95% for the F006, 79-86% for the F007, and
68-83% for the F019.  The total cyanide reductions in the filtrates were in
the range of 78-99% for the F006, 99-99.5% for the F007, and 98-99.9% for the
F019.  The ammonia-nitrogen from the cyanide oxidations reached a maximum of
about 300 mg/L for the F006 at the 240* and 280°C tests, 5000 mg/L for the
F007 2008C test, and about 760 mg/L for the F019 28Q'C test.

      The total cyanide values shown for "cake" samples represent only a small
contribution to the values for the filtrates.  When both filtrate and solids
are considered together, the cyanide destruction values given in the tables
for the filtrates are representative.  For example, the total and amenable
cyanide analysis of the F006 solids was 102 ug/g (ppm) at 200*C.  The cyanide
in the solids would contribute only 1.5 mg/L total cyanide to the 26.3 mg/L
value for the filtrate and the total cyanide destruction remains nearly 78%.

DISCUSSION

      Difficulties were encountered with the initial cyanide analyses for the
three wastes.  The measured total cyanide concentration of the F019 was about
2500 micrograms per gram as received from the waste generator.  A value of
5000 micrograms per gram was reported by another laboratory.  After a four to
one dilution, the cyanide concentration of the waste was about 300 milligrams
per liter---much lower than the calculated value, which was either 1250 or 625
milligrams per liter, depending on which initial value was used.

      For the F006 test, about 40"miHigrams per liter cyanide came from the
diluted filter cake  (ten to one) plus 315 milligram per liter cyanide from
addition of potassium hexacyanoferrate.  The analyses of the slurry indicated
about 120 milligrams per liter cyanide, about 35 percent of the expected
value.

      It was recently noted that the SW-846 method 9010 for cyanide does not
specify the sample size or distillation time.  This could have been a factor
in the variability of the results.  Certain compounds can also interfere with
the analysis  (thiocyanate, sulfides, nitrates) and storage of the sample can
result in inaccuracies in the determination of cyanide.  However, we were not
able to identify the specific reasons for these inconsistencies in the WAO
tests.
                                     99

-------
CONCLUSION:

      Even though there were apparent difficulties with the cyanide analytical
method, the data from the wet-air oxidation treatment tests indicate
significant destruction of the cyanides.  It was also determined that the
oxidized liquor would require subsequent treatment to remove the residual
ammonia, any organic compounds and dissolved heavy metals.

TREATMENT STANDARDS FOR F006, F007, F019

      Wet-air oxidation is the treatment technology proposed for F019. (3)
However, final treatment standards for F006, F019 and F007 wastewaters are
based on the performance of alkaline chlorination for the amenable and total
cyanides, and chromium reduction followed by chemical precipitation using lime
and sulfides and sludge dewatering for the metal.  The total cyanide treatment
standard concentrations required by RCRA are 590 milligrams per kilogram.

REFERENCES

1.    Sittig, H. 1973.  Pollutant Removal Handbook. Noyes Data Corporation.
      Pack Ridge, NO

2.    U.S. EPA, 1986.  "Test methods for Evaluating Solid Wastes," (SW-846),
      Office of Solid Waste, Washington, DC

3.    "Land Disposal Restrictions for Third Scheduled Wastes,"
      November 22, 1989, Federal Register, pp. 48372-48529
                                    100

-------
                                              TABLE 1.  WET-AJR-OXTDATION RESULTS FOR F006

                                                                                       Oxidized product
Parameter

Oxidation temp, °C
Time at temp, min

COD,g/L
COD redaction, %
PH
NH,-N, mg/L
Tonl solids, g/L
Total ash, g/L"

Cyanide, total, mg/L*
Total cyanide reduction,1*
Cyanide amenable, mg/L*
Amenable CN reduction, %

Sulfides, mg/L
Fluoride, mg/L
Arsenic, pg/L
Antimony, jig/L
Barium, mg/L
Beryllium, mg/L
Cadmium, mg/L
Chromium (T), mg/L
Chromium (+6), mg/L
Copper, mg/L
boo, mg/L
Lead, mg/L
Mercury, mg/L
Nkkel, mg/L
Selenium, mg/L
Silver, mg/L
Thallium, mg/L
Vanadium, mg/L
Zinc, rag/L
Autoclave
  Feed
Filtrate
Cake
-
-
3.7
.
11
-
17.3
15.9
122.4
-
73.9
-
200
60
0.166
95.5
99
203.60
1.92
1.77
263
78
26.3
64
200
60
197
.
-
-
100%
85.3%
102 Mg/g
-
102 Mg/g
.
  Filtrate

  240
   60

0.269
   93
  9.7
305.1
  1.25
  1.17

 0.82
   99
 0.82
   99
 Cake

240
 60

 24
                                                             100%
                                                            86.6%

                                                           20 Mg/g

                                                           20 Mg/g
339
<\3.
<5.0
034
<0.01
220.5
348.4
Interference
729.1
1177
1.12
0.42
643.9

-------
                                              TABLE 2. WCT-AK-OXIDATION RESULTS FOR FOOT
Raw
Waste

33370
-
23,667
-
21,750
-
260,770
242,300
10.8
134
136
<12
323
<0.04
7.5
<0.1
7663
0.81
0.0048
8.46
312
<0.05
4180
Dilute
Waste*

16,685
-
11,834
-
10,875
-
130,385
121,150
-
67
0.68
<12
1.62
<0.04
3.75
<0.1
3832
0.40
0.0024
423
15.6
<0.05
305

Filtrate
200
3,060
81.7
117
99.0
114
99.0
110,000
103,600
9.8
5,000
„
<12
0.017
<0.004
324
0.90
863
23
0.0095
038
13.0
<0.05
305

Solids
200
_
-
183ng/g
-
143|lg/g
-
9,500''
8300"
-
-
_
<35
104
<1.15
30.5
IS*
261,440
464
023
1305
90
<1.4
75J90

Filtrate
240
2315
86.1
86
993
86
99.2
86,600
83,100
9.4
3,104
._
<12
0.012
<0.004
333
0.66
553
1.4
0.0106
0.099
92
<0.05
159

Solids
240
•
-
<1.0 |lg/g
-
<1.0 ng/g
-
9,000*
saoo*
-
-
«.
<35
83.4
<1.15
13.5
<3.4
234,600
229
0.14
408
89
<1.4
79270

Filtrate
280
3,578
78.6
58
99.5
55
99.5
762000
66,600
9.0
1,187
'.
<12
0.003
<0.004
3.04
0.18
316
0.80
0.0106
0.067
5J
<0.05
29.8

Solids
280
—
-
<1.0 jig/g
.
<1.0 |ig/g
-
8,800"
7,800*
.
-
—
<32
68.0
<1.07
133
<3.9
219,450
172
0.153
337
116
<13
82^65
Parameter

Oxidation temp, °C

COD, mg/L
COD reduction, %

Cyanide (I), mg/L
Cyanide CD. red., %
Cyanide (A), mg/L
Cyanide (A), red., %

Total solids, mg/L
Total ash, mg/L
pH
NH3-N, mg/L

Sulfide, mg/L
Arsenic
Barium
Cadmium
Chromium (T)
Chromium (+6)
Copper
Lead
Mercury
Nickel
Selenium
Silver
Zinc

* Concentrations calculated on 1:1 dilution of raw wastewater.

* Concentration of suspended solids/suspended ash filtered from oxidized waste.
c IS = Insufficient Sample

-------
                                               TABLE 3.  WET-AIR-OXIDATION RESULTS FOR F019

                                                                                   Oxidized product
Parameter

Oxidation temp, °C
Time at temp, min

COD, g/L
COD reduction, %
pH
NH,-N, mg/L
Total solids, g/L*
Total ash, g/L*

CN, total, mg/L'
Total CN from solids, mg/L
Total CN reduction, %
CN amenable, mg/L*
Amenable CN from solids, mg/L
Amenable CN reduction, %

Sulfides, mg/L
Fluoride,  mg/L
Arsenic,
Antimony,
Barium, mg/L
Beryllium, mg/L
Cadmium, mg/L
Chromium (T), mg/L
Copper, mg/L
Iron, mg/L
Lead, mg/L
Mercury,  mg/L
Nickel, mg/L
Selenium, mg/L
Silver, mg/L
Thallium, mg/L
Vanadium, mg/L
Zinc, mg/L
 Autoclave
   Feed
 10.5

 8.54
 27.7
 41.7
 28.1

293.1


240.9
 38.9
 80.2
 <5.0
  2.7
0.007
 2.11
 1231
0.355
  189
 14.7
 0.35
0.875
 50.8
<0.01
<140
 0.31
4902
   Filtrate

   200
    60
  1.75
  83.3
  7.95
527.50
  2.78
  1.45

  5.07

    98
     5

    98
  24.7
   132
  <5.0
  051
<0.001
 0.013
  1.92
 0.053
  0.13
 0.003
   1.1
 0.007
   250
<0.005
   9.5
 0.006
   4.6
        Cake

       200
        60

       73.7
     98.8%
     83.9%

  22.9 Mg/g
       0.91

  22.9 Mg/g
       0.91
  0.17 Mg/g
   114 Mg/g
   465 Mg/g
   138 Mg/g
  0.32 Mg/g
   102 Mg/g
72567 Mg/g
  48.7 Mg/g
12,000 ug/g
   816 Mg/g
 <0.02Mg/g
    46 Mg/g
   <20Mg/g
  <0.5Mg/g
    <5Mg/g
  <0.5Mg/g
61,000 Mg/g
   Filtrate

   240
    60

  3.04
    68
   7.9
   668
  2.55
  1.26

 0.058

  99.9
  0.02

 >99.9
  30.9
   221
  <5.0
  0.35
<0.001
 0.007
  1.86
 0.046
  0.08
 0.008
    1.0
 0.011
  <5.0
<0.005
    5.0
<0.005
   4.6
         Cake

        240
         60

        160
       100%
      79.6%
    142
        3.07

    142 Mg/g
        3.07
   0.15 Mg/g
    135 Mg/g
    460 Mg/g
    145 Mg/g
   059 Mg/g
     99 Mg/g
 68,590 Mg/g
     34 Mg/g
 11,034 Mg/g
    586 Mg/g
  <0.02Mg/g
    43mg/g
    <20Mg/g
   99.9
38.9
<5.0
<5.0
0.77
<0.01
<0.04
24
0.12
0.08
0.916
0.8
051
<5.0
0.05
5.0
<0.05
15.2
280
60
25.3
_
.
.
100%
89.7%
18 Mg/g
0.57
.
18 Mg/g
0.57
-
053 Mg/g
109 Mg/g
469 Mg/g
166 Mg/g
032 Mg/g
104 Mg/g
74,072 Mg/g
66 Mg/g
15538 Mg/g
584 Mg/g
<0.02ng/g
48 Mg/g
<20Mg/g
<0.5Mg/g
<5Mg/g
<0.5 Mg/g
279.000 Mg/g
* except as noted

-------
                 PASSIVE TREATMENT OF METALS MIME DRAINAGE
                    THROUGHUSE OF A CONSTRUCTEDWETLAND

           by:   S. D. Machemer, P. R. Lemke, T. R. wildeman,
                 R. R. Cohen, R. W. Klusman, and 3. C, Emerick
                 Colorado School of Mines
                 Golden, Colorado 80401
           and   Edward R. Bates,
                 Risk Reduction Engineering Laboratory, u. S.  EPA,
                 Cincinnati, Ohio 45268

                                 ABSTRACT

     For  contaminant  removal,  use of  a  constructed wetland allows one  to
maximize one or two processes over  the many that occur in a natural  wetland.
For acid mine drainage, anaerobic processes mediated by bacteria will  raise
the pH and precipitate the heavy metals as sulfides.   The Big  Five Tunnel
Pilot Wetland was built to test these ideas on a metal mine drainage.   Since
the original construction  in 1987  and  subsequent  modifications  in  1988 and
1989, long term removal of contaminants has been achieved and  much  has been
learned about competing removal processes.   The best removal efficiency has
been with a  subsurface flow wetland that is similar to a trickling filter
flow system.   With this system,  100 % removal of Fe, Cu, and Zn,  increase  in
the pH from below 3 to above 6,  and 25 % removal of Mn has been consistently
achieved.

     This paper has been reviewed in  accordance with the U, S.  Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication,

                                 BACKGROUND

     Under the SITE Emerging  Technology Program, the  U.  S.  environmental
Protection Agency is seeking to evaluate and further develop,  through  pilot
scale, promising new technologies for the remediation of Superfund hazardous
waste  sites.   EPA accepted and is  jointly  funding  a proposal  from the
Colorado School of Mines to evaluate and further  develop the  technology  of
constructed  (man  made)  wetlands for treating toxic metals in mineral mine
drainage.  Geologists have long understood that soils in wetlands are  often
foul because they naturally accumulate  contaminants.   The methods  for  a
wetland to accumulate contaminants include:
     1.    Filtering of suspended and colloidal material  from the water.
     2,    Uptake of contaminants into the roots and leaves of live plants.
     3.    Adsorption  or  exchange  of contaminants  onto  inorganic  soil
     constituents,  organic solids, dead plant  material, or algal material,
     4,    Neutralization  and  precipitation  of  contaminants  through the
     generation of  HCOj'and NH3 by bacterial decay of organic matter.
     5.    Destruction or precipitation of contaminants in the aerobic zone
     catalyzed by the  activity of  bacteria.


                                    104

-------
     6.    Destruction or precipitation  of chemicals in the anaerobic zone
     catalyzed by the activity of bacteria.
With so  many possible removal  processes, a  wetland,  such as  depicted in
Figure 1, is  the typical contaminant treatment system in a natural ecosystem.
In addition,  it operates  in a passive mode requiring no additional reactants
and no continuous maintenance.

WHY USE A CONSTRUCTED WETLAND?

     In  the  last decade, engineers began  to use wetlands for the removal of
contaminants from water  (1,  2) .   In some instances,  natural  wetlands were
used.   However,  a  natural  system  will  accomodate  all the  above  removal
processes and probably will  not  operate  to maximize  a certain process.  If
a wetland is constructed, It can be designed to maximize a specific process
suitable for the removal  of certain  contaminants from water.  Engineering as
well as ecological reasons lead to the choice of constructing a wetland for
contaminant removal rather than using an  existing natural ecosystem.

     As  an example  of constructing a wetland to  maximize specific removal
processes, consider the  bacterial processes that are  items  5 and 6 in the
above list.  Typical microbial mediated reactions that are possible in the
aerobic  zone of a wetland include;
               4 Fe2+  +  02  -f 10 H20   	> 4 Fe(OH)3   + 8 H*
                      2  02   + H2S  	>  SO,"  +  2  H*
                 2 H,O  + 2 N,  + 5 O,   	>  4 NO,"  +  4 H*
Typical microblallly mediated  reactions  that  are possible in the anaerobic
zone of a wetland include:
          4  Fe  4 Fe2
CO,
11
               3 CHjO  +  2 Nj  +  3 H2O   	>   4 NH3  +  3 CO2
                     SO<°  +2 CH2O  	>   H2S  +  2 HC03"
In  these reactions, "CH2O"  is  used to  symbolize  organic material  in the
substrate.
     DAM

              Figure 1.  A model of a typical wetland system.

      It  is  apparent  that  the  anaerobic  reactions are  approximately the
reverse of the aerobic reactions.  Both zones exist in a wetland.  If removal
involves  aerobic  processes,  then the wetland should be constructed so the
water remains on the surface.  If  removal involves anaerobic processes, then

                                    105

-------
the wetland should be constructed so the water courses through the substrate.
In a natural wetland, the water primarily remains on the surface.  Also, note
that the aerobic reactions generate hydrogen ions and the anaerobic reactions
consume hydrogen ions.

     In the important area of microbially mediated  removal, the wetland must
be  constructed  to  maximize  removal  reactions  and  minimize  competing
reactions.  In the case of  removing contaminants  from acid mine drainage,  it
is clear that  removal  processes should consume hydrogen  ions,  consequently
anaerobic processes are emphasized <3, 4) .   The research and development  at
the Big  Five Tunnel site  in Idaho  Springs,  Colorado has concentrated  on
understanding  the chemistry and ecology involved  in  removal  and designing
structures from readily available materials that maximize these processes.

     Although   it  apppears   to  be   "low   technology",     an   intense
interdisciplinary effort and creative engineering skills are needed to design
and perfect  systems that maximize natural  processes.  For more  details  on
what should be considered,  the monographs  by Reed, Middlebrooks  and Crltes
(1)  and  by Hammer  (2)  are excellent  places  to  begin.    For  specific
information  on the  use of wetlands  for  controlling  acid  mine  drainage
problems, the U. S. Bureau of Mines Monograph edited by Kleinmann (3)  is  an
excellent source.  Note that all three  references have been published in the
last  three  years.     The  technology  being  considered is  still  in the
development stage.

                        THE BIG FIVE PILOT WETLMTO

ORIGINAL DESIGN, CONSTRUCTION, AND OPERATION

     The  original  design  is  shown in  Figure 2j and construction  from that
design was  completed  in the  summer  and fall of 1987. The completed  pilot
plant consisted  of  a  3.05  m  wide,  18,3 m long,  and 1.22 m  deep  concrete
structure divided into three individual 6.1m  treatment cells <5,  6). The
cells were separated from each other by wood walls and each  cell was  lined
with 30  mil Hypalon sheets.   The liner isolated  each cell  and prevented
chemical  reactions  between system  components  and  mine drainage.    Mine
drainage flowed  into each  cell  at the upstream  end  through  baskets filled
with washed  10-to 15-cm river rock.   The  rock enclosures were  about  15  cm
long and extended the  full  width  and depth of each cell  to  facilitate flow
through and contact with the substrate.  For monitoring purposes, each cell
was supplied with a outflow drain  at  the downstream end and six access wells
completed at different depths.

     In the original construction, the cells were  filled with three feet  of
the following  substrate materials;   Cell A - mushroom compost;  Cell  B - a
mixture of equal parts of peat moss,  aged manure, and decomposed wood; Cell
C  -  same as B but  with a  15 cm  layer of  5  - 8 cm limestone rock.   From
existing  wetlands  in the  region  were   transplanted  cattails   (Typha.
ancmstifolia and T.  latifolia)-,  sedges  (Carex aguatilis, and C. utriculata),
and rushes (Juncus articus)  into each cell.  Each cell was soaked and tested
with municipal  water  for  one month  and flow of mine drainage  through the
cells was initiated in October,  1987.

     From the beginning, Cell A with mushroom compost was most effective  in
removing  contaminants  and  raising the pH.   Selected values for the mine
drainage input and the cell outputs  are  shown in Table  1.  Mn,  Fe,  Cu, and

                                     IDS

-------
                   OUTLETS
                          •INLETS
                                                                TO
                                                            PORTAL
TOP OUTLET
                                                             INLET
 HYPALON
   LINER
                    SAMPLING
                       WELLS
  ROCK BOX
AT  INLET
Figure 2.  A plan view of the original pilot system at the Big Five Tunnel
           including an expanded cross section of Cell B.

Zn  are  the  primary  heavy  metal  contaminants  and  also  give  excellent
indication of the removal processes that  are operating  (7).   it was found
that removal strongly depended on the loading rate expressed in square feet/
gallon/ minute (6) .  Since the size of the  cell is fixed at 18.6 mz (200 ft"),
the flow is inversely proportional to the loading rate.  The best results for
Cell A during this period were 100  % removal of Cu and Zn, 63 % removal of
Fe, and an increase in pH from 3.0  to 6.2  at a loading rate of 600 ft1/ gal/
min.   In the  first  stage of operations,  Mn  was  not  removed.   The removal
patterns and results from other  experiments performed  on  the substrates (7,
8) gave convincing evidence that the important removal  process  was bacterial
reduction of sulfate in the mine drainage to hydrogen sulfide and subsequent
precipitation of the metals as sulfides.

PROBLEMS WITH THE ORIGINAL DESIGN

     In  acid  mine  drainage,  Fe   (II)   is  continuously  oxidizing  and
precipitating as  Fe(OH),  (4).   Many of  the problems  encountered during
original operation  were  attributed to clogging by ferric hydroxides (6) .
Pipes,  valves, and rock baskets all became clogged and maintenance frustrated
the objective of passive  operation.  Also, when it was  realized that the
important removal processes  were  occurring in the substrate anaerobic zone,
the original design  appeared inadequate for maximizing  contact  of water with
substrate. Figure 2 shows that water flows out from the top of the cell and
this insures a poor flow pattern through  the anaerobic  zone.  Subsequent
phases of  the project  involved  the redesign and construction of cells  to
maximize contact  and subsurface  flow.
                                    117

-------
TABLB 1.  CONCENTRATIONS  (mg/L) OF METALS, PERCENT REDUCTION OF METALS, pH,
AND FLOW RATES (GALLON/MINUTE)  IN THE BIG FIVE MINE DRAINAGE AND WETLAND CELL
OUTPUT WATERS.*
Hater Sample
Mn
red.
INITIAL OPERATION
Mine
Cell
Cell
Cell

Mine
Cell
Cell
Cell
CELL
Mine
Cell
Cell
Cell

Mine
Cell
Cell
Cell
Drainage
A
B
C

Drainage
A
B
C
34
27
33
34

26
25
26
25

21
1
0


4
0
4
A MODIFICATION
Drainage
A
B
C

Drainage
A
B
C
27
22
27
25

30
33
27
29

19
0
7


-10
10
3
Fe red.
Zn
December 11,
32
18 45
24 26
22 32
August 19,
37
20 56
15 59
11 70
February 21
32
12 63
28 13
31 3
May 6. 1989
42
28 33
10 76
9 79
10.6
7.8
9.8
9.6
1988
8.1
<0.1
6.1
5.8
, 1989
9.3
4.5
6.1
7.2

10.4
7.8
6.2
6.4
red.
1987

27
12
9


100
24
28


52
34
23


25
40
38


1
0
0
0

0
0
0
0

0
<
0
0

0
<
0
0
Cu

.02
,44
.89
.91

.91
.17
.55
.38

.56
.05
.82
.26

.76
.05
.36
.46
red.


57
12
10


81
40
58


100
0
53


100
53
39
PH

2.8
4,6
3.1
3.3

2.9
5.5
3.2
3.5

3.0
5.1
3.4
3.5

3.0
3.5
3.0
3.2
flow
rate


1.0
1.0
1.0


0.51
0.24
0.34


0.28
0.31
0.32


0.92
0.83
0.81
August 1, 1989
Mine
Cell
Cell
Cell
CELL
Mine
Cell
Cell
Cell

Mine
Cell
Cell
Cell

Mine
Cell
Cell
Cell
Drainage
A
B
C
32
31
26
32

3
19
0
B MODIFICATION
Drainage
B-Up
B-Down
E

Drainage
B-Up
B-Down
E

Drainage
B-Up
B-Down
E
35
34
23
24

32
31
20
20

31
30
27
28

3
34
31


3
38
38


3
13
10
43
39 9
24 44
18 58
October 3,
46
39 15
8 83
<1 100
November 5,
38
17 55
<1 100
<1 100
January 13,
33
33 0
12 64
10 70
9.4
5,2
6.6
4.8
1989
9.9
9.3
0.8
<0.1
1989
8.7
8.4
6.0
<.l
1990
9.0
8.9
8.1
<.l

45
30
48


6
92
100


3
31
100


1
10
100
0
<
0
0

0
0
<
<

0
0
<
<

0
0
0
<
.75
.05
.46
.09

.66
.59
.05
.05

.61
.48
.05
.05

.59
.49
,44
.05

100
39
88


11
100
100


21
100
100


17
25
100
2.9
4.1
3.1
3.7

3.2
3.5
6.5
6.3

2.9
3.6
5.9
6.5

2.9
3.2
3.2
6.0

0.30
0.48
0.43


0.29
0.16
0.062


0.22
0.24
0.11


0.21
0.20
0.10
*The area of  Cells  A,  B,  and C is 200 ftz; the area of Cells B-Up,  B-Down,
and E is 100  ft2.
                                     108

-------
                         SUBSEQUENT MODIFICATIONS
CELL A MODIFICATION

     The first redesign concentrated on the issue of increasing the contact
of the drainage with the substrate especially in the  anaerobic zone.   This
was accomplished by the  addition  of:   Two walls  running the length of the
cell  to  increase  the flow  path  length by  a   factor  of  three,  and  six
redistribution baffles to  collect water  flow   from  the  top  surface  and
redistribute it to the bottom of the substrate.   This was done on Cell  A and
a cut-a-way view of the redesign  is shown  in Figure 3.    Essentially,  Cell
A was redesigned to be a six segment plug flow reactor <9).
  OVERFLOW
  DISCHARGE
                                                          INLET  BOX
      Figure 3.
                                  USTRIBUTION
                                     BAFFLE

          END BAFFLE

A cut-a-way view of the modification done to Cell A,
     Although the initial structure design was amenable to major changes,  the
results  of  the  redesign  were discouraging.   The desired  plug flow from
segment to segment through the lower part of the substrate was not achieved.
Considerable water flowed across the top of cell segments and leaked from  one
segment to others. When the cell was modified, the original mushroom compost
was removed, stockpiled, and returned to the cell after the remodeling.   It
was   speculated  that  through   this   handling,  substrate   permeability
significantly decreased.   Subsequent  experiments on new and  used  mushroom
compost verified that permeability decreased from 3.0 x  10"3  to 9.2 x  10"B
cm/sec  (10).  Selected values on removal and changes in pH from the  cells
during this phase are shown in Table 1.

     As a result  of this setback,  a laboratory  and bench scale program  was
developed to  determine how well typical soil tests could be adapted to this
highly  organic  substrate.    Especially important were  the development  of
methods  to determine hydraulic  conductivity  that  could  give  reasonable
indications of what to expect in a constructed wetland (10).  Other  tests on
substrates included the determination of specific gravity, bulk density, size
fractions, and percent moisture.

CELL B MODIFICATION

     Results of  laboratory  and bench-scale permeability experiments led to
the modification of Cell  B into  upflow  and downflow cells to  monitor  and

                                    109

-------
evaluate permeability at the pilot scale.  The original cell was divided into
two lined,  identical  cells so individual variables could be tested.   All
features needed to determine  soil  permeability  in the cell  were Included.
A special  feature in  the  design was  inclusion  of a plenum beneath each
subcell for even  distribution of drainage  in the upflow configuration and
oven collection when used as a down flow  cell.  Each cell could be operated
in the upflow or downflow configuration.  Figure 4  is a cut-a-way diagram of
the downflow operation.
                     MINE DRAINAGE -*.
                     RESERVOIR
         SURFACE
         DISTRIBUTION
         SYSTEM
                                                       CELLB
                                                       SUBSURFACE
                                                       DESIGN
     DISCHARGE
                                          ILTEft FABRIC
                                    PLENUM
Figure 4.  A  cut-a-way  view  of the  modification done  to Cell  B.   Two
identical subcells were constructed;  figure shows  a downflow configuration.

     There are a number  of features in this modification that should improve
contact with the substrate.   When operating either as an upflow or downflow
cell, the mine  drainage is  forced through the substrate before discharge.
This configuration is comparable to a trickling filter process Instead of the
plug-flow reactor design in the Cell A modification (9).  Also included in
the modification, was the addition in series of two 150-gallon stock tanks
before the inlets of the subcells.  Figure 4 shows the placement of one of
the tanks. The tanks serve the purpose of  completing the precipitation of
ferric hydroxides  before drainage enters  the subcells  so  that plumbing,
plenums,  and filter fabrics do not become clogged.

     Remodeling of Cell B was completed in August, 1989.  The two subcells
were filled with fresh mushroom compost  to a depth of 0.61 m.  Flow of mine
drainage through the subcells was initiated on September 1.   One subcell was
operated  in  the  upflow  configuration,  the  other  downflow.    Upflow
                                    110

-------
permeability measured  1.3 x 10"* cm/sec,  and downflow permeability measured
3.1 x 1Q~4 cm/sec  (10) .  The development of laboratory and bench-scale methods
for  predicting  permeability  in  actual constructed wetlands  has  proven
successful,

     Values  on removal of Mn, Fe,  Cu,  and Zn and increases in pH for Cell B-
Upflow and Cell B-Downflow are given  in  Table  1.   Just  as  for the original
Cell A design,  removal has been impressive from initiation of flow.  Removal
from  the  downflow system is better than from the upflow, however,  flows
through the  cells have not always been equal so a true comparison has yet not
been completed.  After 3 months,  removal is not as good as  during the first
month.   Recent  laboratory  studies on the  substrate have  shown  that  the
complexing of metals onto organic sites in the substrate can occur and that
this  is  a significant removal  process  at  the start-up  of  a  constructed
wetland (6).  Even so, the Cell  B-Downflow appears to operating better than
Cell A at comparable loading factors.

CELLS D AND E DESIGN,  CONSTRUCTION, AND OPERATION

     When Cell B was remodeled,  Cells D and E were constructed  using the
original  substrate from  Cell B,   Cell D  was  designed to polish discharges
from  anaerobic cells by using aerobic processes.   Features  of  the design
include a shallow depth  (0.50 m)  and a length to width ratio of 10.  It has
been  receiving the discharge from Cell A and  removal of Cu,  Zn,  and Fe is
completed and the pH is  raised to above  6.

     Cell E  was  designed to operate as a  downflow,   subsurface  wetland.
Construction was  completely accomplished with  materials found locally.  It
is approximately 9.3 m2 and the substrate is 0.61 m deep. Subsurface flow is
achieved by flow through landscape fabric into 2.5-cm gravel and subsequent
discharge into a tube on the downflow end.   For this system,,  results have
been  excellent.

     Flow of mine drainage through Cell E was initiated on September 1, 1989.
From  the beginning,  removal  of  Cu,  Zn, and  Fe  has  been 100  %,  pH has
increased to  6.5,  and Mn removal has averaged 25 %.   Removal  results are
given  in  Table 1 and  are shown  in Figure 5.   Laboratory  experiments have
confirmed that sulfate  reduction with  subsequent precipitation  of metal
sulfides  is the predominant removal process in Cell E (7).

                         ADDITIONAL CONSIDERATIONS

     A possible toxic metal removal mechanism would be through accumulation
in leaves, stems, and  roots of wetland plants.  Should this occur, it could
create problems by concentrating toxic metals in the biologically very active
wetland  surface.   However,  at  the Big  Five  Site, careful collection and
analysis  of  wetland  plant samples has revealed only minor accumulation of
toxic metals after two years of  operation.   This  result is consistent with
studies of wetland removal systems used for coal mine  drainage  that  show that
uptake by plants  accounts for at most 5  % of the metal  remova.1  (11) .

      Gradually over  considerable  time, toxic metals  will accumulate in the
anaerobic and aerobic zones of  a wetland.   At  some  point,  these metal
sulfides  and hydroxides mixed with organic substrate will have to be removed
from  the wetland and treated for disposal or be recycled for metal recovery.
However,  the  quantity of sludge  to be thus handled  represents  a small

                                    111

-------
fraction of that which would be produced by more conventional processes  and
this sludge may even have economic potential for metal recovery.
O
uu
z
O
uu
1.0


0.8


0.6


0.4


0.2
                                                          •••I ••«/*%• I Ml •!

                                                          —Hi—
MnE
FeE
CuE
ZnE
SO4E
                      40
                    60    80    100  120   140

                       DAYS
          Figure 5.  Cell E removal data since September 1,  1989.

                                 CONCLUSIONS

     Using  constructed  wetlands  for  wastewater  treatment   is   still  a
developing technology.  However,  the results from the Big Five  Pilot Wetland
that was funded by the  Emerging Technology Program (ETP) of the U. S. EPA
shows promising removal of heavy metals and  increase  of pH for acid mine
drainage.  Conclusions from the project include:
     1.  Toxic metals such as  Cu and  Zn can be removed and the pH of mine
     drainage can be increased on a long term basis,
     2.   The major removal process  is sulfate  reduction  and subsequent
     precipitation  of  the metals  as  sulfides.   Exchange  of metals onto
     organic matter can  be important during the initial period  of operation.

     3.  A trickling filter type of configuration achieves the best contact
     of the water with the substrate.
     4.  Removal efficiency depends strongly on loading factors, in the Big
     Five wetland/  factors above 1,000 feet* per gallon/minute are needed for
     reasonable removal.
     5.  Permeability of the substrate is  a  critical design  variable for
     successful operation.  Using laboratory and bench-scale  tests, a good
     indication of  the  soil permeability in  a  constructed  wetland can be
     determined.
     6.   Solutions to  problems such  as plugging  of plumbing by ferric
     hydroxides and freezing of discharge lines during  winter have  to be
     designed and  constructed into the passive nature of wetlands to achieve
     long term operation.
                                    112

-------
     7.  Eventual removal of precipitated metal sulfides for metal recovery
     or disposal must be included in the operating plan.

                             ACKNOWLEDGEMENTS

     Development of environmental technology doesn't occur without scientists
and engineers taking an interdisciplinary approach and sponsors taking risks.
For this project, success required many people  to  see the  broad picture and
take risks.   Among them  are:   Roger  Olsen  and Richard Chappell  of Camp,
Dresser, and McKee; John Gormley and James Gusek of Denver, Knight, Piesold;
Holly Fliniau, James  Kreissl,  and James Lazorchek  of the U. S. EPA; and Jeff
Deckler and Jeb Love of the Colorado Department of Health.

                                REFERENCES

1.   Reed, S. C.,  Middlebrooks, E. J., and Crites,  R. W.  Natural Systems for
     Waste Management and Treatment.  McGraw-Hill, New York,  1988.  308 pp.

2.   Hammer, D. A.   Constructed Wetlands for  Wastewater Trea.tment.   Lewis
     Publishers, Chelsea, Michigan,  1989.  800 pp.

3.   Kleinmann, R. L.  P.  (ed.), Proceedings of  a Conference on Mine Drainage
     and Surface Mine Reclamation.  Vol.  1. Mine  Water  and Mine Waste.   U.
     S. Department of the Interior,  Bureau of Mines Information Circular 1C
     9183, 1988, 413 pp.

4.   Wildeman, T. R., and Laudon, L. S.   The use  of  wetlands for treatment
     of environmental problems in mining:  Non-coal mining applications.  In;
     D. A. Hammer (ed.), Constructed Wetlands for Wastewater Treatment. Lewis
     Publishers, Chelsea, Michigan,  1989.  p. 221.

5.   Howard,  E.  A.,   Emerick,  J.  C.,   and  Wildeman, T.  R.    The  design,
     construction and initial operation of a research site for passive mine
     drainage treatment in Idaho Springs, Colorado.  In;  D. A. Hammer (ed.),
     Constructed  Wetlands  for  Wastewater  Treatment.    Lewis  Publishers,
     Chelsea, Michigan, 1989.  p. 761.

6.   Howard,  E.  A.,  Wildeman,  T. R.,  Laudon,  L.  S.,  and Machemer,  S.  D.
     Design considerations for the passive treatment of acid mine drainage.
     In;  Proceedings of the Conference "Reclamation, A Global Perspective".
     Alberta Land Conserv. and Reclam. Council Report No. RRTAC 89-2. p. 651.


7.   Machemer, S. D.,  and Wildeman, T. R.  Organic  complexation compared with
     sulfide precipitation as metal removal processes from acid mine drainage
     in a constructed wetland.  Jour, of Contaminant Hydrology,   in review.


8.   Laudon, L. S.  Sulfur Mineralization in a Wetland Constructed to Treat
     Acid Mine Drainage,  Master's Thesis No.  3660, Colorado School of Mines,
     Golden, Colorado, 1988.  58 pp.

9.   Tchobanoglous, G. and Schroeder, E. D.  Water Quality.  Addison-Wesley
     Publishing, Reading MA, 1985.  768 pp.
                                    113

-------
10.  Lemke,  P.  R,    Analyses and  Optimization of  Physical  and  Hydraulic
     Properties of Constructed Wetlands Substrates for Passive  Treatment  of
     Acid Mine Drainage.  Master's Thesis No.  3823, Colorado School of Mines,
     Golden CO, 1989.  78 pp.

11.  Scencindiver, J. C., and Bhjumbla,  D. K., Effects of  cattails  (Thpya)
     on metal removal from mine drainage.  In:  Proceedings of  a Conference
     on Mine Drainage and Surface Mine Reclamation.  Vol.  1.  Mine Water and
     Mine  Waste.  U.  S. Department   of the  Interior,  Bureau  of Mines
     Information Circular 1C 9183, 1988.  p. 359.
                                   114

-------
SOIL VAPOR EXTRACTION TECHNOLOGY ASSESSMENT
James T. Curtis
Tom A. Pedersen
Camp Dresser & McKee Inc.
Cambridge, MA  02142

Chi-Yuan Fan
USEPA RREL-RCB
Edison, NJ  08837
                                  ABSTRACT


     Soil vapor extraction (SVE) systems are being used in increasing
numbers due to many advantages these systems hold over other soil treatment
technologies.  SVE systems appear to be simple in design and operation, yet
the fundamentals governing subsurface vapor transport are quite complex.

     In view of this complexity, an expert workshop was held 1:0 discuss the
state-of-the-art of the technology, the best approach to optimize system
application, and process efficiency and limitations.  As a result of the
workshop, an SVE Technology Assessment report was produced.  This report
discusses the basic science of the subsurface environment and subsurface
vapor flow, site investigations, SVE system design and operation,
monitoring, emission control, and costs.  The report also serves as the
proceedings of the expert workshop.  Additional research activities being
conducted include a field demonstration of a structured SVE system design
approach; a laboratory column study to determine and characterize residuals
following vapor extraction; an assessment of secondary emissions and
regulations governing releases from SVE systemsj cost of SVE implementation
and operation; and a survey of techniques to enhance vapor removal.

     This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved
for presentation and publication.

                               INTRODUCTION


     Soil vapor extraction (SVE) is being used with increasing frequency at
sites throughout the country for the remediation of unsaturatcid zone soils
that have become contaminated with volatile organics.  SVE, also known by
various names such as vacuum extraction, in situ vaporization, or soil
venting, has many positive features that give it an advantage over other


                                    115

-------
soil treatment techniques such as excavation and off-site treatment, soil
flushing, or biotreatment.  Some of the features that make this technology
applicable to a broad spectrum of sites are:

  o  SVE is an in situ technology, so there is a minimum of site
     disturbance; often, business operations at the site need not be
     interrupted;

  o  SVE can treat large volumes of soil at reasonable costs;

  o  SVE is easily installed and uses standard, readily-available
     equipment, which allows for rapid mobilization and implementation of
     remedial activities;

  o  SVE is effective in reducing the concentration of volatile organic
     compounds (VOCs) in the vadose zone, reducing the potential for
     further migration; and

  o  SVE complements groundwater pump and treat techniques, which may be
     instituted concurrently.

     These features, combined with the apparent simplicity of SVE system
design, implementation, and operation, have combined to make SVE systems
one of the fastest growing remediation choices.  This growth has not
necessarily been accompanied by a concomitant expansion of the knowledge
base to design and operate an SVE system properly.  Indeed, the ease with
which SVE systems can be installed and operated belies the very complex
nature of subsurface vapor behavior and transport.

     Much of the technical information regarding the design, construction
and operation of an extraction system is held by the SVE technology
developers and vendors.  Engineering practices, which are often considered
proprietary by the developers and vendors, may be based in large part on
each developer's experiences.  This atmosphere does not encourage rigorous
objective review of design or operating methods and makes it more difficult
to analyze the results of SVE use.  Some vendors claim complete remediation
of a site to a specified level, yet are reluctant to share their methods or
procedures used to determine attainment of those goals.  This lack of
knowledge poses limitation to regulatory agency personnel or others trying
to interpret system operating results.

     It was recognized that gaining a deeper understanding of various
aspects of soil vapor behavior would help researchers, vendors, consulting
engineers, and regulatory agency personnel.  It was the increased use of
these systems and the need for a greater understanding of the principles
that underlie soil vapor behavior and other issues related to SVE that led
ORD, through its Risk Reduction Engineering Laboratory, Release Control
Branch (RREL-RCB) to initiate SVE research efforts.  This paper discusses
the activities thus far, summarizes the results, and highlights the
upcoming research to be completed under this contract.
                                     116

-------
                                 WORKSHOP


     The initial step in these efforts was the convening of an expert
workshop in Edison, NJ on June 28 and 29, 1989.  Workshop participants
included SVE technology vendors, petroleum Industry representatives,
university researchers, consulting engineers, regulatory agency
representatives, and others who were familiar with this technology.  The
workshop had dual objectives.  One goal was to discuss the state of the art
of SVE related topics, such as site characterization, SVE pilot systems,
full scale system design and operation, vapor treatment options, the
setting and attainment of cleanup criteria, and monitoring of SVE progress.
Some presenters discussed actual case studies; others discussed a
structured framework for conducting site  investigations and system design,
the use of modeling to help design extraction systems, and other research
on SVE currently in progress.  The second objective of the workshop was to
discuss additional research needs.  Panel discussions were held with
workshop attendees regarding areas for future research and topics needing
immediate attention.  The regulatory climate was also discussed, Including
suggested remediation standards and methods to determine cleanup
attainment.

     The consensus of those present was that SVE is effective for removing
a wide range of volatile compounds.  Hutzler et al. (1988) reported the
results of an extensive survey of operating SVE systems.  Data reported
included number and type of wells, well spacing and rate of influence, type
of pumps and auxiliary equipment used.  These reports showed that, while
SVE was very often successful in removing volatile contaminants, many times
this technique was used without a complete understanding of the physics and
chemistry behind the process.


                     SVE TECHNOLOGY ASSESSMENT DOCUMENT


     The SVE Technology Assessment Report was produced based on the
information presented at the workshop and subsequent research efforts.
This two-part report contains sections on the principles of soil vapor
behavior} the performance of site investigations prior to SVE use; SVE
system design approaches; system operation; system monitoring; vapor
treatment and cost in Part 1.  Part 2 of  the report serves as the
proceedings to the expert workshop and contains selected papers presented
at the workshop.  Papers were selected to provide coverage representative
of the wide range of topics discussed.  The report was written for an
audience that desires an in-depth understanding of the soil vapor
extraction and treatment process.  A summary of the findings presented in
this document is presented below.

PRINCIPLES OP SOIL VAPOR BEHAVIOR

     Products released into the subsurface environment from UST systems are
acted upon by numerous forces that influence the degree and rate at which
they migrate from the source.  The extent to which the released products
partition into the vapor phase is dependent upon the characteristics of the


                                     117

-------
product, the soil environment, and the elapsed time since the release
occurred.  The manner in which the release product behaves in the
subsurface will have a significant bearing on whether soil vapor extraction
could be an approach for the site under consideration.  Figure 1 is a
nomograph that uses vapor pressure and soil air permeability (with the
length of time since release) to show the applicability of SVE.

Contaminant Characteristics

     The physical and chemical properties of the released product control,
to a great extent, the movement and ultimate fate of that product in the
subsurface.  The degree to which a compound partitions into the vapor phase
is described by that compound's vapor pressure and Henry's law constant.
The soil sorption coefficient, K., describes the tendency of a compound to
become adsorbed to soil.  The solubility describes the degree to which a
product will dissolve into pore water.  The distribution of a product among
the three phases will vary with changes in site-specific conditions and
will also change over time in response to weathering.

     The released product's volatility controls in large part the quantity
present as vapor in the soil pores.  Volatility is perhaps the most
important product characteristic affecting applicability of soil vapor
extraction to that compound.  The parameter that best describes a
compound's volatility is its  vapor pressure.

Vapor Pressure—

     Vapor pressure is the force exerted by the vapor of the chemical in
equilibrium with its pure solid or liquid form.  Generally, compounds with
vapor pressures of less than 10"  mmHg are not volatile and are not removed
by SVI (Dragun, 1988); vapor pressures above 0.5 mmHg are removed to a
significant degree (Bennedsen et al., 1985).  Many gasoline constituents
have sufficiently high vapor pressures that they can be removed by SVE
(Table 1).

Henry's Law—

     Henry's law governs the volatilization from a contaminant in solution,
rather than from a pure product.  Henry's law constant (K. ) may be a more
appropriate partitioning constant outside of the free product zone, where
product is likely to exist in solution with pore water (Stephanatos, 1988).
Compounds with KL  above 0.01 (dimensionless) are suitably volatile for
removal via SVE (Hutzler et al. 1989).  Gasoline, with a composite ^-32,
is particularly well-suited to SVE.

Soil Sorption Coefficient—

     Sorption of contaminant to soil particles and organic matter controls
the distribution of released products in the soil zone and has a very
strong effect on the movement of the product through the vadose zone.  In
many cases the majority of the released product may exist in the sorbed
phase.  A strong relationship exists between the organic carbon content of
the soil and the sorption coefficient.  As soil organic carbon content
increases, the sorption for most products increases.


                                     118

-------
Water Solubility-

     Solubility controls the degree to which a product dissolves into
groundwater and pore water present in the vadose zone.  Soluble products
are more likely to dissolve in infiltrating precipitation and move away
from the source.
                                 t
Contaminant Composition—

     Petroleum hydrocarbons, the most common products released from
underground storage tanks (EPA, 1988), are actually mixtures of many
different compounds.  Each particular type of petroleum product has a
different composition and therefore will behave differently in the
subsurface.  SVE is more applicable to those mixtures with greater amounts
of lighter end fractions (e.g., gasoline) versus heavier-end fractions
(e.g., diesel or heating oils).

Weathering—
   &
     Weathering refers to the changes in the nature of a chemical mixture
after its release into the environment.  The product composition will
change over time and affect the ease with which that product may be removed
via SVE.  The more volatile, soluble, and degradable compounds will be
removed from the mixture, leaving the resultant mixture relatively richer
in less-volatile, less-soluble, and more-refractory compounds.  Figure 2
shows the composition by class for gasoline.  Weathering has applicability
to SVE because this phenomenon also occurs as SVE progresses, as the more
volatile fractions are removed.  Figure 3 shows the change in composition
of extracted vapor at a site in Utah (DePaoli et al., 1989).  As SVE
proceeds, the extracted vapors are progressively enriched with heavier
compound.

Soil Environment

     Similar to the product characteristics, the soil environment
significantly affects the transport and fate of released products and the
likelihood of success with SVE.  The most important properties are
summarized below.

Soil structure—

     Coarse-textured, highly permeable soils are best suited to SVE.  SVE
has also worked successfully, however, in clays and silts, where
interbedded permeable layers exist or macroporosity and secondary structure
exist.

Air permeability—

     The air permeability of the soil is perhaps the single most important
soil parameter with respect to the success of vapor extraction.  The
permeability incorporates the effects of several soil and vapor
characteristics.  The more important soil characteristics are the
stratigraphy, air-filled porosity, particle size distribution, water
content, residual saturation, and the presence or absence of macropores or


                                    tie

-------
preferred flovpaths; important product characteristics include the vapor
viscosity and vapor density.

     Soil water content has a significant effect on the air permeability.
In general, higher water contents reduce the air-filled porosity thereby
decreasing the connected pores through which air can flow by adveetion.

Ganglia formation—

     Ganglia, or isolated globules of product that remain in soil
micropores, may form in soils with fissures, cracks, or other dead end
pores.  A high degree of ganglia formation retards SVE suitability.

Water content—

     SVE is generally more successful at lower moisture contents.  High
water content reduces the air-filled porosity, with a reduction in air
permeability (Figure 4).  If the water content is very low, however,
sorption force to soil increases, leading to reduced volatilisation into
the soil gas (Reible, 1989).  A range of 94 to 98.5 percent relative
humidity in soil gas appears to be optimal for SVE (Davies, 1989).

VaporTransport

     Darcy's law is typically used to describe vapor flow through soils.
Under vacuum conditions, the applied gradients dominate the natural
gradients, and the law simplifies tos

                             i. -  * F.
where

     qa » volume flow per unit area
     Kft « air permeability of the soil
     nm » air viscosity
     Pa « applied vacuum

     Obviously, the air permeability plays a key role in determining SVE
applicability and also in system design.  The air permeability incorporates
many soil characteristics Including porosity, structure, grain size
distribution, water content, and preferred flow paths.  Other things being
equal) SVE applicability is a function of air permeability.

SITE INVESTIGATIONS

     The initial step in considering remediation for a site is the
performance of a site Investigation, including a site history review,
preliminary site screening, detailed site screening, and contaminant
assessment.

     During a site history review, all the available Information that may
be useful is assimilated.  Typically, this information will include
reports, maps, and other data.  Sources for this information often include
federal agencies such as EPA, USGS, SCS, FEMA, and state and local
agencies.


                                   120

-------
     Preliminary site screenings consist of a site visit, limited sampling
(often soil gas surveys), and the development of a site map.  The main
objective is to delineate the extent of contamination.  Equipment used
often includes soil gas survey equipment, and a PID, FID, or gas
chromatagraph (GO).

     A detailed site characterization is performed to obtain data needed
for SVE system design.  Data obtained usually include air permeability,
moisture content, soil type and structure, and other data.  Typically, this
step vould include borings, geophysical techniques, and monitoring veil
installation.

     Soil vapor contaminants are normally characterized via soil GC
analysis.  This step also helps to determine potential SVE effectiveness
and may be of value to set cleanup goals.

     Pilot testing is performed to allow for the determination of specific
soil and contaminant properties that will be used in full-scale system
design.

SYSTEM DESIGN

     SVE systems are designed to maximize the air flow through the zone of
contamination.  Vertical wells, trenches or soil piles may be used as
extraction points.  Wells are used where contamination exists at depth,
while trenches are more appropriate where the water table is near the
surface.  Contaminants may also be removed from stockpiled soil, which may
result when soil is excavated from around a leaking tank.

     The radius of influence of the extraction well is the primary design
variable.  The radius, which is determined by the results of a
site-specific air permeability test or during pilot studies, depends most
highly on the applied vacuum and the air permeability of the soil (Figure
5).  The radius of influence can also be affected by the depth to water,
the position of the extraction vents relative to the surface, or the
presence of an impermeable surface seal.  Air inlet or injection wells are
also used to control the radius of Influence.  Inlet wells allow air to
enter the subsurface at specific points, while injection wells force air
into the soil.  Ground water depression pumps may be used with SVE systems
to reduce ground water upwelllng.

EQUIPMENT

     Most systems contain piping, a vacuum pump or positive displacement
blower as the motive force, vapor pretreatment (air/water separator), and
emission control (Figure 6).  The piping is typically PVC, with the
diameter varying with the design air flow.  Both blowers and vacuum pumps
are used.  Air-water separators can be as simple as a 55-gallon drum or may
include refrigerated condenser units, demisters, level controls, and other
devices.  Emission control may be accomplished using granular activated
carbon, incineration, or catalytic oxidation.  Internal combustion engines
and biotreatment have also proved successful.
                                     121

-------
SYSTEM OPERATION AND MONITORING

     SVE can be operated automatically and controlled via microprocessors,
or can be operated manually.  Cleanup progress is assessed by volume flow
rate and vapor concentration monitoring.  These data are normally converted
to a mass flow rate.  Extraction may be continuous or "pulsed" (i.e.,
intermittent).  Pulsed venting may be more energy-efficient.

     Monitoring of vapor phase concentration and composition in extracted
vapors is done on a regular basis.  The data collected are  used to show the
progress made by SVE.  Monitoring typically shows that removal rates drop
as the cleanup progresses, as the most volatile compounds are removed
first.

     SVE has also been shown to enhance levels of biotreatment in soils,
probably due to the inflow of added oxygen.  DePaoli et al. (1989)
calculated that 27.5% of total removal at one site was due to
biodegradation.

SECONDARY EMISSIONS CONTROL

     There are four common options for treating the vapors removed from the
soil by SVE: granular activated carbon (GAC) adsorption, incineration,
catalytic oxidation, and internal combustion engines (ICEs). Each method
has advantages and disadvantages, and no single method is most appropriate
for all situations.

Granular Activated Carbon (GAC)

     GAC is the most common treatment method for vapors, for various
reasons, including familiarity, ease of implementation and operation,
possibility of regeneration, and its applicability to a wide range of
contaminants, concentrations and flowrates.  GAC generally may not be the
most cost-effective option where mass removal rates are high, due to the
increased costs of carbon regeneration or replacement.

Incineration

     Incineration uses very high- temperatures (1400°F or higher) to destroy
vapor phase contaminants.  This method is good for streams with high
concentrations, since the incineration can become self-sustaining at
vapor concentrations over about 10,000 ppmv.  Below this level,
supplemental fuel must be used.

Catalytic Oxidation

     Catalytic oxidation employs a precious metal formulation as a
catalyst, allowing the reaction to occur at temperatures of 800°F,
significantly below the temperature needed for incineration.  This results
in reduced fuel costs.  The incoming vapor stream concentration is limited
to about 3000 ppmv, because at concentrations above this level the heat of
combustion will destroy the catalyst.  These units are not suited for use
in chlorinated vapor streams due to potential catalyst degradation.
                                    122

-------
Internal Combustion Engines

     In southern California, over 100 internal combustion engine systems
have been permitted for use at SVE sites.  Results show the units provide
good reduction in hydrocarbon concentrations (>99X).  These units are
composed of an automobile engine with a carburetor modified to accept vapor
phase fuel.

COSTS

     The costs of SVE may be divided into site investigation, capital,
operation and maintenance, and monitoring costs.

     Site investigation costs include those associated with gathering site
historical data, performing a preliminary site characterization, and, if
needed, a detailed site characterization and contaminant assessment.  For a
typical leaking UST site like a gasoline service station, the site
investigation phase may be limited in scope and cost.

     Capital costs include the costs of procuring and installing all
equipment, piping, and instrumentation necessary for system operation.
Engineering costs are also usually included in this category.  Major cost
factors are typically the well drilling and installation, the blowers, and
the emission control system, if needed.  Fully-contained units are now
available for purchase or lease.

     Operation and maintenance costs consist chiefly of power for the
blowers and pumps and costs associated with the emission control system.
Emission control units, typically incinerators or granular activated
carbon, can dominate the cost of SVE.  Monitoring costs include laboratory
and field analytical costs such as gas chromatograph analysis, which is
used to determine the efficacy of the SVE system and/or cleanup attainment.

     Typical costs for a complete installation may vary over a wide range,
from as little as $10,000 for a small site to well over $1,000,000 for a
large, multi-acre site.  Gasoline service stations may be cleaned for under
$100,000 in some cases.  On a unit basis, reported costs have ranged from
$10/yd  to over $300/yd , with lower costs associated with larger sites.


                                CONCLUSIONS


     Soil vapor extraction systems are being used in increasing numbers for
the removal of volatile organic compounds from unsaturated zone soils.
These systems operate in situ, so there is a minimum of site disturbance
and site operations need not be interrupted; they can treat large volumes
of soil at reasonable costs; and are relatively easy to install, allowing
for rapid mobilization and implementation.  Although these systems appear
to be simple, however, subsurface vapor flow is quite complex.

     A workshop was held to discuss the state-of-the-art of the technology.
Based on the workshop and subsequent research, an SVE Technology Assessment
report was produced, which discusses basic science of the subsurface


                                     123

-------
environment, subsurface vapor flow, SVE system design and operation,
secondary emission control, and costs.
                                REFERENCES
Bennedsen, H.B. 1987.  Vacuum VOCs from Soil. Pollution Engineering.
  19(2):66-69.

Davies, S.H. 1989.  The Influence of Soil Characteristics on the Sorption
  of Organic Vapors.  Presented at the Workshop on Soil Vacuum Extraction,
  U.S. EPA, ORD, R.S. Kerr Environmental Research Laboratory, Ada, OK.
  April 27-28.

DePaoli, D.V., S.E. Herbes, and M.G. Elliott. 1989.  Performance of an
  In-situ Soil Venting System at Jet Fuel Site.  Presented at the Soil
  Vapor Extraction Technology Workshop, U.S. EPA, ORD, Risk Reduction
  Engineering Laboratory, Edison, NJ.  June 27-28.

Dragun, J. 1988.  The Soil Chemistry of Hazardous Materials.  Hazardous
  Materials Control Research Institute, Silver Spring, MD.

EPA. 1988.  "Underground Storage Tanks; Technical Requirements".  Federal
  Register 53, No. 185, 37082-37212, 23 September 1988.

Hutzler, N.J., B.E. Murphy, and J.S. Gierke. 1989.  Review of Soil Vapor
  Extraction System Technology.  Presented at the Soil Vapor Extraction
  Technology Workshop, U.S. EPA, ORD, Risk Reduction Engineering
  Laboratory, Edison, NJ.  June 27-28.

Reible, D.D. 1989.  Introduction to Physico-Chemical Processes Influencing
  Enhanced Volatilization.  Presented at the Workshop on Soil Vacuum
  Extraction, U.S. EPA, ORD, R.S. Kerr Environmental Research Laboratory,
  Ada, OK.  April 27-28.

Stephenatos, B.N. 1988.  Modeling the Transport of Gasoline Vapors by an
  Advective Diffusive Unsaturated Zone Model.  Proceedings of the
  Conference on Petroleum Hydrocarbons and Organic Chemicals in Ground
  Water:  Prevention, Detection, and Restoration.  Houston, TX. November
  9-11. p. 591-611.
                                     124

-------
    TABLE  1.  CHEMICAL PROPERTIES OF HYDROCARBON  CONSTITUENTS
Chemical Class
04
CS
06
07
08
CO
010
Meno-«romtttiea
06
07
06
08
00
010
Phenol
C1-phenotg
02-phenols
03-phenols
C+phenoto
Mara)
tt^ato
Representative Liquid Density
Chemical (a/on.3)
rvButan*
rvPentarw
n-Hexan*
n- Heptane
n-Oetane
n-Nonane
n-Deean*

Benzene
Toluene
m-XyHne
Ethylbenzen*
1,3.S.Trlmethylbenzene
1,4-Diethylbenzene
Phenol
m-Cr»«ol
2,4»Dimethylprttnol
2,4,6'Trlmethyiphenol
m-Ithylphenol
Indanol
Naphthalene
O.S70
OJ26
0.650
0.884
0.703
0.718
0.730

OJ8S
0.867
OJ64
OM7
OJ6S
0.862
1.068
1.027
OMB
m
1X37
m
\JOiS
Henry's
Law
Constant
(dim.)
26,22
29.77
36.61
44 JO
62.00
m
NA

0.11
0.13
0.12
0.14
0.00
0.10
0,036
0.044
0.046
m
m
m
m
solubility
•2e°c
81.1
41.2
12.6
2JO
om
0.122
0.022

1780
SIS
182
167
72.0
16
82000
23500
1800
m
m
m
30
PureVtpof
Pressure
(mrnHo.)
«20°C
1560
424
121
3S.6
10.6
8.2
0.06

7S.2
21.8
8.18
7,08
1.73
0.607
0,620
0.1 S
0.068
0.012
0.06
0.014
0.063
Vaj or Density
4060
1670
670
105
65.6
22.4
7.4

321
110
36.6
41.1
11.4
S.12
2.72
OJ0
0.30
O.Ofl
OJ3
0.1
0.37
Soil Sorptlon
Conitsnt (toe)
(Ukc)
@25°C
250
320
600
1300
2600
6800
13000

36
. 00 '
220
210
100
1100
110
6.4
m
m
m
m
600
NOTE:  NA - Not available
SOURCE:  Compiled from various published and unpublished sources.
                                  IIS

-------
 VAPOR
PRESSURE
   SVE
LIKELIHOOD
    OF
 SUCCESS
  SOIL AIR
PERMEABILITY
  TIME
 SINCE
RELEASE
I
Butane — ^-
Pentane — ^-
Benzene — ^~

Toluene — ^~
Xylene — ^-
Phenol — ^"
Naphthalene — ^-

Aldicarb — ^-
\

•«« (
1
3
-102

-101

10'1
-ID"2 |
!
-io-3 ;
-10- j
%
r

f
SUCCESS
\/r~D\/
VhKY
LIKELY



SUCCESS
SOMEWHAT
LIKELY
I
^•.
• SUCCESS
x LESS
f LIKELY


ills
^^^^^t^^j
wiiiii
AfiSgR^^^^^K
^^^^^
^^^g^
^^^^
^^^^
^^^^^^m

*M.
*jj$$-^"i
f^i
a.





HIGH
(gravel,
coarse
sand)



MEDIUM
(fine sand)


LOW


Weeks
Months
Years

Weeks

Months
Years
Weeks
Months
Years













        Figure 1. SVE Applicability Nomograph
                      126

-------
                     COMPOSITION OF GASOLINE
  (45%)
                     (10%)
              OLEFINS




              AROMATICS




              ALIPHATICS
                            (6%)
                               (45%)
                             SOURCE: R.L. JOHNSON, 1989
                                                                 (4%)
                                                  (90%)
          FRESH GASOLINE
                            WEATHERED GASOLINE
§


£
ui
Q.
o

UJ
               Figure 2. Effect of Weathering on Gasoline Composition
                       LEGEND
            184
     After: DePaolietal. (1989)
190
212
218
243
490
                           VOLUME OF EXTRACTED GAS (103 n? )
748
            Figure 3. Variation of Hydrocarbon Distribution in Extracted Gas
                                       12?

-------
cc
    too-
    90-
     80-
     70-
     60H
     eo-
     40-
     30-
     20-
     10-
           AIR
      PERMEABILITY
     O A  Surface soil

     • D  Subsurface soil

Source: Adapted from Corey (1957)
               I
               10
            20     30
                                                WATER
                                             PERMEABILITY
 40     SO      60

Water Saturation %
         Figure 4. Air and Water Permeability as a Function of Water Content
                                         128

-------
                   EXTRACTION WELL
                                         SURFACE SEAL
                                                             FLOW LINES
           Figure 5. Relationship of Flow and Induced Vacuum
  ;. SECONDARY
  :\ EMISSION
                               WATER COOLED
                              HEAT EXCHANGER
PRESSURE
 GAUGE
      HEADER
SURFACE SEAL
V


v r '> >( »> >'* 't'r>"'
BENTONITE jf
CEMENT/1^
GROUT
V /
SAND



S
|
'&
$f
1



JV
"-",



H
"?
I
:•>
txt= I
,K_xl 	 	 	 KJL-1 ,
riin Cnfj

0.020 SLOT
t^ SCREEN

PRESSURE
 RELEASE
  VALVE
                                                                SILENCE
                                                                MUFFLER

DM
V-s
Qy—
txj=
l_ 	




BALL
VALVE
_Kjt"| —


AIR-WATER
5EPARATOF
. • •"

r\'~





T
                                GROUNDWATER    '
                                 EXTRACTION  STRAINER
                                                        PUMP
              BLOWER
                                        V
          TO
     GROUNDWATER
       TREATMENT
        SYSTEM
PACK
          Figure 6. Typical Soil Vapor Extraction System Schematic
                                  129

-------
         ALGASORB®: A NEW TECHNOLOGY FOR REMOVAL AND RECOVERY OF METAL
                       IONS FROM GROUNDWATERS

                   By: Dennis W. Darnall,  Sandy Svec  and Maria Alvarez
                       Bio-Recovery Systems,  Inc.
                       P. O. Box 3982, UPB
                       Las  Graces, NM   88003
                                     ABSTRACT
       A  new  sorption process for removing  toxic metal ions from  water has been
developed.  This  process  is based  upon the natural, very strong affinity  of
biological materials,  such  as the cell  walls  of plants  and microorganisms, for heavy
metal ions.   Biological  materials,  primarily algae, have been  immobilized in a
polymer  to  produce a "biological"  ion exchange resin,  called  AlgaSORB®.  The
material  has a remarkable  affinity  for heavy  metal ions and  is  capable  of
concentrating these  ions by a  factor of many  thousand-fold.    Additionally,  the
bound  metals  can  be stripped  and recovered  from  the  algal  material  in  a manner
similar  to  conventional  resins.
       This  new  technology has been  demonstrated to  be an  effective method  for
removing  toxic  metals from  groundwaters.   Metal  concentrations can be reduced to
low  parts per  billion (ppb) levels.   An  important characteristic of the  binding
material  is  that  high concentrations of  common  ions  such  as  calcium,  magnesium,
sodium, potassium,  chloride and sulfate  do not interfere  with  the  binding of heavy
metals.   Waters containing a total  dissolved  solids (TDS)  content of several thousand
and  a  hardness of  several hundred parts per  million  (ppm)  can be successfully
treated to remove  and recover heavy metals.    The process  has  been  demonstrated
under the SITE  Emerging  Technology program for the effective removal of
mercury   from  a   contaminated groundwater.

                  This paper  has   been  reviewed   in accordance with
                  the U.  S. Environmental  Protection   Agency's peer
                  and administrative review  policies  and approved
                  for presentation  and publication.
                                         130

-------
                                   INTRODUCTION
       In recent years  there  has been increased attention  focused on pollution  of
water supplies by  heavy metal  ions.   These metals are toxic in rather low
concentrations  and  can lead  to  acute and  chronic  illness  in  humans and  other
animals.  Past  waste  disposal  practices  have  resulted in serious  contamination  of
the environment.   The major sources of heavy  metals are  leachates   from  legal  and
illegal landfills  and drainage  from old mines.   Several of the sites on the National
Priorities List are  landfill  facilities  that have problems  caused  by  Improper
placement  of heavy metal  wastes  below ground.

       Chemical  treatment using  lime or caustic precipitation  has been  the  most
common method  of removing dissolved metals  from  leachates, mine drainages,  and
contaminated  aquifers.    Other  treatments  include  reverse  osmosis,   electrodialysis,
and  carbon adsorption.   The energy  and chemical  costs  of these various methods
become  major operating expenses  during  the use  of these  treatments.   A  serious
limitation of many  of the  current  treatment  technologies  is the difficulty  and/or
expense  of treating  waters for  removal of  heavy  metals  to allowable drinking
water  levels.    Alternatives for  economical  recovery  of  dissolved  metals  from
contaminated  waters are limited.   The development of rapid, widely  applicable,  low
cost methods for  the  removal, recovery and recycling of  heavy metal ions  from
contaminated  waters at Superfund  sites  is  a  high priority.

       For  a  number of years  Bio-Recovery Systems has been  working  with  a new
sorption process for removing  heavy  metal  ions  from water (1-4).   This sorption
process  is  based  upon  the natural,  very strong affinity of the cell walls  of algae  for
heavy  metal ions.   Algal cells  have  been  immobilized in a  silica  gel polymer
(AlgaSORB®)  and used much as an  ion-exchange  resin.  The algae  are killed in  the
immobilization  process indicating  that sorption  does  not require a  living
organism, and  hence the algal matrix  can be exposed, with  little or  no ill  effects,  to
solution  conditions  which  would normally  kill living cells.    The pores of  the
polymer are apparently large enough  to allow  free  diffusion of  ions  to  the  algal
cells, since similar concentrations  of  metal  ions  are bound by  free  and immobilized
cells.

       AlgaSORB® functions as a  "biological" ion  exchange resin and binds both
metallic  cations  and metallic  oxoanions.   Anions such  as chloride or sulfate are
only  weakly bound or  not bound  at all.   Like ion-exchange resins,  the  algae-silica
system  can  be recycled.  Metal  ions  have been  sorted and stripped  over  as many  as
75  cycles with  no  noticeable loss in  efficiency.   In  contrast to current ion
exchange technology,  however,  another  real  advantage  of  the  algae-silica  matrix
is  that  the components of hard  water  (Ca+^t Mg+2)  or monovalent cations (Na+, K+)
do  not  significantly  interfere  with the binding  of toxic, heavy  metal  ions (2, 3).
The  binding of Ca+2  and  Mg+2  to ion-exchange resins  often  limits   Its  usefulness
                                          131

-------
since  these  ions  are  frequently present in  high concentrations  and  compete  for
heavy  metal  ion  binding.   Frequent  regeneration of  ion-exchange resins is
necessary in  order  to  remove heavy  metal ions from  solution.   Another advantage
of  AlgaSORB® is  that it has a high  affinity for metal  ions.   Effluents  from
AlgaSORB®  columns frequently show that metal ions  can  be reduced to the part  per
billion level.   AlgaSORB® can be  used to remove  aluminum, cadmium, chromium,
cobalt, copper, gold,  iron, lead,  maganese,  mercury,  molybdenum, nickel,
platinum, silver, uranium,  vanadium, and zinc.   Many of these metals  are
hazardous wastes  at   Superfund sites.

       A major advantage of AlgaSORB®  is that the efficiency of heavy metal  ion
removal  is  not diminished by  the presence of organic  compounds  in the  water.   For
example, waters containing  copper ion  and  high concentrations  of  organics  such
as  butyl cellosolve, alcohols,  other  etchers  and halogenated  hydrocarbons  have
been successfully treated  (4).   Humic  and  fulvic  acids  seem  to have little  or no
effect  on metal ion  binding to  AlgaSORB®.   The presence  of organic compounds
limits  the  usefulness of  ion exchange resins  since organics  often   will bind  to  these
resins  and  decrease their metal  binding  capacity.

       In 1988, Bio-Recovery Systems was selected  to participate  in the  Superfund
Innovative Technology Evaluation  (SITE)  Emerging  Technologies   Program.    The
goal of  the project  was to test  AlgaSORB® for the removal  of mercury  from  a
contaminated   groundwater.


                          DESCRIPTION OF  GROUNDWATERS
       A  number of years ago  an industrial process using  mercury resulted  in  soil
contamination  with  elemental  mercury.    The  mercury  subsequently  percolated
through the  soils and  contaminated groundwater.   At  some  point  the  mercury  was
oxidized to the  bivalent oxidation  state and  was  found at  various  concentrations in
the groundwaters depending  upon  the monitoring site.   Currently,  the
groundwaters  are extracted  from  an  upper perched  groundwater table  via  a
drainage gallery.   A facility has  been constructed to treat extracted  groundwaters
by  the use  of precipitation  with  dithiocarbamates,  followed by  polishing with
activated carbon  and a  specialty  ion exchange  resin.    The  water is pumped  from
the gallery  at mercury  concentrations of  0.1-3.0  ppm  and   is currently  treated to
allowable  discharge limits  of 10  ppb  mercury.

       Wells  monitoring  the  ground  water during the  late   1980's  showed seasonal
variations  in  the mercury concentrations.   It appears  that  mercury  levels  decrease
in the dry seasons compared to the  rainy  seasons.  Chemical speciation of the
mercury  in  the  groundwaters  was  not rigorously determined,  but speciation
studies on  soils overlying the  groundwater  indicated  the  predominant  species  was
oxidized inorganic  mercury.   The  composition of other elements in  the
groundwater  seems to  change with  the  seasons  as well,  but  an  average composition
is given in Table  1.   Variations in  mercury content over  a four-year monitoring
period in  waters from  two wells,  which  are well  separated from one another,  are
shown in  Table 2.


                                        132

-------
  TABLE 1. AVERAGE COMPOSITION OF MERCURY-CONTAINING GROUNDWATERS

             Constituent                  Concentrations   (mg/L)
              Chloride                              5,800
              Sodium                                2,900
              Calcium                                460
              Magnesium                             440
              Total Dissolved Solids                  11,000
              pH                                      8.0
          TABLE 2, SEASONAL VARIATION OF MERCURY CONCENTRATION IN
                             MONITORING WELLS
Month/Yr
Oct/1
Nov/1
Dec/I
Jan/2
Mar/2
Apr/2
May/2
Sep/2
Dec/2
Feb/3
Sep/3
Dec/3
Apr/4
May/4
Jun/4
Aug/4
Sep/4
Oct/4
Well 1
(mg/L)
9.60
3.35
0.29
5.50
3.80
10.00
4.20
7.70
6.10
6.20
8.50
2.70
4.00
4.00
4.40
5.80
7.70
13.00
Well 2
(mg/U
0.370
0.293
0.426
0.230
0.390
0.200
0.300
0.370
0.510
0.500
0.240
0.140
.
0.260
0.170
0.180
0.086
0,240
                         EXPERIMENTAL PROCEDURES
    £ Mercury analyses were performed using  the EPA Method  245.1 of cold vapor
atomic absorption spectroscopy (5).  A Perkin  Elmer  Model 3030B AAS instrument
was calibrated daily  for mercury, and a  calibration  verification  record was
maintained using data collected by  the analysis of EPA certified check standards.
Preparation  of standards for mercury  analysis were performed in  accordance with
                                    133

-------
the specifications  in  Methods  for  the  Chemical Analysis of Water and Wastes (5).
Spiked  samples were  analyzed with each  batch  of samples to  determine  if  matrix
interference  existed,  and  frequent  blanks  were  run to  ensure  there  was  no
mercury  carry  over  during  analysis.

Mercury  concentrations  in  groundwaters,  column  effluents  and  regenerating
solutions  were determined  by  linear  regression  calibration  curves  generated from
four  point  standard  calibration analysis,  (5).

       Samples  collected  in  the  field pilot studies  were split and  sent to Woodward-
Clyde Consultants, EER  Technologies and  Bio-Recovery  Systems for mercury
analysis.

       Laboratory  tests on  the efficiency of mercury adsorption on AlgaSORB®
were  determined using  small glass  columns (1.5  cm i.d. x  20 cm)  which contained
the  sorbent.    Mercury-containing  groundwaters  were  pumped  through the
column at flow rates which varied from 6-20  bed volumes  per  hour .   Effluents
from  the  columns were  collected  using a fraction  collector  and mercury  content
was  determined.  Once the columns became saturated  or leaked mercury  above
discharge  limits (10 ppb), the column was stripped  with  10-bed volumes  of a
selected  stripping reagent followed  by  10-bed volumes  of deionized  water.
Analyses  of  stripping  effluents  were  performed  to verify  stripping.

       Pilot  studies  were  conducted  with  a small  portable effluent  treatment  system
which has  two AlgaSORB®-containing  columns  in  series  and which is  capable of
treating flows of up to 0.5 gpm.   On-site pilot testing was conducted on August 30  to
September 1,  1989 and November 6  to December 1, 1989.


                            RESULTS AND DISCUSSION
       Samples  of groundwater were  collected  at  various  times  during  1989.   All
samples  were acidified to pH 2 with nitric acid in the field  prior to transport for
laboratory  studies.  Once the samples  were received at Bio-Recovery  Systems,  the
solutions  were  neutralized  to the  original  pH  with  dilute sodium hydroxide.
Laboratory and  field studies  were  complicated by  the  fact that  over a  10 month
period,  mercury  concentrations changed  by an  order of magnitude.   Table  3  shows
mercury  concentration  variation  over  the sampling   period.

       Different  species  of  algae  can be  immobilized  to produce  different
AlgaSORB®  resins.   Since different bioploymers comprise the cell  walls of different
algae,  some  species  of algae behave  differently  from  others  with respect to metal
ion binding.   Thus, different AlgaSORBs  containing  different  algal species were
tested  for mercury removal  from  the  groundwaters.   However,  these experiments
were  complicated  by  the  fact  that  consistent  mercury removal  performance  was
not observed  using a  single immobilized  alga  on waters collected  at  different times.
For example, Table  4 shows mercury  contents in  effluents  from  columns
containing  AlgaSORB  602.   The results shown  in  Table 4 can not simply be
explained by  variation in  mercury content in  the  influent  waters.   For  example,
                                        134

-------
Column  C  shows  lower leakage  levels of  mercury  than  column B  in the first 40 bed
volumes  of effluent  even  though  the  influent  concentration  of  mercury  was
nearly three times higher in column  C  than column B.   This  suggests that  a
variation  in the  chemical  species of mercury may  occur  with time.


             TABLE 3. MERCURY CONCENTRATIONS IN  GROUNDWATERS
Sample Number
103-13089
176-42089
177-42089
265-070589
343-090189
368-100489
369-100489
PH
8.5
8.0
8.0
7.9
7.8
7.9
7.9
Mercury
Concentration
(f-ig/L)
150
435
144
1120
620
1550
1550
Date
Collected
01-30-89
04-20-89
04-20-89
07-05-89
08-31-89
10-04-89
10-04-89
 TABLE 4. TEST OF AlgaSORB 602 ON MERCURY-CONTAMINATED GROUNDWATERS t

Bed Volumes
of Effluent
1-4
17-20
37-40
54-60
69-72
73-76
93-96
113-116
133-136
149-152
A*
Effluent Hg
(ng/L)
0.5
0.8
1.3
4.0
-
2.2
2.3
3.0
5.0
6.5
B*
Effluent Hg
(ng/U
9.9
10.1
21.8
14.8
-
31.0




C*
Effluent Hg
(Hg/0
1.3
3.4
8.1
27.0
72.5





     t Groundwaters collected at various times were pumped at a flow rate of 6 bed-volumes per hour
       through identical columns containing AlgaSORB 602.  Effluents from each column were collected
       in four bed-volume fractions and analyzed for mercury content.  Influent  flow rate was  six bed
       volumes per hour.

     * Column A influent  water was  collected January 30, 1989, and had  an  influent mercury
       concentration of  150 jigAL. Column B influent water was collected April 20, 1989, and had an
       influent mercury concentration  of 435 ng/*L. Column C influent water was collected July 5, 1989,
       and had an influent mercury concentration of 1120  jig/L,
                                           135

-------
       Over 99 percent  of the  mercury was  stripped from  the columns in Table 4  by
the passage of  10  bed volumes of 1.0 M sodium  thiosulfate through the column  (data
not  shown).

       After examining  several  different AlgaSORB®  preparations and  noting
similar types of variations as shown  in  Table  4, it  was decided to settle on two
different  AlgaSORB®  resins for final testing.   While these two resins could have
been  blended  into  a single column, they  were placed in  two columns  which were
connected in  series and from  which  effluents  samples could be taken from  each
column for mercury  analysis.  Table  5  shows results of  these  experiments.

       Data in Table  5 show  that  the two  columns  arranged in  series were effective
in mercury  removal to below one ppb through passage of 250  bed volumes of
mercury  contaminated  waters  which  contained 15SO ppb  mercury.

       On*site pilot testing of  the AlgaSORB® resins  was performed  from November
6, 1989,  to  December 1, 1989.  Two  columns  (2.54 cm i.d. x 81 cm)  were separately
filled with  AlgaSORB 624 and AlgaSORB  640.  The columns each had a bed volume  of
400  raL  and were  connected in  series.   Mercury-contaminated  waters were pumped
through the two columns  and  two  bed-volume fractions (800 ml)  were collected,
split  and sent  to EER Technologies, Woodward-Clyde  and  Bio-Recovery Systems  for
analysis.   Results of on-site  pilot testing  are  shown  in Table 6.


            TABLE 5. TEST OF AlgaSORB 624 AND AlgaSORB 640 ON  MERCURY-
                         CONTAMINATED GROUNDWATERS*
          Bed Volumes of Effluent                   Effluent  Hg  (u,g/L)
0-12
12-24
24-36
48-60
60-72
84-96
108-112
132-144
168-180
192-204
252-264
288-300
312-324
324-336
0.3
0.2
0.2
0.3
0.5
0.7
0.8
0.9
0.8
0.9
0.6
0.6
2.0
1.9
   *Two columns (1.0 cm i.d. x 37 cm) coupled in series were used.  Groundwaters collected October 4,
   1989, and containing  1550 fig/l_ mercury were passed through the columns at a rate of six bed-volumes
   per minute.  Ten bed  volume fractions were collected and analyzed  for mercury.  Data shown above are
   mercury concentrations in effluents  from the second column.
                                        136

-------
       By  the time  the on-site testing  had  begun  in  November the mercury
concentration in  the  groundwaters had  changed from  about 1500 ppb (in  October)
to about 700 ppb  (see  Table 6).   With  the  exception  of the first fraction collected,
the data in  Table 6  shows  that  over 500 bed  volumes of mercury-contaminated
waters  were  treated  before mercury  in effluents  approached the  10  ppb  discharge
limit.
        TABLE 6.  ON-SITE PILOT TESTING FOR MERCURY REMOVAL FROM
                   GROUNDWATERS t
                                           Mercury  Concentration  fu.g/L't
 Bed  Volumes of                 Bio-Recovery   Woodward Clyde  EEiR  Technologies
    Effluent        Influent      Analysis          Analysis          Analysis
7-8
85-86
163-64
229-230
289-290
313-314
343-344
379-380
415-416
449-450
467-468
503-504
533-534
587-588
	 A1
9.5
5.3
2.1
1.4
1.8
1.9
5.5
2.0
1.8
4.9
4.0
5.8
7.7
10.5
fi* ££H
14.2
8.0
3.6
1.4
2.6
. 2.4
9.3
3.1
3.2
7.8
7.2
9.6
10.3
13.0
7«n
11
<10
<10
<10
<10
<10
10.0
<10
<10
10.0
<10
<10
<10
15
6on
                       600*          770
     t  A portable  water treatment  system  was equipped with two columns connected  in series.  The first
       column was filled with AlgaSORB 624 and the second was filled with AlgaSORB 640.  Groundwaters
       were pumped through the system at a flow rate of 6 bed-volumes per hour.  Effluent samples were
       collected  and sent to Woodward Clyde Consultants, EPA (EER Technologies  Corporation) and Bio-
       Recovery Systems for analysis.

     *  Influent samples were  collected for analysis just prior to 436  and 600 bed  volume fractions of
       effluent were collected.
                                     CONCLUSIONS


       Initial laboratory  testing of  AlgaSORB®  resins clearly showed promise for
mercury  recovery from  contaminated  groundwaters.   Once  the  mercury was  loaded
on  the resin, it could  be stripped  with sodium thiosulfate.   On-site pilot testing
confirmed the  laboratory  studies and  showed  that  AlgaSORB®  resins treated  over
                                          131

-------
500  bed volumes of  groundwaters before mercury  levels  in  effluents  exceeded
discharge  limits.
                                    REFERENCES
     I.  Darnall, D.W., Greene, B.,  Henzl, M., Hosea, M., McPherson, R,, Sneddon, J.
         and Alexander, M.D.,  Binding and Recovery  of GoId(IH) and  Other  Metal
         Ions from Aqueous Solution  by Algal Biomass, Environ. Sci. Technol. 20:
         206, 1986.

     2.  Darnall, D.W., Greene, B.,  Hosea, M,, McPherson,  R., Henzl, M,, and
         Alexander,  M.D., Recovery of Heavy Metal  Ions  by Immobilized Algae, 1m
         R. Thompson  (ed.), Trace  Metal Removal From  Aqueous Solution,  The Royal
         Society  of  Chemistry, London, 1986, p. 1.

     3.  Greene, B.,  McPherson, R. and Darnall,  D,W.,  Algal  Sorbents  for  Selective
         Metal Ion Recovery, In:   J.W.  Patterson  and R. Passimo (eds.), Metals
         Speciation,  Separation and Recovery.   Lewis  Publishers, Chicago,  Illinois,
         1987,  p. 315.

     4.  Daraall, D.W. and Gabel, A., A New  Biotechnology for  Recovery  Heavy
         Metal Ions  from Wastewater, Inj.   Proceedings of the Third National
         Conference  on New Frontiers for  Hazardous Waste Management,  U.S.
         Environmental Protection  Agency,  EPA-600/9-89-072,  1989,  p. 217.

     5.  Methods for the  Chemical  Analysis of  Water and Wastes.   EPA-600/4-79-
         020, U.S. Environmental  Protection Agency,  Revised  March 1983 and
         subsequent  EPA-600/4 Technical Additions  Thereto,  Cincinnati,  Ohio, 1983.
                                        138

-------
          A SITE DEMONSTRATION OF ULTROX UV/OXIDATION TECHNOLOGY

                             Norma Lewis, M.A.
                   U.S. Environmental Protection Agency
                  Risk Reduction Engineering Laboratory
                             Cincinnati,  Ohio

                       Kirankumar Topudurti, Ph.D.
                        Gary Welshans,  Ph.D.,  P.E.
                            Robert Foster,  P.E.
                    PRC Environmental Management, Inc.
                             Chicago, Illinois
                                 ABSTRACT

      This paper presents the field demonstration test activities and test
results of the ultraviolet  (UV) radiation/oxidation technology developed by
On-Site Ultrox International of  Santa Ana,  California.   The  technology
simultaneously uses UV radiation, ozone, and  hydrogen peroxide  to oxidize
dissolved organic contaminants present  in groundwater or wastewater.   The
demonstration was performed  in  February and March of  1989 at the Lorentz
Barrel and Drum  (LB&D)  site  in San Jose, California,  under  the Superfund
Innovative Technology Evaluation (SITE) program.

      The Ultrox system was evaluated for its effectiveness in treating the
volatile organic compounds (VOCs) present in groundwater at the  LB&D site.
Experiments were conducted in which hydraulic  retention time, ozone dose,
hydrogen peroxide dose, UV radiation intensity, and influent pK  level were
varied over a wide range to assess the system's performance.

      The Ultrox system  achieved VOC  removals of  greater than 90 percent.
Most VOCs were removed through chemical  oxidation.  However, for a few VOCs
such  as 1,1,1-trichloroethane  and  1,1-dichloroethane,  stripping  also
contributed toward  removal.   The treated  groundwater met the  applicable
discharge standards  at 95 percent confidence  level  for discharge  into  a
local waterway.  There were no harmful air emissions to the  atmosphere from
the Ultrox system,  which is equipped with an off-gas treatment unit.

      This paper has been reviewed in accordance with U.S.  Environmental
Protection Agency's  peer and administrative review policies and approved
for presentation and publication.


                                    13!

-------
                               INTRODUCTION

      Technologies  to destroy, treat, detoxify, reduce mobility,  or volume,  or
recycle  hazardous waste  materials  are  being  developed and  demonstrated
within the Environmental Protection Agency (EPA).  The Superfund Innovative
Technology Evaluation  (SITE)  Program  was  created  in 1986  to  provide
information on alternative and  innovative technologies.  The  SITE program
also generates reliable performance and cost data  for  these  technologies
from each demonstration, as well as  a broader range of data on each process
from non-SITE activities.

      The On-Site Ultrox International's proposal for ultraviolet radiation
(UV)/ oxidation technology was selected  in 1988 for the SITE program, by the
EPA's Office of Research and Development (ORD) and Office of Solid Waste and
Emergency Response (OSWER).  Upon approval of EPA's Region IX and agreement
between  On-Site  Ultrox  International,  ORD, and OSWER,  the  technology was
demonstrated  at   the  Lorentz Barrel  and  Drum  (LB&D)   site  in San  Jose,
California.

                   LORENTZ BARREL AND DRUM SITE HISTORY

      The  LB&D site located  in San Jose,  Santa Clara  County,  California,
was used primarily  for  drum  recycling operations  from  about 1947  to 1987.
The drums contained residual  aqueous wastes, organic solvents, acids, metal
oxides,  and oils.  The preliminary site  assessment report for the LB&D site
showed that the  ground  water  and soil were contaminated with  organics and
metals (1).

      The  upper  aquifer at the  LB&D site was  selected  as the  waste stream
for  evaluating  the  UV/oxidation technology.   Samples  from the  shallow
aquifer  were  collected  in December 1988 which  indicated volatile organic
compounds  (VOCs) were  present.   VOCs  detected at  high levels  included
trichloroethylene (280  to  920 jUg/L), vinyl chloride  (51 to  146 jUg/L),  and
1,2-trans-dichloroethylene (42 .to 68 jUg/L).  The  pH and alkalinity of the
ground water  were about 7.2  and 600 mg/L as  CaC03, respectively.   These
measurements  indicated  that  the  bicarbonate ion (HC03~), which acts  as an
oxidant  scavenger, was  present at high  levels.  Other  oxidant scavengers,
such as  bromide, cyanide, and sulfide were not detected.

              UV/OXIDATION EQUIPMENT AND PROCESS DESCRIPTION

      Major components of the  Ultrox system include the UV/oxidation reactor
module,  an air compressor/ozone generator module,  a hydrogen peroxide feed
system,  and a catalytic ozone decomposition (Decompozon) unit.  The Ultrox
treatment system uses UV radiation,  ozone,  and hydrogen peroxide to oxidize
organics in water.   An isometric  view  of  the  Ultrox  system is  shown in
Figure 1.
                                    140

-------
      During the operation, contaminated water first comes in contact with
hydrogen peroxide as it flows through the  influent  line to the reactor.  An
in-line static  mixer is used  to  disperse the hydrogen peroxide  into the
contaminated water.  Next,  the water comes in  contact with UV radiation and
ozone as  it flows  through  the reactor at  a specified rate  to  achieve the
desired hydraulic  retention  time.   As the  ozone gas  in the  reactor  is
transferred to the  contaminated water, hydroxyl radicals (OH°) are produced.
The hydroxyl radical formation from ozone  Is catalyzed by UV radiation and
hydrogen peroxide.   Hydroxyl radicals are  known to  react with organlcs more
rapidly than the oxidants, ozone, hydrogen peroxide,  and UV radiation.  Ozone
that is not transferred to the contaminated water will be  present in the
reactor off-gas, which will be ultimately destroyed by the Decompozon unit
before being vented to the atmosphere.

      The demonstration reactor (Model PM-150) has a volume of 150 gallons
and  is  3  feet  long  by 1,5 feet  wide by  5.5 feet high.   The  reactor  is
divided by  5  vertical baffles into  6 chambers  and contains 24 lamps (65
watts each) in  quartz  sheaths.  The UV lamps  are Installed vertically and
are  evenly  distributed throughout  the  reactor  (four lamps  per chamber).
Each chamber has one  stainless steel sparger  that  extends  along the width
of the reactor.  These spargers uniformly diffuse  ozone gas, supplied from
the ozone generator module, from  the base  of  the reactor Into the water.

      The Decompozon unit  (Model  3014 FF) uses  a nickel-based proprietary
catalyst  to convert  reactor  off-gas ozone to  oxygen.   The Decompozon unit
can accommodate flows  of up to 10 standard cubic  feet per  minute and can
destroy ozone concentrations  in  ranges of 1  to  20,000 ppm  (by weight)  to
less than 0,1 ppm.

                         TECHNOLOGY DEMONSTRATION

      The demonstration was designed to meet the  following objectives:  (1)
evaluate  the performance of the Ultrox system in treating the VQCs present
in the ground water at the LB&D site under different operating conditions;
(2) determine the  extent of  VOC  stripping; (3)  evaluate the efficiency of
the Decompozon unit to destroy reactor off-gas ozone; and (4) determine the
operating conditions needed  for  the effluent  to  meet applicable discharge
standards (NPDES),

TESTING APPROACH

      The test  operation was  designed for 11  test  runs.  After these runs,
two additional runs were performed to determine if  the system's performance
was  reproducible.   These  operating conditions are summarized  in Table 1.
All 13 runs were performed over a period  of two weeks.

      Five  operating parameters   including (1) influent pH,  (2) retention
time,  (3) ozone dose,  (4)  hydrogen  peroxide dose,  and (5) UV radiation
intensity were  adjusted within specified ranges.   The  initial operating

                                     141

-------
conditions  (Run 1), were  based on  the results  of a treatability  study
performed by Ultrox on the LB&D site ground water for Region IX.

      Three indicator VOCs were selected for preliminary testing after each
run.  These were trichloroethylene (TCE, a major volatile contaminant at the
site), 1,1-dichloroethane (1,1-DCA), and 1,1,1-trichloroethane (1,1,1-TCA).
The latter two were selected because these VOCs are relatively difficult to
oxidize.  These indicator VOCs were analyzed overnight.  Once the preferred
level  (operating  conditions  at which the  effluent concentrations  of
indicator VOCs  were  below  NPDES limits, and the  relative  operating  costs
were the lowest) was determined from the overnight analysis, that parameter
remained at  that level  for  the remaining runs.   The pH was adjusted by
adding sulfuric  acid in  the  first three runs.  After the  preferred values
were determined for all five  operating parameters,  two runs  (12 and 13) were
performed to verify the reproducibility of the Ultrox system's performance.

SAMPLING AND ANALYTICAL  PROCEDURES

      Air and liquid samples were collected from the Ultrox system at nine
locations.  The  liquid samples were collected from:  (1)  hydrogen peroxide
feed tank; (2) acid feed tank  (when the pH was adjusted);  (3) raw influent
line; (4) influent line  after acid and hydrogen peroxide  addition (when pH
was adjusted);  (5)  mid-point of the reactor; and (6)  effluent  line.  The air
samples were collected from:   (1) ozone gas feed line; (2)  reactor off-gas
line; and (3) treated off-gas.

      For  the  critical  parameters  in this  study  (VOCs  in water),  six
replicate samples  were   collected.    Duplicate  samples were  collected for
several noncritical parameters.  Sampling at the influent port began  about
15 minutes after each run was started.   At other locations in the reactor,
sampling began  after three  retention  times  to allow the system  to  reach
steady state.  All samples for off-site laboratory analysis were preserved
as required before being  shipped to the laboratory.  Details on the quality
assurance and quality control procedures are presented in the Demonstration
Plan and the Technology  Evaluation Report (2,3).

                                 RESULTS

      Results of the Ultrox system are summarized  to present the overall
effectiveness of the UV/oxidation  technology in  removing  VOCs  from  the
ground water at the LB&D site.  The removal efficiencies  and concentration
profiles of all VOCs are not presented  in this paper.

REMOVAL OF VOCS

      Based on overnight analysis performed during the demonstration  (when
two of the six replicates determined the average effluent concentrations for
each indicator VOC) ,  Runs 8 and 9 showed that  the effluent met the discharge
standard at either set of conditions.   Since a lower hydrogen peroxide dose


                                   142

-------
was used in Run  9,  compared to Run  8,  Run 9 was chosen as  the  preferred
operating run.  However,  based  on a complete analysis of the four  remaining
replicates for Run 9  performed  after  the  demonstration,  the mean
concentration of 1,1-DCA was found to be slightly higher than  5  Mg/L,  the
discharge standard for the  VOC.   Since Run 9 had the preferred  operating
conditions during the demonstration,  the verification runs (12 and 13) were
also performed at those conditions.

      A comparison of 95 percent upper  confidence limits was made  for  the
effluent VOCs  in Runs 9, 12, and 13.   These  limits were calculated using  the
one-tailed Student's t-test.  The effluent met the discharge standards  for
all regulated VOCs at 95 percent confidence level in  Runs 12 and 13.  In  Run
9, the mean concentrations  for 1,1-DCA  and 1,2-DCA  exceeded the  discharge
standards.

      Percent  removals  for  the  indicator VOCs  and total  VOCs  decreased
considerably in Run 7 (Figure 2).   This appears to be  due to the  decreased
ozone dose applied in Run 7.  Removal efficiencies for TCE were  higher than
those for  1,1-DCA and  1,1,1-TCA.    This  finding is  consistent  with  the
rationale used in selecting the indicator VOCs.

      To determine the extent of VOC removal attributed to  stripping within
the treatment system, VOC samples were collected from  the reactor off-gas.
A total of 25 samples were  collected during the demonstration.   Four VOCs
frequently  detected  were  TCE,  vinyl  chloride,  1,1,1-TCA and 1,1-DCA.
Reactor off-gas VOC  emission rates were compared to the VOC removal rates
(estimated by difference between  the VOC  input rates  at the influent  and
output rates  at  the  effluent  ports  of the  Ultrox  system).   Results  are
summarized in Table  2.  The  extent of stripping for any particular VOC is
expected to be proportional to  the ratio of air  flow rate to the: water flow
rate.  The ratio for  Runs 1  to 5 is  approximately 2;  for Run 6 and Runs 8
to 13, it is about 4.5;  and for run  7,  it  is 1.  If stripping  contributed
to the total removal  of the four VOCs, the  extent of stripping should be  the
least in Run 7, and the most in Runs 6 and 8 to 13.  The data presented in
the table follow this trend  for three of the four VOCs (except for the vinyl
chloride in Runs 6,  7, and  9).  A quantitative correlation of the  extent of
stripping cannot be made because the operating conditions in each run were
different.   For example,  at a given air  to water flow ratio,  when  oxidation
doses are  varied,  the extent  of oxidation  also varies.   Therefore,   the
extent of stripping will be indirectly affected,

      Henry's law constants  for the  four VOCs are also presented in Table
2.   When comparing  these  constants with the  VOCs,  their volatility  is
expected to increase from left to right:

            1,1-DCA  ---> TCE ---> 1,1,1-TCA ---> vinyl chloride

      However, significant removals for  1,1,1-TCA and 1,1-DCA weire observed
to be due  to  stripping.   Conversely, the extent of stripping was  low  for

                                    143

-------
vinyl chloride and TCE which possess double bonds between the carbon atoms.
Based  on these results,  stripping  is a significant removal  pathway for
compounds that are difficult to oxidize.

DECOMPOZON UNIT

       Ozone  concentrations  in the influent to  and  the  effluent from the
Decompozon  unit were  analyzed in each  run.   The  effluent ozone
concentrations were low (less than 0.1 ppm) for  Runs 1 to 8, approximately
1 ppm in Runs 9 and 10,  and greater than  10 ppm  in Runs 11, 12, and 13.  A
malfunctioning heater in  the Decorapozon  unit from runs  11  to 13 created
high ozone levels (greater than 1 ppm).  The temperature  in the Decompozon
unit should have been 140°F  for the unit  to function properly,  whereas the
temperature  for  Runs 11  to  13 was  only about  80°F.    Ozone  destruction
efficiencies were  achieved  greater than 99.99  percent  in Runs  1  to 10.
Significant VOC removal  occurred when  the unit functioned as designed (Runs
1 through 8).

NON-CRITICAL PARAMETERS

      Non-critical  parameters measured  included semi-volatile organics,
PCBs/pestlcides, and total organic carbon (TOG),  Non-critical parameters
for inorganics  included pH,  conductivity,  and  alkalinity.   Temperature,
turbidity,  residual oxidants,  and  electricity consumption were  also
measured.

      No  semivolatiles  or  PCBs/pesticides  were detected  in  either  the
Influent or effluent.  Minimal TOC removal occurred indicating that complete
oxidation of organics to carbon dioxide and water did not occur.

       Iron  and manganese were  present at  low concentrations, and  no
significant metal removal  occurred.  No  changes  in  alkalinity  and
conductivity were observed after the  treatment.   However, the pH increased
by 0.5 to 0.8 units after the treatment.  The increase in pH is attributed
to  the reaction  between hydroxyl radicals  and bicarbonate ions  (the
predominant form  of alkalinity at the ground water pH,  which  is  7.2)  In
which hydroxyl ions are  produced (4).

      Turbidity increased by  1 to 4  units  after the treatment,  which may
be  due to the  insignificant  amount  of  metal  removal  by oxidation  and
precipitation.  The temperature Increased by approximately 5°F after the
treatment and was due mainly  to  the  heat  from  the UV lamps.   Ozone gas
transfer to the ground water was over  95 percent, with 5 percent remaining
in the reactor off-gas before treatment.  After  the reaction, the residual
ozone and hydrogen peroxide concentrations in the effluent were usually less
than 0.1 ppm.  Electrical consumption averaged about  11 kWh/h of operation.
                                    144

-------
                               CONCLUSIONS

      The  treated ground water met  the discharge standards  for  disposal
into a nearby waterway  at the 95 percent confidence level at a hydraulic
retention time of 40 minutes, an influent pH of 7.2 (unadjusted),  an ozone
dose of 110 mg/L, a hydrogen peroxide dose  of 13  mg/L,  and with all 24 UV
lamps operating.

      The  Decompozon  unit destroyed the  reactor  off-gas ozone to  levels
less than 0.1 ppm (OSHA Standard)  with destruction  efficiencies greater than
99.99 percent.

      Removal efficiencies for TCE and total VOCs were as high as 99 percent
and 90 percent,  respectively.  However, maximum  removal  efficiencies  for
1,1-DCA and 1,1,1-TCA were about  65  percent and 85 percent,  respectively.

      Both chemical oxidation and stripping  occurred  in  the removal of 1,1-
DCA and 1,1,1-TCA.  In general, about 12 to 75 percent of the total  removals
for 1,1,1-TCA, and 5 to 44 percent of the total removals  for  1,1-DCA were
due to stripping. In most of the runs,  stripping for  TCE and vinyl  chloride
was less  than 10 percent.    For  other  VOCs, such as  1,1-dichloroethene,
benzene,  acetone, and 1,1,2,2-tetrachloroethane, stripping was negligible.
VOCs present in the  gas  phase within  the reactor at levels of approximately
0.1 to 0.5  ppm were removed  to  below  detection  levels in the  Decompozon
unit.

      Very  low TOG  removal occurred, which  implies that partial oxidation
of organics took place  in the system, but not complete conversion to carbon
dioxide and water.   However,  in  the GC and GC/MS analyses performed  for
VOCs,  semivolatile  organics, and PCBs/pesticides,  no  new compounds  were
discovered.

                                REFERENCES

1.    CH2M Hill.  Preliminary Site Assessment Report  for the Lorentz Barrel
      and Drum Site.  1986.

2.    PRC  Environmental  Management,  Inc.  and Engineering-Science,  Inc.
      Demonstration Plan for the Ultrox International UV/Oxidati.on  Process.
      Prepared for U.S. EPA,  February 1989.

3.    U.S. EPA.  Technology Evaluation Report SITE Program Demonstration of
      the Ultrox International UV Radiation/Oxidation Technology.  In Press.

4.    Hoigne,  J. and  Bader,  H.  Ozonation   of  Water:   Role  of  Hydroxyl
      Radicals as Oxidizing  Intermediates.   Science  190:782, 1975.
                                    145

-------
                                 Treated Off-Gas
                                              Reactor Off-Gas
        CATALYTIC OZONE DECOMPOSER
                          from Wastewater
                          Feed Tank
                                                               Treated
                                                               Effluent
                                                    ULTROX
                                              UV/OXIDATION REACTOR
                                                           Hydrogen Peroxide
                                                           from Feed Tank
                          OZONE
                        GENERATOR
Ait1
    -Compressed Air
   Air
Compressor
   Figure 1.  Isometric View of Ultrox System.

-------
100 -
90 •

80 '


x§ 70
•f 60 "
I 50
(0
DC 40 -
O
§ 30
20 -
10
0..
















f
/
/
t
/
/
/
/
/
/
/
/
/
/
/
/
/































/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/

































/
/
/
/
/
/
/
/
/
/
/
/












"















7
/
/
/
/
/
/
/
/
/
/
/





























~/
/
/
/
/
/
/
/
/
/
/
/
/





























7
/
i/
/
/
/
/
/
/
/
S
/
/
/
/
/
/
/
s























7
/
/
/
/
/
/
/
/
/
/
























t
/
/
f
/
/
/
/
/
t1
/
/
/>
/
/
/
/
/

































7
/
/
/
/
/
i>
/
/
/
/
/
/
i>
/
/
>

































/
/
/
/
/
/
/
/
^
/
/
/
/
/
/
/
/
/
/
/
/


































/
/
/
/
/
/
/
/
/
/
/
/
/
/
s
/
/
































/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
y


































r
/
/
i>
/
/
/
/
/
/
/
/
/
/
if
/
/
/
/
/
/
/
















i i i i i i i i i i i i
                               8
                       Run Number
nice
E3 1,1-DCA
1,1,1-TCA
                         10   11   12   13
Total VOCs
       Figure 2. VOC Removals in Different Runs.

-------
                  TABLE 1.   OPERATING PARAMETERS MATRIX

Retention Ozone Dose HaCJg
Run
No.
1
2
3
4
5
6
7
8
9
10
11
12
13

Influent
pH
7.2
6.2
5.2
7.2
7.2
7.2
7.2
7.2
7.2
7.2
7.2
7.2
7.2
TABLE 2. EXTENT
Time
(min)
40
40
40
60
20
40
40
40
40
40
40
40
40
(mg/L of
Water)
75
75
75
75
75
110
38
110
110
110
110
110
110
OF VOC STRIPPING IN THE
UV Lamps
Dose (24 {§ 65 W
(mg/L)
25
25
25
25
25
25
25
38
13
13
13
13
13
each)
All On
All On
All On
All On
All On
All On
All On
All On
All On
1-12 On
13-24 On
All On
All On
ULTROX SYSTEM



Run
No.
1
2
3
4
5
6
7
8
9
10
11
12
13


lAir Flow Rate "I
[water Flow Rate]
2.1
2.3
2.1
2.0
2.1
4.5
1.0
4.5
4.5
4.3
4.6
4.4
4.3
Percent

1,1 -DCA
0.0043f 0
7.4
9.1
9.9
7.4
17
16
4.9
23
16
27
44
34
37
Stripping

TCE
.0091f
2.0
3.4
2.7
3.0
3.5
1.2
1.2
7.5
6.6
9.4
24
7.0
26
Contribution

1,1,1-TCA
0.014f
43
34
31
29
29
65
12
85
58
73
>99
76
75
for
Vinyl
Chloride
0.082f
0.013
0.95
0.013
0.010
1.7
0.072
3.1
1.2
0.040
1.1
13
8.9
1.8
f   Henry's law constant of the VOC, atm-m3/mol.




                                  148

-------
                       SITE DEMONSTRATION OF BIOLOGICAL
                   TREATMENT OF GROUNDWATER BY BIOTROL,  INC.
                AT A WOOD PRESERVING  SITE  IN  NEW BRIGHTON, MN

                                      by

                  Mary K.  Stlnson,  ORD/RREL-USEPA,  Edison, NJ
                       William Hahn,  SAIC,  Paramus, NJ
                    Herbert S.  Skovronek, SAIC,  Paramus, NJ
                                   ABSTRACT
      A wood preserving site In New  Brighton,  MN on EPA's National Priorities
List was selected for evaluation of a groundwater treatment for pentachlorophenol
with a fixed-film biological system. The system employs Indigenous microorganisms
but Is also amended with a specific pentachlorophenol-degrading bacterium. The
mobile,  pilot-scale unit used for  the demonstration houses a 540-gallon, three-
stage bioreactor filled with structured PVC packing for blomass support. After
an Initial acclimation period, groundwater from a well  on the site was fed to
the system at  1,  3,  and 5 gpm with  no pretreatment  other than pH adjustment,
nutrient addition, and temperature control. Each flow regime was maintained for
about two weeks while samples were collected for extensive analyses.

      At 5 gpm, the  system was capable of achieving about 96%  removal of the
pentachlorophenol  in  the  incoming   groundwater   and  producing  effluent
pentachlorophenol  concentrations  of about  1  ppm,  which met the  local  POTW
requirement for discharge. At the lower flows (1 and 3 gpm) removal was higher
(about 99%) and effluent pentachlorophenol concentrations were  well below 0.5
ppm.

      Operating  costs,   including power  (pumping  of  liquids  and  heating),
nutrients and caustic, and operator labor, are  reported. This system appears to
be a  compact  and cost-effective  treatment for pentachlorophenol-contaminated
wastewaters.  Pre- and post-treatment such as  for oil or solids removal, may be
required on a site- and wastewater-speciflc basis.

      The results reported in this paper are  preliminary and a full report is
in preparation.  This paper has  been  reviewed  in   accordance  with the  U.S.
Environmental Protection Agency's peer and administrative review policies and
approved for presentation and publication.

                                     149

-------
                                 INTRODUCTION
      Soil and  groundwater  contamination  by chemicals commonly resulting from
wood preserving operations has frequently been found at Superfund sites on the
National Priorities List. Under the Superfund Amendments and Reauthorization Act
of  1986 (SARA),  the U.S.  Environmental  Protection  Agency  was  empowered  to
initiate a S_uperfund Innovative Technology Evaluation (SITE) program to develop,
demonstrate, and evaluate new and innovative technologies  that could be used at
Superfund sites, A method for the destruction or removal of hazardous
chemical species such as pentachlorophenol  (PGP) and creosote-derived polynuclear
aromatic hydrocarbons  (PAHs)  found at wood preserving  sites was  deemed to be
suitable for investigation under this program.

      BioTrol, Inc.  of Chaska, MN offered a biochemical destruction technology
and encouraging claims from earlier,  small-scale  studies  that  indicated  that
efficient removal  of such pollutants from contaminated soil and groundwater could
be  achieved.  While biotreatment  has  a  long history  as  a  cost-effective
destructive method  for  organic  chemicals  in both  industrial and  municipal
wastewaters, it was uncertain whether  such technology would be  effective  at
Superfund sites for  the recalcitrant  chemicals that  might be  encountered as a
result of long term wood preserving operations, specifically pentachlorophenol
and polynuclear aromatic hydrocarbons.

      Subsequently, the BioTrol, Inc. Aqueous Treatment System (ATS) was selected
for investigation under the SITE program.  After considering alternate sites, a
facility recently added to the National  Priorities List  was chosen for a pilot-
scale evaluation of  the technology. The selected  site,  in New  Brighton,  MN, a
suburb  of  Minneapolis,  has  been  used  for  wood  treatment  with  various
preservatives,  including creosote,  pentachlorophenol,  and  chromated  copper
arsenate since the 1920s. Tests at  the  site as part  of a RI/FS indicated that
both  the   soil   and  the  underlying  groundwater   were contaminated  with
pentachlorophenol  and  polynuclear  aromatic hydrocarbons, even  though  these
chemicals were no longer being used in  wood treatment.  The owner  and operator
of the  site, the  MacGillis  and  Gibbs Company, agreed  to  host  the pilot-scale
testing of the BioTrol system.

                              PROCESS DESCRIPTION
      The BioTrol Aqueous Treatment System (ATS) shown in Figure 1 consists of
a. conditioning or temper tank, a heater and heat exchanger, and a  three-stage
fixed-film biological  reactor.   Incoming wastewater  is  first brought  to  the
conditioning tank where the pH is adjusted (if necessary)  to just above 7.0 with
caustic and inorganic nitrogen and phosphorus nutrients are added. After passing
through  the  in-line  heater  and heat  exchanger to  assure  a more  constant
temperature in the vicinity of 70 F,  the wastewater  is introduced to the base
of the  first  of the three bioreaction  chambers (Figure  2).    Each  chamber is
filled  with  an inert  support for  bacterial  growth;  in  the  study  corrugated
polyvinyl chloride sheets were the support medium used (Figure 3) . The influent
is passed up  through  each chamber while air  is injected at  the base  of each
chamber through a sparger tube system, as shown in Figure 2.


                                      150

-------
INFLUENT

PUMP
HEAT EXCHANGER
     EFFLUENT PUMP
                                      TABLE
     NUTRIENTS
             •BLOWER    ^-REACTOR  ^CONTROL

                                         PA N F i ^
           -TEMPER TANK              HANtL^
 Figure 1. BioTrol, Inc. Mobile Aqueous Treatment System (ATS) System.

-------
                                       Nutrients,
                                        Caustic
        liquid Sample
         (W*flW«ter)
                                                                                                               To POTW
                                                                                                                 or
                                                                                                               MacGilUs
                                                                                                                 and
                                                                                                                Gtobs
                                                                                                            Liquid Sampfo
                                                                                                              (Effluent)
Not*: Circtod nuniMn U wnflt point* rafer to Utx*ng
    system
                  Figure 2. Aqueous Treatment  System (ATS)  with  Sampling Points Shown.

-------
                                   FIGURE 3
                      CORRUGATED POLYVINYL  CHLORIDE MEDIA
      The system is acclimated by introducing an indigenous bacterial population
taken from the  local soil. After allowing about one week for acclimation of this
growth  to the  wastewater,   the  system  is  "seeded"  with  an  inoculum  of  a
flavobacterium   specific   to    the    target  contaminant,   in   this   case
pentachlorophenol.  The wastewater containing the  contaminant  is then recycled
through the system to allow the bacterial population to readjust.  When the system
is fully adapted to the wastewater,  once-through processing is ready to begin.

      At the  MacGillis and Gibbs  site  it was  determined  that the: quality of the
groundwater  did not  warrant any  pretreatment,   even  though  it contained  a
significant level of oil  (about 50-60 ppm).  While pretreatment such as oil/water
separation or solids  removal may be needed in other cases,  such decisions must
be site and wastewater specific.  Similarly,  post-treatment decisions also depend
on the  specific  site.  At this  facility,  a decision was made  to install  a bag
filter  to  collect  the  small amount of sloughed biomass that  was anticipated,
primarily so  that pollutants in the sludge  could be measured as part of the EPA
investigation.

                              SITE TESTING PROGRAM
      Working  in collaboration  with the  developer  of  the  process,  it  was
determined that operation of the system at three increasing flow rates, 1,  3,

                                     153

-------
and 5 gpm,  corresponding to residence times of 9, 3 and 1.8 hours, respectively,
each for two weeks, would allow the effectiveness of the process to be determined
at low contaminant loadings and at the design level. In fact, while the screening
data reported  in 1984 as part of  the RI/FS had suggested high concentrations
(-100-200 ppm)  of pentachlorophenol  might be  present in the groundwater, when
two wells were drilled in preparation for the project, a maximum of about 45 ppm
pentachlorophenol and only low levels of polynuclear aromatic hydrocarbons (<1
ppm total PAHs) were found to be present.

      The groundwater  obtained from the selected well, the  influent  to,  the
effluent from, and the two  intermediate  stages of  the bioreactor were monitored
for pentachlorophenol, other semivolatile organics, chloride,  and TOG. Chloride
and TOG were monitored to provide supporting evidence for the vendor's claim that
pentachlorophenol  removal occurred by mineralization to water, carbon dioxide
and salt by the following equation.

         OH
         I
        ^CV
   Cl-C
   Cl-C,
'C-C1
 I
.C-C1
                   +  excess 02  	>  6 C02 + 0.5 H20 + 5 Cl-
         I
         Cl

      Other  parameters  also monitored  to provide a  complete history  of the
groundwater as it passed through the system included total and volatile suspended
solids, oil  and  grease, nitrogen  and phosphorus,  volatile organics,  and heavy
metals. Because  there  is  always concern  when  treating wastewaters containing
chlorinated aromatics, testing was also done for chlorodioxins and furans.

      Because this investigation was  part of the  SITE program and careful and
complete analytical history (and safety) was desirable, carbon adsorption units
were installed on both the aqueous discharge and on the air leaving the covered
reactor chamber. Samplings and analyses were carried out before and after these
units to determine whether significant quantities  of the contaminants were lost
by any route other than biodegradation.

      Finally, static bioassays were  carried out  on  the incoming groundwater,
the influent to the reactor, and the effluent to learn whether the groundwater
was toxic to aquatic species and whether treatment removed the chemical source
of any toxicity.

                                    RESULTS
       From  comparison  of   the   pentachlorophenol   concentrations   for  the
groundwater as removed from  the well  and  the  effluent from the bioreactor,  it
is clear that the BioTrol  system  is capable of achieving about 96% removal  of
pentachlorophenol at the design flow rate,  5 gpm. And, at that flowrate,  final
affluent concentrations,  before carbon polishing, are  approximately 1 ppm. Table 1
summarizes the pentachlorophenol removals at the three different flow rates.

                                     154

-------
              TABLE 1. AVERAGE PENTACHLOROPHENOL REMOVAL BY
                   THE BIOTROL AQUEOUS TREATMENT SYSTEM
Flow
Rate
(gpm)
1
3
5
Ground-
water
(ppm)
42.0*
34.5*
27.5*
Effluent

(ppm)
0.13
0.36
0.99
Percent
(%:
Average
99.8 **
98.7 **
97.6 **
Removal
)
Range
87.4-99.9+
95.8-99.8
79.3-99.4
            * decrease with time may reflect drawdown of aquifer
            ** based on average of daily effluents

     However, it must be noted that as  the analytical results were obtained, it
became apparent  that an unexpected  dilution  phenomenon was occurring  in the
influent chamber where the composite influent samples were taken. The effect was
a  significant   reduction   in  the  apparent   "influent"  concentrations  for
pentachlorophenol (and other parameters) - and, presumably, in the values at the
two intermediate sampling points as  well. Where  these  values should have been
essentially the same  as the values for the groundwater, it was observed that they
were  considerably  lower.  Grab  samples  obtained  by the  vendor between  the
conditioning tank and  the bioreactor and  analyzed for  pentachlorophenol using
another method also confirmed the discrepancy. (In this alternate method, high
pressure liquid chromatography  [HPLC],  the aqueous sample is injected directly
onto a column at ambient temperature and the levels of pentachlorophenol measured
with  a UV detector  at  254 run  and 220 nm.  Although the method is not "EPA-
approved" and was not subjected to  the  extensive  quality assurance used for the
GC/MS method, an abbreviated  evaluation has demonstrated that  the  results are
reliable and comparable  to  those obtained by GC/MS.) It  is believed  that the
differences in concentrations, which were particularly significant at the lower
flow rates, are the result of backmixing from each of the reaction chambers into
the preceding mixing chambers.  Consequently,  the  results  being  presented are
based primarily on  the incoming groundwater as  it was analyzed at the well head
and the  final  effluent  from the bioreactor,   using  EPA  Method  3510/8270,  for
which the Method Detection Limit for pentachlorophenol  is 50 ug/L.

      At the lower flow rates studied,  1 and 3 gpm, pentachlorophenol  removals
(based on the change  from the  groundwater to the  effluent) increase to 99+% and
final pentachlorophenol concentrations  of 0.1  ppm and even less are achievable.
These results are summarized in Table 1.


      The changes in  chloride  and TOC results  (weekly) parallel the decrease in
pentachlorophenol at  all  flows  (Table  2);  however,  they  are not sufficiently
precise  to  provide   more  than supportive  evidence  for  mineralization  of
pentachlorophenol to sodium chloride, water,  and carbon dioxide.
                                     155

-------
    TABLE 2. COMPARISON OF AVERAGE CHLORIDE, TOG, AND PCP RESULTS
Flow
Rate
1
3
5
r 	
I
PCP
-41.9
-34.1
-26.5
	 fO,
1 Cl(f)
1 (P
i . . 	
| +44.2
| +40.5
j +22.0
ange
Cl
pm)
+27
+22
+17
(c
(c)
.9
.7
.6
j_1 *.as 	
| TOC(f)
. . 	 	
1 	
| -24
| -32
| -21
	 i . .i
TOC(c) I
1
i
-11.3 |
- 9.2 |
- 7.0 |
            (f) — found; (c) calculated


     As part of the effort to confirm that pentachlorophenol was being removed
by  biochemical mineralization and  not  by adsorption  on the biosolids  or by
stripping due to the air in the bioreactors, both solids and air emissions were
also monitored.  Although the sludge  trapped in the  bag filter was  found to
contain pentachlorophenol (34 and 170 ppm found in two samples), the amount of
sludge was so  small  that adsorption of  pentachlorophenol on the biosolids and
removal with  the  suspended solids does not represent  a  significant alternate
removal  mechanism.  Thus,   even  if  all   the  suspended  solids  (effluent
groundwater) produced by the system during the twelve days of the 1 gpm run were
trapped in the filter,  this would amount  to only about 7 Ibs of sludge.   Even
with a pentachlorophenol content as high as 170 ppm  (which  was  measured in a
later sample), this would only account for about 0.0012 Ibs of PCP or about 0.02%
of  the  total  pentachlorophenol   input   of  about   6.05  Ibs.     Similarly,
pentachlorophenol was not present above the detection limit  in any of the air
samples obtained over the reactor chamber,  using a modified Method 5 collection
system with an XAD resin trap and an analytical method with a detection limit
of  1.7 ug/cubic  meter or 0.2 ppb.  Therefore,  it does appear that biological
degradation  is,  by far, the primary means of eliminating the pentachlorophenol
from the groundwater.

     Concentrations of the  various polynuclear aromatic hydrocarbons as part of
the semivolatile fraction were below detection limits  in the samples of incoming
groundwater used  in the demonstration program. Two  analyses during  the pre-
demonstration testing indicated total PAHs of 145 and 295 ppb; Consequently, it
is not possible to draw any conclusions as to removal efficiency or mechanism.
However,  several PAHs, including naphthalene and methyl naphthalene at maximum
levels of 34.6 ppb and 47.9 ppb,  respectively, and others at considerably lower
levels,  were found during the modified Method 5 testing of  the air emissions
from the reactor. This  suggests  that  some air stripping  of these constituents
may be occurring.

     Small amounts of various  chlorinated dioxins  were found in the  effluent
(<340 ng/L, using method SW8280) and, particularly, the sloughed biomass sludge,
where one sample did exhibit 1900 ng/g of OCDD isomer. This value is currently
being re-examined. With  one exception, an effluent sample found to contain 62
ng/L, the 2,3,7,8-tetrachlorodioxin of primary concern was not detected in any
of the influent, effluent, or sludge  samples using high  resolution GC coupled
with low resolution MS.

     The incoming groundwater was found to contain low concentrations of several
                                     156

-------
of the heavy metals, including nickel (<91 ug/L), zinc (<32 ug/L), copper (<25
ug/L),  lead (<11  ug/L),  and arsenic  (<6.5 ug/L)  from the  chromated copper
arsenate wood preservative currently used in wood treatment at the site. With
the exception of  one  sample  which is believed to be  an  anomaly,  there was no
change in the concentrations of the metals across the system.

    Acute biomonitoring with  fresh water minnows (96 hr static test) and daphnia
magna  (48 hr  static  test)  demonstrated that  the toxicity in  the  incoming
groundwater or the influent  was  essentially totally removed by the treatment.
LCSO's increased from an estimated low of 0.2% (groundwater/control water) for
the groundwater  to more than 100%  (as calculated from results) in the treated
effluent.

                                     COSTS
      Preliminary cost estimates were  carried out by the vendor for operation
of the pilot plant at MacGillis and Gibbs excluding the ancillary equipment such
as carbon units and  bag  filter but including cost for nutrients,  electricity,
heat, labor and caustic. In addition, costs were extrapolated by the vendor to
a full scale system capable of treating 30  gpm of  a similarly contaminated (-40
ppm pentaehlorophenol) groundwater based on the demonstration study and other
information at their disposal. On  these bases, operating cost at the 5 gpm and
the 30 gpm rate would be $4.24/1000 gallons  and $2,62/1000 gallons, respectively
(Table 3). As shown in the table, certain costs do not  increase at the expected
rate. For example, unit nutrient cost would decrease because of bulk purchase;
electricity cost/gallon treated decreases because it is assumed chat with deeper
bioreactor beds in the  30  gpm unit (8 ft instead of 4 ft)  the energy for the
compressor supplying the  air would be used more efficiently;  operator labor cost
also are not expected to increase  in direct proportion  to the size of the unit.


            TABLE 3.  OPERATING COST OF TREATMENT ($/1000 gal)

                 Cost Item      at 5 gpm     at 30 gpm
nutrients
electricity
heat
labor
caustic
0.042
0.416
1.46
2.08
0.24
0.017
0.216
1.46
0.69
0.24
                  TOTAL         4.24          2.62


     These  costs do  not  include  leasing  or amortization  of  the    capital
equipment, which are approximately $3,200/month (5 gpm mobile),  $30,000 (5 gpm
skid mounted) and $80,000 (30 gpm skid mounted),  respectively.

     Clearly labor and heat (electrical) requirements are the major factors to
consider when  treating  waters at a specific  site.  And,  of course,  any site-
specific pre-  or post-treatment requirements,  such as  oil/water  separation,
solids removal, polishing, air emissions control, etc., would have to be factored
Into the cost calculation for that site.

                                    15?

-------
                                 CONCLUSIONS


      On the basis  of the pilot plant study carried out  at the MacGillls and
Sibhs site  in  Minnesota,  the BioTrol process would be  successful in treating
groundwater or other  pentachlorophenol-contaminated wastewaters (at  -40 ppm
pentachlorophenol) to  levels suitable for discharge to  a  POTW or reuse within
a plant.  One  unforeseen  benefit  of  the  treatment was  that biotoxicity in the
incoming groundwater was eliminated by the treatment.

      Contaminated waters  of different  concentrations can  be accommodated by
increasing or decreasing the throughput rate, recycling a portion of the stream
or by sizing the system differently.  Site-specific factors such as groundwater
temperature, ambient temperature, extent of contamination with oil and/or solids,
etc.» can all play a role in the cost-effectiveness of overall treatment.

     Although  a  secondary  objective  of  the  study  was  to  evaluate  the
effectiveness of the system for removal of polyaromatic nuclear hydrocarbons that
might be present at the site as a result of the use of creosote, the levels of
these constituents in the groundwater used for the study were too low to reach
any conclusions as to removal.
                                     158

-------
                     FACTORS AFFECTING THE RELIABILITY
                         OF OPERATIONS OF THE EPA
                        MOBILE  INCINERATION SYSTEM


             James P. Stumbar, Robert H.  Sawyer, Gopal D.Gupta
                     Foster Wheeler Envlresponse, Inc.
                             Edison, NJ  08837


                    Joyce M. Perdek, Frank J.  Freestone
                      Releases  Control Branch,  USEPA
                             Edison, NJ  08837


                                     ABSTRACT -

    The EPA Mobile Incineration System has been operated, from October, 1985
until April, 1989 at the Denney Farm Superfund Site at McDowell, Missouri, to
demonstrate the high-temperature incineration of hazardous wastes.  These
consisted of a wide variety of solids, liquids and sludges contaminated with
2,3,7,8-tetrachlorodibenzo-p-dioxin and other hazardous compounds.  The system
started with an availability of only 40% but achieved an availability of 80%
during the site closure period.  Based upon the operating experience, the
availability of the system was determined by:  the interaction of various parts
of the system with feed or ash characteristics, basic mechanical design
factors, crew shakedown periods at the start of each operating period, and
interruptions to perform scientific and regulatory tests.  This paper discusses
the factors that determined the availability of the unit,  the corrective
actions taken to deal with adverse feed characteristics and design
deficiencies, and the results of the corrective actions.  The information
contained in this paper is directly applicable to field use of mobile and
transportable incinerators at Superfund and other industrial cleanup sites.
                                    DISCLAIMER


The information in this document has been funded by the U.S. Environmental
Protection Agency under Contract No. 68-03-3255 under the sponsorship of the
Office of Research and Development.  The document has been reviewed in
accordance with U.S. Environmental Protection Agency policy and approved for
publication.  Mention of trade names or commercial  products does not consti-
tute endorsement or recommendation for use.
                                      1S9

-------
                        FACTORS AFFECTING THE RELIABILITY
                             OF OPERATIONS OF THE EPA
                            MOBILE INCINERATION SYSTEM
INTRODUCTION
    Under the sponsorship of the Office of Research and Development of the U.S.
Environmental Protection Agency (EPA), the EPA Mobile Incineration System (MIS)
was designed and constructed to demonstrate high-temperature incineration of
hazardous wastes [1,2,3,4,5,6].  The full scale demonstration of the unit
occurred at the Denney Farm Superfund Site at McDowell, Missouri, from October,
1985 until April, 1989.

    This paper presents an analysis of the factors affecting the reliability of
the operations of the EPA Mobile Incineration System (MIS) during the above
period.  The reliability of the system is measured in terms of availability.
Availability is defined as the actual operating time when feed was entering the
system divided by the total time available for processing feed multiplied by
100.

    The availability of the MIS was determined by several factors consisting
of:  the interaction of the various parts of the system with feed or ash
characteristics, basic mechanical design factors, crew and equipment shakedown
periods at the start of each operating period, and interruptions to perform
scientific and regulatory tests.  As the incineration of the contaminated
materials progressed, adverse feed characteristics and deficiencies in the
design and operation of the MIS were identified.  Corrective actions were taken
to deal with adverse feed characteristics and to remedy design deficiencies.
These actions produced Increases in the capacity and the reliability of the MIS
and a significant reduction in operating costs.  Figure 1 shows that the
availability was increased from 40% to 80% between 1985 and 1989.

    Correction of a major design deficiency or corrective actions to remedy
problems due to adverse feed characteristics often did not show the anticipated
Increase in avail ability.  Although a problem had been corrected by the changes
to the unit, other problems would surface, and only a small increase in
availability was obtained.  This occurred because during the long downtimes
caused by the original problem, maintenance was performed on the entire
system.  This general maintenance sometimes hid these other problem areas.
Specific examples are discussed below.


Description of the MIS

    The MIS, shown in Figure 2, consists of a refractory-lined rotary kiln,
followed by a cyclone, a secondary combustion chamber (SCC), and an air
pollution control system.  The air pollution control system consists of a
quench system, followed by a wet electrostatic precipitator (WEP), a packed bed
                                     160

-------
       100
/~s
K
v^
u
2
h
LJ
_l
I
^
5
H
O
Z
O
UJ
u
u.
        10
         0
              ,	,—!	,	,	j—,	,—^	,—,	j	,	,	!	,	,	j—,	,—1	,   r—|	,	,	j-

         Jul-85   Oct-85  Jan-86  Jun-87  Sep-87  Feb-88   May-88 Aug-88  Nov-88   Feb-89

                                             MONTH
                                      Figure 1.  MIS availability.

-------
                       COMBUSTION GAS ANALYSIS
                         (tHC,HO,,O1,CO,CO>l
   CYCLONE
                                                                STACK GAS ANALYSIS
                                                                 (THC, NO,, O,, CO)
BURNERS
                  SECONDARY
                  COMBUSTION
                   CHAMBER
WEP     MX SCRUBBER
         1 DEUISTER
 BURNER
(RT. SIDE)
              1.0. FAN   DIESEL
                      1.0. FAN
                      DRIVE
    Figure  2   EPA  Mobile  Incineration  System

-------
scrubber and a mist eliminator [5].  Auxiliary systems consist of the feed
handling, ash handling, and air pollution monitoring systems.  The feed
handling system contains a conveyor system, shredder, weigh scale, ram feeder,
and HEPA/Carbon filtration system for fugitive dust and vapor control.  The ash
handling system has alternating gates to minimize air in-leakage into the back
of the kiln.  The ash is discharged into drums.  The air pollution monitoring
system contains instruments to continuously monitor oxygen and carbon dioxide
and to periodically monitor carbon monoxide, unburned hydrocarbons, and
nitrogen oxides.


AVAILABILITY DURING 1985-1986

    During 1985 the overall availability averaged 50% while processing
materials from the first five sites as shown in Table 1.  The chief cause of
downtime or unavailability was the buildup of solids in the SCC,  Approximately
90% of the downtime was attributable to cleaning out the particulate carryover
from the kiln to the SCC.  Other sources of downtime were extensive maintenance
on the Cleanable High Efficiency Air Filter (CHEAP) and feed system jams.  The
addition of the cyclone and the replacement of the CHEAP with the WEP were the
major actions taken to address these concerns.  The addition of the Linde "A"
Burner Systeur helped to reduce the amount of carryover that the cyclone had
to handle by decreasing the kiln exit gas velocities.  Table 2 shows that these
changes were effective in solving these problems.  Based upon the lost
processing time for cleaning the SCC, the availability was predicted to
increase to 80% if the above design changes were implemented.


AVAILABILITY DURING 1987

    During the 14-week operation of the modified system, an overall system
availability of 53% was achieved.  This consisted of a 30.5% availability
during the trial burn and a 54.8% availability during other activities.

    Although slagging or plugging of the SCC was almost eliminated, the
anticipated improvement in availability did not materialize.  New sources of
downtime or unavailability were encountered.  The new causes of downtime were
the slagging of the kiln that accounted for 10.6% and delays caused by testing
and preparation for testing that caused 8.7%.  Other important factors were
mechanical failures accounting for 9.9% and human error accounting for 7.4%.
The causes of downtime are presented in Tables 2 and 3.  The reasons for these
periods of unavailability are discussed below.

Kiln Slagging

    Kiln slagging became a factor in 1987 due to the processing of large
quantities of trash that constituted about 40% of the feed to the HIS.  The
trash present in the feed caused slagging of the kiln walls with a lava-type
hard material that eventually built a dam restricting the flow of material
through the kiln.  This dam was located at about eight feet from the front or
at the midpoint of the kiln.  The materials conducive to the slagging were


                                      163

-------
                                             TABLE 1.   HIS OPERATIONS  SUMMARY
Feed Source or Activity
1985-1966
Heosho and The Farms
Heosho and The Farms
1985-1966 Totals
1987
Heosho and The Farms
RCRA/TSCA Trial Burn4
1987 Totals
1988-1989
Baldwin Park5
Baldwin Park Trash 6
Brominated Sludge Tests
Verona
Maintenance
Brcrainated Sludge
Closure
1988-1989 Totals
1988-1989 Totals w/out
Date of Activity
7/23/85-8/27/85
10/26/85-2/03/86
6/12-9/18
7/28-8/3 8/31-9/1
2/22/88-4/27/88
8/02/88-8/12/88
4/28/88-5/01/88
5/02/88-8/01/88
8/12/88-8/31/88
8/31/88-1/23/89
1/24/89-4/15/89
sludge9
Pounds
Ib Solids
436,000
1.495.000
1,931,000
1,095,000
162.000
1,257,000
911,000
136,000
49,000
3,097,000
0
2,161,000
2.989.000
9,343,000
7,182,000
Processed
Ib Liquids
60,300
107.300
167,600
31,300
7.600
38,900
22,200
3,800
0
3,000
0
1,400
200
30,600
29,200
Avail.1
(X)
38.1
54.0
50.0
54.8
30.5
53.1
39.1
58.9
55.1
59.8
0.0
60.8
80.0
60.78
60.68
Hours
Out
591
1313
1904
995
132
1127
966
104
35
891
0
1331
382
3709
2378
Hours
Online
363
1540
1903
1217
58
1275
620
149
43
1324
0
2061
1530
5727
3666
Total
Hours
954
2853
3801
2212
190
2402
1586
253
78
2215
0
3392
1912
9436
6044
Totals for 1985-1989
                       GRAND TOTAL
12,531,000      237,100
       12,768,000
  X Availability "(Time Processing Solid Feed/Total Time)  x 100.
  The Farms are Denney, Eruin, Rusha and Tally.  Only small quantities  of  trash were processed.
  Included is approximately 475,000 Ib of trash.
  The low availability was due to preparation time and a failed kiln  bearing.
  The low availability was due to problems encountered during the start-up in  February and March, 1988.
  Host of the material was trash consisting of wooden pallets and steel drums  with some soils.
  Due to Decontamination of the feed system the records stop on April 16.   Incineration
  activities continued until April 22, 1989, chiefly on residues from cleaning the feed system.
  Availability calculation excludes the August, 1988 maintenance period.
                                                  164

-------
            TABLE 2.  DESCRIPTION OF PROBLEMS ENCOUNTERED  DURING THE OPERATION
             OF THE MOBILE INCINERATION SYSTEM, CORRECTIVE ACTIONS TAKEN,
                       AND THE RESULTS OF  THE CORRECTIVE ACTIONS
Year Problem Cause Corrective Action Result X Availability ^Availability
Lost Before Lost After
1985/1986


1987



Secondary CoAuation | Paniculate Carryover From
Ctiawber (SCC) Pluninoj Rotary Kiln
Cleanable High iff .
Air Filter < CHEAP)
Feed System
Ram Jans
Feed System
Feed Systen
Ki In Slagging
.
Process Water
Contamination
Mechanical Problems
Buildup on Back of Ram
Blockages at Doctor Blade
Bridging of Wood 8 Steel
Drun Rings
Trash Feed
Hunan Error
Cyclone & Oxygen
Burner - 198?
Wet Electrostatic
Precipitator (UEP)
S CPI Separator
- 1987
Hauptman*
Conveyor - 1987
Roller Dam - 198S
Television & Manual
Size Reduction
- 1988
Reduce Kiln Temp.
Blend Trash with
Soil - 1987
Change Operating
Procedures
Plugging of SCC Solved
Reduced Maintenance
Reduced Frequency
Easier to Keep Clear
Reduced Frequency
and Quicker Response
Reduced Frequency
No Recurrence
44
3.5 '.
2.6 1.
Affected tt
Affected tt
10.6
7.4
0.6
1.5
1.3
roughput
roughput
NA2
0
(continued)
est. = estimated
™ UA — •«*.* AMM.I wjhBW! ^ TktfK iutk MB*A *«n*«l»! mfm *&*!>** t«*«^l A.*** in *minv*Zf*tr* fttm C¥ * w«au*1 1 •§%! I 5 *w LJO«< An**ra m*+*Mft t*t**f ^l%tt aeK «a^4M3 L1AI*A
repaired.

-------
                                                           TABLE  2.   (Continued)
Year
1?88
1987
1988
1987/1988
im




1935-1983
Problem
tci In Slagging
SCC Exit Venturi
SCC Exit Venturi
UEP Rod Problems
Process Water Pipe
Corrosion

Stack Monitoring
Rotary Ki In to sec
Duct Fouling
Compressor Failures
Miscellaneous Failures
Cause
Air Leakage Past Ash Gates
Acid Gas Corrosion (HCl)
Acid Gas Corrosion (HCl)
Frequent Shorting
pH swings

CO meter. New Crew
Brominated Sludge Ash
Loss of Fuel due to
Insufficient Head
year and Tear on unit
Corrective Action
Redesign Ash Gates
Use HaynesR
Alloy 525
Use Cos table Ref .
Stainless Steel
Brace internals
Redesign pH control
and caustic
injection system
Hen CO Monitor £
Training
Install knockout
ports
Use recirculating
puep
Ascertain & stock
critical spare
parts
Result 3
Greatly Reduced Frequency
Three month Life
No problems after 9 *o.
Ho rod failures after 7.5
no.
Improved pH control
Insufficient time to
evaluate corrosion
Good performance over
13 months
Allowed Hot Cleanouts
Prevented loss of diesel
Reduced waiting time
for repair ports
(Availability 3
Lost Before
16.8
6.3
5.8
4.6
3.0 1

9.53
15.
1.6
5-10
(Availability
Lost After
5.0
5.3
0
4
4

1.4
9.4
0
2-3
Unavailability expressed as time lost during the Baldwin Park incineration activity.

-------
       TABLE 3.  CAUSES OF SYSTEM UNAVAILABILITY IN 1987
  Cause                                    Unavailability
                                                  (*)
Kiln Slagging                                  10.6
Mechanical                                      9.9
Testing Delays                                  8.7
Human Error                                     7.4
WEP                                             4.6
Feed Jamming                                    3.5
Electrical                                      2.3
                            Total              47.0
                               18?

-------
plastic materials used to contain the contaminated dirt during transport to the
site and the discarded clothing and material used during cleanup, such as
rubber gloves, plastic suit material, boots, etc.

    During  1987, the chief methods of minimizing slagging of the kiln were to
operate the kiln at the lowest feasible temperature, about 1450°F or 50°F
above the low temperature feed cutoff setting, and to blend as much
contaminated soil as possible with the trash.

    Prior to the modifications only small quantities of trash had been fed
because of the difficulties encountered in processing it.  Processing
difficulties included inability to control kiln temperature and excess oxygen,
and feed system jams [7,8].  The modifications to the feed system and the
addition of the oxygen burner system permitted the feeding of trash but
slagging of the kiln then became a major reliability factor.  Slagging is
caused by the presence of sodium and potassium in the ash.  The mechanism of
slagging is discussed further in References 9 and 10.

    Trash content of the feed also was a major causative factor for the 3.5%
unavailability due to feed jamming.  The trash often jammed at the doctor blade
or restricting dam that was originally used to level the granular material on
the conveyor belt.  The doctor blade worked quite well for granular material
but it created large blockages when materials such as shredded plastic,
clothing, trash, metal, or mud were fed.  As shown in Table 2, a roller was
installed to replace the doctor blade and this reduced the number of jamming
incidents.

Testing Delays

    Delays related to testing accounted for 8.7 % unavailability [11].  These
delays were associated with times required to clean out the system (kiln,
conveyor, water system flush, etc.); to prepare for tests; to hold the system
on hot standby between tests; and to perform special calibrations.  About 25%
of the unavailability related to testing represents the downtime during the
Demister* installation.  The installation of this mist eliminator was
required to prevent high particulate emissions when the unit was processing
feeds containing percent levels of organic chlorides.  This situation was
encountered during the 1987 trial burn [4],

SCC Venturi Failures

    The failure of the InconelR 625 venturi at the exit of the SCC caused
6.3% downtime as shown in Table 2.  This accounted for most of the mechanical
sources of downtime.  After over 3000 hr of operation, some of which was with
highly chlorinated compounds, the exit venturi had completely disintegrated,
and its support plate showed severe stress corrosion cracking.  Based upon the
presence of chlorine in conjunction with the high temperature (2200*F), the
material  for the venturi and diffuser was changed to HaynesR alloy 556.  The
change to this material failed to extend the life of these parts.  A venturi,
fabricated from castable refractory, was installed during the summer of 1988.
This refractory venturi was in excellent shape after 9 months with over 5000
hours of operations when the MIS activity at Denney Farm was completed.

                                      181

-------
Human Error

    Human error caused 7.4% of the unavailability.  This consisted of the time
to decontaminate the process water system when it became contaminated with
furans.  System downtime caused by contamination/decontamination of equipment
is an added cause of unavailability in hazardous waste operations.  This source
of downtime is not normally encountered in equipment processing nonhazardous
materials.

    The furans release occurred in the following manner.  The kiln rotation was
stopped during Test 4 of the trial burn due to the kiln bearing failure.  After
the kiln rotation was stopped, the operators manually shut down the entire
system.  This action stopped the kiln burners, the SCC burners, and the ID
fan.  The ID fan was quickly restarted because of the puffing from the kiln but
the SCC burners were not restarted.  Failure to restart the SCC burners was a
critical error because this created a rapid drop in the SCC temperature.  This
temperature drop caused generation of furans due to incomplete combustion of
PCBs which continued to volatilize from the solid waste remaining in the kiln.
Consequently, these furans contaminated the process water system as shown by
their presence in both the process water sample taken prior to Test 5 and the
water samples taken during Test 5.  Some of these furans then were released
from the stack during Test 5.

    The contamination of the process water system was not discovered until
after Test 5 had been completed the following day.  At this point, further
trial burn testing was postponed and normal operations were continued until it
was pointed out that they should be stopped in order to decontaminate the
process water system.

    The primary lesson learned from this experience is that should the kiln
rotation be stopped with contaminated solid materials in the kiln, the SCC
burners must continue to operate.  The operators were instructed to keep the
SCC burners on whenever the kiln rotation was stopped.  If the SCC burners
should trip, they should be relit following proper procedures including an air
purge.  A second lesson learned from this experience is that if ever there is
any indication that an incident has contaminated the process waiter, the system
operation should be terminated because of the potential for the release of
contaminants from the stack.  The process water should be analyzed to determine
if contamination has occurred, and then be completely decontaminated if
necessary.

Risk Analysis

    Since a release of furans is potentially harmful to the surrounding
population, a risk analysis [4] was conducted to determine the impact of the
furans release.  The risk analysis was conducted using state-of-the-art
methodology and in accordance with EPA procedures.  In addition, the risk
analysis always assumed a "worst case scenario".  The results showed that the
incident was minor.  There was no additional risk to the public: resulting from
the incident.
                                      169

-------
HEP Problems

    The WEP caused 4.6% downtime as follows:  It was difficult to establish the
required high voltage during initial startup of the precipitator.  The
electrodes had to be positioned accurately within each box cell before voltage
could be sustained without breakdown.  After a hot cycle, the WEP again failed
to hold high voltage because of electrode shorting.  The electrodes were again
adjusted and reinforcing angles were added to the rod suspension system.  The
unit performed normally for 20 days, and then all three circuit boards failed
1n the control panel catastrophically.  This caused a three-day shutdown until
new replacement circuit boards could be installed.  No conclusive reason could
be established for the failure, but metal chips from the initial cabinet
assembly were thought to be the cause.  After replacing the boards and turning
off the fogging sprays, the unit sustained satisfactory operation.  It has been
necessary to perform a wash cycle daily or more often in heavy feed conditions
to maintain high voltage.  This wash cycle removes particulate buildup from the
surfaces of box cells.

    The fiberglass rods also were subject to a rupture failure mode that
occurred after repeated short cycles.  The shorting happened usually during
startups after a shutdown of the MIS.  This shorting is attributed to movement
of the rods due to expansion and contraction as the equipment warms and cools.
To overcome both the rod movement problem and the rupture failure problem, the
rod material was changed to Type 304 stainless steel.  The stainless steel rods
were installed in the unit during September 1988.  They solved the rupture
problem but periodic shorting still occurred.


AVAILABILITY DURING 1988 AND 1989

    During 1988 and 1989, the availability of the unit averaged 61%.  This
period was marked by constant improvement in the availability, because many
problem areas such as the SCC venturi nozzle, the kiln ash gates, the WEP rod
failure problem, the pH control system, the diesel fuel delivery system and the
spare parts inventory were addressed.  As shown in Figure 1, availability of
the HIS exceeded 70% from November 1988 until the end of Closure - a period of
6 months.  The 70% level was exceeded even when the unit was processing
brominated sludge, which was difficult to handle because its ash tended to foul
flue gas passages and instruments.  The Closure period was marked by a record
80% availability.

Kiln Ash Gates

    The kiln downtime was caused mainly by slagging.  The ash gates were
discovered as a source of the slagging in June, 1988 as follows.  During June
the rate of slagging had increased,and the kiln caused 37% MIS downtime.  Fuel
flowrates to the kiln also had increased.  Inspection of the ash gates during
troubleshooting of the problem showed that they had warped and did not close
properly.  The failure to close permitted excessive air leakage that cooled the
flue gases at the exit of the kiln.  Since the control thermocouple for the
                                      170

-------
kiln was downstream of the leakage point, It read an estimated 300°F to
400°F lower than the average gas temperature in the kiln.   This caused the
kiln burners to fire harder to achieve the setpoint temperature.   Consequently,
the kiln operated several hundred degrees higher than required.

    Inspection of the ash gate design showed that it could be improved to
provide tighter closure.  A new design was installed.  Following installation
of the new ash gates, the lost time from kiln slagging, which had averaged
16.8% during the first half of 1988, dropped to 5% as shown in Table 2.

pH Control Improvement

    During 1987 the MIS experienced difficulty in controlling the pH in the
process water system.  Corrosion in the process water system was attributed to
excursions when the process water became acidic for brief periods of time.  The
pH control was improved in stages starting with changes in 1987.   The nature of
the problem and the steps taken to solve it are described below.

    The scrubber system for the MIS utilized a caustic solution to neutralize
the HC1 gas formed and to maintain the scrubber water pH in the 7 to 9.5
range.  There were three areas in the MIS where the flue gas was scrubbed:

    -Quench elbow and sump
    -Lower WEP filler region
    -Mass transfer (MX) scrubber

    A 10% to 20% caustic solution was injected into the quench sump and into
the recirculation water flow of the MX scrubber. The initial arid largest
percentage of the neutralization occurred in the quench elbow and sump.  For
that reason the majority of the caustic had to be injected at this location.
The injection of caustic directly into the quench system tank from a line
controller by a solenoid valve located 25 feet upstream caused large swings in
the pH (3-10.5) of the quench sump.  The long injection line between the valve
and the quench sump caused a delay in caustic injection after the valve was
activated.  The pH continued to decrease during this delay. After the valve was
closed, the line caused an uncontrolled trail-off of injection.  The pH
continued to rise during the trail-off.  The swings could be controlled for any
one condition of organic chloride loading by changing the settling of the
timer.  However, when organic chloride loading changed sharply,, the timing
required manual changes.  Also, under heavy loadings the swing;; were still
excessive.  In 1987, the caustic injection valve was placed at the quench sump
to eliminate the delays.  This reduced the uncontrolled swings but did not
completely solve the problem.  Further steps were necessary as described below.

    The pH was controlled with pH sensors which operated in si Up streams from
the recirculation water loops.  These sensors fouled easily and required
frequent cleaning to keep them functional.  Also, the sensors were suitable for
only on/off control of the caustic solution.

    In 1988, the system was changed to modulating control using a Great
LakesK unit, which also featured a flow-through sensor.  This change produced


                                      171

-------
 smooth control  of the pH,  and the flow-through sensors were resistant to
 fouling.  Hence, control was improved and maintenance on the instrumentation
 and the process water system was decreased.

 Spare Parts  Inventory

    During 1988, a study was performed which identified critical spare parts.
 The study was conducted by maintenance experts who checked the operating
 requirements, the equipment manuals, and added their own experience to produce
 a list of critical spare parts.  These parts were purchased and put into
 inventory.   The spare part inventory prevented the downtime that was sometimes
 associated with waiting for a critical spare part to arrive.  This arrangement
 could save anywhere from a few hours to one or two days of time depending upon
 the availability of the failed part.

 Fouling

    The fouling consisted  of: deposits that restricted the flow areas in the
 duct system, the cyclone,  and the quench elbow; deposits that fouled the
 thermocouples,  the kiln oxygen analyzer, and other instrumentation; and
 sedimentation that filled  the quench sump and the CPI separator.  The rate of
 fouling was  highly dependent on the ash characteristics of the feed materials
 [9,10,12].   The ash from brominated sludge, which contained calcium and sulfur,
 formed a consolidated deposit which fouled the ductwork between the kiln and
 the SCC.  A  consolidated deposit also occasionally formed in the quench elbow
 upstream of  the quench nozzles.  This fouling necessitated a system shutdown
 about every  twelve days to remove the deposits.

    For brominated sludge, fouling caused 9.4% downtime (320 hr) but for most
 feed materials  fouling was controlled by the cleaning of the system during
 general maintenance periods.  The installation of clean out ports in the duct
 system was a necessary action to minimize the downtime caused by the ash from
 brominated sludge.

 Instrument Problems

    Instrument  problems accounted for 2.9% downtime.  These consisted of
 instrument failures such as the CO analyzer failure, weigh scale failures,
 broken thermocouples, etc.  Instrument problems were highest immediately after
 start-up during the Baldwin Park campaign.  Instruments contributed 10.3%
downtime (163 hr) during this period.  Much of this downtime (88 hr) was caused
by failure of the CO analyzer and computer problems in the analytical trailer.
After the Baldwin Park campaign, instrument problems occurred when the unit was
processing difficult feeds such as brominated sludge or trash.  When the unit
was processing  soil, instrument problems were minimal.


PROBLEM SUBSYSTEMS

    The systems viewpoint  of availability shows which part of the MIS
contributed most to the downtime.  As shown in Table 4, problems associated


                                      172

-------
with the kiln produced the greatest downtime during 1988-1989.  The 12.9%
downtime accounted for about 30% of the total MIS unavailability.  Most of the
downtime associated with the kiln was due to slagging.  This 1s shown In
Table 5.  The second largest category was downtime that affected the total
system.  These Included periods for general system maintenance and the testing
periods.  The 7.7% downtime affecting the total system accounted for about 18%
of the unavailability.  The duct system caused 5.3% downtime, which accounted
for about 13% of the unavailability.  The duct system downtime was caused
almost entirely by fouling by ash from the bromlnated sludge.  The SCC caused
5.1% downtime, which accounted for 12% of the unavailability.  The SCC downtime
was evenly divided between refractory repairs and the repair of the corroded
metal venturl.  A small percentage was caused by fouling.  The feed system
caused 3.9% downtime, which accounted for 9% of the unavailability.  This was
mainly caused by conveyor problems.

CONCLUSIONS

    The paper has discussed the factors that affected the availability of the
MIS during the operations at Denney Farm; has shown how the areas which caused
downtime were identified and addressed; and has presented the results of the
actions taken to Improve availability.  The conclusions based upon the
operating experience are presented below:

    o     By analyzing and correcting problems, the availability of the MIS was
          increased from 40% to 80%.  During the last 6 months of operation,
          availability exceeded 70%.

    o     Horizontal SCCs can become quickly plugged by carryover of
          participate matter from the kiln.  A cyclone placed between the kiln
          and SCC is effective in preventing this problem.

    o     Kiln slagging caused a significant percentage of the downtime.  Kiln
          slagging can be minimized by operating the kiln at minimum feasible
          temperatures, by blending slag-forming feed materials with large
          quantities of nonslagging feed materials, and by Insuring accurate
          temperature readings to prevent overfirlng.  Prevention of cold air
          leakage upstream of a control thermocouple is very important to
          prevent slagging.

    o     When processing hazardous wastes, Inadvertent contamination of clean
          parts of the system can be an added source of equipment downtime.

    o     Refractory materials were able to withstand the combination of high
          temperatures (>2000 °F) and high hydrochloric acid concentrations
          much better than high nickel alloys.

    o     Control of pH 1s important to prevent corrosion of the process water
          system.  Although the downstream end of the air pollution control
          system remained basic, other parts of the system periodically became
          acidic.  This caused corrosion 1n these parts of the system.
                                      173

-------
                                              TABLE A.  SYSTEM SOURCES OF NIS UNAVAILABILITY
                                                               1988-1969
emulative Downtime by HIS System Expressed as X Downtime
Feed Source or Activity

Baldwin Park
Baldwin Park Trash
Scoping Burn
Verona
Maintenance
Broni rated Sludge
Closure
Average by System
instr.
Air
0.06X
O.OOX
O.OOX
0.09X
O.OOX
4.35X
0.37X
1.59X
All

2.27%
O.OOX
55.13X
6.68X
100.00X
0.68X
3.45X
7.69X
APC

9.52X
O.OOX
O.OOX
0.41X
O.OOX
1.36X
O.OOX
2.08X
Ash

0.57X
O.OOX
O.OOX
0.63X
O.OOX
0.21X
0.05X
0.31X
Ducting

O.OOX
O.OOX
O.OOX
O.OOX
O.OOX
12.91X
4.71X
5.31X
ID Fan

O.OOX
O.OOX
O.OOX
5.51X
O.OOX
1.87X
O.OOX
1.87X
Feed

7.69X
1.58X
O.OOX
6.85X
O.OOX
1.45X
7.27X
3.92X
Kiln

28.75X
39.53X
O.OOX
30.23X
O.OOX
10.18X
2.56X
12.91X
Quench

0.69X
O.OOX
O.OOX
0.60X
O.OOX
3.23X
O.OOX
1.28X
sec

11.35X
O.OOX
O.OOX
9.79X
O.OOX
3.17X
1.57X
5.08X
Total
Unavail.
60.91X
41. 11X
44.87X
40.23X
100.00X
39.40X
19.98X

Note:  The selection of the causes of unavailability are dependent on engineering judgement.  This adds a degree of uncertainty.
       maintenance period caused an unavailability of 5X which included the UEP rod replacement, system cleanouts,  etc.
The

-------
                                    TABLE S.   SOURCES OF HIS UNAVAILABILITY BY TYPE AND PERCENTAGES
                                                              1988-1969
Cumulative Downtime X
Feed Source or Activity

Baldwin Park
Baldwin Park Trash
Scoping Burn
Verona
Maintenance
Broninated Sludge
Closure
Mech.

28.8BX
0.79X
O.OOX
14.04X
O.OOX
8.28X
7.01X
Feed
Jaroning
0.25X
0.40X
O.OOX
O.OOX
O.OOX
O.S9X
2.41X
Ash
Jamming
0.25X
O.OOX
O.OOX
O.OOX
O.OOX
0.03X
O.OOX
Elect.

2.65X
O.OOX
O.OOX
0.54X
O.OOX
1.74X
0.58X
Human

2.21X
O.OOX
O.OOX
0.99X
O.OOX
0.18X
O.OOX
Slagging
Kiln
16.33X
37.55X
O.OOX
17.16X
O.OOX
8.11X
O.OOX
Instr.

10.28X
2.37X
O.OOX
0.90X
O.OOX
2.71X
0.42X
Slagging
sec
0.06X
O.OOX
O.OOX
O.OOX
O.OOX
1.68X
O.OOX
Testing

O.OOX
O.OOX
44.87X
O.OOX
O.OOX
O.OOX
O.OOX
Kaint.

O.OOX
O.OOX
O.OOX
6.59X
100.00X
6.49X
3.40X
Foul ing

O.OOX
O.OOX
O.OOX
O.OOX
O.OOX
9.43X
6.17X
Average by Type
11.99X     0.72X
0.05X
1.25X
0.63X    10.20X
2.92X
0.59X
0.35X     8.94X
4.43X

-------
An adequate spare parts Inventory is important to minimize downtime
associated with obtaining parts which are not available onsite.

Adverse feed characteristics can cause internal fouling of ductwork.
An Incineration unit should be equipped with an ample number of
cleanout ports to facilitate cleanout.

Failures 1n the monitoring system required by the operating permit is
another added source of downtime.

Correction of a problem that 1s causing significant periods of
downtime does not Insure large increases in system availability.  The
problem may be replaced by another problem which is kept under
control by maintenance when the system 1s down.
                            176

-------
                                   REFERENCES


 1.  Yezzi,  J.J., Jr. et al.   Results  of the Initial  Trial  Burn of the EPA-ORD
    Mobile  Incineration Systems"  In: Proceedings of the 1984 National  Waste
    Processing Conference,  ASME, New  York,  New York, pp. 514-534.
 2.  Lovell, R.J., et al.  Trial Burn Testing of the EPA-ORD Mobile Incineration
    System.  EPA-600/D-84-054, Municipal Environmental  Research Laboratory,
    Cincinnati, Ohio, 1984.

 3.  Mortensen, H. et al.   Destruction of Dioxin-Contaminated Solids and Liquids
    by Mobile Incinerations,  EPA-600/2-87-033, Hazardous  Waste Engineering
    Research Laboratory,  Cincinnati,  Ohio,  1987.

 4.  King, 6., and Stumbar,  J.  Demonstration Test Report for Rotary Kiln Mobile
    Incinerator System at the James Denney Farm Site, McDowell, Missouri.  EPA
    Contract 68-03-3255,  Risk Reduction Engineering Laboratory, Cincinnati,
    Ohio, 1988.

 5.  Gupta,  6.D., etal  EPA Mobile Incineration System Modifications, Testing,
    and Operations February 1986 to June 1989  Draft Report, EPA Contract
    69-03-3255, Risk Reduction Engineering Laboratory,  Edison, New Jersey,
    1990.

 6.  Freestone, F.J., et al   Evaluation of On-site Incineration for Cleanup of
    Dioxin-Contaminated Materials.  Nuclear and Chemical Waste Management, Vol
    7, pp 3-20, 1987.

 7.  Miller, R.  Discussion of Feeds Processed in the MIS during July and August
    1985  Private Communication, July 1989.

 8.  Mortensen, H.  Discussion of Feeds Processed in the MIS from October 1985
    to February 1986  Private Communication. July 1989.

 9.  Bryers, R.W.  Deposit Analysis:  Cyclone Riser/Quench Elbow - Denney Farm
    Site, Foster Wheeler Development  Corp., Livingston, New Jersey, 1988.

10.  Bryers, R.W. Examination of Fouling of Convective Heat Transfer Surface by
    Calcium and Sodium Using Micro-Analytical Techniques.  ASME Paper
    86-JP6C-FACT 5, ASME, New York, NY, 1986.

11.  Gupta,  G.D., et al.  Operating Experiences with EPA's Mobile Incineration
    System, In:  Proceedings of the International Symposium on Incineration of
    Hazardous Municipal,  and Other Wastes.   American Flame Research Committee,
    Palm Springs, CA, 1987.

12.  Stumbar, J.P., et al.  Effect of Feed Characteristics on thi Performance of
    EPA's Mobile Incineration System.  Proceedings of EPA 15th Annual Research
    Symposium, EPA, Cincinnati, Ohio, 1989.
                                      177

-------
 CHARACTERIZATION OF INTERNAL INSPECTION PROCEDURES
      & EQUIPMENT FOR UNDERGROUND STORAGE TANKS

           by:   Susan E. Rohland
                 PEI Associates, Inc.
                 Edison, NJ   08837

                 Michael  Borst
                 CDM Federal Programs Corp.
                 Edison, NJ   08818

                 Robert W. Hillger
                 U.S. EPA, REEL, RGB
                 Edison, NJ   08837
                             ABSTRACT

      Efforts are presently  underway to examine the various  tools
and  techniques used for conducting internal inspections.   This study
documents  the  significant  factors  that  are  evaluated during an
inspection.   It  examines the application of each inspection method  by
identifying  the  procedural  steps,  necessary equipment  and
instrumentation  involved, and  the  circumstances  under which the
method  is  performed.   This  paper  presents an overview of  the
inspection  methods  currently being used  by practitioners.
                                178

-------
                           INTRODUCTION

IMPETUS FOR THE STUDY

      Releases from  underground  storage  tanks  (USTs)  are  increasing
in number nationwide.   Internal inspections  fulfill an  important role
in the repair, maintenance, and  release prevention  of  USTs.   Existing
inspection  methods  and  techniques  range  from  visual inspection to
ultrasonic and  magnetic  testing.  A common inspection  method, for
example, involves striking  a  ball peen hammer on  the  interior  wall  of
a tank and  making  a determination on its  thickness  based  on the
ringing sound produced.  The  U.S.  EPA's Risk Reduction  Engineering
Laboratory is  presently  directing  a project  to  characterize internal
procedures and equipment for  performing internal inspections  of
USTs. In this  study, the technical  applicability of these different
inspection techniques, in light of the UST regulations, was
investigated.

      While  the Federal regulations recommend internal  inspection
procedures developed by nationally recognized  organizations, such  as
the American  Petroleum Institute  (API) and  the National Leak
Prevention  Association  (NLPA), for use in  repair or  upgrading  of
USTs (40 CFR 280.33  and 280.21, respectively), internal inspections
are performed for other reasons.   These  circumstances  include:
structural integrity  evaluations, tank  reconditioning,  tank  closure,
and  change-in-service.   In addition to the types of  internal  inspection
referenced  in  the regulations,  many others  have been  developed  by
tank  manufacturers  and  engineering societies and  are  in various
degrees of use.


PROJECT OBJECTIVES

      This study  identifies and characterizes current techniques used
for conducting internal  inspections of underground  storage tanks.
The  study investigated  state-of-the-art (SOTA)  for  internal
inspection  methods.   The  term 'state-of-the-art'  is  defined  as the
                                 179

-------
current  level  of  development  and capability in terms  of procedure,
process,  and technique in current practice.   This  study includes a
review of  existing industry  standards and  practices.   It  also contains
a review of  non-destructive test methods.
                             APPROACH

      The  research effort  associated with this  study  consisted of a
review  of  available  literature and  conversations  with  trade  and
professional  associations, manufacturers,  vendors,  and  independent
consultants.  No sampling or analyses  was performed  in  conjunction
with this study.  No  direct  observation  of  procedures  has  been
undertaken.

LITERATURE SEARCH

      An  extensive  review of current  literature related  to  internal
inspection  of storage tanks   utilized  both  computerized  literature data
bases and  hard copy  documents  found  in engineering  and  technical
libraries at various universities  and the  EPA.   In general,  little
information on internal UST  inspections was  available.   Detailed
documentation  on several types  of inspection  equipment, however,
was  obtained from  non-destructive testing  handbooks  and ASME
Boiler And Pressure  Vessel  Code.

CONTACTS WITH PROFESSIONAL ASSOCIATIONS

      Professional and trade  associations were  the  most  valuable
source of information  on  SOT A  practices.   Types of information
gathered included:   general  descriptions of  the  procedures  and  the
steps involved in performing them,  examples of the  instruments
used  to conduct  the inspection,  and  details on  method performance.
Some basic  information on  field considerations  (such  as  permitting
requirements, portability  of  equipment,  cost  to perform the
inspection  and any  special  considerations for waste generated  during
                                 180

-------
the inspection) was  also collected.  The organizations contacted  are
presented  here:

      •   Petroleum  Equipment  Institute  (PEI)
      •   American  Petroleum  Institute  (API)
      •   Steel Tank Institute (STI)
      •   National  Leak Prevention Association  (NLPA)
      »   Fiberglass Petroleum Tank & Piping Institute (FPT&PI)
      •   National Association of Corrosion Engineers (NACE)
      •   American Society of Testing  & Materials (ASTM)
      •   Association of Composite Tanks  (ACT)
      •   Non-Destructive  Testing  Association (NDTA)
      •   American Society of Mechanical Engineers (ASME)

CONTACTS WITH MANUFACTURERS, VENDORS & CONSULTANTS

      Representatives  of  the industries  which manufacture  tanks,
apply linings, or provide inspection  services were  contacted  to  obtain
details on  particular inspection  methods.  Among  the  field from
which information was obtained  are:  Owens-Corning (an  FRP  tank
manufacturer),  Buffalo Tank Company  (manufacturer  of  steel tanks),
(Bridgeport Chemical  Company (a manufacturer and installer  of
lining resins).   Conversations  with these representatives  focused on
the identification  of procedures  and  equipment currently  in  practice.
Information on  potentially  applicable  or emerging  internal  inspection
methods  and procedures was also gleaned from  these  discussions.

REVIEW OF APPLICABLE STANDARDS

      Since the regulations rely  on existing  inspection methods,
efforts also focused on investigating  standards relevant to UST
inspections.  No  single standard describes procedures for  all  types of
internal  inspections.   Current  standards cover activities  performed
during the installation and  upgrading  of tanks.   Inspection
procedures relating to UST design and construction have  been
developed  by tank manufacturers as  part of  their  quality
                                 181

-------
assurance/quality control protocols.   In general,  these  serve  as pre-
installation  requirements.   Closure-  and cleaning-  oriented
inspections  have only  been marginally  addressed in  the  survey of
information.

API Recommended  Practice  1631. Internal Lining  of Underground
Storage  Tanks

      This  recommended  practice describes  requirements and
procedures for the application of  lining material  to the interior  of
existing  underground  petroleum  storage  tanks.
NLPA Standard 631,  Internal  Inspection. Repair and Lining  of Steel
and  Fiberglass  Storage  Tanks

      This standard  creates specific  requirements to evaluate  the
appropriateness  of issuing a  permit  to  perform  internal inspections
and  provides  criteria  to evaluate  the lining materials  used.

NLPA Standard 632.  Internal  Assessment  of  Steel Tanks 10  Years
and  Older for Corrosion Holes and  Structural  Soundness

      This standard  provides  the minimal internal  inspection
procedures to obtain  accurate assessment  of  the structural   soundness
and  the  amount and  level  of  corrosion prior  to  installing a corrosion
protection system.

ACT 100. Specification  for The Fabrication of FRP Clad/Composite
Underground  Storage Tanks

      This standard provides guidelines  for the  fabrication of FRP
clad/composite  underground storage  tanks  that  meet UL58   "basic
tanks"  standards.
                                 182

-------
FPTPI Recommended Practice-T-89-lT  Remanufacturing  of Fiberglass
Reinforced  Plastic  Underground Tanks

      This recommended practice  addresses the accepted  practices
by  the FRP industry  for remanufacturing  of fiberglass reinforced
tanks,

Underwriters  Laboratory  Standard  1316.  Glass-Fiber-Reinforced
Plastic Underground  Storage Tanks  for Petroleum Products

      This recommended practice  is  designed  for  fiberglas,s-
reinforced  plastic (FRP) USTs  to  contain petroleum products that are
covered by UL.
                          STUDY FINDINGS

OVERVIEW OF INTERNAL INSPECTION METHODS

      Eighteen methods have  been  identified  during the course of
this  study  and are listed  in  Table 1.   Telephone discussions  imply
that  only  a few of these methods are  actively in use as current
practice (visual observation,  hammer testing).   Many  others are
applied infrequently  throughout  the  country.   In addition,  our
research has identified  a few  technologies (e.g.,  acoustic emission) as
being  speculative  for internal UST inspection applications.

UST ACTIVITIES INVOLVING INTERNAL INSPECTIONS

      Internal inspections fall into one of several  categories
depending  upon the activity being performed  on  the UST.   These
inspections are conducted  during installation, routine  maintenance,
repair, lining, change-in-service, or closure.   In  addition, many tank
manufacturers (both  steel and  FRP)  have instituted stringent  in-
house  internal inspection procedures  at  various  stages  of the
manufacturing process.   Some states now require  inspections  by  state
UST personnel or certified  inspectors at the time tanks are  closed.
                                 183

-------
                  Table 1. Distribution of Internal Methods By Type of Activity and Tank Material
METHOD
TANK DISCONTINUTIES
Visual (Human Eye)
Visual (Boroscope) *
Liquid Dye Penetrant
Magnetic Particle
Eddy Current »•
Radiography ••
Acoustic Emission *
Microwave-
TANK TIGHTNESS
Bubble Test
Pressure Change
Vacuum Test
TANK THICKNESS
Ultrasonic
Hammer Testing
TANK UNING
HoBdayTest
Lining Hardness Test
Dry Film Thickness
TANK DEFLECTIONS
Internal Tank Diameter
TANK CONTAMINATION
Tank Cleaning
INSTALLAION
STEEL FRP

X








X
X
X












X








X
X
X








X


UPGRADING
STEEL FRP

X
X
X
X
X
X
X
X





X
X

X
X
X



X

X
X
X


X
X
X



X

X


X
X
X

X

X
CLOSURE
STEEL FRP

X





















X

X





















X
REMANUFACTURING
STEEL FRP

X

X
X
X
X
X
X


X


X








X

X

X



X
X

X
X
X

X


X
X
X



X
Emerging technologies
Applicable technologies not practiced

-------
      Installation  inspections are designed to identify any  structural
damage immediately after  the  tank  is  in  place and to check tank
tightness.   Upgrading  inspection  activities  are designed to  evaluate:
tank lining, tank  discontinuities,  tank  wall thickness,  and tank
deflection.  Inspections at closure assist in ascertaining  the  level of
cleanliness  before removal  or  abandonment-in-place.    A  new
element  in  internal  inspections is the  remanufacture  of  tanks  that
have been taken out of service.   This  is  a very new  and rapidly
evolving  business  area and  information is  still being  collected.

METHODS IN ACTIVE PRACTICE

Visual  Inspection

      This  method  is  used to  detect any  surface discontinuities such
as cracks and  other  porosities  in  both  steel and FRP tanks.  It  may
also be used  to determine the subsurface conditions of an FRP  with
the use of  other  instruments.  The  quality  of the inspection is
primarily affected by  lighting  conditions  in the tank  and the
experience  of  the  inspector.   Visual inspection is  an  inexpensive,
quick  method  of  evaluation  but requires  tank entry.   It  does not
enable the  inspector to  detect  small surface  or subsurface
discontinuities.

Liquid  Dye Penetrant

This  simple,  non-destructive test is performed to  detect
discontinuities (such as  cracks, pits or delaminations) on the surface
of steel or  FRP tanks.  A liquid  penetrant is  applied to a  cleaned tank
surface and allowed to be absorbed by capillary action into
discontinuities.   A developer is then  applied  to draw  the  penetrant
out of cracks.  Depending upon the type  of  penetrant used, the
resulting  indications are  visible under  normal or  ultraviolet light.
The accuracy  of  the inspection method is  heavily  influenced by the
presence  of any  contaminants  which  may interfere with  the
penetrant process.   This method is  also susceptible to moist or  high
humidity  conditions which can inhibit  the capillary action.
                                  185

-------
Magnetic Particle  Test

      This method  uses  an induced  magnetic field  in  a localized  area
of the tank surface as a magnetic powder is  applied.  The flux
leakage  created  by a  surface discontinuity will  attract  the magnetic
powder and result in  a visual indication.   ASTM has  described  a
standard practice for  the  performance  of magnetic particle  tests.
Reliability of  the results is  dependant  upon  the  direction of the
magnetic field and  false  indications may result  if  magnetization is  too
high.  This is a  non-destructive test for  the  interior surfaces of steel
tanks which is most  effective in  detecting discontinuities on  the
surface of the tank.   The requisite  equipment  is relatively  small and
versatile for  field  applications.

Bubble  Testing

      The  principle of a bubble  test is  to locate a  leak by applying  a
solution  that will form bubbles when a gas  passes  through it.   If no
continuous  bubble  formation  is  observed, the UST is  considered
acceptable.   Surface tension  of  the  bubble solution must be
sufficiently low  to  allow for  the detection of small holes.  Particular
attention is given to  tank seams  and  openings (such  as  unused ports).
Bubble tests  can be  performed  using either  direct  pressure or  a
vacuum.   Direct  pressure  tests  are  more common.

Pressure Test

      A positive  pressure  test involves  the  introduction  of a known
pressure of air or other ambient gas to the  tank after  all ports  of
entry have  been sealed off.   In  order for the tank  to  pass the  test,  it
must hold a  designated  level of pressure for a prescribed period of
time.

Vacuum Test

      A negative pressure  test  may  be performed  using  the
principles  described for the  positive pressure  test  in  reverse  fashion.
After all ports of entry have been  sealed, air  is withdrawn  from the
tank  until a  specified negative  pressure  inside  the tank  is achieved.
If the tank maintains  this  pressure  over a designated  period  of  time,
it passes the test.

                                   186

-------
Holiday  Tests

      Holiday  detectors are used  to ascertain  the  presence of
discontinuities  such as voids,  cracks,  thin  spots, foreign inclusions, or
contamination in the coating film.   The principle of the test  is to
connect  an electromagnetic sensitive relay or  solid-state  electronic
relay circuit that  energizes  an audible or visual  indicator.   A  voltage
detector  (low  and high voltage equipment  is  available) is employed
according to the thickness of the lining.

Lining Hardness  Test

      This  test is performed  to verify  whether  an applied  protective
coating film has  cured to  a  hardness that meets  the  manufacturer's
specifications.   An ASTM  standard has  been developed for  the
determination  of  indentation  hardness of  both reinforced  and non-
reinforced  rigid plastics using  a Barcol impressor.   This device
distributes  force,  using a hardened  truncated  steel cone,  to determine
the lining  hardness in a steel  tank.  The method is  fully  portable and
should not  generate  a  hazardous waste  or require additional  safety
considerations  beyond those  associated  with  tank entry.

Dry Film  Thickness  Measurement

      In this  method, an elcometer measures  the  thickness  of non-
ferromagnetic  coating films including paint, electroplating,
galvanizing, powder,  plastic,  and rubber applied to a  ferromagnetic
base.  This instrument can  be used to gauge the thickness of  steel
tank coatings.   Overall accuracy of the procedure  is high;  small  errors
may arise from  inconsistencies in  the surface profile  or  the  substrate
material.   There  is no standard procedure for this  instrument.

Ultrasonic  Testing

      This  method  may be  used to determine  the  wall  thickness  of a
steel or  FRP  UST and may also detect any surface or subsurface
discontinuities.    In this  method,  an  ultrasonic  wave is induced  at the
surface of the tank which propagates through  the  wall.  The  wave
travels  through the tank  medium  at a  constant velocity  dependent
upon the physical properties of the tank.  The wave is reflected back
to  the instrument  when  it  encounters  a  discontinuity or wall
boundary.   The transient  time of the wave is measured by  a  pulse-
echo  instrument  to calculate the wall thickness.   ASTM delineates
                                  187

-------
objectives for performing  the  test.   API  1631 indicates  that
ultrasonic measurements  are  used for UST  inspections,  but provides
no  guidelines for its application.

ffampier Testing

      In this  method, an  inspector strikes a  brass hammer against  a
cleaned tank  surface.   The  vibration and  rebound of the hammer and
the sound produced is  dependent upon  wall  thickness.  There  is  no
proven  method  available  to  evaluate  the  effectiveness of  hammer
testing.  This is considered a non-destructive test for steel tanks
which can  be performed  quickly  at relatively low cost.   It is
imperative,  however, that an  experienced inspector  be engaged  to
obtain the  most  accurate results.   Both API  1631 and NLPA 631
describe hammer testing in  varying levels of detail.

Internal  Tank Diameter

      Precise  measurements  are  taken and a series of calculations  are
performed  to  compute  the internal diameter   of the tank.  From  these
calculations, it can  be  determined whether or not the tank has been
deflected and by what  amount  (percent).   Some standards require  no
greater  than a 1% deflection  to  ensure  structural integrity  of  the
tank.

      Table 2 depicts  the various  circumstances  under  which internal
inspections  are performed and the  types  of  tanks for which the
method is  appropriate.   Throughout the course of the study,  two
methods were identified  as internal  inspection procedures,  but  are
not actively used by  the contractors who provide inspection services.
These methods   are radiography  and eddy current.

EMERGING TECHNOLOGIES

      In keeping an eye to  the  future of  inspections  and how  the
Agency may  assist states and owner/operators in complying with  the
regulations, preliminary information  on emerging methods is  also
being collected.   Such emerging technologies have been  developed  for
use in  other  applications  but are not yet performed  on  underground
storage tanks.  A typical  source  of an  emerging technology is  the
pressurized  vessel  industry.   Examples of such technologies are:
acoustic emission,  the  boroscope, and  microwave.

                                 188

-------
   INSTALLATION
    Steel/FRP

1. Visual Inspection
2. Pressure Test
3. Internal Tank
  Diameter
  (FRP only)
4. Vacuum Test
5. Bubble Test
     UPGRADING
      CLOSURE
 REMANUFACTURING
        Steel

 1. Tank Cleaning
 2. Ultrasonic Test
 3. Visual Inspection
 4. Holiday Test
 5. Dry Rim Thickness
 6. Barcol Test (optional)
 7. Pressure Test
 8. Hammer Test
 9. Liquid Dye Penetrant
10. Magnetic Particle
      FRP

 1. Tank Cleaning
 2. Ultrasonic Test
 3. Vacuum Test
 4. Internal Tank Diameter
 5. Visual Inspection
 6. Holiday Test
 7. Dry Film Thickness
 8. Barcol Test (optional)
 9. Pressure Test
1C. Liquid Dye Penstrarst
     Steel /FRP

1. Tank Cleaning
2. Visual Inspection
        Steel

1. Tank Cleaning
2. Ultrasonic Test
3. Visual Inspection
4. Pressure Test
5. Liquid Dye Penetrant
6. Magnetic Particle
                                                                                                 FRP

                                                                                           1. Tank Cleaning
                                                                                           2. Ultrasonic Test
                                                                                           3. Vacuum Test
                                                                                           4. Visual Inspection
                                                                                           5, Holiday Test
                                                                                           6. Dry Film Thickness
                                                                                           7. Barcol Test (optional)
                                                                                           8. Pressure Test
                                                                                           9. Liquid Dye Penetrant
                                                                                          10. Bubble Test
                    Table 2. Typical Internal Measurement Methods Employed During Inspections

-------
                             CONCLUSIONS

      Available information  suggests that  no compilation of  SOTA
internal  inspection methods  has  been prepared  to  date.   The
identification  and description of these methods in a  single  document
would serve  to  educate owner/operators  in  selecting the  most
appropriate type of  inspection under each  circumstance (i.e.,
installation, closure,  etc.)  during  the operational life of  the tank.  It
would provide state regulatory  authorities  with  information
necessary  to  design inspection programs and  provide  contractors
with  a variety of UST inspection techniques.

      Internal  inspections  fall  into  four categories: installation,
upgrading, closure,   and  remanufacture.    Several  methods  have also
been  identified during the pre-installation phase  of  the tank's  life.
However, since the  focus  of the  study is  the operational life of the
tank,  pre-installation methods  are not emphasized in  this study.

      One obvious finding of this study is that, with the exception of
visual inspections, the  type of inspection performed  depends  on  the
status of the  tank.   For each major activity in  the operational life of
the tank, there are  several  different inspection  methods  which  may
be applied.  Ultimate selection of one method  in  lieu of another
usually  depends  upon  the desired  degree of accuracy  of the results.

      An examination  of  various  applicable standards  has revealed
that  many contain requirements  which  conflict with those of  other
standards.  In some instances, conflicts  between  requirements exist
within the same standard.   For  instance, vacuum  tests  and  tank
deformation tests  are not  required by EPA in  the preamble  to the
regulations, but  the  standards  called  out in the preamble provide
reference to  these  methods.

      Little information is  presently available  on  the impact of tank
size  on the performance of the test.  This is because most methods
require  that  the  test be performed  a certain number of times  within
a specified  surface area of the tank.  The test is  then repeated at
regularly defined  intervals  across the entire surface of the  tank.
Without  the benefit  of further  investigation, this  would suggest  that
tank size has no influence  on  the performance  of the test  or its
results.

                                  190

-------
            RECOMMENDATIONS FOR FUTURE ACTIVITIES

      For the balance  of  this fiscal  year, the scope of this work
assignment will focus on the evaluation of  the  methods identified
earlier in our discussion  in  order to  establish  a full understanding of
their range of applications.   Among  the specific  tasks anticipated are:

      *  Obtaining data  to reveal how or if these inspections are
         being applied by tank  owner/operators  as part  of  a
         comprehensive  maintenance program,

      »  Investigating  the necessity  to develop protocol manuals to
         ensure the standard application  of methods by  vendors
         offering  internal inspection  services.

      *  Studying the  effects of  tank  size on  test results;  particularly
         in  the  categories of tanks  less than  30,000 gallons, tanks
         ranging from 30,000 to 60,000 gallons,  and tanks  greater
         than 60,000  gallons on  capacity.
                        ACKNOWLEDGEMENTS

      The  authors would like  to  acknowledge  the efforts  of  the
following people  in  providing  technical review  or  research  assistance
in the preparation of these  proceedings:  Evan  Fan and Jim Yezzi-EPA,
ORD, RREL, RGB; Paul Mraz, Roy Chaudet, and John Miller-PEI
Associates, Inc.;  John Mazza and Joan Knapp-CDM  Federal Programs
Corporation.
                                 191

-------
    EVALUATION OF INTERNAL LEAK DETECTION TECHNOLOGY
             FOR LARGE UNDERGROUND STORAGE TANKS
                Joseph W. Maresca, Jr., James W. Starr, and Richard F. Wise

                                  Vista Research, Inc.
                            Mountain View, California 94042

                         Robert W. Hillger and Anthony N. Tafuri

                         Risk Reduction Engineering Laboratory
                         U. S. Environmental Protection Agency
                               Edison, New Jersey 08837

                                    ABSTRACT
           The performance standard for tank tightness testing established by the EPA
     regulation requires that the systems used to test underground storage tanks be able to
     detect leaks as small as 0.1 gallons per hour with a probability of detection of 0.95
     and a probability of false alarm of 0.05.  This standard was developed to address
     tanks nominally 37,850 liters (10,000 gallons) in capacity or less, but it also applies
     to tanks as large as 227,140 liters (60,000 gallons). The accuracy of detecting leaks
     in large tanks is not well known, and very little experimental data are available to
     make an assessment.  An experimental field study is being conducted to investigate
     the magnitude of the volume changes that affect the performance of leak detection
     systems based on tank tightness testing in underground storage tanks with capacities
     between 75,710 and 227,140 liters (20,000 and 60,000 gallons).  The experiments
     are designed to estimate the magnitude of the volume changes produced by two of
     the more important sources of noise that need to be compensated for in order to
     obtain good performance: (1) thermal expansion or contraction of the product in the
     tank and (2) structural deformation of the tank itself. Models of these physical pro-
     cesses were developed in a previous study on smaller tanks. The data collected in
     the current experiments will be used to determine whether these models are valid for
     tanks that are up to ten times larger. In addition, the data obtained from several
     thermistor arrays deployed in the tank will be used to estimate the magnitude of the
     horizontal and vertical changes in the temperature of the product in the tank. This
     paper describes the experiments that will be performed and presents an analysis of
     the instrumentation.


                                 INTRODUCTION
     The United States Environmental Protection Agency (EPA) regulation for underground
storage tanks (USTs), published on 23 September 1988, specifies the technical standards  and a
variety of release detection options for rninimizing the environmental impact of tank leakage [1].
With several exceptions, shop-assembled tanks covered by the regulation range in size from sev-
eral hundred gallons in capacity to very large tanks, with no clearly defined upper limit.
Requirements for large, field-erected tanks have not yet been established,
     For the tanks currently covered by the regulation, the EPA performance standards for volu-
metric tank tightness tests require that leaks as small as 380 ml/h (0.10 gal/h) must be detected
with a probability of detection (PD) of at least 0.95 and a probability of false alarm (PFA) of 0.05.
Tanks as large as  227,140 L (60,000 gal) are covered by this regulation. Based on the results of
an extensive experimental program conducted  at the Underground Storage Tank Test Apparatus
in Edison, New Jersey, the EPA concluded that achieving this level of performance is within the
capabilities of the state of the art for tanks with capacities of approximately 37,850 L
(10,000 gal) or less [2], The EPA developed general design recommendations for tank tightness
systems so that this level of performance could be achieved in tests conducted in overfilled and
                                        192

-------
partially filled tanks of this capacity. The EPA also described the features that are required in
order for a tank tightness test to meet its standards.  A leak detection system measures flow rate
due to a leak, but errors associated with (a) temperature changes in the product inside the tank
and (b) structural deformation of the tank itself can interfere with effective detection. The
important features, then, are the ones that compensate for these errors. Experiments on a
30,280-L (8,000-gal) tank showed that an array of five or more equally spaced temperature sen-
sors was  sufficient to compensate for the thermally induced volume changes providing that ade-
quate waiting periods were adhered to after any addition of product, whether this addition
represented a delivery  to the tank or whether it constituted topping of the tank (as is required
when testing an overfilled tank). The addition of product to the tank produced inhomogeneities
in the temperature field that were large enough to prevent an accurate estimate of the mean rate
of temperature. This estimate is required for thermal compensation.  As a means of minimizing
the effect of these thermal inhomogeneities, waiting periods of at least 3 h after topping and 6 h
after a delivery were recommended. Any addition also changes the level of product in the tank
and, therefore, the pressure that is exerted on the tank walls.  This change in pressure causes the
tank to deform. The waiting period required to allow the deformation of the tank to subside
depends on the properties of the  tank and the backfill and native soil surrounding the tank; it
could be  as much as 12 to 18 h.
     Little data are currently available, however, to indicate whether these recommendations are
valid for  the larger tanks covered by the regulation, and therefore it is unclear whether a tank
tightness test that meets the standard established for smaller tanks can achieve the same level of
performance when used to test larger tanks. Based upon previous experimental and analytical
work, we expect that larger tanks will require (1) more temperature sensors and more accurate
estimates of the coefficient of thermal expansion for temperature compensation, (2) longer wait-
ing periods after topping or after a delivery of product to the tank, and (3) an increase in the pre-
cision requirements of the temperature and level measurement systems used by the tank tightness
test methods.  The horizontal and vertical distribution of temperature after topping and after
delivery has not been systematically investigated for large tanks;  without these data it is not pos-
sible to determine the duration of the waiting periods required to let the inhomogeneities in the
temperature field decay and to determine the number of temperature sensors needed for
compensation. No data exist to describe the temporal characteristics of the volume changes pro-
duced by deformation  in large tanks. Until such data are gathered, the duration of the waiting
period required for these volume changes to become negligible is unknown.

                                   OBJECTIVES
     Two sets of experiments will be conducted to investigate the magnitude of the temperature
and deformation effects associated with a tank tightness test; one experiment will be conducted
in a tank  with a capacity of 75,700 L (20,000 gal) and the other in a tank with a capacity of
181,680 L (48,000 gal) or greater.  Based on these tests, an assessment will be made of the abil-
ity of tank tightness tests to meet the EPA standard when such tests are used on larger tanks.
Recommendations for minimizing the errors associated with thermal inhomogeneities and
structural deformation will be made, so that the best possible performance can be achieved.
     This paper describes the experiments that will be conducted and presents the results of an
analysis of the requirements for temperature  and level sensors.  This analysis gives the precision
required  of the sensors and the duration of the data collection period in order for a tank tightness
test on a  181,680-L (48,000-gal) tank to meet the EPA standard.

                      MEASUREMENT METHODOLOGY
     The temperature and deformation effects were characterized in previous work conducted at
the Test Apparatus in tanks of a nominal 30,280-L (8,000-gal) capacity [2].  The current work is
thus focused on extending the experimental base to tanks significantly larger than those installed
at the Test Apparatus.  The previous studies will be used as a baseline against which to compare
                                          m

-------
the results from the experimental data developed in the current work. The currently planned test
program, although more limited in scope than the original study, draws extensively from the
instrumentation and experimental designs developed during the course of this previous work.
     The current test program requires that both product temperature and level measurements be
made for a range of initial conditions in two different-size tanks. A limited range of initial con-
ditions will be generated in each tank, and the tank response then recorded over a period of 6 to
24 hours. Details of specific experimental configurations are discussed in the following sections.
     Experiments will be conducted on a 75,710-L and a 181,710-L (a 20,000-gaI and  a
48,000-gal) UST to estimate the magnitude of the major sources of noise that affect the perform-
ance of internal leak detection tests. The experiments will be conducted at operational UST faci-
lities. The experiments are designed so that the impact on operations at the facility being used
will be minimal. The sensors will be calibrated prior to any field testing and prepared for
immediate implementation at the field-test site. The product temperature changes in the tank
will be examined under three conditions: a nearly full tank, an overfilled tank, and a partially
filled tank.  A special set of experiments will be conducted when the tank is overfilled to deter-
mine the deformation characteristics of the tank. Approximately four days of testing will be
required for each tank.
     The testing program is designed to address the following questions: (1) what are the verti-
cal and horizontal characteristics of the product temperature field, (2) what are the structural
deformation characteristics of the tank system, (3) how long does it take for the temperature
inhomogeneities due to topping and delivery to subside, and (4) what are the requirements of the
temperature and level instrumentation?
     Temperature Field. Accurate temperature compensation requires that a sufficient number
of thermistors be located in the vertical and that the horizontal differences in temperature be
small.  The two arrays of thermistors will be used to make an estimate of the vertical and hori-
zontal differences in temperature before, during, and after the addition of product in each of two
circumstances, a delivery and topping. Data will be collected: (1) after a delivery that brings the
level of product to about 75% of the tank height, (2) after a second delivery that brings the level
of product to  about 95%, and (3) after the tank has been topped such that the level of product is
within the fill tube.  At each level of product, the ability of a tank tightness test to compensate for
the thermally induced volume changes after the spatial inhomogeneities due to product addition
have subsided will be assessed.
     Structural Deformation,  Structural deformation of the tank occurs in response to any
change in the hydrostatic pressure on the* tank.  Previous studies of this phenomenon suggest that
the response of a tank to a step change in pressure (or level) is exponential in form. The details
of the shape of the exponential response are largely influenced by the tank material, type of
backfill, water table level, and local native soil conditions.  Experiments conducted on the EPA
tanks in Edison indicate that a relaxation time constant of three hours and an elasticity constant
of 100 cm2 are not uncommon for the backfill/soil conditions prevalent at the Test Apparatus.
     In a nonleaking tank the time constant of the tank can be determined by producing an
instantaneous level change in the fill tube of an overfilled tank and monitoring the time history of
the volume changes. The time constant of the tank can be determined by fitting an exponential
curve to the cumulative volume changes required to maintain a constant level in the tank.  An
instantaneous level change is produced by inserting  a bar of known volume into the tank, after
waiting for any previous deformation effects to subside.
     Waiting Periods After Topping and Delivery. There  are two effects produced by topping
and delivery that will affect the performance of a volumetric test method: (1) temperature inho-
mogeneities and (2) structural deformation. Both effects diminish with time and so can be mini-
mized if there is a mandatory waiting period between filling/topping and testing.  The effects of
topping and delivery will be determined from the temperature and level data obtained
immediately before and  after product is added to the tank.
                                          184

-------
     Instrumentation Requirements. In large tanks, the instrumentation noise can be a signifi-
cant fraction of the total noise.  One way to overcome high instrumentation noise is to increase
the sampling rate or duration of the test.  The requirements of the instrumentation can be
determined from an analysis that includes the resolution, the system noise, the duration of the
test, and the number of independent samples acquired during a test.

        TEMPERATURE AND LEVEL MEASUREMENT SYSTEMS
     The systems that will be used to conduct the field experiments consist of two types of mea-
suring devices:  temperature and level (height). The temperature measurement system will be
used primarily to monitor the product; the product temperature data will be anzilyzed to estimate
the thermally induced volume changes in the tank. Air temperature will also be measured;  air
temperature data are not central to the experiments or the analysis, however, and will be used
only to characterize weather conditions during the experiments.
     Two types of product level measurements are required. The first is to measure the height
of the product from the bottom of the tank;  a pressure sensor with a precision of 5 cm or less will
be used. Four sensors will be used to measure the change in product level at each average
height. A pressure sensor will be used to measure the level changes when the i:ank is overfilled
into the fill tube, and an electromechanical sensor, an acoustic sensor, or a fiber-optic sensor will
be used to measure level changes when the  tank is partially filled.  The three level sensors that
will be used in a partially filled tank are provided by Vista Research. The choice of the primary
sensor will depend on the type of experiment being conducted and the availability of access
points on the tank. Based upon physical constraints at the test site, efforts will  be made to pro-
vide data redundancy through multiple independent sensors.  The pressure sensor is located at
the bottom of one of the thermistor arrays, and is used for all tests. If possible, at least two of the
other height sensing devices will be located in  and alongside the tubing of the second thermistor
array, or will be inserted in the tank through a separate opening.
     Under ideal conditions, all of the temperature and level measuring sensors will be
deployed in the test tank.  In order to achieve this, the tank must have at least three access points
having diameters equal to or greater than 4  inches. With this arrangement, it should be possible
to install the instrumentation and provide unobstructed access for making the required volume
displacements.  An optimum configuration  would place these ports equally along the horizontal
axis of the tank.  However, because the test tanks will be located at existing commercial installa-
tions, it will be necessary to examine the installed configuration  prior to testing to ensure that the
minimum experimental requirements  can be met.
     To measure product temperature, two arrays of thermistors will be used.  The thermistors,
attached to a stainless steel tube, are spaced at  intervals of 20 cm (8 in.) along the vertical axis of
the tank. To measure air temperature, a single thermistor will be placed in a shaded area outside
the tank. Each array, placed in the tank through a fill hole or other appropriate  connection, is
equipped with a pivoting "arm" that can be  lowered to a horizontal position after the array has
been positioned. The pivot arm provides for the measurement of horizontal thermal gradients
between the tank's centerline and its walls.  A  tank with an inner diameter of 11.5 ft requires 17
vertically  oriented thermistors and a 5.5-ft pivoting arm containing 9 thermistors.
     The data quality objective for the instrumentation to be used in the field tests is based upon
the EPA performance standard for tank tightness tests [1] and is  more fully described 
-------
defined as one standard deviation of uncertainty in each measurement; (3) the resolution of a
sensor is determined from the A/D converter in each instrument and is less than or equal to the
required precision for all the sensors except the acoustic sensor.
Table 1. Specifications of the Measurement Instrumentation
          Sensor
Range     Resolution   Accuracy   Precision   Duration
Temperature
Product
Air
Product Levet
Absolute Pressure
Electromechanical
Fiber Optic
Acoustic

5 to 25°C
5 to 25°C

0 to 3.7 m
1 cm
1 cm
Oto2m

0.0008°C
o.ooo8°e

0.0012 cm
0.00025 cm
0.00010 cm
0.00381 cm

0.05°C
0.5°C

5.0cm
n/a
n/a
n/a

o.oorc
o.rc

5.0cm
0.00064 cm
0.00020 cm
0.00064 cm

Ito2h
< 1 min

< 1 min
2h
Ih
8h
                                  EXPERIMENTS
     The conduct of the field tests is constrained by both the experimental requirements and the
physical and operational limitations imposed by the tank installation. Since the experiments will
be conducted at commercial tank installations (rather than at the UST Test Apparatus), careful
attention has been paid to planning the tests during periods of tank inactivity. In general, the test
sequences have been organized to maximize the amount of data obtained while simultaneously
minimizing the amount of product transfers necessary to establish a particular test condition.
During the course of the tests, the tank is filled incrementally from an initial 30% of capacity all
the way to overfilled, which is the state of the tank during the final testing at the end of the four-
day period. The test sequence is designed to assess the temperature and deformation effects nec-
essary to address the objectives of this study.
     The tests will start on a Thursday morning and end on a Monday morning. It is assumed
that dispensing operations will occur during the business week between 0800 and 2000 and that
no dispensing operations will be conducted over the weekend. The instrumentation will be
placed in the tank during normal operations on Thursday to check out the measurement systems
and to collect some background temperature data.  The tank should be below 50% of its capacity.
A delivery of product sufficient to raise the level to about 75% of tank height is scheduled to
coincide with the close of tank operations. Temperature and level measurements will be made
over the next 12 to 24 h to examine the effects of the delivery. Measurements will terminate
when the dispensing operations begin on Friday morning.  A second delivery of product, which
should nearly fill the tank, is scheduled for the close of business on Friday night. Temperature
and level measurements will be made over the next 24 h to examine the effects of this second
delivery. On Saturday morning, the tank will be topped, raising the level of the product into the
fill tube. The temperature of the added product will be known to differ from that of the product
already  in the tank.  Temperature and level measurements will be made over the next 6 h to study
the temperature effects produced by topping.  Once the deformation has subsided, which may
require an additional 6 to  12 h, a series of experiments will be conducted to estimate the time and
elasticity constants that describe the deformation of the tank.
     The time constant of the deformation of the tank and backfill/soil system will be deter-
mined as follows. A cylindrical bar of known volume will be inserted into the fill tube to pro-
duce an instantaneous level change in the product inside the fill tube. The product in the fill tube
will be releveled at intervals of 5 to 15 min, and the volume change requked to maintain a
constant level in the tank will be recorded. This process will continue for 6 to 12 h.  On Sunday,
the bar will be removed for 6 h and subsequently inserted again for another 6 h. From these level

                                          196

-------
changes the effective time constant of measurements in the fill tube can be determined. These
data can be used to estimate the elasticity constant of the tank/backfill/soil system if the tempera-
ture changes in the tank are small. The volume data obtained from measurements made during
the releveling will be used to determine the time constant of the tank. The magnitude of the
deformation is dependent on the initial  level of product, that is, before the bar has been inserted
or removed from the tank.  Temperature data will be collected to compensate for thermally
induced changes during these tests for structural deformation.

                       INSTRUMENTATION TRADEOFFS
     To meet the EPA performance standard for a tank tightness test, the instrumentation used
to measure temperature and level measurements in large tanks should not inhibit the detection of
a leak of 380 ml/h (0.1 gal/h) and must allow for a PD of 0.95  and a PFA of 0.05  This requires
that the instrumentation noise be sufficiently small that a one-standard-deviation uncertainty in
the measurement of the rate of change of volume is less than 115 ml/h (0.03 gal/h). This
115-ml/h estimate assumes that the instrumentation noise is normally distributed with a zero
mean and that the signal adds linearly with the noise. For instrumentation noise, the assumption
of normality is generally justified. This estimate also assumes that the resolution of the sampled
data is smaller than the standard deviation of the noise.  Thus, the instrumentation or system
noise estimated in terms of volume can be characterized by its standard deviation; the precision
of the instrument is defined by the standard deviation of the instrument noise. If the resolution is
greater than the inherent precision of the instrument, it  is more difficult to characterize the per-
formance of the instrument. To  satisfy this data quality objective, the height and temperature
sensors must have resolution and precision adequate to sense changes of 115 ml/h (0.03 gal/h).
     The sensors, thus, must be able to measure changes to within a one-standard-deviation
uncertainty of 115 ml/h. To estimate the resolution and precision of these sensors, it is necessary
to specify the duration of the test. For tanks less than 37,860 L (10,000 gal) iri capacity (for
which  the EPA performance standard was developed), the duration of a test is usually 1 to 2 h.
This is the minimum amount of time required to make a reasonable estimate of the rate of change
of volume in the tank from level and temperature measurements. Larger tanks, which are usually
only partially filled during  testing, may require tests longer than 1 to 2 h, because the sensors
required to measure level and temperature changes approach the technological limits of non-
laboratory and affordable equipment. At present, most systems that are used to conduct a test on
a partially filled tank are automatic tank gauging systems (ATGSs).  The duration of tests
conducted with ATGSs is typically 4 to 8 h. The precision of the level sensors is typically
between 0.00025 and 0.0025 cm (0.00010 and 0.001 in.). For the current set of experiments, the
duration of the measurements will be based on the resolution and precision of the sensors being
used.

MEASUREMENT OF SMALL LEVEL CHANGES
     Tests conducted in half-filled tanks require the highest degree of precision because the
height-to-volume ratio is lowest at this  product level. A volume-change uncertainty of 115 nnl/h
in a half-filled 181,710 L (48,000 gal),  3.5-m (11,5-ft) underground tank will result in an uncer-
tainty of 0.000174 cm/h (0.000069 in./h) in the corresponding product-level changes. The sensor
precision required to measure a level change of 0.000174 cm/h can be estimated from
                             Sns2 = nS2/(n2:t2-(Zt)2)                                  (1)
where Sm = standard deviation of the slope of the least-squares line in cm/h, S = standard
deviation or precision of the sensor in cm, n = number of independent points (i.e., degrees of
freedom), and t = time in hours.


                                          19?

-------
      Eq, (1) describes the one-standard-deviation error in the slope of a least-squares line fit to a
number of independent points taken over a period of time in terms of the standard deviation of
the ordinate (le., sensor precision).  Eq. (1) can be used to estimate the minimum duration of the
measurement required to obtain the desired Sra, given a sensor with a precision of S. Estimates
made with Eq, (1) are valid providing that (1) the standard deviation of the sensor is greater than
the resolution, and (2) each sample is independent. The standard deviation represents an esti-
mate of the system noise or precision of the sensor.  Eq. (1) can be used to estimate the minimum
duration of a measurement made with a sensor whose precision is known, or it can be  used to
estimate the precision of a sensor given that the duration of the measurement period is specified.
In general, the number of independent samples, and therefore the number of degrees of freedom,
will be significantly less than the number of points acquired by the sensor, because ambient level
and temperature data are highly correlated for periods less than 5 to 15 min.
      Table 2 presents the level sensor precision required to obtain a specified Sm for measure-
ment periods of 1, 2,4, and 8 h. The minimum duration required for sensors with a precision, S,
to obtain a Sm of 0.000174 cm/h can be estimated from Table 2 by finding the precision of the
sensors in the last column of the table and reading the duration from the first column.
Table 2. Estimate of the Precision, S, of the Sensor Estimated for Sm = 0.000174 cm/h for Different Measurement
Periods (A level change of 0.000174 cm/h corresponds to a volume change of 115 ml/h in a half-filled, 181,710-L
(48,000-gal) tank.)
Duration of
Measurement
(h)
1
2
4
8
Number of
Independent
Points (n)
13
25
49
97
Standard Deviation
of Rate of Change
of Sensor (Sj
(cm/h)
0.000174
0.000174
0.000174
0.000174
Standard Deviation
of Sensor (S)
(cm)
0.000196
0.000523
0.001435
0.004000
      For these calculations, it was assumed that the data were sampled once every 5 min. Thus,
it was assumed that there are only 12 degrees of freedom (i.e., 12 independent points) each hour.
If a level sensor with a precision of 0.0025 cm (0.001 in.) is used, the one-standard deviation
uncertainty in the measured level changes, Smf is 0.00226 cm/h (0.00089 in./h) for a 1-h test with
12 degrees of freedom; this is equivalent to a volume change of 1,494 ml/h (0.395 gal/h) in a
half-filled 181,710-L (48,000-gal) tank.  Previous measurements of level and temperature in an
underground storage tank suggested that the number of degrees of freedom might be as low as 3
or 4 each hour. As the number of degrees of freedom decreases, the duration of the measurement
must increase or more stringent requirements must be placed on the precision of the sensor.
      If the resolution of the level sensor is greater than the inherent precision of the measure-
ment system, and the level changes are less than the resolution of the sensor, then the smallest
level change that can be measured with a two-point estimate is a resolution cell divided by the
measurement period. If the level changes are larger than a resolution cell, however, level
changes can be estimated to better than a resolution cell by fitting a least-squares line to the data.
The accuracy of estimating the rate of change depends on the number of resolution cells exce-
eded and the duration of the measurement. A better estimate can be made if we measure the time
at which the level is located at intervals of one-half a resolution cell and fit a least-squares line to
the data. The number of degrees of freedom is equal to the number of resolution cells minus 1.
A more detailed discussion of how to estimate the performance of a system limited by resolution
is given in [3].

                                           198

-------
MEASUREMENT OF SMALL TEMPERATURE CHANGES
      We can estimate, for both half-filled and completely filled tanks, the precision requirement
of the product temperature measurement system assuming an uncertainty of 115 ml/h in the leak
rate, a value of 0.00125/°C for the coefficient of thermal expansion, and a 181,710-L
(48,000-gal), 3.5-m (11.5-ft) tank. A 115-ml/h volume change corresponds to a standard devi-
ation of 0.0010 and a 0.00055°C/h rate of change of temperature in a half-filled tank and a full
tank, respectively. Eq. (1) was used to estimate the thermistor precision required to obtain the
specified standard deviation of the rate of change of temperature, Sm, in both a half-filled tank
and a completely filled tank as a function of measurement period. This calculation also assumes
that the number of independent degrees of freedom was 12 per hour. The results are presented in
Tables 3 and 4.  The minimum measurement period required to obtain a precision of 0.001 °C is
less than 1 h for the half-filled tank and approximately  1.5 h for the completely filled tank. If the
test duration were 2 h or longer, the precision of the thermistors could be higher than 0.001°C.
Table 3. Precision, S, of the Sensor Estimated for Sra = 0.0005°C/h for Different Measurement Periods in a Half-
filled 181,710-L (48,000-gal) Underground Storage Tank

    Duration of          Number of       Standard Deviation      Standard Deviation
   Measurement        Independent       of Rate  of Change         of Sensor (S)
                          Points (n)         of Sensor (S.J
        (h)	CC/h)	(°C)	

         1                   13                 0.0010                   0.0011
         2                   25                 0.0010                   0.0030
         4                   49                 0.0010                   0.0084
	8	97	0.0010	0.0233	

Table 4. Precision, S, of the Sensor Estimated for Sm = 0.0005°C/h for Different Measurement Periods in a Full
181,710-L (48,000-gal) Underground Storage Tank
Duration of
Measurement
(h)
1
2
4
8
Number of
Independent
Points (n)
13
25
49
97
Standard Deviation
of Rate of Change
of Sensor (Sm)
(°C/h)
0.0005
0.0005
0.0005
0.0005
Standard Deviation
of Sensor (S)
<°C)
0.0006
0.0015
0.0042
0.0116
                                     SUMMARY
      Leak detection performance standards currently do not address the impact of increased
tank size on the ability of tank tightness tests to satisfactorily meet regulatory requirements. The
currently planned experimental program has been devised to expand the understanding of large
underground storage tank behavior as it impacts the performance of tank tightness tests.  Upon
completion of the field tests, the suitability of the currently recommended test protocol features
(which were developed during previous evaluations of tank tightness test methods) can be
assessed, and, if necessary, modified accordingly.
      This paper has presented an analysis of the requirements necessary for temperature and
level sensors to detect leaks as small as 380 ml/h (0.1 gal/h) with a PD oifO.95 and a PFA of 0.05.
The analysis shows the tradeoff in test duration as a function of sample interval for independent
samples, resolution, and precision.
                                          189

-------
                                 REFERENCES
1,   U.S. Environmental Protection Agency.  Underground Storage Tanks; Technical Require-
     ments and State Program Approval; Final Rules. Federal Register, 40 CFR Parts 280 and
     281, Vol. 53, No. 185, September 23,1988.
2.   U.S. Environmental Protection Agency.  Evaluation of Volumetric Leak Detection Meth-
     ods/or Underground Fuel Storage Tanks.  Risk Reduction Engineering Laboratory, U.S.
     Environmental Protection Agency, EPA Contract No. 68-03-3409, December 1988.
3.   J. W. Starr and J. W. Maresca, Jr. Quality Assurance Project Plan: Evaluation of Leak
     Detection Methods for Large Underground Storage Tanks.  EPA Contract No.
     68-03-68-03-3409, Vista Research, Inc., Mountain View, California, January 1990.
                                        200

-------
                 SOURCE OF CONTAMINATION ASSOCIATED WITH
                   CLOSURE OF UNDERGROUND STORAGE TANKS
           by:  Warren J. Lyman         Anthony Tafuri
                Andrea E. Sewall        U.S. EPA/RREL/RCB
                Camp Dresser & McKee    Uoodbridge Ave.
                One Center Plaza        Edison, NJ
                Boston, MA  02108
                                 ABSTRACT
    EPA is currently evaluating several technical and regulatory aspects
of underground storage tank (UST) closures.  The purpose of this project
is to evaluate the effectiveness of various tank cleaning procedures used
at closure.  A common cleaning procedure, for example, involves removing
pumpable residuals with a vacuum truck, manual entry into the tank to
scrape residuals off the sides and bottoms (gums and sludge), and then
rinsing with water.

    During the course of this project, with the help of a tank cleaning
company, tank cleaning operations are being observed for several gasoline
and diesel USTs.  Samples of the original residuals (present before
cleaning), the residuals left after cleaning (if any), and aqueous
rinseates are being collected and analyzed for both hazardous
characteristics (e.g., flash point and corrosivity), chemical composition
(including volatile organics, total petroleum hydrocarbons, heavy metals,
and oil and grease), and other characteristics (e.g., pH, BOD and TOC of
the aqueous phase).  Vapor concentrations are also being measured at
various stages in the cleaning process.  Estimates of residuals volumes
are being made both visually and with dipsticks, and in some tanks,  phases
initially present in the UST are detected with a conductivity probe.
                                  201

-------
                                INTRODUCTION
    A  previous  study  showed  that  there  are  significant amounts  of
residuals  left  in  underground  storage tanks (USTs)  at  the  time  closure  is
initiated.   Although 50  to  100 gallons was typical, it was not unusual to
find tanks  at closure with several  hundreds or  even thousands of gallons
of residual liquids.   The handling  and  removal  of  these residuals during
closure -  and the  resulting  tank  cleaning - are being  carried out by a
wide variety of procedures often  dictated more  by  preferences of local
officials  or the selected contractor  than by objective guidelines.  The
cleaning procedures typically  generate  a significant volume of  aqueous
rinseate that also presents  disposal  problems.

    The earlier review of closure activities for USTs  concluded that the
tank cleaning methods currently in  use  appear to be able to satisfactorily
clean  most  gasoline and light  oil tanks.    This assessment,  however, was
made only on the basis of interviews and limited observations.  No data
existed which indicated just how  clean  tanks do get.   Furthermore, the
review found that  there was no generally accepted method of tank cleaning,
nor were there  any generally accepted criteria  for  assessing tank
cleanliness.  Therefore,  it was the purpose of  this study  to monitor
selected tank cleaning methods, to  make quantitative measurements of the
amounts of  residuals  left in USTs before and after  cleaning, and to
characterize the nature of these  residuals  and  any  rinseates generated
during the  cleaning process.

    This field  monitoring, sampling, and analysis program  is currently  in
progress.   Completion of  the field  activities has been hindered, in part,
by the early cold  winter  in  the Boston  area where the work is being
conducted.   Essentially all UST closure activity ceased during  the period
from November to February.  As of mid-February, only three of an
anticipated  twelve USTs have been sampled during closure.  We expect
completion  of the  field activities  in March and April, and the  preparation
of a final  report  shortly thereafter.   As a result, this paper  only
presents an  overview  of the field program and the analyses being
conducted.

                       DESCRIPTION  OF FIELD PROGRAM
COOPERATION WITH A TANK CLEANING/REMOVAL COMPANY

    It was deemed both necessary and desirable to carry out this field
program in cooperation with a company that conducted UST cleaning and
removals as part of its business.  Camp Dresser & McKee (COM) selected
Jet-Line Services, Inc. (Stoughton, MA), a company that offers a range of
                                  202

-------
environmental and remediation services in  the Boston area.  Jet-Line was
also one of the companies that provided significant background information
in the initial survey phase of this program.

    The agreement with Jet-Line calls for  them  to identify UST
closures-of-opportunity, i.e., closure jobs  they have obtained as part of
their normal business that can meet the objectives of the field study (see
below).  If the UST and proposed cleaning  technique are considered to be
of interest to the study, and permission is  obtained from the site
owner/operator, then monitoring and sampling are carried out during the
normal course of the closure.

    In order to minimize health and safety concerns, and to avoid issues
of liability, all measurement and sample collection activities - including
those done by a worker inside the UST - are  conducted by trained Jet-Line
personnel under the general direction of an  on-site COM monitor.
Monitoring and sampling activities also follow  the requirements of a
Quality Assurance Program Plan (QAPP) prepared  by COM.

    The agreement with Jet-Line does not require them to modify their
cleaning techniques to arbitrarily provide this study with a variety of
cleaning methods.  It is expected however, that some variations in
cleaning methods will occur as a result of site-specific conditions
including the nature and amount of residuals found within the tank.

TANK SELECTION GOALS

    The goal of this study is to monitor and sample twelve tanks, as
follows:

       Five (5) leaded gasoline;
       Five (5) diesel fuel (or No. 2 home heating oil); and
       Two (2) waste oil or other petroleum  product.

    Furthermore, we hope to have an opportunity to include one or more
fiberglass reinforced plastic (FRP) tanks along with the steel tanks that
constitute most of the USTs being closed at  present.

    The final mix of tanks included in the study will necessarily reflect
the closures available to Jet-Line within the timeframe imposed on them by
COM.

INFORMATION TO BE COLLECTED

    As noted above,  the principal objectives of this study are to evaluate
the cleaning effectiveness of the closure activities and to generally
characterize the residuals found,  and generated, during closure.   To
support these objectives,  the following information will be collected for
each site included in the study:
                                  20J

-------
    o  Estimates of volumes  of  tank  residuals and secondary wastes;

    o  Hazardous characteristics and chemical composition of  the residuals
       and secondary wastes;

    o  Details of  the cleaning  method used; and

    o  General background information on  the UST and the site that may
       relate to the nature  of  the residuals found in the UST.

Volume Estimates

    During each UST sampling effort, estimates of the volumes of residuals
in the UST before  and after  cleaning (and of secondary waste generated)
will be estimated  from dip stick measurements, simple volume measurements
(e.g., height of waste in 55-gallon drum), pumping rates and durations, or
visual estimates.  Dip stick measurements in cylindrical tanks can be
converted to fluid volumes by standard trigonometric functions.

    Where possible, different volume estimates will be obtained for the
following:

    o  Liquid organic phase:  before and after cleaning

    o  Aqueous phase:  before and after cleaning

    o  Rinseate:  amount used to clean tank

Hazardous Characteristics and Chemical Composition

    Various UST residuals will  be sampled and analyzed for both hazardous
characteristics (e.g., ignitibility and corrosivity) and chemical
composition.  Figure 1 outlines the specific analyses expected for samples
from various types of USTs.

    The analyses specifically include the following:

       Characteristics                  Composition

       Ignitibility                     Volatile Organic Analysis (VGA)
       Corrosivity                      Total Petroleum Hydrocarbons (TPH)
       Reactivity                       Oil and Grease
       Toxicity Characteristic          Biochemical Oxygen Demand (BOD)
         Leaching Procedure  (TCLP)      Total Organic Carbon (TOC)
                                        PH
                                        RCRA Metals (As,  Ba, Cd,  Cr,  Pb,
                                          Hg,  Ag,  and Se)
                                        Polychlorinated Biphenyls (PCBs)
                                  204

-------
    In addition, measurements of vapor concentrations inside the UST will
be made at various stages in the cleaning/closure procedure.

    It is anticipated that modifications to this sampling and analysis
program will be made on a case-by-case basis after considering
site-specific conditions.

Details on Cleaning Method

    The normal cleaning method used by Jet-Line includes the following
basic steps:

    1)  Removal of liquids by suction into a vacuum truck (may leave 1 to
        2 inches of liquid in tank)

    2)  Manual entry of tank to scrape walls and bottom (material is
        removed by suction line and/or shovel); and

    3)  Aqueous rinse of the inside of tank (rinseate is removed by
        suction line)

    Details of the actual procedures used at each tank vill be recorded.
Visual estimates of the cleaning effectiveness will also be recorded.

General Background Information

    Where available, the following information will be recorded for each
UST sampled?

    o  Tank contents (e.g., gasoline, no. 2 fuel oil, waste oil)

    o  Tank dimensions, volume, material of construction, and installation
       details

    o  Tank age and condition upon removal

    o  Tank depth in ground, and depth to water table

USTs EVALUATED TO DATE

    Summary information on the three USTs evaluated to date is provided in
Table 1.  Analysis of samples collected have not yet been reviewed
according to the project QAPP and are thus not presented at this time.
                                   205

-------
               TABLE  1.  INF01MATION  ON INITIAL USTs  EVALUATED

RESIDUALS PRESENT
Site
No.
1
2
3
BEFORE
Organic*
4400
800
94
CLEANING (gal.)
Aqueous
17
None
None
Sludge
Trace
1
70
RESIDUALS PRESENT
AFTER CLEANING
Organic Liquid
1
None
None
(gal.)
Sludge
None
1
None
RINSEATE
GENERATED
(gal.)
25
None**
50

 * All contained No.  2 fuel  oil
** Sawdust  used instead of water  to  clean  inside of  tank

REFERENCES

1. Lyman, tf.J. and  Tafuri, A.,  "Considerations of Underground Storage
   Tanks at Closure," Proc.  15th Annual EPA  Research Symposium,
   Cincinnati, OH,  April 10-12, 1989.
  OASOLIHS TASKS

Fi»ld Ob»«rv«tlon«
and
                                                FU1L OIL OR DIESEL TANK

                                                  _ ri»ld Observations
                                                        Measur«m»ntB
    BEFORE        AFTER
   CLEANING       CLEANING
(Org.  Liquid}   (Aq. Liquid)
	  Pha»»         Phaao
Ignitibility
VOA
RCRJk M«tal»
VOA
RCHA H»talB
|>H
BOD
          Oil I Gr««s«
                                     BEFORE
                                     CLEANING
                            {Org. Liquid)
                                  Phase
                                Ignitibility
                                VOA
                                RCRA'M.tttls
  (Sludga)

Ignitibility
TCLP
VOA
TPH
RCRA Metftls
                                                                      AFTER
                                                                    CLEANING
                                                  (Sludge)    (Aq. Rl
                                                                             nseate)
                                                          Ignitibility
                                                          TCLP
                                                          VOA
                                                          TPH
                                                          RCRA M*tals
                                                                          VOA
                                                                          5PPH
                                                                          RCRA M«t*l«
                                                                          pH
                                                                          BOD
                                                                          TOC
                                                                          Oil t Grease
                              WASTE OIL OR "OTHER" TANK

                               - rield Obsarvations
                                 and Mo«sur«m«ntB

I BEFORE
1 CLEANING
(Org. Liq.) (Sludge)
Ignitibility Ignitibility
Corrosivity Corrosivity
Reactivity Raactivity
VOA TCLP
RCRA M*tals VOA
TPK TPH
PCBS FCBa
RCRA M»t«ls
1 AFTER
1 CLEANING
(siudge) (An. R-fnseate)
Ignitibility VOA
Corrosivity RCRA Metals
Reactivity TPH
TCLP pH
VOA BOD
TPK TOC
RCRA Metals Oil I Groasa
 Pigur* 1.  Flow Chart of Sa»pl»B to bo Collected and the Parameters to b* Analyzed
                                      206

-------
                 REMOVAL OF SOLUBLE TOXIC METALS FROM WATER
       by:  L.P. Buckley, S. Vijayan, G.J. McConeghy and S.R. Waves
            Atomic Energy of Canada Limited, Chalk River Nuclear
            Laboratories, Chalk River, Ontario, Canada KOJ 1JO

      and:  J.F. Martin, Risk Reduction Engineering Laboratory,
            United States Environmental Protective Agency,
            26 V. Martin Luther King Drive, Cincinnati, Ohio 45268
                                  ABSTRACT
    The removal of selected, soluble toxic metals from aqueous solutions has
been accomplished using a combination of chemical treatment and ultrafiltra-
tion.  The process has been evaluated at the bench-scale and is undergoing
pilot-scale testing.  Removal efficiencies in excess of 95-99% have been
realized.

    The test program at the bench-scale investigated the limitations and
established the optimum range of operating parameters for the process, while
the tests conducted with the pilot-scale process equipment are providing
information on longer-term process efficiencies, effective processing rates,
and fouling potential of the membranes.  ¥ith the typically found average
concentrations of the toxic metals in groundwaters at Superfund sites used as
the feed solution, the process has decreased levels up to 100-fold or more.

    Experiments were also conducted with concentrated solutions to determine
their release from silica-based matrices.  The solidified wastes were sub-
jected to EP Toxiclty test procedures and met the criteria successfully.  The
final phase of the program Involving a field demonstration at a uranium tail-
ings site will be outlined.


         This paper has  been  reviewed  in  accordance with the
         U.S. Environmental Protection Agency's peer and admini-
         stration review policies  and approved for presentation
         and publication.
                                      207

-------
INTRODUCTION

    Conventional treatment to remove contaminants from waste disposal sites
or to prevent the spread of contaminant plumes is usually of the brute force
mode, vhere large quantities of contaminated soils are excavated and treated
by incineration, solidification, or other related technologies to remove,
treat or detoxify the contaminants.  In general, these large-scale technolo-
gies need to incorporate bulk material handling capabilities to accomplish
the desired waste management.

    An alternative to such material handling technologies is to remove the
metal contaminants from the groundwater, through treatment to extract con-
taminants, and perhaps recycle the water to the contaminated zone to recover
more contaminant adsorbed to the soils in contact with the groundwater.  The
technology proposed is a combination of chemical treatment and ultrafiltra-
tion to enhance the recovery of toxic metals from the water.  The technology
has been evaluated at the bench-scale level for the removal of radioactive
cations from waste water streams (1,2).  The technology appears suitable for
the removal of toxic metal cations from contaminated groundwater surrounding
industrial waste disposal sites and for the treatment of industrial efflu-
ents.

    In the past year, chemical treatment in combination with ultrafiltration
has been evaluated through funds allocated by the Emerging Technology Pro-
gram of the United States Environmental Protection Agency (USIPA).  The
Emerging Technology Program, part of the Superfund Innovative Technology
Evaluation (SITE), provides guidance and funding for new technologies which
may become tools in the cleanup of environmental pollution.  The two-year
evaluation of the combined chemical treatment and ultrafiltration technique
began by conducting experiments in laboratory-scale equipment, to verify the
process capabilities for the removal of hazardous toxic metals which may be
present in groundwater or industrial wastes.

    Thus far, only four metals have been evaluated with the process.  They
are cadmium, lead, mercury, and arsenic; indicative of metals present in
groundwater associated with hazardous waste disposal sites.  The program to
establish the viability of the process has progressed smoothly, results
achieved from the program will be discussed, and proposed efforts to con-
clude the second year of the contract, which takes the technology through
pilot-scale testing and field trials, are outlined.

TECHNOLOGY DESCRIPTION

    The emerging technology reported here evolved through efforts by Atomic
Energy of Canada Limited (AECL) to remove dilute concentrations of radio-
active species from waste waters.

    The process involves contacting the waste water with water-soluble mac-
romolecular compounds added to the waste solution to form complexes with the
soluble heavy metal ions.  A high molecular weight polymer, generally a
commercially available polyelectrolyte, Is added to the waste solution to
form the selective complexes.  The polyelectrolyte quantities needed to
achieve the separation of metal ions are generally in the parts-per-million
range•
                                      208

-------
    The solution is then processed through an ultrafiltration membrane system
which retains the macromolecular complexes (retentate), vhile allowing uncom-
plexed ions such as Na+, K+, Ca++, Cl" S04=, N03~» etc.,  to pass through the
membrane with the filtered water (permeate).  The filtered water can be re-
cycled, or discharged depending upon the removal efficiency desired. A
removal efficiency approaching 100% can be achieved for metal ions which have
been complexed.

    The technology as described can be applied to separate toxic heavy metal
ions such as Cd, Cr, Hg, Ni, Cu, Zn, As and Pb from leaehates generated in
Superfund sites (3).  Other inorganic and organic materials, if present as
suspended and colloidal solids, can also be removed.  The ultrafiltration
membrane can be chosen with a sharp molecular weight cut-off, to further
enhance separation.  The toxic metals can be concentrated in a small residual
volume which should be amenable to conventional solidification treatment.

BENCH-SCALE TESTS

Separation Studies

    The first year of the test program vas divided into five tasks: first, to
evaluate the major variables using a factorial design having five variables,
each at two levels, with fixed levels of metal contaminants; second, to study
the extent of fouling and the cleaning techniques needed to recover the flux
rate of the membranes; third, to evaluate the system response to changes in
toxic metal concentrations} fourth, to perform a second evaluation of fouling
and the cleaning techniques necessary to operate a continuous system? and
fifth, to conduct a series of tests at three levels for each dominant vari-
able, to optimize the technique for the removal of toxic metals.

    The bench-scale tests were performed using Amicon ultrafiltration stirred
cells.  Cells with volume capacity of 50 mL or 200 mL, and equipped with
rated membranes having molecular weight cut-off of 10 000, were used.  The
metal ions were complexed with polyelectrolytes, i.e., polyethylenimine, with
a molecular weight of 50 000 and Gantrez AN 119, a polymethyl vinyl ether/
maleic anhydride copolymer, of similar molecular weight.  The operating con-
ditions for the test series are summarized in Table 1.

	TABLE 1.  BENCH-SCALE EXPERIMENTAL CONDITIONS	

          Variable                Test Conditions      Test Conditions

     Solution pH                alkaline              acidic
     Membrane Type              polysulfone           cellulose acetate

     Polyelectrolyte            polyethylenimine      Gantrez
     Polyelectrolyte            10 times the

       Concentration, mg/kg       metal concentration    0
     Toluene, mg/kg             1000                     0
                                      209

-------
   Cadmium removal efficiencies indicate that: the separation technique is
more effective at elevated pH values; removal efficiency of about 99% can be
obtained; and, it was difficult to distinguish the minor effects of other
operating parameters such as membrane type, polyelectrolyte type and the
presence of organics (toluene) on the removal efficiencies.

   Removal efficiencies of 90% or higher vere obtained for mercury.  In con-
trast with cadmium, the removal of mercury is less affected by solution pH.
Polyethylenimine appears to be a better complexing agent than Gantrez.  Fac-
tors such as membrane type and toluene had no apparent effect.

   Lead is a special ease, because its starting concentration in the feed
solution was two orders of magnitude higher than the other heavy metal ions
examined in this study.  Like cadmium, it is more effectively removed from
solution at elevated pH values.  However, the removal efficiencies are
equally good when no polyelectrolyte was added to the feed solution, an indi-
cation that most lead is removed by precipitation as hydroxide.

   Relatively poor separation efficiencies of <35% for arsenic were achieved
with and without the addition of either polyelectrolyte.  The reason is that
a major fraction of arsenic is present in solution as an anionic species,
i.e., As043", while most other metal ions are in the form of cations.  The
results obtained for arsenic Indicate the present limitation of the proposed
complexing polymers in the chemical treatment and ultrafiltration technique.

   Polysulfone membranes experienced severe fouling, with permeation rates
declining more than 70% when solutions containing polyethylenlmine were
treated.  Gantrez caused less fouling to the polysulfone membranes, but the
permeation reduction was also quite severe at high Gantrez concentrations.

   Cellulose acetate membranes, in contrast, are essentially free from foul-
ing caused by either electrolyte; only some slight fouling with permeation
rate reductions of about 20-401 was experienced when either polyelectrolyte
was used at a relatively high concentration in the lead test series.  How-
ever, even with the limited fouling observed, the permeation rate of the
cellulose acetate membranes is still five to ten times lower than achieved
with the fouled polysulphone membranes.

   Operating parameters such as pH and the presence/absence of toluene did
not appear to have any impact on the fouling behaviours of the ultrafiltra-
tion membranes in these bench-scale tests.

   Since the polysulfone membranes showed significant fouling in the first
series of tests performed, only this type of membrane was used in a number of
repeated tests to produce similarly fouled membranes, to evaluate the effec-
tiveness of different chemical cleaning solutions. Four chemical cleaning
solutions were compared:  lergazyrae detergent; sodium hydroxide and hydrogen
peroxide$ methanol acidified with hydrochloric acid; and, deionized water as
the control cleaning solution.

   In each test, a membrane was subjected to a pure-water test to obtain the
initial permeation rate.  Ultrafiltration of the test solution containing

                                      210

-------
metal polymer complexes followed, to obtain a water recovery of 90%.  Then,
the following steps were taken:  rinsing with deionized water, cleaning with
one of the chemical solutions, and, finally, re-rinsing with deionized water.
At the end, a pure-water test established the post-test permeation rate.

   The results indicate:

   For a polysulfone membrane fouled by polyethylenimine, acidified methanol
is the best cleaning solution to restore the permeation rate.  Tergazyme
detergent caused further severe fouling to the polysulfone membrane which had
already been fouled by polyethylenimine.  The detergent may contain an ani-
onic surfactant which reacts with the cationic polyelectrolyte, to form an
insoluble product which plugs the membrane pores.  Sodium hydroxide-hydrogen
peroxide did not remove the foulant better than deionized water.

   For a polysulfone membrane fouled by Gantrez, cleaning with deionized
water was able to restore most of the permeation rate.  Overall, it is easier
to remove this polyelectrolyte than polyethylenimine from the fouled polysul-
fone membrane.  Acidified methanol was less effective, while the other two
cleaning solutions are slightly more effective than deionized water.

   The repeated tests performed for the membrane cleaning tests provided an
opportunity to check the reproducibility of the performance of the chemical
treatment ultrafiltration combination technique.  The results indicated that
relatively consistent results such as separation efficiencies, permeation
rates and permeation reduction can be reproduced for tests carried out under
the same operating conditions with different membranes.

   The performance of this separation technique may be affected by the vari-
ability of the feed concentrations of the metal ions.  Varying the dissolved
metal concentration in a feed solution represents a realistic situation where
the separation process is applied using a predetermined set of operating con-
ditions.  Metal separation efficiencies, membrane fouling behaviour and the
effect of chemical cleaning were studied.  Polysulfone membranes were chosen
to allow the study of membrane fouling and cleaning.  An alkaline solution pH
value was selected to get high removal efficiencies for the heavy metals.  No
toluene was added to the feed solution, since the earlier tests did not indi-
cate any adverse effects of separation when toluene was present or absent.

   The test program produced optimum conditions for the dominant variables
and provided additional verification of the process to remove soluble metal
cations from solution.  The addition of excessive amounts of polyelectrolyte
did not enhance the separation observed from the first series of experiments.
Variation of the pH values improved the separation, but again, at increasing
alkalinity levels, the improvement in the separations observed were minimal.
Finally, further evaluation of the polyelectrolyte selection did not produce
large enough differences to dominate the selection for the pilot-scale tests.

Solidification and EP Toxicity Tests

   Next, a solidification study was performed on some of the concentrated
waste products generated from several of the above tasks, specifically to

                                     211

-------
establish the success In producing a quality product which will be strong
enough to survive the structural integrity procedure in the EP Toxicity pro-
cedure, and provide concentrations of the contaminants below the limits set
out in Table 1 of 40 CPR 261.24 (4).  Whether the concentrated waste had a
harmful effect on the matrix stability or its ability to retain heavy metals
had to be resolved.

   There were three matrices chosen, all cementitious-based to immobilize the
concentrated product which will arise from the volume reduction of the waste
solutions treated by the ultrafiltration membranes.  The three matrices were:
ordinary Portland Cement, Type I (OPC)| Aquaset, a proprietary product of
Fluid Tech, Inc.; and a mixture of OPC and blast furnace slag.  Products were
prepared with ten- and twenty-fold increases in metal and polyelectrolyte
concentrations, typically expected from recycle of the waste solutions
through an ultrafiltration system.  The samples were molded and then cured
for periods of seven days.  The samples were released from their molds, each
in turn placed in a compaction tester to establish the structural integrity
of the specimen, and then subjected to the EP Toxicity test procedures (4).

   Results are available at this time for only the OPC specimens.  All
formulated specimens successfully passed the structural stability test,
allowing the specimens to be leached intact.  The test matrix included both
electrolytes and metal concentrations at ten and twenty times the average
feed concentration.  The measurement of metal content of the leachates
indicated the levels were less than the detection limits available for the
analytical Instruments used, and well below the maximum concentration of
contaminants for characteristic of EP toxicity listed 40 CFR 261.24 (4).

PILOT-SCALE TESTS

   A schematic of the pilot-scale test facility Is presented in Figure 1.
The ultrafiltration membranes used were purchased from the Romicon and Amicon
corporations.  The modules (or cartridges) are constructed with hollow fibre
membranes, and made with polysulphone.  The Amicon module has twice the sur-
face area of the Romicon unit.

   A hollow fibre configuration for the ultrafiltration membrane system was
selected because of its many attractive features (5).  These include: high
surface area-to-volume ratio; low holdup volume; low energy consumption; and
relatively easy cleaning by "backflushing".  Minor disadvantages include
higher membrane replacement costs and susceptibility to plugging.

   The choice of polysulphone material for the membrane was based on the high
flows capable with this material, compared to that of cellulose acetate, in
spite of the increased fouling of the polysulphone membranes observed in the
first test phase.  Other reasons for the choice were the operating range
established from the first series of tests, which indicated alkaline condi-
tions were better, and to overcome the distinct possibility of poor lifetimes
expected of cellulose acetate at elevated pH levels.
                                     212

-------
   The pilot-scale  test  program was established to obtain engineering design
data to permit  the  building of the field test unit, and to permit some ini-
tial cost evaluations  to be undertaken.   Tests were of three general types to
evaluate the operation of single membrane modules: first, in a batch node, by
continuous recycle  of  both permeate and  retentate streams to the feed tank,
to establish the separation efficiency;  second, in a concentration mode,
enriching the feed  by  allowing the permeate to be discharged from the system
while recycling the retentate to the feed tank', and third, in a feed and
bleed mode, where fresh  feed is introduced to the system while simultaneously
removing concentrate and permeate, with  the bleed-to-permeate ratio fixed by
the specified recovery.
         PotyeJecfrolyte
            Addition  f
                                Recycle
                                Retentate
                    Feed Tank
1
                               Hollow Flber_
                                Cartridge
                              Throttling of
                                Valve  T
                              —»*	1	
                     Bleed
           Permeate   To Drain
   Pressure
   Gauge

   JL
   «|_L*4.
             T
       Additional
       Chemicals
Permeate
 Tank
     Backfluahing Pump
                    Recirculaflng Pump
                   Figure 1.   Schematic of pilot-scale unit
   The operating  conditions for the three types of pilot-scale tests are
given in Table 2.   Most  experiments covering batch and concentration modes of
operation are complete.   The majority of analytical results foi: the separa-
tion efficiencies  of  the heavy metals are pending.  Preliminary analysis of
some analytical results  for batch-mode tests have revealed removal efficien-
cies In the range  of  85-991 for lead, mercury and cadmium from a combined
alkaline feed stream.  The analysis of organic content of the permeate vater
indicated that most,  if  not all,  of the polyelectrolyte is being removed from
the feed stream by the membranes.  The feed flov rate was 4000 US gal/day and
permeate flux varied  from 30 to  75  US gal/ft2 day.  Unlike the; bench-scale
test resultsi the  initial pilot-scale data showed appreciable loss of metals.
It appears the loss is due to adhesion, adsorption, and free settling of
metal complexes and precipitates within the process equipment.  Methods to
account for and minimize losses,  and to recover the metals, art: being
studied.
                                      113

-------
                TABLE 2. PILOT-SCALE EXPERIMENTAL CONDITIONS
         Variable                   High Value           Low Value


    Solution pH                  alkaline              neutral
    Membrane Surface Area (ft2)     30                   15
    Operating Pressure (psig)       45                   25
    Feed Rate (US gal/day)        6500                 4000
    Polyelectrolyte            10 times the               0
     Concentration (ppm)    metal concentration
    Metal Concentration        10  times              0.5, 0.3, 60
     in feed (ppra)            the low value
[cadmium, mercury and leadj
    The hydraulic behaviour of the new membranes after a few days of usage
produced a steep drop in the permeation rate of about 30%, mainly due to
compaction of the membranes.  After this initial irreversible decrease, the
permeate rate decreased more slowly, eventually reaching a point where back-
flushing or chemical cleaning was necessary to recover the flux to reasonable
levels.  Once removal of the polyelectrolyte begins, additional permeate flux
loss is evident (see Figure 2).  The loss becomes more dramatic when alkaline
feed solution containing lead is concentrated.  A permeate flux loss of up to
75% from membrane compaction, physical ageing by extended operation, and by
plugging of pores by particles, was not uncommon in the pilot-scale tests.
These flow results confirm the initial bench-scale studies performed with
flat membranes.

DEMONSTRATION TESTS

    The demonstration tests are expected to take place at a site near the
town of Elliot Lake, Ontario, about 400 km from the Chalk River Nuclear
Laboratories site.  A substantial quantity of uranium mine tailings was
deposited within the property boundaries during operation of the Nordic Mine
by Rio Algom Limited from 1957 to 1968.  The waste, while mildly contaminated
with residual quantities of uranium, radium and thorium, also contains other
metal ions including iron, calcium, magnesium, aluminum, copper, cobalt,
zinc, lead, nickel, chromium and cadmium.  The tailings impoundment has been
investigated to determine the migration of acidic groundwater seepage.  While
the contaminated groundwater does not constitute a hazard because of its
remote location and limited migration of the heavy metals and radioactive
ions (6), it provides an opportunity to field test the developing technology.
                                     214

-------
ou
C
U,0<
5 »
C/>
D «5n
»«x "^
X 1
IL 40 *
03
| 30
C
| 20
m
^^^^^l B
L • — g 	 g 	 a
\ System Conditions; Inlet Pressure
\ Temperature =
\nono nSystem: (
Dfi
I- M. „
H
I , I , I

= 25 Psig
25±1*C
Id/water
= 7


           Romicon
           —a—
Amicon
    4

 Time (h)

Romicon with PE!
Amicon with PEI
             Figure 2.  Permeation  rates of hollow  fibre  modules

    The project involves building a mobile facility capable of treating about
30 L/min of contaminated groundwater.  After commissioning  the mobile unit,
it would be transported to and operated at the  tailings impoundment.   The
technology will be assessed  to determine its potential to decontaminate a
plume of toxic metal ions released  in  the groundwater from  the tailings
impoundment.

CONCLUSIONS

    Based on the tests performed thus  far, we can state:

    Among the four heavy metals examined, cadmium,  lead and mercury are
    effectively removed (99%), while arsenic removal was  less  effective.
    Separation of heavy metal ions  is  better in an  alkaline medium than in an
    acidic medium and membrane composition is not a major factor.
    The presence of toluene  In feed water does not  seem to  affect the metal
    separation efficiency nor the membrane fouling  rate,  although its effect
    on membrane integrity during extended operation must  not be ignored.
    Initial data from the pilot-scale  studies have  confirmed high separations
    are achievable and design data  are available for the  field test facility.

    In summary, the proposed technology could be applied  to remove selec-
tively toxic metal ions from the Superfund site leachates.  The residual
volume produced in this technique is significantly  smaller  than volumes
                                     211

-------
achieved In other conventional water treatment techniques, such as evapor-
ation, ion exchange and reverse osmosis, vhere large quantities of non-toxic
ions are also included In the concentrated products.  Significant savings may
be realized due to the reduction in the final volume of the stabilized pro-
duct that needs to be disposed of.  Even though it is not a stand-alone
treatment process, this chemieal/ultrafiltration combination method may con-
tribute to a cost-effective scheme which integrates a number of separation
processes to provide treatment for complex waste streams.  The proposed tech-
nology may be utilized to provide in-situ cleanup of contaminated soils, by
extracting the contaminants from the groundwater, and by recycling the water
to the contaminated zone to recover more contaminant adsorbed to the soils.

ACKNOWLEDGEMENT

    The tests described and the resulting data in this paper were obtained
from research conducted by Atomic Energy of Canada Limited and sponsored in
part by the U.S. Environmental Protection Agency, Eisk Reduction Engineering
Laboratory, Cincinnati, Ohio, under Federal Grant number CR-815321-01-0.

11PERBNCES

1.  Le, V.T., Buckley, L.P., and McConeghy, G.J.  Selective removal of
    dissolved radioactivity from aqueous wastes by a chemical treatment/-
    ultrafiltration technique.  Presented at the International Conference
    on Separations Science and Technology, Hamilton, Ontario, October
    1989 j also Atomic Energy of Canada Limited Report AECL-9861, June
    1989.

2.  Buckley, L.P., Le, V.T., McConeghy, G.J., and Martin, J.F.  Selective
    removal of dissolved toxic metals from groundwater by ultrafiltration
    in combination with chemical treatment.   Presented at Haztech Inter-
    national '89 Conference, Cincinnati, Ohio, September 12-14, 1989}
    also Atomic Energy of Canada Limited Report AECL-10030, September
    1989.

3.  Josephson, J.  Implementing Superfund.  Environ. Sci. Technol., Vol.
    20, No. 1, pp.23-28, 1986.

4.  Code of Federal Regulations, Protection of Environment, Title 40, Part
    261, Appendix II, "EP Toxic!ty Test Procedures", U.S. Government
    Printing Office, 1987.

5.  Cheryan, M.  In:  Ultrafiltration Handbook.  Technomic Publishers,
    Lancaster, 1986.  pp 144-151.

6.  Horin, K.A., Cherry, J.A., Dave, N.K., Lim, T.P., and Vivyurka, A.J.
    Migration of acidic groundwater seepage from uranium-tailings
    impoundments, 1.  Field study and conceptual hydrogeochemical model.
    Journal of Contaminant Hydrology, 2? pp 271-303, 1988.
                                     218

-------
                Laboratory Studies of the Thermal Destruction
                      of Toxic Organics in Sewage Sludge

                      Barry Bellinger and Sue L. Mazer
                        Environmental Sciences Group
                   University of Dayton Research Institute
                             Dayton, Ohio 45469

                              Richard A. Dobbs
                    U. S. Environmental Protection Agency
                           Cincinnati, Ohio 45268

                                  ABSTRACT

     Limited field data are available concerning organic emissions from
sewage sludge incinerators.  This is of particular concern because
hydrophobic hazardous organic compounds, such as certain pesticides,
polynuclear aromatic hydrocarbons, and polychlorinated biphenyls, have been
shown to partition onto the sludge during the wastewater treatment process.

     Laboratory thermal decomposition studies were undertaken to evaluate
potential organic emissions from sewage sludge incinerators.  Precisely
controlled thermal decomposition experiments were conducted on sludge spiked
with hazardous organic compounds, on mixtures of pure compounds in absence of
sludge, and on unspiked sludge.

     The following trends were observed:

     •  The sludge exerted little if any effect on the destruction
        of very fragile contaminants.
     •  If a contaminant was intermediate in thermal stability, it was more
        difficult to destroy in the presence of the sludge.
     •  If a contaminant was very stable in the pure mixture, the sludge
        enhanced its ease of destruction.
     •  Fluidized bed systems appeared superior to multiple hearth for
        control of organic emissions.
     •  The majority of the observed decomposition products were from the
        biomass decomposition, not the spiked contaminants.

     While incineration has been used to dispose of sewage sludge since the
1930's, concern regarding potential environmental insult due to organic
emissions from this disposal method has developed only recently (1-5).  Very
little is known about the fate of hazardous organic constituents or even
about total mass emissions of organic material from sewage sludge
incineration.  The data that do exist (6-10) suggest that some compounds


                                     217

-------
emitted from sludge incinerators may present a human health hazard, and the
mechanism of their formation warrants further study.

EXPERIMENTAL

     To study the thermal decomposition of common sewage sludge contaminants,
and to identify products of incomplete combustion (PICs) from contaminated
sewage sludge, thermal decomposition studies were conducted on two high
temperature flow reactor systems, the Thermal Decomposition Unit-Gas
Chromatographic (TDU-GC) System and the Thermal Decomposition Analytical
System (TDAS) (11,12),  Both of these systems incorporate a precisely
controlled fused silica tubular reactor and in-line gas chromatograph (TDU-
GC) or gas chromatograph/mass spectrometer (TDAS).   The systems may be
operated in a pyrolysis or oxidation mode, at temperatures between 300"C and
1100'C, and reaction times of 0.25-6.0 s.

     Thermal decomposition studies were conducted on a relatively clean
sludge spiked with contaminants, on mixtures of pure contaminants (without
sludge), and on unspiked sludge.  The contaminants which were studied were
heptachlor, pentachlorophenol, diphenylnitrosamine, pyrene, butylbenzyl
phthalate, 2,3',4,4',5-pentachlorobiphenyl, azobenzene, and technical-grade
nonylphenol.  The sludge was spiked at about 1 mg/g with the organic
components, except nonylphenol, which was spiked at 10 mg/g.  These levels
are at least an order of magnitude greater than those typically found in
"real world" sludges, but such spike levels were required to minimize
interferences from the sludge matrix.  Reactor temperatures for both systems
were varied over_the range of 300*C-1000°C, and a 2-second gas-phase, mean,
residence time (t ) at temperature was maintained for all experiments.

     A few experiments were performed to simulate the effect of the upper
hearths (drying zone) of multiple hearth incinerators on the destruction of
spiked sludges.  The behavior in this zone was mimicked by gradually heating
the sludge in the insertion region of the TDU-GC while maintaining the
reactor at a low temperature.  Thus, a single sludge sample was placed in the
sample insertion region of the TDU-GC, and was successively heated at
15*C/min in nitrogen over four temperature ranges.   These temperature ranges
were 50*C to 200*C, 2009C to 300'C, 300°C to 400'C, and 400C° to 500°C.  The
gas-phase residence time of desorbed species in the insertion region was less
than one second.  During each of the four sets of experiments, the reactor
was maintained at a nondegradative temperature (300"C).  While the sample was
being heated over a given temperature range, any evolved species were
condensed at the head of the in-line gas chromatographic column.  After these
species were collected, they were analyzed via GC/FID.

RESULTS

     The results of thermal oxidation and pyrolysis of the mixture and the
spiked sludge along with that of the other test are summarized in Table 1 by
the temperature for 99% destruction at t  = 2.0 s (Tgg(2)) in order of
increasing stability under pyrolytic conditions.  Dipnenylnitrosamine was
                                     218

-------
   Compound
   Heptachlor
   Butylbenzyl Phthalate
   Nonylphenol
   Azobenzene
   Pentachlorophenol
   Dlphenylamlne
   2,3',4,4',5-
     Pentachlorobi phenyl
   Pyrene
                                    TABLE 1
                          T99(2)  CC) OF TEST COMPOUNDS
                                   Pyrolysls
                                                         Oxidation
Hlxture
450
500
570
600
690
780
Sludqe
460
500
610
640
870
840
4
+10
0
+40
+40
+80
+60
Mixture
490
400
480
600
640
620
Sludqe
490
480
!>10
620
730
670
A
0
+80
+30
+20
+70
+50
                            870       820    -50
                            1000     >1000    --
760        740   -20
720        680   -40
<3
w
a:
UJ
u
oc
UJ
Q.
H
X
o
Ul
5
   «D
    10
0.1
    » 2.3'.4, 4'.5-PENTACHLOROBIPHEN YL
        o N2.  Mixture
        QAIR, Mixture
        *N_2,  Sludge
        °AIR, Sludge
  0,01
      0    100   200   300   400  SOO   600   7OO   800
                       EXPOSURE TEMPERATURE, "C
      Figure  1.   Thermal  decomposition profiles for 2,3',4,4'
                  pentachlorobiphenyl.
                                                               900  1 000
                                                              ,5-
                                   219

-------
converted to diphenylamine at transport temperatures (<275"C); consequently,
our thermal stability data are for the product, diphenylamine.
     Inspection of Table 1 shows that the test materials are generally more
stable under pyrolysis than oxidation for a given matrix (i.e., spiked onto
sludge versus run in a mixture of pure compounds).  This result is not
surprising in light of numerous other combustion and pyrolysis studies in the
literature, and can generally be accurately interpreted using chemical
reaction kinetic theory (13,14).  However, there is a more interesting
observation concealed in the data.  There is a general  trend of increasing
stabilization of the test compounds by the sludge matrix as the Tgn(2) of the
compound increases.  This holds true except for the two most stable compounds
2,3%4,4',5-pentachlorobiphenyl (PCB) and pyrene, which exhibit
destabilization by the sludge (cf. Figure 1).

     A multitude of PICs were observed to form from the thermal degradation
of the mixture, the unspiked sludge, and the spiked sludge.  From these three
sets of experiments it was clear that some of the PICs formed from the
thermal degradation of the spiked sludge were formed from degradation of
specific contaminants, others were formed from degradation of the sludge
biomass, while others had to be attributed to interactions of the biomass and
tjie contaminants. These- data are summarized in Table 2 along with a statement
of the possible source of the observed PICs.

     The source of the PICs was assigned based on 4 factors: 1) relative
yields from the mixture and sludge, 2) temperatures of maximum yield compared
to temperature for decomposition of possible precursors, 3) general
principles of reaction kinetics, and 4) the large precursor concentration for
the sludge and nonylphenol.  Some precursors for some PICs could be
specifically identified, while other PICs could occur as decomposition
products of several components of the contaminant mixture or the sludge.

DISCUSSION

     The decomposition of heptachlor and butylbenzyl phthalate are expected
to be dominated by fast unimoleeular reactions (15).  In the case of
heptachlor, a concerted elimination of HC1 is expected to be the dominant
destructive pathway although simple carbon-chlorine bond rupture may
contribute.  The rate of this reaction is independent of reaction atmosphere
and thus no (or a very small) change in stability due to the presence of
sludge is observed.  In a similar manner, the decomposition of butylbenzyl
phthalate is expected to have a large contribution from the concerted
reaction to form 1-butene and the corresponding acid.  This is observed in
pyrolysis! however, the oxidation data are not conclusive.  The thermal
oxidation profile has an inflection which indicates the presence of a co-
el uting impurity that may cause the butylbenzyl phthalate to appear
erroneously more stable when spiked into the sludge.

     The four contaminants of intermediate stability (nonylphenol,
azobenzene, pentachlorophenol, and diphenyl amine) are expected to be
                                     220

-------
         Product
                                      8ummwy el Th*mi«J OccompaiMan Product*

                                                                  OxWOen

                                                                   State*
CSM
Swum*
C«HI
C«HI2
C7H8
TeiuM*
IH-Pyr«i-2-on«
C7H14
OMoncdaptntadbn*
•wisanMb
Stnirttahyd*
Oulnen*
CflH1«
Bwuoturan
C3-Swinm
C8H1I
IH-fcvfen*
C9H1I
M»tiyt-1H-*xJtn»
Mtl^ft
Plwiel
                           MO
                           1000

                           7M
                           MO
                           MO
                           TOO
                           MO
                           780
                           1000
                           Tso
                           TOO
                           •00
•SO
700
700
MO
700
625
MO

700
7M
TOO
MO
KS
                                  MO
      MO
                                  m
                                  ISO
                                  800
               MO

               MO
               -Twr
               700
                                          aoo
               •00
               OOP

               880
               MO
               MO
               •SO
                                                                    «M
                                                                    860
                                  MO
•00
7SO
700
700

780

MO

•SO
TOO
050
                   800

                   BOO



                   too
                                         •so
                                         •so
                                  700    *50
                                                              47i
                                                              471
                                                              47S
                                                              471
                                                                             680

                                                                             MO


                                                                             471

                                                                             WO
                                                                             •SO
                                                                             •00
                                                                             600
                                                                            "KB"
                                                                             ta
                                          700
e?-M<«liflyfe«nz«M
TifcMmttifiMufMM
TflM(MaMMv(iM«M
                           •so
                           I2S

                           we
                           S2S
                           1000
                           1000
SOP

TOO
TOO
                           MO
                           POO
                           MO
                           MO
                           MO

                           MO
                           MO
                                          THT
                                          850
                                          880

                                          700
                                          •SO
                                                  •SO

                                                  •SO
                                                  •00

                                                  471
                                          MO
                                          «0
                                          TOO
                                          700

                                          MO
                                          MO
                                          MO
                                                              471
                                                              47t
                                                              •00
                                                             J22.

                                                              no


                                                              680
                                                             400
                                                             800

                                                             MO
                                                             WO
                                                  47*


                                                  MO

                                                  •00

                                                  •so
                                                  MO
                                                  MO

                                                  MO
                                                                                               Soure*
                                                                                             NonyfphMiel
                                                                  Mog^oo,

                                                                 . PCBh»n«i
                                                                  NomphMval
                                                                Mxturtoriludg*
                                                                                               Mxhn
                                                                                             Nonytjphwul
                                                                                             NonylphWMl
                                                                                             Namftohwwl
                                                              Mo»iy«ph«oo( of Slude*
                                                                    MMur*
                                                                                           MxkmorStw^p
                                                                                             Honytphtod
                                                                                               Mwtat
                                                                                                fC8
                                                                                                PCS
                                                                                                PC8
                                                                                                PCB
                                                                                         POP, PCB«Htpl»chk)r
                                                                                           HtpttcMorerPCt*
                                                                                           PCPwH*piieM«r
                                                                                           H«pt«chlofer POP
                                                                                          OtohtflyimJtro**mln*
                                                  221

-------
destroyed by bimolecular pathways.  Abstraction of H by OH or 0 is expected
to dominate under oxidative conditions.  Displacement of the aromatic
substituents from the contaminants by H is expected to dominate under
pyrolytic conditions.  The stabilizing effect of the sludge suggests that it
is a radical scavenger which lowers the reactive radical concentration and
slows the rate of destruction.
     Known scavengers such as benzyl (C6H5CH2) and allyl (C3H5) may be
readily produced from the degradation or the sludge biomass (16).  These
radicals are resonantly stabilized and may therefore exist long enough to
efficiently scavenge H atoms to form toluene and propene.  Scavenging of H
atoms under oxidative conditions also reduces 0 and OH concentration by
reducing the rate of the chain branching reaction H + 0* ---> 0 + OH (17).

     However, one must still rationalize the destabilizing effect of the
sludge on the two most stable contaminants, 2,3',4,4',5-pentachlorobiphenyl
and pyrene.  Both compounds are again expected to decompose by H abstraction
by 0 or OH under oxidative conditions.  Under pyrolytic conditions
2,3',4,4',5-pentachlorobiphenyl is expected to decompose by Cl displacement
by H while pyrene probably decomposes by H abstraction.  In either case, an
increase in radical concentration is required to explain the destabilizing
effect of the sludge.  We have hypothesized that the higher temperature
required for destruction of these species  (>700°C in air and >800°C in
nitrogen) is sufficient to destroy the benzyl and allyl radicals to levels
such that they are not effective scavengers.  Instead, the decomposition of
the sludge provides a large increase in the concentration of reactive
radicals and an increased rate of destruction of the stable contaminants.

     Of the 63 PICs reported in Table 2, we estimate that 41 originated
primarily from decomposition of the contaminants, 10 originated solely from
the sludge, 9 could have contributions from the sludge and the contaminants,
while only 3 could be clearly attributed to sludge/contaminant interaction.
The majority of the observed PICs were chlorinated and non-chlorinated
aromatic species while most of the remainder were chlorinated and non-
chlorinated short-chained unsaturated aliphatic hydrocarbons.  Many of the
observed non-chlorinated aliphatic species could be attributed to nonylphenol
side chain decomposition products.  There was very little evidence of
increased chlorination (the chlorine source being the contaminants) of
hydrocarbon species (primarily from sludge biomass degradation).  This is as
expected based on the large H/C1 ratio in the system due to the excess of
sludge and the generally unfavorable nature of displacement of H by Cl (18).
In fact, numerous chlorinated species which were observed from the
degradation of the mixture were not observed from the spiked sludge.  The H
atoms from the sludge facilitate Cl displacement and lower the concentrations
of chlorinated species.  Only chlorinated benzenes from the degradation of
2,3',4,4',5-pentachlorobiphenyl were major chlorinated products of the spiked
sludge.

     The major, thermally stable PICs were all aromatic in nature.  These
included simple benzenoid species such as benzene, toluene, and styrene which
were attributed primarily to sludge degradation.  Polynuclear aromatic
                                     222

-------
species were also observed to be stable and produced in significant yields,
including naphthalene, indenes, benzo and dibenzo-furans, 9 H-fluorene, and
phenanthrene which are also probably due to sludge degradation.  Nitrile
containing species such as benzonitrile, 9 H-carbazole, and
diphenyldicarbonitrile were among the highest yield stable PICs and
apparently resulted from interaction of diphenylamine or azobenzene
decomposition products and the sludge.

     Compared to previous studies of the thermal degradation of a wide range
of waste types, the spiked sludge resulted in a relatively high yield of
stable PNAs, aromatic nitriles, and chlorinated aromatic species (14,19).
However, we recognized that the wide range of combustion conditions,
especially exposure temperature, in a multiple hearth could significantly
affect emissions.

     Accordingly, a limited set of experiments  was  run to simulate the
stepped increase in exposure temperature of contaminated sludge cascading
down the hearths of the drying zone of a multiple hearth incinerator.
Specifically, it was expected that bed temperatures on the upper hearths
could be high enough to volatilize organic constituents, but that gas phase
temperatures might not be high enough to result in their destruction.  This
simulation was performed by a stepwise increase in the temperature of the
insertion region from 50 to 500°C while the reactor temperature was held
constant at the virtually non-destructive temperature of 300°C.  These data,
summarized in Figure 2, graphically display  that the majority of the
desorbable organics were volatilized below 200*C,  This corresponds to the
upper hearths of a multiple hearth incinerator, which for a countercurrent
flow design, means little destruction would be achieved.

     A comparision of potential emission levels from the multiple hearth
versus the fluidized bed incinerator is shown in Figure 3.  The top
chromatogram depicts volatile species desorbed from the sludge and
transported through the reactor at 300°C, while the bottom chromatogram shows
the remaining species after these volatilized organics have been subjected to
an 800'C reactor temperature.  While the thermal conditions in the top
chromatogram approximate those in the drying zone of the multiple hearth
incinerator, those in the bottom chromatogram are more characteristic of the
thermal regime in the fluidized bed incinerator.  This indicates that the
thermal regime of the fluidized bed incinerator may be much less likely than
that of a multiple hearth to lead to major gaseous organic emissions.

     We were also concerned about the effects of a high bed temperature (and
short reaction time at temperature) combined with a long burnout time but low
reaction temperature.  This would correspond to conditions achieved in a
multiple hearth incinerator when non-volatile components of a highly
contaminated sludge reach the upper hearths of the combustion zone.  Results
of sludge pyrolysis experiments at 500'C and 800*C revealed that large
quantities of cholesterol-like compounds, PNAs and chlorinated PNAs are
formed under these conditions.  The cholesterol-like compounds are likely
present in the sludge (from bile salts, natural and synthetic hormones, and
cell membranes) while the PNAs result from the sludge and sludge/contaminant
                                     223

-------
 Figure 2.

ChromatogranB of organics desorbed
from sludge  in various sample  insertion
region temperature ranges•
                                                   200*C~]00*C
                                                  300*C-400»C
                                                  400*C~JOO'C
 Figure 3.
Comparison of potential  emissions levels
froa a)  drying zone of multiple hearth
versusf  b) fluidlzed bed incinerator.
                                                Mactor at 100 *C
                                                              I
                                                            .1 I
tOO*C
                                                i i i t i i i i i i
                                 224

-------
interaction.  Since these PNAs and chlorinated PNAs are very stable under
oxidative or pyrolytic conditions, they are likely to escape destruction in
the upper regions of the incineration chamber.

Toxic Emissions Control Implications

     Our results suggest that counter-current flow multiple hearth
incinerators, as presently operated, may not be sufficient to destroy most
organic emissions with even a 99% efficiency.  Thus, much higher upper hearth
gas phase temperatures must be realized (>100Q°C) in order to control
emissions of volatiles.  Even with these controls it may be difficult to
control PIC emissions due to the reduced gas phase reaction time experienced
by combustion intermediates.  Although it is important to increase the gas
phase temperature of these systems, it also seems prudent to attempt to
strive for gradual heating of the sludge to minimize possible condensed phase
pyrolysis or flash volatilization (which causes gas phase pyrolysis due to
poor fuel/air mixing) which can result in formation of stable polynuclear
aromatic species (20,21).

     One solution to this dilemma would be to employ a high-temperature
afterburner.  Research studies indicate that a system operating at a minimum
temperature of IQOO'C, t  = 2.0 s, and 50% excess air should be sufficient to
destroy most toxic emissions (14,18,19).

     Regulation and compliance testing would appear to present a difficult
challenge because of the multitude of toxic contaminants that could be
contained in sewage sludge.  One possibility would be to conduct compliance
testing for major emissions identified from lab studies or comprehensive
research studies conducted on full-scale units.  Based on the results of our
lab tests benzonitrile, benzene, naphthalene, and benzaldehyde would be
appropriate compliance test surrogates that are likely to be "major" trace
emissions from any sewage sludge  incinerator regardless of contamination.  If
products specific to the sludge can be identified based on lab studies or
theory, those with the highest concentration and stability should also be
monitored.  The previously developed hazardous waste thermal stability index
could be used to guide these selections (14).

     The field test results on these compounds could then be used in
conjunction with a risk assessment of the stack emissions.  This risk
analysis should contain all possible emissions identified by laboratory
studies, reaction kinetic theory, or field tests.  The emissions results for
the limited number of compounds from the compliance test could then be
compared to theoretical or laboratory based predictions of emission rates  and
used to "calibrate" the emissions estimates for the other estimated
emissions.   In this way, a comprehensive, quantitative risk assessment may be
conducted.

                               ACKNOWLEDGEMENT

     The authors would like to recognize the efforts of our colleagues, Mr.
J. Stalter who assisted in the data acquisition; Ms. D. A. Tirey, and Mr.  R.
C. Striebich who assisted in review and preparation of this manuscript.

                                     225

-------
                                   CREDIT

     This work was partially supported by the US-EPA's Water Engineering
Research Laboratory under Cooperative Agreement CR-811777.

                                   REFERENCES

 1.  Report to Congress on the Discharge of Hazardous Wastes to Publicly
     Owned Treatment Works.  U.S. EPA Office of Water Regulations and
     Standards, Washington, D.C., 1986.

 2.  Fate of Priority Pollutants in Publicly Owned Treatment Works.  EPA-
     440/1-82/303, 1982.

 3.  A Comparison of Studies of Toxic Substances in POTW Sludges.  Final
     Report prepared by Camp, Dresser, and McKee for US-EPA Contract
     No. 68-02-6403, 1984.
 4.  Giger, W., Brunner, P. H.» and Schaffner C.  Science. 225(4462), pp.
     623-625, 1984.

 5.  Second Review of Standards of Performance for Sewage Sludge
     Incinerators.  EPA-450/3-84-010, US-EPA Office of Air Quality Planning
     and Standards, Research Triangle Park, NC, 1984.

 6.  Bridle, T. R., Bumbaco, M. J., and Creseuolo, P. J.  Fate of Polynuclear
     Aromatic Compounds During Sewage Sludge Incineration.  Presented at the
     75th Annual WPCF Conference, New Orleans, 1984.

 7.  Bennett, R, L. and Knapp, K. T.  Environ. Sci. & Tech., 16(12) pp.
     831-836, 1982.

 8,  Draft Final Test Report - Site 01, Sewage Sludge Incinerator SSI-A,
     National Dioxin Study, Tier 4: Combustion Sources.  Prepared by Radian
     Corp. for US-EPA Contract No. 68-03-3148, 1986.

 9.  Final Draft Report - Site 03, Sewage Sludge Incinerator SSI-B, National
     Dioxin Study, Tier 4: Combustion Sources.  Prepared by Radian Corp. for
     US-EPA Contract No. 68-03-3148, 1986.

10.  Draft Test Report - Site 12, Sewage Sludge Incinerator SSI-C, National
     Dioxin Study, Tier 4: Combustion Sources.  Prepared by Radian Corp. for
     EPA Contract No. 68-03-3148, 1986.

11.  Rubey, W. A. and Carnes, R. A.  Rev. Sci. Instrum. 56(9) pp. 1795-1798,
     September 1985.
12.  Rubey, W. A. and Grant, R. A.  Rev. Sc i... I n strum., Vol. 59, No. 2, pp.
     265-269, February 1988.
13.  Edwards, J. B.  Combustion, The Formation and Emissions of Trace
     Species, Ann Arbor Science, Ann Arbor, MI, 1977.

14.  Taylor, P. and Dellinger, B.  A Thermal Stability Based Ranking
     of Hazardous Organic Compound Incinerability, Environmental Sci.
     Technol., (in press, 1989 ).
                                     22G

-------
15.   Benson, S. W. and O'Neal, H. F.  "Kinetic Data on Gas-Phase Unimolecular
     Reactions," NSRDS-NBS 21. National  Bureau of Standards, Washington,
     D.C.,  1970.
16.   Held,  T. J., Dryer, F. L., Brezinsky, K., Pity, W. J., and Westbrook, C.
     K., "The High Temperature Oxidation of Isobutene," Presented at
     Symposium on Chemical and Physical  Processes in Combustion, the Eastern
     Section of the Combustion Institute Clearwater, Fl,  December, 1988.

17.   Westbrook, C. K., "Inhibition of Hydrocarbon Oxidation in Laminar Flames
     and Detonation by Halogenated Compounds," 19th Symposium (International)
     on Combustion, the Combustion Institute, 1982, pp 127-141..

18.   UDRI Draft Report, "Technical Resource Document Minimization and Control
     of Hazardous Combustion By-Products", Co-operative Agreement CR-813938-
     01-0,  US-EPA, December 1989.

19.   Dellinger, B., in Hazard Assessment of Chemicals - Current Developments.
     J. Saxena, ed., Vol. 6, pp. 293-337.
20.   Linak, W. P. et al., JAPCA. 1987, 37, No. 1, 54.

21.   Linak, W. P. et al., JAPCA. 1987, 37, No. 8, 934.
                                      227

-------
                PREDICTED AND OBSERVED ORGANIC EMISSIONS
                      FROM SEWAGE SLUDGE INCINERATION

                         Dabra A. Tirey, Richard C. Striebich,
                          Sue L Mazer, and Barry Bellinger
                        University of Dayton Research Institute
                           Environmental Sciences Group
                                Dayton, Ohio 45469

                                 Harry E. Bostlan
                        U.S. Environmental Protection Agency
                              Cincinnati, Ohio 45268

                                   ABSTRACT

      Samples of sewage sludge burned at three multiple hearth and one fluidized bed
incinerator were subjected to laboratory flow reactor thermal decomposition testing under
both pyrolytlc and oxidatlve atmospheres. The time/temperature conditions of the
laboratory testing were established to simulate as closely as possible full-scale incineration
conditions such that a direct comparison of results could be made.

      The laboratory test results indicated that biomass decomposition products, not toxic
Industrial contaminants, comprised the majority of the emissions. Benzene, toluene,
ethylbenzene, acrylonitrile, and acetonitrile were consistently the most environmentally
significant products of thermal exposure. Quantitative comparison of emission factors
derived from lab and field results for those compounds observed in both studies showed
excellent correlation for the pyrolysis testing. Rank/order correlations were statistically
significant at the 95% confidence level for all four sources. This study suggests pyrolysis
pockets of poor fuel/air mixing in the full-scale Incinerators may have been responsible for
many of the observed combustion by-products.

                                 INTRODUCTION

      Results of recent field tests on sewage sludge  incinerators have shown that a
number of toxic products of incomplete combustion (PICs) may be emitted from multiple
hearth and fluidized bed units (1-3).  This raised some concerns and suggests that a better
understanding of the nature of these emissions be developed  before adequate regulatory
and control strategies be undertaken.
      We have recently participated in a comparison of the results of bench scale thermal
decomposition testing of four sewage sludge samples with results of a full-scale
Incineration test program conducted by Radian Corporation under  US-EPA sponsorship (4-
10).   Three multiple hearth unite and one fluidized bed facility were chosen for examination
(c.f. Table 1.)
                                       228

-------
         TABLE 1. DESIGN PARAMETERS OF SELECTED SEWAGE SLUDGE INCINERATORS

         Design Parameter   Site 1	_Ste2	SitaJ	Ste4
Type Multiple Hearth
6 Hearths
Source Domestic
Exit O2 Cone (min) 10%
Combustion Zone 1 500-1 600°F
(815-8706C)
Multiple Hearth
8 Hearths
Industrial and
Domestic
6%
1300-1 600'F
(704-870°C)
Fluidized Bed
5 ft. sand depth
22 ft. diameter
Municipal
9-12%
13SO-14008F
(732-760«C)
Multiple Hearth
6 Hearths
Industrial and
Domestic
7-1 OX.
1100-1700°F
(593-9SE68C)
                                 EXPERIMENTAL

      Thermal decomposition testing for all four of the sewage sludge samples was
conducted using the System for Thermal Diagnostic Studies (STDS) (10,11).  This
Instrumentation assembly permits the study of the thermal decomposition of gaseous,
liquid, solid, or polymeric substances under very precise conditions.  Thermal
decomposition experiments can be conducted in a variety of atmospheres with residence
times varying from 0.1-10 seconds.  An In-line GC-MS system provides for analysis of the
effluent resulting from thermal exposures.

      Using combinations of insertion region programming and reactor temperature, the
multiple hearth Incinerator simulation was divided into four progressive steps,,   The first
and second steps were designed to simulate the drying zone of the incinerator (i.e., the
sludge was exposed to temperatures ranging from 50-450'C); the third and fourth steps
were designed to simulate the combustion zone of the incinerator (i.e., the sludge was
exposed to temperatures of 450-800°C). GC-MS analysis was performed at the end of
each step In this process. The individual contributions were then summed to provide for
the total emissions coming from laboratory simulated 'incineration1 of the sludge sample, It
was this data which was compared to full-scale emissions data for each of ths three sites.

      Simulation of the fiuidlzed bed unit required only a single step with rapid heating of
the sludge to 700°C in the insertion region  followed by a downstream reactor exposure of
700°C. GC-MS techniques were again employed to perform analysis of the reactor
effluent. Laboratory data from this exposure was then compared with full-scale data from
the single fluidized bed facility.

                                    RESULTS

      Full-scan experiments showed that greater than 99% of the mass emissions from the
thermal decomposition of all four sludges resulted from the decomposition of the biomass
of the sludge Itself. Many long-chain carboxyllc acids, ketones, aldehydes, amlno acids,
etc., were detected.  Even the two most common emissions seen from all four sludges,
benzene and toluene, were more a result of the thermal decomposition of the biological
material present In the sludge than from any Initial contamination in the feed itself. This
phenomenon was observed in previous studies on sludges from different situs (12).
      Sulfur containing compounds were only observed In laboratory tests for Sites 2 and
3. Perhaps the industrial contribution to the influent waste stream in both cases was
responsible for the sulfur-containing emissions.

      The reactor effluent for the laboratory fluidlzed bed sample from Site 3 was notably
much cleaner In terms of number and relative yield of products than observed for the three
multiple hearth samples. This Is as one might expect based on the more uniform
combustion provided for in the simulated fluidized bed experiments.
                                        229

-------
       A comparison of laboratory and field test results are presented in Table 2.  This
comparison Is complicated by the difficulty of conducting chemical analysis for specific toxic
materials within an overwhelming organic matrix as backround. Both laboratory and field
studies were subject to this difficulty.  As a result, the joint program concentrated upon
detection of specific toxics contained  in a generic 'target list' which are enumerated in
Table 2.  Compounds were selected based  upon toxicity, suspected prevalence in sewage
sludge, and representativeness of different classes of potential pollutants.  Compounds
were also added to the 'target list' based upon their detection in a preliminary round of
laboratory testing of sludges from the four test sites which was completed before full-scale
testing was begun. This initial testing allowed for the addition of major 'site specific'
emissions to the generic 'target list*.
      TABLE 2.  COMPARISON OF TARGET LIST COMPOUNDS DETECTED IN FULL-SCALE AND
                                LABORATORY TESTING
Sftef
Field Lab
Acrytonttrilo
Benzene
Carbon Tetrachioride
Chlorobenzene
Chloroform
1,2-DIchloroethane
1 ,2-Dtehloroethene
Ethylbenzene
Methytono Chbride
Tetrachtoroethytene
Toluene
1 ,1 ,1 -Trichtoroethane
Trfchforosthyterie
Vinyl Chloride
AcetonKrlle
2-Butanona{MEK)
Pyridlne
Phenol
Naphthalene
2-Nttrophenol
1 ,2-DfchJorobenzene
1 ,3-DIehtorobenzene
1 ,4-Dfchtorobenzene
Xfl
X
X
X
Xw
X
X
X
X
Xs&w
X
X
X
X
nap)
na
na
X
na

X
X
X
X
Xs





X

X


na(6)
na
na
na
X
na




Site 2
Field Lab
Xw
X
X
Xs
XW<4)

X
Xs
Xs&w
X
Xs
X
X
X
na
na
na
X
Xs
X
X
X
X
X
X





Xs

Xs


na
na
na
na






Sites
Field Lab
X
Xs(3)
Xs
Xs
Xs&w


Xs
Xs&w
Xs
Xs
Xs
Xs

na
na
na
s
na
na
na
s
s
(2)
Xs






Xs
Xs


na
na
na
na

na
na
na


Site 4
Field Lab
X
Xs
X
X
Xs&w
na
Xs&w
Xs
Xs&w
Xs&w
Xs
Xs&w
Xs&w
X
X
X

X
Xs




X
Xs





X
Xs
Xs
Xs

na
X
X

X
X




      1. 'X* Indicates compound was detected In effluent from stack or from thermal testing
      2. ' blank ' indicates compound was analyzed for, but was not found In effluent
      3. Subscript *w' denotes compound was detected In the scrubber inlet water
      4. Subscript's* denotes compound was detected in the sludge feed
      5. 'na' Indicates that compound was not included in target list' for that facility
      6. Vinyl chloride was not included in laboratory testing
Site!
      Previously performed laboratory studies of the pxidatlve pyroiysis of pure
tetrachioroethyiene have shown that carbon tetrachioride is formed as a PIC (13).  These
experiments also showed that given a hydrogen source, trichloroethylene was readily
formed as a PIC from tetrachioroethyiene.  The radical pool generated during the
Incineration of municipal sludge would provide such a source of hydrogen. Thus, if
tetrachioroethyiene were present in a sludge stream, both trichloroethylene and carbon
tetrachioride would likely be detected in the stack effluent. Full-scale data generated in this
program correlated with this laboratory finding; tetrachioroethyiene was detected in the
sludge feed sample collected during the full-scale testing and carbon tetrachioride and
trichloroethylene were found in the stack. However, tetrachioroethyiene was not detected
In the sludge feed sample furnished for  the laboratory study, as evidenced by its absence

                                          230

-------
In the 300°C run; nor were any other chlorinated compounds detected.  This resulted in no
chlorinated products being observed.

      This was a reoccuring problem throughout the test program.  Although precautions
were taken, It Is possible that the small sample furnished for laboratory analysis; was not
truly representative of the sludge feed, or that the volatile organic content of the laboratory
samples was lost during transit or cold storage, or that  microbial anaerobic degradation
occured.

Site 2
      In the feed samples collected during full-scale testing from Site 2, two chlorinated
compounds, methylene chloride and chlorobenzene, were found which were not detected
in the sample provided for laboratory testing.  In previous laboratory experiments we have
found that 1,1-dichloroethyiene, 1,2-dichloroethy!ene, and trichloroethylene were all
formed as products from the pyrolytic decomposition of pure methylene chloride (14).
Thus, it Is not surprising that a facility incinerating sludge contaminated with methylene
chloride had these compounds present in its emissions. Since no methylene chloride was
present in the laboratory sludge feed sample, these products could not be expected to be
detected.  In fact, no chlorine-containing compounds were detected in the sludge sample
provided for laboratory testing. This explains the absence of chlorine-containing
compounds in the effluent from thermal testing of this sample.

      The chlorobenzene detected in the stack emissions at this facility probably resulted
from a combination of two processes. The first is the original contamination found In the
sludge feed. Previous studies have shown chiorobenzene to be a very stable compound,
even when decomposed as part of a mixture (15,16). Thus, the appearance of
chlorobenzene in the stack is probably due, in part, to the ability of this compound to
undergo relatively harsh thermal exposures and escape unscathed. The second
contribution to the emission of chlorobenzene at the  full-scale facility is the possibility that
this compound was formed as a PIC. We have conducted many studies involving the
decomposition of benzene and/or toluene or related  compounds in the presence of
chlorinated species.  Hie free chlorine radicals generated in such a reaction atmosphere
have combined with phenyl radicals to produce chlorobenzene in almost every case (16).
Because a host of chlorine containing compounds were detected in the sludge feed at the
full-scale facility, it is logical that chlorobenzene was detected in the stack.

      Two products seen in full-scale emissions, tetrachloroethyiene and 1,1,1-
trichlorethane, may be the result of a slightly more complex scenario.  Tetrachlorethylene Is
a frequent product observed from the oxygen deficient decomposition of various
chlorinated materials as mixtures or pure compounds (17). It has proven to be a stable
compound when seen as a PIC. In fact, significant levels of tetrachloroethyiene were still
present in one laboratory study at temperatures above 900°C, making this compound one
of the most stable PICs observed in any experiments we have conducted. Thus, it is
conceivable that a decomposition atmosphere containing some of these compounds could
have produced tetrachloroethyiene as a PIC.

      The emission of 1,1,1 -triehloroethane is surprising due to its previously
demonstrated thermally fragile nature (18).  For it to be emitted from the incinerator as a
PIC, it would have to be formed near the exit of the high temperature zone and its
decomposition rapidly quenched.  Recombination of methyl radicals and trichloromethyl
radicals is the most  likely mechanism of formation. Since chloroform is the best source of
trichloromethyl radicals detected, and it was in the scrubber water, it is most likely that the
1,1,1 -triehloroethane was formed outside the combustion zone of the incinerator. The
proposed  reaction requires temperatures high enough to decompose chloroform
(approximately 500°C), presumably from quenching  of hot gases  in the scrubber,  followed


                                        231

-------
by methyl-trlchloromethyl radical recombination.  Alternatively, a liquid phase reaction must
be Invoked.

       Finally, methylene chloride and chloroform were both seen as contaminants In the
scrubber water at the Site 2 facility. This explains their presence in that facility's emissions
during full-scale testing and their absence In the lab tests. Acrylonitrlle was also detected
In the scrubber water at the Site 2 facility, but its presence in the stack cannot be attributed
exclusively to this phenomenon.  This compound was seen at the pollution control device
Inlet for this facility, Indicating that It was formed as a PIC In the Incineration of the Site 2
sludge. Laboratory thermal testing reproduced this finding.
      Many hit list compounds were observed In the full-scale analyses of the sludge feed
at the fluldlzed bed facility. All of these compounds were subsequently detected In
emissions from full-scale testing conducted at tills facility: benzene, carbon tetrachlorlde,
chlorobenzene, chloroform, ethylbenzene, methylene chloride, tetrachloroethylene,
toluene, trlchloroethyiene, and 1,1,1-trlchloroethane. Analysis of the sample provided for
laboratory testing found only benzene, toluene, and very low levels of tetrachloroethylene,
while only benzene and toluene were detected as products In the effluent from laboratory
thermal experiments.

      However, It Is  important to note that the emission factors (in units of mg
emitted/metric ton of combustibles fed) reported from the Site 3 facility show that all but a
few of the compounds were present In the stack at low levels (I.e., less than an emission
factor (EF) of 100).  Only benzene, chloroform, methylene chloride and toluene had EF
values greater than 100. Of these, chloroform and methylene chloride were contaminants
of the scrubber water.  Laboratory testing duplicated full-scale results for those compounds
with field EFs of greater than 100, except for those compounds whose EFs were an
indication of their contamination In the scrubber water.

Site 4
      Full-scale analysis reported  methylene chloride, 1 ,2-dlchloroethylene, benzene,
chloroform, ethylbenzene, tetrachloroethylene, toluene, 1,1,1-trlchlorethane,-
trlchloroethyiene, and naphthalene to be contaminants of the original sludge feed for the
facility at Site 4. All of these compounds except  1 ,2-dichloroethylene were subsequently
observed In stack emissions.  Laboratory analysis of the Site 4 sample only documented
the presence of benzene, toluene, tetrachloroethylene and 1,1,1 -trlchloroethyiene.  Of
these benzene, toluene, and small levels of tetrachloroethylene were seen In the
emissions from laboratory thermal testing.  Naphthalene was also detected In laboratory
testing, although It was not detected In the original sludge feed sample as was the case for
full-scale testing.

      Full-scale testing also showed acetonitrile, acryionltrile, carbon tetrachlorlde,
chlorobenzene, 2-butanone (MEK)  and phenol to be present In full-scale emissions.  These
compounds were undoubtedly formed as PICs. Laboratory testing of the Site 4 sample
reproduced detection of benzene, MEK, acetonitrile, acrylonitrile and phenol.

      The presence of tetrachloroethylene and 1,1,1-trlchloroethane in the full-scale and
the laboratory sludge feed samples (9,10) would predict that carbon tetrachlorlde and
chlorobenzene should  be found In the emissions from full-scale and laboratory testing of
this sludge for reasons outlined previously.  In  fact, full-scale results did show carbon
tetrachlorlde and chlorobenzene to be part of the stack emissions from the facility at Site 4.
However, testing of the laboratory sludge sample from the same site did not include these
compounds.  Comparison of the analytical detection limits between laboratory and field
testing showed that laboratory detection limits  were only slightly higher than those reported


                                         232

-------
for the full-scale testing. Therefore, It Is possible that lower initial concentrations of parent
materials (tetrachloroethylene and 1,1,1-trichloroethane) in the laboratory feed sample did
not permit sufficient conversion to products so that the analytical detection limits could be
surpassed  for detection of carbon tetrachloride or chlorobenzene.

                                    DISCUSSION

      The apparent loss of chlorinated compounds from the sludge samples for whatever
reasons clearly complicated field/lab comparisons.  As a result we conducted a statistical
comparison of the relative  emissions for only those compounds observed in both the
laboratory and field testing. Table 3 presents a ranking of the  relative rate of these
emissions from highest to lowest, as well as a listing of the  relative emission factors (REFs)
for each  of these compounds.  Subjecting the ranked pairs of field/lab data to the Wilcoxin
Rank Sum Test (19) provided  interesting results. In all three cases amenable to statistical
analysis  the rankings of highest to lowest emissions exhibited a statistically significant
agreement. Site 3 could not be examined  in this manner since there were insufficient data.

      TABLE 3.  RANKING OF TOP EMISSIONS FOR COMPOUNDS QUANTrTATED IN FOJLL-SCALE
                               AND LABORATORY TESTING

                               Ranking               Relative Emission Factors
                                                     (relative to benzene)
             Compound          Field   Lab                Field   Lab
             Slta  1:
             Acrylonitrile          2     2                  0.350  0.836
             Benzene            11                   11
             Toluene            3     4                  0.301  0.015
             Ethylbenzene        4     3                  0.101  0.638
             Slta  2:
             Acrylonitrile          1     1                   2.40   2.03
             Toluene            2     2                  2.36   1.66
             Benzene            33                  11
             Ethylbenzene        4     4                  0.100  0.311
             Slt«  3:
             Benzene            11                   11
             Toluene            2     2                  0.160  0.120
             Site  4:
             Acetonitrile          1     3                  5.20   0.130
             Acrylonitrile          3     2                  3.33   0.147
             Benzene            41                   11
             MEK               5     6                  0.330  <0.005
             Ethylbenzene        6     5                  0.122  0.016
             Toluene            2     4                  4.44   0.020

      We also conducted  a comparison of the observed average emission of all
compounds ('hit list* and observed semi-volatiles) in the full-scale testing from all sites
versus predicted emissions based on a large previously generated data base on stability of
organic compounds (20-22). The full-scale destruction efficiencies (DEs) for each  species
were ranked  and compared to the destruction efficiencies predicted based purely on our
previously  developed thermal stability based ranking of toxic organic compound
indnerability  (20). Since many of the emissions were dearly due to PIC formation, a
second comparison was made wherein the predicted contribution due to PIC formation was
Included (21,22).  Principles of elementary reaction kinetics and available lab data were
used to estimate PIC yields (21-23). The impact on measured DE depends on the
concentration of the species in the feed. Those compounds with low feed concentrations
are subject to large  effects from PIC formation, while the measured DE of those compounds
with high feed concentration are only affected by PIC formation if the PIC yield is very high.
Figure 1  presents a correlation summary of these two comparisons: 1) theoretical stability-
                                          233

-------
vs-observed destruction efficiency and 2) theoretical stability plus PIC formation propensity-
vs-observed destruction efficiency.




•o
«
>
o
J3
O







zu
18

16
14
12

10

8

6

4

2
n
• 1*1*1 •• i » * "I'l'i'i %«j""™n
- •" 0
a J
• ThtrmtlStibllllyonly Q 1
O • H
O PICFormUonlndudKl Q 1
a
a
a
• o
• o
o •
o •
D
O •
0 •
o • -
• a
*i«fit*$|t*tii.i*i>
           0     2     4    6     8     10    12   14   16   18   20
                                    Theoretical

      Figure 1 - Theoretically Predicted-vs-Actual Pollutant Emission Ranking


      As can be seen, excellent correlation was obtained when PIC considerations were
Included In the theoretical ranking of thermal stability, whereas poor correlation was
observed when theoretical stability alone was compared with observed stability.  The
Spearman Coefficient of Rank correlation test resulted in a 99.5% confidence interval for
correlation between field data and theoretical predictions when PIC formation was
Included. Only chloroform showed a major deviation.  However, this is not surprising since
chloroform was found In the scrubber water at many facilities. Stripping of this compound
from the water and subsequent emission in the stack accounts for its seemingly high
'observed' ranking.

                          SUMMARY AND CONCLUSIONS

      This was the first time that a direct comparison has been attempted using an actual
waste sample for laboratory analysis which was incinerated at a full-scale facility. In spite
of the complexity of the program, the comparison of results for major P!C emissions was
remarkably good. Future work should include a plan for immediate comparison of the
chemical analysis results for samples collected for  laboratory and field testing to ensure
that a true comparison of emissions between the two studies may be guaranteed (I.e., to
ensure that both studies will be working with the same starting material).

      Emission factors (mass emitted/mass fed units) were used in this study as a basis of
comparison between laboratory and field data. The full-scale emission factors were
generated with data taken at the outlet of a pollution control device, while laboratory data
might have been more analogous to full-scale data taken at the inlet to the pollution control
device,   This is because the product effluent generated  in laboratory thermal testing does
not get 'scrubbed'.   In the future, It would be advisable to compare full-scale emissions at
the inlet of the pollution control device with those obtained in laboratory thermal testing to
                                        234

-------
see if these comparisons will provide a better correlation. This type of comparison was not
possible for this study.

      The laboratory data generated in nitrogen agreed best with full-scale data. This is
consistent with the previously proposed theory that most 'combustion' emissions are
actually due to pyrolysis in small oxygen deficient pockets of gases  (23).   In fact,  a
previously proposed pyrolytic thermal stability index of toxic waste incinerability in
conjunction with estimates of PIC emission predicted full-scale emission test results with
remarkable accuracy (24).

      The development of an a priori 'target list1 is a valuable first step towards
determining potential toxic organic emissions from sewage sludge Incinerators. This
necessarily requires that one suspect a particular compound to be present.  However,
more research is clearly necessary before we can ensure that ail potentially toxic
emissions are being addressed.  One possible technique for developing a more
comprehensive full-scale emissions test approach would be to conduct laboratory thermal
decomposition testing on spiked sludges. By adding higher levels of toxic organic
contaminants and PIC precursors to the sludge, trace toxic emissions can be more  readily
and comprehensively identified to guide full-scale testing and monitoring.


                           Acknowledgement qnd Credit

      The authors acknowledge the contribution of our colleague Mr. J. Staiter who
assisted in obtaining the experimental data and to Mr. M. Palazzolo of Radian Corporation.
for his thoughtful insights. This work was supported under EPA Cooperative Agreement
CR813938-01-0 and sub-contract to Radian Corporation S94223.


                                  REFERENCES

1. Report to Congress on the Discharge of Hazardous Wastes to Public Owned Treatment
   Works.. U.S. EPA Office of Water Regulations and Standards.  Washington, 0. C..
   1986.

2. Final Draft Test Report - Site 03, Sewage Sludge incinerator SSI-B, National Dioxin
   Study, Tier 4, Combustion Sources. Radian Corp. for US-EPA Contract No. 68-03-
   3148.  1986.

3. Draft Final Test Report - Site 01, Sewage Sludge Incinerator SSI-A, National Dioxin
   Study, Tier 4, Combustion Sources.  Radian Corp. for US-EPA Contract No. 68-03-
   3148,.  1986.

4. Bostian, H. E., Grumpier, E. P.. Metals and Organics Emissions At Four Municipal
   Wastewater Sludge Incinerators. Paper presented at the Waste Incineration Section  of
   the Pacific Basin Consortium for Hazardous Waste Management Meeting held in
   Singapore. April  1989.

5. Bostian, H. E., Crumpler, E. P., Palazzoio, M. A., Barnett, K. W.. Emissions of Metals and
   Organics From Four Municipal Wastewater Sludge Incinerators - Preliminary Data. Ja;
   Proceedings of Conference on Municipal Sewage Treatment Plant Sludge
   Management.  Palm Beach, Florida. June  1988.

6. Site 1 Draft Emission Test Report Sewage Sludge Test Program. DCN:87-232-009-60-
   01. Radian Corp. for US-EPA Contract No. 68-02-6999. July 1987.

                                        23S

-------
7.  Site 2 Draft Emission Test Report Sewage Sludge Test Program.  DCN:87-232-009-71-
   01. Radian Corp. for US-EPA Contract No. 68-02-6999. October 1987.

8.  Site 3 Draft Emission Test Report Sewage Sludge Test Program.  DCN:87-232-009-60-
   06. Radian Corp. for US-EPA Contract No. 68-02-6999. October 1987.

9.  Site 4 Rnal Emission Test Report Sewage Sludge Test Program.  DCN:89-239-005-28-
   05. Radian Corp. for US-EPA Contract No. 68-02-4288. September 1989.

10. Final Report entitled, "A Laboratory Investigation of Sewage Sludge Incineration and
   Comparison to Full Scale Incineration Data,". US-EPA  Co-operative Agreement
   CR813938-01-0 and sub-contract to Radian Corporation S94223.  September 1987.

11. Rubey.W.A. and Grant, R.A.. RQV. Scl. Instrum.. Vol. 59(2). pp 265-269. February,
   1988.

12. Project Summary and Final Report entitled, "Potential Emissions of Hazardous Organic
   Compounds from Sewage Sludge Incineration,".  US-EPA Co-operative Agreement
   CR811777. 1987.

13. Tirey, D. A., Taylor, P. H., Dellinger, B..  Gas Phase Formation of Chlorinated Aromatic
   Compounds from the Pyrolysis of Tetrachloroethylene.  Submitted to Combust. Scl.
   Technol..  September 1989.

14. Tirey, D. A., Taylor, P. H., Dellinger, B..  Products of the Incomplete Combustion from the
   High Temperature Pyrolysis of the Chlorinated Methanes.  ID; Proceedings of
   Symposium on Emissions from Combustion Processes: Origin, Measurement, and
   Control. American Chemical Society.  In press 1989.

15. Graham, J. L, Hall, D. L, and Dellinger, B.. Environ. Scl. Technol..  Vol. 20. pp 703-
   710. 1986.

16. Dellinger,  B., Tirey, D. A., Taylor, P. H. and Lee, C. C..  Laboratory Studies of a
   Hazardous Waste incinerability Surrogate Mixture. Paper presented at the 15th Annual
   EPA research Symposium held in Cincinnati, Ohio. April 1989.

17.Taylor, P.  H.and Dellinger, B.. Environ. Scl.Technol..  Vol.22.  pp438. 1988.

18. Benson, S. W. and O'Neal, H. E.. kinetic Data on Gas-Phase Unimolecular Reactions.
   NSRDS-NBS 21. National Bureau of Standards.  Washington, D.C.. 1970.

19. Dickson, W. J.f Massy, F. J., Jr..  Introduction to Statistical Analysis- Jni 2nd ed..
   McGraw-Hill, New York. 1957.  pp 294-295, 384, 468,469.

20. Taylor, P.  H. and Dellinger, B., A. Thermal Stability Based  Ranking of Hazardous
   Organic Compound Incinerability.  Environ. Sci. Technol..  In press, 1989.

21. Dellinger,  B..  Theory and Practice of the Development of a Practical Index of
   Hazardous Waste Incinerability. 1m J. Saxena, (ed.), Hazard Assessment of Chemicals
   - Current Developments. Vol.6.  Hemisphere Pub. Co.. Washington, D. C..1989.  pp
   293-337.
                                       236

-------
22. UDRI Drift Report, Technical Resource Document .Minimization and Control of
   Hazardous Combustion By-Products. Prepared for US-EPA under Co-operative
   Agreement CR813938-01-0. December 1989.

23. Dellinger, 8., Rubey, W. A., Hall, D. L and Graham, J. L. Haz. Waste and Haz. Mat.
   Vol.  3.  No. 2. pp 139-150. 1986.

24. Dellinger, B., Graham, M. D., and Tirey, D. A.. Haz. Wasta and Haz. Mat...  Vol. 3. No. 3.
   pp 293-307.1986.
                                      23?

-------
             CHARACTERIZATION OF MUNICIPAL WASTE
                COMBUSTION ASHES AND LIACHATE
                     FIELD STUDY RESULTS
                       HAIA K.  ROFFMAN
                   AWD TECHNOLOGIES,  INC.
                          ABSTRACT
Incineration of MSW has become an important alternative to the
land disposal of MSW.  Incineration is an effective means of
reducing the volume of MSW and can provide an important source
of  energy.   Ash from  the combustion  of  household waste has
been excluded from regulations under Subtitle C of  RCRA, which
regulated  disposal of hazardous  wastes.   However,  in  some
instances   testing  the   residues  from  municipal   waste
incinerators by  the Extraction  Procedure (EP) Toxicity  test
is  being  required  to  determine if these  residues  would be
classified  as  hazardous waste  and, therefore,  subjected to
disposal  regulations  under  Subtitle C.    Ashes from  MWC
facilities,  on  occasion,  have  exhibited hazardous waste
characteristics as determined by  the  EP  Toxicity  test.  The
debate regarding the representativeness  and  the validity of
this  test  and  the relation  of  these   results  to  actual
leachates from ash disposal facilities has not been settled.

For  this  reason,  EPA and  CORRE  have  cosponsored a study
designed to enhance the  data  base on  the characteristics of
MWC ashes,  laboratory extracts  of MWC ashes,  and leachates
from MWC ash disposal facilities.   Ash samples were collected
from 5 MWC facilities and leachate samples were collected  from
the companion  ash  disposal sites.   These ash  and  leachate
samples  were  analyzed  for the   Appendix  IX  semivolatile
compounds, polychlorinated dibenzo-p-dioxins/polychlorinated
dibenzofurans (PCDDs/PCDFs), metals for which Federal primary
and secondary  drinking water standards  exist, and  several
miscellaneous conventional compounds.  The ash samples were
also subjected to six laboratory extraction procedures  and the
extracts were then analyzed for  the same compounds as the ash
samples. All sampling, laboratory preparation, and laboratory
analysis followed stringent quality assurance/quality control
(QA/QC) procedures.

The major findings  of this study are described  in this paper.

                            238

-------
                         INTRODUCTION
This paper  provides a summary of the findings provided in a
recent  report  which has been prepared for the United States
Environmental  Protection Agency  (EPA) and  the Coalition on
Resource Recovery and the Environment (CORRE).  EPA and CORRE
have cosponsored  this study to enhance the data base on the
characteristics  of  MWC ashes,  laboratory  extracts  of  MWC
ashes,  and  leachates from MWC  ash disposal facilities.

The Coalition on Resource Recovery and the Environment  (CORRE)
was established to provide credible information about resource
recovery and associated environmental issues to the  public and
to public officials.  In providing information,  CORRE takes
no  position as  to  the appropriateness  of  one  technology
compared to others.   CORRE recognizes that successful waste
management  is  an integrated utilization of many technologies
which  taken as a  whole, are  best  selected by  an informed
public  and  informed  public officials.

Incineration of  municipal  solid  waste (MSW)   has  become an
important waste disposal alternative because  it provides an
effective means of  reducing the volume of MSW as well as an
important source  of energy recovery.   Currently,  10 percent
of the United States MSW is incinerated.   Based on the number
of municipal waste combustion  (MWC) facilities being planned
across  the  country,  this percentage is expected to increase
to 16-25 percent by  the  year 2000.

As incineration of MSW has increased in recent years,  so has
concern over its management.   To resolve the  many  legal and
technical issues  surrounding  ash,  Congress  is  considering
several legislative initiatives that would classify municipal
waste combustion (MWC)  ash as a special waste under Subtitle D
of the Resource  Conservation  and  Recovery  Act (RCRA)  and
require the Environmental Protection Agency (EPA)  to develop
special management standards for the full life cycle of ash.
In anticipation of Congressional action, EPA and the coalition
on Resource Recovery and the Environment  (CORRE)  cosponsored
this  study  to  characterize  ash  and  to  gain  a  better
understanding of how it behaves in the environment.
                            239

-------
                     DESCRIPTION OF STUDY
 Combined bottom and fly ash samples were collected from five
 mass-burn MWC facilities and leachate  samples were collected
 from the companion ash disposal  sites.

 The  facilities  sampled  were selected by  CORRE  to meet  the
 following criteria:

   o    The facilities were to be state-of-the-art facilities
        equipped   with  a   variety   of   pollution   control
        equipment.

   o    The facilities were to be located in different regions
        of the United States.

   o    The  companion ash  disposal facilities  were  to  be
        equipped with leachate  collection systems  or  some
        means of collecting leachate samples.

 The identities of the facilities are being held in confidence.

 The  ash  and  leachate  samples collected were analyzed  for  the
 Appendix   IX   semivolatile   compounds,    polychlorinated
 dibenzo-p-dioxins/polychlorinated dibenzofurans (PCDDs/PCDFs),
 metals for which Federal primary and secondary drinking water
 standards exist,  and  several  miscellaneous   conventional
 compounds.   In addition, the ash  samples were analyzed  for
 major components in the  form of oxides.

 The  ash  samples  were  also  subjected   to   six  laboratory
 extraction procedures and the extracts  were then  analyzed  for
 the same compounds as the original ash samples. The following
 six extraction procedures were used during this  study:

   o     Acid Number 1 (EP-TOX)
   o     Acid Number 2 (TCLP Fluid No.  1)
   o     Acid Number 3 (TOLP Fluid No. 2)
   o     Deionized Water  (Method  SW-924), also known  as the
        Monofill Waste Extraction Procedure (MWEP)
   o     eo2 saturated deionized water
   o     Simulated acid rain (SAR)

These extraction procedures have been used separately by a
variety  of  researchers on MWC ashes but  never have all six
procedures been used on the same MWC ashes. This was intended
to compare the analytical results of the extracts  from all six
procedures with  each  other  and with leachate  collected from
the ash disposal sites used by the MWC facilities.

                            240

-------
All sampling, laboratory preparation, and laboratory analysis
followed  stringent  EPA  quality  assurance/quality  control
 (QA/QC) procedures.   The detection limits of the  analytical
methods  used  were  well  below  present  levels  of  human,
environmental, or  regulatory concerns.

The EPA publication "Interim Procedures for Estimating  Risk
Associated   with  Exposures  to   Mixtures  of  Chlorinated
Dibenzo-p-Dioxins  and Dibenzofurans (CDDs and CDFs)" was  used
to evaluate  the  dioxin data.   These procedures use  Toxicity
Equivalency  Factors  (TEFs)  to  express the concentrations  of
the different isomers and homologs as an equivalent amount  of
2,3,7,8-Tetrachloro  Dibenzo-p-Dioxin  (2,3,7,8-TCDD).   The
Toxicity  Equivalents,  as calculated by  using  the TEFs, are
then totaled and compared to the Centers for Disease Control
(CDC)  recommended upper   level  of  2,3,7,8-TCDD  Toxicity
Equivalency of 1 part per billion in residential soil.

The major features of  the five  MWC facilities and ash sites
sampled are  provided in Table  1,  and  Table 2 respectively.
Pertinent information on the operating conditions  of the MWC
facilities,  as  well as information  about  the  air pollution
control equipment  used is also provided  in Table 1.
                STUDY RESULTS AND CONCLUSIONS
Major  findings of  and conclusions  drawn from  the results
obtained  from  the  sampled  ash,  natural  leachates,  and
laboratory  extracts  are summarized in  the  paragraphs which
follow.
                    Ash Analysis  Results
Of the five ash samples (one from each facility) analyzed for
the Appendix IX semivolatile compounds, four samples contained
bis(2-ethylhexyl)phthalate,   three   contained   di-n-butyl
phthalate, and one contained di-n-octyl phthalate.  Two PAHs,
phenathrene and fluoranthene, were detected in only one of the
five ash samples.  These semi-volatile compounds were detected
in the parts per billion (ppb)  range.

The results for the ash samples analyzed for PCDDs/PCDFs are
presented in Table 3.  This table also includes the calculated
Toxicity Equivalents (TE)  for each homolog of PCDD/PCDF.  The
data indicate  that PCDDs/PCDFs were found at  extremely low
levels in each of the ash samples.  The Total TE for each ash

                           241

-------
sample was  well  below the Centers for Disease Control  (CDC)
recommended Toxicity  Equivalency  limit of 1 part per billion
2,3,7,8-TCDD in  residential soil.

All 25  of the ash samples  (five  daily composites from each
facility) were analyzed for the metals listed on the primary
and secondary  drinking water  standards  as well  as  for the
oxides of five najor  ash  components.  Although, the results
from these  analyses  indicate  that the ash is heterogeneous,
this heterogenicity appears to have  been reduced by the care
taken when  compositing the ash samples during  this study.
Data from this study  showed less variability than comparable
data in the literature.

Metals  showing  the   widest  range  of concentrations  among
samples  collected at each facility included  barium   (ZB);
cadmium  (ZB); chromium (ZD,  ZE) ; copper  (ZA,  ZB,  ZC) ; lead
(ZD); manganese (ZA,  ZC) ?  mercury  (ZE); zinc (ZB,  ZD,  ZE); and
silicon dioxide  (ZA).

Metals showing the widest  variation of concentrations between
the facilities included barium (results  for Facility ZC are
lower  than  the   results  for  the other facilities) ;  iron
(results  for each facility  vary   from  all  of the  other
facilities)  ; lead (results for Facility ZD are higher than the
results  for the other  facilities);  mercury   (results  for
Facilities  ZC and ZD are lower  than the results for the other
facilities); sodium   (results  for Facilities  ZD and ZE are
lower than  the  results for  the  other  facilities);  calcium
oxide (the  results for  Facilities ZA and ZB are higher than
the results for  the  other facilities);  and  silicon  dioxide
(the results for Facility ZC are higher than the results for
the other facilities).

Some additional  findings of the ash sampling and analyses are
as follows:

  o   The  ashes are  alkaline with the pH ranging from 10.36
       to 11.85.

  o   The  ashes are rich in  chlorides and sulfates.   The
       total soluble  solids in the ashes  varied from 6,440 to
       65,800 ppm.

  o   The  ashes contained unburnt total  organic carbon (TOC)
       ranging  from  4,060 ppm  (0.4  percent)  to  53,200 ppm
        (5.32 percent).
                            242

-------
                  Leachate Analysis Results
Only four Appendix IX semi volatile compounds were found in the
leachates.   Benzoic acid was found  in two leachate  samples
collected   at  one  site.     Phenol,   3-methylphenol,  and
4-methylphenol were found in the leachate samples from another
site.  All of these compounds were detected at very low levels
(2-73 ppb).

PCDDs/PCDFs of the  higher chlorinated homologs were  found  in
the  leachate  from  one site  only.    This  indicates  that
PCDDs/PCDFs  do  not readily leach  out  of the ash.   The low
levels  found  in the  leachates of  the  one site  probably
originated from  the solids  found within the leachate samples
because these samples were not  filtered nor  centrifuged prior
to analysis.

The metal content in the leachate samples did not. exceed the
EP Toxicity Maximum Allowable Limits established for the eight
metals  in  Section  261.24 of 40  CFR 261.   Indeed,  the  data
indicate that although the leachates are not used as  a  source
of potable water, they are close to being acceptable as  such
as far as the metals are concerned.

Other observations  on the leachate analyses are:

   o     Sulfate  values ranged from 14.4  mg/L  to 5,080 mg/L,
        while  Total  Dissolved  Solids  (TDS)  ranged   from
        924 mg/L to 41,000 mg/L.

   o     The field pH values  ranged  from 5.2  to 7.4.

   o     Ammonia  (4.18-77.4 mg/L) and nitrate (0.01-0,45 mg/L)
        were present in  almost all  leachate  samples.

   o     Total  Organic  Carbon  values  ranged  from  10.6 to
        420 ppm.


                Ash ExtractsAnalysis Results
The  data  obtained  during the  metals analyses  of the  ash
extracts indicate, in general, that the extracts from the EP
Toxicity,  the  TCLP  1,  and the TCLP 2  extraction procedures
have  higher  metals  content than  the  extracts  from  the
deionized water (SW-924),  the saturated CO2 solution, and the
Simulated Acid Rain (SAR)  extraction procedures.

                            243

-------
The EP Toxicity Maximum Allowable Limits for  lead and cadmium
were frequently exceeded by the extracts from the EP Toxicity,
TCLP 1, and TCLP 2 extraction procedures.  One of the extracts
from the  EP Toxicity extraction procedure also exceeded the
EP Toxicity Maximum Allowable Limit for mercury.

None of the extracts  from the deionized water  (SW-924), the
saturated  CO2  solution,  and  the Simulated  Acid  Rain (SAR)
extraction  procedures  exceeded  the  EP  Toxicity  Maximum
Allowable Limits.  In  addition, all of the extracts from these
three extraction procedures also met the Primary and Secondary
Drinking Water Standards  for metals.

Table  4 compares the range of  concentrations of  the metals
analyses of the ash extracts with the range of concentrations
for leachate as  reported  in the literature and the range of
concentrations for the leachates as  determined in this study.
For the  facilities sampled during  this study, the  data in
Table  4 indicate that the extracts  from the deionized water
(SW-924), the saturated CO2 solution,  and the SAR extraction
procedures simulated  the  concentrations for lead and cadmium
in the field leachates better than the extracts from the other
three extraction procedures.

Additional observations are:

   o    Of the  five composite  samples  of the deionized water
        (SW-924)  extracts analyzed  for  the  Appendix  IX
       semivolatile compounds  (one from each facility), only
       one sample  contained  low  levels of benzoic  acid
        (0.130 ppm).

   o    None of  the  extracts  contained PCDDs/PCDFs.   These
       data confirm the findings of  the actual field leachate
       samples that PCDDs/PCDFs are  not leached from the ash.
                            244

-------
            TABLE 1




MAJOR FEATURES OF MWC FACILITIES
Operational
Features
Facility Type
Startup Date
Capacity
Combustion
Temperature
Temperature of
air entering the
boiler
Volume of air
entering boiler
Source of ash
quench water
Air pollution
control
equipment
Approximate
waste
composition
Facilities
ZA
Energy recovery,
continuous feed, reverse-
reciprocating grate.
May 1986
275 tons/day/boiler
2 boilers
1,800-2.000^ at stoker
Under fire: 250*F
Over fire: ambient
Under fire:
70,000-90.000 Ib/hour
Over fire:
41.000 Ib/hour
Floor drains, rain water.
Lime slurry is injected
into flue gas after
economizer, fabric filter
baghouses.
Residential: 40%
Commercial/
Light industrial: €0%
ZB
Energy recovery,
continuous feed,
reciprocating grate.
Early 1987
75- lOOtons/day/faoiler
2 boilers
1.8000F
Under fire: ambient
Overfire: ambient
Under fire:
10,890 cu ft/min
Over fire:
5,900 cu ft/min
Cooling tower and boiler
blowdowns. septic system
discharge, floor drains.
Dry lime is injected into flue
gas after economizer, fabric
filter baghouses.
Fly ash has phosphoric add
added to it and is
agglomerated before being
mixed with bottom ash.
Residential: 80%
Commercial/
Light Industrial: 20%
ZC
Energy recovery,
continuous feed, reverse-
reciprocating grate.
January 1987
400 tons/day/boiler
3 boilers
1. 750-1 .800°F
Under fire: 380°F
Over fire: ambient
Under fire:
34,000 ftf/min
Over fire:
11,000ft3/min
Tertiary effluent from
neighboring sewage
treatment plant.
Electrostatic
precipitators.
Residential: 60%
Commercial/
Light Industrial: 40%
ZD
Energy recovery,
continuous feed,
reciprocating grate.
1975
750 tons/day/boiler
2 boilers
1500-1700°Fftuegasasit
enters superheater
Under fire: ambient
Over fire: ambient
Under fire:
48,000 ftVmin
Over fire:
32,000 ft3/min
Cooling tower and boiler
blowdowns.
Electrostatic precipitators
Residential: 90%
Commercial/
Light Industrial: 10%
ZE
Energy recovery,
continuous feed,
reciprocating grate.
September 1987
750 tons/day/boiler
2 boilers
1,800°F at the grate
Under fire: ambient
Over fire: ambient

Wastewater from plant
processes.
Lime slurry is injected into
flue gas after economizer,
electrostatic precipitators.
Fly ash has water added to
it and is agglomerated
before being mixed with
boitoiri asii.
Residential: 65%
Commercial/
Light Industrial: 35%

-------
TABLE 1
MAJOR FEATURES OF MWC FACILITIES
PAGE TWO
Operational
Features
Amount of
electricity
generated
Amount of
electricity used
internally by
facility
Material
removed from
incoming refuse
Material
removed from
ash
Facilities
ZA
13.1 megawatts/hour
1.7 megawatts/hour
Large appliances, other
unacceptable material
diverted to demolition
landfill.
Ferrous metal removed
from ash at the MWC
facility.
ZB
4.5 megawatts/hour
0.63 megawatts/hour
Large appliances, material
that will not pass through
the boilers.
None.
ZC
29 megawatts/hour
2.5 megawatts/hour
Large appliances,
material that will not
pass through the boilers.
Ferrous metal removed
from ash at the MWC
facility.
ID
35 megawatts/hour
2.5 to 3.5
megawatts/hour
Large appliances,
material that will not
pass through the boilers.
Ferrous metal removed
from ash at the MWC
facility.
ZE
45 megawatts/hour
7 megawatts/hour
Large appliances, material
that will not pass through
the boilers.
Items greater than
to inches in diameter.

-------
                  TABLE 2




MAJOR FEATURES OF MWC ASH DISPOSAL FACILITIES
Operational
Features
Facility Type
Startup Date
Disposal Capacity
Amount of Ash
Disposed
Materials other
than Ash
disposed of
Leachate
Collection System
Cover
Compaction of
Ash


ZA
Monof ill -single clay
liner
1986
83,400 cubic yards
150 tons/day
None
Perforated PVC pipe in a
coarse aggregate
envelope
Final cover -soil and
HDPE
Only as bulldozer spreads
act. in 3tK Kit


ZB
Monofill - double liner
(HDPE and compacted till
soil)
October 1988
90,000- 100,000 tons
60 tons/day
None
Slotted HDPE
Daily cover -sand. Non
working face covered by
plastic to limit leachate
generation
Bulldozer spreads and
i»*>rw%rk-»4-*-4' -%*U t*i O4<*» :«M*>k
lifts.
Facilities
1C
Codisposed facility -
bottom-clay liner
synthetic sidewall liners
Landfill -1984
Ash Disposal -1985
Total capacity 9 million
tons
400,000 tonsfyear.
40% ash (2/3 of ash from
ZC MWC facility).
Non-burnable materials
from 2 MWC facilities.
Overflow from 2nd MWC
facility.
Main header - PVC
collection trenches -
gravel with fabric filter
Daily - native soil and
shredded tires.
Intermediate - native
soils.
final - native soils.
Track mounted


ZD
Monofill - unlined. Ash is
placed over trash
deposited before 1975
1975
Remaining capacity -
990,000 tons (6 years)
450 tons/day
None
None - leachate samples
were collected from well
points installed in the ash
Daily cover -soil.
Intermediate - soil
compacted to 10 6
permeability.
Final -clay or HDPE.
Only as bulldozer spreads
~.u :_ __i«. f-.tt


ZE
Monofill -double liner
(HDPE and clay)
1987
Permitted for 20 years,
approximately 3.8 million
tons
S25 tons/day
None
Slotted HDPE
Daily cover - soil.
Intermediate - soil
compacted to ID'6
permeability.
Final -clay of HDPE.
Vibrating roller.


-------
                                                                   TABLES

                                                             ASH DiOXIN RESULTS
Compound
23,7,8-TCDD
Other TCDD
23.7.8-TCDF
Other TCDF
1,2.3.7.8-PCOD
Other PCDO
1,2,3,7.8-PCDF
23.4,7,8-PeDF
Other PCOF
1.2.3A7,8-HXCDD
1.2.3,6,7.8-HxCDD
1,23.7,8,9-HxCDD
Other HXCDO
1.2.3A7.8-HXCDF
1,2,3.6,7,8-HxCDF
1,2A7A9-HXCDF
2,3,4.6.7.8-HxCDF
Other HxCOF
1,2,3,4.6,7,8-HpCOD
Other HpCDD
U.3.4.6.7.8-HPCDF
1.2.3,4,7,8,9-Hi»CDF
Other HpOJF
OCOD
OCDF
TOTAL TEs
Towcity
Equivalent
Factor
(TEFJW
1
0.01
0.1
0.001
OS
0.005
0.1
0.1
0.001
0.04
0.04
0.04
0.0004
0.01
0.01
0.01
0.01
0.0001
0.001
0.00001
0.001
0.001
0.00001
0
0

Samples (pg/§ or ppt)
ZA-AH-003
Value
10
206
263
1,688
33
317
61
46
484
12
17
28
1S4
74
131
36
5
281
159
140
139
8
51
313
66

Toxidty
Equivalents
10
2.06
263
1.69
16.5
1.59
6.1
4.6
0.484
0.48
0.68
1.12
0.062
0.74
1.31
0.36
0.05
0.0281
0.159
0.0014
0.139
0.008
0.00051
0
0
74.5
ZB-AH-001
Value
24
351
617
3,721
118
759
194
162
1.527
40
34
79
342
336
524
127
54
939
319
288
539
48
197
544
243

Toxicity
Equivalents
24
3.S1
61.7
3.72
59
3.80
19.4
16.2
1.53
1.6
1.36
3.16
0.137
3.36
5.24
1.27
0.54
0.0939
0.319
0.0028
0.539
0.048
0.00197
0
0
211
ZC-AH-003
Value
16
281
236
1,208
71
1,051
64
56
607
66
90
120
925
218
279
193
70
635
1,849
1.511
653
83
254
6,906
563

Toxicity
Equivalents
16
2.81
23.6
1.21
35.5
5.26
6.4
5.6
0.607
2.64
3.6
4.8
0.37
2.18
2.79
1.93
0.70
0.063S
1.85
0.0151
0.653
0.083
0.00254
0
0
119
2D-AH-003
Value
35
541
626
2,633
ND
1,910
151
171
1,736
86
148
194
853
654
660
479
124
1,686
1,555
1384
1,842
119
384
4,519
893

Toxicity
Equivalents
35
5.41
62.6
2,63
0
9.55
15.1
17.1
1.74
3.44
5.92
7.76
034
6.54
6.60
4.79
1.24
0.169
1.56
0.0138
1.84
0.119
0.00384
0
0
189
ZE-AH-003
Value
10
120
176
1,136
35
248
52
43
448
11
It
22
104
95
134
45
20
280
122
0
155
16
44
294
59

Toxicity
Equivalents
10
1.2
17.6
1.14
17.5
1.24
5.2
4.3
0.448
0.44
0.44
0.88
0.042
0.95
134
0.45
0.20
0.028
0.122
0
0.155
0.016
0.00044
0
0
63.7
(<>    Toxicity Equivalency Factors are EPA's current recommended Factors, (EPA, March 1987).
NO   Not detected below 221 pg/g.

-------
                   TABLE 4
COMPARISON OF ASH EXTRACT METAL ANALYSES RESULTS
      WITH LEACHATE METAL ANALYSES RESULTS
Parameter
Arsenic
Barium
Cadmium
Chromium
Copper
Iron
Lead
Manganese
Mercury
Selenium
Silver
Sodium
Zinc
Samples(pg/L)
EPTOX
Extracts
ND-31
23-455
25- 1,200
ND-86
24-5,170
ND-82,000
ND- 19,700
250-8,540
ND-203
ND
ND
33,600-
225,000
67-95,600
TCLP1
Extracts
ND
161-1,850
ND-1,150
ND-8.0
5-858
ND-7,220
ND- 10,550
ND-5,170
ND-3.8
ND
ND
1,000.000-
1,640,000
9.7-79,500
TCLP2
Extracts
ND-60
12-809
ND-1,560
ND-799
5.4-1,400
ND-1 62,000
ND-26,400
3,8-7,370
ND-4,6
ND
ND
38,700-
228.000
26-164,000
CO2 Extracts
ND-53
126-530
ND-354
4.2-98
8.8-620
ND-304
ND-504
ND-2,390
ND-1 55
ND
ND-1 6
24,800-
168,000
5-127.000
DIH2O
Extracts
ND-45
ND-3,050
ND-7.6
6.8-16
12-534
ND-1 15
ND-3,410
ND-20
ND-0.96
ND
ND
24,100-
209,000
5.4-1,340
SAR Extracts
ND
127-3,960
ND-6.0
ND-10
8.5-610
ND-97
ND-3.940
ND-6.4
ND-1.1
ND-23
ND
24,200-
201,000
12-1,290
Leachate
(Literature)*1)
5-218
1,000
ND-44
6-1,530
22-24,000
168-
121,000
12-2,920
103-4,570
1-8
2.5-37
70
200,000-
4,000,000
ND-3,300
Leachate
(CORRE)
ND-400
ND-9,200
ND-4
ND-32
ND-12
ND-1 0,500
ND-54
6.7-18,500
ND
ND-340
ND
14,000-
3,800,000
5.2-370

-------
TABLE4
COMPARISON OF ASH EXTRACT METAL ANALYSES RESULTS
WITH LEACHATE METAL ANALYSES RESULTS
PAGE TWO
Parameter
Aluminum Oxide*
Calcium Oxide*
Magnesium Oxide*
Potassium
Monoxide*
Silicon Dioxide*
Samples (pg/L)
EPTOX
Extracts
ND-1 50,000
592,000-
4,810,000
27,300-
130,000
10,100-
189,000
5,090-98,700
TCLP1
Extracts
ND-62,800
666,000-
2,750,000
55-375,000
14,600-
210,000
379-51,700
TCLP2
Extracts
ND-1 52,000
692,000-
3,540,000
623-137,000
15,100-
1,110,00
820-143,000
COa Extracts
ND-90,700
398,000-
1,920,000
207-59,300
12,300-
155,000
418-71,800
DIH20
Extracts
ND-203,000
141,000-
1,740,000
21-379
13,100-
189,000
402-3,990
SAR Extracts
ND-1 18,000
142,000-
1,800,000
12-430
14,500-
181,000
364-3,770
Leachate
(Literature)d)
NR
21,000
NR
21,500
NR
Leachate
(CORRE)
ND-920
64,600-
8,390,000
14,800-
367,000
10,900-
1,620,000
ND
ND  Not Detected.
NR  Not Reported in the literature.
(D   EPA, October 1987.
*    The ash extracts were analyzed as ions for these compounds and reported as oxides. The leachates were analyzed and are reported as ions for
     these compounds.

-------
    MOBILITYOF DIOXINS AND FURANS ASSOCIATED WITH STABILIZED INCINERATION
                      RESIDUES.IN THE MARINEENVIRONMENT

        by:  Frank J, Roethel, Vincent T, Breslin and Kenneth Aldous*

                          Waste Management Institute
                        Marine Sciences Research Center
                         State University of New York
                       Stony Brook, New York  11794-5000
                                     and
                     *New York State Department of Health
                Wadsworth Center for Laboratories and Research
                              Empire State Plaza
                            Albany, New York 12201
                                   ABSTRACT

     The combustion of municipal solid wastes in "waste to energy"
facilities produces ash known to contain trace amounts of organic compounds
such as polychlorinated dibenzodioxins (PCDD) and polychlorinated
dibenzofurans (PCDF).   Prior investigations have shown that these combustion
by-products can be stabilized using Portland cement to form solid concrete-
like materials.  Small artificial habitats were constructed in Conscience
Bay, Long Island Sound, New York using blocks of stabilized incineration
residue.

     Divers have periodically returned to the site to monitor the
interactions of the stabilized incineration ash blocks with the marine
environment.  Results show that the blocks retain their structural integrity
after prolonged seawater exposure,  PCDDS  and PCDFS . were measured in pg/g
concentrations and are retained within the cementitious matrix of the block
following placement in the sea.  Organisms growing on the blocks were found
to contain no measurable quantities of organics.

     This paper has been reviewed in accordance with U. S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.


                                      251

-------
                                INTRODUCTION
     In urban coastal areas where  landfills are few and increasingly
distant, ocean  disposal of stabilized  incineration residues may provide an
acceptable alternative to current  landfill practices  (1).  Previous studies
have demonstrated  that the stabilization of combustion residues (coal fly
ash, oil ash) using  additives  (lime, sodium carbonate, Portland cement) can
be used to produce solid blocks that are environmentally acceptable in the
sea (2,3).

     In May  1985 a research program was initiated at  the Marine Sciences
Research Center to examine the feasibility of utilizing stabilized
incineration residues for artificial reef construction in the ocean.
Results of these studies showed that particulate incineration residues could
be combined  with cement to form a  solid block possessing physical properties
necessary for ocean  disposal  (4).  The stabilized residues were subjected to
regulatory extraction protocols and in no instance did the metal
concentrations  in  the leachates exceed the regulatory limits for toxicity
(5).  Bioassays revealed no adverse impacts to the phytoplankton communities
exposed to elutri'dte  concentrations higher than could  be encountered under
normal disposal conditions (4).  The success of the laboratory studies
resulted in  securing the necessary permits for the placement of an
artificial habitat constructed using stabilized incineration residue in
coastal waters.

     April 1987 and  again in September 1988, stabilized incineration residue
blocks and cement  blocks, that serve as a reference material, were submerged
in eight meters of water in Conscience Bay, Long Island Sound, New York.
The purpose  of  the September 1988  placement was to confirm the data gathered
from the first  habitat and examine the fate of PCDD's and PCDF's associated
with incineration  residue.  Divers have periodically returned to the reef
site to study the  interactions of  stabilized incineration residue with the
marine environment.  Stabilized incineration residue blocks were retrieved
from the reef site for physical and chemical testing.  Compressive strengths
of the reef blocks were measured to monitor the strengths of the blocks and
after 13 months of submersion experienced a 5% increase in strength.
Samples of blocks  exposed to seawater were analyzed for metals to determine
if these consitiuents associated with particulate residues are effectively
retained within the  stabilized blocks.  To date the data suggest,  that
blocks possess  concentrations of metals nearly identical to preplacement
values supporting  the concept that metals are effectively being retained.
In addition, divers  removed organisms from the surfaces of the blocks to
evaluate the possible uptake of metals.  No metal enrichment was observed in
the tissues of  organisms associated with the ash blocks when compared to
data collected  from  organisms removed from concrete control blocks.

     The present study was designed to evaluate the mobility of dioxins.and
furans incorporated  in blocks of stabilized incineration ash following
placement in the sea.
                                      252

-------
                       ASH SAMPLES AND BLOCK PRODUCTION
     For the second artificial habitat, placed into the sea in September
1988, incineration residues for block making were collected from the
Baltimore Resource Recovery Facility, Baltimore, Maryland in August 1988.
Combined ash, a mixture of bottom ash and fly ash, was collected.  Prior to
block making, the combined ash was sieved to particle sizes l€;ss than 3/8".
The residue >3/8" was crushed with a jaw-crusher and resieved.  Screening
the combined ash was necessary to prevent large particles from damaging the
block making equipment.

     Block manufacturing was conducted in August 1988 at the research
facilities of the Besser Company at the Alpena Community College, Alpena,
Michigan.  Block fabrication employed conventional block making machines
currently used by the industry.

                               REEF PLACEMENT
     The "Narrows" region of Conscience Bay, Long Island Sound was selected
as the site for the in situ investigations of the interactions: of stabilized
incineration residues with the marine environment.  Conscience Bay is a
small embayment immediately west of the Port Jefferson Harbor channel on the
northern shore of Long Island.  Tidal currents within the "Narrows" can
exceed 10 Km/hr (6).  As a result, the sediments of Conscience Bay at the
reef site are predominantly composed of poorly-sorted gravel and coarse
sand.  The salinity within Conscience Bay varies between 25.5 to 29.7 ppt
and the water temperatures range from 0-24,5°C (5).

     Stabilized incineration residue blocks and standard cement blocks were
submerged in about 8 meters depth  (mean high tide) on September 23, 1988.
The blocks were arranged underwater to produce two separate structures, one
incineration residue and one cement, approximately 2 meters apart.  Each
reef was designed to maximize the  surface area exposed to seawater and to
provide numerous crevices to facilitate biological colonization.

     Sample collection on the reef was performed by certified SCUBA divers.
Due to strong tidal currents within Conscience Bay, diving activities were
restricted to forty-five minutes prior to and after slack water.

                       ORGANIC EXTRACTION PROTOCOLS
INCINERATION ASH OR STABILIZED INCINERATION RESIDUE

     Each sample of incineration ash or coarsely ground stabilized
incineration residue is added directly to a glass Soxhlet thimble containing
a layer of dichloromethane-extracted silica gel.  A finely milled,
homogeneous sample is mixed 1:1 with coarse, dichloromethane-e>xtracted,
anhydrous sodium sulfate before addition to a glass Soxhlet thimble.  The


                                      253

-------
sample is,spiked at approximately 60 pg/g (concentration of each congener)
•with the   C._ labeled PCDD/PCDF internal standard mixture, covered with a
plug of hexane-extracted glass wool, and Soxhlet -extracted with benzene for
16 hours.  The extract is concentrated to 25 mL for processing through the
following cleanup steps.

BIOMASS

     Each sample is added to a 150-mL flask and. spiked at approximately 1
ng/g (concentration of each congener) with the   C.„—labeled PCDD/PCDF
internal standard mixture.  The sample is homogenized with 50 ml of 50%
ethanol/eyelohexane using a Virtis Model 45 homogenizer with an
UltraShear  attached.  The homogenate is transferred to a 600-mL beaker with
three 50% ethanol/eyclohexane rinsings.  The suspension is boiled gently for
three rain.  Fifty mL of ethanol and 25 mL of cyclohexane is added and
boiling is continued for 20 min.  The supernatant is filtered.  Fifty mL of
50% ethanol/eyclohexane is added to the biomass remaining in the 600—mL
beaker; it is boiled for five min and filtered.  The filtered solid is
washed two times with 5 mL of cyclohexane, once with 10 mL of ethanol and
then the combined washes and filtrates are concentrated to less than 100 mL.
The extract is divided between two centrifuge bottles using ethanol rinses.
Fifteen mL of cyclohexane and 25 mL of distilled water is added to each.
The bottles are centrifuged for five min at 5000 rpm.  After the organic
layer is removed from each portion, 15 mL of cyclohexane is added, and the
centrifugation process is repeated.  Another 15 mL of cyclohexane is added
to each aqueous layer, and the centrifugation step is repeated.  The combined
extracts are concentrated to about 2 mL.  To remove all the traces of
ethanol, another 50 mL of cyclohexane is added^ and the extract is again
concentrated to about 2 mL and dried over anhydrous sodium sulfate.  The
extract is dissolved in 25 mL of benzene for processing through the cleanup
steps described below.

CLEANUP OF EXTRACTS

     The following steps are used to remove interferences from the extracts
and enrich the analytes for PCDD/PCDF analysis.

Acid/Base Column

     The extract is passed through a mixed acid/base column containing
silica gel, sulfuric acid-treated silica gel, and potassium hydroxide
treated silica gel with three 10-mL portions of hexane, then 30 mL of
hexane.  The first 90 mL is collected as the PCDD/PCDF fraction.

Automated Chromatography Sys tern

     After preliminary cleanup on the acid/base column, the extract is
cleaned up using a sequence of acid alumina, carbon,  and neutral alumina
columns (7).  By using the solvent system listed below, the four columns can
be used in series with no concentration of the extract.  The 90-mL extract
                                      254

-------
from the acid/base column is applied to the acid alumina column, followed
with 30 mL of 3% dichloromethane/hexane and eluted with 70 raL of 50%
diehloromethane/hexane onto the carbon column.  The 70-mL PCDD/PCDF fraction
from the acid alumina column is followed by 50 mL of 10% benzene/hexane in
the forward direction.  The PCDD/PCDF fraction is eluted with 30 mL of 50%
xylene/hexane in the reverse direction onto the neutral alumina column.  The
30-mL extract from the carbon column is followed by 37 mL of 3%
dichloromethane/hexane, and eluted with 75 mL of 50% dichloromethane/hexane.
The extract is concentrated to 100 /zL.

Silica Gel/Micro Diol Column

     The concentrated extract from the automated chromatography cleanup is
applied to a micro column containing a layer of silica gel and a layer of
Bondesil 20H (Analytichem International, Harbor City, CA) .  The extract is
followed by 100 p"L of hexane, then 1.6 mL of hexane.   The first 1.6 mL
represents the PCDD/PCDF fraction.

Final Sample Concentration

     The hexane eluent from the silica gel/micro diol column is concentrated
to 100 /*L, replaced with benzene, transferred to a 0.1 mL tapered vial with
rinses, and concentrated to about 0.5 fiL.  The sides of the vial are washed
down with 20 ftL of benzene and the extract is concentrated to about 0.5 ^*L,
The vial sides are washed down with the recovery standard,   C, ,,-1,2,3,4
-TCDD.  The final volume for GC/MS is 4 fiL,


PCDD/PCDF ANALYSIS

     The Department of Health Laboratory, in Albany,  uses the USEPA Method
8280 "The Analysis of Polychlorinated Dibenzo-p-dioxins and Polychlorinated
Dibenzofurans".  The system used to analyze a sample extract for the
presence of PCDDs and PCDFs is comprised of a Hewlett-Packard (HP) model
5890A Gas Chromatograph (Hewlett- Packard Co., Palo Alto, CA) interface
directly to the ion source of an HP 5970B Mass Selective Detector (MSB), and
an HP ChemStation.  This is a high resolution capillary column gas
chromatography/low resolution mass spectrometry (HRCG/LRMS) technique that
incorporates an HP 9000 computer (work-station), which controls virtually
all instrument parameters and contains the vendor-supplied data-acquisition
and processing software.

     The HP Ultra 2 fused silica capillary column used separates individual
levels of chlorination of PCDDs and PCDFs by retention time into separate
windows and effectively separates 2,3,7,8-TCDD from all other TCDD isomers
for isomer-specific determination of 2,3,7,8-TCDD.

     Data is acquired by selected ion monitoring (SIM).   Two ions of the
molecular ion isotopic cluster that are characteristic of each tetrachloro
through oetachloro PCDD or PCDF level of chlorination are monitored.  The
fragment ions arising from the loss of COCl from the molecular ion are also


                                     255

-------
monitored.  A list of the Ions and the respective time intervals in which
they are monitored is presented in Table 1,

     Software written in this laboratory for the IBM PC and PC-compatibles
(8) is used to process the SIM data files.  TASQ is a collection of
interacting compiled BASIC programs which contain the routines that search
the data and follow the decision-making criteria by which PCDDs and PCDFs
are found and quantitated using criteria described in USEPA Method 8280.

     The mass scale of the MS is calibrated by using perfluorotributylamine
(PFTBA).  The HP 5970B MSB is autotuned daily to ions at m/z 69, 219, and
502 to meet the isotopic ratio criteria.

                          RESULTS AND DISCUSSION
     The results of the PCDD/PCDF analyses conducted on duplicate samples of
the combined incineration ash prior to stabilization is presented in Table
2.  All isomers were detected and found in concentrations typical of
combined incineration ash (9).  The 2,3,7,8-CDD s were detected in
increasing concentrations from the lower to the higher chlorination levels
in each of the ash samples.  TCDD concentrations ranged from 5.3 to 17 pg/g
and OCDD from 1100 to 1400 pg/g.

     Among the 2,3,7,8 CDF s, the hepta isomer was detected in the highest
concentration with a range from 180 to 240 pg/g.  The 2,3,4,8«TCDF
identified in these analyses represent 2,3,7,8 TGDF and co-eluters. The
Ultra 2 GO column used is not isomer specific for 2,3,7,8-TCDF and several
other isomers could co-elute at the same retention time.  One sample showed
no detectable 2,3,7,8 TCDF though we believe it is present.  A shifting of
the retention time (Table 1) just beyond the retention time tolerance would
cause the peak present not to be matched with the standard.

     Table 3 presents the PCDD/PCDF concentrations found in stabilized ash
blocks prior to and following placement in the sea.  Prior to submersion,
all 2,3,7,8 isomers were observed in concentrations lower than mean ash
concentration reported in Table 2.  This is due in part to the addition of
15% Portland cement used as the stabilization additive.   Reduction in
concentrations beyond that explained by dilution is observed for certain
isomers and may be the result of an encapsulation of the ash particles by
the cement thereby reducing the extraction efficiency.

     Following placement of the blocks in the sea, duplicate analyses
indicate  no significant reduction of 2,3,7,8-isomers when compared to
preplacement results.  Milling one sample of the exposed block did result in
observing higher concentrations of most 2,3,7,8-Isomers.  The milled sample
had a mean particle size that ranged between 132 faa. and 65 fan significantly
smaller than the unmilled samples whose particle size ranged between 425 ^m
and 0.25 inch.  Two TCDF isomers were not detected in the milled sample and
as explained earlier may be the result of a shifting of the retention time.
                                     258

-------
     Table 4 presents the PCDD/PCDF concentration observed In the hydroid
biomass collected from both ash and concrete reference blocks on June 8,
1989, after 258 days of submersion.  The only PCDD/PCDF isomer detected
above detection limits in the hydroid biomass samples that were removed from
both the stabilized ash blocks and concrete control blocks was OCDD,  In the
biomass sample collected from the concrete blocks, over 1 ng/g of OCDD was
detected.  It is not unreasonable for OCDD to be detected when no other
isomers are observed since it is the most widely dispersed of all the PCDD's
and PCDF's in the environment.

     The detection limits obtained for the biomass samples were limited by
the GC/MS sensitivity.  The results do, however, place an upper limit on the
amount of PCDD's/PCDF's in the biomass.  The higher detection limits
obtained may have been caused by the low mass of the samples and/or low
recoveries in some samples.  In addition, the extraction method used for the
biomass samples has been validated for terrestrial samples but not for
marine organisms.

                                CONCLUSIONS
     Particulate incineration residues, when combined with Portland cement,
can be successfully stabilized into solid blocks using conventional block
making technology.  The establishment of the Conscience Bay artificial reef
site constructed using stabilized incineration residues has provided a
unique opportunity to study the in situ interactions of stabilized
incineration residue blocks with the marine environment.  Results of this
study, though not reported in this paper, have shown that the stabilized
incineration residue blocks have retained their strengths even after
prolonged seawater exposure.  PCDD's and PCDF's are retained within the
cementitious matrix of the stabilized incineration residue blocks after 7
months of submersion.  In addition, the organisms growing on the surfaces of
the stabilized incineration residue blocks are not accumulating PCDD's or
PCDF's from the blocks.

     To date, no adverse environmental impacts have been observed at the
Conscience Bay reef site due to the presence of stabilized incineration
residue blocks.  Continued monitoring of the reef site is planned and will
provide data to more clearly define the long-term effects of stabilized
incineration residue in the marine environment,

                                REFERENCES
1.   Breslin V.T., Roethel, F.J. and Schaeperkoetter,  V. P.  1988.  Physical
     and Chemical Interactions of Stabilized Incineratin Residue in the
     Marine Environment.  Marine Pollution Bulletin.  19:11, 1988.

2.   Parker, J.H., Woodhead, P.M.J. andDuedall, I.W.   Coal-Waste Artificial
     Reef Program, Phase 3; Vol. 2: Comprehensive Report. EPRI Report
     CS-2009, Electric Power Research Institute, Palo Alto, California,
     1981.  404 pp.

                                     w

-------
3.   Kalajian, E.H., Duedall, I.W., Shieh, C.S. and Wilcox, J.R.  In:
     Proceedings of the Fourth International Conference on Artificial
     Habitats for Fisheries. November 2-6, 1987. Miami, Florida.

4.   Roethel, F.J., Schaeperkoetter,  V., Park, K. and Gregg, R.  The
     fixation of incineration residues: physical and leachate properties.
     Final report to the New York State Legislative Commission on the
     Water Resource Needs of Long Island.  Working Paper No. 26, MSRC, SUNY
     at Stony Brook, New York, 1986.   200 pp.

5.   Park, K.  Leaching behavior of incineration residues from
     municipal solid wastes.  M.S. Thesis, SUNY at Stony Brook, Stony
     Brook, New York, 1987.  90 pp.

6.   Roethel, F.J.  Interactions of stabilized power plant coal wastes
     with the marine environment.  Ph.D. Dissertation,  MSRC, SUNY at Stony
     Brook, New York, 1981.  349 pp.

7.   O'Keefe, P.W., Smith, R.M.,  Hilker, D.R., Aldous,  K.M. and Gilday,
     W.  In:  Chlorinated Dioxins and Dibenzofurans in the Total Environment
     II, (eds.) Keith,  L., Rappe, C.,  and Choudhary, G.,  Butterworth
     Publishers, MA.  1985.  111-124.

8.   Valente, H.,  IBM-XT Processing of Selected Ion Monitoring Data Files
     from a Hewlett-Packard 5970 Mass Selective Detector, Journal of
     Automatic Chemistry,  £(4)-.158-165, 1987.

9.   LIRPB.  Ash Characterization Data Inventory - Task 1.2.  Long Island
     Regional Planning Board, Hauppauge, New York 1987.  97 pp.
                                     258

-------
    TABLE 1.  IONS MONITORED FOR PCDD's AND PCDF's.
Group
Compound
                         Ions
                              Time
  DBF
  MCDF
  DCDF
  DD
  MCDD
168.05,139.05
202.00,204,00,139.05
236.00,238,00,173.00
184.00,155.05
218.00,220.00,155.05
                                                7.01-18.0 min
13
  DCDF
  TrCDF
  TCDF
  DCDD
  TrCDD
  C-2-DCDD

  TCDF
  PeCDF
.  TCDD
  C  -TCDF

37C12-TCDD
  Clf-TCDD
270,00,272.00,209.00
305.90,303.90,242.90
252.00,254.00
286.00,288.00
264.00,266.00
339,90,341.90,276.90
321.90,319.90,258.90
317.90,315,90
333,90,331.90
327.90,262.90
                                                18.0-35.0 min
                                                35.0-47.0 min
  PeCDF
  PeCDD
 *C  -PeCDF
  CJ^-PeCDD

  HxCDF
I""C  -HxCDF

   12"HxCDD

  HpCDF

13«PCDD
13
  OCDF
  OCDD
     -OCDD
355.90,357.90,292.90
351.90,353.90
367.90,369.90

373,80,375.80,312.85
389,80,391.80,328.90
385.85,387.85
401.90,403.90

407.80,409.80,344.80
423.80,425.80,360.80
419.80,421.80
435.80,437.80

443.80,441,70,378.80
459,70,457.70,394.80
455.80,453.70
471.80,469.80
                                                47.0-60.0 min
                                                60.0-72.0 min
                                                72.0-82.0 min
                                                82.0-95.0 min
                           259

-------
     TABLE 2.  PCDD/PCDF CONCENTRATIONS IN COMBINED INCINERATION ASH
ANALYTE
INCINERATION ASH
SAMPLE 1
DET. LIMIT
2378 TCDD
12378 PCDD
123678 HXCDD
123789 HXCDD
123478 HXCDD
1234678 HPCDD
12346789 OCDD
2378 TCDF
12378 PCDF
23478 PCDF
123478 HXCDF
123678 HXCDF
234678 HXCDF
123789 HXCDF
1234678 HPCDF
12346789 OCDF
4.7
7.3
14
12
15
20
44
4.2
5.0
6.9
7.1
8.1
11
11
12
30
CONG.
17
19
48
65
26
450
1400
0
44
54
80
57
68
0
240
180
SAMPLE 2
DET. LIMIT
0
0
1
1
1
3
18
0
0
0
0
1
1
1
3
12
.38
.51
.3
.1
.4
.2

.32
.37
.49
.97
.1
.2
.1
.4

CONG.
5.3
17
31
42
17
340
1100
69
32
17
39
66
69
3.7
180
68
CONCENTRATIONS EXPRESSED IN pg/g
                                   260

-------
                        TABLE 3.   PCDO/PCOF CONCENTRATION*  IN  STABILIZED INCINERATION ASH REEF BLOCKS
ANALYTE
STABILIZED INCINERATION ASH BLOCKS
REEF BLOCK T*0
REEF BLOCK T=192
SAMPLE 1

2378 TCDO
12378 PCDD
123678 HXCOD
123789 HXCDD
123478 HXCDD
1234678 HPCDO
12346789 OCDD
2378 TCDF
12378 PCDF
23478 PCDF
123478 HXCDF
123678 HXCDF
234678 HXCDF
123789 HXCDF
1234678 HPCDF
12346789 OCDF
DET. LIHIT
0.6
0.9
1.7
1.6
1.9
2.4
6.7
0.52
0.73
0.97
0.95
1.1
1.2
1.2
1.7
4.3
CONC.
6.1
14
26
34
17
230
850
66
30
31
48
34
33
2.4
140
57
DET. LIHIT
0.32
0.45
1.2
1.0
1.4
2.1
6.7
0.27
0.34
0.45
0.83
0.93
1.1
0.93
2.4
4.1
CONC.
5.5
14
25
37
15
270
850
59
27
12
37
57
33
3.4
150
56
REEF BLOCK T=192
SAMPLE 2
DET. LIMIT
0.29
0.44
1.7
1.6
1.9
3.7
18
0.28
0.33
0.44
2.7
3.0
3.5
3.3
5.8
12
CONC.
4.7
14
24
42
8.9
240
900
46
27
7,9
39
7i
110
10
140
57
REEF BLOCK T=192
MILLED
DET. LIMIT
7.0
13
16
15
20
2.4
34
6.4
'7.9
12
8.8
10
12
11
13
24

CONC.
6.6
16
30
51
21
230
1300
0
35
53
71
39
49
0
180
76
CONCENTRATIONS EXPRESSED IN  pg/g

-------
                                      TABLE 4.  PCDO/PCOF COHCEHTRAT10HS*  IN  HYOROIO  BIOHASS
ro
s
AHALYTE


2378 TCDD
12378 PCDD
123678 HXCDD
123789 HXCDD
123478 HXCDD
1234678 HPCDD
12346789 OCDD
2378 TCDF
12378 PCDF
23478 PCDF
123478 HXCOF
123678 HXCDF
234678 HXCDF
123789 HXCDF
1234678 HPCDF
12346789 OCDF
BIOMASS REHOVEO FROH
SAMPLE
DET, LIH J
17
25
44
41
49
100
1
;ONC.
0
0
0
0
0
55
520 1100
17
17
24
24
28
34
31
55
320
15
0
0
0
0
0
0
0
0

DET.
43
93
150
140
180
350
1100
72
25
35
26
27
36
32
39
100
CONCRETE BLOCKS
SAHPLE 2
LIH CONC.
0
0
0
0
0
0
870
0
0
0
0
0
0
0
0
0

DET.
43
91
120
120
150
270
620
81
25
36
24
28
35
29
35
71
SAHPLE 3
LIH CONC.
0
0
0
0
0
0
1400
0
0
0
0
0
0
0
0
110
BIOHASS REMOVED FROH
SAHPLE
DET. LIH (
22
82
300
280
340
2500
0
41
10
14
25
28
33
31
44
97
1
:oHC.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
INCINERATION ASH BLOCKS

DET.
150
230
360
330
420
540
2000
140
140
200
150
170
220
190
250
840
SAHPLE 2
LIH CONC.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
           CONCENTRATIONS EXPRESSED  IN  pg/g

-------
                 WATER  MOVEMENT THROUGH AN  EXPERIMENTAL  SOIL  LINER

             I.G. Krapac,  K.  CartwrigFvt, S.V.  Panno,  B.R.  Hensel,
                       K.H. Rehfeldt,* and B.L.  Herzog

                       Illinois State Geological Survey
                         "Illinois State Water Survey
                                Champaign, IL 61820
                                   ABSTRACT
      A field-scale soil liner was constructed to test whether compacted soil
barriers in cover and liner systems could be built to meet the U.S. EPA
saturated hydraulic conductivity requirement (< 1 x 10"7 cm/s).  The 8 x 15 x
0.9 m liner was constructed in 15-cm compacted lifts using a 20,037-kg pad-
foot compactor and standard engineering practices.  Water infiltration into
the liner has been monitored for 1 year.  Monitoring will continue until water
breakthrough at the base of the liner occurs.  Estimated saturated hydraulic
conductivities were 2.5 x 10"9, 4.0 x 10"8, and 5.0 x 10~8  cm/s, based on
measurements of water infiltration into the liner by large- and small-ring
infiltrometers and a water balance analysis, respectively.

      Also investigated in this research effort was the variability of the
liner's hydraulic properties and estimates of the transit times for water and
tracers.  Small variances exhibited by the small-ring flux data suggested that
the liner is homogeneous with respect to infiltration fluxes. The predictions
of water and tracer breakthrough at the base of the liner ranged from 2.4 to
12.6 years, depending on the method of calculation and assumptions made.  The
liner appeared to be saturated to a depth between 18 and 33 cm at the end of
the first of year of monitoring.  Transit time calculations cannot be verified
yet, since breakthrough has not occurred.  The work conducted so far indicates
that compacted soil barriers can be constructed to meet the saturated
hydraulic conductivity requirement established by the U.S. EPA.


                                 INTRODUCTION
      Despite an increased emphasis on recycling, waste reduction, and waste-
to-energy conversion for the alleviation of the waste disposal crisis, land
burial of solid wastes will continue to be an integral part of any waste
management plan (1).  In land burial schemes, compacted soil barriers with low


                                      283

-------
hydraulic conductivity are commonly used in cover and liner systems to control
the movement of liquids and prevent groundwater contamination. Little research
has been done to evaluate the effectiveness of field-scale compacted soil
barriers in retarding the movement of water and leachates. Soil liners have
come under scrutiny by the regulatory and scientific communities because the
hydraulic conductivity of these liners has often been greater than expected
(2).  Although laboratory tests have indicated that liner materials can
achieve the hydraulic conductivity required by the U.S. EPA, studies that
compared field and laboratory data showed that laboratory measurements can
underestimate hydraulic conductivity (3,4,5).

      This study was initiated to provide data on the performance of a field-
scale soil liner using in-situ measurements of construction (e.g., soil
density and moisture content) and hydrogeologic parameters (e.g., infiltration
rates and moisture movement).  This paper summarizes the hydrologic findings
from monitoring data collected during the first year of a long-term project
involving an experimental liner.


                        EXPERIMENTAL  DESIGN AND  METHODS
LINER CONSTRUCTION AND INSTRUMENTATION


      The Batestown Till Member of the Wedron Formation, an illitic glacial
till, was selected as the material with which to construct an experimental
field-scale liner.  The selection was based on the physical properties of the
till (Table 1) and on preliminary field studies that indicated the till could
be compacted such that the saturated hydraulic conductivity was less than 1 x
10"7 cm/s (6,7).  Rigid quality control practices were applied during liner
construction to ensure uniform moisture content and a maximum clod size less
than 15 cm in diameter.  The liner measures 8 x 15 x 0.9 m, and was
constructed in 15-cm compacted lifts using a 20,037-kg pad-foot compactor and
standard engineering practices (8).  A shelter was built over the liner to
maintain a controlled environment (Figure 1).

      The liner was compacted at an average moisture content of 11.5%, 1.5%
wet of optimum as determined by the Standard Proctor test.  The mean dry
density of the liner was 1.84 g/cm3,  93% of the maximum  Standard Proctor
density.

      A pond above the liner was filled in April 1988 and is maintained at a
depth of 29.5 ± 0.2 cm.  Evaporation pans in the pond are used to determine
evaporative water loss, a parameter needed for the calculation of a mass water
balance for the entire liner. When breakthrough occurs,  drainage will be
collected in an underdrain system and by pan lysimeters  installed beneath each
quadrant of the liner.

      Four large-ring (1.5-m diameter) and 32 small-ring (28-cm diameter)
infiltrometers are used to monitor water infiltration into the liner.


                                      264

-------
TABLE 1.  PHYSICAL AND  CHEMICAL PROPERTIES OF THE BATESTOWN TILL
Ks (laboratory)
      recompacted
Ks (field)
      prototype
5.1 x 10'9 ± 1.0 x  10*9 cm/s

3.6 x ID"8 cm/s
Standard Proctor Compaction
Max. Dry Density         1.98 ± 0.00 g/cm3
Opt. Moisture Content    9.9  ±0.41
Particle-Size Distribution
Clay        28.9 ±  2.8% (<  4/wn)
Silt        33.4 ±  2.5% (>4/im to <63/im)
Sand        32.0 ±  1.2% (>63/«n to <2mm)
Gravel       5.8 ±  2.1% (>2mm)

Specific Gravity    2.74 ± 0.01 g/cm3
                        Atterberg Limits
                        Liquid  Limit
                        Plastic Limit
                        Plastic Index
23.3 ± 0.5%
13.1 ± 1.1%
10.1 + 1.2%
                        Clay  Mineral  Composition
                        Illite             63.7 + 0.6%
                        Chlorite            9.7 ± 0.6%
                        Expandables        26.6 ± 0,6%
    South
                                                                        1.5m
                                               North
          soil liner
                                                       drainage pit
                                        pan lysi meter
Figure 1.  Cross  section  of the compacted soil liner, shelter, and
           accessory  components.
                                       265

-------
Tracers  (bromide,  o-trifluoromethylbenzole acid,  m-trifluoromethylbenzoic
acid, or pentafluorobenzoic  acid)  were added to the large-ring infiltrometers.
Pressure head  is monitored using tensiometers installed in 10 nests of six
instruments per nest.   Each  instrument within a nest is located at a different
depth in the liner.  Samples of soil  water are collected using suction cup
lysimeters located in  10  nests  of  six lysimeters  (Figure 2).  Gypsum blocks
were installed in  the  liner  but did  not function  properly because the liner
matric potential was below the  working range of the instruments.   Further
details on liner construction and  instrumentation can be found in Krapac et
al. (8).
DATA ANALYSIS
      Cumulative  infiltration  curves  for each infiltration ring were used to
determine steady-state  infiltrability;  cumulative infiltration volume was
plotted and regressed with  respect  to time since the liner pond was filled
(Figure 3).  The  slope  of the  linear  regression divided by the cross-sectional
area of the infiltrometer represented the "average"  steady infiltration flux
for a 1-year period.  The saturated hydraulic conductivity was calculated from
the infiltration  flux and measured  hydraulic  gradient.

      A vertical  hydraulic  gradient for each  month since the liner pond was
filled was calculated from  tensiometer  data.   Monthly gradients were
calculated by plotting  head at six  different  depths  in  the liner against
elevation.  The data were then analyzed using linear regression; the slope of
the regression line represented the hydraulic gradient  for the entire liner.
                     retaining wall
                              large infiltrometer
                         © /      ©
                                                    @)
                                     catwalk
             A AAAA
                                           ®
            O small Infittrometer
            A tensiometer

            ® evaporation pan
      A

i gypsum block

i lyslmeter
  A


0   2
4 m
Figure 2.  Instrumentation design  of  the  soil  liner.
                                       266

-------
      Analytical and numerical methods were used to predict breakthrough times
of water and tracers at the base of the liner. The saturated hydraulic
conductivity of the liner was also estimated using these methods.
                            RESULTS AND DISCUSSION
GRADIENTS AND INFILTRATION
      The monthly hydraulic gradients ranged from a low of 1.05 in August 1988
to a high of 1.72 in April 1989 with a mean of 1.5 (Figure 3).  When the liner
reaches steady state, the overall gradient based on the depth of the pond and
thickness of the liner will be approximately 1.3.  The variation in gradient
was related to head changes at various depths in the liner.  These head
changes were a reflection of soil tension deviations in the liner resulting
from long-term atmospheric pressure and temperature fluctuations.

      Changes in atmospheric pressure can affect soil pore-water pressure when
some of the pores contain entrapped air, a situation likely in the liner.  An
increase in atmospheric pressure can compress entrapped air, and a pressure
decrease can allow entrapped air to expand and occupy more pore space (9, 10,
11).  A decrease in air volume, caused by increased atmospheric pressure,
would increase tension.  This effect may be more apparent in fine-grained
materials (11).
                           100          200          300
                             Time Since Ponding (days)
                                                               400
Figure 3.  Hydraulic gradients in the liner and cumulative infiltration from
           April 1988 to April 1989 as determined by pressure transducer
           tensiometers and a water balance analysis.
                                      267

-------
      A positive relationship has been shown between temperature change and
pressure head in soil systems (12, 13), and was observed in this study.   Turk
(11) suggested that temperature changes have two long-term effects on soil
moisture: (1) surface tension is affected in the capillary pores near the soil
surface, resulting in drainage of pores when temperature increases; (2)
entrapped air can expand or contract-an increased temperature decreases
tension.

      The average infiltration fluxes were 7.9 x 10"8 cm/s and 5.0 x 10"9 cm/s
for the small-ring (SR) and large-ring (LR) infiltrometers, respectively.  The
infiltration fluxes of the 32 small-ring infiltrometers exhibited a low
variance (coefficient of variation = 2%), suggesting a relatively homogeneous
distribution of the infiltration rate throughout the liner.  The mean fluxes
of the two infiltrometer data sets were statistically different at a 99%
confidence level, as determined by a t-test.  The reason for the difference
between these two data sets is not known.

      Indirect evidence suggested several likely reasons for the lower
infiltration rates of the large rings: (1) water leaked between the grout-soil
interface, (2) the measurement capabilities were exceeded by the low hydraulic
conductivity of the liner material, and  (3) colonies of anaerobic bacteria
could have formed mats on the liner surface inside the rings, reducing the
permeability of the liner surface.

      A mass water balance for the 1-year period was used to determine the
overall flux of the liner.  During this period, 9860 liters of water was added
to the liner pond, and 6600 liters of water was estimated to have evaporated.
An estimated  3260 liters of water infiltrated into the liner.  After the area
of infiltration was corrected to account for the infiltrometers,  the average
flux for the entire liner was 1.0 x 10"7 cm/s.


HYDRAULIC CONDUCTIVITY ESTIMATES


      Darcy's law, the Green and Ampt infiltration approximation (14), and the
numerical model SOILINER (15) were used to calculate the saturated hydraulic
conductivity of the liner.  The three methods produced comparable results with
all conductivity values less than the U.S. EPA requirement of 1 x 10"7 cm/s
(Table 2).

      The application of Darcy's law requires the assumption that all flow
through the liner is saturated.  Darcy's law can be written as V = -K I, where
V is the specific discharge, K is the saturated hydraulic conductivity, and I
is the hydraulic gradient.  Solving for K produces K = -V/I; -V is the
measured steady-state infiltration flux.  The average gradient for the liner
after the first year of monitoring as determined from the tensiometer data was
1.5.  Hydraulic conductivity values for the liner were calculated from three
sets of infiltration data (Table 2).
                                      268

-------
TABLE 2.   HYDRAULIC CONDUCTIVITY VALUES DETERMINED FROM INFILTRATION DATA

Data
set
Small rings
Large rings
Water
balance
Infiltration
flux (cm/s)
7.9 x
5.0 x
1.0 x

lo-8
ID'9
io-7

Hydraulic Conductivity (cm/s)
Darcy's law
5.3 x 10"8
3.3 x 1CT9
6.7 x IO"8

Green -Ampt
4.0 x
2.5 x
5.0 x

io-8
io-9
IO"8

SOILINER
5.7 x
3.5 x
7.2 x

io-8
lO'9
10-*

      The Green and Ampt (14) soil infiltrability approximation differs from
Darcy's law in that of the depth of the wetting front must be known to
estimate a hydraulic conductivity.  This approximation technique assumes that
the wetting front is uniform and sharp, and has a constant matric potential.
The wetted zone is also assumed to have a uniform moisture content and a
constant hydraulic conductivity.  On the basis of these assumptions, the
analytical solution to vertical infiltration is

                  K . 1 [1+
                               LI
where the bracketed term is the hydraulic gradient, h is the ponding depth
(29.5 cm), ${  is the tension  at  the wetting  front  (conservatively estimated  to
be zero), L, is the depth to  the wetting front,  and  i  is the  steady-state
infiltration flux.  Tensiometer data suggested that the wetting front after 1
year of ponding was at a depth between 18 and 33 cm.  Table 2 shows the
saturated hydraulic conductivities based on the Green and Ampt assumption for
the same three data sets used in the Darcy approximation (wetting front depth
conservatively estimated to be 30 cm).  The conductivity values calculated
from Darcy's law were approximately 33% higher than those calculated by the
Green and Ampt method because the gradient term increased from 1.5 (Darcy
estimate) to approximately 2 (Green-Ampt estimate).

      SOILINER  estimated a saturated hydraulic conductivity by determining
the required model input conductivity that would generate a flux value equal
to that measured in the liner. The liner hydraulic gradient was computed by
the model based on the head values at the top and bottom of the liner, the
soil properties,  and the moisture characteristic curve of the compacted liner
soil.  For example, a flux of 7.9 x Iff* cm/s, as determined by the small-ring
infiltrometers, required the modeled soil  column (liner) to have a hydraulic
conductivity of 5.7 x 10"8 cm/s  (Table 2).
                                      26S

-------
TRACER TRANSIT TIME PREDICTIONS


      An objective of this project is to determine the accuracy of transit
time predictive methods.  Water did not break through at the bottom of the
liner during the first year of monitoring, nor have soil-water samples
contained detectable concentrations of the tracers.  After the first year of
monitoring, the predictions cannot be verified, only compared.

      The simple and modified transit time equations are analytical solutions
used to estimate the earliest breakthrough of tracers at the bottom of the
experimental soil liner (5).  The numerical models SOILINER (15) and CHEMFLO
(16) were also used to estimate tracer breakthrough times.  The conductivity
value (7.2 x 10"8 cm/s) used in the predictions was the highest value
estimated by the infiltration data sets (Table 2).  The effective porosity was
assumed to equal the total porosity as suggested by the U.S. EPA (5).  The
porosity (77) was calculated to be 0.33 using the average liner density (1.84
g/cm3)  and a particle density of 2.74 g/cm3.

      The simple transit time equation assumes that the liner has always been
saturated and drains freely at the bottom, that there is steady- state one-
dimensional flow, and that dispersion and adsorption are neglected.  Transit
time was calculated as
                  t „
                                    (h-fd)

where d is the liner thickness (90 cm), h is the depth of the liner pond (29.5
cm), v is the Darcian velocity, and all other parameters are as defined above.
On the basis of this equation, tracer transit time through the liner was
predicted to be 9.9 years.

      The modified transit time equation (17) includes suction potential at
the bottom of the liner, recognizing that the liner is not in a completely
saturated condition.  All other assumptions of the simplified transit time
equation remain the same.  Modifying equation [2] yields

                  t =  2_d
                               Ksat  (h+^+d)


where ^ is suction potential at the bottom of the liner (55 cm, per
tensiometer data).  The modified transit time method estimated tracer
breakthrough in approximately 6.7 years.

      SOILINER predicted the transit time of a non-adsorbed, non-diffused,
non-dispersed particle from the top of the liner to the base.  With input data
representative of the field-scale liner, a transit time of 12.6 years was
obtained.  As with the preceding transit time equations, total rather than
effective porosity was used.
                                      270

-------
      CHEMFLO was used to simulate the effects of dispersion and diffusion on
the transit time of a non-retarded particle.  With dispersion values of 9.0 to
0.2 cm and typical diffusion values for clayey materials (10"6 to 10'7 cm2/s;
18), transit times of 2,5 to 4.6 years were obtained.


WATER BREAKTHROUGH TIME PREDICTION


        The Green-Ampt infiltration model assumes piston flow and can predict
the time it will take for the wetting front to reach a prescribed depth in the
liner.  If the wetting front depth is assumed to be at the base of the liner,
the model predicts the time of water breakthrough,
                  '  } [Lf  -  (h + &)  In (1 +
All parameters are defined above, except that Lf  (the depth to the wetting
front at breakthrough) equals the liner thickness, 90 cm.  #f is  7 cm,  based
on tensiometer data just below the wetting front, and 9l  is 0.21.   A
breakthrough estimate of 2.4 years produced by using the largest hydraulic
conductivity from all the data sets represents the earliest time in which
water will exit the bottom of the liner.  The use of the Green-Ampt estimated
hydraulic conductivity for the entire liner (5.0 x 10  cm/s) produced  a
breakthrough estimate of 3.4 years.

      The tensiometer data suggested that the wetting front was at a depth of
about 30 cm after a year of ponding.  The tensiometer data reflect the actual
liner performance and suggest that water should break through in approximately
3 years if the rate of water movement in the liner is assumed to be constant
with respect to time.  The Green-Ampt breakthrough estimates of 2.4 to 3.4
years for the wetting front to reach the bottom of the liner appears to be
consistent with actual data.


                            SUMMARY  AND  CONCLUSIONS
      On the basis of 1 year of monitoring data, soil liners probably can be
constructed to meet the saturated hydraulic conductivity requirement
established by the U.S. EPA.  Rigid quality control practices will have to be
followed during construction if soil liners are to achieve performance
specifications.  Estimates of the saturated hydraulic conductivity of the
liner using either Darcy's law or the Green-Ampt infiltration model and the
small-ring infiltration data ranged from 2.5 x 10"9 to 6.7 x 10"8 cm/s.  The
numerical model SOILINER estimated a saturated hydraulic conductivity of 5.7 x
10~8 cm/s based on measured fluxes; this value differs less than one order of
magnitude from those conductivities derived from either the small-ring
infiltrometers or mass water balance approach.  The infiltration data from the


                                      271

-------
large- and small-ring inflltrometers suggest that Infiltration fluxes were
relatively homogeneous throughout the liner; however, the large-ring
inflltrometers had significantly smaller infiltration fluxes than did the
small-ring infiltrometers.

      Transit time predictions based on two analytical methods and one
numerical method indicated that a non-dispersed, non-diffused tracer will exit
the bottom of the soil liner between 6.7 and 12.6 years after ponding.  If
diffusion and dispersion were taken into consideration, tracer breakthrough
could occur in 2.5 to 4.6 years.  According to tensiometer data, water
breakthrough should occur in approximately 3 years.  During the first year of
monitoring soil water in the liner no tracers have been detected, and no water
has been collected at the bottom of the liner; therefore these predictions
cannot be verified.


                                ACKNOWLEDGMENTS
      This work was funded in part by the U.S. Environmental Protection
Agency, Risk Reduction Engineering Laboratory, Cincinnati, Ohio, through
Cooperative Agreement CR-812650,  Additional support was provided by the
Illinois Hazardous Waste Research and Information Center and the Illinois
State Geological Survey.


                                  REFERENCES
1.    Birks, D.  The garbage route: choosing all the options.  World Action
      for Recycling Materials and Energy from Rubbish (Warmer) Bulletin.  23:
      10,  1989.

2.    Daniel, D. E. and Brown, K. W.  Landfill Uners: How well do they work
      and what is their future?  In:  J.R. Gronow, A. N. Schofield and R. K.
      Jain (eds.), Land Disposal'of Hazardous Waste: Engineering and
      Environmental Issues, Ellis Horwood Limited, West Essex, England, p.
      235.

3.    Daniel, D. E.  Predicting hydraulic conductivity of clay liners.
      Journal of Geotechnical Engineering. 110: 28i, 1984.

4,    Hirzog, B. L. and Morse, W. J.  Hydraulic conductivity at a hazardous
      waste disposal site: Comparison of laboratory and field-determined
      values. Waste Management and Research.  4: 177, 1986.

5.    U.S. Environmental Protection Agency.  Design, construction, and
      evaluation of clay liners for waste management facilities.  EPA/530-SW-
      86-007F, U.S. EPA, Cincinnati, Ohio, 1988.
                                      272

-------
6.    Albrecht, K. A., and Cartwright, K.  Infiltration and hydraulic
      conductivity of a compacted earthen liner.  GroundWater.  27: 14,  1989,

7.    Albrecht, K, A., Herzog, B. L,, Follmer, L. R., Krapac,  I. G., Griffin,
      R. A., and Cartwright, K,  Excavation of an instrumented earthen liner:
      Inspection of dyed flow paths and soil morphology.  Hazardous Waste and
      Hazardous Materials.  5: 269,  1989.

8.    Krapac, I. G., Cartwright, K., Albrecht, K., Brutcher, D. F.,
      DuMontelle, P. B., Follmer, L. R., Griffin, R. A., Hense'l, B. R.,
      Herzog, B. L., Larson, T. H., Miller, K. W., Panno, S. V., Rehfeldt, K.
      H., Risatti, J. B., Su, W.  Field study of transit time through
      compacted clays.  Draft Final Report, U.S. EPA, Cincinnati, Ohio, 1989.
      241 pp.

9.    Peck, A. J.,  The water table as affected by atmospheric pressure.
      Journal of Geophysical Research.  65: 2383, 1960.

10.   Norum, D.I,, and Luthin, J. N.  The effect of entrapped air and
      barometric fluctuations on the drainage of porous mediums.  Water
      Resources Research.  4: 417,  1968.

11.   Turk, L. J.,  Diurnal fluctuations of water tables induced by
      atmospheric pressure changes. Journal of Hydrology.  26: 1,  1975.

12.   Smedema, L. B., and Zwerman, P. J,  Fluctuations of the phreatic
      surface, 1. Role of entrapped air under a temperature gradient.
      Soil Science.  103: 354,  1967.

13.   Peck, A.J.  Change of moisture tension with temperature and air
      pressure, theoretical.  Soil Science.  89: 303,  1960.

14.   Green, W. H., and Ampt, G. A.  Studies on soil physics:  I. Flow of air
      and water through soils.   Journal of Agricultural Science. 4:  1, 1911,

15.   Johnson, R, A., Wood, E. S., Wood, R. J., and Wozmak, J.  SOILINER
      model-documentation and user's guide (version 1): EPA/530-SW-86-
      006, Public Comment Draft, 1986.  193 pp.

16.   Nofziger, D. L., Rajender, K., Nayudu, S. K., and Su, P. Y.  CHEMFLO:
      One-dimensional water and chemical movement in unsaturated soils.
      EPA/600/8-89/076,  U.S. EPA, Cincinnati, Ohio, 1989, 106 p.

17.   Cogley, D. R., Goode, D. J., and Young, C. W.  Review of the transit
      time equation for estimating storage impoundment bottom liner
      thickness: Appendix A. In Goode and Smith (eds.).  EPA/530-SW-84-
      001, U.S. EPA, Cincinnati, Ohio,  1984.

18.   Freeze, R. A. and Cherry, J. A.  Groundwater.  Prentice-Hall, Inc.
      Englewood Cliffs, NJ, 1979.  604 p.
                                      273

-------
           HAZARDOUS AND TOXIC WASTES ASSOCIATED WITH URBAN
                            STORMWATER RUNOFF
             Robert E. Pitt, P.E., Ph.D., Department of Civil Engineering,
                       University of Alabama at Birmingham
           Richard Field, P.E., Chief, Storm and Combined Sewer Pollution
                  Control Program, U.S. EPA, Edison, New Jersey


ABSTRACT

       This paper summarizes some of the Information obtained during a research
project sponsored and directed by the EPA's Storm and Combined Sewer Research
Program and conducted under a subcontract from Foster-Wheeler/Enviresponse. The
research project examined a variety of organic and metallic toxicants in stormwater and
combined sewer overflow (SCSO) source flows {Pitt and Barren 1990). The study was
designed to use rain events and sample locations to illustrate the variables associated
with toxicant concentrations in urban runoff. An attempt was made to specifically
address the following questions:

        1.   What are the typical toxicant contaminant levels in stormwater?
        2.   What are the origins of these toxicants in stormwater?
        3.   What rain or land use factors affect toxicant concentrations in
             stormwater?
INTRODUCTION

       Stormwater runoff has been identified as a major contributor to the degradation
of many urban streams and rivers (Field and Turkeltaub 1981; Pitt and Bozeman 1982;
Pitt and Bissonnette 1984). Organic and metallic toxicants are expected to be
responsible for much of these detrimental effects, and have been found in urban runoff
discharges during many previous studies (EPA 1983; Hoffman, et al. 1984; Fam, et ai.
1987; Pereira, et al. 1988).
       Table 1 summarizes the estimated discharges of commonly detected organic
and metallic toxicants from all U.S. cities having populations greater than 100,000
population (which total about 15,000 square miles, Dept of Commerce 1980). These
cities will be required to participate in the EPA's stormwater permit program (Federal
Register, December 7,1988). These values are for discharges that are directly entering
the nation's surface receiving waters. This information is based on the Nationwide
Urban Runoff Program (NURP) results of about 100 stormwater outfall samples (EPA
1983). This NURP data is mostly for residential areas, with some commercial area
influences. More recent information  indicates that industrial stormwater discharges can
have many times the concentrations of the toxicants as the areas represented in the
NURP data (Pitt and McLean 1986). In addition, base flows occurring in storm drains
during dry weather that may be contaminated by non-stormwater discharges (such as
industrial waste cross-connections), can also significantly increase these estimated
loadings (Pitt, et al. 1990).  Therefore, the large discharges noted in Table 1 can be
expected to be even much larger, when all urban areas, land uses, and flow regimes
are considered. Most importantly, these are actual discharges as monitored at outfalls,
and not estimated discharges associated with chemical storage or disposal operations.

                                    274

-------
METHODOLOGY

       This research included the collection and analysis of several hundred urban
runoff samples from a variety of source areas and under different rain conditions. A
number of combined sewer overflow and detention pond samples were also included in
the evaluation portion of these first phase activities. This effort was significantly greater
than has been attempted previously for toxic pollutants in stormwater and will enable
several critical questions to be addressed, as stated previously in the objectives.
       Samples were analyzed for many organic pollutants using gas chromatographs
with a mass selective detector (GC/MSD) and with an electron capture detector
(GC/ECD) and metals using a graphite furnace equipped atomic adsorption
spectrophotometer (GFAA). All samples were also analyzed for particle distributions
from about 1 to 100 microns.  All samples were also analyzed using a toxicity screening
technique. All SCSO samples were also filtered to determine the liquid/solid partition
coefficients of the pollutants and the relative toxicities of the filterable and nonfilterable
portions of the samples. The following paragraphs  briefly summarize the sampling and
analyses features of this first phase research effort.

Sampling Effort

       About 300 subsamples were analyzed for organic and metallic toxicants, toxicity
screening, and particle size distributions. All of these samples were partitioned into
filterable and non-filterable  components for complete analyses.
       The relative importance of different source areas (such as roofs,  streets, parking
areas, etc.) in contributing toxicants were examined from field studies conducted as
part of this research. Samples were collected from the most significant potential source
areas in residential, commercial, and industrial land uses. The areas that received the
most sampling attention were parking  and storage areas in industrial and commercial
areas. These areas have been noted in previous studies to have the  largest potential of
discharging toxicants (Pitt and McLean 1986).
       Sheetflow samples were collected during five Birmingham Alabama rains.
Replicate samples from many of the same source areas, but during different rains,
enabled differences due to rain conditions versus site locations to be statistically
evaluated.

Source Area Runoff Grab Samples

       The sheetflow samples were collected using manual grab procedures. Hand
operated pumps created a vacuum in  the sample bottle which then drew the sample
directly into the container through Teflon tubes. About one liter of sample was collected,
split into two containers: one 500-mL glass with Teflon lined lid was used for the organic
and toxicity analyses, and another 500-tnL polyethylene bottle was used for the metal
and other analyses.
       Most of the source area sheetflow samples were obtained from the Birmingham,
Alabama area during the first phase of this project.  However, cooperative researchers
in Seattle, Washington also submitted  a limited number of additional  stormwater
samples for comparison.

CSO Grab Samples

       Twenty combined sewer overflow (CSO) outfall grab samples were collected in
the New York City area for complete analyses. These outfall samples were used to

                                     275

-------
 make a preliminary evaluation of the relative toxicities of CSOs and urban stormwater
 runoff. These samples were collected during four different rains in New York.


 Organic Pollutant Analyses

       The samples were analyzed using a Hewlett Packard 5890 gas chromatograph
 with a 5970B mass selective detector (GC/MSD), and a Perkin-Elmer Sigma 300 gas
 chromatograph with an electron capture detector (GC/ECD). The following lists the
 organic toxicants that were analyzed in these samples:

 Pesticides (detection limit: 0.3 ug/L):

       BHC, heptachlor, aldrin, heptachlor epoxide, endosulfan, DDE, DDD, DDT,
 endrin, and chlordane.

 Phthalate Esters (detection limit: 0.5 ug/L):

       Bis(2-ethyhexyl) phthalate, butyl benzyl phthalate, di-n-butyl phthalate, diethyl
 phthalate, dimethyl phthalate, and di-n-octyl phthalate.

 Polynuclear Aromatic Hydrocarbons (detection limit: 0.5 ug/L):

       Acenaphthene, acenaphthylene, anthracene,  benzo (a)  anthracene, benzo (a)
 pyrene, benzo (b) fluoranthene, benzo (ghi) perylene, benzo (k) fluoranthene,
 chrysene, dibenzo (a,h) anthracene, fluoranthene, fluorene, indeno (1,2,3-cd) pyrene,
 naphthalene, phenanthrene, and pyrene.

 In addition, selected nrtroaromatics, haloethers, and other chlorinated hydrocarbons
 were also analyzed.

 Metallic Pollutant Analyses

       The samples were analyzed using a Perkin-Elmer graphite furnace atomic
 absorption spectrophotometer (GFAA). Standard EPA approved methods were used in
 these analyses. Aluminum, cadmium, chromium, copper, lead, nickel, and zinc were
 analyzed in all samples. The detection limits were about 1 ug/L, except for cadmium
 which had a detection limit of about 0.1  ug/L.
       Low detection limits were necessary for these metal analyses.  In prior studies,
 the total forms of most of the metals were well within the detection limits of standard
 flame atomic absorption spectrophotometer (AAS) procedures, but the filterable
 portions were commonly not detected (Pitt and McLean 1986; EPA 1983). The
 partitioning of the heavy metals between the solid and liquid  phases is an important
 factor in determining the treatability of these pollutants and was therefore an important
 goal of this research. This information is needed to assess the fates of the metals in
 receiving waters and in treatment processes.

Toxic'rty Screening Tests

       A number of previous studies have found high concentrations of toxic pollutants
 in SCSOs. Some urban stormwater runoff studies attempted to use conventional 96-hr
fish bioassay toxicfty tests (such as Pitt 1979), but very few fish died during the tests.
 However, in situ taxonomic studies of urban runoff receiving waters found significant
 evidence of toxic effects from the long-term exposure to these pollutants (such as

                                    276

-------
reported by Pitt and Bozeman 1982, for the same stream as the negative fish bioassay
tests). More recent bioassay tests have used more sensitive organisms and have
detected significant SCSO toxicities (Spiegai et al. 1984 in Syracuse, NY; Mount et al.
1985 in Birmingham, AL; Mount et al. 1986 in Waterbury, ON; and Norberg-King et al,
1988 in San Francisco Bay).
       The objective of this task was to obtain toxietty measurements from a large
number of SCSO and source area samples, along with toxicity measurements
corresponding to different SCSO sample partitions. It was necessary to use a rapid
screening method to examine the relative toxicities for the different samples because of
the time and financial limitations of this first phase of the research program, A series of
special tests were made to compare toxicities of selected sheetflow and CSO samples
to both the screening method and conventional bioassay methods.
       The toxicity testing  procedure that was used (Microtox from Microbics, Inc.)
uses luminescent bacteria to indicate relative toxicities of samples. This procedure was
used to screen all of the samples collected during this project. The partitioned samples
(filterable and non-filtrable for each sample) were all tested for relative toxicity. These
data enabled toxicity comparisons between different source areas, in addition to toxicity
reduction potential for different treatment processes,  to be made. These tests were not
used to determine the absolute toxicities of the samples, but only to examine the toxicity
differences between the different source areas and sample partitions. In addition, about
twenty samples were also analyzed concurrently using a variety of conventional
bioassay techniques, for comparison with the Microtox procedure.

Particle Size Analyses

       Many SCSO treatment processes are very sensitive to the particle size
distributions and settling velocities of the solids (Dalrymple et al,  1975), Wet detention
ponds, catchbasins, grass filters, street cleaning, microscreening, filtration, swirl
concentrators are some of the treatment methods that require a knowledge of particle
size and/or settling characteristics. Additionally, the fate of many toxic pollutants in
receiving waters is also very sensitive to these particle physical characteristics. Without
knowing the specific particle size distributions and settling velocities, the necessary
design information for these controls therefore remains unknown.
       Unfortunately, there art wide variations in the particle distributions for different
source areas (Pitt and McLean 1986), which makes the design of runoff controls having
consistent performance difficult. The objective of this subtask was to obtain a
statistically significant number of individual particle size distributions from many SCSO
source areas.
       A laser particle counter (SPC-510 from Spectrex Corp.) was used to analyze
particle slzt distributions for all of the SCSO and receiving water samples. This
instrument produces particle size distribution plots for particle sizes ranging from 0.5
microns to more than 100 microns. Settling column tests are currently being conducted
to determine the specific gravities of SCSO samples which will enable settling velocities
to be calculated,

Treatablllty Tests

       This project also included  tests to examine the treatability of source area and
outfall samples. Filtration tests, In conjunction with literature information, enabled an
examination to be made of the benefits of typical treatment processes to reduce toxicity
and potential toxic pollutant components of SCSOs. This subtask stressed the fate
mechanisms (partitioning) that can be later related to specific control processes. As an
example, knowing the filterable fraction of the aggregate toxicity of a sample will allow
estimates to be made concerning the maximum treatability of the waste by particle

                                     in

-------
separation processes (such as catchbasins, grass filter strips, and wet detention
ponds). The detailed particle size distributions obtained in this research will enable the
relative benefits of various sediment barrier practices (such as filtration and screening)
to be estimated. Many detailed bench-scale unit processes will be performed during the
current phase of this study, and future work will include pilot- and full-scale tests of
various treatment practices.


DATA OBSERVATIONS

Toxicity  Observations

      The Microtox procedure allowed toxicity screening tests to be conducted on
each sample's total and filtered components. This screening procedure enabled about
300 samples to  be evaluated. The Microtox procedure was not used to determine the
absolute toxicity of the samples, or to show that urban stormwater runoff components
were In fact toxic. The objectives of these analyses were to identify the most toxic
source areas and to identify the approximate toxic reductions possible by complete
separation of the unfiltered pollutants from the mixtures.
      Actual urban stormwater runoff problems that have been monitored are quite
varied, but are probably mostly associated with long-term pollutant exposures,
especially through heavily polluted sediments. Receiving water concentrations during
runoff events and typical laboratory bioassay tests have not shown many significant
short-term receiving water problems (Field and Pitt 1990).
      Each sample was tested as unfiltered and filtered. A Millipore 0.45»micron filter
was  used, under a gentle vacuum, to prepare the filtered  samples. The toxicity, as
determined by the Microtox procedure, was expressed as three values, I} Q (the
percentage light decrease after about 10 minutes of exposure), (35 (the percentage
light decrease after about 35 minutes of exposure), and the EC$Q. The £€50 is the
sample dilution corresponding to a 50 percent light decrease after a 35 minute
exposure. Therefore, only samples that have 135 values greater than 50 were further
tested to determine the £650 values. Higher values of I-JQ ar|d '35' anc' lower fractions
of ECso, correspond to greater toxicities.
      Microbics suggests that light decrease values greater than 60 percent
correspond to "highly" toxic samples, light decrease values between 20 and 60 percent
correspond to "moderately" toxic samples, and light decrease values less than 20
percent correspond to "not" toxic samples. Table 2 shows the percentages of samples
In each category that corresponded to each of these groupings. Also shown on Table 2
are the numbers of samples analyzed in each source area category.
      The category having the largest percentage of highly toxic samples was the
combined sewer overflows. The urban creeks and detention pond effluents had the
largest percentage of samples that can be considered least toxic. The source areas
that  had the greatest toxic responses were the parking and storage areas.
      Tests were conducted  on unfiltered and filtered portions of each sample to
indicate the toxic reduction potential associated with complete separation of the
particulate pollutant  components. When the toxic responses of all of the samples were
compared, it was found that no significant differences in the toxic responses occurred
for the unfiltered versus filtered samples. In  many cases, the filtered samples actually
Indicated greater toxicities than their unfiltered counterparts. This was probably
because of normal experimental errors (found to be about 15 percent through
controlled testing).
      The chemical analyses found that significant portions of the monitored toxicants
were associated with the suspended solids (nonfilterable  residue). Upon sample
filtering,  the concentrations of the toxicants were generally greatly reduced. However,
                                     278

-------
as noted above, the Microtox toxicities of the samples were apparently little affected by
filtration. Either other nonmonitored toxicants were responsible for the indicated
toxicities and were mostly associated with filterable forms, or the toxicants associated
with the suspended  solids had little effect on the test organisms. The chemical analyses
did find significant portions of the toxicants associated with suspended solids that may
form highly toxic sediments in receiving waters. These sediments may adversely affect
receiving water  beneficial uses long after a runoff event.
       In summary,  the Microtox analyses indicated short-term toxicities associated
with filtered pollutants that would not be removed through sedimentation processes. In
contrast, the chemical analyses found significant toxicant concentrations associated
with sediment forming materials that would affect long-term toxic responses and these
could be partially removed through sedimentation and other particle separation
processes.

Suspended Solids Analyses

       Suspended solids (particulate residue), turbidity, pH, and particle size
distributions were obtained for the unfiltered portions of the samples. The runoff from
the paved areas all had relatively low suspended solids concentrations (generally less
than 100 mg/L), while some of the sheetflows from unpaved areas had concentrations
as high as 750 mg/L. The turbidity values varied in a similar manner; they were all quite
low, except for the unpaved areas. Except for roof runoff and storage area runoff, the
pH values were within a typical range of about 7 to 8.5. They were as low as 4.4 for roof
runoff and as high as 11.6 for storage area runoff. The samples representing complex
mixtures of source areas (urban creeks, detention ponds, and CSOs), all had pH values
closest to 7.
       For any one sample, the particle size distributions were generally narrow; the 10
to 90 percent ranges were represented by particle sizes as close as 20 microns apart.
The smallest particle sizes were found for roof runoff. In contrast, landscaped areas and
loading docks had some of the largest particle sizes found.

Organic Toxicant Concentration Observations

       A major portion of the effort of this research project was spent in conducting the
organic toxicant analyses. Table 3 summarizes the organics that were observed in at
least ten percent of the unfiltered samples analyzed. Most of the organic compounds
detected were PAHs. Two ethers were also frequently detected. This list is similar to the
frequency of detection list prepared by the EPA (1983) as part of the  Nationwide Urban
Runoff Program.
       Table 4 contains all of the observed base neutral data, while Table 5 contains all
of the observed pesticide data. Roof runoff, urban creeks, and the CSOs had the
greatest number of observed maximum  organic toxicant values. As noted previously,
the CSO category had the largest percentage of highly toxic samples. The roofs
contained high concentrations of several pesticides, fluroanthenes and a pyrene. A
CSO sample had an extremely high bis (2-ethly hexyl) phthalate concentration of 56
mg/L. Vehicle service areas and parking areas also had several of the observed
maximums.

Heavy Metal Concentration Observations

       Table 6 summarizes the heavy metal observations. In contrast to the organic
analyses, the detection frequencies for all of the metals were very high. Roof runoff had
the highest concentrations of zinc, probably associated with galvanized metal. Parking
areas had the highest nickel concentrations, vehicle service areas had the highest

                                     279

-------
cadmium and lead concentrations, while streets had high aluminum concentrations.
Surprisingly, landscaped areas had the highest chromium and urban creeks had the
highest copper concentrations.
       Many observations of fitterabie metais were also made and are also summarized
on Table 6. Except for storage areas, most of the zinc was associated with the filterable
sample partitions. In contrast, very little of the nickel was found in the filterable sample
partitions. Most of other metals were also found associated with the suspended solids
fraction. Therefore, suspended solids separation processes would be very effective in
removing heavy metals from these source areas, with the exception of zinc. Similarly, if
the metals were not removed before discharge, they would likely contribute to polluted
sediments in the receiving waters.


CONCLUSIONS

       The following paragraphs summarize the major project conclusions, as they
related to the project objectives.

Objective 1: Characterization of Toxic Components in SCSQs

       Overall, about 300 sample components were analyzed to determine toxicant
concentrations in sheetflows and other SCSOs as part of the first phase of this project.
       Most pH values were In a narrow range of 7 to 8,5 and the suspended solids
concentrations were generally less than 100 mg/L The particle size ranges were usually
narrow for any one sample, but the distribution ranges developed using all samples
from a single source area were substantially greater.
       Only a small fraction of the toxic organic pollutants analyzed were frequently
detected. Thirteen organics were detected in more than ten percent of all samples
analyzed. The greatest detection frequencies were for 1,3-dichlorobenzene and
fluoranthene, which had detection frequencies of 23 percent. The organics most
frequently found in these samples were similar to organics most frequently detected In
prior studies conducted elsewhere and were mostly the PAHs, especially fluoranthenes
and pyrenes.
       Roof runoff, urban creeks, and CSO samples had the greatest frequencies of
detection for the organic compounds analyzed. Vehicle service areas and parking
areas had several of the observed maximum organic compound concentrations
observed. Very little evidence was obtained to differentiate the solid/liquid partitioning of
orgsntcs for different source areas.
      The detection limits of the analyses were greater than anticipated and the
frequency of detection was therefore less than if the detection limits were Improved. The
use of larger sample volumes would have reduced the detection limits, which would
result In substantially greater detection frequencies.  Most of the organics were
associated with unfilterable sample portions. In contrast to the organics, the heavy
metals were detected In almost all samples analyzed, Including the filtered samples.
       Roof runoff had the highest concentrations of zinc observed,  probably due to
galvanized roof drainage components. Parking areas had the highest nickel
concentrations, while vehicle service areas had the highest concentrations of cadmium
and lead. Urban creek samples had the highest copper concentrations,  probably due
to illicit discharges.

Objective 2: Relative Toxlcitles of Sheetflows and SCSOs

      The toxlcity and chemical tests were not conducted to demonstrate the toxicity
of urban runoff. Many actual receiving water studies (summarized in the  project report

                                    280

-------
Pitt and Barron 1990; and by Field and Pitt 1990) identified the various problems that
have been associated with urban stormwater discharges. These long-term receiving
water studies have demonstrated that actual urban stormwater problems are quite
varied, and are probably mostly associated with long-term exposures to toxicants,
especially in the sediments, and to habitat destruction associated with high flows and
debris. Receiving water concentrations during runoff events and typical laboratory
bioassay tests have not indicated many significant short-term problems associated with
urban stormwater runoff.
       The Microtox screening tests found that CSOs had the greatest percentage of
samples considered the most toxic, followed by samples obtained from parking and
storage areas. Runoff from paved areas all had relatively low suspended solids
concentrations and turbidities, especially compared to samples obtained from unpaved
areas.
       About 15 percent of all of the unfiltered samples analyzed were considered
highly toxic using the Microtox screening procedure. The remaining samples were
approximately evenly split between being moderately toxic and not being toxic.
       Preliminary data evaluations indicated that variations in observed Microtox
toxicities and organic toxicant concentrations may be greater for different rains than for
the different source areas sampled. As an example, high concentrations of PAHs were
mostly associated with long-antecedent dry-periods.

Objective 3: Partitioning of Toxic Components and Treatability of Toxlcanls in SCSOs

       There were no significant differences in the measured  Microtox toxicities
associated  with the unfiltered samples and the filtered portions of the samples.
However, most of the organics and metals were associated with the suspended solids
of the runoff samples. An exception was for zinc, which was found mostly in the filtered
sample portions. This implies that most of the Microtox measured toxicity was
associated  with filterable forms of the pollutants.
       This preliminary result suggests that simple sedimentation,  even though
removing much of the mass of toxic pollutants from the water, may have minimal
benefits in reducing immediate toxicant effects. Uncontrolled  sedimentation, such as in
lakes or reservoirs, or large rivers, may result in long-term contaminated sediment
problems. However, controlled sedimentation in SCSO control devices allow effective
residue management, including appropriate disposal of the potentially heavily
contaminated sediments, which would minimize downstream receiving water sediment
problems.
       The literature review of the potential transport and fate mechanisms of these
pollutants found that many processes will affect these pollutants. Sedimentation in the
receiving water is the most common fate mechanism because many of the pollutants
investigated are mostly associated with particulate matter. Exceptions included zinc and
1,3-dichlorobenzene which were mostly associated with the filterable sample portions.
Particulate removal can occur in many SCSO control processes, including catchbasins,
swirl concentrators, screens, drainage systems, and  detention ponds. These control
processes allow removal of the accumulated polluted sediment for final disposal in  an
appropriate manner. Uncontrolled sedimentation will occur in receiving waters, such as
lakes, reservoirs, or large rivers. In these cases, the wide dispersal of the contaminated
sediment is difficult to remove and can cause significant detrimental effects. Biological
or chemical degradation of the toxicants in the sediments may occur, but is quite slow
for many of the pollutants in the expected anaerobic  environments. Degradation of the
soluble pollutants  in the water column may occur, especially when near the surface in
aerated waters. Volatilization (evaporation) is also a mechanism that may affect many of
the detected organic toxicants. Increased turbulence and oxygen supplies would
encourage these processes that may significantly reduce pollutant concentrations.

                                    281

-------
Sorption of pollutants onto solids and metal precipitation increases the sedimentation
potential of the pollutants and also encourages more efficient bonding of the pollutants
in soils, preventing their leaching to surrounding waters.


REFERENCES
Dept. of Commerce. Statistical Abstract of the United States, 1980. U.S. Bureau of the
Census. 101st. edition. Washington, D.C., 1980.

Dalrymple, R.J., S.L. Hodd, and D.C. Morin. Physical and Settling Characteristics of
Particulates in Storm and Sanitary Wastewaters. EPA-670/2-75-011. U.S. Environmental
Protection Agency, Cincinnati, Ohio, 1975.

EPA. Results of the Nationwide Urban Runoff Program. Water Planning Division, PB 84-
185552, Washington, D.C., December 1983.

Fam, S., M.K. Stenstrom, and G. Silverman. "Hydrocarbons in Urban Runoff." Journal of
Environmental Engineering, 113:1032-1046,1987.

Field, R. and R. Pitt. "Urban Storm-Induced Discharge Impacts: US Environmental
Protection Agency Research Program Review." Proceedings of the 2nd Wageningen
Conference on Urban Storm Water Quality and Ecological Effects Upon Receiving
Waters, Sept. 1989. Wageningen, The Netherlands. Journal of the IAWPRC: Water
Science and Technology. To be published in 1990.

Field, R., and Turkeltaub, R. "Urban Runoff Receiving Water Impacts: Program
Overview". Journal of Environmental Engineering, 107:83-100,1981.

Hoffman, E.J., G.L Mills, J.S. Latimer, and J.G. Quinn. Urban Runoff as a Source of
Polycyclic Aromatic Hydrocarbons to Coastal Waters. Environment Science and
Technology, 18:580-587,1984.

Mount, D.I., A.E. Steen and T.J. Norberg-King. Validity of Effluent and Ambient Toxicity
for Predicting Biological Impact on Five Mile Creek, Birmingham, Alabama. EPA/600/8-
85/015. U.S. Environmental Protection Agency. Duluth, Minn., December 1985.

Mount, D.I., T.J. Norberg-King and A.E.  Steen. Validity of Effluent and Ambient Toxicity
for Predicting Biological Impact, Naugatuck River, Waterbury, Connecticut. EPA/600/8-
86/005. U.S. Environmental Protection Agency. Duluth, Minn., May 1986.

Norberg-King, T.J., D.I. Mount, J.R. Amato and J.E. Taraldsen. Application of the Water
Quality Based Approach on a Regional Scale: Toxicity Testing and Identification of
Toxicants in Effluents from the San Francisco Bay Region.  National Effluent Toxicity
Assessment Center Technical Report 01 -88. U.S. Environmental Protection Agency.
Duluth, Minn., May 1988.

Pitt, R. Demonstration of Nonpoint Pollution Abatement Through Improved Street
Cleaning Practices. EPA-600/2-79-161, U.S. Environmental Protection Agency,
Cincinnati, Ohio, August 1979.

                                    282

-------
Pitt, R. and M. Bozeman. Sources of Urban Runoff Pollution and Its Effects on an Urban
Creek. EPA-600/2-82-090, U.S. Environmental Protection Agency, Cincinnati, December
1982.

Pitt, R. and P. Bissonnette. Bellevue Urban Runoff Program, Summary Report. U.S.
Environmental Protection Agency and the Storm and Surface Water Utility, Bellevue,
Washington, 1984.

Pitt, R. and J. McLean. Toronto Area Watershed Management Strategy Study; Number
River Pilot Watershed Project. Ontario Ministry of the Environment, Toronto, Ontario,
1986.

Pitt, R,, and P. Barron. Assessment of Urban and Industrial Stormwater Runoff Toxicity
and the Evaluation/Development of Treatment for Runoff Toxicity Abatement - Phase I.
U.S. Environmental Protection Agency, Office of Research and Development, Edison,
New Jersey, 1990.

Pitt, R., M. Lalor, M. Miller, and G. Driscoll. Assessment of Non-Stormwater Discharges
into Separate Storm Drainage Networks - Phase 1. Development of Methodology for a
Manual of Practice. U.S. Environmental Protection Agency, Office of Research and
Development, Edison, New Jersey, 1990.

Spiegel, S.J., E.G. Tifft, C.B. Murphy, and R.R. Ott. Evaluation of Urban Runoff and
Combined Sewer Overflow Mutagenicity. EPA-600/2-84-116. U.S. Environmental
Protection Agency, Cincinnati, June 1984.
Table  1.  Estimated Toxicant Discharges from Large U.S. Municipalities
(15,000 square miles total)


                      Median       Detection         Discharges
                   Cone.  (ug/L)    Frequency  (%)      (tons/year)
Arsenic                7               50                   80
Chromium               30               60                  350
Copper                 35               90                  700
Cyanide                40               25                  200
Lead                  150               95                 3000
Zinc                  150               95                 3000

Bis(2-ethylhexyl)
    phthalate          6               20                   30
Chlordane               i.s             20                    5
Chrysene                1.5             10                    3
Pluoranthene           3               15                   10
Pentachlorophenol      15               20                   7(3
Phenanthrene           1.5             10                    4
Pyrene                 2               15                    «

Source: from 1PA 1983


                                   283

-------
Table  2. General  Toxicity Groupings of Analyzed Samples
roofs
parking  areas
storage  areas
streets
loading  docks
vehicle  service  areas
landscaped areas

urban creeks
detention ponds

CSOs
most
toxic

   8%
  19
  25
   0
   0
   0
  17

   0
  10

  65
moderatly
toxic

    58%
    38
    50
    67
    67
    40
    33

    12
    10

    30
                                             least  number
                                             toxic  analyzed
33%
44
25
33
33
60
50

88
80
12
16
 8
 6
 3
 5
 6

19
12

20
Table 3. Detection Frequencies of the Most Frequently Occurring
Organic Compounds

                               Frequency of detection:
      1,3-Dichlorobenzene
      Fluoranthene
      Pyrene
      Benzo (b) fluoranthene
      Benzo (k) fluoranthene
      Benzo (a) pyrene
      Bis (2-chloroethyl) ether
      Bis (chloroisopropyl) ether
      Naphthalene
      Chlordane
      Benzo (a) anthracene
      Benzyl butyl phthalate
      Phenanthrene
                     23%
                     23
                     19
                     17
                     17
                     17
                     14
                     14
                     13
                     13
                     12
                     12
                     10
                                  284

-------
                                   Table 4.  OtoMrved Base Rautral Compounds
          nuober of
          nonfiltared
          analyoea:
Base Neutrals
roof runoff
     12
parking
   13
storage
   7
streets
   6
loading dock
     3
Bin (2-chloroethyl) ether
1, 3-Dichlorobenzene
Bin (chloroiaopropyl) ether
Hexachloroethana
Bia (2-chloroethoxyl) nathane
Naphthalene
Dl-n-butyl phthalate
Pbananthrene
Anthracene
Benzyl butyl phthalate
Fluoranthene
Bl* (2-ethyl haxyl) phthalate
Pyrene
Benzo (a) anthracvoe
Chryoene
Beozo (b) fluoranthene
Benso (k) fluoranthen*
Benzo (a) pyrene
Benzo (g,h,l) perylene
Di-n-octyl phtbalat*
nitrobenzene
Isophorone
Of ethyl  phthalate
20.3;20.5;86.6
13.6;54.7;88.4
67.8f81.5;147
5S.8

47.7|187
31.2
22.1
23.7
10S
7.«ilS.3|44.8

27.6
16.3
73.1
6.4;14.2»27.5;266
3.4ill.S|12.2|221
10.9;34f51.6;300
1S.3; 23.8
9.6;33.2;60.2
81.4;102;217
40.9|47.2

71.8
12.8;41

£3.3;21.4
1.0;1S.6;93.8

1.0;40.2;79.6
1.8;16.2|54.8
29.3
10;17.5;132
8;11.3;41.6
21.4;19.5;78.3
19.7
16.2
                  1S.1
                  S.4
4.5
31.2
8
0.6
30S
0.7
                  13.9
                  IS.4
                  18.8

-------
                       Table 4.   Observed Base Nautral Compound*
                                                                                                      (cont'd. )
          nonfiltered
          analyse*!
Base
vehicle service
    areas
      4
landscaped
  areas
    5
urban creeks
     4
dotentIon
  pond*
    4
New York CSOl
     19
    overall
detection froq.
                                                                                                                                      overall
                                                                                                                                      a&xlstum
Bis {2-ehloroethyl)  ether
1,3-Dichlorobenzene
Bis (chloroisopropyl)  ether
Hexachloroethano
Bie (2-chloroethoxyl)  methane
Naphthalene
 Acenaphthylene
Pluorena
Dl-n-butyl phthalate
Phenanthrone
Anthracene
Benzyl butyl phthalate
Fluoranthene
Bis (2-ethyl hasyl)  phthalate

Pyrene
Benzo (a) anthracene
Chrysene
Benzo 
-------
                                          Table 5,   Observed Pesticides (pg/L)
                  roof runoff
Nuober of analyaeai   12
Peaticides

alpha BHC
delta BHC
aldrin
DDT
endrin
chlordane
ODD
0.7
1.1
0,7
Q.3,-46.3

O.SiO.9,-2,2


parking storage
13 7



0.3
1.4
0.8.-1.2 1.1

vehicle
loading eerivce
streets dock areas
634





0.8 1.0 0.8


Hew York
CSOa
19
all BHCs:
0.3



0.5
1.2
overall
detection
frequency

all BHCs:
4*
1
4
1
13
1
                                                                                                               overall
                                                                                                               maximum
 0.7
 1.1
 0.7
46.3
 1.4
 2.2
 1.2

-------
T«bl« 6.  Observed Source ATM Hetal Concentration* (fig/14
Source Area Runoff
roof a
detection frequency
median
maximum
% filt. (range, median)
parking areas
detection frequency
median
% filterable
storage areas
detection frequency
median
% filterable
streets
detection frequency
median
maximum
% filterable
loading docks
detection frequency
median
maximum
% filterable
vehicle service areas
detection frequency
median
RHIX jLnmifl
% filterable
Aluminum
non
filfc. flit.
11/12
270
8370
2-»100
12/12
1550
22,500
1-»100
6/6
975
6990

4/4
4,000
10,040
l-»44
3/3
810
930
<<5«
4/4
920
1370

-------
Table 6.  Observed Source Area Metal  Concentrations
(cont'd.)
Aluminum

Source Area Runoff
landscaped areas
detection frequency
medium
maximum
% filterable
New York CSOs
detection frequency
„» median-
oo maximum
% filt. (range, median)
urban creeks
detection frequency
medi an
maximum
% filterable
detention ponda
detection frequency
mpfH an
iSixliuU
% filterable
non
filt.

4/4
2500
4610
34*93
19/19
720
23,030
1
4/4
1600
3250
13*42

4/4
550
1350
6*100

filt.

4/4
1600
1860
(50%)
_



4/4
240
500
(15%)

4/4
200
330
(60%)
Cadmium
non
filt.

3/5
0.04
1.0
(<2!
20/20
1.6
10
2*100
4/4
5
30
(<1!

4/4
0.24
1.0
(<5<

filt.

1/S
<0.1
1.0
5%)
20/20
0.25
5.1
(35%)
0/4
<0.1
<0.1
k)

1/4
<0.1
<0.1
k)
Chromium
non
filt.

5/1S
100
250
2*68
20/20
17.6
130
(
Lead
non
filt.

5/5
9.4
70
<»>
14/20
1.8
7.S
(15%)
0/4
<1
<1
k)

0/4
<1
<1
S%)
Nickel
non
filt.

3/5
30
130
(<5
20/20
11.3
48.2
12*100
3/4
20
70
(<5

3/4
20
70
(<5

filt.

0/5
<1
<1
»)
19/20
5.5
48.2
(50%)
0/4
<1
<1
*)

0/4
<1
<1
*)
Zinc
non
filt. filt.

5/5 5/5
32 32
1160 670
58*100 (100%)
20/20 20/20
96 34
390 80
3*100 (35%)
4/4 4/4
24 19
32 23
53*100 (80%)

4/4 4/4
22 22
25 25
(all 100%)

-------
 DEGRADATION OF CHLORIHATEP BIPHENYLS BY ESGHERICHIA COLI  COMTAIHING CLOMED
        GENESOF THE PSEUDOMOHAS HJTIDA Fl  TOLUEME  GATABOLIC PATHWAY
               bys  Gerben J, Zylstra, Sadhana Chauhan, and David T. Gibson
                    Department of Microbiology and
                    Biocatalysis Research Group
                    University of Iowa College of Medicine
                    Iowa City, Iowa 52242
                                  ABSTRACT

      Pseudomonas putida Fl initiates the degradation of toluene by a
multieotnponent enzyme system designated toluene dioxygenase. The enzyme
system consists of a flavoprotein (reductaseTOL), a ferredoxin
(ferredoxinj0f ), and an iron-sulfur protein (ISP.JQY).  Electrons are
accepted from HADH by reduetase.pQf and transferred through ferredoxin^Qr to
the terminal dioxygenase, ISP~QT.  Reduced ISPmQT catalyzes the addition of
molecular oxygen to the aromatic nucleus stereospecifically to form cis-
dihydrodiols from a variety of aromatic substrates.  The cis-dihydrodiols
are oxidized to catechols which are the substrates for meta ring cleavage
by catechol dioxygenase.  Certain bacteria that degrade polychlorinated
biphenyls (PCBs) utilize an analogous pathway.  Homology studies between
PCB degrading strains and P. putida Fl have indicated that the enzymes
involved are quite similar.  We have constructed strains of Escherichia
coli that express the cloned genes of the P. putida Fl toluene catabolic
pathway under the control of the tac promoter.  These strains were utilized
to show that the initial enzymes .in the toluene catabolic pathway can also
oxidize certain chlorinated biphenyls.

      This paper has been reviewed in accordance with the U.S.
Environmental Protection Agency's peer and administrative review policies
and approved for presentation and publication.
                                      290

-------
                               INTRODUCTION
      Polychlorlnated biphenyls (PCBs) are a class of aryl halides
synthesized by the catalytic chlorlnation of biphenyl.  Depending on the
number of chlorines and their position on the biphenyl ring structure 209
PCB congeners are possible.  They are extremely stable compounds, resistant
to acids, bases, and oxidation, possess high dielectric constants, and are
highly insoluble in water.  Due to these properties, they find extensive
use in industry in capacitors, transformers, dielectric fluids, fire
retardants, and plasticizers.  The very properties which make them
extremely useful in industry has also made them very persistent pollutants
in the environment.  Although PCBs are no longer synthesized, considerable
amounts have previously been introduced into the environment whisre they
tend to persist for long periods of time*  This is due to the resistance of
many individual PCBs to biodegradation.  This is particularly true for
highly chlorinated PCBs (containing more than 5 chlorines) (1,2).  The
lipophlllc nature of PCBs causes them to concentrate in the food chain.
This is of concern to human health particularly since their long term
effects are not fully understood.

      Several reports have now been published on the study of the
biodegradation of PCBs.  There is evidence that particular PCBs can undergo
microbial reductive dehalogenation in anoxic sediments (3,4).  Several
reports have also shown that particular PCBs can be degraded aeroblcally by
mixed or pure bacterial cultures.  Bedard and coworkers have developed an
assay for screening and characterizing bacteria that degrade PCBs ranging
from di- to hexachlorobiphenyls (5).  Two strains, Alcaligenes eutrophus
H850 and Pseudomonas sp. LB400, have shown unique abilities to degrade
highly chlorinated PCBs like penta- and hexachlorinated biphenyls In
addition to the lower chlorinated biphenyls (5).  A 2,3-dioxygenase attack
has been reported as the first step in the degradation of PCBs by several
strains in addition to A. eutrophus H850 and Pseudomonas sp. LB400 (6,7).
Pseudomonas sp. LB400 has been shown to oxidize a wide variety of PCBs and
specific intermediates in the degradative pathway have been Isolated and
identified (8,9).  A novel 3,4-dloxygenase attack on 2,5,2',5'-
tetrachlorobiphenyl has been reported by Nadlm et. al. (8) by both A.
eutrophus H850 and Pseudomonas sp. LB400.  This observation may be
explained by a biphenyl 2,3-dloxygenase with relaxed substrate specificity
or by the presence of a second biphenyl dioxygenase which attacks at the
3,4 position (9).  The Peeudomonas sp. LB400 genes responsible for the
degradation of biphenyl to 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid
have been cloned Into E. coll.  This recombinant strain degrades the same
spectrum of PCBs as Pseudomonas sp. LB400 (10).

      The initial reactions involved in the oxidation of biphenyl (the
aromatic nucleus of PCBs) by Pseudomonas sp. LB400 are shown in Figure 1A.
Analogous pathways are utilized in the degradation of a variety of aromatic
hydrocarbons (11).  Biphenyl is first oxidized to cls-biphenyl dihydrodiol
by the action of biphenyl dioxygenase.  cis-Biphenyl dihydrodiol is further
dehydrogenated to form 2,3-dihydroxybiphenyl.  The latter compound
                                      231

-------
    A.
                  bphA
                 Blphenyl
               Dloxygenase
                           bphB
                        c/s-BIphenyl
                        Dlhydrodlol
                       Dehydrogenase
    bphC
 2,3-DIhydroxy-
 Blphenyl 1,2-
 Dloxygenaae
           CHa
                  CH3
Figure 1,
                todC1C2BA
                 Toluene
                Dloxygenase
                            todD
                         c/s-Toluene
                         Dlhydrodlol
                       Dehydrogenase
    todE
3-Methylcatechol
2,3-Dloxygenase
A» Metabolism of biphenyl Tby Pseudomonas  sp.  400.
B. Metabolism of toluene by Pseudomonas putida Fl.
undergoes ring cleavage by  2,3-dihydroxybiphenyl dioxygenase to form 2-
hydroxy-6~oxo-6-phenylhexa-2,4-dienoie acid.   The genes for these three
enzymes have been designated bphA.  bphB.  and  bghC,  respectively.

      A similar pathway for the  degradation of mononuclear aromatic
hydrocarbons has been  shown in Pseudomonas  putida Fl  (11,  Figure IB).
Degradation of toluene (and other compounds)  by this  microorganism proceeds
through a cis-dihydrodiol.  a 2,3-catecholt  and a ring fission product, all
of which are structurally similar to  intermediates in the  pathway for the
aerobic degradation of biphenyls.   Such similarity also extends to the
structural gene sequences encoding  for particular proteins in the two
degradation pathways.  It has been  shown  previously that the P. putida Fl
genes for eis-toluene  dihydrodiol dehydrogenase (todD)  and 3-methylcatechol
dioxygenase (todE) share 61Z homology at  the  nucleotide level with the P.
pseudoalcaligenes KF707 genes for cis-biphenyl dihydrodiol dehydrogenase
(bphB) and 2,3-dihydroxybiphenyl dioxygenase  (bphC) (12).

      In the present work we investigated whether the P. putida Fl genes
for toluene dioxygenase (todClC2BA),  eis-toluene dihydrodiol dehydrogenase
(todD)» and 3-methylcatechol 2,3-dioxygenase  (todE) are homologous to the
                                      292

-------
analogous Pseudotnonas sp. LB400 genes for biphenyl dioxygenase (bphA). eis-
biphenyl dihydrodiol dehydrogenase (bphB). and 2,3-dlhydroxybiphenyl 1,2-
dioxygenase (bphC).  We also investigated whether such structural homology
extended to functional homology by investigating the ability of enzymes in
the P. putida Fl toluene degradative pathway to metabolize certain PCBs.

                           MATERIALS AND METHODS
BACTERIAL STRAINS, PLASMIDS, AND MEDIA

      P. putida Fl (13) and F39/D (14) have been described previously.  P.
putida F39/D is a todD mutant derived from P. putida Fl by N-methyl-N'-
nitro-N-nitrosoguanidine mutagenesis.  P. putida F39/D therefore
accumulates cis-dlols when incubated with aromatic compounds after
induction.  Pseudomonas sp. LB400, a PCB degrading strain, was previously
isolated from PCB-contaminated soil (15).  The recombinant plasmid pGEM456
(10) contains the Pseudomonas sp. LB400 bphABC genes (responsible for the
first three steps in PCB metabolism by this strain) cloned into pUC18.  The
recombinant plasmid pDTG351 (16,17) contains the P. putida Fl todClC2BADE
genes (responsible for the first three steps in toluene metabolism by this
strain) cloned into the broad host range vector pKT230.  Plasmid pDTG601
and pDTG602 (12) contain the todC!C2BA genes and the todClC2BAD genes,
respectively, cloned into the expression vector pKK223-3 so as to place
gene expression under control of the tac promoter.  Thus, E. coli strains
containing either pDTG601 or pDTG602 will accumulate cis-toluene
dihydrodiol or 3-methylcatechol, respectively, from toluene after induction
with isopropyl-/3-D-thiogalactopyranoside.  E. coli strain JM109 (recAl
endAl g?rA96 thi hsdRl? supE44 relAl A(lae-proAB) [F* traD36 proAB
lacIqZAM151) was used as host for plasmids pDTGSOl and pDTC602.

      Mineral salts base medium (MSB) was prepared as described previously
(18).  Arginine (0.2Z), toluene (vapor phase), or glucose (20 mM) served as
carbon sources.  Thiamine (1 mM) was added to E. eoli JM109 cultures.
Ampicillin (100 jug/ml) was added to plasmid-containing cultures to prevent
plasmid loss.

DNA TECHNIQUES

      Plasmid DNA was prepared by the alkaline-sodium dodecyl sulfate
procedure of Birnboim and Doly (19) as described by Ish-Horowitz and Burke
(20) and purified by centrifugation in a cesium chloride-ethidium bromide
density gradient.  DNA was stored in 10 mM Tris 1 mM EDTA (pH 8.0) at
-20°C.  Plasmid DNA was cleaved with restriction enzymes as recommended by
the supplier (Bethesda Research Laboratories, Inc., Gaithersburg, Md.).
Agarose gel electrophoresis (l.OZ) was conducted in 40 mM Tris, 20 mM
acetate, 2 mM EDTA buffer.  Transfer of DNA from agarose gels to Zeta-probe
nylon membranes (Bio-Rad Laboratories, Rockville Centre, NY) was performed
using an LKB vacugene apparatus as recommended by the supplier (LKB
Instruments, Gaithersburg, MD).  DNA restriction fragments to be used as
probes in Southern blotting experiments were separated by gel
                                     293

-------
electrophoresls and eluted from gel fragments by the procedure of
Vogelattein and Gillespie  (21).  DNA fragments were labeled by the random
priming method of Feinberg and Vogelstein (22).  Southern hybridizations
were performed ae recommended by the nylon membrane supplier (Bio-Rad
Laboratories, Rockvllle Centre, NY).

PCB TRANSFORMATIONS

      £• putlda F39/D was grown on MSB arglnine in the presence of toluene
with shaking at 30°G until the cells reached a turbidity of 1.0 at 600 nm.
The cells were harvested by centrifugation (8,000 X g for 10 min at 4°G),
•washed in 50 mM Na/KPO^ buffer (pH 7.25), and resuspended at a turbidity
(600 nm) of 1.0 in 50 mM Na/KPO^ buffer (pH 7.25) supplemented with
arginine.  E. coll JM109 containing cloned genes of the P. putida Fl
toluene degradative pathway were grown on MSB glucose with thiarnine and
ampicillin with shaking at 37°C.  When the cells reached a turbidity of 0.5
(at 600 nm) IPTG was added to a final concentration of 1.0 mM and growth
continued for 1 hr (one doubling time).  Cells were harvested as above
except that they were resuspended in 50 mM Na/KPO, buffer (pH 7.25)
supplemented with glucose at a turbidity of 2.0 (600 nm).  Cells (50 ml
aliquots) were shaken at 30°C overnight with the appropriate chlorinated
biphenyl (1 mg).  Cell cultures were then extracted three times with an
oqual volume of ethyl acetate (150 ml total volume).  Each culture was
acidified and extracted three more times with equal volumes of ethyl
acetate.  Extracts were dried over anhydrous sodium sulfate and evaporated
to approximately 2.0 ml under vacuum at 30°C.  The extracts were then
evaporated to dryness tinder a stream of nitrogen at room temperature in a
small glass vial.  Each residue was resuspended in 100 n"L methanol.

CHEMICAL ANALYSES

      Thin layer chromatography was performed with a solvent system of
chloroform and acetone (80:20 v/v) using silica gel plastic plates (Eastman
Kodak, Rochester, NY).  Products were localized under ultraviolet light
(254 nm) followed by spraying with Gibb's reagent (2,6-dichloroquinone-4-
chloroifflide, 3.0% in methanol).  Catechols were identified by a strong
purple color and phenols by a bright blue color after heating at 80°C for
10 min.

      Acetonide derivatives were synthesized to confirm the cis-relative
stereochemistry of dihydrodiol metabolites (23).  Dihydrodiols were
dissolved in 2.0 ml dimethoxypropane at 0°C.  A crystal of £-toluene
sulfonic acid was added to initiate the reaction.  Anhydrous sodium
carbonate (100 mg) was added after 1 hr and stirred for 10 min.  Benzene (2
ml) was added and the reaction mixture filtered through a sintered glass
funnel to remove solid material.  The filtrate was evaporated under a
gentle stream of nitrogen and the residue was redissolved in a small
quantity of chloroform.

      Compounds were analyzed by gas chromatography followed by mass
apectroscopy (Hewlett Packard 5890 with MS 5970) equipped with a 25 m Ultra
                                     294

-------
HP-1 column (0.2 mm x 0.33 j/m film thickness).  The injector, oven, and
detector temperatures on the gas chromatograph were 175°C, 70°C, and 280°C,
respectively.  Samples were run with a gradient program starting at 70°C
and ending at 280°C at a rate of 10°C per min.  The carrier ga« was Helium
(1 ml/min).

CHEMICAL SOURCES

      Biphenyl was obtained from Aldrich Scientific (Milwaukee,, WI).  2,3-
Dihydroxybiphenyl was a gift from General Electric Co. (Schenectady, NY).
2-Hydroxybiphenyl, 3-hydroxybiphenyl, 2,3-dichlorobiphenyl, 2,4-
dichlorobiphenyl, 2,5-dichlorobiphenyl, and 3,4-dichlorobiphenyl were
obtained from Ultra Scientific (North Kingstown, RI).  2-Chlorobiphenyl and
3-chlorobiphenyl were obtained from Chemicals Procurement Laboratories,
Inc. (College Point, NY) while 4-chlorobiphenyl was obtained from Aldrich
Scientific (Milwaukee, WI).  cis-2,3-Dihydroxy-l-phenyl-cycloh«xa-4,6-diene
was prepared using a cis-dihydrodiol-accumulating mutant strain of
Bei-jerinckia (B8/36) according to published procedures (24).

                                  RESULTS
SOUTHERN BLOT ANALYSES

      It was previously shown (12) based on nucleotide sequence analysis
that the genes for P. putida Fl cis-toluene dihydrodiol dehydrogenase
(todD) and 3-methylcatechol 2,3-dioxygenase (todE) were 61Z homologous with
the genes for P. pseudoalcaligenes KF707 cis-biphenvl dihydrodiol
dehydrogenase (bphB) and 2,3-dihydroxybiphenyl 1,2-dioxygenase (bphC).  No
direct analysis could be made of the homology between the geneo for P.
putida Fl toluene dioxygenase (todC_lC2BA) and the genes for P.
pseudoalcaligenes KF707 biphenyl dioxygenase (bphA) since no nucleotide
sequence has been published for any biphenyl dioxygenase.  We decided
therefore to analyze whether the biphenyl degradation genes from another
PCB degrading strain, Pseudomonas sp. LB400, were homologous to the toluene
degradation genes of P. putida Fl by Southern blot analysis.  Plasmid
pDTG351 (16,17; Figure 2) contains the todClC2BADE genes.  Based on the
nucleotide sequence (12) the region from the EcoRI site to the first Belli
site contains the genes for the three components of toluene dioxygenase
(todClC2BA) and the region from the first Belli site to the first Xhol site
contains the genes for cis-toluene dihydrodiol dehydrogenase and 3-
methylcatechol dioxygenase (todDE).  Two radioactive probes were prepared
using DNA located between the restriction sites mentioned above.  One probe
contained the todC!C2BA genes and the other contained the todDH genes.
Plasmid pGEM456 (11) cleaved with PstI was the target DNA.  This plasmid
contains the bphABC genes cloned from Pseudomonas sp. LB400 (!]-)•  As can
be seen in Figure 3, the probe prepared from the todC!C2BA genes hybridized
to three PstI fragments (1.9, 1.6, 0.3 kilobase pairs) from pGKM456 while
the probe prepared from the todDE genes hybridized to a single PstI
fragment (2.1 kilobase pairs).  Thus, the P. putida Fl genes for the
conversion of toluene to 3-methylcatechol (todClC2BADE) are structurally
                                     295

-------
«   §1
If   ff
  MM      MM
HM-OiE  M  M«MM
o   u Q-u
y-im  w  now^
   :T  7  IfTfr
                    Vector
              Cl C2  B   A   D  E
              	Cloned Chromosomal Region
                                           L
                                            todC!C2BA Probe
                                                                J
                                                         todDE Probe
 Figure 2.  Restriction map of pDTGSSl.
       genes are indicated.
          The locations of the todClC2BADE
            1  2
                                     Lane 1: 1 kilobase pair DMA standard
                                             (Bethesda Research
                                             Laboratories Inc.,
                                             Gaithersburg, MD).
                                     Lane 2: pGEM456 cut with Pstl.
                                     Lane 3: pGEM456 cut with Pstl and
                                             blotted with the todC!C2BA
                                             genes.
                                     Lane 4: pGEM456 cut with Pstl and
                                             blotted with the todDE genes.
 Figure 3.  Southern blot showing homology between bphABC genes and
       todClC2BADE genes.

 oimilar to the Pseudomonas sp. LB400 genes for the conversion of biphenyl
 to 2,3-dihydroxybiphenyl.

 D1HYDRODIOL FORMATION FROM CHLORINATED BIPHENYLS

       The Southern blot experiments outlined above demonstrated homology
 between the aromatic hydrocarbon degradation genes from P. putida Fl and
 Pseudomonas sp. LB400.  However, structural homology does not imply
 functional homology.  Enzymes may show homology with one another but still
 have different substrate specificities (12).  Therefore, the cis-toluene
 dihydrodiol-accumulating strain P. putida F39/D (derived from strain Fl)
 and E. coli JM109(pDTG601) (containing the todC!C2BA genes) were utilized
 to determine if the substrate range of toluene dioxygenase extended to
                                      296

-------
chlorinated biphenyls.

      The above strains were first tested for their ability to oxidize
biphenyl to cis-biphenyl dihydrodiol.  The two strains were grown, induced,
and incubated with biphenyl as described in the Materials and Methods
section.  Product accumulation was monitored by taking an absorption
spectrum of culture supernatants at appropriate time intervals.  After the
transformation, reaction products were extracted and analyzed.  No residual
biphenyl was detected in the extracts, indicating complete conversion to
product.  The oxidation product cis-2.3-dihvdroxy-l-phenyl-cyclohexa-4.6-
diene was identified using thin layer chromatography and gas chromatography
followed by mass spectroscopy by comparison to the authentic compound.  In
addition, 2-hydroxybiphenyl and 3-hydroxybiphenyl were identified by the
same techniques as the products formed by acid-catalyzed dehydration of the
cis-biphenyl dihydrodiol.  The cis-relative stereochemistry was confirmed
following the preparation and identification of an isopropylidine
derivative (23,24).

      Several simple PCBs were chosen for initial analyses of the ability
of toluene dioxygenase to oxidize chlorinated biphenyls.  The single
chlorinated biphenyls: 2-chlorobiphenyl, 3-chlorobiphenyl, and. 4-
chlorobiphenyl were all readily oxidized to products by toluer.e dioxygenase
utilizing either P. putida F39/D or E. coli JM109(pDTG601) (Te.ble 1).  In
addition, the dichlorinated biphenyls: 2,3-dichlorobiphenyl, 2,4-
dichlorobiphenyl, 2,5-dichlorobiphenyl, and 3,4-dichlorobiphenyl were also
readily oxidized by both organisms.  Gas chromatography followed by mass
spectroscopy of the products formed showed molecular ions characteristic of
dihydrodiols formed from each chlorinated substrate (Table 1).  In
addition, each dihydrodiol formed an isopropylidine derivative indicating a
cis-relative stereochemistry of hydroxyl groups.  Analysis of the acid-
catalyzed dehydration products formed from the cis-dihydrodiols was
conducted by gas chromatography followed by mass spectroscopy.  The results
showed that two phenols were formed from each dihydrodiol (Table 2).  These
products are analogous to the 2- and 3-hydroxybiphenyls formed from cis-
biphenyl dihydrodiol and confirm the dihydrodiol structures of the initial

         TABLE 1.  GAS CHROMATOGRAPHY / MASS SPECTROSCOPY ANALYSIS OF
              CIS-DIHYDRODIOLS FORMED FROM CHLORINATED BIPHENYLS*
Cl-
2-
3-
4-
2,
2,
2,
3,



3-
4-
5-
4-
Rt
17
18
18
19
18
18
20
.1
.4
.2
.2
.9
.9
.1
222(15)
222(11)
222(11)
256(32)
256(39)
256(34)
256(65)
m/e (% of base peak)
204(100)
204(100)
204(100)
227(69)
227(67)
227(73)
238(100)
169(76)
188(57)
188(30)
172(34)
172(68)
172(36)
228(51)
141(43)
169(50)
169(41)
165(100)
165(100)
165(100)
184(51)

141(35)
141(35)
128(58)
128(74)
128(73)
128(69)
Isotope Intensity
P P+2 P+4
100%
100%
100%
100%
100%
100%
100%
28%
23%
35%
76%
73%
74%
70%



18%
10%
14%
18%
*Abbreviations: Cl, chlorine position on biphenyl; Rt, retention time; m/e,
molecular ion; P, parent ion.
                                     297

-------
    TABLE 2.  GAS CHROMATOGRAPHY / MASS SPECTROSCOPY ANALYSIS OF
           PHENOLS FORMED BY ACID-CATALYZED DEHYDRATION OF
         Clg-DIHYDRODIOLS FORMED FROM CHLORIHATED BIPHENYLS*
Isotope Intensity
Cl-
2-

3-

4-

2,3-

2,4-

2,5-

3,4-

_gt
15
16
16
17
16
17
17
18
17
18
17
18
18
19
.1
.9
.0
.8
.0
.9
.3
.9
.0
.6
.0
.5
.3
.8
m/e (Z of base peak)
204(62)
204(100)
204(94)
204(100)
204(100)
204(100)
238(99)
238(100)
238(74)
238(100)
238(85)
238(100)
238(99)
238(100)
169(100)
169(7)
169(100)
169(5)
169(75)
169(4)
203(83)
203(3)
203(70)
202(2)
202(55)
202(3)
203(45)
202(3)
141(28)
141(15)
141(35)
141(17)
141(30)
141(17)
168(100)
168(24)
168(100)
168(24)
168(100)
168(23)
168(100)
168(17)
115(19)
115(10)
115(24)
115(8)
115(20)
115(8)
139(51)
139(22)
139(43)
139(22)
139(51)
139(20)
139(37)
139(12)
P
100Z
100Z
100%
100Z
100Z
100Z
100Z
100Z
100Z
100Z
100Z
100Z
100Z
100Z
P+2
32Z
32Z
28Z
32%
32Z
33Z
64Z
79Z
56Z
70Z
7SZ
65Z
69Z
76Z
I±4 _






11Z
12Z
10Z
12Z
12Z
11Z
13Z
12Z
*Abbreviations: Cl, chlorine position on biphenyl; Rt, retention
time; m/e, molecular ion; P, parent ion.

products.  Ho metabolites were formed when P. putIda F39/D and E. coli
JM109(pDTG601) were incubated with chlorinated blphenyls containing more
than two chlorine substituents.

CATECHOL FORMATION FROM CHLORINATED BIPHENYLS

      In order to further study the substrate range of the enzymes of the
toluene dioxygenase pathway we analyzed the ability of cis-toluene
dihydrodiol dehydrogenase to dehydrogenate the cis-dihydrodiols (Table 3).
In this case the recombinant E. coli strain JM109(pDTG602) was utilized.
This strain contains the cloned genes for both toluene dioxygenase
(todClG2BA) and els-toluene dihydrodiol dehydrogenase (todD).  Therefore,
any product formed from toluene dioxygenase in vivo will be acted upon by
the dehydrogenase to form the corresponding catechol.  A feasibility study
was first conducted with biphenyl,  E. coli JM109(pDTG602) was grown,
Induced, and incubated with biphenyl as described above.  Transformation
products were extracted and analyzed as described in the Materials and
Methods.  The only product detected by either thin layer chromatography or
gas chromatography followed by mass spectroscopy was 2,3-dlhydroxybiphenyl.
Authentic 2,3-dihydroxyblphenyl had properties identical to those of the
bacterial metabolite.

      The monochlorinated and dichlorinated biphenyls which were shown to
be oxidized by toluene dioxygenase to cis-dihydrodiols were also tested
with IS. coli JM109(pDTG602) to determine if catechol compounds could be
formed.  Each of the chlorinated compounds tested showed a single product
with a molecular Ion corresponding to a dihydroxylated product of the
                                     298

-------
    TABLE 3.  GAS CHROMATOGRAPHY / MASS SPECTROSCOPY ANALYSIS OF
            CATECHOLS FORMED FROM CHLORINATED BIPHENYLS*
Isotope Intensity
Gl-
2-
3-
4-
2,
2,
2*
3,



3—
4—
5-
4-
Rt
17
18
18
19
19
19
20
.3
.4
.5
.5
.2
.3
.7
m,
220(91)
220(100)
220(100)
254(95)
254(76)
254(65)
254(88)
le (% of base t>eak)
185(100)
185(86)
185(72)
219(54)
219(61)
219(39)
219(49)
157(17)
157(17)
157(15)
184(100)
184(100)
184(100)
184(100)
128(25)
128(25)
128(21)




P
100Z
100Z
1002
1002
1001
1001
1002
P+2
352
332
372
722
682
692
752
P+4



142
122
112
132
*Abbreviations: Cl, chlorine position on blphenylj Rt» retention
time; m/e, molecular ion; P, parent ion.

starting chlorinated biphenyl (Table 3).  Thus, P. putida Fl is capable of
metabolizing certain chlorinated blphenyls to the corresponding catechols
by the enzymes of the toluene dioxygenase pathway.

                                 DISCUSSION
      £• putida Fl is capable of growing on a variety of aromatic
hydrocarbons as the sole source of carbon and energy for growth.  The
pathway by which this strain degrades such compounds is shown in Figure IB.
Toluene is first oxidized to cis_-l(S),2(R)-dihydroxy-3-methyleyclohexa-3,5-
diene by the action of toluene dioxygenase.  Subsequently this product is
dehydrogenated by cis-toluene dihydrodiol dehydrogenase to form 3-
methylcatechol, the substrate for meta ring cleavage by 3-methyleatecho1
2,3-dioxygenase.  This pathway is chemically analogous to that utilized by
Pseudomonas sp. LB400 in the metabolism of biphenyl and various PCBs
(Figure 1A).  In the present work we have shown that the genes for toluene
dioxygenase (todClC2BA), cis-toluene dihydrodiol dehydrogenase (todD), and
3-methylcatechol dioxygenase (todE) show homology to the analogous genes
from the Pseudomonas sp. LB400 PCS degradation pathway (bphABC).  In
addition, toluene dioxygenase was shown to be able to oxidize several
chlorinated biphenyls to cis-dihydrodiols.  Such products could be acted
upon by cis-toluene dihydrodiol dehydrogenase to form dihydroxy chlorinated
biphenyls (catechols).  The positions of the hydroxyl groups on cis-
biphenyl dihydrodiol and dihydroxybiphenyl formed by toluene dioxygenase
were determined to be on the second and third carbon of one of the aromatic
rings by comparison with known standards.  By analogy, oxidation of the
chlorinated biphenyls tested with toluene dioxygenase in this study
probably occurred at the same positions.  We suspect also that such
oxidation occurred on the aromatic ring that did not contain any chlorine
atoms.  We are currently investigating whether this is actually the case.
The chlorinated biphenyls oxidized by toluene dioxygenase were limited to
those containing one or two chlorine substituents.  Toluene dioxygenase
does not show the broad PCB substrate specificity of the Pseudomonas sp.
LB400 enzyme or the analogous enzyme from Alcaligenes eutrophus H850 (8,9).
                                     299

-------
Consequently, Pseudomonas ap. LB400 and A. eutrophue H850 remain the
organisms of choice for future studies on the degradation of chlorinated
biphenyls that contain several chlorine substituents.

                             ACKNOWLEDGEMENTS
       This work was supported by Cooperative Agreement #CR-816352-01-0 from
the U.S. Environmental Protection Agency, Risk Reduction Engineering
Laboratory, Office of Research and Development,  We thank Dr. P. R. Sferra,
EPA Project Officer, for his interest, support, and suggestions.  6.J.Z. is
the recipient of NRSA award 1 132 (3113091 from the National Institute of
General Medical Sciences.

                                 REFERENCES
1.   Furukawa, K.  Microbial  degradation of polychlorinated biphenyls
     (PCBs).  In: A. M.  Chakrabarty  (ed.), Biodegradation and
     Detoxification of Environmental Pollutants.  CRC Press, Inc., Boca
     Raton, Florida, 1982.  p.  33.

2.   Safe, S. H.  Microbial degradation of polychlorinated biphenyls.  In;
     D,  T. Gibson  (ed.), Microbial Degradation of Organic Compounds.
     Marcel Dekker, Inc., New York,  1984.  p. 361.

3.   Brown, J. F., Jr.,  Bedard, D. L., Brennan, M. J., Carnahan, J. C.,
     Feng, H., and Wagner, R. E.  Polychlorinated biphenyl degradation in
     aquatic sediments.  Science  236: 709, 1987.

4.   Brown, J. F., Wagner, R. E., Feng, H., Bedard, D. L., Brennan, M. J.,
     Carnahan, J. C., May, R. J.  Environmental dechlorination of PCBs.
     Environ. Toxicol. Chen.  6: 579, 1987.

5.   Bedard, D. L., Unterman, R., Bopp, L. H., Brennan, M. J., Haberl, M.
     L., and Johnson, C.  Rapid assay for screening and characterizing
     microorganisms for  the ability  to degrade polychlorinated biphenyls.
     Appl. Environ. Microbiol.  51: 761, 1986.

6.   Ahmed, M. and Focht, D.  D.   Degradation of polychlorinated biphenyls
     by  two species of Achromobacter.  Can. J. Microbiol. 19: 47, 1973.

7.   Furukawa, K., Tomizuka,  N.,  and Kamibayashi, A.  Effect of chlorine
     substitution on the bacterial metabolism of various polychlorinated
     biphenyls.  Appl. Environ. Microbiol. 38s 301, 1979.

8.   Nadim, L. M., Schocken,  M. J.,  Higson, F. K., Gibson, D. T., Bedard,
     D.  L., Bopp, L. H., and  Mondello, F. J.  Bacterial oxidation of
     polychlorinated biphenyls. In;  Proceedings of the 13th Annual Research
     Symposium on Land Disposal,  Remedial Action, Incineration, and
     Treatment of Hazardous Waste.   U.S. Environmental Protection
     Agency, Cincinnati, 1987.  p. 395.


                                     300

-------
9.  Singh, S., Higson, F. K., Nadim, L. M., and Gibson, D. T.  Oxidation
    of polychlorinated biphenyls by Pseudomonas putida LB400.  In:
    Proceedings of the 14th Annual Research Symposium on Land Disposal,
    Remedial Action, Incineration, and Treatment of Hazardous Waste.  U.S.
    Environmental Protection Agency, Cincinnati, 1988.  p. 346.

10. Mondello, F. J.  Cloning and expression in Escherichia coli of
    Pseudomonas strain LB400 genes encoding polychlorinated biphenyl
    degradation.  J. Bacteriol. 171: 1725, 1989.

11. Gibson, D. T. and Subramanian, V.  Microbial degradation of aromatic
    hydrocarbons.  In; D. T. Gibson (ed.)» Microbial Degradation of
    Organic Compounds.  Marcel Dekker, Inc., New York, 1984.  p. 181.

12. Zylstra, G. J. and Gibson, D. T.  Toluene degradation by Pseudomonas
    putida Fit nucleotide sequence of the  todClC2BADE genes and their
    expression in Escherichia coli.  J. Biol. Chem. 264: 14940, 1989.

13. Gibson, D. T., Koch, J. R., and Kallio, R. E.  Oxidative degradation
    of aromatic hydrocarbons by microorganisms. I. Enzymatic formation of
    catechol from benzene.  Biochemistry 7: 2653, 1968.

14. Gibson, D. T., Hensley, M., Yoshioka,  H., and Mabry, T. J.  Formation
    of (+)-cis-2,3-dihydroxy-l-methylcyclohexa-4,6-diene from toluene by
    Pseudomonas putida.  Biochemistry 9: 1626, 1970.

15. Bopp, L. H.  Degradation of highly chlorinated PCBs by Pseudomonas
    strain LB400.  J. Ind. Microbiol. 1: 23, 1986.

16. McCombie, W. R.  Cloning and analysis  of genes involved in toluene
    utilization by Pseudomonas putida PpFl.  In;  Abstracts of the Annual
    Meeting of the American Society for Microbiology.  American Society
    for Microbiology, Washington, D.C., 1984.  p. 155.

17. Zylstra, G. J., McCombie, W. R., Gibson, D. T., and Fineti:e, B. A.
    Toluene degradation by Pseudomonas putida Fl: genetic organization of
    the tod operon.  Appl. Environ. Microbiol. 54: 1498, 1988.

18. Stanier, R. Y., Palleroni, N. J., and  Doudoroff, M.  The aerobic
    pseudomonads: a taxonomic study.  J. Gen. Microbiol. 43: 159, 1966.

19. Birnboim, H. C. and Doly, J.  A rapid  alkaline extraction procedure
    for screening recombinant plasmid DNA.  Nuc. Acids Res. 7: 1513, 1979.

20. Ish-Horowitz, D. and Burke, J. F.  Rapid and efficient coumid cloning.
    Nuc. Acids Res. 9: 2989, 1989.

21. Vogelstein, B. and Gillespie, D.  Preparative and analytical
    purification of DNA from agarose.  Proc. Natl. Acad. Sci.  76: 615,
    1979.
                                     301

-------
22. Feinberg, A. P.  and Vogelstein, B.  A technique for radiolabelling DNA
    restriction cndonuclease fragments to high specific activity.  Anal.
    Bioehem.  132:  6,  1983.

23. Brown, B. R. and MacBride, J. A. H.  Mew methods for assignment of
    relative  configuration to 2.3-trans-flavan-3.4-diols.  J. Chem. Soc.
    p. 3822,  1964.

24. Gibson, D. T., Roberts, R. L., Wells, M. C., and Kobal, V. M.
    Oxidation of biphenyl by a Beilerinckia species.  Bioehem. Biophvs.
    Res. Commun. 50:  211, 1973.
                                     302

-------
USE OF DNA HYBRIDIZATION PROBES IN THE MONITORING OF BIOTREATMENT SYSTEMS

          by:   Richard A. Haugland1, Alan F. Rope1, Candace Strange1,
              John C. toper1,  Henry H.  Tabak2  and  P.R.  Sferra2

             Department of Molecular Genetics,  Biochemistry and
              Microbiology
              University of Cincinnati
              Cincinnati, Ohio  45267

             *Risk Reduction Engineering Laboratory
              U.S. Environmental Protection Agency
              Cincinnati, Ohio  45268
                               ABSTRACT

      As the result of recent technological  advances,  significant potential now
exists for the application of DNA hybridization probes in characterizing complex
microbial  communities.   A  review  is presented of  relevant  methodologies now
available  for  such  analyses as well  as  examples of their use.  In conjunction
with  other currently  ongoing studies  at  the  U.S.  Environmental  Protection
Agency's  Risk Reduction Engineering  Laboratory on the treatability  of toxic
organic compounds, we are initiating studies to  apply DNA probe methods towards
the characterization of resident microbial communities in wastewater treatment
systems.   The objectives  of this work will  include  1)  the development of DNA
hybridization  probes  that  specifically  detect  organisms   with  degradative
capabilities  toward different toxic chemicals in  these communities,  2)  the
establishment  of   hybridization  probe-based  methodologies  for  determining
abundances  of  these  organisms  in  the  communities,  3)  the  assessment  of
correlative relationships between the occurrence of these organisms and actual
degradative  rates  by  the  communities and  4)  the  identification  of adaptive
response mechanisms by  the communities to high concentrations of toxic chemicals.

      This  paper  has been reviewed  in  accordance with  the U.S. Environmental
Protection  Agency's peer  and administrative review  policies  and  approved for
presentation  and publication.
                                     303

-------
                             INTRODUCTION

      Increasingly stringent federal and state regulations concerning the
levels of hazardous organic chemicals which may be discharged from municipal
and industrial wastewater treatment facilities as well as from a vast number
of waste sites are creating new challenges for the design and operation of
effective yet economical treatment systems.  Advances in bioreactor designs
as well as in the identification and development of microorganisms with
priority pollutant degradative capabilities have provided important tools
that can be used to meet these challenges.  Up to now, however, relatively
little effort has been directed towards the characterization of microbial
communities within most waste treatment systems without which optimization
and the potential of bioremediation can not be realized.  This paucity of
characterization may be ascribed to both the high complexity of such
communities in most instances as well as the Inadequacy of traditional
methods for performing such determinations.  The development of new methods
for characterizing microbial communities and the application of these
methods in predicting and describing responses to potentially toxic
compounds would provide a means by which to improve upon the design,
operation and development of waste treatment systems.

      In recent years, considerable research effort has been directed toward
the application of DNA hybridization probes for both the identification and
enumeration of specific microorganisms and genotypes in environmental
settings.  Hybridization probes are functionally single-stranded nucleic
acid molecules (usually DNA) that have been labeled in a manner that allows
their detection. They are used to identify homologous single-stranded DNA or
RNA molecules in heterogeneous mixtures of nucleic acids from various target
sources.  Much of the impetus for these studies has been the recognized need
for monitoring genetically engineered organisms following their release into
the environment (1).  With an ever-increasing number of Isolated genes
coding for characteristic properties of different microorganisms becoming
available as potential probes, the utilization of this technology for
monitoring natural organisms in different environments is also likely to
increase.  Comparative nucleotide sequencing analysis of ribosomal RNAs or
of ribosomal RNA genes is another technology which is rapidly gaining
acceptance for use 1n determinative microbiology (2,3).  While to date,
phylogenetlcally-based hybridization probes generated from this growing data
base have been used only sporadically 1n characterizing microbial
communities 1n environmental samples, the potential applications for such
probes appear  great.

      This report will review various methodologies that are now available
for the use of DNA hybridization probes 1n detecting and enumerating
specific microorganisms or genes 1n environmental samples.  Examples, where
available, will also be presented of applications of these methods 1n the
prediction and analysis of microbial community responses to hazardous
chemicals.  These studies have thus far shared the feature of employing
probes for functional degradative genes.  Current limitations of this
approach, which will be discussed, have led our laboratory to Initiate
feasibility studies on the use of phylogenetlcally-based probes derived from
small subunit ribosomal RNA (ssrRNA) in addition to functional gene probes


                                     304

-------
for performing such evaluations.  Potential advantages and disadvantages of
this new approach will also be discussed.

                             METHODOLOGIES

      Hybridization probes are usually labeled by the incorporation of
radioactive isotopes into their structures, however, a number of nonisotopic
labeling procedures have also been developed.  Hybridization results from
the formation of hydrogen bonds between sequences of homologous base
residues present in the probe and corresponding molecules from the target
source, i.e., guanine bonds with cytosine and adenine bonds with thymine in
DNA, or with uracil in RNA.  Practical demonstration of hybridization
between probe and target source molecules involves fixing the target nucleic
acids to supports, such as nylon membranes prior to incubation with the
probe, in which case probe molecules also become fixed to the support upon
hybridization.  Hybridization is demonstrated and under appropriate
conditions quantified by label retention on the support.

      The majority of DNA probe applications related to environmental
studies have utilized the colony hybridization method.  While several
modifications of the original procedure of Grunstein and Hogness (4) have
been reported, its underlying principles have remained constant.  In this
procedure, portions of microbial colonies which have been cultivated on agar
medium are transferred to the support membranes by a process analogous to
replica plating.  The cells on the membrane are then lysed in situ to
release their DNA contents which are concommitantly denatured to a single-
stranded form and then bound to the membrane.  The membrane is then
incubated with the probe DNA and hybridization of probe molecules to
homologous target sequences from the individual colonies is detected as
described above.  Various applications of this procedure have demonstrated
both its versatility and sensitivity which, using appropriate conditions,
can approach one in 108 colonies (5).   The primary limitation of this
protocol is its requirement for prior cultivation of the microorganisms on a
given artificial medium.  It is widely recognized that only a small portion
of the microorganisms occurring in many natural habitats are amenable to
such cultivation (6,7).  Additional problems associated with this procedure
include variability in the conditions required to obtain satisfactory lysis
of different microorganisms and occasionally high noise to signal ratios
associated with non-specific binding of the probe to cellular debris from
the colonies.

      The limitations associated with detection and enumeration procedures
requiring prior cultivation of microorganisms from environmental samples
have led to the development of methods that allow hybridization of DNA
probes with nucleic acids that have been directly isolated from such
samples.  Methods for the isolation of DNA from soil and sediment samples
which have proven to be particularly refractory to such analyses have
recently been developed.  These methods alternatively employ a preliminary
step in which microbial cells are separated from the samples prior to
isolation of their DNA (8,9) or direct lysis of the cells and isolation of
their DNA from these samples (8,10).  Efficiencies of DNA recovery have been
shown to be higher using the latter procedure (8), however, DNA isolated by


                                     305

-------
both methods has been shown to be amenable to qualitative and quantitative
analysis for the presence of specific organisms or genotypes (9,10).  The
hybridization analyses are characteristically performed in a similar fashion
to colony hybridization analyses wherein the DNA isolated from the soil or
sediment samples is denatured and bound to a membrane support prior to
incubation with the probe.  Quantitation is achieved by comparing the
amounts of label retained by known amounts of total sample DNA to a standard
curve.  The curve is generated from different amounts of the specific target
DNA incubated simultaneously with the probe.  Detection limits for target
sequences using this approach have ranged in different studies from 0.2 pg
to 2 ng/ug total DNA (9,11).  These variations are likely to be in large
part attributable to differences in probe labeling efficiencies resulting
from the use of different labeling methods (9).  Experiments involving the
additions of known numbers of organisms containing target DNA sequences to
soil and sediment samples have also been used to estimate the sensitivity of
these analyses (9,12).  While differences in the ratios of target to non-
target organisms have the potential to significantly affect such estimates,
the values reported in these studies have coincided fairly closely at
approximately 104 cells/gram of sample.

      This degree of sensitivity, while useful for many applications, may be
insufficient for the detection of many microbial strains or genotypes
present in low numbers or in a high background of non-target organisms.
Recent studies have demonstrated, however, that target DNA sequences can be
specifically amplified within heterogeneous DNA mixtures by employing the
polymerase chain reaction or PCR procedure (12).  Target DNA amplification
by this procedure has been shown to increase the sensitivity of DNA probe
hybridization analyses for detecting specific organisms by at least 100-
fold.

                         FUNCTIONAL GENE PROBES

      Hybridization analyses using DNA probes can provide descriptive
information concerning the composition and functional attributes of
microbial communities with a minimum of time and effort.  Consequently,
these analyses would find numerous uses in monitoring alterations and
adaptive responses of microbial communities resulting from their exposure to
toxic chemicals.  Perhaps of greater practical importance, however, is the
potential predictive information available from these analyses for
determining the degradative capabilities of different microbial communities
toward such chemicals.  While it is known that numerous factors can affect
degradative activity, considerable effort has been directed towards
examining the correlation between the occurrence of appropriate genotypes
and actual degradation.  Studies on the relative numbers of colony-forming
organisms that hybridized with a probe consisting of the entire naphthalene
degradative plasmid, Nah 7, showed a 30-fold higher level of such organisms
in a sediment sample that had previously been exposed to a mixture of
hydrocarbons compared with in an unexposed sediment sample from the same
source (13).  These results corresponded with a 15-fold higher rate of
naphthalene mineralization by the microbial community in the exposed sample.
Similarly, changes in the abundance of colony-forming organisms that
hybridized with the Nah 7 plasmid probe were found to closely approximate


                                     306

-------
changes in naphthalene mineralization in a continuously-operated bioreactor
(11).  Other studies have been performed using DNA probes for Hg2+-
resistance(mer) genes.  Genes of this type are involved in the reduction of
Hg2+ to elemental mercury  (Hg°) which results in rapid elimination of this
chemical  by volatilization.  Hybridization analyses using mer gene probes
showed significantly higher frequencies of these genes in DNA samples
directly extracted from Hg2+-contaminated  freshwater and  sediment  samples
than in similar DNA extracts from corresponding uncontaminated samples
(14,15).   These results were shown to correlate with higher rates of Hg°
volatilization and higher frequencies of Hg2+-resistant organisms  occurring
in the contaminated samples.

      Such studies have thus far indicated that hybridization analyses using
functional gene probes may in some but not all instances provide predictive
information concerning the ability of microbial populations to degrade or
otherwise eliminate toxic chemicals.  While as stated earlier, numerous
factors in addition to the frequencies of such genes can affect overall
rates of these processes, current evidence suggests that in most instances
where the results of hybridization analyses have proven to be non-
predictive, other isofunctional but non-homologous genes may have been
present in the microbial communities.  This point can be illustrated in
studies on Hg2+-resistant  organisms  from several  saltwater  communities
(15,16).   In each instance the frequencies of Hg2+-resistant  organisms  as
determined by platings on Hg2+-containing  media were shown  to correlate with
Hg° volatilization  rates by these communities.   Hybridization analyses
showed, however, that the large majority of these resistant organisms did
not contain genes which were homologous to the available mer gene probes.
Similarly, studies on toluene degradation in a groundwater aquifer microcosm
showed that this activity could not be correlated with the frequency of
organisms that hybridized to a toluene degradative gene probe originating
from the Tol plasmid (17).  Other studies have shown,  however, that at least
three distinct pathways for toluene degradation in addition to the pathway
encoded for by the Tol plasmid exist in different bacteria (18-21).  Since
it  is unlikely that the genes specifying these pathways would show homology
to Tol plasmid genes, such organisms would not have been identified in these
studies.   These examples show that the use of functional  gene hybridization
analyses for accurate predictive assessments of a microbial community's
degradative capability requires probes for all of the relevant genotypes
that may be present in the community.  In many instances, this technology
may therefore not yet be applicable due to the unavailability of the
appropriate probes.

                    PHYLOGENETICALLY-BASED PROBES

      Comparative nucleotide sequence analysis of ssrRNAs has become
recognized as a highly useful method for defining phylogenetic relationships
between different organisms.  Such relationships have been inferred from
observations that different portions of these universally-occurring
molecules undergo nucleotide sequence divergence at different evolutionary
rates (3).  An important offshoot of this research has been the recognition
that this variability of nucleotide sequence conservation in different
portions of ssrRNA molecules can form the basis for the construction of DNA


                                     307

-------
hybridization probes, specific for various phylogenetic groupings from the
kingdom down to the subspecies level (22-26).  In addition, these analyses
have provided the basis for the development of new methods that allow the
rapid determination of ssrRNA sequences (26,27).

      Due to the recent nature of these developments, the data base of
sequenced ssrRNAs for different organisms is still relatively small and the
number of studies using hybridization probes generated from this data base
to evaluate microbial communities in natural environments is even smaller.
The applicability of such an approach, however, has been recently
illustrated in an analysis of the responses of various strains of
Bacteroides succinogenes and Lachnospira multipants-like organisms in a
bovine rumen that were exposed to the ionophore antibiotic, monensin, using
several species and group-specific probes (23).  As in hybridization
analyses using directly extracted DNA, the detection and enumeration of
ssrRNA sequences does not require prior cultivation of organisms from the
communities under investigation.  In addition, methods have been recently
described which would allow for the sequencing and construction of probes
for ssrRNAs from non-culturable organisms (26,28).

TOWARDS THE DEVELOPMENT OF A COMPREHENSIVE MONITORING SYSTEM BASED ON DNA
                    HYBRIDIZATION PROBES

      As previously discussed, determinations of functional degradative gene
frequencies show promise as predictive indicators of the degradative
capabilities of microbial communities toward hazardous chemicals, provided
that gene probes for all of the relevant degradative pathways are available.
In instances where such probes are not available, ssrRNA-specific probes for
phylogenetic groups with demonstrated degradative capabilities towards the
compounds in question may be a practical alternative.  In view of the rapid
procedures now available for obtaining ssrRNA sequence information from both
culturable and non-culturable oranisms, and, because this technology does
not require extensive preliminary characterization of the genes involved,
the amounts of time and effort required for the development of such probes
should be relatively small.  Significant questions remain, however,
concerning the usefulness of such probes, particularly in the field of
biodegradation research.  These stem in part from current uncertainties as
to the specificity of such probes, particularly at the sub-species level.
It can be anticipated that ssrRNA sequence information for large numbers of
new microbial species and strains will be added to the presently existing
data base in the near future which should clarify much of this uncertainty.
Of potentially greater importance is the question of whether specific
degradative capabilities are universally associated with specific
phylogenetic groups.  For example, it has been well documented that
degradative genes for a number of toxic and/or synthetic compounds are
carried on transmissible plasmids which do not necessarily occur within all
members of a given phylogenetic group (29).  Conversely, there is presently
a great need to further establish the role of degradative gene transmission
(via plasmids or other means) within microbial populations as a possible
adaptive mechanism in community responses to toxic chemicals.

      Our laboratory has been involved in the isolation, from different


                                     308

-------
activated sludge sources, of microbial strains that show degradative
activity toward several pollutant chemicals.  As shown in Table 1, the
majority of these test compounds are degraded by a number of distinct
microbial strains which in all likelihood do not yet represent all of the


TABLE 1.   NUMBERS OF INDEPENDENT BACTERIAL STRAINS SHOWING CATABOLIC
           ACTIVITIES TOWARD AROHATIC HYDROCARBON TEST COMPOUNDS FROM
                     TWO WASTEWATER TREATMENT SYSTEMS1'2
Test Compound
Toluene
Benzene
Ethyl Benzene
Phenol
o-Cresol
2, 4-Dimethyl phenol
Dimethyl phthal ate
D1 ethyl phthal ate
Di butyl phthal ate
Phthal ic acid

BP3
7
5
5
8
3
2
0
0
0
6
Source of strains
HL4
3
5
2
12
NO
0
2
2
2
ND
1 Catabolic activity established on the basis of growth in the presence of
the test compounds as sole carbon sources.

2 Strains differentiated by comparisons of agarose gel electrophoretic
banding patterns of total genomic DNA fragments following digestion with
restriction endonucleases.

3 Activated sludge from a wastiwater treatment system at the British
Petroleum facility in Lima, Ohio.

4 Activated sludge from a wastewater treatment facility at the Hoffman-
LaRoche facility 1n Nutley, New Jersey.

ND: not determined
                                     309

-------
organisms from these communities with such capabilities.  Our initial goal
will be to establish phylogenetic relationships between these and other
degradative strains as they are isolated based on both traditional
determinative methods such as the API and Biolog test systems as well as by
comparative analyses of their ssrRNA nucleotide sequences.

      The sequence data will subsequently be used to construct DNA
oligonucleotide probes that are specific to each of the phylogenetic groups
identified.  These probes will then be used to determine the distributions
in the original sludge populations of specific degradative phenotypes within
each phylogenetic group.  They will also be used to determine the frequency
of organisms specific to each group within the total sludge microbial
communities and whether these frequencies show correlations with the
degradative activity exhibited by these sludge communities toward the
different test compounds.  Where available, DNA hybridization probes for
relevant functional degradative genes will also be employed in parallel
analyses.

      In addition to these studies, samples of the original sludges will be
used to inoculate continuously-operated bench scale bioreactors fed with
sterilized conventional wastewater and varying loads of the test compounds.
Monitoring of the microbial populations within these reactors will again be
performed by hybridization analyses using phylogenetic and functional gene-
specific probes. These studies will be used to further assess the extent of
correlation between degradative activity and the frequencies of these
organisms and/or genes within the overall microbial communities as well as
to identify adaptive response mechanisms within these communities to high
concentrations of the different compounds.


                               ACKNOWLEDGMENTS

      This work is supported by Cooperative Agreement CR-816337 from the
U.S. Environmental Protection Agency, Risk Reduction Engineering Laboratory,
Office of Research and Development, P.R. Sferra, Project Officer.


                              LITERATURE CITED

1.    Levin, M.A., R. Seidler, A.W. Borquin, J.R. Fowle III, and T. Barkay.
      1987. EPA developing methods to assess environmental release.
      Bio/Technology 5:38-45.

2.    Olsen, 6.J., D.J. Lane, S.J. Giovannoni and N.R. Pace. 1986. Microbial
      ecology and evolution: a ribosomal RNA approach. Ann. Rev. Microbiol.
      40:337-365.

3.    Woese, C.R. 1987. Bacterial evolution. Microbiol. Rev. 51:221-271.

4.    Grunstein, M. and D.S. Hogness. Colony hybridization: a method for the
      isolation of cloned DNAs that contain a specific gene. Proc. Natl.
      Acad. Sci. U.S.A. 72:3961-3965.


                                     310

-------
5,    Maas, R. 1983. An improved colony hybridization method with
      significantly increased sensitivity for the detection of single genes.
      Plasmid 1:417-420.

6.    Atlas, R.M. 1982. Enumeration and estimation of biomass of microbial
      components in the biosphere, p. 84-102. In: R.6. Burns and J.H. Slater
      (ed.), Experimental microbial ecology. Blackwell Scientific
      Publishers, Oxford.

7.    Roszak, D.B. and R.R. Colwell. 1987. Survival strategies of bacteria
      in the natural environment. Microbiol. Rev. 51:365-379.

8.    Steffan, R.J., J. Goksoyr, A.K. Bej and R.M. Atlas. 1988. Recovery of
      DNA from soils and sediments. Appl. Environ. Microbiol. 54:2908-2915.

9.    Holben, W.E., J.K. Jansson, B.K. Chelm and J. M. Tiedje. 1988. DNA
      probe method for the detection of specific microorganisms in the soil
      bacterial community. Appl. Environ. Microbiol. 54:703-711.

10.   Qgram, A., G.S. Sayler and T. Barkay. 1987. The extraction and
      purification of microbial DNA from sediments. J. Microbiol. Meth.
      7:57-66.

11.   Sayler, G.S., R.K. Jain, A. Ogram, C.A. Pettigrew, L. Houston, J.
      Blackburn and W.S. Riggsby. 1986. Applications for DNA probes in
      biodegradation research, p. 499-508. In: F. Megusar and M. Gantar
      (ed.), Perspectives in microbial ecology. Slovene Society for
      Microbiology, Ljubljana.

12.   Steffan, R.J. and R.M. Atlas. 1988. DNA amplification to enhance
      detection of genetically engineered bacteria in environmental
      samples. Appl. Environ. Microbiol. 54:2185-2191.

13.   Sayler, G.S., M.S. Shields, E.T. Tedford, A. Breen, S.W. Hooper,
      K.M. Sirotkin and J.W. Davis. 1985. Application of DNA-DNA colony
      hybridization to the detection of catabolic genotypes in environmental
      samples. Appl. Environ. Microbiol. 49:1295-1303.

14.   Barkay, T. and H. Pritchard. 1988. Adaptation of aquatic microbial
      communities to pollutant stress. Microbiol. Sci. 5:165-169.

15.   Barkay, T., C. Liebert and M. Gillman. 1989. Hybridization of DNA
      probes with whole-community genome for detection of genes that encode
      microbial responses to pollutants: mer genes and Hgz* resistance.
      Appl. Environ. Microbiol. 55:1574-1577.

16.   Barkay, T. 1987. Adaptation of aquatic microbial communities to Hg2+
      stress. Appl. Environ. Microbiol. 53:2725-2732.

17.   Jain, R.K., G.S. Sayler, J.T. Wilson, L. Houston and D. Pacia. 1987.
      Maintenence and stability of introduced genotypes in groundwater
      aquifer material. Appl. Environ. Microbiol. 53:996-1002.


                                     311

-------
18.   Harayama, S., R.A. Leppik, M. Rekik, N. Mermod, P.R. Lehrbach, W.
      Reinke and K.N. Timmls. 1986. Gene order of the TOL catabolic plasmid
      upper pathway operon and oxidation of both toluene and benzyl alcohol
      by the xylA product. J. Bacteriol. 167:455-461.

19.   Zylstra, G. and D.T. Gibson. 1989. Toluene degradation by Pseudomonas
      putida Fl. Nucleotide sequence of the toddC2BADE genes and their
      expression in Eschen'chia coli. J. Biol. Chem. 264:14940-14946.

20.   Winter, R.B., K.-M. Yen and B.D. Ensley. 1989. Efficient degradation
      of trichloroethylene by a recombinant Eschen'chia coli. Bio/Technology
      7:282-285.

21.   Shields, M.S., S.O. Montgomery, P.J. Chapman, S.M. Cuskey and P.M.
      Pritchard. 1989. Novel pathway of toluene catabolism in the
      trichloroethylene-degrading bacterium G4. Appl. Environ. Microbiol.
      55:1624-1629.

22.   Giovannoni, S.J., E.F. DeLong, G. Olsen and N.R. Pace. 1988.
      Phylogenetic group-specific oligodeoxynucleotide probes for
      identification of single microbial cells. J. Bacteriol. 170:720-726.

23.   Stahl, D.A., B. Flesher, H.R. Mansfield and L. Montgomery. 1988. Use
      of phylogenetically based hybridization probes for studies of ruminal
      microbial ecology. Appl. Environ. Microbiol. 54:1079-1084.

24.   Putz, J., F. Meinert, U. Wyss, R.-U. Ehlers and E. Stackebrant. 1990.
      Development and application of oligonucleotide probes for molecular
      identification of Xenorhabdus species. Appl. Environ. Microbiol.
      56:181-186.

25.   Amann, R.I., L. Krumholtz and D.A. Stahl. 1990. Fluorescent-
      oligonucleotide probing of whole cells for determinative, phylogenetic
      and environmental studies in microbiology. J. Bacteriol. 172:762-770.

26.   Barry, T., R. Powell and F. Gannon. 1990. A general method to generate
      DNA probes for microorganisms. Bio/Technology 8:233-236.

27.   Lane, D.J., B. Pace, G.J. Olsen, D.A. Stahl, M.L. Sogin and N.R. Pace.
      1985. Rapid determination of 16S Mbosomal RNA sequences for
      phylogenetic analyses. Proc. Natl. Acad. Sci. U.S.A. 82:6955-6959.

28.   Weller, R. and D.M. Ward. 1989. Selective recovery of 16S rRNA
      sequences from microbial communities in the form of cDNA. Appl.
      Environ. Microbiol. 55:1818-1822.

29.   Rlttmann, B.E., B.F. Smets and D.A. Stahl. 1990. The rolo of genes in
      biological process. Environ. Sci. Technol. 24:23-29.
                                     312

-------
   BIODEGRADABILITY STUDIES OF THE ALASKAN WEATHERED CRUDE OIL CONSTITUENTS
      WITH THE USE OF ELECTROLYTIC RESPIROMETRY AND SHAKER FLASK SYSTEMS
               Henry  H.  Tabak1, Sanjay Desai2, Albert D. Venosa1,
         John  A.  Glaser1, John R. Haines1, and Wipawan Nisamaneepong3
                     'U.S.  Environmental  Protection Agency
                     Risk Reduction  Engineering  Laboratory
                            Cincinnati, Ohio 45268

                Department of  Chemical and  Nuclear Engineering
                           University of Cincinnati
                            Cincinnati, Ohio 45221

                            ng  and  Economics  Resear
                            Cincinnati, Ohio 46268
3Engineering and Economics Research Systems
                                   ABSTRACT

      Bench-scale biodegradability studies of the Alaskan weathered crude
oil constituents were undertaken as part of the overall bioremediation
project for the shorelines of Prince William Sound, Alaska, contaminated by
the Exxon oil spill.  The purpose of the studies was to evaluate the
capability of the indigenous microbial consortium of the sea water and the
coastline and island beach areas in the vicinity of the oil spill, to
biodegrade and/or biotransform the spilled weathered crude oil alkane
hydrocarbon and polynuclear aromatic (PAH) constituents in batch-type
respirometric reactors and shaker flask systems.

      The specific objectives of the current studies were to utilize the
electrolytic respirometry oxygen uptake and shaker flask gas chromatographic
analytical data in order to:  (1) determine the biodegradability of the alkane
and PAH constituents of the weathered crude oil by the indigenous microbiota,
(2) evaluate the capability of varied fertilizer formulations (i.e.,
oleophilic fertilizer - Inipol, slow release water soluble fertilizer -
Sierra, etc.) to serve as nutrients for promoting an effective biotreatment of
these organics, and (3) assess any potential inhibitory effects of the crude
oil constituents on both the indigenous biomass and their biodegradative
activity on readily biodegradable or biogenic substrates.  Additionally,
studies incorporated the use of Valdez, Alaska waste treatment plant oil bilge
biomass to determine its capacity to bio-oxidize the weathered crude oil
constituents.

      The data obtained which demonstrate almost complete utilization of the
alkane hydrocarbon constituents in presence of Inipol EAP 22 fertilizer and
very significant biodegradation of the PAH constituents in the culture media
after 6 weeks of incubation.  Growth data indicate significant increase of
growth of indigenous oil degrading microbiota (increase of 4 to 6 orders of
magnitude) at the end of 6-week incubation period.  Results from shaker flasks
                                     313

-------
based on growth data and on the GC scans and GC/MS analysis data for the
alkane hydrocarbons and PAH constituents of the weathered crude oil
demonstrate similar significant biodegradation of PAHs and complete
dissimilation of alkane hydrocarbons in presence of Inipol in culture media
after 6 weeks of incubation.

      Control respirometric and shaker flask experiments without Inipol  EAP 22
provided data that demonstrate absence of biodegradative activity of the
microbiota on the alkane hydrocarbons and PAHs oil components.


                                 INTRODUCTION

      The U.S. EPA Alaska Bioremediation Project was initiated in the
aftermath of the March 24, 1989, Exxon Valdez oil spill in Prince William
Sound, Alaska.  Hydrocarbon-degrading bacteria are known to exist in the
waters and sediments of the Prince William Sound, but in the presence of large
amounts of degradable petroleum, the growth and activity of the bacteria
becomes limited by the availability of nitrogen and phosphorus.  The objective
of the bioremediation project was to determine if techniques for accelerating
the hydrocarbon bioremediation rates through the use of indigenous,  mixed
populations of hydrocarbon-degrading microorganisms can be utilized in the
cleanup of the spill and to demonstrate a method of enhancing the cleanup of
oil contaminated shorelines by adding nutrients to stimulate the growth of
naturally occurring oil-degrading microbiota.

      The bioremediation project was composed of four complementary field
studies: (1) a beach substrate study which attempted to assess the quantity
and constituents of the oil in the substrate and their change over time (by
determination of decrease of C17:pristane ratio, and the C18:phytane ratio in
case of alkane hydrocarbons and the biodegradability rates of polycyclic
aromatic hydrocarbons); (2) a nutrient addition project to evaluate the
success of applying different nutrient additions (inorganic nitrogen and
phosphorus) that are expected to enhance the biodegradation of the contaminant
oil; (3) a microbiology study to. assess the effect of the fertilizers on the
stimulation of oil-degrading bacteria both in the oil-bearing substrate and in
the interstitial water within the substrate, and to significantly increase the
population of oil degrading microbiota; and (4) a longer term assessment to
evaluate the effects of the treatment on the Prince William Sound beaches and
waters.

      Laboratory bench-scale biodegradability studies of the Alaskan weathered
crude oil constituents using closed systems were undertaken as part of the
overall bioremediation project.

      The specific objectives of the current studies were to utilize the
electrolytic respirometry oxygen uptake and shaker flask gas chromatographic
analytical data in order to: (1) determine the biodegradability of the alkane
and PAH constituents of the weathered crude oil by the indigenous microbiota,
(2) evaluate the capability of varied fertilizer formulations to serve as
nutrients for promoting an effective biotreatment of these organics, and (3)

                                     314

-------
assess any potential inhibitory effects of the crude oil constituents on both
the indigenous biomass and their biodegradative activity on readily
biodegradable or biogenic substrates.  Additionally, samples of the Alyeska
Ballast Water Treatment Plant sludge were evaluated as enhancements to the
natural microbiota for biodegrading the oil.

      Analytical respirometry served as the primary experimental approach for
studying the rates of biodegradation of oil constituents.  Supplementing the
analytical respirometry experiments were batch flask microcosms to provide
additional testing of the biodegradability of the individual aliphatic and
aromatic constituents of the crude oil.

      In addition to the oxygen uptake (consumption) measurements generated
from the respirometric reactors,  GC/MS chromatogram scans of alkane
hydrocarbons and GC/MS analysis of aromatic hydrocarbons (PAHs) in aqueous
culture sample extracts were taken at various time intervals from
respirometric vessels and  shaker flask microcosms.  This was done to
determine which of the oil constituents were biodegraded in the closed reactor
systems.

      The oleophilic fertilizer, Inipol EAP 22, containing urea as the
nitrogen source and lauryl phosphate as the phosphorus source in an oleic acid
external phase, and manufactured by Elf Aquitaine,  was used as the nutrient
fertilizer.  The ability of Inipol to provide an adequate amount of nitrogen
and phosphorus for the microbial degradation of weathered crude oil was
compared to that of the OECD minimal salts medium containing inorganic and
water soluble nutrients.

      These laboratory experiments were conducted to support the field work,
which was much less controllable and much more susceptible to nutrient
washout, re-oiling, tidal cycles and washout, temperature variation, climatic
changes, freshwater/saltwater interactions, etc., all of which would confound
interpretation of the findings.


                                  BACKGROUND


BIODEGRADATION OF PETROLEUM HYDROCARBONS IN MARINE ECOSYSTEMS

      The expansion of petroleum development into new frontiers, such as deep
offshore waters and ice-dominated arctic environments, and the apparently
inevitable spillage which occur during routine operations and as a consequence
of acute accidents have dramatically focused an interest on the problems of
oil pollution of marine ecosystems.

      Most previous reviews concerning the microbial degradation of petroleum
pollutants have been concerned with the marine environment (Atlas, 1977; Atlas
and Bartha, 1973a; Atlas and Schofield, 1975; Bartha and Atlas, 1977; Colwell
and Walker, 1977; Crow et a?., 1974; Floodgate, 1972a, 1972b, 1973, 1976;
Jordan and Payne, 1980; Karrick, 1977; National Academy of Sciences, 1975; Van
der Linden, 1978; ZoBell, 1946, 1964, 1969, 1973).  Atlas's review (1982)
                                     315

-------
expanded the scope to Include consideration of the fate of petroleum
hydrocarbons in freshwater and soil ecosystems.


ENHANCEMENT OF NATURAL BIODEGRADATION OF CRUDE OIL IN MARINE ENVIRONMENT

      The fate of petroleum hydrocarbons in the environment is largely
determined by abiotic factors which influence the weathering, including
biodegradation of the oil.  Factors which influence rates of microbial growth
and enzymatic activities affect the rates of petroleum hydrocarbon
biodegradation.  The persistence of petroleum pollutants depends on the
quantity and quality of the hydrocarbon mixture and on the properties of the
affected ecosystem.  In one environments petroleum hydrocarbons can persist
indefinitely, whereas under another set of conditions>the same hydrocarbons
can be completely biodegraded within a relatively few hours or days.

Dispersants

      Dispersants have been used to treat oil spills.  In some cases the use
of toxic dispersants probably has resulted in greater ecological impact than
the oil spill itself; such was the case in the Torrey Canyon incident (Cowell,
1971; Smith, 1968).  Some dispersants may contain chemicals which are
inhibitory to microorganisms.  Without toxicity, however, dispersion can
enhance petroleum biodegradation.  Mulleins-Phillips and Stewart (1974a) found
that some dispersants enhanced n-alkane degradation in crude oil, but that
other dispersants had no effect.  Gatellier et a?., (1971, 1973) and Robichaux
and Myrick (1972) likewise found that some dispersants inhibited hydrocarbon-
oxidizing populations, whereas others enhanced hydrocarbon-degrading
microorganisms.  Atlas and Bartha (1973b) tested several dispersants and found
that all increased the rate but not the extent of hydrocarbon mineralization.

Nutrients

      There is some confusion and considerable apparent conflict in the
literature regarding the limitation of petroleum biodegradation by available
concentrations of nitrogen and phosphorus in seawater.  Several investigators
(Atlas and Bartha, 1972a; Bartha and Atlas, 1973; Floodgate, 1973, 1979;
Gunkel, 1967; LePetit and Barthelemy, 1968; LePetit and N'Guyen, 1976) have
reported that concentrations of available nitrogen and phosphorus in seawater
are severely limiting to microbial hydrocarbon degradation.  Other
investigators (Kinney et a7., 1969), however, have reached the opposite
conclusion, i.e., that nitrogen and phosphorus are not limiting in seawater.
The difference in results appears to be based on whether the studies are aimed
at assessing the biodegradation of hydrocarbons within an oil slick or the
biodegradation of soluble hydrocarbons.

      Extensive studies have been made which demonstrate enhanced biodegra-
dation of crude oil due to addition of nitrogen and phosphorus [Floodgate
(1973), Bridie and Bos (1971), Atlas and Bartha (1972) Reisfeld et a?.,
(1972), Colwell et a7., (1978), Ward and Brock (1976), LePetit and N'Guyen
(1976), Gibbs (1975), Van Der Linden (1978).  Oil-degrading microorganisms are
ubiquitous and can metabolize a wide range of oil components under diverse

                                     316

-------
environmental conditions (Atlas, 1981; Mulkins-Phillips and Stewart, 1974b),
although the rates of oil biodegradation in the marine environment are
generally slow because of the lack of optimal ratios of carbon, nitrogen and
phosphorus.

      Inoculation "seeding" experiments have repetitively demonstrated that
specific cultures of oil-degrading bacteria fail to enhance the hydrocarbon
degradation capability of natural environments because they typically
disappear from dominant microflora.  The indigenous microflora, on the other
hand, show a capacity for the adaptation to petroleum degradation within days
[Bartha and Atlas (1977); Lee and Levy (1987); Taggar et a/.,  (1983)] under
conditions where the rates of petroleum biodegradation is limited by nutrient
deficiencies.  The beneficial effect of fertilization with nitrogen and
phosphorus has been conclusively demonstrated and offers great promise as a
countermeasure for combating oil spills [Atlas (1981); Delaune et a 7 (1980);
Gibbs and Davis (1970); Olivieri et al. (1976)].  The optimum ratios of
carbon, nitrogen and phosphorus to support maximum oil biodegradation ratios
have been defined by laboratory studies [Atlas and Bartha (1972); Bridie and
Bos (1971)].

      The addition of nitrogen- and phosphorus-containing fertilizers can be
used to stimulate microbial hydrocarbon degradation.  In addition, oleophilic
nutrients have been developed to overcome the problem of the water soluble
nutrients being rapidly removed from the oil-water interface where
biodegradation occurs [Atlas and Bartha (1973c); Sirvins and Angles (1986)].

      Bergstein and Vestal (1978) studied the biodegradation of crude oil in
Arctic tundra ponds.  They concluded that oleophilic fertilizer may provide a
useful tool to enhance the biodegradation of crude oil spilled on such
oligotrophic waters.  Without addition of nitrogen and phosphorus, hydrocarbon
biodegradation was limited.  Atlas and Bartha (1973c) described an oleophilic
nitrogen and phosphorus fertilizer which could overcome limitations of
nitrogen and phosphorus in seawater and stimulate petroleum biodegradation in
seawater.

      Olivieri et a/., (1976) described a slow-release fertiliser containing
paraffin-supported magnesium ammonium phosphate as the active  Ingredient for
stimulating petroleum biodegradation.  They reported that the biodegradation
of Sarlr crude oil 1n seawater was considerably enhanced by addition of the
paraffin-supported fertilizer.  Kator et a7. (1972) suggested the use of
paraffinized ammonium and phosphate salts for enhancing oil biodegradation in
seawater.

      Studies of Tramier and S1rv1ns (1983) generated a unique approach to
render the enhancement of oil biodegradation process in marine environment
operational.  A nutrient Including nitrogen, phosphorus and carbon has been
developed and formulated as an oleophilic microemulslon.  The nutrient has
been tested 1n laboratory and field situations 1n temperate and cold climates.
Biodegradation rates have ranged from 60 percent to 85 percent depending on
conditions.
                                     317

-------
      Bronchart  et a7.  (1985) developed an oil dispersant with oleophilic
nutrients made up of an  oleophilic part (normal paraffine and olefine) and of
a hydrophilic moiety (containing nitrogen and/or phosphorus) which exhibited a
high dispersion efficiency and a quick-starting biodegradation of the
dispersed crude oil.

      In a study of biodegradation of a weathered crude oil emulsion (Halmo,
1985) direct addition of fertilizer to oil on shore enhanced the natural
biodegradation.  An oil-soluble urea fertilizer was at least as efficient as a
water-soluble one containing ammonia and nitrate.  In one year, the paraffins
were totally degraded.   Composting was studied both in aerated full-scale
windrows and in model columns.

      Laboratory tests conducted by Lee and Levy (1987) at 3°C and 20eC
indicate that biodegradation of a light crude oil (condensate) by indigenous
microflora of sandy beach environments can be enhanced by the addition of an
oleophilic nutrient.  Their field trials conducted over a 204-day period in
the intertidal zone of a sandy beach in Atlantic Canada suggest that the most
promising approach to enhancing the biodegradation of oil-stained sandy beach
environments is through  periodic replenishment of nutrients after the
indigenous microflora have adapted to the contaminated sediments.

      Studies of Sveum (1988) and Syeum and LaDousse (1984) demonstrated that
oil exploration in arctic regions will require special oil spill cleanup
methods for shorelines.  The application of fertilizers to speed
biodegradation of oil may hold promise.  Inipol EAR 22, an oil-soluble
fertilizer that is nontoxic and biodegradable, was tested in a series of
experiments on Spitsbergen, Norway, to determine its effectiveness under
various conditions.  Uptake and exchange of nutrients depends on the complex
growth kinetics of the bacteria involved and requires detailed study.

      Field experiments  (Lee and Levy, 1989), in an aerobic intertidal sandy
beach environment (low energy coastal systems), demonstrated that the
biodegradation of Scotian Shelf condensate and Hibernia crude oil can be
substantially accelerated by periodic application of agricultural fertilizers.

                             MATERIALS  AND METHODS


GENERAL EXPERIMENTAL APPROACH

Nutrient Media

      Two nutrient formulations were compared for their ability to support the
growth of hydrocarbon degraders on weathered Prudhoe Bay crude oil.  The
oleophilic fertilizer Inipol EAP 22 contains urea as the nitrogen source and
lauryl phosphate as the  phosphorus source encapsulated in an oleic acid
external phase.  Because of its oleophilic nature,  it enables the
microorganisms to perform their hydrocarbon metabolic activities at an optimum
rate at the oil-water surface, and it acts by stimulating the growth of
hydrocarbon specific bacteria.

                                     318

-------
      The second nutrient solution used in these studies was an OECD synthetic
medium (OECD Guidelines for Testing of Chemicals, OECD, Paris, France, 1981}
consisting of measured amounts per liter of deionized distilled water of (1)
mineral salts solution; (2) trace salts solution, and (3) a solution (150
mg/L) of yeast extract as a substitute for vitaminsolution.

Hicrobial Inoculum

      The microbial inoculum consisted of seawater from Snug Harbor, beach
material  collected from an uncontaminated beach in Valdez, and weathered crude
oil from the spill.  Some vessels and flasks were additionally supplemented
with indigenous biomass from the Alyeska ballast water treatment plant.

      For seawater source inocula, seawater was used as diluent after
appropriate volumes of the stock solutions of the mineral salts medium and
yeast extract was added to reactors in addition to the oil (mousse) sample.
The sludge biomass from the oil bilge waste treatment system was added to the
medium at a concentration of 10 ml or at other desirable concentration levels.
Beach sand/rock was added in a measured amount to the bottom of the
respirometric vessels and shaker flasks to serve as source of the beach
surface and subsurface microbial population.  Regardless of the microbial
inoculum source, seawater served as the diluent liquid after all nutrients
beach sand and the crude oil substrate were added to the respirometric
reactors and shaker flask microcosms.

Test and Control Substrates

      The test substrate used in the study was the Prudhoe Bay weathered crude
oil from the oil polluted beaches of Prince William Sound, containing the
alkane, paraffinic, polycyclic aromatic and asphaltic organic constituents.
Aniline was used as the  biodegradable control reference compound at a
concentration level of 100 mg/L in the toxicity control systems..


ANALYTICAL RESPIROMETRY

      The electrolytic respirometry approach to determine the biodegradability
of the organic test compounds in this study was chosen because of the specific
advantages of the respirometric methods over that of manometric procedures in
tracking oxygen utilization during the exertion of biochemical oxygen demand
(BOD).

      A comprehensive description of the procedural steps of the respirometric
tests and of the experimental design employing test and control systems is
presented elsewhere [OECD Guidelines for Testing of Chemicals (1983), Tabak et
a7.  (1984)].

      For fully automatic data acquisition, frequent recording and storage of
large numbers of oxygen uptake data, the Sapromat B-12 recorders are
interfaced to an IBM-AT computer via the Metrabyte interface system.  The use
of Laboratory Handbook software package allows the collection of data at 15
minute intervals.
                                     311

-------
Experimental Design

      The typical experimental system consisted of duplicate flasks for three
concentrations of weathered crude oil (1,000 mg/L, 300 mg/L and 100 mg/L) each
containing Inipol EAP 22 at 5 percent of the oil weight or OECD nutrient
concentrate solution (2x standard concentration).  All the vessels contained 2
g of uncontaminated beach sand from Valdez, Alaska, and 1,000 mg/L of seawater
collected off-shore at Snug Harbor, an inlet on Knight Island in Prince
William Sound.  In addition, the experimental system incorporated duplicate
flasks containing each 1,000 mg/L oil, 10 mL of Alyeska ballast sludge, 2
grams of unpolluted beach sand and Inipol EAP 22 at 5 percent of oil weight,
diluted with 1,000 mL of seawater.  The respirometric control system consisted
of single flasks representing (1) inoculum control; (2) inoculum/nutrient
control; (3) toxidty control for testing of toxicity of oil to aniline
biodegradatlon; (4) positive reference control (aniline + oil + seawater +
beach sand; (5) abiotic substrate control; (6) no nutrient control; no
Inipol); and (7) substrate/nutrient control.  The vessels containing beach
material, oil and Inipol were charged by first adding the beach sand, then
adding the measured amount of oil onto the sand, followed by adding one
measured amount of Inipol to the oiled rocks, and finally by filling the
vessel with Snug Harbor seawater.  The experimental design of the respirometry
experiments is summarized in Table 1.

      The contents of the reaction vessels were first stirred for an hour to
ensure an endogenous respiration state at the initiation of oxygen uptake
measurements.  The incubation period of the experimental run was extended to
30 weeks.

IncubationTemperature.

      The reaction vessels were incubated at 15eC in the dark (enclosed in the
temperature controlled waterbath) and stirred continuously throughout the run
by magnetic stirrers.  This temperature was chosen since it was estimated, to
be the average summer temperature in Valdez, Alaska.

Respirometric Vessels.

      The reaction vessels used in the study were 1,200 mL capacity flasks
with two side arms at lower and upper positions to allow intermittent sampling
of the aqueous and gaseous phase for residual parent oil component compounds,
intermediate and end products of oil metabolism and the liquid phase to
determine the growth rate of indigenous microbiota.  The total volume of the
culture/medium/nutrients/substrate in the respirometric reactor was 1,000 mL.
The small amounts of liquid sampled throughout the experimental run did not
have any significant effect on the cumulative oxygen uptake values and the
generated oxygen uptake velocity curves.


FLASK EXPERIMENTS

      Flask microcosms were set-up to provide further support of the
respirometric studies.  Concentrations of weathered crude oil as high as

                                     320

-------
10,000 mg/L were used to determine if such levels were inhibitory to the
biodegradation process.  Thirty 2-liter capacity flasks were integrated into
the program consisting of 18 experimental and 12 control  flask microcosm
systems.


Experimental Design

      The experimental system consisted of (1) duplicate flasks for three
concentrations of weathered crude oil (10,000 mg/L, 3,000 mg/L and 1,000 mg/L)
each containing Inipol EAR 22 at 5 percent concentration of the oil weight (2)
duplicate flasks containing 10,000 mg/L weathered crude oil, Iriipol EAP22 at 5
percent concentration of oil weight and Alyeska ballast sludge (10 mL) as an
additional inoculum supplement; and (3) duplicate flasks containing 10,000
mg/L weathered oil, OECD medium concentrate (2x standard concentration) in
place of Inipol EAR 22 as a nutrient and with or without Alyeska ballast
sludge (10 mL) as an additional inoculum supplement.  Each flask equally
contained 20 grams of uncontaminated beach sand from Valdez and 1,000 mL Snug
Harbor seawater.  The flasks were charged with various additives in the same
fashion and order as above.  In addition, the experimental system incorporated
duplicate flasks for three weights of polluted beach sand (5, 10 and 15 g),
each set of duplicates containing Inipol EAP 22 at 2.5, 5.0 and 7.5 mg/L and
1,000 mL of Snug Harbor seawater.  The shaker flask control system consisted
of single flasks representing inoculum controls; inoculum/nutrient controls;
toxicity controls and positive reference controls.  The experimental design of
the shaker flask experiments are summarized in Table 2.

      The flask microcosms were placed on a rotary New Brunswick Shaker and
incubated in a dark constant temperature room at 15*C.


ANALYTICAL METHODOLOGY FOR ALKANE HYDROCARBONS AND POLYNUCLEAR
AROMATIC HYDROCARBONS

      Oil biodegradation in the respirometric reactors was monitored by oxygen
consumption data, microbial growth determination, gas chromatography scans and
GC/MS analysis whereas the biodegradative activity of the indigenous biomass
in the shaker flask was followed by GC/MS analysis, GC scans and by microbial
growth determination.  Gas chromatographic scans and GC/MS analysis of culture
samples from respirometric and shaker flask systems were made at the start of
experiments, and after 6 and 30 weeks of incubation.  Growth data were based
on plate colony counts (cell numbers on marine and weathered crude oil agar
plates).

      Oil characterization methodology measured individual oil components in
the aliphatic and aromatic fractions of oil extracts.  The aliphatic fractions
were measured by gas chromatography using a flame ionization detector.  The
aromatic fractions were characterized by gas chromatography/mass spectrometry.

      A procedure for  sample extraction and fractionation and a methodology
for analysis of aliphatic and aromatic hydrocarbons, which was used in these
studies, was developed by Nisamaneepong  (1989).  The analytical protocol is a

                                     321

-------
modification of the most current procedures used in the analysis of the above
hydrocarbons in oil.  An established and an EPA acceptable QA/QC protocol for
aliphatic and aromatic hydrocarbons was used in the study.

                                    RESULTS

      Data included in this paper represent results obtained by analysis of
culture samples from respirometric vessels and from shaker flask microcosms as
well as results based on oxygen consumption data for the 6-and 30-week
incubation period.  Biodegradation data reported are those generated from
studies that incorporated the use of Inipol EAR 22 as fertilizer nutrient
source as well as that of the OECD synthetic medium.

      Naturally weathered Prudhoe Bay crude oil collected from the oil
contaminated beaches in the Prince William Sound, Alaska, as well as an
artificially weathered crude oil from Prudhoe Bay were subjected to 6C/FID
analysis for alkane hydrocarbons and GC/MS analysis of the aromatic hydro-
carbons (PAHs) before the initiation of the respirometric and shaker flask
biodegradation studies.  Tables 3 and 4 provide a summary of the concentration
levels in ng/g (ppm) of the C-8 through C-38 aliphatic hydrocarbon (including
pristane and phytane) constituents and polycyclic aromatic hydrocarbon
constituents of the naturally and artificially weathered crude oil samples.


ANALYTICAL RESPIROMETRY

      Respirometric studies have provided data based on cumulative oxygen
uptake data, growth data and GC scans and GC/MS analysis data which demon-
strate almost complete biodegradation of the aliphatic hydrocarbons in the
presence of the Inipol EAR 22 fertilizer nutrient, and very significant
utilization of the PAH hydrocarbons in the culture media after six weeks of
incubation.

Oxygen Consumption Data

      Respirometric results of the respirometry experiments based on the
oxygen uptake data from vessels containing 1,000 mg/L oil and Inipol EAP 22 (5
percent by weight of oil) and from the control respirometric vessel containing
only Inipol nutrient as a source of carbon for 66-day incubation periods are
summarized in Figure 1.  The figure displays cumulative oxygen consumption as
a function of time illustrated by oxygen uptake velocity curves for the two
reactor systems.  Oxygen uptake began on both vessels after only 1.5-days lag
period.  Maximum oxygen uptake on Inipol alone occurred by the 9th day,
leveling off at approximately 150 mg/L.  Oxygen uptake rate on the reactor
containing weathered oil and Inipol was multiphasic:  The first 10 days
exhibited the highest uptake rate followed by a slower rate for the next 16
days and a somewhat faster rate for the next 4 days.  Maximum cumulative
oxygen uptake appears to have occurred by the 30th day, showing a typical
plateau of oxygen consumption beyond the 30 days of incubation.  Data
collection continued well beyond the 66 days shown in the figure and results
will be reported subsequently.  Figure 2 provides oxygen uptake as a function
of time in the inoculum control vessel (beach material + seawater, no oil or

                                     322

-------
Inipol) and in the substrate control vessel (oil, no Inipol, sterilized beach
material and seawater).  In the inoculum control system, the endogenous oxygen
uptake curve plateaus at a maximum of 18 mg/L, whereas maximum oxygen uptake
in the substrate control system is 5-6 mg/L for the 66 days of incubation.
Oxygen uptake curves for the no nutrient control vessel (beach material +
seawater + oil, no Inipol) and for the substrate/nutrient control vessel (oil,
Inipol, sterilized beach material and seawater) are shown in Figure 3.  The
maximum oxygen uptake values for the two control systems were ?A mg/L and 5-6
mg/L respectively at the end of 66 days.

      The vessel containing oil, Inipol and the Alyeska ballast water biomass
exhibited an oxygen uptake curve that was almost superimposabla on the non-
supplemented curve.  Thus, in the closed environment of the respirometrie
vessel, no enhancement by an external source of enriched organisms was
detected.

Microbial Growth Data

      Growth data indicate significant increase of growth of the indigenous
oil degrading microbiota (increase of 4 to 6 orders of magnitude) at the end
of 6 week incubation period.  Comparisons of the bacterial colony counts on
marine agar plates, providing microbial concentration of heterotrophic popu-
lation with those on oil agar plates characterizing the growth of hydrocarbon
degraders at weekly intervals, indicate a significant increase of the hydro-
carbon utilizing bacteria over the first 4 weeks of incubation in the
respirometrie reactors in contrast to a much lower increase of the hetero-
trophic population for the same period.  The total heterotrophic microbial
count leveled off after the first three weeks of incubation with a total
increase of 2-3 orders of magnitude, whereas the hydrocarbon utilizing
indigenous cell count leveled off after 6 weeks of incubation.

Analytical Data

      Figure 4 illustrates a comparison of the GC/FID scans of alkane
hydrocarbons in culture samples from respirometrie vessels containing 1,000
mg/L oil + Inipol, taken at the start, 6 weeks and 30 weeks of incubation,
respectively.  These scans demonstrate almost complete utilization of the
alkane hydrocarbons in these experimental systems at the end of 6 weeks of
incubation.  GC/FID scans of alkane hydrocarbons at the end of 6 and 30 weeks
of incubation in experimental flasks containing Alyeska ballast water sludge
biomass enrichment and in control flasks with Aniline + oil + Inipol also
indicate complete aliphatic hydrocarbon utilization after 6 weeks of
incubation.

      Results based on the GC/MS chromatogram scans demonstrate significant
biodegradation of the PAH constituents of the weathered crude oil in
experimental respirometrie vessels containing 1,000, 300, and 100 mg/L
concentrations of oil and 5 percent concentration of Inipol based on oil
weight in experimental Alyeska biomass supplement vessels and in the control
systems.


                                     323

-------
      Figures 5 through 7 graphically summarize the biodegradation data for
the aliphatic hydrocarbons in 1,000 mg/L oil + Inipol respirometric vessels,
1,000 mg/L oil + Inipol + Alyeska ballast biomass supplement vessels and in
the toxicity control reactor (1,000 mg/L oil + Inipol + aniline) by relating
the initial concentration of these alkanes to those present in cultures after
6 weeks of incubation.  Figure 8 compares biodegradation data for the alkanes
in 100, 300, and 1,000 mg/L oil + Inipol flasks for 0, 6 and 30 weeks of,
incubation.

      Biodegradation data for PAH hydrocarbon constituents of oil in respiro-
metric vessels containing 1,000, 300 and 100 mg/L oil + Inipol after 0, 6 and
40 weeks of incubation are graphically summarized in Figure 9.

      Control experiments using the respirometric approach incorporating the
above three oil concentrations without the Inipol EAR 22 nutrient provided
evidence of insignificant levels of bioactivity of the indigenous seawater and
beach microbiota.

      Percent recovery data of surrogate, o-terphenyl from the experimental
and control respirometric flask samples and of the matrix spike of the
alkanes, n-C15, n-C20 and n-C28 from the naturally weathered crude oil
indicate that the percent recovery falls within the range of established QC
limit range (60-140 percent).  Percent recovery data of the surrogates,
naphthalene, acenaphthene and chrysene from experimental and control flask
samples and of the surrogates, acenaphthene and pyrene from the naturally
weathered crude oil and seawater samples were also shown to be within the
established QC limit range.  The above percent recovery values for the initial
and 6-week samples in the established QC limit range suggest that the
analytical data for the alkane and aromatic hydrocarbons have been generated
with a good QA/QC data analysis.


FLASK MICROCOSMS

      Data from flask microcosms are those based on results obtained by
analyses of culture samples for the alkane hydrocarbon and PAH constituents on
the Alaskan weathered crude oil at the initiation of the experiment and after
six weeks of incubation in shaker flasks.

      Results based on 6C/FID scans and GC/MS analysis data for the alkane
hydrocarbon and PAH constituents of weathered oil in the experimental and
control systems under rotary agitation demonstrate significant biodegradation
of the PAH constituents in culture media and almost complete utilization of
alkane hydrocarbon constituents in presence of 5 percent of Inipol of oil
weight after 6 weeks of incubation.  Control experiments without Inipol
fertilizer provided data which demonstrate insignificant biodegradative
activity of the alkane hydrocarbon and PAH hydrocarbon constituents of oil.

      Figure 10 illustrates a comparison of the GC/FID scans of alkane
hydrocarbons in culture samples from shaker flask microcosms containing 10,000
mg/L oil + Inipol, taken at the start and after 6 weeks of incubation
respectively.  These scans again demonstrate almost complete biodegradation of

                                     324

-------
aliphatic hydrocarbons In these experimental systems at the end of 6 weeks of
Incubation.

      Figure 11 summarizes the results of the GC/FID scans of allkane
hydrocarbons from flasks containing 10,000 mg/L weathered crude with no
nutrients, oil with Inipol, and oil with OECD nutrients.  In the no nutrients
control, some minor changes in the alkane fractions are evident after 6-weeks
incubation.  Some of these changes may have been due to biodegradation
resulting from background levels of N and P present in the seawater, oil or
beach material.  The magnitude of the changes, however, was relatively
insignificant.

      The flask containing the oil plus Inipol exhibited complete removal of
all aliphatic components within six weeks.  Even the pristane and phytane
fractions were reduced to undetectable levels.  The flask containing the
minimal salts solution (soluble form of N and P) also exhibited complete
removal of the straight chain aliphatics.  However, there were still
measurable amounts of pristane and phytane remaining at six weeks, although
the levels were significantly reduced from the no-nutrient controls.  These
results suggest that Inipol may be enriching for a different type of microbial
population from that which is selected by the OECD medium.  The Inipol-
enriched organisms are not only able to break down straight chain hydrocarbons
at a very rapid rate but the branched chain components as well.  The QECD-
enriched organisms are also able to degrade the branched chain aliphatics, but
at a reduced rate or longer lag period.

      The GC/MS traces of the aromatic fractions are summarized in Figure 12.
In the no-nutrient control (top of figure), several of the components were
reduced to undetectable levels after six weeks (note fractions, H, Q, S, and
T, corresponding respectively to dibenzothiophene, C3-fluorenes, naphthalene,
and Cl-naphthalene).   The traces from the Inipol and OECD flasks exhibit
virtual complete removal of all aromatic fractions after six weeks incubation.

      The percent recovery data of the surrogate, o-terphenyl, from the
experimental and control shaker flask culture samples and of the surrogate
matrix spikes of n-C15, n-C20 and n-C28 alkane hydrocarbons from same flask at
zero incubation time and after 6 weeks of incubation were observed to be
within the QC limit range for each of the surrogates, thus lending credence to
thi quality of analytical data generated for alkane hydrocarbons.  The percent
recovery of the surrogate PAH compounds from culture media was again observed
to be within the QC limit ranges for each of these surrogates, demonstrating
good QA/QC quality of analytical data for PAH constituents of oil extracted
from culture media samples.

                                  CONCLUSIONS

      Respirometric and shaker flask microcosm biodegradation data demonstrate
a very significant enhancement on biodegradation of alkane and aromatic (PAH)
hydrocarbon constituents of oil contaminating the Alaskan beaches of Prince
William Sound with the use of the oleophilic fertilizer nutrient, Inipol EAP
22 and the OECD synthetic medium (soluble N and P nutrients).

                                     325

-------
      Rapid and virtually complete biodegradation of all aliphatic and
aromatic hydrocarbon components of the weathered crude oil has unequivocally
been demonstrated in the closed laboratory environment of the respirometric
vessels and flask microcosms.

      Almost complete disappearance of these hydrocarbon components occurred
within six weeks with some traces of PAH hydrocarbons remaining in the culture
systems.

      GC/FID chromatographic scans and GC/MS analytical data for the alkanes
and PAH hydrocarbons in culture system samples taken after six weeks of
Incubation indicate complete losses due to biodegradation of these components
of the weathered crude oil.

      The oxygen uptake in respirometric vessels started after only a 1.5-day
lag period, indicating a very active hydrocarbon-utilizing potential of the
indigenous seawater and beach sand microbiota of Prince William Sound.

      The different microbial populations enriched by the two types of
fertilizers suggest that perhaps a combination of Inipol EAP22 and a water
soluble source of nutrients may ultimately be the appropriate manner of
stimulating rapid bioremediation of crude oil contaminating the Alaskan
beaches.

      Results from the Alyeska ballast biomass environments suggests that
enhancement with external sources of microbial populations would not likely
occur and massive inoculum may not be warranted, at least in the Alaskan
bioremediation effort.

      GC/FID and GC/MS chromatographic scans at 1-week intervals would have
provided valuable insight on the rates of disappearance of the individual
alkane and aromatic hydrocarbon components of oil.  Future experimentation
should incorporate more frequent sampling for each analysis.

      No inhibitory effects of crude oil at the concentrations of 100, 300,
1,000, 3,000, and 10,000 mg/L in culture systems were observed on both the
indigenous biomass and on their biodegradation activities on readily
biodegradable or biogenic substrate.

      Analytical respirometry was shown to be a valuable experimental approach
for testing biodegradability of the hydrocarbon constituents of crude oil
contaminating the marine environment and for testing the hydrocarbon
biodegradation potential of the indigenous seawater and beach sand microbial
consortium in the oil contaminated areas.
                               ACKNOWLEDGEMENTS

      The authors wish to thank Mrs. Rena M. Howard and Mrs. Betty A. Kampsen,
secretaries in the U.S. Environmental Protection Agency's Risk Reduction
Engineering Laboratory, Cincinnati, Ohio, for their excellent and timely
wordprocessing skills in preparing this manuscript.

                                     326

-------
                                  REFERENCES

1.    Atlas, R.M.  1977.  Stimulated petroleum biodegradation.  Crit. Rev.
      Microbiol.  5: 371-386.

2.    Atlas, R.M. 1981.  Microbial degradation of petroleum hydrocarbons:
      an environmental perspective.  Microbiol. Rev., 45: 180-209.

3.    Atlas, R.M.  1982.  Assessment of potential interactions of micro-
      organisms and pollutants resulting from petroleum development on the
      outer continental shelf of Alaska.  Final report Contract #03-5-022-85.
      RD No. 28, p. 187.

4.    Atlas, R.M., and R. Bartha.  1972.  Degradation and mineralization of
      petroleum in seawater;  limitation by nitrogen and phosphorus.
      Biotechnol.  Bioeng.  14: 309-317.

5.    Atlas, R.M., and R. Bartha.  1973a.  Fate and effects of oil pollution
      in the marine environment.  Residue Rev.  49: 49-85.

6.    Atlas, R.M., and R. Bartha.  1973b.  Effects of some commercial oil
      herders, dispersants and bacterial inocula on biodegradation of oil in
      seawater, p. 283-289.  In D.G. Ahearn and S.P. Meyers ed., The microbial
      degradation of oil pollutants.  Publication no. LSU-SG-73-01.  Center
      for Wetland Resources, Louisiana State University, Baton Rouge.

7.    Atlas, R.M., and R. Bartha.  1973c.  Stimulated biodegradation of oil
      slicks using oleophilic fertilizers.  Environ. Sci. Techno!.  7: 538-
      541.

8.    Atlas, R.M., and E.A. Schofield.  1975.  Petroleum biodegradation in the
      Arctic, p. 183-198.  In A.W. Bourquin, D.G. Ahearn, and S.P. Meyers ed.,
      Impact of the use of microorganisms on the aquatic environment.  U.S.
      Environmental Protection Agency, Corvallis, Ore.

9.    Bartha, R., and R.M. Atlas.  1973  Biodegradation of oil in seawater:
      limiting factors and artificial stimulation, pp. 147-152,  In D.G.
      Ahearn and S.P. Meyers ed. The microbial degradation of oil pollutants.
      Publication No. LSU-SG-73-01.  Center for Wetland Resources, Louisiana
      State University, Baton Rouge.

10.   Bartha, R., and R.M. Atlas.  1977.  The microbiology of aquatic oil
      spills.  Adv.. Appl. Microbiol.  22: 225-256.

11.   Bergstein, P.E., and J.R. Vestal.  1978.  Crude oil biodegradation in
      Arctic tundra ponds.  Arctic 31: 158-169.

12.   Bridie, A.L., and J. Bos.  1971.  Biological degradation of mineral oil
      in seawater.  0. Inst. Pet. London.  57: 270-277.

13.   Bronchart, R.E.E., J. Cadron, A. Charier, A.A.R. Gillot, W. Veretraete,
      1985.  A new approach in enhanced biodegradation of spilled oil:

                                     321

-------
      enhancement of an oil dispersant containing oleophilic nutrients.  Oil
      Spill Conference, February 25-28, Los Angeles, CA.  pp. 413-462.

14.   Colwell, R.R., and J.D. Walker.  1977.  Ecological aspects of microbial
      degradation of petroleum in the marine environment.  Crit.  Rev.
      Hicrobiol.  5: 423-445.

15,   Colwell, R.R., A.L. Mills, and J.D. Walker.  1978.  Microbial Ecology
      studies of the Hetula spill in the Straits of Magellan.  J. Fish. Res.
      BD. Can.  35: 573-580.

16.   Cowell, E.B. ed., 1971.  The ecological effects of oil pollution on
      littoral communities.  Applied Science Publishers, Ltd., London.

17.   Crow, S.A., S.P. Meyers, and D.G. Ahearn.  1974.  Microbiological
      aspects of  petroleum degradation in the aquatic environment.  Mer
      12: 37-54.

18.   Delaune, R.D., 6.A. Hambrick III, and W.H. Patrick, Jr.  1980.
      Degradation of hydrocarbons in oxidized and reduced sediments.  Mar.
      Pollut. Bull. 11: 103-106.

19.   Floodgate, 6.D.  1972a.  Biodegradation of hydrocarbons in the sea, pp.
      153-171.  In R. Mitchell ed., Water pollution microbiology.  John Wiley
      and Sons, Inc., New York.

20.   Floodgate, 6.D.  1972b.  Microbial degradation of oil.  Mar. Pollut.
      Bull. 3: 41-43.

21.   Floodgate, G.D.  1973.  A threnody concerning the biodegradation of oil
      1n natural waters, pp. 17-24. In D. G. Ahearne and S. P. Meyers ed.,
      The microbial degradation of oil pollutants.  Publication No. LSU-SG-
      73-01.  Center for Wetland Resources, Louisiana State University, Baton
      Rouge.

22.   Floodgate, G.D.  1976.  Oil biodegradation in the oceans, pp. 87-92.  In
      J. M. Sharpley and A. M. Kaplan ed., Proceedings of the Third
      International Biodegradation Symposium.  Applied Science Publishers,
      Ltd., London.

23.   Floodgate, G.D.  1979.  Nutrient limitation, p. 107-118.  In A. W.
      Bourquin and P. H. Pritchard ed., Proceedings of Workshop, Microbial
      Degradation of Pollutants in Marine Environments.  EPA-66019-79-012.
      Environmental Research Laboratory, Gulf Breeze, FL.

24.   Gatelller, C.R.  1971.  Les facteurs limitant la biode'gradation des
      hydrocarbons dans 1'epuration des eaux.  Chim. Ind. Paris 104: 2283-
      2289.

25.   GatelHer, C.R., J.L. Oudln, P. Fusey, O.C. Lacase, and M.L. Priou.
      1973. Experimental ecosystems to measure fate of oil spills dispersed by

                                     328

-------
      surface active products, pp. 497-507.  In Proceedings of Joint
      Conference on Prevention and Control Oil Spills.  American Petroleum
      Institute, Washington, DC.

26.   Gibbs, C.F.  1975.  Quantitative studies in marine biodegradation of
      oil.  I.  Nutrient limitation at 14°C.  Proc. R. Soc. London Ser. B
      188: 61-82.

27.   Gibbs, C.F., and S.J. Davis.  1976.  The rate of microbial degradation
      of oil in a beach gravel column.  Microb. Ecol. 3: 55-64.

28.   Gunkel, W.  1967.  Experimentell-okologische Untersuchungen uber die
      limiterenden Faktoren des microbiellen Olabbaues in marinen Milieu.
      Helgo. Wiss. Merresunters.  15: 210-224.

29.   Halmo, G.  1985.  Enhanced biodegradation of oil.  1985 Oil Spill
      Conference, Los Angeles, California, February 25-28. 531-537.

30,   Jordan, R.E., and J.R. Payne.  1980.  Fate and weathering of petroleum
      spills in the marine environment.  Ann Arbor Science Publications, Inc.,
      Arbor, MI.

31.   Karrick, N.  1977.  Alterations in petroleum resulting from
      physicochemical and microbiological factors, pp. 225-299.  In D.C.
      Mai ins ed., Effects of petroleum on arctic and subarctic marine
      environments and organisms, Vol. 1.  Nature and fate of petroleum.
      Academic Press, Inc., New York.

32.   Kator, H., R. Miget, and C. H. Oppenheimer.  1972.  Utilization of
      paraffin hydrocarbons in crude oil by mixed cultures of marine bacteria.
      Paper No. SPE 4206.  Symposium on Environmental Conservation.  Society
      of Petroleum Engineers, Dallas, TX.

33.   Kinney, P.J., O.K. Button, and D.M. Schell.  1969.  Kinetics of
      dissipation and biodegradation of crude oil in Alaska's Cook Inlet, pp.
      333-340.  In Proceedings of 1969 Joint Conference on Prevention and
      Control of Oil Spills.  American Petroleum Institute, Washington, DC.

34.   Lee, K. and Levy, E.M.  1987.  Enhanced biodegradation of light crude
      oil in sandy beaches.  1987 Oil Conference, April 6-9, Baltimore, MD.
      pp. 411-416.

35.   Lee, K. and Levy, E.M.  1989.  Enhancement of natural biodegradation of
      condensate and crude oil on beaches of Atlantic Canada.  Oil Spill
      Conference, February 13-16, San Antonio, TX.  pp. 479-486.

36.   LePetit, J., and M.H. Barthelemy.  1968.  Les hydrocarbures en mer:  le
      probleme de 1'epuration des zones littorales par les microorganismes.
      Ann. Inst. Pasteur Paris.  114: 149-158.
                                     329

-------
37.   LePetit, 0., and M.H, N'Guyen.  1976.  Besoins en phosphore des
      bacteries metabolisant les hydrocarbons en mer.  Can. J. Microbiol.
      22: 1364-1373.

38.   Hulkins-Phillips, G.J., and J.E. Stewart.  1974a.  Effect of four
      dispersants on biodegradation and growth of bacteria on crude oil.
      Appl. Microbiol.  28: 547-552.

39.   Hulkins-Phillips, 6.J., and O.E. Stewart.  1975b.  Distribution of
      hydrocarbon-utilizing bacteria in northwestern Atlantic waters and
      coastal sediments.  Can. J. Microbiol.  20: 955-962.

40.   National Academy of Sciences.  1975.  Petroleum in the marine
      environment.  National Academy of Sciences, Washington, DC.

41.   Nisamaneepong, W.  1989.  Summary of Method for Saturated Hydrocarbon
      (HC) and Polynuclear Aromatic (PAH) Hydrocarbons Analysis for Seawater
      Samples, p. 5.

42.   Olivieri, FL, P. Bacchin, A. Robertiello, N. Oddo, L. Degen, and A.
      Tonolo.  1976.  Microbial degradation of oil spills enhanced by a slow-
      release fertilizer.  Appl. Environ. Microbiol. 31: 629-634.

43.   OECD, "OECD Guidelines for Testing of Chemicals," Section 3, Degradation
      and Accumulation, Method 301C, Ready Biodegradability: Modified MITI
      Test (I) adopted May 12, 1981 and Method 302C Inherent Biodegradability:
      Modified MITI Test (II), adopted May 12, 1981, Director of Information,
      OECD, Paris, France, 1981.

44.   Reisfeld, A., E. Rosenberg, and D. Gutnick.  1972.  Microbial
      degradation of oil:  Factors affecting oil dispersion in seawater by
      mixed and pure cultures.  Appl. Microbiol.  24: 363-368.

45.   Robichaux, T.O., and H.N. Myrick.  1972.  Chemical enhancement of the
      biodegradation of crude oil pollutants.  J. Petrol. Techno!. 24: 16-20.

46.   Sirvins, A. and Angles, M.  1986.  Development and effects on marine
      environment of a nutrient formula to control pollution by petroleum
      hydrocarbons.  In Strategies and Advanced Techniques for Marine
      Pollution Studies:  Mediterranean Sea, C.S. Siam and H.J.M. Dou eds.,
      Springer, Verlag, Berlin, pp. 357-404.

47.   Smith, J.E.  1968.  "Torrey Canyon" pollution and marine life.
      Cambridge University Press, England

48.   Sveum, M.  1988.  Marine gas oil in arctic shoreline sediments.  SINTEF,
      Division of Applied Chemistry, Trondheim, Norway.  Final Report, NTIS
      ISBN. No. 32-595-4410-5.

49.   Sveum, P., and Ladousse, A.  1989.  Biodegradation of oil  in the Arctic:
      Enhancement by oil soluble fertilizer reduction.  1989  Oil Spill
      Conference, February 13-16, 1989, San Antonio, TX.

                                     330

-------
50.   Tabak, H.H., Lewis, R.F., and Oshima, A.  1984.  Electrolytic
      respirometry biodegradation studies, CEC/OECD ring test of respiration
      method of determination of biodegradability, Ring Test Program 1984.
      EPA Draft Final Report, HERL, U.S. Environmental Protection Agency,
      Cincinnati, Ohio, August 1984.

51.   Taggar, S., Bianchi, A., Julliard, M., LePetit, J., Roux, B.  1983.
      Effects of microbiai seeding of crude oil in seawater in a model system.
      Marine Biology, 78: 13-20.

52.   Tramier, B. and Sirvins, A.  1983.  Enhanced oil biodegradation: a new
      operational tool to control oil spills.  1983 Oil Spill Conference,
      February 28 - March 3, San Antonio, TX.  pp. 115-119.

53.   Van der Linden, A.C.  1978.  Degradation of oil in the marine
      environment, p. 165-200.  In J.R. Watkins ed.. Developments in
      biodegradation of hydrocarbons-1.  Applied Science Publishers, Ltd.

54.   Ward, D.M., and T.D. Brock.  1976.  Environmental factors influencing
      the rate of hydrocarbon oxidation in temperate lakes.  Appl. Environ.
      Microbiol.  31: 764-772.

55.   ZoBell, C.E. 1946.  Action of microorganisms on hydrocarbons.
      Bacteriol. Rev.   10: 1-49.

56.   ZoBell, C.E.  1964.  The occurrence, effects and fate of oil polluting
      the sea.  Adv. Water Pollut. Res. 3: 85-118.

57.   ZoBell, C.E.  1969.  Microbiai modification of crude oil in the sea, pp.
      317-326.  In Proceedings of Joint Conference on Prevention and Control
      of Oil Spills.  American Petroleum Institute, Washington, DC.

58.   ZoBell, C.E.  1973.  Microbiai degradation of oil:  present status,
      problems and perspectives, pp. 3-16.  In D. G. Ahearn and S. P. Meyers
      ed., The microbiai degradation of oil pollutants.  Publication No. LSU-
      SG-73-01.  Center for Wetland Resources, Louisiana State University,
      Baton Rouge.
                                     331

-------
TAKE 1.  OrCXIMOCTJU. OtSIW FOt RISF1ROMTR1C STUDIES
                                                                                  TULE 2.  CXHRINENTAL OESISN OF FLASK MICROCOSMS

Itittlefl 011 CMIC.
ttlllll (H/L
»i, m 1,000
n, «i 300
n, m ioo
t4, f4R 1,000
Cl
U
U UOOO
Cl
Cl 1,000
U 1,000
C7 1.000
T - »«ml
i * mum*
C • Cwtral
Abtlki Mich
, Inljwl Cone, lillut Sirvj
•J/L (*! CS/l!
SO • 2
IS • 2
S 2
SO 10 2
SO - 2
2
SO 2
SO 2

2
SO




StMlUl*
(•L)
1,000
1,000
1,000
1,000
1.000
1,000
1,000
1,000
1,000
1,000
"




An Hint
(•B/U
,
-
-

.
-
ICO
100
.
-
"




Oil Cane. Ii
Fink (BjA)
F1.F1R 10,000
F2.F2R 10,000
F3.F3R 10,000
F4.F4R 10.000
FS.FSR 3.000
FI,F« 1,000
F7.F7R oily stixl
FS.fM •

C2
C3 10,000
C4
CS
Cl
C7
a
Cl 10,000
CIO 10,000
Cll 0»1j Ml*
Cit otly lud
F • Fluk
t • RtplUit*
C • Ctntrel
•V m UMMllllttd
**» - PtTliit** taKA
«,.«. 8.,c» SM.
(•g/l) OEO) («.} (9) {«U
SOO 10 U* 1,000
500
* • 20 U 1.000
-- 4.
ISO 20 U
so
2.S
s.o
7,$
SOO
500
. 4.
SOO

.
. .
« .
,
7.1
20 U
5 P**
10 t
IS P
20 U
20 U
20 U
20 U
20 U
20 U
20 U
20 U
20 U
20 U
15 t
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
,000
IS P ,000



tmt

(«VU










100
100
100
109
.

.
-
-
'




TAKE 1.
           XLIBUTIC COKSTITVCXTS Of AUSXM F*UOHO£
           Ut CIUDC OIL IT K/FID AKAUSIS
                                                                               TABLE 4.
Conctntrition, 119/9 (ppa)
Alkuit HjilrMlrbciK
C 1
c t
C 10
C U
C 12
C 13
C 14
C IS
C If
C 1?
nlSTMl
C 11
HffMK













tl
10
21
U
2J
24
2S
2f

M
»
M
M
C M
NitnriUjr v»ch«r«l Artlflcltlijr Vtithtrm
Alukin CrwH Oil AUlkjn Cndl Oil
31.4
W.S
&3.»
1M
2S4
370
MS
112
144
675
629
732
(0*
117
S»
331
340
22t
252
130
112
111
H.I
03.1
<2.5
<2.S
<2.S
d
68. 0
411
•»
1150
1180
1410
584
1010
72f
SSI
SH
434
7M
127
S74
471
590
711
4M
122
371
221
141
<2.S
<2.3
<2.S
AROMATIC CONSTITUENTS OF ALASKAN PRUDHOE
BAY CRUDE OIL 8T CC/HS ANALYSIS
                                                                                                       dofict nint ion,
                                                                            r*n»ttUc
                                                                                               ~ Mtsirilljt ttllhtrid
                                                                                                   AUikin 011
                                                                  ActnipttlMM   ~~^^~~
                                                                  AemipHthyli»
                                                                  8«njo(g,h,1)P«ryl«n«
                                                                  Chr>Mnt/Btnzo(<)Ajithrtcini
                                                                  Cl-Chrynnti
                                                                  C*-C!irjrjt(i»f
                                                                  CJ-OtbtnzothtophMtt
                                                                  U-Q1btntetli1oplMn»
                                                                  Fluorinthint
                                                                  Cl-FluortntiiiMi/pjFrtnti
                                                                  Fluomit
                                                                  Cl-FluartMf
                                                                  C2-F1usrtMf
                                                                  C3»F1inrtfMi
                                                                  MMMthrm/Mtkricn*
                   8.76
                  < 5
                  < S
                  < S
                   7.43
                   7.2G
                 < 0.9
                 < 8.1
                 < O.S
                 < 10
                  21.0
                  <0.9
                  5«.S
                  32.8
                   4.S*
                 < O.I
                  18.8
                  n.o
                  «a.i
                  S4.0
                 < $
                  11,1
                  62.)
                   124
                  6».S
                  31.9
                  20.1
                  31.0
                  36.S
                  21.0
                   J.7I
                   2,31
                                     iITyiiitHirja
                                 AUikin Oil
	13.1	
 43.6
  S.26
 < S
 < 5
< o.s
  2.
-------
         Oxygen Uptake  on Weathered Crude Oil


        	 on. *
                                                                Oxygen Uptake on Weathered Crude OH

                                                       mocw.UMcOL)     — frnmcujGmnxm.fammsnoii.uatm
    teoo
E
X
O
I20O

1OOO

 BOO

 6OO

 4OO

 200

  O
             1O    2O    30   4O    5O    t»    7O

                      Time, Days
""s*
o»
E

J£
m
+H*
Q.
3

C
•
at
>»
X

ItlU
MO
12O
100

BO

6O

4O

2O

O

-
"
,

1.

-

-

L. -
^^^^^ 	 	 	 -_,-,.
' 	 Ill 	 Ti 	 THiimiiiiUtininifiMniiin 	 il
                                                                    1O    2O    30    4O    SO    6O

                                                                             Time,  Days
         Fteunc 1.
                         MTWM 9T1KU in tISMMNItllK
                     MITM 1.000 Mt/t Or MMNtm CM)N
                 ML HUB iHlnL Wit.
PIWMM t.  CMwunm m*MM •rr«i in
        MMfMfl Cwmrflt MWIMMTMC
                                                                                                  urn
                                        Oxygen  Uptake on  Weathered  Crude OH

                                                            ----- ('§OTJ|lg|«^**
Qt

JC
S
^"
^

C
«
a
*^%
X
O
•DM
MO
120
100

80
SO

4O

Pn
2O 1
0«
	 	 	 ,
1

1
J

I 1
1
J
1

/ 	 ' 	 ~ 	 ^^-^^__^ 1

                                                 20
                                                      3O
                                                            4O
                                                                  SO
                                                                        60
                                                                              70
                                       Fs
                                            $.
                                                    Thne,  Days
                                                      MVMM HTTMtt M HO wrMWT «W
                                                           CMfML *MPtM>Hmfe

-------
r
   Nil - Cra* »lt 1.000 •!/! » l«lp»l Ufa  ' mWII - CnMn •!! 1.000 mifl. t l»lp>l
      - lulllil 0 wtk r»i»Irw«trlc m»«rl«nit
                                         - St«
                                                           ««ftrl««it
                                                        MtirMll - Cntt ill I.OM i
                                                              - nirtjr Mttf rasi
                                                                               nip»i turn
                         FZWIM 4.  COMPARISON OF  GC/FID SCAMS or ALKANE
                                     HYDROCAHBONS IN CULTURE SAHPLCS FKOH
                                     RtSPXRONETRZC  FLAIKJ CONTAINING
                                     1,000 MC/L OIL + INIPOL EW22 AFTER 0,
                                     6 AND 30 WEEKS OF INCUBATION.
ill - Craft) *11 It^M milt. » Inlpri Of U
   - until • «wt Ottar *
                                                     •r
                                                        ilMll . Cm* *1I 10,*00 ^A • Intel IM> S
                                                              - Six MtU ttaUr 
-------
             BKWCQRADATION OF AUPHATIC HYDROCARBONS
             IN  1,000 man. ALASKAN WEATHERED  CRUDE Ott.
            I 0  WEEKS            ^H « WEEKS
                                           BtOOEQRADATION OF  1000 BigO.  CRUDE OH. M
                                              PRESENCE OF BALLAST WATER BIOMASS
                                           0  WEEKS
                                                                      « WEEKS
I
   125O
   1OOO
    75O
    500
    250
       o1-









Jill
.•Hlifcfcfcfc









1
• huh










•I


















II LL. .__
l^DU
tOQO
41
a
a
<* 75O
o
3
• 500
c
O
o
25O









.1,1
















































































1









•la..

                             Alkanes
         FlEUM S,  »10Dt&«»0*T10» Of »11PB»I1C BrOBqC«"10
                  1.000 M&/L Ai»s«*« uiAininio CBUDI on
                                                            Alkanea
                                         FIGUIIC i.  lioefCMOATiDH or MIFHUIIC mronocMiMMi i«
                                                   1,000 HC/l Al«SK»>< WIHTHIHID CRUOI OIL IN
                                              BtODEORAMTION OF WEATHERED CRUDE  OIL M
                                                       THE PRESENCE  OF AMUNE
                                             I START               HB 6 WEEKS
                                    1250
                                    1OOO
                                 a
                                 S
                                 I   75°
                                 I
                                 I   500
                                     25O
Jill
1
                                          CBUI TO II 12 13 14 K 16 t?«* IBPH 1920Z1 2! ?32< 25 ZB» » H .M 3k »
                                          FlGMI 1.  liMfMMMrlMI Or «LirMTIC MnMCMMM in
                                                   l.ttt NC/l AUIIIUUI M«?NE«I* HWI Oil in
                                                   • IfMMMfTllIC rt««» COHTAIIIIIK AHILIHI.

-------
       RESPIROMETRIG STUDIES WITH  INIPOL
         O  WEEKS
                         6 WEEKS
                                          30  WEEKS
i
-------
                 RESPIROMETRIC STUDIES  WITH INIPOL
                 O WEEKS
                                           6 WEEKS
     O

     ed
     »—,
     s=S

-------
DEGRADATION  OF  ALIPHATICS  -- SHAKER  FLASK STUDIES
             O WEEKS
                                     6  WEEKS
leuuu
14OOO
g: IZOOO
ea 1OOOO
~ca 8OOO
"g eooo
g 4OOO
CJ
zooo

14OOO
g; izooo
~£ 8OOO
g eooo
CJ
c§ 4000
zooo
o
14OOO
g: isooo
a 10000
.2
~rt 80OO
g 6000
CJ
o 40OO
2000
n
-
-
-
-


_
-
-
_ 1
<
_
-

-
-
-




I
:•



q
:•







!: n
H ' 'I

kill IJi ;
IO 11 IS 13 1« IO !• 17 PI
n F
'/ .;
i \ $
J> i ^ t ^
i M i
Mis
M M i i
IO 11 1* 13 14 10 !• 17 PI

1 ill
^ _ > j e
| i j | M
a 1 1 1 i h
: H i 1 * «
jjjjjjjj
NC
1


; i: : r Fl
n 111
1 18 PH IO >O 31 33 33 3« 3O
Alkanes
:
* P
'i >
IN = 1 U
1 !• PH 19 30 31 33 33 2« *B
Alkanes

r
!J

j .
ILL ill
NUTRIENT
CONTROL



lid do. _ _
t» 3O 30 33 34 3O 38

INIPOL

LrLn-fL
t« 38 3O 33 3« 3« 38

MINIMAL
SALTS



IFL___
          CO IO II I* 13 14 IO l» 17 P« IS PR 19 SO XI 3> S3 34 25 SO X9 3O 38 34 3« 38
                            Alkanes
    FIGURE 11.
COMPARISON OF IZODEGRADATION DATA FOR
ALKANE HYDROCARBONS IN SHAKER FLASKS
WITH 10,000 MG/L OIL + INIPOL,  WITH
10,000 MG/L OIL + MINIMAL SALTS AND
IN FLASK WITH NO NUTRIENT.

         338

-------
     DEGRADATION  OF  AROMATICS  —  SHAKER  FLASK STUDIES
                      O WEEKS
                                                    9  WEEKS
      i
18OO

1000

1-400

12OO

1OOO

 8OO

 aoo

 *00

 800

   O
                    NO  NUTRIENT
                       CONTROL
                                        Ar orr»«t.lo»
      a
180O

1«00

1-4OO

1800

1OOO

 eoo

 eoo
            800

              O
                       INIFOL
     .1
      a
1800

iaoo
1*00

1*00

1000
 800
 aoo

 •400

 «oo

   O
                       MINIMAL
                         SALTS
A acenaphthylene
B benzo(a)pyr«ne
C benzo(b)f1uort«nthene
0 chrysene/benzo(a)anthracene
I Cl-chrysenes
F dlbenzothlophene
              6 Cl-d1benzoth1ophenes
              H C2-d1benzoth1ophenes
              I C3-d1benzoth1ophenes
              J fluoranthent
              K fluorene
              L Cl-fluorenes
M C2-fluorenes
N C3-fluorenes
0 naphthalene
P Cl-naphthalenes
Q C3-naphthalenes
S C4-naphthalenes
S phenanthrene/anthracene
T Cl-phenanthrenes/anthracenes
U C2-phenanthrenes/anthracenes
V C3-phenanthrenes/anthracenes
W C4-phenanthrenes/anthracenes
X pyrene
      FIGURE 12.
           COMPARISON OP tiootonAOATiON  DATA FOR
           AROMATIC  HYOROCARiONS  XN SHAKER PLAHKS
           WITH 10,000  MG/L OIL + INZPOL,  WITH
           10,000 MG/L  OIL  •*• MINIMAL  SALTS AND  IN
           FLASK  WITH NO NUTRIENT.
                                      339

-------
           BIOLOGICAL TREATMENT OF WASTEWATERCONTAINING
           HAZARDOUS ORGANIC COMPOUNDS:  2-ETHOXYETHANOL
                       Margaret J. Kupferle*
                      University  of  Cincinnati
                      Cincinnati, Ohio  45221

                        Michael  P.  Vitello
                     Foppe Theilen Group,  Inc.
                      Cincinnati, Ohio  45246

                           Lisa M, Brown
                U.S.  Environmental Protection Agency
               Risk Reduction Engineering Laboratory
                      Cincinnati, Ohio  45268


                              ABSTRACT

    Biological treatment is one potential technology for treatment
of RCRA-listed wastes containing hazardous organic compounds in
concentrations up to 1% (10,000 mg/L), provided that the compounds
are not recalcitrant or toxic to the microorganisms in the system.
As part of the USEPA's in-house testing program at the Test &
Evaluation Facility, biological treatment was evaluated as a
treatment technology for 2-ethoxyethanol, a solvent that can occur
as the major constituent in RCRA-listed F005 wastes.

    Respirometry was used to screen 2-ethoxyethanol for toxicity
under aerobic conditions.  Toxicity did not occur in the
concentration range studied  (0 mg/L - 10,000 mg/L), and the data
implied that biodegradation occurred.

    Pilot-scale activated sludge systems were subsequently used to
further evaluate the fate of 2-ethoxyethanol in aerobic systems.
The pilot-scale systems were operated at a flow rate of
approximately 11 liters per day for a 24-hour hydraulic residence
time with a target solids retention time of 10 days.  The systems
were supplied with a synthetic feed which was spiked with 2-
ethoxyethanol, and the concentration of 2-ethoxyethanol was
measured in feed, effluent, waste sludge and reactor off-gas
samples.  At the highest feed concentration studied, the system
reduced the 2-ethoxyethanol concentration from 2284 ± 201 mg/L in
the feed to 18 ± 17 mg/L in the effluent.  No 2-ethoxyethanol was
detected in the off-gas or waste mixed liquor samples.


                                340

-------
    This paper has been reviewed in accordance with the U.S.
Environmental Protection Agency's peer and administrative review
policies and approved for presentation and publication.
                            INTRODUCTION

    The assessment of biological treatment as a viable treatment
technology for 2-ethoxyethanol FOQ5 waste was carried out by
University of Cincinnati  (UC) project staff at the USEPA Test &
Evaluation (T&E) Facility during FY88-89 under EPA Contract No.
68-03-4038.  This research was part of Work Assignment NO.2-WH21.0
entitled "Hazardous Waste Treatment Technology Assessment".  The
work assignment covered in-house bench and pilot scale testing as
part of a larger treatment technology assessment program funded by
the Office of Solid Waste (OSW) through the Hazardous Waste
Engineering Research Laboratory (HWERL) and, later, the Risk
Reduction Engineering Laboratory (RREL).

    Preliminary bench-scale testing was done to screen 2--
ethoxyethanol for treatability/biodegradability.  In these
studies, toxicity due to 2-ethoxyethanol was tested in both
aerobic and anaerobic systems at concentrations up to 1% as 2-
ethoxyethanol,  Oxygen uptake rate was used as the measurement
parameter for the aerobic studies.  Gas production rate was used
as the measurement parameter for the anaerobic studies.  Toxicity
effects were not noted in the tested concentration range and
biodegradation was implied in both cases.  The aerobic studies are
the topic of this paper. The results of the anaerobic studies have
been discussed elsewhere(2).

    Before biodegradation could be confirmed, a reliable analytical
method was required for 2-ethoxyethanol,  Poor purging
efficiencies required the use of direct injection for introduction
of the sample into the gas chromatograph.  SW-846 Method 8015
(GC/FID method for nonhalogenated volatile organics - see  (1)) was
modified to use a column capable of separating trace solvents in
aqueous solutions {Supelco Carbopak B/3% SP1500, 8,0/120 mesh
packing or equivalent).

    While the analytical method for 2-ethoxyethanol in aqueous
solutions was being identified, start-up of two existing activated
sludge pilot plants was initiated,  After method QA/validation
studies were performed, monitoring of the feed and effluent to the
two pilot plants was begun.  Data from the pilot plant studies are
presented and discussed following a brief discussion of the
aerobic bench-scale respirometry studies,
                                341

-------
                        RESPIROMETRY STUDIES

    A Tech-Line respirometer was used to study the toxicity effects
of 2-ethoxyethanol concentration on the oxygen uptake rate of an
activated sludge culture. A schematic diagram of the apparatus is
shown in Figure 1.  The respirometer includes two 4—liter aeration
chambers in a single water jacket tank.  The seed source was
unacclimated activated sludge from a pilot plant at the T&E
Facility which treats screened raw wastewater from the Cincinnati
Metropolitan Sewer District.  Neat 2-ethoxyethanol was obtained as
purified grade liquid  (Cellosolve) from Fisher Scientific Company.
Oxygen uptake rate was measured in milliliters of oxygen per liter
of activated sludge per hour  (mL/L/hr).
                    VENT
                    OR
                    FILL
                                   RECORDER
                                           C02
                                           SCRUBBER
                                AIR PUW
           Figure 1.   Respirometry Experimental Apparatus
    Two types of tests were performed.  One type was the Cumulative
Toxicity Test recommended by the manufacturer of the respiro-
meter (3) .  In the first part of this test, a readily degradable
carbon source was supplied to both the test and blank cells and
2-ethoxyethanol was spiked into the test cell cumulatively.  In
the second part, the maximum amount of test compound was added
immediately after the background uptake rate had been measured.
In analyzing the data, a drop in oxygen uptake rate in the test
cell as compared to the rate in the blank cell indicates toxicity.
An increase implies biodegradation of the test compound.  The
results for the 2-ethoxyethanol test are shown in Table 1. Once
2-ethoxyethanol was spiked into the test cell, the oxygen uptake
rate in the test cell was always greater than in the blank cell.
The rate of uptake declines with time, however, in both cells in
the first part of the experiment as the readily degradable carbon
source and nutrient supplements are consumed.  There appears to
be a shift in the relative difference between the uptake rates
(4.1 to 1.8 mL/L/hr) after a total of 2320 mg/L was added to the

                                342

-------
test cell.  However, this phenomenon is more likely clue to some
factor other than the test compound concentration since in the
second part of the experiment, the rates again showed a 4,0
mL/L/hr increase for the maximum 2-ethoxyethanol concentration of
9300 mg/L.  No toxicity was observed.


     TABLE  1.  RESPIROMETRIC  CUMULATIVE TOXICITY TEST RESULTS
                     FOR 2-ETHOXYETHANOL (2-E)
2-E Added
(mg/L)
0
580
580
1160
2320
4650
9300
0
9300
Cumulative 2-E
(mg/L)
0
580
1160
2320
4640
9290
18590
0
9300
30 min Ne
Test
14.5
16.5
12.2*
10.2
7.3
6.9
6.2
13.2
12.1
t Oxygen I
(mL/L/hr
Blank
14.7
9.7
8.1
6.1
5.5
5.4
4.5
14.0
8.1
Jptake Rate
)
Increase
- 0.2
6.8
4.1
4.1
1.8
1.5
1.7
- 0.8
4.0
  * Oxygen uptake  rate decreases with time and increasing
   2-ethoxyethanol  concentration, but  it is always greater for
   the test  cell  than the blank.  If significant inhibition were
   occurring in the concentration range tested, the oxygen
   uptake rate in the test  cell would  be less than the blank.


    The second type of test supported this observation.   In these
tests, no carbon  source or  feed chemicals other than
2-ethoxyethanol (in the test cell only) were added to the
activated sludge  culture.   Different concentrations of 2-ethoxy-
ethanol were  spiked in the  range from  23 - 1160 mg/L.  Each
concentration was tested in a separate experiment on a different
day using a  fresh sample of activated  sludge.  The initial oxygen
uptake rate  for both the spiked and unspiked cultures was measured
at each concentration- of 2-ethoxyethanol.  These data are
presented in  Table  2.  Note that, once again, the oxygen uptake
rate in the  spiked  cell is  consistently greater than in the
unspiked cell, indicating that the 2-ethoxyethanol is not toxic at
these concentrations.  This also implies that 2-ethoxyethanol is
degradable.   The  use of different samples of activated sludge is
the most likely source of the variability in the data in this set
of experiments.

    The bench-scale respirometry batch experiments were useful
toxicity screening  procedures.  Since  no toxicity was observed in
the range tested  (0-10,000  mg/L), continuous pilot-scale testing
at 2-ethoxy-  ethanol concentrations of 5000 mg/L and 10,000 mg/L

                                343

-------
was originally proposed.  During the acclimation phase, operation
at 5000 mg/L resulted in severe operational problems (bulking and
the production of slimy exocellular polysaceharides).  It was
hypothesized these problems were due to oxygen transport
limitations (relative to available carbon substrate) in the
microbiological floe resulting in the predominance of bulking
organisms and the formation of polysaccharide storage products.
The target test concentrations were subsequently reduced to 500
mg/L and 2500 mg/L to alleviate this problem.
   TABLE 2,  RESPIROMETRIC TOXICITY SCREENING STUDIES: EFFECT OF
             2-ETHOXYETHANOL CONCENTRATION ON OXYGEN UPTAKE RATE
         Amount 2—E Spiked
          (mL)      (mg/L)
       ial Oxygen .Uptake Rate
           fmL/L/hr)
Spiked   Unspiked  Difference
0.1
0.1
0.1
0.2
0.3
0.5
1.0
1.0
1.0
1.5
2.0
3.0
5.0
23
23
23
46
70
116
232
232
232
349
465
698
1160
16.3
19.7
12.8
21.8
15.0
30.0
14.7
26.2
31.3
18.8
30.0
27.5
21.1
8.8
12.2
7.3
8.2
8.8
12.5
5.4
10.0
9.8
8.9
7.4
10.0
8.0
7.6
7.5
5.4
13.6
6.2
17.5
9.3
16.2
21.5
9.9
22.6
17.5
13.1
                ACTIVATED SLUDGE PILOT PLANT STUDIES

DESIGN OF ACTIVATED SLUDGE PILOT PLANTS

    Two existing activated sludge pilot plants were used for this
portion of the  study.  See Figure 2.  A synthetic feed was chosen
over the primary effluent available at the T&E Facility in order
to reduce input variability and to facilitate consistent operation
of the pilot plants.  The basic formula was derived from a mixture
described by Kincannon and Stover(4). 2-Ethoxyethanol served as
the primary carbon source, eliminating glucose and glutamic acid.
At high 2-ethoxyethanol  concentrations, the formula required
nitrogen and phosphorus  supplementation in order to keep the
COD:TKN:P ratio at approximately 150:5:1.  Feed for each pilot
plant was prepared fresh daily from neat reagent grade chemicals
                                344

-------
and was stored in a refrigerator at 4 °C as it was pumped to the
appropriate pilot plant aeration basin.  Samples were collected
from each feed carboy just before the carboy was attached to the
system.

    Effluent from each system was continuously composited in a
refrigerated carboy except when the flow was diverted for
collection of 2-ethoxyethanol samples one time per day.  The
2-ethoxyethanol samples were collected in 40-mL vials with teflon-
lined septa.  The daily composite samples were used for all other
analyses.
                                    Off-GAS SAMPLING TRAPS
   Figure  2.   Activated Sludge Pilot Plant Experimental Apparatus


    Solids were wasted directly from the aeration basins.
Peristaltic pumps were used with timers to deliver the  appropriate
wasting flowrates, while providing  a high enough velocity to avoid
clogging.  Wastes were composited daily by pumping directly to
refrigerated  flasks.  The appropriate waste flowrate for control
of each system was calculated  and set daily based on mixed liquor
and effluent  volatile suspended solids.  Samples for all waste
mixed liquor  analyses were collected from the composites in the
refrigerated  flasks.

    The aeration basin off-gas was  sampled using the charcoal
sorbent tubes specified in NIOSH Method 1403  (5).  The  maximum
flowrate  specified in the method, 0.05 L/min, was used  for a total
of 2 hours per sample, or a maximum volume of 6 L.  A second tube
                                345

-------
was placed in series with the sample tube to quantitate any
breakthrough of 2-ethoxyethanol.


PROCESS MONITORING DATA

    Tables 3 and 4 summarize the routine process monitoring data
for solids retention time  (SRT), total and volatile suspended
solids (TSS/VSS), dissolved organic carbon (DOC) and 2-ethoxy-
ethanol (2-£) for target test concentrations of 500 mg/L and 2500
mg/L,  Mean values with standard deviations for the acclimation
phase  (three SRTs just before the test phase) and for the four-day
test phase itself are presented.

    Because 2-ethoxyethanol served as the primary carbon source in
the feed,  the amount of biomass produced in the aeration basin is
proportional to the 2~ethoxyethanol concentration.  The large
standard deviations for the acclimation phases are due both to
expected fluctuations inherent 1n the acclimation process and to
system upsets resulting from operational problems (temperature
fluctuations and plugging lines).  The DOC removals were
approximately 70% with the exception of the test phase for the
second concentration in which they averaged 86% as system
operation improved. The 2-ethoxyethanol removals were 80 - 90%
with the exception of the test phase for the second concentration
in which they reflected the improvement in system operation and
averaged 99%. Normally a longer recovery period after a system
         TABLE 3.  PILOT PLANT PROCESS MONITORING DATA FOR
                        500 MG/L 2-E SYSTEM

     Parameter                   Acclimation     During Test

                                    10.211.4

     Aeration basin TSS, mg/L      650 ±233        870  ± 60
     Aeration basin VSS, mg/L      571 ±228        793  ± 59

     Effluent TSS, mg/L             12 ±   9          5  ±  2
     Effluent VSS-, mg/L             8 ±   6          3  ±  2

     Influent DOC, mg/L            320 ±  10        305  ±  3
     Effluent DOC, mg/L             91 ±  40         94+5
     % DOC Removal                   71  %            69 %

     Influent 2-E, mg/L            479±18        4 95  ±40
     Effluent 2-E, mg/L             33 ±21         92  ± 18
     % 2-E Removal                   93  %            81 %
                                341

-------
         TABLE 4.  PILOT PLANT PROCESS MONITORING DATA FOR
                        2500 MG/L  2-E  SYSTEM
     Parameter
Acclimation
During Test
Average SRT, days 8
Aeration basin TSS, mg/L 3104
Aeration
Effluent
Effluent
Influent
Effluent
basin VSS, mg/L 2202
TSS
VSS
DOC
DOC
r
r
r
i
mg/L
mg/L
mg/L
mg/L
77
52
1540
481
% DOC Removal
Influent
Effluent
2-E
2-E
i
r
mg/L
mg/L
2255
345
% 2-E Removal
.4
±2136
± 1501
±
±
±
±
68
±
±
85
68
34
50
207
%
290
380
%
8
2065
1494
41
29
1485
208

2284
18

.8
±
±
±
±
±
±
86
±
±
99
156
167
4
5
84
65
%
201
17
%
upset would be desirable prior to test phase data collection, but
initiation of the test phase was constrained by project deadlines.
The amount of 2-ethoxyethanol measured in the effluents from both
systems shortly after seeding was under the detection limit of 10
mg/L.  The 2500 mg/L system returned to this degree of removal for
the last three of the four test phase days, resulting in better
removals than in the 500 mg/L system.


MASS BALANCE CALCULATIONS

    During the test phase for each concentration,  all streams
entering and leaving the system were analyzed for 2-ethoxyethanol
so a mass balance could be used to examine the fate of 2-
ethoxyethanol in the system and to indirectly estimate the amount
of removal attributable to biodegradation.  These streams included
the aeration basin off-gases and wasted mixed liquor in addition
to the influent and effluent.

    Aeration basin off-gases were sampled by passing 6 L of gas
through charcoal sorbent tubes.  The charcoal was then eluted
according to NIOSH Method 1403 (5) with methylene chloride
containing 5% v/v methanol and an internal standard.  The eluent
was analyzed using a gas chromatograph with a flame ionization
detector.  The detection limit for this method was 0.02 mg 2-
ethoxyethanol per tube {or 6 L gas).  No 2-ethoxyethanol was found
                                347

-------
in detectable quantities in the gas samples.  The detection limit
was used for purposes of mass balance calculations.

    Wasted mixed liquor samples were handled two different ways,
depending on the mixed liquor suspended solids concentration of
the sample.  For the system which was fed 500 mg/L 2-
ethoxyethanol, only the filtrate was analyzed because the mixed
liquor suspended solids concentration was low.  No 2-ethoxyethanol
was found in detectable quantities in this sample, and the
detection limit for the direct injection method discussed earlier,
10 mg/L, was used for purposes of mass balance calculations.  For
the system which was fed 2500 mg/L 2-ethoxyethanol, the suspended
solids concentration was high enough that it was necessary to
analyze the solids in addition to the filtrate.  The solids from
50 mL  of sample were collected on filter paper and extracted with
50 mL  of methanol.  The extracts were then analyzed using the
direct injection method.  No 2-ethoxyethanol was detected in the
filtrate or extract.  For the solids, the estimated detection
limit based on recoveries of 2-ethoxyethanol from methanol was
56.8 mg/L.  For mass balance calculation purposes, the total
amount of 2-ethoxyethanol which could have been present in the
mixed liquor sample without being detected was estimated as the
sum of the filtrate extract detection limits, i.e., 66.8 mg/L.

    The mass balances around each system for target 2-ethoxyethanol
concentrations of 500 and 2500 mg/L are illustrated in Figures 3
and 4, respectively.  The mass flowrates were calculated from the
measured 2-ethoxyethanol concentration (mg/L) multiplied by the
measured flowrate (L/min) for the given sample stream.  The term
"apparent biodegradation" represents the difference between the
mass flowrate of 2-ethoxyethanol into the system via the influent
and out of the system via the effluent, off-gases and wasted mixed
liquor streams.  Although this term is primarily due to
biodegradation, it does not strictly represent only biodegradation
because it also incorporates any errors inherent in sampling and
analysis as well as possible losses due to chemical reaction.  The
apparent biodegradation at 500 mg/L was >80.9%.  At 2500 mg/L, it
was >98.8%.  These values represent conservative estimates because
effluent concentrations in both systems were elevated due to
operational upsets in the systems discussed earlier.


                            CONCLUSIONS

    Respirometry and the anaerobic toxicity assay were used to
screen 2-ethoxyethanol for toxicity.  Toxicity did not occur in
the concentration range studied (0 - 10,000 mg/L) and
biodegradation was implied.  In pilot-scale testing of 2-
ethoxyethanol, the activated sludge system reduced the feed
concentration from 495 ± 40 to 92 ± 18 mg/L and from 2284 ± 201
mg/L to 18 ± 17 mg/L.  No 2-ethoxyethanol was detected in the off-
gas samples or in the wasted mixed liquor samples.


                                348

-------
           OFF-GflSES
          <0.002 : 0.0001
            mg/min
                    OFF-6»5IS
                   <0.002 s 0.0001
                     mg/mm
   INFLUENT
  5.593 ±0.417
   mg/mln
 EFFLUENT
0.677 ±0.161
 mg/mln
                              18.8%
             WBfTI MIKEB LIOUOR
            "* 0.007 ±0.0001
                mg/min
       RPPRRENTBIOOEGRflOBTION «>80.9% O2.910 mg/min)
INFLUENT
16.788:2.163
mg/mln


«O.I7,
DERATION
BASIN
                                               <0.37.
                                                  I-
                     Ll
UlflSTIMIKtD LIQUOR
 <0.055 • 0.011
   mg/mln
                 EFFLUENT
                <0.137 -0.111
                 mg/mln
                                                                    <0.8 %
                                             SPFflBtNT SIODECRRDflTIOH »>9«.8% OI6.S94 mg/mlnl
    Figure 3. Mass Balance  for
               500 rag/L System
          Figure 4.  Mass Balance for
                     2500 mg/L System
                               REFERENCES

1.  USEPA Office of  Solid Waste  and Emergency Response.   Test
    Methods for Evaluating Solid Waste, 3rd  Edition.  SW-846.  U.S.
    Environmental Protection Agency,  Washington,  DC, November 1S86.

2.  Bhattacharya, S.K.,  Safferman,  A., and Kupferle, M.J.
    Application of an acidogenic anaerobic process for
    detoxification of 2-ethoxyethanol and 2-nitropropane.   Paper
    presented at American Chemical Society Conference, Miami,
    Florida.  September 1989.
    Arthur,  R.M.  Baltimore Inhibition Test.
    publication, Fond Du Lac, Wisconsin.
                     Tech-Line  Instruments
    Kincannon, D.F.,  Stover, E.L.,  et al., Removal mechanisms for
    toxic priority  pollutants.   JWPCF.  55(2):  157-163,  1983.
    National Institute for Occupational Safety and Health.
    of  Analytical Methods,  3rd  Edition.   1984.
                                   Manual
                                   349

-------
          BENCH-SCALE BIODEGRADATION STUDIES WITH ORGANIC POLLUTANTS
                           USING A WHITE ROT FUNGUS

                      John A.  Glaser and  Henry  H.  Tabak
                     RREL, U.S.  EPA,  Cincinnati,  OH  45268

           Susan  Strohofer, Margaret  Kupferle, and  Pasquale  Scarpino
               Department of  Civil  and Environmental  Engineering
                           University of Cincinnati
                            Cincinnati, Ohio 45221

                                M.  Wilson  Tabor
                       Institute of Environmental Health
                          University of Cincinnati
                            Cincinnati, Ohio 45267


                                   ABSTRACT


      An array of naturally occurring and xenobiotic waste materials have been
shown to be degraded by the white rot fungus, Phanerochaete chrysosporium.
The ability of this fungus to degrade the wide spectrum of organic substrates
is attributable in part to the production of extracellular enzymes commonly
called lignlnases, for their ability to degrade lignin.  The mixture of
ligm'nases produced by P. chrvsosporium is composed of 10-15 peroxidases that
appear among the strongest biological oxidants known.  The production of these
enzymes occurs in a secondary metabolic state that is supported by the
cometabolism of more easily degraded carbon substrates.  The application of
this fungus to treat liquid phase wastes has been designed to use a rotating
biological contactor.  The HyCoR process has been demonstrated for the
treatment of pulp and paper industry waste water through the decolorization
and degradation of color-producing bodies and contaminating chlorinated
compounds.  Bench scale studies have advanced the general operation of this
reactor but scale up has not been undertaken.  This paper describes bench
scale studies devoted to a systematic investigation of operational factors and
their sensitivity for use at pilot scale.
                                  BACKGROUND


      White rot fungi are found throughout the Northern Hemisphere and are
known for their ability to degrade wood components (cellulose and lignin).
Phanerochaete chrysosporium, a white rot fungus, is a species that has been
selected for intense investigation due to its superior ability to degrade
lignin (!)•
      Lignin is a biogenic, high molecular weight polymer, composed of a
series of substituted phenyl propanoic units and is resistant to

                                     350

-------
biodegradation (2).  The biological polymerization process forms lignin
polymers that are disperse and non-homogeneous in composition (2).  For the
fungus to break down the highly variable lignin polymer structure, enzymes of
low specificity and low selectivity are required.  P. chrvsosporium is endowed
with such a complement of enzymes that enable it to degrade lignin.  The
fungal enzyme system is extracellular and is composed of a series of enzymes
called ligninases.  Ligninases are peroxidases that utilize hydrogen peroxide
to oxidize organic substrates.  Due to the low specificity of the ligninase
enzymes systems, P. chrysosporium has been tested for its ability to degrade a
variety of waste components: chlorinated phenols (3), eg. pentachlorophenol
(4,5), polyaromatic hydrocarbons (6,7), polychlorinated biphenyls (8),
polymeric dyes (9), and pesticides (10)-

      The powerful extracellular enzyme system is induced in response to
specific nutrient deprivation in nature.  Proper conditions for the induction
of ligninase production can be simulated in the laboratory and reactors
without lignin as a substrate (11,12).  The metabolic phase responsible for
this shift in activity is distinct from the primary growth of the organism and
is referred to as a secondary metabolic phase (13).

      One waste treatment application of white rot fungi is based on the use
of a liquid bioreactor configured as a rotating biological contactor, the
MyCoR process (14,15).  Other reactor configurations have been investigated
(16).  The MyCoR process has been demonstrated at bench scale for the
treatment of paper and pulp industry wastewater.  The color-producing bodies
and contaminating chlorinated compounds characteristic of these effluents are
removed by the process (15,17).  The original work using this fungus to treat
these effluents was first published in 1977 (20).
                                 INTRODUCTION


      This research program explores the MyCoR process for its potential use
as a hazardous waste site cleanup technology.  The MyCoR technology has been
studied at research scale to establish the proof of concept.  To extend this
line of development, we have chosen the decolorization of Kraft pulp mill
effluent as a means to begin the optimization of operational conditions at the
bench-scale.  This program addresses issues of application practicality to
develop cost-effective and efficient treatment technology.
                                 STUDY DESIGN


      All factors were held constant for the operation of a given reactor
throughout the growth and secondary metabolic phases of each experiment with
the exception of oxygen supply.  In selected experiments, growth was initiated
under an air atmosphere which was replaced with a pure oxygen supply at 48
hour operation. Variables studied included:

                                     351

-------
      (1) Air vs increased oxygen concentration
      (2) Glucose concentrations of 1.0%, 0.5%, and 0.1%
      (3) Nitrogen concentrations to support ligninolytic {secondary phase)
          and non-1igninolytic (primary phase) conditions
      (4) Choice of buffer either 2,2-Dimethyl Succinate (DMS) or Potassium
          Tartrate (KTar)
      (5) Disk rotational speeds were 1 or 6.5 rpm
                              MATERIALS & METHODS
WASTEWATER SOURCE
      Kraft pulp mm effluent from a post-chlorination, base treatment stage
was sampled on a schedule corresponding with the same type wood pulp
processing.  Periodic samples of 10-50 gallons were collected as needed to
maintain a steady source for reactor treatment studies.

REACTOR OPERATION

      Stationary cultures of fungus were maintained on a 3% agar (1% malt, 1%
glucose), pH 4.5.  Cultures were kept both at room temperature and Incubated
at 34° C.

      Bench-scale rotating biological contactors (RBCs) were constructed from
plans provided by Dr. T. Joyce, North Carolina State University (16).  These
reactors were operated in both growth and treatment phases in a constant
temperature room at 33-35° C.   The pH of the reactor contents was  maintained
at 4.25-4.7 through daily adjustments to 4.4-4.5 with IN HC1 or NaOH.

      Inoculation and development of biofilm in the reactors were conducted in
accordance with literature instructions except where changes were investigated
as part of the study (14,15).  Biofilms were grown both on synthetic media and
in the presence of pulp mill effluent supplemented by nutrient and mineral
additions.  Five to seven days for blofilm growth and treatment acclimation
were required before treatment assessment.

REACTOR MAINTENANCE PROCEDURE

      Sample collection and pH adjustment was scheduled on at least a daily
basis.  Removal of liquid and pH adjustment was conducted by means of cannula
and a syringe.  Sample collection and operational time did not change the
Initial liquid volume of 2-L per reactor by more than 10% at completion of
batch runs.  Wet weights of biofilms were measured by upon completion of the
reactor batch runs after draining the liquid phase from each RBC.

      The gas atmospheres were established by allowing free gas exchange with
the ambient atmosphere for air mediated experiments whereas elevated oxygen
concentrations were established by sealing the lid of the RBC with tape and

                                     312

-------
supplying pure oxygen to the reactor.  For the supplied oxygen experiments,
the headspace of the reactors was aggressively flushed with pure oxygen for 5-
15 minutes before the oxygen flowrate was adjusted to a constant rate of
20mL/mi n.

ANALYTICAL TECHNIQUES

      Dissolved oxygen was measured by YSI Model 54A meter with stirring BOD
probe YSI model 5720A.  Ambient atmospheric oxygen concentration was measured
with a Fisher Model 1200 Gas Partitioner.  Spectroscopic measurements of
UV/Vis were measured with a Beckman Model 25 spectrophotometer. Measurements
were made on filtered liquid samples.  Wavelengths were chosen to monitor the
reduction of substrate concentration throughout the batch treatment periods
which ranged from 2 days to 1 week.  Due to the heterogeneity of the Kraft
pulp mill effluent, a single wavelength was considered inappropriate to
measure rates of conversion.  The wavelengths of 580 nm (measuring the yellow-
orange color region), 450 nm (for humic substance detection), 280 nm (for
aromatic structure and lignin fragment detection) were used (19,20).  In pulp
mill wastes, peak absorbance is at 220 nm.  A wide flattened peak is found at
280 nm.  The highly concentrated nature of the effluent required dilution for
spectrophotometric measurements,  A nine percent dilution was found to be in
the linear range to calculate a relative percent concentration based on Beer's
law at 280 nm.  Color was measured as platinum color units (PCU) which is a
standard method developed by the paper and pulp mill industry  (23).

      Treatment capacity was assessed by the total and percent reduction of
the representative wavelength absorbances and color units original zero time
sample (TO) set to equal 100%.  Data for initial experiments were collected by
measuring absorbance at 580 nm only.  Subsequent experiments showed a strong
linear correlation between the visible wavelengths and color units.  The
regression data for this correlation are as follows:

     m = 0.000073   b = -0.02183   R2 = 0.926211  no. observations - 143.

      Consecutive treatment results in the same reactor/biofilm were averaged
to minimize the biological variability (22).  Percent reduction at 280nm
represents reduction of substrates related largely to aromatic fractions and
not total color.  The correlation of color units and absorbance at 280 nm
exhibits a large scatter indicating a lack of linear correlation between color
and UV absorbance.  Absorbances at the 220 nm wavelength fluctuated throughout
the treatment study making assessment of these results difficult.  Following
acclimation, a period of 1 day was considered significant treatment time.
More critical evaluation assessed treatments under 24 hours (Table 1).
                                     353

-------
         TABLE  1.  REMOVAL MEASUREMENTS BY  DIFFERENT MEANS  OF  DETECTION

REACTOR CONDITIONS
GLUCOSE X
1.0
1.0
0.5
0.5
DISK SPEED
(rpm)
1
6.5
6.5
1
Total Average Removal (96)
8 HR
280 nM
10
13
11
5
11
Color Units
39
41
33
27
39
17 HR
280 nM
19
23
19
20
21
Color Units
57
59
55
51
55
      Continuous  liquid-liquid txtraction  (EPA Method 3250) followed by GC/MS
EPA 8270 for chlorophenols.  A Hewlett-Packard SC/MS with HP autosampler and
HP software was used  for peak integration.  For semi-quantisation and
preliminary data  generation GC/FID was used.
                                  DISCUSSION
GROWTH
      No discernible difference in growth or activity was observed in the.
biof11ms inoculated with spores cultivated at room temperature compared to
those at 34° C.  Sufficient numbers of spores for the Inoculation  of 2-6
reactors were grown within 10 days at 34° whereas a 2 weeks  was  required  for
room temperature growth of spores.

      No significant differences were observed in the consecutive  treatment
performance of biofilms developed in the presence of pulp mill effluent and
those developed purely on synthetic media.  An initial growth and  acclimation
period of 1 week was required after start-up in both cases.
                                 354

-------
      Under all conditions of growth and treatment, but not at all times,
black and red coloration of the blofilm developed In the form of localized
"speckled" areas.  Speculation has attributed such coloration to one or more
of the manganese dependent lignlnase enzymes.

      When environments of purified oxygen were established with the Initial
Inoculation of a reactor, a tendency was observed for some of the fungal mass
to group into spherical clumps on the discs.  This was not observed when
blofllm growth was begun under air.

OXYGEN CONCENTRATION

      It 1s known that fungal growth is facilitated by lower oxygen
concentrations (23), but that the lignln degrading system of this fungus 1s
positively influenced by Increasing the oxygen concentration In the transition
between primary and secondary metabolism (24).  No discernible difference in
treatment was noted when cultures were placed initially under art Increased
oxygen atmosphere despite the smoother appearance of the blofllm.  The cost of
oxygen supports the practice of early growth under air for consideration at
pilot and field scale.

      Invariably with all other parameters held constant, more biomass
developed under air than Increased oxygen.  When treatment failed 1n a set of
reactors under increased oxygen, the fungus grew to a weight comparable to
that grown 1n the presence of air.

      Initial acclimation to the pulp mill effluent was found to be slower in
the presence of air than 1n the presence of higher oxygen concentrations.
These reactors decolorized dilute effluents comparably.  The oxygen assisted
reactor systems showed comparable treatment activity for both dilute (< 3000
PCU) and concentrated (>4000 PCU) effluents.  However, reactors with air
atmospheres showed greatly diminished decolorizatlon rates for concentrated
pulp mill effluent than the rate observed for the more dilute effluents (Fig.
1).
                  A
                  k
                  *
                  o
                  r
                  b
                  •
                  n
                  0
                  *

                  8

                  0
                  a
     C
     O

     t
     h
     e
     u
     *
     •
     n
     d
                                2O   80   40   80

                                   Time (hrs)
                                                  eo
 o
TO
                 Figure 1. Effect of oxygen on concentration.

                                 355

-------
      In similar studies, Yin et. al. demonstrated the kinetics  of color
removal by £. chrvsosporium is concentration dependent with effluents
containing <8000 PCU (22).  The effluent concentration used in our studies was
4000-5500 PCU.  Hence, oxygen concentration and its availability are
determinative factors for the biological treatment of these effluents in RBCs.
Therefore, these results suggest strong correlation between oxygen
requirements and pulp mill effluent color concentration.  The relationships
indicated by these observations will be integral to commercial operation of a
treatment system of this design.

GLUCOSE

      It is known that glucose is used de novo to produce hydrogen peroxide
from molecular oxygen.  This product is used by the ligninases in the
oxidative depolymerization of lignin (24).  The dual role for glucose in both
growth and secondary metabolism is supported by the growth and treatment
results described in this study.

      The amount of biofilm growth is not directly proportional to glucose
concentration.  Under increased oxygen concentrations, 0.5% glucose performed
treatment as efficiently as 1.0%.  See tables 1 and 2.  Under air, only with
dilute effluent, the activities of the initial acclimation/treatment batch and
the second consecutive batch were comparable at the two glucose
concentrations.  The treatment of later batches were accompanied with a
decline in the rate of decolorizing activity at the 0.5% glucose level.
However, the extent of decolorization ultimately was maintained by employing
longer treatment periods.  The decolorization rate was maintained at 1%
glucose for dilute effluent feedstocks.  An additional factor of carbon source
is indicated here in terms of concentration, oxygen and kinetics.  A critical
concentration between 0.1 and 0.2% glucose has been identified as necessary
for decolorization with supplied oxygen (26).  Preliminary results show
activity at 0.1% yet at a decreased rate.

NITROGEN

      Experiments using nitrogen-rich media, conditions that do not induce the
lignin degrading secondary metabolic state in non mutant fungal species,
throughout the treatment resulted in an interesting "non-active" condition.
Physical attachment of color bodies was noted on the biofilm for a period of
1-2 days followed by their release back into the liquid matrix.  Substrates
monitored at the 280 nm wavelength also displayed this uptake/release
phenomenon.  This is to be contrasted with conditions of observed "inactivity"
where no initial uptake by the fungus is observed and no treatment occurs
(Fig. 2).  In slower acting systems (i.e. under air), residual attached color
from a previous batch is decolorized upon the addition of a new batch of
effluent containing a fresh supply of glucose (i.e., H20H).
                                     356

-------
                    O.U
                    O.M
                                                           C
                                                           U


                                                           h
                                                           o
                                                           u
                                    Time (hrs)
                        Figure 2.  Performance overview.
      Physical attachment or adsorption of substrate to the fungal biofUm Is
primary to Initiating subsequent degradation by the fungus.  Tho treatment
"Inactivity" observed may be an Inhibition of this first step.  A total
removal was based on measurement of the liquid concentrations of color and
Individual organic substrates.  For example tHchlorophenol concentrations of
18 ppb In the pulp mill effluent were not detected after treatment even In the
case where the desired enzymatic conditions for treatment had not been
established.  This loss of phenol was attributed to the a great retentive
characteristics of the developed non-11gn1nolyt1c biofllm.

BUFFER

      No difference 1n growth or treatment extent was observed when potassium
tartrate (KTar) was used In the place of 2,2-dlmethyl succinate (DMS) buffer
for biofllm development at 1% and 0.5% glucose concentrations with the reactor
set at 1 rpm.  The QMS buffer has been previously used for the optimization of
llgnlnise production (27).  Choice and use of buffer 1s of critical Importance
for economic considerations for larger scale operations.

DISK ROTATIONAL SPEED

      Development of biofllm at 1 rpm Invariably resulted in bridging between
at least two of the discs% whereas this never occurred at 6.5 rpm.  In
addition, 1t appeared that the films at 6.5 rpm were more evenly distributed
with an overall smoother surface.  No difference In treatment activity was
observed although dissolved oxygen concentrations at 6.5 rpm were generally
double those at 1 rpm (Table 2).  The Influence of disk rotational rate maybe
more pronounced for larger scale operations but the effects may be difficult
to translate from bench scale.
                                      357

-------
                      TABLE 2. REACTOR PERFORMANCE TESTING

CONDITIONS
% GLUCOSE
1.0
1.0

0.5
0.5
RPM
1
6.5

1
6.5
TOTAL AVERAGE REMOVAL(%J
UV WAVELENGTH
(280 nM)
28
28
17
25
COLOR UNITS
71
70
49
66

22
29
24
30
46
63
56
39
KRAFT PULP MILL  EFFLUENT VARIABILITY

      Differences  1n  sampling lots of the Kraft pulp mill effluent and changes
in the lot during  the time of storage may explain the variability of treatment
results.  For  a  total of five lots of effluent treated by our system, there
were distinct  changes in treatability.  The sample lots of the effluent can
differ significantly.  Their composition depends on daily mill processes and
feedstocks.  Pulp  rail! effluent for softwood is typically dark in color,
whereas the hardwood  wastewater is less concentrated.  No trend can be
observed to identify  the time interval for greatest treatment activity.
Maximum treatment  effects were observed for both early and late periods 1n the
lifetime of the  biofilms.  An apparent toxicity developed within the effluent
lots over .time,  normally encountered within 2 months after collection of the
effluent supply.   At  this stage, fungal treatment became less effective and
ultimately the fungus would not grow in the presence of that effluent lot.
Toxins showing-activity towards fungi have been previously noted in these
effluents (24).  Bacterial contamination is a possibility and is also
suspected.  Continuing investigation of is problem is required for full
assessment of the  bench scale treatment data.
                                3S8

-------
                       AREAS  OF  CONTINUING  INVESTIGATION
   The kinetic effects of oxygen and glucose concentrations on treatment.
   Influence of substrate for treatment on optimal conditions.
   Effluent variability and its control.
   Application to treatment of chlorinated phenols and polyaromatic
   hydrocarbons.
   Economic evaluation of operational factors.
                               ACKNOWLEDGEMENTS

      Special thanks to Dr. Thomas Joyce and his coworkers at NC State
University for their helpful advice and assistance. We are indebted to Dr.
Wipawan Nisamaneepong for analytical support.  Thanks, Latrice Martin and Steve
Bradshaw.
                                  REFERENCES

1.  Buswell, J. and Odier, E.  1987.  Lignin Biodegradation. Crit. Revs.
    Biotechnol. 6: 1-60.

2.  Kirk, T.K.  1980.  Studies on the physiology of lignin metabolism by the
    White-Rot fungi, pp. 51-64.  In:  T.K. Kirk, T. Higuchi, H.M. Chang,
    (editors), Lignin Biodegradation: Microbiology, Chemistry and Potential
    Applications,  Vol. 2. CRC Press, Inc., Boca Raton, Fla.

3.  Hammel, K.E. and Tardone, P.  1988.  The oxidative 4-dechlorination of
    polychlorinated phenols is catalyzed by extracellular fungal lignin
    peroxidases.  Biochem. 27: 6563.

4.  Mileski, G.J., Bumpus, J.A., Jurek, M.A. and Aust, S.D.  1988.
    Biodegradation of pentachlorophenol by the White Rot fungus Phanerochaete
    chrvsosporium.  Appl. and Environ. Microb.  54: 2885.

5.  Lamar, R., Glaser J., and Kirk, T.K.  1990. Fate of pentachlorophenol
    (PCP) in sterile soils inoculated with the white-rot basidiomycete
    Phanerochaete chrysosporium: mineralization, volatilization, and depletion
    of PCP.  Soil Biol. Biochem. in press.

6.  Sanglard, M.S., Sanglard, D., and Fiechter, A.  1986. The role of
    extracellular ligninase in biodegradation of benzo(a)pyrene by
    Phanerochaete chrvsosporium.  Enzy. Microb. Technol.  8: 209.

7.  Hammel, K.E., Kalyanaraman, B., and Kirk, T.K.  1986.  Oxidation of
    polycyclic aromatic hydrocarbons and dibenzo(p)-dioxins by Phanerochaete
    chrvsosporlum ligninase. J. Biol. Chem.  261: 16948.

                                     358

-------
8.  Eaton, D.C.  1985.  Mineralization of polychlorinated biphenyls by
    Phanerochaete chrvsosporium; a ligninolytic fungus.  Enzy. Microb.
    Technol.  7: 194.

9.  Glenn, O.K., and Gold, M.H. 1983.  Decolorization of several polymeric
    dyes by the lignin-degrading basisiomycete Phanerochaete chrvsosporium.
    Appl. Environ. Microb.  45: 1741.

10. Bumpus, J.A., and Aust, S.D. 1987. Biodegradation of DDT [1,1,1-trichloro-
    2,2-bis (4-chlorophenyl)ethane] by the White Rot fungus Phanerochaete
    chrvsosporium.  Appl. Environ. Microb.  53: 2001.

11. Keyser, P., Kirk, T.K., and Zeikus, J.G.  1978.  Ligninolytic enzyme
    system of Phanerochaete chrvsosporium; synthesized in the absence of
    lignin in response to nitrogen starvation.  J. Bacteriol.  135: 790.

12. Jeffries, T.W., Choi, S., and Kirk, T.K.  1981.  Nutritional regulation of
    lignin degradation by Phanerochaete chrvsosporium.  Appl. Environ. Microb.
    42: 290.

13. Kirk, T.K., Connors, W.J., and Zeikus, J.G.  1976.  Requirement for a
    growth substrate during lignin decomposition by two wood-rotting fungi.
    Appl. Environ. Microb.  32: 192.

14. Chang, H-M., Joyce, T.W., Kirk, T.K., and Huynh, V.B.  1985.  Process of
    degrading chloro-organics by the White Rot fungi.  US Patent No.
    4,554,075.

15. Chang, H-M., Joyce, T.W., and Kirk, T.K.  1987.  Process of treating
    effluent from pulp or paper-making operation. US Patent No. 4,655,926.

16. Huynh, V-B., Chang, H.M., and Joyce, T.K.  1985.  Dechlorination of
    chloro-organics by a White Rot fungus.  Tappi J.  68: 98.

17. Lewandowski, G.A., Armenante, P.M., and Pak, D.  1990.  Reactor Design for
    hazardous waste treatment using a White Rot fungi.  Water Res.  24: 75.

18. Fukuzumi, T., Nishida, A., Aoshima, K., and Minami, K.  1977.
    Decoulorization of kraft waste liquor with White Rot fungi.  Mok. Gak.
    23: 290.

19. Janshedar, H., Brown, C. and Fiechter, A.  1981.  Determination of
    biodegraded lignin by ultraviolet spectrophotometry.  Anal. Chim. Acta.
    130: 81.

20. Alen, R., and Hartus, T.  1988.  UV spectrophotometric determination of
    lignin from alkaline pulping liquors.  Cell. Chem. Technol.  22: 613.

21. National Council of the Paper Industry for Air and Stream Improvement,
    Inc.  1971.  NCASI Technical Bulletin No. 253.  National Council of the
    Paper Industry for Air and Stream Improvement, Inc. New York, NY.

                                     360

-------
22. Yin, C., Joyce, T., Chang, H.  1989.  Kinetics of bleach plant effluent
    decolorization by Phanerochaete chrvsosporium.  J. Biotechnol,  10: 67.

23. Tabak, H.H., and Cooke, W.B.  1968.  The effects of gaseous environments
    on the growth and metabolism of fungi.  Botan. Rev.  34: 126.

24. Bar-Lev, S.S., and Kirk, T.K.  1981.  Effects of molecular oxygen on
    lignin degradation by Phanerochaete chrysosporium.  Biochem and Biophys.
    Res. Comm.  99: 373.

25. Kelley, R.L., and Reddy, C.A.  1988.  Glucose oxidase of Phanerochaete
    chrvsosporiunu pp. 307-316.  In: Wood, W.A., and Kellogg, S.T. (editors),
    Methods in Enzymology.  Vol. 161.  Academic Press, San Diego, CA.

26. Yin, C.F., Joyce, T.K., and Chang, H.M.  1989.  Role of glucose in fungal
    decolorization of wood pulp bleaching effluents.  J. Biotechnol.  10: 77.

27. Fenn, P., and Kirk, T.K.  1979.  Ligninolytic system of Phanerochaete
    chrvsosporium: inhibition by o-Phthalate.  Arch Microbiol.  123: 307.
                                     361

-------
EVALUATION OF EPA SOIL WASHING TECHNOLOGY .FOR REMEDIATION AT UST SITES

Mary E. Tabak, P.E.
William Glynn
Camp Dresser & McKee Inc.
Boston, Massachusetts

Richard P. Traver, P.E.
Chapman, Inc.*
Freehold, New Jersey

                                   ABSTRACT

     The U.S. Environmental Protection Agency through its Risk Reduction
Engineering Laboratory's Release Control Branch has undertaken research and
development efforts to address the problem of leaking underground storage
tanks (USTs).  Under this effort, EPA is currently evaluating soil washing
technology for cleaning up soil contaminated by the release of petroleum
products from leaking underground storage tanks.  Soil washing is a dynamic
physical process which remediates contaminated soil via two mechanisms -
particle separation and dissolution of the contaminants into the washwater.
As a result of the washing process, a significant fraction of the contaminated
soil is cleaned, and can be returned into the original excavation or used as
cleaned "secondary" fill or aggregate material.  Since the contaminants are
more concentrated in the fine soil fractions, their separation and removal
from the bulk soil increases the overall effectiveness of the process.
Subsequent treatment will be required for the spent washwaters and the fine
soil fractions.

     The soil washing program evaluated the effectiveness of soil washing
technology in removing petroleum products (unleaded gasoline, diesel/home
heating fuel, and waste crankcase oil) from an EPA-developed Synthetic Soil
Matrix (SSM) and from actual site soils.  Operating parameters such as contact
time, washwater volume, rinsewater volume, washwater temperature and
effectiveness of additives were investigated.  Further work was conducted to
determine what, if any, effect additives have when added to washwater.  The
additives investigated were CitriKleen (a biodegradable degreasing agent) and
an organic surfactant.  Actual soils from UST sites in Ohio and New Jersey
were washed using the optimum parameters derived for the SSM.

     This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
* Prior to joining Chapman, Inc., in Freehold, NJ, Mr. Traver was the EPA
  Technical Project Monitor for this project.

                                    362

-------
                                 INTRODUCTION


     Based on the Hazardous and Solid Waste Amendments of 1984 and Its Land
Ban Regulationsi EPA has discouraged the excavation and landfill disposal
practices of the past for contaminated soils resulting from leaking
underground storage tanks (USTs).  EPA has encouraged the use of on-slte
treatment technologies, however, problems have plagued the development of
on-site treatment technologies for the treatment of petroleum contaminated
soils.  Technical support is needed to develop effective long-term corrective
actions at leaking UST sites, design cleanup program guidance, aind help
implement state programs.

     The remedial options available for the treatment of contaminated soils
from UST sites are broadly segregated into two main categories, namely those
which remove the contaminants without excavation (in situ techniques) and
those which require excavation of the soil and subsequent cleaning on-site.
The former group of remedial options have not yet been demonstrated for high
efficiency removal of contaminants from the subsurface.  These techniques are
plagued by the uncertainty of soil contamination levels in the subsurface
after treatment.  Soil excavation followed by extensive cleaning of the soil
will ensure a more complete removal of contaminants over in situ techniques.

     On-site soil washing of excavated soils is a viable alternative to in
situ techniques and has been shown to be effective for the cleanup of
hazardous waste contaminated soils (EPA, 1988).  The goal of this effort is to
determine the feasibility of soil washing for cleaning up petroleum
contaminated soils.

     Soil washing is a physical process in which excavated soils are contacted
with a liquid medium, usually water.  The two principle cleaning mechanisms
include the dissolution of the contaminants into the extractive agent and/or
the dispersion of the contaminants into the extraction phase in the form of
particles (suspended or colloidal).  The separation of the highly contaminated
fine soil particles (silts, clay and colloidal) from the bulk of the soil
matrix can result in volume reduction of the bulk soil.  As a result, a
significant fraction of the contaminated soil is cleaned and can be put back
into the original excavation.  Since  the contaminants are more concentrated
in the fine soil fractions, their removal from the bulk soil increases the
overall effectiveness.  Subsequent treatment will be required for the spent
wash waters and the fine soil fractions.

     Under Phase I of this research program, a synthetic soil msitrix (SSM)
containing a range of petroleum products at varying concentration levels was
prepared and subjected to bench-scale performance evaluations of soil washing
technology.  Operating conditions derived from tests using SSM vere used to
evaluate soil washing technology on actual samples from leaking UST sites in
Ohio and New Jersey.

     Prior to preparing the quantities of SSM needed for the bench scale
tests, several bench scale experiments were performed to develop a
dose/response relationship between,the quantity of petroleum product added to

                                    363

-------
the soil matrix and the analysis quantification.  The petroleum products
evaluated during this study include unleaded gasoline, diesel oil and waste
crankcase oil.  The SSM was then blended with a specific quantity of petroleum
product to obtain a predetermined concentration level.  Total petroleum
hydrocarbon (TPH) analysis was performed to verify the concentration levels
for diesel and waste oil, and benzene, toluene, ethylbenzene and xylenes
(BTEX) analysis was performed to verify the concentration levels for gasoline.

     The bench-scale washing experiments were designed to simulate the
EPA-developed pilot-scale Mobile Soils Washing System (MSVS).  Specifically,
the bench-scale experiments were designed to simulate the drum-screen washer
which separates the >2-mm soil fraction (coarse material) from the <2-mm soil
fraction (fines) by use of a rotary drum screen.  A high pressure water knife
operates at the head of the system to break up soil lumps and strip the
contaminants of the soil particles.

                    SYNTHETIC SOIL MATRIX CHARACTERIZATION

     The basic formula for the SSM was determined by others from an extensive
review of Superfund sites and a review of the composition of eastern U.S.
soils (Traver, 1989 and PEI Assoc., 1988).  The SSM was a mixture of clay,
silt, sand, top soil and gravel, prepared by others in two 15,000-pound
batches.

     A review of the existing soil characteristics was made and additional
tests were conducted to further delineate the physical and chemical properties
of the SSM.  The tests included particle size distribution, moisture
retention, Atterberg limits, cation exchange capacity, base saturation,
organic matter, chemical constituents and mineralogy.  Quantification and
assessment of these specific properties will assist the technical community to
understand the differences that may be observed between the performance of
soil washing technology on the SSM and on actual UST site soils.  The results
of the characterization tests are presented in Tables 1-5 and Figures 1 and 2.

     The test results indicated that the SSM is composed of 60 percent sand,
19 percent silt and 21 percent clay as determined by particle size
distribution analysis (Table 1 and Figure 1).  Based on this composition the
SSM would be classified (USDA) as having a sandy clay loam texture.

     The moisture content of the SSM ranged from 33.1 percent at saturation (0
bar) to 8.7 percent at the permanent wilting point (15 bars).  The moisture
content at field capacity (0.1 bar) was 21.0 percent.  The moisture-retention
curve (Figure 2) developed from the moisture content data was indicative of a
finer textured soil.  The moisture content data can be used to evaluate
moisture and chemical characteristics of the SSM.  For example, the amount of
soil water that can be extracted from the SSM under typical environmental
conditions (0 to 15 bars) will be 24.4 percent.  The remaining soil water is
considered as "unavailable", which can only be removed by artificially induced
vacuums or pressures.
                                    364

-------
                              DOSE/RESPONSE TEST

     The purpose of the dose/response test vas to determine the appropriate
amounts of gasoline or diesel fuel vhich should be added to the SSM to obtain
the desired concentration of selected constituents or parameters (e.g., BTEX
or TPH).  Various mixtures of SSM and gasoline or diesel vere prepared and the
soil analyzed for benzene, toluene, ethylbenzene and xylenes (BTEX) and total
petroleum hydrocarbons (TPH).

     The procedure used vas to place known amounts of SSM, water, and either
gasoline or diesel into vide mouth glass jars.  The contents vere mixed by
hand for five minutes, and then the contents vere submitted to a laboratory
for analysis of TPH and BTEX.

     The results of the dose/response tests are shovn in Table 6.  The lab
tests indicate that the soils reach a level of liquid saturation at about 23%
liquid (both vater and gas or diesel).  The tests vere conducted such that the
soils vere all prepared to a 20% vater level.  However, this limited the
amount of gas or diesel vhich could be mixed into the soil mixture.  At 201
vater, the highest achievable BTEX concentration vas about 3000 mg/kg.  For
diesel, at vater content of 2QX, the highest TPH concentration vas 60,000
mg/kg.

     The dose/response curves are plotted in Figures 3 and 4.  A linear reg-
ression of the data yielded the following relationships for dose/responses:

Gasoline:

G = gasoline concentration, mg/kg
B = BTEX concentration (sum of benzene, toluene, ethylbenzene, and
xylenes), mg/kg

                   G . 13.33(8) - 375                                  (1)

The correlation coefficient r, for this equation is 0.998.

Diesel:

D - diesel fuel concentration, mg/kg
T = TPH concentration, mg/kg

                   D » 0.675(T) + 6268                                 (2)

The correlation coefficient r, for this equation is 0.995.

     An additional experiment vas conducted to determine if a higher level of
BTEX could be achieved if a lover moisture content vas used.  In this
experiment, a sample of soil vas moistened to a. 10% vater content and then
saturated vith gasoline.  The results, shovn in Table 6 as "Jar 9", indicated
that the BTEX content vas increased to 4670 mg/kg.
                                    365

-------
     Equations 1 and 2 can be used  to determine  the amount of diesel or
gasoline to add to  the SSM to reach the desired  concentrations.  Based on
these calculations, estimates were  made to determine  the amount of gasoline
and diesel fuel to  add to the SSM to obtain  the  desired concentrations of BTEX
and TPH for the bench scale experiments.  The SSM blends were prepared in the
EPA SSM Blending Facility in Edison, NJ in 50-Ib batches for use ia the bench
scale soil washing  experiments.

                        BENCH SCALE SOIL WASHING TESTS

     The experiments involved washing the SSM spiked  with gasoline, diesel
fuel, or waste crankcase oil under  several operating  conditions to obtain the
sensitivity of various parameters affecting  soil washing efficiency.

     The experiments were conducted by contacting approximately 1400 g of soil
with varying amounts of washwater.   The contact  time  was varied as was the
rinsewater volume.  The washing of  the soils was conducted by shaking the soil
and washwater in a  2-gallon jar in  a shaker  table operating with a stroke and
frequency of 1.6 inches and 4 Hz respectively.   The rinsing of the soils was
performed in a Gilson Vet-Vac Model WV-1 which both rinsed the soils as well
as separated the particles into three fractions  using No. 10, No. 60 and No.
140 sieve trays.  The process of the washing and rinsing yielded five distinct
fractions - the soils on the three  sieve trays,  a washwater, and a rinsewater.
All fractions were  measured for mass (or volume) as well as contaminant
concentration.  A measure of total  BTEX (benzene, toluene, ethylbenzene and
o~, m-, and p-xylenes) was used on  gasoline  spiked soils, and total petroleum
hydrocarbons (TPH)  was used on diesel spiked soils.

     Preliminary screening tests were conducted  on soils spiked with diesel
and gasoline to determine the optimum conditions for  contact time, washwater
volume, rinsewater  volume and washwater temperature.  The parameters tested
are listed in Table 7 along with an assessment of how the parameter affected
removal of TPH or BTEX from the soil and on  the  particle separation during the
washing process.  As shown in this  table, most parameters had minimal effect
on the removals and particle separation for  the  ranges tested.  Increased
contact time did somewhat improve the contaminant removal and particle
separation.  The addition of CitriKleen did not  improve contaminant removal,
and in fact actually decreased the  removals.  Increased temperature did not
improve removals for soils washed with plain water, but did improve the
removals somewhat for washwaters containing CitriKleen or surfactant.  Typical
results obtained for these experiments are shown in Table 8.

     The results of the experiments  indicate that the optimal washing
conditions for SSM  spiked with diesel or gasoline are:  20-to 30-minute
contact time, Isl soil to washwater mass ratio,  3:1 rinsewater to washwater
volume ratio, and ambient temperature for the washwater.  These conditions
resulted in a 90+%  removal of TPH and BTEX in the No. 10 and No. 60 sieve
fractions.  It should be noted that  these conditions  represent the most
cost-effective operating conditions  for bench scale treatment of SSM using
soil washing technology.  Operating conditions for each site soil may vary and
should be determined on a case by case basis.

                                    366

-------
     Experiments were then conducted on site soils from four UST sites located
in Ohio and New Jersey.  The results, shown in Table 9, indicate that the
removals in the 10 and 60 sieve fraction were significantly lower than the
removals achieved for SSM.  The site soils had a significantly different
particle size distribution than the SSM, as shown in Table 10.  The bench
scale soil washing apparatus was able to recover only 55% of the SSM mass
after washing, and 45% was washed off into the washwater and rinsewater.
Since the site soils contained lower amounts of fines (7 to 43% by weight for
the four sites tested), greater amounts of soil were recovered from the
washing process, resulting in lower suspended solids in the washwaters, as
shown in Table 11.

                           SUMMARY AND CONCLUSIONS

     The soil washing experiments involved washing the SSM spiked with
gasoline, diesel fuel, or waste crankcase oil under several operating
conditions to obtain sensitivity analyses on various parameters affecting soil
washing efficiency including:  contact time with washwater, washwater volume,
washwater temperature, and chemical additives to the washwater such as a
surfactant and a degreasing agent (CitriKleen).  The SSM experiments yielded
highly reproducible results.  Since the soils were prepared in the SSM
blending facility, the soil characteristics and contaminant content were
homogeneous throughout the matrix.  Experiments were also conducted using
actual site soils, but due to the heterogeneity of these soils, the
reproducibility of these experiments was much lower than for the SSM.

     The results of the optimization tests using SSM indicated that removals
greater than 90% of petroleum products could be achieved for the SSM.
However, as washing conditions were varied, no significant change in the
removals could be detected for most parameters tested.  The contact time in
the wash cycle affected the percent removal possibly due to improved
mechanical separation of the fines from the larger particles with more shaking
time in the wash step.  Increased washwater and rinsewater volumes did not
significantly improve contaminant removal.  The use of additives such as
CitriKleen and surfactant did not improve removal of contaminants.  Increased
amounts of CitriKleen in the washwater caused decreased removals of
contaminants which may be attributed to enhanced adsorption of contaminants  to
the soil due to the presence of CitriKleen adsorbed onto the matrix.
Experiments using surfactant resulted in considerable foaming which would lead
to operational problems in pilot or full scale operation.

     The experiments using actual site soils resulted in lower removals of
contaminants than for the SSM experiments using the same washing conditions.
The washwaters for the actual site soils also contained much lower suspended
solids than the SSM washwaters.  These results indicate that the primary
mechanism for contaminant removal for the SSM was particle separation.  Since
this mechanism alone was able to account for a high percentage of contaminant
removal, the experiments showed no significant improvement when conditions
were varied.  The only parameters which seemed to improve removals were those
which would improve the particle separation in the soil matrix.
                                    367

-------
     Although the SSM experiments achieved high removals, only 55% of the
washed soil mass was recovered, and a vashvater containing over 20% solids vas
produced.  The vashvaters from the actual site soils experiments had lover
suspended solids, but also resulted in lover contaminant removals from the
soil matrix.  Therefore, enhancement of the solubilization mechanism would be
required to effectively remove contaminants from soils containing a small
fraction of fine materials.  The resultant vashvaters vould also contain lover
suspended solids.  Future vork to evaluate the enhancement of the
solubilization mechanism in soil washing is recommended.

                                  REFERENCES

EPA Office of Solid Vaste and Emergency Response, Assessment of International
    Technologies for Superfund Applications, EPA/540/2-88/003, (1988).

PEI Associates, Inc.  CERCLA BDAT SARH Preparation and Results ofPhysical
    Soil Washing Experiments, Cincinnati, OH (1988).

S.P. Traver, Developmentand Use of EPA's Synthetic Soil Matrix (SSM/SARM).
    U.S. EPA Releases Control Branch, Risk Reduction Engineering Laboratory,
    Edison, NJ (1989).
                                    368

-------
TABLE 1.  USEPA SYNTHETIC SOIL MATRIX PARTICLE SIZE DISTRIBUTION (USDA)
              SOIL FRACTION
  USDA(%)
USCS(%)
Gravel
Sand (Total)
Very Coarse
Coarse
Medium
Fine
Silt
Clay
—
60.0
16.0
8.8
11.7
23.5
19.0
21.0
—
58




15.2
26.8
    TABLE 2.  USSPA SYNTHETIC SOIL MATRIX CHEMICAL CHARACTERISTICS
       PARAMETER
UNITS
     VALUE
       Organic Matter

       pH

       Cation Exchange
       Capacity (CEC)

       Base Saturation

                         Ca
                         Mg
                         K
                         H

       Available Phosphorous
meq/lOOg
                         Weak Bray
                         NaHCO-
       Potassium

       Magnesium

       Calcium
ppm




PP"1

ppm

ppm
     1.3

     8

     21.7


     99%

     86.2
     12.4
     1.3
     0



      20
      31

      112

      324

     3740
                                 3S9

-------
  TABLE 3.   USSPA SYNTHETIC SOIL MATRIX X-RAY FLUORESCENCE
             ANALYSIS - MAJOR CONSTITUENTS  (ppm)
PARAMETER
Silica Dioxide (SiO,)
Calcium Oxide (CaO)
Aluminum Oxide (Al-O,)
Magnesium Oxide (MgO}
Magnetite (Fe203)
Potassium Monoxide (K20)
Titanium Dioxide (TiOj)
Sodium Monoxide (Na-Oj
Phosphoric Acid (PgU,-)
Manganese Oxide (MnO;
Barium Oxide (BaO)
Chlorine (Cl)
Sulfur (S)
SAMPLE
37.9
21.5
7.85
4.03
3.64
1.38
0.44
0.34
0.31
0.15
0.03
0.02
<0.05
DUPLICATE
39.6
22.3
8.1
4.25
3.78
1.44
0.45
0.37
0.33
0.15
0.04
0.02
<0.05
COMMON*
RANGE
49-75
1-70
2-57
0.1-1.0
1-79
0-3.6
0.2-1.7
0-1.0
0-1.2
0-4.0
0-0.3
0-0.0009
0-1.0
  * Sources  Based on Lindsay (1979)
  TABLE 4.  USEPA SYNTHETIC SOIL MATRIX X-RAY FLUORESCENCE
            ANALYSIS - TRACE ELEMENTS (ppm)
PARAMETER
Strontium
Zirconium
Chromium
Vanadium
Zinc
Rubidium
Yttri urn
Wolfram
Lead
Nickel
Copper
Tin
Arsenic
Thallium
Molybdenum
Uranium
Niobium
Cobalt
(Sr)
(Zr)
(Cr)
(V)
(Zn)
(Rb)
00
(¥)
(Pb)
(Ni)
(Cu)
(Sn)
(As)
(T.1)
(Mo)
(U)
(Nb)
(Co)
SAMPLE
369
349
131
57
46
41
43
28
22
14
12
<50
<20
<10
<10
<10
<10
<10
DUPLICATE
390
370
135
58
46
46
40
32
25
15
14
<50
<20
12
<10
<10
<10
<10
COMMON*
RANGE
50 - 1000
60 - 2000
1 - 1000
20 - 500
10 - 300
50 - 500
25 - 250
—
2 - 200
5 - 500
2 - 100
2 - 200
1-50
—
0.2 - 5
—
—
1-40
* Source:  Based on Lindsay (1979)
                            370

-------
 TABLE  5.   USEPA SYNTHETIC SOIL MATRIX - X-RAY DIFFRACTION CLAY MINERALOGY
                       MINERAL
SAMPLE
               Quartz
               Calcite
               Dolomite
               Palgioclase  feldspar
               K-feldspar
               Smectite/vermiculite
               Kaolinite
               Mica/illite
               Chlorite
               Polygorski te/at tapulgi te
               "Unidentified"
 27
 27
 18
  8
  8
  5*
 <5*
 <5
              * Estimated concentration.  Quantification not
                possible due  to  low concentration of mineral.
               TABLE 6.  SOIL WASHING DOSAGE/RESPONSE TESTS
Jar No.
1
2
3
4
9
Gasoline
Concentration
(mg/kg)
6,100
30,200
58,300
14,600
86,000
% Vater
21
20
16
20
10
BTEX
Concen t ra t i on
(mg/kg)
400
2,200
4,420
1,280
4,670
TPH
Concentration
(mg/kg)
NT
NT
390
NT
NT
Dose/
Response
Ratio
15
14
13
11
18
Jar No.
5
6
7
8
Diesel
Concentration
(mg/kg)
6,200
16,000
31,000
47,700
% Mater
20
20
20
20
BTEX
Concentration
(mg/kg)
NT
NT
NT
98.3
TPH
Concentration
(mg/kg)
1,950
13,300
33,500
63,500
Dose/
Response
Ratio
3
1
1
I
NOTES:

 NT - Not Tested
                                   371

-------
     TABLE 7.  EFFECT OF TESTED PARAMETERS ON SOIL WASHING EFFICIENCY OF SSH
VA1IED PARAMETER
Contact Time
Soaking Time
Soil/¥ashwater Ratio
Rinsevater/Vashvater Ratio
CitriKLeen
Washwater Temperature
0.13% CitriKleen/Temp.
0.5% Surfactant/Temp.
RANGE
10-30 min
15-30 min
0.5-2
3-10
0-0.7%
55-180°F
55-180°F
75-120°F
EFFECT ON
CONTAMINANT
REMOVAL
0/+
0
+
0
_
0
•f
0/+
EFFECT ON
PARTICLE
SEPARATION
+
0
0
0/+
0
0/+
0
4-
      NOTES:

      + » improvement
      0 « no effect
      - - negative effect
          TABLE 8.  REMOVAL EFFICIENCIES FOR SSM SOIL WASHING EXPERIMENTS
SOIL
CONTAMINANT
Gasoline
Gasoline
Diesel
Diesel
Diesel
Waste Oil
Waste Oil
INITIAL
CONCENTRATION
2,100 rag/kg BTEX
2,100 mg/kg BTEX
18,000 mg/kg TPH
18,000 mg/kg TPH
18,000 mg/kg TPH
1,100 mg/kg TPH
9,300 mg/kg TPH
WASHING
SOLUTION
77°F Water
77°F 0.5%
Surfactant
77 °F Water
77°F 0.67%
CitriKleen
178eF Water
77 °F Water
77 °F Water
PERCENT REMOVAL
#10 SIEVE
99
93
99
99
99
97
96
#60 SIEVE
99
98
98
97
99
98
98
#140 SIEVE
99
94
98
—
99
85
92
NOTE: All above experiments conducted with the following parameters held
      constant:  30 minute contact time, 1:1 soil/vashwater ratio, 3:1
      rinsewater/washwater ratio.
                                      372

-------
                   TABLE 9.  ACTUAL SITE SOILS REMOVALS


                                           PERCENT REMOVAL
SITE LOCATION
Grove City, OH
(gasoline)
Mahvah, NJ
(gasoline)
Princeton, NJ
(home heating fuel)
Holmdel, NJ
(kerosine)
#10 SIEVE
10
67
16
77
82
86
91
97
#60 SIEVE
49
50
80
89
72
0
54
91
#140 SIEVE
49
14
83
49
70
Notes:

(1) Percent removal calculated based on initial bulk soil concentration
    using the following relationship.

    Percent Removal               Bulk soil cone. - washed sieve
    for Sieve Fraction      *                       fractioncone. x
                                           Bulk soil cone.

(2) All experiments performed in duplicate.


              TABLE 10.  PARTICLE SIZE DISTRIBUTION OF SOILS


                                       PERCENT IN EACH FRACTION
SITE LOCATION
SSM
Grove City, OH
Mahvah, NJ
Princeton, NJ
Holmdel, NJ
#10 SIEVE
13
30
26
23
14
#60 SIEVE
31
50
43
12
67
#140 SIEVE
11
4
21
8
12
WASHWATER
AND RINSEVATER
45
16
10
43
7
                                    373

-------
TABLE 11.  SUSPENDED SOLIDS LEVELS IN WASHWATERS AND RINSEWATERS
                                SUSPENDED SOLIDS (mg/L)
SOIL
SSM
Grove City, OH
Mahwah, NJ
Princeton, NJ
Holmdel, NJ
WASHWATER
228,000
33,000
40,600
110,000
40,600
RINSEWATER
6,070
3,100
3,060
6,180
6,820
    NOTE;  Results are for experiments conducted with 30
           minute contact time, 1:1 soil to vashvater
           ratio, 3:1 rinsevater to washwater ratio.
                              374

-------
 Figure  1



     100
   M
   0>
   c
   •H
  4J
  C
  o
  U
  M
  0
      40 -I
      20
       . 001
U.S.  EPA Synthethic Soil Maitrix
  Particle Size  Distribution  Curve
  01      .1       i

   Grain Size  (mm)
Figure  2
        .01
 U.S.  EPA Synthetic Soil  Matrix

  Moisture Retention Curve
  . i       i

Matric Potential
100
                          375

-------
 Figure 3     Dose/Response Curve  for  Gasoline
     5000
  ^  4000 '
  b>
  0
  -"  3000 -
  U



  O  2000 H
  w
  E4
  ffl
     1000 -
              20000  40000  60000  80000  100000



               Gasoline  Added  (mg/kg)
Figure  4    Dose/Response  Curve for Diesel  Fuel
 u
 c
 o
 u


 w
     60000
     50000-
     40000-
     30000 -
20000-
     10000-
               10000   20000   30000   40000   50000


                 Diesel  Added  (mg/kg)



                           376

-------
     Vacuum-Assisted Steam Stripping to Remove Pollutants from Contaminated Soil
                                   A Laboratory Study

                                          by

               Arthur E. Lord, Jr., Donald E, Hullings and Robert M. Koerner
                             Geosynthetk Research Institute
                                   Drexel University
                                Philadelphia, PA 19104

                                    John E. Brugger
                         Risk Reduction Engineering Laboratory
                         U.S. Environmental  Protection Agency
                                   Edison, NJ  08837


                                      ABSTRACT

       A project is underway to evaluate the vacuum-assisted steam stripping procedure for
removal of organic contaminants from soils. The first two years of the study were devoted to
determination of basic parameters such as steam permeability and pollutant removal efficiency.  A
wide variety of soils were used and the "pollutants" were kerosene and a series of simple organic
chemicals of widely varying boiling point and polarity.  These results were reported in the previous
two Symposiums. The vacuum-assisted steam stripping technique appears to have significant
potential for pollutant removal from soils.  A unique composite geosynthetic cap was developed for
the field vacuum-assisted steam stripping application.

       The present paper concentrates on the ability  of steam stripping to remove  high boiling
point materials. Dodecane (b.p. = 216"C) was used with  a 50% sand - 50% silt soil mixture.
Steam  pressures of 3, 5, 10 and 11.5  psi were utilized. The original level of dodecane (5% by
weight) was reduced by two and one half orders of magnitude in 6 hours of steam stripping at 11.5
psi. This is considerably better than we reported at last year's Conference,  at lower steam
pressures.

  Al'so presented -flare data comparing steam stripping, vacuum stripping, air stripping and heat
stripping of kerosene from sand and a 50% sand -  50%  silt soil  mixture  under  a variety of
temperatures. It appears from the laboratory testing that steam stripping is the most thorough soil
decontamination method.
                           INTRODUCTION AND OVERVIEW

       With contaminated soils at Superfund (and other) sites, it is important that the chemicals be
prevented from reaching the groundwater. Fortunately, in many locations the partially saturated or
vadose zone exists and acts as temporary containment retarding the downward movement of the
pollutant. The remediation options are:

       • Excavation and off-site disposal.
       « Excavation and on-site treatment.
       • Insitu treatment (via a number of possible methods, e.g., biological, physical or
         chemical).

                                        377

-------
        A number of these techniques (and others) have been reviewed in recent articles (1,2).
 These in-situ techniques have been discussed by the authors (3).

        The present study falls in the in-situ treatment category wherein the authors propose to have
 pipes inject steam into the soil  beneath the contaminated zone.  Steam stripping of the chemical
 occurs and when aided by a vacuum at the ground surface brings the contaminants to a collection
 point where they can be properly treated.  A unique aspect of the study is the development of a
 geosynthetic cap assembly consisting of a high transmissivity geotextile and a flexible membrane
 liner (geomembrane). The vacuum is applied to the underside of this liner and the contaminated
 gas and/or liquid moves beneath the liner in the geotextile to the outlet ports.  A schematic diagram
 of a proposed system is given in Figure 1 for reference purposes.

        There have been a few steam stripping soil decontamination studies reported in the literature
 (4,5,6,7). These works were field-oriented, sites-of-opportunity projects with no attempt to look
 at the general problem of the feasibility of steam stripping a wide variety of chemicals from a wide
 variety of soils.

        The present work continues a long term study to determine the ability of vacuum-assisted,
 steam stripping to decontaminate general organic chemical species from a variety of soil types.
                          PREVIOUS WORK ON THIS PROJECT

       The work performed in the first phase of the project has been reported in detail elsewhere
(3,8,9,10). Only a brief review of the results of this work will be given here.  Among the tasks
undertaken, in a wide variety of soils were:

       • Observations were made of the transient steam front movements in two dimensional
         flow.

       • The steam permeabilities were determined in conventional one-dimensional flow.

       • The efficiency of steam stripping kerosene and a number of individual organic chemicals
         from soils was determined.  The chemicals included dodecane, decane, octane, octanol
         and butanol.  The Analytical methods  included either volume separation of the outflow
         material (for kerosene) or gas chromatography (GC) of extractions from the soil (for the
         pure compounds).

       • A steady state analytical model was developed where steam flowed upward to the
         collection cap from pipes embedded in the soil.  Use of this model allowed calculation of
         the decontamination time for a given kerosene spill*.

       • A small scale model of the  geosynthetic cap was fabricated and used to determine  its
         feasibility as a cover assembly during steam stripping.  A schematic diagram of the
         experimental setup is shown in Figure 2.

The results from the above topics indicated that the vacuum-assisted steam stripping technique
showed significant promise as a soil decontamination method for a wide range of soil types and
chemicals.
*Certain objections have been raised concerning the model and its use. Currently work is underway in the
development of an alternate model.

                                         378

-------
   Steam
   Generator
. Injection P oes
                                    TI*
   Pipe Manifold System-
                           PLAN VIEW
                             Vacuum Collection

          Flexible Membrane Liner
Vadose
 Zone

                                                       Valves
                       Needled Nonwoven Geotextile
                                               Injection F'ipes
                        Steam  Distillation
                       ELEVATION VIEW
 Figure   1   -  Proposed  In-Situ   Vacuum-Assisted
              Steam  Stripping  Field Apparatus
                           379

-------
Contaminated
Soil
      Clean Soil
„/
                         To Condenser     „
                          and Vacuum      Geosynthet.c Cap
                                      s (FML and Geotextile)
                                          Steam in
Figure  2  - Pilot  Scale  Experiment  for  Vacuum-Assisted
           Stearn Stripping Using the Geosynthetic Cap
                                380

-------
       As a result of the previous work (including the literature search), certain aspects of the
 problem needed to be addressed in further work. These include:

       • Determining the limits of confidence in the soil extraction/gas chromatography analytical
         procedure.

       • Determining  the effect of steam pressure and  temperature on the  steam stripping
         capability.

       * Comparison of the efficiency of steam stripping versus the  other common in-situ
         techniques, i.e., air stripping, vacuum extraction and heat alone.

       • Determining if the rate of chemical removal changes subsequent to a long delay period
         following partial chemical removal by steam stripping (or other decontamination
         methods).
                                   PRESENT WORK

Accuracy of Chemical Analysis

       As discussed at the last Symposium, there was concern as to the accuracy of the soil
extraction/GC method used in the chemical analysis.  Therefore work was undertaken in this
regard.  As a first step, the GC extraction fluid, ethyl ether was "doped" to given values of three
common organic compounds. The compounds were dodecane, toluene and o-xylene. The levels
of doping were 5 |il, 20 (il and 100 ^.1 in 50 ml of the ethyl ether. The samples were given to our
Chemistry Department co-workers, the amounts of the dopants being unknown to them.  Standard
reduction in volume techniques were used on the extracts to concentrate the chemical, and then GC
analysis was performed.  With very few exceptions (tor 24 of the 27 samples provided) the GC
analysis gave values which were within 20% of the actual doped levels.

       As a second step  (and more difficult assessment) dodecane was introduced into a 50%
sand/50% silt soil sample. The levels of doping were 10, 40 and 100 fAand the soil weight was
50 grams,  hese correspond to 150, 600 and 1500 ppm (by weight) of dodecane in the soil. The
dodecane was mixed as thoroughly as possible into the soil, allowed to sit for 20 hours and then
extracted with the ethyl ether.  Three-fold extraction was used with vigorous hand-mixing during
each extraction. The extracted fluid was reduced in volume and then GC analysis was performed.
The results are shown in Figure 3,  It is seen that poor agreement exists between the GC results
and the doped amounts at the lower levels. This could be due to certain causes:

       •  a certain amount of dodecane becomes attached to :he finer soil particles and thus resists
         extraction

       •  large percentage loss of these very small volumes of dodecane during extraction
         (evaporation, attachment to container walls, stirring rods, etc.).   It was  however
         attempted to minimize all these effects

      The results of Figure 3 may not be completely pertinent to the general question of soil
analysis accuracy (due to some of the problems raised above), but they should certainly be folded
into any final conclusions. It is interesting to note that an average of 35 J4.1 is missing in both the
100 \il and 40 fil doped samples. This may be the amount that is attached to the fine-grain silt
particles. There has  been work done by soil chemists in "spiking"  soils with various priority

                                         381

-------
o
V)

D>

O
0)
>
CD
"O

-------
 pollutants to determine the recovery rate on subsequent extraction for GC analysis (11). Recovery
 rates for the various compounds are variable (some few being non-recoverable), but many are in
 the region of 70-80%,  Unfortunately, no analysis of the soil (grain size distribution, etc.) is given
 (11).

 .Effect of Steam Pressure (and Temperature) on Steam Stripping Efficiency

       In the previously reported work (3,8,9,10) the steam pressure available was limited to the
 house steam and was nominally about 5  psi gauge (with some variations and even inadvertent
 cessations.) A steam generator was therefore purchased allowing well-controlled steam pressures
 from 3 to 30 psi to be available. (A pressure of 11.5 psi was the highest that could be achieved
 because of the large volume of steam which was flowing.) Figure 4 is a schematic diagram of the
 vessel constructed *o allow  a determination of the  effect of steam  pressure (and hence  also
 temperature) on the steam stripping ability of given chemicals from a soil sample. Steam passes
 through the outside dual-wall of the container in order to keep the walls hot and lessens excessive
 steam condensation.

       Steam passes through  the soil sample (6 inch diameter x 2 inches  high) which is contained
 on the top by a porous plate which precludes the soil from being blown out of its position by the
 stream pressure.   The soil temperature  and steam pressures are monitored as per Figure 4.
 Dodecane (at 5 wgt %) was used as the chemical ana the soil was a 50% sand/50% silt mixture.*
 The soil was analyzed for residual dodecane after various steam stripping scenarios. The results of
 this work are shown in Figure 5. Here is plotted the amount of dodecane remaining in the soil as a
 function of time at steam pressures of 3, 5, 10 and 11.5 psi.  (The measured temperature of the soil
 was essentially independent of steam pressure and remained very near to 100°C).

       It is seen in Figure 5 that while the initial removal rates seem to be relatively independent of
 steam pressuref, the higher steam pressures reduce the residual dodecane level much lower at the
 longer times. (We, of course, are assuming here that these relatively large differences are real, and
 are not grossly affected by the problems shown with  the soil extraction/GC analysis method
 discussed earlier). The level of 150 ppm reached at 11.5 psi in 6 hours  is considerably lower  than
 the value of 1000 ppm reported in last year Proceedings (while using ~5 psi steam in a different
 cell).

 Efficiency of Steam Stripping and Other Decontamination Methods

       Figure 6 shows the efficiency of kerosene removal in both sand and the 50% sand/50% silt
 mixture, in the aforementioned small laboratory samples. The steam pressure used here was about
 5 psi and the vacuum was  10  inches Hg for sand and 20 inches Hg for the 50/50 mixture.  In the
 case of sand the kerosene is  completely removed (as far as volumetric separation analysis tells) in a
 very short time. For the 50/50 mixture the volumetric separation analysis  indicates that some 30%
 of the kerosene is remaining after 700 minutes of steam  stripping. However the output fluid  was
 colored indicating, some kerosene content. Carbon Oxygen Demand (COD) analysis of the output
 fluid as a function of time (9) allowed a more reliable curve of kerosene  removal to be ascertained.
The results of this correction process is shown in Figure 7. (It should be emphasized that the COD
correction procedure is not  exact, but it gives a much clearer picture than that represented by the
 unconnected curve.)

       In order to compare steam stripping with other  decontamination techniques, laboratory
experiments were  also performed with vacuum, air and heat stripping of kerosene from sand and
*The properties of this soil can be found in reference 3.
fTne initial removal may simply be a physical "pushing out" of the chemical from the soil.

                                         383

-------
             Thermocouples
Steam out
Steam
Input
l_
m—

a

r1


t 	
Steam
Gage


•

r1 \

1 r

n
--*--•- j ^ *--*- --1
• S S f S S f
s \ % \x\ v
's's-Vs's's'
'/'I ^"^
L
1


~
—
J




L
fc*:%3:%'ft%^|
{; > soil :%%%%:
** *^ ' ^
L 11 11 II t L 11
I II i 1 1 I 1 1 I



1 M ! 1 ! !

I

11,11 	 1 1 n 	 j.i 	 LI,.
j 1 1 1 1 1 1 1 1 1 1 1 1 	
s.

I ,
. (x)


-L.
1 __J











-r. Stea
0"0 Rele
Valv
Jj— N -J
Steam
Gage
Water
VV ale I
Drain

                                Valve
     Figure 4  -  Schematic Diagram  of  Device to Steam
             Strip Soil Sample at Various Pressures
                            314

-------
  120
                i          1         i
             initial concentration is 5% or 50,000 ppm
§ 80
CD
_J

0)
o
03
TO
O
60-
40 H
20-
    0
                                                             1800
                                                                •1500
                                                            •1200
                                                                h900
                                                                -600
                                                                "300
                                                               0
      0
            2468

                          Time (hours)


           Figure 5(a) - Steam Stripping GC Results
10
12
  120
                                                             1800
             initial concentration is 5% or
                                                                  0
              50
                  100     150     200    250

                Pore Volumes of Condensed Steam
 300     350
            Figure 5(b) - Steam Stripping GC Results

                           385

-------
                           50/50 unco erected)
      0    100   200   300   400    500   600    700


                          Time (minutes)



         Figure 6 - Kerosene Removal via Steam Stripping
  100
cc
>
o

e
Q)
oc
V)
o

<5
800
                Corrected via COD measurements
      0    100   200   300   400    500   600   700   800

                          Time (minutes)



  Figure 7 - Kerosene Removal via Steam Stripping (COD corrected)



                        386

-------
the 50/50 soil. In these cases, the soil was dried and outgassed at 500°C for two hours to ensure
stability. The soil sample was then placed in the small laboratory cells (2.5-inch diameter x 6-
inches high) and "doped" to the given level of kerosene (about 20%). The cells were then placed
in a furnace at a given temperature (see Figure 8) and either (a) a vacuum was applied to the top of
the sample (with the bottom inlet open) or (b) compressed air was injected into the top (and hence
out the bottom) or (c) the sample was left open at both ends and simply heated by the furnace. The
samples were dry, except for the kerosene, when the experiments started, so the kerosene removal
could be determined by sample weight loss.*

      Figure 9 shows the results for the vacuum stripping experiments for sand and the 50/50
soil.  Figure  10 shows the results for the air stripping experiments.  Note the vacuum levels and air
pressures used as  given in Figures 10 and 11.  Figure 11 shows the results for kerosene removal
with only temperature elevation being  used (T = 100°C).  In this case no vacuum or air was
applied. In  all cases the average of two runs  was  used and plotted.  If the two runs hud quite
different results,  a third run  was performed  for agreement and then the two consistent runs
averaged.

       It is seen from Figures  9 and 10 that the temperature of the soil is a most important variable
in decontamination. Especially in the case of the 50/50 soil, n temperature of 100°C drastically
raises the removal rate over that at ambient  temperatures.  However heat alone is insufficient to
rapidJy decontaminate the soil as is  shown in  Figure 11. There must be a forced air movement
(either by vacuum or direct air  injection) to allow the heat to act efficiently. That is, natural air flow
is insufficient.

       Figure 12 compares the steam stripping (at 5 psi steam pressure) and ambient temperature
vacuum and air stripping techniques. It is seen that the steam stripping is  clearly superior in
removing the kerosene.  Figure  13 compares the steam stripping and 10()°C vacuum and air
stripping techniques.  The initial rate of kerosene removal is greater for the air and vacuum
processes**, but the ultimate removal is belter for steam  stripping. The lower initial removal rate
in the case of steam may be that Initially  the soil must be heated by the steam (via condensation) to
reach 100°C. In the air and vacuum stripping experiments the furnace is at 100°C when the sample
is put in and may reach 100°C faster than in the case of steam stripping experiments.

       It is very important to  note here that the vacuum used in the field will be very much lower
than those used in these laboratory experiments.  Hence  a very biased comparison is being made
here in favor of the vacuum technique. Also in the case of steam and air stripping it will be most
difficult to  heat  separately  the  large  soil masses  encountered in the  field to the elevated
temperatures, a fact which also biases the comparison in favor of vacuum and air stripping.

Effect of Delay Times in Steam Stripping

       If a dynamic decontamination method (such as steam stripping) is used for a certain time
(not long enough to be essentially complete) and then stopped for a given "delay" period, what will
be the consequences?  If some of the less mobile contaminant can move to more mobile environs
during the delay, is it possible that the new  rate of removal will exceed that  which was present
immediately before the delay? An attempt  has been made to look into this possible effect with
controlled laboratory experiments.
*lt is important to note that the weight change method cannot be used for steam stripping, for the
water introduced adds extra weight to the sample.  The output kerosene must be collected and
monitored by volumetric separation.
**Again this may reflect an initial "pushing out" of the pollutant.

                                        387

-------
FURNACE
                              VACUUM (air out)
                        x^x~ AIR (air in)
                      SOIL
                      AND
                    KEROSENE
                              OPEN
   Figure 8 - Schematic  Diagram of  Apparatus
       Used  for Vacuum,  Air and  Heat Stripping
                         388

-------
                    Vacuum equals 10" Hg
                        I     !     1
100
        50  100  150 200 250  300  350  400  450  500

                       Time (minutes)


       Figure 9{a) - Vacuum Stripping of 100% Sand
                          Vacuum equals 20" Hg
   0     100   200   300   400   500   600   700   800

                       Time (minutes)


       Figure 9(b) - Vacuum Stripping Of 50/50 Soil

                     389

-------
  100
                       Input air pressure equals 3 psi;
                        Air flow equals 2.5 liters/min
      0    50   100  150  200  250  300  350  400  450  500

                           Time (minutes)

            Figure 10(a) - Air Stripping of 100% Sand
  100
T3
CD

O


0>
DC

CD
C
CD
«3
g

CD
                              Input air pressure equals 3 psi
                              Air flow equals 2.5 liters/min
100    200
              300   400    500

                Time (minutes)

Figure 10(b) - Air Stripping of 50/50 Soil

                390
                                              600   700   800

-------
O
E
0)
cc
c/)
O
                100
200
300
400
500
600
700
800
                                       Time (minutes)
              Figure 11 - "Heat Stripping/Kerosene from 100% Sand and 50/50 Soil

-------
  100
   90


   80


   70
o


o>
DC

0)
c
03
W
O

o5
     0     100   200   300   400   500   600   700   800

                         Time (minutes)

     Fig 12 - Comparison of Decontamination Processes

            (Vacuum and Air at 23° C)
  100
     0     100   200   300    400   500   600   700   800

                         Time (minutes)

    Figure 13 - Comparison of Decontamination Processes

           (Vacuum and Air at 100° C)

                         382

-------
       Small cells, of the kind used in earlier reported work (3,8,9,10), were used. The soil was
a 50% soil/50% silt mixture in which was mixed 20% kerosene. The material was vacuum-
assisted steam stripped at a steam pressure of 5 psi and a vacuum of 20 inches of mercury for a
period long enough to remove roughly 50% of the kerosene, at which time the rate of removal of
kerosene had slowed considerably from the initial value. A number of samples were so prepared
and then sealed and stored for subsequent use. Delay times of 15 and 50 days were used.

       Unfortunately, for both the 15-day delay and the 50-day delay, the initial steam condensate
upon resuming treatment was very brown-colored and non-separable, so the simple volumetric
analysis cannot be performed. Some of the kerosene must have gone into the residual water during
the delay.  (No COD analysis was performed here.)
                            SUMMARY AND CONCLUSIONS

       This paper presents the most recent results obtained in a laboratory project to ascertain the
general viability of utilizing vacuum-assisted steam stripping to remove organic chemicals from a
wide variety of soils. Our previous laboratory work had investigated steam flow in soils and the
efficiency of removal of certain organic chemicals and mixtures from soils.  Also a small scale
demonstration of our proposed field set-up, which includes a unique easy-to-use geosynthetic cap
assembly was performed.

       The main thrust of the present work investigated the effect of differing steam pressures on
the removal efficiency of a high boiling point organic compound from a soil with a permeability
typical of many soils encountered in the field.  To this end a new experimental apparatus had to be
constructed. The results indicate that significantly larger amount of dodecane (b.p. 216°C) can be
removed at 10 and 11.5 psi steam pressure than at either 3 or 5 psi. This has direct application to
in-situ decontamination processes, if 10 psi can indeed be maintained at the subsurface steam entry
point.  The  results may also have strong implications for vacuum-assisted  steam stripping of
excavated soils, where much, much higher steam pressures (and temperatures) could be maintained
in a closed treatment vessel.

       It should be mentioned that gas chromatography was the analytical method to determine the
amount of chemical remaining in the soil. "Standard" methods of extracting the chemical from the
treated soil  were employed.  There  were  indications from control work performed to find the
accuracy of the method, that a significant amount of dodecane is bound to the silt particles.

       Work was completed to ascertain the  effect of sample temperature on the efficiency of
vacuum-stripping and air  stripping kerosene from sand and a 50% sand/50% silt soil mixture.  It
was found  that raising the temperature of the soil sample to 100°C drastically enhanced the
decontamination ability of both techniques. (The vacuum used here is of a much larger value than
that used in  the field.)  The two methods were then somewhat competitive with steam stripping
although steam stripping does ultimately remove more kerosene. Steam stripping is vastly superior
to ambient temperature vacuum stripping and air stripping.  Heat alone does not act as a  good
decontamination process — there must be some forced air movement to reach good removal rates.
It, of course, will be extremely difficult to heat the soil in the field to these elevated temperatures
for vacuum or air stripping.

       A preliminary study was performed to ascertain if stopping the steam stripping for a given
delay time would enhance the subsequent start up removal rate.  Unfortunately, after delays of 15
and 50 days the steam condensate was very colored and could not be volume separated.

                                         393

-------
       In general it appears that vacuum-assisted steam stripping is a very attractive method with
 which to decontaminate soils, and it works well on the hard-to-remove high boiling point organics.


                             SUGGESTED FUTURE WORK

       It is important to use a clay and an organic fraction in the soil in future work. These will
 probably bind the organic chemicals and hence make them possibly more difficult to remove with
 the steam stripping, or any other process.

      More high boiling point organic compounds, with varying properties, such as polarity,
 should be investigated. Vacuum and air stripping should be performed at lower vacuum levels and
 air pressures (much lower than those used herein), in order to more realistically compare steam
 stripping and these other decontamination techniques.

       The analytical technique, used to determine the residual chemical remaining in the treated
 soil, should be investigated. A very thorough literature search should be undertaken in this regard.
 It may be necessary to work directly with soil chemists.  (A different extraction method may be
 indicated.)

       Delay time work should proceed with vacuum and air stripping.

       It would seem appropriate at this time to "scale-up" the laboratory model  field unit
 simulation work to ascertain some of the inevitable problems that always arise in applying a new
 technique. No prohibitive problems are foreseen due to its  very good performance at the small
 laboratory size work, but "scale-up" is definitely warranted at this time.


                                ACKNOWLEDGEMENT

       The Drexel authors wish to thank the Risk Reduction Engineering Laboratory of the U.S.
 EPA of Edison, New Jersey for their generous financial support of this work through Cooperative
 Agreement No. CR-ri 13022.


                                     REFERENCES

 1. Kovalic,  J.  M.  and Klucsik, J. E, "Loathing for Landfills Sets  Stage for Innovative
   Hazardous Waste Treatment Technology," Hazard. Mat. and Waste Manag. 5_, 1987, pp. 17-
    18.

 2. Cheremisinoff, P. N., "Update: Hazardous Waste Treatment," Pollut. Eng.  !£, Feb. 1987,
   pp. 42-49.

 3. Lord, A. E., Jr., Koerner, R. M. and Murphy, V. P., "Laboratory Studies of Vacuum-Assisted
   Steam Stripping of Organic Contaminants from Soil," Proc. 14th Annual Conference on Land
   Disposal.  Remedial Action and Treatment of Hazardous Waste. Cincinnati, Ohio, April 1988,
   sponsored by the Risk Reduction Engineering Laboratory, U.S. Environmental Protection
   Agency, Cincinnati, Ohio, pp. 65-92.

4. Hilberts, B. (1985), "In-Situ Steam Stripping," Assink, J. W. and Van Den Brink,  W. J.,
   Eds., Contaminated Soil. Proc.  of 1st Intern. TNO Conf. on Contaminated Soil, Utrecht, The
   Netherlands, Nov. 11-15, 1985, pp. 680-687.

                                        394

-------
5.  Baker, R., Steinke, J., Manchak, E, Jr. and Ghassemi, M., "In-Situ Treatment for Site
    Remediation," Proc. Third Annual Conference on Hazardous Waste Law and Management.
    Seattle, Washington, October 27  and 30, 1986, and Portland, Oregon, October 31  and
    November 1, 1986.

6.  Baum, R., short article describing process appearing in Chemical and Engineering News,
    December 12, 1988. Work done by K. Udell, J. Hunt and N. Sitar at University of California,
    Berkeley and A. Nagtel of Solvent Services, Inc., San Jose, CA.

7.  "In-Situ/Hot Air Soil Stripping," SITE (Superfund Innovative  Technology Evaluation)
    Demonstration Results,  U.S. Environmental Protection Agency,  EPA/540/M5-90/003,
    February 1990.  Paul De Percin, EPA Project Manager.

8.  Lord, A.  E., Jr., Koerner,  R. M., Murphy, V. P.  and Brugger,  J. E., "In-Situ, Vacuum-
    Assisted,  Steam Stripping of Contaminants from Soil," Proc. of Superfund '87. Rth National
    Conference on Management of Uncontrolled Hazardous Waste Sites. November  1987,
    Washington, DC, pp. 390-395.  Sponsored by the  Hazardous Materials Control Research
    Institute, Silver Spring, MD.


9.  Murphy,   Vincent P.,  "In-Situ, Vacuum-Assisted, Steam-Stripping to Remove Volatile
    Pollutants from Contaminated Soils," Masters Thesis in Civil Engineering, Drexel University,
    Philadelphia, PA, June 1988.

10. Lord, A. E., Jr., Koerner, R. M., Hullings, D.  E. and Brugger, J. E., "Laboratory Studies of
    Vacuum-Assisted Steam Stripping of Organic  Contaminants from Soil," Proc. 15th Annual
    Conf. on Land  Disposal. Remedial Action and Treatment of Hazardous Waste. Cincinnati,
    Ohio, April 1989, sponsored by Risk Reduction Engineering Laboratory, U.S. Environmental
    Protection  Agency, Cincinnati, Ohio.

11. Kiang, P.  H. and Grob, R. L., "Development of a Screening Method for Determination of 49
    Priority Pollutants in Soil," J. Environ. Sci. and Health A21. 15-53 (1986).
                                     395

-------
     EVALUATION OF LIMITED-USE CHEMICAL PROTECTIVE CLOTHING
                     USED FOR EPA SUPERFUND ACTIVITIES
                                         by

                   Jack C. Sawicki, Carla Mond and Arthur D. Schwope
                                 Arthur D. Little, Inc.
                                     Acorn Park
                             Cambridge, MA  02140-2390

                                    Susan Watkins
                    Department of Design and  Environmental Science
                                  Cornell University
                                  Ithaca, NY  14853

                                  Michael Gruenfeld
                               Releases Control Branch
                         Risk Reduction Engineering Laboratory
                            Edison, New Jersey 08837-3679
                                     ABSTRACT

        Limited- and single-use garments are used widely by EPA and EPA contractor
personnel for protection from potentially hazardous liquids and solids at Superfund sites.
Such garments are used in a wide variety of tasks and many hundreds of thousands of these
garments are purchased each year.  They are available from hundreds of distributors, in many
different styles and configurations, and with significant differences in cost and quality.  This
study was undertaken as part of an overall effort to develop guidelines to ensure that the most
appropriate product will be procured by those responsible for worker safety. Of particular
concern were garment sizing and fabric strength and stiffness since field personnel frequently
report that garments don't fit and that  they tear, especially in the crotch  and underarm  areas
and across the upper back.

        Essentially all commercial fabrics used for limited-use chemical protective clothing
(CPC) and two experimental fabrics were subjected to tests to measure breaking strength and
flexibility.  Based on these tests and the  observations from the field, minimum values for
breaking strengths of coated and uncoated fabrics, and maximum values for stiffness were
deduced.

        Several standards and specifications describing size and fit parameters for limited-use
CPC were collected and reviewed.  The two most significant conclusions were that (1)
ANSI/ISEA Specification  101-1985 by itself was inadequate for EPA applications and  (2) to
be useful a sizing specification should relate garment size to body dimensions.  From the

                                        386

-------
analysis, a sizing system was developed and used as a basis for a working version of the
procurement guidelines and for recommendations to ANSI/ISEA and ASTM for modification
of existing sizing standards.

        This paper has been reviewed in accordance with the U.S. Environmental Protection
Agency's peer and administrative review policies and approved for presentation and
publication.

                                  INTRODUCTION

        The Environmental Protection Agency (EPA) has primary Federal regulatory
responsibility for hazardous substance and hazardous waste (HS/HW) release, prevention, and
clean-up operations in the United States.  To address the risk to clean-up workers from
HS/HW, the Occupational Safety and Health Administration (OSHA) promulgated 29 CFR
1910.120, Hazardous Waste Operations and Emergency Response, in 1988.  This regulation
requires personal protective equipment (PPE) to be provided for HS/HW activities.

        PPE includes gloves, boots, garments, respirators, and so forth. These components
are combined to provide four, progressively lower degrees of protection: Levels A, B, C, and
D.  Limited- or single-use garments are most commonly used as a part of Level C and,
sometimes, Level B protection.  The advantage of limited- or single-use clothing is that the
costs and uncertainties of garment decontamination are obviated. Because they are usually
disposed of after one or two uses, limited-use garments  must be relatively inexpensive. Many
cost less than $20; although in some special cases, garments costing $70-100 are disposed
daily.  Several hundred thousand, limited-use garments are used annually by the EPA and its
Superfund contractors.

        Typically the garment materials are composed of a woven or nonwoven fabric
substrate onto which is laminated or coated a continuous film of plastic to provide a barrier to
liquids and vapors.  Nonwoven fabrics without the film  are sometimes used when the HS/HW
is a solid.  The material is  stitched,  adhesively bonded, or heat-welded together to produce a
garment. The garment may or may not have pockets, attached hood and boots, elastic bands
at the wrists and ankles, and so forth. The garments are worn over street clothing, which can
range from shorts and a T-shirt in hot climates to bulky jackets and heavy trousers in cold
climates. At present the most commonly used limited-use garments are based on
Tyvek® 1422 spunbonded  polyolefin, a porous, nonwoven fabric (Du Pont)1.  Because it is
porous, the Tyvek 1422 is laminated or coated with polyethylene film or Saranex 23-P® film
(Dow).  Recently, however, other products have been introduced with claims of greater
durability or better resistance to chemicals. These include Barricade® (DuPont), Responder®
(Lifeguard), and Chemrel Max® (Chemron).
           1 Mention of tradenames or commercial products does not constitute endorsement
        or recommendation for use.

                                        397

-------
                                      PURPOSE

        This study was undertaken as part of an overall effort to develop guidelines to ensure
that the most appropriate limited-use garment can be procured by those responsible for worker
safety.  Of particular concern were garment  sizing and fabric strength and stiffness since field
personnel frequently report that garments don't fit and that they tear, especially in the crotch
and underarm areas and across the upper back.

                          DESCRIPTION OF THE PROBLEM

        EPA and EPA contractor personnel  report limited-use CPC tears, especially coveralls
of Tyvek 1422.  Coveralls constructed of Tyvek 1422 coated with polyethylene (PE) or
laminated to Saranex 23-P are more durable, but some failures occur.  Kleenguard®
(Kimberly-Clark) garments are reported by some field personnel  to be  less prone to tearing
than Tyvek 1422, but the fabric is not available in coated forms.  Tearing most frequently
occurs in the crotch and underarm areas and across the upper back. Tearing may be due  to
insufficient fabric or seam strength or incorrect sizing.

        Workers also report  that  garments are tight in the hood, chest, back, seat, armholes,
and thighs, especially when worn over cold  weather clothing.  Tightness may be due to
incorrect sizing or that the worker was not wearing the correct size garment.  Many field
workers wear oversize garments and "field tailor" them with duct tape.  Women and large
men have the worst fit problems.

        Workers reported  that many of the newer fabrics, while apparently stronger and
having better chemical resistance, are stiffer and,  consequently, less comfortable to wear than
garments based on Tyvek  1422 or Tyvek 1422/PE or Tyvek 1422/Saranex 23-P.

        Field personnel reported  that cost is a very significant factor in the selection of
fabrics and garments.  It appears  that there is a price point between $20 and $40 above which
disposal after use is prohibitive and discouraged in actual practice.

                 PHYSICAL PROPERTIES OF LIMITED-USE FABRICS

PROCEDURES

        To gain insights into the strength and flexibility characteristics of the fabrics of
limited-use garments, specimens of seventeen fabrics were subjected to three standard tests.

        ASTM D1682 - Breaking Load and Elongation of Textile Fabrics. In this test, a
        continually increasing load is applied longitudinally to the fabric specimen.  The  load
        at which the fabric breaks is the test result.

        ASTM D1388 - Stiffness of Fabrics. In this test, a strip of fabric is slid in a
        direction parallel to its long dimension, so that its end projects from the edge of a

                                         398

-------
        horizontal surface.  The length of fabric overhanging the edge (divided by two) when
        the end of the fabric is at a specified angle to horizontal is the test result.

        ASTM D3776 - Mass Per Unit Area (Weight) of Woven Fabric.  In this test, the
        weight of a specified area of fabric is measured.

        For ASTM D1682 and ASTM D1388, tests were performed with the required number
of specimens  from two, perpendicular directions on the fabric. The results averaged since
machine and cross directions could not be distinguished for all fabrics.

RESULTS

        Results are presented in Table 1.

        Tyvek 1422 had the lowest breaking strength of any fabric tested, 3.5 pounds,
although the value is not significantly different from those for Kleenguard (9.7) and Enhance
(8.8). The most commonly used coated fabric, Tyvek 1422/PE, had a breaking strength of
13.3 pounds;  Tyvek 1422/Saranex 23-P was essentially the same at 13.6 pounds.

        The experimental fabrics, Tyvek 1443/PE  and Tyvek 1445/PE, showed breaking
strengths of 12.3 and 14.9 pounds, respectively.

        The uncoated fabrics generally exhibited the highest flexibility, with bending lengths
ranging from 1.2 to 1.9 inches (lower bending lengths indicate greater flexibility).  The
coated,  nonwoven fabrics were generally stiffer, with bending lengths ranging from 1.5 to
4.4 inches. Three of the coated,  nonwoven fabrics had curls that made them inappropriate for
the test.

        The coated woven fabrics, Chemtex, Chemgard®, and Neonyl®, were the strongest
and had bending lengths that spanned 0.9  to 2.8 inches.  The coated woven fabrics, however,
have much higher weights and garments fabricated from them are probably too expensive for
use as a day-in, day-out, disposable garment.

DISCUSSION

        The fabrics can be divided into two stiffness categories: bending lengths less than or
greater than 2 inches.  One could infer from subjective comments from  field personnel  that
acceptable garments should be constructed from fabrics having a bending length less than
2 inches.  This is supported by comments from the field that Tyvek  1422/Saranex 23-P
garments are  the stiffest fabrics that are considered acceptable.  Also based on comments
from the field, the minimum acceptable breaking strength may be in the  range of
9-14 pounds.   Information gathered from the field  in this  study would suggest that the
strength of Tyvek 1422 is less than acceptable, that of Tyvek 1422/PE is marginal, and that of
Kleenguard and Tyvek/Saranex 23-P  are acceptable.

                                        399

-------
      TABLE 1.  PHYSICAL CHARACTERISTICS OF LIMITED-USE
           CHEMICAL PROTECTIVE CLOTHING FABRICS
Fabric (Source)
Tyvek 1422 (Du Pont)
Kleenguard (IKimberly-Clark)
Enhance (Abandaco)
Tyvek 1422/PE (Du Pont)
Tyvek 1443/PE (Du Pont)
Tyvek 1445/PE (Du Pont)
Encase H (Abandaco)
PP/Saranex 23-P (Abandaco)
Tyvek 1422/Saranex 23-P
(Abandaco)
Chemtuff (Chemron)
Chemrel (Chemron)
Chemrel Max (Chemron)
Barricade (Du Pont)
Responder (Lifeguard)
Chemtex (Bata-Sijal)
Chcmgard (Rainfair)
Neonyl (Rainfair)
Type*
NW
NW
NW
CNW
CNW
CNW
CNW
CNW
CNW
CNW
CNW
CNW
CNW
CNW
CW
CW
CW
Weight*
1.24 (0.15)"*"+
1.85 (0.1)
1.88 (0.2)
2.18 (0.15)
2.14 (0.2)
2.23 (0.25)
2.33 (0.3)
3.53 (0.25)
3.60 (0.2)
3.60 (0.35)
4.08 (0.45)
5.19 (0.6)
6.39 (0.45)
8.23 (0.9)
9.52 (1.2)
10.40 (1.1)
15.19 (0.6)
Breaking
strength
8.5 (1.7)
9.8 (1.3)
8.8 (3.6)
13.3 (1.6)
12.3 (0.3)
14.9 (5.0)
12.7 (2.7)
13.1 (3.4)
13.6 (1.4)
46.4 (6.2)
21.3 (6.6)
38.9 (7.4)
28.2 (6.0)
38.1 (5.7)
101.6 (12.4)
65.6 (8.1)
117.6 (17.0)
Bending
length"1"
1.2 (0.6)
1.6 (0.2)
1.9 (0.6)
1.8 (0.7)
1.8 (0.6)
2.1 (1.1)
1.5 (0.6)
1.9 (0.3)
1.7 (0.6)
NA1"
NA
NA
2.4 (1.2)
4.4 (0.7)
0.9 (0.1)
1.6 (0.6)
2.8 (0.5)
ASTM D 3776, 'option C; ounces/square yard.
ASTM D 1682, 5 inch/minute constant-rate-of-extension; pounds
* ASTM D 1388, option A; inches
"*"«" INJtimfti*** 4t% r*at¥*f*l'ttiP»e^c Ic cf-an/lsifYl H^visttirkn
_	   —   f
Test not applicable; fabric had permanent curl.
NW = nonwoven; CNW = coated or laminated nonwoven; CW = coated woven

                            400

-------
                   SIZING AND FIT OF LIMITED-USE GARMENTS

        ANSIflSEA2, Standard 101-1985, Men's Limited-Use and Disposable Protective
Coveralls-Size and Labeling Requirements, specifies minimum dimensions for five sizes of
coverall. See Table 2.  Compliance with the specification is voluntary.
             TABLE 2.  ANSIflSEA 101-1985 MINIMUM REQUIREMENTS
Size*
Small
Medium
Large
X-Large
XX-Large
Chest
21-1/2
23-1/2
25-1/2
27-1/2
29-1/2
Leg
inseam
27-1/2
28
29
29-1/2
30
Sleeve
outseam
31-1/2
32-1/2
33-1/2
35
36-1/2
Body
length
35
36
37
38-1/2
39
Sleeve
opening
6-1/2
7
7
7
7
Leg
opening
9-1/2
10
10
10
10
Front
opening
length
29-1/2
29-1/2
30
30-1/2
31
     All dimensions in inches.
        In the summer of 1987, two types of coveralls were obtained from five manufacturers
of limited-use coveralls  and measured according to the instructions of ANSI 101-1985. The
results are shown in Table 3.  None of the Tyvek 1422 garments and two the Tyvek
1422/Saranex 23-P completely met all the dimensions of the ANSI/ISEA standard.  In a few
cases, the garments would appear to be  significantly below the minimum dimensions;
however, the deficiencies could be  considered minor in most cases.  From 1987 to present,
the major manufacturers were increasingly claiming adherence to the ANSI/fSEA
specification.  Failures due to tearing, however, continued to be reported by EPA and EPA
field personnel. This observation suggests that, while manufacturers may have been meeting
the minimum dimensions of the specification, the specification may be inadequate.
           2ANSI - American National Standards Institute; ISEA - Industrial Safety
        Equipment Association.

                                       401

-------
                  TABLE 3. MEASUREMENTS OF LIMITED-USE COVERALLS
Dimension
Chest
Leg inseam
Sleeve outseam
Body length
Sleeve opening
Leg opening
Front opening

Hood width

Hood length
Minimum
ANSI/ISEA
101-1985
25.5*
29.0
33.5
37.0
7.0
10.0
30.0
**

*#

Tyvek 1422
Manufacturer
1
26.2
29.0
34.5
36.1
7.5
11.0
29.8

9.6

14.7
2
28.8
30.2
34.0
37.5
6.2
9.0
28.0

12.75

14.0
3
25.2
27.0
3.5
40.2
7.2
10.2
29.8

11.0

15.3
4
26.8
29.2
33.0
40.2
8.2
10.8
30.8

9.0

15.0
5
27.0
29.2
36.0
35.0
6.8
9.5
28.0

11.0

14.75
1
26.4
26.5
33.4
40.7
7.9
11.0
28.9

11.0

15.2
Tyvek 1422/Saranex 23-P
Manufacturer
2
26.0
28.5
33.8
36.8
7.0
11.0
30.0

11.0

15.5
3
29.2
30.2
36.2
39.0
7.8
11.5
30.2

11.75

14.5
4
28.0
27.0
37.0
43.0
7.2
11.5
26.5

10.0

15.5
5
27.0
28.5
40.0
38.8
6.2
11.5
29.5

9.8

15.5
All measurements in inches.
Not a requirement.

-------
        Consequently, in order to meet the objective of the study (i.e., guidelines that would
aid the process of procuring acceptable garments), attention was directed towards finding a
basis for recommending modifications to the ANSI/ISEA specification.  The effort was
focus sed on other directly related or pertinent standards.  Three were identified:

        MIL-C-29133A  --   Coveralls, Disposable, General Purpose
        MIL-C-87069A  --   Coveralls, Disposable
        BS 5426:1987   --   Workwear and Career Wear

        The British Standard (BS) was unique in that it related garment sizes to body
dimensions. While such an approach is common for commercial clothing (e.g., consider the
size/body dimension charts in mail order  clothing catalogs), it has not been practiced in the
area of chemical protective clothing. The British Standard also contained more sizes as
means for increasing the likelihood of having garments that fit a  wider range of the
population.

        Other observations and conclusions based on the review  of the standards and the
technical literature and the experience of the authors are:

        •      The current range of 5 sizes should be expanded  to 6 sizes to provide a better
               fit range, especially for women and large  males.  Dimensions  for these 6 sizes
               are proposed in Table 4.

        •      The field data suggests that insufficient back body length is included in
               present coverall designs.  Tearing under the arm may be related to inadequate
               back width and armhole width.  Additional back length and ease for the
               armhole is required to fit over winter clothing.

        «      Additional sleeve outseam and leg inseam length  will increase the range of fit
               for tall workers.

        •      A range of sleeve openings will keep excess bulk from  hindering smaller sizes
               and allow easier donning and doffing for  larger males.

        •      An increase in leg  opening size will provide a wider range of fit, especially
               when donned over safety shoes.

        «      An increase in the  difference between front and back width will increase
               range of motion, especially for the arms when reaching.

        •      An increase in the  range of front opening lengths  will ease donning for larger
               males.
                                         403

-------
       TABLE 4. PROPOSED MINIMUM FINISHED DIMENSIONS FOR
                     LIMITED-USE COVERALLS

A.
B.
C.
D.
E.
F.
G.
H.
I.
J.
K.
L.
M.
Dimension
BACK BODY LENGTH
FRONT BODY LENGTH
ARMHOLE WIDTH
SLEEVE OUTSEAM
SLEEVE OPENING
FRONT CHEST
BACK CHEST
LEG INSEAM
LEG OPENING
FRONT OPENING LENGTH
HOOD LENGTH
HOOD DEPTH
NECKLINE LENGTH
XS
38
33
11
32
6
22
23
28
10
29
16
10
16
S
39
34
11
32
6
23
24
29
10
29
16
10
16
M
40
35
12
33
7
24
26
30
11
30
17
12
18
Size
L
41
36
13
34
7
26
29
31
11
30
17
12
19
XL
42
38
14
35
8
28
32
32
12
31
18
14
20
XXL
43
40
15
36
8
30
34
33
12
32
18
14
20
All measurements in inches.
                             404

-------
        •      ANSI/ISEA 101-1985 does not currently include hood dimensions, however,
              hoods are considered desirable for EPA activities, and must Tit when worn
              over hard hats. Hood opening length and hood depth requirements are
              required.

        •      Neckline length requirements are needed to assure fit when worn over winter
              clothing.

        «      The procedures for measuring the dimensions of a garment should be changed
              from that of ANSI/ISEA 101-1985 to those proposed in Table 5.

        In the proposed sizing system, extra-small and small have been optimized for
females, medium represents a compromise for larger females and smaller than  average males,
large represents a slightly larger than average male, and extra-large and extra-extra-large are
optimized for upper-percentile males.  Size extra-extra-large should fit a 95th percentile
worker wearing winter clothing in a temperate climate zone.

                                   CONCLUSIONS

        Improvements described for both fabrics and coverall sizing are a starring point for
continuing efforts to provide improved protective clothing for EPA hazardous waste workers
and guidelines for its procurement.
                                        405

-------
                  TABLE 5. PROPOSED LIMITED USE COVERALL
                          MEASUREMENT PROCEDURES
A.     BACK BODY LENGTH.  Measure from the top of the neckline at the center back
       collar seam to the crotch seam,

B.     FRONT BODY LENGTH. Measure from the top of the neckline at the center back
       point to the crotch seam with the coveralls flat and front side up.

C.     ARMHOLE WIDTH. Establish a line from the base of the armhole which is parallel
       to the center front. Measure up from the armhole base to the top of the sleeve with
       the coveralls stretched flat.

D.     SLEEVE OUTSEAM. Measure from the center back point to the top of the sleeve at
       the wrist edge.

E.     SLEEVE OPENING. Flatten sleeve at wrist end, completely stretching elastic if
       present. Measure from one folded edge to the other.

F.     FRONT CHEST.  Measure from the base of the armhole across the front chest to the
       base of the other armhole. If there is no underarm seam on either sleeve or body of
       coverall, lay the sleeve and body of coverall at an angle where both are flat and
       establish an underarm point at the juncture of the sleeve and torso.

G.     BACK CHEST.  Measure from the base of the armhole across the back to the base
       of the other armhole, including all of the fullness that lies between these two points.
       If there is no underarm seam on either sleeve or body of coverall, lay the sleeve and
       body of coverall at an angle where both arc flat and establish an underarm point at
       the juncture of the sleeve and torso.

H.     LEG INSEAM. Measure from the crotch seam down the leg inseam to the bottom
       edge.

I.      LEG OPENING.  Flatten the leg at the ankle end, completely stretching elastic if
       present. Measure from one folded edge to the other folded edge.

J.      FRONT OPENING LENGTH. Measure from the center back collar base to the
       bottom of the front opening with the coverall flat and front side up.
                                                                        (continued)
                                      406

-------
                               TABLE 5  (continued)
K.     HOOD OPENING LENGTH. Flatten the hood along the center seam so that the
       sides are superimposed.  Measure on a flat vertical line that extends upward from the
       neckline seam to the highest point on the top of the hood.

L.     HOOD DEPTH. Ratten the hood along the center seam so that the left and right
       sides are superimposed.  Measure on a horizontal line from the center front (face)
       edge to the back of the hood at the point of greatest depth.

M.     NECKLINE LENGTH.  With front of coverall facing up, stretch neckline seam flat.
       Measure from one end of seam to the other.
                                      407

-------
              Methods  fco S*galnafce fcha Stress Cgaclc Reaigfcanep of
                Hicrli Density ^Qlv^iiHvl^ii^ ITMIi Sliftftta and S^aftia
                      Yick H.  Halse, Donald E. Hullings,
                  Robert M. Koerner, and Arthur B.  Lord,  Jr.
                       Geosynthetic Research Institute
                              Drexel University
                            Philadelphia,  PA 19104
                                   Abstract
     The stress crack resistance of high density polyethylene  (HDPE)  flexible
membrane  liner sheets and  seams  is evaluated using  three different testing
methods.   The sheet  materials are evaluated using the bent  strip constant
strain  test  (ASTM D-1693)  and the notched constant  load  test  (KCLT) .   The
seams   are  evaluated using  the   seam constant   load test   (SCLT) .    The
experimental  procedures used in each of these methods  is described.
     In addition,  the effects  of  elevated  temperature on the  stress  crack
resistance  of HOPE sheet  is discussed.  The surface temperature of the FML's
can  be  reduced significantly by using some  type of cover materials, thereby
reducing the  stress cracking potential of the FMLfi.

                                 Introduction.
     The stress cracking of  high density polyethylene  {HDPE? geomembrane sheet
and  seams  is  of concern due to a  few failures  at  surface impoundments  where
the  FML%  were  exposed to the ambient  environment.   Presently,  only  one
quality control  test  method, ASTM D-1693 CD, is used to evaluate the stress
cracking of HDPE  sheet materials used as  FMLS«    However,  the currently used
test conditions may not be able to  distinguish the  stress cracking performance
between different materials.    It   is  felt  that  more  quantitative  and
challenging test methods would help to ensure that  the best  possible materials
are  being utilized.   It should also  be noted that there  is no  current  test
method used to evaluate FML seams as to their stress crack  resistance.
     In  addition  to  a  qualification  of  the HDPE materials,  the  ambient

-------
environment where the FML will be  utilized should also be considered.  In some
applications,  such  as  surface impoundments,  the  FML  is  fully exposed  to
ambient environments along the runout length and down the slope to the liquid
level.    The  surface  temperature  of the  FML  becomes  a  concern,  since
temperature is one of the  acceleration  factors  associated  with aging of HOPE
liners and stress crack propagation.  Minimizing the surface temperature and
direct  sunlight  exposure will  probably  extend  the performance  of  these
materials.
     This paper  focuses  on  four  separate topics,  each  of  which will  be
independently presented.  They are as follows:
     *   Bent strip test for HDPE sheet materials (ASTM D-1693).
     *   Notched constant load test (NCLT) for HDPE  sheet materials.
     •   Seam constant load test (SCLT)  for HDPE  seams.
     *   Temperature reduction of exposed HDPE liner materials.
Suggested procedures and recommendations on all  four  topics are included.
                        Bent Strip Test fASTM D-1693>
     The  ASTM D-1693 bent strip test entitled "Environmental  Stress-Cracking
of  Ethylene  Plastics"  is  the only  test  method  currently  used  by  FML
manufacturers to evaluate the susceptibility of polyethylene sheet material to
stress cracking.  The procedure uses small  specimens  (38mm x  13mm)  which  are
taken directly  from the manufactured sheet.   A  central notch 20mm long  and
from 0.30 to  0.40 mm deep is introduced along the length of each of  the  ten
test specimens.   The test specimens  are  then bent  180° and held in the  flanges
of a small channel.  The entire assembly is immersed in a surface-active agent
at an elevated  temperature.   Figure 1  shows a notched  specimen,  the  channel
holder and  a glass  beaker  which can be  used to  incubate  the  set  of  ten
replicate  specimens.   Note that the  test  specimen  in  this configuration
results in a  constant  tensile  strain on the outside surface.   If cracks  are
observed on  the surface before an  arbitrarily defined length  of time,  the
sheet material is considered to be sensitive to stress cracking.   When cracks
occur,  they generally  develop  at  the  sides of the controlled notch near  the
top of the  arch,  growing towards the edge of the specimen at  approximately
right angles to the notch.  They sometimes break  through the entire  thickness
of the specimen in a brittle mode indicative of poor stress crack resistance.

                                   408

-------
                                   16 mm
                   notch
38 mm
         *	+
                          165 mm
                    19 mm
          13 mm
    (a) Test Specimen
                              (b) Specimen
                                                      (c) Test Assembly
     Fig.  I - Test setup  for ASTM D-1693 bent strip  test
                                  410

-------
     Since this test is accelerated by  both  the surface-active  agent  and the
elevated  temperature,  the conditions of these  two  factors must  be  carefully
considered.  A National Sanitation Foundation  (NSF) Committee has recommended
in their  Standard #54  that the conditions of  the  test  be  100%  Igepal  as the
surface-active  agent  and 100°C temperature.   They  also  stipulate that no
failures  occur within  a  500 hours  test duration.   It must be recognized
however,  that  these conditions are  not extremely severe,  and each  of the
conditions is discussed in following:
  •   Concentration - The viscosity of 100% Igepal solution is very high.   The
     liquid is  unlikely to penetrate completely into the notch and expand the
     defects so as  to  assist  in  the crack  propagation.   A  lower  viscosity
     solution with  higher penetration  capability  will facilitate  the crack
     growth more if  there  is such a tendency.
  •   Temperature  -  The  100°C  test  temperature  is  very close to  the  melting
     point of  HOPE  which is  124°C.    Thus  stress  relaxation  in the  test
     specimens can occur much  easier than at  lower temperatures.   In addition,
     for test solutions which contain water, the testing  temperature  must  be
     lower than 100°C to minimize the evaporation of water.
  •   Time  - The duration for an acceptable test is  a  difficult  decision.   if
     the defined testing  time  is  too short,  it certainly  cannot demonstrate
     the true property  of the material.   On  the other  hand,  a long  period  of
     time   may  not  necessarily  be  a  sensible choice  because  of  polymer
     stress-relaxation.  Indeed, there may not be any tensile stress remaining
     in  the  specimen after a  long period  of  testing  time.    The  presently
     defined 500 hours  testing time is probably too short,  since cracking has
     been  observed in some tests at times between 500 and 1000  hours.

     The  recently reconvened  NSF #54  Committee  has also  addressed these
details of the test  and  its current  deliberations  are focused around the use
of  10%  Igepal  in  90%  tap  water,  at  50°C,  for  1000 hours.    They  also
recommended  that  no  failure  should occur in any of ten test  specimens which
are to be taken equally from both the machine and cross-machine; directions of
the sheet material.

                                   411

-------
                       KTnfrehed Constant Load Test  (NCLT)
     This test is  intended to evaluate the stress cracking resistance of  FML
sheets  under  constant load.   As such,  it does not allow for stress-relaxation
to  occur as in the previous test.  For this  reason alone,  the  test  is a more
severe  indicator of the stress  crack  performance of  HDPE sheet materials.
     The test itself has references dating back to 1960(2), which  resulted in
an  ASTM test procedure that  was designated ASTM D-2552.   It has since been
dropped by  ASTM  due to inactivity in the  use  of this  standard.  The  test used
dumbbell shaped specimens at  various  constant  loads  which  were  equal   to
percentages of  the  yield strength of the  material.   The specimens  were
immersed in a surface-active agent at  an elevated temperature.  A  schematic
diagram of  a 20  station  test unit is  shown in Figure  2.   Each station  was
individually equipped with a  timing light  which shuts  off when  a given amount
of  deformation of the  specimen occurred.  Thus  the  test  had three possible
concluding  states:
     «   elastic state     - where the light never shuts off up to an  arbitrary
        time limit, e.g.  1000  hours
     •   plastic state     - where the light shuts off due  to  plastic elongation
        of the specimen
     *   cracked state     - where the light  shuts  off  due to brittle cracking
        of the specimen

Unfortunately, this test as written required  an extremely long period of time
{•»  10,000 hours)  to  conclude.    Furthermore,  the results  were very random.
This was because there was no  controlled imperfection within the specimens.
The crack initiation  in the individual test specimens W8S  dependent upon  the
material's own defects which varied randomly from one specimen  to another.
     A  modification  of this  test to  alleviate  the  above disadvantages   is
proposed.   The specimens  to  be  used in this new test are  made according  to
ASTM D-1822 which  has a constant  central section 120mm long and 30mm wide.   A
controlled notch is then introduced into one side  of the  surface to a depth of
20% of the thickness of the specimen.   Figure 3 shows  the schematic diagram of
a notched test specimen.  To distinguish the new test  from the older  one, this
test is  designated a  notched constant  load test,  or  "NCLT".  The  details  of
specimen preparation have been fully described in  a previous paper(3).
     Test specimens are taken parallel to  the  cross machine direction  of  the
                                   412

-------
Micro Svttch to
   Timers
 Pin
Joi nts
         Shot
         Can
                   I
                               20 Positions
75.20cm
  30*
                                                        a=n=
                                       Specimen
             SIDE VIEW
                         •Traymoved up and
                          do\*nonrackand
                          pinion arrangement
                  FRONT VIEW
           Fig.  2 - Constant stress  loading apparatus consisting
                     of twenty specimen test positions
                 o
           V
     (0.2t)
                     \
                                                 VI 0.2%
                 O
                      60 mm (2.4inch)

               Sid* Tiew of Th» Cut test Specimen
      Front Tier of th* Cut Test Specimen
          Fig.  3 - Schematic diagram  of  Notched  test specimen
                    for notched constant  load test (NCLT)
                            413

-------
 sheet and  then notched perpendicular  to the  specimen length,  i.e.  in the
 machine  direction of the sheet.  Past  testing  has  shown this to be the most
 sensitive direction  for stress  cracking.    The  notched specimens  are then
 prepared with grommets in their end tabs and placed in the test  device which
 is  capable  of testing up to 20 specimens simultaneously as was shown  in Figure
 2.
     The  specimens are subjected  to  a  predetermined  constant  load while
 immersed in a 10%  Igepal/90% tap water  solution at a temperature of 50°C.  The
 load  levels range  from 25% to 70% of the yield stress of the material at 50°C.
 The load increments   are  at  even  5%   intervals.   The  failure  time  of  the
 specimens is  automatically recorded to  the nearest 0.1 hour.  After the test
 is  completed  (the  time of which varies, but is at least long enough to obtain
 brittle  failures in some of the specimens),  the  experimental data  is presented
 by  plotting the applied stress against the failure time on a log-log scale.
 The data points  can  be  visually separated  into  two  linear regions.   The
 initial  region consists of ductile  failures and the terminal region consists
 of  brittle  failures,  see Figure  4.   The time corresponding to the  point  at
 which  the lines intersect is called the  ductile-to-brittle transition time,
 and is designated as "Tt".
    The  transition time of the WML sheet has  been shown to be very sensitive
 to  the   properties of the material  and probably to the manufacturing process
 as  well.  The results  presented in previous paper(3), and reproduced in Table
 1,  indicate that there is  a  wide range of transition times,  from 4  to  600
 hours, between different HOPE  sheets.   Also note  that there is a correlation
between the transition time and the bent strip test results, particularly for
 the poor stress cracking resistance materials.    However, for  materials with
transition times greater than 100 hours, the bent strip test results  were the
 same.   Hence  the bent strip test is a qualitative test, whereas the NCLT is a
quantitative one.
                                   414

-------
   100
                                               ductile region
                                                                         Knee"
M
0)

-------
Times for Five Different polyethylene Sheets (from Bef- 3)

Materials


S-l
S-2
S-3
S-4
S-5
Bent Strip Results
{no . failure/
hours of test)
9/700
6/700
0/2000
5/2000
0/2000
Transition Time
(hours)

4
6
600
28
300
                        Seam Constant Load Teat(SCLTV
     Stress crack resistance  should also be evaluated for the FML  seams  as
well as for the sheet, since  the seams are probably the most vulnerable parts
of  the  entire liner  system.   The  results  published  in a. previous  paper H)
demonstrated that  all types of  HOPE  seams  can be vulnerable  to  stress
cracking.  Hence, a test method Incorporating seams should be considered.
     The proposed test, which is called a seam constant load  test  (SCLT)  is
very similar to the NCLT just described, except that  a section of the seam  is
included  in the  test  specimen.   The  constant  neck section (120mm)  must
accommodate one edge of the seam where it joins to the parent sheet  material.
Note that  there  is  no induced defect  (i.e. there is  no deliberate  notch)  in
the specimen to  initiate  cracking.  The discontinuities around the  seam are
the  focus for  crack initiation.  Figure  5  illustrates   the   types  of
configurations to be used for the major types of HOPE  liner seams.
     Twenty test  specimens, ten  from each  side  of the seam on the  upper and
lower FMLS. are subjected to a load and then immersed  in a  10% igepal/90% tap
water solution  at 50°C.  The  applied  load is recommended to be 45%  of the
yield stress of the sheet material at 50"C.  The  present  duration of the test
is  1000 hours.   The  specimens  will respond in one of three  ways;  elastic,
plastic, or cracked  as described previously.   Based  upon past experience  in
testing of this nature, the following criteria  should  be  considered  for seams
to avoid stress cracking vulnerability:

                                   418

-------
           I
        t
        1
           I
          i
(a)  Fillet Extrusion Seam
(k)  Flat Extr-usion Seam
(c)  Hot Wedge Seam
    Hot Mr Seam
    Ultrasonic Seam
         Fig.  5 - Sketches of seamed test specimens in  tension in
                   a seam constant  load test  (SCLT)
                                 417

-------
     (a)  No brittle failure should occur in any of the 20 test specimens
         within 24 hours
     Cb)  No more than five (5) brittle failures of the twenty test specimens
         should occur within 100 hours.
     (c)  If less than five brittle failures occur within 100 hours,  then the
         average of all brittle failures up to 1000 hours test duration  should
         be greater than 300 hours.
               Temperature  Reduction of Exposed HOPE Geomembrarie
     In all three of the above  described tests,  elevated temperature is used
to  accelerate the crack growth rate.   Chan  and Williams(5)  indicated that
crack  growth rate  is  increased  40%  for  every degree Celsius  rise  in
temperature.   Since crack  growth rate is directly related to brittle failure
one  can  anticipate  a  similar  proportionality.    Therefore  the  surface
temperature of the  liner will probably  have  a  direct  effect  on stress crack
behavior.
     The  temperatures  of  exposed liners  can  be  very  high.    A  maximum
temperature  of 67°C (178°F)  has  been recorded  on  a  black roof  membrane  in
southern Florida(6).  If that is the  surface temperature of an exposed WKL, a
particular crack will grow 1400% faster than one which  is at 30°C temperature.
In addition, ultraviolet degradation will be another action on  the exposed FML
which could have synergistic effects with the elevated temperature.   Choi and
Broutman(7) found 90% reduction in the brittle failure time of HOPE after 250
hours of UV exposure using a  OV fluorescent weatherometer.  The minimization
of one or both of these actions  (temperature and ultraviolet light exposure),
would definitely extend the performance of the FML.
     Toward gaining  insight into surface  temperature moderation,  a  number  of
laboratory trials on exposed  liner have been attempted.   For these experiments
a sheet of HOPE liner was  fitted with a  thermocouple directly  on  its surface
and placed under a heat  lamp.   Three different trials were  examined.  They are
as follows:
     • soil covering
     « sacrificial covering of liner
     • light colored surfaces
                                   418

-------
     Regarding soil covering experiments/ various thickness  of  a well graded
quartz  sand  were  placed above  the  liner  to  note  the changes in  liner
temperature with time.  Figure 6 presents these results.   Here it is seen that
soil  coverings of  more  than two  (2)  inches  are required before  meaningful
temperature reductions accrue.  More from a practical point of view, however,
the thinnest layer  one could place on an FML without danger of damaging it is
approximately  six  (6) inches.   Unfortunately,  soil coverings on  FML's  are
problematic  in  their  long  term  stability.     Instability comes  from
gravitational  forces,  seepage  forces,  precipitation induced  erosion and wind
induced erosion  from the  contained  liquid.   Thus,   without  continuous
maintenance, soil covering of the liner is only a short-term solution at best
and one which is undoubtedly very expensive.
     Several  attempts  at  providing  a  sacrificial covering  of  various
geosynthetic  materials  over the  primary liner  have been  attempted.    The
sacrificial covering  that  has been used is itself  a  liner with a separating
air space  provided by either a geotextile  or a  geonet.   Figure 7  shows  the
results of these  trials  insofar  as  the  temperature  of the  primary  (or
underlying) liner is concerned.   Here it  can be seen that  the geotextiles with
the highest mass per unit area  (hence largest thickness)  provide the greatest
temperature reduction effect even beyond that of  the geonet.   However, such a
system would be very expensive.
     Utilizing  references from the roofing industry,  which is  heavily involved
in flat roof waterproofing using polymeric  liners,  it  was found that  surface
temperatures  are greatly  reduced  using white  colored  liner  materials  as
opposed to black liners,  To  note  such an effect  on HDPE liners,  experiments
were  performed using  a black liner  and  then  the same liner but now  painted
white.  The  results are provided in Figure 8 and are quite startling.   The
difference in temperature between the two trials  is in excess of 25°C <77°F) .
According  to  the previously  stated  reductions in  stress crack growth from
elevated temperatures, this temperature reduction represents  a  1000% increase
in stress crack resistance.
                                   Summary
     Since  stress  cracking poses a concern with exposed FML S  made from HOPE,
methods for evaluating the phenomena  are  necessary.   The current bent strip

                                   419

-------
o
o
e
&.
a
*•
to
k.
e
Q.
        40
        30
        20-f—'
                  60     120    180    240     300     360     420     480
                                        Time (minutes)
                  Fig. 6. - Surface Temperature Effects  on  a  Soil

                            Covered FML for Varying Soil Thicknesses

-------
                                               Control (Sacrificial FML)
                     ' 4 oz/sq.yd. Geotextite  ^~ ~
20
                                                         90
120
                                  Time (minutes)
 Fig.  7.  - Surface Temperature Effects  on     Geosynthetic Covered FML for
     Varying Air Space  Thicknesses Created by Geotextiles and Geonets

-------
o
o
0>
b.

3
**

«
b>

0

O.

E
        20
                                      120
180
240
300
                                 Time (minutes)
                    Fig.  8.  - Surface Temperature Effects Between

                              Black and White Surfaced FMLS

-------
test  (ASTM  D-1693)  should be made more severe  to  better distinguish between
different  HOPE liners.   An additional  test for  HDPE sheet materials,  the
notched constant load test  (NCLT) does not a-llow for the stress relaxation of
the polymer associated with the bent strip test.   The NCLT  determines the
transition  time  between  ductile and brittle  failure  modes.   A  minimum
transition  time  for qualifying HDPE liners  is  currently being investigated.
This same test setup can  be used to  estimate the stress cracking tendency of
seamed samples.   A test procedure called seam  constant  load test (SCLT) has
been  developed  and has  resulted in  a  set of criteria  for evaluating an
acceptable HDPE seam with respect to  stress  cracking potential.
    Since  the stress  cracking  potential  greatly increases  with a  rise in
liner  temperature,  a  study was undertaken in this  regard.   The  methods
examined to minimize the liner's temperature included soil covering of varying
thicknesses, a sacrificial  liner separated  by either geotextile  or a geonet,
and a  white painted FML.   Of  these  alternates  the last one,  i.e.  the white
colored FML, is  very intriguing  and  should  be explored further.   It may well
be that all exposed FMLS   for  longer than  a specified  time  should be limited
in their maximum surface temperature.
                               Acknowledgements
    This  project  was sponsored by the U.  S.  Environmental Protection Agency
under Project Number CR-815692.  We  sincerely appreciate  this  support and of
the involvement of the Project  Officer David A. Carson.
                                  Re,feren,ce
(1)  ASTM D-1693,  "Environmental  Stress  Cracking  of  Ethylene  Plastics",
    Vol.08.02, 1988
(2)  Lander L.L.,  "Environmental Stress Rupture of Polyethylene",  SPE Journal,
    Dec, 1960
(3)  Halse Y.H., Lord A.E.,Jr. and Koerner R.M. .,"Ductile-to-Brittle
    Transition Time in Polyethylene Geomembrane  Sheet,"  will be
    published in Geosynthetic Testing: for Waste  Containment
    Applications. Jan, 1990

                                   423

-------
(4)  Halse Y.H., Koerner R.M., and. Lord. A.E.Jr.,"Laboratory
    Evaluation of Stress Cracking in HOPE Geometnbrane Seams",
    and,,purflbiJ.ity  o£ Geosynthetics,  Koerner,  R.M.,Ed.,  Elsevier Applied
    Science, 1989

(5)  Chan M.K.V. and  Williams J.G.,  "Slow Stable Crack Growth in High Density
    Folyethylenes", Polymer, Vol 24, 1983

(6)  Backenstow  D.E., "Comparison of  White Versus Black  Surfaces for Energy
    Conservation", Published by Carlisle SynTec Systems, Carlisle, Pa,

(7)  Choi S.W. and. Broutman L.J.,  "Dutile-Brittle Transtions for Polyethylene
    Pipe Grade Resins" Eleven glastic Fuel Gas Pipe Syirposjum, 1989, pp. 296
                                    424

-------
                           LANDFILL TEST CELLS
                       by: N. C, Vasuki, P.E., DEE
                             General Manager
                     Delaware Solid Waste Authority
     The Delaware  Solid Waste  Authority (DSWA) is a special Authority of
the  State  of  Delaware  expressly  created  to  manage  the  disposal of
municipal solid waste generated in the State of Delaware.  By the time the
Authority was created  in  1975,tl>  only  7  active  landfill  sites were
receiving waste, and 35  other sites were shut down because they failed to
meet  the  state's  new  landfill   regulations  adopted  in  1974,   Those
regulations  imposed,  as  a  minimum, the requirement of liners, leachate
collection systems, gas vents and  groundwater monitoring.  Only one of the
seven  remaining   sites,  the  Pigeon  Point  Landfill  near  Wilmington,
Delaware, had a leachate collection system.
DSWA PROBRAM

     DSWA initiated a program  to construct  new landfills  in 1978.   Its
first new  landfill was  opened at Sandtown, Delaware on October 23, 1980.
This was also the  first landfill  in the  state to  utilize a  30-mil, PVC
flexible  membrane  liner  (FML).    In  January  1981,  DSWA acquired the
operating responsibility  for the  Pigeon Point  Landfill and subsequently
made extensive  improvements including  an expanded groundwater monitoring
program and a peripheral leachate collection  system  at  the  toe  of the
above-ground mound-type landfill.  In September 198%, the Authority opened
its Southern Solid Waste Facility at Jones Crossroads.   This  facility is
another above-ground mound - type landfillf  however, because  of the high
water table, this mound type landfill utilizes two 30-mil.  PMC FMLS  with
the drainage layer between the two FMLS.

     In  October  1985,  the  Authority  opened  its third new landfill at
Cherry Island.  This facility is constructed on compacted silt  dredged by
the  U.S.  Army  Corps  of  Engineers  from the Delaware River and Port of
Wilmington shipping channels.    This  facility  is  also  an above-ground
mound-type  landfill,  like  the  Cherry  Island  Landfill, with extensive
leachate collection, gas venting and groundwater monitoring systems.

     The Authority's three landfills serve the  entire state's  needs.  In
calendar year  1989, the  three landfills  received 707,000 tons of waste.
Delaware is perhaps the only state with a policy of  neither exporting nor
importing solid waste.
    7 Del. C. Chapter M
                                   425

-------
LANDFILL OPERATIONS

     The  Central  Solid  Waste  Facility  was the first facility where we
could measure leachate production from a new cell.  The  first 9-acre cell
with a simple leachate collection line showed very low volumes of leachate
production.  Rather than pursuing leachate treatment and disposal options,
the  Authority  decided  to  attempt  recycling leachate through the cell,
using Pohland'st=> postulation that leachate recycling would be beneficial
to maintain  the organic  materials degradation  process in the landfills.
Leachate recycling at this cell was reasonably successful  and the concept
Mas extended to the new cells at both the Central and Southern Solid Waste
Facilities.   Between October,  1980 and  February, 1989,  no leachate was
removed  for  disposal.    Therefore,  the  concept  was deemed a success;
however, there were some side effects.   Periodically  spray irrigation of
leachate on the open face created odorous aerosols.  It was not a reliable
method because  of the  vagaries of  weather conditions  and was therefore
abandoned.  Leachate seeps started appearing on the side slopes.  In order
to prevent the leachate escaping outside the cell boundaries, major repair
work was  necessary.   Most seeps appeared because (a) presence of plastic
sheets or bags in the waste impeded the  movement of  leachate through the
fill material, (b) compacted daily cover sometimes acted as a barrier when
the daily cover had  a high  clay content,  (c> short  circuiting resulted
when the movement of leachate was impeded (the path of least resistance is
followed) and Cd) the leachate mound started to build up.

     The attempt to correct  these problems  resulted in  complicating the
water balance  around the  cells.   It was difficult to use these cells to
check the leachate  production  rate  against  the  HELP  model  or the
Mathei—Thornthwaitee**>  water  balance equations.  In view of the impending
avalanche of both state and federal regulations on solid waste  disposal,
. "The
hydrologic evaluation of landfill performance (HELP) model. "Documentation
for Version I, EPA/530-SW~8^-010,  2, U.S.  EPA, Ofc.  of Solid  Waste and
Emergency Response, Washington, DC
<*»> Thornthwaite,  C.W. and J.R. Mather, 1957.  Instruction and Tables for
Computing   Potential   Evapotranspiration   and    the   Water   Balance.
Publications in Climatology^ Laboratory of Climatology 10(3).
                                    428

-------
DSWA thought  it was prudent to test and quantify the benefits of leachate
recycling.   It was  also important  to find  out the  merits of operating
landfills under moist conditions as opposed to maintaining dry conditions.
Since external leachate treatment is quite expensive, leachate circulation
helps reduce the cost  of treatment.   In  fact, most  newly adopted state
regulations and the proposed Subtitle D Regulation of EPA  require prompt
removal  of  leachate  and  eventually  drying  up  the landfi. 11 after the
impervious capping system is installed.

TEST CELL PROGRAM

     The Authority's  objective  was  to  observe  the  difference between
"moist" and  "dry cells".   Although  such studies  have been attempted in
laboratories and in small test  plots,  no  controlled  test  in  a "real"
landfill environment  had been  attempted.   Moisture is needed to support
biological activity within the  landfill.   If a  landfill cell  is capped
with an  impermeable cover  to prevent  infiltration  and  the leachate is
continuously withdrawn, it will  gradually  lose  moisture  and biological
activity will  slow down.   Landfill  gas production  could be  used as an
empirical index of such activity.   If  it  could  be  shown  that "moist"
landfill  cells  offer  a  better  environment  for  continued  biological
degradation of the easily degradable organic fraction of  the solid waste,
as compared to dry landfills, the "post-closure" monitoring period of such
cells could be reduced because the cell will have achieved stability.

     Although a dry cell may appear to be stable, during  the post-closure
monitoring  period   any  cracking  of  the  cap  could  result  in  water
infiltration.  Such an event would  gradually trigger  biologiical activity
again  within  the  cell.    If  the  cell is not carefully monitored, the
increase in biological activity may not  be  noticed  until  :it  becomes a
serious problem.

     In  this   context,  it  is  important  to  measure  the  difference.
Therefore, the Authority decided to construct two landfill cells, each one
acre in  size.   In Test  Cell #1,  provisions are made to recirculate the
leachate while Test Cell #£ will be allowed to become dry.  Each cell will
receive household  solid waste  from the  same select area and, therefore,
contain similar materials.  Initial estimate  showed 11,000  tons capacity
in each  cell, compacted  to the minimum 1,000 Ibs. per cubic yard density
used by the Authority.  The same contractor operating the Authority's main
landfill cell  at its Central Solid Waste Facility operates the test cells
also.

     This program  also  offered  an  unusual  opportunity  to  test other
variables  in  the  landfill  design  such  as liners, leachate collection
systems, etc.
                                    427

-------
CONSTRUCTION DETAILS

     Site preparation was initiated in June, 1987.  Each cell was designed
to accommodate four different  flexible membrane  liners  36 -mil. CPE-R, (2) 36-mil. Hypalon CSPE-R, (3) 30-mil. PVC,
and (4) 30-mil. CPE.  A 30~mil.  PMC liner was used as the secondary liner.
Figure S shows details of settlement plates and instrumentation.  Figure 3
shows the relative location  of  the  test  cells  within  the Authority's
Central Solid  Waste Facility.   Figure  k shows details in cross section.
Figures 5 and 6 show the cross section of Test Cell 1  and 2 respectively.
The primary  and secondary  liners are separated by a drainage layer which
intercepts any leakage  from  the  primary  liners.    Each  cell  has two
leachate collection  lines as  shown in  Figure 2.   Each cell has its own
drainage area.  All the runoff from each  drainage area  is directed  to a
Parshall flume  having a  12-inch throat  width.  Initially, an ultrasonic
flow meter was used to measure the runoff.  Because  of its unreliability,
a  90  degree  V-notch  was  installed with the standard chart recorder to
measure the runoff.

     A standard weather station near the entrance to the landfill  ('4 mile
away)  provides  information  on  humidity, wind speed and wind direction,
solar radiation, temperature and precipitation.    A  special  "acid rain"
collector  is  also  used  to  measure the acidity of the precipitation in
central Delaware.  A data logger is used to store and retrieve  data.  The
construction cost of the cells was approximately $800,000.

TEST CELL OPERATION

     Actual filling  started on  August 1, 1989.  Since the selected waste
shed has garbage collection on Mondays through Fridays, the truck delivery
pattern was adjusted to assure that both cells received waste collected on
each weekday.  Table I provides operating information for the  first three
months of the cell filling program.

CURRENT STATUS

     The cells  have reached  almost 80*4  capacity.  We anticipate closing
out the test cells in Hay, 1990.    By  that  time,  each  cell  will have
approximately 9,000 tons of waste.  We have also attempted the use of foam
as daily cover in both cells  in lieu  of soil  to increase  the volume of
waste in the cells.

     In  September  1989  and  February 1990, waste characterization tests
were conducted.  The information from those tests is shown in Table 2.
                                    428

-------
     The  U.S.   EPA  has   provided  a   grant  to   study  the  chemical
characteristics of the leachate.  The Authority has contracted with Dr. F.
Pohland of  the University  of Pittsburgh to conduct the chemical analysis
and also assist in a review  of  the  data  from  the  program.   Leachate
samples  are  shipped  to  the  University of Pittsburgh in special teflon
bottles.  Some tests such as pH, ORP (oxidation reduction potential), etc.
are conducted in the field.

     A  series  of  time  capsules  in nylon netting, containing materials
shown in Table 3 have been buried at select locations.

     Each capsule and the  materials  in  it  have  been  tagged.    It is
proposed that one set of capsules will be exhumed approxi-
mately 2  years after  the close  of the cells, and the remaining capsules
will be exhumed 5 years after the close of the cells.

     Two sets of leachate samples have  been  sent  to  the  University of
Pittsburgh for  analysis, and  it is too early to report on any findings of
the analysis.  Data on precipitation, leachate production  and runoff will
be used to verify the HELP model predictions.  Bas production will also be
measured and  gas samples  will be  analyzed for  carbon dioxide, methane,
oxygen, nitrogen and carbon monoxide.

     This test  cell set  up affords  an extraordinary opportunity to test
other  hypotheses   on   degradation   of   certain   chemicals   such  as
trichloroethene.   The initial  program will  only compare the quality and
quantity of leachate generated in each cell and the main cell (2^ acres).

     This  paper  has  been  reviewed   in   accordance   with   the  U.S.
Environmental Protection  Agency's peer and administrative review policies
and approved for presentation and publication.

ACKNOWLEDGEMENT

     The Authority  gratefully acknowledges  the grant  assistance of U.S.
EPA and the guidance given by Mr. R. Landreth and Mr. D. Carson.
                                    429

-------
                     TABLE 1.  SUMMARY OF OPERATIONS
                               August 1989   September 1989  October 1989
Test
Cell
ttl
1. WASTE DELIVERED 1269
(Tans)
2. OPERATING DAYS IS
3. TONS/DAY 106
4. PICK UP STOPS 48658
5. LBS./STOP 53
6. VOLUME USED*
(Cu. Yds.) 1773
7. DENSITY
(Lbs./Cu. Yd. 5 1431
8. COVER VOLUME
(Cu. Yds.) 894
9. PRECIPITATION 8.94
Test
Cell
#2
1101
11
100
45677
48
1420
1551
786
2.94
Test
Cell
#1
1059
10
106
50771
42
2226
1210
475
6.16
Test
Cell
#2
1107
10
111
53121
42
2453
1141
513
6.16
Test
Cell
#1
804
10
124
40470
40
1878
1154
F
3.44
Test
Cell
#2
1000
10
148
52E95
38
2366
1085
F
3.44
     (Inches)
*Rusmar Foam - Is an aqueous based foam consisting of approximately 95%
 air, 4K water and IY, solids.  The solids consist of surfactants and
 polymers which are completely biodegradable.
                                    430

-------
      TABLE e.  COMPOSITION OF RESIDENTIAL UASTE

                    September 1989       February 1990
Category	%	%
PAPER
Newsprint
Corrugated Cardboard/Kraft
Glossy Coated/Magazine
Mixed Paper
Total Paper
PLASTICS
HOPE
PET
Films
Mixed Plastics
Total Plastics
GLASS
Slass
Total Glass
METALS
Aluminum (recyclable)
Ferrous
Other Non-Ferrous
Total Metal
ORBANICS
Diapers
Text i 1 es/Rubber /Leather
Yard yaste
Other Organics
Total Organics
INORGANICS
Mixed Inorganics
Total Inorganics
HOUSEHOLD CHEMICALS
Pesticides
Non-Pesticide Poisons
Paint /Sol vent /Fuel
Automotive Products
Dry /Wet Cell Batteries
Miscellaneous Chemicals
Total Household Chemicals
OTHER
Special Wastes (e.g. tires)

13.88
7.89
3. 41
18.93
43.11

0.68
0.70
4.81
3.S9
8.83

3.58
3.58

0.92
4.92
0.03
5.86

1.88
3.73
10.71
21. OS
37.35

1.13
1.13

0.00
0.01
0.11
o.oa
0.01
0.00
0.16

0.00

ie.79
6.71
a. 62
19.64
41.75

0.59
0.84
3.88
4.03
8.73

4.97
4.97

1.39
6.13.
0.04
7.53

3.29
3.98
4.74
SI. 16
33.17

a. 58
a. SB

0.01
o.oe
1.03
0.03
0.11
0.06
1.87

0.00
                        431

-------
               TABLE 3.  MATERIALS  INCLUDED IN TIME CAPSULES
  1.  CORN ON THE  COB
  S.  ORANGE
  3.  CHICKEN/BONES
  4.  STYROFOAM
  5.  MCDONALD'S STYROFOAM CONTAINERS
  6.  WAXED PAPER  CUPS
  7.  FRENCH FRY CONTAINERS
  8.  BIODEGRADABLE DIAPERS  (UNUSED)
  9.  NON-BIODEGRADABLE DIAPERS  (UNUSED)
10. DETERGENT BOXES (SMALL CONTAINERS)
11. DETERGENT BOTTLES (SMALL CONTAINERS)
12. UNBACKED CARPET
13. RUBBER BACKED CARPET
14. NYLON
15. ORLON
16. COTTON
17. PANTY HOSE
IB. HAIR SPRAY CANS
19. NAIL POLISH
EO. SMALL PAINT CANS
HI. TRASH BAGS (UNIFORM MANUFACTURER)
aa. PLASTIC BREAD WRAPPERS WITH LABELS
33. SCOTCH TAPE
34. MASKING TAPE
35. DUCT TAPE
36. PHOTOGRAPHIC FILM
37. PET BOTTLES (POLYETHYLENE TEREPHTHALATE>
38. TFE DISKS (LOW DENSITY TEFLON DISKS)
39. GARDEN HOSE
30. PLYWOOD
31. SOFTWOOD (PINE)
33. IRON REBARS
33. GALVANIZED PIPE
34. LEAD-ACID BATTERIES
35. HOUSEHOLD BATTERIES (D CELLS)
36. CORRUGATED CARDBOARD
37. WAX BACKED CARDBOARD
38. REGULAR CARDBOARD (BINDER TYPE)
39. DELAWARE STATE NEWS
40. NEWS JOURNAL
41. WALL STREET JOURNAL
42. XEROX PAPER
43. DSWA LETTERHEAD
44.  NATIONAL GEOGRAPHIC
45.  TIME/LIFE MAGAZINE
46.  LEAVES
47.  TREE TRIMMINGS
CONSTRUCTION MATERIALS
48.
49.
50.
51.
53.
53.
54.
CPE
CPER
PVC
HYPALON
PVC PIPE
GEOGRID
GEOTEXTILES
                                   432

-------
                             CENTRAL SOLID WASTE FACILITY
                                   TEST CELLS
               LEGEND
    • 4" LEACHATE SAMPLING WELL
           (Bottom Drainage Layer)
    • 4" LEACHATE SAMPLING WELL
               (Solid Waste)
    A 2" DRAINAGE LAYER PIEZOMETER
                                                       • » PUMP STATION
                                                     • m TANK
                                                   	. 4" PERF. PVC
                                                   .-—a 6" PERF. PVC
                                                     -SPRAY IRRIGATION
                                                      CONNECTION BOX
    III PVC
    '•>' CPER
                   FLUME-1
    If HYPALON
    *& GRAVEL
                         	j*sjg	«j»
 TO THE
AREA
LAGOON
     I
   SPRAY
  IRRIGATION
CONNECTION BOX-1
                                    Figure 1

-------
          LEGEND
                                  CENTRAL SOLID WASTE FACILITY
                                        TEST  CELLS
  O    MAIN PUMP STATIONS

(    )  LEACHATE STORAGE TANKS

        DRECIRCULATION CONNECTION BOXEJ
        AND PUMPS

  0    GROUNDWATER MONITORING WELLS

        CLEAN OUT
  0
                 HONEYWELL
                   OR 4500
                  CIRCULAR
                   CHART
                 RECORDER
         STEVENS
US SYSTEMS  TYPE F
3700 SERIES  WATER
ULTRASONIC  LEVEL
FLOWMETER RECORDER
                        HONEYWELL
     LOGIC BEACH  US SYSTEMS  DR 4500
      PORTABLE   3700 SERIES CIRCULAR
     DATA LOGGING  ULTRASONIC  CHART
       SYSTEM   FLOWMETER^RECOROER,
 STEVENS
 TYPEF
 WATER
 LEVEL
RECORDER
                                                                                  ROCKWELL INT.
                                                                                  S-275 DIAPHRAGM
                                                                                  TYPE GAS METER
A, \  ««•« /
  |  ROCKWELL INT.
 I.     S-275
  I   DIAPHRAGM
  % TYPE GAS METEH
 O
                                                                                  \   I   /
                      t/1
                                                                                    SOILTEST
                                                                               SOIL MOISTURE PROBES
                                               SOILTEST
                                          SOIL MOISTURE PROBES
                                                 I
 IU I HE
 AREA "K'ffi *
 LAGOON V" 5
                     	    I      	               ^"**.<^|t<>*,t*t"J»S*£**».i*3ii
                                         TEST CELL #1
                                          TEST CELL
                                                         TO AREA "C"
                                           Figure  2

-------

-------
               CENTRAL SOLID WASTE FACILITY TEST CELLS
    TYPICAL CROSS SECTION OF LEACHATE AND
              GAS COLLECTION SYSTEMS
                                 GAS
                                         SETTLEMENT
                                          PLATES
 SAMPLE TUBE
 FOR ANAEROilC
  LEACHATE
 PUMP
STATION
                                 GAS
                                        SETTLEMENT
                             GAS
                            METER
                          -6"*PERFORATED PVC
                          COLLECTION HEADER
                                    CONCRETE V OUCH
                                    TO PARSHALL FLUME
   SAMPLE TUBE
  FOR ANAEROBIC
    LEACHATE
                          GAS
                         METiR
     PRIMARY LINER
     SECONDARY LINER
                         STONE COLLECTION SYSTEM
                                   CONCRETE V DITCH
                                   TO PARSHALL FLUME
 PUMP
STATION
Figure 4

-------
                       CENTRAL SOLID WASTE FACILITY
         CROSS SECTION VIEW OF TEST CELL  1
         LEGEND
• 4" LEACHATE SAMPLING WELL
      (Bottom Drainage Layer)

• 4" LEACHATE SAMPLING WELL
         (Solid Waste)

A2" DRAINAGE LAYER
 PIEZOMETER
                            LEACHATE -RECYCLING
                            ^60 MjL FILTER CLOTHi
  SAND LEAK DETECTION LAYER
        1' MINIMUM
4" PERFORATED PVC
SECONDARY LINER
  (30 MIL PVC)
                            Figure 5

-------
                     CENTRAL SOLID WASTE FACILITY
  W   CROSS SECTION  VIEW OF TEST  CELL 2
        LEGEND
 4" LEACHATE SAMPLING WELL
     (Bottom Drainage Layer)

 4" LEACHATE SAMPLING WELL
        (Solid Waste)

 2" DRAINAGE LAYER
   PIEZOMETER
GEOTEXTILE LEAK
DETECTION LAYER
SECONDARY LINER
  (30 MIL PVC)
                           'GEONETTENSAR'
                            Figure 6

-------
                         REVISION OF THE HELP MODEL

                    by:  Paul R. Schroeder
                         Environmental Laboratory
                         USAE Waterways Experiment Station
                         Vicksburg, MS  39180-6199
                                  ABSTRACT

     The Hydrologic Evaluation of Landfill Performance (HELP) model was
initially released in June 1984.  Subsequently, two verification studies
were performed, and the model has had widespread use and testing.  The
Hazardous and Solid Waste Amendments were enacted,and new minimum technology
requirements were developed.  Also, new technology has become available for
use in landfills.  These events have yielded numerous recommendations for
revision of the HELP model.  In response, drafts of Version 2 HELP have been
developed, and its final draft is presently being completed.

     The climatological data requirements have been changed to support a
synthetic weather generator and a vegetative growth submodel added to the
HELP model.  The soil data input has been changed to allow specification of
the initial moisture content of the various layers of material in the
landfill profile.  In addition, the default soil database has been updated.

     Small revisions have been made to routines used to compute snowmelt,
surface evaporation, plant transpiration and soil evaporation but the
fundamental methods used are unchanged from the original version.  The
vertical drainage model has been updated to use a Brooks-Corey formulation
for computing unsaturated hydraulic conductivity and vertical drainage.  The
lateral drainage model has been replaced with a new non-dimensionalized
analytical solution of the Boussinesq equation.  Additional submodels are
now being prepared for inclusion in the final draft of this version.  These
include a method for determining the effect of surface slope of runoff, for
estimating leakage through flexible membrane liners and for modeling lateral
drainage from synthetics drainage nets.  Other revisions have also been
made, particularly in the output.

     This paper has been reviewed in accordance with the U.S., Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.  The work described and presented herein,

                                     439

-------
unless otherwise noted,  WSS developed from research conducted under
sponsorship of the Risk Reduction Engineering Laboratory of the U.S.
Environmental Protection Agency by the Environmental Laboratory of the U.S.
Army Engineer Waterways Experiment Station.  Permission was granted by the
Chief of engineers to publish this information.
                                 BACKGROUND

     The Hydrologic Evaluation of Landfill Performance computer program
(HELP) is a quasi-two-dimensional hydrologic model of daily water movement
into, through and out of landfills.  HELP accepts climatological, soil and
design data and utilizes a variety of solution techniques to compute esti-
mates of surface runoff, evapotranspiration, lateral drainage to collection
systems, percolation through liners, average depths of saturation on the
surface of liners and changes in soil moisture storage.  Landfill systems
composed of various combinations of vegetation, cover soils, waste cells,
lateral drainage collection systems and relatively impermeable barrier
soils, as well as synthetic drainage media and liners, can be modeled.  The
model is applicable to open, partially closed and fully closed sites, and
facilitates rapid evaluation of a wide variety of landfill designs.  HELP is
a PC-based tool developed for use by both designers and regulators.

     HELP is a water budget model that was developed to assist in design
evaluations required by the Resources Conservation and Recovery Act (RCM).
Version 1 of HELP was initially released in June 1984 (Schroeder et al.,
1984a} 1984b).  Subsequently, two verification studies were performed, one
comparing the field data from a total of 20 landfill cells from seven sites
with HELP model results (Schroeder and Peyton, 1987a) and the other
comparing experimental lateral drainage results from physical models with
drainage estimates from the HELP lateral drainage submodel (Schroeder and
Peyton, 1987b).  The model was converted to run on IBM compatible personal
computers in 1986] and since then more than 2000 copies have been
distributed, generating extensive use and testing of model.  In addition,
since release of Version 1 of the HELP model, the Hazardous and Solid Waste
Amendments were enacted and new minimum technology requirements were
developed.  Also, new technology has become available for use in landfills.
These events have yielded numerous recommendations for revision of the HELP
model.  In response, drafts of Version 2 of HELP have been developed, and its
final draft is presently "being completed for release as Version 3.  This
paper presents the revisions of the HELP model.


                      MODEL DESCRIPTION AND REVISIONS

     HELP was adapted from the Hydrologic Simulation Model for Estimating
Percolation at Solid Waste Disposal Sites model of the U.S. Environmental
Protection Agency (Perrier and Gibson, 1980j Schroeder and Gibson, 1982) and
the Chemical Runoff and Erosion from Agricultural Management Systems
(CREAMS) (Knlsel, 1980) and Simulator for Water Resources in Rural Basins
(SWRRB) models of the U.S. Department of Agriculture (USDA) Agricultural
Research Service (ARS) (Williams et al., 1985).  The following sections of


                                    440

-------
this paper present  a description of all of the principal hydrologic and
physical processes modeled by HELP and presents the revisions in each,
including anticipated revisions to be included in Version 3.

INFILTRATION

     Daily Infiltration into the landfill is determined indirectly from a
surface-water balance.  Infiltration equals the sum of rainfall and snow-
melt, minus the sum of runoff and surface evaporation.  Runoff and surface
evaporation are in part a function of interception.  Precipitation on days
having a mean temperature below 32 degrees Fahrenheit is treated as snowfall
and is added to the surface snow storage.  Decreases in snow «torage occur
by snowmelt and surface evaporation.

Precipitation

     In Version 2 rainfall data may be synthetically generated in addition
to the options of being specified by the user or selected from the default
data base of historical rainfall data.  The synthetic weather generator will
be described later in this paper.

Snowmelt

     Snowmelt is computed using a degree-day method with 32 degrees F as the
base temperature (USDA, Soil Conservation Service, 1972); how&ver, several
small changes affecting the snowmelt have been made.  The primary change is
that Version 2 uses synthetically generated daily temperatures: in the
procedure instead of daily values interpolated from mean monthly tempera-
tures.  The inches/degree-day snowmelt constant was also increased from 0.06
to 0,10, a value more typical of open areas.  Finally, Version 2 permits a
small quantity of snowmelt to occur at mean daily temperatures between 23
and 32 degrees F to account for the variation in temperature during a day
and for the fact that landfills often have higher soil temperatures due to
heat generated from biodegradation.

Runoff

     HELP models surface runoff using the Soil Conservation Service (SCS)
curve number method, as presented in the Hydrology Section of the National
Engineering Handbook (USDA, SCS, 1972).  However, several small changes have
been made in its application in Version 2.  The relationship between the
curve number for moderately wet conditions, antecedent moisture condition
(AMC) II, and dry conditions, AMC I, has been revised to better cover the
entire range of possible curve numbers.  The soil moisture content which
corresponds to AMC I, yielding the maximum storage retention parameter, has
been increased to compensate for the larger evaporative zone depths used in
Version 2.  For the same reason the weighting factors for computing the
effective soil moisture content used to determine the daily curve number or
storage retention parameter have been changed to weight the near surface
materials slightly less.  Version 3 will include a procedure to adjust the
curve number as a function of surface slope since surface slopes greater
than 20 percent can produce significantly greater runoff.

                                     441

-------
 Surface Evaporation

      Surface  evaporation is  significantly reduced  in  Version  2.   It  is
 limited to  the  smaller of the  potential  evapotranspiration  and  the sum  of
 the  interception and the snow  storage.   Version  1  limited it  to the  smaller
 of the potential evapotranspiration and  the sum  of the  snow storage  and the
 quantity  of the rainfall that  did not  contribute to runoff.

 Interception

      Interception was not modeled in Version 1.  In Version 2 the intercep-
 tion approaches the maximum  interception exponentially  as the rainfall,
 RAIN, increases.  The maximum  Interception is a  function of the quantity of
 above ground  biomass or leaf area index, LAI, and  is  limited  to a maximum  of
 0.05 inches based on the work  of Horton  (1919).  Interception,  IHT,  is
 computed  as follows:

          INT    - INT    [1 " exP 
-------
however, several differences exist in its application in the two HELP
versions.  Version 2 calculates an evaporation coefficient for stage 2 soil
evaporation instead of reading the value from input.  The potential soil
evaporation or stage 1 soil evaporation may also differ because it is
calculated using the potential evapotranspiration and vegetative cover
(above ground biomass) which are computed with different data in Version 2.
The vegetative cover in Version 1 is described in input but computed using a
vegetative growth model in Version 2.  The vegetative growth model will be
described later in this paper.

Plant Transpiration

     HELP models plant transpiration in the manner of the CREAMS and SWRRB
models (Knisel, 1980$ Williams et al., 1985) whereby the potential plant
transpiration is a linear function of the potential evapotranspiration and
the active leaf area index.  Differences between the two versions include
the differences in determining the potential evapotranspiration and the
active leaf area index as well as differences in the selection of an evapo-
rative zone depth and in the manner in which the program limits the plant
transpiration for low soil moisture.  Version 1 suggested conservative
values for humid regions while Version 2 suggests typical values (unre-
stricted by poor capillarity of coarse-grained materials and not enhanced by
exceptionally good capillarity such as that of compacted clay) based on
level of vegetation and geographical area.  Version 2 restricts the plant
transpiration by limiting the maximum quantity of water that can be
transpired from a segment of the evaporative zone on a given day as in the
SWRRB model (Williams et al., 1985); Version 1 limited plant transpiration
in the manner of Saxton et al. (1971), Sudar et al. (1981) and Shanholtz and
Lillard (1970).

Unsaturated Vertical Drainage

     HELP models unsaturated vertical drainage using a unit hydraulic
pressure gradient approach (saturated Darcy's law) approach where drainage
occurs at a rate equal to the unsaturated hydraulic conductivity.  Under
this approach vertical water routing is only downward except in the evapora-
tive zone where water is removed upward by evapotranspiration.  Again,
several differences exist between its application in Versions 1 and 2.
Version 1 assumed a linear relationship between soil moisture and unsatu-
rated hydraulic conductivity where the hydraulic conductivity equaled zero
at a soil moisture equaling the field capacity^ and the hydraulic conduc-
tivity equaled the saturated value at a soil moisture equaling the porosity.
In Version 2 the unsaturated hydraulic conductivity is computed by the
Campbell equation using Brooks-Corey soil parameters to the define the shape
of this power function (Brooks and Corey, 1964j Campbell, 1974).  This
approach incorporates the moisture retention properties (capillarity) of the
soil in the determination.  Version 1 stopped drainage at soil moistures
equaling the field capacity while Version 2 stops drainage at wilting point.
Version 3 will consider soil matrix interactions in the selection of the
moisture content where the drainage from one layer into another will cease.
This addition will handle the situation where a fine-grained material at a
higher water content will not drain into a coarse-grained material at a


                                    443

-------
lower water content since the soil suction of the fine-grained material is
much greater than the coarse-grained material.  In addition, Version 1
assumed free drainage down to the liner while Version 2 will not allow
drainage from one layer at a rate greater than the rate that the layer
before can accept.  Hence, free drainage is not assumed in Version 2, allow-
ing placement of a lower permeability non-liner layer below a layer of
higher permeability.

Percolation through Liners

     HELP assumes percolation (saturated vertical drainage) for all leakage
through soil liners.  Percolation is computed by Darcy's law using the satu-
rated hydraulic conductivity of the liner material.  The head loss gradient
is equal to the average head above the base of the liner divided by the
thickness of the liner.

Leakage through Flexible Membrane Liners

     Leakage through flexible membrane liners (FML) is modeled as a reduc-
tion of the cross-sectional area of flow through the subsoil below the FML.
The rate of flow through the leaking subsoil is computed as the percolation
rate through a saturated soil liner.  This method provides good results for
composite liners but is not very good for just a FML.  Therefore, Version 3
will include an improved leakage model for FML based of the work of Brown
(1987) and Giroud et al. (1989a; 1989b; 1989c).

Saturated Lateral Drainage

     Both versions model lateral drainage using a steady-state analytical
approximation of the numerical solution of the Boussinesq equation (Darcy's
law for saturated lateral flow through porous media coupled with the contin-
uity equation).  However, different analytical approximations are used based
on different approaches in their development.  Version 1 uses a linearized
Boussinesq equation as developed by Skaggs (1983) that was corrected to use
average depth of saturation in the formulation instead of depth of satura-
tion at the crest of the drainage layer or the peak depth along the drain
slope.  Version 2 converts the Boussinesq equation into a nondimensional
form and solves it for two analytical solutions at the extremes in non-
dimensional average saturated depth.  These two solutions are then fitted
with the same value and slope in an approximation that covers the rest of
the range of non-dimensional depths.  The approximation matched the
numerical steady-state solution of the Boussinesq equation within 1 percent
of the predicted drainage rate.  The solution is not linear; therefore,
Version 2 uses a Newton-Raphson method to converge onto the solution of the
nonlinear approximation.  Average depth of saturation is used in the model
since HELP is quasi-two dimensional, composed of two one-dimensional
drainage models (one vertical and the other lateral) coupled at the inter-
face between the liner and the lateral drainage layer.

VEGETATIVE GROWTH

     Version 1 of the HELP model computed daily leaf area index of actively


                                    444

-------
transpiring biomass from interpolating a 10-point description from input of
the LAI as a function of growing season.  The vegetative cover during the
dormant season was modeled as 60 percent of the maximum active LAI.

     Version 2 of the HELP model accounts for seasonal variation in active
and dormant above ground biomass and leaf area index through a general
vegetative growth model.  This model was extracted from the SWKRB model
developed by the USDA ARS (Williams et al., 1985).  The vegetative growth
model computes daily values of biomass and leaf area index based on a
maximum allowable value from input, daily temperature and solar radiation
data, mean monthly temperatures  and the beginning and ending dates of the
growing season.  The maximum value of leaf area index depends on the type of
vegetation, soil fertility, climate and management factors.  The program
supplies typical values for selected covers; these range from 0 for bare
ground to 5.0 for an excellent stand of grass.  HELP maintains a data file
containing mean monthly temperatures and beginning and ending dates of the
growing season for 184 locations in the United States.  Vegetative growth is
a linear function of the available solar radiation during the first 75
percent of the growing season.  Growth can be limited by temperatures below
50 degrees F and low soil moisture.  Vegetative decay is modeled as
exponential decay and is also a function of temperature and soil moisture.
The decay process is modeled continuously, both during the active growing
and dormant seasons.

SYNTHETIC WEATHER GENERATION

     HELP incorporates a routine for generating synthetic daily values of
precipitation, solar radiation and mean temperature.  This routine is part
of the WGEN computer program for weather generation developed by the USDA
ARS (Richardson and Wright, 1984) based on a procedure described by Richard-
son (1981).  The HELP user has the option of generating synthetic daily
precipitation data rather than using default or manually entered historical
data.  Regardless of which precipitation input option is chosen, the program
generates synthetic daily values of mean temperature and solar radiation.
The generating routine is designed to preserve the dependence in time, the
serial correlations and cross-correlation between variables and the seasonal
characteristics in actual weather data at the specified location.

Precipitation

     Daily precipitation is generated using a Markov chain model and a two
parameter gamma distribution model.  The first-order Markov chain model
generates the occurrence of wet or dry days.  In this model, the probability
of rain on a given day is conditioned on the wet or dry status of the previ-
ous day.  A wet day is defined as a-day with 0.01 inch of rain or more.  The
model requires two transition probabilities:  P(W/W), the probability of a
wet day today given yesterday was wet; and P(W/D), the probability of today
being wet given yesterday was dry.  Monthly values of these probabilities
are stored in a data file for 139 cities in the continental U.S.

     A two-parameter gamma distribution function is used to generate the
precipitation amount on wet days.  The two-parameter gamma distribution


                                    445

-------
describes the distribution of daily rainfall amounts.  Monthly values of the
two parameters are also stored in a data file.  The coefficients were
obtained from the WGEN program documentation (Richardson and Wright, 1984).
Normal mean monthly rainfalls are being added to the data file to improve
the generation of daily rainfall in Version 3.

Temperature

     WGEN generates daily values of mean temperature using a weakly station-
ary normal distribution function.  It assumes that daily mean temperatures
vary in annual harmonic cycle (hence, weakly stationary) with the actual
values distributed normally about the daily position on the harmonic cycle.
Nine temperature coefficients and normal mean monthly temperatures are used
to describe the distribution:  the mean of maximum temperatures on dry days,
the mean of maximum temperatures on wet days, the mean of daily minimum
temperatures, the amplitude of the harmonic cycle of daily maximum tempera-
tures, the amplitude of the harmonic cycle of daily minimum temperatures,
mean of coefficient of variation of daily maximum temperatures, mean of
coefficient of variation of daily minimum temperatures, amplitude of coef-
ficient of variation of daily maximum temperatures and amplitude of coef-
ficient of variation of daily minimum temperatures.  These coefficients are
stored in a data file for 184 cities in the U.S. along with the normal mean
monthly temperatures.

Solar Radiation

     WGEN also generated daily values of solar radiation Using a  weakly
stationary normal distribution function.  It assumes that daily solar radia-
tion values vary in annual harmonic cycle (hence, weakly stationary) with
the actual values distributed normally about the daily position on the
harmonic cycle.  Three solar radiation coefficients and the station latitude
are used to describe the distribution:  the mean of solar radiation values
on dry days, the mean of solar radiation values on wet days and the ampli-
tude of the harmonic cycle of daily solar radiation values on dry days.
These coefficients are also stored in a data file for 184 cities In the U.S.
along with the station latitude.

SOIL CHARACTERISTICS

     HELP makes use of many different soil characteristics.  Three soil
characteristics used throughout the program are porosity, field capacity and
wilting point.  Saturated hydraulic conductivity is used in computing
vertical drainage, lateral drainage and liner percolation.  The porosity
used here is an effective value, defined as the volumetric water content at
saturation (volume of water per unit bulk volume of material).  Field capac-
ity is defined conceptually as the water content that occurs after a pro-
longed period of gravity drainage.  Wilting point is defined conceptually as
the lowest water content that can be achieved by plant transpiration.  Field
capacity and wilting point are defined more precisely as the volumetric
water contents at capillary pressures of 1/3 bar and 15 bars, respectively.
These four parameters can be specified directly by the user or selected from
a table of 18 default soil textures.  The table of default soil textures and


                                    446

-------
their soil characteristics values were revised in Version 2 based on data
compiled by lawls et al. (1982).  The default values are mean values for
5350 horizons of 1323 soils from 32 states.

     Version 3 will also require the water content corresponding to the
residual saturation as input for each material.  In addition, the user will
have the option of specifying the other Brooks-Corey parameters, pore size
distribution index and the bubbling pressure, in lieu of the field capacity
and the wilting point.  The porosity will still be required.

     Other soil characteristics are used for specific purposes.  The soil
evaporation calculation makes use of an evaporation coefficient.  This coef-
ficient Indicates the ease with which water can be drawn upward through the
soil by evaporation.  Version 2 computes this value using Ritchie's rela-
tionship (1972) while Version 1 obtains it from input.  The SCS runoff curve
number based on both soil characteristics and vegetative cover Is Input that
is optional when the soil characteristics are defined from  the  table  Of
default soil textures.

SOIL MOISTURE INITIALIZATION

     Version 1 Initialized the soil moisture of all layers below the evapo-
rative zone to equal the field capacity.  All layers In the evaporative zone
were initialized at the average of the field capacity and the wilting point.
All liners were assumed to be saturated.  The user had no option of initial-
izing the water content of the layers.

     Version 2 provides the user an option of Initializing the water content
of all layers except liners which are assumed to be saturated.  If the user
elects to allow the program to initialize the water content , the program
will attempt to initialize the water content of the layers to their steady-
state values.
                                CONCLUSIONS

     Revisions have been made In the modeling of virtually every process
described by the model and in the much of the input data.  Many of the
changes are minor in nature but others are very significant.  The effects of
these changes may be insignificant for some designs and for some climates
but quite significant for others.  Comparisons have been made between the
two versions for a total of 20 cells located in Wisconsin, California and
fentucky and are currently available in draft form.  The water budgets gener-
ated by the two versions for many of the cells were similar but several were
different.  In particular, Version 2 should improve evaluations in arid and
semiarid regions and during the winter in the northern regions.

     Version 2 also greatly expands the capabilities of the HELP model.  It
permits the use of one more liner  and three more layers and allows the use
of larger slopes and longer drainage lengths for lateral drainage.  It also
increases the range of saturated hydraulic conductivity that can be used in
the model to include values for the largest drainage media or synthetic


                                    447

-------
drain nets and tightest clays.  The program includes synthetic weather
generation and modeling of vegetative growth to simplify data requirements.
Version 3 will also adjust the runoff based on surface slope and provide
better modeling of all designs employing flexible membrane liners.
                                 REFERENCES

Brooks, R.H. and A.T. Corey.  1964.  Hydraulic Properties of Porous Media.
Hydrology Paper No. 3, Colorado State University.  27 pp.

Brown, K»W.» J.C. Thomas, R.L. Lytton, P. Jayawikrama, and S.C. Bahrt.  1987.
Quantification of Leak Rates Through Holes in Landfill Liners.  EPA/600/S2-
87-062.  U.S. Environmental Protection Agency, Cincinnati, OH.  151 pp.

Campbell, G.S. 1974.  A Simple Method for Determining Unsaturated Hydraulic
Conductivity from Moisture Retention Data.  Soil Science, Vol. 117, No. 6,
pp. 311-314.

Giroud, J.P. and R. Bonaparte.  1989a.  Leakage through Liners Constructed
with Geonembranes, Part I:  Geomembrane Liners.  Geotextiles and Geomem-
branes. Vol. 8, No. 1, pp. 27-67.

Giroud, J.P, and R. Bonaparte.  1989b.  Leakage through Liners Constructed
with Geomembranes, Part II:  Composite Liners.  Geotextiles and Geomem-
branee, Vol. 8, No. 2, pp. 71-111.

Giroud, J.P., A. Khatami and K. Badu-Tweneboah.  1989c.  Evaluation of the
Rate of Leakage through Composite Liners.  Geotextiles and Geomembranes,
Vol. 8, pp. 337-340.

Horton, R.I. 1919.  Rainfall Interception.  Monthly Weather Review, U.S.
Weather Bureau, Vol. 47, No. 9.  pp. 603-623.

Knisel, W.G., Editor.  1980.  CREAMS, A Field Scale Model for Chemical
Runoff and Erosion from Agricultural Management Systems.  Vols. I, II, and
III, USDA-SEA-AR, Conservation Research Report 26.  643 pp.

Perrier, E.R. and A.C. Gibson.  1980.  Hydrologic Simulation on Solid Waste
Disposal Sites.  EPA-SW-868, U.S. Environmental Protection Agency,
Cincinnati, OH. Ill pp.

Rawls, W.J., D.L. Brakensiek, and K.E. Saxton.  1982.  Estimation of Soil
Water Properties.  Transactions of the American Society of Civil Engineers.
pp. 1316-1320.

Richardson, C.W.  1981.  Stochastic Simulation of Daily Precipitation,
Temperature, and Solar Radiation.  Water Resources Research, Vol. 17, No. 1,
pp. 182-190.

Richardson, C.W., and D.A. Wright.  1984.  WGEN: A Model for Generating
Daily Weather Variables.  ARS-8, Agricultural Research Service, USDA.  83 pp.


                                    44S

-------
Ritchie, J.T. 1972.  A Model for Predicting Evaporation from a Row Crop with
Incomplete Cover.  Water Resources Research, Vol. 8, No. 5,  pp. 1204-1213.

Saxton, K.E., H.P, Johnson, and R.H. Shaw.  1974.  Modeling Evapotranspir-
ation and Soil Moisture.  Transactions of American Society of Agricul-
tural Engineers, Vol. 17, No. 4.  pp. 673-677.

Schroeder, P.R. and A.C. Gibson,  1982.  Supporting Documentation for the
Hydrologic Simulation Model for Estimating Percolation at Solid Waste
Disposal Sites (HSSWDS).  Draft Report, U.S. Environmental Protection
Agency, Cincinnati, OH.  153 pp.

Schroeder, P.R., J.M. Morgan, T.M. Walski, and A.C. Gibson.  1984a.
Hydrologic Evaluation of Landfill Performance (HELP) Model:  Vol. I.  User's
Guide for Version 1.  EPA/530-SW-84-009,  U.S. Environmental Protection
Agency, Washington, DC.  120 pp.

Schroeder, P.R., A.C. Gibson and M.D. Smolen.  1984b.  Hydrologic Evaluation
of Landfill Performance (HELP) Models  Vol. II.  Documentation for Version 1.
EPA/530-SW-84-010,  U.S. Environmental Protection Agency, Washington, DC.
256 pp.

Schroeder, P.R., and R,L. Peyton.  1987a.  Verification of the Hydrologic
Evaluation of Landfill Performance (HELP) Model Using Field Data.  EPA 600/2-
87-050.  U.S. Environmental Protection Agency, Cincinnati, OH.  163 pp.

Schroeder, P.R., and R.L. Peyton.  1987b.  Verification of the Lateral
Drainage Component of the HELP Model Using Physical Models.  EPA 600/2-87-
049.  U.S. Environmental Protection Agency, Cincinnati, OH.  117 pp.

Shanholtz, V.O., and J.B. Lillard.  1970.  A Soil Water Model for Two
Contrasting Tillage Practices.  Bulletin 38, Virginia Water Resources
Research Center, VPISU, Blacksburg, VA.  217 pp.

Skaggs, R.W.  1983.  Modification to DRAINMOD to Consider Drainage from and
Seepage through a Landfill.  Draft Report, U.S. Environmental Protection
Agency, Cincinnati, OH.  28 pp.

Sudar, R.A., K.E. Saxton, and R.G. Spomer.  1981.  A Predictive Model of
Water Stress in Corn and Soybeans.  Transactions of American Society of
Agricultural Engineers, Vol. 24, No. 1.  pp. 97-102.

USDA, Soil Conservation Service.  1972.  National Engineering Handbook,
Section 4, Hydrology.  U.S. Government Printing Office, Washington, D.C.
631 pp.

Williams, J.R., A.D. Nicks, and J.G. Arnold.  1985.  SWRRB, A Simulator for
Water Resources in Rural Basins.  Journal of Hydraulic Engineering, ASCS,
Vol. Ill, No. 6, pp. 970-986.
                                    449

-------
         WASTE MINIMIZATION  ASSESSMENTS  AT  SELECTED  POD  FACILITIES

           by:   James S. Bridges
                 RREL, U. S.  Environmental Protection Agency
                 Cincinnati,  Ohio 45268
                                  ABSTRACT


     The Waste Reduction Evaluations at Federal Sites (WREAFS) Program
consists of a series of evaluations and demonstrations for pollution
prevention and waste reduction conducted cooperatively by the Environmental
Protection Agency (EPA) and other Federal agencies.  The objectives of the
WREAFS Program include: (1) performing waste minimization opportunity
assessments, (2) demonstrating waste minimization techniques or
technologies at Federal facilities, (3) conducting waste minimization
workshops, and (4) enhancing waste minimization benefits within the Federal
community.

     This paper describes the WREAFS Program support of DOD activities in
the pollution prevention area, and provides an overview of current projects
being conducted at the Philadelphia Naval Shipyard, Fort Riley (Kansas)
Army Forces Command, and the Naval Undersea Warfare Engineering Station in
Keyport, Washington.  Also described is a waste minimization research
project with the Air Force.  These DOD waste minimization opportunity
assessments have identified waste minimization opportunities for a range of
industrial and military operations including:  metal cleaning, solvent
degreasing, spray painting, vehicle and battery repair, ship bilge
cleaning, and torpedo overhaul.  The resultant waste minimization
recommendations are source reduction methods including technology, process,
and procedural changes and recycling methods, which focus on reuse or
recycling.

     This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved
for presentation and publication.
                                     450

-------
                                INTRODUCTION


     The purposes of the WREAFS Program are to identify new technologies
and techniques for reducing wastes from industrial processes used by
federal agencies and to enhance the adoption of pollution prevention/waste
minimization (WM) through technology transfer.  New techniques and
technologies for reducing waste generation are identified through WM
opportunity assessments and are evaluated through joint research,
development, and demonstration (RD&D) projects.  The information and data
from these projects are then provided to both the public and private
sectors through various technology transfer mechanisms, including project
reports, project summaries, conference presentations, and workshops.

     The waste minimization opportunity assessments are conducted by an
assessment team that is composed of personnel from EPA, staff from the
Federal facility that is cooperating in the program, and EPA's contractors.
The assessments follow the procedures described in the EPA report Haste
Minimization Opportunity Assessment Manual.  This manual provides a
systematic planned procedure for identifying ways to reduce or eliminate
waste.  The development of this procedure was supported by the Risk
Reduction Engineering Laboratory, U.S. Environmental Protection Agency,
Cincinnati, Ohio.

     As a result of joint waste minimization opportunity assessments
(WMOA's), RD&D projects are identified with recommendations for pollution
prevention.  The demonstration projects are conducted under inter-agency
agreements with joint funding by EPA and the cooperating Federal agency.
Waste minimization workshops and other technology transfer methods are
being used to communicate the results of these projects to the Federal
community and the private sector.

WREAFS  PROGRAM PROCEDURES

     The WREAFS Program procedures used for conducting the waste
minimization assessments are closely related to the WM procedures presented
in the  Manual.  Figure 1 describes the course followed by a typical WREAFS
project.  The assessments consist of four major phases:

(1)  Planning and Organization - organization and goal setting;

(2)  Assessment - careful review of a facility's operations and
     wastestreams and the identification and screening of potential options
     to minimize waste;

(3)  Feasibility Analysis - evaluation of the technical and economic
     feasibility of the options selected and subsequent ranking of options;

(4)  Implementation - procurement, installation,  implementation, and
     evaluation.
                                     451

-------
                          PLANNING  AND ORGANIZATION
                        • Asstflt(tTe«m eonjJifng of fadSly staff.
                          EPA personnel, and EPA confractjf• tfcff,

                        • CondwSWWmetSngatfct federal «ilt
                          to khniry overtl Buessmert goals.

                        • ktenlfy ti* rotes of tft*Teorn members.

                        • Develop tssJgn™nti forTeam members tor
                         eavch major operating unit or fadMy i
                        • Ravin* th» Inbnmfon and data requirements
                          of fttWMfcsheetj and maX* revision* to
                          tailor 8t*k>nn* to ft* current pro]eci
                        • Compieta fi* Worksheets related to piarrtag
                          aixJorgarfzafcn.

                        * Aqutre background Mmmaflon.
                                                                 ASSESSMENT PHASE
                                            ASSESSMENT  SURVEY
                                      • Inspect the « Select the options that bvor 8rt needs ol the
 fecKty for tve fsxuiblily enaJysii phaso.
01
to
                                       FESIBItmr  ANALYSIS PHASE
• Detomrine the technical feasibility o! the selected options

• Determine tht econorric feasibility of the selected options

* Determine and compare tie return on investment of the
 various options.
                                  IDENTIFICATION  OF RESEARCH  OPTIONS
                                 • Review th» oploni and identify specific research
                                   opporbrities.
                                 PREPARATION  OF WM  ASSESSMENT REPORT
                                                                                               DEMONSTRATION  AND EVALUATION PROJECTS
                                                                                                   • Develop tnferagency agreement with joint funding

                                                                                                   • Prepare design of selected WM option and procure
                                                                                                    needed equipment and supplies

                                                                                                   • Implement demonstration of WM option and monitor
                                                                                                   • Evaluate he results of Hit demonstration
                                                                                                        PREPARATION  OF  TECHNOLOGY
                                                                                                            TRANSFER MECHANISMS
                                                                  • Reports

                                                                  • Conference and seminar papers
                                                             Figure 1, An Overview of the WREAFS Program

-------
     Many of the WH opportunities identified during WREAFS projects involve
low-cost changes to equipment and procedures that may be presently employed
at other Federal facilities or within private industry.  These WM
opportunities can often be implemented by the facility without extensive
engineering evaluations.  Some other WM opportunities identified during
these projects will require further study before full implementation can be
realized.  Typically, opportunities requiring further evaluation are those
that have the potential for affecting the process and/or require the use of
new procedures or equipment.  In such cases it may be necessary to conduct
demonstration projects.

Demonstration and Evaluation Projects

     The types of research projects that can be pursued under the WREAFS
Program are those that are expected to advance the knowledge and practice
of waste minimization technologies and methods, and have broad
applicability to Federal facilities and private industry.  Depending on the
nature and state of development of the WM option selected for demonstration
and evaluation, these projects may include:  (1) process design, (2)
detailed design and specification, (3) system procurement, (4) installation
and start-up, (5) monitoring, and (6) reporting.  Some projects may require
bench-scale and/or pilot testing prior to or as a part of the demonstration
project.  Other projects may utilize full-scale equipment directly on the
production line.

                              CURRENT PROJECTS


     Four WREAFS projects are being performed at DOD facilities.  These
projects are in various stages of completion.  The assessment survey has
been completed for several projects and the waste minimization options and
research opportunities have been identified.  Other projects are just
underway.  The following is a description of each project.

PHILADELPHIA NAVAL SHIPYARD

     This project  is being conducted in cooperation with the Environment,
Safety and Health Office of the Philadelphia Naval Shipyard (PNSY).  The
shipyard has an ongoing program for waste minimization.  With their
guidance, several  industrial operations were selected for application of
the new waste minimization procedures.  The shipyard plans to utilize the
results from this project as a guidance tool for evaluating waste
minimization opportunities at other industrial activities at the PNSY.  The
results of this project will be particularly applicable to facilities that
operate aqueous cleaning and spray painting processes.  However, the
procedures employed to  identify and evaluate minimization alternatives are
applicable to most industrial operations.

Facility Description

     The Philadelphia Naval Shipyard is the nation's oldest continuously
operating naval shipyard.  The shipyard now specializes in revitalizing and


                                     453

-------
repairing ships already in fleet.  The Service Life Extension Program
(SLEP) is the shipyard's largest program.  It is a comprehensive, keel-up
restoration and modernization overhaul intended to extend the life of
aircraft carriers by half, at about one third of the cost of a new carrier.

Selected Areas and the MM Options Evaluations

     The procedure used for conducting the WM evaluation followed the EPA
Haste Minimization Assessment Manual.  The industrial activities selected
for this project include areas used for the fabrication and surface coating
of aluminum products, spray painting operations for steel parts, and a
citric acid derusting operation, which is located at the drydocks.

     The aluminum cleaning operation is performed to remove oil and other
materials from the surfaces of aluminum sheets prior to welding.  The
cleaning line consists of four tanks:  two process tanks, and two rinse
tanks.  The cleaning procedure consists of loading aluminum sheets into a
metal basket, hoisting the basket into a process tank, and rinsing in one
of the rinse tanks.  The process water becomes diluted after repeated
operation due to dragout losses and tap water replenishment.  These tanks
also collect floating oil, and the solution becomes contaminated with
suspended solids.  After approximately three months of operation, the
process tanks are pumped to a tank truck and hauled by the contractor for
disposal.  The rinse tanks are operated as non-flowing rinses because of
the low pH of the rinse water and the lack of facilities for
neutralization.  The rinse tanks are disposed of in the same manner as the
process tanks but on a more frequent schedule, usually every two weeks

     Two WM options were evaluated for the aluminum cleaning operation.
First, a dragout reduction and bath maintenance system was proposed.  This
system will include a hand-held spray rinse that will be applied over the
process tanks.  After affecting the parts, the rinse water will drip into
the process tank.  The spray rinse is expected to return 90 percent of the
dragout back to the process tank.  And 2, the acid baths accumulate oil and
solids from the parts, and therefore, returning dragout losses may cause
the process tanks to accumulate contaminants at a faster pace.  These
contaminants may interfere with the cleaning process; thus a bath
maintenance system was recommended.  This system would include an oil
skimmer for floating oil removal and a cartridge filter for suspended
solids removal.  The dragout reduction and bath maintenance systems are
expected to extend the usable life of the baths to one year.

     A 2-stage rinse was proposed as an alternative to the existing rinse
arrangement, where only one rinse tank is used after cleaning.  The 2-stage
rinse would make use of the existing tanks; however, some rearrangement was
proposed to improve the layout of equipment and the efficiency of the
operation.  Using a 2-stage rinse would reduce the frequency of discarding
the rinse water by allowing the first rinse to become more heavily
contaminated.  Then, after the first rinse is discarded and refilled with
fresh water, the sequence of rinsing is changed (i.e., the cleanest rinse
would always be the second rinse).
                                    454

-------
     The spray painting processes are used for small and medium-sized
aluminum parts, steel parts, and columns.  Aluminum parts are ciegreased by
wiping with rags that have been dipped into xylene.  The parts are then
spray painted in a water curtain booth.  The painting process typically
consists of a zinc chromate primer, air drying, a final enamel paint
coating, and air drying.  A new booth water chemical system was; recently
implemented.  Spray painting of steel parts and columns is performed at
various locations.

     Several WM options were evaluated for the two spray painting
operations.  First, an operator training and awareness program option was
evaluated that could reduce the amount of waste paint by controlling the
amount of paint overspray, the amount of unused paint left in the can, and
the amount of paint that is unusable due to partial solidification prior to
use.  Second, equipment changes were evaluated for the current compressed
air paint spraying systems.  Under consideration is the high-volume,
low-pressure (HVLP) method.  The transfer efficiency of HVLP is; reportedly
in the range of 65 to 90 percent.  Compressed air equipment typically
provides a 40 percent or lower transfer efficiency.  Other WM options
evaluated during the assessment included expanding use of a new paint booth
chemical system and use of sludge dewatering equipment that would reduce
the volume of paint sludge and recycle booth water.

     PNSY employs a chemical cleaning process for ships' tanks, bilges, and
void spaces termed the citric acid process.  It is generally performed
while ships are in drydock.  This process is relatively new and it replaces
the mechanical methods of cleaning and derusting metal surfaces.  The
process generates a spent citric acid/triethanolamine (TEA) solution.
Owing to the significant material input costs and the high disposal costs,
a recovery option was considered during the project.

Results at PNSY

     The results of the assessment, which are presented in Table 1,
indicated that the best options in terms of cost savings are the awareness
and training program for paint waste reduction and the changes to the
aluminum cleaning line including dragout reduction, bath maintenance, and
improved rinsing.  These three options offer a combined net savings in
operating costs of $158,680 per year.

     The citric acid/TEA recovery option was identified as a potential
research project,  A preliminary process design was prepared for the
recovery process that included an electrodialytic membrane unit for
separation and removal of dissolved metals.  This type of technology has
been applied to similar chemical solutions; however, its application to
this waste has not been previously demonstrated.

FORT RILEY, KANSAS

     Fort Riley is a permanent U.S. Army Forces Command (FORSCOM)
installation that provides support and training facilities for the 1st
Infantry Division (Mechanized), Non-Divisional Units, and tenant

                                     455

-------
                              TABLE  1.  PHILADELPHIA NAVAL SHIPYARD SUMMARY
                                     OF WM FEASIBILITY ANALYSIS PHASE
Location Process & Waste
Building 990
Aluminum Cleaning
Spent KRC-7X
Spray Painting of Aluminum
Paint Sludge

Used Paint Thinner
Unused Paint
Building 1028
Spray Painting of Steel
Paint Sludge

WM Options

Bath Maintenance
Two Stage Rinse
Booth Chemicals
Paint Sludge Dewat.
Awareness & Train.
Awareness & Train.

Booth Chemicals
Paint Sludge Dewat.
Nature
of
WM Option

Equipment
Equipment
Materials
Equipment
Personnel/
proced .
Personnel/
proced .

Materials
Equipment
Total Cap.
Investment
$

$12,200
3,116
12,190
9,550
24,266
Net Op.
Cost
Savings
$/yr

$44,190
34,590
5,430
3,840
79,900
Payback
Period
yr

0.3
0.1
2.3
2.5
0.3
Projected
Waste
Reduction
Ib/yr

$ 44,035
190,590
-
15,012
unknown
See used ...
paint thinner

3,300
see bldg 990

5,460
-

0.6
-

27,022
-
Drvdocks
Citric Acid Derusting
Cone. Citric Acid/TEA
ED Recovery System   Equipment
76,050
60,720     1.3
124,241

-------
activities.  It Is a U.S. Government-owned, U.S. Government-operated
facility.  Fort Riley provides the U.S. Army with the capability to house
and train an Army division and associated land combat forces as well as to
service Army functions in an assigned Midwest support area.

Selected Areas and the MM Options Evaluated

     The areas selected for assessment at Fort Riley were the Division
motor pools located in various areas on the post.  Routine mechanical
maintenance is performed in these areas for trucks, tanks, armored
personnel carriers, and other vehicles.  Results of the Fort Riley, Kansas
waste minimization assessment identified two waste reduction opportunities
in a multipurpose building used for automotive subassembly rebuilding, lead
acid battery repair, and other Army maintenance operations.

     Battery acid (32-37 percent sulfuric acid) containing trace
concentrations of lead and cadmium is currently drained from both dead
batteries and batteries requiring repairs, e.g., replacement of battery
terminals, and shipped in 15-gallon drums to a hazardous storage facility
at the installation for ultimate disposal as a hazardous waste.  It is
proposed instead that the waste acid be gathered in a holding tank,
filtered to remove any particulates, and adjusted in concentration to 37
percent sulfuric acid (using 60" Baume commercial sulfuric acid) as needed
for reuse in reconditioned or new batteries.  The buildup of dissolved
metal impurities in this recycling system is prevented by purging part of
the acid from the system.  It is assumed in this assessment that 25 percent
of the acid is purged and 75 percent is reused.

     The dirty aqueous alkaline detergent solution for automotive parts
cleaning, which contains trace concentrations of lead, chromium, and
cadmium at a pH >12 as well as the oils, grease, and dirt removed from the
automotive parts, is currently drained to an on-site nonhazardous waste
evaporation pond.  This waste, heretofore regarded as nonhazardous waste,
is currently being reclassified as a RCRA hazardous waste due to its
characteristics (D002, D006, D007, D008) and will have to be disposed of as
a hazardous waste.  The proposed waste minimization option for this waste
stream involves emulsion breaking to remove the tramp oils, filtration to
remove particulates, and addition of fresh alkaline detergent as necessary,
followed by reuse for automotive parts cleaning.  The buildup of impurities
would be prevented by purging 25 percent of the used alkaline detergent and
recycling 75 percent.

Results  at Fort Rilev

     The waste reduction options identified in this study are recycle/reuse
options.  Presented in Table 2 are the results of the cost analysis phase.
A net savings in operating costs is anticipated to be $149,400 per year.
The expected payback periods for the two waste reduction options identified
are very short.  Successful application of these options at Fort Riley
would create the potential for similar waste minimization options in at
least 10 other U.S. Army FORSGOM installations.
                                     457

-------
           TABLE  2.          OF WASTE MINIMIZATION OPTIONS COST
                    ANALYSIS DATA AT FORT RILEY, KANSAS


Operation
Waste Stream

Waste
Minimization
Options
Total
Capital
Invest.
$
Net
Operating
Cost
Savings, $

Payback
Period,
yr
Projected
Waste
Reduction
Ib/yr
Battery Repair
Shop/Waste
Battery Acid

Automotive
Subassembly
Rebuild/Tornado
Parts Washer
Wastewater
Recycle of       15,200    37,400
waste battery
acid

Recycle of       19,800   112,000
Tornado washer
wastewater
0.41
SI,291
0.18     181,395
NAVAL UNDERSEA WARFARE ENGINEERING STATION, KEYPORT, WASHINGTON

     This project is being conducted in cooperation with the Naval Energy
and Environmental Support Activity (NEESA) of Port Hueneme, California, in
coordination with the Environmental Protection Division (Code 075) of the
Naval Undersea Warfare Engineering Station (NUWES) Staff Civil Engineering
Department.  Several departments at NUWES Keyport are involved in an
ongoing program to further the process of waste minimization on the
Station.  The draft version of this waste minimization assessment is
currently under review by the Naval Undersea Warfare Station and detailed
results of the study are not available.

     NUWES Keyport is located within the central Puget Sound area of
Northwestern Washington State.  In 1978, the facility changed names from
Naval Torpedo Station Keyport to NUWES Keyport recognizing that the
functions of the Station had broadened to include various undersea warfare
weapons and systems engineering and development activities.  The principal
activities currently conducted at NUWES Keyport are the design and testing
of torpedoes.

SelectedAreas and the WH OptionsEvaluated

     Potential sources for waste materials include torpedo as well as other
ordnance handling and related activities.  Specific activities on the
Station include welding, plating, machining, and sheet metal work; painting
and stripping; electrical work; carpentry; fuel storage and use; power
production; pest control; sanitary and industrial wastewater treatment; and
associated storm sewer runoff.  These activities generate a variety of
potentially hazardous wastes, including fuel, oil, coolant, hydraulic
                                     458

-------
fluid, and grease; various metals and plating bath liquids; paint and
thinner; Freon, alcohol, mineral spirits, and other solvents; resin; acids
and caustics; chromate and cyanide salts; wastewater treatment sludge; and
detergent.

     Two areas at the Station were examined in detail during the
assessment.  Both areas are used for the maintenance of torpedoes and have
similar operations, processes, and waste streams.  The major waste
generating activities consist of defueling, disassembling, cleaning,
reassembling, and refueling of torpedoes.  The waste materials at these
areas include:  solids, liquids, sludges, solvents, and oils that are
contaminated with Otto fuel, as well as diethylene glycol (DEC), mineral
spirits (Agitene), and cyanide compounds.  Waste minimization options were
recommended for contaminated solids, contaminated liquids, and waste
mineral spirits.

     The waste minimization option for contaminated solids addresses all
clothing that becomes soiled with Otto fuel (OF).  Recommendations for the
reduction of this waste are to segregate contaminated from non-contaminated
clothing, and only dispose of the contaminated portion of a garment.

     Automated parts cleaning is another waste minimization option
recommended to reduce soiled clothing.  In addition, the increased
efficiency of an automated system will serve to reduce the amount of waste
mineral spirits.

     Contaminated liquids can be reduced in the Otto fuel tank draining
area by installing another automated system that increases efficiency and
decreases spills and contaminated solids.

     At one site location, waste mineral spirits are currently pumped on a
regular schedule from deep sink parts cleaners.  This schedule is followed
whether or not the solvents have become fouled.  By monitoring the solvents
and disposing only when fouled, the amount of waste cleaner will be
reduced.

Preliminary Results at Kevport» Washington

     The waste minimization options under consideration for NUWES Keyport
are being evaluated for technical and economic feasibility.  Projected
waste reductions and costs are not available at this time.

U.S. AIR FORCE

      In support of the Department of Defense's waste minimization program,
the Air Force is seeking to obtain information for its chlorinated solvents
recycling program.  Air Force facilities are high-volume consumers of
industrial solvents.  Applications for solvents range from degreasing
aircraft parts and missile guidance systems to cleaning small bearings and
armament material.  The major chlorinated degreasing solvent currently used
by the Air Force is 1,1,1-trichloroethane.  Due to toxicity arid inherent
health effects, the use of all chlorinated solvents may be curtailed unless

                                     4SS

-------
effective waste reduction and recycling programs are developed.  In this
effort, an assessment will be performed for the recycling program at a
selected Air Force base.  The project is being performed in cooperation
with Auburn University.  Objectives for the study include developing
mitigation technology for minimizing the accumulation of toxic contaminants
in the recycled product and addressing concerns regarding solvent
inhibitors and additives.  The results of the study will be used to
formulate a model program and establish protocols for operating solvent
recycling technology service programs.

                          SUMMARY AND CONCLUSIONS


     Waste minimization opportunity assessments conducted under EPA's
WREAFS program are an excellent mechanism for promoting pollution
prevention research at DOD facilities.  At the three DOD sites audited,
waste minimization opportunities have been identified for the following
hazardous waste streams:  aluminum cleaning solutions, paint sludge, paint
thinner, unused paint, citric acid derusting solutions, waste battery acid,
parts washer wastewaters, contaminated clothing, parts cleaning solutions,
waste Otto fuel, and waste mineral spirits.

     Pollution prevention recommendations have been made for the
Philadelphia Naval Shipyard and Fort Riley that will reduce hazardous waste
generation by a minimum of more than 630,000 pounds per year, with an
annual savings of $383,530 at the two sites.  In addition, a number of
these opportunities will make excellent demonstration and evaluation
projects.

     The WREAFS program can be used as a catalyst to implement pollution
prevention strategies and to identify research opportunities that would not
be otherwise identified.  Through this program EPA can assist DOD in
transmitting new technical information to all of its facilities through
project reports, workshops, seminars, and conferences.
                                     460

-------
  CHEMICAL SUBSTITUTION  FOR 1.1.1-TRICHLOROETHANE AMD METHANOL
                   IN MANUFACTURING OPERATIONS


     by:                  Lisa M. Brown
                         Johnny Springer
              U.S.  Environmental  Protection Agency
              Risk Reduction Engineering Laboratory
                         Cincinnati,  Ohio

                          Matthew Bower
                       APS Materials, Inc.
                           Dayton, Ohio


                             ABSTRACT


     Hazardous wastes are generated from cold solvent degreasing
operations used in many industrial processes.   The spent solvents
are managed under Subtitle C of the Resource Conservation and
Recovery Act (RCRA).  With the land ban of spent solvents,
disposal has become increasingly difficult.  As a result,
industries began investigating ways to avoid using RCRA-listed
cleaning solvents.  EPA's Pollution Prevention Research Branch
along with APS Materials, Inc., a small metal finishing company,
participated in a joint research project to evaluate the
substitution of a dilute, terpene-based cleaner for 1,1,1-
trichloroethane (TCA) and methanol, hazardous wastes F001 and
F003 respectively, in their degreasing operations.

     This paper presents the results of a study evaluating the
waste reduction/pollution prevention that can be achieved by
substituting dilute limonene solutions for TCA and methanol in
the cleaning of orthopedic implants  (e.g.  metal knee and hip
replacements).  This paper describes the original cleaning
process, the modifications made to the process in using the
dilute limonene solution, and the sampling plan used in
evaluating the effectiveness of the solution.   The paper presents
qualitative results of the sampling tests and an economic
evaluation of plant modifications.

     This paper has been reviewed in accordance with the U.S.
Environmental Protection Agency's peer and administrative review
policies and approved for presentation and publication.   Mention
of trade names of commercial products does not constitute
endorsement or recommendation for use.

                               461

-------
                          INTRODUCTION


     Passage of the 1984 Hazardous and Solid Waste Amendments
(HSWA) to the Resource Conservation and Recovery Act (RCRA) of
1976 has redirected the U.S. environmental policy towards waste
minimization to improve the quality of the environment.    In its
efforts to pursue the objectives set forth by Congress in the
HSWA  to RCRA, the USEPA has established a national comprehensive
pollution prevention program that includes information gathering,
research and development, demonstration, support of state and
local government pollution prevention programs, training and
education, technology transfer activities, pollution prevention
assessments, and extensive communications with universities and
the general public.  Implementation of programs to achieve
several of these objectives is accomplished through research
conducted by the Pollution Prevention Research Branch of the Risk
Reduction Engineering Laboratory.  This research addresses the
intent of the Amendments to reduce the release and transport of
hazardous, toxic, and nonhazardous materials through the air,
water and solid media.  The research is of significant benefit to
the USEPA, states, waste generators, and the general public since
results of this research will assist in reducing the generation
of pollutants that threaten both public health and the
environment.  The principal goal of the Pollution Prevention
Research Branch is to encourage the identification, development,
and demonstration of processes and techniques that result in less
waste being generated in order to promote a more rapid
introduction of effective pollution prevention techniques into
broad commercial practice.

     1,1,1-trichloroethane (TCA) is used ms a cold solvent
degreasing agent in many industrial degreasing processes.  In
1986,  TCA  was identified as a hazardous waste that must be
managed under Subtitle C of the Resource Conservation and
Recovery Act.  As a result of this action, industries began
looking for ways to avoid the use of TCA cleaning solvents. The
EPA decided to target the metal finishing industry for
participation in a joint research project to examine the
possibility of substituting a terpene-based cleaner for TCA in
degreasing operations.  APS Materials, Inc., a facility in
Dayton, Ohio participated in the research project.  APS
Materials, Inc. is a metal parts finishing company which
generates TCA and methanol waste from cold solvent degreasing
operations associated with their plasma spray deposition process.
                               462

-------
                 PLASMA SPRAY DEPOSITION PROCESS

     The plasma spray deposition process has emerged as a major
means to apply a wide range of materials on diverse substrates.
The deposition process is accomplished with the use of a plasma
gun.  In the plasma gun, an electric arc is formed between
positive and negative electrodes via an electric discharge
initiated by direct current.  The discharge gives rise to a
breakdown of the dielectric nature of the gas, making it
conductive.  The net result is a gaseous collection of energetic
electrons and ionized molecules known as a plasma.  The plasma
exits as a high velocity flame through the nozzle of the gun.  A
powdered feedstock is injected into the flame via a carrier gas
(usually argon).  The injected powder accelerates, melts, and is
carried at sonic velocities to the substrate on which the
particles impact and solidify rapidly, at rates about one million
degrees per second, building a well adhered protective
coating.(1)

     While APS Materials, Inc. employs the fundamental plasma
spray deposition process, a few changes were made to better
accommodate the plasma spray work performed by their company.
First, APS Materials performs its plasma spraying in an inert
atmosphere chamber.  This is done for cooling and to prevent the
titanium powder used in many of its coating applications from
becoming oxidized thus forming brittle coatings.  APS Materials
also uses helium in the spray gun as a mix gas and to adjust the
heat level and arc length.

     Typically, the plasma spray deposition process requires only
a small amount of substrate preparation.  However, because APS
Materials is involved in plasma spraying parts that must perform
in such hostile environments as aircraft engines  (aircraft parts)
and the human body (orthopedic implants), they must be assured
that the plasma sprayed coating is securely adhered to the
substrate.  For this reason, parts that arrive at APS for coating
undergo a thorough cleaning process prior to the application of
the plasma spray coat.


                       PROCESS DESCRIPTION
ORIGINAL PROCESS

     In the APS biomedical parts division, the company primarily
coats cobalt/molybdenum parts and titanium parts with a porous
titanium alloy.  By using plasma spray technology, the porosity
of the coating is controlled so that growing bone will attach to
the metal surface.  In order to achieve a strong and adhesive
coating, the parts were cleaned with TCA or methanol (TCA for
cobalt/molybdenum and methanol for titanium).  TCA is more
economical than methanol but weakens titanium over time.  The
cleaning process consists of several steps.  Initially, the parts

                               463

-------
received undergo a visual inspection for any gross defects.  The
parts are then partially masked with tape, exposing only the
surfaces that will receive spray coating.  Next, they are grit
blasted to roughen the surface of the part for the application of
the spray coat.  After the grit blast has been completed, the
masking tape is removed.  The part is then immersed in a small
pail containing TCA or methanol.  The pail is placed in an
ultrasonic bath containing warm water for 15 minutes.  Solids
from grit blasting, oil and grease from the manufacturing and
handling of the parts, and any adhesive residuals from the
masking tape are removed in this cleaning process.  After the
ultrasonic bath, a graphite masking suspension is applied to the
part on surfaces where the plasma spray coating is not wanted.
The part is then plasma sprayed and cleaned again to remove
excess titanium and the graphite mask.

     As a check system, APS runs small one-inch diameter disks of
the same composition as the part to be coated - called "test
buttons" - through the same cleaning and coating process.  The
test buttons are placed on a tensile strength testing machine
which measures the tension required to separate the coating from
the substrate as a quality control measure.

     Many wastes are generated during the preparation of the part
for spray coating, with TCA and methanol being the wastes of
primary concern.  Waste TCA and methanol were being generated at
the rate of 1/2 barrel per month.  Disposal of these solvents was
becoming more and more difficult.

DESCRIPTION OF INITIAL BENCH SCALE EXPERIMENTS

     DuSqueeze, a product of DuBois Chemicals, was selected as a
possible substitute for TCA and methanol because of its disposal
qualities.  Disposal of dilute solutions of DuSqueeze could be
accomplished by flushing it to a sanitary or industrial sewer
according to local sewer use permit requirements.  The
feasibility of substituting a dilute, terpene-based cleaner
(DuSqueeze) for TCA and methanol was determined by assessing the
tensile strength of the plasma coating bonds made after cleaning
with DuSqueeze.  Five tests were performed, four on plasma coated
test buttons to assess the tensile strength of bonds made after
cleaning with DuSqueeze as compared to the tensile strength of
bonds made after cleaning with methanol and TCA, and one test to
determine if any limonene remained on the buttons after being
cleaned.  In the first test, four titanium test buttons were
placed in a stainless steel beaker containing a 20:1 dilute
solution of DuSqueeze and water.  The solution was agitated for
20 seconds.  The test buttons were then placed in a stainless
steel beaker containing deionized (DI) water which was agitated
for 20 seconds.  The test buttons were then blow-dried and plasma
sprayed.  The tensile strength of the bond between the plasma arc
coating and the substrate was measured using a Tinius Olsen
tensile tester.

                               464

-------
     In the second test, four titanium buttons were placed in an
ultrasound bath containing a 50:1 dilute solution of DuSqueeze
for 10 minutes.  Nes?t the/ buttons were placed in a stainless
steel beaker containing DI water for thirty seconds.  The
titanium buttons were blow dried for sixty seconds and then
plasma sprayed.  The tensile strength of the bonds were then
tested in the same manner as the first test.  The third test
followed the same procedure as test two, using a 100:1 dilute
solution of DuSqueeze.  In the fourth test the buttons were
cleaned by the same process as the third test, but the buttons
were analyzed for residual limonene and were not plasma sprayed
and tensile tested.  In the fifth test, cobalt/molybdenum buttons
were used instead of the titanium buttons with the test protocol
identical to the third test.

MODIFICATIONS TO EXISTING SYSTEM

     APS purchased a heated ultrasound bath with a timer for the
conversion.  However, when this ultrasound bath malfunctioned, a
heater was added to the old ultrasound bath.  The TCA/methanol
cleaning system did not require a DI water rinse, so a DI water
system was purchased along with a stainless steel bath and
immersion heater.  With the new cleaning system, the parts take
longer to dry, so a heat gun was purchased to speed-up the drying
process.


                      SAMPLING AND ANALYSIS


     The overall purpose of the sampling and analysis project at
APS Materials was to support a purely qualitative judgement of
the cleaning capabilities of the substitute cleaning solution
(i.e., DuSqueeze).  The sampling and analysis protocol for this
project was set up in three parts; sampling spent solutions of
methanol and TCA, sampling the terpene-based cleaning solution
after modifications were made to the cleaning system, and
developing data for a comparative analysis of plasma coating
bond strengths between the coatings of test buttons that were
cleaned with methanol/TCA prior to coating and the coatings of
test buttons that were cleaned with the terpene-based solution
prior to coating.

PRE-MODIFICATION SAMPLING

     The first part of the sampling process was performed prior
to any modifications.  This sampling was performed in order to
determine the type and amounts of contaminants found in the
cleaning solvents.  Samples of the methanol and TCA cleaning
solutions were taken and analyzed for oil and grease, dissolved
solids, suspended solids, titanium metal and cobalt metal.  This
sampling also established the baseline performance for methanol
and TCA.  The samples were taken by mixing the material in a
plastic bucket and then pouring a sample from the bucket through

                               465

-------
a glass funnel into a glass bottle.  The data derived from this
sampling served as a bench mark for the- ensuing DuSqueeze
sampling.

POST-MODIFICATION SAMPLING

     The second part of the sampling scheme was performed after
the modifications were made to the system in order to determine
the effectiveness of the terpene-based solvent (DuSqueeze) in
cleaning the parts.  Sampling of the cleaning solution was
performed throughout a typical operating cycle.  Samples were
recovered at the beginning of a bath cycle (i.e., when the tank
contents were completely replaced with fresh cleaning solution)
to establish baseline concentrations.  A second sample was taken
midway through the effective life of the cleaning solution.  A
final sample was recovered prior to removing the spent solution
from the dip tank.

     One liquid sample was collected during each sampling episode
and split into two aliquots.  One aliquot was placed in a 1000-
ml linear polyethylene bottle with a screw-cap lid.  This sample
was used to analyze for dissolved/suspended solids and the two
specific metals.  The second aliquot was used to test for oil and
grease and was placed in a 1000-ml glass bottle with screw-cap
lid.  Before use, the sample containers were soap-and-water
washed, rinsed thoroughly, and then soaked in acid (nitric acid
for plastic, sulfuric acid for glass) for several hours.  The
bottles were then rinsed thoroughly with tap water, distilled
water, and deionized distilled water respectively.  They were
air-dried and stored with their caps in place.

     Preservation procedures were performed on the liquid samples
immediately after sample collection.  The pH of each liquid
sample was measured using pH indicator paper.  Acid was added to
each sample until the pH was reduced to 2.0.  The samples that
were analyzed for dissolved/suspended solids and metals were pH-
adjusted using nitric acid.  Sulfuric acid was used for
preserving the oil and grease samples.

     In addition to taking samples of the cleaning solution, wipe
samples were taken of the cleaned parts.  Wipe samples were taken
to evaluate the cleaning efficiency of the solution over time by
analyzing for residual contaminants  (oil and grease)  on the
parts.  One wipe sample was taken from the cleaned metal parts
during each sampling interval to determine if there was a
residual of oil and grease.  The wipe sample was performed using
sterile, uncontaminated cloth.  Sterile gloves were worn to
prevent contamination of the cloth with oil and grease.  The
wiping procedure was consistent for each sample.  A glass
container of sufficient volume was used to hold the cloth after
sampling.  Three wipe samples were taken over the life of  the
DuSqueeze cleaning solution, to coincide with the three liquid
samples described above.

                               466

-------
      Analysis for metals was performed using inductively-coupled
 plasma atomic emission spectroscopy (ICP).  Oil, grease and
 dissolved/suspended .solids were analyzed using gravimetric
 analytical techniques.  Spikes and replicate analyses were also
 done to check for accuracy and precision and to identify the
 presence of any matrix effects associated with sample preparation
 or measurement.  Data were then combined and statistically
 evaluated.

      The analysis of plasma coating bond strength compared
 current data collected by APS Materials regarding the strength of
 coatings applied after parts were cleaned with DuSqueeze and
 historical data of bond strength resulting from parts cleaning
 with TCA and methanol.  Data generated two months prior and two
 months following the conversion to DuSqueeze were, used for this
 comparison.


                       RESULTS  AND DISCUSSION
 BENCH SCALE EXPERIMENTS

      The before and after tensile strength results were
 comparable.  Overall, the bonding strengths were actually
 slightly better for the dilute limonene cleaner (see Table 1).
 No residual limonene was detected (detection limit 1 ppm) for
 cleaner at 100:1 dilution.

TABLE 1.  TENSILE STRENGTH TEST RESULTS FOR BENCH SCALE EXPERIMENTS

 Test Buttons                 Cleaning         Tensile
 	Agent	strength (psi)	


 titanium                      methanol         6300+/-1260

 titanium                      DuSqueeze*       7000+/-570

 cobalt/molybdenum               TCA            5150+/-1990

 cobalt/molybdenum             DuSqueeze*       5400+/-1290

 *Tensile strengths measured for test button cleaned with
 various dilutions of DuSqueeze showed no trends or statistical
  differences, so values shown include all measurements.

 ANALYSES FOR IN-PLANT OPERATIONS

      The initial tests for contaminants in methanol and TCA used
 for cleaning yielded the results shown in Table 2.  The samples
 for these analyses were taken when the baths were considered
 spent, just prior to being dumped.

                                46?

-------
TABLE 2.  RESULTS OF ANALYSES OF SOLVENT SAMPLES FOR CONTAMINANTS
Test
Methanol fmcr/L!).
TCA fma/L }
Dissolved solids
Suspended solids
Oil and Grease

Metals

Cobalt
Titanium
    1
   33
  911
 29
  9
141
                       ND
    0.021
* Method detection limit is O.ol

     The amounts of oil and grease found in the wipe samples,
shown in Table 3, were very low at about 1 mg or less.  The
increase in oil and grease from the bath dump as compared to the
fresh bath was very small for one sample and was less than the
fresh bath in the second bath dump sample.  This latter result
could have resulted from the wiping technique.  In any case, the
parts seem to be cleaned just as well at the time the bath is
dumped as when the bath is fresh.

TABLE 3.  RESULTS OF ANALYSES FOR OIL & GREASE ON PARTS CLEANED
               WITH 100:1  DILUTE SOLUTION DUSQUEEZE
Test
                 Oil and Grease
                    Total Ma
Wipe Sample, Fresh Bath
Wipe Sample, Mid-life Bath
Wipe Sample, End-life Bath
BLANK
                        1.0
                        0.4
                        1.2
                         ND*
*  Method detection limit is 0.3

     Table 4 shows results from analyses for residual limonene on
the parts.  Limonene was not detected in the rinse samples, thus
indicating that all of the limonene was removed during dragout
and subsequent drying of the parts.

TABLE 4.  RESULTS OF ANALYSES FOR RESIDUAL LIMONENE ON PARTS
          CLEANED WITH 100:1  DILUTE SOLUTION DUSQUEEZE
Test
                 limonene
                 concentrat ion
                 Total Ucf/sample
Rinse Sample, Fresh Bath
Rinse Sample, Mid-life Bath
Rinse Sample, End-life Bath
BLANK
                     ND(<0.3)
                     ND(<0.65)
                     ND(<0.3)
                     ND(<0.2)
                               468

-------
     In comparing the results in Table 5, it is noted that
dissolved solids and oil and grease were much higher in the fresh
bath and the bath used to 'clean parts only prior to plasma
spraying (Dumpfl), than in the bath used also for cleaning after
plasma spraying  (Dump!2), while the reverse was true for the
suspended solids.  The graphite in the bath may affect the
DuSqueeze cleaning solution to create these differences.

TABIiE 5.  RESULTS OF ANALYSES OF 100:1 DILUTE DUSQUEEZE SOLUTION
                         FOR CONTAMINANTS
Test
Fresh Bath
ma/L
Dumpfl
ma/L
Dump f 2
ma/L
Dissolved solids       3650         3010           837
Suspended solids        ND*            ND*          19
Oil and Grease          37.0         30.8           15.1

Metals

Cobalt                   0.019        0.18          0.081
Titanium                ND#            ND#          1.65

* Method detection limit is2
f Method detection limit is 0.047

     In comparing the DuSqueeze cleaning solution with the
previous methanol and TCA samples, it is noted that the oil and
grease levels in the DuSqueeze are much lower than the other
cleaning solvents.  Suspended solids for the DuSqueeise are lower
than the previous solvents except for the sample containing
graphite which is roughly equivalent.  Dissolved solids for
DuSqueeze are much higher than the other solvents.

     The higher dissolved solids may reflect the fact that the
DuSqueeze is an emulsifying agent which converts the oil and
grease to dissolved solids.  This would explain the lower oil and
grease levels for DuSqueeze.

 TABLE 6.  TENSILE STRENGTH TEST RESULTS FOR IN-PLANT OPERATIONS

Coating/Substrate             Cleaning             Tensile
	Agent	Strength  (psi)

titanium / titanium           methanol               5560+/-600

titanium / titanium           DuSqueeze              7180+/-610

titanium / cobalt-moly          TCA                  5820+/-370

titanium / cobalt-moly        DuSqueeze              5330+/-1560
                               469

-------
     Although the data generated by the sampling and analysis
program, shown in Table 6, indicate  that the terpene-based
cleaner adequately cleaned the parts for this process, since wipe
samples were not taken for the original process, a statement of
comparison between the former and present cleaning techniques is
not feasible.

ECONOMIC ANALYSIS

     Although the old ultrasound bath was in use at the time of
the test, economic analysis is shown for the system that APS is
now operating.

     - Capital Cost:  Ultrasound with heater  $1425
                      5-gallon stainless rinse vessel  $38
                      Immersion heater  $105
                      Heat Gun  $75
                      DI water system installation  $150
                      TOTAL - $1793

     - Annual operating costs:  7.8-11.8 gal DuSqueeze  $150
                                1825-2920 gal DI water  $700

     - Cost savings from avoided purchases:
                                330 gal TCA  $1650
                                120 gal methanol $1000

     - Cost savings from avoided disposal: 6 barrels  $3000

     - Net Cost savings:        $4800 a year

     - Payback $1793/$4800 = 0.37 year, 4.5 months


                           CONCLUSIONS


     In summary, it has been determined that a terpene-based
cleaner can adequately clean metal parts without adversely
affecting the performance of the plasma-arc coating application.
The use of a terpene-based cleaner in place of methanol and TCA
has proven to be an environmental and economic success.
Elimination of the disposal problems associated with methanol and
TCA coupled with the maintenance of plasma-arc coating quality
makes the use of terpene-based cleaners attractive to other
plasma spray coating processes as well as other metal
cleaning/coating operations.  The annual cost savings as well as
the short payback period also make the cleaner attractive from an
economic standpoint.


                               470

-------
                           REFERENCES


1.  Herman, Herbert, "Plasma Spray Deposition Processes", MRS
    Bulletin, p. 60 - 68, December 1988.

2.  Test Methods for Evaluating Solid Waste.  SW-846, Third
    Edition, U. S. Environmental Protection Agency, November
    1986.

3.  Environmental Monitoring and Support Laboratory.  Methods for
    Chemical Analysis of Water and Wastes.  EPA-600/4-29-020,
    U.S. Environmental Protection Agency, Cincinnati, Ohio, March
    1983.
                               411

-------
                         WASTE MINIMIZATION ASSESSMENT CENTERS

                             by: F. William Kirsch and Gwen P. Looby

                          Industrial Technology and Energy Management
                              UNIVERSITY CITY SCIENCE CENTER
                                       3624 Market Street
                                    Philadelphia, PA 19104
                                        (215)387-2255
                                          ABSTRACT

       In 1988, University City Science Center (Philadelphia, Pennsylvania) began a pilot project to assist
small and medium-size manufacturers who want to minimize their formation of hazardous waste but who lack
the In-house expertise to do so. Under agreement with the Risk Reduction Engineering Laboratory of the U.S.
Environmental Protection Agency, the Science Center's Industrial Technology and Energy Management
(ITEM) division established two  waste  minimization  assessment centers (WMACs) at Colorado State
University In Fort Collins and at the University of Tennessee In Knoxvllle. During the second program period
of the project, a third WMAC established at the University of Louisville (Kentucky) has begun to conduct
assessments.

       Each WMAC Is staffed by engineering faculty and students who have considerable direct experience
with  process operations  In manufacturing plants and who also have the knowledge and skills needed to
minimize hazardous waste generation. The waste  minimization  assessments are conducted at  no out-of-
pocket cost to the client.  Several site-visits are required for each client served. The WMAC staff  locate the
sources of hazardous waste in each plant and identify  the current disposal or treatment methods and their
associated costs. They then  Identify and analyze a variety of ways to reduce or eliminate the waste. Specific
measures  to achieve that goal are recommended and the essential supporting technological and  economic
Information Is  developed.  Finally,  a confidential  report  which  details  the  WMAC's  findings and
recommendations  including cost savings, implementation costs, and payback times Is prepared for each
client manufacturer.

       This presentation will discuss results from the first period of this project.


       This paper has been reviewed in  accordance with the U.S. Environmental Protection Agency's peer
and administrative review policies and approved for presentation and publication.
                                             472

-------
                         WASTE MINIMIZATION ASSESSMENT CENTERS


                                        INTRODUCTION


       The amount of hazardous waste generated by industrial plants has become an increasingly costly
problem for manufacturers and an additional stress on the environment. One solutfon to the problem of
hazardous waste is to reduce or eliminate the waste at its source.

       University City Science Center (Philadelphia, Pennsylvania) has begun a pilot project to assist small
and medium-size manufacturers who want to minimize their formation of hazardous waste but who lack the
in-house expertise to do so. Under agreement with the Risk Reduction Engineering Laboratory of the U.S.
Environmental Protection  Agency, the Science Center's Industrial Technology and Energy Management
(ITEM) division initially established two waste minimization assessment centers (WMACs) at Colorado State
University in  Fort Collins  and at the University of Tennessee in Knoxville.   Each WMAC is staffed by
engineering faculty and students who  have considerable direct experience with process  operations in
manufacturing plants  and  who also have the knowledge and skills needed to  minimize hazardous waste
generation. During late 1989, a third WMAC at the University of Louisville was established.

       During the initial period of this pilot project, each of the two WMACs conducted six assessments for
small and medium-size manufacturers at no out-of-pocket cost to the client. Each client had to meet the
following criteria:

       •       Standard  Industrial Classification Code 20-39
       •       Gross annual sales of not more than $50 million
       •       No more than 500 employees
       •       Lack of in-house expertise in waste minimization

       The potential  benefits of the pilot project include minimization of the amount of waste generated by
manufacturers, reduced waste treatment and disposal costs for participating plants, valuable experience for
graduate and undergraduate students who participate in the program, and a cleaner environment without more
regulations and higher costs for manufacturers.

       All told, the measures recommended by the two WMACs in these 12 plants accou nted for an identified
cost saving of $1.28 million/year.  This paper describes how the cost savings were found and identifies the
specific measures designed to reduce waste formation and emissions from two of these plants. However,
equally detailed accounts  of the other ten plants could also be prepared.


                               METHODOLOGY OF ASSESSMENTS


       The waste minimization assessments require several site-visits to each client served. In general, the
WMACs followthe procedures outlined in the Waste Minimization Opportunity Assessment Manual. July 1988.
The WMAC staff locate the sources of hazardous waste in each plant and identify the current disposal or
treatment methods and their associated costs.  They then identify and analyze  a variety of ways to reduce
or eliminate the waste. Specific measures to achieve that goal are recommended and the essential supporting
technological and economic information is developed. Finally, a confidential report which details the WMAC's
findings and recommendations including cost savings, implementation costs, and payback times is prepared
for each client manufacturer.
                                               473

-------
                                            RESULTS


PLANT No. 1 (SIC 2851)

       This plant produces paints, coatings, stains, and surface-treating products at an overall rate of about
1.1 million gallons/year for regional distribution on a schedule of 2080 hours/year for 52 weeks. Its operations
primarily Involve blending and mixing of raw  materials, followed by product testing and packaging and by
cleaning of vessels and lines.  Color separation in the product is obviously important, and each lot must meet
a variety of other customer specifications.

PJapt.Operatic n s

       Individual lots of water-based and solvent-based paints are mixed in a variety of tanks from 200 to
1000 gallons' capacity.  Ingredients for this initial step  include (for water-based) water, latex, resins,
extenders, and dispersed pigments. For solvent-based paints the materials are generally similar in type, but
obviously solvent replaces water  and latex, and the other new ingredients include plasticizers, tints, and
thlnners.

       After batches are made up they are transferred to so-called let-down tanks, where additional water
(or solvent), resins, preservatives, anti-foaming agents, thinners, and bactericides are added.  Testing of
batches encompasses at least color, viscosity, and gloss, and those lots which meet specifications are filtered
and charged to cans for labeling, packaging,  and shipping.

Waste Generation and Existing Waste Management Practices

       The principal waste streams arethe result of equipment cleaning, especially from water-based paints.
For example, rinsing the let-down tanks ordinarily requires about 35 gallons of rinse water, but that value
Increases to 53 gallons if light paint is to be blended after a dark predecessor. The hazardous nature of water
rinses is due to mercury from the bactericide  in the paint.

       In some  instances, rinse water from the mixing tanks is held in 500-gallon tanks and used in the let-
down tanks (instead of fresh  water) to formulate future batches of water-based paint.  The rinses are
separated according to the color intensity of paint in the tanks from which they were derived. For example,
rinses from white paint formulation amount to about 70% of the total, and they are invariably used again.

       Waste rinses not used again are piped to holding and flocculation tanks, to which alum is added to
lower the pH, in which some solid  is precipitated by adding flocculant, and from which supernatant liquid is
removed for re-use in other paint formulations.

       Tanks used for solvent-based paints are rinsed with mineral spirits at a rate of about 5 gallons/400-
gallon tank. These washings  are sent off-site for recovery, followed by recycling or sale as fuel.

       In addition to re-use of rinse water and recovery  of solvent, this plant has adopted the following
measures to reduce waste generation:

       •       Cleaning equipment before the paint dries and hardens.

       •       Eliminating hazardous  materials, except for mercury in  the bactericide added to outdoor
               water-based paint.

       •       Avoiding hazardous container waste by purchasing the  bactericide in water-soluble bags
               which dissolve during paint formulation.

       •       Scheduling batch  formulations so that light ones precede dark ones and thereby reduce the
               total volume of rinses.
                                               474

-------
        •       Reducing the inventory of raw materials to avoid degradation and spoilage and to assure
               high-quality product that can be sold, rather than low-quality paint which adds to the burden
               of waste disposal.

        •       Using bag filters to collect dust.

Waste Minimization Opportunities

        Table 1 summarizes the principal sources of waste, their amounts, the management method applied,
and the  associated costs.

        Table 2 offers a brief description of each recommended WMO (Waste Minimization Opportunity) and
of current plant practice, together with savings and cost data.  Considered individually, the three WMOs
recommended could  save  over $22,000/year, which represents about 25% of current waste management
costs. Each has a simple payback time less than one year.

PLANT  No. 2 (SIC 3443)

        This  plant manufactures aluminum  brazed oil  coolers for use in heavy  equipment.  It produces
approximately 59,290 units each year.

Manufacturing Operations

        The raw materials  used in the production of the oil coolers include aluminum in sheet and coil form,
aluminum castings and extrusions, tubes, fittings, brackets, caution labels, and plastic plugs.

        The following steps are involved in production:

        •       Shearing,  punching, and forming operations to fabricate the oil coolertanks, headers, air fins,
               sides, and oil turbulator fins.

        •       Degreasing  of oil cooler tanks,  headers,  sides, fittings,  and  brackets.  The solvent
               Chlorothene (95% 1,1,1 - trichloroethane)  is used in an open-air, steam-heated vapor
               degreaser. The unit is equipped with a refrigeration unit which condense s Chlorothene vapor
               and  minimizes evaporative losses to surrounding plant air.

        •       Recycling of  spent  Chlorothene to the degreasing operation using an on-site  still.
               Chlorothene is continuously circulated between the degreaser and a steam-heated solvent
               recovery still.  Still bottoms containing spent Chlorothene, water, and  oil are shipped off-
               site as hazardous waste.

        •       Assembly of oil coolers.

        •       Brazing of assembled  oil coolers to join the internal and external coil fin surfaces for
               enhanced heat transfer.  The oil coolers are first preheated in a gas-fired oven at 1020° F for
               15 minutes. They are then dipped into an electrically-heated molten salt bath containing a
               sodium chloride-based compound, lithium chloride, and aluminum fluoride for 1  1/2 minutes
               at 1128° F and dipped in a water quench tank.  Sludge from the salt bath and quench tanks
               is disposed in the outdoor on-site sand filter bed.  Solids remaining in the filter are landfilled
               on company property; water is fed  to the settling pond and eventually discharged to a  river.

        •       Cleaning of oil coolers to remove all residual salt, expose copper cells  (which could cause
               corrosion  failure),  and condition metal surface prior to painting.  The  following steps are
               involved in the cleaning:

               •  submersion in a 2% nitric acid bath (1-2 hour residence time)
                                               475

-------
                TABLE  1.   SUMMARY OF WASTE GENERATED   (SIC 2851)
                                   WASTE TREATMENT
                               WASTE DISPOSAL
HASTE STREAM
   Amount
 Cost
Amount
Cost
HAZARDOUS LIQUID WASTE

A.   Water-based Waste:

     Equipment cleaning
     by water washing


B.   Solvent-based Waste:

     Equipment cleaning
     by solvent washing
26.700 gal
(Hg water
and paint)
$3.740       26.700  gal    $48,040
              off-site
                           27.200  gal
                             (mineral
                             spirits)
                             off-site
                         $37,080
     TOTALS
26,700 gal     $3.740
            53.900 gal   $85,120
                                    476

-------
                           TABLE 2.   SUMMARY OF WASTE  HINJHIZATIOH OPPORTUNITIES  RECOHHEHDED   (SIC  2851)
WHO No.
Present Practice
Proposed Action
Cost Savings
             Hater rinses remove paint
             from tanks and pipes
             About IS gal solvent per
             batch of paint is druuned
             and  sent off-site for
             disposal.
              A  bactericide containing
              mercury  is being used in
              water-based paints.
                                 Install  a pipe-cleaning system
                                 consisting of 3 different-sized
                                 foam plugs or "pigs"  to be sent
                                 throughout the pipes  by
                                 compressed air.  Paint is thus
                                 forced from the lines and to the
                                 canning line filter.   The use of
                                 water and amount of waste are
                                 lower.  (This HMO is applicable
                                 to non-white paints.)

                                 Use a solvent recovery system
                                 based upon distillation and ship
                                 the small amount of remaining
                                 solid to a hazardous waste
                                 disposal site.

                                 Eliminate the bactericide from
                                 water-based  interior paints and
                                 substitute an organic material.
                                 (This MHO is applicable to non-
                                 white paints.)  There  is no cost
                                 difference between these
                                 additives.
                                        Estimated waste reduction
                                        Estimated cost reduction
                                        Estimated implementation cost
                                        Simple payback
                                        Estimated waste reduction
                                        Estimated cost reduction
                                        Estimated implementation cost
                                        Simple payback
                                        Estimated waste reduction
                                        Estimated cost reduction
                                        Estimated implementation cost
                                        Simple payback
                                 1.780 gal/yr
                                 $11.110/yr
                                 11.600
                                 2 months
                                 3,300 gal/yr
                                 15.420/yr
                                 $4.950
                                 11 months
                                 3.100 gal/yr
                                 $5.580/yr
                                 none
                                 immediate

-------
                 cold water rinse
                 dipping in NaOH caustic soda etching solution
                 hot water (102°F) rinse
                 cold water rinse
                 dipping in a 50% nitric acid bath
                 2 cold water rinses
                 dipping in a chromic acid wash
                 2 de-ionized water rinses
                 drying in a natural gas-fired oven

        •       Treatment of hazardous spent process solutions and contaminated rinse water streams. The
               liquids are treated in a neutralization tank with lime for pH control and flocculant to enhance
               removal of suspended solids. The solution leaving the tank is pumped to a clarifier which
               removes solids and allows filtered water to flow to the settling pond.  A solids-rich stream is
               pumped to a sludge-thickener settling tank for secondary sedimentation. Supernate from the
               settling  tank is transferred to the sand filter beds for final water removal before on-site
               landfilling of solids.

        •       Treatment of effluent from the chromic acid and de-ionized rinse water washes.  These
               hazardous waste streams contain chromium in hexavalent form. The streams are treated to
               obtain a sludge containing  less toxic trivalent chromium compounds.  Several chemical
               agents are added to the waste to produce relatively insoluble compounds which are recovered
               on the sandfilter beds and disposed in the landfill. The liquid is pumped to the settling pond
               and is eventually released to the river.

        •       Painting of oil coolers. The coolers  are dipped in a paint-filled tank,  allowed to drip after
               immersion, and transferred to a spray booth for additional spray painting. Paint is collected
               on floorcoverings (plastic sheet or cardboard) and in spray booth filters and disposed of daily
               in barrels which are sent to an off-site landfill.

Existing Waste Management Practices

        The plant has taken the following steps in managing its hazardous wastes.

        •       The plant owns and operates a landfill for its private use.

        •       Chromium reduction from hexavalent to trivalent form is performed in-house.

        •       A refrigeration unit and a solvent recovery still have been added to  the degreasing unit to
               minimize evaporative loss and liquid waste.

        •       The plant constantly monitors its waste-stream effluents and has installed its own hazardous
               waste treatment facility.

        •       Water-based paints are currently used.

        •       The plant has a designated professional staff person based at corporate headquarters who
               periodically visits satellite plant locations to provide assistance in  both hazardous waste
               monitoring and management techniques.

Waste Minimization Opportunities

        The waste currently generated by the plant, the source of the waste, the quantity of the waste, and
the annual treatment and disposal costs are given in Table 3.

        Table 4 shows the opportunities for waste minimization that the WMAC recommended for the plant.
The waste In question, the minimization opportunity, the possible waste reduction and associated savings,
and the implementation cost along with the payback time are given.


                                               478

-------
                                  TABLE  3.   SUMMARY OF WASTE GENERATED   
-------
                             TABLE 4.   SUHHARY  OF HASTE  HIHIHIZATIOH  OPPORTUNITIES RECOHHENOED  (SIC 3443}
                                                                  Part I
Haste
Minimization Opportunity
  Waste Reduction        Net Annual    Implementation    Payback
Ouantity	Per Cent    Savings	Cost	Years
Evaporation of Chlorothene
from the degreaser unit.
Still bottoms from the
on-site solvent
recycling still.
 Evaporation  of
 Chlorothene  and
 Chlorothene  contained
 in the  still  bottoms.

 Sludge  from  the  water
 quench  tank  in the
 brazing process.
 Sludge from the salt
 bath and water quench
 tanks in the brazing
 process.
Install a conveniently removable   3.263 gal/yr
cover on the vapor degreaser
tank to reduce evaporative losses.
Cover the tank except during times
when parts baskets are being
lowered into or taken out of the
tank.

Reduce the amount of lubricants        30 gal/yr
used during metal-working and
the openness of machine work
areas to decrease the amount of
oil picked up by parts during
processing, thereby minimizing
the amount of degreasing required.

Replace  the vapor degreaser system  6.600  gal/yr
with an  ultrasonic cleaning system
which  utilizes biodegradable
detergents.

Hodify  the  procedure  for dipping    23,171  Ib/yr
the  coolers in the salt bath to
minimizt earry-ovtr to the water
quench tank.  Achieve maximum
 salt removal  by  gently vibrating
 or shaking  the parts  baskets and
 subjecting  the parts  to  a  hot
 air blast.

 Replace molten salt  bath  brazing   411.934 Ib/yr
 with vacuum brazing.   Vacuum
 brazing is suitable  for  SOX
 of this plant's  products.
                 SOS
S17.180CA)
  $220
0.01
                 20S
  I.OIO(A)
                                              290
                                                            0.3
                 99S
 20.450(8)
                           20,520(0
SO.000
                  43.BSD
 2.4
                 2.1
                  SOS     203.440(0
                  720.640
                 3.5

-------
                            TABLE  4.   SUHHARY OF HASTE MINIMIZATION OPPORTUNITIES RECOHHENDED   (SIC  3443)
                                                                Part II
Waste
Minimization Opportunity
   Waste Reduction        Net Annual    Implementation    Payback
 Quantity      Per Cent    Savings           Cost          Years
Paint-contaminated
cardboard and plastic
sheets.
Reduce paint loss by
installing a low-pressure
air-jet system over the
paint dipping area to blow
excess paint downward into
tank.  Install an IR paint-
drying  lamp to prevent
dripping when coolers are
moved to the spray booth area.
2.180 Ib/yr
         4.350  (D)
                  2.490
                                           0.6
Paint-contaminated
filters and cardboard
and plastic sheets.
Install an electrostatic spray
paint system for application of
the  oil cooler second coat of
paint in order to reduce
overspray loss.
3.513 Ib/yr
36$
11.200 (D)
13.200
                                           1.2
 Paint-contaminated
 filters  and cardboard
 and  plastic sheets.
Discontinue  the practice of
painting  oil  coolers which
will  be re-painted by
the  customer.
4,90S Ib/yr
50$
59,720 CD)
28.440
0.5
 CA)   Includes  cost  savings attributed  to the avoided purchase of Chlorothene.
 (B)   Total  savings  have  been  reduced by the cost of detergents required.
 tC)   Includes  cost  savings attributed  to the avoided purchase of salt bath constituents.
 (D5   Includes  cost  savings attributed  to the avoided purchase of paint supplies.

-------
                                   DISCUSSION OF RESULTS


       The two plants described have annual waste management costs of $188,370 and the WMACs were
able to recommend to them a series of cost-saving measures which add up to $359,980 per year. The savings
total exceeds the aggregate waste management cost because, in one of the two plants, the measures
recommended save more than the cost of waste treatment, disposal, and recycling.

       For example, in plant No.2, only a $13,485 waste management cost is presently associated with paint-
contaminated filters and cardboard and plastic sheets but elimination of the  painting is estimated to save
almost $60,000 per year. Typically, savings in raw materials costs of 25 to  80% can be achieved  by the
recommendations offered.  It is not hard to see how this kind of savings can add up  to more than the costs
of waste management in some plants.

       The cost-saving approach taken with these results is generally conservative, because the WMOs
address only the avoidance of raw materials costs and the reduction of present and future costs associated
with waste treatment and disposal. Not claimed are the savings related to: possible changes in emission
standards, any liability incurred from waste management practices, and costs  arising from employee health
and safety problems. It should also be noted that each WMO identified is treated as an isolated individual
measure and no consideration has been given to effects occurring because of interactions among WMOs.

       The WMAC program is continuing to serve eligible manufacturers, and future results are expected
to allow more generalizations about the amounts and types of industrial waste materials encountered in the
nation's plants.

       It  seems reasonable to conclude from recent WMAC program experience that small and  medium-
size manufacturers:

       »       Have recognized many of their waste generation problems and have undertaken a variety of
               actions to address them.

       »       Are receptive toward practical  quality  assistance, offered in their  plants objectively and
               competently, that can help them to choose cost-effective waste minimization opportunities
               over and above what they have been able to achieve on their own.

Acknowledgment

       The authors wish to express their appreciation to EPA's Risk Reduction Engineering Laboratory for
support of the Waste Minimization Assessment Centers and the opportunity to prepare this paper. They also
want to acknowledge the waste minimization assessments performed and reports prepared by Dr. Richard
J. Jendrucko and Ms. Phylissa S. Miller at the University of Tennessee and by Dr. Harry W. Edwards, Dr. C.
Byron Winn, Mr. John R, Bleem and Mr. Michael Kostrzewa at Colorado State University. Their work provided
the data upon which this paper is based.
                                             482

-------
    VETERANS AFFAIRS HOSPITAL AND HOSPITAL WASTE MINIMIZATION CJiSE STUDIES

              by:  Kenneth R. Stone
                   Risk Reduction Engineering Laboratory
                   U.S. Environmental Protection Agency
                   Cincinnati, Ohio 45268
                                   ABSTRACT

     The U.S. Environmental Protection, Agency has instituted a broad pollution
prevention research program through the Office of Research and Development to
support continued environmental improvements throughout the nation.  The
Agency is also responding to the national concern in regards to the generation
and disposal of medical wastes.  Recently, EPA's Risk Reduction Engineering
Laboratory (RREL) produced the "Guide to Waste Minimization in Selected
Hospital Waste Streams" (1) with the cooperation of the California Department
of Health Services (hereafter referred to as the "California Study").  The
California Study serves as a manual for conducting waste minimisation
assessments at surgical and general medical hospitals to reduce the generation
of hazardous wastes from chemotherapy and antineoplastic chemicals,
formaldehyde, photographic chemicals, radionuclides, solvents, nercury,
anesthetic gases and other waste chemicals.

     In order to effectively implement its pollution prevention programs, the
EPA is also investigating how the departments and agencies within the Federal
community can help each other reduce their generation of wastes.  As a part of
these efforts, RREL provides staff and support to conduct waste minimization
assessments under the Waste Reduction Evaluations and Assessments at Federal
Sites (WRIAFS) Program.  Under the WREAFS program, the U.S. Department of
Veterans Affairs Cincinnati - Fort Thomas Medical Center 
-------
    VETERANS AFFAIRS HOSPITAL AND HOSPITAL WASTE MINIMIZATION CASE STUDIES
                                 INTRODUCTION
     The California Study and the DVA-Cin Study represent RREL's initial
efforts in the research of hospital wastes.  Together, the studies profile
those wastes that are unique to me'dical care facilities: hazardous and bio-
hazardous wastes, wastes generated from patient care, and medical laboratory
wastes,  RREL did not attempt to study wastes generated in other areas of the
hospitals (e.g., office settings, cafeteria, plant maintenance), as these
wastes are common to many non-medical facilities which have been examined in
prior studies.

     This paper will outline the waste profile of medical care facilities and
suggest waste minimization options based upon the results of both the
California and DVA-Cin Studies.  This will include a discussion of research
needs/opportunities derived from the DVA-Cin Study,
                             THE CALIFORNIA STUDY
HAZARDOUS WASTE PROFILE

     While the volumes of hazardous wastes generated are small in comparison
to an industrial facility, hospitals do employ a wide variety of toxic
chemicals and hazardous materials for numerous diagnostic and treatment
purposes.  Based on the assessments of three hospitals under the California
study, the highest volume of hazardous waste generation comes from the use of
chemotherapy and antineoplastic chemicals, followed by spent photographic
chemicals and formaldehyde solutions used for disinfecting equipment.
Briefly, the wastes studied include:

     Chemotherapy and Antineoplastic Chemicals - Antineoplastic, or cytotoxic,
agents are typically kept on hand in quantities sufficient to last two weeks.
To produce chemotherapy solutions, chemicals are mixed under a hood which re-
circulates air through a filter.  Only a small percentage of these wastes
contain concentrated amounts of chemotherapy compounds.  Much of the waste is
associated with lightly contaminated items such as personal protective
clothing and gauze pads.  An average of 2 to 8 cubic feet of chemotherapy
wastes were generated weekly by the hospitals surveyed.  These wastes were
either transported off-site to a Class I landfill or incinerated as hazardous
waste.  It should be noted that individual States have differing regulations
                                      414

-------
on how such waste is to be disposed.  What is allowed in one State may be a
violation in another.  Therefore, the waste minimization recommendations that
are made in this paper have to be taken in light of the medical waste
regulations of the State in which the hospital is located.

     Formaldehyde - Formaldehyde is used in pathology, autopsy, dialysis,
embalming, and nursing units.  For use in dialysis, formaldehyde is generally
purchased as a 37 percent solution  (formalin) that will be diluted with
filtered, de-ionized water to a final formaldehyde concentration of 2-4
percent.  Formaldehyde is used to disinfect membranes in dialysis machinery
and, in other departments, to preserve specimens.  Effluent is commonly
discharged to a sewer, although in some States this may be considered an
illegal practice.

     Photographic Chemicals - Photographic developing solutions consist of
three parts: developer, stop bath, and fixer.  The developer normally contains
approximately 45 percent glutaraldehyde.  Acetic acid is a component of stop
baths and fixer solutions.  The fixer will contain 5-10 percent hydroquinone,
1-5 percent potassium hydroxide, and less than 1 percent silver.  Silver-
containing effluent is typically passed through a steel wool filter or
electrowinning unit to recover the metal.  The remaining aqueous waste,
containing approximately 1.4 percent glutaraldehyde, 0.3 percent hydroquinone,
and 0.2 percent potassium hydroxide, is typically discharged to the sewer.

     Radionuclides - Radioactive wastes are generated in nuclear medicine and
clinical testing departments.  At the hospitals surveyed, radioactive
materials in nuclear medicine were held on-site until they decayed to non-
hazardous levels.  In clinical testing laboratories, solvents were used for
radioactive tagging.  Wastes at the hospitals were generated at the rate of
800 cubic centimeters per week.  Radioactive wastes were transported off-site
to a landfill.

     Solvents - Solvent wastes are generated in small amounts in various
departments: pathology, histology, engineering, embalming, and laboratories.
A variety of halogenated and non-halogenated compounds are used, but, in the
hospitals surveyed, the most frequently used solvents were non-halogenated:
xylene, methanol, and acetone.  While acetone and raethanol wastes are usually
evaporated and/or discharged to a sewer, xylene is handled as a hazardous
waste.  Solvent wastes are typically recycled or transported off-site for
incineration.  However, some solvent wastes become absorbed into the specimen
and then must be treated as infectious wastes.  In the past, small quantities
of solvent waste would be routinely disposed via lab packs to landfills.
However, high disposal costs, long term liability and regulatory
limitations make this an undesirable disposal alternative.

     Mercury - Mercury wastes are primarily generated by broken; or obsolete
equipment.  Spilled mercury can be recovered and reused if uncontaminated,
however, spillage is not frequently recovered and no mercury spill kits were
present in any of the surveyed hospitals.

     Anesthetic Gases - Nitrous oxide and the halogenated agents halothane
(Fluothane), enflurane (Ethrane), isoflurane (Forane), and other substances
are used as inhalation anesthetics.  Nitrous oxide is supplied as a gas in
cylinders and used containers are returned to the supplier for refill.  The

                                      485

-------
halogenated agents are supplied in liquid form, in glass bottles.  Once empty,
the bottles are treated as hazardous waste.  Waste anesthetic gases are vented
from the operating room directly to the outside, or through a charcoal filter.
Spent charcoal filters are transported off-site and disposed of as hazardous
waste.
HAZARDOUS WASTE MINIMIZATION OPTIONS
     As a result of the hospital assessments, a series of waste minimization
options were developed for the wastes categorized above, as well as some other
wastes not addressed by this paper.  These options are briefly portrayed in
Table 1.

     While the waste minimization options listed on Table 1 respond to
specific waste streams, better operating practices are essential to hospital-
wide waste reduction.  Better operating practices are procedures and
institutional policies that result in a reduction of waste, and are exhibited
through such pollution prevention measures as:

     Waste Stream Segregation - It is important to keep hazardous waste
segregated from non-hazardous waste, since all materials that come into
contact with hazardous waste becomes hazardous.  Hazardous chemical wastes
should be kept separate from infectious wastes.  Further, dilution of
hazardous waste should be avoided, as it only increases the volume of waste
that will have to be treated as hazardous.  Finally, recyclable materials
should be segregated from non-recyclable waste.

     Monitoring Procurement and Product Flow - By centralizing the purchasing
and dispensing of drugs and other hazardous chemicals, a facility can more
readily recognize the occurrence of unnecessary waste and spillage, while
assessing opportunities to minimize usage.  Monitoring drug and chemical flows
from receipt to disposal can be achieved through te use of automated data
systems and bar-coding similar to that used in supermarkets.  In addition to
improving tracking and control, such a system may also provide cost savings by
allowing the hospital to carry lower stocks in inventory efficiently.

     Integrating Individual Departments with Waste Management Responsibilities
- Apportion waste management costs to the departments generating the wastes,
and require users of chemicals with limited shelf life to use up old stock
before ordering new and to report on expired stock.

     Training Employees and Providing Necessary Equipment - Employees should
be trained in hazardous materials management and waste minimization.
Employees need to be aware of chemical hazards, how to prevent spills, how to
provide adequate maintenance of equipment, and how to remediate spills quickly
and safely.
                                      486

-------
                     TABLE 1.  WASTE MINIMIZATION METHODS
                      FOE GENERAL AND SURGICAL HOSPITALS
WASTE CATEGORY
           WASTE MINIMIZATION METHOD
Chemotherapy and
 Antineoplastics
Formaldehyde
Photographic Chemicals
Radionuclides
Solvents
Mercury
Waste Anesthetic
> Optimize drug container sizes in purchasing.
> Return outdated drugs to Manufacturer,
> Centralize chemotherapy compounding location.
> Minimize waste from compounding hood cleaning.
> Provide spill cleanup kits.
> Segregate wastes.

> Minimize strength of formaldehyde solutions.
> Minimize wastes from cleaning dialysis machinery.
> Use reverse osmosis water treatment to reduce
    dialysis cleaning demands.
> Capture waste formaldehyde.
> Investigate reuse in pathology, autopsy labs.

> Return off-spec developer to manufacturer.
> Cover chemical tanks to reduce evaporation.
> Recover silver efficiently.
> Use squeegees to reduce bath losses.
> Use counter-current washing.

> Use less hazardous isotopes whenever possible.
> Segregate and label radioactive wastes, and store
    short-lived radioactive wastes on-site until decay
    permits disposal as general trash.

> Substitute less hazardous cleaning agents.
> Reduce analyte volume requirements.
> Use pre-mixed kits for tests involving solvent
    fixation.
> Use calibrated solvent dispensers for routine tests.

> Substitute electronic devices for mercury-containing
    devices.
> Provide spill cleanup kits and personnel training.
> Recycle uncontaminated mercury wastes.

> Employ low-leakage work practices.
> Purchase low-leakage equipment,
> Maintain equipment to prevent leaks.
                                      48?

-------
                               THE DVA-CIN STUDY
     The fact that hospital and medical care costs have risen dramatically
throughout this decade is commonly attributed to the cost-plus-fee
reimbursement structure of medical insurance.  Under this system, the health
care facility is able to pass on direct costs, a portion of overhead costs,
and service fees to each patient for eventual reimbursement by an insurance
carrier following an established rate scale.  Since the insured patient does
not feel the full impact of the cost of services received, the health care
facility has had little direct incentive to reduce those costs, and probably
even less incentive to investigate the cost benefits from incorporating
pollution prevention opportunities.  However, DVA facilities are not
reimbursed for health care services; each facility operates under a budget
fixed by the Department of Veterans Affairs for the fiscal year.  Therefore,
individual facilities such as DVA-Cin are very sensitive to cost, since
achieving cost savings translates into an ability to extend their services.

     Since the California Study had previously emphasized the opportunities
for hazardous waste minimization, RREL and DVA-Cin chose to look for polluton
prevention alternatives for minimizing the discarded medical supply
wastestream.  That the VA-Cin is uniquely suited to such a study is directly
attributable to its cost sensitivity.  The need to deliver services under a
fixed budget has led DVA-Cin to both adopt environmentally clean practices on
its own, and to continue clean practices that cost-reimbursement hospitals had
abandoned.  For example, the DVA-Cin Medical Center carefully segregates its
waste in order to minimize the volume that will have to be transported by the
infectious waste hauler (unit costs for infectious waste disposal far exceed
those of general refuse).  Also, taking advantage of its access to a DVA
operated laundry in Dayton, Ohio, DVA-Cin still uses durable cloth gowns and
drapes instead of the disposable paper variety.  Hospital staff who indicated
having prior work experience with other Cincinnati area hospitals maintained
that DVA-Cin's consumption of disposable gowns and drapes was the lowest in
the metropolitan area.


DISPOSABLE MEDICAL SUPPLIES


     fhe use of disposables in hospitals and the medical profession has
increased steadily over the past thirty years as devices and items constructed
of aetal, glass, and fabric have been replaced with plastic and paper
•aterials intended to be used once, and then discarded.  There are four major
factors that explain this preference for disposables:

     Health and Safety - Products that arrive prepackaged and pre-sterilized
reassure medical professionals of their integrity, and eliminate the burden of
monitoring inhouse re-sterilization procedures.

     Cost - Disposables lower inhouse labor costs and manpower associated with
cleaning, re-sterilizing and wrapping durable devices and products.


                                      488

-------
     Convenience - Such things as disposable operating room  (OR) packs provide
all the sterile materials needed for a specific operation, reducing OR prep
time.

     Space Constraints - Only needed materials are kept in stock, and there is
no need to review inventory for aging and obsolete items.
DISPOSABLE WASTE PROFILE
     The majority of waste generated by a hospital consists of disposable
products.  According to DVA-Cin personnel, approximately 80 percent of the
hospital's supplies are disposed after a single use.  The DVA-Cin saw a change
from reusables to disposables 10-15 years ago and an additional increase in
the use of disposables in the last 2-3 years due to concern by hospitals over
both patient safety and staff occupational exposure to the AIDS virus.
Therefore, the increase results from greater usage of existing disposable
supplies  (i.e., single-use sponges for patient surgery, and disposable gloves
and masks worn to protect hospital staff) rather than from the use of newly
developed disposable items.

     This section will profile the major disposable items ordered by these
DVA-Cin departments: Laboratory Services; Surgery; Surgical Intensive Care
Unit (SICU); 5 South (a patient floor); Medical Intensive Care Unit (MICU);
Hemodialysis; and the Outpatient Clinic.

     Laboratory Services - This department performs analyses on specimens
taken from patients.  In a 9-month period ending June 30, 1989, the laboratory
had conducted 41,097 venipunctures, 9,935 bacterial cultures, 4,,730 blood
cultures, 854 fungal cultures, and 815 tuberculosis cultures.  The Laboratory
consists of 4 areas: (1) Hematology, (2) Clinical chemistry, (3) Microbiology,
and (4) Histopathology.

          11 Hematology Laboratory - Hematology draws and analyzes blood
     samples from 50-60 patients daily.  The technicians visit the patients to
     draw samples and then return to the laboratory to conduct the analyses.
     Cloth gowns are worn while blood is drawn and then replaced with a second
     cloth gown for lab work.  All gowns are laundered for reuse.

          Hematology generates two 30-gallon bags of infectious waste each
     day.  It is rendered non-infectious via autoclaving and disposed of as
     general trash.  Sharps  (needles, broken glass) are placed in sharps
     containers and those containers are collected by housekeeping staff for
     weekly incineration.

          2)_ Clinical Chemistry Laboratory - Clinical Chemistry conducts blood
     serum and urine analyses on samples drawn by the hematology technicians.
     Approximate waste generation rates for the principal disposables are:

              > Glass test tubes - 2,100 per week
              > Glass sample cups - 2,000 per week
              > Dry reagent slides - 21,000 per week

                                      489

-------
          Like Hematology, Clinical Chemistry generates just under two 30-
     gallon bags of autoclaved waste each day.  The laboratories are adjacent
     and share the same autoclave.  The only disposable medical supply items
     that are reused are cuvette rings, used to test blood coagulation.
     Cuvette rings are washed and reused 5-10 times before disposal.  This is
     done because of the high unit cost ($2.30) of a cuvette ring.

          3J_ Microbiology Laboratory - This section of the laboratory produces
     the greatest amount of discarded supplies by weight because most of its
     wastes are glass products.  At least three 30-gallon bags of autoclaved
     waste are produced each day.  The waste profile primarily consists of:

              > Petri dishes with Agar culture media - 1500 per week
              > Blood culture bottles and Contaminated Slides
              > Vitek Cards - 225 per week
              > Paper towels, gowns and disposable gloves (for tuberculosis
                  isolation)

          Mo disposables are reused.  Petri dishes are not reused because they
     are difficult to clean properly and safely, and preparing media on-site
     is not cost effective.

          4) Histopathology Laboratory - This laboratory is responsible for
     analyzing tissue specimens and body parts from surgery and the morgue.
     Histopathology produces no more than one 5-gallon bag of autoclaved waste
     per day.  Pathological wastes are incinerated on-site.  Disposable
     specimen containers containing formaldehyde are autoclaved and then
     incinerated on-site.

     Surgery Department - The Surgery handles approximately 15 cases daily.
According to DVA officials, DVA-Cin is one of the last hospitals in Cincinnati
that continues to use woven gowns.  The greatest volume of medical supplies
disposed of after a single use are eKam gloves and surgical sponges.  Surgical
sponges had been reused in the past, but are now disposed after a single use
due to concerns over the AIDS virus.  Surgery also utilizes operating room
packs that are prepared with all the disposable products necessary for a
specific type of operation.  The packs are generally used in full, although
sometimes specific items may not be used.

     The Surgery Department generates between one and two 30-gallon bags of
blood and body fluid waste per case, or 15-30 bags per day.  Approximately 70
percent of this work is estimated to be contaminated paper waste.  According
to DVA-Cin officials, other Cincinnati hospitals generate three 30-gallon bags
of waste per case; DVA-Cin maintains its lower waste generation rate through
their continued use of wovens.  Surgery carefully segregates wastes as they
are generated.  Wastes to be disposed of as blood and body fluid waste must be
"grossly contaminated" (i.e., soaked, or dripping with blood).  However, DVA-
Cin feels that surgeons and nurses ought not be burdened by waste segregation
duties during surgery, and are likely to curtail this activity.

     Sharps are placed in sharps containers, which are then clear-bagged
against leakage, and sent to the on-site incinerator.  Blood and body fluid is
taken by an infectious waste hauler for off-site treatment and disposal.  All
other waste is general trash.

                                      490

-------
     Surgical Intensive Care Unit (SICU) - Almost all of SICU's eight beds are
occupied on a regular basis.  Cloth gowns are worn by patients and staff and
are laundered for reuse.  Procedure trays are re-sterilized on-site and
reused, but SICU staff would like to go to the disposable packs like those
used in surgery.

     Blood and body fluid waste generated by SICU consists mainly of suction
liners and tubes.  Foley bags and chest tubes are flushed of their fluids and
placed in general trash.  I-V bags go directly into general trash.  Sharps are
boxed and incinerated on-site.

     Waste is segregated into three categories: (1) sharps; (2) blood and body
fluids, and; (3) general trash.  Blood and body fluid wastes are strictly
segregated into one-to-two 30-gallon bags per day.  However, for those
patients requiring isolation, SICU may generate as much as ten 5-gallon bags
of medical waste per day for each patient.  The number of patients in
isolation varies.

     5. South; Patient Floor  - 5 South has 36 beds, of which 29-32 are
occupied at any given time.  5 South provides pre- and post-operative care,
including administration of medications and changing dressings.  In total, the
aedical and surgical patient floors have 106 beds, of which 78 are occupied at
any given tiae.  Cloth gowns are generally worn on patient care floors,
although disposable gowns are used whenever cloth is unavailable.

     Waste is segregated into three categories: (1) sharps? (2) blood and body
fluids, and; (3) general trash.  5 South generates one-to-two 30-gallon bags
of blood and body fluid waste per day.  In practice, nurses often dispose of
non-blood and body fluid waste in the blood and body fluid waste container as
a matter of convenience.

     Medical Intensive Care Unit/Cardiac Care Unit  (MICU/CCU) - the HICU/CCU
has eight beds, all of which are constantly occupied.  MICU/CCU reuses woven
gowns and pressure bags.  Pressure bags are used to introduce blood to a
patient, and will be  cleaned out for reuse.

     Waste is segregated into three categories: (1) sharps; (2) blood and body
fluids, and; (3) general trash.  The assessment team again observed waste
being dropped in blood and body fluid containers that did not need to be
there; empty disposable urinals were observed in the blood and body fluid
containers.

     Hemodialysis - This unit has 9 treatment stations.  Treatment occurs in
shifts with a capacity to treat 55 patients each week.  Treatment takes about
5 hours.  Nearly all products are disposable, including aprons and masks.  As
is common practice in many hospitals, disposable dialyzers are re-sterilized
and reused approximately 20 times before disposal.  The practice of reusing
disposables in health care is controversial and will be discussed further in
this paper.

     At least four 30-gallon bags of blood and body fluid waste are generated
each day.  Most of the disposable items are discarded in the blood and body
fluid containers.  Sharps are handled as previously indicated.


                                      4S1

-------
     Outpatient Clinic - The clinic services approximately 500 patients each
day.  The services include: surgical procedures, medical exams, chemotherapy,
dermatology, urology, plastic surgery, orthopedics, and ear, nose and throat.
Plastic-coated paper gowns are worn for chemotherapy procedures (disposed of
as cytotoxic waste) and often for other outpatient treatment and procedures.
Reusable wovens would include sheets, pillow cases, towels and blankets.
Badly soiled linens are often discarded rather than laundered.  Gomco suction
apparatus, suture removal sets and scalpels are all reused.

     The Outpatient Clinic fills one 30-gallon bag of blood and body fluid
waste each day.  Chemotherapy wastes are packaged in white plastic containers,
and eventually transported to final disposal off-site by a licensed cytotoxic
waste hauler.  Sharps are handled as previously indicated.


DVA-CIN WASTE GENERATION COMPARED TO OTHER HOSPITALS


     As indicated earlier, DVA-Cin does not generate waste in the quantities
common to hospitals having access to cost-plus-fee reimbursement mechanisms.
On average, hospitals generate between 0.5 and 4 pounds of infectious waste
per patient each day (3).  The DVA-Cin facility produces approximately 0.6
pounds of infectious waste per patient each day, placing it at the low end of
the spectrum.

     There are inconsistencies in how hospitals from different States define
what is infectious waste.  For example, DVA-Cin classifies its laboratory
waste as general trash after autoclaving, whereas a hospital in New Jersey
would continue to list such waste as infectious, despite the autoclave
treatment.  Inflating the DVA-Cin's quantity of infectious waste to reflect
lab wastes would raise the generation rate to 0.87 pounds per patient each day
- still quite low in comparison to other hospitals.  DVA-Cin's continued use
of wovens is likely the primary reason for this lower rate.


                POLLUTION PREVENTION OPPORTUNITIES AT HOSPITALS
     The purpose of this section is to identify and evaluate the opportunities
to minimize waste in a hospital setting.  In addition to responding
specifically to the waste profiles already presented, it is important to
discuss issues affecting pollution prevention decision-making: the
benefits/costs of disposable versus reusable products; the reprocessing of
disposable items intended for single-use; the factors that continue to promote
reliance on disposables; and the ability to implement better operating
practices.  The weight of State regulation also bears heavily upon this
decision-making process, but is best addressed through the efforts of
individual facilities formulating pollution prevention programs that comply
with both legal and health requirements.
                                      492

-------
CHOOSING BETWEEN DISPOSABLE AND REUSABLE PRODUCTS


     As stated earlier in this paper, there are four major factors supporting
the medical professions preference for disposables: health and safety; cost;
convenience; and space constraints.  With the advance of technology, intricate
devices are mass-produced and sold as single-use items, prepackaged and
sterilized to relieve the hospitals of such quality assurance concerns.  Labor
and reprocessing costs are relieved, being replaced by the apparently lower
costs of treatment, destruction, or disposal.  Packs of disposable goods are
custom-fitted, used and discarded, alleviating OR prep time and simplifying
inventory control.

     An excellent example of this decision-making process is found in the
demise of hospital laundries.  Reimbursement of medical services on a cost-
plus basis provided the incentive to introduce new products and services to
ease hospital workloads.  This created a situation wherein funds were not
allocated to upgrade traditional services, because there was little incentive
to modernize operations and streamline procedures.  In the case of hospital
laundries, they were experiencing a rising demand for all linen products as
inpatient services increased during this period.  Antiquated laundry
operations were incapable of meeting the new demand, becoming unable to
efficiently process and sterilize the soiled linens.  Acquiring disposable
linens ensured an adequate supply of products, relieved an overburdened
laundry, and provided cost savings by allowing hospitals to abandon or further
downgrade this service, rather than invest in capital improvements.  The cost-
plus reimbursement method for medical services had led the hospitals into
allowing formerly efficient laundries to lapse into a condition in which the
most cost effective solution resulted in the greatest generation of solid
waste  (4).  The assessment team for the VA-Cin Study suggested that the
medical center's extraordinary use of linens and access to the Dayton laundry
was a significant factor in explaining DVA-Cin's very low waste generation
rate.
REUSING SINGLE-USE DEVICES


     Hospitals and other health care facilities have attempted to reduce costs
by reprocessing disposable, single-use devices/products  (see Table 2 for a
listing of the most commonly reused disposable products).  Although the issue
of reusing disposable devices is highly debated, health  care professionals
agree that if a product is to be reused it must be as functional, sterile, and
safe as when new.  In making this decision, health care  professionals must
consider the possibility of disease transmission or infection, assumption of
product liability, decreased reliability, and cost.

     These concerns impact the decision on what may be reused.  The less
critical an item, the more likely it can be reused (5).  For example, because
a bedpan is considered to be a non-critical item by the  Center for Disease
Control (CDC) and the risk of infection or disease transmission is minimal,
reuse would be considered.  However, an arterial embolectomy catheter would be
considered critical and the potential risks from reuse great.  The hospital
will always opt for health and safety over any issue of  economics or ecology.

                                      4S3

-------
          TABLE 2 -  DISPOSABLE MEDICAL DEVICES REPORTED  TO BE REUSED6
                             (in  Descending  Order of  Frequency)
Heraodialyzers                (46%)
Cardiovascular catheters
  and guidewires              (31%)
Respiratory therapy
  breathing circuits            (18%)
Biopsy needles                (17%)
Cautery devices               (16%)
Anesthesia breathing
  circuits                     (14%)
Endotracheal tubes            (10%)
Suture staple removers         (9%)
Syringes                      (9%)
Orthopedic appliances         (7%)
Suction canisters              (7%)
Trachea! tubes                (6%)
Bovie cords                   (5%)
Esophageal thermometers      (4%)
External pacemaker
  electrodes                   (4%)
Arterial catheter
  needles                     (2%)
Aseptic irrigating
  syringes                     (2%)
Shunt connectors              (2%)
Sterile skin scribes             (2%)
Cfaolangiographic
  catheters                    (1%)
Esophageal stethoscopes        (1%)
Pacemakers
Pulmonary nebulizers
SIdn staplers
Urinary catheter plugs
Allen needles
Arterial embolectomy
  catheters
Condensing bottles
Operating room clamps
Ear syringes
Face tents
Gastric pH  monitors
Hypodermic needles
Javid tubes
Oxygen masks
Microscalpels
Stone baskets
Surgical gloves
Triadaptors
Tracheostomy tubes
Urethral stents
Urinary bags
(1%)
(1%)
(1%)
(1%)
                                           494

-------
FACTORS AFFECTING CONTINUED RELIANCE OK DISPOSABLES
     When considering reusable, or durable, products and reuse of disposable
products as a means for reducing the rate of waste generation and its
associate costs, infection control is the primary limiting factor.  This paper
has noted the increased attention that medical professionals are paying to
this issue in the wake of public concern over the AIDS virus and other blood-
borne pathogens.  The CDC's Universal Precautions state that all blood and
body substances must be treated as potentially infectious.  As the first line
of defense against pathogen transmissions, the medical community employs
physical barriers to prevent contact with body substances: gloves, protective
clothing, masks and eye protection.  Single-use items intended for personnel
protection provide hospitals with added assurance against accidental
transmissions because they are used once, rendered non-infectious through
autoclaving and either hauled off-site as general trash, or incinerated on-
site.  This eases the quality control burden for the hospital.

     In any case, the barriers are employed but once, and that is what counts
to the peace of mind of both doctor and patient.  And, from a more pragmatic
viewpoint, it is clear that, when considering the increasing frequency of AIDS
in urban areas and the seriousness of all infectious diseases, health
facilities must first ensure the sterility of a product or devicei and
then consider the opportunity for pollution prevention.

     Another obstacle to converting from a single-use back to a durable
product or device is that it may simply no longer exist in that form, or be
too expensive to employ.  The DVA-Cin procurement office indicated that
disposable products had, in many instances, completely eliminated the market
for the durable good.  As a result, the durable version is either no longer
available, or can be acquired through special-order supply companies that may
be unable to guarantee long term availability and unable to provide
sufficiently large quantities.  This situation in turn drives up the cost of
the durable version, potentially making it cost prohibitive.


POLLUTION PREVENTION OPPORTUNITIES AT DVA HOSPITAL


     A variety of disposable devices ranging from syringes to hemodialyzers,
from Petri dishes to bedpans, contribute to the growing waste streams
generated by health care facilities.  In order to successfully reduce waste,
it is important for hospitals to reconsider the situations in which single-use
devices/products are used and evaluate whether the disposable is still the
•best option.   If  reusable goods provide  comparable reliability, sterP-ty
and safety, it would be reasonable to consider going back to the reusable.  In
those instances wherein the hospital is reprocessing single-use devices, it
would appear that the durable version is an even more attractive substitute
because the disposable has not relieved the hospital of the burden of re-
sterilization, quality assurance, or labor costs.

     Because of the diversity of the areas toured at DVA-Cin, it would be best
to discuss pollution prevention opportunities by ward.  The major disposable
items in each ward will be reviewed.

                                      495

-------
     Laboratory Services - Nearly all items used in Laboratory Services are
disposable.  All glass products  (e.g., test tubes, sample cups, Petri dishes,
slides, pipettes, pipette tips) are autoclaved and disposed after a single use
as a matter of safety.  Although glass could be re-processed for reuse,
immediate treatment and disposal lessens handling time, decreasing the chance
of exposure to accidents and spills.  Plastic products  (e.g., pipettes,
pipette tips, test tubes, testing items, specimen bags, cuvette rings, etc.)
are autoclaved and disposed after a single use.  Of these, only cuvette rings
are reused, due to their high cost.  Since the only good durable substitutes
would be made of glass, and glass would not be re-processed for the reasons
given, substitution in this case would only increase the weight of the
wastestream.

     However, it was noted that the Microbiology Lab disposed of 1500 Petri
dishes each week.  Because they are glass, there is an opportunity to re-
process the dishes and prepare them with new media.  While they are difficult
to clean, and Agar preparation is very labor intensive, it is important to
consider developing an opportunity to have the dishes re-processed off-site.
Such an alternative would allow the lab to continue functioning without the
disruption from implementing inhouse reprocessing activities and significantly
decrease the weight and volume of the waste generated by the lab.  Within a
two-mile radius of DVA-Cin there are eight other large hospitals (Holmes,
Christ, Good Samaritan, Bethesda, Deaconess, and the University of Cincinnati
hospitals, Children's Hospital Medical Center, and Shriner's Burns Institute).
It would seem to be a reasonable entrepreneural opportunity for some business
to provide reprocessing services since it would have easy access to these main
facilities.  There are several more hospitals in the Greater Cincinnati area,
but the ones listed are located in a way that facilitates the potential for a
sharing of services.

     Surgery - Because DVA-Cin already uses woven gowns, drapes, and
instrument wraps, the greatest volume and weight of disposables in medical
waste from surgery are made up of surgical sponges and exam gloves.  While
waste sponges should continue to be considered potentially infectious, it
would be worthwhile to investigate whether sponges are being also used for
purposes better suited for absorbent, reusable towels  (i.e., cleanup
activities).  The CDC advises that all sharps, including syringes, be disposed
rather than re-processed, and therefore this paper follows that
recommendation.

     The current use of wovens over disposables by DVA-Cin greatly reduces the
potential volume of waste.  Additional opportunity for substitution in this
area is limited due to professional caution over health and safety, as well as
cost considerations.  It should be noted that the OVA laundry in Dayton does
serve several facilities.  As in the case of the Petri dish reprocessing
recommendation, this may be an area in which several community hospitals may
share a service and thus minimize their investment/capital costs.

     SICU/MICU/CCU - The major disposable products used are catheters, tubing,
auctioning equipment, I-V bags, needles and syringes.  Catheters, tubing and
suctioning equipment come into contact with body fluids during use and must be
treated as potentially infectious.  In accordance with Universal Precautions,
needles and syringes are destroyed in a medical incinerator.  However, I-V

                                      496

-------
bags never come into contact with body fluids and remain uncontaminated during
use.  Therefore, plastic I-V bottles could be safely reused for a single
patient, and should be considered as a substitute for the I-V bags.

     5 South - Patient Floors - Outpatient Clinic - The disposable products
regularly used on the patient floors include suctioning equipment, tubing,
catheters, blood transfusion equipment, chucks, and dressing supplies.
Because of the inherent contact with blood and body fluids, these products are
assumed to have a high risk of disease transmission.  The only pollution
prevention option recognized is in the use of chucks.  Chucks act as linen and
surface protectors, absorbing blood and body fluids in order that the reusable
linens will not become grossly soiled and that surfaces will be easier and
safer to clean.  (Chucks are used in the laboratories as well to contain small
spills at work stations.)  Chucks are present throughout the hospital, and
DVA-Cin may want to review the use of chucks to determine whether their
availability has led to use in situations where they are not needed.

     Hemodialysis - The major disposable products in this ward are I-V bags,
tubing, gloves and dialyzers.  Most of these items have been discussed, and it
has been noted that the dialyzers are reused approximately 20 times before
their disposal.  The reuse of dialyzers has been found to be a common practice
in health care institutions.  An informal survey on the reuse of disposables,
conducted at the 1984 Georgetown University International Conference, showed
46 percent of the respondents reporting the reuse of this item in their
institutions (6).

     With respect to high-tech items, it is believed that hemodialyzers are
the only devices which have been studied in sufficient depth to show that
function is not impaired through reuse (7).  With this technical knowledge as
evidence of safety and reliability, hospitals are able to write policies
allowing dialyzer reuse as a waste reduction option.


ADDITIONAL OBSERVATIONS ON DVA-CIN'S POLLUTION PREVENTION EFFORTS


     The DVA Hospital has already realized many of the waste reduction
opportunities arising from product substitution and waste segregation
practices.  The hospital's standard use of wovens is a significant part of the
reason DVA-Cin's waste generation rates are so low in comparison to industry
average.  The use of wovens in surgery accounts for the fact that DVA-Cin
produces 50% to 65% of the waste normally produced during operations in
Cincinnati area hospitals.

     There has been some discussion on the reprocessing of glassware.  The
recycling of glassware from sodalime  (e.g., pasteur pipettes) may greatly
reduce the volume and weight of a hospital's current wastes.  However, a large
percentage of the glassware used in laboratories is made of borosilicate which
cannot be recycled with general consumer waste glass.  Also, despite the
reliability of disinfection from autoclaving there is a stigma ascribed to
medical waste that may restrict or eliminate recycling as a pollution
prevention option.  Community recycling centers should be contacted regarding
their policies for accepting waste glass from health care facilities.

                                      49?

-------
     The Outpatient Clinic uses both woven and plastic-coated paper gowns.
For most treatments the woven gown is a safe, reliable barrier against
infection and should be the primary gown worn by the Outpatient staff.  The
use of the paper gowns are best limited to those treatments, such as
chemotherapy, which require more stringent infection control and increased
personnel protection.
                    RESEARCH AND DEVELOPMENT OPPORTUNITIES
     In addition to assisting in the identification of pollution prevention
opportunities for health care facilities, a major concern for RRIIi in
conducting these studies has been to look for those areas in which research
and development may support advancing new alternatives.  In learning of the
concerns, difficulties and successes of the health care profession, RREL hopes
to expand IPA's experience in the medical waste area and provide a solid basis
for planning future research.  Suggestions for further research in the health
care industry are presented below:

     Evaluate Reuse Potential in Single-Use Devices - As stated earlier in
this paper, hemodialyzers are reused because they have been studied and
evaluated closely to determine that such reuse does not impair their function,
nor compromise patient safety.  There are indications that other disposable
products may not as yet have been studied in sufficient depth to make reliable
determinations of their suitability for reuse.  A cooperative effort could be
established between EPA and representatives of the health care community to
undertake this research and provide substantive data to either support or
reject reuse considerations for the items listed in Table 2.  Research data of
this kind would give health care professionals a firm basis on which to make
such decisions, as well as opening up a potential to uncover cleaner
alternatives to some of the disposable products being reused.

     Quality Assurance - There are also legal and ethical considerations
associated with the reuse of disposables.  Among which are the manufacturers
disclaimers of warranty for reuse.  There is agreement that manufacturers can
offer a higher assurance of sterility than an individual health care facility.
Research conducted by the EPA in cooperation with health care professionals,
other Federal agencies (such as the Food and Drug Administration), and trade
associations can form the basis for developing a protocol for reuse, giving
hospitals a standard under which to set down operating procedures and
institutional policies.

     Hidden Cost Factors - There appears to be some confusion in comparing the
relative costs of disposables versus reusables.  The unit cost of a disposable
does not represent the full cost of actually using that product.  Disposal
costs are becoming an ever more important factor as landfill and incineration
regulations become increasingly more stringent.  For its part, the reusable
also carries storage and handling costs.  The EPA may wish to conduct
analytical studies in conjunction with health care facilities in order to
quantify these costs as an aid in decision making.
                                      498

-------
     Development of Reprocessing Capacity - Two pollution prevention
alternatives cited by this paper involved reprocessing services that have been
diminished by the disposable revolution.  Space and labor constraints, coupled
with the general availability and convenience of disposables, appear to be the
major obstacles to on-site reprocessing of durable materials.  However, as
health care cost containment gains increasing importance, reprocessing may
become cost effective for some items.  The potential for promoting some
reprocessing capability should be explored, particularly in those areas
exhibiting a high density of medical facilities.

     Developing a Reusable Market - Certain bills in Congress to amend the
Resource Conservation and Recovery Act (RCRA) will require that Federal
agencies meet certain objectives for use of recyclable products..  The SPA and
DVA should consider working together in developing procurement guidelines for
the DVA which will stimulate the production and distribution of reusable and
recyclable products.


                                  CONCLUSION
     In the case of the DVA study, the Assessment Team was impressed by the
difficult challenges undertaken by the hospital professionals to perform their
duties of human care while attempting to minimize the impact of those
activities on the environment.  Follow-on discussions indicate that this is a
dynamic process for DVA-Cin, as they develop initiatives in training,
information sharing and cooperation with other Federal agencies.

     For its part, the EPA hopes to learn from future cooperation with DVA,
seeking the health care professionals' advice and guidance in planning and
implementing research programs to respond to the needs of the m&dical
community in the areas of hazardous waste, infectious waste, and other waste-
streams.  Opportunities to reduce these wastes do exist, and additional
opportunities will be uncovered through research.  Research will also provide
the data on which to make operational decisions of benefit to health care
facilities, while favoring environmental considerations.
                                       409

-------
                                  REFERENCES
1)   Guide to Waste Minimization in Selected Hospital Waste Strearns, USEPA
     Office of Research and Development and California Department of Health
     Services, November, 1989.

2)   Hospital Pollution Prevention Case Study, USEPI Office of Research and
     Development, 1990.

3}   Characterization of_ Medical Waste Generation and Treatment and Disposal
     Practices in New York and New Jersey, USEPA Region II, January 30, 1989.
     And Rutala and Sarrubi, "Management of Infectious Waste from Hospitals",
     Infection Control, 1983.

4)   "Reusable Linens": An Economical Alternative to Disposables", Hospital
     Material Management Quarterly, February 1984, pp. 7-26.

5)   According to the Center for Disease Control, a critical item is one that
     will enter the vascular system or any sterile area of the body.  An item
     is semi-critical if it comes into contact with only intact mucous
     membranes.  A non-critical item comes into contact only with intact skin.

6)   "Reuse of Disposable Medical Devices in the 1980s", Proceedings of the
     International Conference.  Institute for Health Policy Analysis,
     Georgetown University Medical Center, 1984, Appendix B.

7)   "Single Use or Reuse: What's the Answer?", OR Manager, October, 1985,
      p. 6.
                                      500

-------
                  A TEN YEAR REVIEW OF PLASTICS RECYCLING

                 by:    S.  Garry Howell
                       Risk Reduction Engineering  Laboratory
                       Cincinnati, Ohio 45268
                                  ABSTRACT

     A short history of the practice of plastics recycling as practiced in
the United States and Europe for the past ten years indicates that much
progress has been made in educating the public sector about the
environmental damage done by indiscriminate disposal  of plastic Items.
Recent legislation has made the collection of some discarded plastic items
more efficient, and has provided economic incentives to recover and reuse
waste plastics.  The methods of collection, separation, cleaning, and
fabrication of plastic wastes into useful and saleable items are discussed,
Also discussed will be examples of products made from recycled plastic.
                                    SOI

-------
                                INTRODUCTION


     Considering the enormous quantities of plastic material  produced in
this country, as shown in Figure 1, it is little wonder that discarded
plastics have become a  significant portion of municipal waste streams.
Figure 2 shows the composition  of municipal  waste in the United States; the
quantities given are percent by weight , for while plastics constitute only
6.5% (the General Accounting Office estimates 7.2%) by weight, the inherent
low densities of the resins and low bulk densities of plastic products such
as bottles are said by some to compose 32% of the volume of landfilled
waste. A large part of the waste is discarded packaging of one type or
another.  The low price of plastic shrink wrap, blister packs, bottles, ad
infinitum, has resulted in a throwaway mentality; we discard items which
appear to have an almost infinite life if landfilled,  and sometimes present
a hazard to marine life.  Incineration is an option for most plastics,
which with the exception of some of the halocarbons, have a high fuel
value; incineration recovers only a part of the energy content of the
materials since the energy consumed in resin manufacture is wasted.  Modern
plastic materials are almost universally derived from petroleum. Recovering
and recycling plastics means that in addition to the hydrocarbon content
saved, a large part of the energy expended in production is also saved.

     A common perception is that plastics are a relatively recent
invention.  In reality, the first synthetic plastic, Celluloid, was
developed over 120 years ago (1869), and Bakelite over 80 years ago, in
1909.  The volumes of these and subsequently developed plastic materials
such as cellulose acetate and poly(vinylchloride) (PVC) were relatively
small, and since they were made into semidurable items such as electrical
equipment or insulation, motion picture film, etc., they were not
considered a nuisance when discarded.  The "Plastics Age" might be
considered as beginning in the late 1940's when the high volume production
of low density polyethylene (LDPE) began.  The rapid growth of the industry
is illustrated in Figure 1, which includes all plastics, thermosetting
(once formed or molded they cannot be reshaped by heat), and thermoplastics
(which within reason can be recycled by remelting, and reshaped into
different objects).  Production volumes of the more common thermoplastics
are shown in Figure 3, along with the amounts of the four major types
recycled in 1988.

     This review will primarily be concerned with thermoplastics,
particularly those most often fabricated into disposable items which have a
very short usable life and are discarded shortly after purchase.  The
sources of wastes, means of recovery from waste streams, technologies for
converting the waste material, and uses and markets for the recycled
plastics will be discussed.

TYPES OF THERMOPLASTICS

     Thermoplastics are generally divided into two groups.  The first
group, often called commodity plastics, would include the polyethylenes
(PE), polypropylene  (PP), poly(vinylchloride) (PVC), polystyrene (PS),

                                     502

-------
their copolymers, and poly(ethyleneterephthalate) (PET).  These products
are low priced and produced in huge quantities; prices range from $0.50 to
$0.70 per pound.  The combination of low price and physical properties
explains their widespread use in disposable packaging, in one-time use
agricultural and architectural coverings, and in some engineering
applications where their physical properties are acceptable.

     The second group is called engineering plastics.  Characterized by
higher strength, resistance to heat, and impact resistance, they have
become widely used to replace metals in many automotive, appliance,
aerospace, and industrial products.  Prices of these range from $0.85 to
more than $10.00 per pound.  Their applications to more durable goods and
lower production volumes result in less of a pollution problem and will not
be discussed at length in this review.  An exception is PET, which has
excellent clarity, mechanical properties, and heat resistance.  PET could
be said to straddle the line between commodity and engineering plastics, as
it is the plastic used in the ubiquitous clear plastic soft drink bottles,
some hot filled food products, and in many mechanical applications as well.
PET can also be made into fiber (Dacron, Kodel, etc.) and films (Mylar,
Cronar).

                        RESIN MANUFACTURE
     The great majority of modern synthetic plastics are derived from
petroleum or natural gas.  Polyethylenes and polypropylene belong to a
class called polyolefins, since their starting materials (monomers)  are
olefins made by thermally cracking the larger hydrocarbon molecules in
natural gas liquids or petroleum fractions.  Poly(vinylchloride) (PVC)
might also be described as an oleflnic material, since it differs from
polyethylene only by having every second hydrogen on the chain replaced by
an atom of chlorine.  Generalized structures of the polyolefins and PVC are
shown below.
POLYETHYLENE
                                                    H      H
                                                    C	C
                                                    H      H
                     heat, pressure, catalyst *        polyethylene


* Low density polyethylene  (LOPE) is produced by contacting ethylene
monomer and an organic peroxide catalyst  (or more  properly, initiator) at
pressures of 15,000 to 50,000 psi and temperatures of 3008 to 575°F.
                                    503

-------
POLYPROPYLENE
      H
      C
      H
H
C
H
H
CH
H
           propylene
           heat, pressure, catalyst-*
H     H
C	C
      H
                                CH3


                               polypropylene
** Polypropylene is polymerized with  Ziegler-Natta type catalysts similar
to those used for high density polyethylene but at lower temperatures and
pressures.

POLY(VINYLCHLORIDE)
             H
      C! — C
       H
       C
       H
                               H     H
                           — C — C
                               Cl     H
          vlny chloride
            heat, pressure, catalyst***
                             poly (vinylch lor Ida)
*** Poly(vinylchloride)  polymerizations  are carried out with peroxides,
persulfates, or redox catalysts  at  relatively low temperatures and
pressures, the type of catalyst  and method of polymerization vary according
to the intended end use  for  the  product.

(C « carbon, H = hydrogen, Cl  =  chlorine)

     The polyethylenes can be  thermally cracked  or depolymerized  back to
lower molecular weight polymers,  or under  severe conditions, to oily
materials or to carbon and hydrogen.   This process has been proposed as a
method of recycling but  will not be discussed here.  PVC can also be
thermally cracked, with  the  evolution  of hydrogen chloride gas (HC1), which
is poisonous and corrosive.  In  common with other organic polymers,  they
can also be depolymerized or otherwise affected  by the ultraviolet portion
of sunlight and oxygen in the  air.   This is a slow process, and is made
even slower by special inhibitors (antioxidants  and ultraviolet
stabilizers) added before fabrication  into the final  product.   Without
these inhibitors many uses for plastics  (such as telephone cable jacketing
exposed to sunlight and  heat)  would be limited.

     High density polyethylene is catalyzed by heterogeneous catalysts
which are organometals such  as triethyl  aluminum combined with titanium
tetrachloride at 180°-250°F  and 150-200psi.
                                    504

-------
POLY(ETHYLENETEREPHTHALATE)

     PET is also a petroleum derived  product;  a copolymer of terephthalic
acid, (which is made  by  oxidizing para xylene), and ethylene glycol, or
other glycols such as diethylene  glycol  or cyclohexane diethanol.  A
simplified reaction sequence showing  the direct esterification reaction
(there are actually three  different processes all arriving at the same end
product) is shown here:
      H O C < ' V COH +
      terephtallc acid
I I  /—\ I  I
c <) c-o-c
   N	/       H
                c-o -

                H
             2    2


poly(ethylenoterephthalate)
     The fibers Dacron, Kodel,  etc.,  made from PET are chemically identical
to the resin made  into bottles.  Some attempts to incorporate them into the
recycling chain will  be covered in a later section.

POLYSTYRENE

     General purpose  polystyrene is a clear,  somewhat brittle resin.  Most
disposable polystyrene items are of resins made by the following reaction
sequence:
                                                  H
                                                  C
                 H
                 C
                 H
                • tyr«n»     h««t, pr*s*ur«, catalyst"**

**** A peroxide  such as benzoyl  peroxide.
                                                  poly«tyr«n»
     The familiar foamed  beverage cups,  insulated containers, and home
insulation  are  chemically identical  to the clear drinking cups used on
airlines.   The  foamed materials are made by incorporating a hydrocarbon
such as n-pentane during  the polymerization process.
                          SOURCES OF PLASTICS WASTE
DOMESTIC
     As indicated  in  Figure  2,  in 1984 the percentage of plastics in
municipal  trash  was about 7.2% by weight, or about 9 million tons - almost
                                    505

-------
double that present in 1976.  By volume, plastics amount to about 32%,
according to Ramani Narayan of Purdue University (1).  No assay of the
volumes of individual polymers (PET, PE, PVC, etc.) appears to be entirely
accurate, but it seems fair to assume that overall, the ratios of the
plastics made into disposable items such as bottles, packaging films and
containers, etc. would approximate the ratios of resins sold for these
purposes.  In 1988, over 16 billion pounds of plastics were used in
packaging.  This excludes the 935 million pounds of PET (polyethylene
terephthalate) sold for soft drink and specialty bottles (Z).

AGRICULTURAL

     The main uses of plastics in agriculture are in mulch films, feed
bags, fertilizer bags, and in temporary tarpaulin-like uses such as covers
for hay, silage, etc.  A growing use is temporary covering for fields that
are being fumigate'd.  This film and the various bags would appear to be
prime candidates for recyling, as they are reasonably clean and easily
accumulated.  Mulch and covering film are usually discarded after they have
begun to deteriorate, generally as the result of being exposed to sunlight.
Reprocessed resins produced with a high proportion of these films will have
poor physical properties.

INDUSTRIAL

     Industrial  and construction use of disposable plastics is difficult to
quantify as some plastic film is used for temporary enclosures and then
left in place as a vapor barrier, while some is removed and discarded.
Polystyrene foam is likewise widely used as insulation, with an almost
indefinite lifetime, but many industrial products are shipped with foamed
"popcorn" or molded packing, almost all of which is discarded.  The low
densities of these foams (1-2 lbs/ft3  vs.  soil  at  70) make  them  especially
undesirable in landfills, while at the same time making hauling costs to
recyclers prohibitively high.  Much machinery, lumber, plywood, bagged
goods such as cement and mortar are shipped with polyethylene film covers.
The majority of this could be easily recycled.  Shrink wrapped cases are
replacing corrugated paper cartons in many grocery items.  Groceries should
be prime candidates to initiate recycling programs, since they could serve
as collection centers as well as being generators themselves.

"FAST FOOD" CONTAINERS

     Foamed polystyrene is widely used in food packaging (meat and produce
trays) at the retail grocery level, but the largest part of this 725
million Ib/yr market is in "fast food" containers.  The volume of waste
generated from this 725 million pounds is tremendous, for the density of
the foamed material is only 2 to 3 pounds per cubic foot, which would
indicate an uncompressed bulk of 13 to 20 million cubic yards occupying
space in our landfills.  A consortium of chemical and oil companies has
announced the formation of the National Polystyrene Recycling Company to
build five regional plants to recycle polystyrene (3).  The five plants,
said to cost $14 million, plan to collect the used material from
restaurants, hospitals, schools, and other big users in 30 states.  The


                                     506

-------
consortium is only planning to recycle 25% of the scrap available; there is
obviously room for others to enter the market, though finding large, easy
to collect volumes of scrap may be difficult.

MILITARY

     No reliable estimate of the amounts of recyclable plastics used by
domestic military installations is presently available.  Contact is now
being made with the Department of Defense to quantify their usage.  Because
of the concentration of these facilities, recycling should be easier than
for domestic waste.

AUTOMOBILE WRECKING

     The amount of plastics used in automobile manufacturing has been
growing at approximately a 9% rate over the past ten years.  Even in 1979
(the approximate date of production of autos now being scrapped) the amount
of plastics used in automobiles was 787,000 tons (4).  It would thus appear
that a large part of this could be recovered.  An investigation of this was
done in 1975 by the Bureau of Mines (5), but only a crude sink/float
technique was used to separate the plastics fraction, and little interest
was generated since good homogenous materials could not be produced.
Interest is growing in identifying the various plastics components in
automobiles with a molded or stamped-in label so that salvagers could more
easily segregate the types as they strip the vehicle prior to crushing and
grinding.  An evaluation of an innovative process to separate the various
kinds of resins from finely ground waste, such as from automobile grindings
is planned as a joint project between EPA and a large plastic producer.

WIRE AND CABLE

     A large source of recyclable plastic should be the electrical and
telephone cables from demolished buildings.  Until  recently, these were
burned to recover the copper or aluminum; now that strict emissions
controls are required on these incinerators, physical separation of the
plastic insulation would appear to be a viable option.  One site where wire
and cable was processed to recover the metal conductors is reputed to have
over 100,000 tons of mixed plastic "fluff" in a pile covering over 7 acres.
The amount of plastic recoverable from wire and cable is not known, but 433
million Ibs. of PVC, 386 million Ibs. of LDPE, and 125 million Ibs. of HOPE
were sold for insulation in 1988.

                            COLLECTION/RECOVERY


     Segregation and collection of waste plastics from agricultural
operations, industrial, and construction sites should not be difficult if
it were economically attractive to the generator.  These operations
typically generate relatively large quantities of discarded plastics in
small areas, so segregation would appear to be entirely feasible.
                                    SO?

-------
                               DOMESTIC WASTE


     A successful plastics recycling operation  must have a dependable
source of waste which is not overly contaminated with wet garbage, which
makes sorting and washing difficult.  The recycling plant should not be too
far from the collection points, since the cost of hauling the bulky
materials such as empty bottles or polystyrene foam would greatly increase
operating costs.  An efficient collection system is therefore mandatory,
and should operate over a large enough area (preferably urban) to supply an
adequate amount of feedstock.  Thus the collection of reprocessible
plastics from urban waste has proven to be the most difficult and costly
aspect of polymer recycling.  In the United States, early attempts to hand
sort municipal garbage in bulk by spreading it over a moving conveyer belt
proved to be unsuccessful, mainly due to the difficulty of finding
employees willing to perform the sorting for low wages.  The present trend
is for states and municipalities to enact legislation requiring
homeowners, and in some instances businesses, to presort trash.  Under this
system, separate containers are furnished for wet garbage, newspapers,
glass, metals, and plastics.  Several states, Michigan among them, have
enacted "bottle laws" which require purchasers to pay a deposit on all
beverage containers, whether or not they were meant to be returnable.  This
has resulted in greatly increased recycling of PET bottles; but other types
of containers such as PE milk and detergent containers, bags, film, etc.,
have been relatively unaffected although these items constitute about 80%
of all plastics trash.  The City of Cincinnati recently started a voluntary
recycling system, and has found that 65% of households are participating;
they had only predicted 35%.  It is apparent that the public is aware of
the need to recycle.  In Cincinnati, the only plastics that the recycler,
BFI, will take are PET bottles, (easily identifiable, being clear, usually
with an opaque base cemented to them), and polyethylene milk bottles, which
are mostly translucent white.  In mixed wastes, these would be worth the
extra effort to separate, as they can be easily hand sorted from a conveyer
belt.  In Europe, PVC bottles are widely used for wine and edible oils, and
some attempts have been made to separate them for recycling.  PVC bottles
are usually clear, and differentiating them from PET might be difficult for
an untrained person.

     As mentioned previously, a special case which is well worth mentioning
are the foamed polystyrene fast food trays and cups which have become a
litter problem, as well as occupying a disproportionate amount of landfill
volume.  Resin producers have formed at least two joint ventures to recycle
these discards into usable products (3).

     In contrast to American practice, the separation and collection of
recyclables in Europe is much more efficient and has a long history.
Kunststoffe German Plastics, a technical magazine which covers the
production, testing, and fabrication of plastics, devoted a large part of
its May 1978, issue to plastics recycling (7).  Since Kunststoffe is aimed
primarily  at the production and fabrication of plastics, the means used to
                                     508

-------
collect and separate plastics are not covered in detail.  The articles
covered in this issue discussed other aspects which will be covered later
in this review.
                          SEPARATION TECHNOLOGIES
MUNICIPAL WASTE
     Municipal  waste from which the "wet garbage" has been separated is the
feed to several recycling operations in European countries.  A flow chart
of a Dutch process which had been operating for a year in 1978, is shown in
Figure 4 (8).  The TNO system operates at 15-20 tonnes (33,000-44,000 Ibs)
per hour, and is designed to recover metals, plastics, and papers from a
mixed waste stream using a number of different operations.  Figure 5 is a
Sankey diagram of the TNO process, indicating the percent recovery of the
various input materials.  It is apparent from the flow chart that resorting
to mechanical means to separate the various fractions is very difficult;
one can count fourteen major pieces of capital equipment, exclusive of the
conveyers, etc. required to carry material from one operation to the next.
The final product is merely called "bales of plastic", apparently a mixture
of all the overhead from the zigzag air classifier.

     A process developed by SINTEF in Norway (9) is shown in Figure 6.
Note the similarities to the TNO process as far as the types of equipment
employed in the first stages.  Both preshred or grind the refuse, use
trommels to screen it, then zigzag air classifiers to separate the light
from the heavy fractions.  The SINTEF process takes the overhead from the
air classifier and passes it to a simple wash tank holding 250 liters (66
gal.)  of water where if is washed and separated into sink/float fractions.
In reality, this separation is relatively easy; after the ground particles
of plastic are freed of dirt, oil, grease and adhering paper, the
polyolefins (density <1) will float, while the PVC, other plastics, and wet
paper, etc. whose densities are greater than water, will sink.  While the poly-
 olefins all have a density less than one, and will be separated as a
mixture of HOPE, LDPE and PP, using the mixture to make even as prosaic as
a garbage bag is difficult.  This problem will be covered in a later
section.

     Even if plastic waste has been presorted at the curb, some final
classification is usually done to segregate the more valuable plastics from
the main stream.  One such system is described in a paper  recently
presented at Recyclingplas IV  (10).  The AKW plastics recycling plant which
has been operational in Blumenrod, West Germany since July 1988, is said to
have  its entire output sold. The Blumenrod plant, in the Coburg district,
collects waste from a population center of about 280,000 people.  Three
collection systems are employed in the district.  One, (Figure 7-1) has a
number of larger receptacles located in heavily used areas such as near
parks, city centers, etc.  The receptacles are marked for glass and for
paper, with the non recyclable material segregated.  Further separation
must  be done in a "Material Recovery Facility" (MRF).  The amount of
material collected from these points, while appreciable, is not nearly


                                     509

-------
enough to justify the construction of a recycling facility, so a second
system of curbside collection is also practiced (Figure 7-2).  Here the
recyclables are all placed in a single bin which is taken to the MRF.  This
system, which is presently used in  several U.S. cities and proposed for
others, requires the householder to separate the waste paper, metal, and
plastic into a "green bin", which greatly facilitates the sorting at the
recovery facility.  This system does not require the householder to make
the decision as to what constitutes the most desirable recyclables, but
does greatly decrease the amount of final sorting required.  The third, and
preferred system, requires some education of the householder, as a three
bin system is in place, is shown at the bottom of Figure 7.  In all three
cases a final sorting is performed at the recycling plant.

     Diagrams of some of the early recycling operations indicate that
little or no presorting was done.  Even if the sink/float separation
mentioned in the previous paragraph were used on a granular waste plastic
stream containing say, polyethylenes, polypropylene, foamed polystyrene,
solid polystyrene, PVC, and PET, the polyethylenes and polypropylene
(polyolefins) plus a large portion of the foamed polystyrene would float.
Attempts to fabricate useful items fom this mixture would probably be
futile, since the polyolefins and polystyrene are incompatible.  The
bottoms would have a mixture of all the others, and attempts to fabricate
useable products from them would also yield articles of both poor strength
and appearance.

     It is apparent that an efficient recovery operation must segregate the
various types of plastics either by hand sorting before grinding, or by
some mechanical means.  Figure 9 is a pictorial presentation of the AKW
sorting plant where mixed waste is hand sorted on moving conveyers before
grinding.  Although some attempt has apparently been made to presort at the
collection points, metals, cardboard, paper, wood, and plastics are hand
sorted again.  In this illustration, only two classes of plastics are
defined, rigid and film, although the text of Reference 10 uses the terms
plastic film and plastic bottle material.  Since 99+%  of plastic films are
polyolefin, no problem would likely result from compounding them into items
for resale if an intensive mixer were used.  Figure 9 does not indicate it,
but the plastic bottles should be further separated into three groups,
since not only are the three major types incompatible, but they represent
products of higher economic value if they are recovered in a relatively
pure form.

     Of particular interest to most American recyclers are PET bottles, as
they are easily identifiable, being clear, usually with an opaque base
cemented to them, while polyethylene bottles are mostly translucent white.
Both of these are worth the extra effort to separate, as they can be easily
hand sorted from a conveyer belt.  In Europe, PVC bottles are widely used
for wine and edible oils, and some attempts have been made to separate them
for recycling.  PVC bottles are usually clear, and differentiating them
from PET might be difficult.
                                    510

-------
     After separation or sorting,  the plastic fractions are finely ground
or shredded (in the earlier TNO and SINTEF processes the total stream was
ground then separated in an air classifier) this would appear to be a
relatively straightforward operation, but in practice  can be very
difficult.  Plastics are tough, and tend to tear rather than break cleanly.
Also, when grinding a waste stream, there are pieces varying in size from
one inch (25.4 mm) to pieces several yards or meters long; thicknesses may
be as little as 0.001 in. (0.0254 mm) or up to 0.5 in. (12.7 mm) this
variability requires grinding equipment with extremely close tolerances,
and the presence of dirt, grit, or other contaminants causes these to
change quickly.  Grinding also results in heat buildup which can change
knife clearances, and make the plastic more resilient and harder to cut
cleanly.  Some grinders operate with a gas stream flowing through the
grinding chamber (some specialized operations use liquid nitrogen) which
serves the dual purpose of cooling and conveying the material.  The AKW
system grinds the waste under water, which cools more efficiently and aids
in washing the plastic particles.

                    MECHANICAL SEPARATION/CLASSIFICATION


     Zig zag or cyclone air separators used in the earlier European
processes are not very efficient for classifying particles of low density
with high surface to volume ratios such as granulated plastics.  This,
coupled with the small density differences between the several types of
plastics (approximately 0.9 to 1.5gm./ml) yield an overhead product that is
a mixture of the feed material.  The AKW process uses hydroclones, which
are cyclones operating with liquids as the conveying media, as shown in
Figure 8.  The hydroclones complete the washing, and separate the solids
into a heavy fraction which in their operation is primarily PVC and PS
(polystyrene), and a light fraction composed of the polyolefins (low and
high density PE and PP).  These fractions are then dewatered in a
centrifuge and dried, both operations being carried out in conventional
commercially available equipment.  An overall view of the AKW MRF is shown
in Figure 9.

                                 PET BOTTLES


     Another special case in the municipal waste classification is the
ubiquitous PET soft drink bottle.   In 1979, 1.5 billion of these containers
were produced, consuming about 150 million Ibs. of resin  (11). PET is now
used in many other types of containers, from salad oil to peanut butter;
the 1989 usage is expected to be 1.02 billion Ibs.  These bottles are
easily identified in waste (and often as roadside litter), and the recycled
resin has a high resale value.  These facts have resulted in PET becoming
the most profitable of plastics recycling efforts, aided by mandatory
recycling laws such as the one in Michigan.  Another incentive is offered
by machines which return cash for each bottle inserted; the machines grind
the bottles for periodic collection by the recycler.
                                    511

-------
     Host PET bottles have two piece bodies, the clear or slightly tinted
bottle proper, and a base molded of high density polyethylene.  The caps
and seal rings (which stay on the body after opening) are aluminum with an
ethylene-vinyl acetate gasket or liner.  Earlier bottles had a glued on
paper label, which are now being superseded by shrunk on polyethylene film
labels.  Several recyclers have been in operation since 1980.  Most of them
grind the bottles, air separate the paper and fines,then use a sink/float
process in water to separate the PE base from the PET and aluminum.  The
Centerfor Plastics Recycling Research (CPRR), has a pilot plant in
Piscataway, NJ which uses an electrostatic separator to remove aluminum
after the above separation steps (12).  CPRR is offering the process for
license at a nominal cost.

                              PRODUCT FINISHING


     Although some plastics fabricators have equipment capable of extruding
or molding articles from the fluffy granulates, the majority are used to
handling pelletized resins.  It is therefore preferable to pass the dried
granulate to an extruder to melt blend or homogenize the mixed plastics and
to form small pellets which are much more easily fed to the final
fabrication step.  Plastics extruders are deceptively simple in appearance
and operation, consisting of a heated cylinder enclosing a long auger like
screw.  The screw is specially configured for each type of plastic, and
sometimes has special mixing flights, interruptions  for the removal of
volatiles from the plastic melt,etc. The design of extruders and their
ancillary equipment is a complex mixture of art and science, and they must
be chosen with great care.

     Extruders used to blend and pelletize the granulate are usually fitted
with a "crammer feeder", which is an auger to force the fluffy granulate
into the feed section of the extruder proper.  As the mi.xed plastic
particles are  forced through the heated cylinder, they are heated both by
conducted heat from the cylinder, and by friction as they are compressed by
decreasing the depth of the screw flights.  The frictional heat is of such
magnitude that most extruders only use external heat for startup; while
running they must be cooled by air or water on the external barrel surface,
and water into the center of the screw.  Feedstocks such as the granulated
mixed plastics wastes often have some water or other volatiles in them
despite the predrying step, and may require devolatilization in the
extruder.  This is accomplished by decompressing the melt near the middle
of the barrel length, where a vacuum is applied.  The melt is then further
homogenized, passed through a fine screen, and exits the extruder through a
multi holed die, where it is cooled, solidified, and chopped into pellets.

     Much emphasis has been placed on homogenizing the molten plastic at
this stage of the operation,  Non crystalline polymers of similar chemical
structures can often be easily melt blended and molded or extruded into
useful products of reasonably high quality.  PET is an excellent example;
it is routinely recycled into molded products or carpet fibers.  Mixtures
of crystalline polymers such as the polyolefins are much more difficult to
homogenize, since each of the three components has a different melting


                                   512

-------
point (low density polyethylene 110° C.,  high density 120°,  and
polypropylene 140° C.).  When such  a mixture is melted by merely raising the
temperature without intensive mixing, the crystallites or spherulites of
each specie are not broken up, with the result that a thin film or fiber
will have small particles (gels) of polymer or dissimilar polymers such as
PVC, oxidized materials^ etc. visible to the naked eye.  Such impurities
sometimes act as stress risers, weakening the product, and at the least are
unsightly.  It is therefore extremely important that a good homogenous
blend of any recycled product be made at the recycling plant.  Many good
homogenizing extruders are commercially available, and while not able to
make virgin quality product from scrap, will still make good recycled
material.

     To this point, only regrinding and pelletizing in preparation for
reprocessing has been considered.   Another possibility for recycle exists,
which is to depolymerize the resins back to their monomers and either
remake them into .polymers, or find other uses for the molecular
fragments,13' 14>    some of which may not be polymerizable.   Polyolefins may
be thermally depolymerized or "cracked", yielding a mixture of liquid and
gaseous products.  The latter will usually consist of the original monomer
(ethylene, propylene, etc.); but most of the mixture will be composed of
many higher olefins unless extremely severe conditions are used. Halogen
containing polymers such as PVC will evolve hydrogen chloride (HC1) when
heated, become brittle, and finally insoluble.  Polyesters such as PET are
easily depolymerized by hydrolysis in glycols or even water at high
temperatures and appropriate pressures.  Eastman Chemical Products Inc. has
demonstrated that they may be depolymerized in propylene glycol (16), then
reacted with acids such as maleic to produce a thermoset polyester resin.
Goodyear and duPont are also doing research using the depolymerization
approach.

                     APPLICATIONS  FOR RECYCLED PLASTICS


     Pending development of better means of sorting the various types of
plastic materials  into relatively pure fractions, it appears that at the
present only two types could be considered for making into high quality
objects, PET and HOPE.  High density polyethylene (HOPE), from such objects
as discarded milk containers, can be easily distinguished from other
components of a waste stream, or sorted at the curb.  After the bottles are
reground and cleaned they may be recompounded into pellets for remolding or
extrusion.  The cost of such recompounded material is 25-35% less than
virgin resin, and  if strict quality control is performed by the
recompounder, physical properties may be in the same range as the virgin
polymer.  It would seem that several non appearance automotive items could
be made from reprocessed material, such as windshield washer fluid
containers, fender and trunk liners, etc.  Procter and Gamble has pioneered
among consumer goods companies in the development of recycled plastics
packaging.  Recycled HOPE is used to make test quantities of one gallon
jugs for CheerR detergent.   These  containers are made in a three layer
                                    513

-------
extrusion; virgin resin is used for the outer appearance layer, reclaim for
the center, and again a layer of virgin for the inside in contact with the
product.

     Polyethylene terephthalate (PET) is already being recovered in such
purity that it is suitable for some fiber uses such as pillow or cushion
fillers.  Procter and Gamble has tested clear PET bottles for Spic and
SpanR Pine cleaner.   There are a number of automotive or industrial parts
which could be made of reprocessed PET, either clear (radiator overflow
tanks) glass filled (filler caps, terminal blocks) or compounded with other
resins such as low density polyethylene (LDPE) which improves
processibility, lowers moisture absorption, and also elevates some
mechanical properties.  (LDPE/PET blends have lower moisture absorption
than PET alone).

     Most lead-acid automobile batteries are contained in polypropylene
(PP) cases.  Cases from batteries being reclaimed for their metal value are
ground to 1/8" pieces, washed, mixed with a foaming agent, and extruded
into I beams for bed rails (17). To quote from this article in Modern
Plastics "One rather curious factor emanating from this reclamation project
is a morphological change in the scrap PP which is thought to stem from the
residual lead and sulfur imparted by the battery environment.  The precise
nature of the chemistry has not been determined, but the result is definite
improvements in such properties as impact strength, flexural modulus, and
heat deflection."  The above phenomena might also be related to results
reported by Zamorsky and Muras (20) when PP was repeatedly extruded and the
melt viscosity only decreased by 4.4%, while the carbonyl index during
weathering was relatively unchanged.

     The example of the reclaimed PP given above is unusual, for reclaimed
materials usually have physical properties inferior to virgin resin.
Nearly all molders of products such as battery cases will use  resins of
about the same molecular weight and molecular structure.  When
recompounded, these polymers will retain essentially the same structure, so
they will be physically compatible, and retain most of their original
properties.  Clear plastic materials picked from domestic trash on the
other hand, will appear to be uniform, but in reality the many manufactured
items will be made from a variety of resins having vastly different
molecular weights, copolymers (such as ethylene-vinyl acetate), pigments,
and other additives.  Despite thorough washing in systems such as the AKW
process mentioned above, some of the products were used to package motor
oil, vegetable shortening, or other products which absorb into the
polymeric matrix and change its properties.  This sensitivity to
contamination is much more evident in crystalline polymers, a class which
includes the polyolefins (polyethylene, polypropylene, etc.).

     Assuming that the reclamation process is efficient enough to separate
plastics from the other constituents in a waste stream, and that further,
these are separated by some mechanical method which yields two streams, one
having a density of less than 1.0, the other above 1.0, a number of
marketable items could be produced. A few such items and sources of raw
materials are shown in Table 1.


                                    514

-------
                           CONCLUSIONS


     The continued buildup of plastic wastes in landfills has become a
major source of concern for many sectors of society.  Plastics
manufacturers and their trade organizations have begun to realize that they
must help diminish the problem, and turn their considerable technical
skills to this end.  Cooperation between these organizations, academia, and
EPA has been excellent, and together, the amount of plastics recycled
should be greatly increased.
                                    515

-------
      Product
              TABLE  1

PRODUCTS MADE FROM RECYCLED PLASTICS

      Starting Materials
Reference
Structural Concrete
(walls, etc.)

Dunnage
(loose or molded)

Drainage Pipe
Bed Rails
Fence Posts or
Floors for Farm
Buildings.

Floor Mats, Wheel
Chocks, Pads, Truck
Bed Liners,
Drainage Pipe
Telephone Conduit

Composite Board
Polyester Polyols
(raw material for
polurethane foams)

Garbage Bags
   Foamed polystyrene (PS) +
   Portland cement.

   Reground foamed PS food trays,
   refoamed, molded.

   Foamed PS food trays, iron oxide,
   nucleator for azo foaming agent.
   "PVC blister packs, bottles,  circuit
   board holders -f virgin PVC.

   Recycled lead-acid battery cases,
   0.5% foaming agent.

   Polyolefin waste (min 50%), up to
   20% PVC, 20% paper +fearning agent.
   Copper contaminated wire & cable
   scrap.
   Plastics auto scrap, sawdust,
   (or chopped glass) 5% MDI-isocyanate
   binder.

   PET bottle scrap or polyester fiber
   from discarded clothing, chemically
   digested.

   Polyolefin waste from household
   and industrial sources
    8


    8


    12




    17


    13
    14



    15



    16


    18
Greenhouse Film
   Clean reclaim film blended with
   linear low density PE
                                                                 19
                                    516

-------
   U.S. PLASTICS PRODUCTION
        BILLIONS OF POUNDS
  1978
            20
Source: Modern. Plastics
  38.248
 40
FIGURE 1.
                  517
60
80

-------
     COMPOSITION OF MUNICIPAL WASTE
                    Percent by Weight
                               Paper & Paperboard41.0%
Yard Waste 17.9%
         Metals 8.7%
                Glass 8.2%
                                          Miscellaneous .6%
         Plastics 6.5%

      Food Waste7.9%
Rubber, Leather, &c8.1%
     Somce.ll.S.EPA
                       FIGURE 2

-------
    PRODUCTION OF MORE COMMON PLASTICS
          vs. VOLUME  RECYCLED  (1988)
      Low Density PE

      High Density PE

       Polypropylene

  polyolefins recycled
         Polystyrene

            recycled I 0.08$
   P o ly(vin ylchloride)

      "     recycled

   Poly(ethylenetere-

          phthalate)

      "     recycled
Source: Modern. Plastics
1    4    •    «

  BILLION POUNDS

   MQURI 3
                          518

-------
                       HOUSEHOLD REFUSE
                                                                                 FIGURE  4

                                                       TNO/ESMIL   WASTE   SORTING   PROCESS
                                                                                                                                                           20
    I feeojng bunker
    2 Shi«tf(flor
    3 Feeding (xirimt
    * Bui tssWff«1
    5 Ben cofvr,t<. 2
    6 Migner>c stjjjrjiof
    7 B«ll con,e,n - 3

    S Sh'eoaer to* tins
   10 8e« convey** - *
   11 Magnetic itpaiati
   12 Bw convejer - 5
   13 mmmef imn (ouihef)
3  U F««*ng Donne,
a  IS Beilcony^n.6

   " floto«iocit-1

   19 Bw-ton»tyM -1
   20 Fin - t
   21 Cyclone. 1
   22 Rotor LOCK • 2
   23 Rolling $Gfetn

   25 But con»tn». 9
   26 Co«*tyei cnol*

  28     I
  29 6« ton.tyti •
  30 «oio» leek - 3
  31 Fin-2
  32 Cyaoni.2
  33 Boio> La* • 4
  34 But' lor pasK
  35 Ro'lfif CQftvVytf
  36 Zig-itg ciMsAt
  3/ f etovig Byn»n
  18 Ben convevw -
  39 0»yet
  40 Gu burnic
  41 Fin - 3
  42 Cyoont. J
  43 ROW Lock
  44 Hood
  4S Fin - 4
  46 Cydone.4
  4J Role* LOCk . 6
  48 Bueriof pi»e»
 4i Roller coiwtytt - 2
                                                                                                                                                                          28
                                                                                                      i	•	~	
                                                                                                      I   BALES  OF PLASTIC

-------
                          Solid household refuse
input
magnet
mill
zig-zag classifier
Siflvi
f\
i
moistening i 	
zig-zag classifier
dryer
alter separator
San

)< •

L

!
:l

v r
\ 36%
K / heavy traction
»^. i r
>„.,
25%
Sine iraoon
3" V, "" '
..
t*
i:
K
.•»
"»
u

il . 8%


iL ^

J
z
T
.
v is%
>paoef fraction
" approx. 10%
k«y Diagram Esmil - Recycling • TNO Systnm masiure (airary)
Sas»d on wwqnt oafeacitages
Based on trie average composition ol me refuse, trie plant gives ine
ion owing results:
degree of separation degree of
on airary oasis cleanliness
paper
piastic folio
steel-lins
(me fraction
70 • 80%
70 - 80%
95%
•-•
approx. 5% plastic ano straw
09% sano
1 % mainly paper
5%
-
Source: NATO/CCMS 123
FIGURE  5
                                      521

-------
                SCREEN
                           WATER
 SHREDDED
  REFUSE
              ACCEPT
                                                ACCEPT
       FROM CYCLONE
                                                    DRYER
   GRINDER

w
J
1,
ASHING TANK

1

MEC
ncu
                                DEWATERING
                                                                      CYCLONE
                                                                        TO
                                                                    PURIFICATION
                                                                             UGH*
                                                                            FRACTION
                                                                       AIR-
                                                                    CLASSIFIER
                                                               HEAVY
                                                              FRACTION
                                                                    CYCLONE
                                                                 TO GRANULATION
PROCESS   FOR   MECHANICALLY  SEPARATING   PLASTICS

                        FROM  DOMESTIC   WASTE

                                     (S1NTEF)
                                 FIGURE 6
                                          522

-------
         AKW Apparate -f- Verfahren GmbH
     Post-Consumer Waste Collection Systems
               (Consumer Pre-Sorted)

 Central Collection Points
    Direct to Processing Facility 10-23%
                       Remaining
                       Waste Volume
                       90-75%
'Green Bin' System with MRF
   .	(MATERIAL RECOVERY FACILITY)
   To Processing
      Through
Master Recovery System    Remaining
     up to 45%         Waste Volume
	    up to 55%

        +               ^

                                Oetn iin
Multiple Bin System without MRF
   To Processing
 Through After-Sorting
     uo to 45%
Remaining
Waste Volume
up to 55%
                   FIGURE  7
                                ?3Otr CUss 4
                                 3in .M«al Bin
                      523

-------
                         AKW Apparate + Verfahren GmbH
                                         Sorting Plant
                                         for Green Dustbin and
                                         Industrial Waste
en
•sa>
                                                            Industrial Wasle


                                                          I lousehold
   Pre Sorted Paper
                                FIGURE 8

-------
         AKW Apparate + Verfahren GmbH
              HYDROCLONE
                       Overflow
  Feed
Internal Vortex
External Vortex
                    Underflow
                 FIGURE 9
                  525

-------
                           REFERENCES

1.   Science News 135 282, May 6, 1989
2.   Modern Plastics 66 5, p.42,  1989
3.   Environment Reporter June 16, 1989
4.   Modern Plastics 56 1,  p.69, 1979
5.   Valdez, et al Bureau of Mines Report of Investigations RI 8091  1975
6.   Goldbaum, Ellen Chemical Week 144 19 p.9 1989
7.   Kunststoffe 68 5, p.23 1978
8.   Waste Plastic Recycling Information Exchange Volume II, Contributed
  •   Papers, NATO/CCMS Report 123, p. 3-15 1981
9.   Ibid, p. 4-1
10.  Stolzenberg, Andreas  Paper presented at Recyclingplas IV
     Washington, D.C. May 24, 1989
11.  Modern Plastics 57 4, p.82 1980
12.  Ibid. 64 2, p.15 1987
13.  Barnes, B.A. et al Chemical Engineering 87 12,
     pp50-51 1980
14.  Modern Plastics 57 4, p. 82 1980
15.  Ibid. 57 5 p.64 1980
16.  Carlstrom, W.L. et al Modern Plastics 62
     5, p. 100 1985. Also U.-S. Patents 4,223,068 and 4,417,001
17.  Modern Plastics 55 6, p. 61 1978
18.  Brochure Published by AKW GmbH Hirschau, W. Germany
19.  Ram, A. and Getz, S. J. Applied Polymer Science 29
     2501-15 August, 1984
20.  Zamorsky, Z. and Muras, J. J Polymer Degradation and Stability 14
     41-51 1986
                                    526

-------
              EVALUATION OF EPA WASTE MINIMIZATION ASSESSMENT

              by:  Mary Ann Curran
                   Kenneth R.  Stone
                   RREL, U.S.  Environmental  Protection Agency
                   Cincinnati, Ohio  45268
                                  ABSTRACT
     EPA's research efforts to encourage the use of waste minimization
opportunity assessments is presented in this paper.  The early stages of
EPA research centered on the development of the EPA-recommended procedure
for conducting an assessment, and its use at a number of facilities.  This
paper will demonstrate the value of the waste minimization assessment for
discovering and developing opportunities to minimize wastes by presenting
briefs on assessments recently conducted at several private concerns.

     These private concerns include a photo lab, a truck manufacturer, a
large church facility, a nuclear power generation facility, and a graphic
controls manufacturer.  These assessments are at various stages of
completion.  The status and results of each assessment are presented.

     Based on completed assessments and discussions with company
representatives, the EPA-recommended procedure as described in the "EPA
Waste Minimization Opportunity Assessment Manual" is evaluated for its
usefulness.  Reluctance in initiating an assessment has been encountered,
but when technical support was provided to conduct an assessment, company
representatives expressed strong satisfaction with the results.

     This paper has been reviewed in accordance with the U.S.  Environmental
Protection Agency's peer and administrative review policies and approved
for presentation and publication.  Mention of trade names or commercial
products does not constitute endorsement or recommendation for use.
                                     527

-------
                                INTRODUCTION
     While the word "assessment" often raises the fear that what is being
talked about is an environmental audit, the assessment team is not looking
for incidences of facility non-compliance.  Their purpose is to examine a
process and its components for inspiration to develop techniques that would
enhance the cleanliness of a particular process or operation.  To
accomplish this goal, certain team members must have technical background
appropriate to the type of process they are assessing.  Therefore, a
knowledge of RCRA compliance and SARA Community Right-to-Know regulations
is not required of the assessment team.

     Conducted by an in-house team or with an independent outside
consultant, a waste minimization opportunity assessment (WMOA) is simply a
structured review of a process or operation to lead to identified opportun-
ities for waste reduction or recycling.  Its focus can be broad or narrow.
Often, it is more effective to select a few areas for intensive assessment
than to attempt to cover all waste streams and processes at once.

     The EPA has published "The Waste Minimization Opportunity Assessment
Manual" (EPA/625/7-88/003) for conducting a waste minimization assessment.
This manual is available at no cost from the EPA's Pollution Prevention
Research Branch, Cincinnati, Ohio 45268.  The procedure recommended in the
manual is outlined in Figure 1.  WMOA's are an extremely effective way to
improve a facility's operations, from both an economic and environmental
standpoint.

     The following sections on industry assessments,  "church assessments",
and New Jersey assessments represent descriptions of some of the assessment
efforts currently being conducted by the EPA.

                            INDUSTRY ASSESSMENTS


     The focus of the industrial assessments effort has been on locating
small and medium-sized facilities which may not have the immediate
resources or expertise to do what is necessary to reduce their waste  and
would benefit significantly from Agency support.  Toward this goal,
assessments have been conducted at a mini-photo lab and a truck
manufacturing facility.  Both hazardous and non-hazardous wastes are
included  in the assessments.

     Details on the mini-photo lab and the truck manufacturer assessments
are provided below.

MINI-PHOTO LAB

     After an assessment in August 1989, the assessment team identified
five waste minimization options they considered applicable to the waste-
streams of interest.  Following is a brief description of these options.
                                     528

-------
    Option  1  - Wash  Water  Control  - Wash water  is  used  for  color  film
    development  and  the  B&W  paper  process.  The wash water  is  turned
    on  each production day at  approximately 7:00 a.m. and shut off at
    7:00  p.m.  Water use is  therefore  continuous during the day,
    however,  production  is not.  The waste minimization option
    consists  of  a  simple timer control  system consisting of a  switch,
    timer and solenoid valve.  The operator would  punch a button  on
    the switch to  activate the timer.   In turn, the activated  solenoid
    would allow  water to flow  for  a preset time period.

    Option  2  - Silver Recovery/Metal Replacement Cartridges -  Silver-
    is  found  (as light-sensitive silver halide) in spent photographic
    chemicals and  wash waters  as a result of removing the emulsion on
    films and papers. A metal replacement cartridge is a widely-used
    device  for silver recovery.  It can be used alone or in
    conjunction  with other recovery technologies.   In this  case,  the
    spent process  solutions  which  contain significant amounts  of
    silver  would be  plumbed  to a single pipe.   Two cartridges  would be
    used  to allow  for high capacity while maintaining a high recovery
    rate.

    Option  3  - Silver Recovery/Electrowinning  - An electrowinning unit
    passes  a  direct  current  through a  concentrated silver solution
    from  anode to  cathode  causing  the  silver to plate out onto the
    cathode in nearly pure metallic form.  A wide  range of  equipment
    is  commercially  available  for  electrowinning.  Using
    manufacturer's literature  as a basis, it is expected that  up  to
    two batches  (4 gallons each) can be treated each day.   During the
    average batch, 1.13  troy oz. of silver would be recovered  within
    4.5 hours.

    Option  4  - Silver Recovery - This  option is based on using the
    electrowinning device  in Option 3, with metal  replacement
    cartridges used  to  polish  the  effluent.  The average effluent will
    be  desilvered  from  500 mg/Lto approximately 10 mg/L, using only
    one cartridge.

    Option  5  - Bleach Fix  Recovery - The recommended method for bleach
    fix recovery is  desilvering with two metal  replacement  cartridges.
    This  requires  three  steps: 1)  silver recovery, 2) restoring
    bleaching ability by aerating  ferrous-EDTA  complex  to oxidize back
    to  ferric-EDTA,  and  3) replenishment of chemicals lost  through
    carry-over with  the  film or paper. Approximately 75% of the
    recovered bleach fix solution  can  be reused while 25% should  be
    discarded to prevent contaminant build-up.

    Total capital  investment,  net  operating cost,  and payback  period  for
each option  are  shown in Table  1.   The  owner  of the lab  has  received a copy
of the final assessment  report  and  is  taking  the recommended options under
advisement.
                                    52S

-------
       TABLE 1.  SUMMARY OF MINI-PHOTO LAB WASTE MINIMIZATION OPTIONS
                                 TOTAL CAP.       NET OP. COST    PAYBACK
WASTE MINIMIZATION OPTION	INVESTMENT. $    SAVINGS, S/YR   PERIOD. YR
Wash Water Control               $  675            $ 1,436        0.47

Silver Recovery Using             1,071              1,325        0.81
Metal Replacement Cartridges

Silver Recovery Using             3,510              1,414        2.48
Electrowinning

Silver Recovery Using             3,667              1,757        2.08
Electrowinning with MRC
Tailing

Recycle of Bleach Fix and         1,571              2,508        0.63
Silver Using MRCs


TRUCK MANUFACTURER

     This truck manufacturing facility produces 34 trucks (tractor-trailer)
per day.  The production processes are primarily assembly and painting.
The current quantities of generated wastes and the associated disposal
costs for the first three quarters of 1989 are given below:

                                         Amount        Cost of
                                          fib)         Disposal

     Waste Paint                        184,860        $12,957
     Pretreatment Sludge                 71,020        $ 9,134
     Undercoating                         3,375        $ 2,560
     Degreasing Solvent (Chlorinated)    13,060        $ 5,431
     Used Oil                            28,275        $   105
     Paint Sludge                       474,960        $15,132
     Housekeeping                         3,800        $ 1,428


     The above figures represent a sharp decrease from recent years.   The
facility has instituted a number of waste minimization measures and cost
reduction methods related to good waste management practices.

     A site visit was conducted in January 1990 to begin the assessment.
Although this facility has made major strides in waste minimization,  the
assessment team feels there are additional opportunities which may have
significant impact.  The following are targeted areas which will be
investigated further throughout the assessment and feasibility phases.
                                    530

-------
     Spray Painting - Air-assisted airless spray equipment is used for most
     spray painting.   This method is a distinct improvement over
     conventional  compressed air spray painting,  however,  alternatives
     exist which may improve transfer efficiency.   Increasing the transfer
     efficiency reduces the volume of paint used and reduces volatile
     organic carbon (VOC)  emissions.

     Phosphating - An automated phosphating (conversion coating) process
     and electro-coat (E-coat)  is used for small  and medium-sized parts.
     This line consists of several processing and  rinsing  steps.  The rinse
     water is piped to a chemical treatment plant  where it is combined with
     paint booth wastewater.  The resultant sludge is disposed as a
     hazardous waste.

     It may be possible to avoid waste treatment of the phosphating rinse
     water by using an ion exchange recycle system, thereby also reducing
     water usage.   Furthermore, the current wastewater treatment process,
     which uses large amounts of ferric chloride,  may be altered, resulting
     in reduced sludge generation.

     Degreasing of Rail Frames - The rail  frame,  or chassis, is degreased
     prior to spray painting using a chlorinated solvent (90% 1,1,1,-
     trichloroethane/10% methylene chloride).  The spent solvent is
     distilled (350-400 gallons per day) and reused.  Waste minimization
     options may include chemical substitution, procedural changes, or
     improvements to the recycle process.

     The assessment team is completing the feasibility phase and a draft
report is expected in May 1990.

                             CHURCH ASSESSMENT


     This study of a church facility looked at daily office operations,
special functions, general maintenance and an on-site pre-school. As would
be expected, churches are not normally large waste sources, however, they
are a tremendous source of social awareness.   It is anticipated that this
assessment and suggestions for waste minimization will result in wide-
spread distribution through the church's governing bodies and congregation.
This information will impact not only other churches, but also people's
activities at home and at work.

     The location of the church assessment was the Mt. Washington
Presbyterian Church  (MWPC) which is about fifteen miles east of downtown
Cincinnati.  With a  1990 budget of $615,000 and a $3,000,000 renovation and
expansion project, this 2,000-member church represents a substantial insti-
tutional facility.  The church has a very aggressive Recycling Committee
which has been active in collecting recyclable material for the community.

      The site visit was made in December 1989,  The specific areas of
concern included building and grounds maintenance, pre-school, social
activities, kitchens, administrative offices, and new building expansion.

                                    531

-------
     Predictably, the largest waste generated by the church is white paper.
However, there are numerous other cleaners, paints, lawn materials (e.g.,
weed killer), etc., that are used in significant quantities.  The final
report describing waste minimization options is expected to be available by
the summer of 1990.

                           NEW JERSEY ASSESSMENTS


     A pilot project with the New Jersey Department of Environmental
Protection (NJDEP), entitled "Assessment of Reduction and Recycling
Opportunities for Hazardous Waste (ARROW)," will allow the State to
evaluate waste minimization techniques and conduct assessments at
approximately thirty facilities within New Jersey,  The objective of the
site selection is to cover ten industries (three sites in each) to develop
industry-specific information through the assessment activities.

     Through a subcontract with NJDEP, the New Jersey Institute of
Technology (NJIT) is locating sites and performing the assessments by
following the EPA-reeommended procedure outlined in the EPA manual.
Participation in the program by facilities is on a voluntary basis.  To
date, response to the program has been enthusiastic,and 14 companies are
lined up for assessment work.  Four site visits have been completed and the
assessment reports are being prepared.  Brief descriptions of two of the
companies visited and potential waste minimization options follow below.

NUCLEAR POWER GENERATION FACILITY

     Interestingly, the bulk of the wastes from this electrical power
generation facility is from construction and maintenance activities when
power generation is shut down.  Three major sources of waste streams were
identified by the assessment team: operations, maintenance, and site
services.  After analysis of costs and waste generation quantities, the
assessment team targeted opportunities for reduction in the levels of off-
spec materials and containers of partially used materials which go to waste
treatment and disposal.  Several waste reduction options were identified,
such as improved project estimation and planning of material procurement,
dispensing, and stocking; incentives to contractors for waste reduction;
and improved security to protect against wastes imported to the site.

GRAPHIC CONTROLS MANUFACTURER

     This facility manufactures pens and markers for automatic recording
devices and inks for use in these devices.  The waste generation data
indicate that the operation for ink formulation and preparation contribute
to the bulk of the hazardous waste generation.  Some options leading to
reduced waste generation include reduction in quantities of rinse water
used in the cleaning of equipment; improved scheduling of colors and types
of batches of inks to reduce cleaning between batches; increased use of
mechanical cleaning of tanks to supplement water cleaning; and changes in
ink preparation procedure such as the utilization of a large ink base which
                                     532

-------
could be tinted to the appropriate color in smaller batches as the need
arose using small amounts of tinting color.

     NJIT continues to work with facilities who show a strong interest in
waste minimization and have volunteered to participate in the ARROW
program.  This effort will continue through August 1991.

                      ASSESSMENT  PROCEDURE  EVALUATION


     The EPA's assessment procedure itself is of interest regarding its
usefulness in identifying waste minimization opportunities.  As expected,
this issue is difficult to quantify, however, some general comments on the
effectiveness of the manual can be made from the assessments that "have been
completed within the EPA's waste minimization opportunity assessment
program.

     Discussions between EPA contractors performing waste minimization
opportunity assessments and company representatives led to identifying
three issues on how a company may be delayed in initiating an assessment
effort.  These three issues are as follows:

     1)   Implementation of the Manual's procedure is based on the
          establishment of firm corporate commitment.  This commitment is
          needed in order to devote necessary time and resources.
          Management may not be approachable if specific information is not
          available to them on what the company can expect in return for
          their investment.

     2)   To someone who is not familiar with the manual, it appears
          cumbersome and lengthy and looks like a lot of work in filling
          out the forms.  This impression may cause users to lay the Manual
          aside until a later time.  The user may or may not return to the
          Manual.

     3)   Occasionally there is the fear that the completed manual forms
          would become available to the regulatory agencies and become
          incriminating information.  This unfounded fear prevents people
          from filling out the forms which is a critical step in
          understanding a facility's waste generation and management
          practices.

     In situations where outside assistance was received in performing
assessments, company representatives expressed strong satisfaction with the
results.  They perceived the amount of time  involved in the exercise to be
appropriate.  The most valuable result was  the identification of ways to
lower costs, which  in turn helped convince management of the profitability
to be  incurred in an effective waste minimization program.

     It is reasonable to conclude that many manufacturers recognize their
waste generation problems and have undertaken actions to address and
improve them.  Toward this end, they are very receptive to any method such

                                     533

-------
as the one presented in the Manual which will assist them in meeting their
waste management objectives.

                                 CONCLUSION
     This paper describes several waste minimization success stories
arising from the EPA's Pollution Prevention Research Branch in Cincinnati,
Ohio.  The programmatic approach has been to go to industry to determine
the manual's implementation and to transfer technical pollution prevention
impacts throughout the community, especially to small and medium-sized
businesses which may not otherwise have the resources to pursue pollution
prevention initiatives on their own.  Furthermore, it is clear that EPA's
program has focused on practical approaches to already existing processes
and facilities.  Such an approach begs the question: What about the future?

     EPA's assessment program will continue to aid in the establishment of
a knowledge pool of individuals technically-oriented to pollution
prevention.  The assessment process is becoming an integral part of
business management practices, much as safety concerns have become routine.
Beyond these assessments, the Agency's pollution prevention research
programs must turn to identifying clean practices, clean products and
processes.  With the cooperation of representatives from private concerns,
EPA anticipates broad potential for research in alternative technologies
and products that lower risks to the environment and our future heritage.
                                     534

-------
WASTE  MINIMIZATION ASSESSMENT PROCEDURE
              Need to minimize waste
           PLANNING AND ORGANIZATION
          ASSESSMENT PREPARATION STEP
               ASSESSMENT STEP
           FEASIBILITY ANALYSIS STEP
                IMPLEMENTATION
              Successful implementation



                FIGURE  1
                       535

-------
                      SLURRY-BASED BIOTREATMENT OF
                  STILL BOTTOMS SORBED ONTO SOIL FINES


                 Robert C. Ahlert. PhD, PE, Dist. Prof.
                 David S. Kosson, PhD, Asst. Prof.
                 William V. Black, Doctoral Candidate
                 Juelide Suenter, Graduate Assistant

                  Chemical & Biochemical Engineering
                          Rutgers University
                  P.O. Box 909, Piscataway, NJ 08855
                         (201) 932-3399/4346

                                with

                 John E. Brugger, PhD, Project Officer
                 Risk Reduction Engineering Laboratory
                 U.S. Environmental Protection Agency
                          Edison, N.J. 08837

                              ABSTRACT

Research focus is on the biodegradation of organic contaminants on fine
fractions, separated from whole soil.  Leachates or aqueous extracts from
whole soil are of importance, also.  Waste activated sludge from
conventional secondary treatment provides mixed microbial populations for
semi-continuous slurry reactor configurations.  To date, operations have
been carried out under aerobic conditions.  However, at some point, a
sequence of aerobic/anaerobic steps will be evaluated.  Whole soil samples
were obtained from a site immediately adjacent to an impoundment used for
disposal of still bottoms.  This waste is the residue of benzene, toluene
and xylenes (BTX) production by distillation of a high molecular weight
petroleum feed stock under strong acid conditions.  Chlorinated aromatic
wastes have been detected in the otherwise predominantly hydrocarbon
materi al.

Contaminants are a diversity of simple aromatic and extremely complex multi-
ring compounds.  Thus, relative concentrations of representative compounds
are monitored during biodegradation.  In many instances, the analytical
method of choice is gas or liquid chromatogram peak height, as a function of
carrier, instrument temperature program and reference residence time.  Feed
rate and the addition of phenol as co-substrate are manipulated experimental
varaiables.  Significant reductions of marker species concentrations, often
to levels below detection limits, are observed with low, medium and some
high molecular weight PAH-type compounds, as well as the chlorinated
species.

This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.


                                    536

-------
                          INTRODUCTION

     Distillate bottoms and sludges from a benzene, toluene and xylene (BTX)
production process were impounded in two lagoons, over several decades.  The
process consisted of cracking naphtha in the presence of sulfuric acid.
Possible contaminants include, but are not limited to, simple aromatic
species, polyaryls species, polynuclear aromatic hydrocarbons (PAHs), as well
as sulfonated derivatives of these compounds.  One impoundment exceeds ten
acres in extent and contains an estimated 100,000 cubic yards of residues.
The contents of the lagoon have separated into several distinct layers that
include, in bottom-to-top sequence, a solid organic layer, a layer of viscous
organic matter, an oil-like layer, and a floating aqueous layer.  Many of the
waste species have slight solubilities in water and/or affinities for some of
the constituents of soil; these have migrated into and through the soil
immediately adjacent to and underneath the lagoon.  It is the contaminated
soil from such adjacent areas that is the subject of this investigation.

     The EPA has determined that concentrations of distillate bottoms
contaminants must be significantly reduced before the site can be closed.
Many methods of soil treatment are possible; these include chemical
extraction, incineration, and biological treatment.  A number of workers have
demonstrated that compounds typically found in distillate bottoms (aromatics,
polyaryls, and PAHs) are biodegradable under aerobic conditions.  Mass
transfer of these compounds from the soil to the microorganisms is usually
the most significant barrier to biodegradation processes.


     A biotreatment process using slurried soil fines has been developed.
The slurry-based system aids mass transfer by facilitating greater contact
between contaminated soil and microorganisms.  Obtaining the slurry of fines
involves a pretreatment process in which whole soil is fractionated; the
highly contaminated fractions are prepared as a suspension of soil fines for
subsequent biodegradation experiments [1],  Separating the soil fines reduces
the mass of material requiring additional treatment.  The biological
treatment process utilizes a mixed microbial culture in a slurry-based system
operated in a semi-continuous manner, under aerobic conditions.

                         LITERATURE REVIEW

     The distillate bottoms contaminants are a complex mixture of polycyclic
aromatic hydrocarbons, including aromatic compounds, such as benzene,
toluene, and xylene (BTX), and a complex mixture of bi- and polyaryl species,
such as bis(p-tolyl)ethane and l,2-dimethyl-4-phenylmethylbenzene.  Bjorseth
(1983) describes PAHs as fused five and six member ring compounds in which
the interlocked rings have at least two carbon atoms in common [2],  Bi- and
polyaryl compounds consist of two or more aromatic rings connected by a
single bond from a carbon atom of aromatic ring to a carbon atom of another
or to a carbon chain that connects to another aromatic ring.

    BTX and sixteen PAHs are considerd priority pollutants by the EPA.
Mammalian enzymes are thought to metabolize many of these compounds in


                                     S37

-------
carcinogenic substances (Cerniglia, 1984)[3].  For this reason, their
presence in soil and potential mobility into groundwater is a matter of great
concern. A more detailed discussion of the properties and hazards of PAHs is
given by Enzminger and Ahlert [4].

     Mackay and Shiu (1977) measured the aqueous solubilities of a number of
PAHs at 25C [5]. Table 1 lists the solubilities of some of these compounds. As
would be expected, solubility decreases significantly with increasing
molecular weight. The aqueous phase concentrations of these and other organic
compounds in a system composed of contaminated soil and water is strongly
dependent upon adsorptive/desorptive partitioning characteristics with
respect to soil constituents, water and biomass. Dzombak and Luthy (1984)
discussed the relation between soil organic carbon content and sorption [6].
They cited Lambert's (1965) studies demonstrating that adsorption of neutral
pesticides correlated with the organic content of soil [7]. The more organic
matter present, the greater the adsorption capacity of soil for organic and
nonpolar compounds. Water molecules compete poorly with nonpolar molecules
for hydrophobic surface sites in soil organic matter  [6].
           Table 1: SOLUBILITIES OF TYPICAL PAHs IN WATER [5]
Number
Compound of
Naphthalene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Chrysene
Benzo(a)pyrene
Dibenz(a,h) anthracene
Rings
2
3
3
3
3
4
4
4
5
5
Molecular
Weight
128
154
166
178
178
202
202
228
252
278
Molecular
Formul a
C10H8
C12H10
^13^10
^14^10
C14H10
C16H10
C16H10
C18H12
C20H12
C22H12
Solubility
fug/L)
31700
3930
1980
1290
73
260
135
2.0
3.8
2.5

     The solubility and rate of desorption influence bioavailability and have
significant effects on the rate and extent of biodegradation.  Biological
studies have been conducted on a number of PAHs and aromatic species
demonstrating the abilities of a variety of microorganisms to metabolize
these compounds.  Therefore, making the distillate bottoms'contaminants more
readily available should have a positive effect on the biodegradation of
these compounds.
                                     538

-------
     Bauer and Capone (1985) isolated bacterial populations from intertidal
marine sediments capable of aerobic transformations of naphthalene and
anthracene [8].  Herbes and Schwall (1977) studied the microbial
transformation of these two compounds as well as benzo(a)anthracene and
benzo(a)pyrene [9],  They found that two-and three-ring compounds were
readily transformed within two weeks while benzo(a)athracene showed little
and benzo(a)pyrene showed virtually no biodegradation after a month,  Bauer
and Capone (1988) later showed, when a microbial population is preexposed to
an individual PAH, the subsequent rate of degradation of that compound is
enhanced [10].  They also found that preexposure can enhance the rate of
degradation of other PAHs, indicating that selected populations have a broad
specificity for PAHs and/or possess common pathways for PAH metabolism.
Shocken and Gibson (1984) used Beijerinckia sp, grown on biphenyl to co-
oxidize acenaphthene and acenaphthylene [11].  Hihelcic and Luthy (1988)
compared the microbial degradation of naphthol, naphthalene and acenaphthene
under aerobic, anaerobic and denitrifying conditions [12].  All three
compounds were degraded quickly under aerobic conditions and more slowly
under denitrifying conditions; only naphthol was degraded under anaerobic
conditions.

     Within the last two years, evidence for biodegradation of higher
molecular weight compounds has been reported.  Shiaris (1989) measured rates
of transformation of benzo(a)pyrene in Boston Harbor [MA] sediments, and
observed half-lives of 53.7 and 82.3 days [13].  Heitkamp, et al. (1988)
observed the degradation of pyrene into dihydrodiols and pyrenols by
Hycobacterium  sp.  [14].  Heitkamp and Cerniglia (1989) later showed that in
addition to transforming pyrene, Mycobacteria play a significant role  in the
mineralization of benzo(a)pyrene [15].

                          MATERIALS AND METHODS

     The soil  slurry used for present experiments was prepared by the method
described by Ahlert, Kosson and Black (1989) [1].  The final separation step
was changed from filtration through a ten-micron mesh to gravity settling
followed by siphoning the slurry from the top of the settling vessel.  This
modification enabled larger quantities of slurry to be prepared at one time.
The initial extraction, carried out in a rapidly stirred five-gallon plastic-
lined pail, did not at first generate the third tar-like organic phase
observed in smaller scale experiments.   This was corrected by returning to
screening.  Gravity settling led to great variability in contaminant levels
and soil particle size distributions.

This screening method isolates from 5000 to  18000 milligrams of organic
material per kilogram of whole dry soil.  Assuming the tarry materials is 92%
carbon and recovery is approximately 50%, the TOC of the whole dried soil is
of the order of 20000 mg/kg or about 2%.  The Total Volatile Solids (TVS) of
whole dried soil is 49000 mg/kg or 4.9%.  Considering the apparent strength
of binding and the difficulty of separation, this agreement seems excellent
Five percent is probably an approximate upper bound on the contaminant level
of whole soil.
                                     539

-------
     Bioreactors have a working volume of two liters and are filled initially
with 50% (one liter) soil slurry, 33% (two-thirds liter) of aerobic culture
and 17% (one-third liter) of mineral salts medium.  The slurry contains
between 30 and 40 g/L of soil fines.  If soil fines represent about 40% of
the contaminants on whole soil, this corresponds to 60 to 80 mg of
contaminants per liter.  The composition of the mineral salts medium is that
required to achieve the following concentrations of salts in the final
reactor mix: ammonium sulfate, 500 mg/L; magnesium sulfate, 100 mg/L; ferric
chloride,  0.5 mg/L; manganese sulfate, 10 mg/L; potassium phosphate (dibasic)
(cometabolism experiments) 100 mg/L.  Humidified air is sparged into the
bottom of the reactors.  Effluent gas passes through a condenser and
condensate is returned periodically  to the reactor.  Reactor pH is maintained
between 6.5 and 7.0 by automatic addition of 0.15 M sodium hydroxide by means
of a pH controller.

     The reactor feeding procedure involves withdrawing 600 mL of well-mixed
reactor contents and replacement with 600 mL of slurry stock, seed and
mineral medium.  Feed corresponds to the initial composition in the reactor,
consisting of 50% (300 mL) soil slurry, 33% (200 mL) of water and 17% (100
mL) of mineral salts medium.  The initial feeding frequency was daily;  this
was subsequently changed to every other and, eventually, every third day.
This protocol was used for semi-continuous experiments with no cosubstrate.
The feeding frequency for experiments in which phenol was used as a
cosubstrate with the soil contaminants, was every third day.

     Concentrations of the soil contaminants were determined by extracting
samples of the reactor contents (25 mL) with cyclohexane (5 mL), followed by
High Performance Liquid Chromatograph (HPLC) analysis.  Samples were
extracted once; replicate assays were carried out on each extract.  An
extraction of the stock soil slurry indicated removal of some contaminants to
levels below the HPLC detection limit, i.e., reductions of 75 to 90%.  Serial
extractions of the soil slurry demonstrated that the first extraction removed
at least 90% of the extractable contaminants.  Parallel extractions must be
carried out to determine the accuracy of this separation procedure.

     HPLC analysis confirmed the presence of numerous aromatic species and
higher molecular weight compounds.  Chromatographs display a virtual
continuum of peaks with retention times between 8.3 and 22 minutes.  Many
compounds were identified by matching retention times with a Standard.
Subsequent GC/MS analysis identified a number of aromatic and biaryl
species.  The names and structures of some of the most abundant species are
shown in Figure 1.  The two largest peaks in the solvent extract were found
to be l,l-bis(p-tolyl)-ethane and l,2-dimethyl-4-(phenylmethyl) benzene.
These identifications were assigned to the two large, corresponding peaks in
the HPLC chromatographs.  In most cases, reactor performance was referred to
HPLC retention times, as distinguished from specific compound identities.
                                     540

-------
        Naphthalene
     (CloHt,  MW-128)
                                             CH3—CH'
                 Benzene,  l,l'-ethylidenebis
                      (CUH14, MW~182)
         CH3  CH3
  1,1' -Biphenyl, 2,2' -Dimethvl-
       {C14HU,  MW-182)
CH-
CH,
                   Benzene,  1,2-dimethyl-
                     4-(phenylmethyl}-
                      (Cl$Hw, MW==196)
             CH3


             CH;J
          CH2-CH=CH2
Naphthalene,  1,2,3-trimethyl-
         4-propenyl
     r1-Di-
-------
                                RESULTS

VOLATILITY STUDIES

     A study was carried out to determine which, if any, volatile or semi-
volatile compounds were being stripped from the slurry.  HPLC analysis
indicated that benzene is completely removed in four days, while
concentrations of toluene remain relatively constant.  The large peak
corresponding to the xylene isowers and ethyl benzene masked the naphthalene
peak at the beginning of the study.  After one day, the peak corresponding to
xylene/ethyl benzene had essentially disappeared, exposing the naphthalene
peak.  Naphthalene and other medium molecular weight compounds decreased
between 30 and 50% during the four-day study.  Non-volatile, high molecular
weight compounds exhibited reductions of 10 to 20%, indicating a second
phenomena such as indigenous microbial activity was playing a role in
contaminant reduction.

SEMI-CONTINUOUS FERMENTATIONS - WITHOUT COSUBSTRATE

     Results are reported for Reactor 2.  They can be grouped into three
categories: a) complete or almost complete degradation [Figures 2 and 3], b)
no detectable degradation [often characterized by extremely irregular
chromatograph patterns - also, see Figure 4], and c) significant degradation
after extended lag [see Figure 5 with lag exceeding 20 days].  Naphthalene,
xylene, and ethyl benzene levels were below HPLC detection limits, after a
bioreaction interval as short as one day.  The magnitude of this reduction
Indicates that, after accounting for stripping, microbial reactions still
play a significant role in the reduction of these compounds, especially
naphthalene.  Other compounds that were virtually completely degraded, with
little if any evidence for stripping, were species with HPLC retention times
of 11.94 min, 13.26 min (l,2-Dimethyl-4-phenylmethyl-benzene), 13.86 min,
15.16 min [l,l-Bis-{p-to1yl)-ethane], and 15.87 min.  Compounds displaying
partial reduction had HPLC retention times of 10.61 and 14.42 min,
respectively.

     The compound with a retention time of 12.47 min displayed a period of
low reaction followed by recovery.  This may have been the result of a shift
in microbial population distribution; wash-out of one class of organisms
followed by rapid growth of another class may have occurred.  Compounds
having molecular weights of at least 210, i.e., those with retention times of
at least 16.00 min, exhibited scattering indicative of a total lack of
biodegradative activity.  Extending the biodegradation interval, i.e.,
reducing the feeding frequency, increased the extent of biodegradation of
heavier compounds - those with corresponding retention times of 14.47, 15.16
and 15.87 min.

     Extracts of reactor contents exhibited significant variability.
Inadequate mixing of the stock slurry prior to reactor feed preparation
probably resulted in erratic contaminant levels in reactor solutions.  It is
extremely important that the stock solution be mixed thoroughly in all future
                                     542

-------
             8000
                                                  Immediately After Feeding
                                                  After Biodegradatdon Interval
             6000
  Absolute
  Peak Area
 (Dim.  units)
             4000-
             2000-
                                    OO 'I EMarO-O  ['"O-Cn
                  0         10        20         30
                                       Time   (Days)
   Figure  2:  Relative  Contamination vs  Time for Reactor  2
               with Ho  Cosubstrate Added.    Results  for peak
               with HPLC Retention Time  of  13.26  minutes,
               tentatively  identified as  1,2-dimethyl-4-
               phenyjmethyl-benzene.
50
             4000
                                                 Immediately After Feeding
                                                 Mter Biodegradation  Interval
            3000
  Absolute
 Peak Area
(Dim.  units)
            2000-
             1000-
                 0         10         20
                                      Time   (Days)
    Figure  3:   Relative  Contamination  vs  Time  for  Reactor  2
                 with No  cosubstrate Added.    Results  for peak
                 with HPLC  Retention Time  of  13.86  minutes.
                               543

-------
            7000
            6000-
  Abaolute
 Paak Area
(Dim.  units)
5000
                                                 Immediately After Feeding
                                                 After Biodegradation Interval
                                     20        30
                                      Time   (Days)
                                               40
                                                       50
  Figure 4:  Relative  Concentration  vs Time  for Reactor  2
              with No  Cosubstrate Added.    Results  for peak
              with Retention Time of  14.42  minutes.
            10000
             8000
  Absolute
  P«ak  Area
(Dim.  units)
 6000-
 Figure  5:
             4000-
             2000-
                                                 Inmediately After Feeding
                                                 After Biodegradation Interval
                            10
                          20
                          Time
                                  30
                              (Days)
Relative Concentration  vs  Time  for Reactor  2
with   No  Cosubstrate Added.    Results  for peak
with  HPLC  Retention Time  of  15.16 minutes,
tentatively  identified  as  1,1-Bis(p-tolyl)ethane,
                               544

-------
experiments.  Variations in reactor volume due to fluctuating aeration and
humidification of influent air, affected the relative mass of reactor
contents removed prior to feeding.  In turn, this caused variations in
contaminant concentrations immediately after feeding and after the
biodegradation interval.  The variations described above resulted in the
following deviations from optimal reactor performance: 1) poor reactor
agitation caused soil settling and the high soil content in the bottoms of
the reactors (where samples are taken) led to unusually high concentrations
to be measured after the degradation interval 2) either a low soil content in
feed solution caused by poor mixing of the stock slurry solution or water
mixed in with cyclohexane extract and subsequent improper HPLC sampling and
analysis, resulted in unusually low compound concentration measurements in
samples taken immediately after feeding.

SEMI-CONTINUOUS FERMENTATIONS - WITH PHENOL AS COSUBSTRATE

     Fermentations with phenol present as cosubstrate were run in duplicate
and are referred to as Reactors 2 and 3. The Reactor 2 experiment with phenol
was carried out in the same reaction system as the Reactor 2 without
substrate. As for the experiment with no cosubstrate, low and medium
molecular weight compounds exhibited significant (often total) reductions -
retention times of 8.49, 10.61, 11.94, 12.47, 13.26 and 13.87 min.
     With phenol added, Reactor 2 displayed substantial degradation of the
compound with an HPLC Retention Time of 14.42 min, early in the experiment.
Reactor 3 may have had some initial activity, also.  In later stages of
reactor operation, neither system displayed significant activity.  These
observations may be attributable to i) a shift in microbial distribution or
11) depletion of phenol and consequent loss of a cometabolic contribution.
After an initial lag of 20 to 30 days, l,l-Bis(p-tolyl)ethane [BTA] appeared
to be degraded completely, with or without the addition of phenol.  However,
an intermediate period of little or no conversion was observed and the length
of this period of inactivity was significantly longer  (an order-of-magnitude)
with phenol present.  This effect was probably a consequence of microbial
acclimation or population displacement. High molecular weight compounds, with
HPLC retention times exceeding 15.87 min, exhibited virtually no reduction.

     Increasing feed concentrations at the end of the phenol experiment were
attributed  to the remaining slurry stock (bottom of the jar) having a much
higher  soil content than was used for most of the earlier feed solution
preparations.   It appears that the addition of phenol did not appreciably
improve the performance of the bioreactors for degradation of hydrocarbon
contaminants of high molecular weight.  It did appear to have some influence
on the  degradation of species of mid-range molecular weight, i.e., BTA.

                           CONCLUSIONS

     Polynuclear  aromatic, polyaryl and chlorinated hydrocarbons are
associated  with the  finer fractions of contaminated soil, i.e., natural
organic matter  and clays.  Slurry preparation, with suspension of fines and
settling of larger particles, leads to contaminant enrichment of the fines.


                                     545

-------
     A separate tar-like fraction appears during suspension and settling.

     Aliquots taken from large volumes (master batches) of soil fines
suspensions are highly variable.  Samples drawn from slurry bioreactors, for
assays of relative contaminant concentrations, dispersed microbial
populations and nutrient concentrations, are variable, also.

     Microbial degradation rates are strongly correlated with contaminant
molecular weight.  Over periods of days, slurry reactions yield three
distinct results: complete degradation, complete degradation after several
days lag, and no detectable reduction in contaminant concentration.

     Species with molecular weights exceeding 200, based on HPLC detection
limits and retention times, resist microbial reaction.

     Phenol appears to have little effect as a cosubstrate.

                          REFERENCES

1.  Ahlert, R.C., D.S. Kosson and W.V. Black, Preliminary Results on the
Aerobic/Anaerobic Biochemical Reactor for the Mineralization of Organic
Contaminants Bound on Soil Fines, 15th Annual EPA Research Symposium,
Cincinnati (OH), April 1989.


2.  Bjorseth, A., Handbook of Polycyclic Aromatic Hydrocarbons, Marcel
Dekker, Inc., New York (NY), 1983.

3.  Cerniglia, C.E., Microbial Transformation of Aromatic Hydrocarbons, in
Petroleum Microbiology. R.M. Atlas (Ed.), Macmillan, New York (NY), 1984.

4.  Enzminger, J.D., R.C. Ahlert, J.V. Lepore, C. Gleason and C. Dreyer,
Development of Mixed Aerobic Cultures for Degrading Coal Tar, AIChE Summer
National Meeting, Boston (MA), August 1986.

5.  Mackay, D. and W.Y. Shiu, Aqueous Solubility of Polynuclear Aromatic
Hydrocarbons, J. Chem. and Engineering Data, Vol. 22(4), Pps. 399-402, 1977.

6.  Dzombak, D.A. and R.G, Luthy, Estimating Adsorption of Polycyclic
Aromatic Hydrocarbons on Soils, Soil Science, Vol. 137(5) Pps. 292-308, 1984.

7.  Lambert, S.M., R.G. Porter and H. Schieferstein, Movement and Sorption of
Chemicals Applied to the Soil, Heeds, Vol. 13, Pps. 185-190, 1965.

8.  Bauer, J.E. and D.G. Capone, Degradation and Mineralization of the
Polycyclic Aromatic Hydrocarbons Anthracene and Naphthalene in Intertidal
Marine Sediments, Appl. and Env. Micro., Vol. 50(1), Pps. 81-90, 1985.

9.  Herbes, S.E. and L.R. Schwall, Transformation of Polycyclic Aromatic
Hydrocarbons in Pristine and Petroleum-Contaminated Sediments, Applied and
Environmental Microbiology. Vol. 35(2), Pps. 306-316, 1978.


                                     546

-------
10. Bauer, J.E. and D.G.Caponi, Effects of Co-Occurring Aromatic Hydrocarbons
on Degradation of Individual Polycyclic Aromatic Hydrocarbons in Marine
Sediment Slurries, ADD!, and Env. Micro.. Vol. 54(7), PPs. 1649-1655, 1988.

11. Schocken, M.J. and D.T. Gibson, Bacterial Oxidation of the Polycyclic
Aromatic Hydrocarbons Acenaphthene and Acenaphthylene, Applied and
Environmental Microbiology, Vol. 48(1), Pps. 10-16, 1984.

12. Mihelcic, J.R. and R.G. Luthy, Degradation of Polycyclic Aromatic
Hydrocarbon Compounds under Various Redox Conditions in Soil-Water Systems,
Applied and Environmental Microbiology, Vol. 54(5) Pps. 1181-1187, 1988.

13. Shiaris, M.P., Seasonal Transformation of Naphthalene, Phenanthrene, and
Benzo(a)pyrene in Surficial Estuarine Sediments, Applied and Environmental
Microbiology. Vol. 55(6), Pps. 1391-1399, 1989.

14. Heitkamp, M.A. et al,  Pyrene Degradation by a Hvcobacterium sp. ;
Identification of Ring Oxidation and Ring Fission Products, Applied and
Environmental Microbiology, Vol. 54(10) Pps. 2556-2565, 1988.

15. Heitkamp M.A. and C.E. Cerniglia, Polycyclic Aromatic Hydrocarbon
Degradation by a Mycobacter.1 um sp. in Microcosms Containing Sediment and
Water from a Pristine Ecosystem, Applied and Environmental Microbiology. Vol.
55(8), PPs. 1968-1973, 1989.
                                     54?

-------
                        ANAEROBIC TREATMENT OF LEACHATE

                            Sanjoy K. Bhattacharya
                Civil  Engineering Department,  Tulane University
                            New Orleans, LA  70118

                               Richard A. Dobbs
                     RREL,  U.S.  EPA,  Cincinnati,  OH  45268

                                Rao V.R.  Angara
                      NSI Technology .Services  Corporation
                            Kansas City, KS  66115


                                   ABSTRACT


      An upflow anaerobic filter  (volume: 800 gal. and empty bed contact time;
3.7 days) followed by a conventional activated sludge system (HRT: 5.6 hr. and
SRT: 8 days) was used to determine the fate and effect of hazardous leachate
in anaerobic-aerobic systems.  The leachate flowrate to the anaerobic filter
was 0.15 gpm.  Screened raw municipal wastewater (0.85 gpm) and the filter
effluent were added to the conventional activated sludge system.  Selected
organic compounds and heavy metals were spiked to the synthetic leachate (C-
source: acetate, propionate, and  butyrate, total feed COD: 1600 mg/L) to
simulate a typical hazardous leachate.

      The results showed that selected heavy metals at design concentrations
did not cause any adverse effects on the anaerobic process.  Among the organic
toxicants the removals of chlorobenzene and toluene were low (15-25%) whereas
the removals of acetone, methyl ethyl ketone, methyl isobutyl ketone,
trichloroethylene, methylene chloride, and 1,1-dichloroethane exceeded 90
percent.  The removal of all added organic compounds by the combined
anaerobic-aerobic process exceeded 90 percent.  Hence, an upflow anaerobic
filter preceding a conventional activated sludge process appears to be
effective in treating hazardous leachates.
                                    S48

-------
                                 INTRODUCTION


     The objective of this research was to obtain information on the treat-
ability of hazardous leachates in publicly owned treatment works (POTWs).  The
hypothesis was that a combined anaerobic-aerobic biological treatment process
should be able to biodegrade most of the organic compounds present in
hazardous leachates.  It is well known from literature that some organics such
as chlorinated hydrocarbons can be biodegraded efficiently by anaerobic
processes.  Anaerobic reductive halogenation has been recognized as a possible
biotransformation mechanism for chlorinated aliphatic hydrocarbons (1-4).

      Aromatic volatile organic compounds, whether homocyclic or heterocyclic,
present a special problem to anaerobes because of their ring structure.  Evans
(5) concluded that the benzene nucleus is first reduced and then cleaved by
hydrolysis to yield aliphatic acids which in turn could enter common anaerobic
pathways leading to the final products of carbon dioxide and methane.  Through
the use of  0-labeled water,  it  has  been  shown  that  oxygen  from water  is
incorporated into anaerobic fermentative transformation of toluene and benzene
(6).  The same researchers proposed degradation sequences for the
transformation of toluene and benzene by mixed methanogenic cultures (7).  In
contrast, Schink (8) reported no indication of methanogenic degradation for a
number of compounds, including toluene and benzene.

      A summary of recent work on anaerobic processes for transformation of   -
volatile compounds is presented in Table 1.


        TABLE  1.  ANAEROBIC  TRANSFORMATION  OF  VOLATILE ORGANIC  COMPOUNDS


Compound                          Product (s).                    Reference

chloroform                        carbon dioxide                1
carbon tetrachloride              carbon dioxide                1
                                  chloroform                    8, 9
1,1-dichloroethane                chloroethane                  10
1,2-dichloroethane                *                             11
1,1,1-trichloroethane             1,1-dichloroethane            9
                                  *                             1, 12,  13
1,1,2,2-tetrachloroethane         1,1,2-trichloroethane         1
chloroethene                      carbon dioxide                3
dichloroethene                    chloroethene                  3, 12,  13
trichloroethene                   dichloroethene                3, 9, 13, 14
tetrachloroethene                 trichloroethene               13
benzene                           carbon dioxide                6, 7, 15
                                  methane
toluene                           carbon dioxide                6, 7, 15
	methane	

* Products not identified

                                     S4S

-------
      Recent process and serum bottle studies with semi volatile and/or
nonvolatile organic compounds are summarized 1n Table 2.
             TABLE 2.  RECENT INVESTIGATIONS OF ANAEROBIC PROCESSES
Invest 1 gator (s)
Bl urn et al .
(1986)
Study
anaerobic filters
with coal conversion
wastewater constituents
Pollutants
phenol ics
Reference
(16)
Kuhn et al.
   (1985)
Lema et al.
   (1987)

Suidan et al
   (1987)
Boyd et al.
   (1983)

Hickey et al
   (1987)
Johnson & Young
   (1983)
She!ton et al
   (1984)
Vinkataramani
et al.
   (1986)
laboratory aquifer
columns to simulate
groundwater infiltration

bench-scale digesters
with landfill leachate

anaerobic filters
with coal gasification
wastewater (synthetic)

serum bottle assays to
determine fate

serum bottle assays to
monitor CH,,  H2  from
pulse addition
dichlorobenzenes
dimethyl benzenes
 unspecified
 organics

 aromatics
 phenolics
serum bottle assay in
synthetic sludge for
inhibition and recovery

serum bottle assay to
determi ne degradabi11ty
and pathway
 chloroform
formaldehyde
 bromoethanesulfonic  acid
 trichloroacetic  acid

 semi-volatile
 priority
 pollutants

 phthalates
serum bottle assay of    not specified
high-strength industrial
landfill leachate
(17)



(18)


(19)



(20)


(21)




(22)



(23)



(24)
      The fate of both volatile and semivolatile organics in anaerobic
digestion of primary and secondary sludge containing sorbed compounds has been
described (25).

      Several studies have been conducted on the removal and fate of specific
organics by conventional primary/activated treatment (26-32).  A combined
                                    550

-------
anaerobic-aerobic process should have the advantages of both processes in
terms of capability to biodegrade specific compounds.  The goal of this
research project was to verify this hypothesis for biological treatment of
hazardous leachate.
                             MATERIALS  AND  METHODS


      Selection of leachate constituents to be used for experimental fate
studies was based on survey results of actual hazardous waste sites.  The
organic compounds and heavy metals which were representative of a "typical"
hazardous waste leachate are shown in Table 3 along with the recommended
concentrations to be tested.
                        TABLE 3. LEACHATE CONSTITUENTS
            CONSTITUENT                          CONCENTRATION.  ug/L

            Acetone                                   10,000
            Methyl ethyl ketone                        5,000
            Methyl isobutyl ketone                     1,000
            Chloroform                                10,000
            Trichloroethylene                            400
            Methylene chloride                         1,200
            1,1-Dichloroethane                           100
            Chlorobenzene                              1,100
            Toluene                                    8,000
            Ethyl benzene                                 600
            Dibutyl phthalate                            200
            Bis (2-ethylhexyl) phthalate                 100
            Nitrobenzene                                 500
            Trichlorobenzene                             200
            Phenol                                     2,600
            Aluminum                                  30,000
            Copper                                       100
            Iron                                     500,000
            Lead                                         400
            Magnesium                                180,000
            Manganese                                 65,000
            Nickel                                     1,500
            Zinc                                       1,000
      The leachate constituents were spiked to a synthetic feed to simulate
real life hazardous leachate.  The synthetic feed prepared in tap water had a
design Biological Oxygen Demand (BOD) or Chemical Oxygen Demand (COD) of
approximately 2,000 mg/L derived from added acetic and propioriic acids.


                                     551

-------
      The anaerobic system consisted  of  an 800-gallon  upflow  anaerobic  filter
packed with plastic media and  had  a flowrate of  about  216 gpd  (0.15 gpm).   The
empty bed contact time was 3.7 days.  The effluent of  the filter was mixed
with about 0.85 gpm of raw wastewater.   The total flow of 1 gpm was added to
the primary clarifier of the activated sludge  system.  The solids retention
time in the aeration tank was  about 8 days and the hydraulic  retention  time
was 5.6 hours.  A schematic of the anaerobic-aerobic process  is shown in
Figure 1.
           Synthetic feed
Gas    Raw wastewater 0.85 gpm
                                                                   1.0 gpm
            Figure 1.  Schematic diagram of anaerobic-aerobic process.
      Following an acclimation period of three months, the system was spiked
for 12 weeks.  Composite aqueous samples were collected during 24-hr periods
(such periods are referred to as EVENTS), once every week for six weeks.
Samples were analyzed for volatile organic compounds using Method 1624
Revision B (33).  Method 1625 Revision B (34) was used to determine
semi vol at lies.  Metals were measured by flame atomic absorption
spectrophotometry (35) .

      While the anaerobic-aerobic system was being built, an anaerobic
toxicity assay (ATA) was performed using serum bottles (50 ml culture) to
assess potential adverse effects of the selected compounds on the anaerobic
process.  Acetate enrichment culture was used as the seed organisms.
                                    552

-------
                             RESULTS AND DISCUSSION
      The  results of the ATA showed that  the heavy metals did not cause  any
toxicity  to the batch  anaerobic acetate enrichment culture in the serum
bottles.   The organic compounds, however,  showed toxicity even at 50% of  the
design  concentrations listed in Table  1,  Chloroform appeared to be the most
likely  cause for the toxicity based on reported inhibition to anaerobic
processes (36,  37).  The design concentrations  of the other organic compounds
were  not  expected to cause toxicity.   When the  ATA was performed with the
individual  compounds, only the serum bottles with 5 mg/L or greater
concentration of chloroform showed toxicity. Although more work is required
to prove  that chloroform (5 mg/L or higher) caused toxicity to the anaerobic
filter, it was  decided to spike the feed to the pilot system with all
compounds (in Table 1) except chloroform.

      The measured concentrations of volatile leachate compounds are shown  in
Table 4 for the initial sampling event.


        TABLE 4. MEASURED CONCENTRATIONS  OF LEACHATE           (EVENT 1)
                             CONCENTRATION OF TOXICS
LEACHATE COMPOUNDS
VOC
Acetone
Methyl ethyl ketone
Methyl Isobutyl ketone
Tricloroethylene
Methylene chloride
1 , 1-Dichloroethane
Chlorobenzene
Toluene
Ethyl benzene
LEACH
3625
1388
446
144
573
36.8
425
843
227
LEACH
(OUP)
2906
1166
422
131
512
33.1
387
719
214
FIL
EFF
78.
66.
48.
3.
105
7.
337
714
116

7
6
9
2

6



FIL
EFF
(OUP)
94.1
76.0
52.8
3.8
125
8.2
356
774
122
RAW
WW
662
35.8
11.2
1.1
8.3
1.7
8.9
25.9
73.8
PRI*
INF
575
40.4
17,0
1.4
22.8
2.6
58.2
129
80.1
PRI
EFF
711
104
16.1
1.0
16.3
1.8
37.8
172
108
SEC
EFF
50.7
7.7
4.2
0.9
4.8
0.8
1.6
2.0
2.8
PRI
SLU
405
33.0
14.6
1.2
17.8
2.8
56.2
273
103
SEC
UAS-
0.0
13.7
3.4
0.0
9.9
2.4
4.9
5.0
0.6
LEACH-LEACHATE    FIL EFF-FILTER EFFLUENT    RAW W-RAW VASTEWATER    PRI EFF-PRIMAKY EFFLUENT
SEC EFF-SECONDARY EFFLUENT    PRI SLU-PRIMARY SLUDGE    SEC WAS-SECONDARY HASTE ACTIVATED SLUDGE
       Measured concentrations for duplicate samples were generally  in good
agreement considering the complex matrices involved.  Based on measured  flow
rates,  the masses of volatiles in the  Influent,  effluent, and sludges were
calculated.  The results of the calculations are shown in Table 5.
                                     513

-------
               TABLE 5.  CALCULATED MASS OF LEACHATE COMPOUNDS IN
                       SLUDGE AND  LIQUID  PHASES  (EVENT 1)
                                 MASS OF TOXICS (gm)
LEACHATE COMPOUNDS
VOC
Acetone
Methyl ethyl ketone
Hithyl Isobutyl kotono
Triehloroathylena
Hathyl ena chloride
1,1-Dlchloroethane
Chlorobenzene
Toluene
Ethylbenzena
LEACH
2
1
0
0
0
0
0
0
0
,96
,13
.36S
,118
.469
,030
.348
.689
.186
LEACH
(DUP)
2.38
0.953
0.345
0.107
0.419
0.027
0.316
0.587
0.175
FIL
EFF
0.064
0.054
0.041
0.003
0.086
0.006
0.276
0.584
0.095
FIL
EFF
(DUP)
0.077
0.062
0.043
0.003
0,102
0.007
0.291
0.633
0.099
RAW
WW
3.07
0.16,6
O.OS2
0.005
0.038
0.008
0.041
0.120
0.342
PR!*
INF
3.13
0.220
0.092
0.008
0.124
0.014
0.317
0.704
0.436
PRI
EFF
3.88
0.566
0.087
0.005
0.089
0.010
0.206
0,938
0.591
SEC
EFF
0,276
0.042
0,023
0.005
0.026
0.004
0.009
0.011
0.016
PRI
SLU
0.087
0.007
0,003
0.000
0.004
0.001
0.012
0,059
0.022
SEC
WAS
0.000
0.002
0.000
0,000
0.001
0,000
0.001
0.001
0.000
      Based on the mass  data shown In Table 4, the percent removals were
calculated for Event  1.   The percent removals obtained for the first event are
summarized in Table 6.
           TABLE 6.  PERCENT REMOVAL OF LEACHATE COMPOUNDS (EVENT 1)
                         ANAEROBIC
AEROBIC
LEACHATE COMPOUNDS

VOC


Acetone
Methyl ethyl ketone
Methyl Isobutyl ketona
Trichloroethylane
Methylene chloride
1 , 1-Dlchloroethane
Chlorobenzene
Toluene
Ethylbenzene
%
FILTER
REMOVAL


97.4
94.4
88.2
97.4
78.8
77.4
14.7
4.7
46.1
X
PRIMARY
REMOVAL


-23,7
-157.1
5.4
29.7
28.7
30.6
35.0
-33.3
-35.4
X
TOTAL
REMOVAL


91.2
80.9
75.5
36.7
79.0
71.1
97.2
98.5
96.4
X
ADSORBED



2.8
4.0
3.8
3.3
4.0
6.4
4.0
8.4
5,1
X
VOLATILIZED,
BIODEGRADED,
OR CHEMICALLY
TRANSFORMED
88.4
77.0
71.7
33.4
74,9
64.6
93.2
90.0
91.3
      Comparable percent  removals were calculated for all sampling events for
the volatile organics  and results are tabulated in Table 7.
                                     554

-------
          TABLE  7.  REMOVAL OF  VOLATILE OR6ANICS  BY  ANAEROBIC  FILTER


LEACHATE COMPOUNDS

Acetone
Methyl ethyl ketone
Methyl isobutyl ketone
Tri chl oroethyl ene
Methyl ene chloride
1,1-Dichloroethane
Chlorobenzene
Toluene
Ethyl benzene

DESIGN
FEED CONC
M9/L
10,000
5,000
1,000
400
1,200
100
1,100
8,000
600
AVERAGE
MEASURED
FEED CONC
W/L
4,289
1,269
355
162
637
54
408
755
246
AVERAGE
MEASURED
FIL EFF
M9/L
78
69
45
2
58
11
338
789
116
AVERAGE
PERCENT
FILTER
REMOVAL
98
95
87
99
91
80
17
-4
55
      Except for the aromatic solvents (chlorobenzene, toluene, and
ethyl benzene) the removals of VOCs in the anaerobic filter were 80% or higher.
Trichloroethylene and acetone showed the highest removals, viz., 99% and 98%,
respectively.  The design feed concentration shown in Table 7 was not achieved
in the feed to the anaerobic filter.  This was most likely due to losses of
the compounds in the recirculation tank since duplicate samples showed good
agreement throughout the study.

      Average percent removal of semivolatile compounds by the anaerobic
filter were calculated and are presented in Table 8.


         TABLE 8.  REMOVAL  OF  SEMIVOLATILE ORGANICS  BY  ANAEROBIC FILTER


LEACHATE COMPOUNDS

Di butyl phthalate
Bis (2-ethylhexyl)
Nitrobenzene
Tri chlorobenzene
Phenol

DESIGN
FEED CONC
m/i
200
phthalate 100
500
200
2,600
AVERAGE
MEASURED
FEED CONC
M9/L
38
108
45
161
1,678
AVERAGE
MEASURED
FIL EFF
0gA
BDL
36
BDL
45
1,105
AVERAGE
PERCENT
FILTER
REMOVAL
...
67
—
72
34
      The effluent concentration of dibutyl phthalate and nitrobenzene were
below detection limit (BDL).  Dibutyl phthalate and bis (2-ethylhexyl)
                                    S55

-------
phthalate were removed by the anaerobic filter.  Phenol was not effectively
removed by the filter.

      Removal of metals was generally high.  Table 9 shows the removal
achieved by anaerobic treatment.
                TABLE 9. REMOVAL OF METALS BY ANAEROBIC FILTER
LEACHATE COMPOUNDS
 DESIGN
FEED CONC
  W/L
 AVERAGE
MEASURED
FEED CONC
  Itg/l
AVERAGE
MEASURED
FIL  EFF
AVERAGE
PERCENT
 FILTER
REMOVAL
METALS
Aluminum
Copper
Iron
Lead
Magnesium
Manganese
Nickel
Zinc
30,000
100
500,000
400
180,000
65,000
1,500
1,000
3,449
257
253,200
418
40,500
20,906
1,647
597
251
BDL
7,941
20
32,580
1,860
896
BDL
93
—
97
95
20
91
46
_ _ _
BDL: BELOW DETECTION LIMIT CONCENTRATION
---: NOT CALCULATED DUE TO BDL CONCENTRATIONS
      Metal removal in the anaerobic filter is most likely due to chemical
precipitation and sorption on the biomass.  Since magnesium has high
solubility, removal was only 20%.

      Removal of volatile organics by the combined anaerobic-aerobic treatment
system is presented in Table 10.
                                    556

-------
              TABLE  10.  REMOVAL OF VOLATILE  ORGANICS  BY  COMBINED
                           ANAEROBIC-AEROBIC SYSTEM
VOLATILE ORGANICS            AVERAGE  REMOVAL       STANDARD DEVIATION
Acetone
Methyl ethyl ketone
Methyl Isobutyl ketone
Trichloroethylene
Methyl ene chloride
1,1-Dlchloroethane
Chlorobenzene
Toluene
Ethyl benzene
94.9
89.1
58.1
72.9
66.5
73.6
93.2
85.4
92.8
3.4
5.6
19.1
26.6
13.5
17.5
7.6
22.3
3.5
      Total removal was based on the influent to the anaerobic filter and the
effluent from the secondary clarifier of the activated sludge system.
Chlorobenzene, toluene, and ethyl benzene, which were not efficiently removed
by the anaerobic filter, showed 93, 85, and 93 percent removal by the combined
anaerobic filter-activated sludge treatment system.  Thus, the advantage of
the combined process has been demonstrated.  The concentrations of
semivolatiles and metals in the secondary effluent were too low to calculate
the removals for the combined system.
                                  CONCLUSIONS


      The combined anaerobic-filter activated sludge treatment system offers
advantages for improved treatment of hazardous leachate constituents over
either process alone.  An anaerobic process ahead of a conventional
primary/activated sludge system offers a control technology that can
potentially avoid or reduce air emissions due to chlorinated aliphatic
solvents and other volatiles which can be anaerobically degraded.  Heavy
metals at high concentrations did not cause toxicity to the anaerobic filter.
Chloroform at S.O mg/L was inhibitory to the anaerobic filter.  Additional
research is needed to define the acceptable level of chloroform that can be
tolerated.  High removals of most organics was achieved by the anaerobic
filter.  Total removals by the combined process were also high.  The combined
system appears to be a feasible process for treatment of hazardous leachates.
                                    557

-------
                                  REFERENCES
 1.   Bouwer, E.J., and McCarty, P.L.  Appl. Environ. Microblol.  45: 1286,
      1983.
 2.   Parsons, F., Wood, P.R., and DeMarco, J.  J. Amer. Water Works Assoc.
      76(5): 56, 1984.
 3.   Vogel, T.M., and McCarty, P.L.  Appl. Environ. Microbiol.  49: 1080,
      1985.
 4.   Barrio-Lage, G., Parson, F.Z., Nassar, R.S., and Lorenzo, P.A.  Environ.
      Sci. Techno!.  20: 96, 1986.
 5.   Evans, W.C.  Nature.  270: 17, 1977.
 6.   Vogel, T.M., and Grbic-Galic, D.  Appl. Environ. Microbiol.  52: 200,
      1986.
 7.   Grbic-Galic, D., and Vogel, T.M.  Appl. Environ. Microbiol.  53: 254,
      1987.
 8.   Schink, B.  FEMS Microbiol. Ecol.  31: 69, 1985.
 9.   Parsons, F., Barrio-Lage, G., and Rice, R.  Environ. Toxicol. Chem.  4:
      739, 1985.
10.   Strand, S.E., and Shippert, L.  Appl. Environ. Microbiol.  52: 203,
      1986.
11.   Stucki, G., Krebser, V., and Leischger, T.  Experientia.  39: 1271,
      1983.
12.   Vogel, T.H., and McCarty, P.L.  Environ. Sci. Techno!.  21: 1208, 1987.
13.   Hallen, R.T., Pyne, J.W., Jr., and Molton, P.M.  Extended Abstract
      Annual Meeting of the American Chemical Society.  1986.
14.   Kleopfer, R.D., Easley, D.M., Haas, B.B., Jr., and Deihl, T.G.  Environ.
      Sci. Technol.  19: 277, 1985.
15.   Schink, B.  FEMS Microbiol. Ecol.  31: 69, 1985.
16.   Blum, O.W.J., Hergenroeder, R., Parkin, G.F., and Speece, R.E.  J. WPCF.
      58(2): 122, 1986.
17.   Kuhn, E.P., Colberg, P.J., Schnoor, J.L., Wanner, 0., Zehnder, A.J.B.,
      and Schwarzenbach, R.P.  Environ. Sci. Technol.  19(10): 961, 1985.
18.   Lema, J.M., Ibanez, E., and Canals, J.  Environ. Technol. Letters.  8:
      555, 1987.
                                     558

-------
19.   Suidan, M.T., Fox, P., and Pfeffer, J.T.  Water Sci. Tech.  19: 229,
      1987.

20.   Boyd, S.A., Shelton, D.R., Berry, D., and Tledje, J.M.  Appl. Environ.
      Microblol.  46(1): 50, 1983.

21.   Hickey, R.F., Vanderwielen, J., and Switzenbaum, H.S.  Wat. Res.
      21(11): 1417, 1987.

22.   Johnson, L.D., and Young, J.C.  J. WPCF.  55(12): 1441, 1983.

23.   Shelton, D.R., Boyd, S.A., and Tledje, J.M.  Environ. Sci. Techno!.
      18(2): 93, 1984.

24.   Venkataramani, E.S., Ahlert, R.C., and Corbo, P.  Aerobic and Anaerobic
      Treatment of High-Strength Industrial Waste Residues.  Proceedings of
      41st Purdue Univ. Industrial Waste Conference.  249, 198S.

25.   Dobbs, R.A., Govind, R., Flaherty, P.A., Crawford, T.L. Siddiqui, K.,
      and Austern, B.M.  Fate and Effects of RCRA and CERCLA Toxics In
      Anaerobic Digestion of Primary and Secondary Sludge.  Presented at the
      15th Annual EPA Research Symposium, Cincinnati, Ohio.  April 10-12,
      1989.

26.   Fate of Priority Pollutants in Publicly Owned Treatment Works.  Volume
      2, EPA 440/1-82-303, U.S. Environmental Protection Agency, July, 1982.

27.   Fate of Priority Pollutants in Publicly Owned Treatment Works.  Volume
      1, EPA 440/1-82-303, U.S. Environmental Protection Agency, September,
      1982.

28.   Hannah, S.A., and Rossman, L.  Monitoring and Analysis of Hazardous
      Organics in Municipal Wastewater - A Study of Twenty-five Treatment
      Plants.  Presented at the Seminar on Hazardous Substances in Wastewater.
      Toronto, Ontario, November 3, 1982.

29.   Petrasek, A.C., Kugelman, I.J., Austern, B.M., Pressley, T.A., Winslow,
      L.A., and Wise, R.H.  J. WPCF.  55(10): 1286, 1983.

30.   Hannah, S.A., Austern, B.M., Eralp, A.E., and Wise, R.H.  J. WPCF.
      58(1): 27, 1986.

31.   Hannah, S.A., Austern, B.M., Eralp, A.E., and Dobbs, R.A,.  J. WPCF.
      60(7): 1281, 1988.

32.   Bhattacharya, S.K., Angara, R.V.R., Hannah, S.A., Bishop,, D.F., Jr.,
      Dobbs, R.A., and Austern, B.M.  Fate and Effects of Selected RCRA and
      CERCLA Compounds In Activated Sludge Systems.  Presented at the 15th
      Annual EPA Research Symposium, Cincinnati, Ohio.  April 10-12, 1989.

33.   Method 1624 Revision B.  Volatile Organic Compounds by Isotope Dilution
      SC/MS.  Federal Register, 49, No. 209: Friday, October 26, 1984, p. 175.

                                    559

-------
34.   Method 1625 Revision B.  Semivolatile Organic Compounds by  Isotope
      Dilution GC/HS.  Federal Register, 49, No. 209: Friday, October 26,
      1984, p. 184.

35.   Methods for Chemical Analysis of Water and Wastes, EPA 625/6-74-003.
      Methods Development and Quality Assurance Research Laboratory.  National
      Environmental Research Center, Cincinnati, Ohio.

36.   Yang, J.  The Effects of Cyanide and Chloroform Toxicity on Methane
      Fermentation.  Ph.D. dissertation, Drexel University, Philadelphia, PA.
      (1981).

37.   Hovlous, J.C., Wagg, 6.T., and Conway, R.A.  Identification and Control
      of Petrochemical Pollutants Inhibitory to Anaerobic Processes.  EPA R-2-
      73-194, U.S. Environmental Protection Agency.
                                     560

-------
     THE EFFECTS OF TEMPERATURE, CONCENTRATION OF SUBSTRATE AND MICROBIAL
INOCULUM AND THE SOURCE OF SLUDGE BIOMASS ON THE KINETICS OF BIODEGRADATION


              Henry H. Tabak1, Sanjay Desai2, and Rakesh Govind2


                     ^.S.  Environmental  Protection Agency
                    Risk  Reduction  Engineering  Laboratory
                            Cincinnati,  Ohio   45268

                Department of Chemical  and Nuclear Engineering
                           University of Cincinnati
                            Cincinnati,  Ohio   41221


                                   ABSTRACT

     A developed multi-level respirometric protocol  is presented for the
determination of the biodegradability, microbial acclimation to toxic
substrates and first order kinetic parameters of biodegradation and for
estimation and quantitation of Monod kinetic parameters (um,  K and Y)  of the
toxic organics, in order to correlate the extent and rate of their
biodegradation with a predictive model based on chemical properties and
structure of those organic pollutant compounds.  Respirometric biodegradation
kinetic data are provided  for representative RCRA alkyl benzenes, phenolic
compounds, phthalate esters and ketones and CERCLA leachate organics.

     Studies determined the effects of the source of sludge biomass,
temperature and concentrations of microbial inoculum and of toxic substrate on
the generated biodegradation kinetic parameter data in order to make an
assessment of their constancy under a variety of environmental conditions.
Sludge biomass from totally domestic, domestic/industrial  and fully industrial
treatment systems was used in the respirometric study to evaluate the effect
of the source of microbiota on the kinetics of biodegradation.  Temperatures
of 15°C, 25°C, and 35"C were used as culture incubation temperatures and the
substrate concentration levels (20, 40, 60, 80 and 100 m/L) were used to
determine the effects of temperature and substrate concentration on the
kinetic parameter values.  Data on the effects of the environmental conditions
of the biokinetic parameters for representative toxic organic pollutant
compounds are discussed.

     A predictive biodegradation structure-activity model  based on the group
contribution approach was  developed with the use of respirometrically
generated biodegradation kinetic data.  The predictive biodegradation fate
model will closely predict the results found experimentally.   Fate of
structurally related toxic organics may be anticipated without the time and
expense of experimental studies.

                                     561

-------
                                 INTRODUCTION

     It has been estimated that 50,000 organic chemicals are commercially
produced in the United States and a large number of new organic chemicals are
added to production each year (Blackburn and Troxler, 1984). The presence of
many of these chemicals in the natural ecosystem is a serious public health
problem. The 1986 amendments of Comprehensive Environmental Response,
Compensation and Liability Act (CERCLA), known as the Superfund Amendments and
Reauthorization Act (SARA) mentioned that cost effective treatments and
recycling must be considered as an alternative to the land disposal of wastes.
Among the alternative methods available, bioremediation and incineration are
the two technologies which destroy the wastes but the cost of bioremediation
is estimated to be only about one third the cost of incineration.

     Biodegradability and toxicity of organic chemicals determine their
behavior in the natural ecosystems and during the treatment of wastewater. The
high diversity of species and metabolic efficiency of microorganisms suggests
that they play a major role in the ultimate degradation of these chemicals.
Microorganisms use these carbon and energy sources for growth, converting them
into carbon dioxide, water, ammoniacal nitrogen and new cell material and thus
maintain the biological equilibrium in nature. Hence, there is need for data
on the biodegradability of organic chemicals to assess their fate in natural
ecosystems and in engineered environments. Information regarding the extent
and the rate of biodegradation of organic chemicals is very important in
evaluating the relative persistence of the chemical in the environment, which
in turn is important for regulating its manufacture and use.  Thus, there is a
need to develop correlations and predictive techniques to assess
biodegradability of toxic organics.  Lack of an adequate database on
biodegradation kinetics has prevented the development of such techniques.

     The main objective of this study was to develop a protocol based on
oxygen consumption to generate data base on biokinetic constants of
representative priority pollutants, RCRA and CERCLA toxic organics.
Experiments were conducted using electrolytic respirometry to generate oxygen
consumption data from which biokinetic parameters were developed.  The
representative compounds included 20 benzenes, 7 phthalates, 5 polycyclic
aromatic hydrocarbons, 4 nitrosoamines, 8 phenols, 4 ketones, p-nitroaniline
and benzyl alcohol with aniline as the reference compound.

     The environmental factors investigated were temperature, concentration of
biomass and substrate and the source of biomass.  These effects on
biodegradation kinetic rates will subsequently be related to the structural
properties of the compound.  This in turn will facilitate prediction of the
extent and rate of biodegradation of organic pollutant chemicals from the
knowledge of their structural properties.  The specific organic compounds,
used for the study of the effects of environmental factors on kinetic
parameters were:  benzene, toluene, ethylbenzene, phenol, resorcinol, 2,4-
dimethyl phenol, dimethyl phthalate, diethyl phthalate, butyl phthalate and
butyl benzyl phthalate.
                                     162

-------
                                  BACKGROUND

     An extensive review on the use of oxygen consumption measurements in
routine examination of sewage, in control of sewage treatment processes and in
the application to specific wastes or compounds has been presented by Jenkins
(1960), Montgomery (1967) and King and Dutka (1986).  Techniques used to
evaluate biodegradation kinetics were reviewed in detail by Howard, et a7.
(1975, 1981) and Grady (1985).  These techniques utilize continuous, fed-batch
and batch type reactors for providing data from which kinetic parameters can
be evaluated.

     The success of batch techniques in obtaining intrinsic kinetic parameters
(dependent only on the nature of the compound and the degrading microbial
community and not on the reactor system used) was indicated in the studies of
Simkins and Alexander (1984, 1985), Robinson and Tiedje (1983), Cech et a7.
(1984), Braha and Hafner (1987), Tabak, et a7. (1984), Larson and Perry (1981)
and Paris and Rogers (1986).

     Measurement of oxygen consumption through electrolytic respirometry is a
batch type technique which has been shown to be very promising for automating
data collection associated with biodegradation and intrinsic kinetic
parameters.  Studies performed by Dojlido (7), Larson and Perry (22), Tabak et
a7. (34), Oshima et a7. (27), Gaudy et a7. (9, 10) and Grady et a7. (13, 14,
15) have utilized the electrolytic respirometry approach to generate
biodegradation kinetic data.

     Dojlido (7) divided the oxygen uptake curve into seven different phases
and then proposed an empirical model for each phase and evaluated the
biodegradability and toxicity of a test compound by measuring empirical rate
constants and time interval associated with each phase.  Various phases were
distinguished by identifying inflection points in the curve through the plots
of the logarithm of the slope versus time.

     Tabak et a7. (34) and Oshima et a7. (27) sought to capitalize on
Dojlido's method of identifying inflection points in order to quantify more
fundamental rate coefficients.  In order to identify more fundamental (and,
therefore, intrinsic) kinetic coefficients, they used the generalized concept
of oxygen uptake by Gaudy and Gaudy (8).  Substrate removal was divided into
exponential and declining phases separated by an inflection point, and the
endogenous phase.  They coupled substrate removal and cell growth to oxygen
consumption by imposition of an electron balance and consequently were able to
evaluate n  from oxygen consumption data up to the inflection point.  They
proposed the use of the lag time as an indicator of how difficult it is to
achieve acclimation to a test compound and their respirometric studies were
carried out with an unacclimated biomass.

     The justification for using respirometry to obtain intrinsic kinetic
coefficients lies in the concept of oxygen consumption as an energy balance
[Busch et a7.  (4); Gaudy and Gaudy (8)].  This concept states that all of the
electrons available in a substrate undergoing biodegradation must either be
transferred to the terminal acceptor or be incorporated into new biomass or
soluble microbial products.  If the concentrations of the substrate, the

                                     563

-------
products and the biomass are all expressed in units of chemical oxygen demand
(COD), then the oxygen uptake can be calculated in a batch reactor from an
equation relating oxygen uptake to substrate, biomass and soluble products
[Grady et a?. (13)].  Furthermore, since biomass growth and product formation
are proportional to substrate removal, this suggests that an oxygen uptake
curve can provide the same information as either a substrate removal curve or
a biomass growth curve.  This latter concept has recently been used by Gaudy
et a7. (12, 14) to calculate biodegradation kinetics.  Specific growth rates
obtained from growth studies as slopes of plots of Ln (biomass concentration)
or Ln(X) versus time at different substrate concentrations, compared favorably
with those obtained from exponential phase of respirometric oxygen uptake
curves as slopes of plots of Ln (d oxygen uptake/d time or Ln (dOu/dt)  versus
time.

     Studies of Grady et a7. (12, 14) have demonstrated that is is possible to
determine Intrinsic kinetics of single organic compounds by using only
measurements of oxygen consumption in respirometric batch reactors,  with the
use of computer simulation techniques and non-linear curve fitting methods,
intrinsic kinetic parameters were obtained from oxygen consumption data and
were shown to be in agreement with those obtained from traditional measurement
of substrate removal (DOC, SCOD, 14C)  or  cell  growth.
                             MATERIALS AND METHODS


EXPERIMENTAL APPROACH

Elgctrolvtic Respirometry

     The electrolytic respirometry studies were conducted using an automated
continuous oxygen uptake and BOD measuring Voith Sapromat B-12 (12-unit
system) electrolytic respirometer-analyzer.

     The nutrient solution used in these  studies was an OECD synthetic medium
(OECD Guidelines for Testing of chemicals, OECD, Paris, France, 1981)
consisting of measured amounts per liter  of deionized distilled water of (1)
mineral salts solution; (2) trace salts solution, and (3) a solution (150
mg/L) of yeast extract as a substitute for vitamin solution.

     A comprehensive description of the procedural steps of the respirometric
tests and of the experimental design employing test and control systems is
presented elsewhere [OECD Guidelines for  Testing of Chemicals  (1983), Tabak et
a?. (1984)].

     For fully automatic data acquisition, frequent recording  and storage of
large numbers of oxygen uptake data, the  Sapromat B-12 recorders are
interfaced to an IBM-AT computer via the  Metrabyte interface system. The use
of a Laboratory Handbook software package allows the collection of data at 15
minute intervals.

                                     564

-------
Determination of Substrate Biodegradabil itv from Oxygen Uptake Data

     In this study, biodegradation was measured by three approaches:  the
first, as the ratio of the measured BOD values in mg/L (oxygen uptake values
of test compound minus inoculum control - endogenous oxygen uptake values) to
the theoretical oxygen demand (ThOD) of substrate as a percent; the second as
a percentage reduction of the original amount of test compound as measured by
dissolved organic carbon (DOC) changes [OECD Guidelines for Testing of
Chemicals (Method DGXI 283/82, Revision 5) (26)]; and the third, as a
percentage reduction of the original amount of test compound as measured by
specific substrate analysis.

     Graphical representation of percent biodegradation based on the BOD/ThOD
ratio were developed against time for each test compound. The experimental DOC
data for the initial samples and samples from reaction flasks collected at the
end of experimental run were used to calculate the percent biodegradation
based on the percent of DOC removal in the culture system.

Determination of Kinetic Parameters of Biodegradation

     The Monod equation, relating cell growth to biomass and substrate
concentration and the linear law, relating cell growth to substrate removal
are the most popular kinetic expressions which can provide adequate
description of growth behavior during biodegradation of substrate.  The Monod
relation states that cell growth is first order with respect to biomass
concentration (X) and mixed order with respect to substrate concentration (S)
by the equation


                          dx/dt = (Sigo/oc,  +  s)                          (i)


Cell growth is related to substrate removal by the linear law by the equation


                             dX/dt - -Y(dS/dt)                             (2)


      The kinetics of biodegradation were evaluated by quantifying ^,,  Ks  and
Y kinetic parameters expressed in the equation for the rate of substrate
removal . rs:


                                                                         (3)


where X is the concentration of biomass capable of utilizing the organic
substrate and Y is the biomass yield coefficient for the compound and in the
Honod equation;
                                        . + Ss)


                                     565

-------
(1f the compound is not inhibitory to its own biodegradation) or by the
Ha] dang T eg uat lot):


                       d -  /imSs/Ks + S.  + (S^)                        (5)


if the compound is inhibitory.  In these equations, nm is the maximum specific
growth rate, Ks is the half saturation coefficient, Kj  is the inhibition
coefficient and S is the concentration of substrate.


Determination of Rates of Exponential and Declining Growth

      The firstorder kinetic rate constants (specific growth rate parameters)
were determined by the linearization of the BOD curves or transforming  the
typical BOD curve to the linear function of time t, by the relationship of log
dOy/dt to t, which gives straight lines expressing the exponential  and
declining endogenous phases of the BOD curve.  The slope of  the Ln(d oxygen
uptake/dt) versus t  give specific rate constants of the exponential growth
phase (n values) and the declining growth phase (p' values)  of the BOD  curve
as described by Dojlido (7), Tabak et a7. (34), Oshima et a7. (27), and Tabak
et a7. (35, 36).

      Acclimation time values (t0) and the time values for the initiation and
termination of the declining growth phase (tj and t2)  for each test compound
were determined from linearized expressions of BOD curves.


Estimation of Monod Kinetic Parameters

      Based on the concept of oxygen uptake as energy balance (Gaudy and
Gaudy, 1971), if the concentrations of the substrate, the products and  the
biomass are all expressed in BOD  units, then the oxygen  uptake (Ou at any time
in the batch reactor may be calculated [Grady et a7. (13)] from


                    Ou - (Sso - Ss) - (X-X0)  - (Sp-Spo)                    (6)


where S  ,  S  and X0 are the concentrations of substrate, products  and cells,
respectively, at time zero, and Ss and S  are the  concentrations  of substrate
and product, respectively, at time t.

      To apply equation (4) for the determination of kinetic coefficients,
equations must be available which express the concentrations of soluble
substrate (Ss), soluble product (Sp)  and  biomass  (X)  as functions of time.
For batch reactors, those equations are:
                                     566

-------
                        dSs/dt = -(/^Y^X/K, + S.)                       (7)

                      dSp/dt - Yp(-dS/dt)  = (YpM,/Y)SsX/{Ks + S.)           (8)

                 dX/dt = »mSsX/(Ks + Ss) - KsbX/(Ks  + S.)                   (9)
where Ss = soluble substrate concentration; S  = soluble product
concentration; Y  = product yield; and b = decay coefficient.

      In equation  (9) the effect  of decay  is negligible when the substrate
concentration is high, but as substrate concentration decreases, the
importance of the decay term increases and, finally, when substrate
concentration approaches zero, the decay term approaches traditional first
order rate.

      Equations (7),  (8) and (9)  are solved simultaneously and  the values of
substrate, biomass and product concentrations obtained over time are
substituted in equation (6) to calculate theoretical oxygen uptake.


A.    Determination of Y Constant

      The initial estimate of yield coefficient, Y can be obtained by
identifying the start of the plateau, which is the point at which substrate
concentration is approximately equal to zero. Using equations (2) and  (8),
equation (6) can be expressed as


                  Ou = (S0  -  S)  -  Y(S0 - S) - Yp(S0 - S)

                     - (1 - Y - Yp) (S0  -  S)                               (10)


At the plateau the substrate concentration is very low compared to S0,  so that
it can be neglected. This simplifies equation (10) and on rearranging  this
simplified equation becomes:


                            Y = (S0 ' Oupt)/So - YP                         


      Where 0  t is the cumulative oxygen update  value at the initiation  of
plateau, S0 is the concentration of substrate at time zero,  and Yp  is the
product yield coefficient.  Y constant can then  be estimated from equation
(11).

      A vertical line is drawn at  the point of intersection of  the tangents of
the exponential and plateau phases of the  curve.  The oxygen uptake value
obtained at the point of intersection of the vertical line (drawn through
intersection of tangents) and oxygen uptake curve is the 0 t value -
[cumulative oxygen uptake value at the initiation of plateau].

                                     567

-------
      The 0  t value  is then  substituted  into equation  (11)  ,  for  Y
determination.  Y  is soluble product concentration formed, divided by initial
substrate concentration. In  this study,  product yield  (Yp) was negligible.


B.    Determination  of tim Constant

      The initial estimate of n  is obtained by assuming that S » Ks in
equation (1) which simplifies this equation. Integrating this simplified
equation


                                X = X0 exp(0mt)                             (12)


Using equation (2) substitute S in terms of X in equation  (6) (with  Sp and Spo
both equal to zero)  and then rearranging


                             X = [0U/(1/Y -  1)]  + X0                        (13)


Substituting equation (12) in equation  (13), and taking natural log  gives an
expression from which #m can be estimated.


                      In [X0 + 0U/(1/Y -1)] - In X0 + nmt                   (14)


A plot of ln[X0 + 0U/(1/Y - 1)]  versus time,  t,  will give a straight line with
slope Mm.

      Hm - the maximum specific growth rate can be determined from
experimental oxygen  uptake curve plot  in the following manner:


(1)   Values of the  change of Ou with time (dOu/dt) or slopes are determined
      along the entire experimental oxygen uptake  curve.

(2)   These,dOu/dt (slope) values are then plotted against the cumulative Ou
      values for each time interval.

(3)   The slope of the developed linearized form of oxygen uptake curve is the
      estimated nm value.


C.    Determinationof K... Constant

      The oxygen uptake value Out associated with 1/2 of the estimate of  um is
used in equation

                                     561

-------
                         S - S0  - Out/(l - Y - Yp)  = !
-------
biomass (XQ)  concentration must be carefully measured in COD units.   The ratio
of the two values must lie in a certain range in order to allow independent
evaluation of /i ,  Ks  and  Y [Simkins  and  Alexander  (30,  31)].   Grady's  studies
(12, 13) have snown that a SS(/X0 ratio  of around 20 works well.  The value of
Y  may be estimated by determining the residual  stable SCOD concentration
after substrate depletion (plateau area).  It is numerically equal to the
residual SCOD divided by the initial SCOD.  The value of the decay
coefficient,  b, may be determined by fitting to the oxygen consumption curve
after the plateau when the only activity contributing to oxygen consumption is
endogenous metabolism and cell decay.  Once X ,  Sso, Y  and b are known, #m,  Ks
and Y may be determined by non-linear curve fitting techniques [Grady (12, 13,
14)].  Assuming a set of values for nm,  Ks and Y,  theoretical  oxygen  uptake  is
calculated using equations (6) - (9).  The error between the theoretical and
the actual oxygen uptake curve is calculated and then used to obtain new
estimates.  The procedure is repeated until  minimum error is found.

      The technique involves the calculation of a theoretical oxygen
consumption using oxygen uptake equation (6) and equations for substrate,
product and biomass concentrations  (7)  - (9) with assumed Monod parameters.
The residual  sum of squared errors  (RSSE) associated with the difference in
calculated and experimental oxygen uptake values is used to obtain new
estimates.  The above procedure is repeated until a minimum RSSE is found.

      The Grid Search technique was selected as a most suitable non-linear
curve fitting technique for application in the determination of the kinetic
parameters from oxygen uptake data, because it can allow easy discrimination
between local minima and the global minimum RSSE.  This technique enables a
comparison between the calculated and experimental oxygen uptake data.  The
value of Y is fixed.  For this value of Y, a pair of iim and K  which  give RSSE
is found on a fi'.Ks plane.  The above procedure  is repeated  with  other values
of Y.  Values of fim,  Ks and Y  which  give minimum RSSE associated  with  the
difference in calculated and experimental oxygen uptake data constitute the
best values of the kinetic parameters.

      The values of /tm, Ks and Y developed from  grid  search  technique,  which
when substituted into equations (7) - (9), will provide X and S values, which
(when substituted into oxygen uptake equation 4) will  in turn provide
calculated oxygen uptake values at the  region of the plateau, closest to the
experimental  oxygen uptake values, with a minimum RSSE, will constitute the
best quantitative kinetic parameter values.

      A more comprehensive description  of the procedural steps in the
determination of rates of exponential and declining growth as well as in the
estimation and quantitation of Monod kinetic parameters for toxic organic
compounds from oxygen consumption data  has been given elsewhere by Tabak et
a7. (1989a, b, c).
                                     S70

-------
Modifications for Volatile and Sparingly Soluble Compounds

Volatile Compounds--

      The equations (7) - (9) are valid for a non-volatile compound present at
a concentration below its solubility limit, so they have to be modified for
volatile and sparingly soluble compounds. The respirometer reactors being
sealed, when compound is added it partitions itself between gas and liquid
phases. As biodegradation proceeds and the substrate concentration decreases
in liquid phase, it is transferred from the gas phase to the liquid phase to
maintain equilibrium in accordance with Henry's law. If S is the total
substrate concentration in the system, SL is the substrate concentration in
the liquid phase, and V  and VL  are  volumes  of  gas  and  liquid  phases
respectively, then
                              SL - S/(l + HVg/VL)                           (17)


where H is the dimensionless Henry's Law Constant, specified by the  following
equation


                   H  =  [Henry's  Law  Constant  in  atm/moles/m3!               (18)
                                       RT

R is the universal gas constant and T  is the temperature  in "K.

     Equation 17 was derived to quantify the partitioning between the gas  and
liquid phases.  The  use of Henry's  Law Constant is justified since the molar
concentration of the compound is  approximately  10~4,  and hence  the activity
coefficient in the liquid phase is  equal to  1.0.


Let SL1  =  concentration of substrate in liquid phase at time t1

    SLZ  -  concentration of substrate in liquid phase at time t2


then using equation  (7)


     $L2 • SLi ' UlSAiXti/Wt/dC. + SL1)] + (SL1 - SL2)(HVg/VL)


As dt tends to zero  the above equation results  in


                   dSL/dt = - (MAX/Y)/[(KS + SL)(1 + HVg/VL)]               (19)


Substituting equation  (17) in equation (19)

                                     571

-------
                          dS/dt = -  (MmSX/Y)/(Ks' + S)                      (20)


where Ks' = Ks(l + HVg/VL).  Similarly equations  (8) and  (9) change to
                         dS/dt - (Yp^SX/YVdV + S)                      (21)

                dX/dt - [(/imSX)/(Ks' + S)]  -  [(Ks'bX)/(Ks' + S)]            (22)


Sparingly Soluble Compounds--

     A sparingly soluble compound will dissolve  into the solution  phase  up  to
its solubility limit and the undissolved part will exist as  a  different  phase.
As the compound undergoes biodegradation in  solution phase more  of the
undissolved part dissolves  into the solution, keeping  its concentration  in
solution phase constant at  the solubility  limit. Therefore,  degradation
proceeds at a constant substrate concentration as long  as total  substrate
concentration in the system is greater than  its  solubility limit.  Now, if the
compound is also volatile then depending on  Henry's law constant it will
partition itself into the gas phase and liquid phase. When the microorganisms
degrade this compound from  the solution phase, then compound from  the
undissolved part dissolves  into solution phase.  This keeps the compound
concentration in the solution phase and gas  phase constant.  If S is the  total
substrate concentration (per unit volume of  liquid phase) in the system  and Ss
is its solubility limit then the concentration in solution phase and gas phase
will be constant as long as S > Ss/(l + HV/VL).  So if  substrate  concentration
is held constant at Ss in equations (7),  (8) and (9)  then they can  be solved
analytically and the solutions are


                       X =  X0 exp[(/y>s -  bKs)t/(Ks + Ss)]                   (23)


                    S = S0  - [(nmSsm/(nmSs - bKs)](X  - X0)                 (24)


                     Sp = [(YpMmSs/Y)/(/imSs  - bKs)](X -  X0)                  (25)


Thus, equations (23),  (24)  and (25) should be used when the  total
concentration of the substrate, S is greater than Ss(l + HVg/VL)  and equations
(20), (21) and (22) should  be used when the  total substrate  concentration,  S
is less than or equal to Ss(l + HV/VL).
                                     572

-------
                            RESULTS AND DISCUSSION

     Respirometric biodegradability, biokinetic and Monod kinetic data for
selected RCRA alky! benzenes, phenols, phthalates and ketones were generated
in these studies.  The electrolytic respirometry oxygen uptake data for the
test compounds, the control reference compound aniline, the inhibition and
endogenous control systems were generated revealing the lag phase (acclimation
phase), the biodegradation (exponential) phase, the different bio-reaction
rate slopes (characteristic of the test compound) as well as the plateau
region at which the biooxidation rate reaches that of the endogenous rate of
microbial activity.

     According to the OECD interpretation of biodegradation (BOD) data, test
compounds were judged to be easily biodegradable if they showed a high level
of oxygen uptake within 10 days after the observed level of biodegradation
first exceeds 10 percent (the time at which 10 percent biodegradation is
achieved is considered the lag or acclimation period).  In addition, the
results of biodegradation are valid if the control reference substrate,
aniline achieved 60 percent biodegradation within a period of 28 days.

     Based on the biokinetic equations relating growth rate of microbiota in
presence of above compounds, the substrate utilization rate, and rate of
oxygen uptake (BOD) curves, specific growth rate kinetic parameters
(biodegradation rate constants) were derived as slope values of the linearized
plots (plots of the log of dQ^/dt)  of exponential  and declining growth phases
of the BOD curve.  The acclimation time values (t0),  and time values for the
initiation and the termination of the declining growth phases (ti and t2)  for
the test compounds and aniline were also generated.

     The estimations of the Monod kinetic constants for benzene, phenol,
phthalate, and ketone compounds were determined directly from experimental
oxygen uptake curves without the consideration of initial growth and growth
yield assumption.

     Subsequently the Monod kinetic constants were quantitated with the use of
computer simulation methods coupled with a non-linear regression technique
(Grid Search Method).  The values of urn, Ks and Y were developed from this
technique with the use of oxygen uptake equation and equations for substrate,
product and biomass concentration and through a comparison of experimental and
theoretical (calculated) oxygen uptake curves, with a minimum RSEE value
associated with them.

     The biodegradation of alkyl benzenes, phenols, phthalates and ketones and
the reference compound, aniline at 100 mg/L concentration levels in synthetic
medium containing 30 mg/L sludge biomass (dry weight basis) was measured with
the use of oxygen consumption values and the ratios of BOD values to the
theoretical oxygen demand (ThOD) of substrate.  The oxygen uptake data were
generated over a period of 20-40 days.

     The oxygen uptake rate revealed that all of them were biodegradable at
the concentration level of 100 mg/L and under the experimental conditions used

                                     573

-------
and according to OECD interpretation of biodegradability of the compounds
under such environmental conditions.

QUANTITATION OF MONOD KINETIC CONSTANTS

     Table 1 summarizes the quantitative kinetic parameter values for the
alkyl benzenes, phenols, phthalates and ketones and for aniline and provides
the average percent error values between the experimental oxygen consumption
values and the theoretical (calculated) oxygen uptake data (derived from best
fit parameter values).  The average percent error between the experimental and
calculated oxygen uptake data was calculated by


               Absolute  Value  of [(100/n) 2  ((Ou)e - (Ou)p)/(0u)e]


where 'e' and  'p' denotes experimental and predicted values respectively. The
biokinetic constants for volatile and non-volatile compounds were calculated
using equations (19) -  (22).  For sparingly soluble compounds equations (23) -
(25) were used until the total concentration was greater than modified
solubility limit and equations  (19) - (22) used thereafter.

     The values of biokinetic constants given in Table 1 are the average of
two replicates except for aniline, which had eight replicates, so for aniline
the standard deviation  is given along with the mean values.  The representa-
tive oxygen uptake curves are given in Figures 1-4, which illustrate the
agreement between the experimental (asterisks) and the calculated (solid
lines) data. The three  biokinetic constants have major impacts on different
parts of the curve. The maximum specific rate constant n,  affects the
exponential part of the curve and examination of data sets revealed that
generally good agreement was obtained in this region. The oxygen uptake at the
plateau is affected mainly by the yield coefficient Y. The predicted curves
agreed well with experimental curves at this point, so one could be confident
about yield values. The half saturation constant Ks,  defines the shape of the
curve before the plateau. The agreement in this region was not very good for
all the test compounds, so one would tend to be less confident about the
values of Ks.

     The variability of oxygen  uptake data between replicates of the compounds
is typical for biodegradation of toxic compounds (Urano and Kato, 1986; Grady,
et a7., 1989 and Gaudy, et a/., 1988) and this appears to be inherent in the
use of natural microbiota (Grady and Lim, 1980). Even for the same compound,
different patterns of oxygen uptake have been observed several times
throughout this study as seen in Figures 1-4.

EFFECT OF DIFFERENT ENVIRONMENTAL PARAMETERS ON RATE CONSTANTS

     To study  the effects of different parameters; temperature, concentration
(substrate as well as biomass)  and source of biomass, 10 compounds were
selected from  the list  of 32 which were biodegraded. The criteria for
selection was  that they had to  be present in at least three of the four USEPA
lists; RCRA list, Superfund list, Priority pollutant list and Best

                                     574

-------
Demonstrated Available Technology (BOAT).  The ten compounds selected based on
these criteria were benzene, toluene, ethyl benzene, phenol, resorcinol, 2,4-
dimethyl phenol, dimethyl-, diethyV, dibutyl-and butyl-benzyl phthalates.

     A statistical analysis was performed to find whether a kinetic rate
constant was similar or different under various treatments for each parameter
(Walpole and Myers, 1972). One way analysis of variance (ANOVA) was used for
this purpose. The theory for one way ANOVA assumes that various treatments
have a common variance. Hence, Cochran's test was used to check the equality
of variances for different treatments.  The value of f, which is the value of
the random variable F having the F distribution was calculated and compared
with the critical value at 95 percent confidence level, denoted by the
variable F.  If the value of f was less than the critical value, F, then the
hypothesis that the kinetic parameters under all treatments are equal is
valid.  The Cochran's equality of variance test, requires the calculation for
the parameter g and comparing its value with the critical value G at 95
percent confidence level (Walpole and Myers, 1972).

Effect of Substrate Concentration

     The experiments were conducted at 25°C and at substrate concentrations of
20, 40, 60, 80 and 100 mg/L, using sludge from Little Miami wastewater
treatment plant  (Cincinnati) as inoculum.  The duration of each of the
experiment was about 20 days.  Duplicate runs were made for each of the five
concentrations for all the compounds.  The results of these experiments are
presented in Table 2.  In Table 2 F and G are critical values and f and g are
calculated values at 95% confidence level for ANOVA and Cochran's test of
equality of variances.

     The F-test showed that nm  values are similar at all the concentration
levels for all the compounds. This suggests that Mm is independent of initial
substrate concentration. Similar trends have been confirmed for ethylene
dibromide, aniline and m-nitro phenol (Aelion, et a/., 1989), at low
concentrations of phenol  (Scott, et a/., 1983) and for phthalates and
polyaromatics  (Grady, et al., 1989a). Fannin, et al. (1981) reported that the
degradation rate decreased with increase in concentration for phenol. They
studied phenol degradation in the concentration range 100-400 mg/L and phenol
is  inhibitory to microbial degradation beyond 100 mg/L (Grady, et al., 1989),
so  degradation rate decreases with increasing concentration beyond 100 mg/L.
Variation in lag time is not more than one day for a substrate concentration
range of 20 -100 mg/L, which indicates that  at least in this concentration
range lag time  is independent of substrate concentration.


Effect of Biomass Concentration

     The experiments were carried out at 25"C, with a substrate concentration
of  100 mg/L and  biomass concentrations of 20, 30 and 40 mg/L.  Activated
sludge from the  Little Miami wastewater treatment plant  (Cincinnati) was used
as  inoculum. The duration of experiments was 10-20 days. Experiments were done
using duplicates for all the compounds at each concentration level of biomass.

                                     575

-------
The results of these experiments are presented in Table 3 along with the
statistical analysis.

     The F-tests shows that the specific rate constant /L,  is independent of
initial biomass concentration. Aichinger (1989) has found a similar trend for
phthalates and polyaromatic hydrocarbons. He has used a method similar to the
one used in this study to estimate kinetic parameters from oxygen uptake data.
Fannin, et al, (1981) and Nabivanets, et al. (1975) reported that increasing
the amount of bacterial inoculum caused a significant linear increase in the
degradation rate. They had calculated the rate by dividing percentage degraded
by time period (%/hr). This was the simple approach of calculating the rate
without taking the biomass concentration into account. Their findings could be
explained simply:  when larger bacterial populations are initially present in
the inoculum, faster substrate consumption would result.  As shown in Table 3
there is a distinct decrease in lag time with an increase in initial biomass
concentration for all the compounds except phthalates, which may be due to the
presence of a larger number of microorganisms at the start of the experiment.


Effect oftemperature

     To study the effect of temperature, experiments were carried out at
substrate concentration of 100 mg/L and at three different temperatures 15°,
25* and 35*C, using activated sludge from the Little Miami wastewater
treatment plant (Cincinnati) as inoculum.  The duration of experiments was 10-
20 days.  Experiments were conducted with duplicate samples of all the
compounds at each temperature.  Microorganisms were not acclimated to the
temperature of 15* or 35 *C before the start of the experiment. The results of
these experiments are presented in Table 4.  At 15°C benzene, toluene and
ethyl benzene were not mineralized completely at the cessation of oxygen
consumption. The residual material were probably the transformation products,
although no studies were done to identify them.  These products may have been
inhibitory to further degradation of three benzenes.

     The maximum specific rate constants for all the compounds increased with
an increase in temperature over a temperature span of 15°-35eC. The effect of
temperature on the rate constant was described by the Arrhenius equation


                               j^  - A exp(-E/RT)


where     A    -    frequency factor (hr'1)

          E    -    activation energy (cal/mol)

          R    -    gas constant (1.986 cal/°K x g mole)

          T    -    absolute temperature (°K)
                                     576

-------
Arrhenius plot of the temperature data showed good agreement for different
compounds (Figure 5).  Table 4 lists the frequency factor and the activation
energy of some of the test compounds.  These activation energies of test
compounds except for phenol and diethyl phthalate are in the range reported
for ordinary enzyme reactions and they indicate that the degradation of these
compounds in wastewater treatment plant is not thermodynamically hampered
(Larson, et a7., 1981).  Several researchers have reported that degradation
rates are directly proportional to temperatures (Johnson and Heitkamp, 1984',
Hoi1 is, 1976 and Stephenson, et a?., 1983).  Furthermore, the results in Table
4 indicate that lag time decreases with increase in temperature.


Effect of Sludge Source

     The experiments were conducted with sludges from four different sources;
Little Miami wastewater treatment plant (Cincinnati), Test and Evaluation
Facility of USEPA (Cincinnati), Monsanto plant (Addyston, OH) and BP America
plant (Lima, OH).  The substrate and biomass concentration used was 100 and 30
mg/L respectively.  Duplicate samples were run for all the compounds with
sludge from each source.  The results of the experiments are given in Table 5.

     Statistical analysis showed that the kinetic constants obtained with
sludges from domestic, domestic-industrial and industrial sludge sources were,
except for a few exceptions, in fairly good agreement and were shown to fall
within a similar range as far as microbial activities are concerned.  In this
study correlations were made between microbial populations representing
domestic and industrial waste treatment sludge, and exceptions will occur
which could suggest that kinetic constant may be unique to the biomass
representing some particular industrial waste treatment system.  In such
exceptions, the history and composition of microbial community may play an
important role in the rate of biodegradation but the extent of biodegradation
does not depend on the culture used.  Data, however, support that in most
cases, there was a fair correlation between specific rate constant for any
compound biodegraded by microbiota from different sources of sludge, and that
the extent of biodegradation (% biodegraded) for all the test compounds) was
similar.  An even better correlation between kinetic constants were observed
in studies  (not reported here) relating sludges from different domestic waste
treatment systems only  (studies In progress).

     These data corroborate the work of Lewandowski (1988) on POTW treatment
of industrial organic wastes.  His experimental data suggest an unexpected
capacity of microorganisms from a domestic treatment plant (Livingston) to
degrade chlorinated hydrocarbons (2-chlorophenol, 2,6-dichlorophenol, and 2,4-
D) at rates that were virtually the same as those for the plant handling
industrial waste (PVSC).  Even the phenol degradation rates were the same,
although PVSC had experienced significant prior exposure.  In addition, the
dominant microbial populations were also very similar, both before and after
phenolic exposure.  Since the operational characteristics for these two plants
are very different, these results raise the possibility that plant-to-plant
differences in response to hazardous wastes may not be as severe as some have
imagined.

                                     577

-------
     Mierobial adaption to the compounds tested was quite rapid, resulting in
a 2- to 5-fold increase in the average degradation rates after only the second
exposure 1n a batch reactor.  When multiple substrates were tested, the
individual compound degradation rates in the mixture were very similar to the
degradation rates as sole carbon source.


                                  CONCLUSIONS

     Electrolytic respirometry approach involving natural sewage, sludge and
soil microbiota is a very successful, accurate, cost effective and less labor
intensive method for generating biodegradability/ inhibition and biokinetic
data.  This procedure measures the biodegradability and determines the
Intrinsic kinetic parameters of biodegradation by using measurements of oxygen
consumption in respirometric batch reactors and by applying non-linear curve
fitting techniques to oxygen uptake data.  The method is very promising for
automating data collection associated with biodegradation and the intrinsic
kinetic parameters.  The values of the kinetic parameters for single organic
compounds in synthetic media determined from oxygen consumption data are
similar to those obtained from traditional measurement of substrate removal
(DOC, SCOD, 14C)  or  cell growth thus  can  be  used  for  the  assessment  of  the
biodegradative fate of toxic organics in waste treatment systems.

     The electrolytic respirometry, with its automatic collection of data, is
a simple and reliable method to study biodegradation of toxic compounds,
including slightly soluble and volatile compounds. Experiments were conducted
for 56 organic compounds with aniline as the reference compound. The results
showed that 32 compounds degraded under the experimental conditions and for
these compounds Honod parameters were calculated. The course of biodegradation
of volatile and slightly soluble compounds could be modeled with the Monod
model.

     Experiments were also conducted to study the effect of temperature,
substrate concentration, biomass concentration and source of sludge on the
kinetic rate constant. The rate constant increased with increase in
temperature and was described by the Arrhenius equation. It proved to be
independent of the initial substrate and biomass concentration used. The
kinetic constants are not usually dependant on history and composition of
microbial community but there are a substantial number of exceptions. The
extent of biodegradation does not depend on the culture used.

     Wiggins, et a?. (1987) and Ventullo and Larson (1986) had suggested that
lag time in aquatic systems may reflect the time for multiplication of the
initially small population of active organisms to a critical level and
increased activity per cell.  If growth of a small degrading population to
significant levels account for the observed adaptation response, then at
higher concentration one might expect more rapid growth and therefore reduced
lag periods.  In this study it was found that in the concentration range of
20-100 mg/L, lag time proved to be independent of substrate concentration and
it decreased with increase in initial biomass concentration, which supports
the above hypothesis.

                                     578

-------
     Developed biokinetic data, unaffected by the varying environmental
conditions, will enable ultimate determination of the fate of organic
compounds of similar molecular structure to those experimentally studied by
way of the established predictive treatability models based on structure/
activity relationships.  A possible relationship between the kinetic
parameters and the effect of different factors on these parameters and the
structural properties of toxic organics can facilitate prediction of the
extent and rate of biodegradation in wastewater treatment systems.

     A biodegradation kinetic data base will enable the classification of
biodegradability and the inhibitory effects on the sludge biomass of the toxic
organic RCRA and CERCLA pollutant compounds.  The biokinetic data will provide
the basic pilot and full scale treatability information and an accurate
assessment of the need for pretreatment for these toxic pollutants.


                                ACKNOWLEDGEMENT

     The authors wish to thank Mrs. Rena M. Howard, Secretary in the U.S.
Environmental Protection Agency's Risk Reduction Engineering Laboratory,
Cincinnati, Ohio, for her excellent and timely wordprocessing skills in
preparing this manuscript.


                                  REFERENCES

1.    Aelion, C.M., R.C. Dobbins, and F.K. Pfaender.  1989.  Adaptation of
      aquifier microbial communities to the biodegradation of xenobiotic
      compounds: Influence of substrate concentration and preexposure.
      Environ. Toxicol. Chem., 8: 75.

2.    Aichinger, 6.  1989.  Determination of biodegradation kinetics of
      selected phthalates and polyaromatic hydrocarbons using electrolytic
      respirometry.  MS Thesis, Environmental Systems Engineering, Clemson
      Univ., Clemson, SC.

3.    Braha, A. and F. Hafner.  1987.  Use of lab batch reactors to model
      biokinetics.  Water Res.  21: 73.

4.    Busch, A.W., L. Grady, Jr., T.S. Rao and E.L. Swilley.  1962.  Short-
      term total oxygen demand test.  J. Water Pollut. Control Fed.  34: 354.

5.    Fannin, T.E., M.D. Marcus, D.A. Anderson and H.L. Bergman.  1981.  Use
      of a fractional design to evaluate interactions of environmental factors
      affecting biodegradation rates.  Appl. Environ. Microbiol. 42: 936.

6.    Cech, J.S., J. Chudoba and P. Grau.  1984.  Determination of kinetic
      constants of activated sludge microorganisms.  Wat. Sci. Tech. 17:  259.

7.    Dojlido, J.R.  1971.  Investigation of biodegradability and toxicity of
      organic compounds. EPA-600/2-79-163.  U.S. Environmental Protection
      Agency, Cincinnati, Ohio.

                                     579

-------
8.    Gaudy, A.F., Jr. and E.T. Gaudy.  1971.  Biological concepts for design
      and operation of the activated sludge process.  EPA-17090 FQJ.  U.S.
      Environmental Protection Agency, Cincinnati, Ohio.

9.    Gaudy, A.F., Jr., A.F. Rozich, S. Garniewski, N.R. Moarn and A.
      Ekambaram.  1987.  Methodology for utilizing respirometric data to
      assess biodegradation kinetics.  In Proc. 42nd Purdue Annual Industrial
      Waste Conf., Lewis Pub., Chelsea, MI, p. 35.

10.   Gaudy, A.F., Jr., A. Ekambaram and A.F. Rozich.  1988.  A respirometric
      method for biokinetic characterization of toxic wastes.  Paper presented
      at the 43rd Annual Industrial Waste Conference, Purdue University, West
      Lafayette, Indiana.

11.   Grady, C.P.L., Jr.  1985.  Biodegradation:  Its measurement and
      microbiological basis.  Biotech. Bioengineer, 27: 660.

12.   Grady, C.P.L., Jr. and H.C. Lim.  1980. Biological wastewater treatment:
      Theory and Application, Mercel Dekker, New  York.

13.   Grady, C.P.L., Jr., J.S.Dang, Harvey, D.M., A. Jobbagy and Wang, X-L.
      1989.  Determination of biodegradation kinetics through use of
      electrolytic respirometry.  Wat. Sci. Tech., 21: 957.

14.   Grady, C.P.L., Jr. J. S. Dang, D. M. Harvey, and A. Jabbagy.  1989.
      Evaluation of biodegradation kinetics with  respirometric data.
      Submitted to Water Pollut. Control Fed.

15.   Grady, C.P.L., Jr., G. Aichinger, S. F. Cooper and M. Naziruddin.  1989.
      Biodegradation kinetics for selected toxic/hazardous organic compounds.
      Hazardous Waste Treatment: Biosystems for Pollution Control, AWMA. (in
      press).

16.   Hollis, C.G.  1976.  Thermal effects on biodegradation of pollutants in
      water.  Research Report No. 53, Dept. of Biology, Memphis State Univ.,
      TN.

17.   Howard, P.H., S. Banerjee and A. Rosenberg.  1981.  A review and
      evaluation of available techniques for determining persistence and rates
      of degradation of chemical substances in environment:  An update of the
      1975 Report.  EPA-560/5-81-011.  U.S. Environmental Protection Agency,
      Cincinnati, Ohio, National Technical Information Services No. PB84-
      168731.

18.   Howard, P.H., J. Saxena, P. R. Durkin and L.-T. Ou.  1975.  Review and
      evaluation of available techniques for determining persistence and
      routes of degradation of chemical substances in the environment.  EPA-
      560/5-75-006.  U.S. Environmental Protection Agency, Cincinnati, Ohio,
      National Technical Information service No.  PB243825.


                                     580

-------
19.   Jenkins, D.  1960.  The use of manometric methods in the study of sewage
      and trade wastes.  In P.C.G. Isaac (ed.), Waste Treatment, Pergamon
      Press, Oxford, p. 99.

20.   Johnson, B.T. and M. A. Heitkamp.  1984.  Environmental and chemical
      factors influencing biodegradation of phthalic acid esters in fresh
      water sediments.  Environ. Pollut. (Series B), 8: 101.

21.   King, E.F. and B. I. Dutka.  1986.  Respirometric techniques.  In:  G.
      Britton and B. J. Dutka (eds.),  Toxicity Testing using Microorganisms.
      Vol. I, CRC Press, Boca Raton, Florida, p. 76.

22.   Larson, R.J. and R. L. Perry.  1981.  Use of electrolytic respirometer
      to measure biodegradation in natural waters.   Wat. Res., 15: 697.

23.   Levandowski, G.  1988.  POTW treatment of industrial organic wastes.
      Final Report, Project BICM-6, New Jersey Institute of Technology,
      Newark, New Jersey, p. 104.

24.   Montgomery, H.A.C.  1971.  The determination of biochemical oxygen
      demand by respirometric methods.  Water Res., 1: 631.

25.   Nabivanets, B.I., I. G. Garasevich and V. I. Kachan.  1976.  Effect of
      various factors on glucose decomposition rate in natural waters.
      Gidrobiol Zh, 11, p. 17, Biological Abstracts (1976) Group 5C - Effect
      of pollution.

26.   Organization of Economic Cooperation and Development.  1983.  OECD
      guidelines for testing of chemicals.  EEC directive 79/831, Annex V,
      Part C: Methods for determination of ecotoxicity, 5.2 Degradation,
      biotic degradation, manometric respirometry, Method DGX1, Revision 5,
      p. 1-22.

27.   Oshima, A., H. H. Tabak and R. F. Lewis.  1985.  The evaluation of
      biological treatability and removability of toxic organic chemicals by
      respirometry.  In-house report, U.S. Environmental Protection Agency,
      Water Engineering Research Laboratory, Cincinnati, OH.

28.   Paris, D.F. and J. E. Rogers.  1986.  Kinetic concepts for measuring
      microbial rate constants  : Effects of nutrients on rate constants.
      Appl. Environ. Microbiol., 51: 221.

29.   Robinson, J.A. and J. M. Tiedje.  1983.  Nonlinear estimation of Monod
      growth kinetic parameters from a single substrate depletion curve.
      Appl. Environ. Microbiol. 45: 1453.

30.   Simkins, S. and M. Alexander.  1984.  Models for mineralization kinetics
      with the variables of substrate concentration and population density.
      Appl. Environ. Microbiol. 47: 1299.
                                     581

-------
31.   Simkins, S. and H. Alexander.  1985.  Nonlinear estimation of the
      parameters of Honod kinetics that best describe mineralization      of
      several substrate concentrations by dissimilar bacterial densities.
      Appl. Environ. Hicrobiol. 50: 816.

32.   Stephenson, T., T.N. Lester and R. Perry.  1983.  The influence of
      transient temperature changes on the biodegradation of nitrilotriacetic
      acid in the activated sludge process: A pilot plant study.  Environ.
      Pollut. (Series A), 32: 1.

33.   Tabak, H.H., S.A. Quaye, C.I.Hashi, C.I. and Barth, E.F.  1981.
      Biodegradability studies with organic priority pollutant compounds.  J.
      Water Poll. Control Fed. 53: 1503.

34.   Tabak, H.H., R.F. Lewis and A. Oshima.  1984.  Electrolytic respirometry
      biodegradation studies, OECD ring test of respiration method for
      determination of biodegradability.  In-house Report, U.S. Environmental
      Protection Agency, Municipal Environmental Research Laboratory,
      Cincinnati, OH.

35.   Tabak, H.H., Desai, S. and Govind, R.  1989a.  The determination of
      biodegradability and biodegradation kinetics of organic pollutant
      compounds with the use of electrolytic respirometry.  Proceedings of EPA
      15th Annual Research Symposium:  Remedial Action, Treatment and Disposal
      of Hazardous Waste (in press).

36.   Tabak, H.H., S. Desai and R. Govind.  1989b.  Determination of
      biodegradation kinetics of RCRA compounds using respirometry for
      structure-activity relationships.  Proceedings of the 44th Annual Purdue
      Industrial Waste Conference, West Lafayette, Indiana, May 9-11, 1989 (in
      press).

37.   Tabak, H.H., S. Desai and R. Govind.  1989c.  Biodegradability and
      biodegradation kinetics assessment for toxic organics using oxygen
      uptake measurements.  Proceedings of the 12th U.S./Japan Conference on
      Sewage Treatment Technology, Cincinnati, Ohio, October 12-13, 1989  (in
      press).

38.   Ventullo, R.M. and R.J. Larson.  1986.  Adaptation of awuatic microbial
      communities to quaternary ammonia compounds.  Appl. Environ. Microbiol.,
      51: 356.

39.   Urano, K. and Z. Kato.  1986.  Evaluation of biodegradation ranks of
      priority organic compounds.  J. Hazardous Materials, 13: 147.

40.   Walpole, R.E. and R.H. Myers.  1972.  Probability and statistics for
      engineers and scientists.  MacMillan Pub. Co., New York, New York.

41.   Wiggins, B.A., S.H. Jones and M. Alexander.  1987.  Explanations for the
      acclimation period preceding the mineralization of organic chemicals in
      aquatic environments.  Appl. Environ. Microbiol., 53: 2617.

                                     582

-------
 TABLE 1.   QUANTITATIVE HOMO KINETIC PARAMETER VALUES
                                                                              TABLE Z.  RESULTS ILLUSTRATIH6  EFFECT OF  SUBSTRATE CONCENTRATION
Compound
Anlllnt

|*BZ*Q£ft
3*nztn«
Ethyl btnztnt
Tolutna
o-Xyltni
••Iyl«nt
p-Xyltitt
CiMnt
Butyl btnztnt
sftc-Butyl bcnzint
ttrt-Butyl btnzMl
l-Pti«nyl htxaiu
MithilitM
OlMthyl phthiUU
OUthyl phthiUtt
Dlpnpyl phmlatt
Dlfcityl phthiUU
HwMlc
PKtfNl
o-Crtttl
•-Crttel
p-Crtiol
2.4-OlnUi/l phmol
CttKhol
ta>orc1nol
Acstent
2-ftlttMM
4-NMHyl-2-M*tMOM
timiartm
K,
(hr'1)
0.2S3t
0.015
0.282
0.21«
0.523

0.123
0.140
0.194
0.10*
0.110
0.021
0.135

0.317
0.211
0.220
0.180
0.2*3
0.2*4
0.542
0.321
0.301
0.495
0.3*2
0.217
0.23*
0.240
O.lOt
K.
«9A
2».93»
2.61
6.57
10.07
7.75
.
0.75
2.47
10.03
10.00
10.00
10.00
7.47

7.00
7.4*
6.50
7.8*
3.00
1.7S
(.50
(.00
(.M
5.40
5.87
1.00
*.n
7.47
10. M
Y
«9/-g
0.43.
0.05"
0.27
0.34
0.36

0.26
0.36
0.23
0.23
0.62
0.22
0.53

0.49
0.36
0.23
0.38
0.30
0.53
0.44
0.3*
0.32
0.28
0.21
0.55
0.45
0.43
0.54
Av<).
X arror
6.2±2.4

S.32
8.77
5.62

12.94
13. 4]
19.21
18.29
0.12
18.91
8. 52

8.87
7.89
6 65
6.95
4.51
9.52
6.3*
8.04
5.21
8.19
5.32
5.C2
3.M
4.51
12.2*

Stnzent




Tolutnt




Ethyl
Mnzint



Phtnol




2.4-OlMthyl
pniml

R»ore1nol



Olnthyl
phthttnt


OUthyl
phthil «tt



Otbutyl
pftttaUtt



*L d
20
40
60
BO
100
20
40
60
80
100
20
40
60
SO
100
20 0
40 0
60 0
SO 0
100 1
60 2
(0 2
100 2
40
60
80
100
44
(0
1*
10*
20
40
(0
M
100
20
40
60
M 2
100 2
lyl hr
0 0.
S 0.
S 0.
7 0.
5 0.
8 0.
0 0.
2 0.
0 0.
2 0.
3 0.
7 0.
0 0.
0 0.
0 0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0 0.
0 0.
0 0.
0 0.
0 0
.0 0
.0 0

135
284
277
261
282
494
490
615
667
522
260
240
220
250
216
1M
2(4
381
317
283
1(0
258
303
540
517
5M
312
404
2tt
111
317
131
171
1(2
2*7
2(2
0*4
1(1
120
201
1*0
Y
, 0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0
0
0

3;:
3d
211
211
2:'
3;'
3J
4(1
J;i
31.
3t
3<
30
2Ji
3<
U
U
40
311
3Si
31
211
3i:
31.
3!'
31.
11:
il
411
441
411
111
4*1
4i'
4li
311
51
5(1
4
t.n
2.2
ts Conc.(«j/L) f F g G
30 40
0.212
0.27
6.57
3.5
0.522
0.34
7.71
2.20
0.216
0.34
10.07
3.0
0.2*1
0.10
1.00
1.0
0.3*3
0.12
7.0*
2.2
0.195
0.27
5.88)
1.1
0.117
0.4*
7.00
2.1
0.2*1
O.M
7.4*
2.0
0.1M
0.17
7.M
2.0
0.201 $.15 9.55 0.67 0.96(9
0.11
7.29
1.0
0.442 0.79 0.87
0.31
(.9*
1.85
0.18* 0.14 O.M
0.32
9.84
2.8
0.1SI 2.51 0.91
0.14
4.37
0.8
0.21* 2.57 0.41
0.14
(.00
2.0
0.11* 0.21 0.84
0.11
4.H
1.0
0.207 1.17 0.91
0.4*
7.17
2.7
0.141 (.11 O.M
0.40
(.40
2.0
0.21* 1.47 O.M
O.M
7.M
2.0
Compound
Btnzint



TollMIM



Ethyl
btnnn*


HWMl


2.4-OlMthyl
ptaml


KtKHXln.1


Olwttiyl
pMJulit*

Olttlnrl
pMtwltt*


pMtaiou


Piruttir
K.
Y
K
'49
H,
Y
K
ll,
7
K,

7
K,

.
1
K,
l«9
7
|f
'•9
7
&.
7
t
1*9
*
K
'••
T«
IS
0.261
0.22
7.44
8.5
0.514
0.32
7.07
6.5
0.210
0.12
10.51
6.5
0.220
0.2*
3.50
2.5
0.2(7
0.21
7.01
8.5
0.3*0
0.22
5.00
2.5
0.2N
0.41
(.00
i.50
0.22*
O.M
t.M
5. SO
0.172
0.1*
9.44
(.0
•Mritun |
21
0.282
0.27
6. 57
l.i
0.523
0.36
7.75
2.2
0.216
0.34
10.07
3.0
0.281
0.10
3.00
1.0
0.303
0.12
7.00
2.2
0.3*5
0.21
5.87
1.1
0.117
0.4*
7.00
2.50
0.2*2
0.34
7.4*
2.00
0.1M
0.37
7.M
2.0
35
0.292
0.29
7.44
2.5
0.530
0.35
(.27
2.0
0.221
0.36
9.15
2.3
0.332
0.34
4.87
0.8
0.334
0.27
5.M
1.0
0.420
0.2*
5.M
1.0
0.111
0.41
(.**
2.00
0.114
0.15
7.17
l.M
O.lt*
O.M)
».27
1.1
A E
hr'' kci!/«ol
0.14 0.94



0.81 0.28



0.46 0.44


27.74 2(.38


7.21 18.81



l.M 11.71


1.41 O.M

20.72 23.14


0.44 0.71


                                                              583

-------
             AnKne
             Conen.  : 100 rog/L
  160.00 -i
   120.00 -
 0>
_*
 o
"a. BO.OO
ID

 C
 0)
 CT»
 >> 40.00
 ?c
o
      n 00 - Tt i i I I I i t i iiit i i i n|IT i rrrnrvf'i-' i i i »•» in
         000     ., 5.00      10.00       1500      70 (Ml

                        Time,  hour:;

Figure 1. Oxygen uptake curves for biodegradation of aniline.


   100.00 i


                          Dimelhyl  Phlholate
                                                               100.00 -
                                                                80.00
                                                                 60.00
                                                             O
                                                             •4-f
                                                             CL
                                                             Z)
                                                                40.00
                                                             X

                                                            O 20-°°
                                                                           B«nzoic ock)
                                                                           Concn.  :  100  mg/L
                                                                 0.00 'pi I I I I I I I I I I I I I i I i i i i i i i i t i ! i i ! i r ri-ri-i-i i
                                                                     0,00        2.00        4.00        600        HOC)

                                                                                    Time,  hours

                                                            Figure 2. Oxygtn uptake curves  for biodegradation of benzole acid.

                                                                160.00 -i
                                                              cn
                                                              E

                                                             bJ
                                                                120.00 -
                                                             9-  80.00


                                                             •z.
                                                             LJ
                                                             X
                                                             O
                                                                 40.00-
     0.00
         0.00
                   4.00       8.00       12.00
                        TIME,  hours
                                                                  o.oo
4-Melhyl-2-Penlanone



           V
                                                                     o.oo
                        16.00
                                                                                 4.00       8.00       12.00
                                                                                      TIME,  hours

Figure 3. Oxygen uptake curves for dimethyl phthalate at  100 mg/L.    Figure 4. Oxygen uptake curves for 4-raethyl-2-pentanone at 100 wg/L.

-------
t
    -OB

    -O.Q -

      -1 -

    -1.1 -

    -1.2 -

    -1,3 -

    -1.4 -

    -1.8 -

    -l.§ -

     -1.7 -
                                          Anftwluft plot
_,B _j	,	,	,	,	r	j	,	,	
   OOO324      O.O0328       O.0033Z       0.00338      O.O034

                                           »A
                                                                       T	1	
                                                                           OO0344
                                                                                          O 00343
Q    Banian*              +   R«*orcliw»l                   »    Dtbutyl phlhololo
                Figure 5. Arrheniys plot for benzene,  resorcfnol and  dibutyl  phthalate.

CMPOUM
I«Mtru
Talmiw
EUiyl
MMM*
KKMl
2,4-OlHthyt
gMMl
faMIClrat
OiBlthyl
•ktlnltt*
01 (Owl
tktkiltti
(MIUIM*
TMLE
P«riMt*p
1
fr
txttnt
1
•sttnt
•xttnt
f-
UtMt
txtmt
?•
«tMt
UUM
f"
«MM
$. iEsuirs w
LUtlt Mtut
0.212
0.27
M!O
O.IN
0.34
7.7J
0.210
0,34
10,07
M.O
oito
3.W
H.O
0.303
O.lt
7.00
M.O
o!ii
5.17
17.1
0.117
?!oo
M.f
o.m
O.M
7.50
M.S
0.110
0.17
Ml
H.O
: ana OF
TIE
o.m
0.23
1.77
13.0
0,371
O.M
6.52
4Z.O
0.130
0.10
2.M
9S. »
0.1 54
O.M
9. SO
H.O
0.171
0.34
3.00
91.0
O.lil
0,31
4.M
U.I
o.m
0.41
2.00
M.O
0.140
0.47
2.00
90.0
0.211
0.11
7.1J
•4.1
sunn SOU«CE
Noftsinte
0.122
0.22
7.S4
51. S
0.141
4^51
(0.9
0.240
0.26
7.17
M.O
o!}»
s.oo
94.0
0.1M
0.30
1.50
17.0
0.3S1
0.11
*,77
«8.0
O.llt
0.40
s.zi
0.114
O.M
10.00
If.O
O.US
0.41
4.12
M.O

>P
O.S72
0.22
9.24
52.0
0.771
0.29
a. 13
<4.0
0.499
0.2S
10.12
99.0
0.31S
0.22
(.99
0.113
0.21
7.00
M.O
0.194
0.30
l.SO
91.0
0.440
0.47
S.OO
M.S
0.411
0.32
7.41
M.I
0,4(7
O.M
s.u
H.I
                                               585

-------
                          CHEMICAL COMPATIBILITY OF
         GEOTBXTILES. GEONETS AND PIPES WITH HAZARDOUS WASTE

                         by: P. Cassidy, M. Mores, D. Kerwick
                              Southwest Texas State University
                              San Marcos TX  78666

                                           and

                              K, Versehoor, D. White
                              Texas Research International
                              Austin TX  78733
                                     ABSTRACT
       Despite the advanced technology displayed by the United States chemical industry,
there exists little comparable technological capacity to deal with the billions of tons of
improperly disposed toxic chemical  waste.  Due to this lack of technology, research in the
field of geosynthetics  used in hazardous waste containment began  approximately ten years
ago. One problem facing the advancement of modern geosynthetic technology is the lack
of consistent testing methods.  For this purpose, a project was undertaken to develop a
technological base for predicting the long-term durability and compatibility of geotextiles,
geonets and pipes  in contact with hazardous waste.
       In the initial phase of this project, an extensive literature search was conducted on
existing information to investigate the performance and failure criteria for these
geosynthetics. From this literature review,  an  article has been prepared on geotextiles,
geonets,  and  pipes used in landfill sites, types  of polymeric materials, exposure and
characterization methods.  In addition to focusing  on the available information of the
chemical compatibility of geosynthetics, this review has  identified  some areas of current
and future study,  such as the use of analytical techniques to determine the molecular
interaction between geosynthetic materials and leachate components.
       The second phase, currently in the planning stage, will address some of the
significant questions and problems of chemical compatibility testing of geosynthetics.
Some- of these concerns are the determination of test methods that will reflect field
performance and development of novel accelerated-aging techniques. Handling of
chemicals and the equipment required for the establishment of safety procedures is another
concern.  A more  complete understanding of the failure  mechanisms will allow for better
design and material choices in future landfill sites.  New guidelines have been
recommended for the  evaluation of geotextiles, geonets and pipes used in  hazardous waste
facilities. These recommendations are based upon the information obtained in the first
phase of this study and will be the  subject  of the  experimental work in phase two.
       This paper  has been reviewed in accordance with the U. S. Environmental
Protection Agency's peer and administrative review policies and approved for presentation
and publication.
                                          586

-------
                                  INTRODUCTION
       The need to develop a technological base for predicting the long-term durability and
compatibility of geotextiles, geonets and pipes with hazardous waste  has been evident for
some time.  A  cooperative study consisting of two phases was initiated to determine the
inadequacies  of present practices among engineers in this industry.  The first of two phases
included a comprehensive survey which enabled those involved to define the technology
base presently available to evaluate the chemical compatibility of these geosynthetic
materials. The following general objectives were proposed:

       *     Providing an up-to-date review of data produced within the government,
             industry and academia to assess the chemical compatibility of geosynthetic
             products other than geomembranes used in construction of hazardous waste
             landfills;

       *     Recommendation of improved test methods to assess chemical compatibility
             of geotextiles,  geonets and pipes;

       *     Recommendation of criteria for evaluation of chemical compatibility data for
             material selection and permit application review purposes.

In addition to the proposed objectives, the final recommended test methods and evaluation
criteria should include:

       *     Measurement of physical or molecular-level characteristics that can  reflect the
             product's ability to perform its design function;

       *     Minimization of testing laboratory cost and elimination of tests which do not
             reflect field performance or are redundant;  and

       *     Production of meaningful  results within a reasonable testing time.

                                 PHASE I
The initial phase consisted of an extensive information gathering process that included
compilation of a database and subsequent preparation of a review article, assembly of data
from mail surveys, phone and personal interviews with manufacturing organizations, testing
laboratories and facility designers.  This  information gathering allowed for the
establishment of performance and failure criteria, review of analysis and existing data in
order to define the Phase n test plan.  Specifically  this portion of the project provided:

       *      A table of chemical resistance recommendations for geosyntheties;

       *      A survey of methods currently used by environmental engineering firms and
              commercial testing labs to assess chemical compatibility;
                                           587

-------
       *     The suggestion of a generalized procedure that would incorporate current
             testing protocols utilized by labs now performing chemical compatibility
             testing; and

       *     An experimental plan for Phase n to further define laboratory procedures  and
             evaluation criteria.

Overall, as a result of this work, polymer degradation phenomena will be further
understood which will establish new modes of technology exchange among academia,
industry and the government regulatory community.

TEST METHODS

       With the exception of the EPA's method 9090, no standardized methods exist for
analyzing the durability of geosynthetic materials before or after exposure to the
environment or aggressive chemicals.  In the past, the durabilities of the geosynthetics have
traditionally been assessed on the basis of mechanical property  test results. The evaluation
of microstructural or molecular changes which can cause changes in bulk properties has
been performed, but not extensively.  Analytical techniques developed for the evaluation of
polymer properties can aid in the-determination of changes in the molecular structures of
the base  polymers used to manufacture the geosynthetic(l).

Specialized Analytical Techniques

       Analytical techniques have found limited use hi this field so far, but hold great
promise for the future.  These specialized techniques elucidate the relationship between
microstructural changes in the base polymer and physical and mechanical property changes
of geosynthetics.  The information obtained will provide a more complete understanding of
the interaction between geosynthetic materials and the leachates they contain.  From this a
more accurate prediction of the service life-time of geosynthetics can  be deduced.  These
techniques, described below, include analytical methods commonly used to characterize
components on the  molecular level.

Thermal  Analysis ~ Thermal analysis includes a range of techniques for determining the
temperature dependence of the polymer property changes. Mass,  heat capacity, heat of
chemical or physical change, mechanical response and volume are the most commonly
observed physical property changes examined by  thermal analytical methods.  A good
indicator of structure-behavior relationships in polymer applications, this measurement of
property  changes versus temperature has prompted the development of these techniques
specially for polymer work(2).

Spectroscopy — Spectroscopic techniques provide information on the compositional and
structural characteristics of a polymeric material.  Infrared (IR)  spectroscopy is particularly
useful in the identification of characteristic functional  groups and molecular configurations
by simple inspection and reference comparison.  The sample is subjected to infrared
radiation of successively decreasing frequencies.  A series of spectral  bands is generated,
each of which correlates to  a particular frequency or range of frequencies where the
organic (functional groups in die plastic) material absorbs radiation.  These characteristic

                                           S88

-------
absorption frequencies indicate certain functional groups in the polymer(l).  Comparative
IR techniques  can be a sensitive method to detect changes in the polymer.

Chromatography ~ Chromatographic methods allow the  separation, isolation and
identification of closely related components in complex  mixtures. Due to the diversity of
these  separation techniques, other analytical methods do not possess the ability to identify
components from mixtures with such accuracy.  Chromatography identifies components in
the gaseous, liquid or solid state,  which may include substances  that have been absorbed
into the geosynthetic material itself.  This valuable information can  be used to identify
various components of leachates or products that result from polymer degradation.

Microscopic Analysis — The use of magnification to evaluate geosynthetic materials is
provided by various  types of microscopic techniques.  Microscopy supplies rapid, direct
observational information  about the surface as well as the internal microstructure and  defect
distribution within the materials.  Microscopic analysis is  also a  valuable tool for the
examination of geosynthetic  materials before and after exposure  to aggressive chemical
media.

Viscosity Measurements — Relative changes in molecular  weight may be monitored by melt
flow index (MFI)  testing and dilute solution viscosity (DSV) testing.  Both of these tests
measure either the melt or solution viscosity of the  polymer used in the geosynthetic
product. These measurements are  related to the average  molecular weight of the polymer
which in turn  is related to the properties of the polymer.

CURRENT PRACTICE

      According to RCRA Part B permit applications on file at several EPA regional
offices, until very recently, the general practice has  been to cite  either manufacturers'
recommendations or general chemical resistance  data from technical references,  such as
plastics' applications' handbooks,  to demonstrate the compatibility of the geosynthetics
other  than  liners, proposed with a  specific waste of known composition.  These referenced
data or recommendations  were submitted along with  Method 9090(3) results for the
proposed geomembrane liners in the original Part B application.

      More recently, the trend has been to perform laboratory testing for all geosynthetics
to be  used in  the proposed containment facility design.  This testing may  be
"piggybacked" with the chemical compatibility testing required for geomembranes as
required by Method  9090. Several commercial laboratories  have performed a significant
number of chemical  compatibility  studies in which geotextiles, geonet, pipe or a
combination of these were tested  against a  specific waste  leachate in a manner analogous
to Method 9090 as applied to  liners.

Product Sampling Procedures

      Variability in properties among samples of geosynthetics  must be understood, since
chemical compatibility  testing  depends on the  ability to  compare properties measured before
and after chemical immersion exposure.
      Geotextiles, in particular, demonstrate  significant variability within a given roll or

                                          589

-------
product lot in weight, thickness and mechanical properties.  This is a normal characteristic
and relates to the manufacturing methods used.  This variability in properties must be
understood,  since chemical compatibility testing depends on the ability to compare
properties measured before and after chemical immersion exposure.  Interpretation problems
could result if, for  example,  a  much thicker section of geotextile were cut into samples and
tested at zero days, while samples tested at a later immersion interval were cut from a thin
area.

       This  problem has been recognized by the commercial laboratories, and four ways of
eliminating or controlling variability are in use.  These methods have been developed so
that changes observed before and after  exposure may be attributed to the degradation rather
than variations within the material.

       The first method involves recording the weight and thickness of each  specimen to
be included  in the  immersion testing.  Each specimen is assigned a coded identification
number and individual specimen weights and thickness  are recorded into a database which
will be kept  for reference throughout the project. A numbered  tag is attached to each
specimen.  Specimens are not screened; each weighed specimen is included in the test
matrix regardless of weight or  thickness.  Prior to assigning specimens to exposure baths or
intervals, all cut, weighed and  tagged samples are "shuffled" physically so that weights and
thicknesses throughout the range are evenly distributed.  When test data are reported,
original weights and thickness  from the unexposed specimen database are included with the
test results.   Widely varying  data may  be identified and considered  in the context of any
apparent trends in the data.  This method is often preferred, but is not used  exclusively.

       The second  method involves  a pre-screening of cut specimens. First, average roll
values for weight and thickness are  developed by measuring and weighing 20 specimens
cut randomly from various parts of the roll.  Then,  specimens  for each physical  property  to
be measured at each exposure interval are cut from the original roll. The quantity to be
cut for each  property is increased over  that needed for the actual testing by a factor of 1.5
or more.  After cutting, each individual specimen is weighed.  If the weight falls outside
one standard deviation from  the established roll values, that specimen is rejected and
discarded.  Only specimens falling within one standard  deviation for both weight and
thickness are tested.  Pre-screened specimens are not marked or identified, and weights are
not recorded.

       The third method involves cutting all specimens required for baseline and immersion
interval testing from one selected region across  the width and along the length of the
supply roll of geotextile.  Weight and thickness compliance with minimum roll  value
specifications is verified before testing  begins.   This method recognizes that due to the
manufacturing procedures  used, product weight and  thickness tend to vary in  a consistent
way across any given roll.

       The fourth method consists of testing seven replicates for each mechanical property
test, and rejecting the two of the seven  which produce data outside a reasonable standard
deviation so that the reported test value is an average of five replicates.  No  pre-screening
or sample selection is done prior to testing.
                                           590

-------
       Of the six testing laboratories contacted, five reported having significant experience
with evaluating chemical compatibility of geotextiles.  Of these five, one uses the first
method, two the second, one the third and one the fourth for addressing the problem of
geotextile variability.  Rationale for the first method is that an  assessment of product
variability can be obtained as a "by-product" of chemical compatibility tesdng, since
weights and thicknesses are recorded and reported. Also, the recorded physical data may
prove useful in the later data evaluation, for example, to explain data  which show
inconsistencies or unexpected trends. Rationale for the second, third and fourth methods is
that chemical compatibility testing is intended to determine whether a  specific leaehate
interacts  with a geosynthetic product in any measurable way; since this issue is unrelated to
product variability,  the best approach is to eliminate this variability as much as possible,
thereby simplifying data analysis.

       The sampling problems  described are less critical for geonet and pipe products.   For
these, special procedures analogous to those described above have not been developed.
Product sampling procedures arc based on requirements of the  various test method
standards.

Immersion and Leaehate Handling Procedures

       Immersion bath design,  leaehate preparation procedures  and leaehate: handling for
geonet, geotextile and pipe compatibility tests are commonly performed following Method
9090 guidelines.  It has been found that these general procedures, originally developed for
geomembranes, are equally suited for the other products, and by using the same procedures
and equipment a consistent test exposure protocol may be  applied to all materials included
in the  test plan. Equipment differs  in design and construction  details; however, each
commercial  laboratory includes the  basic elements as  required by Method 9090  to maintain
leaehate immersion baths at two temperatures, while providing  for uniform dispersion
within  the tank or vessel, exposure of all specimen surfaces, and containment of
evaporating  volatiles.

Test Method Selection Criteria and Test  Method Modifications

       Since standardization has been lacking for chemical compatibility testing of
geotextiles, geonet and pipe, the selection of test methods for evaluation of property
changes caused by  leaehate exposure has been  left up to the individual facility owner or
designer.  From an engineering standpoint it is desirable to learn whether design or
functional performance properties of the geosynthetic  component are affected by extended
leaehate exposure.   However, it has not been practical in most  cases to measure
performance  properties,  and it has become standard practice to  select index tests for
inclusion in chemical compatibility  test programs. The goal in selecting and evaluating
results  of index tests is  to determine whether the leaehate is interacting with the
geosynthetic  product in  a way  that  might degrade the material  or its ability to perform an
intended design function over the long term. Though index test results are not  directly
transferable  to field performance, they provide  a scale by which degradation can be
monitored.

       Criteria used to select index or performance tests for inclusion  in chemical

                                           591

-------
compatibility studies can be summarized as follows:

       *     The test must provide reproducible results and be relatively independent of
             experimental factors or operator-introduced errors.

       *     The test must require relatively small individual specimens so that leachate
             volume can be minimized, and to allow a sufficiently large number of
             replicates (especially critical for geotextiles).

       *     The test should not measure a  property or simulate an exposure that would
             be completely uncharacteristic of the  proposed design function.


       Chemical compatibility testing as outlined in Method 9090 has been applied to the
evaluation of geotextiles, geonet and pipe in  practice. The introduction of a standard
chemical compatibility test method expanded  to include these geosynthetic products will be
welcomed by participants in the commercial geosynthetic testing market, as well as by
designers and users of the geosynthetics.

       Each laboratory surveyed reported that, in practice, standard index and  performance
test methods are modified to meet the special problems associated with chemical
compatibility testing.  A summary of test method modifications reported is given below for
each class of geosynthetic in this study.

       As noted in the previous section, varying numbers  of replicates are tested for
determination of index properties. For geotextiles, two of five  laboratories test five
samples; one tests seven, then rejecting two which generate data outside a reasonable
standard deviation, and two test ten replicates per determination.  One laboratory
recommends evaluating permittivity of geotextiles using common samples, i.e.  samples that
are not discarded after testing, but re-immersed and retested at  each succeeding interval.

       For chemical compatibility testing there  is general  agreement to test in  only one
direction (usually the machine direction) for those index methods which are
direction-oriented.  One laboratory recommends  use of a modified grab test substituting a
2-inch-wide specimen for the standard 4-inch grab sample.

       Monitoring weight change as a function of exposure, as required for liners in
Method 9090, is difficult for geotextiles  since it would require  that each geotextile
specimen be fully  dried prior to weighing and subsequent destructive testing.   Some labs
have attempted to  get around this problem by including specimens specifically assigned for
weight and dimensional determinations which are not destructively tested.  However,
problems associated with drying of geotextiles have been difficult to resolve due to the
physical nature of the products.

       Four of six labs responding reported using a non-standard strip  tensile test for
evaluating geonets during chemical  compatibility studies.   Strip width varies from two to
three inches; two laboratories test across three parallel strands.  One laboratory applies the
tensile grab test standard, ASTM D 1682(4),  for geotextiles to geonet.   In general, only

                                           592

-------
one of the two possible directions is tested.  The strip tensile test is preferred to standard
methods such as wide width tensile strength  because of sample size.  One laboratory uses a
quality control specification developed by a geonet manufacturer  for strip tensile testing of
net in one direction.  No  ASTM standard exists for narrow strip  tensile testing of geonet.

       One lab applies the ASTM D 751(5)  ball burst method (using a 1-inch ball) for
evaluation of geonet in chemical compatibility studies.  One laboratory uses a special
hydraulic transmissivity test device which accepts 4-inch square geonet specimens  (a more
generally accepted dimension for hydraulic transmissivity determinations is 12-inch square,
although the  4-inch size is compliant  with the ASTM D 4617<6), the standard for
transmissivity).

       The most common test applied to pipe in waste containment applications is pipe
stiffness: ASTM D 2412(7). However, the standard test requires long and unwieldy section
of pipe.  The laboratories surveyed have used 3- to 5-inch long cut sections of  pipe for
chemical compatibility testing.  Even  with shorter specimens, pipe specimens are bulky and
require  large volumes of leachate for  immersion; the number of replicates  which can be
immersed is limited to two to four, depending on pipe dimensions and internal procedures.

       Two laboratories have reported using  tensile bars cut from walls of solid pipe for
determining tensile strength as a function of  exposure according to ASTM D 638(8) or a
similar method. This approach is not  suitable for corrugated or slotted pipe because of the
uneven  wall profile.

       This section has provided an overview of how chemical compatibility testing as
outlined in Method 9090  has been applied to the evaluation of geotextiles, geonet and pipe
in practice. It can be seen that many widely varying adaptations arc currently in  use.  The
introduction of a standard chemical compatibility test method expanded to include these
geosynthetic products will be welcomed by participants in  the commercial geosynthetic
testing market, as  well as by designers and users of the geosynthetics.

EVALUATION AND INTERPRETATION OF CHEMICAL COMPATIBILITY  TEST
  DATA

       Current practices used within the geosynthetics industry to determine chemical
compatibility of geosynthetics reveal inconsistencies and inadequacies for  testing geotextiles,
geonets and pipes.  The problem facing the design and regulatory community is to find a
way to assess the data produced from these testing programs. Evaluation  of useful lifetime
through accelerated aging tests (accelerated life testing-ALT) has been sorely neglected due
to the complexity  of data interpretation and the  lack of understanding  of these tests as
applied to geosynthetic chemical compatibility.   The interaction of geosynthetic  polymer
and fibers with chemical wastes may  result in a change in physical properties of the
geosynthetics, either from chemical attack or from a solvation process caused by mutual
solubility of the waste material with the geosynthetic.  However, the designer need only be
concerned if the design function  of the geosynthetic can no longer be  fulfilled after
exposure to the waste or leachate.

       Earlier in  this paper, it can be  seen that widely varying opinions were received.

                                           593

-------
This is a complex issue, and there is probably some merit to each of the arguments
presented.  The SWTSU survey team concluded that to interpret the significance of data
does in fact require an informed analysis of the compatibility results in light of
considerations specific to the site, facility design, and synthetic and natural materials
proposed.  The compatibility results  themselves should be considered together as a  whole
and a decision to accept or reject the material on one physical  test value  at one isolated
time interval or temperature should be avoided.  That is to say, observed degradation
should be consistent across  several measured properties or test  values if it is to be
considered significant. Statistical analysis of tabulated compatibility data  may be
appropriate in many cases.

RECOMMENDATIONS

       The  specific methods used to evaluate these geosynthetic products vary between the
various laboratories.  Standards  for chemical compatibility testing of geosynthetic products
should be within  the framework of consensus standards organizations such as ASTM
(particularly Committee D-35),  and that regulators cite these industry standards when they
become available. Also, analytical techniques such as thermal  analysis and  spectroscopic
methods should be performed before and after a chemical compatibility test.  There are two
reasons for doing this:  (1)  to provide additional data to assess any effect of leachate
immersion on the base polymer, and (2) to provide a fingerprint for future assurance that
the correct material was  installed.

       Testing should provide valuable data not within the scope of existing standard test
methods.  Also, using appropriate modifications  to existing test methods  and non-standard
methods greatly reduces  the volume  of leachate  required for chemical compatibility testing,
thus reducing the cost burden imposed on the facility designer.  In addition,  industry
should develop standard  methods designed to meet the needs for index tests which  provide
full physical  characterization of these geosynthetics using small samples.  The evaluation of
chemical compatibility data is an involved process that requires consideration of details
specific to the site, facility design, waste components and the materials tested.

                                       PHASE H
       The second phase of this research involves investigating and recommending a testing
protocol for geosynthetic chemical compatibility testing.  Although this investigation  is
only in the planning stage, the section below gives a summary of the questions and
concerns being addressed by this plan.

       These questions and concerns have been grouped into three categories. The first
category  is that of field performance versus laboratory or  index tests. There are two major
concerns  here.  One, what  amount of failure in an index test constitutes a field failure and
two, which index test i s the most effective for each type of geosynthetic material.  The
next category involves accelerated aging.  Test conditions must be developed so that aging
or degradation can be observed in a short period of time.  Subsequently, some means of
extrapolation must be applied to provide a prediction of the service life-time of the
geosynthetic under normal  conditions.  The last category is to establish a correlation

                                          584

-------
between data obtained about the molecular interactions and the properties of the bulk
material.  Such a correlation would help in several areas including QA/QC, early detection
of degradation and provide a more complete understanding of degradation process.

                              ACKNOWLEDGEMENTS
       The authors appreciate the support for this project by the USEPA, under the
direction of Robert E. Landreth, Risk Reduction Engineering Laboratory, Office of
Research and Development, Grant # CR  815495. Special thanks are given to Deborah
Koeck, Keith Brewer, Kurt French, Kristine Ludwig and Matt Adams for technical
assistance.

                                   REFERENCES
1.  Tisinger, L. G.,  "Microstructural Analysis of the Durability of a Polypropylene
    Geotextile," Geosynthetics '89 Conference Proceedings, San Diego, CA, February
    21-23, pp. 513-524.

2.  Algers, M. S., "Polymer Science Dictionary," Elsevier Science Publishers Ltd., Essex,
    England,  1989, p. 474.

3.  USEPA, "Method 9090: Compatability Tests for Waste  and Membrane Liners,"
    EPA SW-846, Test Methods for Evaluating Solid Waste, USEPA, Washington, D. C.,
    1986.

4.  Standard Test Method for Breaking Load and Elongation of Textile Fabrics.
    ASTM D 1682-64, ASTM, Philadelphia, PA, 1964.

5.  Standard Test Method for Coated Fabrics.  ASTM D 751-79, ASTM,  Philadelphia, PA,
    1979.

6.  Standard Test Method for Constant Head Hydraulic Transmissivity (In-Plane  Flow) of
    Geotextiles and  Geotextile Related Products. ASTM D 4716-87, ASTM, Philadelphia,
    PA, 1987.

7.  Standard Test Method for Determination of External Loading Characteristics  of Plastic
    Pipe by Parallel-Plate Loading. ASTM D 2412-87,  ASTM, Philadelphia, PA, 1987.

8.  Standard Test Method for Tensile Properties of Plastics. ASTM D 638-82, ASTM,
    Philadelphia, PA, 1982.
                                          595

-------
                      STABILIZATION/SOLIDIFICATION FOR
                        TREATMENT OF SUPERFUND SOILS
                  by:  Franklin R. Alvarez and Richard P. Lauch
                       Risk Reduction Engineering Laboratory
                       U.S. Environmental Protection Agency
                       Cincinnati,, Ohio  45268

                       Michael M. Arozarena
                       PEI Associates, Inc.
                       Cincinnati, Ohio  45246

                       Marshall W. Allen
                       International Technology Corporation
                       Knoxville, Tennessee  37923
                                  ABSTRACT


     The overall objective of this research was to determine the effective-
ness of the stabilization/solidification process in reducing the mobility (as
measured by the Toxicity Characteristic Leaching Procedure) of metals, pesti-
cides, and polychlorinated dibenzofurans in actual Superfund soils.  Another
objective of the study was to determine how well the stabilization/solidifi-
cation process performed on actual Superfund soil versus a synthetic soil
matrix.

     Superfund soils from the Tri-State Metals Plating site in Columbus,
Indiana, and the Syncon Resins Site in Kearney, New Jersey, were selected for
bench-scale stabilization/solidification testing at the International Tech-
nology Corporation (IT) Technology Development Laboratory in Knoxville,
Tennessee.  Screening tests were conducted to determine optimum water ratios
and binder ratios with three commonly used binders—port!and cement, lime/fly
ash, and kiln dust.  Water-to-total-solids ratios of 0.30 to 0.45 and binder-


                                      596

-------
to-soil ratios of 0.5, 0.75, and 1.0 were used for testing.  Samples prepared
in triplicate were allowed to cure under controlled conditions for 28 days
and then subjected to unconfined compressive strength (UCS) testing.  The
minimum binder-to-waste ratio for each binder type that exhibited a UCS value
greater than 50 psi was selected for chemical analysis.  The chemical anal-
yses were performed by IT Corporation and a contract laboratory program (CLP)
laboratory (Ecology and Environment, Inc. in Buffalo, New York).

     Analytical results are presented for the untreated and stabilized/solid-
ified Tri-State Metals Plating and Syncon Resins soil, volatile organic com-
pound (VOC) headspace analyses during sample curing, and the portland cement,
kiln dust, and lime/fly ash binders.  Both untreated and treated Tri-State
soils were subjected to total waste analysis (TWA) and toxicity character-
istic leaching procedure (TCLP) metals; total organic carbon (TOC); humic
acid content; grain size distribution; and cation exchange capacity.  Both
treated and untreated Syncon Resins soils were analyzed for TWA and TCLP
metals, pesticides, and polychlorinated dibenzofurans; TOC; humic acid con-
tent; grain size distribution; and cation exchange capacity.  Binders were
analyzed for TWA and TCLP metals.  Also reported are UCS values, bulk densi-
ties, and curing conditions.

     This paper also includes a comparison of results from test:; conducted
during Phase I, which involved stabilization/solidification of a synthetic
soil matrix, with results from testing conducted on actual Superfund site
soils (Phase II).  Differences in experimental procedures (e.g., blending,
curing, VOC headspace sampling), analytical methodology, and physical/chemical
results are discussed.

     This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer review and administrative review policies and ap-
proved for presentation and publication.

                               INTRODUCTION


     Section 3004 of the Resource Conservation and Recovery Act (RCRA) of
1976, as amended by the Hazardous and Solid Waste Amendments (HSWA), mandates
a ban on the land disposal of hazardous wastes unless they have been treated
to prevent the migration of hazardous constituents into the environment.  The
U.S. Environmental Protection Agency (EPA) must set land disposal restriction
(LDR) treatment standards, based on best demonstrated available technology
(BOAT) and performance levels, that substantially diminish waste toxicity or
greatly reduce the likelihood of constituent migration so as to minimize
short- and long-term threats to human health and the environment.

     Land disposal restrictions apply to soil and debris when they are con-
taminated with a restricted RCRA hazardous waste.  Because of the complex
nature of many soil and debris matrices, meeting treatment standards for
wastes mixed with soil and debris may be difficult.  Therefore, EPA is under-
taking a rulemaking that will set LDR treatment standards specificially for
soil and debris.
                                      597

-------
BACKGROUND

     Under Phase I of EPA's BOAT research program (conducted from April to
November 1987)» a surrogate soil containing a wide range of chemical contam-
inants typically occurring at Superfund sites was prepared for use in bench-
scale or pilot-scale performance evaluations of five available treatment
technologies:  1) physical treatment (soil washing), 2) chemical treatment
using a potassium polyethylene glycol reagent (KPE6), 3) thermal desorption,
4) stabilization/solidification, and 5) incineration.  Under Phase II of
EPA's BOAT research program (conducted from April 1988 to March 1990), con-
taminated soils from several Superfund sites were used to reevaluate four of
these treatment technologies at the bench scale:  soil washing, KPEG, low-tem-
perature thermal desorption, and stabilization/solidification.  Incineration
was not evaluated in Phase II because of the expense involved in conducting
pilot-scale evaluations and the large quantity of data available on incinera-
tion of hazardous waste.

     A previous EPA report (1) presented the performance evaluation results
obtained during Phase II for physical treatment (soil washing), chemical
treatment (KPEG), and thermal desorption, and compared results obtained
during Phases I and II.  This paper addresses the performance evaluation
results obtained during Phase II for stabilization/ solidification.  A com-
parison of Phase I and Phase II experimental results for stabilization/solidi-
fication is also provided.

PURPOSE AND OBJECTIVES

     The purpose of the paper is twofold:  1) to present the data generated
by the stabilization/solidification of Superfund soil from Syncon Resins near
Kearney, New Jersey, and Tri-State Metals Plating in Columbus, Indiana, and
2) to compare the results obtained on actual Superfund soil (Phase II) to the
results obtained on synthetic soil (Phase I).

     EPA specified several objectives in the performance of these stabiliza-
tion/solidification tests:

     1)   To calculate the percent reduction of Teachable metals, pesticides,
          herbicides, and furans.

     2)   To determine the concentration of metals, pesticides, herbicides,
          and furans in the stabilized/solidified soil TCLP extracts.

     3)   To estimate VOC emissions from the stabilization/solidification
          process.

     4)   If possible, to draw conclusions from a comparison of Phase I and
          Phase II stabilization/solidification data.
                                      598

-------
                           EXPERIMENTAL PROCEDURES

     The purpose of the Phase II  program was  to evaluate the effectiveness  of
the stabilization/solidification  treatment technology on two Superfund site
soils—the Tri-State Metals Plating site at Columbus, Indiana,  and the Syncon
Resins site at Kearney, New Jersey.  Three generic binding agents (portland
cement, kiln dust,  and a 1:1 mixture of lime/fly ash) were used to prepare
soil/binder blends, and the performance of each was evaluated with respect  to
unconfined compressive strength development and TCLP leaching of contaminants.
No proprietary chemicals were used.

     All physical and chemical  analyses were  performed by IT Corporation in
Knoxville, Tennessee, and Ecology and Environment, Inc., in Buffalo,  New
York.  Analytical procedures and  results of Quality Assurance testing are
included in the full project report (2).

DESCRIPTION OF BINDERS AND SOILS

     The portland cement binder was a standard Type I cement manufactured by
the Dixie Cement Co. of Knoxville, Tennessee.  The kiln dust used was Pozza-
lime Calcic Powder distributed by Mineral By-Products, Inc. of  Marietta,
Georgia.  The composition of the  kiln dust was specified as 25  to 45  percent
CaO, 15 to 30 percent CaC03 and lesser amounts of oxides of aluminum, iron,
silicon, and magnesium.  The fly  ash used was IFA Stabilizer, a power plant
fly ash distributed by Western Ash Co. of Phoenix, Arizona, and composed of
nearly equal amounts of Si02 and  CaO.  The fly ash was added to an equal
weight of reagent grade Ca(OH)2 powder (Mallinkrodt, LOT 4195 KCHX) and mixed
thoroughly.

     The Syncon Resins soil consists primarily of alluvial sand, silt, clay,
and detritus.  Its texture is heterogeneous and spongy, and it  has a  nonuni-
form liquid holding capacity.  This soil is contaminated with heavy metals,
pesticides, herbicides, and furans.

     The Tri-State Metals Plating soil is the product of loamy  glacial out-
wash and calcareous sand and gravel.  Its clay-like texture is  interspersed
with granular particulate of greatly varying  sizes.  This soil  is contam-
inated with heavy metals.  The physical and chemical data on these soils are
presented in Tables 1, 2, and 3.

                                   RESULTS


     Tables 4 and 5 summarize the TCLP data obtained for untreated and treated
Tri-State Metals Plating and Syncon Resins soils.  The percent  reduction was
calculated for each pollutant using the following equation:

                         (1 -   ) x 100 = c
                                      599

-------
            TABLE 1.  SYNCON RESINS SOILS
Parameter
*
Volatiles, pg/kg
Methyl ene chloride
Toluene *
Semivolatiles, wg/kg
Pyrene
Bis(2-ethylhexyl )phthalate
Pesticides/herbicides/PCBs, pg/kg
Gamma-BHC (lindane)
4,4'-DDE
4 ,4 '-DDT
2,4-D
Dioxins/furans, pg/kg
TCDD
PeCDD
HxCDD
TCDF
PeCDF
HxCDF
Metals, mg/kg
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Selenium
Silver
Thallium
Vanadium
Zinc
TOC, mg/kg
CN, mg/kg
Sulfide, mg/kg
pH, S.U.
Result
4.
69 BT
68 B

1,600
780

1,100
8,600..
88.0001"1"
280
c
ND (O.ll)9
ND (0.21)
ND (0.60)
12.9
18.4
12.7

<4.49
581
166
0.57
1.29
65.3
91.5
224
0.75
20.7
<0.18
0.74
<6.29
42.2
167
57,800
0.31
194
6.58
Only results greater than the quantitation limit (five
times the detection limit) are reported.

This flag is used whenever the analyte is found in the
blank as well as a sample.

Above quantitation limit, estimated value.

ND = not detected at the specified detection limit (meth-
od detection limit).

                          BOO

-------
        TABLE 2.  TRI-STATES METALS PLATING SOIL
Parameter
Result
Volatiles, yg/kg

  Methylene chloride
  Toluene

Semivolatiles, pg/kg

Pesticides/herbicides/PCBs (yg/kg)

  Aroclor-1016

Dioxins/furans, pg/kg
  TCDD
  PeCDD
  HxCDD
  TCDF
  PeCDF
  HxCDF

Metals, mg/kg
   14 B
   27
     tt
   ND
  550
ND (0.19)
ND (0.13)
ND (0.16)
ND (0.12)
ND (0.12)
ND (0.17)
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Selenium
Silver
Thallium
Vanadium
Zinc
TOC, mg/kg
CN, mg/kg
Sulfide, mg/kg
pH, S.U.
<3.01
8.12
44.4
<0.21
49.9
506
99.2
267
<0.12
159
<0.17
<0.37
<6.07
12.8
79.0
9430
67.8
72
7.20
* Only results greater than the quantisation limit (five
  times the detection limit) are reported.
  This flag is used whenever the analyte is found in the
  blank as well as a sample.
t ND = not detected at specified detection limit (method
  detection limit).
                            601

-------
        TABLE 3.  GRAIN SIZE DISTRIBUTION OF CANDIDATE TEST SOILS
                              (tot. percent)
Site soil
*
Syncon Resins

Tri -State Metals1"

Coarse sand
(>0.5 mm)

47.8
29.8
27.3
26,0
Fine sand
(0.05-0.5 mm)

29.1
47.0
36.3
36.2
Silt
(0.002-0.05 mm)

15.1
15.0
24.9
26.6
Clay
(<0.002 mm)

8.0
8.2
11.5
11.2
The soil is primarily alluvial sand, silt, clays, and detritus (landfill
soil).

The soils at this site are developed over loamy glacial outwash and
calcareous sand and gravel.
                                    602

-------
              TABLE 4.   CONCENTRATIONS OF METALS IN TCLP LEACHATES OF TREATED AND UNTREATgD ALIQUOTS
                      OF TRI-STATE METALS SOIL AND CORRESPONDING PERCENT REDUCTION VALUES
Analyte
Antimony
Arsenic
Barium
Beryl 1 i urn
Cadmi urn
o» Chromium
CD
*** Copper
Lead
Mercury
Nickel
Selenium
Silver
Thallium
Vanadium
Zinc
Untreated
soil
Cone., mg/L
<0,060
<0.005
1.016
<0.002
1.273
0.687
0.063
0.100
<0.0002
0.512
<0.005
<0.010
<0.005
<0.010
0.337

Portl and
Cone., mg/L
<0.060
<0.005
0.84b
<0.002
<0.005
0.267
0.080
0.036
<0.0002
0.045
<0.005
<0.010
<0.005
<0.010
0.036

cement
% reduction
NA1"
NA
16.8
NA
>99.6
61.1
0
64.0
NA
91.2
NA
NA
NA
NA
89.3
Treated
Lime/fly
Cone., mg/L %
<0.060
<0.005
2.24
99.5
82.7
0
54.0
NA
95.7
NA
NA
NA
NA
89.6

Kiln
Cone., mg/L
<0.060
<0.005
0.931
<0.002
<0.005
0.354
0.081
0.252
<0.0002
0.044
<0.005
<0.010
<0.005
<0.010
0.033

dust
% reduction
NA
NA
8.4
NA
>99.6
48.5
0
0
NA
91.4
NA
NA
NA
NA
90.2
* Data presented and percent reduction  calculated  are based on the average concentration of three replicate
  samples.

  Not applicable.  Percent reduction  cannot  be calculated when analyte is not detected  in untreated soil.

-------
    TABLE 5.   CONCENTRATIONS  OF METALS AND ORGANICS  IN TCLP OF LEACHATES TREATED AND UNTREATED ALIQUOTS
                     OF SYNCON RESINS SOIL AND CORRESPONDING PERCENT REDUCTION VALUES*
Analyte
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Selenium
Silver
Thai 1 i urn
Vanadium
Zinc
Untreated
soil
Cone., mg/L
<0.060
1.007
1.007
<0.002
0.006
0.014
0.116
0.092
<0.0002
0.052
<0.005
<0.010
<0.005
0.018
1.07
Treated soil
Portland
Cone., mg/L
<0.060
0.033
1.847
<0.002
<0.005
0.016
0.194
<0.008
<0.0002
0.045
<0.005
<0.010
<0.005
<0.010
0.061
cement
% reduction
NA1"
96.7
0
NA
>16.7
0
0
>91.3
NA
13.5
NA
NA
NA
>44.4
94.3
Lime/fly
Cone., mg/L %
<0.060
0.019
1.107
<0.002
<0.005
0.025
0.378
0.071
<0.0002
0.025
<0.005
<0.010
<0.005
<0.010
0.020
ash
reduction
NA
98.1
0
NA
>16.7
0
0
22.8
NA
51.9
NA
NA
NA
>44.4
98.1
Kiln
Cone., mg/L
<0.060
0.016
1.39
<0.002
<0.005
0.010
0.258
0.101
<0.0002
0.022
<0.005
<0.010
<0.005
<0.010
0.043
dust
% reduction
NA
98.4
0
NA
>16.7
28.6
0
0
NA
57.7
NA
NA
NA
>44.4
96.0
(continued)

-------
TABLE 5 (continued)


Analyte
Lindane,
yg/liter
DDT,
pg/ liter
DDE,
pg/ liter
DDD,
yg/ liter
PCDF,
ng/ liter
Untreated
soil
Cone., mg/L
<0.5

<2^

<1.07

7.8

<4.84ft

Treated soil

Portland
Cone., mg/L
<0.2i

<0.40

<0.50

0.38

<6.46ft


cement
% reduction
NA

NA

NA

95.1

NA


Lime/fly
Cone., mg/L %
<0.12

<0.23

<0.23

<0.23

<15.7ft


ash
reduction
NA

NA

NA

97.0

NA


Kiln
Cone., mg/L
<0.10

<0.20

<0.20

<0.55

<5.32ft


dust
% reduction
NA

NA

NA

92.9

NA

 * Data presented and reduction efficiencies  calculated  are  based on the average concentration of three
   replicate samples.
   Not applicable.  Cannot calculate percent  reduction when  analyte is not detected in untreated soil.
   Average concentration of four replicate samples.

-------
where

     a = pollutant concentration (e.g., mg/L) in TCLP extract of treated
         sample
     b = pollutant concentration (e.g., mg/L) in TCLP extract of
         untreated sample
     c = percent removal

     The "percent reduction" value indicates the stabilization efficiency;
i.e., how effectively the stabilization treatment sequestered or bound the
pollutant, thereby yielding a TCLP leachate with a lowered concentration of
the pollutant.  Note that no correction was made for dilution by addition of
binder and water.

     The Tri-State Metals Plating soil  had detectable concentrations of 7 of
the 15 metals analyzed for (barium, cadmium, copper, lead, nickel, chromium,
and zinc).  Reduction efficiency cannot be calculated for any analyte not
detected in the untreated TCLP extract; therefore, reduction efficiencies are
reported only for these seven metals.  Stabilization/solidification was
effective on cadmium (greater than 99.5 percent reduction), nickel (greater
than 91.2% reduction), and zinc (greater than 89.3% reduction).  The data
indicate that reduction efficiencies did not vary significantly with the
binders used on these three metals.  Reduction efficiencies reported for
chromium and lead varied among binders.  Lime/fly ash was most effective in
stabilizing chromium (82.75$), followed by port!and cement (61.1%), and kiln
dust (48.5%).  Portland cement was most effective in stabilizing lead
(64.0%), followed by lime/fly ash with 54.0 percent; no reduction was
experienced with the kiln dust binder.   All three binders were ineffective  in
stabilizing barium and copper.  Actually, the binders themselves contribute a
significant concentration of barium to the extracts.  Also, compared with the
total metal analyses shown in Table 2,  only very small quantities of metals
are leaching from the untreated soil.

     The Syncon Resins soil had concentrations of 9 of the 15 metals analyzed
for (arsenic, barium, cadmium, chromium, copper, lead, nickel, vanadium, and
zinc).  The untreated TCLP extract concentrations of cadmium, chromium, and
vanadium were so low that conclusions cannot be drawn from the available data
as to the effectiveness of the stabilization/ solidification process.  Stabili-
zation/solidification was effective, however, on arsenic (greater than 96.7%)
and zinc (greater than 94.3%).  The data indicate that reduction efficiencies
did not vary significantly among the binders on these metals.  Reduction
efficiencies reported for lead and nickel did vary among binders.  Portland
cement was most effective in stabilizing lead (greater than 91.3%), followed
by lime/fly ash (22.8%); kiln dust showed no reduction.  Kiln dust was the
most effective in stabilizing nickel (57.7%), followed by lime/fly ash (51.9%);
Portland cement demonstrated only 13.5 percent reduction.  Ail three binders
were ineffective in stabilizing barium and copper.  Again, all three binders
actually contribute a significant concentration of barium to the extracts.
Also, compared with the total metal concentation present (as shown in Table
1), only very small quantities of metals were leaching from the untreated
soil.
                                     60S

-------
     Four pesticides were analyzed for in the untreated and treated Syncon
Resins TCLP extracts:  lindane (gamma-BHC), DDT, DDE, and ODD.  Because
lindane was not detected in the untreated TCLP extract, reduction efficiencies
could not be calculated.  In the case of DDT and DDE, mixed data were ob-
tained on the untreated TCLP extracts, which makes it difficult to draw any
conclusions as to the effectiveness of the stabilization/solidification
process.  Triplicate samples of the untreated soil for DDT and DDE showed
nondetected, estimated concentrations below the detection limit, and low
detected concentrations.  Based on these results, reduction efficiencies
cannot be calculated for DDT and DDE.  Reduction efficiencies reported for
ODD were 97.0, 95.1, and 92.9 percent for lime/fly ash, portland cement, and
kiln dust, respectively.

     The TCLP analyses for four replicate untreated Syncon Resins soil indi-
cated that furans (PCDF) were not detected in three of the four samples and
detected at 10.7 ng/L in the fourth sample.  As was the case with several of
the metals, lindane, DDT, and DDE, the reduction efficiency of a pollutant
that is not detected in the untreated sample TCLP extract cannot be calculated.

        COMPARISON OF SARM AND SUPERFUND SOIL TCLP ANALYTICAL RESULTS


     A second objective of this project was to compare analytical results
obtained during Phase I on synthetic soil with results obtained during Phase
II on Superfund soil.  To arrive at a valid comparison of these results, one
must review data on similarly contaminated soils.  A review of the analytical
data on the SARM soil and the soil from Tri-State Metals Plating and Syncon
Resins indicated that the SARM II (Table 6) and the Syncon Resiins soils
(Table 1) were the most similar for purposes of a performance comparison.

     Tables 7, 8, and 9 present TCLP analytical results for the untreated and
treated SARM II and Syncon Resins soils with the binders portland cement,
lime/fly ash, and kiln dust, respectively.  These tables also present reduction
efficiencies to permit a comparison of the effectiveness of stabilization/soli-
dification on SARM II and Syncon Resins soil for specific analytes and binders.
The TCLP extract analytes in common in both the SARM II soil and the Syncon
Resins soil were arsenic, cadmium, chromium, copper, lead, nickel, and zinc.
As shown in Table 7, with portland cement as the binder, similar reduction
efficiencies were obtained for lead (78.6% vs. 91.2%) and zinc (96.3% vs.
94.3%); however, very dissimilar reduction efficiencies were obtained for
copper (93.2% vs. 0%) and nickel (90.0% vs. 12.6%).  Arsenic and chromium
reduction efficiencies could not be compared because these metals were not
detected in the untreated SARM II soil, and cadmium reduction efficiencies
could not be compared because of the low concentration in the Syncon Resins
untreated soil.

     Results obtained with lime/fly ash (Table 8) indicate that reduction
efficiencies for zinc (99.9% vs. 98.1%) compare well; however, results for
other analytes such as copper (96.6% vs. 0%), lead (>78.6% vs. 22.6%), and
nickel (>90.0% vs. 51.7%) were not comparable.  As with portland cement,
arsenic, cadmium, and chromium reduction efficiencies could not be compared.
                                      607

-------
     Kiln dust reduction efficiencies (Table 9) demonstrate that results ob-
tained for zinc compare well (94.6% vs. 96.0%); however, reduction efficiencies
for copper (89.9% vs. 0%), lead (47.1% vs. 0%), and nickel (>90.0% vs. 57.2%)
were very dissimilar.  As with portland cement and lime/fly ash, reduction
efficiencies for arsenic, cadmium, and chromium could not be compared.
         TABLE 6.  RESULTS OF TOTAL WASTE ANALYSIS FOR SARM SAMPLES
Analyte
    SARM I
High organic,
  low metal
              SARM 2
           Low organic,
            low metal
              SARM 3
           Low organic,
            high metal
          SARM 4
       High organic,
        high metal
Volatiles, pg/kg

  Acetone
  Chlorobenzene
  1,2-dichloroethane
  Ethyl benzene
  Styrene
  Tetrachloroethylene
  Xylene

Semivolatiles, pg/kg

  Anthracene
  Bis(2-ethylhexyl)
   phthalate
  Pentachlorophenol

Inorganics, mg/kg
    3,150
      330
      380
      350
      000
      710
    4,150
3,
1,
      940

      600
      135
230
  9.2
  3.9
 74
 26
 16
210
               275

                34
                62
220
  8.9
  3.1
100
 24
 13
150
                  265

                  140
                   15
13,000
   270
   830
 2,500
   540
   540
 3,700
              775

              500
               78
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Zinc
18
17
27
193
190
27
392
18
23
37
260
240
32
544
904
1,280
1,190
9,650
15,200
1,140
53,400
810
1,430
1,650
13,300
16,900
1,380
28,900
  Source:  Weitzman, L., and L. E. Hamel.   Evaluation of Solidification/
  Stabilization as a BOAT for Superfund Soils.  Acurex Corporation.  EPA
  Contract No. 68-03-3241, September 1988.
                                     608

-------
  TABLE 7.   COMPARISON OF THE  EFFECTIVENESS  OF  PORTLAND  CEMENT AS  BINDER  IN
                   SARM II SOIL  VERSUS  SYNCON RESINS  SOIL
                        (TCLP  data  reported  in  mg/L)
Analyte
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Zinc

Untreated
ND* (0.15)
0.73
ND (0.01)
0.89
0.70
0.40
14.6
SARM II
Treated
B/S = 0.7
ND (0.15)
ND (0.01)
0.03
0.06
0.15
0.04
0.54
Syncon Resins
%
reduc-
tion
NAf
>98.6
NA
93.2
78.6
90.0
96.3
Untreated
1.007
0.0062
0.0143
0.116
0.092
0.0516
1.070
Treated
B/S - 0.5
0.0331
ND (0.005)
0.0156
0.194
0.0081
0.0451
0.0614
%
reduc-
tion
96.7
>19.4
0
0
91.2
12.6
94.3
  Not detected at concentration indicated in  parentheses.
  Not applicable when analyte is not detected in  untreated material.
   TABLE 8.   COMPARISON OF THE EFFECTIVENESS  OF  LIME/FLY  ASH  AS  BINDER  IN
                   SARM II SOIL VERSUS SYNCON RESINS  SOIL
                        (TCLP data reported in mg/L)
Analyte
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Zinc

Untreated
ND* (0.15)
0.73
ND (0.01)
0.89
0.70
0.40
14.6
SARM II
Treated
B/S = 2.0
ND (0.15)
ND (0.01)
ND (0.01)
0.03
ND (0.15)
ND (0.04)
0.02
Syncon Resins
%
reduc-
tion
NAf
>98.6
NA
96.6
>78,6
>90.0
99.9
Untreated
1.007
0.0062
0.0143
0.116
0.092
0.0516
1.070
Treated
B/S « 0.5
0.0186
ND (0.005)
0.0246
0.378
0.0712
0.0249
0.0198
%
reduc-
tion
98.2
>19.4
0
0
22.6
51.7
98.1
  Not detected at concentration indicated in parentheses.
* Not applicable when analyte is not detected in untreated material.
                                     609

-------
         TABLE 9.  COMPARISON OF THE EFFECTIVENESS OF KILN DUST
           AS BINDER IN SARM II SOIL VERSUS SYNCON RESINS SOIL
                      (TCLP data reported in tng/L)
Analyte
Arsenic
Cadnri um
Chromium
Copper
Lead
Nickel
Zinc

Untreated
ND* (0.15)
0.73
ND (0.01)
0.89
0.70
0.40
14.6
SARM II
Treated
B/S = 1.0
ND (0.15)
ND (0.01)
0.05
0.09
0.37
ND (0.04)
0.78
Syncon Resins
%
reduc-
tion
NAf
>98.6
NA
89.9
47.1
>90.0
94.6
Untreated
1.007
0.0062
0.0143
0.116
0.092
0.0516
1.070
Treated
B/S = 0.75
0.0165
ND (0.005)
0.0105
0.258
0.101
0.0221
0.0431
X
reduc-
tion
98.4
>19.4
26.6
0
0
57.2
96.0
Not detected at concentration indicated in parentheses.

Not applicable when analyte is not detected in untreated material.
                                    610

-------
                                 CONCLUSIONS


     The results of this study indicate that the stabilization/solidification
process can effectively stabilize Superfund soils containing arsenic, cadmium,
nickel, zinc, and the pesticide ODD when Portland cement,  lime/fly ash,  or
kiln dust is used as the binder.  Stabilization efficiencies for these metals
ranged from >89.3 percent for zinc to >99.5 percent for cadmium.  The stabil-
ization efficiency obtained for ODD was 97.0 percent with  lime/fly ash as the
binder.  A comparison of total waste analyses with TCLP for untreated soil
shows that only a small portion of the total contaminants  leaches from the
soil.

     The binder selection had a great impact on the stabilization efficien-
cies of some metals, e.g., chromium, lead, and nickel.   Lime/fly ash proved
to be the most effective binder for stabilizing chromium in the Tri-State
Metals soil, with an 82.7 percent reduction efficiency, whereas portland
cement and kiln dust stabilized only 61.1 and 48.5 percent, respectively.
Portland cement was the most effective binder for treating lead (64.0%
removal), followed by lime/fly ash (54.0%).  No stabilization was experienced
with kiln dust.  Tri-State Metals Plating soil is described as a loamy,
glacial till outwash soil over sand and gravel.

     Portland cement proved to be the most effective binder for stabilizing
lead in the Syncon Resins soil with a stabilization efficiency of greater
than 91.3 percent; lime/fly ash and kiln dust were ineffective.  Kiln dust,
however, was the most effective binder for stabilizing  nickel in the Syncon
Resins soil, followed closely by lime/fly ash; stabilization efficiencies
were 57.7 and 51.9 percent, respectively.  Portland cement was ineffective in
stabilizing nickel in the Syncon Resins soil.  Syncon Resins soil is described
as unconsolidated, consisting primarily of alluvial sand,  silt, clay, and
detritus.

     Results on both the Syncon Resins and Tri-State Metals Plating soils
indicate that all three binders were ineffective in stabilizing copper.

     It was impossible to report stabilization efficiencies for a number of
metals, pesticides, and polychlorinated dibenzofurans analyzed under this
study because the untreated TCLP extract concentrations were below detectable
limits.

     From an operational standpoint, several conclusions can be drawn.
Overall, the portland cement binder performed the best  by yielding a soil/
binder product exceeding 250 psi, which is the upper limit of measurement
with the available testing equipment.  This was true for both Syncon Resins
and Tri-State Metals Plating soils and for all binder-to-soil ratios (0.5,
0.75, and 1.0).  The lime/fly ash binder performed almost as well on the
Tri-State soil.  On the Syncon Resins soil, this binder produced a nearly
linear rise in compressive strength from 50 to 250 psi  with increasing
                                     611

-------
binder-to-soil ratios.  The kiln dust binder performed the least well5 the
unconfined compressive strength of the highest binder-to-soil ratio barely
met the minimum acceptable level of 50 psi.
     The following are some additional specific conclusions drawn from the
use of stabilization/solidification for treatment of Syncon Resins and Tri-
State Metals Plating soils:
Syncon Resins Soil—
     0  Portland cement binder
          1)  Uniform blends with good compressive strength were achieved at
              the lowest binder-to-soil (B/S) ratios,
          2)  Desired product was achieved throughout the 0.3 to 0.4 water-
              to-total solids (W/TS) ratios, although the 0.4 blend had free-
              standing liquid at the 0.5 B/S ratio.
          3)  Use of a 0.3 W/TS ratio is recommended to ensure cohesive
              mixing of soil and to prevent solids from settling.
     0  Kiln dust
          1)  Textured blends were achieved with moderate, slow-developing,
              compressive strength.
          2)  Desired product was achieved at 0.38 to 0.4 W/TS ratio, but
              compressive strength did not increase significantly with
              increasing B/S ratios.
          3)  Use of a 0.4 W/TS ratio is recommended to enhance blend homo-
              geneity.
     0  Lime/fly ash mixture
          1}  Problems were encountered in achieving uniform blends; thorough
              mixing was not always effective.
          2)  Acceptable product was achieved at 0.4 to 0.48 W/TS ratios.
          3)  Compressive strength was achieved faster at the higher B/S
              ratios.
          4)  Use of a 0.45 W/TS ratio is recommended to enhance the
          thoroughness of mixing.
                                     812

-------
Tri-State Metals Plating—
     0  Portland cement

          1}  Uniform hydrated appearance was achieved at a W/TS ratio as low
              as 0.2.

          2)  Desired product was achieved throughout the 0.2 to 0.4-W/TS
              range.

          3}  All blends achieved greater than 50 psi eompressive strength
              after 24 hours.

          4)  Use of a 0.3 W/TS ratio is recommended to enhance mixing.

     0  Kiln dust

          1)  Thoroughness of blending is the key factor in achieving uniform
              blend product.

          2)  Compressive strength of 50 psi was achieved in 0.2 to 0.4 W/TS
              range.

          3)  Use of a 0.4 W/TS ratio is recommended to ensure thorough
              blending.

     0  Lime/fly ash mixture

          1)  Blends achieved uniform silty, monolithic product.

          2)  Desired product was achieved with 0.2 to 0.4 W/TS blends.

          3)  All blends achieved greater than 50 psi eompressive strength
              after 48 hours.

          4)  Use of a 0.4 W/TS ratio is recommended to enhance thoroughness
              of mixing.

     For a comparison of pollutant stabilization efficiencies obtained with
the SARM soil and those obtained with actual Superfund soil, the selected
SARM and Superfund soils had to have a similar chemical and physical composi-
tion.  A review of the available data indicated that SARM II (low metals, low
organics) and Syncon Resins soils were most similar.  The stabilization tests
indicated that the SARM II soil and Syncon Resins soil exhibited similar
stabilization efficiencies for zinc regardless of which binder was used.
Stabilization efficiencies for lead were similar when Portland cement was
used as the binder.  The stabilization efficiencies for other metals (e.g.,
copper and nickel) were very dissimilar when portland cement, lime/fly ash,
and kiln dust were used as binders.  Stabilization efficiencies, for lead were
also dissimilar when lime/fly ash and kiln dust were used.
                                     613

-------
     It is apparent from the data presented in this study that the stabiliza-
tion/solidification process could not be evaluated to its full potential.   A
number of analytes were not detected or were at too low a concentration for
stabilization efficiencies to be calculated.  In future treatability studies,
it is recommended that the investigator spike the untreated soil  with the
compounds of interest in concentrations that will test the limitation of the
technology as opposed to the limitation of the analytical methodology.

                                 REFERENCES


1.   PEI Associates, Inc., and IT Corporation.  Evaluation of Alternative
     Treatment Technologies for CERCLA Soils and Debris (Summary  of Phase  I
     and Phase II).  Prepared for the U.S. Environmental  Protection Agency,
     Risk Reduction Engineering Laboratory, Cincinnati, Ohio, under Contract
     No. 68-03-3389.  September 1989.

2.   PEI Associates, Inc.  Stabilization/Solidification Treatment of Super-
     fund Soils for RCRA Land Disposal Restrictions.  Prepared for the U.S.
     Environmental Protection Agency, Risk Reduction Engineering  Laboratory,
     Cincinnati, Ohio, under Contract No. 68-03-3389.  February 1990.
                                     614

-------
                     TREATABILITY OF TRIBUTYLTIN IN POTWs

                    by:  lRichard  A.  Dobbs,  lHenry H. Tabak,
                        2Rakesh Govind and 2Devendra Atnoor

             rRisk Reduction  Engineering  Laboratory, United States
             Environmental  Protection  Agency,  Cincinnati,  OH  45268

               Department of Chemical and Nuclear Engineering,
                University of Cincinnati, Cincinnati,  OH 45221


                                   ABSTRACT
      Organotin contaminated wastewater generated during drydock and ship
cleaning operations must be treated before discharge to the environment,  A
study was conducted to assess the treatability of tributyltin in a
conventional primary/activated sludge municipal wastewater treatment plant.
Experiments were conducted to evaluate the treatability and fate of
tributyltin by both abiotic and biotic mechanisms.

      Sorption isotherms were measured for tributyltin and its degradation
products on primary sludge, mixed-liquor solids, and digested sludge.
Partition coefficients ranged from 1,585 to 13,200 for monobutyltin and
tributyltin, respectively.  The organotin compounds were not removed by
stripping.

      Both aerobic and anaerobic biotic mechanisms were studied for
degradation of tributyltin.  Steady state aerobic degradation levels in the
range of 75-95% were observed with sludge retention times of 8 and 18 days.
In anaerobic studies, biological methane potential and anaerobic toxicity
assays showed no effect on the performance of methanogens.
                                    615

-------
                                  INTRODUCTION

      Tributyltin  (TBT)  is used as an antifouling toxicant in boat paints and
as wood preservative, molluscicide and  insecticide.   It  has become popular
because of its excellent antifouling properties, long life-time, and lack of
corrosion problems.   During the life cycle of paints TBT is slowly released to
the aqueous environment.  During drydock and ship cleaning activities
wastewater streams contaminated with TBT are produced.  Treatment is desirable
before the wastewaters are discharged to the environment.

      Physical-chemical  and biological treatment processes are available for
treatment of organotin-contaminated wastewaters.  In general, biological
processes are more economical and easier to apply.  However, since TBT and its
degradation products  are acutely toxic to many microorganisms, mollusks,
insects, and higher plants (1,2), treatability studies were undertaken to
assess the potential  of  publicly-owned treatment works  (POTWs) to remove
organotin compounds.  Treatment of TBT-contaminated wastewater at municipal
wastewater treatment  plants is advantageous for the following reasons: (I)
POTWs are readily available all over the country, (2) substantial dilution is
available which minimizes toxic impacts of TBT compounds, (3) a diverse
biomass is available which enhances the potential for degradation of toxic
compounds, and (4) nutrients and co-metabolites required for successful
operation of biological  processes are present.

      Studies in the  literature suggest that biodegradation of organotin
compounds occurs and  is  a major degradation pathway in aquatic and sedimentary
environments.  Degradation of TBT in soil has been reported (3).  The effects
of organotin-contaminated wastewater on the activated sludge process was
studied using continuous bench-scale reactors and fish bioassays (4).
Performance and stability of a municipal treatment facility treating organotin
concentrations up to  75 /ig/L demonstrated that the full-scale activated sludge
process was not adversely affected by the contaminant (5).  None of the
studies characterized the tributyltin fate or specification in the activated
sludge process.

      The present study was" conducted in three phases.  In the first phase
analytical methodology to recover TBT and related compounds from both sludge
and aqueous matrices was investigated.  The second phase assessed the role of
abiotic mechanisms (stripping and sorption) on removal of TBT.  The final
phase evaluated both  aerobic and anaerobic mechanisms of biodegradation.

                            ANALYTICAL METHODOLOGY

      Analytical methods used in the studies cited did not allow determination
of degradation and ultimate fate of organotin compounds during treatment by
the activated sludge  process.  Recent advances in analytical techniques have
made this possible.  The analytical method used in the present study was a
modification of techniques described by Matthias (6) and Muller (7).  The
basic methods described were based on simultaneous hybridication/extraction
with gas chromatography  flame photometric detection.  Application of the


                                     616

-------
method to raw wastewater or sludge samples resulted in severe emulsion
problems.  The modification used in the present study included a sonification
step on the combined extracts to break the solvent-water emulsion prior to
drying with sodium sulfate.  Sodium borohydride was used for hybridization and
methylene chloride was used as the extracting solvent,  Di-n-propyltin-
dichloride or diphenyltin dichloride was used as an internal standard.

                      EXPERIMENTAL  RESULTS  AND DISCUSSION

ABIOTIC STUDIES (STRIPPING)

      The major abiotic mechanisms for removal of most organic compounds in
conventional primary/activated sludge wastewater treatment plants are
volatilization (surface desorption) or air stripping and sorption on solids
and sludges.

      In order to assess the fate of a particular tributyltin derivation in
wastewater one must consider the dissociated active form, the TBT cation
(Bu3Sn+)  and  its major metabolites  presumably  formed  by  progressive
debutylation to inorganic tin (i.e., tributyltin (TBT) Bu3Sn+ ->  dibutyltin
(DBT)Bu2Sn2+ -> monobutyltin (MBT) BuSn3* -> inorganic tin  Sn4+).

      Although these cationic species are not expected to be affected by the
volatilization or air  stripping mechanisms, tests were conducted in bubble
columns to validate this assumption.  Two one-liter bubble columns (length to
diameter ratio = 12) similar to the design of Mackay (8) were acid washed,
rinsed thoroughly, and filled with a solution containing 0.5 mg/L of
tributyltin chloride,  dibutyltin dichloride, and monobutylin trichloride -
after 20 hours the columns were dumped and refilled with the same test
solution in an attempt to eliminate glass surface adsorption during the
experimental run.  Nitrogen from a cylinder was passed through a pressure
regulator, flow controller, gas saturation bottle and into the bottom of the
fritted bubble column.  Nitrogen was used as the stripping gas to avoid the
possibility of air oxidation.  Flow rate was measured at 70 ml/minute with a
soap bubble flow meter.  The column was maintained at 23*C with a water
jacket.  Stripping was continued over a 20-hour period.  Samples were
withdrawn from the column for analyses at 0, 4, 13, and 20 hours.  Analytical
data are summarized in Table 1.

                    TABLE  1.   STRIPPING  DATA FOR ORGANOTINS

 Time (hr)        TBT  (mg/L)        DBT (mg/L)        MBT (mg/L)
0
4
13
20
0.50
0.47
0.47
0.45
0.52
0.49
0.48
0.47
0.51
0.49
0.49
0.47
      TBT=tributyl tin   DBT=dibutyl tin    MBT=monobutyltin



                                     617

-------
      Stripping data  in Table  I have shown that TBT and its degradation
products are nonvolatile.  The observed small changes in concentration with
time are most likely  due to continued adsorption of the organotins on the
walls of the glass bubble column.

              ABIOTIC STUDIES (MEASUREMENT OF SORPTION ISOTHERMS)

      In dilute systems typical of most environmental situations, the
partition coefficient Kp for sorption on wastewater solids can be defined as
follows:
                                   Kp = CS/CL                      (1)

where: Cs - concentration of pollutant in the solid phase and CL =
concentration of pollutant in the liquid phase.

      The usual method for measuring partition coefficients is  to determine a
sorption isotherm.  Slurries of the wastewater solids are prepared in
different initial concentrations of the toxic organic chemical  being studied.
After the systems reach equilibrium, the concentration in both  the solution
and solid phases are  measured  (solid phase concentration can be calculated
based on the difference between the initial concentration added and the
solution concentration at equilibrium).  Data are fitted to the Freundlich
equation which can be written:
                               X/M  = KCen                           (2)

where:  X  « C0-Ce which  is  the amount  of solute  sorbed  from a
              given volume of solution
        C0 » initial  concentration of solute added
        Ce - concentration of solute at equilibrium
        M  - weight of solids added to the solution
  K and n  - empirical constants

Data are fitted to the logarithmic form of Equation (2) which can be written
as follows:
                          log X/M = log K + n log Ce               (3)

For dilute solutions  this equation yields a straight line with  a  slope of n
and an intercept equal to the value of K (X/M at Ce = 1.0) when X/M is plotted
as a function of Ce on logarithmic paper.  The intercept is an  indicator of
sorption capacity and the slope of sorption intensity.  K is often referred to
as the sorption or adsorption coefficient.  To obtain the partition
coefficient (K )  when X/M is in mg/gm and Ce is  in  mg/L,  the following
equation is used:
                                  Kp = X/M_
                                       Ce/1000                    (4)

(Equation (4) also applies when X/M is in #g/gm and Ce is in ng/L).

      A detailed experimental protocol for determination of the partition
coefficient or sorption capacity of wastewater solids for toxic organic
                                     618

-------
compounds has been described previously (9).  The moisture content of the
wastewater solids was determined by drying a portion to constant weight at
110°C.  A stock slurry that contained distilled water and wastewater solids
was prepared to give the desired concentration of solids for the isotherm
test.  Sludge slurries were transferred to 500 ml bottles to yield 0.2 g dry
organic solids per liter.  Organotin compounds were dosed at :iO, 50, 200, 500,
and 1,000 /xg/L initial concentration for primary, mixed-liquor, and digested
sludge.  Bottles were capped and stirred for 18 hours.  The phases were
separated via centrifugation and aqueous and solid phases analyzed by the
analytical method described earlier.  Sorption data for TBT, [)BT, and MBT are
shown in Tables 2-4.
                  TABLE 2.  SORPTION DATA FOR TRIBUTYLTIN
Type of Sludge
Primary
sludge
Mixed
Liquor
Sludge
Digested
sludge
Jg/L
10.0
50.0
200.0
500.0
1000.0
10.0
50.0
200.0
500.0
1000.0
10.0
50.0
200.0
500.0
1000.0
c
JgA
3.0
13.0
71.0
106.0
398.0
5.0
21.0
54.0
120.0
487.0
3.0
20.0
39.0
98.0
313.0
X
W/L
7.0
37.0
129.0
394.0
602.0
5.0
29.0
146.0
380.0
513.0
7.0
30.0
161.0
402.0
687.0
X/M
M9/9
35.0
185.0
645.0
1970.0
3010.0
25.0
145.0
730.0
1900.0
2565.0
35.0
150.0
805.0
2010.0
3435.0
log Kp
4.07
4.15
3.96
4.27
3.88
3.70
3.84
4.13
4.20
3.72
4.07
3.88
4.31
4.31
4.04
Avg.log Kp
4.07
3.92
4.12
                                    619

-------
TABLE 3.  SORPTION DATA FOR DIBUTYLTIN
Type of SI
Primary
sludge
Hixed
Liquor
sludge
Digested
sludge

Type of
sludge
Primary
sludge
Hixed
Liquor
Sludge
Digested
sludge
C0
udge fig/I
10.0
50.0
200.0
500.0
1000.0
10.0
50.0
200.0
500.0
1000.0
10.0
50.0
200.0
500.0
1000.0
TABLE
m/i
10.0
50.0
200.0
500.0
1000.0
10.0
50.0
200.0
500.0
1000.0
10.0
50.0
200.0
500.0
1000.0
m/i
6.0
33.0
80.0
296.0
596.0
4.0
31.0
110.0
148.0
595.0
5.0
29.0
160.0
325.0
547.0
X
/*g/L
4.0
17.0
120.0
204.0
404.0
6.0
19.0
90.0
352.0
405.0
5.0
21.0
40.0
175.0
453.0
4. SORPTION DATA
m/i
7.0
30.0
161.0
360.0
709.0
6.0
33.0
77.0
320.0
793.0
8.0
32.0
173.0
361.0
726.0
X
ra/L
3.0
20.0
39.0
140.0
291.0
4.0
17.0
123.0
180.0
207.0
2.0
18.0
27.0
139.0
274.0
X/M
/*g/g
20.0
85.0
600.0
1020.0
2020.0
30.0
95.0
450.0
1760.0
2025.0
25.0
105.0
200.0
875.0
2265.0
log Kp Avg. log Kp
3.52
3.41
3.88 3.58
3.54
3.53
3.88
3.49
3.61 3.72
4.08
3.53
3.70
3.56
3.10 3.48
3.43
3.62
FOR MONOBUTYLTIN
X/M
pg/g
15.0
100.0
195.0
700.0
1455.0
20.0
85.0
615.0
900.0
1035.0
10.0
90.0
135.0
695.0
1370.0
log Kp Avg. log Kp
3.33
3.52
3^08 3.31
3.29
3.31
3.52
3.41
3.90 3.48
3.45
3.12
3.10
3.45
2.89 3.20
3.28
3.28
                    120

-------
      All  the organotin  compounds showed strong sorption characteristics.
Average log K  values were 4.04 ± 0.10 for TBT;  3.59 ± 0.12  for DBT; and 3.3 ±
0.14 for MBT tor sorption  on  all three types of wastewater solids  tested.  The
sorption data for TBT  on mixed liquor solids from Table 2 is plotted according
to the logarithmic form  of the Freundlich equation (Equation 3)  and  is  shown
in Figure 1.
     10000p—
   CO
   CO


   O)
   ^x
   O)
   3
      1000
        100
         10
                                                       K • 6.18
                                                       n •  1.07
                                                       r • 0.96
                    i   i i  i i i i i
                                           i  i i i i i _ i _ i  i  i  i i i i
1
                                                                  1000
                      10                 100

                             Ce, ug/l

Figure  1.  Sorption isotherm for TBT on mixed liquor  solids.
                                     621

-------
      The isotherm is typical  of  those  obtained  for all  the organotin
compounds on all three types of wastewater  solids  studied.

B10TIC STUDIES  (INHIBITION TESTING)

      Knowledge of inhibitory  characteristics  of TBT to  the microorganisms is
important for determining fate of TBT in  activated sludge  processes.
Electrolytic respirometry studies were  used to determine the inhibition
characteristics of the compounds  under  consideration using  unacclimated
biomass.  The studies were conducted using  an  automated, continuous oxygen
uptake measuring Voith, Sapromat  B-12 respirometer.   The instrument consists
of a temperature controlled waterbath,  a  recorder,  and a cooling  unit.   The
waterbath contains the measuring  unit shown in Figure 2.
             C                B               A


          Figure 2.  Electrolytic respirometry (measuring unit).
               A. Reaction vessel,          1. Magnetic stirrer,
               _ _            .           2. Sample (250 ml).
               B. Oxygen generator.         _ _  .    ..  . .
                                         3. Carbon dioxide absorber.
               C. Pressure indicator.         4. Pressure indicator.
                                         5. Electrolyte,
                                         6. Electrodes.
                                         7. Recorder.
The recorder shows the digital  indication of oxygen  uptake  and constructs a
graph of these values for each  measuring unit.   The  recorder is also connected
to an IBH-AT computer, which  collects  data every fifteen  minutes.   The cooling
unit constantly recirculates  water  to  maintain  constant temperature in the
waterbath.  Each measuring  unit consists of a reactor vessel  with  a CO^
absorber mounted in  a stopper,  an oxygen generator,  and a pressure indicator.

-------
This measuring unit is interconnected by tubes, forming an air sealed system,
so that the atmospheric pressure fluctuations do not adversely affect the
results.  The magnetic stirrer ensures effective exchange of gases by
providing vigorous agitation.  The activity of the microorganisms in the
sample creates a vacuum which, via a pressure indicator, triggers the oxygen
generator.  It supplies the required amount of oxygen by electrolytic
dissociation of a copper sulfate solution in 5% sulfuric acid.  The quantity
of the sample, the amperage for electrolysis, and the speed of the synchronous
motor are so adjusted that, with a sample of 250 ml, the digital counter
indicates the oxygen uptake directly in mg/L.  Carbon dioxide generated is
absorbed by soda lime, the nitrogen/oxygen ratio in the gas phase above the
sample is maintained throughout the experiment, and there is no depletion of
oxygen.

      The nutrient solution used in these studies was a synthetic medium
formulated by the Organization for Economic Cooperation and Development (OECD)
and was prepared from the stock solutions shown in Table 5.

                    TABLE 5.  COMPOSITION OF OECD NUTRIENT
                    SOLUTION FOR ELECTROLYTIC RESPIROMETRY
      StocK Solution
Nutrient
Weight/L
A.
B.
C,
D.
E.



F.
KH2P04
K2flPO,
Na-HP04.2H?0
NH:CI
MgS04.7H20
CaCl,
FeCl3.6H20
MnS04.4H20
H3B03
ZnS04.7H20
(NH)6Mo70,4
FeCl, EDTA
Yeast Extract
8.5 g
21.75 g
33.4 g
2.5 g
22.5 g
27.5 g
0.25 g
39.9 mg
57.2 mg
42.8 mg
34.7 mg
100.0 mg
150.0 mg
Nutrient solution prepared by diluting 10 mL A and 1.0 mL B-F to one liter in
distilled water.
Aniline, at a concentration of 100 mg/L, was used as the reference substance.
The concentrations of tributyltin used for inhibition studies was 50, 100,
500, and 1000 Mg/L.  Tributyltin stock solution was prepared in acetone and
the reactor vessels were coated with appropriate amounts of this solution to
achieve desired concentration.  Acetone was evaporated before aidding the
nutrient solution and make-up deionized water.

      The microbial biomass was activated sludge from the Little Miami
wastewater treatment plant in Cincinnati, Ohio, which receives primarily
municipal domestic wastewater.  The activated sludge sample was brought to an
                                      S23

-------
endogenous phase  by aerating for 24 hours.  The sludge biomass  at  30 mg/L was
added to the medium.   The synthetic medium was brought to  a  final  volume of
250 ml.

      The experimental  system for inhibition studies consisted  of  duplicate
flasks for the  reference substance aniline, toxicity controls  (tributyltin
plus aniline at 100 mg/L),  and inoculum control.  An unacclimated  microbiota
was used.  The  contents of the reaction vessels were first stirred for an hour
to ensure an endogenous respiration state at the initiation  of  oxygen uptake
measurements.   The  contents were then transferred to reaction vessels coated
with tributyltin  and aniline was added.  The reaction vessels were incubated
at 25*C in the  dark,  enclosed in the temperature controlled  waterbath, and
stirred continuously throughout the run.

      Oxygen uptake of the reference substance aniline and the  toxicity
controls, containing tributyltin at four different concentrations, was
followed over a period of 7 days.  The oxygen uptake data  is graphically
represented in  Figure 3.
  e>
  OL
  3
  X
  O
200


180-


160-


140-


120-


100-


80-


60-


40-


20-


 01
   0
CONTROL

ANILINE (100 mg^)

ANILINE+TBT(0.05m94>

ANILINE»TBT(0.1m<»a}

ANILINE+TBT(0 Smgrf)

ANILINE »TBT (1
                          TIME (days)


              Figure 3.  Oxygen uptake data for inhibition studies
                                      624

-------
Analysis of the culture samples at the end of the resplrometrie  run  revealed
80% to 90% degradation of tributyltin.  The oxygen uptake  data shown in  Figure
3 demonstrate that TBT was not inhibitory to the  degradation  of  aniline  up  to
concentrations of 1000 #g/L.   However, there was  an impact on the  lag time
prior to the onset of degradation which increased with  increasing  TBT
concentrations.

BIOTIC STUDIES (CONTINUOUS FLOW SWISHER REACTORS)

      The Swisher reactor simulates the flow pattern of an aeration  basin and
a clarifier.  A schematic diagram of the reactor  is shown  in  Figure  4,
               Activated carbon ,-
             Effluent out
                                                      Influent in
                                                     Activated
                                                     sludge
                                                       Air in
             Figure 4. Schematic of Swisher Flow Reactor.
      The larger arm contains the biomass which is recycled  in  the  reactor  by
air introduced in the inlet.  The smaller arm acts as a clarifier.   A series
of 300 ml continuous flow Swisher reactors were set up to investigate the
steady state biodegradation of TBT.  Sufficient air was introduced  to maintain
a dissolved oxygen concentration of 3.0 ng/l.  Influent was  pumped  to the
reactors at a flow rate of 1.2 liters per day resulting in a hydraulic
                                     625

-------
retention time of  6  hours  (liquid  volume  in  the  reactor was  200 ml).   Solids
were wasted at regular  intervals to  maintain the desired solids retention  time
(SRT).  The reactors were  charged  with  activated sludge from the Mill  Creek
sewage treatment plant.  The  units were fed  wastewater spiked with  TBT at  1,0,
12.5, 50, and 100  #g/L.  Two  wastewater feeds, natural  and synthetic,  were
used.  Primary effluent  from  the Mill Creek  treatment  facility was  used as
natural feed and synthetic feed was  prepared in  the  laboratory.  Composition
of the synthetic feed is given in  Table 6.
                  TABLE  6.   COMPOSITION  OF  SYNTHETIC  FEED  FOR
                             CONTINUOUS SWISHER  FLOW REACTORS
Bottle
Bl



B2
B3
B4
B5
B6



B7
B8
Chemical Concentration (q/L)
Phosphate Buffer
NH4C1
KH,P04
K2flP04
Na2HP04
Magnesium sulfate
Calcium chloride
Ferric chloride
Sodium bicarbonate
Organics
Yeast extract
Bactopeptone
Meat extract
Urea
Fish meal

1.7
8.5
21.75
33.4
22.5
36.0
0.25
75.0

55.0
50.0
50.0
50.0
50.0
Volume added (mL)
20.0



20.0
20.0
20.0
80.0
20.0



20.0
10.0
      The above chemicals were  added  to  four  liters  of distilled water  to  form
synthetic feed to the Swisher flow  reactors.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ^^11. |^^^^^—||,,||^^^^^^^_||f^^^^^^_||^^|_—^^,,, I,,,— MI   IIIIIIIIIM-|^^_|,|—^^,^

      Mixed liquor volatile  suspended solids, chemical oxygen  demand (COD),
total Kjeldahl nitrogen  (TKN),  total  organic  carbon,  ammonia nitrate, and
nitrite were measured to document proper operation of the  continuous flow
reactors.  The experiment was conducted  at two  SRTs  (8 and 18  days).  Samples
were analyzed for TBT and its degradation compounds  to evaluate the
acclimation and steady-state performance of the continuous flow reactors.
Results obtained at steady state are  presented  in Table  7.
                                      626

-------
                  TABLE 7.  STEADY STATE DEGRADATION OF TBT IN
                            CONTINUOUS FLOW SWISHER REACTORS
Feed type   Initial Cone.     Percent Biodegradation
            TBT, jig/L         8 day SRT         18 day SRT
Natural



Synthetic



100
50
12.5
1.0
100
50
12.5
1.0
63.1 + 2.5
72.3 + 4.4
72.0 +12.5
75.0 ±10.0
83.4 + 2.0
94.9 + 4.0
83.0 + 8.8
85.0 +15.0
86.0 + 2.0
97.0 + 3.0
89.5 + 8.0
98.0 + 2.0
90.0 + 2.5
95.0 + 4.0
94.5 + 4.0
88.9 +11.0
      Steady state degradation data in Table 7 for the continuous flow Swisher
reactor demonstrate that substantial biodegradation was achieved with both
natural and synthetic wastewater feeds.  Biodegradation increased with the
longer SRT.  The major biodegradation product was MBT with lesser amounts of
DBT.  Mass balance calculations indicated inorganic tin was also produced.

BIOTIC STUDIES (SHAKER FLASK EXPERIMENTS)

      Initially an attempt was made to measure the kinetics of the TBT
degradation reaction using electrolytic respirometry.  TBT at 10, 20, and 30
mg/L was added to the reaction flasks.  Microbiota used was taken from the
continuous flow Swisher reactors.  After 7 days in the run no biological
activity was observed in the respirometric flasks.  As a result kinetic data
were determined in shaker flask studies.  These studies were conducted by
contacting TBT at desired concentrations with the acclimated biomass in a
batch mode and determining the change in concentration of TBT with time.

      In the experiments, 250 mL flasks were filled with 100 mL of feed
(synthetic and natural wastewater separately), spiked with 50 /jg/L and 100
/ig/L of TBT.  Acclimated biomass from Swisher reactors was then added to the
flasks.  Actual concentrations were determined at zero time by analyzing the
contents of the flasks.  The flasks were then analyzed at varying times for
TBT and its degradation products.

      The biokinetic constants were determined assuming a constant biomass
concentration.  This assumption is reasonable on the basis that the amount of
degradable substrate is very small compared to the amount of biomass already
present and hence the increase in the biomass concentration due to TBT
degradation will  be small.  Kinetic constants were determined on the basis of
Monod's kinetic model assuming negligible bacterial decay constant (b).  The
kinetic constants k (rate constant) and K  were determined from the slope and
intercepts of the least squares straight line fit of -dt/ds vs. 1/s where s is
the concentration of TBT and -ds/dt is the rate of change of concentration.
                                     627

-------
Values of k and Kc for different concentrations are listed in Table 8.
                 5
                      TABLE 8.   MONOD'S KINETIC CONSTANTS
                                   8 day SRT
Feed type Inlet concentration Ks

              TBT (/zg/L)    jigmol/L
                        oqmolTBT
                    (mg bacteria - hr)
Natural
Synthetic
50.0
100.0
50.0
100.0
1.34
1.39
1.25
1.32
0.33
0.33
0.31
0.32
0.92
0.93
0.92
0.85
                                  18 day SRT
Natural


Synthetic
 50.0
100.0

 50.0
100.0
1.34
1.47

1.41
1.35
0.32
0.35

0.34
0.33
0.76
0.81

0.86
0.95
Ks and k were obtained from a plot of -dt/ds vs 1/s assuming constant
bacterial concentration.
                        - dt- JL 1
                          ds
                      kX,
      Data presented in Table 8 demonstrate that the biodegradation kinetic
parameters for TBT were independent of SRT over the interval of 8-18 days.  In
addition, the parameter values for both natural and synthetic wastewater feeds
were essentially the same.

                  BIOTIC STUDIES  (ANAEROBIC SCREENING OF TBT)

      Anaerobic treatment processes are widely used for biological
stabilization of concentrated wastewater sludges.  The objective of the
screening study was to determine  whether or not the anaerobic process was
adversely affected by sludges containing sorbed TBT.  A batch anaerobic
bioassay technique based on the method developed by Owen (10) was used to
determine the biochemical methane potential (BMP) and the anaerobic toxicity
assay (ATA) for TBT.

      The BMP assay was conducted in 250 ml reagent bottles and the ATA assay
was conducted in 125 ml reagent bottles with rubber serum caps of appropriate
sizes.  Bottles were gassed at a  flow rate of approximately 0.5 L/min with 30%
                                     628

-------
COp and 70% N2,  then,  stoppered  and  equilibrated  at  an  incubation  temperature
35 C prior to introducing samples, defined medium and inoculum.  Defined
medium was prepared using the recipe given in the Table 9 and stored at 4°C
until needed.

Solution
SI
S2
S3
S4








S5
S6
S7








TABLE 9. STOCK SOLUTIONS
FOR DEFINED MEDIUM
Compound Concentration g/L
Sample
Resazurin
(NH.)2HP04
CaCl,.2H,0
NH4CT
MgCl,.6H?0
KC1
MnCl2.4H20
CoCl2.6H20
Na,MoO,.2H?0
H3io3
CuCl2.2H20
ZnCl2
FeCl3.4H20
Na2S.9H20
Biotin
Folic acid
Pyridoxine hydrochloride
Riboflavin
Thiamin
Nicotinic acid
Pantothenic acid
B12
p-aminobenzoic acid

1.0
26.7
16.7
26.6
120.0
86.7
1.33
2.0
0.17
0.38
0.18
0.14
370.0
500.0
0.002
0.002
0.01
0.005
0.005
0.005
0.005
0.0001
0.005
                  Thioctic acid              0.005
      The defined medium contained nutrients and vitamins for mixed anaerobic
cultures.  Resazurin was added to detect oxygen contamination and sodium
sulfide was added to provide a reducing environment.  The final assay
concentrations of nitrogen, phosphorus, and alkalinity were respectively: 12
mg/L as N, 19 mg/L as P, and 2500 mg/L as CaC03.

      The defined medium was equilibrated at assay temperature, inoculated,
and transferred into serum bottles.  For BMP assay, inoculation was
accomplished anaerobically by inserting a gas flushing needle into the neck of
media flasks while adding 200 mL of seed organisms to 1800 mL of defined
medium.  A 20% by volume inoculum was used.

      Figure 5 shows a schematic diagram of the procedure for anaerobic
transfer of defined medium into serum bottles.
                                     129

-------
   Graduated
   100 ml
   pipet
J Flushing
   Defined
   medium
               Flushing
               needle
                                LJ
                                                            30 % Carbon dioxide
                                                            70 % Nitrogen
                                                          Suction
          Magnetic stirrer
               Serum bottle      Water seal
       Figure 5,  Schematic diagram of apparatus for anaerobic

                   transfer of defined medium.
      Care was taken to minimize the air to the anaerobic bottles.  After the
bottles were capped, they were zeroed with a syringe and incubated at 35°C.

      In the BMP assay, seed blanks were prepared without addition of the
organic substrate.  Samples were anaerobically added to the bottles before the
transfer of the inoculated defined medium and triplicates prepared for TBT at
10, 500, and 1000 pg/L.

      In addition to the seed blanks, a spike of acetate and propionate was
added to each bottle in the ATA assay,  A control with only the spike was also
prepared.  Each bottle contained 75 mg acetate and 26.5 mg propionate.
                                      630

-------
      Gas measurements  for  the  first few days was critical  for these studies.
Gas produced was measured with  a calibrated pressure transducer.


      Figure 6 shows  the cumulative methane gas production  for ATA studies  and
as Figure 7 shows the cumulative gas production in BMP studies.
_§


c



3
•0

2
Q.


e

f

'55
E


$
      a


      u
                                                                blank



                                                                control
          10-
                                                      800
            Figure 6.  Cumulative methane production (ATA),
                                     S31

-------
     c
     o
    TD
     O
     E

     *
     3


     O
           Blank


           lOugfl


           500ug/I


           1000 ug/1
          20-
                      200
                                400
                                           600
800
                             Tim* (hour*)

          Figure 7.  Cumulative methane production (BMP).
      Cumulative methane production in the ATA Test was  not  affected by up to
1000 fig/L of TBT as shown in Figure 6.  Methane production in the BMP assay
actually increased with increasing TBT concentration due to  degradation of the
additional alcohol solvent which was added to the  samples with the higher
doses of TBT.  The magnitude of this effect is shown in  Figure 7.  Both ATA
and BMP assays have shown that TBT is neither toxic nor  inhibitory to
anaerobic microbiota.
                                      632

-------
                                  CONCLUSIONS

      TBT and its degradation products were strongly sorbed to wastewater
solids.   Inhibition of normal biological treatment was not evident up to the
solubility limit of TBT,  Organotin compounds were biodegradable under
acclimated conditions.  Sorption on wastewater solids would be a dominant
removal  mechanism under unacclimated conditions.  TBT sorbed on wastewater
solids did not inhibit anaerobic processes.  Fate of TBT and its degradation
products in anaerobic treatment has not been determined.
                                     633

-------
                                  REFERENCES

1.    Magulre, R.J., Chau, Y.K., Bengert, 6.A., Hale, E.J., Wong, P.T.S., and
      Kramer, 0. -  Occurrence of Organotin Compounds in Ontario Lakes and
      Rivers.  Environ. Sci. & Techno!. 16> 698 (1982).

2.    Bushong, S.J., Hall, L.W., Hall, S.W., Johnson, W.E., and Herman, R.L.-
      Acute Toxicity of Tributyitin to Selected Chesapeake Bay Fish and
      Invertebrates.  Water Research 2£ (8): 1027 (1988).

3.    Barug, D. Hicrobial Degradation of Bis-  (tributyltin) Oxide.
      Chemosphere. Ifi, 1145 (1981).

4.    Argaman, Y., Shelby, S.E., Jr., and Hucks, C.E., David W. Taylor Naval
      Ship Research and Development Center Report SME-81-52, Annapolis, MD
      (1981).

5.    Avendt, R.J., and Avendt, J.B. -  Performance and Stability of Municipal
      Activated Sludge Facilities Treating Organotin Contaminated Wastewater.
      David W. Taylor Naval Ship Research and  Development Center Report SME-
      CR-29/82, Annapolis, MD (1982).

6.    Matthias, C.L., Bellama, J.M., Olson, 6.J., and Brinckman, F.E. -
      Comprehensive Method for Determination of Aquatic Butyltin and
      Butylmethyl tin Species at Ultratrace Levels Using Simultaneous
      Hybridization/Extraction With Gas Chromatography - Flame Photometric
      Detection.  Environ. Sci. & Techno!. 20_, 609 (1986).

7.    Muller, M.D. - Comprehensive Trace Determination of Organotin Compounds
      in Environmental Samples Using High Resolution Gas Chromatography With
      Flame Photometric Detection.  Anal. Chem. 59, 617 (1987),

8.    Mackay, D., Shiu, W.Y., and Sutherland,  R.P. - Determination of Air-
      water Henry's Law Constant for Hydrophobic Pollutants.  Environ. Sci. &
      Techno!. 13, 333 (1979).

9.    Dobbs, R.A., Jelus, M», and Cheng, K., -  Partitioning of Toxic Organic
      Compounds on Municipal Wastewater Treatment Plant Solids.  EPA/600/D-
      86/137.  July 1986; NTIS PB68-218427.

10.   Owen, W.F., Stuckey, D.C., Healy, J.B.,  Jr., Young, L.Y., and McCarty,
      P.L. - Bioassay for Monitoring Biochemical Methane Potential and
      Anaerobic Toxicity.  Water Research 13,  485 (1979).
                                     634

-------
                  SUPERFUND TECHNICAL SUPPORT ACTIVITIES
                        by:   Dr.  Benjamin  L.  Blaney
                               U. S. Environmental Protection Agency
                               Cincinnati, Ohio 45268
                                ABSTRACT
    The  Technical Support Branch of the Superfund Technology
Demonstration Division of the Risk Reduction Engineering Laboratory,
provides technical support to the EPA Regions through three programs:  1.
Superfund Technical Assistance Response Team(START), 2.  the Treatability
Assistance  Program and, 3.  the Technology Support Project.

    These programs offer engineering assistance on site-specific
remediation problems through the use of RREL technology teams with
expertise in the areas of incineration, materials handling,
solidification/stabilization, extraction, chemical, biological, and
aqueous treatment.  The Branch also performs a number of technology
transfer activities in order to provide engineering data and other
information on Superfund site remediation to EPA Regions, and Program
Offices, and to others. Technology transfer documents include treatment
technology protocols and briefs, guidance documents on technical issues
associated with RI/FS development, treatment technology vendors lists and
treatment technology data bases.  The Branch also coordinates Regional
requests for RREL to perform in-house treatability studies.
                                    635

-------
           "Guide for Conducting Treatability Studies under CERCLA"
                             Jonathan S. Herrmann
                     U.S. Environmental Protection Agency
                            Cincinnati,  Ohio  45268

                                      and

                               Judy L. Hess!ing
                              Gregory D. McNeily
                             PEI Associates, Inc.
                                Chester Towers
                            Cincinnati,  Ohio  45246
      Systematically conducted, well-documented treatability studies are an
important component of the remedial investigation/feasibility study (RI/FS)
process under the Comprehensive Environmental Response, Compensation, and
Liability Act (CERCLA).  These studies provide valuable, site-specific data
necessary to aid in the selection of the remedy.  This guide, which is being
issued in interim final form, focuses on treatability studies conducted before
the Record of Decision (ROD) has been signed; however, treatability studies
may well continue into the remedial design/remedial action (RD/RA) phase.

      The guide describes a three-tiered approach for conducting treatability
studies that consists of:  1) laboratory screening, 2) bench-scale testing,
and 3) pilot-scale testing.  Depending on the information gathered during site
characterization and technology screening and the data gaps that exist,
treatability studies may begin with any tier (e.g., bench-scale testing) and
may skip tiers that are not needed  (e.g., laboratory screening followed by
pilot-scale testing).

      The guide also presents a stepwise approach or protocol for conducting
treatability studies for determination of the effectiveness of a technology
(or combination of technologies) in remediating a CERCLA site.  The steps
include:  (1)  establishing data quality objectives, (2) selecting a
contracting mechanism, (3) issuing the Work Assignment, (4) preparing the Work
Plan, (5) preparing the Sampling and Analysis Plan, (6) preparing the Health
and Safety Plan, (7) conducting community relations activities, (8) complying
with regulatory requirements, (9) executing the study, (10) analyzing and
interpreting the data, (11) and reporting the results.
                                     636

-------
            IMPLEMENTING A WASTE MINIMIZATION  PROGRAM AT THE
            A.M. BREIDENBACH  ENVIRONMENTAL RESEARCH CENTER

               by:    Robert  L. Gould
                     Joseph  W. Till man
                     Science Applications  International Corporation
                     Cincinnati, Ohio  45203

                     Brian A.  Westfall
                     U.S. Environmental  Protection Agency
                     Cincinnati, Ohio  45268
                               ABSTRACT

  A permanent waste minimization  program  has  been  initiated  at  EPA's
A.M.  Breidenbach Environmental Research Center in Cincinnati, Ohio.
The Waste Minimization Opportunity Assessment Manual (EPA/625/7-
88/003) generic procedures were modified as needed to provide a basis
for evaluation of wastes generated in the laboratories and offices at
the Center.  The waste minimization assessment team gathered
information through a site visit and follow-up correspondence.
Results of the assessment will be used to document waste minimization
opportunities, strengthen the waste minimization program
implementation plan, and identify research needs and potential
demonstration projects for waste minimization at a laboratory and/or
office facility.
                                 637

-------
    BIOLOGICAL AND PHYSICO-CHEMICAL REMEDIATION OF A MERCURY
                    CONTAMINATED HAZARDOUS WASTE


                          by:  Conly L. Hansen
                               David K. Stevens
                               Gour S. Choudhury
                               Ravinder Menon
                               Utah State University
                               Logan, Utah 84322
                                  ABSTRACT

   Biological and physico-chemical removal of mercury from contaminated water and
sediments involves enzymatic reduction of ionic mercury (Hg2+) to the elemental form
(Hg°) and subsequent volatilization or precipitation. A continuous biological process for
removing Hg2+ from water has been tested in laboratory scale.  Greater than 99% of up to
100 mg/L influent Hg2+ can be removed using the process. A simplified schematic of the
process is given below.

                   Hg(0) vapor
 100 mg/L
 Hg(2+) '
Bioieactor
1.5mg/L
Total Hg
as Hg(0)
J
1
Centrifuge
<0.01 mg/L.
Total Hg
                                              Sludge that is
Hg2+ is reduced at a rate of 1.7 - 2.0 mg/L-h using an acclimated, mixed strain of mercury
resistant bacteria improved over the years in our laboratory. Diluted whey permeate (10: 1)
is added in equal volumes with the Hg2+ contaminated water as an inexpensive nutrient
source for the bacteria. Final disposition of most of the biologically reduced mercury is
controlled by aeration rate, hydraulic retention time and Hg2+ loading rate. Volatilized Hg°
is trapped by adsorption on activated carbon (AC) or chemically impregnated AC. Hg
adsorbed on AC can be desorbed by heating. Desorbed vaporous Hg can be condensed for
reuse.
                                       638

-------
    STEAM STRIPPING AND BATCH DISTILLATION FOR REMOVAL AND/OR RECOVERY OF
                         VOLATILE ORGANIC COMPOUNDS

                   by:  Sardar Q. Hassan
                        Department of Chemical Engineering
                        University of Cincinnati
                        Cincinnati, Ohio  45221-0171

                        Dennis L. Timber!ake
                        Risk Reduction Engineering Laboratory
                        U.S. Environmental Protection Agency
                        Cincinnati, Ohio  45268
                                  ABSTRACT
     The Risk Reduction Engineering Laboratory of the U.S. Environmental
Protection Agency Office of Research and Development is conducting treatment
technology assessment for the RCRA listed hazardous wastes  at its
Cincinnati Test and Evaluation (T&E) Facility.  As a part of this program,
pilot-scale testing is being conducted to evaluate the performance of steam
stripping and batch distillation for removal and/or recovery of volatile
organic compounds.  A 2-inch I.D. steam stripping column with packing
equivalent to 20 theoretical plates has been selected for the testing.  The
batch distillation unit has a 125-gallon batch still and a 6-inch I.D.
column with packing equivalent to 50 theoretical plates.  At present, these
units are in different stages of testing and shakedown.  To date, steam
stripping runs have been performed with toluene, isopropanol and 2-
nitropropane.  Different computer programs, for prediction of vapor-liquid
equilibrium data and column design and/or evaluation have also been
reviewed.  Selected programs have been modified for use at the T&E Facility.
A bench-scale unit is currently being set up to experimentally determine
required vapor-liquid equilibrium data.
                                     639

-------
      OZONE/ULTRAVIOLET LIGHT TREATMENTOF FERRI-CYANIDE COMPLEXES

             by:  Mark J. Briggs
                  PEI Associates, Inc.
                  Cincinnati, Ohio 45246

                  Sardar Q. Hassan
                  Department of Chemical Engineering
                  University of Cincinnati
                  Cincinnati, Ohio 45221-0171

                  Dennis L. Timberlake
                  Risk Reduction Engineering Laboratory
                  U, S. Environmental Protection Agency
                  Cincinnati, Ohio 45268
                                ABSTRACT

     Recent concern over the quantity of total cyanide being disposed of
in landfills and discharged to surface water has prompted the U.S. EPA
Office of Solid Waste to begin investigating alternative treatment
methods for electroplating wastewater.  UV aided ozonation of ferri-
cyanide complexes is one of the technologies currently being evaluated
at the US EPA Test and Evaluation Facility in Cincinnati.  Ferri-cyanide
complex is a major cyanide source remaining after the destruction of
free cyanide by alkaline chlorination.  UV aided ozonation had been
reported to have promise for treatment of ferri-cyanide complexes.  In
the current work the effects of different parameters, such as UV light
intensity, temperature, ozone flow rate, and initial ferri-cyanide
concentration, on the destruction rate of ferri-cyanide complex is being
studied.  This work will also include the determination of the effect of
ozone activation by UV light in a separate chamber prior to ozone
introduction into the reactor containing the ferri-cyanide solution.
                                  840

-------
                          THE  USEPA/WRITE  PROGRAM

                           by:  Paul M. Randall
                                Johnny Springer
                                U.S. Environmental Protection Agency
                                Cincinnati, Ohio  45268
                                 ABSTRACT

     In 1989, the U.S. Environmental Protection Agency (USEPA) entered
into cooperative agreements with six (6) states: California, Connecticut,
Illinois, Minnesota, New Jersey, and Washington to participate in a new
pollution prevention research program called Waste Reduction Innovative
Technology Evaluation (WRITE).  In this three (3) year program, the states
work with industry to demonstrate and evaluate source reduction and
recycling technologies to reduce or eliminate waste.  Technology
evaluations will include both engineering effectiveness and cost analysis.
This presentation will focus on two (2) of these WRITE programs:  Illinois
and New Jersey.

Illinois

     In June of 1989, the Illinois Hazardous Waste Research and
Information Center (HWRIC) and the USEPA began a three (3) year program.
Some projects under consideration include substituting water-based and
nontoxic liquid cleaners in flexographic printing, soybean-based inks in
off-set printing, closed loop recycling and substituting zinc hydroxide
for zinc cyanide in an electroplating process and other projects.
Discussion will focus on the approach, methods, and technologies.

New Jersey

     In August of 1989, the New Jersey Department of Environmental
Protection (NJDEP) and USEPA began this three (3) year project.  Some of
the technologies being considered are: an electroplating wastewater
recycling system, substitution of a colloidal compound cleaner for
degreasing operations, and other projects.  Discussion of the project will
include project objectives and approach, obstacles encountered and
technologies under consideration.
                                    641

-------
         AMERICAN INSTITUTE FOR POLLUTION PREVENTION

                  by«  Thomas R. Hauser
                       University of Cincinnati
                       Cincinnati, Ohio ^5221
                           ABSTRACT

     The Environmental Protection Agency, jointly with the
University of Cincinnati, formed the American Institute for
Pollution Prevention (AIPP) in June 1989.  The institute's
primary goal is to generate broad support from the public and
private sectors and assist EPA in achieving the widespread and
expeditious adoption of pollution prevention concepts.

     The EPA approach to pollution prevention is somewhat
unique among all other EPA environmental protection efforts
in that it is not a regulatory-based, adversarial type program
but rather must achieve success by relying on information
transfer and persuasion with respect to both the motivation
of pollution generators to incorporate pollution prevention
practices and how, from a technical viewpoint, pollution
prevention can be achieved,

     AIPP members have been nominated by their respective
professional societies or trade associations and have
individually demonstrated numerous accomplishments in the
field of pollution prevention.  They have all agreed to serve
voluntarily because of their continuing interest in
environmental protection.  The AIPP has formed four Councils
to address its goals in the areas of economics, education,
implementation and technology.
                              642

-------
                    U.S. EPA INCINERATION RESEARCH FACILITY UPDATE
                      by:    Johannes W. Lee
                            Acurex Corp
                            Jefferson, Arkansas 72079

                            Robert C. Thurnau
                            U.S.  Environmental  Protection Agency
                            Cincinnati, Ohio  45268
                                        ABSTRACT

     In the last few years, the EPA Incineration Research Facility (IRP, formerly CRF) In
Jefferson,  Arkansas has conducted incineration tests at an accelerated pace to meet the
data needs of the Agency. In FY'89, its physical  plant underwent a major expansion. A new
building, completed in July 1989, increased the  enclosed space to 15200 square  feet and
provided a facility more consistent  with the increasing level of activities.

    Concurrent with building construction,  the rotary kiln incinerator system (RKS)  was
reconfigured to improve several aspects of its operation. These included implementing more
ideal sampling ports at the afterburner exit; relocating  the scrubber system to allow easy
connection of alternate air pollution control systems (APCS); replacing the secondary APCS
by one of  improved design; and making arrangements for an automatic control system for
the entire RKS.

    Ongoing research at the IRF focused on providing support to Regional Office remediation
efforts by  conducting treatability tests. Two test series were completed in FY'89. One looked
at the applicability of incineration  in decontaminating arsenic-contaminated soil  at  a
Region 1 Superfund site. The other studied the treatability of the contaminated-soils at two
Superfund sites in Region 9.

    Another series of parametric tests continued the study on  the distribution of  toxic
trace metals in incinerator systems.  A previous series was conducted with a venturi/packed
column scrubber for air pollution control. The recent series was  conducted with an ionizing
wet scrubber.

    Planned test programs include POHC  incinerability  ranking  evaluation; a listed waste
BOAT evaluation; continued trace metal distribution  studies with alternate APCDs;  further
testing to  support Regional remediation effort; and third party  testing as provided for by
the Federal Technology Transfer Act.
                                           643

-------
    PEVELOPMENT AND DEMONSTRATION OF A PILOT-SCALE DEBRIS
                                WASHING SYSTEM


                           by:    Michael L. Taylor
                                  Majid A. Dosani
                                  John A. Wentz
                                  Avinash N. Patkar
                                  PEI Associates, Inc.
                                  Cincinnati, Ohio

                                  Naomi  P. Barkley and Donald E. S arming
                                  U. S. Environmental Protection Agency
                                  Risk Reduction Engineering Laboratory
                                  Cincinnati, Ohio

                                  Charles Eger
                                  U. S. Environmental Protection Agency
                                  Atlanta, Georgia


                                     ABSTRACT

       A large number of hazardous waste sites in the United States are littered with metallic,
masonry, and other solid debris which may be contaminated with hazardous chemicals [e.g.,
polychlorinated biphenyls (PCBs), pesticides, lead, or other metals] and in some cases clean-up
standards have been established (10 p.g PCBs/100 cm2 for surfaces to which personnel may be
frequently exposed). Although the majority of debris at Superfund sites does not possess the
potential for reuse, the debris could, following decontamination, either be returned to the site as
"clean fill" in lieu of transporting the debris offsite to a hazardous waste landfill or, in the case of
metallic debris, be sold to a metal smelter.

       During previous phases of this project we have developed a technology for performing on-
site decontamination of debris. Both bench-scale and pilot-scale versions of a Debris Washing
System (DWS) have been designed and constructed. The DWS utilizes an aqueous solution
which is applied to the debris during a high pressure spray cycle followed by a turbulent wash
cycle. The aqueous cleaning solution is recovered and reconditioned for use concurrently with the
actual debris cleaning process;  therefore the quantity of process water utilized to clean die debris is
minimized.

       In this paper, we present the results obtained during a field demonstration of the DWS.
Data will be presented  which are indicative of the effectiveness of the system for removing PCBs
from the surfaces of metallic debris, as well as the efficiency of the closed-loop solution
reconditioning system which is built into the DWS. The performance of the DWS under adverse
(sub-freezing) temperature at a hazardous waste site located in Hopkinsville, Kentucky will also
be discussed.
                                         644

-------
             ENGINEERING HANDBOOK FOR HAZARDOUS WASTE INCINERATION
Leo Weitzman, PhD
LVW Associates, Inc.                      C. C. Lee, PhD
Durham, NC                                Risk Reduction Engineering Laboratory
                                          EPA, Cincinnati, OH
Justice Manning, PE
Center   for  -Environmental  Research      Sonya Stelmack
Information                               Office of Solid Waste:
EPA, Cincinnati, OH                       EPA Washington, DC
      One of the most difficult, and time-demanding, tasks for a permit writer
is to evaluate and  interpret  incinerator designs and to determine whether the
construction  and testing  of  a  hazardous waste  incinerator  will   cause  an
unreasonable impact  on health and the environment.  In addition, Individuals must
understand the fundamentals of incinerator  construction and operation  to design
trial burns  and to evaluate  the results.    This paper describes  the revised
Engineering  Handbook  for Hazardous Waste  Incineration,  a document due  to be
published in late  1990 which will  provide guidance on the  basic principles,
concepts and designs of hazardous waste  incinerators.   The Engineering Handbook
will also give  methods that  can be used  to  estimate the performance  of the
incinerator and its ancillary equipment.

      The Engineering Handbook  is an extensive  revision and  update of SW-889,
a document by the same  name which was published in 1981.   It is a compilation
of  information  from  the open  literature  describing  current  techniques  and
equipment used for hazardous waste incineration.  The document is not intended
to be a  detailed design  handbook, rather it is intended as  a reference  for those
individuals who are associated with the regulatory aspects of the subject.  It
is  specifically aimed  at permit  writers and   regulators  who are  generally
unfamiliar with the field and need  a central  source  of basic information.  As
such, its main emphasis is on the evaluation of equipment and their designs to
assess how well  new incinerators, whose designs are  being evaluated  prior to
construction, are likely to achieve their function in a safe and environmentally
sound manner. The Handbook also contains fundamental thermodynajnic and kinetic
information and techniques that  can  be  used  to  evaluate,  design and operate a
facility.

      This work is being conducted under EPA contract 68-C8-0011, WA 1-30 to the
Eastern Research Group, Inc.   LVW Associates, Inc., the Environmental Research
Division  of  Acurex  Corporation and  John  Richards,  are  participating  as
consultants  and subcontractors.   The work is being  performed  under  the joint
guidance and sponsorship of the Office of Solid Waste and the Office of Research
and Development of the U.S. Environmental Protection Agency.

                                     645

-------
        AN ASSESSMENT OF THE CURRENT STATE OF THE ART OF
                   BIO/PHOTODEGRADABLE PLASTICS

                         by:   R. B. Sukol
                              J, Poles
                              J. Miller
                              C. Chambers
                              PEI Associates, Inc.
                              11499 Chester Road
                              Cincinnati, OH 45246

                              H. Haxo
                              Matrecon, Inc.
                              815 Atlantic Avenue
                              Alameda, CA  94501


                               ABSTRACT

    The focus of this project has been to assist EPA in developing parameters to
support Federal regulation of the disposal  of consumer-use plastic products such
as polyethylene, polyvinyl chloride (PVC), and polystyrene. Although plastics
recycling is increasing, major concern still exists regarding the fate of these
products and  their ecological effects in the soil and the marine environment.
Objective methods are needed to define degradation, deterioration, and
decomposition and to judge the relevance and value of engineering tests.

    The authors recommend that 6-pack carrier rings be manufactured from
photodegradable materials, and they conclude that insufficient research has been
conducted to evaluate engineering tests.  The project also includes a review of
State and local regulation of plastics.
                                   646

-------
              FIELD TESTS OF HYDRAULIC FRACTURING
                 TO INCREASE FLUID FLOW IN SOILS

               L.C. Murdoch, G. Losonsky, I. Klich, P. Cluxton


                       Center Hill Research Facility
             Department of Civil and Environmental Engineering
                          University of Cincinnati
                          Cincinnati, Ohio 45224

      Fluid flow in the subsurface is a basic requirement of many methods of
in-situ remediation,  including pump and treat, soil flushing, vapor extraction,
bio-remediation, and steam stripping. Hydraulic fracturing is a method of
increasing fluid flow to improve the effectiveness of those m-situ methods,
especially in areas where flow is limited by material of low permeability.
Basically, hydraulic fracturing involves injecting fluid into a well until pressures
exceed a critical value and fracturing begins. Sand is then pumped into the
fracture to hold it open and provide a high permeability channelway. The
process was developed more than fifty years ago to increase the yields of oil
wells. Last year at the Symposium we reported results from our first field test
where we used oil-field equipment to show that it was possible to create
hydraulic fractures at shallow depths in soil—conditions typical of many
contaminated sites.

      Shortcomings of the 1988 test led us to redesign equipment and methods
of hydraulic fracturing at shallow depths.  The new method, which employs
equipment similar to that used for grouting, was field tested in June  1989 with
the creation of 23 fractures at two sites in Cincinnati, Ohio. Both sites are
underlain by overconsolidated, silty-clay till.  Nineteen of the fractures were
excavated to examine their size, thickness and orientation.  The excavated
fractures all contained sand and they reached a maximum thickness of one inch
and a maximum length of 28 feet.  As  many as four flat-lying fractures, stacked
one on top of another at spacings as close as 6 inches, were created from the
same borehole. Venting of the fractures to the ground surface was reduced
compared with results from 1988.  Steady-state rates of inflow into boreholes
intersecting hydraulic fractures were 3.2 and 9.1 greater than steady rates into
boreholes in unfractured till, according to results of tests conducted using a
constant-head permeameter.
                                  647

-------
       BARRIER EQUIVALENCY OF LINER AND CAP MATERIALS

                           by:      David E. Daniel
                                    The University of Texas at Austin
                                    Austin, Texas 78712
                                ABSTOACT

     The objective of this project is to investigate materials that might be used as
hydraulic barriers in liners or covers to replace or augment a low-permeability,
compacted soil liner.  In the first year of investigation, potentially equivalent barrier
materials were identified, and available data for the materials were collected,
tabulated, and analyzed. Additional data that would be needed to determine the
equivalency of the barriers were identified. In the next year of study, tests will be
performed to aid in determining the equivalency of the materials to a low-
permeability, compacted soil liner.

     The materials that were investigated included blended soils, amended soils,
manufactured blankets of clay, waste materials, concrete, sprayed-on materials, and
manufactured membrane/clay composites. The materials fell into two categories: 1)
materials that would have to be custom-designed and evaluated for each project, and
2) manufactured, "off-the-shelf1 materials, the properties of which would not have
to be repeatedly studied.  Future emphasis will be placed on the second category of
materials. The greatest amount of data and experience in the field was found for
thin blankets of bentonite attached either to geotextiles or membranes. These
materials show promise as equivalent barriers, but more data are needed to define
characteristics of seam overlaps, puncture resistance, resistance to desiccation and
frost action, and similar performance parameters.
                                   648

-------
                                   ABSTRACT

                     An Upcoming  S.I.T.E.  Demonstration:
                AWD Technologies,  Inc., "AquaDetox/SVE System"


                                Gordon  M. Evans
                     U.S.  Environmental Protection Agency
                     Risk  Reduction  Engineering  Laboratory
                               Cincinnati,  Ohio


      As part of the Agency's Superfund Innovative Technology Evaluation
(SITE) Program, a demonstration of AWD Technologies "AquaDetox/SVE System"
will be conducted during the summer of 1990.  The process.is an automated
system which combines a vacuum  steam  stripping tower (the AquaDetox unit)
with a soil vapor extraction (SVE) unit into a closed-loop system.  According
to the developer, the system is capable of removing more than SO of the 110
volatile compounds listed in CFR 40, July 1, 1989, by the U.S. EPA.  Within
the AquaDetox system, contaminated groundwater enters the top of a stripping
tower that operates under a moderate vacuum.  Steam is injected from the
bottom.  Contaminants are stripped from the water, condensed, and collected
for later recycling.  The SVE Systen removes contaminated soil-gases frota the
vadose sone through a network of extraction wells.  These soil-gases are then
exhausted to a series of granular activated carbon beds for hydrocarbon
removal.  The cleaned gases are then reinjected into the ground.  The ability
of this system to clean contaminated groundwater and soil-gases in a closed-
loop system eliminates the possibility of air emissions, and thus the need to
obtain air quality permits.

      The demonstration will occur as a part of an ongoing remediation at the
Lockheed Aeronautical Systems facility in Burbank, Ca. {part of the San
Fernando Superfund Site).   Among the innovative design features is the
addition of an automatic regeneration system for the granular activated carbon
beds.  Steam is used from the AquaDetox unit to strip the hydrocarbons from
the beds.  This vapor is sent back to the AquaDetox unit, condensed, and is
also collected for recycling.  One of the points which the SITE Program will
investigate is the developer's claim that the system enjoys a cost advantage
over dual-stage stripping systems.
                                    649

-------
       GUIDANCE ON METALS AND PIC EMISSION CONTROLS FROM HAZARDOUS
                       WASTE INCINERATION

                         Shiva Garg
              U.S. Environmental Protection Agency
                       401 M Street S.W.
                       Washington, D.C.
                           ABSTRACT

       THe U.S. Environmental  Protection Agency (EPA)  is
currently working on  regulations  to control emissions of heavy
metals and Products of incomplete Combustion (PICs)  when
hazardous waste is incinerated in incineraters,  boilers or
industrial furnaces like  cement kilns.   The EPA has also
published guidance manuals for use by the permit writers in the
Regions and the States so that they can implement the proposed
controls in the interim,  until the final rules are promulgated,
using omnibus authority provided  to the permit writers under
RCRA section 3005(c)(3).   The  methodology to control
carcinogenic metals  (arsenic,  beryllium, cadmium and chromium)
emissions has been provided such  that the incremental lifetime
risk to the maximum exposed individual (MEI)  does not exceed
10" . The control for PIC emissions is provided by using
Carbon Monoxide (CO)  and  Total Hydrocarbon (THC)  emissions from
the stacks as surrogates.  The  CO  must be monitored continuously
and must not exceed 100 parts  per milliom by volume (ppmv)
corrected to 7% oxygen and on  a dry basis.   In cases where the
CO exceeds the above  level,  the compliance must be demonstrated
by monitoring THC emissions which should not exceed 20 ppmv.
Also, the THC should  be monitored continuously in addition to CO
during the life of the permit  in  those cases.

    The regulatory proposal for hazardous waste incineration in
boilers and industrial furnaces was published in the Federal
Register (54 FR 43718) on October 26,  1989 and for incinerators
in the Federal Register  (55 FR 17862)  on April 27, 1990.  The
final promulgation is expected in 1991 for both the proposals.
                               650

-------
           Reduction of Transient Pnflfefrom a Rotary Kiln
                 The Role of the Secoimfary (^>mbusti^>ni
                                          by
                                    P.M.  Lemieux
                                     W.P. Linak
                                    J.A. McSorley
                         U.S. Environmental Protection Agency
                    Air and Energy Engineering Research Laboratory
                           Research Triangle Park, NO 27711
                                    J.O.L. Wendt
                          Department of Chemical Engineering
                                 University of Arizona
                                   Tucson, AZ 85721
                                      J.E. Dunn
                          Department of Mathematical Sciences
                                University of Arkansas
                                 Payetteville, AR 72701
                                     ABSTRACT

       Experiments on a 73 kW (250,000 Btu/hr) rotary kiln incinerator simulator equipped with a
58.4 kW (200,000) Btu/hr afterburner/control temperature tower were performed to examine the
effect of the secondary combustion chamber on reducing transient puffs generated during the batch
charging of surrogate solid waste.  In this preliminary set of tests, the effects of temperature,
oxygen and residence time were examined in the absence of flame contact and turbulent mixing.
A first generation deterministic model is being developed  to examine the response of the
secondary combustion chamber to a transient input.  Preliminary data indicate that not only does
the secondary chamber provide a reduction in the absolute  magnitude of a puff, but also
considerably alters the composition of the puff, with respect to both chemical species and biological
activity.
                                          651

-------
                        THE USE OF ELECTROKINETICS
                   FOR HAZARDOUS HASTE SITE REMEDIATION

                   by:  Joseph T. Swartzbaugh, Ph.D.
                        Barbara L. Cormier, P.E.
                        Robert E. Mentzer, P.E.
                        Andrew W. Weisman
                        PEER Consultants, P.C.
                        Dayton, Ohio 45432

                        Denis Nelson
                        Superfund Technology Demonstration Division
                        Risk Reduction Engineering Laboratory
                        Cincinnati, Ohio 45268 •


                                 ABSTRACT

    This study was performed to assess the state-of-the-art of
electrokinetically-enhanced contaminant removal in soils.  Prior research
efforts, both laboratory and field, have demonstrated that electroosmosis
has the potential to be effective in facilitating the removal of certain
types of hazardous wastes.  Particularly encouraging results have been
achieved with inorganics in fine-grained soils where more traditional
removal alternatives such as pump and treat are less effective.

    Although the results of various studies suggest that electrokinetics
is a promising technology, further testing is needed at both the
laboratory and field levels to fully develop this technology for site
remediation.  A conceptual test program is presented based on best
available data which incorporates system design and operating parameters
used in previous applications of this technology.
                                    652

-------
        GROUND PENETRATING RADAR EQUIPPED FIELD ROBOTIC VEHICLE
                          FOR SURVEYING CERCLA SITES


             by:          James F. Osbom
                          Daniel A. Christian
                          Field Robotics Center
                          The Robotics Institute
                          Carnegie Mellon University
                          Pittsburgh, PA 15213


                                     ABSTRACT

   For years, robotics and automation have increased productivity in manufacturing industries
through standardization and repeatability. Core robotic technologies have now progressed to the
point that robots are moving into the field and  offering the  same benefits.  One such
implementation is subsurface mapping of waste sites - the process of spatially correlating non-
invasive measurements in order to graphically represent the geometry of buried structures.
Through a cooperative agreement with the US EPA, the Field Robotics Center is developing a
subsurface mapping system based on a mobile robot that deploys ground penetrating radar (GPR)
transducers.

   This prototype Site Investigation Robot (SIR) maintains knowledge of its location on site and
carries a scanning mechanism that provides the local positional accuracy necessary for GPR
imaging. GPR data are digitized and stored fa memory for subsequent digital filtering, specialized
geophysical processing and image enhancement.  These processes transform the radar data into a
subsurface map which is displayed on a computer graphics terminal and stored in the site database.
While we have demonstrated these techniques for two dimensional GPR images, the focus of our
current research is development of processing techniques that exploit the three-dimensional nature
of GPR and create 3-D subsurface maps.
                                         653

-------
         DETOXIFICATION OF RELEASED VAPORS/PARTICULATES BY ENTRAPMENT
                          IN CHEMICALLY ACTIVE FOAMS
                  by:  Patricia M. Brown
                       Foster Wheeler Enviresponse, Inc.
                       Edison, New Jersey  08837


                       Ralph H. Hiltz
                       MSA Research Corporation
                       Pittsburgh, Pennsylvania  15230


                       John E. Brugger, Ph.D.
                       U.S. Environmental Protection Agency
                       Edison, New Jersey  08837

                                   ABSTRACT

    In the wake of the Bhopal disaster, there has been increasing public and
governmental concern over the possibility of toxic gas and vapor releases
from chemical plants.  Title III of SARA, "Superfund Reauthorization and
Amendment Act of 1986", specifically addresses emergency preparedness for
such releases.  Industry is under intense pressure to provide additional
safeguards to prevent "gas" clouds from threatening populated areas.

    Many ways to prevent releases are already in use by industry ranging
from safety-conscious plant design, through safety alarms and interlocks, to
backup devices such as scrubbers and flares.  However, once a gas or vapor
leak occurs, the options for its control are few.

    Under study is the use of "foam scrubbing," a novel option for
controlling emergency releases of airborne toxics.  In this approach, foam
is generated with conventional equipment using the contaminated air.  The
foaming solution contains neutralizing agents and may require a special
surfactant system for compatibility.  With the airborne materials
encapsulated in the foam, a large interior liquid surface area is available
for gas, vapor, aerosol, or particulate sorption.  Neutralization agents
present in the bubble walls can then react with the entrapped toxic gas or
vapor to render it innocuous.  The foam collapses to yield a processable
liquid.

    This proposed technology has potential applications to emergency
control; among these are fixed installations in plants and portable units
for field use by emergency response teams.


                                     654

-------
         ASSESSMENT  OF MATERIALS HANDLING TECHNOLOGIES


                           by:    Majid A. Dosani
                                  Michael L. Taylor
                                  John M. Miller
                                  PEI Associates, Inc.
                                  Cincinnati, Ohio

                                  Naomi P. Barkley
                                  Donald E. Sanning
                                  U. S. Environmental Protection Agency
                                  Risk Reduction Engineering Laboratory
                                  Cincinnati, Ohio


                                     ABSTRACT

       A typical Superfund or other hazardous waste site contains hazardous chemicals which are
frequently intermingled with remnants of razed structures as well as contaminated soil, gravel,
concrete, and metallic debris. Specialized materials handling and classifying technologies are
needed to deal with the large quantities of these various types of contaminated materials present at
the site prior to, or in conjunction with decontamination and disposal operations. The objective of
this study is to summarize the types of debris, material, and contaminants found at Superfund and
other hazardous waste sites in EPA Region V and the materials handling equipment and general
procedures that have been implemented  to perform site restoration and cleanup.

       This paper provides information  on state-of-the-art materials handling equipment and
procedures that would be useful for addressing difficult, site-specific, materials  handing
problems.  The capabEMes, performance, applicability for hazardous waste site work, and costs
for a variety of materials handling techniques  will be discussed. Case studies for sites within EPA
Region V will be presented which include detailed information concerning debris, material,
contaminants found on site, and the specific materials handling needs and problems encountered.
In addition, the materials handling technologies currently used in the United Kingdom and West
Germany will be briefly summarized.
                                         ess

-------
                   METAL TREATMENT BY ADSORPTIVE FILTRATION

                  by:  Mark M. Benjamin
                       Department of  Civil  Engineering  FX-10
                       University of  Washington
                       Seattle, Washington  98195
                                   ABSTRACT

      Treatment of solutions containing toxic metals  at Superfund  sites  is
complicated by the wide range of concentrations which may be  present  and the
complexity of the matrix.  Although highly concentrated metals may be
treatable by conventional precipitation processes, these processes may not
meet the stringent effluent discharge requirements.   Also, metals  are often
present in relatively low, but still unacceptable concentrations in the
effluent from treatment processes  for removal of organic contaminants.

      The process being tested in  this project is one which combines  an
efficient sand filter with an adsorption process capable of lowering  metal
concentrations well below levels achievable by precipitation  alone.   The
process uses a conventional sand filter in which the  sand is  first coated
with iron oxide.  The Iron oxide surface has a strong pH-reversible affinity
for the metals.  The fact that the oxide is physically attached to sand
grains rather than a suspended material makes it much easier  to work  with,
and allows the filtration of particulate metal contaminants to proceed
simultaneously with adsorption from the soluble phase.  Finally, the  process
is applicable to anionic metals such as chromate, selenite, and arsenate,
which are difficult to precipitate efficiently.
                                      656
  U&GOVBVMEMrPHNnNQOFFCE: 1SBO-748-1SB/2OIS2

-------
United States
Environmental  Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use. $300
                                                                       Please make aN necessary changes on the above label, detach
                                                                       or copy and return to the address in the upper left-hand comer.

                                                                       If you do not wish to receive these reports CHECK HERE G:
                                                                       detach, or copy this cover, and return to the address In the
                                                                       upper left-hand comer.
                                                                     EPA/600/9-90/037

-------