SEPA
           United StatQS
           Environmental Protection
           Agency
            Office of Research and
            Development
            Washington, DC 20460
EPA/600/R-93/040
April 1993
Nineteenth Annual
RREL Hazardous Waste
Research Symposium

Abstract Proceedings

-------

-------
                                                 EPA/600/R-93/040
                                                      April 1993
19TH ANNUAL RREL HAZARDOUS WASTE RESEARCH SYMPOSIUM

                    ABSTRACT PROCEEDINGS
                          Coordinated by:

               Science Applications International Corporation
                     Ft. Washington, PA 19034

                      Contract No. 68-C2-0148
                     Work Assignment No. 1-1
                      Work Assignment Manager:

                         Emma Lou George
                  U.S. Environmental Protection Agency
                 Risk Reduction Engineering Laboratory
                        Cincinnati, OH  45268
             RISK REDUCTION ENGINEERING LABORATORY
              OFFICE OF RESEARCH AND DEVELOPMENT
             U.S. ENVIRONMENTAL PROTECTION AGENCY
                       CINCINNATI, OH 45268
                                                   Printed on Recycled Paper

-------
                                 FOREWORD
      Today's rapidly developing technologies and industrial practices frequently carry
with them the increased generation of materials, that if 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 is responsible  for  planning,
implementing,  and managing research, development, and demonstration programs to
provide an authoritative, defensible engineering basis in support of the policies, programs,
and regulations of EPA with respect to drinking water, wastewater, pesticides, toxic
substances, solid and hazardous waste, and Superfund-related activities. This publication
is one of the products of that research and provides a vital communication link between
researchers and users.

      These Abstract Proceedings from the 1993 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.
                           E. Timothy Oppelt, Director
                      Risk Reduction Engineering Laboratory

-------
                                   NOTICE
      The abstracts presented in these Proceedings (with the exception of those from
the Hazardous Substance Research Centers) have 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.
                                      in

-------
                                  ABSTRACT
      The Nineteenth Annual Risk Reduction Engineering Laboratory (RREL) Hazardous
Waste Research Symposium was held in Cincinnati, Ohio, April 13-15,1993. 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
and five Hazardous Substance Research Centers.

      These Proceedings are organized into three sections.  Sections A and B contain
extended abstracts of the paper presentations.   Section C  contains abstracts of the
poster displays.  Subjects include  remedial action, treatment, and control technologies
for waste disposal, landfill liner and cover systems, underground storage tanks, municipal
solid  waste and residuals  management  and  demonstration and  development of
innovative/alternative treatment technologies for hazardous waste. Alternative technology
subjects include pollution prevention, thermal destruction of hazardous wastes, field
evaluations, existing treatment options, emerging treatment processes, and biosystems
for hazardous waste destruction.   Drinking water treatment/management, corrosion,
organics removal,  health  effects,  and  ultrafiltration   are  also  addressed  in  the
presentations.
                                       IV

-------
                           ACKNOWLEDGEMENTS
      The 19th Annual RREL Hazardous Waste Symposium was planned in partial
fulfillment of Contract No. 68-C2-0148, Work Assignment No. 1-1 by Science Applications
International Corporation (SAIC) under sponsorship of the U.S. Environmental Protection
Agency. Emma Lou George of the Risk Reduction Engineering Laboratory (RREL) was
the Work Assignment Manager responsible for coordinating this project. The conference
program  and activities  were planned by  a committee consisting of the following
individuals:  Emma Lou George, Randy Parker, Jeff Adams, and Lou Garcia of RREL and
Lisa Kulujian of SAIC.  Special acknowledgement is given to Thomasine Bayless and
Robert Stenburg of RREL for reviewing and coordinating this proceedings and to Jo-Ann
Hockemeier and Denise Rambo of SAIC for coordinating many of the Symposium
activities.

-------
                             TABLE  OF CONTENTS

                                       Session A

Evaluation of In Situ Subsurface Bioremediation with Laboratory Microcosms, Laboratory
Macrocosms, and In Situ Macrocosms
       Wendy J. Davis-Hoover, U.S. EPA, RREL 	
Chlorinated Aliphatic Hydrocarbon Biodegradation by Methanotrophic Bacteria
       Perry L. McCarty, Stanford University	
Development of Bio-Scrubber for Removing Hazardous Organic Emission from Soil, Water, and
Air Decontamination Processes
       Paul T. Liu, Aluminum Company of America	13

On-Site Chemical Analysis:  The Key to Most Cost-Effective Remediation and Decision-Making
       Albert Robbat, Jr., Tufts University	18

Field -Experiences with the EPA/RREL Volume Reduction Unit
       Rich A. Griffiths, U.S. EPA, RREL 	21

Three-Dimensional Characterization of Groundwater Contamination
       Dennis McLaughlin, Massachusetts Institute of Technology	27

Pneumatic Fracturing at Low Permeability Formations
       John R. Schuring, New Jersey Institute of Technology	32

Field Study of In-Situ Trichloroethylene Degradation in Groundwater by Phenol-Oxidizing
Microorganisms                                ,
       Lewis Semprini, Stanford University	37

Evaluation of an Ultrasonic Cleaning System as a Replacement for CFC-Based Solvents in a Metal
Parts Cleaning Operation
       Paul M. Randall, U.S.  EPA, RREL	42

On-Site Solvent Recovery
       Ivars J. Licis, U.S. EPA, RREL	46
                                                                                        I '
The RREL Lead Paint Abatement Program
       George T. Moore, U.S. EPA,  RREL  	50

Innovative Clean Technologies Demonstration
       Kenneth R. Stone, U.S. EPA, RREL	53

Reduction of Chemical Waste  Assessment of Selected Overhaul/Repair Processes
       Tom T. Walker	56

Demonstration of The Toronto Harbour Commissioner's Soil Recycling Project
       Teri L. Richardson, U.S. EPA, RREL  	61

Results of the EPA MITE Program: Evaluation of Landfill Mining Technology
       Lynnann Hitchens, U.S. EPA, RREL	67
                                            VI

-------
Automatic Sortation Process for Post-Consumer Plastic Containers
       Diana R. Kirk, U.S. EPA, RREL	72

Two Treatment Options for Chromated Copper Arsenate Wood Preserving Residues
       Ronald J. Turner, U.S. EPA, RREL 	77

Discovery and Development of Anaerobes for Bioremediation
       James M. Tiedje, Michigan State University	84

Features of BTEX-Degrading Microorganisms from Oxygen-Limited (Hypoxic) Environments
       Ronald H. Olsen, University of Michigan Medical School	 89

Effects of Nitrogen Source on Crude Oil Biodegradation
       Brian A. Wrenn, University of Cincinnati	94

Field Treatability Trials for Fungal Treatment of Soils Contaminated with Wood Treating Waste
       John A.  Glaser, U.S. EPA, RREL	99

Aerobic, Phenol-Induced Biotransformation of TCE
       Gene F. Parkin, The University of Iowa	,104


                                        Session  B

Hazardous Waste Incinerator Emissions Resulting from Waste Feed Cutoffs
       Marta K. Richards, U.S. EPA,  RREL	109

Metals Emissions/Control from the Burning of Hazardous Waste
       W. Randall Seeker, Enery and Environmental Research Corporation	 114

Prediction of Dust Generation  from Handling Powders in Industry
       Marc Plinke, University of North Carolina at Chapel Hill	119

Soil Detoxification Using Solar Technology: The Result of EPA's Part of the Tri-Agency Effort
       Paul Gorman, Midwest Research Institute	124

Comparison of Thermal Treatment PCB Data from Two Superfund Sites
       Larry R. Waterland, Acurex Environmental Corporation  	129

SITE Demonstration of Chemical Waste Management "X*TRAX" Thermal Desorption Process
       Paul R. dePercin, U.S. EPA, RREL	134

Evaluation of Corrosion Inhibitors to Reduce Lead in the Drinking Water in Buildings
       Thomas J. Sorg, U.S.  EPA, RREL			137

Point-of-Entry Drinking Water Treatment System Applications at Hazardous Waste Sites
       James A. Goodrich, U.S. EPA, RREL	.142

The Removal of Ionic Contaminants from Drinking Water
       Thomas F. Speth, U.S. EPA, RREL	 153

A Closed Loop Approach for Removal and Destruction of Chlorinated Aromatics from Soil
       Shubhander Kapilla, University of Missouri-Rolla	 158
                                            VII

-------
Feasibility of Removing Uranyl, Thorium and Radium from Kaolinite by Electrokinetics
       Yalcin B. Acar, Louisiana State University	161

Evaluation of the Biogenesis Soil Cleaning Process for Organic Contaminants
       Annette Gatchett, U.S. EPA, RREL	166

Pollutant Fluxes to Aquatic Systems via Coupled Biological and Physicochemical Bed-Sediment
Processes
       L.J. Thibodeaux, Louisiana State University  	,	168

Laboratory and Modeling Investigations of Surfactant Enhanced Aquifer Remediation
       Linda M. Abriola, The University of Michigan	173

Procedures for Selection and Design of a Cap for In-Situ of Treatment of Contaminated Sediment
       Danny D. Reible, Louisiana State University	177

Optimization of Soil Vapor Extraction for Remediation of Gasoline-Contaminated Soil and Ground
Water
       Chien T. Chen, U.S. EPA, Releases Control Branch	182

SITE Demonstration of Resource Conservation Company's B.E.S.T. Process
       Mark C. Meckes, U.S. EPA, RREL	188

Characterisation and Treatment of Contaminated Soils Using Mineral Processing Techniques
       Peter Wood, Warren Spring Laboratory	192

Hydraulic and Impulse Fracturing to Enhance Remediation
       Larry Murdoch, University of Cincinnati	197

An Integrated Chemical and Biological Treatment (CBT) System for Site Remediation
       Robert L. Kelley, Institute of Gas Technology	202

SITE Demonstration of the Chemical Waste Po*WW*ER™ Evaporation and Catalytic Oxidation
System
       Randy A. Parker, U.S. EPA, RREL	,206

Bioremediation of Chlorophenol Contaminated Aquifers
       Kenneth J. Williamson, Oregon State University 	211

Preliminary Assessment of Fate Mechanisms of Chlorinated Organic Compounds in a Sediment-
Water Estuarine System
       Spyros G. Pavlostathis, Georgia Institute of Technology	216

Transport and Entrapment Behavior of Immiscible Organic Waste Chemicals in Heterogeneous
Aquifers: Implications on Modeling, Recovery and Remediation
       Tissa H. Illangasekare, University of Colorado	220


                                          Posters

A Beach Microcosm for the Study of Oil Biodegradation
        Kevin L. Strohmeier, University of Cincinnati	224

Acid Extraction Treatment System
        Steven W. Paff, University of Pittsburgh Applied Research Ctr.	225
                                             VIII

-------
An Evaluation of Drinking Water Samples Pre-Disinfected with Chlorine Dioxide Using
Chemical, Microbiological and Mutagenicity Endpoints
       Kathleen Schenck Patterson, U.S. EPA, RREL .	226

Anaerobic Bioprocessing of Nitroaromatic Compounds
       Dane Higdem, J.R. Simplot Company	227

Application of Computer Simulation for Introduction of Pollution Prevention in Industry
       Jordan Spooner, U.S. EPA, RREL	228

Density-Driven Advection During Laboratory Leaching Tests
       W.W. Slack, University of Cincinnati	229

Electron Beam Treatment for Superfund Site Remediation
       W.J. Cooper, Florida International University	230

Electronic Component Cooling Alternatives:  Compressed Air and Liquid Nitrogen
       Johnny Springer, Jr., U.S. EPA, RREL	231

Evaluation of Corrosion  Rates of Lead Pipes and Copper Pipes by Polarization Measurement
       Marvin Gardels, U.S. EPA, RREL	232

Evaluation of Material Recovery Facilities for Municipal Solid Waste Recycling
       Lynnann Hitchens,  U.S. EPA, RREL	233

Fluid Extraction-Biological  Degradation of Polyaromatic Hydrocarbons From Town Gas Soils
       David Rue, Institute of Gas Technology  	234

Microbial Removal of Lead from Contaminated Soil
       Wendy Davis-Hoover, U.S. EPA, RREL	235

Nutritional Requirement and Buffering Capacity of Various Phosphate Sources for Oil Degrading
Cultures  in Sea Water
       Edith L. Holder, University of Cincinnati  	236

Operating Parameters to Minimize Emissions During Rotary Kiln Emergency Safety Vent
Openings
       Paul M. Lemieux,  U.S. EPA, AEERL  	237

Perox-Pure(TM) Chemical Oxidation Treatment
       Norma M. Lewis, U.S. EPA, RREL	238

Predicting Resistance of Chemical Protective Clothing Using Neural Computing Techniques
       William G. Wee, University of Cincinnati	239

Predicting Solidification/Stabilization Performance Using Neural Computing Techniques
       William G. Wee, University of Cincinnati	240

RD&D Under the Tidewater Interagency Prevention Program (TIPPP)
       Kenneth R. Stone, U.S. EPA, RREL	241

RREL Remedy Screening Program
       Eugene Harris,  U.S. EPA, RREL  	242

RREL Underground Storage Tank Research Program Overview
       Anthony N. Tafuri, U.S. EPA, RREL	243

                                             ix

-------
Remediation of Solids with Toxic Organics and Metals - The Mel DAS Process
       Srivats Srinivasachar, PSI Technology Company	245

Reverse Osmosis:  Pesticide and Organics Removal from Drinking Water
       Carol Ann Frank, U.S. EPA, RREL	246

Risk Reduction Engineering Laboratory (RREL), Drinking Water Technology Transfer Activities -
U.S. EPA
       Walter A. Feige, U.S. EPA, RREL 	247

SOWAT:  Sequence Optimizer for Wastewater Treatment
       William G. Wee, University of Cincinnati	248

Site Remediation Technical Resource Documents
       Ben Blaney, U.S. EPA, RREL  	249

Slow-Release Oxygen Source for Bioremediation in Subsurface Soils
       Stephen J. Vesper, University of Cincinnati	250

The Leachate-Recirculating Municipal Solid Waste Landfill
       Debra R. Reinhart, University of Central Florida  	251

The U.S. EPA Incineration Research Facility
       J.W. Lee, Acurex Environmental Corporation	252

Thermally Induced Water Movement Within Geomembrane Covered Clay Liners
       Robert Norton, Iowa State University  	253

Use of Secondary Lead Smelting Technology for the Reclamation of Lead from Lead-Containing
Superfund Sites
       Steven W. Paff, University of Pittsburgh Applied Research Center	254

X-Ray Treatment of Organic Wastes  in  an Aqueous Matrix
       Randy Curry, Titan/Pulse Sciences, Inc	255

-------
   EVALUATION OF IN SITU SUBSURFACE BIOREMEDIATION WITH LABORATORY MICROCOSMS.
                   LABORATORY MACROCOSMS. AND IN SITU MACROCOSMS
                                 Wendy J. Davis-Hoover, Ph. D.
                         US EPA, Risk Reduction Engineering Laboratory,
           Municipal Solid Waste Research Management Branch, Soils & Residuals Section,
                           5995 Center Hill Ave., Cincinnati, Ohio 45224
                           PHONE: 513-569-7206  FAX: 513-569-7879

                      Stacy Pfaller, K. Pete  Paris and Stephen J. Vesper, Ph.D.
              University of Cincinnati, Department of Civil and Environmental Engineering
                           5995 Center Hill Ave., Cincinnati, Ohio 45224
                           PHONE: 513-569-7898 FAX: 513-569-7445
INTRODUCTION
    Bioremediation of contaminants from subsurface soil is a natural but often slow process (1).
Frequently, the public and the polluter prefer this method of treatment to limit exposure and for financial
considerations.  However the natural process is often too slow to allow for its use in reducing risks of
contaminant exposure. We have been evaluating methods to enhance the rates of subsurface
bioremediation.  We are specifically developing ways to enhance in situ subsurface bioremediation
because this eliminates many risks of exposure and is more cost effective than on site bioremediation.
    In order to evaluate enhancers of subsurface bioremediation, we need to have an  effective model of
in situ bioremediation. Currently, researchers are using several models including small microcosms,
columns of loose soil  and  even soil slurries.  One example is the small serum bottle microcosm model
(2). However, for in situ  subsurface bioremediation, many contaminated soils  have a moisture content
around 19-26 percent, a temperature of 12 degrees C, and they still have their soil structure. This work
demonstrates our ventures to model the environment of an ecosystem of an intact soil so that we may
accurately study the effects of enhancement of bioremediation rates.
METHODOLOGY
       LABORATORY MICROCOSMS

   Laboratory microcosms were prepared as described (2) with the following modifications.  In
summary, 10 grams of native Center Hill soil from the same soil horizon (BJ as the intact soil core in the
macrocosms (as discussed below) were aseptically weighed and added to sterile 60 ml serum bottles,
stoppered with sterile 1 cm thick butyl rubber stoppers.  Sterile deionized water was added to the same
concentration of moisture as the laboratory macrocosms stated below. These experiments were
performed with  at least three replicates. The microcosms were incubated at 12 degrees C in the dark
(covered  with aluminum foil). After acclimation of 14 days, each microcosm was spiked with .00155 ml
of 100 %  propylene giycol (PPG) (a surrogate contaminant) (in 1:1 deionized water) /10 grams of soil,
recapped, covered, and reincubated at 12 degrees C. This day was  defined as day zero.

       LABORATORY MACROCOSMS

   Laboratory macrocosms were fabricated out of stainless steel to contain a mold volume of 0.01415

-------
 cubic meters (0.5 cubic feet). They were designed to vary the oxygen level (by changing the oxygen
 concentration in the water added to the bottom of the chamber), pressure, and temperature and to
 collect samples of carbon dioxide. Both intact and packed soil cores were studied in these chambers.
    An intact core was removed  in one piece at the B1f B.,, and Bx horizons (7.6 cm, 12.7 cm, and 7.6
 cm, respectively (3 inch, 5 inch,  and 3 inch, respectively)) of Center Hill Soil Site BT1. After the first
 experiment, the soil core was wrapped in sterile saran wrap 5 times to simulate the constraints on
 oxygen concentrations in the subsurface soils and thus minimize the effects of atmospheric oxygen on
 the edges of the soil core.  The soil core was then placed into the macrocosm and the top was sealed.
 The bottom of the core sits on a stainless steel support grid that is kept in contact with a 2.5 liter
 reservoir. The reservoir was maintained at 2-8 ppm dissolved oxygen by recirculation thru silicone
 tubing (Masterflex).  Incubation was at 12 degrees C, in the dark, for 7 to 14 days for acclimation.
    The packed soil core was made from the surrounding soil at the same horizon as the intact soil core.
 It was packed to the same density as the intact soil with a targeted bulk density (dry) of 1508 kg/m3
 (94.1 pounds/ cubic foot).  It was packed in a 19 L (5 gallon) plastic bucket. The bucket was cut away
 from the column, and the column was shaved down to the appropriate size for the macrocosm. This
 core was handled as described above for the intact core.
    After this time of acclimation, both cores were spiked with PPG at the same concentration as the
 microcosms. This was done with a perforated tube (a 6mm (1/4 inch),  open at the bottom and
 perforated for the  first  12.7 cm (5 inches)), which was driven 21 cm (7 inches) into the center of each
 macrocosm.  The second experiment required the use of a pump at the rate of 15  ml/hour. The tubes
 were removed, the holes grouted, the macrocosms were sealed and this day was called day zero.

       IN SITU MACROCOSMS (SUBCOSMS)

    In situ macrocosms were developed to study subsurface bioremediation "in the original place". They
 were made of stainless steel with a rounded top, 20.3 cm  (8 inch) sides, 2:2.9cm (9 inch) diameter open
 bottom, a valve at the top for carbon dioxide collection, and an opening at the top for PPG spiking and
 temperature probe. Five to 10 cm  (2 to 4 inches) from where the macrocosm soil was taken, we
 removed soil down to the interface of the B, and Ba horizons (61 cm from the surface (24  inches)).  We
 moistened the soil overnight and pushed the four subcosms into place.  Two subcosms were spiked
 with PPG (same concentration) using a perforated tube as described above.  The other two subcosms
 remained unspiked as a negative control. This day was defined as zero. The temperature probes and
 tubing for carbon dioxide collection were added to the subcosms and the subcosms  were  covered again
 with soil to the surface of the surrounding intact soil.

       SACRIFICING OF THE CHAMBERS

   Although generated carbon dioxide was collected from the macrocosms and subcosms about every
 other day, the other parameters were evaluated only at the end of the experiment, the point of sacrificing
 of the chambers. The soil from the microcosms was sampled by taking a grab sample.  Soil from the
 macrocosms were analyzed by taking 9 corings in the shape of a +  (one in the center (position A), 2
 north (one next to the center core (Position B, direction  N), another next to that but farther from the core
 (Position C, direction N),, two south (etc.),  two east, and two west).  Each core was then aseptically
divided into 9 parts and every 3rd part was analyzed for: 1. PPG, 2. FDA and soil moisture, and 3. soil
pH, degrading, and total organisms. Three composite samples were taken from the remaining soil.  The
subcosms were dug up, turned over, cored and analyzed as described above for the  macrocosms.
criterion:
       CRITERION OF SUCCESS: The presence of bioremediation was determined with the following
              Microbial Activity
                     Carbon Dioxide Generation:  Carbon dioxide generation was not measured in

-------
                             the microcosms. Carbon dioxide was pushed out of the macrocosms
                             with nitrogen, and after day 40 pulled out of the subcosms with a 1 L
                             vacuum flask at a vacuum of ^380 mm of mercury and regulated with
                             needle valves.  After trying several methods (carbon dioxide results for
                             macrocosm experiment 1 is  not discussed  here due to unreliability of
                             carbon dioxide measurements.) , we used the method of Haines and
                             Molsny (3).  Briefly, the carbon dioxide was collected in a series of 5
                             test tubes of 3-20 ml of KOH (usually 0.06 M) with alizarin red S
                             indicator. The pH and temperature of the KOH was measured, and the
                             results placed in the computer program of  Haines and Molsny (3) to
                             obtain the milligrams of carbon dioxide produced.
                      Total culturable organisms: Growth on R2A performed as per Dang et al. (4).
                      Total degrading bacteria: MS and PPG: Growth on  Minimal Salts medium in the
                             presence of PPG was performed as described by Vesper et al. (5).
                      Fluorescein Diacetate hydrolysis Activity (FDA): This was performed on soil
                                            samples as described by Schnurer and Rosswall (6).
              Concentration of Contaminate: PPG: The soil samples were analyzed for the presence
                      of PPG using  the technique of Vesper et al. (5).
              Soil Moisture: Soil moistures were obtained by using a standard EPA test procedure (7).
              pH: Soil pH was determined  on  the soil samples using the method described in (7)
   The results were analyzed by using a General Linear Models Procedures from Statistical Analysis
System (Gary, NC) (SAS-PC).
RESULTS

       LABORATORY MICROCOSMS

               Microcosms seem to differ greatly from the soil environments from which they came.  As
evidence, FDA activity and the numbers of PPG degrading bacteria are greater, compared to packed
core microcosms. Compared to intact soil core macrocosms, microcosms seem to alter the
environment in all parameters tested. The microcosms have more degrading bacteria, more FDA
activity, higher pHs, but drier conditions.

       LABORATORY MACROCOSMS

               Bioremediation of PPG was evaluated in packed versus intact soil columns.  As
expected, the soil pH and moisture in the two columns were not statistically significantly different (at 0.05
level). Microbial activity expressed in the FDA procedure and the numbers of PPG degrading and total
bacteria showed no significant differences in the two columns.  The level of contamination with PPG was
also statistically significantly decreased at the 0.0001 level in the packed compared to the intact soil.
The packed column had  more total numbers of bacteria and more PPG.
               In another trial, bioremediation of PPG was evaluated in packed versus intact soil
columns. Again as expected, the soil moisture in the two columns were not statistically significantly
different (at 0.05 level), however neither was the concentration of PPG. Microbial activity expressed in
the FDA procedure, the total  numbers of bacteria, and the PPG degrading bacteria, showed significant
differences  (0.0001, 0.01, and 0.01 level )in the two columns. However, the soil pH  was also statistically
significantly different at the 0.0001 level. The packed column had more FDA activity, more total and
degrading bacteria and a higher pH.
               An in another trial, as expected the soil pH and moisture in the two columns were not
statistically significantly different (at 0.05 level). Microbial activity expressed in the numbers of PPG
degrading bacteria show no significant differences in the two columns nor did the levels of PPG.

-------
 However, microbial activity indicated by the numbers of total and degrading bacteria was statistically
 significant at the 0.01 and  .001 level (respectively) and the FDA activity at the .0001 level.  This indicates
 that the packed column had more activity. The packed column had more numbers of total and
 degradative bacteria and higher levels of FDA activity.
               The carbon dioxide generation data can be seen in Figure 1 and demonstrates that
 there appealed to be no difference in packed compared to the intact soil core in the amount of carbon
 dioxide generated.

        IN SITU MACROCOSMS (SUBCOSMS)

    In comparing PPG spiked  subcosms to unspiked control subcosms, there was no statistically
 significant differences in the number of total or PPG degrading bacteria, nor in the FDA level.  The
 amount of PPG was, as expected, significantly different, but so was the moisture (.001) and pH (0.01
 level). The spiked soils had fower moisture and higher pH.
    The carbon dioxide generation data (see Figure 2) demonstrates that the subcosms that received the
 PPG responded by producing more carbon dioxide.  The increased activity after day 40 was probably
 due to the change in the collection procedure (the vacuum was applied).
    The spiked subcosms seem to be similar to packed microcosms in the area of number of total
 bacteria, FDA activity, PPG levels and pH. They are significantly different in the areas of numbers of PPG
 degrading bacteria (0.05 level) and moisture (0.01 level). The spiked subcosms have more degraders,
 and are wetter.  In comparison to the intact macrocosms, the spiked subcosms are not different in the
 number of degrading bacteria, FDA activity, and PPG concentrations. They are significantly different in
 the total number of organisms (0.01 level), moisture (0.001) and pH (0.01).  The spiked subcosms have
'greater total numbers of bacteria, less moisture, and  a higher pH than the intact soil macrocosm.

        COSTS

    The costs for the microcosm are much less than the costs for either the macrocosm or the
subcosm. However, it is impossible to  obtain enough sample from 1 microcosm to perform the analysis
.that can be done from either the macrocosm or the subcosm, thus 170 more microcosms have to be
"tested at one time.  Although the macrocosms are slightly larger than the subcosms, there is enough soil
to perform the analysis.  The estimated cost to build a stainless steel macrocosm is $1,200.00;
subcosm, $500.00; and the'glass microcosms, about $2.00 apiece. The subcosms have an economic
advantage in that they do not need to be held in an electronically controlled temperature room
 (refrigeration).
'CONCLUSIONS
    In examining the soil moisture and soil pH results, the differences were statistically significant but
these parameters were so stable that the differences do not appear to be biologically important. This
must be considered when evaluating these results.
    The microcosms appear to simulate a different environment than the macrocosms or the subcosms.
Their increased activity may be due to an increased oxygenation since the soil was disrupted and has
no structure.  The macrocosm results indicate that the rate of bioremediation in an intact soil core is less
than that in the packed soil column. If the intact core simulated the true subsurface soil and its
ecosystem, these results would indicate that using packed soil cores or microcosms would inflate
projected rates of bioremediation.
    Although the bottom  of the subcosm is completely open, the generation of carbon dioxide in the
spiked subcosms is comparable to the generation in the macrocosms (Rgures 1  and 2). This would
indicate that the bacteria, gases,  and contaminants are not lost to the soil below.  Although we could
                                             4

-------
expect to not have bacterial movement on a large scale in the intact soils, the loss of the gases and
contaminants might be expected. This is encouraging as in situ microcosms have advantages in
studying contaminated soils at a contaminated site.
REFERENCES
1. Leahy, J. G. and Colwell, R.R. Microbial Degradation of Hydrocarbons in the Environment. Microbiol.
       Reviews 54(3):305-315,1990.

2. Sinclair, J. L, and Lee, T. R. Biodegradation of Atrazine in Subsurface Environments. EPA/600/S-
       92/001, U.S. Environmental Protection Agency, Cincinnati, Ohio, 1992. 8 pp.

3. Haines, J.R., and D. Molsny. Calibrating Aqueous Equilibrium for KOH, CO2. Draft USEPA Risk
       Reduction Engineering Laboratory report.

4. Dang. J. S., Harvey, D. M., -Jaobbagy, A. and Grady Jr., C.P.L Eyaluation of biodegradation kinetics
       with respirometric data. J. Water Pollut. Control Fed. 61:1711-1721, 1989.

5. Vesper, S. J., Davis-Hoover, W. J., and Murdoch, L C.  Submitted for publication.

6. Schnurer, J. and Rosswall, T. Fluorescein diacetate hydrolysis as a measure of total microbial activity
       in soil and litter. Appl. Environ. Microbioi. 45:1256-1261, 1982.

7. United States Environmental Protection Agency, Office of Solid Waste and Emergency Response. Test
       Methods for Evaluating Solid Waste, Physical/Chemical Methods, Third edition, SW-846,
       September 1990.
FOR MORE INFORMATION
                                  Wendy Davis-Hoover, Ph. D.
                         US EPA, Risk Reduction Engineering Laboratory,
           Municipal Solid Waste Research Management Branch,  Soils & Residuals Section,
                           5995 Center Hill Ave., Cincinnati, Ohio 45224
                           PHONE: 513-569-7206  FAX: 513-569-7879

-------
                               FIGUEE 1.

       INTACT VERSUS PACKED SOIL CORES: IMPACT ON C02 GENERATION
           2500
        =  2000
        Q  1500
        x
        g
           1000



            500



              0
o
£D
CC

O






1


I  -500
        O
            	 Packed, exp. 3

            	Intact, exp. 3

            	 Packed, exp. 2

            	 Intact, exp. 2
              -40     -20     .0       20       40


                             DAYS AFTER CONTAMINATION
                                               60
50
                                FIGURE 2.


IN SITU MACROCOSMS:  CONTAMINATION'S IMPACT ON COS  GENERATION
          2000
        e
        o
       =   1500
           1000
       g   500
       o
          -500
                                             CONTAMINATED

                                            • UNCONTAMINATED
              20    30     40     50     60    70     80


                              DAYS AFTER CONTAMINATION
                                                  90
 100

-------
       CHLORINATED ALIPHATIC HYDROCARBON BIODEGRADATION
                        BY METHANOTROPHIC BACTERIA

                  Larry Smith, Thomas Hemysson, and Perry L. McCarty,
                   Western Region Hazardous Substance Research Center
                             Department of Civil Engineering
                                  Stanford University
                             Stanford, California 94305-4020
     Methanotrophic bacteria, which oxidize methane for energy, have been found capable of
oxidizing chlorinated aliphatic hydrocarbons (CAHs) by cometabolism. The goal of this project is
to obtain a better basic understanding of the relationship between the relative concentrations of
methane and CAHs, and the overall CAH transformation rates. There are many current attempts to
devise treatment processes for degrading CAHs. However, there is a lack of basic understanding
of the factors affecting reaction rates for cometabolized compounds. It has been hypothesized that
oxidation rates for methane and CAHs can be described by a competitive inhibition model. Here,
the rate limiting step is the oxidation of either methane or CAHs by methane monooxygenase
(MMO). To evaluate this hypothesis, reaction coefficients for methane and TCE alone are being
evaluated'using a mixed methanotrophic culture originally derived from the Moffett Field aquifer.
Also, reaction rates and possible competitive inhibition between different CAHs that are likely at
contamination sites is being evaluated

METHANE/TCE INTERACTIONS

     TCE transformation rate by the methanotrophic mixed culture was found to be enhanced by
the presence of a high concentration of methane (5-6 mg/L) at a high TCE concentration (5-9
mg/L). This result was contrary to the expectation that the presence of methane, through
competitive inhibition, would cause a reduction in the TCE transformation rate due to the strong
affinity  of MMO for methane (Ks = 0.03 mg CH4/L). Additional experiments were conducted at
similar methane and biomass concentrations, but with  lower TCE concentration (0.9 mg/L). In
these experiments, the rate of TCE transformation was found to be lower in the presence of
methane (Figure 1), thus suggesting the competitive inhibition does exist.  It appears here that the
fraction  of the total energy reserves consumed and the fraction of cells inactivated were lower than
in the high TCE case, thus diminishing the importance of these factors and apparently leaving
competitive inhibition as the dominant process.  The processes occuring and their interrelationships
are obviously much more complex than originally anticipated.

Mixed  Culture Growth Conditions

     A  two-reactor treatment system for TCE destruction by methanotrophic bacteria is being
evaluated (Figure 2) (Alvarez-Cohen and McCarty, 1991). The system consists of a growth
reactor which continuously produces the active culture for addition to a TCE-contaminated waste
stream in a separate transformation reactor.  This reactor configuration eliminates TCE
transformation product toxicity during cell growth, benefits from the controlled growth conditions
in the growth reactor, and benefits from the absence of competitive inhibition and more favorable
reaction rate kinetics of a plug flow transformation reactor rather than a completely mixed reactor.

     Two aspects of growth reactor performance were studied, reactor retention time and nutrient
concentration. First, 2-, 4-, and 9-day residence times (SRT, solids retention time, days) were
studied.  The results (Table 1) indicate that although the shorter SRT's afforded higher growth
yields, the TCE transformation activities were significantly diminished, leading to lower TCE
transformation yields at the shorter SRT's.

-------
      The second operational factor studied was nitrate (N) concentration in the nutrient medium
 fed to the growth reactor. Many organisms are known to be capable of storing energy by
 producing polymers such as poly-beta-hydroxybutyrate (PHB) from excess carbon source when a
 shortage of a required nutrient occurs. Two reactors were operated at 2-day SRT, one of which
 was fed growth medium with  1/2 as much nitrogen as used in the standard growth medium. The
 9-day growth reactor operation was unchanged and used as a reference case. The results of the
 growth reactor data and TOE transformation characteristics are summarized in Table 2 and indicate:

      •  N starvation resulted in a much higher growth yield in the 2-day SRT reactor.
      •  Comparison of the 2-day SRT reactors indicates that N starvation improved the Ty.
      •  Both 2-day SRT reactors had much higher Tc than observed in the earlier 2-day reactor
         experiment described above and in Table 1.
      •  All three reactors had high TC by historical standards for this culture.
      •  TCE transformation by the 9-day SRT reactor culture was impaired by formate addition, a
         result contrary to previous results and literature reports, but consistently reproduced in
         this culture between June, 1992 and the present.
      •  The activity of the mixed culture toward TCE increased dramatically during 1992 (Tc
         increased by a factor of almost 3 over the former typical value of 0.036 g/g for the culture
         without formate), showing that the the activity of a mixed culture with time can be quite
         variable.
 PHB

     The TCE transformation capacity with resting (not fed methane or formate) methanotrophic
 cells has been suggested to be a function of the cells poly-b-hydroxybutyrate (PHB) content
 (Henry and Grbic-Galic, 1991).  PHB serves as an internal source of reducing power that can
 maintain the activity of MMO for TCE epoxidation. The PHB content of methanotrophs has been
 shown to increase under nutrient deficiency.  This study was instigated in order to determine more
 explicitly whether by increasing the PHB content of cells, a higher transformation capacity (Tc)
 could be obtained for resting cells.

 Procedures

     Mixed culture from the various CSTRs described above were used in these studies. Cell
 concentration was determined by absorbance measurements at 600 nm, and calibration with dry
 mass determinations as described above. PHB content was quantified by filtration of cells,
 digestion with sodium hypochlorite, extraction with chloroform, washing and drying of residue,
 heating in the presence of sulfuric acid to hydrolyze PHB, and measuring absorbance at 235 nm
 after the procedure by Ward and Dawes, Anal. Biochem. 52: 607 (1973).

     Methanotrophs have two forms of MMO, a particulate form and a soluble form. The soluble
 form was found in previous studies to be the active form for TCE transformation by the culture
 used here. Soluble MMO can oxidize naphthalene to napthanol, which can be readily measured by
 absorbance at 530 nm.  The relationship between cell PHB content and the rate of napthalene
 oxidation was determined in batch culture.  TCE oxidation rates as functions of PHB content were
 also determined by the batch method as described above.

     In order to manipulate the PHB  content in cells over a broader range than was available with
the CSTR cultures, a small sample of culture was incubated over night under batch conditions with
different methane-to-nitrogen ratios. Here, bottle headspace contained 15% methane, and the
solution nitrate concentration was varied among the several bottles used.

-------
Results                                                              '

     The PHB content of the 9-day CSTR (designated LMA) was measured over a 40 day period,
and was found to be highly variable, ranging from 2 to 8% of cell dry weight.  Cell density over
this period was relatively constant (2,500 to 3,100 mg/1). The naphthalene oxidation rate by
resting cells was found to be directly proportional to PHB content (Figure 3), thus suggesting the
importance of the internal energy reserves on the rate obtained. Another finding was that the
naphthalene oxidation rate per unit mass of cells varied little and was essentially independent of
reactor operation or cell PHB content if 20 mM sodium formate was added.  This addition, which
provided a readily-available external source of reducing power,  always resulted in the highest
naphthalene oxidation rates. The coefficient of variation of naphthalene oxidation rate in the
presence of 20 mM formate was only 15% over a two-week period.

     Cells were grown in batch cultures with different nitrate contents. Cultures with lower nitrate
content produced cells with higher PHB content as anticipated. Naphthalene oxidation rates by
cells with and without formate addition are summarized in Table 3.  Oxidation rates with low-
PHB-content cells were only 12% of the rate when formate was present, but were 64% of the rate
with formate when using high PHB content cultures.

     TCE transformation capacity was determined as a function of cell PHB content, with and
without 20 mM formate addition.  The results in Table 4 indicate the TC was indeed higher in
resting cells with higher PHB content The results also demonstrate that formate is a better source
of reducing power, and when it is  present, the TC of cells is about the same. Previous studies
under this project have similarly demonstrated that maximum TC was obtained in the presence of
formate.  The exception to this is the currently operated 9-day SRT reactor, which has an especially
high TC without formate, and a value with formate that is lower, but typical of that from past
performance.

Discussion

     These results indicate that Tc, and the related value, Ty, are functions of two primary factors.
First is the availability of reducing power to supply the energy needs for regeneration of MMO.
This reducing power can be supplied to resting cells by internal energy reserves, such as PHB, or
by external sources, such as formate.  However, if sufficient reducing power is available to cells,
then TC is limited by transformation product toxicity.  That is, cells are killed by products of TCE
oxidation, and for this reason, a given mass of cells can only transform a certain amount of TCE
before they are expended. TC and Ty can thus be increased by increasing either the internal or
external energy reserves, or, in some manner, reducing the toxicity of the transformation products.
Internal reducing power reserves can be increased through nitrogen starvation, but this obviously
has a limit as nitrogen  starvation can reduce bacterial growth rates and cellular yields.
REFERENCES

Alvarez-Cohen, L. and P. L. McCarty. 1991. Two-Stage Dispersed-Growth Treatment of
      Halogenated Aliphatic Compounds by Cometabolism. Environmental Science and
      Technology, 25(8): 1387-1393.

Henry, S. M. andD. Grbic-Galic. 1991. Influence of Endogenous and Exogenous Electron
      Donors and Trichloroethylene Oxidation Toxicity on Trichloroethylene Oxidation by
      Methanotrophic Cultures from a Groundwater Aquifer. Appl. Environ. Microbiol., 57(1):
      236-244.

-------
 ACKNOWLEDGEMENTS

      This study 'was supported by a Swedish Forest and Agricultural Research Council post
 doctoral fellowship to T. Henrysson, by the Gas Research Institute through a subcontract with
 Radian Corporation,  and by the Westinghouse Savannah River Company, all through the EPA
 sponsored Western Region Hazardous Substance Research Center under agreement R-815738. It
 has not been subjected to Agency review and therefore no offical endorsement should be inferred.
 Table 1. Summary of results of study of the effect of growth reactor SRT on culture TCE
          transformation characteristics.
Culture characteristic
Tc (mg TCE/mg ceUs)
Ty (mg TCE/mg CH4)
Cell concentr. (mg TSS/L)b
Y (mg cells/mg CELO
Growth Reactor SRT (days)
2
+formate
0.031
0.012
-formate
0.028
0.011
1450
0.40
4
+formate
0.009
0.003
-formate
0.006
0.002
1350
0.37
9
+formate
-formate
>0.080a
>0.02ia
2200
0.26
aTotal extent of TCE transformation capacity not determined due to complete transformation of
 TCE before exhaustion of transformation capacity.

     concentration measured as total suspended solids concentration.
Table 2. Summary of results of study of the effect of growth reactor nutrient nitrogen feed
          concentration on culture TCE transformation characteristics. Mean values for three
          experiments.
Culture characteristic
Tc (mg TCE/mg cells)
TC standard deviation
Ty (mg TCE/mg CH4)
Cell concentr. (mg TSS/L)a
Y (mg cells/mg CH4)
Growth Reactor SRT / NO3-N concentration
2 d/ 165 mg N/L
+formate
0.128
0.021
0.031
-formate
0.077
0.010
0.018
1150
0.24
2d/82mgN/L
+formate
0.122
0.035
0.061
-formate
0.060
0.010
0.030
1260
0.50
9d/165mgN/L
+formate
0.058
0.005
0.015
-formate
0.102
0.034
0.026
2720
0.25
aCell concentration measured as total suspended solids concentration.
                                          10

-------
Figure 1. Experimental results showing evidence of competitive inhibition
          at low TCE concentration.
                 Liquid Recycle
                                  Cell Separator
                                                 Treated
                                                 Effluent
                                             Transformation
                                                Reactor

                                               Plug Flow
               Microbial Growth Reactor
                    Complete Mix
                                        Influent Wastewatcr
Figure 2.  Schematic of proposed treatment system.
                       11

-------
                 0.41
                 0.0
y = O.OS1 + O.OSSx  R*2 = 0.92
                                   2      34      5
                                      PHB-contenf: (%)
                                  6 -    7
Figure 3. Correlation between the PHB content and the normalized naphthalene oxidation
         rate.  The naphthalene oxidation rate without formate was normalized to the rate
         with 20 mM formate in each sample.
Table 3.  Napthalene oxidation rates and PHB-contents after cell incubation in batch
          reactors at different methane-to-nitrogen ratios.
Addition
Nitrate
None
PHB
content
0.9%
7.4%
Naphthalene oxidation
rate (u>mol/mg«dav)
Without With
formate formate
038
1.36
3.12
2.14
Normalized
rate
0.12
0.64
Table 4. TCE transformation capacities for cultures from reactors with different operating
         conditions and different PHB contents.
               Reactor         PHB           TCE transformation

                             content          , capacity (mg/mg)

                                           Without         With

              	formate	formate
                uas
                LMA
0.7%

4.7%
0.056

0.091
0.11

0.12
                                   12

-------
      DEVELOPMENT OF BIO-SCRUBBER FOR REMOVING HAZARDOUS ORGANIC EMISSION
                 FROM SOIL. WATER AND AIR DECONTAMINATION PROCESSES

                              R. L. Gregg, H. Sabol and Paul K.T. Liu
                              Media and Process Technology Group
                                 Aluminum Company of America
                                      1135 William Pitt Way
                                     Pittsburgh,  PA 15238
                                        (412) 826-3711
INTRODUCTION

   Biofiltration, in its most general sense, is the removal and decomposition of contaminant from gases into
nonhazardous substances through the use of micro-organisms. Bio-filters are believed to be the most
economical way to treat the low level contaminants (up to several thousand ppm) in gas streams.

  For efficient operation, the filter media must meet several requirements:

       Provide optimum environmental conditions for the resident microbes
       Consist of uniform pore size and particle structure (for low bed pressure drop, minimizing gas
       channeling, high reactive surfaces)
       Minimal bed compaction (minimize maintenance, media replacement)

  Composition of an existing commercially available biofilter, generally satisfies the first requirement by
providing sufficient nutrients for the micro-organisms (typically bacteria), except for particularly refractory
contaminants (i.e., chlorinateds) (1). The problem with composting, however, is the huge space
requirement compounded by continual loss of effective surface area during biomass build up (slothing).

  An activated carbon-based biofiltration module, a bio-scrubber, has been developed to improve the
existing bio-filtration systems. These synthetically produced filters address the current deficiencies of
composting and other naturally occurring media-based biofiiters. Its advantages are:

       Very high surface area to bed volume ratios
       Low pressure drops
       Minimal pressure drop loss due to slothing
       Much smaller bed requirements (allows the use of compact filters only)
       High water retention in the microporosity (long shelf life while not in use, during start up/shut
       down, minimal requirement for additional water addition)

   In addition, activated carbon media beds provide one more key separation mechanism for biofiiters,
adsorption of gases onto the carbon. This provides the following advantages:

       Increased surface concentration of contaminants
       Removal of hydrophobic gases that would not typically be absorbed into the aqueous phase
       Allow the biofilter to be efficient at higher concentrations of contaminants

   The above attributes also could result in enhanced biodegradation of substances that would not
typically be efficiently biodegraded in a biofilter, providing additional applications for the technology.
                                             13

-------
METHODOLOGY

  A bench-top bioscrubber testing unit including the biofilters and gas supply, was assembled in the lab.
Five glass columns with 1" ID and 2'L were packed with selected activated carbons ( US Mesh 12x30) as
filters. Three sampling ports were installed along the length of the column for gas sampling and pressure
drop measurement, as shown in Figure 1. Air containing 10 ppm of toluene was prepared by diluting a
custom-premixed canister containing 500 ppm of toluene with additional air.  Flow rates ranging from 0.5
to 2 liter/min were controlled with MKS mass flow meters and controllers. Both feed and effluents were
sampled with a Precision gas tight 1 ml syringe, then analyzed by GC (Model 3400 Varian) with a 8' x 1/8"
stainless steel 60/80 Carbopack B, 1% SP-1000 column, with a detection limit of 0.1 ppm.  Pressure drop
was measured with a Monagahelic pressure gauge (0 to 100" water). Any excess biomass was removed
as required by manually removing, gently washing and replacing the carbon bed. Inorganic nutrient,
required for biological growth, was fed to the column down f tow at a 0.1 mL/hr rate.

RESULTS

  The columns have been consistently degrading the contaminant for a period of 11  months. They have
achieved a > 95% removal efficiency within the first 5 to 10 inches of the carbon bed, creating a stationary
mass transfer zone with an empty bed contact time (EBCT) of 1 to 4 seconds depending on flow. This
unusual performance indicates the effectiveness and efficiency of the bio-scrubbers developed.

  Three columns have been operating since 3/23/92 until the present (February 1993).  All columns were
fed with 0.5 liters/minute of air containing 10 ppm of toluene as a target concentration from 3/23 to
9/10/92. The actual feed concentration fluctuated from 10 to 20 ppm for most of  that time. During this
period, no toluene breakthrough was observed at Port A, indicating the effective  mass transfer zone was
less than 5 inches. More importantly, the mass transfer zone remained stationary for the entire time
period.  Biodegradation of toluene evidently was effective and complete, showing no signs of
accumulation of contaminants or of the metabolic by-products.

  After the successful demonstration of the concept, several additional operating conditions were
studied. The flow rates for Columns A and B were increased to 1  liter/minute and then 2 liters/minute
while Column C remained at the original flow of 0.5 liters/minute to act as a control. The EBCT under the 2
liter/minute flow rate is about 2  seconds.  During this period (9/10/92 to 10/9/92), both Columns A and B
show some breakthrough ranging from 0 to 5 ppm (Figures 3 & 4). No toluene was detected at Port B in
each column. The effective mass transfer zone was estimated to  be about 7.5 inches and remained
stationary for the entire period.  Column B was further challenged  by increasing the flow rate to 4 liters/min,
equivalent to 1 second of EBCT from 10/9/92 to 1/21/93. No contaminant breakthrough at Port B was
observed for the majority of the experimental period. In certain instances, i.e., on December 30 and
January 18,1993, trace breakthroughs were observed, but the column rapidly recovered to its typical
efficiency, and the breakthroughs were possibly due to channeling of the flow. The mass transfer zone
was estimated to be approximately 10 inches at this flow rate. The flow rate of Column B next was reduced
to 0.5 liters/minute; no toluene breakthrough was detected at Port B as had been observed previously.
The recovery of the column to the original mass transfer zone indicates that the increase from 5 to 10" is
possibly due to the degradation kinetics vs. linear velocity of the contaminant. Therefore, the mass
transfer zone is concluded to be stationary throughout the entire study.

  The biofilter seemed able to absorb the fluctuation of the feed ranging from 5 to 15 ppm most of the
time.  The fluctuation observed in the feed was not reflected in the effluent measured at Ports A and B,
indicating that activated carbon effectively acts as a sink to adsorb the fluctuation.

  The biomass generated and accumulated in the filter was expected. Biomass  was visually detected
occupying the interparticle space. This build-up would essentially result in a pressure drop increase.
                                              14

-------
Periodic removal of the biomass manually was practiced to maintain a minimal pressure drop throughout
the operation period. While the excess biomass was removed from the column, sufficient amounts of
biomass were retained on the carbon to maintain effective biodegradation when the bed was replaced.
The biofilter efficiency was not reduced due to the biomass removal.

  A pilot-scale test unit, capable of handling 4 CFM of flow, was designed and constructed. It is a stand-
alone unit with compressor, backwashing capabilities, and an inorganic nutrient supply/recycle system.  A
pilot test is currently underway at a customer site utilizing a discharge stream containing a low level BTEX
concentration. Results will be reported when available in the future.

CONCLUSION

  The engineered biofilter, bioscrubber, has been tested on a bench-top scale for about 11 months. It
consistently demonstrates an effective and efficient removal for a low level organic contaminant, toluene,
in air. Column length ranging from 5 to 10 inches was required to confine the wavefront of the filter for the
empty bed contact times ranging from 1 to 4 seconds.  The continuous bio-regeneration on the carbon
has been demonstrated consistently allowing the use of a shallow bed. The feed fluctuation was
cushioned effectively by the use of the active media as a sink. A pilot study is being conducted at an oil
refinery site  to demonstrate the technology/process.

REFERENCES

1.      Leson, G. and Winer, A. M. Biofiltration: An Innovative Air Pollution Control Technology for VOC
        Emissions. Air and Waste Management Assoc. 40 (8Y. 1045-1054
 FORE MORE INFORMATION:

               EPA Technical Manager:
Ms. Naomi Barkley
Environmental Protection Agency
Risk Reduction Engineering Lab
Cincinnati, OH 45268
(513) 569-7854
                                              15

-------
                                  2 Inchei I.D.
                Influent
             Zone 6-6 Inches
                                                  Total column
                                                  length • 12
                                                  Inches
                 Figure 1. Schematic of Bench-top Bio-Scrubber Unit
•«• PORT A/Col C     S POFTTB           -^ PORTC           * INFLUENT

Figure 2: Removal of 10 ppm Toluene From Air With A Bioscrubber- Column C
                                    16

-------
           53533?^
           ,i ,i i  i i .1 i j i i  i L^^^^^^^^^^^I j i J»J,iJ.J_ JL i JL ^L *LiL,fcL^^.^^ «L «L «i^ «L «L A A A A A A *\ /% A A i i J* i i «•••«*»•*
-e-PORT A              •&  PORTS              ^ PORTC              * INFLUENT


    Figure 3: Removal of 10 ppm Toluene From Air With A Bioscrubber-Column B
           0   ITiTTTi ITI'I i'tii n 1111111111111111111111111111II11111111111
               c-<^aoog^a. njaoi vr •
               c^rfT^-^icscffii t* t» t« ^rooor^tNc^iqi ^ ^i ^cMr^^Or^i-^c^KSSi tii | • ^ ^rocp' t" t* t« |csr
       •«• PORT A             -B-PORTB              ^ PORTC              » INFLUENT


       Figure 4: Removal of 10 ppm Toluene From Air With A Bioscrubber- Column A
                                             17

-------
                               ON-SITE CHEMICAL ANALYSIS:
       THE KEY TO MOST COST-EFFECTIVE REMEDIATION AND DECISION-MAKING

                                       .Albert Robbat, Jr..
                 Chemistry Department, Trace Analytical Measurement Laboratory,
                             Tufts University, Medford, MA 02155
                                         617-627-3474
Introduction
        Much is speculated about hazardous waste site assessment and cleanup costs. For example, some
have projected that by the year 2020 over 60,000 hazardous waste sites will require cleanup by state and
federal agencies with the U.S. Environmental Protection Agency (EPA) managed sites costing in excess
of $150 billion andttie Departments of Energy and Defense (DOE and DOD) facility and landfill cleanups
estimated to exceed $2 trillion and $100 billion, respectively (1-3).  These efforts require a great deal of
analytical data to  support site  assessment and  remediation decision making.   For instance, site
investigation, management and control activities depend heavily on obtaining chemical information about
the site with analytical costs accounting  for up to 70% of the initial site investigation and 50% of the
overall cleanup costs. DOE estimates that it will spend between $15-45 billion on analytical services over
the next 30-years; with current sample volume for hazardous waste constituent analyses exceeding 400,000
samples/year and some of the larger DOE sites  spending as much as $10 million annually (4).

       The current approach of collecting field samples and sending them to the laboratory for chemical
analysis is time-consuming, inefficient, costly and can pose significant risk to workers and the community.
The typical data turnaround time  for EPA's Contract Laboratory Program (CLP) exceeds 90-days, while
private sector purchased laboratory services can surpass 30-days unless steep sample surcharges are paid,
i.e., 500% to 100% for 2-day  to 14-day turnaround times.  Based on this approach, on-site decision
making is non-existent  and, the  time it  takes between listing a "potential" hazardous site and actual
cleanup, unreasonable.  Sample manipulation during collection, transport and storage can  affect sample
integrity and thus,  analyte concentration; for  example, for  volatile organic  compounds  (VOCs)  the
discrepancies between higher field  detected concentrations and laboratory analyses is well-recognized
(5,6).

       Field analytical  methods range from  using  laboratory  instruments  and EPA  standardized
procedures operated in  transportable, temperature  and clean air controlled laboratories  to screening
methods employing hand-held  vapor sniffers which provide non-specific compound detection.  These
extremes in field analytical capability trade off measurement accuracy, precision and time for cost.  Much
has been written about the advantages of providing on-site detection based, in large part, upon speculation
and intuitive arguments with little published  documenting data quality between  field and laboratory
measurements for either small or  large data sets (7-13).  Nonetheless,  Cornell (7) has suggested that on-
site screening analyses  can provide cost  savings of up to 70% when compared to current geotechnical
engineering  practices.  -Despite all  of the seeming  advantages embodied in rapid on-site detection  of
pollutants, little progress has been made with regard to their routine acceptance.

       Toward this end, we have developed field-practical thermal desorption gas chromatography/mass
spectrometry (TDGC/MS) methods  that can provide screening level and more quantitative  analyses (14-
17).  The instrument, powered by battery in the field or electrical service was operated without any special
housing conditions.  In these studies, selected ion monitoring  (SIM) mass spectrometry provided highly
reliable screening level data in < 3-min/sample and more quantitative information in 10 to 15-min/sample
when operated  in either the SIM or total ion  current/selected ion extraction  (TIC/SIE)  modes.  For
                                               18

-------
example, poorer  SIM measurement  precision,  ~ 40%  RSD,  was found for polycyclic aromatic
hydrocarbons (PAH) as compared to TIC/SIE, < 30% RSD; with the latter providing data consistent with
CLP requirements. In contrast, the SIM provided both screening level and quantitative analyses for PCBs.
We have developed methods for VOCs as well as phenols, phthalates and chlorinated pesticides, studies
include:

       •       field TDGC/MS (Bruker Instruments) technology and methods have been demonstrated
               through EPA's  Superfund Innovative Technology  Evaluation programs at  several
               Superfund sites with interlaboratory studies accomplished through the cooperation of the
               Region 1, Hazardous Waste Division (see references).

       •       rapid screening TDGC/MS analytical data for ~ 200 samples over 6-days was to evaluate
               data quality in the service mode and to provide contaminant profiles at the Yellow River
               Road Superfund site, Baldwin, Florida in cooperation  with RSKERL-Ada, OK.

       •       2-month study, Fort Devens, MA, in which 300 samples were analyzed by screening and
               quantitative TDGC/MS analysis to determine data quality over an extended period of time
               during actual SI/RI investigations.  A site visualization software package provided 3-D
               contaminant profiles. Work is in progress to assess costs and to provide a case study that
               will begin to quantitate field  implementation and  benefits in  support of the  Army's
               USATHAMA activities in contrast to the current hand-waving arguments made by many
               as to the "perceived vs real"  benefits.

       •       analyzed over 200 VOC samples in 10-days  (actual  volume limited by constraints in
               sample collection). The data was used by Argonne National Laboratory to characterize
               a site  in 3.5-weeks.  Earlier site characterization using the so-called "phased approach"
               conducted by ANL at another site required 3-months and 3 trips to get to the same point
               as the Tufts/Argonne expedited site characterization study.

       •       analyzed over400 samples during a PCB  excavation project for Tenneco in 5.5-weeks.
               Initial engineering estimates indicated 4-month time period for cleanup.  We were asked
               to  conduct  a technology  study  after  the 8-months to  determine data  quality  and
               cost/benefits in which the engineering firm estimated an additional 5-months to complete
               cleanup.  On-site analysis in conjunction with real-time decision making  resulted in
               completion of the cleanup effort in 5-months.

       These and other studies will be presented to document cost-effectiveness when interactive sample
collection, analysis and decision-making are employed to provide expedited site characterization.
References

1)     Hazardous Waste Remediation: The Task at Hand, The University of Tennessee,
       Waste Management Research and Education Institute Knoxville, Tennessee.
2)     Hazardous Waste: DOD Estimates for Cleaning Up Contaminated Sites Improved
       but Still Constrained, GAO/NSIAD-92-37.
3)     Hazardous Materials Control Research Institute, Focus, February 1992, pp. 11.
4)     Laboratory Management Division, "Analytical Services Program Five-Year Plan,"
       Office of Environmental Restoration and Waste Management (EM) Department of
       Energy, January 29, 1992.
                                              19

-------
5)     Hewitt, A.D., Mlyares, P.H., Leggett, D.C., and Jenkins, T.F., Envir. Sci. & Tech., 26(10), pp.
       1932-1938.
6)     Hewitt,  A.D., "Review  of Current and Potential  Future  Sampling  Practices  for  Volatile
       Organic Compounds in  Soil", Proceedings  from  the  16th Annual Army  Environmental
       Research and Development Symposium, June 23-25,  1992, Williamsburg, VA.
7)     Cornell, F.W., "Site Investigations: The Role of Field Screening and Analysis
       Devices, "Proceedings of R&D 92 National Research and Development Conference
       on the Control of Hazardous Materials, February 1992, pp. 187-192.
8)     Fribush, H.M., Environmental Testing &  Analysis, 1992, May/June, pp.35.
9)     Mardsen, P., Amer. Environ. Lab., October 1990, pp. 42-45.
10)    Fribush, H., and Fisk, J.F. Amer. Environ. Lab., October 1990, pp.29-33.
11)    Williams, L.R., Amer. Environ. Lab., October 1990, pp. 4-5
12)    Williams, L. R. Proceedings for The First and Second International Symposium "Field Screening
       Methods for Hazardous Wastes and Toxic Chemicals", February 12-14, 1991, Las Vegas, NV.
13)    Chudyk, W., Environ. Sci. Technol, 1989, 23(5), pp. 504-507.
14)    Robbat, A. Jr., Liu T-Y and Abraham B.  M., Anal Chem, 1992, 64, pp 358.
15)    Robbat, A. Jr., Liu T-Y and Abraham B.  M., Anal Chem, 1992, 64, pp 1477.
16)    Robbat, A. Jr., Liu C-G, Liu T-Y, J. Chromatography, 1992,  625, pp 277.
17)    Robbat, A. Jr., Liu C-G, Liu T-Y, J., and Abraham A., submitted  Environ. Sci. and Technol.,
       1993.
                                              20

-------
              FIELD EXPERIENCES WITH THE EPA/RREL VOLUME REDUCTION UNIT
                                       Richard A. Griffiths
                              Risk Reduction Engineering Laboratory
                              U.S. Environmental Protection Agency
                                   2890 Woodbridge Avenue
                                   Edison, New Jersey  08837
INTRODUCTION
EPA personnel completed two major field experiments in 1992 on soil washing using the Risk Reduction
Engineering Laboratory's Volume Reduction Unit (VRU).  The first was conducted at a former wood-
treatment site in Pensacola, Florida, and the second was at a pesticides-contaminated site near Denver,
Colorado.
The VRU is a mobile, pilot-scale soil washing system that consists of process equipment and support
utilities mounted on two trailers.  The process trailer contains equipment for soil feeding, organic vapor
recovery, soil washing and screening, gravity separation, flocculation, and water clarification. An oil-fired
boiler on the process trailer supplies steam for thermal desorption of VOC's. The utility trailer contains
water storage tanks, a water heater, water filters, and carbon adsorption drums for recycling the process
water. An air compressor, small electric generator, and equipment storage compartments are on this
trailer. The VRU is described in more detail by Masters et al. (1).
THE ESCAMBIA WOOD-TREATMENT SITE
The wood-treatment site is the former Escambia Treating Company facility in Pensacola. The soil is very
sandy, with less than 5% silt- and clay-sized particles.  The primary contaminants are pentachlorophenol
(PCP) and creosote. The site has been excavated to groundwater. More than 250,000 cubic yards of
soil has been stockpiled on tarps adjacent to the excavation.


In late  1991, the EPA's Environmental Response Team (ERT) arranged bench-scale tests for soil washing
and showed that PCP and carcinogenic creosote compounds could be reduced to target levels (30-ppm
PCP, 50-ppm carcinogenic creosotes, 100-ppm total creosote). Total creosotes were not reduced to
target levels; however, the results showed sufficient promise that the ERT decided to proceed with pilot-
scale tests to obtain better performance and cost information.


RREL and ERT personnel  with support from Foster Wheeler Enviresponse Corporation and Roy F.
Weston Company conducted 20 pilot-scale tests at the site during  July 1992.  Only the soil handling and
soil washing/screening subsystems of the VRU were used. Wastewater/slurry was collected in plastic
tanks for future tests of UV/peroxide and other treatment processes.
                                               21

-------
Method

Soil was taken from a location several meters below grade on the side of the excavation, mixed using a
Bobcat and shovels, and hand-screened to remove components larger than 0.25 inch.


Tergitol NP-10 nonionic surfactant by Union Carbide was used, based on its success in the bench-scale
tests and its low cost.


Each test Included a 30- to 60-minute stabilization period followed by 120 minutes of run time. A typical
run consumed 300 pounds of soil and 240 gallons of water.  Eight-ounce samples of the feed soil and
the washed soil and one-liter samples of the wastewater/slurry were collected every 30 minutes.


The VRU does not currently perform a final rinse of the washed soil.  To evaluate rinsing, a sample of
washed soil was placed on a 120-mesh screen over a beaker, and approximately one liter of clean water
was poured through the sample.  This additional step was made for five runs.


Results

The mixed feed soil proved to be very consistent, with average concentrations of 150 ppm PCP, 71 ppm
carcinogenic creosote, and 1200  ppm total creosote.


The test results are summarized in Table  1. The contaminant removals were generally high, and the
cleanup criteria were easily met.  The target residual of 30 ppm for PCP was exceeded by a large
margin. The target residuals of 100 ppm total creosote and 50 ppm carcinogenic creosotes were also
exceeded with room to spare.


Of the primary independent process variables, the concentration of surfactant had the  greatest effect on
removal of creosotes.  The pH of the wash solution had the next greatest effect on removal of
creosotes.  The effects of temperature and  liquid:solid ratio were minor.


The pH of the wash solution had  the greatest effect on removal of PCP, but the effect was not strong.
The other variables had insignificant effects on the removal of PCP.


Rinsing after washing proved to be highly beneficial, reducing residual PCP below detection limits in all
five cases and reducing carcinogenic creosotes below detection limits in four out of five.  Total
creosotes were also significantly reduced in all five cases.


The process slurry stream was found to contain 1% to 2% solids. Most contaminant levels were
nondetect or below linear detection limits.  These levels are  one to two orders of magnitude lower than
expected. We speculate that the contaminants were tightly  bound to the fines in the slurry and were not
easily extractable, and QA checks support this hypothesis.
                                                22

-------
The VRU ran smoothly with only a few operational problems during the course of the study.  The most
important was foaming of the wet solids exiting the 100- mesh screen when 0.2% or more surfactant was
used at ambient temperature.  This caused poor dewatering by the vibratory screens. Excess water in
the washed soil increased from approximately 10% by weight of the soil stream to 50% to 110% during
the problematic runs.
THE SAND CREEK PESTICIDES SITE
The pesticides-contaminated site is the Sand Creek Superfund site in Commerce City, Colorado.  The
soil contains approximately 31% fines smaller than 200 mesh.  The contaminants are heptachlor, dieldrin,
aldrin, DDT, and numerous other common pesticides.


At the request of the Remedial Project Manager from EPA Region 8, RREL personnel with support from
Foster Wheeler Enviresponse Corporation conducted 23 VRU tests at the site during September 1992.
The Region tracked several contaminants, but chose to concentrate on heptachlor and dieldrin.  The
risk-based target levels were 0.3 ppm for heptachlor and 0.1 for dieldrin, but Region 8 also established a
variance for greater than 90% removal based on the planned use of the site.
Method

Soil for the tests was taken from several different depths.  Three surfactants were tested:
SDS (sodium dodecylsulfate)                 anionic
Adsee 799 + Witconol NP-100  (50/50)        nonionic
Tergitol NP-10                               nonionic
Witco Chemical
Witco Chemical
Union Carbide
A mixture of 66% SDS and 34% Tergitol was also tested.
VRU screens with 60 and 200 mesh were selected. However, all soil larger than 200 mesh was
composited after the run and analyzed as "coarse" fraction.  Except for a few "control" runs, all runs
used pH 10 and 130 °Fwashwater.


Results

The test conditions and results of washing the coarse fraction are summarized in Table 2.  Contaminant
concentrations in the feed soil varied widely, because the level of contamination varied significantly with
depth.


The risk-based target levels were not met in any of the runs. The 90% removal target was met in several
runs. A wash plus rinse typically produced 95% removal.
                                              23

-------
Use of surfactant improved removals, though the effect was not so great as it was in the Pensacola
tests. The pH of the wash solution had no effect The data are not sufficient to show the effect of
temperature.


The wastewater/slurry was found to contain 20% to 22% solids. Since 31% of the feed was fines, a
substantial amount of fines must have remained in the +200-mesh fractions.
Again, foaming was the most important operational problem.
REFERENCES
Masters, H., B. Rubin, R. Gaire, and P. Cardenas. EPA's Mobile Volume Reduction Unit for Soil Washing.
jn: Proceedings of the Seventeenth Annual RREL Hazardous Waste Research Symposium, Remedial
Action, Treatment, and Disposal of Hazardous Waste, EPA/600/9-91/1002,1991
FOR MORE INFORMATION:
       telephone:
Richard A. Griffiths
Releases Control Branch (MS-104)
U.S. Environmental Protection Agency
2890 Woodbridge Avenue
Edison, New Jersey 08837
908-321-6629
                                              24

-------
 in
•P
 in
 o>
EH

 (0
 E
 (D
 O
 in
w

 m
rH
 O
 o

 in
 c
 0)
 (1)
X!
•P
 W
-P
01
0)
 (C
CO
 Q)
       o\o
       i—i

       co
       o
       S
       w
       CO
       Q
       H
       CO
       H
       CO
      EH
      H
      Q

      O
      U

      EH
      CO
      W
      EH
    C
   •H
 (U  U
 I \  £_l
 O  (C
 w u
 o
 0) rH
 r<  (0
U 4->
    O
   EH
   04
   U
   04
    O
 (1)  M
-P  (0
 O  U
 tn
 o
 Q)  i-H
 S-i  n]
U  -P
    O
    EH
                 U
                 04
             K
              Cb
             -P
              O
              (C
             CO
             4J
              C
              0)
              o
             u
                 =»fc
                                     A     A     A
                                                                                                        A
                                     A     A     A
                         •  •   •   • c   •  c  •  c	c
                       cMcoeon    oj     n     r\iy3^j«n'*voni<4'
                                                                          C   •   •  •   •  C
                                                                             rg r^ o CN
                                                                                    H
                                                      in
                       H«Ht-HrHrHrHrHHH
                                                                             in            in
                                                                   HrHrH    HrHt-H
                       oinooooominooinoooinooininooininin
                       CMCocNOjrvjojrgeocoofNjeocMCMrsieoocNcoeoocNeococo
                       H    HHHt-HH        HH    HHH     HrH         rHH
                       '*«aJCV]rJCMrHrHHrHrHOOOOOOOO

                       ooooooooooooooooooooooooo
                                                         XJ X!
                              (C (T)     (d
                                 HrHCNCM        H     HHHrHHrH            H        H
                                                                     *
                                                                                                                     U  0)
                                                                                                                     M 4J
                                                                                                                     0)  O
                                                                                                                     A  U
                                                                                                                     -P  O
                                                                                                                     O  0)
                                                                                                                         ^
                                                                                                                     •H  O
                                                                                                                         n]
                                                                                                                     C-P
                                                                                                                     •H  o
                                                                                                                        -P
                                                                                                                     H
                                                                                                                     ••  e
                                                                                                                     vo  ft
                                                                                                                         a
                                                                                                                     w
                                                                                                                     > o
                                                                                                                     C H
                                                                                                                     3
                                                                                                                     rl   -
                                                                                                                         (/]
                                                                                                                     W  0)
                                                                                                                     •H 4J
                                                                                                                     X3  O
                                                                                                                     -P  U
                                                                                                                         O
                                                                                                                     C  0)
                                                                                                                     •H  M
                                                                                                                         o
                                                                                                                     T>
                                                                                                                     0)  U
                                                                                                                     W-H
                                                                                                                     3  C
                                                                                                                         0)
                                                                                                                     W  CT>
                                                                                                                     (0  O

                                                                                                                     *5
                                                                                                                     ^  o
                                                                                                                     H  M
                                                                                                                     ••  (0
                                                                                                                     fM  O
                                                                                                                     CO
                                                                                                                         a.
                                                                                                                         Ol
                                                                                                                  0)
                                                                                                                  w
                                                                                                                  c
                                                                                                                 •H
                                                                                                        (0
                                                                                                        rt   ^
                                                                                                           Ol
                                                                                                        •o u
                                                                                                        •H O4
                                                                                                        i-H
                                                                                                        O  g
                                                                                                        u  a
                                                                                                                     •H  o
                                                                                                                      3  in
                                                                                                                  O-H
                                                                                                  o
                                                                                                  fl)  X!
                                                                                                 4->  in
                                                                                                  a)  (a

                                                                                                 r,
                                                                                                  c
                                                                                                  II
                                                                                                                      0)  0)
                                                                                                                         0)
                                                                                                                         0)
                                                                                                                         C7>
                                                                                                                      II  
-------
                  0)
 W
4J
 W
 0)
 a)
 M
o
TJ
 (tJ
co
CO
Q
H
CO
                  O
                  (0
           Q)
           S3
                                              o\
                                                        CM
                                                                      CO CM
in CM co
                                          CM
                                                    'HHHHCMHCMCMH
o

 Q)
 O
           a)
                                                                   HHHHHHHHCNJ03HH
                                                                                                                 a)
u
 Q)
           o
           ro
          •P
                         Or>-OOOOOOOOCMCMOOOOOOOOOOOO
                                                                                                                 O
                                                                                                                 0)
  w
                                                                                         O  «J   *
                                                                                                                 Q)
                                                                                                                 (1)
                                                                                                                 Q)
                                                                                                              II  (fl
                                                                      HrHHHHHHOJCMCMCM
                                                                26

-------
                        THREE DIMENSIONAL CHARACTERIZATION OF
                              GROUNDWATER CONTAMINATION
                                      Dennis Mclaughlin
                                         LynnB. Reid-
          Ralph M. Parsons Laboratory for Water Resources and Environmental Engineering
                          48-209 Massachusetts Institute of Technology
                          Cambridge, MA 02139, USA (617) 253-7176
                                         Shuguang Li
                     Department of Civil  Engineering, Portland State University
                          Portland, Oregon 97205, USA (503) 725-5543

1. INTRODUCTION
Monitoring and remediation of groundwater  contaminant plumes is greatly complicated by the
heterogeneity of the subsurface environment and the cost of obtaining subsurface concentration
measurements. Recent field studies have demonstrated that the physical and chemical properties of
natural subsurface environments vary dramatically over surprisingly small distances. Such heterogeneity
causes groundwater to flow in tortuous paths which transport contaminants in complex ways.  Clearly, we
cannot expect to clean up contaminant plumes unless we can 1) locate them and 2) predict how they will
respond to remedial efforts.
The key to our approach to characterizing contaminant plumes is to combine information obtained from
groundwater models and field measurements. Although field measurements provide important
information about local conditions, they are  usually too limited to provide a good synoptic picture of
subsurface conditions. Conversely, models tell us much about the physical and chemical processes
which control solute movement but they generally do not account for the geologic heterogeneity
encountered at real field sites.  When models and measurements are combined we can obtain a
physically reasonable site characterization which reproduces the heterogeneity revealed by field
observations.

2. METHODOLOGY
Although the concept of combining measurements and model predictions sounds reasonable,  it may not
be obvious just how this should be done in practice. Here we adopt a statistical approach based on
Bayesian estimation theory (details are described in a series of  recent papers by Graham  and  Mclaughlin
(1,2,3) and Li and Mclaughlin (4)). Our approach assumes that random fluctuations in geological
variables such as hydraulic conductivity induce random fluctuations in dependent variables such as
groundwater velocity and solute concentration. The dependent variable fluctuations are not arbitrary but
are, rather, a result of the physical laws which govern groundwater flow and transport. These laws
enable us to  derive approximate expressions for the spatially-variable means and covariances of
hydraulic head, pore velocity and solute concentration (1,2). The statistical moments of solute
concentration provide a concise description  of subsurface conditions at a contaminated site.
                                              27

-------
When field data become available they may be used to update or condition the concentration statistics
derived from a stochastic transport model (1,2). The improved concentration estimate at a given time
and location is simply a weighted combination of the prior estimate and the  measurements. The
uncertainty in an estimated concentration value (as measured by its variance) decreases when it is
updated with new measurements. In practice, the degree of uncertainty reduction depends on the
measurement locations and on the heterogeneity of the site.  Maps of the concentration variance can be
used to guide the placement of future monitoring wells or indicate the degree of uncertainty about site
conditions.
The results presented in this abstract were generated with a new Bayesian estimation technique called
the nonstationary spectral method. This technique expresses random fluctuations in contaminant
concentration or hydraulic head in terms of a "transfer function" which generally depends on time,
location, and wave number (6). It is significantly more efficient than competing numerical methods such
as classical Kalman filtering or Monte Carlo simulation (5).

3. RESULTS

3.1  Description of Field Site
Over the past few years we have applied the site characterization procedure described above to a coal
tar disposal site in northern New York, U.S.A. The shallow unconfined groundwater aquifer at this rural
site is composed of unconsolidated glacial materials.  Groundwater discharge occurs to the east (about
400m from the contaminant source) into seeps feeding small tributaries of the Hudson River. The site is
contaminated with coal tar by-products leaching from several shallow deposits buried in the early 1960s.
We focus on naphthalene, primarily because of its relatively high solubility, low volatility and small
sediment sorption rate (4). The naphthalene plume appears to have reached a steady-state condition,
with chemicals leaching from the buried wastes at a rate comparable to the flux into the seeps.
The site investigation was performed by students from the Parsons Laboratory at the Massachusetts
Institute of Technology and two consulting firms specializing in subsurface investigations. In total, over
100 piezometers, 18 multi-level samplers, and 45 monitoring wells were installed at the site.  In
addition, 18 estimates of hydraulic conductivity were obtained from laboratory analyses and slug tests.
Figure 1 illustrates the vertical distribution of naphthalene measured in monitoring wells and multi-level
samplers along a cross-section near the plume centerline. The naphthalene plume sinks considerably
along  its length; it is clearly three-dimensional.

3.2 Modeling Results
Three-dimensional characterization results from the coal tar site are illustrated in Figure 2. Model
parameters for the three-dimensional flow and contaminant transport models were obtained by
assuming  a constant hydraulic gradient in space and time.  The spatially uniform prior mean of log
hydraulic conductivity was derived from site-wide averages; fluctuations about this mean were assumed
to be exponentially correlated.  The "prior" mean concentration plume obtained from an unconditional
stochastic analysis is smooth and regular, similar to plumes produced by deterministic models.
                                            28

-------
Figure 2 illustrates the updated estimate of the plume obtained after conditioning with 117 concentration
measurements. The plume now exhibits irregularities similar to those typically observed at
heterogeneous sites.  Li (6) presents a three-dimensional map of the concentration standard deviation
obtained before and after updating.  The updated uncertainty is dramatically reduced from the prior
case, reflecting the valuable contribution of field measurements

4. CONCLUSIONS
This abstract describes a new technique for characterizing contamination in heterogeneous aquifers.
The technique has been applied to a field site in New York.  The mean and variance of variables such
as contaminant concentration are obtained by combining Bayesian estimation concepts with a
computationally efficient numerical algorithm called the nonstationary spectral method.  The resulting
updated plume reflects the influence of point measurements yet remains consistent with the underlying
physical principles governing subsurface transport.
Application of the technique to a coal tar waste site provide three-dimensional maps of a naphthalene
steady-state plume. Remedial activities or complicated chemical behavior such as source removal,
extraction and injection wells, contaminant sorption and/or chemical transformations can all be included
in the stochastic approach outlined above; our future work will address many of these issues. We
believe that our coal tar application demonstrates that stochastic estimation methods can provide the
information needed to make groundwater remediation more effective.

5. REFERENCES
1.    Graham, W. and D. McLaughlin, Stochastic analysis of non-stationary subsurface transport 1. Un-
      conditional moments, Water Resources Research, 25(2), 215-232, 1989a.
2.    Graham, W. and D. McLaughlin, Stochastic analysis of non-stationary subsurface transport 1.
      Conditional moments, Water Resources Research, 25(11), 2331-2355,1989b.
3.    Graham, W. and D. McLaughlin, A stochastic model of solute transport in groundwater: Applica-
      tion to the Borden, Ontario tracer test, Water Resources Research, 27(6), 1345-1360,1991.
4.    Hyman, J., Three-dimensional characterization of groundwater contamination with the multi-level
      sampler, S.M. Thesis, Massachusetts Institute of Technology, May, 1990.
5.    Li, S.G. and D. McLaughiin, A non-stationary spectral method for solving stochastic groundwater
      problems, Water Resources Research, 27(7), 1589-1605,1991.
6.    Li, S.G. A non-stationary spectral method for solving stochastic groundwater problems, Ph.D.
      thesis, Massachusetts Institute of Technology, 1993.

6. FOR MORE INFORMATION
Professor Dennis McLaughlin
48-209 Parsons Laboratory,  Massachusetts Institute of Technology
Cambridge, MA 02139, USA
dennis@maya.mit.edu (Internet)
                                             29

-------
                                                        0)
                                                         re
                                             1
                                                         S
                                                         8
                                                         S
                                                        _a>
                                                        I
                                                        I
                                                        1
                                                         (0
                                                         o>
                                                         0>
                                                         O)
30

-------
                                            o


                                            J3

                                            Q.
                                           ,0)
                                            re

                                           f
                                            «J
                                            0)
                                            §•£

                                            O  Q.

                                            ** w
                                            *• T"
                                            o ,_
                                           IS

                                            co to"
                                            o o
                                           '•5 £
                                            ^ g
                                            J> o
                                            to A

                                            §2

                                           |f

                                            ™ E
                                           ¥ 2
                                            M

                                            OJ  u

                                            o  §

                                            =  V
                                            O)  O
                                            il J2
31

-------
  PNEUMATIC FRACTURING OF LOW PERMEABILITY  FORMATIONS
                         John R. Schuring and Paul C. Chan
                 Hazardous Substance Management Research Center
                         New Jersey Institute of Technology
                            Newark, New Jersey 07102

                     John J. Liskowitz and Conan D. Fitzgerald
                          Accutech Remedial Systems, Inc.
                              Cass Road at Route 35
                               Keyport, N.J.  07735

                                   Uwe Frank
                                    U.S. EPA
                             2890 Woodbridge Avenue
                          Edison, New Jersey 08837-3679

INTRODUCTION

       Pneumatic fracturing is an innovative technology which enhances the in situ
removal and treatment of volatile organic compounds (VOC's) in low permeability soil
and rock formations.  The process may be generally described as injecting air into a
contaminated geologic formation at a pressure which exceeds the natural in situ stresses,
and at a flow rate which exceeds the permeability of the formation. This causes failure of
the medium and creates a fracture network radiating from the injection point. Once
established, the fractures increase the flow rate of vapors or liquids through the
formation, and make the contaminants more accessible. The effect of pneumatic
fracturing on sedimentary rock formations such as sandstone and shale is shown
conceptually in Figure 1.

       The principal objectives of pneumatic fracturing are reduction in treatment time,
and extension of available technologies to more difficult geologic conditions. Pneumatic
fracturing is designed to be integrated with  other in situ treatment technologies such as
vapor extraction, thermal injection, and bioremediation. Initial applications focused on
enhancing treatment of the vadose zone, but recently the technology is being extended into
the saturated zone. The pneumatic fracturing system has also been modified to deliver
biological supplements (e.g. nutrients, buffers, and microorganisms) directly into the
fractured formation to enhance in situ bioremediation.

       During August 1992, the U.S. EPA sponsored a Superfund Innovative Technology
Evaluation (SITE) field demonstration of the Pneumatic Fracturing Extraction (PFE)
process at a site in Hillsborough, N.J. PFE is an integrated remedial process for cost-
effective removal and treatment of contaminants from geologic formations with low to
moderate permeabilities.  The demonstration was performed by Accutech Remedial
                                       32

-------
Systems, Inc. of Keyport, N.J., and the Hazardous Substance Management Research
Center at New Jersey Institute of Technology (NJIT) in Newark, N.J.  McLaren Hart of
Warren, N.J. provided site support, and served as project consultant. Science
Applications International Corporation (SAIC) of Paramus, N.J. performed the
technology evaluation.

       The project site, which was formerly occupied by a manufacturing facility, is
underlain by bedrock consisting fractured siltstone and shale known locally as the
Brunswick Formation. As a result of past industrial activities, trichloroethylene (TCE)
and other VOC's contaminated the bedrock forming a source area in the vadose zone. In
spite of the fractured nature of the bedrock, feasibility tests at the site showed the
formation permeability was very low, and conventional vapor extraction would be
ineffective.  Costly excavation and removal of the source area, or encapsulation, were the
options under consideration prior to the decision to apply pneumatic fracturing.
                                           AIR       Cyllnd«r«
                                          SOURCE  /  2. Compr*tior »/
                                                    Rectivtr Tank
                                                            SEDIMENTARY
                                                            ROCK
                                    SEE DETAIL A
                     BEFORE FRACTURE
                                            AFTER FRACTURE
            EXIST.
            DISCONTINUE!
            (TYP)
            JOINT FILLING
                                   JOINT FILLING
                                   PARTIALLY
                                   CLEARED
                                  DETAIL  "A"
                        EFFECT OF FRACTURING ON ROCK DISCONTINUITIES
   Figure 1.  Pneumatic Fracturing Concept in Sedimentary Rock Formations
                                        33

-------
METHODOLOGY

       The general evaluation approach was to monitor the changes in subsurface air flow
and mass removal rate which resulted from the pneumatic fracture injections. A central
well was installed which served as both the fracture well and the main vapor extraction
well.  It was surrounded by seven monitoring wells which ranged from 7.5 to 20 feet from
the central well. All wells were cased to a depth of about 8 feet, but were left uncased in
the treatment zone to assure maximum connection with the formation. All wells
terminated at a depth of about 20 feet which was approximately 5 feet above the ground
water table.  To examine the effects of geologic structure on fracturing, monitoring wells
were located with consideration to formation strike and dip.

       Prior to any fracturing activities, baseline behavior was established by extracting
air from the  central well and recording the flow rate and TCE concentration. Cumulative
samples were collected in tedlar bags and analyzed with a field gas chromatograph.
Extraction tests were run with the outlying monitoring wells capped to measure vacuum
radius of influence, and also with selected monitoring wells uncapped to evaluate the
effects of passive air inlet.  In addition, short duration extraction tests were run on each
monitoring well to evaluate air permeability, and to check for interconnectivity between
wells.

       A series of four fracture injections were then made at successively deeper levels
ranging from 9.0 to 16.4 feet below grade. Each injection lasted about 20 seconds, and
was applied  to a discrete two foot interval. The PFE system consists of an HQ injector
connected to a bank of compressed air cylinders via an automated manifold system. The
PFE system  delivers the injected air at controlled flow rates and pressures which are
adjusted according to geologic conditions and depth.

       During injection, pressure gages were mounted on the outlying monitoring wells to
monitor the  lateral extent of the fractures. Another method used to estimate fracture
radius was measurement of ground surface heave using electronic tiltmeters. Twelve
biaxial tilmeters were placed in a cross pattern  around the fracture well which were
interconnected to a PC controlled data aquisition system.  Direct examination of the
effects of fracture injections on the bedrock formation were made with a borehole video
camera. The condition of the fracture well was recorded with the camera before
fracturing, and again after fracturing.

       After completion of fracturing, extraction tests were again run at the central well
to evaluate changes in flow rate and TCE concentration under capped and uncapped
conditions.  Air flow rates at the outlying monitoring wells were also checked.  The results
from these tests were then compared with baseline values to evaluate the effects of
fracturing. The total duration of the PFE field  experiments was about two weeks.
Following the PFE tests, additional experiments were performed to investigate the effects
of Hot Gas  Injection (HGI), the results of which will not be discussed in this paper.
                                        34

-------
 RESULTS

        The post-fracture measurements at the central extraction well showed substantial
 increases in formation permeability and contaminant removal rate as a result of fracturing.
 Based on comparative extraction tests before and after fracturing, air flow for the capped
 well condition was increased by 8 times, and TCE mass removal increased approximately
 9 times. For comparative test runs made with selected monitoring wells uncapped
 (passive inlet), the improvements were even greater.  Extracted air flow rate increased
 about 175 times and TCE removal rate increases about 25 times compared with baseline
 values. It is noted that all results are preliminary as of 2/1/93, and the Technology
 Evaluation Report is due in Spring 1993.

        All of the peripheral monitoring wells exhibited improved vacuum communication
 with the central fracture/extraction well, and showed increased permeability. Test results
 for the monitoring wells are summarized in Table 1.  As indicated, sensible vaccum at
 most of the monitoring wells increased substantially.  Extracted air flows from the
 monitoring wells also increased: 5.9 to 15.1 times at a 10 ft. radius, and 3.1 to 12.0 times
 at a 20 ft. radius.  It is noted that baseline air flows for most of the wells were below the
 minimum reading, so actual change ratios were even greater.

        The monitoring well data suggest the effective radius of the fracturing was at least
 20 feet. Surface heave data recorded by the tiltmeters showed that fractures propagated
 up to 35 ft., but these larger radii were not confirmed with flow and vacuum
 measurements. A review of both' tiltmeter and monitoring well data suggest that the
 fracture orientation was predominantly horizontal, with a slight directional preference
 along the geologic strike of the formation. Overall, fracture patterns were relatively
 uniform, although some variations were observed with regard to depth and direction.
                  Table 1. Summary of Monitoring Well Results
Monitoring
Well No.
FMW-1
FMW-2
FMW-3
FMW-4
FMW-5
FMW-6
FMW-7
Distance To
Fracture Well
(ft.)
10
10
10
10
20
7.5
20
Geologic
Orientation
Strike
Off Strike & Dip
Dip
Strike
Strike
Dip
Dip
Vacuum Influence (in. H2O)
Pre-
fracture
7.5
4.5
23
14
0
44
0
Post-
fracture
92
93
92
92
77
90
81
Change
Ratio
+12.3
+20.7
+4.0
+6.6
(+77)
+2.0
(+81)
Extracted Air Flow (SCFM)
Pre-
fracture
0.62*
0.62-0.88
0.62*
0.62*
0.62*
0.88
0.62*
Post-
fracture
5.2-6.4
5.2-7.0
5.1-9.4
5.7-8.1
5.5-7.5
4.8-7.1
1.9-2.0
Change
Ratio
8.3-10.2
5.9-113
8.2-15.1
9.2-13.1
8.8-12.0
5.5-8.1
3.1-3.2
NOTES: 1. All values are for 136 in. H2O (10 in. Hg) vacuum applied at pump.
      2. (*) denotes flow gage was at minimum reading. Actual pre-fracture flow was less.
      3. Data for vacuum influence taken 30 minutes after start-up.
      4. All data is preliminary as of 2/1/93. Technology Evaluation Report is due out Spring 1993.
                                         35

-------
These variations were attributed to the geologic heterogeneities present in the formation,
and other influences such as surface structures and utilities.  Direct examination of the
pneumatically induced fractures with the borehole video camera confirmed that the
permeability enhancements were largely due to dilation of existing geologic
discontinuities, although some new fractures were also noted.

       Analytical test results of the extracted effluent showed unusally high
concentrations of several other chlorinated hydrocarbons and benzene which had only
been detected in trace amounts before fracturing. This result suggests that the pneumatic
fractures opened new pathways to pockets of contamination, in addition to increasing
formation air flow. It is noted that perched water was unexpectedly encountered in the
formation throughout the demonstration which may have adversely affected air flow, and
may also have removed TCE and other VOC's during dewatering control.
CONCLUSIONS

       The demonstration showed that PFE was very effective in enhancing the
permeability and VOC removal rate from a low permeability siltstone/shale formation. A
greatly extended vacuum radius of influence was also observed, which will result in a
reduction of the number extraction wells required to remediate the site. The application of
PFE should decrease remediation time, and in the case of this site, eliminate the need to
excavate or encapsultate the source area.
ACKNOWLEDGEMENTS

Grateful acknowledgement is given to the following research assistants for their significant
contribution to this project: Firoz Ahmed, Thomas Boland, Yuan Ding, Trevor King,
Deepak Nautiyal, Norman Ng, and Anthony Vandeven.
                                       36

-------
           Field  Study of In-situ  Trichloroethylene  Degradation in Groundwater
                           by Phenol-Oxidizing  Microorganisms

                       Gary D. Hopkins,  Lewis Semprini, and Perry L. McCarty

                      Western Region Hazardous Substance Research Center
                                 Department of Civil Engineering
                          Stanford University, Stanford, CA 94305-4020
                                        415-723-4131

INTRODUCTION

    Recent research has demonstrated that aerobic microorganisms grown on phenol or toluene can
initiate the cometabolic oxidation of chlorinated aliphatic compounds (CACs) to stable nontoxic
endproducts [1,2,3,4]. Such microorganisms possessing toluene oxygenase (TO) [2,4] have good
potential for bioremediating aquifers contaminated with CACs and their anaerobic and abiotic
transformation products.

    Previous evaluations at the Moffett test site of enhanced in-situ biodegradation of CACs focused on
stimulating indigenous methanotrophic bacteria through methane and oxygen addition [5]. In these tests
approximately 20% of the TCE added was degraded as *rt was transported through a two meter
biostimulated zone. TCE injection concentrations ranged from 50 to 100 ug/l of TCE. Recent in-situ
studies at the test site demonstrated that TCE was more effectively degraded by a phenol-utilizing
population [6].  Here 80% of the TCE added was degraded in the 2 meter biostimulated zone while
injecting 12.5 mg/l phenol, 35 mg/l DO, and 40 u,g/l TCE.

    The objective of the current work was to determine how effective the phenol-utilizers were at
degrading TCE over concentrations ranging from 62.5 to 1000 ug/l. In addition, microcosm studies were
performed to evaluate the process over a similar concentration range.


METHODOLOGY


    Microcosm studies were performed at the field site with columns fed oxygenated groundwater from
the test zone. The 1 meter, 3" I.D., columns were filled with 1/4" stainless-steel distillation packing as a
biological support medium. The columns were inoculated with microbes in groundwater from the test
zone. Five columns were batch fed every other day with 12.5 mg/l phenol, 30 mg/l DO, and TCE
concentrations ranging from 0 to 1000 ug/l. The concentrations of phenol, DO and TCE of the column
fluids were determined from samples collected during the column exchange using analytical methods
described previously  [6].

    The methodology for the in-situ tests was similar to that used in our previous studies [5, 6].  The
experiments consisted of a series of stimulus-response tests. The stimulus being'the injection of
groundwater amended with the chemicals of interest, and the response being the concentration history of
the chemicals at the monitoring  locations. Experiments were performed under the induced gradient
conditions of injection and extraction.  In this study TCE was injected along with phenol and DO. This
permitted the TCE concentration effect to be studied in greater detail. Phenol was pulse injected at 100
mg/l for  1 hr in an 8 hr pulse cycle, resulting in a time averaged concentration of 12.5 mg/l during the first
1000 hr  of the test.  The  initial TCE concentration was 62 u.g/1. The TCE concentration was gradually
raised by doubling the concentration after effective transformation had been achieved at the lower
concentration.  After one week of operation, the injected TCE concentration was raised to 125 ug/l, then
to 250 u.g/1  after another  week, and 500 u.g/L during the subsequent week, and then 1000 u.g/L following
that.  The time averaged  phenol injection concentration was then raised to 19 mg/l for one week, and then
increased again to 25 mg/l, or twice the initial concentration, while maintaining TCE concentration at
                                            37

-------
1000 u.g/1. Bromide was added as a conservative tracer as a basis for comparison for determining
transformation extents.


RESULTS

Microcosm Studies

     The microcosm studies showed that the extent of TOE degradation observed in the columns was
essential independent of the concentration applied. Approximately 50% of the TCE applied was removed
in the columns. TCA which was present as a background contaminant was not removed.  Phenol was
effectively utilized in all the columns, and the uptake of phenol and DO was not affected by TCE
concentration. Thus TCE transformation was effective and appeared to be controlled by a rate
process(es), and not reaction stoichiometry or toxicity, up to a concentration of 1000 \ig/L TCE.

In-sHu Studies

     Rgure 1 shows the normalized concentration breakthrough of TCE at the three observation wells
SSE1, SSE2, and SSE3, located 1, 2.2, and 3.8 meters from the injection well. The normalized
concentration represents the observed concentration divided by the injection concentration at the time of
the observation. The sequential breakthrough of TCE at the SSE1, SSE2, and SSE3 is shown, with the
combined processes of adyective transport, sorption, and biodegradation all contributing to the observed
response. The smoothed lines through the raw data are running averages. This is especially helpful for
well SSE1 where the pulsed addition of phenol induced competitive inhibition causing fluctuations in
TCE concentration. The damping effect of transport and sorption combined with the lack of phenol
results in much smaller fluctuations at the SSE2 and SSE3 wells.
                                     TCE Injection Concentration (ng/1)
                                  200
                                            400        600


                                             TIME (HR)
1000
       Rgure 1.  Normalized breakthrough of TCE at observation wells during biostimulation with
                  12.5 mg/I phenol. TCE injection concentrations range from 68 to 1000 u.g/1.
                                             38

-------
 For TCE injection concentrations ranging from 62.5 \ig/\ to 500 jig/1 the extent of breakthrough is similar
 despite the increases in injection concentration. Bromide tracer tests conducted over this period show
 complete breakthrough of the conservative tracer, C/Co = 1, indicating the tower degree of TCE
 breakthrough results from removal due to bfodegradation. Less breakthrough was observed at the SSE2
 well compared to the SSE1, indicating continued removal with transport through the test zone
 Approximately 70 and 85 percent TCE removal was observed at wells SSE1 and SSE2, respectively
 Upon increasing the TCE injection concentration to 1000 ug/L, the normalized TCE  concentration
 increases with the values at the SSE2 well approaching those at the SSE1 well.

     While injecting 1000 u.g/1 TCE the phenol concentration was increased to 19 mg/l at 1008 hr and then
 to 25 mg/l at 1176 hr.  The results are illustrated in Rgure 2. TCE removal increased following the increase
 in phenol injection.  Since no controls are available it is not possible to conclude that the degree of
 improvement results from phenol concentration increases or other factors such as better adaptation to
 TCE degradation.
                                Phenol Injection Concentration (mg/l)
                800
900
1000      1100       1200


      TIME (HR)
                                                                  1300
     Rgure 2. TCE concentration response resulting from increased phenol addition.

     Phenol concentration was measured in the injection stream and at all monitoring locations.  Phenol
was frequently detected at SSE1, but the concentrations found were generally less than 0 5 mg/l while
12.5 mg/l was injected.  However, detection at SSE2 and SSE3 were infrequent. The measurement
method at the field site had a visible response to phenol concentration as low as 10 u.g/1, but the
quantifiable detection limit was about 25 u.g/1.  The phenol concentrations at the SSE2 and SSE3
monitoring locations were generally too low to give a visible response, and thus are presumed to be below
20 u.g/1, and probably below 10 u.g/L Phenol removal was excellent, and probably greater than 99.9
percent in this system.

     The TCE results are summarized in Table 1. The percentage removals listed are based on average
values at the end of the period following the change in concentrations.  Removal up to 89 percent was
                                             39

-------
observed at the SSE3 well, with 12.5 mg/l phenol added.  Removals were similar with up to 500 u.g/1 TCE
injected. This indicates the removal was first order with respect to TCE concentration.  Upon increasing
the TCE concentration to 1000 u.g/1, the removals decreased to 77 percent with 12.5 mg/l phenol
injected. The lower percentage removal suggests either that this TCE concentration is nearer the KS
value resulting in deviations from first order kinetics, or else TCE transformation product toxicity was
beginning to have a measurable effect.
Table 1.  Average removal efficiencies for TCE at various well locations, and transformation yields.
                                        Percentage Removal
Phenol
Added
mg/l
12.5
12.5
12.5
12.5
12.5
19
25
TCE
Added
ug/l
62
125
250
500
1,000
1,000
1,000

SSE1
60
68
70
68
78
75
78

SSE2
78
82
82
84
85
82
85

SSE3
89
87
88
88
90
85
90
Transformation
Yield
g TCE/g phenol
0.0044
0.0087
0.018
0.035
0.062
0.045
0.036
     When the phenol injection concentration was increased TCE transformation also increased. At the
highest phenol concentration used, about 90% of the TCE was removed at an injection concentration of
1000 u.g/1. Thus, this study demonstrated that a high removal efficiency can be achieved in-srtu by phenol
and DO addition, even with relatively high TCE concentrations.

     Transformation yields are also presented in Table 1.  The highest yield observed in the field was
0.062 g TCE/ g phenol. This yield was obtained while injecting 12.5 mg/l phenol and 1000 u.g/1 TCE.
Lower yields are obtained at lower TCE concentrations, or at higher phenol injection concentrations. The
maximum yield in the field is approximately 50 percent of that observed in the laboratory with a Moffett
derived mixed culture, which is encouraging.


CONCLUSIONS

     An indigenous phenol-utilizing population  effectively degraded TCE up to 1000 u.g/L At the highest
phenol injection concentration up to 90% of the TCE added at 1000 u.g/1 was degraded in a three meter
biostimulated zone.

     Microcosm studies performed under conditions similar to the field tests agreed qualitatively with the
in-srtu tests.  The results are promising, indicating that microcosm studies are of use in evaluating the
potential for cometabolic in-srtu treatment at contaminated sites. Future studies at the site will explore a
range of contaminants including vinyl chloride,  chloroform, and 1,1 -dichloroethylene. The injection of a
non-competitive source of reducing power such as formate and its effect on removal efficiency will be
explored. Studies with toluene as a growth substrate will  also be performed for a comparison with phenol
results.
                                              40

-------
REFERENCES

1.    Nelson, M. J. K., Montgomery, S. O., O'Neill, E. J., and Prttchard, P. H.  Aerobic Metabolism of
      Trichloroethylene by a Bacterial Isolate. Appl. Environ. Microbiol. 52:383-384,1986.

2.    Nelson, M. J. K., Montgomery, S. O., Mahaffey, W. R. and Prftchard P. H.   Biodegradation of
      Trichloroethylene and Involvement of an Aromatic Biodegradative Pathway. Appl. Environ.
      Microbiol. 53:949-954. 1987.

3.    Nelson, M. J. K., Montgomery, S. O., and Prttchard, P. H.   Trichloroethylene Metabolism by
      Microorganisms That Degrade Aromatic Compounds. Appl. Environ. Microbiol. 54: 604-606,1988.

4.    Wackett, L. P., and Gibson, D. T.  Degradation of Trichloroethylene by Toluene Dfoxygenase in
      Whole-Cell Studies with Pseudomonas putida F1. Appl. Environ. Microbiol. 54:1703-1708,1988.

5.    Semprini, L., Roberts, P. V., Hopkins, G. D. and McCarty, P. L.   A Field Evaluation of in-situ
      Biodegradation of Chlorinated Ethenes: Part 2, Results of Biostimulation and Biotransformation
      Experiments. Ground Water 28:715-727. 1990.

6.    Hopkins, G.D., Semprini, L., and McCarty, P.L  Microcosm and In-situ Field Studies of Enhanced
      Biotransformation of Trichloroethylene by Phenol-Utilizing Microorganisms, submitted to Appl.
      Environmental Microbiol. 1992.
                                            41

-------
                    EVALUATION OF AN ULTRASONIC CLEANING SYSTEM

                      AS A REPLACEMENT FOR CFC-BASED SOLVENTS

                         IN A METAL PARTS CLEANING OPERATION
                                       Paul M. Randall
                             U.S. Environmental Protection Agency
                             Risk Reduction Engineering Laboratory
                             Pollution Prevention  Research Branch
                               26 West Martin Luther King Drive
                                    Cincinnati, Ohio  45268
                                       (513) 569-7673
                                            and
                                        Paul B. Kranz
                       Erie County Department of Environment and Planning
                              Division of Environmental Compliance
                                      95 Franklin Street
                                   Buffalo, New York  14202
                                       (716) 858-7897
 INTRODUCTION

     "   For many years, the use of chlorofluorocarbon-based solvents( CFC's) in vapor degreasing has
 been the accepted standard for cleaning parts. The reasons for the universal acceptability of this type of
 process include the efficiency and ease with which parts are cleaned and the subsequent compliance with
 quality control standards for cleanliness of the parts or materials cleaned. In recent years, however, the
 disadvantages of this technology have become increasingly apparent.  For instance, the CFC processes'
 generate fugitive emissions which result in reporting requirements under SARA Title III.  Also, there are
 concerns about employee health and safety and  its increased cost in recent years.  Furthermore, CFC's
 are targeted for eventual elimination because of their ozone depleting characteristics.

        The replacement of a conventional CFC-based parts cleaning system with an aqueous alkaline
 system utilizing ultrasonics was evaluated to determine:

 o      The economics associated with cleaning oil and dirt from metal parts in comparison to a CFC-base
        vapor degreasing process.

 o      The use and generation of hazardous materials from an ultrasonic cleaning system.

 o      The performance of an aqueous ultrasonic cleaning system separating the surface contamination
        from the cleaned components.

       Based upon preliminary tests by the industrial participant, a full scale replacement of CFC-base
parts cleaning was performed at their site in Buffalo, New York.
                                            42

-------
METHODOLOGY

        Ultrasonic cleaning consists of immersing a part in a liquid medium, agitating that medium with high
frequency ( 18 to 20 kHz) sound for a brief time, rinsing with clean water, and drying.  The underlying
mechanism behind this process involves microscopic bubbles in the liquid medium imploding or collapsing
under the pressure of agitation, thus producing shock waves. These shock waves impinge on the surface
of the part and through a scrubbing action displace or loosen particulate matter from that surface. The
process by which these bubbles collapse or implode is known as cavitation.

        In this project, stainless steel, aluminum, and copper parts coated with standard screw oils, water-
based coolants, in-house shop dirt, and metal shavings are cleaned in a modular design of cleaning and
rinsing tanks. The liquid medium is a heated alkaline solution.  Previously, cleaning activities involved the
use of two types of freon-based solvents that generated more than 10,000 Ib of fugitive emissions annually
from two vapor degreasers and two workbench stations.

        For 2 wks in January 1992, the ultrasonic cleaning system was evaluated for 131 batches of parts
ranging from  large tubes to pins and from one to several thousand parts per batch.  Because this was
considered typical production, the results could be extrapolated to an annual basis. Average cleaning times
and chemical addition  requirements were tracked, and subjective quality control inspections were done on
each batch.

        Cleaning and rinse tanks were sampled and analyzed to determine contaminant loading.  Oil and
grease and total organic carbon were evaluated to determine the extraction efficiency of the wash solution
and to assist  in calculating contaminant loading to the publicly owned treatment works( POTW).

RESULTS

        Prior  to 1989, parts cleaning activities conducted involved the use of two types of freon-based
cleaning solvents, Genesolv D and Genesolv 5535. After 1989, the use of these cleaning solvents were
gradually phased out and a larger ultrasonic cleaning system was installed in August 1991 for a total cost
of $  44,411 . A log of cleaning activity was kept for two weeks during January 1992.  The log described
the types and number of parts being cleaned as well as the time each  batch of parts spent in the various
stations. During this two week period, 131 batches of parts  were cleaned which engineers described as
a typical production run. The average time per batch spent in this unit was approximately 8 minutes. The
size  and number of parts cleaned  per batch ranged from large tubes to pins, and from one to several
thousand units.  The data averaged 186 parts cleaned per batch or cycle. This production  run was
extrapolated to an annual basis.

        Table 1  summarizes the waste reduction  potential for phasing out the cleaning solvents and
installing the ultrasonic cleaning system. Based on 1990 data, 12,471 Ibs/yr of freon fugitive emissions and
still bottoms were eliminated and 1.134 million gals/yr of cooling water( normally sewered). The trade off
for installing the ultrasonic system was the generation of 1050 lbs/yr( 3.5 changeovers/ year) of spent
cleaning bath, 567,000 gals/yr of rinsewater and 450 Ibs/yr of an oil/aqueous concentrate.

        The overall operating cost of the ultrasonic cleaning system is substantially less than that of the
CFC-based system, mainly because of the marked decrease in raw material costs( see Table 2).  Power
costs for the  ultrasonic cleaning system are significantly higher than for the CFC-based system.  Labor
costs are comparable and throughput  is nearly equal.  Overall,  an  annual  savings  of $  27,178 was
estimated based on this data. Based upon the initial capital costs for purchase and installation of this
system, a payback period  of 1.6 yrs could be expected.  From the sampling and analysis results of the
cleaning and rinse waters, it was calculated that the ultrasonic system produced a maximum TOC sewer
discharge of approximately 311 Ib/yr, for an estimated BOD loading of 600 Ib/yr.  Based on rinsewater flows
to the sewer,  this loading would contribute a concentration of approximately 70 mg/L BOD, a concentration
lower than the typical concentrations of 250 mg/L (over which a POTW will generally assess a surcharge.
                                                43

-------
                  TABLE 1. SUMMARY OF WASTE REDUCTION POTENTIAL
                                                           Amount per year
  Freon Vapor Degreasers

  1. Fugitive emissions & still
    bottoms

  2. Cooling Water
      12,471 Ibs
1,134,000 gals( 2 units )
  Aqueous Ultrasonic Cleaning System

  1. Cleaners

  2. Rinsewaters

  3. Oil/Aqueous concentrate
      1,050 Ibs

     567,000 gals

       450 Ibs
                 TABLE 2.  OPERATING COSTS SUMMARY ( dollars per year)

Raw Materials
Power Costs
Sewer Costs
Off-site disposal
Water Costs
Labor Costs
Total
Vapor Degreasing
33,939
1,559
6,200
370
1,780
8,205
52,053
Ultrasonic Cleaning
1,203
8,087
6,200
200
890
8,295
24,875
CONCLUSIONS

       The use of an ultrasonic parts cleaning system in lieu of a CFC based vapor cleaning system
results in clear benefits in terms of waste and cost reductions. These can be summarized as follows:

o      Freon use cut by approximately 20 drums/yr
                                             44

-------
o      Waste reduced by over 12,000 Ib/yr when fugitive emissions are considered in the total

o      Annual costs for cleaning parts are cut by 48% from $ 52,053 to $ 24,875 based on 1990 costs
       and estimates

o      Transport and fate of wastes changed from predominately uncontrolled air emissions of Freon to
       predominantly cutting and cleaning rinsewaters that can be sewered. In terms of concentration of
       about 300 Ibs/yr TOG, an approximate BOD loading of 600 Ibs could be estimated. In the flow of
       rinsewater used,  this would contribute a concentration of about 70 mg/L BOD, which is lower than
       typical concentrations of 250 mg/L over which a POTW will generally assess a surcharge. If this
       were considered on a total basis, the  cost of this loading might be on the order of $200 annually.


REFERENCES

1.      USEPA,  Facility Pollution Prevention Guide,  EPA/600/R-92/088,  Office of  Research  and
       Development, Risk Reduction Engineering Laboratory, Cincinnati, Ohio,  May 1992.

2.      USEPA, Waste Minimization in Metal Parts Cleaning. EPA/530-SW-89-049, Office of Solid Waste
       and Emergency Response, Washington, D.C., August 1989.

3.      USEPA. Alternatives for CFC-113 and  Chloroform in metal cleaning. EPA/400/1-91/019, U.S. EPA,
       Air and Radiation ( ANR-445), June 1991.
                                             45

-------
                                ON-SITE SOLVENT RECOVERY

                                         Ivars J. Licis
                                          U. S. EPA
                             Risk Reduction Engineering Laboratory
                                 26 W. Martin Luther King Drive
                                    Cincinnati, Ohio 45268
                                        (513)569-7718

                                       Arun  R. Gavaskar
                                            Battelle
                                       505 King Avenue
                                  Columbus, Ohio 43201-2693
                                        (624) 424-3403
INTRODUCTION
   This investigation was a Pollution Prevention study, evaluating the benefits and capabilities of using
small scale, solvent recovery/reuse systems. The study is in support of the 33/50 program for reducing
17 high priority, toxic wastes. It is part of the Waste Reduction Innovative Technology Evaluations
(WRITE) program of the Pollution Prevention Research Branch and performed in cooperation with the
state of Washington, Department of Ecology.

   As identified by the 33/50 program, which endeavors to obtain voluntary reductions of 33% by 1992,
and 50% by 1995, the list of 17 includes solvents  generally used for cleaning and  degreasing of metal
products, tools and equipment. The list of companies in this category is long (over a hundred
thousand) and includes small users or users producing relatively small amounts of releases and wastes.
The economies of scale have been a significant deterrent from having the smaller generators practicing
the capture and recycling of their spent solvents and fugitive releases.

   The main objective of this research was to evaluate the application of small-scale hardware and
processes that hold promise of good performance, good economics and the reduction of waste and
releases on a multi media basis.

   Three systems were identified, tested, and evaluated. These included: a small  atmospheric batch
still, a vacuum heat-pump unit and a Low-Emission Vapor Degreaser (LEVD).

METHODOLOGY

   The specific objectives for the atmospheric and vacuum units were to obtain information on the
quality of the distilled product, compare it to "virgin" material and determine its ability to perform the
required job. In the case of the LEVD,  the objective was to determine the ability to clean machined parts
to prescribed standards, characterize the emissions and wastes produced and compare to those  in
standard practice.

   For all three systems, cost information was gathered for all known criteria and  compared to
operations under previous procedures.

   Each of the two liquid distillation systems were tested as installed at a manufacturing location.  The
LEVD unit was located at the plant fabricating the degreasers. Standard ASTM methods were used to
determine visual  appearance, color, non-volatile matter, specific gravity, water content, pH, total acid
acceptance, acidity, metal corrosion, purity, conductivity,and absorbance. A flame ionization detector
(FID) was used for measuring the degreaser working chamber concentrations prior to opening the lid
                                              46

-------
and checking ambient solvent concentration levels in the vicinity of the LEVD.

Atmospheric Batch Distillation

       A model LS-55D, Finish Thompson unit was tested on spent methyl ethyl ketone (MEK) used
for flushing out spray paint lines during color changes.  The recycled solvent was reused for the same
purpose, and the residue was shipped off as hazardous waste. For this evaluation, a 55 gallon drum of
spent solvent was processed in a 12 hour period. Data were collected on samples of virgin solvent,
spent solvent and recycled solvent.

Vacuum Heat-Pump Distillation

       The unit tested was Model 040, manufactured by Mentec AG in Switzerland and supplied in the
U.S. by Vaco-Solv Chicago, Inc. This  configuration features a single stage, single pump for both
vacuum distillation and compression/condensation, without additional, external heating or cooling. This
model is suitable for solvents with boiling points up to 80° C.  Data were taken
on the distillation of spent Methylene Chloride (MC) used for cold immersion cleaning of wires and
cables.

Low-Emission Vapor Degreasing

        Low-emission vapor degreasing (LEVD) is a technology currently used in Europe, where vapor
degreasers are regulated as a point source.  Previous studies on conventional open-top vapor
degreasers have showed that a large part of the solvent (over 90% in some cases) is lost through air
emissions.  Reductions in solvent loss can be realized through better cooling, more freeboard,
improved work handling and tank covers.

    Air emissions are mainly workload-related, caused either by dragout of solvent on the workload
itself (and subsequent vaporization) or by disturbance in the air-vapor interface during entry and exit of
the workload. Other sources of air emissions are convection and diffusion during startup, operation,
idling, shutdown, and, to a small extent, equipment leaks.

    The LEVD unit tested in this evaluation was a Model 83S (Size 1, 330-1100 Ibs. of steel parts) which
is manufactured in the U.S. by Durr Automation, Inc., and used perchloroethylene (PCE). After the  unit
is loaded, the lid is hermetically sealed.  Following the degreasing cycle, the liquid on the parts is
evaporated and the vapor is condensed and returned to the sump.  Vapor remaining in the
compartment is adsorbed on carbon and desorbed back to the sump.  The system can also be
operated as a distillation unit  to clean  and recycle the liquid solvent in the sump. For the distillation
cycle, the unit is operated without a load.

RESULTS

Atmospheric Unit

        The recycled solvent was considered suitable for the intended use.  In appearance, color and
absorbance, the virgin and recycled samples from the atmospheric unit were closely similar.  Specific
gravity, conductivity and acidity values of the recycled samples fell between those of the spent and
virgin samples.  MEK purity of the recycled sample was approximately 85%. The difference between
this value and the 99% virgin MEK is believed to be composed of volatile paint thinners, and therefore
not a problem for this application. Approximately 5% water was identified as a constituent of the
recycled MEK.  The main source of this water was believed to be a small leak in the water cooled
                                               47

-------
condenser. While of concern, this constituent was not causing any problem in the painting operation
during the test period.

    The use of this process reduced the amount of solvent waste sent to disposal from 880 to 262
gal/yr. The process requires use of cooling water at 18,360 gal/yr.  Additionally, this water is reusable
as process water.  Energy use was included as an additional operating cost.  While not quantified here,
the consequence of using electrical energy also produces wastes created through energy production.
The process added approximately 1300 Kw-hrs/yr.

    The economic evaluation revealed a payback period of less than 2 yrs.  The purchase price of the
unit was around $13,000, with estimated savings of $10,000 per year.

Vacuum Unit

    The MC purity of the recycled solvent was 86%. This compares favorably with the "virgin" sample
purity of 90%.  The "virgin" material used was previously recycled MC, supplied via vendor. The quality
indicators for the vacuum unit distillate were generally similar to those determined for the atmospheric
unit. The pH of the vacuum distillate was 7 and the corrosion test for steel and aluminum was negative.


    The product distilled via the vacuum unit was considered satisfactory for the purpose. From each
55 gals of spent solvent, the process recovered 48 gals of distillate and produced 3 gals of still bottoms
to be sent to disposal as RCRA hazardous waste. An additional, very small amount of vacuum pump oil
is produced at each routine oil change and combined with other used oil generated by the facility.
Four gallons of solvent were lost as air emissions.  Running the unit slower (per manufacturer's
recommendations) could reportedly recover most of the MC. Running time per 55 gal batch was 12
hrs.
To operate at the increased rate, an air condenser has been added. To accomodate the increased
vapor concentrations in the distillate discharge, a modification has been made to vent the uncondensed
vapor on the roof (but in compliance with applicable regulations).
    The price of the unit was $23,000, the overall cost savings were $18,300/yr. and the payback period
was under 2 years. The 985 Kw-hr of electricity is included as an expense but not as a source of
quantified waste.

LB/D Unit

    The LEVD was demonstrated as a feasible alternative to existing vapor degreasing practice.  The
sealed design of the LEVD, sofvent and vapor recovery, recycling and reuse provide a significant
decrease in solvent losses, compared to conventional, open top degreasers.

    With the use of FID's measuring concentrations inside the working chamber prior to the lid opening,
and ambient readings around the process area, an estimate of 0.0013 Ib/cycle of PCE releases was
calculated.  At a cycle of approximately 60 minutes,  this is a loss of 0.0013lb/hr to clean a 560 Ib. load
of steel parts.  An equivalent, conventional degreaser, sized for 560 Ib. of steel parts/hr with a 4.5 sq. ft.
opening would release 1.617 Ib. of solvent /hr. during continuous operation.

    Overall calculation of PCE losses resulted in 2646 Ibs. per year for the conventional unit and
approximately 4 Ibs for the LEVD.  Overall cost comparisons resulted in a cost savings of $ 25,067 over
the conventional degreaser. With a purchase price of $210,000, payback would be around 9 years.
However, a number of other factors can be significant. The type of metal (thermal diffusivity other than
steel), shape of parts and their capacity to trap solvent, the cost of  auxiliary equipment needed for the
                                              48

-------
conventional degreaserto attain compliance (and others) can significantly change the economics.
These are discussed in some detail in the full report.
    The LEVD also affords greater production flexibility, when pollution prevention is an objective.
Unlike a conventional degreaser, potentially, there are no significant idling losses between loads or
downtime losses during shutdown.

CONCLUSIONS

   All three technologies evaluated in this study demonstrated good potential for pollution
prevention/waste reduction of three of the substances identified as high priority pollutants by the 33/50
program.

   The two on-site solvent distillation technologies reduced large volumes of hazardous solvent waste
to a recyclable product for the given applications.

   All three of the technologies evaluated produce a net gain

   The advantages due to reduced liabilities resulting from producing less hazardous waste are
additional benefits. These were not directly calculated.

   The equations for determining cost benefits are sensitive to changes in many of the parameters.
Specific applications need to be calculated
individually.

   The technologies evaluated in this study are aimed at the short range solutions as prescribed by
the 33/50 program. With the idea of making improvements, cost effectiveness of longer range solutions
should be investigated also. Examples, such as preventing mechanical parts from becoming soiled, or
changing the process to one that does not use oil, or does not require oil (or other fouling substances)
to be removed could be alternate and more cost effective methods in the long run.
                                             49

-------
                     THE RREL LEAD PAINT ABATEMENT PROGRAM
                John Burckle,  George T.  Moore,  Pamela St.  Aimee

                U.S. EPA/RREL - 26 W. Martin Luther King Drive
                            Cincinnati,  Ohio  45268
                                (513) 569-7506
     Elevated blood lead levels in children has become a priority public
health concern in the United States.  Dr. James 0. Mason, Assistant Secretary
for Health, has testified to Congress that "Lead is the number one
environmental poison for children".  Flaking lead-based paint is the major
source of lead exposure to children.  Approximately 57 million housing units,
which represents 74% of all privately owned and occupied housing units in the
U.S., are contaminated with lead-based paint inside or outside the structure,
at or above the action level threshold established by the Lead-Based Paint
Poisoning Prevention Act.  Approximately 9 million of these units are occupied
by families with children under seven years of age.  Of these, 3.8 million
units have either peeling lead-based paint or excessive lead-bearing dust, or
both; these are considered priority hazards.  Based on HUD projections, it
would cost about $10,000 per unit for priority sites -- this is equivalent to
$38 billion 1990 dollars overall cost.  At the present industry capacity of $2
billion per year, it would take 20 years to achieve abatement of all priority
sites.  The RREL lead paint abatement program is designed to address these
priority problems in order to reduce the exposure of children to sources of
lead contamination.  The program addresses lead paint abatement, lead in soil
originating from lead paint, and the problem of disposal of wastes containing
lead paint.

LEAD PAINT ABATEMENT

     There are a number of existing and new technologies for paint removal.
Many have been applied to lead paint abatement.  Several have entered the
market and some, still under development, are considered emerging
technologies.  The technologies are based either on chemical softening,
mechanical mechanisms such as abrasion and blasting with various media, and
thermal systems which loosen the adherence between the paint film and
substrate.  The U.S. Air Force has developed systems based on plastic media
and carbon dioxide pellet blasting to remove paint from aircraft.  The U.S.
Army has investigated alternate paint stripping formulations for refitting
various military hardware.  The U.S. EPA, as part of it's efforts in pollution
                                       50

-------
prevention,  is looking into these and other technologies developed in response
to the DOD needs by private industry for applications in the private sector.
     Although the technologies differ,  there are several considerations to be
applied in common in the evaluation of all  of these technologies for lead
paint
abatement, including the need to collect and dispose of the 1ead'removed
during abatement, protection of abatement workers and the surrounding
environment from fugitive emissions during  the remediation, the establishment
of monitoring techniques to determine when  "clean" levels are reached, and
economics.
     The primary thrust of this program (there are others in the agency with
other goals) is the development of improved systems for higher
productivity/lower cost removal of lead-based paints from homes.  A second,
and equally important concern is to replace organic chemical solvents which
contribute to the volatile organic chemical emission problems of
photo-chemical pollution and the greenhouse effect with "clean technologies"
for paint removal.

     Our research program is directed towards comparing removal techniques
based upon emerging technologies with more  established practices.
             Conventional            Emerging
             Encapsulation           Blasting
             Enclosure               Blasting
             Replacement             Blasting
             Chemical strippers
Systems
w/chemical softeners
w/lamp heating
     We have committed to a program to test and assist in the development of
promising systems for this application.  A number of criteria are of concern.
These are:
       Removal Effectiveness
       Cost
       Generation of Hazardous Waste
       Recyclability of Reclaimed Wastes
       Damage of Sensitive Substrate Materials
       Generation of Fugitive Emissions in the Workplace
       Operability in the Urban

LEAD SOIL ABATEMENT TECHNOLOGY

     For any given treatment technology, the cost of the treatment is
relatively fixed and a function of the process scale.  The predominant
variable cost becomes the cost of transporting the soil to the treatment site
and returning clean soil to the abatement site.  Therefore, we must explore
opportunities to minimize transport costs.  One obvious approach is to employ
treatment systems at, or near to, the abatement site.  In this scenario, the
contaminated soil is treated and returned to its original site area.

     Based on our experience with Superfund technologies, there are four basic
technologies which may be applied.  These are chemical separation and
recovery, mechanical separation and recovery, thermal treatment with
volatilization or vitrification and solidification/stabilization methods.
                                     51

-------
     The chemical techniques require leaching the tnetals into solution,
separation from the soil particles, and subsequent recovery by selective means
or by bulk precipitation.  Mechanical techniques rely on physical property
differences such as specific gravity, size, floatability, or magnetism.
Physical separation and chemical leaching techniques have been highly
developed within the mining industry for minerals separation concentrations
down to about 0.5% metals content.  Thermal treatments have been developed to
fume off volatile metal oxides or incorporate them into glass slags which are
stable under water leach conditions.  Solidification/stabilization processes
based on incorporation of the metals into cementatious or basic silica
environment to prevent leaching are practiced for hazardous waste treatment
and disposal.

     Systems which permit the recovery of the lead in a form amenable to
recycle to the economy through the existing industrial infrastructure are
considered most viable at this time.  These systems achieve a substantial
separation of the lead for the soils and are based on chemical, physical or
thermal fuming processes.  The thermal slagging process produces a high-lead
content slag which may be useful as an aggregate material for asphaltic
concrete or landfilled.  The products from the  solidification/stabilization
process would also require landfill ing.  The options requiring movement and
subsequent treatment of the contaminated soils off site are viewed, for the
purposes of this effort, as the least desirable because they involve longer
distance healing and trigger the move towards higher handling costs.

WASTE MINIMIZATION

     A pollution prevention goal is to develop abatement processes
which permit and promote the recycling of the lead.  In phase of the program
we will be evaluating the efficacy of existing and and experimental
technologies to recycle, treat, or dispose of lead-paint abatement material,
contaminated soil and paint-aggregate from bridge cleaning.
     A consortium of government and industry laboratories is being organized
to participate in tests to characterize the waste and determine their
treatability.  Based on the results of initial pilot efforts, more specific
concentration/recovery technology studies will be proposed including the
potential of new smelting technologies to handle contaminated soil and debris.
                                      52

-------
                 Innovative Clean Technologies Demonstrations

                               Kenneth R.  Stone
                     Risk Reduction Engineering Laboratory
                       26 West Martin Luther King Blvd.
                            Cincinnati,  Ohio  45268

INTRODUCTION

    Since its inception, the Pollution Prevention Research  Branch (PPRB) of
the Risk Reduction Engineering Laboratory has sponsored  programs and projects
in support of small businesses, Federal, State and local  institutions, and
academia.  The Innovative Clean Technologies Project (ICTP)  was conceived as
a pollution prevention outreach effort to small businesses,  where innovators
could apply to the EPA for grants to support developing  new concepts for
reducing waste generation in a variety of processes and  industries.   The Waste
Reduction Evaluations at Federal Sites (WREAFS) program  reached across Federal
departments to develop a network of cooperation within the  Federal community.
The Waste Reduction Innovative Technology Evaluation (WRITE) program went
directly to State and local institutions to develop and  document technology
for pollution prevention.  Now that these programs have  matured, they offer
products that, in combination with products from ancillary  projects, have a
great potential for crossover between the client organizations and waste
generators in the public and private sectors.

METHODOLOGY

    The purpose of this presentation is to provide a few examples of PPRB
activities that have been tested on a limited scale and are now being
developed into on-site demonstrations at operating Federal  facilities such as
airbases, Navy stations and Army centers.  One example is provided from each
of the three programs:  ICTP; WREAFS; and WRITE.

Innovative Clean Technologies Project

    Under RREL's Innovative Clean Technologies Program,  Earth Safe Industries
received a grant to develop and demonstrate a substitute for formaldehyde in
the biological sciences and medical field.  This product, NoToX, has been
validated as an embalming fluid.   In addition to NoToX's antimicrobial
capabilities,  initial findings  suggest performance characteristics superior  to
formaldehyde, with anatomical specimens having a more.pliable texture amenable
to separating  layers of tissue  and muscle fiber for anatomical study.
Occupational exposure to formaldehyde and phenol fumes is eliminated.

    Recently,  additional testing suggests that NoToX may also have application
as a substitute for xylene  in histopathology.  This project will demonstrate
the performance of NoToX as a histological  substitute for xylene at Lackland
AFB.  The demonstration is  envisioned as a  blind test between xylene  and
NoToX, conducted under  EPA  quality assurance  protocols.  At the request of the
                                       53

-------
Air Force, the environmental and energy impacts from the use of NoToX as an
embalming fluid will be analyzed as well.

Waste Reduction Evaluations At Federal  Sites Program

    Recently, a base-wide Pollution Prevention Opportunity Assessment (PPOA)
conducted at the NASA Langley Research  Center (NASA-LaRC) resulted in the
design of a clean product demonstration.   The purposes of the NASA-LaRC PPOA
were: to establish baseline information on NASA-LaRC activities; to identify
opportunities where reduction in resource  use and emissions might be achieved;
to identify potential alternatives for  comparison to operating practices; to
help guide the development of new, clean  technologies on-site.

    This project is focused on the production of non-refractory composite
materials and aircraft structures made  from those materials.  At Federal
facilities and in commercial plants, advanced composite materials are
typically produced through hot melt and solution prepregging (The F-22
Advanced Tactical Fighter (ATF) features  larger expanses of composite material
produced under the hot melt process).

    NASA-LaRC's Polymeric laboratory has  developed a process called, "Dry
Powder Towpreg", in which a composite material can be manufactured without the
environmental burdens resulting from the  use of solvents or energy demands of
the solution and hot-melt technologies.  Also, unlike the typical polymer
resins, the powdered towpreg does not need refrigeration.

    This project will compare the three processes (hot melt, solution and dry
powder), using life cycle methodology to  ascertain the energy and
environmental impacts of each, from the production of the polymer resins
through the manufacture of the material.   Final products will be tested to
determine comparable performance characteristics.

     One such project rs focused on the production of non-refractory composite
materials.  Basic composites are graphite  or carbon fibers impregnated with
some type of polymer through a process  called, "prepregging."  Polymer resins
normally have a shelf-life and must be  refrigerated.   At Federal  facilities and
in commercial plants, advanced composites  are typically produced through
solution and hot-melt prepregging.  In  solution prepregging, polymer resin is
placed in a solvent carrier and applied to the fiber.  The solvent volatilizes
off as toxic air emissions (some solvent  molecules also become trapped in the
polymer matrix).  Hot melt prepregging  avoids much of the solvent usage, but
there is still the question of solvents trapped in the resin being released
when the resin is catalyzed by the heating process.

Waste Reduction Innovative Technology Evaluation Program

    Under the WRITE program, RREL has researched the use of Cold Compressed
Air and liquid nitrogen Cryogun processes  as substitutes for CFCs in cleaning
electronic parts and components.  Through  the use of nitrogen, the Cryogun
process shows promise for several  electronics applications and the research to
                                       54

-------
date indicates a need to refine the heat exchanger configuration of the test
nozzles to better control temperature.

     This project will demonstrate Cold Compressed Air technology in testing
electronic components during manufacture for, and maintenance in DoD
operations.  This test will collect data in four areas: accuracy in detecting
a failed component through cooling; electrostatic discharge risk; cooling rate
and absolute temperature drop; and sound levels.

     The Cryogun using liquid nitrogen to cool circuit boards for testing has
several heat exchanger attachments to provide flexibility in the application.
In this part of the study, data will be collected on the absolute temperature
drop and cooling rate per unit area of the various attachments to the gun to
provide performance guidelines for the user community.  Also, the pollution
prevention impacts of the various attachments will be studied to determine how
well each minimizes the consumption of liquid nitrogen.

CONCLUSION

    The ICTP, WREAFS and WRITE projects have conducted assessments,
evaluations, research and demonstrations across a variety of industrial
activities and chemical operations that can benefit each others clients.  In
addition to conducting RD&D activities with Federal agencies, the WREAFS
program provides technology transfer  by accessing-other PPRB projects and
programs for products applicable to WREAFS clients.  PPRB project officers
lecture classes at the Air Force Institute of Technology (AFIT) with the
objective of providing DoD facility and environmental managers with
information leading to opportunities  to reduce the generation of wastes.
Through a variety of contacts  and networks such as the AFIT curriculum, PPRB
is able to access facilities eager to demonstrate cleaner products  and
technologies on site.

For More  Information on  the Towpreg and NoToX Projects:

Kenneth R. Stone
Risk Reduction Engineering Laboratory
26 West Martin Luther King Blvd.
Cincinnati, Ohio  45268
513/569-7474

For More  Information  on  the Cold Compressed  Air and Liquid Crygun Projects:

Johnny Springer, Jr.
Risk Reduction  Engineering Laboratory
26 West Martin  Luther King Blvd.
Cincinnati, Ohio  45268
513/569-7542
                                       55

-------
                   REDUCTION OF
                                         WARTT?
         ASSESSMENT  OF SELECTED nV/FRHAUl /RFPATR  PROCESSES
               Tom T. Walker
               Pollution Prevention Division
               Directorate of Environmental Management
               Oklahoma City Air Logistics Center
               Tinker Air Force Base ,  Oklahoma 73145
                 (405) 734-7071
INTRODUCTION
     Pollution prevention by eliminating the source  is a
preferred strategy shared by the US Environmental Protection
Agency and the Department of Defense.   A joint Research,
Development,  and Demonstration project was initiated to implement
this strategy at Tinker Air Force Base, Oklahoma City in  1992.
The EPA Risk Reduction Engineering Laboratory and the Oklahoma
City Air Logistics Center, Directorate of Environmental
Management,  managed the project.

     An assessment of the overhaul and repair processes at Tinker
Air Force base was completed as a first step in the program. The
objective of the project was to review the major chemical waste
generating processes at Tinker Air Force Base and identify
alternative processes that would minimize the generation  of
pollutants while continuing to meet overall mission requirements.

     The Oklahoma City Air Logistics Center at Tinker Air Force
Base is one of five Air Logistic Centers within the Air Force
Materials Command and employs approximately 20,000 people. The
center includes an extensive industrial complex dedicated to the
maintenance of the Air Force KC-135 ,  E-3, B-52 and B-l aircraft.
Jet engines and aircraft related accessories are also overhauled
and repaired at the center. The scope  of chemical processes used
at Tinker range from the wipe on of solvents taken from six ounce
containers to automated handling of parts through treatment in
large tanks of chemicals.

      Tinker Air Force Base was an ideal site for the project
because of the variety and number of chemical processes on the
base combined with Tinker's determination to implement
alternative processes to reduce the generation of chemical waste
by eliminating the source.

METHODOLOGY

     The process assessment project was contracted to Battelle.
Five major chemical Waste generating categories were identified
for assessment. They were chlorofluorocarbons, electroplating,
component cleaning, painting-depainting and vapor degreasing. The
                               56

-------
assessment was
base!ine data,
identification
development of
recommendati on
carried out with six major tasks. The gathering of
alternative process characterization,
of alternatives and review by Tinker engineers,
an evaluation method, assessment and
of alternatives.
      The alternative processes were identified through
literature search, industry and government installation survey,
and the experience of the Risk Reduction Engineering Laboratory,
Battelle, and Tinker engineering personnel. Battelle used a
method of multiple attribute evaluation to compare alternatives
to the baseline processes. There were no laboratory or prototype
tests made to compare processes.

RESULTS

     Information gathered during the investigation was recorded
in spreadsheet form and named the Roadmap. The Roadmap defines
shops, chemicals and the programming of alternatives. Over 350
process applications are described  in the Roadmap. Alternative
processes already programmed were noted and emphasis was given to
processes without defined alternatives.
     The investigation of processes resulted  in the generation of
twenty six separate reports which describe alternative methods to
perform the required overhaul -repai r processes. The following  is
a list of the reports grouped by category with a brief
description of the recommended alternative processes or
chemi cal s .
           Deplpt.ing P.hpmir.al A 1 -he-mat, i we»<;  for Vapor

     Trichlorotrif luoroethane - Fuel  Control Overhaul
          Alternative recommended:  Final  rinse with hydrocarbon
blend or alcohol .

     Trichlorotrifluoroethane - Engine Blade and  Vane  Overhaul
           Alternative recommended:  Aqueous .Cleaners with
agitation or power spray washers.

     1,1,1, Trichloroethane  - Tubing and  Cable Production
          Alternative recommended:  Alkaline cleaners with  power
flushing or ultrasonic agitation.

       1,1,1, Trichloroethane  - Wheel  Bearing Overhaul
          Alternative Recommended:  Hydrocarbon blend with
agitation.

          ri7nnp  Ppplp-hing  Chemicals u<;pd  as Aerosol

     Chlorof luorocarbons -  electronics  spray  freeze
                                 57

-------
           Alternative recommended: Liquid nitrogen or compressed
 3 1 r* «

      Chlorof 1 uorocarbons - Electronics Dusters
                Alternative recommend: Compressed Air

      Chlorof 1 uorocarbons - Contact Cleaners.
           Alternative recommended: NSN 6850-00-585-9145
 Mil-C-83360 qualified products using HCFC - 22 or Carbon dioxide
 as  propel! ant.

       Chi orofl uorocarbons - Corrosion Preventative Compounds
           Alternative recommended: Stoddard solvent,  aliphatic
 hydrocarbons with carbon dioxide as propellent.  Propane as
 propellant.

      De-icing  Compounds   R-12  propel! ant
           Alternative Recommended:  Pump spray or carbon dioxide
 as  a  propellant.

      Chlorof 1 uorocarbons  Lubricants
           Alternative Recommended:  NSN  9150-00-458-0075 has
 qualified  products.

                 Dpnlptinn r.hpmiralc; -r
     Trichlorotrifluoroethane/ 1,1,1,  trichloroethane
           Alternative  Recommended:  Electromechanical devices-
hydrocarbon  solvents-  Cables and  harness  aqueous cleaners-
Indicators and  Transducers- Alcohol-   Precision Bearings-
hydrocarbon  solvents.

     Trichlorotrifluoroethane - Oxygen  Equipment
           Alternative. Recommended:  Exterior of LOX converters-
steam cleaning  and  alkaline cleaners.  Follow Naval Sea Command
test program for  interior cleaning  alternatives.

     Mil-L-  63460 Lubricant/Cleaner-
           Alternative  Recommended:  Follow  test program at US Army
Picatinny  Arsenal for  qualified replacements.

     Chi orofl uorocarbon Refrigerant
           Alternatives Recommended: HCFC compounds.

           PI at ing

     Cyanide Solutions   Silver Plating
          Alternative  Recommended: Silver  Strip - Succinimide or
Lactide based solutions.   Silver Strike -  Succinimide and Lactide
solutions. Copper Strike- Pyro-phosphate solutions.
                                58

-------
     Perchloroethylene Vapor Degrease- Wax Plating Maskant
Removal
          Alternative Recommended:  Latex maskant as a substitute,
Mixture of water and perchloroethyl ene.

     r.nmpnncmt. (".loaning _ CnnVfirsi QH Coating

     Chromate Conversion - Aluminum
          Alternative Recommended:  No rinse solutions for
aircraft- modified application technique to reduce volume of
chromate solutions used.

     Phenol /Methyl ene Chloride  Carbon Removal
          Alternative Recommended:  n-methyl -2-pyrrol idone;
methyl -2-pyrrol i done; dibasic esters; cyclic monoterpenes;  and
alkyll esters.

     Methylene Chloride    Engine Exciter overhaul
          Alternative Recommended:  Heat softening of  Epoxy  Foam;
Alkaline Cleaner flux remover; Abrasion cleaning of internal
residues.

     Perchloroethylene Vapor Degrease  Oil Cooler Overhaul
Process
          Alternative Recommended:   Aqueous cleaner with power
flushing.

     Recycle Alkaline Cleaner Process
          Alternative to disposal:  Continuous  filtration  to
extend life of solution; Removal of carbonates and regeneration
of hydroxides with calcium oxide.

              i nt.i n/Pa i nt.i n
     Methylene Chloride  Depainting
          Alternative Recommended: Benzyl  alcohol  has  been
 implemented to remove polysulfide primer  and  topcoat.  High
 pressure water blast for depatnting aircraft.

     Methyl Ethyl Ketone  Radome Depainting
          Alternative Recommended: Add protective  barrier to
 structure  during manufacture and use media blasting.

     Orthodichlorobenzene/Cresyl ic Acid   Heat exchanger/oil tank
 cooler overhaul .
          Alternative Recommended:  n-methyl -2-pyrrol idone.

     Chromated Sealants
          Alternatives Recommended:  polysulfide polymer

     Painting
                                 59

-------
          Alternative Recommended: Consolidated  Paint  facility to
better control emissions.

CONCLUSIONS

     A suitable alternative was identified for most processes.
Most of the alternatives identified will require testing and
development before they can be implemented on a  production scale.
All of the processes reviewed have wide usage in Department Of
Defense industrial facilities and in private industry. Cyanide
process alternatives and some solvent uses required no new
equipment but require testing.  Several of the process
alternatives eliminate substantial amounts of chemical waste.

     The Risk Reduction Engineering Laboratory and Tinker's
Directorate Of Environmental Management are planning technology
demonstrations for several  of the alternative processes.

     The results have been valuable for programming pollution
prevention alternatives at Tinker Air Force Base and has
contributed to a program that will eliminate the use of
chlorofluorocarbons as solvents.
                                 60

-------
DEMONSTRATION OF THE TORONTO HARBOUR COMMISSIONER'S SOIL RECYCLING PROJECT


                              Teri L. Richardson
                     U.S. Environmental Protection Agency
                     Risk Reduction Engineering  Laboratory
                        26 W. Martin Luther King Drive
                            Cincinnati, Ohio  45268
                                 (513)569-7949

INTRODUCTION

      Treatment trains can offer a practical, cost saving approach to
remediating a hazardous waste site,  especially if organic and inorganic
contamination co-exist. Individual technologies, each with a specific
function, can be combined in such a manner that all contamination is
effectively treated and the need to treat or landfill the product is
lessened or eliminated.

      The Toronto Harbour Commissioner's  (THC)  soil recycling demonstration
project, which began in January 1992 and continued for approximately nine
months, illustrated the use of three integrated pilot-scale treatment
technologies: soil washing, metals removal by chelation, and bioremediation.
Three soils, designated as A,B and C were processed through the units. A
fifty-five tons per day pilot plant was housed inside a temporary RUBB Fabric
Building and was designed to treat organic and metal contamination in soil.
The actual process sequence is set according to the specific contaminants in
the soil. The goal of the THC demonstration project was to demonstrate that
the soils located within the Port Industrial District of Toronto's waterfront
can be treated to meet or exceed the Modified Ontario Ministry of the
Environment  (MOE) Criteria Levels for Industrial Soils and additional target
criteria set by THC, without using an incineration process. It is anticipated
that a full-scale plant will be designed and installed in order to perform a
three-year,  85 tons/hours remediation project of the Toronto Harbour front
area to treat an estimated 2,200,000 tons of contaminated soil.

      The U.S. Environmental Protection Agency  (EPA), in cooperation with the
THC  conducted a field demonstration of the THC Soil Recycle Treatment Train
from April 14-16, 1992.  During this period, soil B, which was excavated from a
site that had previously been used for metals finishing and refinery and
petroleum storage, was being processed in the pilot-plant. The objective of
the SITE demonstration project was to evaluate the THC claims that the
technology will meet the following criteria:  1) produce clean gravel and sand
fractions that will  meet the.THC criteria for organic and inorganic compounds
and 2) produce a fine  soil  fraction that will meet the THC criteria for
organic  and  inorganic  compounds after metals removal and/or biological
treatment.

      The SITE program,  which was authorized in the 1986 Superfund Amendments,
was established to promote  joint participation  between technology developers
and EPA  to  accelerate  the  introduction of hazardous waste remediation
alternatives into the  marketplace.   The primary vehicle  for achieving the
goals of the program is  a  technology demonstration that  is designed to
generate pertinent cost  and engineering data for a specific technology. During
a  SITE  demonstration,  the  developer  is generally responsible for operating the
technology  at a  selected location.   EPA's primary  responsibilities include
writing  a  demonstration  plan, conducting  sampling  and analysis operations, and
technology  transfer  activities.

       A SITE demonstration project typically requires a  few days to several
months  to  complete,  depending on the type of process being evaluated  and the
amount  of  waste  that has to be  processed  in  order  to obtain useful results.
For  this demonstration,  the THC determined that a  minimum of six months would
                                      61

-------
be required to evaluate the treatment train.

METHODOLOGY

      The THC treatment train consists of three remediation technologies; a
soil washing unit that is used to separate the soil into uncontaminated coarse
material and highly contaminated fines, a metals removal process that uses
chelating agents to preferentially remove selected metals from a slurry, and
bioslurry reactors that aerobically digest the organic contaminants in the
slurry.  Clean products from the soil washing unit and the bioreactors can be
combined and reused as clean backfill on industrial land.

Soil Washing

      Two soil washing technologies were examined during the THC demonstration
project: a Bergmann USA attrition soil washer and a high pressure soil washing
unit.  The Bergmann Soil Washing System was the first technology in the
treatment sequence during the period in which the SITE demonstration was
conducted. The Bergmann Soil Washing System is a separation technology that
separates metal and organic contaminated waste into clean and concentrated
product streams through the use of physical and chemical mechanisms. It
consists of trailer transportable modular units and can be assembled on site.
The pilot-scale unit that was evaluated during both demonstrations has a
treatment capacity of 5-10 tons of bulk contaminated soil per hour. A full
scale system has a maximum treatment capacity of 55 tons/hour.

      Four process streams are generated: clean gravel (>0.24 in.), coal/peat
(>0.0025; <0.24 in.), clean sand (>0.0025; <0.24 in),  and contaminated fines
(<0.0025 in).

      A rotary trommel washer removes particles larger than 0.24 inch as a
clean gravel fraction.  Particles with diameters less than 0.24 inch pass
through a screen in the trommel washer into a holding tank.  There are three
oil skimmers on the holding tank to remove any free oil from the water.  The
remaining soil (particles smaller than 0.24 inch)  and washwater are pumped to
a separation cyclone where the contaminated fines (<0.0025 inch) are separated
from coarser soil particles (>0.0025 in.). The fines are pumped to a lamellar
separator, that is composed of thin separation plates, and then to a gravity
thickener.  The coarse soil is pumped to the attrition scrubbers.  There are
three attrition scrubbing cells.  The slurry of the soil and washwater is
scrubbed in the first cell, then pumped to the second and third cells for
additional scrubbing. Paddles on the shafts agitate the soil particles and
cause them to rub against each other and thereby scrub the fine particles and
contaminants from the surfaces of the soil particles.   When the process is
used alone, a combination of acids,  bases, detergents or surfactants could be
used in the attrition cells to aid in dislodging or dissolving contaminants.
However, the use of subsequent treatment processes limits the types of
chemicals that can be used.

      Scrubbed soil particles and washwater from the attrition scrubbers are
pumped to another separation cyclone in order to separate sand particles
(>0.0025 inch) from the process water and the remaining fines.  The coarse
soil stream is then passed through a density separator to remove low density
materials such as coal, wood and peat from the heavier soil particles.   The
coal, wood and peat are collected separately as a potentially contaminated
waste stream.  The cleaned coarse soil particles are discharged by conveyor to
a collection bin and are combined with the gravel from the trommel washer for
return to the original site.  Process washwater with contaminated fines
smaller than 0.0025 inches flows through a lamellar separator to remove the
fines from the process water and are then pumped to a sludge thickener.  The
contaminated slurry from the sludge  thickener is fed into two large holding
tanks at the front end of the metals removal unit or directly to the bioslurry
reactor process,  depending on the concentration of metal contamination in the
feed soil.
                                       62

-------
Contaminated process water is removed from the lamellar separator and sludge
thickener,  treated in a conventional water treatment system, and is discharged
to an outdoor storage pond.


Metals Removal Process

      A metals removal process that uses chelating agents to remove metals was
selected by the THC to treat the contaminated slurry from the soil washing
process. The composition of the slurry was approximately 24% solids by weight
and 76% process water.  Initially, the slurry is fed into two large holding
tanks at the front of the metals removal process.  An undisclosed mild acid is
added to the slurry to desorb and solubilize any metal contaminants from the
soil particles.  The slurry is then pumped into a screw type tubular reactor
where it comes into contact with metals chelating agents that have an affinity
for attracting specific metal contaminants.  The slurry, void of metals, is
pumped to a holding tank where it is mixed with an oxidant prior to treatment
in the biological system.

      The chelating agent with the contaminant metals flows countercurrent to
the slurry and is fed through a second tubular reactor where a mild acid is
used to break the bond between the chelating agent and the contaminant metals.
The chelating agent is then recycled to the first reactor to be reused.  The
metals/acid mixture is pumped to a unit in which the metals are removed on
plates by electrolysis, and the acid is returned to the holding tanks for
reuse.

Biological Treatment System

      The biological treatment system that was evaluated combines chemical
hydrolysis and biodegradation of hydrocarbons and consists of a series of
three 20,000 gallon CSTR aerobic upflow bioreactor tanks in which organic
contaminants are treated.  The unit throughput is 10 gpm of slurried fines or
up to 10 tons of contaminated fines per day on a dry weight basis. It is
designed to biodegrade all hydrocarbons present in the feed soil. The slurry
is fed from the holding tanks into the bioreactor tanks, where bacteria are
fed air and, if necessary, co-nutrients such as urea or phosphoric acid
solutions, to maintain optimum bacterial activity. Constant mixing and
suspension of fines is provided by submerged pumps and upflow of air.  When
the organics content in the slurry has been reduced to below the specified
criteria levels, the slurry is pumped to an aerobic digester where the
bacteria digest themselves.  This is allowed to continue until the measured
oil and grease level in the slurry meets MOE guidelines.  The decontaminated
slurry is then pumped to a holding tank before it passes through three
hydrocyclones where it is partially dewatered.  The clean, partially dewatered
fines are collected in a holding bin and can be combined with the clean soil
from the wash plant.

Site Description

      Soils that were processed as part of the THC Soil Recycling
Project were from the Toronto Port Industrial District  (PID).  The soil is
characterized as a silty sand. Specific industries that have operated within
the PID include refineries, oil terminals, coal storage, and processing and
metal recyclers.  'As a result, the majority of the contamination in the soil
is light and heavy hydrocarbons, and heavy metals.  In most instances, the
metals and hydrocarbon contamination co-exist and cannot be separated.  The
type and concentration of each contaminant is site specific and highly
variable.
                                        63

-------
RESULTS

Soil Washing

      The THC criteria and the average influent soil contaminant
concentrations are presented in Table 1.  Concentrations are the result of
averaging daily composite sample results  during a three day sampling period.
Two composite samples were taken each day.

                 TABLE 1. Concentration of Selected Parameters
               in the Feed Soil THC Attrition Soil Wash Process
Soil
% Moisture
Oil & Grease
mg/kg
TRPH mg/kg
Cu mg/kg
Pb mg/kg
Zn mg/kg
Naphthalene
mg/kg
Benzo(a)pyrene
mg/kg
THC
Criteria
	
10,000
	
225
750
600
8.0
2.4
Feed
Soil
9.18
8,200
2,540
18.3
115
72.5
. 11.15
(1.92)
                ( )  Indicates1 value reported is below quantitation
                limit but above detection limit.  Value should
                be considered an estimate.

      Results from pre-demonstration sampling indicated contamination levels
much higher than those that were actually processed.   The only feed soil
parameter that exceeded the criteria was naphthalene.

      The soil washing unit produced a gravel and a sand that met the
criteria.  Removal rates for oil & grease,  total  recoverable petroleum
hydrocarbons (TRPH), naphthalene, benzo(a)pyrene, were 65 percent or more for
the gravel product.   Removal rates for the sand fraction were 75 percent or
greater.  Metal contaminants were concentrated in the fines slurry.

      The performance characteristics of the soil washing process are given in
Table 2.
                                       64

-------
                  TABLE 2.  Selected Feed and Product
              Characteristics of the Soil Wash Process
                          Feed   Clean  Coal/Peat
                          Soil  Gravel   Fraction
Clean    Cont.
 Sand    Fines
Percent of Feed Based
Metals Removal Process
Percent of Feed Based
on THC Overall Analysis
Oil & Grease mg/kg
TRPH mg/kg
Copper mg/kg
Lead mg/kg
Zinc mg/kg
Naphthalene mg/kg
Benzo ( a ) pyrene

	
8233
2542
18.3
115
82.5
11.15
(1.91)
11.5
10.5
3333
814
6.4
45.3
46
2.62
(0.58)
1.6
2.5
38,100
11,850
32.9
406
210
64
(14.5)
68.1
70.2
2183
621
13.8
46
34.1
2.05
(0.53)
18.8
16.7
40,000
14,000
83.1
522
344
51.7
(10.0)
      The level of metals contamination in Soil B that were encountered were
so low that it was not necessary to use the metals removal process.  However,
limited data was subsequently obtained when another,  more contaminated PID
soil, was being processed.  The removal efficiencies that were achieved based
on the metals concentrations in the influent versus the effluent are given in
Table 3.
           TABLE 3.  Selected Heavy Metals Data for Removal of Metals
            From the Liquid Stream by the THC Metals Removal Process
Metal
Copper
Lead
Nickel
Zinc
Influent
mg/kg
51.1
( 49. 2-53. 2) -1
100.5
(94.2-112)
11.7
(10.7-12.7)
277
(264-294)
Effluent
mg/kg
1.8
(0.9-3.0)
29.0
(13.5-46)
3.3
(0.9-7.3)
101
(53-183)
Removal %
96
71
71
63
                     of Values
Bioslurry -Process

      The data that was evaluated from the bioslurry process were the result
of sampling the discharge from two bioslurry reactor batches.   The first
batch, identified as 2c,  was accumulated between March 25-31,  1992 and was
discharged from the bioslurry reactor on April 1, 1992.   The retention time
from the time of nutrient addition to discharge was forty-one days.   Seven
                                       65

-------
grab samples were collected.  The second batch, designated as  2a,  was
accumulated from April 14 through April 22,  1992.  The retention time was
thirty days.  Seven grab samples were collected.  A comparison of the inlet
and outlet concentrations for batch 2a is summarized in Table  4.  The results
of batch 2c are not directly related to the SITE inlet concentrations,  but the
results were similar.

               TABLE 4.  THC Bioslurry Reactor Inlet and Outlet
                    Concentrations for Selected Parameters

Oil & Grease
mg/kg
TRPH mg/kg
Cu mg/kg
Pb mg/kg
Zn mg/kg
Benzo ( a ) pyrene
mg/kg
Naphthalene
mg/kg
Contaminated
Fine Slurry
40,000 .
14,000
83.1
522
344
(10. O)1
57.7
Bioslurry
Batch 2a Discharge
25,300
5,440
83.6
548
343
(2.6)
<13.02
Apparent
Removal
36
61
—
-5
—
74
75
  1. ( ) indicates that the value is oeiow cne quantisation limit ror
     the procedure.  Value shown is estimated.
  2. Based on average quantitation limit for seven samples.

      For both batches, the THC criteria for oil and grease was exceeded.  The
oil and grease value was approximately 5% for batch 2c and 2.5% for batch 2a.
The THC criteria for oil and grease is 1%. The developer indicated that the
high values could be attributed to biological fats and oils being extracted
from the biomass during analysis.

      The target criteria of 2.4 mg/kg was also exceeded in both batches for
benzo(a)pyrene.  Batch 2c had an average value of 3.2 mg/kg, whereas the
average value for batch 2a was 2.6 mg/kg.  The levels of benzo(a)pyrene in the
discharge were above the analytical detection limit, but below the
quantitation level.  Therefore, the reported values are considered to be
estimates.

CONCLUSIONS

      The THC treatment train provides a method for the removal of organic and
inorganic contamination in soil.  The data that was evaluated as part of the
SITE demonstration indicates that the soil washing process can effectively
remove 65 percent or more of the contamination in the feed soil and
concentrate the metals in the fines slurry.  The metals removal process, which
is used to further treat the fines, can remove 60 to 90 percent of the metals.
The bioslurry reactor process removed approximately 70 percent of the PAH
compounds.
                                        66

-------
     RESULTS OF THE EPA MITE PROGRAM: EVALUATION OF LANDFILL MINING TECHNOLOGY

                                       Lynnann Hitchens
                                            US EPA
                                     5995 Center Hill Road
                                     Cincinnati, Ohio 45224
                                         (513) 569-7672

                                      Edward L. von Stein
                                       CalRecovery, Inc.
                                         5 Science Park
                                 New Haven, Connecticut 06511
                                         (203) 786-5250

                                       George M. Savage
                                       CalRecovery, Inc.
                                     725C Alfred Nobel  Drive
                                      Hercules, CA 94547
                                         (510) 724-0220

INTRODUCTION

       The US EPA established the Municipal Solid Waste Innovative Technology Evaluation (MITE)
Program to provide municipalities and the public sector with information on new and developing solid
waste management technologies. The MITE program provides a framework within which technology
developers have the opportunity to demonstrate the effectiveness of their technology or process in the
field. The technology developer is responsible for funding and directing the demonstration of the
technology; EPA funds and directs the technical and cost evaluations and publishes the results.

       The landfill mining technology was submitted to the MITE program by the Collier County Solid
Waste Department, and was selected in  May 1991. Landfill mining uses basic excavation and solid
waste processing operations to reclaim and recover landfilled materials.  Collier County was  one of the
first groups either public or private sector to develop and apply this technology.  Since the inception of
their operation, a number of other landfills have attempted similar projects.  Landfill mining can meet a
number of fairly divergent objectives.

  •     Recovering cover soil and other potentially recyclable materials

  •     Decreasing the footprint of a landfill and the acreage that requires closure and post-closure care.

  •     Using the high energy value material in a waste-to-energy combustor.

  •     Removing material from an  uniined landfill  with the goal of lining and reusing the space.

       The Collier County Solid Waste Department launched their operation solely to reclaim used
cover soil  and degraded material from a closed portion of the Naples Landfill. Their operation simply
used a series of screening steps to separate the soil-like fraction which was then used on the active part
of the landfill. Material not suitable for reuse as cover was placed back in the landfill. The use of the
reclaimed  cover soil on the active portion of the landfill represented a cost savings to Collier County on
the purchase of cover soil.

       The MITE program evaluation looked at an expansion to the existing operation that would seek
                                               67

-------
 to recycle ferrous, aluminum and plastic.  Several objectives were established for the MITE evaluation:

  •     Evaluate the unit operations and their ability to produce process streams of required purity.

  •     Evaluate the soil fraction in comparison with Florida State compost standards to determine its
        applicability as a soil amendment.

  •     Evaluate the marketability of the product materials such as ferrous, plastic, and aluminum.

  •     Determine the cost of operation.

        In pursuing these objectives, a one week time period was devoted to monitoring the
 system/process flows and obtaining the necessary samples.
METHODOLOGY

        The system evaluated was designed by the Collier County Solid Waste Department. Figure 1
depicts the process flow diagram, identifying unit operations as well as product streams.
                                 Soil
      Solid
Ferrous
   *    (S3)
Residue
                                                                         (S4/S5)
                            Pfastics
         Non-processibles
                           (37)
            Finger    Heavies
            Screen            (S9)
            Unders
                  (S8)
            Figure 1.  Process Flow Diagram of the Collier County Landfill Mining System.

       The process line has four separate unit operations to provide the required separations: a coarse
grizzly screen and fine trommel screen, a ferrous magnet and an air knife.  The grizzly screen, with bars
having an opening of six inches, separates the non-processible material (S7) from the feedstream.  This
larger material is landfilled. All remaining material  (less than six inches) is conveyed to the trommel.
The purpose of the trommel is to separate the soil fraction (S1) from the remaining material.  The
trommel has 3/4 inch openings through which the soil and degraded material pass. This stream is
usable as cover on the active part of the landfill. Analytical testing was performed on samples of this
material for comparison to Rorida State compost regulations.

       The oversized material (> 3/4 inch) moves from the trommel, onto a conveyor,  and through the
final two unit operations, separating it for possible recycling. A ferrous magnet is used to separate
ferrous material (S3) and the remainder of the material then enters the air knife.
                                               68

-------
       The purpose of an air knife is to perform a density separation by blowing, with high speed air,
the lighter, smaller material from the unit, while the denser, heavier and larger material falls to the bottom
and is collected. At the entrance to the air knife all material passes over a vibrating finger screen with
3/4 inch  openings, through which heavy material falls. These finger screened unders (S8) are collected
for disposal.  The remaining material then enters the fluidizing section which was modified to produce
three additional process streams.  The high speed air stratifies the material to enhance the removal of
smaller size, high density refuse particles.  The large heavy material (S9) is bottom discharged and the
lighter material that is blown through the air knife was separated into a moderately light fraction (S4/S5)
containing mostly aluminum, glass and small plastic fragments, and a "superlight" fraction (S2) which
was essentially all plastic film.

       The original process design also contained an eddy current separator between streams S3 and
S4/S5. The purpose of the eddy current separator was to remove the aluminum fraction but after a few
test runs it was determined that the eddy current separator used was incompatible with capacity
requirements and particle size of the infeed stream.  The soil-coated heterogenous mix did not permit
the equipment to operate properly, so it was removed.

       Both a mass balance and  a composition analysis were  performed on all product streams.
During operation all feed material and all process streams were weighed. Additionally, the individual
product streams were sampled and characterized.

       A sample of material was taken from a collection roll-off container and laid out on the sampling
grid.  Random sub-samples of material were taken and shipped for analysis. The remainder of the
sample was characterized according to the fourteen solid waste categories:
        Paper and paperboard
        Yard waste
        Food waste
        Aluminum
        Glass
Plastic                «
Unidentifiable          .
Ferrous metal          •
Non-ferrous metal      •
Textiles
Rubber/leather
Non-processible
Inerts (soil)
        Each category was weighed to obtain the purity of each product stream. This characterization
was performed on product streams S1, S2, S3 and S4/S5.

        An air quality survey was also performed and conducted concurrently with the product sampling
and characterization. Twelve individual air sampling episodes were conducted over four days.  The air
samples were analyzed for the following parameters:  total and  respirable particulates, total microbial
agents,  selected metals and fibers.


RESULTS

        There were four primary product streams: soil (S1), plastic (S2), ferrous (S3), and residue
(S4/S5). Because of the inability of the eddy current separator to function with the residue (S3) as feed,
a separate aluminum stream was not obtained; however, the  residue stream contained a significant
amount of aluminum.  Mass balance data were taken over seven days of operation.  A total of 292 tons
of mined material was processed with the system having an average processing rate of 13 tons per
hour.

        Sampling and characterization of the product streams (S1, S2, S3, S4/S5) occurred over four
days of processing.  Table 1 shows the average composition of each of the primary process  streams.
This table illustrates the relative purity of each product stream.
                                               69

-------
       Table 2 presents laboratory analysis for the soil fraction as compared to the heavy metals as
regulated by Florida State Compost Concentration Code 1, lowest limit set of the four classes of
compost.  In addition to this analysis, the soil fraction underwent analysis for 34 parameters. Among
those were:  bacteriological agents, moisture content, pH, ammonia, TKN and nitrate nitrogen,
phosphorus, potassium, total and volatile solids, BOD and COD, asbestos fibers and a range of heavy
metals.
                TABLE 1. AVERAGE COMPOSITION OF THE PRODUCT STREAMS
                                        (% WEIGHT)
 Component
 Paper & Paperboard
 Plastics
 Yard Waste
 Ferrous Metals
 Rubber/Leather
 Textiles
 Wood
 Food Waste
 Aluminum
 Glass
 Inerts (Soil)
 Non-Ferrous Metals
 Unidentifiable
 TOTAL
 Soil (S1)
(59.39%)*


    0.6
    0.3
    1.1
    0.0
    0.0
    0.0
    0.0
    0.0
    0.0
    1.9
   94.2
    0.0
    2.0

  100.0
Ferrous (S3)
   (1.73%)*
       1.2
       9.0
       0.2
      81.5
       0.0
       1.2
       2.0
       0.0
       0.5
       0.2
       1.6
       0.1
       2.3

     100.0
Plastic (S2)
  (2.42%)*


     14.3
     74.5
      1.5
      0.1
      0.0
      4.4
      0.2
      0.0
      0.7
      0.1
      0.9
      o;o
      3.4-

    100.0
Alum/Residue
      (S4/S5)
     (7.69%)*

       15.2
       24.3
        9.7
        0.4
        4.9
        8.3
       15.2
        0.1
        5.5*
        3.3
        8.0
        0.7
        4.3

      100.0
 * These numbers represent the average weight percent of each stream as a fraction of the total
 amount of mined material.
 " Separation of aluminum was not possible with the existing equipment.  This represents the majority
 of the aluminum
                   TABLE 2.  COMPARISON OF HEAVY MEATAL LIMITATIONS
                           WITH RECOVERED SOIL FRACTION (S1)
 Metals (mg/kg dry wt)
 Cadmium
 Lead
 Mercury
 Zinc
 Chromium
 Nickel
 Copper
          Florida*     Recovered Soil Fraction S1*

          <15                             1.7
         <500                           56.0
          N/A                             0.2
         <900                          197.5
          N/A                           13.8
          <50                             3.9
         <450                           32.0
 " Florida's Heavy metal Criteria for Compost; Code 1.  Source: FAC Chapter 17-709.550(1 )(e)
 * Average of 4 samples collected daily for each of the consecutive days
                                             70

-------
CONCLUSIONS

       The average mass balance data and the accompanying statistical analyses show that all targeted
product streams exceeded the study objectives of a 90% confidence interval with a standard error of
less than 25%. The average output/input was 90.2%. These results indicate a satisfactory closure on
the mass balance.  The system had an average hourly processing rate of 13 tons per hour, based on the
four days of processing run times and quantities of material.

       Results of the air quality survey indicated that dust, gypsum fiber, metals concentrations were
well below permissible exposure limits based on workplace standards, but that microbial agents may
have been released during processing.

       Key to this study was to determine the quality of the recovered materials. Table 2 indicates  that
the soil fraction (S1) contained levels of metals well below the Florida State  compost requirements for
Type A Compost.   However this material would not meet the State requirement for foreign matter. The
soil fraction made up approximately 60% of the total material mined.

       One to three pound quantities of mined recyclables were sent to several possible buyers from
the representative markets. The general reaction to the ferrous, aluminum, and plastic samples was the
same for each evaluator: the mined materials would require extensive cleaning and pre-processing
before they could be competitive with their source-separated counterparts.  Without such pre-
processing, the materials would  only be suitable for lower quality markets than those for source
separated materials where such  markets exist.1


REFERENCES

1.      CalRecovery, Inc.  Landfill Mining Technology Evaluation EPA/MITE Demonstration Program.
        Draft  Report, 1992. 119 pp.

2.      Collier County Solid Waste Department. Landfill Reclamation Technology Transfer Information.
        Naples, Florida,  1991.

3.      New  York  State Energy  Research and Development Authority.  Town of Edinburg Landfill
        Reclamation Demonstration Project. Energy Authority Report 92-4,  Albany, New York, 1992.
        148pp.

4.      Rule  17-709: Criteria for the Production and Use of Compost Made from Solid Waste, State of
        Florida Department of Environmental Regulation, 1989.
                                               71

-------
         AUTOMATIC SORTATION PROCESS FOR POST-CONSUMER PLASTIC CONTAINERS
                                          Diana R. Kirk
                                            U.S. EPA
                              Risk Reduction Engineering Laboratory
                                     5995 Center Hill Avenue
                                      Cincinnati, Ohio 45224
                                         (513) 569-7674
                          David V. Bubenick and Charles N. Faulstich, Jr.
                                        wTe Corporation
                                         7 Alfred Circle
                                       Bedford, MA 01730
                                         (617) 275-6400
INTRODUCTION
       Plastics, which are manufactured from petroleum and natural gas, are a rapidly growing segment
of the municipal solid waste stream.  Plastics comprise 8 percent of the waste stream by weight and 20
percent by volume but only 2 percent is recycled. Recycling offers a multitude of benefits such as:
conserving landfill capacity, minimizing disposal costs, saving the refining and some of the
manufacturing energy used in the production of virgin resins when the oil costs are high.


    Rigid containers, which make up almost half of the plastics in the waste stream, are comprised of
different resins. Greater value and broader applications for plastics are achieved with homogeneous
resins, therefore plastics must be separated by polymer type with minimum contamination. The
significant differences in material properties require the separation of plastic containers by resin prior to
reclamation.  Up until recently, manual sorting has been the only method commercially available for
separating whole plastic bottles by resin type.  Manual sorting has been found to be less than ideal,
owing to the labor and training expense and susceptibility to error. Highly automated plastics sorting
systems have thus been designed to replace manual sorting with the goals of increasing the separation
efficiency and product quality while reducing the recovery cost.


    This project evaluates the Rutgers Center for Plastics Recycling Research (CPRR) system for the
automatic sortation of post-consumer plastic containers.  This evaluation provides an objective analysis
of the ability of the system to successfully identify and separate five types of plastic containers, and a
determination of the system's potential use in full-scale commercial applications.
METHODOLOGY
   The Rutgers CPRR automated plastics sortation system integrates equipment that separates rigid
plastic containers into five types: polyvinyl chloride (PVC); clear polyethylene terephthalate (PET); green
PET; clear (or natural) high density polyethylene (HOPE); and opaque HOPE.  The system uses x-ray
                                               72

-------
fluorescence to separate out PVC, and optical methods for distinguishing between PET and HOPE
containers.


    Uncrushed post-consumer plastic containers are manually fed into a feed bin.  The feedstock will be
proportioned according to a typical, average composition of curbside recyclables: 24.7% for natural
HOPE; 39.6% for opaque HOPE; 5.6% for PVC;  22.8% for clear PET; and 7.2% for green PET.  In order
to effectively identify and separate the containers by type, they must be presented in an orderly and
uniform manner. To achieve this goal, an inclined cleated conveyor with variable speed transports the
plastic containers past a counterclockwise rotating paddle wheel which diverts oversized containers and
natural HOPE from the flow stream.  All the other containers fall from the upper end of the ascending
conveyor onto a silo where they are gravity-fed  via two feed chutes on another conveyor.


    Two feed chutes start the initial aligning of containers. The containers are further aligned and
singulated on a conveyor that consists of three  belts moving at constant differential speeds with an
angled partition to help singulate further onto another conveyor.


       The last conveyor is equipped with a series of three detector/ejector stations for separating
individual container types. The ejector component is identical for each station.  The air ejectors have
two nozzles, angled toward each other in order to create a  pressure wave to divert the container across
the conveyor into its respective recovery station. The  dual air nozzle configuration is intended to
minimize any spin imparted to the container as  may be the  case with single, highly focused air jets.  The
sensors which detect each polymer type are mounted on the conveyor side wall. The first detector uses
x-ray fluorescence to detect the chlorine atom found only in PVC containers. The next two detectors
use optical sensors for distinguishing between clear PET and green PET. The remaining containers,
which are essentially opaque HOPE, are  not positively separated but simply collected at the end of the
conveyor. After each test is conducted,  bottles in bins are counted and checked for "imposter" bottles.


   The project's success depends on meeting the objectives set forth. The primary objective is to
determine the maximum feed rate, with 95% confidence, to  assure 100% recovery of PVC and 90% for
other container types, while meeting product contamination requirements assumed typical of the
industry. These are standards considered appropriate for commercial-scale systems, within limits of
contamination.  Two secondary objectives are to determine the mechanical reliability, availability and
maintainability of the overall system, and finally, to determine the potential for commercial scale
operation by establishing a proof of concept for the technology employed.


    The test plan involved running three series of tests to satisfy the project's objectives.  Single
composition, short term tests were run to evaluate the recovery of the five container types without any
interference from mixed containers.  Mixed composition, short term tests were run to evaluate the
effectiveness of separating containers by type from a mixed stream of known composition.  Extended
tests were run to verify the optimum operating and performance results obtained from the short-term,
mixed tests, can be sustained over a longer operating period, characterized by random mixing that more
closely simulates a commercial operation.


    The system evaluation will be based  on 100 short-term tests consisting of 75 single composition
tests and 25 mixed composition tests. The 75 single composition tests consist of 5 replications each at
50, 75 and 100 containers per minute for natural HOPE,  PVC, clear PET, green  PET and opaque HOPE
                                                73

-------
 rigid containers. The 25 mixed composition, short term tests, also based on five replications each, are
 performed at feed  rates of 30, 50, 75, 100 and 120 bottles per minute.  An extended test, one week in
 duration, Is conducted once the optimal feed rate from the single composition tests is determined.
 RESULTS
   The mixed composition, short term test results, together with the observations of the characteristics
 of the system resulted in the selection of a bottle feed rate of approximately 73 bottles per minute for the
 extended tests.  At 8 bottles per pound, this is approximately 550 pounds per hour.  Downtime occurred
 due to equipment failure (i.e. paddle wheel flaps falling off and singulating conveyor drive failing),
 causing the evaluation to include only two to three replicates for each test instead of five replicates.
 Reliability, availability, and maintainability data obtained were limited because of the pilot-scale nature of
 the system.


    The mixed composition extended test results showed that only natural HOPE and clear PET met the
 established recovery accuracy objectives. Average recovery accuracies were 99.1 percent for natural
 HOPE, 89.3 percent for PVC, 92.8 percent for clear PET, 86.5 percent for green PET, and 89.4 percent
 for opaque HOPE. The average overall recovery accuracy of the system was 92.4 percent.
 Contamination levels in all of the products exceeded the assumed limits representative of commercial
 level specifications.


    Most of the problems encountered with the system design dealt with materials handling. The
 sensors were sometimes unable to detect the different sorts of plastics due to the positioning of the
 bottles. For example, the laundry detergent  bottles with handles are sometimes mistaken by the sensors
 as clear PET. When a laundry detergent bottle with a  handle passes by the clear PET sensor, the eye of
 the sensor detects the space where the handle is located and the bottle is shot into  the clear PET bin.
 The spacing between bottles on the conveyor seems to be a major factor in detection of the plastic. For
 example, if one bottle lies over another one on the conveyor, this may cause the clear PET sensor to
 classify the bottle as clear PET when the overlap of the bottles causes a gap.  Video taping the PET
 station during an extended mixed composition test provided a means for documenting the types of
 identification and recovery mistakes occurring. Through slow-motion viewing of the  video tape, the
 exact delivery time and type of bottle was recorded, from which  the separation distances were
 calculated. During one extended mixed composition test, fifty percent of the mistakes occurred during a
 time interval equal to eight percent of the total test time. It was discovered that the mistakes are
 attributed to variable spacing between bottles. The errors observed at the PET station included
 undetected and unrecovered PET, clear PET being swept with contaminant bottle into recovery bin, or
 single contaminant recovery. This emphasizes the importance of singulation or more importantly how
 the plastic containers are fed onto the conveyor for accurate detection.
CONCLUSIONS
    In the course of processing the bottles, many types of identification/separation mistakes were
observed. Some mistakes were caused by deficiencies of the material handling system, in that bottles
were too closely spaced or not properly aligned for proper detection. The material handling system had
a greater effect than sensor performance on the overall performance of the system. Recovery accuracies
were not found to be directly proportional to the feed rates, but appeared to be influenced more by the
                                               74

-------
capacity of the feed system to deliver properly singulated bottles.  This was evidenced by the video
tape analysis of the clear PET station during one extended test which showed that one-half of the
mistakes were bottles captured as the result of inadequate bottle spacing. The high feed and recovery
rates for the single composition tests also demonstrated the significance of spacing  between bottles.
The single composition tests did not have interferences from different bottle weights and sizes and so
appropriate spacing provided adequate detection of the plastic types.


   Although the recovery accuracy and product contamination data from this pilot scale system failed to
meet the commercial-grade standards adopted for this evaluation, the results do confirm that the
technology employed to identify and separate plastic containers by resin  is valid. The importance of
proper bottle presentation was documented via the video tape analysis. Video taping one of the
extended tests proved invaluable in providing insight into diagnosing the system's singulation
performance and errors in station recovery.  The most critical mistake occurs from poor singulation and
causes two bottles to be swept off the belt together. The swept bottle  can be located in front of or in
back of the proper bottle. Separation distances greater than 2 inches are considered adequate to avoid
sweep mistakes. Abrasion or deterioration of the bottle surface was observed to cause clear and green
PET to avoid identification on occasion.  System design  modifications are required to optimize recovery
and  minimize errors.


    Such fundamental design changes include a  multiple pass system which will improve recovery by
allowing more than one opportunity to detect and remove the target compound.  PVC recoveries can be
enhanced by upgrading the PVC detector to ASOMA 2 which is  relatively insensitive to container-to-
sensor distance. Also,  the feed system should be redesigned so that a more uniform bottle feed rate is
achieved in the system, generating proper spacing  between bottles for enhanced recovery.  This can  be
done by increasing chute and conveyor skirting clearances to prevent feed jams from occurring.  It can
also be accomplished by providing a secondary  bin for containers not  removed at the natural HOPE
station, therefore eliminating big spaces on the cleats. Large spacing between bottles causes the
system to attain non-uniformity and lower feed rates.  Additional segments could be added to the
singulating conveyor for further singulation and improved spacing.  Each subsequent conveyor in a
series of conveyors should operate at a higher speed than that of the conveyor before it to take
advantage of any prior  bottle singulation.


    The basic principles of x-ray flourescence, optical detection,  air-jet  ejection, and mechanical oversize
separation were demonstrated to be valid methods for detecting and separating whole, uncrushed
plastic containers from  a mixed plastic stream. The deficiencies of the material handling system and the
lack of a state-of-the-art PVC detector severely limited the performance of the overall system in terms of
recovery and feed rate. Even with these limitations, the system  still achieved an overall recovery rate of
92 percent at a throughput rate of 73 bottles per minute.  Implementing the recommended
improvements could significantly improve system performance.
 REFERENCES
 1.    U.S. EPA, 1992. Characterization of Municipal Solid Waste in the United States: 1992 Update.
      EPA/530R-92-019, Solid Waste and Emergency Response, Washington D.C.
                                                75

-------
2.   wTe Corporation, MITE Program for Automatic Sortation Process for Post-Consumer Plastic
     Containers, Draft Report, 1993.


3.   wTe Corporation, MITE Program for Automatic Sortation Process for Post-Consumer Plastic
     Containers, Draft Quality Assurance/Quality Control Project Plan, 1993.


4.   wTe Corporation, MITE Program Evaluation for Automatic Sortation Process for Post-Consumer
     Plastic Containers,  Monthly Technical Progress Reports, 1992.
                                              76

-------
               TWO TREATMENT OPTIONS FOR CHROMATED COPPER ARSENATE
                                WOOD PRESERVING RESIDUES

                                       Ronald J. Turner
                              U.S. Environmental Protection Agency
                                26 West Martin Luther King Drive
                                    Cincinnati, Ohio  45268
                                         513-569-7775

                                             and

                                       Mary Beth Foerst
                                        IT Corporation
                                      11499 Chester Road
                                    Cincinnati, Ohio  45246
                                         513-782-4700

INTRODUCTION

    Pentavalent arsenical compounds of varying formulations have been used as wood preservatives in
substantial quantities for over 50 years.  Currently, there are five arsenical wood preservative formu-
lations listed in the American Wood Preservers Association Standards: Types A, B, and C chromated
copper arsenate (CCA), ammoniacal copper arsenate, and ammoniacal copper zinc arsenate. Type C
wood preservative contains 34 percent arsenic, 18.5 percent copper, and 47.5 percent chromium.  Type
C is the predominant arsenical wood preservative used in the United States for applications such as
decks, docks, foundation and marine piling, fences, and utility poles (1).

    Arsenical wood preservatives are applied by pressure  processes.  Residuals consist of wastewaters,
drippage, sediments, spent formulations, and filter screenings.  Such wastes are classified under the
Resource Conservation and Recovery Act (RCRA)  as EPA Hazardous Waste Code F035.  Inorganic
wood preserving processes typically have no net generation of wastewaters because the water is recy-
cled back to the work tank.

    The U.S. Environmental Protection Agency (EPA) is evaluating performance data from treatment
alternatives including stabilization and recovery via leaching of CCA wood preserving wastes; these
wastes possess hazardous characteristics for arsenic and chromium.  The  standards for F035 are
scheduled for proposal in May, 1993.  These standards will not address the disposal of CCA-treated
wood.  Until promulgated, F035 residues will be classified as D004 (arsenic) and D007 (chromium).

WASTE CHARACTERISTICS OF ACTIVE FACILITIES

    Three CCA wood treatment facilities were visited for purposes of sampling and analyzing their recy-
cle  and waste streams.  Metals data for wastes streams from these facilities are summarized in  Table 1.
The highest arsenic concentration in any of the residual solids was 150,000 mg/kg; the highest chromi-
um  concentration was 90,000 mg/kg.  Copper was also present in concentrations up to 10 percent.
Trace quantities of acetone and n-butyl alcohol were the principal volatile organic  compounds detected
in the drippage and treatment cylinder sumps.
                                              77

-------
    The facilities generally manage their process residuals by contracting with a commercial waste
 removal  company or relying on the suppliers of the inorganic preservatives for waste removal. There are
 currently three such suppliers in the United States.

 SOUDIFICATION/STABILIZATION OPTION

    Treated and untreated F035 samples composited from several CCA wood preserving sites were
 obtained from the Chemical Waste Management (CWM) facility at Emelle, Alabama.  The untreated mate-
 rial was mixed with a backhoe in a rolloff container and samples were collected.  Untreated samples
 from CWM were also transported to  the Waterways Experiment Station (WES), U.S. Army Corps of
 Engineers, Vicksburg, Mississippi; and the U.S. EPA Test & Evaluation (T&E) Facility, Cincinnati, Ohio,
 for separate solidification/stabilization (s/s) or metals extraction tests. The arsenic concentration values
 of the untreated material ranged from 21,000 to 77,000 ppm.

    The results of the s/s tests are summarized in Tables 2 and 3. These data show that conventional
 s/s with  cement, kiln dust, or lime/fly ash binders (Table 2) did not effectively immobilize the arsenic or
 chromium to the 5.0 mg/L regulatory limit characteristic values as determined by the Toxicity
 Characteristic  Leaching Procedure (TCLP).  However, the addition of ferrous sulfate by CWM to reduce
 the hexavalent chromium to trh/alent chromium resulted in substantially lower leachable concentrations
 of arsenic in the cement-stabilized products (arsenic:  9 mg/L by WES vs. 0.12 mg/L by CWM;
 analyses conducted by same laboratory).

 METALS LEACHING OPTION

    Copper, chromium, and arsenic may be leached from F035 residual solids under proper conditions.
 This option may be preferable to s/s and land disposal if the separated metals are recovered for reuse.
 Bench-scale extraction tests were conducted by EPA at the T&E,  and by Lewis Environmental, Inc.,
 under a Small  Business Pollution Prevention Grant. The latter  process uses strong sulfuric acid leaching
 followed  by water washing and carbon adsorption to separate the metals for recovery or for potential
 reuse in the wood  preserving process. The T&E tests used 1M solutions of sulfuric acid, hydrochloric
 acfd, sodium hydroxide, and ammonium hydroxide.  The metals data from the inhouse leaching study
 are shown in Tables 4 and 5. The results reported by Lewis indicate that the samples pass the TCLP
 tests for arsenic and chromium in the acid-leached F035 wood preservative waste (2).  Because the T&E
 tests were conducted under much lower concentrations of extraction agents, metals  extraction was not
 as efficient.  A 3-hour extraction with sulfuric acid produced a residue that passed the TCLP.

    Although it appears technically feasible to recover the metal values from F035 residues, there is
 presently little  interest from metal recycling facilities in accepting  wastes with large amounts of arsenic.
The CCA wood preserving chemical suppliers and wood preservers are encouraged to  continue their
 efforts towards reuse.  Arsenic fixation processes other than s/s may also be applicable (e.g., Cashman
 Process), but additional studies may  be required.  Recycling of waste CCA-treated wood into useful
 products is another area for research.

 REFERENCES

 1. Baldwin, W. J.  Reuse of Wood Preservative That Contains Arsenic. EPA/600/R-92/105.  8/92.

2. Lewis, T. Potential  For Recovery of CCA From F035 Wood  Preserving Operations EPA/600/R-
   92/105.  8/92.
                                               78

-------
LU
CD
Z  -K
I-H OO
—I LU
0. tH
2 H-
< H-l
00 —I
    W
LOO
CO «t
O lo-
ll.
    CD
S, Z
OM
o; H-
    o
00 O
 LU
 _J
 CO
                  oo
                   a
                   a:
                   00
                   a
                   o    -
                   CO CT> Cn
                      JxJ T-
                   _l\ CU
                   <: ens
                   I— i     S-
                   oo    -a
                   LU
                   a:
1
LU "
LU
                   0:0-

                   LU t
                   a:
                       LU
                       0
                       LU
                                                             in   c\j
                                                                                       LOOO
                    '
                        c\icvi
              oo
                                        '
                                       o

                                       §

                                       m

         ,_
         •
         "'
o             mo
«d-CsJLf)     -CO
  >   i    o   1-1     ^
CM   i-~     •    i   CM
 i     .   o   m    i
oov     .0
o             oo
1^-                  00
o
0

                                                 ..
                                                 "2
                     cs   _.
                      '
          1     I
         CVJO
         r-tO
              0

              PO
  "CO
  -    i
     CM
       •
     O
                        LO
                                               V
                                  o   o

                                  §   2

                                                        0
                                                        0
                          0
                          OO
                                  CO  •— 1
                                          i— <>-
                                                                        <_>
                                                                                             —
                                                                                        — 1
                                                IT)
                                             o cu
                                             re S-
                                             y^  ,1J
                                             en i

                                             cu co
                                             c o
                                             o a.

                                                cu
 o -a

 co  co
 i-
 to   •> co
 CU  CO  C
 >> O. CU
    E  cu
CM  3  i-
 V  CO  O
       CO
-a  s-
 C  CD  S-
                                             *5=
                                                    CD

                                              +" C +•»
                                             O  CU
                                             CO  E E
                                                 +-> O
                                              eu  to s-
                                              i-  co t*-
                                              cu  s-
                                              3  4J CO
                                                    -o
                                             •o  co ••—
                                              cu  -c i—
                                             4->  4-> O
                                             •r-     CO
                                              CO   *•
                                             •r-  CO l—
                                              ^*  ^^ fl3
                                                 tO 3
                                              co  CU "O
                                              CU  S- -r-
                                             •i—  4-1   CO CU

                                             i—  CU

                                              0*0 .
                                              fO    .
                                             <+-  o
                                                 CO '
                                              cu
                                              cu
                                              L.  CO CO
                                             -C  CO CU
                                              +J  CUr-
                                                 U CL
                                              CU  O E
                                             .C ^ tO
                                             H— ca. co
                                                                                                              s-  o
                                                                                                                        -a
                                                                                                                        cu
                                                                                    ot-
                                                                                    O) -i- q-
                                                                                    s-     o

                                                                                    CO  J-  CO
                                                                                    •i-  CO +J
                                                                                       +->  C
                                                                                    S-  to  CO
                                                                                    CU  S  3
                                                                                                  3  CO 4->
                                                                                                      (O  to
                                                                                                  ^3§
                                                                                                  to in  o
                                                                                                     co
                                                                                                  cu o  cu
                                                                                                  CO I
 to  cu
 O J=<+-
 CU I—  O

        CO
 CU   •  £=
 4-»  C  O
 CO  O  -r-
 fO T-  4->
 3 +->  «J
     (0  S-
 CO i—  4->
 3  3  C
 O  E  CU
 cu  s-  u
 3  O  C
- cr*4—  o
 to     u
     cu
 4-  >  4->
 O -i-  C

 C  (O  U
 o  >  •>-
 •i-  S-<*—
 4J  CD  -i-
 to  co  c
 s-  cu  en
 cu  t-  -r-
 c  a. co
 cu
 en 2  ^^*
     CD  co
 4->  C  CO
 CU     4->
 c  en re
     c u
 o  T- -r-
  s:  x-a
     •i- C
  eu  E -r-

  io  c co

     '*" (O
                                                                                        cu
                                                                                     i— CO  >jx:  (O
               CU C 4J
               > 
-------
Si
2
=s
=j UJ
           EC
           c/3



           ^S
           w
             00
           z s
CO
  UJ
W



1
cc
H
UJ CC
          w

          g
          UH
          W
          ca
                  CN
               u  s:
               < u
                            80

-------
o
• s-
X 0
-
1— co o; a:
Z Q£  <^
o CNJ •* o g
o CT! 2 <=> o
v "* v o

o o 2 oo
' "* ° " 0
v ~ co"

s => >-
=3 p— • o: cs:
•— < s LU =>
S O Q- Q 0
Q Q^ Q. e:C Qi
e£ ^ O LU LU
O O O 	 IS

i ° 5 i §




-*. LO _. rvi
* o 1 
                                                                                                             a.
                                                                                                          i- re
                                                                                                          O v>
                                                                                                          0) i—
                                                                                                          i-,
                                                                                                          c  re
                                                                                                          re -o
                                                                                                          c  
                                                                                                          •r-  S-
                                                                                                          S  3
                                                                                                          re j=
                                                                                                           O i— t
                                                                                                           O 10
                                                                                                           re  o
                                                                                                             If
                                                                                                          •a to
                                                                                                           a>
                                                                                                          4->   •
                                                                                                           re i —
                                                                                                          i —  O  •
                                                                                                           3 > CO
                                                                                                           a>\ a;
                                                                                                           cu  s-r-
                                                                                                           S-  O) CL
                                                                                                             •M e
                                                                                                          4-  co re
                                                                                                           o -r- co
                                                                                                              CJ>
                                                                                                           co  cu
                                                                                                          r— 0£.£2
                                                                                                           cu     re
                                                                                                           > i— S-
                                                                                                           cu  re 01
                                                                                                          i —  S-
                                                                                                              cu i—
                                                                                                           i--o re
                                                                                                           o  a> s-
                                                                                                             u_ a>
                                                                                                           co     >
                                                                                                           c  s- cu
                                                                                                           o  cu co
                                                                                                          •r-  O.
                                                                                                          4->    (4-
                                                                                                           re  co o
                                                                                                           &-  re
                                                                                                          +J     co
                                                                                                           c  cu cu
                                                                                                           a>  s- -M
                                                                                                           o  re -i—
                                                                                                           c:     co
                                                                                                           o  c o
                                                                                                           o  o a.
                                                                                                              co e
                                                                                                           S-r- O
                                                                                                           =J  t. O
                                                                                                           E  re
                                                                                                          •i-  Q. OJ
                                                                                                           X  E 5-
                                                                                                           re  o cu
                                                                                                          s:  o s
             c
             re
             re
             o
             o
             re

             CO
             re

             •a
          •a cu
          CU 4->
          NI re
          re  en
          c  cu
          re  s-
          o  o
          zz

           I   I
                                                            81

-------
•*
S
m
s
o
O -J ^
** LL £
O— « V> Z ""> O W"! — '


°oor^S ~" "^ — • rn
CN CD j>»* ^S -^ C3 C3
^^ CJ *"•'• ^^ ^™^ ^i^

2°°§g«;gQQS
(NOdo'^dZZo



§31131111


^tNggcnQQ-os
S ~" o o °^ Z Z o d


= 2g|S|||§




S 2 § s i 1 1 § |





n S § b ^
CJ *^ >-, C erf C5
g 1 s s g D*i
SSSoSaoSz
c^ajQoia.<w
<«oo8sS«
                                                                            *
                                                                            o
                                                                         .-2 §
                                                                         «i C
                                                                         s *o
                                                                         c*-  C
                                                                         •3  «
                                                                         co  cu
                                                                            O
                                                                           •o
                                                                     «    *«-
                                                                               CU ro
                                                                         00 w
                                                                         .2 <=>
                                                                         cu
     03 H
     CO 5
    JJ
     •*-» o
     3 <2
     is
    i§
    •a o
    e ^
    §=3
    a s
    Ui *-•
    CU 03
    a c

    S.-2
    « 'o
    > 73
    CO O
                                                                               « O


                                                                              fl
                                                                               S"S
                                                                              W Js
           T3
            03
           •a
           •a
           CU
            03
            Q.
           a
           's
           o
           CO
                                                                                CU
                                                                                Q.
JS
O
03
CU
           T3
           C
           03
                                                                                     B
                                                                                     Q.
                                                                                     4-*
                                                                                     73

                                                                                     ."S
                                                                                     'o
                                                                                     CO
                                                                     +_*  1—• **•*  ^rf
                                                                     •3  _ B  <*;
                                                                     S   - '£  "3
                                                                     S  -S is
                                                                                     •o
                                                                                     cu
                                                                                     E

                                                                                     I
                                                                                     cu
                                                                                     Q.

                                                                                     a
                                                                                     cu
                                                                     5   a C   -r
                                                                     ^   fli O   17
                                                                                       »—»
                                                                                       a
                                                                                       o
                                                                                       cu

                   •8-i
                                                                                               I
                                                                                             43  a
                                                                                             if
                                                                                             CU  O
                                                                                         CJ
                                                                                       CO  CO

                                                                                       O-

                                                                               a S
                                                                                          oo
                                                                                       o  o
                                                                                       ZZ
                                                                                       II  II
                                                                                       Q oi
                                                                                       ZZ
                                           82

-------
     en

     z
     U4
     DC
     111
     O.
     O
in

m

CQ
DC
o
u.

<


I
           HI

           s
           O
                                 O


                                 <
                            Orj
                            m vj
                 5   !-J.




                 <«1
                 W J



                 < tu
                    Z
                    W
                    O


                    O
                    Z

                    5
                    u


                    3
W
O



i
H-4
E
U

W
                                E
                                O
                                DC
O
Kt

Z
                                sc
                                0
                                ^
                                s
                                 DC
                                 o.
                                      K
                                 E
                                 a
            DC
             a
      DC
       a
                                 o.

                                                            =qQQQQ
                                                                                 o
                                                                                 • ~*
                                                                                 t_l
                                                                                 o
                                                                                 o
                                                                                 o
                                                                                 •o
                                                                                 •o
                                                                                 c
                                                                            s§
                                                                            O TJ
                                                 53  i
                                                                            S.s
                                                                            IH C
                                                                           .0 O
                                                                              S
                                                    c
                                                    a
                                                    x

                                                 £  2
                                                 ft}  r*


                                                 B  fi
                                                 00 t3

                                                 .S g
                                                                                 2 -o
                                                                                 5s
tn


15
*s a


"o c
en rt
00 §
C f

*M U.
S ,2


IE
S o.
1- nn
                                                    8
ments

atural
                                                                                   2    S

                                                        W ^
                                                           A
                                                        •o c

                                                        5'a
                                                                                         « o

                                                                                         °2s
                                                                                        •sd
                                                                                         o o
                                                                                        Z Z
                                                                                         n H

                                                                                        Q OH
                                                                                        zz
                                             83

-------
           DISCOVERY AND DEVELOPMENT OF ANAEROBES FOR BIOREMEDIATinN)

                    James M. Tiedje, John F. Quensen, III, Joann Chee-Sanford,
                                      Stephen A. Boyd
                    Michigan State University, 540 Plant & Soil Sciences Building
                              East Lansing, Michigan, 48824-1325

  INTRODUCTION

     Chlorinated organic chemicals probably constitute half of the environmental organic pollutant
  problems in the world today.  The ability of aerobic microorganisms to degrade these chemicals is
  often limited by chlorine blocking the site of enzymatic attack. Consequently, some of these
  compounds are persistent in the environment.  In the last  decade, the ability of anaerobic
  microorganisms to reductively dechlorinate some of these chemicals  has become more widely
  recognized. This process has the following  advantages:  (i) reduction in the degree of chlorination
  making the product more  susceptible to mineralization by aerobic microorganisms if it is not
  completely degraded by the anaerobic community, (ii) reduction  in toxicity of the parent compound
  and (iii)  the relative ease  of establishment of  the appropriate in situ conditions  conducive of
  dechlorination in many environments that contain these pollutants, i.e., principally, establishment of
  anaerobic conditions which usually occurs following water saturation of soils and sediments.  Hence
  reductive dechlorination is important and often an economically feasible technology  to  be
  considered for remediation of chlorinated pollutants.

     Bioremediation is a much simpler technology to manage if the pollutant provides a strong natural
  selection for growth of the biodegrading organisms.   The more challenging situation  is  for co-
  metabolic processes in which  substrates other than the pollutant are  needed to  stimulate a
  secondary and often indirect  biodegradation process.   Which  case  describes  reductive
  dechlorination?  Evidence suggests that  reductive  dechlorination does provide  a  selective
  advantage to dechlorinating  organisms (1,2),  although the resultant growth advantage is not as great
  as for many aerobic organisms that grow on the pollutant chemical.  The advantage of reductive
  dechlorination to anaerobes is not as a carbon  and energy source, but as an electron acceptor
.  (oxidant).

     It is now becoming the rule rather than the exception to find that an organochlorine compound is
  reductiveiy dechlorinated.  More than 50 organochlorine compounds have now been shown to be
  reductively dechlorinated, including many organochlorine pesticides (e.g., DDT, heptachlor alaclor
  2,4-D),  chlorinated solvents  (e.g.,  PCE,  TCE, CHCI3, TCA),  and arylhalides  (e.g.,  PCS,
 chlorobenzenes, chloroguaiacols, chloroanilines)  (3,4,5).  From these studies and  others the
 following important generalizations can be made:

      1.    Microorganisms capable of reductive  dechlorination seem to  be relatively ubiquitous in
 nature, at  least in anaerobic environments.   The presence of  these organisms in aerobic
 environments is not yet well investigated.

      2.    Reductive dechlorination requires anaerobic conditions, with a few exceptions.

      3.    Reductive dechlorination is a relatively slow process, with rate measurements often in
 terms of days to weeks, but it is not so slow as to be considered unfeasible, especially for in situ
 treatment.

      4.    In  general, there appears to be a specificity between  particular dechlorinating
 populations and particular chemicals. There is yet  no evidence for broad spectrum dechlorinators.

      5.    Most dechlorinating cultures are consortia (mixed natural communities) and it has been
 very difficult to isolate in pure culture active dechlorinators from such communities  Thus fermentor
 scale growth for inoculum to stimulate bio remediation is not yet feasible
                                           84

-------
 METHODOLOGY

    Studies on PCB dechlorination have been conducted under three related conditions: (i) extent
 of in situ dechlorination (6), (ii) mixing sediments with indigenous PCB dechlorinating populations
 with other sediments or PCBs, and (iii) eluting PCB-degrading populations from sediments and using
 this population to inoculate other PCB contaminated soils or sediments.  The latter approach also
 proved that the observed transformation of PCBs to lesser chlorinated products was dependent on
 the microbial inoculant (7). The PCB analytical data are from congener-specific analyses of the PCB
 components which are then corrected to molecular weight and summarized for all products based on
 those with chlorines in the meta+para positions versus those with chlorine in the ortho position (7).

 RESULTS

    The rate of PCB dechlorination is affected by a number of factors. Those that have been studied
 can be summarized as follows. PCB concentration: Optimum rates of PCB dechlorination usually
 occur for concentrations in the range of several hundred to 1,000 ppm (w/w) of sediment. Below 50
 ppm, dechlorination is often very slow or non-measureable.  Bloavaliability:  PCBs may be
 dissolved in the organic  phase of soils or sediments and are perhaps also protected under waxy
 layers of  aged organic matter and do not reach biodegrading organisms.  The requirement for
 moderately high concentrations of PCBs for a more rapid dechlorination may be partially a result of
 the lesser availability of the lower PCB concentrations.  Inhibitors:  PCBs are often not the sole
 contaminant present. Other contaminants such as oil and grease, heavy metals, and solvents can be
 toxic to organisms and restrict or eliminate dechlorination. We have found large concentrations of oil
 and grease to be the most commonly encountered inhibitors of dechlorination in  sediments.
 Temperature: Temperatures of  12 and 25 "C support Aroclor 1242 dechlorination. Temperatures
 of 37  *C or above showed no  dechlorination, providing further  support for the microbial  and
 enzymatic nature of this process.  Dechlorination at 12 °C  is an important result  because this
 temperature is a reasonable environmental temperature in temperate regions.  Nutrients: electron
 donors and mineral nutrients are needed for microbial populations. In an anaerobic environment
 mineral nutrients  are rarely limiting,  but quality electron donors  may  become limiting in deep
 sediments because these zones are usually isolated from a new supply of carbon from the water
. column. Research has shown that new carbon additions do stimulate PCB dechlorination  rates at
 least in some cases. This has been achieved in two ways:  First, adding  a readily available carbon
 source to sediments such as methanol, glucose, acetone or acetate stimulated dechlorination rates
 by a factor of two to three (8). The second approach is by mixing or other physical disruption of
 sediments so that native carbon is made more available. In this case, added carbon usually provides
 no further enhancement of the dechlorination rate.  A number of investigators have made many
 attempts  to add various carbon, mineral and  co-factor amendments in an  attempt to stimulate
 dechlorination.  So far, only marginal increases in dechlorination  have  been noted at best.  For
 implementation of an invasive bioremediation process, stimulation of the PCB dechlorination rate by
 an order of magnitude likely would be necessary.

    As is true for all Aroclors, chlorine is preferentially removed from meta and para positions while
 chlorines in the ortho position are preserved (7).   Thus, complete dechlorination of PCBs is not
 common,  although Van Dort and Bedard (9) have  recently shown that dechlorination can be
 observed from the ortho position if the microbial  community is first induced by adding an ortho
 enriched congener.  The PCB dechlorination,  although incomplete, does  result in risk reduction.
 The coplanar PCBs are the components of commercial mixtures with the greatest demonstrated
 toxicity.  These are structurally similar to 2,3,7,8-tetrachlorodibenzo-p-dioxin and exhibit dioxin-like
 toxicity.  However, this toxicity is dependent on chlorines in the  meta and para positions of the PCBs,
 and it is these chlorines that we would expect to be removed. To investigate this directly,  we
 measured the  dechlorination of a  toxic tetrachlorobiphenyl  (34-34-CB)  together with other
 tetrachlorinated biphenyls present in Aroclor 1242.  All three tetrachlorobiphenyls measured  were
 removed at equivalent rates. We have since confirmed that toxicity is reduced by toxicity bioassays.
 Thus, reductive dechlorination does reduce PCB risk, even though complete degradation does not
 occur.  It should also be  pointed out, however, that the ortho enriched PCBs are not major
 components of Aroclor  mixtures and, therefore, have not been subject to  extensive toxicology
 testing.
                                            85

-------
    Despite reduction in toxicity, legal statutes may dictate removal of RGBs beyond that achieved by
 anaerobic organisms.  Because the  dechlorinated products can be metabolized by aerobic
 microorganisms (10) it is feasible to sequentially couple anaerobic and aerobic treatment. Substantial
 reduction of  PCBs  by  the sequential treatment has been demonstrated in the laboratory.
 Successful field application of this concept depends upon the following:  (i) extensive dechlorination
 of the PCBs so that the aerobic organisms can more efficiently mineralize the residual PCBs, (ii) most
 known aerobic PCB degraders grow on biphenyl and not the chlorinated PCBs, so the substrate to
 enrich the appropriate aerobic PCB-degrading population may not be present.  Since biphenyl
 degrading enzymes are close relatives  of those involved in toluene and naphthalene  degradation,
 co-contamination with petroleum derivatives  may actually aid the aerobic phase  of  the PCB
 treatment, and (Hi) successful distribution of air or h^C^ in the anaerobic matrix to support sufficient
 aerobic metabolism. General Electric researchers have now successfully demonstrated PCB removal
 using the sequential anaerobic-aerobic concept in a pilot field study.

    The most highly chlorinated Aroclor,  1260, which is completely resistant to aerobic  microbial
 attack  has also been shown to be subject to the dechlorination by anaerobic bacteria.  Figure 1
 shows the lesser chlorinated products  produced when a sediment containing Aroclor 1260 was
 inoculated with an anaerobic  microbial enrichment from a PCB contaminated  sludge  industrial
 lagoon.

    Remediation by reductive  dechlorination  involves insuring that the  contaminated soils  or
sediments are anaerobic.   In the case of sediments, this may naturally be the case. In the case of
soils, it may be necessary to flood the soils if leaching can be prevented or may be partially achieved
by adding a cheap  available carbon source.  For a habitat that is naturally anaerobic, such as
sediments, it has been important to estimate the existing rate  and  extent of dechlorination.
Remediation schemes may be:  (i) to continue the in situ incubation, (ii) to attempt to stimulate the
dechlorination  rate by  mixing the sediment and/or adding soluble carbon, or (iii) to cap the site to
ensure anaerobiosis and containment.
        so
        40-
        30-
      o
      "o
        20-
        10-
OWeek

20 Week
            Mono    01     Trl   Tetra   Ponl.-i   Hexa   Hepta

                                         Homolog
                                                       Octa
                                                             Nona
                                                                   Oeca
           Figure 1.  Dechlorination of 5,000 ppm of Aroclor 1260 in a sediment
           inoculated with a microbial population enriched from a PCB-
           contaminated industrial sludge lagoon.
                                           86

-------
    Remediation of soil by reductive dechlorination has been less well studied.  However, in the one
case we have studied, a soil contaminated with transformer oil (Aroclor 1254) was dechlorinated if
PCB dechlorinating organisms from the Hudson River sediment were added to the  soil and
incubated under anaerobic conditions (Figure 2). The transformer soil was also mixed 1:1 with clean
upstream Hudson River sediment, which may have provided nutrients to the soil community.  If the
soil was simply made anaerobic, very little dechlorination was observed within the  24-week
incubation. The dechlorination of this naturally contaminated soil occurred at an equivalent rate
constant to that found for Aroclor 1254 added to Hudson River sediment. Thus, the soil remediation
was equal to the best rate of PCB removal achieved in sediment.

CONCLUSIONS

    Reductive dechlorination of chlorinated organic compounds, including PCBs, is a remediation
process that  should  be considered  as it is often easy to  achieve  and may be the cheapest
bioremediation technology.  Before reductive dechlorination of PCBs can  be a widely accepted
treatment, it must be optimized and field tested so that the target concentrations can be achieved.
The greatest needs are a means to enhance the rate of reductive dechlorination and schemes to
insure success of the subsequent aerobic metabolism in a coupled anaerobic/aerobic process. In
many cases,  it  may also be necessary to  enhance the bioavailability of the PCBs to  the PCB
degraders. Nonetheless, research progress in recent years has now made it worthwhile to field test
methods for PCB bioremediation.
       ri
       O
                      Trnuf ormor Soft
•r*
ra
«
P«
n
0)
4
o
2
u
0)
03
O

3-


2-

1-

n
*f*iuuorui«r *
* *~""-*\ _»__
^-—


"^


T-* Uata «
*-* Ortho
0 I . | i ,
0 8 16


-T HI





c Para

1 1
24
                                               3-
                                                1-
                                               o-J
                                             Transformer Soil
                                                           1
                                                           8
                                                    16
Weeks of Incubation
                                                    Weeks of Incubation
           Figure 2. Anaerobic reductive dechlorination of a soil contaminated by
           a transformer spill of Aroclor 1254. The left figure shows dechlorination
           when the soil was made anaerobic and the right figure shows
           dechiorination when the soil was inoculated with PCB-dechlorinating
           organisms eluted from the Hudson River and mixed 1:1 with clean
           Hudson River sediment. Chlorines substituted in the ortho position
           (A) and meta + para positions (T)
                                           87

-------
 ACKNOWLEDGMENT


    The authors' work is supported by the General Electric Corporation and The Great Lakes - Mid
 Atlantic Environmental Protection Agency Hazardous Substance Research Center.

 REFERENCES


 1.    Dolfing, J., and J.M. Tiedje. 1987.  Growth yield increase linked to reductive dechlorination in a
      defined 3-chlorobenzoate degrading methanogenic co-culture.  Arch. Microbiol  149-102-
      105.
 2.
3.
4.
5.
6.
7.
8.
9.
 Mohn, W.W.,  and J.M.  Tiedje.   1991.  Evidence for chemiosmotic coupling of reductive
 dechlorination  of ATP synthesis in Desulfomonile tiedjei.  Arch. Microbiol. 157:1-6.


 Kuhn, E.P.,  and J.M.  Suflita.   1989.   Dehalogenation  of pesticides  by anaerobic
 microorganisms in soils and groundwater - a review, p. 111-180. In Sawhney and Brown (eds.)
 Reactions and Movements of Organic Chemicals in Soils. Soil Sci. Soc. Am and Am Soc
 Agron., Madison, Wl.


 Mohn, W.W., and J.M. Tiedje.  1992. Microbial reductive dehalogenation.  Microbial Rev (in
 press).


 Vogel, T.M., C.S. Griddle, and P.L. McCarty. 1987. Transformations of halogenated aliphatic
 compounds. Environ.  Sci. Technol. 21:722-736.


 Brown, J.F., D.L. Bedard, M.J. Brennan, J.C. Carnahan, H. Feng, and R.E. Wagner  1987
 Polychlorinated biphenyl  dechlorination in aquatic sediments. Science 236:709-712.


 Quensen, J.F.,  III, J.M. Tiedje, and  S.A. Boyd.   1988.  Reductive dechlorination of
 polychlorinated biphenyls by anaerobic microorganisms from sediments. Science 242752-
 754.


 Nies, L., and T.M. Vogel.  1990.  Effects of organic substrate on dechlorination of Aroclor 1242
 in anaerobic sediments.  Appl. Environ. Microbiol. 56:2612-2617.


Van Dort, H.M.,  and D.L. Bedard.  1991.   Reductive ortho and meta dechlorination of a
polychlorinated biphenyl  congener by anaerobic microorganisms.  Appl. Environ. Microbiol.
57:1576-1578.
10.  Abramowicz, D.A. 1990. Aerobic and anaerobic biodegradation of PCBs- A review  Grit Rev
     Biotech.  10:241-251.
                                          88

-------
   FEATURES OF BTEX-DEGRADING MICROORGANISMS FROM OXYGEN-LIMITED (HYPOXIC^
                                      ENVIRONMENTS

                                       Mark D. Mikesell
                                       Jerome J. Kukor
                                       Ronald H. Olsen

                           Department of Microbiology and Immunology
                              University of Michigan Medical School
                                   Ann Arbor, Ml 48109-0620

INTRODUCTION

       Strategies for the bioremediation of hydrocarbon-contaminated aquifers rely mainly on providing
oxygen to the contaminant plume, easing the most commonly encountered limitation on biodegradation,
namely, the inadequate supply of an electron acceptor. This may be accomplished by adding air or pure
oxygen or by adding hydrogen peroxide as a source of oxygen. The utility of adding gaseous oxygen is
limited by its expense and its low solubility, and by the fact that oxygen consumption in contaminant
plumes can be rapid (Wilson et al, 1986).  Whether the source of oxygen is the added gas or the
decomposition of hydrogen peroxide, a major problem is the reaction with metal oxides, leading to
reduced permeability of the aquifer and plugging of delivery systems (Lee et al., 1988). Since nitrate can
serve as an electron acceptor at reduced oxygen concentrations, resulting in denitrification, the possibility
of exploiting denitrifying bacteria for the bioremediation of groundwater contaminated with aromatic
hydrocarbons is attractive (Britton, 1987).  The high water solubility of nitrate and the comparative ease
with which it can be delivered to contaminated zones, in contrast to oxygen, enhance its potential
feasibility as a candidate for addition to a contaminated aquifer for the purpose of stimulating in situ
biodegradation.

       Our work developed out of an interest in the behavior of bacteria resident in benzene, toluene,
ethyibenzene and xylene(s) (i.e., BTEX) contaminated aquifers. We observed that such environments are
oxygen-limited, yet continuous monitoring over extended periods suggested that BTEX contaminants
were disappearing and the contaminant plume may be diminished. Such observations suggested that
significant microbial populations were active in situ. Moreover, laboratory studies with microcosms
suggested the ascendancy of bacterial populations under hypoxic (i.e., oxygen limited) conditions whose
growth and BTEX-degrading.activities were associated with the reduction of nitrate under conditions
whereby  oxygen concentrations were less that twenty percent of saturation.

        To evaluate factors affecting the effectiveness of using denitrification as a component in a
groundwater bioremediation strategy, we investigated bacteria from a hydrocarbon-contaminated aquifer.
As part of a commonly-used type of remediation strategy, hydrocarbon-contaminated groundwater is
pumped through beds containing granular activated carbon (GAG) particles (Voice, 1989). It is well
known that in this sort of "pump and treat" operation GAG systems adsorb aquifer bacteria as well as
contaminants such as BTEX, and that biodegradation by adsorbed bacteria can be credited with
extending the duration of usefulness of such systems (Bouwer & McCarty, 1982; DeLaat & Bouanga,
1985). We were given access to samples of such a GAG adsorption system, and since it was likely that,
due to the presence of organic contaminants (BTEX), oxygen levels were depleted in the aquifer as well
as in the GAG beds, we were interested in examining the GAG bacteria for activity under hypoxic,
denitrifying conditions as well as under aerobic conditions.  We also isolated bacteria from BTEX-
contaminated plumes a three sites and determined their characteristics as for the bacterial isolates which
obtained from GAG.

METHODOLOGY

       Granular activated carbon (GAG)  samples were obtained from an aquifer treatment system
composed of two trains operated in parallel, each with three stages arranged in series. This system had
been receiving petroleum-contaminated groundwater containing BTEX for approximately three years.
Samples of about 200 mi of GAG slurry were collected aseptically from each of the six GAG beds.
                                            89

-------
       Aquifer core samples were obtained using a Waterloo Cohesionless-Aquifer Core Barrel sampling
device. This sampling method permitted collection of Aquifer material in a 1.5 m core barrel from which a
portion was removed from each end, giving approximately 0.9 m of intact core suitable for microbiological
analysis. The core samples were sealed and stored in an upright position at 4 C until the enrichment
cultures were prepared.

       Bacteria able to utilize benzene, toluene, ethylbenzene, and p-xylene were isolated directly from
the GAC samples under aerobic and hypoxic (denitrifying) conditions. GAC was mixed with sterile water
(1:1 wt.: vol.) on a vortex mixer for 1 min.; after settling, 0.1 ml of the resulting supernatant was spread on
the surface of solid minimal medium containing (per liter) 20 g purified agar.0.1 g NH4SC-4, 0.02 g
MgSO4'7H2O, 0.01 g CaCIa, 5 mg FeSO4'7H2O, and 25 mg NaMoO4.2H2O, and buffered at pH 6.8 with
40 mM potassium phosphate buffer.  For hypoxic incubations the medium contained nitrate at 10 mM
(140 mg/l nitrate-N). Aerobic incubations with volatile substrates (BTEX) took place inside glass
dessicators with the hydrocarbon supplied as vapor by saturating a piece of filter paper (2x2 cm) with
100 u.l of pure compound. For the comparatively more soluble, non-volatile substituted aromatic
compounds such as the cresols and catechois the substrate was incorporated into the medium at a
concentration of 0.05%. Hypoxic incubations were performed in GasPak jars (BBL Microbiology Systems,
Cockeysville,  MD) with the substrate provided in the same manner as in the aerobic incubations.
Pseudomonas putida mt-2, a nondenitrifier containing the TOL plasmid pWWO, was included as a
biological control for hypoxic conditions. The hydrocarbons (BTEX) were provided individually as the sole
source of carbon. Aerobic and hypoxic plates were incubated for 48 and 170 hours, respectively. Of the
colonies appearing on each plate, eight were selected, picked and streaked for purification under the
same conditions as in the initial incubation. The colonies were selected on the basis of colony
morphology as well as the extent of growth so that a variety of organisms could be obtained. After a third
streaking for purity each isolate (eight on each substrate under each of the two incubation conditions,
hypoxic and aerobic) was incubated with each of the other substrates (benzene isolates incubated with
toluene, ethylbenzene, and p-xylene, etc.). Finally, the aerobic isolates were incubated with all substrates
under denitrifying conditions, and vice versa.  The extent of growth for the isolates was judged by visual
inspection of standard plate inoculations; growth was rated on a scale of one to ten, a score of one
corresponding to negligible growth, a ten denoting lush growth.

       Strain characterization was facilitated by use of Rapid NFT strips (Analytab Products, Inc.,
Plainview, NY). All aromatic hydrocarbons and other growth substrates were obtained from Aldrich
Chemical Co. (Milwaukee Wl) and were used without further purification.

       For incubation in liquid culture, serum bottles (160 ml capacity) with Teflon-lined rubber stoppers
and aluminum crimp seals (all from Wheaton, Millville, NJ) were used. The composition of the medium
was the same as that used for the direct plate isolations (above) but without agar.  After the addition of 0.5
g GAC particles to 100 ml medium the bottles were flushed for 10 minutes at approximately 1.5 l/min with
C-2-free gas mixture (85% N2,10%H2,  5 % CC>2). Before sealing, toluene was added to a concentration
of 1 mM (92 mg/l). The toluene was monitored by high  performance liquid chromatography and was
replenished when the concentration fell below 200 uM.  When the toluene had been replenished five
times the enrichments were transferred (10% v/v)  to freshly prepared medium and the process continued.
Isolates were obtained by inoculating solid medium and incubating hypoxically as described above.

       Batch incubation of pure cultures was performed in serum bottles as described above, but the
bottles contained 150 ml of minimal medium and were flushed with C-2-free gas for 15 minutes after
addition of an inoculum to an optical density (O.D.425)of 0.1. The inoculum was obtained by growing the
strain on solid minimal medium containing nitrate (10 mM) and incubated with toluene vapors for 48h..
The batch cultures were incubated with 250 uM each of benzene, toluene, ethylbenzene, and p-xylene
and sampled with a syringe at 1d, 3d, and 6d.  The initial dissolved oxygen concentration in hypoxic batch
cultures of approximately 2 mg/l was determined by polarographic C-2 measurement using a Clark 02
electrode (Yellow Springs Instrument Co., Yellow Springs, OH).
                                             90

-------
       Aromatic hydrocarbons and potential degradation products were analyzed by reverse-phase high
performance liquid chromatography.  Separation was performed on a 25 cm x 4.6 mm Spherisorb C-\ Q
column (Phase Separations, Norwalk CT) using a mobile phase consisting of methanol and 1 % aqueous
acetic acid in varying proportions depending on the separation requirements.  For benzene, toluene,
ethylbenzene,  and p-xylene, for example, good resolution was achieved using 70% methanol at 1.5
ml/min.  Compounds were detected by UV absorbance and identified by co-chromatography with
analytical standards.

RESULTS

       GAG Studies:  Inoculation of minimal medium with GAG slurry supernatant resulted in the
formation of numerous colonies utilizing each of the four hydrocarbons under both aerobic and  hypoxic
conditions. Numbers of isolated colonies were similar on all four substrates, ranging from 2 x 103 to > 1 x
105 organisms/ml of GAG slurry supernatant. A more precise determination of total numbers of BTEX-
utilizing bacteria is not possible, since our procedure for detaching ceils was not intended to be
exhaustive. Some samples yielded as many as four colony types while others appeared to contain a
practically pure culture, based on the colony morphology of the primary isolates. At this stage,  no clear
pattern emerged to distinguish the  isolates obtained on the four substrates or the two incubation
conditions. Also, the two GAG trains and the three stages in each train appeared to be similar in the
numbers and types of organisms obtained.The isolation protocol resulted in the selection of eight isolates
on each substrate for each set of growth conditions (e.g., toluene/aerobic, toluene/hypoxic, etc.).  We
tested each of the isolates for growth under both conditions and on all four substrates (i.e., B, T, E or X),
yielding results in the form of a set of tables.  From the data in these tables it was clear that he  substrate
used for the initial isolation and purification of strains did not always support the strongest growth.  For
example, benzene isolates generally grew better with toluene or ethylbenzene. This may be attributable
to varying strength of the compounds as effectors for the degradative pathway(s)  involved (R.H. Olsen et
al., unpublished results).  When the results for all the isolates, aerobic and hypoxic, are taken in
aggregate ethylbenzene stands out clearly as the best growth substrate, followed by toluene, with
benzene and p-xylene the least effective. In the work described by Ridgeway et at. (1990) initial isolation
was performed using gasoline, a mixture containing hundreds of aromatic and aliphatic hydrocarbons.
Almost 75% of their isolates were able to grow on toluene, which was followed by p-xylene and
ethylbenzene in frequency utilized.  These gasoline-degrading isolates were generally limited to two or
three structurally related compounds.

        We now began an enrichment strategy to obtain a strain from the GAG samples  that was able to
grow at reduced Q£ concentrations using toluene as a carbon source and nitrate as an electron acceptor.
This procedure resulted in the isolation of a bacterial strain able to metabolize BTEX under oxygen-limited
denitrifying conditions. This strain  has been identified as a Pseudomonas fluorescens and is designated
CFS215. Although several substitued aromatic compounds have been screened as growth substrates,
we have not yet determined the pathway(s) used for BTEX metabolism in this strain.

The nitrate dependence for metabolism of BTEX by strain CFS215 under hypoxic  conditions is illustrated
in the Figure below.
                                                              fey*
                                           Incubation tim«
                                              91

-------
       Clearly each of the BTEX compounds examined is metabolized under hypoxic conditions when
nitrate is present and not with ammonia only. Further studies are being directed at more complete
measurements including monitoring oxygen, nitrate, nitrite, and the hydrocarbon over time, allowing more
complete understanding of the process for CFS215 and other bacteria which obtained from these studies.

       Core studies: We also studied bacteria isolated from soil cores withdrawn from BTEX
contaminant plumes. Samples for the isolation of these subsurface bacteria from BTEX-contaminated
aquifers were incubated in mineral salts buffer overnight to elute bacteria from particles. These eluates
were then plated on solid mineral salts medium followed by incubation aerobically or under hypoxic
conditions in the presence of BTEX vapors. Aerobic cultures were incubated 48 h and hypoxic cultures 1
week. Data representative of three sites sampled (See table below)show that the total number of colony
forming units (CPU) is uniform throughout the cores and that similar counts were obtained for bacteria
able to use BTEX as a sole carbon source growing aerobically. This does not imply, however, that
identical isolates occur on both nutrient agar (for total count) and mineral salts-BTEX medium since many
microorganisms growing on the latter often show suboptimal or no growth on complex nutrient medium. A
distinctive feature of the microbial counts is suggested by the development of significant populations able
to grow under hypoxic conditions during the 4-week incubation period. The ascendancy of these
populations correlates also with the overall BTEX degradation activity for the corresponding microcosms.
Furthermore, denitrifying activity corresponded to BTEX degradation and the emergence of BTEX
degrading bacteria. Three correlates, then, obtained from such analysis: BTEX degradation, denitrifying
activity, and the enrichment of CFUs growing hypoxically on the mineral salts-BTEX medium.  Similar
results obtained from the other two sites examined.

                           Denitrification and Viable Counts for Site "A"
                    Viable Count (CFUVml enrichment culture
        Total   Aerobic
Depth,  BTEX   Total

 (m)    score3   t=0 wk
                          Aerobic
                          BTEX
Hvpoxic with BTEX
                          t=0 wk   t=0 wk   t=1 wk   t=2 wk   t=4 wk
   Denitrificationk

0 wk 1 wk  2 wk 4 wk
1.7
2.7
3.4
4.7
6.3
7.8
0
2
4
2
0
0
3x1 06
2x1 06
1x106
6x1 05
1x1 06
4x1 05
.3x1 06
2x1 06
8x1 05
1x106
8x1 05
7X1 05
<10
<10
<10
<10
<10
<10
<10 1x102
<10 1x1 O4
2x1 O3 6x1 O4
<10 6x1 O3
<10 <10
<10 <10
8x1 03
2x1 04
4x1 05
7x1 04
<10
<10
...
+ + + +
++ ++ ++ ++
+ + + +
-
-
a BTEX score is a measure of BTEX degradation activity; the higher number indicatesgreater
  degradation; a score of 0 indicates no degradation compared to sterilized controls.
b -, no apparent nitrate disappearance, no nitrite produced; +, ca. 0.5 mg/l nitrate N; ++, ca. 1 mg/l nitrite
  N.

CONCLUSIONS

       A diverse population of BTEX-degrading bacteria developed on the surface of GAC particles
utilized in a pump-and-treat restoration scheme for a petroleum-contaminated aquifer. We isolated and
partially characterized many bacteria able to utilize benzene, toluene, ethylbenzene, and p-xylene under
either aerobic or hypoxic (2 mg I"1 02) conditions. Many of the isolates were characterized by a broad
                                             92

-------
substrate range, and activity under both aerobic and hypoxic conditions was dependent on the substrate
used for isolation, with ethylbenzene and toluene most likely to yield the strains with the broadest range of
activity. Using an enrichment culture technique we isolated a strain of P. fluorescens with which we
demonstrated nitrate-dependent BTEX metabolism under hypoxic conditions.

       These results, together with results published elsewhere, indicate that strategies for in situ
bioremediation may require less oxygen than has been assumed to be necessary, providing nitrate is
present to serve as an alternate electron acceptor when conditions become hypoxic.

       We have also shown that bacteria able to utilize BTEX under hypoxic, denitrifying conditions can
be isolated from hydrocarbon-contaminated aquifers, and that the occurrence of these bacteria is
correlated with hypoxic BTEX degradation and denrtrification activity. Whereas aerobic BTEX degraders
were distributed uniformly in the sampled profile, bacteria able to utilize BTEX under hypoxic conditions
appear to be stratified in zones within the plume at each of the three sites examined.

REFERENCES

Bouwer E. J., & McCarty P. L. (1982) Removal of trace organic compounds by activated carbon and fixed-
       film bacteria. Environ. Sci.  Technol. 10:836-839.
Britton L. N. (1987) Aerobic denitrification as ah innovative method for in situ biological remediation of
       contaminated subsurface sites. Report ESL-TR-88-40. Air Force Engineering Services Center,
       Tyndall Air Force Base, FL
DeLaat J., & Bouanga F. (1985) Influence of bacterial growth in granular activated carbon filters on the
       removal of biodegradable and non-biodegradable organic compounds. Water Res.19:1565-1568.
Lee M. D., Thomas J. M., Borden R. C.,  Bedient P. B., Wilson J. T., & Ward C. H. (1988) Biorestoration
       of aquifers contaminated with organic compounds. CRC Crit. Rev. Environ. Control 18:29-89.
Ridgeway H. F., Safarick J., Phipps D,. Carl P., & Clark D. (1990) Identification and cataboiic potential of
       well-derived gasoline-degrading bacteria from a contaminated aquifer. Appl. Environ. Microbioi.
       56:3565-3575.
Voice T. C. (1989) Activated carbon adsorption. In Freeman H. M. (Ed) Standard Handbook of Hazardous
       Waste Treatment and Disposal (pp 6.3-6.21). McGraw-Hill Book Co., New York.
Wilson J. T., Leach L. E., Henson M. & Jones J. N. (1986) In situ biorestoration as a ground water
        remediation technique. Ground Water Monit. Rev. 7:56-64.
                                             93

-------
               EFFECTS OF NITROGEN SOURCE ON CRUDE OIL BIODEGRADATION

                    Brian A. Wrenn, Miryam Kadkhodayan, and Makram T. Suidan
     University of Cincinnati, Civil and Environmental Engineering Department, Cincinnati, OH 45221

                               John R. Haines and Albert D. Venosa
   U.S. Environmental Protection Agency, Risk Reduction Engineering Laboratory, Cincinnati, OH 45268


INTRODUCTION
       The use of fertilizers to stimulate biodegradation of oil in natural environments is not new.  Many
researchers have examined the fate of oil in various environments including soils, water, and seawater, and
they have concluded that adding nutrients, especially nitrogen and phosphorus, can stimulate
biodegradation of oil. Mineral nutrients (e.g., KNO3, NH4N03, K2HPO4, MgNH4PO4) and organic nutrients,
such as urea, paraffin-supported mineral nutrients, and octyl phosphate, are the most common
compounds used for bioremediation.  Few researchers, however, have studied different nutrient types
concurrently.

    Observations made in our laboratory suggest that the source of nitrogen added to oil-degrading
enrichment cultures may have a powerful  effect on degradative ability. We have routinely used NH 4N03 as
the nitrogen source, but recently we observed that KNO3 permitted greater degradation of oil as measured
by oxygen consumption.  Preliminary results showed that the pH of flasks containing NH 4NO3 was two to
three units lower than flasks supplied with KNO3.  These experiments were undertaken to demonstrate the
effect of nitrogen source on biodegradation of crude oil and to determine the likely cause of the pH shift.


METHODOLOGY
    The effects of different nitrogen sources and pH adjustments on oil degradation were measured in
analytical respirometers. The  respirometers,  Model WB512 from N-CON Systems, Larchmont,  NY, supply
pure O2 to 500 mL reaction flasks in response to pressure drops created by O2 consumption. A computer
monitors the amount of oxygen-delivered and records the data.  The respirometers are equipped with a
temperature controlled water bath and magnetic stirrers. Each flask has a CO 2 trap that contains a
standardized KOH solution, which can be changed as often as necessary through a valve in the cap. The
KOH solution is changed when the pH indicator, Alizarin Red S (10 mg/L) changes color.  The  amount of
CO2 absorbed by the KOH solution can be calculated from its initial normality and  its final pH.

    Experiments to determine the effects of nitrogen source on crude oil biodegradation were conducted
in artificial seawater (2) at 20 °C. In a preliminary experiment, the ability of an oil-degrading enrichment
culture to grow on light Arabian crude oil (LA) in sea salts medium with either NH 4NO3, KNO3 (each at 100
mg N/L), or NH4CI (6.25,12.5, 25, and 50mg N/L) was examined.  The pH of each flask was measured
daily, and half of the flasks were adjusted  as necessary by adding sterile 1 rxl KOH.  The experiment was
run for two weeks.

    The relationships among  the nitrogen source, culture pH, and the kinetics of oil-dependent oxygen
consumption and growth were studied in greater detail in a subsequent experiment. Six treatments were
examined. Oil-degrading enrichment cultures were grown in a sea salts medium that contained either
NH4CI or KNO3 (50 mg N/L).  A nitrification inhibitor (nitrapyrin) was added to half of the NH 4CI cultures to
inhibit acid production by autotrophic n'rtrifiers.  Half of the cultures in each nitrogen source treatment
group (i.e., KNO3, NH4CI, and  NH4CI plus nitrapyrin) were adjusted to a pH  between 6 and 7 by daily
                                               94

-------
addition of an appropriate volume of a sterile NaOH solution. The pH of the remaining cultures was not
adjusted but was monitored. Additional samples were removed periodically for measurement of NH 3,
NO'2/NO'3, and heterotrophic bacterial numbers.

    Ammonia was measured using an ion selective electrode with an Orion 720A pH/mV meter. Nitrite
and nitrate were measured using U.S. EPA method 353.1  (1). Because divalent cations interfere with the
hydrazine reduction step in the nitrate analysis, calcium and magnesium were removed by precipitation at
pH 11 prior to the analysis.

    Bacterial population density was measured by total heterotrophic plate counts on Difco marine agar
2216 (Difco Laboratories, Ann Arbor,  Ml). Plates were incubated six days at 20  °C before counting.


RESULTS

Effects of Nitrogen Source on Culture pH, Oxygen Consumption, and Growth

    The effects of several common nitrogen sources on crude oil biodegradation in a poorly buffered sea
salts medium, as measured by oxygen consumption, are summarized in Table 1. The amount of oxygen
that was consumed at the end of the experiment (O „) is proportional to the mass of oil that was degraded.
Cultures that contained ammonia had a substantially lower final pH than did cultures that were supplied
with nitrate as the sole source of nitrogen, unless they were regularly neutralized by addition of KOH. In
most cases, the pH of ammonia-containing cultures became low enough to inhibit oil biodegradation, but
pH-adjusted, ammonia-containing cultures performed as well as cultures that were supplied with KNO 3.
Thus, the inhibitory effect of ammonia on oil biodegradation appears to be a result oWts effect on the
culture pH.
 TABLE 1. EFFECT OF NITROGEN SOURCE ON OIL BIODEGRADATION (OJ AND CULTURE pH AFTER
                          14 DAYS OF GROWTH IN SEA SALTS MEDIUM
Nitrogen
Source
NH4NO3
KNO3
NH4CI
NH4CI
NH
-------
pH. The cultures that were supplied with KNO3 exhibited only minor fluctuations in pH, and oxygen
consumption proceeded equally well in pH-adjusted and unadjusted cultures. In fact, pH adjustment was
not required in any of the nitrate-containing cultures. The pH of cultures that contained NH 4CI, however,
decreased sharply after two to three days. Oxygen uptake by cultures that contained ammonia was
strongly inhibited if the pH was not regularly adjusted by addition of base. Oxygen consumption in the
pH-adjusted ammonia-containing cultures was similar to that in the nitrate cultures, but the lag time was
shorter.  Nitrification inhibitor had no effect: the NH4CI and NH4CI plus nitrapyrin cultures behaved in an
almost identical manner.

     Bacterial growth,  as measured by total heterotrophic plate counts, was also affected by culture pH.
The growth rates and maximum bacterial density were approximately equal in all of the KNO 3 cultures, but
pH-adjusted, NH4Cl-containing cultures achieved maximum cell densities that were nearly ten-fold higher
than those of the unadjusted  cultures. Furthermore, the heterotrophic bacterial density declined after ten
days in the unadjusted ammonia-containing cultures, but the maximum  population density was stable in
the pH-adjusted, ammonia cultures and the KNO 3 cultures until the experiment was terminated.  The
biomass yields were approximately the same in pH-adjusted ammonia cultures and in the cultures that
contained KNO3. Comparison of the behaviors of the NH4CI and KNO3 cultures implicates ammonia as the
source of the acid that caused the observed pH reductions. Additional support for this conclusion is
provided by the relationship between the observed pH changes in the pH-adjusted ammonia cultures and
the time course of ammonia consumption. Acid production coincided with the  period of active ammonia
uptake, and the culture pH stabilized after ammonia was completely consumed. Thus, acid production in
these oil-degrading cultures was directly related to the presence (and presumably to the metabolism) of
ammonia. Acid production was not related to oxygen consumption (i.e., oil degradation), which continued
for at least two weeks  after the ammonia was exhausted, or to growth, which occurred equally well in
pH-adjusted ammonia cultures and in KNO 3 cultures.

The Role of Nitrification in Ammonia-Associated Acid Production

     The inability of nitrapyrin to affect acid production in ammonia-containing cultures argues against the
hypothesis that nitrification is the cause of the observed pH changes. Nevertheless, nitrate production was
observed in all of the treatments that contained NH 4CI.  In each case, the first appearance of nitrate above
its background level coincides with the first observable reductions in pH. In the pH-adjusted cultures, the
nitrate concentrations rapidly stabilized at approximately twice their initial concentrations. In the cultures
that were not pH adjusted, nitrate accumulated to six to seven times its initial concentration.

     The difference in the extent to which nitrate accumulated in the pH-adjusted and -unadjusted cultures
is probably related to the ability of oil-degrading bacteria in each culture to use nitrate as a nitrogen source.
In the  pH-adjusted cultures, ammonia was rapidly consumed and  the culture conditions favored continued
oil-degradation and growth.  Thus, heterotrophic bacteria in these cultures probably were able to use most
of the  nitrate that was  formed by nitrification. The nitrate that accumulated in the pH-adjusted cultures
reflects a balance between the rate at which nitrate was  produced by nttrifiers and the rate at which it was
consumed by heterotrophs.  The low pHs of the cultures that were not  regularly neutralized, however,
inhibited all  metabolism, including growth and consumption of ammonia. It is unlikely that utilization of
n'rtrate by heterotrophs occurred in these cultures, because the ammonia concentration remained very high
throughout the experiment (ammonia represses expression of the genes involved in assimilatory nitrate
reduction).  Therefore, the nitrate that accumulated in the cultures that were not pH adjusted probably
reflects the total amount that was produced.
     Although nitrification occurred in the cultures that were supplied with ammonia and the timing of
nitrate and acid production coincided, insufficient nitrate was produced to account for the observed acid
                                                 96

-------
 production. The reaction stoichiometry, neglecting growth, that relates production of acid and nitrate by
 nitrifiers is:
2O2 -—
NO
H2O
                                                    2H*
Thus, two moles of acid equivalents are produced per mole of nitrate that is formed by oxidation of
ammonia. Based on this ratio and the amount of nitrate that accumulated, less than 1 % of the acid that
was observed in the pH-adjusted cultures can be attributed to nitrification. More nitrate accumulated and
less acid was produced in the cultures that were not pH adjusted, but less than 5% of the acid observed in
these cultures can be attributed to nitrification. Since the nitrate that accumulated in the cultures that were
not pH-adjusted is probably equal to the amount that was produced, nitrification cannot be the source of
the ammonia-associated acid.

Effect of Nitrogen Source on Lag Time

     In poorly buffered media, acid production that is associated with ammonia metabolism inhibits oil
biodegradation.  Therefore, nitrate is a superior nitrogen source when the buffering capacity of the
oil-contaminated environment is low.  Because nitrate must be reduced to ammonia before it can be
assimilated into biomass, however, ammonia is a preferred nitrogen source for many bacteria.  This
preference is observable in the oil biodegradation kinetics.  Oil biodegradation started more quickly in  the
pH-controlled, ammonia cultures than it did in the nitrate cultures.  Rapid oxygen consumption usually
started after three to four days in cultures that were supplied with ammonia, but five to six days usually
passed before biodegradation began  in the nitrate cultures. Once oil biodegradation began, the rates of
oxygen consumption were similar in nitrate and pH-controlled ammonia cultures.


CONCLUSIONS

     We have identified two major effects of the nitrogen source on crude oil biodegradation in
respirometer flasks. In poorly buffered media, acid  production that is associated with ammonia
metabolism can reduce the culture pH to a level that inhibits oil biodegradation.  When the culture pH is
controlled, either by regular addition of base or by providing sufficient buffering capacity in the medium,
both nitrogen sources support extensive biodegradation of crude oil, but  biodegradation starts more
quickly in the presence of ammonia than in the presence of nitrate. This difference can be important in
some circumstances.  For example, when nutrients can be washed out or diffuse away from the
contaminated area, maximum efficiency of nutrient utilization will be achieved when biodegradation starts
quickly.

     Although the source of the ammonia-associated acid was not identified, it appears to be directly linked
to ammonia metabolism by the oil-degrading enrichment culture that was  used in these experiments. Our
experience suggests that this phenomenon is widespread,  because all of the oil-degrading enrichment
cultures that are maintained in our laboratory exhibit some degree of acid  production in the presence of
ammonia. Ammonia-associated acid production can limit the rate and extent of oil biodegradation in
environments that have limited buffering capacity and limited dilution effects. Bioremediation of poorly
buffered environments that are contaminated with oil or refined petroleum products will probably be  more
effective if a nitrate-based fertilizer is used or if sufficient buffering capacity is supplied in the fertilizer
mixture.
     Our original hypothesis - that nitrification was the cause  of the pH changes that we observed - is not
supported by our data.  A small amount of nitrate was produced in the ammonia-containing cultures, but
this could account for only a small fraction of the acid that was produced.  Due to the extremely long
                                                97

-------
generation times that are characteristic of these organisms, it seems unlikely that autotrophic nitrifiers
could be present in these oil-degrading enrichment cultures. Formation of nitrate by heterotrophic
oxidation of ammonia has been reported, and such a process might be responsible for the traces of nitrate
that were observed in these cultures.

    The results of this research clearly demonstrate that oil biodegradation can be strongly affected by the
source of nitrogen provided to support bacterial growth. Ammonia and nitrate each have advantages and
disadvantages, and these should be considered when choosing a fertilizer mixture to enhance
bioremediation at specific sites. Biodegradation of crude oil begins more quickly when nitrogen is supplied
as ammonia than when it is supplied as nitrate.  Ammonia utilization, however, can be accompanied by
acid production, and under some conditions this acidity can inhibit the rate of oil biodegradation or cause it
to cease entirely. Thus, supplying an appropriate nitrogen source might be as important to successful
bioremediation as supplying enough nitrogen.


REFERENCES
1. USEPA. Methods for Analysis of Water and Wastes. EPA 600/4-79-020 Method 353.1, 1979.

2. Macleod, R.A., and Onofrey, E.  Nutrition and metabolism of marine bacteria. II. Observation on the
        relation of seawater to the growth of marine bacteria.  Journal of Bacteriology 72: 661 -667, 1956.
                                                98

-------
                    FIELD TREATABILITY TRIALS
                               FOR
 FUNGAL TREATMENT OF SOILS CONTAMINATED WITH WOOD TREATING WASTE
               John A.  Glaser1, Richard T. Lamar2,
             Mark W. Davis  , and Diane M. Dietrich
              •'•U.S.  Environmental Protection Agency
              Risk Reduction Engineering Laboratory
                  26 W. Martin Luther King  Dr.
                     Cincinnati, Ohio 45268
                          (513) 569-7568

                  U.S.  Department of Agriculture
                   Forest Products  Laboratory
                      1 Gifford Pinchot Dr.
                    Madison, Wisconsin 53705
INTRODUCTION
     Past investigations of soil treatment systems using lignin-
degrading fungi have largely been confined to laboratory or
bench-scale studies. Recently, a project consisting of two phases
ie. treatability study in the fall of 1991 and a demonstration in
the summer Of 1992 were conducted at an abandoned wood treating
site in Mississippi to evaluate fungal treatment effectiveness
under field conditions. The study site in Brookhaven MS, located
60 miles south of Jackson, was identified as a removal action
site for EPA Region 4.  The fungal treatment studies reported
herein were conducted at Brookhaven because the site
characteristics were suitable for conducting field
investigations, not to promote considerations of fungal treatment
as one of the treatment options for the site.  While the wood
treating facility was in operation, two process liquid lagoons
were excavated and the sludge was mounded above the ground
surface in a RCRA hazardous waste treatment unit. The excavated
material provided the contaminated soil for both phases of the
project. The demonstration phase was undertaken as a SITE Program
Demonstration Project.

METHODOLOGY

     The first-phase study was designed to evaluate the ability
of three different fungal species to degrade pentachlorophenol
(PCP) in soil  (Table 1). The soil pile was sampled and analyzed
for PCP and creosote components prior to developing the test
                                99

-------
site. Analysis of the laboratory results identified sections of
the pile with PCP concentrations of less than 700 mg/kg. These
section were used to supply the contaminated soil for the first-
phase treatability study. The data from the second-phase
demonstration is currently being evaluated. The results of the
demonstration will be reported later.
Table 1. Brookhaven Soil Pile Characteristics
FACTOR
pH

CEC (meq 100/g)

Base Saturation  (%)

Total Nitrogen  (%)

Total Carbon  (%)
VALUE


 3.8

 8.87

54.8

 0.04

 2.17
     The test location was constructed on an uncontaminated
portion of the wood treating site. A base for the test plots was
formed by changing the elevation with clean soil to promote
better drainage conditions. Soil beds, 3 m by 3 m ( 10 ft by 10
ft) . were constructed of galvanized sheet metal. A leachate
collection system was'installed to accumulate the liquid
discharge from all test plots at a central location for testing
and treatment. After installation of the leachate system, 25 cm
(10 in.) of clean sand was layered into each test plot followed
by a 25 cm  (10 in.) lift of contaminated soil.

      The contaminated soil was sized through a 2.5 cm (1 in)mesh
screen using a Read Screen All shaker screen having a 8.4-m3 (10-
yd /hr) capacity. The soil was deposited in separate piles on a
polyethylene tarp. Further homogenization was accomplished by the
mixing of different portions of screened soil. The mixed soil was
then applied to the treatment plots using a front end loader.
Woodchips were added to the soil plots to provide a substrate
that could sustain growth of the fungi.

     Inoculum was developed jointly with the L.F. Lambert Spawn
Co. of Coatesville, PA. The prepared inoculum and inoculum
carrier were shipped to the site by refrigerated transportation.
A total of 10 plots were used in the study.  The experimental
design  (Table 2) consisted of a randomized complete block (RGB)
without replication and a balanced incomplete block (BIB) with
treatments replicated four times. Six of the plots were allocated
                                 100

-------
to the RGB design and four to the BIB design. The BIB plots were
subdivided into 1.5 m by 1.5 m (5 ft by 5 ft) subplots.
Table 2.  Feasibility Study Experimental Design
Treatment'
   Fungus/Contro1
    1

    2

    3

    4
    7

    8

    9


   10
 P. chrvsosporium
    P. sordida

 P. chrysosporium/
    T. hirsuta

Inoculum Carrier
    Control

  No Treatment
    Control
    T. hirsuta      12.5%

 P. chrvsosporium   16.2%
Inoculum
Loading
(Dry Wgt.)
  6.3%

 12.5%

 12.5%

  6.3%
  6.3%

  12.5%
                    12.5%(Day 0)
                     3.7%(Day 14)
 Wood Chip Control
a Wood chips were added to each plot at a 2.5% loading level with
the exception of the No Treatment Control that received no
amendment.

     After inoculation with fungi, each plot was irrigated and
tilled with a garden rototiller. The tiller was cleaned as it was
moved between the plots to prevent cross contamination between
treatments and controls.  Soil moisture was monitored on a daily
basis throughout the study and maintained. Ambient and soil plot
temperatures were recorded throughout the study daily. Plot
tilling was scheduled on a weekly basis for the duration of the
study. A time series analysis of treatment performance was
accomplished by sampling the plots before application of the
treatments, immediately after treatment application, and then
after 1, 2, 4, and 8 weeks of operation.
                                101

-------
RESULTS

     The study was conducted over a 2-month period between
September 18 and November 13, 1991. The greatest removal of PCP
(Table 3) was achieved in the plot inoculated with Phanerochaete
sordida. Over the 42-day period of study, this treatment produced
nearly 90% transformation of PCP from the contaminated soil
initially having a pH of 3.8 and PCP concentrations of 673 mg/kg.

Table 3. Pentachlorophenol Transformationa
Treatment   Fungus/Control
                         PCP
                     Trans formed
           PCP Initial
           Cone.(mg/kg)
    1

    2

    3

    4


    5


    6


    7

    8

    9

   10
 p.  chrvsosporium
    p. sordida

 p. chrvsosporium/
    T. hirsuta

Inoculum Carrier
    Control

  No Treatment
    Control

    T. hirsuta

 p. chrvsosporium

        it

 Wood Chip Control
15%

67%

89%

23%



14%



15%


55%

52%

55%

 0%
 576

1017

 673

 615



 687


 737



 334

 333

 360

 471
aThe extent of transformation  indicated in this table  is
developed from the forty-two day treatability study  conducted  in
September-October 1991.


     Removal data for the  creosote constituents  (PAHs)  are
presented in Table 4 for the treatment using £.. sordida.
Concentration decreases of the three- and four-ring  PAHs were
consistently greater for the fungal treatment than for the
controls. Larger ring PAHs persisted in treatment and  control
plots.
                                102

-------
Table 4. Creosote Constituent Transformation3
Compound
Acenapthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Initial Cone.
(mg/kg) No
429
225
941
684
972
572
Benzo [a] anthracene 74
Chrysene
90
Treatment
Plot
49
75
69
57
23
10
11
6
% Decrease
Carrier
Plot
68
57
49
48
42
22
13
14
P. sordida
Plot
95
95
90
85
72
52
24
33
a Initial concentrations were based on soil samples taken 1 day
after treatment application. Each value for initial concentration
and percentage decrease is the mean of 24 observations.

CONCLUSIONS

     The excellent performance of P. sordida in biotransforming
PCP parallels previous field study experience with P.
chrysbsporium at a Wisconsin site. However, at Brookhaven, P.
sordida achieved the greatest percentage removal. P. sordida is
an organism that resides in the soil under normal conditions. The
soil conditions encountered at the Brookhaven site are not
optimal for the cultivation of microorganisms, and the
concentrations of PCP found in the Brookhaven soil are beyond the
limits to sustain suitable growth for many microorganisms. All of
these factors strongly suggest that fungal treatment using £._
sordida has exceptional promise for processing pentachlorophenol
and other difficult to degrade pollutants. Fungal treatment
technology should be ready for commercial application in the next
12-18 months. Future challenges to making the technology more
cost effective are developing improved and cheaper inoculation
techniques and extending application of the technology to other
organic pollutant classes.

FOR MORE INFORMATION:

                          John A.  Glaser
               U.S.  Environmental  Protection Agency
              Risk Reduction Engineering Laboratory
                   26 W.  Martin Luther King Dr.
                      Cincinnati,  Ohio 45268
                          (513)  569-7568
                                 103

-------
         AEROBIC, PHENOL-INDUCED BIOTRANSFORMATION OF TCE

                              Mathew M. Shurtliff
                                  CH2MHill
                               P.O. Box 147009
                           Gainesville, FL 32614-7009
                                (904)331-2442

                                Gene F. Parkin
                   DepL of Civil and Environmental Engineering
                             The University of Iowa
                              Iowa City, IA 52242
                                (319) 335-5655

                               David T. Gibson
                              DepL of Microbiology
                             The University of Iowa
                              Iowa City, IA 52242
                                (319) 335-7980

NOTE:  An extended version of this paper is being submitted for publication to Applied
        and Environmental Microbiology

INTRODUCTION

    Trichloroethylene (TCE) is a volatile chlorinated aliphatic compound that has been
commonly used as a solvent TCE is also a commonly found groundwater contaminant
that has been shown to induce hepatocellular carcinomas in mice and is a suspected human
carcinogen. In response to the potential human health effects, TCE has been identified as a
priority pollutant by the USEPA, and a maximum contaminant level (MCL) of 5 [ig/L for
public drinking water supplies has been promulgated. In 1986, an estimated 1.8 million
persons in the United States were exposed to drinking water containing concentrations of
TCE greater than the MCL; an estimated 62% of this exposure came from groundwater
sources (1).
    In anaerobic (reducing) environments, TCE has been shown to be biotransformed via
reductive dechlorination to vinyl chloride, which is a well-documented, potent carcinogen.
A review of the literature indicates that organisms possessing ammonia monooxygenase,
propane monooxygenase, methane monooxygenase, two types of toluene monooxygenase,
and toluene dioxygenase have been shown to cometabolically degrade TCE (2). In
addition, methanotrophic consortia and a toluene-oxidizing pure culture have been shown
degrade radiolabeled TCE to nonvolatile aqueous products and carbon dioxide. Previous
research in our laboratory showed that a phenol-enriched, mixed culture had significant
potential to biodegrade TCE (3).
    The research described herein further investigated aerobic, phenol-induced TCE
degradation by a phenol-enriched, mixed culture. The objectives of this research were to
describe the kinetics of phenol and TCE utilization using the Monod relationship, to
perform mass-balances on TCE fed to a reactor containing these organisms maintained in
continuous culture, and to study the products of TCE degradation by these organisms using
radiolabeled TCE.

METHODOLOGY

    The TCE-degrading bacteria were acclimated and maintained at 20°C in a continuous
culture reactor (chemostat). Filter flasks with a side arm at the base of the flask were used
                                      104

-------
as reactors. Flasks containing freshly-prepared feed solution were autoclaved prior to use.
Peristaltic pumps transported media into the reactor and culture out of the reactor. A
syringe pump added TCE to the reactor below the fluid level to minimize volatilization of
TCE. In mass balance experiments, an abiotic control reactor containing sterilized
deionized water was operated under identical flow and temperature conditions.
    The hydraulic retention time of each reactor was maintained at five days. Each reactor
was aerated by a continuous flow of 95% pure oxygen through the headspace to minimize
the volatilization of TCE that would occur in a bubble-diffusion aeration system. The
primary feed solution for the mixed culture consisted of phenol as the sole carbon and
energy source (420 mg/L), ammonia as the nitrogen source, trace metals, and a phosphate
buffer to maintain a neutral pH (2).  TCE feed solutions were prepared in 100-mL
volumetric flasks by dissolving 60 |iL of pure TCE in deionized water. TCE solutions
were dispensed into 9-mL glass vials and sealed with no headspace using teflon-lined
screw caps. The individual vials were stored at 4°C until they were drawn into the feed
syringes for use.
    Aqueous TCE concentrations were determined by two methods:  head-space analysis
and extraction with iso-octane. A Hewlett-Packard Series 5890A gas chromatograph
equipped with a VOCOL capillary column and 63Ni electron capture detection. TCE in the
effluent gas from the chemostat was collected using activated-carbon-containing, ORBO-32
tubes (Supelco).  TCE was desorbed using carbon disulfide, diluted with iso-octane, and
analyzed by GC.  14C-labeled CO2 from  14C-labeled TCE transformation was collected in
high-pH traps and measured using a Beckman LS 6000IC Scintillation Counter. Biomass
concentrations were estimated by measuring volatile suspended solids (VSS). Phenol was
measured using a colorimetric assay involving the condensation of 4-aminoantipyrine with
phenol under alkaline conditions in the presence of an oxidizing agent (potassium
ferricyanate).  Details are given elsewhere (2).
    The kinetics  of phenol removal were described by the Monod relationship, which
consists of the following expression:

                            dP   kPX
                           " dt ~ KS+P

where k is the maximum phenol removal rate constant (mg phenol/mg biomass-day), P is
the phenol concentration (mg phenol/L),  X is the concentration of biomass (mg
biomass/L), Ks is the Monod half-velocity constant (mg phenol/L), and t is time.The
kinetics of bacterial growth are described by the expression:
                              dX/dt   YkP
                                      KS+P
                                            -b
where (I is the net bacterial growth rate (day1), b is the bacterial decay constant (day1), k,
P, and Ks are the Monod parameters described above, and Y is the yield coefficient, which
describes the mass of biomass produced per mass of phenol utilized (mg biomass/mg
phenol). Batch experiments performed to determine k, Ks  Y, and b for phenol utilization
were initiated by adding phenol feed solution to 500 mL of^ bacterial culture to obtain an
initial phenol concentration of approximately 400 mg/L.  Phenol and VSS measurements
were taken, at various times and this data used to determine values for k, Ks, Y, and b.
    A second-order expression was used to determine the kinetics of TCE (S) removal:
                           — - k'SX
                           dt  ~ki>X
which is first order with respect to TCE concentration, and second order overall.
Experiments were performed in closed batch systems with resting cells and TCE
                                     105

-------
concentration was monitored over time. VSS was measured before and after each
experiment to ensure that it did not change significantly.

RESULTS

    Several types of experiments were conducted to assess the kinetics and transformation
products of phenol-induced TCE biotransformation by the mixed culture: abiotic studies to
assess TCE sorption to biomass, mass-balance experiments using the chemostat, phenol
removal kinetics experiments, TCE removal experiments, and TCE mineralization
experiments using 14C-labeled TCE.
    Abiotic Sorption Experiments.  These experiments were conducted in batch systems
(62-mL serum vials sealed with aluminum crimp-top caps) containing organisms taken
from the chemostat and killed either by autoclaving or by adding sodium azide to a final
concentration of 1%. TCE was measured before and after 48 hours of mixing with an
orbital shaker. Results from three autoclaved and four azide-sterilized  systems gave TCE
recoveries of greater than 98% indicating that sorption to biomass was  not a major TCE
removal mechanism.
    Mass-Balance Experiments.  Mass-balance experiments were conducted using an
abiotic control reactor fed only feed solution containing phenol and TCE but no bacteria and
the chemostat described above. Conditions were identical except the chemostat had phenol-
utilizing bacteria present. The abiotic reactor was used to estimate abiotic losses from the
system other than sorption to biomass. The difference between the total mass of TCE
entering the reactors and the total mass of TCE leaving the reactor was  labeled as "percent
not accounted for."  The "percent not accounted for"  for each mixed-culture, mass-balance
experiment was compared to the average "percent not accounted for" determined from
several control reactor mass balance experiments performed at comparable influent TCE
concentrations. The amount of TCE biodegraded was estimated by subtracting the control
reactor "percent not accounted for" from the mixed culture reactor "percent not accounted
for."  Results from six mass-balance experiments are summarized in Table 1.

           Table 1. Summary of mixed-culture, mass-balance experiments1.
Experimental
Set
1
2
3
4
5
6
Influent
TCE,mg/L
23.7 ± 8.0
26.7 ± 0.9
31.1 ±2.1
61.3 ±4.1
64.5 ± 5.1
66.0 ± 7.0
Influent
Phenol, mg/L
420
210
100
420
210
100
Biomass Range,
mg/L
144 - 160
60-80
30-50
144-160
60-80
30-50
Estimated
% Biodegraded
92.0 ± 3.3
92.7 ± 4.3
68.5 ± 5.1
57.8 ± 2.9
70.3 ± 3.0
33.7 ±11.5
^Experimental sets consisted of five individual mass-balance experiments. Conditions in
the mixed-culture reactor: DO > 10 mg/L, pH = 7.0, temp = 20°C.


    Phenol Kinetics Experiments. Four batch kinetic experiments were performed to
determine the Monod kinetic parameters Y, k, and Ks. Three additional batch experiments
were performed to determine Y alone. Two batch experiments were performed to
determine b. Values determined were: Y = 0.61 mg VSS/mg phenol, k = 9.3 mg
                                      106

-------
phenol/mg VSS-day, and b = 0.12 day1. A value of 3.3 mg phenol/L was estimated for
Ks.
    TCE-Removal Kinetics.  TCE removal kinetic experiments were performed at four
initial aqueous TCE concentrations.  At least six separate experiments were performed for
each initial aqueous TCE concentration. Rate constants were determined from standard
first-order linearization plots of TCE removal for each experiment; linear regressions from
the first-order plots were used to confirm that under the conditions tested the observed
removal of TCE was indeed first-order with respect to TCE. Results of the experiments are
summarized in Table 2.

                     Table 2 Results TCE removal experiments
Initial aqueous
TCE,mg/L
1.5
5.3
14.6
27.5
k'
L/mg VSS-day
0.073 ± 0.035
0.028 ± 0.0052
0.0055 ± 0.00068
0.0032 ± 0.00079
n = # of tests
10
8
8
6
    TCE Mineralization Experiments. A total of ten mineralization experiments were
performed; results and averages are shown in Table 3. Results are reported as the percent
of total radiolabeled TCE added to the reactor found in fractions associated with the carbon
dioxide traps, biomass, and filtrate from the reactor.

                   Table 3.  Results of mineralization experiments
Experimental
1
2
3
4
5
6
7
8
9
10
Average
Std. Dev.
% recovered
asCO2
26.6
18.6
19.3
20.7
21.1
20.2
30.8 ,
21.5
22.8
20.4
22.2
3.7
% recovered
with biomass
8.9
8.7
8.8
8.8
8.5
8.5
8.7
9.3
8.6
9.1
8.8
0.3
% recovered
infiltrate
40.0
46.9
47.2
48.0
46.4
45.0
32.8
37.5
36.6
40.9
42.1
5.3
% recovered
total
75.5
74.2
75.3
77.5
75.9
73.7
72.3
68.3
68.1
70.4
73.1
3.3
                                     107

-------
SUMMARY AND CONCLUSIONS

    This research investigated aerobic, phenol-induced TCE degradation by a phenol-
enriched, mixed culture.  Mass-balance experiments were performed to determine the
effects, of varying influent phenol and TCE concentrations on TCE biodegradation. Abiotic
experiments were performed to determine the extent of sorption of TCE to biomass.
Phenol and TCE utilization kinetics were studied to determine reaction rates, and
mineralization experiments with 14C-labeled TCE were performed to study TCE
transformation products. The following conclusions are drawn from this research:
   1.  Sorption of TCE to biomass was negligible and was not a significant TCE removal
      mechanism in the system studied.
   2. The Monod kinetic parameters of phenol utilization were determined to be:
      Y = 0.61 ± 0.06 mg VSS/mg phenol, k = 9.3 ± 5.6 mg phenol/mg VSS-day, Ks <
      3.3  mg phenol/L, and b = 0.12 ± 0.005 day 1-
   3. High levels of TCE biodegradation were accomplished when the influent
      phenol/influent TCE ratio was sufficiently high. As the phenol/TCE ratio decreases
      from a threshold value, TCE biodegradation levels decrease in what appears to be a
      linear fashion. Influent phenol/TCE ratios can be a significant design parameter used
      in designing aerobic, phenol-induced, pump-and-treat bioreactors for TCE removal.
   4. Resting-cell TCE removal rates decreased with increasing initial aqueous TCE
      concentration.
   5. Under the conditions studied, an estimated 22.2% of the degraded mass of TCE was
      mineralized to carbon dioxide, 8.8% was incorporated into biomass, and 42.1% was
      transformed to nonvolatile products.  It is most likely that the remaining 26.9% was
      transformed to volatile products other than carbon dioxide.

REFERENCES

1.  Cothem, C. R., W.A. Coniglio, and W.L. Marcus. Estimating risk to human health.
    Environ. Sci. Technol.  20:111-116,  1986.

2.  Shurtliff, Mathew M., G.F. Parkin, and D.T. Gibson. Kinetics and Transformation
    Products of Aerobic, Phenol-Induced Trichloroethylene Degradation. Submitted to
    Appl. Environ. Microbiol., 1993.

3.  Coyle, C.G., G.F. Parkin, and D.T. Gibson. Aerobic, Phenol-Induced TCE
    Degradation in Completely Mixed, Continuous Culture Reactors.  Biodegradation (in
    press).

ACKNOWLEDGEMENTS

    This work was  primarily funded by the Region 7 & 8 Hazardous Substance Research
Center of the U.S. Environmental Protection Agency (EPA) at Kansas State University. .
Additional support was provided by the Biotechnology Byproducts Consortium of the
University of Iowa, which is supported by a grant from the U.S. Dept. of Agriculture.

FOR MORE INFORMATION

    Gene  F. Parkin
    Dept.  of Civil and Environmental Engineering
    The University of Iowa
    Iowa City, IA  52242
    (319)  335-5655
                                   108

-------
    HAZARDOUS WASTE INCINERATOR EMISSIONS RESULTING FROM WASTE FEED CUTOFFS

                                      Marta K. Richards
                             U.S. Environmental Protection Agency
                             Risk Reduction Engineering Laboratory
                                 26 W. Martin Luther King Drive
                                    Cincinnati, Ohio 45268
                                       (513) 569-7783

                                            and

                         W. Eddie Whitworth, Jr. and Larry R. Waterland
                               Acurex Environmental Corporation
                                 Incineration Research Facility
                                  Jefferson, Arkansas 72079
                                       (501) 541-0004

INTRODUCTION

       In July 1990, the Environmental Protection Agency (EPA) and the Occupational Safety and
Health Administration (OSHA) established a joint Task Force to review safety and health issues at
hazardous waste incineration facilities nationwide. The Task Force inspected 29 facilities. In addition to
noting a number of OSHA and EPA regulation and hazardous waste management permit violations,
EPA noted a significant number of waste feed cutoff (WFCO) episodes at about half the incinerators
inspected. At these facilities, the incinerator was being routinely operated at conditions sufficiently close
to permit limits that small upsets caused routine WFCOs. This has caused some concern because the
premise has been that operation outside of  permit conditions has unacceptable environmental impacts.
Thus, if incinerator operation routinely drifts  outside  of permit conditions, the question arises as to what
adverse emissions impacts, if any, occur.

       The purpose of this test program, performed at EPA's Incineration Research Facility (IRF), was
to address this question.  Specifically, the test program was designed to evaluate whether increases in
hazardous constituent trace metals, hazardous constituent organics, and HCI emissions occur with
repeated WFCO episodes. Tests were conducted to evaluate the incremental emission impact of
triggering WFCOs by two modes of air pollution control system (APCS) failure (exceeding a permit
condition) and by exceeding the incinerator CO emission limit. The APCS failure test objective was to
determine whether incremental emissions of trace metals and HCI occur with WFCOs triggered by the
APCS failure.  The elevated-CO-emissionstest objective was to determine whether incremental principal
organic hazardous constituent (POHC) emissions occur with WFCOs triggered by operation with high
flue gas CO levels.

METHODOLOGY

       All tests were performed in the rotary kiln incineration system (RKS) at the IRF. This system
consists of a primary combustion chamber, a transition section,  and a fired afterburner chamber. After
exiting the afterburner, flue gas flows through a quench section followed by a primary APCS. The
primary APCS for these tests consisted of a venturi scrubber followed by a packed-column scrubber.
Downstream of the primary APCS, a backup secondary APCS, comprised of a demister, an activated-
carbon adsorber, and a high-efficiency paniculate air (HEPA) filter, is in place.
                                             109

-------
       The test waste fired throughout the test program was composed of a mixture of organic liquids
comprising the POHCs and a hazardous constituent trace metals solution added to a clay absorbent
material.  The organic liquid used was a mixture of 76% toluene by weight and 12% each of
chtorobenzene and tetrachlorethene. The synthetic waste was packaged into plastic-bag-lined
fiberpack drums and fed to the rotary kiln via the fiberpack drum ram feed system.  The same mixture
of organic liquids was also co-fired through the kiln burner via a liquid nozzle in the  burner.  Five
hazardous constituent trace metals were added to the solid synthetic waste by incorporating a
concentrated aqueous solution of metals into the synthetic waste mixture. When added to the
clay/organic liquid mixture, the resulting synthetic waste contained arsenic at 31 mg/kg, barium at
310 mg/kg, cadmium at 16 mg/kg, and chromium and lead at 39 mg/kg.

       The WFCOs planned for testing were venturi scrubber pressure drop upsets (shutting off the
scrubber system induced draft fan), scrubber liquor flow upsets (shutting off the scrubber liquor
recirculation pump), and excessive CO emissions upsets. One test was performed for each of the
scrubber upsets (Tests 2 and 3). Two test conditions resulting in excessive CO emissions were tested.
The first test condition (Test 4a) used a reduced combustion air flow to produce excessive CO. The
second (Test 4b) used an elevated charge size to produce excessive CO.  Duplicate baseline condition
tests (Tests 1a and 1b), with no WFCOs, were also completed.  A tow feedrate baseline test (Test 5)
was added to the test program to provide baseline test results at the decreased overall feedrates
corresponding to Tests 2 and 3 which had routinely interrupted feed of extended duration with each
WFCO episode.

       Routine scrubber exit CO spikes were indeed experienced during the high CO tests, as desired.
Twenty-two spikes were experienced during the volatile organic emissions sampling period during
Test 4a, seven of which drove the scrubber exit CO monitor to a full-scale reading of 650 ppm. Less
frequent CO spikes were experienced for Test 4b.  Eleven scrubber exit spikes occurred over the
volatile organic emissions sampling period during this test. Only two of these drove the scrubber exit
CO monitor to full scale. The other spikes were about 200 ppm or less.

       In addition to obtaining synthetic solid and liquid waste feed, kiln ash, and pre- and  post-test
scrubber liquor samples, the sampling protocol for all tests included sampling the flue gas at the
afterburner exit and at the scrubber system exit for trace metals using the EPA multiple metals train,
and particulate and HCI using Method 5. In addition, the scrubber exit flue gas was sampled for the
volatile POHCs using Method 0030.  The stack downstream of the secondary APCS was also sampled
for particulate and HCI, using Method 5.

       The Method 0030 samples were  analyzed for the volatile POHCs and the multiple metals train
samples was analyzed for the five test trace metals. In addition, the synthetic solid waste feed, kiln ash,
and pre-  and post-test scrubber liquor samples were analyzed for the test POHCs and trace metals.

RESULTS

       Test data showed that, over all seven tests, scrubber exit flue gas toluene concentrations
ranged from 12 to 88 jig/dscm, tetrachtoroethene concentrations ranged from 1.5 to 14 jug/dscm, and
chtorobenzene concentrations ranged from 1.5 to 9.8 /ng/dscm.  The highest concentrations for all three
POHCs were measured in the repeat baseline test, Test 1 b.  The lowest toluene concentration was
measured in the high CO from decreased air flow test, Test 4a.  The lowest tetrachtoroethene and
chtorobenzene concentrations were measured in Test 3, the scrubber liquor flow failure test. However,
flue gas concentrations for these two POHCs were comparably tow in Test 4a. Flue gas concentrations
for all three POHCs in the other high CO test, Test 4b, were comparable to those measured in the
baseline test, Test 1a
                                             110

-------
       Corresponding DREs as measured at the scrubber exit ranged from 99.9982% to 99.99977%
for toluene, from 99.9986% to 99.99981% for tetrachtoroethene, and 99.9989% to 99.99981% for
chlorobenzene. Again, the lowest DREs for all three POHCs occurred in the repeat baseline test,
Test 1 b.  The highest DREs for all three POHCs occurred in the high CO, reduced air flow test, Test 4a.
POHC DREs for the high CO, increased charge mass test, Test 4b, were comparable to, though
uniformly greater than those measured in Test 1 a, the baseline test. The test data clearly indicated that
repeatedly exceeding an instantaneous 100 ppm CO limit at the exit of the APCS via the two
mechanisms tested did not cause increased POHC flue gas concentrations or emission rates, nor
decreased POHC ORE.

       Test program trace metal data also showed no significant test-to-test variations in scrubber exit
flue gas metal concentrations for any of the test metals.  Thus, within the range of the duplicate baseline
test metals concentration variability, it appears that none of the WFCO operating modes tested
significantly affected scrubber exit flue gas metal concentrations.  Metals emissions rates were similarly
apparently not significantly different for the WFCO tests compared to the baseline tests.

       Metals partitioning data showed that barium and chromium were relatively nonvolatile in all of
the tests.  Greater than 98% of the barium and 96% of the chromium was accounted for by the kiln ash
discharge. Less than 1 % of the barium and chromium was measured in the scrubber exit flue gas, with
about 1% of the barium and 1% to 4% of the chromium measured in the scrubber liquor. Arsenic and
lead were more volatile in the tests, though still predominantly nonvolatile Nominally 80% to 90% of the
arsenic was measured in the kiln ash discharge for all tests except the tow feedrate baseline test,
Test 5. Between about 5% and 10% of the arsenic was measured in the scrubber exit flue gas for all
except Test 5. Slightly larger fractions of arsenic were measured in the scrubber liquor. Nominally 80%
to 95% of the lead was measured in the kiln ash discharge for the three tests having reliable solid
sample lead concentration data About 5% to 10% of the tead was measured in the scrubber exit flue
gas, and 1% to 12% in the scrubber liquor.  Cadmium exhibited quite volatile behavior in all tests. Only
20% to 40% of the cadmium was accounted for in the  kiln ash discharge.  Most of the cadmium, 50% to
70%, was accounted  for in the scrubber exit flue gas.  Only 10% to 27% of the cadmium was
accounted for in the scrubber liquor.

       The above distributions reflect overall test program results. The test data show no repeatedly
significant difference in metals distributions from test to test for any of the metals, within the degree of
data variability exhibited in the two baseline tests and/or the precision of the measurements. This
suggests that metals partitioning among the incinerator discharges was relatively unaffected by the
different operating conditions leading to repeated WFCOs that were tested.

       Apparent scrubber collection efficiency data showed  that the venturi/packed column scrubber
system was nominally 70% to 90% efficient in collecting barium and chromium. Arsenic apparent
collection efficiencies were perhaps as tow as 45% to 64% Cadmium apparent collection efficiencies
were tower, in the 13% to 35% range.  Lead apparent collection efficiencies were highly variable. Within
the range of variability in the data, however, no test-to-test differences in collection efficiencies were
apparent. This suggests that none of the repeated WFCOs tested affected scrubber metals collection
efficiencies.

       The paniculate emission data from the tests showed  that afterburner exit paniculate levels
ranged from 48 to 98 mg/dscm, corrected to 7% O2I for the two baseline, the two scrubber failure, and
the tow feedrate baseline tests, Tests  1a, 1b, 2, 3, and 5. Afterburner exit paniculate levels were
apparently increased for the two high CO WFCO tests.  This  is likely due to a combination of increased
kiln ash entrainment into the combustion gas,  as well as some flue gas soot formed during the high CO
tests. Significant soot formation was evident during both high CO tests in visual observations of the kiln
combustion gas and the substantial darkening of the scrubber liquor with collected soot.
                                             Ill

-------
        Scrubber exit paniculate levels were reduced to the 8- to 15-mg/dscm range at 7% O2 for
Tests 1 a, 1 b, 2, 3, and 5. Scrubber exit paniculate levels were essentially the same for the two
scrubber failure tests as for the two baseline and the low feedrate baseline tests. Thus, the two
scrubber failure modes tested with WFCO resulted in no apparent increased paniculate emissions.
Scrubber exit paniculate emissions for the two high CO WFCO tests were slightly higher, at 17 and
26 mg/dscm at 7% O2.  These are most likely the result of the higher scrubber inlet loadings for the two
tests.  The scrubber paniculate removal efficiencies measured were not significantly different from test
to test, ranging from 77% to 89%.

        Afterburner exit HCI levels ranged from 325 to 1,130 ppm with the test to test variations in
chlorine feedrate as determined by waste feedrate and chlorine content.  Scrubber exit HCI levels were
reduced to 0.5 to 1.5 ppm. Relatively constant scrubber HCI collection efficiencies, at 99.8% to 99.9%,
were measured.  The two scrubber failure modes tested, with attendant WFCO,  apparently did not
result in increased HCI emissions or decreased scrubber HCI collection efficiency.

CONCLUSIONS

       Test program results show that none of the incinerator operating modes tested, which resulted
in repeated WFCOs, caused increased hazardous constituent or HCI emissions. The DREs for the
three test POHCs, toluene, tetrachloroethene, and  chtorobenzene, ranged from 99.9982% to 99.99981%
over the test program.  The lowest DREs were measured in one of the two baseline tests. The highest
DREs were measured in the repeated CO spike test in which CO spikes were produced by reducing
the combustion air flowrate to the incinerator from  the baseline test levels. The other CO spike test, in
which CO spikes were produced by increasing the solid waste feed batch charge mass and heat
content to an overcharge situation, had POHC DREs comparable to the other baseline test.  POHC
DREs for the two scrubber system failure tests, one in which venturi scrubber pressure drop was
repeatedly reduced by shutting off the scrubber induced draft fan (fan failure) and the other in which
scrubber liquor flow was stopped by shutting off the scrubber system recirculatton pump, were also
comparable to those measured in the baseline tests.

       Within the variability in the test-to-test trace metal data, none of the repeated WFCO operating
modes tested resulted in increased flue gas metals emissions. Scrubber exit flue gas concentrations
and emissions rates of arsenic, barium, cadmium, chromium, and lead for the two scrubber failure and
two high CO WFCO tests were not significantly different than for the baseline tests. Trace metal
distributions among the three incinerator discharge streams, kiln ash, scrubber liquor, and scrubber exit
flue gas, and scrubber trace metal collection efficiencies were not significantly different from baseline to
WFCO tests, again within the test-to-test data variability and the precision of the  metals analysis
methods.

       Scrubber exit flue gas HCI concentrations and emission rates varied only with the feedrate of
chlorine in the wastes fed to the incinerator. Scrubber HCI collection efficiencies were 99.8% to 99.9%
for all tests and were not reduced with any scrubber failure or high CO operating mode tested.

       Scrubber exit flue gas paniculate levels were in the 8- to 15-mg/dscm at 7% O2 range for all
tests except the two high-CO WFCO tests. Scrubber exit flue gas paniculate levels for the two
scrubber-failure WFCO tests were tower than levels measured in the two baseline tests. Scrubber exit
paniculate levels were higher, at 17 and 26 mg/dscm, for the two high-CO WFCO tests. However,
these were the result of increased scrubber inlet paniculate levels for these two tests.  Scrubber
paniculate collection efficiency was relatively constant at 77% to 89% from test to test.
                                             112

-------
       The higher inlet paniculate levels for the high CO WFCO tests were likely due to a combination
of increased entrainment of solids from the kiln into the kiln exit combustion gas, and soot formed
during the incomplete combustion environment resulting in the CO spikes. The increased entrainment
results from the high intensity puff of incompletely combusted organic associated with kiln overcharging
used to produce the high CO.

       Overall, test results suggest that the permit requirement to terminate waste feed whenever a
permit-specified operating limit is exceeded is an appropriate protection against increased incinerator
emissions of POHCs, trace metals, and HCI.  Only paniculate emissions increases, of perhaps double
the baseline, routine operation levels, were measured in these tests.  Further, these increases were not
associated with ARCS failures, but with increased ARCS inlet paniculate levels arising from combustion
conditions associated with repeated CO spikes.

        Barium and chromium exhibited nonvolatile behavior in all the tests performed. Greater than
98% of the barium and 96% of the chromium measured in incinerator discharges was accounted for in
the kiln ash. Arsenic and lead exhibited more volatile behavior. The kiln ash discharge accounted for
80% to 90% of the arsenic measured in the discharges for all tests except the tow feedrate baseline
test, Test 5. Nominally 80% to 95% of the lead measured in the incinerator discharges was in the kiln
ash for the three tests for which reliable solid sample lead concentrations were measured. Cadmium
was more volatile in all tests; only 14% to 43% of the cadmium in the discharges was accounted for in
the kiln ash. These observations are consistent with past IRF metal partitioning experience.

        For more information contact Marta K. Richards, the EPA Technical Project Manager, at:

               Risk Reduction Engineering Laboratory
               U.S. Environmental Protection Agency
               Cincinnati, OH  46268
               (513) 569-7783
                                              113

-------
        METALS EMISSIONS/CONTROL FROM THE BURNING OF HAZARDOUS WASTE

              R. George Rizeq, Wyman Clark, David W. Hansell, and W. Randall Seeker
                         Energy and Environmental Research Corporation
                                  18 Mason, Irvine,  CA 92718
                              Tel. (714)859-8851, Fax. (714)859-3194


INTRODUCTION

       The behavior of metals in combustion systems is an area of interest to the EPA and to the
industry.  Toxic metals emissions constitute a threat to human health and the environment.  Much is known
about the behavior of metals in combustion systems. The study of metals behavior has been an ongoing
project supported by the EPA contract: Engineering Analysis of Hazardous Waste Thermal Destruction
(68-CO-0094 and its predecessors) for a number of years.  Continuing support is also being provided
through EPA Contract 68-W3-0001, Technical Support for RCRA Hazardous Waste Control Program.  In
1988, Energy and Environmental Research Corporation (EER) summarized the knowledge in the field in
the document: Prediction of the Fate of Toxic Metals in Hazardous Waste Incinerators (1).

       In FY91 EER began work on a project to expand and update this document, creating a "Metals
Bible". When completed, this comprehensive document will combine the analysis of test data with the
understanding of the fundamental processes to address  all aspects of metals behavior in combustion
systems. It will address various wastes including hazardous, mixed, medical, municipal, industrial, military,
and other wastes; and it will address various combustion  systems including incinerators, boilers, industrial
furnaces, and other combustion systems. For these wastes and technologies the "Metals Bible" will
cover a range of processes influencing metals behavior including partitioning of metals to the various exit
streams of the process (e.g., bottom ash, air pollution control equipment ash, emissions), transport of
the emitted metals through the atmosphere, and the resulting health effects. This information will be
related to problems of practical importance including improved methodologies for control, testing, and
monitoring of metals. The FY91 efforts were summarized in the FY91 Final Report (2).

        In FY92 EER continued this effort, specifically addressing those aspects of the "Metals Bible"
related to the behavior of metals in hazardous waste combustion systems. This extended abstract is a
summary of the FY92 efforts which are described in more detail in the FY92 Final Report (3).
                         •.
       Objectives. The objectives of the FY92 efforts were to: 1) update the thermodynamics  database
w'rth the latest literature data on the thermochemicai properties of metals, and revise metals volatility
tendencies to better predict metals emissions; 2) compile available data on the behavior of metals in
hazardous waste combustion systems nationwide, and review and enter the data into the metals
emissions database; and 3) assess the implications of the new data on the current understanding of
metals behavior and compile information related to metals behavior including testing methodologies and
current regulations and guidance.

METHODOLOGY

       This program was conducted in two parallel major tasks which were performed simultaneously.
These two tasks included the development and application of:  1) a rnetals emissions database and 2) an
improved thermodynamic model of metals behavior. The metals emissions database task encompassed
the following:

   •   compilation of available data on metals behavior  from hazardous waste incinerators (HWIs)
       nationwide;
   •   assessment of the quality and completeness of  each data set;
   •   modification of EER's air toxics emissions database to allow entry of data from HWIs;
   •   assembly of the data into the database for analysis; and
   •   use of metals database to analyze data from trial burn reports.
                                           114

-------
For more effective analysis of the field data, there was a need to update and upgrade the
thermodynamics database used in EER's metals partitioning model. In this task, EER:

       conducted a literature search to obtain the latest thermochemical properties of metals;
       updated the thermodynamics database to include data for all toxic metals of interest;
       expanded the database to include additional chemical species for many metals;
       reassessed metal volatility tendencies and studied the effect of major operating parameters on
       metals behavior;
       expanded the study of chromium and arsenic volatility/speciation; and
       evaluated the effect of earth metals present in the waste on toxic metals volatility.

       To assess the implications of the analysis performed under the two major tasks, a study was
conducted to compare the predicted versus measured uncontrolled metals emissions from a full scale
hazardous waste incinerator facility.  The primary results from this program are presented in the following
section.

RESULTS

       Thermodynamics.  In the combustion process, metals in the waste or fuel are subjected to high
temperatures. Depending on the thermodynamic properties of the metals and on the local conditions, a
portion of the metals may react and/or vaporize.  Metals which vaporize are swept away with the
combustion gas and, as the gas cools, tend to condense into or onto the surfaces of very fine particles
which are in the size range that is least effectively removed in most air pollution control devices. The
ability to predict and analyze the metals vaporization process is largely dependent on the accuracy and
the completeness of the data on the thermodynamic properties  of metal species.  In FY92, EER
upgraded its thermodynamics database, adding data for a number of new species and increasing the
number of compounds for the toxic metals as illustrated in Figure 1.
  o 8Q-(

 | 70^
  ca
 •J40
  0)
 fao
 110
  ZJ
y""7
           Sb    As   Ba    Be    Cd   Cr    Hg     Ni    Pb    Se    Ag    TI
                        Old Thermo
                                     New Thermo
                        Figure 1  .  Updated thermodynamics database.
                                           115

-------
        Several significant changes were noted in comparing predictions of the behavior of metals based
 on the updated thermodynamics with those based on the previously used thermodynamics database:

                                          W
                                            1E-8-
    1E-9-
                                     1E-10.
                                           o
                                           o
                                           CM
                                                                        Cr(total)
                                                                           Cr(+6)
             o
             co
o
o
co
o
o
co
o
o
o
o
o
CM
o
o
o
o
CO
o
o
oo
                                                       Temperature (K)
o
o
a
                                              Figure 3
                       Volatility of chromium in a
                       chlorinated environment.
                                             116

-------
        1E-7
1E-6
    1E-5      1E-4      1E-3      1E-2
Measured Uncontrolled Emissions (gr/sdcf)
1E-1
1E+0
             Figure 4 . Comparison between predicted and measured uncontrolled
                       emissions of metals from the APTUS/EPA test (4); Cond.-1.
controlled and uncontrolled metals and particulate emissions results. A system of data quality indicators
has been built into the database so that the accuracy of the emissions results can be assessed.  In
addition, the database is designed to support a wide range of activities including:

       emission factor development;
       mass balance (enrichment and partitioning analyses);
       complex statistical analysis;
       parametric studies;
       air pollution control equipment and incinerator emission control assessment;
       data quality assessments;
       at the stack risk evaluations; and
       regulatory development.

       Based on a survey of EPA regions, it is estimated that approximately 60 incinerators and 50
boilers and industrial furnaces have conducted source tests and collected metals emissions data while
burning hazardous waste. By the end of FY92, data from 11 commercial hazardous waste incinerators
(comprising over half of the commercial incinerator population) have been collected and entered into
EER's metals emissions database. The availability of such data in a relational database format allows
comparisons of emissions from different facilities and the development of correlations to identify those
parameters which influence metals emissions.

CONCLUSIONS

       The main conclusions drawn from the work  performed in FY92 and summarized in this extended
abstract are:

    •  The predictions of metals volatility during waste combustion have been improved due to the
       addition of data for many toxic metal species to the thermodynamics database. Equilibrium
                                            117

-------
        predictions of metals volatility based on the improved thermodynamics database are in better
        agreement with field measurements, particularly for arsenic, chromium, and nickel.

    •   The methodology for the prediction of uncontrolled emissions of metals from waste combustion
        devices seems to work well in comparison to a full scale HWI measured data. This suggests that
        using this modeling approach can be effective in planning tests and optimizing operating
        conditions prior to conducting actual tests on HWIs and on boilers and industrial furnaces.

        Additionally, a number of preliminary conclusions  have been reached based on analysis of the
data from the metals emissions database:

    •   Concentrations of volatile metals emitted from incinerators with fabric filters are generally lower
        than from incinerators with venturi scrubbers.

    •   Volatile metals tend to be more highly enriched in captured particulate matter than do nonvolatile
        metals.

    •   Nonvolatile metals are more likely to partition to the bottom ash than are volatile metals.

    •   Arsenic emissions tend to increase with increasing primary chamber temperature.

    •   The ratio of hexavalent to total chromium emissions ranges from 0.1 to 100%, with the higher
        numbers coming from facilities which have low total chromium emissions.

    •   The ratio of pure metal to PM emissions is generally below 10%; however, mercury emissions
        have been observed to exceed PM emissions in facilities with very low PM emissions.

Because of the limited amount of data currently in the database, these conclusions are still preliminary.  In
the future it is planned to collect and include into the database all available incinerator and boiler and
industrial furnace metals data.

REFERENCES

(1)  Barton, R.G., Maly, P.M., Clark, W.D., Seeker, W.R.,  and Lanier, W.S.  Prediction of the Fate of
    Toxic Metals in Hazardous Waste Incinerators. Final Report, EPA Contract 68-03-3365, U.S.
    Environmental Protection Agency, Washington, D.C., 1988.

(2)  Rizeq, R.G., Clark, W., and Seeker, W.R. The Behavior of Metals in Waste Combustion Devices.
    FY91 Final Report, EPA Contract 68-CO-0094, Work Assignment 0-3, U.S. Environmental Protection
    Agency, Cincinnati, Ohio, 1992.

(3)  Rizeq, R.G., Hansell, D.W., Clark, W., Pooler, L, and Seeker, W.R. The Behavior of Metals in
    Hazardous Waste Combustion Systems.  FY92 Final Report, EPA Contract 68-CO-0094, Work
    Assignment 1-1, U.S. Environmental Protection Agency, Cincinnati, Ohio, 1992.

(4)  Radian Corporation. A Performance Test on a Spray Dryer, Fabric Filter, and Wet Scrubber System.
    Draft Test Report, DCN: 89-232-011-034-06, U.S. Environmental Protection Agency, Washington,
    D.C., 1989.

ACKNOWLEDGEMENTS

        The authors gratefully acknowledge the support of, Dr. C.C. Lee, the EPA project officer, and
Mr. Shiva Garg, the EPA Work Assignment Manager. For more information on this project, they may be
contacted at:  C. C. Lee, U.S. EPA Risk Reduction Engineering Laboratory, 26 West Martin Luther King
St., Cincinnati, OH 45268, Tel. (513) 569-7520; or Shiva Garg, U.S.  EPA Office of Solid Waste, 2800
Crystal Drive, Crystal City, Arlington, VA 22202, Tel. (703) 308-8859.
                                              118

-------
        PREDICTION OF DUST GENERATION FROM HANDLING POWDERS IN INDUSTRY

     M. Plinke, D. Leith, and M. Boundy, Department of Environmental Sciences and Engineering,
    117 Rosenau Hall, CB #7400, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599,
                                    Tel. 919 966 3851, and
   Thomas Holdsworth, U.S. EPA, Office of Research and Development, Risk Reduction Engineering
                         Laboratory, Cincinnati, Ohio, Tei. 513 569 7675
INTRODUCTION

    This paper presents the results of a two-year investigation of factors that affect dust generation. The
success of the work reflects collaboration between researchers at the U.S. EPA Office of Research and
Development, MS and PhD students from the University of Karlsruhe, Germany, and students, staff, and
faculty from the University of North Carolina, with significant input from the Ecological and Toxicological
Association of the Dyestuffs Manufacturing Industry (ETAD), an international industry group concerned
with generation of dusts from powdered dyes. The work performed has been useful to the U.S. EPA
Office of Toxic Substances (OTS) for Pre-Manufacture Notification (PMN) reviews, in which the potential
of materials to generate toxic dusts must be assessed.  In addition, the work should be  useful to industry
as it identifies the properties of materials as well as the properties of processes that affect dust
generation in industry.

    This work has resulted in three papers already published or accepted  for publication in technical
journals (1,2,3) one paper submitted for publication (4),  and at least two additional papers that will be
submitted for publication in the coming year (5,6.) It has  been  the subject of diploma theses for four
students at the University of Karlsruhe, the MSEE thesis for two students  at the University of North
Carolina, and for one student a PhD thesis conducted at the University of North Carolina for submission
at the University of Karlsruhe.

    Powders and granulated solids are used throughout industry.  Handling these materials at transfer
points, bagging, or dumping stations generates respirable dust that may affect worker health or create a
safety problem.  To evaluate this hazard, the dust concentration in the worker's breathing zone must be
determined. This research investigates factors that affect dust generation to obtain a fundamental
understanding of the dust process by which dust is generated.
       Energy Input
Separation Forces

  Binding Forces
                                                           Dust Generation
 Figure 1  Model of dust generation

    Figure 1 shows the approach taken here. Dust is generated when energy is introduced to a granular
 material, creating separation forces that drive the particles of material apart. An example of a separation
 force is the impact or inertia force with which falling granules hit a solid surface. To generate dust, these
 separation forces must overcome interparticle binding forces, or cohesion forces, that hold the granular
 material together. Our hypothesis, Eq. (1), asserts that G/, the amount of dust generated with size V
 depends on the fraction of particles with size "f in the test material, Fracj, and on the ratio of the
 separation forces to the interparticle binding forces.
 Dusf Generation Rate, G/ =(Frac/ )°
         (Separation Forces)b
     (Interparticle Binding Forces)0
(1)
                                            119

-------
 METHODOLOGY
    To evaluate the coefficients a, b, and c in Eq. (1), three test methods had to be developed: one to
 measure the interparticle binding forces, another to measure the particle separation forces, and a third to
 measure the amount and size distribution of the dust generated during a fall of material. Conventional
 laser diffraction measurements were used to determine the size distribution of the materials.  With
 values known for all terms in the equation, the unknown coefficients could be evaluated.

    (1) Interparticle Binding Forces -  The Peschl Rotational Split Level Shear Tester or Peschl tester
 (IPT, Vaduz, Liechtenstein) was used to determine interparticle binding forces. It consists of a
 measurement cell with two vertically-stacked rings (base and top) with an inside diameter of 50 mm.
 The sample is placed inside the cell, compacted, and then sheared  by rotation of the sample in the base
 ring against the stationary, upper part of the sample in the top ring.  A shear plane develops horizontally
 in the middle of the sample between the two rings. The torque necessary for the top ring to resist the
 rotational movement of the base ring is transmitted to a pressure transducer.  This torque is proportional
 to the interparticle binding forces. The dependent variable used here to describe interparticle binding
 forces is cohesion: the resistance of a bulk material against shearing normalized by the shear area and
 without normal strain in the bulk material.

    (2) Particle Separation Forces -  Particle separation forces were determined by measuring the impact
 of material falling at a constant rate onto a pile. The pile was built on a plate set on a balance, so that
weight change could be recorded overtime. Impact force could be determined by noting the difference
 between pile weight while material was falling vs. pile weight while no  material was falling. The impact
force divided by material flow rate was used to represent impact or separation forces.

    (3) Dust Generation - Figure 2 shows the apparatus used to measure dust generation.  This
apparatus has two sections: the dust generation section on the left and the dust measurement section on
the right. Separation of the generation and measurement sections enables independent examination of
dust generation and dust transport mechanisms.  Granular material falls from a known height at a mea-
sured rate and passes through a hole centered on the lid of the hopper. Within the hopper, falling
material hits a pile of the same material built under the same experimental conditions. Air'is entrained
with material as it enters the.hopper. From the hopper the entrained dusty air is drawn through a wide
slot into the measurement area by fan "B". In the dust measurement area fan "A" circulates air through
an elutriation column and an air return channel. The elutriation column prevents particles larger than
25 urn in aerodynamic diameter, and thus not respirable, from reaching the column top where an
impactor separates particles smaller than 25 urn into seven size categories.
                                   Fan
                                                  Impactor
                 Hole of
            Adpstable Diameter
                  Vent
 Cutoff Valve

  Elutrlatton Column


/ Air Return



  Fan"B"
                      Receiving Hopper       \      Large Particles Settle Out
                                  Baffles Distribute Air Evenly


                     -Generation Section—»1»—Measurement Section—*
   Figure 2  Schematic diagram of test apparatus
                                             120

-------
    Table 1 lists variables tested in this work, and the levels at which experimental measurements were
made.  In all, over 800 pilot-scale tests were conducted in this investigation.

              TABLE 1  VARIABLES INVESTIGATED IN DUST GENERATION TESTS
Parameter
Dependent
Variable
Drop Height
Flow Rate
Material
Moisture Content
Particle Size
Range of Material
Expected
Influence on
Separation
Forces
Separation
Forces
Binding /
Separation
Forces
Binding
Forces
Binding
Forces
Range for
Binding Force
Measurement
Cohesion
N/A
N/A
TiC-2 , Limestone,
Glass Beads,
Lactose
0 to 5 %
d<5nm, 525|im
Range for
Separation Force
Measurement
Impaction
0.25m to 1.5 m
0.1kg/sto0.6 kg/s
TiC-2 , Limestone,
Glass Beads,
Lactose
0 to 5 %
d<5^m, 525nm
Range for
Dust Measurement
Dust Generation Rate
0.25 m to 1 .5 m
0.1 kg/s to 0.6 kg/s
TiC-2 , Limestone,
Glass Beads,
Lactose
0 to 5 %
d<5jim, 525nm
    As shown in Table 1, the material flows tested in this work are representative of processes on an
industrial scale. As a result, the apparatus shown in Figure 2 and used for this work was necessarily
large: about 2 m tall and 1 m x 2 m in cross-section.  This apparatus is inappropriate for tests on
materials that are toxic or valuable, as the tests reported here required handling many kg of material.
Accordingly, a miniaturized version of the dustiness test apparatus was developed (3), which requires
only a few hundred grams of material, and which is completely enclosed. Some experiments with the
miniaturized apparatus have been conducted already, and are the subject of a poster presented at this
meeting. Further investigations to establish the equivalence of results for the full-scale apparatus and
the miniaturized apparatus are now underway.
RESULTS

Binding Forces
                                                     Moisture Content
                                                     Mass Median Diameter
                                          Moisture Content
                                        Mass Median Diameter
Figure 3.  Influence of material moisture content and particle size on cohesion
                                              121

-------
    As shown in Figure 3, increasing moisture content increased cohesion for materials that do not
absorb or react with water in the moisture range investigated. The rate of this increase depends on the
material and its size distribution.  The higher the specific surface of a material, the slower its increase in
cohesion with increasing moisture content.  Figure 3 also shows that fine particles had higher cohesion
than coarse particles, which suggests that coarse  particles are easier to separate. Thus, bulk materials
consisting primarily of coarse particles should be more dusty than  materials composed primarily of fine
particles, although large particles may not remain airborne for long.  The four materials investigated
displayed different cohesion at similar moisture contents and size  distributions.  For example, under dry
conditions limestone had lower cohesion than titanium dioxide, which suggests from Eq. (1) that
limestone should be more dusty than titanium dioxide, everything  else being equal.

Separation Forces
                                                   Drop Height
                                                   Material Flow Rate
                                        Drop Height
                                      Material Flow Rate
Figure 4.  Influence of material drop height and material flow rate on separation forces

    As shown in Figure 4, increasing drop height greatly increased impaction or separation forces in all
cases. Increasing the flow rate did not affect separation forces. The impaction varied greatly between
different materials. Titanium dioxide created the highest impaction,  followed by lactose, limestone and
glass beads.  Processes in which easy flowing, low-density particles  are dropped from a low drop height
create the lowest separation forces on the particles.

Dust generation
                                                          Cohesion
                                                          Impaction
                                     Binding Force / Cohesion
                                    Separation Force / Impaction
Figure 5.  Influence of binding and separation forces on dust generation
                                              122

-------
    Figure 5 shows that dust generation increased with increasing separation forces caused by a higher
energy input such as an increase in drop height.  Dust generation decreased when the interparticle
binding forces were increased by the addition of water to the parent material or by low particle size.  A
statistical model for dust generation rate, G/, was developed as postulated in Eq. (1), using binding and
separation forces and material size distribution as independent variables. The R2 value for this equation
was 0.7. Thus, with knowledge of binding and separation forces, together with the size distribution of the
material, the amount and size distribution of dust particles generated by handling a granular material can
be predicted.

    Additional work is currently under way in several areas.  Whereas the results presented here are for
tests conducted with a large-scale test apparatus, the effectiveness of a miniaturized apparatus for
measuring dust generation is already under study as discussed above.  Although the work presented
here is for dust generated when granular material falls into a pile, separation forces should be
characterized for mixing, grinding, and other common industrial processes as well. During the last year,
theoretical work has been initiated to predict binding forces, and hence dust generation, using only the
physical and chemical properties of a material.  This approach will eliminate the need for measuring
interparticle binding forces using the Peschl  tester.

CONCLUSIONS

    Reducing dust exposure to workers is important for industry and government. This paper describes a
method to estimate the amount and size distribution of respirable dust that is generated when powders
fall through still air onto a pile. The work postulates that dust generation rate for particles of any size is
related to:  (1) the fraction of particles with that size in the parent material, (2) interparticle binding forces
within the powder, arid (3) interparticle separation forces caused by the process.  Statistical and graphical
analyses of these results show the importance of each factor in a model for dust generation.  This model
and two simple tests enable the prediction of dust generation in industrial processes that involve material
falling into a pile. This work lends insight into material or process modifications that are effective at
reducing dust generation in industry; in addition, it will help government agencies assess the  potential
dustiness of powders and granular materials.

REFERENCES:


1.    M. Plinke  and D. Leith, D.  Hdlstein, and M. G.  Boundy, Experimental Examination of Factors that
      Affect Dust Generation. Am. Ind. Hva. Assoc. J. 52 (12): 521-528 1991.

2.    M. Plinke, R. Maus, and D. Leith, Experimental Examination of Factors that Affect Dust
      Generation Using the MRI and Heubach Testers, Am.  Ind. Hyg. Assoc. J. 53 (5) 325-330 1992.

3.    B. Cawley, D. Leith, A Bench Top Apparatus to Examine Factors that Affect Dust Generation.
      Accepted  by ADD!. OCCUD. Environ. Hyg. (1993)

4.    M. Plinke, D. Leith, V. Schurmann, and R. Hathaway, Experimental Examination of Interparticle
      Binding Forces. Submitted to Powder Technology (1993)

5.    M. Plinke, D. Leith, Experimental Examination of Particle Separation Forces. Planned for
      publication (1993)     '

6.    M. Plinke, D. Leith, Prediction of Dust Generation from Handling Powders in Industry.  Planned for
      publication Am. Ihd. Hyg. Assoc. J.. (1993)

FOR MORE INFORMATION:

Thomas Holdsworth, U.S. EPA, Office of Research and Development, Risk Reduction Engineering
Laboratory, Cincinnati, Ohio,  Tel. 513 569 7675
                                               123

-------
                     SOIL DETOXIFICATION USING SOLAR TECHNOLOGY:
                  THE RESULT OF ERA'S PART OF THE TRI-AGENCY EFFORT

                                         Paul Gorman
                                            Ed  Ball
                                          John  Jones
                                       Pam Murowchick
                                   Midwest Research Institute
                                     425 Volker Boulevard
                                  Kansas City, Missouri 64110
                                        816-753-7600
 INTRODUCTION
        As part of the Tri-Agency effort, the Environmental Protection Agency (EPA) contracted with
Midwest Research Institute for design, construction, and testing of a  Minipilot Solar Reactor System
with liquid organic feed.  The system was designed to enable study of the destruction of organic
compounds and emission of products of incomplete combustion (PICs) when combustion of the liquid
organic feed was augmented with concentrated solar input.

        This project did not involve use of solar energy for desorption of organics from soil.  Rather, it
was directed to use of solar energy for destruction of organics that would be desorbed from soil by
conventional low temperature thermal desorption, with recovery of the desorbed organics in liquid form.
Advantages of this concept are that it utilizes conventional desorption equipment and enables soil
desorption operations to be continuous, independent of solar availability.  Also, this concept enables the
solar destructor to be relatively  small  since it needs to process only the amount of organics recovered
as liquid from the soil desorption operations.

        Testing of the minipilot solar reactor system is intended to include determination of destruction
of principal organic hazardous constituents (POHCs) and emission of PICs, when operating with
concentrated solar input using facilities at the National Renewable Energy Laboratory (NREL).
However, it was first necessary to conduct tests without solar input, in order to determine if it would be
possible to quantify significantly lower levels of emissions that might be provided by solar input.

        Testing of the system with solar input has not yet been carried out,  but testing  without  solar
input has been completed. Those results are summarized in the paper, along with  results from tests
done with input of ultraviolet (UV) light.

METHODOLOGY

        The minipilot solar reactor system was designed to operate in two modes:  single or dual
chamber. In single-chamber mode, the liquid feed is combusted in one chamber that is equipped with
a quartz window.  In dual-chamber mode, the liquid is combusted in one chamber and  the gases then
enter the second chamber which is equipped with the quartz window.

        The entire system consisted of many components including feed and control of liquid organic
and combustion air, and effluent gas monitoring and cleanup (see Figure  1). It was fully equipped with
process monitoring instruments, flame detectors and other safety shutdown  interlocks, continuous
emission monitoring of CO, O2,  and THC, and computerized data logging  and control.
                                              124

-------
               Compressed Air •
            Liquid Organic Feed
                                                             1st Chamber
iber

1   L


  2nd
                                                                         Chamber   i

                                                                         	3   C1
Gas Effluent
 Discharge
                                                                      Gas
                                                                     Cooler
                                Carbon     • Alumina    •  Water
               Gas Pump          Bed        NaOH     Separator                   M.mv
                                           Pellet Bed


        Figure 1.  Simplified schematic diagram of Minipilot Solar Reactor System.
_Quartz
 Window
        a.tts wen
36.000   MU"™1
                                          -13.875 DIA QUARTZ WINDOW


                    Figure 2.  Solar detoxification reactor—section view.
                                            125

-------
        Design of the system was based on a liquid organic feed rate of 11 grams per minute and an
air feed rate of 10 standard cubic feet per minute. Each reactor chamber was 10.5 inches (in) inner
diameter by 36 in long with approximately 4-in-thick internal insulation and 4-in-thick external insulation
(see Figure 2). Operating temperatures could be varied over a range of 1400° to 2000°F.

        Composition of the synthetic liquid organic feed to be used in most of the nonsolar and solar
testing was selected by the Tri-Agency group.  Its composition was:
        No. 2 fuel oil
        o-Dichlorobenzene (DCB)
        Carbon tetrachloride (CCI4)
        Naphthalene
        Toluene
 40%
 40%
 10%
 5%
 5%
100%
       After initial operation of the system and preliminary testing with an on-line gas
chromatograph/electron capture detector, work proceeded into conducting tests using EPA
Methods 0030 and 0010 (volatile organic sampling train [VOST] and Modified Method 5 [MM5]) to
quantitate emissions of the POHCs as listed above and emission of PICs. As noted previously, the
primary purpose of these tests was to quantitate emissions in the nonsolar mode in order to determine
if significantly lower emissions could be quantitated when operating with solar input.  This was
determined by comparison of the measured emissions with detection limits of the methods and blank
levels.

RESULTS

       A series of 10 tests, without solar input, was carried out using the EPA methods (VOST and
MM5). This series included tests in both the single- and dual-chamber modes at two waste feed and
air input flow rates (referred to as "low flow" and "high flow").  Four of these tests were done with input
of UV light.  All tests involved analysis of samples for POHCs and PICs, and some tests also included
analysis of dioxins/furans.

       Results from the 10 tests are summarized in Table 1. Major conclusions made from these
results are presented in the next section.

CONCLUSIONS

       Major conclusions drawn from the test results were as follows:

               Destruction/removal efficiency for the four  POHCs were all above 99.999% in these
               nonsolar tests.

               Based mainly on comparison of measured emissions with blank levels, it would be
               possible to quantify a significant decrease  in  emissions of volatile and semivolatile  PICs
               and two of the POHCs (CCI4 and DCB). However, it would not be possible to  quantify
               any significant decrease in emissions of the other two POHCs (toluene and
               naphthalene).

              Tests with artificial  UV light indicated a  significant decrease in emissions of volatile
               PICs when operating  in the single-chamber mode, but no significant decrease  for
              semivolatile  PICs or for the four POHCs.
                                             126

-------
        $j
       s*

                                                       ^, CQ
           s
1
Q
                        o) in  o in
                                          cvj in

                                          oi
                                                00) „'
                                                             •--,

                                                             Q.Q B

                                                             O c> O
                                                              O xj- to
                                                              O O CD

                                                              odd
                                                                      p p
                                                                      tr TJ-

                                                                      v v
                                    cq T- o>

                                    coin
o
co
2

cr
a:

e

s
tr
8
o
cc
CO
UJ
cr
            = 00 O o CM m.
                                                   15  S:
                        °>.oi  me

                        88  a'

           gir- r-

           X
                 •;; •»
                                                             §
    I
                        -8  §!
                        §8  s <

                        S.I  g|

                        88  g><
                                                                      V V
                                                                      q o

                                                                      M- TT

                                                                      V V
                                                                     P p


                                                                      V V
    •a
             igtS^ffi
             'rSiHWia
            j-t-1*
                >"c"i>iii>
              CM o O O

              Szai^:
                                             IKS  $',
                                                             *- -r- CM
                                                             CO r~ m



                                                             pop

                                                             odd
                                          CM I

                                          5!
                                                                     S^in  cocMin  2g
                                                                     •^•rrco  qcrjoj  St
                        88  s
                                                                                   o o
                                                                                   o o
                                                                                   d P
ca
<
                                    ,o,ra
                                                O) O) O,  o u*
                                                a. a. a.  co P

                                                       •D oi
           S\
                                   2  o

                                   IIE
                                   -Pas
                                    i ^ o
                                    i|-(-
                                         lo

                                         t51
                                         ro ^o
                                         v> £=


                                         jS  i


                                         II  |

                                              CO
                                              CO
                                            g m

                                            S|5

                                            |l^

                                            Sfl
                                                       §0

                                                       co a.
w o

i5x
                                                    g
I £_

l|i§
3it|i:
                                                                                    6

                                                                                    u
                                                                                    co

                                                                                    Igfe
ii 7= .=

sfi
QZCO
       C r-   <=
                                                                                         O


                                                                                       fl

                                                                                       r <">

                                                                                       °«
                                                                                       5 CO

                                                                                       2S
                                     S'|
                                     0) Q.
                                     |§-

                                   |1 s
                                   ||g
                                   •ss-i
                                                                                    S^SS

                                                                                    5 _C j^CO

                                                                                    M 2 p'S

                                                                                    S 3 0.6
                                                                                    S « ct-
-------
       Based on these results and conclusions, it was recommended that work proceed on carrying
out tests with solar input, as planned at NREL, as soon as funding can be made available.

FOR MORE INFORMATION

       C. C. Lee or George  Huffman
       U.S. Environmental Protection Agency
       Risk Reduction Engineering Laboratory
       26 West Martin Luther King Drive
       Cincinnati, Ohio 45268
       513-569-7520 or
       513-569-7431
                                            128

-------
        COMPARISON OF THERMAL TREATMENT PCS DATA FROM TWO SUPERFUND SITES

                                       Larry R. Waterland
                               Acurex Environmental Corporation                             ..  ,
                                       555 Clyde Avenue
                                 Mountain View, California 94042
                                         (415)961-5700

                                             and

                              Marta K. Richards and Howard O. Wall         .     ,
                              U.S. Environmental Protection Agency
                              Risk Reduction Engineering Laboratory
                                 26 W. Martin Luther King Drive
                                     Cincinnati, Ohio 45268
                                 (513) 569-7783   (513) 569-7691

INTRODUCTION

        One of the primary missions of EPA's Incineration Research Facility (IRF) is to support Regional
Offices in evaluations of the potential of incineration as a treatment option for contaminated soils and
sediments at Superfund sites.  Two priority sites for which pilot-scale incineration treatability studies were
completed over the past two years were the New Bedford Harbor (NBH)  Hot Spot Operable Unit in
Region  1 and the Scientific Chemical Processing (SCP) site in Region 2.

        The NBH site near New Bedford, Massachusetts, contains approximately 10,000 yd3 of
contaminated marine sediments in a 5-acre area of the harbor. This area has been identified as the Hot
Spot Operable Unit and incineration of the dredged sediment has been selected as the treatment option.
The SCP site was operated by SCP  in the 1970's for the handling, treating, and disposing of a wide
variety of chemical wastes. These, and perhaps former, activities at the 5.9-acre site, contaminated the
soil and underlying groundwater with a wide variety of contaminants.  The primary contaminant at both
sites is PCBs, and both sites are contaminated with cadmium, chromium, copper, and lead. The  SCP
site soils are additionally contaminated with several volatile organic constituents as well as other toxic
trace metals.

        The objectives of the test programs performed at the IRF using site materials were to:  confirm
that conventional incineration can achieve the required 99.9999% destruction and removal efficiency
(ORE) for the contaminant PCBs and result in an ash discharge that is not PCB-contaminated; determine
the distribution of the contaminant trace metals among the incinerator discharge streams, including the
metals'  leachability from the kiln  ash; determine the effects of incineration temperature on PCB
destruction and metal distributions;  measure the effectiveness of a conventional air pollution control
system  (APCS)  of the type available at the IRF for collecting paniculate and trace metals; and determine
if the kiln ash from incineration treatment can be disposed of as nonhazardous solid waste by virtue of
the ash not being a TC hazardous waste.

METHODOLOGY

        Both test programs were performed in the rotary kiln incineration system (RKS) at the IRF. The
IRF RKS consists of a primary combustion chamber,  a transition section, and a fired afterburner
chamber. After exiting the afterburner, flue gas flows through a quench section followed by a primary
APCS.  The primary APCS for both test series consisted of a venturi scrubber followed by a packed-
                                              129

-------
column scrubber. Downstream of the primary ARCS, a secondary ARCS, consisting of a demister, an
activated-carbon adsorber, and a high-efficiency particulate air (HEPA) filter, is in place.

        For both test programs a number of 55-gal drums of dredged sediments or excavated soils were
shipped to the IRF, where they combined to form one test feed material for each test program.  Before
testing the combined sediments were repackaged into 1.5-gal fiberpack containers for feeding the RKS
via the ram feeder in place on the system.  Because neither the NBH site sediments nor the SCP site
soils contained sufficient PCB contamination (5,300 mg/kg in the NBH sediments and 1,290 mg/kg in
the SCP soil) to be able to establish 99.9999% ORE at typical RKS feedrates of 55 to 68 kg/hr (120 to
150 Ib/hr), it was decided to spike the soils to higher PCB concentrations to provide a margin in the
ability to establish 99.9999% ORE. The material used to spike the soils was an Askarel transformer fluid
composed roughly of 70 to 75% Aroclor 1242 and 25 to 30% Aroclor 1254. The PCB spike was added
to the fiberpacks during packaging.

        For both test programs three incineration tests were performed, one at a nominal kiln exit gas
temperature of 816°C (1,500°F) and two at a nominal kiln exit gas temperature of 982°C (1,800°F).  For
all three tests for each material the afterburner temperature was nominally 1,204°C  (2,200°F).  The
major difference  in incinerator operating conditions between test series was the kiln solids residence
time, which was 0.5 hr for the NBH tests and 1 hr for the SCP tests.  One of the tests at  high kiln
temperature for both series included a period of operation feeding both native (unspiked) and spiked
sediments/soil.  The period of operation with unspiked soil/sediments was denoted Test 2a, the period
with spiked soil/sediments Test 2b in the following  discussion.

RESULTS

        Table 1 summarizes the PCB contents of each incineration test sample. As noted in the table,
the spiked NBH sediment feed contained 4.59% PCB.  The kiln ash resulting from the incineration of the
sediments (both spiked and native), without dewatering, had substantially reduced, though still
significant, PCB contents.  The kiln ash for the spiked wet sediment feeds contained between 130 and
250 mg/kg PCB.  Interestingly, within the range of the variability of the data, the higher kiln temperature
tested for Tests 2b and 3 did not  result in significantly lower kiln ash PCB concentrations than the lower
temperature tested in Test 1. The kiln ash resulting from native sediment feed incinerated without
watering also contained significant PCB levels at 100  mg/kg.

        For the SCP tests, the spiked soil (Tests 1, 2b, 3) contained about 4% PCB compared with the
native soil content of 0.12% (Test 2a). The kiln ash-from the high kiln temperature tests (Tests 2a, 2b,
and 3) contained no detectable PCB.  However, the kiln ash from the low kiln temperature test (Test 1)
contained 56 mg/kg.  Taking results from the two tests together, it appears that even the high kiln
temperature condition is insufficient to achieve complete PCB decontamination with short, 0.5 hr, solids
residence time.  Longer solids residence times of 1 hr do not lead to complete PCB decontamination at
the low kiln temperature tested. Both high kiln temperature (982°C [1,800°F]) and  long  solids residence
time (1 hr) seem required for complete PCB decontamination.  This observation not withstanding, all test
conditions gave greater than the required 99.9999% PCB ORE, as shown in Table 1.

        Table 2 summaries the test trace metal distributions among the three incineration system
discharges, the kiln ash, scrubber liquor, and scrubber exit flue gas, for each test for the four common
site material trace metal contaminants.  Quite consistent trace metal partitioning results are seen for the
two materials tested. For all tests chromium exhibited relatively nonvolatile behavior. The kiln ash
discharge represented the predominant fraction  of the discharged amount, from 88  to 98.5%. Chromium
distributions were unaffected by kiln temperature for the NBH tests, and showed  a slight  decrease in kiln
ash fraction with increased kiln temperature for the SCP tests.  Copper exhibited greater volatility,
though was still relatively nonvolatile. Again, copper's distribution were unaffected by kiln temperature
                                              130

-------
                          TABLE 1. PCB ANALYSIS RESULTS
      Parameter measured
Test 1
Test 2a
Test 2b
Tests
NBH Tests
Average kiln exit gas temperature
  °C                             824
  (°F)                            (1,516)
PCB concentration
  Sediment feed, %                 4.59
  Kiln ash, mg/kg                  220
  Scrubber liquor, /yg/L             < 1
  Scrubber exit flue gas, //g/dscm     0.98
PCB ORE, %                      99.999938

SCP Tests
Average kiln exit gas temperature
          981
          (1,797)
          0.615
          100
           985
           (1,805)
           4.59
           250
           984
           (1,803)
           4.59
           130
                        <0.35        0.75

                        > 99.999980   99.999947
°c
(°F)
PCB concentration
Soil feed, %
Kiln ash, mg/kg
Scrubber liquor, //g/L
Scrubber exit flue gas, //g/dscm
PCB ORE, %
818
(1,504)

4.07
56
<3.3
<0.28
> 99.999981
986
(1,807)

0.12
<0.33

987
(1,808)

4.01
<0.33
<3.3
<0.14
> 99.999990
983
(1,802)

4.24
<0.33
<3.3
<0.14
> 99.999990
*_ = Sample not collected.
                                         131

-------
                   TABLE 2. NORMALIZED TRACE METAL DISTRIBUTIONS
NBH Tests
Test 1 2b
Kiln exit gas temperature, °C 824 985
°F (1,516) (1,805)
Apparent
3
984
(1,803)
distribution,
1
818
(1,504)
% of metal
SCP Tests
2b
987
(1,808)
measured

3
983
(1,802)

   Cadmium

    Kiln ash
    Scrubber exit flue gas
    Scrubber liquor
     Total
   Chromium
    Kiln ash
    Scrubber exit flue gas
    Scrubber liquor
     Total
   Copper
    Kiln ash
    Scrubber exit flue gas
    Scrubber liquor
     Total
   Lead
    Kiln ash
    Scrubber exit flue gas
    Scrubber liquor
     Total
61
23
16
100
88
4
8

100
83
6
11

100
53
17
30

100
8
36
56

100
92
2
6

100
82
10
8

100
19
55
26

100
19
42
39

100
92
2
6

100
89
5
6

100
23
42
35

100
61
32
7

100
98.5
0.1
1.3

100
93
4
3

100
56
26
18

100
5
61
34

100
92
1
7

100
66
22
12

100
9
72
19

100
3
63
34

100
91
3
6

100
83
11
6

100
18
54
28

100
for the NBH tests, and showed a slight decrease in kiln ash fraction with increased kiln temperature for
the SCP tests.

       Cadmium and lead were moderately volatile at the low kiln temperature tested for both test
series; about 50 to 60% of the discharged amount of each was accounted for in the kiln ash. However,
both were significantly more volatile at the high kiln temperature in both test series.  The kiln ash
fractions for both metals were decreased.

       For the NBH tests, flue gas particulate levels at the scrubber exit ranged from 70 to
101 mg/dscm, corrected to 7% O2. These levels were below the 180 mg/dscm at 7% O2, hazardous
waste incinerator performance standard. Measured  scrubber HCI collection efficiencies were 98.8 to
99.9%, just at or above the incinerator performance  standard of 99%.
                                             132

-------
       In contrast, the scrubber exit paniculate level for the SCP tests was 88 mg/dscm for the low kiln
temperature test.  Levels were significantly higher for the high kiln temperature tests, at 240 to
310 mg/dscm at 7% O2. Evidently the paniculate generated by the high kiln temperature test condition
was quite difficult to collect, even more so than the low kiln temperature test paniculate. The scrubber
exit paniculate levels measured for Tests 2b and 3 exceed the hazardous waste incinerator performance
standard of 180 mg/dscm at 7% O2. Scrubber HCI collection efficiencies for the two high kiln
temperature tests were 99.8 to 99.9%.  This level of control efficiency meets the incinerator performance
standard.  Scrubber exit HCI levels were about a factor of 10 higher for the low kiln temperature tests.

CONCLUSIONS

       Greater than 99.9999% PCB ORE was achieved in the incineration treatment of both the NBH
sediments and the SCP soils at both incineration temperatures tested. However, when NBH sediments
that had not been dewatered were incinerated at a kiln solids residence time of 0.5 hr, the treated
sediments (kiln ash) were still PCB-contaminated, regardless of kiln temperature.  The same was true for
the treatment of SCP soils at low kiln temperature and  increased kiln  solids residence times of 1 hr.
Complete PCB decontamination of the SCP soils occurred only at the 1 hr solids  residence time under
the high kiln temperature tested. Apparently the combination of high kiln temperature and 1 hr solids
residence time is needed to effect complete PCB decontamination of these materials.

       Of the contaminant trace metals, chromium and copper were relatively nonvolatile. The kiln ash
discharge accounted for 66 to 98.5% of the measured discharged amounts of these metals. These
fractions were not affected  by the range of kiln temperatures tested for the NBH material; a slight
decrease in  kiln ash fraction for both metals occurred with increased  kiln temperature for the SCP soil.
Cadmium and lead exhibited relatively volatile behavior, and increasingly so at the high kiln exit gas
temperature. The kiln  ash discharge contained 53 to 56% of the lead and 61% of the cadmium
accounted for in the discharges at the low kiln temperature. These fractions decreased to 3 to 19% for
cadmium and the 9 to 23% for lead at the high kiln temperature.

       Reported results suggest that incineration would be an effective treatment option for the NBH
site sediments.  However, sediment dewatering before  incineration and/or incineration at high kiln
temperature and solids residence times of up to 1  hr might be required to yield a treated sediment not
contaminated by PCBs.  Incineration under the high kiln temperature  conditions tested might also be an
effective treatment option for the SCP site hot layer soil. However, the paniculate control capabilities of
a wet scrubber APCS of the design in place at the IRF  will likely not be sufficient to meet the hazardous
waste incinerator paniculate emission standard.

       For  more information contact the NBH test Technical Project Monitor  (TPM) Marta K.  Richards,
at (513) 569-7783 or the SCP TPM,  Howard O. Wall, at (513) 569-7691. The address for both is:

            Risk Reduction Engineering Laboratory
            U.S. Environmental Protection Agency
            Cincinnati, OH  45268
                                            133

-------
            SITE Demonstration of the Chemical Waste Management
                   "X*TRAX" Thermal Desorption Process
                              Paul R. de Percin
                     Risk Reduction  Engineering  Laboratory
                     U.S. Environmental Protection Agency
                            Cincinnati, Ohio 45268
Abstract
In May 1992 a field demonstration of the CWM X*TRAX thermal desorption process
was performed at the ReSolve superfund site in North Dartmouth, MA under the
USEPA Superfund Innovative Technology Evaluation (SITE) program.  PCB
contaminated soils and wetland sediments were successfully treated to less
than 1 mg/kg.


Introduction

      The X*TRAX SITE Demonstration was conducted in May 1992, after a proof-
of-process test at the"Re-Solve Superfund Site in North Dartmouth, MA.  The
system is being used to treat approximately 35,000 tons of soil and sediment
contaminated with PCBs at the site.  For this application, the soil is
screened to remove objects larger than 1 inch in size.

      During the Demonstration, about 215 tons of soil were treated at an
average feed rate of 4.9 tons per hour, a residence time of 2 hours and an
average treated soil temperature of 732°F.   Soil  PCB concentrations ranged
from 181 to 515 mg/kg.  Average flow rates for the carrier gas and process
vent gas were 700 and 37 acfm, respectively.


Technology Description

      The X*TRAX Model 200 Thermal Desorption System developed by Chemical
Waste  Management, Inc.  (CWM) is a low-temperature process designed to
separate organic contaminants from soils, sludges and other solid media.  The
X*TRAX Model 200 is fully transportable and consists of three semitrailers,
one control room trailer, eight equipement skids, and various pieces of
movable equipment.  The equipment requires an area of about 125 feet by 145
feet.
                                      134

-------
      The X*TRAX system is a thermal  and  physical  separation process;  it does
not involve incineration.  Contaminated solids are fed into an externally
heated rotary dryer where temperatures range from 750 to 950°F.   Evaporated
contaminants are removed by a recirculating  nitrogen carrier gas that  is
maintained at less than 4% oxygen  to  prevent combustion.  Solids leaving the
dryer are sprayed with treated cooling water to help reduce dusting when the
treated solids are returned to their  original  location and compacted in place.
The nitrogen carrier gas is treated to remove and recover dust particles,
organic vapors and water vapors.

      An eductor scrubber removes  the dust particles and 10 to 30% of  the
organic contaminants from the carrier gas.   Scrubber liquid collects in a
phase separator from which sludge  and organic liquid phases are pumped to a
filter press, producing filter cake and filtrate.   The filtrate is then
separated into organic liquid and  water phases.  Most contaminants removed
from the feed solids are transferred  to the  organic liquids or the filter
cake.  The filter cake is typically blended  batchwise with feed solids and
reprocessed in the X*TRAX system,  while the  concentratted organic liquids are
typically treated or disposed of off  site.

      Carrier gas exiting the scrubber passes through two condensers in series,
where it is cooled to less than 40°F.  The condensers separate most of the
remaining water and organic vapors from the  gas stream.  Organic vapors are
recovered as organic liquids; water is treated by carbon adsorption and either
used to cool and reduce dusting from  treated solids, or treated and
discharged.  About 5 to 10% of the nitrogen  gas exits the system through a
process vent, passing through a particle  filter and carbon adsorption  system
before being discharged to the atmosphere.   The volume of gas released by the
X*TRAX system is about 100 to 200  times less than the amount released  by an
equivalent capacity incinerator.


Demonstration Results

      The Demonstration included three identical tests, each lasting 6 hours.
During each test, solid, liquid and gas samples were collected from feed soil,
treated soil, filter cake, filter  press filtrate,  condensed aqueous liquids,
water used to wet treated soil, and process  vent gases before and after
activated carbon treatment.  Condensed organic liquids were collected  before
the start of the first test and after completion of the third test.
Extensive analyses were performed  under rigid quality assurance procedures.
Key findings from the X*TRAX SITE  Demonstration are:

1)    X*TRAX successfully removed  PCBs from  feed soil and met the site-
      specific treatment standard  of  25 mg/kg for treated soils.  PCB
      concentrations in all treated soil  samples were less than 1 mg/kg
      and the average concentration was 0.25 mg/kg.  The average PCB
      removal efficiency was 99.9%.

2)    Polychlorinated dibenzo-p-dioxins (PCDD) and polychlorinated
      dibenzofurans (PCDF) were not formed within the X*TRAX system.

3)    Organic air emissions from the  X*TRAX  process vent were negligible
      (0.4 grams/day).  PCBs were  not detected in vent gases.
                                     135

-------
4)    X*TRAX effectively removed other organic contaminants from the
      feed soil.  Concentrations of tetrachloroethene, total recoverable
      petroleum hydrocarbons, and oil  and grease were all reduced to
      below detectable levels in the treated soil.

5)    Metals concentrations and soil physical  properties were not
      altered by the X*TRAX system.
                                      136

-------
           EVALUATION  OF  CORROSION  INHIBITORS TO REDUCE LEAD IN THE
                          DRINKING  WATER  IN BUILDINGS

            Thomas J.  Sorg, Darren A.  Lytle,  and Michael  R.  Schpck
                     U.S. Environmental  Protection Agency
                    Risk Reduction Engineering Laboratory
                       Drinking Water  Research  Division
                                (513)  569-7370

INTRODUCTION

     The U.S. Environmental Protection Agency (U.S. EPA)  promulgated its final
drinking water regulations for lead in June 1991  (1).  The  new rule is a more
complex rule and applies to all community water systems  and nontransient,
noncommunity systems.   The new regulations established a  maximum contaminant
level goal (MCLG) for lead at zero and replaced the interim maximum
contaminant level (MCL) of 0.05 mg/L,  (the enforceable level)  with a treatment
technique requirement. The new treatment requirement includes  optimal
corrosion control source water treatment, public education,  and lead service
line replacement.  These requirements  are triggered by instances in which an
action level for lead of 0.015 mg/L (1L sample) is exceeded in the 90th
percentile of the samples collected at the customers tap.  In  other terms, the
regulations require treatment action to be taken by the water  utility if more
than 10 percent of the monitoring samples exceed 0.015 mg/L.   The action level
is not an MCL, but it represents a level  at which a utility must take some
action to reduce lead in its drinking  water.  Although the  goal  of the
regulation is to reduce lead levels to as low as possible,  the 0.015 mg/L
action level is commonly used as an acceptable  level.

     Lead is not normally found in drinking water leaving the  water treatment
plant.  The principal  source of lead  in drinking water at the  consumer's tap
is the materials used in the distribution and home plumbing system that
contain lead, such as lead based solder,  brass  faucets and  fixtures, and lead
service lines.  When lead containing materials  are exposed  to  drinking water,
the dissolution of lead can occur.   The degree  of dissolution  is dependent
upon water quality and standing water  contact time.

     Numerous studies have been conducted to document  elevated lead levels in
the drinking water in household outlets.   Exposure to  high  lead levels in
drinking water, however,  is not limited to only the household  environment.
Elevated lead levels have been frequently found in the drinking water in
office buildings or other similar facilities whose plumbing systems contain
lead based materials.   Of particular concern is the occurrence of high lead
levels in school buildings.

     The concern for lead in drinking  water has resulted  in monitoring of lead
from outlets in many types of buildings.   In situations where  high levels have
been found, various solutions to decrease the lead levels have been
implemented.  Some of the control measures that have been used are individual
treatment devices on the outlets with  high lead levels,  replacement of the
lead containing plumbing materials, and water quality  modifications to reduce
the leaching of lead into the drinking water.
                                       137

-------
     Modification of the water quality by either changing the corrosivity of
the water with pH and/or alkalinity adjustment or the addition of chemical
corrosion inhibitors are two methods listed in the lead regulations as
acceptable treatment techniques.  Published information on the application and
effectiveness of these methods in building water distribution systems is very
limited, however.

     In 1990, a federal building in the Washington, D.C. area was found to
have a serious problem of high levels of lead in many of the outlets in the
facility.  Special monitoring tests suggested that the source of lead was lead
based solder and brass faucets and fixtures.  Various options were considered
to solve the problem with the addition of a corrosion inhibitor to the
buildings water supply being selected as the most practical and economical
method.

     Because several corrosion inhibitors are available, a pilot study was
conducted to determine the effectiveness of three different inhibitors.  This
paper describes the study and results.

METHODOLOGY

     The facility is a research building containing many laboratory outlets
(faucets) as well as the normal drinking water fountains found in any office
type building.  Although the building was approximately four years old, the
building was unoccupied during the four year period, except for maintenance
employees.  The delay in occupancy was the result of a series of structural
defects that required extensive time to correct.

     Because the water system had been virtually unused, a monitoring study
was conducted in two wings of the building to determine the effect of water
usage to reduce lead levels and the overall extent of the lead problem.  Nine
outlets in each of two wings of the building were selected to be sampled twice
a week during the water use study.  The water at each selected outlet was run
four times a day, five days a week, for a half hour each run period.
Overnight standing samples (250 ml) were collected from each sample outlet
each Tuesday and Friday and analyzed for lead, copper and zinc.  Water samples
were also collected and analyzed for a number of general water parameters,
such as pH, alkalinity, phosphate, silicate, ammoniac, sulfate, nitrate,
chloride, iron, calcium, potassium and magnesium.

     After approximately eight months, the water use study showed no
significant reduction in lead levels indicating a need to provide chemical
treatment of the water supply.  Three corrosion control  inhibitors were
selected as possible treatment chemicals; zinc orthophosphate, a generic
orthophosphate (referred to as "calcium" orthophosphate) and sodium silicate.


     The chemical feed study was conducted in a similar manner as the water
usage study.  Three wings of the facility were isolated and a chemical  feed
system installed in a utility chase of each wing to introduce the corrosion
inhibitor chemical into the drinking water system.  The feed system consisted
of a chemical feed pump and a 100 gallon chemical storage tank.  The chemical
feed pump fed the corrosion inhibitor chemical into the cold water line only
                                      138

-------
on demand according to usage.  Chemical  feed settings were adjusted to deliver
approximately 3.0 mg/L PO^   in the wings using the orthophosphate chemicals
and 30 mg/L Si02 in the wing using the silicate  chemical.

     Nine outlets in each wing were selected to  be sampled on Tuesday and
Friday.  Water was run through each tap  for a half hour four times a day, five
days a week.  Overnight standing samples of 250  ml were collected and analyzed
for lead, copper and zinc.  Additional  samples were collected on a less
frequent basis and analyzed for other water parameters as was done for the
water use study.  The chemical feed study lasted approximately six months
(mid-November to mid-April).

RESULTS

     Various studies have shown that water usage (aging) will result in
decreased lead levels in new plumbing systems with time.  The aging effect
with most waters normally requires two years or  more.  The eight month water
usage study indicated a very slight decreasing trend.in lead levels, but
because of a very wide variation in the  data it  was readily apparent that
usage alone would not solve the problem in a short period of time.  An example
of the lead results from one sampling tap is shown in Figure 1.

     The chemical feed study lasted just over five months.  The results of the
three corrosion inhibitors during the monitoring period are shown in Figure 2.
The data indicate that all three chemicals were  very effective in reducing the
lead concentration in the drinking water.  At the end of the study, the
majority of the weekly samples from all  three systems were less than 20 M9/L-
U.S. EPA recommends an action level of 20 /ng/L for lead in 250 mL samples
collected from drinking water fountains  and other type of outlets in schools
and other buildings.  Variation in the data existed from outlet to outlet and
from week to week.  Spikes of up to 500  jug/L were occasionally found and
assumed to be particles of solder or brass.  Because of these variations,
statistical analyses of the data is difficult to apply with any reasonable
degree of certainty.

CONCLUSIONS

     The results of the water usage study indicated that decreasing the levels
of lead in the drinking water from a new building by normal water usage may
take a very long time, in the case of this building and its drinking water
quality substantially longer than eight months.

     Routine sampling will likely show a very wide variation in lead
concentration from outlet to outlet and day to day.  Spikes can be expected
and are probably the results of particles of lead based material such as
solder or brass.  With system usage, this problem should decrease in time.

     Corrosion  inhibitors can be very effective in decreasing lead levels in
plumbing systems within a reasonable length of time, as shown in this case
study of approximately six months.  The results  of this study showed that all
three inhibitors were effective.  Consequently,  the selection of the corrosion
control chemical will likely be based on secondary factors.  Some of these
factors include the nature of the chemical and its impact on water quality
                                       139

-------
such as pH change, increase in zinc, sodium, phosphate, or silicate.  Other
factors for consideration are the cost of the chemical, material handling,
monitoring, and chemical storage.

REFERENCES

1.   Drinking Water Regulation:  Maximum Contaminant Level Goals And National
     Primary Drinking Water Regulations For Lead And Copper; Final Rule.
     Fed. Reg. 56(140)26460-25717, June 7, 1991.
                                      140

-------
•d
 03
    600
    500 --
    400 --
    300
    200
    100
        0     50    100    150   200   250   300
                       Time, days
Figure 1.  Test results from water usage  study (Rm. 3325).
    100
                        O  Zinc orthophosphate     	
                        •  Calcium orthophosphate
                        A  Sodium silicate
               25    50     75    100   125    150
                       Time, days
Figure 2.  Test results from chemical corrosion inhibitor
study (Rm. 1412, 2315, and 2412).
                            141

-------
                POINT-OF-ENTRY DRINKING WATER TREATMENT SYSTEM
                    APPLICATIONS AT HAZARDOUS WASTE SITES
                              James  A.  Goodrich
                           Environmental  Scientist
                     Systems and Field  Evaluation Branch
                      Drinking Water Research Division
                    Risk Reduction Engineering Laboratory
                    U.S. Environmental  Protection Agency
                        26 W. Martin Luther King Drive
                           Cincinnati,  OH   45268
INTRODUCTION
     Several small systems and individual homeowners have been faced with
the task of treating  their groundwater that has been contaminated with various
organic contaminants.  Contamination is such that the locations described in
this presentation have been designated Federal Superfund sites undergoing
either emergency or  remedial actions.  These sites have utilized point of
entry (POE) water treatment devices to treat their groundwater.  The devices
used include:  single and dual granular activated carbon (GAC) columns; air
stripping in-series  with GAC, and ozone/UV followed by GAC.  Cost (capital
and operating), contaminant removal performance, GAC breakthrough, and
disinfection by-product formation are all important factors presented.

     Hazardous waste sites often involve small communities or neighborhoods
within a larger community.  As such, these small systems have small produc-
tion, small revenues, small budgets, but big problems.  They are not capable
of taking advantage  of economies-of-scale because certain types of services,
such as maintaining  a chlorinator,  must be provided no matter how few the
connections.  Because of limited revenues, very often only part-time opera-
tors can be hired with little money available for training and certifica-
tion.  Small systems normally do not have a large pool of trained engineers
and scientists to deal with complex equipment or to deal with constantly
changing treatment needs.  The most significant requirements for small
systems are low construction and operating costs, simple operation,
adaptability to part-time operations, low maintenance, and no serious
residual problems.  These problems are exacerbated when individual
homeowners or small communities are faced with the prospect of having to
treat water deemed hazardous, and not just merely out of compliance.

METHODOLOGY

     Whole house POE treatment is an alternative to centralized treatment
technology for individuals and small systems.  In fact, it may be the only
alternative given the cost of installing a new distribution system with
central  treatment or connecting to a distant water supply.  Small  systems
may find POE devices less costly to buy, and easier to install and maintain
than a custom-designed central plant.
                                    142

-------
     Treatment technologies such as granular activated carbon (GAC),
aeration, reverse osmosis, ion exchange, and ozone/ultraviolet light have
been widely adapted to treating water for the entire house.

     The U.S. Environmental Protection Agency (USEPA) views the use of
point-of-use (POU) and POE differently.(l)  The USEPA is willing to accept
POE treatment as an available technology for complying with drinking water
regulations, but not POU devices.  In the November 1985 Federal Register,
the USEPA proposed that POU and POE treatment not be considered Best  .
Technology Generally Available (BAT) but be considered acceptable technology
to meet Maximum Contaminant Levels (MCLs), provided certain conditions were
met.(2)  This proposal was made because of difficulties associated with
monitoring compliance and effective treatment performance comparable to
centralized treatment.  In the 1987 Final Rule, POU and POE treatment
devices were not designated as BAT because:  1) of the difficulty in
monitoring the reliability of treatment performance and controlling their
performance in a manner comparable to central treatment, 2) these devices
are generally not affordable by large metropolitan water systems, and 3) not
all of the water is treated in the case of POU devices which can lead to
volatile organic chemical  (VOC) exposure through indoor air transport by
showers or dermal contact.(3)

RESULTS

     Several Superfund sites have been employing both POU and POE devices
for a few years.  Various  states have been utilizing POU/POE for their
remediation programs.  In  POU/POE applications, the predominant contaminants
being removed are the chlorinated solvents including trichloroethylene,
tetrachloroethylene, 1,1,1-trichloroethane,  1,2-dichloroethane, and trans-
1,2-dichloroethylene.(4)   Also being treated are waters contaminated by
petroleum products, aldicarb, ethylene dibromide (EDB), or radon.(5)  Table
1 summarizes the contaminants of concern and their influent levels.  The
removal efficiencies provided by the various systems range between 86 and
99+ %.
                     TABLE  1.  SUMMARY OF EXISTING DATA
                         POE WATER TREATMENT STUDY(4)
SITE NAME
& LOCATION
State of
Maine
State of
Maine
POE SYSTEM
Diffused air
stripping
Diffused air
stripping, or
packed tower
CONTAMINANTS
Gasoline and
No. 2 Fuel Oil
Radon
MAX. NO. POE
INFLUENT SYSTEMS
INSTALLED
240,000 /ig/L 100
400,000 pC/L NA
                                    143

-------

SITE NAME
& LOCATION
Suf fol k
County,
New York
Cattaraugus
County,
New York
Green County,
New York
Onendaga
County,
New York
York County,
Pennsyl -
vania
Berks
County,
Pennsyl -
vania
Adamstown,
Mary! and
Monroe
County,
Pennsyl -
vania
TABLE 1.
POE WATER
POE SYSTEM
Carbon cell
2 Carbon
Cells
2 Carbon
Cells
Packed Tower
Pref ilter,
Carbon cell,
UV light
Pref ilter,
Carbon cell,
UV light
Pref ilter,
Carbon cell,
UV light
Prefilter,
Carbon cell ,
UV light
SUMMARY
TREATMENT
OF EXISTING DATA
STUDY(4) (CONT.)
CONTAMINANTS MAX.
INFLUENT
Aldicarb
TCE00
pCE(b>
TCE
1,2-DCE
l,l-DCA(d>
TCE
1,2-DCE
TCE
PCE
DCA
TCE
1,1,1-TCA
1,1 -DCA
1,2-DCE
TCE
1,2-DCE
PCE
500 /ig/L
3,600 /ig/L
79,500 /ig/L
690 /ig/L
4,600 /ig/L
1,700 /ig/L
23,000 /ig/L
1,000 /ig/L
50 /tg/L
Ce) 570 /ig/L
520 /ig/L
44,000 /ig/L
210 /tg/L
570 /ig/L
7,000 /ig/L
290 /ig/L
30 /ig/L

NO. POE
SYSTEMS
INSTALLED
3,000
37
6
5
2
2
6
28
18
22
144

-------
                     TABLE 1.   SUMMARY OF EXISTING DATA
                    POE WATER  TREATMENT STUDY(4) (CONT.)
SITE NAME POE SYSTEM CONTAMINANTS
& LOCATION
Florida 2 Carbon Napthalene
Cells Total hydro-
carbons
Benzene
Ethyl benzene
l,2-DCA
Toluene
Xyl ene
MAX. NO. POE
INFLUENT SYSTEMS
INSTALLED
12 /ig/L 11
220 /ig/L
210 /ig/L
38 /ig/L
89 /ig/L
8 Mg/L
63 /zg/L
Polk and
Jackson
Counties,
Florida

Byron,
Illinois
Prefilter,
2 Carbon
Cells,
UV light

Prefilter,
2 Carbon
Cells
Ethylene
dibromide
TCE
PCE
(EDB)
           800 /ig/L
           500 /ig/L
           130 /ig/L
850
 10
Elkhart,
Indiana


Uniontown,
Ohio
Pref ilter,
2 Carbon Cell
Packed Tower
Aeration
Packed Tower
Aeration
TCE
s,
Carbon tetra-
chloride
Vinyl Chloride
Chloroethane
5,000 /ig/L


7,500 /ig/L
7 /ig/L
2 /ig/L
60


1
9

(a) - Trichloroethylene
(b) - Tetrachloroethylene
(c) - trans-1,2-dichloroethylene
                              (d) - 1,1-dichlorethane
                              (e) - 1,1-trichloroethane
                              (f) - 1,2-dichloroethane
     Most of the GAC treatment systems utilized at Superfund sites are of
the two-filter design.  The second carbon filter acts as a backup for the
first; if the first filter should reach breakthrough, the second carbon
filter will provide continued treatment and protection.  When carbon
replacement is necessary, the first filter is usually removed from service,
the second filter is moved to the primary position in the series and a new
filter is placed in the secondary position.  This procedure permits optimal
use of the capacity life of the activated carbon and provides satisfactory
treatment of the contaminated water.
                                    145

-------
      The  POE 6AC systems  identified  at  Superfund  sites  in USEPA Region 3
 consist of a prefilter;  a flowmeter,  two  carbon cells in series, and  a UV
 light.  Each carbon cell contains 1.5 to 2.0 ft  of granular  activated  carbon.
 Where extremely high  levels of VOCs exist (e.g.,  44,000  ug/l TCE),  several of
 the systems have a POE packed-tower air  stripper connected to the beginning of
 the treatment  train,   the air  stripper  removes most of  the  VOCs from  the
 influent  water and  thereby extends the  effective  life of the carbon filters.

      The  US  EPA Region 3  communities  currently using POE units are being
 administered by Federal  personnel  under emergency response  legislation.
 However,  this  "emergency"  has  now  been  underway in some cases for more than
   three  years.  The  communities did  not  rank high enough to be placed on
 the National  Priority List (NPL) and  thus become  Superfund  sites eligible
 for remedial  action.   This lack of NPL  listing has created  a situation where
 the federal  personnel  do  not have  a mechanism to  turn over  responsibility
 for the POE  units.

      The  following  are examples in EPA  Region 3 utilizing POE treatment
 technology.  The Cryochem Metal Company site has  19 GAC/UV  POE units  to
 remove TCE.  This is  an  interim remedy  but the homeowners seem to prefer the
 POE units over any  other  solutions.   The Kemberton site has used 23 GAC POE
 units and will  soon be connected to a public water supply in accordance with
 the State of Pennsylvania  consent  decree.  The Reticon  site is undergoing
 the Remedial Investigation phase and  is employing 6 POE units.  The Bendix
 site  is using  19 POE  units for TCE removal and is undergoing a State  of
 Pennsylvania led remedial  investigation.  The USEPA will probably agree to
 the State's decision  in this case  also.   In another case, the risk
 assessment indicated  the  few homeowners affected  would  be at a higher risk
 if they connected to  a central system because of  the total  trihalomethane
 (THMs) levels.   Although  in compliance, the risk  from 80-90 ug/l of THMs
 found in  the central  system exceeded  that of the  15-25 ug/l of TCE found in
 the homeowners'  wells.  The responsible party also prefers  to monitor the
 several POE  units in'definetly because of the lower capital  cost,  rather than
 pay to connect  to the  distant distribution system.(5)

      To date,  five  separate Superfund actions have been taken in and  around
 Elkhart,  Indiana. Elkhart is a diversified industrial community manufacturing
 Pharmaceuticals, band instruments,  recreational vehicles, and injection molded
 plastics  and foams.  A large number of  industries  in the Elkhart area use or
 have  used TCE  or other organic solvents in their  processes.

      The  levels  of  contamination were as high as  19,380 ug/l of TCE.
 Drinking  water  from these  contaminated  wells constituted an  immediate and
 significant threat  to  the  residents of  the affected households.  In
 addition,  a real threat is present from the inhalation and absorption of
water contaminated  with TCE in levels above 1500  ug/l according to the
Agency for Toxic Substance and Disease  Registry (ATSDR).

      Carbon tetrachloride  (CCLJ  was  also  found  in this  area.   As  a result,
more  than  800 residents were placed on  bottled water delivery.  It was
decided to extend city water mains to those affected areas because of the
widespread contamination.
                                    146

-------
     In total, approximately 14,500 feet of water main were installed
ranging in size from 1.5 to 12 in., and 301 homes and 7 businesses were
connected to the municipal system.  In addition, 11 homes with minor
contamination not adjacent to a water main were given POU water filters.
Two homeowners refused to have city water hooked up to their homes despite
repeated efforts to convince them of the threats posed by the VOC
contamination.

     This, however, was not the limit of Elkhart's experience with POU/POE
water treatment devices.  In June of 1986, a private citizen on the west
side of town had his well water analyzed.  The results indicated severe
contamination with both TCE and CCL.   Levels  of 800 /tg/L and 488 /ig/L,
respectively, were found.  The USEPA sampled 88 private wells and found CCL4
as high as 6,860 /ig/L and TCE as high as 4,870 jug/L.  The USEPA then decided
to install GAC POE units because of the time and distance involved in
attempting to extend the city's water mains.  In total, 54 POE filters were
installed and 22 POU filters were selected for  homes with slight contamination
near those known to have very high concentrations as a safeguard if and when
the plume expands.(6)

     Two of the homes were equipped with packed tower air strippers with two
GAC POE units in series.  The air stripper has a 40:1 air-to-water ratio and
operates at a rate of 5 gpm.  The air stripper is packed with 1 inch
diameter polypropylene cylinders.  The entire unit installed cost about
$4000.  There have not been any microbiological problems encountered to
date, although flushing has been recommended when the unit remains idle for
more than one day.  The unit can have a UV light for post-GAC disinfectant
installed.

     Granular activated carbon isotherm calculations have proven unreliable
in predicting breakthrough for the GAC POE units.  Significant under and
over estimating of time to breakthrough has been encountered.  Competitive
effects are very evident in a dual GAC unit monitored in Elkhart.  Isotherm
data estimated breakthrough for chloroform at approximately 225,000 gallons
but it was estimated to actually have broken through at approximately
130,000 gallons.  The data in Table 2 indicate the risk for homeowners if
special monitoring had not been undertaken and carbon change-out had only
been scheduled on a routine basis.

     The amount of water treated before breakthrough has ranged from 25,000
to more than 300,000 gallons.  Carbon replacement has cost approximately
$510/tank and $40 each for the sediment prefliters.  There are currently
several drums of spent carbon awaiting a decision on how they are to be
disposed.
                                     147

-------
                 TABLE 2.   ELKHART,  INDIANA BREAKTHROUGH DATA
Treated
Volume
(aallons)
173,990


188,210


198,370


Contaminant
Chloroform
Carbon tet
TCE
Chloroform
Carbon tet
TCE
Chloroform
Carbon tet
TCE
Influent
Concentration
taa/U
373
=4,400
162
356
=4,300
138
308
=3,900
111
Effluent
Concentration
(ua/U
30
112
< 1
58
224
< 1
36
309
< 1
     The State of Connecticut  is  using  an advanced oxidation  POE  unit to
treat a well contaminated with  PCE.  The well serves four  homes.  The unit
utilizes GAC following contact with ozone/UV in combination to ensure complete
removal of the PCE.  The ozone/UV process removed an average  of 89 % of the
raw water PCE levels originally found in the range of 300-600 /jg/L.  The GAC
unit removed the remaining  PCE.   No ozonation by-products  were detected.(7)

POE COST EVALUATIONS

     Two general cases are  considered here.  In the first  case, it is
assumed that for a small community a central treatment plant  exists with a
distribution system already in  place.   Because of organic  chemical contami-
nation in the water supply,  GAC treatment will be used.  For  what size
community (number of households)  is it  more cost effective to use a GAC POE
device rather than adding GAC treatment to the central plant? Assumptions
include a household water usage of 80 gallons/per capita/day  with an average
of 3.3 people per household.

     The POE system consists of two contactors in-series,  each adsorber has
about 2 cu ft of GAC, providing 4.1 minutes EBCT with a design loading of 4
gpm/ft .   Three  contaminants are considered  at removals of 95-99%, 1,2-
dichloropropane (1,2-DCP),  trichloroethylene (TCE), and dibromochloropropane
(DBCP).  Carbon replacement  for the POE systems ranged from a one- to two-
year frequency.   POE costs  include an installed cost of $2,000 with a pay-
back period of 10 years at  8% interest.   Costs for routine maintenance,
sampling and analysis were  $350/year.  Carbon replacement  costs varied by
contaminant.
                                    148

-------
     Costs associated with 6AC treatment at the central  plant vary with
system capacity and type of contaminants being removed.   Carbon service
lives ranged from 70 to 250 days for the synthetic organic contaminants
examined.  Replacement of spent carbon with virgin 6AC was assumed.  Capital
costs were amortized at 8% for 20 years.  Table 3 presents the costs of POE
and central treatment employing GAC for removal of three organics as a
function of the number of households involved.  When a very small number of
households are involved, the utilization of POE treatment systems appears to
be cost effective compared with installing additional treatment at the cen-
tral plant.  In this case, when 20 or more households are involved, treat-
ment at the central plant is more economical for all three contaminants.
          TABLE 3.   COST OF POE VERSUS  THAT  OF A  CENTRAL  SYSTEM FOR
                     REMOVAL OF ORGANIC CHEMICALS BY  GAC
Cost $/Year/Household
No. of
Households
10
15
20
25
50
DBCP
Central
1325
954
760
639
380

POE
775
775
775
775
775
TCE
Central
1332
960
766
646
385

POE
815
815
815
815
815
1.2-DCP
Central
1356
985
790
670
410
POE
900
900
900
900
900
     The  second general cost comparison case concerns a situation  in which
homes  in  a community rely on private wells for water supply.  Because of
contamination to  the area's groundwater, homes must either  install  POE
systems or be hooked-up to an existing central treatment plant.  The cost-
effectiveness of  the central system option is in part a function of the
number of households involved, total length of pipe and pumping required  for
connections, size and economics of the existing central plant, and  additional
treatment required  by the central plant.

      In this scenario, a 1.6-mgd  conventional treatment plant serving about
10,000 people was assumed to exist nearby, providing water  at a delivered
cost  of about $1.70/1000 gallons.  In order to provide water from  the
central system to the homes using private wells, additional distribution
pipes  and components are required.  Costs associated with the additional
distribution system needed for hookup were allocated to the homes  involved
for evaluating  trade-offs between central water  supply  versus POE systems.
A combination of  6-inch and 8-inch ductile-iron pipes, fittings, and valves
were  assumed, with  installed costs amortized  at 8% over 30  years.   The  POE
cost  was  assumed  to be $900/year/household, reflecting a complete  carbon
replacement  frequency of  1 year.
                                     149

-------
     Figure 1 compares annual household costs of POE systems to those of
central water supply as a function of the total length of pipe and com-
ponents required for connection of the specific number of homes involved, as
indicated by each curve.  For example, if 10 homes are involved, central
water supply appears to be more economical when the total length of pipe
required for all connections is less than 2000 ft.  For 10 homes, if more
than 3000 ft. of distribution connection is needed, then POE systems may be
cost-effective  (>$1200/year for central supply versus $900/year for POE).
Near the break-even points, i.e., about 2000 ft. of pipe for 10 homes, site
specific issues will dictate the cost trade-offs.  For example, if the
central plant size is 6 mgd rather than 1.6 mgd, then the household cost for
central supply may be reduced by $20 to $30/year.  However, if GAC treatment
must be added to the conventional central plant then household costs may
increase by $40 to $50/year.  Also, if the new distribution components
require additional pumping then POE may look more economically attractive.
In the scenario given in Figure 1, the general cost trade-offs indicate that
if the average length of service lines required for connection to central
supply per house are significantly greater than 200 feet then POE may be a
cost-effective alternative.^)

CONCLUSIONS

     Given the analyses presented, when evaluating treatment options and
alternatives for small systems and private homeowners, decision-makers will
have to consider the potentially very high costs such as those of (1) pipe
installation, repair, rehabilitation, or replacement and (2) long-term
central treatment and maintenance versus POE maintenance and monitoring.  In
either case, some type of water quality district, water company, or
maintenance contract would have to be created to satisfy the federal
regulations.
                                    150

-------
VN. °
\ uj
\ O
\
'

--^.
" -..^
""~t"^*-- """"""'"•--^
'"••-. ""^X.
""""-•.. \
^^ ^^ * • ^^w
^"^ \
--, ^\""x
^x^ \
•» ^ \
N N
N
S
X
N
S
N
V) O N
% o in o m o o
3 £- T- CM CM U> f-
o ...
.c i ; ! i
•»; ! Ill
o
z
i — i — i — i — i — i — l — i — 1 — i 	





\
\
\
\
\ \ •
\ \ :
» ,
\ \ -
\ \ x '
•••X \ \ -


"
iVN\\\
s X ••\ • \
x '•• \ \
\ \ -A 1
\ \ -. » \ i
\VA\I
, \,V\\I


z
LU
00
^J» oo
vl' 01
UJ
•«• ^3
*^^ ~pr
O  O
^r <->
DL
*- ^
0 =*
£ to
O) o
C .^
°? 0
_J 0.
"
CO
0 ^
^
C9
U.


O
0
^^
oooooooooo
§gg§S§§S2S
CM   T-   T-   T-   T-   T-
                               151

-------
REFERENCES

 1.  Panel Discussion.  Home Water Treatment:  Is It Feasible?. Journal of
     the American Water Works Association. 78(10): 20-31, 1987.

 2.  National Primary Drinking Water Regulations:  Volatile Synthetic
     Organic Chemicals.  Federal Register.  50(219): 46880-46932, 1985.

 3.  National Primary Drinking Water Regulations:  Synthetic Organic
     Chemicals; Monitoring For Unregulated Contaminants.  Federal Register.
     52(130): 25690-25717, 1987.

 4.  Chambers, C. D. and Janszen, T. A.  Point-of-Entry Drinking Water
     Treatment Systems For Superfund Applications.  EPA/600/S2-89/027, Risk
     Reduction Engineering Laboratory, Cincinnati, Ohio, Project Officer:
     Mary K. Stinson, 1990.

 5.  Goodrich, J. A., Stevens, T., and Walsh, C. C.  Small Systems Meet
     Superfund Challenge with Point-Of-Entry Treatment Units.  In:  (
     Proceedings of Hazardous Materials Control/Superfund '91.  Washington,
     DC, 1991. p. 283.

6.   Bianchin, S. L.  Point-of-Use and Point-of-Entry Treatment Devices Used
     at Superfund Sites to Remediate Contaminated Drinking Water.  In:
     Proceedings of Conference on Point-of-Use Treatment of Drinking Water,
     Cincinnati, Ohio, Water Engineering Research Laboratory, Cincinnati,
     Ohio, October 6-8, 1987, EPA/600/9-88/012, 1988.
     Goodrich, J. A., Adams, J. Q., Lykins, Jr., B. W., and Clark, R. M.
     Safe Drinking Water from Small Systems:  Treatment Options.  Journal
     American Water Works Association.  84(5): 49-55, 1992.
of
                                    152

-------
              THE REMOVAL OF IONIC CONTAMINANTS FROM DRINKING WATER

                                      Thomas F. Speth
                                   Environmental Engineer
                              Drinking Water Research Division
                            U.S. Environmental Protection Agency
                                 26 W. Martin Luther King Dr.
                                    Cincinnati, OH 45268
                                       513-569-7208

INTRODUCTION

    Carbon adsorption is a treatment technique that can remove a broad-spectrum of contaminants
from water.  When designing a treatment unit, design decisions are often made without regard to the
target compound's chemical characteristics.  One characteristic that has been known to affect
carbon adsorption is ionic character. Muller et al. (1,2) has to date completed the most definitive
attempt to model ionic species in carbon adsorption systems.  Others (3,4) have published adsorption
data on ionic compounds.  The most worthwhile conclusion of the above work is that pH can  have a
substantial effect on carbon adsorption capacity.  Generally,  positively charged ions will have higher
capacities at higher pHs, and negatively charged ions will have higher capacities at lower pHs.
Complicating adsorption calculations are the charge speciation of compounds at different pHs. All
these factors must be understood before GAC adsorber designs are made.

    The  USEPA has the responsibility of regulating contaminants in drinking water through the Safe
Drinking Water Act. Currently 55 organic compounds are being regulated. Regulated compounds
are typically broken up into two groups: volatiles and pesticides.  Volatile compounds are nonionic.
Pesticides, however, have a wide range of chemical characteristics depending on their purpose.
Two compounds that are being regulated have been found by the USEPA to have unique treatment
characteristics.  These are diquat and glyphosate. Diquat (1,1-ethylene-2,2-bipyridyiium dibromide)
is an ionic herbicide that is used as a non-crop weed killer, a general aquatic herbicide, a pre-
harvest top killer, or a desiccant of seed crops. Diquat inhibits the electron transfer in both the
cytochrome chain of the mitochondria and the electron transfer chain of chloroplasts. Diquat  has a
net plus two charge at neutral pH.  Its pKas are estimated at  11.7 and 12.2.  It is soluble in water and
is nonvolatile.  Its maximum contaminant limit is currently set for 20 ppb.

    Glyphosate is a non-selective, broad-spectrum, post-emergent herbicide. It is the only regulated
organic compound that has oxidation designated as the "Best Available Technology" for treatment.
Glyphosate is zwitterionic (5) with pKa values of <2.0, 2.6, 5.6, and  10.6. At high pHs, it has a -3
charge. At low pHs, it has a +1 charge. Its amphoteric character is determined by the phosphonate,
carboxyl, and amine groups.  Glyphosate is nonvolatile. Glyphosate's herbicidal activity comes from
its inhibition of the 5-enolpyruvylshikimic acid-3-phosphate synthase which normally leads to the
biosynthesis of aromatic amino acids (6). Glyphosate's maximum contaminant limit is currently set
for 700 ppb.

    Both diquat and glyphosate have been shown  to adsorb to clays (5,7). However, neither diquat
nor glyphosate has shown the ability to hydrophobically complex with natural organic matter.  This
presentation will look at how the adsorption of diquat and glyphosate changes with pH and the
presence of nonionic compounds. Also, other treatment alternatives will be presented for
glyphosate.
                                            153

-------
METHODOLOGY

    Carbon adsorption studies were completed with F-400 carbon (Calgon Corp., Pittsburgh, PA)
and 250 ml PTFE-capped glass bottles.  The isotherms were equilibrated in an end-over-end tumbler
for three weeks.  The procedure is identical to that published earlier (8). Distilled-water isotherms for
diquat were run at pHs of 2.0, 3.0, 4.0, 5.5, 6.1, 6.7, 8.0, 8,9, 9.5, and 10.1.  The isotherms
completed at pH 6.1, 6.7, and 8.0 utilized a 0.001 M phosphate buffer. The isotherms completed at
8.9, 9.5, and 10.1  utilized a 0.001 M boric acid buffer. At low pHs, no buffer was used because of the
desire to avoid organic buffers that could affect the isotherm results.

    The glyphosate jar tests were run to simulate a full-scale coagulation, flocculation, and
sedimentation treatment  train. Standard jar-test equipment and procedures were employed (9).
Alum was used as the primary coagulant. The alum doses were 1.0, 2.0, 5.2, 10.4, 15.6, 20.9,
31.3,and 41.7 mg/L. No  coagulant aids or PAC were used. All the jar-test studies were run  at the
same mixing conditions.  The rapid-mixing speed was set for 100 rpm for 2 minutes. The subsequent
flocculation and sedimentation mixing speeds were: 30 rpm for 10 minutes, 20 rpm for 10 minutes,
10 rpm for 10 minutes, and 0 rpm for 60 minutes.  Turbidity, temperature, pH, and contaminant
samples were taken at the end of sedimentation.                   ,

    Diquat was analyzed with a Hitachi U-2000 spectrophotometer at its absorption maximum of 308
nm.  Cis-1,2-dichloroethene was analyzed with EPA Method 601 for purgeable halocarbons.
Glyphosate was analyzed by a high-performance liquid-chromatography method with post-column
derivatization.  All materials were either glass, stainless steel, or Teflon.

RESULTS AND DISCUSSIONS

    Figure 1  displays diquat isotherms over a wide pH range.  At  high pHs, the carbon's capacity
for diquat is very high. With decreasing pH, the capacity drops very sharply.  The isotherms at pH
2.0, 3.0, and 4.0 show a  negative slope which is not possible. The negative slopes are attributed to
either experimental scatter or competitive effects between diquat and hydrogen ions. Because each
isotherm has data that covers a liquid-phase concentration of 400  ng/L, the capacities can be
compared at this concentration.  As can be seen in Figure 1, the capacity drops from 286,000 ng/g
at a pH of 9.5 to 9.72 jig/g at a pH of 2.0. The capacity at a pH of 9.5, therefore, is more than
29,000 times more than that at a pH of 2.0. This difference is much greater than that found for other
compounds studied (1,2,4).

    The maximum adsorption cajaacity corresponds to a compound's pKa (4). Because the pKa of
diquat is estimated at 11.7 and 12.2, the highest capacity would be expected to be at a pH  of 10.1.
However, because of the scatter in the data for the isotherms at pHs 8.9, 9.5,  and 10.1, it is  difficult
to determine if this is true for diquat.

    Figure 2 shows a bisolute cis-1,2-dichloroethene (DCE)/diquat isotherm completed at a pH of
10.1.  Also shown are the Ideal Adsorbed Solution Theory  (IAST) predictions for diquat and  DCE
(10).  The initial diquat concentration was 3.84 mg/L and the initial DCE concentration was 0.982
mg/L. As can be seen, there was little reduction in diquat  capacity due to competitive adsorption
from DCE. This was expected because diquat is more strongly adsorbed than DCE at this pH.  The
IAST prediction confirmed these results. A large reduction in capacity for DCE was expected
because of diquat's strength  of adsorption. This is shown by the IAST prediction for DCE.  However,
the actual  reduction in DCE capacity due to competition from the diquat was very small as shown by
the data. One possible explanation is that a certain percentage of the diquat at a pH of 10.1 was in
nonionic form. This nonionic diquat may have competed against DCE for the same adsorption sites,
while the ionic diquat may have adsorbed to  different adsorption sites.  As a check, the initial
                                             154

-------
concentrations of diquat used in the IAST model were reduced to one-tenth of the actual initial
concentrations. The resultant IAST predictions for the three bisolute isotherms were good.  This may
indicate that approximately ten percent of the diquat is in nonionic form which can compete with
DCE for adsorption sites.

    The theory that diquat is approximately ten percent nonionic at a pH of 10.1  implies several
things.  First of all, to achieve the single-solute diquat capacities shown in Figure  1, and the bisolute
diquat capacities shown in Figure 2, ionic diquat would have to be strongly adsorbed.  This is
because the adsorption capacities shown in Figures 1 and 2 are greater than ten percent of the initial
diquat solution. The ionic diquat, therefore, must compete for different adsorption sites than DCE.
The adsorption of diquat at these sites must be highly dependent on pH as indicated by the data at
lower pHs.

    Another explanation for the bisolute adsorption data is that a small percentage of the adsorption
sites attract both ionic diquat and DCE, but most of the adsorption  sites selectively adsorb diquat or
DCE. This could occur in addition to the previous explanation that  a certain percentage of the diquat
is in nonionic form.  The selectivity of adsorption sites could be explained by micropore adsorption in
which the micropores are accessible to DCE,  but not to diquat because of its larger size. The
macropores would contain the adsorption sites that are accessible to both  diquat and DCE. The low
capacities at low pHs can be explained by taking the charge of the carbon into account in the  same
fashion  as explained earlier.  The diquat can be repelled from the carbon due to the net positive
charge of the carbon surface.

    As  previously mentioned, glyphosate has oxidation designated "Best Available Technology" for
treatment. This was based on work that showed that glyphosate's  capacity in Ohio River water was
much different than what was predicted (11).  The ionic speciation of glyphosate was most  likely
responsible. The resultant insecurity in adsorption capacity, and the demonstrated ability of oxidants
such as chlorine and ozone to destroy glyphosate precipitated the decision to name oxidation  as
"best available technology".  It must be stated that the oxidation byproducts of glyphosate are not
known.

     Other technologies also.showed an ability to remove glyphosate.  An interesting relationship of
glyphosate removal with turbidity removal was discovered from jar-test data. As turbidity was
removed below 2 NTU, glyphosate began to be removed.  This suggests that the glyphosate removal
required the presence of organic matter that was associated with the turbid matter.  Therefore, it is
postulated that the glyphosate attaches to very small floe particles that  are only removed when
coagulation  conditions are optimal for turbidity removal.  The ability of alum coagulation to remove
glyphosate is entirely consistent with the work of Sprankle et al. (5) who have suggested that
glyphosate forms organic/metal/glyphosate complexes.

CONCLUSIONS

     The ionic herbicide diquat was shown to have adsorption  capacities that vary greatly with pH.
At a pH of ,9.5, diquat had more than 29,000 times its capacity than at a pH of 2.0. Bisolute
isotherms with diquat and cis-1,2-dichloroethene showed that Ideal Adsorbed Solution  Theory did not
accurately predict multicomponent adsorption.  Adjustments that attributed the reduction in cis-1,2-
dichloroethene capacity to the presence of a  small amount of nonionic  diquat made up for the
discrepancies.  The comparison of these results to diquat's single-solute capacities indicated that
ionic diquat was strongly adsorbed to adsorption sites that did not  attract cis-1,2-dichloroethene.
The Ideal Adsorbed Solution, Theory prediction discrepancies also could be accounted for by
assuming that only a small percentage of the adsorption sites could adsorb both ionic diquat and
cis-1,2-dichloroethene.  In this case, the selective adsorption may have been associated with
                                              155

-------
adsorption sites located in micropores that were accessible to cis-1,2-dichloroethene, but not to
diquat, because of diquat's larger size.

    The ionic herbicide glyphosate was found to be easily oxidized with chlorine and ozone. The
oxidation byproducts, however, are not known. The glyphosate jar-test studies found that as
turbidity was removed, so was glyphosate. The majority of the glyphosate removal happened as the
turbidity was reduced below 2 NTU. Alum was necessary for glyphosate removal in the jar-test
studies.

    This report has not been subject to agency review and therefore does not necessarily reflect
USEPA policy.  Mention of trade names of commercial products does not constitute endorsement or
recommendation for use by the USEPA.

REFERENCES

1.  Muller G., Radke C.J., and Prausnitz J.M. Adsorption of Weak Organic Electrolytes From Dilute
    Aqueous Solution onto Activated Carbon - Part I. Single-Solute Systems. J. Colloid Interface
    Sci. 103, 2, 466-483, 1985a.

2.  Muller G., Radke C.J., and Prausnitz J.M.  Adsorption of Weak Organic Electrolytes From
    Dilute Aqueous Solution onto Activated Carbon - Part II. Multisolute Systems. J. Colloid
    Interface Sci.  103, 2, 484-492, 1985b.

3.  Rosene M.R.  and Manes M. Application of the Polanyi Adsorption Potential Theory to
    Adsorption from Solution on Activated Carbon -10. J. Phys. Chem. 81,1651-1657,1977.

4.  Ward T.M. and Getzen F.W. Influence of pH on the Adsorption of Aromatic Acids on Activated
    Carbon.  Envir. Sci. Technol. 4,1, 64-67,1970.

5.  Sprankle, P.,  Meggitt, W.F., &  Penner, D. Adsorption, Mobility, and Microbial Degradation of
    Glyphosate in the Soil, Weed Science, 23:229-234,1975.

6.  Malik, J., Barry, G., & Kishore, G.  The Herbicide Glyphosate, BioFactors, 2:1:17-25,1989.

7.  Weber J.B., Perry P.W., and Upchurch R.P. The Influence of Temperature and Time on the
    Adsorption of Paraquat,  Diquat, 2,4-D, and Prometone by Clays, Charcoal, and an Anion-
    Exchange Resin.  Soil Sci. Soc. Proc. 29,1965. p. 678-688.

8.  Speth T.F. and Miltner R.J. Adsorption Capacity of GAG for Synthetic Organics. Jour. AWWA,
    82(2), 72-75,  1990.

9.  Miltner, R.J., Baker, D.B., Speth, T.F., & Fronk C.A. Treatment of Seasonal Pesticides in
    Surface Waters, Jour. AWWA, 81:1:43-52,1989.

10. Myers, A.L & Prausnitz, J.M. Thermodynamics of Mixed-Gas Adsorption, Am. Institute Chem.
    Eng. J. 11:121-127,1965.

11. Speth, T.F. The  Removal of Glyphosate From Drinking Water. Accepted for publication in J.
    ASCE - Envir.  Eng., December, 1992.
                                             156

-------
w
 6
 8
 o
1
 o
CO
     10000
      1000 -
       100 -
         1-
       0.01
      o.ooi
pH=lO.l
pH=9.5
pH=8.9
pH=8.0
pH=6.7
pH=6.1
pH=5.5
pH=4.0
pH=3.0
pH=2.0
         0.001
                          0.01
                          0.1
                                                                           10
                       Liquid-Phase Cone (mg/L)
                Figure 1. Diquat isotherms completed in distilled/deionized water at various pHs.
      100O
   o
 W
 a
 o
 u
       100 r
        10-
 43
  i
        0.1
             Diquat Single-Solute Isotherm
                                                     • Diquat Blsohtte Data
                                                    • D Diquat IAST Prediction
                                                     • DCE Bisolute Data
                                                     O DCE IAST Prediction
             DCE Single-Solute Isotherm
            Initial DCE Conc.=0.982 mg/L
            Initial Diquat Cone.=3.84 mg/L
            DistiUed/Deionized Water, pH=10.1
            F-400 Carbon
          O.O1
                                0.1
                                                                           10
                       Liquid-Phase  Cone  (mg/L)

           Figure 2. Comparison of EAST predictions to bisolute data for diquat and cis-l,2-dichloroethene.
                                    157

-------
                         A CLOSED LOOP APPROACH FOR REMOVAL
                     AND DESTRUCTION OF CHLORINATED AROMATICS
                                         FROM SOIL

                Shubhender Kapila, M.H. Liu, K.R. Ryoo, R.K. Puri, and A.A. Elseewi1
                        Center for Environmental Science and Technology
                                    Department of Chemistry
                                  University of Missouri - Rolla
                                     Roila, Missouri  65401
                               (314) 341-6187 or (314) 341-6162
                   1 Environmental Division, Southern California Edison Company
                               P.O. Box 800, 2224 Walnut Grove
                                  Rosemead, California  91770

INTRODUCTION

       Remediation of contaminated soil is a challenging task especially when contaminants are
recalcitrant chemicals. Chlorinated aromatics as a class are some of the most toxic and persistent
xenobiotic contaminants. Members of this class include chemicals such as polychlorinated dibenzo-p-
dioxins (dioxins), polychlorinated biphenyls (PCBs), chlorinated phenols and  chlorinated pesticides.
These chemicals are considered synonymous with environmental contamination in the popular press.
The only widely accepted technology for destruction of these compounds in soil is high temperature
incineration. This process, while highly efficient, is very expensive and alternate approaches for
decontamination need to be researched and implemented. The report presents preliminary results for an
alternative approach.  The approach combines supercritical fluid extraction with carbon adsorption and
oxidative regeneration. This combination permits reuse of solvent and adsorbent with minimum external
energy input and essentially zero toxic discharge.

       Supercritical fluid extraction (SFE) is receiving increasing attention as a decontamination
technique.  It has been demonstrated that high desorption efficiency can be obtained with SFE for a
number of non-polar or moderately polar contaminants.  The major advantages of SFE lie in rapid
equilibration and ease with which the contaminants can be separated from the critical fluids, thus allowing
the reuse of fluids. The simplest means for this separation is to decrease the density of the fluid by
lowering the pressure. This approach, while simple, is unsatisfactory for large scale decontamination due
to the substantial energy input required for fluid recompression. Furthermore, a separate detoxification
step is necessary for ultimate disposal of the contaminants.  The energy cost of decompression can be
reduced by use of activated carbon.  The use of cartoon can be made economical through an oxidative
carbon regeneration process. The objective of the present study was to explore applicability of this
integrated approach.

METHODOLOGY

       The study involved the following set of experiments:

               1. Partition Experiments

               2. Breakthrough Experiments

              3.  Carbon Regeneration Experiments

              4. Bench Scale SFE-COR System

1. Partition Experiments

       The objective of these experiments was to optimize extraction parameters.  The experiments
were carried out with a static SFE system which was interfaced directly to a Gas  Chromatograph (GC) or a
                                               158

-------
Liquid Chromatograph (LC). The details of the system have been provided in earlier reports (1-4). All
experiments were carried out with soil fortified with selected contaminants. The contaminant
concentrations were varied over a 1 to 500 ppm range.  Extraction parameters such as fluid density,
pressure, temperature and equilibration period were optimized. The effects of matrix parameters, i.e.,
moisture content, surface area, texture and organic matter content were also examined.

2. Breakthrough Experiments

        One of the critical questions in the reuse of the fluid is related to the adsorption efficiency of
Granular Activated Carbon (GAG).  To ascertain this, a serial adsorbent trap configuration was used.
Known concentrations of contaminants in critical fluid were obtained by dispersing contaminant on glass
beads and placing beads in the critical fluid container. After a set equilibration period,  contaminants in
critical fluid were passed through the adsorbent trap. A mass balance approach was used to determine
breakthrough and adsorption efficiencies.

3. Carbon Regeneration Experiments

        The oxidative regeneration was used for destruction of adsorbed contaminants. This process
termed COR has been investigated in our laboratory. The details of these studies have been reported
earlier (5).

4.  Bench Scale SFE-COR system

        An integrated bench scale system is being fabricated in the laboratory. The system is designed
for operation in the static as well as the dynamic modes. It consists of an extractor, a trap and a recycling
pump. The liquified fluid is pumped through a heating extraction coil where its temperature is brought
above the critical temperature, supercritical fluid is then made to pass through the matrix held in the
extraction tube housed in a thermostatic bath. The fluid, along with extracted contaminants, passes
through the GAC trap. The contaminants are adsorbed on the carbon and the "clean" fluid is liquified for
reuse.

RESULTS AND CONCLUSIONS

        Partition experiments revealed that removal efficiencies ranging between 75-99 % can be readily
obtained. The optimal extraction parameters for carbon dioxide and nitrous oxide were found to be in the
near critical region. Highest extraction efficiencies were obtained with  addition of  polar modifiers.
Methanol was the modifier of choice; its introduction even at relatively  low levels (1-2%) improved
extraction efficiencies to >95%. This improvement is related to wetability of the surface and weakening of
the electrostatic interaction between contaminant molecules and polyphenoiic components of soil organic
matter.

        Breakthrough studies with  pentachlorophenol and dioxins as model compounds showed that all
compounds extracted with supercritical fluid were adsorbed on GAC in the first trap. Complete
destraction, i.e., z_ 99.99%, was obtained through the regeneration process with consumption of 
-------
       4. Liu, M.H., Kapila, S., Nam, K.S., and Elseewi, A.A., J. Chromatoar.. in press

       5. Cody, C.C., Kapila, S., Manihan, S.E., Larsen, B.W., and Yanders, A.F., Chemosphere, 20 (10-
12): 1959-1966, 1990.

FOR MORE INFORMATION:

       Dr Shubhender Kapila, Center for Environmental Science and Technology, Department of
Chemistry, University of Missouri - Rolla, Rolla, Missouri 65401, (314) 341-6187 or (314) 341-6162.
                                             160

-------
   FEASIBILITY OF  REMOVING  URANYL. THORIUM AND  RADIUM  FROM KAQLINITE
                                   BY  ELECTROKINETICS

                                         Yaicin B. Acar
                                  Civil Engineering Department
                                   Louisiana State University                     :
                                    Baton Rouge, LA 70803
                                    Phone:  (504) 388-8638            ,   ;,

                                 Robert J. Gale and Alberto Ugaz
                                    Department of Chemistry
                                   Louisiana State University
                                    Baton Rouge, LA 70803
                                    Phone:  (504)388-3010

                                        Robert E. Marks
                                    ELECTROKINETICS Inc.
                         The Louisiana Business and Technology Center
                              Suite 102. LSU, South Stadium Drive
                                    Baton Rouge, LA 70803
                                    Phone:  (504) 388-3992
INTRODUCTION
   Electrokinetic Soil Processing is an emerging technology in remediation of soils contaminated by
inorganic and certain organic species (1-11). The method uses DC currents in the order of m A/cm 2 of
electrode area to remove/separate inorganic/organic contaminants from soils. It is envisioned that the
technique will find different applications in construction of barriers opposing advective-dispersive
transport of contaminants in clay liners, diversion schemes for waste plumes, and for injection of grouts,
microorganisms and nutrients into subsoil strata . The feasibility of the technique and its cost-efficiency  in
removing different chemical species from fine-grained soils prompted the need to investigate the
feasibility of removing radionuclides  (12).

  Radioactive contamination is a particularly serious problem facing the nation and the world. In the United
States alone, there are 33 radioactively contaminated sites listed or proposed for listing on the National
Priorities List (NPL). Contamination in most of these sites was caused by uranium mining and milling,
commercial radium industry, or Federal nuclear weapons research and production program (12,13).
Furthermore, the Department of Energy (DOE) weapons research, development, and production mission
generated  large quantities of radioactive and mixed wastes during the past forty years. A general
description of the sites listed on the NPL and other sites resulting from DOE mission are discussed by
Acaretal. (12).

  Only excavation and land encapsulation have been used to remediate radiologically contaminated
sites. The rapid use of disposal space and rising cost of land disposal prompt the need to find new,
innovative remediation technologies. Therefore, recent  remediation techniques for radioactive materials
have focused on separation/concentration of the radioactivity from the innocuous material and
containment/stabilization of the radioactive matrix. EPA has initiated the Volume Reduction and Chemical
Extraction  (VORCE) project to investigate technologies that can reduce the volume of soil contaminated
with radioactivity at Superfund sites. The use of electrical currents to remediate radionuclide contaminated
soils (Electrokinetic Soil Processing) has been included for investigation as a potential VORCE technology
                                              161

-------
  The objectives of this cooperative agreement between Electrokinetics Inc. (EK) and the USEPA Office
of Research and Development are to investigate the feasibility of removing radionuclides from soils by
application of electrical currents in Phase I of the project and to conduct a pilot scale study of electrokinetic
remediation process in Phase II. This cooperative agreement supported under the SITE-03 program was
initiated in September 1,1990 and Phase I continued until March, 1992. Currently, Phase II of the
program is on-going. Most sites are partially saturated and in these sites, the process will be used by
supplying an anode fluid.  It is found necessary to validate the hypothesis that the process will still be
efficient in moving the fluid   in the anode compartment across to the cathode when the soil is partially
saturated. The specific objectives of Phase I were to investigate if the electrokinetic soil processing can
be used in soils with partially saturated conditions and to assess removal efficiency of uranium (U), thorium
(Th), and radium (Ra) from saturated kaolinite specimens by the electrokinetic soil processing technique.
In accomplishment of the  objectives of Phase I, this paper presents a summary of the findings of the
bench-scale laboratory test conducted for radionuclide  removal.


METHODOLOGY

  One-dimensional laboratory tests are planned to evaluate the removal of uranium, thorium, and radium.
One-dimensional studies prevent introduction of geometrical considerations, and therefore, such tests
simplify assessment of the feasibility and efficiency of removal. The concentrations of the radionuclides
are taken to be representative of average field concentrations reported for radionuclides.  EPA/520/189-
004 and EPA/540/288-002 indicate that 1,000 pCi/g of  uranyl and thorium are typical of mean
concentrations for field conditions.  Kaolinite was used for the preparation of laboratory specimens as this
mineral has an adsorption capacity and a low permeability typical of fine-grained  deposits.  The electro-
osmotic flow efficiency is high in this mineral.  Kaolinite  radionuclide mixture will be compacted or
consolidated and tested in one-dimensional flow conditions.

  The testing program can be divided into two groups; saturation tests and radionuclide removal tests.
Uranium, thorium and radium were selected in bench-scale tests conducted for radionuclide removal. As
the process is better understood,' new studies are initiated at Electrokinetics Inc. EK to investigate other
aspects; such as investigation methods to enhance removal and development  of cost-effective
electrodes. The testing program, characteristics of the test materials, the details of the  equipment, and
the procedures employed in specimen preparation and testing in the saturation and radionuclide removal
studies in this Phase  I study are presented in EK-BR-009-0292 (14).
 RESULTS
   Saturation and fabric changes are not expected to  significantly influence the flow conditions in one-
 dimensional electrokinetic soil processing. The results indicate that the pore fluid supplied at the anode
 section, in partially saturated soils and under one-dimensional conditions, will flush across the specimen
 under electrical gradients resulting in flow conditions,  efficiency and pH profiles similar to those obtained
 in fully saturated specimens. Compacted specimens can be used in laboratory studies of electrokinetic
 soil processing. Partially saturated conditions in such specimens will not influence the final steady state
 flow conditions in the tests with contaminants. Even when an electrode which permits egress and ingress
 of water is used in electrokinetic soil processing, settlements are expected in fine-grained deposits due
 to consolidation of the soil surrounding the anode. The lower the hydraulic conductivity of the deposit,
 the more pronounced will be the effect of this consolidation. Recent studies (15) investigating the
 feasibility of using the technique in unsaturated sands reconfirm the results of this study.
                                                  162

-------
  Uranium removal tests at 1,000 pCi/g of activity demonstrated that the process efficiently removed
uranium from Georgia kaolinite (Figure 1). Removal efficiency decreased from the anode towards the
cathode, due to the increase in pH. A yellow uranium hydroxide precipitate was encountered in sections
close to the cathode and on the cathode. Formation of this precipitate increases the resistivity , the
voltage gradients and energy expenditure.  In response to the need to  develop methods to remove this
precipitate in soil sections close to the cathode and in order to increase average removal  rates, tests are
conducted by introducing low levels of acetic acid in the cathode compartment to neutralize the hydroxyl
production and allow migration of the uranyl ions into the cathode compartment.  Acetic acid introduction
at the cathode successfully  prevented the precipitation encountered close to  the cathode and 95% of
uranium was flushed out of the specimen.

  Thorium  removal tests at 50-300 pCi/g of activity have demonstrated that thorium is only removed by
migration from the leading anode sections of the cell.  Unlike the case for uranium, no deposits were
found on the cathode itself. These results may be due to the high ionic charge on the Th+4 ion, which
enhances its adsorption and makes desorption by hydrogen ion a more difficult task.  Better
electroosmotic flows can be achieved by acidifying the soil section adjacent to the cathode, but thorium
removal efficiency was still  very poor. Complex species that can reduce the net ionic charge on the
thorium species are necessary in order to efficiently remove thorium by electrokinetic processing.

  Radium tests at 1000 pCi/g of activity have demonstrated minimal removal across the test cells.  The
only rational explanation is  the extremely low solubility of the precipitated radium sulfate species.  Similar
to the results for thorium, we anticipate that a complex species has to be found  to avoid this insolubility
problem.
CONCLUSIONS
  It is established that electrokinetic process will efficiently move the anode fluids across  partially
saturated specimens. Uranium removal tests at 1,000 pCi/g of activity demonstrated that the process
efficiently removed uranium from Georgia kaoiinite. The use of very low concentration acetic acid at the
cathode compartment to neutralize the formation of the base  by the electrolysis reactions successfully
avoided the precipitation problem in removal of the uranyl ion and a 95% removal efficiency was achieved
for the uranyl ion at 1,000 pCi/g of Georgia kaolinite.

  Thorium removal tests have demonstrated that thorium is only removed by migration from the leading
anode sections of the cell. Complex species that can reduce the net ionic charge on the thorium species
are necessary in order to efficiently remove thorium by electrokinetic processing.

   Radium tests at 1,000 pCi/g of activity have demonstrated minimal removal across the test cells.  The
only rational explanation is the extremely low solubility of the precipitated radium suifate species. Similar
to the results for thorium, we anticipate that a complex species has to be found to avoid this insolubility
problem.

  The tests further demonstrate that the uranium, thorium and radium species have distinct and different
chemical properties and that for the latter two contaminants, a simple electrochemical treatment without
additional pre-, concurrent-, or post- chemical treatment will not be satisfactory. Radium, even in extremely
small concentrations, does not appear to exist as a mobile, charged ionic species, and it is likely present
as its  extremely insoluble, nonionized sulfate. It may be possible to dissolve this  species by complexation,
if a suitable complexant that functions at relatively high acid strengths (pH<2) can be found. Thorium
which is present as a highly charged ion at low pH, Th+4. would require 4 H+ ions to desorb it from the
surface  sites on the clay and it readily forms gelatinous, nonconducting hydroxides in the vicinity of the
                                                 163

-------
                                1021 pCi/g
                                1005 pCi/g
                                979 pCi/g
                                955 pCi/g
                                929 pCi/g
                                1428 pCi/g
                                0.2       0.4       0.6        0.8

                                 NORMALIZED DISTANCE FROM ANODE
         Figure 1 . Removal Efficiency for Uranyl Ion in Unenhanced Electrokinetic Remediation
                           (Open Symbols are for Shorter Duration Tests)

cathode  Such hydroxides can plug the pores in the soil deposit, increase the resistance of the system to
current flow as well as diminish the ionic migration and electroosmotic advection. It is essential to use
species which will improve the solubility of species, decrease the overall charge and decrease the
 adsorptive strength of thorium.

  In general for a mobile cationic species such as Pfa2+ ion or UO22+ ion, we have found that 5 to 10% of
the contaminant may remain in the 10% section of a soil that is adjacent to the cathode. The PH patterns
 ndiSte that the ions are precipitated as hydroxides in this region. A simple and more complete way to
achteve removal would be to add a dilute mineral acid to the effluent at the cathode upon completion of
the process. This prospect of post-chemical treatment coupled with a re-electrolysis proved to be
successful in removing the uranyl ion at a 95% efficiency across the soil.
REFERENCES
 1  Acar Y B Gale  R J., and  Putnam, G. Electrochemical Processing of Soils: Theory of pH Gradient
 Development by Diffusion and Liner Convection. .Innmgl of Fnvirnnmentai Science anj Health, A25 (6).
 687-714, 1990.

 2 Acar Y B Named, J., Gale, R. J., and Putnam, G. Acid/Base Distribution in Electro-Osmosis. Bulletin
 ni the Transportation Research. Soils Geology and Foundations, Geotechnical Engineering 1990 ,
 Record  No. 1288:23-34, 1991.

 3 Acar Y B   Hamed  J. Electrokinetic Soil Processing in Remediation/Treatment; Synthesis of Available
 n"atn Rj.iiPtin of the Transportation Research. Energy and Environmental Issues 1991, Record No. 1312:
 152-161, 1992.

 4 Acar Y B  Li  H., Gale R. J. Phenol Removal from Kaolinite by Electrokinetics. .Innrngl of GeofechnicaJ
 p'pr,infl'ftrinn."ASCE ,118 (11):1 837-1 852, 1992.
                                                164

-------
5. Acar, Y. B., Alshawabkeh, A., Gale, R. J. Fundamentals of Extracting Species from Soils by
Electrokinetics. Waste Management, Pergamon Press, London, 17 (3):     , 1993.

6. Acar, Y.B., Alshawabkeh, A.  Modeling Conduction Phenomena in Soils Under an Electric Current. 1m
Proceedings of the XIII. International Conference on Soil Mechanics and Foundation Engineering,
Balkema Publishers, Rotterdan, Netherlands, January 1994.

7. Alshawabkeh, A.,  Acar, Y.B.  Removal of Contaminants from Soils by Electrokinetics: A Theoretical
Treatise. Journal of Environmental Science and Health. A27 (7):1835-1861,,1992.

8. Banerjee, S., Horng, J., Ferguson, J. F., and Nelson, P. O. (1990), "Field Scale Feasibility of
Eiectrokinetic Remediation," Unpublished  Report Presented to USEPA, Land Pollution Control Division,
RREL.CR 811762-01, 122 p.

9. Hamed, J., Acar, Y. B., and Gale, R. J. Pb(ll) Removal from Kaolinite Using Electrokinetics. Journal of
Geotechnical Engineering. ASCE, 112(2):241-271, 1991.

10. Lageman, R., Pool, W., and Seffinga, G. Electro-Reclamation Theory and Practice. Chemistry and
Industry. Society of Chemical Industry, London, (9):241-271, 1991.

11. S;hapiro, A. P., Renauld, P., and Probstein, R. Preliminary Studies on the Removal of Chemical
Species from Saturated Porous Media by Electro-osmosis. Physicochemical Hydrodynamics. 11 (5/6):
785-802, 1989.

12. Acar, Y. B., Gale  R. J., Leonard, C. QAPP-Phase I; Project EKSITE. Quality Assurance Project Plan
submitted to Office of Research and  Development,  RREL, Electrokinetics Inc., Baton Rouge, Louisiana,
EK-BR-002-0390, 1990,  126 p.

13. Frankena, Frederick and Joann Koelln . Radioactive Waste as a Social and Political Issue: A
Bibliography. Series-AMS Studies in  Modern Society; Political and Social Issues; 21, AMS Press, New
York.

14. Acar, Y. B., Gale, R. J., Ugaz, A., Puppala, S., Leonard, C. Feasibility of Removing Uranium, Thorium
and Radium from Kaolinite by Electrochemical Soil Processing. EK-BR-009-0292, Final Report-Phase  I of
EK-EPA Cooperative Agreement CR816828-01-0,  Electrokinetics Inc., Baton Rouge,  Louisiana, 1992,
243 p.

15.Lindgren, E. R., Mattson, E. D., Kozak,  M. W. Eiectrokinetic Remediation of Contaminated Soils, im
Proceedings of the Department of Energy Workshop on Electrokinetics, Atlanta, Jan. 22-23, 1992.
FOR  MORE  INFORMATION
Randy A. Parker, Office of Research and Development, Risk Reduction Engineering Laboratory, EPA, 26
West Martin Luther King Drive, Cincinnati, OH 45268. Phone: (513) 569-7620, Fax: (513) 569-7620.
                                               165

-------
                EVALUATION OF THE BIOGENESIS SOIL CLEANING PROCESS
                               FOR ORGANIC CONTAMINANTS

                                       Pinaki Banerjee
                             PRC Environmental Management, Inc.
                                     1921 Rohlwing Road
                                  Rolling Meadows, IL 60008

                                       Annette Gatchett
                             U.S. Environmental Protection Agency
                             Risk Reduction Engineering Laboratory
                             26 West Martin Luther King Boulevard
                                     Cincinnati, OH 45268
       The BioGenesis Enterprises, Inc. (BioGenesis), soil washing technology was demonstrated as
part of the U.S. Environmental Protection Agency's (EPA) Superfund Innovative Technology Evaluation
(SITE) program in November 1992. The demonstration was conducted at a petroleum refinery in
Minnesota. Approximately 2,000 cubic yards of soil at the site were contaminated with crude oil. Total
recoverable petroleum hydrocarbon (TRPH) concentrations were detected at up to 30,000 parts per
million (ppm).

       The BioGenesis Soil Cleaning Process consists of two stages.  In the first stage, contaminants
are transferred from the soil matrix to liquid phases using a proprietary surfactant solution.  The end
products of the first stage include treated soil, contaminated wastewater, and an oil-solvent phase.  In
the second stage, the surfactant solution enhances biodegradation of residual contamination in soil and
wastewater.  The oil-solvent phase is recovered for recycling or disposal. The end products of
microbial degradation are expected to be nontoxic inorganic compounds.

       Three runs, each consisting of 18 cubic yards of soil, were treated at the refinery site over 2
days. Each batch of soil was washed twice with water containing the BioGenesis solution at a
temperature of 180°F.  Mixing time, solution concentration, and mixing intensity were kept at constant
operating conditions.  TRPH concentrations were monitored in treated and contaminated soils, water,
and wastewater.  Results of chemical analyses and field measurements collected during
demonstrations at the refinery site will be evaluated to determine the following: (1) TRPH removal
efficiency in the treatment system, (2) whether or not the treatment system's performance is
reproducible, (3) the extent of biodegradation of TRPH, and (4) treatment costs.
                                             166

-------
          CONTAMINATED
              SOIL
                                    WASH UNIT
                                                    FILTER
                                                  1  UNIT
                                                          EJTLUENT FROM
                                                            WASH  UNIT
                                                                          TO WWTP
                                                            TREATED SOIL
                                                        MAKE-UP WATER
i
5
                       BIOGENESIS  SOIL WASHING PROCESS
                                        167

-------
               POLLUTANT FLUXES TO AQUATIC SYSTEMS VIA
  COUPLED BIOLOGICAL AND PHYSICOCHEMICAL BED-SEDIMENT PROCESSES
          D.D. Reible, L.J. Thibodeaux, K.T. Valsaraj and J.W. Fleeger
                          Louisiana State University
                           Baton Rouge, LA 70803
                              (504) 388-1426
Introduction

      A three year project was proposed to evaluate the coupled effects of
physicochemical and biological processes on contaminant transport from
sediments.  Sediment-dwelling organisms representing two functional groups, one
a sediment stabilizer (a mucus-generating snail, Physella virgata) and the other a
sediment destabilizing  bioturbator (a conveyor-belt feeder, Tubifex tubifex). were
to be introduced to contaminant inoculated sediment in specially designed
laboratory microcosms or cells.  The flux of contaminants to the water column
due to the sediment processing by the organisms was to be contrasted with each
other and with the physicochemical transport processes currently under study.

      In the first year  of the proposed project, the experimental procedure was to
be finalized and experiments focused on Tubifex tubifex. Different densities of the
organism were to be examined as well as the effect of snail addition or removal.
Pyrene had been selected as a model chemical tracer of the particle movement on
the basis of hydrophobicity and low vapor pressure and analytical procedure
development was underway.

Progress

      Initial experiments to develop and test analytical procedures and evaluate the
bioturbation potential of Tubifex tubifex have been conducted. Two complete
experiments have been conducted since initiation of the project in February. In
both experiments, hydrophobic contaminants were added to a freshwater
sediment.  These sediments were then placed in  15 experimental microcosms and
Tubifex tubifex were introduced at densities between 0 and  100 organisms per cell
(75 cm2 surface area). Water containing nutrients was forced through the cells at
flowrates of the order of 200-1000 cm3/day.  During the first experiment, the
sediments were inoculated with pyrene only while during the second experiment,
pyrene, phenanthrene  and dibenzofuran were used as model contaminants.  In  both
                                    168

-------
experiments, contaminant inoculation was conducted by plating a known quantity
of solid contaminant onto the walls of a jar and tumbling for 2-3 weeks with damp
sediment. Sediment samples were analyzed by Soxhlet extraction to determine
uniformity of the sediment contamination. Water was added to beakers containing
sediment samples and tumbled for 7 days to allow determination of equilibrium
partitioning.

      These initial experiments have assisted in the continued improvement of the
experimental procedures. The ability to maintain the viability and activity of the
organisms for periods of weeks to months in the microcosms has been
demonstrated.  Animal densities of up to 100 per cell  (>1 per cm2) have been
maintained during an experiment and  up to 300 per cell have been maintained in
control cells.  Survival ranged between 95-98% of the animals introduced to the
cells during an experiment.  This number does not include initially dead  animals
that were replaced one day after the start of each experiment. The number of dead
animals introduced to the sediment at the start of the  experiment was always less
than 20% and was negligible if recently acquired  animals were introduced into the
cells. During the second experiment, three contaminants (pyrene, phenanthrene
and dibenzofuran) were introduced into the sediment rather than a single
contaminant (pyrene) at lower concentrations. Animal mortality and the initial
replacement rate of animals was higher during this experiment. Lower
concentrations will be used in subsequent experiments but three contaminants
representing a range of hydrophobicity will continue to be used.

      Analytical procedures have not developed as rapidly as animal handling
procedures. Water samples are extracted using hexane, concentrated by
evaporation and analyzed by high performance liquid chromatography (HPLC). A
great deal of analytical method development has resulted in the identification of
procedures that maintain high solvent extraction efficiencies  and reproducible
quantitation levels in the part-per-billion range.  Problems were experienced,
however, with low-level contamination of the glassware used in the analysis.
After considerable evaluation, cleaning procedures have  been modified to reduce
the significance of this problem.  Regular spiked replicates and blanks are also part
of the analytical procedures. To speed the analysis of samples and improve
reproducibility,  an automatic sampler was added to the HPLC.

      Figure  1  shows the results of the first of the experiments. The average
chemical flux of pyrene from the sediment is shown as a function of days into the
experiment. The data suggest that the presence  of the organisms results in a 2-4
fold increase in the pyrene flux.  Some data indicate,  however, that the
contaminant fluxes are well-correlated with water flow rate through the cell,
suggesting water side mass transfer resistances control.  In addition, a theoretical
model of the water flow over the sediment also suggests that water side
resistances should control.  Under  such conditions, the influence of  organism
                                    169

-------
30.00-
I

**  20.00

..o
CO
^c,
X       te
J=  10.00-
 0.00
     0
                                  50 Animals
                                  25 Animals
                                   a1
                                  No animals
           a
                        nr
                   10    15    20
                         time (days)
25    30.    35
         Figure 1  Average Fluxes for pyrene
                in the first experiment
                          170

-------
density on overall contaminant flux from the sediment would be expected to be
small. The increase indicated in Figure 1 at higher organism density may represent
the effect of organism activities on water mixing. Thus it remains uncertain
whether the data collected to-date represents the effect of organisms on
contaminant movement under field conditions, which are generally sediment-side
mass transfer resistance controlled.

      The concentration of pyrene in the animals at the conclusion of the
experiments was approximately 2% by dry weight.   Given a log K00 of about 5 for
pyrene, this is consistent with essentially ail of the  dry weight of the organisms
being carbon available for pyrene sorption.  The average mass of pyrene in a single
organism was  about 36 JJQ. By comparison the average mass  of pyrene collected
in the cell outlet over the entire experiment is only about .1 fjg. Thus the 25-100
organisms per cell sorbed much more  pyrene than was released to the water
column.  However, the 36 fjg  of pyretie in the organisms at the organism at the
conclusion of an experiment is estimated to be 10-20% of the pyrene contained in
the 0.5-1  cm3 of sediment processed  by an organism over the course of the
experiment. Thus most  of the pyrene in the sediment processed by the animals
reached the surface but  desorption into the water column was significantly slowed
by either water-side or intraparticle transport resistances.  The large body burden
of the organisms also  suggests that predation on contaminated animals may be an
important mechanism  for transport of hydrophobic  contaminants into the water
column.

      A mathematical model has  been developed to investigate the movement of
hydrophobic contaminants by conveyor belt feeders such  as Tubifex tubifex.  The
model is essentially a  box model with separate well-mixed zones at the sediment-
water interface, in the aerobic sediment layer and in the anaerobic sediment layer.
The geometry of the model is simple but the bulk of the effort associated with its
development has been the location and incorporation of organism growth and
activity as a function of environmental conditions,  e.g. temperature.
Unfortunately, the model does not adequately describe a situation where water-
side processes control or when diffusive processes dominate sediment side
transport.  Improvements to this model are under development.

      A theory has been developed and presented  in a paper  at the Summer
National Meeting of the  American Institute  of Chemical Engineers which suggests
that the importance of water-side resistances decreases with  time and increases
with hydrophobicity of the contaminant.  The second experiment included less
hydrophobic chemicals than pyrene in an attempt to overcome this problem.  The
second experiment  was also conducted with  higher overlying  water flowrates in
the microcosms to further reduce water-side mass  transfer resistances. The
results of this experiment, however, remain inconclusive. Attempts  to employ a
significantly higher  water flowrate resulted in analytical problems in that the water
                                    171

-------
samples were too dilute to accurately quantify contaminant concentration.  Current
efforts are directed toward refinement of the analytical techniques to resolve this
problem.  In an experiment about to be initiated, the flowrate through the cells will
be increased by a factor of 25 and the water volume collected for sample will be
increased by a similar amount. A continuous liquid-liquid extractor with a built in
slow-dry concentrator will be used to increase the sample concentration by a
factor of about 20 over current procedures. The equipment is suggested by EPA
for priority pollutant sample extraction. The extraction will be via petroleum ether
which was selected for its extraction effectiveness for poiynuciear aromatics and
its high vapor pressure which eases the evaporation-step. After extraction, the
organic concentrate will be exchanged with acetonitrile using a micro-Snyder
column.  Through this process an initial sample of 1 L of water will be
concentrated into a 1  ml solution with acrylonitrile. Analysis will continue to be
via HPLC. It is expected that the higher flowrates allowable by this concentration
of the sample will negate any water-side mass transfer resistances. Both
preliminary experimentation and the theoretical model suggest that the sediment-
side should  control the contaminant flux under these conditions. The expected
endpoint of  the first year studies remains the estimate of the hydrophobic organic
flux enhancement by  the presence of Tubifex tubifex at various densities in
laboratory microcosms.

Proposed Work

      The focus of the Year 02 activities will be a comparison of the action of
Tubifex tubifex with the action of the  Physella virgata.  It is expected that no new
analytical problems are posed by the presence of the other organism either
together or  separately from the Tubifex tubifex.  The experimental plan for the year
includes

      1.     Completion of Tubifex tubifex experiment currently being  initiated.
      2.     Repetition of previous experiment (or similar) to insure reproducibility.
      3.     Experiment with Physella virgata to  define baseline behavior
      4.     Combined experiment with delayed addition of Physella virgata to
             determine combined effect.

      Coupled with the measurements of water concentrations of the
contaminants will be  measurements of dissolved organic carbon and, subject to the
development of appropriate techniques, measurements of the  partitioning between
the dissolved organic carbon and water. In addition, the position of Tubifex tubifex
in the food  chain will  be evaluated to assess movement of contaminants into the
water column by predation.  This information is expected to provide a  solid
database for the development of mathematical models of the bioturbation process
for the test  organisms.
                                    172

-------
        LABORATORY AND MODELING INVESTIGATIONS OF SURFACTANT ENHANCED
                                   AQUIFER REMEDIATION

            Linda M. Abriola, Kurt D. Pennell, Timothy J. Dekker, and Walter J. Weber, Jr.

                         Department of Civil & Environmental Engineering
                                        1351 Beal St.
                                  116 Engineering Building 1A
                                  The University of Michigan
                                  Ann Arbor, Ml 48109-2125
                                        313-763-1464

INTRODUCTION

       The contamination of groundwater by organic solvents and other petroleum products has
become a major environmental issue throughout the United States.  These compounds frequently enter
the subsurface as a separate organic phase or nonaqueous phase liquid (NAPL).  Following a NAPL
spill, the organic phase will migrate downward as a result of gravitational and capillary forces, with a
portion of the NAPL being retained in soil pores as immobile globules due to interfacial forces.
Although considerable effort has been directed toward the study of NAPL transport in porous media, the
development of viable remediation technologies remains a formidable task.  It is widely recognized that
existing pump and treat remediation technologies are ineffective and extremely costly due to the low
aqueous solubility of most NAPLs.

       Surfactant enhanced aquifer remediation (SEAR) has been proposed as an alternative method
for removing residual  NAPLs from contaminated aquifers. Two potential SEAR approaches may be
distinguished based on (a) micellar solubilization of NAPLs and (b) mobilization of  entrapped NAPLs.
The first approach capitalizes on the capacity of aqueous surfactant solutions to enhance the solubility
of hydrophobic organic compounds. Surfactants are amphiphilic compounds, possessing both a
hydrophilic and lipophilic portion.  Above the critical micelle concentration (CMC), surfactant monomers
aggregate to form micelles consisting of a lipophilic core surrounded by a hydrophilic mantle.  The
dramatic enhancement of the aqueous solubility of NAPL above the CMC is due to the incorporation of
hydrophobic  compounds within surfactant micelles (1, 2).  Aqueous surfactant solutions could be
utilized in conventional pump and  treat remediation schemes to enhance the solubility of NAPLs,
thereby increasing the rate of NAPL recovery.  The second approach involves the  mobilization of NAPL
due to reductions in the interfacial tension between the organic and aqueous phases. Surfactants have
been utilized by the petroleum industry to achieve ultra-low interfacial tensions for enhanced oil
recovery.  This approach could be employed for aquifer restoration, provided that the  mobilized NAPL
could be recovered from groundwater formations.

       To date, surfactant flushing experiments have achieved mixed results and  standard procedures
for the design and implementation of SEAR technologies have yet to be established.  A fundamental
requirement for the development of such procedures is the ability to mathematically model SEAR
processes. However, relatively few controlled SEAR studies have been performed which provide the
data necessary for  model development and critical evaluation. Thus, the objective of this research was
twofold: (a) to conduct detailed surfactant flushing experiments and (b) to develop and validate a
numerical  model capable of predicting surfactant enhanced solubilization of entrapped NAPLs. The
model will provide a means for interpreting laboratory experiments, investigating the factors influencing
NAPL recovery, and evaluating alternative SEAR strategies.
                                            173

-------
METHODOLOGY

       Three soil column experiments were conducted to measure the ability of polyoxyethylene (POE)
(20) sorbitan monooleate to solubilize residual dodecane in Ottawa sand (20-30 mesh) (3).  Borosilicate
glass columns (4.8-cm i.d.) were packed with air-dried Ottawa sand and saturated with deionized water.
Dodecane saturations of approximately 17% were achieved by injecting 14C-labeled dodecane into the
water-saturated column, followed by the displacement of 15 pore volumes of water. Pulses of tritiated
water, a nonreactive tracer, were pumped through the soil column before and after dodecane
entrapment to determine the effect of the residual organic liquid on hydrodynamic dispersion. A 4%
solution of POE (20) sorbitan monooleate was then introduced into the soil column to solubilize residual
dodecane.  The concentration of dodecane in the column effluent was determined by  standard liquid
scintillation counting (LSC) procedures.  The apparent solubility of dodecane in aqueous surfactant
solutions and surfactant sorption by Ottawa sand were determined in batch reactors.

       A transient NAPL solubilization model was developed which incorporates transport equations
for both the surfactant and the solubilized organic liquid (4).  The two partial differential equations were
coupled by the solubility of the NAPL, represented as a linear function of the surfactant concentration
above the CMC. The rate of mass transfer between the entrapped NAPL and aqueous phases was
described using a lumped mass transfer coefficient based on a  linear driving force model.  Surfactant
sorption by the solid phase was represented by the Langmuir equation. The transport equations were
successfully implemented in the model using a weighted residual finite element scheme based on the
Galerkin method.  The entrapped NAPL was assumed to exist as discrete spheres, with a range of
diameters at each location in the domain.  An estimate of the median size of dodecane globules was
obtained from styrene polymerization experiments conducted in the same Ottawa sand (5). Changes in
the size and mass of the NAPL spheres were calculated at each time step, and the corresponding
values of NAPL saturation were updated continuously. Model input parameters for dodecane
solubilization and surfactant sorption were obtained from the batch studies, while column dispersivity
values were derived from tritiated water breakthrough curves.

RESULTS

       The solubility of dodecane in aqueous solutions of POE (20) sorbitan  monooleate increased
linearly over the range of surfactant concentrations studied (Figure 1). In a 4% surfactant solution the
solubility of dodecane was approximately 6 orders-of-magnitude greater than its solubility in water (3.7
X 10"3 mg/L).  The dramatic and linear enhancement in dodecane solubility above the CMC is
consistent with solubilization data reported for other hydrophobic organic chemicals of environmental
concern (1,2).

       Following the injection of a 4% solution  of POE (20) sorbitan monooleate, the concentration of
dodecane in the soil column effluent increased to approximately 500 mg/L (Figure 2).  Although this
represents a sizable increase in the solubility of dodecane, the effluent concentration was 7 times less
than the equilibrium value (3,500 mg/L) determined in batch studies.  To further investigate the impact
of  nonequilibrium solubilization on dodecane recovery, (a) the pore-water velocity was varied and (b)
flow was interrupted during each column experiment.  The results of a representative  column
experiment conducted at pore-water velocities ranging from 6 to 22 cm/hr and flow interruption periods
of 3.5 to 100 hrs is shown in Figure 2.  The steady-state effluent concentration of dodecane  was
reduced from approximately 500 mg/L to 150 mg/L as the pore-water velocity was increased. In
addition, the effluent concentration of dodecane approached the equilibrium value as the duration of
flow interruption was increased to 100 hours.  These data are indicative of rate-limited, rather than
instantaneous, solubilization of residual dodecane.  The interruption of flow and reduction in  interstitial
velocity allowed for greater contact times between the surfactant solution and  residual dodecane,
thereby increasing the effluent concentration of dodecane.
                                            174

-------
       The NAPL solubilization model was utilized to simulate the soil column experiments described
above. Comparisons of simulated and measured dodecane effluent concentrations are shown in Figure
2. The initial appearance of dodecane in the column effluent was accurately predicted by the model.
The slight retardation in dodecane breakthrough was attributed to sorption of the surfactant.  Thus, the
use of batch-derived Langmuir adsorption coefficients provided reasonable estimates of surfactant
sorption and the corresponding breakthrough of solubilized dodecane.  Rate-limited solubilization of
dodecane was represented by a coefficient that accounted for mass transfer during both steady-state
and interrupted flow. Periods of flow interruption were simulated using a time-varying superficial
velocity and a zero flux boundary condition.  The model was able to capture the change in steady-state
effluent concentrations of dodecane, as well as the position and height of the effluent peaks following
flow interruption.

CONCLUSIONS

       The results of batch solubilization experiments demonstrate the sizable capacity of POE (20)
sorbitan monooleate to enhance the solubility of dodecane, which can be attributed to the incorporation
or partitioning of dodecane within the hydrophobic core of the surfactant micelles. The injection of a
4% surfactant solution into contaminated soil columns dramatically improved the recovery of residual
dodecane.  Effluent dodecane concentrations were, however, considerably less than  the equilibrium
solubility.  This discrepancy was attributed to rate-limited solubilization of residual dodecane. Evidence
for mass transfer rate limitations included: (a) the increase in effluent dodecane concentrations following
flow interruption and (b) the reduction in steady-state effluent concentrations as the pore-water velocity
was increased.  A one-dimensional solubilization simulator was developed to model surfactant
enhanced solubilization. The model was able to accurately represent rate-limited micellar soiubilization
of residual dodecane in soil columns.  Model results suggest that under rate-limited conditions the
selection of optimum pumping strategies should be based, in part, on the trade-off between the total
flushing time and the required volume of surfactant solution.

REFERENCES

1.     Edwards, D.A., Luthy, R.G., and Liu, Z. Solubilization of polycyclic aromatic hydrocarbons in
       micellar nonionic surfactant solutions. Environ. Sci. Technol. 25(1): 127-133,  1991.

2.     Kile, D.E. and Chiou, C.T. Water solubility enhancements of DDT and trichlorobenzene by
       some surfactants below and above the critical micelle concentration. Environ. Sci, Technol.
       23(7): 832-838, 1989.

3.     Pennell, K.D., Abriola, L.M., and Weber, W.J., Jr. Surfactant enhanced solubilization of  residual
       dodecane in soil columns 1. Experimental investigations, submitted to Environ. Sci. Technol.

4.     Abriola, L.M., Dekker, T.J., and Pennell, K.D. Surfactant enhanced solubilization of residual
       dodecane in soil columns 2. Mathematical modeling, submitted to Environ. Sci. Technol.

5.     Powers, S.E., Abriola, L.M., and Weber, W.J., Jr. An experimental investigation of nonaqueous
       phase liquid dissolution in saturated subsurface systems: Steady mass transfer rates.  Water
        Resour. Res. 28(10): 2691-2705, 1992.

FOR  MORE INFORMATION:   Linda M. Abriola, The University of Michigan, Department of Civil &
Environmental  Engineering, 1351  Beal St., 116 Engineering Building 1A, Ann Arbor,  Ml 48109-2125,
Tel: 313-764-9406.
                                             175

-------
I*+UUU

12000


5*10000
0)
£ 8000
S
CD
"8 6000
Q
4000




2000





n
,8''
X
X
•• X
X
X
X
~ X
X
X
"" X

X »
^ ^
x^'
X
xO''
X
X
X
0'
X
X
X
X
X
" 1 1 1 1 1 1 1 1
§
•§
Q
        0      20     40    60     80     100    120    140   160   180
                           Surfactant Concentration (g/L)
  Figure 1. Solubility of dodecane in aqueous solutions of POE (20) sorbitan
           monooleate at 22 C.
                                       10
                                  Pore Volumes
  Figure 2. Measured and predicted effluent concentrations of dodecane following
          the injection of a 4% solution of POE (20) sorbitan monooleate.
                                 176

-------
                  PROCEDURES FOR SELECTION AND DESIGN OF A CAP FOR
                     IN-SITU TREATMENT OF CONTAMINATED SEDIMENT

            Danny D. Reible, Greg J. Thoma, Kalliat T. Valsaraj, Louis J. Thibodeaux
                             Department of Chemical Engineering
                                  Louisiana State University
                                   Baton Rouge, LA 70803
                                      (504) 388-3070

                                     Dennis Timberlake
                                       US EPA- RREL
                                   Cincinnati, Ohio 45268
                                      (513) 569-7547

INTRODUCTION

       Most methods for the remediation of contaminated sediments involve the removal and surface
treatment of the  sediment.   Effective removal is difficult and expensive,  however, as is surface
treatment.   Removal and  surface treatment also provides  many  opportunities  for  release  of
contaminants to the ecosystem.  Resuspension, dissolution and evaporation all occur during dredging.
Temporary or permanent  storage in confined disposal facilities leads to contaminant releases via
effluent discharge, evaporation,  leachate  and runoff.  In addition, no surface  treatment process is
100% effective nor results in 100% containment of the pollutants during treatment.  To avoid these
problems, in-situ capping of contaminated sediments with a clean sediment has been proposed. If the
cap can be engineered to avoid erosion and movement, this can be an extremely effective method of
containing hydrophobic contaminants in the sediment.  A cap increases the effective diffusion path,
significantly reducing the  diffusive flux of contaminant.  More importantly, however, the cap will
separate the benthic organisms that populate the upper 5-10 cm of the original sediment from the
contaminants. Bioturbation, the reworking of sediments by the movement and feeding habits of these
organisms,  can  be an  important  mechanism for the transport of particles  and  any  associated
contaminants, and the presence of a cap of sufficient thickness effectively eliminates this mechanism.
In addition,  capping can effectively eliminate the introduction of contaminants to the food chain by
reducing the exposure of the benthic organisms.  Reduction of bioavailability has generally been one
of the first reasons to consider capping of contaminated sediment.

        The selection of capping as a remedial alternative depends on a variety of factors that are not
currently well-defined. The initial concern is for the integrity of  any cap. Although a cap  can always
be replaced after  erosion due to a storm event or after  slow loss of the capping layer, the preference
would be for long-term stability of the cap. The remaining considerations in the selection and design
of an in-situ cap for treatment of contaminated sediments involve
        1)      insuring adequate depth of cap to eliminate significant bioturbation of the contaminated
               layer and maintain sufficient effective cap depth to contain the  pollutant,  and,
        2)      appropriate  selection  of  cap materials to  maximize containment  of  the  desired
               pollutants.
The US Army Corps of Engineers have developed experimental procedures to define the effective depth
of bioturbation for a particular capping project (1,2).  They have not developed analytical tools for the
quantitative analysis of  effective cap depth or other design criteria. Wang et al. (3) described some
simple  models developed on the basis of preliminary laboratory experiments that could be used to
determine  the required  effective cap depth and the resulting  containment of pollutant within the
contaminated sediment. Although the models adequately described the breakthrough and steady state
flux through the  laboratory  caps, the models did not consider the transient sediment contamination
levels or the effect of colloidal material in the porewater.
                                             177

-------
       It is toward these weaknesses that the current work is directed. The primary objective is the
development and laboratory verification of simple analytical tools for the design of a sediment capping
layer. These tools allow determination of the rate of the transient colloidally enhanced diffusion and
the effect of water side mass transfer resistances. The revised models are used to compare to the
experimental data of Wang et al. (3). Preliminary results with different chemical and sediment systems
will be summarized in the presentation.

METHODOLOGY

       The basic experimental procedures follow those of Wang et al. (3).  These experiments entail
the use of a sediment, artificially contaminated with a tracer compound, in small laboratory capping
cells. Freshwater sediments from area streams and lakes are currently employed in the experiments.
In the initial experiments by Wang et al. (3), two sands, a local lake sediment and Tao River sediment
were employed. In those experiments, the  sediments  were contaminated by equilibration with an
aqueous  solution of 2,4,6  trichlorophenol  (TCP).   In the  current  experiments, the sediment is
contaminated with polyaromatic hydrocarbons (PAHs),  e.g. pyrene, phenanthrene and  dibenzofuran,
at concentrations of about 200 mg/kg. This is accomplished by dissolving the compounds in hexane
and evaporating the hexane from a glass container, leaving the contaminants plated to the walls of the
container. Wet sediment is then added to the container and tumbled  for two to three weeks.

       The contaminated sediments were loaded into specially constructed flow cells. A skimmer was
used to control the sediment layer depth. A selected capping material (e.g. uncontaminated sediment)
was then placed on top again using a skimmer to control layer depth.  Water  was then pumped  over
the surface of the exposed sediment or cap, and the effluent water analyzed for the contaminant of
interest.  TCP was measured by UV spectrophotometry.  PAHs in the current experiments are extracted
from the  aqueous effluent with hexane and analyzed via high performance liquid chromatography. The
concentration measurements provided a  direct indication of the contaminant leaching  rate from the
capped sediment. The details of the experimental procedure are described  in Wang et  al. (3).

       The measured effluent concentration and contaminant flux were compared to predictions of
three analytical  diffusion models.
       i)     Base case (no cap)- diffusion with water side resistances
       ii)     One layer model- underlying sediment assumed to maintain  initial concentration
       iii)     Two layer model- transient diffusion modeled in cap and sediment layers
All models predict the flux through the sediment-water interface as  a result  of diffusive processes
within the sediment.  Retardation of diffusion by sorption to sediment particles and enhancement of
diffusion by  colloidal  particles in the porewater are included in the models.  Advection is assumed to
be negligible as would likely be the case in any site considered for capping.  Bioturbation typically
results in much more rapid chemical transport than diffusion and therefore these models apply only to
the non-bioturbed portion of any sediments or cap.  The one layer model is similar to that presented
by Wang et al. (3) except allowance is made for water-side mass transfer resistances. The two layer
model is previously unpublished  and describes  the transient diffusion within an initially uniformly
contaminated sediment layer and an initially uncontaminated capping  layer.

RESULTS

       Baseline experiments without a sediment cap were used to develop and evaluate models of the
physico-chemical transport processes in the contaminated sediment. The flux of TCP during one such
experiment is shown in Figure 1.  The  contaminant flux continually decreased as a result of the
depletion of TCP in the surface layer of the sediment.  In the laboratory, analyses of the experiments
are complicated by the presence  of water-side mass transfer resistances that may not be important
under field conditions.   The model predicted fluxes are shown for the case of no water-side mass
                                            178

-------
s-
-
O)
    2500
    2000
    1500
X

3  1000
LL
CL
o

     500
        0
           o\
observed
kl = inf
kl = 2.55
          0     5     10    15    20    25    30    35

                          TIME (days)	
     Figure 1.  Comparison of predicted to observed TCP flux for experimental

              uncapped system

              kl - water side mass transfer coefficient
                               179

-------
transfer resistances (kl-»oo) and for the case of water-side mass transfer resistances estimated from
a model of the laminar flow over the sediment surface (kl = 2.55). The derivation of the water-side
mass transfer coefficient for the later case is shown in Thoma et al. (4). Note that the water-side mass
transfer resistances are relatively unimportant for the TCP experiments. A theory presented by Thoma
et al. (4), however, suggests that water-side resistances could be important for more hydrophobic
contaminants.

       Figure 2 shows the observed TCP flux from the four capped sediments during the experiments
of Wang et al. (3).  During all experiments, the fluxes increased during the experiment until the capping
layer was saturated with the contaminant. The observed flux then tended to decrease slowly due to
the depletion of contaminants from the underlying sediment.  The time of the maximum flux from a
given capped sediment was dependent on the depth  and sorptive characteristics of the cap.  The
slowest dynamics were exhibited by the most sorptive cap, that from University Lake. The partition
coefficient for TCP in this capping material was observed to be about 28 compared to less than 5 for
the Tao River sediment and about 1 for the two sands. As a  result the maximum flux was observed
more than 30 days into the experiment compared to less than 5 days for the other materials.

       The two layer model predicted this qualitative behavior well as shown in Figure 2 by the solid
line. The model results are based on separate estimates of model parameters and therefore represents
an unfitted prediction.  The prediction of an effective diffusion coefficient, however, depends on the
pure water diffusivity, the porosity and tortuosity of the medium, the concentration and partition
coefficient for the colloidal material in the porewater, and the  partition coefficient of the contaminant
between the sand or cap  and  porewater. Because of the uncertainty  in these parameters,  best fit
curves to the data are also shown.  As shown in the figure,  essentially equivalent fits are observed
whether the cap properties or the underlying contaminated sediment properties are fit to the data. The
best fit modeling parameters are generally within the range of uncertainty of these parameters.

CONCLUSIONS

       The generally good agreement between the model predictions and fits to the experimental data
of Wang et al. (3) suggest that the key contaminant migration processes are described by the models.
The models  are  currently being used  to evaluate experimental data  collected from sediments
contaminated with pyrene, phenanthrene and dibenzofuran and  to predict the  behavior of  other
contaminants under idealized field  conditions.  Preliminary results of  these  investigations will be
presented.

       The project is expected to significantly enhance the understanding of the in-situ containment
and isolation of contaminated sediments by capping. Simple design tools that can identify conditions
under which capping is preferred or contraindicated are being developed and tested.

REFERENCES

1.      Brannon, J.M., R.E. Hoeppel, T.C. Sturgis, I. Smith and D. Gunnison, "Effectiveness of capping
        in isolating contaminated dredged  material from biota  and  the overlying  water";Technical
        Report D-85-10 US Army Engineer Waterways Experiment Station: Vicksburg MS, 1985
2.      US Army Engineer Waterways Experiment Station, Environmental Effects of Dredging Technical
        Notes, EEDP-01 -3,4,9  , 1987 &1988.
3.      Wang, X.Q., L.J. Thibodeaux, K.T. Valsaraj and D.D.  Reible, Environ. Sci. Techno/. 1991, 25
        (9), 1578-1584.
4.     Thoma,G.J., D.D. Reible, K.T. Valsaraj and L.J. Thibodeaux, Environ. Sci. Techno/. 1992,
        submitted.
                                            180

-------
O)
E
500
450
400
350
300
250
200
150
Too
 50
  0
                   10     15
                   Time(days)
                   10     15
                              20
                                        25
                                                        10
15  20  25
 Time (days)
15  20  25
                                                                        30   35
                   observed
                   predicted
 Figure 2.       Comparison of predicted to observed TCP flux for experimental
                capped systems (A) Balsam sand  cap; (B) Tao River cap;
                (C) Quartz sand cap;  (D) University Lake cap
                CS- contaminated sediment (Kd.Dw-partition coefficient.water diff.;
                                         181

-------
                         OPTIMIZATION OF SOIL VAPOR EXTRACTION FOR
             REMEDIATION OF GASOLINE-CONTAMINATED SOIL AND GROUNDWATER

                                          Chien T. Chen
                                U.S. EPA, Releases Control Branch
                          2890 Woodbridge Avenue, Edison, NJ 08837-3679
                                         (908)906-6985

                                       Richard L Johnson
                       Department of Environmental Science and Engineering
                     Oregon Graduate Institute, 19600 N.W. Von Neumann Drive
                                    Beaverton, OR 97006-1999
                                         (503)690-1193

 INTRODUCTION

        Soil vapor extraction (SVE) is a widely used in-situ technology for removing volatile organic
 compounds (VOCs) from contaminated soil (1).  This technology can treat large volumes of soil at
 reasonable costs. SVE systems are relatively easy to install and use standard, readily available
 equipment (2).  However, overall removal efficiencies are limited by high molecular weight fraction of
 gasoline constituents within the vadose zone and all constituents in groundwater. Practical techniques
 to overcome these limitations and thereby increase the value  of SVE technology are required.

        A series of remediation experiments were conducted  to remove gasoline from a large
 experimental sand aquifer (LESA) at the Oregon Graduate Institute (OGI). The first SVE operation was
 conducted to evaluate the operational parameters.  Subsequent experiments were conducted to evaluate
 the optimization of the SVE process  by:  1) manipulation of the groundwater level to increase the
 air-gasoline contact area, 2) pulsed pumping of the SVE system to allow time for diffusing-iimited mass
 transfer, and 3)  air sparging into the saturated zone, and injection of air into the capillary fringe to
 enhance volatilization.

 METHODOLOGY

        The experiments were conducted in the LESA, a 3-D concrete tank (35' x 30' x 15' deep) filled
 with medium grade sand. The LESA (3) is shown in Figures 1 & 2 together with  a gasoline distribution
 resulting from a typical spill.  Two hundred to 400 liters (L) of  synthetic gasoline were released into the
 LESA over a period of several days at a point 0.5 meter (m) below ground surface and 3 m above the
 water table. An array of 5 vapor extraction wells was installed at one end of the LESA and 5 passive
 injection wells were located at the other end.  These wells and the associated plumbing were 2 inch
 schedule 40 polyvinyl chloride (PVC) pipes. The ground surface was covered with an impermeable
 plastic sheet and the extraction wells were connected to a regenerative blower. The concentrations of
 extracted hydrocarbons (HCs) and carbon dioxide (COj) were closely monitored. The water level was
fluctuated during the SVE process to examine the extent of the improvement of process efficiency. A
network of glass and video-camera access tubes (installed in the aquifer) were used for direct
observation of the gasoline (dyed bright red) in the medium. Four sparging wells (16' long x 2" diameter
schedule 40 PVC pipe with the bottom 5' screened) were installed near the center of the tank. During
the SVE process, the air flow rate was monitored by the injection and detection of a tracer (sulfer
hexafluoride [SF6i]) at a number of points throughout the aquifer.  The extracted vapors were collected
in evacuated stainless steel cylinders to which an internal standard had been added. The concentrations
of HCs in each cylinder were analyzed by capillary gas chromatography using a flame ionization
detector.
                                              182

-------
  PASSIVE AIR
INJECTION WELL
            ,GLASS OBERVATION WELL

                       SVEWELL
                          «
                          «
                          •
|~100LOF
i RESIDUAL
                                            • 36
• n
                                                      74
             SPARGE WELL
                                     ~ 300 L OF PRODUCT
                GROUNDWATER FLOW
            Figure 1. Cross-section view of the Aquifer after the
               Release of 400 L of the Synthetic Gasoline
                           •30 FEET-
                                    SPARGE WELL »1
                     LNAPL
                                                 GROUNDWATER
                                                   FLOW
                                               & CLASS WELL


                                               D VAPOR/GROUNDWATER

                                                 BUNDLES
                              SVE WELLS
            Figure 2. Plan View of the Aquifer after the Release
                   of 400 L of the Synthetic Gasoline
                                183

-------
RESULTS

        In the course of this project, two spills of synthetic gasoline followed by SVE and SVE enhanced
processes were conducted.  The results are described as follows:

                                         FIRST SPILL

        Two hundred liters of the mixture as shown in Table 1 were spilled into the aquifer.  The vapor
concentration and the distribution of the product in the subsurface were then tracked on a continuous
basis.  Due to the small amount of spilled material, only a small portion of the gasoline reached the
water table.  Most of the gasoline was retained within the unsaturated zone. The initial distribution of the
gasoline was estimated to be as follows:

        60% as residual in the unsaturated zone
        32% as floating product on the water table
        8% as residual in the saturated zone

               TABLE 1.  VOLUME PERCENTAGE (%) OF THE CONSTITUENT IN THE
                      SYNTHETIC GASOLINE USED FOR THE FIRST SPILL
Compound
n-Pentane
Isopentane
Methylcyclopentane
2-Methylpentane
3-Methylpentane
n-Hexane
Cyclohexene
Volume
Percentage
1
1
9
13
20
46
1
Compound
Toluene
Neptane
2,2,4-Trimethylpentane
Styrene
Cumene
Decane
Methyl t-butyl ether
Volume
Percentage
2
1
1
1
1
1
1.
Soil Vapor Extraction

   The SVE was started 6 months after the spill.  The volumetric flow through the pump was maintained
at 100 cubic feet per minute under standard conditions (scfm).  The concentration of HCs in the
extracted vapor dropped quickly, especially that in the upper half of the unsaturated zone where the HCs
disappeared in two hours.  In the lower half (0.5 m above the water table), the HC concentration
decreased to 20% in the same period of time.

Manipulation of the Groundwater Level

   The water table was dropped approximately 0.5 m over the course of 1 week before the SVE was
started again.  The preliminary results showed that water level manipulations had little effect on removing
mass from the groundwater.

Blodearadatlon

   CO2 concentrations within the aquifer increased continuously after initiation of the spill.  The flux of
CO2 out of the system suggested that about 0.1 kg/day of HCs were being degraded for 250 days. The
analysis of CO2 was discontinued at that time.
                                              184

-------
                                       SECOND SPILL

   Two hundred liters of synthetic gasoline were spilled into the aquifer. The composition of this
synthetic gasoline is shown in Table 2. As had been the case for the first spill, most of the gasoline was

               TABLE 2. VOLUME PERCENTAGE (%) OF THE CONSTITUENT IN THE
                     SYNTHETIC GASOLINE USED FOR THE SECOND SPILL
Compound
n-Pentane
Isopentane
Methylcyclopentane
2-Methylpentane
3-Methylpentane
Volume
Percentage
2
1
9
13
20
Compound
n-Hexane
Benzene
Toluene
2,2,4-Trimethylpentane
Styrene
Volume
Percentage
45
1
3
3
3
retained in the unsaturated zone.  Because of this, a second 200 L was released into the tank one week
after the first spill. The distribution of gasoline at the conclusion of the 400 L release is shown in Figures
1 &2.

Soil Vapor Extraction

       Ten days after the spill, SVE was started. The air flow rate was set at 80-100 scfm. The
concentration of HCs in the extracted vapor dropped quickly from about 2,000 parts per million by
volume (ppmv) at the beginning to 500 ppmv in 40 hours, then to about 150 ppmv about six days after
the SVE was initiated. The vapor concentrations then remained "relatively constant" at that level. At the
end of 18 days, the pump was turned off. The total mass of hydrocarbons recovered during those 18
days was 48 kilograms (kgs).

Pulsed Pumping

       After 4 days of not pumping, the pump was started again. One hour after the pump was
restarted, the HC concentration increased to 1,670 ppmv.  Two days later it dropped to about 100 ppmv
and decreased slowly thereafter. Pumping was continued for five more days at which time the
concentration of HCs was about 60 ppmv and the system was shut down. The total amount of gasoline
recovered during this period was 22 kgs.  The estimated gasoline retained in the unsaturated zone was
almost completely removed.

Air Sparging

       The water table was raised approximately 1 m to entrap all of the free product below the water
table.  Four air sparging (AS) wells (Figure 2)  provided air under the trapped residual using a
continuous-duty oil-free compressor.  A total of five SVE/AS experiments were conducted on well #1. In
all these experiments, the HCs were not observed until about 30 minutes after the operation was
initiated; the maximum concentration was reached about one hour later.  The concentrations then
dropped quickly. At the end of the fifth experiment, the concentration dropped to zero. A total of about
70 hours of SVE/AS were conducted and 2.3  kgs of HCs were removed. The same experiments were
conducted on wells #2, #4 and #3 except that only one sparging each was needed for these wells.
The total recovery of HCs from these three wells was 0.05, 1.5 and 0.4 kgs respectively. A total of 4.2
kgs from the remaining 230 kgs was removed by SVE/AS.  The concentrations of various constituents
                                              185

-------
from well #4 are shown in Figure 3.,  The hexane shown in the figure is the total of methylcyclopentane,
2-methypentane, 3-methylpentane and n-hexane.
                                         SPARGE WELL 4
                                                 STYRENE

                                                   ISOOCTANE

                                                    BENZENE
                       U
                                                               28
                                         TIME (HOURS)
                      Figure 3. The Concentrations of Various Compounds in
                        the Extracted Vapor when Sparging Well #4 was On
Infection of Air into the Capillary Fringe

       With the water table lowered back to the level of the original release, air was injected into the
capillary zone through the sparge wells at 5 scfm (Figure 2). The extraction rate of HCs increased to a
maximum of 1  gm/min in about 2 hours.  It then decreased gradually to about 0.08 gm/min in about 50
hours. The total mass recovered from each well was about 1 kg.  Following this, the water table was
lowered another 0.6 m and the same SVE with air injection was applied In the new capillary zone.  The
total mass recovered from the two wells was 20 kgs. The other two wells did not recover appreciable
mass.

CONCLUSIONS

       The following conclusions can be drawn from the results to date:

       o SVE can readily remove gasoline constituents from the unsaturated zone In a short period of
          time, but not from the capillary or saturated zone.

       o Pulsed pumping can decrease pumping time; however, it should be examined in greater detail
          to optimize the pulse interval.

       o Air sparging removed hydrocarbons trapped In the saturated zone; however, the removal
          efficiency,was poor. This may be due to the design of the sparging wells which had very
          small radii of Influence (4).  More research Is necessary to improve the design and application
          of air sparging.

       o Injection of air Into the capillary zone did enhance SVE removal of gasoline; however, better
          design of the injection wells may increase overall efficiency.
                                               186

-------
Acknowledgement

       This research work was funded by the United States Environmental Protection Agency (EPA)
under the Cooperative Agreement No. CR816947-01 to the Oregon Graduate Institute (OGI).  The
experimental work was conducted by William Bagby and Matthew Perrot of OGI to whom
acknowledgement is made. The authors also would like to thank Anthony N. Tafuri, S. Krishnamurthy,
and Esperanza Renard of EPA for their, helpful discussion and critical review.

REFERENCES

1. Environmental Protection Agency, Innovative Treatment Technologies: Semi-Annual Report,
   EPA/540/2-91/001, Office of Solid Waste and Emergency Response, Technology Innovation Office,
   Washington, D.C., 1992. 79 pp.

2. Pedersen, T.A. and Curtis, J.T. Soil Vapor Extraction Technology Reference Handbook,
   EPA/540/2-91/003, U.S. Environmental Protection Agency, Cincinnati, Ohio, 1991. 316 pp.

3. Johnson, R.L., Bagby, W.,  Perrot, M. and Chen, C.T. Experimental Examination of Integrated Soil
   Vapor Extraction Techniques. Jrr Proceedings of the 1992 Petroleum Hydrocarbons and Organic
   Chemicals in Ground Water:  Prevention, Detection, and Restoration. The American Petroleum
   Institute and The Association of Ground Water Scientists and Engineers, Washington, D.C., 1992. pp
   441-452.

4. Loden, M.E.  A Technology Assessment of Soil Vapor Extraction And Air Sparging,
   EPA/600/R-92/173, U.S. Environmental Protection Agency, Cincinnati, Ohio, 1992. 63 pp.
                                              187

-------
     SITE DEMONSTRATION OF RESOURCES CONSERVATION COMPANY'S
                        B.E.S.T. PROCESS

                         Mark C. Meckes
              U.S. Environmental Protection Agency
              Risk Reduction Engineering Laboratory
                 26 West Martin Luther King Drive
                     Cincinnati,  Ohio  45268
                          (513) 569-7348

              Thomas J. Wagner and Joseph W. Tillman
         Science Applications International Corporation
                     635 West Seventh Street
                            Suite 403
                     Cincinnati,  Ohio  45203
                          (513) 723-2600

INTRODUCTION

     The technology demonstration segment of  the Superfund
Innovative Technology Evaluation (SITE)  program assesses the
effectiveness of developing technologies to treat waste streams
from uncontrolled hazardous waste sites.  This presentation
summarizes the activities and results of a pilot-scale evaluation
of Resources Conservation Company's (RCC)  Basic Extractive Sludge
Treatment (B.E.S.T.)  process, conducted in cooperation with the
Great Lakes National Program Office and the U.S. Army Corps of
Engineers.

     The B.E.S.T. process uses organic solvents, usually
triethylamine (TEA),  to extract organic contaminants from soils,
sludges, and sediments.  The process is nondestructive and
functions as a separation technology by segregating materials
into three fractions,  oil, water, and solids.

     The pilot plant used for this evaluation is designed to be
batch fed and can treat up to 150 pounds of solids per batch.  It
consists of four basic operations:  extraction, solvent recovery,
solids drying, and water stripping.  Extraction is carried out in
two separate vessels.   Chilled TEA is miscible with water.
Therefore,  cold extractions will dewater solids and remove some
organics.  Cold extractions  (<50°F)  are  conducted on high water
content materials (i.e. sludge)  in a premix tank.  Solids are
batch fed into the tank?  the tank is purged with nitrogen to
eliminate a flammable atmosphere;  chilled TEA is then added to
                               188

-------
the tank and mechanically mixed with the solids.  After five
minutes, the mixer is stopped and the solids are allowed to
settle.  Fluids are decanted from the tank and pumped to a
centrifuge.  Centrate is pumped to the solvent recovery system.
Solids from, the centrifuge and the premix tank are moved to the
extractor/dryer.

     The same series of events which occurred in the premix tank
now take place in the extractor/dryer.  However, this vessel is
heated up to 170°F by injecting steam into its jacket.   Water has
limited solubility in TEA at temperatures above 100°F,  and TEA
has increasing affinity for organic compounds as the temperature
is increased.  Consequently, the fluid which remains after
agitation in the extractor/dryer contains less water.  The number
of extraction cycles necessary to decrease the concentration of a
specific organic contaminant to a given cleanup level is
determined by bench scale treatability tests.

     Solvent/oil/water mixtures obtained from the extractions are
pumped to a solvent evaporator.  Here the fluids are heated to
boiling, leaving the oil fraction as a still bottom.
Solvent/water vapors are directed to a condenser (operating at
100°F)  where a heterogeneous condensate of solvent  and water is
formed.  This mixture flows into a solvent decanter, where the
now immiscible water and solvent phases separate.  Solvent is
decanted and pumped to the solvent storage tank for reuse.  Water
is drained from the decanter into a water receiver tank.  TEA is
recovered from this water by direct contact steam stripping.

     Solids drying is conducted in the extractor/dryer.  Steam is
applied to the jacket of the extractor/dryer which raises the
temperature of the solids to 170°F.   The solids are then agitated
by the vessel's paddle impellers, to increase heat transfer and
reduce drying time.  After the bulk of the solvent is removed,
steam is directly injected into the vessel.  The solvent and
steam form an azeotrope which is directed to a condenser.  Dry
solids, with a moisture content of about 5 weight percent, are
then discharged through a port located on the bottom of the
vessel.

METHODOLOGY

     Sediment samples were obtained from two separate areas of
the Grand Calumet River in Gary, Indiana.  The sampling locations
were selected based upon previous site data provided by EPA
                                189

-------
Region V.  The first area to be sampled was designated transect
28 and was known to have concentrations of polychlorinated
byphenyls (PCBs) below 50 rag/kg.  The second area to be sampled
was designated transect 6.  PCB concentrations in the sediments
at this location were known to be >50 mg/kg.

     Sediment samples were collected from a given location and
transported to the demonstration site.  Here the sample was,
passed through an eighth inch screen, into a large tank where it
was homogenized with a portable mixer.  Each mixed sediment was
placed in containers, weighed, and stored for processing.

     Samples of the two sediments were sent to RCC for bench
scale treatability testing.   RCC used the results from these
bench scale tests to specify operating conditions to be used
during pilot scale testing.  The experimental design allowed for
five separate batches of each sediment sample to be treated by
the B.E.S.T. process.  Therefore a total of ten batches were
processed.  RCC was allowed to optimize process operations over
three of the five batch runs, the remaining two runs were to be
operated under the same operating conditions which RCC designated
as "optimum".

     Samples were collected from the untreated and treated
solids, product water, product oil, recycled solvent, and vent
gases.  Analyses included semivolatile organics, oil and grease,
PCBs, and TEA for solid and liquid samples.  Vent gases were
analyzed for TEA.

RESULTS

     Samples obtained from transect 28 contained 41% moisture,
and were designated sediment A.  Samples from transect 6
contained 64% moisture, and were designated sediment B.  For
sediment A, seven extraction cycles were used.  The optimized
operating conditions included two cold extractions  (42-53°F) ,
followed by a warm extraction  (97-110°F) ,  followed by three hot
extractions  (152-167°F),  and one warm extraction (118-121°F) .
Seven extraction cycles were also used for sediment B.  However,
for treatment of this sediment two cold extractions  (28-53°F)
were followed by five hot extractions  (146-170°F) .

     Table 1. shows the average contaminant concentrations  in
the untreated and treated sediment samples, and the removal
efficiencies attributed to solvent extraction.  The fact that
                                190

-------
high removal efficiencies (>96%)  of the organic contaminant
fraction were obtained following treatment of the sediments was
expected, and was consistent with RCC's claims.

        TABLE 1.  SEDIMENT A AVERAGE REMOVAL EFFICIENCIES
Parameter
Total PCBs, mg/kg
Total PAHs, mg/kg
Oil and Grease, mg/kg
Untreated
12
550
7000
Treated
0.04
22
111
Percent
Removed
99.7
96.0
98.4
     Table 2.   reveals that the results obtained for treatment of
sediment B were consistent with the results shown in Table l.
Furthermore, initial contaminant concentration were more than an
order of magnitude greater in sediment B.  Removal efficiencies
of total PAHs from sediment B were 99.3% as compared to 96.0% for
sediment A.  This higher removal efficiency is attributed to
higher initial PAH concentrations in sediment B.

        TABLE 2.  SEDIMENT B AVERAGE REMOVAL EFFICIENCIES
Parameter
Total PCBs, mg/kg
Total PAHs, mg/kg
Oil and Grease, mg/kg
Untreated
425
70,900
128,000
Treated
<1.6
510
1460
Percent
Removed
>99.6
99.3
98.9
CONCLUSIONS

     Evaluation of the demonstration test data indicates that the
B.E.S.T. process was very effective in removing organic
contaminants from test sediments obtained from the Grand Calumet
River.  All vender claims made prior to the demonstration
regarding percent removal of organics, solvent residual in
process products, and system mass balances were met.

FOR MORE INFORMATION: Contact the author
                                191

-------
         CHARACTERISATION AND TREATMENT OF CONTAMINATED SOILS
                   USING MINERAL PROCESSING TECHNIQUES

                             Peter Wood and Mike Pearl

                             Warren Spring Laboratory
                                Gunnels Wood Road
                                    Stevenage
                                  Herts SG1 2BX
INTRODUCTION
       Within the U.S. EPA's Emerging Technology Program, Warren Spring Laboratory (WSL) is
investigating the application of mineral processing techniques for the characterisation and
treatment of contaminated soils.  The aim of these investigations is to  enable the
separation, isolation or concentration of contamination leaving cleaned non- or less
contaminated material for re-use or simple disposal. The contaminated material can then
go for further treatment or disposal by other techniques, if successful this approach will be
particularly useful for the treatment of large volumes of materials with  low levels of
contamination and will minimise the amount of waste material needing to be treated or
landfilled.

       The two year program of investigations consists of laboratory studies designed to
characterise the distribution of contamination within the soil and to evaluate techniques  for
achieving a separation of more from less contaminated fractions.  Parameters  that may
reflect a variation in contaminant concentration within a soil include particle size, particle
density, particle magnetic susceptibility and particle surface properties such as
hydrophobicity. Consequently determining the distribution of contamination on the basis of
these parameters will enable the use of appropriate separation techniques from mineral
processing that utilise these same parameters.

       The full range of investigations will include preliminary evaluation at laboratory
scale of soil from three sites contaminated with heavy metals, PAH's and hydrocarbons to
determine the contaminant distribution and evaluate potential treatment methods. This  will
then be followed by further investigations to determine a possible treatment route for soil
from one of the three sites and will culminate in a pilot scale operation to prove the
treatment route with up to 50 tonne of soil. Much of the strength of the program is that it
will be evaluating the successes of different processes such as gravity based separation
techniques, magnetic separation and froth flotation.

       This paper describes the laboratory characterisation techniques: size fractionation,
density fractionation, magnetic separation  and froth flotation.  It then presents selected
results obtained from the characterisation of one specific soil and discusses a potential
treatment route. The soil examined  is from the site of a demolished gas works where the
main contamination was determined as 807 mg/kg lead, 32 mg/kg arsenic, 205 mg/kg
polyaromatic hydrocarbon (PAH) and 637.0 mg/kg petroleum hydrocarbon.
                                         192

-------
METHODOLOGY
       WSL has adapted a number of associated laboratory techniques to characterise the
distribution of contaminants within the soil and so permit evaluation of the potential for
separation of contaminants from contaminated materials. These techniques are as follows:
       size separation
       size and specific gravity separation
       separation by froth flotation
       magnetic susceptibility assessment.
       Size Separation.  This procedure is designed to provide an indication of the
contaminant distribution and concentration within the various size fractions of the soil. The
method adopted for grain size determination involves screening,  or sieving, with water at a
variety of sizes and the use of a hydrocyclone to achieve separation at the finer size.

       Size and Density.  This procedure provides a matrix of varying size and density
values within which the distribution and concentration of contaminants can be determined.
To ascertain whether there is any preferential partition to specific gravity fractions a series
of so called "sink-float" tests are carried using  heavy liquids on a number of size fractions.
Contaminated material is suspended in these heavy  liquids.  Particles less than the density
of the liquid float whilst those heavier sink.  The densities of the liquids can be altered by
addition of a diluent.  Tests are carried out at a number of densities to ascertain
preferential partitioning. The heavy liquids used for separation are generally organic
liquids such as bromoform (density 2.89 g.ml'1), tetrabromoethane (density 2.964 g.ml'1)
and di-iodomethane (density 3.325 g.ml'1).

       Froth Flotation.  This procedure is designed to evaluate the distribution and
concentration of contaminants within certain predetermined size ranges that are naturally
hydrophobic, or can be made so, and can therefore be removed  from the soil by froth
flotation procedures.  The tests are conducted on the < 2.00 mm material with an option of
removing, or otherwise, the finest material (< 0.010 mm) prior to  testing. The flotation tests
are carried out as batch processes in laboratory flotation machines.  During testing,
material is slurried for a given period with a variety of reagents, using acid and alkali to
control and maintain the desired pH. Initially the naturally hydrophobic materials are
floated by adding only a frothing agent.  Passage of air through the slurry then produces a
froth (Froth 1) which is scraped off into a collecting vessel.  Once the froth no longer
contains any soil material the air is switched off.  A quantity of collector is then added to
make further particles hydrophobic and the air turned on to produce a second froth (Froth
2).  Further tests can be carried out with different reagent types and concentrations, and
different operational conditions such  as time for collection/conditioning or pH to produce
further froths (e.g. Froth 3). The final non-floating material is called the residue.

       Magnetic Separation.  This test is designed to evaluate the distribution and
concentration of contamination associated with ferromagnetic,  paramagnetic and non-
magnetic particles within pre-determined size ranges of the soil.  The laboratory separator
                                            193

-------
used for the magnetic characterisation is the Franz Isodynamic Magnetic Separator. Tests
are carried out over a number of field strengths which are varied by altering the DC current
around the magnets.  Initially all material is run through the apparatus at low current and
the products collected. The non-magnetic product is then repassed at a higher current and
so on.  In this way a number of products are obtained. It is worth noting that the first run is
carried out with no current ie residual magnetism. The  magnetic products from this first
run are the ferromagnetics. Particles separating  at higher currents are either small
amounts of ferromagnetics in a larger non-magnetic particle, or paramagnetic particles.
RESULTS
       Only selected results are presented in order to indicate a potential treatment route
for the gas works soil.  Table 1 indicates the distribution of contamination determined by
size density characterisation of the gas works soil.  Three significant aspects can be seen:

* the > 2.00 mm size that represents 29 % by weight of the soil is relatively free of heavy
metals (22 mg/kg of arsenic and 299 mg/kg of lead) and PAH (51  mg/kg) but does contain
80 % of the petroleum hydrocarbons in the soil (1900 mg/kg)

* the < 0.010 mm size that represents 24 % by weight of the soil is relatively free of
petroleum hydrocarbons (37 mg/kg) but does contain 34 % of the  arsenic (45 mg/kg), 38
% of the lead and 77 % of the hydrocarbons (647 mg/kg) in the soil

* within the remaining size ranges (2.0-0.5 mm, 0.5-0.075 mm, 0.075-0.045 mm and 0.045-
0.010 mm) the two heaviest density fractions ( 2.8-3.3 p and > 3.3 p) represent
approximately 3.5 % by weight and contain 11 % of the arsenic, 25 % of the lead and an
unknown quantity of the PAH and petroleum hydrocarbons.
                                           194

-------
O
X
o
LU

D
03


m
m

S
o
o
III

m








£
E

'13
«*3
C
8
3
c5
c
E
CO
c
<3








O)
o
5
'55
§
Q

03


vO
E
3
0 0
Q. c
8
0
x
u
c
o
O




^

_
CO
03
1

o^
o
O
CO
< 0
o
0


^*


Q.


E

CO
o
o
UJ



o
o
05
00
CO
CO


to



CM
o
T~



Cn
CO
05
T~



CM

00
en
00



Is

a
o
CM
A
CO
05
CO




OO
O5
CM
O
CO
in


12



T- 
T— T—
CM
CO T— ^ O OO
00 O5 CO CO CO
T— CO T— T- 00




CM in oo CM oo
CM T-  Tf CO CO N.
T- CO CD CO CO
CM CO T- CM ^}-
T—



^" CO ^. CO ^J"
CM -r- 
O T- O f- CO




T— OS ^ OO CO
r^ o en CM ^3-
en oo oo r^ m
T- in
in as g- co CD
^* O5 TT O5 00
O T- O O CO




CO 00 CM CO CM
"*• 1— CO O CM
CM
ocM£jnS
d ^ o' cs in

CO OO
0 ^ °? CO
CM °. °°. 00 =
V CM CM A CO

in in
O 0
0 0
8
CO




CM
,_
CM


in
in



CO O CD 00 0
T- T- f-* CO M-
O IO Q T- t*.




in CM oo rj- o
CM r** oo o ^
co •* j- CM in
T- CM CO T-
in ^- oo CM en
T- O t^- O O)
0 CO 0 CM CO
T— T—



CM CM en r«. r^.
CM CO CO CD CO

T- N- co in i—
CM CM CM CO T-
o co o a en

oo oo

CM °. °°. CO =
V CM CM A CO

in o
So
d d
9
,_:




CO
OJ
rf
f:


^
CD


CM
|s^
CO



CM
CD
CM
fp
^^
CO
CO



in

T-
^



"ra

o
o
V

o
o



g

o
a


§
CM



o
o
1~


§

a
o



CM
CO


o
o
T—


To


03
                                               195

-------
CONCLUSIONS
       The results from the laboratory characterisation of this soil indicate that there is
some potential for the physical treatment of the soil to produce concentrates of soil
contamination.  One possible option for a treatment route is outlined below:

1 - screening out of all material > 2.0 mm.  The removal of this size would remove 29 % of
the original weight as a material relatively free of heavy metals and would be necessary
before any further processing is undertaken.  However, if this fraction is rich in
hydrocarbons as the results suggest then it would then be available for further treatment
(for example by a washing process or by biological techniques)
2 • hydrocyclonlng to remove the < 0.010 mm material. This would remove 24 % of the
weight of the soil which would contain 77 % of the PAH's, 38 % of the lead and 34 % of
the arsenic. Any petroleum hydrocarbons in this fraction would also be removed
3 - gravity separation to remove the > 2.8 g.ml'1 density fractions from the 2.00-0.010 mm
sized material.  This would remove a further < 4 % of material containing a further 25 % of
the lead and 11 % of the arsenic.

       Four products would be produced by this process:

* a relatively clean coarse (> 2.00 mm) material but with petroleum hydrocarbons that may
degrade naturally or may need to be further treated

* a relatively contaminated fine (<0.010 mm) material

* a relatively contaminated high density medium sized (2.0-0.010 mm) material

* a relatively clean treated material representing 43 % by weight of the original soil and
containing 26 mg/kg arsenic, 493 mg/kg lead, 75  mg/kg  PAH and 274 mg/kg petroleum
hydrocarbons.

        The full range of work so far undertaken indicates that some contaminants within a
soil exhibit a selective distribution that is reflected in parameters such as soil particle size,
density, magnetic susceptibility and surface property differences.  In some  instances these
 selective distributions are sufficient to suggest that mineral processing equipment has
 potential for the treatment of the soil by separating grossly contaminated from less
 contaminated particles.

 For More Information: Mary K Stinson. Risk Reduction Engineering Laboratory. USEPA.
 2890 Woodbridae Avenue (MS-1041 Edison.  New Jersey 08837-3679: Tel: (908^ 321-
 6683.
                                            196

-------
      HYDRAULIC AND IMPULSE FRACTURING TO ENHANCE  REMEDIATION

                                        Larry Murdoch
                    Mark Kemper, Allen Woif, Elizabeth Spencer, Phil Cluxton
                       Center for GeoEnvironmental Science and Technology '
                                     University of Cincinnati
                                      5995 Center Hill Rd.
                                     Cincinnati, Ohio 45224
                                         513-569-7847

        Meager rates of fluid flow are a major obstacle to in situ remediation of low permeability soils.
This paper describes methods designed to avoid that obstacle by creating fractures to increase the
effective permeability and change paths of fluid flow in soil. The most well-known method, hydraulic
fracturing, involves injecting fluid  at modest rates and pressures during several tens of minutes. A new
method, termed impulse fracturing, involves injecting fluid at fast rates and great pressures during
several  tenths of a second.

HYDRAULIC FRACTURING
        Hydraulic fracturing has been used for more than 50 years to stimulate the yield of wells
recovering oil from rock at great depth, and it has recently been shown that hydraulic fracturing will
stimulate the yield of wells recovering liquids and vapors from soil at shallow depths.  Recent efforts
have focused on conducting field  evaluations of air flow and vapor extraction using hydraulic fractures.
        The utility of hydraulic fractures is by no means limited to well stimulation.  Relatively large
volumes of solid compounds can be delivered to the subsurface as granular materials filling hydraulic
fractures. My colleagues have developed a solid compound  (Davis-Hoover and others, 1991), which
slowly releases oxygen, that we inject into hydraulic fractures along with slowly releasing nutrients to
stimulate in situ aerobic biodegradation of organic compounds in soils. In addition, it is feasible to fill
hydraulic fractures with metal catalysts, which Gillum and Burns (1992) have proposed as a method of
degrading a wide range of halogenated organic compounds.
Method
        Hydraulic fracturing begins by injecting fluid into a borehole at a constant rate until the pressure
exceeds a critical value and a-fracture is nucleated. Coarse-grained sand, or some other granular
material, is injected  as a slurry while the fracture grows away from the borehole. Guar gum gel, a
viscous fluid, is commonly used to facilitate transport of the sand grains into the fracture. Injection
pressure of 200 to 400 kPa (30 to 60 psi) is typically required to initiate a hydraulic fracture at 3 m depth,
but propagation only requires 100 to 170 kPa (15 to 25 psi).  After pumping, the fracture is propped open
by the sand and the guar gum gel is decomposed by an enzyme added during injection.  In fine-grained
soil, we generally create hydraulic fractures beneath casing that is driven to depth with a hammer.
Lateral pressure of the soil seals the casing during injection and the casing can  be driven to greater
depth to create another fracture.  Stacks of gently dipping hydraulic fractures have been created with
vertical  spacing of 15 cm to 30 cm using the driven casing method.
Fracture Form
        Details vary considerably, but hydraulic fracturing generally produces a single parting, with
multiple fractures requiring repeated operations. The soil parting typically takes on one of two forms: a
steeply  dipping feature that climbs rapidly to the ground surface, or a gently dipping feature that can be
10 m or more in maximum dimension.  The gently-dipping form is equant to slightly elongate in plan and
dips toward the parent borehole.  In some cases, the fracture is nearly flat-lying in the vicinity of the
borehole and the dip increases to approximately 20° at some distance away (Fig. 1), whereas in other
cases the fractures appear to maintain a roughly uniform dip from the borehole to the termination.
        In nearly every case, the fracture has a preferred direction of propagation so that the borehole is
off the center of the fracture. The preferred direction of propagation is commonly related to the
distribution of vertical load at the ground surface, with the fractures propagating toward regions of
diminished vertical load. Beneath sloping ground, therefore, it is possible to anticipate the preferred
                                              197

-------
direction; it is typically downslope. Beneath level ground, we have used vehicles to artificially load the
ground surface and influence the propagation direction away from the vehicles.
       Hydraulic fractures have been created between 1.2 and 9.1 m below the ground surface during
our research. All of them have been in silty clay, and most have  been in glacial drift that probably was
overconsolidated.  Maximum dimensions increase with depth, but are in the range of 6 to 11 m.
Between 0.14 and 0.34 m3 (5 to 12ft3  with 1ft3 = 100 Ibs) of sand are injected into a fracture.  The
average thickness of sand in a fracture ranges from 0.5 to 1 cm.  The largest fracture that has been
characterized was 17 m in maximum dimension, the most voluminous contained 1  m3 of sand, and the
thickest was 2.5 cm.

IMPULSE FRACTURING
       Creating useful hydraulic fractures is currently limited to overconsolidated soils, so to address
this limitation we recently have been developing a method of creating sand-filled fractures with high
pressure impulses of water.  This investigation is in its early stages and the results  are preliminary;
nevertheless, we have included this section because those results indicate that impulse fracturing should
be a versatile and effective tool in low permeability soils.

Method
       Impulse fracturing involves deforming soil with pulses of water generated by the Hydraulic
Impulse Technique (HIT).  The HIT device (Kinnan, 1986) is a high-pressure hydraulic intensifier that
creates a pulse of fluid of approximately 0.5 L, which is discharged through a narrow nozzle in less than
300 ms.  Injection pressure is 60 MPa to 70 MPa (8500 to 10000 psi) and fluid velocity is on the order of
150 to 460 m/s (500 to 1500 ft/s).
       Granular material, such as sand, is introduced into the fluid pulse and carried into the
subsurface.  Configurations have been evaluated that allow granular solids to be emplaced in  impulse
fractures while the nozzle is pointed vertically at the ground surface or below an open borehole,  or while
the nozzle is inclined to the axis of either a vertical or a directional borehole.

Fracture Form
       The general deformation created by a single impulse includes a cylindrical hole and fractures
either parallel or normal to the axis of the hole.  The results of two early field tests; one where the nozzle
was directed vertically downward at the ground surface, and another where the nozzle was inclined to the
axis of a directional borehole characterize some of the capabilities of this technique.  The vicinity of the
each test was excavated to map the resulting deformation.
       During the vertical tests, five pulses from the HIT device  created an axial hole  as much as 5 cm
in diameter and 130 cm deep.  Fractures were created parallel to the axis and  extending along the full
length of the hole. They were 30 cm long and extended into soil on either side of the hole (Fig. 2).
Coarse-grained sand to gravel (0.5 to 0.8 cm diameter) was injected with each pulse, completely filling
the axial hole and filling most of the fracture as well.
       The other test consisted of firing pulses as a nozzle was pulled along a directional (nearly
horizontal) borehole. The nozzle was inclined to the borehole axis and oriented in the horizontal plane,
and eight pulses were spaced along the borehole (Fig. 3). A bentonite slurry was injected to mark the
resulting fracture.  This resulted in an open cavity 2.5 to 5 cm thick within 30 cm of the borehole. A
bentonite-filled fracture extended from the open cavity to approximately 1.2 m from the borehole.

RESULTS
        Pilot-scale demonstrations of the effects of hydraulic fracturing have recently been evaluated by
the USEPA SITE program and the results will be presented in a soon-to-be-available report.  Highlights
of two of the demonstrations will be described below, with the details available in the SITE report.

Air Flow Study
       A preliminary vapor extraction test at the USEPA Center Hill Research Facility has recently been
conducted using fractured and conventional wells in uncontaminated ground. The site  is underlain by
stiff, brownish-gray, silty-clay till, which contains rock fragments below 3 m. Hydraulic fractures were
created at depths of 1.5, 3 and 4.6 m and extend to radial distances of 3 to 4.6 m,  according to the
results of uplift monitoring. The fracture at 4.6 m was not used during the tests described here.  A multi-
                                               198

-------
 level recovery well was completed with separate casing for each fracture. Another well was created in
 unfractured ground as a control using the same completion techniques as described above.
        Vapor extraction was conducted at the fractured and unfractured wells using conventional
 blowers. This site was uncontaminated, so the purpose was to examine the effects of hydraulic fractures
 on air flow in low permeability soil.
 Distribution of suction head
        Suction diminishes abruptly with distance away from the conventional well, whereas it decreases
 relatively gradually with distance from the fractured well. For example, 0.25 kPa (1 inch H2O) of suction
 head occurs between 0.6 and 0.9 m from the conventional well, whereas it occurs 7.6 m from the
 fractured well. Suction in soil typically increased after rainfall.
 Well yield
        The yield of the conventional well with 9.5 kPa (38 inches H,O) of suction remained roughly
 constant at 140 ml_/s ( 0.3 cfm) throughout the 40 days of testing.  In contrast, the yield of the well
 intersecting hydraulic fractures was between 1.4 and 2.1 Us (3 and 4.5 cfm) for most of the test, roughly
 an order of magnitude more than from the conventionai well. Yield from the fractured well decreased
 abruptly following rainfall, and then increased after water was removed from the recovery well. This
 dependence of yield on rainfall was seen at all the sites where we conducted tests, although the
 magnitude of the effect was diminished by installing a system to regularly dewaterthe vapor extraction
 wells.

 Theoretical modeling
        A theoretical steady-state analysis of the flow of air to a fracture was conducted using AIRFLOW,
 a 3-dimensional code developed by Craig Joss and Art Baehr. We assumed that the fracture was
 shaped like a flat-lying circular disk, 1.5m below the ground surface, and that it was 3 m in radius,  0.6
 cm thick and filled with coarse-grained sand  (the actual fracture tapered near its periphery and was
 slightly greater in radius). The  ground surface was assumed to be at atmospheric pressure and an
 impermeable layer was assumed to be at a depth of 6.1 m. The permeability of the till was estimated at
 10'9  cm2, based on calibration simulations using data from a conventional well screened at 1.5 m.  The
 permeability of the fracture was unknown, so we evaluated a wide range of permeability to compare to
 the field observations. The results of the simulation are shown along with field data obtained by applying
 25 kPa (100 inches H20) of suction head to the fracture at 1.5 m depth.
        Both well yield and the distribution of suction are strong functions of the permeability of the
 fracture (Fig. 4). The suction head observed in the field is predicted approximately by a fracture whose
 permeability is between 3  x W6 and  5 x 10"6 cm2.  The observed suction is slightly greater than
 predicted at radial distances greater than 4.6 m, but the uplift data indicate that the fracture is probably
 bigger than the assumed 3 m.  The yield observed in the field was approximately 3.3 L/s (7 cfm), which
 is  predicted for a fracture whose permeability is 7 x 10"6 cm2.
 Solvent Recovery Study
        A pilot-scale test of hydraulic fracturing during a vapor extraction remediation was conducted at
 a site in the Chicago area containing solvents that were spilled during filling of a storage tank.  Well
 yields, pressures, and contaminant concentrations were obtained approximately from two fractured wells
and from one conventional well three times a week for 21 weeks beginning in July, 1992.  The wells
 intersecting hydraulic fractures yielded contaminants at roughly steady rates throughout the test (Fig. 5).
 Mass yield estimated at 0.10 kg/day was observed from RW3 early in the test, however, at later times
 the yield diminished to 0.08 kg/day at RW3 and similar yields occurred from RW4 throughout the test.
 The  yield from RW2 was typically on the order of 0.003 kg/day, although during one week-long interval it
 apparently increased to 0.06 kg/day.

 CONCLUSIONS
        Hydraulic and impulse fracturing techniques are capable of creating sand-filled fractures that
 increase the rate of fluid flow in soil. During  hydraulic fracturing, the most widely known of the two
 techniques, fractures are created by injecting sand-laden slurry at modest rates and pressures. In
                                              199

-------
contrast, during impulse fracturing fluid is injected at high rates and pressures to crack subsurface

       Field tests show that hydraulic fractures can increase the volumetric yield of vapor extraction
recovery wells by more than an order of magnitude. At a site contaminated by solvents, the mass yield
of hydraulically fractured wells is 10 to 30 times greater than the yield from conventional wells.
Moreover, the suction is distributed over a wider area using hydraulic fractures compared to conventional
wells. At one site underlain by silty clay till, a reference suction head of 2.5 cm was observed 0.6 to 0.9
m from a conventional recovery well, whereas it was observed 7.6 m from a well intersecting hydraulic

 ra  uresiihose results can be predjcted using established theoretical analyses of flow of a compressible
fluid to a permeable layer.

ACKNOWLEDGEMENT                                                    .  .
       This work has been supported by the both the SITE Program and the Municipal Solid Waste
Research Management Branch, Soils and Residual Section of the Risk Reduction Engineering
Laboratory of the USEPA under contract #68-09-0031 . We thank Mike Rouher, Naom. Barkley, and
Wendy Davis-Hoover with the EPA for their guidance and support. This paper has not been reviewed by
the USEPA however so no endorsement should be implied. Frank Kinnan, director of Underground
Research, developed the HIT device and we appreciate his cooperation with our investigation.  Craig
Joss and Art Baehr provided us with a copy of their code AIRFLOW before it was officially  released.
Ron Hess and Elliot Duffney of the Xerox Corporation provided access to their site and data.
 Davis-Hoover, W.J., LC. Murdoch, S.J. Vesper, H.R. Pahren, O.L Sprockel, C.L. Chang, A. Hussain,
        W A Ritschel. Hydraulic fracturing to improve nutrient and oxygen delivery for in situ
        bioreclamation  In situ Bioreclamation, Butterworth-Heinemann, Boston, 67-82, (1991).
 Gillham RW andDR  Burris. In situ treatment walls -Chemical dehalogenation, denitrification, and
        bio'augmentation. Proceedings of Subsurface Restoration Conference, Dallas, Texas, June 21-
        OA fifi RR  ("\ QQ?^
 Kinnan F Techniques for establishing inground support footings and for strengthening and stabilizing
       ' the soil at inground locations.  U.S. Patent 4621 950. Nov. 1 1 , (1 986).

 Information: Wendy Davis-Hoover, Ph.D., USEPA, RREL, Work Assignment Manager. 513-569-7206
  Figure 1. Idealized hydraulic fracture created at shallow depths in overconsolidated silty clay.
                                                200

-------
      section view
Figure 2. Fracture created by firing downward with
          the HIT device. Zone A is fracture filled
          with gravel to thicknesses of 0.2 to 0.4
          inch.  Zone B is sparsely filled with gravel.
          Zone C is filled with water, but no gravel.
                                                                       plan vie
                                           Figure 3.  Deformation created by a series of pulses
                                                     of bentonite slurry at various points along a
                                                     horizontal bore. Nozzle was oriented
                                                     horizontally and produced an open
                                                     horizontal cavity and fracture.  The cavity
                                                     was 1 to 2 inches high and approximately
                                                     12 inches deep.
                      Suction >l veil:  100 Inches
                      Till Perm««bllily:  10-' cm*
                 5    10   16   ZO   25   30
                   Distance from well (H)
               Cl
               Dtpth: 5 II
               R.dlui: 10 It
               Ap«rtur»; 0.2S Inch
    ID'7  ID"4  ID"3  10"'  10'3

Froctwre Permeebilily (cm )
                                                                                                      150
                                                                 Figure 5. Solvent recovered as a function of
                                                                           time from fractured wells (RW3
                                                                           and RW4) and from a conventional
                                                                           well (RW2).
    Figure 4a.  Suction as a function of radial distance at a depth of 4 ft, showing field data and
               theoretical results for various permeability of fracture. 4b.  Yield as a function of
               fracture permeability.  Yield without a fracture is 0.23 cfm.
                                                      201

-------
      AN INTEGRATED CHEMICAL AND BIOLOGICAL TREATMENT (CBT) SYSTEM FOR SITE
                                        REMEDIATION

                     Robert L. Kelley*, Vipul Srivastava* and Naomi P. Barkleyt

                                  'Institute of Gas Technology
                                     3424 South State St.
                                       Chicago, II 60616
                                        (312) 949-3809

                        tU.S. EPA Risk Reduction Engineering Laboratory
                                26 West Martin Luther King Drive
                                     Cincinnati, OH 45268
                                        (513) 569-7854

Introduction

       Institute of Gas Technology (IGT) has developed treatment technologies which enhance
bioremediation by integrating   chemical and biological treatment (CBT) processes for remediation of
contaminated soil and sludge. The treatment system combines two remedial techniques: 1) chemical
oxidation as the pre-treatment, and 2) biological treatment using aerobic and anaerobic biosystems
either in sequence or alone, depending on the waste.  The CBT process uses mild chemical treatment
to produce intermediates that are biologically degraded, reducing both the cost and risk associated with
the more severe process. The CBT process can be applied  to a wide range of organic pollutants,
including alkenes, chlorinated alkenes, aromatics, substitute  aromatics, and complex aromatics.
Applicable matrices include soil, sludge, groundwater, and surface water.

       In chemical treatment/oxidation, metal salts and hydrogen peroxide are used to produce the
hydroxyl radical, a powerful oxidizer.  The reaction of the hydroxyl radical with organic contaminants
causes chain reactions, resulting in modification and degradation of organics to biodegradable and
environmentally benign products. These products are later destroyed in the biological step.  The
contaminated material is treated chemically by a chemical reagent degrading the organopollutants to
carbon dioxide, water, and more biodegradable, partially oxidized intermediates. In the second stage of
the CBT process, biological systems are used to degrade the hazardous residual materials, as well as
the partially oxidized material from the first stage. Chemically treated wastes are subject to cycles of
aerobic and anaerobic degradation if aerobic or anaerobic treatment alone is not sufficient.

       This project is a joint U.S. EPA, IGT, Gas Research  Institute (GRI), and private industry funded
program that has been evaluating this technology primarily for PCB- and PAH-contaminated soils. The
expected result of this program will be a field demonstration  of the CBT technology at a gas industry
waste site. The  research plan is aimed at demonstrating the superiority of the integrated CBT process
over other established or developing processes for remediation of soils and gas condensate
contaminated with PCBs and PAHs.  The principal objectives of this program are to identify 1) key
engineering and process parameters/conditions, 2) key economic factors, 3) strengths of the CBT
process, and 4) possible limitations of the CBT process. Using optimal treatment conditions, PCB-
contaminated soils and gas condensates are first treated with Fenton's reagent and then bioremediated
under anaerobic and/or aerobic conditions to determine the advantage of chemical pretreatment to
bioremediation.  Contaminated soils contain 100 to 1,000 ppm of the contaminants, and the treatment
end point goal is below 5 ppm.  Finally, a preliminary economic analysis is being conducted to
determine the overall feasibility and to determine if the  project should proceed to a pilot-scale
demonstration.
                                             202

-------
Methodology

       PCB Stardards A mixture of 10 congeners, the Dry Colors Manufacturers Association (DCMA)
PCB mixture (Supelco, Supelco Park, Pennsylvania) was used as a representative mixture of PCBs in
some experiments. A PCB master standard was made for use in other experiments using commercially
obtained Aroclor 1242, 1248, 1254, and 1260 standards (Ultra Scientific, North Kingstown, Rhode
Island).

       Biological Culture  Aerobic cultures were isolated and purified on TGY medium (tryptone
glucose yeast extract medium). After growth, these cultures were evaluated for their ability to grow in
the presence of biphenyl as the sole source of carbon.  All grew in APS broth in the presence of
biphenyi except one culture.

       PAS broth was prepared by adding 77.5 ml of PA concentrate (composition per liter of dH2O:
K,HPO4, 56.77g; KH2PO4, 21.94; and 27.61 g) and 50 mg of yeast extract to 910 ml of glass-distilled
water.  After autoclaving and upon cooling, 10 ml of sterile PAS 100 X salts (composition per liter of
dH2O:  MgS04, 19.5g; MnSGyH2O, 5g; FeSCy7H2O,  1g; and CaCI2-2H2O, 0.3g along with several
drops of concentrated H2SO4 per liter to prevent precipitation of basic salts) was added.  Biphenyl was
supplied by the addition of sterile, molten biphenyl to  the autoclaved medium.  For PAS plates, 15g of
agar were added per liter of PAS broth.

       Anaerobic Cultures were cultivated from sediments, muds, and sludges under anaerobic
conditions suitable for the development of anaerobic microbial communities.  The inocula sources
include natural systems that are pristine and those that are contaminated with PCBs and/or other
organic materials.

       The enrichments are prepared using a modified Hungate technique.1"3  The culture medium is
revised anaerobic mineral  medium (RAMM) as described by Shelton and Tiedje.4  Substrates added to
act as an election source include cellulose, glucose, yeast extract, ethanol, and acetate (final
concentration set at approximately 0.2 g/L).  The anaerobic consortia  obtained will  be assayed for the
ability to reductively dehalogenate various congeners often comprising PCB mixtures.

       14C-Labeled Experiments In some experiments PCB degradation was monitored using  14C-
labeled compounds using previously described methods.  14C-labeled  compounds are added to the
clean soil in a 125-ml flask sealed with a stopper. H2O2 is added to the flask via a syringe (or over an
extended period using a syringe pump) while the contents of the flask are stirred or shaken at 150  rpm
for the desired length of time (up to 16 hours).  The headspace of the flask is continuously flushed with
air at a flow rate of 10 ml/min. The UCO2 produced by the oxidation reaction is swept out of the flask
by the air and trapped in a basic ethanolamine solution (Carbosorb TM; Packard Instrument Company).
The trap can be removed at desired intervals and the solution combined with a liquid scintillant
(Permafluor V TM; Packard Instrument Company) and counted in a liquid scintillation counter (LSC)  to
quantify the amount of CO2 produced.  A "no-treatment control" is tested for comparison.

       PCB Analysis Quantitative and qualitative PCB determinations were made using a Hewlett
Packard 5890-GC (Hewlett Packard Co., Palo Afto, California) equipped with a 63Ni electron capture
detector (ECD) (Hewlett Packard model no. 19233), a HP-7673A autosampler, and a HP 3396A
integrator.  A Hewlett Packard Ultra 2, capillary column, 50 m long with an internal diameter of 0.2  mm
and 0.33 urn film thickness was used.  Chromatographic conditions include:  column head pressure, 60
kPa, carrier gas,  helium; linear velocity, 30 cm/sec; splitless injection; injector temperature, 220°C;
detector temperature, 350°C; detector make-up gas, Ar/CH4 (95%:5%). The temperature program
consisted of an initial column temperature starting at  100°C for 1 minute; rate A, 20°C/min to 150°C;
rate B, 3°C/min to 300°C, held for 6.5 minutes.
                                             203

-------
Results

       PCS Degradation

       Preliminary studies of PCB degradation with MC-Iabeled PCBs in aqueous suspension indicate
that 30-40% of 2,2',4,4'-biphenyi in mineralized to CO2. Equally as important, 30-35% of
2,2',4,4'-biphenyl is modified to a water soluble, more easily degradable product. These studies have
shown that the chemical pretreatment works best with the following conditions: 1)10 mM Fe++, 2) 5%
H2O2 and 3) pH = 4. As expected the higher the chlorination of the biphenyi, the lower the amount of
mineralization. Chlorine groups in the 3-position inhibited mineralization by 50% when compared with
the 2-position.  Chlorines in the 4-position appear to be slightly more susceptible (-8%) to chemical
degradation than chlorines in the 2-position. This difference is statically significant as measured by +
the standard  deviation. This study indicated that the majority of PCB mineralization occurs in the first
45-60 minutes of the reaction in aqueous suspension.

       With unlabeled DCMA mixture  of PCBs, as much as 99.9% of some PCB congeners were
removed, and 88.8% of total PCBs were degraded with 5% H2O2 (Table 1). The degradation of PCBs
increases as the reagant concentration increases; and PCBs  with less than 5  chlorine groups are more
susceptible than PCBs with greater than 5 chlorine groups. This pattern of attack was predictable
based on chemical  reactivity. This pattern also complements the anaerobic and aerobic biodegradation
of PCBs in that  it aggressively attacks  the intermediate biphenyi compounds with 3-5 chlorine groups.
              Table 1. SUMMARY OF CHEMICAL DEGRADATION OF DCMA MIXTURE BY
                              FENTON'S REAGENT
              Reagent
              Concentration, %
Total
% Degradation

     < 5-Ci
                 1
                 5
                10
 80.3
 84.5
 88.8
     98.1
     99.9
     99.7
>5-Cl

  63.8
  66.4
  76.0
        Currently, experiments with actual PCB-contaminated sludges have shown significant chemical
degradation of PCBs. As much as 85% of some mono-chlorinated congeners can be removed in a one
week incubation with aerobic cultures, and anaerobic cultures have shown great promise.

        PAH Degradation

        Previous bench-scale studies conducted with a variety of PAH contaminated soils and the field-
scale experiments conducted with a PAH-contaminated soil have shown that the CBT process improves
the rate as well as the extent of PAH removal from these "weathered" soils.  In this previous research,
a field demonstration  has been initiated after   promising treatability tests were established.  This field
experiment was devised to determine the effectiveness of the integrated process using a prepared-bed
land treatment system. Sixteen equal-sized field plots  (1.2 X 3.6 m) containing 15 to 20 cm in depth
are being tested under four major, different conditions. Data from this field demonstration has become
available and the results  are very encouraging.
                                           204

-------
       The field experiment showed that the integrated treatment system resulted in about 50%
greater removal of total PAHs and 90% greater removal of carcinogenic PAHs than those of
conventional biotreatment. The integrated system's results exceed the treatment goals which had been
established and were reached in 42 days.
Conclusions
       PAH Degradation

              Integrated Chemical Biological Treatment (CBT) process has been proven effective in
              field tests with PAH-contaminated soils.

       •      CBT process can improve the rate, as well as the extent, of PAH degradation.

              CBT process is more effective than the conventional bioremediation for soils that
              contain high PAH levels, high silts and clay, or both.

       PCB Degradation

       •      PCBs can be dechlorinated, as well as mineralized, by Fenton's reagent.

       •      Anaerobic dechlorination was successfully achieved with PCB mixtures.

              An integrated biological (anaerobic)—chemical (Fenton's)—biological (aerobic)
              treatment offers potential improved removal of PCB.
References

1.      Balch, W. E. and.Wolfe, R. S., "New Approach to the Cultivation of Methanogenic Bacteria: 2-
        Mercaptoethanesulfonic Acid (HS-CoM)-Dependent Growth of Methanobacterium Ruminantium
        in a Pressurized Atmosphere." Appl. Environ. Microbil. 32. 781-91  (1976).

2.      Bryant, M. P., "Commentary of the Hungate Technique for Culture of Anaerobic Bacteria," Am.
        Jour. Clin. Nutri. 25:1324-28 (1972).

3.      Hungate, R. E., "A Roll Tube Method for Cultivation of Strict Anaerobics," Methods Microbiol. 3,
        117-32(1969).

4.      Shelton, D. R. and Tiedje, J. M., "General Method for Determining Anaerobic Biodegradation
        Potential," ADD!. Environ. Microb. 47. 850-57 (1984).
 For Further Information

        EPA Project Manager:
        Naomi Barkely
        U.S. EPA
        Risk Reduction Engineering Laboratory
        26 West Martin Luther King Drive
        Cincinnati, OH  45268
        513-569-7854
Technology Developer Contact:

Robert Kelley
Institute of Gas Technology
3424 South State Street
Chicago, IL  60616-3896
312-949-3809
Fax: 312-949-3700
                                         205

-------
           SITE DEMONSTRATION OF THE CHEMICAL WASTE
     PO*WW*ER  EVAPORATION AND CATALYTIC OXIDATION SYSTEM
                       Randy A. Parker
               Environmental Protection Agency
                    26 West M.L. King  Drive
                    Cincinnati, Ohio 45286
                        (513)  569-7271
INTRODUCTION
     The PO*WW*ER   system developed by Chemical Waste
Management, Inc.  (CWM), treats a variety of wastewaters by
reducing the volume of aqueous waste and catalytically
oxidizing volatile  contaminants.  The  system consists
primarily of  (1) an evaporator that reduces the influent
wastewater volume,  (2) a  catalytic  oxidizer that oxidizes the
volatile contaminants  in  the vapor  stream,  (3)  a scrubber
that removes acid gases produced during oxidation,  and  (4)  a
condenser that condenses  the vapor  stream leaving the
scrubber.

     The technology was evaluated under the Superfund
Innovative Technology  Evaluation (SITE) program in  September,
1992 at the CWM Lake Charles Treatment Center  (LCTC)  in Lake
Charles, Louisiana.  CWM's Engineering and Technology group
designed and built  the pilot-scale  PO*WW*ER M process  at the
Lake Charles Treatment Center.  CWM's  sister company, ARI
Technologies, Inc.  (ARI),  is responsible for marketing,
engineering, and manufacturing future  PO*WW*ER   process
units.  The Lake-Charles  Treatment  Center has facilities that
include a hazardous waste landfill,  a  high-capacity
stabilization unit, and drum managing  and decanting
facilities.
  Figure 1. Schematic diagram of the PO*WW*ER™ pilot plant.

                              206

-------
METHODOLOGY

       Evaporation and catalytic oxidation are the core
processes of this technology.   Figure 1 shows the flow
diagram of the PO*WW*ER  system pilot plant.  Aqueous waste
to be treated by
the PO*WW*ER  system pilot plant is stored in a stainless
steel feed tank.  The feed tank is equipped with an agitator
mounted on the top ofthe tank to mix additives into the feed
waste and to keep solids in the feed waste in suspension.  To
control foaming in the evaporator,  an anti-foam agent can be
added to the waste in the feed tank or sprayed directly into
the evaporator.   The pH of the waste is also monitored and
adjusted in the feed tank.  Once the feed is suitable for
treatment, it is pumped to the evaporator.

  The evaporator consists of three main components:  heat
exchanger, vapor body, and entrainment separator.  As feed
waste is pumped to the evaporator,  it combines with heated
process liquor.   The liquor waste is then further heated in a
vertical shell-and-tube heat exchanger which has two passes,
four tubes per pass.  Heat is supplied by steam generated in
a boiler.  Liquid waste flows through the tube side of the
heat exchanger and steam passes on the shell side.  In the
heat exchanger,  the liquid is heated to the boiling
temperature, although boiling does not occur in the tubes
because of the back pressure in the system.

     After passing through the heat exchanger, the liquid
waste enters the vapor body where boiling occurs and vapor is
released.  The vapor consists of mostly water and volatile
components, both organic and inorganic.  The liquid level in
the vapor body is monitored through sight glasses located on
the side of the vapor body vessel.   The liquid level in the
vapor body is controlled by the feed rate and the brine purge
rate.  A portion of the concentrated brine is removed
periodically from the heat exchanger in batches.  When the
vapor temperature reaches a value corresponding to a specific
brine boiling point created by a specific brine
concentration, some of the brine is drained by gravity into a
55-gallon waste drum.  The vapor exits the vapor body to an
entrainment separator, which separates mist droplets and
particles from the vapor stream, and the remaining heated
process liquor is recirculated.
                                  TM
     The next step in the PO*WW*ER  process is oxidation of
the volatile organics and inorganics in the vapor stream from
the evaporator.   The process is designed to operate with a
catalyst in either a fluidized or static bed mode.

     In a full-scale system the oxidizer consists of three
                             207

-------
main pieces of equipment:  a recuperative heat exchanger, a
reactor heater, and a catalytic reactor.   In a full-scale
PO*WW*ER  unit, the inlet vapor would be preheated along
with oxidation air in a recuperative heat exchanger with
vapor exiting the catalytic reactor.  The pilot-plant at
CWM's Lake Charles Treatment Center, however, does not have a
recuperative heat exchanger.  The preheated vapor is heated
to oxidation temperature at the reactor heater, a direct-
fired propane burner.  Propane and air are introduced in the
burner for proper combustion and oxidation to occur.  Air is
fed to the system by a compressor.  The heated vapor then
enters the catalytic reactor and passes through the catalyst
bed where oxidation of volatile materials contained in the
vapor stream takes place.  Oxidation products include carbon
dioxide (CO,) ,  water,  hydrogen chloride (HCL),  sulphur oxides
(SOX)  and  nitrogen oxides (NOX)

     The third step of the PO*WW*ER™ process includes
scrubbing the vapor stream to neutralize the acid gasses
produced in the oxidizer, and condensing the process vapor to
yield the product water.  After oxidation, the vapor stream
exits the reactor vessel and passes through a scrubber.  The
scrubber consists of a packed bed. in which the vapor passes
countercurrently through a caustic solution.  The scrubber
vapor is cooled and condensed in a shell-and-tube condenser.
The product condensate is transferred to a stainless steel
product tank.  The noncondensible gases are vented to the
atmosphere.

     Four primary objectives were identified for the SITE
demonstration of the PO*WW*ER  process:  (1) assess the
ability of the process to concentrate wastewaters with a low
solids concentration into a slurry, (2) assess the ability of
the process to remove volatile organic and inorganic
contaminants from wastewaters, (3) determine if the product
concentrate is nontoxic to aquatic organisms and, (4)
determine if the noncondensible gas stream meets proposed air
permit requirements.

RESULTS

     During the demonstration, landfill leachate from the
Lake Charles Treatment Center was treated using the
PO*WW*ER  pilot plant at a processing  rate of  0.21 gallons
per minute (gpm).  Six test runs were conducted with unspiked
leachate, and three runs were conducted using spiked
leachate.  The leachate was spiked with the following
compounds: methylene chloride, tetrachloroethene, toluene,
phenol, mercury, cadmium, copper, nickel, and iron.

     During each test, samples were collected from the feed
waste, product condensate, and brine.  Continuous emission
                             208

-------
monitoring (GEM)  was performed on the noncondensible gas
stream.  Feed waste, product condensate and brine samples
were analyzed for total suspended solids (TSS), total
dissolved solids (TDS)  ammonia, cyanide, VOC's and SVOC's.
Samples of the feed waste and product condensate were also
analyzed for acute toxicity.  GEM of the noncondensible gas
stream included monitoring for total nonmethane hydrocarbons,
CO2/  NOX, SO2,  and oxygen.   Additional analyses were also
performed to characterize further the feed waste, product
condensate, brine and the noncondensible gas stream.
                                TH
     The ability of the PO*WW*ER  system to concentrate
aqueous waste was evaluated by volume reduction and
concentration ratio achieved.  The volume reduction, defined
as the ratio of feed waste volume over brine volume was about
20.  The concentration ratio, defined as the ratio of total
solids concentration in the brine over the total solids
concentration in the feed waste was about 32.

     The  feed waste contained concentrations of VOCs ranging
from 270 to 110,000 micrograms per liter (/Ltg/1) ; SVOCs
ranging from 320 to 29,000 [J.g/1; ammonia ranging from 140 to
160 milligrams per liter (mg/1); and cyanide ranging from 24
to 36 mg/1.

     The average CO emissions from tjie noncondensible gas
vent ranged from 1.1x10"  to  3.92x10" pounds per  hour (Ib/hr)
and the 60-minute maximum CO emissions ranged from less than
1.27x10" to 4.28xlO~ Ib/hr.  The average SO2 emissions were
less than 5.1x10" Ib/hr and  the 60-minute maximum SO2
emissions ranged from less than 5.1x10"  to 8.36x10"  Ib/hr.
the average NOX .emissions  ranged from 3.15x10"  Ib/hr to
4.57x10" Ib/hr, and the 60-minute maximum NOX emissions
ranged from 3.26x10  Ib/hr to 4.85x10   Ib/hr.

CONCLUSION

     Preliminary results of the SITE demonstration of the
PO*WW*ER  system pilot-plant indicates that no VOCs, SVOCs,
ammonia or cyanide, all of which were present in the feed
waste, were not detected in  the product condensate.   The non-
condensible vent gas emissions for CO, SO2,  and NOX met the
regulatory requirements for the Lake Charles Treatment Center
site.
                 TM
     The PO*WW*ER  system removes sources of  feed waste
toxicity, but does not necessarily produce nontoxic product
condensate because of the unadjusted pH, temperature,
hardness, and salinity of the condensate were outside the
optimal ranges for growth and survival of test organisms.

     Key findings of the SITE demonstration including
                            209

-------
analytical results will be discussed in detail in the EPA
Applications Analysis Report (AAR) and the EPA Technology
Evaluation Report (TER).  The SITE demonstration results will
also be summarized in the Demonstration Summary Report and
videotape.


For More Information Contact:
                       Randy A. Parker
               Environmental Protection Agency
                   26 West M.L. King Drive
                    Cincinnati,  Ohio  45286
                        (513)  569-7271
                             210

-------
              BIOREMEDIATION OF CHLOROPHENOL CONTAMINATED AQUIFERS

                              Sandra L. Woods, Associate Professor
                              Jonathan D. Istok, Associate Professor
                                Kenneth J. Williamson, Professor

                                 Department of Civil Engineering
                                      Apperson Hall 202
                                    Oregon State University
                                   Corvallis, OR 97331-2302
                                         503-737-2751

INTRODUCTION

       Pentachlorophenol (PCP) is an important environmental pollutant due to its toxicity to a wide variety
of organisms and wide distribution in the environment.   PCP has been used extensively as a wood
preservative and pesticide. Hundreds  of wood treating sites have been contaminated with PCP with many
of these sites have been placed on the National Priority List  for remediation.

       Anaerobic transformation of chlorophenois occurs by reductive dechlorination, a process by which
chlorines are replaced with hydrogen.  This process depends upon the parent compound, the microbial
consortium, and various environmental factors.  Reductive dechlorination  is of environmental importance
because the lesser-chlorinated metabolic products are generally less toxic and more easily degraded by
aerobic bacteria. However, the less-chlorinated metabolic products are more mobile in aquifers because
of their increased water solubility.

       Aerobic transformation of chlorophenois occurs by hydroxylation of the chlorines. This process also
depends upon the parent compound, the microbial consortium, and various environmental factors. Aerobic
degradation of chlorophenois is a co-metabolic process with  a primary electron donor required in addition
to the chlorophenois.

       The transport of chlorophenois in aquifers depends on physical, chemical, and biological processes
such as  advection, dispersion, sorption, and biodegradation.   Models are needed  to integrate these
processes to allow the prediction of the fate and transport of  chlorinated phenols. Of special interest was
the inclusion of several types of sorption/desorption processes including  equilibrium/nonequilibrium and
linear/nonlinear conditions.

METHODOLOGY

       Anaerobic  chlorophenol  biotransformation  pathways and removal  rates were determined  for
methanogenic consortia fed 5,300 mg acetate per liter, 3.4 u.M PCP, and nutrients for periods between 10
days and 9  months. The reactor was brought  to steady-state with respect to retention time, substrate
removal,  gas production,  Eh,  and pH.  A 9.5-liter continuous-flow "mother" reactor served as a source of
inocula for the degradation experiments conducted in smaller batch reactors.  Batch experiments were
conducted in a 2.5-liter glass reactor.  The batch reactor was continuously monitored for Eh, pH, and gas
production.  Liquid samples were periodically removed and analyzed for chlorophenois by acetylation and
extraction in hexane (1)(2).

       Aerobic chlorophenol biotransformation  removal rates were determined for  aerobic consortia fed
acetate (500 mg/L as COD), nutrients, and eight chlorophenois (4-chlorophenol [4^CP]; 2,4-dichlorophenol
[2,4-DCP]; 3,4-dichlorophenol [3,4-DCPJ; 3,5-dichlorophenol  [3,5-DCP]; 2,3,5-trichlorophenol [2,3,5-TCP];
2,4,5-trichlorophenol [2,4,5-TCP]; 3,4,5-trichlorophenol [3,4,5-TCP]; and 2,4,6-trichlorophenol [2,4,6-TCP])
                                              211

-------
at 0.1 mg chlorophenol per liter each. The 8-liter mother reactor was continuously aerated to provide
oxygen. Batch experiments were conducted in a 2.5-liter glass reactor and continuously monitored for pH
and dissolved oxygen.  Liquid samples were periodically removed and analyzed for chlorophenols (as
above) and acetate.

       A numerical model using the time-centered finite difference method of Crank-Nicolson (3) was used
to describe one-dimensional chlorophenol transport in a soil column. The systems of non-linear equations
were solved by a combination of the successive averaging technique and the Newton-Raphson iteration.
The program was used to  predict breakthrough curves for several types of sorption processes and the
results compared to measured breakthrough curves from laboratory experiments using Chehalis soils in
30-cm long columns. The model allowed prediction of the type of sorption processes occurring from the
shape of the breakthrough  curve.

       A mathematical model including advection, dispersion, sorption, and anaerobic biodegradation was
developed to predict fate and transport of chlorophenols in an aquifer based  upon the  results of the
laboratory biodegradation and column sorption experiments. Acetate was assumed as the electron donor
and degradation was modelled  using Monod kinetics and a "macroscopic bulk concentration" for the
chlorophenols. The model tracked the production of all metabolic products and their subsequent transport,
sorption, and degradation.

RESULTS

       The consortia exposed to PCP for only 10 days dechlorinated PCP principally at the ortho position.
However, after acclimation  PCP was biodegraded to form all three tetrachlorophenol (TeCP's), as well as
3,4,5-, 2,4,5-, and 2,3,5-TCP, as shown in Figure 1. The "sum" represents the sum of PCP, all TeCP and
the three TCP's;  the fall off after 0.6 days resulted from the production of DCP's.  These experiments
indicated that the consortium is capable of dechlorination at the ortho, para, and meta positions of  PCP.
Further experiments in which the batch reactor was fed singularly the  various metabolic products of PCP
resulted in the pathways shown in Figure 2.
                               O.2      0.4      0.6      0.8      1.0      t.2
                      Figure 1.  Progressive reductive dechlorination of PCP
                          by a PCP-acclimated methanogenic consortium.
                                              212

-------
          Rgure 2. Observed reductive dechlorination pathway, for PCP and its metabolites

       The aerobic degradation of the tri-, di-, and mono-chlorophenols was observed to be zero order
as illustrated for 2,4,6-TCP in Figure 3. The aerobic consortia was capable of degrading 4-CP, 2,4-DCP
3-4-DCP, 3-5-DCP, 2,4,6-TCP, 2,3,5-TCP, 2,4,5-TCP and 3,4,5-TCP at rates  ranging from 0.02 to 0.2
u.M/L-hr for all the compounds.except 4-CP. The degradation rate of 4-CP was greater than 10 times larger
that the other compounds.
                             10   15    20   25   30    35   40   45~
                                         Time (Hours)
50
                    Figure 3.  Co-metabolic degradation of 2,4,6-trichlorophenol
                                             213

-------
       Forthe case of sorption only, the model described closely laboratory breakthrough curves for PCP,
TeCP's, TCP's, DCP's, and MCP's (Figure 4). The model predictions were based upon sorption distribution
coefficients estimated from water solubility data (4).
                o
               O

               o
                   1.0 -i
                   0.8 -
                   0.6 -
                   0.4 -
                   0.2  -
                   0.0
                 o-O "O-O	0  0 „» -»
                  /   	 TeCP ANAL
                 {    	PCP  ANAL
                /    00000 TeCP NUM
               /    xxxxx PCP  NUM
                              100    200     300     400    500     600

                                       PORE  VOLUMES
                 o
                o
                o
l.O -
0.8 -
0.6 -
0.4 -
0.2 -
n n -


ur --"tr
/
/ /
/

-------
CONCLUSIONS
                                                                                         1 >• -   V
       For reductive dechlorination of PCP, ortho dechlorination was observed most frequently with eight
of nine possible ortho-dechlorination products observed. In contrast, only two para-dechlorination products
were observed out of seven possible, and four meta-dechlorinations of nine possible.

       Aerobic degradation of tri-, di-, and mono-chlorophenols is reasonably rapid and suggests that
these compounds will be removed  in aquifers  under  oxygen rich  conditions.  All  of the  anaerobic
dechlorination products can be successfully removed under aerobic conditions.

       The numerical  model for one-dimensional transport and sorption was able to  describe the
movement of chlorophenols through soil columns.

REFERENCES

1.     Voss, R.H., J.T. Wearing, and A. Wong.  A Novel  Gas Chromatographic Method for the
       Analysis of Chlorinated Phenolics in Pulp Mill Effluents.  In L.H. Keith (ed.), Advances in
       the Identification and Analysis of Organic Pollutants in Water, 2: 1059-1095, ann Arbor,
       Mich.,  1981.

2.     National Council  of  the  Paper Industry  for Air  and  Stream Improvement  (NCASI).
       Experience with the Analysis of Pulp Mill Effluents for Chlorinated Phenols using an Acetic
       Anhydride Derivitization Procedure. Stream Improvement Technical Bulletin No. 347, June,
       1981.

3.     Ungs,  M.J.,  L. Boersma,  and S. Akratanakul.  The Numerical Analysis of Transport of
       Water and Solute through  Soil and Plants. Agricultural Experiment Station, Special Report
       753, Oregon State University,  1985.

4.     Jones,  P.A.    Chlorophenols and  their  Impurities  in  the Canadian Environment,
       Environmental Impact Control Directorate, Environmental Protection Service, Environment
       Canada, Minister of Supply and Services Canada,  1981.

FOR MORE INFORMATION

       For more information, contact:

                      Dr. Sandra Woods, Associate Professor
                      Department of Civil Engineering
                      Apperson Hall 202
                      Oregon State  University
                      Corvallis,  OR  97331-2302
                      503-737-6837
                                               215

-------
 PRELIMINARY ASSESSMENT OF FATE MECHANISMS OF CHLORINATED ORGANIC COMPOUNDS
                        IN A SEDIMENT-WATER ESTUARINE SYSTEM

                      Spyros G. Pavlostathis, Peter Gess and Mark Prytuia
                                 School of Civil Engineering
                      Georgia Institute of Technology, Atlanta, GA 30332
                                      (404) 894-9367

INTRODUCTION

       Contaminated bottom sediments,  previously thought of as natural sinks for contaminants,
are now viewed as the source of contamination.  Although research on contaminated sediments
has been intensified  in the last decade, we still lack fundamental knowledge of the processes and
process interactions  which govern the fate of sediment-bound toxic chemicals.

       Bayou d' Inde, a tributary of the Calcasieu River in Southwestern Louisiana, was chosen for
study because recent investigations and monitoring of this tributary have shown that high-
molecular-weight chlorinated organic compounds (benzenes and butadiene) and metals  (chromium,
copper, lead, mercury, nickel) persist in the sediments (1). The objectives of this study are:  (a)
Characterization of selected  sites in the contaminated sediment-water estuarine system; (b)
Assessment of the rate and  extent of biotransformation of the chlorinated organic compounds; and
(c) Investigation of the abiotic release and binding mechanisms responsible for the fate  of the target
organic compounds. The work presented here pertains to preliminary sampling and characterization
of both water and sediment  samples, examination of potential fate mechanisms and a preliminary
biotransformation study of hexachlorobenzene (HCB) and hexachloro-1,3-butadiene  (HCBD).

METHODOLOGY

       Water and sediment  samples were collected from three stations in the Bayou d' Inde
tributary and physical/chemical analyses were performed.  Standard methodologies were used for
most of the analyses (2) (3)  (4).  Contaminant extraction from sediment samples  was performed by
using the Soxhlet method with a mixture  of 1:1  (v/v) acetone:hexane. Hexane was used for the
liquid-liquid extractions of water samples. The chlorinated organic contaminants were identified
and quantified with a gas chromatography unit equipped with a photoionization detector in tandem
with an electrolytic conductivity detector.  Confirmation of these compounds was also  achieved by
the use of a GC/MS  unit.

       Interactions between iron, natural organic matter and the target contaminants have been
examined in order to identify the conditions and potential effect of these physico-chemical
interactions on the binding and transport of the target contaminants. Abiotic experiments are
under way in order to quantify these interactions by taking into account the conditions  of the test
site to the degree possible (e.g.,  salinity and temperature).

       The biotransformation of HCB and HCBD under mixed sulfate-reduction and methanogenic
conditions was monitored.  Batch microcosms were constructed with sediment and water samples
from the test site. HCB and HCBD were added in separate microcosms. The effect of  sediment
organic carbon as well as that of added, external carbon on the  biotransformation of the target
compounds was assessed. The metabolic rates of sulfate-reduction and methanogenesis were
quantified and compared with those achieved in the presence or absence of HCB  and HCBD.
Duplicate microcosms were  constructed with a sediment/water slurry (final solids concentration of
25 g/L) and the following amendments: I, nothing added; II, electron donor;  III, HCB;  IV, HCB
plus sodium azide (Control I);  V, HCB plus electron donor;  VI, HCBD;  VII, HCBD plus  sodium
                                            216

-------
azide (Control II);  and VIII, HCBD plus electron donor. A mixture of lactate and acetate was
provided as external electron donor (final concentration 10 mM each). The final concentration of
the added organic contaminants was 44 and 36 mg/L for the HCB and HCBD, respectively.


RESULTS

       The results of analyses of two water samples (Station 1 and 3) and three sediment samples
(Station 1, 2 and 3) are given in Tables 1 and 2, respectively.  Nitrate and phosphate were not
detected.  The following chlorinated organic compounds were identified in sediment extracts from
Station 1 (in jug/kg dry weight basis):  dichlorobenzene, 3900; trichlorobenzene, 2570;
tetrachlorobenzene, 5507 pentachlorobenzene, 3440; HCB, 8890; and HCBD, 7020. Therefore,
HCB and HCBD are the predominant contaminants in the sediment under study. The less
chlorinated organic contaminants are probably dechlorination products of HCB. The water samples
when extracted with hexane did not yield any detectable amounts of HCB  or HCBD. When the
purge-and-trap technique was used, volatile organic  compounds (VOC) -- such as mono-, di- and
trichlorobenzes as well as HCBD - were detected in water samples.
  Parameter
TABLE 1.  RESULTS OF WATER ANALYSES

                          Station  1
Station 3
PH
Alkalinity, mg/L as CaC03
Total dissolved solids, mg/L
Dissolved volatile solids, mg/L
Dissolved organic carbon, mg/L
Total dissolved carbon, mg/L
Chloride, mg/L as CI"
Sulfate, mg/L as SO42"
Metals, fjg/L
Cr (total)
Cu
Fe (total)
Mn
Pb
Chlorinated VOC, jjg/L
Chlorobenzene
Dichlorobenzene
Trichlorobenzene
Hexachlorobutadiene
7.8
78
9440
8440
5.2
30.6
5520
547

108
25
375
99
205

0.20
0.13
0.14
0.66
7.5
70
NM*
NM
5.1
27.3
4950
552

106
20
252
67
197

0.10
0.05
0.13
0.37
 NM, not measured
                                           217

-------
                        TABLE 2.  RESULTS OF SEDIMENT ANALYSES
  Parameter
Station 1
Station 2
Station 3
pH
Alkalinity, mg/L as CaC03
Water content, %
Volatile solids, % dry weight
Total organic carbon, % dry weight
Metals, mg/kg dry weight
Cu
Fe (total)
Mn
Pb
6.6
2040
61.8
10.2
4.7

143
24400
184
12
6.9
1320
54.4
6.8
3.3

132
15800
154
8
6.9
2000
55.7
8.7
4.6

28
15080
- 126
8
       Duplicate microcosms were sacrificed after 22 and 36 days of incubation. Hexane
extraction and quantification  of the chlorinated organic compounds were carried out.  In microcosm
V (amended with electron donor), over 86% and almost complete disappearance of HGB was
observed after 22 and 36 days of incubation, respectively. In contrast, ca. 36 and 87% of HCB
was depleted in microcosm III (without external electron donor) at the two incubation periods,
respectively (Fig. 1).  The predominant dechlorination products were  di- and trichlorobenzenes.
Similar results were obtained in microcosms amended with HCBD,  For example, after 22 days of
incubation, more than 94% of the  initial HCBD was dechlorinated in microcosms VI and VIII,
respectively.  A large number of chlorinated compounds were detected at retention times lower
than that of HCBD.  Identification of these compounds will be achieved by use of a GC/MS unit.

       Sulfate-reduction was pronounced in all three microcosms amended with electron donors
(II, V, and VIII). After 22 days of incubation, about 82 to more than  93% of the initial sulfate was
depleted.  In contrast, only 3 to 18% sulfate depletion was detected  in microcosms not amended
with electron donors  (I, III, and VI) for the same incubation period. The occurrence or lack of
sulfate-reduction was also reflected by the presence or absence of hydrogen sulfide gas.  All
microcosms amended with electron donor achieved high gas production, except microcosm VIII
(amended with electron donor and HCBD) which produced only 27%  of the gas measured in the
other two microcosms. Methane was not detected in the gas of microcosm VIII.


CONCLUSIONS

       Nitrate was not detected in both sediment and water samples. Therefore, denitrification
was not assessed for its biotransformation potential of the target compounds.  The metabolic  rates
of both sulfate-reduction and methanogenesis were insignificant in microcosms not amended  with
external electron donors.  The addition of electron donors enhanced the metabolic rates of these
two processes as well as the biotransformation of HCB and HCBD. The predominant dechlorination
products of HCB were di- and trichlorobenzenes. Lower molecular weight chlorinated compounds
were formed as a result of dechlorination of HCBD. The  addition of HCBD led to inhibition of
methanogenesis but did not affect sulfate reduction.  HCBD and/or its dechlorination products may
be responsible for the observed reduction in gas production. The sediment-bound contaminants
could  potentially be transformed via biological reductive dechlorination, if sufficient electron donors
are  available to accelerate the biotransformation rate.
                                            218

-------

         UJ

         IL8  3
         N

         UJ
         CD  o
         O  2
         IT
         O
         O
                            DCONTROL
                         0                 22                36
                           INCUBATION  TIME (Days)
Figure 1.      Biotransformation of HCB in three microcosms: Control (poisoned); III and V (live
             without and with external electron donor, respectively)
REFERENCES

1.     Cunningham P. et al.  Toxics Study of the Lower Calcasieu River.  PB90 226150/AS,
      National Technical Information Service, Springfield, VA, 1990.

2.     American Public Health Association. Standard Methods for the Examination of Water and
      Wastewater,  17th ed., APHA-AWWA-WPCF, Washington, DC, 1.989.

3.     Page, A. L. (ed.).  Methods of Soil Analysis - Part 2: Chemical and Microbiological
      Properties, 2nd ed., American Society of Agronomy, Inc., and Soil Science Society of
      America, Inc., Madison, Wl, 1982.

4.     U. S. Environmental Protection Agency. Test Methods for Evaluating Solid Waste. SW-
      846, 3rd ed., EPA/OSWER, Washington, DC., 1986.
                                        219

-------
  TRANSPORT AND ENTRAPMENT BEHAVIOR OF IMMISCIBLE ORGANIC WASTE CHEMICALS IN
  HETEROGENEOUS AQUIFERS: IMPLICATIONS ON MODELING. RECOVERY AND REMEDIATION.
                                     TissaH. Illangasekare
                  Department of Civil, Environmental and Architectural Engineering
                                 University of Colorado, Boulder
                              Colorado 80309-0428; (303) 492 6644
INTRODUCTION
     Industrial solvents  and sludge,  electroplating  bath  solutions, petroleum refinery wastes,  wood
preservation wastes and corrosive wastes are just  some of the hazardous chemicals in the form of
nonaqueous phase liquids (NAPL) which are generated as by products from industrial and manufacturing
activities. Also, liquids such as gasoline, diesel and jet fuels fall into this category of NAPL contaminants
because of their soluble constituents.   Nonaqueous phase liquids (NAPLs) are present at an increasingly
large number of waste sites across the country, and they are proving to be some of the most difficult
contaminants to successfully remediate. The primary characteristic that makes NAPLs unique is that they
have relatively low aqueous solubilities, and thus may exist as an immiscible phase in the subsurface for a
very long time, thus acting as a continual source of contamination as their soluble components slowly but
continuously vaporize or leach out. This is in contrast to the highly miscible fluids, which have been studied
in great detail and are comparatively easy to track and to be modelled.

     Our own laboratory studies and those by other researchers have demonstrated that soil heterogeneities
can cause lateral spreading, preferential flow and pooling of the organic fluids. Studies have also shown
the initiation of unstable flow due to the presence of heterogeneities. Instabilities can cause the organic to
infiltrate the soil as one or more "fingers", rather than a uniform front.

     The focus of the research presented here-has been on the development of fundamentally valid
descriptions  of the subsurface movement of NAPLs that can be used to provide useful guidance  in the
development of models for regulation, site characterization and remediation design and planning.

METHODOLOGY

     The primary goal was to obtain through laboratory experimentation an improved understanding of the
fundamental  processes associated with the entrapment, movement and mobilization of nonaqueous phase
organic waste fluids (NAPLs) in naturally heterogenous groundwater aquifers.  The knowledge gained was
used to evaluate existing numerical simulation models and to develop new models.  The models when
properly validated can then be used to evaluate and design effective recovery and remediation technologies
for aquifers contaminated with organic waste chemicals.

     This research has examined a number of individual problems associated with multiphase subsurface
contamination. Laboratory investigations were conducted in columns and large soil tanks.  Existing models
and models which have been developed by the author and  co-workers were evaluated using laboratory
data. Laboratory techniques to determine the parameters used in these models were developed.

     Spill experiments with lighter than water fluids (LNAPLs) were conducted in a 32'ft long tank equipped
with an automated dual-gamma attenuation system for the insitu measurement of phase saturations. After
the spill, various direct recovery strategies were implemented to evaluate their feasibility and effectiveness.
Attempts were made to develop a new type of numerical models which are capable of modeling NAPL sharp
fronts, and NAPL entrapment.
                                             220

-------
      Controlled experiments in soil columns using various soil and chemical types were conducted to
 develop descriptions of  NAPL entrapment at the pore scale based on easily measured macroscopic
 parameters (1). Water flooding experiments were conducted to test the effectiveness of mobilization methods
 for the removal of NAPL entrapped in the pores.

      Experiments with dense fluids were conducted in a 6 ft high two-dimensional tank and a 4 ft high three-
 dimensional tank. Contaminant  migration patterns under heterogenous conditions were observed and
 measured. Three-dimensional spill experiments were designed to investigate the effects and sensitivity
 of various parameters on finger shape and vertical finger propagation.  An image processing system was
 used to visualize the unstable displacement of a wetting fluid by a dense non-wetting fluid. The data were
 used to describe observed fingering as a function of fluid density contrasts, absolute viscosity differences,
 surface tension and pore scale  perturbations  and to test the statistical concepts of self-similarity and
 percolation theory.

      One dimensional column experiments were conducted to examine the entry pressure  phenomena
 between two soil layers. Of particular interest was what, if any, influence the interface between the two soils
 has on the entry pressure.

 RESULTS                           '

     The results show the importance of defining the controlling soil parameters and heterogeneities at
 different scales of interest, namely: the pore-scale., site-scale and regional-scale. The pore parameters which
 have to be defined and measured at the microscopic scale determine the residual entrapment, mobilization
 and stability conditions which produce fingering. The large tank experiments demonstrated quite dramatically
 the critical role played by the spill site-scale heterogeneities in defining the migration paths and the modes
 of entrapment (2). These  heterogeneities resulted in much larger entrapment saturations compared to the
 residual pore saturations. This observation  required  the investigation and quantification of the basic
 processes which control the "macro-scale" entrapment resulting from the macro- or site-scale heterogeneities.
 The site scale heterogeneities are not only defined by the average soil properties of individual formations but
 also the properties of the interfaces between the formations. These heterogeneities in combination with the
 interface effects produce preferential flow paths for both the chemical and water. The ineffectiveness of
 some of the pump and treat and insitu treatment schemes could be attributed to the failure to understand and
 quantify these site scale heterogeneities using the traditional site characterization techniques.

     Figure 1 shows sample results from a spill simulation conducted in the large tank. The effects of
 heterogeneities in both the unsaturated and saturated zones on the entrapment behavior of a LNAPL are
 shown.

     Our own investigations using both qualitative and quantitative laboratory data generated in large soil
flumes have  identified some of the inadequacies of existing models in simulating NAPL transport in
 heterogeneous soil systems (3). Sudden variations in aquifer properties resulted in stability and convergence
 problems in numerical models Artificial contaminant fronts created as a result of these numerical problems
 resulted in unrealistic plume configurations which were not observed in the laboratory. The models were
shown to fail when simulating fronts.  The models also failed to recognize and predict the preferential flow
 processes which were observed in the laboratory. In addition, none of these models have the capability to
 model the transport behavior of DNAPLs, specifically when the movement occurs through fingering as has
 been observed in the laboratory.

     Figure 2 shows a sample  result from a spill experiment where a DNAPL was allowed to flow in
saturated soil. The figure shows multiple finger development.
                                             221

-------
                                                                                  NAPL SATURATION



                                                                                         50%-80%
Rgure 1. Laboratory simulated distribution of a light organic fluid in a heterogeneous aquifer after a spill.
Figure 2. Laboratory simulated distribution of a dense organic fluid in saturated soil after a spill.
                                                 222

-------
     The experimental and modeling results also suggest the need to develop better laboratory and field
testing techniques to obtain both macroscopic and  site-scale parameters which  characterize aquifer
heterogeneities. An innovative computer automated soil characterization technique which uses a flow pump
and a precision pressure transducer was developed (4).


CONCLUSIONS

     An extensive laboratory data base was developed to provided an improved qualitative understanding
of the fundamental physics of flow, entrapment, mobilization and mass transfer associated with immiscible
fluid in soils, particularly in heterogeneous systems.  This data was used to evaluate existing simulation
models and to identify necessary improvements to models.

     The results of these experiments demonstrated the critical role played by aquifer heterogeneities in
immiscible fluid migration and entrapment. The model testing suggests the need to develop better models
which could simulate the processes  occurring at the interfaces of heterogeneities and to  develop site
characterization procedures to capture the properties of heterogeneous formations.
REFERENCES
1.   Szlag, D.C. and Illangasekare, T.H. Entrapment of nonaqueous phase liquids: correlation with
    macroscopic parameters, Submitted to Groundwater, 1993.

2.   Illangasekare, T.H., Yates, D.N., Armbruster, E.J. and Reible, D.D. Effect of heterogeneity on transport
   and entrapment of nonaqueous phase waste products in aquifers: an experimental study, submitted to
   Water Resources Research, 192

3  Illangasekare, T.H., Yates, D.N., Armbruster, E.J. and Wald, J. Laboratory evaluation of limitations of
   conventional numerical multiphase flow models in porous media. IQ: Computational Methods in Water
   Resources IX, Vol 2. Mathematical Modeling in Water Resources, Ed. Russel, Ewing, Brebbia, Gray and
   Finder, Computational Mechanics and Elsevier, 289-296,1992.


4.  Znidarcic, D., Illangasekare, T.H. and Manna, M., Laboratory testing of and parameter estimation for
   two-phase flow problems, in: Proc. of the Geotechnical Engineering Congress, Ed. McLean, Campbell
   and Harris, ASCE, New York, 1078-1089,1991.
FOR MORE INFORMATION

Tissa H. Illangasekare
Department of Civil, Environmental and Architectural Engineering
University of Colorado, Boulder
Colorado 80309-0428; (303) 492 6644; email:lllangas@bechtel.edu.colorado
                                             223

-------
A BEACH MICROCOSM FOR THE STUDY
     OF OIL BIODEGRADATION
     Strohmeier1,
Albert D. VenosaS
             Kevin L.
             Makram T.  Suidan1, and John R. Haines*.
        Department  of Civil and Environmental Engineering
                     University of  Cincinnati
                      Cincinnati,  OH  45221
                          (513) 569-7179
               2US Environmental  Protection Agency
              Risk Reduction Engineering Laboratory
                      Cincinnati,  OH  45268

     Microcosms have been constructed and operated that model
biodegradation of crude oil in the intertidal zone of sand and
gravel beaches.  The heart of the system is a computerized
respirometer that measures biological activity by means of
continuous tracking of oxygen consumption.  A 7.5 cm by 30 cm
column of coarse sand is contaminated with hydrocarbon and
inoculated with acclimated microorganisms.  Artificial seawater
flows through the column intermittently, simulating waves, and a
pump controls tidal inundation.  This flow-through system
encourages washout of organisms and nutrients and removes water
soluble metabolites.  It more closely resembles real world
conditions than do closed, batch systems.

     The fates of carbon, oxygen, and nutrients are followed
closely throughout the course of an experiment.  Substrate carbon
is loaded at a known concentration, and removal is tracked by
effluent carbon analysis and by trapping carbon dioxide in a
potassium hydroxide solution.  The mass balance of carbon is
completed by extracting and quantifying the residual hydrocarbon.
Oxygen use is tracked by the respirometer; incomplete degradation
is assessed by chemical oxygen demand, completing the oxygen mass
balance.  Nutrient use is quantified by differences between feed
and effluent concentrations of the nutrients.

     This system has been designed to provide better estimates of
field biodegradation rates of crude oil.   The  material balances
provide useful information on the  fates of specific compounds and
fractions of oil over time following a spill. The automated
microcosms will aid in determining optimal nutrient and inoculum
application rates and frequencies  for field studies and oil
spills.
               224

-------
                          ACID EXTRACTION TREATMENT SYSTEM

                                      Stephen W. Paff
                           Center for Hazardous Materials Research
                                    320 William Pitt Way
                                    Pittsburgh, PA 15238
                                      (412) 826-5320
      The Acid Extraction Treatment System (AETS) is a means of removing heavy metals from
contaminated soils and solids.  AETS represents a significant extension of existing "soil washing"
techniques in the Netherlands,  which have been directed primarily at remediation of hazardous
organics contamination. Initial  bench-scale studies of AETS have been very promising and have
achieved removal efficiencies of 29 to 99% for various heavy metals in soil.

      Applications  of AETS in the United States may include: (1) remediation of soils and/or solids
contaminated with  heavy metals; (2) post-treatment of high, heavy metal content ash residue from
incineration processes; and  (3)  treatment of industrial and municipal wastewater treatment sludge
containing heavy metals.

      The objective of the E02/AETS project is to determine AETS' effectiveness and commercial
viability in reducing the concentrations and/or leachability of heavy metals in soils to acceptable
levels.

      To date, CHMR has performed a combination of laboratory and bench-scale testing for acid
extraction.  The first year project effort was focussed primarily on laboratory-scale testing.  The
results of the laboratory-scale testing were  promising, with heavy metal removal efficiencies of
between 29% and  99% for such heavy metals as arsenic, cadmium, copper, lead, nickel,  and zinc.

      CHMR used the information gained during the first year to design and construct a pilot-scale
unit capable of processing between 20 and 100 kg of soil per hour. The results from this system
have also been promising, with removal efficiencies of between 70 and 90%, and typical  TCLP
reduction in the 90% range."

      CHMR will continue testing the pilot-scale unit through the remainder of 1992,  on a variety
of soils obtained from  Superfund sites across the United States, and under several different process
conditions.
                                           225

-------
   flN FVA1.UATION OF DRINKING WATER SAMPLES PRE-DISINFECTED WITH  CHLORINE
   ""        USING  CHEMICAL.  MICROBIOLOGICAL  AND MUTAGENICITY ENDPOINTS
    Kathleen Schenck Patterson, Lucille M. Garner, Robert  G.  Miller and
                          Benjamin W. Lykins, Jr.

                    U.S.  Environmental  Protection Agency
                   Risk Reduction Engineering Laboratory
                      Drinking Water Research Division
                    Systems  and Field Evaluation  Branch
                       26 W.  Martin Luther King Drive
                           Cincinnati,  Ohio 45268
                (513/569-7947, 569-7417, 569-7454, 569-7460)

     Many drinking water utilities are considering alternatives  to  the
exclusive use of chlorine (C12)  for disinfection in order to comply with
current and proposed federal  regulations regarding acceptable  levels  of
disinfection by-products.  Consequently, alternative treatment processes  are
being evaluated.

     At a pilot-scale drinking water treatment plant in  Evansville,  Indiana,
three studies were conducted in which river water was either  treated  with
liquid chorine dioxide (CIO,), gaseous  CIO,,  C12 or was not disinfected.  In
the first study, those streams disinfected" with CIO, were  subsequently
treated with a reducing agent (FeCl2) and Cl,.   In the second  study, the
reducing agent was omitted from the liquid CIO, stream.  In the  third study,
chloramine  (NH2C1) was substituted for C12 as the secondary disinfectant  in
the stream  treated with gaseous C102.  For each treatment  stream total
organic carbon  (TOC), total  organic halide (TOX)  and microbiological  con-
taminants were determined.   XAD resin concentrates were  also  prepared for
mutagenicity testing. in the  Ames  assay.

     Mutagenic  activity was  detected in all  the water  samples.  In the first
two studies, the  levels of mutagenic activity present  in samples treated
  ... «•   * i _.  ___ ___.._ /"* "I rt   .CA^Im-is-t^J Ut t * OT  t.m\s*n ^*r«**^Mrt4''4'^lI\/ 4"r\Q O O tTlO
                                                                       alone.
UWU Ol*UUlCOj U1IC I C V C I «J W I  iMMW\A*j*-iii*- V4^***i** w^ f* • w v. • ~  ... — —...r	- -	
with liquid or gaseous C102 followed by C12 were  essentially the same.
These samples were slightly less active than samples treated with  C12 	
In the third study, the addition of NH2C1  following gaseous C102 reduced the
level of mutagenicity observed by 50% compared to the  levels present in
samples treated with either liquid C102 and Cl, or C12  alone.   The concen-
trations of TOX present in the water samples showed a  pattern  similar to
that of the mutagenicity data.  No significant bacterial contamination was
observed in water samples treated with either C102 and a secondary  disin-
fectant (Cl, or NH2C1)  or C12  alone.  Thus  the  substitution  of NH,C1 for C12
as a secondary disinfectant may be beneficial  in  drinking water treatment.

     This abstract does not necessarily reflect  EPA policy.

     For more information contact Kathleen Schenck  Patterson,  USEPA, RREL,
DWRD, SFEB, 26 W. M. L. King  Drive, Cincinnati,  Ohio 45268,  513/569-7947.
                                    226

-------
        ANAEROBIC BIOPROCESSING OF NITROAROMATIC COMPOUNDS

                           Dane Higdem
                      J. R. Simplot Company
                           P.O.  Box 912
                       Pocatello, Id 83204
                           (208)234-5367

                  Ronald L. and Don L. Crawford
         Center for Hazardous Waste Remediation Research
                  Food Research Center Room 202
                        Moscow,  Id 83843
                           (208)885-6580

     The J. R. Simplot Company,  in conjunction with the University
of Idaho,  has  developed a  unique anaerobic process for degrading
nitroaromatic  compounds.    These  compounds  include dinoseb,  a
herbicide  that was  used  as  a  devining   agent  on  potatoes and
legumes, and TNT, an explosive used for ordinance. The process uses
an anaerobic microbial consortium that has been shown  to mineralize
the nitroaromatic compounds.

     Once  anaerobiosis is  established (Eh <-200mV), the anaerobic
consortium degrades  dinoseb and  TNT completely without forming
polymerized intermediates that are found with aerobic  composting or
land farming.  Aerobic treatments do not destroy the nitroaromatic
compounds; rather, they promote polymerization of the intermediate
metabolites into aminoaromatic compounds that can be as toxic, if
not more so, than the parent compound.  In  addition, other natural
soil  microorganisms  can   cleave  the  azo  polymer  linkage,  re-
releasing  these toxic compounds to the soil and water.

     Tests have shown this novel system degrades target compounds
faster and to  a greater extent  than existing technologies.   The
J. R. Simplot Company,  in conjunction with the University of Idaho,
has developed a simple bioenrichment process for  the degradation of
nitroaromatic  hydrocarbon  contaminated soil.    This inexpensive
technology does not  require specialized  anaerobic  bioreactors or
other specialized equipment.  The simplicity of the system allows
for tremendous flexibility in  regard to site size,  allowing the
smallest  agricultural site to the  largest defense site  to be
economically treated.
                              227

-------
                      APPLICATION OF COMPUTER SIMULATION
            FOR INTRODUCTION OF POLLUTION PREVENTION IN INDUSTRY

                                    Jordan Spooner
                                    Radojka Olbina

                          Pollution Prevention Research Branch
                          Risk Reduction Engineering Laboratory
                          U.S. Environmental Protection Agency
                              26, W. Martin Luther King Drive
                                  Cincinnati,  OH 45268

                            Phone: (513) 569-7422; 569-7526

                                       Abstract

    Several computer supported methods are being developed and can be applied to extract
valuable information from available data sets to predict new data from underlying theories. One of
these methods, computer simulation (computer-supported modeling) is a technique which
allows direct observation of systems' behavior under a variety of conditions. Computer simulation
can provide support of bench, pilot and/or full plant model tests as a tool for model sensitivity and
uncertainty analysis. It has been used  as a valuable decision-making tool in wide variety of
technical and non-technical fields.

    In this Pollution Prevention Research Branch project, on-line and conventional information
sources were searched to acquire data on the various applications of computer simulation, with a
focus on industrial (e.g., chemical, biotechnological) and energy production, and transportation.
The data found have been organized  into the following reference (bibliographic and referral)
databases:

(1) Computer simulation technical application; and
(2) Commercial computer simulation software.

By organizing these data into databases and then retrieving and analyzing information using the
methodology of structuring of data into systems, research and development trends will be
determined.

    The preliminary results of the project indicate that the number of pollution prevention
applications of  computer simulation is  small in comparison to other technical applications. The
leading technical application appears  to be for  solving of wastewater treatment, biotechnological,
Industrial and especially chemical production problems. Within these fields, computer simulation is
most often used for process optimizatbn and economic analysis.
                                            228

-------
            DENSITY-DRIVEN ADVECTION DURING LABORATORY LEACHING TESTS

                            S. Proffitt, J. E. Isenburg, and W. W. Slack
                                     University of Cincinnati
                                      5995 Center Hill Road
                                     Cincinnati, Ohio, 45224
                                        (513)569-7885

        We added pH indicator dyes to the leachant of a standard American National Standards Institute
(ANSI) 16.1 test, which is commonly used to characterize material that has been treated by solidification
/ stabilization processes, and obtained results that contradict a critical assumption of the ANSI test. The
test was conducted by suspending a two-inch cube of solidified clay, sand, and portland cement in a
glass tank filled with synthetic acid rain that was a mixture of sulfuric and nitric acids with pH 4.2. As the
alkaline components of the solid neutralized the  acidic leachant, the dye rendered the fluid a vivid
purple.  Striking color patterns revealed significant density-driven advection around the sample. A video
camera recorded these movements. This recording shows the formation of a thin (approximately 1
millimeter thick) layer of alkaline fluid (pH ~ 11)  around the sample.  This fluid, densified by leaching
products such as calcium, potassium, and sodium, drained down the side of the sample, drizzled off the
bottom edges of the sample and through the fresh acid underneath the sample, and accumulated in the
bottom of the glass tank.  Leaching continued by this mechanism until the dense alkaline fluid completely
covered the sample. Afterwards the interface between dense alkaline fluid  and fresh acid  became
diffuse and the fresh acid above the sample slowly became alkaline, indicating continued but slow
leaching.

        Existence of these advection mechanisms carries two significant implications.  On one hand in
the laboratory, the alkaline boundary constitutes  additional resistance to leaching relative to the
assumptions of the ANSI test and causes the traditional ANSI analysis to underestimate diffusivity for the
sample. On the other hand in the field, density-driven advection away from a buried monolith could
markedly exceed diffusive transport through porous media, resulting in more rapid release of
contaminants.

       For More Information: P. M. Erickson, USEPA, RREL, 5995 Center Hill Road  Cincinnati Ohio
45224, (513)569-7884.                                                                '
                                            229

-------
             ELECTRON BEAM TREATMENT FOR SUPERFUND SITE REMEDIATION

                             W.J. Cooper, M.G. Nickelsen, K. Lin
                               Drinking Water Research Center
                                Florida International University
                                     Miami, FL33199
                                       305-348-3049

                                  T.D. Waite, C.N Kurucz
                                    University of Miami
                                  Coral Gables, FL 33124
                                       305-284-3467

                                       David  V. Kalen
                           High Voltage Environmental Applications
                                       P.O. Box 8358
                                     Miami, FL 33124
                                       305-253-9143

   High energy electron beam irradiation has been studied extensively for the ultimate destruction of
toxic organic chemicals at Superfund sites. Six compounds - tetrachloroethylene, trichloroethylene,
chloroform, benzene, toluene, and phenol - commonly found at Superfund sites have been studied at
three different concentrations, three different pH levels  and in the presence and absence of solids.
Additional studies have also been conducted for methylene chloride and carbon tetrachloride.  In all
of the studies we have shown that the process effectively destroys the parent compounds and that
for the halogenated organic compounds there are no haiogenated intermediates.  We can account for
100% of the organohalogens from the parent compound as inorganic chloride ion.  The reaction by-
products that have been identified are formic acid and formaldehyde with trace quantities of other low
molecular weight organic compounds,  acids and aldehydes.  However, none of these are formed at
greater than 10% of the influent solute concentration,  and  that is only at the highest influent
concentrations of the solutes.

    Because the process produces both highly oxidizing (OH-, hydroxyl radicals) and highly reducing
{e'.,,, aqueous  electron and H-, hydrogen atom) at the  same  time and  at  approximately the  same
concentration, this process very effectively removes ail toxic organic compounds.  The process is  a
flow-through process as a result of the very rapid reaction between the radicals and toxic organics of
interest.  The process is essentially pH-independent in the pH range 3-11 and the presence of  3-5%
solids had no adverse effect on the removal efficiency  of the  solutes.  No additional chemicals are
required for pretreatment.  The process is temperature independent and produces no sludge and no air
emissions.  Because  electron beams deliver extremely high dose rates, the efficiency is high. Overall
process efficiency is 72% for conversion of electrical power into chemical energy and this decreases
the cost per unit volume. Typical treatment costs vary depending on solute concentration and matrix
from as low as $0.50 per thousand gallons to $50 per thousand gallons.

    For additional information, please contact Franklin Alvarez, Office of Research and Development,
 US Environmental Protection  Agency, 26 West Martin  Luther King Drive, Cincinnati, Ohio 45268,
 (513)569-7631.
                                            230

-------
                  ELECTRONIC COMPONENT COOLING ALTERNATIVES:
                      COMPRESSED AIR AND LIQUID NITROGEN
                             Johnny Springer, Jr.
                        Waste Minimization,  Destruction
                        and Disposal Research Division
                     Risk Reduction Engineering  Laboratory
                            Cincinnati, Ohio 45268
                                (513) 569-7542

                                 Steve  Schmitt
                                   Battelle
                                505 King Avenue
                           Columbus,  Ohio  43201-2693
                                (614) 424-4104

                             Captain  Vern  Mil hoi en
                             Newark Air Force Base
                              Newark, Ohio 43057
                                (614) 522-8053
                                   ABSTRACT
     The goal of this study is to evaluate tools used to troubleshoot circuit
boards with known or suspected thermally intermittent components.  Failure
modes for thermally intermittent components are typically mechanical defects
such as cracks in solder paths or joints, or broken bonds such as
interconnections inside integrated circuit packages or capacitors.  Spray cans
of refrigerants (CFC-12 and HCFC-22) are commonly used in electronics
manufacturing and repair businesses for this purpose, and will serve as the
benchmark for the evaluation.

     A promising alternative technology that will be evaluated in this study
is a compressed air tool which provides a continuous stream of cold air that
can be directed towards specific components.  Another alternative technology
that will be considered is a dewar that dispenses cold nitrogen gas as the
cooling agent.  The critical parameters which will  be measured for each
cooling method to provide a basis for comparison are accuracy, electrostatic
discharge risk, cooling capability, technician safety, pollution prevention
potential, and economic viability.

     Newark Air Force Base will be the site for evaluating the various
technologies.  Electronic circuit boards from a variety of Air Force Systems
are tested and repaired on a daily basis at this facility.  A percentage of
these circuit boards demonstrate thermally intermittent failure modes and will
be used for comparison testing.
                                      231

-------
         EVALUATION OF CORROSION RATES OF LEAD AND COPPER PIPES
                       BY POLARIZATION MEASUREMENT

                 Marvin C.  Garde!s  and Michael  R.  Schock
                  U.S. Environmental  Protection Agency
                  Risk Reduction Engineering Laboratory
                    Drinking Water Research Division
                       26 W. Martin Luther King Dr.
                         Cincinnati,  Ohio 45268
                            Tel: 513-569-7217
      Corrosion rate monitors based on polarization resistance have been
used in industrial facilities for several  years.   Some use two matched
test electrodes; others use three electrodes.   In all, solution
resistivity adversely affects measurement accuracy.  The advantage of
corrosion rate monitoring is that the data provides a continuous record
while coupon monitoring provides the average of the corrosion rate over
the entire exposure period.  Accurate corrosion rates provide data for
water treatment changes to maintain preset criteria.

      Corrosion rates for lead were obtained using three short sections
of lead pipe as the electrodes; a one hundred foot length of pipe was
used for the leaching studies.  Each study was designed to obtain data in
both flowing and stagnant water.  The electrochemical methods can only
obtain data in the presence of laminar flow.  Past literature suggested
using a high amount of sodium bicarbonate to reduce lead solubility in
alkaline water. Using this method only created more erratic information.
A computerized study on lead solubility indicated that the ideal pH and
alkalinity was 9.2 and 25-30 mg as CaC03/L.   Results obtained by linear
polarization at these adjustments gave very low values.  The leaching
data correlated with the electrochemical method.  These data are
illustrated.

      Similar analysis has been started using copper pipes.  TWo small
loops were built using a ten liter reservoir, small sections for linear
polarization, and a thirty inch copper pipe for leaching data.  Two
different monitors are being used; one using three electrodes previously
used on lead studies, and the other with two electrodes.  The object  is
to monitor the sample both with and without inhibitors. The same water
(Cincinnati tap water) was used in each reservoir.  Both loops were
started at the same time.  The monitors were exchanged in the loops for  >
comparison; they gave similar results.  In each loop an imbalance reading
has occurred which is often referred to as "pitting tendency".
Additional data will be collected.


For more  information:  Marvin C. Gardels, U.S.EPA,  26 W. Martin Luther
King Dr., Cincinnati, OH 452658  (513) 569-7217.
                                    232

-------
EVALUATION OF MATERIAL RECOVERY FACILITIES FOR MUNICIPAL SOLID WASTE RECYCLING

                                        Lynnann Hitchens
                                            US EPA
                                      5995 Center Hill Road
                                      Cincinnati, OH 45224
                                         (513) 569-7672
        Increasing environmental concern over the disposal of municipal solid waste in the
United States has prompted municipalities to increase recycling activities in order to
conserve natural resources and minimize the amount of solid waste requiring disposal.
Material Recovery Facilities, or "MRFs," have become a major component in solid waste
management plans. For the purpose of this study, a MRF is defined as a facility that
receives, via a commercial collection vehicle, commingled recyclables for the purpose of
separation, contamination removal, and preparation for market. These materials can
include but are not limited  to:  steel, aluminum, glass and plastic containers, newspaper
and other types of paper, and corrugated cardboard. According to the 1992-1993 MRF
Yearbook, 257 facilities are either in operation, being designed, or under construction
in the United States.  This number represents a 147% increase and is expected to grow
substantially in the coming years.

        This evaluation is being conducted under the auspices of the Municipal  Solid  Waste
Innovative Technology Evaluation (MITE) Program. The purpose of the MITE program is to
provide objective evaluations of solid waste management technologies and transfer this
information to municipalities and the public sector. The MRF evaluation will include six
facilities, differing in size, ownership, geographic area, and separation scheme.  The
purpose of this evaluation is document the potential  hazards that may be present in a
materials recovery facility, the true cost of operation, and conduct a mass balance over a
fixed period of time. The evaluation will consist of the following four components:

  •      Sampling of the air quality within and surrounding the facility.
  •      Material mass balance and calculation of real recycling rate.
  •      Energy balance, including the collection,  separation, and transportation
        requirement.
  •      Documentation of the operational costs.

        The results of this study will be used to provide municipalities and solid  waste
authorities with information on successful programs, and to assist in economical and
environmentally safe facility design and operation.

        For more information:  Lynnann Hitchens, US EPA, Risk Reduction  Engineering
Laboratory, 5995 Center Hill Road, Cincinnati, Ohio  45268, (513) 569-7672.
                                              233

-------
 FLUID EXTRACTION-BIOLOGICAL DEGRADATION OF POLYAROMAT1C HYDROCARBONS FROM
                                      TOWN GAS SOILS
                              David M. Rue and J. Robert Paterek*
                                  Institute of Gas Technology
                                    3424 South State Street
                                    Chicago, IL 60616-3896
                                       (312) 949-3947*

                                      Annette M. Gatchett
                              U.S. Environmental Protection Agency
                             Risk Reduction Engineering Laboratory
                                 26 W. Martin Luther King Drive
                                     Cincinnati, OH 45268
                                        (513) 569-7697
      The Fluid Extraction Biological Degradation (FEED)  Process developed by the Institute of Gas
Technology extracts hydrocarbon contaminants from soil and then biologically degrades the pollutants to
harmless products. The process was developed as an environmentally benign technology to safely and
economically degrade pollutants, such as polyaromatic hydrocarbons (PAHs).  This process overcomes
the limited bio-availability of these compounds while in soils.

   In the first stage of the FEBD process, contaminants are extracted  from a solid matrix using
supercritical carbon dioxide with or without a modifier. More than forty (40) laboratory-scale extraction
tests have been conducted with three contaminated soils using both carbon dioxide and carbon dioxide
containing five percent methanol  (modifier).  At optimum extraction conditions, between 91.5 and 99
percent of a suite of sixteen one to six ring PAHs was removed  from the soils.  Extractions have been
carried out at temperatures from  90F to  170F and  pressures from  1100 to 4000  psig.  High  levels of
extraction were achieved for all  compounds studied with  extraction  level  decreasing with increasing
molecular size.   Extraction levels  were greatest at  115F and increased with  increasing  pressure.
Methanol addition increased the extraction efficiency of the larger four to six ring compounds.

   Biodegradation of the extracts has  been successful in  batch and semi-continuous batch  reactors.
Reactor  types studied include constantly stirred  tank reactors and solids phase  column  reactors.
Bacterial cultures pre-grown in  PAHs have shown no lag phase before  PAH degradation commences.
The IGT bacterial cultures can  degrade total PAHs concentration ranging from 50 ppm to greater than
300 ppm. Upper concentration  limitations are being determined.  Rates of removal in extracts containing
300 ppm total PAHs reached their maximum during the first 72 hours of incubation.  Growth is supported
during this degradation period as indicated by an increase  in bacterial protein.  Percentages of growth
due to PAHs and the solvent are not known to date.

   For more information: Annette M. Gatchett, U.S. Environmental  Protection Agency, Risk Reduction
Engineering Laboratory, 26 W. Martin Luther King Drive, Cincinnati, OH 45268, (513) 569-7697.
                                             234

-------
                  MICROBIAL REMOVAL OF LEAD FROM CONTAMINATED SOIL
                                Wendy J. Davis-Hoover, Ph. D.
                        US EPA, Risk Reduction Engineering Laboratory,
          Municipal Solid Waste Research Management Branch,  Soils & Residuals Section,
                          5995 Center Hill Ave., Cincinnati, Ohio 45224
                           PHONE: 513-569-7206  FAX: 513-569-7879

                    Rebecca J. Donovan-Brand, and Stephen J. Vesper, Ph.D.
             University of Cincinnati, Department of Civil and Environmental Engineering
                          5995 Center Hill Ave., Cincinnati, Ohio 45224
                           PHONE: 513-569-7898  FAX: 513-569-7445
   Heavy metals have been effectively removed from waste waters of various sources for many years
using a number of microorganisms or components of microorganisms.  We are interested in removing
lead from contaminated soil. We have isolated a strain of Pseudofnonas aeruainosa from the vicinity of
a lead mine which is able to accumulate lead from soil and solid media.

  Yeast malt extract (YM) and R2A were contaminated with several concentrations of lead carbonate. A
piece of sterile Whatman number 1 filter paper was then placed on the agar surface and inoculated with
1 ml  of 10 7 CFU/ml of P. aeruainosa culture or 1  ml of sterile broth (as the negative control).  The
plates were placed in a 29°C incubator for various times depending on the experiment. The filter papers
were then removed and the filter paper and the residual medium were digested separately and the lead
content analyzed by atomic absorption spectroscopy (EPA Method 239.1). The amount of lead in the
sterile controls were subtracted from the treated to determine the amount of lead biologically removed.

  In  experiments using YM  medium, we found that 72% of the lead removed was removed in the first
week of incubation. The remaining 28% was removed by the end of the second week. Further
incubations for two more weeks did not result in additional lead removal.

  When the YM medium contained 1% lead carbonate, only 0.1% was removed biologically.  However, if
the concentration of lead carbonate was 0.1% then 9% was removed, and for 0.05% 17% was removed,
and for 0.01% then 33%  and for 0.005% then 66 % was removed. It appears that removal of lead may
be limited by the biomass available.  When the medium used  in this experiment was R2A, there was
essentially no removal of lead from the plates.

  We have examined YM media to determine which components were important in the lead removal
process. We did this by  removing the medium component and in other experiments by adding twice
the amount of the normal concentration in the medium. When dextrose was eliminated from the
medium, the lead removal was eliminated. However, doubling the concentration of dextrose in the
medium only increased the  lead removal about 26% above the standard medium.  We are in the process
of optimizing the medium for lead removal.

  In  experiments using soil  from an urban Cincinnati, Ohio neighborhood which contains 7,000 ppm
lead, this isolate was able to remove up to 10% of the lead when the soil was lowered to a pH of 5 but
only 8% at a pH of 6 and 5% at a pH of 6.8.
                                             235

-------
     NUTRITIONAL REQUIREMENT AND BUFFERING CAPACITY OF VARIOUS PHOSPHATE
                     SOURCES FOR OIL DEGRADING CULTURES IN SEA WATER
                                       Edith L. Holder
                                       John R. Haines
                                      Albert D. Venosa

                                    University of Cincinnati
                       Department of Civil and Environmental Engineering
                                    Cincinnati, OH 45221

                             Risk Reduction Engineering Laboratory
                             U.S. Environmental Protection Agency
                                    Cincinnati, OH 45268

       One of the priorities of EPA's oil spill bioremediation research is to define the optimum nutrient
requirements for biodegradation. The traditional phosphate source used in bacterial culture is
orthophosphate. In sea water, however, most of the orthophosphate precipitates as insoluble salts of
calcium and magnesium. This insolubility raises questions about the bioavailability of phosphorus and
suitability of orthophosphate as a buffer. This study was conducted to investigate several forms of
phosphate to find a more soluble source, to determine the minimum concentration of P needed to obtain
optimal growth of oil degraders,  and to define the buffering capacity of various phosphates in sea water.

       The phosphate sources that were studied were orthophosphate (PO,,"3), pyrophosphate (P-P/),
tripolyphosphate (PaO^-5) and urea phosphate (CH4N2O • H3PO4). The concentration used was 400 mg/L
as P unless otherwise specified. The amount of precipitate formed by each phosphate source in artificial
seawaterwas compared using orthophosphate as a  reference.  Pyrophosphate produced 140% of the
reference precipitate by weight,  and urea phosphate 90% of precipitate by weight.  Tripolyphosphate
produced no measurable precipitate at 400 mg/L as P, but at 800 mg/L it produced 40% of the reference
precipitate. Each source was titrated against 0.1 N HCI in both seawater and distilled water to test its
buffering capacity and range.  Orthophosphate had 9 times more buffering capacity than either
pyrophosphate or tripolyphosphate and double the capacity of urea-phosphate between pH 6 to 8.5. The
buffering capacity of pyrophosphate and tripolyphosphate were skewed upwards above pH 8.5 compared
to distilled water.

       Biodegradation studies were carried out in laboratory batch respirometers containing artificial
seawater with both orthophosphate and tripolyphosphate, Light Arabian crude oil, and a mixed
consortium of oil degraders from the Texas shoreline. With NH4CI as the nitrogen source, the pH of the
culture declined substantially  in the presence of tripolyphosphate. With KNO3 as the nitrogen source,
which does not result in a significant change in culture pH, the physiologically limiting concentration of P,
as determined by the maximum oxygen uptake achieved, appeared to be between 10 and 20 mg/L as P.
The cultures with tripolyphosphate consumed more O2 and evolved more CO2 (approximately 15%
difference in the 20 and 40 mg/L cultures), indicating that the phosphorus in tripolyphosphate may be
more bioavailable than in orthophosphate.

       For More Information: Albert D. Venosa, U. S. EPA, Cincinnati, OH 45268
                                             236

-------
                  OPERATING  PARAMETERS  TO  MINIMIZE  EMISSIONS
             DURING ROTARY KILN  EMERGENCY SAFETY  VENT  OPENINGS
                               Paul M. Lemieux and William P. Linak
                     Air and Energy Engineering Research Laboratory, MD-65
                             U.S. Environmental Protection Agency
                               Research Triangle Park, NC  27711

                                      Carin DeBenedictis
                         U.S. Environmental Protection Agency, Region 4
                                     345 Courtland St. N.E.
                                      Atlanta, GA  30365

                                       Jost O.L. Wendt
                              Department of Chemical Engineering
                                      University of Arizona
                                      Tucson, AZ  85721

                                       James E. Dunn
                              Department of Mathematical Sciences
                                     University of Arkansas
                                     Fayetteville, AR  72701

                                          Abstract

    A subset of hazardous waste incinerators must include emergency safety vents (ESVs).  ESVs (also
called dump stacks, vent stacks, emergency by-pass stacks, thermal relief valves, and pressure relief
valves) are regarded as true emergency devices. Their purpose is to vent combustion gases directly from
the combustion chambers to the atmosphere in the event of a failure of other system components. This is
done for operator safety as well as to protect the incinerator and other downstream equipment. ESVs are
common and typically required for rotary kiln and hearth incinerators which process a portion of their waste
load as bulk solids or contained liquids introduced continuously or in batch charges. Research has been
performed at the U.S. EPA on an 81 kW rotary kiln incinerator simulator examining optimum settings of kiln
operating parameters so as to minimize emissions during an ESV opening event. Initial results indicate
that alteration of operator-controllable kiln parameters during the onset of an ESV opening event can have
a significant effect on emissions of both organics and HCI.  A low air flow rate, possibly equal to the flow
rate induced by the natural draft coupled with air in-leakage, results in lower concentrations of both
organics and HCI. Rotational speed appears to have slightly different effects on organics and HCI.
Whereas emissions of HCI are minimized at a very low or non-existent rotational speed, emissions of
organics exhibit a minimum at a low (but non-zero) rotational speed, with increasing emissions at both zero
and high rpm. The use of a small afterburner to simulate a stack flare during an ESV event dramatically
reduced organic emissions.
                                          237

-------
                  PEROX-PURE™  CHEMICAL OXIDATION TREATMENT

                             Norma M. Lewis
                                 U.S. EPA
                       26 W. Martin Luther King Dr.
                          Cincinnati, Ohio 45268
                               513-569-7665
      The perox-pure™  chemical  oxidation treatment technology was  developed
by Peroxidation Systems, Inc. (PSI), to destroy dissolved organic contaminants
in water.  The technology uses ultraviolet (UV) radiation and hydrogen
peroxide to oxidize organic compounds present in water at parts per
million(ppm) levels.  This treatment technology produces no air emissions and
generates no residue, sludge, or spent media that require further processing,
handling, or disposal.   The technology uses medium pressure mercury vapor
lamps to generate UV radiation.  The principal  oxidants in the system,
hydroxyl radicals, are produced by direct photolysis of hydrogen peroxide and
UV wavelengths.

      The perox-pure™  technology  has  been  used to treat  leachate,  ground
water, and industrial wastewater all containing a variety of organic
contaminants, including chlorinated solvents, pesticides, polynuclear aromatic
hydrocarbons, and petroleum hydrocarbons.  The technology was applied at
Lawrence Livermore National Laboratory Site 300 in,Tracy, California in
September 1992.  The technology was conducted in three phases.  Phase 1
consisted of eight runs, Phase 2 consisted of four runs, and Phase 3 consisted
of two runs.  About 40,000 gallons of ground water contaminated with volatile
organic compounds (VOC) was treated.  The principal  ground water contaminants
were trichloroethene (TCE) and tetrachloroethene (PCE), which were present at
concentrations of about 1,000 and 100 micrograms per liter (ug/L),
respectively.  Ground water was pumped from two wells into a 7,500 gallon
bladder tank to minimize any variability of influent characteristics.  Treated
ground water was stored in two 20,000 gallon steel tanks and analyzed before
being discharged.
                                      238

-------
  PREDICTING  RESISTANCE OF  CHEMICAL  PROTECTIVE CLOTHING  USING NEURAL
                                COMPUTING  TECHNIQUES
                                 Ypng-LinHu, William G. Wee
                       Artificial Intelligence and Computer Vision Laboratory
                        Department of Electrical and Computer Engineering
                                    University of Cincinnati
                                    Cincinnati, Ohio 45221
                                   "Phone: (513) 556-4778
                                            and
                                      John J. Coleman
                             Risk Reduction Engineering Laboratory
                              US Environmental Protection Agency
                                    Cincinnati, Ohio 45224
                                        ABSTRACT

       In compliance with the Premanufacture Notification (PMN)  program,  EPA's Office  of Toxic
Substances must assess potential hazards posed by the manufacture of new chemicals. Approaches to
assess clothing performance have emphasized the prediction of permeation behavior.  Neural Computing
Networks are capable of "learning" the best way to combine pertinent input information into predicting the
related outcomes without specific knowledge of the input-output relations.  Predicting resistance of
chemical protective clothing seems a good application for neural networks.  The objective of the project is
to use neural computing technique to predict resistance of chemical protective clothing polymers to
various chemical compounds and to compare these predictions with other methods

       Back propagation neural networks were used to predict the steady-state permeation rates of
chemicals/clothing polymers. Experimental values of steady-state permeation  rates were used as the
target values corresponding to the physical and chemical properties of chemicals which can be found in
tables arranged  by "chemical names" and/or by Chemical Abstracts Service(CAS) Registry Numbers. In
the current neural network systems the values of molecular weight, vapor pressure, and density were
used as inputs.  The steady-state permeation rates were the outputs of the networks. These synthesized
neural networks' results were, in turn, compared with calculated results using conventional methods.

       Preliminary results of using neural computing to predict the  resistance of chemical  protective
clothing polymers (butyl rubber and natural rubber) to various chemical compounds are very satisfactory.
When compared to the existed EPA approaches, the neural computing networks provide much better
estimation of permeation rate. The improvement in accuracy and predictability over the ADL/OTS UNIFAP
and Equation of State (EOS) procedures is  as much as one to two orders of magnitude. The neural
network approaches appear feasible for the protective clothing problem.
For More Information:
                          John J. Coleman
                          RREL/USEPA
                          5995 Center Hill Rd
                          Cincinnati, OH 45224
                          Phone: (513)  556 7464
                                         239

-------
  PREDICTING SOLIDIFICATION/STABILIZATION PERFORMAN
-------
   RD&D under the Tidewater Interagency Pollution Prevention Program (TIPPP)

                               Kenneth R.  Stone
                     Risk Reduction Engineering Laboratory
                       26 West Martin Luther King Blvd.
                            Cincinnati, Ohio  45268


    The concept of TIPPP is to take advantage of the capabilities of well-
defined communities to develop an integrated multi-media pollution prevention
plan with results that are transferable to other communities.  TIPPP will
sponsor a number of joint EPA/DoD/NASA Community R&D projects that require
demonstration before being accepted within the public and private sectors.

    Four bases are participating in this study: Fort Eustis (Army); Sewell's
Point Naval Base; Langley Air Force Base;  NASA-Langley Research Center.
Research plans have been developed with the Bases to guide decision-making for
future cooperative RD&D.  Further PPOAs and RD&D efforts continue after the
TIPPP plans have targeted the areas of concern.  In these instances, each RD&D
project will be developed and initiated separately between ORD and the
requesting base.

    For example, as a result of the cooperation between NASA and EPA/ORD on
the Langley Research Center's plan, ORD is funding a technology demonstration
of an innovative process for making composites.  At Federal facilities and in
commercial plants, advanced composite materials are typically produced through
hot melt and solution prepregging (The F-22 Advanced Tactical Fighter (ATF)
features larger expanses of composite material produced under the hot melt
process).  This project will compare the three processes (hot melt, solution
and dry powder), using-life cycle methodology to ascertain the energy and
environmental impacts of each, from the production of the polymer resins
through the manufacture of the material.  Final products will be tested to
determine comparable performance characteristics.

    Pollution prevention projects are currently being planned with Ft. Eustis
and Langley AFB.  Naval Base Norfolk is developing pollution prevention
projects through the Naval Energy and Environmental Support Activity.
                                       241

-------
r
                                         RREL  REMEDY SCREENING PROGRAM

                                                Eugene Harris3
                                                (513)  569-7862
                                                 Joan  Mattox9
                                                (513)  569-7624
                                                      and
                                University of'Cincinnati  Staff  at Center Hillb
                                                (513)  569-7885
                                          aU.  S.  EPA RREL Laboratory
                                           Cincinnati, Ohio  45268
                                bCivil  and  Environmental Engineering  Department
                                           University  of  Cincinnati
                                            Center Hill Laboratory
                                            5995  Center Hill Road
                                           Cincinnati, Ohio   45224
                                                   Abstract

                      The RREL START screening program has entered its third year at the
                United States Environmental Protection Agency in Cincinnati, Ohio.  The
                Screening Program is designed to assist the Regional Project Managers in the
                EPA regions with the decisions about the technology to be used  in the clean-up
                of National Priority List sites.  The program provides the technical
                experience of the teams of EPA experts and a state-of-the-art laboratory
                facility in which to perform the tests.  The result is unbiased  indicator data
                that should help in the remedy selection process.

                      We presently have the ability to perform eight standard protocols.
                These include Solidification/Stabilization of Inorganics, Thermal Desorption,
                Chemical Dehalogenation, Biotreatability, Soil Washing,  Soil Flushing. Solvent
                Extraction, and Soil'Vapor Extraction.  We will be adding Solidification/
                Stabilization of Organics and we are considering the possibility of adding
                Vitrification.  The screening program is designed for  rapid turnaround,  i.e.
                no more than three months from the time a sample is received until  the report
                is complete.

                      We have now performed these protocols on soils from more  than a dozen
                sites from Alaska to Florida.  We anticipate testing many more  sites this year
                and continuing to improve all aspects of the program.
                 For more information:
EPA Work Assignment Manager
Eugene Harris,  START,  RREL, WMDDRD
26 W.  Martin Luther King Drive
Cincinnati,  Ohio   45268
(513)  569-7206
                                                    242

-------
            RREL UNDERGROUND STORAGE TANK RESEARCH PROGRAM OVERVIEW

                              Anthony N. Tafuri
                     U.S. Environmental  Protection Agency
                    Risk Reduction Engineering Laboratory
                            2890 Woodbridge Avenue
                              Edison,  NJ   08837
                                (908)  321-6604

     The overall goal  of RREL's Underground Storage Tank (UST) Research
Program is to identify and evaluate reliable, cost-efficient techniques and
equipment for preventing, detecting,  and locating leaks in underground tanks
and pipelines, and for cleaning up contamination  at leaking UST sites.

     Early R&D activities at RREL's full-scale UST Test Apparatus in Edison,
NJ concentrated on evaluating and verifying the performance of internal leak
detection methodologies for tanks and pipelines.   Results provided the basis
for technically defensible and achievable regulatory standards and data for
industry to design and develop leak detection systems that meet or exceed
these standards.

     Federal requirements for leak detection and  corrective action at UST
sites went  into effect in December 1988 at 750,000 facilities nationwide.
RREL participated in a coordinated research effort to develop and evaluate
standard test procedures for determining the performance of the most common
types of leak detection methods.  During this period, a national program for
certifying  leak detection methods was begun.

     As leak detection systems were installed at UST sites, an increasing
number of releases were discovered.  With this increase came a corresponding
need for quicker, cheaper and higher quality cleanups.  Accordingly, RREL
shifted its research emphasis to site remediation.  Efforts to locate the
sources of  small leaks in UST systems more accurately, especially in
pressurized systems, were initiated; a new methodology for selecting cleanup
technologies at leaking UST sites was developed;  and technologies such as soil
vapor extraction (SVE), soil washing, thermal desorption, and soil reuse were
investigated.

     More recently, SVE, coupled with other components such as air sparging
and biotreatment, and  improved product-removal techniques for light
nonaqueous-phase liquids are being investigated as an integrated systems
approach to cleaning up  leaking UST sites.  Ex-situ treatment technologies,
such as thermal desorption and biological treatment in soil piles, are also
being investigated.  Other technologies, such as steam stripping, radio
frequency heating, soil flushing, chemical oxidation, and solvent extraction
will be screened for their application to leaking UST sites, and those that
appear to be the most  promising will be selected for further development and
evaluation.  Research  efforts will also address contamination in fractured
rock media.
                                       243

-------
     RREL's research activities are also addressing the compliance needs of
the regulated community in areas such as "large" tanks, USTs containing
chemicals, cathodic protection, flexible pipelines, new installations, etc.
New and improved techniques for determining whether the system is leaking or
is about to leak, and for locating leaks, will also be developed and
evaluated.

For More Information:
Anthony N. Tafuri (908) 321-6604
USEPA, Risk Reduction Engineering Laboratory
2890 Woodbridge Avenue, Edison, NJ   08837
                                       244

-------
     REMEDIATION OF SOLIDS WITH TOXIC ORGANICS AND METALS - THE MelDAS PROCESS

                 Srivats Srinh/asachar, Joseph R. Morency and Barbara E. Wyslouzil
                                    PSI Technology Company
                                20 New England Business Center
                                      Andover, MA 01810

       Forty percent of Superfund sites have both heavy metal and organic contaminants. While high
temperature treatment, such as incineration, is effective for remediating organic wastes, it creates
problems when metals are present; airborne emissions of metal aerosols and teachable metals in the
captured flyash.  PSI Technology Company's  (PSIT) metals immobilization and decontamination of
aggregate solids (MelDAS) process addresses both these problems. This technology was accepted Into
the SITE Emerging Technology Program in July 1991.

       In one version of the MelDAS process, the contaminated soil is treated in an thermal unit (e.g.
incinerator) in conjunction with sorbents.  High temperatures present during processing destroy the
organics, while simultaneously encapsulating the metals into a form that is non-leachable.

        Bench-scale testing, completed in August of 1992, focussed on surrogate wastes with lead and
arsenic. Three experimental tasks were performed; determination of metal vaporization rates, adsorption
of arsenic vapor from flue gases, and immobilization of metals associated with flyash.  The vaporization
experiments examined the effect of temperature, HCI concentration and CO/Cfe ratio on vaporization
rates. The arsenic vapor adsorption experiments, configured to simulate a baghouse, identified several
sorbents that were effective. The immobilization of leachable metals present in fly ash was achieved by
processing the ash in combination with sorbents. Test results indicate that a much lower amount of the
sorbent (1-10  % of soil) is required for effectively immobilizing the metals than previously anticipated.

       Pilot-scale testing of the MelDAS process is being conducted at the rotary kiln simulator at EPA's
Air & Energy Engineering Research Laboratory in Research Triangle Park, NC. The soil being tested, is
from the Baird & McGuire site in Holbrook, MA,  and contains lead, arsenic, and pesticides (e.g.
methoxychlor).  Results from these tests will enable the design of a cost-effective thermal system for
treating wastes with organics and metals.
       For More Information:

       Mark Meckes
       U.S. EPA, Risk Reduction Engineering Laboratory
       26 West Martin Luther King Drive
       Cincinnati, OH 45268
       (513) 569-7348
                                            245

-------
     REVERSE OSMOSIS: PESTICIDE AND ORGANICS REMOVAL FROM DRINKING WATER

                                             by

                                      Carol Ann Frank
                               Environmental Protection Agency
                               Drinking Water Research Division
                                   26 West M. L. King Dr.
                                   Cincinnati, Ohio 45268
                                       (513)569-7592

       Approximately 21 billion pounds of pesticides have been applied to United States farmlands since
1964. In agricultural regions, high pesticide concentrations occur in surface and groundwaters because of
spring runoff or leaching. In addition large volumes of solvents are manufactured in the US each year. Many
of these pesticides and solvents, which are synthetic organic contaminants (SOC's), have been detected in
public water systems.   Because certain of these compounds  pose health risks, the  United States
Environmental Protection Agency (USEPA) has a mandate under the Safe Drinking Water Act Amendments
of 1986 to regulate them by establishing Maximum Contaminant Levels (MCL's).

       When drinking water supplies are contaminated and alternative sources of water are unavailable,
treatment of existing sources becomes imperative. The Drinking Water Research Division within the Risk
Reduction Engineering  Laboratory of the USEPA,  Cincinnati,  is responsible  for evaluating various
technologies that may be feasible for meeting the MCL's. Reverse Osmosis (RO) is commonly defined as
diffusion, with applied pressure, through a semipermeable membrane. Historically it has been associated with
removal of salts and other inorganic compounds from water. Thin-film-composite membranes, introduced
in the  1970's, have shown promise for removing certain low-molecular-weight organics  (Molecular
Weight<200). Thus, over twenty volatile compounds, as well as pesticides, were treated using several types
of reverse-osmosis  membranes.

       In an attempt to understand possible compliance problems, an investigation was conducted on river
watercontaining Alachlor, Atrazine, Cyanazine, Linuron, Metolachlor, Metribuzin, and Simazine. The purpose
of this study was to determine to  what extent reverse osmosis was able to remove pesticides from the
Sandusky River at the Tiffin, Ohio, Water Treatment Plant. In addition, pilot-scale studies were conducted
using several different types of polymeric  membranes to remove pesticides from spiked groundwater. In-
house studies also were conducted on spiked  distilled water and spiked groundwater to investigate the
removal of over 20 SOC's, using various reverse osmosis membranes.

        For More Information: Carol Ann Frank, United States Environmental Protection  Agency, Risk
 Reduction Engineering Laboratory, 26 West M.L. King Dr., Cincinnati, OH 45268, (513) 569-7592.
                                          246

-------
                RISK  REDUCTION  ENGINEERING  LABORATORY  (RREU.
               DRINKING WATER TECHNOLOGY ACTIVITIES  -  U.S.  EPA
                     Walter A.  Feige  and Clois J.  Slocutn
                    U.  S.  Environmental Protection Agency
                        26 W.  Martin Luther King Drive
                            Cincinnati, OH  45268
                      Tel. 513/569-7496 and 513/569-7281
     The poster conveys by pictures and words the mission of EPA's Drinking
Water Research Division (DWRD), located in Cincinnati, Ohio.  A mission
statement, as well as pictures representing the activities of DWRD's four
Branches, is presented.  Handouts will be distributed that will denote
specific research areas and list the Division's contact persons by expertise.

     DWRD plans, coordinates and conducts a national program to provide the
technology necessary to help prepare the primary and secondary regulations for
drinking water.  The Division has the responsibility for a program that
integrates chemistry, engineering, microbiology, and cost-effective techniques
(acquisition, treatment, distribution, and support services) for assuring the
delivery of safe .drinking water to. reduce the risk of chemically and
microbiologically induced health effects to the public.

     The Division conducts both intramural and extramural research to
establish practices for the control and removal of contaminants, and for the
prevention of water quality deterioration during storage and distribution in
the most economical manner.  The Division operates its own inhouse pilot
facilities, chemistry, and microbiology laboratories, as part of the
intramural program.

     Major areas of research include disinfectants, disinfection by-products,
corrosion control, small system technologies, distribution systems, bacteria,
viruses and protozoa.

     For More Information: Walter A. Feige, U.S. EPA, 26 W. Martin Luther King
Drive, Cincinnati, OH  45268, Tel. 513/569-7496.
                                    247

-------
                            SOWAT: SEQUENCE OPTIMIZER FOR
                                WASTEWATER TREATMENT
                 S. Krowidy, W.G. Wee, Dept. of Electrical & Computer Engineering
                R.S. Summers, M. Suidan, Dept. of Civil & Environmental Engineering
                                   University of Cincinnati
                                 J.J. Coleman & L. Rossman
                                     USEPA, Cincinnati
                                        ABSTRACT

       SOWAT takes the concentrations of the contaminants in the wastewater along with its BOD, COD,
pH and suspended solids. SOWAT employs a two phase approach to solve the wastewater treatment
problem. The first phase, called analysis phase, is designed as a knowledge acquisition tool. The second
phase, called synthesis phase, is designed to generate the treatment trains by using the knowledge
extracted from the analysis phase.

       The analysis phase analyzes the database and integrates it with other expert knowledge including
treatment ordering and exclusion rules. The treatability properties of each technology are obtained from
analyzing the treatability database. To account for the unreliability of the data we use fuzzy relationships to
describe treatability properties of the technologies. As a result the effluent concentrations  of the
compounds are obtained as fuzzy membership functions. The synthesis phase takes a set of input
concentrations and generates treatment trains using the results from the analysis phase with heuristic
search approach. It develops several alternative sets of treatment trains to reduce the concentrations of
the contaminants and parameters to a specified limit. These alternative set of treatment trains are
generated in the increasing order of cost. SOWAT is built with the RREL (Risk Reduction Engineering
Laboratory) treatability database and integrated with expert knowledge. SOWAT is developed  under an
MS-DOS environment.
                                                  USER
                                               INTERFACE
  Database
       	1

  Expert
Knowledge
                     ANALYSIS
                       PHASE
SYNTHESIS
  PHASE
Output Train
                           Rgure 1. Block Level Design of SOWAT.
For More Information:
                          John J. Coleman
                          RRELAISEPA
                          5995 Center Hill Road
                          Cincinnati, Ohio 45224
                                           248

-------
                              SITE REMEDIATION TECHNICAL
                                 RESOURCE DOCUMENTS

                                        Ben Blaney
                                     U.S. EPA. RREL
                                   Cincinnati. OH 45268
                                       513-569-7406
       RREL has a variety of technology transfer activities to provide information on contaminated site
remediation to EPA Remedial Program Managers (RPMs), RCRA Corrective Action staff and other site
remediation managers. The Laboratory has produced a number of engineering technical resource
documents, including:

       •      Treatability study guidance  documents,

       «      Documents on remedial options for common categories of contaminated sites (e.g.,
              wood preserver sites), and

       •      Engineering bulletins and issue papers.

This poster will describe these and other RREL site remediation technology transfer activities, providing
samples of each and means to order documents.
                                            249

-------
      SLOW-RELEASE OXYGEN SOURCE FOR BIOREMEDIATION IN SUBSURFACE SOILS

                                     Stephen J. Vesper3*
                                       (513)569-7898
                                    Lawrence C. Murdoch3
                                       (513) 569-7897
                                         Sam Hayes3
                                       (513) 569-7698
                                   Wendy J. Davis-Hoover*3
                                       (513) 569-7206
                        aCivil and Environmental Engineering Department
                                    University of Cincinnati
                                    Center Hill Laboratory
                                    5995 Center Hill Road
                                   Cincinnati, Ohio  45224
                               bU. S. EPA Center Hill Laboratory
                                    5995 Center Hill Road
                                    Cincinnati, Ohio 45224

                                    Corresponding author
                                          Abstract

       Two solid peroxides, calcium peroxide and sodium percarbonate, were encapsulated in
polyvinylidene chloride to determine their potential as slow-release oxygen sources during
biodegradation of contaminants in subsurface soils. In laboratory studies under aqueous conditions, the
encapsulated peroxides were estimated to release oxygen over about a two month period.  These
investigations, employing two bacterial isolates, demonstrated that peroxide toxicity was markedly
reduced for the microencapsulated compounds.  The reduction in toxicity demonstrated by the
encapsulated peroxides may be due, in part, to the decreased rate of release of hydrogen peroxide. In
laboratory studies, the encapsulated sodium percarbonate was used to provide oxygen as an electron
acceptor for microorganisms during the biodegradation of propylene glycol. In 30 days at 12°C
(subsurface soil temperature), the concentration of propylene glycol was reduced 10-fold, the number of
propylene glycol degrading organisms increased 10-fold, and fluorescein diacetate hydrolysis increased
4-fold when compared to controls. Acidic soil conditions (pH 4.7) were also neutralized to a pH of about
8.3 by the encapsulated peroxides.
For more information:
EPA Work Assignment Manger
Wendy Davis-Hoover, RREL, WMDDRD
Center Hill Research Facility
5995 Center Hill Road
Cincinnati, Ohio  45224
(513) 569-7206
                                          250

-------
             THE LEACHATE-RECIRCULATING MUNICIPAL SOLID WASTE LANDFILL

                                      Debra R. Reinhart
                        Civil and Environmental Engineering Department
                                  University of Central Florida
                                      P.O. Box 162540
                                   Orlando, Florida 32816
                                       (407)  823-5783

       Moisture addition has been demonstrated repeatedly to have a stimulating effect on
biological stabilization of waste placed in landfills. Leachate recirculation appears to be the most
effective method to increase moisture content in a controlled fashion. The advantages of leachate
recirculation include distribution of nutrients and enzymes, pH buffering, dilution of inhibitory
compounds, recycling and distribution of methanogens, liquid storage, and evaporation
opportunities.  It has been suggested that leachate recirculation could reduce the time required for
landfill stabilization from several decades to two to three years thus minimizing opportunity for long
term adverse environmental impact.

       While the effectiveness of leachate recirculation has been well documented  in lysimeter
studies, methods to accomplish it in the field are still evolving.  Prototype methods which have
been utilized include spraying, surface ponds, vertical injection  wells with and without wicks, and
horizontal surface infiltration devices.                         ,

       Because it appears to have greater potential for accelerating waste stabilization than any
other single mechanism, leachate recirculation continues to be pursued at field scale throughout the
world. The US EPA is actively involved in research addressing leachate recirculation,  providing
funds for programs at two sites, the Mill Seat Landfill in Monroe County, New York and the
Southwest Landfill in Alachua County, Florida.

       Successful application of leachate recirculation requires that all aspects of moisture
management and landfill processes be considered. For example, adequate leachate storage during
early operational phases is essential until sufficient waste is present to accept the accumulated
leachate.  As experience is gained  in designing and operating leachate recirculation systems, their
use will become routine and widely accepted by regulators and operators.

For More Information:  David A. Carson, Office of Research and Development, RREL, 26 West
                      Martin Luther King Drive, Cincinnati, Ohio 45268
                                             251

-------
                   THE U.S. EPA INCINERATION RESEARCH FACILITY

                                 J. W. Lee and S. Venkatesh
                              Acurex Environmental Corporation
                                 3900 NCTR Road,  BIdg. 45
                                  Jefferson, Arkansas 72079
                                       (501) 541-0004

                                       R. C. Thurnau
                Waste Minimization, Destruction, and Disposal Research Division
                            Risk Reduction Engineering Laboratory
                                    Cincinnati, OH 45268
                                       (513) 569-7692

       The U.S. EPA's Incineration Research Facility (IRF) located in Jefferson, Arkansas, houses
bench-scale and pilot-scale incinerators including a rotary kiln system (RKS) and a thermal
treatability unit (TTU). The RKS consists of a 2.0 MMBTU/hr rotary kiln primary chamber, a gas-
fired 1.5 MMBTU/hr afterburner, and primary and  secondary air pollution control systems.  It is
capable of incinerating solid and liquid wastes. The TTU is a small commercial pathological waste
incinerator converted to allow, thermal  treatment of liquid and solid wastes in a semi-batch
continuous mode.  Quartz trays (8-in x 10-in x 2-in) containing up to 2 Ib of test materials each can
be subjected to thermal treatment conditions for 20 min to 2 hr.

       In FY92, the IRF completed three test programs in the RKS, including a treatability study of
two sludges contaminated with hazardous organics and trace metals from the Bofors-Nobel
Superfund site; a treatability study of PCB-contaminated soil from the Scientific Chemical
Processing Superfund site; and a research program to  study the effects of waste feed  cut-offs on
particulate, HC1, trace metal, and organic constituent emissions from incinerators.  These test
programs supported EPA Regional offices and the Office of the Solid Waste.

       Work with the TTU provided data for remedy screening of thermal treatment technologies.
One test program in FY92 was an evaluation of the thermal treatability of organics, pesticides and
trace metals in contaminated soil from the Popile and American Creosote Superfund sites in support
of the Superfund Technical Assistance Response Team (START) program. Another test program
assisted the U.S. Army Corp of Engineers, Waterways Experiment Station to evaluate the thermal
treatability of hazardous organic- and PCB-contaminated sediments from New  York Harbor.

       Ongoing test programs at the IRF include an extensive study of the effectiveness of organic
decontamination and the fate of contaminant metals when a contaminated soil is treated in a rotary
kiln operated at low to moderate temperatures associated with thermal desorption processes; and an
evaluation of the incinerability of wastes which simulate the Westinghouse Savannah  River Plant
low-level radioactive wastes.

       Results from the previous year's tests and current year activities will be highlighted in the
poster presentation.

       For more information contact R. C.  Thurnau,  the EPA project officer  at the above address.
                                          252

-------
                     THERMALLY INDUCED WATER MOVEMENT WITHIN
                          GEOMEMBRANE COVERED CLAY LINERS

                   Robert Horton, Joseph G. Benjamin, and Ibrahim N. Nassar
                                  Department of Agronomy
                                    Iowa State University
                                      Ames. IA 50011
                                       (515) 294-7843
       Water transfer in closed compacted soil columns was observed for two soil types exposed to
two different nonisothermal conditions (low and high surface temperature amplitude).  Clarinda clay and
Fayette silty clay loam soils were wetted to their optimum soil water contents. The wetted soils were
packed into PVC columns having 0.077-m diameter and 0.300-m length.  A mechanical device (Instron,
Model 1125) was used to compact Clarinda and Fayette  soil to bulk densities of 1.45 and 1.67 mg/m3,
respectively. The low temperature amplitudes were less  than 10 degrees centigrade while the high
temperature amplitude was about 15 degrees centigrade. A numerical model of coupled heat and
water transfer was used to predict soil water and temperature distributions. The model predicted soil
temperature well in comparison with observed values for both high and low temperature amplitudes.
Under low temperature amplitude, predicted values of soil water content for Clarinda and Fayette soils
compared similarly to the observed values. Under high temperature amplitude, the model
overestimated water transfer in comparison with observed values.

       For more information: Brunilda Davila, U.S. EPA, RREL, 5995 Center Hill Avenue, Cincinnati,
OH 45224 (513-569-7849).
                                           253

-------
                     USE OF SECONDARY LEAD SMELTING TECHNOLOGY
                              FOR THE RECLAMATION OF LEAD
                         FROM LEAD-CONTAINING SUPERFUND SITES
                                      Stephen W. Paff

                           Center for Hazardous Materials Research
                                    320 William Pitt Way
                                    Pittsburgh, PA 15238
                                      (412) 826-5320

      Lead is one of the most commonly found waste materials at Superfund sites. The most
common method employed to remediate the lead is stabilization and disposal.  The Center for
Hazardous Materials Research (CHMR) and Exide Corporation  (a major secondary lead smelter) are
investigating the potential for using secondary lead smelters for the recovery of lead from  battery
cases and other materials removed from Superfund sites. The purpose  of this investigation is to
determine if lead can be reclaimed from these materials, which typically contain lead in
concentrations of 1 % to 10%.  Ultimately, the project is intended to develop an economical
alternative to disposal.

      As part of the investigation, CHMR/Exide identified several prospective Superfund sites with
lead-containing waste battery cases, dross, slag, lead oxides, and debris. In addition, CHMR/Exide
identified other sources of lead including  residential housing rehabilitation and bridge deck blasting
operations.  CHMR/Exide processed materials from three Superfund sites, as well as one residential
house, in the blast and reverberatory furnaces at Exide's Reading, Pennsylvania secondary lead
recycling facility.  Lead was successfully reclaimed from all materials processed.  CHMR measured
the feed rates, effect on the furnaces, effect on lead quality and quantity and effect on emissions
from the smelter to determine lead recovery economics.

      Results to date indicate that it is possible to reclaim lead from a variety of materials using
secondary lead smelting technology.- CHMR/Exide  have determined that recovery of lead in a
smelter is an economical alternative to landfilling in most cases for battery cases, lead dross, slag
and other lead materials containing over 10% lead.  The economics of recovering lead from
demolition materials from housing have not been determined.

      CHMR/Exide will  continue to perform experiments to  determine the economics of lead
recovery, maximum feed rates into a furnace and full breadth of the use of the technology.
Materials which may be tested during the second year include lead-containing bridge paint
abrasives, slag, battery cases intermingled with dirt and debris, and possibly lead-contaminated
soils.
                                            254

-------
             X-RAY TREATMENT OF ORGANIC WASTE IN AN AQUEOUS MATRIX

                    Randy Curry, Carl Eichenberger, Heinz Lackner, Bob AKes
                                  Titan/Pulse Sciences, Inc.
                                    600 McCormick Street
                                   Sari Leandro, CA 94577
                                       (510) 632-5100

                                   Esperanza Piano Renard
                              United States Environmental Agency
                             Risk Reduction Engineering Laboratory
                          Releases Control Branch, Edison, NJ 08837

                                   John Bayless, Consultant
                                     John R. Bayiess Co.
                                    20325 Seaborad Road
                                      Malibu, CA 90265

                                   Gilbert Yang, Consultant
                                   Department of Chemistry
                                University of Southern California
                                    Los Angeles, CA 90089

       The application of x-ray treatment to soil and water contaminated with volatile and semi-volatile
organic compounds (VOC/SVOC) is currently being evaluated under the Superfund Innovative
Technology Program (SITE). The x-ray (bremsstrahlung radiation) treatment technology was developed
by Titan/ Pulse Sciences Inc. (PSI).

       The technology is based on the in-depth deposition of ionizing radiation.  The interaction of ener-
getic photons with matter generate secondary electrons within the contaminated waste material. The
electrons break up the water molecules forming OH radicals which break up the VOC and SVOC con-
taminants. The resultant by-products are believed to be in the form of water, carbon dioxide and oxygen.

       The physical mechanism by which VOCs and SVOCs are removed is primarily dependent on the
matrix or substrate. For example, the primary reactant in oxygenated Water is the hydroxyl radical. This
chemical reaction is being evaluated to understand the reaction and to determine scaling of the x-ray
treatment process for pilot demonstration.  Several compounds commonly found in both water and soil
are currently being evaluated. These are trichloroethylene (TCE), tetrachloroethylene (PCE), benzene,
toluene, and carbon tetrachloride.

       An electron beam, 55 nsec in duration, is generated with a linear induction accelerator (LIA). X-
rays are generated using a 1.2 -1.4 MeV, 800 A LIA, the high energy pulse of electrons is directed onto a
high-Z tantalum target which converts the electron pulse into x-rays (bremsstrahlung radiation).

       Preliminary results of the compounds evaluated in an aqueous matrix are reported, concentra-
tions of 1600-2100 ppb TCE, 160-240 ppb benzene have been completely mineralized at an x-ray dose
of less than 20 and  6-9 Krads respectively.  Carbon tetrachloride appears to require substantially higher
irradiation doses than TCE or benzene. A175 ppb carbon tetrachloride concentration was reduced to 14
ppb with a x-ray dose of 200 Krads.  Other compounds are being investigated, additional results will be
mentioned at the  poster session..

       For more information contact Randy Curry, Titan/PSI, 600 McCormick St., San Leandro, Ca.,
94577(510)632-5100.
                         »U.S. GOVERNMENT PRINTING OFFICE:! 99 3 -750 -002/EOmit
                                           255

-------

-------

-------

-------

-------
United States
Environmental Protection Agency
Center for Environmental Research Information
Cincinnati, OH 45268
Official Business
Penalty for Private Use
$300
EPA/600/R-93/040
















C O. =S
t* o 3
-1
o 3
g w o
* 8 3
m
•^ ?"
2? 
II
1^
5 O
W X
|S
8 x
5" m
3m
P
ll?
||g
BoS.
3-S »
S^l
to
Q
38
3-i
o g.
*^1
0* 
-------