vvEPA
United States
Environmental Protection
Agency
Risk Reduction
Engineering Laboratory
Cincinnati, OH 45268
EPA/600/9-90/006
Feb. 1990
Research and Development
Remedial Action,
Treatment and Disposal
of Hazardous Waste
Proceedings of the
Fifteenth Annual
Research Symposium
-------
-------
EPA/600/9-90/006
Feb. 1990
REMEDIAL ACTION, TREATMENT, AND DISPOSAL
OF HAZARDOUS WASTE
Proceedings of the Fifteenth Annual Research Symposium
Cincinnati, OH, April 10-12, 1989
Sponsored by the U.S. EPA, Office of Research & Development
Risk Reduction Engineering Laboratory
Cincinnati, OH 45268
Coordinated by:
JACA Corp.
Fort Washington, PA 19034
and
PEI Associates
Cincinnati, OH 45246
Project Officers:
Eugene F. Harris
John Glaser
Teri Shearer
Cincinnati, OH 45268
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OH 45268
-------
NOTICE
These Proceedings 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.
-11-
-------
FOREWORD
Today's rapidly developing technologies and industrial practices fre-
quently 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 pro-
tecting 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 demonstra-
tion programs. These provide an authoritative, defensible engineering
basis in support of the policies, programs, and regulations of the EPA with
respect to drinking water, wastewater, pesticides, toxic substances, solid
and hazardous wastes, and Superfund-related activities. This publication
is one of the products of that research and provides a vital communication
link between researchers and users.
These Proceedings from the 1989 Symposium provide the results of proj-
ects recently completed by RREL and current information on projects pre-
sently 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 com-
munity. The 1989 symposium attracted over 900 attendees from industry,
Federal and State agencies, consulting firms, and universities. The 1990
symposium is planned for April 3, 4, and 5 in Cincinnati, OH.
E. Timothy Oppelt, Director
Risk Reduction Engineering Laboratory
-iii-
-------
ABSTRACT
The Fifteenth Annual Research Symposium on Remedial Action, Treatment,
and Disposal of Hazardous Waste was held in Cincinnati, OH, April 10-12,
1989. The purpose of this Symposium was to present the latest significant
research findings from ongoing and recently completed projects funded by
the Risk Reduction Engineering Laboratory (RREL).
These Proceedings are organized in four sections: Sessions A, B, and
A/B consist of paper presentations. Session C contains the poster
abstracts. Subjects include remedial action treatment and control tech-
nologies for waste disposal, landfill liner and cover systems, personnel
protection, underground storage tanks, and demonstration and development of
Innovative/alternative treatment technologies for hazardous waste. Alter-
native technology subjects include thermal destruction of hazardous wastes,
field evaluations, existing treatment options, emerging treatment pro-
cesses, waste minimization, and biosystems for hazardous waste destruction.
-------
CONTENTS
SESSION A
Page
Development and Validation of a Surrogate Metals Mixture
R.G. Barton, Energy and Environmental Research Corporation. 1-10
Incinerating Ethylene Dibromide and Dinoseb Stocks
Donald A. Oberacker, U.S. Environmental Protection Agency.. 11-20
Pyrolytic Thermal Degradation of a Hazardous Waste Incinerability
Surrogate Mixture
D.A. Tirey, University of Dayton Research Institute 21-31
Incinerability Ranking of Hazardous Organic Compounds
Robert E. Mournighan, U.S. Environmental Protection Agency* 32-42
A Prototype Baghouse/Dilution Tunnel System for Particulate
Sampling of Hazardous and Municipal Waste Incinerators
P.M. Lemieux, U.S. Environmental Protection Agency 43-49
Evaluation of Alternative Treatment Technologies for Hazardous
Wastes from Acrylonitrile Production
E. Radha Krishnan, PEI Associates, Inc 50-63
The Role of Site Investigation in the Selection of Corrective
Actions for Leaking Underground Storage Tanks
Myron S. Rosenberg, Camp Dresser & McKee Inc 64-82
Summary of the Results of EPA's Evaluation of Volumetric
Leak Detection Methods
Joseph W. Maresca, Jr., Vista Research, Inc 83-98
An Outreach Process: Case Histories of Underground Storage
Tank Corrective Actions
William M. Kaschak, COM Federal Programs Corporation. 99-108
Considerations of Underground Storage Tank Residuals at Closure
Warren J. Lyman, Camp Dresser & McKee, Inc 109-123
Laboratory Studies of Vacuum-Assisted Steam Stripping of Organic
Contaminants from Soil
Arthur E. Lord, Jr., Drexel University 124-136
Low Temperature Thermal Desorption for Treatment of Contaminated
Soils Phase II Results
Richard P. Lauch, U.S. Environmental Protection Agency.... 137-150
-v-
-------
SESSION A (Continued)
Page
Detection of Macro Defects in Soil-Bentonite Cutoff Walls
Andrew Bodocsi, University of Cincinnati .. 151-163
A Field Test of Hydraulic Fracturing in Glacial Till
L.C. Murdoch, University of Cincinnati.. 164-174
Computer-Based Methods of Assessing Contaminated Sites: A Case
History • ,c ,„
W.G. Harrar, University of Cincinnati 1/5-185
Results and Preliminary Economic Analysis of an APEG Treatment
System for Degrading PCBs in Soil
John A. Wentz, PEI Associates, Inc 186-200
Destruction of Cyanides in Electroplating Wastewaters Using
Wet Air Oxidation
H. Paul Warner, U.S. Environmental Protection Agency 201-208
Determining Cost Effective Approaches to the Environmental
Control of Electroplating Operations
John 0. Burckle, U.S. Environmental Protection Agency..... 209-227
-vi-
-------
SESSION B
Page
PCB Degradation: Status and Directions
P.R. Sferra, U.S. Environmental Protection Agency 228-240
The Development of Recombinant Bacteria for Polychlorinated
Biphenyl Degradation
Frank J. Monde!lo, GE Research and Development Center 241-250
Treatment of Wood Preserving Soil Contaminants by White Rot Fungus
John A. Glaser, U.S. Environmental Protection Agency 251-263
Biological Treatment of Petrochemical Sludges
Stephen D. Field, Louisiana State University 264-272
The Determination of Biodegradability and Biodegradation
Kinetics of Organic Pollutant Compounds with the Use of
Electrolytic Respirometry
Henry H. Tabak, U.S. Environmental Protection Agency ..273-296
Prediction and Modeling of "Biodegradation Kinetics of Hazardous
Waste Constituents
Rakesh Govind, University of Cincinnati....... 297-311
Preliminary Results on the Anaerobic/Aerobic Biochemical Reactor
for the Mineralization of Organic Contaminants Bound on Soil Fines
Robert C. Ahlert, Rutgers University 312-329
Fate and Effects of RCRA and CERCLA Toxics in Anaerobic Digestion
of Primary and Secondary Sludge
Richard A. Dobbs, U.S. Environmental Protection Agency 330-339
Fate and Effects of Selected RCRA and CERCLA Compounds in
Activated Sludge Systems
San joy K. Bhattacharya, University of Cincinnati 340-349
Compatibility of Flexible Membrane Liners and Municipal Solid
Waste Leachates
Henry E. Haxo, Jr., Matrecon, Inc 350-368
Geosynthetic Concerns in Landfill Liner and Collection Systems
Robert M. Koerner, Drexel University 369-378
Attenuation of Priority Pollutants Codisposed with MSW in
Simulated Landfills
Frederick G. Pohland, University of Pittsburgh 379-395
Site Demonstration of Hazcon Process
Paul R. de Percin, U.S. Environmental Protection Agency 396-408
-VT1-
-------
SESSION B (Continued)
Site Demonstration of the Terra Vac In Situ Vacuum Extraction
Technology
Peter A. Michaels, Foster Wheeler Enviresponse, Inc 409-426
The Office of Research & Development WRITE Program
Ivars J. Licis, U.S. Environmental Protection Agency 427-436
Solidification/Stabilization as a Best Demonstrated Available
Technology for Resource Conservation and Recovery Act Wastes
R. Mark Bricka, Department of the Army 437-447
Volatile Emissions from Stabilized Waste
Leo Weitzman, Acurex Corporation 448-458
Technologies Applicable for the Remediation of Superfund
Radiation Sites
Ramjee Raghavan, Foster Wheeler Enviresponse, Inc 459-469
-viii-
-------
SESSION A/B
Page
RRELevant QA Diagnostics: Major Findings in FY 88
Guy F. Simes, U.S. Environmental Protection Agency 470-479
Effect of Feed Characteristics on the Performance of EPA's
Mobile Incineration System
James P. Stumbar, Foster Wheeler Enviresponse, Inc.. 480-498
Long-Term Field Demonstration of the Linde® Oxygen Combustion
System Installed on the EPA Mobile Incinerator
Min-Da Ho, Union Carbide Industrial Gases Inc 499-514
In Place Treatment of Contaminated Soil at Superfund Sites: A
Review
M. Roulier* U.S. Environmental Protection Agency 515-525
RREL Expert Systems Project: Developing Tools for Hazardous
Waste Management
Jay E. Clements 526-534
Assessment of Chemical and Physical Methods for Decontaminating
Buildings and Debris at Superfund Sites
Michael L. Taylor, PEI Associates, Inc 535-556
-IX-
-------
SESSION C
Page
Immobilization Mechanisms in Solidification/Stabilization Using
Cement/Silicate Fixing Agents
L.G. Butler, Louisiana State University 557
Waste Reduction Evaluations at Federal Sites
James S. Bridges, U.S. Environmental Protection Agency 558
Demonstrate Computer Assisted Engineering (CAE) Techniques
for Remedial Action Assessment
P.R. Cluxton, University of Cincinnati. 559
Hydraulic Mechanisms of a Multiple Soil Layer Cover
Richard C. Warner, University of Kentucky 560
Evaluation of Solidification/Stabilization Treatability Studies
at the United States Environmental Protection Agency Center
Hill Facility
Edwin F. Barth, U.S. Environmental Protection Agency 561
The EPA Manual for Waste Minimization Opportunity Assessments
Mary Ann Curran, U.S. Environmental Protection Agency..... 562
The U.S. EPA Combustion Research Facility
Johannes W. Lee, Acurex Corporation 563
Evaluating the Cost Effectiveness of SITE Technologies
Gordon M. Evans, U.S. Environmental Protection Agency..... 564
Soliditech Site Demonstration
Walter E. Grube, Jr., U.S. Environmental Protection Agency 565-566
BioTrol Soil Washing System
Steven B. Valine, BioTrol, Inc 567
Separation of Hazardous Organics by Low Pressure Reverse Osmosis
Membranes
M.E. Williams, University of Kentucky 568-582
Testing of a Leachate Treatment System Based on a Wood
Degrading Fungus
John A. Glaser, U.S. Environmental Protection Agency 583
Assessment of KPEG Treatment for PCB Contaminated Soils
Alfred Kernel, U.S. Environmental Protection Agency 584
State-of-the-Art Field Hydraulic Conductivity Testing of
Compacted Soils
Joseph 0. Sai, K.W. Brown & Associates, Inc
585
-x-
-------
SESSION C (Continued)
Page
Review of Soil Washing Technologies for Soils Contaminated
with Heavy Metals
Carl Gutterman, Foster Wheeler Enviresponse, Inc 586
Rotary Kiln Incineration
V.A. Cundy, Louisiana State University 587-588
Biological Treatment of Chiorophenol-Contaminated Groundwater
Thomas J. Chresand, BioTrol, Inc.... 589
Computerized Management and Dissemination of Information for
Research and Development Operations at the Technical Information
Exchange (TIX) Edison, NJ
May Smith, Enviresponse, Inc 590-591
The EPA Treatability Database
Stepahanie A. Hansen, Radian Corporation 592-593
Chemical Treatment of Metals in Wastewaters and Sludges at
the T&E Facility
Douglas W. Grosse, U.S. Environmental Protection Agency 594
Biological Degradation of Chlorinated Phenoxy Acids
R.A. Haugland, University of Illinois at Chicago..... 595
Use of Foam Technology for Control of Toxic Fumes During
Excavation at Superfund Sites
Ramjee Raghavan, Foster Wheeler Enviresponse, Inc......... 596
Soil Washing—Removal of Semivolatile Organics Using Aqueous
Surfactant Solutions
Edward Coles, Foster Wheeler Enviresponse, Inc 597
Preparation for SITE Demonstration of a Powdered Activated Carbon
Treatment (PACT) Unit
John F. Martin, U.S. Environmental Protection Agency 598
-XI-
-------
-------
DEVELOPMENT AND VALIDATION OF A SURROGATE METALS MIXTURE
R. G. Barton, W. D. Clark, and W. R. Seeker
Energy and Environmental Research Corporation
Irvine, California
C. C. Lee
Risk Reduction Engineering Laboratory
U. S. Environmental Protection Agency
Cincinnati, Ohio
ABSTRACT
The U.S. Environmental Protection Agency (EPA) is developing regulations
to control the burning of metal-bearing wastes in hazardous waste
incinerators. These regulations affect the majority of the incinerators in
the U.S. since nearly every waste contains at least trace quantities of some
toxic metals. Recent research indicates that metal emissions from
incinerators are controlled by a number of parameters including metals type,
incinerator temperature and waste chlorine content. This makes it difficult
to define appropriate permit conditions which will guarantee acceptable
emissions of all toxic metals given the data obtained from current trial burn
procedures. A surrogate metals mixture was developed to provide a coherent,
defensible method for evaluating the behavior of metals in incinerators. The
mixture consists of four surrogate metals whose behavior can be used to
predict the behavior of toxic metals. Components of the surrogate metals
mixture were selected based on five criteria - volatility, toxicity,
abundance, chemical species, and cost.
A test series was planned to verify the appropriateness of using a
surrogate mixture and determine if the metals selected were acceptable. The
test program involved spiking a synthetic waste with both the surrogate
metals and toxic metals, burning the waste under a variety of operating
conditions and comparing the surrogate metals behavior with the behavior of
toxic metals.
A validation test was carried out at the EPA's Combustion Research
Facility (CRF). The results from this test were analyzed and the validity
of the surrogate metals approach was assessed.
-------
INTRODUCTION
The EPA is developing regulations to control the burning of metal
bearing wastes and has issued guidance on metals related permitting
procedures for hazardous waste incinerators (1). The guidance is based on
dispersion modeling and worst case risk assessment. Federal and State
officials are implementing the guidance. In addition, States are instituting
their own regulations restricting metals emissions. This body of regulations
is likely to affect the majority of the incinerators in the U.S. since nearly
every waste contains at least trace quantities of some toxic metals.
Recent research indicates that toxic metals emissions from incinerators
are controlled by a number of parameters, including incinerator temperature
and waste chlorine content (2). In the light of these findings, it will be
difficult to define appropriate permit conditions which will guarantee
acceptable metals emissions from waste incinerators given the data obtained
from current trial burn procedures.
CURRENT REGULATORY APPROACH
The approach that will be used to regulate metals emissions from
combustion devices burning metal-bearing wastes has not yet been fully
defined. Current guidance calls for the establishment of three different
evaluation criteria or "tiers" (1). A facility which can meet the criteria
associated with any one of the tiers will be allowed to operate. The
criteria are based on dispersion modeling and worst case risk assessment.
Under Tier III, site specific dispersion modeling is used to determine
allowable emissions based on predicted ambient concentrations. Tier II
establishes allowable emissions limits based on reasonable worst-case
dispersion modeling. Tier I sets limits on metals feed rates. The Tier I
rates are back-calculated from Tier II emission limits by assuming all of the
metals in the waste are emitted.
In general, if a facility can qualify under any one of the three tiers,
it will be permitted to burn the waste in question. It will not be allowed
to burn a waste with higher metals concentrations. The current guidance does
address the potential impact of operating conditions.
RATIONALE FOR MODIFYING CURRENT APPROACH
Tier I is adequate but very conservative. In many cases, operators will
be unable to meet the criteria set forth and will be forced to use either
Tier II or Tier III. However the procedures associated with Tiers II and III
are subject to a number of limitations which make it difficult to set
appropriate permit conditions given the data to be gathered. The principal
limitations are as follows:
t Waste sampling is often difficult and unreliable due to the non-
homogeneous nature and low metals content of most wastes.
• Behavior of many metals is not well characterized.
-------
• Emission of toxic metals may present a danger during the trial
burn. '
• Tests involving a large number of metals are costly.
PROPOSED MODIFICATION
Because of the significant disadvantages associated with monitoring all
of the toxic metals present in a hazardous waste during a trial burn, an
alternative approach was developed. This approach is centered on use of a
mixture of non-hazardous surrogate metals. The mixture consists of four
non-toxic metals which are added to wastes prior to incineration during a
trial burn. The emissions of these four metals are representative of the
emissions of all metals.
Use of a surrogate mixture has a number of advantages. The emissions
are not dangerous because non-hazardous metals are used in the mixture.
Metals whose behaviors are well defined are introduced into the waste in a
known physical and chemical form. Thus the interactions experienced by the
metals will be known and the relationship between the behavior of the metals
during the trial burn and during actual operation will be known. In
addition, use of the same metals in all trial burns will allow the
establishment of a database on metals behavior. Since only four metals need
be spiked and monitored, the costs associated with using the mixture would be
only slightly greater than those associated with running a current trial
burn.
There are disadvantages associated with the surrogate mixture approach.
The most significant is the fact that the concept has not been
experimentally verified. While in theory it is possible to use a few metals
to indicate the behavior of all metals, in practice each metal may behave
differently. Unanticipated interactions and kinetic limitations can have
different effects on different metals. A second disadvantage is the
significant amount of engineering analysis which is required to determine
permit conditions based on data obtained from trial burns in which surrogate
metals are used. A final disadvantage is that the technique neglects the
actual physical form of a metal in the waste. Metals in the mixture are
always added in the same physical and chemical form while the metals in the
waste may be present in a different form. The physical and chemical form of
the metals may have an important impact on their behavior.
SELECTION OF SURROGATE METALS
APPROACH
Components of the surrogate metal mixtures were selected based on the
following criteria:
• Metals volatility
• Toxicity
-------
• Abundance
• Chemical Species
• Cost
SELECTION OF METALS
Volatility
The particles which have the greatest potential to escape capture
contain relatively high concentrations of metals which vaporize in the
incinerator. The quantity of a metal emitted by an incinerator is directly
related to the volatility of the metal. The surrogate metals should reflect
the range of volatilities. Table 1 lists toxic metals ranked based on the
temperature at which their effective vapor pressure is 1 x 10'6 atm. The
effective vapor pressure is the combined vapor pressures of all of the metal
species that would be present in the incinerator at equilibrium.
Toxicitv
It is desirable to use nontoxic surrogate metals in trial burns. For
this analysis, a metal is classified as toxic if it is listed in 40 CFR 261,
Appendix VIII. All other metals are considered nontoxic. Table 2 lists
selected nontoxic metals ranked according to the temperature at which their
effective vapor pressure is 1 x 10'6 atm. Surrogate metals can be selected
from this list. One surrogate metal should be volatile at all reasonable
temperatures and one should be nonvolatile at all reasonable temperatures. A
metal which is volatile at low kiln temperatures and a metal which is
volatile at high kiln temperatures would also be selected.
Abundance
In order for surrogate metals to negate the effects of waste
variability, the amount added must be much greater than the quantity
originally in the waste. Thus, metals commonly found at high concentrations
in the waste should not be used as surrogate metals. If they were used, the
cost of doping would be prohibitively high and the metal and its media would
alter the characteristics of the waste. It was assumed that the surrogate
metals should be doped at a level ten times greater than the expected metal
concentration in the waste. An analysis of the impact of various mixture
media (such as water and #2 fuel oil) on the characteristics of typical
wastes and the solubility of metal compounds in the media indicates that no
more than 5000 ppm of any one metal should be added. Thus metals selected as
surrogates should be those which are not typically present in wastes at
concentrations higher than 500 ppm. However, since the cost of adding metals
increases with the amount added it is desirable to choose metals which are
present in concentrations much less than 500 ppm.
Chemical Species
One of the key assumptions associated with the selection of surrogate
metals is that all metals related reactions are very fast. Any metal
-------
Table 1. Volatility of toxic metals (waste chlorine concentration = 10 %)
Metal
Lead
Mercury
Arsenic
Osmi urn
Thai 1 i urn
Cadmium
Selenium
Silver
Antimony
Nickel
Barium
Beryl 1 i urn
Vanadium
Chromium
Temperature*, K
260
290
310
310
410
490
590
900
930
970
1180
1330
1610
1880
Temperature*, °F
5
60
90
110
280
420
610
1160
1220
1280
1660
1930
2430
2930
Principal Vapor Species
PbCl4
Hg
As203
0564
T10H
Cd
Se02
AgCl
Sb203
NiCl2
Bad 2
Be(OH)2
V02
Cr02/Cr03
Temperature at which the effective vapor pressure is 1 x 10'6 atm. The
effective vapor pressure is the combined vapor pressures of all the
species involving the metal of interest which are present at equilibrium
under given conditions.
Table 2. Volatility of nontoxic metals (waste chlorine concentration =
10 %)
Metal
Boron
Copper
Cesium
Lithium
Potassium
Sodium
Bismuth
Iron
Zinc
Calcium
Strontium
Tin
Magnesium
Silicon
Titanium
Tantalum
Thorium
Zirconium
Temperature*, K
< 250
< 400
770
780
840
870
890
890
1050
1130
1190
1320
1410
> 2000
> 2000
> 2000
> 2000
> 2000
Temperature*, °F
< -10
< 260
930
950
1050
1100
1150
1150
1430
1580
1690
1920
2070
> 3140
> 3140
> 3140
> 3140
> 3140
Principal Vapor Species
H3B306
CuoCls
CsCl2/Cs2Cl2
Li2CT2/LiCl
KC1/K2C12
NaCl/Ra2Cl2
Bi
FeCT2
Zn/ZnO
CaCl2
SrCl2
SnO
MgCl2
SiO/Si02
Ti02
Ta02
Th02
ZrO
Temperature at which the effective vapor pressure is 1 x 10"6 atm. The
effective vapor pressure is the combined vapor pressures of all the
species involving the metal of interest which are present at equilibrium
under given conditions.
-------
present in the waste will quickly react to form the species favored by
thermodynamic equilibrium. Thus the species in which each surrogate metal is
introduced into the waste is generally unimportant.
However, care must be taken not to introduce elements which could alter
the species favored by equilibrium or form other inorganic pollutants. For
example, the presence of chlorine leads to the formation of metal chlorides.
Metal chlorides are generally more volatile than other metal species. In
addition, chlorine leads to the formation HC1. Thus, if a metal were intro-
duced as a metal chloride in a waste which does not normally contain chlor-
ine, metals and HC1 emissions would be greater than they would be under
normal operating conditions.
Organometallic compounds warrant special consideration. When a metal is
present as an organometallic compound it is generally much more volatile than
when present in inorganic form. Thus if a waste contains an organometal then
the surrogate mixture should also contain an appropriate organometal.
Cost
The cost of each surrogate metal species is also an important
consideration. The costs of the chemicals vary greatly depending on the
particular species of interest. In general, inorganic salts are relatively
inexpensive ranging from $ 9.00 to $ 136.00 per pound. Organometallic
compounds are usually more expensive ranging from around $ 9.00 to over
$ 114,000.00 per pound.
POTENTIAL MEDIA
The medium used to introduce the metals into the waste must meet three
criteria:
a It must be capable of dissolving the metal species.
• It must be relatively safe and easy to handle.
t It should not alter the characteristics of the waste
significantly.
Water and dilute aqueous acids are good media for introducing a surrogate
mixture to a solid waste or an aqueous based waste. Aqueous solutions can
not be used in conjunction with organic liquid wastes which are burned iri
liquid injection incinerators. The water would interfere with the
atomization and flame patterns. In those cases an organic fuel, like number
2 fuel oil, can be used. The fuel oil would be soluble in the organic waste
and thus would not interfere with the combustion processes. Wastes which
contain organometallic compounds will also require the use of organic media
since organometallic compounds are usually insoluble in water.
RECOMMENDED SOUPS
The selection of surrogate metals based on the criteria discussed above
is summarized in Table 3. The costs shown in the tables are based on retail
prices for small lots. The price for the quantities required for a trial
burn will probably be less.
,6
-------
The nontoxic metals chosen for use in trial burns are:
Copper
Bismuth
Strontium
Magnesium
Copper is very volatile under common combustion conditions. Bismuth and
strontium exhibit intermediate volatility and are usually only encountered in
very low concentrations in wastes. Magnesium is not volatile at reasonable
incinerator conditions. All of the metals are available as soluble nitrates
which range in cost from $ 14.00 to $ 18.00 per pound.
VERIFICATION TESTS
OBJECTIVES
A series of tests are needed to verify the ability of surrogate mixtures
to represent the behavior of many different metals. A number of assumptions
are associated with the development of surrogate mixtures. These assumptions
are listed in Table 4. The verification tests must assess the validity of
these assumptions. In addition, the tests should examine the specific
mixture proposed and determine if it is appropriate.
PRELIMINARY RESULTS
A series of tests using the surrogate metals mixture were carried out at
the EPA's Combustion Research Facility (CRF). The tests involved burning a
synthetic waste spiked with both hazardous metals and the non-toxic surrogate
metals in a pilot scale rotary kiln. The impacts of chlorine concentration,
kiln temperature and afterburner temperature on metals emissions were
examined. Metals emissions data from the tests are not yet available.
However, information on the composition of the residual ash is available.
Figure 1 shows the impact of chlorine concentration and kiln temperature on
the normalized enrichment of each metal in the ash. Enrichment is defined as
the ratio of the concentration of a metal in the ash to the concentration in
the feed. Magnesium is not expected to vaporize under the conditions
encountered in the kiln. Thus, the enrichment of each metal was normalized
by the enrichment of magnesium to account for dilution effects. Normalized
enrichments close to 1.0 indicate that little vaporization has occurred while
values close to 0.0 indicate that most of the metal has vaporized.
The surrogate metals cover the range of toxic metals behavior well.
Copper and bismuth are both very volatile. Strontium exhibits intermediate
behavior, while magnesium is non-volatile. However, it must be emphasized
that these are preliminary results. Additional data will be analyzed as it
becomes available.
SUMMARY
A surrogate metals mixture can be used to aid in the analysis of metals
behavior in waste incinerators. The surrogate mixture can be used to
facilitate the following activities:
-------
Table 3. Summary of non-toxic surrogate metals selection.
Metal
Compound
Metal Volatility Cost**
Temperature , °F $/g
Abundance
of metal , ppm (3)
Inorganic: Medium - Dilute Nitric Acid
Copper
Bismuth
Strontium
Magnesium
Cu(N03)2
Bi(N03)3
Sr(N03)2
Mg(N03)2
< 260 0.03
1150 0.04
1690 0.01
2070 0.02
1.5 - 550 (4)
ND
1.08
5.0 - 590
Organometallic: Medium - Number 2 Fuel Oil
Metal
Copper
Bismuth
Strontium
Magnesium
Compound
Copper(II) 2,4 -
pentanedionate,
CuC10H14°4
Triphenyl Bismuth,
Strontium 2,4 -
pentanedionate,
Magnesium 2,4 -
pentadionate,
Mg(C5H702)2
Compound
Boiling Point, °F
ND
M.P. = 172
M.P. = 428
ND
Cost**
$/g
0.14
1.08
0.52
0.25
ND - No Data
effective vapor pressure is the combined vapor pressures of all the
species involving the metal of interest which are present at equilibrium
under given conditions.
**
1987 retail price for small lots.
could be significantly less.
The actual price for a trial burn
Table 4. Assumptions made during surrogate metals mixture development.
• The initial form of a metal is important only when extremely
volatile metal, compounds are present.
§ Dispersed metals will vaporize to a larger extent than agglomerated
metals.
• Metal interactions will have a negligible impact on behavior.
• The use of a limited number of metals is sufficient to indicate the
behavior of all toxic metals.
8
-------
LU
( .U
0.7!
t
NORMALIZED ENRICHMEI
p o
9 ro en
in et
L
rc
„ t^
^B
^m
•••
i
] T * 1600°F, 3X C1
I T » 1700°F, 3X C1
1
Wi
I
V/////////////////////A
••
I
1
1
'/////////////////////////A
mm
Y/////////////////////////A
-
As
cd
Ba Sr
Cr
Figure 1. Preliminary analysis of results from the Combustion Research
Facility tests involving surrogate metals mixture. Surrogate
metals are underlined.
-------
• Determination of the maximum allowable concentrations of
toxic metals in incinerated wastes.
• Establishment of permit conditions
t Evaluation of APCD capabilities.
e Establishment of a common base of information on the behavior
of metals in incinerators.
REFERENCES
(1) Versar Inc., "Guidance on Metals and Hydrogen Chloride Controls for
Hazardous Waste Incinerators," U.S. EPA, Office of Solid Waste, March
1988.
(2) Barton, R.G., et al., "Prediction of the Fate of Toxic Metals in Waste
Incinerators," 13th National Waste Processing Conference. Philadelphia,
May 1-4, 1988.
(3) Fennelly, P., et al./'Environmental Characterization of Waste Oil
Combustion in Small Boilers," Hazardous Waste. 1 (4), 1984, p 489.
(4) Kristensen, A.,"Operating the Rotary Kiln Incinerators at Kommunekemi,"
Hazardous Waste and Hazardous Materials. 2 (1), 1985, p 7.
10
-------
INCINERATING ETHYLENE DIBROMIDE AND DINOSEB STOCKS
by: Donald A. Oberacker
USEPA, RREL
Cincinnati, OH 45268
and
Carol Stangel
USEPA, OPP
Washington, DC 20460
ABSTRACT
Pursuant to the Federal Insecticide, Fungicide, and Rodenticide Act
(FIFRA), pesticides for which registrations have been suspended due to
imminent hazard and then finally cancelled based on findings of unreasonable
adverse effects on human health and the environment can no longer be mar-
keted and used in the U.S. for their intended purposes. In several cases to
date, after initiating such actions to terminate the use of a pesticide, the
U.S. Environmental Protection Agency (EPA) has been obligated under Section
19 of FIFRA to indemnify owners of the suspended and cancelled pesticide and
to accept products for interim storage and/or safe disposal. The 1988
amendments to FIFRA will shift the responsibility for and the cost of
disposal of suspended and cancelled pesticides to their manufacturers, in
the future. EPA will continue to be involved in the recall, storage and
disposal process but will assume an oversight role. However, at present,
EPA must fulfill its pre-FIFRA '88 obligation to complete the disposal of
ethylene dibromide (EDB), dinoseb and 2,4,5-T and silvex pesticides.
This paper summarizes EPA's progress in achieving the proper destruc-
tion of two pesticide inventories, ethylene dibromide (EDB) and dinoseb, at
commercial hazardous waste incineration facilities. Over three hundred
thousand gallons of EDB were incinerated during 1988 and in early 1989
using a special technique involving the presence of sulfur during
incineration that prevented the release of bromine to the atmosphere.
EPA is currently initiating a similar program that will result in the
incineration of two to four million gallons of dinoseb. In this case,
the major technical concern is the potential for NOx generation due to
the nitrogen content of dinoseb.
The characteristics of EDB and dinoseb, the disposal options con-
sidered for treatment and disposal of each pesticide, the results of
11
-------
pilot or field-scale testing, the incineration method ultimately
selected, and incinerator performance testing results are also addressed
in this paper.
INTRODUCTION
After EPA suspended and cancelled the registrations of the
pesticides ethylene dibromide (EDB) and dinoseb, and was required by
FIFRA to accept stocks of these pesticides for disposal, the Agency
essentially was placed in the position of "waste generator" of these
materials, which represented a large quantity of RCRA hazardous wastes.
While this was by no means the first occasion that the government or EPA
became a waste generator, the chemical characteristics of EDB and
dinoseb did present interesting challenges in terms of seeking
environmentally safe, viable, and economic disposal methods. Neither of
these pesticides had ever been commercially disposed of in any
significant quantity or by any commercial scale, carefully evaluated and
proven hazardous waste treatment methodology. Except for
unsubstantiated claims that these or similar materials had been
incinerated in small amounts by the commercial incineration or chemical
manufacturing industry, essentially no meaningful data existed to guide
EPA in disposing of the EDB and dinoseb stocks for which the Agency was
responsible.
Besides incineration, EPA carefully considered a host of other
treatment options including chemical detoxification, materials recovery,
deep-well injection, and distillation techniques. These other options
all held some potential for treating EDB and dinoseb, but each was
deemed to require extensive sub-scale development and evaluation to
demonstrate its effectiveness. All of the non-incineration options
appeared to need performance testing, scale-up, design, and permitting
activities most of which would be very time and cost intensive. Deep
well injection was rejected for both pesticides (except for weak
rinsates from EDB containers) due to uncertainties in predicting long-
term environmental effects, though it appeared to be the least costly
disposal solution. The various chemical detoxification and distillation
schemes studied for both EDB and dinoseb were ultimately ruled out for
reasons of long process equipment development, projections of
uncertainties in the permitting process, and the likelihood that any
recovered chemicals of value would not be of assured purity and
marketability.
High temperature incineration held a seemingly more immediate,
safe, and economically competitive potential provided answers were found
for EDB's bromine control issue and for dinoseb's tendency to generate
nitrogen oxides or NOx. This paper describes EPA's pilot- or field-
scale incineration tests which investigated these particular performance
issues and also generated data on destruction and removal efficiency
(ORE) and particulate emission performance on these two pesticides.
Following these tests which proved successful, EPA proceeded with
12
-------
incineration as the selected treatment/disposal technology for both EDB
and dinoseb.
EDB CHARACTERISTICS
Ethylene dibromide or EDB is a liquid halogenated hydrocarbon
which was registered as a pesticide in 1948 and suspended in 1983 and
1984. It was used largely in agriculture as a pre-plant soil fumigant
and to fumigate stored grain. Human exposure to EDB resulted primarily
from grain-based food products, and mounting evidence beginning as early
as 1975 regarding EDB's carcinogenic, mutagenic, and adverse
reproductive effects ultimately led to suspension, cancellation, and
indemnification actions under FIFRA by 1985. Between 1985 and 1988, EPA
made indemnity payments for and/or accepted for disposal a total of
approximately 329,000 gallons (3.7 million pounds) of EDB pesticides.
As with most pesticides, there existed a variety of individual EDB
products formulations with a range of concentrations of active and inert
ingredients (associated solvents or vehicles, etc.) as well as a variety
of sizes and types of containers, both pressurized and non-pressurized.
Major constituents of EDB pesticide products included ethylene dibromide
(1.5 to 50 percent or more), ethylene dichloride (up to 45 to 60%),
carbon tetrachloride (16 to 80%), carbon disulfide (0 to 16%), sulfur
dioxide (dissolved, 0 to 3%), chloropicrin (0 to 38%), and small amounts
of diesel oil, naphtha, and pentane, etc., all expressed in terms of
individual component weight percentages.
The most critical incineration characteristic of EDB was its
bromine content. The EDB molecule itself (C2H4Br2) is approximately 85%
bromine by weight. Unlike chlorine which readily combines with hydrogen
to form scrubbable HC1, past experience with brominated compounds shows
that thermal destruction will normally result in significant (and
visible) bromine gas (Br2) emissions from an incinerator stack.
DINOSEB CHARACTERISTICS
Dinoseb pesticides have been used for several decades primarily as
contact herbicides to control broadleaf weeds, but also as desiccants to
dry vegetation on food crops in the fields and facilitate harvesting of
vegetable and seed crops, etc. The active ingredient dinoseb is an
organo-nitrogen compound (2-sec-butyl-4, 6-dinitrophenol) manufactured
and formulated into over two dozen varieties of water and/or oil diluted
forms, all liquid in nature, except for the "technical" or "parent acid"
forms which are dry solids. Some of dinoseb's challenging
characteristics relative to disposal, beyond it's high nitrogen content,
are the explosive nature of the dry solids, the tendency for certain of
the water-mixed formulations to precipitate solid salts of dinoseb upon
exposure to sub-40°F ambient temperatures, and the tendency for
evaporation of alcohol or other low boiling point vehicles if handled in
open containers. However, water or oil dilution readily controls the
13
-------
explosion hazard issue for solids. In addition, many of the
formulations contain significant amounts of sodium, calcium, and inert
material, characterizing dinoseb as a "salt waste" and raising
incineration issues of refractory attack and drawing attention to
residue and particulates. No heavy or toxic metals are involved, nor do
dinoseb products (now wastes) contain chlorine.
The human health risks associated with dinoseb are of similar
concern to the Agency as those of EDB. Dinoseb was suspended by EPA
under an emergency order in October, 1986 and then cancelled in June of
1988 because of evidence indicating that it may cause birth defects,
male sterility, cataracts, damage to the immune system, and ecological
problems. Limited evidence also establishes dinoseb as a possible
carcinogen. To date, EPA has received requests for disposal of some 2.7
million gallons of liquid dinoseb and about 50,000 pounds of solid
materials, and EPA anticipates ultimately receiving requests for
disposal of a total of as many as 4 million gallons of liquids.
Returning to the nitrogen issue, dinoseb's active ingredient
molecule contains nearly 11% nitrogen.by weight. Various diluted or
formulated stocks contain from 1 to 6% nitrogen by weight, although the
average,nitrogen concentration of all stocks combined is closer to 1%.
Host of the stocks are of the water-based types.
ERA'S INCINERATION PROGRAMS
EDB
Initially, EPA considered incineration of EDB to be a low
feasibility disposal option due to the potential for bromine gas
release, as noted above. Instead, the Agency first selected the option
of chemical treatment for the detoxification of EDB's major constituents
and recovery of chemical feedstocks. Process development (including
decanning of EDB stocks) was pursued under an EPA contract in the 1985-
87 time frame. Except for the decanning activity, this effort proceeded
unsatisfactorily and the projected completion time and costs were soon
deemed intolerable due to unforeseen process equipment scale-up
problems.
In the fall of 1987, the EDB incineration option was revisited as
a direct result of an unsolicited proposal EPA received from a major
commercial hazardous waste incineration firm. Proposed was the concept
that the incineration of EDB along with adequate concentrations of
sulfur (as S, S02, etc.) in the hot zone of the combustion chamber would
encourage virtually complete chemical conversion of Br2 to hydrogen
bromide or HBr, which then should be scrubbed at high efficiency in the
incinerator's emission control system. The effectiveness of the sulfur
process, it was assured, had been functionally demonstrated on EDB
materials in one of the proposer's incinerators in the past as verified
by plume opacity observations, but no detailed stack gas verification
analyses had been performed. On the issue of how sulfur enters the
14
-------
reactions to promote HBr, EPA found that published literature existed on
older laboratory studies which measured reaction rates, but these were
conducted at temperatures much lower than those in an incinerator's
environment. The literature also offered no answer as to what chemical
reactions take place, and therafore offered no verification of the
proposer's bromine control solution.
EPA elected to evaluate the sulfur/bromine conversion process at
field-scale by conducting a detailed test burn using test quantities of
EDB, co-fired with a sulfur source in the form of 10% dilute sulfuric
acid. This test burn, conducted in December, 1987, was a complete
success and the results are described below.
During the week of December 7, 1987, EPA conducted a detailed
field-scale test burn of ethylene dibromide (EDB) at a permitted
RCRA/TSCA commercial incineration facility owned by Rollins
Environmental Services (RES), Incorporated in Deer Park (Houston),
Texas.
The three objectives for the test burn were:
1. To confirm the ability of the incinerator to achieve the
required levels of destruction and removal efficiency (ORE)
for the principal hazardous components of the EDB materials.
2. To verify the effectiveness of sulfur addition to the
combustion chamber to force the formation of hydrogen bromide
(HBr).
3. Assess the compatibility of EDB co-firing with normal waste
disposal operations at the facility.
The pesticide test burned at the Rollins site consisted of 20,000
gallons of an EDB/ethylene dichloride (EDC)/and carbon tetrachloride
(CCl^) mixture and 5,000 gallons of an EDB/chloropicrin formulation.
Scoping tests, which were brief initial test firings of the EDB
pesticides, were included as part of the test burn program to
immediately test the sulfur concept and to select pesticide flow rates
to be used in the more detailed test burn. Scoping tests showed good
performance with various combinations of the two types of formulations,
but the test burn itself involved only non-chloropicrin material. The
approximately 22,000 gallons of material remaining after completion of
the scoping and test burn program (including chloropicrin) was also
incinerated at Rollins during the several days following the test burn.
The following is a summary of the test conditions used during the
EDB test burn (1):
EDB Waste Stream Composition (by weight)
EDB 10.8 percent
EDC 44.5 percent
15
-------
CC1,
j 42.8 percent
pesticide flow rate into incinerator - 49.7-50.7 Ib/min.
(fed to rotary kiln)
net waste flow to incinerator - nominally 300 Ib/min.
including a sulfur stream as noted below, PCB and RCRA
solid and liquid waste streams
kiln temperature
afterburner temperature
stack Oo level
stack CD level
stack C02 level
stack NOX level*
stack S02 level
stack flow rates
1780 to 2000°F
2230 to 2250°F
10 percent
16-19 ppm
8-9 percent
47-63 ppm (primarily NO)
42-46 ppm
39,800 to 43,000 dscf
*During scoping tests, the chloropicrin material (38-40% EDB) was
briefly fired at up to 40 Ib/min. and NOX increased to 70-90 ppm
in the stack.
During the EDB test burn, the incinerator achieved destruction and
removal efficiencies and satisfied other regulatory standards as follows
(as determined by VOST and M-5 methods):
DRE:
EDB
EDC
ecu
Particulate Emissions:
Bromine Level in Stack:
Sulfur Feed:
in excess of 99.9999 percent
in excess of 99.99999 percent
99.999 to 99.99999 percent
0.0081 to 0.0123 grains per dscf
@ 7% C02
non-detectable (detection limit
4-5 micrograms
per dscf)
10-25 Ib/min. of a 10% sulfuric acid
solution (fired into the kiln next to
the EDB gun)
During the scoping runs, the sulfur stream was intentionally
stopped several times for brief periods (each stoppage lasted a few
seconds only). A visible brownish plume resembling typical bromine
fumes would issue forth from the stack whenever the sulfur stream was
stopped. These momentary emissions, coupled with the reliable lack of
visible or detectable bromine emissions with the sulfur present,
demonstrated the effectiveness of the reaction in which Br2 is converted
to HBr by sulfur and then scrubbed in the air pollution control device.
Mass balance calculations to account for bromine at all entrance and
exit and transient mass accumulation points associated with the
16
-------
incinerator showed that essentially all bromine was captured in the
scrubber water streams.
DINOSEB
As EPA considered disposal options for dinoseb, incineration
immediately emerged as a priority option for this RCRA P-020 type of
waste. Still, EPA felt it necessary to conduct pilot-scale tests to
evaluate and quantify the ORE, NOx and particulate generation and
control issues. A secondary option, distillation for recovery of
potentially marketable chemicals was considered but not pursued due to
the uncertain marketability of recovered materials plus the questionably
long facility design, construction, and permitting time which would be
required.
EPA conducted two pilot-scale incineration test programs on
dinoseb formulations during 1987, one at EPA's Research Triangle Park
(RTP) facility in North Carolina and one at a contractor's facility in
Tulsa, OK operated by the John Zink Company. Sampling and analyses and
reporting for both studies were conducted by an EPA contractor, Acurex
Incorporated. The results of these pilot-scale studies are summarized
below:
Tests at RTP
Objectives:
to determine ORE, NOx, and particulate emissions for the
"Dynamyte 5" formulation of dinoseb, one of formulations
with the highest concentration of dinoseb
to obtain data with and without NOx control via special
burning techniques
The Dynamyte 5 pesticide had the following characteristics:
dinoseb
diesel #2 oil
xylene
inerts
heating value (HHV)
nitrogen content
54.4%
4.04%
32.5%
9.06%
13,076 BTU/lb.
6.63% by weight
The tests at RTP utilized a low-NOx burner/package boiler
simulator device with a nominal heat release capacity of 3 million
BTU/hr. Dynamyte 5 was fired as received at approximately 16.7 gallons
per hour in all tests. The firing techniques consisted of:
Dynamyte 5 as the only fuel, with no special techniques to
reduce NOx
17
-------
Dynamyte 5 as the only fuel, with air staging to reduce NOx
Dynamyte 5 fired alone in the primary chamber, with both air
staging and natural gas-fired reburning in secondary chamber
to reduce NOx.
The results showed the following (2):
Temp, (dea. F) Temp, fdea. F)
NOx
conventional
firing
firing with
air staging
air staging
primary
chamber
1306
2450
2104
secondary
chamber
1122
1171
1205
ppm
(corrected
to 7% 02)
2998
85
88
and reburning
ORE of dinoseb
greater than 99.99% in all tests
particulate emissions .024 to.045 gr. dscf (or 54 to 101 mg/dscm)
@ 7% 02 in all tests
Tests at John Zink
Objectives:
to incinerate a mixture of dinoseb blended from all
inventories of different types of formulations
proportioned approximately according to known volumes
awaiting disposal
to confirm ORE performance and quantify particulate
emissions
to quantify NOx emissions under typical RCRA and TSCA
operating temperatures
to experience the handling, blending, and feeding
characteristics of injecting a blend of dinoseb
pesticides into a typical incinerator burner nozzle
The dinoseb mixture used for the Zink test had the following
overall characteristics as fired:
percents by weight: 82% Dyanap; 16% Dynamyte 3;
and 1.6% Dynamyte 5
18
-------
net dinoseb and dinoseb salts: 21.2% by weight
sodium ananap (Naptalam): 19.0% by weight
sodium hydroxide: 5.1%
water (primarily) and other inerts: 40.6%
xylene, alcohols, diesel oil, etc.: 13.5%
overall heating value (HHV): 4473 BTU/lb.
nitrogen content: 1% approximately, by weight
The testing results at the Zink facility resulted in the following
findings (3):
dinoseb mixture firing rates: 4 to 47 gallons per hr.
ORE of dinoseb POHC: in excess of 99.999%
NOx emissions from natural gas alone: 92-150 ppm @ 7% 02
NOx emissions when firing dinoseb: 112-836 ppm @ 7% 02 at
1750F from low to high flow rates
NOx emissions when firing dinoseb: 274-307 ppm @ 7% 09 at
2200F with low flow rate
NOx emissions when firing dinoseb utilizing patented
"Noxidizer" system: 40 ppm
particulate emissions before scrubber: .014 to .305
grains/dscf @ 7% 02
particulate emissions after scrubber: .0025 to .0079
grains/dscf @ 7% 02
Status of pesticide disposal
Based upon the field-scale EDB and the pilot-scale dinoseb
incineration work described above, the Agency has been proceeding with
disposal of both pesticides through incineration. The current status as
of early 1989 is that almost all of the EDB inventories have been
incinerated and the entire task should be completed in early spring.
Regarding dinoseb, EPA awarded incineration disposal contracts to two
commercial incineration firms, Chemical Waste Management, Inc. and
Rollins Environmental Services, Inc., at the end of 1988. Dinoseb
incineration demonstration tests are anticipated shortly at up to five
different sites. The demonstration tests will involve firing
representative quantities of dinoseb pesticide waste coupled with
observations of routine performance parameters plus NOx measurements.
19
-------
The demonstration burns, if successful, are to be followed by routine
incineration disposal. EPA estimates that at least 12 months time will
be required to treat all of the dinoseb inventories for which the Agency
has disposal responsibility.
REFERENCES
1. White, M.O., et al. RES(Tx) EDB Test Burn Program Emission Test
Results, Final Reports, Vol. I and II, Alliance Project No. 5-879-
999, Alliance Technologies Corporation, Bedford, Massachusetts,
June, 1988. Vol. I (199 pp) and Vol. II (296 pp).
2. Linak, W. Results to Date of AEERL Dinoseb Tests, unpublished. EPA
memorandum from William Linak, AEERL/CRB-Research Triangle Park,
North Carolina to E. T. Oppelt, RREL/OD-Cincinnati, April 4, 1988.
12 pp.
3. Wool, H., Villa, F. and Mason, H. Test Report for the Trial Burn of
Dinoseb in a Pilot-Scale Incinerator. Acurex Corporation, Mountain
View, California, September 20, 1988 (EPA and NTIS report numbers to
be assigned). 19 pp.
20
_
-------
PYROLYTIC THERMAL DEGRADATION OF A HAZARDOUS WASTE
INCINERABILITY SURROGATE MIXTURE
Presented at the 15th Annual EPA Research Symposium on Land Disposal,
Remedial Action, Incineration, and Treatment of Hazardous Waste
Cincinnati, Ohio
April .1.0-12, 1989
D.A. Tirey, B. Dellinger, P.M. Taylor
Environmental Sciences Group
University of Dayton Research Institute
300 College Park
Dayton, Ohio 45469-0001
and
C.C. Lee
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
26 W. Martin Luther King Dr.
Cincinnati, Ohio 45268
ABSTRACT
This report presents preliminary data from a study whose objective is to
develop a single mixture of organic compounds that can be consistently used to
demonstrate incinerator performance and compliance with EPA regulations The
proposed mixture consists of six compounds (sulfur hexafluoride, trifluoro-
chloromethane, chlorobenzene, pentachlorobenzene, acetonitrile, and tetra-
chloroethylene) which are expected to be very difficult to destroy under various
incineration failure conditions. Our initial studies have been conducted in the
absence of oxygen, thus simulating a localized waste/oxygen mixing failure mode
of an incinerator. With the notable exception of sulfur hexafluoride, agreement
between predicted and observed relative POHC stability was observed. Benzene
hydrogen cyanide, benzonitrile, and naphthalene have been identified as possible
surrogates for PIC formation.
21
-------
INTRODUCTION
Due to the public concern over possible emissions of toxic organic
compounds, there has been considerable debate concerning how to best
determine the performance of a hazardous waste incinerator. The current
regulatory approach requires demonstration that a facility can destroy or remove
principal organic hazardous constituents (POHCs) in the waste feed with a
destruction and removal efficiency (ORE) of 99.99% (or 99.9999% for
polychlorinated biphenyls and polychlorinated dibenzo-p-dioxins). Since it is
impractical to test the performance of an incinerator for every possible organic
component of the waste, considerable effort has been expended to develop an
"incinerability11 index. Several such indices have been developed and applied to
predicting the relative incinerability of toxic organics with varying degrees of
success.1
Unfortunately there is only limited data to confirm that any of these indices
are working successfully, although currently available comparisons indicate that a
thermal stability-based approach appears to have considerable promise.1'2
Furthermore, these indices have not always been uniformly applied in the
permitting process, at times causing delays in the facility being certified for
operation. In addition, any time there is a special type of waste being burned, there
is concern that some toxic component (which may not be on EPA's Appendix VIII
list) is more difficult to incinerate than the POHCs selected for ORE performance
testing during the trial burn.
As a result, there has been considerable interest in developing a
standardized ORE surrogate mixture that could be used to determine the
performance of any incinerator.3 Each of the components of this mixture would test
the capabilities of the incinerator by testing the incinerator's failure modes. A failure
mode of an incinerator may be defined as an operating condition where a fraction
of the material experiences a departure from the spatially and temporally averaged
optimum operating conditions that results in an increase in toxic emissions. In
general, modes of failure can be thermal (e.g., quenching, temperature gradients,
heat transfer limitations), temporal (e.g., uncertainties in residence time
distributions due to plug flow assumptions), or fuel-oxidant mixing (e.g., poor micro-
mixing resulting in "pyrolysis pockets") related. The fuel-oxidant mixing failure has
been suggested as the most important, although all failure modes likely contribute
to increased emissions of undestroyed POHCs as well as products of incomplete
combustion (PICs),2-5
One type of temporal or residence time failure is the possibility of slow
vaporization of condensed phase materials. However, it has been shown that this
phenomenon would have minimal impact on the relative incinerability of POHCs.6
What would be more crucial in this case are the thermal failure modes that may
exist by virtue of quenching of hot gases in the post flame zone by cool secondary
air It is conceivable that sufficiently low temperatures may be reached such that
oxidation reactions involving POHCs are slow enough to result in measurable
emissions. This has never been conclusively demonstrated in full-scale
22
-------
Acetonitrile was selected as an example of a non-chlorinated compound that
has been shown to be stable under both oxidative and pyrolytic conditions.2'7 This
compound was also chosen as a model compound of the generally very stable
group of nitrile-containing compounds.2 Chlorobenzene was selected as an
example of a chlorinated compound that has been shown to be very stable under
pyrolytic conditions,2 thus representing a mixing failure test of an incinerator.
Although there was some concern over PIC formation distorting its apparent ORE,
this compound was included in the mixture at a feed concentration thought to
minimize this possibility. Pentachlorobenzene is a good example of a chlorinated
compound which should exhibit high thermal stability under oxidative conditions.7
It was thus included in the surrogate mixture to test the thermal failure mode of an
incinerator. Tetrachloroethylene was selected as an example of a perchlorinated
compound which is stable under pyrolytic conditions and moderately stable under
oxidative conditions.2-7 Trifluorochloromethane and sulfur hexafluoride were
included to exclusively test the thermal quenching failure mode of an incinerator.
These compounds should represent the most stable components of the mixture
with destruction occurring above 1000°C.8'9 Toluene was included as a suitable
solvent material of moderate thermal stability for this complex organic mixture.
EXPERIMENTAL APPROACH
Experiments were conducted using the Thermal Decomposition Analytical
System (TDAS). The TDAS is a closed, in-line thermal instrumentation system
capable of accepting a solid, liquid, or gas-phase sample, exposing this sample to
a highly controlled thermal environment and then performing an analysis of the
effluents resulting from this exposure. Its design and operation have been
described in previous reports and publications.10'11
Due to the low vapor pressure of pentachlorobenzene under ambient
conditions, a liquid phase organic mixture was considered the most viable
approach to produce a relatively stable, homogeneous sample within the target
concentrations for each component. The sample was thus prepared by dissolving a
specific quantity of solid into the liquid phase and purging the liquid with the two
gases for 30 minutes. Samples were ultimately prepared with little or no
headspace, a prerequisite for keeping the integrity of the gases dissolved in the
liquid intact. Experiments performed before beginning data acquisition revealed
that samples prepared in this manner and stored at room temperature could be
used as long as 48 hours without appreciable losses of sulfur hexafluoride or
trifluorochloromethane detected. High purity (> 98%) samples of each component
were used without additional purification in preparing the mixture.
The gas-phase mixture concentration in the tubular fused silica flow reactor
was maintained at a concentration of ~1 x 10'3 moles/liter using a manual liquid
injection technique. For each experiment, a volume of 0.2 \i\ was injected into the
system at a rate of ~0.05 (xl/s using a small sub-microliter syringe. All components
23
-------
Incinerators, but it seems wise to address this possibility in the development of a
surrogate mixture.
Sensitivity analyses indicate that relative POHC thermal stability is most
sensitive to oxygen concentration.2 This leads us to believe that an appropriate
surrogate mixture should include components of high thermal stability for both
oxidative and pyrolytic conditions. Although generally less sensitive, the relative
thermal stability of POHCs also varies as a function of temperature. Thus, the
surrogate mixture should also contain components of high relative thermal stability
when based on both low temperature (~700°C) and high temperature (~100u°C)
destruction efficiency measurements. From a kinetic viewpoint, this means we
should select mixture components whose primary mechanism of decomposition
covers all known reaction classes and which are examples of slow reactions within
each class.
Table 1 presents the components of a low chlorine content surrogate mixture
that were selected based upon the above criteria and which were tested in this
study.3
Table 1
Proposed Incinerability Surrogate Mixture
Component
Acetonitrile
Chlorobenzene
T99 (2)a
1000b
990b
Destruction
Mechanism
C-H bond rup.,
H abstraction
C-CI bond rup.,
Conc.d
(wt. %)
11.4
19.3
Failure Mode
Represented
Mixing and
Thermal
Quenching
Mixing
Pentachlorobenzene 940b
Sulfur Hexafluoride 1090°
Tetrachloroethylene 895b
Trifluorochloromethane 1010°
Toluene (solvent) 890b
Cl displacement
G-CI bond rup.,
Cl displacement
S-F bond rup.
Cl displacement
C-CI bond rup.
3.5
0.1
13.2
1.0
51.5
Mixing and
Thermal
Quenching
Thermal
Quenching
Mixing and
Thermal
Quenching
Thermal
Quenching
Mixing
C-H bond rup.,
H abstraction
Footnotes:
a. For mixtures of compounds with elemental composition of
CaHsCli and waste/oxygen equivalence ratio of 3.0.
b. Based on experimental data.
c. Estimated using unimolecular reaction rate theory.
d. Relative concentrations based on waste mixture with an elemental
composition of C3H3 2Cl0.3N0i15F0 015S0.oo3-
24
-------
of the sample were simultaneously vaporized in the insertion region which was
held isothermally at 275°C for all experiments.
Precise two-second thermal exposures were conducted in helium carrier
(99.99+% purity) over a temperature range of 300-1000°C. Duplicate experiments
were conducted along the temperature profiles to determine data reproducibility.
Blank runs were conducted daily to determine detection limits.
The effluents resulting from thermal exposure were cryogenically focussed
at the head of a 30 m x 0.25 mm ID fused silica capillary column (0.25|i DB-5
stationary phase) held at -60°C. The GC oven temperature was programmed to
290°C at 10°C/min. Helium was used as the GC carrier gas and a mass selective
detector in the full scan mode (35 to 400 m/e) was used for effluent quantitation.
The lightest of these effluents (35 ^ m/e <, 132) were not amenable to this technique
and were analyzed using "on the fly" mass spectrometry. Mass spectral response
curves were generated for each POHC. Calibration of mass spectral response for
the myriad of PICs generated has not yet been completed. Tentative PIC
identifications and their semi-quantitative formation and destruction curves were
determined by mass spectral data analysis.
RESULTS
The individual thermal decomposition profiles of the six POHCs are
presented in Figure 1. The thermal stability index (Tgg (2)) for each of the POHCs
and its theoretically predicted versus observed relative stability are displayed in
Table 2. Except for the unexpected low stability of sulfur hexafluoride, the predicted
and observed relative stabilities are in agreement. The differences in absolute
thermal stabilities in Tables 1 and 2 reflect the differences in test conditions, i.e., 0 •
3.0 versus oxygen-free and H/CI ratios of 3:1 versus 10:1, respectively.
As illustrated in Figure 1, for greater than 99% destruction, the relative
stability of tetrachloroethylene and pentachlorobenzene reversed. This rare
thermal decomposition behavior was not observed for the other components in the
mixture. Despite the fact that the decomposition profiles of these two compounds
crossed at approximately 875°C, their thermal decomposition behavior was the
most reproducible (±5%) of all components in the mixture. The thermal behavior of
the light gases was less reproducible (±20%), due to the lack of controlled,
cryogenically trapping during the data acquisition procedure. Similarly, acetonitrile
was somewhat difficult to quantitate with the GC column employed because of the
polarity of this compound. The extreme tailing behavior of this compound made
accurate quantitation of the peak area difficult. Because of this, the detection limit
for this compound was necessarily high and a fairly high level of uncertainty (±10-
20%) in data was observed.
25
-------
10
Sulfur Hexafluoride
Trlfluorochloromethane
Acetonitrile
Tetrachloroethylene
Chlorobenzene
Pentachlorobenzene
500
600 700 800
TEMPERATURE (°C)
900
1000
Figure 1. Thermal decomposition behavior of proposed surrogate mixture in
flowing nitrogen. Mean, gas-phase residence time = 2.0 s.
TABLE 2
Experimental versus Predicted POHC Thermal Stability
Compound
Predicted
Experimental
Exp. T99 (2)
Sulfur Hexafluoride
Trifluorochloromethane
Chlorobenzene
Acetonitrile
Pentachlorobenzene
Tetrachloroethylene
1
2
3
4
5
6
3
1
2
4
5
6
-920
1010
975
-910
870
860
Table 3 lists the tentative PIC identifications and their relative semi-
quantitative concentrations for temperatures greater than 700°C. Inspection of this
table shows that numerous complex PICs were produced from the mixture of seven
simple starting materials. Several general observations were evident. First, the PIC
with the greatest frequency of occurrence and concentration was benzene.
Second, the trend in product formation was toward unsaturated, conjugated
compounds which were thermodynamically favored because they are resonance
stabilized, e.g., polynuclear aromatic hydrocarbons, (PNAs). Almost exclusively,
these products were the only PICs detected at high reactor temperatures other than
stable lighter gases such as hydrogen chloride and methane. This trend toward
unsaturated, conjugated products with increasing reactor temperature was also
one observed in previous studies. The relative paucity of halogenated compounds,
e.g., benzyl chloride, trichlorobenzene, tetrachlorobenzene was consistent with the
large H/CI ratio of this mixture. 100+20% of the chlorine was accounted for as
26
-------
hydrogen chloride at 1000°C. The remainder of the observed PICs can be
generally characterized as substituted benzenes. The substituent groups included
methyl, ethyl, ethenyl, ethynyl, propynyl, phenyl, chlorophenyl, and nitrite.
TABLE 3
Tentative PIC Identifications
and Relative Semi-Quantitative Concentrations
Tentative
Identification 700
Hydrogen Chloride XX
Methane
Hydrogen Cyanide
Benzenepropanenitrile
Benzene XX
Ethylbenzene
Ethenylbenzene
Ethynylbenzene
Isocyanobenzene or Benzonitrile
Propynylbenzene
Methylbenzonitrile
Benzyl Chloride X
Trichlorobenzene
Naphthalene or Azulene
Quinoline or Isoquinoline
Tetrachlorobenzene
1-Pheny|naphthalene X
1,1'-Biphenyl or
Ethenylnaphthalene
Methyl-1,1'-biphenyl isomers
1 , 1 '- ( 1 , 2-Ethanediyl) bis-benzene
Acenaphthalene br Biphenylene
Naphthalenecarbonitrile
Chlorobiphenyl
9H-Fluorene
Chloropentafluorobenzene
Phenanthrene or
9-Methylene-9H-fluorene
2-Phenylnaphthalene
1 -Phenyimethylene-1 H-lndene
Pyrene or Fluoranthene
Ci3HsCl4 isomers
1 1 H-Benzofluorene (a or b isomers)
Triphenylene or Chrysene or Naphthacene
or Benzophenanthrene
Benzofluoranthene or Benzopyrene
or Benzacephenanthrylene
750
XX
XX
XXX
XX
XX
XX
X
XX
X
XX
XX
XX
XX
X
XX
XX
XX
XX
Temperature (*C)
800 850 900
XX
X
XX
XXX
XX
XX
XX
XXX
XX
XX
XX
XX
XX
XX
XX
X
XXX
XXX
XX
XX
XX
XX
XX
XX
XX
X
XX
XXX
XX
XX
XX
XX
XXX
X
XX
XXX
XX
XX
XX
X
XXX
X
XXX
XXX
XX
XXX
XX
XX
XX
XX
NA
NA
xxxx
XX
XXX
XX
XX
XXX
X
-XXX
XX
XX
X
XX
XXX
XX
XXX
XX
XX
950
XXX
NA
NA
xxxx
XX
XXX
XX
XXX
XXX
XX
XX
XX
XX
XXX
XX
XXX
XX
XX
1000
XXX
NA
NA
xxxx
XX
XXX
XXX
XXX
XX
X
,
XXX
XXX
XX
XX
Legend:
NA: GC/MS data was not obtained at these temperatures.
Concentration based on GC/MSD integrated response:
0.1% <. X < 0.25% of total mass injected.
0.25% <; XX < 2.5% of total mass injected.
2.5% ^ XXX < 25.0% of total mass injected.
XXXX ^ 25.0% of total mass injected.
27
-------
DISCUSSION
Perhaps the most surprising result of this study was the apparent instability
of sulfur hexafluoride, being undetectable at a temperature of 900°C. This result
contradicts previous studies where this compound was not destroyed (at the 99%
level) below temperatures of ~1100°C.8'9-12 Experiments were performed to
determine whether its apparent instability was possibly due to reactions in the ion
source of the MSD. Pure sulfur hexafluoride was exposed to 300°C (a non-
degradative temperature) and 900°C (the temperature at which sulfur hexafluoride
was below detection limits in the mixture data) and no decomposition was
observed. Sulfur hexafluoride in the presence of helium was similarly tested with
no observed reaction. The experiments were repeated with successive additions of
carbon dioxide, hydrogen chloride, water vapor, and methane without any
observed decomposition.
The only fluorine-containing product detected was a very small peak
tentatively identified as pentafluorochlorobenzene. Two other products observed
as PICs in a previous study of sulfur hexafluoride mixture decomposition,8 sulfur
tetrafluoride and hydrogen fluoride, were not detected. However, the concentration
of these compounds may have been below analytical detection limits due of the
low sulfur hexafluoride concentration in the feed (see Table 1). Because of the lack
of mass balances for fluorine, we are continuing to assess this anomalous result.
We plan to study simple binary sulfur hexafluoride/organic mixtures to determine if
bimolecular thermal decomposition reactions in the flow reactor can account for the
observed profile. We also plan to verify our results using different detectors
including an electron capture and flame photometric detector.
The relative stability of the remaining POHCs accurately reflect their
predicted mechanism of destruction. Trifluorochloromethane decomposes by
unimolecular C-CI bond rupture, which is a relatively slow reaction due to its high
bond dissociation energy (A = 3.16 x 1014 s'1, Ea = 85 kcal/mole).8 Under pyrolytic
conditions, chlorobenzene likely decomposes by Cl atom displacement by H,13 the
latter being the dominant reactive species in the system as illustrated in Table 4.
The observation that pentachlorobenzene was more fragile is due, in part, to the
larger number of chlorines available for displacement. The rate of the displacement
reaction at 1085 K has been measured as ~4.0 x 108 liter/mole-s per Cl atom.13
Table 4
Reactive Species Equilibrium Mole Fraction
at Elevated Temperatures14
Specie
Cl atoms
H atoms
700°C
6.99E-1 1
3.59E-10
800°C
8.68E-10
4.65E-09
900°C
7.04E-09
3.91 E-08
1000°C
4.13E-08
2.37E-07
All bonds are quite strong in acetonitrile and the most likely pathways of
destruction were methyl radical displacement by H atoms forming hydrogen
cyanide and H abstraction by methyl radicals yielding methane.15 The stability
reported here compares favorably with that reported in the literature when one
28
-
-------
accounts for the initiation of radical chain reactions by other more fragife
compounds which generate reactive H and Cl atoms.15 Tetrachloroethylene has
recently been studied as a pure compound in this laboratory.16 The most likely
mechanism of destruction in this mixture was Cl displacement by H atoms,
although trichloroethylene was not observed as a PIC. The lower C-CI bond (85
kcal/mole versus 95 kcal/mole for chlorobenzene) facilitates Cl displacement,
resulting in its being the most fragile compound in the mixture. There was also
evidence that toluene reformed at elevated temperatures. Initial decomposition by
C-H bond rupture of the side chain results in the formation of stable benzyl radicals.
Their recombination with high concentrations of H atoms may be responsible for
the observed reformation.
Figure 2 presents a plot of integrated response for all detectable organic
PICs and POHCs as a function of temperature. This graph indicates that PIC
emissions are greater than POHC emissions at temperatures above 800°C. It also
displays how the total PIC response correlates with benzene formation and
destruction, suggesting that benzene may be an appropriate surrogate for organic
PIC emissions in full-scale incinerators.
HI
500
600 700 800
TEMPERATURE (°C)
900
1000
Figure 2. Comparison of POHC decay curve versus semi-quantitative PIC (and
benzene) formation and destruction curves.
At 700°C, the major organic PICs are benzene and benzenepropanitrile.
The former was probably produced by electrophilic displacement of the toluene
methyl substituent by H atoms, the latter by recombination of benzyl and
ethanenitrile radicals. Each of these two species was probably formed by C-H bond
homolysis in the initial decomposition of the parent molecules. The ethanenitrile
29
-------
species was apparently also present at higher temperatures as evidenced by the
detection of naphthalenecarbonitrile. At temperatures above 700°C, the presence
of nitrile radicals was evidenced by the formation of benzonitrile and
methylbenzonitrile. Benzyl chloride was a relatively low-yield recombination
product as a result of the low chlorine concentration in this mixture.
As the temperature was raised to 750°C, more substituted benzenes began
to dominate. One potential route to the formation of ethyl-, ethenyl-, and propynyl-
benzene likely involved displacement of the methyl substituent from toluene. It is
Interesting to note the formation of ethynylbenzene at temperatures greater than
900°C. This was likely due to the higher temperature required for ethynyl radical
formation from the decomposition of a suitable aromatic precursor. At temperatures
of 800-900°G, aromatic substitution products, e.g., biphenyl and chlorobiphenyl,
were produced from toluene displacement reactions involving phenyl and
chlorophenyl radical (produced from toluene and chlorobenzene, respectively).
Tetrachlorobenzene was formed as a dechlorinated product of penta-
chlorobenzene at 800-850°C and this molecule, in turn, dechlorinated to
trichlorobenzene at higher temperatures.
The remaining products at the highest temperatures were benzene and
numerous PNAs. A mechanism suggested in the literature for the formation of
PNAs involves vinyl radical chain displacement reactions followed by cyclization.17
At higher temperatures, increasingly unsaturated PNAs were formed, which may
ultimately lead to soot formation. Besides benzene, the most notable PIC was toxic
hydrogen cyanide, produced from methyl radical displacement by H atoms. This
compound appeared at 850°C and persisted at higher temperatures.
SUMMARY
We plan to continue our study of this mixture under other failure mode
conditions. It is encouraging that the predicted and observed POHC stability are in
agreement under pyrolytic conditions, with the notable exception of sulfur-
hexafluoride. Benzene, hydrogen cyanide, benzonitrile, and naphthalene have
been identified as appropriate PIC surrogates. However, the relative absence of
chlorinated PICs suggests that the mixture elemental composition should be
modified to increase chlorinated PIC formation. These compounds are desirable to
provide surrogates for testing incinerators burning highly chlorinated hazardous
waste.
ACKNOWLEDGMENTS
We gratefully acknowledge M. Tissandier for assisting in data acquisition
and J. Kasner for performing the equilibrium calculations. This research was
partially supported by the US-EPA under cooperative agreement CR-813938.
30
-------
REFERENCES
1. Dellinger, B., Graham, M. and Tirey, D. Hazard. Waste Hazard. Mater., 3,
293 (1986).
2. Taylor, P. H., Dellinger, B. and Lee, C.C. Environ. Sci. Technol., in review,
1989.
3. Dellinger, B., Taylor, P.H. and Lee, C.C. "Development of Hazardous Waste
Incinerability Surrogate Mixtures," Presented at the 2nd Annual National
Symposium on Incineration of Industrial Wastes, San Diego, CA, 1988.
4. Tsang, W. "Fundamental Aspects of Key Issues in Hazardous Waste
Incineration," ASME Publication 86-WA/HT-27, 1986.
5. Lee, K.C.JAPCA.SS, 1542(1988).
6. Chang, D., et al. "Relationships Between Laboratory and Pilot-Scale
Combustion of Some Chlorinated Hydrocarbons," Proceedings of AlChE
Summer National Meeting, Denver, CO, 1988.
7. Dellinger, B., Torres, J., Rubey, W.A., Hall, D.L., Graham, J.L., and Carnes, R.A.
Hazard. Waste Hazard. Mater., 1, 137 (1984)
8. Taylor, P.H. and Chadbourne, J.F. JAPCA, 2Z, 729 (1987).
9. Tsang, W. and Shaub, W. in Proceedings of the 2nd Conference on
Management of Municipal, Hazardous and Coal Wastes, p.241, 1983.
1 0. Rubey, W.A. "Design Considerations for a Thermal Decomposition Analytical
System," EPA-600/2-80-098, U.S. Environmental Protection Agency,
Cincinnati, OH, 1980.
11. Rubey, W.A. and Carnes, R.A. Rev. Sci. Instrum. 56. 1795 (1985).
12. England, W., et al. in Proceedings of the 79th APCA Annual Meeting,
paper 86-61.1, 1986.
13. Tsang, W. "High-Temperature Chemical and Thermal Stability of Chlorinated
Benzenes," Presented at the International Flame Research Committee
Symposium on the Incineration of Hazardous , Municipal, and Other Wastes,
Palm Springs, CA, 1987.
14. Reynolds, W.C. "STANJAN Equilibrium Program, Version 3.0," Department of
Mechanical Engineering, Stanford University, Stanford, CA, 1986.
15. Lifshitz, A., Moran, A., and Bidani, S. Int. J. Chem. Kin., 1£, 61 (1987).
16. Taylor, P.H. and Dejlinger, B. "Development of a Thermal Stability Based Index
of Hazardous Waste Incinerability," Fiscal Year 1988 Report prepared for EPA
Cooperative Agreement CR-813938, C.C. Lee, Project Officer, December
1988.
17. Cole, J.A., et al. Combust. Flame, 56, 51 (1984).
31
-------
INCINERABILITY RANKING OF HAZARDOUS ORGANIC COMPOUNDS
by: Robert E. Mournighan
Marta K. Richards
Howard 0. Wall
Technology Research Section
Thermal Destruction Branch
Waste Minimization Destruction and
Disposal Research Division
U.S. Evironmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Since EPA became involved with the regulation and permitting of hazardous
waste incinerators, developing a reliable measure of incinerator performance has
been one of its goals. The use of a thermal stability index to rank Principal
Organic Hazardous Constituents (POHCs) has facilitated the evaluation of
incinerators. The development of data to support the thermal stability rankings
has been an ongoing effort.
This paper discusses the results of a parametric study in which temperature
and oxygen concentrations were evaluated relative to destruction and removal
efficiency (ORE) of the test mixture. The compounds used in the test mixture
were: toluene, chlorobenzene, tetrachloroethylene, tetrachlorobenzene and sulfur
hexafluoride.
32
-------
INTRODUCTION
Since the regulation of hazardous waste by the USEPA, both the Agency and
the regulated community have been searching for means of evaluating incinerator
performance. The concept of a "trial burn," a series of tests designed to
evaluate the ability of an incinerator to process waste to certain standards,
was developed; and surrogates were introduced (1) for hazardous and solid waste
and for Principal Organic Hazardous Constituents, or POHCs. In essence, the
incinerator's ability to destroy waste was based on its performance in destroying
the POHCs.
Destruction and Removal Efficiency (DRE) is used by EPA to express
incinerator performance. The DRE of a POHC is calculated as follows:
DRE = (Wln_-Wput) x 100
u-
win
(2)
Where in and out are the mass flow rates of the POHC
input and output (at the stack) respectively.
As regulations and policies regarding trial burns developed, it became
apparent to all concerned that DRE determinations were rapidly becoming extremely
expensive and time consuming. It is not unusual to see trial burn costs
approaching 2% of the capital costs of the facility. Largely responsible for
these costs was the requirement that several types of POHCs be evaluated,
resulting in the use of multiple sampling trains. The process of obtaining
representative samples could take a week, or longer. Sample processing and
analysis, also expensive, would drag out the time-span between trial burn and
results to over three months. The objective for regulators and regulated alike
was to search for a less expensive approach. The approach was similar to that
taken for POHCs and solid waste: find a surrogate for the waste being burned
and develop a cheap, quick analytical method for the analysis of the surrogate.
Several compounds, e.g., Sulfur hexafluoride (SFg), Freons^, and fluorocarbons,
were put forth as candidates.(3)
Because of its high thermal and chemical stability and the fact that it
is inexpensive and non-toxic, SFg received much attention as a surrogate. Quick,
relatively inexpensive, and reliable methods were developed for its analysis,
and initial evaluation began in 1984.(4) A system employing a gas chromatograph,
equipped with an electron capture detector (6C/ECD), was used to obtain SFg
concentrations in stack gas every 2 to 4 minutes, a much shorter time than the
sampling time required for sampling trains. Equipment set-up, operation, and
calculation of results could be done in 1-1/2 days (4), a far cry from three
months.
Since that time, a great deal of effort, time and money has been spent,
both in the United States and Canada (5), to exhaustively evaluate SFg as a POHC
or waste surrogate.
During the same time period, the concept of using a standard POHC mixture
(a "POHC Soup") for trial burns, by employing a group of compounds with a range
of chemical and physical characteristics, was developed. A paper describing the
33
-------
rationale and results of that effort is being given at this conference.(6)
Both of these efforts have been researched extensively at the laboratory
stage (5, 6, 7). An evaluation program at pilot scale for the USEPA Combustion
Research Facility (CRF) was proposed in 1987. The CRF has two pilot-scale
(3 x 10° Btu/hr) incinerators, a liquid injection system and a rotary kiln.
The SFg/"POHC Soup" program was performed in the liquid injection system. This
paper describes the ORE results and discusses the implication of this work and
of other researchers.
Interest in assuring and demonstrating compliance with RCRA permit
conditions has also concerned the hazardous waste community. Trial burns are
prohibitively expensive to carry out on a routine or even annual basis, with
not much gained in the process. This approach still leaves "gaps in coverage,"
where incinerator performance may be unknown. It became apparent, after initial
research at EPA, that residence time, temperature and turbulence were not the
only factors to consider in the evolution of incineration criteria. The Toxic
Substance Control regulations (8) for PCB incineration reflected that
realization, and stipulated not only time and temperature, but also a minimum
stack Op concentration as well as a minimum combustion efficiency of 99.9% are
required.
This paper concentrates on the evaluations of SF6 and the "POHC Soup" as
trial burn surrogates.
EXPERIMENTAL BACKGROUND
The experimental program was executed at the USEPA Combustion Research
Facility's Liquid Injection System (LIS). Figure 1 is a sketch of the unit,
showing each of the elements that make up the incinerator and the air pollution
control system.
The LIS is fired with liquid waste, and propane is used to maintain
temperature control. Combustion air is supplied by forced-draft fan. The waste
mixture containing the "POHC Soup" was stored in a stainless steel vessel and
pumped continuously to the LIS.(9) SFg was injected into the liquid feed stream
as a gas, where it dissolved into the liquid phase just prior to incineration.
Sampling of the stack gas for both SFg and the POHCs was done at the exit
of the air pollution control systems, just after the ionizing wet scrubber (IWS).
Even though the gas passes through additional gas cleanup, sampling further
downstream was not conducted. The final gas cleanup is unique to this facility
and is not standard for commercial and private incinerators. Table 1 shows the
experimental conditions and the ORE results for the SFg and the POHC components.
Oxygen and incinerator temperature were varied as part of a designed
experiment. All other variables, such as waste composition and flow rates, were
kept constant.
The compounds used in the test mixture were toluene (70%), chlorobenzene
(10%), pentachlorobenzene (10%) and tetrachloroethylene (10%). This combination
of compounds resulted in a mixture which contained both volatile components and
34
-------
•4-J
I/)
U
0)
•o
•r-
3
O)
O)
35
-------
semivolatiles. Stack sampling of these components was accomplished with the
VOST (USEPA Method 0030) and Modified Method 5 (USEPA Method 0010). The SF6
sampling method was proportional gas sampling, with the sample being fed directly
to 6C/ECD.
Table 1. Data Table - Temperature, Oxygen and
DREs (number of nines)
Afterburner
Exit
Tempera- Exit
Experiment ture, °C Oxygen %
ORE
Tetra-
chloro- Chloro- Pentachloro-
SFg* ethylene* Toluene* benzene* benzene*
1
2
3
4
5
6
7
8
9
10
*Calcu
1030
1114
943
1274
1091
945
1310
1077
1175
1105
lated ORE by
2.3
3.3
5.6
1.3
5.1
8.2
4.5
9.2
8.0
4.7
the formu
4.35
3.59
3.33
7.00
5.31
4.17
5.33
3.44
5.70
3.96
la ORE
5.40
5.52
6.14
5.51
5.74
6.66
5.34
5.85
6.70
5.16
= -Log
6.19
6.21
7.10
6.11
5.85
6.68
5.59
5.74
7.55
6.39
1-DRE
100
6.11
6.04
6.38
6.40
4.14
6.66
5.80
6.28
6.54
5.68
7.80
7.52
7.34 '
7.30
6.98
7.41 •
7.49
7.35
7.43
7.44
ANALYSIS OF RESULTS - SFg
To gain as much information about the behavior of the POHC/SFg DREs with
respect to the two variables, a regression analysis was performed for each data
set, one for each compound. Table 2 lists the results of the regression
analysis, the model used and the correlation coefficient.
Figure 2 is a contour plot with independent variables, temperature and
afterburner exit oxygen concentration, on the Y and X axis respectively. The
SFg ORE (expressed in the number of nines) is displayed as contours of the
regression surface. This figure demonstrates why it is so hard to make solid
judgments about the relationship between ORE and a single variable. As the
figure clearly illustrates, whether ORE goes up or down with increasing Op
depends on the operating temperature. At 950°C, ORE increases from about
3-nines to nearly 4-nines as oxygen increases. At 1350°C, the same change
results in the opposite effect, dropping the ORE from greater than 7-nines to
less than 5-nines ORE as oxygen increases.
At this point, it is stressed that the regression analysis and the
resulting figures developed from it should only be used in general, not
36
-------
SF6 ORE VS T, O2
1400
T
^1300H
P
E
R 1200
A
T
U 1100-
R
E
1000
900
01
x
+
i 1 1 1 1 r
23456
i r
7 8 9 10
O2 CONCENTRATION
3-9's
6-9's
ORE
+ 4-9's
x 7-9's
*• 5-9's
Figure 2. SF6 ORE vs. T, 02.
37
-------
predicting absolute ORE figures for specific compounds. Until additional data
are developed, these relationships should only be applied to the CRF incinerator
and can only be used to describe the relationship between dependent (ORE) and
independent (Temp.- and" Op) variables within the range of the' experiment.
Extrapolation of the results outside the range is risky.
Compound
Table 2. Data Table
Model
SFg 1st Order
Toluene 1st Order
Tetrachloroethylene 1st Order
Chlorobenzene 2nd Order
Pentachlorobenzene 2nd Order
Correlation
Coefficient
0.77
0.45
0.75
0.76
0.68
In the laboratory work done on the thermal decomposition of SFg at the
University of Dayton Research Institute (UDRI) (7), it was determined that SFg
destruction was independent of oxygen concentration, indicating the mechanism
to be unimolecular decomposition. Although Figure 2 shows an effect of 02 on
SFg ORE, it may not mean that Q£ specifically causes the changes. What is not
depicted is that as the 02 concentration increases, air to fuel ratios and
incinerator residence times change concurrently. The variation in SF6 ORE may
be related to those effects and is not necessarily inconsistent with the UDRI
results.
Figure 3 is a plot of the SFg DREs versus temperature with 3 data sets
illustrated. The data shown as diamonds were the SFg DRE-temperature
relationship calculated from the data supplied in References o and 7. This was
calculated for a 2-second residence time and represented the DRE temperature
relationship for SFg exclusive of any other physical or chemical effects.
The data plotted as Xs 'were data from the work conducted at the Alberta
Environmental Centre (5) in which the effect of the presence of refractory on
SFg DRE was evaluated. As one can see, the presence of refractory in the reactor
had a marked effect on the DRE-temperature relationship, reducing the required
temperature by some 200°C.
The third data set plotted in Figure 3 was the CRF data, and shows the
regression line for those data points. Here, as with the Canadian data, there
is a marked difference in SFg DRE-temperature behavior. While some of the
difference could be attributed to the presence of refractory and residence time
changes in the incinerator, the authors feel that there is more to it than that.
Since the laboratory data (UDRI) were taken with non-flame systems, in which the
substances underwent an extremely narrow temperature distribution, SFg DRE should
not be expected to' be similar, since in flame conditions, SFg would "see" or
experience a wide range of temperatures in the incinerator environment. At any
rate, the behavior is not the same, and more investigation is. warranted.
38
-------
s
7
6
5
4
3
2
SF6 THERMAL DECOMPOSITION
PILOT/LAB COMPARISON
ORE (-LOG(1-DRE/100))
D
D
n
D
n
800 900 1000 1100 1200 1300
TEMPERATURE, DEG C
1400
REF 5
CRF REGRESSION
REF 5 and 7
CRF DATA
Figure 3. SF6 Thermal decomposition pilot/lab comparison.
39
-------
POHC RANKING
Figure 4 is a contour plot of the relationship between conditions necessary
to achieve 6-nines SFg ORE and that of the individual POHCs. It also illustrates
why looking at one variable at a time can be confusing. At low Oo
concentrations, tetrachloroethylene requires only 900°C for 6-nines while at 4.5%
Oo> 1250°C is required, surpassing that for chlorobenzene, which was more stable
at the lower Q£ concentrations. Pentachlorobenzene DREs were never below
6-nines and therefore are not plotted.
Ranking the POHCs is shown in Figure 4. This figure illustrates that
rankings can change, depending on combustion conditions. This is also true
depending on the nature of the experimental device.
Table 3 is a list of rankings of POHCs derived from Figure 4 and from data
supplied by UDRI in Reference 9. Although not definitive, the research into
POHCs and POHC rankings have produced results which seem to make surrogate use
even more questionable than when it was first suggested.
Table 3. POHC Ranking'With 02 Present9
LIS Incineration
UDRI Low 0? (1.5%) High 09 (5%)
SFg
Tetrach 1 oroethy 1 ene
Toluene
Pentachlorobenzene
Chlorobenzene
1
2
5
3
4
1
4
2
5
4
2
3
1
5
3
being most stable; 5 being least stable.
CONCLUSIONS
o SFg is a limited surrogate. Data showing reactivity with refractory
and difficulty with using it may have reduced its apparent value
o POHC ranking Js;not absolute and depends on combustion conditions
o Toluene is themost stable1'of the orgahics at elevated temperature
and 02 levels, while SFg is most stable at intermediate and lower
levels.
RECOMMENDATIONS
o Modify the use of surrogates for trial burns, and specify
minimum combustion cbnditions
o Choose a minimum number of POHCs for trial burns
o Use the most stable POHCs at low oxygen values, since that condition
is present in most incinerator failures
o Develop method for continuous monitoring of toluene and benzene,
for performance monitoring.
40
-------
RELATIVE POHC STABILITY
T
E
M
P
E
R
A
T
U
R
E
C
l,^UU
1,300-
1,200-
1,100-
1,000-
900-
Qf\f\
* • . ' '
*.'
„,*'
A*
.••"i*
• ' " " ^** ++
NJ/T^ j_~r
++"f
4++
+
4-
i
Dn ++ .'..••'
>DnapDn . ' . ....
+• DDDQ
+ DnnDaannaaDDannnnaD •
1 2 3 4 5 6 7 8 9 10
O2 CONCENTRATION
6-9'S POHC ORE
SF6 + TCE * TOL ° CLBZ
Figure 4. Relative POHC stability.
41
-------
REFERENCES
1. Oppelt, E.T. Incineration of Hazardous Waste - A Critical Review. JAPCA.
37: 558, 1987.
2. 40 CFR Part 264.343.
3. Tsang, W.M. and Shaub, W.M. Surrogates as Substitutes for Principal
Organic Hazardous Constituent Validation of Incinerator Operation. In:
Proceedings of the Second Conference on Municipal, Hazardous and'Coal
Wastes, Miami, FL 1983. p. 241.
4. England, W.G., Rappolt, T.J., Teuscher, L.H., Kerrin, S.L. and Mournighan,
R.E. Measurement of Hazardous Waste Incineration Destruction and Removal
Efficiencies Using Sulfur Hexafluoride as a Chemical Surrogate. In:
Proceedings of the 79th Annual APCA Meeting, 1986. 106:162097W. .
5. Pandompatam, B., Liem, A.J., Frenette, R. and Wilson, M.A. Effect of
Refractory on the Thermal Stability of SF6. JAPCA. 39: 310, 1989.
6. Del linger, B. Testing and Evaluation of a POHC/PIC Incinerability
Mixture. U.S. EPA 15th Annual Research Symposium, Cincinnati, Ohio, April
April 10-12, 1989.
7. Taylor, P.H. and Chadbourne, J. Sulfur Hexafluoride as a Surrogate.
JAPCA. 37: 729 1987.
8. 40 CFR 761.70a.
9. Waterland, I.E. and Staley, L.J. Pilot Scale Listing of SFg As A
Hazardous Waste Incinerator Surrogate. To be Presented at the 82nd
National Meeting of the Air and Waste Management Association, Dallas, TX,
1989. Paper 89-23B.4.
42
-------
A PROTOTYPE BAGHOUSE/DILUTION TUNNEL SYSTEM
FOR PARTICULATE SAMPLING ".". -
OF HAZARDOUS AND MUNICIPAL WASTE INCINERATORS
' 'by -'• .-'• , ,-" .' , •-;,.
P.M. Lemieux,
J.A. McSorley,
.' W.P. Linak., . " ".,. '.',.'..
United States Environmental Protection Agency
Air and Energy Engineering Research' Laboratory
Research Triangle Park, NC 27711
, ABSTRACT .
EPA's Air and Energy Engineering Research Laboratory (AEERL) has
developed a prototype baghouse/dilution tunnel sampling system. This system
was designed originally for the sampling of flue gas particulate from fossil
fuel combustors, but has been modified to obtain samples of particulate matter
from hazardous and municipal waste incinerators. _.- Samples collected by this
sampling system are to be used for health effects, studies. , The sampling
system simulates the flue gas quenching processes 'occurring'upon emission from
stack to the atmosphere. A nominal 10:1 dilution-• with ambient air promotes
nucleation of vapor-phase organic compounds and condensation on existing
particulate matter. This unit is able to sample 2.8 dscm/min (100 cfm) of
effluent. At the allowable particulate loading rate of 180 mg/dscm stipulated
by RCRA regulations, this sampler is able to capture approximately 20 g of
sample in 1 hour. At this rate, it is feasible to generate kilogram-sized
particulate samples that are adequate for bioassay directed fractionation
and/or mouse skin painting carcinogenicity tests. Replicate samples can also
be obtained, so that duplicate health effects tests can be performed, a luxury
not normally available. It is also possible to sample semi-volatiles using
XAD-2 either upstream or downstream of the baghouse.
-------
INTRODUCTION
Exhaust gases from combustion sources typically consist of a mixture of
Nar H2O, COs, O2f CO, NOX, acid gases, particulate matter, and organic
compounds. When the hot stack gases contact the atmosphere and cool from
approximately 200°C (400°F) to ambient conditions, the particulate can act as
nucleation sites for the condensation of semi-volatile hydrocarbons. These
particles, when respired, can potentially deposit toxic, carcinogenic, or
mutagenic material into the lungs of humans.
A substantial database has been accumulated on the mutagenicity of
various, combustion emissions, including coal-fired power plants, diesel
engines, and woodstoves. It is desirable to extend the database to include
hazardous and municipal waste incinerators in order to place their health
effects into perspective with those of other combustion emissions.
The difficulty in acquiring representative samples from incinerators for
the purpose of performing health effects research is compounded by the need
for large samples. Stack-based particulate samples from incinerators may
typically contain approximately 1% extractable organic material by mass.
Whereas chemical analyses require microgram-sized quantities of extractable
organic material, microbial bioassays require milligram-sized samples, and
animal cancer studies require gram-sized samples. In other words, 1 kg of
particulate is needed to provide sufficient extractable organic material to
perform animal studies. At typical stack particulate loading rates of 180
mg/m3, it is obvious that the small sampling systems currently in use are not
practical.
BACKGROUND
In 1982, EPA designed and constructed a large particulate sampler for
use in its synfuels program. The sampler, a combination baghouse and dilution
tunnel system, was designed to pull 2.8 dscm (100 cfm) of sample from a stack,
dilute it with 28 dscm (1000 cfm) of filtered ambient air to simulate
atmospheric quenching, and pass the diluted effluent through a fabric filter
to capture the particulate. The particulate could then be removed from the
filter bags by mechanical rapping or pulse jet. The sampler was designed and
constructed, but never used, because the synfuels program was discontinued
soon after the unit's construction. When the question of health effects
sampling of incinerators arose in 1987, this unit was resurrected and
modified.
APPARATUS
The apparatus, shown in Figure 1, consists of five main components: the
ambient filter housing, the sample probe and delivery line, the dilution
tunnel, the baghouse, and the blower. All parts of the system except the
blower are made from 304 or 316 stainless steel. The entire system can be
placed on a 48 ft (14.6 m) long flatbed trailer for transportation to field
test sites. The system can also be oriented in a number of different
configurations to allow for space limitations when the unit is to be mounted
on the ground.
44
-------
The ambient filter housing contains, in series, a particulate filter, a
charcoal filter, a High Efficiency Particulate Air (HEPA) filter, and a 10 kW
air preheater. This filtering system removes particulate matter and organics
from the ambient dilution air, and raises the temperature so that the
resulting diluted gas mixture is 5-10°C (10-20°F) above the dew point.
The sample itself is withdrawn from the stack using a stainless steel
probe and a flexible 7.62 cm (3 in.) Outside diameter (OD) stainless steel
sample line. A 3 in. NPT nipple is required on the stack to be sampled.
Flanges are used to connect the sample line to the dilution tunnel.
The dilution tunnel is made from 20 cm (8 in.) OD x 3.35 m (11 ft) long
tubing, and contains a number of ports for external sampling. Thermocouples
are positioned before; and after dilution. A 3-way valve provides for
continuous emission monitor (CEM) sampling before or after dilution. There
are also four 3-in. NPT ports that can be used for XAD sampling, pitot tube
measurements, or sling psychrometry.
The baghouse is 71 cm (28 in.) OD and 106 cm (42 in.) tall. It contains
a Gore-Tex filter cartridge (model #38293-2). An air jet above the cartridge
provides a reverse pulse of high pressure air or nitrogen. A Pyrex bulb at
the bottom of the baghouse catches the particulate matter that is pulsed off
the filter. The dilution tunnel connects with the baghouse tangentially, so
that a cyclonic effect is produced, driving some of the large particles to the
bottom of the baghouse without impinging on the filter.
The sample and the dilution air are drawn through the system, by an
induced draft blower, capable of providing 1.49 kPa (60 in. water) of static
head pressure. A gate valve isolates the blower from, the rest of the'system
so that the blower does not need to be shut off during pulsing.
OPERATION
The baghouse/dilution tunnel sampler has fairly substantial power
requirements. A 208 V, 3-phase, 90 A power supply is required to run the
blower and the preheater. Aside from the electricity requirements, the unit
is self-contained. It can be operated from an 80 kW generator if sufficient
local power is unavailable.
For the proper operation of this unit in cold weather, it is absolutely
critical to ensure that the temperatures of no surface exposed to the sample
gas falls below the dew point. Incinerator stack gas typically contains HC1,
which, when dissolved in water, reacts with the stainless steel sample system.
In the case of adverse ambient weather conditions, it is necessary to insulate
some or all of the exposed metal pieces, especially the sample delivery line.
Prior to taking a sample, it is necessary to flush the system with preheated
ambient air in order to raise the temperature of all exposed parts to at least
45°C (110°F) . Heated tape is also used to preheat the sample delivery line to
100°C (212°F). . .
A pitot tube can be inserted into the dilution tunnel to periodically
monitor the gas velocity. The dilution ratio can be calculated either
directly by CO2 ratios before and after dilution or indirectly by a
calculation based on the temperatures before ,and after dilution. The gas
velocity, along with the dilution ratio, can be used to calculate sample flow
and particulate loading rates.
45
-------
Periodic monitoring of the dew point .of the stack gases in the dilution
tunnel by sling psychrometry is necessary. It is quite simple to monitor wet
bulb temperature in the dilution tunnel due to the high velocity of the
diluted gas. The air preheater should be adjusted so that the diluted gas
temperature is. always slightly "above the dew point.
The dilution tunnel vacuum and the pressure drop across the baghouse are
monitored continuously. By adjusting'the blower gate valve so that the system
vacuum is approximately 6'.07 kPa (3 in. of'water), the flow rate through the
system can be held constant. The initial pressure drop is approximately 0.12
kPa (5 in. of water) with a new filter, -and rises to approximately 1.12 kPa
(45 in. of water) when the filter becomes loaded with particulate. Pulsing
the system will knock particulate off the filter, arid the pressure drop will
fall back to approximately 0.5 kPa (20 in. of'water)." After the particulate
drops into the bulb, the sample is removed and ' isolated, and the bulb is
replaced. ,' , . ....'...'..
RESULTS
To date,'the EPA prototype'baghouse/dilution tunnel sampler has-been
used at several different facilities,"with varying degrees of success. First,
the sampler was operated on the Rotary Kiln Incinerator Simulator at the
Environmental Research Center in Research "Triangle Park, NC. Experiments were
performed examining transient puffs resulting from batch charging of plastics.
Two samples, each weighing approximately 100 g, were acquired, and the EPA's
Health Effects Research Laboratory (HERL) performed Salmonella mutagenicity
(Ames) bioassays on the organic material extracted from the particulate. A
typical dose-response curve is shown in Figure 2. HERL is performing animal
cancer studies on the organic extracts at the present time.
The sampler was then-tested at several full-scale municipal waste
incinerators. During these tests, water condensation problems^ were
discovered. Because the flue gas from municipal waste combustion ^typically
contains HC1, the condensation of water in the stainless steel sampling system
was detrimental to the usefulness of the samples. When water condensed on^the
walls, it would absorb HC1 from the gas. stream, and create an acidic solution.
Aqueous HC1 can leach chromium (Cr) and other metals out of the stainless
steel. Cr present in bioassay samples ruins the bioassay because the Cr is so
toxic that it kills the bacteria before they can revert (mutate). Since the
sampler was initially designed for sampling fossil-fuel combustors, the
extremely high water content present in effluents from municipal waste
combustion (15-35% water by weight) was unforeseen. In the presence of low
ambient temperatures, the only way to ensure that the dilution air would
remain above the dew point was to provide external heating to the dilution
air. A 10 kW air preheater was installed after these tests.
After modifications, the baghouse/dilution tunnel was then used to
sample another full-scale municipal waste incinerator. Although the unit
operated properly, and water condensation was minimized, very little sample
was collected. Simultaneous filter samples from the EPA's Source Dilution
Sampling System (SDSS) recovered no material either. It is possible that the
particulate matter was so small that it passed through the filter, or that the
particulate matter was an inorganic fume rather than flyash. The Gore-Tex
filters are rated to capture 100% of the particulate matter greater than 0.6
Urn. Scanning electron microscopy indicated that the residue deposited on the
stack sample probe was mostly composed of particles smaller than 1 pm.
46
-------
The system was also used to sample fugitive emissions at an aluminum
smelter. In one of-the process buildings at the smelter, a high concentration
of vapor phase organic material and a considerable amount of particulate
matter were present. The system was able to collect a 35.5 g sample of
particulate matter, which was within 10% of the calculated capture, based on
reported particulate loading and estimated flow rates. The analyses for these
final two tests are being performed at the current time.
CONCLUSIONS
Earlier successful tests indicate that this system has potential for use
as^a high-volume particulate sampler for health effects studies. Problems do
exist, but they appear to be surmountable. This unit may not be suitable for
sampling at facilities with very small particulate . «1 |Jm) or with very high
water content (>30% by weight) . It also appears to be more attractive to
install a baghouse/dilution tunnel system in a permanent or semi-permanent
fashion at a facility. Even at extremely high sample volumes, several days or
even weeks of sampling may be necessary to ensure that sufficient particulate
matter is collected for subsequent health effects studies.
ACKNOWLEDGEMENTS
Special thanks to David M. De Marini of HERL- for providing the sample
mutagenicity data.
This paper has been reviewed in accordance with the US EPA's peer and
administrative review policies and approved for presentation and publication.
47,
-------
AMBIENTFILXERHOUSING
Fartkulate Filter
DM— Sample Gate Valve
- Thermocouple
Charcoal Filter
BLOWER
Blower Gate Valve
o
SAMPLE DELIVERY
\
Sample Line'
DILUTION TUNNEL
O
Probe •
• Thermocouple
Figure 1: Diagram of Prototype Baghouse/Dilution
Tunnel Sampling System
48
-------
1000 r
20 40 60 80
Dose (gg/plate)
100
Figure 2: Dose-Response Curve Derived from Salmonella TA98 Bioassay of
Particulate Matter Extract from Prototype Baghouse/Dilution Tunnel System
During the Combustion of Polyethylene in EPA's Rotary Kiln Incinerator
Simulator " '
49
-------
EVALUATION OF ALTERNATIVE TREATMENT
TECHNOLOGIES FOR HAZARDOUS WASTES
FROM ACRYLONITRILE PRODUCTION
E. Radha Krishnan
PEI Associates, Inc.
Cincinnati, Ohio
Ronald J. Turner
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Cincinnati, Ohio
ABSTRACT
Three aqueous waste streams from acrylonitrile production have been
included in the list of hazardous wastes under the U.S. Environmental Protec-
tion Agency's land disposal restrictions program: 1) the bottom stream from
the wastewater stripper (EPA Hazardous Waste No. K011); 2) the bottom stream
from the acetonitrile column (EPA Hazardous Waste No. K013); and 3) bottoms
from the acetonitrile purification column (EPA Hazardous Waste No. K014).
The listing constituents for K011, K013, and K014 include acrylonitrile,
acetonitrile, acrylamide, and hydrocyanic acid.
The waste streams contain suspended solids consisting primarily of spent
metallic oxide catalyst particles and acrylonitrile polymers. Current prac-
tice in industry is to mix the aqueous waste streams in settling ponds/ tanks
where the suspended solids are separated as a sludge, and to dispose the
liquid stream by deep well injection. Because the sludge is derived from
hazardous wastes K011, K013, and K014, it has the same listing constituents
as the aqueous waste streams. The sludge is generally disposed in offsite
landfills.
This paper presents the results from: 1) a bench-scale evaluation of wet
air oxidation as a treatment technology for the mixed K011/K013/K014 waste
stream, and 2) a field investigation to evaluate the performance of rotary
kiln incineration as a treatment technology for the sludge. The wet air-
oxidation tests were conducted at Zimpro Passavant's research facility in
Rothschild, Wisconsin, and the incineration tests were conducted at the John
Zink Company's test facility in Tulsa, Oklahoma.
50
-------
INTRODUCTION
The Risk Reduction Engineering Laboratory (RREL) of the U.S. Environ-
mental Protection Agency (EPA) is providing the Office of Solid Waste (OSW)
with data on various hazardous waste treatment technologies to assist in the
development of land disposal restriction standards under the Resource Conser-
vation and Recovery Act (RCRA). Three aqueous wastes from acrylonitrile
pr2d™Jl?n (which are 1dentified by EPA Hazardous Waste Numbers K011, K013,
and K014) result from the purification of product streams in the acryloni-
trile production process and are defined as follows:
K011: Bottom stream from the wastewater stripper.
"
K013: Bottom stream from the acetonitrile column.
K014: Bottoms from the acetonitrile purification column.
In addition to these wastes, a K011/K013/K014 sludge is also generated
at acrylonitrile production plants by separation of the suspended solids from
the mixed aqueous wastes.
The EPA is required to'set treatment standards for -acrylonitrile produc-
tion wastes based on the best demonstrated available technology (BOAT), as a
prerequisite for the placement of treatment residues in land disposal facil-
ities.. The effective date of the treatment standard is June 8, 1989. In
1987-88, EPA:RREL conducted a program consisting of 1) sampling and analysis
to characterize K011/K013/K014 aqueous and sludge waste streams from several
acrylonitrile producers, and 2) treatability testing of wet air oxidation and
incineration on the mixed K011/K013/K014 aqueous waste stream and sludge
respectively, from one acrylonitrile production facility. This paper pre-
sents the results of these tests. - » •
WASTE CHARACTERIZATION - ,
There are six facilities in the United States that are involved in the
production of acrylonitrile and could generate K011, K013, and K014 listed :
wastes. A few facilities do not purify the crude acetonitrile stream pro-
duced in the process and, hence, do not generate K014. The listing con-
stituents for K011, K013, and K014 include acrylonftrile, acetonitrile,
acrylamide, and hydrocyanic acid. , , ?
51
-------
The K011 waste stream represents the largest single pollutant load
within the acrylonitrile production process. Typically, the K011 waste
stream contains about 100 to 4,000 ppm of cyanide, 40 to 3,000 ppm of
acetonitrile, 0.2 to 8,000 ppm of acrylonitrile, 1,000 to 2,000 ppm of
acrylamide, and less than 200 ppm of acrolein. In addition to the primary
contaminants listed above, this stream also contains approximately 4 percent
suspended solids. The suspended solids consist largely of spent metallic
oxide catalyst particles and polymeric acrylonitrile. Also, the K011 stream
contains about 10 percent dissolved sulfates. Typical generation rates at
acrylonitrile plants for the K011 waste stream vary from 100 to 200 gallons
per minute.
The K013 waste stream typically contains about 26 to 60 ppm of cyanide,
less than 35 ppm of acetonitrile, less than 10 ppm of acrylonitrile, less
than 120 ppm of acrylamide, and less than 1 ppm of acrolein. This waste
stream constitutes greater than 99 percent water. Typical generation rates
for the K013 waste stream at acrylonitrile plants vary from 100 to 200 gal-
lons per minute.
Primary pollutants in the K014 waste stream are acetonitrile and cya-
nide. Typically, the K014 waste stream contains 1,000 to 60,000 ppm of
acetonitrile, and 5 to 5,000 ppm of cyanide. Typical generation rates for
the K014 waste stream at acrylonitrile plants vary from 5,000 to 20,000
gallons per day.
In a typical acrylonitrile production facility, the aqueous waste
streams (K011, K013, and K014) are comingled prior to their ultimate dis-
posal. The mixed waste is sent to settling ponds/tanks where the suspended
solids are removed as an underflow sludge. The liquid effluent is disposed
of by deep well injection. The sludge generation rate per plant varies from
100 to 250 tons per year. The predominant practice in industry is to peri-
odically dispose of the K011/K013/K014 sludge in offsite landfills. The
approximate concentrations of the major constituents in the K011/K013/ K014
sludge are as follows:
Constituent
Silicon, molydenum, iron, and aluminum oxides
Acrylonitrile polymers
Inert shell, dirt, gravel, and kiln dust
Water
Acrylonitrile
Acetonitrile
Acrylamide
Cyani de
Concentration
range
2 to 50%
2 to 25%
10 to 40%
10 to 50%
0.4 to 1 ppm
0.7 to 3 ppm
2 to 3 ppm
900 to 2000 ppm
The sludge has a heating value of 2,400 to 3,000 Btu/lb.
52
-------
TEST FACILITY DESCRIPTIONS
Several technologies, or combinations of technologies, may be applicable
for treatment of the mixed K011/K013/K014 aqueous wastes and sludge from the
acrylonitrile production process. Some applicable technologies for treatment
of the listed organic constituents and cyanide present in the mixed K011/
K013/K014 aqueous wastes include (1) wet air oxidation, (2) wet air oxidation
followed by biological treatment, and (3) critical fluid extraction followed
by additional cyanide treatment of the aqueous phase and recycling or incin-
eration of the solvent phase. An applicable technology for treatment of the
cyanide or organic components in the K011/K013/K014 sludge is rotary kiln
incineration. The incineration technology may be combined with stabilization
for further treatment of the metals present in the incinerator ash.*
In 1988, EPA-RREL conducted a bench-scale evaluation of wet air oxida-
tion technology for treatment of the mixed K011/K013/K014 aqueous wastes and
a pilot-scale evaluation of rotary kiln incineration technology for treatment
of the K011/K013/K014 sludge. The mixed aqueous wastes and the sludge for
the tests were obtained from the same acrylonitrile production facility. The
wet air oxidation tests were conducted at Zimpro Passavant's research facili-
ty in Rothschild, Wisconsin, and the incineration tests were conducted at the
John Zink Company's test facility in Tulsa, Oklahoma.
Zimpro Wet Air Oxidation Treatment System
Wet air oxidation is a liquid-phase, oxidation process in which an
aqueous solution or suspension of organic and/or oxidizable inorganic com-
pounds are oxidized to carbon dioxide and other innocuous end products. The
waste stream is thoroughly mixed with a gaseous source of oxygen (usually
air) at temperatures of 175 to 327°C (347-621°F) and pressures of 300 to
3,000 psig. Elevated temperatures enhance the solubility of oxygen in aque-
ous waste thus providing a strong driving force for oxidation. Elevated
pressures are required to control evaporation by maintaining water in the
liquid state.
The bench-scale wet air oxidation tests on the combined K011/K013/K014
aqueous waste stream were carried out in 0.5-liter capacity titanium auto-
claves at Zimpro's research facility. During each test, 125 ml of raw waste
was charged into the autoclave. The autoclave was pressurized to 810 psig
with air to provide about 120 percent of the sample chemical oxygen demand
(COD) air requirements. The autoclave was then placed in a heater-shaker
mechanism, heated to the desired temperature, and maintained at the tempera-
ture for the designated residence time. The autoclave was cooled with tap
water, immediately after the completion of the residence time, and the off-
gases were analyzed by gas chromatography for total hydrocarbons and methane.
A series of three autoclave oxidations were performed at temperatures of
200°, 240°, and 280°C (392% 464°, and 536°F) and a liquid residence time of
60 minutes. A total of four autoclave "runs were made at each condition and
the oxidized material combined before analyses.
U.S. Environmental Protection Agency, Office of Solid Waste, Washington,
D.C. Best Demonstrated Available Technology (BOAT) Background Document
for K011, K013, and K014. December 1988.
53
-------
John Zink Rotary Kiln Incineration System
Figure 1 presents a schematic of the John Zink pilot-scale rotary kiln
incineration system used for conducting tests on the K011/K013/K014 sludge.
The test system consisted of a rotary kiln and kiln afterburner for combus-
tion, a ram feeder for feeding waste contained in fiber packs into the kiln,
a water quench for kiln ash, and a venturi-scrubbing system for flue gas
treatment. In addition, the Oklahoma State Department of Health required
that an additional afterburner, a fume incinerator, -be used during all tests
involving hazardous waste to provide further thermal treatment of the primary
combustion system's flue gas. , The minimum heat input to the rotary kiln
incineration system (including waste and natural gas supplementary fuel) was
2 million Btu/h and the maximum was 3 million Btu/h. The kiln provided a
1-hour solids' residence time at 0.25 rpm. The kiln afterburner provided a
flue gas residence time of 2.2 seconds at 2000°F. Temperatures were con-
trolled by manually changing the supplementary fuel feed rate and/or the
combustion air flow rate.
Continuous monitoring and recording of C02, CO, 02, and temperature were
conducted at the exit of the kiln afterburner at a point prior to the solids
separator.
The hot combustion gases leaving the kiln afterburner entered a cyclonic
solids separator, which removed large pieces of fly ash that might damage
downstream equipment. The gases then entered an adjustable-throat venturi
scrubber, which was followed by a droplet separator consisting of a cy-
clonic-flow knockout chamber and a vent stack. A 15 percent sodium carbonate
solution was periodically added to the scrubber water at a rate of 0.8 gal-
lons per minute for a total of 10 to 15 minutes during each of the incinera-
tion tests to maintain the pH of the scrubber recycle water between 6 and 7.
The scrubber did not require any blowdown during the tests.; Makeup water,
however, was continuously added to the scrubbing system to compensate for
evaporation losses.
A total of three incineration tests, each of 2 to 2\ hours .duration,
were conducted on the K011/K013/K014 sludge. The waste feed rate for all the
tests was 400 Ib/h; this feed rate corresponded to^6ne,: 10-1 b fiber pack with
K011/K013/K014 sludge being fed to the kiln every 90 seconds'. Each sludge
fiber pack was spiked with 2 weight percent of chlorobenzene to evaluate
destruction removal efficiency (ORE) for the incineration tests. Chloro-
benzene was selected as a surrogate principal organic hazardous constituent
(POHC) for ORE evaluation because of 1) stack gas sampling arid analytical
problems associated with the least incinerable POHC in the sludge (i.e.,
acetonitrile), and 2) the close ranking of incinerability parameters (i.e.,
heat of combustion and thermal stability) for chlorobenzene and acetonitrile.
During the three tests, the rotary kiln incineration test system operated
under near steady state conditions at average kiln operating temperatures of
1800°F and kiln afterburner temperatures of 2000°F. The pressure drop across
the venturi scrubber averaged about 25 inches of water. Stack gas oxygen
concentrations averaged about 4.2 percent, and CO concentrations were less
than 10 ppm.
54
-------
(O
•o 10
E -(->
(O (/I
0}
•O 4J
O)
o; c
M- O
•r~
TD •»->
C res
>••-
to
0)
+-> I—
S-
o o
S-
-------
Following the incineration tests, stabilization testing was conducted at
the U.S. Army Waterways Experiment Station (WES), Vicksburg, Mississippi, on
a composite kiln ash sample from the three tests.
TEST PROTOCOLS
Wet Air Oxidation Tests
The following four aqueous streams were sampled for the combined K011/
K013/K014 bench-scale wet air oxidation tests:
K011/K013/K014 aqueous waste feed
Product from oxidation at 200°C (392°F) and 810 psig
Product from oxidation at 240°C (464°F) and 810 psig
Product from oxidation at 280°C (536°F) and 810 psig
The raw waste and the treated waste samples from each of the three tests
were analyzed for the organic listing constituents of K011/K013/K014 as well
as total and amenable cyanide content. Chemical analyses for the listed
organic compounds (with the exception of acrylamide) and cyanide were per-
formed according to methods, from U.S. EPA's Test Methods for Evaluating Solid
Wastes, Third Edition, SW-846, November 1986. Acrylamide was quantitated by
high pressure liquid chromatography with ultraviolet detection. Additional
analyses were also conducted on all four samples for sixteen metals and
volatile organic priority pollutant compounds.
Rotary Kiln Incineration Tests
Figure 1 identifies the five streams that were sampled for the rotary
kiln incineration tests of the K011/K013/K014 sludge:
Sample Point
A
B
C
D
E
Description
K011/K013/K014 sludge waste feed
Kiln bottom ash
Fly ash from solids separator
Scrubber water before waste feed
Scrubber water durina waste feed
Eight different K011/K013/K014 sludge feed samples (A) were submitted
for analysis, two from the first test and three from each of the subsequent
two tests. The samples represented feed to the kiln at 1-hour intervals.
Six different kiln ash samples (B) were collected: two from each test. The
samples were collected at 1 to li hour intervals to enable the waste repre-
sentative of the ash sample to be treated for 1 to U residence times before
ash sample collection. One fly ash sample (C) was collected after the com-
pletion of the three tests. One sample of the pretest scrubber water (D) was
taken immediately prior to initiation of the first K011/K013/K014 sludge
incineration test. Six scrubber water samples (E) were collected: two
during each test. One sample was obtained near the midpoint of the test, and
56
_
-------
one near the end of the test period. Comprehensive analyses were performed
on the samples for the following parameters: volatiles, semivolatiles,
metals, water quality parameters, organochlorine pesticides, phenoxyacetic
herbicides, organophosphorous herbicides, polychlorobiphenyls (PCBs), dioxins
and furans. Table 1 identifies the specific constituents that were detected
above the practical auantitation limit (PQL) in any of the sludge feed, ash,
or scrubber water samples. Samples were prepared and analyzed in accordance
with SW-846 methods. Acrylamide, one of the organic listing constituents of
the K011/K013/K014 sludge, could not be analyzed by standard SW-846 methods;
high pressure liquid chromatography was used for acrylamide analysis.
Stack sampling was conducted during each of the three tests at a loca-
tion after the venturi scrubber control system, but prior to the fume in-
cinerator, to determine the concentrations and mass emission rates of the
following:
0 Volatile organics, including surrogate POHC (chlorobenzene) for ORE
evaluation
0 Particulate matter, hydrochloric acid (HC1) and total cyanide (CM)
Metals
Volatile organic emissions were measured in accordance with the Volatile
Organic Sampling Train (VOST) protocol (SW-846, Methods 0030 and 5040).
Particulate, HC1, and CN concentrations and mass rates were measured with a
modified EPA Method 5 sample train. Metal emissions were collected with a
modified EPA Method 12 sampling train.
Three different samples from stabilization testing on the composite kiln
ash were submitted for metals analysis.
TEST RESULTS
Wet Air Oxidation Tests
Table 2 presents the analytical results for the organic listing con-
stituents and cyanide for each of the three wet air oxidation tests. The
combined K011/K013/K014 aqueous waste feed for the test contained 575 ppm of
acetonitrile, 24.4 ppm of acrylonitrile, and 270 ppm of acrylamide. The
amenable and total cyanide contents of the waste feed were 1,011 and 1,277
ppm, respectively. The data from the tests showed that appreciable oxidation
of acetonitrile only occurred at 280°C (536°F), corresponding to an acetoni-
trile concentration of 90 ppm in the oxidized liquor. This represents an 84
percent reduction relative to the feed. Acrylonitrile was detected at 0.47
ppm in the liquor from oxidation at 200°C (392°F), but nondetectable at the
higher temperatures. The acrylamide results indicated reductions of 96 to 98
percent in the oxidized samples, corresponding to residual concentrations of
11 to 5 ppm. Amenable and total cyanide destruction efficiencies approached
57
-------
TABLE 1. ANALYTES DETECTED IN SAMPLES FROM K011/K013/K014
SLUDGE INCINERATION TESTS
Volatiles
Metals
Water quality
parameters
Organophosphorous
pesticides
Acetonitrile
Acrylonitrile
Chloroform
Methylene Chloride
1,1,1-Trichloroethane
Trichloroethylene
Acetone
Benzene
Styrene
Silver,,
Aluminum'
Arsenic
Barium
Beryllium
Cadmium
Chromium
Copper
Iron
Molybdenum
Nickel
Lead
Antimony
Silicon
Vanadium
Zinc
Hydrogen sulfide
Sulfate
Total organic carbon
Total cyanide
Chloride
Fluoride
Phorate
Ethyl parathion
TABLE 2. ANALYTICAL RESULTS FROM BENCH-SCALE WET AIR
OXIDATION TESTS ON K011/K013/K014 AQUEOUS WASTE
Analyte
Acetonitrile
Acrylonitrile
Acryl amide
Feed
ppm
575
24
270
Cyanide (amenable) 1011
Cyanide (total)
a Product from
Product from
0 Product from
Not detected.
1277
oxidation
oxidation
oxidation
Oxidized
, Liquor No. 1 ,
ppm
690
.4 0.47
11
37
39
at 200°C and 810
at 240°C and 810
at 280°C and 810
Oxidized .
Liquor No. 2 ,
ppm
560
NDd
10
11.4
13.6
psig.
psig.
psig.
Oxidized
Liquor No. 3,
ppm
90
NDd
5
12.9
15.5
58
-------
99 percent at the two higher oxidation temperatures, corresponding to cyanide
residual levels of 11.4 to 15.5 ppm. There was no significant difference in
the residual cyanide levels at 240°C (464°F) and 280°C (536°F). Higher
operating pressures than those used in the bench-scale wet air oxidation
tests should result in further reductions of residual cyanide levels.
Molybdenum was the metal present at the highest concentration (23.6 ppm)
in the waste feed. Other metals present in the feed at concentrations above
1 ppm include barium (15.4 ppm), zinc (1.8 ppm), and nickel (1.08 ppm). The
products from oxidation contained approximately 22 ppm of molybdenum, ap-
proximately 3 ppm of barium, up to 54 ppm of zinc, approximately 3 ppm of
nickel, and up to 391 ppm of copper. The higher concentrations of zinc,
nickel, and copper in the oxidized products relative to the feed are attri-
buted to possible solubilization of metals from the solids in the feed.
Off-gases from the tests contained about 25 ppm of total hydrocarbons
and a trace of methane. The oxygen concentration in the off-gases was main-
tained in excess of 6 percent for all three tests, and CO concentrations were
below the detectable level of 100 ppm.
Rotary Kiln Incineration Tests
Table 3 presents the analyses for the detected volatile organic com-
pounds and total cyanide in K011/K013/K014 sludge feed samples from each of
the three incineration tests. The K011/K013/K014 sludge used for the tests
contained low levels of acetonitrile (up to 2.7 pg/g) and acrylonitrile (0.95
ug/gh characteristic of sludges generated from physical treatment of aqueous
wastes from acrylonitrile production. Acrylamide was not detected in the
sludge feed above its PQL of 6.5 yg/g. Other volatiles detected at high
levels in the K011/K013/K014 sludge feed were benzene and styrene.
Tables 4 and 5 present the incinerator ash and scrubber recycle water
analyses, respectively, for the volatile organic compounds and total" cyanide
in samples from each of the three incineration tests. No volatiles were
detected in the ash and scrubber recycle water samples. Lower detection
limits could not be achieved for the volatile analytes in the ash samples
because of analytical interferences. Total cyanide concentrations were
reduced from levels as high as 2,000 pg/g (ppm) in the sludge feed to levels
of less than 22 pg/g in the ash. No cyanide was detected in the scrubber
recycle water samples. As with the feed, no acrylamide was detected in the
ash and scrubber recycle water samples above the PQL of 6.5 pg/g and 1125-
mg/liter, respectively.
The major metals found in the K011/K013/K014 sludge feed samples were
aluminum (up to 1100 pg/g), barium (up to 200 pg/g), chromium (up to 200
ug/g), Iron (up to 4000 pg/g), nickel (up to 470 pg/g), and zinc (up to 210
pg/g). Molybdenum was also detected in the feed samples at levels as high as
17,000 pg/g. The major metals in the kiln ash were molybdenum (up to 45,000
ug/g), iron (up to 32,000 pg/g), aluminum (up to 6,700 pg/g), and nickel (up
to 2,700 pg/g). The ash Toxicity Characteristic Leaching Procedure,(TCLP)
extracts contained the following metals above their respective PQLs: ar-
senic, barium, chromium, copper, lead, nickel, selenium, and zinc.^Nickel
was the metal with the highest ash TCLP value at concentrations of up to 14.0
mg/liter (ppm). For the scrubber recycle water samples, the major metals
59
-------
r
TABLE 3. VOLATILES AND CYANIDE ANALYSES OF K011/K013/K014
SLUDGE FEED FOR ROTARY KILN INCINERATION TESTS
Analyte
Test No. 1, Test No. 2,
yg/g yg/g
Test; No., 3,
yg/g
Volatiles-
Acetonitrile
Acrylonitrile
Chloroform
Methylene chloride
1,1,1-Trichloroethane
Trichloroethylene
Acetone
Benzene
Styrene
Total cyanide
0.870
0.410
0.032
0.034
0.045
0.016
NDa
57
16
1,200
1.200
0.540
0.030
0.240
0.029
0.014
0.095
48
15
2,000
2.700
0.950
0.042
0.210
0..032
0.018.
NDa
55
18
1,500
Not detected.
60
-------
TABLE 4. VOLATILES AND CYANIDE ANALYSES OF K011/K013/K014
INCINERATOR ASH FOR ROTARY KILN INCINERATION TESTS
Analyte
Test No. 1,
yg/g
Test No. 2,
yg/g
Test No. 3,
Volatiles
Acetonitrile
Acrylonitrile
Chloroform
Methylene chloride
1,1,1-Trichloroethane
Trichloroethylene
Acetone
Benzene
Styrene
Total cyanide
<500
<500
< 10
<250
< 10
< 10
<250
< 10
< 10
10
<500
<500
< 10
<250
< 10
< 10
<250
< 10
< 10
12
<500
<500
< 10
<250
< 10
< 10
<250
< 10
< 10
22
61
-------
TABLE 5. VOLATILES AND CYANIDE ANALYSES OF K011/K013/K014
SCRUBBER RECYCLE WATER FOR ROTARY KILN INCINERATION TESTS
Analyte
Test No. 1,
mg/liter
Test No. 2,
mg/liter
Test No. 3,
mg/liter
Volatiles
Acetonitrile
Acrylonitrile .
Chloroform
Methylene chloride
1,1,1-Tri chloroethane
Trichloroethylene
Acetone
Benzene
Styrene
Total cyanide
<0.5
<0.5
<0.01
<0.25
<0.01
<0.01
<0.25
<0.01
<0.01
<0.5
<0.5
<0.01
<0.25
<0.01
<0.01
<0.25
<0.01
<0.01
<0.5
<0.5
<0.01
<0.25
<0.01
<0.01
<0.25
<0.01
<0.01
62
-------
present were iron (up to 99 rug/liter), molybdenum (up to 260 rag/liter),
silicon (up to 16 mg/liter), and zinc (up to 4 nig/liter).
With one exception, no organophosphorous pesticides, organochlorine
pesticides/PCBs, organochlorine herbicides, dioxins or furans were detected
above the PQL in any of the-,K011/K013/K014 sludge feed, ash, or scrubber
water samples. One sludge feed sample contained 0.11 pg/g of phorate and
0.27 yg/g of ethyl parathion.
Stack emissions data showed a ORE greater than 99.99 percent for the
surrogate POHC, chlorobenzene. Particulate concentrations corrected to 7
percent oxygen ranged between 0.10 and 0.31 grains per dry standard cubic
foot (gr/dscf) for the three tests, with an average of 0.21 gr/dscf; these
results exceed the required RCRA limit of 0.08 gr/dscf at 7 percent oxygen.
The failure to meet the RCRA particulate limit was attributed to poor per-
formance of the scrubber on the pilot-scale rotary kiln incineration test
system. Full-scale air pollution control systems should be able to meet the
RCRA particulate limit. HC1 emission rates ranged between 0.01 and 0.02
Ib/h, considerably below the 4.0 Ib/h RCRA limit. Cyanide was not detected
in the stack emission samples. Iron and molybdenum were the metals that
exhibited the highest concentrations in the stack gas, corresponding to mass
emission rates of up to 0.5 Ib/h and 0.1 Ib/h, respectively.
Complete results from the kiln ash stabilization tests are currently not
available.
CONCLUSIONS
Wet air oxidation and incineration appear to be effective treatment
methods for the K011/K013/K014 aqueous wastes and sludge, respectively, based
on the results of the tests conducted by EPA-RREL. The wet air oxidation
tests demonstrated significant reduction of the K011/K013/K014 organic list-
ing constituents and cyanide at the highest oxidation temperature of 280°C
(536°F). It is believed that residual organic constituents from wet air
oxidation would be amenable to biological treatment. Residual metals in the
oxidized liquor may be treated by chemical precipitation. Operating data
collected during incineration tests of the K011/K013/K014 sludge show that
rotary kiln incineration is an appropriate treatment technology for the
sludge. The ORE performance standard of 99.99 percent was achieved for the
surrogate POHC, chlorobenzene. The incineration tests demonstrated signifi-
cant reductions of organic constituents and cyanide in the ash. Metals in
the residual ash will require subsequent treatment by stabilization pro-
cesses.
63
-------
The Role of Site Investigation
in the Selection of Corrective Actions
for Leaking Underground Storage Tanks
by: Myron S. Rosenberg
David C. Noonan
Camp Dresser & McKee Inc.
One Center Plaza
Boston, MA 02108
and
Anthony N. Tafuri
Chi-Yuan Fan
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Woodbridge Avenue
Edison, NJ 08837
and
Iris Goodman
U.S. Environmental Protection Agency
Office of Underground Storage Tanks
401 M Street, S.W.
Washington, DC 20460
ABSTRACT
There are numerous sites across the nation where soil treatment
technologies are being applied to clean up soil contaminated with petroleum
hydrocarbons from leaking underground storage tanks (USTs). Developing an
accurate understanding of subsurface conditions at a site (i.e., a site
investigation) increases the likelihood that a given soil treatment
technology will be effective at a given site. This paper presents an
approach for conducting a site investigation including identifying what
information about the subsurface environment and the released petroleum
product is needed and how it can be used.
Worksheets are provided, to help the reader make a preliminary
determination as to where in the unsaturated zone most of the petroleum
hydrocarbons are likely to be and where they are likely to move. Critical
success factors, or factors for determining how effective a given soil
treatment technology will be in cleaning up a particular site, are
provided. Worksheets that contain the critical success factors for two
soil treatment technologies (soil venting and biorestoration) are provided
in this paper so that a cross-comparison of site conditions can be
reviewed.
64
-------
INTRODUCTION
Cleaning up petroleum hydrocarbons in the unsaturated zone is a real
world problem at every site where releases have occurred. The science
associated with soil treatment technologies is not as well documented or
understood as that associated with water treatment technologies. Despite
this uncertainty, however, soil contamination must always be addressed by
engineers on behalf of the owners and operators at an UST site, and some
kind of soil cleanup is typically necessary when a release occurs.
Furthermore, the proposed approach for cleaning up the unsaturated zone
must be approved by the regulatory agencies charged with monitoring the
cleanup.
The question at hand is how to move ahead with cleaning up the
unsaturated zone by selecting a treatment technology that is likely to be
effective given the uncertainty surrounding the removal of petroleum
hydrocarbons from soil. The purpose of this paper is to provide an
approach for evaluating the likely effectiveness of soil treatment
technologies in the face of incomplete or uncertain data. Our approach
builds on the present state of knowledge of soil treatment with a focus on
utility; what technologies will likely work given our (limited)
understanding of conditions at the site.
Intuitively, certain basic information is known about the site.
Information can be used to develop what we have termed a "site
investigation".
This
A good understanding of the conditions in the unsaturated zone is
essential to selecting an appropriate soil treatment technology. A good
site investigation includes enough information to answer the following
questions:
• What was released? Where? When?
• Currently, where in the unsaturated zone is most of the petroleum
likely to be?
• How much petroleum product is likely to be present in different
locations and phases?
« How mobile are the constituents of the contaminant, and where are
they likely to travel and at what rate?
Answers to these questions—or at least "ballpark estimates" of the
answers — can sometimes be easily developed. Often an estimate is
sufficient to gain a relative understanding of the potential effectiveness
of a given technology at a given site. The need for site-specific measure-
ments of important parameters can never be eliminated, because there is no
substitute for accurate, site-specific, field data. However, it is
possible to combine actual measurements with literature values for other
65
-------
parameters (for which no field data are presently available) to make at
least qualitative assessments of site conditions and to judge which
corrective action technologies are likely to work.
OVERVIEW OF APPROACH
Figure 1 shows the three basic components of the approach: 1) site
investigation, 2) technology selection, and 3) monitoring and follow-up
measurements. Figure 1 also shows which questions are answered in each
step. Critical success factors, or CSFs as shown on Figure 1, are
parameters that are likely to determine the effectiveness of a given
technology in a given situation. They are discussed in detail later in
this paper.
Site Investigation
A site investigation begins with basic information about the release
itself. Basic information about the release may be obtained by asking:
• What contaminants were released?
• How much was released?
• Was the release slow or instantaneous?
• How long ago did the release stop?
• When was the release detected?
Information about the site is gathered next, and usually involves
searches of records and use of professional expertise. Typical information
sought about the site includes:
Soil Temperature
Soil pH
Rainfall, Runoff, and Infiltration Rate
Soil Surface Area
Organic Content
Soil Porosity
Particle density
Bulk density
Hydraulic Conductivity
Permeability ,
Field Capacity
Soil Water Content
Local Depth to GW
If information regarding these critical parameters is unavailable for
the site of interest, there are various sources of data that can provide
default values for making ballpark estimates of these critical parameters.
66
.
-------
KEY
QUESTION
WHAT WAS
RELEASED?
WHERE IN THE
,UN3ATURATED
' ZONE IS IT?'1
HOW MUCH IS
THERE?
WHERE IS IT
GOING?
WHAT ARE THE
TECHNOLOGIES?
WHAT ARE EACH
TECHNOLOGY'S
ADVANTAGES?
WHAT ARE THE
CRITICAL SUCCESS
FACTORS?
.WHAT WILL. WORK
'AT MY SITE? ,
EVALUATION OF
SUCCESS*
SELECTION
ACTION
SITE INVESTIGATION
EVALUATE PHASE(S) OF
CONTAMINANT(S) IN SOIL
SELECT TECHNOLOGY
BASED ON YOUR
SITE AND HOW IT MATCHES THE
TECHNOLOGIEtfCSFs
CE
PERFORMANCE MONITORING
SAMPLING & MEASUREMENT
OFSITECONDTIONS
COMPARE TO PRE-ESTABLISHED
CLEAN-UP GOALS
YES
NO
« CONTINUE TO MONITOR
Figure 1 . An Overview of the Approach
67
-------
In addition to release-related and site-related information, a site
investigation should include an understanding of the physical and chemical
nature of the contaminants released. By drawing the relevant physical and
chemical characteristics, it is possible to estimate how the contaminant
may partition in the subsurface (what phase it is likely to be in), how
readily it will move away from the site as a vapor or liquid, and whether
it is likely to degrade significantly over time. Contaminant-specific
parameters, which can be found in many chemical handbooks, include:
Unweathered Composition
Pore Vapor Pressure
Water Solubility
Liquid Viscosity
Liquid Density
Vapor Density
Soil Sorption Coefficient
With this information in hand, some basic determinations can be made
as to "which "phase" most of the contaminants are in.
For simplicity, it was assumed that petroleum hydrocarbons in the
unsaturated zone can exist only in three phases (as shown on Figure 2): as
contaminant vapors in the pore spaces (vapor phase), as residual liquid
trapped between soil particles (liquid phase), or as liquid dissolved in
the pore water that surrounds soil particles (dissolved phase). Note that
for recent petroleum releases (less than 1 year ago), most of the petroleum
is likely to be in the residual saturation phase with somewhat smaller
portions existing in the vapor and dissolved phases. Although only 3
phases are considered in this paper, petroleum hydrocarbons can be found in
other phases as well. Recent research done by EPA's RREL Office in Edison,
New Jersey identified as many as 14 different conditions under which
petroleum could be found in the subsurface.
Finally, critical parameters for evaluating the mobility of contami-
nants exist. Mobility is used here as a general term to indicate how
readily a contaminant moves into air and water. Vapor pressure and
solubility are the primary indicators used in determining product mobility.
Vapor pressure is an indicator of how easily contaminants will volatilize
into air. Solubility is an indicator of a contaminant's affinity for
water. The mobility of the contaminants in the unsaturated zone is very
important. Mobility directly affects the choice of corrective action
technology.
Technology Selection
The second step of the approach involves selecting a technology based
on the collected information. Critical success factors (CSFs), or factors
that are likely to determine the effectiveness of a given technology in
cleaning up the unsaturated zone in a given situation, can be identified
for each technology. Critical success factors for soil venting, for
example, include the volatility of the contaminants, soil temperature,' and
moisture content of the soil. Conditions at any site can be compared with
these factors to determine which soil treatment technologies are most
68
-------
GROUNDWATER'
DRY SOIL
PARTICLES
PORE
SPACES
WET SOIL
PARTICLES
UNSATURATED
ZONE
CAPILARY2ONE
SATURATED
ZONE
BEDROCK
PETROLEUM
VAPORS
WPORE
SPACES
RESDUAL
PETROLEUM
TRAPPED
BETWEEN
PARTICLES
PETROLEUM
DISSOLVED
WSOIL
MOISTURE
CLEAN SOIL
SOIL CONTAMINATED BY
PETROLEUM RELEASE
Figure 2. Representation of Three Different Phases in which
Petroleum can be Found in Unsaturated Zone
69
-------
likely to be effective. For this paper, CSFs are presented for soil
venting and biorestoration, two soil treatment technologies used at leaking
UST sites. Note that the CSFs are not equally important. Depending on
site conditions, it is likely that some CSFs will be much more important
than others. It is not possible to prioritize the CSFs for all conditions
at all sites.
Final selection of soil treatment technology will likely be based on
other criteria, however, such as what cleanup criteria are being used and
how much time is available to clean up the site. In urgent situations,
such as where a municipal water supply well is threatened, the soil would
likely be excavated; excavation can be undertaken quickly. If the cleanup
is not urgent, then other technologies such as biorestoration (which can
take as long as a year to be effective) might be preferred. Certain soil
treatment technologies may not be able to achieve the cleanup goals if the
concentrations being used for cleanup criteria are set too low.
Performance Monitoring and Follow-up Measurements
After a technology (or technologies) is selected and installed, it is
important to track its performance and monitor its effectiveness during
cleanup activities. Monitoring and follow-up are essential because of the
uncertainty surrounding subsurface conditions. If a technology's
performance is poor, it may be necessary to reexamine the data collected
during the site investigation and their inherent assumptions. Poor cleanup
performance could be attributed to a misinterpreted or incomplete site
investigation which resulted in an incorrect selection of the technology.
This feedback loop is an important step in the entire cleanup process.
RESULTS OF EFFORT
Contaminant Location and Migration
The rate and degree of partitioning of the residual liquid into the
vapor phase and dissolved phase depend on site-specific and
contaminant-specific factors, as well as time. A contaminant's volatility
(as measured by vapor pressure) is an indicator of how easily it will move
into the air. A contaminant's solubility is a measure of how easily it
will dissolve in water (including pore water).
Vapor analysis and soil sampling can be undertaken to determine in
which phase most of the contamination resides. Typically, soil is analyzed
for hydrocarbon concentrations of bulk liquid, while soil gas is sampled
for evidence of hydrocarbon vapors. General "rules-of-thumb" for
determining the phase which contains the contamination are presented in
Table 1. Vapor sampling at the site provides information into these rules
of thumb, and later, information that can be used to evaluate the soil
treatment technologies.
The final step in a site investigation focuses on the mobility of the
contaminants. Knowing where contaminants are likely to move, and how
likely they are to move, provides insight into how mobile the contaminants
are. The success of a corrective action plan depends on the contaminants'
mobilization. Mobilizing contaminants means being able to move
contaminants from one phase into another that can be more directly removed
70
-------
TABLE 1
RULES-OF-THUMB* FOR DETERMINING IN
WHICH PHASE CONTAMINATION CAN BE FOUND
Evidence of residual liquid contamination;
High concentrations (>1% by weight) of contaminants in
several soil analyses; (i.e., petroleum makes up >1% of
the weight of soil sample);
High concentrations (>10% by volume) of pure chemical
vapor density in several soil gas analyses (i.e.,
contaminant vapors are above 100,000 parts per million).
Evidence of contaminant vapors;
- Presence of residual liquid contaminants;
Significant concentrations in several soil gas analyses.
Evidence of pore-water contamination;
Significant concentrations of contaminants in several
analyses of pore water or groundwater;
Presence of residual liquid contaminants and a
significant soil moisture content.
^Sampling is required to apply these rules of thumb.
71
-------
with a given treatment technology. For example, soil venting works by
causing a disequilibrium between phases. An equilibrium is typically
established among all three phases when a petroleum release occurs.
Removing contaminant-saturated air with a soil venting system causes an
equilibrium shift. As clean (non-contaminated) air replaces the
contaminant-saturated vapors that are removed, contaminants remaining as
residual liquid will volatilize into the fresh air, seeking to establish
equilibrium. As the process continues, more and more contaminant in the
residual liquid will be "mobilized" into the vapor phase, where it can be
captured by the soil venting system.
The migration of petroleum hydrocarbons into and out of the two
dominant phases, as residual liquid and as vapors, is governed by specific
chemical and environmental factors. Chemical and environmental factors
that can influence a liquid's mobility in the unsaturated zone are listed
in Table 2 (grouped as soil-related and contaminant-related). In Table 2,
items are listed with both qualitative descriptors (high, medium, and low),
and corresponding quantitative ranges of values. Although the quantifica-
tion ranges are somewhat subjective, they have been arranged so that the
right-hand column indicates "high mobility" and the left-hand column "low
mobility."
Using Table 2 to evaluate conditions at a site of interest, it is
possible to get an understanding of the relative mobility of liquid
contaminants at the site. If the preponderance of factors at a site fall
in the right-hand columns of Table 2, liquid contaminants would likely be
more mobile and likely to migrate than if most factors matched those in the
left hand column. Where site-specific values are unavailable, default
values can be substituted as appropriate. Additional work is needed to
prioritize or rank the CSFs because in same cases one CSF could override a
preponderance of other CSFs on one side of the table or the other. Work
continues to refine and improve the tables and to make them as usable as
possible.
Vapors are generally mobile in
mobility greatly depends on the air-
porosity less that portion filled by
Several other factors also influence
zone as listed in Table 3. Table 3
i.e., by comparing parameters from a
listed in the table, it is possible
mobility of vapor phase contaminants
the unsaturated zone. The degree of
filled porosity of the soil (the total
water or liquid contaminants).
vapor transport in the unsaturated
is structured similarly to Table 2,
given site of interest with those
to get an understanding of the relative
at that site.
Contaminant vapors may be mobilized (and subsequently removed) by
several natural or induced processes or driving forces. These include:
• Bulk transport due to pressure gradients (e.g., from vacuum
extraction wells);
• Bulk transport due to vapor density gradients (which could result,
for example, if the contaminant vapor has a significantly
different density than air or from temperature gradients);
72
-------
TABLE 2
FACTORS TO EVALUATE THE EXTENT
OF MIGRATION OF LIQUID CONTAMINANTS
FACTOR
UNITS
SITE OF
INTEREST
INCREASING MIGRATION ^
RELEASE RELATED
• Tim* Sine* Last Release
Months
SITE- RELATED
• Hydraulic Conductivity
• Soil Porosity
• Soil Surfsos AIM
• Liquid Contaminant Content
• Sell Temperature
• Roek Fractures
• Water Content
CONTAMINANT- RELATED
• Uquld Viscosity
• Uquld Density
cm/sec
%
ma/g
%
«fc
—
%
cPofce
3
g/cfn
Long
(>12)
o
Medium
(1-12)
O
Short
(<1 )
O
Low
(<10's)
O
Low
(<10)
O
High
(>50)
O
Low
(<10)
0
Lew
(«10)
O
Absent
O
Wgh
(>30)
O
Medium
(10'5-10'9)
O
Medium
(10-30)
0
Medium
(5-50)
O
Medium
(10-30)
O
Medium
(10-20)
O
O
Medium
(10-30)
O
High
(>20)
O
Lew
<«D
O
Medium
(2-20)
O
Medium
(1-2)
O
High
(>10'»)
O
High
(>30)
0
Low
(«5)
0
High
(>30)
O
High
(>20)
O
Present
O
Low
(<10)
O
Low
(<2)
0
High
(>2)
O
73
-------
TABLE 3. FACTORS TO EVALUATE THE EXTENT OF
MIGRATION OF CONTAMINANT VAPORS
FACTOR
UNITS
SITE OF
INTEREST
INCREASING MIGRATION ^
•
'SITE -RELATED
• Air Filled PoroaKy
• Total Poreatty*
• Water Content*
• Depth Below Surface
%
%
%
nwtorc
Low
(<10)
o
Low
(<10)
O
High
(>30)
O
Deep
(>10)
O
CONTAMINANT- RELATED
• Vapor Density
g/m
Low
(<50)
O
* the total porosity bsc that fraction rifled wtth water equals th« *lr flltod poroclty
Medium
(10-30)
0
Medium
(10-30)
O
Medium
(10-30)
O
Medium
(2-10)
O
Medium
(50-500)
O
High
(>30)
O
High
(>30)
0
Low
(<10)
O
ShaBow
(<2)
O
High
(>500)
0
74
-------
• Sweep flow due to the in-situ generation of gases or vapors (e.g.,
vapors volatilizing from liquid contaminant or gases generated by
microbial biodegradation of contaminants);
• Molecular diffusion due to concentration gradients.
The current scientific understanding of these processes spans the
range from well-documented to rudimentary and hypothetical. Molecular
diffusion is, perhaps, the best understood and the easiest to address
experimentally and, hence empirically. However, in some circumstances it
may not be the most important process governing vapor mobility.
A soil with medium to high water-filled porosity will tend to
immobilize vapors; however, if one applies a vacuum extraction system,—
inducing soil gas flow — the soil will tend to dry out and this will
increase vapor mobility over the treatment period.
Technology Selection
The focus of the approach thus far is developing an understanding of
what was released from the UST, where most of the contamination resides
(i.e., in which phase most of the contamination exists), and how likely the
contamination is to migrate (i.e., move into and out of other phases).
With the site investigation completed, the next step becomes one of
selecting a treatment technology that will be effective given the
conclusions drawn from the site investigation. Many of the parameters
determined during the site investigation serve as indicators as to how well
a given soil treatment technology will perform. These indicators of likely
performance are called CSFs.
Tables 4 and 5 show the critical success factors for soil venting and
biorestoration. The three right-hand columns of each table provide values
for the CSFs that suggest whether a technology is "less likely," "somewhat
likely," or "more likely" to be effective. A column is provided for the
user to write down the values for the CSFs at the site of interest. These
values can then be compared with the values in the right hand column, and
the corresponding circles checked off or darkened. When completed, these
tables provide valuable insight as to which technologies are likely to be
effective and, more important, which are not. A brief description of each
of the technologies and their related CSFs is provided below.
Soil Venting
Soil venting is a general term that refers to any technique that
removes contaminant vapors from the unsaturated zone. Venting may occur
passively (with no energy input) or actively. Passive venting, which is
often used at sanitary landfills for methane gas removal, consists of
perforated pipes sunk into the contaminated area that provide an easy path
to the atmosphere. These vents sometimes have a wind-driven turbine at the
outlet to provide a slight draft.
More effective is active venting, where a pressure gradient is induced
to move vapors through the soil. Most common is vacuum extraction
technology, where extraction wells are placed near the release site and a
75
-------
TABLE 4.
WORKSHEET FOR EVALUATING THE FEASIBILITY OF
SOIL VENTING BEING EFFECTIVE AT YOUR SITE
CRITICAL SUCCESS
FACTOR
UNITS
SITE OF
INTEREST
SUCCESS
LESS
LIKELY
SUCCESS
SOMEWHAT
LIKELY
SITE RELATED
• Dominant
Contaminant Phase
• Soil Temperature
• Soil Hydraulic
Conductivity
• Moisture Content
* Geological
Conditions
• Soil Sorptlon Capacity
• Surface Area
• Depth to around water
Phase
C
cm/tec.
%
—
cma/g
nt
Sorbed to soil
o
Low
(<10)
0
Low
(< ID"5)
O
Moist
(>0.3)
0
Heterogeneous
O
High
(>50)
O
Low
(<3)
O
Uquld
0
Medium
(1(T- 20)
O
Medium
(10-4-10^
O
Moderate
(0.1 to 0.3)
O
o
o
Medium
(3-15)
O
SUCCESS
MORE
UKELY
Vapor or
Liquid
0
high
(>20)
O
high
(>10^
O
Dry
(<0.1)
o
Homogeneous
O
Low
(1S)
0
CONTAMINANT- RELATED
• Vapor Pressure
• Solubility
mmHg
mg/L
OTHER CONSIDERATIONS
• Co«t to tram S 1 5 to 160 p*r cubic yird.
• dp** 01 nmeving rouMnoi or oKtom.
• Air «nuiar* wii lk*y n*M to M mitt wBi QAC.
Low
(<10)
O
HlQn
(>1000)
O
Medium
(1010100)
O
Medium
(100-1000)
O
a tr— i-
r*Qn
(>100)
O
Low
(<100)
O
• •TiMmnt«nl»dar*on-ifti
• CM tnuH b« «Mn e MBU «viaHeni MOM* ««pan
• ClMnjp MM «M n m t» Mcmnogy * net
^propn»«<»n«n«m»fB»neyf»«ar»«l»n»«o«a
76
-------
TABLE 5.
WORKSHEET FOR EVALUATING THE FEASIBILITY
OF BIORESTORATION BEING EFFECTIVE AT YOUR SITE
CRITICAL SUCCESS
FACTOR
RELEASE • RELATED
• Tim* Sine* Release
SITE RELATED
• Dominant
Contaminant Phasa
• Soil Temperature
• Soil Hydraulic
Conductivity
• SollpH
• Molatura Content
CONTAMINANT- RELATED
• Solubility
• Biodagradablllty
• Raf rectory Index
• Fual Typa
UNITS
Months
SITE OF
INTEREST
SUCCESS
LESS
LIKELY
Short
<8)
O
o«y
«y>
Low
«100,
Low
(<001)
No. 6 Fual OH
(Heavy)
O
Vapor
O
Medium
(10». 20*C)
O
Medium
(10-5.10^
O
O
Moderate
(0.1 to 0.3)
Medkim
(100 to 1000)
O
(0.01 to 0.1)
No. 2 Fuel Oil
(Medium)
O
Long
(>12)
O
Dissolved
O
Ngh
(>20*C)
O
Hgh
(>io-»)
0
(6-8)
O
Moist
<1)3)
ttgh
(>1000)
Hgh
(>ai,
Gasoline
(Light)
O
OTHER CONSIDERATIONS
• Cost is from $60 to $125 per cubic yard.
• Completely destroys contaminants under optimal conditions
• Effectiveness varies depending on subsurface conditions
• Biologic systems subject to upset
• Public opinion sometimes against putting more chemicals in ground
• Difficult to monitor effectiveness
• Minimizes heath risk by keeping contaminarte in ground and on ste
• Takes long time to wortt— not for emergency response
77
-------
vacuum is applied to the wells. The soil gas is drawn through the soil to
the extraction well and brought to the surface.
Venting removes contaminants in the vapor phase, but also affects to
a limited extent residual liquid contaminants. Dissolved contaminants are
unaffected. Hydrocarbons typically are found in all three phases, and an
equilibrium is established, with a certain fraction existing in each phase.
The portion found in each phase depends on both the particular compound and
the local conditions. If conditions change, the equilibrium will shift,
and contaminants will transfer between phases to re-establish an
equilibrium. ".
The success of a vacuum extraction program depends both on the
properties of the contaminants and the properties of the soil. Compounds
with high Henry's Law constants will be removed to a greater degree by
vacuum extraction than the others, so this technology is more effective for
volatile compounds.
The water solubility of each contaminant will also affect the success
of venting, although this factor is relatively less important than those
listed above. Highly soluble compounds will tend to exist predominantly
dissolved in pore water, with less in the vapor phase. Vacuum extraction
tends to dry out the soil, however, and over time dissolved contaminants
will likely volatilize and be removed. Work continues on ranking the CSFs
to improve the usability of the tables.
Soil properties also greatly influence the success of soil venting.
Table 4 lists several soil properties to consider. Most important is the
soil hydraulic conductivity. Soil with low permeability, such as clay,
restrict the movement of vapors through the soil and towards wells.
Contaminants can still be removed from low permeability soil by soil
venting, but the process requires more closely spaced wells or a greater
vacuum. Most soils have preferential flow paths that are responsible for
most of the soil's permeability. These flow paths, which result from
things such as root intrusions, prevent the vapors from coming into
intimate contact with all of the contaminated soil, thus decreasing the
effectiveness of the technique.
Other important properties include soil temperature and moisture
content. The ambient temperature of the soil has a strong effect on the
volatility of the contaminant. As temperature rises, vapor pressure and
Henry's Law constant rise dramatically. For this reason, soil venting
would be expected to be more successful in areas where soil temperature is
high. In some cases, air is heated prior to injection to raise the
temperature of the soil and increase volatilization.
The moisture content has two effects on the soil. First, because soil
with a high water content has relatively less air-filled porosity, higher
water content leads to a lower air permeability and therefore a lower
removal rate. Second, pore water can absorb (dissolve) contaminants from
the vapor phase, which serves to retard the removal of contaminant vapors.
This is especially true of contaminants with low vapor pressures and low
Henry's Law constants. Dry soil is thus better suited to in-situ stripping
than wet soil. The vacuum extraction process tends to dry out the soilj
78
-------
over time the air permeability will increase and the dissolved contaminants
will volatilize, both of which tend to increase the degree of removal.
Carbon content is related to vapor pressure and can serve as a
substitute indicator. Compounds with lower carbon content (C, to C •) can
be vented more effectively than compounds with high carbon content (C0 to
C16). 9
Once Table 4 has been completed, other factors, such as cost, must be
included in the evaluation before making a final selection of a technology.
In general, soil venting is a relatively inexpensive technique compared
with other alternatives, especially when large volumes of soil must be
treated. The capital costs of venting consist basically of the extraction
and monitoring well construction, one or more blowers and housing, pipes,
valves, fitting, and other hardware, and electrical instrumentation.
Operations and maintenance costs consist of labor, power, maintenance, and
monitoring. Venting wells constructed of two-inch diameter slotted PVC
pipe cost approximately $20 per linear foot (0.3 m) for a twenty foot
depth. Vacuum pump sizing depends on local soil conditions and the volume
to be treated. Operating costs vary depending on time of operation and
local utility rates. The actual costs of soil venting at a Florida site
were estimated to be $106,000 (capital) and $68,000 (annual O&M). Air
treatment, if necessary, would have more than double costs. These figures
correspond to roughly $20 to $60/yd ($25 to $78/m3). Air emission control
via GAC is usually assumed to double the total capital cost of the cleanup.
Soil venting programs are relatively easy to implement and may be installed
and started in two to four weeks. This time is devoted to determining the
extent of contamination, designing the system, acquiring pumps and piping,
and installing the equipment.
A venting program will typically be operated for six to twelve months.
The removal rate is usually highest at the beginning of the program (once
the vacuum is established) and falls off after the most volatile
contaminants are removed. Volatilization from dissolved and sprbed
contaminants then becomes rate-limiting, and the system's effectiveness may
decline dramatically.
Biorestoration
Biorestoration is the process of adding nutrients to enhance the
natural biodegradation processes of soil microbes. It may also involve the
addition of specially-adapted microbes to the subsurface, but this is not a
common procedure. Experience has shown that at sites where the release
occurred long ago, native soil microbe populations are very large, and the
use of introduced microorganisms should not be necessary. Indigenous
microbes are expected to be as efficient at degradation as
specially-acclimated microorganisms. Acclimated microorganisms should be
thought of only as a convenient method of quickly increasing microbial
populations. Specially-acclimated microorganisms must compete with native
populations and be able to move from point of injection to the location of
contamination, and maintain selectivity for the contaminant of interest.
Oxygen is the single most important ingredient in a successful
biorestoration program. Although biodegradation may continue to occur
anaerobically, the lack of oxygen severely limits the rate of cleanup.
79
-------
Oxygen may be introduced by pumping air into the unsaturated zone, e.g.,
soil venting. In addition to oxygen, soil microbes also require
macronutrients (nitrogen and phosphorous) to survive and prosper. These
nutrients are typically added in order to facilitate biodegradation.
by
Slightly alkaline soil pH is optimum for biodegradation, but anything
in the range of 6.0-8.0 is considered acceptable. Redox potential is
important for various anaerobic microbes. The ratio of oxidized materials
to reduced materials in soil establishes an electrical potential, called
the redox potential. The temperature of the soil environment will also
affect the rate of degradation. Warmer temperatures generally result in
higher rates of degradation. While biodegradation has been shown to occur
over a wide temperature range, the range of 20° to 35 °C seems optimal.
Also, microbes generally have a low tolerance for severe temperature
changes .
A knowledge of the hydraulic conductivity of the soil and the site
hydrogeology in general is extremely important in assessing the feasibility
of biorestoration for a particular site. Even when all other factors are
positive, in-situ biorestoration will not be successful if a low hydraulic
conductivity prevents the added nutrients and oxygen from contacting the
zone of contamination. The residence time should be short enough so that
the oxygen concentration is sufficient throughout the site for microbes to
degrade all of the organic compounds. Also, the geochemistry of the
subsurface could inhibit adequate mixing if reactions (such as metal oxide
precipitation) clog the soil. The soil microbes themselves may clog the
soil and decrease the hydraulic conductivity.
Table 5 lists the critical success factors for biorestoration. By
comparing the parameters at the site of interest with those in this table,
a general understanding can be obtained of the suitability of
biorestoration at that site. A preponderance of CSFs that match the
right-most column would indicate that biorestoration is likely to be
effective at that site.
As with soil venting, several other considerations must be included in
the evaluation before a final selection can be made.
The costs of biorestoration for the unsaturated zone vary widely and
are difficult to quantify and compare. Also, most reported costs refer to
cleaning up groundwater rather than the unsaturated zone. One estimate
gave costs of $60 to $123 per cubic yard ($78 to $160 per cubic meter).
Unit costs for larger volumes are generally lower due to economies of
scale.
A biorestoration program can be set up relatively quickly, but it may
take several months for the microbes to become adjusted and start
significant degradation if the contaminant release is recent, or if
non-indigenous microorganisms are used. The system may need to be
"fine-tuned" (i.e., varying the levels of oxygen and nutrients added) to
operate efficiently. The start of a biorestoration program may be delayed
due to the drilling of injection and extraction wells, the design and
procurement of construction of the oxygenation equipment, and the need for
injection permits. Other important factors are listed at the bottom of
Table 5.
80
-------
SUMMARY
Table 6 presents a summary of the critical success factors for several
treatment methods and excavation. Table 6 includes only the objective
(scientific) factors that could affect the choice of technologies at a
specific site. Economic, political, regulatory, and other potentially
controversial factors often are as important as these objective factors.
Therefore, this summary table is useful for direct comparison of the
technologies on the grounds of technical feasibility only. When potential
treatment technologies have been narrowed to those that appear technically
feasible, other considerations (cost, public perception, etc.) will likely
affect the final selection.
The intent has been to provide an approach for selecting a soil
treatment that is based on scientific understanding to the extent possible
yet recognizes the uncertainties associated with cleaning up the
subsurface. Several limitations should be noted when using the approach
proposed herein:
• Not for Emergency Response - It is assumed that all necessary
emergency responses have been taken, that the source of the
release (e.g., tank or supply line) has been identified and
repaired, and that proper notification of government agencies
(local, state and federal) has taken place.
• Unsaturated Zone Coverage Only - This paper addresses site
investigation and corrective action for contamination in the
unsaturated zone only. Guidance is not provided for contamination
of the saturated zone. In some instances, an assessment of the
unsaturated zone might proceed regardless of potential or actual
groundwater contamination, but it should not be assumed that the
groundwater is uncontaminated. The presence of a floating
contaminant layer on the water table, or a contaminant plume in
the groundwater, may ultimately affect the selection of the
unsaturated zone corrective action or actions. Integrated
guidance for site assessment, corrective action, and evaluation
for both the unsaturated and saturated zones contamination is not
addressed in this document.
• Focus on Petroleum Hydrocarbons as Contaminants - Because
petroleum products comprise by far the greatest percentage of
materials stored in USTs, most assumptions made in this paper are
biased toward these materials. Special focus is given to
gasoline.
The proposed approach outlined in this paper shows promise for the
future. EPA, recognizing some of the inherent weaknesses in the state of
knowledge regarding soil treatment, continues to pursue research aimed at
removing some of these weaknesses and uncertainties. EPA's Office of
Underground Storage Tanks and Office of Research and Development are moving
forward with research projects and studies that focus on increasing the
understanding of how soil treatment can be made more effective and
efficient.
81
-------
Z
O
t-
>
u
X
UJ
O (/)
Ij EC
zffl
SOIL
FLUSHING
BIO-
STORATION
c
A
tn
1 CRITICAL SUCCESS
FACTORS
1 RELEASE -RELATED
1
I
1
f?
1
1
1
1 TIME SINCE LAST RELEASE
8
•j-
H
|
1
1
1
i
1 VOLUME OF SPILL
8
t
1
1
1
1
*»,
I
"E
1 VOLUME SOIL CONTAMINATED
e
L
t
i
|
o
2,
i
o
2.
I
o
2,
I
1
1 HYDRAULIC CONDUCTIVITY
1
1
1
i
i
p
I SOIL TEMPERATURE
1
1
!
i
§
t
§
I SOIL SORPTKJN CAPACITY
1
1
1
Mgh(>30)
S
v
1
*
I SOIL MOISTURE CONTENT
1
1
1
p
p
1
5
i
i
i
i
i
o
\
1
Is
IJ
UJ
i
"v
1
t
1
t
*
1 SOIL CARBON CONTENT
J_l
V
«
ff
2.
1
i
?
i
i
M
UJ
£
§
J
: 5
'
I
1
1 PHOXtMTiY TO LANDFILL
1
Htgh(>1000)
1
i
(
}
i
1 CONTAMINANT- RELATED
CONTAMINANT SOLUBILITY
1
1
6-
^
1
I
I
i
i
1
1
1
WBh(>0.1)
^
i
\
i
1
v
I
(7
^
1
i
1
1
(7
*
i
i
*
*
i
•
i
i
i
i
i
UJ
o
z
DC
2
CC
S FOR OPTIMUM
SS FACTOR!
UJ
8
F CRITICAL SI
O
c
<
s
1 a
i
*
82
-------
SUMMARY OF THE RESULTS OF EPA'S EVALUATION
OF VOLUMETRIC LEAK DETECTION METHODS
by: Joseph W. Maresca, Jr., James W. Starr, and Robert D. Roach
Vista Research, Inc.
100 View Street
P.O. Box 998
Mountain View, CA 94042
and , :
John S. Farlow and Robert W. Hillger
U.S. Environmental Protection Agency
Releases Control Branch
Risk Reduction Engineering Laboratory
Edison, NJ 08837 "
ABSTRACT
A United States Environmental Protection Agency (EPA) research program evaluated the
current performance of commercially available volumetric test methods for the detection of '
small leaks in underground gasoline storage tanks. The evaluations were performed at the EPA
Risk Reduction Engineering Laboratory's Underground Storage Tank Test Apparatus in Edison,
New Jersey.
The methodology used for evaluation made it possible to determine and resolve most of
the technological and engineering issues associated with volumetric leak detection, as well as to
define the current practice of commercially available test methods. The approach used
(1) experimentally validated models of the important sources of ambient noise that affect vol-
ume changes in nonleaking and leaking tanks, (2) a large database of product-temperature
changes that result from the delivery of product to a tank at a different temperature than the
product in the tank, and (3) a mathematical model of each test method to estimate the perform-
ance of that method. The test-method model includes the instrumentation noise, the configura-
tion of the temperature sensors, the test protocol, the data analysis algorithms, and the detection
criterion.
Twenty-five commercially available volumetric leak detection systems were evaluated.
The leak rate measurable by these systems ranged from 0.26 to 6.97 L/h (0.07 to L84 gal/h),
with a probability of detection of 0.95 and a probability of false alarm of 6.05. Five methods
achieved a performance less than 0.57 L/h (0.15 gal/h). Only one method was able to detect
leaks less than 0.57 L/h (0.15 gal/h) if the probability of detection was increased to 0.99 and
the probability of false alarm was decreased to 0.01. The measurable leak rates ranged from
0.47 to 12.95 L/h (0.12 to 3.42 gal/h) with these more stringent detection and false alarm
parameters.
The performance of the methods evaluated was primarily limited by test protocol, opera-
tional sensor configuration, data analysis, and calibration, rather than by hardware. The exper-
imental analysis and model calculations suggested that substantial performance improvements
could be realized by making only procedural changes. With modifications, it is estimated that
the majority of the methods should be able to achieve a probability of detection of 0.99 and a
probability of false alarm of 0.01 for leak rates between 0.19 L/h and 0.57 L/h (0.15 gal/h).
83
-------
INTRODUCTION
Leaking underground storage tank systems represent a serious environmental threat.
There are over two million underground storage tank systems in the United States. Estimates of
the fraction of these tank systems that may be leaking range from 10 to 25%. Records from
past release incidents indicate that, without the use of release detection, a release can become
substantial before it is detected (1,2).
There are many commercially available methods for detecting leaks in underground stor-
age tanks. Those which are the most widely used in the petroleum industry are in a category
called volumetric tank tests (also known as "precision," "tank tightness," or "tank integrity" tests).
The premise of a volumetric tank test, and hence its name, is that any change in the volume of
fluid within a tank can be interpreted as a leak. Detection of these leaks is difficult because
there are many physical mechanisms which produce volume changes that can be mistaken for
leaks. Most of the volumetric tanks tests on the market today claim the ability to detect leaks
as small as 0.19 L/h (0.05 gal/h). (This is the "practice" recommended in the National Fire
Protection Association (NFPA) Pamphlet 329 (3) for volumetric tests in tanks less than 47,316 L
(12,499 gal) in capacity.) These volumetric tank tests, however, do not specify the reliability of
their test results in terms of probability of detection (PD) and probability of false alarm (PFA)
against this 0.19-L/h leak rate. The 1984 Hazardous and Solid Waste Amendments to the
Resource Conservation and Recovery Act of 1976 charged the United States Environmental Pro-
tection Agency (EPA) with developing regulations for the detection of releases from under-
ground storage tanks. The new regulations, released in September 1988, state that all volumetric
tank test methods must have the capability of detecting leaks as small as 0.38 L/h (0.1 gal/h)
with a PD of 0.95 and a PFA of 0.05. These are only minimum standards, and the tank
owner/operator may want better protection against the possibility of a testing mistake.
Development of technically sound and defensible regulations required that both the threat
to the environment and the technological limits of release detection be known. The threat to
the environment is extremely difficult to define because the amount of petroleum that is haz-
ardous to the environment is site specific.
A performance standard that is based on the current technology will minimize the uncon-
trolled release of petroleum product. Unfortunately, the data required to formulate a realistic
regulatory policy were incomplete or nonexistent before the study described in this paper was
undertaken. While many commercial leak detection methods are available and can be used to
detect small releases, the performance of these methods was unknown. Very little evidence,
theoretical or experimental, had been provided by the manufacturers to support their perform-
ance claims. The limited evidence available prior to these evaluations suggested that the major-
ity of these methods were not reliably meeting these claims. In 1986, the Risk Reduction
Engineering Laboratory (RREL) (formerly the Hazardous Waste Engineering Research
Laboratory) of the EPA initiated an experimental research program to evaluate the current
practices. Participation in the program was voluntary, and the manufacturers of 25 commer-
cially available methods elected to participate.
84
-------
The specific objectives of the program were to produce the technical data necessary to
support the development of release detection regulations; to define the current practice of com-
mercially available systems; to make specific recommendations to improve the current practice;
and to provide technical information that would help users select suitable methods for testing
the integrity of underground storage tanks. This paper summarizes the results of the EPA
research program. A detailed description of the Edison evaluations is presented in a two-
volume EPA technical report (4). Volume I of the report contains the objectives and the chro-
nology of the experiments, a thorough explanation of the engineering principles underlying the
experiments, and a comprehensive analysis of the results. Volume II includes the individual
evaluation reports written for each of the 25 test methods.
VOLUMETRIC TEST METHODS
A volumetric test method measures the change in product volume that results from a leak
in the tank; a leak may manifest itself as a release of product from the tank or an inflow of
ground water into the tank. Figure 1 gives an overview of the test procedure used by most test
methods. These methods measure product level and product temperature. The product-level
and temperature data are converted to product volumes. The volume changes produced by
thermal changes of the produced are then subtracted from the product volume data. The result-
ing temperature-compensated flow rate is then compared to a predetermined flow rate for that
method called the threshold. If the flow rate exceeds the threshold, the tank is declared
leaking. If not, the tank is declared tight. The test procedure usually incorporates one or more
waiting periods after any addition of product to the tank to minimize the effects of structural
deformation and horizontal inhomogeneities in the product-temperature field. While the details
of the actual instrumentation, measurement protocols, data reduction and analysis algorithms,
and detection criteria differ from method to method, the testing approach is essentially the same
for all methods.
Volumetric tank tests can be divided into two categories. In the first, the tank is filled to
capacity, and in the second, the tank is partially filled. In filling a tank to capacity the opera-
tor does not stop until the level of the product reaches the fill tube (or a standpipe located
above grade); hence, the term "overfilled" is applied to these tests. Overfilled-tank tests can be
further categorized according to those conducted under nearly constant hydrostatic pressure (i.e.,
those with a nearly constant product level) and those conducted under variable hydrostatic pres-
sure (i.e., those with a fluctuating product level).
In a constant-level test product is added or removed in order to maintain a constant fluid
level in the tank's fill tube or standpipe. To conduct a successful test, it is necessary, once a
tank has been filled and then again after it has been topped off prior to testing, to observe a
waiting period long enough to ensure that the tank has expanded to its maximum capacity and
the temperature fluctuations have subsided. Then, if the fluid level is kept constant, the tank
will neither expand nor contract during the test, and measured volume changes will accurately
represent actual volume changes. Partially-filled-tank tests are generally considered constant-
level tests. Because the surface area of the product is spread across the width and length of the
tank, any level changes will be quite small, regardless of the size of the associated volume
85
-------
change. Unless product is added or removed during the test, causing the tank to deform, a
partially filled tank behaves in the same manner as an overfilled tank given that the appropriate
waiting periods are observed.
Grois Adjustment ol Product Level
(Product Delivery)
Wait for Temperature
and Structural Deformation
to Become Small
Equipment Setup
Fine Adjustmant of Product Level
(Topping the Tank)
Product Level and Temperature Measurement
Figure 1. Overview of volumetric tank test procedure.
In a variable-level test, the fluid level is allowed to fluctuate. When such a test is con-
ducted in an overfilled tank, the surface area of the product is extremely small ~ it is usually
limited to one or more small diameter openings (e.g., fill tube or vent pipe). Any volume
changes will thus be seen as significant height changes. Unless the deformation characteristics
of the tank being tested are known (as well as those of the backfill and surrounding soil) it is
not possible to distinguish between the volume changes due to a leak and those that normally
occur in a nonleaking tank. These deformation characteristics are not known at the time of the
test, and it is impractical to measure them. There is, consequently, a high risk of large errors
in variable-level tests (5).
HOW PERFORMANCE IS DEFINED
The accuracy of a test method in ascertaining whether the tank is leaking or not is what is
meant by the performance of the method. Performance is defined by the test method's PD and
Ppx for each leak rate that the method claims to be able to detect. The PD refers to the test's
chances of correctly identifying a leaking tank compared to its chances of failing to detect a
leak that is actually present. The PFA refers to a test's chances of declaring a leak when in fact
86
-------
none exists. The key to how well a test method performs is its ability to discriminate between
the volume changes produced by a leak (i.e., the signal) and other volume changes that normally
occur in a tank (i.e., the noise). The latter can either mask a leak, or mimic a leak and thus be
confused with one. The noise is the sum of the product-level or product-volume changes pro-
duced by the measurement system itself, by the environment, and by the operational practice.
The major sources of environmental noise are produced by product temperature changes, vapor
pockets, structural deformation of the tank system, evaporation and condensation from the
product surface and tank walls, and internal and surface waves. The noise field is particularly
large immediately after delivery of product to the tank or topping of the tank during a leak
detection test.
The performance of a detection system can only be determined once the fluctuation level
(product-level or product-volume changes) at the output of the measurement system is known
with and without the signal present. For any test method, the statistical fluctuation of the noise
is observed in the histogram of the volume-rate results created by plotting the measured volume
rates from a large number of tests conducted (1) over a wide range of conditions, (2) with many
systems on one or more nonleaking tanks, and (3) by many different operators. The histogram
indicates the probability that a particular volume rate will result from a test on a nonleaking
tank. The histogram of the noise is developed experimentally. The histogram of the signal-
plus-noise is usually developed from a model that indicates how the signal adds to the noise.
A complete specification of system performance requires a description of the PD and the
PFA at a defined leak rate and an estimate of the uncertainty of the PD and PFA. If, in addi-
tion, a frequency of testing is specified, then the limits of the threat to the environment, the
confidence with which these limits can be met, and the costs associated with mistakes in testing
can be defined.
The threshold is often confused with the leak rate to be detected. The EPA release
detection regulation requires volumetric test methods to have a minimum detectable leak rate of
0.38 L/h (0.1 gal/h). In order for a test method to meet this requirement, its threshold must be
less than 0.38 L/h. If the threshold is equal to the leak rate to be detected, the PD is only 0.50.
Choosing the right balance between the PD and PFA is a very difficult task. Missed
detections result in the release of product into the ground and the consequential contamination
of the nation's major source of potable water. False alarms lead to the expense of additional
testing and/or the repair or replacement of tanks that are not leaking. It is fair to expect tank
owners to interpret this balance in terms of financial considerations. The clean-up costs result-
ing from a missed detection must be weighed against the cost of unnecessary testing and repairs
resulting from a false alarm.
EVALUATION APPROACH
A three-step procedure was used to conduct the evaluations (9-12). The first step was to
develop and experimentally confirm models of the important sources of noise that control the
performance of each test method. If the total noise field is accurately modeled, the sum of the
volume contributions from each noise source will be equal to the product-level changes in a
nonleaking tank. As part of the modeling effort, two large databases, reflecting the different
product temperature conditions which could be experienced during field testing, was obtained to
87
-------
simulate a test performed after a delivery of approximately 15,000 L (4,000 gal) of product at
one temperature to a 30,000-L (8,000 gal) storage tank half-filled with product at another tem-
perature. Temperature data were collected for a range of temperature differences of -jJO°C
between new and existing product. One database corresponds to tests in a tank as the product
temperature naturally attempts to come into equilibrium with the surrounding backfill and
native soil, and the second to tests that deliberately mix or circulate the product in the tank.
The second step was to develop and validate, for each leak detection method, a model that
mathematically described it. The test-method model includes the precision and accuracy of the
instruments; the test protocol; the data collection, analysis and compensation algorithms; and the
detection criterion. The model, in turn, was validated in two steps. First, each manufacturer
was required to review the model for accuracy and to concur that it accurately represented the
method before the evaluation was allowed to continue; and second, the manufacturer was
required to participate in a three-day program of tank-test and calibration experiments at the
UST Test Apparatus. The manufacturer used his own testing crews and test equipment for the
three days of testing. Methods that were not operational at the time of the tests, or that were
different from those with which their respective manufacturers had concurred, were not evalu-
ated.
Finally, a performance estimate for each method was made by combining, in a simulation,
the test-method model approved by the manufacturer, the product-level measurements estimated
from the noise models, and the temperature database. The performance of a test method was
evaluated by repeatedly simulating the conduct of a tank test in order to develop a histogram of
the noise. Operational effects and deviations from the prescribed protocols during the three-
day field testing program were also examined and discussed.
0.2
I
I 0.1
8T
u_
0.0
-400
R3
1
0
Volume Rate - ml/h
400
Figure 2. Histogram of the noise generated for the five-thermistor test method.
Examples of the noise histogram and performance curves illustrative of the output gener-
ated for each test method evaluated are shown in Figures 2 and 3 for a hypothetical test method
which tests when product is near the top of the tank, using an array of five equally spaced,
88
-------
volumetrically weighted thermistors. It was assumed that the temperature and product-level
sensors used by this method had sufficient precision to measure ambient product-volume
changes that were less than 0.04 L/h (0.01 gal/h). The data were sampled once per minute and
the duration of the test was 1 h. The only source of noise considered in the simulation was
thermal expansion or contraction of the product.
I e-'°
1000 2000 MOO
Threshold • mtfh
•UOO -2000 -IOM 0 1000 2000 MOO
Threshold • ml/h
S.
o.oc o.io o.u
Probability ol False Alarm
O.OS 0.10 016
Probability at False Alarm
Figure 3. Examples of performance curves for the five-thermistor test method. (A) Pj> vs.
Threshold, (B) PPA vs. Threshold, (C) PD vs. PPAfor flow out of the tank and (D) PD vs. PPAfor
flow into the tank.
The performance is presented in three displays, the first of which is a plot (Figure 3A) of
the probability of detection versus detection threshold for a family of leak rates with flow into
and out of the tank (positive and negative volume rates). The second display is a plot (Figure
3B) of the probability of false alarm versus threshold. The third display shows the probability
of detection versus the probability of false alarm for a family of leak rates. Note that the third
display is separated into two plots, one for outflows (Figure 3C) and one for inflows (Figure
3D).
89
-------
Performance curves were generated that were based on the simulated noise and signal-
plus-noise histograms. For high levels of performance, the PD and PFA are estimated from the
tails of the histogram. With limited data, good estimates of the PD and PFA are sometimes
difficult to make. In this study, the performance estimates were typically based on 50 to 200
independent realizations of the manufacturer's test. To reduce the uncertainty in the perform-
ance estimates at the higher PDs and lower PFAs, an exponential curve was fit to the tails of the
histogram. The PD and the PFA obtained from the curve were used to estimate performance; an
estimate of the uncertainty of the PD and the PFA was also made.
UNDERGROUND STORAGE TANK (UST) TEST APPARATUS
The evaluations were performed by the Risk Reduction Engineering Laboratory (RREL)
at the EPA's UST Test Apparatus located in Edison, New Jersey. The Test Apparatus is envi-
ronmentally safe and was designed and built to evaluate the performance of in-tank leak detec-
tion systems. Construction was completed in August 1986. The Test Apparatus consists of two
2.43-m (8-ft)-diameter, 30,000-L (8,000-gal) underground storage tanks installed in a
pea-gravel backfill material; one is a steel tank coated with plastic, and the other is a fiberglass
tank. -Two above-ground tanks are used to heat or cool product for simulation of a delivery to
the underground tanks. With this combined apparatus, different product temperatures, product
levels, and leak rates can be generated and accurately measured. The apparatus can be and was
configured to investigate each source of noise experimentally. In November 1987, a pressurized
pipeline system was added to the Test Apparatus, but was not used in this evaluation.
To address the overall project objective, a set of data quality objectives was established at
the beginning of the program and was adhered to throughout the data collection (11). The data
quality objectives for this project were established to evaluate the 0.19-L/h (0.05-gal/h) per-
formance claim. The precision and accuracy of the product-level and temperature data col-
lected at the UST Test Apparatus were specified so as to evaluate the performance of each test
method at a leak rate of 0.19 L/h with a probability of detection of 0.95 and a probability of
false alarm of 0.001, a more stringent requirement than either the draft (PD of 0.99, PFA of
0.01) or the final (PD of 0.95, PFA of 0.05) EPA release detection standard. This requires that
the precison of the instruments used to measure temperature and product level and the accuracy
of the constants used to convert temperature and product level to volume must have a total
uncertainty of less than 0.04 L/h (0.01 gal/h) when the data are combined to estimate the
temperature-compensated volume rate. The UST Test Apparatus instrumentation, calibration
procedures, and data quality analyses after each test were designed to verify that the data were
meeting the data quality objectives.
RESULTS OF THE EDISON EVALUATION
The names of the 25 commercially available volumetric test methods that were evaluated
by the EPA at the UST Test Apparatus are presented, along with their manufacturers, in
Table 1. The performance of 19 of the 25 test methods is summarized in Table 2. Proper
interpretation of the quantitative performance estimates given in Table 2 for overfilled-tank-
test methods requires the use of the last column in Table 2. Table 3 lists the names of the
methods for which no quantitative estimates could be made and the reason.
90
-------
Table 1. PARTICIPANTS COMPLETING THE EPA VOLUMETRIC TEST METHOD EVAL-
UATION PROGRAM
Test Method Name
Test Method
Manufacturer
Telephone
Number
AES/Brockman Leak Detecting System
Ainlay Tank 'Tegrity Tester
Automatic Tank Monitor and Tester
(AUTAMAT)
Computerized VPLT Tank Leak Test-
ing
System
DWY Leak Sensor
EZYCHEK
Gasoline Tank Monitor (GTM)
Gilbarco Tank Monitor
Inductive Leak Detector 3100
INSTA-TEST
Leak Computer
Leak-O-Meter
LiquidManager
LMS-750
MCG-1100
Mooney Leak Detection System
OTEC Leak Sensor
PACE Leak Tester
Petro Tite
Portable Small Leak Detector (PSLD)
S.M.A.R.T.
Tank Auditor
Tank Monitoring Device (TMD-1)
Tank Sentry II
TLS-250 Tank Level Sensing System
Associated Environmental Systems
Soiltest, Inc.
Exxon Research and Engineering Co.
NDE Technology, Inc.
DWY Corp.
Horner Creative Products, Inc.
Tidel Systems
Gilbarco, Inc.
Sarasota Automations, Inc.
EASI, Inc.
Tank Audit, Inc.
Fluid Components, Inc.
Colt Industries
Pneumercator Co., Inc.
L & J Engineering, Inc.
The Mooney Equipment Co., Inc.
OTEC, Inc.
PACE (Petroleum Association for
Conservation of the Canadian Envi-
ronment)
Heath Consultants, Inc.
TankTech, Inc.
Michael & Associates of Columbia,
Inc.
Leak Detection Systems, Inc.
Pandel Instruments, Inc.
Core Laboratories, Inc.
Veeder-Root Co.
(805) 393-2212
(312) 869-5500
(201) 765-3786
(213) 212-5244
(715)
(517)
(214)
(919)
(813)
(219)
(619)
(619)
(813)
(516)
(312)
(504)
(715)
(416)
735-9520
684-7180
416-8222
292-3011
366-8770
239-7003
693-8277
744-6950
882-0663
293-8450
396-2600
282-6959
735-9520
298-1144
(617) 344-1400
(303) 757-7876
(803) 786-4192
(617) 740-1717
(214)660-1106
(512) 289-2673
(203) 527-7201
In Table 2, test methods are arranged by alphabetical order in three categories:
(1) partially filled-tank-test methods, (2) constant-level overfilled-tank-test methods, and
(3) variable-level overfilled-tank-test methods. The reason for this arrangement is that
performance is largely controlled by characteristics particular to each of these categories. For
example, vapor may be trapped in an overfilled-tank-test but not in a partially-filled-tank test;
evaporation and condensation will be an important source of error in a partially-filled-tank test
but not in an overfilled-tank test; in overfilled tank tests, the tank must be topped prior to the
test, an action that carries with it the risk of degrading the performance, while in a
partially-filled-tank test topping is not necessary.
The first column in Table 2 lists the name of each method. The second, third, and fourth
columns show the mean, standard deviation, and number of the simulated temperature-
compensated tank test data that were used to estimate performance. Using these data, perform-
ance curves were generated for each method. Two interesting results, shown in column 5, and
columns 6 and 7, were derived from these curves. The fifth column presents the
91
-------
Table 2. ESTIMATES OF THE PERFORMANCE OF 19 VOLUMETRIC TEST METHODS
PD & PFA Smallest Smallest
to Detect Detectable Detectable
Number a 0.38 L/h Leak Rate Leak Rate Waiting
of (0.10 for for Period
Standard Simulat gal/h) PD=0.9S PD=0.99 After
Test Method Name
Mean
(L/h)
Deviation ed Tests
(L/h)
Partially-Filled-Tank
Gasoline Tank Moni-
tor
(GTM)i
Gilbarco Tank Moni-
tor1
Inductive Leak
Detector
3100
Tank Sentry II
TLS-2501
0.105
0.016
0.055
-0.093
-0.016
0.408
0.075
1.012
0.154
0.142
Overfilled-Tank Test
Leak Computer
MCG-110
Petro Tite
0.005
0.206
0.002
0.096
0.119
0.209
13
59
45
23
46
Leak Rate PFA=0.05
(PD,PFA)
(L/h)
PFA=0.01
(L/h)
Topping
(h)
Test Methods
0.73,0.21
0.96,0.003
0.72,0.33
0.89,0.16
0.15,0.001
1.35
0.26
4.23
0.58
0.51
1.91
0.47
9.54
0.89
0.90
N/A
N/A
N/A
N/A
N/A
Methods/Nearly Constant Level
132
97
25
0.97,0.04
0.97,0.09
0.79,0.21
Overfilled-Tank Test Methods/Variable
AES/Brockman Leak
Detecting System2
Ainlay Tank 'Tegrity
Tester
Computerized VPLT
Tank Leak Testing
System
EZY CHEK
Leak-O-Meter
LiquidManager
Mooney Leak Detec-
tion
System
PACE Leak Tester
Portable Small Leak
Detector (PSLD)
S.M.A.R.T.
Tank Auditor
-0.167
0.076
0.023
0.048
-1.060
0.307
-0.266
0.143
-0.192
-0.033
1.048
0.910
0.470
0.230
0.184
2.072
0.168
0.551
0.810
0.871
0.366
1.107
112
284
99
399
231
79
196
245
135
81
207
0.45,0.34
0.50,0.31
0.66,0.19
0.86,0.15
0.57,0.49
0.80,0.14
0.47,0.38
0.37,0.32
0.63,0.32
0.58,0.32
0.57,0.43
0.32
0.36
0.80
Level
6.79
2.97
1.08
0.62
6.96
0.75
3.13
6.97
3.05
2.25
6.27
0.65
0.86
1.11
12.95
3.93
1.84
0.93
10.80
1.25
4.58
11.12
5.60
3.43
•12.52
Variable
Variable
Variable 3
0
2
0
1
0
0.75
Variable
0
0
0 or 12
0
1
2
S
These test methods were employed in a special precision test mode rather than in their
normal operating mode as automatic tank gauging systems (ATGS)
Data analysis algorithms for this method had to be modified in order to determine per-
formance.
5 to 8 min/1,000 gal of product in tank at high level and 0 h when product level is
dropped to low level for testing.
92
-------
performance that the method should be able to achieve in actual tank tests. This performance is
expressed in terms of the probability of detection and probability of false alarm for leak rates
as small as 0.38 L/h (0.1 gal/h), using the manufacturer's detection criterion as defined at the
time of the evaluation. (The majority of the manufactuers used a detection threshold of 0.19
L/h (0.05 gal/h) to determine whether the tank was leaking or not.) The sixth and seventh
columns present the potential performance of the method in terms of the smallest detectable
leak rate that might be achieved for a PD of 0.95 and a PFA of 0.05, and a PD of 0.99 and a PFA
of 0.01, respectively. This performance estimate does not employ the manufacturer's detection
threshold; instead, a threshold was selected which yields a probability of false alarm of 0.05 and
0.01, respectively. These quantitative performance estimates should not be used unless the
reader first understands what was used to generate them.
There were no quantitative performance estimates made for 6 of the 25 methods in the
evaluation program; they are listed in Table 3, along with the reason why no performance esti-
mate was made. The reason was one of three. In two cases, the manufacturer's test crew could
not perform a satisfactory tank test during the 72-hour period allotted to them during the field
tests at the UST Test Apparatus. In three cases, the data obtained during the field tests clearly
indicated that the method was not behaving as the manufacturer had said it would. In general,
these methods used an integrating temperature-compensation measurement approach, which had not
been experimentally validated adequately by the manufacturer prior to the evaluation. In the
final case, the Test Apparatus was not properly configured for all of the field tests, preventing
an adequate field test of the method; in this latter case, however, the temperature-compensation
scheme of this method had also not been experimentally validated prior to the evaluation.
Table 3. LIST OF TEST METHODS NOT EVALUATED
Test Method Name
Reason
Autamat
DWY Leak Sensor
Insta-Test
LMS-750
OTEC Leak Sensor
TMD-1
Operational principles could not be verified
Operational principles could not be verified
Did not successfully conduct a tank test
Operational principles could not be verified
Improper configuration of Test Apparatus
Did not successfully conduct a tank test
INTERPRETATION OF EVALUATION RESULTS
Three general remarks apply to nearly all of the methods evaluated. First, the majority of
methods exhibited a bias; the magnitude of this bias is evidenced in the size of the mean of the
data in Table 2. In general, the performance of methods with a bias can not be accurately
estimated, because unless the physical mechanisms producing the bias are known and can be
quantified (so that the bias can be removed), performance can change from test to test. The
bias was arbitrarily removed for the estimates presented in Table 2. If the bias is large (i.e., if
it represents a large percentage of the leak rate to be detected), the method should be consid-
ered suspect.
93
-------
Second, experimental estimates of the precision and accuracy of each method's instrumen-
tation were derived from data obtained from a calibration of the level and temperature mea-
surement systems; this calibration was done as part of the evaluation. It is assumed that the
instruments used in actual practice are accurately and routinely calibrated, and that the
precision and accuracy of these instruments are equal to or better than the precision and accu-
racy used in the performance estimate. It was observed during the evaluation field tests that the
calibration procedures that many of the manufacturers used or claimed to use were generally
inadequate, and that in many cases the instruments had never been satisfactorily calibrated at
all.
Third, deviations from the protocol alter performance; during these evaluations, perform-
ance was seen to improve as well as to suffer as a result of changes in protocol. In order to
make the performance estimates, therefore, it was assumed that the test protocol as given by the
manufacturer is always followed precisely. This implicitly assumes that only the best test crews
are used to execute a test, that is, the type of crew that participated in the evaluation program.
The remaining remarks apply specifically to test methods in certain categories. Not all
sources of error were included in the performance estimates presented in Table 2, and as a
consequence, the actual performance of a method may be poorer than the performance shown
here.
The effects of evaporation and condensation were not included in the estimates for meth-
ods that test in partially filled tanks. In general, these effects are small, but in some circum-
stances they can be large enough to cause testing errors.
The effects of a product delivery are included in all of the performance estimates for
methods that overfill the tank and maintain a constant level of product. However, the degrada-
tion in performance that results from topping the tank during an overfilled-tank test were not
included in the performance estimates, nor were the effects of any product-level changes that
are required before starting a test. The effects of topping are to produce spatial inhomogenei-
ties in the product-temperature field due to the addition of product at a different temperature
than exists in the tank and changes in tank volume due to structural deformation. Waiting
periods for temperature fluctuations and structural deformation to subside are usually incorpo-
rated into test protocols, but in general these waiting periods were found to be too short. The
short waiting periods shown in Table 2 suggest the magnitude of the problem. Thus, in actual
practice, performance could be significantly reduced in comparison to that presented in Table 2.
In all overfilled-tank tests the potential exists for trapping vapor in the top of the tank or
in its associated piping. The effects of trapped vapor were not included in the performance
estimates in Table 2 for either one of two reasons: most manufacturers claim to be able to
remove vapor before a test begins and without experimental evidence to the contrary, this claim
was accepted as true; and even if trapped vapor were included in the estimates, it would be
difficult to do this accurately, because the distribution of the volume of trapped vapor is
unknown. That vapor will be trapped, however, is almost inevitable, and the performance esti-
mates shown in Table 2 will be reduced if this vapor is not removed before a test.
94
-------
The same effects that have the potential to degrade the performance of constant-level
tests also impact the performance of variable-tests. In addition, when the product level within
the fill tube or standpipe is allowed to fluctuate during a test, it is nearly impossible to convert
product-level measurements to volume. Test methods that allow this situation to occur have a
basic flaw in protocol and should be considered suspect (4,5).
STATISTICAL SUMMARY OF PERFORMANCE RESULTS
The estimates of potential performance from Table 2 are summarized in Tables 4 and 5.
Table 4 summarizes the methods1 actual performance using the manufacturers' detection criteria.
Table 5 summarizes potential performance using a detection criterion that produces a probability
of false alarm of either 0.05 or 0.01. This performance is expressed as the smallest leak rate
than can be detected with a probability of 0.95 or 0.99. Table 5 suggests that the methods are
divided into two distinct levels of performance.
An estimate of potential performance based on the experimental and theoretical work per-
formed during the program is presented in Table 6. For many methods, as discussed in [4], a
significant increase in performance can be achieved by means of protocol changes alone. Of
course, the actual performance improvement would depend on the specific changes made by the
manufacturer.
Table 4. ESTIMATES OF MANUFACTURERS' PERFORMANCE IN TERMS OF PD and PFA
FOR DETECTION OF A LEAK RATE OF 0.38 L/h (0.1 gal/h) USING THE MANUFAC-
TURERS' DETECTION CRITERION
0.90 *
0.65 -
0.35 <
0.10-
PD
-------
The temptation to use only those methods that were ranked highest in this evaluation
should be avoided for two reasons. First, Table 6 suggests that with modifications most of the
methods should be able significantly increase their performance. Since more than a year has
elapsed since the evaluations were performed, many methods have made changes that should
improve their performance. Second, the quantitative estimates presented in Table 2 alone are
not sufficient to assess the performance of the method.
Table 6. ESTIMATE OF THE PERFORMANCE OF VOLUMETRIC TEST METHOD EVAL-
UATED AT THE UST TEST APPARATUS AFTER TWO LEVELS OF MODIFICATIONS
EXPRESSED IN TERMS OF THE SMALLEST LEAK RATE THAT CAN BE DETECTED
WITH A PD OF 0.99 AND A PFA OF 0.01
Number of Methods Having
This Detectable Leak Rate
Detectable Leak Rate
(L/h)
0.19 < LR < 0.57
0.57 < LR < 0.95
0.95 < LR < 1.32
Evaluation
Results
1
5
2
After Minor
Modification
6
13
After Protocol and
Equipment Modifi-
cations
12
7
SUMMARY
An important EPA-sponsored research program has been completed that has evaluated and
made estimates of the performance of commercially available volumetric leak detection methods
as they existed in the period March through July 1987. The performance estimates assumed that
the test procedures were implemented by competent test crews using calibrated equipment. For
each method evaluated, recommendations were made, as required, to improve performance.
This two-year project has determined and resolved key technological and engineering issues
associated with this general type of leak detection. The following objectives were accomplished:
(1) evaluation of the performance of 25 currently available volumetric systems for detection of
leaks in underground gasoline storage tanks; (2) generation of technical information important in
the development of EPA's underground storage tank regulations; (3) development of specific
recommendations that will allow manufacturers to improve the current practice of each method;
and (4) development of basic information to assist the test users in selecting a method that meets
EPA's new regulatory requirements for underground storage tanks. A summary of some of the
key conclusions of this research project are provided below.
Current performance is significantly less than claimed by most manufacturers. Of the 25
commercially available volumetric leak detection systems evaluated, most presently perform at a
level that is considerably lower than the common industry claim of 0.19 L/h (0.05 gal/h). There
are two reasons for this discrepancy between vendor claims and actual performance. First, in
almost all instances, the measurements made by EPA during this project appears to be the first
systematic evaluation of the test method. Second, the performance estimates were presented in
terms of a probability of detection and a probability of false alarm, a format that most man-
ufacturers have not previously used to quantitatively describe performance.
96
-------
EPA regulatory requirements for performance are met by five of the methods evaluated.
Many more should meet the requirements after modifications. By and large, the leak detection
systems evaluated were limited by protocol and operation practice rather than by hardware. In
general, such limitations can be overcome by rather modest modifications to testing procedures,
such as waiting 3 h after topping the tank and before starting a test; major equipment redesign
is not necessarily required.
Volumetric test methods should not be selected solely on the ranking contained in this
paper. Because, as discussed in [4], the performance of many test methods can be significantly
improved with only minor modifications, some low-performance methods are expected to meet
or exceed the level of high-performance methods in the immediate future. As a consequence,
and because many methods have already incorporated many of the recommended changes, the
current ranking of test method performance implied by this paper is not particularly significant
in the selection of reliable methods.
Tank testing is complex, but a high level of performance can be achieved if several key
principles are followed. Those systems that did well in this evaluation had adequate spatial
sampling of the vertical temperature profiles of the product in the tank; incorporated adequate
waiting periods after product delivery and/or topping the tank (in tests that overfill tanks) to
allow the tank deformation and the spatial inhomogeneities in the product temperature field to
become negligible; maintained a nearly constant hydrostatic pressure head during the test; made
an experimental estimate of the height-to-volume conversion factor; and used sound data analy-
sis algorithms and detection criteria. Performance of a test method suffered significantly when-
ever one of these aspects of testing was ignored or poorly implemented. In general, any method
will perform poorly and provide results that are difficult to interpret if it: (1) fails to maintain
a constant hydrostatic head during the test; (2) does not accurately estimate the height-to-
volume conversion factor; (3) tops the tank and begins to test almost immediately, or (4) waits
an insufficient period of time after product delivery before beginning the test. Most
manufacturers recognized the need to wait after product delivery, but they did not appear to
fully appreciate the magnitude of the degradation that occurs when the waiting period after
topping (in the overfilled-tank test methods) is not long enough.
Reliable tank testing takes time. The total time required for the methods evaluated at the
UST Test Apparatus to complete a reliable tank test, from delivery of product to removal of the
equipment from the testing site, is generally 12 to 24 h. The time required is controlled by the
waiting periods after product delivery or after topping the tank. The waiting periods can be
minimized by incorporating data analysis algorithms into the test protocol which identify the
point at which these two effects have become negligible.
Two inevitable outcomes of this research project will be (1) rapid improvements in per-
formance based on changes implemented by manufacturers and (2) increases in test users' and
regulators' expectations concerning verification of future performance claims for all volumetric
methods. Manufacturers of many of the methods that were evaluated by RREL have already
begun to make the changes necessary to improve their systems' performance and to verify the
new performance claims.
97
-------
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
U.S. Environmental Protection Agency. Part 280 —- Underground Storage Tanks; Pro-
posed Rules. Federal Register, Vol. 52, No. 74, 1987.
U.S. Environmental Protection Agency. Part 280 — Technical Standards and Corrective
Action Requirements for Owners and Operators of Underground Storage Tanks. Federal
Register, Vol. 53, No. 185, 1988.
National Fire Protection Association. Underground Leakage of Flammable and Combus-
tible Liquids. NFPA Pamphlet 329, National Fire Protection Association, Quincy, Massa-
chusetts, 1987.
Roach, Robert D., James W. Starr, and Joseph W. Maresca, Jr. Evaluation of Volumetric
Leak Detection Methods for Underground Fuel Storage Tanks. Final Report, Vol. I,
EPA/600/2-88/068a and Vol. II, EPA/600/2-88/068b. U. S. Environmental Protection
Agency, Cincinnati, Ohio, 1988.
Roach, Robert D., James W. Starr, Christopher P. Wilson, Daniel Naar, Joseph W.
Maresca, Jr., and John S. Farlow. Discovery of a New Source of Error in Tightness Tests
on an Overfilled Tank. In: Proceedings of the Fourteenth Annual Research Symposium.
Hazardous Waste Engineering Research Laboratory, Office of Research and Development,
U.S. Environmental Protection Agency, Cincinnati, Ohio, 1988.
Maresca, Joseph W., Jr. A method of determining the accuracy of underground gasoline
storage tank leak detection devices. In: Proceedings of the Underground Tank Testing
Symposium, Petroleum Association for Conservation of the Canadian Environment,
Toronto, Ontario, 1982.
Maresca, Joseph W., Jr., Christopher P. Wilson, and Noel L. Change, Jr. Detection per-
formance and detection criteria analysis of the tank test data collected on the U.S. Envi-
ronmental Protection Agency national survey of underground storage tanks. Final Report,
Vista Research Project 2013. Vista Research, Inc., Palo Alto, California, 1985.
Maresca, Joseph W., Jr., Noel L. Chang, Jr., and Peter J. Gleckler. A leak detection
performance evaluation of automatic tank gauging systems and product line leak detectors
at retail stations. Final Report. American Petroleum Institute, Vista Research Project
2022, Vista Research, Inc., Mountain View, California, 1988.
Wilson, Christopher P., Joseph W. Maresca, Jr., Harold Guthart, John A. Broscious,
Shahzad Niaki, and Douglas E. Spitstone. A Program Plan to Evaluate Underground Stor-
age Tank Test Methods. Vista Research Project 2011, Vista Research, Inc., Palo Alto,
California, 1985.
Starr, James W., John A. Broscious, Shahzad Niaki, John S. Farlow, and Richard Field.
An approach to evaluating leak detection methods in underground storage tanks. In:
Proceedings of the 1986 Hazardous Material Spills Conference. U.S. Environmental Pro-
tection Agency, St. Louis, Missouri, 1986.
Starr, James W. and Joseph W. Maresca, Jr. Protocol for Evaluating Volumetric Leak
Detection Methods for Underground Storage Tanks. Technical Report, Contract No.
68-03-3244, Enviresponse, Inc., Livingston, New Jersey, and Vista Research, Inc., Palo
Alto, California, 1986.
Maresca, Joseph W. Jr., Robert D. Roach, James W. Starr, and John S. Farlow. U.S. EPA
evaluation of volumetric UST leak detection methods. In: Proceedings of the Thirteenth
Annual Research Symposium. Hazardous Waste Engineering Laboratory, Office of
Research and Development, U.S. Environmental Protection Agency, Cincinnati, Ohio,
1987.
98
-------
AN OUTREACH PROCESS; CASE HISTORIES OF UNDERGROUND
STORAGE TANK CORRECTIVE ACTIONS
by: William M. Kaschak, P.E.
Harold E. Lindenhofen
COM Federal Programs Corporation
Fairfax, Virginia 22033
Robert W. Hillger
Richard A. Griffiths
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Edison, New Jersey 08837
ABSTRACT
The U.S. Environmental Protection Agency's (EPA) regulations for
underground storage tanks (USTs) require corrective action to be taken in
response to leaking USTs. However, the level of experience of personnel in
the EPA regions, the states, and the local environmental agencies vary
widely. EPA is expanding its previously developed Case History File
database to facilitate technology transfer among the personnel involved in
UST corrective action. This information will allow UST personnel to obtain
the immediate benefit of the experiences of other people involved in UST
and other cleanup activities.
The original Case History File contains reports filed by On-Scene
Coordinators (OSCs) and Remedial Project Managers (RPMs) about the
technical, administrative, financial, and other aspects of spill and/or
waste site cleanups that they have handled. The File consists of a
database section and a narrative section. The database section allows
menu-controlled computerized searches to be made in any of 27 categories.
The narrative section contains detailed reports on the response actions
taken at the site. The narrative section is organized into 10 subsections:
General Information, Chemical Information, Effects of the Incident, Site
Characteristics, Containment Actions, Removal/Cleanup Actions, Treatment
Actions, Disposal Actions, and Operational Considerations.
The original File has been modified to incorporate additional data
relevant to USTs, such as methods of detection, causes of UST leaks,
tank/piping construction, etc. New reports are being added as the EPA's
Edison office receives them from the states and EPA regions.
This paper provides an overview of the UST Case History File, describes
how the data were collected, analyzes the initial data including discussion
of various technologies used for UST site cleanups, and provides a sample
scenario of a database search for an UST-related incident.
99
-------
INTRODUCTION
After December 22, 1988, the U.S. Environmental Protection Agency (EPA)
regulations (40 CFR 280) require corrective action for leaking underground
storage tanks (USTs). However, the level of experience among EPA, state,
and local response personnel vary considerably in this new fieldj and what
constitutes appropriate corrective action is not always clear. Presently,
no mechanism is in place to facilitate the transfer of technological
information among the federal, state and private communities concerning UST
corrective actions. To clarify this matter and to improve technology
transfer among response personnel, EPA expanded its previously developed
Case History File (File) database for hazardous material spills and waste
site remedial actions to include information on UST corrective actions.
The File is a component of EPA's Computerized On-Line Information
System (COLIS), which is maintained by the Environmental Emergency Response
Unit-Technical Information Exchange (EERU-TIX) contractor, Enviresponse,
Inc.^at EPA's Edison, New Jersey facility. The intent of the File is to
facilitate technology transfer among response personnel who need to select
site-specific corrective actions. The File is an easy-to-use informational
tool that eventually will include a significant amount of case history
data.
s The UST File is an on-line computerized system with a database section
and a narrative section. The database section allows searches to be made
using any combination of 27 different criteria, such as EPA Region, state,
hazardous substance, hydrology, UST construction, corrective action
technologies, etc. The narrative section of the File contains detailed
information in a plain-text format. The on-line system is very easy to use
and files can be created in only a few minutes.
Most of the data in the database and the narrative sections have been
obtained from after-action reports submitted by federal/state On-Scene
Coordinators (OSCs) or Remedial Project Managers (RPMs). The Case History
File will allow the review of technical information, such as past perform-
ance, cost, practicality, reliability, and other factors, to facilitate the
selection of cost-effective corrective actions. The objectives of the
project were to expand the existing database to accommodate information
relevant to UST corrective actions, and to collect case history data on UST
corrective actions involving a range of site conditions, locations, and
technologies.
CASE HISTORY FILE
The Case History File's data are organized into two sections: a
database section and a narrative section. The database section allows
searches by using key words. The user selects one of 27 available search
criteria shown in Table 1. Upon selection of a search criterion, either a
100
-------
new menu of key words appears for further selection or the user is required
to input certain information. An example of this would be for search
criteria 12, detection method. The user could choose external detector,
internal detector, inventory records, sight, smell, tank tightness, taste,
or others. In the case of criterion 19, depth to groundwater, the user
would be prompted for a range such as 15 to 20 feet.
TABLE 1. DATABASE CRITERIA OF THE CASE HISTORY FILE
1 - Incident number
2 - Date of incident
3 - Date of report
4 - Type of incident*
5 - U.S. EPA Region
6 - State
7 - NPL rank
8 - Site name
9 - Chemicals*
10 - Quantity
11 - Origin*
12 - Detection method*
13 - Main effects
14 - Resources affected
15 - Area affected
16 - Population affected
17 - Topography
18 - Hydrology*
19 - Depth to groundwater*
20 - Annual precipitation*
21 - Ground materials
22 - UST construction*
23 - Site uses
24 - Containment*
25 - Removal/cleanup*
26 - Site treatment*
27 - Disposal*
*Search criteria of specific interest to UST OSCs and RPMs
The narrative section of the Case History File contains detailed
reports in a text format and is organized into 10 subsections: General
Information, Chemical Information, Effects of the Incident, Site
Characteristics, Containment Actions, Removal/Cleanup Actions, Treatment
Actions, Disposal Actions, and Operational Considerations.
A user starts a search in the database section. When the user
specifies -a search category and a key word or numeric value for that e
category, the system creates a user's file that contains the incidents that
match the user's criterion. The user may create up to 10 files that have
10 different criteria and may view the data for the incidents in any one of
these 10 files at any time.
DATA COLLECTION
Data collection began with the 10 U.S. EPA regional UST coordinators,
who recommended states in which to begin the initial data collection
efforts. A checklist was developed and sent to 28 states to collect
initial information on sites, and was used as a screening tool to reduce
the sites to a manageable number for the initial data collection efforts.
The checklist was tailored to ensure that the initial sites selected would
provide a good representation of diverse hydrogeological conditions,
environmental settings, geographical areas, and corrective action tech-
nologies. Information on tank locations, leak quantities, hydrogeological
data, etc. was also contained in the checklist.
101
-------
After-action reports were sent to 24 states and the District of
Columbia. Site visits were made to 14 ongoing corrective actions in 7
states. A total of 50 after-action reports were collected from 16 states,
the District of Columbia, and 7 local offices.
ANALYSIS OF DATA SUBMITTED BY STATES
This section presents an analysis of the data collected during the
initial project activities. The objectives of the project were to focus on
petroleum USTs and completed or ongoing corrective actions. The infor-
mation presented is limited to the narrative sections of the Case History
File that emphasize UST technology transfer: site characteristics,
immediate corrective actions, long-term corrective actions, free product
removal, and operational considerations.
SITE CHARACTERISTICS
The site characteristics subsection includes information that describes
the site geology/hydrogeology and the various site investigation techniques
that have been undertaken to quantify subsurface conditions. The site
characteristics are critical to the corrective action selection process.
The results of the survey pertaining to site characteristics are limited to
a discussion of the topic of field sampling and analysis techniques.
Hydrogeological studies of varying degrees of complexity were performed
at 40 of the 50 sites. The most common technique used to define the extent
of contamination consisted of soil borings, installing monitoring wells,
and groundwater sampling. Modeling was used at only 4 sites to assist with
plume identification. Soil gas surveys were the most common field
screening technique that was used at 7 sites as part of the overall study.
The complexity of the field studies ranged from having 6 shallow soil
borings to installing 100 monitoring wells. Indicator compounds were used
at 38 sites, and most of the sites used more than one. The predominant
indicator compounds used were benzene, toluene, xylene (BTX) and benzene,
toluene, ethyl benzene, and xylene (BTEX).
IMMEDIATE CORRECTIVE ACTIONS
The section on immediate corrective actions contains information on
initial actions to mitigate the impact of a sudden or newly detected
release. Such actions are usually initiated within a few hours to a few
days from the time the release is discovered, but may take from a few hours
to several months to complete. Immediate corrective actions focus on
source control.
Immediate response actions were reported at 46 out of 50 sites. The
predominant technologies used were contaminated soil removal (at 17 sites)
and emptying the tank (at 11 sites). Removal of free product, replacing
the piping, and removing the tank were implemented at 9 sites. Additional
technologies used at several sites were tank testing, tank excavation,
installation of monitoring wells, pipe repair, explosivity monitoring, and
tank replacement. The distribution of immediate corrective action
102
-------
technologies reported for the 46 sites is presented in Table 2. Immediate
corrective actions at each site usually consisted of several technologies
used in combination. Therefore, the total number of incidents (115) is
greater than the total number of sites reporting immediate corrective
actions (46).
TABLE 2. IMMEDIATE CORRECTIVE ACTION TECHNOLOGIES
Technology
Incidents Reported
Remove Soil
Empty Tank
Remove Free Product
Remove Tank
Replace Pipe
Test Tank
Excavate Tank
Repair Pipe System
Install Monitoring Wells
Explosivity Monitoring
Replace Tank
Other
TOTAL
17
11
9
9
9
8
6
5
5
4
3
29
115
LONG-TERM CORRECTIVE ACTIONS
Long-term corrective actions are those actions undertaken to mitigate
the more extensive effects of contamination on public health and the
environment. They are not clearly separable from the immediate actions
taken or the actions taken to recover free product that has been lost.
Long-term corrective actions are applied to both the groundwater and the
soil and have been reported for 41 of the 50 sites.
Groundwater
Groundwater treatment was reported at 41 sites and involved 4 primary
techniques. Each technology and the occurrence of that technology are pre-
sented in Table 3. Long-term corrective actions for groundwater vary
greatly in complexity.
TABLE 3. GROUNDWATER TREATMENT TECHNOLOGIES
Technology
Incidents Reported
Air Stripping
Activated Carbon
Interceptor Trench
Gravity Separator
TOTAL
23
19
11
8
61
103
-------
In certain instances, the groundwater was extensively treated with a
well-planned series of steps. Preliminary separation was accomplished
either with a dual pump system or immediately after the recovery with an
oil/water separator. The effluent water was then treated with air
stripping, followed by carbon adsorption, and eventually discharged to the
sewer system. The applications, as a general rule, used a combination of
technologies rather than a single technology. A discussion of the various
technologies is presented below.
Air Strippers—Air stripping is a method of removing dissolved volatile
organics, such as BTEX compounds, from the water. Air is forced over thin
layers of water within a column packed with material to encourage the
volatilization of contaminants from the water into the air. Air strippers
are the most commonly used technology. Nearly half (23) of the unit
operations used air stripping.
Activated Carbon Treatment—Activated carbon treatment entails adsorption
of the organics present in the water or in the air in a stationary carbon
bed. The carbon will eventually become saturated with organics and stop
adsorbing any additional contaminants. The carbon is then replaced. The
spent carbon is either regenerated or treated as a hazardous waste and
disposed of.
Activated carbon units were used at 19 sites. At 8 of these sites,
activated carbon treatment followed air stripping as either a step to
achieve maximum cleanup levels or to remove pollutants from the air
discharge.
Interceptor Trenches—Interceptor trenches are used primarily for free
product recovery. The trench intercepts the groundwater and free product
in lieu of pumping from wells for recovery. The use of interceptor
trenches is limited, however, by groundwater depth.
Gravity Separators—A gravity separator is a coarse method of separating
free product from groundwater. The lighter free product phase floats on
the water and can be removed for independent disposal.
Soil
Like groundwater treatment, soil treatment technologies vary in com-
plexity. The major difference appears in the relative number of techniques
employed for the corrective action. Overall, 29 of the 50 sites had some
sort of corrective action listed for soil. All but 2 of these used
excavation of contaminated soil as either the primary or secondary
remediation tool. Other corrective actions utilized in the soil treatment
were aeration, biological treatment, and incineration. The number of
incidents using the various technologies for soil treatment is shown in
Table 4.
104
-------
TABLE 4. SOIL TREATMENT/DISPOSAL TECHNOLOGIES
Technology
Incidents Reported
Excavation
Aeration
Biological
Incineration
TOTAL
25
10
3
2
40
FREE PRODUCT REMOVAL
Free product removal generally refers to the recovery of product that
is located beneath the surface in large enough quantities for recovery by
mechanical methods. Free product removal was reported at 42 of the 50
sites. At 18 sites, free product removal consisted of several technologies
(2-3) used in combination. Pump systems were utilized at 26 sites.
Recovery systems were reported at 14 sites, while trenches were reported at
5. Additional devices used for free product removal include the vacuum
truck, skimmer, gravity separation tank, sorbents, and incineration. The
length of time required to complete the actions ranged from as little as 6
hours to as long as 3 years.
OPERATIONAL CONSIDERATIONS
This section of the report discusses operational considerations such as
permits, public involvement, administrative issues, and cost information.
The subtopics of major interest that will be discussed a.re permits and
costs.
Twenty-six after-action reports indicated that some type of permit was
required. The majority of required permits were National Pollutant Dis-
charge Elimination System (NPDES) or groundwater permits. At 6 sites, air
quality permits were required. There were no reported project slippages
due to the inability to obtain the required permit(s).
Of the 50 after-action reports, only 19 provided any information on
site remediation costs. The most expensive cleanup cost was $1.2 million,
and involved the installation of 14 wells in overlapping capture zones
where water was withdrawn and treated by an air stripper to remove volatile
organic compounds. The payment of cleanup cost was predominantly borne by
the site owner (38 out of 39 reported instances).
SAMPLE CASE HISTORY FILE DATABASE SEARCH
To illustrate the operation of the Case History File, the following is
a sample database search to demonstrate how an OSC would create files to
review the actions that other UST personnel have taken when confronted with
a similar incident. The scenario is for a gasoline tank that has ruptured
in a drinking water aquifer recharge area and released an unknown quantity
of gasoline into the environment. The OSC is interested in reviewing
105
-------
information on sites that have used either of two cleanup
technologies—bioremediation or air stripping—for remediation.
The OSC initiates the search by selecting UST incidents to create
fileO. He then selects the chemical, gasoline, to search and create filel.
Files 0 and 1 are then combined to create file2, which is UST incidents
with gasoline. Figure 1 provides a graphical presentation of what is in
file2 and how the OSC would complete a search of the Case History File and
create specific files based upon the criteria selected. The OSC begins to
search by the primary areas of interest: treatment technologies and
hydrology of the site. These searches result in files 3, 4, and 5 for air
stripping, bioremediation, and aquifer recharge, respectively. To narrow
the field of files to review, the OSC would combine files using "AND/OR"
logic. In this scenario, the OSC is interested in sites that have used
either air stripping or bioremediation and combines files 3 or 4 to create
file6. The OSC further refines the field by combining fileS and file6 to
create file?, which would be a site in an aquifer recharge area that has
either air stripping or bioremediation as the treatment technology. In
this scenario, only one site meets all of the search criteria, the City of
Farmington, New Mexico site, which actually employed both treatment
technologies.
At this point of the file search, the OSC is ready to review the files.
Upon the review of file?, the OSC could review any of the files created and
may want to review some or all of file6 to see where air stripping or bio-
remediation had been previously used. The OSC may also desire to review
other sites in fileS to see what other technologies may have been used.
The OSC could also search by additional criteria to create new files to
review.
For any of the files, the complete after-action reports could be
reviewed on the system. The initial display on the computer screen for the
City of Farmington incident would be a one-page Abstract of Incident as
presented in Figure 2. The Abstract of Incident is a summary of the
database information for the site. After the Abstract of Incident is
reviewed, the user may then review the narrative sections, which provide
more specific details of the site and actions taken.
CONCLUSIONS
The level of detail and the questions answered among the 50 Case History
File After-Action Reports varied widely. Many of the questions could not be
answered from the file material alone. Follow-up discussions on the Case
History File After-Action Reports were required in numerous instances.
The responses to the Case History File After-Action Reports indicated
that the information requested is thorough in capturing the information that
would be useful in responding to leaking USTs. The state personnel also
acknowledged their willingness to work with a computerized database system
and indicated a need for such a system to promote technology transfer and to
educate new staff. The Case History File was demonstrated at a national UST
106
-------
SAMPLE DATA BASE SEARCH FOR A SPECIFIC UST
INCIDENT
File 2
UST incidents with Gasoline
Circle Mobil, SC
Cftyof Farrnlngton, NM
City of Cypress, CA
Chevron, CA
Thompson Grocery, LA
Stadium Mobil, MA
Molan Oil Co., MO
Exxon Site, MD
Caltrans, CA
Week Brothers Mobil, IA
Wans Service Station, MA
Crossroads USA, IW
Boll GAS, MO
Texaco USA., FL
Southland, TX
Super America, MN
Armour Oil, CA
D.C. DES, WA DC
W.W. Service Center, AZ
Boron Gas Farmington, Ml
White Oil, CO
Total number of Sites
associated with a specific
file
/
/
4
4
4
4
4
/
4
/
4
4
4
4
•f
4
/
•f
4
4
4
21
Fil^3
Air
Stripping
/
4
4
4
4
4
4
4
4
9
File 4
Bio-
remediation
/
4
4
4
4
5
FileS
Aquifer
Recharge
4
4
4
3
File 6
File 3 or
File 4
4
4
4
4
4
4
4
4
4
4
4
11
File?
File 5 and
FileS
4
1
FIGURE 2
107
-------
CASE HISTORY FILE
Abstract of Incident
Incident
Region
Site name
Substance
Quantity
Origin
Effects
Resources
UST const
Geography
Hydrology
Ground
Site uses
Containmt
Removal
Treatment
Disposal
102 Type : UST Inc. date
6 State: NM Rpt. date
City of Farmington NPL rank
Gasoline CAS #
30,000 gallons DOT #
UST installation, error Detection
Soil and water contamination Area
Groundwater, drinking water Popula
steel
15 May 85
02 Feb 88
N/A
71-43-2
1114
inventory
100 acres
5
valley Precip : 8.2 in.
aquifer recharge GW depth : 10 ft.
sand, gravel, bedrock, asphalt surface, concrete surfaced
commercial, residential, rural, parkland
groundwater control
empty tank, replace tank, remove tank, excavate soil, exc
biodegradation, air stripping, groundwater extraction, vo
land farming, evaporation, treatment, soil washing, injec
FIGURE 2. SCREEN DISPLAY OF AN ABSTRACT OF INCIDENT
FOR THE CASE HISTORY FILE FOR INCIDENT 102
workshop held in Santa Fe, New Mexico in November 1988, as one of several
tools presented for use by state personnel in managing their UST programs.
The major conclusions drawn from the analysis of the data and
information reported in the narrative section of the Case History File are:
o
o
o
o
o
o
Time for cleanup varied from a few hours to years.
Indicator compounds are used to quantify extent of contamination.
Immediate corrective actions were implemented at 46 of the 50 sites.
Predominant groundwater treatment technologies were air stripping and
activated carbon.
Permits were required, but did not slow down the response.
Contaminated soils were usually excavated from the site.
The Case History File provides an excellent forum for technology
transfer. The initial data collection efforts of the UST incidents have
been well received by the state personnel.
108
-------
CONSIDERATIONS OF UNDERGROUND STORAGE TANK RESIDUALS AT CLOSURE
by: ¥arren J. Lyman
Camp Dresser & McKee Inc.
Boston, MA 02108
and Anthony Tafuri
U.S. Environmental
Protection Agency
Edison, NJ 08837
ABSTRACT
The U.S. Environmental Protective Agency (EPA) is currently evaluating
several technical and regulatory aspects of underground storage tank (UST)
closures. A key concern is the manner and extent of tank cleaning that is
appropriate and feasible when a tank is removed from service. The
objective of this work was to obtain a thorough technical/scientific
understanding of UST residuals at closure: their origins,
physical/chemical properties and ease of removal by different cleaning
methods. The information generated will be used as an aid in the
regulatory process and will be useful to those implementing/overseeing
closure activities.
Information was obtained via phone contacts with knowledgeable
individuals including tank cleaning companies, published and unpublished
literature, site visits and worksheets completed by state UST program
representatives. The investigation was limited to underground storage
tanks containing gasoline-and diesel.
Gasoline and diesel USTs were found to have significant quantities of
residuals in them at closure, typically tens to a few hundreds of gallons.
However, although there is little explicit guidance available, tank
cleaning and removal companies are apparently capable of removing most of
these residuals with fairly simple cleaning techniques. Objectives and
quantitative evaluations of the effectiveness of the cleaning techniques
are not possible at present because of the nearly complete lack of anything
but anecdotal data.
INTRODUCTION
The magnitude of the environmental contamination problem - actual and
future potential - presented by underground storage tanks (USTs) has only
become clear in recent years. Hidden leaks and unreported spills have
probably occurred at tens of thousands of USTs and would continue at
thousands more without intervention. Federally mandated intervention will
109
-------
come as a result of EPA's Final Rule published in the Federal Register on
September 23, 1988. Many states and local communities have also passed
their own set of UST regulations. The direct or indirect effect of these
rules will be the closure and removal - or abandonment in place - of
hundreds of thousands of USTs over the next 10-20 years.
A key question related to UST closure involves the quantity and compo-
sition of residual liquids in the UST and the dangers to human health and
the environment that would result from their improper handling or release.
In August, 1988, EPA's Risk Reduction Engineering Laboratory (RREL)
initiated a study to obtain basic scientific and technical information on
UST closure activities, with special attention on UST residual and cleaning
procedures used to remove them. Specific tasks included: (1) characteriza-
tion of UST residuals (e.g., quantities and composition); (2) an evaluation
of removal and disposal alternatives for UST residuals; (3) an events study
of UST closure activities using the Darning management approach; and (4) a
sampling and analysis of USTs before and after cleaning operations.
This paper covers only work carried out in support of the first two
tasks. The approach taken to obtain desired information included:
o Phone contacts with over seventy knowledgeable individuals
including tank cleaning companies
o Observations of four tank cleaning/closure operations
o Review of the limited amount of published information available
o Review of the answers to a worksheet/questionnaire filled out by
state UST program representatives (and others) at an UST workshop
(total of 23 responses)
o Carrying out a few simple engineering calculations related to
residuals generation and closure costs
INVESTIGATION OF UST RESIDUALS
QUANTITY
Overview
Underground storage tanks containing either gasoline or diesel can
usually be emptied to. within 4-6 inches (10-15 cm) of the tank bottom.
This distance usually dictates residual quantity; e.g., for a 10,000-gallon
tank, this distance translates into about 100-200 gallons* of materials
*Because gallons is a more common measure of volume and UST capacity, these
units are used in this paper. For conversion to metric: 1 gal = 3.785
liters.
110
-------
remaining on the tank bottom. Much larger amounts are frequently found,
usually associated with abandoned tanks or with leaking tanks into which
groundwater has flowed.
Information from Phone Survey
During the phone survey , a total of 18 estimates was obtained
regarding the volume of residuals for an average 10,000-gallon tank,
containing either gasoline or diesel. Additional estimates of fuel oil
residuals, for comparison purposes, and rinseates (resulting from cleaning
operations) were also obtained.
Gasoline — The volume of residuals found in gasoline tanks at any one
site can vary significantly, from 0 to 10,000 gallons. While these ranges
are in fact possible, they are not typical. On the average, the twelve
sources reporting volumes of gasoline UST residuals quoted residual
quantities from 0 to 1,000 gallons as typical. The mean of the values
reported was 160 gallons, while the median of the values reported was 75
gallons.
Diesel — Most respondents qualitatively agreed that diesel tanks
contained more residuals than gasoline tanks, all other things being equal.
However, the six quantitative responses of volume of diesel residuals
obtained during the survey ranged from 0 to 200 gallons, with a mean value
of 58 gallons. The median value was slightly higher and similar to that
estimated for gasoline tanks: around 75 gallons.
Fuel Oil — Most respondents qualitatively agreed"that fuel oil tanks
produced the greatest amount of residuals, in comparison with gasoline and
diesel oil tanks. The two quantitative responses of volume of fuel oil
residuals obtained were 500 and 1,000 gallons, averaging to 750 gallons,
significantly higher than gasoline and diesel.
Rinseates — As one would imagine, the volume of rinseates generated
during the cleaning procedures can vary widely with the type of cleaning
procedure used. The three phone-survey values of rinseate volumes ranged
from 100 to 3,300 gallons, with an average of 1,200 gallons. Again, as one
would expect, these volumes are significantly higher, in fact one order of
magnitude higher, than the residuals themselves. As is noted below, the
American Petroleum Institute's (API) Recommended Practice 1604 calls for
the tank to be filled nearly to the top for cleaning and/or vapor removal
purposes. This practice would generate much greater volumes of rinseate.
Detailed Information from One Company
In connection with an UST sediment characterization project for the
State of Minnesota, Delta Environmental Consultants, Inc., examined the
files of one tank cleaning and removal company that kept detailed records
of the depth and volume of residuals in each UST removed. The range of
residuals volumes found for gasoline, fuel oil and waste oil tanks is shown
in Figure 1. A statistical summary is provided in Table 1.
Ill
-------
Ill
g
Q
CO
Ul
CC
Q
LU
Q
j
2
U.
O
CC
LU
CO
6 20 " 40 60 " 80 " 100 " 140 180 250 400 600 800 >1000
200 300 1,000
0 20 40 60 80 100 140 180 250 400 600 800 >1000
200 300 1,000
10
0 20 40 60 SO 100 140 180 250 400 600 800 >1000
200 300 1,000
VOLUME OF RESIDUALS FOUND IN TANK (gal)
(NOTE CHANGES IN SCALE AT 100,200,300 AND 1000 gal)
Source: Delia Environmental Consultants (1988)
FIGURE 1. QUANTITY OF RESIDUALS FOUND IN UST's BY ONE
MINNESOTA COMPANY (9-1-87 TO 8-30-88)
112
-------
TABLE 1. QUANTITY OF RESIDUALS FOUND IN USTs BY ONE
MINNESOTA COMPANY (9-1-87 to 8-30-88)*
Number of Tanks
Avg. Tank Capacity (gal)
Avg. Residuals Volume (gal
Median Residuals Volume (gal)
Gasoline
214
5,800
49
-20
Fuel
Oil
221
5,900
81
-40
Waste Oil
151
3,600
162 (93**)
-50
information derived from raw data in reference 3.
**Excluding one tank with 9,375 gallons of residuals.
ORIGIN, NATURE AND COMPOSITION
Overview
Based primarily on anecdotal data and some rough calculations, it is
estimated that 70-90% of the gasoline and diesel residual consists of the
product itself, probably of somewhat diminished purity. The remaining
10-30% consists mostly of water (with numerous dissolved constituents);
product-related residuals (e.g., gum, sediment, tars); rust and scale (in
steel tanks); dirt and other foreign objects; and a small, but
disproportionately- important mass of microorganisms. The importance of
the microorganisms comes from the significant internal corrosion that can
be due to the action of sulfate-reducing bacteria.
Location in Tank
The typical location of residuals within an UST is diagrammed in Figure
2. While most of the residuals will reside on the bottom of the tank, the
presence of some side-wall scale and gum has also been observed. The
bottom residuals, while containing some gum, scale and grit, are usually
pumpable liquids, but might properly be considered sludges with low solids
content. The settleable solids are frequently found pushed slightly to the
side of the UST (e.g., to the 5- and 7-o'clock positions as viewed end on)
due presumably to turbulence during filling operations. Finally, in tanks
that are installed with a slight end-to-end tilt, as is proper, a greater
fraction of the residuals will be found at the low end of the tank.
Composition of Solids/Sludge Fraction
Excluding larger debris (e.g., broken dip sticks, beverage cans, rubber
hoses), the solids/sludge content of UST residuals is probably mostly dirt,
rust (tank scale), and high molecular weight organic material (e.g., tars
and gums). Visual observation of samples scraped from the bottoms of USTs
during manual cleaning - after all easily pumpable material had been
removed - indicated about a 50-percent solids content. Some data from
analyses of such samples for benzene, toluene, ethyl benzene, xylene and
total hydrocarbons are shown in Table 2.
113
-------
Underground Storage Tank with most Product
Removed in Preparation for Closure
Residual Fuel
(approx. 4 to 6 inches)
Sediment, grit, gum
(dirt, rust particles, or fuel sediment) Water Layer
Thickness of sludge may be enhanced (probably < 1 inch)
at 5- and 7-o'clock positions in vicinity
of fill tube.
FIGURE 2. SCHEMATIC OF UST RESIDUALS
114
-------
TABLE 2. BENZENE, TOLUENE, ETHYL BENZENE, XYLENE, AND
TOTAL HYDROCARBONS IN THE SOLID/SLUDGE PORTION
OF UST RESIDUALS3
SarnpL
Numbei
MDLC
1
2
2d
3
4
4d
5
6
7
7d
8
Parent
* Material .
c of Sample B<
Gasoline
Mixed Oil
Mixed Oil
Mixed Oil
Mixed Sludge
Mixed Sludge
Mixed Sludge
tt6 Fuel Oil
tt2 Fuel Oil
#2 Fuel Oil
Dried Residual
Concentration (rag/kg)
snzene
0.12
110
5.1
4.9
39
1.9
1.5
190
1.0
3.8
4.0
7.1
Toluene
0.12
270
11
10
100
12
8.4
310
2.4
11
10
160
Ethyl
Benzene
0.12
30
1.8
1.8
13
6.2
4.4
44
1.0
1.9
1.7
41
Total
Xylene Hydrocarbons
0.12
140
8.8
8.9
67
19
13
210
6.0
9.4
8.3
210
1.0
1700
120
110
800
270
180
2400
86
109
110
1800
Data from reference 3.
Sample Descriptions;
1. Sludge at bottom of gasoline storge tank
2. & 3. Drying mixed oil residuals in two different tanks
4. & 5. From mixed sludge drums
6. From drum of #6 fuel oil residuals
7. From drum of #2 fuel oil residuals
8. Composite of dried residual from a number of open tanks
CMDL = Method Detection Limit
Duplicate analysis ,
115.
-------
Water Content
The phone survey also provided significant evidence for the presence
of a layer of waiter at the bottom of many, if not most, USTs. It is a
common practice for USTs in service, for example, to check for the presence
of water (and sediment) with a dip stick prior to refilling the UST. The
end of the dip stick is coated with a special paste that changes color on
contact with water. A rule of thumb for some gas stations is to limit the
depth of water to about one inchj greater amounts are pumped out prior to
refilling the tank. (One inch of liquid in a 10,000 gallon tank represents
about 12-18 gallons.) At stations where frequent monitoring was not
undertaken, or where water input rates were high, the volume of water at
the bottom of the tank could clearly be much higher. Also, abandoned tanks
are frequently found to have large volumes of water due to precipitation
runoff into open fill tubes or groundwater leakage in through holes formed
by corrosion.
Residual water would contain a significant amount of dissolved-
hydrocarbons (~100-300 mg/L), dissolved salts (e.g., Na , Cl , Fe , HC03~,
Pb ) and other soluble components or additives present in the fuels (e.g.,
ethanol, methyl-t-butylether (MTBE), detergents). The composition could
lead to such water being classified as a hazardous waste and to a
requirement for pretreatment prior to discharge to any sewer.
There is one mechanism by which water can accumulate in USTs solely by
inputs dissolved in the product delivered to the site. The water present
in solution in the product delivered to an UST is likely to be near the
solubility limit and, in summer at least, warmer than ground temperatures.*
As the fuel enters the UST and cools down, the solubility limit is lowered
causing some water to come out of solution and form, or add to, a separate
aqueous phase.
The order of magnitude of the contribution of this phase separation
process is estimated to be one gallon of water per tank refill for
gasoline USTs. Water is also suspected to enter USTs via entry of moist
air through the open fill tube followed by condensation.
This UST water may play a significant role in the internal corrosion of
steel tanks. Several surveys have shown that internal corrosion of steel
tanks is a fairly common occurrence although external corrosion is roughly
three times more important. Significantly different mechansims may be at
work in internal and external corrosion, although the presence of water is
probably necessary in both cases. For internal corrosion, this water may
be present as a condensate on the tank walls or as a layer on the bottom.
*The solubility limits for water in gasoline and in diesel are not known
precisely but are probably on the order of 1,000 mg/L and 100 mg/L,
respectively. Significantly larger amounts might be present in solution
if the fuels have hydrophilic additives such as ethanol or MTBE.
116
-------
Tank Rust or Scale
The phone survey and a limited literature review provided evidence that
steel tanks are likely, over time, to shed rust particles (iron oxide,
Fe2°3^ an<* *ron scale on the ir»side. This internal corrosion may be caused
by galvanic action or bacterial action (see subsection on Microorganisms).
Some of the rust and scale may remain on the tank walls while portions
will drop and accumulate on the bottom. The total volume of side and bottom
scale is thought to be relatively small, perhaps no more than one liter.
A rough estimate of the amount of rust that might accumulate in an UST
can be derived from a calculation in which 0.1% of the mass of the steel
tank is assumed to be converted from Fe to Fe^O,,. This leads to an
estimate of about 9 kg (20 Ib) of rust in a 10,000-gallon tank.
Microorganisms < .
Like water, microorganisms appear to be fairly ubiquitous in petroleum
storage and distribution_systems. While they may appear to be present in
large numbers (10 to 10 organisms per liter), their mass is small. At
times, however, large floes can be formed which can clog fuel lines and
fuel filters.
The microorganisms involved include several varieties of bacteria and
fungi. Of special import are the class of sulfate-reducing bacteria that
can cause significant corrosion to iron and steel products. These bacteria
are strict anaerobes that perform anaerobic respiration by oxidizing
certain organic compounds or H2, and reducing sulfate, and often other
reduced sulfur compounds, to hydrogen sulfide. The sulfide can then react
with iron to form an iron sulfide precipitate (FeS) which can contribute to
the solid portion of UST residuals.
Corrosion is usually evidenced as pits below microbial mats. There are
numerous theories as to the biochemical and chemical basis for the
corrosion, but no current agreement on any one. No data were found on
rates of microbial corrosion to be expected in USTs.
Microorganisms do need water to thrive and, in storage tanks, are
usually found at the fuel-water interface. The mix of hydrocarbons, water
content, oxygen content (low for anaerobes), nutrient content and pH are
all important factors in the growth of these microorganisms. They
apparently thrive better in kerosene and fuel oils than in gasoline.
Ignitability
o
Delta Environmental measured the flash points of UST residuals. In
general, their results indicated that gasoline sludge, scale and dried
residuals, and mixed sludge had flash points near 50-60°F and can be
flammable. Residuals from higher boiling fuels (e.g., sample nos. 2, 6 and
7 in Table 2) had flash points above 140°F, the point above which materials
are not considered hazardous because of ignitability.
117
-------
EP Toxicity
Delta Environmental also evaluated the EP Toxicity of the UST
residuals. In this procedure, concentrations of eight heavy metals in an
aqueous extract of the waste are compared with specified limits. EP
Toxicity analyses were below regulatory limits for all metals except for
one lead sample.
REMOVAL AND DISPOSAL
CLOSURE PRACTICES
Phone survey and analysis of worksheet questionnaires (from the EPA/UST
Workshop ) showed that tank removal, involving some form of tank cleaning,
was most commonly practiced. Second was closure in place, which involves
filling the tank with an inert material (e.g., clean sand) after removing
the residuals. In a few instances, tank closure involved a change in
service for the tank (i.e., a change in the liquids stored in the UST).
The following sections cover only closure in which tank removal is a part,
and focus on the cleaning practices.
CLEANING PROCEDURES
Procedures Used
A variety of tank cleaning and removal procedures appear to be in use,
although many are variations of a simple, logical theme. Many of the steps
in these procedures are dictated by safety considerations* and by state and
local regulations, rather than by a direct concern for strict tank
cleanliness. Most procedures involve an initial pumping of residuals with
a suction line and a subsequent rinse with water followed by rinseate
removal. The water rinse may involve: (1) filling the tank with water;
(2) rinsing with spray from a "garden" hose [low pressure]; (3) rinsing
with high pressure water; (4) steam hosing; and (5) possible use of a
detergent. The American Petroleum Institute's recommended procedures (API
1604) call for filling the tank with water followed by sequential removal
of floating product and water.
Several tank cleaning companies cut a manhole into the UST, allowing a
man to enter and physically remove bottom grit and (with a "squeegee")
liquids adhering to the side walls. Some companies consider this procedure
too dangerous, especially for gasoline tanks; the practice is prohibited in
some areas. A summary of the basic UST cleaning and removal steps in the
API 1604 procedure, and in other procedures described in the phone survey,
is provided in Table 3.
*Prevention of human exposure to toxic chemicals, fires and explosions, and
spillage.
118
-------
TABLE 3. SUMMARY OF UST CLEANING AND CLOSURE PROCECURES
SOURCE*
A,B,C
API 1604**
COMMENT
E
H
1. Prepare workers and area for safe operations.
2. Drain product piping into tank; also cap or
remove product piping.
3. Remove liquids and residues from tank.
4. Excavate to top of tank.
5. Remove tank piping, pumps and other fixtures.
6. Purge tank of flammable vapors.
7. Fill tank with water until floating product
nears the fill opening; remove floating product.
8. Pump out water.
9. Test tank atmosphere for flammable or
combustible vapor concentrations.
10. Plug or cap all accessible holes except 1/8"
vent hole.
11. Excavate and remove tank.
1. Empty tank as much as possible.
2. Triple rinse, with high pressure water (gasoline
tank) or detergent (diesel tank).
3. Inert tank with NZ or C02-
4. If necessary, enter tank and physically remove
sludge. 2
5. Punch 6 holes, each 1 ft , to render tank
useless.
6. Remove tank from ground.
1. Empty tank as much as possible.
2. Purge tank.
3. Cut opening(s) in tank.
4. Rinse, pump out rinseate.
5. Remove tank from ground.
Proprietary process involving pumping fuel out of
tank, filtering through vacuum, spraying fuel back
into tank through nozzle, pumping, filtering, etc.
May take numerous cycles to clean tank.
Warm water and detegent used as rinse agent;
high-pressure not used because of safety hazard.
1. Empty tank as much as possible.
2. Inert with C02.
3. Cut opening in tank.
4. Worker enters tank, physically removes any
sludge or scum.
5. Remove tank from ground.
119
-------
TABLE 3. SUMMARY OF UST CLEANING AND CLOSURE PROeEDURES (Cont'd)
SOURCE*
COMMENT
K
H
1. Empty tank as much as possible.
2. Purge tank.
3.
4.
5.
6.
1.
2.
3.
4.
5.
1.
2.
3.
4.
5.
6.
1.
2.
3.
4.
1.
2.
3.
4.
Cut 2-ft opening in tank.
Worker enters tank: squeegees side and bottoms;
scrapes sides and bottoms; washes with water.
Rinseate is pumped out.
Tank is removed from the ground.
Empty tank as much as possible.
Purge tank.
If tank has a manhole: rinse with caustic
pH) detergent.
Pump out residuals and rinseate.
Remove from ground, lay on its side.
Empty tank as much as possible.
Inert with C02*
Cut opening in tank.
Physically clean residuals.
Inert tank.
Remove from ground.
Empty tank as much as possible.
Triple rinse with high pressure steam.
Inert with C02.
Remove from ground.
Empty tank as much as possible.
Remove from ground.
Cut manhole in tank.
Worker physically removes residuals.
(high
*Each letter (A, B, C...) represents a different source that was
interviewed in the phone survey. Most of the soures are tank
cleaning and removal companies who are describing their own
standard procedures. In two instances, the procedures are those
specified by a county agency.
P
**Basic steps in API Recommended Practice 1604. Several details
relating to safety and regulatory compliance have been omitted
for brevity.
120
-------
With some companies, it is common to put both initially-pumped
residuals and aqueous rinseate into the same tank truck (for off-site
treatment and disposal). Other companies segregate the residuals from the
rinseate thus facilitating subsequent product recovery and/or treatment.
The volume of rinseate generated appears to range from a low of 100-200
gallons per tank to about one third of the tank's volume, except for the
API 1604 procedure, which calls for filling the tank with water.
Cleaning Effectiveness
There are no data which provide objective evidence of the degree of
cleanliness achieved by the procedures used. For tanks that are
subsequently reused as scrap metal (i.e., crushed or cut-up and then
remelted), a modest amount of retained residuals may be environmentally
acceptable. Worker protection may be the more stringent basis for
regulation. For tanks that are filled in place or landfilled, the
residuals remaining after cleaning operations (retained residuals) are
likely to pose only a small-to-negligible risk of adverse environmental
impact. This would be related to the small volume of retained residuals,
limited environmental mobility for most constituents, and limited
toxicological significance for the bulk of the constituents.
In practice, tank cleanliness (i.e., the absence of sludge, scale,
sediment and liquid product or rinse water) is "determined" by a number of
methods including:
o Visual inspection of the tank
o Visual inspection of a wipe sample
o Analysis of the cleaning rinseate
o Adherence to a standard cleaning procedure (e.g., API 1604 or the
cleaning company's own standard procedures)
The frequency of inspection (of cleaned tanks) by state or local
officials appears to be quite low.
TREATMENT AND DISPOSAL OF SECONDARY WASTES
Secondary wastes from UST cleaning operations fall into several groups:
o "Pure" residuals removed from the UST (and stored separately) before
any cleaning is undertaken; this is expected to be 70-90% fuel
product (i.e., gasoline or diesel)
o Liquid/sludge residuals; What is left after "pure" product is
removed; may also have enhanced solids concentration if tank walls
were manually scraped
o Rinseate; usually resulting from a simple water wash sprayed into
the UST. Detergents may be used
121
-------
° Combined wastes; Some cleaning companies pump the pure residuals
and rinseate into the same tank truck at the UST site
o Debris; including large items found in tank as well as discarded
cleaning rags and protective clothing used during cleaning
Very little data were obtained on the quantity of secondary wastes
beyond what is known about the amounts of "pure" residuals usually found in
USTs (approx. 100 gal for 10,000-gal tank). The amount of wash water used
varies widely from as little as 100 gal to amounts approximating one third
of the tank volume. If the API Recommended Practice 1604 is adhered to
strictly, a nearly full tank of wastewater would be generated.
No reliable data were obtained on the composition of these wastes
beyond what was provided in Table 2. It is fairly easy to speculate,
however, given the general understanding of the composition of "pure"
residuals and the extent of dilution by water in the cleaning process. A
"combined waste", for example, might roughly be a 50/50 mixture of fuel and
water; detergent would be present if used in the cleaning process. Without
scraping and thorough pumping, a significant fraction of the side scale and
gum, and bottom sediment, might not be transferred from the UST to the
secondary wastes. All waste groups are likely to have emulsion
characteristics, i.e., small droplets of one phase dispersed throughout the
other phase. The use of detergents would increase the degree of
emulsification.
Very little information has been obtained on actual treatment and
disposal of UST secondary wastes. However, an initial separation of
hydrocarbon liquids from aqueous phases (in a large settling tank) is
common. The aqueous phase often ends up, after varying degrees of
pretreatment, in a sanitary sewer leading to a municipal biological
treatment plant. The hydrocarbon phase is often treated as a waste oil
(i.e., shipped to an oil refiner), burned for its heat content, incinerated
or, less likely, drummed and landfilled.
Most states providing information on the worksheets at the EPA/UST
Workshop indicated that UST residues and cleaning by-products were
considered hazardous wastes and thus had to be handled according to
appropriate state and Federal rules for such material.
CONCLUSIONS
This study has shown that there are significant amounts of residuals
left in USTs at the time closure is initiated. The handling and removal of
these residuals during closure - and the resulting tank cleaning - are
being carried out by a wide variety of procedures often dictated more by
preferences of local officials or the selected contractor then by objective
guidelines. The cleaning procedures typically generate a significant
volume of aqueous rinseate that also presents disposal problems.
122
-------
A significant fraction of the residuals consists of recoverable fuels,
and fuel recovery is practiced by a number of tank cleaning contractors.
The residuals (before or after removal of recoverable fuels) are
potentially hazardous as a result of both ignitability and toxicity
characteristics, yet in some states the material is classified as a waste
oil and not a hazardous waste.
The tank cleaning methods currently in use appear to be able to
satisfactorily clean most gasoline and light oil USTs. However, no data
exist indicating just how clean the tanks do get, no standard or generally
accepted method of tank cleaning has been identified, nor have any criteria
been developed for tank cleanliness for any of the various tank disposal
alternatives.
Monitoring of selected tank cleaning methods, quantitative measurements
of the amounts of residuals left in the tank after cleaning, and a
characterization of the rinseate generated are planned for the next
portions of this study.
REFERENCES
1. The phone survey data cited in this paper are documented in the
following report: "Evaluation of the Technical Aspects of UST
Closure," Interim Report, EPA Contract No. 68-03-3409, Work Assignment
No. 16, September, 1988.
2. "Removal and Disposal of Used Underground Petroleum Storage Tanks," API
Recommended Practice 1604, second edition, American Petroleum
Institute, Washington, D.C., 1987.
3. Delta Environmental Consultants, Inc. (St. Paul, MN), "Tank Sediment
Characterization and Disposal Report," report to Minnesota Pollution
Control Agency, St. Paul, MN, November, 1988.
4. "Making it Work: Workshop for State Tank Program Managers,"
Conference/workshop sponsored by the U.S. EPA's Office of Underground
Storage Tanks, held in Santa Fe, NM, November 15-17, 1988.
123
-------
LABORATORY STUDIES OF VACUUM-ASSISTED STEAM STRIPPING OF ORGANIC CONTAMINANTS
FROM SOIL
by: Arthur E. Lord, Jr., Robert M. Koerner, and Donald E. Hullings
Geosynthetic Research Institute
Drexel University
Philadelphia, PA 19104
and
John E. Brugger
Risk Reduction Engineering Laboratory
Superfund Technology Demonstration Division
U.S. Environmental Protection Agency
Edison, NJ 08837
ABSTRACT
A long-term research and demonstration project is underway to
investigate vacuum-assisted, steam stripping of organic chemicals from
contaminated soil and to develop a field unit to steam strip and collect such
pollutants. In previous work, an analytical model was developed for steam
stripping in the field. The model involved steam flow to the surface from
pipes embedded below the ground surface. The data needed to implement the-
model were obtained from experiments that involved vacuum-assisted steam
stripping of kerosene from a variety of soil types (from sands to silts) in
small scale laboratory experiments. This approach was used to determine the
time to decontaminate a given kerosene spill in a particular soil. Small
scale pilot studies were also made of the field unit employing a unique
geosynthetic, vacuum cap assembly. The results of this first phase showed
that kerosene (and also gasoline) could be quite effectively removed from a
wide variety of soils with the vacuum-assisted, steam stripping technique.
Also, the geosynthetic cap (geotextile plus geomembrane) anchored in the
soil, performed quite well in confining and collecting the steam and
contaminant.
The chemical analysis method in the first phase of the work used simple
volume separation of the collected kerosene and water fractions. Gas
chromatography (GC) is presently employed to analyze the residual chemical
content in the soil after steam stripping. The efficiency of vacuum-assisted,
steam stripping of alkanes (octane, decane and dodecane) from a range of
soils from sands to silts, was determined. Octane, which has a relatively
high vapor pressure (b.p. = 126°C) could be reduced from an initial content
of 5-10%, to less than 10 ppm in a few hours. Dodecane, which has a
124
-------
relatively low vapor pressure (b.p. = 216°C) was reduced to about 1000 ppm in
the same time frame. Decane which has an intermediate vapor pressure (b.p. =
174°C) was reduced to 80 ppm. Octanol, which has vapor pressure
characteristics between decane and dodecane (b.p. = 195°C) and is somewhat
polar, was reduced to 10 ppm. Butancl, which has vapor pressure
characteristics similar to octane (b.p. = 118°C) and is strongly polar was
reduced to lower than 1 ppm.
As a result of this work, the utility of vacuum-assisted, steam
stripping methods to decontaminate soils at Superfund (and other) sites is on
a somewhat firmer technical basis than before. This paper identifies future
research work, such as determining the exact nature of the relation between
compound vapor pressure and polarity on the ability to steam strip the
compound from the soil.
INTRODUCTION AND OVERVIEW
With contaminated soils at Superfund (and other) sites, it is important
that the chemicals be prevented from reaching the groundwater. Fortunately,
in many locations the partially saturated or vadose zone exists and acts as
temporary containment retarding the downward movement of the pollutant. The
remediation options are:
• Excavation and off-site disposal
• Excavation and on-site treatment
• In-situ treatment (via a number of possible methods, e.g., biological,
physical or chemical)
A number of these techniques (and others) have been reviewed in recent
articles (1,2). The in-situ techniques have been discussed by.the authors
(3) .
The present study falls in the in-situ treatment category wherein the
authors propose to have pipes inject steam into the soil beneath the
contaminated zone. Steam stripping of the chemical occurs and when aided by a
vacuum at the ground surface brings the contaminants to a collection point
where they can be properly treated. A unique aspect of the study is a
geosynthetic cap assembly consisting of a high transmissivity geotextile and
a flexible membrane liner (geomembrane). The vacuum is applied to the
underside of this liner and the contaminated gas and/or liquid moves beneath
the liner in the geotextile to the outlet ports. A schematic diagram of a
proposed system is given in Figure 1 for reference purposes.
There have been three steam stripping, soil decontamination studies
reported in the literature, as far as the authors are aware (4,5,6). [One
company mentions steam stripping in their advertisements for hazardous waste
site remediation (7).] The field works are site specific projects, and
understandably make no attempt to look at the general problem of steam
stripping a wide variety of chemicals from a wide variety of soils.
The present work involves a long term study to determine the ability of
vacuum-assisted, steam stripping to decontaminate general organic chemical
125
-------
species from a variety of soil types.
PREVIOUS WORK ON THIS PROJECT
The work performed in the first phase of the project has been reported
elsewhere (3,8,9). Only a brief review of the results of this work will be
given here. Among the tasks undertaken, in a wide variety of soils were:
• Observations were made of the transient steam front movements in two
dimensional flow. (The sample size was 2 ft x 2 ft x 1 in. and steam
at about 5 psi pressure was used.)
• The ste,am permeabilities were determined in conventional one-
dimensional flow. Results are given in Table 1. (The sample size was
2.5 in. diameter X 6 in. high.)
TABLE 1. STEAM AND WATER PERMEABILITIES
Soil Standard Water Steam Permeability Steam Permeability
Permeability Water Permeability
Sand (%) Silt (%)
100 0
75 25
50 50
25 75
0 100
(cm/sec)
1.38 X10-3
2.06 X 10-4
8.82 X 10-5
9.61 X 10-5
3.6 X 10-6
(cm/sec) —
1.70 X 10-4 Q.13
3.14 X 10-5 0.15
2.78 X 10-5 o.31
2.15 X 10-5 0.22
possible steam fracture
• The efficiency! of steam stripping of kerosene (and gasoline) from
various soil types was determined in the same cells where the steam
permeabilities were determined. Results are shown in Figures 2 and 3.
• A. steady state analytical model was developed where steam flowed
upward to the collection cap from pipes embedded in the soil (see
Figure 4 for the model) . The model, used together with the
permeability and stripping efficiency data described previously,
allowed a determination of the time to decontaminate a given kerosene
spill. The results using the model are given in detail in reference 3.
• A small scale model of the geosynthetic cap was used to determine its
feasibility as a cover assembly during steam stripping. The cap
consists of a geotextile and geomembrane which is anchor-trenched to
the soil. A schematic diagram of the experimental setup is shown in
Efficiency means the rate and degree of removal of the contaminant.
126
-------
Figure 5. Results are shown in Figure 6 for vacuum-assisted steam
stripping in beach sand for various levels of kerosene saturation.
(Dimensions given on diagram.) No noticeable escape of steam or
kerosene was detected during the experiments, indicating the cap
performed well in containing and collecting the steam and kerosene.
The work in Phase I pointed out at least three definite needs:
• Our- "chemical analytical technique" involved separation of the water
and the kerosene (or gasoline) and measuring the amounts of each
volumetrically. Material lost to outgassing and that condensed in the
lines is not counted as output and hence a more precise analytical
technique is needed. A major effort is now underway to determine the
amount of material remaining in the soil after vacuum-assisted steam
stripping. Gas chromatography (GC) is the technique being used.
• Kerosene and gasoline are mixtures of very many compounds of widely
varying properties. Therefore, in order to understand the process more
completely, it is desirable to determine the efficiency of
vacuum-assisted, steam stripping of individual compounds. This is now
being pursued with a series of alkanes of differing vapor pressures,
and also a series of polar organics of differing polarities. It is
felt that vapor pressure and polarity are two very important
parameters in determining the ability to steam strip a particular
compound.
• The method shown in Figures 1 and 4 is only one possible means of
injecting the steam. Also to be considered (more in keeping with the
field work performed so far) is the flow of steam from vertical pipes
with perforations over part of their lengths. It is possible that a
vertical vacuum tube(s) will be placed in the soil for steam and
contaminant collection. Here the cap will be a secondary collection
system and mainly function to reduce any unwanted Vaporizations tinto the
ambient.
The first two points (GC and simple compound studies) are addressed in
this paper.
PRESENT WORK
RESULTS WITH INDIVIDUAL COMPOUNDS
Volumetric Measurement of Steam Stripping Efficiency
Volumetric analysis of the steam stripping ability of various organic
compounds (and kerosene) were determined. Typical results are shown in Figure
7, Here is shown the apparent percentage removal in a 50% sand - 50% silt
soil for dodecane, decane, octane, octanol and kerosene. It is obvious from
Figure 7 (and similar results for the other soil types) that a more rigorous
method of determining stripping efficiency is needed. For example, the
stripping in octane is essentially complete after one hour, therefore the
127
-------
apparent removal efficiency of 40% is certainly artificially low.
Rather than continuing on with detailed accounting for all the chemical
steam stripped, a more simple, reliable, precise method was sought with which
to determine the stripping efficiency. The method chosen was gas
chromatography (GC) analysis of the chemical remaining in the soil after
steam stripping. Results using GC are discussed in the next Section.
Gas Chromatocrraphic Determination of Residual Chemicals in the Sn-i 1
Standard gas chromatographic (GC) techniques were employed. Three-fold
extraction (from the steam-stripped soil samples) with methylene chloride was
used. Some actual GC results for the residual chemical are shown in Table 2
for decane. The results and those to be shown in Figure 8 show that vacuum-
assisted steam stripping is a very rapid and efficient means for removing
organics from soils. The amount of decane remaining in the soil unfortunately
does not follow any trend as a function of silt content. This observation
also applies for all the other chemicals (dodecane, octane and octanol).1
Therefore, for want of a better approach, the overall results for the various
chemicals will be presented as the average (over the five soil types)
remaining in the soil after some particular time period of steam stripping.
Figure 8 gives such a presentation. Here is plotted the average amount
of chemical remaining in the soil after five hours of steam stripping versus
the vapor pressure of the-pure chemical at 100°C. For the alkanes, it is seen
that the removal is a very strong function of the vapor pressure. The low
vapor pressure dodecane is relatively much more difficult to remove than the
high vapor pressure octane. However even dodecane, with a very low vapor
pressure is still reduced to only 0.1% of its initial value in the soil, via
vacuum-assisted steam stripping.
It is interesting to compare our results for the alkane steam stripping
with the theoretical value for the simple steam distillation of a two
component immiscible liquid mixture (10). The theoretical alkane-water ratio
in the steam-distilled mixtures is
alkane
W „
water
M P
alkane alkane
M P
water water
(1)
where
W
alkane
W
water
M
alkane
M
water
"alkane
ratio of weights stripped (distilled)
molecular weight of the alkane
molecular weight of the water
vapor pressure of the pure alkane at the combined boiling point
It may be that more chemical remains in the finer grain soils, but it may be
more difficult to remove with the GC extraction solvent.
128
-------
rH
0
IS
H
Oi
H
B
CO
J2
w *»^
EH f?
W ft
* 2
g &
is
£ o
g CO
nj
s! °
H -Q
g CO
Bfi T|
O £,
g|
Q 5.
H
CO
§
CM
H
CQ
£
h_
?*
-H
G
-H
g
d)
PS
0\°
CO
EH
CO ft
§
-H
O
CO
t7>
O
in
H
3
w
§j3
O
,_fl
0) -H
M O
O CO
J
rH
-H
o
CO
a>
i— i
i*
CO
^ (O CO ^ C3 rH ^d1 C3 C^CM CM
OrHrHrHrHHOrHrHOO
cQooSScM
i-H rH
vorovovor>CMC~c3^'<0o
rH-*C~r~-^inrHin<^rHrH
cocoooooooomm
"* ^ -ST rH rH i-H CM
CMt~rOCMCMOOCOint~COOO
CN]'CvIrocnrororo'a<'<3J
T-irHr^c^-LOLOLOr^r^
rHCMC-OOCO^rHCT^OlO^O
rH rH
»r-4
•rl
o
o
CO
g
m
Cn
a
o
•H
i-H
rH
"2
(D
a
o
M
a)
ft
0)
G
n)
o
0)
•c
o
CO
(d
M
tn
*
>l CO 0
M +J -P
fi (1) 'D
-rl M 1>
H 3 a>
fl> CO C
ft S «
g "H
goo
_« n s
o n
*O O O
s ^*
M1 ri
* G 0
-S-H-.H
,_1 III
0 rH 0
o m
M M
0) CM O
C*^ ' i)
• id
o "
">• g; * -g
•y -H co x!
"^ M (d 4J
•jj g 4J O
3 id ^^
T3 0) /-) t,
0 +J S 3
M CO 5
gj IH -p M-I
o ^
0) 3 >i
X! <1> 0 0
.p o o (d
G (d M
tu m 3
—t (i) rO (d
.3 cu w o>
4-^ rH 4J
(0 rj
O Q) co >i
-H g d) H
O *rH* |-| fl)
G 4J (0
•H 0 -H
m U O
G S 0)
(d M M
^ 3 ft
w ° „,
>°^ §
t_ (~{ £j
o **
f^ -H a) qj
•P +J -H
rH -rH -H Ti
129
-------
"water
= vapor pressure of pure water at the combined boiling point
The very simple argument will be tested here, that the amount of alkane
remaining in the soil will be inversely proportional to the theoretical
weight ratio of alkane to water (that is, the larger ratio favors quicker
removal). Table 3 gives this theoretical ratio (Eqn. 1), the inverse ratio of
the theoretical alkane weights (based on octane) and the experimental ratio
of remaining chemical in the soil. The trend (comparing the second and third
columns of Table 3) is certainly in the correct direction, but relative sizes
are considerably off, i.e., the strict ideas of steam distillation of
immiscible liquid mixtures does not apply to steam stipping of alkanes from
soil. This is certainly not surprising or unexpected as the steam stripping
in soils is a much more complicated process.
In the case of the alkane-based alcohols (which have some solubility in
water), it is seen that low vapor pressure compounds (e.g. octanol) can be
very effectively steam stripped from a wide variety of soils. The butanol has
been reduced below detection limits (1 ppm) by the steam stripping technique.
The reason for the very efficient removal of the alcohols (versus the
alkanes) is not understood at present and needs to be investigated further.
Octanol is mildly polar (dielectric constant = 3.4) and butanol is quite
strongly polar (diel. const. = 20). The degree of polarity is certainly one
of the major differences between the alcohols and the alkanes. ^
CONCLUSIONS
It appears that vacuum-assisted, steam stripping of a number of organic
chemicals, in particular alkanes and alkane-based alcohols from a wide
variety of soils is quite feasible. The residual chemical (in most cases) can
be reduced to below 100 ppm in a relatively short treatment time. It is hoped
that the work presented here leads to a better understanding of the mechanism
and future potential of the process in the very important area of in-situ
soil decontamination. This research has showed the importance of vapor
pressure and polarity in determining the steam stripping ability.
The unique geosynthetic cap comprised of a geotextile, for rapid lateral
flow of steam and contaminant, and a geomembrane, for containment of the
steam and contaminant, appears to act quite effectively. When anchored into
the soil at the edges and connected to a vacuum, the geosynthetic cap
performed well in the steam stripping of kerosene from sand.
More research work remains to be done. In particular:
• More chemicals need to be investigated to determine the general
ability of steam stripping.
• A clay fraction should be added to the soils - clays will bind certain
chemicals more than the coarser grained materials.
•••It may be that the alcohols are more difficult to extract (with the GC
solvent) from the soils than the alkanes - and hence appear (artificially) to
be more effectively removed.
130
-------
• The GC methods 's reliability in regard to the steam-stripped samples
must be pursued further.
• Work should proceed in developing the unique geosynthetic cap assembly
- and testing it on a larger scale.
• The use of the technique on excavated soils should be considered.
REFERENCES
1. Kovalic, J. M. and Klucsik, J. F., "Loathing for Landfills Sets Stage for
Innovative Hazardous Waste Treatment Technology," Hazard. Mat. and Waste
Manag. £, 1987, 17-18.
2. Cheremisinoff , P. N., "Update:
M, Feb. 1987, 42-49.
Hazardous Waste Treatment," Pollut. Eng.
3. Lord, A. E., Jr., Koerner, R. M. and Murphy, V. P., "Laboratory Studies of
Vacuum-Assisted Steam Stripping or Organic Contaminants from Soil, " Proc.
14th Annual _ Conf. on Land Disposal. Remedial Action and Treatment of
Hazardous Waste, Cincinnati, Ohio, April, 1988, sponsored by the Risk
Reduction Engineering Laboratory, U.S. Environmental Protection Agency,
Cincinnati, Ohio.
4. Hilberts, B. (1985), "In-Situ Steam Stripping," Assink, J. W. , Van Den
Brink, W. J., Eds., Contaminated Soil, Proc. of 1st Intern. TNO Conf. on
Contaminated Soil, Utrecht, The Netherlands, Nov. 11-15, 1985, pp.
680-687.
5. Baker, R., Steinke, J., Manchak, F., Jr., and Ghassemi, M., "In-Situ
Treatment for Site Remediation, " Proc. Third Annual Conference on
Hazardous Waste Law and Management, Seattle, Wash., October 27 and 30,
1986, and Portland, Oregon, October 31 and November 1, 1986.
6. Baum, R., short article describing process appearing in Chemical and
Engineering News, December 12, 1988, pp. 24-25. Work done by K. Udell, J.
Hunt, and N. Sitar at University of California, Berkeley and A. Nagtel of
Solvent Services, Inc., San Jose, CA. Also described by Kelley, K. P. in
Haz. Mat. World, January 1989, pp. 12-14.
7. Advertisement of GeoCon Inc., Pittsburgh, PA, - appearing in Hazardous
Material Control, Volume 1, #4, July-August 1988.
8. Lord, A. E., Jr., Koerner, R. M., Murphy, V. P. and Brugger, J. E.,
"In-Situ, Vacuum-Assisted, Steam Stripping of Contaminants from Soil, "
Proc. - af — Superfund _ '87r 8th National Conference on Management of
Uncontrolled Hazardous Waste Sites, November 1987, Washington, DC, pp.
390-385. Sponsored by the Hazardous Materials Control Research Institute,
Silver Spring, MD.
131
-------
9. Murphy, Vincent, P., "In-Situ, Vacuum-Assisted, Steam-Stripping to Remove
Volatile Pollutant from Contaminated Soils," Masters Thesis in Civil
Engineering, Drexel University, Phildelphia, PA, June 1988.
lO.Prutton, C. F. and Maron, S. H., Physical Chemistry, MacMillan Co., NY
(1975), pp. 175-177.
ACKNOWLEDGEMENTS
This project is funded by the U.S. Environmental Protection Agency under
Cooperative Agreement CR-813022-01. The Drexel authors offer our sincere
appreciation to the Agency for their support. Thanks ar due Dr. Frank Davis,
Ping Chen and Bang Chi Chen of the Chemistry Department at Drexel for their
involvement with the GC work.
TABLE 3. THEORETICAL WEIGHT RATIOS OF ALKANE TO WATER IN STEAM DISTILLATION
COMPARED TO RESIDUAL ALKANE IN SOIL AFTER STEAM STRIPPING
W
alkane
W
octane
W
water
W
alkane j
W
alkane
W
octane j
theor.
theor.
expt'1.
Boiling Point of
(Alkane/Water) Mixture
octane 2.83
decane 0.86
dodecane 0.39
1
3.30
7.25
1
8
100
90°C
97.5°C
99.5°C
132
-------
«HINJECTION PIPCS
PIPE MANIFOLD SYSTEM —a
PLAN VIEW
VALVES
VACUUM COIUCTIOB
MCHDBTRCKH
riOUlU HCMIMM LIMU
4 MIOUB WINWD»E« CUICXTIIC
C— 81 tUCOHmiTC
t -A t
HUM BltTILlATIUH
IDUICIION PlPta
. CRQUIM WA1ER TMLE
ELEVATION VIEW
Figure 1 - Schematic Diagram of Proposed In-Situ Vacuum-Assisted Steam
Stripping Field Apparatus
100
z
Ui
tn
o
200
400 600
TIME (minutes)
eoo 1000
Figure 2 - Steam Stripping Efficiencies of Kerosene from the Various Soils in
the One-Dimensional Cells (Volumetric Analysis)
133
-------
STEAM STRIPPING FOR GASOLINE
• 0
50 100 ISO
TIME (min)
200
Figure 3 - Steam Stripping Efficiencies of Gasoline from Two Soil Types
(Volumetric Analysis),
ground surface
flow lines
equi-pressure
contours
steam pipe
Figure 4 -Model Used in Steam Stripping Calculations
134
-------
Contaminated
Soil
To Condenser
and Vacuum
Clean Soil
Geosynthetic Cap
'(Geotextile and Geomembrane)
m
••
14"xl4"
Soil Box
Steam in
Figure 5 - Schematic Diagram of Pilot Scale Experiment for Vacuum-Assisted,
Steam Stripping Using the Geosynthetic Cap
100
I
O
I
K
I
e
10% Saturation
25% Saturation
50% Saturation
1000
TIME(min)
2000
Figure 6 - Results for Steam Stripping Kerosene Contaminated Beach Sand Using
Pilot Scale Geosynthetic Cap of Figure 5 (Volumetric Analysis)
135
-------
REMOVAL EFFICIENCY FOR 50% SAND
TIME (hours)
Figure 7 -Results for Steam Stripping Various Chemicals from 50% Sand/50%
Silt Mixture (Volumetric Analysis)
SOIL STEAM STRIPPED
FOR 5 HOURS
•§• 1000
CL
Q.
HI
O
O
100-
O
E
ui
HI
cc
LU
8
cc
UI
Dodocane
Butanol
0 100 200 300 400 500
VAPOR PRESSURE AT 100 C (mm Hg)
Figure 8 - Results of Average Residual Chemical in Soil After Steam Stripping
for 5 Hours (Gas Chromatography Analysis)
136
-------
LOW TEMPERATURE THERMAL DESORPTION FOR
TREATMENT OF CONTAMINATED SOILS
PHASE II RESULTS
Richard P. Lauch and Robert C. Thurnau
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
Ed AT perin and Arie Groen
IT Corporation
Knoxville, TN 37923
Barbara B. Locke and Catherine D. Chambers
PEI Associates, Inc.
Cincinnati, OH 45246
ABSTRACT
Performance of the low temperature thermal desorption process for re-
moving volatile contaminants from soils was evaluated. The data obtained
were necessary to assist EPA in the study of alternatives for treating Super-
fund soils. Soils from two Superfund sites were selected for treatment. The
effect of temperature and residence time was determined using a tray-type
furnace. Temperatures of 350°F and 550°F, and residence time of 30 minutes
were tested. The differences in concentration before and after treatment of
volatile and semivolatile organic compounds were used as a measure of treat-
ment effectiveness. Metal concentrations before and after treatment were
also determined. Results from these tests on actual Superfund soils (Phase
II) were also compared to earlier results of tests on synthetic soils (Phase
I). The Phase II results showed that over 87 percent of volatile organic
compounds, and over 79 percent of semivolatile organic compounds were removed
at the 550°F temperature.
INTRODUCTION
The thermal treatment of solids has been practiced for many years to
effect a chemical change in the solid or to separate components based o.n a
physical property such as vapor pressure. The application of thermal treat-
ment to hazardous waste problems has utilized both physical and chemical
processes to decontaminate soils or other solids containing hazardous
constituents.
Historically, thermal treatment has been most commonly practiced in
direct-fired incineration systems which heat contaminated solids to high
temperatures. These systems are effectively used to decontaminate solids
that contain hazardous organic compounds. The organic constituents on the
137
-------
heated solids are removed and/or rendered nonhazardous through a combination
of chemical reactions and physical transformations. The heated solids are
discharged from the incineration system after the organic contamination has
been removed. The organic compounds removed from the solids are then de-
stroyed. Treatment of contaminated solids by this technique is currently
practiced on a commercial scale on a wide variety of solid waste problems.
The thermal desorption process takes advantage of thermal driving forces
to remove organic contamination while avoiding typical incineration process-
ing conditions which are expensive or have negative public perception.
Thermal desorption is conducted at lower operating temperatures, offering
significant fuel savings over high temperature incineration. The heat re-
quired for thermal desorption is provided by indirect heating of the soils as
opposed to direct-fired heating of solids in an incineration process. This
greatly reduces the quantity of off-gases which must be cleaned prior to
discharge. This design aspect not only reduces the cost of subsequent air
pollution control but also facilitates the design of a closed system with no
visible plume. Thus treatment of contaminated soils by thermal desorption is
potentially more cost effective on low level organically contaminated soils.
Thermal desorption has been successfully tested, at both the bench and
pilot scale, on a wide range of solids contaminated with organics (EG&6 1988;
IT Corp., 1986). The organic compounds that have been successfully removed
from different soil types include: polynuclear aromatic compounds from soils
contaminated by coal gas manufacturing plants, polychlorinated biphenyls
(PCBs) from soils contaminated by spills of oils, and priority pollutant list
compounds from surrogate soils spiked with these compounds (PEI and IT 1987,
Szabo, et al. 1988). In these tests the technology was found to be effective
in removing the organic contamination to the desired levels. The specific
treatment conditions for these compounds varied according to the chemical and
physical properties of each contaminant and the matrix containing the con-
tamination.
Because the nature of the process is physical and chemical, it lends
itself well to widely varying soil types and soil characteristics as might be
found at Superfund sites. Often the nature of the soils found at Superfund
sites varies in contaminant depth, geographical site location, and existing
mineralogical conditions. The types of contaminants also vary widely but
they are usually compounds with moderate vapor pressures. Compounds with
lower vapor pressures can also be removed using higher treatment tempera-
tures. Figure 1 provides a general idea at what temperature several organic
compounds will vaporize; however, vapor pressures will vary depending on
whether the compound is the sole contaminant or there is a mixture of con-
taminants. It is interesting to note that the final results of this study
did, for the most part, show the highest percent removals for compounds with
the highest vapor pressures (see Figure 1).
It is the ability to treat a variety of soils with varying types and
levels of contamination without other pretreatment that makes thermal desorp-
tion an attractive technology for treatment of Superfund soils. The treated
soil often requires no further treatment and can be immediately returned to
the site.
138
-------
en
X
E
E,
UJ
rr
ID
CO
CO
UJ
rr
DL
DC
O
CL
600
760
700 -
600 —
500 -
400
300 -
200
100
= Trichloroethylene
= Toluene
Q = Tetrachloroethylene
^ = Pentachlorobenzene
| | = Hexachlorobenzene
TEMPERATURE, °C(°F)
Figure 1. Vapor pressure of selected volatile and
semivolatile organic compounds.
An additional design aspect of thermal desorption which facilitates
application to the wide range of Superfund site problems is the flexibility
to desian the off-gas treatment for specific site requirements. The off-gas
from the thermal desorber can be incinerated in a high temperature secondary
combustion chamber if it is desirable to destroy the organic contamination. .
EXPERIMENTAL DESIGN
In the Phase I study, conducted from May to September, 1987, the effec^
tiveness of low temperature thermal desorption on spiked synthetic soils was
evaluated. Samples of soil were thermally treated in static trays in an
electric oven for specified periods of time. Removal of organic constituents
was measured after treatment at different test temperatures for a duration of
30 minutes. The effectiveness of thermal treatment was measured by analysis
of the treated residue for the known contaminants in the soil samples.
The experimental approach for Phase II testing was similar to Phase I.
The test runs for Phase II consisted of each Superfund site soil being treated
139
-------
at two different temperatures (350°F and 550°F). Each temperature was eval-
uated by conducting four test runs. A test run consisted of two separate
batch runs, with the residue from the two batches being composited into a
single test run sample.
The experimental procedures for Phase II testing are described in the
following Experimental Apparatus and Experimental Procedure sections. The
soils used were from two Superfund sites: the Berlin-Farro Site and the Old
Mill Site. The Berlin-Farro Site soils consisted generally of glacial till,
while the Old Mill Site soils were a glacial silty clay with sand, gravel,
and boulders. Table 1 presents the physical analysis of the two site soils
used in Phase II and the synthetic matrix used in Phase I. Generally, soils
lower in clay content (and conversely higher in sand content) should be
easier to treat with thermal desorption because the organic contaminants are
not as tightly bound to the sand as to the clay.
TABLE 1. PHYSICAL ANALYSIS OF TEST SOILS
Phase II
Berlin-Farro
Old Mill
Phase I
Synthetic soil
matrix
Coarse sand
(>0.5 mm)
4.2
3.3
28.2
37.2
7
6
Fine sand
(0.05-0.5 mm)
Silt
(0.002-0.05 mm)
Clay
(<0.002 mm)
35.2
36.1
40.3
35.6
48
48
33.4
34.1
22.6
19.6
33
33
27.2
26.5
8.9
7.6
12
13
EXPERIMENTAL APPARATUS
The experimental apparatus used in Phase II was the same equipment used
in Phase I. The primary piece of test equipment was a Lindberg furnace,
Model 51848, with an electronic temperature controller and 1600 watt heater
system. The o^en has double-shell construction with interior surfaces made
of Moldathernr ', a molded aluminum-silicate insulation material. This oven
is capable of operating up to 1100°C and has a relatively fast heat-up rate
due to its low mass. The interior space is approximately 10 cm (3.9-in.)
wide x 11 cm (4.3-in.) high x 21 cm (8.3-in. deep). A loose block (1/2 in.
thick) of Moldatherm is placed on the bottom of the oven to provide addi-
tional separation between an object placed in the oven and the hot interior
surface of the oven.
140
-------
The oven was continuously purged during each test by nitrogen from an
Incoloy (3/8 in.) tube inserted through the back wall. The purge gas was
directed against the back wall to promote preheating and distribution. The
purge gas flow rate, which was measured using a standard rotameter, was
maintained at approximately 90 cc per minute. This flow rate resulted in a
complete turnover of the oven environment every 20 minutes.
Two thermocouples were used to measure temperature. One of these was an
NBS traceable, type K, sheathed thermocouple placed approximately 3 cm above
the soil at the center of the oven. This was the thermocouple used to measure
the "test temperature". The other thermocouple was used to measure tempera-
ture within a soil layer during selected test runs. These thermocouples, the
oven temperature indicator, and the purge gas rotameter were calibrated
according to standard engineering practices. Temperatures were recorded
using a Cole Farmer 3-pen recorder, Model 595. The oven, nitrogen purge
line, and thermocouple arrangement are shown in Figure 2.
INTERIOR OF
OVEN CHAMBER
OVEN INDICATOR
THERMOCOUPLE
PURGE
TEST THERMOCOUPLE
SOIL THERMOCOUPLE
GAS EXIT AT DOOR SEAL
Figure 2. Schematic of interior of static tray test oven with the tray inserted.
141
-------
A specially made tray was used to contain the soil within the oven. The
tray which weighed approximately 430 grams, was 8.9 cm (3.5-in.) wide x 3.3
cm (1.3-in.) high x 19.3 cm (7.6-in) long and made of Incoloy to resist
oxidation. A separate Incoloy lid was used to cover the tray while cooling.
TRAY TEST PROCEDURE
A single procedure was established for all thermal treatment tests; four
replicates were run for each sample. The detailed tray test procedure is
delineated in the EPA draft report entitled "Alternative Treatment Technology
Evaluation of CERCLA Soil and Debris." Briefly, the procedure entails trans-
ferring approximately an 80 gram sample of soil to a clean tray and spreading
the soil to achieve a uniform layer on the bottom of the tray, usually about
2.5 to 3 mm deep. The tray is then inserted into the oven at ambient tem-
perature and heated to the target temperature. When the prescribed residence
time at the target temperature is reached, the oven is shut off and the
sample is removed and allowed to cool for about an hour. The sample is
weighed and an aliquot is transferred to a VGA vial. The vial is properly
sealed with a Teflon-lined septum cap and the sample is sent for analysis.
The tray is cleaned and prepared for the next run.
RESULTS
Soil samples from the Berlin-Farro Site and the old Mill Site were
analyzed for all volatile organics listed on the Hazardous Substance List
(HSL). Only the compounds that were detected are reported (see Tables 2 and
3). Only those semivolatile organics detected in the untreated soils were
analyzed for in the treated soils. Berlin-Farro soils were analyzed for the
semivolatile organics listed in Table 2. Samples were also analyzed for
pyrene and bis-(2-ethylhexyl)phthalate; however, the results were inaccurate
because of analytical problems at the laboratory and therefore, were riot
reported in Table 2. No semivolatile organics were detected in the untreated
Old Mill Site soils. Aroclor 1260, however, was present and was analyzed for
in the treated soils (see Table 3). TCLP extracts of both soils were ana-
lyzed for arsenic, barium, chromium, copper, lead, nickel, vanadium, and
zinc.
The results of the thermal treatment test runs show that thermal desorp-
tion is effective in removing organic contaminants from Superfund soils. The
treatment of Berlin-Farro soil is shown in Table 2. The reduction of the
volatile organic compounds was typically greater than 90 percent at tempera-
tures of 350°F and 550°F. The exception to this result was the behavior of
2-butanone at 350°F. The concentration of 2-butanone appears to increase 18
percent from the initial concentration. This compound, however, was detected
in all of the blanks corresponding to analysis of this sample; therefore,
this apparent increase probably resulted from laboratory contamination.
The percent reduction of semivolatile organic compounds from BerlinFarro
soil was slightly lower than for the volatile compounds. This is logical be-
cause of the higher vapor pressure of the volatiles. The analytical results
are summarized in Table 2. Data indicated an increased concentration of
pentachlorobenzene and hexachlorobenzene at the 350°F test temperature. As
142
-------
LU
CS
^C
CfC
UJ
^»
•a:
•»— '
_i
i—i
o
CO
0
a:
C£
^c
u_
l
«£!.
1— <
ce:
LU
CO
u_
o
^
o
1 — 1
1—
o.
a:
0
CO
LtJ
a
wJ
<^
s:
O**
UJ
^L.
H-
LL.
O
CO
J—
— J
,r J
oo
LU
a:
O
1— <
Zp
CO
o
LU
o
^"
Q£
<^
Sfe
it.
•_J
oo
*
C\l
LU
_J
CO
CO
a>
4-j
c
wj
— <
c:
0
4-9
CJ
3
•a
CD
ct:
^
CO
*^^
co
^3.
-1
0
•r—
^ ^
•^
"^-^
co
=1
co
«^£
*^^
CO
3.
•t
^~
•^
o
CO
i.
0)
+J
ro
CO
a.
CM CO CT> «3- •-•
f**^ co cr* cr» cr>
A A
O LO CO »>» LO
CO CM CM CM
V V
*• ^
CO "5j" CM ^O ^*
«-l CO CTi CT» tTi
*— ' A A A
CM
co co co cr» ro
«d- CM CM t-l CM
co v v v
CD f>- O CO f»»
cr> «a- oo co co
CM t— 1 CM «3- CO
co a>
u c
E c .c ^
IO (I) 4J CO
Ol f
S- 4-> O O
O O) O) i_ 4->
coo — •
OJ 0 S- •—
t— c o J= a, co
•r- CO f— (J C O)
4-> +J J= CO O) C
CO 3 O t. 3 O)
r— CQ «r- 4-» r— r—
o i &- a> o >>
> CM t— 1— 1— X
4
f
c
C/)
0
•r—
c:
CO
Ol
S-
o
olatile
(U
CO
5(-
CJ CO LO LO
o o
o o
CO CO
A •
co co
V V
P-. co
o
0 0
CO LO
•>!- O
CO
o
o o
O 0
en o
•V A
T— 1 ^O
•sf
0)
c
0!
•f—
T3
ro
ro
C X
Ol (U
Q- 3C
OJ
3
CO
'o
>1
•a
0)
a>
S-
S-
o
1 *
o
«
c
o
•f—
CO
[ ^
c
(U
o
c
o
o
r—
to
•^
^_J
•r—
C
•r-
ps
O
i.
**~
d)
CO
c
CO
^Z
u
^5
CO
CO
"Q
O)
4->
O
a.
QJ
s_
o
•r-
4->
O
3
•o
OJ
C£
i— 1
•
t/1
O)
CO
a>
^
4-»
C
CD
CO
a
^
•r—
C
O
CO
2!
ro
C
O
•^~
4^
ro
s_
4->
C
QJ
U
C
O
u
c
reases i
u
c
1— <
•
t"
rO
^.
y^
0
CO
S-
o
**^
ro
QJ
g^
4^
C
•r-
•o
QJ
0
(U
(U
HO
c
o
a.
o
o
CM
t
>.
s-
o
4->
CO
s_
o
/-I
CO
r—
Q.
_J
CJ
a>
4_>
^3
QJ
C
QJ
to
QJ
S-
o.
CO
CO
CO
3
CO
QJ
4J
CO
•r*
CO
LU
co
• A
t*
o
CO
o
-Q
ro
QJ
-t-5
^^
.a
•o
QJ
•a
0)
ai
o
X
QJ
CO
ro
3-
CO
CO
*^
ro
C
ro
QJ
J3
tj
O
4.)
•r-
c
•r—
C
O
1 <
o
QJ
(1)
HO
O •
J= -0
4-J Q,
QJ 4->
E CO
QJ 3
J= U
1— r-
CO
O
HO a,
(U J3
^ ^
CO 4->
i — O
3 C
U C
r— ro
ro O
0
' C
4-> O
O-r-
C 4J
U
II 3
O O)
«3-
143
-------
LU
CD
§
LU
^>
2.
_J
O
CO
7;
i— i
2:
o
_J
o
Lu
O
z:
0
»— i
i—
o_
ce:
o
CO
LU
0
^
ex.
LU
n:
|— .
Lu
O
00
f—
_l
=3
CO
LU
f^
U
•z.
1
0
LU
O
>—
(V^
•a:
§F
^^
co
•
co
LU
^J
CD
cC
i—
c
o
•r~*
O
3
•o
Di
&«
Lu cn
O *s
O *-s>
LO cn
LO a.
t— i
c
o
•r-
«^_>
U
3
"^
(U
C£
LU cn
O -*"*
O "-^
LO cn
CO 3.
cn
*o -^
cn
re 3.
(U
t. «
C *f™
=> o
CO
t-
CO
*4^
fli
(Ss
re
t-
Q-
CM CO CO LO
C7% C7l LO CM
A A
LO LO r^> oo
CM CM LO «5l-
V V
CO O CM O
CM CM r*«* r^^
CO LO CO LO
1*^ CO «* 00
i-H CM
O CM CM O
O LO LO LO
^- CO t-l CM
CM
co cu
u c
•r- CO O) • — .
C C .C t—
re cu 4-> re
cn.c cu ••->
t. -4-3 O O
O CO i. +>
o o - —
CJ J_ i—
•i- r— O C CO
+j J= re co c
re u s- 3 co
r- -i- +J i— r—
O J- CO O >>
s»t-^t-x
LO
CM
A
CO
o"
o
r-4
SM«* '
o
•z.
CM
ff-~^
o
LO
*a^^
o
o
o
CO
o
o
o
CM
CO
u
•r-
C
re
cn
i.
o
CO O
r— LO
••- CM
4"^ i~H
re
i— S-
o o
> r—
•r- 0
E 0
i-
CO «C
CO
3
4J
to
o
£
>^
J2
-o
U
CO
s-
s_
o
o
4_)
o
c
.0
re
.}_>
c
(U
u
c
o
o
^—
re
•r-
^J
•t—
C
•r-
ps
£
**~
CO
cn
c
re
.£=
O
&«
CO
re
•o
CO
j_
o
0.
i.
c
o
•^
4_)
u
3
•o
on
t-H
•
CO
o>
CO
CO
4_)
C
CO
i.
re
Q.
c
•r-
C
O
CO
CO
i.
re
c
o
•f—
re
^
4^
C
CO
o
c
o
u
c
•1—
CO
CO
CO
re
CO
i-
o
c
CM
•
*~**
^^
•r—
E
•f—
|—
C
o
•1^
o
CO
CO
•o
-o
o
.c
^_>
CO
^VM^
CO
•I-*
u
CO
•40
CO
•o
^^
0
II
o
o
t— 1
*H^t
0
CO
144
-------
can be seen in Figure 1, the vapor pressure of these two compounds at 350°F
is low. As the sample dries out in the heating process it looses mass, thus
concentrating any non-volatilized contaminants. The apparent increase in
concentration of these two compounds at 350°F may be a result of this phenom-
enon. It should be noted here that several problems were associated with the
semivolatile analyses, and these data should be used with caution. Specifi-
cally, hold times for all semivolatile analyses were exceeded and method
detection limit requirements were not met.
The test results for the Old Mill soils, presented in Table 3, are
similar to the Berlin-Farro soil data. The volatile organic data trended
toward reduced levels of volatile organics at higher treatment temperatures.
The data for the Old Mill site soils showed an increase in concentration of
Aroclor (PCB) after treatment at 350°F, but greater than 95 percent reduction
at 550°F. Again, the apparent increased concentration of Aroclor at 350°F is
likely a result of the concentrating of the compound as the sample mass
decreased during the heating process. Ardor is successfully volatilized at
550°F.
Tables 4a and 4b summarize the results of TCLP extract metals analyses
for the two site soils. Total metals concentrations in both soils are also
given in the tables. No significant reductions were noted, nor was reduction
of metals expected as a result of the low temperature desorption process.
Increases in particular metals concentrations are a result of sample weight
loss as the sample dries out during the heating process. This behavior is
comparable to that described above for the semivolatiles and PCBs at 350°F.
The concentration of metals in the TCLP extract compared to the total con-
centrations of metals in the soil shows that very few metals are leaching out
of the soil and, therefore, further stabilization treatment to bind the
metals would probably not be required for these soils.
The test results for the synthetic soils are presented in Table 5. The
data show that a high percentage of volatiles were removed at lower tempera-
tures and that semivolatiles required the higher temperature of 550°F for
more complete removal.
The experimental results from the tray tests involves the discussion of
three separate issues. These issues are: 1) the Superfund site samples used
in the Phase II tests; 2) the experimental activities; and 3) the analyses of
the thermally treated samples.
The Superfund site soil samples used in the Phase II study were received
at the Illinois Institute of Technology Research Institute (IITRI) in five-
gallon containers. Inspection of the soil indicated the samples had not been
homogenized (presence of roots, stones, and discontinuity of color) prior to
shipment to IITRI. A decision was made at that time to not homogenize the
samples in order to prevent the loss of volatile components.
The experimental activities for the Phase II investigation used the same
experimental approach and similar experimental procedures as Phase I and
other EPA-sponsored thermal treatment evaluations. The experimental program
145
-------
TABLE 4a. SUMMARY OF METALS ANALYSES FOR
THERMALLY TREATED BERLIN-FARRO SITE SOILS
Parameter
Arsenic
Barium
Chromium
Copper
Lead
Nickel
Vanadium
Zinc
Total analysis
untreated
(mg/kg)
13
83
36
'117
32
22
20
63
TCLP
Untreated
0.004
0.331
0.015
0.257
0.027
0.087
ND (0.014)3
0.157
Concentration
350DF
0.0092
0.359
0.013
0.317
0.03
0.104
ND (0.014)3
:7 0.223
(mg/1)
550" F
0.0042
0.318
0.023
0.2
0.025
0.097
ND (0.014)3
0.499
TABLE 4b. SUMMARY OF METALS ANALYSES FOR
THERMALLY TREATED OLD MILL SITE SOILS
Parameter
Arsenic
Barium
Chromium
Copper
Lead
Nickel
Vanadium
Zinc
Total analysis
untreated
(mg/kg)
7.0
45
8.5
71
65
9.0
5.0
169
TCLP
Untreated
. 0.007
0.337
ND (0.01)3
0.065
0.103
0.046
ND (0.014)3
1.26
Concentration
350°F
0.0092
0.258
ND (0.01)3
0.089
0.038
0.064
ND (0.014)3
0.921
(mg/1)
550°F
0.0052
0.343
ND (0.01)3
0.092
0.074
0.096
ND (0.014)3
1.80
w
2 Detected in blank.
q
ND (value) = Not detected (method detection limit).
146
-------
LU
1—
LU
QC
_J
__J :
1
LU
t-
O tt>
U_ O)
(O
to s-
LU O>
to >
>- res
2: _J
o
o to
z: o
0 P
rv* LU
SI
CD >-
CO
a;
«=C
E: .,
to
LO
LU
CQ
^£
• '
t
U
2 C
•o o
o ^^
o *>».
LO O)
LO a.
u
=3
T3 C
OJ O
^ '•£
*S
LL. O>
O ^^
CD "-^
LO O>
CO 3.
,
U
3 C
•a o
QJ *r™
Q£ <4~'
^
U_ CD
O -^
CD -•*.
LO C7)
r-l a
CD
•0 J^
+-> o>
«™<
»^
(I)
S- •>
=> o
I/)
S-
OJ
0)
E
(O
J-
o.
CO CTk CTk CTt O*t CT) CT>
CTt CT) CTk CX> O> CTt C71)
A A A A
o •«* LO o*-< «— i in
CD V V f*"*.^-"^ CM
T-H LO LO
•t ^ -"- -
i-H Q O
.......
co en 01 CD crt CTi en
en en en en en en en
A
o en co o o r-i o
O LO CO CO CM LO
f»» «3- f-i- en
* V
o
to
en en en en en en en
A A
o o o o o
O O-l O-« O O
o o *— » o -— - •-* o
•> «LO «LO « »
O CO ^— -*O *— **d" O
CM »-« Q CO Q CM
z 1-1 z co
O O O O O O 0
ooooooo
CD O O O O O O
o CM «a- vo i— i co r>.
CO CM O t— 1 LO CM f^
CO CO CO «—t VO >
-l-> .C
O) 0) <1) -f->
c o c a)
O) i- tt) O.
Nl O N S- O)
C i— C O r—
tt) ^ "5 ^ O) !c 4J
tt) C O -i- C (_) O) fO
^™ O i- CD r*~ tt) O 1 >, S- S_ O) O
4-> tt>i— CM X: >,+J i— >
i— eC t_> <-H LU to (— x* E
O O)
> to
en en LO
en en co
A A
o o o
O CD O
i-H «— 1 O
A A «t
to to co
t-l r-i LO
V V
i-l O CO
CO «* CO
o o o
o o o
o o o
en co LO
LO CM CO
CO LO CM
I— 1 «— 1
•-H r>. co
CM «-»«—!
0 O 0
0 00
o o o
-'• ft fl ft
o o o
O O CO
00 rH CO
A A
LO CM
j
0 CD O
§o o
o o
i-T t-" 7
»-H CM CO
CM LO CO
r-^ CM
^•^ ^M.
r— O
>> c
X OJ
tt> JT
^z o.
i — O
tt) >> O) 1-
C JC •«-> O
tt) 4^ (0 r—
rd 1 ro O
1- CM .C «O
4-> to .C C
C ••- Q. tt)
«C CQ d.
•
+J
c
0
4J
IJ
tt>
"O
•o
o
.c
" -E-
•a
4J
o
tt)
tt)
•a
1 %
o
2:
II
a>
3
1
CD.
frH •'"
147
-------
was carefully structured to obtain the desired level of precision to allow
accurate interpretation of the test results. The Quality Assurance Project
Plan sought to control and measure experimental precision by requiring a
total of four replicate runs for each test condition on each Superfund soil.
The results from Phase II show similar reduction trends for volatile
organics as the Phase I results as illustrated in Figure 3. Treatment of the
actual Superfund site soils (Phase II) was not as efficient as the treatment
of the synthetic soils (Phase I). The surrogate soils, however, were highly
contaminated; a high percentage removal is easier to obtain when the original
concentration is very high. Furthermore, the difference in removal
efficiency is likely to be related to the natural aging and weathering of the
contaminated Superfund soils used in Phase II. The synthetic soils used in
Phase I were spiked and then treated within a relatively short time frame.
Generally, however, volatile organic compounds were readily removed, even at
350°F. The differences in percent reduction for some of the organic results
are due to differences in the detection limits reported by the analytical
laboratory. Higher detection limits mathematically reduce the percent
reduction for specific analytes.
Although the quality of the semivolatile data was lacking, the trend for
semivolatile removal was very similar for both Phase I (surrogate soil) and
Phase II (actual Superfund soil). Semivolatiles required treatment at 550°F.
Figure 4 shows the percent reduction in semivolatile organics compared between
Phase I and Phase II. The data points from the Phase II tests bracket the
line drawn to represent the results of the Phase I test and illustrate the
similarity in effectiveness of thermal desorption in both tests.
CONCLUSIONS
Low temperature thermal desorption was tested at the bench scale on two
different Superfund site soils. Tests were run at 350°F and 550°F, residuals
were analyzed for volatile and semivolatile organic compounds, and TCLP ex-
tracts were analyzed for metals. The results indicate that low temperature
thermal desorption removes over 86 percent of volatileorganics at 350°F and
over 87 percent at 550°F. An average of over 79 percent of semivolatile
organics were removed at 550°F; semivolatile organic reduction at 350°F was
inconsistent and highly related to the vapor pressure of the compounds.
Analyses indicate that this treatment process is not effective in reducing
the concentration of metals.
The results of the Phase II testing (Superfund site soils) of low tem-
perature thermal desorption are similar and seem to support the Phase I
results (synthetic soil matrix). That is, significant reduction of volatile
and semivolatile organic compounds in contaminated soils was accomplished by
the low temperature thermal desorption process. As was expected, thermal
desorption appeared to be more effective in treating the synthetic soil
matrix (Phase I) than the actual Superfund soils (Phase II). This behavior
is likely a result of the natural aging, weathering, and biological processes
to which the Superfund soils were exposed.
148
-------
0>
•5 100.
_CO
5
CO
•5
o>
o>
CO
CO
D
111
oc
95
90
85
80
Synthetic Soil #1
(high organic/low metal)
Berlin-Farro Site Soils
Old Mill Site Soils
I I I I I I I I I I 1 I I
100 200 300 400 500 600 700
TEMPERATURE, °F
Figure 3. Reduction efficiency of low temperature thermal desorption
as a function of temperature for volatile organics.
*-^ 100-
-------
Because sample hold times were exceeded and method detection limits for
semivolatile analyses were exceedingly high, the quality of the semivolatile
data is questionable. The trends for reduction that were indicated, however,
are smiliar to those seen in Phase I of this study. Volatile organic, PCB,
and TCLP extract analyses all met the quality assurance requirements for this
project.
Bench scale tests of low temperature thermal desorption indicate that
this technology is promising for effective treatment of soils contaminated
with volatile and semivolatile orgam'cs and PCBs. Further investigation of
this technology at the engineering or pilot scale may be warranted.
REFERENCES
EG&6 Idaho. "Thermal Desorption/UV Photolysis Process Research, Testing and
Evaluation Performed at Johnston Island for the U.S. Air Force Installation
Restoration Program". Document No. AD-A195613. 1988.
IT Corporation. Laboratory Investigation of Thermal Treatment of Soil Con-
taminated with 2,3,7,8-TCDD, U.S. Environmental Protection Agency, Edison,
New Jersey. July 1984.
IT Corporation. "Technology Demonstration of Thermal Desorption/UV Photo-
lysis Process for Decontaminating Soils Containing Herbicide Orange".
Presented at the Spring 1986 American Chemical Society Conference, New York,
New York. 1986.
PEI Associates, Inc. and IT Corporation. "Low-Temperature Thermal Treatment
of Surrogate CERCLA Soils - Bench-Scale Tests." Prepared for U.S. EPA Risk
Reduction Engineering Laboratory, Cincinnati, Ohio under Contract No.
68-03-3389, October 1987.
PEI Associates, Inc. and IT Corporation. "Alternative Treatment Technology
Evaluation of CERCLA Soils and Debris (Phase II Results), Draft Report."
Prepared for U.S. EPA Risk Reduction Engineering Laboratory, Cincinnati,
Ohio, under Contract No. 68-03-3389. February 1989.
Perry, R.H. and D. Green, (ed.). Perry's Chemical Engineers' Handbook.
edition. McGraw-Hill Book Company, New York, New York. 1984.
6th
Szabo, M. F., R. D. Fox, and R. C. Thurnau. "Application of Low-Temperature
Thermal Treatment to CERCLA Soils", Proceedings 14th Annual Hazardous Waste
Research Symposium, Cincinnati, Ohio. May, 1988.
150
-------
DETECTION OF MACRO DEFECTS IN SOIL-BENTONITE CUTOFF WALLS
by
Andrew Bodocsi, Richard M. McCandless and Koon Wah Ling
University of Cincinnati
Department of Civil & Environmental Engineering
Cincinnati, Ohio 45224
ABSTRACT
Two methods for the detection of buried hydraulic defects
or "windows" in cutoff walls are described. The first method
can be used with existing cutoff walls, and is based on the
monitoring of groundwater levels, employing a system of
standpipes spaced at intervals along the wall. The method was
tested with a model cutoff wall, where a series of standpipes
were used to measure the drawdown curves for various
artificial "windows" of known geometry and location. The same
conditions were then examined using a numerical groundwater
transport model in order to validate the numerical model for
use in a parameter study of field-scale "window" detection
cases.
Preliminary results include calculations of the maximum
field spacing of standpipes that can be used to detect point
source leaks or "windows" of specified size or discharge rate
for a) varying differential head conditions across the cutoff
wall, and b) varying ratios of horizontal to vertical
permeability of the in situ soils adjacent to the cutoff wall.
The second method is applicable to cutoff walls under
construction. The detection system in this case included an
impervious liner on the downstream side of .the wall,
perforated vertical standpipes attached to the upstream side
of the liner at intervals, and strips of geotextile that were
wrapped around the standpipes and were extended and attached
to the liner on each side of each standpipe. This system was
capable of detecting the approximate location of a "window" in
the model cutoff wall, and the existence and approximate
location of a tear or a hole in the liner.
151
-------
INTRODUCTION
Soil-bentonite cutoff walls may be used to contain
leachate generated at hazardous waste sites as a temporary
containment measure alone, or as a part of permanent remedial
work. A preceding study by these authors (1987) showed that
during the placement of the soil-bentonite backfill into a
model cutoff trench, bentonite slurry may be entrapped in a
recessed corner (at the base of a step) or in narrow fissures
in the bottom of the trench. The study also demonstrated that
sediment on the bottom of the cutoff trench will most likely
be entrapped by the advancing soil-bentonite backfill and
result in highly pervious sand "windows" in a cutoff wall.
According to Evans et. al., (1985) there is also a high
likelihood that pockets of bentonite slurry are entrapped at
many locations within the backfill during construction.
Regardless of the mechanism, as-built or long-term chemically-
induced hydraulic defects, or "windows", in cutoff walls
represent continued environmental risk.
This paper describes the detection of buried "windows" by
two distinct techniques. The first method can be used with
existing cutoff walls and is based on monitoring the drawdown
in groundwater levels using standpipes spaced at intervals
along the wall. In this research the piezometric drawdown
curves for various artificial "windows" of known location and
size were measured for a model cutoff wall. The same
conditions were then examined using the MODFLOW Groundwater
Transport Model (McDonald and Harbaugh, 1988) in order to
validate the numerical model. Once validated, MODFLOW was
used for a parameter study of field-scale window detection
cases. Results are reported as the maximum field standpipe
spacings that can be used to detect in situ point-source
"windows" of specified size or leakage discharge for various
differential head conditions across the wall, and for varying
site soil permeability conditions.
The second method tested could be installed on the
downstream side of new cutoff walls during construction. The
system modeled consists of a continuous impervious membrane
liner, perforated standpipes attached to the upstream side of
the liner, and strips of filter fabric wrapped around the
standpipes and also attached to the liner. The system can be
used to locate "windows" in the cutoff wall and holes or tears
in the liner.
152
-------
WINDOW DETECTION BY DRAWDOWN MEASUREMENTS
ADJACENT TO AN EXISTING CUTOFF WALL
EQUIPMENT, PROCEDURES AND RESULTS
Figure 1 is an end view of the flume apparatus used in
this phase of the testing. In this model, a solid plexiglass
partition-" was used to simulate an impervious soil-bentonite
cutoff wall. The partition was provided with orifice plates
at two locations to create various size "windows", or holes.
Each orifice plate contained sixteen 23.8 mm (15/16 in.)
diameter holes, and each hole could be independently opened or
closed to give a specified size "window" opening. A medium
masonry sand with a horizontal permeability of 2 x 10~2 cm/sec
was used to simulate the in situ site sand outside the
hazardous waste containment area. The water table in the sand
represented the unpolluted groundwater table. Flow through
the orifice plate "windows" was caused by differences in
hydraulic head across the partition (model cutoff wall) that
were controlled by overflow ports in the downstream reservoir.
Several 12.7 mm (1/2 in.) o.d. perforated standpipes were
placed within the sand adjacent to the plexiglass partition
(simulated cutoff wall) as also shown in Figure 1. The water
levels in the pipes were measured by a 2 mm (1/8 in.) o.d.
flexible nylon tube lowered into the pipes and connected to a
pressure measuring transducer on the other end.
Downstream reservoir Upstream reservoir
Site Zone (Sand)
152 mm
30S mm
7fl mm
Overflow
port «
(typical)
a 10
o
~l\l: Stondplpe • .' * '•.,
{Varied locations) '.-
Continuous water
- supply
. Constant head
overflow port
Perforated Inflow
panel
Drain fitting
Note:. 25.4 mm " \'
Figure 1. End View of Window Detection Model for the Case of
Existing Cutoff Walls.
153
-------
Numerous drawdown tests were conducted on this model with
a variety of combinations of window openings and differential
head conditions. Typical of the experimental drawdown curves
is the curve shown in Figure 2. The curve drawn through the
triangle symbols was measured for a composite window opening
(one orifice plate) of 51.5 sq. cm. (8.0 sq. in.) and a head
difference of 190 mm (7.5 in.). The curve drawn through the
asterisk •:symbols (*) was computed for" the same conditions
using MODFLOW. Figure 3 shows similar results but for a
double window case (both orifice plates used).
Having close agreement between the experimental and
numerical results as shown in Figures 2 and 3, the numerical
model was then used for a more extensive parameter study of
field-scale "window" detection scenarios. Specifically, the
MODFLOW model was applied to varying size "window" openings,
various differential head conditions, varying ratios of
horizontal to vertical permeability of the site sand (kh/kv),
and different ratios of the permeability of the "window -
forming" material to that of the site sand (kw/ks). For each
set of chosen parameters, the field drawdown curve was
computed and then plotted to scale. Using a practical
accuracy of +/- 2.5 cm (+/-. 1 in.) for standpipe readings
measured under field conditions, and the graphical
construction technique shown in Figure 4, the maximum
allowable horizontal spacing of standpipes along a cutoff wall
was established for each window configuration studied. The
results are summarized both graphically and in tabular form.
Figures 5a through 5d present the maximum standpipe
spacings that could be used and still detect the stated
SOO-i
§450
o
o
Nl
UJ
D_
400-
350-
Notes:
1. window size — SI.5 aq cm
2. differenUol head - 190 mm
***** computed head
/=4AAA observed head
i i i i i 11 | 11 i 111 11 i | 11 i i i i M i |'i i t iTTTi i fi'i'i i i i i i'T f
0 500 1000 1500 2000 2500
DISTANCE ALONG. CUTOFF WALL A30S (mm)
Figure 2. Typ.ical Comparison of Measured and Predicted Drawdown
Curves for a Single "Window".
154
-------
50CH
§450^
o
400-
350
Notes:
1. window size «» 51.5 sq cm
2. differentiol head = 190 mm
3. two windows case
***** computed head
A&&A& observed head
I I I I I I 1 I 1 1 I 1 I I I I 1 I 1 I 1 I I I I I I
1500 2000 2500
DISTANCE ALONG CUTOFF WALL AXIS (mm)
Figure 3. Typical Comparison of Measured and Predicted Pravdown
Curves For Two Adjacent '^Windows11 .
•o
oi
o
o
s
CL
Maximum Standpipe Spacing, S
Ji25 mm
Tangent Line
to Computed
Drawdown Curve
Lower Confidence Limit
for Field Groundwater
Level Measurements
Distance along Cutoff Wall Axis
Figure 4. Graphical Construction Technique to Estimate Maximum Allow-
able Spacing of Standpipes Assuming a Practical Accuracy of
± 2.5 cm (± 1 in.) For Water Level Measurements.
155
-------
1U.U -
CO
0)
Q.
Q.
TJ
a 5.0 -
•*-»
CO
«4-
^J
i '-
s. :
CO
00-
W» \J
--**
,--"•"
«*•*"""'
•**
— — ^"*^
^ — — """"""^
^ * Notes:
_, — ^^ 1. differential head
«* 1.525 m
2. Kw/Ks « 1
— *
^ _- — — -" AWAQ - 20 I/day
j.— — "" sW«***Q « 80 I/day
. _ — — »«"" Q - 160 I/day
*• »*«««Q - 310 I/day
. . A
23456789 10
Kh/Kv of Native Soil
Figure 5a. Maximum Standpipe Spacing For a.Differential Head of
1.525 m (5.0 ft.) to Detect Various Size 'Windows' as a
Function of kh/kv of the Site Soils.
15.0 -
CO
CD
•§10.0
"i
s.
CO
OJ
s.
o
CO
5.0 -
0.0
Figure 5b.
*-
__ — *
Notes:
1. differentia! head
= 3.050 m
2. Kw/Ks » 1
•*••• Q
40 I/day
160 I/day
- 320 l/doy
-, 620 I/day
2
3
4
Kh/Kv
'4'
of
• 1 1 1 1 1 1
6
Native
1 1 1 1 1
7
Soil
1 1 1 1 1
8
,,,,,
1
0
Maximum Standpipe Spacing For a Differential Head of
3.050 m (10.0 ft.) to Detect Various Size 'Windows' as a
Function of kh/kv of'the Site Soils.
156
-------
20.0 -i
w.
Q)
Q.
'5.
TJ
iio.o
•*-»
CO
D)
I
Q.
CO
0.0
___•---*•""
Notes:
1. differential head
-* - 4.575 m
2. Kw/Ks - 1
/VWAQ - 65 I/day
*****Q = 250 l/doy
i Q - 490 I/day
| Q - 930 I/day
TT
2
1 1 1 1 1
3
4
Kh/Kv
5
of
6
Native
•y
Soil
-r-| ri
8
TT'ri
9
i i i 1
10
Figure 5c. Maximum Standpipe Spacing For a Differential Head of
4.575 m (15.0 ft) to Detect Various Size 'Windows' as a
Function of kh/kv of'the Site Soils..
20.0 -
CO
CD
Q.
"5.
•g
Notes:
1. differential head
= 6.100 m
2. Kw/Ks = 1
£10.0 -
o>
'o
a.
CO
0.0
Figure 5d. Maximum Standpipe Spacing For a Differential Head of
6.10 m (20.0 ft.) to Detect Various Size 'Windows' as a
Function of kh/kv of the Site Soils.
*- -
A
1 2
. — - — A- —
Vi""*1"
Kh/Kv
•
5 6
of Native
— — A
7 8
Soil
A/WWQ ~ 85 I/day
MHMM Q «= 650 I/day
»»••• Q « 1250 I/day
9 10
157
-------
CD
I-
w II
a
c o
BJ.H
•P 4J
CO.H
ft
CO
S
fij C
sl
Q)
rH
•S
DIFFERENTIAL HEAD H (m)
o
o
H
*
10
in
c^
in
•
«*
o
in
o
•
CO
in
oa
to
•
H
£~
•3 £
to
ft CO
CO
Discharge
Q (L/day)
2
Is
ta
aco
co
Discharge
Q (L/day)
BU
* A
(0
04 CO
CO
Discharge
Q (L/day)
o>
IS
(0
ftCO
CO
Discharge
Q (L/day)
EC ~
H < &
S w
in
CO CO
CM CO
H H
in o o o o
co r» <* in m
H n vo oj
H
o\
eg in o
H
in
H
in o o o o
vo n in en n
H N ** o\
co in
O H
n vo
H
H
o o o o o
** CO VO
§£
Sfi
£*
r; W
•H C
O BJ
p<
., o
f-l
H
g»ll
c
»fl
:*
o IH
T|
|J|
2
g^
|
fe
[Zj
1*1
H
Q
o
0
,_)
VO
in
in
••*
o
in
o
CO
in
in
H
id
ft CO
CO
a) ~^
P*&
h rt
RJ-O
tt
W ^
•H
Q oe
&
^
•H e
0^
ft CO
CO
a) ^
os
M (d
(OtJ
'•S>
to •— •
•H
Q ty
£_
C ^-»
•Si
ft CO
CO
Q) -^
pn>»
££
•s>
ra v'
•H
o a
&i
c ^
•H g
0-2-
ft CO
CO
<1)^
0>>i
M n)
id -o
^>
01 >-*
•H
Q ~»
8< e
W 0
H § O1
S W
vo vo
M
H
O O O O O
o o o in o
O H H H O
H n ^f co in
H
o vo
eo
CO
«* O CO O\
H H H
o o o o o
in vo n r» vo
r- in H H H
H n vo H
H
** p
n r»
H
vo
r« in
H H
o o o o o
o o o o- **
in o H o •«*
H CM <* t»
rf O H CM
CM CO vo CM
CM «# en co r«
H co
158
-------
leakage in liters/day as a function of kh/kv ratio (ratio of
horizontal to vertical permeability of the site soil). In
each figure, kw/ks equals unity for the general case where
site sand has been trapped to form the "window". Each graph
was prepared for the .case of a single window of varying size
and for differential head conditions that varied from 1.525 m
to 6.1 m (5ft. to 20 ft.), as measured between the prevailing
groundwater table on the outside, and the leachate level
inside the cutoff wall. Note that since the reported
standpipe spacings are for a single window, the spacings are
conservative and would also be sufficient to detect multiple
windows i.e., the presence of more than one window would
broaden the drawdown curve while the spacing of standpipes
remained the same (compare Figures 2 and 3).
Tables 1 and 2 give the maximum standpipe spacings and
flow rates in liters/day for four different assumed head
conditions and five different window sizes. For the results
shown in Table 1, values of kw/ks = 1 and kh/kv = 2 were
assumed. Table 2 presents results for an assumed kw/ks equal
to 10, and kh/kv of 1. The kw/ks ratio of 10 implies that the
entrapped material in the "window" is a bentonite slurry with
very high permeability. When the results in the two tables
are compared, it can be seen that a "slurry window" gives a
much greater flow than a sand-filled "window", and a larger
allowable maximum spacing between standpipes.
As an example of using either the graphs or the tables,
assume that as a result of a field pump test, a "window" that
allows approximately 300 liters/day leakage is believed to
exist at a site where the ratio of horizontal to vertical
permeability of the natural soils is 2.0. Also assume that
there is only one window, that it consists of entrapped sand
(thus kw/ks = 1.0), and that from field measurements the head
difference is known to be 3.0 meters. In order to detect the
location of the "window", it is necessary to know the maximum
allowable spacing of standpipes along the cutoff wall
perimeter. From Figure 5b the maximum spacing of standpipes
that could be used is 7.3 m (24 ft.). A corresponding
"window" size of 1860 sq. cm. is indicated in Table 1. Figure
5b-shows that if a site sand with a kh/kv equal 4 (a typical
ratio for natural stratified soils) prevails at the waste
site, then the maximum standpipe spacing could be increased to
9. m (29 ft.).
It should be noted that the sensitivity of a standpipe
system can be improved by increasing the difference between
the head and tail water levels. This could be achieved in
practice on a periodic monitoring basis by temporarily pumping
down the leachate level in the containment zone.
159
-------
WINDOW DETECTION USING A SYSTEM OF INTERMITTENT
STANDPIPES AND IMPERVIOUS LINER
Figure 6 shows the front and end elevations of a
different window detection model used for this phase of
testing. The end elevation shows a plexiglass wall used to
model an"' impervious liner that has been installed on the
downstream side of a soil-bentonite cutoff wall.. Vertical
standpipes were attached to the upstream side of the liner
(plexiglass wall) at 114 mm (4.5 in.) spacing. The perforated
standpipes were 12.7 mm (1/2 in.) o.d. and extended the full
depth of the model cutoff wall, each having a drain valve at
the bottom. Wrapped around and attached to each pipe was a
50.8 mm (2 in.) wide strip of free-draining geotextile. The
model cutoff wall was 102 mm (4 in.) thick and contained a
50.8 mm (2 in.) diameter by 102 mm (4 in.) long horizontal
sand "window" which extended completely through the cutoff
wall. The upstream side of the cutoff wall was supported by a
perforated plexiglass wall, which separated the wall from the
upstream "leachate" reservoir, but let the "leachate" flow
through. The water in the reservoir represented the leachate
contained by the cutoff wall.
Medium sand window
through backfill
Soll:bentonlte backfill
13mm SS
Vertical
Generalized
plezomeirlc
profile • — _
Indicating
location of
burled window
..
t/e tjrji.
* • • '
\'\:
T?
PP/
•'0 '
•
r "
' '•**7i
]Lf
i * * •
1 ':-]
13 Drain valves
I3i
Front Eievatlc
Fliter-fabr
standpl
I
1; n
\
9
f
— 102— j-7fl-|
.*- Fabric wrapped standplpe
• • *»
: .* •*•
fftSJtff?*
, , * *
• •"*."" * •
AiS&^l';
rf,
~^~
^.Overflow port
Constant head
'"water reservoir
^.Sombentonlte
-^backfill
Medium sand
-"window
ponfll
11 V -"•
3 Drain valve
Note: All dimensions In
25.4mm "1*
End Elevation
Figure 6. Schematic of Standpipe/Liner Window Detection Model for
New Cutoff Walls.
160
-------
The typical window detection test started by draining all
standpipes through their individual drain valves. After
closing the drain valves, the reservoir was filled with water
(model leachate). This initiated flow through the window.
The leachate flowed through the window, but was stopped by the
impervious liner (the plexiglass wall) on the downstream side.
However, very slowly, the leachate would flow laterally along
the interface between the liner and the soil-bentonite
backfill toward the closest geotextile strips and, once
reaching the fabric strips, into the two closest standpipes.
After a few hours elapsed time, the two closest standpipes had
filled up with leachate to the level of the upstream
reservoir, while the other standpipes remained empty, or had
filled to a much lower level. The dashed line in Figure 6
shows the leachate levels in the standpipes at the conclusion
of the test. From the peak in leachate levels in the
standpipes one could estimate the most probable location of a
buried window.
The same system could also be used to detect leaks in the
impervious liner. To demonstrate this, a 6.4 mm (1/4 in.)
diameter hole, representing a tear in the liner, was drilled
in the downstream plexiglass wall. To start a leak detection
test, all standpipes were filled with water to the leachate
level in the upstream reservoir. After a few hours, the water
levels in the standpipes adjacent to the tear were observed to
drop off, while those in the other standpipes remained high.
The geotextile strips adjacent to the tear facilitated
drainage from the standpipes to which they were attached.
Consequently, by finding no drop in the water levels in a row
of standpipes, one could conclude that there were no tears in
the liner. Conversely, if there are drops in the water level
in two adjacent pipes, then one could strongly suspect a tear
or hole in the liner in the zone between the two pipes.
The two general detection scenarios described above would
be initiated either by draining (pumping) the standpipes in
the case of a window in the backfill, or filling the
standpipes to check for a tear in the membrane liner. The
time required to observe the response of the standpipes could
vary from days to weeks depending upon numerous variables
including standpipe spacing, the permeabilities of the site
and backfill soils and the size and location of the defect to
be detected. Future work will focus on the relationship
between these (and other) variables under field scale
conditions.
161
-------
CONCLUSIONS
Cost-effective performance monitoring of soil-behtonite
cutoff walls could be implemented by one of the two techniques
described herein. The spacing of conventional standpipes
required to detect a point-source leak can be determined in
cases where the subsurface geology and hydrology are well
characterized, and where a site-specific design study has
defined a maximum tolerable leakage discharge.
In the case of existing cutoff walls where leaks are
suspected, a series of standpipes installed downgradient
adjacent to a portion of the cutoff wall could be combined
with localized pumping tests to locate the lateral position of
a window within a few meters. The estimated location of the
defect could then be confirmed after seasonal variations in
groundwater levels and other site-specific factors were taken
into account. Once located, the defect could be eliminated by
in situ deep soil mixing or some other suitable technique.
In the case of new cutoff walls, a permanent standpipe/
liner leak detection system could be designed and installed as
part of the long-term remedial strategy. The system .could
also serve as a post-construction Quality Control check to
determine compliance with job specifications.
ACKNOWLEDGEMENT
The research described herein was supported wholly by the
Waste Minimization, Destruction and Disposal Research Division
(WMDDRD) of the U.S.EPA Risk Reduction Engineering Laboratory
(RREL). The authors wish to thank Project Officer Joseph K.
Burkart and Work Assignment Managers Herbert R. Pahren and
Walter E. Grube for their administrative and technical
support. We also thank Graduate Assistants Steve Liatti and
Carl Huntsburger and the UC contract management team at Center
Hill for their contributions to this study.
162
-------
REFERENCES
McCandless, R. M., Bodocsi, A. and Ling, K. W., 1987.
Evaluation of Slurry Cutoff Wall/Permeable Barrier System With
Oraanics. Interim Report: EPA Contract No. 68-03-3379; Work
Assignment #0-1, 75 pp.
Evans, J. C., Lennon, G. P. and Witmer, K. A., Analysis of
Soil-Bentonite Backfill Placement in Slurry Walls; Proceedings
of the Sixth National Conference on the Management of
Uncontrolled Hazardous Waste Sites, Washington, D. C.,
November, 1985.
McDonald, M. G., and Harbaugh, A. W., 1988. A Modular Three-
Dimensional Finite-Difference Ground-Water Flow Model.
(MODFLOW), U.S. Geological Survey Techniques of Water-
Resources Investigation, Book 6, Chapter Al.
163
-------
A FIELD TEST OF HYDRAULIC FRACTURING IN GLACIAL TILL
L.C. Murdoch
Center Hill Research Facility
University of Cincinnati
5995 Center Hill Rd.
Cincinnati, Ohio 45224
ABSTRACT
Hydraulic fracturing, a method of increasing fluid flow within the subsurface,
will improve the effectiveness of several remedial techniques, including pump and
treat, vapor extraction, bio-reclamation, or soil flushing. The basic method is
straightforward: the casing of a well is perforated and fluid is injected until pressures
exceed a critical value, fracturing the material enveloping the well. Sand pumped
into the fracture holds it open and provides a high-permeability channelway suitable
for either delivery or recovery.
Although it has been used to improve the recovery of oil for more than half a
century, hydraulic fracturing has apparently never been used to improve the
remediation of contaminated sites. The goal of this research, therefore, is to
evaluate the feasibility of using hydraulic fractures for remediation. To do so, we
must first establish whether hydraulic fractures can be created under conditions
typical of waste sites. Then, if they can be created, we will assess their use in the
practice of remediation. Our initial investigations consist of theoretical analyses of
both the creation of fractures and the flow of groundwater to or from fractures, and
bench-scale experiments of the hydraulic fracturing process in soil. Recently, we
have conducted a field test of the method at shallow boreholes in an
uncontaminated site.
_ Preliminary results of the bench-scale experiments, the theoretical
investigations, and the field test are all encouraging. In the experiments, clay-rich
colluvium was remolded and consolidated into rectangular blocks of various water
contents. The blocks were loaded in a triaxial cell and hydraulic fractures were
created by injecting dyed fluid through a tube resembling a borehole. Hydraulic
fractures have been successfully created in all tests conducted thus far. Remarkably,
even extremely soft samples of saturated, loosely consolidated clay were readily
fractured in the bench-scale apparatus.
Particularly interesting results have come from the field experiment, in which
ten hydraulic fractures were created at shallow depths (roughly 2m) in a tight, silty-
clay till. Subsequent excavation and mapping have yielded three-dimensional images
of the fractures. In general, they were slightly elongate in plan and they dipped
gently (14° to 25°) toward the borehole. All but two of the tests ended when the
fractures vented at the ground surface. The largest one covered 90 m2 in plan and
extended 13.5 m from the borehole when it vented. More typically, however, the
fractures covered roughly 20 m2 in plan and extended 5 to 8 m from the borehole. A
maximum thickness of 1 cm of sand was observed in the excavated fractures.
164
-------
INTRODUCTION
Removing contaminants from tight soil or rock is a particularly difficult
problem because most available techniques of remediation require fluid flow either
into, or out of, the contaminated region. Where low permeabilities inhibit flow,
remediation will at best be slow, at worst it will be impossible. Hydraulic fracturing
is a method of increasing fluid flow within the subsurface that should improve the
effectiveness of a variety of remedial techniques. Recovery rates in pump and treat,
vapor extraction, or soil flushing systems, for example, could be increased by
hydraulic fracturing. Moreover, hydraulic fractures will facilitate the delivery of
treating materials, such as nutrients for microorganisms in bio-restoration systems.
The basic technique of hydraulic fracturing is straightforward: fluid is
injected into a borehole until pressures exceed a critical value, fracturing the
enveloping material. Sand pumped into the fracture holds it open and provides a
high-permeability channelway for either delivery or recovery. Although it has been
used to improve the recovery of oil for more than half a century, hydraulic fracturing
has apparently never been used to improve the remediation of contaminated sites.
At the Center Hill Research Facility (Department of Civil and
Environmental Engineering, University of Cincinnati), we are evaluating the
feasibility of using hydraulic fractures to remediate contaminated sites. The
feasibility study was funded by the USEPA in May 1987, and we will complete the
study in September 1989;
Theoretical analyses, lab experiments, and field testing comprise the scope of
the currently funded project. We have conducted theoretical analyses of the
processes involved in creating hydraulic fractures and the effects that fractures will
have on the flow of groundwater at a site. According to the theoretical results,
shallow hydraulic fractures are expected to have a sheet-like form, measuring
several tens of meters or more in length and up to several cm in thickness. The
orientation of the fractures will be governed primarily by the state of stress in the
subsurface. In over-consolidated soil or bedrock, high lateral stresses will result in
sub-horizontal fractures, whereas in normally consolidated soil or fill, weak lateral
stresses will result in sub-vertical fractures. These findings are supported by field
observations and lab experiments.
Results of theoretical analyses of groundwater flow, based on standard
methods used by hydrologists, indicate that hydraulic fractures could significantly
increase the yields of recovery wells. Immediately after fracturing, the yields could
increase tenfold or more. The yield diminishes with time, but even after a long time
the yield from a fractured well is more than twice that of an unfractured well,
according to the analyses. Observations at many oil wells indicate that hydraulic
fracturing increases yields by amounts similar to those in the analyses (Howard and
Fast, 1970).
Laboratory experiments have been conducted to determine the soil
conditions required to create a hydraulic fracture. Hydraulic fractures have been
created in all the samples tested thus far. Remarkably, even soft clay that is poorly-
165
-------
consolidated and saturated with water can be readily fractured. Based on the lab
tests, we expect that hydraulic fractures can be created in most naturally-occurring
materials.
The most exciting results have come from a field test, where 10 hydraulic
fractures were created from shallow boreholes (2 m depth) in uncontaminated
glacial till. The purpose of this paper is to describe our field test of hydraulic
fracturing. The description includes the geologic setting, method used to create the
fractures, and the forms of the fractures determined from excavation.
Editorial restrictions permit only a summary of the field test in the following
pages. An extensive description of the test, however, will be available in our final
project report to be submitted in September 1989.
THE FIELD TEST
A field test of hydraulic fracturing was conducted at a site 12 km north of
downtown Cincinnati on the western side of the valley of Mill Creek, a southerly-
flowing tributary of the Ohio River. The site is on the southeastern side of an area
owned by the ELDA Company, who currently uses it as a municipal landfill.
_ Glacial till, which is probably of Illinoian age, underlies the test site. The till
is unlithified; that is, it is uncemented and readily softens or crumbles when
moistened. Hydraulic fractures were created in two stratigraphic units within the till.
The lower unit is massive, dense (bulk s.g.: 2.29) silty-clay containing 10 to 20
percent pebbles and cobbles. The upper unit consists of beds of silty clay and
irregular graded beds of silt, sand and gravel. The beds are flat-lying and they are
typically several dm in thickness. Seven of the hydraulic fractures were created in
the lower, massive unit, and three (2,12 and 13) were created in the bedded unit.
The in-situ state of stress was measured'at the ELDA site using testing
equipment developed for this project. At the depth of initiation of the fractures
(roughly 2 m), the vertical stress is 35 KPa, whereas the horizontal stress is 340 KPa.
Roughly horizontal hydraulic fractures were expected due to the large ratio (roughly
10:1) or lateral to vertical stresses.
The till was unsaturated, but contained small amounts of local perched water
in some gravel lenses. Water contents in the silty-clay were 11 to 13 percent by
weight. Saturated hydraulic conductivity of the silty clay is between 1.5 x lO"6 cm/sec
and 1.9 x 10'7 cm/sec, whereas in silty sands and gravel it is between 1.0 x 10"5
cm/sec and 3.5 x 10"5 cm/sec, according to in-situ measurements made using a
borehole permeameter. Other details of the characteristics of the till will be in the
final report.
BOREHOLES
Eleven boreholes were drilled along a narrow strip trending roughly NE.
166
-------
Most borings are approximately 10 m from their neighbors, except at the southern
end of the site where four borings are clustered together, roughly 2.5 m from one
another (Fig. 1). Depths to the bottom of casing range between 1.64 m and 1.95 m
for most (nine of 11) boreholes. The other two boreholes were deeper; 2.72 m and
3.81 m. The boreholes were 7.5 cm in diameter. ,
The boreholes were designed for the purpose of creating hydraulic fractures
in over-consolidated material. In general, a borehole consisted of a steel tube
cemented into a boring, and open at both ends. A basket was fixed to the lower end
of the casing to prevent cement from plugging the bottom of the borehole. The open
boring extended several dm below the basket and was partly filled with fragments
(cuttings) of till. A narrow notch, oriented normal to the axis of the borehole, was
cut in the wall of the boring several cm below the bottom of the casing (Fig. 2). The
notches extended 4 cm into the till.
Roughly horizontal hydraulic fractures were expected to develop in the till,
and the design of the boreholes was intended to nucleate a horizontal fracture at the
notch. Excavations of the fractures, however, indicated that the notches were
ineffective at nucleating hydraulic fractures (vertical fractures developed in the walls
of the open boring). We conclude that deeper notches will be required to nucleate
horizontal hydraulic fractures at the borehole.
METHOD OF FRACTURING
Hydraulic fractures were created by Halliburton Services, a subcontractor,
using equipment designed to hydraulically fracture oil wells. The equipment consists
of a truck containing a blender, a centrifugal and a positive displacement pump; two
trucks containing sand; a truck containing water; and a van containing monitoring
and control equipment.
During our hydraulic fracturing tests, water was pumped from a water truck
and mixed with sand and chemicals-dye and a gel-in the blender. A centrifugal
pump was used for most of the injection of the mixture of water, sand and chemicals
into the boreholes. Occasionally, the positive displacement pump was used when
pressures in excess of 480 KPa were required to initiate fracturing. A backhoe was
driven next to each borehole and the blade lowered onto the wellhead to prevent
the casing from lifting during injection.
Pumping rate ranged from 0.075 to 0.227 m3/min (20 to 60 gpm), which is
approximately the lower limit that could be maintained by the equipment. The
duration of the tests ranged from 2 to 10 minutes, and the average volume of
injected water was 0.6 rrr (150 gals). Sand was mixed with water at ratios of 0.1 to
0.2 by volume. The volume of sand in fracture 13 was 0.12 m3 (4.3 ft3), based on
calculations using many measurements of the thickness of the fracture.
MONITORING
A variety of techniques of monitoring the process of hydraulic fracturing
167
-------
FIGURE 1. The site of the hydraulic fracturing field test.
Portland cement
3—Basket
Notch
Boring partly filled
with cuttings
FIGURE 2. Schematic of a borehole used in the test.
168
-------
have been developed within the petroleum industry (a review of monitoring
techniques is contained in our first interim report). Such techniques are crucial
because they can provide information on the size, shape, location, and orientation of
a hydraulic fracture as it is formed. This information is required to design the proper
use of a fracture in delivery or recovery. Moreover, real-time monitoring can be
used to ensure that only the intended material is cut by the hydraulic fracture.
Three different monitoring methods were evaluated during the field test. The
simplest method involves recording the injection pressure as a function of time.
Under ideal conditions, pressure increases during initial injection and then
decreases abruptly as the fracture begins to propagate (Haimson and Fairhurst,
1970). Accordingly, the peak injection pressure is commonly regarded as an
indication of the beginning of fracturing. We measured injection pressure as a
function of time in the field and found that the form of the function is similar to the
form that occurs when hydraulic fractures are created in rock: pressure increased to
between 240 and 760 KPa, and then decreased abruptly. We conclude, therefore,
that the onset of hydraulic fracturing in soil can be determined by monitoring
injection pressure (our lab experiments indicate that the fracture actually begins to
grow slightly before the pressure reaches a maximum).
Another monitoring technique involves measuring the deformation of the
ground surface over a hydraulic fracture. The measured deformations are compared
to the results of a theoretical model that determines surface deformations as a
function of the geometry of an idealized fracture at depth (e.g. Pollard and
Holzhausen, 1979; Davis, 1983). We tested this technique by measuring
deformations using highly accurate tiltmeters obtained from Applied Geomechanics
Inc., Santa Cruz, CA. The equipment yielded strong signals, indicating relatively
large tilts. We are currently inverting the tilt data and comparing the results to the
actual geometry of the fractures as a detailed evaluation of this monitoring
technique.
The electric geophysical methods Mise-a-la-masse, Dipole-Dipole, Wenner,
and Spontaneous potential were evaluated as techniques of monitoring hydraulic
fractures at shallow depths (Steirman, 1984). Mise-a-la-masse yielded anomalies
associated with the hydraulic fractures, and it shows promise as a monitoring
technique.
HYDRAULIC FRACTURES
A backhoe was used after the fracturing operation to dig networks of
trenches in the vicinity of the boreholes, exposing the hydraulic fractures on the
trench walls. In general, the fractures are elongate in plan and the parent borehole
lies near one end of the fracture. Most of the fractures vented to the surface and the
long axis of each fracture lies on a line between the borehole and the vent. The
major axes, measured from borehole to vent, range from 1.8 to 13.5 m, and the
average is roughly 6 m (Table 1). Accordingly, the length of the fractures averages 3
times more than the depth of their initiation.
The areas covered by the fractures range from 2.2 to 90 m2, and the typical
area is between 20 and 30 m2. The two deeper fractures (2 and 4) did not vent and
our excavations were insufficient to determine their sizes. We were unable to create
169
-------
a fracture at one well (8) because sand plugged the casing.
All of the hydraulic fractures created beneath level ground dip toward their
parent borehole, and the dip angles are remarkably similar, ranging from 12° to 25°
(Table 1). One hydraulic fracture (13) was created beneath sloping ground, and it is
nearly horizontal.
The three-dimensional forms of the hydraulic fractures were determined
from detailed cross-sections mapped on the walls of trenches cutting the fractures.
The maps and cross-sections will be presented in our final report, but they will be
summarized here in the form of an idealized hydraulic fracture (Fig. 3). The
idealized form for fractures beneath level ground consists of the following zones:
Zone lr Adjacent to Borehole: Vertical fracture containing the axis of the open
part of the borehole (Fig. 3). The strike of the vertical fracture is
perpendicular to the major axis of the fracture. The vertical fracture
changes orientation abruptly to a sub-horizontal fracture within one to
several dm of the borehole.
Zone 2. Vicinity of Borehole: Sub-horizontal fracture extending as much as 2 m
from borehole (Fig. 3). The maximum extent of this zone is roughly equal to
the depth of initiation of the fracture. The sub-horizontal fracture either
terminates or changes orientation abruptly to Zone 3.
Zone 3. Majority of the Hydraulic Fracture: Planar to trough-like fracture
dipping shallowly toward parent borehole (Fig. 3).
Zone 4. Vent: Steeply-dipping fracture intersecting the ground surface. Strike of
the fracture is parallel to the strike of the fracture in Zone 1 (Fig. 3). The
fracture at the vent is several dm to 1 m in length and extends to a depth of
several dm.
The transition between zones is abrupt and marked by a sharp change in
orientation of the fracture. Zone 2 is absent from some of the fractures, but the
other three zones occur in all the fractures created beneath level ground.
The idealized fracture is asymmetric in plan with respect to the borehole.
The long axis of the fracture is on the side opposite from the backhoe that was used
to prevent movement of the casing during injection.
HIGHLIGHTS AND SHORTCOMINGS OF THE TEST
The successful creation of hydraulic fractures in unlithified material is the
main highlight of the test. This result, combined with the results of our lab
experiments, suggests that hydraulic fractures can be created in a variety of near-
surface settings.
The fractures that were created are several times larger than anticipated.
Moreover, we expect that the maximum size attainable by a hydraulic fracture will
increase with increasing depth of initiation. This is important because the area
170
-------
Load due to
backhoe
Zone 2
Horizontal frx
Zone 1
Vertical frx
FIGURE 3. Oblique view of an idealized hydraulic fracture created during the field test.
Based on exposures of ten hydraulic fractures in the walls oftrenches.
TABLE 1.: SIZES AND DIPS OF HYDRAULIC FRACTURES
Fracture#
2
4
5
6
7
8
9
10
11
12
13
Depth
(m)
2.77
3.84
1.64
1.85
1.83
1.89
1.75
1.83
1.67
1.98
1.83
Plan Area
[m
unknown
unknown
13.
28
2.2
20
12
9
30
90
Major axis
in Plan (m)
unknown
unknown
3.6
6.4
1.8
5.5
3.3
4.1
8.2
13.5
Average
Dip ,
shallow
shallow
25°
14°
variable
17°
22° and 25°
24° „
12°
sub-horiz.
171
-------
covered by a fracture will play a key role in how effective it is in improving delivery
or recovery.
We should caution, however, that the orientation and size of a hydraulic
fracture will be strongly dependent on the properties (primarily the in-situ state of
stress) at a particular site. Site assessment in general, and measurements of the in-
situ state or stress in particular, will be crucial in evaluating where hydraulic
fracturing could be beneficial. We are developing a site assessment tool that
measures in-situ stress in soil by creating a small hydraulic fracture.
The major shortcoming of the test was a paucity of sand propping some of
the hydraulic fractures. The fractures will be ineffective in improving recovery if
they lack sand to prop them open. We know that it is possible to inject sand because
one fracture (13) contained several cubic feet of sand, reaching thicknesses of one
cm. In other fractures, however, sand was nearly absent even though it was mixed
into the injection fluid.
The lack of sand is probably due to methods and equipment used to create
the hydraulic fractures. Following fracturing, sand was found in pipes extending
from the blender truck to the well head. We suspect that the sand settled out in the
pipes during pumping and never reached the well head, and this suspicion is
confirmed by mass balance calculations at well 13. The equipment used during the
test was designed to create fractures several orders of magnitude larger than the
ones we created. Different equipment, designed specifically to create small fractures
at shallow depths, should improve the placement of sand.
DISCUSSION
The orientations of the hydraulic fracture in the four zones are consistent
with a conceptual model based on experiments and fracture mechanics. According
to the conceptual model, we expect a hydraulic fracture to propagate at an
orientation that requires the least expenditure of energy. Typically, hydraulic
fractures grow normal to the direction of least principle compression in the
enveloping material. In over-consolidated till, such as at the ELD A site, lateral
compression exceeds the vertical stress due to weight of the overburden.
Accordingly, under ideal conditions (uniform loading of an isotropic material of
infinite extent) we expect sub-horizontal fractures in the till. Sub-horizontal
fractures were observed in Zone 2, but elsewhere the fractures were dipping,
suggesting that conditions are different from the ideal.
Several conditions at the test site cause the fractures we created to differ
from fractures created under those ideal conditions. Fluid pressures acting on the
wall of the open segment of the borehole affect the stresses local to the borehole. As
a result of the fluid pressures, the circumferential stress at the wall of the borehole
diminishes, potentially to values less than the vertical stress. Apparently the fluid
pressure resulted in local stresses that favored the nucleation of a vertical fracture in
the wall of the borehole (the fractures in Zone 1). Vertical fractures in the walls of
open boreholes were created in our lab experiments, as well as in the lab
experiments of others (Haimson and Fairhurst, 1970; Medlin and Masse, 1979).
172
-------
Shear stresses develop at the tip of a horizontal fracture when it grows to a
length roughly equal to its depth beneath the ground surface. This type of
mechanical interaction between the hydraulic fracture and the ground surface will
cause the fracture to turn upward and propagate toward the ground surface (Pollard
and Holzhausen, 1979; Narendran and Cleary, 1983). Apparently, mechanical
interaction results in the inclination of hydraulic fractures in Zone 3 (e.g. Narendran
and Cleary, 1983; fig. 9), and perhaps in Zone 4.
The asymmetry of the fractures with respect to the borehole seems to be due
to vertical loading caused by the backhoe-fractures propagated away from the.
backhoe. Accordingly, it could be possible to control the direction or the long axis of
a shallow fracture by artificially loading the overlying ground surface.
We are currently developing a theoretical model, based on principles of
fracture mechanics and fluid mechanics, that will explain the development of the
forms of hydraulic fractures observed in the field experiment.
CONCLUSIONS
The conclusions of the field test include results related to the fractures that
were produced, the methods used to monitor the fractures, and the equipment used
in the fracturing process.
1. Fracturing: Hydraulic fractures can be created at shallow depths in glacial till.
The fractures are elongate in plan and dip gently toward their parent
borehole. The maximum dimensions of the fractures are several times
greater than the initiation depths.
2. Monitoring: Injection pressure can be used to determine the onset of
hydraulic fracturing in till. Tiltmeters can be used to monitor the growth of
hydraulic fractures at shallow depths. Electrical geophysical methods show
promise as monitoring tools.
3. Equipment: Equipment used to create hydraulic fractures at oil wells can be
used to create hydraulic fractures at contaminated sites. New equipment,
designed specifically for creating hydraulic fractures at shallow depths in
soil or rock, should perform better than equipment used by the oil industry.
Field tests planned for FY 1989 are designed to evaluate new equipment
used to create hydraulic fractures in soils. One of the systems that we are developing
is based on equipment used by construction and geotechnical contractors. A
pneumatic rig is used to drive a casing to a desired depth. Then, a specially-designed
point on the casing is withdrawn and a hydraulic fracture is created at the bottom of
the open casing. The point is then replaced and the casing driven to a greater depth
where another fracture is created. In this manner, we expect to be able to create
multiple hydraulic fractures at various depths from a single boring.
173
-------
ACKNOWLEDGEMENTS
This project was funded by USEPA Project 68-03-3379-2-8, and I appreciate
the support and guidance of the current USEPA Work Assignment Managers, H.
Pahren and M. Roulier. Previous WAMs, D. Keller and D. Ammon, have also
played important roles in the success of the project.
The field test would have been impossible without the cooperation of John
Stark,the manager of the ELD A Landfill. I also thank Mark Roberts, Halliburton
Services; Don Steirman, the University of Toledo; and Gary Holzhausen, Applied
Geomechanics, for then- efforts in the project.
I appreciate the help of Joe Wilmhoff, who prepared the illustrations.
REFERENCES
Davis, P.M. Surface deformation associated with a dipping hydrofracture. J.
Geophys. Res., v. 88, pp. 5826-5834,1983.
Haimson, B. and C. Fairhurst. In-situ stress determination at great depth by means
of hydraulic fracturing, in Rock Mechanics—Theory and Practice,
Proceedings llth Symposium on Rock Mechanics, pp. 559-584,1970.
Howard, G.C. and C.R. Fast. Hydraulic Fracturing. SPE AIME, New York, 1970.
Medlin, W.L. and L. Masse. Laboratory investigation of fracture initiation pressure
and orientation. Soc. Pet. Eng. J. (April, 1979, pp. 124-144,1979.
Narendran, V.M. and M.P. Cleary. Analysis of growth and interaction of multiple
hydraulic fractures. SPE Paper 12272, presented at 7th SPE Reservoir
Simulation Symposium, pp. 389-398,1983.
Pollard, D.D. and Gary Holzhausen. On the mechanical interaction between a fluid-
filled fracture and the Earth's surface. Tectonophysics, 53, pp. 27-57,1979.
Steirman, DJ. Electrical methods of detecting contaminated groundwater at the
Stringfellow waste disposal site, Riverside County, California. Environ. Geol.
Sci., 6, pp. 11-20,1984.
174
-------
COMPUTER-BASED METHODS OF ASSESSING CONTAMINATED SITES:
A CASE HISTORY
W.G. Harrar, L.C. Murdoch, P.R. Cluxton, M.S. Beljin
Center Hill Research Facility
University of Cincinnati
5995 Center Hill Rd.
Cincinnati, Ohio 45224
ABSTRACT
A Computer-Assisted Engineering (CAE) system, based on an AT-style
microcomputer, has been developed to evaluate data from hazardous waste sites.
Commercially available software packages are used to store, manipulate, analyze,
and graphically represent site information. Utility programs, written specifically for
the project, are used to transfer data from one software package to another. System
capabilities include the generation of maps and cross-sections showing the geology,
hydrology, and distribution of contaminants; the calculation of volumes or masses of
contaminated material; and the modeling of ground water flow and contaminant
transport.
The Queen City Farms Superfund site has been characterized and evaluated
using the CAE system. Maps and cross-sections of the geology, hydrology, and
distribution of contaminants at the site were created. A conceptual model of the
groundwater flow and advective transport of the contaminants was established
based upon the maps and cross-sections. Preliminary numerical modeling of the
groundwater flow and advective transport of contaminants was conducted. The
results of the numerical models confirm our conceptual model and predict a
possible scenario for the migration of ground water contaminants. The volume of
contaminants in the soil was also calculated, and maps of the top and bottom
surfaces of the contaminated zone were generated to aid possible excavation.
175
-------
INTRODUCTION
Investigation and remediation of hazardous waste sites results in large
amounts of information that can be stored, analyzed, and displayed using a personal
computer. We have developed a computer system intended for that purpose under
the Computer Assisted Engineering (CAE) project funded by the USEPA. The
system can be used throughput the remedial process. It is designed to characterize a
site from data obtained during a remedial investigation and feasibility study, to
estimate the volumes of contaminated material, to aid in the design of remedial
procedures, and to evaluate the effectiveness of the implemented remedial actions.
In the following paper, we describe the hardware and software components, and
some of the basic capabilities that have been developed. A summary of one of the
Technical Assistance projects serves as an example of the CAE system.
DESCRIPTION OF THE COMPUTER SYSTEM
The early stages of the. project involved selecting hardware and software, and
iping basic capabilities of the system. Additional capabilities have been
developedin response to requests from EPA regional offices.
HARDWARE
The components of the CAE system are based on an AT-style personal
computer containing 3.6 Mb of RAM and a 40 Mb hard disk. Peripheral
components include two digitizing tablets, a drafting plotter, a screen camera, and a
high resolution color graphics display.
SOFTWARE
The CAE system consists primarily of readily available software that we have
tailored to the analysis of problems at contaminated sites. A Database
Management System (DBS) stores information from site investigations. A
Geographic Information System (GIS) performs spatial analyses. Ground water
Modeling Programs (GMP) analyze me movement of water or chemical
contaminants in the subsurface. A Computer-Aided Design and Drafting (CADD)
package receives information from elsewhere in the system and renders it as elegant
drawings.
Other software has been written during the project to perform tasks that
were impossible to accomplish using commercial software. The transfer of data
between various software packages, for example, typically requires a program that
translates the output format of one package into the input format of another.
176
-------
USES OF THE SYSTEM
The DBS is the heart of the CAE system. It contains any information
generated during site investigation or remediation that is required for analysis of the
site. Accordingly, the format of the database is flexible enough to permit recovery of
the types of data needed during the analyses, yet is compact enough to facilitate the
data entry. The formats that we use vary slightly from one site to another depending
on the detail of sampling, and on the type or data that are stored. An example
format for a site sampled at several different times consists of one record, or row,
for each contaminant analyzed at each well. The records are divided into fields, or
columns, containing information on (a) Well identification, (b) contaminant name,
(c) contaminant concentration, and (d) sampling date. Maps of the distribution of
each contaminant during each sampling period can be constructed using this format.
The CAE system is capable of generating maps and cross-sections that
characterize hazardous waste sites. In many cases, the maps and cross-sections are
virtually identical to those created by hand. In general, however, they are intended
only as first approximations, the details of which will be modified by trained
investigators. The interpolation of continuous fields from data taken at points is
fundamental to site characterization, and it is a primary role of the GIS. Inferring
fields, or plumes, of contamination from analyses of samples taken at wells is one
example of interpolation that is used extensively.
Another application of the GIS is creating spatial or three-dimensional
models of site geology or hydrology. The models can be sliced to generate cross-
sections or maps.
The GIS is powerful tool for inferring continuous fields from point data. The
difference between the GIS and other contouring packages lies in the control that
the user has on the interpolation procedure. Inferring concentrations in regions
where neighboring wells lack contamination is one difference between the GIS
method and a contouring package that uses linear interpolation. We use the GIS to
infer that points whose nearest neighboring well is uncontaminated are themselves
uncontaminated. In contrast, the linear interpolation package infers that points
near uncontaminated wells are contaminated at concentrations equalling a weighted
average of concentrations at neighboring wells. Several other methods, such as
kriging, are currently available for generating continuous fields from point data. The
other methods will yield slightly different results compared with those presented
here.
The level of confidence of a line on a map or cross-section is commonly
indicated by demarking the line as solid, dashed, or dotted. We have developed a
method of assigning line style based on how close the line is to a data point
containing known values. Lines that are relatively close to a point are solid, those at
intermediate distance are dashed and those at great distance are dotted. The
distance used to discriminate between line styles can be rigorously determined using
a value we term the proximity index. The proximity index, Pi, is defined as the
average minimum distance between points divided by the square root of two.
According to this definition, a Pi of 1 is the minimum distance required to
completely cover an area sampled on a square grid. Solid lines are used in areas of
Pi < 1, dashed lines for 1 < Pi < 2 and dotted lines for Pi > 2. The GIS allows us to
incorporate the proximity index into the creation of maps and cross-section. The
reader should keep in mind that the proximity index is intended only to indicate the
177
-------
relative distance to a point of known value; the level of confidence associated with
that distance must be determined from the geology, hydrology, and other factors.
Contaminated soils must be excavated at many sites during remediation. The
GIS is currently being used to design strategies of excavation. For example, it is
used to estimate the volumes of soil of various concentrations that will be generated.
Moreover, the GIS produces maps of the depths of excavation required to remove
soil contaminated with concentrations greater than allowable by EPA guidelines.
Flow of ground water and contaminant transport are analyzed using
numerical and analytical models. A typical flow analysis involves generating files of
the location of saturated intervals from well-logs in the DBS, and inferring the
locations of saturated zones using the GIS. These data are then used to determine
boundary and initial conditions in a flow model.
Analyses of flow and transport are conducted first to characterize the site.
Subsequent analyses, based on those done during characterization, are used to
assess potential remedial action procedures.
APPLICATION OF THE CAE SYSTEM: THE QUEEN CITY FARMS SITE
The CAE system has currently been used to analyze data from a half dozen
sites, with each site requiring one or several system capabilities. The Queen City
Farms site is one example that illustrates many of the capabilities, ranging from the
creation of maps and cross-sections to the analysis of ground water hydrology and
the calculation of excavation parameters.
The site is located near Seattle, in King County, Washington. The primary
source of contamination is thought to be from ponds where industrial and municipal
wastes were stored during the 1950s and 1960s. The ponds are on the southeastern
and southwestern edges of the Queen City lake (Figure la). Other potential sources
of contamination include sludge ponds and a leachate treatment systems associated
with the Cedar Hills Landfill, and the Cedar Hills landfill itself (Figure la). Figure
la was created by digitizing several base maps of the site.
DISTRIBUTION OF CONTAMINANTS
Contamination has been identified as a black, oily compound in soil, and
aromatic hydrocarbons, metals, sulfates, and nitrates in ground water. We received
data from the office of USEPA Region X on the concentration of contaminants in
wells sampled during three periods from fall 1986 to spring 1987.
Maps were created showing the distribution of contaminants after each
sampling period. For example, Figure Ib is a map of the concentration isopleths for
methylene chloride in the shallow aquifer between 3 Dec. 1986 and 8 February
1987. It shows a maximum concentration of 260 ppb in the vicinity of the Queen
City ponds, and a secondary maximum of roughly 200 ppb south of the sludge
178
-------
Queen City Forms Site and Cedar Hill Landfill
King County. Washington
BASE MAP
Leachate Treatment
System °
CITY FARMS
DISPOSAL SITE
Absent
o Sample Location
Queen City Farms Site and Cedar Hill Landfill
King County. Washington
M«thylen« Chloride Concentration (ppb) Winter 1086-87
V X,
\
1
I •
i:
i :
100 - •
• 50 '"
FIGURE 1. a. Base map of study area showing potential sources of
contamination and end points of cross-section line B-B'. b. Methylene
Chloride concentrations (maximum = 260 ppb) in shallow aquifer for
sampling period 3 December, 1986 to 8 February, 1987.
179
-------
disposal ponds. A zone of lesser contamination of methylene chloride occurs north
of the Cedar Hills landfill.
The map (Figure Ib) was created in the GIS by analyzing data from 16
monitoring wells. The nearly straight isopleths near the sludge pond result from the
absence of contaminants in wells west of the sludge ponds and on the east-central
side of the study area (see previous section). The proximity index for this map is 725
feet, so the solid lines indicate areas that are within 725 feet of a monitoring well;
dashed lines are between 726 and 1450 feet; and dotted lines are greater than 1451
feet from a monitoring well. A final copy of the map was created by transferring
data from the GIS to the CADD.
GEOLOGY AND HYDROLOGY
The site is underlain by Recent alluvium and deposits of Pleistocene glacial
•• . , • A* 1 1 • Pi _ ._ J A* 11 /TT?i _ — —. ^\ T— XWA-M A***i1 4-\-i r\ \*
-------
181
-------
Conceptual Model
The till forming the base of the shallow aquifer is absent in an area south of
the Queen Qty Farms site (Figure 1). This is important because the upper surface
of the till dips toward that area. Thus, shallow ground water flows towards the area
of absent till, where it presumably flows from the shallow aquifer downward to
recharge the intermediate aquifer.
The piezometric surface of the intermediate aquifer slopes generally
northeastward except beneath the recharge area where it is mounded. Flow
directions are inferred to be roughly radial from the recharge area and northeasterly
elsewhere.
The piezometric surface of the deep aquifer slopes generally northeastward.
The elevation of the piezometric surface of the deep aquifer was 10 to 30 feet above
that of the intermediate aquifer during the times when measurements were made.
Accordingly, flow in the deep aquifer is inferred to be generally northeastward, and
there should be minor flow upward from the lower to the intermediate aquifer.
Migration of contaminants due to advective transport in ground water is
inferred to be generally southward in the shallow aquifer, downward to the
intermediate aquifer through the zone of recharge, and generally northeastward in
the intermediate aquifer. Contamination of the deep aquifer is possible if the
hydraulic heads in the intermediate aquifer locally exceed those in the deep aquifer.
Numerical Models
Numerical modeling was conducted to test our conceptual model and to
evaluate the possibility of flow from the intermediate to the deep aquifer in the area
of recharge. Flow rates through the area of absent till were estimated using two
techniques: a flow rate of 1.2 ft3/sec was estimated by averaging the flow rate
calculated from several cross-sections of existing water levels in the shallow aquifer;
a flow rate of 5.6 ft3/sec was estimated by measuring the area draining to the
recharge zone and multiplying by the maximum rainfall, 6.0 in/month.
Those recharge rates (and other parameters that will be described in our
final report) were used in a 3-D ground water model, MODFLOW, (McDonald and
Harbaugh, 1983) to calculate steady-state heads in the intermediate and deep
aquifers. Streamlines derived from the head calculations indicate that flow in the
intermediate is generally toward the northeast, but locally radiates from the zone of
recharge. Flow in the deep aquifer is also to the northeast with local effects at the
zone of recharge (Figure 3). These results support our conceptual model.
Analyses indicate that recharge results in mounds in the local distributions of
head in the lower aquifers, and the relative heights of the mounds are sensitive to
the rate of recharge. The lesser recharge rate results in a mound in the
intermediate aquifer that is lower than the piezometric surface of the deep aquifer,
implying that flow is upward through the silt bed separating the aquifers. In
contrast, the greater recharge rate results in a mound in the intermediate aquifer
that is higher than the piezometric surface of the deep aquifer. Accordingly,
182
-------
Shallow Aquifer
N 0 500
''////'Till absent
Intermediate Aquifer
Deep Aquifer
N o soo
Flow direction
Potential source of
contamination
Possible extent of
contamination
FIGURE 3. Flow patterns in the shallow, intermediate, and deep aquifers; the
latter two are generated from computed heads with maximum estimated
recharge rate applied as a point source in the intermediate aquifer. Shaded
area in deep aquifer depicts possible extent of contaminant migration.
183
-------
downward flow from the intermediate to the deep aquifer is implied for the greater
recharge rate. Additional modeling suggests that recharge rates in excess of 1.6
ft3/sec could result in flow from the intermediate to the deep aquifer.
We conclude that it is possible for the mound in the intermediate aquifer to
exceed the height of the piezometric surface of the deep aquifer. As a result, the
migration of contaminants from the intermediate to the lower aquifer is possible,
according to the preliminary modeling done using the CAE system. This possibility
is based on preliminary modeling and its validation will require obtaining additional
data at the site.
The modeling described above is intended solely as a demonstration of the
CAE system. The capabilities of the CAE system involve the use of interpolation
procedures and models that have been tested and are generally accepted by the
scientific community. To date, however, the accuracy of the maps generated with
the system have yet to be varified by field checking. For this reason, we are
currently unable to estimate how well the interpolated fields will predict, for
example, the distribution of contaminants in the field, and we advise the reader to
keep this uncertainty in mind when interpreting the maps. We suggest that our
maps and models, or any other equally preliminary maps and models, be thoroughly
checked in the field before they are used with confidence. The calculations and
conclusions presented here are not necessarily those of the USEPA.
SUMMARY
An AT-style computer system has been tailored to fit the needs of storing,
analyzing and displaying data from the investigation and remediation of hazardous
waste sites. The system is designed to generate maps and cross-sections of the site
geology, hydrology, and distribution of contaminants. As such, the system is
designed to characterize a site from data obtained during a remedial investigation
and feasibility study, to estimate the volumes of contaminated material, to aid in the
design of remedial procedures, and to evaluate the effectiveness of the implemented
remedial actions.
Models of ground water flow and transport are integrated into the system to
expedite the preliminary analysis of existing hydrology, or the evaluation of
contaminant recovery.
As an example of the CAE system capabilities the Queen City Farms
Superfund site has been characterized and evaluated. Maps and cross-sections of
the geology, hydrology, and distribution of contaminants in ground water at the site
were created. A conceptual model of the site hydrogeology and advective transport
of contaminants in a three aquifer system was developed based on the maps and
cross-sections.
The site in underlain by Recent alluvium and Pleistocene glacial outwash,
stratified drift, and till. The hydrology consists of a shallow, unconfined aquifer
overlying a till of low permeability; an intermediate, unconfined aquifer overlying a
silt layer; and a deep aquifer that is semi-confined by the silt layer. The till forming
the base of the shallow aquifer is locally absent in an area south of the Queen City
184
-------
Farms site. Migration of contaminants due to advective transport in ground water
is inferred to be generally southward in the shallow aquifer, downward to the
intermediate aquifer through .the zone of recharge, and generally northeastward in
the intermediate aquifer. The elevation of the piezometric surface of the deep
aquifer, which slopes northeastward, was higher than that of the intermediate
aquifer when measurements were made. Contamination of the deep aquifer is
possible if the hydraulic heads in the intermediate aquifer locally exceed those in
the deep aquifer.
Preliminary numerical modeling of the ground water flow and advective
transport of contaminants was conducted to test our conceptual model and
determine critical recharge values to the intermediate aquifer that induce ground
water flow downward into the deep aquifer. The results of the numerical modeling
confirm our conceptual model and predict a possible scenario for the migration of
contaminants in the ground water. A value of 1.6 fir/sec was determined as a
critical recharge rate that induces flow into the deep aquifer.
The volume of contaminants in the soil was also calculated, and maps of the
top and bottom surfaces of the contaminated zone were produced to aid possible
excavation.
The capabilities of the CAE system are available to regional offices of the
USEPA under the Technical Assistance Program. Direct inquiries to Gene Harris,
RRELSTDD.
ACKNOWLEDGEMENTS
Appreciation goes out to Herbert Pahren, Eugene Harris, and former project
WAM's Douglas C. Ammon and Douglas Keller for their guidance and cooperation.
Thanks goes to personnel at the EPA Region X, Seattle, Washington for providing
the data. Joseph Wilmhoff contributed his drafting skills.
REFERENCE
McDonald, M.G., and Harbaugh, AW. A modular three-dimensional groundwater
flow model. Open-File Rep. 83-875. U.S. Geol. Surv., 1983. 528 pp.
185
-------
RESULTS AND PRELIMINARY ECONOMIC ANALYSIS
OF AN APE6 TREATMENT SYSTEM FOR
DEGRADING PCB'S IN SOIL
by:
John A. Wentz,
Michael L. Taylor, Ph.D,
William E. Gallagher
PEI Associates, Inc.
Cincinnati, Ohio
D.B. Chan, Ph.D, P.E.,
Naval Civil Engineering Laboratory
Port Hueneme, California
Charles J. Rogers
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio
ABSTRACT
This paper describes the system and operational procedures utilized as
well as results obtained when the APE6 chemical dechlorination process was
scaled up to field-scale and employed to dechlorinate PCB-contaminated soil
on the Island of Guam, U.S.A. The APEG system consisted of a steam jacketed,
mixer, steam generating plant, and condensate collection system. Approxi-
mately 15 cubic yards of soil in batches of 1.5 to 2 cubic yards each with
average initial PCB concentrations of 3430 ppm Aroclor 1260 were KPEG treat-
ed. PCB concentrations of treated soil were reduced by more than 99.999
percent with no individual PCB congener exceeding 2 ppm. The demonstration
proved the efficacy of the APEG process to chemically dechlorinate PCB con-
taminated soil without the use of DMSO or TMH. Field-scale demonstrations
are being planned for Fall 1989 with modified reagents and optimized operat-
ing parameters where the APEG chemical dechlorination process is estimated to
cost $200-300 per ton of contaminated soil.
INTRODUCTION
Halogenated chemical contaminants such as chlorinated dibenzodioxins
(PCDD's), chlorinated dibenzofurans (PCDF's), and polychlorinated biphenyls
(PCB's) have contaminated soil, water, and other matrices in various loca-
tions throughout the United States and the world. Because many of these
halocarbon contaminants have been found to be highly toxic in laboratory
animal studies, human exposure is undesirable. To date, only limited dis-
posal or treatment options are being developed for ^these contaminants and the
matrices they contaminate—particularly soil. The large quantities of con-
taminated soil have created a need for a safe, cost-effective, cleanup
186
-------
process as an alternative to the current practices of secured on-site stor-
age, Class B landfill ing, or incineration;
In 1978, at the Franklin Research Center in Philadelphia, Pennsylvania,
a reagent was identified and successfully utilized to dechlorinate PCB's in
oil.1 The reagent consisted of an alkali metal hydroxide (AOH) and poly-
ethylene glycol (PEG) mixture which became known generically as APEG (alkali
polyethylene glycolate). The U.S. Environmental Protection Agency's Risk
Reduction Engineering Laboratory (U.S. EPA/RREL) initiated further develop-
ment of the APEG chemical dechlorination process for PCB oils to include
soils contaminated with PCB's, PCDD's, and other potentially toxic, halo-
genated aromatic compounds. Initial laboratory findings indicated that
PCB-contaminated soils could be decontaminated and .that further investigation
of the process including assessment for full-scale service was warranted.
A typical laboratory-scale procedure for dechlorinating PCB-contaminated
soil entails mixing potassium hydroxide (KOH) and PEG-400 (average molecular
weight 400) to formulate the reagent known as potassium polyethylene glyco-
late (KPEG). The KPEG reagent is mixed with the contaminated soil, heated to
150°C, and held at that temperature for 1 to 4 hours to allow completion of
the reaction. Excess reagent is decanted, the soil neutralized with sulfuric
acid and rinsed two or three times with water, and the decontaminated soil
discharged.
The PCB's, PCDD's, and PCDF's are dechlorinated in a reaction with the
APEG mixture. The reaction of AOH with PEG-400 produces an alkoxide (ROA)
(see Equation 1) that, in turn, reacts with a chlorine atom on the aryl ring
to produce an ether and chloride salt (AC1) (see Equation 2). Replacement of
the chlorine atom on the aryl ring with an ether linked PEG detoxifies the
compound.2 The dechlorination process is described in general terms in
Equations 1 and 2:
ROH + AOH •> ROA + HOH
ROA + ArCl
n
ArCln_1OR
AC1
(Eq.. 1)
(Eq. 2)
Early APEG reagent formulations included solvents such as dimethyl sulfoxide
(DMSO) and triethylene glycol methyl ether (TMH). The DMSO and TMH were
believed to serve as cosolvents to the APEG formulation to enhance reaction
rate kinetics by improving rates of extraction of the aryl halide compound
into the alkoxide phase.3'4 Later findings, subsequent to the first pilot-
scale APEG chemical dechlorination demonstration on PCB-contaminated soils,
indicated that DMSO and TMH could be removed from the APEG formulation with-
out hindering the dechlorination process or extending the reaction time.
In June 1987, a pilot-scale APEG chemical dechlorination demonstration
was performed on a PCB-contaminated site. The pilot-scale demonstration was
one of the earliest attempts to dechlorinate PCB-contaminated soil at pilot
scale using a reactor. The system consisted of a reaction .vessel, electrical
heating elements, and a condensate collection system to collect moisture
driven off of the soil and KOH solution at the elevated temperatures. The
187
-------
reactor vessel consisted of a 16-inch-diameter pipe, 40 inches long, that was
loaded with approximately 35 Ib of RGB-contaminated soil per batch. A pre-
mixed reagent formulation of DMSO, TMH, and PE6-400 was added to the soil,
then 45 percent KOH solution added separately. The treatment parameters
utilized mimicked those used at laboratory scale. Initial PCB concentrations
of the batches ranged from 133 to 7013 ppm and averaged 1990 ppm. PCB con-
centrations of the treated soil batches ranged from 1.09 to 12.4 ppm and
averaged 5.6 ppm, representing an overall PCB destruction rate of more than
99.7 percent.
The satisfactory reduction of PCB's in the soil at the pilot scale lead
to the U.S. EPA/RREL's decision to design, construct, and demonstrate the
efficacy of a larger KPEG chemical dechlorination system. The proposed
field-scale system would be capable of treating 1 to 2 cubic yards of con-
taminated soil per batch at a remote location.
EXPERIMENTAL METHODS
DESCRIPTION OF SELECTED SITE FOR FIELD-SCALE KPEG DEMONSTRATION
The U.S. Navy Civil Engineering Laboratory (NCEL) and U.S. EPA/RREL
agreed upon a U.S. Navy site for the field-scale demonstration. The U.S.
Navy Public Works Center (USN PWC) site on the Island of Guam, U.S.A, was
selected when analytical results of the collected soil samples indicated
average PCB concentrations of 2500 ppm with "hot spots" as high as 45,860 ppm
(4.58 percent). Soil contamination found mainly in a nearby storm drainage
ditch resulted from leaks from a transformer rework building that had been
used as early as World War II. The waste PCB oil that was stored outside
leaked and was carried by surface runoff into the ground.
In preparation for the field-scale KPEG treatment demonstration, a 60-ft
by 40-ft metal building was constructed on a 100 ft by 100 ft concrete pad
and was used to stockpile the 20 cubic yards of excavated PCB-contaminated
soil. The excavated soil was screened mechanically to separate particles
1/2-inch and smaller. Of the 20 cubic yards, approximately 15 cubic yards
passed the 1/2-inch screen and were stockpiled for treatment. The remaining
5 cubic yards consisted of coral and rock ranging from 1/2 to 12 inches in
diameter. The oversized material was stockpiled separately for subsequent
special processing.
BRIEF OVERVIEW OF MIXER SELECTION
The type of reaction vessel used for the pilot-scale demonstration
alluded to above was not conducive to the order of magnitude of size for the
proposed field-scale demonstration. A mixer system was required that would
provide sufficient capacity and mixing capabilities for the KPEG/soil mixture
as well as provide efficient heat transfer. The mixer was selected based
upon the demonstrated mixing range and heat transfer efficiency as determined
by mixer manufacturer facility tests and scale-up potential.
188
-------
Various mixers were evaluated at several mixer manufacturer's facil-
ities. At each facility, prototype mixers were charged with noncontaminated
soil, DMSO, TMH, PEG-400, and 45 percent KOH. The mixers were turned on and
heated with either hot oil or steam through the external jackets. Physical
and operational data were collected, and the mixer was selected that afforded
the widest range of mixing and greatest heat transfer efficiency.
DMSO REMOVAL FROM KPEG FORMULATION
During the design phase of the field-scale KPEG system, the U.S. EPA/
RREL initiated laboratory treatability studies to determine the KPEG chemical
dechlorination effectiveness on PCB-contaminated soil without DMSO or TMH.
U.S. EPA/RREL laboratory results indicated that DMSO and TMH could be removed
from the KPEG formulation without hinderance to the chemical dechlorination
process or extension of reaction time. The removal of DMSO from the formula-
tion was also appropriate from a health and safety point of view. The ex-
cellent solvent characteristics of DMSO, coupled with the known rapid rate of
skin penetration by DMSO, posed serious concerns for workers in the presence
of compounds such as PCB's and PCDD's.
FIELD-SCALE KPEG TREATMENT SYSTEM
A block flow diagram of the KPEG chemical dechlorination system designed
for the demonstration on the USN PWC site on the Island of Guam, U.S.A., is
provided in Figure 1. The diagram illustrates that the mixer was the primary
component of the system where the chemical dechlorination process occurred
and was supported by ancillary equipment to make the system functional. An
extensive pipe network was required for interconnecting the auxiliary systems
to each other and to the mixer. To reduce the complexity of the piping
network, a centralized pipe rack was installed. Figure 2 provides a layout
of the site plan of the KPEG demonstration in Guam. The site plan illus-
trates the location of the equipment, pipe rack, and exclusion zone. The
site plan indicates that the majority of the auxiliary systems were located
outside the exclusion zone to provide easy access. Only equipment contacting
the soil in its contaminated state or required to be located near the mixer
because of physical limitations was located within the exclusion zone.
Mixer
The selected mixer was designed with a total capacity of 793 gallons
(106 cubic feet) and a working capacity of 490 gallons (65 cubic feet). The
mixer was equipped with a 2-speed, 75-horsepower motor and gear box capable
of providing mixer shaft speeds of 30 and 60 rpm. All potentially wetted
parts were comprised of 316 stainless steel to prevent corrosion from chemi-
cal attack. The mixer was provided with an 8-inch-diameter shaft that ran
the length of the mixing cylinder, which was supported at each end by posi-
tive-flow nitrogen purged seals. Extending radially from the shaft were arms
with plows that maintained a 5/8-inch tolerance from the wall. A maximum
tolerance of 5/8 inch between the plow and wall was recommended by the manu-
facturer without substantially sacrificing mixing and heating efficiency by
creating a dead zone where caking could occur. This 5/8-inch tolerance also
established the maximum allowable particulate size of 1/2 inch to prevent
particulate jamming between the plow and wall.
189
-------
CO
o
.0
Eo
_. CD
So
O) —
ELt-O
COCO
"55
LE
CD
LU
QL
CO
CD
190
-------
CO
O
1
•c
o
w co
£51
=3 m
!
,0)
CO
o
CO
LL
CD
LJLJ
d.
^
CO
3
O
o
o
191
-------
Figure 3 is a diagram of the selected mixer and designed features.
Three 2-inch and one 1-inch flanged liquid charge ports were located along
the top of the mixer for PE6-400 and sulfuric acid addition. One 1-inch
charge port was provided at each headwall of the mixer for water addition.
12-inch Teflon-coated rupture disc was provided at the top of the mixer as a
precautionary measure against over-pressurization of the mixer. A 2-inch
vent was also provided that vented to the condensate collection system.
Solids were loaded into the top of the mixer through a 20-inch by
24-inch rectangular flange. A 16-inch flanged screen assembly with a 2-inch
drain was provided at the bottom of the mixer for reagent draining following
treatment. Treated soils were discharged from the mixer through an 8-inch
air-operated ball valve located on the bottom center of the mixer.
0T\e mixer cyTinder was provided with a steam jacket rated at 80 psi
(156°C). A manifold was provided across the top and bottom of the steam
jacket. During heating, steam entered the top manifold, traversed downward
through the jacket, and exited the bottom manifold. The manifolds were
designed to serve as part of the cooling system as well. Rearrangement of
the valves allowed for upflow of cooling water through the jacket.
Platform
The entire mixer/motor assembly was mounted on a platform to elevate the
discharge port sufficiently to allow for placement of a soil-collection
hopper underneath the mixer for discharge. The platform was designed with
catwalks around the mixer for access. The catwalks folded down for transport
in order that legal road widths were not exceeded. The platform was designed
with an integrally mounted jib crane for lifting drums of soil and dry KOH
for charging into the mixer.
Liquid Reagent Loading
The liquid reagent loading system consisted of a pallet scale and air-
operated diaphragm pump. PEG-400 was placed on the scale and tared. The
positive displacement pump was used to charge proper quantities of PEG-400
into the mixer.
Heating System
The heating system was a leased package steam generating plant. Design
calculations based upon approximated soil and moisture content and reagents
indicated that a 600-lb-per-hour, 80-psi unit was required to heat the mixer
contents from ambient temperature to 150°C within a 4-hour timeframe. Greater
pressure steam (higher temperature) could not be used because the steam
jacket rating on the mixer was 80 psi. The mixer steam manifold included a
steam pressure relief valve specified at 80 psi.
Nitrogen System
The nitrogen system was provided in the design as part of the safety
consideration in the event DMSO and TMH were not removable from the KPEG
formulation. The nitrogen system consisted of a pressure regulator and flow
192
-------
m
2
til
CO
CO
til
CO
CO
193.
-------
controller. Nitrogen was purged into the mixer through the seals to displace
ambient air. The removal of DMSO meant the nitrogen was no longer required
and its use was discontinued when it was learned the seals would not become
damaged without the nitrogen flow.
Condensate Collection System
The condensate collection system was designed to collect and condense
moisture vapors vented from the mixer while at elevated temperatures. The
condensate collection system consisted of a 2-inch vent line connected to a
knock-out tank that was used to remove any solids which may have been thrown
into the vent line from the rigorous mixing action. The knock-out tank
vented to a fan-cooled condenser where 9600-cfm ambient air was blown over
the condenser coils. The fan condenser drained into a condensate collection
tank, which originally vented into a secondary condenser consisting of a
copper coil submerged in an ice bath. Restrictions in the line size at the
ice condenser created back-pressure on the system, which necessitated its
removal. The condensate collection tank was then vented directly to an
activated-carbon column for collection of any remaining volatilized organics.
Process Cooling Water System
The hot treated soil contained within the mixer cylinder required cool-
ing prior to further processing. The process cooling-water system consisted
of a 1250-gallon water tank and centrifugal pump assembly. Piping on the
mixer manifolds was revalved to allow cooling water to upflow through the
mixer jacket and return to the water tank. The process cooling-water system
also provided the feedwater directly into the mixer cylinder when the soil
was rinsed in an attempt to recover reagent.
Reagent Collection System
The reagent collection system was included in the design to attempt to
recover and reuse a portion of the KPEG dechlorination reagent. The reagent
collection system consisted of two 1-inch liquid charge ports on the mixer
headwa-lls, a 16-inch screened flange assembly with a 2-inch drain on the
bottom of the mixer, and two 500-gallon steel tanks immediately adjacent to
the mixer. Ideally, following treatment, the soil was washed by pumping
process cooling water directly into the mixer through the liquid charge ports
and allowing mixing. The reagent would be allowed to drain through the
screen assembly and 2-inch drain line into the 500-gallon tanks for potential
reuse. The first attempt at the soil wash and reagent drain proved futile.
Subsequent batches were not washed or drained of KPEG reagent prior to neutra-
lization. Alternative methods for KPEG reagent recovery would be required
should reagent recovery prove essential to the economics of the process.
Neutralization System
The neutralization system consisted of a pallet scale, sulfuric acid,
and drum pump. A stoichiometric quantity of sulfuric acid was pumped into
the mixer to neutralize the known quantity of KOH. The high calcium carbonate
(CaC03) content of the soil required additional acid to reduce the mixer
content pH to a range within 6 to 9.
194
-------
OPERATION OF THE FIELD-SCALE KPEG TREATMENT SYSTEM
The mixer was charged with 3400 Ib of PCB-contaminated soil, 1555 Ib of
PEG-400, and 285 Ib of KOH. The mixer was turned on to high speed (60 rpm)
to mix the soil and reagents. The vent line from the mixer to the condensate
collection system was opened, the fan condenser turned on, and the steam
generating plant ignited. Eighty psi (156°C) steam was circulated through
the mixer jacket until the mixer contents reached 150°C, as indicated by the
thermocouple readouts. Steam pressure was reduced to 70 psi (150°C) and the
temperature and mixing maintained for a 4-hour period. At the completion of
the 4-hour period, the steam generator and mixer were shut down. The fan
condenser was turned off and the contents allowed to cool overnight.
Following overnight cooling, the mixer contents had dropped from 150° to
90°C. Additional cooling was performed by recirculating cooling water from
the process cooling water system in the upflow manner through the mixer
jacket until the mixer contents were cooled to 50°C. The cooling water
remained on and a stoichiometrically calculated quantity of sulfuric acid was
pumped into the mixer in 20-1b increments. Because of the known presence of
high CaC03 concentrations in the Guam soil, additional sulfuric acid was re-
quired to adjust the pH to within a range of 6 to 9. Samples were collected
from the sample collection port on the mixer, and the slurry pH was measured.
Additional 20-1b increments of sulfuric acid were added, and the pH measure-
ment process was repeated until the pH was within the 6 to 9 pH range. The
strong exothermic reaction during acid addition reelevated the temperature of
the mixer contents. The cooling water continued to pass through the mixer
jacket until the mixer temperature was returned to 45°C. During the entire
cooling process, cooling water initially at an ambient temperature of 25°C
was elevated to 40°C, which represented a significant transfer of heat away
from the mixer.
A soil collection hopper was placed under the mixer discharge valve.
The air-operated valve was controlled from the mixer control panel mounted on
the mixer, which was accessible from the platform catwalk. The mixing action
internal to the mixer cylinder directed the contents to the discharge port.
After the treated soil was collected in the soil-collection hoppers, the
hopper lids were securely fastened. The soil collection hoppers were stored
in a secured area on site, awaiting analysis.
RESULTS
VERIFICATION OF PCB DECHLORINATION
On-site analyses of untreated and treated soil were performed by using
extraction and gas chromatographic mass spectrometric (GC/MS) methods adapted
from EPA Methods (SW 846, 3rd Edition). Corroborative analyses on duplicate
samples of the untreated and treated soil and on collected condensate and
recoverable reagent samples were performed by an independent laboratory in
the United States by GC/MS as well. Table I provides the analytical results
195
-------
CD
UJ
O_
vx
a
z:
*f-
uT
to
UJ
r- 1
O
0
•%
CO
H-4 CO
O _1
co »— t
o
E co
=3 C£
CD O
feo
*T-» tn«J
3E >
j— • ifj
25
QiCJ
Hc^
g§
Q. t— 1
1—
0 0
^^ *•— *
UJ
UJ
C£ UJ
O_ CD
i— « z:
UJ
a o
z: a:
rD UJ
O OL.
u_
Q
co z:
CQ
O f—
Q. z:
UJ
U_ CD
OeC
UJ
co o;
«££
o
>— 4
1—
cc
J— •
UJ
C-?
z:
CD
o
t-H
UJ
CQ
£
i—
C -P
•r- C
C CD
O "O CftjQ ^~s
•r- C tO tO to
•P 3 — - 0
O Q. O Z5
o o
0 I.
•a
c s= cu
o 3 -p .a--^
•r- O CO CO to
•P 4- e/l _l +J
co c >->^-^
cn c co to
fO 0 _J -P
H-> •!- (O
c: -p cn-o
(DOC
O 3 -r- CM
s- -a to
cu cu 3 •
a. s- — 6
z:
,1 ^
•t-*
c c;
•r— CU
CQ -f->
O «tJ
P^ CL)
t(_ +j
o
cn
c c
'•P 0
fO •—
S- i—
•P 0
i-
C C
+ CM
-|*
JD •
(O O
—i z:
—1 O
•^
x:
o •
•P 0
fCj p^
ca
r- 1
CD
•
0
t-H
CO
CO
CD
r--.
in
in
01
^•^
CM
r-l
CO
I—I
CM
in
i — i
CM
r— 1
*~^
vo
CM
CO
~-^
F-H
*^ff
0
Ol
o>
T-H
si-
t-H
•
O
^—^
co
r5
t-H
CD
CO
Ol
Ol
o>
0
t— 1
ai
Ol
r-H
CO
CM
oo
CO
o
CM
vo
LO
o
^
CO
0
vo
t-H
Ol
Ol
Ol
vo
co
CM
LO
t-H
vo
o
sl-
co
0
Ol
co
1 1 1 1
t-H O - — •
co in si-
co co in
t-H CM
t-H
oo co r»
Ol Ol t-H
t-H CD VO
01 01 r*~ i
Ol Ol Ol
Ol Ol Ol
CO t-H I — VO
CD co r-x in
co co t^ co
st- t-H OO CO
o> 01 co st-
t-H t-H St- CO
^-^* ^— ^
CM CM
CO r-H O *-*
r** in vo un
r-~ vo CM — '
CO CO CO
o o o o
«d" t-H t-H si"
LO O •* t-H
st- in vo p^
i i i i i
^__^
o
«-H O CD O
t-H t-H i— 1
J*^
— CD. O Q
o z: z: z:
z:
st- O Ol sl-
1 OO VO VO VO
VO O LO VO
Ol OO CD 00
vo
"•—* 1 1 1 1
st- co in in vo
Ol s|- 00 t— i VO
sf O O O 0
^_^
LO 1 1 1 I
oo oo si-
O CM CO si" Ol
vo • • • •
CO LO Si- CO St-
00
s • r** r^ rv*
oo a: i i i
1 VO t-~ CO
fli
^*
.
CJ
o.
CO
ID
C
o
?^
CO
O
CO
r~"
UJ
i |
(JU> .
to
-
^>
CO
CO
^>
>,
i.
o
•p
CO
%-
o
J3
CO
r—
•P
to
1
C
UJ
a:
o;
a.
Ill
l-l-l
CO
«_
*
t— I
,
o
^~-
^^
o
•p
cb
O
^\
-If
CO
,
rtl
tu
to
(U
c
cu
A
0)
,__ ^
I""
"*Z'
X
c
^
o
i-
o
JD.
.
-.^J
1—4
C
o
« %
*r*
to
i.
o
Q.
S~
0
0
cn
o
ai
*-~
c-
o
•p
CO
c»
S-
O)
-p
*~H
s
CM
. • .
0
^~
*^.: ;
o
•p
CO
0
CO
+
TJ
ai
3
C
•r-
•P
O
O
196
-------
Q.
(/)
I/)
cu w -o
CO J= CU O
S- -4-> -C .C
4J f_ +J
c •> cu
CU CU E
o s- •
c: o oj o
O M- •«-
O CU >> i-
.S- S- -M
CQ CU O CU
O JE -t-> E
CX +•> CO O
r— ••> O 4->
CO CO -Q O
•i- CO CU
+J J= _J
•i- OJ
C 3 >,
•i- o .a
s_
CO .C T3
.E -tJ cu
I— C
t-H -I- C
CO O
• V) +•> T-
•o cu .a +->
cu.f: o 3
•1->.,0 i—
O 4-> to O
CU CO CU
I <*-
>> "O .
s- >> cu -^
S- +J O
O «r-
CU f
S- CX
S- CO
o s-
o a>
1
o
4-
OJ
-o
>v CU
s- •«->
o o
4-> CU
CO S-
cu
I
o
cu
o
U
cu
co T 6
s- s- >,+->
CO O S- CO
_ fl
E O
0-H
s- w
Sd
cu
o
o
O)
V cyj™
._ _ cu co -a
s- s- ^ en s-
O O +-> CO
ja o -o
CO .C $= •
_| +J 4-> CO ^O
O T- «5f
>,C 3 COO
ja o
co
-o
c
(O
OJ
in
0)
o
•)->
CO
CQ
C
O)
CO
(U
s
es
c
cu
CO
ex
•o
cu
c
o
o
UJ
s
•o s-
CU>cO
S- f-t
§L»
cu s-
S- O
vt ca
c s-
o o
•r- JD
4-> CO
CO —1
O) •!-> CO
S-' O *^^
CO. CO
ex s- vt
E 4-> -O
O X O
o cu .c:
CO
o -o
c cu
O -M
o s-
o
cu ex
jr: cu
I— s-
E CD O!
cu c: s:
JZ •!-
3 a.
S UJ
o -a
r- cu E
c o
CU T- S-
S- CO 4-
co -l->
ja "O
c/) O CU
cu ••->
3 CU CX
i— S- CO
CO CU T3
> S CO
(n
•I—
to
ca
CO
cu
ca
o
CX
T3
s_
co
CO
(/)
o
U
o
cu
en i/i
CO O
S- CX
co E
> o
CO
cu
3
s
•o
.cu
-p -I-"
•r- CO
•P £
i— 4-»
CO
CO t—
ca
CX
r
4-> -O >5
CO i—
S- N CU
o >> +J
i— ca
CU CO C
CD C -r-
ca ca -a
S- S-
CU +J O
> o c
c
o
o
CO
cu
•o
cu
•*->
o
CO
cu
cu
c
o
II
a
cu
I
o
to
cu
o
CO
ca
CM CO
CO
197
-------
of the pre- and post-KPEG-treated soils from both laboratories, as well as
the percent PCB reduction, PCB concentration in the collected condensate, and
PCB concentration in the neutralized KPEG reagent. The PCB concentrations of
the untreated soil from the corroborative analyses were reported as re-
covery-corrected. On-site PCB concentrations of the untreated soil were not
recovery-corrected when reported, thus justifying the consistent discrepency
between the two laboratory facilities.
The PCB analyses in Table I indicate that the lowest PCB reduction was
99.58 percent, while the majority of the reduction rates exceeded 99.9
percent. Low concentrations of PCB were identified in the collected conden-
sate. For each 1.5-cubic-yard batch of soil that was treated, approximately
50 gallons of condensate was collected. Using Batch 4 as an example, initial
PCB concentration of the 3400 Ib of soil charged into the mixer was 3778 ppm;
thus, the total quantity of PCB can be calculated by Equation 3:
Total Ib PCB in soil = Ib soil x
(Eq. 3)
Placing Batch 4 values into Equation 3 indicates that 12.8 Ib of PCB were
contained within the batch. Analysis of the condensate from Batch 4 indi-
cated a PCB concentration of 13.81 ppm. The total quantity of PCB in the
condensate can be calculated by Equation 4:
Total Ib PCB = Gal condensate x pCB
in condensate 1 x 10
> (ppm)
> 4)
Again using values obtained from Batch 4, the total quantity of PCB's trans-
ferred from the soil to the condensate was 0.006 Ib of PCB. Therefore, the
total quantity of PCB's transferred from the soil in the mixer to the conden-
sate was less than 0.04 percent of the overall quantity, assuring that the
reduction of PCB's in the soil was not the result of PCB relocation into the
condensate by steam stripping.
The collected condensate was passed through an activated carbon system
to remove the residual PCB's. All analyses of carbon-treated condensate were
reported as nondetectable. Treated condensate was collected and transported
to the sanitary sewer for discharge.
The results presented in Table I indicate the efficacy of the KPEG
treatment for the dechlorination of PCB's. The operation of the system,
which was capable of treating 1.5 cubic yards per batch utilizing equipment
that is readily available for scale-up, indicates that scale-up to full scale
is conceivable, assuming favorable economics. The operation of the system
was performed without major mechanical or operational problems. Therefore,
full-scale design need only enlarge the system and incorporate minor changes
to improve operations, particularly the materials handling aspects of the
treatment process.
DISCUSSION OF PROCESS ECONOMICS
The economics of the KPEG chemical dechlorination process must indicate
a favorable advantage when compared with other treatment and disposal
198
-------
practices for the development of the process toward a full-scale system.
The demonstration performed on the RGB-contaminated soil in Guam was
performed solely to demonstrate the feasibility and potential of the KPEG
technology. The system was purposefully designed to be labor-intensive for
two reasons: (1) as a demonstration focusing on the efficacy of the KPEG
reagent to dechlorinate the PCB, the system was not automated in order that
capital expenditures be reduced, and (2) the use of hands-on operation would
allow for observations and overall intimate knowledge of the system to be
acquired. To reduce capital costs even further, all system equipment that
could be leased was utilized in place of purchased equipment. These lease
charges were initially incorporated into the operational costs per unit of
contaminated soil, which artificially elevated the operational costs. Other
factors included in the initial operation costs were all labor, associated
per diem, and automobile charges; major equipment items leased from the U.S.
mainland; major equipment items leased while in Guam; diesel fuel; process
chemicals; electrical consumption; and contractor profit and overhead.
Taking into account a weekly expenditure (excluding unusually uncharacteris-
tically high costs associated with performing work in Guam) where six batches
of 1.5 cubic yards (3400 Ib) of PCB-contaminated soil were treated per batch,
the operational costs were calculated by PEI to be $1700-1800/ton of PCB-
contaminated soil.
The system utilized in Guam only demonstrated the potential use of KPEG
reagent for PCB dechlorination. During the demonstration, no attempt was
made to optimize the reagent formulation or operating parameters. Since
returning from Guam, U.S. EPA/RREL and PEI have continued laboratory treat-
ability studies to optimize the reagent formulation and operating parameters.
From the data concerning a modified KPEG reagent formulation and reduced
constraints upon operation, as determined by the U.S. EPA/RREL laboratory op-
timization studies, a full-scale, portable, self-supportive treatment system
has been preliminarily costed. The estimated capital expenditures for a
full-scale system, including equipment and construction costs, are $3.5 to
$4.5 million. The system would theoretically be capable of dechlorinating 72
tons of PCB-contaminated soil daily, with a more realistic throughput of 54
tons per day. Assuming that all equipment is purchased outright and mate-
rials handling is automated to reduce excessive labor, operational costs are
estimated to be $200 to $300/ton. This cost includes all direct labor and
indirect living costs for an out-of-town operational crew, diesel fuel con-
sumption, and chemical usage. This cost, however, does not include exca-
vation of the contaminated soil or placement back onto the ground following
treatment; therefore, the overall cost to the client will be slightly higher.
The purpose for presenting the operational costs without including the
excavation or replacement costs was so that a cost comparison can be made
with the current treatment practice of incineration for PCB-contaminated
soils. A telephone poll of several independent PCB-permitted incinerators
throughout the country indicated an incineration charge averaging $1713 per
ton of PCB-contaminated soil. This incineration charge does not include
excavation; loading into Department of Transportation (DOT) .approved drums,
since the majority of incinerator facilities will not accept contaminated
199
-------
soil in bulk; transportation of soil to the incinerator site; ash disposal;
or any cost associated with replacement of excavated soil with clean fill.
CONCLUSIONS
Based on the projected cost estimates that have been performed, the APEG
chemical dechlorination process appears to be superior in economics, assuming
the optimizations obtained in the laboratory are suitable for full-scale
scenarios. Advantages of the APEG process are that transportation costs of
contaminated soil and replacement costs of clean fill are eliminated, since
treatment is performed on site and treated soil should be suitable for place-
ment back onto the ground. Prior to initiating design of the full-scale
system, additional field-scale demonstrations will be performed utilizing the
optimized parameters determined by U.S. EPA/RREL, NCEL, and PEL These
field-scale demonstrations are currently being planned for Fall 1'989.
ACKNOWLEDGEMENTS
PEI is grateful to the technical support and guidance provided through-
out the course of the KPEG. field-scale demonstration and on-site sample
analysis on the Island of Guam, U.S.A., by Dr. Alfred Kernel and Mr. Harold
Sparks of the U.S. EPA Risk Reduction Engineering Laboratory, Cincinnati,
Ohio. PEI is also grateful to Mr. Gorman Dorsey for use of the USN PWC FENA
laboratory where on-site sample analysis was performed as well as Mr. Jess
Lizama, Supervisory Environmental Engineer of the USN PWC site, Guam, for
arrangements of construction equipment, supply of utilities, and USN PWC
personnel for assistance of systems assembly and disassembly.
REFERENCES
1. laconianni, F. J. Destruction of PCBs—Environmental Applications of
Alkali Metal Polyethylene Glycolate Complexes. Prepared for U.S. En-
vironmental Protection Agency, Hazardous Waste Engineering Research
Laboratory, Cincinnati, Ohio. Franklin Research Center, Philadelphia.
May 31, 1985.
2. DeMarini, D. M., J. E. Simmons. Toxicological Evaluation of By-Products
From Chemically Dechlorinated 2,3,7,8-TCDD. Accepted for publication in
Chemosphere, 1989.
3. Peterson, R. L., E. Milicic, and C. J. Rogers. Chemical Destruction/
Detoxification of Chlorinated Dioxins in Soils. In: Incineration and
Treatment of Hazardous Waste, Proceedings of the Eleventh Annual Re-
search Symposium. EPA/600/9-85/028, September 1985.
4. Peterson, R. L. 1986 Method for Decontaminating Soil, U.S. Patent
Number 4,574,013, March 4, 1986.
200
-------
DESTRUCTION OF CYANIDES IN ELECTROPLATING WASTEWATERS USING
WET AIR OXIDATION
by: H. Paul Warner
USEPA
Cincinnati, OH 45268
ABSTRACT
Many of the technologies normally applied for the destruction of cyanide
in waste streams containing cyanides and metals result in the generation of
sludges which contain high concentrations of cyanide (200 to 1000 mg/kg), most
of which is strongly complexed with constituent metals. In order to reduce
the total cyanide content of these sludges, we investigated the possibility of
treating the original wastestream by a technology which would destroy the
cyanide, both free and complexed, prior to precipitation of the metals. This
paper presents the results of the application of wet air oxidation for cyanide
destruction prior to sludge generation. Experience and engineering judgment
strongly suggest that this technology could also be applied to liquids and
sludges generated by other technologies which contain high concentrations of
cyanides.
INTRODUCTION
Pursuant to Section 3004 (m) of the Resource Conservation and Recovery
Act (RCRA), enacted as part of the Hazardous and Solid Wastes Amendments
(HSWA) on November 8, 1984, the Environmental Protection Agency (EPA) is
investigating alternatives for treating cyanide-containing electroplating and
heat treating wastes prior to placement in landfills. The Agency has
previously established treatment standards for metals in the sludges from
these wastes with the First Third Listed Hazardous wastes (53 FR 31137, August
17, 1988). Treatment standards for the wastewater from these wastes were soft
hammered with the First Third rule. The Agency is now developing standards
for all of these wastes and will, with additional regulations, set standards
for cyanide. Cyanide standards were reserved by the Agency in the First
Thirds rule. This paper will discuss the application of wet air oxidation as
one of the technologies which can be applied for treatment of cyanide-
containing electroplating, specifically, spent plating bath (F007).
For the discussion that follows, it must be pointed out that data
related to the raw waste selected for treatment have been claimed as
Confidential Business Information (CBI) by the generator and cannot be used in
201
-------
this presentation. However, by using constituent concentrations of a
"typical" F007 waste, a relatively accurate evaluation of this treatment
process can be made. The constituent concentrations of this "typical" F007
waste will be the average of other F007 wastes as found in the Agency's
proposed background document for the regulation of cyanides (1).
It should also be pointed out that prior to the pilot scale work, a
bench scale run was made which, due to the corrosivity of the raw waste,
mandated the use of titanium as the material of construction for the treatment
system. Only the smallest scale system was contructed of titanium and
therfore was chosen for this project.
APPROACH
Alkaline Chlorination is one of the most commonly applied treatment
technologies for F007 wastes, and is usually followed by precipitation,
clarification, and finally dewatering of the generated sludge. In a well
operated system, the final liquid discharge meeting effluent discharge
concentration regulations may be directly discharged to a surface stream or
Publicly Owned Treatment Works (POTW). However, the dewatered sludge will
normally contain, along with varying concentrations of regulated metals, high
concentrations of total cyanide (200 to 1000 mg/kg). These cyanide
concentrations would more than likely restrict continued land disposal of the
dewatered sludge without additional treatment. Taking into consideration the
necessity of disposal, alternative treatment techniques were evaluated. Wet
air oxidation of F007 wastes has been successfully demonstrated by
Zimpro/Passavant in Casmalia, California. Their technology was selected for
further evaluation, and for the generation of data in support of the Agency's
regulatory program.
WET AIR OXIDATION PROCESS (2)
Wet air oxidation is the liquid phase oxidation of organics or
oxidizable inorganic components at elevated temperatures and pressures.
Oxidation is brought about by combining the wastewater with a gaseous source
of oxygen (usually air) at temperatures and pressures in the range of about
175° to 327°C (360° - 620°F) and 2069 to 20,690 kPa (300 - 3,000 psig),
respectively. The solubility of oxygen in aqueous solutions is enhanced at
elevated pressures, and the elevated temperatures provide a strong driving
force for oxidation.
Wet air oxidation has been demonstrated at bench-scale, pilot-scale, and
full-scale as a technology capable of breaking down hazardous compounds to
carbon dioxide and other innocuous end products. Cyanide in electroplating
wastes is converted to carbonate and ammonium ions when oxidized as shown by
the reactions:
2 MTaCN + 02 + 4 H,0 = Na2COa + (NH^COj
The major processing steps in the wet air oxidation -process are
wastewater pressurization/air compression, preheat, reaction, cooling,
depressurization, and liquid/gas separation. Figure 1 is a flow diagram of
the wet air oxidation process.
202
-------
r
in
CO
0
o
o
RJ
13
X
o
<
*"*
©
c
(0
>
C0
in
in
rt
CL
Q.
E
N
O
o
4^
co
E
-------
The wastewater or slurry is brought to system pressure by a high
pressure pump. Air from a compressor may be added directly to the waste or to
dilution water and preheated to raise the temperature of the mixture at the
reactor base such that the exothermic heat of reaction will increase the
mixture temperature to the desired maximum. Preheating can be accomplished by
an external heat source as shown in Figure 1 or by the reactor effluent.
Startup energy is provided by the external heat source to the preheater or an
auxiliary heater. Residence time for the oxidation reaction is provided by
the reactor; the temperature of the wastewater-air mixture rises as the
reaction occurs. The reactor effluent is cooled with cooling water (as shown
in Figure 1) or with the wastewater-air mixture. Cooling is usually to about
95° to 135°F (35-57°C). A control valve reduces the pressure of the oxidized
liquor-spent air mixture. The gas phase is disengaged from the liquid phase
in the separator vessel. Off-gas from wet oxidation systems is usually
treated to reduce the concentration of hydrocarbons. Wet scrubbing, which is
commonly used to cool the gas stream, results in some reduction of
hydrocarbons. Adsorption columns and afterburning provide additional organic
emissions reduction.
The overall F007 treatment system at Zimpro/Passavant may be described
in three operations: (1) a blending step to control feed parameters; (2 wet
air oxidation process; and (3) treatment of oxidized liquor.
Feed Blending Operation
Four fifty-five gallon drums of F007 waste were mixed in a water-heated
stainless steel tank to ensure a homogenous feed composition. The waste
required heating to about 110° to 130°F (43-54°C) to maintain its liquid
state. Below that temperature range, the waste crystallized because of the
high concentration of sodium carbonate, which made handling very difficult.
After the waste was thoroughly mixed, the drums were refilled and placed in a
hot water bath, maintained at 110°F (43°C). As waste was needed in the
performance of the test run, a drum of waste was removed from the water bath,
thoroughly agitated with a portable mechanical mixer, and then pumped into the
treatment system feed tank. The feed tank is equipped with a heating coil,
mechanical mixer, and a recycle line to ensure the feed is maintained at the
proper temperature and is homogenous.
Wet Air Oxidation Treatment
The steady-state operating conditions of the Zimpro/Passavant wet air
oxidation process are shown in Table 1. All the parameters listed in Table 1
are key operating parameters; however, the single-most important parameter
used in determine steady-state was maintaining the reactor outlet temperature
of about 450°F (232°C).
Operating phases of the wet air oxidation process include: warm-up
{with tap water followed by waste feed), stabilization, steady-state, and cool
down (with tap water). The warm-up period with tap water typically requires
about four hours followed by an additional two hour warm-up with the
waste feed. Waste feed warm-up continues until the operating conditions were
204
-------
within approximately 5 percent of the desired conditions, which signals the
beginning of the stabilization period. During the stabilization period the
operating conditions were continually fine tuned. Typically, 1 to 3 residence
times ( i.e., about 1 to 3 hours) were allowed before steady-state is begun.
The steady-state period continues as long as needed to collect all the
required samples or until the system falls outside steady-state conditions
specified in Table 1. Following steady-state, the system is switched to tap
water and cooled down over a period of about 6 to 8 hours.
TABLE 1. SELECTED OPERATING PARAMETERS OF THE WET AIR OXIDATION PROCESS
Operating parameter
Steady-State conditions
Reactor inlet temperature
Reactor outlet temperature
Reactor pressure
Waste feed rate
Dilution water feed rate
Dilute nitric acid feed rate
High-pressure air injection rate
Residual oxygen content of the off-gass
430° - 470°F
440° - 480°F
1700 psig
2.5 - 3.0 gal/h
2.5 - 3.0 gal/h
0.5 - 0.6 gal/h
60 -80 scfh
16 - 20%
In the wet air oxidation treatment process dilution water (i..e., tap
water) is pumped at a rate of 2.72 gal/h (1745 psig) and combined with 1.1
fta/min of compressed air prior to passing through an oil preheater (Figure
1). The preheater heats the tap water/air stream to about 520°F (271°C) as it
enters the base of the pilot-scale titanium reactor at 1710 psig.
Approximately 2 feet from the bottom of the 3-in. i.d., 15-ft-long reactor,
the waste feed is pumped into the reactor at a rate of 2.75 gal/h. Total
influent flow rate to the reactor is 5.47 gal/h. Heat tapes, spaced evenly
along the length of the reactor at 15-in intervals, along with the heat of
reaction maintain the temperature at about 450°F (232°C). The oxidized liquor
exits the reactor through one of two exit ports. One port is used as the
reactor outlet for the oxidized liquor, and 18 percent nitric acid is pumped
at a rate of 0.55 gal/h (1735 psig) through the other port to remove any
carbonate plugging at the exit port. Every four hours through the test
run, the valves controlling the exit ports are reversed to ensure that the
reactor does not plug. From the reactor, the oxidized liquor passes through a
tube-in-tube water cooler that brings the temperature down-to about 106°F.
After the cooler, the liquor passes through a pressure control valve that
returns the wastewater to atmospheric pressure. The wastewater then enters a
gas/liquid separator. The oxidized liquor is collected from the bottom of the
205
-------
separator and the off-gas passes through a caustic scrubber containiing a 5
percent NaOH solution. The oxygen content of the off-gas is monitored
continuously with an Oj Meter, and a dry gas meter measures the off -gas flow
rate prior to release to the atmosphere. Off -gas samples before and after the
scrubber are analyzed for methane, and total hydrocarbons by a gas
chromatograph .
Treatment of Oxidized Liquor
In order to discharge the oxidized liquor to the POTW, the wastewater
required neutralization and metals precipitation. This was accomplished by
the following process by use of nitric acid and sodium sulfide:
1.
2.
4.
5.
The oxidized liquor was pumped into a 100-gallon holding tank.
Nitric acid was added for pH adjustment. Two types of acid were
used: a) spent acid from the reactor acid wash or b)
concentrated, 38 degree Baume1 acid.
Sufficient acid was added to lower the pH to 8.0 to 8.5.
NOTE: Care must be taken when the acid is added because a large
quatitity of carbon dioxide is liberated and the reaction is
somewhat violent.
Sulfide in the form of sodium hydrosulfide (NaHS) or sodium
sulfide (NaS) -was added to the tank and mixed.
The sulfide precipitates are very difficult to filter; therefore,
diatomaceous earth was added as a filter aid to the slurry to
improve filterability.
A sample of slurry from the tank was filtered and a chip of
sulfide was added to the filtrate. If the filtrate remained
clear, the metal ion precipitation was complete. If a brown
precipitate appeared, more sulfide was required. Additional
sulfide was added t'o the holding tank until the filtrate remained
clear when a sulfide chip was added.
A small plate and frame filter press was used for the filtering.
The press cloths were precoated with diatomaceous earth prior to
filtering the sulfide solids to improve filterability and prevent
the sulfide solids from blinding the filter cloth. Filter cake
generated from the treatment of the oxidized liquor was disposed
of at a hazardous waste landfill.
Treatment Operational Problems
The most significant operational problems were plugging and/or scaling
within the oil preheater and reactor . These problems were anticipated prior
to the initial run due to high suspended and dissolved solids concentrations
206
-------
in the raw waste and did, in fact, result in the termination of the initial
run. There were other operation problems (valve and guage failure), however
they were more than likely due to the fact that the pilot systems had not been
operational for over a year. After the initial run termination, modifications
to the reactor (removal of internal mixing baffles), the oil preheater
(repiping around the preheater), and further dilution of the raw waste,
allowed quite successful completion of the second run. The sampling program
required 24 hours of steady state operation. An engineering representative of
Zimpro/Passavant estimated that the system could be operated for from 7 to 10
days before requiring a complete system acid purge.
Following the completion of the sampling period, the treatment system
was allowed to cool and then the reactor was broken down to visually check the
plugging/scaling effects of the second run. A brown, sand-like deposit was
found at the bottom of the reactor. This material did not adhere to the
reactor walls, but was deposited loosely at the reactor bottom. This would
indicate that it was held in suspension in the reactor during operation and,
given sufficient operating time, would probably discharge from the reactor
with the oxidized liquor (3).
A second type of material found in the reactor was a scale deposit on
the reactor walls. This material adhered to the reactor walls and would not
discharge from the reactor during operation. Removal of this scale would be
required by the previously mentioned acid purge after a yet undetermined
operating period. Neither the loose material in the bottom of the reactor nor
the reactor scale should prevent effective treatment of F007 waste by the wet
air oxidation technique. However, it could lower the heat transfer
coefficient.
RESULTS AND CONCLUSIONS
As previously mentioned, in order to evaluate the efficiency of wet air
oxidation in this presentation, a "typical" F007 waste constituent
characterization had to be generated for the feed stream (raw waste data
protected by CBI). Incorporating this "typical" waste data in the calculation
of percent removal for cyanide, 99.9% removal is observed for total and
amenable cyanide as seen in Table 2. The destruction of cyanide in the wet
oxidation process increased ammonia concentrations from minimal in the raw
waste to an average of 5400 mg/1 in the oxidized liquor. Cyanide was not
detected in the off-gas scrubber water above its practical quantitation limit
of 0.25 mg/1.
As seen in Table 2, a considerable amount of the copper and zinc is
apparently removed by the wet oxidation process. A mass balance around the
process, including concentrations of the metals in the scale, bottom solids,
and the acid wash following completion of the test run, reveals that the
metals are concentrated in the scale and the bottom solids. In the mass
balance calculations, metals concentrations from the actual raw waste were
used. However, as previously pointed out, the raw data are not presented
because of a claim of CBI by the generator.
207
-------
TABLE 2. PERCENT REMOVALS ACROSS OXIDATION/PRECIPITATION PROCESS
F007*
Raw Feed
mg/1
CN, 33000
total
CN, 31000
amenable
Cu 5000
Zn 10000
Oxidized Percent Filter Press Percent
Liquor Removal Filtrate Removal
mg/1 mg/1
2.51 99.9 1.65 99.9
0.05 99.9 ND 100
802 84.0 2.45 99.9
6.47 99.9 2.45 99.9
*Rounded averages from other F007 wastes.
ND=Not detected.
Conclusions from this study are as follows:
1.
2.
REFERENCES
Document 1.
Report 2.
Report 3.
Wet air oxidation is an effective treatment method for the
destruction of cyanides in F007 wastes, including complexed
cyanides.
Wet air oxidation, when followed by sulfide precipitation of
metals, is an effective treatment system for complete
treatment of F007 wastes.
Engineering judgment and years of experience predict that
other cyanide wastes containing metals, both liquids and
sludges, could effectively be treated, after appropriate
concentration or dilution, by the wet air
oxidation/precipitation technology.
USEPA, Office of Solid Waste, Proposed Best Demonstrated
Available Technology (BDAT) Background Document for Cyanide
Wastes, December 1988.
USEPA, 1988 of, Office of Solid Waste, On Site Engineering
Report of Treatment Technology Performance and Operation for
Wet Air Oxidation of F007 at Zimpro/Passavant, Incorporated
in Rothschild, Wisconsin, Washington, DC.
Zimpro/Passavant, Final Report for the Pilot Plant
Demonstration Study on Wet Air Oxidation of F007
Electroplating Cyanide Wastes, Rothschild, Wisconsin, June
1988.
208
-------
DETERMINING COST EFFECTIVE APPROACHES TO THE ENVIRONMENTAL
CONTROL OF ELECTROPLATING OPERATIONS
by: John 0. Burckle
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Cincinnati, OH 45268
ABSTRACT
The U.S. EPA has sponsored a number of studies over the past few years
to determine improved approaches available to achieve more cost-effective
control of electroplating operations. These studies indicate that a
multimedia process systems approach is likely to be the most effective in
attaining the environmental pollution goals required by regulatory agencies
while minimizing costs to do so. The approach involved requires a number of
steps. These include: (1) the careful analysis of the plating processes
utilized and definition of the characteristics and the pathways of pollutant
generation; (2) the analysis of the contribution of pollution prevention
techniques to minimize the quantity of wastes formed; (3) selection of a
combination of pollutant controls and ultimate disposal technologies which are
allowable under the FWPCA, CAA and RCRA; (4) and finally, the development of
a route to achieving the required environmental protection goals at least
cost. This approach is based upon an iterative process which takes into
consideration the capital and operating costs of a number of alternative
pathways consisting of various combinations of pollution prevention, control,
and ultimate disposal alternatives. The target sought is the scenario in
which the multimedia environmental goals are obtained for all aspects of the
operation at the minimum annualized cost for capital and operating expenses.
An acceptable target, from the perspective of cost, is that scenario in which
the annualized cost for installation of pollution prevention and control
technologies, including the costs of residual management, is offset by the
savings achieved in raw materials wastage and waste treatment and disposal
costs for the existing waste management case.
209
-------
INTRODUCTION
This paper attempts to summarize the results of the work we have
conducted over the last several years to address pollution control in the
electroplating sector on a multi-pollutant, multimedia basis. Currently the
Agency promulgates standards on a media-specific basis, i.e., for air, water,
and hazardous wastes. The current standards-setting process does not provide
a "systems analysis" approach to minimizing residual pollution discharged to
the environment for all media or all process pollutants. The requirements for
each media are considered separately and in a stepwise fashion. It was the
purpose of our efforts to develop an integrated, systematic approach to the
selection of various techniques for the reduction and treatment of wastes
which would optimize the effectiveness of reducing the overall multimedia
impacts.
The discharge of wastewaters generated in electroplating processes into
receiving waters or sewers are regulated at the Federal level under the
Federal Water Pollution Control Act. Regulations to achieve the effluent
limitations established for electroplating operations require the use of Best
Applicable Technology, "BAT", for existing sources and New Source Performance
Standards, "NSPS", for new sources. The effect of these regulations on the
treatment of electroplating wastewaters are illustrated in Figure 1. The
National Pollutant Discharge Elimination System, "NPDES", requires a permit
for any discharge into a receiving waterway (or tributary) and imposes
requirements for pretreatment of wastes discharged to sewers.
SKIMMED OILS
'ASTE TOXIC ORGANICS
HAUL OR
RECLAIM
^ TREATED
EFFLUENT
CONTRACTOR
REMOVAL
Figure 1. Treatment Required by the Federal Water Pollution Control Act
(BAT Level-NSPS is the same except for requirements for
recovery of Cadmium)
210
-------
The disposal of residual wastes to the land is regulated under provisions
of the Hazardous and Solid Waste Amendments of 1984. These amendments of the
Resource Conservation and Recovery Act require that the Agency ban RCRA wastes
from land disposal unless it is determined that land disposal is more
protective of the public health and the environment than available alternative
technologies. This act also requires that the Agency assess and report on
strategies to increase the use of recycling and waste minimization techniques.
Under the requirements of this Act, the Agency has established the "Best
Demonstrated Available Technology" for electroplating wastes and is
investigating the contribution which can be made through implementation waste
minimization techniques. .
Historically, we have advanced from a time when no treatment was required
to a time of zero discharge under the FWPCA, with pollutants removed from the
air and water being discarded onto the land to leach back into the ground and
surface waters. The enactment of the Hazardous & Solid Waste Amendments with
its "Landfill Ban" and "waste minimization" provisions has brought us full
cycle. There is now no place left to go for the disposal of the pollutants
and other hazardous wastes. There is now little alternative to taking action
to reduce the production of toxic pollutants through product substitution and
process changes, to utilize recycle and reuse opportunities, and, as a last
alternative, to detoxify as best we can any residuals which remain prior to
land disposal.
OVERALL CONSIDERATIONS
The major source of pollution generation in a well-operated and maintained
electroplating shop is the plating chemicals which are lost from the plating
bath when the plated parts are removed. This is known as "drag-out". In
plating operations which do not take measures to control drag-out, plating
chemical losses range from 50 to 90% -- that is from 50 to 90% of the plating
chemical is removed from the bath in the form of drag-out and rinsed down the
drain. The drag-out is removed from the plated workpiece by rinsing in water.
Free, or once through rinsing had been the usual practice until recently. It
has been recognized by those "in the trade" that this type of rinsing produced
by far the major waste stream requiring extensive treatment to meet the
requirements of the FWPCA.
A number of approaches are available to control the polluted wastewaters
generated in electroplating. The most obvious is the treatment of the
wastewater to remove pollutants to levels acceptable for discharge into the
environment. However, wastewater treatment can prove very expensive when
applied to conventional electroplating processes owing to large wastewater
volumes resulting from free rinse operations.
The size and cost of the equipment required to remove pollutants to levels
acceptable for discharge depend, in large part, upon the wastewater volumetric
flow rate in addition to the inlet and outlet concentrations of the polluting
substance. Where chemical processes are used for pollutant reduction, greater
quantities of wastewater treatment chemicals are required per unit of
pollutant removed as the wastewater concentration decreases. This results in
211
-------
proportionally greater sludge formation. Inefficient design and operation,
therefore, significantly affects the interrelated factors of materials
consumption, pollution control, and costs.
Because of increasing prices for electroplating chemicals, process water,
pollution control systems and reagents, sewage charges, and residual waste
disposal, it is highly desirable to incorporate certain in-plant changes into
existing plating operations. These in-plant changes result in the dramatic
reduction of the production of wastewaters requiring treatment and disposal.
Reduction of wastewater flow with recovery and recycle of plating bath
chemicals has become the preferred technique, not only for enhanced control
of environmental pollution, but also for the significant economic advantages
inherent in the application. Wastewater flow reduction can be achieved
through two approaches (Figure 2). The first approach involves certain
in-plant process changes such as application of techniques to reduce drag-out,
minimize rinse water usage, and recycle rinse water directly to the plating
bath. The second involves application of controls at the electroplating bath
to concentrate plating chemicals from rinse waters for recycle to the bath
such as evaporation, ion exchange, or electrolyte recovery. Those techniques,
when systematically implemented, result in significant cost savings in bath
chemicals, process water, wastewater treatment, sewage, and waste disposal.
Workpiece
• -»—
Chemical
recycle
Bath
purification
n' r
L__J
Plating
bath
•^—
•*~
i r
i • i.
L_>f
Recovery
unit
B— 4j
L.4
1 r
i r
S 1
1 -J
Rinse tanks
«—
Rinse recycle
Makeup
^~ water
(•) Closed Loop
Workpiece
••*
Chemical
recycle
Bath
™ purification
1 1
1 1
1 1
l_ J p.
Platin
bath
i r
'1
Recovery
unit
•
1 r
H h
I.J
Rinse tanks
•i
"~1 r
i
L_J
L
(b) Open Loop
To waste
treatment
Makeup
water
Figure 2. Waste Minimization Approaches
212
-------
The resulting cost savings may be so significant that a plater cannot
ignore their economic potential even if pollution control was not recognized.
As shown in Table 1(1), the replacement cost for plating chemicals is a major
operating cost item. How many platers can afford to discard 50 to 90% of
their chemical raw materials? In addition to providing chemical savings and
reducing water use costs, in-plant changes provide a basis for a pollution
control system design. Waste treatment equipment needs, whether wastewater
concentrating techniques such as ion exchange or conventional end-of-pipe
treatment systems, often will be reduced significantly by in-plant changes.
And the potential impact on pollution control is also significant.
TABLE 1. ESTIMATED COSTS OF PLATING CHEMICAL LOSSES
Cost
Plating Chemical Rep!
Nickel
As NiS04
As NiCl2
Zinc cyanide as Zn(CN)2
Using C12 for cyanide oxidation
Using NaOCl for cyanide oxidation
Chromic acid as H2Cr04
Using S02 for chromium reduction
Using NaHS03 for chromium reduction
Copper cyanide as Cu(CN)2
Using C12 for cyanide oxidation
Using NaOCl for cyanide oxidation
Copper sulfate as CuS04
acement
1.19
1.05
2.00
2.00
1.18
1.18
2.62
2.62
0.88
Treatment
0.20
0.30
1.03
2.02
0.51
0.84
1.02
2.30
0.20
Disposal
0.35
0.41
0.57
0.57
0.52
0,52
0.50
0.50
0.36
Total
1.74
1.76
3.60
4.59
2.21
2.54
4.14
5.42
1.44
Wastewater concentration of 100 mg/L assumed.
Disposal of dewatered sludge (20 percent solids).
MINIMIZING DRAG-OUT
Drag-out losses can be minimized either by reducing the amount of plating
solution which leaves the plating bath, or by recycling plating chemicals in
the rinse water to the plating batlr . Many devices and procedures can be
used successfully to reduce drag-out. These techniques usually are
employed to alter viscosity* chemical concentration, surface tension, speed
213
-------
of withdrawal, of the workpiece from the plating tank, position of the
workpiece on the plating rack, and temperature. Also drag-out tanks are used
for capturing lost plating solution and returning it to the bath. Most
drag-out reduction methods are inexpensive to implement and are repaid
promptly through savings in plating chemicals. An additional savings many
times the cost of the changes will be realized once a pollution control system
is installed. The reduced drag-out will decrease the need for treatment
chemicals and subsequently, the volume of sludge produced. The
cost-effectiveness of each of these methods is discussed below.
There are also several simple methods of drag-out recovery that should be
considered. Four simple drag-out recovery methods are: drainboard, drip tank,
spray rinse, and air knife. The drain board is the simplest method of
drag-out recovery. It can capture drips of plating solution as racks and
barrels are transferred between tanks (Figure 3). Not only do drain boards
save chemicals and reduce rinse water requirements, they also prevent
unnecessary floor wetting.
Workpiece
Drip bar
Concentrated
solution
Figure 3. Drain Board
A drip tank is an ordinary rinse tank that, instead of being filled with
water, simply collects the drips from racked parts and barrels after plating
and before rinsing. When a sizable volume of solution has been collected in
the drip tank, it can be returned to the plating bath. Using a drip tank
tends to restrict the potential use of a rinse tank. As will be discussed,
an additional rinse tank used as a drag-out tank or in a series arrangement
may be more beneficial. The determining factors are the volume of drag-out
and the evaporation in the plating bath.
Spray rinses are ideal for reducing drag-out from plating tanks on
automated lines. As the workpiece is withdrawn from the plating solution,
214
-------
the workplace, is rinsed by a spray returning as much as 75% of the drag-out
chemicals back to the plating tank. Spray rinsing is best suited for flat
parts, but will reduce drag-out effectively on any plated part. The volume
of spray rinse may not exceed the volume of surface evaporation from the
plating tank.
An air knife reduces drag-out in much the same way as a spray rinse, and
is particularly useful when the surface evaporation rate in the plating bath
is low. Air knives reduce the volume of drag-out adhering to the workpiece
by subjecting the workpiece to a high-velocity stream of air. The drag-out
is returned to the plating bath without changing its concentration.
RINSE WATER MANAGEMENT
The rinsing operation can account for as much as 90 percent of a shop's
water usage. Therefore, careful management and conservation of water in the
rinsing operation offers the greatest opportunity for significant reductions
in water consumption and wastewater operation. To achieve the desired degree
of cleaning, the rinsing operation must include: (1) turbulent motion between
the workpiece and rinse water, (2) adequate time of contact between the
workpiece and rinse water, and (3) contact with rinse water of sufficiently
low concentration to effect the dilution of the plating bath washed off the
surface of the workpiece. These three principles apply to all rinsing
operations, including those using flow-through or still rinse tanks. They
can be implemented in a number of ways, depending upon the feasibility for
application to a specific plating line.
The amount of make-up water required to dilute the rinse solution depends
on the quantity of chemical drag-in from the upstream rinse or plating tank,
the concentration of chemicals in the rinse water, and the contacting
efficiency between the workpiece and the water. It is important that
sufficient turbulence and adequate contact time be employed for all rinsing
operations to minimize the introduction of excess quantities of make-up
rinsewater. The maximum allowable concentration becomes a very important
parameter when the other two parameters are satisfied. In fact maximum
allowable concentration is the governing factor with respect to water use in
a well-operated shop.
Reductions in rinse water volumes can be achieved through use of a number
of conservation techniques including the control of fresh make-up water
introduction based upon dissolved solids content of the rinse water, rinse
water recycling, multiple rinse stages, use of other rinse techniques, and
minimization of drag-in.
Use of a simple method of water conservation is becoming more widespread.
It involves the reuse of rinse water at two or more rinse tanks where the
contaminants in the rinse water after a processing step do not detract from
the rinse water quality at another station. This method is applied most often
to the rinses following acid dips and alkaline cleaners. For example, instead
of using 19 liters per minute of rinse water in each rinse tank for a total
of 38 liters per minute, the rinse water used following an acid dip can be
215
-------
reused as rinse water directly after an alkaline cleaner. This practice will
reduce the water use for these two tanks by 50 percent. In most cases,
contamination does not appear to be a problem. In fact, the rinsing following
the alkaline cleaner appears to improve. The mass transfer action
attributable to diffusion is accelerated as the concentration of alkaline
material at the interface between the alkaline drag-out film and the
surrounding water is reduced by the chemical reaction there. Also, alkaline
solutions usually are more difficult to rinse off than acid solutions because
of the higher viscosities, so neutralization aids in this respect(3).
Other reuse arrangements can be employed where the less contaminated
overflow from critical or final rinsing operations is reused for intermediate
rinse steps, such as acid dips and alkaline cleaning steps. The rinse water
following a nickel plating bath can be routed to the rinse tank following the
acid dip. This rinse water, in turn, could be routed to the alkaline cleaner.
Choosing the optimal configuration requires analysis of the particular
rinse water needs. Interconnecting rinsing systems might make operations more
complicated, but the cost advantage justifies the extra attention required.
The benefits from reusing rinse water are limited, however, because that
method of conservation cannot be applied to all rinse operations. Methods
exist, however, that can be applied more widely, and that result in more
dramatic water savings. The three most common methods are parallel and
countercurrent rinsing and the use of still rinse tanks.
Workplace ••*•
Overflow pipes
Air agitation
Figure 4. Three-Stage Countercurrent Rinse Arrangement
216
-------
Using a countercurrent rinse tank arrangement, the plater can achieve even
greater water savings than with the parallel system. In this arrangement
(Figure 4), make-up water flows into the rinse tank farthest away from the
plating tank and moves (countercurrently to the work-flow) toward the rinse
tank closest to the plating tank either by gravity or by pumping. The
workpiece is dipped in the least pure water first and in the cleanest water
last. The quantity of chemicals entering the final rinse will be
significantly smaller than that entering a single-tank rinse system. The
amount of rinse water required for dilution will be reduced the same degree.
The relation between volume of rinse water required as a function of
initial concentration in the plating bath, required concentration in the final
rinse tank, and number of rinse tanks is shown in Figure 5. For example, a
typical Watts-type nickel plating solution contains 270,000 milligrams per
liter of total dissolved solids (Cp), and the final rinse must contain no more
than 37 milligrams per liter of dissolved solids (Cn). The ratio of Cp/Cn is
7,300; hence, 7,300 liters of rinse water would be required for each liter of
process solution drag-in, assuming a single-tank rinse system. By installing
a two-stage rinse system, the same degree of dilution is achieved with only
86 liters of water per liter of process solution drag-in, a reduction in rinse
water consumption of almost 99%. The mass flow of pollutants leaving the
rinse system remains constant.
100.000 E~
10.000 =-
1.000 —
100
• iiitml i i 11 mil 11 11 iiill i i i inn! i i 11 mil
10
100 1.000 10.000 100.000
RINSE RATIO
Figure 5. Rinse Water Dilution vs. Rinse Ratio for Multitank
Countercurrent Rinsing
217
-------
A three-stage countercurrent rinse arrangement would further reduce water
consumption to 76 liters/liter of drag-in. The resulting cost savings by
going from a one-stage to a three-stage rinse system would include reducing
water use and sewer fees and reducing the size of the required waste treatment
systems. The investment cost to add two additional rinse tanks is highly
site-specific. For manual plating operations, the major factor affecting cost
is the availability of space in the process area. For automatic plating
machines, the cost of modifying the unit to add additional stations may be as
high as $20,000 per station. Rubber-lined , steel open-top tanks with
appropriate weir plates and nozzles cost anywhere from $1,000 to $3,000
depending on the cross-sectional area required for the workpiece.
A parallel rinse tank arrangement using three rinse tanks is illustrated
in Figure 6. With the parallel feed system, each tank is individually fed
with fresh water. The rate of water flow to each tank should be the same to
obtain the optimal water savings. In this case, each rinse tank receives a
fresh water feed and discharges the overflow to waste treatment. The rinse
ratio required for a parallel rinse arrangement is defined by r = n (Cp/Cn)1/n.
If the rate given in Figure 5 for a countercurrent rinse system with the same
number of rinse tanks is multiplied by the number of rinse tanks, the parallel
rinse water rate can be estimated. Rinse water rates are significantly higher
for parallel rinsing than for countercurrent (series).
Drag-out
Plating
bath
Water
To
waste
treatment
Figure 6. Three-Stage Parallel Rinse Arrangement
RINSING RECOVERY SYSTEMS
The drag-out losses from the plating process can be significantly reduced
by relatively low cost process modifications which lead to an integrated
rinsing-recovery system. This system will offer substantial savings in
plating chemicals where it can be employed. These alternatives are usually
considered only after steps have been taken to minimize drag-out and rinse
water usage.
218
-------
The techniques for drag-out and rinse water management can be used to
formulate a strategy for simple recovery systems using multiple rinse tanks
and a minimum of additional equipment. The strategy takes advantage of the
need to resupply the plating bath losses, particularly the return of the more
concentrated solutions of dragged-out plating chemicals to the plating bath
to make up for water lost by surface evaporation. The amount of chemicals
actually recovered depends on the amount of chemicals lost from the plating
tank, the number of rinse tanks used, the concentration of chemicals permitted
in the final rinse tank, and the rate at which rinse water can be recycled to
the plating tank.
Of these, the rate at which rinse water can be recycled to the plating
tank is usually the most critical; it is primarily dependent on the amount
of surface evaporation from the plating tank. If the evaporation rate can be
matched to the required rinse water rate, the entire volume of rinse water
could be returned to the plating bath. This set-up is referred to as a
closed-loop recovery system. In a closed-loop rinse water system the only
chemical loss is from the drag-out after the last rinse tank, which has a
dilute concentration of plating chemicals. A closed-loop system may be
impractical when:
o a very low final rinse concentration is required and only
achievable through a larger number of rinse stages;
o excessive drag-out is unavoidable because of product configuration;
o plating tank surface evaporation is minimal.
The tank arrangement shown in Figure 7, which consists of a drag-out tank
followed by a flow-through rinse tank, is the simplest recovery system. The
drag-out tank is a rinse tank that initially is filled with pure water.
Evaporation
Workpiece
Plating
bath
Figure 7. Recovery with a Drag-Out Tank
219
-------
The drag-out rinse collects a significant portion of the process solution
carried on the parts, rack or barrel. Air agitation must be used to aid the
rinsing process because there is no water flow within the tank to cause
turbulence. The presence of a wetting agent is helpful(3). As the plating
line is operated, the salt concentration" increases as more work passes through
the rinse tank. Periodically, the strong solution in the drag-out tank is
returned to the plating tank. The volume returned is limited to the volume
made available in the process tank by evaporation.
As a rule, the use of a drag-out tank will reduce chemical losses by 50
percent or more. The efficiency of the drag-out -tank arrangement can be
increased significantly by adding a second drag-out tank. Use of a two-stage
drag-out system usually reduces drag-out losses by 70 percent or more.
The use of drag-out tanks usually results in less water savings than does
parallel or series rinsing. The operational procedure used with drag-out
tanks is responsible for this effect. The rinse water in the drag-out tank
increases in plating salts concentration until a portion is returned to the
plating bath to compensate for evaporative losses. The concentration of salts
in the drag-out tank can reach as high as 75 percent of the plating bath
concentration. Consequently, a significant water flow in the rinse following
the drag-out tank would be necessary to meet the maximum allowable
concentration.
The low final concentration problem can be overcome in many cases by
operating the final rinse as a free (running) rinse, and using the upstream
tanks as a countercurrent rinse-and-recycle system. Using this approach,
significant drag-out recovery can be realized while providing a final rinse
with a low level of contaminants. Figure 8 shows an automatic
rinse-and-recycle system with a running rinse. Level-control devices in the
plating and rinse tanks control the flow of rinse water through the system.
Makeup waten
ILC1
Surface
evaporation j
Workpiece
Plating bath
Two-stage recovery rinse
LC •> Level Control
1
1_
1 1
1
1
1
Running rinse
To waste
treatment
Figure 8. Automated Rinse-and-Recycle System
220
-------
100
80
a 60
40
20
0.2
0.4
0.6
0.8
O
1.0
02 4 6 8
RECYCLE RINSE RATIO (r)
Notes:
n = number of counter-flow rinse tanks in recovery use.
r = recycle rinse (gal/h) + drag-out (gal/h).
Recycle rinse set equal to surface evaporation from batch.
Figure 9. Drag-Out Recovery Rate for
Rinse-and-Recycle Systems
10
the percent recovery of drag-out for such a system as function of the recycle
ratio, defined as the volume of recycled rinse divided by the volume of
drag-out in a given time is shown in Figure 9, (the recycle rate is assumed
to be equal to the evaporation rate). As shown, such systems can recover from
40 to 100 percent of the drag-out.
A relatively new application of multiple rinse tanks is the
drag-in/drag-out configuration (Figure 10). With the drag-in/drag-out system,
the rinse tank preceding the plating bath (drag-in tank) is connected to the
recovery rinse (drag-out tank) following the bath; the recovered drag-out
solution is circulated by a pump. The concentrations of salts in the drag-in
and drag-out tanks remain about equal. When a rack or barrel is processed,
it drags in plating solution to the plating tank, thereby increasing recovery.
The drag-in/drag-out system finds application with plating baths that have
a low evaporation rate. The recycle ratio, which determines recovery
efficiency, is calculated as the volume of recycled rinse plus the volume of
drag-out divided by the volume of drag-out. The recycle ratio, therefore, is
greater with a drag-in/drag-out system than a common recovery tank. If the
evaporation rate is low, the difference between the recycle ratios for common
recovery and drag-in/drag-out systems is significant. When evaporation rate
is high, the difference is less.
221
-------
Recirculate
Rinse
Workplace
Figure 10. Drag-In/Drag-Out Recovery Arrangement
Various other rinsing configurations could be developed by adding tanks.
The choice of a best arrangement is difficult because of the trade-offs
involved between further reducing chemical losses and further reducing the
rinse flow rate. Obviously, the value of the lost chemicals is a significant
cost. Chemical losses also result in additional rinse water and waste
treatment chemical requirements and more sludge.(4) Although complex, the
evaluation and selection of a multiple rinse tank system can be accomplished
by analyzing each rinsing configuration and comparing cost factors, such as
water, sewer, and waste treatment. The results of the evaluation will enable
the plater to determine whether a multiple rinse tank arrangement is
beneficial and to identify the most appropriate configuration.
WASTEWATER TREATMENT
When the techniques for flow reduction have been implemented as far as
possible, wastewater treatment techniques must then be considered to reduce
the remaining pollution burdens to produce a wastewater suitable for discharge
and a concentrated waste for recycle or for waste treatment.
There are a number of processes used to concentrate plating chemicals in
rinse waters for recycle to the plating bath. These processes include
evaporation, reverse osmosis, ion exchange, electrodialysis, and electrolyte
recovery. Several new technologies are being investigated. These are coupled
transport, reversible gel absorption, and freeze crystallization. These
222
-------
processes differ in selectivity for the concentration of the chemicals that
are dissolved in the rinsewater. Raw wastewaters may be concentrated or
treated directly by one of several precipitation techniques to reduce the
dissolved solids content of metals to a level which will allow discharge of
the cleaned wastewater. There are several concentration techniques available,
etc. The concentrated solutions may be recycled to recover the plating
chemicals of value or treated to remove the metals via chemical precipitation,
electrolytic recovery of the metal, etc.
Plating shop wastewaters can be classified into seven groups, each group
relating to a specific type of generic pollutant removal problem. These are
treatment of acidic metal-bearing wastewaters, recovery of precious metals,
destruction of cyanide, control of hexavalent chromium, treatment of complexed
metals, removal of oil and grease, and control of toxic organics. It is
considered best practice to segregate wastewaters containing oil and grease,
toxic organics, precious metals, cyanides, hexavalent chrome, and complexed
metals, as these require pretreatment before precipitation (Figure 1). While
a broad range of alternative technologies exists for treating the wastewaters
from these processes, treatments most widely practiced today are based on the
following technologies.
Common acidic, metal-bearing wastewaters are usually chemically treated,
usually with lime or alkali-earth hydroxide to achieve adjustment of pH and
precipitation of metals. The precipitates are separated from the liquid phase
by flocculation and clarification. Further separation may be achieved by
filtration, where required. A large range of alternative treatment
technologies, which provide wastes which are more environmentally stable than
hydroxide sludges or which provide for removal and capability for recycling,
are also available. These techniques, which are addressed briefly in the
following section, are more costly for wastewater treatment; however, some may
prove useful in controlling overall processing costs when the impact of waste
management costs incurred to comply with land disposal regulations are fully
incorporated into the overall system.
The recovery of precious metals, the reduction of hexavalent chromium,
the removal of oily wastes, and the destruction of cyanide must be
accomplished prior to metals removal. Oils and greases are removed by gravity
separation and skimming of free oils followed by chemical and emulsion
breaking and subsequent skimming for the removal of emulsified oils.
Cyanide bearing wastes are treated with an oxidizing agent (ozone or
chlorine) to destroy the cyanide in the wastewater. Cyanide, as well as being
a highly toxic pollutant, will complex metals such as copper, cadmium and zinc
and prevent efficient removal of these metals. Wastewaters containing
hexavalent chrome are treated with a reducing agent to reduce the chrome to
the trivalent form which can then be precipitated from solution by hydroxide.
Following separate stream treatment, the effluents are combined and the
metals are removed by precipitation and subsequent clarification. Most metals
precipitate as hydroxides although some, such as lead and silver,
preferentially form other compounds (e.g., carbonates or chlorides). The
223
-------
optimum pH for precipitation is generally in the range of 8.9-9.3, although
it will vary somewhat depending on the specific waste composition. The use
of coagulants or flocculants to enhance the effectiveness of clarification
is also required.
Chelating agents are often used in electroplating operations. When metal
chelating agents are present, the wastewater streams containing complexed
metals must be segregated and separately treated. Chelating agents react with
dissolved metal ions to form "chelate complex", which is usually quite soluble
in neutral or slightly alkaline solutions. A waste stream containing metal
complexes may then be separately treated by adjusting the pH to 11.6 to 12.5
with lime to break the complex. Other alternatives may be used such as
sulfide precipitation, ion exchange, or starch xanthate precipitation.
CONCLUSIONS
Applications of waste minimization principles is the most important step
in achieving the desired environmental protection goals. Those principles,
briefly outlined in this paper, are now broadly used because they offer
significant cost savings potential through the conservation of process
materials, reduction in waste treatment requirements, and, in many situations,
the elimination of large quantities of wastes requiring land disposal. Waste
minimization techniques in practice offer a six-month to four-year payback,
usually a 15 to 20 percent return on investment, and a 50 to 90 percent
reduction in chemical wastage and in wastewater operation. The techniques
should be evaluated in the sequence illustrated in Table 2 and compared cost-
wise to the present plant performance based on the data derived from the plant
audit.
After the various alternatives of waste minimization are evaluated and
the generation of a waste requiring land disposal is found to be unavoidable,
it is then necessary to select an optimum wastewater treatment technology from
among the many alternatives. The relative attributes of several alternatives
are given in Table 3. The selection process is not necessarily straight
forward as the technical and economic considerations are affected by a number
of site specific factors. For example, in comparing hydroxide precipitation
processes, the costs of using caustic soda and magnesium hydroxide appear
similar. However, use of magnesium hydroxide may be more desirable when
disposal costs rise because of haul distance as less sludge is produced per
pound of metal precipitated. Also in comparing hydroxide and sulfide
precipitation, the reagent costs are similar for caustic and soluble sulfide
while most disposal cost for soluble sulfide is twice that of caustic soda.
However, sulfide precipitates containing high levels of copper, lead, and
zinc, especially in association with precious metals, may be of value to
primary or secondary metal smelters and refiners. If so the disposal cost is
eliminated, and there may even be a net positive cash flow generated.
224
-------
In addition, the wastewater treatment technique employed should be
selected to produce a waste compatible with the treatment standards obtained
from the application of the "Best Demonstrated Available Technology required
under the provisions of the land disposal restrictions of the Solid and
Hazardous Waste Act of 1984. Standards to be applied to F006 wastes
containing regulated metals have been promulgated based upon a BOAT using
solidification in a matrix of cement kiln dust. However, cyanide-containing
wastes are currently regulated under the "soft hammer" provisions of ^he rule
(Federal Register Vol .53, No.l59/Wednesday, August 17, 1988/pgs 31138-31222).
These provisions require that disposal facilities meet the minimum
requirements of RCRA 3004(o), i.e.rdouble liner, leachate collection system,
and ground water monitoring - or equivalent performance as provided in 3004
TABLE 2. SELECTING TECHNICAL OPTIONS FOR ENVIRONMENTAL
_ CONTROL OF ELECTRQPI ATING OPERATIONS _
1. Process Audit
a. Chemicals Used
b. Processes used
c. Pathways of pollutant generation - multimedia
d. Characterization of pollutant generation - multi pollutant
2. Waste Minimization Techniques
a. Drag-out minimization
b. Rinse water minimization
c. Rinsing recovery systems
3. Waste Water Treatment
a. Concentration techniques
b. Other recovery techniques
c. Precipitation techniques
4. Hazardous Waste Treatment
a. Land Disposal
b. "Soft Hammer" Provisions
5. Environmental Comparison Test
Does the results of controls selected comply with the .
requirements of the CAA, FWPCA, and RCRA?
6. Cost Comparison Test
Does the pathway selected yield an affordable or optimal
cost when compared with the possible pathways?
This consideration will have an effect on the cost figures for sludge
disposal given in Table 2. These cost values are for conventional level
disposal as practiced prior to the "land ban provisions". As the provisions
for disposal under the land ban, including interim soft hammer requirements,
are added, the disposal costs are likely to escalate. This escalation will
make non- and low-sludge producing waste treatment alternatives even more
attractive. Higher cost processes which are capable of producing a waste
which can be delisted also may become more cost-competitive.
225
-------
TABLE 3.
Cost ($/lb of metal)
Precipitating
Aaent
Hydroxide
Sodium
Calcium
Magnesium
Sulfide
Soluble
Insoluble
Calcium
Carbonate
Sodium
Sodium bi-
Calcium
Sludge Chemical
Hb/lb metal) Usaae
150
300
150
300
900
300
300
200
200
0.32
0.06
1.20
0.36
0.34
0.27
0.30
0.40
0.30
- 0.37
- 0.07
- 1.30
- 0.48
- 0.45
- 0.34
- 0.35
- 0.48
- 0.35
Sludge*
Disoosal
2.00
4.00
1.10
4.00
12.00
4.00
2.60
2.60
2.60
Total
2.40
4.10
2.40
4.50
12.50
4.35
3.00
3.10
3.00
Other
Sodium biohydrate 90
Dithiocarbonate 490
Starch xanthate 1,000
Adsorption
Ferric chloride
Alum
Electrodeposition
Electrowinning 0
Special electrodes 0
16.00
25.00
77.00
0.46
5.80
0
0
1.20
6.10
17.20
31.40
0
0
2.00
6.00
based upon a dewatering cost of one cent per pound of raw sludge (1 to 2%
solids by weight) and a disposal cost of ten cents per pound of dewatered
siudge.
without credit for resale of recovered metals into the secondary metals
scrap market.
REFERENCES
Environmental Pollution Control Alternatives: Reducing Water Pollution
Control Costs in the Electroplating Industry; EPA 625/5-85-016;
September, 1985.
Control and Treatment Technology for the Metal Finishing Industry:
In-Plant Changes; EPA 625/0-82-008; January, 1982.
226
-------
Kushner, Joseph B.; Water and Waste Control for the Plating Shop.
Cincinnati, OH, Gardner Publications, 1976.
Roy, Clarence.; "Methods and Technologies for Reducing the Generation of
Electroplating Sludges." In U.S. Environmental Protection Agency and
American Electroplaters' Society, Inc. (cosponsors), Second Conference
on Advanced Pollution Control for the Metal Finishing Industry.
EPA 600/8-79-014. NTIS No. PB 297-453. Feb. 1979.
This paper was adapted by John Burckle from references 1 and 2 above.
227
-------
PCB DEGRADATION: STATUS AND DIRECTIONS
by: P. R. Sferra
Risk Reduction Engineering Laboratory
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
There is varied experimental activity to develop effective means of
destroying PCBs that are polluting the environment as hazardous wastes in
soils, sediments, water, and in storage. The recognized potential of
achieving environmentally harmless clean-up of PCBs by the use of
biological processes has resulted in increasing research activity. This
paper reviews the status of PCBs as pollutants and the current knowledge Of
microbial degradation of PCB compounds and the attempts to develop
technology for their control. Current information on degradation by mixed
cultures, pure cultures, aerobic and anaerobic processes and the degradative
relationships between various microorganisms and substrate structure are
discussed. Emphasis is on the importance of knowledge of breakdown pathways
resulting from metabolic action of microorganisms on PCBs. This information is
vital to the successful development through genetic engineering techniques of
strains with improved capability for degradation. Pertinent metabolic activity
and related genetic processes are presented as a basis for the rationale for
ongoing research utilizing applied genetics to develop microbial strains that
will be super degraders of the PCBs polluting the environment.
228
-------
INTRODUCTION
Polychlorinated biphenyls (PCBs) have been of worldwide concern as
polluting chemicals for over 20 years since Jensen (39) in 1966 first
reported their presence in wildlife samples. Because of their abundance,
persistence, and apparent hazard to wildlife and human health these
chlorinated hydrocarbons have been subjected to considerable study and as a
consequence by this time more than 1100 research articles on various
aspects of PCBs have been published.
Commercial production of PCBs, which began in 1929, was accomplished
by the synthetic process of direct chlorination of biphenyl with anhydrous
chlorine and a catalyst of iron filings or ferric chloride. Control of the
reaction conditions resulted in varied degrees of chlorination. With
chlorines replacing hydrogens, a large number of possible congeners can be
produced. Thus, a polychlorinated biphenyl is one of 209 compounds having
the formula: C12H10_nCln including 3 possible monochlorobiphenyls. The
209 congeners can be subdivided into 10 groups of homologs according to the
degree of chlorination and the number of isomers per homolog can vary from
1 to 46. PCBs were manufactured in the United States and Great Britain
under the tradename Aroclor, in Japan under the names Kanechlor and
Santotherm, in Germany Clophen, in Italy Fenclor, and in France Phenoclor
and Pyralene. The commercial products are complicated mixtures of chloro-
biphenyls. The different Aroclors are given a 4-digit designation that
represents the type of molecule and the weight percent of chlorine.
Aroclor 1242 consists of 12 = chlorinated biphenyl and 42 = 42% chlorine
per weight.
Two characteristics of the commercial products complicate attempts to
understand their behavior in the environment. One is that any one of the
commercial products may consist of a great number of different PCB
congeners, therefore the task of destroying PCBs is not as simple as
destroying one particular type of molecule. In addition, polychloro-
dibenzofurans (PCDFs) have been found in microgram per gram levels in
commercial PCB mixtures; these impurities therefore may be responsible
for at least some of the toxicological effects attributed to the
commercial PCB mixtures.
The utility of PCBs as industrial materials is based mainly upon their
chemical and physical stability and their electrical insulating properties.
These compounds have been used as dielectric fluids in capacitors and
transformers, heat transfer fluids, hydraulic fluids, lubricating and
cutting oils, dedusting agents, additives to pesticides, printing inks,
paints, pesticides, copying paper, carbonless copy paper, adhesives,
sealants, and plastics. Because of their properties the PCBs do not
readily degrade in the environment (22).
From 1929 through 1976 the world production of PCBs was about 1.3
billion pounds. Monsanto, in the U. S., produced about 1.25 billion
pounds before they stopped production in 1977. Production elsewhere
continued through at least 1983 with total world production through 1980
estimated to be about 2.4 billion pounds (22). It is also estimated that
229
-------
in the U. S. about 404 million pounds of PCBs are still in use and
accessible, about 262 million pounds are still in use and generally not
accessible, and about 585 million pounds are in storage, landfills, the
environment, or have been destroyed. The PCBs still in use and accessible
are in transformers,
capacitors, and other electrical equipment. Within this last category are
PCBs in sediments, soil, vegetation and animals, the atmosphere, and fresh
water totaling about 24 million pounds, another 13 million pounds are in
the oceans, about 385 million pounds are in landfills and other storage,
and an estimated 162 million pounds has been degraded (46). Several major
contaminations involving PCBs resulting from localized discharges have become
notorious. These occurred in the Hudson River, Waukegan, Illinois, New
Bedford, Massachusetts, and Bloomington, Indiana. Also, Yusho, an incident of
mass food poisoning western Japan in 1968 affected more than 1600 people who
had consumed rice-bran oil contaminated with PCBs.
PCBs have been found in marine mammals, marine and freshwater fish,
shellfish, birds, bird eggs, adipose tissue in humans, human milk,
terrestrial animals, wastewaters, drinking water, household products, and
in foods. PCBs are both hydrophobic and lipophilic and because they are
soluble in lipids they tend to accumulate in organisms, especially those
high in the food chain. Studies of the toxic effects of PCBs on organisms
have found inhibition of cell division, reduction in RNA levels, reduction
in chlorophyll index, reduced growth rate, reduced cell population size,
and inhibition of carbon fixation (27).
The amounts of PCBs in the environment are considered to be
significant, a situation perceived to be dangerous both to wildlife,and
human health. As a consequence there has been considerable effort to
determine how the polluting PCBs can be destroyed and those still in
service prevented from becoming new hazardous wastes. The Risk Reduction
Engineering Laboratory (RREL) of the U.S. Environmental Protection Agency has
expended considerable effort in various aspects of the PCS pollution problem.
Physical, chemical, and biological means of destroying PCBs are being
investigated under RREL sponsorship. Basic knowledge of the anabolic and
catabolic capability of the living cell by means of its enzyme systems leads
to the expectation that cost-effective highly efficient biological control of
hazardous pollutants such as PCBs with minimum adverse impact upon the
environment can be developed. Mondello and Yates (51) will report in these
proceedings on recent progress made in PCB biodegradation research under RREL
sponsorship.
DISPOSAL AND DESTRUCTION OF WASTE PCBs
In 1982, Fradkin and Barisas (26) assessed the available and emerging
technologies for the disposal of PCBs and PCB-contaminated materials and in
1988, Carpenter arid Wilson (15) published a technical and economic
evaluation of processes for the removal of PCBs from sediments. Exner
(23) and Weitzman (78) also prepared information on the destruction or
decontamination of PCBs. Non-biological treatment technologies for PCB
wastes include wet air oxidation, supercritical water oxidation, catalytic
230
-------
dehydrochlorination, sodium-based nucleophilic substitution, alkali metal
potassium glycolate-based nucleophilic substitution, UV/photolysis,
microwave plasma, pyrolysis, removing and concentrating by heated air
stripping, extraction, adsorption, vitrification, stabilizing, and
bottom recovery by dredging (15).
In 1973, publications by Gibson et al.(36), Wallnoffer et al. (76),
Catelani et al. (17), Ohmori et al. (55), and Ahmed and Focht (3, 4) reported
on the degradation of biphenyl and PCBs by single strains of bacteria. Since
that time it has become evident that PCBs can be degraded by pure cultures and
by naturally occurring populations of microorganisms. Data have also been
collected to determine the conditions that affect the degradation of PCBs and
to determine the metabolic pathways of PCB degradation by microorganisms.
Factors such as temperature, nutrients, oxygen, and pH all influence the
ability of the organism not only to function normally but function as a
degrader of xenobiotic compounds such as PCBs. Other factors, related to the
PCBs, such as volatility, water solubility, emulsification, adsorption, and
degree of chlorination of the PCBs all influence the degradative activity of
the microorganism (28).
Furukawa et al. (35) determined relationships between chemical structure
and breakdown of biphenyl and PCBs. General conclusions from their results
were as follows: (1) As chlorine substitution increases, the degradation rate
markedly decreases, (2) PCBs containing 2 chlorines in the ortho position of a
single ring (2,6-) and each ring (2,2'-) show a high resistance to degradation,
(3) PCBs with all chlorines on a single ring are generally degraded faster than
those with the same number of chlorines on both rings, (4) PCBs with 2 chlorines
at the 2,3- position of one ring are more susceptible to microbial degradation
at least compared with other tetra- and pentachlorobiphenyls, and (5) ring
cleavage occurs with a nonchlorinated or lesser chlorinated ring of the biphenyl
molecule.
Tucker et al. (74) determined that the rate of PCB biodegradation by
activated sludge decreased with increasing chlorine content of commercial
mixtures. Their studies compared the biodegradation rates of biphenyl and
the Aroclors 1221, 1016, 1242, and 1254 showing that the average percent
degradation of the commercial mixtures is directly related to their weight
percent chlorine content (Aroclor 1016 contains 41% chlorine). Further,
metabolism of PCBs by bird and mammal species decreases in rate as the number
of chlorines in the biphenyl molecule increases.
In 1973, Gibson et al. (36) isolated a species of Beijerinckia that
utilizes biphenyl as a sole source of carbon for growth. This organism
metabolized biphenyl to cis-2,3-dihydroxy-l-phenylhexa-4,5-diene (cis-2,3-
dihydro-2,3-dihydroxybiphenyl). This reaction is catalyzed by 2,3-
dioxygenase. A further intermediate, catalyzed by dihydrodiol
dehydrogenase, was identified as 2,3-dihydroxybiphehyl. Continued
research (16, 17, 18, 47) verified these metabolic steps and determined
that meta cleavage, catalyzed by 2,3-dihydroxybiphenyl dioxygenase, between
Cl and C2 of the 2,3-dihydroxy compounds yields 2-hydroxy-6-oxo-6-
231
-------
phenylhexa-2,4-dienoate which can degrade to benzole acid. Bacterial
strains are known to convert PCBs into the corresponding chlorobenzoates.
Mondello and Yates (51) expand on this subject of metabolic pathway to degrade
PCBs.
Mlcrobial degradation of PCBs has been and is being studied by a number
of biochemists, microbiologists, analytical chemists, and molecular
geneticists. The widespread presence of PCBs in the environment is of concern
to governments and the public worldwide thus there has been an abundance of
research on various aspects of PCBs, e.g., ecology, toxicology, biochemistry,
and applied genetics. There has been increasing effort to provide a
technology utilizing some process of biodegradation to destroy these polluting
compounds leaving no harmful effect upon the environment. The living cell is
capable of highly efficient enzyme-catalyzed reactions and man is on the verge
of mastering the use of cells, most likely specially developed bacteria, to
clean up hazardous wastes such as the PCBs. Table 1 lists a few references to
publications on a variety of aspects of the PCB problem and Table 2. consists
of a list with references of most of the species and strains of microorganisms
that have been used in PCB biodegradation research. These publications can
provide excellent background material on the subject.
Table 1. PCB Research References
SUBJECT: REFERENCE
SUBJECT: REFERENCE
Activated sludge: 40, 74, 75
Anaerobic strains: 59
Anaerobe strain B-206: 71, 72
Analog enrichment: 24
Analysis, all 209 congeners: 52
Analytical chemistry: 22
Animal toxicology: 42
Attenuation
by earth materials: 37
Biodegradation
of congeners: 7, 9
Biodegradation
of hydroxybiphenyls: 43
Biphenyl metabolism: 47
Carcinogenic
and other chronic effects: 41
Chemistry: 58
Early report: 57
Effect of glucose uptake: 64
Effect on nitrification: 65
Environmental
dechlorination: 10, 12
Freshwater
microbial populations: 64, 67
General: 21, 27, 61, 63
Human exposure: 20
Marine bacteria: 14
Metabolism
and biochemical toxicity: 49
Molecular toxicology: 62
PCB intake in humans: 45, 78, 81
Persistence: 5
Photochemistry: 60
Progenitive manifestation: 44
Rapid assay: 8
Screening: 70
Structure: 69
Treatment technologies: 26
232
-------
Table 2. Microbial taxa tested for PCS degradative ability
TAXON
Achromobac ter spp.
Achromobacter sp. strain B-218
Achromobacter sp. strain BP
Achromobacter sp. strain pCB
Acidovirans group
Acinetobacter sp. strain P6
32, 34, 43
Acinetobacter sp. strain 4-CB1
Alkaligenes sp.
Alcaligenes sp.
Alcaligenes eutrophus strain H850
Alcaligenes faecalis strain Pi434
Alcaligenes odorans
Alcaligenes denitrif icans
Arthrobacter sp. strain BIB
Arthrobacter sp strain M5
Aspergillus flavus
Azotobacter sp. strain 4CB
Bacillus brevis sp. strain B-257
Beijerinckia sp. strain B8/36
Corynebacterium sp. strain MB1
Escherichia coli strain TB1
Escherichia coli strain FM4560
Nitrobacter agilis
Nitrosomonas europaea
Pseudomonas spp.
Pseudomonas sp. strain 1008
Pseudomonas sp. strain H1130
Pseudomonas sp. strain HBP1
Pseudomonas sp. strain JB1
Pseudomonas sp. strain LB400
Pseudomonas sp. strain MB86
Pseudomonas sp. strain Pi304
Pseudomonas sp. strain Pi939
Pseudomonas aeruginosa PA01161
Pseudomonas cepacia strain H201
Pseudomonas cepacia strain Pi704
Pseudomonas cepacia strain RJB
Pseudomonas paucimobilis Ql
Pseudomonas pseudoalcaligenes KF707
Pseudomonas putida
Pseudomonas putida
Pseudomonas putida strain LB400
Pseudomonas putida strain LB410
Pseudomonas testosteroni strain H128
Pseudomonas testosteroni strain H336
Pseudomonas testosteroni strain H430
Rhizoous -iaoonicus
80
48
3
3
76
1
1
31
32
7
76
19
19
43
30
53
38
48
36
11
51
50
65
65
80
66
76
43
56
50
6
76
76
29
76
76
76
73
29
17
16
54
76
76
76
76
77
REFERENCE
, 4
, 4
, 2, 13, 25, 30,
, 2
,34
, 9, 51, 54, 76, 82
, 37
, 37
, 7, 9, 76
, 51, 82
, 33
, 33, 73
, 18
, 68, 76
233
-------
1.
2.
REFERENCES
Adriaens, P., and D.D. Focht. Biodegradation of 4,4'-Dichlorobiphenyl
(4,4'-DCBP) by a Coculture of Two Acinetobacter Species. Ann. Meeting
of the American Society for Microbiology, Miami Beach, FL. p.222, 1988.
Adriaens, P., and D.D. Focht. 4-Chlorobenzoate Metabolism by
Acinetobacter sp. 4-CB 1 and its Relevance to Biodegradation of 4,4'-
Dichlorobiphenyl. In Environmental Biotechnology. Reducing Risks from
Environmental Chemicals through Biotechnology. G.S.Omenn, Ed.,
Life Sciences, Plenum Press, New York. Vol. 45, p.442, 1988.
Basic
3. Ahmed, M. and D.D. Focht. Oxidation of Polychlorinated Biphenyls by
Achromobacter pCB. Bull. Environ. Contam. & Tox., 10(2):70-72, 1973.
4. Ahmed, M., and D.D. Focht. Degradation of polychlorinated biphenyls
by two species of Achromobacter. Can. J^ Microbiol. 19:47-52, 1973.
5. Ballschmiter, K., M. Zell, and H.J. Neu. Persistence of PCB's in the
Ecosphere: Will Some PCB-Components "Never" Degrade? Chemosphere,
2:173-176, 1978.
6. Barton, M.R., and R.L. Crawford. Novel Biotransformation of
4-Chloro-biphenyl by a Pseudomonas sp. Appl. and Environ. Microb.,
54(2):594-595, 1988. "
7. Bedard, D.L., M.L. Haberl, R.J. May, and M.J. Brennan. Evidence for Novel
Mechanisms of Polychlorinated Biphenyl Metabolism in Alcaligenes eutrophus
H850. Appl. and Environ. Microb., 53(5):1103-1112, 1987T
8. Bedard, D.L., R. Unterman, L.H. Bopp, M.J. Brennan, M.L. Haberl, and C.
Johnson. Rapid Assay for Screening and Characterizing Microorganisms for
the Ability to Degrade Polychlorinated Biphenyls. Appl. and Environ.
Microb.. 51(4);761-768. 1986. ~~
9. Bedard, D.L., R.E. Wagner, M.J. Brennan, M.L. Haberl, and J.F. Brown, Jr.
Extensive Degradation of Aroclors and Environmentally Transformed Poly-
chlorinated Biphenyls by Alcaligenes eutrophus H850. Appl. and Environ.
Microb.. 53(5):.1094-1102, 1987.
10. Brown, J.F., Jr., D.L. Bedard, M.J. Brennan, J.C. Carnahan, H. Feng, and
R.E. Wagner. Polychlorinated Biphenyl Dechlorination in Aquatic
Sediments. Science, 236:709-712, 1987.
11. Brown, J.F., Jr., R.E. Wagner, D.L. Bedard, M.J. Brennan, J.C. Carnahan,
R.J. May, and T.J. Tofflemire. PCB Transformations in Upper Hudson
Sediments. Northeast. Environ. Science, 3(3/4):167-179, 1984.
12. Brown, J.F., R.E. Wagner, H. Feng, D.L. Bedard, M.J. Brennan, J.C.
Carnahan, and R. J. May. Environmental Dechlorination of PCBs. Environ.
Toxicol. and Chem., 6:579-593, 1987.
234
-------
13. Brunner, W., F.H. Sutherland, and D.D. Focht. Enhanced Biodegradation of
Polychlorinated Biphenyls in Soil by Analog Enrichment and Bacterial
Inoculation. J. Environ. Qual., 14(3);324-328, 1985.
14. Carey, A.E., and G.R. Harvey. Metabolism of Polychlorinated Biphenyls by
Marine Bacteria. Bull. Environ. Contain. Toxicol., 20:527-534, 1978.
15. Carpenter, B.H., and D.L. Wilson. Technical/Economic Assessment of
Selected PCB Decontamination Processes. J._ Haz. Mat., 17:125-148, 1988.
16. Catelani, D., and A. Colombi. Metabolism of Biphenyl. Structure and
Physicochemical Properties of 2-Hydroxy-6-oxo-6-phenylhexa-2,4-dienoic
Acid, the Meta-cleavage Product from 2,3-Dihydroxybiphenyl by
Pseudomonas putida, Biochem. J., 143:431-434, 1974.
17. Catelani, D., A. Colombi, C. Sorlini, and V. Treccani. Metabolism of
Biphenyl. 2-Hydroxy-6-oxo-6-phenylhexa-2,4-dienoate: The Meta-Cleavage
Product from 2,3-dihydroxybiphenyl by Pseudomonas putida. Biochem. J.,
134:1063-1066, 1973. , .
18. Catelani, D., C. Sorlini, and V. Treccani. The Metabolism of Biphenyl
by Pseudomonas putida. Experientia, 27(10):1173-1174, 1971.
19. Clark, R.R., E.S.K. Chian, and R.A. Griffin. Degradation of Poly-
chlorinated Biphenyls by Mixed Microbial Cultures. Appl. and Environ.
Microb., 37(4):680-685, 1979.
20. Cordle, F., P. Corneliussen, C. Jelinek, B. Hackley, R. Lehman, J.
Mclaughlin, R. Rhoden, and R. Shapiro. Human Exposure to
Polychlorinated Biphenyls and Polybrominated Biphenyls. Envirbn.
Health Persp., 24:157-172, 1978. ...,.•
21. Dagley, S. Microbial Degradation of Organic Compounds in the Biosphere.
American Scientist, 63:682-689, 1975.
22. Erickson, M.D. Analytical Chemistry of PCBs. Butterworth Publishers,
Boston, MA, 508 pp.,1986.
23. Exner, J.H., Summary of Polychlorinated Biphenyl Treatment Alternatives.
In Detoxication of_ Hazardous Waste, J.H. Exner, Ed., Ann Arbor Science
Publishers, Inc., Ann Arbor, MI, pp. 119-120, 1982.
24. Focht, D.D. Analog Enrichment Decontamination Process. U.S. Patent.
Patent Number: 4,664,805. U.S. Patent and Trademark Office. 1987.
25. Focht, D.D., and W. Brunner. Kinetics of Biphenyl and Polychlorinated
Biphenyl Metabolism in Soil. Appl. and Environ. Microb., 50(4): 1058-
1063, 1985.
i
26. Fradkin, L., and S. Barisas. Technologies for Treatment, Reuse, and
Disposal of_ Polychlorinated Biphenyl Wastes. Argonne National Laboratory
ANL/EES-TM-168, 47 pp., 1982.
235
-------
27. Furukawa, K. Microbial Degradation of Polychlorinated Biphenyls
(PCBs). In Biodegradation and Detoxification of Environmental Pollutants,
A.M. Chakrabarty, Ed., CRC Press, Inc., Boca Rafon, FL, pp. 33-57, 1982.
28. Furukawa, K. Modification of PCBs by Bacteria and Other Microorganisms
In PCBs and the Environment. J.S. Waid, Ed., CRC Press, Inc., Boca
Raton, FL, Vol. II, pp. 89-100, 1986.
29. Furukawa, K., and N. Arimura. Purification and Properties of 2,3-
Dihydroxybiphenyl Dioxygenase from Polychlorinated Biphenyl-Degrading
Pseudomonas pseudoalcaligenes and Pseudomonas aeruginosa Carrying the
Cloned bphC Gene. J. Bact., 169(2):924-927, 1987.
30. Furukawa, K., and A.M. Chakrabarty. Involvement of Plasmids in Total
Degradation of Chlorinated Biphenyls. Appl. and Environ. Microb.,
44(3):619-626, 1982.
31. Furukawa, K., and F. Matsumura. Microbial Metabolism of
Polychlorinated Biphenyls. Studies on the Relative Degradability of
Polychlorinated Biphenyl Components by Alkaligenes sp. J_^ Agric. Food
Chem.. 24(2):251-256, 1976.
32. Furukawa, K., F. Matsumura, and K. Tonomura. Alcaligenes and
Acinetobacter Strains Capable of Degrading Polychlorinated Biphenyls.
Agric. Biol. Chem., 42(3):543-548, 1978.
33. Furukawa, K, and T. Miyazaki. Cloning of a Gene Cluster Encoding
Biphenyl and Chlorobiphenyl Degradation in Pseudomonas
pseudoalcaligenes. J. Bact., 166(2):392-398, 1986.
34. Furukawa, K., N. Tomizuka, and A. Kamibayashi. Effect of Chlorine
Substitution on the Bacterial Metabolism of Various Polychlorinated
Biphenyls. Appl. and Environ. Microb., 38(2):301-310, 1979.
35. Furukawa, K., K. Tonomura, and A. Kamibayashi. Effect of Chlorine
Substitution on the Biodegradability of Polychlorinated Biphenyls.
Appl. and Environ. Microb., 35:223, 1978.
36. Gibson, D.T., R.L. Roberts, M.C. Wells, and V.M. Kobal. Oxidation of
. Biphenyl by a Beijerinckia Species. Biochem. and Biophys. Res. Comm.,
50(2):211-219, 1973.
37. Griffin, R.A., and E.S.K. Chian. Attenuation of Water-Soluble
Polychlorinated Biphenyls by Earth Materials. U. S. Environmental
Protection Agency Research Report EPA-600/2-80-027, 93 pp., 1980.
38. Hernandez, B.S., and D.D. Focht. Metabolism of Biphenyls, Chlorinated
Biphenyls, and Ethylene by Azotobacter 4CB in N-free Media. Ann. Meet. Am.
Society for Microbiology, Miami Beach, FL, p. 222, 1988.
39. Jensen, S. Report of a new chemical hazard. New Sci. , 32:612, 1966.
40. Kanekb, M., K. Morimoto, and S. Nambu. The Respone of Activated Sludge
to a Polychlorinated Biphenyl (KC-500). Water Res., 10:157-163, 1976.
236
-------
41. Kimbrough, R.D.. The Carcinogenic and Other Chronic Effects of
Persistent Halogenated Organic Compounds. In Health Effects of
Halogenated Aromatic Hydrocarbons. Ann. N. Y. Academy of Sciences, The
New York Academy of Sciences. New York. 320:415-418, 1979.
42. Kimbrough, R., J. Buckley, L. Fishbein, G. Flamm, L. Kasza, W. Marcus,
S. Shibko, and R. Teske. Animal Toxicology. Environmental Health
Perspectives, 24:173-185, 1978.
43. Kohler, H.-P.E., D. Kohler-Staub, and D.D. Focht. Degradation of 2-
Hydroxybiphenyl and 2,2'-Dihydroxybiphenyl by Pseudomonas sp. Strain
HBP1. Appl. and Environ. Microb., 54(11):2683-2688, 1988.
44. Krockel, L., and D.D. Focht. Construction of Chlorobenzene-Utilizing
Recombinants by Progenitive Manifestation of a Rare Event. Appl. and
Environ. Microb., 53(10):2470-2475, 1987.
45. Kuwabara, K., T. Yakushiji, I. Watanabe, S. Yoshida, K. Yoyama, and N.
Kunita. Increase in the Human Blood PCB Levels Promptly Following
Ingestion of Fish Containing PCBs. Bull. Environ. Contam. Toxicol., ,
21:273-278, 1979.
46. Lauber, J.D.. Disposal and Destruction of Waste PCBs. In PCBs and
the Environment, J.S. Waid, Ed., CRC Press, Inc., Boca Raton, FL, Vol.
Ill, pp. 83-151, 1987.
47. Lunt, D., and W.C. Evans. The Microbial Metabolism of Biphenyl.
Biochemical Journal, 118:54p-55p, 1970.
The
48. Masse, R., F. Messier, L. Peloquin, C. Ayotte, and M. Sylvestre.
Microbial Biodegradation of 4-Chlorobiphenyl. A Model Compound of
Chlorinated Biphenyls. Appl. Environ. Microbiol., 47(5):947-951, 1984.
49. Matthews, H., G. Fries, A. Gardner, L. Garthoff, J. Goldstein, Y. Ku,
and J. Moore. Metabolism and Biochemical Toxicity of PCBs and PBBs.
Environ. Health Persp., 24:147-155, 1978.
50. Mondello, F.J. Cloning and Expression in Escherichia coli of Genes
Encoding Polychlorinated Biphenyl Degradation from Pseudomonas Strain
LB400. J_._ Bact., 171(3) :1725-1732, 1989.
51. Mondello, F.J., and J.R. Yates. The Development of Recombinant Bacteria
for Polychlorinated Biphenyl Degradation. Proceedings Fifteenth Annual
Hazardous Waste Symposium, Remedial Action, Treatment and Disposal of
Hazardous Waste, In press.
52. Mullin, M.D., C.M. Pochini, S. McCrindle, M. Romkes, S.H. Safe, and
L.M. Safe. High-Resolution PCB Analysis: Synthesis and Chromatographic
Properties of All 209 PCB Congeners. Environ. Sci. Technol., 18(6):468-
476, 1984.
237
-------
53. Murado, M.A., MaC. Tejedor, and G. Baluja. Interactions Between Poly-
chlorinated Biphenyls (PCBs) and Soil Microfungi. Effects of Aroclor-1254
and Other PCBs on Aspergillus flavus Cultures. Bull. Environ. Contam. &_
Tox., 15(6):768-774, 1976.
54. Nadim, L.M., M.J. Schocken, F.K. Higson, D.T. Gibson, D.L. Bedard, L.H.
Bopp, and F. J. Mondello. Bacterial Oxidation of Polychlorinated
Biphenyls. Proceedings of the Thirteenth Annual Research Symposium, Land
Disposal, Remedial Action, Incineration and Treatment of Hazardous Waste,
EPA/600/9-87/015, pp. 395-402, 1987.
55. Ohmori, T., T. Ikai, Y. Minoda, and K. Yamada. Utilization of polyphenol
and polyphenol-related compounds by microorganisms. Agric. Biol. Chem.,
37:1599, 1973
56. Parsons, J.R., and D.T.H.M. Sijm. Biodegradation Kinetics of Polychlor-
inated Biphenyls in Continuous Cultures of a Pseudomonas Strain.
Chemosphere, 17(9):1755-1766, 1988.
57. Peakall, D.B., and J.L. Lincer. Polychlorinated Biphenyls. Another Long-
Life Widespread Chemical in the Environment. BioScience, 20(17):958-964,
1970.
58. Pomerantz, I., J. Burke, D. Firestone, J. McKinney, J. Roach, and W.
Trotter. Chemistry of PCBs and PBBs. Environ. Health Persp., 24:133-146,
1978.
59. Quensen III, J.F., J.M. Tiedje, and S.A. Boyd. Reductive Dechlorination
of Polychlorinated Biphenyls by Anaerobic Microorganisms from Sediments.
Science, 242:752-754, 1988.
60. Ruzo, L.O., M.J. Zabik, and R.D. Schuetz. Photochemistry of Bioactive
Compounds: Photoproducts and Kinetics of Polychlorinated Biphenyls. J_^
Agr. Food Chem. , 22(2):199-202, 1974.
61. Safe, S.H. Microbial Degradation of Polychlorinated Biphenyls. In Micro-
bial Degradation of Organic Compounds, D.T.Gibson, Ed., Marcel Dekker,
Inc., New York, NY, pp.361-369, 1984.
62. Safe, S., L.W. Robertson, L. Safe, A. Parkinson, S. Bandiera, T. Sawyer, and
M.A. Campbell. Halogenated Biphenyls: Molecular Toxicology. Can. J._
Physiol. Pharmacol., 60:1057-1064, 1982.
63. Sahasrabudhe, S.R., and V.V. Modi. Microbial Degradation of Chlorinated
Aromatic Compounds. Microbiological Sciences, 4(10):300-303, 1987.
64. Sayler, G.S.,, L.C.. Lund, M.P. Shiaris, T.W. Sherrill, and R.E. Perkins.
Comparative Effects of Aroclor 1254 (Polychlorinated Biphenyls) and
Phenanthrene on Glucose Uptake by Freshwater Microbial Populations.
Appl. and Environ. Microb., 37(5):878-885, 1979.
65. Sayler, G.S., M.P. Shiaris, W. Beck, and S. Held. Effects of Polychlor-
inated Biphenyls and Environmental Biotransformation Products on Aquatic
Nitrification. Appl. and Environ. Microb., 43(4):949-952, 1982.
238
-------
66. Sayler, G.S., M. Shon, and R.R. Colwell. Growth of an Estuarine
Pseudomonas sp. on Polychlorinated Biphenyl. Microb. Ecol., 3:241-255,
1977. ;
67. Shiaris, M.P. , and G.S. Sayler. Biotransformation of PCB by Natural
Assemblages of Freshwater Microorganisms. Environ^ Sci. Technol.,
16(6):367-369, 1982. ;
68. Singh, S., F.K. Higson, L.M. Nadim, and D.T. Gibson. Oxidation of Poly-
chlorinated Biphenyls by Pseudomonas putida LB400. Proceedings of the
Fourteenth Annual Research Symposium, Land Disposal, Remedial Action,
Incineration and Treatment of Hazardous Waste, EPA/600/9-88/021, pp. 346-
354, 1988.
69. Sissons, D., and D. Welti. Structural Identification of Polychlorinated
Biphenyls in Commercial Mixtures by Gas-Liquid Chromatography Nuclear
Magnetic Resonance and Mass Spectrometry. J._ Chromatogr., 60:15-32, 1971.
70. Sylvestre, M.. Screening Method for the Isolation of Polychlorinated
Biphenyls Degrading Bacteria. 80th Annual Meeting American Society for
Microbiology, Miami Beach, Fla., May 11-16, 1980. Abstracts of the Annual
Meeting of the American Society for Microbiology, Abstract 236, Vol. 80,
1980.
71. Sylvestre, M., and J. Fauteux. A New Facultative Anaerobe Capable of
Growth on Chlorobiphenyls. J. Gen. Appl. Microbiol., 28(l):61-72, 1982.
72. Sylvestre, M., R. Masse, F. Messier, J. Fauteux, J.-G. Bisaillon, and R.
Beaudet. Bacterial Nitration of 4-Chlorobiphenyl. Appl. and Environ.
Micro., 44(4):871-877, 1982. ::[~
73. Taira, K., N. Hayase, N. Arimura, S. Yamashita, T. Miyazaki, K. Furukawa.
Cloning and Nucleotide Sequence of the 2,3-Dihydroxybiphenyl Dioxygenase
Gene from the PCB-Degrading Strain of Pseudomonas paucimobilis Ql.
Biochemistry, 27:3990-3996, 1988.
74. Tulp, M.T.M., R. Schmitz, and 0. Hutzinger. The Bacterial Metabolism of
4,4'-Dichlorobiphenyl and Its Suppression by Alternative Carbon Sources.
Chemosphere, 1:103-108, 1978.
75. Unterman, R., D.L. Bedard, M.J. Brennan, L.H. Bopp, F.J. Mondello, R.E.
Brooks, D.P. Mobley, J.B. McDermott, C.C. Schwartz, and O.K. Dietrich.
Biological Approaches for Polychlorinated Biphenyl Degradation. In
Environmental Biotechnology. Reducing Risks from Environmental
Chemicals through Biotechnology. G.S. Omenn, Ed., Plenum Press, New York,
pp. 253-269, 1988.
76. Wallnofer, P.R., G. Engelhardt, S. Safe, and 0. Hutzinger. Microbial
Hydroxylation of 4-Chlorobiphenyl and 4,4'-Dichlorobiphenyl. Chemosphere,
2:69-72, 1973. '
239
-------
77. Watanabe, I., T. Yakushiji, K. Kuwabara, S. Yoshida, K. Maeda, T.
Kashimoto, K. Koyama, and N. Kunita. Surveillance of the Daily PCB Intake
from Diet of Japanese Women from 1972 Through 1976. Arch. Environ.
Contam. Toxicol., 8:67-75, 1979.
'k
78. Weitzman, L. Treatment and Destruction of Polychlorinated Biphenyls and
Polychlorinated Biphenyl-Contaminated Materials. In Detoxication of_
Hazardous Waste, J.H. Exner, Ed., Ann Arbor Science Publishers, Inc., pp.
131-142, 1982.
79. Wong, P.T.S., and K.L.E. Kaiser. Bacterial Degradation of Polychlorinated
Biphenyls II. Rate Studies. Bull. Environ. Contam. & Toxic., 13(2):249-
256, 1975.
80. Yakushiji, T., I. Watanabe, K. Kuwabara, S. Yoshida, K. Koyama, and N.
Kunita. Levels of Polychlorinated Biphenyls (PCBs) and Organochlorine
Pesticides in Human Milk and Blood Collected in Osaka Prefecture from 1972
to 1977. Int. Arch. Occup. Environ. Health, 43:1-15, 1979.
81. Yates, J.R., and F.J. Mondello. Sequence Similarities in the Genes
Encoding Polychlorinated Biphenyl Degradation from Pseudomonas Strain LB400
and Alcaligenes eutrophus H850. J^ Bact., 171(3):1733-1735, 1989.
240
-------
The Development of Recombinant Bacteria for
Polychlorinated Biphenyl Degradation
Frank J. Mondello and James R. Yates
GE Research and Development Center
S chenec tady, N.Y.
ABSTRACT
Pseudomonas strain LB400 and Alcaligenes eutrophus strain H850 have
previously been demonstrated to degrade an unusually wide variety of
polychlorinated biphenyls (PCBs). The genes encoding the PCS degradative
enzymes (the bph genes) from these organisms were isolated using the
broad-host-range cosmid vector pMMB34, and found to be expressed in
Escherichia coli. A comparison of the PCB degradative capabilities of the
wild-type and recombinant strains was conducted using resting-cell assays.
Significant improvements in the activity of the recombinant strains were
observed after plasmid modifications. The degradation of a variety of PCB
mixtures including Aroclor 1242 was found to be comparable for LB400 and the
recombinant strain FM4560.
DNA:DNA hybridization analysis was used to determine that the genes
encoding PCB degradation in LB400 and H850 are genetically distinct from
those in a variety of other organisms. This indicates the existence of at
least two different classes of genes for PCB metabolism. The availability of
DNA probes for the LB400/H850 class of bph genes will make it possible to
determine the fate of recombinant strains or plasmids containing these genes
in bioremediation processes.
INTRODUCTION
Polychlorinated biphenyls (PCBs) consist of a biphenyl molecule
containing from 1 to 10 chlorines, making it possible to produce 209
different PCB congeners (1). Commercially, PCBs were used and discarded as
complex mixtures (known as Aroclors or Kanechlors), which contain 60-80
different congeners (1). Effective processes for the biodegradation of
environmental PCBs will therefore require organisms capable of attacking a
wide variety of these congeners. Few PCB degrading strains isolated thus far
are capable of degrading a broad range of highly chlorinated PCBs. Two
strains with outstanding PCB degrading ability are Pseudomonas strain LB400
and Alcaligenes eutrophus H850, which can degrade PCB molecules containing up
to six chlorines (2,3).
The PCB degradative abilities of LB400 and H850 are similar. In both
organisms this process involves two different types of oxidative attack (4).
At least one type of attack is performed by enzymes of the bph pathway, which
are known to be involved in the metabolism of biphenyl (Figure 1). The first
enzyme in this pathway is a 2,3-dioxygenase which inserts oxygen at carbon
241
-------
positions 2 and 3 of biphenyl. This type of activity has been found in all
known PCS degrading bacteria and results in the formation of 2,3-dihydrodiol
(5). Dihydrodiol dehydrogenase then converts this compound to 2,3-dihydroxy-
biphenyl which undergoes extradiol cleavage by 2,3-dihydroxybiphenyl
dioxygenase to form a yellow meta-cleavage product. This compound is
converted by a hydrase to benzoic acid (1,4,6).
bprtA
,OH
COOH
P bphD
COOH
Clx
Clx
Clx
Clx
Clx
Figure 1. Degradation of Biphenyl and PCBs by the 2,3- Dioxygenase Pathway.
Gene designations: bphA, biphenyl 2,3-dioxygenase; bphB, dihydrodiol
dehydrogenase; bphC, 2,3-dihydroxybiphenyl dioxygenase; bphD, meta-cleavage
product hydrase
LB400 and H850 also contain a much less common 3,4-dioxygenase activity,
which adds oxygen atoms at carbon positions 3 and 4 (4). This activity
may explain the unusually broad congener specificity of these organisms,
enabling them to degrade congeners recalcitrant to degradation by other
organisms. For example, many PCB degrading bacteria are unable to degrade
2,5,2',5'- tetrachlorobiphenyl, presumably because all of the 2,3-ring
positions are blocked by chlorines. This congener is readily attacked by
LB400 and H850, resulting in the formation of a 3,4-dihydrodiol (4). It is
not yet known if 3,4-dioxygenase activity is encoded by bphA of the
2,3-dioxygenase pathway.
The exceptional degradative capabilities of LB400 and H850 make the
isolation and analysis of the bph genes from these organisms an important
first step towards the development of new strains with enhanced abilities to
degrade PCBs. The purpose of this report is to summarize recent progress
made toward this goal. A more detailed description of some of this work will
appear in the March 1989 issue of the Journal of Bacteriology.
CLONING THE LB400 AND H850 GENES FOR PCB DEGRADATION
Genomic libraries of LB400 and H850 were constructed using the
wide-host-range, mobilizable cloning vector pMMB34 (7), and introduced into
an Escherichia coli host via standard procedures. Ampicillin resistant
colonies were tested for the presence of 2,3-dihydroxybiphenyl dioxygenase
activity. This enzyme is encoded by the bphC gene of the biphenyl/PCB
degradative pathway, and catalyzes the conversion of 2,3-dihydroxybiphenyl to
a yellow compound, 2-hydroxy-6-oxo-6-phenyl hexa-2,4-dienoate (4). Colonies
expressing bphC were identified by their accumulation of visible quantities
242
-------
of this product. Clones from both LB400 and H850 were found to express
2,3-dihydroxybiphenyl dioxygenase activity. The bphG-containing plasmids
from LB400 were designated pGEM410, 420, 430. Those from H850 were
designated pGEMSOO, 810, 830, and 850. These strains were examined for the
presence of additional enzymes of the PCB/biphenyl pathway by using a series
of rapid screening procedures (4,6). Each of the recombinant plasmids
contained several bph genes and many encoded all of the enzymes required to
convert PCBs to chlorobenzoic acids (Table 1). These data indicated that the
genes for PCS degradation in both LB400 and H850 were closely linked. A
similar result has been reported by Furukawa and Miyazaki for bphA, B, and C
of Pseudomonas pseudoalcaligenes KF707 (8). — -
TABLE 1. BPH-PATHWAY GENES ENCODED BY RECOMBINANT PLASMIDS
Genes Expressed
PLASMID
pGEM410
pGEM420
pGEM430
pGEMSOO
pGEMSlO
pGEM830
pGEM850
pMMB34
bphA bphB bphC bphD
+ + > • + +
Enzymatic activity present (+) or absent (-)
PCB METABOLISM BY RECOMBINANT STRAINS ;
E. coli strains containing pGEM plasmids were tested for their ability
to degrade polychlorinated biphenyls in resting cell assays. The strain with
the highest activity was FM4100 (containing plasmid pGEM410) and was chosen
for further study. Incubation of this strain with either 2,3-, 2,5-, or
2,2'-di- chlorobiphenyl resulted in the accumulation of,2,3-, 2,5- and
2-chlorobenzoic acid, respectively. This unequivocally demonstrated the
expression of at least the first four enzymes of the 2,3-dioxygenase pathway
of PCB metabolism.
The ability of LB400 and H850 to degrade 2,5,2',5'-tetrachlorobiphenyl
(2,5,2',5'-CB) sets them apart from other PCB degrading strains. This
congener has no free adjacent 2,3 ring positions, and is not readily oxidized
by the 2,3-dioxygenase enzymes of most PCB degrading bacteria (2). LB400 and
243
-------
H850 catabolize 2,5,2',5'-CB via a 3,4-dioxygenase to a compound identified
as 3,4-dihydroxy-3,4-dihydro-2,5,2',5'-tetrachloro- biphenyl (3,4-dihydro-
diol) (4). Degradation of 2,5,2',5'-CB by FM4100 resulted in the production
of a metabolite with a GG profile identical to that of purified
3,4-dihydrodiol (provided by D.T. Gibson) demonstrating the existence of
3,4-dioxygenase activity in the recombinant strain.
MAPPING STUDIES
Digestion of pGEM410 with the endonuclease EcoRI results in the
formation of nine DNA fragments whose sizes and relative order are 15.5, 6.1,
2.9, 0.8, 2.0, 6.7, 2.9, 2.3, and 0.5 kilobase pairs (kb). The relative
positions of these fragments were determined by analyzing subclones produced
from an incomplete EcoRI digest of pGEM410 DNA. The partially-digested DNA
was ligated into pUC18 and this mixture was transformed into E. coli.
Plasmids were isolated from several transformants and subjected to
restriction analysis with the enzyme EcoRI. These subclones contained
overlapping multi-fragment segments of the original recombinant plasmid.
Alignment of these fragments resulted in the map shown in Figure 2.
DC DC CC DC
0 O Ol O
O O 0 O
in uj nil uj
ft^^^MMMMM^^^^^^MMMBMMB^^^^H^^^^^^^^^W^tHM^B^H
6.1 2.9 0.8 2.0
SC37
SC46
SC46A
SC56
O/^CT
DC CC DC DC DC
O O O O O
0 O O O O
UJ UJ UJ UJ 111
2.9 6.7 2.3 0.5
II
J
II I
II I
II II
I III
I I I
II I
III
Figure 2. Mapping the EcoRI sites of the pGEM410 insert. The positions of
these sites are indicated at the top. The inserts of several subclones were
mapped and aligned as shown. Subclone designations are shown at the left.
244
-------
Recombinant strains containing these subclones were tested for various
enzymatic activities associated with biphenyl metabolism. An examination of
the activities associated with these fragments revealed that the bph genes
were grouped together within a 12.4 kb region of DNA. A partial restriction
map of this region is shown in Figure 3.
2.9kb
2.3kb
0.5kb
VECTOR
bph B,C
bph D
Figure 3. Partial restriction endonuclease map of the region encoding PCB
degradation in pGEM410. Gene designations are as described in Figure 1.
COMPARING PCB DEGRADATION BY RECOMBINANT AND WILD-TYPE STRAINS
The PCB degrading abilities of two recombinant strains, FM4110 and
FM4560, were compared to that of LB400. FM4110 is E. coli strain TB1
containing the original recombinant plasmid pGEM410. FM4560 is E. coli
strain TB1 containing a derivative of pGEM410. This derivative was produced
by inserting the 2.9 and 6.7 kb EcoRI fragments (encoding the initial three
enzymes of the Bph pathway) into the vector pUClS.
Resting cell assays were used to test the ability of FM4110 and FM4560
to degrade PCB mixes IB and 2B (9). The recombinant strains were grown using
succinate as the carbon and energy source since they will not grow on
biphenyl. PCB degradation by FM4110 was significantly lower than that by
LB400 for many tetra-, penta- and hexachlorinated congeners (Table 2).
Furthermore, PCBs with chlorines at both para ring positions were not
attacked by FM4110.
FM4560 showed a substantial increase in PCB degrading ability over
FM4110, and demonstrated activity similar to that of LB400 for a wide variety
of tetra-, penta-, and double-para substituted PCBs (Table 2) . FM4560 was
also similar to LB400 in its ability to degrade Aroclor 1242, a complex PCB
mixture containing 60-80 different congeners. In resting cell assays with 10
ppm of Aroclor 1242, FM4560 and LB400 degraded 85 and 91% of the PCBs,
respectively. As shown in Figure 4, the patterns of depletion demonstrated
by the two strains appear nearly identical. These results demonstrate the
ability of this recombinant strain to degrade a PCB mixture that is. an actual
environmental contaminant.
245
-------
TABLE 2. DEGRADATION OF PCB CONGENERS BY LB400 AND RECOMBINANT STRAINS
PCB
Congener
2,3
2,2'
2,4'
2,5,2'
2,5,4'
2, 5, 2', 5'
2, 3, 2', 3'
2,3, 2', 5'
9 5 3' 4'
^» J » J »**•
2452' 5'
i. ,*-t-, -i , ^ , _/
234?' 5'
^ , .J ,*+, Z. , J
9 4 5 9 ' 3'
^»^» J ,^ » °
4,4'
2,4,4'
2, 4,3', 4'
2, 4,2', 4'
3,4, 3', 4'
2, 4, 5, 2', 4', 5'
LB400
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
****
***
*****
***
*****
**
***
Percent Depletiona
FM
4110
*****
*****
*****
*****
*****
****
***
***
***
*
-
- -
_
-
-
-
-
—
FM
4560
*****
*****
***** '•'
*****
*****
*****
*****
*****
*****
*****
***
***
**
****
*
**
-
—
testing Cell Assays- Mix IB, 2B. Percent depletion: ***** - 80-100%;
**** - 60-79%; *** = 40-59%; ** = 20-39%; * = 10-19%; - = 0-9%.
B
JLL
.ju-M-jJ WU~
Figure 4. Degradation of Aroclor
1242 by E. coli FM4560 and
Pseudomonas LB400. (A) Aroclor 1242
(10 ppm) incubated with mercury
killed cells; (B and C) Aroclor 1242
incubated at 30 C for 24 h with cells
(optical density at 615 nm of 1.0)
of FM4560 and LB400 respectively.
246
-------
COMPARING THE GENES FOR PCB DEGRADATION FROM DIFFERENT ORGANISMS
The PCB-degrading abilities of Pseudomonas strain LB400 and A. eutrophus
H850 are significantly different from those of other bacteria. This is
reflected both in the wide range of congeners they can degrade and in the
presence of 3,4-dioxygenase activity. This might indicate that the bph genes
of LB400 and H850 are similar to each other, and substantially different from
those in other PCB degrading strains. This possibility was examined using
DNA:DNA hybridization experiments.
Genomic DNAs from nine different bacterial strains were examined for
sequences similar to the bph genes of LB400. The selected strains varied
widely in PCB degrading ability. This group included representatives from
four different genera, and six species (see figure legend for a list of the
organisms tested). Figure 5 shows the results of an experiment in which a
plasmid containing the bphAB and C genes of LB400 was used as a probe. This
probe hybridized to fragments of the A. eutrophus H850 genome. Similar
results were obtained using probes containing bphD (data not shown). Probes
with other fragments of LB400 DNA (i.e., non-bph gene fragments) did not
hybridize to H850 DNA. Genomic DNAs from the other PCB degrading strains did
not have sequences that hybridized to these probes. Therefore, the bph genes
of these strains cannot be very closely related to those of LB400 and H850.
These data indicate that there are at least two distinct varieties of genes
for PCB degradation. Also, similarities in the LB400 and H850 bph genes
explain the very similar PCB-degrading capabilities of the two organisms.
FUTURE PROSPECTS
Recombinant strains of E. coli that express the LB400 and H850 bph genes
have been constructed. While significant improvement in the PCB degrading
ability of these strains has already been achieved through genetic
manipulations, further studies designed to increase and regulate the
expression of the LB400 bph genes are currently in progress.
The discovery that the entire bph pathways of LB400 and H850 were
genetically distinct from those of many other PCB degrading strains was
surprising. All of the organisms examined are capable of converting PCBs to
chlorobenzoic acids and it might be expected that some of the genes in these
pathways would be similar to those of LB400. As bph genes from other
organisms become available we will attempt to examine their relationship to
those of LB400, and attempt to correlate specific biochemical activities with
specific types of bph genes.
The recombinant strains have thus far been examined only for their
ability to degrade PCBs in resting cell assays. Since it is important to
evaluate the degradation of PCBs as they are actually found in the
environment, the abilities of these, and future, recombinant organisms to
degrade PCBs on soil will be examined in the laboratory.
Previous studies using LB400 to degrade PCBs on contaminated soils have
been promising, but it is likely that degradation could be increased if cell
survivability could be improved (10). This could be accomplished by using a
wide-host-range mobilizable plasmid to introduce the cloned LB400 bph genes
into an organism indigenous to a contaminated soil. The resulting strain may
247
-------
Figure 5. Autoradiogram of Southern blot where several PCB-degrading strains
were examined for bph genes similar to those of LB400. All DNAs were digested
with EcoRI and probed with pGEM415 (containing bphAB and G). pGEM415,
positive control; C600, E. coli (negative control); LB400, Pseudomonas sp.;
H850, Alcaligenes eutrophus; MB1, Corynebacterium sp.; H336 Pseudomonas
testosteroni; H430, P. testosteroni; Pi434, Alcaligenes fecalis; H1130,
Pseudomonas sp. (acidovarians group); H201, Pseudomonas cepacia.
248
-------
combine the superior PCB degrading ability of LB400 with the survivability of
the indigenous organism.
The fate of recombinant strains and plasmids in the environment is a
question which must be addressed for any bioremediation process utilizing
genetically engineered organisms. Studies to assess the environmental
transfer, mobility, and persistence of recombinant molecules often employ DNA
hybridization methods because they are highly sensitive and specific. The
availablity of DNA probes for the LB400/H850 bph genes makes it possible to
use these techniques to determine the stability of the cloned genes in the
environment and to detect them even in organisms where they are not
expressed.
ACKNOWLEDGEMENTS
This work was supported by Grant CR812727 from the U.S. Environmental
Protection Agency, Office of Research and Development, Hazardous Waste
Engineering Research Laboratory, Cincinnati, OH. The authors thank Dr. P.R.
Sferra, EPA Project Officer for his interest, support and suggestions.
REFERENCES
1. Rochkind, M.L., Blackburn, J.W., andG.S. Sayler. Chlorinated biphenyls.
In, Microbial decomposition of chlorinated aromatic compounds. United
States Environmental Protection Agency, Cincinnati OH, 1986. 129 pp.
2. Bedard, D.L., R. Unterman, L.H. Bopp, M.J. Brennan, M.L. Haberl, and C.
Johnson. Rapid assay for screening and characterizing microorganisms
for the ability to degrade polychlorinated biphenyls. Appl. Environ.
Microbiol. 51: 761, 1986.
3. Bopp, L.H. Degradation of highly chlorinated PCBs by Pseudomonas strain
LB400. J. Ind. Microbiol. 1: 23, 1986.
4. Nadim, L., M.J. Schocken, F.J. Higson, D.T. Gibson, D.L. Bedard, L.H.
Bopp, and F.J. Mondello. Bacterial oxidation of polychlorinated
biphenyls. Proceedings of the 13th Annual Research Symposium on Land
Remedial Action, Incineration, and Treatment of Hazardous
395, 1987.
Disposal,
Waste, p
A.M.
Furukawa, K. Microbial degradation of polychlorinated biphenyls.
Chakrabarty, (ed.), In, Biodegradation and detoxification of
environmental pollutants, CRC Press, Inc., Boca Raton, Florida, p. 33,
1982.
Mondello, F.J. and L.H. Bopp. Genetic and cell-free studies of PCB
biodegradation in Pseudomonas putida LB400. Proceedings: Biotech USA
1987. Online International Inc. p. 171, 1987.
Frey, J., M. Bagdasarian, D. Feiss, F.C.H. Franklin, andJ. Deshusses.
Stable cosmid vectors that enable the introduction of cloned fragments
into a wide variety of Gram-negative bacteria. Gene 24: 299, 1983.
249
-------
8. Furukawa, K., and T. Miyazaki. Cloning of a gene cluster encoding
biphenyl and chlorobiphenyl degradation in Pseudomonas
pseudoalcallgenes. J. Bacteriol. 166: 392, 1986.
9. Bedard, D.L., M.L. Haberl, R.J. May, and M.J. Brennan. Evidence for
novel mechanisms of polychlorinated biphenyl metabolism in Alcaligenes
eutrophus H850. Appl. Environ. Microbiol. 53: 1103, 1987.
10. Unterman, R., D.L. Bedard, M.J. Brennan, L.H. Bopp, F.J. Mondello,
R.E. Brooks, D.P. Mobley, J.B. McDermott, C.C. Schwartz and D.K.
Dietrich. Biological approaches for PCB degradation. In: Omenn
et.al., (eds.), Reducing Risks From Environmental Chemicals Through
Biotechnology. Plenum Press, London. 1988. p. 253.
250
-------
Treatment of Wood Preserving
Soil Contaminants
by White Rot Fungus
John A. Glaser
United States Environmental Protection Agency
Risk Reduction Engineering Laboratory
26 Martin Luther King Drive
Cincinnati, Ohio 45268
Rich Lamar, Diane Dietrich, and T. Kent Kirk
United States Department of Agriculture
Forest Products Laboratory
1 Gifford Pinchot Drive
Madison, Wisconsin 53706
ABSTRACT
A wood degrading fungus, Phanerochaete chrysosporium has
been the object of considerable attention for its potential
application to hazardous waste degradation. The development of a
field soil treatment technology based on this fungus has been the
focus of an intense research program. Early stages of this work
sought to determine ways to assist the growth of the fungus in
soil, an environment not known to be sought by this fungus. Once
general methodology was established to promote its sustained
growth then studies were pursued to quantitatively determine the
extend of biodegradation attributable to the fungus using
surrogate soils(artificially contaminated with wood preserving
waste constituents). Work with pentachlorophenol and related
polyaromatic hydrocarbons has paved the way to undertake limited
field trials during the North American growing season of 1989.
The related bench scale experimental work and some information
concerning the field trials will be presented.
Introduction
The use of specialized or selected microorganisms to degrade
soil bound contamination has received considerable interest. This
situation is attributable in part to the interest of those
responsible for cleanup actions to use cost effective and
environmentally compatible means to treat wastes. In spite of
these very promising aspects, biological detoxification, as a
site cleanup technology, must be recognized as a developing
technology. Biological treatment has excellent credentials in the
areas of municipal and industrial wastes but awaits development
for the treatment of mixtures of more toxic and persistent
chemicals found as components of hazardous waste sites.
251
-------
Wood Treating Waste [1]
A major hazardous waste problem confronting regulatory
authorities in the United .States is the waste associated with the
wood treatment industry. Depending on the age of a facility, at
least three technologies have contributed to the accumulated
waste. Creosote treatment was followed by pentachlorophenol which
was replaced with copper chromated arsenite. Each of these
technologies present its special conditions for cleanup. Creosote
is derived from coal tar production and usually contained a host
of compounds ranging from straight aromatic compounds to
polyaromatic species including smaller quantities of aromatic
nitrogen bases and an array of phenolic compounds. The lower
vapor pressure components of this mixture contribute to the
residuals found at such sites. Pentachlorophenol, a potent
fungicide, is a major contaminant at wood treating site that
exhibits significant toxicity towards microflora. The analysis of
wastes derived from this technology have identified other
potential toxic components. For our current development efforts,
we have narrowly focussed on a significant portion of the waste
including the major contributors that are polycyclic aromatic
compounds and phenols. It is necessary to limit the scope of
contamination treatable by this technology to permit the
development activity to be achieveable in a reasonable time
frame.
Bacteria vs Fungi [2]
Microorganisms (both bacteria and fungi) are known to
possess a variety of detoxification skills [3]. Xenobiotic
chemical pollutants generally do not provide sufficient energy to
sustain many microorganisms. The biological degradation of such
substrates occurs as part of a cometabolic activity, where the
organism's growth is maintained by specific substrates and the
detoxification activity of other materials non-growth substrates
ensues. Many bacteria and fungi can accomplish simple
transformations on organic substrates but often fail to complete
the conversion of toxicant substrate to carbon dioxide or
generate a toxic intermediate that can impair the growth of the
microorganism. The use of bacterial communities recognizes these
deficiencies through the combined use of many species where the
abilities of one species supplants the inadequacies of another.
Since the collective action of these communities is important to
treatment success, it is important to protect them from
environmental effects that may adversely affect the communities
[4].
Fungi have not been investigated to any extent for use as
degraders of waste materials until recently [5]. Sewage treatment
252
-------
operations steered clear of filamentous fungi due to processing
problems and the possibility that such fungi may be pathogenic.
Exceptions to these generalization do exist [6]. A wood rotting
basidiomycetes, Trametes versicolor, was studied, twenty-five
years ago, in an attempt to quantify its ability to degrade
chlorinated phenols [7]. Wood preservative chemicals have been
found to "be degraded by fungi [8,9].
A wood degrading fungus, Phanerochaete chrysosporium
characterized by fast growth and easy reproductive cycles
degrades an extensive list of hazardous waste consistuents under
laboratory conditions. This ability to degrade hazardous
pollutants appears to corellate well with the fungus1 ability to
degrade lignin, a complex natural polymer composed of
phenylpropane units that is resistant to decay by many
microorganisms. This lignin degrading ability is attributable to
a complex mixture of enzymes secreted by the fungus to the
extracellular medium. The enzymes are peroxidases that utilize
hydrogen peroxide from complementary enzyme systems to perform
the initial oxidative conversion of pollutant substrates.
Some of the more common substructures of lignin resemble the
chemical structure of many persistent organic compounds
contaminating the environment. This structural similarity gave
sufficient reason to pursue application of a white rot fungus,
Phanerochaete chrysosporium to the biodegradation of hazardous
waste constituents [10].
Phanerochaete chrysosporium is a filamentous, white, wood
rotting fungus and has been classified as a member of the
Hymenomycetes subclass of Basidiomycetes [11]- To distinguish
them from bacteria, fungi are eukaryotic, ie. they possess a
nuclear membrane and as microorganisms are considered to be
plantlike without chlorophyll having no photosynthetic abilities
[12].
Carbon Substrate Degradation by Wood Rotting Fungi
White rot fungi are primary wood degraders in nature[10].
They excel in their ability to recycle carbon of wood origin when
compared with brown rot fungi. The naturally occuring polymers of
cellulose and lignin are degraded by these fungi forming the
major sources of carbon to assist fungal growth. Lignin is by far
the more difficult to degrade due to its composition as a
heteropolymer formed from the cross linking of three precursor
cinnamyl alcohols and cannot serve as the sole carbon source for
growth of the fungus [13]. The fungus must be able to switch its
ability degrade these various polymers as the concentration of
polymer varies with the composition of the wood. This ability for
non mutant strains of P. chrysosporium is controlled by the
253
-------
absence of certain nutrients. Nitrogen deficiency is generally
used to induce this secondary metabolic cycle of lignin
utilization [14] .
The enzyme systems responsible for the initial attack on
lignin require unusual abilities due to the complexity and
resistance of the lignin structure. The 600-1000 k-dalton size
range for lignin is far too large to enter the cells of
microorganisms by known transport systems. An enzyme system
permitting the microorganism to overcome this limitation would
most likely be extracellular, non-specific (due to the
heterogenity and large molecular weight -of the substrate), and
resistant to protease destruction. It is important to realize
that the ability of P. chrysosporium to degrade lignin by these
extracellular enzymes occurs in a secondary metabolic cycle. The
fungus uses cellulose as its primary growth substrate but when
large quantities of lignin are encountered or certain nutrients
are not present the secondary metabolic cycle is entered [15].
The extracellular lignin degrading enzymes serve to fragment
lignin into pieces that can be assimilated by the fungus. This
conceptualization of degradation activity stresses the importance
of the individual enzyme's wide range activity and function. The
intracellular enzyme components complete the conversion of the
lignin fragments into carbon dioxide.
Lignin cleavage reactions are catalyzed by a hemoprotein
ligninase [16-18]. Hydrogen peroxide is consumed in this
reaction that degrades lignin indicating a peroxidative
mechanism. The generation of hydrogen peroxide has been
attributed to three different enzymes: glucose oxidase [19],
pyranose-2-oxidase [20], and methanol oxidase [21]. The glucose
oxidase enzyme, considered the major contributor to hydrogen
peroxide production, may be located in unique periplasmic
microbodies. Stoichiometries of product formation as well as
hydrogen peroxide and oxygen uptake are consistent with a radical
pathway [16]. These results established the one-electron
oxidative mechanism as the primary extracellular oxidative
pathway for P. chrysosporium.
Degradation Studies of Waste Constituents
Radiorespirometric studies of the degradation of [U
pentachlorophenol in aequous media indicated .that the substrate
was rapidly converted to carbon dioxide. Enzyme studies showed
that pentachlorophenol is converted to the 1,4-
tetrachlorobenzoquinone by the fungus [22.]. The quinone was
difficult to quantify due to its propensity to form, charge
transfer complexes with cellular materials. Further elucidation
of the metabolic pathway is in progress. Several aromatic
254
-------
hydrocarbons, benzo[a]anthracene, pyrene, anthracene,
benzo[a]pyrene and perylene, (constituents of creosote), were
converted to carbon dioxide by the fungus in liquid
culture[17,18]. This latter finding serves to differentiate the
fungus from bacterial species since few bacteria have the ability
to utilize the higher molecular weight aromatic polycyclics.
Life Cycle of Phanerochaete chrysosporium
To adequately harness the striking abilities of the wood
rotting fungi, it is necessary to understand their life cycle in
order to optimize the treatment process. The life cycle of
Hymenomycetes fungi is characterized by many structures formed
during vegetative, sexual, and asexual reproductive phases [12].
The fungal mycelium, a mass of interwoven filamentous
hyphae, is usually submerged in growth medium when cultured in
liquid. The mycellium passes through three distinct stages of
development. The primary mycellium growth phase is not vigorous.
Once secondary mycellium is formed subsequent growth is
frequently different from the primary mycellium. As the mycellium
tissues organize and specialize the tertiary phase is initiated.
Secondary and tertiary mycellia comprise the vegetative segment
of the life cycle. The vegetative phase is the longest and
dominant growth phase. The highest concentration of extracellular
enzymes are secreted during the vegetative phase. Eventually the
tissues of the tertiary mycellium differentiate into fruiting
bodies that are shed depending on environmental conditions.
Asexual reproduction, continued maintenance of current degrading
abilities, can occur anytime during the vegetative growth phase.
P. chrysosporium produces asexual spores prolifically and at all
stages of the life cycle [23].
It has been shown that P. Chrysosporium produces at least
ten extracellular hemoproteins and roughly half have ligninase
activity [24]. The heterogeneity among the various extracellular
proteins produced by P. chrysosporium points to possible
functional differences among them important to pollutant
degradation.
Soil Detoxification Technology Development
The general success of liquid phase biodegradation studies
with the fungus stimulated speculation that this microorganism
may be an appropriate candidate for the treatment of contaminated
soils. Attempts to innoculate environmental matrices with non-
native microorganisms have met with varying degrees of success
[25]. The elucidation of optimal practices leading to successful
innoculation of contaminated environmental materials remains to
be discovered [26]. At the outset of this research, P.
255
-------
chrysosporium was not known to inhabit the soil. Due to this
general lack of knowledge of the habitat, a rather cautious
research effort was engaged to determine the ability of the
fungus to inhabit and thrive in the soil. Early work, indicated
that P. chrysosporium did not grow well in non-sterile soils;
this may be attributable in part due to ineffective competition
with the indigenous microflora. These results were anticipated
since the soil is not the normal habitat of P^ chrysosporium.
Lately, it has been found that growth within the soil can be
accomplished through the use of larger quantities of inoculum
[27,28].
Recent research has assessed the effects of selected soil
types, temperatures, pH, and water potentials on the growth of
the fungus in sterile and non-sterile soils. Three well
characterized soils(topsoil and subsoil) were used in this work
(Table I). The effect of soil type, temperature, and water
potential on the growth of P. chrysosporium in three sterile
soils was evaluated in a factorial experiment. Soils were
sterilized by fumigating with methyl bromide to avoid confounding
effects from native microflora. Growth of the fungus was
evaluated at five soil temperatures ranging from 25 to 39°C and
four water potentials ranging from -0.03 to -1.5 MPa. The extent
of growth was determined by measuring the amount of ergosterol
that could be extracted from two sub-samples of the soil from
each test at the end of a two week incubation period and reported
as ug. of ergosterol/ g. of soil [29,30].
Growth of the fungus was the greatest in the Marsham,
intermediate in the Xurich and least in the Batavia soil. The
same trend was observed when a visual assessment system of growth
was used. Soil water potential had a significant affect on the
growth of the fungus in soil. As the .soil water potential was
increased (corresponds to decrease in soil water content), fungal
growth decreased. Water potential is another easily controlled
soil factor [31]. Growth of the fungus between 25 and 35°C was
unaffected but significantly decreased at 39°. These results do
not agree with earlier work using the visual growth estimation
technique. The difference may be attributable to an increase of
sporulation of the fungus at the soil surface with increased
temperature leading to a biased measurement of growth. Soil
temperatures under field conditions can be controlled by
selecting the normal warm months for operation and by soil
solarization. Biomass accumulations as well as growth habit of P.
chrysosporium were greatly influenced by soil type.
Growth Measurement
The measurement of growth has presented major problems for
256
-------
this study. Assessment of growth in the early stages of these
studies was done by means of visual estimation on a ranking
basis. Due to the three dimensional growth patterns of the fungus
in a solid substrate such as soil, it was necessary to find a
more reliable means to determine the extent of fungal growth. The
normal means of assessing growth in bacterial systems have no
application to the present study. Based on work investigating the
infestation of cereal grains by fungi, a mycosteroid, ergosterol
(ergosta-5:6,7:8,22:23-trien-3-ol) has been employed as a
quantitative means to determine growth of the fungus [32]. It has
been shown to correlate well with the visual estimation
technique.
The application of the fungus to soil treatment is the
remediation of wood treating sites. Target pollutants identified
for treatment at these sites are pentachlorophenol(PCP) and the
major aromatic hydrocarbon contaminants found in creosote
(napthalene, anthracene, and phenanthrene). Creosote has been
extensively characterized and new substrates will be added to the
mixture when deemed necessary [33]. The degradative ability of
the fungus in the soil has been evaluated through the measurement
of evolved labelled carbon dioxide. Disappearance of the parent
compound was monitored by GC or HPLC techniques. Separation of
the soil into solvent extractable, humic acid, fulvic acid and
humin fractions permitted material balance evaluations [34].
The degradation of 14C [UL]-pentachlorophenol was studied
over an eight week period in the three soils. Mineralization,
volatile losses, extractable PGP in the soil, and soil residuals
containing bound PGP and transformed products were measured to
develop a tight material balance [35]. A very small percentage of
the total 14C was accounted for. by mineralization and
volatilization. Both mineralization and volatilization were
significantly greater in innoculated than in non-innoculated
cultures of the three soils. The extractable quantities of PGP
were greatly reduced by innoculation with P. chryspsporium. The
greatest rate of PGP removal due to fungal activiy was found with
the Marsham soil. Extractable PGP after 14 days was about 2 ppm
of the original 50 ppm spiked amount and that was reduced
roughly 1 ppm after an additional 14 days of treatment. Decreases
of PGP concentration in the control tests are attributable to
several potential causes: abiotic avenues of degradation or loss,
irreversible binding to the soil or degradation due to the
regrowth of native organisms.
The closure of material balances for these experiments is
made possible by the careful and detailed analysis of the soil
fractions for labelled carbon content. The fate of ^C PGP in the
soil was determined by analysis of the recoverable carbon label
from: an organic extractable fraction, the soil organic matter
257
-------
(humic and fulvic fractions), and the non-extractable humin
fraction. Combustion analysis was used to assay the amount of ^C
associated with humic and fulvic acid fractions, and the non-
extractable humin fraction.
The percent of total 14C recovery ranged from 55 to 84% over
a 56 day period in the Marsham soil tests. Volatilization losses
were less than 3% for the three soils. There was a significant
amount of labelled carbon activity associated with nonextractable
fractions, indicating that there is possibly incorporation in the
soil material of the pollutant substrates during its metabolism.
Bollag has shown the possible polymerization reactions between
pentachlorophenol and soil chemicals such as syringic acid [36].
Polymeric forms of a series of pollutants were constructed by
Haider and Martin, who showed that P. chrysosporium would degrade
these higher molecular weight materials but at a slower rate than
the parent pollutant substrate [37].
Future research in the soil application will include small
scale treatment of selected pollutants at environmentally
significant concentrations, the evaluation of admendments on
primary and secondary metabolism, and the delivery of oxygen
within the soil to the growing fungus.
258
-------
Table I. Physical and Chemical Characteristics of Test Soils
Macro Features
Soil Type
Texture
Horizon
Cation Exchange
Capacity (meq/lOOg)
Base Saturation (%)
PH
Organic Matter(%)
Nitrogen(%)
Batavia
silty clay/loam
Bt2
17
29.5
5.4
0.5
0.05
Marsham
sandy loam
A
38
66.4
6.8
12.0
0.46
Xurich
sandy loam
A
14
24.0
,7.1
39.0
0.18
Trace Constituents(ppm)
Calcium
Magnesium
Potassium
Phosphorous
Boron
Mnaganese
Zinc
Sulfur
1950
850
145
75
0.6
12.5
1.9
9.7
4900
1650
90
17
1.3
9.0
12.2
11.3
1675
640
80
17
0.8
77.5
6.4
3.9
259
-------
References
1. U.S. Department of Agriculture. 1980. The Biological and
Economic Assessment of Pentachlorophenol, Inorganic Arsenicals,
Creosote vol 1 and 2. Technical Bulletin 1658-1 and 1658-11.
2. Griffin, D.M. 1985. A Comparison of the Roles of Bacteria and
fungi. In Bacteria in Nature Vol 1. (ends) E. R. Leadbetter and
J.S. Perlmutter, pp. 221-255. New York: Plenum.
3. Wainwright, M. 1988. Metabolic Diversity of Fungi in
Relation to Growth and Mineral Cycling in Soil - A Review.
Trans. Br. Mycol. Soc. 90 150-170.
4. Slater, J.H. and D. Lovatt. 1984. Biodegradation and the
Significance of Microbial Communities. I_n Microbial Degradation
of Organic Compounds (ed) D.T. Gibson, pp. 439-4
5. Eaton, D.C., Mineralization of polychlorinated biphenyls
by Phanerochaete chrysosporium; a ligninolytic fungus. Enzyme
Microb. Technol.. 1985, 7, 194-196.
6. Bollag, J.-M. 1972. Biochemical Transformation by Soil Fungi.
Grit. Rev. Microbiol. 2, 35-58.
7. Lyr, H., Enzymatische detoxifikation chlorieter phenole,
Phytopathol. Z., 1963, 47, 73-83.
8. Duncan, C.G., and Deverall, F.J., Degradation of wood
preservatives by fungi, Appl. Microbiol.. 1964, 12, 57-62.
9. H.H. Unligli, Depletion of pentachlorophenol by fungi, Forest
Prod. J. 18 45-50 (1968).
10. Bumpus, J.A., and Aust, S.D., Biodegradation of
Environmental Pollutants by the White Rot Fungus Phanerochaete
chrysosporium; Involvement of the Lignin Degrading System.
BioEssays, 1987, 6, 166-170.
11. Burdsall, H.H. and Eslyn, W.E., A new Phanerochaete with a
chrysosporium imperfect state. Mycotaxon. 1974, 1, 123-33;
12. Deacon, J.W., Introduction to Modern Mycology, Blackwell
Scientific Publications, Oxford, 1984, pp.1-24.
13. Kirk, T.K. and Shimada, M., In Biosynthesis and
Biodegradation of Wood Components, ed. T. Higuchi, Academic
Press, N.Y., 1985, pp 579-605.
260
-------
14. Kirk, T.K.
of Organic Compounds,
1984, pp. 399-438.
Degradation of
ed. D.T.
Lignih. In Microbial Degradation
Gibson, Marcel Dekker, New York,
15. Kirk, T.K. and R,L. Farrell. 1987. Enzymatic "Combustion":
The Microbial Degradation of Lignin. Ann. Rev. Microbiol. 41,
465-505.
16. Hammel, K.E., Tien, M., Kalyanaraman, B., and Kirk, T.K.,
Mechanism of oxidative C-C cleavage of a lignin model dimer
by Phanerochaete chrysosporium ligninase: Stoichiometry and
involvement of free radicals. J. Biol. Chem., 1985, 260,
8348-53.
17. Hammel, K.E., Kalyanaraman, B., and Kirk, T.K., Oxidation of
polycyclic aromatic hydrocarbons and dibenzo[p]dioxins by
Phanerochaete chrysosporium ligninase. J. Biol. Chem., 1986,
261, 16948-52.
18. Haemmerli, S.D, Liesola, M.S.A., Sanglard, D., and Feichter,
A., Oxidation of benzo(a)pyrene by extracellular ligninases,
from Phanerochaete chrysosporium, J. Biol. Chem., 1986, 261,
6900.
19. R.L. Kelley, and C.A. Reddy
Phanerochaete chrysosporium. In
BIOMASS Part
S.T. Kellogg
B, Lignin,
PP
Pectin,
307-315.
1988. Glucose Oxidase of
Methods in Enzymology, Vol 161,
and Chitin. (eds) W.A- Wood, and
20. Vole. J. and K.-E. Eriksson. 1988. Pyranose-2-Oxidase from
Phanerochaete chrysosporium In Methods in Enzymology, Vol 161,
and Chitin. (eds) W.A. Wood, and
BIOMASS Part
S.T. Kellogg.
21. Eriksson,
Phanerochaete
B, Lignin, Pectin,
pp. 316-321.
K.-E., and A. Nishida. 1988. Methanol Oxidase of
chrysosporium In Methods in Enzymology, Vol 161,
BIOMASS Part
S.T. Kellogg
B, Lignin, Pectin, and Chitin. (eds) W.A. Wood,
pp. 322-326.
and
22. Hammel, K.E. and P.J. Tardone. 1988. The Oxidative 4-
Dechlorination of Polychlorinated Phenols is Catalyzed by The
Extracellular Fungal Lignin Peroxidases. Biochemistry 27, 6563-
6568.
23. Gold, M.H, and Cheng, T.M., Conditions for fruit body
formation in the white-rot basidiomycete Phanerochaete
chrysosporium. Arch. Microbiol., 1979, 121, 37-41.
261
-------
24. Kirk, T.K., Groan, S., Tien, M., Murtaugh, K.E., and Farrell,
R.L., Production of multiple ligninases by Phanerochaete
chrysosporium; Effect of selected growth conditions and use of
mutant strain. Enz. Microb. Tech.. 1986, 8, 27-32.
25. Zaidi, B.R., Stucki, G., and Alexander, M., Low Chemical
Concentration and pH as Factors Limiting the Success of
Innoculation to Enhance Biodegradation, Environ. Toxicol.
Chem., 1988, 7, 143.
26. Goldstein, R.M., L.M. Mallory, and M. Alexander. 1985.
Reasons for Possible Failure of Innoculation to Enhance
Biodegradation. Appl. Environ. Microbiol. 50, 977-983.
27. Lamar, R.T., Larsen M.J., Kirk, T.K., and Glaser, J.A.,
Growth of the white-rot fungus Phanerochaete chrysosporium in
soil. In Land Disposal, Remedial Action, Incineration and
Treatment of Hazardous Waste Proceedings of the 13th Annual
Hazardous Waste Symposium, EPA/600/9-87/015, \T. S~T
Environmental Protection Agency, Cincinnati, Ohio, 1987, pp. 419-
429.
28. Lamar, R.T., Larsen, M.J., Kirk, T.K., and Glaser, J.A.,
Effect of Biotic and Abiotic Soil Factors on Growth and
Degradative Activity of the White-Rot Fungus Phanerochaete
chrysosporium Burds. In Chemical and Biochemical
Detoxification of Hazardous Waste,(ed.) J.A. Glaser, Lewis
Publishers, Ann Arbor, MI., 1988. (In Press)'.
29. Matcham, S.E., B.R. Jordan, and D.A. Wood. 1985. Estimation
of Fungal Biomass in Solid Substrate by Three Independent
Methods. Appl. Microbiol. Biotechnol. 21, 108-112.
•30. Seitz, L.M., D.B. Sauer, R. Burroughs, H.E. Mohr, and J.D.
Hubbard. 1979. Ergosterol as a Measure of Fungal Growth.
Phytopathology 69, 1202-1203.
31. Sommers, L.E., Gilmour, C.E., Wildung, R.E., and Beck, S.M.,
The effect of water potential on decomposition processes in
soils. In Water potential relations in soil microbiology,
SSA Special Publication Number 9,Soil Sci.Soc.of Amer.,
Madison, WI., 1981, pp.97-117.
32. Weete, J.D., and D.J. Weber. 1980. Lipid Biochemistry of
Fungi and Other Organisms. Plenum Press, New York.
33. Nestler, F.H.M., U.S. Dept. Agri. For. Ser. Res. Pap. FPL
195. U.S. Department of Agriculture,Forestry Service,Forest
Products Laboratory, Madison, WI.
262
-------
34. Stevenson, F.J. 1982. Humus Chemistry, Chapt 2. pp
Wiley, New York.
26-54.
35. Mayaudon, J. 1971. Use of Radiorespirometry in Soil
Microbiology and Biochemistry, In Soil Biochemistry Vol 2 (eds)
A.D. Mclaren and J. Skujins. pp. 202-256.
36. Bollag, J.-M. and S.-Y. Liu, Coploymerization of halogenated
phenols and syringic acid, Pest. Biochem. and Physiol. 23,
261 (1985).
37. Haider, K.M. and J.P. Martin. 1988. Mineralization of 14C-
Labelled Humic Acids and Humic-Acid Bound 1^C-Xenobiotics by
Phanerochaete chrysosporium. Soil Biol. Biochem 20, 425-429.
263
-------
BIOLOGICAL TREATMENT OF PETROCHEMICAL SLUDGES
Stephen D. Field, Associate Professor of Civil Engineering
Andrew K. Wojtanowicz, Assistant Professor of Petroleum Engineering
Louisiana State University
Baton Rouge, Louisiana 70803
ABSTRACT
The research results of using a microbial population acclimated to
petrochemical waste sludges to degrade the complex hazardous organic sub-
strates found in these sludges are presented. The results obtained include
the determination of microbial population stability, rates of growth and
substrate utilization for selected single compound substrates1and two API
separator sludges. Specific growth rates as high as 0.38 hr~ have been
obtained for an API separator sludge treated in a batch reactor loaded with
five percent oil (v/v). Cultural stability appears to have been established
after three months of continuous loading in complete-mixed stirred tank
reactors for two different API separator sludges; microbial composition
differs for each of the API sludges. These results are being included in a
growth rate model to relate specific growth rate to substrate concentration
and desorption rates of organic compounds from the inorganic solid consti-
tuents as a function of microbial culture activity. The results of this
research are the initial step towards a detailed description and evaluation
of biological treatment of complex organic hazardous waste sludges.
INTRODUCTION
Hazardous waste sludges from the petroleum industry (e.g., API separator
sludges, DAF floats, storage tank still bottoms) and contaminated soils from
hazardous waste spills and at 'Superfund' sites lack technical and cost
effective treatment methods. Petroleum industry wastes primarily rely on
landfarming for degradation of the petroleum compounds and sludges. This
process suffers from low loading rates (less than 10% oil in the cell), long
treatment periods (1-3 years), high odor emission, and potential for ground-
water pollution. In addition, landfarming is considered a land disposal
method by the USEPA and falls under the land ban requirement. Therefore, an
efficient alternate technology needs to be developed. In a similar manner,
effective soil cleansing of petrochemical components at Superfund and other
264
-------
spill sites is lacking; incineration is not an optimal technology for soils
and fixation/stabilization still requires removal and disposal in a secure
landfill.
The expected promulgation of a stricter test for defining hazardous
wastes by EPA, and the existing "land disposal ban" will bring substantially
greater amounts of wastes under the hazardous waste regulatory authority.
Increased public opposition to off-site transport and incineration of
hazardous wastes will require the development of alternative, economical
treatment methodologies such as bioengineered systems, in order to meet the
goals of a clean environment and economic growth. The support of the
development of bioengineered systems to treat wastes on-site offers potential
economic benefits through cost savings to the petroleum and petrochemical
industry, as well as a methodology to treat hazardous wastes.
The isolation and development of biological cultures in the laboratory
to degrade toxic and bioinhibitory materials has been previously estab-
lished. However, there is a dearth of concurrent development of engineering
data and information necessary to translate the laboratory performance of the
microbial populations to successful full-scale applications.
A microbial culture has been developed which has the ability to survive
in and consume natural crude-oil long-chain hydrocarbons and processed hydro-
carbons, including diesel oils. This culture, which relies on the cometabo-
lism of several species of bacteria, fungi and protozoa, has has degraded 80
to 90% of the oil present in slurries containing as much as 24% oil and 24%
solids. This success in degrading high strength petroleum and petrochemical
wastes has been tested and found to be reproducible for selected USEPA
classified hazardous sludges (e.g., API separator sludges, silt pond sludges)
generated in the petroleum processing industry.
Previous research demonstrated the effectiveness of the microbial
populations for the biodegradation of oilbased drilling muds and oilfield
production wastes (1,2,3). Specifics of these studies included investigation
of aerobic biodegradation of drill cuttings containing 58% oil v/v and 48%
solids w/w and an oilfield production sludge containing 13% oil v/v and 30.4%
solids w/w. Results indicated that the hydrocarbons were tightly bound to
the soil and clay structure. The process of aerobic biodegradation, however,
quickly destroyed the bonding mechanisms and released the oil for rapid
biodegradation.
The microbial culture under investigation has demonstrated excellent
potential for degrading Principal Organic Hazardous Constituents (POHC's) in
waste oil and API separator sludges. The microbes have been found to
effectively consume long-chain hydrocarbons (up to C-36), aromatics and
chlorinated hydrocarbons. Two API separator wastes containing 26-28% oil
w/w, 30% solids w/w and significant amounts of up to 19 different POHC's,
including: benzene, phenol, chloroform, 1,2-dichloroethane, ethylbenzene,
chlorobenzene and toluene, have also been significantly biodegraded. In
similar tests of a 1000 mg/L chlorinated pesticide substrate and a 14000 ug/L
benzene substrate, the population exhibited microbial viability and proli-
feration under the toxic environments.
265
-------
The overall objective of this research program is to examine the feasi-
bility of using a heterogeneous microbial population acclimated to petroleum
and petrochemical waste sludges to degrade the complex hazardous organic
substrates found in these wastes. This paper presents the preliminary
results on the determination of: microbial population stability for two API
separator sludges, and specific growth rates and substrate utilization
measured for an API separator sludge.
MATERIALS AND METHODS
The measurement of specific growth rates and substrate utilization rates
were performed in batch reactors using sealed 250 ml erlenmeyer flasks
incubated in a constant temperature shaker bath. Procedures are outlined in
previous publications (4,5). From the specific growth rates and substrate
utilization rates, Monod growth kinetic coefficients (maxium specific growth
rate and half-velocity constant) and yield coefficients were able to be
calculated (6).
Two API separator hazardous wastes were used for continuous loading
experiments. These materials were designated as BRE API Separator and PLS
API Separator sludges, and were obtained from different sources. Samples of
these sludges contained oil-rich and water-rich layers above the bottom
sludge. Removal of the surface oil and water left oil saturated settled s
sludge for use in the continuous reactor program. This pretreatment was
instituted since a commercial processing route in a continuous reactor
processing system would typically use a preseparator to remove the free oil
and surplus water. The two continuous culture aerobic systems have been
operated for over a six month period and periodically monitored for substrate
removal and microbial population assay. Microbial population assays included
microorganism identification and enumeration, and metabolic capability by
testing for percentage of organisms in the total population that can metabo-
lize various substrates. Details of the reactors' specifications operation
and testing protocol have been published elsewhere (4). A summary of key
operational data is presented in Table 1.
RESULTS AND DISCUSSION
The results of microbial growth and respiration measurements for the BRE
separator oil are presented in Figure 1. The microbial growth and respira-
tion rates increase with increasing substrate concentration; the maximum
growth rate measured was 0.38 hr~ for a substrate concentration of five
percent oil (v/v). This growth rate is well above the required growth for
effective treatment to reduce the oil content of typical waste API separator
sludges^ For comparison, Gaudy et al. (7) was able to obtain growth rates of
0.35 hr on sucrose using municpal sewage organisms. The upper range of
growth in activated sludge systems treating municpal sewage is approximately
0.42 hr (8). An average true yield coefficient (gm TOG cell mass pro-
duced/gm TOG substrate consumed during log growth) of 0.28 was determined
(5). This value is similar to those reported for heterogeneous populations
grown on mixed priority pollutant substrates (9).
266
-------
TABLE 1. OPERATING DATA FOR TWO CONTINUOUS CULTURE REACTORS
USING BRE AND PLS HAZARDOUS WASTE SLUDGES
Parameter
BRE Sludge
PLS Sludge
Settled Sludge Feed, g/d
Feed Sludge Density, gm/ml
Feed Sludge pH
Oil Content, v/v %
Reactor Volume, ml
Mean Cell Residence Time, d
Redox Potential, mv
Reactor pH range
10
1.87
6.3
3.7
1000
5
+295
6.5 - 7.2
10
1.11
7.1
6.7
1000
5
+355
5.7 - 7.2
The growth curves for the API separator sludge substrates demonstrate a
typical lag phase prior to log growth. During this period, the free phase
hydrocarbons floating on top of the water phase can be seen to be emulsified
and made readily available to the microbial population. This period of time
seems to be somewhat variable, typically ranging from 10 to 20 hours. This
lag phase can be shortened by using stock cultures which are presently in
log growth to initiate the experiment.
The growth rate results presented must be viewed as encouraging but
preliminary; they are being repeated for verification. In addition, other
substrates are being tested to examine the effectiveness of biological
treatment for these types of wastes. In particular, the attendant principal
organic hazardous constituents in the separator sludges currently are being
monitored for degradation in the API separator oil experiments.
The microbial cultures grown in continuous culture loading experiments
for the two different API separator sludges showed adaptation to the specific
sludge source. The BRE API separator sludge generally was the more difficult
to achieve growth and substrate removal, even though the kinetic experiments
demonstrated excellent growth rates. The initial microbial culture screening
techniques revealed no actinomycetes growth, two fungal populations, four
varieties of free swimming protozoa, and bacteria tentatively identified as
Pseudomonas fluorescens, and strains of either Azotobacter, Azomonas, or
Beijerinckia. The PLS separator sludge culture differed by demonstrating the
presence of actinomycetes, and only one fungal population. The ciliates and
bacteria were similar to the BRE separator culture.
267
-------
cn
*-•
u
6000 -i
5000 -
4000-
3000 -
2000-
1000 -
a 0.195
« 0.595
n 1.055
* 5.095
' • I " " I1' " I |n i u ii i n 1111111111
20 25 30 35 40 45 50 55 60
Time (hrs)
10 n
31
*->
«r-
W
c
CO
o
15
o
*->
a
O
1 -
.01
Q 0.1 S5 (1=0.04
+ 0.595 (1=0.11
» 1.0% (1=0.12
<• 5.0S5 |l=0.3S
20 25 30 35 40 45 50 55 60
Time (hrs)
Figure 1. Respiration and growth measurement
for BRE-sludge oil.
Examination of the common fungal population to both separator sludge
sources has been tentatively identified as Phialophora j eanselmei, a dark
green fungus common to polluted waters. The other fungal growth present in
the BRE separator sludge culture has yet to be identified.
Observations after one month of loading reveal population changes in
both separator sludge systems, but no reduced efficiency of microbial
consumption. The BRE separator fed culture shows slower growths on solid
media, no growth on oil droplets in solution, one dominate ciliate type and
different fluorescent pseuomonads than initially identified, and two to
three fungi types. The BRE separator culture shows no actinomycetes present.
268
-------
The PLS culture grows faster on solid media, demonstrates the growth on oil
droplet surfaces, has three different kinds of ciliates all of which are
different from the BRE population, different fluorescent pseudomonads, and
the same single fungal population. These populations are being freeze dried
and stored for further investigation at a later date, when the evolution of
the culture to a steady state population can be completely investigated from
an ecological standpoint.
From these observations, the cultures are exhibiting a dynamic state of
population diversity. The overall reactor performance in fterms of waste
treatment efficiency has not been significantly afffected, however. The
implications are that the continuous loading reactors need to be continued
until a stable population develops in terms of both numbers and species
diversity, and reactor preformance monitored to detect any effects on process
performance associated with these changes. It also is evident that waste
type will influence the species composition. This logically would be antici-
pated. The extent of such changes need to be considered and evaluated,
especially for those wastes exhibiting variable composition.
The results of the endeavor to explore some of the possible substrates
available to the microbial populations using solid media are presented in
Table 2. These results reflect only a portion of the microbial populations'
capabilities since not all organisms identified in the cultures would be
expected to survive on pour plate culture techniques. The results indicate
that metabolically, both cultures possess the ability to metabolize similar
substrates. These substrates include the standard nutrient agar, tryptic
'soy agar and dextrose for comparison. As is evident in Table 2, a wide
variety of substrates are available as a sole carbon source including the
cyclic compounds benzene, toluene and xylenes, the long chain hydrocarbon
dotriacontane (C-32), and methylene chloride. No growth was attainable for
pentachlorophenol or para-cresol using solid media at concentrations of 800
and 1000 mg/L, respectively. Growth was attained at 50 mg/L pentachloro-
phenol on solid media. It is evident that the higher concentrations are
toxic to the populations on solid media. Separate liquid cultures fed the
pentachlorophenol and para-cresol at 800 mg/L and 1000 mg/L, respectively,
were capable of supporting growth. Monitoring of the pentachlorophenol
liquid culture chloride concentration with time showed 60 percent of the
pentachlorophenol was dechlorinated within five days, verifying substrate
consumption. Chloride production was used to verify consumption of methylene
chloride in liquid culture also.
Reactor performance in terms of polycyclic aromatic hydrocarbon (PAH)
compounds destroyed for the BRE sludge reactor are presented in Table 3.
Average PAH reduction was approximately 90 percent with the reactor operating
on a 5 day hydraulic and mean cell residence time. The more readily degrad-
able PAH's (napthalene, fluorene, phenanthrene and anthracene) were 85 to
99 percent removed in the reactor. The more difficult substrates (pyrene,
benzo (a) anthracene, chrysene) were removed with 75 to 86 percent effective-
ness. These removal efficiencies were achieved with light loading rates
(approximately 130 mg/kg PAH in feed sludge) but are significant in that this
was achieved with an abundance of more readily available substrates present.
269
-------
TABLE 2. MICROORGANISM GROWTH ON SELECTED ORGANIC CARBON SOURCES
FOR BRE AND PLS CONTINUINS CULTURES
Carb'on
Source
Concentration
mg/1
Agar
Media*
Liquid
Media*
Nutrient Agar
Tryptic Soy Agar
Dextrose
2-2-4 Trimethyl pentane
Benzene
Toluene
Methylene Chloride
Mixed Xylenes
Pentachlorophenol
Pentachlorophenol
P-Cresol
Succinic Acid
Dotriacontane
BRE Separator Oil
PLS Separator Oil
500
690
700
870
660
860
800
50
1000
1250
1250
525
525
(+) Growth attained, (-) no growth
270
-------
TABLE 3. PAH REMOVAL EFFICIENCIES IN CONTINUOUS
LOADED REACTOR
PAH
Percent Removed
Napthalene
Fluorene
Phenanthrene
Anthracene
Fluoranthrene
Pyrene
Benzo(a)anthracene
Chrysene
Benzo(a)fluoranthene
Benzo(a)pyrene
Benzo(g,h,i)perylene
99
96
86
85
71
75
86
82
85
87
97
Total oil content of the feed was 100 fold higher (approximately 16,000 mg/
kg). Increased loading rates and biokinetic parameter measurements are being
made to examine for maximum process efficiencies for PAH destruction.
CONCLUSIONS
The results obtained thus far on the biological treatability of petro-
leum and petrochemical wastes are very encouraging. The growth rates
obtained on API separator sludge oil are well within the requirements for
efficient process design for reducing oil content of the sludges. Substrate
utilization at low concentrations indicates that process design can achieve
quality effluent and waste solid streams. Further work needs to be expended
on the fate of the principal hazardous constituents present in the sludges.
Additional work on complex substrate availability, substrate destruction,
kinetics of reactions for process design and scale up, and environmental
growth limitations need to be addressed. Microbial stability appears not to
be a concern for a specific waste stream. Continuous loading experiments
demonstrate an ability to maintain an effective population over time.
271
-------
Population composition has been observed to change with different waste
sources; this factor needs to be addressed further to determine the potential
implications for those waste streams that have temporally variable composi-
tion.
ACKNOWLEDGEMENTS
This work was partially supported by the LSU Hazardous Waste Research
Center. Lee Forbes and Robert Marks, graduate assistants in civil engi-
neering, are gratefully acknowledged for their data collection efforts on
which most of this paper is based.
REFERENCES
1. Marks, R. E., Field, S. D., Wojtanowicz, A. K., "Biodegradation of
Oilfield Production Pit Sludges," Proceedings of the 42nd Industrial
Waste Conference, Lewis Pub. Inc., Chelsea, MI, (1988).
2. Marks, R. E., Field, S. D., Wojtanowicz, A. K. "Oil Reduction in Spent
Drilling Muds by Biotreatment," paper presented at the Third National
Conference on Drilling Muds, Norman, OK, May, (1987).
3. Marks, R. E., Field, S. D., Wojtanowicz, A. K., "Biodegradation of
Oil-Based Drilling Muds and Production Pit Sludges," Journal Energy
Resources Technology, 110, 183-188, (1988).
4. Field, S. D., Forbes, L., Marks, R. E., Wojtanowicz, A. K. "Biological
Treatment Oilfield and Petrochemical Hazarodus Wastes," Proceedings of
the Engineering Foundation Conference on Biotechnology Applications in
Hazardous Waste Treatment, Longboat Key, FL, Oct. 1988 (in press).
5. Forbes, L. "Biodegradation of Petroleum and Petrochemical Wastes," MS
Thesis, Dept. Civil Engineering, Louisiana State University, Baton
Rouge, LA, 88 p., Dec. 1988.
6. Braha, A., Hafner, F., "Use of Lab Batch Reactors to Model Kinetics,"
Water Research, 73-81, (1987).
7. Gaudy, A. F. Jr., Yang, P,. Y., Bustamante, R., Gaudy, E. T.,
"Exponential Growth in Systems Limited by Substrate Concentration,"
Biotechnology Bioengineering, 15, 589-596, (1973).
8. Medcalf and Eddy, Inc., "Wastewater Engineering: Treatment, Disposal,
Reuse," 2nd Edition, McGraw Hill, NY, (1979).
9. Kincannon, D. F., Stover, E. L., "Determination of Activated Sludge
Biokinetic Constants for Chemical and Plastic Industrial Wastewater,"
EPA-600/2-83-073, USEPA, Ada, OK, (1984).
272
-------
THE DETERMINATION OF BIODEGRADABILITY AND BIODEGRADATION KINETICS OF
ORGANIC POLLUTANT COMPOUNDS WITH THE USE OF ELECTROLYTIC RESPIROMETRY
Henry H. Tabak
U.S. Environmental Protection Agency
Office of Research and Development
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
and
Sanjay Desai and Rakesh Govind
Department of Chemical and Nuclear Engineering
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
Electrolytic respirometry involving natural sewage, sludge and soil
microbiota is being applied to the fate studies of priority pollutant and
RCRA toxic organics to generate data on their biodegradability and on
biodegradation/inhibition kinetics. This paper discusses the experimental
design and procedural steps for the respirometry biodegradation and toxicity
testing approach for individual organics or specific industrial wastes. The
discussion also includes a review of the electrolysis BOD measuring system
inherent in electrolytic respirometry and the factors affecting
respirometric determination and measurement of respiration rate.
A developed multi-level protocol is presented for determination of the
biodegradability, microbial acclimation to toxic substrates and first order
kinetic parameters of biodegradation (n and n') and for estimation of the
Monod kinetic parameter (JL, Ks and Y ) of toxic organic compounds, in order
to correlate the extent and rate of biodegradation of these organics with a
predictive model based on chemical properties and structure of these
compounds.
Respirometric biodegradability/inhibition and biodegradation kinetic
data are provided for representative RCRA alky! benzenes, phenolic
compounds, phthalate esters, and ketones.
273
-------
INTRODUCTION
Electrolytic respirometry is attaining prominence in biodegradation
studies and is becoming one of the more suitable experimental methods for
measuring the biodegradability and the kinetics of biodegradation of toxic
organic compounds by the sewage, sludge and soil microbiota and for
determining substrate inhibitory effects to microorganisms in wastewater
treatment systems.
Biodegradation of toxic and hazardous organic compounds holds a great
promise as an important fate mechanism in wastewater treatment and in soil
detoxification. Information about the extent and rate of biodegradation is
a prerequisite for informed decision making on the applicability of the
biodegradation approach. Unfortunately, relatively little quantitative data
are available from which engineering judgement can be made, because of the
large effort required to assess biodegradation kinetics.
Current research in our laboratories has shown that it is possible to
assess biodegradation kinetic parameters from oxygen uptake data, obtained
through the use of electrolytic respirometry. This methodology greatly
reduces the work and expense involved in evaluation of biodegradation
kinetics. The ongoing biodegradation studies are concerned with the
generation of biokinetic database so that it can be ultimately used to
establish a possible correlation between molecular substrate configuration
(chemical/physical characteristics) and biomass activity (kinetic
parameters) as an index of biodegradation. The experimental respirometry
testing is also providing data on the concentration levels of toxic organics
inhibitory to microbial activity.
Initially, the inter-laboratory, ring test, Organization of Economic
'Cooperation and Development (OECD) studies at the EPA laboratory,
Cincinnati, Ohio, were undertaken to develop confirmatory respirometric
biodegradability testing procedure. Respirometric biodegradability and
biokinetic data are provided for the selected non-inhibitory and non-
adsorbing compounds, tetrahydrofuran, hexamine, pentaerythritol, 1-napthol,
sodium benzene sulphinate, thioglycolic acid and the biodegradable reference
compound, aniline.
Subsequently, similar electrolytic respirometry studies were initiated
to determine biodegradation kinetic parameters for selected representative
toxic compounds of varied classes of organics included in the Priority
Pollutant, RCRA list, and to demonstrate presence of any inhibitory effects
of these organics of specified concentration levels on the sludge biomass
and on the metabolism of biogenic compounds.
The objectives of the present study were to utilize the electrolytic
respirometry oxygen uptake data to: (1) determine the biodegradability of
selected RCRA alkyl benzenes, phenols, phthalates, and ketones, (2) generate
information on their acclimation time (lag) values (t0) and the initiation
and termination time values for the declining growth phase (t^ and t2); (3)
determine their first order kinetic parameters of biodegradation (specific
growth rate constants for the exponential growth phase (/*) and for the
declining growth phase (#'); (4) estimate the Monod kinetic parameters (Mm,
274
-------
Ks and Yg) of these compounds without initial growth or growth yield
assumptions; (5) demonstrate presence of any inhibitory effects of these
compounds on the metabolism of biodegradable reference compound, aniline;
and (6) to correlate the extent and rate of biodegradation of these
compounds with a predictive model based on chemical properties and structure
of these compounds.
The purpose of this study was to obtain information on biological
treatability of the benzene, phenol, phthalate and ketone organics in
wastewater treatment systems which will support development of an EPA
technical guidance document on the discharge of the above organics to POTWs.
Respirometric biodegradability, biokinetic and inhibition data are provided
for the selected RCRA benzene, phenol, phthalate and ketone compounds.
BACKGROUND
MEASUREMENT OF OXYGEN CONSUMPTION
Measurement of oxygen consumption is one of the oldest means of
assessing biodegradability. Time consuming manual measurement of oxygen
uptake (dilution BOD measurements) was replaced gradually by a more direct
and continuous respirometric method for measurement of oxygen consumption in
biochemical reactions, for use in routine examination of sewage and in
control of sewage treatment process.
A rather comprehensive review of the use of respirometers for the study
of sewage and industrial waste and their application to water pollution
problems was published by Jenkins in 1960 (21). Montgomery's (29) review on
respirometric methods summarized the design and application of respirometers
for determination of BOD.
The application of respirometry was gradually directed to research
studies to assess the toxicity and bipdegradation of specific wastes or
compounds, to evaluate factors affecting biological growth and to provide an
insight into nitrification reaction. Of the commercial respirometers which
have been developed for respirometric studies, the electrolytic
respirometers were shown to be most applicable for measurement and
quantitation of biodegradation activity because they automatically produce
oxygen as needed, thereby eliminating some of the limitations of other
techniques and allowing output data to be collected automatically for direct
recording and processing (3, 7, 26, 30, 53, 56, 57). A recent detailed
review of respirometric techniques and their application to assess
biodegradability and toxicity of organic pollutants was published by King
and Dutka (23).
RESPIROMETRIC BIODEGRADABILITY TESTING
Most uses of electrolytic respirometry in biodegradability testing have
been for screening purposes to measure the extent of biodegradation as a
percentage of the theoretical oxygen demand exerted in some time period (3,
6, 7, 17, 27, 50, 52, 54, 55). A more recent study by Painter and King (38)
concluded that a procedure based on electrolytic respirometry was reliable
275
-------
for assessing biodegradability, and could serve as an adequate Level I
screening test for biodegradability (33).
A considerable amount of studies using electrolytic respirometry to
determine the biodegradability of wastes and specific organics is available
in published literature and significant data on biodegradation of pollutants
based on oxygen uptake have been generated (1, 15, 16, 20, 24-26, 28-32, 37,
40, 44, 48, 49, 51, 58).
There are many techniques that have been used to evaluate
biodegradation kinetics and these were reviewed in detail by Howard et al.
(18, 19) and Grady (11)- These techniques utilize continuous, fed-batch and
batch type reactors for providing data from which kinetic parameters can be
evaluated. The use of batch systems in biotechnology and biological
wastewater treatment represents a less labor intensive, less expensive and
much faster way to model biokinetics.
The kinetic parameters obtained by the above techniques should be
intrinsic, that is, dependent only on the nature of the compound and the
degrading microbial community and not on reactor system used for data
collection. If this condition is satisfied then the parameters obtained can
be used for any reactor configuration and can be used.in mathematical models
to estimate the fate of toxic organics.
Batch techniques are successful in obtaining intrinsic kinetic
parameters by applying non-linear curve fitting techniques to single batch
substrate removal curves, provided initial conditions are selected with
proper care (Simkins and Alexander (42, 43); Robinson and Tiedje (41); Cech
et al. (5); Braha and Hafner (2)]. Batch systems can be used with either
acclimated or unacclimated biomass for providing kinetic data and require
that samples be taken at discrete time intervals during the course of
biodegradation [Tabak et al. (45); Larson and Perry (25); Paris and Rogers
(39)].
Measurement of oxygen consumption through electrolytic respirometry is
a batch type technique which has been shown to be very promising for
automating data collection associated with biodegradation and intrinsic
kinetic parameters.
USE OF ELECTROLYTIC RESPIROMETRY TO GENERATE BIOKINETIC DATA
Whereas the electrolytic respirometry is becoming the most commonly
employed method for automatically collected data associated with
biodegradation, with the exception of the studies performed by Dojlido (6),
Larson and Perry (25), Tabak et al. (46), Oshima et al. (35, 36), Gaudy et
al. (9, 10), Grady et al. (12-14) there have been very few investigations
into the use of respirometry to generate biodegradation kinetic data.
Larson and Perry (25) showed that the electrolytic respirometer can be used
to measure biodegradation of complex organics in natural waters when
specific analytical methods or radio!abeled materials are unavailable.
However, they used empirical kinetic expressions which were system specific.
Dojlido (6) divided the oxygen uptake curve into seven different phases
and then proposed an empirical model for each phase and evaluated the
276
-------
faiodegradability and toxicity of a test compound by measuring empirical rate
constants and time interval associated with each phase. Various phases were
distinguished by identifying inflection points in the curve through the
plots of the logarithm of the slope versus time.
Tabak et al. (46) and Oshima et al. (36) sought to capitalize on
Dojlido's method of identifying inflection points in order to quantify more
fundamental rate coefficients. In order to identify more fundamental (and
therefore intrinsic) kinetic coefficients they used the generalized concept
of oxygen uptake by Gaudy and Gaudy (8). Substrate removal was divided into
exponential and declining phases separated by an inflection point, and the
endogenous phase. They coupled substrate removal and cell growth to oxygen
consumption by imposition of an electron balance and consequently were able
to evaluate um from the oxygen consumption data up to the inflection point.
They proposed the use of the lag time as an indicator of how difficult it is
to achieve acclimation to a test compound and their respirometric studies
were carried out with an unacclimated biomass.
The justification for using respirometry to obtain intrinsic kinetic
lies in the concept of oxygen consumption as an energy balance [Busch et al.
(4); Gaudy and Gaudy (8)]. This concept states that all of the electrons
available in a substrate undergoing biodegradation must either be
transferred to the terminal acceptor or be incorporated into new biomass or
soluble microbial products. If the concentrations of the substrate, the
products and the biomass are all expressed in units of chemical demand
(COD), then the oxygen uptake can be calculated in a batch reactor from an
equation relating oxygen uptake to substrate, biomass and soluble products
[Grady et al. (12)]. Furthermore, since biomass growth and product
formation are proportional to substrate removal, this suggests that an
oxygen uptake curve can provide the same information as either a substrate
removal curve or a biomass growth curve. This latter concept has recently
been used by Gaudy et al. (9, 10) to calculate biodegradation kinetics.
Specific growth rates obtained from growth studies as slopes of plots of
Ln(X) versus time at different substrate concentrations, compared favorably
with those obtained from exponential phase of respirometric oxygen uptake
curves as slopes of plots of Ln (doydt) versus time.
Studies of Grady et al. (12, 13) have demonstrated that it is possible
to determine intrinsic kinetics of single organic compounds by using only
measurements of oxygen consumption in respirometric batch reactors. With
the use of computer simulation techniques and non-linear curve fitting
methods, intrinsic kinetic parameters were obtained from oxygen consumption
data and were shown to be in agreement with those obtained from traditional
measurement of substrate removal (DOC, SCOD, 14C) or cell growth.
MATERIALS AND METHODS
EXPERIMENTAL SETUP
The electrolytic respirometry studies were conducted using an automated
continuous oxygen uptake and BOD measuring Voith Sapromat B-12 (12 unit
system) electrolytic respirometer-analyzer. The instrument consists of a
temperature controlled waterbath, containing measuring units* a recorder for
277
-------
digital indication and direct plotting of the decomposition velocity curves
of organic compounds; and a cooling unit for the conditioning and continuous
recirculation of waterbath volume. The recorder shows the digital
indication of oxygen uptake and constructs a graph for these values of each
measuring unit. The cooling unit constantly recirculates water to maintain
constant temperature in waterbath. Each measuring unit as shown in Figure 1
is comprised of a reaction vessel with a carbon dioxide absorber mounted in
a glass joint flask stopper, an oxygen generator and a pressure indicator.
This measuring unit is interconnected by hoses, forming an air sealed
system, so that the atmospheric pressure fluctuations do not adversely
affect the results.
The activity of the microorganisms in the sample creates a vacuum which
is recorded by the pressure indicator, which triggers the oxygen generator.
The pressure conditions are balanced by electrolytic oxygen generation. The
quantity of the sample, the amperage for the electrolysis and the speed of
the synchronous motor are so adjusted that, with a sample of 250 ml, the
digital counter indicates the oxygen uptake directly in mg/L. The C02
generated is absorbed by soda lime. The nitrogen/oxygen ratio in the gas
phase above the sample is maintained throughout the experiment and there is
no depletion of oxygen. The recorder-plotter concomitantly constructs an
oxygen uptake graph for the selected values. The oxygen generators of the
individual measuring units are electrolytic cells which supply the required
amount of oxygen by electrolytic dissociation of a copper sulfate solution
combined with sulfuric acid.
The nutrient solution used in these studies was an OECD synthetic
medium (33, 34) consisting of measured amounts per liter of deionized
distilled water of (1) mineral salts solution; (2) trace salts solution; and
(3) a solution (150 mg/L) of yeast extract as a substitute for vitamin
solution.
The microbial inoculum was an activated sludge from The Little Miami
wastewater treatment plant in Cincinnati, Ohio, receiving municipal
wastewater. The activated sludge sample was aerated for 24 hours before use
to bring it to an endogenous phase. The sludge biomass was added to the
medium at a concentration of 30 mg/L total solids. Total volumes of the
synthetic medium in the 500 mL capacity reactor vessels were brought up to a
final volume of 250 mL.
The test and control compound concentrations in the media were 100
mg/L. Aniline was used as the biodegradable reference compound, at a
concentration of 100 mg/L.
The typical experimental system consisted of duplicate flasks for the
reference substance aniline and the test compounds, a single flask for the
physical-chemical test (compound control), a single flask for toxicity
control (test compound plus aniline at 100 mg/L each) and an inoculum
control. The contents of the reaction vessels were preliminarily stirred
for an hour to ensure endogenous respiration state at the initiation of
oxygen uptake measurements. Then the test compounds and aniline were added
to it. The reaction vessels were then incubated at 25°C in the dark
enclosed in the temperature controlled waterbath and stirred continuously
throughout the run. The microbiota of the activated sludge used as an
278
-------
Inoculum were not preacclimated to the substrates. The incubation period of
the experimental run was between 28 to 50 days. A more comprehensive
description of the procedural steps of the respirometric tests is presented
elsewhere (33, 46).
For fully automatic data acquisition, frequent recording and storage of
large numbers of oxygen uptake data, the Sapromat B-12 recorders are
interfaced to an IBM-AT computer via Metrabyte interface system. The use of
Laboratory Handbook software package allows the collection of data at 15
minute intervals.
DETERMINATION OF SUBSTRATE BIODEGRADABILITY FROM OXYGEN UPTAKE DATA
In this study, biodegradation was measured by two approaches: the
first, as the ratio of the measured BOD values in mg/L (oxygen uptake values
of test compound minus inoculum control - endogenous oxygen uptake values)
to the theoretical oxygen demand (ThOD) of substrate as a percent; the
second as a percentage of the test compound as measured by dissolved organic
carbon (DOC) changes (OECD Guidelines for Testing of Chemicals) (34).
Graphical representation of percent biodegradation based on the
BOD/ThOD ratio were developed against time for each test compound. The
experimental DOC data for the initial samples and samples for reaction
flasks collected at the end of experimental run were used to calculate the
percent biodegradation based on the percent of DOC removal in the culture
system.
DETERMINATION OF KINETIC PARAMETERS OF BIODEGRADATION
The first order kinetic rate constants were determined by the
linearization of the BOD curves (oxygen uptake of test compound minus oxygen
uptake of inoculum control), which gives straight lines expressing the
exponential and declining endogenous phases of the BOD curve as shown in
Figure 2. The slope of the Ln(6 oxygen uptake/St) versus t as described by
Dojlido (6), Tabak et al. (46), Oshima et al. (35, 36), and Tabak et al.
(47). give specific rate constants of the exponential growth phase (/* values)
and the declining growth phase (#' values) of the BOD curve.
Acclimation time values (t0) and the time values for the initiation and
termination of the declining growth phase (t1 and t2) for each test compound
were determined from linearized expressions of BOD curves.
The estimations of the Monod Kinetic parameters, maximum specific
growth rate constant, 0m, half saturation constant, Ks and growth yield
constant, Y were determined directly from experimental oxygen uptake curves
without the consideration of initial growth and growth yield assumption
[Jobbagy et al. (22); Tabak et al. (47)].
A. Determination of Yg Constant
Y - the true yield parameter or the ratio of growth of biomass to
substrate utilization, can be obtained for the experimental oxygen uptake
curve at the initiation of the plateau of the curve as shown in Figure 3. A
vertical line is drawn at the point of intersection of the tagents of the
279
-------
exponential and plateau phases of the curve. The oxygen uptake value
obtained at the point of intersection of the vertical line and oxygen uptake
curve is the Oupt value - oxygen uptake value at the initiation of plateau.
The Oypt value is then plugged into the equation Y = (S_-0 t/S0)-Y
(where S0 is initial concentration of substrate and Y is soluble product
concentration formed divided by initial substrate concentration) for Y_
determination. 9
B. Determination of um Constant
l^ - the maximum specific growth rate can be determined from
experimental oxygen uptake curve plot in the following manner:
(1) Values of the change of Ou with time (60u/6t) or slopes are determined
along the entire experimental oxygen uptake curve as shown in Figure 4.
(2) These 60u/6t (slope) values are then plotted against the cumulative 0
values for each time interval, as shown in Figure 5.
(3) The slope of the developed linearized form of oxygen uptake curve is
the estimated /tm value.
C.
Determination of K. Constant
Ks - the half saturation constant or the substrate concentration at
which the specific growth rate is 1/2 the maximum specific growth rate can
be obtained from the experimental oxygen uptake curve in the following
manner:
Value of 0 t can be calculated from the plot of (50u/5t) versus Ou
provided the Ks value is 1 or less (insignificant in comparison to S0
value) and the plot contains a linear section with the slope p, as
shown in Figure 5.
Other (SOJSt) versus 0 plot in which the slope deviates from nm
because of larger Ks values (more significant in comparison to S ) is
illustrated in Figure 6.
The value of 60 /6t is determined at the intercept of the straight line
developed from the plot of 60u/5t versus Ou (Figure 5) which contains a
linear section with slope 0T.
Beginning with the value of 1/2 the intercept value, another straight
line (b) is constructed with the slope 1/2 that of the slope of
original line (a) whose slope is /im.
At the point where line (b) intercepts the declining experimental curve
of the plot, a vertical line from that point of interception can
provide the value of Out on the x axis.
(6) This Out value is then used in the determination of K, with the use of
(1)
(2)
(3)
(4)
(5)
the equation
- (Out/(l-Y-YJ) =
p gy
Where S0 = initial substrate concentration
St = substrate concentration at time t
Yp - soluble product concentration formed (i.e. intermediate
metabolites) divided by the initial substrate concentration
280
-------
(7) When the Out, Ya, YD, and S0 values are plugged into the equation, value
of St can be calculated - which is the value of Ks (in systems where Ks
value is 1 or less).
RESULTS AND DISCUSSION
Respirometric biodegradability, biokinetic and Monod kinetic data for
selected RCRA alkyl benzenes, phenols, phthalates and ketones are reported
in this paper. The electrolytic respirometry oxygen uptake data for the
test compounds, the control reference compound aniline, the inhibition and
endogenous control systems were generated revealing the lag phase
(acclimation phase), the biodegradation (exponential) phase, the different
bio-reaction rate slopes (characteristic of the test compound) as well as
the plateau region at which the biooxidation rate reaches that of the
endogenous rate of microbial activity. Figure 7 illustrates a
representative oxygen uptake curve for aniline and the endogenous controls.
Figure 8 shows the replicate pentaerythritol oxygen uptake curves and the
toxicity control (pentaerythritol plus aniline) curve. Figure 9 illustrates
a representative graphical treatment of the percent biodegradation of
pentaerythritol with time, which was developed for each test compound (OECD
studies).
Based on the biokinetic equations relating growth rate of microbiota in
presence of above compounds, the substrate utilization rate, and rate of
oxygen uptake (BOD) curves, specific growth rate kinetic parameters
(biodegradation rate constants) were derived as slope values of the
linearized plots (plots of the log of DO/dt) of exponential and declining
growth phases of the BOD curve. The acclimation time values (t0), and time
values for the initiation and the termination of the declining growth phases
(tj and t2) for the test compounds and aniline were also generated.
The estimations of the Monod kinetic parameters for benzene, phenol,
phthalate, and ketone compounds reported here, were determined directly from
experimental oxygen uptake curves without the consideration of initial
growth and growth yield assumption.
RESPIROMETRIC STUDIES WITH SELECTED RCRA ALKYL BENZENE COMPOUNDS
The biodegradation of benzene, toluene, ethyl benzene, m- and
p-xylenes, tert-butyl benzene, sec-butyl benzene, butyl benzene, cumene, 1-
phenyl benzene and the reference compound, aniline at 100 mg/L concentration
by 30 mg/L sludge biomass (as measured by oxygen consumption by sludge
microbiota in mg 02/L) was followed over a period of 20 days. The
electrolytic respirometry oxygen uptake and BOD curves were generated and
graphical treatment of the percent biodegradation was established for each
compound. Figure 10 demonstrates typical oxygen uptake and BOD curves for
p-xylene and p-xylene + aniline and Figure 11 illustrates graphically the %
biodegradation of p-xylene with time.
The percent biodegradation data based on the BOD/ThOD ratios for
benzene, toluene, ethyl benzene, m- and p-xylene and the reference compound,
aniline, are summarized in Table 1. All of the above alkyl benzene
compounds were shown to be biodegradable substrates at concentration levels
281
-------
of 100 mg/L when exposed to 30 mg/L of activated sludge biomass under the
environmental conditions of the respirometric testing procedure, and within
the period of 20 days of incubation.
The toxicity test control flask respirometric data revealed no
inhibitory effects by these test compounds at the 100 mg/L concentration
levels on the bio-oxidation of aniline by sludge microbiota.
Table 2 summarizes the bio-kinetic data for the benzenes studied,
showing the specific growth rate constants for the exponential growth phase
0* values) and for the declining growth phase (/{' values) of the linearized
form of the BOD curves of these compounds, as well as the t0, t, and t?
kinetic parameters. Figure 12 shows a typical plot of Ln(60u/5t) vs. time
for toluene, from which the kinetic parameters were determined.
Table 3 summarizes the Monod kinetic parameter
these benzene compounds.
Ks, Y ) data for
9
RESPIROMETRIC STUDIES WITH SELECTED RCRA PHENOLIC COMPOUNDS
The biodegradation of phenol, resorcinol, o-, m- and p-cresols,
catechol, 2,4-dimethyl phenol and the reference compound aniline at 100 mg/L
concentration levels and exposed to 30 mg/L biomass was followed over a
period of 20 days.
All of the phenols were shown to be biodegradable substrates under the
conditions of the respirometric testing procedure. The toxicity test
control flask respirometric data revealed no inhibitory effects by these
compounds at the 100 mg/L levels on the biodegradation of aniline by the
sludge biomass.
Table 4 summarizes the bio-kinetic data for the phenols studied,
showing the specific growth rate constants as well as the t0, t,, and t2
kinetic parameters. Table 5 provides the Monod kinetic parameter data for
these phenolic compounds.
RESPIROMETRIC STUDIES WITH SELECTED RCRA PHTHALATE ESTER COMPOUNDS
Evaluation of the biodegradability and determining of bio-kinetics of
degradation of phthalate compounds, dimethyl phthalate, diethyl phthalate,
dipropyl phthalate and butyl benzyl phthalate was achieved with use of
respirometric oxygen uptake data.
All of the above phthalates were shown to biodegradable under the
conditions of the respirometric tests and were shown not to exhibit any
inhibitory effects at the 100 mg/L levels on aniline biodegradation by the
sludge microbiota.
Tables 6 and 7 summarize respectively the biokinetic (first order) and
Monod kinetic parameter data for the selected phthalate esters under study.
282
-------
RESPIROMETRIC STUDIES WITH SELECTED RCRA KETONE COMPOUNDS
Respirometric oxygen uptake data from the studies with the selected
ketone compounds, acetone, 2-butanone, 4-methyl-3-pentanone and a cyclic
ketone, isophorone were utilized to determine their biodegradability and
biodegradation kinetic parameters.
All the ketones were shown to be biodegradable at 100 mg/L
concentration levels in media containing 30 mg/L biomass and did not exhibit
any toxicity to aniline biodegradation at these concentrations.
Tables 8 and 9 summarize respectively the first order and Monod kinetic
parameter data for these ketones.
CONCLUSIONS
The experimental data of respirometric studies with several classes of
organic compounds definitely demonstrate that it is possible to measure the
biodegradability (percent biodegradation - as a ratio of BOD to ThOD) and to
determine the kinetics of degradation of single organic compounds by using
only measurements of oxygen consumption in respirometric batch reactors.
The values of the kinetic parameters determined from oxygen consumption data
were demonstrated to be similar to those based on the measurements of
substrate removal and those made with cell growth data.
The generated data on biodegradation, biodegradation rates and
substrate inhibition kinetics through the use of electrolytic respirometry,
will enable the classification of biodegradability of toxic priority
pollutant and RCRA toxic organic compounds and ultimate projection of the
fate of organic compounds of similar molecular structure to those
experimentally studied by way of the established predictive treatability
models based on structure-activity relationships.
With the electrolytic respirometry approach, data base of the removal
of the above compounds by biodegradation fate mechanism can be adequately
generated to support the development of predictive models on fate and
removal of toxics in industrial and municipal waste treatment systems. A
possible relationship between the kinetic parameters and the effect of
different factors on these parameters, as determined through electrolytic
respirometry and the structural properties of the organic pollutant, can
eventually facilitate prediction of the extent and the rate of
biodegradation of organic chemicals in the field of wastewater treatment
systems from the knowledge of the structural properties of the pollutant
organics.
A preliminary predictive biodegradation - structure/activity model
based on the group contribution approach was developed from the generated
biodegradation kinetic data (first order kinetic parameters) with the use of
electrolytic respirometry. It is expected that the model will closely
predict the results found experimentally. In this way, the fate of other
organic compounds may be anticipated without the time and expense of
experimental work.
283
-------
The electrolytic respirometry biodegradation studies will provide basic
pilot scale treatability information and data which will be used to confirm
methods to predict treatability and the need for pretreatment of
structurally related pollutants (e.g., by structure, anticipated
treatability properties, etc.). This study will thus provide a more
extensive list of pollutants than was covered by experimental data, for
consideration in guiding the Agency to predict the fate of such compounds
without costly experimental testing.
1.
2.
3.
4.
5.
REFERENCES
Arthur, R.M. Twenty years of respirometry. In:
Thirty-Ninth Annual Industrial Waste Conference.
West Lafayette, IN, 1984.
Proceedings of the
Purdue University,
Braha, A. and Hafner, F. Use of lab batch reactors to model
biokinetics. Water Res. 21(1): 73, 1987.
Bridie, A.L.A.M. Determination of Biochemical oxygen demand with
continuous recording of oxygen uptake. Water Res. 3: 157, 1969.
Busch, A.W., Grady, L., Jr., Rao, T.S., and Swilley, E.L. Short-term
total oxygen demand test. J. Water Pollut. Control Fed. 34: 354,
1962. ~
Cech, J.S., Chudoba, J., and Grau, P. Determination of kinetic
constants of activated sludge microorganisms. Water Sci. Technol.
259, Amsterdam, 1984.
17:
6. Dojlido, J.R. Investigation of biodegradability and toxicity of
organic compounds. EPA-600-2-79-163. U.S. Environmental Protection
Agency, Cincinnati, Ohio, 1979.
7. Fuhs, G.W. Some factors affecting biochemical oxygen demand as
determined in manometric or manostatic devices. Wasser Abwasser-
Forschung. 5: 161, 1968.
8. Gaudy, A.F., Jr. and Gaudy, E.T. Biological concepts for design and
operation of the activated sludge process. EPA-17090 FQJ. U.S.
Environmental Protection Agency, Cincinnati, Ohio, 1971.
9. Gaudy, A.F., Jr., Rozich, A.F., Garniewski, S., Moran, N.R., and
Ekambaram, A. Methodology for utilizing respirometric data to assess
biodegradation kinetics. Paper presented at the 42nd Annual Industrial
Waste Conference, Purdue University, West Lafayette, Indiana, 1987.
10. Gaudy, A.F., Jr., Ekambaram, A., and Rozich, A.F. A respirometric .
method for biokinetic characterization of toxic wastes. Paper
presented at the 43rd Annual Industrial Waste Conference, Purdue
University, West Lafayette, Indiana, 1988.
11. Grady, C.P.L., Jr. Biodegradation: Its Measurement and
Microbiological Basis. Biotechnol. Bioenaineer. 27: 660, 1985.
284
-------
12.
13.
14.
15.
16.
17.
18.
19.
20.
21,
22.
Grady, C.P.L., Jr., Dang, J.S., Harvey, D.M., Jobbagy, A., and Wang,
X.-L. Determination of biodegradation kinetics through use of
electrolytic respirometry. Water Sci. Techno!. 21: 957, Brighton,
1989.
Grady, C.P.L., Jr., Dang, J.S., Harvey, D.M., and Jobbagy, A.
Evaluation of biodegradation kinetics with respirometric data. Water
Pollut. Control Fed. 1989: (submitted).
Grady, C.P.L., Jr., Aichinger, G., Cooper, S.F., and Naziruddin, M.
Biodegradation kinetics for selected toxic/hazardous organic compounds.
Hazardous Waste Treatment: Biosystems for Pollution Control, AWMA,
1989: (in press).
Halbartschlarger, J., Kohler, H., Szwerinski, H., and Bardtke, D.
Investigations on the biological degradation of chlorinated
hydrocarbons using dichloromethane (methylene chloride) as an example.
Gwf-Wasser/Abwasser. 125(H.8): 380, 1984.
Hickey, C.W. and Nagels, J.W. Modifications to electrolytic
respirometer systems for precise determination of BOD exertion kinetics
in receiving waters. Water Res. 19: 463, 1985.
Howard, P.H. and Banerjee, S. Interpreting results from
biodegradability tests of chemicals in water and soil. Environment.
Toxicol. Chem. 3: 551, 1984.
Howard, P.H., Banerjee, S., and Rosenberg, A. A review and evaluation
of available techniques for determining persistence and routes of
degradation of chemical substances in the environment: An update of
the 1975 Report. EPA-560/5-81-011. U.S. Environmental Protection
Agency, Cincinnati, Ohio, National Technical Information Services No.
PB84-168731, 1981.
Howard, P.H., Saxena, J., Durkin, P.R., and Ou, L.-T. Review and
.evaluation of available techniques for determining persistence and
routes of degradation of chemical substances in the environment. EPA-
560/5-75-006. U.S. Environmental Protection Agency, Cincinnati, Ohio,
National Technical Information Service No. PB243825, 1975.
Huang, J.Y.C., Cheng, M.D., and Mueller, J.T. Oxygen uptake rates for
determining microbial activity and application. Water Res. 19(3):
373, 1985.
Jenkins, D. The use of manometric methods in the study of sewage and
trade wastes. In: P.C.G. Isaac (ed.), Waste Treatment. Pergamon
Press, Oxford, 1960. p. 99.
Jobbagy, A., Grady, C.P.L., and Tabak, H.H. Characterization of
biodegradation through respirometry: Graphical analysis and
theoretical considerations. Clemson University Research Report,
Clemson, South Carolina. Water Res. 1989: (to be submitted).
285
-------
23. King, E.F. and Dutka, B.I. Respirometric techniques, in: G. Britton
and B.J. Dutka (eds.), Toxicity Testing using Microorganisms. Vol. 1.
CRC Press, Inc. Boca Ruton, Florida, 1986. p. 76.
24. Klecka, G.M. Fate and effects of methylene chloride in activated
sludge. Appl. Environ. Microbiol. 44: 701, 1982.
25. Larson, R.J. and Perry, R.L. Use of the electrolytic respirometer to
measure biodegradation in natural waters. Water Res. 15: 697, 1981.
26. Liebman, H. and Offhaus, F. Volumetric BOD measurements with the help
of 'Sapromat" a new apparatus for determining 5-day BOD and toxicity.
Abwassertechnik. 17: 4, 1966.
27. Madden, M. and Tittlebaum, M. Oxygen uptake rates associated with
biological treatment of pentachlorophenol wastewater. J. Environ. Sci.
Health. A19(3): 321, 1984. ~
28. Manios, V. and Balis, C. Respirometry to determine optimum conditions
for the biodegradation of extracted olive press-cake. Soil Biol.
Biochem. 15(1): 75, 1983.
29. Montgomery, H.A.C. The determination of biochemical oxygen demand by
respirometric methods. Water Res. 1: 631, 1967.
30.
31.
32.
33.
34.
35.
Montgomery, H.A.C., Oaten, A.B., and Gardiner, O.K. An automatic
electrolytic respirometer--Its construction and use. Effluent Water
Treatment J. 11: 23, 1971. ~ ~~~
Nochi, K. Oxygen consumption due to decomposition of chemical
substances. J. Water Waste. 22(11): 1285, 1980.
Nochi, K. Oxygen consumption due to decomposition of chemical
substances. J. Water Waste. 26(7): 751, 1984.
OECD, "OECD Guidelines for Testing of Chemicals", Section 3,
Degradation and Accumulation, Method 301C, Ready Biodegradability:
Modified MITI Test (I) adopted May 12, 1981 and Method 302C Inherent
Biodegradability: Modified MITI Test (II), adopted May 12, 1981,
Director of Information, OECD, Paris, France, 1981.
OECD, "OECD Guidelines for Testing of Chemicals", EEC Directive 79/831,
Annex V, Part C: Methods for Determination of Ecotoxicity. 5.2
Degradation. Biotic Degradation. Manometric Respirometry. Method
DGX1, Revision 5, 1983. p. 1-22.
Oshima, A., Tabak, H.H., and Lewis, R.F. Electrolytic respirometry
biodegradation study - OECD ring test biodegradation method evaluation
and development. EPA Draft Report, MERL. U.S. Environmental
Protection Agency, Cincinnati, Ohio, 1985.
286
-------
36. Oshima, A., Tabak, H.H., and Lewis, R.F. The evaluation of biological
treatability and removability of toxic organic chemicals by
respirometry. EPA Draft Manuscript, MERL. U.S. Environmental
Protection Agency, Cincinnati, Ohio, 1985.
37. Pagga, U. and Gunthner, W. Biodegradation and toxicity studies with
microorganisms: A comparison of laboratory tests with a treatment
plant model system. In: Proceedings of the International Symposium
on Principles for Interpretation of the Results of Testing Procedures
in Ecotoxicology. Valbonne, France, 1980.
38. Painter, H.A. and King, E.F. Environment and quality of life-ring test
programme 1983-84 - Assessment of biodegradability of chemicals in
water by manometric respirometry. Final Report, Contract No.
XI/W/83/238; Directorate-General Environment, Consumer Protection and
Nuclear Safety, Commission of the European Communities, Report No.
EUR9962EN, 1985.
39. Paris, D.F. and Rogers, J.E. Kinetic concepts for measuring microbial
rate constants: Effects of nutrients on constants. ADD!. Environ.
Microbiol. 51: 221, 1986.
40. Rigin, V.I., Golovin, Y.G., and Tyuneva, G.S. Determination of the
biochemical oxygen demand of natural waters and effluents with the
electrolytic generation of oxygen. Khimiva 1 Teknologiva Vod.v. 4(2):
180, 1982.
41. Robinson, J.A. and Tiedje, J.M. Nonlinear estimation of Monod growth
kinetic parameters from a single substrate depletion curve. Appl.
Environ. Microbiol. 45: 1453, 1983.
42. Simkins, S. and Alexander, M. Models for mineralization kinetics with
the variables of substrate concentration and population density. Appl.
Environ. Microbiol. 47: 1299, 1984.
43. Simkins, S. and Alexander, M. Nonlinear estimation of the parameters
of Monod kinetics that best describe mineralization of several
substrate concentrations by dissimilar bacterial densities. ADD!.
Environ. Microbiol. 50: 816, 1985.
44. Simpson, J.R. and Nellist, G.R. Development and use of a large-volume
automatic respirometer. Water Pollut. Contr. 69: 596, 1970.
45. Tabak, H.H., Quave, S.A., Mashni, C.I., and Barth, E.F.
Biodegradability studies with organic priority pollutant compounds.
JWPCF. 53(2): 1503, 1981.
46. Tabak, H.H., Lewis, R.F., and Oshima, A. Electrolytic respirometry
biodegradation studies, CEC/OECD ring test of respiration method of
determination of biodegradability, Ring Test Program 1984. EPA Draft
Final Report, MERL. U.S. Environmental Protection Agency, Cincinnati,
Ohio, 1984.
287
-------
47. Tabak, H.H., Desai, S., Govind, R., and Grady, C.P.L. Evaluation of
biodegradability and biodegradation kinetics of organic pollutant
compounds with the use of electrolytic respirometry. Presented at the
61st Annual Conference of Water Pollution Control Federation, Dallas,
Texas. October 2-6, 1988.
48. Therien, N., and Ilhan, F. Relating BOD5 with on-line oxygen uptake
rate measurements using automatic respirometers in view of process
monitoring and control. In: Procedures and Practices in Activated
Sludge Process Control. Vol. 3. 1982. p. 113.
49. Urano, K. and Kato, Z. Evaluation of biodegradation ranks of priority
organic compounds. J. Hazard. Materials. 13: 147, 1986.
50. Verstraete, W., Voets, J.P., and Vanlocke, R. Three-step measurement
by the Sapromat to evaluate the BOD5, the mineral imbalance and the
toxicity of water samples. Water Res. 8: 1077, 1974.
51. Wojnowska-Baryla, I. and Young, J.C. Measuring the effect of
biocatalytic additives on treatment process performance. J. WPCF.
55(11): 1373, 1983.
52. Yoshimura, K. and Masuda, F. Biodegradation of Sodium Alkyl
Poly(oxyalkylene)-sulfates. Am. Oil Chem. Soc. J. 59: 328, 1982.
53. Young, J.C., Garner, W., and Clark, J.W. An improved apparatus for
biochemical oxygen demand. Anal. Chem. 37(5): 784, 1965.
54. Young, J.C. and Baumann, E.R. Demonstration of the electrolysis method
for measuring BOD. Draft Final Report to the U.S. Environmental
Protection Agency from Iowa State University for Grant WP16020 DUN and
2-80036, 1972.
55. Young, J.C. and Affleck, S.B. Long-term biodegradability tests of
organic industrial wastes. In: Proceedings of the 29th Industrial
Waste Conference. Purdue University, Extension Series 145, West
Lafayette, Indiana, 1974. p. 154.
56. Young, J.C. and Baumann, E.R. The electrolytic respirometer-I.
Factors affecting oxygen uptake measurements. Water Res. 10: 1031,
1976.
57. Young, J.C. and Baumann, E.R. The electrolytic respirometer-II. Use
in water pollution control plant laboratories. Water Res. 10: 1141,
1976.
58. Young, J.C. Biochemical oxygen demand: Measurement and application.
Iowa State University, Ames, Iowa, Final Report, 1977.
288
-------
3 *-t a m
W V *i U
« H *• -rt
e Sirs
« g-s
O -rt C 3
•MUM
H U DIM
«. « >«4>
n a HM
tK*> U O Pi
(••rt
•* a
o a«au
N
«.-»
U
o
1
4JO tl C «l
M m J]*H 41 H
MM >• V
u »•» o « f<-a u
'ft • « O *« *1 r^ 4)
6 MU 0.41 41 U
•-i« n
-------
FIGURE 4 PLOT OF OXYGEN UPTAKE RATE
( Ou VERSUS TIME )
10 20 30 40 50 .60 70
UM-0.1. KS-1. YC-0.4. XO-5. SO-100
FIGURE 5 Ptot showing (AQu/At) versus Out curve containing a
ifnoAP Oa/*tu*vn tAjHh a *tts**>ui.» 4* WA UMA^ L _._&* A.'
6.0
4.5
4.0
as
£3.0
g2£
02.0
15
to
0£
0.0
linear section with a slopes mto be used In estimating
the values of K Jn cases whore K_values are nealioible
In comparison t
-------
500
en
10 20 30 40
ELAPSED TIME (Days)
FIGURE 7 OXYGEN UPTAKE DATA ON ANILINE
500
400
« 300
z
a
u
U)
— — — PenlMqrthrilol A .
—— Ptntxrirthfitol B .
PentMrythrilol It Aniline
200
100 •
"TO 20 30 40
ELAPSED TIME <0ayo>
FIGURE 8 BIOLOGICAL OXYGEN UPTAKE CURVE
r'°U (Run 1 Sample No. 10-12)
100
a
a
m
Paitaajthritol A
— — — PenlMnrthritol 0
ELAPSED TIME (days)
FIGURE 9 BIODEGRATION (% BOD REMOVAL) CURVE
( PENTAERYTHRITOL )
291
-------
1. Inoculum Control
2. p-Xjtcn.
3. p-Xylm BOD
4. p-X;l
-------
TABLE 1. SUMMARY OF RESPIROMETRIC BIODEGRADATION DATA FOR SELECTED BENZENES
PERCENT BIODEGRADATION (BASED ON % BOD REMOVAL)
Time
(days) Aniline
0.0 0.0
1.0 2.58
2.0 3.43
3.0 3.94
4.0 14.95
5.0 102.0
6.0 102.0
7.0 102.0
8.0 102.0
9.0 102.0
10.0 102.0
11.0 102.0
12.0 102.0
13.0 102.0
TABLE Z.
COMPOUNDS
Aniline
(Experiment 1)
(Experiment 2)
Benzene
Ethyl benzene
Toluene
p-Xylene
m-Xylene
tert-Butyl benzene
sec-Butyl benzene
Cumene
Butyl benzene
1-Phenyl hexane
Benzene
0.0
2.11
2.11
4.15
4.97
71.5
74.22
81.85
93.37
93.89
95.45
95.45
96.13
97.46
SUMMARY OF
ThOD
for 100
310
310
308
317
313
317
317
322
322
320
322
326
To! uene
0.0
2.78
8.81
85.11
89.04
94.34
97.92
100.0
101.18
103.48
103.48
103.48
103.48
103.48
BIO-KINETIC DATA
mg (days)
4.00
4.00
4.50
4.00
2.00
3.90
2.00
4.40
3.50
2.40
3.30
4.00
Ethyl benzene
0.0
2.27
2.27
2.99
3.78
4.29
40.25
73.91
73.78
83.94
. 88.86
88.86
89.65
91.00
FOR SELECTED
BENZENE
t; t2
(days) (days)
4.70
4.65
4.87
4.21
2.20
4.22
2.35
5.12
4.00
2.79
3.92
4.55
4.83
4.79
5.00
4.83
2.42
4.83
2.50
5.70
5.70
3.00
4.56
5.15
m-Xylene
0.0
1.35
2.14
69.68
69.68
76.68
82.58
84.67
87.03
87.79
90.12
90.15
90.97
91.80
COMPOUNDS
/I
(day-1)
2.78
3.80
8.57
8.33
8.75
9.94
6.60
1.21
0.78
2.31
2.42
1.85
p-xylene
0.0
1.58
1.58
1.58
9.4
70.38
82.68
95.58
99.49
101.76
104.7
105.78
106.7
108.6
*'
(day-1)
3.29
8.60
25.30
44.80
14.93
4.50
29.60
1.69
0.73
2.24
2.42
1.90
It - specific growth rate constant for exponential growth phase of BOD curve.
p' » specific growth rate constant for declining growth phase of BOD curve.
293
-------
TABLE 3. SUMMARY OF MONOD KINETIC PARAMETER DATA FOR SELECTED BENZENE COMPOUNDS
COMPOUNDS
Aniline
Benzene
Ethyl benzene
Tolouene
p-Xylene
•-Xylene
tert-Butyl benzene
sec-Butyl benzene
Cunene
Butyl benzene
1-Phenyl hexane
Lag Time
-------
TABLE 5. SUMMARY OF MONOD KINETIC PARAMETER DATA FOR SELECTED PHENOLIC COMPOUNDS
COMPOUNDS
Aniline
Phenol e
Resorcinol
p-Cresol
o-Cresol
m-Cresol
Catechol
2,4-dimethyl phenol
Lag Time
(t0)
days
4.00
1.00
1.50
1.00
1.20
1.44
0.85
2.00
Y9
mg biomass
mg substrate
0.38
0.58
0.48
0.33
0.41
0.46
0.49
0.39
Mm
(day-1)
6.15
9.82
12.22
6.11
4.10
7.97
12.80
5.62
mgs/l
6.10
9.43
6.31
27.78
16.41
17.62
43.87
14.07
H - maximum specific growth rate.
K, - half saturation constant; concentration of substrate at
Yg - growth yield, mg biomass formed/mg substrate coriiumed.
TABLE 6. SUMMARY OF BIO-KINETIC DATA FOR SELECTED PHTHALATE ESTER COMPOUNDS
COMPOUNDS
ThOD t0 t, t2 it v'
for 100 mg (days) (days) (days) (day-1) (day-1)
Aniline
Dimethyl phthalate
Diethyl phthalate
Dipropyl phthalate
Butyl benzyl phthalate
310
168
195
211
226
4.00
3.46
2.00
2.40
2.00
4.70
3.98
2.97
2.87
2.28
4.83
4.25
3.30
3.40
2.80
2.78
2.76
2.16
2.04
4.12
3.29
4.71
2.92
2.00
2.33
specific growth rate constant for exponential growth phase of BOD curve.
specific growth rate constant for declining growth phase of BOD curve.
295
-------
TABLE 7. SUMMARY OF MONOD KINETIC PARAMETER DATA FOR SELECTED PHTHALATE ESTER COMPOUNDS
COMPOUNDS Lag Time
(t0)
days
Aniline
Dimethyl phthalate
01 ethyl phthalate
Dlpropyl phthalate
Butyl benzyl phthalate
H * maximum specific
4.00 .
3.46
2.00
2.40
2.00
Y j^
ma biomass (day-1)
mg substrate
0.38 6.15
0.43 7.07
0.46 3.00
0.48 5.78
0.61 7.80
«g'/i
6.10
41.68
11.67
15.81
36.25
growth rate.
K, « half saturation constant;
Y, - growth yield, ng
biomass
concentration of substrate at 0j/2.
formed/ing substrate consumed.
TABLE 8. SUMMARY OF BIO-KINETIC DATA FOR SELECTED KETONE COMPOUNDS
COMPOUNDS ThOD t0 t, t, n n'
for 100 ng (days) (days) (days) (day-1) (day-1)
Aniline
Acetone
2-Butanone
4-Methyl -2-pentanone
Isophorone
310
221
244
272
278
4.00
3.70
2.00
1.85
22.30
4.70
3.99
2.20
2.24
23.70
4.83
4.18
2.35
2.35
25.40
2.78
2.45
2.41
2.31
0.73
3.29
3.98
4.98
4.80
0.38
p « specific growth rate constant for exponential growth phase of BOD curve.
It' • specific growth rate constant for declining growth phase of BOD curve.
TABLE 9. SUMMARY OF MONOD KINETIC PARAMETER DATA FOR SELECTED KETONE COMPOUNDS
COMPOUNDS
Aniline
Acetone
2-Butanone
4-Hethyl -2-pentanone
Isophorone
Lag Tine
(to)
days
4.00
3.70
2.00
1.85
22.30
ma biomass
ng substrate
0.38
0.36
0.39
0.45
0.43
(day-1)
6.15
4.86
5.11
6.40
1.57
*s
«>9/l
6.10
9.76
10.79
24.70
27.42
/i - naxlnua specific growth rate.
K, - half saturation constant; concentration of substrate at
Y, - growth yield, *g bloaass formed/eg substrate consulted.
296
-------
PREDICTION AND MODELING OF BIODEGRADATiON KINETICS OF
HAZARDOUS WASTE CONSTITUENTS
Rakesh Govind and Sanjay Desai
Dept. of Chemical & Nuclear Engineering
University of Cincinnati
Cincinnati, OH 45221
and
Henry H. Tabak
RREL, U S Environmental Protection Agency
Cincinnati, OH 45268
ABSTRACT
Biodegradation is the most important mechanism in controlling the concentration
of chemicals in an aquatic system because it can mineralize toxic pollutants to
innocuous forms. So the fate of organic chemicals in an aquatic environment, is
dependant on their susceptibility to biodegradation. Because of a large number of
chemicals, it will be expensive and labor intensive to gather this information in a
reasonable amount of time. Hence, there is need for a prediction method to obtain
these data.
Experiments using electrolytic respirometer, were conducted to collect oxygen
consumption data of some RCRA compounds and first order kinetic rate constants
were obtained. Using these data along with those from literature, a structure-
activity relationship was developed.
297
-------
INTRODUCTION
It has been estimated that 50,000 organic chemicals are commercially produced in
the United States and a large number of new organic chemicals are added to the
production each year (1). The presence of many of these chemicals in the environment
is a serious public health --•-•--- -•• • •• • .......
disposal techniques.
ilth problem. Their presence could be attributed to inadequate
Since many of these 'hazardous chemicals can be detected in wastewater, their
fate in wastewater treatment system is of great interest. Of the many factors that
affect the fate of these compounds, microbial degradation is probably the most
important (2). The high diversity of species and the metabolic efficiency of
microorganisms suggest that they play a major role in the ultimate degradation of
these chemicals (3). Biodegradation can eliminate hazardous compounds by
biotransforming them into innocuous forms, degrading them by mineralization to
carbon dioxide and water. So the information regarding the extent and the rate of
biodegradation of these chemicals is very important for regulation of their
manufacture and use. Due to a large number of these chemicals, gathering of this
information in a reasonable amount of time will be both expensive and labor intensive.
Thus, the only practical way out is to develop correlations and predictive techniques
to assess biodegradability (4). Lack of an adequate database on biodegradation kinetics
prevents the development of such techniques.
EXPERIMENTAL TECHNIQUES
There are many techniques for collection of biodegradation data and these are
reviewed in great detail by Howard, et al. (5) and Grady (6). Experimental techniques
for biodegradation data collection fall into tnree broad categories : continuous, fed-
batch and batch reactor systems.
Continuous culture reactors require an acclimated biomass. They also require long
transitional time intervals for reaching quasi steady state condition (7). So it is time
consuming, tedious and expensive. However, the analysis of data obtained is simpler
because the equations for continuous reactors, operating at steady state, reduce to
algebraic equations that are easily solved. This technique is used more for evaluating
parameters for the design of treatment systems rather than for biodegradation kinetics
(8). Because of the microbial competition, each continuous reactor will have a unique
microbial community associated with it (9); hence the evaluated kinetic parameters are
system specific and are not intrinsic.
Fed—batch reactors have also been used to estimate the kinetic parameters of
biodegradation. In these reactor configurations, a quantity of biomass is added into a
reactor and a substrate stream is continuously added in negligibly small amounts with
respect to the reactor volume. Because of small amounts of both substrate and
microorganisms, a pseudo steady state is achieved. So this technique reduces the time
required and also alleviates the problems associated with changes in the composition
of the microbial community. This configuration requires an acclimated biomass because
the microbes must be capable of responding instantaneously to the input of substrate.
However, this technique cannot be used to determine the Monod biokinetic parameters,
but it is an excellent procedure to determine the parameters for treatment plant
design or operation.
The use of batch cultures in biotechnology and biological wastewater treatment
298
-------
represents a less expensive and much faster way to model biokinetics in fermenters
and in activated sludge tanks (7). The batch method commonly used for a large
number of compounds (10) is one in which the substrate of interest, at different
concentrations, are inoculated with the small amount of biomass. Then the increase in
biomass concentration in each reactor is measured as a function of time. Another
technique focuses on the substrate removal rather than the microbial growth. This is
commonly used in engineering studies. The batch reactors are inoculated with large
quantities of biomass and the substrate removal is measured as a function of time.
Both of these techniques have been widely used using general measures as five day
BOD and COD. Batch reactor can be used with either acclimated or unacclimated
biomass. It requires that samples be taken at discrete time intervals during the course
of biodegradation. So if unacclimated biomass is used, the number of samples required
may be large depending on the acclimation time. Tabak, et al. (11) have collected
degradability and acclimation data on 96 compounds by static culture screening
procedure and culture enrichment process. Larson and Perry (12) and Paris, et al. (13)
have done biodegradation studies with unacclimated biomass in the batch reactors and
have evaluated the kinetic parameters.
In the above procedure, the number of data points collected are less because of
manual sampling. This can be avoided by monitoring oxygen consumption as an indirect
measure of biodegradation using an electrolytic respirometer. The automatic data
collection and recording allows sufficient accumulation of data, so the reliability of
the kinetic parameters evaluated is maximized.
In the respirometer methods of BOD measurement, wastewater samples are kept
in contact with the gas phase source of oxygen. Oxygen uptake by the microogranisms
over a period of time is measured by the changes in volume or pressure of the gas
phase. An alkali is included in the apparatus to absorb carbon dioxide produced during
biodegradation. Samples are usually incubated at constant temperature and are kept
away from light. The latest development in respirometric techniques has been the
advent of an electrolytic respirometer. It supplies oxygen, produced from electrolysis
of water, to the air space above the sample in a completely sealed reaction vessel.
The production of oxygen is triggered due to the pressure changes in the reaction
vessel. Studies by Larson and Perry (12), Young and Baumann (14), Tabak, et al. (15)
and Dosanjh and Wase (16) have shown that the electrolytic respirometer eliminates
most of the technical difficulties associated with other methods for determining BOD.
It is particularly useful for the rate studies because it provides both, a continuous
record of oxygen uptake and it maintains an unchanging atmosphere over the sample
regardless of the length of the test.
EXPERIMENTAL SET-UP
The electrolytic respirometer Sapromat B—12, consists of a temperature controlled
waterbath, which contains the measuring units, a recorder for digital indication and
direct plotting of the oxygen uptake curves and a cooling unit. The waterbath has 12
reaction flasks, each connected to the recorder. Each unit as shown in figure 1
consists of a reaction vessel C, with a carbon dioxide absorber (sodalime) 3, mounted
in a stopper, an oxygen generator B, and a pressure indicator A; The vessels
interconnected by hoses, form a sealed measuring system so that the barometric
pressure fluctuations do not adversely affect the result. The magnetic stirrer 1, in the
sample 2, to be analyzed, provides vigorous agitation, thus ensuring an effective
exchange of gases. The activity of the microorganisms in the sample creates a
reduction in the pressure which is recorded by the pressure indicator. It controls
both, the electrolytic oxygen generation and the indication and the plotting of the
299
-------
measured values.
The consumption of oxygen by microorganisms creates a reduction in pressure in
the reaction vessel. As a result, the level of 0.5% sulphuric acid in the pressure
indicator rises and comes in contact with the platinum electrode. This completes the
circuit and triggers the generation of oxygen by electrolytic cell. The oxygen gas is
provided to the reaction vessel, alleviating the negative pressure. So, the level of
electrolyte in the pressure indicator drops down, breaking contact with the electrode.
This switches off the electrolytic cell. The amount of oxygen supplied to the sample is
recorded directly in milligrams per litre by the recorder. The recorder is connected to
an IBM AT which records data from the measuring units every 15 minutes.
Figure 1. Diagram of a Measuring Unit
A. Pressure indicator
B. Oxygen generator
C. Reaction vessel
1. Magnetic stirrer
2. Sample (250ml)
3. CC>2 absorber
4. Pressure indicator
5. Electrolyte
6. Electrodes
7. Recorder
BIODEGRADATION KINETICS
A considerable amount of information concerning biodegradation kinetics is
available in the published literature. Early literature shows widely differing kinetic
rates in different studies. The evaluation and prediction of the extent and rate of
biooxidation is affected by methodological and experimental factors. Regardless of the
different assumptions involved in the measurement of biodegradation rates, it is
300
-------
generally considered associated with microbial cell growth and so most of the models
for it are the same as those used to model growth and substrate removal.
Although many models have been proposed for microbial growth, the Monod
relation is the most popular kinetic expression (17). Monod model, in combination with
the linear law for substrate removal can provide an adequate description of microbial
growth behavior. It states that the cell growth is first order with respect to the
biomass concentration (X) and mixed order with respect to the substrate concentration
(S)
dX/dt = (S u™ X) / (Ks + S) . . . . . (1)
Cell growth is related to the substrate removal by the linear law
dX/dt = - Yg (dS/dt) (2)
The kinetic parameters of interest are maximum specific growth rate |xm, half
saturation constant Ks, (it is the concentration of substrate when u,=0.5jxm), and the
yield coefficient Yg. The Monod equation has two limiting cases. When the substrate
concentration is much greater than the saturation constant the term (S/KS+S)
approaches 1.0 and cell growth and substrate removal are zero order with respect to
substrate concentration. When the substrate concentration is much smaller than the
saturation constant, the term (S/KS+S) approaches (S/KS) and cell growth and
substrate removal are first order with respect to the substrate concentration. Many
researchers have used either of the above two approaches. The characteristics of
these kinetic expressions have been discussed by Sim kins and Alexander (18).
The electrolytic respirometer has been mostly used to measure the extent of
biodegradation as a percentage of the theoretical oxygen demand over a certain period
of time. Several researchers have tried to extract kinetic parameters from oxygen
uptake data. Larson and Perry (12) used empirical kinetic expressions which were
system specific. Dojlido (19) divided the oxygen uptake curve into seven different
phases and then proposed an empirical model for each phase and evaluated the
biodegradability and toxicity of a test compound by measuring empirical rate constants
and time intervals associated with each phase. Tabak, et al. (15) divided the substrate
removal region of oxygen uptake curve into two regions, separated by an inflection
point. The first period was called exponential phase where it was assumed that the
substrate was not limiting and cell growth was occuring, while the second period was
called declining phase and here it was assumed that the substrate was limiting. So
work done to obtain kinetic parameters from oxygen uptake curve has been empirical.
STRUCTURE-ACTIVITY RELATIONSHIP
The structure—activity relationships (SAR) have been widely used in pharmocology
and medicinal chemistry. The different methods and models used are free energy
models, Free-Wilson mathematical model, discriminant analysis, cluster analysis, pattern
recognition, topological methods, and quantum mechanical methods.
The free energy model of Hansch, et al. (20) is widely used. They incorporated
octanol—water coefficient, log P, in the linear free energy relationship as a measure of
lipophilicity. This provided a general SAR model for biological activity. The success of
this model has led many researchers to include additional physicochemical parameters
and properties, structural, topological and molecular indices. Using similar principles
other researchers have proposed models to include more complex relationships between
301
-------
the bioactivity and the chemical structure or the properties. Martin (21) has dicussed
these models.
Free and Wilson proposed a mathematical model to assess the additive effects of
substituents and quantitatively estimate their magnitude (22). According to their model
the structure of a compound is composed of different groups or a core which is
substituted in various positions, resulting in a series of linear equations of the form
where BA is biological activity, X: is the j**1 group, aj is the contribution of the jtn
group and (3 is the overall average activity. These linear equations are solved by least
square method for a; and (3. This method requires a large number of compounds for a
meaningful analysis and it will breakdown if there are interactions between different
groups. Fujita and Ban (23) suggested that BA should be expressed as log(l/C), where
C is the concentration of the compound that produces a constant biological response.
This modified Free— Wilson model is in common use in medicinal chemistry.
Discriminant analysis is used where only semiquantitative or qualitative data have
to be evaluated. In this method, a linear combination of parameters called linear
discriminant function is formed, which classifies the observations. Martin (21) has
discussed the background of this method and has given examples. Principal component
method is used as a preliminary step in multiple regression analysis of the Hansch
type (24). Factor analysis is used to gain insight into the structure of
multidimensional data set and involves manipulation of the eigenvectors of variance—
covariance matrix of the dataset (25). Cluster analysis is used to group similar
substituents when various combination of parameters are considered (26).
Pattern recognition is used to examine structural features and chemical
properties for the patterns associated with different biological activities (27). In this
method, a set of descriptors is generated for each compound and then suitable
algorithm is applied to find some combinations and weight of the descriptors which
give a perfect classification for a set of compounds. This classification is then applied
to another set of compounds of known classification and performance is judged by the
percentage of correct prediction.
Various methods have been proposed to relate topology of the molecule with its
biological activity. Verloop (28) has proposed a method to treat directionality of steric
effects. A computer representation of a compound is created and then measured by
tangential planes which results in five STERIMOL parameters. Kier and Hall (29) have
used molecular connectivity index, a number calculated from graph theoretical
principles for SAR correlation. Blankley (30) has reviewed all these and other methods
used in SAR.
In the field of biodegradation there are several studies which have attempted to
correlate some physical, chemical or structural property of a chemical with its
biodegradation. Based on the type and the location of the substituent groups, Geating
(31) developed an algorithm to predict biodegradation. Qualitative relationships for
different compounds have been investigated by others, but quantification is required
for regulatory purposes.
In the literature, first order rate of biodegradation or five day BOD of chemicals
have been correlated with physical or chemical properties. Paris, et al. (32,33,34)
established a correlation between second order biodegradation rate constant and the
302
-------
van der Waal's radius of substituent group, for substituted anilines and for a series of
para-substituted phenols. Wolfe, et al. (35) correlated second order alkaline hydrolysis
rate constant and biodegradation rate constants for selected pesticides and phthalate
esters. Several workers have observed a correlation between biodegradability and
liphophilicity, specifically octanol/water partition coefficients (log P). Paris, et al. (36)
found a good correlation between biodegradation rate constant and log P, for a series
of esters of 2,4—dichlorophenoxy acetic acid. Banerjee, et al. (37) obtained a similar
relationship for chlorophenols. Vaishnav, et al. (38) correlated biodegradation of 17
alcohols and 11 ketones with octanol—water coefficients using 5—day BOD data.
Pitter(39) has found a dependence of biodegradation rate on electronic factors, like
Hammett substituent constant, for a series of anilines and phenols. Dearden and
Nicholson (40) have correlated 5—day BOD with modulus difference of atomic charge
across a selected bond in a molecule for amines, phenols, aldehydes, carboxylic acids,
halogenated hydrocarbons and amino acids. A direct correlation between the
biodegradability rate constant and the molecular structure of the chemical has been
used by Govind (41) to relate the first order biodegradation rate constant with the
first order molecular connectivity index and by Boethling (42) to correlate the
biodegradation rate constants with the molecular connectivities for dialkyl esters,
carbamates, dialkyl ethers, dialkyl phthalate esters and aliphatic acids.
GROUP CONTRIBUTION APPROACH
Using a group contribution approach, a very large number of chemicals of
interest can be constituted from perhaps a few hundered functional groups. The
prediction of the property is based on the structure of the compound. According to
this method the molecules of a compound are structurally decomposed into functional
groups or their fragments, each having unique contribution towards the compound
property.
The biodegradability rate constant k, is expressed as a series function of
contribution a;, of each group of the compound. The first order approximation of this
series function representing biodegradation rate constant can be expressed as
Ln (k) = 2
(3)
where Nj is the number of groups of type j in the compound, oj is the contribution
of group of type j and L is the total number of groups in the compound.
CALCULATION OF GROUP CONTRIBUTION PARAMETER, a
Using equation 3 for each compound, a linear equation in ot's is constructed. For
a given dataset this generates a series of linear equations which are solved for ex's
using regression analysis.
If there are n a's and m compounds (n
-------
m
S = 2 { Ln(ki)
1=1
j }
(4)
The estimates of a's should be such that it produces the least value of S. These
estimates are obtained by differentiating equation 4 with respect to a and setting it
equal to zero
m
= -22{[Ln(ki)-2Njjaj]Nki} =
(5)
This will generate a series of n linear equations which are solved for a's. If N is the
matrix of coefficients of a's and Y is the vector of Ln(k) values then the solution
vector a is given by
a=(N'N)~1 (N'Y)
where ' denotes the transpose of matrix.
MODEL VERIFICATION
The study of Urano and Kato (43) was used to determine the contribution of
several different groups. This study was selected because it involved testing of a large
number of chemicals using a consistent method. The requirement of consistent method
is necessary to ensure that the difference in rate of biodegradation of different
compounds is due to the difference in their chemical structure and not the test
conditions. The range and average of Lnfk) values, for compounds used for this model,
are given in Table 1. The average values were used for calculation of the group
contribution parameters given in Table 2.
The biodegradation experiments were carried out for cresols, phenol, 2,4— dimethyl
phenol, 2— butanone, acetone, 1— phenyl hexane and butyl benzene using an electrolytic
respirometer Sapromat B— 12 (Voith— Morden, Milwaukee, Wl). The chemicals were from
Aldrich chemical company with 99+% purity. These compounds were used to validate
the model, but were not used in the calculation of the group contribution parameters.
The experimental conditions were : temperature 25°C, biomass concentration 30 mg/L
and compound concentration lOOmg/L. The biomass was obtained from Little Miami
wastewater treatment plant in Cincinnati, which receives 95% domestic waste. The
biomass was aerated for 24 hours and then used for the experiment. The
biodegradation rate constants for these compounds were determined using the
following empirical first order rate equation given by Urano and Kato (44).
dBOD/dt = k BOD.
Integrating this equation
log (BOD) = kt + constant
(6)
(7)
is obtained. The above equation was used for only the rising part of the BOD versus
time curve, and initial data and data for endogenous phase were neglected. Note that
the selection of the rising part of the BOD curve was arbitrary and was based on
visual inspection of the BOD curve. The biodegradation rate constants were also
predicted using the group contribution parameters of Table 1. The comparison of these
values is given in Table 3 and Figure 2. The % error in BOD values were calculated
304
-------
-------
by
% Error = (1/n) 2 {(BODe - BODp)/BODe}*100
where subscripts e and p denote experimental and predicted values respectively and n
is the total number of data points used. Figure 3 shows the best and worst case of
prediction obtained. The experimental conditions in both the studies were similar
except the nature and the source of biomass.
RESULTS AND DISCUSSION
The average error in the prediction of BOD values varies from 13% to 85%. It is
important to emphasize that there are several reasons for the rather large discrepancy
between the predicted and experimental values. These reasons are as follows :
1. The data used for calculating the group contribution parameters were developed
by Urano and Kato (refer Table 1). The experimental BOD curves, obtained from
electrolytic respirometry, were approximated by an exponential relation (refer
equations 6 and 7}. The range of curve to which this equation was applied was
selected arbitrarily. This resulted in some error between the experimental data and the
calculated value of the kinetic constant, k. Furthermore, a range of k values were
calculated by Urano and Kato due to differences in the replicate samples. In our
calculation of the group contribution parameters, an average value for the k values
was used. This introduced another error in the input data for the group contribution
estimates.
2. The group contribution method described here is the first order approximation. It
assumes that the contributions of the groups are independent of each other. So the
rates predicted are within an order of magnitude.
3. Biomass used in both the studies were different. Urano and Kato had acclimated
their biomass to the compound before using it in the experiment, while our biomass
was from domestic wastewater treatment plant and was not acclimated to the
compounds used in the study.
Inspite of these errors, the prediction is within an order of magnitude. With
availability of more data and using a detailed kinetic model rather than a first order
exponential equation, the prediction error may be reduced considerably.
ACKNOWLEDGMENT :This research was supported by co-operative agreement CR
812939—01 from U. S. Environmental Protection Agency.
REFERENCES
1. Blackburn, J.W., Troxler, W.L. Prediction of the fates of organic chemicals in a
biological treatment process — An overview. Environ. Prog. 3: 163,1984.
2. Howard, P.H., Saxena, J., Sikka, H. Determining the fate of chemicals. Environ.
Set. Tech. 12: 398,1978.
3. Alexander, M. Biodegradation of chemicals of environmental concern. Science 211:
132,1980.
4. Strier, M.P. Pollutant treatability : A molecular engineering approach. Environ.
Sci. Tech. 14: 28,1980.
5. Howard, P.H., Banerjee, S., Rosenberg, A. Review and evaluation of available
techniques for determining persistance and routes of degradation of chemical
substances in environment. 560/5—81-011, U.S. Environmental Protection Agency,
306
-------
o
Q
O
m
o
o
m
s^»
0>
o
Q
O
m
Q
o
m
•s^X
o»
3
0.9
0.8 -
0.7 -
0.6 -
0.5 -
0.4 -
0.3 -
0.2 -
0.1 -
0
0.05
0.8
0.7 -
0.6 -
0.5 -
0.4 -
0.3 -
0.2 -
0.1 -
2,4-Dimethylphenol
0.15
0.25
0.35
0.45
1-Phenylhexane
0.2
—1—
0.4
0.6
Time, days
d Experimental Data + Predicted Data
Figure 3. Best and worst case of prediction
307
-------
1981.
6. Grady, C.P.L., Jr. Biodegradation : Its measurement and microbiological basis.
Biotech. Bioengg. 27: 660, 1985.
7. Braha, A., Hafner, F. Use of lab batch reactors to model kinetics. Wat Res. 21:
73,1987.
8. Chudoba, J., Cech, J.S., Farkac, J., Grau, P. Contro of activated sludge
filamentous bulking ; Experimental verification of a kinetic selection theory. Wat.
Res, 19: 191, 1985.
9. Bull, AT., Slater, J.H. In: AT. Bull, J.H. Slater feds.), Microbial Interaction and
Communities, Vol. 1., Academic Press, New York, NY, 1982.
10. D'Adamo, P.C., Rozich, A.F., Gaudy, A.F., Jr. Analysis of growth data with
inhibitory carbon sources. Biotech Bioengg. 26: 397,1984.
11. Tabak, H.H., Quave, S.A., Mashi, C.I., Barth, E.F. Biodegradability studies with
organic priority pollutant compounds. JWPCF 53: 1503,1983.
12. Larson, R.J., Perry, R.L. Use of electrolytic respirometer to measure
biodegradation in natural waters. Wat. Res. 15: 697,1981.
13. Paris, D.F., Rogers, J.E. Kinetic concepts for measuring microbial rate constants
: Effects of nutrients on rate constants. Appl. Environ. Microbiol. 51: 221,1986.
14. Young, J.C., Baumann, E.R. The electrolytic respirometer — II — Use in water
pollution control plant laboratories. Wat. Res. 10: 1141,1976.
15. Tabak, H.H., Lewis, R.F., Oshima, A. Electrolytic respirometry biodegradation
studies, OECD Ring test of respiration method for determination of
biodegradability. Draft Manuscript, MERL, U.S. Environmental Protection Agency,
Cincinnati, OH, 1984.
16. Dosanjh, M.K., Wase, D.A.J. Oxygen uptake studies on various sludges adapted to
a waste containing chloro—, nitro—, and amino—substituted xenobiotics. Wat. Res.
21: 205,1987.
17. Howard, P.H., Banerjee, S. Interpreting results from biodegradability tests of
chemicals in water and soil. Environ. Tox. Chem. 3: 551, 1984.
18. Simkins, S., Alexander, M. Models for mineralization kinetics with variables of
substrate concentration and population density. Appl. Environ. Microbiol. 47:
1299,1984.
19. Dojlido, T.R. Investigations of biodegradability and toxicity of organic compounds.
EPA-600/2-79-163, U.S. Environmental Protection Agency, Cincinnati, OH, 1979.
20. Hansch, C. On the structure of medicinal chemistry. J. Med. Chem. 19:1, 1976.
21. Martin, Y. C. Quantitative Drug Design : A Critical Introduction, Dekker,
NewYork, 1978.
22. Free, S. M., Wilson, J. W. A mathematical contribution to structure-activity
studies. J. Med. Chem. 7:395,1964.
23. Fujita, T., Ban, T. Structure-activity studies of phenylamines as substrates of
biosynthetic enzymes of sympathetic transmitters. J. Med. Chem. 14:148, 1971.
24. Mayer, P. P. Drug Design, Ed., E. J. Ariens, Medicinal Chemistry, Vol. IX,
p.187, Academic Press, NY, 1980.
25. Weiner, M. L., Weiner, P. H. A study of structure-activity realtionships of a
series of diphenylaminopropanols by factor analysis. J. Med. Chem. 16:655, 1973.
26. Hansch, C., Leo, A. J. Substituent Constants for Correlation Analysis in
Chemistry and Biology, Wiley (Interscience), NY, 1979.
27. Dunn, W. J., Ill, Wold, S. Structure—activity analysed by pattern recognition :
The asymmetric case. J. Med. Chem. 23:595,1980.
28. Verloop, A., Hoogenstranten, W., Tipker, J. Drug Design, Ed., E. J. Ariens, Vol.
VII, p.165, Academic Press, NY, 1977.
29. Kier, L. B., Hall, L. H. Molecular Connectivity in Chemistry and Drug Research,
Academic Press, NY, 1976.
30. Blankley, J. C. Quantitative Structure-Activity Realtionships of Drugs, Ed.
308
-------
Topliss, J. G., p.l, Academic Press, NY, 1983.
31. Geating, J. Literature study of the biodegradability of chemicals in water. EPA—
600/2-81-175, U.S. Environmental Protection Agency, 1981.
32. Paris, D. F., Wolfe, N. L. Realtionships between properties of a series of anilines
and their transformation by bacteria. Appl. Environ. Microbiol. 53:911,1987.
33. Paris, D. F., Wolfe, N. L., Steen, W. C. Structure-activity relationships in
microbial transformation of bacteria. Appl. Environ. Microbiol. 53:971,1987.
34. Paris, D. F., Wolfe, N. L., Steen, W. C., Baugham, G. L. Effect of phenol
molecular structure on bacterial transformation rate constants in pond and river
samples. Appl. Environ. Microbiol. 45:1153,1983.
35. Wolfe, N.L., Paris, D.F., Steen, W.C., Baugham, G.L. Correlation of microbial
degradation rates with chemical structure. Environ. Sci. Tech. 14: 1143, 1980.
36. Paris, D. F., Wolfe, N. L., Steen, W. C. Microbial transformation of ester of
chlorinated carboxylic acids. Appl. Environ. Microbiol. 47:7,1984.
37. Banerjee, S., Howard, P. H., Rosenburg, A. M., Dombrowski, A. E., Sikka, H.,
Tullis, D. L. Development of a general kinetic model for biodegradation and its
application to chlorophenols and related compounds. Environ. Sci. Tech. 18:416,
1984.
38. Vaishnav, D. D., Boethling, R. S., Babeu, L. Quantitative structure—
biodegradability relationships for alcohols, ketones and alicyclic componds.
Chemosohere. 16:695, 1987.
39. Pitter, P. Correlation between the structure of aromatic compounds and the rate
of their biological degradation. Collection Czecoslovak Chemical Comm. 49:2891,
1984.
40. Dearden, J.C., Nicholson, R.M. The prediction of biodegradabilities by the use of
quantitative structure—activity relationships : Correlation of biological oxygen
demand with atomic charge difference. Pestici. Sci. 17: 305, 1986.
41. Govind, R., Treatability of toxics in wastewater systems. Hazardous Substances 2:
16,1987.
42. Boethling, R.S. Application of molecular topology to quantitative structure—
biodegradability relationships. Environ. Tox. Chem. 5: 797, 1986.
43. Urano, K., Kato, Z. Evaluation of biodegradation ranks of priority organic
compounds. J> Hazardous Mtl. 13: 147, 1986.
44. Urano, K., Kato, Z. A method to classify biodegradabilities of organic compounds.
J. Hazardous Mtl. 13:135, 1986.
309
-------
TABLE 1. DATA USED TO CALCULATE GROUP CONTRIBUTION PARAMETERS
Compound
Ethyl alcohol
Butyl alcohol
Ethylene glycol
Acetic acid
Propionic acid
n— Butyric acid
n— Valeric acid
Adipic acid
Methyl ethyl ketone
Hexamethylenediamine
n— Hexylamine
Mono ethanol amine
Acetamide
Benzene
Benzyl alcohol
Toluene
Acetophenone
Hydroxybenzoic acid
Aminobenzoic acid
Aminophenol
Ln(k)
-2.9004 ~
-3.0791 "
-3.3524 "
-2.6037 "
-2.6736 "
-2.7031 ~
-2.6310 -
-2.8134 "
-3.4738 ~
-4.3428 "
-2.8647 "
-3.3242 "
-2.9957 ~
-2.8647 "
-2.7806 ~
-2.6037 "
-3.1701 "
-2.3538 "
-2.6592 "
-3.2442 "
-3.1465
-3.3242
-3.6496
-2.7181
-2.9759
-3.0576
-2.6593
-3.1235
-3.6889
-4.5099
-3.0576
-3.3814
-3.0576
-2.9759
-3.1701
-2.8647
-3.5404
-2.9188
-2.9374
-3.2968
Average Ln(k)
-3.0159
-3.1942
-3.4900
-2.6593
-2.8134
-2.8674
-2.6451
-2.9565
-3.5755
-4.4228
-2.9565
-3.3524
-3.0262
-2.9188
-2.9565
-2.7257
-3.3382
-2.6363
-2.7984
-3.2702
310
-------
TABLE 2. GROUP CONTRIBUTION PARAMETERS
Group
Methyl
Methylene
Ketone
Amine
Acid
Hydroxy
Aromatic CH
Aromatic Carbon
CH3
CH2
CO
NH2
COOH
OH
ACH
AC
-1.3667
-0.0438
-0.5073
-1.4654
-1.3133
-1.7088
-0.5016
1.0659
TABLE 3. COMPARISON OF ACTUAL AND PREDICTED Ln(k) VALUES
Compound
o— Cresol
m— Cresol
p— Cresol
Phenol
2,4— Dimethylphenol
2— Butanone
Acetone
Butylbenzene
1— Phenylhexane
Actual
Ln(k)
-2.6878
-2.3694
-2.4647
-3.0006
-2.8460
-3.1326
-3.1161
-3.1285
-3.3971
Predicted
Ln(k)
-2.9501
-2.9501
-2.9501
-3.1509
-2.7493
-3.2845
-3.2407
-2.9402
-3.0278
Error %
41.71
27.94
57.10
42.28
13.13
50.57
13.86
28.78
85.31
311
-------
PRELIMINARY RESULTS ON THE ANAEROBIC/AEROBIC
BIOCHEMICAL REACTOR FOR THE MINERALIZATION OF ORGANIC
CONTAMINANTS BOUND ON SOIL FINES
Robert C. Ahlert, PhD, PE, Dist. Prof.
David S. Kosson, PhD, Asst. Prof.
William V. Black, Graduate Student
Chemical & Biochemical Engineering
Rutgers University
P.O. Box 909, Piscataway, NJ 08855
ABSTRACT
The goal of the overall research program, a part of which is
discussed in this paper, is to demonstrate a sequence of
aerobic/anaerobic microbial process steps for degradation of
contaminated soil fines and slurries of soil fines. Toward this end,
it must be possible i) to assay individual organic species and total
contaminant organic carbon in soils of varying properties, ii) to
separate whole soils into fractions according to particle size, and
iii) to assay [as in i)] reactor slurries containing suspended soil
particles, microbial culture and dissolved, dispersed and sorbed
organic contaminants and metabolites. These techniques are required
to define the nature of the contamination, devise operating
conditions to facilitate microbial contact, and assure complete
mineralization of target contaminations and "clean" residuals.
The first major section of this paper describes the development of
analytical methodology for whole soil and soil fractions; in
parallel, techniques for mixing/homogenizing, fractionation and
extraction used in sample preparation are discussed. It has been
possible to separate soil fines and some "bulk" organic matter. A
large part of the total organic chemical contamination is due to
sorption and physical pore interactions with the fine particle
fractions [clay minerals and humic substances] of whole soil.
A second section describes microbial degradation experiments.
Systems and procedures for microbial reactions were designed and
implemented to accomodate the properties and behavior of target
substrates. Both shake flask and fermentation reactions are being
carried out on slurries of soil fines. Low molecular weight
polynuclear and chlorinated aromatic hydrocarbons are readily
biodegraded.
312
-------
INTRODUCTION
Distillation bottoms and sludges from benzene-toluene-xylenes [BTX]
production were impounded for several decades. The production process
consisted of the catalytic cracking of naphtha, in the presence of fuming
sulfuric acid, and distillation. Therefore, lagoon contents include
naphtha-related compounds, distillate residues and compounds resulting from
reactions of these species and sulfuric acid. Possible contaminants
include, but are not limited to, simple aromatic species, polynuclear
aromatic hydrocarbons [PAHs], phthalates, as well as sulfonated derivatives
of these compounds. Many of these species have slight solubility in water
and/or an affinity for some of the constituents of soil and have migrated
into and through the soil immediately adjacent to the lagoon.
During this study, soil samples were obtained from the containment area
surrounding the lagoon. The impoundment has been designated a CERCLA-NPL
site; it exceeds ten acres in extent and contains an estimated 100,000 cubic
yards of residues. The contents of the lagoon have separated into several
distinct layers that include, in bottom-to-top sequence, a solid mixture of
organic and inorganic substances, a tar-like layer, a layer of viscous
organic matter, and a floating aqueous layer.
ANALYTICAL APPROACH
Initially, soil samples are mixed and homogenized. The resulting material
is air-dried and sieved through a 3-cm brass screen to remove debris, rocks
and gravel; this procedure also breaks up macro-agglomerates. A second
sieving, with a 5-mm screen, improves homogeneity, enhances mixing and helps
toward analytical reproducability. Direct solvent extractions of
homogenized, contaminated soil utilize methanol, cyclohexane, or methylene
chloride. Methanol has relatively high polarity, cyclohexane is a model
cyclic compound, and methylene chloride is a moderately polar, volatile
compound with broad solvent capabilities.
Gas chromatography [GC] is used to identify and quantify compounds in soil
extracts. USEPA Test Methods 602 and 610 are employed in these analyses.
Method 602 is used to assay aromatic species in GC column effluent with a
photoionization detector [PID] in series with an electrolytic conductivity
detector [ELCD]. Method 610 is employed to detect PAHs and phthalates,
utilizing a flame ionization detector [FID]. Standard solutions are assayed
in sequence with solvent solutions to match retention times for compound
identifications. Peak areas are used to construct standard curves,and
provide a basis for determination of contaminant concentrations. Compound
concentrations are calculated from both PID and FID output to check
analytical consistancy.
Experimental Methods
Soil contaminant levels were initially estimated to fall between 2 and 5 %
on a dry weight basis. Direct solvent extractions are carried on varying
masses of soil with the goal of limiting contaminant concentrations to
about 100 mg/L in extract solutions. This target concentration was adopted
313
-------
to avoid overloading the detectors. Soil masses varying from 0.15 to 0.38 g
are extracted with a fixed volume of solvent.
Duplicate amber serum bottles, each with label, septum and aluminum cap, are
weighed with a Mettler PE3600 balance. Bottles are 100 ml in volume Soil
is added until target weights are attained; approximately 50 ml of methanol,
cyclohexane or methylene chloride is added to both bottles. Bottles are
sealed with the septa [Teflon-coated neoprene] and reweighed. Experimental
errors include the small discrepancies in obtaining target soil masses and
measuring solvent volumes. In general, these are accounted for in
concentration calculations. A second form of the experiment is carried out
to facilitate compound identification. It is the same in all respects,
except that 20 g of soil are added; since quantification is not desired,
duplicates are not performed.
Extraction vessels are shaken for approximately one hour. This time was
found to be adequate in earlier studies; however, it assumes that only
readily reversible, high-rate sorption processes are involved. Higher
energy binding processes and sorbate trapped by capillary forces would not
participate in such short-term partitioning. Extract solutions are filtered
through 0.2-um MSI Cameo II 25-mm disposable syringe filters, into duplicate
5-mL vials. Samples are stored at 4°C, to minimize volatilization losses,
and enclosed to exclude light and avoid photolytic chemical reactions.
Method 602 utilizes a 1.8-m long by 2-mm ID stainless steel GC column packed
with 100/200 mesh Supelcoport, coated with 5 % SP-1200 and 1.75 % Bentonite-
34. Oven temperature is held at 50°C for 2 min; a 6°C ramp takes the oven
to 90° for a final period of 23 min. The PID has detection limits of 1 to
10 pg for unsaturated carbon bonds found in aromatic compounds. The ELCD
has detection limits of 0.1 to 1 pg when used in the detection of
chlorinated compounds or to verify PID results. Sample size is 2 uL.
Method 610 is applicable to PAHs and phthalates. It utilizes a 1.8-m long
by 2-mni ID glass GC column packed with 100/200 mesh Chromosorb W-AW-DCMS,
coated with 3 % OV-17. Oven temperature is held at 100°C for 4 min; a 8°C
ramp takes the oven to a final temperature of 280°C. The FID has detection
limits of 10 to 100 pg. Sample size is 5 uL.
Chemical Oxygen Demand [COD] is determined for some contaminated soil
samples. This procedure is identical to that described in Standard
Methods". The COD has some value for comparison with carbon in identified
species, to estimate extraction and identification efficiencies.
Analytical Results
Extracts generated in Experiment 12888 were distinctly different in color.
After filtration, the cyclohexane extract was translucent orange, methylene
chloride gave an opaque brown liquid, and the methanol solution was clear
and tan. This appeared to be evidence for susbtantial variation in
extraction efficiency. In addition, the high polarity of methanol leads to
destruction of soil aggregates; thus, methanol is capable of extracting
contaminants held in micro- and macro-pores by capillarity and interfacial
tension. Solvents can be compared on the basis of the mass of naphthalene
314
-------
extracted. No naphthalene was extracted by cyclohexane; methanol and
methylene chloride extracted 1,089 and 1,403 mg/kg dry soil, respectively.
Thus, for naphthalene, solvent power varied considerably.
The PID sees cyclohexane and impurities in methylene chloride. The ELCD
detects chlorinated compounds and is overloaded by methylene chloride
solutions. The consequence is chromatogram baseline fluctuations and large
peak area integration inaccuracies. Methanol was the only solvent suited to
Method 602. Methylene chloride was used in conjunction with Method 610.
Eight major organic compounds were identified and quantified in methanol
solution. In order of decreasing concentration [mg/kg], they are:
naphthalene - 1,090; 1,2-dichlorobenzene - 360; toluene - 150; xylene
isomers - 145; benzene - 113; and, ethyl benzene - 28. The eight species
account for 64 % of the total peak area of chromatographic responses.
Benzene, toluene and the xylenes are primary products of naphtha
distillation. GC residence times for standard solutions and extract
solutions, with PID detection, varied less than 0.004 sec for this group.
Naphthalene is also a major component of naphtha; residence times differed
by 0.004 sec. Ethylbenzene [EB] is formed by catalytic reaction of benzene
with ethylene, an olefin found in industrial naphtha. In BTX production,
sulfuric acid is the catalyst. EB residence times varied by 0.008 sec.
The appearance of 1,2-dichlorobenzene was signaled by the PID and verified
by the ELCD, a halogen-specific detector. Chlorinated compounds are not
normally found in naphtha nor are they produced by sulfuric acid catalysis.
The presence of this compound may indicate disposal to the impoundment from
another manufacturing source or a spill clean-up activity. Two substantial,
unidentifiable peaks were encountered with Method 610; neither was observed
with Method 602. These peaks correspond to compounds that are believed to
be sulfonated aromatic hydrocarbons; operating temperatures for Method 602
preclude elution of such higher boiling species. Naphthalene sulfonic acids
represent compounds of higher molecular weight and boiling point, requiring
increased GC oven temperatures [Method 610] and extended residence times.
Standard solutions
[see Table la] and
solvent [see Table
Method 610 for the
Similarly, Figures
purgable aromatic
is the naphthalene
included seven purgable aromatic compounds in methanol
fifteen PAHs in a 50:50 methanol:methylene chloride mixed
lb]. Figures 1 and 2 are gas chromatograms obtained by
PAH standard solution and methanol extract, respectively.
3 and 4 are GC outputs obtained by Method 602 for the
standard and the methanol extract, respectively. Figure 5
case run with a modified version of Method 602.
COD has been measured for several soil samples. This assay is used to
determine the total oxygen required to fully oxidize all reduced species; it
does not distinguish between contaminant organic carbon and hydrogen, soil
organic matter, and metals in reduced or partially oxidized states. The COD
of sieved, air-dried, unextracted soil corresponds to about 50 mg C/kg soil.
In comparison, the total organic carbon [TOC] associated with the eight
quantified contaminant species cited above is approximately 1,6 mg C/kg soil
or 3.2 % of the COD carbon equivalent. This is not a reflection of
extraction efficiency. However, CODs run on soil samples taken at points
315
-------
remote to the disposal lagoon average close to 50 mg C/kg soil. Thus,
extraction efficiency with methanol is probably relatively good.
Table la
Purgable Aromatic Standard Mixture [602-M]
(in Methanol)
Compound
Benzene
Toluene
Ethyl benzene
Chlorobenzene
1,2-Di chlorobenzene
1,3-Di chlorobenzene
1,4-Di chlorobenzene
Concentration (mg/L)
2000
2000
2000
2000
2000
2000
2000
Table Ib
Polynuclear Aromatic Hydrocarbon Standard Mixture [610-M]
fin 50:50 Methanol:Methv1ene Chloride)
Compound
Acenaphthene
Fluoranthene
Naphthalene
Benzo(a)anthracene
Benzo(a)pyrene
Benzo(b)f1uoranthene
Benzo(k)f1uoranthene
Chrysene (93 %)
Acenaphthylene
Anthracene
Benzo(g,h,i)pyrene
Fluorene
Phenanthrene
Dibenzo(a,h)anthracene
Indeno(l,2,3-cd)pyrene
Pyrene
Concentration (mq/L)
1000
200
1000
100
100
200
100
100
2000
100
200
200
100
200
100
100
316
-------
Method 610
figure i
6.2S
13 7.62
ftuorene.
12.72. Phenanthrene and
!
16.86
16.58
20-
18.48
19.4S
. 17 8en20 (a)anthracene and
22.42
ffTamb)fiuoranth&ne4i\d Senze
-------
5--
10-
15-
20-
Method 610
2
, 16. 69
) 17.03
P 17,53
J&.20
.^*
>Z1*ft»
•8«2^
23. 3Z
318
-------
seoo-
O
-P
U
«<">-
1
2JOO-
JOO
9200
Chromatograph for Standard
Solution of Purgable Aromatics
Method 602
Figure 3
Ch lorobenzene
1 , 4-dichlorobenzene
1 ,3 -diehlorobenzene
1 ,2-dichlorobenzene
6ZOO-
S-i
O
O
JJ
Pi
Q
M
PH
zeno-
zon
Benzene
Toluene
Ethyl benzene
Chlorobenzene
-1,4-dichlorobenzene
-1,3-dichlorobenzene
"T"
T
'V _______
18
1 , 2-dichlorobenzene
'T'
T"
"I"
3£
T"
319
-------
C37S-
Hethanol Extract
Method 602
Figure 4
tn
O
4J
O
\.
~r
10
.-'~\--.
Naplithalene —-'"
1 , 2-dichlorobenzene
25
I
55
320
-------
3SBII -i
o rsn-
4J
o
0)
0)
o .,
Naphthalene by Modified
Method 602
Figure 5
Naphthalene-
j
B'iWI-l
U
'QJ
4J
(1)
.f
1C
"1 "
15
"1
20
Standard
r
T"
4.0
Naphthalene
:i
A.
5
14
._, r—=_n.
IS 20 • • ZS
Methanol Extract
T
30
T-
321
-------
MICROBIAL APPROACH
The solubilities of PAHs in aqueous media decreases with increasing
molecular weight; for four rings or more, saturation concentrations are very
low and difficult to measure. Within this context, rates of aerobic
metabolism of PAHs are insignificant. In addition, this class of aromatic
compounds has a high affinity for several constituents of natural soils,
i.e., soil organic matter [humic substances] and clay minerals. PAHs are
lipophilic, i.e., prefer sorptive association and/or dissolution in organic
rather than hydrophilic phases. In addition, many multi-ring molecules can
migrate into stable positions inside clay mineral structures, either
diffusing between laminae in mica-like structures or entering the crystal
lattice directly. The latter process is a form of clathration. These
preferred soil components are, in general, the smaller particle fractions of
the soil system. Thus, it is possible to introduce PAHs into a microbial
systems in slurry form, with substrate(s) bound to particulate matter.
A slurry form of bioreactor has a number of possible thermodynamic phases
present, including: aqueous medium with dissolved substrate and nutrients,
substrate bound to soil particles, suspended single and clustered cells,
substrate emulsions and colloids, and cells attached to the exterior and
macro-pore surfaces of dispersed particles. Sampling and analytical
methodology are difficult throughout the design and implementation of
biodegradation experiments.
An earlier Section dealt with the problems of accurate chemical contaminant
identification and quantification for soils prior to chemical or biochemical
reaction. Often tightly-bound PAHs must be extracted from a soil mass, with
uncertain efficiency or recovery. Method verification is critical. Extract
solution is assayed by GC or High Performance Liquid Chromatography [HPLC].
Multiple phases must be sampled and extracted, during and after biochemical
transformations. The possible appearance of intermediate or final
metabolites, i.e., incomplete mineralization, adds to the complexity of mass
balances for substrate species. COD is a useful tool for quasi-continuous
monitoring of the progress of slurry-type bioreactor systems. In well-
aerated aqueous systems, metals are not a serious factor. However, variable
cell mass and natural soil organic matter severely limit this approach.
Before slurry reaction can be undertaken, whole soil must be reduced to
several particle size fractions. Contaminant PAHs favor finer particles,
thus, fractionation is a useful way of concentrating these compounds prior
to degradation experiments. A reproducible method for soil classification
has been developed and has been demonstrated to lead to the desired
concentration of target substrates, as follows. Whole soil is separated
into three phases: a tar-like organic phase, a coarse sandy phase and an
aqueous suspension of soil fines. The latter is suited to slurry reactor
experiments. Experiments consist of small-scale studies in shake flasks,
performed in matrix formats, and fermentation studies in reaction vessels of
larger volume.
322
-------
Background
The rates of microbial assimilation of PAHs have been demonstrated to be
functions of solubility, molecular weight, number of six-member rings, ,
degree and type of substitution, as well as environmental conditions, such
as temperature, pH and oxygen concentration. The solubility of
unsubstituted PAHs,in water, drops sharply as the number of rings increases.
It rapidly diminishes to levels that are too low to support significant
biological activity; see Table 2 for data. Compounds of six or more rings
have vanishing solubility in water.
The number and type of substituents on or in a PAH molecule have a marked
influence on solubility. The solubilities of phenols, nitrogen
heterocyclics, polynuclear polyols, sulfonates and other mono- and poly-
substituted PAHs are often significantly higher than the basic hydrocarbons.
Therefore, substituted compounds are more likely to be observed as solutes
in contaminated groundwater. Also, surfactants increase PAH solubility.
However, these compounds complex or "react" with the high molecular weight
polynuclear species to create a composite hydrophilic exterior. The result
is either a stable emulsion, colloidal suspension or micro-dispersion; it
cannot properly be classed as dissolution in the thermodynamic sense of a
homogeneous liquid phase. Sodium laurylsulfate increases the solubility of
2- to 7-ring PAHs by 2 orders-of-magnitude or more.
Biodegradation of 2- to 3-ring PAHs by pure microbial cultures has been
demonstrated; naphthalene, phenanthrene and anthracene have been shown to be
assimilated quantitatively. Higher molecular weight compounds, i.e.,
benzo(a)anthracene and benzo(a)pyrene, can be degraded to simpler
intermediates in the presence of supplementary carbon sources or
cometabolites, i.e., biphenyl and succinate.
Bacteria concentrate, grow, and form bioslimes in aqueous boundary layers at
liquid/liquid and liquid/solid interfaces. Organic cosolvents can transport
PAHs to such interfaces and increase the rate of biodegradation. There is
no information to support microbial metabolism of solid PAHs; similarly,
there is little data on the bioreaction of PAHs sorbed on nonreacting
surfaces.
Compound
Table 2. Solubilities of PAHs in Hater
Mol. Wt.
Naphthalene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(a)pyrene
Dibenzanthracene
128
154
166
178
178
202
202
252
278
Solubility (ug/L)
31,700
3,200
2,000
1,300
73
260
140
4
# of Rings
2
3
3
3
3
4
4
5
5
323
-------
Experimental
i) Apparatus
Aerobic biodegradation experiments are carried out in 60-mL Ehrlenmeyer
flasks, on a laboratory shake device, or in 3-L [working volume]
fermentation vessels. The shaker studies are arranged in matrix format with
the following common composition: 30 ml soil slurry, 20 ml inoculum and 10
ml nutrient medium [see Table 3]. The composition of the inoculum is varied
to provide a basis for the evaluation of substrate volatility losses,
reactor surface wetting and sorption onto biomass as mechanisms of substrate
disappearance. The 20 ml "inoculum" is live culture; in sterile controls,
the 20 ml of seed is replaced by autoclaved culture or deionized water.
Controls are intended to illustrate the extent to which volatilization and
inorganic surface sorption influence sub-strate fate. Given that biomass
rendered unviable by autoclaving retains substantial sorption capacity,
autoclaved culture is designed to investigate this loss pathway. Also, this
control provides a zero-time or baseline measurement for carbon or oxygen'
demand. The pH of the composite solution is adjusted to 7.15 by addition of
a mixture of solid potassium monobasic and dibasic phosphates.
Fermentation studies are carried out with a working volume of 3 L prepared
in the same ratio of 3:2:1 for soil slurry:inoculum:nutrient medium as in
the shake flask studies. The reactor is sparged with air at the rate of 5
to 6 L/min; it is stirred at 300 rpm. The pH is maintained at 7.15 by
periodic additions of 0.25M sodium hydroxide regulated by a pH controller.
Samples are taken at 24-hr intervals, to monitor the course of reaction.
Separate studies utilizing deionized water are used to evaluate
volatilization losses.
Table 3 - Medium Composition
Constituent Concentration [ma/Li
(NH4)2S04
MgS04.7H20
FeCl3.6H20
MnS04.H20
CaClo
1,500
100
0.5
10
7.5
324
-------
ii) Soil Slurry
Soil slurry is prepared by a sequence of homogenization, extraction, and
fractionation steps. This procedure is designed to create a suspension of
soil fines that displays minimum variation from batch-to-batch. Whole soil
samples are homogenized by passing the air-dried material through a 5-mm
screen, quartering the resulting solids cone through the apex, segregating
the quarters, and sieving each quarter to form a new cone. Soil is sieved
three times. A prescreening with a 3-cm sieve removes rocks and
miscellaneous debris and, also, serves to break-up larger clumps of packed
soil.
Homogenized soil [84 g on a dry basis] is extracted with 350 ml of water at
pH 7. Extraction separates the contaminated soil into three phases: a tar-
like [smell, sticky, viscous, etc.] organic phase corresponding to 0.65 -
0,75 % of the initial dry mass; a mixture of larger, heavier particles
[sand]; and, an aqueous supension [slurry] of fine soil particles. The
aqueous suspension is separated from the settleable solids by screening
through a 10-micron sieve; filtered solids are washed with 650 ml of water
to dilute the filtrate slurry to 1 liter and the final concentration. The
fractionation procedure has been found to retain approximately 65 % of the
initial dry mass of soil on the sieve. The final slurry of fine particles
is stable and does not show any evidence of settling under experimental
conditions.
iii) Inoculum
The aerobic inoculum is obtained as waste sludge from the Somerset/Raritan
Valley Sewage Authority.
iv) Nutrient Medium
The nutrient medium is prepared from a conventional recipe for aerobic
cultures and does not employ a primary or supplementary carbon source; see
Table 3. The soil slurry supports microbial activity without the use of a
co-substrate; none is added.
v) Analytical
The progress of substrate conversion is monitored by COD and GC analyses.
Samples are prepared for the latter by extraction with methylene chloride at
8 ml of solvent for 20 ml of slurry. The remainder of the assay is carried
out in accordance with USEPA Method 610. A 1.8-m long by 2-mm ID glass
column is packed with a stationary phase consisting of 100/200 mesh
Chromosorb W-AW-DCMS coated with 3 % OV-17. Oven temperature is held at
100°C for 4 min; a 8°C/min ramp increases the temperature to 280°C. An FID
is used to determine residence times and peak areas. Gas pressures are 70,
40 and 65 psi for nitrogen, hydrogen and dry air, respectively. A 5-uL
sample is injected into the GC.
325
-------
Results
Table 4 summarizes moisture contents and COD analyses for whole soil
and the fractions generated by repeated sieving, fractionation and
extraction. COD determinations are referred to 1 kg of air-dried whole
soil. The COD balances vary about 15 % for a typical set; this is a
consequence of an error of at least ± 5 % in this measurement. Slurry,
diluted with wash water, has a COD of 7.1 g 02/L.
Table 4 - Moisture Content and COD
Sample Type Composition F%1 COD Fa Oo/kol
Whole Soil
Water
Large Particles
Soil Fines
Tar-like Residue
SIurry
16
53
31
89.6
0
trace
19.4
84.5
When an active microbial inoculum is combined with a slurry of soil fines, a
lag period of approximately 6 hr is observed. Acid production in shake
flasks and fermentations, and carbon dioxide generation by fermentations,
are not observed until after the lag phase. There is no loss of COD, as
might accompany volatilization or sorption. It is assumed that COD
attributable to biomass remains unchanged during the experiment, i.e.,
growth is negligible. The reduction of COD in the flasks inoculated with
live cultures is indirect evidence for substrate mineralization, as opposed
to physical uptake (sorption) by the biomass. The results of a shake flask
matrix study are summarized in Table 5. Flasks contain a working volume of
60 ml; total reaction time is 80 hr. The slurry has an initial soil fines
concentration of 30 g/L and a COD of 7.0 g 02/L; inoculum COD is 4.75 g
02/L.
Table 6 sumarizes results of a larger scale fermentation study. Reactor
working volume is 3 L and reaction time is 68 hr. The slurry has an initial
soil fines concentration of 45 g/L and a COD of 7.5 g 02/L; as in the shake
flask illustration, inoculum COD is 4.75 g 02/L. Initial reactor COD was
calculated to be 15.9 g 02; an experimental determination gave 18.9 g 02.
The difference is probably measurement error, due to the several phases
present, i.e., cells', contaminated fines, suspended tarry material and
dissolved PAHs. A volatilization loss study was carried out with the
fermentor. The COD of air sparged slurry did not change in 4 days; it
remained at 3.1 g 02/L.
326
-------
Table 5 - Typical Shake Flask Matrix Study
Initial COD [mg 02]
Final COD [mg 02]
COD Change [%]
COD Change [%]
{corrected for inoculum)
Final Dissolved TOC
[mg C]
[mg C/L]
COD Equivalent [mg 023
Live
Inoculum
310
204
-34
-50
8
131
18
Autoclaved
Inoculum
310
318
nil
nil
8
131
18
No
Inoculum
210
240
+14
+14
8
133
18
Table 6 - Fermentation Study
COD [mg 02]
21 hrs
13200
COD Change [%] 30
COD Change [%] 39
(corrected for inoculum)
Final Dissolved TOC
[mg C] 318
[mg C/L] 106
COD Equivalent [mg 02] 848
46 hrs
10200
46
61
315
105
840
68 hrs
6900
63
84
288
96
768
Figures 6 and 7 are gas chromatographs for the fermentation liquor
after 21 hrs and at the end of the experiment. The characteristics of
the soil slurry without medium or culture added are described in Figure 8.
The contents of the fermentor and the original soil slurry were extracted
with methylene chloride; the volume ratio was 5:2 for aqueous
suspensionrsolvent. Figures 6 and 7 show definite declines in the number
and size of peaks, especially those corresponding to low molecular weight
PAHs.
327
-------
Figure 6: 21 Hours
(nt.n
til?.91
i •••M. 31
!«I*.*f MM
7.211
t.flZ
j.irr
Figure 7: 68 Hours
II.Jt
»•.**
'*•"
*n« •-.
1.972
"tt'.t J
•'.I »
I?.9 9
J.r 1
:.i i
•I.5-.I
i.in
3..-JI
^Figure 8: Original Slurry
If.11
1.1?
]!'«•
a»ii.ti v* /.jjj
ICIH.^f vv 11.59:
ll-f.-|l« vv Il'll9
n*i,:i w i!*,
-------
CONCLUSIONS
1) Bench-scale shake flask studies are performed with slurries of soil
fines and mixed microbial seed. COD, corrected for the presence of
the inoculum growing at a trivial rate, is reduced by 50 % in
approximately 80 hours. Similarly, TOC is measured on settled [filtered]
aqueous phase and remains low throughout. The latter assays are a
reflection of limited hydrocarbon solubilities.
2) Larger-scale fermentations are carried out in 3-liter, stirred, air-
sparged reactors. Inoculum and nutrient medium are mixed with slurry.
Biodegradation is monitored by assays on samples of aqueous dispersions
and measurement of carbon dioxide generation rates. COD reductions
exceed 84 % in 68 hours.
3) Whole soil is separated into size fractions to characterize contaminant
distribution by soil constituent type and particle size. It is possible
to separate whole soil into larger, settleable particles [primarily sand
and silt], slurried fines, a clarifiable aqueous phase and a bulk organic
phase. Recovery of initial whole soil COD in the slurry and a tar-like
organic phase is nearly quantitative.
4) Analytical techniques have been developed and demonstrated with whole
soil, soil fractions, slurries of fines and filtered liquids. These
techniques are essential to the identification of contaminant species and
quantification of individual and total contaminant concentrations.
Assays are necessary to define initial and intermediate conditions and to
demonstrate that contaminant destruction [mineralization] by microbial
reaction is effective and approaches completeness.
329
-------
FATE AND EFFECTS OF RGRA AND CERCLA TOXICS
IN ANAEROBIC DIGESTION OF PRIMARY AND SECONDARY SLUDGE
by: Richard A. Dobbs, 2Rakesh Govind, 2Peter A. Flaherty,
3Thomas L. Crawford, 2Kaniz Siddiqui, ^arry M. Austern
xRisk Reduction Engineering Laboratory, United States
Environmental Protection Agency, Cincinnati, OH 45268
Department of Chemical and Nuclear Engineering,
University of Cincinnati, Cincinnati, OH 45221
3Department of Civil and Environmental Engineering,
University of Cincinnati, Cincinnati, OH 45221
ABSTRACT
Municipal wastewaters have been shown to contain toxic organics, many of
which are anthropogenic. Sorption onto solids is one of the primary means of
removal of these chemicals. This study investigates the steady-state fate and
effects of organic priority pollutants sorbed on primary and secondary sludge
during subsequent treatment in anaerobic digesters. The investigation
consisted of two separate studies of pollutants: volatile organics chosen from
the RCRA list, and semi-volatile organics from the CERCLA list. Simulating
typical digestion processes, three bench-scale, semi-continuous, complete-mix
units with a total volume of 40 liters each were operated with a solids
retention time of thirty days at 35°C. Primary and secondary sludge were
combined in equal weight ratios prior to feeding to the digesters.
Conventional operating parameters were monitored for the one control and two
organic-spiked digesters to assess differences in performance.
The fates of seven of the volatile compounds which showed consistent
behavior, including several chlorinated aliphatics and ethylbenzene, were
determined by purge and trap gas chromatographic analyses. Steady-state fates
of twelve of the semi-volatile organics, which included dichlorobenzenes,
phthalates, p-cresol and lindane, were determined by gas chromatographic/mass
spectroscopic analyses.
330
-------
INTRODUCTION
Municipal wastewaters have been found to contain a number of toxic
organics (1). In response, effluent regulations have also restricted the
particular pollutant discharge levels, thus increasing the interest in the
fate of specific toxic organic compounds in wastewater treatment processes.
One area which presents potential problems is the handling and disposal of
sludge. Sorption of toxics onto solids is one of the fundamental removal
mechanisms of pollutants from wastewater. Previous investigations have shown
that organic compounds tend to accumulate in sludge at concentrations several
orders of magnitude higher than the influent concentration (2,3). One study
reported some organic component concentrations in sludge greater than 10,000
times that found in the aqueous phase (4). A dilute toxics-laden wastewater
could potentially generate hazardous primary and/or secondary sludge, creating
more difficult and costly final disposal options such as landfilling or
incineration as a hazardous waste.
Anaerobic digestion is often utilized in the stabilization and reduction
of wastewater treatment sludge. A survey of ninety-eight municipal wastewater
treatment plants in the United States found that seventy-three incorporated
anaerobic digestion as a means of sludge stabilization and volume reduction
(5). Anaerobic treatment systems offer several advantages over aerobic
treatment in terms of lower energy requirements, higher process loading, and
the potential energy recovery in the form of methane gas production.
Anaerobic processes may also control the discharge of volatile compounds to
the atmosphere via biotransformation or capture of the volatilized material
with the off gas.
Research on the fate of toxic organic material in aerobic systems has been
extensive, yet relatively little work has been conducted to determine the
fates of organic toxics in typical anaerobic processes. Investigations
generally have been conducted with bench or pilot scale processes, serum
bottles with digester sludge, and serum bottle studies with specific or
enriched anaerobic cultures. Most studies have been conducted by spiking the
compound or compounds directly into the influent process stream or serum
bottle. This investigation, however, attempts to better typify actual
processes by determining the fate and effects of compounds already sorbed onto
the digester feed sludge. The goal of this study is to determine the steady-
state fates of several volatile and semi-volatile organic priority pollutants
in the anaerobic digestion of primary and secondary aerobic sludge, and to
determine the effects of these toxic sludge on the digestion process.
EXPERIMENTAL METHOD
AEROBIC PILOT PLANT
The United States Environmental Protection Agency (USEPA) has evaluated
the removal and fate of selected toxic organic compounds during primary-
activated sludge treatment of municipal wastewater (6). Volatile and semi-
volatile organics were investigated under separate studies with the same
aerobic treatment system. Two parallel pilot plants were studied: one
continuously spiked to allow for biomass acclimation, and the other
331
-------
Intermittently spiked to simulate unacclimated operation. The treatment
systems chosen as a representative municipal wastewater treatment plant design
consisted of primary clarification followed by conventional plug-flow
activated sludge treatment. Each system was operated with a sludge retention
time of 4.0 ± 0.3 days over a period of six months. Screened and degritted
raw wastewater from the Cincinnati Mill Creek Treatment Plant was used as feed
for the aerobic systems. Mill Creek is a combined residential/industrial
treatment facility.
Volatile and semi-volatile toxics studied were selected from the RCRA and
CERCLA lists of organic priority pollutants, respectively. A miscible mixture
of the compounds listed in Table 1 (volatiles) and Table 2 (semi-volatiles)
was used to spike the raw wastewater influent to the pilot plant at 0.25 mg/L
of each compound for both studies. The primary and secondary sludge from the
continuously spiked aerobic system were used as feed to the bench-scale T
anaerobic digesters.
ANAEROBIC BENCH-SCALE DIGESTERS
Three bench-scale digesters (see Figure 1) were constructed from
Plexiglass cylinders with quarter-inch thick walls. Influent and effluent
ports were constructed from polyvinyl chloride pipe. One-inch valves were
used at each port location to provide for sludge feeding and digester mixed-
liquor withdrawal. *A Plexiglass cup with approximate capacity of one liter
was threaded at the top of the inlet pipe to act as a funnel during digester
feeding. A long thermistor probe was inserted through the top endplate to
TABLE 1. VOLATILE ORGANIC COMPOUNDS IN RAW WASTEWATER SPIKE
Acetone
Atrazine
Carbon tetrachloride
Chloroform
Cyclohexanone
1,2-dichloroethane
1,2-Dichloropropane
2,4-Dimethylphenol
2,4-Dinitrophenol •
Ethylbenzene
Furfural
Methylene Chloride
Methyl ethyl ketone
Methyl isobutyl ketone
4-Ni tropheno1
Phenol
Tetrachloroethylene
Tetrahydrofuran
Toluene
1,1,1-Trichloroethane
1,1,2-Trichloroethane
Trichloroethylene
TABLE 2. SEMI-VOLATILE ORGANIC COMPOUNDS IN RAW WASTEWATER SPIKE
1,2-Dichlorobenzene
1,3-Dichlorobenzene
1,4-Dichlorobenzene
1,2,4-Trichlorobenzene
Nitrobenzene
1,3-Dinitrobenzene
2,6-Dinitrotoluene
E-Cresol
Hexachloroethane
Hexachloro-1,3-butadiene
N-nitrosodiphenyl amine
Dimethyl phthalate
Diethyl phthalate
Di-n-butyl phthalate
Butyl benzyl phthalate
(Bis) 2-ethylhexyl phthalate
Napthalene
4-Chloroaniline
Lindane
Dieldrin
332
-------
measure digester temperature. All fittings were sealed with silicone caulking
and each unit was pressure tested to insure air tightness. The gas
delivery/mixing system was constructed from one-eighth inch stainless steel
tubing and stainless steel fittings. The H-shaped sparger was placed one-half
inch above the endplate and levelled.
On the outside of the digester, a two-inch long, stainless steel tube was
inserted through the upper endplate to act as the digester gas outlet port.
Butyl rubber tubing was inserted over the inlet and outlet stainless steel
tubes with hoseclamps. The tubing ran from the gas recycle outlet port to a
gas pump and from the gas pump to the gas recycle inlet port. A glass tee was
inserted between the gas recycle outlet port and the pump to allow for the
escape of digester off-gases to a Tedlar gas collection bag. A stainless
steel tee with a septum and cap was placed in line with the gas collection bag
for withdrawal of gas samples for analysis.
BUTYL RUBBER
STOPPER
TEMPERATURE
PROBE
STAINLESS
STEEL TUBING
SPARGER
BUTYL
RUBBER SEAL
(TYPICAL)
r BRASS
SVAGELDK
VALVE
(TYPICAL)
GAS RECYCLE
PUMP
TO GAS
COLLECTION
BAG
61 cm
SIDE
VIEW
EFFLUENT
PORT
30.5 cm —
Figure 1. Schematic of anaerobic digester for volatile
and semi-volatile studies.
DIGESTER OPERATION
During startup for both studies, the three 40-liter digesters were seeded
with 30 liters of anaerobic biomass from the Mill Creek plant. The digesters
were cleaned out and reseeded between volatile and semi-volatile experiments.
Initially, all three digesters were fed one liter of unspiked primary and
secondary sludge for three or four days to achieve start-up. After start-up,
digester 1 continued to receive unspiked sludge from the Mill Creek plant,
while digesters 2 and 3 were fed sludge from the continuously-spiked aerobic
333
-------
pilot plant. The feed rate of one liter per day resulted in a 30 day solids
retention time. The digesters were maintained at 35 + 1°C in an environmental
room. Off-gases were recirculated through the sparger for 15 minutes every
six hours to keep the contents of the digesters well mixed. Conventional
parameter evaluations were performed biweekly to monitor digester biomass
acclimation and the approach of steady state. Steady state operation was
achieved after seven weeks for the volatiles study and five weeks for the
semi-volatiles study, after which analysis for the spiked organics began.
The daily draw and feed proceeded after digester biomass was mixed
sufficiently uniformly distribute the cpntents. One liter of mixed liquor was
purged from the effluent port and then, reintroduced through the feed port to
insure representative sampling. A second one liter sample was withdrawn and
used for the analysis of conventional parameters and organic compounds. One
liter of feed sludge was then added to the digester through the influent port.
FEED SLUDGE
For the volatile organics study, unspiked feed sludge for digester 1 were
gathered weekly from the Mill Creek treatment plant to act as control.
Primary sludge was drawn from the gravity thickeners and diluted to 4% total
solids (if necessary) with clarified primary influent. Secondary sludge was
collected from the return activated sludge line, thickened by means of a
perforated bowl centrifuge and diluted with the resulting centrate to 4% total
solids. Both unspiked sludge were then combined in equal volumes and the
resultant feed sludge was stored at 4°C in a large carboy until needed for
digester feeding.
Volatile spiked feed sludge for digesters 2 and 3 were gathered daily to
minimize component loss. The primary sludge was collected from the clarifier
of the continuously-spiked aerobic pilot plant. Thickening was achieved by
overnight settling at 4°C in completely-filled glass containers to avoid loss
of volatiles in the headspace. Secondary sludge was drawn from the waste
activated sludge line of the pilot plant and thickened by overnight settling
in glass jars at 4°C followed by centrifugation. After achieving 4% total
'solids, the spiked primary and secondary sludge were combined in equal volumes
and stored in glass containers at 4°C (with no headspace) until needed for
digester feeding. Slight losses of volatiles occurred during processing,
however all feed sludge were analyzed as fed to the digesters.
For the semi-volatile organics study, unspiked sludge were taken weekly
from the Mill Creek facility and spiked sludge from the T&E pilot plant.
Unspiked primary sludge was gathered from the gravity thickeners and diluted
to 4% total solids, when necessary, with primary influent. Secondary sludge
was gathered from the thickened secondary sludge line (after polymer addition
but prior to gravity-belt thickening). Return activated sludge was used for
dilution when necessary. Both primary and secondary sludge were combined in a
1:1 volume ratio to yield the final feed sludge for digester 1.
Semi-volatile spiked sludge for digesters 2 and 3 were gathered biweekly
from the pilot plant, as the danger of component loss by volatilization was
not significant as that for the volatile organics. Primary sludge was
thickened by overnight settling at 4°C in large Nalgene carboys. Secondary
sludge was thickened by the perforated bowl centrifuge and diluted with the
centrate to 4%. Both spiked sludge were combined in equal volume ratios prior
to storage at 4°C in a large carboy.
334
-------
DIGESTER PERFORMANCE
Conventional parameters were measured for the sludge feed, mixed-liquor,
and gas samples from each of the three digesters in order to assess digester
performance. Digester temperature, pH, and gas production were measured
daily. Chemical oxygen demand, alkalinity, volatile fatty acids, total and
volatile suspended solids, total Kjeldahl nitrogen, ammonia nitrogen, and
total and organic phosphorus were measured biweekly throughout the study.
Analyses of off-gases for methane and carbon dioxide content were performed
weekly.
For organics analysis, the primary/secondary sludge mixtures used as
digester feed were analyzed daily for the volatile organics or composited
daily and analyzed weekly for the semi-volatile organics. Mixed liquor was
collected for analysis once a week. A portion of the mixed liquor was
centrifuged, and the centrate was collected for analysis. For the volatiles
study, gas samples were collected and analyzed weekly. Gas samples for the
semi-volatiles study were analyzed near the end of the testing period.
ANALYTICAL METHODS
Conventional parameters were evaluated by EPA Methods (7). All analyses
were performed in duplicate, when possible, for each sample. Volatile organic
compounds were analyzed using EPA Method 601 (8) and EPA Method 602 (9). The
analytical system was composed of a Tracer LC2 Sample Concentrator, a Tekmar
Model ALS Automatic Laboratory Sampler, and a Tracer 585 Gas Chromatograph
with a Nelson Analytical 900 Series Interface to an IBM Personal Computer AT.
Digester off-gas samples were analyzed for specific volatile organics by
injecting a 5-cc sample directly into the purging chamber with a gas-tight
syringe. From this point, the methods used for gas analysis were identical to
those described above for the aqueous samples.
Semi-volatile organics analysis was performed by EPA Method 1625 (10).
Analytical equipment included a Varian gas chromatograph and an Incos 50 mass
spectrometer. Off gas was analyzed as described above. For those compounds
not quantifiable in the spectra, gas-phase concentrations of the semi-volatile
compounds were estimated by pure component Henry's Law constants evaluated at
35°C. The chemical component in the gas was assumed to be in equilibrium with
the mixed-liquor centrate.
Mass balance calculations were performed for each of the sampling events
from the sludge feed, mixed liquor, mixed-liquor centrate, and gas
concentrations, multiplying by the appropriate volumetric flows and gas
productions for each digester yielded the mass flows. The amount of component
sorbed on solids was determined by mass difference between the mixed liquor
and the mixed-liquor centrate. Fates of the specific organics were then
determined.
RESULTS AND DISCUSSION
CONVENTIONAL DIGESTER PARAMETERS
Conventional parameters are reported in Table 3 to document operation and
performance of all three digesters during the 7-week test period for the
335
-------
volatile organics study. Parameters for digesters 2 and 3 were very similar,
and are thus presented as an average for the spiked system. Results in Table
3 indicate that all digesters functioned within the limits of normal operation
over the course of experimentation. Of particular interest is the methane
content of off-gases from the digesters. As shown, the difference in methane
production (per gram VSS destroyed) between the spiked and control digesters
is only 13%. Further, for each digester the methane produced per gram COD
reduced was above the theoretical value (11) of 0.35 liters(at STP), again
signalling normal operation. The total gas production from each digester was
also similar, with a difference of 6% between the spiked and unspiked units.
Conventional operating parameters for the control and spiked digesters
during the 9-week test period of the semi-volatile organics study are
presented in Table 4. As with the volatiles study, the operational variations
TABLE 3. SUMMARY OF ACCLIMATED DIGESTER CONVENTIONAL OPERATING PARAMETERS
FOR RCRA VOLATILES STUDY
Parameter:
Digester
CONTROL
SPIKED
CONTROL
SPIKED
Temp. (C) 35.1 ± 0.8
pH 7.34 ± 0.07
TSS (mg/L) 29,200 ± 1,000
VSS (mg/L) 13,600 ± 1,500
COD (mg/L) 22,000 ± 3,600
TEA (mg/L) 3,560 ± 100
VFA (mg/L) <50
TKN (mg/L) 895.0 ± 40.0
TP (mg/L) 51.0 ± 4.0
GAS (Lgjp/d) 14.0 ± 3.6
CH4 (L/gVSS) 0.62 ± 0.02
CH4 (L/gCOD) 0.41 ± 0.07
35.2 ± 0.8
7.32 ± 0.12
33,200 ± 2,500
15,300 ± 2,200
24,000 ± 4,000
3,610 ± 320
<50
860.0 ± 25.0
28.0 ± 9.0
13.1 ± 3.8
0.71 ± 0.11
0.39 ± 0.17
...
39,400 ± 3,500
24,500 ± 2,100
42,300 ± 5,700
675 ± 80
360 ± 90
200.0 ± 45.0
42.0 ± 4.0
...
42,700 ± 4,700
24,700 ± 6,400
44,000 ± 9,500
1,360 ± 515
470 ± 195
175.0 ± 25.0
22.0 ± 5.0
— — —
TABLE 4. SUMMARY OF ACCLIMATED DIGESTER CONVENTIONAL OPERATING PARAMETERS
FOR CERCLA SEMI-VOLATILES STUDY
Digester
Parameter: CONTROL SPIKED
Feed
CONTROL
SPIKED
Temp. (C)
PH
TSS (mg/L)
VSS (mg/L)
COD (mg/L)
TBA (mg/L)
VFA (mg/L)
TKN (mg/L)
TP (mg/L)
GAS (Lsxp/d)
CH< (L/gVSS)
CHA (L/gCOD)
35.1 ±
7.08 ±
31,000 ±
17,400 ±
30,700 ±
4,150 ±
<50
709.1 ±
87.9 ±
12.1 ±
0.70 ±
0.48 ±
0.8
0.03
1,400
1,100
9,400
150
123.1
16.9
2.1
0.12
0.25
32
17
35.2 ± 0.8
7.07 ± 0.03
1,600
800
4,300
170
800
700
27,400 ±
3,670 ±
<50
569.2 ± 93.9
23.9 ± 5.8
10.7 ± 1.4
0.65 ± 0.12
0.34 ± 0.09
43,600
30,200
55,900
560
890
4,800
3,000
12,600
120
220
181.0 ± 43.1
88.2 ± 20.6
43,500 ± 5,400
28,700 ± 3,000
49,700 ± 8,900
450 ± 140
570 ± 170
90.8 ± 22.3
19.1 ± 5.7
336
-------
between the two spiked digesters were slight, therefore parameters are
tabulated as an average of both. No major operational difference between the
spiked and control digesters is seen. Methane production was near the
theoretical value of 0.35 l(STP)/g COD destroyed for all digesters. Average
daily gas production from the control and test units differ by 12%. An
increase in gas production from digester 1 near the end of the test period is
attributed to an increase in COD in the control feed, while the COD of the
spiked feed and spiked units' gas production remained relatively constant. As
with the volatiles study the semi-volatile spike does not appear to have
hindered digester operation.
FATE OF CHEMICAL COMPONENTS
The specific fate of each compound attaining steady-state conditions is
summarized in Table 5 according to treatment mechanism. Volatile and semi-
volatile concentrations in samples from the two spiked digesters were averaged
together in the mass balance calculations. For both studies, it was found
that weekly pollutant concentrations, especially in the sludge feed, varied
considerably in some instances. Fluctuation in the feed concentration was due
to variability in the background chemical concentrations in the raw wastewater
and the method of the feed preparation. An additional (and unquantifiable)
complication in assessing the fate of individual components is that some of
the degradation products are also components of the spike mixture (i.e.,
tetrachloroethylene to trichloroethylene). Finally, it should also be
emphasized that accurate analytical data for specific components in an
anaerobic digester matrix are difficult to obtain. In spite of these
difficulties, the experimental results provide useful information on the fate
of sorbed organic toxics in anaerobic digesters.
For the volatile organics study, eleven of the compounds originally spiked
into the raw wastewater were consistently present in the digester feed sludge.
Other compounds in the original spike were either not present in the feed or
were present sporadically or in trace quantities. Steady-state conditions
were achieved during the test period for seven of these volatile compounds,
shown in Table 5. Biodegradation was the most significant removal mechanism
for all of these compounds except 1,2-dichloropropane. As would be expected,
volatilization of the components into the digester off-gas was also a primary
removal mechanism. Four other compounds, including chlorobenzene, chloroform,
1,2-dichloroethane, and toluene, initially accumulated in the digesters, but
all showed evidence of degradation by the end of the testing period.
Analytical results of the semi-volatile organic analysis showed twelve
compounds exhibiting steady-state behavior. Degradation was the predominant
removal mechanism for lindane, dibutyl phthalate, butyl benzyl phthalate, and
2,-cresol were almost completely degraded. The other compounds, while
generally showing significant biotransformation, tended to sorb or remain
sorbed on the solid material. Results of off gas analysis showed only the
dichlorobenzenes, trichlorobenzene and hexachloro-1,3-butadiene to be
quantifiably present. As expected from the high Kows of the components,
solubilization was not significant except for 4-chloroaniline. As mentioned
previously, some components may biotransform into some of the other spiked
components, thus affecting the apparent fates. Of note here is 1,2,4-
trichlorobenzene, which has been shown to degrade to 1,2-dichlorobenzene (12).
337
-------
SUMMARY
The performance of the anaerobic digesters were not hindered by either the
volatile or semi-volatile organic spiked sludge. Gas production and methane
content for spiked and unspiked systems were nearly identical. All parameters
were within typical operating limits. Aliphatic volatile compounds were
significantly degraded or otherwise volatilized. Removal of volatile organics
by sorption was generally not a significant mechanism. The semi-volatile
compounds tended to sorb onto the mixed-liquor solids or be degraded.
Solubilization and volatilization generally did not play key roles. For both
studies, all compounds showed some evidence of degradation. Anaerobic
digestion is a viable method of treating toxic-laden primary and secondary
sludge, however, more research is necessary to determine specific fates of
individual compounds in actual digestion processes.
ACKNOWLEDGEMENT
This work was performed jointly by the USEPA and the University of
Cincinnati under cooperative agreement no. CR812939-01.
TABLE 5: STEADY-STATE FATES OF SPIKED ORGANIC COMPOUNDS
I. VOLATILES:
Fate Mechanism Average
SOL VOL SORB DEC FEED LOAD
(%) (%) (%) (%) (mg/kg)
1,1,2-Trichloroethane
1,1,1-Trichloroethane
Trichloroethylene
Tetrachloroethylene
Methylene Chloride
Ethylbenzene
1,2-Dichloropropane
II. SEMI-VOLATILES:
0
0
0
0
0
0
5
0
3
8
8
14
26
46
0
2
0
1
3
14
10
100
95
92
91
83
72
39
39
64
84
37
33
10
14
144 ± 45
61
20
25
61 ± 14
Lindane
Butly benzyl phthalate
Di-n-butyl phthalate
2-Cresol
4- Chloroaniline
Napthalene
1 , 3-Dichlorobenzene
Bis (2-ethylhexyl) phthalate
1,2, 4-Trichlorobenzene
Hexachloro -1,3 -butadiene
1 , 2-Dichlorobenzene
1 , 4-Dichlorobenzene
0
1
1
6
31
4
3
5
3
1
4
4
0
0
0
0
0
4
10
0
5
8
14
16
2
2
3
20
34
65
61
69
66
73
66
68
98
97
96
74
34
27
26
26
26
18
16
13
490
290
270
13
12
230
320
1100
750
1100
280
275
± 140
± 190
± 140
± 6
± 3
± 40
± 110
± 700
± 320
± 560
± 70
± 60
SOL—solubilized(amount in mixed-liquor centrate) VOL=volatilized
SORB-sorbed on solids DEG=degraded FEED LOAD=mg organic/kg dry wt. solids
338
-------
REFERENCES
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Hannah, S. A., and Rossman, L. Monitoring and analysis of hazardous
organics in municipal wastewater--a study of twenty-five treatment
plants. Paper presented at the Seminar on Hazardous Substances in
Wastewater, Toronto, Ontario. November 3, 1982.
Hannah, S. A., Austern, B. M: , Eralp, A. E., and Dobbs, R. A. J. WPCF.
60(7): 1281, 1988.
Dobbs, R. A., Jelus, M., and Cheng, K. Partitioning of toxic organic ''
compounds on municipal wastewater treatment plant solids. EPA/600/D-
86/137, U.S. Environmental Protection Agency, Cincinnati, Ohio, 1986.
Malz, F. Heavy metals and chlorinated hydrocarbons in sewage sludge. In;
Proceedings of the Workshop on Removal of Chlorinated Hydrocarbons and
Heavy Metals from Wastewater by Advanced Treatment Systems. IUPAC
Commission on Water Chemistry (VI.6), Frankfurt, FRG, 1986.
Sludge handling and disposal practices at selected municipal wastewater
treatment plants. MCD36. U.S.Environmental Protection Agency, Office of
Water Program Operations, Washington, DC, 1977.
Bhattacharya, S. K., et al. Fate and effects of selected RCRA and CERCLA
compounds in activated sludge systems. Paper presented at the 15th
Annual EPA Research Symposium, Cincinnati, OH. April 11, 1989.
EPA methods for chemical analysis of water and wastes.
U.S. Environmental Protection Agency, 1979.
EPA-600/4-79-020.
Test methods - Methods for organic chemical analysis of municipal and
industrial wastewater: Purgeable halocarbons - Method 601. EPA-600/4-
82-057. U.S. Environmental Protection Agency, 1982.
Test methods - Methods for organic chemical analysis of municipal and
industrial wastewater: Purgeable aromatics - Method 602. EPA-600/4-
82-057. U.S. Environmental Protection Agency, 1982.
EPA Method 1625 - Semivolatile organic compounds by isotope dilution
GC-MS. Federal Register. 49(209): 1984.
Parkin, G. F., and Owen, W. F. Fundamentals of anaerobic digestion of
wastewater sludge. J. Environ. Eng. 112(5): 879, 1986.
Tsuchiaya, T. and Yamaha, T. Reductive dechlorination of 1,2,4-
trichlorobenzene by Staphylococcus epidermidis isolated from intestinal
contents of rats. Agric. Biol. Chem.. 48: 1080, 1984.
339
-------
FATE AND EFFECTS OF SELECTED RCRA AND CERCLA
COMPOUNDS IN ACTIVATED SLUDGE SYSTEMS
by
Sanjoy K. Bhattacharya, Rao V.R. Angara
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221
Sidney A. Hannah, Dolloff F. Bishop, Jr.
Richard A. Dobbs, and Barry M. Austern
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Two separate studies were conducted to investigate the removal and fate
of 28 selected RCRA compounds (0.25'tng/l of each compound) and 19 selected
CERCLA compounds (0.5 mg/1 of each compound) in conventional activated sludge
treatment. In each study, two pilot-scale (35 gpm) activated sludge systems
(SRT: 4 days for RCRA study and 8 days for CERCLA study) were operated in
parallel at the USEPA Test & Evaluation Facility in Cincinnati, Ohio. One
system was spiked continuously with either RCRA or CERCLA toxics to produce
an acclimated biomass; the other was spiked intermittently with the same
toxics and sampled to determine performance under unacclimated conditions.
The selected RCRA or CERCLA compounds did not cause any adverse effects on
COD and SS removals. The concentrations of organics (RCRA study) in the air
emissions indicated that the chlorinated aliphatic solvents were essentially
volatilized into the plant air emission stream, whereas the aromatic volatile
benzenes were substantially degraded. Additional work is planned to attempt
to reduce the analytical variability encountered in these studies.
340
-------
INTRODUCTION
A survey of publicly owned treatment works (POTW) showed that concentra-
tion of priority pollutants in the influent wastewater to many of these
plants exceeded the allowable concentrations for these chemicals (1).
Petrasek, et al. (2) studied the fate of 22 toxic organics in wastewater
treatment plants. They reported that a typical POTW significantly (up to
90%) reduced the concentrations of most of these compounds. However certain
compounds were present in the activated sludge effluent in relatively high
(20-30 jug/L) concentration. Hannah, et al. (3) investigated the comparative
removal of priority pollutants by six biological and physical-chemical treat-
ment processes. They reported that activated sludge process provided the
best results. A further review of the literature indicated that only limited
data are available for many priority pollutants.
In this study, the removal and fate of selected RCRA and CERCLA toxic
organic pollutants were evaluated with two pilot-scale activated sludge
systems fed municipal wastewater at the USEPA's Test and Evaluation Facility
in Cincinnati, Ohio. The 28 RCRA (semivolatile and volatile) and 19 CERCLA
(semivolatile only) chemicals spiked into the systems are shown in Table 1.
The selected RCRA and CERCLA toxics were spiked into the raw wastewater in
two separate test periods.
EXPERIMENTAL SYSTEMS AND TESTING APPROACH
The twq activated sludge systems were operated at a flow rate of 35 gpm
and a hydraulic retention time (HRT) of 7.5 hours. An operational sludge
retention time (SRT) of 4 days was used in the RCRA study period. In the
CERCLA study period, the SRT was 8 days. Each compound was spiked at 0.25
mg/L for the RCRA study and at 0.5 mg/L for the CERCLA study. The operating
conditions and design characteristics for the two systems used in the study
are given in Table 2. Both RCRA and CERCLA studies were performed with an
acclimated (continuous addition of toxicants) and an unacclimated (intermit-
tent spiking of toxicants) system. These systems were operated in parallel.
To sample from the air space above the primary clarifier, the units were
covered and vented through a duct to the roof. An air sweep equivalent to a
5 kilometer per hour wind was maintained over the surface of the primary
clarifier by exhausting air at 14,000 liters/min. The aeration basin was
fitted with an air tight cover and the off-gas was also vented to the roof.
Air flow in the aeration basins averaged 5,600 liters/min.
341
-------
TABLE 1. RCRA AND CERCLA TOXIC ORGANIC POLLUTANTS
RCRA Study Period
CERCLA Study Period
acetone
cyclohexanone
furfural
2-butanone
4-methyl-2-pentanone
tetrahydrofuran
carbon tetrachloride
chlorobenzene
chloroform
1,2-di chloroethane
1,2-dichloropropane
methylene chloride
tetrachloroethylene
trichloroethylene
1,1,1-tri chloroethane
1,1,2-tri chloroethane
ethyl benzene
toluene
total xylenes
bis(2-ethylhexyl) phthalate
butyl benzyl phthalate
1,4-dichlorobenzene
2,4-dimethylphenol
2,4-dinitrophenol
naphthalene
nitrobenzene
4-nitrophenol
phenol
1,2-dichlorobenzene
1,3-dichlorobenzene
1,4-dichlorobenzene
1,2,4-tri chlorobenzene
nitrobenzene
1,3-di ni trobenzene
2,6-dinitrotoluene
p-cresol
4-chloroaniline
hexachloroethane
hexachlorobutadi ene
dimethyl phthalate
diethyl phthalate
dibutyl phthalate
butyl benzyl phthalate
bis(2-ethylhexyl) phthalate
naphthalene
lindane
dieldrin
Automated analytical procedures were used for the conventional pollu-
tants (COD, BOD, NH4-N, N03-N and TKN) and 6C/MS procedures were used for
the toxic organic compounds. RCRA samples were analyzed by a contract
laboratory (PEI Associates Inc., Cincinnati, OH). Air samples were collected
in stainless steel canisters, and were analyzed by GC/MS. From these data,
masses of each RCRA compound stripped during the sampling event were calcu-
lated. Sludge and liquid samples were also analyzed by GC/MS according to
approved USEPA methods (4). Semi-volatile RCRA compounds were extracted from
the samples using continuous liquid-liquid extraction. Prepared portions of
the extracts were injected into the GC/MS for analysis. The semi-volatile
CERCLA compounds were analyzed following Method 1625 (5). Analyses were per-
formed at the EPA, RREL. Details of the analytical procedures were reported
elsewhere (6).
342
-------
TABLE 2. OPERATING CONDITIONS AND DESIGN CHARACTERISTICS OF
THE PILOT SYSTEMS
I.
II.
III.
IV.
Design Flow
= 2.2 Us
= 191 m3/d
Primary Clarifiers - Diameter = 2.9 m
Weir Diameter = 2.8 m
Surface Area =6.8 m2
Surface Overflow Rate =28.0 m3/m-d
Aeration Basins - L:W:D = 5.4 m:3.0 m:3.6 m
Surface Area = 16.3 n£
Volume = 59.7 m3
Hydraulic Residence Time = 7.5 hrs.
Secondary Clarifiers - Diameter = 3.6 m
Surface Area = 10.4 n£
Surface Overflow Rate =18.4 m3/mz.d
Three tests (sample collection events) were performed during the RCRA
study period. For the CERCLA study, eleven tests on the continuously spiked
(acclimated) system and 4 tests on the intermittently spiked (unacclimated)
system were performed.
RESULTS
The presence of the spiked toxic organics in the wastewater produced no
major adverse effects on the treatment of conventional pollutants. Average
removals of conventional pollutants in the pilot systems during the two
studies were between 94 and 97 percent for SS and between 81 and 88 percent
for COD (Table 3). In the RCRA study period nitrification in the activated
sludge processes produced average NH4-N reductions between 76 and 81
percent. In the CERCLA study, the NH4-N removal was between 88 and 98
percent. The CERCLA toxics (0.5 mg/L) did interfere with nitrification in
the acclimated system.
Substantial variability occurred in the reported results with some toxic
compounds, especially in the RCRA study period. Table 4 lists average
measured concentrations of the selected RCRA organics in wastewater and
sludges for the acclimated system. The difference between the concentration
of most toxics in the spiked wastewater feed and primary effluent was very
low. The primary sludge showed enhanced concentrations of the two phthalates
and naphthalene along with reduced concentrations in primary effluent indica-
ting adsorption onto sludge solids. Four compounds (tetrahydrofuran, 1,2-
dichloroethane, methylene chloride and 1,1,2-trichloroethane) were present in
343
-------
TABLE 3. AVERAGE PERCENT REMOVALS OF CONVENTIONAL POLLUTANTS
DURING THE RCRA AND CERCLA STUDIES
Acclimated System
% Removal Standard
Deviation
Unacclimated System
% Removal Standard
Deviation
RCRA Study
SS
COD
NH4-N
CERCLA Study
97
82
76
4
7
19
97
81
81
3
8
18
SS
COD
NH4-N
95
88
88
3
4
14
94
87
98
6
10
3
the secondary effluent stream in high concentrations (between 95 and 140
ug/L) indicating poor removals of these organics. Five other compounds
(cyclohexanone, furfural, 2,4-dimethyl phenol, 2,4-dinitrophenol and 4-
nitrophenol) were not evaluated due to inconsistent results. The average
removals of the toxic compounds in the RCRA study are summarized in Table 5.
The removal of RCRA organics with primary treatment was between 3 and 44
percent. The total removal was between 36.6 and 99.0 percent. The calcu-
lated percent stripped for the individual volatile compounds varied from 1 to
139 percent. No air analyses were performed for the semivolatiles. Bio-
degradation of a compound was estimated by subtracting the measured removals
by adsorption and stripping of the compound from the total removal. The
estimated biodegradation was between -42 and 97 percent. A negative bio-
degradation indicated inconsistent mass balance and the problems of
estimating biodegradation using this approach. Biodegradation appeared to be
the predominant removal mechanism for the polar solvents, and the aromatic
volatiles (e.g., toluene: 72%, xylenes: 66%, chlorobenzenes: 60%). The
unacclimated system (e.g., biodegradation of toluene: 56%) showed similar
results and no significant advantage of acclimation was observed (6).
Table 6 lists the average concentrations of the CERCLA organics. Five
compounds (1,2,4-trichlorobenzene: 89 ug/L, 2,6-dinitrotoluene: 125 ug/L,
p-cresol: 156 ug/L, lindane: 198 ug/L, and dieldrin: 99 ug/L) showed high
concentration in the secondary effluent (Table 7). The average removals of
the toxic compounds in the CERCLA study are summarized in Table 7. The
344
-------
s:
Oi
co co
0 -Zi
i — i ZD
K
CH. CO
O LU
o: co
o
o: LJ_
ll_ °
CD LU
CD
CO ==C
O LU
i — i ^*
£ «5"
•y* LU
LU 1 —
O CO
Z >-
O CO
0
Q
LU LU
CD 1 —
LU >— l
•=C 0
CO UJ
UJ
tD
a:
UJ
<;
CkC O
Q H-
3: <
I— >
to uj
o
UJ
1
g 0
C> 1—
z ^
00 UJ
o
UJ
C3
C£.
%
UJ
5
z
II
II
II
II
II
II
II
II
II
II
II
II
—1 II
-^ 11
O> II
D II
11
11
II
II
II
II
II
II
II
II
II
II
—1 II
^ II
a> u
n n
n
n
n
n
n
n
u
u
u
n
u
ii
n
n
—i u
^, n
CD II
D II
II
11
II
II
II
II
II
II
II
II
II
II
||
II
-J II
^« II
a> n
=i II
n
u
ii
n
n
n
ti
u
it
n
u
It
n
-i u
•-- II
O) II
3 II
II
II
II
II
II
II
jj
II
II
It
II
It
II
II
II
II
II
II
II
II
II
It
II
11
II
II
II
II
tl
II
II
II
OJ
«— OJ CO
•<»• %o oj
OJ «-
«— ro in
JO 0 ,0
o in ro
CM
0s *"~ *O
ro o CM
CO
CM CO CM
SCO NO
«-
ro co N-
*- CO CO
*—
N- --* co
*»j" ~ ro ro
v± «— CM
c
OJ
4-J
c
C 0-
C Q) CM
OJ C «
O Q) OJ C >-
c JT t- m j=
O O D *J -M
4-> — ' M- D 0)
. 3 « i
<£ CJ u. CM ^f
CM
1
st
co
OJ
tn
OJ
in
OJ
OJ
o-
1
C
m
L.
D
M-
o
|
0)
t_
4-*
Q)
in
1
«-
ro
o
*~
JJ
j:
1
C.
O
^1
2
4J
0)
1—
c
OJ
o
ro
s.
o^
in
OJ
»-
ro
o
O
O
.r:
0
Q
i
OJ
sf
o
CO
o
OJ
o
OJ
Nf
in
R
OJ
OJ
o
in
OJ
c
1
o
t.
o
u
Q
OJ
' R ~* -
co -o ro
ro o o
in o in
t- OJ OJ
ro sj- ro
>o in r^
«— OJ OJ
oj ro oj in
r^ o in in
o »-
r^ -o m co
*~
co sf -o ro
o- o- co ro
ro «— «— oj
i? 5 in 5
sf «- «- O
ro *— o ^
-d- oj oj oj
• ai o
O JS — ' t-
-• 4J X 0
jr 4) .c — •
O O 4J JZ
c. v o
QJ O O "^
C — • c- t-
CJ JZ O H-
— ' U — ' <
>* (0 J= «-
JZ L. U *
01 O tl
5E 1— 1— «-
O OJ
ro
fe M
in ^~
•o o
r- OJ
r- OJ
1C
OJ . OJ
0 O
OJ «—
Sf !J
^ t>
c
ro
o co
S. x± <— c\j . , t\j
' ^_ ]
1
«— O CD O CD *—
1
. .
1
1
>t in in in in in
»•$• o« N- ' 'CM co *o
CM ro . in ro o
T-
.
(
f
in ro in to ro CD
o ro o- ro oj o
i— «— CM
ro ^o ro *— r^- CD
ro fM ro NQ >d- co
-•
•
.
•
o CM CM *4- ON . in
OJ vO O -
ro
_.
03 .
JZ
-M
.C
Q. . OJ N C O
x ^: c .— ' t- >. O 0) C C
f-v>ni^c_c~aj-UOQ4J O-M O
1 >- 1 - 1 i .C L- ••- C
yj-+j^j--J--st O. *J Z
-------
CO
o- —
l— l CO
2: 2:
— >
5:
OH-
— °°
^C L fJ
S1-
Q£ CO
1 1 l ^t^
o o
CO LU
—I CD
11
z: LU
it
.11. 1
-
II 0
II I—
II
II
II
!!
n
n
u
u
H
IS »
ii s
ii »_
n a:
It Q.
n
ii
n
ii
ii
11
n
n
n
n
n
it
n
it
n
it
ii
ii
n
ti
u
u
u
11
u
II
tl
n
H
"
1 1
| REMOVAL STANDARD |
(%) DEVIATION |
1 1
[REMOVAL STANDARD
(%) DEVIATION
1
| REMOVAL STANOARt
| (%) DEVIATIOI
& 1
STANDAF
DEVI ATI
_,
^
o *!
UJ
g I
«C t—
0 <
^c ^>
1— UJ
CO O
— I
o S
UJ
C£
Ul
z
II
ii in t- NO
it
it
it
ii
n
n
n
ti
n
n
ll oj -si- in
II ON ON ON
II
II
II
II
II
II
II
II *~ "~ *~
II
It
II
II
II
It
II
II
II
ll ro <— oj
it
ii
ii
ii
u
It •
tl
n *~
it
ll
it
it
u
ll
ll
It OJ T- «-
II
II
II
It
ll NO co in
u oj oj ro
u
u
u
ii
u
HO CD in
UN. in S-
jl ON ON ON
II
II
II CO T— CO
II NO «— 0
ll oj ro NO
ii
ll
it
ll
it
u
IIO 00
II NO -.
II C J= t- JC
ii w — < H- D a!
Ii 8 S. D ? ?
It
OJ
O
t_
o
CO
1 —
NO
i
-
in
ro
£
sf
^_
OJ
ro
ON
O
ON
0
•J-
OJ
•M
CO
CD
JCZ
4->
•>.
C
OJ
J.3
D
CO
ro c\j vQ c\J
.
(\J CM *— T—
«— in in o
ro ro N. CD -
O O O o
*O OJ f\J *—
S- >d" N" r—
ro N- ro o
co NQ -j- ro
• O O
i . . Ji t_ .~ C
"^ *^«. "^ (D 't- ^ .C
T— OJ OJ Z Z -J- CL
I
I
1
1
1
I
1
1
1
1
1
1
1
I
J
1
)
1
I
1
1
1
t
1
I
1
1
1
1
t
1
1
1
:
|
I
I
1
1
I
1
I
1 CD
f
l Crt
l
1 (U
1 4-»
(D
f >
I "-
t O
I CU
1 *-*
t M
1 03
I 13
t +
I
t I-
1 E
t ••-
1 f-
1 Q_
1
t
t
( 0)
! f
) — •
1 C/)
1
1 *
1
t
I
1
1
r
t
i
l
i
i
l
i
i
l
i
i
l
i
i
i
l
i
l
t
f
o
o
-------
1
u_
co
i— t CO
CD
O &
ii ^^
i— >
A? ^j
UJ UJ
CJ 1—
o >-
o co
LU Q
CD UJ
fff |— •
Dig
UJ S
^> i — i
0
UJ
ca
r
ii
u
n
u
n
n
n
u
ii
u
it
it
n
n
u
it
u
n
n
ii
u
n
u
n
n
u
n
n
n
n
n
ii
u
n
u
n
n
n
n
n
n
u
u
n
u
u
n
n
n
it
n
n
u
n
u
n
n
n
u
n
n
n
ii
n
n
n
n
n
n
n
n
u
u
u
n
!!
jj
j!
II
Jj
U
II
II
It
11
It
'j
II
II
ii
SECONDARY SLUDGE
PRIMARY SLUDGE |
1
EFFLUENT |
1
>-
a
Q 11
II
II
UJ It
(D —1 II
K ~O> II
UJ 3 II
> II
II
t— UJ II
CO O II
II
UJ II
0 _J II
ce. 01 n
UJ 3 II
> tl
< it
n
Z II
Q O II
II
h- LU II
CO O II
II
UJ II
C3 —I II
< — II
UI 3 II
> II
< II
II '
II
II
II
II
II
II
It
II
II
II
II
LU II
CM ro
ro in
in co
o ro
ro in
- <*^ *^ ^ *~~ N- ^O
fo «— *sj~ in «— in in
^- r— N-
CM«-inNOCQ COS-O
NO-J-O>OCO OfMO
O «— in ^a- «— N- -o
«- ^- ro
in QS ON r**- ^f ^o *o ^~
«— «— rocoo mco«—
o O* o NO in co
sjO^*oOi>fs- N-O>O-
coco^ro>o ru <— **
tn^fON-NO fMN-in
T— «— m 0* o in NO
-r -j-ro-st
rONOcor^-cM CO^NQ
cviin-4-OJOj r^irorvj
rONOororo rotoo
m-ON-coco <— in ON
ONN-ONC\JON rOf\Jf>
roin*4-"J'ro -vf-^d-in
UJ
-~i
UJ -
< < _I _J Q. X
*— 31 ^ «4 UJ
CO 3= 1— H- >- -J UJ
CZ Q. D_ 2 3= UJ
O -J UJ h- -J 2
CJi— rct-_jfMj-
to
c
i
S
*
347
-------
II
II
II
II
II
!! ~
n
n
n
n
n
n
n
.
CO
0
co
-
uj co N
°"Q |i
LU LU M
^S£ "
LU I
0
CO
e£
It
1—
_j
z
*~*
UJ
>
g
o:
•K
<
O
1—
«I
S
CL
UJ
z:
o
E
LU
o:
g
«c
o
z
UJ
a
&?
g
i—
<
>
MJ
*£
s^
g
§
>
UJ
0
s«
g
1—
UJ
o
II
II
11
11
11
II
II
II
II
It
It
II
II
II
II
II
11
11
II
II
11
|j
11
ii
II
1!
ii
n
n
n
n
it
ii
n
n
n
n
n
n
n
n
n
n
n
n
n
"
H
||
II
11
II
II
II
II
II
II
II
II
II
II
O
to
0.
o
CM
ro
•*
-3-
ro
o
f-
tr-
*—
in
CM
ca
m
0
a_
UJ
z
LU
Z
LU
g
3:
0
a
e
o
vO
,_
in
O
CO
CO
o
0
o
in
ro
CO
ca
0
a:
o_
UJ
z
UJ
IM
Z
UI
g
0
a
o
&
0
N.
CM
in
ro
in
o
*o
,_
ro
,„.
^3-
CO
ca
0
ee.
o_
UJ
LU
IM
UJ
g
0
Q
^
CO
•J-
CM
N.
ro
CO
1^
«-
in
CO
o
in
CM
N.
CM
LU
Z
Si
z
LU
CO
O
g
O
en
t-
1
CM
^—
ro
c>
r-
^_
<0
CM
CO
CM
O
CO
CM
O.
>o
UI
UI
IM
Z
UJ
m
o
h-
z
r^
ro
N.
§
co
o
o
o>
LJJ
UI
IM
Z
UI
CO
o
1—
z
a
i
ro
^
CO
ro
r-
in
in
CM
ro
in
CM
S
in
co
ro
ro
CM
LU
z
LU
3
O
t—
O
1—
z
1
CM
CM
O
•a
,-
CM
O
CM
O
R
_i
o
LU
C£
U
Q.
co
IS
,-
CM
,_
5Q
o
So
CM
in
ro
si-
,_
CM
LU
_J
Z
O
o
(J
^^
*
*
o
o
^~"
JUt
*
4c
O
in
CM
«-
fe
CM
^O
ro
^.
ro
UI
|
i—
LJJ
O
o
_J
3:
0
1
CO
o
CM CO CO ^t
v- ro CM
f^~ "^ CO *O
co ** i
•*3" T— in i
i
i
i
i
i
i
i
UJ 1
h- 1
_J I
UJ - I
CL X 1
UJ 1
_J 3T 1
>- — 1 LLJ 1
rsl >- z I
Z 3= HI 1
UI 1— — J -SS. \
03 LU < LU — 1
—1 CM 1— - N-^ 3= Q _J |
t- i-, < ^-. •-• i
CO CO Z -J O 1
1
1
1
1
1
1
1
1
1
1
i
i
i
i
i
i
i
i
i
i
t
i
i
i
i
i
i
i
i
i "
i
i
i
i
i
i o
1 O)
1 TJ
1 3
1 (/)
\
\ TJ
1 — •
1 *-• >. (D
1 - O t_ — •
i
1 4-« O
.^ •*- TJ
U I_
a. a. .
TI *J "c
3 O O
-^ 1—
a> *
* *
* * *
_
,
,
.
.
•
-
' ,
-
.
'
,
348
-------
removal of CERCLA organics in primary treatment with the acclimated system
was between 3.5 and 79.0 percent. The total removal varied between 55.9 and
98.1 percent. Biodegradation was estimated by subtracting the removal by
adsorption from the total removal of a particular compound. The extent of
biodegradation varied between 28 and 100 percent. The unacclimated system
also exhibited similar removals. Biodegradation values were similar for both
acclimated and unacclimated systems (e.g., naphthalene: 79% and dimethyl
phthalate: 85%). Like the RCRA study, no significant advantage of acclima-
tion was observed for the CERCLA compounds (6). In the CERCLA study period,
the amounts of organics found in the complex primary sludge samples were
substantially lower than the measured removals across the primary process.
Due to the analytical variability encountered in these studies, additional
work has been planned.
CONCLUSIONS
The following conclusions were drawn from this study:
1. The polar solvents and aromatic volatiles were biodegraded to a great
extent. For example, toluene exhibited 72 percent and total xylene showed
66 percent biodegradation.
2. A significant amount of chlorinated aliphatic solvents may be volatilized
from an activated sludge system. The percent stripped varied between 1
and 139.
3. Pesticides (lindane and dieldrin) were removed by both adsorption onto
primary and secondary sludge and biodegradation in the secondary tank.
REFERENCES
4.
5.
Fate of priority pollutants in publicly owned treatment works, 1,
EPA440/1-82/303, USEPA Effluent guidelines div.. WH-552, Wash. D.C,
Sept. 1982.
Petrasek, A. C, et a!., "Fate of toxic organic compounds in wastewater
treatment plants", Journal of Hater Pollution Control Federation 1983,
1286-1296.
Hannah, S.A, et al., "Comparative removal of toxic pollutants by six
wastewater treatment processes", Journal of Water Pollution Control
Federation 1986, 27-34.
USEPA Methods for Evaluating Solid Waste, SW846, 3rd edition, Nov. 1986.
Federal Register. Guidelines establishing test procedures for the analysis
of pollutants under the clean water act; 40 CFR, 136, 49, No. 209, 1984.
Bhattacharya, S.K., et al., "Removal and fate of RCRA and CERCLA toxic
pollutants in wastewater treatment", Final Report, Contract No. 68-03-4038.
349
-------
COMPATIBILITY OF FLEXIBLE MEMBRANE LINERS AND
MUNICIPAL SOLID WASTE LEACHATES
by: Henry E. Haxo, Jr.
Matrecon, Inc.
Alameda, CA 94501
ABSTRACT
This paper describes the results of a survey of the open technical
literature relating to the composition of currently produced municipal solid
waste (MOT) leachate, the compatibility of flexible membrane liners (FMLs)
with such leachate, and the results of limited experimental work on the
absorption of organics from dilute aqueous solutions that simulate MSW
leachates.
The results of the survey revealed little information on the compati-
bility of FMLs currently being used in the construction of disposal facili-
ties with MSW leachates. Some information was available from studies per-
formed in the 1970s. The little information that was available on the
composition of MSW leachates currently being produced indicates (1) that
concentrations of potentially polluting species in current leachates appear
to be low and (2) that current leachates may contain more potentially pol-
luting organics which are less biodegradable and which may be more aggressive
to FMLs than leachates generated during the 1970's.
Because of uncertainties regarding the compatibility of FMLs and MSW
leachates and the absorption of organics from dilute solutions, such as
leachates, limited laboratory experiments were performed to measure the
partitioning of selected organics from dilute aqueous solutions to various
types of FMLs. The results indicate that, depending on the similarity of
the solubility parameters of the organics and the FMLs, some of the organics
even at low concentrations can partition from the water in which they are
dissolved and significantly swell an FML.
350
-------
COMPATIBILITY OF FLEXIBLE MEMBRANE LINERS AND
MUNICIPAL SOLID WASTE LEACHATES
INTRODUCTION
Under Subtitle "C" of the Resource Conservation and Recovery Act (RCRA),
the EPA requires all materials that are used in constructing hazardous waste
disposal facilities and may come in contact with leachate to be tested for
chemical resistance to the specific leachate. At the present time, the EPA
is considering extending this requirement for chemical resistance testing to
all materials which are used in constructing municipal solid waste (MSW)
landfills and which may come in contact with leachate.
Previous studies indicate that commercially-available flexible membrane
liners (FMLs) are chemically resistant to typical MSW leachate (1, 2). These
studies were conducted with leachate generated in simulators containing
freshly collected and shredded MSW. Analyses of that leachate showed in-
organic salts, volatile organic acids, and suspended solids as the only
contaminating species. However, specific analyses of additional organics do
not appear to have been perfromed. In addition, adding small amounts of
organics from small quantity generators and nonhazardous industrial wastes
to an MSW stream could generate a leachate containing measurable quantities
of organic compounds that could chemically attack an FML and reduce its
service life.
Even though commercially available FMLs are generally chemically
resistant, some could be adversely affected by specific constituents of
the leachate. The effects can depend on the type of constituent, its con-
centration in the leachate, and the specific type of FML. Furthermore, the
effects of the constituents can be synergistic and can vary with time as the
concentrations change with the aging of the waste. Organic molecules [as
indicated by such analyses as volatile acids, volatile solids, total organic
carbon (TOG), and chemical oxygen demand (COD)] may be damaging to some FMLs.
For example, specific organics can cause some types of FMLs to swell, to
become softer and more permeable, and to lose in mechanical properties, such
as tensile strength and tear resistance; thus, exposure to some organics can
allow FMLs to be more easily torn and damaged. Other constituents of a
leachate, e.g. water, can also cause some FMLs to swell.
Determining the chemical resistance of FMLs and other materials of
construction is necessary when the waste liquid or leachate is. known to
contain constituents that are aggressive to these materials and when the
351
-------
concentrations of these constituents are sufficiently high. The current
method for determining chemical resistance of FMLs to waste liquids and
leachates for permitting purposes under RCRA is EPA Method 9090. However,
this method as presently performed may not be adequate for assessing FML
resistance to MSW leachate due to the instability of the leachate during the
required 120-day exposure at the temperatures (23° and 50°C). To assess FML
resistance, the FMLs under test should be exposed to leachates that are in an
in situ condition.
This paper describes the results of a survey of the open technical
literature relating to information that would be appropriate to a discussion
on extending the requirement for chemical resistance testing of materials
used in constructing MSW landfills, particularly FMLs. The information
sought in the survey included data on the composition of currently generated
MSW leachates, the compatibility of FMLs with such leachates, and basic
information regarding dilute aqueous solutions. Results of limited experi-
mental work on the partitioning of selected organics from dilute aqueous
solutions to various types of FMLs are also presented in this paper. Because
MSW leachates are essentially complex dilute solutions, these results can
supply information on the tendency of FMLs to absorb organics from leachates
and, ultimately, on the compatibility between FMLs and MSW leachates.
SURVEY OF THE SCIENTIFIC AND OPEN TECHNICAL LITERATURE
The information sought in the survey included the following:
o Data on MSW leachate composition and characteristics, particularly
currently generated leachate.
o Data from the compatibility testing of FMLs with MSW leachates.
o Information on the preservation and stabilization of MSW leachates,
particularly at temperatures somewhat higher than 23°C so that these
leachates can be used in conventional compatibility tests, e.g. EPA
Method 9090.
o Data on the resistance of FMLs and polymeric compositions to dilute
aqueous solutions of organics and inorganics. Such solutions simu-
late MSW leachates in many respects.
o Basic information from physical chemistry regarding such factors as
the chemistry of dilute aqueous solutions of organics, solubility
parameters, constituent activity, partitioning of dissolved organics
between phases, and transport of organic species.
The information resulting from this literature search is summarized in the
following subsections.
352
_
-------
Composition of MSW Leachate
The literature search revealed that most of the available information
on the composition and characteristics of MSW leachate had been generated in
the 1970s when the disposal of MSW was a principal concern of the EPA (3, 4,
5, 6, 7, 8, 9, 10, 11, 12). The data reflect the characteristics of leach-
ates generated in both laboratory and pilot-scale projects and in actual
full-scale MSW landfills; these data include information such as chemical
oxygen demand (COD), biological oxygen demand (BOD), total organic carbon
(TOG), hardness, pH, electrical conductivity (EC), total dissolved solids
(TDS), trace metal analysis, volatile organic acids, and dissolved inor-
ganics. Almost no specific information on organic species was found, except
for information on the organic acids, even though many are potentially pol-
luting or aggressive to FMLs and other polymeric .construction materials.
Much of the information on leachate composition was considered in developing
the chapter on wastes for the Technical Resource Document, SW-870, "Lining of
Waste Impoundment and Disposal Facilities" and its subsequent revision (2,
13). Overall, the information indicates the following:
o MSW leachate is a complex mixture of inorganics, organics, and
bacteriological constituents usually generated in anaerobic environ-
ments in MSW landfills. MSW leachate is generally mildly acidic,
and many of the constituents are biodegradable.
o Leachates from different MSW landfills vary widely; the precise compo-
sition is waste- and site-specific and depends on such variables as
type of waste, amount of infiltrating water, age of the landfill, and
pH. Table 1 presents typical results of analyses of MSW leachates
that were obtained and reported in the 1970s.
o MSW leachate is highly oxidizable and unstable and subject to rapid
changes in composition on removal from the in situ anaerobic environ-
ment in which it is usually generated. Even when cooled and sealed
in bottles, the composition of leachate will change for a limited
time (4).
Data in the open literature on the composition of recently generated
MSW leachates are limited; data that are available indicate the presence of
priority pollutants, aromatic hydrocarbons, and other constituents which
may be absorbed by FMLs (14, 15, 16). Table 2 presents a statistical
analysis of data on the concentrations of organic constituents of leachates
generated between 1980 and 1985. These data include data from 15 case
studies performed by the EPA (15); data from landfill leachate sampling by
Minnesota and Wisconsin, and data from the National Pollutant Discharge
Elimination System (NPDES) discharge permits for leachates from landfills
in New Jersey. Even though only a relatively small number of facilities
were surveyed, the data resulting from those studies were considered reli-
able by the EPA (16). The concentrations of most of the organics are low
and, even if they partition to the FMLs, the amounts may not be sufficiently
large to cause significant changes in properties of the FMLs, even after long
exposures.
353
-------
TABLE 1. COMPOSITION OF MSW LANDFILL LEACHATES GENERATED BEFORE 1980
Concentration of Constituents. (mg/L), Except pH and Electrical Conductivity
Constituent
BOD5
COD
TOC
Total solids
Total dissolved solids
Total suspended solids
Total volatile acids as acetic acid
Acetic acid
Fropionic acid
Butyric acid
Valeric acid
Organic nitrogen as N
Ammonia nitrogen as N
Kjeldahl nitrogen as N
Total phosphorus
PH
Electrical conductivity (ftmho/cm)
Total alkalinity as CaC03
Total acidity as CaCC>3
Total hardness as CaC03
Metals and anions:
Arsenic
Boron
Cadmium
Calcium
Chloride
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Mercury
Nickel
Phosphate
Potassium
Silica
Sodium
Sulfate
Zinc
Reference
No. 17
• • *
42,000
• • •
36,250
• • •
• • •
• • •
• • •
• • •
• • •
• • •
• • •
950
1,240
• • •
6.2
16,000
8,965
5,060
6,700
• • •
* • •
• • •
2,300
2,260
• • •
• • •
1,185
» • •
410
58
* • •
• • •
82
1,890
• • •
1,375
1,280
67
Source
Reference
No. 18
13,400
18,100
5,000
12,500
• • •
85
9,300
5,160
2,840
1,830
1,000
107
117
• • •
• • •
5.1
• • •
2,480
3,460
5,555
• • •
• • •
• • •
1,250
180
• • •
• • •
185
• • •
260
18
• • •
• • •
1.3
500
• • •
160
...
...
of Data
Reference
No. 19
• • •
1,340
• • •
• • *
• • •
• • •
333
• • •
• • •
• • •
• • •
* • •
862
• • •
• • •
6.9
• • •
* • •
• • •
• • •
0.11
29.9
1.95
354.1
1.95
<0.1
<0.1
4.2
4.46
233
0.04
0.008
0.3
• • •
• • *
14.9
748
<0.01
18.8
Reference
No. 4
81-33,360
40-89,520
256-28,000
0-59,200
584-44,900
10-700
• •
• •
• •
• •
• •
• • •
0-1,106
• • •
0-130
3.5-8.5
2,810-16,800
0-20,850
i • • «
0-22,800
• • •
• • *
0.03-17
60-7,000
4.7-2,467
• • •
0-99
0
0-2,820
17-15,600
0.09-125
• • *
• • •
• • •
28-3,770
• • •
0-7,700
1-1,558
0-370
Based on: Reference No. 2, p A-5.
354
-------
TABLE 2. CONCENTRATIONS OF ORGANIC CONSTITUENTS/
MSW LEACHATES GENERATED AFTER 1980a
(Units in ppb)
Constituent
Acetone
Benzene
Bromome thane
1-Butanol
Carbon tetrachloride
Chlorobenzene
Chloroethane
Bi s ( 2-chloroethoxy )me thane
Chloroform
Chloromethane
Delta BHC
Dibromome thane
1 ,4-Di chlorobenzene
Dichlorodifluoromethane
1 , 1-Di chloroethane
1 ,2-Dichloroethane
Cis 1 , 2-dichloroethene
Trans 1 , 2-dichloroethene
Di chloromethane
1 ,2-Dichloropropane
Diethyl phthalate
Dimethyl phthalate
Di-n-butyl phthalate
End r in
Ethyl acetate
Ethyl benzene
Bis(2-ethyl hexyl) phthalate
Isophorene
Methyl ethyl ketone
Methyl isobutyl ketone
Naphthalene
Nitrobenzene
4-Nitrophenol
Pentachlorophenol
Phenol
2-Propanol
1,1,2, 2-Te trachloroe thane
Tetrachloroethene
Tetrahydrofuran
Toluene
Toxaphene
1,1, 1-Trichloroethane
1,1, 2-Trichloroethane
Trichloroethene
Trichlorof luoromethane
Vinyl chloride
m-Xylene
p-Xylene and o-Xylene
Minimum
140
2
10
50
2
2
5
2
2
10
0
5
2
10
2
0
4
4
2
2
2
4
4
0
5
5
6
10
110
10
4
2
17
3
10
94
7
2
5
2
0
0
2
1
4
0
21
12
Maximum
11,000
410
170
360
398
237
170
14
1,300
170
5
25
20
369
6,300
11,000
190
1,300
3,300
100
45
55
12
1
50
580
110
85
28,000
660
19
40
40
25
28,800
10,000
210
100
260
1,600
5
2,400
500
43
100
100
79
50
Median
7,500
17.
55
220
10
10
7.5
10
10
55
0
10
7.7
95
65.5
7.5
97
10
230
10
31.5
15
10
0.1
42
38
22
10
8,300
270
•8
15
25
3
257
6,900
20
40
18
166
1
10
10
3.5
12.5
10
26
18
aThis table was developed by the U.S. EPA, Office of Solid Waste, Economic
Analysis Branch.
Source: Reference No. 16.
355
-------
Compatibility of FMLs with MSW Leachate
The only data on the compatibility of FMLs with MSW leachate found in
the literature survey had been developed by Haxo et al (1, 2, 13). In this
work the major types of liner materials that were available and used for
containment in the early 1970s were evaluated on exposure to MSW leachate for
up to 56 months. The leachate used in these studies was generated in pilot-
scale MSW landfill simulators filled with a shredded residential solid waste.
As the work was part of an exploratory research program performed before
Method 9090 was developed, no immersion testing that approximates EPA Method
9090 or other short-term exposures were conducted. All of the exposure
testing was performed at 23°C or less with leachate that was continually being
generated in the landfill simulators. Complete analyses of the leachate were
not performed, though several parameters, i.e. solids, pH, COD, and total
volatile acids (TVA), were followed. Overall, the changes in the physical
properties of the FMLs resulting from 56 months of exposure were relatively
minor. It should be noted, however, that the leachate generated for this
study probably did not contain aromatics, chlorinated hydrocarbons, and other
volatile organics that are known to affect hydrocarbon FMLs. Consequently,
these results may not be completely applicable to leachates being generated
currently in active landfills.
Preservation of MSW Leachates for Use in Compatibility Tests
An important factor in conducting a compatibility test is that the
composition of the test leachate reflects the composition of in situ
leachate. Thus, the test leachate should maintain a constant~~concentration
of those constituents that could affect an FML during extended service.
Because MSW leachate is highly oxidizable and unstable and is subject to
change almost immediately after removal from a service environment, which
generally is anaerobic, methods of preserving, stabilizing, or sterilizing
leachates need to be developed for possible use in performing exposure tests
to assess the compatibility of FMLs with MSW leachate.
For analytical purposes, refrigeration has been used to protect leach-
ates from bacteriological changes before they can be analyzed. However, EPA
Method 9090, the standard test for determining the compatibility of an FML
with a waste liquid, is performed at higher temperatures (23° and 50°C);
therefore, refrigeration can only be used to store the leachate before use
in exposure testing.
No information was located in the open literature indicating any methods
of long-term stabilization of MSW leachates that could be used in the com-
patibility testing of FMLs at 23° and 50°C.
Analysis of a single sample of stored MSW leachate indicated that stabi-
lization of the leachate may be possible with sterilization. A 1-gal bottle
of MSW leachate, which had been generated in October 1976 in a pilot-scale
simulator (1) and had been stored in a sealed brown glass bottle, was found
in September 1988 to be only slightly changed in composition since 1976.
356
-------
There was no indication of volatile chlorinated hydrocarbon organics in
the leachate.
If the EPA should require compatibility testing of FMLs with MSW leach-
ates in accordance with EPA Method 9090, additional research is needed to
develop adequate means of preserving or sterilizing MSW leachates to prevent
changes in leachate composition during the test, except for those changes
resulting from absorption by the FML under test.
Applicable Information on the Chemical Resistance of FMLs
Solubility parameters are used in polymer science and technology to
assess the solubilities of polymers in different organic solvents (20).
Data generated in a recent study (21, 22) on the solubility parameters of FMLs
are applicable to assessing the compatibility of FMLs and MSW leachates if the
composition of the leachate is well established. In that study, the solubility
parameters of polymers used in manufacturing FMLs were determined using their
equilibrium swelling in a variety of different-organics, the solubility
parameters of which are well documented in the literature (23, 20, 24, 25).
However, the solubility parameters of an FML are only one property of an FML
that can affect the magnitude of its swelling when it is in contact with a
leachate or waste liquid; for example, crosslinking, crystallinity, and filler
content of the FML compound can significantly affect the amount of swelling.
Applicable Information from Physical Chemistry on Dilute Solutions
When a substance such as an organic solvent is added to a two-phase
system such as a polymer and water It will, in general, distribute at equi-
librium with different concentrations in the two phases (26). The equi-
librium concentrations are related to relationship between the solubility
parameters of the phase and those of the solvent. The ratio of the con-
centrations at equilibrium of the solute in the two phases, also known as the
distribution coefficient, remains essentially constant over a wide range of
concentrations. Therefore, in the case of an organic distributed between
water and a polymer, a decrease in the concentration of the organic in the
aqueous phase would eventually result in a decrease in its concentration in
the polymer phase. This characteristic has been found to be applicable to
dilute aqueous solutions in contact with a polyethylene FML. Limited data
have been reported on the distribution coefficients of selected organics in
dilute aqueous solutions between the water and selected polyethylene FMLs
(22, 27, 28).
EXPERIMENTAL WORK ON DISTRIBUTION COEFFICIENTS OF FMLS
On reviewing the data obtained in the literature survey, it was obvious
that there were areas of uncertainty which should be resolved before valid
recommendations could be made as to the type of testing needed to assess the
long-term compatibility of a specific FML with a specific MSW leachate.
357
-------
Inasmuch as only a limited amount of experimental work could be per-
formed in this study, it was decided to measure the absorption of three
organics by different FMLs from dilute aqueous solutions. Samples of FMLs
based on four different polymers, including a linear low-density polyethyl-
ene (LLDPE), polyvinyl chloride (PVC), chlorinated polyethylene (CPE), and
chlorosulfonated polyethylene (CSPE), were placed in vapor-tight cells con-
taining a dilute, unsaturated aqueous solution containing three different
organics. The concentrations of the organics in the aqueous solutions were
monitored until they had become relatively constant at which time the cells
were opened and the FML samples analyzed to determine the concentration of
the organics in the FML samples. Analyses of both the solutions and the FMLs
for the organics were performed using gas chromatographic (GC) procedures.
Experimental details and results are presented in this section.
Gas Chromatography Procedures
A Perkin-Elmer Sigma Three Series gas chromatograph with a flame ioni-
zation detector was used for the GC analyses. The instrument was fitted with
an open-capillary column coated with polyethylene glycol. Details of the
GC analyses are presented in Table 3.
The concentrations of the organics in the aqueous solutions were
determined by injecting samples removed from the test cells directly into
the GC. These cells featured septums through which the aqueous solutions
could be sampled during the test to determine whether equilibrium had been
reached.
The concentrations of the organics in the FML samples were determined
by headspace GC. In this procedure, an FML specimen containing absorbed
organics is placed in a small vapor-tight can provided with a septum
through which vapors from the specimen can be sampled. The can is placed
in an oven at 105°C and heated for approximately an hour. A sample of the
vapors is drawn from the can and injected into the GC. The FML specimen is
removed from the sampled can and placed in a new can which is then heated
in a 105°C oven for approximately an hour. Once again, the vapors inside
the can are sampled and injected into the GC. The process of heating,
sampling, and injecting is repeated until no organics are detected in the
sampled vapors by the GC. The amount for each organic were summed to
represent a total for the organic in the sample.
The concentrations of the organics in the injected samples were
calculated by comparing peak height values resulting from the GC analyses
with calibration curves. The calibration curves for the aqueous solutions
were determined by injecting 1 /*L of various solutions of known concentra-
tions of the different organics into the GC column. Injections of each
standard were performed five times to ensure reproducibility of injection
techniques. The calibration curves for the headspace GC analyses were
prepared by analyzing a specific 100 /xL volume of vapor from headspace cans
filled with different amounts of the organics.
358
-------
TABLE 3. GAS CHROMATOGRAPHY CONDITIONS FOR LEACHATE
AND VAPOR ANALYSIS
Condition
Setting
Detector range
Injector and detector temperature
Initial temperature
Initial holding time
Final temperature
Final holding time
Temperature program rate
Detector
Column^
Chart speed
Carrier
Specimen size:
Liquid
Vapor from headspace
Attenuation
x 10
250°C
60°C
1 min.
120°Ca
1 min.
20°C/min.
Flame ionization:
H2 30 cc/min.
02 40 cc/min.
Polyethylene glycol coated
fused silica open tube
(FSOT) capillary: 0.53 mm
in diameter and 15 m in
length
30 cm/hour
Helium, 10 cc/min.
1 ML
100 /uL
32 (for 500 ppm concentration)
down to 4 (for low concentration)
aCan vary with the solvent mixture from 120°C to 260°C, the maximum
temperature for the column. ,
^Trade name Superox (Altec), Megabore DB WAX (J and W).
359
-------
Selection of Volatile Organics
For this study, three types of volatile organics representing a range
of chemical characteristics were desired, including a volatile containing
oxygen, a volatile that was a chlorinated solvent, and a volatile that was
an aromatic. The three selected were methyl ethyl ketone (MEK), trichloro-
ethylene (TCE), and toluene. All of these organics have been observed in
MSW leachates. Properties of these solvents are presented in Table 4.
TABLE 4. ORGANICS USED IN ABSORPTION EXPERIMENTS
WITH DILUTE AQUEOUS SOLUTIONS
Organic
Toluene
Trichloroethylene
Methyl ethyl ketone
Mole-
cular
weight
92.13
31.40
72.10
Density
at 20°C,
K cm""
0.866
1.476
0.805
Boiling
point ,
°C
110.6
87.2
79.6
Vapor
pressure
at 25°C,
mm Hg
31.96
80.30
100.0
«o
8.9
9.2
9.3
Solubility
parameters3
<5d See Reference No. 29.
cChemical Abstract Services' number.
Immersion Test Cells
The immersions were performed in vapor-tight cells consisting of 8-oz
glass jars with ground polished top edges, and Teflon-lined phenolic resin
tops (Figure 1). Each of these tops was fitted with a Swagelock sampling
port and a Teflon-lined silicone rubber septum for withdrawing samples for
GC analysis. Special arrangements were made in the cells to suspend the FML
samples in the solutions.
Experimental Procedure
The work performed in this study developed out of previously reported
work (22, 27, 28). In the earlier study, samples of a polyethylene FML were
placed in the same type of test cells as used in the present work. These
cells were then filled with a series of dilute but saturated aqueous solu-
tions, each of which contained a single organic. These solutions were
prepared with an excess of the organics to maintain saturation. After
analyzing the aqueous solutions and FML samples at the end of the immersions,
the distribution coefficients (i.e. the ratio of the concentration of the
organic in the FML to its concentration in the aqueous solution) were"cal-
culated. It was noted, however, that the amounts of organics absorbed by the
FML approached the amounts the FML absorbed when immersed in neat organics.
It appears that the water had acted as a permeable medium between the FML and
360
-------
the reservoir of excess organics which allowed the organics to be absorbed by
the FML until saturation of the FML was reached. These conditions are not
representative of those in a landfill due to the improbability of an excess
of the organic being maintained until saturation was reached in the installed
FML.
Teflon
Septum
\
Teflon-lined
Screw cap
Swagelock
Assembly
Washer
Nut
TOP ASSEMBLY
Jar with
ground and
polished
edge
8 OZ JAR
Figure 1. Schematic of the immersion test cell.
In the experimental work conducted in the present study, the aqueous
solutions were prepared at concentrations less than saturation and the
number of FMLs in test were increased to include the four basic polymer
types. Four immersion test cells were each filled with an aqueous solution
containing a mixture of 500 ppm (on a weight basis) each of MEK, TCE., and
toluene. Samples of CPE, fabric-reinforced chlorosulfonated polyethylene
(CSPE-R), LLDPE, and PVC FMLs were placed in separate test cells, and the
concentrations of the volatile organics in each cell were followed by GC.
In all cases, the initial concentrations of the MEK in solution showed
little change and, thus, little distribution to the FMLs, whereas the
toluene and TCE showed drops in concentration and thus a transfer from the
water to the FMLs. The cells were dismantled and the respective FMLs
were analyzed for the volatile organics by headspace GC.
Experimental Results
The results of analyzing the aqueous solutions and the FMLs at the end
of the immersion are presented in Table 5. The results show that there was
a significant increase in the weight of the FML samples; however, the dif-
ference between the weight gains of the FML samples and the total amount
of organics detected in each FML sample indicates that not all of the weight
gains could be attributed to the absorption of the organics. Thus, consider-
able water had also been absorbed by the specimens. In addition, it is also
361
-------
TABLE 5. ABSORPTION BY IMMERSED FMLS OF A MIXTURE OF ORGANICS3
FROM DILUTEb AQUEOUS SOLUTIONS
Parameter
Contents of cell at beginning
of experiment
Amount of water in cell, g
Amount of organics in cell:
MEK, mg
TCE, mg
Toluene, mg
Total
Original weight of FML
specimen, g
FML specimen at end of
immersion
Swollen weight , g
Weight gain, g
Weight gain, %
Headspace analysis of
swollen FML specimen
Amount in swollen FML
specimen:
MEK, mg
TCE, mg
Toluene , mg
Total
% of original weight
of FML
Concentration of organics
in swollen FML specimen
MEK, ppm
TCE, ppm
Toluene , ppm
PVCC
579
225.5
112.8
112.8
112.8
338.4
3.4580
3.668
0.210
6.1
1.35
54.0
55.6
110.95
3.21
370
14,720
15,200
LLDPEC
580
226.6
113.3
113.3
113.3
339.9
2.5337
2.626
0.093
3.7
0.08
36.3
44.5 .
80.88
3.19
30
13,800
16,900
CPEC
581
228.4
114.2
114.2
114.2
342.6
5.4670
6.246
0.779
14.2
1.16
63.7
65.5
130.36
2.38
190
10,200
10,490
CSPE-RC
582
225.2
112.6
112.6
112.6
337.8
6.1601
6.702
0.542
8.8
0.55
34.3
38.8
73.65
1.20
80
5,120
5,790
continued .
362
-------
TABLE 5. CONTINUED
Parameter
FML specimen after head-
space analysis
PVCC
579
LLDPEC
580
CPEC
581
CSPE-RC
582
Weight.g 3.4501 2.5310 5.4551 6.1253
Loss in weight (based on
original weight), % 0.23 0.11 0.22 0.56
at end of experiment
Amount in aqueous solution:
MEK, mg
TCE, mg
Toluene , mg
Total
Concentration of organics
in aqueous solution (CH_Q):
MEK, ppm
TCE, ppm
Toluene , ppm
Total amount of organics
in celld:
MEK, mg
TCE, mg
Toluene , mg
Total
Organic accounted for at
end of experiment , %
Distribution coefficient
(CFML/CH2o):
MEK
TCE
Toluene
112.5
48.3
39.2
200.0
500
214
174
114
102
95
311
92
0.74
68.8
87.4
110.1
87.5
80.4
278.0
486
386
355
110
124
125
359
106
0.06
35i.8
47.6
91.4
41.6
27.9
160.9
400
182
122
93
105
93
291
85
0.48
56.0
86.0
111
46.6
38.1
195.7
493
207
169
111
81
77
269
80
0.16
24.7
34.3
aMethyl ethyl ketone (MEK), trichloroethylene (TCE), and toluene.
CPVC » polyvinyl chloride; LLDPE = linear low-density polyethylene;
CPE = chlorinated polyethylene; CSPE-R = chlorosulfonated polyethylene
(fabric-reinforced).
^Initial concentration of each organic was 500 ppm.
of organics in water and in the FML.
363
-------
possible that some of the organics not accounted for at the end of the experi-
ment had been absorbed by the FML but were not recovered during the headspace
GC procedure. Regardless, the results do show a large partitioning of the
organics from the water to the FMLs and, furthermore, they show that the FMLs
varied considerably in their absorption of the different organics. The total
amounts of organics and water that were absorbed ranged from 3.7% for the
polyethylene to 14.2% for the CPE. This experiment was repeated with similar
overall results. The results of very limited physical testing, which was
performed on percut specimens immersed in the second experiment, consistently
indicated moderate losses in tensile property values.
CONCLUSIONS
The results of the literature survey and the limited experimental work
that was conducted in this study indicate that:
o There is little information in the literature regarding the compo-
sition and characteristics of current generation leachates and the
compatibility of FMLs with these leachates.
o The limited data available indicate that MSW leachates currently
being generated may contain low concentrations of organics having low
biodegradability; in addition, depending on their concentration, some
of these organics are aggressive towards FMLs.
o It is questionable whether EPA Method 9090, as presently designed,
can result in a consistent and realistic assessment of the compati-
bility of an FML with a MSW leachate without modification of the
exposure cells to maintain anaerobic conditions and a standard means
of assuring stabilization of the leachates so that little change in
concentrations of the constituents during the required four months is
ensured.
o The use of distribution coefficients that represent the distribution
of dissolved organics between aquous and polymeric phases appears to
be applicable to assessing the compatibility of FMLs and other poly-
meric construction materials with MSW leachates, which are basically
dilute aqueous solutions of organic and nonorganic solutes. Thus, the
amount of organics absorbed by an FML from a dilute aqueous solution
with which it is in contact may be estimated. For example, at low
concentrations of organics in solution, the amount of the organics
that an FML will absorb will diminish as the concentration in the
leachate is lowered.
o The solubility parameters of an FML and the individual organic are
important tools in assessing the level of absorption of the organics
by the FML and in determining the distribution coefficient of an
organic between an aqueous solution and an FML.
364
-------
RECOMMENDATIONS
It is recommended that further experimental research be performed:
o To develop quantitative data regarding the distribution of dissolved
organics between various FMLs and aqueous solutions or actual MSW
leachates in which the FMLs are immersed.
o To determine threshold concentrations of various organics found
in MSW leachates, below which levels absorption will not signif-
icantly affect the properties of FMLs and other materials of con-
struction used in liner systems.
o To assess possible interactions between organics in dilute aqueous
solutions and the effects of the absorbed organics on permeability
and properties of FMLs.
o To perform a round-robin series of exposure tests performed in
accordance with EPA Method 9090 to establish error and bias in
the test method.
o To perform more in-depth analyses of MSW leachates than have
normally been made to determine the presence of specific organics
that swell various FMLs. These organics include chlorinated
solvents, aromatic solvents, and some aliphatic solvents, all of
which which are known to deteriorate the properties of polymeric
compositions.
ACKNOWLEDGMENT
The work on this project was performed by Matrecon under Work Assignment
No. 0/01 of Subcontract 771-87 of Environmental Protection Agency Contract
68-03-3413 to PEI Associates, Inc.
REFERENCES
1. Haxo, H. E., R. M. White, P. D. Haxo, and M. A. Fong. 1982. Final
Report: Evaluation of Liner Materials Exposed to Municipal Solid Waste
Leachate. NTIS No. PB 83-147-801. U.S. Environmental Protection
Agency, Cincinnati, OH.
2. Matrecon, Inc. 1988. Lining of Waste Containment and Other Impoundment
Facilities. EPA-600/2-88-052. [SW-870, second revised edition.] U.S.
Environmental Protection Agency, Washington, B.C. 991 pp.
3. Chian, E. S. K., and F. B. DeWalle. 1976. Analytical Methodologies for
Leachate and Gas Analysis. In: Proceedings of a Research Symposium on
Gas and Leachate from Landfills: Formation, Collection, and Treatment.
EPA-600/9-76-004 (NTIS No. PB-251-161). U.S. Environmental Protection
Agency, Cincinnati, OH. pp 44-53.
365
-------
4. Chian, E. S. K., and F. B. DeWalle. 1977. Evaluation of Leachate
Treatment. 2 volumes. EPA-600/2-77-186 a,b. U.S. Environmental
Protection Agency, Cincinnati, OH.
5. Dunlap, W. J., D. C. Shew, J. M. Robertson, and C. R. Toussaint. 1976.
Organic Pollutants Contributed to Groundwater by a Landfill. In:
Proceedings of a Research Symposium on Gas and Leachate from Land-
fills: Formation, Collection, and Treatment. EPA-600/9-76-004 (NTIS
No. PB-251-161). U.S. Environmental Protection Agency, Cincinnati,
OH. pp 96-110.
6. Ham, R. K. 1975. Milled Refuse Landfill Studies at Pompano Beach, FL.
Approx. Range, Three Cells Aged One Year. 21 pp.
7. Ham, R. K. 1976. Solid Waste Degradation Due to Shredding and Sludge
Addition. In: Proceedings of a Research Symposium on Gas and Leachate
from Landfills: Formation, Collection, and Treatment. EPA-600/9-76-
004 (NTIS No. PB-251-161). U.S. Environmental Protection Agency,
Cincinnati, OH. pp 168-176.
8. Ham, R. K., K. Hekimian, S. Katten, W. J. Lockman, R. J. Lofty, D. E.
McFaddin, and E. J. Daley. 1979. Recovery, Processing, and Utili-
zation of Gas from Sanitary Landfills. EPA-600/2-79-001. U.S.
Environmental Protection Agency, Cincinnati, OH. 133 pp.
9. Pohland, F. G. 1975. Sanitary Landfill Stabilization and Leachate
Recycle and Residual Treatment. EPA-600/2-75-043. U.S. Environmental
Protection Agency, Cincinnati, OH. 105 pp.
10. Pohland, F. G. 1976. Landfill Management with Leachate Recycle and
Treatment: An Overview. In: Proceedings of a Research Symposium on Gas
and Leachate from Landfills Formation, Collection, and Treatment.
EPA-600/9-76-004 (NTIS No. PB-251-161). U.S. Environmental Protection
Agency, Cincinnati, OH. pp 159-167.
11. Pohland, F. G., D. E. Shank, R. E. Benson, and H, H. Timmerman. 1979.
Pilot Scale Investigations of Accelerated Landfill Stabilization with
Leachate Recycle. In: Municipal Solid Waste: Land Disposal. Proc.
5th Annual Res. Sympos. EPA-600/9-79-023a. U.S. Environmental Pro-
tection Agency, Cincinnati, OH. pp 283-295.
12. Pohland, F. G., W. H. Cross, and J. P. Gould. 1987. The Behavior and
Assimilation of Organic and Inorganic Priority Pollutants Codisposed
with Municipal Refuse - A Progress Report. In: Proceedings of the
Thirteenth Annual Research Symposium: Land Disposal of Hazardous Waste.
EPA-600/9-87-015. U.S. Environmental Protection Agency, Cincinnati, OH.
pp 26-37.
13. Matrecon, Inc. 1983. Lining of Waste Impoundment and Disposal Facili-
ties. SW-870 Revised. U.S. Environmental Protection Agency, Washington,
D.C. 448 pp.
366
-------
14. Kmet, P., and P. M. McGinley. 1982. Chemical Characteristics of Leach-
ate from Municipal Solid Waste Landfills in Wisconsin. In: Proceedings
of the 5th Annual Madison Conference of Applied Research and Practice on
Municipal and Industrial Wastes, September 22-24, 1982. Dept. of Eng.
and Applied Science, University of Wisconsin Extension, Madison, WI.
pp 225-254.
15. EPA. 1986. Muncipal Landfill Case Studies (unpublished). These
studies were prepared by PEI, SRW, and ICF. U.S. Environmental Pro-
tection Agency, Office of Solid Waste, Washington, D.C. Cited in: U.S.
EPA. 1986. Subtitle D Study - Phase I Report. EPA/530-SW-86-054.
U.S. Environmental Protection Agency, Office of Solid Waste, Washington,
D.C.
16. EPA. 1986. Subtitle D Study - Phase I Report. EPA/530-SW-86-054.
U.S. Environmental Protection Agency, Office of Solid Waste, Washington,
D.C.
17. Wigh, R. J. 1979. Boone County Field Site. Interim Report, Test Cells
2A, 2B, 2C, and 2D. EPA-600/2-79-058. U.S. Environmental Protection
Agency, Cincinnati, OH. 202 pp. (NTIS PB-299-689).
18.
19.
20.
21.
22.
23.
Breland, C. G. 1972. Landfill Stabilization with Leachate Recir-
culation, Neutralization, and Sludge Seeding. CE-756A6. School of
Civil Engineering, Georgia Institute of Technology, Atlanta, GA. 80
pp.
Griffin, R. A., and N. F. Shimp. 1978. Attenuation of Pollutants in
Municipal Landfill Leachate by Clay Minerals. EPA 600/2-78-157 (NTIS
PB 287-140). U.S. Environmental Protection Agency, Cincinnati, OH.
146 pp.
Barton, A. F. M. 1983. Solubility Parameters and Other Cohesion
Parameters Handbook. CRC Press, Boca Raton, FL.
Haxo, H. E., and P. J. Pick. 1986. Determination of the Solubility
Parameters of FMLs for Use in Assessing Resistance to Organics. In:
Proceedings of the Twelfth Annual Solid Waste Research Symposium: Land
Disposal, Remedial Action, Incineration and Treatment of Hazardous
Waste. EPA/600/9-86/022. U.S. Environmental Protection Agency,
Cincinnati, OH. pp 132-145.
Haxo, H. E., T. P. Lahey, and M. L. Rosenberg. 1988. Factors in Asses-
sing the Compatibility of FMLs and Waste Liquids. EPA/600/2-88/017
(NTIS No. PB 88-173-372/AS). U.S. Environmental Protection Agency,
Cincinnati, OH. 143 pp.
Barton, A. F. M.
75(6):731-753.
1975. Solubility Parameters. Chemical Reviews
367
-------
24. Leo, A. and C. Hansch. 1970. Linear Free-Energy Relationships Between
Partitioning Solvents Systems. Journal of Organic Chemistry 36(11):
1539-1544.
25. Leo, A., C. Hansch, and D. Elkins. 1971. Partition Coefficients and
Their Uses. Chemical Reviews, Vol. 71, No. 6. pp. 525-554.
26. Daniels, F., and R. A. Alberty. 1961. Physical Chemistry. 2nd
Edition. John Wiley and Sons, NY.
27. Haxo, H. E. 1988. Transport of Dissolved Organics from Dilute Aqueous
Solutions Through Flexible Membrane Liners. In: Proceedings of the
Fourteenth Annual Solid Waste Research Symposium: Land Disposal,
Remedial Action, Incineration and Treatment of Hazardous Waste, May
9-11, 1988. U.S. Environmental Protection Agency, Cincinnati, OH.
21 pp.
28. Haxo, H. E., and T. P. Lahey. 1988. Transport of Dissolved Organics
from Dilute Aqueous Solutions Through Flexible Membrane Liners.
Hazardous Waste and Hazardous Materials. 5(4):275-294.
29. Riddick, J., and W. Bunger. 1970. Techniques of Chemistry, Volume
II - Organic Solvents, Physical Properties and Methods of Purification.
Wiley-Interscience, NY.
368
-------
GEOSYNTHETIC CONCERNS IN LANDFILL LINER AND COLLECTION SYSTEMS
by: Robert M. Koerner, Arthur E. Lord, Jr., and Yick H. Halse
Geosynthetic Research Institute
Drexel University
Philadelphia, PA 19104
ABSTRACT
The use of geosynthetic materials in landfill liner and closure systems
is commonplace. This use includes the containment of all types of solid waste
materials, e.g., hazardous, municipal, industrial, ash, low level
radioactive, and heap leach, ores. The geosynthetic materials involved include
the following;
• geomembranes or flexible membrane liners (FMLs),
• geotextiles as filters and cushions,
• geonets as drains,
• geogrid reforcing elements, and
• geocomposites for various uses.
While it is felt that current systems can indeed be designed and constructed
with confidence, there are certain aspects in need of further investigation
and/or clarification. This paper highlights several of them. Included are an
investigation of FML behavior in anchor trenches, stress cracking behavior of
HOPE seams, flow rate reduction in geonets, and particulate and biological
clogging of geotextile filters. Some amount of data is presented on each
topic but, clearly, the paper represents research-in-progress from the point
of view of definitive guidelines. The goal of each individual topic is to
generate a data base so as to avoid long-term problems when using
geosynthetics in association with waste containment facilities.
INTRODUCTION AND OVERVIEW
Current waste containment facilities represent a bevy of geosynthetic
materials. Included in the group are FMLs which form the primary and/or
secondary liners against escaping liquids, geonets which are commonly used as
drainage materials in primary and/or secondary leachate collection systems
and geotextiles used as filters to allow leachate to enter into drainage
collectors. When the waste containment facility is doubly lined, as are many
hazardous, municipal and industrial landfills, the cross section of Figure 1
369
-------
is common. Here one can itemize the following geosynthetics and their primary
functions (progressing from the waste downward);
• geotextile filter beneath the waste,
• geonet as a primary leachate collector on side slopes,
• geotextile filter around perforated pipe drains,
• geotextile cushion above primary FML,
• the primary FML (i.e., a geomembrane),
• geotextile separator between primary clay and geonet drain,
• geonet as secondary leachate collector,
• the secondary FML (i.e., a geomembrane), and
• geotextile cushion against soil subgrade
Thus as many as nine (9) geosynthetics are used in the liner system shown.
When adding the various geosynthetics placed in the cap, or closure, of such
facilities, it is easily seen that these materials play a key role in the
proper functioning of the system. This is tantamount to stating that they
must be properly evaluated, selected, designed, and installed.
Toward this end we have selected to investigate a number of areas which
are felt to be of concern. They are = identified on Figure 1 and will be
described in the sections to follow.
1. FML Behavior in Anchor Trenches - As shown in Figure 1 at Location "1",
FMLs will typically come out of a lined facility and terminate around the
perimeter of the site. There is a short horizontal runout length and then a
vertical drop into an anchor trench. The lengths involved can be estimated by
available design models (1,2) but their evaluation awaits field verification.
As an intermediate step we have constructed a large scale test facility to
model the situation in the laboratory. A steel reinforced wooden box of
internal size 3 ft. height by 3ft. width by 6 ft. length has been constructed
which can contain various anchor trench configurations, see Figure 2. The
jacking system to which the outside of the FML is attached can be oriented so
as to exert a downward force thereby simulating an actual situation. Normal
Stress is exerted by an air-bag system to mobilize overburden pressures up to
1500 lb/ft2. While not necessary for the perimeter anchor trench shown in
Figure 1, these high normal stresses are mobilized on the interior berms of
discrete cells when a zoned landfill expands laterally and then vertically.
Strain gages are attached to the FML test specimens at four locations.
The initial data indicate that the largest FML stresses are in the portion of
the FML closest to the pullout force. These stresses rapidly dissipate with
increasing distance into the test box as seen in the curves of Figure 2
moving from locations Gl through G4. Obviously, soil friction plays an
important role in this stress dissipation. Also note that the FML in the
vertical anchor trench is essentially non-stressed, i.e. gage number G4. This
has been typical of tests conducted to date. The planned experiments include
30 and 60 mil HDPE, 40 mil LLDPE, 30 mil PVC and 36 mil CSPE-R. Normal stress
will vary from 100 lb/ft2 to 1500 lb/ft2. Results will be compared to the
various design models so as to verify, modify or refute the existing
literature.
370
-------
2. Stress Cracking Behavior of HOPE Seams - ASTM defines stress cracking
as, "an internal or external rupture in a plastic caused by tensile stress
less than its short-term mechanical strength." Thus any mechanism which cause
premature failure of the FML can fall into this general category, e.g.,
environmental stress cracking, residual stress cracking, scratches and cracks
resulting from grinding marks, geometrical irregularities, and/or fatigue
failure. Many (if not all) of these situations are encountered at HDPE liner
seam locations.
An evaluation of HDPE seams was initially addressed using a modified
version of ASTM D-2552 test method which immerses dumbbell-shaped test
specimens in a surface wetting agent at an elevated temperature. A number of
HDPE seam types were evaluated under different constant stress levels
resulting in a large number that cracked (3), see Tables l(a) and (b) for
test results of 168 and 1000 hour test durations, respectively.
TABLE l(a) - LABORATORY STRESS CRACKING RESULTS FOR 168 HR. DURATION
Type of Seam
fillet extrusion
flat extrusion
hot wedge
hot air
ultrasonic
Number of
Tests
179
40
60
80
40
Elastic
74
15
25
27
28
Results
Plastic
13
19
19
22
5
Cracked
92
6
15
31
7
Percent
Cracked
51
15
27
39
25
TABLE l(b) - LABORATORY STRESS CRACKING RESULTS FOR 1000 HR. DURATION
Type of Seam
fillet extrusion
flat extrusion
hot wedge
Number of
Tests
20
20
20
Elastic
14
11
8
Results
Plastic
1
0
0
Cracked
5
9
12
Percent
Cracked
25
45
60
To be emphasized is that these laboratory tests are index tests wherein
stress relaxation of the polymer cannot occur and that they are conducted
under very harsh conditions. The extent of the problem in the field is not
known, although it is known to exist at a few surface impoundments where the
FML was exposed to the atmosphere <4>, see Figure 1 at Location "2".
Work is currently ongoing in evaluating various seaming procedures
(e.g., temperature, pressure and time), determining residual stresses,
evaluating cracking initiation mechanisms, measuring crack growth rate in
various HDPE formulations, and attempting to optimize seam geometry. The
ultimate goal is to assess the magnitude of the stress cracking problem, its
optimization and/or its remediation.
371
-------
3. Flow Rate Reduction of Geonets - The intrusion of a geotextile (or any
other material) into the flow apertures of a drainage net will surely reduce
its flow rate capability. While a number of cross sections can be envisioned,
the situation of a clay over a geonet with a geotextile separator between
them is of great concern, see Figure 1 at Location "3".
Flow rate tests using a cross section consisting of a geonet between two
60 mil HDPE liners versus the same type of geonet between a 60 mil HOPE liner
beneath it and a geotextile/clay layer above it have been conducted. The
geotextile used as the separator was a needle punched, continuous filament,
nonwoven fabric of 8.0 oz/yd2 mass per unit area. The test procedure was
performed in accordance with ASTM D-4716. Results are given in Table 2 where
one can see flow rate decreases of 19% to 41% depending upon the hydraulic
gradient and applied normal stress. Clearly, intrusion of the geotextile/clay
is occurring as it spans the open apertures of the geonet. Note that the
geotextile acts in a membrane reinforcement mode and must do so for the
design lifetime of the system. If it fails, the clay will immediate extrude
into the geonet thereby blocking all flow. Thus long term creep tests are
warranted for many situations.
TABLE 2 - FLOW RATES (IN GAL/MIN-FT) AND REDUCTIONS (IN %) FOR DIFFERENT
GEONET DRAINAGE CROSS SECTIONS
Normal
e>4-
(Ib/ft2)
5000
10,000
15,000
Cross
HDPE (both sides)
GT-Clay (one side)
Difference
Reduction
HDPE (both sides)
GT-Clay (one side)
Difference
Reduction
HDPE (both sides)
GT-Clay (one side)
Difference
Reduction
Hydraulic Gradient "i"
i = 0.25
1.6
JUJi
0.3
19%
1.4
•J-l
0.3
21%
1.2
ILu2
0.3
25%
i = 0.50
3.7
2^1
1.0
27%
3.4
2^3.
1.1
32%
3.2
JU&
1.4
44%
i = 1.00
7.1
±*&
2.3
32%
6.4
AJ.
2.3
36%
5.6
3.3
2.3
41%
4. Parfciculate Clogging of Geotextile Filters - Shown in Figure 1 at
Locations "4" is a geotextile covering a primary leachate collection geonet
on the sideslope and drainage gravel on the bottom. In both cases the
geotextile must act as a filter. Hence, it must provide adequate permeability
(thus open voids) and soil retention (thus tight voids), i.e., a balanced
void structure must be achieved. While these considerations can be adequately
handled by proper design, there is a third necessary condition, that being
prevention of long-term clogging.
372
-------
The customary approach toward the assessment of geotextile clogging is
by the gradient ratio test, CW-02115. This U.S. Army Corps of Engineers
developed test, however, was directed toward sandy soils and woven,
monofilament geotextiles of relatively high open areas (5). For the very fine
sediment carried by leachate, the authors prefer the use of long-term column
tests (6,7). Here the actual cross section being considered is simulated
using site-specific cover soil, the candidate geotextile and actual (or
simulated) leachate. The long term flow through this cross-section is
measured over a period of time. As seen in Figure 3, the flow rate initially
decreases until a transition time is reached. This is due to soil compaction
and an initial "tuning" of the geotextile to the upstream soil and the
permeating liquid. Beyond this point, the flow rate either becomes constant
(thus equilibrium is established), becomes uncertain (which requires
continued testing), or decreases to zero (hence the system is clogged).
A broad base study is ongoing in this regard using four geotextiles
(woven monofilament, nonwoven melt-bonded, lightweight needle-punched
nonwoven, and heavyweight needle-punched nonwoven); four conditions above the
geotextile (no soil, Ottawa sand, clayey silt and a combination of sand and
clayey silt); and two permeating liquids (water and sediment laden water).
Upon establishing one of the trends shown in Figure 3, the system will be
epoxy-set, properly sectioned and microscopically viewed so as to understand
the mechanisms involved. The goal of this study is a laboratory test matrix
illustrating the behavior and proper design against fine particulate clogging
of geotextile filters.
5. Biological Clogging of Geotextile Filters - To be sure, municipal
landfill leachates have many viable (active) bacteria present in them, see
Figure 4. The geotextile filters shown in Figure 1 at Locations "5" interface
with these bacteria. Columns of the type shown in Figure 3 have been set up
at six municipal landfills under aerobic conditions resulting; in the flow
rate trends shown in Table 3(a). Relatively large reductions over the elapsed
times are indicated, e.g., 50% reduction (or higher) occurred in one-third of
the test materials. A separate phase of the study was also aimed at anaerobic
bacteria clogging of geotextiles which produced the results of Table 3 (b) .
Here the flow rate reductions are much less than noted previously, but are
nonetheless quantifiable in all cases of 7 months or longer. See reference 9
for additional details.
This study is also ongoing into a second phase where backflushing with
water and with biocide-treated water will be performed. The resulting flow
rate response will be evaluated. If flow is reconstituted to its original, or
near to original, flow rate the recommendation will be forthcoming to
incorporate such procedures into the design of the collection system. Such
designs are certainly possible and, if justified, should be implemented.
SUMMARY AND CONCLUSIONS
Numerous types of geosynthetic materials are currently seeing widespread
use in waste containment facilities. Use of these materials has led to a
certain degree of confidence which is augmented by the installation of
redundant liner and leachate collection systems. Due to their rapid
373
-------
co
in
m
d>
-H r-i
S W
•P
w to
\p a)
(0 4J
0)
4J T)
(0
0) 0)
^•^
M tn
S -H
O H
H H
Cu Cn
Cu o
I rH
id H
* 3
<4H
o
-H -P
1-1 10
X> <1)
(0 JJ
<4H *
H
fa
O
HI
co
CO
«.
I rH
df> dP O O
IT) O O O
I rH t-l H
I I I
dp dp dp dp O U
m o o o --^ ~~.
I rH rH rH S 2
O O
S3 ^
dP dp 6IP o\»
o in o o
I I I I
co to w to
rH O
rH rH CT>
in
r- r-
00 CO 00 00
rH rH in
00 OO
OO
eg in vo CN
rH H rH
VO
rH eg ro i1 in vo
I I I I I I.
a<
CO
Q.
•o
o
II M
_ <1>
B 4J
a.
H
IS CU
374
-------
development, however, there are areas where additional inquiry is warranted.
Some of these areas have been the focus of this paper.
• FML anchor trench behavior and mobilization mechanisms will soon have
a meaningful data base via large scale laboratory evaluation. A
rigorous critique of available design methods will result.
• Stress cracking of HDPE field seams appears to be strongly related to
the workmanship involved in making the seam. Strict construction
quality control (10) should alleviate the concern for most situations
except possibly the exposed berms of surface impoundments. In these
areas proper design with sufficient expansion and contraction
capability must be considered.(11)
• Flow rate reduction of geonets via intrusion of materials above and
below the net's apertures or openings is very quantifiable. Proper
design of the geotextile or geomembrane above and below the net is
necessary. The laboratory simulation test is available to assess the
adequacy of the design.
• Fine precipitate clogging of geotextile filters associated with
leachate collection systems is of concern for those leachates with
high particulate content. A laboratory simulation test is proposed and
a broad based series of experiments is ongoing. If such clogging is of
serious concern, backflushing might be required for remediation and
reconstitution of flow.
• Biological clogging of geotextile filters associated with leachate
collection systems is a likely concern for municipal landfills.
Clearly, biocide treatment either within the geotextiles or in the
backflushing liquid as proposed above are possible remedial measures.
Work is ongoing to evaluate these various remediation schemes.
ACKNOWLEDGEMENT S
This project is funded by the U.S. Environmental Protection Agency under
Project No. CR-813953-01. Our sincere appreciation is extended to the Agency
and in particular to the Project Officer, Robert E. Landreth.
REFERENCES
1. Koerner, R. M., Designing with Geosynthetics, Prentice Hall Publ. Co.,
Englewood Cliffs, NJ, 1st Ed. 1986, 2nd Ed. 1989 (to appear).
2. Richardson, G. N. and Koerner, R. M., "Geosynthetic Design Guidance for
Hazardous Waste Landfill Cells and Surface Impoundments," EPA Contract No.
68-03-3338, 1987, GRI, Drexel University, Philadelphia, PA.
3. Halse, Y. H., Koerner, R. M. and Lord, A. E., Jr., "Laboratory Evaluation
of Stress Cracking in HDPE Geomembrane Seams," Proc. Aging and Durability
of Geosynthetics, Dec. 1988, GRI, Drexel University, Philadelphia, PA.
375
-------
4. Peggs, I. and Carlson, D. S., "Stress Cracking of Polyethylene
Geomembranes: Field Experience," Proc. Aging and Durability of
Geosynthetics, Dec. 1988, GRI, Drexel University, Philadelphia, PA.
5. Haliburton, T. A. and Wood, P. D., "Evaluation of U.S. Army Corps of
Engineers Gradient Ratio Test for Geotextile Performance," Proc. 2nd Int.
Conf. on Geotextiles, Las Vegas, NV, Aug. 1-6, 1982, IFAI, pp. 97-101.
6. Koerner, R. M. and Ko, F. K., "Laboratory Studies on Long-Term Drainage
Capability of Geotextiles," Proc. 2nd. Int. Conf. Geotextiles, Las Vegas,
NV, Aug. 1-6, 1982, IFAI, pp. 91-95.
7. Halse, Y. H., Koerner, R. M. and Lord, A. E. Jr., "Filtration Properties
of Geotextiles Under Long Term Testing," Proc. ASCE/PennDOT Conf. on
Advances in Geotechnical Engineering, Hershey, PA, Apr. 1987, pp. 1-13.
8. Rios, N. and Gealt, M. A., "Biological Clogging Growth in Landfill
Leachate Collection Systems," Proc. Aging and Durability and
Geosynthetics, Dec., 1988, GRI, Drexel University, Philadelphia, PA.
9. Koerner, G. R. and Koerner, R. M., "Biological Clogging in Leachate
Collection Systems," Proc. Aging and Durability of Geosynthetics, Dec.
1988, GRI, Drexel University, Philadelphia, PA.
10. Rollin, A., "Factors Influencing Geomembrane Seam Quality," Proc.
Geosynthetics '89, San Diego, CA, IFAI.
11. Peggs, I. D., "Failure and Regain of Geomembrane Lining System,"
Geotech. Fabrics Report, Vol. 6, No. 6, 1988, pp. 13-16.
GEOTEXTILE FILTERS
P-FML
S-FML
Figure 1 - Cross Section of Double Lined Landfill Facility Often Used for
Municipal/Industrial Solid Waste Disposal
376
-------
2'-0"
D
01
n
* ^
10 •
400 -jw —
__ - J» H
200-1
nJ
'fa
!y
*• ••"
> o
a
a
+
o
*
p
D
J
O
a
«
!l>"J'™"
Q
D
3
H
a
*
u
*
c;
"~ET~
»
Q
*
E
Q
Fl
c
...
1 2
TRUE STRflIN (7o)
El G1
* G2
a G3
« G4
- Pullout of 60mil HOPE with 350 Ib/ft2 Surcharge
Figure 2 - Details of Test Facility and FML Test Results for Anchor Trench
Pullout Experiments
377
-------
TIME
Figure 3 - Long-Term Flow Column and Usual Trend in Flow Rate Data versus
Time
Figure 4 - Viable Bacterial Titer of Each of the Leachate Samples as
Determined by Multiplying Viability by the Total Direct Count TDC,
after Rios and Gealt<8)
378
-------
ATTENUATION OF PRIORITY POLLUTANTS CODISPOSED
WITH MSW IN SIMULATED LANDFILLS
by: Dr. Frederick G. Pohland
Department of Civil Engineering
University of Pittsburgh
Pittsburgh, PA 15261
and
Dr. Wendall H. Cross, Dr. Joseph P.
Gould and Ms. Debra R. Reinhart
School of Civil Engineering
Georgia Institute of Technology
Atlanta, GA 30332
ABSTRACT
Organic and inorganic priority pollutants codisposed with shredded
municipal solid waste (MSW) in ten pilot-scale simulated landfill columns,
operated under single pass leaching or leachate recycle, were capable of being
attenuated by microbially-mediated landfill stabilization processes. The
results of detailed investigations have indicated that inorganic heavy metals
(Cd, Cr, Hg, Ni, Pb and Zn) were subject to a complex array of attenuation
mechanisms within the MSW matrix, including precipitation, encapsulation,
complexation, reduction, adsorption and ion exchange. Similarly, the major
classes of organic priority pollutants (aromatic hydrocarbons, halogenated
hydrocarbons, pesticides, phenols and phthalate esters) were attenuated mainly
by sorption, volatilization and bioassimilation and release of identifiable
by-products within the leachate and gas transport phases.
Collectively these in situ processes constitute the assimilative capacity
of a landfill for priority pollutants, the magnitude of which is dependent on
microbial viability during the sequential phases of landfill stabilization as
affected by loading intensity and contact opportunity. Data are presented to
demonstrate some of the principal assimilative mechanisms as well as the
efficacy of in situ process control through leachate and gas management. Based
upon these findings, a landfill management strategy is proposed for landfills
receiving inputs of organic and inorganic priority pollutants.
379
-------
INTRODUCTION
There exist few more complex and challenging problems than those associated
with the management of municipal and industrial solid wastes. Because these
wastes often contain toxic and hazardous substances, they may become a threat
to human health and the environment unless proper treatment and disposal
techniques are employed. Of the array of options available for such treatment
and disposal, codisposal in sanitary landfills is probably the most prevalent,
considering the magnitude of household and small quantity generator sources.
Therefore, it is imperative that the efficacy of codisposal of hazardous wastes
with MSW be properly assessed, and that decisions on codisposal be approached
in as scientifically and technically sound a manner as possible.
One approach to such an assessment is to examine the progress of waste
stabilization with time and under various operational conditions, either in the
laboratory or in the field. Laboratory or pilot-scale investigations are often
more cost effective and permit evaluations with greater operational control and
less parametric variability. Hence, pilot-scale simulations were chosen for a
3-year investigation on the behavior and fate of selected inorganic and organic
priority pollutants codisposed with shredded MSW under the influence of single
pass leaching or leachate recycle. This paper presents a progress report on
some of the research results to date.
EXPERIMENTAL PROCEDURE
Column Construction and Loading
Since the behavior and fate of toxic inorganic and organic compounds within
a landfill setting are controlled by the related effects of various mass trans-
fer and removal mechanisms, including solubilization, sorption, volatilization,
chemical transformation and bioassimilation, it was necessary to develop an
experimental protocol sufficient to achieve an adequate description and control
of the waste mass and the gas and leachate transport phases. Accordingly,
shredded MSW was selected as the primary waste matrix, and was augmented by
heavy metal sludge and selected classes of organic priority pollutants. In
addition, to provide a range of loadings under both single pass leaching and
leachate recycle, 10 simulated landfill columns were constructed and prepared
for operation as indicated in Figure 1 and Table 1.
The simulated landfill columns were constructed from 0.3-cm steel
cylinders, 0.9 m in diameter and 3 m in height. The columns were lined with 30-
mil HOPE which was supported by an underdrain system composed of 20 cm of gravel
and 2.5 cm of sand. The underdrain system allowed collection of leachate for
sampling, discard (single pass columns) or recirculation (recycle columns)
through a perforated pipe distribution system installed after loading. Moisture
addition (or recycle) was facilitated through this pipe distribution system.
A leachate sight glass was also connected to the underdrain system as well
as the headspace under the cover. Gas generated during landfill stabilization
380
-------
1.22m
1.83m
0.61m
LEGEND
1 GAS METER
2 TEMPERATURE INDICATOR
3 GAS SAMPLING VALVE
4 GAS TRAP
5 CHECK VALVE
6 PRESSURE GUAGE
7 DISTRIBUTOR ARM
8 RECYCLE PUMP
9 FLANGE
10 THERMOCOUPLE
11 HOPE LINER
12 IN-LINE FILTER
13 STEEL
14 LEACHATE DRAIN
15 LIQUID SAMPLE PORT
16 LIQUID LEVEL CONTROL
17 GRAVEL, SAND, AND
GEOTEXTILE LAYERS
18 GEOTEXTILE, SAND,
GEOTEXTILE, AND GRAVEL
LAYERS
19 110 V AC
20 110 V AC
TO PUMP
21 110 V AC FROM
LIQUID LEVEL CONTROL
22 VENT TO ATMOSPHERE
23 SHREDDED REFUSE
® BALL VALVE
RECYCLE UNIT
SINGLE PASS UNIT
Figure 1. Construction and Operational Features of Simulated Landfill Columns
was collected from this system and measured with an automatic gas-displacement
meter.
Each of the 10 columns received 42 individual 9-kg batches of shredded MSW
placed and compacted over an 8-hour period. Columns 1 and 2, designated control
columns (Table 1) , received only shredded MSW and an equivalent amount of saw-
dust also added to the other test columns because of its use to facilitate
homogeneity in the sludge-loaded columns. This loading strategy provided five
pairs, each pair identically loaded, but operated with either single pass
leaching or leachate recycle.
Batches of metal sludge were prepared by mixing with sawdust and augmenting
with metal oxides as necessary to deliver low, medium and high loadings (Table
1) . The organic priority pollutants (Table 1) were mixed together and added at
the surface of the first 30 cm of compacted MSW to increase the potential for
detection during operations. After column loading, an 8-cm layer of washed pea
gravel was added to the surface of the MSW, the columns were sealed, and initial
381
-------
3
I
53
M
£
•§
o-g
•. a
il
II
* =8 *
ro M
OJ M (Nl c3
>«.
fl) —* «4
O) CD O) O)
lit
'1!
- 2.
I-S
1'
a. ro
§
8.-I
ITS
382
-------
moisture (tap water) was introduced to satisfy field capacity and encourage
leaching. Once leaching commenced, the five single pass columns received
periodic moisture additions equivalent to local annual rainfall (123 cm),
whereas the five recycle columns were subjected to leachate recycle with
moisture additions only of quantity necessary to replace that removed for
sampling and analysis.
Analytical Methods
Leachate and gas samples were analyzed routinely for both gross parameters
and priority pollutants once sufficient quantities were produced to permit
analysis. Leachate samples were subjected to immediate analyses for pH,
conductivity, chemical oxygen demand (COD), and oxidation-reduction potential
(ORP), supplemented by biochemical oxygen demand (BOD5) , total organic carbon
(TOG), volatile organic acids (TVA), alkalinity, nutrients (N,P), selected
anions (Cl~, SOA~2, S"2, Br") and cations (Ca+2, Mg+2, Na+, K+, Li+) , heavy metals
(Cd, Cr, Fe, Pb, Hg, Ni, Zn) and organic priority pollutants (Table 1) according
to standard techniques. Gas samples were analyzed for C02, 02, N2, H2, CHA and
volatile organic priority pollutants using GC and GC-MS techniques.
EXPERIMENTAL RESULTS AND DISCUSSION
Since the primary purpose of the research investigations was to evaluate
the fate of selected organic and inorganic priority pollutants cpdisposed with
MSW under the influence of single pass leaching and leachate recycle, emphasis
here is placed on these pollutants as the simulated landfills progressed through
acid formation and into the methane fermentation phase of landfill stabiliza-
tion. Accordingly, experimental data are presented to reveal potential
attenuating mechanisms operative during these phases and contributing to the
overall assimilative capacity of the landfill environment.
Progress of Landfill Stabilization
The progress of landfill stabilization can be described by certain
indicator parameters (1). For purposes here, gas production and leachate COD
and TVA concentrations and pH were selected as indicated for the single pass
and recycle columns in Figure 2. As illustrated in this figure, acid formation
(acid phase) became prominent and was intentionally sustained for the majority
of the report period. Reductions in leachate strength (COD and TVA) or
increases in gas production, as well as some increase in pH, occurred only after
the onset of methane fermentation (methane phase), which was initiated by
digested sludge seeding with neutralization and took effect on about Day 720.
Thereafter, leachate and gas quality changed dramatically, particularly for the
control columns (1 CR, 2 C) without heavy metal and organic priority pollutant
codisposal.
The differences in results between the individual single pass and recycle
test columns also are indicative of loading implications, particularly with
respect to the heavy metals, and the washout by single pass operations as
contrasted to leachate recycle. The apparent release of leachate constituents
through washout contributed to a decrease in concentrations, but did not greatly
383
-------
W S0.n
+j
o
J 40.
3
15 30.-4
n 20.-
10.-
ti
I
0.
100.
»-o 1 C
o-o 5 0
*-* 70L
*-* 9 OM
«—f 10 OH
o-o 2 C
o-a 3 0
A-A 4 OL
«—» 5 OM
*-* 8 OH
smr.LC
o I C
aao
• 4 01
« 5 OW
• 8 OH
200. 400. 600. 800. 1000.
Time since loading, days
1200. 0.
200. 400. 600. 800. 1000.
Time since loading, days
1200.
Figure
2. Changes in Gas Production and Leachate COD, TVA and pH during
Simulated Landfill Investigations
384
-------
enhance the extent of stabilization in the single pass columns where heavy
metal and organic priority pollutants were admixed. With recycle, leachate
concentrations and gas production tended to stabilize initially and then
decreased or increased, respectively, when methane fermentation ensued. There-
fore, the containment for conversion of these leachate constituents to gas, by
better contact and controlled in situ treatment, emerged as a major advantage
of recycle over single pass operations.
Mechanisms of Change in Inorganic Priority Pollutants (Heavy Metals)
A principal focus of these investigations was the behavior and fate of the
toxic heavy metals in terms of their mobility in the leachate (or gas) phases
as landfill stabilization progressed. Therefore, special emphasis was placed
on analyses of these metals and potential attenuating mechanisms.
Hydroxide Precipitation--
Of the heavy metals added to the sludge loaded columns (Table 1), only
Cr+3 was subjected to significant solubility control by hydroxide. Based upon
its extremely low solubility (pKso = 30.5), precipitation of the chromic ion as
the hydroxide, Cr(OH)3, at equilibrium concentrations of chromium on the order
of 1 mg/L, would be attained even at pH levels as low as 5.0. Therefore, even
during the acid phase of landfill stabilization (pH 5.5 to 6.0), the mobility
of trivalent chromium would be limited by hydroxide precipitation. This control
was demonstrated by leachate chromium analyses (Figure 3), although some, initial
fluctuation occurred for both the single pass and recycle columns. These
fluctuations coincided with early fluctuations in pH and, no doubt, were
affected somewhat by the respective column operations and the possibility of
short-circuiting. Moreover, since leachate chromium concentrations decreased
sharply to very low levels during the acid phase and into the methane phase,
coupled with the absence of other important precipitants for chromium in typical
landfill leachates, hydroxide precipitation was considered the primary
attenuating mechanism for this heavy metal.
30.
-------
Redox Processes --
The ORP is particularly important in defining the chemical character of
the landfill environment. As illustrated in Figure 4, negative leachate redox
values provide reducing conditions that mediate the behavior of many of the
codisposed inorganic species. The impact of this condition can be either
directly through a modification of the nature of the pollutants with a change
in their mobility, or indirectly by reductive generation of a potent
precipitant.
•400 L
SINGLE PASS
ate
030
• 4 01
+ S OU
• tl Oil
800. 900. 1000. 1100.
Tlmo since loading, days
1200.
0.0
700.
800. 900. 1000. 1100.
Time since loading, days
1200.
Figure 4. Changes in Leachate ORP, Sulfate and Sulfide during
Simulated Landfill Investigations
386
-------
An important example of the latter indirect case is the reduction of
sulfate to sulfide and subsequent precipitation of very sparingly soluble metal
sulfides. This is an extremely important attenuating mechanism in the landfill
environment since many heavy metals are precipitated and removed from the
leachate, particularly when leachate recycle provides an additional mechanism
by filtration through the waste matrix. However, the efficiency of attenuation
is also a function of the availability of sulfides, primarily provided by in
situ microbially-mediated reduction of sulfates.
As also illustrated in Figure 4, sulfate reduction to sulfide tended to
coincide with the onset of methane fermentation, most noticeably in the least
loaded recycle columns. The associated impact of available sulfide on the
leachate lead and zinc concentrations is dramatically illustrated in Figure 5.
Both metals rapidly decreased in concentration in correspondence with the
reduction of sulfate to sulf ide. Therefore, the important attenuating mechanism
for these heavy metals is precipitation as PbS (pKso = 28.0) and ZnS (pKso =
24.8).
Similarly, a direct impact of reducing conditions on attenuation of toxic
heavy metals is also apparent with mercury (Figure 5). Following an initial
peak, leachate mercury decreased rapidly to concentrations below 100 /zg/L and
eventually stabilized at about 10 /ig/L. However, in spite of the exceptionally
low solubility of mercuric sulfide (pKso = 52.5), leachate mercury concentra-
tions tended to exhibit no additional decrease with sulfate reduction.
Collectively, these observations are consistent with reduction of mercuric ion
to metallic mercury, which has been shown to have a water solubility of 5 to 30
jzg/L (2). This behavior has also been documented for bottom sediments (3),
where similar reducing conditions prevail. Hence, reduction to volatile mercury
and subsequent sorptive containment within the waste matrix would constitute
important attenuating mechanisms for this heavy metal.
Other Precipitants --
During the acid phase, when significant sulfate reduction and sulfide
generation were not in evidence, control of metal solubility may have involved
such generally abundant anions as chloride, sulfate, carbonate and possibly
phosphate. Within the landfill environment, these potential precipitants may
have only transient significance, particularly where leachate recycle and
promotion of sulfate reduction and reducing conditions are provided under
controlled operating conditions.
Heterogeneous Physical-Chemical Processes --
Codisposal of alkaline metal sludges with MSW presents the potential for
generation of chemical microenvironments within the waste matrix, and the
opportunity for a complex array of heterogeneous (liquid/solid) reactions of
major importance to the overall attenuation of heavy metals. These micro-
environments provide a source of alkalinity and acid neutralizing capacity
which will lessen leachate metal mobility, while the hydroxide sludge will react
with leachate anions such as sulfide, phosphate and sulfate to develop an
encapsulating layer capable of impeding dissolution of the sludge metals into
the leachate. Therefore, these alkaline microenvironments will contribute
387
-------
BtCYCtC.
1 CD
6 OS
7 OtR
9 OUR
10 OHR
0.
200. • 400. 600. 800. 1000.
Time since loading, days
1200.
0.
2000.'
n 700. 400. 600. 800. 1000. 1200.
1500. -
1000,-
500.
0.
200. 400. 600. 800. 1000.
Time since loading, days
1200.
Figure 5. Changes in Leachate Lead, Zinc and Mercury during
Simulated Landfill Investigations
effectively to in situ attenuation during the acid phase of landfill stabiliza-
tion when it is most needed, whereas this contribution will diminish as
encapsulation proceeds and other attenuating mechanisms, such as sulfide
precipitation, gain prominence.
The MSW itself also provides abundant surface area for sorptive inter-
actions with leachate constituents. These processes include physical adsorption
on the solid surface, ion exchange (particularly on soil), chemisorption by
388
-------
complexation with insoluble ligands associated with the MSW, and physical
containment in transiently stagnant void volume liquid. In addition,
complexation of metals by soluble ligands, particularly with moderate to high
molecular weight humic-like substances, may enhance sorption by incorporating
the metal into a relatively hydrophobic molecule. Current analyses on leachate
aromatic hydroxyl concentrations and molecular weight distributions are being
provided to help determine the significance of this latter attenuating
mechanism.
Mechanisms of Change in Organic Priority Pollutants
Since the landfill environment provides numerous possibilities for
contaminant transport and transformation, an equally important focus of these
investigations was the behavior and fate of the organic priority pollutants
during both acid and methane phases of landfill stabilization. In this case,
mobility and possible attenuation were not as clear as for the heavy metals.
However, scrutiny of the data (Figures 6 through 8) suggests that the 12
compounds added to the test columns (Table 1) can be divided into four general
groups in terms of their relative mobility and reactivity.
Four of the compounds, dieldrin, hexachlorobenzene, bis-2-ethylhexyl-
phthalate and 7-1,2,3,4,5,6-hexachlorocyclohexane (lindane) essentially have
not emerged in either the leachate or gas phases although in the case of
lindane, three samples from over 600 analyzed revealed its presence at about 10
to 20 /ig/L. Three of the compounds, dibromome thane, 2-nitrophenol and
nitrobenzene, appeared early in appreciable concentrations and then decreased
to below detection limits (Figure 6), particularly after methane fermentation
had been established. In contrast,1,2,4-trichlorobenzene has been detected
regularly in the leachate (Figure 7) , but at relatively low levels (< 1 mg/L) .
The remaining four compounds, naphthalene, 1,4-dichlorobenzene, 2,4-dichloro-
phenol and 1,1,2-trichloroethene, appeared in the leachate in varying
concentrations throughout the project period (Figure 8).
Headspace gas analysis has indicated the presence of only three of the test
compounds; trichloroethene fairly regularly at 1 to 20 jug/m3, and naphthalene
and dichlorobenzene sporadically at equally low concentrations. Moreover,
evidence of biodegradation of at least two of the added organic priority
pollutants has been obtained. Bromide ion (Figure 9) has been detected in the
leachates from all test columns; slowly leaching from the single pass columns
due to washout, while remaining relatively constant in the leachates of the
recycle columns (-150 mg/L). Since no leachate bromide was detected for the
control columns, and appeared from the test columns coincident with the decrease
in dibromomethane (Figure 6), microbially-mediated debromination would be
suggested. Likewise, vinyl chloride (as high as 300 /*g/L) has been detected in
the headspace gas of the test columns after reductions in leachate
trichloroethene were observed, suggesting partial bioconversion to this known
intermediate (4).
Primary Attenuating Mechanisms --
In addition to the suggestion of bioconversion of some of the organic
priority pollutants into detectable intermediates and reaction products, several
389
-------
SINGLE PASS.
o 2 C
0 3 O
» 4 OL
+ 5 OU
* B OH
0.
200. 400. 600. 800. 1000.
Time since loading, days
1200. 0.
200. 400. 600. 800. 1000.
Time since loading, days
1200.
Figure
6. Changes in Leachate, Dibromomethane, 2-Nitrophenol and Nitrobenzene
during Simulated Landfill Investigations
other potential attenuating mechanisms could be envisioned. Inspection of the
concentration data, as well as the mass of each added compound released from
the test columns (Table 2) , provides some basis for deducing possible mechanisms
controlling relative mobility in the leachate or gas. Although £here does not
appear to be a variation in perceptible patterns of release of the organic
priority pollutants between the single pass and recycle columns, the operational
procedure in either case would determine whether the leached compounds would
390
-------
UJ
1
c_>
LJJ
O
i
CO
o
O£.
Q
O
Ul
0
&
111
OK
01 CO
UJ 5
GO —1
CO O
O O
(O — 1
rf II
rz* s^
r— O
^ Z
| ^
_I _j
O
a. a
UJ
^ <
|i
a.: co
UJ
u =
3 o
0£ 0£
O "-
CO
UJ
(__
_J
13
Q
*>
o
LL
O
CO
UJ
3
i—
ujapug;
a^BjBtpqd
3T19 Z U9 q
OJOTUOBX9H
suBpufT;
euazuaq
— o aoxqoTax— -17 * 3 ' j
•[ouaqdoj^Tji-^
auazuaq
••OJt OTUO"E(T™^7 T
auazuaaoaaxM
3Ua dBN
+
d OTUOTfT-17'7
aUBUaaOJOTUOTJT
*
A^j^uapi uuinjog
II II II II II
II It II II II
ii ii ii *4~ r- ii
" O
V
^/
it t i O '"* O *•* O '•^
in *^ CM CM N. %o
o o o
t i ii ^D ^^ C3 ^^ O ^^
v^ vx v^
^D ^^ CZ3 *"* C3 *"*
^D ^— CD O* ^5 r1*
ii ii C\l *^ O • O •
» » S.X >^
O ^N 0 ^ O '^
ii ii «— ^" r^in in'O
»— • in » N- •
C5 ^% ^D ^^ O ^^ C3 ^^ CD ^^
CDO OO CD -J- O O OCM
^T ,-T ^^ ^^
O • *"* O ^N O *^
SN- O sf O *-
»- . ro •
v— vx C? ^^ iN- ^^
. . .
§x*x ^3 ^N CO ^^
CO ON- O CM
*~ ^ O. ^f O> ^
^D ^^ ^D ^^ CD ^^
ON- O CD O CO
ii ii OK) CD jTO CD CM
•S* %O -*
** to ro
0 CJ O O S
T- CM ro st in
it 11 ii ii ii
11 ii it 11 ii
^•n' " XX
II II 11 *Q «— ""d" «—
• O • O
«- O O
N.X V
^^
^D ^^ ^D ^* CD ^"* CD *"* CD ^^
%oro I'-NO -OCM oro ro«—
ro • o CM • sa- • «- •
§**\ O ^^ O *"* O '•^ O *•*»
^o co CM o «~ in ro r*- ^3
-* ~
O«— OTO OCM OCM 0*4-
»t«— »*— •>•:— -*v- ».«—
CM-4- ON- O-« OCO Oo «- in • to • •* • IO,'~
N- O^-* v— v-' v-:^^ in
o *** o ^^ o ^^ o ,^* o ^^
ot- OTO OTO o*r oin
o« in*— ^is" ^o*" r^«—
t*» T- in co ^ *O t*-
•r- v^ T- a- CON- ^CM or- OCM
^^ ^^ TO t^> ^O
in co TO «- • CM
a OL ai
CZ —1 31 £ 3=
o o o o o
-o N. co o o
•
c
3
"o
o
~o
L.
c
o
0
c
"^
T3
UJ
CA
o
(/)
w
l_
o
• • M-
't-> "D
CO ?S 3
01 t- >
£ ° ~
O O> 55
•M
c -c c
c 0) "ra
- O l_ •!->
fi. 0 J=
S Q. Q-
O
-------
3.0
.2,0
s>
I
B9
J3
oW
o
0.0
RECYCLE.
Q 1 CR
O 6 OR
» 7 OLR
+ 9 OUR
« 10 OHR
SINGLE PASS.
D 2 C
030
* 4 OL
+ 5 OM
« 8 OH
0.
200. 400. 600. 800.
Time since loading, days
1200.
200. 400. 600. 800. 1000.
Time since loading, days
1200.
Figure 7. Changes in Leachate 1,2,4-Trichlorobenzene during
Simulated Landfill Investigations
escape due to washout (single pass) or be more likely subjected to additional
attenuation opportunity with recirculation (recycle).
Evidence to date indicates that the organic priority pollutants will be
influenced by the equilibrium conditions established at the point of contact
between the gas, liquid and solid phases. Supplemental studies with the test
compounds have demonstrated that other leachate constituents have enhanced the
solubility of the more hydrophobic molecules, but that physical sorption was
a primary attenuating mechanism. Moreover, a strong correlation of sorbability
existed with the octanol/water partition coefficient (Figure 10). Therefore,
sorption of the test compounds on the waste matrix could be attributed to
hydrophobic interactions and removal from the aqueous phase onto nonspecific
surface sites, particularly such MSW constituents as fats, oils, waxes, biomass,
humic-like substances, lignin, plastics and leather.
In the final analysis, only soluble, less hydrophobic compounds would be
expected to elute from the wastes in the test columns. Hence, compounds would
be expected to emerge in the approximate order of increasing affinity for the
waste. However, the degree of MSW stabilization and the corresponding
microbially-mediated phase (acid or methane) would influence this pattern,
i.e. , pH changes could impair sorption of ionizable compounds and the more
stable the waste, the greater the mobility. Hence, upon completion of the
stabilization process through methane fermentation and into final maturation,
the liquid (leachate) transport phase should be removed to preclude continued
fractionation of many of the residual organic priority pollutants from the waste
matrix. With leachate containment and recycle, such operational control can be
easily accommodated.
SUMMARY AND CONCLUSIONS
Simulated landfill columns, containing shredded MSW and any array of
inorganic and organic priority pollutants have been shown to possess a
significant assimilative capacity due to a variety of attenuating mechanisms.
392
-------
200.
400. 600.' 800. 1000
Time since loading, days
1200. 0.
240. 480. 720. 960.
Time since loading, days
1200.
Figure 8. Changes in Leachate Naphthalene, 1,4-Dichlorobenzene,
Trichloroethene and 2,4-Dichlorophenol during
Simulated Landfill Investigations
393
-------
400.
300.-
o
D Soil/Sediment Data
(Karlckhoff, 1981)
-b-
R»lu«« Beat Fit Curv* Plu* 86 Percent
Confidence Interval
D
1234
Log Octanol/Water Partition Coefficient
Figure 10. Relationship Between Log Octanol Water Partition
Coefficient and Log Sorption Partition Coefficient
Normalized to Carbon Content.
394
-------
Metal mobility was minimized by microbially-mediated physical and both
homogeneous and heterogeneous chemical processes, including hydroxide
precipitation, direct and indirect reductive removal, sorption, and retention
in stagnant intersticial void liquid. Associated specific attenuative
mechanisms, such as encapsulation, tended to function during operational phases
when other mechanisms, such as sulfide precipitation, were unavailable.
Organic priority pollutants appeared to be removed primarily by sorption
and/or biodegradation, with identifiable metabolic intermediates and reaction
products being detected. The sorptive capacity of MSW results in extremely Ipng
contact times within the landfill setting, and enhanced opportunity for
acclimation and further degradation of more recalcitrant compounds.
Leachate recycle offers the advantage of not only retaining leached
materials within the landfill system, but also providing better contact and
redistribution of components contained in the leachate, thus facilitating and
beneficiating the overall efficiency of stabilization.
Landfills containing MSW and operated with leachate recycle, can provide
opportunities for codisposal of organic and inorganic priority pollutants with
reasonable assurance that mobility and potential release can be controlled to
safeguard human health and the environment.
REFERENCES
1. Pohland, F. G. Dertien, J. T., and Ghosh, S. B. Leachate and Gas Quality
Changes During Landfill Stabilization of Municipal Refuse. In: Proc. 3rd
Intl. Symp. on Anaerobic Digestion, Boston, MA, 1983.
2. Hughes, W. L. A Physicochemical Rationale for the Biological Activity of
Mercury and Its Compounds. Ann. N.Y. Acad. Sci. 65:454, 1957.
3. Mercury in the Environment. An Epidemiological and Toxicological Appraisal.
L. Fribery and J. Vostal, [Eds]. CRC Press, Cleveland, OH, 1972.
4. Wilson, B. H., Smith, G. B. and Reas, J. F. Biotransformation of Selected
Alkylbenzenes and Halogenated Aliphatic Hydrocarbons in Methanogenic Aquifer
Material: A Microcosm Study, Environmental Science and Technology 20, 997-
1002,,1986.
5. Karickhoff, S. W. Semi-Empirical Estimation of Sorption of Hydrophobic
Pollutants on Natural Sediments and Soils, Chemosphere, 10(8), 833, 1981.
395
-------
SITE DEMONSTRATION OF HAZCON PROCESS
Paul R. de Percin
SITE Demonstration and Evaluation Branch
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio
ABSTRACT
In October 1987 the HAZCON stabilization process was tested
and evaluated at the Douglassville, Pa. superfund site. There
was extensive physical and chemical testing of the untreated and
treated waste samples. Stabilized wastes were also stored
underground on site as a field durability test. After 270 days
(July 1988) these wastes were dug up and the samples obtained
were tested for any changes in chemical and physical properties.
The wastes were reburied and will be sampled again in July 1989.
The durability of the treated samples was determined by
comparing-the 28-day and the 270-day test results. Test results
of the HAZCON 270-day samples indicated a slight loss in long-
term durability, but because of the data scatter this may not be
statistically important. There was some loss in UCS strength in
the low oil and grease samples (<5%) , but the high oil and grease
samples (16-25%) the UCS strength increased. TCLP leachate
concentrations for lead remained low (ppb) ; and the VOC, BNA and
PCB leachate concentrations appeared lower than the 28-day sample
results. A noticeable decrease in the porosity of the samples
was observed. Overall, there was not a major change in sample
characteristics and some durability was demonstrated.
396
-------
INTRODUCTION
The HAZCON solidification/stabilization treatment process
was demonstrated at the Douglassville, Pennsylvania superfund
site under the Superfund Innovative Technology Evaluation (SITE)
program in October 1987. Reliable field performance and cost
information were the major objectives of this demonstration. A
secondary objective was to study the long-term durability of the
treated waste under field conditions. The treated wastes
resulting from the HAZCON demonstration were stored on-site,
underground and isolated from the untreated waste.
One aspect of solidification/stabilization processes is that
the waste is not destroyed, but macro- and micro-encapsulated
(trapped) or fixated (reacted) in/to the stabilizer matrix.
Thus, in future years there is potential for the contaminants to
leach from the treated waste as the stabilizer degenerates.
Long-term durability or lifetime treatment effectiveness is a key
questions both regulatory agencies and responsible parties have
about stabilization.
After being stored on-site for nine months (July 1988),
sample cores were bored from the treated wastes and the samples
were tested using physical and chemical procedures. The sampling
and the sample testing procedures were the same as the procedures
used to obtain and test the one month samples. One and nine-
month sample test data were compared to determine if any
degradation of the treated waste was evident.
BACKGROUND
The Douglassville, Pa superfund site is a former oil
recovery facility covering about 50 acres near Pottstown, Pa.
There are six contaminated areas: two large lagoons referred to
as lagoon north (LAN) and lagoon south (LAS), an oily filter cake
storage area (FSA) , an oil drum storage area (DSA) , an area where
generated sludge was landfarmed (LFA), and the plant facility
area (PFA). The major contaminants at this site are oil and
grease, semivolatile organics (BNAs) and lead; minor contaminants
were PCBs and volatile organics (VOCs).
The HAZCON proprietary solidification process involves the
mixing of hazardous waste material and cement with a patented
nontoxic chemical called Chloranan. Chloranan is claimed to
neutralize the inhibiting effect that organic contaminants
normally have on the crystallization of cement-based materials.
For this treatment, the wastes were immobilized and bound into a
hardened, leach-resistant concrete-like mass. Waste from the
Douglassville superfund site was selected because of the high
levels of oil and grease (up to 25%) and lead (up to 2.2%) in the
FSA waste. This combination was considered very difficult to
397
-------
process successfully.
During the HAZCON SITE demonstration at Douglassville, five
cubic yards of treated waste was produce for each of five wastes.
The sixth waste (LAS) generated, twenty-five cubic yards of waste
as a test for equipment reliability. Waste was mixed with the
Chloranan at a ten to one (10:1) ratio and with the portland
cement at a one to one (1:1) ratio. Water was added as needed.
A 1:1 ratio of waste to cement is high, but it was believed
necessary because of the high level of organics and heavy metals
in the FSA waste. These mixing ratios were not changed for the
other wastes. The treated waste was poured into 1-cu-yd molds
(as a slurry) and allowed to harden. The molds were stripped
from the hardened waste after one to two days.
During the waste processing, the holes created excavating
the feed waste were enlarged to accommodate burial of the treated
waste blocks. Before the blocks for a particular waste area were
buried, the hole was lined with plastic (to prevent seepage of
contaminated water into the hole) and a one foot layer of clean
soil was deposited. After the blocks were placed into the
excavation hole, additional soil was added to cover the blocks.
Stakes were planted to identify the location of each block.
SAMPLING AND ANALYSIS PROCEDURES
One month and nine months after the blocks had been buried,
samples were collected from two blocks in each area for physical
property tests and from one block for chemical tests (soil
analyses and leaching tests) and weathering tests.
Samples of solidified waste were collected using a rotary
rig with diamond-tipped core barrel. The outside diameters for
the cores were 7 cm for unconfined compressive strength,
permeability, and leaching tests; and 4.5 cm for wet/dry and
freeze/thaw tests. These multiple samples of different diameters
required the boring of at least two holes in each of twelve
blocks, two blocks in each area. Cores were removed from the
core barrel and sealed in aluminum foil, placed in glass jars
closed with a custody seal, placed in zip-loc bags, and stored
with ice packs during storage and shipment to the laboratory.
Both standard methods from SW-846 and experimental test
methods were used to evaluate the core samples. The test
procedures are listed in Table 1 below:
398
-------
Table 1
Physical Tests
Chemical Tests
Micro Scale Tests
Moisture
Bulk Density
Unconfined Compressive Strength (UCS)
Wet/Dry Weathering
Freeze/Thaw Weathering
Permeability - Falling Head
Leaching - TCLP - VOCs, BNAs, PCBs
Priority Pollutant Metals - Pb, Cr, Ni,
Zn, Cu, Cd
Total Oil and Grease
X-Ray Diffraction
Scanning Electron Microscopy
Unconfined compressive strengths were performed on the Wet/Dry
test specimens and controls after the twelve-cycle weathering
tests, except for LAS where a permeability test was performed on
the test specimen. Unconfined compressive strengths were
performed on the Freeze/Thaw test specimens and controls after
the twelve-cycle weathering tests, except for DSA and FSA where
permeabilities were performed on the test specimens. The listed
metals were the only metals detected in significant quantities.
TEST RESULTS
The results are presented in three parts; physical, chemical
and microstructural properties. The physical properties include
moisture content, bulk density, permeability, unconfined
compressive strength, and weathering effects. The chemical
analyses were for treated soil and TCLP leachate analyte
concentrations for volatile organic compounds (VOC), base
neutral/acid extractables (BNA), six priority pollutant metals,
and polychlorinated biphenyls (PCB). Microstructural analysis
included X-ray diffraction and.scanning electron microscopy.
It should be noted that the 28-day (i.e., 1 month) sample data is
from the report "Technology Evaluation Report, SITE Program
Demonstration Test, HAZCON Solidification, Douglassville,
Pennsylvania - Volume I (EPA/540/5-89-001a) .
Physical Properties
The results of the physical tests are summarized in Tables
2, 3 and 4; and are discussed below.
399
-------
The moisture content ranged from 8.8% by wt for PFA-1 to
20.4% by wt for FSA-3. All the values are approximately 1% to 2%
by wt less than for the 28-day samples, with the largest
reduction for FSA. This may indicate that the blocks continued
to cure thus reducing the free moisture content.
The bulk densities were essentially equivalent to those
obtained in the 28-day cores. They ranged from 1.51 g/ml for FSA
to 2.05g/ml for PFA. The values at LAS and LAN increased by 5%
by wt and 3% by wt, respectively. This increase is most likely
due to sample variations.
The permeabilities ranged from 2.2xlO"10 cm/sec at DSA-3 to
2.3X10"7 cm/sec at DSA-1. These values are in the same general
range as the 28-day cores but appear to be slightly larger. The
apparent increases in permeability are larger than the decreases,
and eight of the .twelve samples increased. This trend is not
conclusive as data on permeability variation within individual
blocks is needed. This apparent increase could be explained by a
small breakdown in the porous structure of the blocks, which
would allow interconnection of the pores and a path for water
flow.
The unconfined compressive strength (UCS) ranged from 230
psi at FSA-3 to 1170 psi at PFA-3. Comparing the results to
those of the 28-day cores, nine of twelve samples had a lower
UCS. The largest losses occurred for the areas with low oil and
grease content. Averaging all of the values across the six plant
areas shows an overall decrease of about 20%. This apparent
•reduction in UCS is consistent with an increase in permeability,
if the solid structure is deteriorating.
The oil and grease content in the treated soil ranged from
0.48% by wt at DSA-3 to 12.1% by wt for FSA-1. All values are
consistent with 28-day cores. The samples are also consistent on
a material balance basis to the untreated soil oil and grease
concentrations.
The wet/dry weathering tests weight losses of the test
specimens were slightly greater than for the 28-day samples, with
losses ranging from 0.7% to 1.7% by wt. The unconfined
corapressive strength for the long-term and 28-day samples were
the same order of magnitude. A permeability was performed on the
weathered test specimen from LAS-LM1 and was 5.5x10" cm/sec
which agreed with the unweathered long-term sample but is greater
by a factor of 30 compared to the unweathered 28-day cores.
The freeze/thaw weathering test weight losses of the test
specimens was nearly double that of the 28-day cores, with values
ranging from 0.99% by wt at PFA-3 to 3.19% by wt at LAS-LM1. The
loss for the controls was only slightly greater than for the 28-
400
-------
day core controls. Therefore, unlike the wet/dry samples, the
corrected cumulative weight loss (wt. loss of test specimens
minus wt loss of control each divided by its dry starting weight)
for the long-term samples were larger than for the 28-day cores.
This indicates, along with UCS _and permeability results, that the
freeze/thaw cycles at Douglassville during the winter of 1987-
1988 may have damaged the internal structure of the test blocks.
The unconfined compressive strengths of the 12-cycle weathered
test specimens ranged from 430 psi at LFA-5 to 860 psi at LAN-1.
These values on average are larger than the UCS values from the
28-day weathered cores. However, individual UCS values are so
scattered that a comparison of UCS values between long-term and
28-day cores shows them to be approximately equal.
Chemical Properties
The results of the chemical tests are summarized in Tables 5
and 6; and are discussed below.
The total PCBs ranged from <0.64 mg/kg at DSA-3 to 5.9 mg/kg
at FSA-1. These values are considerably lower than anticipated
based upon untreated soil sample analysis performed during the
Demonstration Test. Analytical difficulties were encountered.
Analysis on 28-day core samples were not performed. PCBs were
not detected in the TCLP leachates.
The VOC content ranged from 210 ug/kg at DSA-3 to 70,800
ug/kg at FSA-1. These values are considerably lower than i the
28-day cores where the values ranged from 3,490 ug/kg at LAN-1 to
108,700 ug/kg at FSA-1. The VOC content in the TCLP leachate
ranged from 13 ug/1 for LAN-1 to 493 ug/1 for LAS-LM1. A
calculation of the migration potential, which is defined as
analyte weight in the leachate divided by analyte weight in the
solid, provided erratic results, but appear to be equivalent to
the 28-day cores. The Demonstration Test results showed
equivalent migration potentials for the treated and untreated
soils. Therefore, no immobilization of VOCs could be seen.
The total metals content in the long-term treated soil
samples was equivalent to those obtained for the 28-day cores
with lead values ranging from 980 mg/kg at DSA-3 to 8,600 mg/kg
at FSA-1. As previously, the primary metal contaminant is lead.
Leachate results from the TCLP leach test showed total metals
content ranging from 3 ug/1 at DSA-3 to 120 ug/1 at LFA-5. The
FSA-1 results were quite low, 39 ug/1, wit only lead detected in
the leachate. A calculation of migration potential produced
values equivalent to those for the 28-day samples, which showed
high immobilization of the metals. The migration potential for
lead ranged from 2.1x10 ug leachate/ug soil for LAN-1 to
6.3x10 ug leachate/ug soil for LFA-5.
401
-------
The quantity of BNAs for the long-term soil samples was
considerably less than reported for the 28-day cores. However,
the detection limits used by the laboratory contractor were a
factor of 50 to 100 higher due to laboratory difficulties. For
example, at FSA-1, no BNAs were_ detected in the long-term samples
while the 28-day cores showed a value of 210 mg/kg. However, the
long-term and 28-day samples had TCLP leachate concentrations
that were equivalent. The long-term leachate values ranged from
35 ug/1 for PFA-3 to 3,109 ug/1 for FSA-1, the latter being about
the same as that of the 28-day cores. Therefore, it appears that
no further immobilization of these organics occurred.
The oil and grease levels in the long-term sample leachates
were below 1.0 mg/1, which is less than for the 28-day samples,
where the values were primarily between 2.0 mg/1 and 4.0 mg/1.
Oil and grease was detected in the TCLP leachates only for LAN-1
and FSA-1. The other four samples showed values below the
detection limits for oil and grease of 0.4 mg/1. It appears that
the mobility for oil and grease is equivalent for the treated and
untreated soils. This is an apparent improvement over the
earlier results.
Analysis of clean soil samples taken from the backfill used
around the buried blocks in each plant area showed no
distinguishable change in its measured physical properties or
organic content. No VOC, BNA or PCBs were detected in the soil.
Microstructural Analysis
Analyses of thirteen samples were performed on a
microstructural scale. The samples were studied by scanning
electron microscopy (SEM), optical microscopy (OM), and X-ray
diffraction (XRD). The type of information to be obtained from
the tests were:
Microscopy - crystal appearances, porosity,
fractures, and the presence of unaltered soil
crystalline structure of the
The results can be summarized as follows:
X-ray Diffractometry
hydration products.
The samples were morphologically poorly defined material
compared to hardened portland cement/soil mixtures which do
not contain waste material.
The porosity of the long-term samples were low, having
decreased noticeably from examinations of the 28-day
samples. Decreased porosity suggests that either
microstructures vary significantly or that dehydration
402
-------
reactions continued after the previous samples were
analyzed.
Mixing appears to be poor as observed previously. Brownish
colored aggregates and opajgue particles, likely to be
untreated waste material, could be observed in both long-
term samples and those of the 28-day cores.
X-ray diffraction analyses continue to support that
encapsulation is the major process contributing to
stabilization of these contaminated soils.
As in the 28-day samples, previously examined, unhydrated
clinker was detected in the long-term samples, suggesting
poor mixing and/or incomplete hydration.
CONCLUSIONS
Based on the comparison of the long-term and 28-day sample
test results, the following conclusions were drawn:
* The priority pollutant metals were immobilized, with
the long-term results equivalent to the 28-day results.
* The VOCs and BNAs leaching did not change.
* The clean soil backfill appears not to have been
contaminated by the buried blocks.
* A small deterioration in physical properties appears to
have occurred. This was concluded from a series of
small changes, which are as follows:
a decrease in UCS for samples of low oil and
grease content
an increase in permeability of the samples
increased weight losses of test specimens and
controls during the wet/dry and freeze/thaw
weathering tests
large increase (double) in weight loss of
freeze/thaw test specimens.
* The microstructural results are similar to earlier
results except that the porosity was low having
increased noticeably from the earlier samples.
The general conclusion is that the 9-month test results were
403
-------
similar to the earlier result, but where they deviated the
results appear to show a small reduction in the physical
properties.
Further long-term samples .are scheduled to be collected in
June 1989 and analyzed as described in this study. Comparison of
three sets of data will allow a stronger conclusion to be drawn
about the long term durability of solidification/stabilization
processes.
404
-------
trt
S
i
u_
0
IU
ot
IU
&
£
PHYSICAL
(M
IU
•J
1
I
§
« X
i
•—
"2 .i **•
ill
S h. ?
a s I
5 o *-
s
p^ V
15
1*6
£ -
II
?£
-. i
O ci
L
i
o. <
!>.»«« COM CMK> MN «-O M^
•<) ^T ep O> eo •- u>K. NCM ino
ineo «7in •- ro o>o KM o- S
M N» 40 *• ^" co ** N. •o in
cMo (O O> 03 •- -O
•*in -or-- M<- •* m coo inin
•-•- ^-»- CMCM »-»- «- •-»-
MK inp RS inia SS «l8
»- -* in h- CMCM coco (O'- eo I--
OO Kl CM OO -J- O -* 1M OO
JQO •*•"* '"t KO OO OM
£ S 3 S E 5 S S S S 35
57* o •» 0 in
l^>* v*i tin to <
-------
TABLE 3. WEATHERING TEST RESULTS - WET/DRY
'
Plant
Area
DSA-3
LAH-1
FSA-1
LFA-5
PFA-3
LAS-LH-1
Average
Long-term
Ut. Loss, X
Specimen Control
1.50 1.33
0.80 0.66
0.86 0.80
1.71 1.40
0.91 0.73
1.42 1.00
1.20 0.99
UCS, psi
Specimen Control
830 470
890 500
340 430
250 230
2290 1270
(a) 830
920 622
a) Permeability performed on weathered sample resulted
Plant
Area
DSA-3
LAH-1
FSA-1
LFA-5
PFA-3
LAS-LH-1
Average
TABLE
Long-term
Wt. Lose, X
Specimen Control
1.85 1.73
1.55 0.62
1.48 1.04
2.89 1.28
0.99 0.88
3.19 1.00
1.99 1.09
28-day
Ut. Loss, X UCS,
Specimen Control Specimen
0.90 0.88 1150
0.74 0.77 180
0.93 0.75 340
1.53 1.15 1230
0.73 0.66 1170
1.05 0.84 400
0.98 0.84 739
in 5.5x10"8 cm/sec.
psi
Control
750
610
330
500
1190
450
638
4. WEATHERING TEST RESULTS - FREEZE/THAW
UCS, psi
Specimen Control
(a) 1530
860 420
(b) 520
430 870
750 620
470 770
628 788
28-day
Ut. Loss, X UCS,
Specimen Control Specimen
1.29 1.24 660
1.07 0.60 370
0.53 0.53 400
2.17 1.20 520
0.58 0.49 980
0.95 0.73 210
1.10 0.80 523
psi
Control
1020
250
210
320
350
400
425
a) Panaeability performed on weathered sample resulted in 4.0x10' on/sec.
b) Permeability performed on weathered sample resulted in 7.1x10 cm/sec.
406
-------
S
1
CD
U
OL
i
UJ
X-
_J
_J
u
X
uu
ac
o
in
ui
i
i
00
CM
E
4)
i
+ S ci
f — • O
£55
* 8 s
8.
a.
u
c
X
*
c
5
8
S - S
— ex:
*rf **- O
« *-. ro
&• C 4)
01 C —
— *••
x 8. S
c
c
g
>
0.
CJ
»—
c
§
te
***
Cfl
(J
OL
J
Q.
•O
O
»
5
««
3
41
**
a
u
a
±
a,
3
_^
to
|
i.
T3
O
3
^
5?
«T
a
£
S
O*
o>
^
(0
"8
*J
0
«>
u
ff
V
a
f
a
a
.
"o
(A
U
1
I
in to o o in
•". ^ S S R S
iX o o o d o
§T- &t |W
O Q O
*^ X> •"
— *
0 O 0 0 0 0
in O- o K) in ry
Kl »» h- O> to «»
in" m to K? in" ^
3 «M *-
Kl 00 tO
AJ ro o *o -^ h*
• •••••
>» 0 0 U\ 0 to o> o o to
^r •- o o^ o o
o o o o o »-
«- Q Q O O O.
CM oo oo in in in
g" S
333333
>O Kl t> •- K IO
?.* in «- to •*
Ki »- -- in to 5
• 111*1
5 3 35 2 2 S
o -J u. -J a. -J
0)
0
«5
i.
^
u
0
C
o
i
2
0
1
S 0
•v. w
! oo a
*•» u
s -
§•
DC -O
S "
— M
I I
~a » ^
_o S
1 0
f. ,C
84* •
C
i- .•«- 5
0 ° 8
0 X i
M U
0 <->
« » 2
g S —
a 01.
> JC 0
1- *1 i
ii M U
•*" CB
•0 I ^
a u ^
£ a «
•a 5 - cT
CM O0
O ** C
SO. C 0
S St.
— — u a
o.
1 £ i1 i
f «D 0 0
1 si M
s - * Is
« 33 * ~*
£ 00 •«• CD
5 s 5 ^ «^
? C C ° 0
•^ O o 01*
0 — ~- .CO
oa u u n •*.
" " • £
an 80
3 xx t i£
. * :
407
-------
TABLE 6. LONG-TERM CHEMICAL ANALYSES - 8HA AND LEAD
Plant
Are*
DSA-3
LAN-1
FSA-1
IFA-5
PFA-3
LAS-LH-1
BHAs in
treated soil, mg/kg
DO* (12.15)
BD (17.90)
BD (534.2)
1.1 (36.70)
0.91 (18.45)
BO (39.60)
BKA in TCLP
leachate, ug/l
43 (14)*
1,260 (1.409)
3,109 (4,083)
60 (169)
35 (55)
216 (689)
Migration
potential - BNA,
ug teachate
ug solid
- (•)
- (.62)
- (.15)
1.09 (-)
.077 (.53)
• (.31)
Lead in
treated
soil, mg/kg
980 (570)*
3,800 (3,010)
8,600 (10,200)
3,800 (2,700)
3,400 (4,900)
4,300 (3,850)
Lead in TCLP
leachate, ug/l
3 (7)
4 (5)
39 (400)
120 (50)
38 (11)
23 (51)
Migration potential
lead, (104)
ug leachate
ug solid
0.61 (1.69)*
0.21 (0.36)
0.91 (8.0)
.6.32 (5.40)
2.24 (0.61)
1.30 (3.13)
* Values In parentheses are from 28 day samples as reported in the Technology Evaluation Report.
+ Below detection limits. The detection limits were quite high due to interferences.
408
-------
SITE DEMONSTRATION OF THE TERRA VAC IN SITU VACUUM EXTRACTION TECHNOLOGY
by: Peter A. Michaels
Foster Wheeler Enviresponse, Inc.
Edison, NJ 08837
Mary K. Stihson
USEPA, RGB, RREL
Edison, NJ 08837
ABSTRACT
Terra Vac Inc's vacuum extraction system was the subject of a SITE
program demonstration test in Grovel and, Massachusetts. The site chosen was
contaminated with volatile organic compounds, mainly trichloroethylene, which
were used as degreasing solvents in an operating machine shop on the site.
The eight-week test run produced the following results:
o extraction of 1,300 Ib of VOCs • ..
o a steady decline in the VOC recovery rate with time
o a marked reduction in soil VOC concentration in the test area
o an indication that the process can remove VOCs from clay strata
The system operation proved to be very reliable. Upon achievement of a
steady operation, the only stoppages occurred in order to replace spent
activated carbon canisters with fresh canisters.
409
-------
INTRODUCTION
This SITE program demonstration test was planned to determine the
effectiveness of Terra Vac Inc's vacuum extraction technology in the
removal of volatile organic compounds from the vadose zone. The location
of the test was on the property of an operating machine shop. The property
is part of a Superfund site and is contaminated by degreasing solvents,
mainly trichloroethylene.
OBJECTIVES
The main objectives of this project were:
The quantification of the contaminants removed by the process.
The correlation of the recovery rate of contaminants with time.
o
o
o
The prediction of operating time required before obtaining site
remediation.
o The effectiveness of the process in removing contamination from
different soil strata.
APPROACH
The objectives of the project were achieved by following a
demonstration test plan which included a sampling and analytical plan. The
sampling and analytical plan contained a quality assurance project plan.
This QAPP assured that the data collected during the course of this project
would be of adequate quality to support the objectives.
The sampling and analytical program for the test was split up into a
pretest period, which has been called a pretreatment period; an active
period; midtreatment; and a posttreatment period.
The pretreatment period sampling program consisted of:
o soil boring samples taken with split spoons
o soil boring samples taken with Shelby tubes
o soil gas samples taken with punch bar probes
Soil borings taken by split spoon sampling were analyzed for volatile
organic compounds (VOCs) using headspace screening techniques, purge and
trap, GC/MS procedures, and the EPA-TCLP procedure. Additional properties
of the soil were determined by sampling using a Shelby tube, which was
pressed hydraulically into the soil by a drill rig to a total depth of 24
feet. These Shelby tube samples were analyzed to determine physical
characteristics of the subsurface stratigraphy such as bulk density,
particle density, porosity, pH, grain size, and moisture. These parameters
were used to define the basic soil characteristics.
410
-------
Shallow soil gas concentrations were collected during pre-, mid-, and
posttreatment activities. Four shallow vacuum monitoring wells and twelve
shallow punch bar tubes were used at sample locations. The punch bar
samples were collected from hollow stainless steel probes that had been
driven to a depth of 3 to 5 feet. Soil gas was drawn up the punch bar,.
probes with a low-volume personal pump and tygon tubing. Gas-tight .50-ml
syringes were used to collect the sample out of the tygon tubing.
The active treatment period consisted of collecting samples of:
o wellhead gas
o separator outlet gas
o primary carbon outlet gas
o secondary carbon outlet gas
o separator drain water
All samples with the exception of the separator drain water were r
analyzed on site. On-site gas analysis consisted of gas chromatography
with a flame ionization detector (FID) or an electron capture detector
(ECD). The FID was used generally to quantify the trichloroethylene (TCE)
and trans 1,2-dichloroethylene (DCE) values, while the ECD was used to
quantify the 1,1,1-trichloroethane (TRI) and the tetrachloroethylene (PCE)
values. The use of two detectors, FID and ECD, was necessitated by high
concentrations of TCE in the extracted well head gas. Owing to the high
TCE concentrations, most of the samples injected on the ECD had to ,be
diluted. Even with dilution factors of 333 to 1, the TCE concentration on
the ECD would exceed the linear range of the detector, thus necessitating
the use of two detectors.
The separator drain water-was analyzed for VOC content using SW846
8010. Moisture content of the separator inlet gas from the wells was
analyzed using EPA Modified Method 4. This method is good for the
two-phase flow regime that existed in the gas emanating from the wellhead.
Table 1 lists analytical methods used for this project. !
The posttreatment sampling essentially consisted of repeating
pretreatment sampling procedures at locations as close as possible to the
pretreatment sampling locations. . ... .
The activated carbon canisters were sampled, as close to the center of
the canister- as possible, and these samples were analyzed for VOC content
as a check on the material balance for the process. The method used was
P&CAM 12.7, which consisted of desorption of the carbon with CS2 and
subsequent gas chromatographic analysis.
PROCESS DESCRIPTION
The vacuum extraction process is a technique for the removal and.
411
-------
TABLE 1. ANALYTICAL METHODS
Parameter
Analytical method
Sample Source
Grain size
PH
Moisture (110°C)
Particle density
Oil and grease
EPA-TCLP
TOC
Headspace VOC
VOC
VOC
VOC
VOC
VOC
VOC
ASTM D422-63
SW846* 9040
ASTM D2216-80
ASTM D698-78
SW846* 9071
F.R. 11/7/86,
Vol. 51, No. 216,
SW846* 8240
SW846* 9060
SW846* 3810
GC/FID or ECD
GC/FID or ECD
SW846* 8010
SW846* 8010
Modified P&CAM 127
SW846* 8240
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil gas
Process gas
Separator liquid
Groundwater
Activated carbon
Soil borings
*Third Edition, November 1986.
412
-------
venting of volatile organic constituents (VOCs) from the vadose or
unsaturated zone of soils. Once a contaminated area is completely defined,
an extraction well or wells, depending upon the extent of contamination,
will be installed. A vacuum pump or blower induces air flow through the
soil, stripping and volatilizing the VOCs from the soil matrix into the air
stream. Liquid water is generally extracted along with the contamination.
The two-phase flow of contaminated air and water flows to a vapor liquid
separator where contaminated water is removed. The contaminated air stream
then flows through activated carbon canisters arranged in a parallel-series
fashion. Primary or main adsorbing canisters are followed by a secondary
or backup adsorber in order to insure that no contamination reaches the
atmosphere. Figure 1 illustrates the layout of wells and equipment.
EQUIPMENT LAYOUT AND SPECIFICATIONS
Specifications are given in Table 2 for the equipment used in the
initial phase of the demonstration. This equipment was later modified when
unforeseen circumstances required a shutdown of the system. The
vapor-liquid separator, activated carbon canisters, and vacuum pump skid
were inside the building, with the stack discharge outside the building.
The equipment was in an area of the machine shop where used cutting oils
and metal shavings had been stored.
Four extraction wells (EW1-EW4) and four monitoring wells (MW1 - MW4)
were drilled south of the shop. Each well was installed in two sections,
one section to just above the clay lens and one section to just below the
clay lens. The extraction wells were screened above the clay and below the
clay. As shown in Figure 2, the well section below the clay lens was
isolated from the section above by a bentonite portland cement grout seal.
Each section operated independently of the other. The wells were arranged
in a triangular configuration, with three wells on the base of the triangle
(EW2, EW3, EW4) and one well at the apex (EW1). The three wells on the
base were called barrier wells. Their purpose was to intercept
contamination, from underneath the building and to the side of the
demonstration area, before this contamination reached the main extraction
well (EW1). The area enclosed by the four extraction wells defined the
area to be cleaned.
INSTALLATION OF EQUIPMENT
Well drilling and equipment setup were begun on December 1, 1987. A
mobile drill rig was brought in, equipped with hollow-stem augers, split
spoons, and Shelby tubes. The locations of the extraction wells and
monitoring wells had been staked out previously based on contaminant
concentration profiles from a previously conducted remedial investigation
and from bar punch probe soil gas monitoring.
Each well drilled was sampled at 2-foot intervals with a split spoon
pounded into the subsurface by the drill rig in advance of the hollow stem
auger. The hollow stem auger would then clear out the soil down to the
depth of the split spoon, and the cycle would continue in that manner to a
depth of 24 feet. The drilling tailings were shoveled into 55-gallon drums
413
-------
MW1
1 Schematic diagram of equipment layout.
414
-------
2//PVCPIPE
— BENTONITE
3'
SAND
SCREENING
12.67'
« GROUT
" BENTONITE
SCREENING
24'
Fig. 2 Schematic diagram of an extraction well
415
-------
TABLE 2. EQUIPMENT LIST
Equipment
Number required
Description
Extraction wells
Monitoring wells
Vapor-liquid
separator
Activated carbon
canisters
Vacuum pump skid
Holding tank
Pump
4 (2 sections each)
4 (2 sections each)
1
Primary: 2 units in
parallel
Secondary: 1 unit
2" SCH 40 PVC 24' total depth
2" SCH 40 PVC 24' total depth
1000-gal capacity, steel
Canisters with 1200 Ib of carbon
in each canister - 304 SS
4" inlet and outlet nozzles
25 HP motor - positive displace-
ment lobe type blower - 3250 rpm
2000-gal capacity - steel
1 HP motor - centrifugal
for eventual disposal. After the holes were sampled, the wells were
installed using 2-inch PVC pipes screened at various depths depending upon
the characteristics of the soil in the particular hole. The deep well was
installed first, screened from the bottom to various depths. A layer of
sand followed by a layer of bentonite and finally a thick layer of grout
were required to seal off the section below the clay lens from the section
above the clay lens. The grout was allowed to set overnight before the
shallow well pipe was installed at the top of the grout. A layer of sand
bentonite and grout finished the installation.
VOC REMOVAL FROM THE VADOSE ZONE
The permeable vadose zone at the Groveland site is divided into two
layers by a horizontal clay lens, which is relatively impermeable. As
explained previously, each extraction well had a separate shallow and deep
section to enable VOCs to be extracted from that section of the vadose zone
above and below the clay lens. The quantification of VOCs removed was
achieved by measuring
o gas volumetric flow rate by rotameter and wellhead gas VOC
concentration by gas chromatography
416
-------
o the amount of VOCs adsorbed by the activated carbon canisters by
desorption into C$2 followed by gas chromatography.
VOC flow rates were measured and tabulated for each well section separ-
ately. The results of gas sampling by syringe and gas chromatographic
analysis indicate a total of 1,297 Ib of VOCs were extracted over a 56-day
period, 95% of which was trichloroethylene. A very good check on this
total was made by the activated carbon VOC analysis, the results of which
indicated a VOC recovery of 1353 Ib; virtually the same result was obtained
by two very different methods.
One view of the reduction in VOC concentrations in the vadose zone can
be seen from examining the three-dimensional shallow soil gas plots. Soil
gas was collected during pretreatment, midtreatment, and posttreatment from
punch bar probes and shallow vacuum monitoring wells. The collection
points were located on a coordinate system with extraction well 1 as the
origin (0,0). Each collection point has an x and y coordinate, and TCE
concentrations are plotted on a "Z" scale, which gives a three-dimensional
plot across the grid. Values of "Z" between data points and around the
grid are generated by the Kriging technique, which uses given data points
and a regional variable theory to generate values between and around sample
locations. Kriging is the name given to the least squares prediction of
spatial processes and is used in surface fitting, trend surface analysis,
and contouring of sparse spatial data.
A total of twelve shallow punch bar tubes were utilized along with the
four shallow vacuum monitoring wells. The punch bars were driven to a
depth of 3 to 5 feet, and as with the vacuum wells, soil gas was drawn up
the punch bar probes with a low-volume personal pump and tygon tubing.
50-ml gas-tight syringes were used to collect the sample out of the tygon
tubing. The gas samples were analyzed in the field trailer using gas
chromatographs with flame ionization detectors and electron capture
detectors.
The soil gas results show a considerable reduction in concentration
over the course of the 56-day demonstration period as can be seen from
Figures 3 and 4. This is to be expected since soil gas is the vapor halo
existing around the contamination and should be relatively easy to remove
by vacuum methods.
A more modest reduction can be seen in the results obtained for soil
VOC concentrations by 6C/MS purge-and-trap analytical techniques. Soil
concentrations include not only the vapor halo but also interstitial liquid
contamination that is either dissolved in the moisture in the soil or
existing as a two-phase liquid with the moisture.
Table 3 shows the reduction of the weighted average TCE levels in the
soil during the course of the 56-day demonstration test. The weighted
average TCE level was obtained by averaging soil concentrations obtained
every two feet by split spoon sampling methods over the entire 24-foot
depth of the wells. The.largest reduction in soil TCE concentration.
occurred in EW4, which had the highest initial level of contamination.
417
-------
MU2
a
a
0
Fig. 3 Pretreatment shallow soil-gas concentration
418
-------
MAP VIEW
MW2
MU3
MUJ4
Fig. 4 Posttreatment shallow soil-gas concentration
419
-------
TABLE 3. REDUCTION OF WEIGHTED AVERAGE TCE LEVELS IN SOIL
(TCE Cone, in mg/kg)
Extraction Well
Monitoring Well
1
2
3
4
Pretreatment
Posttreatment
1.10
14.75
227.31
0.87
0.34
8.98
84.50
1.05
% Reduction
1
2
3
4
33.98
3.38
6.89
96.10
29.31
2.36
6.30
4.19
13.74
30.18 . .
8.56
95.64
69.09
39.12
62.83
EW1, which was expected to achieve the greatest concentration reduction,
exhibited only a minor decrease over the course of the test. Undoubtedly
this was because of the greater-than-expected level of contamination that
existed in the area around MW3 that was drawn into the soil .around EW1.
The decrease in the TCE level around MW3 tends to bear this out.
EFFECTIVENESS OF THE TECHNOLOGY IN VARIOUS SOIL TYPES
The soil strata at the Groveland site can be characterized generally as
consisting of the following types in order of increasing depth:
o medium to very fine silty sands
o stiff and wet clays
o sand and gravel
Soil porosity, which is the percentage of total soil volume occupied by
pores, was relatively the same for both the clays and the sands.
Typically, porosity over the 24-foot depth of the wells would range between
40% and 50%. Permeabilities, or more accurately hydraulic conductivities,
ranged from 10"* cm/sec for the sands to 10"° cm/sec for the clays,
with corresponding grain sizes equal to 10"1 mm to 10"3 mm.
Pretest soil boring analyses indicated in general that most of the
contamination was in the strata above the clay lens with a considerable
quantity perched on top of the clay lens. This was the case for EW4, which
420
-------
showed an excellent reduction of TCE concentration in the medium to fine
sandy soils existing above the clay layer, with no TCE detected in the clay
in either the pretest or posttest borings (see Table 4). One of the wells,
however, was an exception. This was MW3, which contained the highest
contamination levels of any of the wells, and was exceptional in that most
of the contamination was in a wet clay stratum. The levels of
contamination were in the 200-1600 ppm range before the test.
After the test, analyses of the soil boring adjacent to MW3 showed
levels in the range of ND-60 ppm in the same clay stratum. The data, as
shown in Table 5, suggest that the technology can desorb or otherwise
mobilize VOCs out of certain clays.
From the results of this demonstration it appears that the permeability
of a soil need not be a consideration in applying the vacuum extraction
technology. This may be explained by the fact that the porosities were
approximately the same for all soil strata, so that the total flow area for
stripping air was the same in all soil strata. It will take a long time
for a liquid contaminant to percolate through clay with its small pore size
and consequent low permeability. However, the much smaller air molecules
have a lower resistance in passing through the same pores. This may
explain why contamination was generally not present in the clay strata, but
when it was, it was not difficult to remove. Further testing should be
done in order to confirm this finding.
CORRELATION OF DECLINING VOC RECOVERY RATES
The vacuum extraction of volatile organic constituents from the soil
may be viewed as an unsteady state process taking place in a nonhomogeneous
environment acted upon by the combined convective forces of induced
stripping air and by the diffusion of volatiles from a dissolved or sorbed
state. As such it is a very complicated process, even though the equipment
required to operate the process is very simple.
Unsteady state diffusion processes in general correlate well by
plotting the logarithm of the rate of diffusion versus time. Although the
representation of the vacuum extraction process presented here might be
somewhat simplistic, the correlation obtained by plotting the logarithm of
the concentration of contaminant in the wellhead gas versus time and
Obtaining a least squares best fit line was reasonably good. This type of
plot, shown in Figure 5, represents the data very well and is more valid
than both a linear graph or one plotting concentration versus log time, in
which a best fit curve would actually predict gas concentrations of zero or
less.
Looking at the plots for EW1, shallow and deep, equations are given for
the least squares best fit line for the data points. If the vacuum
extraction process is run long enough so that the detection limit for TCE
on the ECD, which is 1 ppbv, is reached, the length of time required to
reach that concentration would be approximately 250 days on the shallow
well and approximately 300 days on the deep well.
421
-------
TABLE 4. EXTRACTION WELL 4:
TCE REDUCTION IN SOIL STRATA
Depth
ft
0-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
16-18
18-20
20-22
22-24
Description Permeability TCE Cone, com
of strata
Med. sand w/gravel
Lt. brown fine sand
Med. stiff It. brown fine sand
Soft dk. brown fine sand
Med. stiff brown sand
V. stiff It. brown med. sand
V. stiff brown fine sand w/silt
M. stiff grn-brn clay w/silt
Soft wet clay
Soft wet clay
V. stiff brn med-coarse sand
V. stiff brn med-coarse w/gravel
cm/sec
1°1
10-J
10-5
10";
lo-J
lo-J
lO-J
IQ-f
10"°
10"!
lo-J
lO'3
pre
2.94
29.90
260.0
303.0
351.0
195.0
3.14
ND
ND
ND
ND
6.71
post
ND
ND
39
9
ND
ND
2.3
ND
ND
ND
ND
ND
TABLE 5. MONITORING WELL 3:
TCE REDUCTION IN SOIL STRATA
Depth
ft
0-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
16-18
18-20
20-22
22-24
Description
of strata
M. stiff brn. fine sand
M. stiff grey fine sand
Soft It. brn. fine sand
Lt. brn. fine sand
Stiff V. fine brn. silty sand
Silty sand
Soft brown silt
Wet green -brown silty clay
Wet green -brown silty clay
Wet green-brown silty clay
Silt, gravel, and rock frag.
M. stiff It. brn. med. sand
Permeability
cm/sec
1Q-5
10"?
10"!
10-J
lo-J
10-J
10-J
10 Q
1° "I
10"?
10-J
ID'4
TCE Cone.
pre
10.30
8.33
80.0
160.0
ND
NR
316.0
195.0
218.0
1570.0
106.0
64.1
ppm
post
ND
800
84
ND
63
2.3
ND
ND
62
2.4
ND
ND
422
-------
CROVELAND/TERRA-VAC DEMONSTRATION
_ 1000
£
CL
O
cr
LU
CJ
O
C>
Ld
O
°'1
0.0.1
EXTRACTION WELL #1
SHALLOW
CURVE COEFFICIENT- R2=0 62.
0 20 40 60 80 100
DAY OF ACTIVE TREATMENT
Fig. 5 Wellhead TCE concentration vs. time
423
-------
PREDICTION OF TIME REQUIRED FOR SITE REMEDIATION
The soil concentration that would be calculated from the wellhead gas
concentration using Henry's Law is included in the last column of Table 6.
Calculations for the predicted soil concentrations were made assuming a
bulk density of the soil of 1761 kg/nr3, a total porosity of 50%, and a
moisture content of 20%. The calculated air filled porosity of the soil is
approximately 15%. Henry's constant was taken to be 0.492 KPa/m3-gmol at
40°F.
Given the nonhomogeneous nature of the subsurface contamination and
interactions of TCE with organic matter in the soil, it was not possible to
obtain a good correlation between VOC concentrations in wellhead gas and
soil in order to predict site remediation times. Henry's Law constants
were used to calculate soil concentrations from wellhead gas concentrations
and the calculated values obtained, correcting for air filled porosity,
were lower than actual soil concentrations by at least an order of
magnitude (see Table 6).
TABLE 6. COMPARISON OF WELLHEAD GAS VOC
CONCENTRATION AND SOIL VOC CONCENTRATION
Extraction Well
TCE concentration
in wellhead gas
ppmv
TCE concentration
in soil ppmw
Predicted by
Henry's Law
ppmw
IS
ID
2S
2D
3S
3D
4S
9.7
5.6
16.4
14.4
125.0
58.7
1095.6
54.5
7.2
ND
20.4
20.9
18.0
9.1
0.11
0.07
0.20
0.17
1.5.3
0.74
12.49
Before one can attempt to make a rough estimation of the remediation
time, a target value for the particular contaminant in the remediated soil
must be calculated. This target concentration is calculated by using two
mathematical models, the Vertical and Horizontal Spread ModelW and the
Organic Leachate Model^a'. The mathematical models allow the use of a
regulatory standard for drinking water in order to arrive at a target soil
concentration.
The VHS model is expressed as the following equation:
Cy = C0 erf (Z/(2(azY)°-5)) erf (X/(atY)°-5)
424
-------
where:
Cy = concentration of VOC at compliance point (mg/1)
CQ = concentration of VOC in leachate (mg/1)
erf = error function (dimensionless)
Z = penetration depth of leachate into the aquifer
Y = distance from site to compliance point (m)
X = length of site measured perpendicular to the direction of
ground water flow (m)
at = lateral transverse dispersivity (m)
az = vertical dispersivity (m)
A simplified version of the VHS model is most often used, which reduces
the above equation to:
cy = cocf
where:
Cf = erf (Z/(2(azY)°-5)) erf (X/(atY)°-5), which is
reduced to a conversion factor corresponding to the amount of
contaminated soil
The Organic Leachate Model (OLM) is written as:
C0 = 0.00211 cs°'678S°-373
where:
concentration of VOC in leachate (mg/1)
concentration of VOC in soil (mg/1)
solubility of VOC in water (mg/1)
This
The regulatory standard for TCE in drinking water is 3.2 ppb.
regulatory limit is used in the VHS model as the compliance point
concentration in order to solve for a value of the leachate concentration.
This value of leachate concentration is then used in the OLM model to solve
for the target soil concentration.
Once the target soil concentration is determined, a rough estimation of
the remediation time can be made by taking the ratio of soil concentration
to wellhead gas concentration and extrapolating in order to arrive at a
wellhead gas concentration at the target soil concentration. The
calculated target soil concentration for this site is 500 ppbw. This
corresponds to an approximate wellhead gas concentration of 89 ppb for
EW1S. The equation correlating wellhead gas concentration with time (see
Figure 5) is then solved to give 150 days running time.
After 150 days the vacuum extraction system can be run intermittently
to see if significant increases in gas concentrations occur upon
restarting, after at least a two day stoppage. If there are no appreciable
increases in gas concentration, the soil has reached its residual
425
-------
equilibrium contaminant concentration and the system may be stopped and
soil borings taken and analyzed.
(a) EPA Draft Guidelines for Petitioning Waste Generated by the Petroleum
Refinery Industry, June 12, 1987.
ACKNOWLEDGMENTS
The authors wish to thank Mr. James S. Ciriello, formerly of the U.S.
Environmental Protection Agency, Region I, Boston, Massachusetts at the
time of the project, for his efforts during the course of this project. A
special note of gratitude is to be given to Mr. Thomas Quinlan of the
Valley Manufactured Products Company, Inc. for his special support and
cooperation that helped make this project a successful one.
426
-------
THE OFFICE OF RESEARCH & DEVELOPMENT
WRITE PROGRAM
Ivars J. Licis
Research Program Manager
Risk Reduction Engineering Laboratory
and
M. Lynn Apel
Chief, Process Engineering Section
Risk Reduction Engineering Laboratory,
U.S. Environmental Protection Agency
Cincinnati, OH 45268
Passage of the 1984 Hazardous and Solid Waste Amendments
(HSWA) to the Resource Conservation and Recovery Act (RCRA) of
1976 marked a strong change in the policies of the United States
concerning the generation of hazardous and nonhazardous wastes.
In addition to authorizing very stringent treatment and disposal
regulations, the Amendments also indicated, as the Nation's top
waste management priority a redirection toward "waste
minimization" as a preferential strategy for encouraging
improvement in environmental quality.
To carry out the intent of the Amendments to reduce the
generation of waste, the U.S. Environmental Protection Agency
(EPA) has developed a comprehensive pollution prevention program
addressing the release and transport of hazardous, toxic, and
nonhazardous materials through air, water, and solid media. The
EPA pollution prevention program includes information gathering,
research and development, demonstration, support of state and
local government waste minimization and pollution prevention
programs, training and education, technology transfer activities,
waste minimization assessments, and extensive communication with
universities, state and local governments, and the general
public.
This paper describes one of the major pollution prevention
research programs being undertaken by EPA, the Waste Reduction
Innovative Technology Evaluation [WRITE] Program. The WRITE
Program has been designed to identify, evaluate, and demonstrate
new ideas and technologies that lead to waste reduction. The
WRITE Program involves the cooperative efforts of the EPA, state
and local governments, private industry, universities, and other
organizations to encourage the research and development of
427
-------
effective techniques and technologies for multimedia pollution
prevention.
BACKGROUND
During the last decade, waste management and disposal
regulations have changed drastically. During the 1970's, the
seriousness of the hazardous waste problem prompted the Congress' *
to enact the Resource Conservation and Recovery Act in 1976. By
the early 1980's, it had become apparent that even well-regulated
land disposal could cause significant environmental damage. The
Amendments to RCRA resulted in the banning of the land disposal
of over 400 chemicals and hazardous wastes. During the 1980's,
significant effort was devoted to research, development, and
commercialization of hazardous waste treatment technologies.
Although some efforts were successful in substantially mitigating
the problem of existing waste, it became apparent that treatment
alone would not solve all of the problems associated with the
increasing amounts of waste being generated.
Congress recognized this fact while drafting the national
policy on hazardous wastes and directed EPA to report on the
feasibility and desirability of developing mandatory requirements
to compel the adoption of pollution prevention techniques. The
1984 HSWAs to RCRA define the national policy on hazardous waste
management as follows:
"The Congress hereby declares it to be the
national policy of the United States that,
wherever feasible, the generation of
hazardous waste is to be reduced or
eliminated as expeditiously as possible.
Waste that is nevertheless generated should
be treated, stored, or disposed of so as to
minimize the present and future threat to
human health and the environment." [1]
In 1986, EPA responded to the Congressional request with a
Report to Congress on the Minimization of Hazardous Waste [2].
In this report, the Agency defined waste minimization as:
"The reduction, to the extent feasible, of
hazardous waste that is generated or
subsequently treated, stored or disposed of.
It includes any source reduction or
recycling activity undertaken by a generator
that results in either (1) the reduction of
total volume or quantity of hazardous waste
or (2) the reduction of toxicity of
hazardous waste, or both, so long as the
reduction is consistent with the goal of
428
-------
minimization of present and future threats
to human health and environment."
In January 1989, the Agency issued a Pollution Prevention
Policy Statement [3]. This policy encourages organizations,
facilities and individuals to fully utilize source reduction and
recycling practices and procedures to reduce risk to public
health, safety and the environment. Source reduction is
emphasized as the preferred approach since it eliminates or
significantly reduces the quantity of pollutants generated,
thereby significantly reducing the potential risk to health and
the environment. As a second choice, recycling reduces the
amount of waste to be treated or sent for disposal.
Additionally, many of the waste reduction efforts have proven to
be both more economical, in the long run, as well as beneficial
to the environment producing a win-win situation. On this basis,
source reduction and recycling are the primary two elements in
the Waste Management Hierarchy preceding treatment and disposal.
In this context, source reduction has been defined as the
reduction or elimination of waste at the source, usually within a
process. Source reduction measures include process
modifications, feedstock substitutions, improvements in feedstock
purity, housekeeping and management procedural changes, increases
in the efficiency of equipment, and recycling within a process.
Likewise, recycling has been defined as the use or reuse of a
waste material as an effective substitute for a commercial
product or as an ingredient or feedstock in an industrial
process. It includes the reclamation of useful constituent
fractions within a waste material or the removal of contaminants
from a waste to allow it to be reused [4].
EPA'S POLLUTION PREVENTION PROGRAM
In its Report to Congress, the Agency explored various
technical, economic, and policy issues relevant to the reduction
and recycling of hazardous and nonhazardous wastes, and concluded
that it would be counterproductive for EPA to establish a
mandatory program for waste minimization at this time. The
Report to Congress stressed that the most constructive role
government could assume is to promote voluntary waste
minimization by providing information, technology transfer, and
assistance to waste generators. The Agency proposed a waste
minimization program to encourage industry to accelerate efforts
to reduce the generation of wastes through implementation of
process changes and/or the incorporation of recycling methods.
In its efforts to pursue the objectives set forth by
Congress, EPA has established a national program to effect waste
minimization, or what has since become included in the term
"pollution prevention". Pollution prevention is a term that has
been used more frequently within the last year by several
organizations including EPA to describe techniques, practices, or
procedures implemented by the private and public sectors to
429
-------
prevent the generation of pollutants. As it is used today,
"pollution prevention" has replaced the term "waste minimization"
which was generally applied to reducing the generation of
hazardous wastes. Through elimination of this term, that may be
perceived as closely tied to RCRA, EPA is emphasizing that it's
pollution prevention policy has applicability beyond the RCRA
hazardous waste context.
EPA's pollution prevention program is based on several
elements including: (1) promoting waste reduction by transferring
technical information to firms and state and local government
technical assistance officers; (2) encouraging the adoption of
waste reduction by identifying, evaluating, and demonstrating
appropriate technologies and by promoting the use of waste audits
or assessments; (3) keeping Congress advised on national progress
of waste reduction techniques and practices; (4) fostering
development of state and local government pollution prevention
programs; (5) developing outreach and communication programs with
the goal of raising awareness of the benefits of waste reduction
practices within government, industry, and the public.
To accomplish some of these objectives and encourage the
identification, development, and demonstration of processes and
technologies that result in less waste being generated, EPA's
Office of Research and Development (ORD) has initiated several
multimedia pollution prevention research programs. These
programs are designed to be the cornerstone of the Agency's
pollution prevention research and demonstration effort and will
provide much needed data on new technologies to the generating
sectors. Addressed under these programs are hazardous,
nonhazardous, industrial and municipal wastes. Design and
implementation of these programs is being undertaken by the
Pollution Prevention Research Branch (formerly the Waste
Minimization Branch) of the Risk Reduction Engineering Laboratory
of ORD in Cincinnati, Ohio.
The Agency has also established a Pollution Prevention
Office (PPO) within the Office of Policy, Planning and Evaluation
in Washington, D.C. The Pollution Prevention Office is
responsible for award and oversight of the Source Reduction and
Recycling Technical Assistance grants, establishing a National
Pollution Prevention Information Clearinghouse, and publishing a
quarterly newsletter on pollution prevention topics. PPO
together with ORD and the Agency's media-specific offices (i.e.,
Office of Solid Waste, Office of Toxic Substances, etc.) will
develop and implement EPA's multi-media pollution prevention
program.
THE WRITE PROGRAM
Reducing the generation of industrial and other wastes can
be achieved in many ways. Process chemistry can be changed;
potential waste streams can be recycled within a manufacturing
430
-------
process or back into the process; process technology and/or
equipment can be modified to produce products more efficiently,
resulting in less waste; plant operations, i.e., "housekeeping"
methods, can be changed or controlled to produce fewer and
smaller waste streams or less waste in general; changes in raw
materials (feedstocks) can lead to fewer waste streams or
less-hazardous waste streams; finally, changes in the end
products from manufacturing operations can, in some instances, be
made so as to affect the types and quantities of wastes emitted.
The EPA WRITE Program outlined below is designed to identify,
evaluate and enhance the application of technologies and
practices containing these attributes.
The WRITE Program is a research program designed to
identify, evaluate, and/or demonstrate the use of innovative
engineering and scientific technologies to reduce the volume
and/or toxicity of wastes produced from the manufacture,
processing, and use of materials. The WRITE Program is broad in
technical scope and addresses the reduction of pollutants across
all environmental media: air, land, surface water, and
groundwater. Attention is directed toward methodologies with the
potential for reducing the quantity and/or toxicity of waste
produced at the source of generation, or to achieve practicable
onsite reuse or recycling of waste materials. Strong
consideration is given to the applicability of a technique on an
industry-wide basis and across industries. Industries of primary
interest under the WRITE Program include chemical, fabricated
metal, electronic, printing and publishing, lumber, petroleum,
transportation, food, and textile [5].
The objectives of the WRITE Program are:
0 To establish reliable performance and cost
information on pollution prevention techniques by
conducting evaluations or demonstrations of the more
promising innovative technologies.
0 To accomplish an early introduction of waste
reduction techniques into broad commercial practice.
0 To encourage active participation of small- and
medium-sized companies in evaluating and adopting
pollution prevention concepts by providing support to
these companies through State and local government
agencies.
To encourage the transfer of knowledge and technology
concerning pollution prevention practices between
large, medium-sized, and small industries.
0 To provide solutions to important chemical-,
wastestream-, and industry-specific pollution
prevention research needs.
431
-------
For the purposes of this program, innovative waste
minimization technologies are defined as those technologies that
represent an alternative approach to conventional waste
management methods (i.e., alternatives to incineration, land
disposal, treatment, etc.). This approach involves substantially
reducing the volume or toxicity of waste generated at the source
and/or recycling or reusing the waste materials. In some cases,
the technologies may incorporate novel modifications to a
component of an existing process or operation, or the adaptation
of an existing technology to a new process.
To date the WRITE Program encompasses approximately 12
research and demonstration projects totalling over $4.0 million.
Under these projects, approximately 30 waste reduction
technologies will be evaluated throughout the next three years,
and several long-term waste- and industry-specific research
studies will be undertaken. These activities are conducted under
three subprograms of the WRITE Program; the WRITE Pilot Program
with State and Local Governments, the WRITE Program With
Industry, and the WRITE Research Program [see WRITE Program
Chart).
WRITE PILOT PROGRAM WITH STATE AND LOCAL GOVERNMENTS
At the moment, the "WRITE Pilot Program with state and
Local Governments" is the largest subprogram under WRITE and
addresses immediate information transfer needs between government
and industry. Through the joint efforts of EPA and various state
and local governments, technical and economic evaluations of
source reduction and recycling technologies are being conducted
of manufacturing and processing operations across approximately
twenty industries. This joint approach was chosen because State
and local government officials are often more familiar with local
industrial practices and regional manufacturing and economic
interests that can affect the potential success and widespread
applicability of proposed pollution prevention technologies.
States currently participating in this program include
California, Connecticut, Illinois, Minnesota, New Jersey, and
Washington.
Under this program, $100,000 per year is provided by EPA to
each participating state/local government. The state/local
government also contributes additional matching funds ranging
from 33% to 50% of the cost of the research. An average of five
waste reduction technologies will be evaluated by each
state/local government during the 3-year period of this pilot
program. Waste reduction technologies evaluated under this
program are based on several selection criteria. These include:
(1) type of waste minimization technology, (2) status of
development, (3) unique nature of the technology, (4)
application, (5) source reduction performance capability, (6)
extent of process modification, (7) cost effectiveness of the
technology, (8) process safety and health considerations, (9)
432
-------
cost to the EPA and state/local government, and (10)
legal/contractual issues. A worth assessment model is then used
as one of the decision tools to evaluate and rank potential waste
minimization technologies.
The technical and economic evaluations conducted for each
technology include an in-depth study of the process, a literature
review of comparable processes, material and energy balance
computations, a field demonstration, and determination of cost
estimation parameters including itemization of capital and
operational costs, calculation of the payback period and return
on investment. A summary of the types of information collected
during a technical and economic evaluation of a waste
minimization technology under this program is shown in Table 1.
The example concerns the modification of a cold solvent cleaning
process. In the cold cleaning of ball bearings with solvents,
using a two-step countercurrent cleaning sequence can increase
the cleaning efficiency. It can also substantially reduce the
solvent requirement and, hence, the waste generation. This
process does not involve substantial equipment modification.
Material balance calculations indicate a waste reduction of 50
percent and a 33 percent reduction in fresh solvent requirements
[6].
WRITE PROGRAM WITH INDUSTRY
The "WRITE Program With Industry" focuses on evaluations of
waste reduction technologies currently in use or under
development by large industries. One of the objectives of this
program is to encourage the transfer of knowledge and technology
concerning pollution prevention practices between large,
mid-size, and small industries. Under the "WRITE Program With
Industry", evaluations of waste reduction technologies are
performed directly with industrial firms or through industrial
trade associations and/or technical societies.
WRITE RESEARCH PROGRAM
In addition to evaluation programs, the WRITE Program has a
research subprogram. This subprogram focuses on pollution
prevention research needs, i.e., the generation of data to allow
the future demonstration of emerging new pollution prevention
techniques. Projects under this component of the WRITE Program
address various technical obstacles to waste reduction and to
chemical-, wastestream-, and industry- specific waste
minimization issues. These research efforts are conducted with
industrial firms, universities, other government agencies,
technical societies, and industrial trade organizations.
While emphasis to date has been on industrial processes
the area of focus is being enlarged to include techniques for
pollution prevention for any source and especially those that are
identified with posing significant risk in terms of high toxicity
433
-------
or high threat or those producing large amounts of pollution.
CONCLUSIONS
The WRITE Program will continue to expand and address
pollution prevention research needs within government and
industry. A number of challenging areas have been identified and
are receiving attention. An effort is being launched to provide
a way to prioritize needed work. This is especially significant
with the change of focus from the industrial and hazardous waste
area to all sources of pollution. Another area deals with the
means by which technologies can be best evaluated in terms waste
reduction and cost when the differences between the old and the
new technology involve different pollutants at differing
toxicities, concentrations and exposure pathways.
Additional information concerning this and other EPA pollution
prevention research programs can be obtained from the Pollution
Prevention Research Branch of the Risk Reduction Engineering
Laboratory, U.S. EPA, Cincinnati, Ohio.
REFERENCES
1.
4.
6.
U.S. Congress. Hazardous and Solid Waste Amendments.
Washington, D.C. 1984.
U.S. Environmental Protection Agency. Report to
Congress, Minimization of Hazardous Wastes. U.S. EPA,
Office of Solid Waste, Washington, DC., EPA/530-SW-
86-033. 1986.
U.S. Environmental Protection Agency. Pollution
Prevention Policy Statement. Federal Register, Vol,. 54,
No. 16, January 26, 1989, p. 3845.
Freeman, H.M., and J. Lounsbury. The U.S. EPA Hazardous
Waste Minimization Program. Proceedings of American
Institute of Chemical Engineers Annual Meeting,
Washington, D.C. 1988.
Apel, M.L., H.M. Freeman, M.F. Szabo, S.H. Ambekar.
Guidance Document for the WRITE Pilot Program With State
and Local Governments. U.S. EPA, Risk Reduction
Engineering Laboratory, Cincinnati, OH. 1988.
Jacobs Engineering. Waste Minimization Audit Report:
Case Studies of Minimization of Solvent Waste from
Parts Cleaning and From Electronic Capacitor
Manufacturing Operations. U.S. EPA, Risk Reduction
Engineering Laboratory, Cincinnati, OH,
EPA/600/52-87/057. 1987.
434
-------
TABLE 1. SUMMARY OF ENGINEERING EVALUATION FOR COLD SOLVENT
CLEANING PROCESS MODIFICATION [4]
Type of
Application
Stage of
Development
Unique Nature of
Technology
Applications of
cold solvent
Performance
Need for
Modification
Process modification
Demonstration
First-of-a-kind demonstration
Reduces hazardous waste generation in the
cleaning operations of the parts cleaning
industry, which is a medium/small-scale
operation.
Achieves 50% waste reduction by reducing the
fresh solvent requirement by 33%
Requires essentially only minor equipment
modification.
Cost Effectiveness Added capital costs = $ 600
of Technology Net operating savings = $ 380 per year
Payback period = 1.6 years
Note: Net operating savings include savings
resulting from reduced waste disposal,
reduced solvent requirement, and operation
and maintenance expense.
Safety & Health
Properly designed system is considered safe.
Metal cleaning systems are routinely used in
industry without any safety or health
problems.
435
-------
CC
O
O
cc
CL
111
5E
UJ
UJ
DC
cn
CL
CQ
CO
CO
CO
o<^
ZDOO
CQ
UJZ>
CO
O
rr
"3 «3
< ^
LU<
CODC
cc
«
i=Ol-
w
DCt±
CD CO
OZ>
DCQ
[r
DC
CM
; Q ^ LJLJ uj
• ^ {/J O ft"
3 pc N o S
: LU
ZZQi
CO
O)
LLJ
DC
CD
CO
'a
Q.
into
DC
CD
S:
cc
UJ
,
Q
cc
2 Q
> y
o!
Q O
436
-------
Solidification/Stabilization as a
Best Demonstrated Available Technology for
Resource Conservation and Recovery Act Wastes
by
R. Mark Bricka and M. John Cullinane, Jr.
Environmental Laboratory
Department of the Army
Waterways Experiment Station, Corps of Engineers
P.O. Box 631
Vicksburg, Mississippi 39180-0631
ABSTRACT
Early in 1987, the U.S. Army Engineers Waterways Experiment Station
(WES) was tasked by the U.S. Environmental Protection Agency (USEPA) to
investigate solidification/stabilization (S/S) as a Best Demonstrated Avail-
able Technology (BOAT). Under this program EPA supplied WES with a number
of Resource Conservation and Recovery Act (RCRA) listed hazardous wastes for
evaluation.
Nine listed wastes were processed using three generic S/S processes:
Portland cement, kiln dust, and lime/flyash. The physical and leaching
characteristics of the treated materials were evaluated. Physical charac-
teristics were evaluated using the unconfined compressive-strength (UCS)
test. The leaching characteristics of the wastes were evaluated using the
Toxicity Characteristic Leaching Procedure (TCLP). The TCLP extracts were
analyzed for a total of 22 metals. No attempt was made to analyze for semi-
volatile and volatile compounds in the TCLP extracts, due to the absence of
these compounds in the raw wastes. The effectiveness of the S/S processes
in immobilizing the metals was evaluated by comparing the TCLP extract
result of the raw (unsolidified/unstabilized) waste to the TCL |