&EPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/540/R-95/532
September 1995
Bioremediation of
Hazardous Wastes
Research, Development,
and Field Evaluations
-------
EPA/540/R-95/532
September 1995
BIOREMEDIATION OF HAZARDOUS WASTES:
Research, Development, and Field Evaluations
Biosystems Technology Development Program
Office of Research and Development
U.S. Environmental Protection Agency
U.S. Environmental Protection Agency
Ada, OK; Athens, GA; Cincinnati, OH; Gulf Breeze, FL;
and Research Triangle Park, NC
-------
Disclaimer
The information in this document has been funded wholly or in part by the U.S. Environmental
Protection Agency (EPA) and has been reviewed in accordance with EPAs peer and administrative
review policies and approved for presentation and publication. Mention of trade names or commer-
cial products does not constitute endorsement or recommendation for use.
-------
Contents
Page
Executive Summary ix
Introduction xi
Section 1: Bioremediation Field Initiative 1
Intrinsic Bioremediation of Trichloroethylene at the St. Joseph Aquifer/Lake Michigan Interface:
A Role for Iron and Sulfate Reduction
Jack Lendvay, Mike McCormick, Peter Adriaens, University of Michigan, Ann Arbor, Ml 3
Modeling Intrinsic Remediation as Ground-Water Discharges to a Lake:
The Trichloroethylene Plume at St. Joseph, Michigan
Sean M. Dean, Nikolaos D. Katopodes, University of Michigan, Ann Arbor, Ml 6
The U.S. Environmental Protection Agency's Development of Bioventing
Gregory Sayles, U.S. EPA, Cincinnati, OH 8
Section 2: Field Research 11
A Review of Intrinsic Bioremediation of Trichloroethylene in Ground Water at Picatinny Arsenal,
New Jersey, and St. Joseph, Michigan
John T Wilson, Don Kampbell, James Weaver, Barbara Wilson, U.S. EPA, Ada, OK;
Tom Imbrigiotta, Ted Ehlke, U.S. Geological Survey, Trenton, NJ 13
Intrinsic Bioremediation of a Gasoline Plume: Comparison of Field and Laboratory Results
Morton A. Barlaz, Melody J. Hunt, Sreenivas Kota, Robert C. Borden, North Carolina State
University, Raleigh, NC 17
Toxicity Effects on Methanogenic Degradation of Phenol in Ground Water
Barbara A. Bekins, E. Michael Godsy, Ean Warren, U.S. Geological Survey, Menlo Park, CA 20
A Multiphase, Multicomponent Numerical Model of Bioventing With Nonequilibrium Mass Exchange
Linda M. Abriola, John R. Lang, Klaus M. Rathfelder, University of Michigan, Ann Arbor, Ml 22
Aromatic Hydrocarbon Biotransformation Under Mixed Oxygen/Nitrate Electron Acceptor Conditions
Liza P. Wilson, Peter C. D'Adamo, Edward J. Bouwer, The Johns Hopkins University,
Baltimore, MD 24
-------
Contents (continued)
Page
Nutrient Transport in a Sandy Beach
Brian A. Wrenn, Makram T. Suidan, B. Loye Eberhardt, Gregory J. Wilson, University of
Cincinnati, Cincinnati, OH; Kevin L. Strohmeier, Environmental Technologies and Solutions,
Inc., Covington, KY; Albert D. Venosa, U.S. EPA, Cincinnati, OH 26
Bioremediation of Crude Oil Intentionally Released on the Shoreline of Fowler Beach, Delaware
Albert D. Venosa, John R. Haines, U.S. EPA, Cincinnati, OH; Makram T. Suidan, Brian A.
Wrenn, B. Loye Eberhardt, Miryam Kadkhodayan, Edith Holder, University of Cincinnati,
Cincinnati, OH; Kevin L. Strohmeier, Environmental Technologies and Solutions, Inc.,
Covington, KY; Dennis King, Kingstat Consulting, Fairfield, OH; Bennet Anderson, Delaware
Department of Natural Resources and Environmental Control, Dover, DE 29
Dynamics of Oil Degradation in Coastal Environments: Effect of Bioremediation Products and Some
Environmental Parameters
Marirosa Molina, Rochelle Araujo, U.S. EPA, Athens, GA; Jennifer R. Bond, DYNCORP,
Athens, GA 32
Progress Toward Verification of Intrinsic Cobioremediation of Chlorinated Aliphatics
Mark Henry, Michigan Department of Natural Resources, Oscoda, Ml 35
Development and Capabilities of the National Center for Integrated Bioremediation Research and
Development (NCIBRD)
Mark Henry, Michigan Department of Natural Resources, Oscoda, Ml 36
Phytoremediation of Petroleum-Contaminated Soil: Laboratory, Greenhouse, and Field Studies
M. Katherine Banks, A. Paul Schwab, Kansas State University, Manhattan, KS 37
Intrinsic Bioremediation of Fuel Contamination in Ground Water at a Field Site
Don H. Campbell, Robert S. Kerr Environmental Research Laboratory, Ada, OK;
T.H. Wiedemeier, Parsons Engineering Science, Inc., Denver, CO; J.E. Hansen, U.S. Air Force
Center for Environmental Excellence, Brooks Air Force Base, TX 38
Section 3: Performance Evaluation 39
Detoxification of Model Compounds and Complex Waste Mixtures Using Indigenous and Enriched
Microbial Cultures
K.C. Donnelly, Jeannine L. Capizzi, Ling-Yu He, Henry J. Huebner, Texas A&M University,
College Station, TX 41
Assessing the Genotoxicity of Complex Waste Mixtures
Larry D. Claxton, Virginia S. Houk, Sarah H. Warren, Thomas J. Hughes, and Susan E.
George, U.S. EPA, Research Triangle Park, NC 43
iv
-------
Contents (continued)
Page
Section 4: Pilot-Scale Research 45
In Situ Bioremediation of Trichloroethylene With Burkholderia Cepacia PR1: Analysis of Parameters
for Establishing a Treatment Zone
Richard A. Snyder, University of West Florida, Pensacola, FL; M. James Hendry, John R.
Lawrence, Environment Canada, Saskatoon, Canada 47
Characterization of Trichloroethylene-Degrading Bacteria From an Aerobic Biofilter
Alec Breen, Todd Ward, Ginger Reinemeyer, John Loper, Rakesh Govind, University of
Cincinnati, Cincinnati, OH; John Haines, U.S. EPA, Cincinnati, OH 51
Anaerobic/Aerobic Degradation of Aliphatic Chlorinated Hydrocarbons in an Encapsulated
Biomass Biofilter
Rakesh Govind, P.S.R.V. Prasad, University of Cincinnati, Cincinnati, OH; Dolloff F. Bishop,
U.S. EPA, Cincinnati, OH 54
Operation and Optimization of Granular Air Biofilters
Francis Lee Smith, George A. Serial, Makram T. Suidan, Amit Pandit, Pratim Biswas, University
of Cincinnati, Cincinnati, OH; Richard C. Brenner, U.S. EPA, Cincinnati, OH 57
Abiotic Fate Mechanisms in Soil Slurry Bioreators
John A. Glaser, Paul T. McCauley, U.S. EPA, Cincinnati, OH; Majid A. Dosani, Jennifer S. Platt,
E. Radha Krishnan, IT Corporation, Cincinnati, OH 61
Design and Testing of an Experimental In-Vessel Composting System
Carl L. Potter, John A. Glaser, U.S. EPA, Cincinnati, OH; Majid A. Dosani, Srinivas Krishnan,
Timothy A. Deets, E. Radha Krishnan, IT Corporation, Cincinnati, OH 64
Integrated Systems To Remediate Soil Contaminated With Wood Treating Wastes
Makram T. Suidan, Amid P. Khodadoust, Gregory J. Wilson, Karen M. Miller,
University of Cincinnati, Cincinnati, OH; Carolyn M. Acheson, Richard C. Brenner, U.S. EPA,
Cincinnati, OH 66
Biological Treatment of Contaminated Soils Using Redox Control
Margaret J. Kupferle, Tiehong L. Huang, Yonggui Shan, Maoxiu Wang, Guanrong You,
University of Cincinnati, Cincinnati, OH; Gregory D. Sayles, Carolyn M. Acheson, U.S. EPA,
Cincinnati, OH 68
Development of a Sulfate-Reducing Bioprocess To Remove Heavy Metals From Contaminated
Water and Soil
Munish Gupta, Makram T. Suidan, University of Cincinnati, Cincinnati, OH; Gregory D. Sayles,
Carolyn M. Acheson, U.S. EPA, Cincinnati, OH 71
Development of Techniques for the Bioremediation of Chromium-Contaminated Soil and Ground Water
Michael J. Mclnerney, Nydia Leon, Veronica E. Worrell, John D. Coates, University of
Oklahoma, Norman, OK 73
-------
Contents (continued)
Page
Bioremediation of Chlorinated Pesticide-Contaminated Sites Using Compost
James C. Young, Jean-Marc Bollag, Raymond W. Regan, Pennsylvania State University,
University Park, PA 75
Reductive Electrolytic Dechlorination
John W. Norton, Jr., Makram T. Suidan, University of Cincinnati, Cincinnati, OH; Carolyn M.
Acheson, Albert D. Venosa, U.S. EPA, Cincinnati, OH 76
Biological Ex Situ Treatment of Soil Contaminated With Polynuclear Aromatic Hydrocarbons
Carl L. Potter, U.S. EPA, Cincinnati, OH; Roy C. Haught, IT Corporation, Cincinnati, OH 77
Effectiveness of Gas-Phase Bioremediation Stimulating Agents (BSAs) for Unsaturated
Zone In Situ Bioremediation
James G. Uber, Ronghui Liang, R. Scott Smith, University of Cincinnati, Cincinnati, OH;
Paul T. McCauley, U.S. EPA, Cincinnati, OH 78
Section 5: Process Research 81
Monitoring Crude Oil Mineralization in Salt Marshes: Use of Stable Carbon Isotope Ratios
Andrew W. Jackson, John H. Pardue, Louisiana State University, Baton Rouge, LA 83
Mercury and Arsenic Biotransformation
Ronald S. Oremland, U.S. Geological Survey, Menlo Park, CA 86
Monod Degradation Kinetics of Quinoline in Natural and Microbially Enriched Methanogenic Microcosms
E. Michael Godsy, Ean Warren, Barbara A. Bekins, U.S. Geological Survey, Menlo Park, CA 87
Stimulating the Biotransformation of Polychlorinated Biphenyls
John F. Quensen, III, Stephen A. Boyd, James M. Tiedje, Michigan State University, East
Lansing, Ml 89
Bioaugmentation for In Situ Co-metabolic Biodegradation of Trichloroethylene in Ground Water
Junko Munakata Marr, Perry L. McCarty, Stanford University, Stanford, CA; V. Grace Matheson,
Larry J. Forney, James M. Tiedje, Michigan State University, East Lansing, Ml; Stephen
Francesconi, Malcolm S. Shields, University of West Florida, Pensacola, FL;
PH. Pritchard, U.S. EPA, Gulf Breeze, FL 93
Biodegradation of Chlorinated Solvents
Larry Wackett, Lisa Newman, Sergey Selifonov, University of Minnesota, St. Paul, MN;
Peter Chapman, Michael Shelton, U.S. EPA, Gulf Breeze, FL 96
Biological and Nutritional Factors Affecting Reductive Dechlorination of Chlorinated Organic Chemicals
Dingyi Ye, National Research Council, Athens, GA; W. Jack Jones, U.S. EPA, Athens, GA 98
Predicting Heavy Metal Inhibition of the In Situ Reductive Dechlorination of Organics at the Petro
Processor's Superfund Site
John H. Pardue, Louisiana State University, Baton Rouge, LA 102
vi
-------
Contents (continued)
Page
Effect of Primary Substrate on the Reduction of 2,4-Dinitrotoluene
Jiayang Cheng, Makram T. Suidan, University of Cincinnati, Cincinnati, OH; Albert D. Venosa,
U.S. EPA, Cincinnati, OH 104
Surfactants in Sediment Slurries: Partitioning Behavior and Effects on Apparent Polychlorinated
Biphenyl Solubilization
Jae-Woo Park, John F. Quensen, III, Stephen A. Boyd, Michigan State University, East
Lansing, Ml 108
Partial Characterization of an Anaerobic, Aryl, and Alkyl Dehalogenating Microorganism
Xiaoming Zhang, National Research Council, Athens, GA; W. Jack Jones, John E. Rogers,
U.S. EPA, Athens, GA 109
Microbial Degradation of Petroleum Hydrocarbons in Unsaturated Soils: The Mechanistic Importance of
Water Potential and the Exopolymer Matrix
Patricia A. Holden, James R. Hunt, Mary K. Firestone, University of California, Berkeley, CA 111
Metabolic Indicators of Anaerobic In Situ Bioremediation of Gasoline-Contaminated Aquifers
Harry R. Beller, Martin Reinhard, Alfred M. Spormann, Stanford University, Stanford, CA 113
Contaminant Dissolution and Biodegradation in Soils Containing Nonaqueous-Phase Organics
Larry E. Erickson, L.T. Fan, J. Patrick McDonald, George X. Yang, Satish K. Santharam,
Kansas State University, Manhattan, KS 114
Protein Expression of Mycobacteria That Metabolize Polycyclic Aromatic Hydrocarbons
David E. Wennerstrom, University of Arkansas for Medical Sciences, Little Rock, AR;
Carl E. Cerniglia, National Center for Toxicological Research, Jefferson, AR 115
Section 6: Hazardous Substance Research Centers 117
Co-metabolic Biodegradation Kinetics of Trichloroethylene in Unsaturated Soils
Karen L. Skubal, Peter Adriaens, University of Michigan, Ann Arbor, Ml 119
The Effect of Water Potential on Biodegradation Kinetics and Population Dynamics
Astrid Hillers, Peter Adriaens, University of Michigan, Ann Arbor, Ml 120
Anaerobic-Aerobic Bioventing Development
Gregory Sayles, U.S. EPA, Cincinnati, OH; Munish Gupta, Makram T. Suidan, University of
Cincinnati, Cincinnati, OH 121
Development of Co-metabolic Bioventing: Laboratory Tests
Gregory Sayles, U.S. EPA, Cincinnati, OH; Alan Zaffiro, Jennifer Platt, IT Corporation,
Cincinnati, OH 123
vii
-------
Contents (continued)
Page
Evaluating the Environmental Safety of Using Commercial Oil Spill Bioremediation Agents
Jeffrey L. Kavanaugh, University of West Florida, Pensacola, FL; C. Richard Gripe, Carol B.
Daniels, U.S. EPA, Gulf Breeze, FL; Rochelle Araujo, U.S. EPA, Athens, GA; Joe E. Lepo,
University of West Florida, Pensacola, FL 125
UNIFAC Phase Equilibrium Modeling To Assess the Bioavailability of Multicomponent
Nonaqueous-Phase Liquids Containing Polycyclic Aromatic Hydrocarbons
Catherine A. Peters, Princeton University, Princeton, NJ 126
Field Evaluation of Pneumatic Fracturing Enhanced Bioremediation
Sankar N. Venkatraman, David S. Kosson, Rutgers University, Piscataway, NJ;
Thomas M. Boland, John R. Schuring, New Jersey Institute of Technology, Newark, NJ 127
Solids Suspension Characteristics Related to Slurry Biotreatment Performances
J.-W Jim Tzeng, Paul T. McCauley, John A. Glaser, U.S. EPA, Cincinnati, OH 128
VIM
-------
Executive Summary
The U.S. Environmental Protection Agency's (EPA's)
Office of Research and Development (ORD) hosted the
eighth annual Symposium on Bioremediation of Hazard-
ous Wastes: Research, Development, and Field Evalu-
ations in Rye Brook, New York, August 8-10,1995. More
than 250 people attended, including leading bioreme-
diation researchers, field personnel from federal, state,
and local agencies, and representatives from industry
and academia. Three speakers opened the symposium
with introductions and background information on biore-
mediation research.
Fran Kremer, Coordinator of the Bioremediation Field
Initiative and Symposium Chairperson, began by intro-
ducing several members of the scientific steering com-
mittee of the Biosystems Technology Development
Program (BTDP). The BTDP draws on ORD scientists
who possess unique skills and expertise in biodegrada-
tion, toxicology, engineering, modeling, biological and
analytical chemistry, and molecular biology.
George Pavlou, Deputy Director of EPA's Region 2
Emergency and Remedial Response Division in New
York City, presented a summary of recent advances in
bioremediation technology. Mr. Pavlou noted that new
techniques have increased the availability of in situ con-
taminants to biological degradation, and they have also
increased the range of contaminants that can be treated
biologically. EPA scientists have learned that microor-
ganisms are not only able to transform simple hydrocar-
bons but also such toxic and resistant contaminants as
chlorinated aromatics and heavy metal salts. Mr. Pavlou
provided a regional perspective on these developments,
describing how bioremediation has been put to use in
Region 2.
Timothy Oppelt, Director of the National Risk Manage-
ment Research Laboratory in Cincinnati, Ohio, dis-
cussed the future of bioremediation on the national
level. He explained that bioremediation has been essen-
tial to the development of cost-effective cleanup tech-
nology and that it has already been used at more than
450 sites. This technology could potentially save hun-
dreds of millions of dollars in future cleanup costs. While
bioremediation is now predominantly used for petroleum
decontamination, it will be applied to a wider range of
sites in the future. Mr. Oppelt concluded by describing
the reorganization of research under ORD and by warn-
ing that funding shortages may hinder the development
of bioremediation technology.
The 33 papers and 22 posters presented at the confer-
ence highlighted recent program achievements and re-
search projects aimed at bringing bioremediation into
more widespread use. Taken as a whole, these topic areas
represent a comprehensive approach to bioremediation
of hazardous waste sites. The presentations were or-
ganized into five key research and program areas:
Bioremediation Field Initiative: This initiative was in-
stituted in 1990 to collect and disseminate perform-
ance data on bioremediation techniques from field
application experiences. The Agency assists regions
and states in conducting field tests and in carrying
out independent evaluations of site cleanups using
bioremediation. Through this initiative, tests are un-
der way at Superfund sites, Resource Conservation
and Recovery Act corrective action facilities, and un-
derground storage tank sites. Three papers pre-
sented at the symposium were devoted to this key
program area.
Field Research: Once a bioremediation approach has
proven effective in a laboratory or pilot-scale treata-
bility study, it must be monitored and evaluated at a
field site. The objective of this level of research is
to demonstrate that the particular bioremediation
process performs as expected in the field. For most
bioremediation technologies, certain key factors con-
cerning applicability, such as cost effectiveness, can-
not be thoroughly evaluated until the approach is
scaled up and field tested. Ten papers and two post-
ers provided information on recent field research.
Performance Evaluation: Performance evaluation in-
volves assessing the extent and rate of cleanup for
particular bioremediation processes as well as moni-
toring the environmental fate and effects of contami-
nants and their biological byproducts. Two papers
and one poster addressed this area.
Pilot-Scale Research: Pilot-scale research provides
information on the operation and control of bioreme-
diation technologies and the management of proc-
ess-related residuals and emissions to enable the
full-scale application of a technology. Given the ex-
panding base of experience with various bioremedia-
tion methods, the need for pilot-scale research is
IX
-------
increasing. Ten papers and three posters were pre-
sented concerning research based on microcosms of
field sites.
Process Research: Process research involves isolat-
ing and identifying microorganisms that carry out
biodegradation processes as well as developing tech-
niques for modeling and monitoring such processes.
This research is fundamental to the development of
new biosystems for treatment of environmental pol-
lutants in surface waters, sediments, soils, and sub-
surface materials. Nine papers addressed this critical
area, focusing on the role of metals and chlorinated
organics in bioremediation. In addition, six poster
presentations discussed process research.
Hazardous Substances Research Centers: In addition
to presentations on research being carried out under
the BTDP, the symposium included nine poster pres-
entations from the EPA Hazardous Substance Re-
search Centers (HSRC). The scientists and engineers
involved in HSRC conduct EPA research sponsored
by the following centers: the Northeast Hazardous
Substance Research Center (Regions 1 and 2), the
Great Lakes and Mid-Atlantic Hazardous Substance
Research Center (Regions 3 and 5), the South/
Southwest Hazardous Substance Research Center
(Regions 4 and 6), the Great Plains and Rocky Moun-
tain Hazardous Substance Research Center (Re-
gions 7 and 8), and the Western Region Hazardous
Substance Research Center (Regions 9 and 10).
-------
Introduction
Bioremediation is one of the most promising technologi-
cal approaches to the problem of hazardous waste. This
process relies on microorganisms such as bacteria or
fungi to transform hazardous chemicals into less toxic
or nontoxic substances. There are several reasons why
such biological transformation is often more attractive
than direct chemical or physical treatment. Microorgan-
isms typically:
Directly degrade contaminants rather than merely
transferring them from one medium to another.
Employ metabolic degradation pathways that can ter-
minate with benign waste products (e.g., carbon di-
oxide and water).
Derive the food energy necessary to degrade con-
taminants from the contaminants themselves.
Can be used in situ to minimize disturbance of the
cleanup site.
For these reasons, microorganisms can be effective,
economical, and nondisruptive tools for eliminating haz-
ardous chemicals. Until recent years, however, the use
of bioremediation was limited by the lack of a thorough
understanding of biodegradation processes, their ap-
propriate applications, their control and enhancement in
environmental matrices, and the engineering tech-
niques required for broad application of the technology.
Because the U.S. Environmental Protection Agency (EPA)
believes that bioremediation offers an attractive alterna-
tive to conventional methods of hazardous waste cleanup,
it has developed a strategic plan for its acceptance
and use by the technical and regulatory communities.
The Agency's strategic plan is centered on site-directed
bioremediation research to expedite the development
and use of relevant technology. EPAs Office of Research
and Development (ORD) developed an integrated Bio-
remediation Research Program to advance the under-
standing, development, and application of bioremedia-
tion solutions to hazardous waste problems threatening
human health and the environment.
Related bioremediation studies are being carried out at
five EPA Hazardous Substance Research Centers
(HSRCs) under the direction of ORD's National Center
for Extramural Research and Quality Assurance
(NCERQA). EPA was authorized to establish these cen-
ters by provisions in the 1986 amendments to the Su-
perfund law calling for research into all aspects of the
"manufacture, use, transportation, disposal, and man-
agement of hazardous substances."
EPAs bioremediation research efforts have produced
significant results in the laboratory, at the pilot scale, and
in the field. The many accomplishments include aquifer
restoration, soil cleanup, process characterization, and
technology transfer. Research also focuses on extend-
ing the range of substances that can be treated with
biological agents. This symposium was held to present
and discuss recent developments in bioremediation re-
search undertaken during 1994 and 1995 under the
Biosystems Technology Development Program.
In this document, abstracts of paper and poster presen-
tations from the symposium are organized within five
key research and program areas:
Bioremediation Field Initiative
Field research
Performance evaluation
Pilot-scale research
Process research
The last section of this document includes abstracts of
presentations on bioremediation research performed as
part of the HSRC program.
XI
-------
Section 1
Bioremediation Field Initiative
The Bioremediation Field Initiative is one of the major components of EPA's Bioremediation
Research Program. The Initiative was undertaken in 1990 to expand the nation's field experience
in bioremediation techniques. The Initiative's goals are to more fully assess and document the
performance of full-scale bioremediation applications, to create a database of current field data on
the treatability of contaminants, and to assist regional and state site managers using or considering
bioremediation. The Initiative is currently tracking bioremediation activities at more than 400 sites
under government and private-sector jurisdiction, in both the United States and Canada. Perform-
ance evaluations are currently being conducted at nine sites, three of which were reported on at
this symposium.
Data were presented from work at the St. Joseph, Michigan, Superfund site on the use of iron as
an electron acceptor and the potential for natural attenuation of chlorinated solvents. A presentation
was given on the modeling of the natural attenuation of solvents at the ground-water and lake
interface. Studies also were carried out on the design and field applications of bioventing in the
bioremediation of jet fuel spills.
-------
Intrinsic Bioremediation of Trichloroethylene at the St. Joseph Aquifer/
Lake Michigan Interface: A Role for Iron and Sulfate Reduction
Jack Lendvay, Mike McCormick, and Peter Adriaens
University of Michigan, Ann Arbor, Michigan
Introduction
The anaerobic aquifer at the St. Joseph, Michigan, Na-
tional Priorities List (NPL) site was contaminated with
trichloroethylene (TCE), which has been shown to have
dechlorinated to cis- and trans-dichloroethylene (DCE),
vinyl chloride (VC), ethylene, and ethane. These prod-
ucts occur as a result of natural attenuation processes,
presumably under methanogenic conditions (1).
The flux of all alkyl halides into Lake Michigan is of major
public concern because of the suspected carcinogenic-
ity of VC. As the plume moves toward the aerobic sur-
face water, the dominant redox conditions can be
expected to change because of wave action and vertical
seepage, which promote the interchange of oxygen-rich
lake water and anaerobic ground water. This presenta-
tion provides preliminary results of laboratory investiga-
tions geared toward determining the prevailing redox
processes and the potential for natural attenuation of
chlorinated solvents at the interface.
Background
Three of the most important redox processes in the
natural anaerobic environment are the coupling of the
oxidation of organic matter to iron (Fe) (III) reduction, to
sulfate reduction, and to methanogenesis. These three
processes are considered mutually exclusive which, in
the anaerobic subsurface environment, results in the
development of spatially or temporally distinct redox
zones. It has been demonstrated in aquatic sediment
and aquifer samples that Fe(lll)-reducing bacteria can
outcompete sulfate reducers, as well as methanogens,
for organic matter (2, 3).
Mineral-bound Fe(lll) has been shown to contribute sig-
nificantly to the total oxidation capacity of both pristine
and contaminated aquifers, as it often represents the
most abundant anaerobic terminal electron acceptor (4,
5). The speciation of iron in aquifer solids is greatly
influenced by microbial processes, particularly under
redox conditions favoring sulfate- and Fe(lll)-reduction.
Depending on the temporal or spatial succession in the
development of subsurface redox conditions, Fe(ll) pro-
duced by iron reducers precipitates as iron sulfides once
sulfate-reducing conditions develop, or as iron oxides
such as magnetite (Fe3O4) in the absence of sulfide.
Alternatively, biogenically produced sulfide may precipi-
tate as FeS(1_x) (mackinawite) after reductive dissolution
of Fe(lll) minerals. The iron sulfide and iron oxide min-
erals thus formed may then contribute significantly to the
reduction capacity of aquifer solids and, in turn, play a
major role in the fate of organic contaminants. The
presence of these precipitated minerals is direct evi-
dence for past or present iron- or sulfate-reducing con-
ditions.
Materials and Methods
A three-stage iron analysis was performed on sediments
collected from the same depths in sampling wells 55AB
(upstream) and 55AD (near shore) to evaluate the oc-
currence of oxidized and reduced iron minerals, and to
calculate inorganic reducing equivalents present in aqui-
fer materials.
Bioavailable iron (Fe(lll)): Microbial Fe(lll)-reduction
has been shown to predominantly use amorphous
oxyhydroxides and goethite as terminal electron ac-
ceptors. Quantitation of these minerals can be ap-
proximated by extracting sediments according to
Lovley and Phillips (6).
Ferrous monosulfides and amorphous iron oxides
(Fe(ll)): Reduced Fe(ll) minerals resulting from mi-
crobial iron and sulfate reduction, predominantly fer-
rous sulfides and iron oxides, can be quantified using
a 24-hr extraction with 0.5 M hydrochloric acid (HCI)
(7). As this extraction removes bioavailable iron as
well, FeS can be approximated by substration.
Siderite, crystalline iron oxide, and magnetite (Fe(ll)):
This extraction represents the precipitated ferrous
-------
iron fraction in the absence of sulfate reduction. Wet
sediment samples were extracted with 5 M HCI for a
21-day period, and analyzed according to Heron and
Christensen (5).
This information was then used in conjunction with avail-
able data on ground-water chemistry (pH, redox, car-
bonate, sulfate, sulfide) and compiled in a chemical
equilibrium model (MINEQL+) to predict speciation of
iron. Precipitation of ferrous iron solids was based on
stability constants from the literature.
Results
The analysis of samples from wells 55AB and 55AD
indicated the presence of similar concentrations of
bioavailable Fe(lll) (5 mmol equiv./kg) (Table 1).
Whereas precipitated iron oxides and ferrous monosul-
fides increased dramatically toward the shore, however,
solids representative of the absence of sulfate reduction
significantly decreased (12 versus 5 mmol equiv./kg
aquifer material, after subtraction of bioavailable iron).
This observation, based on iron extraction data, was
confirmed by chemical modeling, which predicted a pre-
dominant occurrence of siderite in the sampling point
farthest from the lakeshore and increasing occurrence
of iron sulfide and mackinawite closer to shore and into
the lake. Iron- and sulfate-reducing activity was rela-
tively easily stimulated in these sediments.
Ongoing dechlorination experiments under sulfate- and
iron-reducing conditions, using TCE and VC, have re-
sulted in the production of t-DCE from TCE under iron-
reducing conditions. Whetherthe production of the trans
rather than the cis isomer is indicative of an abiotic
dechlorination mechanism (by Fe(ll)) remains to be elu-
cidated. VC did not dechlorinate under sulfate-reducing
conditions during the 2 months monitored to date.
Information available from previous field studies in col-
laboration with the Robert S. Kerr Laboratories (Dr. John
Wilson) suggested that neither methanotrophic nor
methanogenic activity predominates in the contaminant
plume underthe shoreline and Lake Michigan, based on
redox potential and oxygen and methane measure-
ments. Redox potentials and an increase in soluble iron
concentrations between both shore samples and below
the Lake Michigan bottom, however, suggest that iron
reduction may be the dominant process in regions near
plume emergence (8).
References
1. Wilson, J.T., J.W. Weaver, and D.H. Kampbell. 1994. Intrinsic
bioremediation of TCE in ground water at an NPL Site in St.
Joseph, Michigan. In: U.S. EPA Symposium on Intrinsic Bioreme-
diation of Ground Water. EPA/540/R-94/515. Washington, DC.
2. Barcelona, M.J., and T.R. Holm. 1991. Oxidation-reduction capaci-
ties of aquifer solids. Environ. Sci. Technol. 25:1,565-1,572.
3. Chapelle, F.H., and D.R. Lovley. 1992. Competitive exclusion of
sulfate reduction by Fe(lll)-bacteria: A mechanism for producing
discrete zones of high-iron ground water. Ground Water 30:29-36.
4. Heron, G., T.H. Christensen, and J.C. Tjell. 1994. Oxidation ca-
pacity of aquifer sediments. Environ. Sci. Technol. 28:153-158.
5. Heron, G., and T.H. Christensen. 1995. Impact of sediment-bound
iron on redox buffering in a landfill leachate-polluted aquifer (Vejen,
Denmark). Environ. Sci. Technol. 29:187-192.
Table 1. Determination of Iron and Sulfate Equivalents in St. Joseph Sediments
Sample Location and
Analysis
Well 55AB
1-hr/0.5M HCI
1 -day/0.5 M HCI
1-hr/0.25 M HCI and 0.25 M
NH2OH HCI
21 -day/5.0 M HCI
Ion chromatography
Well 55AD
1-hr/0.5 M HCI
1 -day/0.5 M HCI
1-hr/0.25 M HCI and 0.25 M
NH2OH HCI
21 -day/5.0 M HCI
Ion chromatography
Chemical Species
Analyzed
Iron Extractions
Limited Fe(ll)
Limited Fe(ll)
Limited Fe(ll) and
Bioavailable Fe(lll)
Limited Fe(ll)
Sulfate Extractions
Soluble sulfate
Iron Extractions
Limited Fe(ll)
Limited Fe(ll)
Limited Fe(ll) and
Bioavailable Fe(lll)
Limited Fe(ll)
Sulfate Extractions
Soluble sulfate
Chemical
Concentration
(mg/kg soil)
219.2 ±15.8
21 0.6 ±30.6
269.6 ± 29.7
881 .8 ± 66.2
(mg/L)
25.8 ± 3.5
(mg/kg soil)
333.4 ± 11.5
445.5 ± 16.9
270.8 ± 8.2
735.9 ± 48.7
(mg/L)
29.3 ± 4.2
Redox Equivalents
(mmol equiv./kg soil)
3.9 ± 0.3
3.8 ± 0.6
4.7 ± 0.5
15.8 ±1.2
(mmol equiv./L)
2.2 ± 0.3
(mmol equiv./kg soil)
6.0 ± 0.2
8.0 ± 0.3
4.9 ± 0.2
13.2 ±0.9
(mmol equiv./L)
2.4 ±0.4
-------
Lovley, D.R., and E.J.P. Phillips. 1987. Rapid assay for microbially 8. Adriaens, P., J.V. Lendvay, N. Katopodes, S. Dean, J.T. Wilson,
reducible ferric iron in aquatic sediments. Appl. Environ. Microbiol. and D. Kampbell. 1995. Intrinsic bioremediation of chlorinated sol-
53:1,536-1,540. vents at the St. Joseph, Michigan Aquifer/Lake Michigan Interface.
EPA/600/R-95/012. In: Proceedings of the 21st Annual RREL Re-
Heron, G., C. Crouzet, A.C.M. Bourg, and T.H. Christensen. 1994. search Symposium, Abstract.
Speciation of Fe(ll) and Fe(lll) in contaminated aquifer sediments
using chemical extraction techniques. Environ. Sci. Technol.
28:1,698-1,705.
-------
Modeling Intrinsic Remediation as Ground-Water Discharges to a Lake:
The Trichloroethylene Plume at St. Joseph, Michigan
Sean M. Dean and Nikolaos D. Katopodes
Department of Civil and Environmental Engineering, University of Michigan,
Ann Arbor, Michigan
Introduction
Contamination of ground water by chlorinated solvents
is widespread and has been in the forefront of public and
regulatory concern for the last decade. As a result,
considerable research efforts, both in the laboratory and
in the field, have addressed the potential for using bio-
logical processes to degrade these pollutants via either
in situ or onsite bioremediation technologies. Natural
attenuation has been observed to be responsible for
removal or partial transformation of both chlorinated and
nonchlorinated organic contaminants. At several sites,
naturally occurring reductive dechlorination has been
found to be responsible for the anaerobic transformation
of trichloroethylene (TCE) to lesser chlorinated interme-
diates, such as c- and t-dichloroethylene (DCE) and
vinyl chloride (VC), and to ethylene (1, 2). Because
aquifers become oxygenated near groundwater and
lake interfaces, concerns have been raised with respect
to surface water contamination by VC. In the current
study, a modeling approach addresses the fate of TCE
and its lesser chlorinated transformation products at this
anaerobic/aerobic interface. The modeling effort de-
scribes and predicts the fate of the chlorinated solvents
at and near the interface, taking into account ground-
water flow rates and microbial degradation rates, as well
as the oxygenating effects of wave action and lake water
intrusion near the shore line.
Numerical Model
The numerical model for the simulation of the hydrody-
namic, physicochemical, and biological processes that
take place at the lake-aquifer interface is validated
based on specific data from an application site at St.
Joseph, Michigan. Although the hydrodynamic proc-
esses are truly three-dimensional, most of the phenom-
ena of interest, such as migration of TCE, DCE, and VC
into the lake and transfer of dissolved oxygen into the
aquifer from water infiltrating through the surf region,
can be modeled by a two-dimensional model on the
vertical plane.
Model Components
Due to the significant time-scale difference between the
near-shore circulation and wave runup and breaking in
the lake compared with the flow in the porous media,
two separate models are constructed for the corre-
sponding hydrodynamic phenomena. The resulting flow
fields are then integrated in a single mass transport and
contaminant fate model. All three components of the
model are two-dimensional, covering a vertical plane
extending from a location inland where uniform flow and
mass flux are observed in the aquifer to a distance
inside the lake where most near-shore current activity
has diminished. The various modules are verified by
analytical solutions and intermodel comparisons.
Ground-Water Module
In the porous media, a finite-element module for variably
saturated flow has been constructed on a vertical plane.
This module uses pressure heads calculated by the lake
module as a boundary condition at the lake-aquifer in-
terface. For the St. Joseph site, this module has been
used in conjunction with two separate grids to take
advantage of naturally defined boundary conditions.
A coarse grid has been developed extending from Lake
Michigan to a ground-water divide approximately 900 m
inland. Zero flux boundaries are prescribed at the
ground-water divide and an underlying clay layer. A
seasonally varying flux boundary condition at the ground
surface reflects recharge from rainfall. To focus on the
lake-aquifer interface region, a refined grid has been
developed extending from Lake Michigan to a point
approximately 100 m inland. One year's output from
running this module on the coarse grid defines a sea-
-------
sonally varying inland boundary condition for the refined
grid.
Lake Module
The near-shore/free-surface flow simulation is based on
the numerical solution of the Navier-Stokes equations
by means of the finite-element method. For turbulent
flow, a widely accepted two-equation closure model is
employed together with certain approximations near the
bed and free-surface boundaries. At high Reynolds
numbers, an upwind formulation known as the Petrov-
Galerkin method of weighted residuals is introduced for
the suppression of nonlinear instabilities. The model can
predict the vertical structure of the flow from the seep-
age face between the aquifer and the lake to the free
surface. Wave action is incorporated, and special atten-
tion is focused on wave runup and breaking. The beach
is assumed to be a porous bed so that water from the
surf and break region is allowed to infiltrate and reach
the aquifer.
For the complete formulation of the problem, bed per-
meability resulting in seepage through the surf region
would be computed by simultaneous solution of the
free-surface flow problem in the lake with the associated
unsaturated flow problem in the subsurface domain. In
this model, bed seepage is introduced as a boundary
condition. This eliminates the difficulty of having to deal
simultaneously with two time scales without affecting the
robustness of the model. The bed seepage is averaged
over time to provide an interface boundary condition to
the ground-water flow module.
Contaminant Transport Module
The contaminants are assumed to be well mixed later-
ally. The contaminant fate and transport module uses a
finite-element model to solve the two-dimensional trans-
port equation based on the flow fields computed by the
ground-water and lake modules. This accounts for con-
taminant transport due to advection and dispersion in
the aqueous phase and for interphase mass transfer
due to sorption and volatilization.
Microbial transformations are incorporated using modi-
fied Monod kinetics to describe a source/sink term in
transport equation. Rate constants have been estimated
by a parallel experimental effort focusing on microbial
interactions. The microbial biomass is assumed to be
immobile below a limiting concentration at which slough-
ing occurs.
Conclusion
A two-dimensional finite element has been developed to
simulate the transport and biodegradation of chlorinated
solvents at and near ground-water and lake interfaces.
Example simulations consider the effects that factors
such as heterogeneities in the porous media, uncertain-
ties in parameter estimation, and varying recharge
through the beach have on the location and concentra-
tion of a plume of chlorinated solvents.
References
1. Freedman, D.L., and J.M. Gossett. 1989. Biological reductive
dechlorination of tetrachloroethylene and trichloroethylene to eth-
ylene under methanogenic conditions. Appl. Environ. Microbiol.
55(9):2,144-2,151.
2. Vogel, T.M., and P.L. McCarty. 1985. Biotransformation of
tetrachloroethylene to trichloroethylene, dichloroethylene, vinyl
chloride, and carbon dioxide under methanogenic conditions. Appl.
Environ. Microbiol. 49(5):1,080-1,083.
-------
The U.S. Environmental Protection Agency's
Development ofBioventing
Gregory D. Sayles
National Risk Management Research Laboratory,
U.S. Environmental Protection Agency, Cincinnati, Ohio
Research conducted in the mid to late 1980s by the U.S.
Air Force (1, 2), researchers in the Netherlands (3-6),
Texas Research Institute (7, 8), Battelle Memorial Insti-
tute (2, 9, 10, 11), Utah State University (11), and the
U.S. Environmental Protection Agency (EPA) (12),
among others, suggested that delivering air to the
vadose zone to promote biodegradation could be a low-
cost means of cleaning fuel-contaminated vadose zone
soils. This approach was motivated by the attempt to
solve two different remediation development problems:
Soil vacuum extraction for treatment of contaminated
vadose zones involved costly off-gas treatment and
only removed the volatile fraction of the contamination.
Oxygen delivery to the vadose zone to promote aero-
bic biodegradation using the approaches attempted
in promoting biodegradation in ground water (namely,
delivering oxygen saturated water or aqueous solu-
tions of hydrogen peroxide or nitrate to the contami-
nated area) was neither efficient nor cost-effective.
A process was needed that could deliver oxygen by
introducing air into the vadose at a rate that minimized
volatilization of the contamination. Several groups simul-
taneously developed what is now known as bioventing.
EPA recognized the potential cost savings of such a
technology over traditional remediation approaches and
began an aggressive bioventing development program
in 1990. The mission of the program, in essence, was
to develop bioventing so that it could be applied at as
many contaminated sites as possible. To date, EPAs
program has demonstrated or is currently developing
the use of bioventing for the following situations:
For operation with air injection.
In cold climates.
With soil warming.
For jet fuel/aviation fuel.
For nonfuel contaminants such as acetone, toluene,
polycyclic aromatic hydrocarbons (PAHs), and trichlo-
roethylene (TCE).
Table 1 provides a list of EPAs involvement in bioventing
research and development. The cumulative knowledge
of EPA, the Air Force, and Battelle Memorial Institute
regarding bioventing of fuel-contaminated sites was dis-
tilled in Principles and Practices Manual for Bioventing,
to be released in late 1995.
The next frontier for aerobic bioventing is the application
of the process to sites contaminated with chlorinated
solvents. EPA is currently involved in two laboratory and
field projects to develop "co-metabolic bioventing."
Co-metabolic bioventing is the promotion of the aerobic
biodegradation of chlorinated solvents, such as TCE, in
the vadose zone by delivering oxygen and, if necessary,
a volatile co-metabolite to the contaminated site. EPA
projects will consider two scenarios: 1) the co-metabolite
is a co-contaminant of the chlorinated solvent, and thus
only air must be delivered; and 2) the co-metabolite
must be delivered with the air stream and must therefore
be volatile.
In summary, EPA and its collaborators, primarily the Air
Force, the U.S. Coast Guard, and Battelle, has been
successful in developing bioventing into an inexpensive,
robust process, applicable to the cleanup of many con-
taminated sites.
References
1. Miller, R.N. 1990. A field-scale investigation of enhanced petro-
leum hydrocarbon biodegradation in the vadose zone combining
soil venting as an oxygen source with moisture and nutrient ad-
ditions. Ph.D. dissertation. Utah State University, Logan, UT.
-------
Table 1. Summary of EPA Bioventing Research and Development
Project/
Location
Dates
Contaminants Scale
Approach/Results
References
Coast Guard 1990-1991 Aviation fuel Lab, pilot
Station, Michigan
Hill AFB, Utah
Eielson AFB,
Alaska
Reilly Tar SF
Site, MN
Greenwood
Chemical SF
Site, Virginia
Dover AFB,
Delaware
1990-1994 Jet fuel (JP-4) Full
1991-1994 Jet fuel (JP-4) Lab, pilot
1992-
PAHs
Pilot
1993-1995 Acetone, Pilot
toluene, others
1995-
Principles and 1995
Practices Manual release
TCE
Fuels
Lab, pilot
Two
volumes
First test of air-injection bioventing. Showed 13-16
that air injection near the water table could
induce biodegradation in the vadose zone and
in ground water.
First full-scale air-injection bioventing. Showed 17
that low-rate air injection could supply
biodegradation oxygen demand and produce
no measurable surface emissions of volatile
organic compounds.
Showed that bioventing in cold climates is 18
feasible and that simple soil warming
techniques can increase the rate of
biodegradation when bioventing, thereby
decreasing the time required for remediation.
Attempting to show that PAHs at wood treating 20
sites can be remediated with bioventing.
Attempting to show that bioventing of sites N/A
contaminated with non-fuel, aerobically
biodegradable organics and in low permeability
soils is feasible.
Attempting to show that TCE can be treated N/A
with bioventing if the necessary co-metabolite
is present, either as a co-contaminant or
delivered as a volatile organic in the injected
air stream.
Volume 1: Principles of Bioventing N/A
Volume 2: The Practice of Bioventing
2. Miller, R.N., C.C. Vogel, and R.E. Hinchee. 1991. A field-scale
investigation of petroleum hydrocarbon biodegradation in the
vadose zone enhanced by soil venting at Tyndall AFB, Florida.
In: Hinchee, R.E., and R.F. Olfenbuttel, eds. In situ bioreclama-
tion. Stoneham, MA: Butterworth-Heinemann. pp. 283-302.
3. Staatsuitgeverij. 1986. Proceedings of a Workshop (March 21-
21). Bodembeschermingsreeeks No. 9. Biotechnologische
Bodemsanering. Rapportnr. 851105002. ISBN 90-12-054133, Or-
dernr. 250-154-59; Staatsuitgeverij Den Haag, The Netherlands.
pp. 31-33.
4. van Eyk, J., and C. Vreeken. 1988. Venting-mediated removal of
petrol from subsurface soil strata as a result of stimulated evapo-
ration and enhanced biodegradation. Med. Fac. Landbouww.
Riiksuniv. Gent 53(4b):1,873-1,884.
5. van Eyk, J., and C. Vreeken. 1989. Model of petroleum minerali-
zation response to soil aeration to aid in site-specific, in situ
biological remediation. In: Jousma et al., eds. Groundwater con-
tamination: Use of models in decision-making. Proceedings of
the International Conference on Ground-water Contamination.
Boston, MA/London, England: Kluwer. pp. 365-371.
6. van Eyk, J., and C. Vreeken. 1989. Venting-mediated removal of
diesel oil from subsurface soil strata as a result of stimulated
evaporation and enhanced biodegradation. In: Hazardous waste
and contaminated sites, Envirotech, Vienna. Vol. 2, Session 3.
ISBN 389432-009-5. Essen, Germany: Westarp Wiss. pp. 475-485.
7. Texas Research Institute. 1980. Laboratory-scale gasoline spill
and venting experiment. Interim Report No. 7743-5:JST American
Petroleum Institute.
8. Texas Research Institute. 1984. Forced venting to remove gaso-
line vapor from a large-scale model aquifer. Final Report No.
8210I-F:TAV American Petroleum Institute.
9. Hinchee, R.E., and M. Arthur. 1991. Bench-scale studies of the
soil aeration process for bioremediation of petroleum hydrocar-
bons. J. Appl. Biochem. Biotech. 28/29:901-906.
10. Hinchee, R.E., and S.K. Ong. 1992. A rapid in situ respiration test
for measuring aerobic biodegradation rates of hydrocarbons in
soil. J. Air Waste Mgmt. Assoc. 42(10):1,305-1,312.
11. Dupont, R.R., WJ. Doucette, and R.E. Hinchee. 1991. Assess-
ment of in situ bioremediation potential and the application of
bioventing at a fuel-contaminated site. In: Hinchee, R.E., and R.F.
Olfenbuttel, eds. In situ bioreclamation: Applications and investi-
gations for hydrocarbon and contaminated site remediation.
Stoneham, MA: Butterworth-Heinemann. pp. 262-282.
12. Wilson, J.T., and C.H. Ward. 1986. Opportunities for bioremedia-
tion of aquifers contaminated with petroleum hydrocarbons. J.
Ind. Microbiol. 27:109-116.
13. Ostendorf, D.W, and D.H. Kampbell. 1990. Boioremediated soil
venting of light hydrocarbons. Haz. Waste Haz. Mat. 1(4):319-334.
14. Kampbell, D.H., and J.T Wilson. 1991. Bioventing to treat fuel
spills from underground storage tanks. J. Haz. Mat. 28:75-80.
15. Kampbell, D.H., J.T. Wilson, and C.J. Griffin. 1992. Performance
of bioventing at Traverse City, Michigan. In: U.S. EPA. Sympo-
sium on Bioremediation of Hazardous Wastes. EPA/600/R-92/126.
pp. 61-64.
-------
16. Kampbell, D.H., C.J. Griffin, and F.A. Blaha. 1993. Comparison
of bioventing and air sparging for in situ bioremediation of fuels.
In: U.S. EPA. Symposium on Bioremediation of Hazardous
Wastes: Research, Development, and Field Evaluations.
EPA/600/R-93/054. pp. 61-67.
17. Sayles, G.D., R.C. Brenner, R.E. Hinchee, and R. Elliott. 1994.
Bioventing of jet fuel spills II: Bioventing in a deep vadose zone
at Hill AFB, Utah. In: U.S. EPA. Symposium on Bioremediation
of Hazardous Wastes: Research, Development and Field Appli-
cations. EPA/600/R-94/075. pp. 22-28.
18. Sayles, G.D., R.C. Brenner, R.E. Hinchee, A. Leeson, C.M. Vogel,
and R.N. Miller. 1994. Bioventing of jet fuel spills I: Bioventing in
a cold climate with soil warming at Eielson AFB, Alaska. In: U.S.
EPA. Symposium on Bioremediation of Hazardous Wastes: Re-
search, Development and Field Applications. EPA/600/R-94/075.
pp. 15-21.
19. Leeson, A., R.E. Hinchee, J. Kittel, G. Sayles, C. Vogel, and R.
Miller. 1993. Optimizing bioventing in shallow vadose zones in
cold climates. Hydrological Sci. J. 38(4).
20. McCauley, P.T., R.C. Brenner, F.V. Kremer, B.C. Alleman, and
D.C. Beckwith. 1994. Bioventing soils contaminated with wood
preservatives. In: U.S. EPA. Symposium on Bioremediation of
Hazardous Wastes: Research, Development and Field Applica-
tions. EPA/600/R-94/075. pp. 40-45.
10
-------
Section 2
Field Research
Field research is essential for evaluating the performance of full-scale bioremediation precesses
and for conducting accelerated testing on technologies that are appropriate for scaled-up applica-
tion. For example, problems associated with the use of bacteria used in the laboratory include
optimizing the activity of the organism under site conditions and defining the risks associated with
introducing a non-native microorganism to the site. The objective of this level of research is to
demonstrate that the particular bioremediation process performs as expected in the field. Re-
searchers at the symposium provided information on several ongoing field experiments.
Field studies conducted at St. Joseph, Michigan, and Rocky Point, North Carolina, sought to
determine the extent of intrinsic bioremediation in subsurface contaminant plumes. EPA found
extensive dechlorination of trichloroethylene contamination by endogenous microorganisms at St.
Joseph. At Rocky Point, researchers found considerable agreement between laboratory models of
hydrocarbon biodegradation and field observations.
At Pensacola, Florida, researchers obtained data from a site contaminated with creosote to
construct a model of toxic inhibition of bioremediation. Anothergroup of scientists constructed a set
of models to better understand the processes of soil vapor extraction and bioventing.
Research also was performed on the use of a mixed oxygen/nitrate electron acceptor condition to
degrade aromatic hydrocarbons.
Finally, a group of presentations dealt with the use of bioremediation to clean petroleum contami-
nation in beaches and wetlands. Researchers performed studies to determine how often nutrients
should be applied to beaches to support bacterial growth, as well as directly studied the process of
bacterial petroleum degradation on beaches. Another study examined what environmental influ-
ences might affect the biodegradation of petroleum in wetland areas.
There were two field research poster presentations. The first of these described the National Center
for Integrated Bioremediation Research and Development (NCIBRD) and its work at Wurtsmith Air
Force Base in Michigan. The second presentation described use of plant growth to enhance
bioremediation in the field.
11
-------
A Review of Intrinsic Bioremediation of Trichloroethylene in Ground Water at
Picatinny Arsenal, New Jersey, and St. Joseph, Michigan
John T. Wilson, Don Kampbell, James Weaver, and Barbara Wilson
U.S. Environmental Protection Agency, Ada, Oklahoma
Tom Imbrigiotta and Ted Ehlke
U.S. Geological Survey, Trenton, New Jersey
Reductive dechlorination occurs frequently in large
trichloroethylene (TCE) plumes. TCE is transformed
largely to cis-dichloroethylene (cis-DCE), then to vinyl
chloride, and finally to compounds that do not contain
organic chlorine. This abstract evaluates the rate and
extent of natural reductive dechlorination of TCE in two
large plumes with similar properties.
Description of the Plumes
Both plumes originated in a release of liquid TCE. The
plume at St Joseph, Michigan, originates from an indus-
trial park, while the plume at the Picatinny Arsenal, New
Jersey, originates in a release from a degreasing vat at
a plating shop. Cross sections of the plumes are de-
picted in Figures 1 and 2. Both plumes have high con-
centrations of TCE in the core of the plume (over 25,000
u,g/L), are devoid of oxygen or nitrate, contain low con-
centrations of iron (II) and methane (generally less than
10 mg/L), and have relatively low concentrations of
sulfate (generally less than 15 mg/L). Both plumes have
concentrations of dissolved organic carbon that are ele-
vated over background. The ground water in both
plumes is cold (near 10°C). The water is hard, with pH
near neutrality.
Both plumes discharge to surface water. The interstitial
seepage velocities of the plumes are very similar. The
seepage velocity of the TCE plume at St. Joseph (cor-
rected for retardation) is near 0.1 m/day, while the ve-
locity of the plume on the Picatinny Arsenal varies from
0.3 to 1.0 m/day. For purposes of calculation, 0.3 m/day
is used in this abstract.
Monitoring
The plume at St. Joseph was characterized by four
transects that extended across the plume, perpendicular
to ground-water flow. At each point in each transect,
water was sampled in 1.5-m vertical intervals extending
from the water table to a clay layer at the bottom of the
aquifer. Each transect contains at least 20 sampling
points. Table 1 compares the average concentration of
TCE, cis-DCE, and vinyl chloride in each transect, as
well as the highest concentration encountered. The
most distant transect was sampled from the sediments
of Lake Michigan. The plume was encountered approxi-
mately 1.5 m below the sediment surface, 100 m from
the shore line.
The plume of TCE at the Picatinny Arsenal is monitored
by a series of well clusters installed along the centerline
of the plume. Table 1 presents data from the monitoring
well in a cluster that had the highest concentration of
TCE. The data were collected in 1989.
Extent of Attenuation
Dechlorination in the plume at St. Joseph is extensive.
Vinyl chloride and cis-DCE accumulated near the spill,
then were degraded as the plume moved downgradient
(Table 1). Dechlorination in the plume at Picatinny Arse-
nal was also extensive. Comparing the location of the
highest concentration with the point of discharge,
dechlorination destroyed approximately 90 percent of
the TCE. Vinyl chloride and cis-DCE did not accumulate
to an appreciable extent. Because the plume at Picat-
inny Arsenal discharged to surface water before
dechlorination was complete, the U.S. Army installed
and continues to operate a pump-and-treat system on
the plume.
Comparison of Attenuation due to
Dilution and Dechlorination
The plume at St. Joseph has high concentrations of TCE
at its core, while the concentration of chloride in the
aquifer is low. This makes it possible to estimate the
13
-------
2001
SCALE
VERTICAL EXAGGERATION 1:10
NORTH
PARKING LOT
CLAY
CLAY-
...--CLAY.
,,,"- "
CLAY
Figure 1. Cross section of the plume at St. Joseph, Michigan, as it leaves the industrial park and enters the sediments under Lake
Michigan. Concentrations are in ug/L total chloroethenes.
- 690 EXPLANATION
670 -
660 -
650 -
640 -
200
- 650
- 640
200 400 800 800 1,000 1.200 1,400 1,600 1,800 2.000
DISTANCE; FROM SOURCE, IN FEET
VERTICAL EXAGGERATION x21
DATUM IS SEA LEVEL
Concentration of
trlohloroethylene, in
micrograms per liter
[:^;>;::| <100
\//\ 100-1,000
|s^j 1,000-10,000
|'"3J >10,000
Sampling point
Figure 2. Cross section of the plume at Picatinny Arsenal, New Jersey, as it moves from its source near Building 24 and discharges
at Green Pond Brook.
contribution of dilution by comparing the accumulation
of chloride from reductive dechlorination to attenuation
of chloroethenes. Table 2 portrays the accumulation of
chloride and reduction of total organic chlorine along the
flow path. Table 2 compares water from the most con-
centrated sample in each transect. Based on KOC rela-
tionships and the fraction of organic carbon in the aqui-
fer, approximately 60 percent of the TCE in the aquifer
should be in solution. TCE was largely depleted, and
sorption of cis-DCE and vinyl chloride in the aquifer
14
-------
Table 1. Attenuation of TCE in Ground Water With Distance From the Source and Residence Time in the Aquifer (1-3)
Average cone. (ug/L)
Highest cone. (y.g/L)
Location
St. Joseph
Picatinny
Distance From Source
(m)
130
390
550
855
240
320
460
Time in Aquifer
(y)
3.2
9.7
12.5
17.9
2.2
2.9
4.2
TCE
6,500
63,000
520
8,700
15
56
<1
1.4
25,000
10,000
1,400
cis-DCE
8,100
723,000
830
9,800
18
870
<1
0.8
220
35
310
Vinyl
Chloride
930
4,400
450
7,660
106
205
<1
0.5
4
1
6
Table 2. Comparison of the Relative Attenuation of TCE, cis-DCE, and Vinyl Chloride With the Attenuation of Chloride in the
Plume at St. Joseph, Michigan (1, 2)
Highest concentration
Distance From
Source (m)
Background
130
390
550
855
Chloride Ion
(mg/L)
14
55
109
71
57
Organic Chlorine
(mg/L)
104
15
0.8
<0.1
TCE
4,000
8,700
11
1.4
c-DCE
128,000
9,800
828
0.8
Vinyl Chloride
4,400
1,660
205
0.5
should be minimal. We will assume that the organic
chlorine in ground water represents the pool of chlorine
available for dechlorination to chloride.
Near the source, the concentration of chloride plus po-
tential biogenic chloride minus background chloride was
145 mg/L. Only 38 percent of this quantity was actually
chloride. Total organic and inorganic chlorine attenuated
with distance downgradient. By the time the plume
reached the lake, the concentration of total chlorine
(minus background) was 43 mg/L, which is significantly
higher than background. Apparently the plume was at-
tenuated three- to four-fold due to dilution. Total attenu-
ation of chloroethenes was at least 100,000-fold.
Kinetics of Reductive Dechlorination in
Ground Water
Table 3 compares first-order rate constants calculated
between transects in the plume at St. Joseph and be-
tween monitoring wells in the plume at Picatinny Arse-
nal. Field-scale estimates of rates are also compared
with attenuation in microcosms constructed from mate-
rial collected along the flow path at Picatinny Arsenal.
There is surprising agreement in the rates of dechlori-
nation of TCE within the same plume, between plumes,
and between microcosm studies and field-scale esti-
mates. Nine separate estimates vary less than an order
of magnitude. The rates of degradation of vinyl chloride
and cis-DCE were comparable to the rates of degrada-
tion of TCE (Table 3).
The rates of attenuation in the two plumes are as slow
as humans experience time. In particular, they are slow
compared with the time usually devoted to site charac-
terization. In plumes with a long residence time, on the
order of decades, however, they have significance for
protection of waters that receive the plumes.
15
-------
Table 3. Rates of Reductive Dechlorination of TCE, cis-DCE, and Vinyl Chloride in Ground Water (residence time refers to time in
the segment of the plume being described, or incubation time of microcosms) (1-5)
Apparent Loss Coefficient (1/yr)
Location
Field Scale Estimates
St. Joseph
Picatinny
Laboratory Microcosm
Picatinny
Distance From
Source (m)
1 30 to 390
390 to 550
550 to 855
240 to 460
320 to 460
0 to 460
240 to 320
0 to 250
Studies
240
320
460
Time From
Source (yr)
3.2 to 9.7
9.7 to 12.5
12.5 to 17.9
2.2 to 4.2
2.9 to 4.2
0.0 to 4.2
2.2 to 2.9
0.0 to 2.3
2.2
2.9
4.2
Residence
Time (yr)
6.5
2.8
5.4
2.0
1.3
4.2
0.7
0.5
0.5
0.5
TCE
0.38
1.3
0.93
1.4
1.2
1.0
0.64
0.42
0.21
cis-DCE
0.50
0.83
3.1
Produced
Produced
1.6
0.5
0.52
9.4
3.1
Vinyl
Chloride
0.18
0.88
2.2
Produced
Produced
References
1. Semprini, L, P.K. Kitanidis, D.H. Kampbell, and J.T. Wilson. An-
aerobic transformation of chlorinated aliphatic hydrocarbons in a
sand aquifer based on spatial chemical distributions. Water Re-
sources Res. In press.
2. Wilson, J.T., J.W Weaver, and D.H. Kampbell. 1994. Intrinsic
bioremediation of TCE in ground water at an NPL site in St.
Joseph, Michigan. In: Symposium on Intrinsic Bioremediation of
Ground Water. EPA/540/R-95/515. pp. 154-160.
3. Martin, M., and I.E. Imbrigiotta. 1994. Contamination of ground
water with trichloroethylene at the Building 24 site at Picatinny
Arsenal, New Jersey. In: Symposium on Intrinsic Bioremediation
of Ground Water. EPA/540/R-95/515. pp. 143-153.
4. Ehlke, T.A., I.E. Imbrigiotta, B.H. Wilson, and J.T. Wilson. 1991.
Biotransformation of cis-1,2-dichloroethylene in aquifer material
from Picatinny Arsenal, Morris County, New Jersey. In: U.S. Geo-
logical Survey Toxic Substances Hydrology ProgramProceed-
ings of the Technical Meeting. Water-Resources Investigations
Report 91-4034. pp. 689-697.
5. Wilson, B.H., T.A. Ehlke, T.E. Imbrigiotta, and J.T. Wilson. 1991.
Reductive dechlorination of trichloroethylene in anoxic aquifer ma-
terial from Picatinny Arsenal, New Jersey. In: U.S. Geological Sur-
vey Toxic Substances Hydrology ProgramProceedings of the
Technical Meeting. Water-Resources Investigations Report 91-
4034. pp. 704-707.
16
-------
Intrinsic Bioremediation of a Gasoline Plume: Comparison of Field and
Laboratory Results
Morton A. Barlaz, Melody J. Hunt, Sreenivas Kota, and Robert C. Borden
North Carolina State University, Raleigh, North Carolina
Introduction
Assessing the potential for natural bioremediation in the
subsurface is complicated by site-specific conditions
and the methods used to estimate biodegradation rates.
Controlled laboratory experiments often are necessary
to verify biological loss of a compound and to assess
factors that influence biodegradation. The effect of re-
moving samples from such a stable environment and
placing them in laboratory microcosms, however, is not
understood. In situ columns have been used to measure
biodegradation on a limited basis, and little is known
about their reliability. In this paper, we use laboratory
microcosms and in situ column experiments to estimate
intrinsic biodegradation rates of benzene, toluene, ethyl-
benzene, and xylene (BTEX) isomers in the subsurface.
Site Background
This research was conducted at a petroleum-contami-
nated aquifer in the southeastern coastal plain near
Rocky Point, North Carolina. The plume is characterized
by negligible dissolved oxygen and redox potentials of
-100 to -200 mV due to intrinsic biodegradation of BTEX
(1). The dominant electron acceptors within the plume
are sulfate and iron (1). The midpoint of the plume is
characterized by high dissolved iron (Fe)(ll) (greater
than 40 mg/L) and low SO4-2 concentrations (less than
4 mg/L). Toluene and o-xylene are nearly depleted (less
than 20 u,g/L), whereas high quantities of benzene,
ethylbenzene, and m-,p-xylene remain (greater than
500 ug/L).
Experimental Methods
Laboratory Microcosms
Multiple replicate microcosms with no headspace were
constructed in an anaerobic chamber under aseptic con-
ditions using blended aquifer sediment and ground
water recovered under anaerobic conditions. Microcosm
preparation was designed to simulate ambient condi-
tions to the maximum extent possible. Microcosms
were spiked with approximately 10,000 u,g/L BTEX
(2,000 u,g/L of each compound) and incubated in an-
aerobic containers stored at the ambient ground-water
temperature, 16°C. Because it cannot be distinguished
from m-xylene by the analytical procedure used, p-
xylene was not added. BTEX loss was monitored by
destructively sampling three live and three abiotic micro-
cosms at monthly intervals for 300 days. A final time
point was taken 100 days later (after 400 days).
In Situ Columns
The in situ columns were similar to a system used
previously (2). Each column consisted of a 1-m long
chamber where sediment and ground water were iso-
lated from the surrounding aquifer. Two sets of columns,
Group A and Group B, were installed at the midpoint
area of the plume. Each set contained three individual
columns: two live and one abiotic control. After installa-
tion, the columns were filled with anaerobic ground
water containing BTEX, which had been recovered from
nearby wells. The contaminant concentrations in
water added to Group A were 1,300, 40, 1,800, 700,
and 15 u,g/L for benzene, toluene, ethylbenzene,
m-,p-xylene, and o-xylene, respectively. The approxi-
mate concentrations of compounds added to Group B
columns were 200, 200, 600, 1,600, and 800 u,g/L for
benzene, toluene, ethylbenzene, m-,p-xylene, and
o-xylene, respectively. The abiotic control columns were
prepared by adjusting the pH to less than 2 with hydro-
chloric acid (HCI). Tracer tests were conducted on all
columns before each experiment began to ensure that
they were properly installed.
Results
A distinct order of compound disappearance was meas-
ured in the laboratory incubations: m-xylene degrada-
tion began with no lag period, followed by toluene,
17
-------
10000
1000 F
"" Abiotic Ben/,ene
Q Benzene
Toluene
Ethylbenzene
400
- m-Xylene
~ o-Xylene
Figure 1. BTEX biodegradation in laboratory microcosms.
o-xylene, benzene, and ethylbenzene (Figure 1). The
rate of m-xylene loss slowed once toluene loss be-
gan; once toluene and o-xylene were below 20 u,g/L
(120 days), the rate of m-xylene loss increased. The
aquifer material was obtained in an area of the plume
where toluene and o-xylene concentrations were very
low (less than 50 u,g/L) but significant quantities of m-,p-
xylene remained (greater than 1,000 u,g/L). Thus, the
microbial population appeared to have an initial prefer-
ence for m-xylene, but switched to toluene and o-xylene
after a 22-day acclimation period. Benzene began to
biodegrade once m-xylene was depleted and was at or
below 10 u,g/L in all microcosms at the final sampling
(403 days). First-order decay rates (K) were determined
during the time of loss for each compound (Table 1).
All live and abiotic in situ columns exhibited an initial
concentration decrease of several hundred micro-
grams/per liter between the injection water and the first
sample taken from the chamber. This initial loss is attrib-
uted to sorption. After the sorption loss, the initial com-
pound concentrations were less than 500 u,g/L in most
I
u
§
U
100
10
ll
Live Column -1
Live Column - 2
Abiotic Column
j I
50
100
150
Days
200 250
300
Figure 2. m-,p-Xylene biodegradation in Group A in situ
columns.
columns. The concentrations of hydrocarbons in the
abiotic columns remained fairly constant or declined
slowly after the initial decrease, indicating biological
activity or short circuiting did not occur in the control
columns.
In Group A, benzene and m-,p-xylene exhibited signifi-
cantly higher losses in the live columns relative to the
abiotic columns. The concentration of m-,p-xylene de-
creased in the live columns after an initial lag of 85 to
121 days (Figure 2). Benzene concentrations remained
constant in both live columns for 155 days, after which
time decreases attributed to biological activity were
measured (data not shown). Initial toluene and o-xylene
concentrations were too low (less than 50 u,g/L) to ac-
curately measure concentration changes. Figure 2
shows the measured m-,p-xylene loss in the Group A
columns and illustrates the timeframe used to calculate
the decay rates. Results from the two live columns were
pooled to estimate the live decay rate. In Group B,
significant biological loss of toluene occurred with no
apparent lag time. The short sampling period of 75
days was not adequate to measure losses of the other
compounds.
Table 1. Comparison of Microcosm and In Situ Column
Biodegradation Rates3
Compound
Benzene
Toluene
m,p-xylene
o-xylene
Ethylbenzene
Laboratory Rate
(percent -day"1)
(time interval in
days)
2.37 (184 to 403)
4.46 (22 to 120)
2.04 (0 to 1 84)
5.59 (37 to 120)
0.19 (0 to 403)
In Situ Column Rate
(percent -day"1)
(time interval in
days)
0.41 (121 to 251)
1.15(13to75)
1.43(121 to 251)
NS
NS
Rate calculated is the difference between the live and abiotic loss
rates assuming first-order model.
NS = The difference between the live and abiotic loss rate is not
significant at the 95-percent confidence level.
Comparison of In Situ Columns and
Laboratory Microcosms
Biological loss for three of the five BTEX compounds
occurred over similar periods in the laboratory and in situ
experiments. In both cases, toluene degradation was
followed by m-,p-xylene and benzene. This order is
consistent with previous field investigations (1). Ethyl-
benzene loss was minimal in the laboratory microcosms
during the 400 days of incubation, and no ethylbenzene
degradation was measured during the 7 months of in
situ monitoring. Loss of o-xylene was not observed in
the Group B columns, but fairly rapid depletion concur-
rent with toluene loss was measured in the laboratory.
The initial concentration of o-xylene (less than 500u,g/L)
18
-------
was possibly too low to stimulate in situ degradation, or
the 75-day monitoring period could have been too short.
Although the monthly sampling frequency was consis-
tent for both types of measurements, the length of moni-
toring was shorter in the in situ columns due to the
limited sample volume available. Thus, direct compari-
son of decay rates between the two types of measure-
ments is difficult. Given these limitations, the measured
rates are comparable in both columns and microcosms.
The slightly lower decay rates measured in the in situ
columns may be due to the lower initial concentrations
used in these experiments. Biological decay was dem-
onstrated in the controlled column and microcosm ex-
periments. Use of in situ columns could provide a prac-
tical link between laboratory evaluations and full-scale
field studies.
References
1. Borden, R.C., C.A. Gomez, and M.T. Becker. 1995. Geochemical
indicators of natural bioremediation. Ground Water 82(2):180-89.
2. Gillham, R.W., R.C. Starr, and D.J. Miller. 1990. A device for in situ
determination of geochemical transport parameters; two biochemi-
cal reactions. Ground Water 28(6):858-862.
19
-------
Toxicity Effects on Methanogenic Degradation of Phenol in Ground Water
Barbara A. Bekins, E. Michael Godsy, and Ean Warren
U.S. Geological Survey, Menlo Park, California
Introduction
At an abandoned creosote works located near Pensa-
cola, Florida, the shallow ground water is contaminated
with phenolic compounds, heterocyclic compounds, and
polyaromatic hydrocarbons. Based on the use of unlined
disposal ponds during the 80 years of operations at the
plant, the contaminants have probably been present in
the ground water for several decades. A methanogenic
consortium in the aquifer is degrading some of the com-
pounds, and the concentrations of the degradable frac-
tion drop to less than 1 percent of the source values by
100 to 150 m downgradient from the nonaqueous phase
(1). In spite of the long exposure time to a continuous
source, the results of acridine orange direct counts and
most probable number determinations indicate a low,
uniform microbial density (1). The continued existence
of low microbial numbers suggests that some factor is
limiting growth. Several possibilities, such as microbial
transport, nutrient limitation, predation, and toxicity,
have been examined. Of these, toxicity appears to be
the most promising explanation. The toxicity of creosote
compounds to various organisms has been studied for
a long time (2). Very little work has been done, however,
on the effects of creosote on methanogenic consortia
known to be especially sensitive to toxic compounds (3).
The results of our work indicate that the dimethylphenols
and methylphenols present in the ground water at this
site inhibit the degradation of phenol. Furthermore, in-
corporating these inhibition effects into a one-dimen-
sional model of the aquifer predicts a steady-state
degradation profile and zero net growth of the active
methanogenic consortium.
Toxicity Assay Results
A serum bottle assay similar to that described by Owens
et al. (3) was performed to determine which of the
compounds present in the ground water might be toxic
to the methanogenic consortium. These compounds
were grouped by class and added at three concentration
levels equivalent to 1.5, 1.0, and 0.5 times the highest
measured field value. The classes of compounds tested
were 1) indene, benzothiophene, 2-methylnaphthalene,
biphenyl, flourene, and 2-naphthol; 2) 2-, 3-, and 4-
methylphenol; 3) 2,4- and 3,5-dimethylphenols; 4) qui-
noline and isoquinoline; 5) 2(1H)-quinolinone and
1(2H)-isoquinolinone; and 6) all of the preceding com-
pounds combined. Duplicate 100-mL serum bottles con-
taining the target compounds, enriched methanogenic
culture derived from the aquifer, mineral salts, and phe-
nol as the growth substrate (at the highest concentration
observed in the field) were prepared in an anaerobic
glove box and capped. The volume of gas produced in
each bottle was monitored by allowing a wetted glass
syringe inserted through the septum to equilibrate with
atmospheric pressure. Figure 1 shows the gas produc-
tion in the bottles containing Mixtures 2, 3, 4, and 6 at
the 1.5 times concentration as well as a control contain-
ing only phenol. Methylphenols and dimethylphenols
showed a substantial toxicity effect, whereas nitrogen
heterocycles had a smaller effect and polyaromatic hy-
drocarbons had no measurable effect (data not shown).
--»-- Methyphenols
- Dimethylphenols
-* Nitrogen heterocycles
x All combined
Control
Figure 1. Gas production by the aquifer microbes with phenol
as the growth substrate and various inhibitor concen-
trations equal to 1.5 times the maximum value ob-
served in the aquifer. The results are not shown for
quinolinone, which was similar to quinoline, or for
polyaromatic hydrocarbons, which were similar to
the positive control.
20
-------
In the bottles with concentrations equal to those in the
field, dimethylphenols had a substantial effect, and
methylphenols had a smaller effect. In the 0.5 times
concentration bottles, methylphenols and dimethylphe-
nols had a slight effect. Preliminary results showing the
buildup of fatty acids suggest that the intermediate steps
in the degradation process are being inhibited.
Model Results
The monitoring of ground-water concentrations of the
degradable compounds for more than 12 years shows
that the concentration profiles are constant in time. The
existence of a steady-state degradation profile of each
substrate together with a low, uniform microbial density
indicates that the microbial numbers do not change with
time. In theory, the functional form of the Monod growth
expression cannot be balanced by a constant decay
rate. To address this problem, toxicity effects are incor-
porated into the following equations for one-dimensional
substrate transport with degradation and microbial
growth:
-V^-T?
0.07
S
Ksfl
sc] _ s2
*~K \+S+~K
KC I ft/
(Eq. 1)
dB
dt ''
B
(Eq. 2)
where R is the retardation factor, S is phenol concentra-
tion, v is the flow velocity, D is dispersion, \im is the
maximum growth rate, Y is the yield, B is biomass, 9 is
porosity, Ks is the half saturation constant, K, and KC are
haldane and competitive inhibition constants, Sc is the
concentration of the inhibiting compound, and kj is the
biomass decay or maintenance rate. When only the
toxicity of phenol is incorporated using the Haldane
inhibition model, the predicted growth is about 50 per-
cent lower but still much higher than the only published
decay rate (4). In addition, the equations do not produce
a steady-state solution. Incorporating the effect of the
dimethylphenol toxicity produces a steady-state solution
for phenol and microbial concentrations that matches
the character of the data (Figure 2). The values of the
parameters used in the solution are given in the figure
100 200
Distance (m)
100 200
Distance (m)
Figure 2.
Solution to Equation 1 for phenol after 2,000, 6,000,
and 10,000 days (left), and coupled solutions to Equa-
tion2 for microbe concentrations at the same times.
The values of the parameters used were R = 1.01,
S(0) = 26 mg/L, v = 1.0 mid, D = 1.0 m2/d,|im = 0.111,
Y = 0.013, Emit = 0.005 mg/L(1.6 x 10s per 100g),6 =
0.38, Ks = 1.33, KI = 250 mg/L, Kc = 0.52 mg/L, Sc = 23
exp(x2/(2 (47)2)) mg/L (an empirical fit to the observed
dimethylphenol concentrations), and kd =0.0326 d"1
caption. The model results show that the aquifer con-
centrations take about 6 years to evolve to a steady
state, while the microbial population takes about 25 years.
The population of aquifer microorganisms oscillates
as they adjust their distribution to account for two
competing effects: 1) the maximum concentration of phe-
nol nearthe source should lead to maximum growth and
substrate utilization there, and 2) the maximum concen-
trations of dimethylphenols near the source lead to
maximum inhibition of growth and substrate utilization.
The result is a tradeoff between a location where the
growth substrate concentrations are higher versus one
farther from the source where the inhibitor concentra-
tions are lower. The final microbial concentrations stabi-
lize at about an order of magnitude higher 50 m from the
source than immediately adjacent to it.
References
1. Godsy E.M., D.F. Goerlitz, and D. Grbic-Galic. 1992. Methano-
genic biodegradation of creosote contaminants in natural and
simulated ground-water ecosystems. Ground Water 30:232-242.
2. Mayfield, P.B. The toxic elements of high-temperature coal tar
creosote. 1951. Proc. Am. Wood-Preserver's Assoc. 47:62-85.
3. Owen, W.F., D.C. Stuckey, J.B. Healy, L.Y. Young, and PL.
McCarty. 1979. Bioassay for monitoring biochemical methane po-
tential and anaerobic toxicity. Water Research Res. 13:485-492.
4. Bekins, B.A., E.M. Godsy, and D.F. Goerlitz. 1993. Modeling
steady-state methanogenic degradation of phenols in groundwa-
ter. J. Contam. Hydrol. 14:279-294.
21
-------
A Multiphase, Multicomponent Numerical Model of Bioventing With
Nonequilibrium Mass Exchange
Linda M. Abriola, John R. Lang, and Klaus M. Rathfelder
Department of Civil and Environmental Engineering, University of Michigan,
Ann Arbor, Michigan
Introduction
Soil vapor extraction (SVE) and bioventing (BV) are
common remediation practices for unsaturated soils
contaminated with volatile organic compounds (VOCs).
These methods have been demonstrated to be effec-
tive at comparatively low costs. The efficiency of these
techniques is known to be restricted by soil charac-
teristics; by mass transfer limitations between phases,
including liquid/solid, liquid/gas, and liquid/microorganism
phases; by the availability of oxygen; and by system
design and operation parameters (1). Assessment of
SVE/BV systems is often hindered by the complex inter-
play of physical, chemical, and biological processes.
Consequently, design and operation of these systems
are typically based on engineering experience and/or
simple design equations. Numerical models of SVE/BV
systems can be valuable tools for the investigation of
the effects of various processes on system performance
and for optimal system design. In this work, a numerical
model is presented that has been specifically developed
to incorporate the complete range of processes occur-
ring at the field scale and to include interphase mass
transfer rate limitations.
Model Formulation
Three fluid phases are modeled: gas, aqueous, and a
nonaqueous phase liquid (NAPL). The gas and aqueous
phases may flow simultaneously in response to applied
pumping/injection or density gradients. The movement
of these phases is described by standard macroscopi-
cally averaged flow equations (2). The NAPL phase is
assumed to be at an immobile residual saturation.
Changes in NAPL saturation, therefore, result solely
from interphase mass transfer.
The NAPL may be a mixture of an unrestricted number
of organic components. The gas phase is assumed to
be composed of nitrogen and oxygen (the two major
constituents of air), water vapor, volatile components of
the NAPL, and a single limiting nutrient. The aqueous
phase is composed of water, oxygen, soluble compo-
nents of the NAPL, and the limiting nutrient. Sorption to
the soil particles is restricted to components of the
NAPL. The migration of each component in each phase
is described by standard macroscopically averaged
transport equations (2).
Quantification of the biotransformation processes fol-
lows the conceptual approach of Chen et al. (3). Biode-
gradation is assumed to occur only within the aqueous
phase by an indigenous, spatially heterogeneous, mixed
microbial population that is present as attached micro-
colonies. There is no biomass transport or detachment
or sloughing of the microcolonies, and biomass growth
does not affect permeability. Monod-type kinetic expres-
sions are employed to describe biophase utilization of
substrates, oxygen, and a limiting nutrient, as well as
growth of the microbial population. Additionally, a mini-
mum biophase concentration reflecting the indigenous
population is maintained when growth is restricted due
to oxygen, substrate, or nutrient limitations.
A linear driving force expression is used to model non-
equilibrium interphase exchange. Interphase partition-
ing processes included in the model are: volatilization
and dissolution of components from the NAPL;
gas/aqueous exchange of oxygen, water vapor, and the
components of the NAPL; sorption of the NAPL compo-
nents to the soil particles through the aqueous phase;
and rate-limited uptake by the biophase of oxygen,
substrate (components of the NAPL), and the limiting
nutrient.
Numerical Solution
The flow and transport equations are solved in two
space dimensions (vertical cross section or radial ge-
ometry) using a standard Galerkin finite element method
22
-------
with linear triangular elements. A set-iterative scheme is
used for computational efficiency. The sets of coupled
flow, transport, and biodegradation equations, as well as
multiple equations within sets, are decoupled and
solved sequentially. Decoupling is accomplished by lag-
ging, either by one iteration or one time step, the cou-
pling terms which are phase density and interphase
mass exchange. Iteration within and between equation
sets is performed to account for nonlinearities and en-
sure solution accuracy. Numerical solutions of the flow,
transport, and biodegradation equations have been in-
dependently verified with analytical solutions and inter-
model comparisons. A detailed description of the model
and example simulations are presented in Lang et al. (4).
Demonstration of SVE and BV Simulations
Hypothetical field-scale SVE and BV systems are pre-
sented to demonstrate model capabilities. The modeled
scenario involves the remediation of a residual NAPL
distributed within a layered soil system. Here the
nonuniform initial NAPL distribution was generated with
a multiphase flow model. NAPL contamination is present
in both the unsaturated zone and capillary fringe.
SVE operations are examined by simulating relatively
large pumping rates to an extraction well positioned in
the center of the residual NAPL zone. Removal effi-
ciency in the test simulations is shown to be sensitive to
mass transfer rates, permeability contrasts, the initial
NAPL distribution, and, to a lesser extent, pumping rate
and well screen position.
A BV operation is also modeled by simulating small gas
injection rates at a well located in the center of the
contamination zone. Removal efficiency is shown to be
sensitive to flow rate, interphase mass transfer, biode-
gradation rates, and NAPL distribution.
An example of the complex interplay between chemical
and biological processes in BV systems is demonstrated
in simulation results shown in Figure 1. Here contami-
nant removal is compared for BV systems run at a
comparatively low flow (0.1 pore volumes/day) and high
flow (1 pore volume/day). Total mass removed (Figure
1a) is greater at high flow. Due to nonequilibrium inter-
phase partitioning occurring at the high flow, however,
the difference in mass removed is less than the propor-
tionate difference in flow rate. The rate of contaminant
interphase partitioning also affects the quantity of mass
removed by biodegradation (Figure 1b), which is far
greater at high flow. At low flow, interphase partitioning
of the contaminant is approximately at equilibrium, re-
sulting in downgradient aqueous concentrations that are
greater than an inhibitory threshold. Consequently,
biodegradation in the low-flow scenario is restricted to
the region upgradient of the NAPL contamination
zone. Nonequilibrium partitioning at high flow rates
produces downgradient aqueous concentrations be-
low the inhibitory threshold, resulting in enhanced
biodegradation. This greater degradation in the
high-flow system produces a reduction in the con-
taminant mass in the gas phase arriving at a down-
gradient extraction point (Figure 1c).
Conclusions
A numerical model of SVE/BV systems has been devel-
oped that incorporates the complete compositional and
biological processes representative of field conditions.
Example simulations demonstrate the model capabili-
ties and illustrate the complex interplay of chemical,
physical, and biological processes occurring in SVE/BV
systems.
References
1. Rathfelder, K., J.L. Lang, and L.M. Abriola. 1995. Soil vapor ex-
traction and bioventing: Applications, limitations, and future re-
search directions. Reviews of Geophysics IUGG Quadrennial
Report. In press.
2. Abriola, L.M. 1989. Modeling multiphase migration of organic
chemicals in groundwater systemsA review and assessment.
Environmental Health Perspectives 83:117-143.
3. Chen, Y.-M., L.M. Abriola, P.J.J. Alverez, P.J. Anid, and T.M. Vogel.
1992. Modeling transport and biodegradation of benzene and tolu-
ene in sandy aquifer material: Comparisons with experimental
measurements. Water Resources Res. 28(7):1,833-1,847.
4. Lang, J.L., K.M. Rathfelder, and L.M. Abriola. 1995. A multiphase,
multicomponent numerical model of bioventing with non-equilib-
rium mass exchange. In: Proceedings of the Bioremediation Sym-
posium, San Diego, CA. April.
15
10
(a) Total Mass Removed
(b) Biodegradation
(c) Gas Phase Recovery
10
20 30 40
time (day)
high flow
- low flow
20 30
time (day)
10
20 30
time (day)
40 50
Figure 1. Contaminant removal versus time for hypothetical BV scenarios at comparatively low and high flow rates.
23
-------
Aromatic Hydrocarbon Biotransformation Under Mixed Oxygen/Nitrate Electron
Acceptor Conditions
Liza P. Wilson, Peter C. D'Adamo, and Edward J. Bouwer
Department of Geography and Environmental Engineering, The Johns Hopkins University,
Baltimore, Maryland
Introduction
Biodegradation of contaminants associated with sedi-
ments and ground water under mixed oxygen/nitrate
electron acceptor conditions may prove to be more suc-
cessful and feasible than remediation under strict aero-
bic or anaerobic conditions. In particular, the low level of
oxygen may allow subsurface microorganisms to attack
the aromatic rings of many organic compounds (using
oxygenases), with nitrate serving as the electron ac-
ceptor to complete the degradation. Providing nitrate to
the subsurface is less expensive than maintaining aero-
bic conditions, and, as nitrate is highly soluble, it is
easier to maintain a residual concentration in ground
water.
A laboratory investigation is being conducted to provide
a better understanding of the effect of dual oxygen/ni-
trate electron acceptor conditions on the biodegradation
of monocyclic and polycyclic aromatic hydrocarbon mix-
tures in aqueous solution. The specific objectives of the
research are 1) to quantify the stoichiometry and kinetics
of biodegradation of a mixture of aromatic hydrocarbons
under microaerophilic conditions (defined as less than
or equal to 2 mg/L O2), and 2) to assess the relative
efficacy of bioremediation under microaerophilic condi-
tions compared with strict aerobic or denitrification con-
ditions in the laboratory, using batch microcosms and
aquifer sediment columns.
Microaerophilic Biodegradation by an
Enrichment of Aquifer Bacteria
The microaerophilic biodegradation of a mixture of aro-
matic compounds was investigated by varying com-
bined concentrations of oxygen and nitrate. Batch
microcosms were prepared using a liquid enrichment of
aquifer bacteria as inocula; a mixture of benzene, tolu-
ene, ethylbenzene, m-xylene, naphthalene, and phen-
anthrene as substrate; and oxygen and nitrate as elec-
tron acceptors.
The results of this study indicated that the level of oxy-
gen had a significant affect on the extent of biodegrada-
tion of most of the aromatic hydrocarbons. Analysis of
the consumption of electron acceptors indicated that
both nitrate and oxygen acted as electron acceptors
during biodegradation of the mixture of aromatic hydro-
carbons. Denitrification may be inhibited by oxygen lev-
els above 1 mg/L (1). In this study, toluene and
naphthalene biodegradation was favored at microaero-
philic oxygen levels between 1.5 and 2 mg/L. The data
(measurements of oxygen and nitrate not shown here)
suggested that oxygen and nitrate were used sequen-
tially to biodegrade naphthalene and toluene, respec-
tively (i.e., denitrification was inhibited until oxygen was
depleted)(Table 1).
At lower levels of oxygen (0.5 to 1 mg/L), toluene and
ethylbenzene biodegradation was favored. The mecha-
nism for biodegradation of toluene and ethylbenzene at
very low oxygen levels (less than or equal to 1 mg/L)
Table 1. Aromatic Hydrocarbons Degraded Under the
Various Combinations of Oxygen and Nitrate
Investigated (degradation is removal greater than or
equal to 10 percent relative to killed controls)
Oxygen (mg/L)
Nitrate
(mg/L)
10
50
150
400
0
T
T
T
T
0.
T,
T,
T,
T,
,5
E
E
E
E
1 1
T,
T, E T,
T, E T,
T, E T,
.5
N
N
N
N
2
B,
T, N B,
T, N B,
T, N B,
T,
T,
T,
T,
E,
E,
E,
E,
8
m-X,
m-X,
m-X,
m-X,
N,
N,
N,
N,
P
P
P
P
B = benzene, T = toluene, E = ethylbenzene, m-X = m-xylene,
N = naphthalene, and P = phenanthrene
24
-------
Days
Figure 1. Conversion of benzene to intermediates, cells, and
carbon dioxide.
may be quite different than at higher levels (greater than
or equal to 1.5 mg/L). The data indicated that at low
levels (less than or equal to 1 mg/L) of oxygen, nitrate
played a role in biodegradation of both toluene and
ethylbenzene. Evidence for simultaneous utilization of
nitrate and oxygen has been documented (2). Oxygen
levels below 1 mg/L may not inhibit denitrification and
may actually be beneficial by increasing cell numbers
(3, 4).
Benzene was recalcitrant under denitrifying and mi-
croaerophilic conditions. Extensive benzene mineraliza-
tion, however, was observed under aerobic conditions
(Figure 1). Although the majority of benzene biodegra-
dation occurred in the presence of oxygen, partial trans-
formation of the parent compound to intermediates and
carbon dioxide was observed in the absence of oxygen.
Microaerophilic Biodegradation in the
Presence of Sediments
A laboratory study of the stoichiometry of microaero-
philic biodegradation in the presence of sediments is
being conducted. Aquifer sediments may affect the
stoichiometry of microaerophilic biodegradation by ex-
erting an additional oxygen demand from natural or-
ganic matter or reduced metals, and by providing sur-
faces for microbial attachment that are not present in
sediment-free microcosms. In this study, aquifer sedi-
ments were used as inocula in microcosms instead of a
liquid enrichment of aquifer bacteria. Initial results indi-
cate that toluene biodegradation under denitrifying con-
ditions occurs after a significant lag time in microcosms
containing sediment as inocula when compared with
denitrification in liquid enrichment microcosms. This de-
lay in denitrifying activity is likely due to the development
of a sufficient denitrifying population. No biodegradation
of aromatic compounds was observed under microaero-
philic conditions (microaerophilic oxygen consumed by
sediment demands). Biodegradation under aerobic con-
ditions (7 mg/L), however, exceeded what was observed
in the liquid enrichment microcosms. Studies of sedi-
ment microcosms and columns are ongoing.
Acknowledgment
This research was made possible by the generous sup-
port of the U.S. Environmental Protection Agency
Robert S. Kerr Environmental Research Laboratory
(Project CR-821907).
References
1. Kuhn, E.P., J. Zeyer, P. Eicher, and R.P. Schwarzenbach. 1988.
Anaerobic degradation of alkylated benzenes in denitrifying labo-
ratory columns. Appl. Environ. Microbiol. 54(2):490-496.
2. Ottow, J.C.G., and W. Fabig. 1985. Influence of oxygen aeration
on denitrification and redox level in different batch cultures. In:
Caldwell, D.E., J.A. Brierly, and C.L. Brierly, eds. Planetary ecol-
ogy. New York: Van Nostrad Reinhold. pp. 427-440.
3. Hutchins, S.R. 1991. Optimizing BTEX biodegradation conditions.
Environ. Toxicol. Chem. 10:1,437-1,448.
4. Su, J., and D. Kafkewitz. 1994. Utilization of toluene and xylenes
by a nitrate reducing strain of Pseudomonas maltophilia under low
oxygen and anoxic conditions. FEMS Microbiol. Ecol. 15:249-258.
25
-------
Nutrient Transport in a Sandy Beach
Brian A. Wrenn, Makram T. Suidan, B. Loye Eberhardt, and Gregory J. Wilson
Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, Ohio
Kevin L. Strohmeier
Environmental Technologies & Solutions, Inc., Covington, Kentucky
Albert D. Venosa
Risk Reduction Engineering Laboratory, U.S. Environmental Protection Agency, Cincinnati, Ohio
Introduction
Bioremediation of beaches that are contaminated with
crude oil is expected to be limited by the availability of
nutrients. Addition of nutrients, such as nitrogen and
phosphorous, has been shown to stimulate the rate of
crude oil biodegradation in the beach environment (1,
2). For bioremediation to be effective, the nutrient con-
centration in contact with the oil-contaminated beach
material must be high enough to allow hydrocarbon-de-
grading bacteria to grow at their maximum rates.
Oil often strands in the intertidal zone (3). When this
occurs, washout will probably dominate the nutrient resi-
dence time in the bioremediation zone. The periodic
flooding that occurs when the tide rises is one of several
important nutrient transport mechanisms. Wave action
can also affect nutrient transport in the intertidal region
of beaches (4), as can the flow of fresh ground water
from inland (5, 6). The objective of this research was to
characterize the nutrient transport rates for a low-energy
sandy beach. The results were used to determine the
nutrient application rate that is required for effective
bioremediation in this type of environment.
Methods
The tracer study was conducted on a long, uniform
stretch of sandy beach south of Slaughter Beach, Dela-
ware, near the southern end of Delaware Bay. The
beach is composed largely of coarse sand with some
gravel, which lies on top of an impermeable peat layer.
The beach has a slope of approximately 11 percent in
the upper intertidal zone.
Eight replicate plots (5 m x 10 m) were established in
the upper intertidal zone. The tops of the plots were
placed approximately at the spring high-tide line. The
plots were divided into two blocks of four plots each. A
water soluble tracer, either lithium or sodium nitrate, was
applied to one plot in each block at each of four different
stages in the tidal cycle (Figure 1): spring tide (full
moon), falling midtide (i.e., as the tide proceeded from
spring to neap), neap tide (last quarter moon), and rising
midtide (i.e., as the tide proceeded from neap to spring).
(Spring tide is the point in the lunar tidal cycle at which
the difference between high and low tide is greatest.
Neap tide is the point at which the difference is small-
est.) Lithium nitrate was used as the tracer for the
O
D
-5
2 -10
-15
neap tide
falling mid-tide rising mid-tide
spring tide
-72 0 72 144 216 288 360 432
time (hrs)
Figure 1. The actual and expected tidal elevations are shown,
with the start times for the four tracer transport ex-
periments. The elevations of the tops and bottoms of
the plots are also shown. All elevations are relative
to a benchmark that was placed behind the beach in
a dune area well above the maximum high tide line.
26
-------
spring- and neap-tide experiments, and sodium nitrate
was used for both midtidal experiments. The tracer was
dissolved in fresh water (the initial concentration was
20 g/L) and applied to the plots with a sprinkler sys-
tem. Tracer was always applied at low tide, and the
first samples were collected shortly afterward.
The spring- and neap-tide plots contained multiport
wells that were used to collect water samples from
discrete depths within the beach. These wells were
placed along a transect through the middle of the plots
at 2.5-m intervals, beginning 2.5 m above the top of the
plots and continuing to 5 m below the bottom of the
plots. The wells had sample ports every 6 in. The top
port of each well was installed 3 in. below the beach
surface. Sand samples were collected from randomly
selected positions within five subsections that were
marked off in each plot. The locations of these subsec-
tions corresponded to the positions of the wells in the
spring and neap-tide plots.
Sand samples were collected as 10-in. cores using a
5-in. auger. Samples that were analyzed for lithium were
kept as two separate 5-in. cores: an upper sample (0 to
5 in.) and a lower sample (5 to 10 in.). The cores taken
for nitrate analysis, on the other hand, were composited
into a single 10-in. (0 to 10 in.) sample. Lithium was
extracted into 1 M ammonium acetate and analyzed by
atomic absorption spectrophotometry (7). Nitrate, which
was extracted and analyzed in the field, was analyzed
by the cadmium reduction autoanalyzer method follow-
ing extraction into 2 M potassium chloride (KCI) (8).
Results and Discussion
The average lithium concentrations in the two 5-in. core
samples for the spring- and neap-tide tracer experi-
ments are shown in Figure 2. The concentrations meas-
ured in samples collected from the five subsections of
each plot were pooled to obtain the plot averages that
are shown in this figure. The plot averages for each of
the replicate plots are shown independently in this fig-
ure. The average lithium concentration was reduced to
zero very rapidly following the spring-tide application
(Figure 2A), but it was washed out more slowly after the
neap-tide application (Figure 2B). Washout of nitrate
following the two midcycle applications behaved simi-
larly; nitrate was washed out quickly following the falling
midtidal application, and it persisted for a relatively long
time following the rising midtidal application (data not
shown). Although the initial lithium concentrations
were higher in the samples from the upper 5-in., it
disappeared more rapidly in this region than it did from the
5- to 10-in. region.
It is clear from Figure 1 that one of the major differences
among the four tracer experiments is the extent to which
the plots were covered with water at high tide. Wave
action also contributed to plot coverage, and the total
90
80
70
60
50
40
30
20
10
100
90
80
70
0-5 inches
5-10 inches
B
0 - 5 inches
time (hrs)
Figure 2. Lithium concentrations in sand samples collected in
the bioremediation zone following tracer application
at spring tide (A) and neap tide (B). The concentra-
tions for the upper (0 to 5 in.) and lower (5 to 10 in.)
samples are plotted separately for each experiment.
coverage is the sum of both effects. The cumulative
effects of plot coverage on nutrient retention are shown
in Figure 3, in which the ratio of the remaining sand
nitrate concentration to its initial concentration is plotted
as a function of the maximum extent of plot coverage
that had occurred before collecting each sample (e.g., if
1.00 1
0.75 -
O
S 0.50 ;
O
0.25
0.00 ' ' ' ' '
0.00
0.25
0.50
0.75
1.00
maximum plot coverage
Figure 3. Relationship between the fraction of nitrate remaining
and the maximum extent to which water had covered
the plots before the samples were collected. Data
from all four tracer experiments are included in this
plot.
27
-------
the first high tide following tracer application covered 75
percent of the plot with water and subsequent high tides
covered only 50 percent, the maximum coverage for all
of the data collected after the first high tide is reported
as 75 percent). Data from all four tracer experiments,
calculated as described for the lithium data plotted in
Figure 2, were used to construct this plot. Figure 3
shows a strong correlation between the maximum ex-
tent of plot coverage and the remaining nitrate concen-
tration, suggesting that nutrient retention in the
bioremediation zone of a sandy beach can be predicted
based solely on the extent of water coverage.
The simplest explanation for the results shown in Figure
3 is that the tracers become diluted by mixing with bulk
seawater when the plots are covered by the rising tide,
and they are washed away when the tide recedes. This
explanation is consistent with a model for nutrient trans-
port in a beach in Prince William Sound, Alaska (9).
Pore-water data, however, show that the tracer move-
ment is predominantly downward into the beach (Figure
4). These data are plot averages, calculated as de-
scribed above, for all pore-water samples collected from
each depth within the plot. Samples collected from wells
outside of the plots were not used to compute these
averages.
Figure 4 shows that nutrient is probably removed from
the bioremediation zone by advective flow through the
porous matrix of the beach, not by mixing with bulk
seawater.
Conclusions
Effective bioremediation requires a sufficient supply of
the growth-limiting substrate to be available to the bac-
teria responsible for biodegradation. For bioremediation
of oil-contaminated beaches, nutrients such as nitrogen
Figure 4.
Average lithium concentrations in pore water col-
lected at discrete depths below the beach surface for
three time points during the spring tide tracer experi-
ment. All samples were collected at high tide.
and phosphorus are expected to be most important (1,
2), and washout is expected to be the dominant nutrient
removal mechanism. In many cases, it will be desirable
to optimize nutrient application rates to minimize costs
and to reduce the opportunity for eutrophication of
adjacent bodies of water. Our data suggest a simple
method for determining the frequency required for
nutrient application.
Nutrient retention in the bioremediation zone of the in-
tertidal region of a sandy beach varies with the lunar
tidal cycle. Our tracers were washed out of the bioreme-
diation zone very quickly when they were applied during
and shortly after the spring tide (when the high tide
reached its maximum elevation), but they persisted
through several tidal cycles when applied around neap
tide, when the lowest high tides occurred. Our data
suggest that the differences in nutrient retention time are
related to the maximum extent to which our experimen-
tal plots were covered by water at high tide. Total cov-
erage appears to be more important than coverage due
to the tide alone. Therefore, wave activity and tidal
elevation must both be considered to determine the
appropriate fertilization frequency. When water-soluble
fertilizers are used, visual inspection of the extent to
which the contaminated area is covered by water
during high tide is a reliable alternative to expensive
and time-consuming chemical analyses for nutrient
concentration.
References
1. Lee, K., G.H. Tremblay, and E.M. Levy. 1993. Bioremediation:
Application of slow release fertilizers on low-energy shorelines. In:
Proceedings of the 1993 Oil Spill Conference, American Petroleum
Institute, Washington, DC. pp. 449-454.
2. Pritchard, PH., and C.F. Costa. 1991. EPA's Alaska oil spill biore-
mediation project. Environ. Sci. Technol. 25:372-379.
3. Payne, J.R., and C.R. Phillips. 1985. Petroleum spills in the marine
environment: The chemistry and formation of water-in-oil emul-
sions and tar balls. Chelsea, Ml: Lewis Publishers.
4. Brown, A.C., and A. McLachlan. 1990. Ecology of sandy shores.
New York, NY: Elsevier.
5. Glover, R.E. 1959. The pattern of fresh-water flow in a coastal
aquifer. J. Geophys. Res. 64:457-459.
6. Nielsen, P. 1990. Tidal dynamics of the water table in beaches.
Water Resources Res. 26:2,127-2,134.
7. American Public Health Association. 1989. Direct air-acetylene
flame method 3111B. In: Standard methods for the examination of
water and wastewater, 17th ed., Washington, DC.
8. Keeney, D.R., and D.W Nelson. 1982. Nitrogeninorganic forms.
In: Page, A.L., R.H. Miller, and D.R. Keeney, eds. Methods of soil
analysis, Part 2. Chemical and microbiological properties. Madi-
son, Wl: American Society of Agronomy, Inc. pp. 643-698.
9. Wise, W.R., O. Guven, F.J. Molz, and S.C. McCutcheon. Nutrient
retention time in a high permeability, oil-fouled beach. J. Environ.
Engin. In press.
28
-------
Bioremediation of Crude Oil Intentionally Released on the
Shoreline of Fowler Beach, Delaware
Albert D. Venosa and John R. Haines
U.S. Environmental Protection Agency, Cincinnati, Ohio
Makram T. Suidan, Brian A. Wrenn, B. Loye Eberhart, Miry am Kadkhodayan, and Edith Holder
Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, Ohio
Kevin L. Strohmeier
Environmental Technologies and Solutions, Inc., Covington, Kentucky
Dennis King
Kingstat Consulting, Fairfield, Ohio
Bennet Anderson
Delaware Department of Natural Resources and Environmental Control, Dover, Delaware
Introduction
A major factor contributing to the equivocal findings of
past field studies (1-5) was that conclusions were usu-
ally based on comparisons between one large treatment
plot and one large control plot. The problem with this
type of experiment is that no replicate plots are estab-
lished to provide a basis for estimating experimental
error. The collection of numerous subsamples from one
treatment plot and one control plot, termed pseudorep-
lication (6), is statistically invalid for drawing inferences
on treatment effects because no experimental error can
be computed. An experiment lacking replication is an
uncontrolled experiment because it does not control for
among-replicate variability inherent in the experimental
material, introduced by the experimenter, or arising from
chance occurrences. To eliminate uncontrollable and
unknown environmental factors that could skew results
in one direction, several replicate plots must be set up
in random fashion on the beach surface. The experi-
mental approach described herein was carefully de-
signed to allow for valid and statistically authentic
comparisons between treatments.
The goals of the study were 1) to obtain sufficient sta-
tistical and scientifically credible evidence to determine
whether bioremediation with inorganic mineral nutrients
or microbial inoculation enhances the removal of crude
oil contaminating mixed sand and gravel beaches, and
2) to compute the rate at which such enhancement takes
place.
Materials and Methods
The plan was to maximize the effectiveness of bioreme-
diation by maintaining a certain level of nitrogen, in the
form of nitrate (agricultural grade sodium nitrate), and
phosphorus, in the form of tripolyphosphate (sodium
tripolyphosphate), in contact with the degrading popula-
tions so that they would be able to grow at their maximal
rates at all times. In a previous study (7), we had shown
that the minimum nitrate-N concentration needed by oil
degraders to grow on hydrocarbons at an accelerated
rate under semicontinuous flow conditions was approxi-
mately 1.5 mg/L. It was also known that, if the incoming
tide completely submerged the plot, the levels of nitrate
in the interstitial pore-water diluted to undetectable limits
(8). Thus, to maximize bioremediation during spring
tides, we reasoned that nutrients would have to be
applied every day. To achieve the target 1.5 mg/L inter-
stitial pore-water concentrations, we assumed a 100-
fold safety factor to account for dilution. The amount of
nitrate-N needed under these circumstances was thus
calculated to be about 55 g/m2, applied once daily to
each plot.
The approach used to assess treatment effects in the
field study was a randomized complete block (RGB)
29
-------
design with repeated measures. Five areas of beach
were selected based on the homogeneity of geomor-
phology within each area. Each area ("block") was large
enough to accommodate four experimental units or test
plots. The blocks were situated in a row on the beach
parallel to the shoreline. Three treatments were tested
on oiled plots: no-nutrient control, water-soluble nutri-
ents, and water-soluble nutrients supplemented with a
natural microbial inoculum from the site. The inoculum
was grown by isolating a mixed culture from the site and
adding it to a 55-gal drum containing Delaware Bay
seawater, the same Bonny Light crude oil, and the same
nutrient mix used on the beach. A fourth treatment, an
unoiled and untreated plot, served as a background
control for microbiological characterization and baseline
bioassays. The four treatments were randomized in
each of the five blocks so that whatever inferences could
be ascertained from the data would be applicable to the
entire beach, not just the test plots.
Results and Discussion
The mean interstitial nitrate-N concentrations measured
over the course of the 14-week investigation were:
0.8 + 0.3 mg/L in the unamended control plots, 6.3 +
2.7 mg/L in the nutrient-treated plots, and 3.5 + 1.7 mg/L in
the inoculum-treated plots. These results indicate that
background nutrient levels on Fowler Beach were high
enough to sustain nearly maximum oil degrader growth.
Figure 1 is a summary of the hopane-normalized alkane
(Figure 1) and aromatic (Figure 2) oil components re-
maining after the first 8 weeks of the study. In regards
to the alkanes, statistical analysis of variance (ANOVA)
showed that the difference between both treated plots
and the control plots were highly significant at Weeks 2,
4, and 8 (p < 0.01) but not at Week 6. Differences
between the nutrient-treated and inoculum-treated plots
were not significant at any time. Clearly, substantial
natural biodegradation was taking place on the control
4 6
time, weeks
10
4 6
time, weeks
10
Figure 2.
Decline in hopane-normalized total PAHs during the
first 8 weeks.
Figure 1. Decline in hopane-normalized total alkanes during
the first 8 weeks.
plots without addition of nutrients. This observation is
consistent with observations made on the background
nutrient levels existent on Fowler Beach. Addition of
nutrients significantly enhanced the natural biodegrada-
tion rates of the alkane fraction, but not to the extent
expected on a less eutrophic beach.
With respect to the aromatic components, results of the
ANOVA revealed no significant differences among any
of the treatments at Weeks 0, 2, 4, and 6, although
substantial biodegradation of the polycyclic aromatic
hydrocarbons (PAHs) occurred on all plots. At Week 8,
statistically significant differences between the treated
and untreated plots were evident. Most of the disappear-
ance occurred among the two- and three-ring PAHs and
the lower alkyl-substituted homologues (data not
shown). The four-ring PAHs began to show evidence of
biodegradation during the eighth week of the study.
These results suggest that biostimulation may not al-
ways be necessary to promote bioremediation if suffi-
cient nutrients are naturally present at a spill site in high
enough concentrations to effect natural cleanup. The
evidence suggests that nutrient application to maintain
a residual nitrate concentration in the interstitial waters
at high enough levels to sustain maximum biodegrada-
tive metabolism resulted in a significant enhancement of
alkane and, to a lesser extent, aromatic biodegradation
over natural attenuation. Bioaugmentation (i.e., supple-
mentation with a bacterial inoculum indigenous to the
area), however, did not appear to result in further en-
hancement.
References
1. Pritchard, P.M., and C.F. Costa. 1991. EPA's Alaska oil spill biore-
mediation project. Environ. Sci. Technol. 25:372-379.
2. Sveum, P., and A. Ladousse. 1989. Biodegradation of oil in the
Arctic: Enhancement by oil-soluble fertilizer application. In: Pro-
ceedings of the 1989 International Oil Spill Conference. Washing-
ton, DC: American Petroleum Institute.
30
-------
3. Sveum, P. and Ladousse, A. 1989. "Biodegration of Oil in the Artie:
Enhancement by Oil-Soluble Fertilizer Application." In: Proceed-
ings of the 1989 International Oil Spill Conference. Washington,
DC: American Petroleum Institute.
4. Rosenburg, E., R. Legmann, A. Kushmaro, R. Taube, R. Adler, and
E.Z. Ron. 1992. Petroleum bioremediationa multiphase prob-
lem. Biodegradation 3:337-350.
5. Lee, K., G.H. Tremblay, and E.M. Levy. 1993. Bioremediation:
Application of slow-release fertilizers on low-energy shorelines. In:
Proceedings of the 1993 International Oil Spill Conference. Wash-
ington, DC: American Petroleum Institute.
6. Hurlbert, S.H. 1984. Pseudoreplication and the design of ecologi-
cal field experiments. Ecol. Monographs 54(2):187-211.
7. Venosa, A.D., J.R. Haines, M.T. Suidan, B.A. Wrenn, K.L.
Strohmeier, B.L. Eberhart, E.L. Holder, and X. Wang. 1994. Re-
search leading to the bioremediation of oil-contaminated beaches.
In: Symposium on Bioremediation of Hazardous Wastes: Re-
search, Development, and Field Evaluations, June 28-30, 1994,
San Francisco, CA. EPA/600/R-94/075. pp. 103-108.
8. Venosa, A.D., M.T. Suidan, B.A. Wrenn, J.R. Haines, K.L.
Strohmeier, E.L. Holder, and B.L. Eberhart. 1994. Nutrient appli-
cation strategies for oil spill bioremediation in the field. In: Twen-
tieth Annual RREL Research Symposium, March 15-17, 1994,
Cincinnati, OH. EPA/600/R-94/011. pp. 139-143.
31
-------
Dynamics of Oil Degradation in Coastal Environments: Effect of Bioremediation
Products and Some Environmental Parameters
Marirosa Molina and Rochelle Araujo
U.S. Environmental Protection Agency, Athens, Georgia
Jennifer R. Bond
DYNCORP, Athens, Georgia
Introduction
Oil extraction, refining, and transshipment are often lo-
cated in coastal regions, putting wetland systems at risk
of exposure to spilled oil and oil products. The inacces-
sibility of wetlands and the fragile nature of those eco-
systems preclude mechanical cleanup of oil, making
bioremediation a preferred option. Moreover, the high
level of indigenous microbial activity suggests a poten-
tial for biodegradation, especially if fertilizer additions
can relieve environmental nutrient limitations.
Bioremediation strategies that have been proposed for
oil spills in wetlands include fertilization, solubilization of
oil, and bioaugmentation with oil-degrading bacteria.
Although bioaugmentation has been demonstrated to be
effective in engineered systems, the ability of introduced
organisms to establish themselves in the complex web
of microbial relationships that characterize wetlands is
questionable. Moreover, supplies of oxygen and nutri-
ents may be insufficient for concurrent degradation of oil
and natural substrates. In this research, we propose a
system for assessing the efficacy of bioremediation in
wetlands and for testing the effectiveness of several
bioremediation products. We also present data on the
dynamics of the bacterial populations and the relative
rates of degradation of natural substrates and oil in
wetlands.
Methods and Materials
Sediment microcosms were constructed using glass col-
umns (10-cm internal diameter, 20-cm length) fitted with
fritted glass supports and filled with homogenized marsh
sediments from Sapelo Island, Georgia. Seawater or
artificial seawaterwas adjusted to a brackish salinity of
20%o and exchanged on a tidal basis. An artificially
weathered Alaska North Slope (ANS) crude oil was
added at low tide to cover the sediment surface to a
depth of 0.5 mm.
Bioremediation products consisted of bacterial prepara-
tions enriched in oil degraders, nutrients, surfactants, or
combinations thereof. Products were applied in a man-
ner consistent with the manufacturer's recommenda-
tions. In addition to commercial products, inorganic
nutrients were added to microcosms to test the potential
response of indigenous bacteria. Ground Spartina al-
terniflora, a salt marsh grass common to the coast of
Georgia, was used as an alternate natural organic sub-
strate. The grass (4.5 g) was added to the sediment
surface in an amount equivalent to a 1-yr standing stock
of aboveground biomass.
After 3 months, the residual hydrocarbons were ex-
tracted and analyzed by gas chromatography/mass
spectrometry (GC/MS). Efficacy was assessed on the
basis of reduced concentrations of specific components
of oil, including straight-chained and branched alkanes
and aromatics, expressed as ratios to conserved inter-
nal markers. Microbial populations were estimated using
modified heterotrophic and oil most probable number
(MPN) techniques (1). Heterotrophic bacteria were also
quantified using standard plate counts.
Deoxyribonucleic acid (DMA) samples were extracted
from each microcosm using a modification of the method
of Tsa and Olson (2). Target DMA was filtered onto
maximum-strength Nytran membranes. Sediment DMA
was loaded onto the membranes in triplicate. The mem-
branes were hybridized with a Pseudomonas23S rRNA
oligoprobe (Group I). Detection was carried out using
Rad-Free Lumi-Phos 530 substrate sheet and exposure
to x-ray film for 3 hr at room temperature. The signal was
quantified using densitometry.
32
-------
Results and Discussion
The alkane fraction of oil was degraded in the presence
or absence of bioremediation products, as indicated by
the absence of C13 and C14 compounds and the greatly
diminished peaks for C13 through C33, including the
branched alkanes, pristane, and phytane. Addition of
inorganic nutrients to sediments containing only indige-
nous bacteria resulted in greater depletion of the alkane
fraction than did additions of products composed of
surfactants or bacterial enrichments. Moreover, in the
presence of surfactants, the extent and range of degra-
dation was less than that in the control treatment, sug-
gesting inhibition of microbial activity.
The aromatic fractions of the oil (Figure 1) were de-
graded to a lesser extent in all treatments than were the
alkane fractions (data not shown), although the rank
order of the treatments was the same. Both the range
of compounds degraded and the extent to which they
were degraded were significantly greater in the treat-
ment containing nutrients than in the presence of the
commercial products.
The abundance of oil-degrading bacteria (Figure 2) was
not consistent with the extent of oil degradation ob-
served. In the absence of bioremediation products, the
number of oil degraders (1.51 x 103 bacteria/g sediment)
after 3 months of experimentation was not appreciably
different from that observed in the initial Sapelo Island
sediment (1.38 x 103). Even at such low numbers, how-
ever, the microbial community was capable of degrading
the alkane fraction of oil, especially when fertilized with
inorganic nutrients. No enhanced degradation resulted
from the addition of a product purported to be en-
riched in oil-degrading bacteria, regardless of the rela-
tively high numbers of total heterotrophs and potential
oil-degrading bacteria that the treatment yielded.
Sediment layer
Ratio to C-2 Chrysenes
D NUTRIENTS
OILDEG.
D CONTROL
D SURFACT.
521 OIL
O O
O O O
Figure 1. Residual aromatics in surface layer of microcosms
after 3 months in the presence of bioremediation
agents.
5.33
Middle
Bottom
3.25
4.92
4.86
4.86
14.92
2468
Log bacteria/g sediment
10
D Nutrients D Added Oil Degraders Control
B Surfactants E3 Sapelo Is. Sed.
Figure 2. Distribution of oil degraders in wetlands microcosms
after 3 months of exposure to bioremediation agents.
The inhibition of oil degradation by the addition of sur-
factant was consistent with the microbial numbers,
which show that, in the presence of the surfactant, the
abundance of oil degraders was less than in the treat-
ments containing nutrients or added oil degraders. The
heterotrophic bacteria, however, were not affected by
the addition of surfactants, as indicated by the numbers
in this treatment being comparable with the numbers
obtained in the presence of added nutrients (data not
shown). The surfactant product could have been used
as a carbon or nutrient source by the heterotrophic
bacteria, thus increasing their numbers. Furthermore,
the surfactant-based product may be inhibitory to the
indigenous oil degraders, which were only enhanced in
the presence of inorganic nutrients.
Enumeration of oil-degrading bacteria gives no informa-
tion about the source of the active bacteria because the
oil-degrading bacteria may have been of sediment or
product origin. DMA hybridizations with 16S rRNA
oligoprobes indicated that Pseudomonas Group I was
significantly reduced in the presence of the microbial
product. This suggests that addition of the bacterial
preparation suppressed some indigenous populations,
including perhaps the indigenous oil-degrading bacteria.
None of the other treatments produced numbers of
33
-------
Pseudomonads that were significantly different from
those observed in the initial sediment; data on other
sediment microbial groups will be forthcoming.
Under nutrient-amended conditions, both alkane and
aromatic fractions were degraded relative to the 521 oil;
however, additions of Spartina alterniflora detritus de-
creased the extent of degradation for both fractions. The
results for aromatic constituents are shown in Figure 3.
Given that microbial activities in wetlands ecosystems
are usually limited by oxygen and nutrients, oil degrada-
tion may compete with the degradation of natural or-
ganic substrates for these substances. In addition, the
hemicellulose fraction of lignocellulose constitutes a
readily degradable carbon source that may compete
with the utilization of petroleum hydrocarbons by the
indigenous microbial community.
In the treatments containing no oil, the number of oil
degraders was constant over the 3-month period (ap-
proximately 6.31 x 10s bacteria/g of sediment), which
indicates that addition of nutrients only does not in-
crease the population of degraders. The two treatments
containing oil produced an increase of oil degraders
during the second month of experiment (Figure 4), which
may correspond to the initial, rapid degradation of the
alkane fraction. By the third month, the oil degraders had
decreased, possibly reflecting the slower rate of degra-
dation of aromatic compounds.
The numbers of heterotrophic bacteria were higher in
the presence of Spartina and oil together than in any of
the other treatments. Although the numbers of oil de-
graders in both treatments containing oil were very simi-
lar, the extent of oil degradation was significantly
different. The diminished oil degradation in the presence
of oil and Spartina suggests that the combination of
substrates may antagonize the activity of indigenous oil
degraders, either directly or by competition for nutrients
and/or oxygen. Given the abundance of Spartina alterni-
flora in coastal regions of the United States, the interac-
tions between oil-degrading and other heterotrophic
bacteria and the impact of natural substrates should not
be overlooked in the design of remediation strategies.
References
1. Brown, E.J., and J.F. Braddock. 1990. Sheen screen, a miniatur-
ized most-probable-number method for enumeration of oil-degrad-
ing microorganisms. Appl. Environ. Microbiol. 56:3,895-3,896.
2. Tsai, Y, and B.H. Olson. 1991. Rapid method for direct extraction
of DMA from soil and sediment. Appl. Environ. Microbiol. 57:1,070-
1,074.
Ratio to C-2 Chrysenes
DOil
D Oil + Spartina
H 521 oil control
^ -C -C p .c :3 ^ ;3 i *--- -i-1 ^ ^ y Q_D_-C
"fl-fl- v^9llllsSSS6o
""^ fificiCLn-n^^ccc
0 ° ° ^ ^ Q- Q- s % £ £
z
i *
OO Q O
T|- C\l CO
666
Figure 3. Residual aromatics in surface layer of microcosms
after 3 months in the presence of a natural organic
substrate.
Sediment + Oil
B Oil degraders EH Heterotrophic bacteria
Sediment + Spartina + Oil
H Oil degraders D Heterotrophic bacteria
Figure 4. Heterotrophic bacteria and oil degraders in surface
layer of microcosms after 3 months of exposure to
oil.
34
-------
Progress Toward Verification of Intrinsic Cobioremediation of
Chlorinated Aliphatics
Mark Henry
Michigan Department of Natural Resources, Oscoda, Michigan
A plume consisting of chlorinated aliphatic and aromatic
hydrocarbons mixed with JP-4 (benzene, toluene, ethyl-
benzene, and the xylenes; BTEX) from a former Air
Force fire-fighting training area shows evidence of con-
tinual natural bioremediation. This site is being charac-
terized and monitored by the National Center for
Integrated Bioremediation Research and Development
(NCIBRD) at Wurtsmith Air Force Base, a decommis-
sioned installation located in lower northeast Michigan.
Wurtsmith is bounded by the Au Sable Riverto the south
and west and by Van Etten Lake to the north and east.
The base sits on a 20-m bed of homogeneous glacial
alluvial sand and gravel aquifer underlain by a thick clay
aquitard. The average ground-water depth in the study
area is 6 m. The hydraulic conductivity has been re-
ported to be 3e-3 cm/sec (v = 2.6 m/day) at the site,
which has resulted in a narrow 50 m x 300+ m plume.
The plume is monitored on a quarterly basis through the
use of dedicated bladder pumps installed in 37 monitor-
ing wells at the site. Local ground-water elevations are
continually recorded by a datalogging network of re-
corder wells.
The site has been characterized through 2 years of
quarterly sampling of the well and pieziometer network,
as well as direct analysis of continuous cores (gathered
by resident Geoprobe sampling equipment) across the
plume. This information is supplemented by a weekly
monitoring of the vertical temperature profile of the site
and periodic soil gas profiles.
In general, the site has a large amount of residual
fuel/solvent residing near the interface of the water
column and capillary fringe, extending at least 125 m
from the source. Soil gas measurements in the vadose
zone near the source indicate that the interstices con-
tain approximately 65 percent methane, 30 percent
carbon dioxide, ppb level hydrogen sulfide and nitrogen,
and virtually no oxygen. The ground water beneath the
free product has almost no dissolved oxygen and has
depressed redox potential, increased electrolytic con-
ductivity, depressed pH levels, and increased concen-
trations of reduced iron. BTEX levels steadily decrease
over the length of the plume. While perchloroethylene
(PCE) and trichloroethylene (TCE) levels are significant
(greaterthan 1,500 mg/kg) in the solids, only trace levels
are found in the ground water. As the dissolved plume
moves downgradient, the predominant chlorinated spe-
cies are cis-1,2-dichloroethylene and vinyl chloride.
On the fringe of the contamination, BTEX metabolites
such as m,p-toluic acid and salicylic acid have been
identified.
The disappearance of TCE, PCE, and BTEX and the
appearance of bacterial metabolites of these com-
pounds over the length of the plume suggest that these
contaminants are being bioattenuated within the same
plume. Changes in redox potential, temperature, and pH
support this assumption. It remains to be seen whether
or not these processes are interrelated, and are perhaps
influenced by bacterially mediated iron or manganese
reduction.
35
-------
Phytoremediation of Petroleum-Contaminated Soil:
Laboratory, Greenhouse, and Field Studies
M. Katharine Banks
Department of Civil Engineering, Kansas State University, Manhattan, Kansas
A. Paul Schwab
Department of Agronomy, Kansas State University, Manhattan, Kansas
Common environmental problems associated with the
pumping and refining of crude oil are the disposal of
petroleum sludge and pipeline leaks. Contaminants are
often treated by incorporation into the soil. If the soil is
frequently tilled and fertilized, soil microorganisms will
be stimulated and organic contaminants biodegraded.
Unfortunately, the biodegradation rate of more recalci-
trant and potentially toxic contaminants, such as the
polycyclic aromatic hydrocarbons (PAHs), is rapid at first
but declines quickly. Biodegradation of these com-
pounds is limited by their strong adsorption potential and
low solubility.
Recent research suggests that vegetation may play an
important role in the biodegradation of toxic organic
chemicals in soil. The establishment of vegetation on
hazardous waste sites may be an economical, effective,
low-maintenance approach to waste remediation and
stabilization. The use of plants for remediation may be
especially appropriate for soils contaminated by organic
chemicals to depths of less than 2 m. The beneficial
effects of vegetation on the biodegradation of hazardous
organics are two-fold: organic contaminants may be
taken up by the plant and accumulated, metabolized, or
volatilized; and the rhizosphere microflora may acceler-
ate biodegradation of the contaminants.
Completed greenhouse studies indicate that vegetative
remediation is a feasible method for cleanup of surface
soil contaminated with petroleum products. Afield dem-
onstration is necessary, however, to exhibit this new
technology to the industrial community. In this project,
several petroleum-contaminated field sites have been
chosen in collaboration with three industrial partners.
These sites have been thoroughly characterized for
chemical properties, physical properties, and initial con-
taminant concentrations. A variety of plant species have
been established on the sites, including warm and cool
season grasses and legumes. Soil analyses for the
target compounds overtime indicate that the interaction
between plants and rhizosphere microflora significantly
enhances remediation of the contaminated soils. Con-
tinued monitoring will allow us to assess the efficiency
and applicability of this remediation approach.
37
-------
Section 3
Performance Evaluation
In an effort to evaluate the performance of various bioremediation technologies, researchers assess
the extent and rate of cleanup for particular bioremediation methods. They also study the environ-
mental fate and effects of compounds and their byproducts because remediation efforts at a
contaminated site can produce intermediate compounds that can themselves be hazardous. Thus,
another important aspect of performance evaluation projects involves assessing the risk of potential
health effects and identifying bioremediation approaches that best protect public health.
To this end, EPA's National Health and Environmental Effects Research Laboratory (NHEERL)
developed an integrated program to address 1) the toxicity of known hazardous waste site contami-
nants, their natural breakdown products, and their bioremediation products; 2) the development of
methods to screen microorganisms for potential adverse health effects; 3) the potential for adverse
effects when chemical/chemical chemical/microorganism interactions occur; and 4) the develop-
ment of methods to better extrapolate toxicological bioassay results to the understanding of potential
human toxicity.
Two performance evaluation papers were presented at the symposium studying the toxicity of
hazardous waste mixtures before and after bioremediation. One group of microorganisms (Pseudo-
monas aeruginosa and Phanerochaete chrysosporium) were found to be unable to significantly
degrade or eliminate the mutagenic activity of a mixture of several aromatic chemicals.
39
-------
Detoxification of Model Compounds and Complex Waste Mixtures Using
Indigenous and Enriched Microbial Cultures
K.C. Donnelly, Jeannine L. Capizzi, Ling-Yu He, and Henry J. Huebner
Texas A&M University, College Station, Texas
Introduction
Biological degradation of organic compounds may be
considered an economical tool for remediating hazard-
ous waste contaminated environments. While some en-
vironments may be too severely contaminated for initial
in situ treatment to be effective, most contaminated
media will use some form of biological degradation in
the final treatment phases. Traditional studies monitor
the success of bioremediation by the loss of a parent
compound or class of chemicals. Prior studies, however,
have observed both an increase and decrease in toxicity
as a result of degradation (1-3). This study uses both
genotoxicity bioassays and chemical analysis to monitor
the efficacy of biological degradation of model com-
pounds and a complex chemical mixture in soils.
Methods
A Weswood sandy loam soil was allotted in 7-kg portions
into 40 stainless steel boxes. Four soil treatment cate-
gories included: 1) no amendment; 2) amendment with
three model chemicals, i.e., benzo(a)pyrene (BAP),
2,4,6-trinitrotoluene (TNT), and pentachlorophenol
(PCP); 3) amendment with a wood preserving bottom
sediment waste (WPW); and 4) amendment with both
the model chemicals and WPW. Additionally, each of the
four soil treatment categories were divided into three-
box subunits and inoculated with either an indigenous
culture, a Pseudomonas aeruginosa culture, or a
Phanerochaete chrysosporium culture. Thus, a total of
40 boxes were prepared, including the soil with no
chemical amendment and the soil with indigenous, bac-
teria, or fungi inocula (Nl, NB, NF); a wood preserving
waste amended soil with each of the three inocula (Wl,
WB, WF); soil amended with the model chemicals and
each of the microbial inocula (Ml, MB, MF); soil
amended with both the model chemicals and wood pre-
serving waste and each inolula (Cl, CB, CF); and steril-
ized soil receiving each of the chemical amendments
(St-N, St-W, St-M, St-C). The soil boxes were stored in
a greenhouse under constant temperature and humidity,
and monitored to maintain moisture content. Samples
were collected on Days 0, 90, 180, and 360 posttreat-
ment. The solvent was extracted with methylene chlo-
ride and methanol, then evaluated for genotoxicity in the
Sa/mone//a/microsome assay. Benzo(a)pyrene and
pentachlorophenol were quantified using gas chroma-
tography (GC), whereas TNT was quantified using high-
performance liquid chromatography (HPLC).
Results
The data presented in Table 1 describe the quantity of
solvent-extractable organics recovered from the soil
treatments using methylene chloride and methanol. The
data indicate that smaller quantities of organics were
recovered from the treatments with indigenous organ-
isms on 0 and 30-day sampling periods. The soil treat-
ment with the wood preserving waste, model chemicals
(Cl), and indigenous organisms was not appreciably
changed during the 360-day incubation. The waste and
bacteria treatment (WB) yielded an average of 32 mg/g
solvent-extractable organics on Day 0 and 20 mg/g on
Day 360. Overall, none of the treatments amended with
wood-preserving waste displayed an appreciable reduc-
tion over the initial 360 days of treatment.
The extract of the unamended Weswood soil induced an
average negative mutagenic response at Days 0 and 90.
In general, treatments with the model chemicals, wood
preserving waste, or combined treatments induced an
average positive mutagenic response at all time points.
With metabolic activation, the methylene chloride ex-
tract of the MB treatment induced an average of 167 net
TA98 revertants on Day 0 and 122 net revertants on Day
90. A slight increase in the mutagenic response was
observed in the majority of extracts of samples collected
on Day 180. The methylene chloride extracts of the
waste amended soils (Wl, WB, and WF) collected on
Day 180 induced responses with activation that ranged
from 46 net revertants (Wl) to 63 net revertants per mg
41
-------
Table 1. Solvent Extractable Organics Recovered From Soil Treatments (mg residue/g soil)
Time (days after treatment)
Treatment3 0
30
60
90
180
Wl = wood preserving waste with indigenous bacteria.
WB = wood preserving waste with P. aeruginosa.
WF = wood preserving waste with P. chrysosporium.
Cl = combined model chemicals and wood preserving waste with indigenous bacteria.
CB = combined model chemicals and wood preserving waste with P. aeruginosa.
CF = combined model chemicals and wood preserving waste with P. chrysosporium.
StC = sterilized soil with combined model chemicals and wood preserving waste.
b Values are the mean of three replicates ± standard deviation.
360
Wl
WB
WF
Cl
CB
CF
StC
19.80 ±3.2b
32.25 ±1.1
21 .04 ±0.9
19.1 7 ±3.5
26.45 ±1.6
23.36 ± 2.0
27.05
23.18 ± 1.6
27.25 ± 2.5
28.35 ± 2.0
29.26 ± 1 .7
25.42 ± 0.9
31.51 ±0.6
30.31
29.45 ± 2.2
26. 10 ±2.8
30.84 ± 4.6
28. 16 ±3.2
26.35 ± 1 .0
29.90 ± 1 .9
32.88
24.18 ±0.8
23.81 ± 1.2
23.79 ± 2.0
26.67 ± 2.7
23.54 ± 0.4
26.1 7 ±0.2
36.60
20.28 ±1.8
25.24 ± 3.8
21.19± 1.7
24.06 ± 0.7
23.97 ± 1 .8
21.71 ±2.6
26.95
18.25 ±0.8
19.78 ±1.6
18.14± 1.0
20.21 ± 0.4
19.62 ±1.1
19.57 ±0.9
28.78
of residue (WB). The methylene chloride extracts of
the combined waste and model chemical soils (Cl,
CB, and CF) collected on Day 180 induced responses
that ranged from 71 net revertants (CB) to 76 net
revertants per mg of residue (CF and Cl). Most of the
methanol extracts induced a stronger response than
was observed in the methylene chloride extract. At
least one solvent-extractable fraction from each of the
soils collected on Day 180 induced a positive mu-
tagenic response.
Discussion
Approximately 1 year after the application of both wood
preserving waste and model chemicals to a Weswood
soil, the level of solvent-extractable organics was not
appreciably reduced. Although chemical analysis sug-
gested reductions in the model chemicals in some of the
treatments, contaminant concentrations were all detect-
able after 360 days. The data do suggest that the pres-
ence of oils and hydrophobic chemicals in the waste
amended soil may have limited the availability of chemi-
cals for microbial degradation. Changes in mutagenicity
over the initial year of the study indicate that the bioavail-
ability of chemicals was increased in some treatments.
These results indicate that 360 days of treatment was
insufficient to eliminate the mutagenic activity of soils
amended with either the model chemicals or the wood
preserving waste. Data collection will continue for an
additional 720 days to monitor the chemical and toxico-
logical changes associated with each treatment.
References
1. Aprill, W, R.C. Sims, J.L. Sims, and J.E. Matthews. 1990. Assess-
ing detoxification and degradation of wood preserving and petro-
leum wastes in contaminated soil. Waste Manag. and Res.
8:45-65.
2. George, S.E., D.A. Whitehouse, and L.D. Claxton. 1992. Gentoxic-
ity of 2,3,4-trichlorophenoxyacetic acid biodegradation products in
the Salmonella reversion and lamba prophage-induction bioas-
says. Environ. Toxicol. Chem. 116:733-740.
3. Donnelly, K.C., P. Davol, K.W. Brown, M. Estiri, and J.C. Thomas.
1987. Mutagenic activity of two soils amended with a wood-
preserving waste. Environ. Sci. Technol. 21:57-64.
42
-------
Assessing the Genotoxicity of Complex Waste Mixtures
Larry D. Claxton, Virginia S. Houk, Sarah H. Warren, Thomas J. Hughes, and Susan E. George
Environmental Carcinogenesis Division, National Health and Environmental Effects Research
Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina
Introduction
Some chemicals associated with environmental spills
and hazardous waste sites can induce permanent ge-
netic alterations in all organisms. These changes, called
mutations, can have deleterious effects on individuals
and their descendants. In human populations, mutations
are known to increase the incidence of cancer and
genetic diseases and may play a role in numerous other
diseases. In nonhuman biota, mutations may alter the
balance of an ecosystem or change the virulence of
pathogens.
In spite of 20 years of research concerning environ-
mental mutagenesis, current knowledge concerning the
identity and effects of most environmental mutagens
(including genotoxic carcinogens) is limited, and data on
their effects on other living organisms and the ecosys-
tem are practically nonexistent. Because their mode of
action theoretically does not depend on a threshold limit,
genotoxicants are a class of toxic substances that
should be examined when found in even low concentra-
tions in the environment.
The major objective of this paper is to review the U.S.
Environmental Protection Agency's (EPA's) research in
the area of genetic toxicology as it relates to bioreme-
diation. Readers can referto a completed manuscript (1)
for more in-depth information, examples of studies and
data, and a discussion of the regulatory context.
Research Targeted by the Risk
Assessment Paradigm
The complexity of environmental situations complicates
the monitoring, evaluation, and risk assessment of haz-
ardous waste sites and spills. The composition of pollut-
ant mixtures may consist of a few or thousands of
individual compounds, and remediation efforts produce
additional components that add to the complexity of
evaluating risk. EPA research targets efforts that
enhance the risk assessment process. Research, there-
fore, can be placed into the four categories associated
with the risk assessment process: hazard identification,
exposure assessment, dose-response assessment, and
risk assessment modeling and calculation methods.
Hazard Identification Research
Hazard identification research strengthens the risk as-
sessment process by developing methods that are more
reliable and more cost effective, and, when possible,
give some sense of relative toxicity when identifying
toxic agents in the environment. This research involves
the development of new assay methods (e.g., the spiral
mutagenicity assay), the enhancement of existing as-
says (e.g., the modification of the prophage assay), the
integration of bioassay and analytical chemistry meth-
ods, and the modification of methods for use with com-
plex environmental mixtures rather than with single
environmental chemicals.
Exposure Assessment Research
Generally, a hazardous substance must come into con-
tact with a specific biological component (e.g., deoxyri-
bonucleic acid) before a toxic response can occur.
Measuring or estimating the potential level of contact
between a toxicant and its reactive target (or a surrogate
for that target) is exposure assessment. Exposure as-
sessment is done on a population and not an individual
basis. Quantitative toxicological assays and analytical
chemistry methods have been developed to determine
relative amounts of toxicity potential existing in different
environmental (e.g., treated versus untreated) sites, to
determine the relative bioavailability of environmental
toxicants and their metabolites, and to follow the change
in environmental concentrations of toxicants over time.
In collaboration with other EPA laboratories, the National
Health and Environmental Effects Research Labora-
tory (NHEERL) has demonstrated the usefulness of
bioassays to enhance and calibrate exposure assess-
ment methods.
43
-------
Dose-Response Assessment
Dose-response assessments quantitate the potency of
an environmental agent(s) for a specific outcome. Al-
though the NHEERL bioremediation program does not
support whole animal carcinogenicity and mutagenicity
studies, the Environmental Carcinogenesis Division
(ECD) of NHEERL does develop, in support of other
programs, information relevant for bioremediation haz-
ard and risk assessments. In addition, ECD has devel-
oped a statistical software system for analyzing, in a
quantitative manner, short-term mutagenicity tests.
Risk Assessment Modeling Research
When information is limited or only a general charac-
terization is required, a qualitative assessment can be
made; however, there is often a need to be as quantita-
tive as possible in risk assessments. ECD, therefore, not
only develops quantitative data but also examines
mechanisms and developing models to help scientists
understand to what extent the toxic effect is relevant to
human populations. Mechanistic research, for example,
may demonstrate that rodents respond to a toxin in a
manner not relevant to humans, alleviating the need for
a quantitative human risk assessment. Likewise, mecha-
nistic research may demonstrate whether or not syner-
gism should be considered when evaluating a
multichemical site.
The purpose of the presentation given at this conference
is to catalog NHEERL bioremediation research, to show
its relevance, and to provide examples of how the infor-
mation obtained will be useful in future assessments.
Reference
1. Claxton, L.D., V.S. Houk, and S.E. George. 1995. Integration of
complex mixture toxicity and microbiological analyses for envi-
ronmental remediation research. In: de Serres, F.J., and A.
Bloom, eds. Ecotoxicology and human health. Boca Raton, FL:
Lewis Publishers. In press.
44
-------
Section 4
Pilot-Scale Research
By studying bioremediation processes under actual site conditions on a small scale, researchers
can gather critical information on issues such as operation, control, and management of residuals
and emissions before moving to full-scale research. This is a critical intermediate step in which the
success of laboratory experiments are further tested in an expanded but controlled setting.
Pilot-scale evaluations covered many different tools for bioremediation, including biofilters, compost
piles, and slurry bioreactors. One paper studied the optimization of biofilters for use in removing
volatile organic contaminants from the air, while another sought to establish the optimal operating
conditions for compost media. Finally, a third paper found that many organic contaminants become
concentrated in the foam produced in slurry bioreactors and suggested that this effect could be
utilized to isolate contaminants from the rest of the slurry.
Two papers focused on the use of combinations of aerobic and anaerobic conditions to degrade
recalcitrant chlorinated wastes. Work was presented on biofiltration with gel beads that provide an
oxic environment at their surface and an anoxic environment at their centers. The authors found
that such filtration was highly effective at degrading trichloroethylene (TCE). Asystem was evaluated
for land treatment that involved switching between anaerobic and aerobic conditions.
Research continued on the integration of physical and biological processes to clean up organic
wastes. One paper showed how wood preserving wastes can be first removed from their soil matrix
by washing and distillation processes and then degraded in a bioreactor.
Two presentations covered research into optimizing the biodegradation of TCE. The first was the
result of the study of a variety of parameters that might affect the success of field-scale TCE
treatment. The next characterized several potentially useful bacteria isolated from a biofiltration
device.
Two papers investigated the use of microorganisms to clean up inorganic wastes. One technique
mentioned involves the biological reduction of sulfates into sulfides. This technique helps to remove
metal cations from solution. The other technique involves the biological reduction of metal cations
into less harmful oxidation states. The authors of the final paper tested the ability of microorganisms
to reduce chromium (VI).
Poster presentations described several pilot-scale research projects. The poster projects within this
category involved the testing of a compost bioreactor, a biofilm-electrode reactor, and a laminar-type
flow reactor.
45
-------
In Situ Bioremediation of Trichloroethylene With Burkholderia Cepacia PR1:
Analysis of Parameters for Establishing a Treatment Zone
Richard A. Snyder
Center for Environmental Diagnostics and Bioremediation, University of West Florida,
Pensacola, Florida
M. James Hendry and John R. Lawrence
National Hydrology Research Institute, Environment Canada, Saskatoon, Canada
Introduction
The use of chloroethenes, including trichloroethylene
(TCE), has led to an extensive contamination of ground-
water resources in the United States. In situ bioremedia-
tion of this contaminant is being pursued with the
aerobic microorganism Burkholderia cepacia (formerly
Pseudomonas cepacia) G4 (1). Mutant strains G4
PR12s and PR131 constitutively produce a toluene ortho-
monooxygenase (TOM) that mineralizes TCE. The gene
for this enzyme is located on the degradative plasmid
TOM of B. cepacia (2). The constitutive expression of
this gene is the result of secondary transposition follow-
ing Tn5 insertion mutagenesis, which also confers
kanamycin (km) resistance to these bacteria. PR12s
contains a single entire Tn5 in the chromosome and an
IS50R of Tn5 in the plasmid. PR131 bears an entire Tn5
and an IS50R in the TOM31c plasmid. The IS50 elements
are at nearly the same locations and are thought to be
responsible for the constitutive expression of torn.
This project involves ground-water flow control with
sealed sheet piling to funnel ground water through a
narrow gate area in which treatment technologies can
be applied. Parameters for the establishment of a bio-
logical treatment zone with PR1 have been investigated
in the laboratory. Transport and survival characteristics
of the bacterium have been examined in ground water
and sediment from a targeted release site (the aquifer
under the Canadian Armed Forces Base, Borden, On-
tario). Monitoring techniques have been developed for
tracking PR1 populations in the treatment zone and
determining the extent of dispersal and survival beyond
the treatment zone.
The functional integration of non-native bacteria in natu-
ral microbial communities and maintenance of bacterial
populations above normal environmental background
concentrations provides a challenge for both microbial
ecologists and applied microbiologists. Monitoring of
population dynamics and trophic interactions is critical
for successful bioaugmentation applications. Risk as-
sessment associated with uncontained biotechnological
introductions also requires careful monitoring of the
survival and dispersal of released microorganisms and
altered genes. Although releases of non-native or re-
combinant bacteria have not been reported to result in
adverse environmental effects to date, there is a respon-
sibility to ensure that released microorganisms will be
constrained by the selective pressures of the target
environment.
Tracking
Selective media provide a first step in tracking the or-
ganism. Phenol, o-cresol, and phthalic acid combined
with kanamycin (km) were tested for growth of PR1 and
isolates from the Borden aquifer. Growth of PR1 was
optimal on 20 mM phthalate medium. In aquifer sedi-
ment slurries, numbers of native bacteria isolated on this
medium range from 0 to less than 5 x 105 colony forming
units (CPU) and direct epifluorescence microscopy
counts of bacterivorous protists are less than 8 x 102
ml"1 in unamended incubations.
A monoclonal antibody (mab) specific to G4 lipopolysac-
charide (IPS) (3) has been used both in direct im-
munofluorescent counts and to confirm CPU of PR1 on
phthalate plates. Details of the production and testing of
this mab can be found in Winkler et al. (3).
Polymerase chain reaction (PCR) primers targeting the
Tn5 insertion junctions have been developed and
tested. The primers were designed based on the
47
-------
assumption that the insertion points would be unique for
PR1. The PCR products resulting from this primer set
are of slightly different size for PR12s and PR131, reflect-
ing the slightly shifted location of the IS50R in the plas-
mid in these two strains. Attempts to extract indigenous
bacterial deoxyribonucleic acid (DMA) from Borden
aquifer sediments and to amplify product from these
extractions with the PR1 primers have failed. We have
demonstrated that PCR amplification with these primers
will work with whole, intact cells added directly to the
PCR reaction. The current limit of detection for positive
amplification from cell suspensions (equivalent to pore
water samples) is 1 x 103 cells/mL"1, but from sediment
slurries the detection limit drops an order of magnitude.
These results indicate that this assay will be most useful
for large volumes of pore water samples collected on
Sterivex filters (Millipore Corp.) for extraction of DMA (4).
Potential host range for the TOM plasmid from PR131 in
24 random isolates from Borden aquifer sediments (R2A
medium; 5) was assessed by direct filter matings. Over-
night cultures of PR131 grown in lactate medium and
aquifer isolates in R2A medium were pelleted and resus-
pended in R2A medium to an optical density of 1.0 @
600 nm. One milliliter of donor and recipient were mixed
and filtered on a 0.2 u,m pore-size polycarbonate filter.
Filters were placed on R2A plates and incubated over-
night at 30°C. Filters were then removed and vortexed
in R2A broth and suspensions plated on selective media
predetermined by antibiotic screening. Out of 24 iso-
lates, 42 percent were positive for TOM31c transfer by
PCR detection, and 80 percent of these were positive
for torn activity as indicated by a positive transformation
of trifluoromethyl phenol (TFMP) (1) and mineralization
of TCE in standing cell assays. These data indicate a
wide potential host range within the target environment
for TOM, highlighting the need to ensure tracking capa-
bility for both the organism to be introduced and its
associated genetic elements. These native bacterial
strains bearing functional TOM plasmids may also be
better adapted for bioremediation application in the tar-
get environment than PR1.
Analysis of PR1 Transport in Borden
Aquifer Material
Transport characteristics of PR1 in Borden sand are
important in determining the retardation and dispersal of
cells within a treatment zone and in the downstream
aquifer. Cell retardation is important to allow contami-
nated water to flow past the inoculated populations of
PR1. A series of experiments was carried out at three
scales: 3.8-cm, 10-cm, and 40-cm length columns
packed with sterile and nonsterile Borden sand and
commercially available silica sand. Artificial ground
water (AGW) was used as the solute. All columns were
packed understanding water to minimize air entrapment
and tamped with a glass rod to attain a uniform bulk
density of 1.8 g/cm"3 and a porosity of 0.4. Ground-water
velocity was set to the approximate velocity of water in
the target site. Bacterial suspensions were pulsed
through for 1 void volume, and the effluents were col-
lected in sterile vials with a chromatography fraction
collector. Numbers of PR1 were determined by plating
on selective media. Chloride (Cl) ion as a conservative
tracerwas measured with Ag/AgCI electrodes calibrated
against Cl ion analysis by ion chromatography.
Both the Cl tracer and bacterial breakthrough curves
displayed similar patterns between columns, indicting
good replication. All bacterial breakthrough curves ex-
hibited a notable pulse of bacteria corresponding to the
breakthrough of the Cl tracer. Reduced peak concentra-
tions of bacteria relative to breakthrough concentrations
of Cl (greaterthan 99 percent) indicate irreversible sorp-
tion of PR1 onto the geologic media. Well-defined tailing
was also observed overthe duration of the experiments,
indicating reversible sorption of PR1 to the geologic
media. Peak heights and tailing were three orders of
magnitude lower in Borden material than silica sand,
possibly due to clays and iron coatings on Borden
sands. Breakthrough of PR1 was not affected by sterili-
zation of the sediment orthe addition of a cotransported
bacterium. These results were integrated with predation
loss rates of PR1 added to Borden sediment in slurry
microcosms to develop a predictive model with both
physical and biological parameters for the transport and
fate of the organism within the Borden aquifer.
Survival of PR1 in Aquifer Microcosms
Much of the target environment consists of anaerobic
saturated sediments. Survival of the organism beyond
an oxygenated target zone will likely depend on its ability
to withstand conditions of little or no oxygen. Plate
counts of a suspension of cells with Nitrogen (N2) gas
flushed through the head space indicate little effect on
PR131 viability through 6 days. Afterthis point, culturable
numbers drop precipitously but maintained a low popu-
lation level through 25 days of anaerobic conditions.
No PR1-specific viruses could be isolated from the tar-
get environment. Samples of aquifer sediment were
incubated with growing PR1 cells as an enrichment, and
supernatants were tested on PR1 overlay plates to look
for cleared viral plaques.
Survival of PR1 above 1 x 107 cells/mL"1 is of interest in
establishing needed inoculation densities for effective
bioremediation (6). Survival below this concentration
and long-term integration of the organism into the native
microbial community was of interest for risk assess-
ment. A series of experiments has been conducted in
ground water, sediment slurries, and flow columns to
examine population dynamics of PR1 and the bac-
terivorous protists from the Borden aquifer that are the
primary vector for loss of PR1. Use of a monoclonal
48
-------
antibody (mab) to PR1 IPS (3) allows enumeration of
PR1 after PR1 numbers have been reduced to the
background level of phthalate utilizers in the system.
The loss of PR1 cells and corresponding tracking data
by mab indicates that with increasing inoculation den-
sity, PR1 populations are sustained for longer periods
prior to rapid loss. This delay in the loss of cell numbers
can be attributed to exceeding the maximum response
of the bacterivorous protists in the system.
Data on the time of sustained PR1 populations and the
loss rate of PR1 cells after this delay were compiled for
all sediment slurry incubations. The delay in rapid loss
of PR1 was incorporated into a regression analysis of
time of sustained PR1 populations above 1 x 107 as a
function of inoculation density. A half-life estimation by
regression analysis incorporates loss rate data after the
period of sustained population numbers as a function of
inoculation density. This former analysis provides an
estimated lifetime for a pulse of PR1 into the treatment
zone. The latter analysis provides an extinction coeffi-
cient for cells in the downstream aquifer.
An aquifer sediment column has been established to
test the response of PR1 within the flow regime of the
target system. A commercial spring water (GMW) was
used as the diluent. A chromatography column was fitted
with cut gas chromatography (GC) vials closed with
Teflon septa to provide sampling ports. Teflon tubing and
fittings were used throughout the setup. Tygon tubing
connected to a constant temperature recirculating bath
was used to jacket the column and maintain 15°C. Flow
was controlled at the column outflow to the flow rate in
the target environment (2 cm/day"1). Cells and TCE were
added by syringe pump. An overflow ensured constant
supply of diluent, cell suspension, or TCE solution to the
top of the sediment. Pore-water samples were taken by
syringe (2:300 ul each), and used for plate counts,
direct counts of bacteria and protists, and TCE analysis
(alternate sampling periods).
With an inoculum of 1 x 108 cells/mL"1 for 1 void volume,
a population of greater than 107 CFU/mL"1 pore water
was maintained for 5 days at in the top portion of the
column. Column data expressed as depth profiles show-
ing a roughly linear decrease in PR131 numbers with
distance traveled through the column from Days 1
through 5, and the combined effect of predation and
elution decreased numbers in the upper portion of the
column at Days 8 and 10. By Day 15, the pulse has been
eliminated at the upper and lower portions of the col-
umn, leaving residual cells in the central portion. Bac-
terivorous protists increased in proportion to the
numbers of PR1 added. These organisms form resistant
cysts on sediment surfaces when food is not available.
Decreases observed over time for protist numbers are
likely due mainly to encystment of these organisms after
depletion of PR131. This results in a reservoir of
bacterivores capable of responding to subsequent addi-
tions of bacteria.
Integration of PR1 Into Stable Microbial
Consortia
Persistence of a non-native bacterium introduced into
an environment is dependent on the ability of that organ-
ism to find refuges from predation and compete with
native bacteria. One such refuge may be in biofilms. To
examine the ability of PR1 to integrate into biofilms, PR1
was introduced into existing biofilms developed from
Borden aquifer material and into a defined microbial
community (including protists) grown with input of TCE
and dibutyl phthalate. PR1 was found not to stick to
pre-existing biofilms of Borden aquifer origin, but suc-
cessfully integrated into the degradative community
growing in the presence of TCE and dibutyl phthalate.
The addition of these substances apparently provided a
competitive advantage to PR1 in this system. One year
post-inoculation, PR1 was located in biofilms from this
system by fluorescent monoclonal antibody and scan-
ning confocal laser microscopy. PR1 was found through-
out the biofilms as scattered cells and microcolonies.
This information suggests that in the presence of TCE it
may be possible to maintain and grow PR1 within a
treatment zone in the target aquifer.
Modeling of Transport and Fate of PR1 in
Borden Aquifer
Using the transport data and loss rates due to predation,
preliminary modeling exercises have been conducted.
These initial approximations assume grazing losses will
be the same for pore water and attached bacteria, and
do not address excystment/encystment processes of
the protists, but do incorporate protist grazing as a
dynamic response to variable bacterial density. In all
simulations, predation reduces the bacterial concentra-
tions in the effluent from a modeled transport column.
With an increased predation constant, peak break-
through numbers are reduced, but tailing is affected
more significantly. Tailing, which can be attributed to the
process of reversible sorption, is of concern for transport
of bacteria to greater distances in geologic media.
Thus, reductions in tailing by the protist response re-
duces the concern of offsite migration of the introduced
microorganism.
Conclusions
Data collected from these laboratory studies indicate
that biological interactions, particularly with bacterivor-
ous protists, will limit the survival of PR1 introduced into
the system and may provide for a natural containment
of the bacterium. Maintenance of PR1 numbers high
enough for mineralization activity will likely require re-
peated additions of sufficient cells to exceed protist
49
-------
maximum response. Preliminary data on TCE additions
suggest that TCE may impede protist activity until it is
mineralized, providing a gradient of protist response in
proportion to TCE removal. Ultimately, the utility of labo-
ratory analyses will be gauged against data from the
field release.
Acknowledgments
The authors thank M.S. Shields and S.C. Francisconi for
advise and assistance in the research. This work was
supported by the U.S. Environmental Protection
Agency's Gulf Breeze Environmental Research Labora-
tory through cooperative research agreement 822568 to
RAS.
References
1. Shields, M.S., S.O. Montgomery, S.M. Cuskey, P.J. Chapman, and
P.M. Pritchard. 1991. Mutants of Pseudomonas cepacia G4 defec-
tive in catabolism of aromatic compounds and trichloroethylene.
Appl. Environ. Microbiol. 57:1,935-1,941.
2. Shields, M.S., Reagin, M.J., Gerger, R.R., Cambell, R., Somerville,
C.C. 1995. TOM: A new aeromatic degradative plasmid from Burk-
holderia (Pseudomonas) cepacia G4. Appl. Environ. Microbiol.
61:1,352-1,356.
3. Winkler, J., K.N. Timmis, and R.A. Snyder. 1995. Tracking the
response of Burkholderia cepacia G4 5223 PR1 in aquifer micro-
cosms. Appl. Environ. Microbiol. 61:448-455.
4. Somerville, C.C., IT. Knight, W.L. Straube, and R.R. Colwell. 1989.
Simple, rapid method for direct isolation of nucleic acids from
aquatic environments. Appl. Environ. Microbiol. 55:548-554.
5. Reasoner, D.J., and E.E. Geldreich. 1985. Anew medium for the
enumeration and subculture of bacteria from potable water. Appl.
Environ. Microbiol 49:1-7.
6. Krumme, M.L., K.N. Timmis, and D.F. Dwyer. 1993. Degradation
of trichloroethylene by Pseudomonas cepacia G4 and the consti-
tutive mutant strain G4 5223 PR1 in aquifer microcosms. Appl.
Environ. Microbiol. 59:2,746-2,749.
50
-------
Characterization of Trichloroethylene-Degrading Bacteria
From an Aerobic Bio filter
Alec Breen, Todd Ward, Ginger Reinemeyer, John Loper, and Rakesh Govind
University of Cincinnati, Cincinnati, Ohio
John Haines
U.S. Environmental Protection Agency, Cincinnati, Ohio
Introduction
The microbial community colonizing a vapor-phase
biofilter was examined to determine the population(s)
capable of trichloroethylene (TCE) degradation. The
community had been exposed to low levels of TCE
continuously for 24 months and maintained degradation
in the absence of a canonical co-metabolite. Although
low levels of autotrophic ammonia-oxidizing bacteria
were present, nitrapyrin inhibitor studies suggested that
alternative bacteria were responsible for TCE oxidation.
In addition, replacement of ammonia with nitrate did not
affect TCE degradation. Incubation of biofilter biomass
in a toluene or benzene atmosphere resulted in a turbid
culture within 2 to 3 days. In light of these observations,
aromatic hydrocarbon-oxidizing bacteria were pursued
as putative candidates mediating TCE degradation.
A significant fraction of the culturable heterotrophs
(greater than 80 percent) were capable of growth on
toluene or benzene. This study describes the naturally
occurring TCE-degradative populations that became es-
tablished over time in the biofilter. Individual isolates
were tested for TCE-degradative capacity under several
growth conditions. The pure cultures tested were all
capable of co-metabolic TCE degradation. These organ-
isms had persisted in the biofilter regardless of condi-
tions that would exert a negative selective pressure due
to generation of the TCE-derived epoxide during aerobic
TCE degradation. Several isolates were selected for
further study. Initial sampling of the biofilter yielded three
isolates: Rhodococcus sp. TA1, Pseudomonas putida
TA2, and Nocardia sp. AR1. Both the Rhodococcus and
the P. putida could be repeatedly isolated from the biofil-
ter. Two other organisms, P. putida DC1 and Burk-
holderia cepaciaGRZ, were isolated more recently. This
consortium appears relatively resistant to the toxic ef-
fects of TCE oxidation at the concentrations used in the
biofilter.
Background
Halogenated aliphatic compounds are a major class of
industrially important chemicals that have become sig-
nificant environmental contaminants with mutagenic and
carcinogenic potential. A widespread ground-water con-
taminant, TCE can undergo co-metabolic oxidation by
a variety of physiologically diverse bacteria (1). Co-
metabolic TCE degradation by aromatic hydrocarbon
utilizing bacteria was originally reported by Nelson et al.
(2). Since that time, efforts to employ these organisms
to ameliorate TCE contamination problems in situ and
in reactors have been conducted. This study examines
the toluene-degrading bacteria surviving in a vapor-
phase biofilter. Although only a mineral salts medium
was supplied to the biofilter, toluene oxidizers survived
and TCE degradation was maintained at a level of 20
percent of a 21-ppmv gas stream.
Experimental Methods
Biofilter and Sampling
The biofilter consisted of ceramic plates in a stainless
steel casing. The initial inoculum was a municipal sludge
sample that was acclimated to a volatile organic com-
pound (VOC) mixture (benzene, toluene, ethylbenzene,
and TCE) fora period of 3 months. At this point, all VOCs
except TCE were removed. The biofilter was operated
at a gas-flow rate of 520 mL/min, and had an empty-bed
residence time of 1.9 min. TCE inlet concentration was
21 ppmv. A mineral salts solution was applied to the
biofilter at a flow rate of 357 ml/day. Biofilter sampling
was conducted by opening the biofilter and scraping
biomass off the ceramic matrix. VOC-degrading bacteria
51
-------
were isolated by incubation of biofilter material in a
mineral salts medium, with the appropriate VOC sup-
plied in the vapor phase.
TCE Mineralization Assays
Degradation experiments using 14C-TCE were con-
ducted in 20.0 ml vials, with teflon-lined silicone septa
closures allowing injection into the vial. An inner vial
containing 0.4 N NaOH served as a CO2 trap. Sterile
control vials were subtracted from experimental values
when determining conversion of TCE to CO2 or soluble
products. All data represent a mean value of triplicate
vials.
Biochemical and Genetic Characterization of
Biofilter Bacteria
Aromatic hydrocarbon utilizing bacteria were isolated as
follows. Biofilter biomass was inoculated into mineral
salts medium, and the medium was exposed to toluene
or benzene vapor as a sole carbon source for 48 hr.
Cultures were then plated onto mineral salts medium
and grown in a toluene or benzene atmosphere for 10
days. Organisms that appeared were picked and
streaked onto mineral salts plates and grown again with
the carbon source in the vapor phase. The isolates were
then checked for purity and TCE mineralization capabil-
ity. Isolates were sent to Microbial ID, Inc. (Newark,
Delaware), for fatty acid methyl ester (FAME) analysis.
16S rDNA sequencing was carried out by amplifying the
27 to 321 base pair region of the 16S rDNA gene by
polymerase chain reaction (PCR). The primers used are
forward '5AGAGTTTGATCCTGGCTCAG-3' (positions
27-46) and reverse '5AGTCTGGACCGTGTCTCAGT-3'
(positions 321-301). Both forward and reverse DNA
strands were sequenced. Gene probes for charac-
terized toluene/TCE co-metabolic oxygenases were ob-
tained from other researchers. The todABC probe was
obtained from Dr. D.T Gibson, and the tbu probe was
obtained from Dr. A. Byrne and Dr. R.H. Olsen (3, 4).
Results and Discussion
The predominant bacteria in the biofilter were shown to
be degrading TCE by toluene/benzene oxygenase co-
metabolic route. The biofilter community had not been
exposed to these compounds for over 24 months, yet
these organisms persisted and were shown to be the
key population mineralizing TCE. These isolates are of
interest because they arose spontaneously from a natu-
rally occurring population and were maintained in the
continual presence of TCE. It might be expected that
TCE-oxidizing organisms would be selected against in
such a system (5). Mixed cultures mineralized 14C-TCE
(data not shown) and exhibited very little diversity when
plated out onto a mineral salts medium and grown with
toluene or benzene vapor as a sole carbon source.
These plates generally had only one or two colony
morphotypes shown to be a P. putida (designated TA2)
and a Rhodococcus (designated TA1). A Nocardia sp.
(designated AR1) was isolated during an early sampling
time but was not reisolated. Two additional organisms,
B. cepacia GR3 and P. putida DC1, were more recently
isolated.
The predominant toluene degrader isolated from the
biofilter was TA2. Growth on glucose, succinate, ortryp-
tophan completely inhibited TCE mineralization by TA2.
A gene probe for the alpha subunit of the tbu toluene
monooxygenase strongly hybridizes with TA2 (Figure 1).
Figure 1. Slot blot analysis of biofilter DNA extracts probed
with the tbu monooxygenase alpha component:
A) P. putida DC1, B) B. cepacia GR3, C) P. putida
TA2, D) Nocardia sp. AR1, and E) Rhodococcus
sp. TA1.
This probe will also hybridize to two other toluene
monooxygenases, tmo and torn (data from our labora-
tory and unpublished information from M.S. Shields,
University of West Florida, and R.H. Olsen, University
of Michigan; therefore, the mode of toluene oxidation is
not definitively established. Slot blot analysis of biofilter
isolates probed with tbu is shown in Figure 1.
Rhodococcus sp.I'M did not hybridize strongly to either
the tbu or tod toluene oxygenase probes, indicating the
uniqueness of its toluene oxygenase. In addition to TA1
and TA2, three other TCE-co-metabolizing organisms
from the reactor were investigated, and their properties
are listed in Table 1. All organisms are being evaluated
to determine basal levels of TCE catabolic activity as
toluene is depleted. The effect of acclimation to alterna-
tive substrates on TCE degradation is also being
examined.
52
-------
Table 1. Properties and Growth Substrates of TCE-Co-metabolizing Biofilter Isolates
Organism Oxidase lndolea Benzene Ethylbenzene Phenol o-Cresol m-Cresol p-Cresol
TA1 - - + + + + + + + + + + + + + + + +
TA2 + + + + + + + + - + +
DC1 + - + + + + + +
GR3 + + + + + + + + - +
a Conversion of indole to indigo.
References 3- Byrne, A.M., J. Kukor, and R.H. Olsen. 1995. Sequence of the
gene cluster encoding toluene-3-monooxygenase from Pseudo-
monas pickettii PKO1. Gene. In press.
1. Ensley, B.D. 1991. Biochemical diversity of trichloroethylene me- 4. Wackett, L.P., and S.R. Householder. 1989. Toxicity of trichlo-
tabolism. Arm. Rev. Microbiol. 45:283-299. roethylene to Pseudomonas putida F1 is mediated by toluene
dioxygenase. Appl. Environ. Microbiol. 55:2,723-2,725.
2. Nelson, M.K.J., S.O. Montgomery, E.J. O'Neil, and P.M. Pritchard. S.Alvarez-Cohen, L, and PL. McCarty. 1991. A co-metabolic
1986. Aerobic metabolism of trichloroethylene by a bacterial iso- biotransformation model for halogenated aliphatic compounds ex-
late. Appl. Environ. Microbiol. 52:383-384. hibiting product toxicity. Environ. Sci. Tech. 25:1,381-1,387.
53
-------
Anaerobic/Aerobic Degradation of Aliphatic Chlorinated Hydrocarbons in an
Encapsulated Biomass Biofilter
Rakesh Govind and P.S.R.V. Prasad
Department of Chemical Engineering, University of Cincinnati, Cincinnati, Ohio
Dolloff F. Bishop
National Risk Management Research Laboratory, U.S. Environmental Protection Agency,
Cincinnati, Ohio
Introduction
During the past decade, public awareness and concern
about the quantity and diversity of persistent (recalci-
trant to degradation) synthetic chemicals produced by
industry have increased. Release of these chemicals
into the environment is inevitable, and hence there is a
strong need to control, direct, and improve the proc-
esses for degradation of these chemicals. Large differ-
ences in the rates and mechanisms of biodegradation
of various compounds under oxic and anoxic conditions
exist. Consequently, sequential anoxic and oxic condi-
tions, enabling cooperation of anaerobic and aerobic
bacteria, are desirable for rapid complete mineralization
of many polyhalogenated compounds (1-6).
Encapsulation of Biomass in Hydrogels
The goal of this research is to use hydrogel-encapsu-
lated bacteria for simultaneous creation of oxic and
anoxic zones inside the hydrogel bead. Further, the oxic
and anoxic zones created inside the hydrogel bead can
be successfully used to mineralize chlorinated com-
pounds such as trichloroethylene (TCE) and perchlo-
roethylene (PCE). Hydrogel beads with encapsulated
bacteria can be used to mineralize chlorinated com-
pounds present in air, ground water, and soil, and can
also be used to promote ecological competitiveness of
laboratory-grown cultures that are specially adapted for
biodegradation of specific environmental pollutants.
Preparation of New Gel Material (7)
The gelation procedure used is as follows: Silica sol,
obtained commercially as LUDOX colloidal silica SM
grade, is mixed with 1 to 3 percent sodium alginate in
the following proportion: silica sol (90 wt percent to
99wtpercent): 1 toSpercentsodiumalginatesolution
(1 wt percent to 10 wt percent). The mixture, after ad-
justment of pH between 7 and 8 by 5N HCI, is mixed
with active aerobic and anaerobic biomass cells, with
the initial cell wt percent varying from 2.0 to 10.0. The
mixture is stirred, then poured into a petri dish to a depth
of 5 mm. Calcium chloride solution (0.1 molar) is poured
on top of the mixture in the petri dish. The silica sol and
sodium alginate mixture immediately gel due to the
diffusion of calcium forming calcium alginate on the
outer surface, and the gel is allowed to cure from 10 min
to 24 hr. During the curing process, the pH of the silica
sol decreases, thereby forming silica gel with pockets of
calcium alginate inside the silica gel. The survival of
active cells is maximized by using a combination of silica
sol and sodium alginate. Once the silica sol has formed
silica gel, stainless-steel wire mesh cylinders 2.5 mm in
diameter and 5.0 mm long (open at both ends) are
pushed into the gel layer, thereby enclosing the gel
inside the wire mesh. The use of the stainless steel
mesh gives the silica gel/calcium alginate bead struc-
tural strength so that the beads can be packed in a bed
without compaction.
Experimental Studies Conducted
A 40-mL bioreactor (1.9 cm inner diameter) consisting
of a jacketed cylinder was constructed from borosilicate
glass. The reactor was packed randomly with the gel
beads. Air was passed at a controlled rate through the
bioreactor, and nutrient solution was trickled down from
the top of the bioreactor counter current to the air flow.
The air was contaminated with chlorinated pollutants,
such as TCE or PCE, using a syringe pump that injected
the liquid contaminant into the air line through a septum.
The concentration of the contaminant in the air stream
54
-------
was varied within the following range: toluene 0 to
100 ppmv; TCE 0 to 25 ppmv; and PCE 0 to 25 ppmv.
The reactor temperature was maintained at 25°C by circu-
lating water from a constant temperature bath through the
jacket of the bioreactor. Nutrient solution was trickled
down the bioreactor at a flowrate of 1 liter per day, and
the nutrient composition was as follows: KhfePCX, (85 mg/L),
K2HPO4 (217.5 mg/L), Na2HPO4-2H2O (334 mg/L), NH4CI
(25 mg/L), MgSO4-7H2O (22.5 mg/L), CaCfe (27.5 mg/L),
FeCI3-6H2O (0.25 mg/L), MnSO4-H2O (0.0399 mg/L),
H3BO3 (0.0572 mg/L), ZnSO4-7H2O (0.0428 mg/L),
(NH4)6Mo7O24 (0.0347 mg/L), FeCI3-EDTA(0.1 mg/L), yeast
extract (0.15 mg/L), and formate (50 mg/L). Results obtained
are shown in Table 1.
Table 1. Percent degradation of TCE in the bioreactor at
various air flowrates. Inlet TCE concentration was
25 ppmv, and nutrient flowrate was 1 liter/day.
Air Flowrate
(mL/min)
Percent Degradation
of TCE
35
40
50
60
65
100.0
67.2
40.7
22.1
10.8
Carbon and chlorine balances were made by monitoring
the increase in carbon dioxide in the exit air, and in-
crease in chloride ion concentration in the exit nutrients
was analyzed by an ion chromatograph. The chlorine
balance was developed at steady-state conditions within
an error band of 15 percent of the calculated increase
in chloride ion concentration.
The proposed degradation pathway was shown to be a
partial dehalogenation in the anoxic zone followed by
oxic biodegradation of the anoxic degradation products
in the outer aerobic zone of the gel bead. The anoxic
zone was created due to oxygen consumption in the
aerobic zone by the oxic degradation of the partially
dehalogenated products as they diffused out from the
anoxic zone.
A mathematical model was developed to describe the
diffusion of TCE and oxygen, and consumption of oxy-
gen due to aerobic degradation of the dehalogenated
products. At the outer surface of the gel bead (denoted
by dimensionless position of 1.0) the oxygen concentra-
tion is about 8 mg/L due to presence of air outside the
bead. As oxygen diffuses inside the gel bead, it is con-
sumed due to aerobic degradation of the dehalogenated
products diffusing outwards. At some point in the interior
of the gel bead, oxygen is completely consumed produc-
ing an anoxic zone in the interior portion of the gel bead.
It is in this anoxic zone that dehalogenation of TCE
occurs. The formate in the nutrient medium is rapidly
absorbed by the gel bead and provides the organic
carbon source needed for partial dehalogenation of TCE
in the anoxic zone by anaerobic microbiota. Other po-
tential carbon sources for anaerobic microorganisms
are acetate and other carboxylic acids.
Experiments also were conducted with perchlo-
roethylene (PCE) at an inlet concentration of 25 ppmv.
Results obtained are shown in Table 2. Chloride ion
balances were obtained at steady-state to prove that
complete mineralization of PCE had occurred. Each
experiment had to be conducted for over 5 days to
achieve a stable exit concentration of chloride ion in the
exit nutrients. No other by products were observed in the
exit gas phase at the above operating conditions.
Table 2. Percent degradation of PCE in the bioreactor at
various air flowrates. Inlet PCE concentration was
25 ppmv, and nutrient flowrate was 1 liter/day.
Air Flowrate
(mL/min)
10
15
20
30
50
Percent Degradation
of PCE
100.0
86.7
72.4
41.8
12.8
Conclusions
The hydrogel-encapsulated biomass reactor is capable
of biodegrading trichloroethylene (TCE) and perchlo-
roethylene (PCE) through an anaerobic/aerobic degra-
dation mechanism. Experimental results indicate that
the degradation of TCE and PCE is complete, and the
empty-bed gas phase residence time for complete
removal is less than a few minutes. Further studies
are ongoing to quantitate the transport parameters
and apply the process for treatment of TCE or PCE
in ground water.
References
1. Abramowitz, D.A. 1990. Aerobic and anaerobic biodegradation of
PCBs: A review. Crit. Rev. Biotechnol. 10:241-251.
2. Beunink, G., and H.J. Rehm. 1988. Synchronous anaerobic and
aerobic degradation of DDT by an immobilized mixed culture sys-
tem. Appl. Environ. Microbiol. 151:95-100.
3. Beunink, J., and H.J. Rehm. 1990. Coupled reductive and oxida-
tive degradation of 4-chloro-2-nitrophenol by a co-immobilized
mixed culture system. Appl. Microbiol. Biotechnol. 29:72-80.
4. Fathepure, B.Z., and T.M. Vogel. 1991. Complete degradation of
polychlorinated hydrocarbons by a two-stage biofilm reactor. Appl.
Environ. Microbiol. 57:3,418-3,422.
5. Fogel, S., R.L. Lancione, and A.E. Sewall. 1982. Enhanced biode-
gradation of Methoxychlor in soil under sequential environmental
conditions. Appl. Environ. Microbiol. 44:113-120.
55
-------
6. Kastner, M. 1991. Reductive dechlorination of tri- and tetrachlo- 7. Bishop, D.F., and R. Govind. 1995. New hydrogel material for
roethylenes depends on transition from aerobic to anaerobic con- degradation of persistent pollutants in immobilized film bioreactors.
ditions. Appl. Environ. Microbiol. 57:2,039-2,046. U.S. Patent Application.
56
-------
Operation and Optimization of Granular Air Biofilters
Francis Lee Smith, George A. Serial, Makram T. Suidan, Amit Pandit, and Pratim Biswas
Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, Ohio
Richard C. Brenner
Risk Reduction Engineering Laboratory, U.S. Environmental Protection Agency, Cincinnati, Ohio
Introduction
Since enactment of the 1990 amendments to the Clean
Air Act, the control and removal of volatile organic com-
pounds (VOCs) from contaminated air streams have
become major public concerns (1). Consequently, con-
siderable interest has evolved in developing more eco-
nomical technologies for cleaning contaminated air
streams, especially large, dilute air streams. Biofiltration
has emerged as a practical air pollution control (ARC)
technology for VOC removal (2-4). In fact, biofiltration
can be a cost-effective alternative to the more traditional
technologies, such as carbon absorption and incinera-
tion, for removal of low levels of VOCs in large air
streams (5, 6). Such cost-effectiveness stems from a
combination of low energy requirements and microbial
oxidation of the VOCs at ambient conditions.
Our biofiltration research has focused on expanding the
range of application of biofiltration technology to the
treatment of high VOC loadings at consistently high
removal efficiencies. The preliminary period of our re-
search was dedicated to the pilot-scale comparison of
three different types of biological attachment media: a
patented peat mixture and two synthetic inorganic me-
dia, one channelized and the other pelletized. The biofil-
ters containing the latter two media were operated as
trickle bed air biofilters (TBABs), called such because
the media received a steady application of water. After
18 months of testing, the pelletized medium (Celite
6-mm R-635 Bio-Catalyst Carrier) was demonstrated to
be significantly better than the other two for handling
high VOC loadings (7-9). Subsequent work to evaluate
the performance and behavior of biofilters using the
R-635 pelletized medium produced two significant find-
ings: first, that an increase in the biofilter operating
temperature permits a significantly higher practical VOC
loading (i.e., a significantly smaller required media vol-
ume), and second, that biofilter performance decreases
substantially with the buildup of back pressure due
to the accumulation of biomass within the media bed
(10,11).
Working exclusively with this pelletized medium, our
continued research focused on the development of
strategies for long-term operation with high VOC load-
ings at sustained high-removal efficiencies. This re-
search effort demonstrated that this objective could be
achieved using a biomass removal and control strategy
employing periodic backwashing of the media with
water. Backwashing (the upflow washing of the fluidized
media with water) gently removes excess biomass from
the media, circumventing the problems noted earlier for
this medium. A second finding of this research was that
NO3-N as the sole nitrogen source was superior to
NH3-N. The use of NO3-N resulted in lower volatile
suspended solids (VSS): chemical oxygen demand
(COD) and VSS:N ratios. In other words, for a given set
of operating conditions, less biomass is produced and
less nitrogen is consumed. Finally, it was also observed
that both the recovery of the VOC removal efficiency
with time after backwashing (unsteady state) and the
VOC removal efficiency with depth (at near steady state)
were superior when using NO3-N.
This paper discusses the continuing research being per-
formed for the development of biofiltration as an effi-
cient, reliable, and cost-effective VOC ARC technology.
The objectives of this effort were to investigate the re-
moval efficiencies of TBABs under high toluene loadings
and low residence times, and to evaluate the associated
development and control of excess biomass with time.
The biofilter operational period between backwashings
was evaluated to determine its effect on the stability of
biofilter performance. Backwashing variables, including
backwashing frequency and backwashing duration,
were evaluated.
57
-------
Methodology
The biofilter apparatus used in this study consisted of
four independent, parallel biofilter trains, each contain-
ing 3.75 ft of pelletized Celite 6-mm R-635 biological
attachment medium. A detailed schematic, equipment
description, and typical system operation are given else-
where (8). Each biofilter had a circular cross section with
a 5.75-in. internal diameter (ID). The air feed was mass-
flow controlled, and the VOC (liquid toluene) was me-
tered by syringe pumps into the air feed stream. Each
biofilter was fed a buffered nutrient feed solution con-
taining all necessary macro- and micronutrients with a
sodium bicarbonate buffer, described elsewhere (8). For
each biofilter, the sole nitrogen nutrient source was NCV
N. The flow directions of the air and nutrients were
downward. All biofilters were insulated and inde-
pendently temperature controlled at 32.2°C.
Results
Each biofilter was loaded with clean, sterilized pellets
and seeded with backwashing water from a similar,
previous run. For each biofilter, a significant dip in per-
formance occurred from Days 17 to 26 due to an error
in preparing the nutrient solution that resulted in feeding
insufficient NO3-N. After the target VOC loading was
achieved, at a COD:NO3-N ratio of 50:1, it was main-
tained for the duration of the run.
Three different backwashing strategies were tested on
all biofilters sequentially. After restart following back-
washing, effluent samples were collected to determine
the recovery of the VOC removal efficiency with time.
On days when no backwashing occurred, samples were
collected along the length of the bed to determine the
VOC removal efficiency with respect to depth.
Biofilter A
This biofilter was started up at 50 ppmv toluene influent
concentration, 1.33 min empty bed residence time
(EBRT), and 21 mmol NO3-N per day. On Day 17, the
biofilter was backwashed for the first time. Detailed
schematic and backwashing descriptions are given
elsewhere (12). The procedure used was to recycle 70
L of 32.2°C tap water through the bed, bottom to top, at
a rate of 57 L/min to induce full media fluidization at a
bed expansion of about 40 percent. At the end of the
backwashing period, the media was flushed at the same
rate with another 50 L of clean, 32.2°C tap water. For
this first backwashing strategy, the period and frequency
were 1 hr twice per week. The target influent VOC
concentration (500 ppmv toluene) and loading (6.2 kg
COD/m3 day) were reached on Day 53. On Day 129, the
second backwashing strategy was started using a pe-
riod and frequency of 2 hr twice per week. On Day 171,
the third and final backwashing strategy was started
using a period and frequency of 1 hr every 2 days. The
performance of Biofilter A is shown in Figure 1.
Biofilter B
This biofilter was started up at 50 ppmv toluene influent
concentration, 0.67 min EBRT, and 21 mmol NO3-N per
day. On Day 17, the biofilter was backwashed for the
first time using the first backwashing strategy of 1 hr
twice per week. The target influent VOC concentration
(250 ppmv toluene) and loading (6.2 kg COD/m3 day)
were reached on Day 35. On Day 129, the second
backwashing strategy of 2 hrtwice perweekwas begun,
and on Day 171, the third strategy of 1 hr every 2 days
was begun. The performance of Biofilter B is shown in
Figure 2.
Biofilter C
This biofilter was started up at 50 ppmv toluene influent
concentration, 1.0 min EBRT, and 21 mmol NO3-N per
day. On Day 17, the biofilter was backwashed for the
first time using the first backwashing strategy of 1 hr
twice per week. The target influent VOC concentration
(250 ppmv toluene) and loading (4.1 kg COD/m3 day)
800
100
6.2 kg COD/m
Influent Cone
T Effluent Cone.
o Percent Remova
100 150
Sequential Date, days
Figure 1. Performance of Biofilter A with backwashing.
58
-------
700
600
500
1 400
300
200
EBRT: 0.67 minutes
6 2 kg COD/m3.day
Influent Cone
» Effluent Cone.
a Percent Removal
100
80
60
40
20 .a
50 100 150
Sequential Date, days
200
250
Figure 2. Performance of Biofilter B with backwashing.
were reached on Day 53. On Day 129, the second
backwashing strategy of 2 hrtwice perweekwas begun,
and on Day 171, the third strategy of 1 hr every 2 days
was begun. The performance of Biofilter C is shown in
Figure 3.
Biofilter D
This biofilterwas started up at 50 ppmv toluene influent
concentration, 2.0 min EBRT, and 21 mmol NO3-N per
day. On Day 17, the biofilterwas backwashed for the
first time using the first backwashing strategy of 1 hr
twice per week. The target influent VOC concentration
(500 ppmv toluene) and loading (4.1 kg COD/m3 day)
were reached on Day 53. On Day 129, the second
backwashing strategy of 2 hrtwice perweekwas begun,
and on Day 171, the third strategy of 1 hr every 2 days
was begun. The performance of Biofilter D is shown in
Figure 4.
Conclusions
The performances of the four biofilters with respect to
the three backwashing strategies were similar, although
clearly affected by both the loading and the residence
time. The effectiveness of the three strategies increased
from the first through the third strategy. This shows that
both backwashing duration and frequency are very im-
portant parameters for control of the biofilters' VOC
removal efficiency. The third and best strategy, however,
actually had less total backwashing time per week. At
the higher loading, the greater than 90 percent removal
efficiencies of both biofilters were unexpectedly high for
the third backwashing strategy but below the sustained
99.9 percent achieved by the lower loaded biofilters. It
can also be seen that for a given loading, the perform-
ance at the lower EBRTs is more sensitive to the back-
washing strategy employed. Both of these effects were
anticipated; what was not anticipated was that this
pelletized medium would perform so well for any back-
washing strategy at a loading of 6.2 kg COD/m3 day.
These findings, as well as biofilter recovery of perform-
ance after backwashing, will be presented.
Acknowledgment
This research was supported by Cooperative Agree-
ment CR-821029 with the U.S. Environmental Protec-
tion Agency.
800
700
60°
400
U>
8
t-1 200
100 -
0
4.0
. 3.5
§ 3.0
cd
s
EBRT: 1.0 minutes
4.1kg COD/m3.day
Influent Cone,
T F.ffluent Cone.
a Percent Removal
r
riL
w~%*-»
80
60
40
I
I
1.5
0.5
0,0
50
100 150
Sequential Date, days
Figure 3. Performance of Biofilter C with backwashing.
59
-------
800
cf 500
o
1
g 400
a
o
^ 300
u
_3
° 200
100
0
4.0
3.5
8 3.0
I"
& 2.0
0.5
0.0
EBRT: 2.0 minutes
4.1 kgCOD/m3.day
Influent Cone.
T Effluent Cone.
D Percent Removal
100
80
60
40 £
>*
20 |
e
100 150
Sequential Date, days
200
250
Figure 4. Performance of Biofilter D with backwashing.
References
1. Lee, B. 1991. Highlights of the Clean Air Act Amendments of
1990. J. Air Waste Manag. Assoc. 41(1):16-31.
2. Leson, G., and A.M. Winer. 1991. Biofiltration: An innovative air
pollution control technology for VOC Emissions. J. Air Waste
Manag. Assoc. 41(8):1,045-1,054.
3. Leson, G., F. Tabatabal, and A.M. Winer. 1992. Control of haz-
ardous and toxic air emissions by biofiltration. Paper presented
at the Annual Meeting and Exhibition of the Air & Waste Manage-
ment Association, Kansas City, MO, June 21-26.
4. Ottengraf, S.P.P. 1986. Exhaust gas purification. Rehn, H.J., and
G. Reed, eds. In Biotechnology, Vol. 8. Weinham, Germany: VCH
Verlagsgesellschaft.
5. Ottengraf, S.P.P. 1986. Biological elimination of volatile xenobiotic
compounds in biofilters. Bioprocess Eng. 1:61-69.
6. Severin, B.F., J. Shi, and T. Hayes. 1993. Destruction of gas
industry VOCs in a biofilter. Paper presented at the IGT sixth
International Symposium on Gas, Oil, and Environmental Tech-
nology. Colorado Springs, CO, November 29 - December 1.
7. Smith, F.L., G.A. Serial, P.J. Smith, M.T. Suidan, P. Biswas, and
R.C. Brenner. 1993. Preliminary evaluation of attachment media
for gas phase biofilters. Paper presented at the U.S. EPA Sym-
posium on Bioremediation of Hazardous Wastes: Research, De-
velopment, and Field Evaluation. Dallas, TX, May 4-6.
8. Serial, G.A., F.L. Smith, P.J. Smith, M.T. Suidan, P. Biswas, and
R.C. Brenner. 1993. Development of aerobic biofilter design cri-
teria for treating VOCs. Paper no. 93-TP-52A.04. Presented at
the 86th Annual Meeting and Exhibition of Air & Waste Manage-
ment Association. Denver, CO, June 13-18.
9. Serial, G.A., F.L. Smith, P.J. Smith, M.T. Suidan, P. Biswas, and
R.C. Brenner. 1993. Evaluation of biofilter media for treatment of
air streams containing VOCs. In: Proceedings of the Water En-
vironment Federation 66th Annual Conference and Exposition,
Facility Operations Symposia, Volume X. pp. 429-439.
10. Serial, G.A., F.L. Smith, M.T. Suidan, P. Biswas, and R.C. Bren-
ner. 1994. Evaluation of the performance of trickle bed biofilters
Impact of periodic removal of accumulated biomass. Paper no.
94-RA115A.05. Presented at the 87th Annual Meeting and Exhi-
bition of Air & Waste Management Association. Cincinnati, OH,
June 19-24.
11. Smith, F.L., G.A. Serial, M.T. Suidan, P. Biswas, and R.C. Brenner.
1994. Pilot-scale evaluation of alternative biofilter attachment me-
dia for the treatment of VOCs. Paper presented at the U.S. EPA
Symposium on Bioremediation of Hazardous Wastes: Research,
Development, and Field Evaluations. San Francisco, CA, June
28-30.
12. Smith, F.L., M.T. Suidan, G.A. Serial, P. Biswas, and R.C. Bren-
ner. 1994. Trickle bed biofilter performance: Biomass control and
N-nutrient effects. Paper no. AC946004. Presented at the Water
Environment Federation 67th Annual Conference and Exposition,
Facility Operations Symposia. Chicago, IL, October 15-19.
60
-------
Abiotic Fate Mechanisms in Soil Slurry Bioreactors
John A. Glaser and Paul T. McCauley
National Risk Management Research Laboratory, U.S. Environmental Protection Agency,
Cincinnati, Ohio
Majid A. Dosani, Jennifer S. Platt, and E. Radha Krishnan
IT Corporation, Cincinnati, Ohio
Introduction
Soil slurry treatment of contaminated soil has been
shown to offer a viable technology for soil bioremedia-
tion. This technology, however, has not sufficiently pro-
gressed to be a durable, reliable, and cost-effective
treatment option (1).
The use of aggressive mixing energy to provide condi-
tions for improved contact between soil contamination
and microorganisms capable of degrading the contami-
nation is the hallmark of slurry treatment technology. A
more complete description of pollutant mass transfer
during the treatment phase is required that includes
treatment fate mechanisms attributable to biotic and
abiotic processes. Losses attributable to abiotic means
can be overlooked in field application of the technology,
because limited questions can be successfully ad-
dressed at field scale. Discussions with U.S. Environ-
mental Protection Agency (EPA) regional personnel and
inspection of active field-scale soil slurry bioreactor op-
erations have identified operational problems, such as
foaming, that could result in possible abiotic loss (2).
Field bioslurry operations have adopted various ap-
proaches to reduce foaming: 1) addition of defoaming
agents, 2) reduction of the rotational speed of the agita-
tor, and 3) reduction of gas flow through the bioreactor
system. The foaming phenomenon is generally consid-
ered a nuisance, rather than a potential beneficial re-
moval mechanism. Where pollutants have a specific
gravity less than water once desorbed from the slurried
soil, the pollutants would rise to the surface, as in a
flotation process. One of our working hypotheses was
that foam formation could be related to this pollutant
release process. If this analysis has merit, it is possible
that the operational strategy used in the field is counter-
productive, because a separated contaminant phase is
re-entrained with partially cleaned soil material.
We have conducted two bench-scale slurry bioreactor
treatability studies at the EPA Testing & Evaluation Fa-
cility in Cincinnati, Ohio, which were designed to assess
operating factors leading to foam formation, and to iden-
tify the most advantageous means to deal with foaming.
The initial study was previously presented as a general
treatability study for treatment of creosote contamination
in a soil (3). During this previous study, foaming became
a major problem for operation. Use of a defoamer con-
trolled foaming conditions, as did reduction of the mixer
rotational speed and gas flow in the more extreme
cases. Subsequent studies devoted specifically to in-
vestigating the causes and conditions of foaming and
using a different batch of soil from the same site as the
earlier study showed little foaming at the beginning of
the study.
Methodology and Experimental Designs
Foam Study (First Study)
A soil from St. Louis Park, Minnesota, contaminated with
polynuclear aromatic hydrocarbons (PAHs) was used to
assess the importance of foaming conditions to the
performance of bench-scale slurry reactors. The design
of the bench-scale experimental bioslurry reactor has
been reported (3). Operational slurry volume was 6 L,
representing 75 percent of the total reactor volume.
To evaluate the conditions and causes of foam forma-
tion, a subsequent study was designed. This investi-
gation used the monitoring conditions specified
for the treatability study, and was conducted us-
ing six bench-scale slurry reactors. Each reactor
was loaded at 30 percent solids, with an initial volume
61
-------
of 6 L. The study design employed two reactors that
were permitted to develop foam, two reactors in which
foam formation was suppressed through the use of a
defoamer, and two reactors in which formaldehyde was
used to suppress biological activity. The study duration
was 1 week, with foam sampling based on the cumula-
tive production throughout the study period.
Foam and Scale Study
Experimental variables selected for the foam and scale
study were 30 percent soil solids loading, two treatment
conditions of condensed foam (removed), and foam
retention within the reactors. Each condition was repli-
cated with a single replicated control based on the foam
retention condition. The foam-retained conditions were
maintained through the use of a dispersant (Westvaco,
Reax 100M). The soil was classified to a minus 3/16 in.
dimension; no provisions were developed for prelimi-
nary dispersion of the soil solids or sand exclusion.
Scale was collected from the reactor walls after the
mobile soil solids were removed. The scale was strongly
fixed to the reactor wall, and some effort was required
to chip the scale away from the reactor.
General Reactor Operation Condition
The control reactors were operated under abiotic condi-
tions to serve as bioinactive control reactors. Formalde-
hyde was used as a biocide in these reactors and
maintained at 2 percent residual concentration.
The following monitoring and operating conditions were
held constant for the reactors:
Dissolved oxygen: greater than 2 mg/L
pH: range of 6 to 9
Ambient temperature: recorded daily
Treatment duration: 10 weeks
Nutrients: C:N:P ratio = 100:10:1
Antifoam: as needed to control foam
Results
In a previous treatability study, a high solids control
reactor showed the greatest amount of foaming (3). The
amount of foaming was surprising because foam forma-
tion was expected to be related to the formation of
biosurfactants by the microbiota. Addition of formalde-
hyde to the control reactor was the only other explana-
tion forthe foam formation observed. Other reactors had
foaming problems, but this reactor was very noticeable
by contrast. Higher solids loading was also observed to
contribute to foam formation.
In contrast to the earlier studies in which foaming was
observed, the foaming study showed little foam forma-
mg/kg
2-& 3-PAHs
t-PAHs
c-PAHs
1474
2311
3786
1016
7950
11905
19855
5367
Foam Cone
Factor
5.4
5.2
5.2
5.3
100 1000 10000 100000
Thousands
Dlnit.Conc. DFoam Cone.
7 DAY STUDY
Figure 1. Foam composition and concentration factor (Study 1).
tion. A small amount of foaming occurred on the first day
of operation. Figure 1 shows the increased concentra-
tion of the t-PAH analytes in the foam, which is five times
greater than the concentration found in the initial sus-
pended soil slurry. A second attempt to evaluate foam
formation is shown in Figure 2. These data show a
decrease in the foam concentration factor, and are prob-
ably more realistic than the first foam study results. The
second study was also designed to evaluate the depo-
sition of t-PAH analytes as part of a scale formed in the
reactor. Figures 3 and 4 show the results of t-PAH
deposition in the scale under conditions in which the
foam was condensed (removed from the reactor) and
retained within the reactor. From inspection of the re-
sults, it is clear that the higher molecular weight compo-
nents of the t-PAHs were deposited in the scale in
quantities 10 to 90 percent above the initial suspended
slurry concentration. A mass balance analysis will be
presented that puts the importance of these abiotic fate
mechanisms into perspective.
Differences in physical characteristics of the soil and
operation of the bioslurry reactor between the two stud-
ies may have contributed to decreased foam formation
in the foam study. Although soil for the same site was
used for both the treatability study and foam study, the
batch of soil used forthe foam study was coarser (less
than 1/4 in.) which may have resulted in lower PAH
2-& 3-PAHs
4-to 6-PAHs
t-PAHs
c-PAHs
36
mg/kg
1848
1147
2995
812
1467
2806
4273
2077
OQ 300a 2600 20OO 15OO 1OOO 500 1OO 1O»
c
D nit. Cone. dFoam Cone
Foam Cone
Factor
0.8
2.4
1.4
2.6
10000
14 DAY STUDY
Figure 2. Foam composition and concentration factor (Study 2).
62
-------
mg/kg
2-& 3-PAHs
4-to 6-PAHs
1848
1147
2995
812
1208
2008
3215
1568
Scale Cone
Factor
0.6
mg/kg
1.7
1.1
1.9
I 3000 2SOO 2OOO 15OO 10OO 5OO 100
Dlnil.Conc. D Scale Cone.
10OOO
14 DAY STUDY
Figure 3. Scale concentrations for foam condensed reactors
(Study 2).
concentrations in the foaming study and to decreased
foam formation. Furthermore, the air flow rate for re-
actors in the foam study (1 ft3/min) was approximately
20 percent of that used in the treatability study (5
ft3/min), which also may have contributed to decreased
foam formation.
Conclusions
Foam formation continues to be an unpredictable and
poorly understood event associated with slurry treat-
ment. The results of ourstudies are based on the bench-
scale reactor and may exaggerate the abiotic fate
mechanisms due to high surface-to-volume ratio consid-
erations. The concentration effects associated with foam
formation indicate that foam removal may be desirable
1717
1013
2730
730
2337
1312
3649
976
Scale Cone
Factor
1.4
1.3
1.3
3OOO 25OO 2000 15OO 10OO 5OO ID
100 10OO 10OOO
Dlnit.Conc. CDScale Cone. 14 DAY STUDY
Figure 4. Scale concentrations for foam retained reactors
(Study 2).
to optimize slurry reactor performance. Future studies
will endeavor to evaluate foam separation as part of the
slurry process.
References
1. U.S. EPA. 1990. Engineering bulletin: Slurry biodegradation.
EPA/540/2-90/016. Cincinnati, OH.
2. Jerger, D.E., S.A. Erickson, and R.D. Rigger. 1994. Full-scale
slurry phase biological treatment of wood-preserving wastes at a
Superfund site. Draft manuscript.
3. Glaser, J.A., M.A. Dosani, P.T. McCauley, J.S. Platt, E.J. Opatken,
and E.R. Krishnan. 1994. Soil slurry bioreactors: Bench scale stud-
ies. In: U.S. EPA Twentieth Annual RREL Research Symposium:
Abstract Proceedings. EPA 600/R-94-011. Cincinnati, OH. p. 127.
63
-------
Design and Testing of an Experimental In-Vessel Composting System
Carl L. Potter and John A. Glaser
U.S. Environmental Protection Agency, Cincinnati, Ohio
Majid A. Dosani, Srinivas Krishnan, Timothy A. Deets, and E. Radha Krishnan
IT Corporation, Cincinnati, Ohio
Introduction
The goal of this compost research is to evaluate the
potential use of compost systems in remediation of soils
contaminated with hazardous chemicals. We have de-
veloped bench-scale composters to evaluate factors
controlling compost treatment at large scale. We are
currently studying the ability of compost microorganisms
to biodegrade polynuclear aromatic hydrocarbons
(PAHs) in in-vessel reactors located at the U.S. Environ-
mental Protection Agency's Test & Evaluation (T&E)
Facility in Cincinnati, Ohio.
Composting differs from other ex situ soil treatment
systems in that bulking agents are added to the compost
mixture to increase porosity and serve as sources of
easily assimilated carbon for biomass growth. Aerobic
metabolism generates heat, resulting in significant tem-
perature increases that bring about changes in the mi-
crobial ecology of the compost mixture.
Optimal conditions for composting may vary depending
on many factors, but generally aerobic conditions with
45° to 55°C (mesophilic temperature range), 40 to 60
percent moisture, and a carbon-to-nitrogen ratio of 20:1
to 30:1 have been considered best. Mesophilic com-
posting in the range of 35°C to 50°C might prove to be
the most effective at destroying certain wastes. Main-
taining temperature below 50°C, however, may not al-
ways be cost effective if cooling requires too much
energy.
In an active compost pile, temperature can easily ex-
ceed 55°C, and temperatures above 70°C have been
reported. When the temperature exceeds 55°C, called
the thermophilic stage, most bacteria are killed. Organ-
isms capable of sporulation, such as some bacteria (2)
and fungi (3, 4), will sporulate and remain dormant until
aerobic activity slows; the temperature falls back into the
mesophilic range when they re-emerge.
Reactor Design
Ten 55-gal, insulated stainless steel compost reactors
have been fabricated to provide the closely monitored
and controlled conditions required for treatability stud-
ies. These fully enclosed, computer-monitored, bench-
scale reactors hold about 1/4 yd3 total compost mixture.
The reactor units stand upright with air flowing vertically
up through the compost mixture for 23 hours per day.
Enclosed units permit on-line analysis of oxygen, carbon
dioxide, and methane at inlet and exit locations. A data-
logging system accumulates data and transmits them to
the PC-based central data system that monitors and
controls each reactor. XAD traps in the exit line of each
composter permit trapping of volatile organic com-
pounds (VOCs) for analysis.
The bottom of each reactor contains a conical collection
system for periodic sampling of any leachate leaving the
reaction mixture. The space above the leachate collec-
tion system holds 2 in. of gravel. Mass balance studies
on soil contaminants are possible by direct sampling of
the reaction mixture at different depths through bung
holes in the lid, together with capture of VOCs and
leachate leaving the reactor.
Periodic determination of compost moisture content in
each reactor unit permits adjustment of total moisture
content in the compost matrix to 40 to 50 percent.
Moisture condensers inside compost units promote re-
tention of moisture within the reactor. Otherwise, with
typical airflows, each unit could lose significant amounts
of water daily. If moisture falls below 40 percent, a water
distribution system inside the reactor may be used to add
water to the reaction mixture without opening the reactor.
The cylindrical reactor design permits mixing of reactor
contents by rolling each unit on a drum roller at desired
intervals. Mixing can be used to break up anaerobic
pockets and to avoid packing of the compost mixture. All
64
-------
reactors are mixed simultaneously by placing them on
rollers over a modified conveyor belt that forces the
reactors to turn in unison. Baffles inside the reactors
promote mixing during rolling.
Insulation between the reactor core and outer shell re-
duces heat loss from the reactor during aerobic activity.
Heating coils provide the option of warming the reactor
to accelerate composting during startup. Each compos-
ter houses five thermocouples connected to a central
computer for on-line temperature measurements. Ther-
mocouples reside at four equally spaced locations within
the compost mixture, and a fifth thermocouple tracks
ambient temperature outside the reactor. If the mean
temperature of the middle two reactor thermocouples
exceeds a predetermined high value, the computer
switches that unit to high air flow (60 L/min) to cool the
reaction mixture. After the high-temperature unit cools
to a specified low temperature, the computer switches
the unit back to low air flow (5 L/min) to reduce further
heat loss from the reaction mixture.
Methods
Current studies focus on defining acceptable operating
conditions and process characteristics to establish suit-
able parameters for treatment effectiveness. Parame-
ters of interest include aeration, moisture dynamics,
heat production, and physical and chemical properties
of the compost mixture. Growth of microorganisms and
disappearance of parent compounds serve as indicators
of parameter suitability.
A 24-day treatability study, using field soil from the Reilly
Tar Pit Superfund site near Minneapolis, Minnesota, was
conducted to evaluate performance of the compost re-
actor system. The soil was contaminated with creosote and
contained 22 PAHs that were measured during the study.
The study design included five replicated treatment con-
ditions involving different ratios of corn cobs to soil and
different airflow rates in 10 reactors. Soil/bulking agent
compositions evaluated in this study were 50:50 (four
reactors) and 30:70 (two reactors) ratios of corn cobs to
soil (50 percent soil and 70 percent soil, respectively).
Selected airflow rates were 5 and 10 L/min.
Results and Discussion
Temperatures in reactors with 50 percent soil and
moisture content of about 50 percent or less climbed to
the upper mesophilic and lower thermophilic ranges.
Temperatures in reactors with moisture content above
53 percent failed to increase much above 30°C. This
might indicate that higher moisture content restricted air
flow through the compost mixture, resulting in insuffi-
cient aerobic activity to attain high temperatures. Reac-
tors with air flows of 5 and 10 L/min exhibited similar
temperatures within the compost mixture.
Reactors with 70 percent soil in the compost remained
relatively cool throughout the entire run, never reaching
the mid to upper mesophilic temperature range. These
reactors tended to maintain higher moisture content
throughout the study. Fewer corn cobs to absorb excess
moisture in the mixture may have resulted in flooding of
the pore space, blocking of air flow through the mixture,
and reduced drying.
Total heterotrophic populations increased from a range of
107-108 to 109-7.6 x 1010 (60- to 300-fold increases) in
reactors during the first 24 hr of composting. Heterotroph
counts ranging from 1.6 x 109 to 1.4 x 1010 remained after
24 days in reactors with 50 percent soil, but had returned
to around 2 x 108 in reactors with 70 percent soil.
Small PAHs (two to three rings) were reduced by aver-
ages of 50 and 30 percent in compost mixtures of 50
and 70 percent soil, respectively, after 24 days. Large
PAHs (four to six rings) were not decreased under any
treatment condition after 24 days. Continued evaluation
of the compost mixture will provide more information on
the long-term ability of composting to destroy large PAHs.
Future investigations will include application to pen-
tachlorophenol and other soil contaminants yet to be
specified. Evaluation of pollutant mass balance and
biotransformation products is an important aspect of
future research.
To judge the abilities of microorganisms to degrade
hazardous wastes in soil under various composting con-
ditions, emphasis will be placed on diagnosing popula-
tion changes throughout treatment and identifying
microbial species responsible forbiodegradation of con-
taminants. Early microbiological studies have focused
on enumerating total microorganisms and determining
the presence of PAH degraders. Future studies will fo-
cus on characterizing changes in biological activity dur-
ing the four stages of composting, and on identifying the
microbial species responsible for significant biodegrada-
tion of PAHs during each composting stage. Reappear-
ance of fungi and other mesophiles (e.g., Actinomycetes)
during the cooling stage is also of interest.
References
1. Nakasaki, K., M. Sasaki, M. Shoda, and H. Kubota. 1985. Change
in microbial numbers during thermophilic composting of sewage
sludge with reference to CO2 evolution rate. Appl. Environ. Micro-
biol. 49(1):37-41.
2. Strom, P.P. 1985. Identification of thermophilic bacteria in solid-
waste composting. Appl. Environ. Microbiol. 50(4):906-913.
3. Fogarty, A.M., and O.H. Tuovinen. 1991. Microbiological degrada-
tion of pesticides in yard waste composting. Microbiol. Rev.
June:225-233.
65
-------
Integrated Systems To Remediate Soil Contaminated With Wood Treating Wastes
Makram T. Suidan, Amid P. Khodadoust, Gregory J. Wilson, and Karen M. Miller
Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, Ohio
Carolyn M. Acheson and Richard C. Brenner
National Risk Management Research Laboratory, U.S. Environmental Protection Agency,
Cincinnati, Ohio
Introduction
Approximately 15 percent of Superfund Records of De-
cision (RODs) are directed towards sites contaminated
with wood treating wastes (1). Several types of pollut-
ants characterize these sites, including pentachlorophe-
nol (PCP), creosote, polycyclic aromatic hydrocarbons
(PAHs), other hydrocarbons, and heavy metals such as
copper, chromium, arsenic, and zinc (2). A process (Fig-
ure 1) that integrates soil washing with sequential an-
aerobic and aerobic biotreatment is being developed to
Wash Solution Recycle
Fresh Wash
Feed
Soil Feed
Water
Feed
Outlet Liquid
Figure 1. Integrated soil treatment process.
cost-effectively remediate soil contaminated with these
wood treating wastes. Soil washing facilitates degrada-
tion by mobilizing the target compounds and expanding
the range of feasible remediation technologies (3). To
reduce costs and the volume of PCP-bearing liquid, the
soil wash liquid is concentrated via distillation, and the
recovered ethanol and water is recycled to the first soil
washing unit. The remainder of the wash solution is
initially bioremediated in an anaerobic environment.
Mineralization of the target compounds is completed
aerobically (4). Process development began by inde-
pendently evaluating soil washing and target compound
bioremediation. PCP-contaminated soils were the initial
focus, but this work is currently being extended to in-
clude soils contaminated with both PCP and PAHs.
Based on preliminary results, the integrated process will
meet the target cleanup level in 73 to 55 percent of the
RODs directed towards PCP remediation, resulting in
soil with a residual PCP level of 8 to 13 mg/kg, respec-
tively (2).
Soil Washing/Solvent Extraction Studies
An equimass (50 percent) mixture of ethanol and water
(5) was found to be the optimal solution to remove PCP
from a variety of spiked soils in a bench-scale soil
washing process. This soil washing method removes
PCP at levels comparable with those achieved through
the analytical techniques of sonication and soxhlet ex-
traction. Starting with initial spike levels of 85 to 100
mg/kg, 70 to 100 percent of the PCP added to the soil
was removed by washing, depending on soil particle
size, contamination age, and soil washing format. PCP
is extracted from soil in a 30-min contact time. The
availability of residual PCP on soils of 20 x 40, 100 x
140, and greater than 200 U.S. meshsize has been
evaluated through a serial procedure: soils were
66
-------
washed with 50 percent ethanol solution, rinsed with
water, and finally treated by soxhlet or sonication extrac-
tion using methanol/methylene chloride. Less than 4
percent of the residual PCP (less than 0.6 mg/kg) was
removed from the soil by the final sonication or soxhlet
extraction, demonstrating the limited availability of the
residual PCP. The solvent washing of soil with mixtures
of water and ethanol is also being investigated for PAH-
contaminated soils, using four compounds on the U.S.
Environmental Protection Agency's list of priority pollut-
ants as model compounds: naphthalene, acenaphthene,
pyrene, and benzo(b)fluoranthene (6). A more ethanol-
rich mixture may be required to effectively mobilize PAHs
from soil.
A sequential soil washing train is being optimized for
PCP-contaminated soils in which ex situ soil washing is
performed with the 50 percent ethanol solution in three
batch-wash stages. After washing the soil for 30 min in
each stage, the washed soil is recovered from the soil-
solvent slurry via vacuum filtration of the slurry, and a
fresh batch of solvent is added to the soil in the next
stage. Preliminary design data indicate that a series
of three 20-mL solvent washes will clean 5 g of soil
(1:12 soil:solvent ratio) as effectively as a single ex-
traction of 100 ml cleans 1 g of soil (1:100 soil:solvent
ratio). Additional optimization will further decrease the
soil:solvent ratio.
Biological Treatment Studies
In the integrated process, the distillate bottoms will be
fed to an anaerobic fluidized-bed granular activated
carbon (GAG) reactor. Two of these reactors were con-
structed and operated for over 40 months, evaluating
variables such as PCP loading and reactor empty bed
contact time (EBCT) (7). The reactor volume is 10 Lwith
a 1-L recycle loop. Based on this evaluation, the follow-
ing optimal operating variables were identified: EBCT,
2.3 hr; ethanol loading, 33.3 g/day (loading rate 6.3 g
chemical oxygen demand/L day); and PCP loading,
4.8 g/day (loading rate 0.55 g/L day). When the GAG
reactor operated at an EBCT of 2.3 hr, on a molar basis,
greater than 99.97 percent of the influent PCP was
dechlorinated to monochlorophenol (MCP). In addition,
data from the extraction of the reactor GAG during the
operating period indicated negligible accumulation of
PCP on the surface of the GAG. An aerobic fluidized-bed
GAG reactor will polish the effluent from the anaerobic
GAG reactor to attain complete mineralization of PCP.
Operation of the aerobic reactor has recently been
initiated.
An additional anaerobic fluidized-bed GAG reactor has
been constructed to evaluate the biotreatment of chemi-
cally synthesized solutions of the four PAHs and PCP in
ethanol. Greater than 99 percent transformation of the
influent PCP concentration of 100 mg/L has been
achieved in the reactor, while operating the reactor with
an EBCT of 9.3 hr. The reactor effluent data represent
greater than 99 percent removal for naphthalene, ace-
naphthene, and pyrene, and greater than 90 percent
removal for benzo(b)fluoranthene in the reactor.
References
1. U.S. EPA. 1994. Innovative treatment technologies: Annual status
report, 6th ed. EPA/542/R-94/005. Cincinnati, OH.
2. U.S. EPA. 1992. Contaminants and remedial options at wood pre-
serving sites. EPA/600/R-92/182. Cincinnati, OH.
3. U.S. EPA. 1990. Soil washing treatment. Engineering bulletin.
EPA/540/2-90/017. Cincinnati, OH.
4. Khodadoust, A.P., J.A. Wagner, M.T. Suidan, and S.I. Safferman.
1994. Solvent washing of PCP contaminated soils with anaerobic
treatment of wash fluids. Water Environ. Res. 66:692.
5. U.S. EPA. 1986. Microbial decomposition of chlorinated aromatic
compounds. EPA/600/2-86/090. Cincinnati, OH.
6. Keith, L.H., and WA. Telliard. 1979. Priority Pollutants I: A per-
spective view. Environ. Sci. Technol. 13:416.
7. Wilson, G.J., J.A. Wagner, A.P. Khodadoust, M.T. Suidan, and R.C.
Brenner. 1994. The evaluation of empty bed contact time on the
biodegradation of pentachlorophenol using an anaerobic GAC
fluidized-bed. In: Proceedings of the National Conference on En-
vironmental Engineering: Critical Issues in Water and Wastewater
Treatment. American Society of Civil Engineers, p. 624.
67
-------
Biological Treatment of Contaminated Soils Using Redox Control
Margaret J. Kupferle
Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, Ohio
Gregory D. Sayles
National Risk Management Research Laboratory, U.S. Environmental Protection Agency,
Cincinnati, Ohio
Tiehong L. Huang, Yonggui Shan, Maoxiu Wang, and Guanrong You
Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, Ohio
Carolyn M. Acheson
National Risk Management Research Laboratory, U.S. Environmental Protection Agency,
Cincinnati, Ohio
Introduction
Land treatment is a well-understood, cost-effective
means of conducting aerobic biological treatment of
soils contaminated with aerobically biodegradable com-
pounds, such as petroleum. Common contaminants in
soil also include highly chlorinated organics that are not
readily biodegraded aerobically, such as pentachlo-
rophenol (PCP), polychlorinated biphenyls (PCBs), and
1,1,1-trichloro-2, 2-bis(p-chlorophenyl)ethane (DDT).
These compounds may, however, be efficiently de-
graded using a sequential anaerobic/aerobic treatment
strategy. A cost-effective process to treat soils con-
taminated with these highly chlorinated contaminants
is needed. A modified type of prepared-bed land treat-
ment that incorporates variable redox states (i.e., an-
aerobic and aerobic phases) is being evaluated in this
project. The first pilot-scale study, using PCP-con-
taminated soil from the American Wood Products site
in Lake City, Florida, is in progress at the U.S. Envi-
ronmental Protection Agency's Test and Evaluation
(T&E) Facility in Cincinnati, Ohio.
Methodology
The pilot-scale study is being conducted in soil pans that
simulate a prepared-bed land treatment unit with perme-
ate collection. Each Plexiglas enclosure contains four
soil pans suspended in a controlled-temperature water
bath maintained at 20 + 2°C. Each soil pan (13 in. x
13 in.) is loaded with contaminated soil to a depth of
8 in. over a graded gravel underdrain system, which is
separated from the soil layer by a coarse mesh stainless
steel screen. During the anaerobic phase, permeate
recycle is optional from the underdrain to the top of the
pan at flowrates of 5 to 20 mL/min. Anaerobic conditions
are maintained in the soil by flooding the pans with clean
creek water from the site. Aerobic conditions are pro-
duced and maintained in the soil by draining the water
and tilling the soil in a manner consistent with landfarm-
ing techniques.
Initial Anaerobic Phase
Based on a review of the literature, which is dominated
by studies evaluating spiked soils, three variables were
selected for study in the initial anaerobic phase: 1) per-
meate recycle vs. no recycle; 2) addition of a supple-
mental organic source (ethanol or anaerobic sewage
sludge); and 3) soil PCP contamination level. The ex-
perimental design was a three-factor analysis of vari-
ance with replication. The amounts of soil, ethanol,
anaerobic digester sludge (32.8 g dry solids/L, 60 per-
cent volatile solids), and site water initially added to each
of the 24 pans are summarized in Table 1. In situ oxida-
tion-reduction potential (ORP) probes were placed in
one pan representative of each of the 12 treatments.
Four probes were buried in each pan, two in opposite
corners approximately 1 in. from the soil/gravel interface
and two in the remaining corners approximately 1 in.
from the soil/water interface near the top of the reactor.
68
-------
Recycle flow rates were maintained at 8 to 10 mL/min.
The pH of the water flooding the pans was measured
in situ each week. The soil in the pans was sampled and
analyzed for PCP and its less chlorinated phenolic me-
tabolites, hydrocarbons, and percent moisture on a
monthly basis. Hydrocarbons were present in the site
soil in significant quantities (1,000 to 2,500 mg/kg dry
soil) because diesel fuel serves as a carrier for PCP in
wood treating operations. The water flooding the pans
was analyzed for PCP and less chlorinated phenolic
metabolites each time a soil sample was collected.
Aerobic Phase
After 6 months, all of the pans except four of the sludge-
amended pans (2D, 3A, 4B, and 5C) were converted to
aerobic conditions. Afterthe free waterwas drained from
the soil, the soil was tilled three times a week for 4
weeks, until it dried to less than 10 percent total mois-
ture. When the conversion phase was completed, some
of the pans were continuously supplied with air at a low
flow rate (in addition to weekly tilling of all pans) and/or
amended with poultry manure (see Table 1). Each
month, the soil in the pans is sampled and analyzed for
PCP and its less chlorinated phenolic metabolites, as
well as hydrocarbons. Moisture content and water addi-
tion volumes are measured weekly to maintain moisture
content in a constant range.
Results
Anaerobic conditions were established in the soil pans
after the first week. The measured ORPs ranged be-
tween -150 to -500 mV (versus Ag/AgCI reference elec-
trode). No apparent correlation was found with respect
to probe depth, soil type, or treatments. The soil sample
PCP concentration data showed no significant amount
of PCP removal in any of the treatments after 6 months.
Changes in PCP concentration in the flood water of
several of the pans were noted, however. After2 months,
the PCP concentration in the flood water of the soil pans
containing sludge dropped from 15 to 55 mg/L to less
than 0.5 mg/L. After 4 months, the PCP concentration
also dropped to less than 0.5 mg/L in the two replicate
pans with low-PCP soil treated without recycle or sup-
plemental organic source. The less chlorinated phenolic
metabolites were not detected as intermediates in the
flood water from any of these pans. No degradation of
hydrocarbons was noted in any of the pans in the initial
anaerobic phase, but degradation has occurred in the
aerobic phase. The presence of hydrocarbons may have
interfered with the bioavailability of PCP in the initial
anaerobic phase. To test this hypothesis, aerobically
treated soil will be reconverted to anaerobic conditions
once the hydrocarbon concentration has been reduced.
Another possibility is that appropriate anaerobic PCP
degraders are not present in sufficient quantities in the
soil pans. A bench-scale study using soil from the
same source as the pilot-scale study has been initi-
ated to investigate the effect of amendment with PCP-
acclimated culture.
Conclusions
Adaptation of the pilot-scale land-treatment units to an-
aerobic operation has been evaluated. Flooding the soil
with water successfully creates a low redox (anaerobic)
state. The in situ ORP probes constructed for the project
work well. Monthly sampling intervals and the analytical
techniques used adequately characterize system be-
havior. The tilling strategy used in the conversion from
anaerobic to aerobic operation was successful. The
presence of significant amounts of hydrocarbon co-
contamination may have affected PCP degradation,
suggesting that a more appropriate treatment sequence
may be aerobic-anaerobic-aerobic. This observation re-
inforces the importance of technology evaluation with
soils characteristic of those found at actual sites.
69
-------
Table 1. Soil Pan Operation Summary
Pan3
6B
1C
2B
6Cf
1Af
3B
3C
2Cf
4C
5Cf
4Bf
6D
2A
1Df
5A
4Df
3Df
5B
6A
5Df
1B
2Df
3Af
4A
Soil
Lowd
36
36
36
36
36
36
36
36
36
36
36
36
-
-
-
-
-
-
-
-
-
-
-
(kg as is)
High6
-
-
-
-
-
-
-
-
-
-
-
-
35
35
35
35
35
35
35
35
35
35
35
35
Initial Anaerobic Phase Treatments'3
Ethanol Sludge Water
(mL) (L) (L)
16
16
16
16
17.2 - 16
17.2 - 16
17.2 - 16
17.2 - 16
5.16 11
5.16 11
5.16 11
5.16 11
16
16
16
16
17.2 - 16
17.2 - 18h
17.2 - 16
17.2 - 16
4.88 12
4.88 12
4.88 10.5h
4.88 12
Aerobic Phase
Treatments0
Recycle
(mL/min)
-
-
8-10
8-10
-
-
8-10
8-10
8-1 Og
-
8-1 Og
-
-
-
8-10
8-10
-
-
8-10
8-10
-
-
8-1 Og
8-1 Og
Air
(mL/min)
-
10
-
10
-
10
-
10
-
Not
Not
10
-
10
-
10
-
10
-
10
-
Not
Not
10
Manure
(g/pan)
-
750
750
-
-
750
750
-
-
converted
converted
-
-
750
750
-
-
750
750
-
-
converted
converted
Pan location for treatments (Pans A-D in Boxes 1-6) randomly assigned for statistical purposes.
b Initial anaerobic phase from August 11, 1994, to March 2, 1995.
c Aerobic phase from March 2, 1995, to July 7, 1995.
d Soil from 5 ft depth at site containing approximately 250 mg PCP per kg dry soil.
e Soil from 12 ft depth at site containing approximately 650 mg PCP per kg dry soil.
f In situ ORP probes added to pan during initial anaerobic phase.
g Recycle was set at 9 mL/min initially but was discontinued after the first week due to extremely low flow (less than 1 mL/min) through
sludge-amended soils.
h Amount of water required to maintain a constant depth of 2 in. above soil surface varied somewhat.
70
-------
Development of a Sulfate-Reducing Bioprocess To Remove Heavy Metals From
Contaminated Water and Soil
Munish Gupta and Makram T. Suidan
Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, Ohio
Gregory D. Sayles and Carolyn M. Acheson
U.S. Environmental Protection Agency, Cincinnati, Ohio
Introduction
Acid mine drainage is characterized by low pH (1.5 to
3.5) and high concentrations of sulfate and dissolved
heavy metals. Bacterial sulfate reduction has been iden-
tified as a potentially cost-effective process for removing
metals from mine drainage (1, 2). Sulfate-reducing bac-
teria convert sulfate to sulfide using an organic carbon
source as the electron donor. The sulfide precipitates
the various metals present in the wastewater, yielding a
very low concentration of dissolved metals in the efflu-
ent. In this study, acetic acid was used as the carbon
source for two reasons: it is relatively inexpensive, and
being an acid, it can effectively leach out metals from
contaminated soils such as mine tailings.
Reactor Selection
To effectively treat metal-contaminated wastewater, a
reactor must establish an anaerobic environment to sup-
port sulfate reduction, resulting in metal precipitation as
metal sulfides, and must provide an efficient clarifier to
remove metal precipitates from the effluent. Because
sludge (metal precipitates and biomass solids) would
accumulate and eventually clog the reactor, ease of
sludge removal or cleaning is an important considera-
tion in selecting a reactor. Two reactors were evaluated:
an upflow anaerobic filter packed with plastic Pall rings
and an anaerobic sludge blanket reactor. To clean the
reactor, sludge can be removed from the bottom of the
sludge blanket reactor. The same technique can be
used for the filter; however, it may be more difficult due
to packing material. The cleaning of the filter, therefore,
was an additional aspect of research.
Reactor Operation and Performance
Two filters (A and B) and one sludge blanket reactor
were operated at a temperature of 30°C and at a pH of
7.2, optimal pH for sulfate-reducing microorganisms.
The feed concentration of metals (shown in Table 1)
used in this study were among the highest concentra-
tions observed at mines in Montana and Colorado.
After an initial acclimation, Filter A was fed the metals
listed in Table 1, while Filter B and the sludge blanket
reactor were fed iron at a concentration equal to the
sum of the molar concentrations of all the metals fed
to Reactor A. Table 1 shows characteristic effluent
Table 1. Influent and Effluent Concentrations (mg/L) for Filter A
Effluent Concentration
Influent
Constituent
Iron
Zinc
Manganese
Copper
Cadmium
Lead
Arsenic
Acetate
Sulfate
Total sulfide
Influent
Concentration
840
650
280
130
2.3
2.1
1.5
3000
5000
0
Filtered
0.09
0.14
5.4
0.02
0.02
0.005
0.01
11.0
800
31
Total
0.497
0.310
5.60
0.022
0.019
0.005
0.01
71
-------
concentrations for Filter A, operating at a hydraulic
detention time of 5 days.
During this period, the level of sludge rose above the
packed bed in Filter A. To investigate whether sludge
withdrawal from the bottom would control the sludge
height in the filter, 1 L of sludge was removed. Sludge
withdrawal lowered the sludge level and the filter con-
tinued to operate efficiently, with less than 1 percent
change in effluent conditions. A similar situation in Filter
B was also corrected in the same fashion. Sludge with-
drawal from the bottom can, therefore, control the accu-
mulation of sludge and prevent clogging of the filters.
The sludge blanket reactor did not perform very well as
a clarifier. Although the soluble iron concentration in the
effluent was less than 0.25 mg/L, the total concentration
was as high as 25 mg/L and varied between 18 and
22 mg/L. This reactor had a high concentration of total
suspended solids in the effluent compared with Reac-
tors A and B. It was concluded that this type of reactor
was not effective in clarification and was unable to meet
the requirements. Therefore, the operation of the sludge
blanket reactor was discontinued.
Conclusions
Compared with an anaerobic sludge blanket reactor, an
upflow anaerobic filter packed with Pall rings was found
to be a very efficient reactor for the treatment of water
contaminated with heavy metals. The filter, unlike the
sludge blanket reactor, worked very well as a clarifier,
and all metals except manganese were reduced to a
concentration close to drinking-level standards. Sludge
withdrawal from the bottom of the filter can be used to
remove accumulating sludge, and, therefore, the filter
can be operated continuously. Ongoing work will evalu-
ate the performance of the filters as a function of hydrau-
lic retention time, lower temperatures, and pH. Sludge
removal frequency will also be optimized.
References
1. Dvorak, D.H., R.S. Hedin, H.M. Edenborn, and RE. Mclntire. 1992.
Treatment of metal-contaminated water using sulfate reduction:
Results from pilot-scale reactors. Biotech. & Bioeng. 40:609-616.
2. Kuyucak, N., D. Lyew, P. St. Germain, and K.G. Wheeland. 1991.
In situ bacterial treatment of AMD in open pits. Presented at the
Second International Conference on the Abatement of Acidic
Drainage, Montreal, Canada, September 16-18.
72
-------
Development of Techniques for the Bioremediation of Chromium-Contaminated
Soil and Ground Water
Michael J. Mclnerney, Nydia Leon, Veronica E. Worrell, and John D. Coates
University of Oklahoma, Norman, Oklahoma
The potential for biotic Cr(VI) reduction in samples from
a Cr(VI)-contaminated aquifer (Elizabeth City, North
Carolina) was evaluated by inoculating aquifer material
into anaerobically prepared mineral salts medium that
did not contain chemical reductant. In inoculated micro-
cosms, the Cr(VI) concentration decreased after 5 days
incubation at 25°C, and almost all of the Cr(VI) was gone
after 25 days (Figure 1). Little or no change in Cr(VI)
0 5
10 15 20 25 30 35 40 45 50 55
Time (days)
- uninoculated
- autaclaved+HgCI2
- autoclaved
- 2Xautoclaved+HgCI2 -
2X autoclaved
- nonsterile
Figure 1. Biological reduction of Cr(VI) with aquifer material.
concentration was observed in uninoculated controls, or
in sterile controls prepared by autoclaving, boiling, orthe
addition of HgCI2 or chloramphenicol. Hydrogen re-
duced the lag time before Cr(VI) reduction occurred but
did not markedly affect the rate of Cr(VI) reduction (Fig-
ure 2). The addition of other exogenous electron donors
such as glucose, acetate, formate, or benzoate did not
affect the rate or lag time associated with Cr(VI) reduc-
tion in microcosms compared with controls that lacked
an exogenous electron donor. The addition of phenol,
lactate, and ethanol to microcosms inhibited Cr(VI) re-
duction. Subsequent addition of Cr(VI) to microcosms
with benzoate as the electron donor decreased the lag
time and increased the rate of Cr(VI) reduced compared
with that observed initially.
10 15
Time (days)
--
sterile
-*-
no donor
-B~
acetate
--
hydrogen
Figure 2. Effect of electron donors on Cr(VI) reduction with
aquifer material.
The effect of sodium sulfate, sodium nitrate, amorphous
ferric hydroxide (each at 10 mM) on Cr(VI) reduction
was tested with benzoate-amended microcosms. The
presence of sulfate and nitrate inhibited the reduction of
Cr(VI) compared with microcosms that did not receive
any of the three additional electron acceptors. Sulfide
levels remained unchanged during the course of the
experiments. In bottles with nitrate, nitrite accumu-
lated after 8 days and decreased after 16 days. Ferric-
hydroxide-supplemented microcosms reduced Cr(VI)
to a much greater extent than unsupplemented control
cultures; ferrous iron production coincided with Cr(VI)
reduction.
Two facultative bacteria that can reduce Cr(VI) were
isolated from Elizabeth City aquifer material, and one
bacteria was isolated from an aquifer underlying a land-
fill in Norman, Oklahoma, using a mineral salts medium
with 5 mM benzoate, 500 u,M Cr(VI). All three isolates
are gram-negative, motile rods that grow singly, in pairs,
and in branched chains. On agar medium, the isolates
formed shiny, smooth, pink colonies and produced a
diffusible green pigment. Upon initial isolation, Cr(VI)
was rapidly reduced (Figure 3); however, the rate and
73
-------
Table 1. Electron Donors That Support Cr(VI) Reduction by
Strain NLB
10
15 20 25
Time (days)
+ benzoate
no donor
killed cells
Figure 3. Cr(VI) reduction by strain NLB.
extent of Cr(VI) reduction decreased with repeated
transfer of the culture in benzoate-Cr(VI) medium. The
use of Cr(VI) was dependent on the presence of an
electron donor and an active inoculum. In addition to
benzoate, other substrates supported Cr(VI) reduction
(Table 1). Increases in cell numbers were observed
when the electron donor and Cr(VI) were both pre-
sent. In the absence of Cr(VI) or electron donor, little
or no increase in cell number was observed: less than
6x 10s cells/ml.
Four observations supported the conclusion that the
decrease in Cr(VI) concentration was a biologically me-
diated reduction process: 1) Cr(VI) concentrations de-
creased faster and to a greater extent in nonsterile
versus sterile microcosms; 2) phenol, ethanol, and lac-
Additions
Fumarate
Purine/Pyrimidine mix3
p-Toluic acid
Lactate
Phenoxyacetate
Malate
Benzoate
Phenol
Ethanol
No addition
Cr(VI)
Reduced
(|iM)
414
337
184
169
166
80
68
59
32
0
Increase in Cell
Numbers
([cells/mL] x 106)
0
31.8
30.1
62.8
56.3
3.8
142.9
19.3
7.8
0
Mix contains 0.5 mM each: adenine, guanine, thymine, uracil.
tate inhibited Cr(VI) reduction in microcosms; 3) re-
peated additions of Cr(VI) to microcosms decreased the
lag time and stimulated the rate of Cr(VI) reduction; 4)
bacteria were isolated and capable of using Cr(VI) as an
electron acceptor. Iron hydroxide stimulated Cr(VI) re-
duction in microcosms, most likely by an indirect mecha-
nism involving the production of ferrous iron. The extent
and rate of Cr(VI) reduction by aquifer microcosms was
not affected when exogenous electron donors, with the
exception of hydrogen, were added. This indicates that
the aquifer material had sufficient levels of endogenous
electron donors to support Cr(VI) reduction.
74
-------
Bioremediation of Chlorinated Pesticide-Contaminated Sites Using Compost
James C. Young, Jean-Marc Bollag, and Raymond W. Regan
Pennsylvania State University, University Park, Pennsylvania
Many sites throughout the United States are contami-
nated with chlorinated pesticides. Of particular interest
to this project are those sites contaminated with chlor-
dane and toxaphene. One objective is to determine the
feasibility of using compost as a culture medium for
mediating the biodegradation of these pesticides. A sec-
ond objective is to determine major pathways of chlor-
dane and toxaphene biodegradation that lead to
mineralization. These objectives are particularly chal-
lenging because chlordane and toxaphene each consist
of several chlorinated cyclic hydrocarbons that individu-
ally may follow different biodegradation pathways, or
may be only partially dechlorinated.
Biodegradation of chlordane and toxaphene and other
chlorinated pesticides is expected to require an organic
co-substrate as a carbon source for the growth of accli-
mated microorganisms that enzymatically are capable
of dechlorinating the pesticides through reductive or
oxidative reactions. Co-substrates considered for use in
field applications include milk solids, sugar, blood meal,
sewage solids, methane, or, in the current project,
compost.
The test program includes the development and op-
eration of a pilot-scale compost reactor that contains
a mixture of 10 percent municipal-sludge compost,
10 percent spent-mushroom compost, 40 percent
grass, and 40 percent alfalfa hay to provide an environ-
ment suitable for the culture of chlordane- and toxaphene-
degrading microorganisms. This compost is used to amend
various contaminated-soil matrices followed by analysis of
the fate of the pesticide. Residual pesticides are moni-
tored using gas chromatography, thin layer chromatog-
raphy, and mass spectroscopy. Test parameters include
soil type, compost-soil ratio, moisture level, oxidation-
reduction potential, pH, presence of sulfates and ni-
trates, and the effect of supplemental soluble organic
co-substrates.
75
-------
Reductive Electrolytic Dechlorination
John W. Norton, Jr., and Makram T. Suidan
Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, Ohio
Carolyn M. Acheson and Albert D. Venosa
National Risk Management Engineering Laboratory, U.S. Environmental Protection Agency,
Cincinnati, Ohio
Reductive dehalogenation is the only known mechanism
to biologically degrade some highly chlorinated organic
compounds, including pentachlorophenol (PCP) (1),
and occurs primarily in anaerobic environments (2). A
biofilm-electrode reactor (BER) was constructed to
evaluate PCP dechlorination as a function of ethanol
concentration and the presence of an electrical current.
The BER was operated as follows: current, 20 mA;
hydraulic retention time, 0.38 days; PCP feed, 5 mg/L;
ethanol feed range, 0 to 100 mg/L. The best observed
dechlorination occurred when 5 mg/L PCP and 25 mg/L
ethanol were fed to the reactor. The effluent under these
conditions contained 0.013 mg/L PCP, 0.26 percent of
the feed concentration. At lower ethanol levels, PCP was
not as effectively dechlorinated. The trichlorophenols (TCP)
and dichlorophenols (DCP) displayed a two- to three-
fold increase in effluent concentration as the substrate
ethanol was decreased, particularly at concentrations less
than 10 mg/L. The monochlorophenols (MCP), however,
reached a maximum of 0.014 mM at ethanol concentra-
tions of 10 to 25 mg/L. The total dechlorination de-
creased significantly when the ethanol was removed
from the feed, indicating that the ethanol stabilized
dechlorination.
After characterizing the ethanol requirements in the sys-
tem, the role of the current in the dehalogenation of the
PCP was evaluated by turning off the current. Electrical
current was shown to play a necessary role in dechlori-
nation, although it is unknown whether this role was the
result of the hydrogen generation or the low reducing
potential surface formed on the cathode of the anode-
cathode cell. Two trials of reactor operation without cur-
rent were conducted. Each current removal trial caused
a reduction in PCP dehalogenation, demonstrated by
the successive appearance of the higher chlorinated
phenols in the effluent. The two trials displayed very
different temporal behavior, however. The first removal
resulted in a quick rise in effluent PCP concentrations,
increasing one order of magnitude in a few hours. Fol-
lowing the second current removal, a much slower ap-
pearance of chlorinated phenols was observed; the
effluent PCP concentration increased one order of mag-
nitude in approximately 6 days. During each trial, after
the current was reapplied, the system recovered. Each
trial showed a recovery pattern similar to the failure
preceding it, the first trial showing a quick recovery and
the second trial showing a much slower recovery. The
causes of the different behaviors have not been charac-
terized. We are presently evaluating the role of electrical
current by varying the current while keeping the ethanol
concentration constant at 10 mg/L.
References
1. Mohn, W, and J. Tiedje. 1992. Microbial reductive dehalogenation.
Microb. Rev. 56:482-507.
2. Suflita, J., A. Horowitz, D. Shelton, and J. Tiedje. 1982. Dehalo-
genation: A novel pathway for the anaerobic biodegradation of
haloaromatic compounds. Science 218:1,115-1,117.
76
-------
Biological Ex Situ Treatment of PAH-Contaminated Soil
Carl L. Potter
U.S. Environmental Protection Agency, Cincinnati, Ohio
Roy C. Haught
IT Corporation, Cincinnati, Ohio
The goal of this project is to evaluate the potential of
biological ex situ soil treatment systems (biopiles) to
remediate soils contaminated with hazardous chemi-
cals. A laminar-type flow pilot-scale reactor with a vol-
ume of 3 yd3 has been constructed at the U.S.
Environmental Protection Agency's Test & Evaluation
(T&E) Facility in Cincinnati, Ohio. Laminar-type flow
from one side of the reactor to the other may provide
even aeration to all areas of the reactor while avoiding
the use of pipes inside the reactor. This design greatly
facilitates loading and unloading of the reactor and is
readily scalable to larger systems.
Passing smoke through the reactor for visual observa-
tion of flow indicated uniform, laminar-type flow through
the empty reactor. Further testing involved filling the
reactor with vermiculite or a synthetic soil, flushing with
argon, and then passing airthrough the reactorto evalu-
ate air flow through this uniform solid matrix. Oxygen
probes, located at 27 positions within the reactor, indi-
cated rapid and uniform air saturation of the system.
Analysis of gas flow through an empty reactor and
through uniform matrices allowed evaluation of reactor
performance without confounding effects of soil inhomo-
geneities that may lead to nonuniform aeration of the
reactor space.
The reactor uses pulsed air flow through the pile to
permit maximum distribution of air within the soil. Air is
driven into the soil during pulse action, then allowed to
diffuse in all directions during the rest interval.
Soil contaminated with polynuclear aromatic hydrocar-
bons (PAHs) from the Reilly Tar Pit Superfund site in St.
Louis Park, Minnesota, has been brought to the T&E
Facility for research on soil aeration and effectiveness
of this ex situ reactor design for biological treatment of
contaminated soils. Micronutrients were adjusted to
100:20:1 phosphorus:carbon:nitrogen, and 0.5 percent
by weight cow manure was added to the soil. A10-week
treatability study is under way to evaluate disappear-
ance of parent PAHs and microorganism population
changes in this reactor system.
77
-------
Effectiveness of Gas-Phase Bio re mediation Stimulating Agents (BSAs) for
Unsaturated Zone In Situ Bioremediation
James G. Uber, Ronghui Liang, and R. Scott Smith
Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, Ohio
Paul T. McCauley
Water and Hazardous Waste Treatment Research Division, National Risk Management Research
Laboratory, U.S. Environmental Protection Agency, Cincinnati, Ohio
Background
Successful in situ bioremediation in the unsaturated
zone requires that water, oxygen, nutrients, primary sub-
strate, and perhaps co-metabolites be available to the
microorganism via physical transport mechanisms. Any
of these substances may be called a bioremediation
stimulating agent (BSA), given that a shortage of any
one may adversely affect the performance of an in situ
bioremediation system. Other potential BSAs include
substances (e.g., surfactants) that are not ordinarily re-
quired for microbial growth but that may enhance sub-
strate or nutrient bioavailability.
Much work has focused on engineering approaches to
deliver BSAs at field scale. Little research has been
conducted, however, to evaluate which in situ delivery
approaches are best for transporting BSAs to microor-
ganisms. Given the complexity of two-phase (gas/water)
or three-phase (gas/water/nonaqueous-phase liquid)
fluid and contaminant transport in the unsaturated zone,
considerable uncertainty exists about the ultimate distri-
bution of BSAs in contaminated soils. Further, microbial
growth processes affect fluid and contaminant transport
not only through biochemical reactions but also through
a spatial-temporal influence on fluid permeability. (Plug-
ging of pore spaces by microorganisms can reduce
wetting fluid permeability by greater than 99 percent.)
The present study will identify in situ BSA delivery strate-
gies that are most likely to achieve a uniform BSA spatial
distribution and, therefore, most likely to improve biore-
mediation field performance. As a byproduct of this
work, the project aims to identify and measure the fun-
damental physical and microbial processes that affect
bioremediation performance enhancement through BSA
delivery methods.
Because of strong capillary forces that affect the distri-
bution and movement of wetting fluids in unsaturated
soils, gas-phase BSAs are more likely to achieve uni-
form in situ spatial distribution. It is, in fact, well known
that movement of water in the unsaturated zone often
occurs in discrete fingers that occupy a small fraction of
the total pore space. Relatively little is known, however,
about the characteristics of in situ gas-phase BSA trans-
port, including physical factors that may lead to complex
and undesirable flow patterns (e.g., interactions of water
saturation and air permeability), chemical transport fac-
tors that may limit gas-phase BSA spatial distribution
(e.g., BSA solubility), and dynamic microbial factors that
may affect BSA transport in the field (e.g., microbial BSA
utilization rates and plugging that leads to heterogene-
ous effects on air/water permeability). Because of the
promise of gas-phase BSAs and the significant un-
knowns regarding their effectiveness, this project will
focus on effective gas-phase addition of nutrients, co-
metabolites, oxygen, and moisture.
Objectives
The objectives of this work are to:
Evaluate the effectiveness of field systems for gas-
phase delivery of BSAs to the unsaturated zone for
enhancing in situ bioremediation performance. These
BSAs include nutrients (organophosphates), co-
metabolites, surfactants or solvents, and water vapor.
Identify and measure the physical and microbial fac-
tors affecting the bioavailability of gas-phase BSAs
in the unsaturated zone, including uneven spatial dis-
tribution of BSAs at the pore- and core-scales and
complex changes in unsaturated zone air permeabil-
ity caused by microbial-growth dynamics.
78
-------
Develop visible light tomography (VLT) systems that
allow visualization of in situ unsaturated zone physi-
cal and microbial processes for controlled evaluation
of alternative BSAs and delivery systems.
Use the VLT systems to evaluate alternative BSAs
for remediation of aged contaminated soils in control-
led but realistic environments.
Accomplishments
We have designed and constructed two different labo-
ratory systems for observing dynamic fluid distribution
in the unsaturated zone under simulated BSA delivery.
A three-dimensional column system has been designed
to collect data on fluid migration through discrete fingers
in disturbed and undisturbed soil cores, and will be used
to measure the limitations on BSA delivery caused
by fingering under a variety of soil conditions and
fluid application rates. This work will be completed
by December 1995. The columns are 30 cm in diameter
and comprised of stackable 10-cm sections separated
by 1-mm spacers. The columns rest on a base that
allows manipulation of the bottom pressure boundary
condition, and the side boundary condition is manipu-
lated through 1-mm gaps between rings. Water is ap-
plied uniformly to the top surface of the soil columns via
a carefully designed air-atomizing nozzle. After the fluid
flow is developed, a dye mixture marks the locations of
any preferential flow pathways. The pathways will be
exposed at the surfaces of each 10-cm ring, and the
complete three-dimensional character of each pathway
will be recorded. Different color dyes will be used to
investigate the persistence of individual fingers when
fluid application is cycled, allowing the soil to drain to
varying water contents between application.
A two-dimensional, vertical, thin-slab visible light to-
mography (VLT) system has been designed to visualize
and measure the interactions between gas-phase BSA
and liquid-phase flow in a controlled environment. The
system will also serve as a bioremediation simulator to
measure the effectiveness of various gas-phase BSAs,
and to visualize dynamic microbial-growth processes
under simulated in situ bioremediation conditions (with
and without BSA addition). The vertically oriented cham-
ber dimensions are 1 m x 2 m x 1 cm. The top boundary
will be either open or closed to the atmosphere, and will
be capable of having controlled amounts of liquid added
uniformly over the slab length. Side boundaries will be
either closed or open to the atmosphere, thereby provid-
ing the ability to control gas-phase BSA injection or
extraction (simulating the operation of BSA injection or
extraction wells). The bottom boundary will be either
open to the atmosphere or, via a manifold, will simulate
water table conditions.
The advantage of the thin-slab system is the ability to
visualize the complex flow and microbial processes oc-
curring in the unsaturated zone under simulated in situ
conditions. A bank of high frequency fluorescent lamps
will illuminate the system from the back. Because light
transmission is related to water saturation, the water
distribution can be easily visualized without the use of
dyes. Fluorescent gases will be investigated for visuali-
zation of the gas movement, as will color-marked gas-
and liquid-phase pH indicator solutions. Data will be
recorded via a CCD camera and a data acquisition
system so that actual fluid flow and microbial processes
can be recorded and visualized. The CCD camera sys-
tem will collect data at a spatial resolution on the order
of the pore scale (approximately 0.5 to 1.0 mm).
79
-------
Section 5
Process Research
Process research involves isolating and identifying microorganisms that carry out biodegradation
processes and the environmental factors affecting these processes. It also deals with the develop-
ment of techniques for modeling and monitoring biodegradation. Through this research, scientists
establish the building blocks of new biosystems for treatment of pollutants in surface waters,
sediments, soils, and subsurface materials. Thorough evaluation is critical at this level of research
because a firm scientific foundation can facilitate the scaling up of a promising technology. Process
research is being conducted on many environmental pollutants.
Two research projects quantified the extent of biodegradation of organic compounds using carbon
isotopes. The methods employed in this research can be used both in the laboratory, where target
compounds are isotopically labeled, and in the field, where contaminants contain different concen-
trations of isotopes than the surrounding environment. The methods were applied in the measure-
ment of petroleum degradation and the demethylation of organometallic compounds.
Another project determined the kinetic rate constants of anaerobic degradation of quinoline by
methanogenic bacteria. This system is potentially of great practical significance because quinoline
is found in wastes at wood preserving plants.
Several papers focused on the use of microorganisms to degrade alkyl halides and polychlorinated
aromatics. Research included the characterization of a bacterial enzyme (toluene 2-monooxy-
genase) that can degrade trichloroethylene and of a strain of Pseudomonas cepacia G4 that
expresses the enzyme constitutively. Further research covered the effects of varying environmental
conditions on the bioremediation of chlorinated compounds.
Other process research projects studied heavy metal inhibition of the bioremediation of polychlori-
nated aromatics and the effects of different primary substrates on the reduction of 2,4-dinitrotoluene.
Primary substrates are used to sustain the growth of microorganisms when the target contaminant
cannot act as a food source.
Several poster presentations at the symposium involved process research. The presentations
covered methods for monitoring bioremediation, the use of surfactants in sediments, and charac-
terization of an anaerobic dehalogenating microorganism. Another presentation dealt with the
potential of Mycobacterium to mineralize polycyclic aromatic hydrocarbons.
81
-------
Monitoring Crude Oil Mineralization in Salt Marshes:
Use of Stable Carbon Isotope Ratios
Andrew W. Jackson
Department of Civil and Environmental Engineering, Louisiana State University,
Baton Rouge, Louisiana
John H. Pardue
Wetland Biogeochemistry Institute, Louisiana State University, Baton Rouge, Louisiana
Introduction
The ability to monitor mineralization of hydrocarbons is
of prime importance in a successful remediation strat-
egy. Hydrocarbon mineralization must be ensured, be-
cause hydrocarbons can be sorbed, transformed, or
buried, or otherwise be undetected but still pose threats
to the existing system ecology. One successful tech-
nique has been monitoring changes in oil composition
relative to a stable, nondegradable compound (1, 2).
Two disadvantages to this method exist, however: 1) its
inability to demonstrate mineralization instead of trans-
formation, and 2) its inability to measure absolute oil
degradation, because only "resolved" compounds are
quantified.
A promising new technique for the detection and quan-
tification of hydrocarbon mineralization is the use of
stable carbon isotope ratios (3). Carbon dioxide gas
ratios vary from of 12C to 13C depending on the source
of the gas. Crude oils are more depleted in 13C, and thus
the mineralization of oil produces CO2 with lower 8-13C
values. Oil has a 8-13C value of-29 to -32 (0/00) depend-
ing on the source of the oil. Salt marshes are predomi-
nantly colonized by C3 plants, and CO2 evolved from
these soils has 8-13C ratios of-14.4 to -17.7 (0/00) (4).
If biodegradation is occurring in a contaminated salt
marsh, the 8-13C value of the produced CO2 should
decrease due to the presence of 13C-depleted CO2 from
the crude oil. If this occurs, it would be possible to qualify
and quantify hydrocarbon degradation by measuring
total CO2 production and changes in the 13C signature
of CO2 produced from the marsh.
Theoretical
The rate of CO2 produced from each carbon source can
be easily computed using three equations describing
CO2 production and the 8-13C signature:
R0+Ri=R, (Eq. 1)
RO R, (Eq. 2)
(Eq. 3)
where R0 and R| are the rates of CO2 production from
the crude oil and indigenous carbon sources respec-
tively, and R, is the total rate of CO2 production. S0 and
S| are the 8-13C signatures of the crude oil and indige-
nous carbon, and S, is the measured 8-13C signature of
the produced CO2. S0,S|, St, and R, are experimentally
determined. R0 and Rm can then be determined from
Equations 2 and 3. This assumes that CO2 is generated
from only these two carbon pools and that there is no
addition of atmospheric CO2.
Results and Discussion
Kinetic Experiments
Microcosm studies showed rapid and nearly complete
(greater than 90 percent reduction in the hopane ratio)
degradation of parent alkanes in the fertilized treatments
under completely mixed, aerated conditions. In the un-
fertilized treatment, less than a 10 percent reduction was
observed in the hopane ratio of the alkanes (Figure 1 A).
Polycyclic aromatic hydrocarbon (PAH) degradation
83
-------
900
tn
g
'
or
0)
(0
Q.
o
X
I I
CL
-g- 3 -
O
-*?
£ 2"
X.
1 1 -
9
S 0
0 °
Fertilized (1)
Fertilized (2)
BH Unfertilized (1)
^B Unfertilized (2)
I
m \
_ 1 1 !
1 a
Jjj
||
1
2345
Time (Weeks)
Figure 3. Rates of CO2-C mineralized from crude oil in fertilized
and unfertilized salt marsh soils (calculated from iso-
tope dilution equations).
84
-------
threne, however, appear to be stable over the period of
this experiment.
The CO2 production rates and the 8-13C ratios measured
were used to calculate the CO2 produced from crude oil
(Figure 3). No mineralization of crude oil was detected
until Week 2, and the majority of mineralization appears
to begin at Week 5. The fertilized treatments appeared
to show higher mineralization rates before the unfertil-
ized and to mineralize at a more even prolonged rate.
The unfertilized treatments have a more intense rate
of mineralization but for only one sampling date.
Amendments of fertilizer inconclusively increased deg-
radation, as evidenced by hopane ratios of specific oil
components.
The importance of the 8-13C data is the data's ability to
calculate mineralization rates directly. They measure the
final product, while monitoring hopane ratios only meas-
ures the disappearance of the parent compound, not
mineralization. These experiments support the ability to
use 8-13C ratios in conjunction with CO2 production to
qualitatively and quantitatively monitor crude oil degra-
dation.
References
1. Bragg, J.R., R.C. Prince, E.J. Harner, and R.M. Atlas. 1994. Nature
368:413-418.
2. Bragg, J.R., R.C. Prince, E.J. Harner, and R.M. Atlas. 1993. In:
Proceedings of the 1993 International Oil Spill Conference, pp.
435-447.
3. Aggarwal, P.K., and R.E. Hinchee. 1991. Environ. Sci. Technol.
25:1,178-1,180.
4. Chmura, G.L., R.A. Socki, and R. Abernethy. 1987. Oecologia
74:264-271.
85
-------
Mercury and Arsenic Biotransformation
Ronald S. Oremland
U.S. Geological Survey, Menlo Park, California
This presentation will cover our recent findings with
regard to bacterial processes affecting 1) methylmer-
cury demethylation and 2) the dissimilatory reduction of
arsenic (As) (V).
Methylmercury Oxidative Degradation
Potentials in Contaminated and Pristine
Sediments of the Carson River, Nevada
Sediments from mercury-contaminated and uncontami-
nated reaches of the Carson River, Nevada, were
assayed forsulfate-reduction, methanogenesis, denitri-
fication, and monomethylmercury (MeHg) degradation.
Demethylation of 14C-MeHg was detected at all sites, as
indicated by the formation of 14CO2 and 14CH4. Oxidative
demethylation was indicated by the formation of 14CO2
and was present at significant levels in all samples.
Oxidized/Reduced demethylation product (ORDP) ra-
tios (e.g., 14CO2/14CH4) generally ranged from 4.0 in
surface layers to as low as 0.5 at depth. Production of
14CO2 was most pronounced at sediments surfaces that
were zones of active denitrification and sulfate-reduc-
tion, but was also significant within zones of methano-
genesis. In a core taken from an uncontaminated site
having more oxidized, coarse-grained sediments, sul-
fate-reduction and methanogenic activities were very
low, and 14CO2 accounted for 98 percent of the product
formed from 14C-MeHg. No relationship was apparent
between the degree of mercury contamination of the
sediments and the occurrence of oxidative demethyla-
tion. Sediments from Fort Churchill, the most contami-
nated site, however, were most active in terms of
demethylation potentials. Inhibition of sulfate reduction
with molybdate resulted in significantly depressed
ORDP ratios, but overall demethylation rates were com-
parable between inhibited and uninhibited samples.
Addition of sulfate to sediment slurries stimulated
production of 14CO2 from 14C-MeHg, while 2-bro-
moethane- sulfonic acid blocked production of 14CH4.
These results reveal the importance of sulfate-reducing
and methanogenic bacteria in oxidative demethylation
of MeHg in anoxic environments.
The Dissimilatory Reduction of As(V) to
As(lll) in Anoxic Sediments and as an
Electron Acceptor for Growth of Strain
SES-3
Anoxic sediment slurries amended with millimolar levels
of As(V) achieved a complete reduction of this oxyanion
to As(lll) upon incubation. As reduction was enhanced
when slurries were provided with the electron donors H2,
lactate, or glucose, although no effect was achieved with
acetate or succinate. Aerobically incubated slurries did
not reduce As(V), nor did formalin-killed or autoclaved
controls. Even though acetate did not stimulate As re-
duction, the oxidation of 2-14C-acetate to 14CO2 in an-
oxic slurries could be coupled with the abundance of
As(V). The selenium (Se) (VI) respiring anaerobe strain
SES-3 was found to be capable of achieving growth by
carrying out the dissimilatory reduction of As(V) to
As(lll). Although growth parameters were meager (e.g.,
Ym = 0.53 g cells/mole lactate; maximal cell density =
9.2 x 107 cells/ml), the ability to reduce As(V) to As(lll)
was constitutive and occurred rapidly in either selenate-
or nitrate-grown cells. These results suggest that the
reduction of As(V) to As(lll) in nature may be achieved
by bacteria-like strain SES-3 carrying out dissimilatory
As(V) reduction.
86
-------
Monod Degradation Kinetics of Quinoline in Natural and Microbially Enriched
Methanogenic Microcosms
E. Michael Godsy, Ean Warren, and Barbara A. Bekins
U.S. Geological Survey, Menlo Park, California
Introduction
The methanogenic biodegradation of nitrogen-contain-
ing heterocyclic compounds found in wastes from petro-
leum refineries, coke operations, coal tar production,
and wood preservation has not been studied in detail.
Quinoline, the largest single component in creosote (1),
is first oxidized to 2(1 H)-quinolinone, which is then de-
graded to CH4 and CO2. In this study, the Monod no-
growth kinetic constants for the oxidation of quinoline
and the Monod kinetic constants for the methanogene-
sis of 2(1 H)-quinolinone were determined under natural
and microbially enriched methanogenic conditions using
nonlinear regression analysis (2). In microcosms simu-
lating natural aquifer conditions, it was necessary to
model the oxidation and subsequent methanogenesis
independently (1). In microbially enriched microcosms,
the two must be coupled to include the biomass in-
crease from the methanogenesis of 2(1H)-quinolinone
as below:
dQ
dt ''
Kn+Q
Qn
dQn
_ _
dt Kn+Q Y(Ks+Qn)
dX
~df~~
U,max*a
Ks+Qn
where
Q =
Umax
Y =
quinoline, mg-L"1
no-growth oxidation constant, day"1
active biomass, mg-L"1
one-half saturation no-growth coefficient,
mg-L"1
maximum specific growth rate, mg-L"1-day"1
2(1H)-quinolinone, mg-L"1
one-half saturation coefficient, mg-L"1
growth yield, mg biomass-mg substrate"1
Materials and Methods
The study site is located adjacent to an abandoned
wood preserving plant within the city limits of Pensacola,
Florida (1). The wood preserving process consisted of
steam pressure treatment of pine poles with creosote
and/or pentachlorophenol. For more than 80 years,
large but unknown quantities of waste waters (consist-
ing of extracted moisture from the poles, cellular debris,
creosote, pentachlorophenol, and diesel fuel from the
treatment processes) were discharged to nearby sur-
face impoundments. These impoundments were unlined
and in direct hydraulic contact with the underlying sand-
and-gravel aquifer. Contamination of the ground water
resulted from the accretion of wastes from these im-
poundments. Methanogenesis in the aquifer was simu-
lated using microcosms containing approximately 3 kg
of freshly collected anaerobic aquifer material in a 4-L
glass serum bottle with 2.5 L of prereduced anaerobi-
cally sterilized mineral salts solution. Approximately
40 mg/L of quinoline was added, simulating a concen-
tration similarto that found in the aquifer (1). The micro-
cosms were prepared, incubated, and sampled in an
anaerobic glove box containing an O2-free atmosphere
maintained at 22°C to 24°C. Microbially enriched micro-
cosms were prepared as above but were batch fed for
three cycles by removing 50 percent of the liquid volume
and replacing that volume with fresh mineral salts con-
taining enough quinoline to bring the final concentration
back to 40 mg/L. After the last feeding cycle, the liquid
culture was removed from the sediment, resulting in
liquid-only culture which was batch fed for six more
cycles.
Substrate concentrations were determined at approx-
imately 4-day intervals by high-performance liquid
chromatography. Total biomass concentrations were
determined at approximately 20-day intervals by total
protein using the Coomassie brilliant blue staining pro-
cedure of Galli (3).
87
-------
Results and Conclusions
The oxidation of quinoline, a reaction that is endergonic
(4), is modeled using derived Monod no-growth kinetic
constants. This oxidation is uncoupled with the degra-
dation of 2(1H)-quinolinone in microcosms simulating
natural conditions, as shown by the complete and
stoichiometric oxidation to 2(1H)-quinolinone before the
onset of methanogenesis. In microbially enriched micro-
cosms, however, the oxidation of quinoline is linked to
the degradation of 2(1H)-quinolinone; the increase in
biomass from methanogenesis must be included in the
equations describing the oxidation of quinoline (Figure
1). The kinetic values derived from the microbially en-
riched, all-liquid microcosm experiments were not sig-
nificantly different from those values from sand-filled
natural microcosms (Table 1). The Monod kinetic con-
stants for both the oxidation and subsequent methano-
genesis are representative of values describing
substrate utilization in an oligotrophic and somewhat
hostile environment (4).
It is still unclear, however, what number of microbial
populations are involved and to what extent each of the
populations influences the steps in the biodegradation
of quinoline. This uncertainty can be seen by the high
concentration of biomass capable of the oxidation of
quinoline in natural microcosms, suggesting that the
ability to oxidize quinoline is not unique to just this
consortium but may be common to many of the individ-
ual members of the creosote-degrading consortia. The
enrichment procedure has altered the microbial popula-
tion of the natural microcosms by potentially removing
all of the microorganisms that can oxidize quinoline but
are not directly involved in the methanogenesis of 2(1 H)-
quinolinone. These results suggest that as long as the
culture is derived from the contaminated aquifer, enrich-
Table 1. Monod Kinetic Constants ±95 Percent Confidence
Interval for Parameters Determined by Nonlinear
Regression for Both Natural and Microbially
Enriched Microcosms
Kinetic Constant
An, day1
/<, mg-L-1
Umax, day1
KS, mg-L'1
Y, mg-mg-1
Starting biomass, mg-L"1
Oxidation
Starting biomass, mg-L"1
Methanogenesis
Natural
Microcosms
0.31 ±0.06
2.0 ± 1.4
0.09 ± 0.06
11.4±0.6
0.03
17.3
0.003
Microbially
Enriched
Microcosms
0.29 ± 0.02
7.58 ± 4.0
0.14 ±0.07
33.1 ± 11.5
0.07
1.39
1.39
ment does not alter the kinetics of quinoline oxidation
and subsequent methanogenesis of 2(1H)-quinolinone.
The size of the various active microbial populations,
however, must be known before fate-and-transport mod-
eling can be attempted.
References
1. Godsy, E.M., D.F. Goer-lite, and D. GrbiCEGaliCE 1 992. Anaerobic
biodegradation of creosote contaminants in natural and simulated
ground water ecosystems. Ground Water 30:232-242.
2. Monod, J. 1 949. The growth of bacterial cultures. Ann. Rev. Micro-
biol. 3:371 -394
3. Galli, R. 1987. Biodegradation of dichloromettiane in waste water
using fluidized bed bioreactor. Appl. Microbiol. Biotechnol. 27:206-
21 3.
4. Godsy E.M. 1993. Mettianogenic biodegradation of creosote-de-
rived contaminants in natural and simulated ground water ecosys-
tems. Ph.D. dissertation. Stanford University, Stanford, CA. p. 155.
Quinolme
* Quinoline
D 2(lH)-Quiiiolinnne
A Biomass
Model
40 60
Time (days)
Figure 1. Quinoline oxidation and 2(1H)-quinolinone methano-
genesis in microbially enriched laboratory micro-
cosms.
88
-------
Stimulating the Biotransformation of Polychlorinated Biphenyls
John F. Quensen, III, Stephen A. Boyd, and James M. Tiedje
Michigan State University, East Lansing, Michigan
Introduction
The discovery that poly chlorinated biphenyls (PCBs)
can be reductively dechlorinated by microorganisms un-
der anaerobic conditions has stimulated interest in the
development of a sequential anaerobic/aerobic biotreat-
ment process for their destruction. While the aerobic
degradation of PCBs is generally limited to congeners
with four or fewer chlorines, the anaerobic process can
dechlorinate more highly substituted congeners, pro-
ducing products that are aerobically degradable. In-
deed, all products from the anaerobic dechlorination of
Aroclor 1254 (1) have been shown to be aerobically
degradable by one or more strains of aerobic bacteria
(2). Also, the high proportion of monochlorinated
biphenyls that can accumulate as a result of anaerobic
PCB dechlorination may serve to induce PCB-degrad-
ing enzymes in aerobic microorganisms (3). More
highly chlorinated congeners can be aerobically co-
metabolized but are not inducing substrates (4).
A greater understanding of the factors controlling the
anaerobic dechlorination of PCBs is necessary before a
successful sequential anaerobic/aerobic biotreatment
process can be developed for PCBs. In particular, it is
important to determine how to stimulate more rapid and
complete dechlorination in areas where the natural rate
and/or extent of dechlorination is limited. The general
approach we have taken is to identify the most probable
site-specific factors limiting in situ PCB dechlorination,
then to apply treatments to alleviate the limitation(s).
During the past year of this project, we have focused on
enhancing the dechlorination of PCBs present in River
Raisin (Michigan) and Silver Lake (Massachusetts)
sediments.
River Raisin Sediment Experiment
In a previous project, we found that little in situ dechlori-
nation of the PCBs present in River Raisin sediment
collected near Monroe, Michigan, had occurred. PCB-
dechlorinating microorganisms were found to exist in the
sediment, however. The sediment supported dechlori-
nation in laboratory assays when spiked with additional
PCBs and inoculated with PCB-dechlorinating microor-
ganisms (meaning inhibitory compounds were not pre-
sent), and the PCBs already present in the sediments
were bioavailable because they were dechlorinated un-
der the conditions of our treatability assay. In fact, indi-
vidual congeners in the contaminated sediment
decreased 30 to 70 percent in 24 weeks at rates nearly
identical to rates for the same congeners freshly spiked
into noncontaminated sediments.
The treatability assays were conducted using air-dried
River Raisin sediments. The sediments were slurried
with an equal weight of non-PCB-contaminated sedi-
ments and reduced anaerobic mineral medium (RAMM).
The slurry was then inoculated with microorganisms
eluted from Hudson River sediment to ensure that PCB-
dechlorinating microorganisms were present, and with
2',3,4-trichlorobiphenyl (34-2-CB) in a small volume of
acetone. The 34-2-CB was added because the addition
of a single PCB congener (or other halogenated aro-
matic compound) can sometimes "prime" the dechlori-
nation of PCBs already present in a contaminated
sediment (5). Non-PCB-contaminated Red Cedar River
sediments were added to provide a source of unidenti-
fied nutrients. The RAMM included essential mineral
salts and a chemical reductant (Na2S) to lower the initial
redox potential. We have conducted separate experi-
ments with slurries made from both air-dried and always
wet River Raisin sediments to help determine what as-
pect of our treatability assay fosters the dechlorination
of the PCBs present in these sediments.
Materials and Methods
Air-Dried River Raisin Sediments
With slurries made from the air-dried sediments, the
factors considered were 1) addition of 34-2-CB, 2) addi-
tion of the mineral salts in RAMM, 3) addition of Na2S,
and 4) addition of the non-PCB-contaminated sediment.
89
-------
The River Raisin sediment was added to Balch tubes
(1 g per tube). An additional 1 g of non-PCB-contami-
nated Red Cedar River sediment was added to the
appropriate treatments. Inocula for each treatment were
prepared by eluting PCB-dechlorinating microorgan-
isms from Hudson River sediments with a medium ap-
propriate to the treatment (i.e., with or without mineral
salts, with or without reductant), and 7 ml of an inocu-
lum was added to each tube using an anaerobic tech-
nique. 34-2-CB in a small volume of acetone was added
to one treatment, while the rest received the same vol-
ume of acetone. The tubes were sealed with Teflon-lined
rubber stoppers and aluminum crimps. Autoclaved treat-
ments served as controls. A tube was sacrificed for each
sample. Triplicate samples were taken at 8-week inter-
vals, extracted, and analyzed for PCBs using capillary
gas chromatography with electron capture detection.
Wet River Raisin Sediments
The same four factors described above plus the neces-
sity of inoculating with Hudson River microorganisms
were considered in an experiment with River Raisin
sediments that had been kept wet since the time of
collection. Portions of the sediment were mixed with the
appropriate medium (i.e., with or without salts, reduc-
tant, or inoculum) in a small Erlenmeyer flask on a
magnetic stirrer in an anaerobic chamber. Red Cedar
River sediments, acetone with or without 34-2-CB, and
PCB-dechlorinating microorganisms eluted from Hud-
son River sediments were also added as appropriate to
each treatment. Portions (7 ml) of the slurries were then
dispensed to Balch tubes, and the tubes were sealed
with Teflon-lined rubber stoppers and aluminum crimps.
The sampling and analytical procedure was the same
as the experiment with slurries made from air-dried
sediment.
Results and Discussion
Decreases in the concentrations of certain PCB conge-
ners in the live samples relative to the autoclaved con-
trols were used to compare the effectiveness of the
various treatments. These congeners (245-25-CB and
235-24-CB in chromatographic peak 42, and 34-34-CB
and 236-34-CB in peak 49) were chosen because each
peak represents more than 2 mole percent of the con-
geners initially present, and because they could not
have been formed in significant quantities from the
dechlorination of other PCBs present.
In the experiment with slurries made from the air-dried
sediment, approximately 50 percent of the congeners
present in each indicator peak were dechlorinated in the
treatment receiving 34-2-CB (Figure 1). No dechlorina-
tion was apparent in any of the other treatments.
In the experiment with slurries made from wet sedi-
ments, no inoculation was required for PCB dechlorina-
Decrease of 245-25-CB/235-24-CB
cd u
> c
o o
'
.S 0.5
oo No amendment
3E 34-2-CB
a£> Trace sails
v-^? Sodium Kulfi.de
oo Red Cedar sediment
8 16
Incubalioii Time (Weeks)
Decreasn of 236-34-CB/34-34-CU
> a
0-0
oo No amendment
E-E 34-2-CB
za Trace salts
v-^f Sodium sulfide
3-^ Red Cedar sediment
Figure 1.
8 16
Incubation Time (Weeks)
Decrease of indicator PCB congeners due to dechlor-
ination in slurries prepared from air-dried River Rai-
sin sediments.
tion, and some dechlorination occurred in all treatments.
Thus, merely making a slurry from the River Raisin
sediments appears to have stimulated some dechlorina-
tion of the PCBs present in them. The most extensive
dechlorination, however, occurred in the treatment re-
ceiving 34-2-CB, showing that dechlorination could be
enhanced by "priming." Somewhat surprisingly, the ad-
dition of Red Cedar River sediments inhibited dechlori-
nation (Figure 2); perhaps the additional organic matter
served as a sorptive sink for some of the PCBs.
Silver Lake Sediment Experiments
Although there is evidence that the PCBs present in
Silver Lake sediments have undergone in situ dechlori-
nation, these sediments do not support PCB dechlorina-
tion in laboratory experiments. These sediments have
high concentrations of several metals, especially zinc,
copper, lead, and chromium. We suspect that the metals
are present mainly in a reduced state in situ and become
partially oxidized and therefore more toxic to dechlori-
nating microorganisms during the sediment handling
90
-------
Decrease of a45-25-CB/235-24-CD
1.0-
*-. O
01
> a
o o
I-
K o
a
o
^
a,
ee No amendment
3-3 34-2-CB
Aa Trace salts
^-^ Sodium sulfide
Q& Red Cedar sediment
16
Incubation Time (Weeks)
Decrease of 236-34-CB/34-34-CB
G
tn O
s-, c
C
o o
D.
o
Du
&-^D No amendment
B-B 34-2-CB
&A Trace salts
v-^7 Sodium sulfide
oe Red Cedar sediment
Figure 2.
Incubation Time (Weeks)
Decrease of indicator PCB congeners due to dechlor-
ination in slurries prepared from wet River Raisin
sediments.
required to set up dechlorination experiments. We also
have previously shown that high concentrations of zinc
can inhibit PCB dechlorination even after highly reduced
conditions are restored. The experiments reported here
were designed to stimulate dechlorination by reducing
the bioavailability of toxic metals through chelation or
precipitation in both a model system and in Silver Lake
sediments.
Materials and Methods
General Procedure
Anaerobic sediment slurries containing PCBs were in-
oculated with a PCB-dechlorinating microbial consor-
tium eluted from PCB-contaminated Hudson River
sediments. Treatments consisted of the addition of met-
al salts and/or amendments to precipitate or chelate
metals. Autoclaved slurries served as negative or sterile
controls, while untreated slurries served as positive con-
trols. Triplicate samples were sacrificed at 4-week inter-
vals, solvent extracted, and analyzed for PCBs using
capillary gas chromatography with electron capture de-
tection. The course of PCB dechlorination was followed
by calculating the average meta plus para chlorines for
each treatment versus incubation time. No dechlorina-
tion from the ortho positions was evident. Dechlorination
patterns were evaluated by assessing changes in spe-
cific congener concentrations overtime.
Model System
Anaerobic slurries of non-PCB-contaminated Hudson
River sediment were spiked with Aroclor 1242 (500 |j,g/g
sediment) and inoculated with PCB-dechlorinating mi-
croorganisms eluted from PCB-contaminated Hudson
River sediments. We consistently observed dechlorina-
tion of the Aroclor 1242 in such preparations. Zinc (Zn)
or Lead (Pb) (as chloride salts) was added at solution
concentrations of 500 |j,g/mL to induce metal toxicity.
Amendments of FeSO4, ethylene diamine triacetic acid
(EDTA), and citrate were added individually to samples
before incubation to test their effectiveness in alleviating
the toxicity of Zn and Pb.
Silver Lake Sediment Slurries
Anaerobic slurries of Silver Lake sediments were spiked
with 34-2-CB and inoculated with Hudson River micro-
organisms. The 34-2-CB was added so that we could
monitor the dechlorination of a freshly added PCB con-
gener in addition to the PCBs already present in the
sediment. Experimental treatments consisted of the ad-
dition of FeSO4, EDTA, and citrate, as in the model
system described above.
Results and Discussion
In the model system, ZnCI prevented Aroclor 1242
dechlorination while PbCI decreased the extent of
dechlorination. EDTA, citrate, and FeSO4 amendments
all reversed the inhibitory effect of PbCI while EDTA and
FeSO4 eliminated the inhibition by ZnCI. In fact, FeSO4-
amended treatments exhibited more extensive dechlori-
nation than the unamended positive controls (i.e., those
without PbCI or ZnCI additions). Apparently, the FeSO4
greatly stimulated dechlorination from para positions. In
all non-FeSO4-amended slurries exhibiting dechlorina-
tion, dechlorination occurred primarily from the meta
positions to yield ortho and para substituted products
(pattern M). But in FeSO4-amended treatments, the ma-
jor products were 2-CB, 2-2-CB, and 26-CB, indicating
that dechlorination occurred from both meta and para
positions (pattern C). We have often noted that the para
dechlorination activity present in Hudson River sedi-
ments is lost during storage of the sediments. It appears
that addition of FeSO4 somehow "rescues" this dechlori-
nation activity.
91
-------
In the Silver Lake sediment slurries, the added 34-2-CB
was dechlorinated in citrate- and FeSO4-amended slur-
ries, but not in EDTA-amended slurries. There was no
indication of further dechlorination of the PCBs already
present in the sediments.
References
1. Quensen, J.F., III, S.A. Boyd, and J.M. Tiedje. 1990. Dechlorination
of four commercial polychlorinated biphenyl mixtures (Aroclors) by
anaerobic microorganisms from sediments. Appl. Environ. Micro-
biol. 56:2,360-2,369.
2. Bedard, D.L., R.E. Wagner, M.J. Brennan, M.L. Haberl, and J.F.
Brown, Jr. 1987. Extensive degradation of Aroclors and environ-
mentally transformed polychlorinated biphenyls by Alcalignes eu-
trophus H850. Appl. Environ. Microbiol. 53:1,094-1,102.
3. Masse, R., F. Messier, L. Peloquin, C. Ayotte, and M. Sylvestre.
1984. Microbial biodegradation of 4-chlorobiphenyl, a model com-
pound of chlorinated biphenyls. Appl. Environ. Microbiol. 41:947-
951.
4. Furukawa, K., F. Matusumura, and K. Tonomura. 1978. Alcali-
genes and Acinetobacter strains capable of degrading polychlori-
nated biphenyls. Agric. Biol. Chem. 42:543-548.
5. Bedard, D.L., H.M. VanDort, R.J. May, K.A. DeWeerd, J.M. Prin-
cipe, and L.A. Smullen. 1992. Stimulation of dechlorination of Aro-
clor 1260 in Woods Pond sediment. In: General Electric Company
research and development program for the destruction of PCBs,
11th progress report. Schenectady, NY: General Electric Corpora-
tion Research and Development, pp. 269-280.
92
-------
Bioaugmentation for In Situ Co-metabolic Biodegradation of
Trichloroethylene in Ground Water
Junko Munakata Marr and Perry L. McCarty
Stanford University, Stanford, California
V. Grace Matheson, Larry J. Forney, and James M. Tiedje
Michigan State University, East Lansing, Michigan
Stephen Francesconi and Malcolm S. Shields
University of West Florida, Pensacola, Florida
P.M. Pritchard
U.S. Environmental Protection Agency, Gulf Breeze, Florida
Introduction
Trichloroethylene (TCE), a common ground-water con-
taminant, has been found to be fortuitously degraded
(co-metabolized) by organisms grown on a variety of
substrates (1, 2). Addition of such substrates can lead
to two significant problems in ground-water aquifers.
First, organisms stimulated by substrate addition may
be unable to degrade TCE. Second, the most promising
compounds for inducing TCE degradation, phenol and
toluene, are themselves hazardous substances and
therefore cause regulatory concern. To address the first
issue, aquifers may be bioaugmented with wild-type
strains known to be effective at TCE degradation. Per-
haps ideally, both problems can be overcome through
the use of mutant strains known both to degrade TCE
efficiently and to grow on a nontoxic substrate. Labora-
tory studies were conducted to investigate these two
alternatives.
Laboratory Studies
Bacterial Cultures
The wild-type strain evaluated was Pseudomonas
cepacia G4 (G4), a strain isolated from a holding pond
at an industrial waste treatment facility in Pensacola,
Florida (2). This organism co-metabolizes TCE using
toluene ortho-monooxygenase (TOM), which is induced
by phenol or toluene (3). The mutant used was P.
cepacia G4 PR130i (PR1), a chemically induced mutant
of G4 that constitutively expresses TOM while grown on
substrates such as lactate. A more complete description
of PR1 is presented at this meeting.
So/7 Microcosms
Small-column microcosms (17 cm3) were constructed
using aquifer material from a test area at Moffett Federal
Air Station. Column fluids were exchanged every 2 to 3
days by pumping 10 ml of solutions held in gas-tight
glass barrel syringes through the column influent port
with a syringe pump. At the start of each fluid exchange
period, 1 ml of bacterial culture was added to the mi-
crocosms followed by 9 ml of oxygenated ground water
containing about 200 |j,g/L TCE and/or primary sub-
strates. Microcosm effluent samples were collected dur-
ing each exchange for analysis.
Detection of Bacteria
A deoxyribonucleic acid (DMA) probe specific for both
strains of G4 was constructed using polymerase chain
reaction (PCR) to amplify segments of G4 DMA between
repetitive extragenic palindromic (REP) sequences. The
REP-PCR reaction can be performed directly on envi-
ronmental samples and therefore does not require ex-
traction of DMA before amplification. The method was
tested using ground-water and sediment samples con-
taining indigenous bacterial populations with and with-
out added G4, with parallel plate counts on R2A agar.
93
-------
Results
Soil Microcosms
The 6.5-mg/L, phenol-fed, nonbioaugmented column
followed a pattern similar to that observed in the field (4)
and consumed approximately 60 |j,g/L TCE relative to a
nonfed control. Columns augmented with induced G4
without a primary substrate achieved similar levels of
TCE degradation. With the addition of 15 mg/L lactate,
degradation increased to 100 |j,g/LTCE in G4-amended
columns, but no such increase occurred in a PR1-
amended column. In these lactate-fed columns, the G4
was pregrown on phenol while the PR1 was pregrown
on lactate. When columns amended with either G4 or
PR1 were fed 6.5 mg/L phenol, 130 |j,g/L TCE was
degraded. The results are summarized in Figure 1.
Detection of Bacteria
The REP-PCR probe was able to detect 10 colony
forming units (CPUs) of G4 against a background of 10^
nontarget CPUs contained in 1 |j,L of template (Figure
2). The probe's sensitivity compares favorably to other
PCR-based detection methods (5).
Log10 CFU nontarget mixture
+10CFUG4
Figure 2. Sensitivity of G4 REP-PCR products using the strain
G4 GF13 probe; agarose gel electrophoresis of REP-
PCR reactions and hybridization of the Southern blot
to GF13.
Application of REP-PCR to aqueous effluent samples
from the soil microcosms produced mixed results (Fig-
ure 3). G4 was not detected in the control or phenol-only
microcosms and gave a strong signal in phenol- and
lactate-fed microcosms augmented with G4, as antici-
o
s
o
c
o
o
LU
P
a)
3
O
0)
I
200 -
8
1
Control-
> 1
1 1
1 1
W 0)
1 1-
^ CO
CD E>
A °
£ i
1
i « i
CD
_£
o « & i §.
B ffl °-
fluent
fluent
1
1
o
a>
a>
0
_co
S
-------
G4 123488789 1ft
Figure 3. Detection of G4 in column effluents, ninth column
exchange (Day 20); agarose gel electrophoresis of
REP-PCR reactions of column effluents and probe
GF13 hybridization of the Southern blot.
pated. In microcosms containing unfed G4 or phenol-fed
PR1, however, responses were weak or absent. The
reason for the latter results has not yet been determined.
Conclusion
Bioaugmentation with G4 or PR1 and phenol feed pro-
vides a means for enhancing native activity toward TCE.
Addition of phenol to aquifers could be avoided by sim-
ply adding G4 previously induced forthe TCE-degrading
enzyme. Lactate enhanced activity of preinduced G4
toward TCE. In PR1-augmented systems, however, lac-
tate did not support the same level of activity toward
TCE as did phenol. G4 and PR1 were identified in
constructed samples with a sensitivity of 1 in 104. De-
tection in aqueous column samples gave some unex-
pected results, the causes for which remain to be
elucidated.
References
1. Wilson, J.T., and B.H. Wilson. 1985. Biotransformation of trichlo-
roethylene in soil. Appl. Environ. Microbiol. 49:242-243.
2. Nelson, M.J.K., S.O. Montgomery, E.J. O'Neill, and P.M. Pritchard.
1986. Aerobic metabolism of trichloroethylene by a bacterial iso-
late. Appl. Environ. Microbiol. 55:383-384.
3. Shields, M.S., S.O. Montgomery, S.M. Cuskey, P.J. Chapman, and
PH. Pritchard. 1991. Mutants of Pseudomonas cepacia G4 defec-
tive in catabolism of aromatic compounds and trichloroethylene.
Appl. Environ. Microbiol. 57:1,935-1,941.
4. Hopkins, G.D., J. Munakata, L. Semprini, and PL. McCarty. 1993.
Trichloroethylene concentration effects on pilot field-scale in situ
groundwater bioremediation by phenol-oxidizing microorganisms.
Environ. Sci. Technol. 27:2,542-2,547.
5. Thiem, S.M., M.L. Krumme, R.L. Smith, and J.M. Tiedje. 1994.
Use of molecular techniques to evaluate the survival of a micro-
organism injected into an aquifer. Appl. Environ. Microbiol.
60:1,059-1,067.
95
-------
Biodegradation of Chlorinated Solvents
Larry Wackett, Lisa Newman, and Sergey Selifonov
University of Minnesota, St. Paul, Minnesota
Peter Chapman and Michael Shelton
U.S. Environmental Protection Agency, Gulf Breeze, Florida
General Scope of Research
Research is being conducted on the bacterial metabo-
lism of chlorinated aliphatic compounds, with a focus
on oxidative mechanisms of biodegradation. Pseudo-
monas cepaciaG4 oxidizes trichloroethylene (TCE) and
related chlorinated alkenes with relatively little loss of
activity over time (1), and the molecular basis of this
observation is being elucidated. In vivo experiments are
delineating the substrate range and concentration limits
of P. cepacia G4 for chlorinated solvents. In vitro experi-
ments are defining the properties of toluene 2-monooxy-
genase, the enzyme catalyzing the oxidation of TCE.
Purification and Properties of Toluene
2-Monooxygenase
Toluene 2-monooxygenase activity was monitored in
vitro via a sensitive radiometric assay using [14C]-tolu-
ene (2). Chromatography of cell-free extracts revealed
that this was a three-component oxygenase system. All
three components have now been purified to homoge-
neity. In vitro reconstitution of the three proteins and
reduced nicotinamide adenine dinucleotide (NADH)
yielded an active enzyme system that oxidizes toluene
to ort/70-cresol and this, subsequently, to 3-methyl-
catechol. One component is a flavoprotein containing
a 2Fe2S cluster that accepts electrons from NADH
(Table 1). A second component is a low molecular
weight protein that stimulates activity but has no obvious
redox-active functional group (Table 1). The largest
component has an a2p2Y2 subunit structure (Table 1).
This component is implicated as the hydroxylase com-
ponent as it alone will oxidize toluene in the presence of
dithionite + methyl viologen + O2 or hydrogen peroxide.
The hydroxylase component contains four to six iron
atoms per holoenzyme. Spectroscopically, this compo-
nent resembles the soluble methane monooxygenase
hydroxylase component from Methylosinus trichospor-
iumOBZb (3).
Table 1. Molecular Properties of Purified Components
Small
Property Hydroxylase Component Reductase
Subunit structure
Subunit molecular
masses (kDa)
Molecular mass (kDa)
Gel filtration
Native PAGE
SDS-PAGE
Calculated (aa
quantitation)
Metal content
Iron content (mol/mol)
Inorganic S" content
(mol/mol)
FAD (mol/mol)
Pi
Absorption maxima
Specific activity
(units/mg)
Percent recovery
(apy)2
5.4, 37.7, 13.5
216
190
211
210
5.3
ND
ND
4.5
282 nm
1.7a
40
Monomer
10.4
19.3
10.5
10.4
ND
ND
ND
4.3
277 nm
79.4a
27
Monomer
40.0
45.8
41.8
40.0
2.3
2.9
1.2
5.8
270, 341 ,
423, 457
nm
512.0b
30
unit is defined as 1 nmol [ C]-toluene/min at 23°C.
b One unit is defined as 1 |imol cytochrome c reduced/min at 23°C.
ND = not detected.
= not determined.
PAGE = polyacrylamide gel electrophoresis
SDS = sodium dodecylsulfate
FAD = flavin adenine dinucleotide
pi = isoelectric point
96
-------
In Vivo Studies With P. Cepacia G4
P. cepacia G4 was shown to grow on aromatic ring
compounds other than toluene and phenol. P. cepacia
also oxidized non-growth-supporting aromatic and ali-
phatic substrates. Examples of the aromatic substrates
that were investigated in some detail include naphtha-
lene and indene. The oxidation of TCE by P. cepacia G4
has been studied in detail. The major oxidation product
is glyoxylic acid. The effects of TCE on P. cepacia G4
also were studied to determine how resistant the organ-
ism is to variable concentrations of TCE. Unlike Pseudo-
monas putida F1, P. cepacia G4 was not detectably
toxified by low concentration of TCE or by metabolites
generated by oxidative mechanisms. High TCE concen-
trations, however, exerted a solvent effect that could
markedly depress cell division rates and even cause
cell death.
References
1. Folsom, R.R., P.J. Chapman, and P.M. Pritchard. 1990. Phenol and
trichloroethylene degradation by Pseudomonas cepacia G4: Kinet-
ics and interactions between substrates. Appl. Environ. Microbiol.
56:1,279-1,285.
2. Yeh, W.-K., D.T. Gibson, and T.N. Liu. 1977. Toluene dioxygenase:
A multicomponent enzyme system. Biochem. Biophys. Res.
Comm. 78:401-410.
3. Fox, B.C., W.A. Froland, J.E. Dege, and J.D. Lipscomb. 1989.
Methane monooxygenase from Methylosinus trichosporium OB3b:
Purification and properties of a three-component system with high
specific activity from a type II methanotroph. J. Biol. Chem.
264:10,023-10,033.
97
-------
Biological and Nutritional Factors Affecting Reductive Dechlorination of
Chlorinated Organic Chemicals
Dingyi Ye
National Research Council, Athens, Georgia
W. Jack Jones
U.S. Environmental Protection Agency, Athens, Georgia
Introduction
Halogenated organic chemicals are of major public con-
cern because these compounds are usually toxic and
persistent in the environment, and they tend to accumu-
late in soils, sediments, and biota. Polychlorinated
biphenyls (PCBs) and organochlorine pesticides are en-
vironmental pollutants of great concern due to a history
of heavy use, toxicities, persistence in the environment,
and wide distribution in environmental media.
PCBs are a mixture of chlorinated biphenyls consisting
of 209 possible congeners. These compounds were
widely used for almost 50 years, with several hundred
million pounds having been released into the environ-
ment. An organochlorine pesticide of concern is the
insecticide toxaphene, a complex mixture of chlorinated
camphenes. Toxaphene consists of more than 177 de-
rivatives and was heavily used in the United States
before 1982. Estimates indicate that about 233,688 met-
ric tons of toxaphene was manufactured in the United
States from 1964 to 1982. Toxaphene, like PCBs and
other organochlorines contaminants, is a global pollut-
ant. Many studies have shown toxaphene to be rela-
tively persistent and bioaccumulated by biota (1).
The objectives of the present research were to study
factors affecting anaerobic transformation of PCBs and
organochlorine pesticides (e.g., toxaphene) and to de-
velop techniques to enhance their in situ bioremediation.
The preliminary goals were 1) to characterize the an-
aerobic microbial dechlorination of PCBs in Sheboygan
River, Wisconsin, sediment, 2) to examine the effects
of nutrients, Fe°, and electron carriers on dechlorination
of PCBs, and 3) to examine the anaerobic biotransfor-
mation of toxaphene using indigenous and PCB-
dechlorinating microorganisms.
Toxaphene Biotransformation
Materials and Methods
A Hudson River (HR) pasteurized enrichment culture
capable of reductive dechlorination of PCBs was used
for initial toxaphene experiments. The enrichment cul-
ture was originally pasteurized at 85°C for 15 min and
subsequently transferred at monthly intervals (1 percent
v/v transfer, repasteurized at 90°C for 10 min at each
transfer). The inoculum was serially diluted, and the
highest dilution (10"6) retaining PCB-dechlorination ac-
tivity was inoculated to the medium without PCBs; this
culture was used as the inoculum for PCB and
toxaphene biotransformation studies. Two milliliters of
the PCB-dechlorinating inoculum was anaerobically
transferred to 28-mL culture tubes containing 2 ml re-
vised anaerobic mineral medium (RAMM) and 1 g of
sterile, uncontaminated (toxaphene-free) pond sedi-
ment. Toxaphene was subsequently added at a final
concentration of 500 u,g/g dry sediment. Control sedi-
ment samples were autoclaved three times (121°C for
1 hr each time) on consecutive days, with addition of
toxaphene occurring on Day 4. All cultures were incu-
bated at 25°C in the dark.
Results and Discussion
Anaerobic transformation of toxaphene by the pasteur-
ized HR enrichment was evident as indicated by
changes in the gas chromatography (GC) isomer-
distribution patterns. GC chromatograms of the auto-
claved control and a sample inoculated with the HR
pasteurized inoculum after 7 months of incubation are
presented in Figure 1. Several peaks representative of
toxaphene isomers are numbered to facilitate compari-
son of the control and experimental chromatograms.
Only a very minor change was observed for a late
98
-------
Figure 1. GC profile of toxaphene in A) autoclaved and B) live
experimental microcosms (inoculated with enrich-
ment from Hudson River) after 28 weeks of anaerobic
incubation. Numbered peaks are for reference only.
eluting peak (#39) in the control and live experimental
samples; thus, this peak was chosen as an internal
reference peak to which other peaks were normalized
(based on peak height). In comparison with the sterile
control, appreciable changes in many peaks were noted
in live samples after 7 months of incubation. The per-
centage changes in specific peaks for the live sample
compared with the sterile control are presented in Figure
2. Increases in some peaks were observed, indicating
accumulation of some dechlorination products. Some
early eluting peaks also decreased, however, suggest-
ing that in addition to reductive dechlorination, anaero-
bic degradation may have occurred. The possible
anaerobic degradation of toxaphene will be further in-
vestigated by identifying polar transformation products.
Changes in toxaphene isomer-distribution patterns were
also observed following GC analyses of autoclaved con-
trols (data not shown). As mentioned previously, sterile
controls were prepared by autoclaving sediments at
121°C for 1 hr on 3 consecutive days. Thus, it is highly
unlikely that the observed transformations in the sterile
controls were biologically mediated. The observed
changes are likely due to either abiotic (chemical)
o
0
CO
1
8
(D
Changes (tt
^
IQU
100 -
50 -
-50-
-100 -
c
''I
)
I
1 1
.
10
J Increasing Peaks
l.ll , ,
1 ' "ll'i'i I'jT |l
Decreasing Peaks
20 30 4
Peak Number
Figure 2. Percentage change (relative to control sample) of se-
lect toxaphene peaks in live experimental microcosm
after 28 weeks of anaerobic incubation.
transformation or enhanced sorption of toxaphene to the
autoclaved soil matrix. It has been previously reported
that specific toxaphene isomers were transformed in
sterile sediments and in a sand-Fe(ll)/Fe(lll) system (2).
Phototransformation of toxaphene has also been
reported (2). On the other hand, the high KOC (soil or-
ganic carbon partition coefficient) value reported for
toxaphene suggests that the chemical mixture should be
strongly sorbed to soil particulates (1); any differences
in KQC among the isomers may influence the toxaphene
isomer distribution patterns in long-term sediment incu-
bations. Thus, the changes noted in the isomer distribu-
tion pattern (after correcting for abiotic transformations
observed in sterile controls) are likely isomers of
toxaphene that were subject to transformation by the
inoculated microorganisms and that were relatively re-
sistant to abiotic transformation. In these inoculated ex-
perimental cultures, CH4 production was not observed,
and a mete-directed dechlorination of amended PCB
congeners (Aroclor 1242) was confirmed in separate
experiments. These results suggest that the HR pas-
teurized enrichment culture was capable of anaerobi-
cally transforming toxaphene. The HR pasteurized
enrichment is easily maintained and cultivated and,
therefore, may be of potential use in the remediation of
toxaphene-contaminated soils. Additional studies are
under way to evaluate the effectiveness of this enrich-
ment culture for remediation of historically contaminated
soils.
PCB Biotransformation
Materials and Methods
PCB biotransformation experiments were performed
with PCB-contaminated (approximately 500 ppm) She-
boygan River (SR) sediment. For abiotic transformation
experiments, PCB-contaminated sediment was slurried
with anoxic site water inside an anaerobic glove box,
99
-------
homogenized, then amended with Fe°. Of the slurry
containing 1 g sediment (dry weight), 1.8 ml was trans-
ferred to replicate 28-mL serum tubes. Half of the tubes
were spiked with 300 u,g Aroclor 1242 as an available
PCB source. Fe°, pyrite, and degassed, sterile distilled
water were then added. Control samples consisted of
autoclaved sediment slurries as described above. All
cultures were incubated at 25°C in the dark.
An experiment to assess the effect of pasteurization on
microorganisms eluted from SR sediment was con-
ducted as described by Ye et al. (3). Finally, eluted
microorganisms from historically contaminated (PCB)
SR sediment were subjected to pasteurization (85°C for
20 min) and used as inocula to assess their potential for
reductive dechlorination of amended Aroclor 1242,
1248, and 1254. Aroclors were added individually at a
final concentration of 500 u,g/g dry sediment.
Results and Discussion
Fe -Amended Experiments
Several studies document the Fe°-mediated reductive
dechlorination oftrichloroethylene (TCE) and other chlo-
rinated compounds (4). Our preliminary results of Fe°-
amended SR sediment slurries, however, indicated that
no dechlorination of PCBs occurred after anaerobic in-
cubation for 2 weeks at 20°C in either live or sterile
samples. Further, no evidence of reductive dechlorina-
tion of PCBs was observed in the SR sediment slurries
spiked with Aroclor 1242, indicating that bioavailability
of PCB congeners was not a limiting factor for the
Fe°-mediated dechlorination.
A prolonged incubation time is usually necessary to
achieve biologically mediated reductive dechlorination
of PCBs. Thus, it was not surprising to find no evidence
of dechlorination in the live experimental samples, es-
pecially because the SR sediment had been stored at
2°C to 4°C for approximately 1 yr before use. It is likely
that a significant amount of time is necessary for the
dechlorinating population to recover to a level to affect
significant dechlorination. Additional experimental re-
sults with SR sediment (without nutrient amendment)
indicated that approximately 4 weeks of incubation was
required before detectable PCB dechlorination was
observed.
Investigators at the U.S. Environmental Protection
Agency Athens Research Laboratory have recently
demonstrated Fe°-mediated reductive dechlorination of
other halogenated compounds. We have, however, no
evidence of PCB dechlorination in Fe°-amended sam-
ples under similar experimental conditions. These re-
sults suggest that PCBs are more resistant to chemical
A
B
-
uliJL
Figure 3. GC profiles of Aroclor 1254 in A) pasteurized, B) auto-
claved, and C) live experimental microcosms (inocu-
lated with SR-eluted microorganisms) after 12 weeks
of incubation.
(abiotic) dechlorination than other chlorinated com-
pounds examined to date.
Pasteurization of SR Sediment
CH4 production was not observed in experimental mi-
crocosms inoculated with the pasteurized microorgan-
isms from SR sediment; nonpasteurized cultures,
however, were actively methanogenic. These observa-
tions are consistent with previous pasteurization experi-
ments using HR sediments as inocula (3). The
pasteurized cultures preferentially removed meta chlo-
rines, while the untreated cultures removed both meta
and para chlorines from selective PCB congeners. In the
present study, dechlorination of Aroclor 1254 was ob-
served by the untreated inocula after 6 weeks of incu-
bation; however, no significant dechlorination of Aroclor
1254 was evident in experiments inoculated with the
pasteurized inocula after 12 weeks of incubation (Figure
3). These results suggest that anaerobic, spore-forming
microorganisms in the SR sediment exhibit similar
dechlorinating pathways as the microorganisms in the
100
-------
HR sediment. Stimulation Of in Situ PCB dechlorination 2. Williams, R.R., and IF. Bidleman. 1978. Toxaphene degradation
may be possible through the addition of a suitable spore in estuarine sediments. J. Agric. Food chem. 26:280-282.
germinant or growth Substrate. 3. Ye, D., J.F. Quensen, III, J.M. Tiedje, and S.A. Boyd. 1992. An-
aerobic dechlorination of polychlorolbiphenyls (Aroclor 1242) by
pasteurized and ethanol-treated microorganisms from sediments.
RGfGTGnCGS Appl. Environ. Microbiol. 58:1,110-1,114.
4. Matheson, L.J., and P.G. Tratnyek. 1994. Reductive dechlorination
1. Saleh, M.A. 1991. Toxaphene: Chemistry, biochemistry, toxicity of chlorinated methanes by iron metal. Environ. Sci. Technol.
and environmental fate. Rev. Environ. Contam. Toxicol. 118:1-85. 28:2,045-2,053.
101
-------
Predicting Heavy Metal Inhibition of the In Situ Reductive Dechlorination of
Organics at the Petro Processor's Superfund Site
John H. Pardue
Department of Civil and Environmental Engineering, Louisiana State University,
Baton Rouge, Louisiana
Introduction
Transition metals and synthetic organic compounds are
common co-contaminants at waste sites that are candi-
dates for biological treatment. The inhibition of microbial
decomposition of natural organic matter by certain tran-
sition metals has been widely documented (1); however,
the inhibition of anaerobic degradation processes (e.g.,
reductive dechlorination) is poorly understood.
Inhibition Characteristics
Inhibition characteristics of a model heavy metal (cad-
mium, Cd) on a model chlorinated aromatic (2,3,4-
trichloroaniline, 2,3,4-TCA) was determined in the
laboratory. Laboratory microcosm experiments were
conducted in three anaerobic flooded soils with varying
properties. Dechlorination of 2,3,4-TCA to monochlo-
roanilines occurred when total pore-water Cd concen-
trations were below a critical threshold level. Inhibition
occurred across a continuum of Cd concentrations in
several soils, but a completely inhibited threshold con-
centration was readily identified (Figure 1). Dechlorina-
tion kinetics and metabolites differed with soluble metal
concentration. Speciation of soluble Cd was necessary
to predict whether inhibition would occur, particularly in
the presence of high concentrations of organic ligands
such as humic acids (Table 1). Estimation of metal pools
using selective extractions and measurement of acid-
volatile sulfide (AVS) provided additional information but
did not adequately predict whether inhibition of de-
chlorination would occur. These results demonstrated
the importance of quantification and speciation of
pore-water metals in predicting potential inhibition of
anaerobic biodegradation reactions such as reductive
dechlorination.
a
o £?
.H e
s s
? "3
S §
=s fe.
1.4
1.2
1.0
0.8
0.6 -
0.4 -
0.2
00 -
o
o
0 0 ° 0
BLH
0 Marsh o
Rice 0
o
*
o
, CuDO , - C>-T
1 10 100 1000 10000
Soluble Cd concentration ((.ig/L)
Figure 1. Normalized 2,3,4-TCA dechlorination rates (k/kcontroi)
versus soluble Cd concentrations in three flooded
soils: bottomland hardwood (BLH), rice paddy, and
freshwater marsh.
Table 1. MINTEQA2 Results From Pore Water of
Representative Rice and Marsh Soil Suspensions
(estimated pore-water humic acid concentrations
were 1 mg/L in the rice soil and 55 mg/L in the
marsh soil)
Soil
RS
MS
Equilibrated
Total Mass 2,3,4-TCA
Soluble Distribution, Dechlorination
Cd (mg/L) Cd Species % Inhibited
0.195 Cd+2
CdCI+
CdSO4 (aq)
Cd-Humate
0.350 Cd+2
Cd-Humate
43.7 +
5.9
1.6
48.5
1.0
98.7
102
-------
Site History
The Petro Processor's, Inc., site is a high-priority Super-
fund site near Baton Rouge, Louisiana. The site served
as a chemical waste pit from the early 1960s to the late
1970s. An estimated 60,000 tons of chlorinated organic
waste, primarily hexachlorobutadiene and hexachlo-
robenzene (HCB), was deposited in several unlined,
diked pits. A spill event resulted in contamination of
stream sediments in an adjacent bottomland hardwood
wetland. Heavy metal contamination is contiguous with
chlorinated organic contamination (primarily HCB) in
these sediments. These sediments are the site of a
bioremediation field trial directed at enhancing reductive
dechlorination of HCB (2).
Predicting Heavy Metal Inhibition
Sampling is being conducted to determine if metal inhi-
bition of reductive dechlorination can be predicted in the
field at the Petro Processor's site. Characteristics of
inhibition of model compounds (described above) are
being used to develop a strategy for predicting inhibition.
Parallel laboratory studies are being used to confirm that
these same inhibition characteristics would be observed
for HCB in the sediments. Laboratory studies using
2,3,4-TCA indicated that noninhibited soils could be
adequately predicted using the AVS/SEM (simultane-
ously extracted metal) concept. This concept has been
used to predict the toxicity of metals to benthic organ-
isms (3). In studies with 2,3,4-TCA, soils in which molar
metal concentrations exceeded molar AVS were not
always inhibited, requiring further metal speciation and
prediction of "free," uncomplexed metal concentrations
using MINTEQA2. Spatial and seasonal information of
AVS and SEM and observations of lower chlorinated
benzene samples are being collected in the field. Sec-
tions of the bayou where molar SEM exceeds molar AVS
are undergoing further metal speciation studies.
References
1. Duxbury, T. 1985. Ecological aspects of heavy metal responses in
microorganisms. Adv. Microbiol. Ecol. 8:185-235.
2. Constant, W.D., J.H. Pardue, R.D. DeLaune, K. Blanchard, and
G.A. Breitenbeck. 1995. Enhancement of in situ microbial degra-
dation of chlorinated organic waste at the Petro Processor's Su-
perfund site. Env. Progress 14:51-60.
3. Di Toro, D.M., J.D. Mahony, D.J. Hansen, K.J. Scott, M.B. Hicks,
S.M. Mays, and M.S. Redmond. 1990. Toxicity of cadmium in
sediments: The role of acid volatile sulfides. Environ. Toxicol.
Chem. 9:1,489-1,504.
103
-------
Effect of Primary Substrate on the Reduction of 2,4-Dinitrotoluene
Jiayang Cheng and Makram T. Suidan
Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, Ohio
Albert D. Venosa
Risk Reduction Engineering Laboratory, U.S. Environmental Protection Agency, Cincinnati, Ohio
Introduction
2,4-Dinitrotoluene (DNT) is one of the priority pollutants
(1) commonly found in munitions wastes. It is recalci-
trant to biological treatment in aerobic processes (2),
such as the activated sludge system, but can be de-
graded (3) in a sequential anaerobic/aerobic biosystem.
2,4-DNT is completely transformed to 2,4-diaminotolu-
ene (DAT) with ethanol as the primary substrate in an
anaerobic reactor. Subsequently, 2,4-DAT is readily min-
eralized (3) in an aerobic reactor. 2,4-DNT can not be
transformed in the anaerobic reactor (4) without a pri-
mary substrate. In this study, the anaerobic biotransfor-
mation of 2,4-DNT with ethanol, methanol, acetic acid,
or hydrogen as primary substrate was investigated. The
effect of the primary substrate on the reductive transfor-
mation of 2,4-DNT was also studied.
2,4-DNT-transforming anaerobic cultures were accli-
mated with 2,4-DNT and ethanol, methanol, or acetic
acid as the feed organic substrates in three chemostats.
The concentrations of 2,4-DNT and the primary sub-
strates in the feed to the three chemostats are listed in
Table 1. The chemical oxygen demand (COD) loading
for all three chemostats was the same. Minerals and
nutrients were added to the chemostat feed to support
bacteria growth. Na2S 9H2O (50 mg/L) was added to
maintain a reducing environment in the chemostats. The
pH and the temperature in the chemostats were main-
tained constant at 7.2 and 35°C, respectively. The hy-
draulic retention time in the chemostats was 40 days.
Table 1. Concentrations of Substrates in Feed for the
Chemostats
Chemostat
Ethanol Methanol Acetic
Fed Fed Acid Fed
2,4-DNT, mg/L
Primary substrate, mg/L
91.7
500
91.7
696
91.7
978
2,4-DNT was completely biotransformed to 2,4-DAT in
all three chemostats. All the primary substrates (ethanol,
methanol, and acetic acid) were converted to methane
and carbon dioxide.
After steady-state operation was achieved in the chemo-
stats, the mixed cultures from the chemostats were used
as the inocula for the batch tests to determine the kinet-
ics of anaerobic biotransformation of 2,4-DNT with dif-
ferent primary substrates. The cultures were then
transferred into the batch reactors in an oxygen-free
anaerobic chamber at 35°C. The pH and the tempera-
ture in the batch reactors were maintained the same as
those in the chemostats. Different initial concentrations
of 2,4-DNT were used in the batch tests. To determine
the co-metabolic mechanism of the biotransformation of
2,4-DNT and the primary substrate, hydrogen was also
used as the primary substrate in the batch tests.
Results and Discussion
All the batch tests were run in duplicate, and the devia-
tions of the results were less than 8 percent. The kinetics
of anaerobic biotransformation of 2,4-DNT with ethanol,
methanol, and acetic acid as the primary substrates are
illustrated in Figure 1 (a), (b), and (c), respectively.
2,4-DNT was completely biotransformed to 2,4-DAT via
4-amino-2-nitrotoluene (4-A-2-NT) or 2-amino-4-nit.ro-
toluene (2-A-4-NT) under anaerobic conditions, regard-
less of the primary substrate. The rate of the
biotransformation of 2,4-DNT and the intermediates
(4-A-2-NT and 2-A-4-NT), however, was much higher in
the presence of ethanol than that in the presence of
either methanol or acetic acid. When ethanol was used
as the primary substrate, hydrogen was produced
during the acetogenesis of ethanol. The hydrogen
then served as the electron donor for the reduction of
2,4-DNT to 2,4-DAT. The bacteria also used ethanol for
their growth. When methanol or acetic acid was used as
104
-------
0.10
0.10
0.00
2,4-DNT
4-A-2-NT
2-A-4-NT
2,4-DAT
1400
200
400
600
800
1000
1200
1400
4-A-2-NT
2-A-4-NT
2,4-DAT
0 200 400 600 800
Time, Hrs
1000
1200
1400
Figure 1. Anaerobic biotransformation of 2,4-DNT with (a) ethanol, (b) methanol, or (c) acetic acid as the primary substrate.
105
-------
the primary substrate, the substrates were used in the
biosystem to support the growth of the bacteria that
transformed 2,4-DNT to 2,4-DAT. Neither methanol nor
acetic acid was degraded until 2,4-DNT, 4-A-2-NT, and
2-A-4-NT were completely transformed to 2,4-DAT. The
hydrogen for the reductive transformation of 2,4-DNT
and the intermediates was probably from bacterial en-
dogenous decay. 2,4-DNT was not biotransformed with-
out a primary substrate (4) in both chemostat and batch
reactors. 2,4-DNT itself cannot support the growth of the
bacteria, and the primary substrate is necessary for
maintaining the biological activities to transform 2,4-
DNT to 2,4-DAT. The rate of the biotransformation of
2,4-DNT was very low in the initial stage of the process,
indicating that 2,4-DNT inhibited its own biotransforma-
tion. The presence of 2,4-DNT and its intermediates also
exhibited inhibition to the bioconversion of the primary
substrate (ethanol, methanol, or acetic acid). The higher
the initial concentration of 2,4-DNT, the longer was this
period of inhibition to the conversion of the primary
substrate. Ethanol, methanol, and acetic acid were rap-
idly converted by the bacteria after 2,4-DNT and its
biotransformation intermediates were completely trans-
formed to 2,4-DAT.
To prove that hydrogen was the electron donor for the
reductive biotransformation of 2,4-DNT, the same batch
test was conducted with hydrogen as the primary sub-
strate. The results are shown in Figure 2. In the control
reactors without 2,4-DNT, hydrogen with CO2 in the
reactors was immediately converted to methane (Figure
2a). When 2,4-DNT was initially present in the reactors,
hydrogen was first consumed for the biotransformation
of 2,4-DNT to 2,4-DAT, and for supporting the growth of
the bacteria. Methane was produced from the excess
hydrogen after 2,4-DNT, 4-A-2-NT, and 2-A-4-NT were
completely transformed to 2,4-DAT. This phenomenon
indicates that 2,4-DNT, 4-A-2-NT, and 2-A-4-NT also
inhibited the hydrogen-utilizing methanogenesis. The
higher the initial concentration of 2,4-DNT, the more
hydrogen was consumed for 2,4-DNT biotransformation
and the less hydrogen was left for methane production
(Figure 2b-f).
References
1. Keither, L.H., and W.A. Telliard. 1979. Priority pollutants I. A per-
spective view. Environ. Sci. Technol. 13:416-423.
2. McCormick, N.G., J.H. Cornell, and A.M. Kaplan. 1978. Identifica-
tion of biotransformation products from 2,4-dinitrotoluene. Appl.
Environ. Microbiol. 35:945-948.
3. Berchtold, S.R., S.L. VanderLoop, M.T. Suidan, and S.W. Maloney.
1995. Treatment of 2,4-dinitrotoluene using a two-stage system:
Fluidized bed anaerobic GAC reactors and aerobic activated
sludge reactors. Water Environ. Res. In press.
4. Cheng, J., Y. Kanjo, M.T. Suidan, and A.D. Venosa. 1995. Anaero-
bic biotransformation of 2,4-dinitrotoluene with ethanol as primary
substrate: mutual effect of the substrates on their biotransforma-
tion. Submitted for publication in Wat. Res.
106
-------
o
S
o
*O
O
P,
2.0 0.02
20 40 60 80 100 120
M
2.0
- 1,5
- 1.0
- 0.5
0.0
G
2 0.10
20 40 60 80 100 120
(e)
0.00
0.01
0.00
0 20 40 60 80 100 120
0.20
0.0 0.00
O
CD
O
O
s-c
CX
CD
0.0
0 20 40 60 80 100 120
(f)
cd
C/3
d
o
o
OJO
o
0.0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
Time, Hrs
2,4-DNT 4-A-2-NT 2-A-4-NT 2,4-DAT
o Hydrogen Consumed ° Methane Produced
Figure 2. Anaerobic biotransformation of 2,4-DNT with hb as the primary substrate.
107
-------
Surfactants in Sediment Slurries: Partitioning Behavior and Effects on Apparent
Polychlorinated Biphenyl Solubilization
Jae-Woo Park, John F. Quensen, III, and Stephen A. Boyd
Michigan State University, East Lansing, Michigan
It is generally believed that the biodegradation of poorly
water soluble compounds in soil or sediment systems is
limited by low bioavailability due to strong sorption of the
compounds to natural organic matter. The use of surfac-
tants to increase the apparent water solubility of such
contaminants has often been suggested as a way of
increasing their bioavailability to degrading microorgan-
isms. A possible limitation of this approach is that solu-
bility enhancement is much greater above the critical
micelle concentration (CMC) of the surfactant than be-
low it, and these supra-CMCs are often toxic or inhibitory
to bacteria. A few reports, however, indicate that sub-
CMC concentrations of surfactants may enhance the
anaerobic dechlorination of aromatic compounds. The
goals of our present research efforts are to determine if
sub-CMCs of surfactants can enhance the microbial
dechlorination of polychlorinated biphenyls (PCBs) and,
if so, by what mechanism(s). We have determined the
partitioning behavior of several surfactants in soil and
sediment slurries and their effects on PCB solubilization.
These experiments were undertaken to determine if an
increase in the apparent aqueous solubility of PCBs by
sub-CMCs of these surfactants is a plausible mechanism
for any observed enhancement of PCB dechlorination.
The sorption of four commercial nonionic surfactants
(Triton X-100, Triton X-405, Triton X-705, and Tween 80)
onto the Red Cedar River sediment used in our PCB
dechlorination assays was evaluated. Sorption iso-
therms were plotted, and Freundlich isotherms of the
form Cs=KCen were fitted to the experimental data
where Cs is the sorbed concentration of the surfactant
(mg/kg), Ce is the aqueous concentration of the surfac-
tant (mg/L), and K and n are constants. K and n values
ranged from 1.193x 10'4to 1.009 x 10'3 and from 0.232
to 0.696, respectively. The Red Cedar River sediment
thus shows orders of magnitude less surfactant sorption
than has been reported for soils, as shown by the low K
value.
The distribution coefficients of three PCB congeners at
sub- and supra-CMC surfactant concentrations (up to
four times the CMC) were determined using [14C]labeled
PCBs. The aqueous-phase PCB concentrations in-
creased at all surfactant concentrations tested com-
pared with the sediment-water system without
surfactants. Notably, this included an increase in the
aqueous-phase concentrations of PCBs even at the
lowest surfactant concentration tested (0.05 times
CMC), especially for the inherently less soluble hexa-
and tetra-CBs by Tween 80. In fact, Tween 80 increased
the solubility of 2,2',4,4',5,5'-CB by a factor of 3.3 at
25 percent of its CMC, and by a factor of 6.3 at 75
percent of its CMC.
The low sorption of the surfactants by Red Cedar River
sediments has important consequences for PCB solubi-
lization. Surfactant monomers sorbed to soils or sedi-
ments will increase the total organic matter content of
the solids and act as an additional sorptive phase. Con-
sequently, if surfactants strongly sorb to the sediments,
they may actually decrease the aqueous phase concen-
tration of nonionic compounds such as PCBs. When the
mass of sorbed surfactant is small, however, as in the
case of the Red Cedar River sediment, most of the
surfactant mass exists in the water, and the PCB solu-
bilization effect of aqueous phase surfactant micelles
and monomers dominates the sorptive capabilities of
sediment-associated surfactant and native organic mat-
ter. Therefore, the aqueous phase concentration of PCB
increases even when relatively small amounts of surfac-
tants are added to the system. While these solubility
enhancements are small relative to those that occur
above the CMCs of these surfactants, the increased
solubility may be enough to significantly increase the
rate of PCB dechlorination, especially for the more chlo-
rinated and less water soluble congeners.
108
-------
Partial Characterization of an Anaerobic, Aryl, and
Alkyl Dehalogenating Microorganism
Xiaoming Zhang
National Research Council, Athens, Georgia
W. Jack Jones and John E. Rogers
U.S. Environmental Protection Agency, Athens, Georgia
Introduction
To better understand controls and pathways of anaero-
bic biotransformation of organic pollutants in contami-
nated environments, pure culture studies are beneficial.
To date, only a few strains of anaerobic dehalogenating
microorganisms have been isolated and characterized.
Among these, Desulfomonile tiedjei\s probably the most
widely studied (1). In this study, we report the partial
characterization of an anaerobic bacterium capable of
both aryl and alkyl reductive dehalogenation.
Results and Discussion
An anaerobic bacterium, designated as strain XZ-1, was
isolated from freshwater pond sediment near Athens,
Georgia. Isolate XZ-1 is a sporeforming, motile rod ca-
pable of reductive dehalogenation of chlorophenols.
Electron acceptors, including sulfite, thiosulfate, and ni-
trate (but not sulfate), stimulated growth in the presence
of yeast extract and pyruvate. None of the following
supported growth or dehalogenation of chlorophenol
(CP): glucose, fructose, galactose, rhamnose, cello-
biose, xylan, ribose, citrate, fumarate, acetate, peptone,
tryptone, casein hydrolysate, and casamino acids. The
addition of 1 mM carbon dioxide reduced the lag
time before growth. No growth was observed in the
presence of 4 percent air or higher. Growth was
completely inhibited by pentachlorophenol (PCP)
(>32u,M), 2,3,4,5-tetraCP (> 8 u,M), 3,4,5-triCP
(>16uM), 3,5-diCP(>120uM), 2,4-diCP (> 500 uM),
and 2-CP (>4,OOOu,M). The generation time of isolate
XZ-1 was 1.8 hr at pH 7.5 (optimal) and 30°C.
Isolate XZ-1 removed orffto-chlorines from all ortho-
chlorine-containing phenols tested (e.g., 2-CP and pen-
tachlorophenol). Hydrogen, formate, ethanol, pyruvate,
and yeast extract served as electron donors for dehalo-
genation of CPs. Only pyruvate and yeast extract,
however, stimulated growth either in the absence or
presence of electron acceptors, including 3-chloro-4-hy-
droxyphenylacetate (an analog of ortho-CP). The aryl
dehalogenation activity was inducible, and induction
was inhibited by addition of chloramphenicol to cell
suspensions. Experiments with D2O demonstrated
that water was the exclusive proton source for aryl
dehalogenation of chlorophenols. Proton nuclear mag-
netic resonance (NMR) studies indicated that hydrogen
was incorporated at the same position where an ortho-
chlorine was removed. Product solvent isotope effects
were 5.4 and 8.5 for dechlorination of 2,3-diCP and
2-CP, respectively. An increase in the assay temperature
reduced the product solvent isotope effect in 2,3-diCP
dechlorinations.
Cell suspensions of isolate XZ-1 also were capable of
reduction of 2,4,6-trinitrotoluene (TNT, 46 ppm) to
2,4,6-triaminotoluene via 2-amino-4,6-dinitrotoluene,
4-amino-2,6-dinitrotoluene, 2,6-diamino-4-nitrotoluene,
2,4-diamino-6-nitrotoluene, and several unidentified
intermediates. The TNT transformation pattern was dif-
ferent in aryl dehalogenation-induced cells and non-
induced cells. The identified intermediates of TNT
reduction accumulated to lower levels in the induced
cells than in the noninduced cells. Addition of pyruvate
stimulated TNT transformation. Heat-treated cell sus-
pensions exhibited only traces of TNT transformation
activity either with or without addition of pyruvate. Cell
suspensions of isolate XZ-1 also metabolized chloram-
phenicol in the presence of pyruvate. No intermediate(s)
of chloramphenicol transformation has been identified
to date.
Both noninduced and aryl dehalogenation-induced cell
suspensions of isolate XZ-1 dechlorinated tetrachlo-
roethene to trichloroethene (TCE). A comparison of aryl
109
-------
and alkyl dehalogenation rates in noninduced and in- and to characterize further the physiology, nutrition, and
duced cells suggests that at least two enzymes are phylogeny of this anaerobe.
responsible for the two activities. Aryl dehalogenation-
induced cells also slowly dechlorinated TCE to cis-1, Reference
- IC oroe ene. 1. Mohn, W.W, J.M. Tiedje. 1992. Microbiol. Rev. 56:482.
Additional studies are underway to identify the range of
transformation activities of dehalogenating isolate XZ-1
110
-------
Microbial Degradation of Petroleum Hydrocarbons in Unsaturated Soils:
The Mechanistic Importance of Water Potential and the Exopolymer Matrix
Patricia A. Holden
Department of Environmental Science, Policy, and Management, University of California,
Berkeley, California
James R. Hunt
Department of Civil Engineering, University of California, Berkeley, California
Mary K. Firestone
Department of Environmental Science, Policy, and Management, University of California,
Berkeley, California
Background
Total soil water potential, ₯, is the potential energy per
unit volume of unsaturated soil water and is commonly
reported in -MPa (1). Matric water potential, *₯m, is the
largest component of *P in most soils and arises from the
interaction of soil water with soil surfaces. Matric poten-
tial determines water film thickness in soil, and thus
controls gas-phase mass transferthrough soil pores and
solution-phase mass transfer through water films. Bac-
teria in biofilms are in equilibrium with water in their
environment, and adaptation to a given soil water poten-
tial or to a changing water potential condition during
wetting and drying will affect intrinsic bacterial physiol-
ogy and biofilm characteristics (2). Because of its role in
both mass transfer and bacterial reaction rates, soil
water potential is an important environmental factor con-
trolling petroleum biodegradation rates in unsaturated
soils.
Project Framework
Our biodegradation model for oil constituents at the
biofilm scale contains the following parameters that will
vary as a function of *P:
qm = intrinsic molar removal rate per area of
biofilm (moles/#-t)
Ks = intrinsic half saturation constant (moles/L3)
Lb = total biofilm thickness (L)
De = effective diffusivity of contaminant through
biofilm (L2/t)
pm = number density of bacteria per mass of
biofilm (#/m)
pb = density of biofilm or mass of biofilm per
biofilm volume (m/L3)
Csat = aqueous solubility of petroleum hydrocarbon
(moles/L3)
Experimental protocols include determining each pa-
rameter as a function of *P and determining the overall
removal rate in unsaturated soil as a function of *P. We
are also examining how physicochemical properties of
the bacterial matrix are altered with *P to effect hydro-
carbon solubility and spreadability.
Biofilm Reactors and Preliminary Results
Phenanthrene, hexadecane, and methyl-decalin are the
selected test substrates representing the three major
classes of petroleum constituents. Pristane is the con-
servative tracer. Polyethylene glycol, a nonpermeating
solute with a molecular weight of 8,000 (PEG 8,000), is
used to set matric water potential in well-mixed and
biofilm culture systems.
Custom-designed biofilm reactors for developing
biofilms under unsaturated conditions have been con-
structed and are being tested using various growth sub-
strates. Transmission electron micrographs taken
through biofilms grown under ^-controlled conditions
reveal architectural changes, specifically cell packing
and morphological, with *P. Preliminary diffusion studies
suggest that diffusional mass transfer through biofilms
is related to the ^-condition during growth.
111
-------
References 2- Harris, P.P. 1981. Effect of water potential on microbial growth and
activity. In: Parr, J.F., W.R. Gardner, and L.F. Elliott, eds. Water
potential relations in soil microbiology, SSSA special publication,
1. Jury, W.A., W.R. Gardner, and WH. Gardner. 1991. Soil physics, VoL 11> Number 6, Madison, Wl: Soil Science Society of America.
5th ed. New York: John Wiley & Sons. PP- 733-740-
112
-------
Metabolic Indicators of Anaerobic In Situ Bioremediation of
Gasoline-Contaminated Aquifers
Harry R. Seller, Martin Reinhard, and Alfred M. Spormann
Department of Civil Engineering, Stanford University, Stanford, California
Bioremediation is one of a limited number of options for
restoring aquifers contaminated with the hazardous aro-
matic hydrocarbons that occur in unleaded gasoline,
such as benzene, toluene, ethylbenzene, and the xyle-
nes (BTEX). Considering the cost and technical difficulty
associated with introducing oxygen into some aquifers,
in situ bioremediation using indigenous, anaerobic bac-
teria merits serious consideration for some contami-
nated sites. A major impediment to the acceptance of in
situ bioremediation is the difficulty of demonstrating that
decreases in the concentrations of BTEX in ground
watertruly represent bacterial metabolism of these com-
pounds rather than abiotic processes such as sorption,
dilution, or volatilization.
Work in our laboratory has included the characterization
of byproducts of alkylbenzene metabolism by pure and
mixed anaerobic cultures (1, 2). This research, which
has focused on sulfate-reducing cultures, has involved
the extensive use of gas chromatography/mass spec-
trometry for metabolite characterization. We have re-
cently integrated such laboratory findings with field data
from a controlled-release experiment conducted at the
Seal Beach Naval Weapons Station in California.
Based on the concordance of laboratory studies of an-
aerobic bacteria and field observations from the aquifer
in Seal Beach, we propose a group of compounds in-
cluding benzylsuccinic acid, benzylfumaric acid (or a
closely related isomer), and the o-, m-, and p-methyl
homologs of these compounds as biogeochemical indi-
cators of in situ anaerobic alkylbenzene metabolism in
gasoline-contaminated aquifers. Under the controlled
conditions of the field study, a strong correspondence
was observed between the disappearance of alkylben-
zenes from ground water overtime and the appearance
of associated metabolic byproducts. This correspon-
dence was both qualitative (i.e., only products specific
to the metabolism of toluene, o-xylene, and m-xylene
were observed, and only these three hydrocarbons were
depleted) and quantitative (i.e., metabolic byproduct
concentrations tended to increase as the associated
alkylbenzene concentrations decreased).
References
1. Seller, H.R., M. Reinhard, and D. GrbiCE-GaliCE. 1992. Metabolic
byproducts of anaerobic toluene degradation by sulfate-reducing
enrichment cultures. Appl. Environ. Microbiol. 58:3,192-3,195.
2. Seller, H.R. 1995. Anaerobic metabolism of toluene and other
aromatic compounds by sulfate-reducing soil bacteria. Ph.D. dis-
sertation. Stanford University, Stanford, CA.
113
-------
Contaminant Dissolution and Biodegradation in Soils Containing
Nonaqueous-Phase Organics
Larry E. Erickson, L.T. Fan, J. Patrick McDonald, George X. Yang, and Satish K. Santharam
Kansas State University, Manhattan, Kansas
Several models have been developed to describe the
dissolution, adsorption to soil, and biodegradation of
nonaqueous-phase contaminants (i.e., hydrocarbons) in
the subsurface, and remediation times have been esti-
mated for various conditions (1-6). The significant rate-
limiting factors determining the required bioremediation
time appear to be the rates of transport of the electron
acceptor or oxygen and organic contaminants in pores
and soil aggregates in the vicinity of the hydrocarbon
phase. The contaminants' solubilities in the aqueous
phase affect their dissolution and transport. Contami-
nant dissolution and transport are more rapid than oxy-
gen transport for more water-soluble compounds such
as toluene, and less rapid for less soluble compounds
such as pyrene. As long as the nonaqueous phase is
present, the higher the solubility of a compound, the
greater the extent of removal by pump-and-treat opera-
tions rather than by oxygen-limited biodegradation. The
sizes of aggregates and hydrocarbon blobs significantly
affect remediation time, which has been found to be
proportional to the square of the characteristic length of
the blob.
The available experimental data for pyrene and anthra-
cene fit well with the results of simulation obtained with
one of the models. Besides dissolution, adsorption,
desorption, and biodegradation, this model takes into
account the hydrocarbon-phase size distribution; more-
over, it expresses the rate of biodegradation by Monod
kinetics.
References
1. Yang, X., I.E. Erickson, and L.T. Fan. 1993. Transport properties
of toluene as a non-aqueous phase liquid in ground water. In:
Proceedings of the 8th Conference on Hazardous Waste Re-
search. Manhattan, Kansas: Kansas State University, pp. 313-330.
2. McDonald, J.P., C.A. Baldwin, I.E. Erickson, and L.T. Fan. 1993.
Modeling bioremediation of soil aggregates with residual NAPL
saturation. In: Proceedings of the 8th Conference on Hazardous
Waste Research. Manhattan, Kansas: Kansas State University, pp.
346-365.
3. Gandhi, P., L.E. Erickson, and L.T. Fan. 1995. A simple method to
study the effectiveness of bioremediation aided, pump-and-treat
technology for aquifers contaminated by non-aqueous phase liq-
uids, I. Single component systems. J. Haz. Mat. 39:49-68.
4. Gandhi, P., L.E. Erickson, and L.T. Fan. 1994. A simple method to
study the effectiveness of bioremediation aided, pump-and-treat
technology for aquifers contaminated by non-aqueous phase liq-
uids, II. Multi-component systems. J. Haz. Mat. In press.
5. Yang, X., L.E. Erickson, and L.T. Fan. 1994. Astudy of dissolution
rate-limited bioremediation of soils contaminated by residual hy-
drocarbons. J. Haz. Mat. In press.
6. Santharam, S.K., L.E. Erickson, and L.T. Fan. 1994. Modeling the
fate of polynuclear aromatic hydrocarbons in the rhizosphere. In:
Proceedings of the 9th Annual Conference on Hazardous Waste
Remediation. Manhattan, Kansas: Kansas State University, pp.
333-350.
114
-------
Protein Expression in Mycobacteria That Metabolize Polycyclic
Aromatic Hydrocarbons
David E. Wennerstrom
University of Arkansas for Medical Sciences, Little Rock, Arkansas
Carl E. Cerniglia
National Center for Toxicological Research, Jefferson, Arkansas
Three species of mycobacteria have been isolated from
petroleum contaminated soil (Mycobacterium sp.
PYR-1) (1) or coal gassification sites (Mycobacterium
sp. PAH135 and M. gilvum) (2, 3). These organisms
have potential application in the bioremediation of poly-
cyclic aromatic hydrocarbons (PAHs) because each can
mineralize various PAHs, including naphthalene, phen-
anthrene, pyrene, and fluoranthene. The present study
was initiated to investigate the molecular basis for the
degradation of PAHs by these species of mycobacteria.
To determine part of the physiological response of the
organisms to the presence of a metabolizable PAH in
the environment, we have analyzed the expression of
proteins by each organism in response to pyrene using
two-dimensional sodium dodecylsulfate-polyacrylamide
gel electrophoresis (SDS-PAGE). For each organism,
the pattern of separated proteins was distinct, and pro-
teins increased in expression following addition of the
PAH. Major proteins increased in induced cells of Myco-
bacterium sp. PYR-1 had approximate masses of 105,
79, 53, 42, and 15 kDa. In comparison, three proteins
were induced in Mycobacterium sp. PAH135 (95, 70,
and 53 kDa) and in M. gilvum (72, 27, and 15 kDa). To
determine whether increased expression of these pro-
teins is associated with metabolism of pyrene in Myco-
bacterium sp. PYR-1, uninduced cells were incubated
with the PAH for varying periods up to 8 hr, and the
amounts of pyrene metabolism and protein expression
were quantified by high-performance liquid chromatog-
raphy (HPLC) analysis and densitometry of proteins
detected in two-dimensional SDS-PAGE gels, respec-
tively. After a delay of about 1 hr, uninduced cells
metabolized all of the pyrene within 8 hr. The 79 kDa
protein, undetectable in uninduced cells, was expressed
at 1.2 percent of proteins within 2 hr and was fully
expressed at about 2 percent of total protein within 4 hr.
Partial characterization of this protein by N-terminal se-
quencing and hybridization of a synthetic oligonu-
cleotide probe corresponding to the amino acid
sequence to Bamlll-digested Mycobacterium sp. PYR-1
deoxyribonucleic acid (DMA) show that this protein is
similar to the kaIG gene product (catalese-peroxidase)
expressed in many other mycobacteria. Kinetics of in-
creased expression of the 15 kDa protein followed those
for the 79 kDa protein. In contrast, the 42 kDa protein
was not increased until 6 hr and was not fully expressed
even at 8 hr after addition of pyrene. A variant of the
organism was isolated that failed to metabolize pyrene
and fluoranthene added to soft agar overlays or in liquid
cultures. The variant retained the ability to metabolize
naphthalene and phenanthrene. None of the proteins
studied was induced in this organism after exposure to
3 u,g/ml_ pyrene for 24 hr. These results indicate that
additional components are required for metabolism of
pyrene and fluoranthene compared with those for meta-
bolism of naphthalene and phenanthrene in Mycobac-
terium sp. PYR-1. Our results suggest that the proteins
studied are associated with metabolism of pyrene in
induced cells of this organism. These results provide
fundamental information about the proteins expressed
by these mycobacteria during PAH degradation. Clearly,
this information will be important for future application of
these mycobacterial strains as inoculants in the biore-
mediation of PAH-contaminated sites.
115
-------
References 3- Boldrin, B., A. Tiehm, and C. Fritzsche. 1993. Degradation of phen-
anthrene, fluorene, fluoranthene, and pyrene by a Mycobacterium
1. Heitkamp, M.A., W. Franklin, and C.E. Cerniglia. 1988. Microbial sp. Appl. Environ. Microbiol. 59:1,927-1,930.
metabolism of polycyclic aromatic hydrocarbons: Isolation and
characterization of pyrene-degrading bacterium. Appl. Environ. Mi-
crobiol. 54:2,549-2,555.
2. Grosser, R., D. Warshawsky, and J.R. Vestal. 1991. Indigenous
and enhanced mineralization of pyrene, benzo(a)pyrene, and car-
bazole in soils. Appl. Environ. Microbiol. 57:3,462-3,469.
116
-------
Section 6
Hazardous Substance Research Centers
The Hazardous Substance Research Centers (HSRCs) conduct research on bioremediation
under the direction of ORD's National Center for Extramural Research and Quality Assurance.
Research is sponsored by the following centers: the Northeast Hazardous Substance Research
Center (Regions 1 and 2), the Great Lakes and Mid-Atlantic Hazardous Substance Research Center
(Regions 3 and 5), the South/Southwest Hazardous Substance Research Center (Regions 4 and
6), the Great Plains and Rocky Mountain Hazardous Research Center (Regions 7 and 8), and the
Western Region Hazardous Substance Research Center (Regions 9 and 10).
The symposium's poster session included presentations on the co-metabolic biodegradation kinet-
ics of trichloroethylene in unsaturated soils; the effect of water potential on biodegradation kinetics
and population dynamics; developments in anaerobic and aerobic bioventing; developments in the
treatability protocol for co-metabolic bioventing; the environmental safety of commercial oil spill
bioremediation agents; the effectiveness of gas-phase bioremediation stimulating agents for un-
saturated zone in situ bioremediation; protein expression of mycobacteria that metabolize polycyclic
aromatic hydrocarbons; a field evaluation of pneumatic fracturing enhanced bioremediation; and
the solids suspension characteristics related to slurry biotreatment performance.
117
-------
Co-metabolic Biodegradation Kinetics of Trichloroethylene in Unsaturated Soils
Karen L. Skubal and Peter Adriaens
University of Michigan, Ann Arbor, Michigan
The ability of methanotrophicand heterotrophic bacteria
to aerobically transform chlorinated solvents is well es-
tablished. Methane monooxygenase (MMO) and aryl
monooxygenase enzymes, produced by these microor-
ganisms respectively during growth on primary sub-
strates, catalyze the cooxidation and dehalogenation of
chlorinated ethenes including trichloroethylene (TCE)
and vinyl chloride. Bioventing may prove useful for
stimulating co-metabolism and achieving in situ reme-
diation of vadose zone soils contaminated with chlorin-
ated alkenes. This possibility motivated an investigation
of co-metabolic dechlorination by indigenous microbial
populations in soils collected from Wurtsmith Air Force
Base (AFB) in Oscoda, Michigan.
Contaminated aquifer and vadose regions at Wurtsmith
AFB contain perchloroethylene, TCE (up to 1 ,OOOu,g/L),
trans-dichloroethylene, vinyl chloride, dichlorobenzenes,
and benzene, toluene, ethylbenzene, and xylene (BTEX)
compounds. High methane concentrations have also
been detected in soil gas at the site, indicating poten-
tially favorable conditions for methanotrophic bacteria.
Sandy soils from several depths are being characterized
and studied in aerobic batch microcosm systems at
room temperature to discern the relative importance of
methanotrophic and heterotrophic organisms, and to
optimize methods for their stimulation. Methanotrophs
are supplied with oxygen and methane, while het-
erotrophs are supported on toluene as the primary in-
ducing substrate. A range of environmentally relevant
concentrations is studied, and following an acclimation
period TCE is added at approximately one-tenth the
level of primary substrate. The effect of soil moisture on
biodegradation kinetics is examined by comparing mi-
crocosms containing soil maintained at the local water
content of 4 percent to microcosms containing saturated
soil. In addition, substrate degradation by soil-derived
cultures is monitored in liquid medium without soil.
Bacterial growth on methane and toluene has been
stimulated, and ongoing work will evaluate optimum
primary to co-metabolic substrate ratios and elucidate
the effect of moisture content on TCE co-metabolism in
soil systems. Through development of a simple method-
ology for screening soils and microbial populations in-
digenous to a particular site, this study may clarify the
potential of bioventing to enhance chlorinated solvent
transformation in unsaturated zones containing mixed
wastes.
119
-------
The Effect of Water Potential on Biodegradation Kinetics and
Population Dynamics
Astrid Millers and Peter Adriaens
Department of Civil and Environmental Engineering, University of Michigan,
Ann Arbor, Michigan
Although bioventing is currently being applied in the
field, much remains to be learned about the underlying
parameters controlling biological degradation kinetics in
these systems. These parameters need to be system-
atically studied to improve modeling and design of
bioventing applications. In this investigation, the impact
of subsurface moisture content on biokinetic parameters
is studied, and the applicability of biological kinetics
obtained in saturated batch systems to the unsaturated
zone is evaluated. Specific emphasis is placed on study-
ing the effects of water potential on oxygen availability,
microbial metabolism, and growth.
Mixed culture studies with indigenous microorganisms
derived from the unsaturated zone at the Wurtsmith field
site, Oscoda, Michigan, have been performed in batch
systems. No degradation of toluene was detected at a
field moisture of about 3 percent (by weight) even after
a month of incubation. Moisture contents between 12
and 16 percent moisture exhibited the fastest degrada-
tion of toluene. Differences in biodegradation kinetics
observed as a function of moisture content and inde-
pendent of population shifts are being verified using
pure cultures of a toluene-degrading microorganism,
isolated from the same unsaturated soil samples.
Water potential, the thermodynamic variable expressing
water activity and therefore water availability for the
microorganisms, is used as the experimental variable
rather than the gravimetric moisture content. Varying
water contents of the soil as a result of drying due to
airflow in bioventing operations influence the different
components of the water potential in the soil matrix. The
osmotic and matric water potential components are
studied separately in their effect on bacterial growth,
energy production, and degradation kinetics. Bacteria
isolated from an unsaturated zone below Wurtsmith Air
Force Base are grown in liquid culture on toluate at
different concentrations of membrane diffusable solutes
(NaCI) and nondiffusable solutes (polyethyleneglycol,
PEG). PEG is used to simulate the effect of matric
potential independent of the effect of mass transfer limi-
tations resulting from moisture changes in porous me-
dia. Salt additions to the liquid medium resulted in higher
growth rates of toluate degraders up to 0.2 M NaCI and
increased CO2 evolution. The amount of adenosine
triphosphate produced appeared to be independent of
the salt addition. Studies will be extended to assess
growth in homogenous solids of defined pore structure,
and mixed population studies will be performed to as-
sess the separate effect of population shifts. The results
from these studies will serve as a model for water po-
tential induced microbial stress in unsaturated soil hori-
zons during bioventing.
120
-------
Anaerobic-Aerobic Bioventing Development
Gregory D. Sayles
National Risk Management Research Laboratory,
U.S. Environmental Protection Agency, Cincinnati, Ohio
Munish Gupta and Makram T. Suidan
Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, Ohio
Motivation
Surface spills and leaking pipelines and underground
storage tanks can result in vadose zone soils contami-
nated with hazardous chemicals. The vadose zone must
be cleaned to minimize contamination of ground water
and emissions of volatile organic chemicals to the at-
mosphere.
Aerobic bioventing is now a common approach to treat-
ing aerobically biodegradable contaminants (e.g., fuels)
in the vadose zone. Many highly chlorinated aromatics
and aliphatics can be destroyed microbiologically, most
rapidly by sequential anaerobic-aerobic treatment. Usu-
ally, the biochemical pathway providing the highest rate
for the initial steps of microbial destruction of the highly
chlorinated organic is anaerobic reductive dechlorina-
tion. Once partially dechlorinated, the resulting com-
pounds typically degrade faster under aerobic, oxidizing
conditions. For example, perchloroethylene, wood treat-
ing wastes containing pentachlorophenol (PCP), poly-
chlorinated biphenyls (PCBs), and many chlorinated
pesticides can be at least partially dechlorinated by micro-
bial consortia under anaerobic conditions. The resulting
dechlorinated product often can be destroyed microbi-
ologically under aerobic conditions.
For treatment of large volumes of soil at depth, the most
cost-effective approach is likely to be an in situ process
that takes advantage of the above well-known anaerobic
and aerobic biological activities. Our approach to this
challenge is to develop "anaerobic/aerobic bioventing."
Approach
Anaerobic-aerobic bioventing requires development of
a new process: anaerobic bioventing. Conditions must
be established in the contaminated vadose zone to
induce anaerobic microbial biodegradation. Anaerobic
conditions may be established by injecting nitrogen into
the vadose zone to displace all oxygen. Avolatile cosub-
strate in the nitrogen stream may be needed to induce
the soil microorganisms to consume all the oxygen and
to establish a low oxidation-reduction potential (ORP).
The low ORP should induce dechlorination of the con-
taminant. The cosubstrates must be volatile and con-
sumable by anaerobic microorganisms (e.g., ethanol,
acetone, hydrogen).
Once the contaminant is fully or mostly dechlorinated,
aerobic bioventing is initiated by injecting air into the
vadose zone. The aerobic microorganisms should then
complete the mineralization of the contaminant.
To move the technology to the field, the following ques-
tions are currently being addressed with pilot-scale
tests:
Can venting with nitrogen at a low rate establish
adequate anaerobic conditions for dechlorination in
unsaturated soil? What is the cost to supply nitrogen
at the required flow rate?
Can a cosubstrate introduced with the nitrogen
stream promote the dechlorination of target com-
pounds such as PCP in unsaturated soil?
What are the most effective volatile cosubstrates, bio-
logically and economically, to promote anaerobic
dechlorination? (Several will be evaluated.)
Will switching to an aerobic environment (by replac-
ing nitrogen and the primary substrate with air in the
injection stream) promote mineralization of the
dechlorination byproducts?
What classes of compounds are amenable to anaero-
bic/aerobic bioventing? (Several will be evaluated.)
121
-------
Will the addition of hydrogen gas (H2) to the nitrogen
stream aid in the development of reducing environ-
ments and promote dechlorination?
Initial experiments were conducted beginning in June
1995. The tests were conducted in pilot-scale soil col-
umns. Two 4-ft long, 4-in. diameter glass columns were
built to simulate an in situ horizontal column of unsatu-
rated soil. Gas sampling ports were placed along the
length of the column. The column was filled with sand
inoculated with secondary effluent to add biomass and
moisture. Essential inorganic nutrients were also added.
Nitrogen gas passed through an oxygen scrubber was
used to flush the columns of oxygen. At field scale, the
nitrogen could be supplied by tank or, perhaps more
economically, by separating it from atmospheric air us-
ing onsite molecular sieves. The first cosubstrates
evaluated were ethanol and acetone.
122
-------
Development of Co-metabolic Bioventing: Laboratory Tests
Gregory D. Sayles
National Risk Management Research Laboratory,
U.S. Environmental Protection Agency, Cincinnati, Ohio
Alan D. Zaffiro and Jennifer S. Platt
IT Corporation, Cincinnati, Ohio
Background
The objective of this project is to accumulate the infor-
mation necessary to write a protocol for determining
site-specific treatability of co-metabolic bioventing. "Co-
metabolic bioventing" is a bioventing-like process that
delivers air and a volatile co-metabolite to a vadose
zone contaminated with chlorinated solvent (e.g.,
trichloroethylene) to induce in situ biodegradation of the
contaminant.
The process attempts to induce the following biologically
catalyzed (unbalanced) reactions:
Cosubtrate + O2 => CO2 + H2O
TCE + 02 + H20 => C02 + HCI
where the cosubstrate is chosen because it stimulates
the productions of enzymes in the microbial culture that
oxidize trichloroethylene (TCE). Known volatile cosub-
strates include methane, propane, butane, toluene, jet
fuel, gasoline, and isoprene. The co-metabolic biovent-
ing system delivers the cosubstrate and oxygen from
gas injection wells to induce the in situ biodegradation
of the contaminant TCE.
The Remediation Technology Development Forum
Workgroup on In-Situ Bioremediation of Chlorinated
Solvents, a research team with members from industry
and government, plans to conduct field demonstrations
of three in situ bioremediation technologies for the
cleanup of chlorinated solvents at two sites. Laboratory
studies using site-contaminated soils are supporting the
field research. The three technologies are 1) co-metabolic
bioventing of TCE and dichloroethylene contaminated
vadose zone soils, 2) accelerated anaerobic biotreat-
ment of chlorinated ethylene contaminated ground
water, and 3) intrinsic bioremediation of chlorinated eth-
ylene contaminated aquifers. Dover Air Force Base
(AFB), in Delaware was selected as the first site.
One task under the co-metabolic bioventing project was
established to develop a protocol to test site-specific
treatability of co-metabolic bioventing. The U.S. Envi-
ronmental Protection Agency's National Risk Manage-
ment Research Laboratory has assumed primary
responsibility for this task. Information from laboratory
testing using soil from Dover AFB and two to four other
sites will be used to generate the recommended proto-
col. The efficacy of the protocol will be evaluated by
comparing treatability test results with results of at least
two field demonstrations.
Approach
To simulate co-metabolic bioventing in a closed reactor,
a biodegradation test using soil from Dover AFB was
conducted in the following manner:
Reactors were closed bottles containing 100 g of soil.
Soil was unsaturated and not mixed.
Cosubstrate (toluene or propane) was added to the
bottle.
No moisture, nutrients, or microorganisms were
added.
Killed controls were established.
Two sets of bottles were established:
- Reactors that were sacrificed in triplicate at various
times during the test.
- Reactors that were monitored by automated respi-
rometry.
123
-------
Contaminant and cosubstrate loss with time were
monitored with the sacrificial reactors.
Oxygen use and carbon dioxide production with time
were monitored by the respirometer.
The biodegradation test was conducted during June and
July 1995. Earlier tests conducted with clean soils
spiked with TCE and cosubstrates indicated that care
must be taken in the reactor design to minimize abiotic
losses of TCE. Bottles that expose the soil and its
atmosphere only to glass and Teflon showed significant
(greater than 40 percent) loss of TCE in killed controls
within a week. This loss was attributed to sorption into
the Teflon gaskets. We have redesigned the bottles to
maximize glass surfaces and virtually eliminate Teflon
surface area. Preliminary tests showed that these reac-
tors minimized abiotic TCE loss very well.
124
-------
Evaluating the Environmental Safety of Using Commercial Oil Spill
Bioremediation Agents
Jeffrey L. Kavanaugh
Center for Environmental Diagnostics and Bioremediation, University of West Florida,
Pensacola, Florida
C. Richard Cripe and Carol B. Daniels
Gulf Ecology Division, U.S. Environmental Protection Agency, Gulf Breeze, Florida
Rochelle Araujo
Ecosystems Research Division, U.S. Environmental Protection Agency, Athens, Georgia
Joe E. Lepo
Center for Environmental Diagnostics and Bioremediation, University of West Florida,
Pensacola, Florida
The use of commercial bioremediation agents (CBAs)
for reducing the ecological impact of oil spills raises
several risk assessment questions. The presence of
petroleum hydrocarbons may contribute some toxicity;
CBAs, with their associated chemical constituents (e.g.,
nutrients, dispersants, enzymes), microbes, and inert
ingredients, may add to this toxicity either directly or
indirectly through decreases in dissolved oxygen or in-
creases in particulates. In addition, interaction of CBAs
with oil may have other environmental effects, either
through increasing the amount of petroleum hydrocar-
bons available to aquatic organisms (i.e., through bio-
surfactant activity) or by generation of toxic metabolites.
A related issue is whether use of a CBA could reduce
the toxicity of the oil (an efficacy issue).
A tiered approach, with increasing complexity, cost, and
effort, has been proposed to address the environmental
safety of CBA usage. Originally developed for assessing
effluents, 7-day chronic estimator tests using a fish
(Menidia beryllina) and a crustacean (Mysidopsis bahia)
were adapted to evaluate CBAs; the tests utilize end-
points of survival, growth, and, in the case of the mysids,
a measurement of egg production. Tier II evaluates the
toxicity of the CBA, alone and in the presence of a
water-soluble fraction of oil, to provide baseline informa-
tion on CBA toxicity and potential synergism with petro-
leum hydrocarbons. Tier III examines effluents from
flow-through test systems that model a variety of aquatic
habitats (open water, beach, marsh) to assess toxicity
under more realistic conditions, where a CBA and oil are
allowed to interact.
Data are presented on the toxicity of a variety of CBAs
classified by vendors as microbial, nutrient, enzyme,
dispersant, and "other." In the flow-through test systems,
the CBAs exhibited relatively low toxicity, either by them-
selves or in the presence of an artificially weathered oil.
During a particular period, an apparent interaction be-
tween one CBA and oil appeared to increase toxicity in
the marsh system. Toxicity reduction in the sand com-
ponent of the beach test system could not be developed
into an efficacy endpoint because very small quantities
of oil produced measurable effects on a benthic
amphipod.
125
-------
UNIFAC Phase Equilibrium Modeling To Assess the Bioavailability of
Multicomponent Nonaqueous-Phase Liquids Containing
Polycyclic Aromatic Hydrocarbons
Catherine A. Peters
Department of Civil Engineering and Operations Research, Princeton University,
Princeton, New Jersey
This work is part of a project to evaluate bioremediation
of contaminants that are nonaqueous-phase liquid
(NAPL) mixtures of polycyclic aromatic hydrocarbons
(PAHs). This poster presents the first phase of this work,
aimed at gaining a thorough understanding of multicom-
ponent NAPL/water-phase equilibria for very complex
mixtures. This provides basic information about maxi-
mum bioavailable concentrations of PAHs.
The thermodynamics of multicomponent NAPL/water-
phase equilibria can be described with knowledge of the
mixture composition, NAPL-phase activity coefficients,
and pure solute aqueous solubilities. This analysis in-
volves application of the UNIFAC model to predict
NAPL-phase activity coefficients for constituent com-
pounds in four different tar materials for which detailed
composition data are available. This group contribution
method has proven to be useful for complex mixtures
such as coal tars because a mixture is represented by
a relatively small number of functional groups, making
thermodynamic analysis using excess Gibbs energy
models tractable. Forthis work, the molecular structures
of the uncharacterized portions of the tars are approxi-
mated through nonparametric regressions of functional
group characteristics with molecular weight. The UNI-
FAC model was found to predict nearly ideal behavior
for most tar constituents. The activity coefficients range
from 0.14 (quinoline) to 1.27 (ethylbenzene), but the
vast majority of the constituents are predicted to have
activity coefficients between 0.9 and 1.1.
These results provide a firm theoretical basis for making
an assumption of solution ideality for many tar constitu-
ents (i.e., Raoult's law). The robustness of this conclu-
sion is indicated through comparable results across
different tar materials, and through a sensitivity analysis
to the estimated characteristics of the uncharacterized
fractions. These results, in conjunction with laboratory
measurements of PAH biodegradation rates (for individ-
ual compounds and for multiple substrate NAPL sys-
tems), will eventually be integrated into a mathematical
model describing the rate of biotransformation of PAH-
containing NAPL contaminants and the dynamics of the
composition of the NAPL residual.
126
-------
Field Evaluation of Pneumatic Fracturing Enhanced Bioremediation
Sankar N. Venkatraman
Department of Chemical and Biochemical Engineering, Rutgers University,
Piscataway, New Jersey
Thomas M. Boland and John R. Schuring
Department of Civil and Environmental Engineering, New Jersey Institute of Technology,
Newark, New Jersey
David S. Kosson
Department of Chemical and Biochemical Engineering, Rutgers University,
Piscataway, New Jersey
In situ bioremediation is often limited by the rate of
transport of nutrients and electron acceptors (e.g., oxy-
gen, nitrate) to the microorganisms mediating the proc-
ess, particularly in soil formations with moderate to low
permeability. To overcome these rate limitations, an in-
vestigation was conducted to integrate the process of
pneumatic fracturing with bioremediation. Pneumatic
fracturing is an innovative technology that uses high-
pressure air to create artificial fractures in contaminated
geologic formations, resulting in enhanced subsurface
air flow and transport rates. Following fracturing, the
pneumatic fracturing system can also be used to inject
electron acceptors and other biological amendments
directly into a formation to stimulate biodegradation. The
specific bioremediation process evaluated in this project
used amendment injections and low-rate in situ vapor
extraction to provide oxygen and other supplements,
which resulted in the formation of aerobic, denitrifying,
and methanogenic biodegradation zones, spatially dis-
tributed with increasing distance from the fracture inter-
faces. A "countercurrent" bioremediation process was
thus established with respect to the diffusion of contami-
nants towards the fracture interface.
Afield pilot demonstration of the integrated technologies
was carried out at a gasoline refinery site over a
20-month period. Initial site characterization indicated
the presence of BTX at concentrations of up to 1,500
mg/kg soil, as well as other hydrocarbons. The soil at
the site was overconsolidated clayey silt with very low
permeability. The site was pneumatically fractured fol-
lowed by periodic injections of subsurface amendments
over a period of 50 weeks. Results demonstrated that
fracturing increased subsurface permeability by an av-
erage of 36 times. Information gained from periodic
vapor sampling indicated that following subsurface in-
jections, the production of carbon dioxide was enhanced
due to increased biological activity. Following a lag
phase, the methanogens became active, and an in-
crease in methane production was observed. There was
no carbon dioxide or methane detected in the prede-
monstration vapor samples. The mass of carbon con-
verted to carbon dioxide and methane was used as an
independent measure for the depletion of total carbon.
Based on this balance, the Generated to Cbi0degraded ratio
was computed to be 3.3:1, indicating that other carbon
sources in gasoline also served as substrate and par-
ticipated in the biodegradation process. After 1 year of
process operation and monitoring, soil samples ob-
tained from the site indicated a 79-percent reduction in
soil-phase BTX concentrations, and over 85 percent of
the BTX reduction was attributed to biodegradation.
127
-------
Solids Suspension Characteristics Related to Slurry Biotreatment Performances
J.-W. Jim Tzeng, Paul T. McCauley, and John A. Glaser
National Risk Management Research Laboratory, U.S. Environmental Protection Agency,
Cincinnati, Ohio
Introduction
Slurry biotreatment has been demonstrated to be an
effective process for bioremediation of contaminated
soils, sediments, and sludges (1). Solid-phase biotreat-
ment such as composting or formal land treatment units
cannot compete with slurry treatment for the extent of
treatment in short timeframes. Slurry biotreatment has
been commonly conducted in reactor systems such as
agitated tanks or lined lagoons. Slurry systems provide
conditions of improved contact between the pollutant
and the microorganisms responsible for the desired
biotransformation. The extent of particle suspension by
agitation is a crucial factor controlling treatment effi-
ciency. Power input from an impeller system dictates not
only the homogeneity of the slurry medium but also the
degree of particle suspension. Equally importantly, the
power input requirement is an important economic factor
in assessing the feasibility of bioslurry treatment for
particular solids to be treated (2). Very little attention is
given, however, to this important component of the treat-
ment system. This work presents our current technical
efforts on identifying the conditions for optimal slurry
agitation for bioremediation of contaminated solids.
Results
Four flow regimes in terms of particle suspension char-
acteristics have been identified; in increasing order of
impeller power input, they are: nonsuspension,
semisuspension, off-bottom suspension, and complete-
suspension regimes. Experimental results indicate that
unique relationships exist between the flow regimes and
power input. In addition, kinetic energy rather than im-
peller rotational speed dictates particle suspension dy-
namics in a slurry medium. A flow regime map (Figure
1) is constructed using power input as the primary pa-
rameter. At extremely low power inputs, particles remain
stationary and settle on the tank bottom. As the power
input exceeds a certain value, the upper layer of the
settled particles starts to become mobile. With a further
N on-Suspension
Semi-Suspension
Favorable
Operating
Range
0)
c
1
0
_J
in
TJ
0
CO
Power Input
Figure 1. Flow regime map of particle suspension in slurry agi-
tation systems.
increase in power, the layer of settled particles reduces
in thickness, and eventually all particles are mobilized
with a portion of particles moving along the tank bottom.
Such a state corresponds to the minimum off-bottom
suspension as conventionally reported in the literature.
The suspended particles tend to fall back to the tank
bottom, however, due to insufficient momentum trans-
ferred from the liquid medium to particles. Dynamics of
particles in this regime can be described by an up-and-
down motion, and the particle distribution is nonuniform
along the height of the slurry tank. An increase in the
impeller power input reduces the degree of nonunifor-
mity, and the particle distribution becomes rather uni-
form as the power input exceeds the minimum
complete-circulation value.
Location of impellers (e.g., bottom clearance) greatly
affects the particle suspension. A substantial reduction
in power input required for both on-bottom and off-bot-
tom particle suspension is obtained as an impeller is
placed near the tank bottom. The conventional design
of agitated tanks, with bottom clearance equal to the
128
-------
impeller diameter or tank radius, is inadequate for par-
ticle suspension applications.
Conclusion
Feasibility of slurry bioreactors for bioremediation of
contaminated solids depends on the energy consump-
tion required to achieve adequate fluid mixing and par-
ticulate suspension in the slurry medium. Four particle
suspension regimes are identified with respect to the
power input to the slurry medium in an agitated tank.
Solids properties affecting the suspension behavior in-
clude 1) size and shape distribution, 2) density differ-
ence, and 3) solids loading. Operation under the
complete suspension regime for achieving maximum
uniformity of particle suspension may not be necessary
because the treatment efficiency may only be marginally
improved. This is because mass transfer resistance be-
tween particles and the bulk liquid phase is not the only
rate-limiting step in the soil-slurry treatment process.
References
1. U.S. EPA. 1990. Engineering bulletin: Slurry biodegradation.
EPA/540/2-90/016. Cincinnati, OH.
2. Muskett, M.J., and A.W. Nienow. 1988. Capital vs. running costs:
The economics of mixer selection. In: The Institution of Chemical
Engineers, ed. Fluid Mixing III.
129
------- |