United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
E PA/603/R-99/012
February 1999
Biotransformation of
Gasoline-Contaminated
Groundwater Under
Mixed Electron-Acceptor
Conditions
-------
EPA/600/R-99/Q12
February 1999
Biotransformation of Gasoline-Contaminated
Groundwater Under Mixed Electron-Acceptor
Conditions
by
Jeffrey R. Barbara, Barbara J. Butler, arid James F. Barker
University of Waterloo
Waterloo, Ontario, CANADA
Cooperative Agreement CR-821887
Project Officer
Stephen R. Hutchins
Subsurface Protection and Remediation Division
National Risk Management Research Laboratory
Ada, OK 74820
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
Notice
The U.S. Environmental Protection Agency through its Office of Research and Development partially
funded and collaborated in the research described here under Cooperative Agreement No. CR-821887,
Award No, 2160401 to the University of Waterloo, Ontario, Canada. It has been subjected to the Agency's
peer and administrative review, and it has been approved for publication as an EPA document. Mention of
trade names or commercial products does not constitute endorsement or recommendation for use.
All research projects funded by the U.S. EPA are required to participate in the Agency Quality
Assurance Program. This research was conducted under an approved Quality Assurance Project Plan
(June, 1993) and Field Sampling Plan (April, 1995). For some of the laboratory and field activities,
modified experimental objectives, unanticipated difficulties encountered during set up, and equipment
availability necessitated deviations from procedures specified in these documents. Actual procedures used
in the laboratory and field experiments performed in this study are described in detail in Chapter 3 and
Appendix C, and are consistent with common research practices as reported in the peer-reviewed
literature. Information on the Plans and documentation of the quality assurance activities and results are
available from the Principal Investigator.
-------
Foreword
The U.S. Environmental Protection Agency is charged by Congress with protecting the Nation's land,
air, and water resources. Under a mandate of national environmental laws, the Agency strives to formulate
and implement actions leading to a compatible balance between human activities and the ability of natural
systems to support and nurture life. To meet these mandates, EPA's research program is providing data
and technical support for solving environmental problems today and building a science knowledge base
necessary to manage our ecological resources wisely, understand how pollutants affect our health, and
prevent or reduce environmental risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's center for investigation
of technological and management approaches for reducing risks from threats to human health and the
environment. The focus of the Laboratory's research program is on methods for the prevention and control
of pollution to air, land, water, and subsurface resources; protection of water quality in public water systems;
remediation of contaminated sites and ground water; and prevention and control of indoor air pollution. The
goal of this research effort is to catalyze development and implementation of innovative, cost-effective
environmental technologies; develop scientific and engineering information needed by EPA to support
regulatory and policy decisions; and provide technical support and information transfer to ensure effective
implementation of environmental regulations and strategies.
This project represents a cooperative effort between the University of Waterloo and the Environmental
Protection Agency. This report summarizes research conducted using both laboratory batch microcosms
and field-scale sheet-piling cells to evaluate whether bioremediation of monoaromatic fuel hydrocarbons
can be enhanced using mixed rather than single electron acceptors. The studies focused on nitrate for
anaerobic bioremediation and oxygen for aerobic bioremediation, and experiments were designed to test
the hypothesis that low levels of oxygen may enhance biodegradation of more recalcitrant compounds
(such as benzene) under denitrifying conditions. The findings from this project are directly applicable to the
field-scale remediation of subsurface environments contaminated by petroleum hydrocarbons.
linton W. Hall, Director
Subsurface Protection and Remediation Division
National Risk Management Research Laboratory
iii
-------
Acknowledgments
Laboratory assistance and analytical work were provided by Kim Hamilton, Shirley Chatten. Marianne
Vandergriendt, Tracy Fowler, and Dru Heagle. The multilevel TDR probes were designed and built by Dave
Redman. Technical assistance in the field was provided by Paul Johnson, Sam Vales, Rick Johnson, Jesse
Ingleton, and a large group of graduate and undergraduate students. Everton de Oliveira and Leanne
Murdie in particular were integral in the construction of the infrastructure and in subsequent data collection
at CFB Borden. Chris Green, Sam Piperakis, Che McRae, John Miller, Randy Pagan, Bill Lawlor, Dave
Bertrand, Sebastian Kirstine, Matt Bogaart, and Don Forbes also helped in the field. Dr. Steve Hutchins of
NRMRL contributed both technical and logistical support throughout the project, including technical
assistance on various aspects of the research, coordination of analytical work at NRMRL on samples from
our field site, and hospitality and assistance during the first author's four month residence at the Lab.
Finally, the report was improved by constructive comments from three anonymous reviewers.
The gasoline used in this study was kindly provided by the American Petroleum Institute, with the
assistance of Dr. Bruce Bauman. Dr. Mike Barcelona of the National Center for Integrated Bioremediation
Research and Development, University of Michigan, extended the services of his laboratory for the analysis
. of organic acids in Borden groundwater. Additional analytical work on groundwater and core-extract
samples from the Borden field site was performed at NRMRL, as well as the entire 14C mineralization
experiment. The contributions and technical support of these organizations are gratefully acknowledged.
-------
EXECUTIVE SUMMARY
Controlled releases of API 91-01 gasoline within steel sheet-piling cells at CFB Borden and laboratory
microcosm experiments were used to investigate the biotransformation of the aromatic hydrocarbons
benzene, toluene, ethylbenzene, xylene isomers, trimethylbenzene isomers and naphthalene (referred to
as BTEXTMB in this report) under mixed NO.; / 0,, electron-acceptor conditions. The main objective of the
research was to evaluate nitrate-based bioremediation as a remedial technology in a gasoline source area.
Dissolved oxygen was added to potentially enhance the mass loss of the soluble compounds that are
recalcitrant under denitrifying conditions. The controlled field experiment also provided some insight into
the interactions that occur between electron acceptors and soluble organics when a hydrocarbon source
area is flushed with electron acceptors to stimulate in situ biotransformation.
In laboratory microcosm experiments, the effect of limited (microaerophilic) 02 was found to depend
on the concentrations of aromatic hydrocarbons and other carbon compounds present in the system.
When aqueous concentrations of the aromatics were low (10x dilution of gasoline-saturated water), and
there were no other sources of labile carbon, the mass of O in a microcosm was fairly large relative to the
mass of carbon, and aerobic mass losses were observed. Notably, however, benzene mass losses were
typically minimal under these conditions. On the other hand, in gasoline-contaminated microcosms less
extensive mass losses were observed, presumably as a result of 02 consumption by other gasoline
hydrocarbons. When aqueous concentrations of the aromatic hydrocarbons were increased to gasoline-
saturated levels to reflect field conditions, negligible losses of the aromatics were observed despite rapid
consumption of microaerophilic 02. Under these conditions, the mass of 02 was probably too low to
observe any losses, even if the aromatics were utilized in preference to other gasoline hydrocarbons.
Nitrate utilization under anaerobic conditions was observed in most laboratory experiments, but rates
were low relative to 02 consumption, and losses were limited to toluene, ethylbenzene, and less
consistently, m-xylene. After in situ exposure, NO, -reducing activity was not inhibited in gasoline-
contaminated aquifer material, but the labile aromatic hydrocarbons were apparently not the preferred
substrates in this carbon-rich environment. In contrast, other laboratory experiments with gasoline-
contaminated material showed that benzene, toluene, ethylbenzene, m-xylene. p xylene, 1,2,4-
trimethylbenzene, and naphthalene would degrade at the expense of O.,. Under mixed electron-acceptor
conditions, patterns of 02, and N03, and aromatic-hydrocarbon concentrations suggested that 02 and
N03~ were used sequentially; most aromatic-hydrocarbon biotransformation occurred early, likely at the
expense of microaerophilic 02, with additional losses of toluene and ethylbenzene occurring under
denitrifying conditions over longer time periods. When the initial concentrations of the aromatic
hydrocarbons were low, there was a beneficial effect of dual electron acceptors: mass losses in
microaerophilic / N03 microcosms were more extensive than in comparable microaerophilic and anaerobic,
denitrifying microcosms. This showed that under certain conditions the extent of mass loss could be
maximized by the presence of these two electron acceptors. However, in experiments with gasoline-
contaminated aquifer material, which were more representative of in situ conditions, mass losses were
either very small or negligible under mixed microaerophilic/ N03 conditions.
API 91-01 gasoline was released into two treatment cells (70 L per cell) in the Borden aquifer to create
gasoline-contaminated source areas below the water table, and then water amended with different
combinations of electron acceptors was flushed vertically through the cells under highly-controlled flow
-------
conditions. The Nitrate Cell received a mixture of microaerophilic 02 and NOg", and the Control Cell
microaerophilic O, only. Aromatic-hydrocarbon and electron-acceptor utilization was monitored during both
flushing and static periods over a 13 month period.
Data from multilevel piezometers showed that dissolved 02was consumed rapidly to a non-zero
threshold concentration in both treatment cells. Because dissolved O, was not detected at sampling ports
located 60 cm below ground surface, and water was injected at about 50 cm bgs, 02 was apparently
consumed within the first 10 cm of the vertical flowpath. In contrast to the rapid 02 consumption, NO,
utilization was low, but the production of NO; suggested that some biological N03-reduction had been
induced. A mass balance indicated that only 12% of added N03 was consumed over the 174-day flushing
experiment. Dissolved aromatic hydrocarbon concentrations remained fairly high (near gasoline-saturated
levels) in both cells throughout the experiment. The depletion of these compounds was generally
consistent with the dissolution of a multicomponent liquid (i.e., relatively rapid depletion of soluble
compounds such as benzene) with no clear evidence of preferential removal of labile compounds from
microbial activity. After the experiments were completed cores were collected from the cells to measure the
mass of BTEXTMB remaining in the residual gasoline and mass balances were completed. In terms of
total BTEXTMB, 81% and 83% of the initial mass was recovered in the Control and Nitrate Cells,
respectively, which correspond to roughly 2,500 g of unrecovered mass per cell. These losses probably
resulted from a combination of physical losses and error associated with the mass balance procedure, with
only minor contributions from biotransformation.
Mass balance results were used to estimate the amount of aromatic-hydrocarbon mass loss that could
reasonably be attributed to biotransformation. The results suggested that the mass of microaerophilic 02
injected into the treatment cells was too low to observe any losses even if all of the 02 was consumed in
mineralization reactions with aromatic compounds. Similarly, given the limited NO., utilization, the mass
loss of compounds such as toluene that are labile under denitrifying conditions was likely very small relative
to the size of the toluene pool in the Nitrate Cell. Consequently, although there was evidence from
metabolite formation that some biotransformation of aromatic hydrocarbons had occurred, mass losses
appeared to be quite low in both cells. Given these results, gasoline dissolution was the dominant mass
removal mechanism from the treatment cells.
The laboratory microcosm experiments, microbial characterization results, and field data suggested that
with in situ exposure, the aerobic microbial population in the Borden aquifer acclimatized to the gasoline
phase and associated high aqueous concentrations of BTEXTMB. However, on the basis of the small
quantity of 0„ available for mineralization reactions relative to the size of the carbon pool and the limited
utilization of NOa, the majority of the mass of recalcitrant compounds (i.e., benzene) would have been
flushed into the aquifer without undergoing appreciable biotransformation. Therefore, in this aquifer there
appeared to be no advantage associated with the microaerophilic / N03 amended cell relative to the control
during the short flushing period investigated in this study; NOa utilization may have continued further
downgradient, however, providing some benefit to an engineered or intrinsic remediation strategy. Despite
these conclusions, it is possible that mixed electron acceptors would be more effective in other
circumstances. For example, it is conceivable that this approach, particularly with respect to the effects of
microaerophilic 02, would be more effective during the latter stages of an enhanced bioremediation project
when source-area concentrations were lower. Similarly, although microaerophilic 02was not found to be
beneficial in our experimental system, the contribution to in situ mass loss could be significant In other
applications. For example, mixed electron acceptors would potentially be effective for downgradient plume
control using a reactive wall or other semi-passive remedial technology.
vi
-------
TABLE OF CONTENTS
NOTICE ; ii
FOREWORD iii
ACKNOWLEDGMENTS iv
EXECUTIVE SUMMARY vi
CHAPTER 1. INTRODUCTION 1
1.1 Objectives 1
1.2 Overview of Experimental Approach 2
1.2.1 Laboratory Experiments 2
1.2.2 Field Experiments 4
1.3 Background Literature 4
1.3.1 Biotransformation Under Single Electron-Acceptor Conditions 4
1.3.2 Biotransformation Under Mixed Electron-Acceptor Conditions 6
CHAPTER 2. STU D Y ARE A 9
CHAPTER 3. EXPERIMENTAL METHODS ; 11
3.1 Aquifer Core Collection . 11
3.2 Laboratory Microcosm Experiments 11
3.2.1 General Set Up Procedures 11
3.2.2 Acetylene Block 13
3.3 ,4C Mineralization Experiment 13
3.4 Microbial Characterization 14
3.4.1 Enumerations 14
3.4.2 Microbial Dehydrogenase Activity 14
3.5 Borden Field Experiment 15
3.5.1 Instrumentation 15
3.5.2 Gasoline Injection 16
3.5.3 Experimental Design 21
3.5.4 Field Sample Collection 23
CHAPTER 4. LABORATORY EXPERIMENTS 27
4.1 Microcosm Experiments: Pristine Borden Sand 27
4.1.1 Experiment 1: Effect of Oxygen, Nitrate Concentration, and Nutrients 27
4.1.2 Experiment 2; Comparison of BTEX Biotransformation Under
Denitrifying Conditions in CFB Borden, Eglin AFB and
Park City Aquifer Microcosms 31
4.1.3 Experiment 3: Biotransformation of BTEX in Gasoline-Contaminated
Groundwater Under Denitrifying, Microaerophilic and
Mixed Electron-Acceptor Conditions 36
4.1.4 Experiment 4: Effect of BTEXTMB Concentration Under Microaerophilic
Oxygen/Nitrate Conditions 46
4.1.5 Experiment 5: Effect of Oxygen Concentration Under High BTEXTMB
Concentration Conditions 48
4.2 Microcosm Experiments: Gasoline-Contaminated Borden Sand..... 51
4.2.1 Experiment 6: Extent of Biotransformation Under Various Substrate and
Mixed Electron-Acceptor Conditions 51
vii
-------
4.3 Microbial Characterization Results 57
4.3.1 Pristine Aquifer Material 57
4.3.2. Gasoline-Contaminated Aquifer Material 60
4.4 Discussion and Conclusions 62
CHAPTER 5. FIELD EXPERIMENTS 65
5.1 Overview of Results 65
5.1.1 Flow Characteristics 65
5.1.2 Dissolved Oxygen and Nitrate 65
5.1.3 Organics .70
5.1.4 Nitrite Production 78
5.1.5 Metabolite Production 78
5.2 Mass Balance Calculations 81
5.2.1 Dissolved Oxygen 81
5.2.2 Nitrate 82
5.2.3 BTEXTMB 83
5.3 Discussion and Conclusions 86
6.0 CONCLUSIONS AND IMPLICATIONS 91
REFERENCES 93
APPENDIX A. GASOLINE CHARACTERISTICS 99
APPENDIX B. TRACER TEST 107
APPENDIX C. ANALYTICAL METHODS 119
APPENDIX D: AQUIFER MONITORING RESULTS 123
viii
-------
LIST OF TABLES
1-1 Summary of Major Experimental Activities 3
2-1 Chemistry of Unamended Groundwater at the Experimental Site 10
3-1 Summary of Microcosm Designs Used in the Study 12
4-1 Summary of Major Conclusions from Laboratory Experiments 28
4-2 Design Summary, Experiment 1 29
4-3 Nitrate-N and Nitrite-N in Anaerobic Microcosms on Day 253 34
4-4 initial and Final BTEX, D.O., and Nitrate/Nitrite Concentrations in Denitrifying Microcosms 34
4-5 Park City, Eglin AFB Subsurface Materials 34
4-6 Microbial Enumeration of Borden, Park City and Eglin AFB Aquifer Materials 36
4-7 Design Summary, Experiment 3 37
4-8 Design Summary, Experiment 4 46
4-9 Percent of Individual Aromatic Hydrocarbons Remaining in Active, Low-BTEXTMB
Concentration Microcosms Relative to Sterile Controls 47
4-10 Design Summary, Experiment 5 49
4-11 Microbial Enumerations and Hydrocarbon Degrading Activity for Nine Replicates
Selected for Reamendment after Day 163 52
4-12 Design Summary, Experiment 6 52
4-13 Percent of Individual Aromatic Hydrocarbons Remaining in Active, Low- and
High-BTEXTMB Concentration Microcosms Relative to Sterile Controls 56
4-14 Dissolved Nitrous Oxide Concentrations in Selected Microcosms Containing
Contaminated Aquifer Material from the Nitrate Cell 56
4-15 Microbial Enumerations of Borden Cores 58
4-16 Experimental Design: ETS Activity in Pristine Aquifer Material 59
4-17 Experimental Design: ETS Activity in Pristine and Contaminated Aquifer Material 61
5-1 Hydrocarbon Component Classes in API 91-01 Gasoline and Gasoline-Contaminated
Core Extract Samples 78
5-2 Detected Organic Acids and Phenols in API 91-01 Gasoline and Groundwater . 79
5-3 Concentrations in mg/L of Selected Redox-Sensitive Constituents in the Experimental
Cells and Injection Water 80
5-4 Nitrate Mass Balance Results for 174-Day Flushing Experiment 83
5-5 Aromatic Hydrocarbon Mass Balance 84
A-1 Mole Fractions of Identified Compounds in API 91-01 Gasoline 101
A-2 Characteristics of API 91-01 Gasoline 104
A-3 Comparison of Measured Concentrations of Aromatic Hydrocarbons
in API 91-01 Gasoline 104
A-4 Measured and Calculated Concentrations of Aromatic Hydrocarbons in Water
Equilibrated with API 91-01 Gasoline at 10°C 105
B-1 Comparison of Bromide Concentrations Determined from a Bromide Electrode
and Ion Chromatography (IC) , 109
B-2 Bromide Tracer Test Results 117
D-1 Environmental Monitoring Downgradient of Wastewater Treatment Mound and
Treatment Cells 124
tx
-------
LIST OF FIGURES
2-1 Location map of the study area at CFB Borden (from Oliveira (1997)) 9
3-1 Plan view of treatment cells with surveyed locations of instrumentation 16
3-2 Plan view of study area 17
3-3 Vertical profiles of water saturation prior to the gasoline injections 18
3-4 Vertical profiles of total dissolved BTEXTMB in the Nitrate Cell on November 22-26, 1995,
approximately one month after gasoline injection 19
3-5 Vertical profiles of total dissolved BTEXTMB in the Control Cell on November 22-26, 1995,
approximately one month after gasoline injection 20
3-6 Injection system schematic for the Nitrate Cell 22
3-7 Record of injection rates over the 174-day flushing experiment 23
3-8 Schematic of groundwater sample collection apparatus 24
4-1 Aerobic BTEX biotransformation in the presence or absence of NOs and/or
NH4-N and P04-P 30
4-2 Aerobic BTEX biotransformation under NO,-free, N,P-free conditions 30
4-3 BTEX biotransformation in (a) anaerobic sterile control and
(b) anaerobic NO,-free microcosms 32
4-4 BTEX biotransformation in selected anaerobic microcosms 33
4-5 Comparison of BTEX biotransformation in Borden, Eglin and Park City
microcosms under denitrifying conditions 35
4-6 Biotransformation of BTEX in gasoline-contaminated groundwater under microaerophilic
and mixed electron-acceptor conditions 38
4-7 Biotransformation of BTEX in gasoline-contaminated groundwater under
denitrifying and anaerobic, N03 -free conditions 39
4-8 Headspace O, (a) and dissolved 02 (b) in sterile and active microaerophilic microcosms 40
4-9 Nitrate and nitrite in (a) mixed electron acceptor and (b) denitrifying microcosms 42
4-10 BTEX biotransformation in denitrifying, mixed electron-acceptor, and microaerophilic
Borden microcosms - normalized BTEX 44
4-11 BTEX biotransformation in denitrifying, mixed electron-acceptor, and microaerophilic
Borden microcosms - dissolved O,,. N03 and NOa" 45
4-12 Nitrate, N02", D.O., and normalized total BTEXTMB concentrations in
Experiment 4 microcosms 47
4-13 Nitrate, N02. D.O., and total BTEXTMB concentrations in Experiment 5 microcosms 50
4-14 BTEXTMB concentrations in individual replicates for each of the treatment groups in
Experiment 5 50
4-15 Normalized D.O. and total BTEXTMB concentrations in sterile microcosms
with and without gasoline-contaminated aquifer material from the
Nitrate Cell (Experiment 6) 53
4-16 Nitrate, N02, D.O., and total BTEXTMB concentrations in active and sterile control
microcosms with gasoline-contaminated aquifer material from the Nitrate Cell (Experiment 6)
(a) Aerobic (high 02) treatment
(b) Microaerophilic (low-02) treatment
(c) Microaerophilic, 10x BTEXTMB dilution treatment 55
4-17 Effect of gasoline-contaminated groundwater on ETS activity in pristine Borden material 59
-------
4-18 Numbers of denitrifiers, aerobic heterotrophs, and benzene-toluene degraders in
pristine and contaminated aquifer material 60
4-19 Accumulation ot INT formazan as a measure of ETS activity in pristine and
contaminated aquifer material 62
5-1 Injection and extraction D.O. concentrations 66
5-2 D.O. concentrations at 60- and 180-cm bgs ports during the 174-day flushing experiment 66
5-3 Vertical profiles of D.O. at various times during and after the 174-day flushing experiment 67
5-4 Injection N03~ concentrations 68
5-5 Nitrate concentrations in injection water and 60- and 180-cm bgs ports during the
174-day flushing experiment 69
5-6 Nitrate concentrations during the static period between flushing experiments 69
5-7 Concentrations of dissolved aromatic hydrocarbons in samples collected from
extraction-well ports 71
5-8 Concentrations of aromatic hydrocarbons in samples collected from the extraction-well
expressed as percentages of concentrations in water equilibrated with fresh API 90-01
gasoline 71
5-9 Vertical profiles of benzene and total BTEXTMB before, during, and after the
flushing experiments 72
5-10 Concentrations of dissolved aromatic hydrocarbons at individual 180-cm bgs ports in the
Nitrate Cell during the 174-day flushing experiment 73
5-11 Concentrations of dissolved aromatic hydrocarbons at individual 180-cm bgs ports in the
Control Cell during the 174-day flushing experiment.... 74
5-12 Corrected toluene, ethylbenzene, and N03 concentrations from a selected piezometer in the
Nitrate Cell that was sampled during the static period 75
5-13 Vertical profiles of residual BTEXTMB in the Nitrate Cell 76
5-14 Vertical profiles of residual BTEXTMB in the Control Cell 77
A-1 Calculated vs. measured BTEXTMB concentrations in gasoline-saturated water 105
B-1 Normalized bromide concentrations in samples of injection and extraction water during
bromide tracer test (Nitrate Cell) 110
B-2 Normalized bromide concentrations in samples of injection and extraction water during
bromide tracer test (Control Cell) 110
B-3 Measured bromide breakthrough curves at 60-cm bgs ports in the Nitrate Cell 111
B-4 Measured and calculated bromide breakthrough curves at 120-cm bgs ports in the
Nitrate Cell 112
B-5 Measured and calculated bromide breakthrough curves at 180-cm bgs ports in the
Nitrate Cell .113
B-6 Measured bromide breakthrough curves at 60-cm bgs ports in the Control Cell 114
B-7 Measured and calculated bromide breakthrough curves at 120-cm bgs
ports in the Control Cell 115
B-8 Measured and calculated bromide breakthrough curves at 180-cm bgs
ports in the Control Cell 116
-------
CHAPTER 1. INTRODUCTION
1.1 Objectives
In North America leaking gasoline storage tanks are one of the major sources of groundwater
contamination. Gasoline and other fuels contain several regulated compounds, most notably benzene,
a suspected human carcinogen. Spilled gasoline is typically trapped in the vicinity of the water table
as immobile, non-aqueous phase liquid (NAPL). Because soluble gasoline constituents partition to
groundwater, the NAPL is a long-term source of groundwater contamination. Consequently, over the
past two decades remediation of gasoline spills has become an important groundwater quality issue.
The aromatic hydrocarbons benzene, toluene, ethylbenzene, xylene isomers, trimethylbenzene
isomers, and naphthalene (referred to as BTEXTMB in this report) constitute the majority of the mass
that partitions to groundwater. These constituents are more toxic and mobile than other fuel
constituents.
Engineered systems incorporating free product recovery in conjunction with in situ bioremediation
have often been used to remediate fuel spills. Providing sufficient oxygen (02) to the contaminated
area is the major limitation of this approach. The problem of adding sufficient 02 is more severe if
residual fuel is present near the water table. If a plume is biodegrading intrinsically, 02 can only be
replenished by the relatively weak dispersive mixing at the plume boundaries (MacQuarrie et al.,
1389), and consequently anaerobic conditions can persist in the core of the plume. For both
engineered and intrinsic remediation approaches, alternate electron acceptors may therefore play an
important role in limiting contaminant migration. Nitrate (N03) has been investigated as a possible
alternate electron acceptor for aromatic hydrocarbon biotransformation. It is very soluble, less reactive
in anaerobic environments, and provides a high energy yield to denitrifying bacteria; as such it may
be useful as a replacement or supplement to 02 in oxygen-limited environments. As the sole electron
acceptor, N03 also has limitations, most significantly the frequent recalcitrance of benzene. In
addition, NO, is a regulated compound, and therefore its fate must also be considered.
To circumvent the limitations of single electron acceptors, recent research has focused on
enhanced bioremediation approaches that rely upon mixed electron-acceptors. In this study we
investigated the use of N03 and 02 mixtures to bioremediate dissolved-phase and residual gasoline.
Dissolved oxygen was added to potentially enhance the mass loss of compounds that are recalcitrant
under denitrifying conditions. It has been hypothesized (e.g., Wilson and Bouwer, 1997) that so-called
"microaerophilic" dissolved 02concentrations (defined here as concentrations below 2 mg/L) could be
utilized for initial oxidation of recalcitrant compounds such as benzene. This would yield partially-
oxidized intermediates susceptible to oxidation under anaerobic, denitrifying conditions further
downgradient, and result in enhanced mass removal. At the field scale, however, there are potential
limitations, such as O, utilization by bacteria growing on non-target organic compounds and abiotic 02
demand from reduced metal species. To date, relatively little research has been performed under
mixed electron-acceptor conditions, and both the fundamental biological processes and field-scale
controls are poorly understood.
The main objective of the research was to evaluate nitrate-based bioremediation as a remedial
technology in a gasoline source area. It was hypothesized that a denitrifying population capable of
rapid aromatic-hydrocarbon biotransformation would develop in the source area in response to
extended NO exposure. The field research was conducted in the Borden aquifer located at Canadian
Forces Base Borden, Ontario. Two controlled gasoline spills were used to generate source areas, and
then water amended with different combinations of N03; and 02 was flushed through the gasoline-
contaminated regions to evaluate the extent of mass loss of the soluble, plume-forming aromatic
1
-------
hydrocarbons. Although both laboratory and field work were performed, the emphasis of the research
was to perform an in situ evaluation of source-area bioremediation under controlled, dynamic
conditions.
The laboratory data were used primarily in support of the field experiment, i.e., to provide data that
could be used in the design of the field experiment, and in the interpretation of in situ results. In
addition, by examining the extent of biotransformation under a wide range of electron-acceptor and
organic substrate conditions, the laboratory studies provided supplementary data that could not be
obtained at the field scale. A third purpose of the laboratory studies was to characterize the microbial
population in the aquifer material and investigate the extent of acclimation to the gasoline spills over
the course of the study. These disparate lines of field and laboratory evidence were then used to
evaluate the effectiveness of this remedial approach in the Borden aquifer. A detailed description of
the experimental activities, including the chronology, rationale, and specific objectives of each
experiment, is provided below in Section 1.2.
In addition to the research performed at the University of Waterloo, this project involved substantial
collaboration with the U.S. EPA National Risk Management Research Laboratory (NRMRL) at the
Robert S. Kerr Environmental Research Center in Ada, Oklahoma. This included experimental work
with Borden sand performed at NRMRL, analytical work (i.e., metabolites, determination of
hydrocarbon classes) on water and core-extracts obtained from the treatment cells at Borden, and
on-site collaboration, both at NRMRL (first author), and at a U.S.EPA research project at Eglin AFB,
Florida. The results from the nitrate-based bioremediation project at Eglin AFB will be reported
elsewhere.
All research projects funded by the U.S. EPA are required to participate in the Agency Quality
Assurance Program. This research was conducted under an approved Quality Assurance Project Plan
(June, 1993) and Field Sampling Plan (April, 1995). For some of the laboratory and field activities,
modified experimental objectives, unanticipated difficulties encountered during set up, and equipment
availability necessitated deviations from procedures specified in these documents. Actual procedures
used in the laboratory and field experiments performed in this study are described in detail in Chapter
3 and Appendix C, and are consistent with common research practices as reported in the peer-
reviewed literature. Information on the Plans and documentation of the quality assurance activities
and results are available from the Principal Investigators.
The report is organized as follows: The Borden field site is described in Chapter 2. Descriptions
of the field and laboratory methods used in the study are included in Chapter 3. Results and
discussion of laboratory experiments, including both microcosm studies and microbial characterization
work, are provided in Chapter 4. Results and discussion of the field experiment are provided in
Chapter 5. The overall conclusions and implications of the research are discussed in Chapter 6.
1.2 Overview of Experimental Approach
The major laboratory and field activities are summarized in Table 1-1, and described briefly below.
Discussions in Sections 1.2.1 and 1.2.2 describe the experiments in chronological order and also
include the rationale behind each series of experiments.
1.2.1 Laboratory Experiments
Prior to the field experiments, a series of preliminary laboratory experiments were performed using
pristine Borden aquifer material. The first experiment (Experiment 1) was performed to obtain
preliminary data on the effects of dissolved 02, and varying concentrations of N03 , as well as the
addition of inorganic nutrients (NH4-N, P04-P), in this aquifer material. This experiment utilized neat
BTEX (approximately 10 mg/L total BTEX). A second microcosm experiment (Experiment 2) was run
to compare denitrifying activity in Borden aquifer material with two U.S. EPA research sites (Eglin
AFB, Florida and Park City, Kansas) where nitrate-based bioremediation has also been evaluated.
The next microcosm experiment (Experiment 3) was performed to determine whether BTEX
biotransformation could be enhanced under mixed electron-acceptor conditions by comparing the
extent of mass loss under denitrifying, microaerophilic, and mixed microaerophilic / NO^ conditions.
This experiment utilized a tenfold dilution of gasoline-contacted water for a more realistic
representation of anticipated field conditions. A microcosm experiment with 14C ring-labelled aromatics
was also completed to determine the extent of mineralization under anaerobic, denitrifying conditions.
Enumerations of aerobic heterotrophs and denitrifiers in samples from several cores, and assays of
2
-------
Table 1-1. Summary of Major Experimental Activities. Activities are Listed in Chronological Order. Chronology is
Approximate Because Some Preliminary Experiments were Run Concurrently. Report Sections in which
Results are Discussed are also Shown, Along with Microcosm Experiment Number, where Appropriate.
Report
Section
Experiment
Objectives
Description
4.3.1
Preliminary microbial
characterization of pristine
Borden sand
Density of requisite
microorganisms, effect of
gasoline-saturated groundwater
on aerobic microbial activity
Pristine Borden sand; aerobic
heterotrophs, denitrifiers;
aerobic dehydrogenase activity
4.1.1
Preliminary microcosm study
(Experiment 1)
Effect of nutrients, 02, and NO:
concentration on extent of
biotransformation
Pristine Borden sand; neat
BTEX mixture; e acceptors:
N03 only, high O2only, mixtures
4.1.2
Microcosm study, microbial
enumerations
(Experiment 2)
Comparison of BTEX-degrading
activity under denitrifying
conditions in Borden aquifer
material with other petroleum-
hydrocarbon contaminated sites
Aquifer material from Borden,
Eglin AFB, FL, and Park City,
KS; neat BTEX mixture; e
acceptors: NOa"only
4.1.3
Microcosm study
(Experiment 3)
Comparison of BTEX
biotransformation under
microaerophilic, denitrifying, and
mixed e- acceptor conditions
Pristine Borden sand; low
substrate conc. (1/10 gasoline-
saturated); local gasoline
source; e"acceptors: NO," only,
low O, only, N037 low 02
mixture
4.1.4
Microcosm study
(Experiment 4)
Extent of biotransformation
under microaerophilic 02/NCV
conditions (high substrate
concentrations)
Pristine Borden sand; variable
substrate conc. (1/10 to
gasoline-saturated); API 91-01
gasoline; e~ acceptors: N037
low Oz mixture
4.1.5
Microcosm study, microbial
enumerations
(Experiment 5)
Extent of biotransformation
under fully aerobic conditions
(high substrate concentrations)
Pristine Borden sand; high
substrate conc. (gasoline-
saturated); API 91-01 gasoline;
e acceptors: NOs high 02
mixture, high Op only
5.0
Field demonstration
In situ demonstration of gasoline
source area remediation under
mixed e- acceptor conditions
Nitrate-amended and
unremediated control treatment
cells; API 91-01 gasoline
source area; e' acceptors: N03
/ low 02 mixture, low 02 only
4.2.1
Follow up microcosm study
(Experiment 6)
Confirmation of field results;
effects of long-term gasoline
exposure on aromatic-
hydrocarbon degrading activity
Gasoline-contaminated sand
from Nitrate Cell; variable
substrate conc. (1/10 to
gasoline-saturated); e'
acceptors: N037 variable Oa
mixtures, unamended;
acetylene block
4.3.2
Microbial enumerations and
activity assays
Microbial characterization;
demonstration of adaptation
response
Gasoline-contaminated and
pristine locations; aerobic
heterotrophs, denitrifiers,
benzene-toluene degraders;
aerobic dehydrogenase activity
3
-------
microbial activity, were also performed prior to the field experiment to obtain baseline information on
the numbers and activity of the indigenous microbial populations.
After the gasoline was spilled in the field, it became apparent that aromatic-hydrocarbon
concentrations would be higher than a tenfold dilution of gasoline-saturated water throughout the field
treatment cells, and so an additional series of microcosm experiments was performed using gasoline-
saturated groundwater (about 100 mg/L total aromatics) and pristine aquifer material. The first such
experiment (Experiment 4) was performed to compare the response of the aquifer material to high
(gasoline-saturated water) and low (10x dilution of gasoline-saturated water) concentrations under the
mixed electron-acceptor conditions of the field trials (microaerophilic 02 plus N03). Because the
added 02 was rapidly consumed in Experiment 4, a second experiment (Experiment 5) was then
performed to determine whether high aqueous concentrations of aromatics were degradable under
fully aerobic conditions.
After the flushing experiments were completed, cores were collected from the treatment cells
amended with N03 to investigate the extent of biotransformation in aquifer material that had been
exposed to gasoline for nearly two years. A final microcosm experiment (Experiment 6) was
performed to investigate the response in gasoline-contaminated aquifer material under various
electron-acceptor and substrate conditions, and to confirm the results from the treatment cell that was
amended with N03 and 02. Additional microbial enumerations and activity assays were then
performed to evaluate the extent of acclimation to the gasoline contamination.
1.2.2 Field Experiments
Flushing experiments began in May, 1996, approximately six months after the gasoline-
contaminated zones were created, and continued for 174 days. During this period groundwater
amended with electron acceptors was injected continuously into the treatment cells. Target vertical
groundwater velocities and cell residence times were about 25 cm/day and 10 days, respectively. The
Nitrate Cell received mixed electron acceptors (NO„ and dissolved 02), and the Control Cell dissolved
02 only to investigate N03" as an alternate electron acceptor in a low O environment. To perform
mass balance calculations, flow rates, electron acceptor and dissolved BTEXTMB concentrations were
measured periodically. In November, 1996, pumps were shut off and the cells were sampled
periodically under static conditions for organics and electron acceptors. In May, 1997, flushing
experiments were initiated again for 24 days to re-establish conditions similar to the previous year.
During this period, a nutrient solution was pumped into both cells to investigate whether microbial
activity was nutrient limited, and the injection concentration of dissolved 02 was increased. For clarity,
these phases are referred to in this report as "the 174-day flushing experiment", "the static period",
and "the 24-day flushing experiment". Groundwater samples were also collected near the end of the
experiment for analysis of partially-oxidized intermediates (metabolites). In July and August, 1997,
cores were collected for microbial characterization work, laboratory microcosm studies with exposed
aquifer material, and a mass balance on the aromatic compounds to determine the extent of
remediation.
1.3 Background Literature
1.3.1 Biotransformation Under Single Electron-Acceptor Conditions
Aerobic Conditions. Under aerobic conditions, petroleum hydrocarbons, including aromatic
hydrocarbons, are readily degraded by indigenous groundwater microorganisms (e.g., Barker et al.,
1987). Aerobic degradation rates are controlled at the plume scale by the aquifer properties that
control 02 transport to the contaminated area (MacQuarrie et al., 1989), rather than by limitations of
microbial metabolism. As a remediation technology, the major limitation of aerobic biotransformation
is the inability to deliver sufficient 02 to the contaminated area. Oxygen replenishment is limited by its
low solubility (ca. 10.9 mg/L at 10°C) and high reactivity with reduced species such as iron (Fe)
(Morgan and Watkinson, 1992).
At low concentrations, the kinetics of dissolved 02 utilization may dramatically limit the rate of
oxygen uptake and aromatic-hydrocarbon degradation. It has been commonly observed that the rate
of 02 uptake is independent of concentration at high 0, concentrations, but below some critical value,
uptake rates become dependent on concentration (e.g., Johnson, 1967). The concentration of
dissolved 02 has also been observed to limit substrate degradation rates (Larson et al., 1981; Shaler
and Klecka, 1986; Chiang et al., 1987; Leahy and Olsen 1997). The threshold or critical 02
4
-------
concentration is often characterized by the half-saturation constant (KJ. which is defined as the
concentration at which the respiration rate is one half the maximum rate, Shaler and Klecka (1986)
compiled Kd0 values for various oxygenases responsible for either ring fission or initial hydroxylation of
the aromatic ring. They found that these values were high (up to 2.2 mg/L) relative to those for
metabolism of small, easily-degraded compounds such as glucose or acetate, for which 02 is required
mainly as the terminal electron acceptor for the cytochrome oxidase. They postulated that high
concentrations were consistent with the dual role of 02 as both cosubstrate (i.e., addition of oxygen
atoms to the aromatic ring during initial oxidation steps) and electron acceptor in this type of
transformation reaction. On the other hand, microorganisms adapted to a low 02 environment may
synthesize enzyme systems designed to more efficiently utilize 02 (e.g., a monooxygenase system
with lower half-saturation constants) (Leahy and Olsen, 1997). Wilson and Bouwer (1997) found that
critical 02 concentrations, compiled from sixteen studies, ranged from 0.013 to 1.5 mg/L, with the
lower values generally corresponding to utilization of simple substrates such as glucose.
if broadly applicable, these findings have important implications for a remedial approach that relies
upon low levels of dissolved 02 to initiate oxidation of recalcitrant organics. If the half-saturation
concentrations for 02 uptake are high, rates of substrate utilization will begin dropping at relatively high
02 concentrations, and substantial threshold 02 concentrations may persist, or the remaining 0„ may
be utilized by other strains growing on simpler non-target substrates. More work is needed to clarify
this issue, however, because even a small mass-loss enhancement may contribute significantly to the
overall success of a remediation program.
Denitrifying Conditions. Several studies suggest that N03 addition is a potentially viable
bioremediation technology (e.g., Kuhn et al., 1988; Hutchins et al., 1991a, 1991b; Barbara et al.,
1992). With NO, as sole electron acceptor, however, results among various studies have not been
consistent, and several potential limitations have been identified.
Perhaps the major limitation associated with NOg addition is the frequent recalcitrance of benzene
under anaerobic conditions (Berry-Spark et al., 1986; Hutchins, 1991a, 1991b; Barbara et al., 1992).
On the other hand, benzene has been shown to biodegrade under denitrifying conditions in the Major
et al. (1988) study which utilized Borden aquifer sediment. While the possibility of experimental
artifact (e.g., microcosm leakage) cannot be ruled out, benzene loss was not observed in active,
anaerobic controls incubated under identical conditions. In other studies with Borden aquifer material
(Berry-Spark et al. 1986; Barbara et al., 1992), as well as at other sites (Hutchins, 1992), it was noted
that the addition of acetylene gas substantially inhibited aromatic hydrocarbon biotransformation. In
the Major et al. (1988) study, however, BTX-degrading activity was less affected by the addition of
acetylene. Their data showed that the accumulation of nitrous oxide corresponded to the period when
BTX was declining, and that N03 was required for BTX disappearance. These results raise the
possibility that the experimental design of Major et al. (1988) selected for a distinct denitrifying
population with the metabolic capability to biodegrade benzene.
Previous studies have also shown varying levels of removal of the toluene, ethylbenzene, and the
xylene isomers. For instance, using aquifer sediment from Park City, Kansas, Hutchins et al. (1995)
found that toluene, ethylbenzene, m-xylene, and p-xylene were biodegraded under denitrifying
conditions to below 5 pg/L in batch microcosms. In contrast, in the field experiment performed by
Barbara et al. (1992), a toluene threshold concentration of 50 to 100 pg/L persisted throughout the
experiment, and ethylbenzene and xylenes removal was on the order of only 50% of injection
concentrations, despite the continued presence of NO.;. Because the aromatic hydrocarbons are
regulated compounds, threshold concentrations are potentially problematic.
A second limitation involves the incomplete mineralization of monoaromatic hydrocarbons under
limited O or anaerobic conditions. Partially-oxidized compounds may form under both aerobic (Barker
et al., 1987) and anaerobic conditions (Cozzarelli et al., 1995; Barbara et al., 1992; Cozzarelli et al.,
1990), but they appear to be most persistent, and therefore accumulate, in 02-depleted environments.
Metabolite production under denitrifying conditions has been demonstrated in laboratory studies with
pure cultures (Evans et al., 1992; Kuhn et a!., 1988). These compounds are mobile and
geochemicatly reactive. If persistent, their presence may adversely affect bioremediation systems
based on anaerobic, N03_ utilization.
5
-------
Another potential problem is the degradation of non-target organic compounds in preference to
aromatic hydrocarbons. Hutchins et al. (1991a) noted that the extent of BTEX mass loss was lower in
aquifer material contaminated with JP-4 jet fuel relative to uncontaminated material spiked with BTEX.
The N03 demand was much larger in the JP-4 contaminated material, suggesting that non-target
organics were being utilized in preference to BTEX. Similar observations were made by Reinhard et
al. (1995) in the hydrocarbon-contaminated Seal Beach aquifer. Barbara et at. (1992) also arrived at
similar conclusions, although in that study it appeared that NO was being used either as an
assimilatory source of N and/or as an electron acceptor to oxidize naturally-occurring organic matter in
preference to low concentrations of BTEX. Although the biotransformation of other fuel constituents
can be considered a positive result, insufficient microbial growth on the mobile, regulated compounds
may limit this technology.
1.3,2 Biotransformation Under Mixed Electron-Acceptor Conditions
Background. The major anticipated advantage of biotransformation under mixed electron-
acceptor conditions is enhanced mass loss, particularly of compounds such as benzene that are
recalcitrant under denitrifying conditions. Because much less work has been done under these
conditions, however, the advantages and limitations discussed in this section are still quite speculative,
particularly for in situ applications where an indigenous microbial population mediates reactions.
From research on their population ecology and growth strategies, most denitrifiers are facultatively-
anaerobic, heterotrophic bacteria that grow readily under aerobic conditions and prefer to utilize 02 as
electron acceptor. Historically there has been considerable debate regarding the effect of 02 on
denitrifying activity. The conventional view was that denitrifying activity did not begin until 02 was
nearly depleted (Tiedje, 1982; Tiedje, 1988), and many researchers considered denitrification a strictly
anaerobic process. There is also considerable recent evidence that denitrification proceeds in the
presence of substantial amounts of O (Krul, 1976; Robertson and Kuenen, 1984; Lloyd et al., 1987;
Bonin and Gilewicz, 1991; Lloyd, 1993; Patureau et al., 1994; Carter et al., 1995). Although aerobic
denitrification is now a well established phenomenon, the regulating mechanisms and physiological
significance of the process are still not well understood, and the biochemical diversity of denitrifiers
makes generalization difficult. Wilson and Bouwer (1997) provide a comprehensive review of the
aerobic denitrification literature.
Under mixed O,,/ NO conditions, 02 appears to be the most important variable. If the dissolved
02 concentration is initially high, then there will be considerable aerobic biotransformation, but
denitrifying activity will be reduced or completely inhibited. Current research indicates that the critical
dissolved Oz concentration above which denitrification is completely inhibited varies over a broad
range (0.02 to 7.7 mg/L) (Wilson and Bouwer, 1997). The critical concentration is thought to be
species, enzyme, and substrate specific, and probably dependent on growth conditions. Alternatively,
if the dissolved 02 concentration is low, rates of aerobic respiration will be low or negligible, and
denitrifying activity will probably dominate. As discussed in Section 1.1.2, the critical O concentration
for supporting aerobic degradation of aromatic substrates appears to be in the range of 1-2 mg/L, with
considerable variation among species and substrates. It should be noted that additional complexity
may be present in situ] dissolved 02 concentrations measured in bulk pore water may not be
indicative of the 02 levels in microsites or biofilms.
If the 02 concentration is initially high (e.g., 10 mg/L), N03 could be used to enhance aerobic
microbial activity and growth by serving as an assimilatory source of nitrogen (Van 'T Riet et al.,
1968), but, as discussed above, dissimilatory NO., reduction would probably be inhibited. Once
dissolved 0? concentrations fell below the critical value for inhibiting denitrifying activity, N03 could
then be utilized as the electron acceptor. The major advantage of high initial 02 concentrations would
appear to be relatively extensive aerobic biotransformation. A second possible advantage would be a
larger population of denitrifiers resulting from aerobic growth of facultative anaerobes (Su and
Kafkewitz, 1994). It should be noted, however, that the regions of aerobic and denitrifying activity
would probably be separated spatially at the plume scale, diminishing the benefit of this effect, with
most 02 depletion occurring near the point of injection, and N03" depletion over a much longer,
downgradient flow path.
Of greater interest for nitrate-based bioremediation is the behavior of a mixture consisting of N03
and microaerophilic dissolved 02 concentrations. Under microaerophilie conditions, it is less likely that
6
-------
the concentration of dissolved 02 would be inhibitory to denitrifiers, As a consequence, denitrifying
activity could occur in the presence of low levels of 02; the rate, however, may be significantly
reduced. There are two potential advantages to a mixture consisting of N03 and microaerophilic 02.
If the target compound is degradable under denitrifying conditions, NOs can be used by facultative
anaerobes to relieve the electron-acceptor deficit imposed by 02 consumption (Mikesell et al1993;
Leahy and Olsen, 1997); this circumvents the problem of re-oxygenating the subsurface. More
importantly for bioremediation of aromatic hydrocarbons, it is hypothesized (Britton, 1989; Wilson and
Bouwer, 1997) that the 02 could participate in reactions and contribute to the complete oxidation of
otherwise recalcitrant compounds such as benzene.
Based on the current literature, the benefits associated with microaerophilic 02 are unclear. If the
concentration of dissolved 02 is below the critical value to support 02-linked respiration, then there
would be negligible aerobic degradation of recalcitrant compounds like benzene (Section 1.3.1). Even
if all of the 02 was consumed (i.e., the Kdo was low), the stoichiometry of the reaction would appear to
constrain the extent of aerobic oxidation. For example, if the aerobic reaction proceeds to CO,,
complete utilization of 2 mg/L O, mineralizes only 0.65 mg/L benzene to C02. Benzene mass loss
would be reduced further if the 02 consumption was spread among other aromatic hydrocarbons. In
fact, in laboratory experiments with Borden sand, m-xylene is commonly biodegraded first under
aerobic conditions (Section 4). Alternatively, some researchers have hypothesized that, rather than
supporting aerobic respiration, low levels of dissolved 02 could be utilized only as a substrate for
oxygenase enzymes (Britton, 1989), This mechanism has been proposed as a means of initiating
aromatic ring oxidation, yielding greater quantities of partially oxidized intermediates susceptible to
further oxidation by a denitrifying pathway. There is currently no evidence that the so-called "sparing
effect" has enhanced the biotransformation of aromatic compounds in situ.
From an engineering perspective, a mixture of microaerophilic 02 / N03" is the optimal combination.
This circumvents the problems associated with adding high 0„ concentrations and relies mostly on
soluble N03 as the oxidant. At the field scale, however, there are potential limitations to adding
microaerophilic 02 concentrations. First, in most contaminated aquifers multiple substrates are
present, including aromatics, other hydrocarbons, and natural organic material. It is conceivable,
therefore, that the 02 added to an aquifer will be utilized completely by bacteria growing on non-target
compounds. Second, the abiotic 0? demand must be determined. As noted by Kennedy and
Hutchins (1992), if the aquifer is initially anaerobic, reduced metal species may exert a large 02
demand, leading to scavenging of 02 intended for bioremediation, and formation plugging. Finally,
some strains of denitrifiers have been shown to be very sensitive to dissolved 02 (Hernandez and
Rowe, 1987), so it is possible that even a low concentration of 02 in situ may inhibit denitrifying
activity.
Application to Fuel-Contaminated Sites. The published laboratory studies on
biotransformation of aromatic hydrocarbons under mixed electron acceptor conditions have yielded
ambiguous results. Major et af. (1988) found that BTX mass loss in microcosms with Borden aquifer
sediment was slightly enhanced in the presence of both NO..- and 02 relative to losses in aerobic
microcosms. They speculated that the NO„ alleviated a nitrogen limitation during aerobic metabolism,
or that denitrification was occurring within anaerobic microsites. In a pure culture study, Su and
Kafkewitz (1994) found that Pseudomonas maitophilia was capable of degrading toluene and xylene
isomers in the presence of N03~ and 2 percent 02. Miller and Hutchins (1995) used laboratory
columns to study BTEX removal from three different aquifer sands under NO, only and then N03 / 02
conditions. In two of the columns, adding low levels of 02 did not enhance BTEX removal, and in the
third, O, had an inhibitory effect. Under all conditions, benzene was completely recalcitrant. Hutchins
et al. (1992) also used laboratory columns to study the effects of various combinations of N0g70 on
the removal of BTEX from aquifer material. They found that adding N03 to the column with low u2
decreased TEX breakthrough by an order of magnitude, demonstrating that N03 was needed to
increase substrate utilization. There were no adverse effects associated with low levels of 02.
Benzene removal was low and independent of electron acceptor conditions. In contrast, Wilson et al.
(1995) observed benzene degradation in the presence of 2 mg/L dissolved 02 and NO.;, but only
toluene degradation under anaerobic, denitrifying conditions.
Anid et al. (1993) investigated BTEX removal in aerobic columns amended with either hydrogen
peroxide or NO.;. In columns amended with NOs, effluent dissolved 02 concentrations dropped from
7
-------
9 mg/L to 2 mg/L and 164 mg/L N03 was consumed. Benzene was recalcitrant in these columns, but
in additional experiments with N03 and <1 mg/L Oz» approximately 25 percent of added benzene
appeared to biodegrade. In this aquifer material, strict anaerobic conditions did not appear to be
required for N03 reduction, although the formation of anaerobic microsites cannot be ruled out. In
addition, benzene removal was enhanced under anaerobic rather than aerobic conditions. Leahy and
Oisen (1997) showed that N03 enhanced the rate of toluene utilization by denitrifying strains after the
dissolved 02 fell below a critical concentration. Denitrifying strains were able to maintain a higher rate
of toluene utilization by switching to denitrifying activity when the availability of dissolved O was low.
Similarly, Mikesell et al. (1993) and Hutchins (1991a) both demonstrated in the laboratory that, under
limited 02 conditions, biotransformation of certain aromatics could be enhanced by the presence of
N03. In the Hutchins (1991a) study, benzene losses were observed in microcosms amended with
both 02 / NOg" after the apparent removal of 0?.
Since the late 1980s, only a handful of nitrate-based bioremediation field studies have been
completed (Sheehan et al., 1988; Hutchins et al., 1991b; Battermann and Meier-Lohr, 1995; Hutchins
et al., 1995; Reinhard et al., 1995). Hutchins et al. (1991b) added NO, and 02 via an infiltration
gallery to an aquifer in Traverse City, Michigan contaminated with JP-4 jet fuel. Both electron
acceptors were consumed, and more N03 was consumed than required for BTX degradation,
indicating that other compounds were being utilized under denitrifying conditions. Toluene and
xylenes degradation appeared to be stimulated in the aquifer, but benzene removal was apparently
due only to flushing. Hutchins et al. (1995) describe the performance of a nitrate-based
bioremediation system applied to a petroleum spill in Park City, Kansas. In this study, NO./ utilization
was confirmed but site heterogeneities obscured the performance evaluation. Although laboratory
data indicated that the aromatic hydrocarbons would biodegrade under denitrifying conditions, this
could not be confirmed in the field. This study showed the problems that can arise in evaluating the
extent of bioremediation in heterogeneous aquifers.
Battermann and Meier-Lohr (1995) describe the performance of a large-scale NO,, plus 02
bioremediation system at an abandoned refinery site contaminated with residual hydrocarbons
(dissolved BTEX concentrations 10 to 100 mg/L). After three years of operation, about 300 metric
tons of hydrocarbons were removed, of which 80 percent were attributed to biotransformation and
20 percent to flushing. During pilot-scale tests at the same site, Battermann et al. (1994) found that
hydrogen peroxide was attenuated near the infiltration point. Nitrate, on the other hand, was
distributed over much larger contaminated areas; in the first year of operation, approximately
100 mg/L N03 was consumed over a 50-day residence time.
These field studies have suggested that ambient dissolved 02 in the injected water does not inhibit
denitrifying activity, but there were no observable benefits associated with 02 either. The extent of
aerobic biotransformation near source areas or the effects of lower concentrations of O further
downgradient are not easily assessed in uncontrolled field situations. Moreover, it is difficult to
determine the contribution of flushing as a mass removal mechanism. The highly-controlled field
experiments used in this study provided additional information on the in situ utilization of 02 in this type
of bioremediation system.
8
-------
CHAPTER 2. STUDY AREA
The field research was conducted at CFB Borden, Ontario (Figure 2-1). The research area is
located in an abandoned sand pit. The depth to groundwater varies from ground surface to about
1.5 meters below ground surface depending on location within the pit and the season. The site has
been studied extensively over the past 17 years. The geology and local hydrogeology were
investigated by MacFarlane et al. (1983). The hydrogeology and groundwater chemistry of the sand
pit were discussed in detail by Mackay et al. (1986).
The sand pit is underlain by a relatively homogeneous, unconfined sand aquifer composed of
interbedded fine- to medium-grained glaciolacustrine sand. Analysis of a bulk sample indicates that
the aquifer material is composed of 58% quartz, 19% feldspars, 14% carbonates, 7% amphiboles, and
CFB BORDEN
DIEPPE
RDAD /
STREAM
LANDFILL
100 r.
Figure 2-1 Location map of the study area at CFB Borden (from Oliveira (1997)).
9
-------
2% chlorite (Mackay et al., 1986). The aquifer is about nine meters thick, extending from ground
surface to a clay layer. Detailed coring has shown that the deposit contains discontinuous lenses of
sand that vary from 0.02 to 0.1 m in thickness and 2-5 m in length (Sudicky, 1986). The mean
hydraulic conductivity is 7 x 10"3 cm/sec, with variations between layers occasionally exceeding two
orders of magnitude (Sudicky, 1986). The mean weight-fraction of organic carbon (foc) was estimated
at 0.00018, and the porosity at 0.33 (Mackay et al. 1986). The spatial variability of porosity based on
36 samples was found to be small (coefficient of variation = 0.05).
A dense, woody peat layer was encountered at the experimental site at a depth of about 2.7 m
below ground surface. Based on cores and other drilling information, this layer was about 10 cm thick,
fairly continuous, and overlain by about 25 cm of dark gray, dense silty sand. Elevated dissolved
methane (CH.) and sulfide (HS ) were frequently detected in groundwater samples collected near the
peat layer, and cores of the peat sequence had a sulfide odor and were dark gray with orange
oxidation zones on the upper and lower boundaries. To avoid pumping water across this sequence,
all experiments were conducted within the top 2.5 m of the aquifer. As a low-hydraulic-conductivity
layer, this sequence appeared to limit the flux of water from the underlying aquifer when the flushing
experiments were in progress.
Groundwater flows in a northeasterly direction at about 9 cm/day, and remains near 10°C
throughout the year. A leachate plume originating from an abandoned municipal landfill is present at
the base of the unconfined aquifer, at depths ranging from five to seven meters below ground surface.
Groundwater above the leachate plume is unaffected by the landfill.
Based on location's of previous experimental activities, the site has never been exposed to
aromatic hydrocarbons. Although the presence and depth of the landfill leachate plume beneath the
site was not determined, groundwater chemistry data collected from a shallow well (2.49 m depth)
indicated that the landfill leachate, if present, was below the experimental zone. Dissolved 02 levels in
shallow groundwater were unexpectedly low at about 0.2 mg/L; in other regions of the sand pit,
shallow groundwater has substantially higher concentrations of dissolved oxygen (e.g., Barbara et al.,
1994). There was no detectable Fe or N03 in background groundwater, and dissolved organic carbon
was low (<2 mg/L). Dissolved Mn was detected, however, and may have exerted an abiotic 02
demand. The chemistry of the groundwater is summarized in Table 2-1.
Table 2-1. Chemistry of Unamended Groundwater at the Experimental Site. Samples Collected May, 1996 from
Shallow, Upgradient Supply Well. Concentrations in mg/L.
Parameter
Concentration
Si
2.88
CI
3.58
so4
6.31
N03
<0.1
per
<0.2
Br 4
<0.1
HCO,
258
Ca 3
80
Mg
3.55
Na
4.21
K
1.16
NH.
0.82
Fe4
<0.1
Mn
1.24
pH
7.48
DO
0.2
DOC
<2
BTEXTMB
n.d.1
Temp (°C)
12.5
1 n.d. - not detected at individual compound method detection limit {Appendix C).
10
-------
CHAPTER 3. EXPERIMENTAL METHODS
The procedures used in the laboratory arid field experiments are described in this chapter.
Descriptions of the procedures used for laboratory microcosm experiments, microbial characterization of
aquifer materials, the controlled gasoline spills in the field, and the design and operation of the field
experiments are included here. A description of the physical and chemical characteristics of the API
91-01 gasoline is included in Appendix A. A description of procedures and results of the field tracer test
designed to evaluate the flow of injected water in the treatment cells is provided in Appendix B. Appendix
C includes descriptions of laboratory sampling and analytical procedures, as well as analytical procedures
used for field samples. Appendix D includes results of the groundwater sampling performed downgradient
of the wastewater treatment mound and treatment cells to monitor potential releases to the aquifer.
3.1 Aquifer Core Collection
Aquifer sediment was collected at various times during the study for different purposes. Core material
was used for logging the stratigraphy, microcosm and microbial characterization studies, and mass
balances on the gasoline injected into the field treatment cells. Cores were obtained in 5.08 cm (2 in)
diameter aluminum tubes using either a piston or core catcher. The coring system is described in detail in
Zapico et al. (1987).
Aluminum core tubes designated for microbiology experiments were rinsed with methanol and flamed
in the field prior to use. After the cores were collected they were sealed tightly, transported within 4 hr to
the University of Waterloo, and stored at 4°C. In preparation for laboratory experiments, aquifer material
was removed from core tubes either in an anaerobic chamber or a sterile, laminar flow cabinet (Nuaire,
Inc., Model NU-408FM-300), depending on in situ 0„ concentrations, and refrigerated in sterile mason
jars. Material adjacent to core-tube walls was discarded or used in sterilized control microcosms.
3.2 Laboratory Microcosm Experiments
3.2.1 General Set Up Procedures
All microcosm experiments in this study had one of two basic designs. The most common approach,
designated Design 1, consisted of a series of crimp-sealed microcosms (60 ml glass hypovials). For each
sampling event, a set of microcosms was sacrificed for analysis. A set contained replicate microcosms for
each treatment group in the experiment. Generally there were three to five different treatment groups,
including a sterile control, per experiment. The second approach, designated Design 2, utilized a smaller
number (typically 1-3 bottles per treatment) of 100 ml screw-top bottles designed for repetitive sampling
through a threaded mininert™ valve (Dynatech Precision Sampling Corp., Baton Rouge, LA). For each
sampling event, a sample of headspace gas was obtained from these microcosms. This design was used
for the first series of preliminary experiments only (Section 4.1.1). A summary of the microcosm designs
is provided in Table 3-1. Sterile equipment and aseptic technique were used in all aspects of microcosm
preparation and sampling.
Each microcosm received 20-25 g of homogenized aquifer material, groundwater spiked with
organics, nutrients, and electron acceptors, if required, and was sealed with a Teflon™ lined septum and
an aluminum crimp seal, or mininert™ valve. All Design 2 microcosms contained a headspace, from
which gas-phase samples were collected. Depending on the intended 02 status of Design 1 microcosms,
a 2-4 ml headspace was present. Preliminary work with this type of set-up indicated that a pure 02 or air
filled headspace was the most reliable means of supplying dissolved 02. If 02 was not required, Design 1
microcosms were prepared with no headspace.
11
-------
Table 3-1, Summary of Microcosm Designs Used in the Study
Type
Description
Design 1
60-ml glass hypovials; crimp-sealed; 20-25 g aquifer material; 2-4 ml pure
oxygen or air headspace present in some microcosms; analytes measured
in fluid phase (organics, e~ acceptors); set of replicates sacrificed for each
sampling event
Design 2
100-mI screw-top bottles; threaded mininert™ valves; 20-25 g aquifer
material; anaerobic or air headspace present; analytes measured in
headspace gas (organics, O?), and fluid phase (e acceptors); single or
replicate bottles sampled repetitively over time
The source of the organic substrates was either gasoline or pure solutions of benzene, toluene,
ethylbenzene and xylene isomers (6:4:2:2 mg/L concentration ratio). To prepare gasoline-saturated water,
an appropriate volume of aerobic Borden groundwater was first gassed for at least 3 hr with either sterile
nitrogen gas or sterile air, depending on the treatment. An aliquot of this water was transferred to a sterile
separatory funnel to prepare gasoline-saturated groundwater. Gasoline was added to yield a 10:1 water
to gasoline volume ratio. The separatory funnel was then shaken manually three times (5 min each).
After 24 hr, gasoline-saturated water was removed from the funnel to glass bottles for dispensing to
microcosms. The dissolved 02 concentration of the prepared water was determined prior to dispensing to
microcosms. If lower concentrations were required for an experiment, a 10x dilution of the gasoline-
saturated water was performed using an aliquot of the gassed, hydrocarbon-free groundwater. Measured
concentrations in gasoline-saturated water ranged from 70-130 mg/L, depending on the experiment.
Concentrations differed because different brands of gasoline were used.
The set-up procedures were similar for all experiments. Microcosms were prepared either in the
anaerobic chamber or in the sterile flow cabinet. For anaerobic or low O conditions, aquifer material was
typically dispensed within the anaerobic chamber to avoid any contact with air. For aerobic conditions,
aquifer material was dispensed into microcosms in the flow cabinet. Aquifer material was dispensed into
sterile control microcosms several days in advance to allow for sterilization (1/2 or 1 hr autoclave run on
three consecutive days). After all the aquifer material was dispensed, contaminated groundwater and
amendments were added. Microcosms were then filled completely with appropriate solutions and crimp
sealed. If required, microcosms received inorganic nutrients (NH,CI and KH2P04) from concentrated,
sterile, anaerobic stock solutions to yield 5 mg/L as N, and 2 mg/L as P. Nitrate, where required, was
added as a concentrated stock solution of KN03 to yield concentrations of 5-50 mg/L as N. Sterile
controls received an additional 0.5-0.6 ml of 10% (w/v) sodium azide solution or 0.25 ml of a 4% HgCI2
solution. In some cases, the nutrients and N03* were spiked into the groundwater before dispensing to
microcosms.
Aerobic Design 2 microcosms were prepared with aerobic groundwater in equilibrium with
atmospheric 0?. Design 1 microcosms requiring dissolved 02 were moved to the sterile flow cabinet after
construction, decrimped, and the appropriate volume of water removed with a sterile syringe. "Low"
dissolved 02 microcosms, also referred to as microaerophilic microcosms, typically received a 2-4 ml
ambient air headspace, and "high" dissolved oxygen microcosms, also referred to as aerobic microcosms,
contained a 4-ml pure 02 headspace. Pure 02 was added by purging the headspace at a rate of
approximately 300 ml/min for 30 seconds and quickly replacing the crimp seal. It should be noted that
partitioning of the aromatic hydrocarbons to a headspace of this size is minimal. For example, partitioning
calculations based on Henry's Law indicate that only 2% of the total mass of benzene would partition to a
4-ml headspace.
We attempted to establish "microaerophilic" conditions in two ways: First, by mixing aerated and 02-
stripped (by N2-purging) waters in a 50:50 ratio, and second, by introducing a small air-filled headspace
into otherwise anaerobically prepared microcosms, similar to the method of Hutchins (1991a). The first
method should have provided an initial D.O. of 3.5-4 mg/L, based on numerous D.O. measurements of
aerated water. However, the method essentially failed, as the available 02 may have been consumed by
the initial microcosm sampling on day 1. Alternatively, because day 1 BTEX levels in these and the
12
-------
strictly-denitrifying microcosms were similar, the D.O. may have actually been consumed or lost during
microcosm preparation in the anaerobic chamber. The microcosms prepared in this manner behaved
identically to those set up under anaerobic conditions, and are not discussed further. The second method
was used for all experiments requiring microaerophilic conditions. It provided a total mass of about
1.0 mg 02 (assuming 20% by volume in air) in a microcosm. Measured initial aqueous concentrations in
sterile controls were ca. 4.5 mg/L (Section 4.2.1). It is acknowledged that the initial D.O. concentration in
microcosms prepared in this manner exceeded microaerophilic levels as defined previously in this study
(2 mg/L or less). However, to obtain observable results it was necessary to provide a mass of 02
sufficient to drive aerobic reactions in this static system. A strict adherence to the operational definition of
microaerophilic conditions (2 mg/L initial concentration with no headspace), while ideal, was not tractable
with our experimental methods. Considering the limitations associated with establishing microaerophilic
conditions, we concluded that the air headspace provided the best analogue to the dynamic system in the
field, where a continuous injection concentration of ca. 2 mg/L 02 could be maintained.
Sterile aqueous controls (no aquifer material) were also prepared (Experiment 6) to investigate the
rate of diffusive loss of dissolved 02 and BTEXTMB. Design 1 microcosms were prepared aseptically with
sterile groundwater and amended with 90 mg/L neat BTEXTMB, 0.6 ml sodium azide solution, and
dissolved 02. Two sets of microcosms were prepared: microaerophilic and fully aerobic. Microaerophilic
microcosms contained a 2-ml air headspace, and were incubated in the anaerobic chamber. Aerobic
microcosms contained a 4-ml pure-O, headspace and were incubated in the laboratory cupboard. These
microcosms were stored with other microcosms prepared for Experiment 6 and periodically sampled over
a 159-day period to observe the loss of dissolved 02 and organics from sterile microcosms.
After set-up microcosms were stored in the dark at room temperature (23±2°C) either in an anaerobic
chamber (Lab-Line Instruments, Inc., Model 6550) supplied with a 1% COa, 2.5% H20, 96.5% N2 mixed
gas, or a laboratory cupboard (aerobic microcosms). Room temperature incubation was unavoidable
because the temperature within the anaerobic chamber could not be controlled. Experimental
temperatures were therefore roughly twice in situ temperatures. Our previous experience with Borden
aquifer material suggests that temperature does not affect experimental results with respect to compound
degradability; compounds that degrade at room temperatures also degrade at groundwater temperatures.
Reaction rates are, however, faster at room temperature.
3.2.2 Acetylene Block
The cause of the rapid NO. utilization observed in the final microcosm experiment (Experiment 6;
Section 4.2) was investigated by assaying for denitrifying activity. Because acetylene inhibits the
reduction of NaO to N2, the accumulation of nitrous oxide in the presence of acetylene is considered strong
evidence of denitrification (Tiedje, 1982). Acetylene gas was added to duplicate microcosms from each
treatment group that contained N03" by injecting 1% (v/v) through the septa, shaking the vial for 1 min,
storing inverted for 1/2 hr to equilibrate the aqueous phase with the headspace, and then quickly replacing
the pierced septa. Prior to injection, the acetylene was passed through a series of flasks containing
distilled water to remove any acetone that may have been present. These microcosms were incubated for
15 days and then analyzed for N20 and acetylene, as well as dissolved O . N03. nitrite (NO,), and
BTEXTMB.
3.3 14C Mineralization Experiment
The conversion of benzene, toluene, m-xylene, and o-xylene to CO (mineralization) under anaerobic,
denitrifying conditions was investigated with standard 14G02 trapping techniques. The radiolabeled
compounds (Sigma Chemical Co., St. Louis, MO) were obtained from pure solutions: [U-14C]benzene,
[ring-U-14C]toluene, m-[ring-U-14C]xylene, and o-[ring-U-uC]xylene. Details of microcosm preparation,
sampling, and analyses used in this experiment have been described previously (Hutchins, 1993). In
brief, 10 g of pristine Borden sand with no known prior exposure to hydrocarbons was dispensed to 60-ml
serum bottles designed for repetitive sampling. Each microcosm received a single radiolabeled aromatic
compound (3-9 mg/L aqueous concentrations), 20 mg/L NOa"-nitrogen if required, N and P as nutrients,
and anaerobic water (distilled water mixed with groundwater from a spring near Ada, OK). Sterile controls
received both mercuric chloride (250 mg/L) and sodium azide (500 mg/L). After spiking with the
appropriate amendments, each microcosm was sealed without headspace using a Teflon™-lined butyl
13
-------
rubber septum, mixed, and incubated at room temperature in an anaerobic glove box. Three replicate
microcosms with corresponding sterile and unamended (no NO.;) controls were prepared for each of the
four radiolabeled compounds.
Microcosms were sampled five times over a 63 day incubation period. All sampling was done in the
anaerobic glove box. The headspace created by removing water for analysis was eliminated by adding
sterile 6-mrh glass beads. Microcosms were then resealed, mixed, and again incubated in the glove box.
Samples were analyzed for aromatic hydrocarbons, N03, and NO,. as described in Hutchins (1993).
The distribution of the radiolabel was determined with a modification of the method used by Grbic-Galic
and Vogel (1987). This method involves measuring the sum of UC02 and nonvolatile intermediates,
nonvolatile intermediates only, and then calculating 14COz by difference. Total 14C activity in the aqueous
phase was also estimated. This approach accounts for radiolabel distribution, and therefore extent of
biotransformation, in the aqueous phase only; activity of the solids was not determined.
3.4 Microbial Characterization
3.4.1 Enumerations
Aerobic Heterotrophs. Enumerations of viable, aerobic heterotrophic and denitrifying
microorganisms were conducted by the standard spread plate and most-probable-number (MPN)
methods, respectively, using 0.1% Na pyrophosphate (pH=7.0) to suspend the aquifer material, and
phosphate-buffered saline (1.18 g Na2HP04, 0,22 g NaH2P04@H20, 8.5 g/L NaCI, pH 7) to dilute the
suspension as required. R2A medium was used for the aerobic heterotrophic plate counts (HPCs)
(Reasoner and Geldreich, 1985). R2A medium is a relatively low-nutrient medium that was developed for
the enumeration of microorganisms in potabie water. All plates were prepared in triplicate for each
dilution, and incubated at room temperature for up to 30 days.
Denitrifiers. For the denitrifier MPN procedure, 18-ml vials were filled with 12 ml 1/10-strength
nutrient broth (Difco Laboratories, Detroit, Ml), and amended with 2 mM KN03 and 0.17% Noble agar
(Difco). Immediately prior to inoculation, the medium was deaerated by placing the vials in flowing steam
for 5-10 min, then quickly cooled to room temperature and inoculated. Inoculated vials were sealed with
sterile, slotted butyl rubber stoppers (Wheaton), and the headspace of each vial was flushed with sterile
nitrogen for 30 sec by loosening the stopper slightly and inserting a sterile syringe needle into the stopper
slot. Vials were then sealed permanently with an aluminum crimp lid. Finally, 0.6 ml acetylene gas was
injected into the headspace using a sterile 1-ml syringe fitted with a membrane filter (0.2 pm pore size).
All vials were prepared in triplicate and incubated at room temperature for up to 61 days. After incubation
0.1-0.2 ml of culture fluid was removed from each vial and tested with diphenylamine reagent for NO
and/or N02 (Tiedje, 1982). Denitrification was confirmed in approximately 10% of vials with depleted! N03
by analyzing for the presence of accumulated N20. A 2-ml sample of microcosm headspace gas was
analyzed for N„0 as described in Appendix C.
Aerobic Benzene-Toluene Degraders. A suspension of aquifer material was prepared by
aseptically adding 10 g (wet wt) of aquifer material to 90 ml 0.1% Na pyrophosphate solution (pH 7). The
suspension was shaken for 10 min at 400 rpm on a rotary shaker, then diluted further in phosphate-
buffered saline. One ml atiquots of selected dilutions were added to triplicate tubes of a mineral medium
(Furukawa et al., 1983) in screw-capped test tubes (10 ml medium/tube). Each tube was then amended
with 1(jL of a neat, filter-sterilized benzene/toluene mixture (1:1 concentration ratio) using a micropipettor.
The tube was closed and shaken to dissolve the hydrocarbons. Tube caps were covered with a layer of
Parafilm and plastic wrap to minimize losses of volatiles during incubation. Resulting maximum aqueous
concentrations were about 50 mg/L for both benzene and toluene, although actual concentrations were
probably lower due to partitioning into the tube headspace, and to losses to the atmosphere during the
amendment procedure. Tubes were incubated at room temperature for 77 days, and during the
incubation, they were shaken periodically to keep the medium oxygenated. At the end of the incubation,
tubes were scored for growth (culture turbidity) by visual inspection, and an MPN of benzene-toluene
degraders was determined from the appropriate 3-tube MPN table (Mayou, 1976).
3.4.2 Microbial Dehydrogenase Activity.
The electron transport system (ETS) test developed by Trevors et al. (1982) allows for a comparison
of microbial activity in soil or sediments under varied incubation conditions. This test measures aerobic,
14
-------
microbial dehydrogenase activity in the aquifer material. Experiments were run to compare activities
between different sample locations under similar substrate conditions (e.g., gasoline-saturated water),
and between different substrate conditions for the same sample location. The test is based on the
reduction of water-soluble 2-(p-iodophenyl)-3-(p-nitrophenyl)-5-phenyl chloride (INT) to methanol-
extractable red iodonitrotetrazolium formazan (INT-formazan) by the microbial dehydrogenase activity of
the aquifer sediment.
Ten grams (wet weight) of aquifer material were incubated in crimp-sealed 60 ml hypovials. Active,
blank and sterile control treatments were included in each experiment. Active treatments received 1 ml of
0.4% (w/v) sterile INT solution, one or more carbon sources (gasoline-saturated water, 1/10 gasoline-
saturated water, sterile 0.2% (w/v) glucose solution, or sterile R2A broth), nutrients (5 mg/L as N and
2 mg/L as P), and sterile distilled water, as required, so that each flask or vial received a total of 2.5 ml or
3.0 ml liquid. Sterile controls contained aquifer material that had been autoclaved for 1/2-1 hr on three
consecutive days, and received 0.5 ml of 10% (w/v) sodium azide solution in place of sterile water. Blank
treatments received water in place of INT solution. Each treatment was prepared in duplicate.
Experiments were performed under aerobic conditions.
Flasks were sampled several times over ca. a 30 day incubation period by removing approximately
1 g of sand slurry, extracting the slurry with 5 ml methanol, and measuring the INT-formazan content
spectrophotometrically (Trevors et aL 1982). An INT-formazan standard curve was generated using
standards consisting of reagent INT-formazan in methanol. Blank-corrected INT-formazan content was
reported on a dry weight basis.
3.5 Borden Field Experiment
3.5.1 Instrumentation
The field trials were performed within 2m by 2m by 3.5m deep Waterloo Barrier™ sealable, sheet-
piling cells. Cells were installed by vibrating the individual sections of sheet piling to the target depth. The
joints were then sealed with bentonite grout to isolate the interior of the cell from the surrounding aquifer.
The base of the cell was open to the underlying aquifer.
Two treatment cells were used in these experiments: a Nitrate Cell (N0a702 amended) and a Control
Cell (02 amended). Each cell contained instrumentation for groundwater sampling, water level
measurements, gasoline injection, dewatering, geophysical measurements, and the addition and
extraction of water. Cell instrumentation is shown on Figure 3-1. Most instrumentation was installed by
advancing a steel casing to depth, washing the cuttings from the inside of the casing with flowing water,
lowering the instrument, and then removing the steel casing. The formation was then allowed to collapse
into the annular space. Drive points were vibrated into the ground using a pneumatic hammer.
The five multilevel piezometers in each cell were constructed of a series of 3.2 mm (1/8 in) stainless-
steel tubes soldered to a 2.54 cm (1 in) stainless-steel center tube. Each piezometer had nine sampling
ports, spaced 30 cm apart, between 30 and 270 cm bgs (below ground surface of cell). To avoid creating
a preferential vertical flow path along the piezometer wall, the 3.2 mm tubes were placed inside the
2.54 cm center tube so that sample ports were flush with the center tube outside wall. Each cell
contained five 2.86 cm (1-1/8 in) diameter stainless-steel drive point piezometers (depth 2.4 m; screen
length 18 cm), used primarily for water level measurements. Each cell also contained three access tubes
for geophysical measurements. The access tubes were constructed of 5.08 cm (2 in) diameter Schedule
40 PVC pipes (2.4 m depth) with solid 3.2 mm (1/8 in) stainless-steel rods embedded into the outside of
the pipe. The rods were used to measure water content prior to the gasoline spill with time domain
reflectometry (TDR).
To flush water vertically through the cells, both injection and extraction wells were also installed.
Extraction wells were constructed of 2.86 cm (1-1/8 in) diameter stainless steel tubing flush-soldered to a
50-cm long drive-point screen; tops of screens were located 200 cm bgs. One centrally-located well was
used in each cell for extraction of water (Figure 3-1). Water was injected into a high hydraulic conductivity
layer located near the top of each cell. This layer was used as a means of distributing injected water
horizontally without the use of multiple injection wells. It was installed by excavating the cells to ca.
60 cm, levelling the surface, and then backfilling with 8 cm of coarse sand, followed by 8 cm of 0.95 cm
15
-------
Treatment Cells
C3U3
" P24D
P23C
CJU8
2 m
P23E
[7>v C3U2
P23A
PZ48
P23E
Nitrate Cell Control Cell
fir :> Injection Wells
Extraction Well
B —; Multilevel Piezometers
£? > Drive-point wells
—> Gasoline Injection Well
<> —^ Multilevel TDR Probes
FIGURE 3-1 Plan view of treatment cells with surveyed locations of instrumentation.
(3/8 in) diameter pea gravel. A commercial filter fabric was then placed over the pea gravel and the cells
backfilled with 5 cm of coarse sand followed by native Borden sand to surface. Prior to backfilling, a
5.08 cm (2 in) PVC injection well with a 30.5 cm (12 in) well screen was installed in each cell to a 65-cm
depth so that the screened interval straddled the pea gravel layer. Both of these wells gradually lost
transmitting capacity over the first 3.5 months of the experiment, and were replaced by stainless steel
drive point wells with No. 10 slot well screens which operated for the remainder of the experiment (Figure
3-1).
A centrally-located drive-point gasoline injection well (depth 150 cm bgs; screen length 18 cm) and a
1.9 cm (3/4 in) diameter PVC dewatering well (depth 250 cm bgs; screen length 86 cm) were also
installed in each cell (for clarity not shown on figure).
After the equipment was installed, plastic tarps were placed on cell surfaces as vapor barriers. Gaps
around well casings were sealed with roofing tar or silicone. These sealants were not in contact with
injected water or gasoline. The outer edges of the cells were sealed by packing a wedge of thick
bentonite grout along the crenulated wall of the sheet piling. The grout was then covered with sand to
slow desiccation. A monitoring and pumping-well network was also installed downgradient of the cells to
detect releases and contro! the plume, if necessary (Figure 3-2). Periodic sampling of downgradient
multilevel piezometers indicated that gasoline releases did not occur (Appendix D), To prevent infiltration
of rainwater, a greenhouse was constructed over the cells.
3.5.2 Gasoline Injection
The objective of the spills was to emplace a source of gasoline below the water table as a spatially-
uniform, residual phase so abiotic losses (volatilization, physical removal during high water table events)
and preferential flow of water around gasoline-contaminated zones would be minimized. To emplace the
16
-------
Site Plan
Regional
Groundwater
Flow Direction
Biopfle Well
0
Wastewater
Treatment Mound
Greenhouse
A ®
Supply
Treatment
Cells
4 m
® Supply Wei:
® Monitoring Well
b Bundle Piezometer
a pump Well
FIGURE 3-2 Plan view of study area.
sources, the water table was lowered, the gasoline gravity-fed into the gasoline injection wells, and the
water table allowed to recover to the pre-spill elevation.
Seventy liters of API 91-01 gasoline (Appendix A) were injected into each cell during October, 1995.
Prior to injection, the ambient water table was approximately 70 cm bgs. The water table was lowered
using pumps connected to the four corner drive point wells, and the PVC dewatering well. The Control
Cell was dewatered for 29 hrs and the Nitrate Cell for 47 hours prior to injection (injections occurred on
successive days). In both instances target water table depths (Nitrate Cell: 175 cm bgs; Control Cell:
180 cm bgs) were reached and maintained for several hours prior to gasoline injection to give the
moisture content profile time to respond to the falling water table and approach a condition of static
equilibrium. The target water table depth was chosen so that the top of the capillary fringe would be
roughly coincident with the base of the gasoline injection well. The thickness of the capillary fringe is
about 30 cm in the Borden aquifer (Nwankwor et al., 1992). Water content profiles collected with the
multilevel TDR probes 1-3 hrs before injection show that the top of the capillary fringe was slightly higher
than anticipated in both cells (Figure 3-3). A description of the methods used to collect and process the
TDR data is provided by Oliveira (1997).
Gasoline was gravity-fed into the gasoline injection wells from a sealed, polyethylene tank. The
durations of the injections were 5 hr 45 min for the Nitrate Cell, and 7 hr 45 min for the Control Cell. A
partially-clogged flow meter was responsible for the slower rate of injection into the Control Cell.
Groundwater extraction from the center drive-point well (screen depth 222-240 cm bgs) continued
throughout the injection. However, these wells were incapable of sustaining the required extraction rate,
and the water table rose ca. 10 cm in both cells during the injections. Extraction wells were turned off
immediately after the injection of gasoline was completed. After about two weeks, the water table had
17
-------
Q.
Screen
I fc/ 9
I Expected Top of Capillary Fringe
_C
Q.
0
Q
Water Saturation, Sw(%)
0 20 40 60 80 100
0
0.4
0.8
1.2
1.6
Nitrate Cell
-o— C4U8
O- C4U5
-A— C4U1
Gasoline sj
Injection Well
'-Screen „
Expected Top of Capillarv Fringe
FIGURE 3-3 Vertical profiles of water saturation prior to the gasoline injections. Profiles were obtained 1 -3 hours
before injection with multilevel TDR probes.
recovered to pre-injection levels. Because the injection procedure involved dewatering, a trapped air
phase was probably also present below the water table. The mass of 02 in the trapped air phase was
estimated and incorporated into mass balance calculations {Section 5.2). The extent of volatilization that
occurred during gasoline injection could not be quantified, but based on relatively short contact times with
the atmosphere during injection and analysis of groundwater from gasoline-contaminated zones one
month after injection (see below), volatilization losses of target compounds appeared minor.
Detailed measurement of the in situ magnitude and structure of the trapped gasoline and air phases
was considered beyond the scope of the research. A detailed characterization of their distribution was
therefore not attempted, but a general understanding of the distribution of the injected gasoline was
developed from field observations, groundwater samples, and core extract data. The distribution of
aqueous-phase BTEXTMB concentrations during November, 1995, one month after injection, is shown in
Figures 3-4 and 3-5. These figures suggest that the gasoline phase did spread radially outward from
injection wells as intended, and was present mainly in the 120 cm to 60 cm bgs depth interval. The
gasoline contaminated zones were thickest in the vicinity of the injection wells, extending to a depth of
approximately 150 cm bgs. These initial observations were confirmed by the results of the post-
experiment coring (Chapter 5). However, complete trapping below the water table apparently did not
occur; after recovery to the pre-injection head, a thin layer of gasoline was observed on the water table in
the Control Cell. This indicated that the 110 cm recovery following injection did not trap all of the gasoline
below the water table. During the November sampling round, it was also observed that a mobile gasoline
phase (i.e., above residual saturation) was present near some of the sampling ports; some samples
contained a gasoline/water mixture, as indicated by field observations and concentrations well above
those in gasoline-saturated water (Figures 3-4 and 3-5). Overall, however, the injections were successful
in emplacing a source below the water table, and the gasoline distributions were suitable for meeting the
goals of the experiment.
During the following Spring snowmelt, groundwater with a sheen of gasoline was observed above the
surfaces of the cells. The amount of gasoline removed from the cells during this event could not be
determined, but visual observations and samples of the standing water suggested that losses were minor.
In addition, the high water table may have trapped some gasoline above the gasoline-contaminated zone
18
-------
BTEXTMB (mg/L)
0 100 200 300 400
2,5 T-
PZ4C
BTEXTMB (mg/L)
0 250 500 750 1000
PZ4D
BTEXTMB (mg/L)
0 100 200 300 400
0.5
E
Cl
w
Q
PZ4E
2.5
BTEXTMB (mg/L)
0 100 200 300 400
0
' ! - '
o r
0.5
0.5
_ 1
„ 1 ¦
E
E.
& 1.5
£ 1.5 •
Q-
Q- '
O <
o M
° 2 ,
Q 2 T
<
>
o
2.5 <
> PZ4B
2.5
<
~
T
3
3 -
BTEXTMB (mg/L)
100 200 300 400
PZ4A
FIGURE 3-4 Vertical profiles of total dissolved BTEXTMB in the Nitrate Cell on November 22-26,1995, approximately
one month after gasoline injection. The water table depth was approximately 50 cm bgs. Note scale
change for piezometer PZ4D.
19
-------
BTEXTMB (mg/L)
0 100 200 300 400 500
0
0.5
1
1,5 6
B7or*
FZoLp
BTEXTMB (mg/L)
0 1000 2000 3000 4000
PZ3D
BTEXTMB (mg/L)
0 100 200 300 400 500
BTEXTMB (mg/L)
0 100 200 300 400 500
0
0.5
1
-B 1-5
Cl
a p
2.5
PZ3B
BTEXTMB (mg/L)
0 100 200 300 400 500
PZ3A
FIGURE 3-5 Vertical profiles of total dissolved BTEXTMB in the Control Ceil on November 22-26,1995, approxi-
mately one month after gasoline injection. The water table depth was approximately 50 cm bgs. Note
scale change for piezometer PZ3D.
20
-------
contacted by amended water during the flushing experiments. These regions were cored at the end of the
experiment and any mass that was present was included in mass balance calculations.
3.5.3 Experimental Design
Injection Procedures. To obtain data that could be used for a mass balance, injection systems
were designed to flush water continuously through the cells at a constant flux. A schematic of the
injection system is shown in Figure 3-6. A vertically-downward flow field was created by injecting into the
pea gravel layer (50 cm bgs), and extracting from the extraction well (screen; 200-250 cm bgs)
(Figure 3-6). The length of the vertical flowpath was about 1.75 m. Constant lower head conditions were
maintained by water level controllers (SSAC Inc., LLC5 Series) which automatically cycled the extraction
pumps on and off. The injection pumps operated continuously. By adjusting the elevation of the lower
heads, upper heads in both cells were maintained at about 50 cm bgs under constant injection conditions,
despite fluctuating water table elevations in the aquifer.
During the first week of the experiment, water was obtained from a shallow upgradient well located in
a flooded area. This well appeared to be pumping surface water which was fully oxygenated. To obtain
groundwater with lower dissolved 02, a shallow PVC well (2.49 m total depth) located approximately 6 m
upgradient of the cells was used as the water supply for the remainder of the experiment (Figure 3-2).
To inject the same dissolved oxygen concentration into both cells, a single 0.635 cm (1/4 in) diameter
polyethylene tube was installed in the well as the intake line. This line was then split with a Y-connector,
and fines were run to individual injection peristaltic pumps (Masterflex L/S Series). The Control Cell
received no additional amendments prior to injection; the Nitrate Cell received NO, via a Swagelok®
fitting located downstream of the injection pump.
Dissolved oxygen was controlled by placing an aerator tube in the supply well. The aerator was
located just above the water intake tube to minimize incorporation of bubbles. The target dissolved 02
concentration range was 2-4 mg/L. This arrangement provided reasonable control over dissolved 02
concentrations, but also appeared to stimulate growth of unidentified bacteria (possibly Fe oxidizers) in the
supply well, injection tubing, and injection wells. During the final 24-day flushing experiment (May 1997),
the injected dissolved O concentration was increased to ca. 5 mg/L by switching the water supply from
groundwater to the pond adjacent to the site.
The injection N03 concentration was controlled by pumping a continuously-stirred, concentrated Non-
stock solution into the injection flow line, A dedicated peristaltic pump (Masterflex L/S Series) was used to
feed stock solution into the line. The pumping rate was set to yield a 100x dilution. The stock tank held
40 L of solution and was replenished every two weeks. The initial target injection concentration was
150 mg/L NOs, but because utilization was low, the target concentration was lowered to 100 mg/L on day
67. During the 24-day flushing experiment, the possibility that biotransformation was nutrient-limited was
investigated by pumping a diluted, modified Bushnell-Haas (MBH) medium (HK?P04, H2KP04, NH4N03,
MgS04-7H20, CaCI2*2H20, FeCL«6H?0) (Mueller et al., 1991) into both cells using a concentrated stock
solution and dedicated feed-pump as above. Because of pump malfunctions, the solution was pumped in
intermittently over the 24 day period. Actual injected concentrations are not known because complete
dissolution of the concentrated stock solution could not be maintained.
Injection flow rates were set by calibrating the peristaltic pump controllers, and measured periodically
(n=22) using a stopwatch and graduated cylinder. During the first two months of the experiment, target
flow rates were adjusted periodically to achieve steady flow conditions and the desired residence time.
Target flow rates and actual measurements are shown on Figure 3-7. To calculate a mean flow rate
representative of the entire 174-day flushing experiment, a mean and standard deviation were calculated
for each of the three target injection-rate periods (Figure 3-7). An overall mean was then calculated from
the individual mean injecti 1 on rates, and a pooled standard deviation from the individual standard
deviations. These calculations yielded rates of 250 ± 14 ml/min and 237 ± 23 ml/min for the Nitrate and
Control cells, respectively. The uncertainty is the standard deviation of the temporal distribution of
injection rates. These injection rates correspond to residence times of about 9 days, or 20 treatment-cell
pore volumes over 174 days. On the basis of a measured rate of 200 ml/min (n=1), an additional three
pore volumes were removed from each cell during the 24-day flushing period the following spring.
21
-------
Injection System
Peristaltic
Pump
Compressed
Air
Peristaltic
Pump
-5
1 OOx Dilution]
Concentrated
Nitrate Solution
Upgracfent
Supply Well
(Location not to scale)
Upper Head
Lower Head
Sample
Port
30 -
60 -
90 -
120-
150-
180-
210-
240-
270-
300-
(cm)
0 j ^
•t-
S
Peristaltic
Pump
Sample
Fori
i>
'Treatment
Water Level
Controller
Pea Gravel
Layer
Gasoline-
Contaminated
Zone
Peat Layer
® > Injection Wells
A — > Extraction Well
~ Multilevel Piezometers
0 Drive-point wells
VERTICAL EXAGGERATION : 1,1
FIGURE 3-6 Injection system schematic for the Nitrate Cell. Cross section of cell shows selected instrumentation to
illustrate positions of injection/extraction wells and multilevel piezometer ports. The system for the
Control Cell was identical except for N03 addition equipment.
22
-------
400
1ET
j= 300
-------
After pumping ca. 25 ml of water to flush the piezometer tubes of stagnant water, dissolved 02
concentrations were measured colorimetrieally in the field (indigo-carmine method (Gilbert et a'.., 1982)).
Samples were collected under low-flow conditions from a flow-through sample cup attached to the end of
the peristaltic-pump outlet tube (Figure 3-8). Samples were typically colorless, and free of turbidity. To
obtain a sample, a CHEMetrics Vacu-Vial® 02 ampoule was inserted in the flowing stream of water and
filled. The concentration was read in a CHEMetrics VVR spectrophotometer, using an 02-speeific filter
that automated sample quantification. Two filters were used in this study: a 0-15 mg/L range filter, and a
0-2 mg/L range filter. Each filter contained a pre-programmed calibration curve, and did not require daily
calibration. The accuracy of the calibration was tested in the laboratory using air-sparged water (aerobic)
and a 2% (w/v) sodium sulfite solution (02-free). The ambient temperature of the laboratory was 24°C.
This test yielded concentrations of 7.54+0.18 mg/L (n=5) for the aerobic water, and 0.0+0.0 (n=5) for the
Oa-free water. At this temperature the concentration of dissolved 02 in equilibrium with air (one
atmosphere total pressure) is 8.25 mg/L (Drever, 1992). The majority of the field samples were analyzed
with the 0-2 mg/L range filter. The specified method detection limit (MDL) for this filter was 0.05 mg/L
dissolved 02.
If required, samples for inorganic parameters were then collected from the outlet tube of the peristaltic
pump (Figure 3-8). Nitrate samples were collected in 18-ml scintillation vials and preserved with two
drops of formaldehyde. Iron and SO/ samples were collected in 30-ml glass syringes and then filtered
(0.45 (jm) into scintillation vials. Samples for Fe analysis were acidified with 2-3 drops of concentrated
hydrochloric acid. Groundwater samples for dissolved CH4 analysis were collected without headspace in
40-ml screw-top glass vials. All of these samples were shipped back to the University of Waterloo for
analysis (Appendix C).
Total dissolved sulfide concentrations were also measured in the field using a CHEMetrics
colorimetric analysis (MDL=0.06 mg/L) in a manner similar to D.O. Total alkalinity as CaC03 was
measured in the field with a Hach Digital Titrator (Model 16900).
Injection/Extraction Samples. Concentrations of electron acceptors injected into the cells were
determined by frequent (daily to bi-weekly) sampling of injection water for N03, N02, and dissolved 02,
FLOW THROUGH
SAMPLE CUP
INORGANICS
k D.O,) X
1/8-INCH
TEFLON TUBING
1/8-INCH
S.S PIEZOMETER
TUBE .
S.S. SWAGELOK
FITTINGS
S.S. HEAD
& TUBING
PERISTALTIC
PUMP
CLAMP
18-ML CRIMP
TOP VIAL
FIGURE 3-8 Schematic of groundwater sample collection apparatus. Dissolved Oz samples collected from sample
cup under flowing stream of water.
24
-------
To provide a baseline, the injection stream was also sampled at feast once for dissolved organic carbon
(DOC), pH, temperature, alkalinity, Fe, SO/', HS", and BTEXTMB. Similarly, extraction water was
analyzed regularly for N03\ NO?, dissolved 02, and BTEXTMB, and less frequently for the inorganic
parameters listed above. Samples of injection water were collected directly from the end of the injection
tubing, and extraction water from a sampling port (Figure 3-6). Collection procedures and preservatives
were as described in the preceding paragraphs.
Metabolites. Evidence for biotransformation was obtained by analyzing samples for hydrocarbon
metabolites. These compounds (aromatic and aliphatic acids and phenolic compounds) result from
incomplete mineralization of various hydrocarbon constituents. Three sets of samples were collected, all
under static conditions; May 1997, early June 1997, and late June 1997. The first two sets of samples
were shipped to the National Center for Integrated Bioremediation (NCIBRD) at the University of Michigan,
Ann Arbor for analysis, and the third to the NRMRL, Ada, Oklahoma.
Samples were collected from two locations within each cell, and the groundwater supply well. A fresh
sample of API 91-01 gasoline was also analyzed by NRMRL for these compounds. Samples were
collected from the peristaltic pump outlet tube in 500 ml amber bottles, preserved with either KOH to
pH=11 or 1% (w/v) trisodium phosphate solution, and refrigerated until shipment.
Aquifer Cores From Treatment Ceils. When the field experiments were completed, cores were
collected to estimate the mass of BTEXTMB remaining in the treatment cells. Core subsamples were
immersed in solvent immediately after collection to obtain an estimate of total contaminant mass in the
aquifer sample, including mass in the sorbed, aqueous, and residual gasoline phases. Because a non-
wetting phase liquid at residual saturation in a homogeneous sand occurs as disconnected singlet and
multi-pore blobs (Chatzis et al., 1983), which are readily transferred to a contacting solvent phase, a one-
step extraction was used in this study: Prior to analysis, gasoline was extracted into the methanol phase
by shaking samples for 15 min at 300 rpm on an orbital shaker, and then sonicating for 3 min to remove
any emulsification that may have been present in the methanol. This method was similar to other
published field solvent-extraction approaches (Ball et al., 1997; Vandegrift and Kampbell, 1988). For
aquifer samples with low concentrations of organics, the efficiency of the extraction, particularly of the
sorbed phase, is critical for obtaining accurate results (Ball et al., 1997). In this experiment, on the other
hand, there was a large absolute amount of mass (ca. 7 kg BTEXTMB recovered in each cell), and a high
percentage of this mass was necessarily present within the gasoline phase; the sorbed and aqueous
phases contributed negligible mass to the total estimate. Based on the above reasoning, if a negative
bias was associated with the extraction technique (i.e., less than 100 % recovery), it was probably small
relative to the total mass in the sample, as well as other uncertainties in the mass balance.
Grids consisting of nine equal areas were used to determine core locations. One core was collected
from the center of each area. An additional core was collected from a random location near the center of
each cell. Core runs were typically from ground surface to 152 cm (5 ft) with recoveries averaging
117 cm and 124 cm from the Control and Nitrate cells, respectively. A deeper interval, 152 cm to 305 cm
(5 ft to 10 ft), was also obtained from the cores at cell centers (cores 3J and 4K). For most cores eight
uniformly-spaced subsamples were collected, corresponding to one subsampie per 15-cm interval. Mass
balance calculations were therefore based on approximately 80 sample locations per cell.
Subsamples of aquifer material were collected within 15 minutes of core extraction. To obtain these
samples, a 2-cm wide strip of the core barrel was removed with a circular saw to expose the aquifer
material. At each sampling location the outer layer of sand was scraped away, and a ca. 10-g plug of
sand was obtained by inserting the barrel of a dedicated, plastic 10-cc syringe into the core. The plug of
sand was then quickly extruded into a pre weighed 40-ml screw-top glass vial containing 15 ml methanol,
re-capped, and refrigerated until analysis. Sample bottles were weighed before and after methanol
addition to accurately determine the mass of methanol. After sample collection, bottles were weighed
again to determine the wet weight of the aquifer sample. Methanol was dispensed into sample bottles no
more than 36 hours prior to use, and stored in the refrigerator to minimize volatilization. There were no
samples with visible loss of methanol. Samples were analyzed one week after collection.
25
-------
To estimate depletion of other gasoline constituents, a total of six additional subsamples of core
material were collected from randomly-selected locations in both cells for a semiquantitative analysis of
gasoline components (i.e., grouped as alkanes, aromatics, bicycloalkanes, naphthenes, olefins, and
PNAs). The core subsamples were extruded into 40-ml screw-top glass vials containing 25-ml deionized
water and refrigerated until shipment to the NRMRL for analysis. For comparison, fresh API 91-01
gasoline and an aliquot of Borden sand spiked with API 91-01 were also analyzed for these component
groups. Details of the analytical method are provided elsewhere (Hutchins et al., 1998).
26
-------
CHAPTER 4 LABORATORY EXPERIMENTS
VnPl IjMki^ I Km 1 1 Tr« Lrt DV/llrt I It I BwMiiMkl till IVI Beiii I i I
The results of the laboratory experiments are summarized in this chapter. Each of the microcosm
experiments listed in Table 1-1 is described in detail here. Section 4.1 includes all of the microcosm
experiments performed with pristine Borden sand. Experiments 1, 2, and 3 were performed before the
field experiment design was finalized. Experiments 4 and 5 were set up after the field experiment was
designed and the gasoline spilled at Borden, but before the field data collection was initiated. Section 4.2
includes the follow-up microcosm experiment performed with aquifer material collected from the Nitrate
Cell. Because of the large amount of detailed information, a results and discussion section is provided for
each individual microcosm experiment described in these sections. Section 4.3 contains results of all of
the microbiological characterization work performed over the course of this study. The major conclusions
from all of the microcosm experiments and microbiological characterization studies are summarized in
Table 4-1. The final section (Section 4.4) includes a general discussion of the experimental work
performed in the laboratory.
In discussing these experiments, the term "denitrifying activity" is used in a general sense to define
any dissimitatory N03 -reducing activity occurring under anaerobic conditions. The term "denitrification" is
only used when an assay was performed to confirm that NO., was reduced to gaseous end products.
Except where noted in Section 4.2, N03 depletion under anaerobic conditions is assumed to be a
dissimilatory process where NCV is utilized as an electron acceptor. This assumption is consistent with
previous experiments in which Borden aquifer material was amended anaerobically with aromatic
hydrocarbons and NO," and denitrification was confirmed with an assay (e.g., Barbara et al., 1992).
4.1 Microcosm Experiments: Pristine Borden Sand
Because this study included a controlled release of gasoline in the field, Borden aquifer material from
an existing spill was not available for preliminary laboratory experiments. Thus the experiments that
provided baseline information on electron-acceptor and aromatic hydrocarbon utilization were performed
with aquifer material that had no known prior exposure to hydrocarbon contamination. The microbial
community in a low-carbon environment such as the Borden aquifer may require time to adapt to the
unfavorable conditions created by a hydrocarbon spill (Chapelle. 1993). It was anticipated therefore that
the hydrocarbon-degrading capabilities of contaminated aquifer material would increase with exposure.
This provided an opportunity to compare aromatic-hydrocarbon and electron-acceptor utilization in pristine
and contaminated Borden sand for evidence of acclimation of the indigenous microbial population.
4.1.1 Experiment 1: Effect of Oxygen, Nitrate Concentration, and Nutrients
Biotransformation of aromatic hydrocarbons under NOq-reducing conditions was previously
demonstrated in the Borden aquifer, in the deeper, anaerobic zone affected by leachate emanating from
an abandoned landfill (Barbaro et al., 1992). The current field experiment was designed to take place in
the upper, pristine aquifer, which is generally aerobic in character, although low dissolved 02 zones are
present. This laboratory experiment was designed to evaluate the potential for aromatic hydrocarbon
utilization at the expense of NO.; and O in this upper aquifer material. Treatments were set up to screen
for 02-supported and N03-supported Bl EX degradation, to investigate the extent of mass loss under a
range of N03" concentrations, and to determine whether added inorganic nutrients (N and P) influenced
BTEX biotransformation.
Design 2 microcosms (Section 3.2.1) were prepared using pristine Borden sand obtained in the
saturated zone approximately 1 m to 1.5 m below ground surface, and were amended with an aliquot of a
neat BTEX mixture (B;T:E:p-X:m-X:o-X in ratios of 3:2:1.i.i:1) to about 12.4 mg/L BTEX. Microcosms
designed for repeated sampling were advantageous for this application because a broad range of
experimental conditions could be evaluated simultaneously in an experiment of practical size. The
27
-------
Table 4-1. Summary of Major Conclusions from Laboratory Experiments
Report Section
Experiment
Major Conclusions
4.1.1
Effect of 02, NOg"
concentration, arid inorganic
nutrients
(Experiment 1)
Aerobic BTEX biotransformation initiated readily in
Borden aquifer material; biotransformation under
denitrifying conditions more limited, with only toluene
and ethylbenzene degrading consistently; addition of
inorganic N and P was needed for optimal aerobic
degradation of 12.4 mg/L BTEX, but had no discernable
effect on anaerobic activity
4.1.2
Comparison of BTEX
biotransformation under
denitrifying conditions in CFB
Borden, Eglin AFB and Park
City aquifer microcosms
(Experiment 2)
Aromatic hydrocarbon biotransformation under
denitrifying conditions was very limited in pristine
Borden material compared to two other petroleum-
hydrocarbon contaminated sites
4.1.3
Biotransformation of BTEX in
gasoline-contaminated
groundwater under
denitrifying, microaerophilic,
and mixed electron-acceptor
conditions
(Experiment 3)
Patterns of BTEX, 02, and N037N02 concentrations
suggested that 02 and NOs were used sequentially
under mixed electron-acceptor conditions; most BTEX
biotransformation occurred early, likely at the expense
of microaerophilic O,: additional TEX losses occurred
later under mixed electron-acceptor, and denitrifying
conditions; benzene losses under microaerophilic
conditions were minimal
4.1.4 Effect of BTEXTMB
concentration under
microaerophilic / N03"
conditions in pristine Borden
aquifer material
(Experiment 4)
4.1.5 Effect of 02 concentration
under high BTEXTMB
concentration conditions in
pristine Borden aquifer
material
(Experiment 5)
4.2.1 Extent of biotransformation
under various substrate and
mixed electron-acceptor
conditions in gasoline-
contaminated Borden aquifer
material extracted from Nitrate
Cell
(Experiment 6)
4.3.1 and 4.3.2 Microbial characterization
results
Negligible BTEXTMB losses at gasoline-saturated
aqueous concentrations, likely at the expense of
microaerophilic 02; patterns of N03" uptake suggested
denitrifying population in pristine material inhibited by
high substrate concentrations
Aerobic biotransformation at gasoline-saturated
aqueous concentrations highly variable in pristine
aquifer material: may indicate patchy distribution of
aerobic populations tolerant of high substrate
concentrations
Large 02 demand and BTEXTMB mass loss relative to
pristine aquifer material; nitrate possibly used as N
source during aerobic degradation; relatively rapid
continued N03" utilization following Oz depletion, but no
discernable utilization of aromatlcs; differences
between pristine and contaminated aquifer period
suggested acclimation had occurred with large potential
G? demand; minor mass loss under microaerophilic /
N03" conditions consistent with in situ observations
Numbers of culturable aerobic heterotrophs and
denitrifiers in Borden aquifer variable but consistent with
other pristine shallow aquifers; microbial numbers in
treatment cells moderately elevated relative to pristine
background locations; aerobic dehydrogenase activity
assays suggested that microbial activity was partially
inhibited by gasoline phase
28
-------
Table 4-2, Design Summary, Experiment 1
Treatment
NO.-N (rmg/L)
Inorganic Nutrients'
Replicates
Aerobic
Active
0
+ N, P
1
Active
0
-N, P
1
Active
25
+ N, P
4
Anaerobic
Sterile2
25
+ N, P
2
Sterile
25
-N, P
1
Active
0
+ N, P
2
Active
0
-N, P
1
Active
5
+ N, P
2
Active
5
-N, P
1
Active
25
+ N, P
7
Active
25
- N, P
1
Active
25
+ mineral salts
2
Active
25
- mineral salts
1
Active
50
+ N, P
1
Active
50
-N, P
1
1 N, P: 5 mg N + 2 mg P per L, as NH«CI and KH2PO4 mineral salts: 2 mg K?HPC>4,2 mg KH2PO4,
3.3 mg NH^CI, 0.4 mg MgSa.7HzO, 0.04 mg CaCl2.2H=0,0.01 mg FeCb.6H£> per L (0.8 mg P
+ 0.86 mg N per L).
2 autoelaved aquifer solids + poisoned (0.01 % HgCh) groundwater.
experimental design is summarized in Table 4-2. The design called for multiple replicates for each
treatment, but during microcosm preparation, stock solution was incorrectly added to some of the aerobic
microcosms. As a result, an aerobic treatment that contained NOs-N but was lacking N and P was not
prepared. Consequently, utilization of N03 -N as a nutrient during aerobic biotransformation could not be
evaluated in this experiment. Strong evidence for assimilatory NO / reduction was obtained in a
subsequent experiment (Experiment 6; Section 4.2.1). Microcosms were sampled periodically over a
250 day period.
Microcosms incubated within the anaerobic chamber were removed for sampling, which was
conducted by syringe under flowing argon gas, then returned to the chamber. It was hoped that use of
threaded rather than older push-type mininert™ valves might minimize microcosm leakage, but problems
were encountered with some vials, presumably as they were passed through the evacuation chamber of
the anaerobic glovebox after sampling. All subsequent laboratory experiments were therefore conducted
using sacrificial (Design 1) microcosms. The HgCI2 solution used in sterile control microcosms was found
to interfere with the N03" analysis, and therefore, concentration data were not obtained for controls.
Consequently, a sodium azide solution was used in sterile control microcosms in all subsequent laboratory
experiments (Experiments 2 through 6). On the basis of these later experiments, NOs utilization was
negligible in sterile controls.
Results and Discussion. As expected, based on earlier CFB Borden studies (e.g., Barker et al.,
1987), BTEX concentrations decreased rapidly in active, aerobic microcosms to below detection by day 8
(Figure 4-1). A second BTEX amendment on day 32 (Figure 4-1, Arrow) was also rapidly metabolized in
NH4-N and P04-P-containing microcosms, in both presence and absence of NQ3. Inorganic nutrient
addition was necessary to obtain maximum degradation under the experimental conditions evaluated
here. In the absence of N and P, only slow biotransformation occurred (Figures 4-1 and 4-2), with m- and
29
-------
aerobic
BTEX
added
o 0.8
Q
^ 0.6
X
LU
0.4
£
0.2
¦as-
30
40
10
20
0
Time (days)
—<¦— sterile o -N03.+NP -w- -N03.-NP v +N03,+NP
FIGURE 4-1 Aerobic BTEX biotransformation in the presence or absence of N03 and/or NH4-N and P04-P. Active,
+N, P microcosms were reamended with 1 p.L of BTEX on day 32 (arrow).
aerobic, -N03, -NP
700
600
N, P added
g> 500
X
LU 400
I—
m
g 300 -
ts
a,
-g 200:
S <
^ 100
100
150
200
250
300
0
50
Time (days)
B d T —Ethb —»c p-X —x ¦ m-X o o-X
FIGURE 4-2 Aerobic BTEX biotransformation under N05-free, N,P-free conditions. NH4GI and KH2P04 were added
on day 253 (arrow).
30
-------
p-xylene declining first (Figure 4-2). Addition of N and P to this microcosm on day 253 (Figure 4-2, Arrow)
resulted in total BTEX depletion within 6 days, providing clear evidence that microbial activity was
restricted in the absence of inorganic nutrients. Note that in these microcosms 02 was available in
excess.
In contrast, far less BTEX biotransformation occurred under anaerobic conditions, although the
addition of N03 did stimulate some aromatic hydrocarbon depletion. No BTEX loss was observed in
anaerobic, N03-free microcosms (Figure 4-3), but ethylbenzene degraded to below detection in all
NOg-amended microcosms, and toluene degradation was also observed in some, but not all
N03-amended microcosms (Figure 4-4). The other aromatic hydrocarbons appeared recalcitrant. As
noted, microcosm leakage was a problem; therefore Figure 4-4 shows data only from selected
microcosms that appeared to maintain integrity. Ethylbenzene degradation began in the N03 -amended
microcosms after a lag period of 50-70 days. In contrast, when toluene loss occurred a lag period was
not apparent, but a plateau effect was always observed, where -0.5-2% of the initial toluene persisted for
an extended period before declining to below detection (e.g., Figure 4-4b).
The presence or absence of inorganic nutrients (either N and P, or the mineral salts suite) did not
affect anaerobic BTEX behavior (Figure 4 4b vs 4-4d). Presumably this was because the total
hydrocarbon loss in even the most active N03 -amended microcosms was small compared to that in the
aerobic BTEX-degrading microcosms. In addition, N03" may have satisfied an assimilatory N requirement
in these microcosms. Evidence of a nutrient limitation, therefore, was not observed. Nevertheless,
addition of NH4-N and P04-P was continued in subsequent microcosm experiments, to ensure that
inorganic nutrient limitation would not impede anaerobic hydrocarbon biotransformation in the event of
greater degradative activity. Provision of NH4 N was intended to satisfy biomass production requirements
so that NOg- would be available as a potential electron acceptor (Hutchins, 1991b).
Nitrate concentration had no discernable effect on ethylbenzene biotransformation, except when no
N03" was provided. Eventual toluene biotransformation to undetectable levels was observed in seven of
11 microcosms amended with 25 mg/L NO,-N, but in none of three 0 mg/L N03-N microcosms, three
5 mg/L N03-N microcosms or two 50 mg/L N03-N microcosms. On this basis, of the concentrations tested
in this experiment, the 25 mg/L level of N03-N was judged optimal. After the day 253 BTEX analyses
were completed, all of the monitored microcosms were sacrificed and analyzed for pH and N03 / NO,/
concentrations. The final microcosm pH ranged from pH 7.0 to pH 7.6. The NOg7 NO,/ data are
summarized in Table 4-3.
A few 25 mg/L NO..-N amended microcosms were analyzed on day 0, but then incubated
continuously within the anaerobic chamber until sacrificed for BTEX, D.O., N03 and N02 analysis on
days 71 or 134 (Table 4-4). Aromatic hydrocarbons lost due to passage through the glovebox evacuation
chamber should therefore have been minimal; hence these microcosms confirmed the occurrence of
toluene and ethylbenzene biotransformation under denitrifying conditions. The removal of these
compounds is consistent with previous results obtained under anaerobic, denitrifying conditions
(Barbara et al., 1992).
Overall this initial screening experiment provided results that were consistent with previous
experiments. In the Borden sand, BTEX biotransforms readily in the presence of 02, and, in the presence
of excess 02, the extent of biotransformation can be limited by inorganic- nutrient availability. The
experimental setup did not allow for a determination of whether a particular nutrient, phosphorous or
nitrogen, limits the reaction. Under anaerobic, denitrifying conditions, only toluene and ethylbenzene
biotransform at an appreciable rate relative to the length of the incubation period, and there is no
discernable increase in mass loss in the presence of inorganic nutrients.
4.1.2 Experiment 2: Comparison of BTEX Biotransformation Under Denitrifying Conditions in CFB
Borden, Eglin AFB and Park City Aquifer Microcosms
In this experiment, the BTEX-degrading capacity under denitrifying conditions of pristine Borden
aquifer material was compared to that of aquifer material obtained from two petroleum hydrocarbon-
exposed sites; Park City, Kansas (petroleum-hydrocarbon mixture), and Eglin AFB, Florida (jet fuel). The
subsurface microbiota from both US sites are known to degrade aromatic hydrocarbons under denitrifying
conditions (e.g., Kennedy and Hutchins, 1992; Thomas et a!., 1995).
31
-------
sterile
800
uu 500
I—
« 200
100 150 200
Time (days)
300
B —E3— T
EthB v— p-X x m X
o-X
800
700
m 500
CD
400
® 300
co 200
anaerobic, no nitrate-N
0
50
100
150
Time (days)
200
250
300
T
-X-
m
FIGURE 4-3 BTEX biotransformation in {a) anaerobic sterile control and (b) anaerobic N03-free microcosms. Plotted
values are the mean of the three microcosms (two +N, P, one -N, P).
32
-------
5 mg/L nitrate-N, +N,P
25 mg/L riitrate-N, +N,P
800
500
x
LU
H-
m
a>
a
CO
w 300
100 150 200
Time (days)
250 300
800
? 500
» 300
100 150 200 250 300
Time (days)
-B-T
¦ Ethb -V- p-X -X- m-X -6- o-X
B -B-T
. Ethb -V- p-X -X- m-X -Q- o-X
50 mg/L nitrate-N, +N,P
25 mg/L nitrate-N, -N,P
S> 500
& 300
100 150 200
Time (days)
, B -B-T
. Ethb -*J- p-X m-X o-X
c£ 400
100 150 200
Time (days)
250 300
_B_ j -ir- Ethb p-X m-X -e- o-X
FIGURE 4-4 BTEX biotransformation in selected anaerobic microcosms: (a) 5 mg/L N03-N+N,P, (b) 25 mg/L N03-
N.+N.P, (c) 50 mg/L N03-N, +N, P, (d) 25 mg/L N03-N. -N,P.
33
-------
Table 4-3. Nitrate-N and Nitrite-N in Anaerobic Microcosms on Day 253
Mcrocosm Condition NQ3-N (mg/L) NQ?-N (mg/L)
Sterile, 25 mg/L NO..-N amended (n = 3)
n.t,5
n.t.
Active, 0 mg/L NQrN amended (n = 3)
n.d.2
n.d.
Active, 5 mg/L NO3-N amended (n = 3)
<0,063
0.06 (s.d. =0.10)
Active, 25 mg/L NQs-N amended (n = 9)
15.45 (s.d. = 1,9)
0.38 (s.d. = 0.19)
Active, 50 mg/L NOa-N amended (n = 2)
38.77 (s.d. = 1.01)
0.45 (s.d. = 0.33)
1 n.t., not tested.
2 n.d, none detected
3 small peakdetected, but belcw quantifiable limit.
Table 4-4. Initial and Final BTEX, D.O., and Nitrate/Nitrite Concentrations in Denitrifying Microcosms
Time Heads pace Gas BTEX (jig/L)
(days)
D.O.
(mg/L)
NOb-N
(mg/L)
NCfe-N
(mg/L)
Ben
Tol
Eben
p-X
m-X
o-X
Microcosm 1
0 716
471
253
211
207
143
n.t.
n.t.
n.t.
2 n.t.1
n.t.
n.t.
n.t.
n.t.
n.t
0.54
n.t.
n.t
71 427
8
172
157
159
112
0.24
15.4
n.d.
Microcosm 2
0 697
460
251
210
208
147
n.t.
n.t.
at.
71 624
2
180
188
191
116
0.32
8.1
ad.2
Microcosm 3
0 698
462
249
207
205
142
at.
n.t.
at.
134 665
n.d.
n.d.
147
150
93
n.d.
6.5
0.13
1 n.t., not tested.
2 n.d. none detected.
Table 4-5. Park City, Kansas and Eglin AFB, Florida Subsurface Materials
Park City, Kansas Sample labelled 6QG38/39; collected 25 ft below grade, several feet below significant
contamination; coarse sand ,
Eglin AFB, Florida Sample labelled 80KC6; collected from 5.5-6.3 ft below grade, downgradient of
residual contamination source; fine sand
34
-------
Design 1 microcosms were constructed with pristine Borden aquifer material, and aquifer material
provided by the U.S. EPA NRMRL from the Eglin AFB, Florida and Park City, Kansas sites. Information
available for the latter two materials is summarized in Table 4-5. Groundwater was not available from the
US sites, so Borden groundwater amended to about 12.4 mg/L BTEX with a neat BTEX mixture (as in
section 4.1.1) was used in all microcosms. All microcosms were provided with 25 mg/L N03-N, 5 mg/L
NH4-N and 2 mg/L PO.-P. One sterile control and three active microcosms were prepared for each
material.
Results arid Discussion. Benzene was recalcitrant in all microcosms, and o-xylene was also
persistent, although there was evidence that o-xylene was subject to cometabolism in the Eglin and Park
City materials (Figure 4-5). The most striking feature of Figure 4-5 is the lack of activity in the Borden
microcosms compared with Eglin and Park City microcosms. Substantial toluene depletion (to a plateau
sterile
Borden
900
en 600
S 400
55 300
20 40 60 80
o> eoo
600
u 500
5 300
Rark City
d> 600
H-1 500 Q
20 40 60 60
Time (days)
20 40 60
Time (days)
¦ B ~ T A Ethb A p-X x ro-X o o-X
FIGURE 4-5 Comparison of BTEX biotransformation in Borden, Eglin and Park City microcosms under denitrifying
conditions. Plotted values are the mean ± s.d. of three replicate microeoms (active) or three single
microcosms from each site (sterile control).
35
-------
of ca. 10 pg/L headspace toluene) did occur, but in only one of the three active Borden microcosms. This
experiment was continued for 168 days, but all compounds persisted in the three active Borden
microcosms, with the exception of the one toluene-degrading microcosm, which continued to degrade
toluene upon reamendment on day 106 (data not shown)
Reamendment of single Eglin and Park City microcosms with toluene on day 106 led to further
declines in o-xylene levels, although interestingly, reamendment of other microcosms with ethylbenzene
(Eglin) or p-xylene (Park City) did not stimulate further o-xylene losses (data not shown). The TEX loss
pattern differed in Park City and Eglin microcosms; the former were least efficient in p-xylene removal,
whereas ethylbenzene was transformed most slowly by the Eglin microbiota (Figure 4-5). The degree of
activity in the Eglin and Park City microcosms confirmed that the provided N, P levels and 25 mg/L NCX-N
would support anaerobic TEX biotransformation at the concentrations tested here.
Aerobic heterotrophic plate counts for the three aquifer materials were fairly similar, but the denitrifier
counts for these particular samples were 10-100x larger in the exposed materials than in the pristine
Borden material (Table 4-6). As discussed in Section 4.3,1, denitrifiers recovered from pristine Borden
aquifer material sometimes exceeded 104/g. Nevertheless, robust hydrocarbon-degrading activity,
comparable to that in the Eglin or Park City aquifer material, was not observed under denitrifying
conditions in any of our experiments with pristine Borden sand. These data demonstrated, therefore, that
the size of the denitrifier population, as represented by denitrifier counts, was not a reliable indicator of
aromatic-hydrocarbon degrading activity in pristine Borden aquifer material.
In this experiment, denitrifying activity in aquifer material that had been exposed to hydrocarbons was
compared with material with no known prior exposure. It was therefore not unexpected that TEX losses
were more extensive in the exposed materials. Given the weak denitrifying activity observed in the
laboratory, and assuming that the results in static laboratory microcosms are reasonably representative of
a dynamic field setting (Barbara et al., 1992), the Borden aquifer would appear to be a poor candidate for
nitrate-based bioremediation. However, we hypothesized that more extensive hydrocarbon-degrading
activity would develop in situ after a period of exposure to gasoline hydrocarbons, 02, and NO,-. The
laboratory data obtained with exposed Borden aquifer material are presented in Sections 4.2 and 4.3, and
discussed in Section 4.4.
4.1.3 Experiment 3: Biotransformation of BTEX in Gasoline-Contaminated Groundwater Under
Denitrifying, Mlcroaerophilic and Mixed Electron-Acceptor Conditions
This experiment was performed to compare the extent of BTEX biotransformation under denitrifying,
microaerophilic, and mixed electron-acceptor conditions. Design 1 microcosms were prepared with
pristine Borden aquifer material, diluted (approximately 10x) gasoline-saturated groundwater, NO., and
inorganic nutrients. Sets of replicate microcosms were prepared under either anaerobic denitrifying,
microaerophilic, or mixed electron-acceptor (i.e., microaerophilic O.. plus NO, ) conditions. This
experiment was performed before the API 91-01 gasoline was acquired. Consequently, the gasoline-
saturated groundwater was prepared with a locally-available unleaded gasoline, which was amended with
pure benzene to provide a BTEX ratio similar to that investigated by Barbara et al. (1992). The initial
Table 4-6. Microbial Enumeration of Borden, Park City and Eglin AFB Aquifer Materials
Sample
Heterotrophic Plate Count'
Denitrifier Count2
CFU/a drv wt < s.d.)
MPN/a drv wt (95% confidence)
Borden
1.1 x 10s (3.8 x104)
2.8x10" (24- 1.5x103)
Park City
9.5 x10s (7.2x10")
2.6 x 10s (4.5 x 10s-1.5 x 1(f)
Eglin AFB
1 1 X.J
9.9x10" (1.6x10")
4.6x 103 (8.2x 10s-1.5x 1(f)
1 plated on R2A medium.
2 3-tube MPN using KNO3-1/10NB medium.
36
-------
Table 4-7. Design Summary, Experiment 3
Treatment1 Replicates Sampling Events
Sterile, denitrifying, 50 mg/L NO3-N
3
7
Sterile, microaerophilic, 50 mg/L NO3-N
3
7
Active, denitrifying, 25 mg/L NO3-N
3
7
Active, denitrifying, 50 mg/L NOs-N
3
7
Active, microaerophilic, 25 mg/L NOs-N + air
3
7
headspace
Active, microaerophilic, 50 mg/L NCb-N + air
3
7
headspace
Active, anaerobic, 0 mg/L NOs-N
3
7
Active, microaerophilic, 0 mg/L NO3-N + air headspace
3
7
1 All microcosms contained 5 mg/L NI-k-N, 2 mg/L POi-P. A 4-ml air heads pace was present as required.
BTEX concentration of the groundwater in the microcosms was approximately 13 mg/L. Microaerophilic
conditions were established with a 4-ml air headspace (Section 3.2.1). For each treatment group multiple
replicate microcosms were sampled periodically over a 348 day period for BTEX, dissolved Oa {modified
Winkler method (Appendix C)), NO,, and NO?. The experimental design is summarized in Table 4-7.
Results and Discussion (i). BTEX results from the microcosms are shown in Figures 4-6 and 4-7.
Benzene persisted in all microcosms for the entire 348 days of monitoring, except for small initial losses
under microaerophilic conditions. Substantial, rapid TEX losses occurred between days 1 and 8 in all
active microaerophilic microcosms (Figure 4-6). The patterns of compound losses were typical of aerobic
Borden microcosms (Section 4.1.1). m-Xylene was degraded to near or below detection levels, and
substantial p-xylene and ethylbenzene losses also occurred. The considerable variability in the data is a
common result when analyzing replicate samples of Borden aquifer material over time, but this factor
makes the interpretation of early toluene and o-xylene behavior more uncertain. These early compound
losses appeared to be more restricted in the N03 -free microaerophilic microcosms (Figure 4-6d) than in
those amended with N03 (Figures 4-6b and 4-6c). For example, -45% of the p-xylene remained in the
former, versus 5-10% in the latter. After day 8, degradative activity largely ceased in the microaerophilic
and mixed electron acceptor microcosms (Figure 4-6). No BTEX biotransformation was observed in
denitrifying microcosms over 140 days of incubation (Figure 4-7).
Sampling was suspended on day 140 because of the lack of BTEX-degrading activity. Approximately
7 months later (day 348 in Figure 4-6, Figure 4-7), another set of microcosms was sacrificed. The use of
mean values obscures the results of this sampling round somewhat, because the results differed for
individual microcosms within each set of three replicates. As apparent in some of the figures (e.g.,
toluene in Figures 4-6b, 4-6c, and 4-7c), further compound losses were initiated between day 140 and
day 348 in individual microcosms. The six microaerophilic / NOs microcosms analyzed on day 348
showed partial toluene degradation. Residual toluene ranged from 15 to 60% and from 25 to 64% of that
present in the day 348 sterile controls, for the microaerophilic / 25 mg/L NOa-N and microaero-
philic 150 mg/L N03-N conditions, respectively. Ethylbenzene was totally depleted in one of these
microcosms, but the actual ethylbenzene fate was difficult to discern because concentrations varied
substantially among replicates at all sampling events after day 1 (Figure 4-6). It is suspected that this
compound sometimes degraded when microaerophilic O was present to tens of pg/L which then
persisted, and sometimes failed to degrade, so that hundreds of pg/L persisted. Three of the six
denitrifying microcosms analyzed on day 348 had also degraded some toluene, but not ethylbenzene. All
other compounds remained recalcitrant in both the microaerophilic / NO, and denitrifying microcosms.
37
-------
sterile, microaerophilie, 50 mg/L NOa-N
1.25
1.25
microaerophilie, 25 mg/L NOs-N
"o" 0.75
O
o
X
LLl
0.5
0.25
0 50 100 150 200 250 300 350
Time (days)
r>- 0.75
0 50 100 150 200 250 300 350
Time (days)
microaerophilie, 50 mg/L N03~N
microaerophilie, nitrate-free
1.25
3
O
X
LLJ
I—
m
0.75
0.25
0 50 100 150 200 250 300 350
Time (days)
1.25
o 0.75
O
X
Hi
00
0.25
0 50 100 150 200 250 300 350
Time (days)
Ethb p-X —X—m-X —&—o-X
FIGURE 4-6 Biotransformation of BTEX in gasoline-contaminated groundwater under microaerophilie and mixed
electron-acceptor conditions. Plotted values are the mean ± s.d. of three replicate microcosms.
38
-------
sterile, 50 irg/L NOrN
25 mg/L N03-N
1.25
0.75
p
o
X
LJJ
H 0.5
m
0.25
1.25
0 50 100 150 200 250 300 350
Time (days)
• 0.75
S 0.5
0 50 100 150 200 250 300 350
Time (days)
50 mg/L N03-N
nitrate-free
1.25 •
x
iii
OS
CO u-3
0 50 100 150 200 250 300 350
Tire (days)
1.25
0 50 100 150 200 250 300 350
Time (days)
-B —a—T —A^Ethb —A— p-X —x— m-X —e—o-X
FIGURE 4-7 Biotransformation of BTEX in gasoline-contaminated groundwater under denitrifying and anaerobic, NO,
-free conditions. Plotted values are the mean + s.d. of three replicate microcosms.
39
-------
headspace Oa
16
14
12
c 10
-------
The results of 02 monitoring (Figure 4-8) support the conclusion that early BTEX losses in the
microaerophilie and the mixed electron-acceptor microcosms were oxygen-linked. One difficulty with the
use of an air-filled headspace to provide 02 is that the D.O. level in the water of active microcosms cannot
be measured continuously, and therefore is not known except at discrete measurement times. Oxygen
partitioning between the gaseous, liquid and solids phases will be affected continually by microbial 02
consumption, which will not likely be constant. In this experiment, both the headspace 02 and dissolved
02 contents were measured up to day 140 (Figure 4-8). Measurements on day 1 indicated a mean 02
concentration of 12.6% (s.d.=0.67, n=12) in the headspace of sterile and active microaerophilie
microcosms. The mean D.O. recorded for the same microcosms was 1.02 mg/L (s.d.=0.67. n=12).
However, it is clear in comparing D.O. data for the sterile and the active microaerophilie conditions
(Figure 4-8b) that a microbial 02 demand was already exerted by the day 1 sampling event. In active
microaerophilie microcosms the D.O. decreased to a mean level of 0.27 mg/L (s.d.=0.10) by day 8, and
persisted at a similar level thereafter (Figure 4-8b). D.O. was maintained at ca. 1.7-1.8 mg/L up to day 22
in the microaerophilie, sterile controls, but decreased to a mean of 0.98 mg/L by day 62 and continued to
slowly decrease thereafter. This may have reflected diffusion across or leakage around the septa and
gradual equilibration with the atmosphere of the anaerobic chamber (See Section 4.2.1).
The NO,-N and N02-N concentrations measured during the experiment are shown in Figure 4-9.
Losses of N03 were apparent. The mean NO -N concentrations of the microaerophilie / N03-N
microcosms on day 348 were 60% and 72% of initial concentrations for the 25 mg/L N03-N and 50 mg/L
N03-N conditions, respectively (Figure 4-9a); the mean NOs-N concentrations of the denitrifying
microcosms were 38% and 76% of initial concentrations for the 25 mg/L NO,-N and 50 mg/L N03-N
conditions, respectively (Figure 4-9b). Nitrite accumulation was minimal, but small amounts of N02" were
occasionally recorded in these microcosm sets, with the exception of the microaerophilie / 50 mg/L N03-N
microcosms (Figure 4-9). These results suggest that some denitrification occurred in the N03-amended
microcosms, which could have supported the late (i.e., after day 8) TEX losses that were observed under
mixed electron-acceptor (Figures 4-6a and 4-6b) and denitrifying (Figures 4-7a and 4-7b) conditions, but
were not apparent under microaerophilie, NO./-free (Figure 4-6d), or anaerobic, NO.-free conditions
(Figure 4-7d).
The overall patterns of BTEX and electron-acceptor utilization under the conditions investigated in this
experiment were most consistent with sequential utilization of 02 and NO.; as electron-acceptors.
Relatively rapid aerobic BTEX losses were followed by slower depletion of toluene and ethylbenzene
under anaerobic, denitrifying conditions. However, a comparison of Figures 4-6b, 4 6c, and 4-6d (i.e.,
mixed electron-acceptor conditions and microaerophilie conditions) is revealing, because the early
(between days 1 and 8) BTEX losses were more restricted under microaerophilie conditions. It's unlikely
that nutrient limitation in the microaerophilie microcosms explains the observed difference; previous
experiments (Section 4.1.1,Figure 4-1) demonstrated that the NH4-N and P04-P present in both treatments
exceeded assimilatory requirements for aerobic degradation of about twice the BTEX mass present in
these microcosms. One possible explanation for the difference between these treatments is the
occurrence of the so-called sparing effect in the mixed electron-acceptor microcosms (see Section 1.3.2;
Hutchins, 1991 a). After the addition of 02 to aromatic hydrocarbon rings, the partially-oxidized TEX
metabolites could have been rapidly metabolized by denitrifying microorganisms. Because considerably
less 02 is required to partially-oxidize the ring, this would 'spare' more of the limited 02 for further initial
oxidation of the parent compounds; hence more parent compound losses would potentially occur under
these conditions. The sparing effect could not occur in the NO,-free microcosms, because the only
electron-acceptor available to support continued degradation of any partially-oxidized TEX metabolites
would be 02. The possibility that this phenomenon might occur in the field was one of the reasons for
undertaking this study (Section 1,3.2).
The occurrence of the sparing effect cannot be unequivocally proven or disproven with the available
data because aromatic hydrocarbon metabolites were not monitored. But in this aquifer material the
explanation seems very unlikely. The sluggish hydrocarbon-degrading activity of the Borden denitrifying
community observed throughout this study suggests the occurrence of a concerted, rapid process of
aerobic oxygenase plus denitrifier metabolism resulting in rapid TEX losses is implausible. Moreover, one
can estimate the N03" demand potentially exerted by these hypothetical, partially-oxidized TEX
41
-------
nitrate/nitrite - mixed electron acceptor
60
50
ra
30
20
10
350
250
300
100
150
200
Time (days)
nitrate/nitrite - denitrifying
60
50
~ 20
10
0
300
350
250
200
100
150
50
Time (days)
—¦— ster, 50 rrg/L N03-N (N03) —*— 25 nng/L N03-N (NOG) —•— 50 rra'L N03-N (N03)
Ster, 50 mg/L N03-N (N02) 25 mg/L N03-N (N02) 50 rra'L N03 N (N02)
FIGURE 4-9 Nitrate and nitrite in (a) mixed electron acceptor and (b) denitrifying microcosms. Solid symbols and
lines = N03"-N; open symbols and dashed lines = NOz-N. Plotted values are the mean ± s.d. of three
replicate microcosms.
42
-------
intermediates, which is roughly similar to the NO, that would be required to support mineralization of the
equivalent amount of parent TEX compounds, and compare the estimated demand to observed losses.
This calculation indicates that total mineralization of the -1665 Fg/L TEX depleted in the mixed electron-
acceptor microcosms between days 1 and 8 would require roughly 8 mg/L N03 as an electron acceptor, if
complete denitrification to N2 is assumed. However, no NOg depletion was measured in these
microcosms over this interval, nor was NO detected on day 8 (Figure 4-9a). Alternatively, the mass of
O, in these microcosms appeared to be adequate for complete mineralization of 1665 pg/L TEX.
Assuming no biomass production, the aerobic reaction is
C47 H4() + 57 02 -» 47 C02 +20 W20 4-1
Equation 4-1 shows that the mass of 02 in a 4-ml air headspace (ca. 1.2 mg) was sufficient to
oxidize roughly 400 pg TEX, which exceeds the observed losses of 76 pg TEX (1665 pg/L x 0,046 L) over
this period. The Of mass would be adequate even if it was conservatively assumed that ca. 75% of the
02 was consumed by day 1 (Figure 4-9). These calculations demonstrate that the sparing effect does not
have to be invoked to explain the BTEX data, which is consistent with the apparent lack of early N03
utilization. However, it remains unclear why in this experiment early mass losses appeared to be more
extensive in the mixed electron-acceptor microcosms.
Results and Discussion (ii). This experimental design was repeated (without the anaerobic,
N03-free microcosms), with 25 mg/L N03-N and 10 mg/L N03-N, and core material that proved to be
more active under denitrifying conditions. For brevity, results for the active, 25 mg/L N03-N condition only
are shown (Figures 4-10 and 4-11). Trends in this experiment followed those of the experiment described
in the previous section (i.e., rapid aerobic TEX degradation followed by markedly slower, more restricted
hydrocarbon losses at the expense of N03), but in this experiment early aerobic losses were similar in the
microaerophilic and microaerophilic / N03 microcosms (Figures 4-10b and 4-10c). In addition, stronger
activity was observed under denitrifying conditions, which was initiated within less than 20 days
(Figure 4-10a). Dissolved oxygen was depleted rapidly over the first 7 days in those microcosms where it
was available, and a low-level plateau (<0.5 mg/L), equivalent in the microaerophilic and denitrifying
microcosms, was maintained thereafter (Figure 4-11 a).
The zero-order rates of N03-N depletion between days 1 and 68 were 0.15 and 0.13 mg/L/day for the
denitrifying and microaerophilic / 25 mg/L N03-N microcosms, respectively. Nitrate loss continued
thereafter, but at a much slower rate (0.03 and 0.02 mg/L/day) (Figure 4-11b). Peak NO, accumulation
was observed over the day 20-36 interval (Figure 4-11b) although N02 levels never exceeded about 2
mg/L, suggesting that further denitrification to gaseous products occurred. The similarity of N03" / N02~
behavior under mixed electron-acceptor and strictly denitrifying conditions indicates that denitrifying
activity was not particularly influenced by the initial presence of microaerophilic 02.
In this experiment toluene degradation was detected in the microaerophilic microcosms between days
20 and 166, i.e., after cessation of aerobic biotransformation (Figure 4-10c). This activity probably
occurred under SO/ -reducing conditions, with the Borden groundwater serving as the source of S042". A
sulfide odor was noted in these microcosms on days 68 and 166, and the solids were noticeably grayish.
Analysis of water from the microcosms sacrificed on day 166 showed that the mean SO/ concentration
was 0.18 mg/L (s.d. = 0.32) in the three microaerophilic, N03 -free microcosms, whereas S042" averaged
12.9 mg/L (s.d. = 0.54) in all other microcosms (n = 18). In addition, D.O. was below detection limits in
some of these microcosms (Figure 4-11a), which was rarely the case in N03-amended microcosms.
Toluene biotransformation under S042 -reducing conditions has been observed in other studies (Beller
etal., 1992).
The mass losses that were observed in this experiment under anaerobic NO, -reducing conditions
were larger than in any other laboratory experiment performed in this study. Nonetheless, the total BTEX
losses observed after 166 days still did not exceed 60% of the initial concentration of ca. 13 mg/L under
microaerophilic conditions (with or without N03), and 45% under strictly denitrifying conditions, m-xylene
and ethylbenzene were the only compounds that (sometimes) degraded to concentrations below detection
levels. For example, in microaerophilic / 25 mg/L N03-N microcosms, BTEX remaining on day 166
(expressed relative to the day 1 concentration) included 74% of the benzene, 2% of the toluene, 0% of the
43
-------
25 mg/L nitrate-N
1,5
1.25
1
a 0.75
0.5
0.25
0
200
0
100
Time (days)
150
microaerophilic, 25 mg/L nitrate-N
1.25
O 0.75
0.5
0.25
200
100
150
50
0
Time (days)
microaerophilic, nitrate-free
1.25
o
O
9. 0.75
x
LU
0.5
I—
~a
0.25
100
150
200
0
50
Time (days)
~ T a"; Bhb p-X K -¦ m-X o - c>xj
FIGURE 4-10 BTEX biotransformation in denitrifying, mixed electron-acceptor, and microaerophilic Borden micro-
cosms - normalized BTEX. Plotted values are the mean ± s.d. of three replicate microcosms.
44
-------
3.0
2.5
2,0
O)
13
,T3
0.5
0.0
0
20
40
60
80
100
120
140
160
180
Time (days)
¦ ¦ - sterile, 25 mg/L N03-N -a— sterile, microaeropnilic, 25 mg/L N03-N
~ 25 mg/L N03-N o microaerophilic, 25 mg/L N03-N
x microaerophilic
nitrate/nitrite
z 15
0 BM-
50
100
Time (days)
150
3o>
E
z
2 £
200
—¦—sterile, 25 mg/L N03-N (N03) —*— 25 mg/L N03-N (N03)
—•— microaerophilic, 25 mg/L NQ3-N (N03) - - -a- - - sterile. 25 mg/L N03 N (N02)
.. - A- - - 25 mg/L N03-N (N02) - o- - - microaerophilic, 25 mg/L N03-N (N02)
FIGURE 4-11 BTEX biotransformation in denitrifying, mixed electron-acceptor, and microaerophilic Borden micro-
cosms: (a) dissolved CL (b) NO," and NOa. In (b), solid symbols and lines = NO/-N; open symbols and
dashed lines = N02 -N. Plotted values are the mean ± s.d. of three replicate microcosms.
45
-------
ethylbenzene, 5% of the m-xylene, 42% of the p-xylene, and 60% of the o-xylene. In contrast, the Eglin
AFB and Park City materials (see Section 4.1.2) reduced the total BTEX content by about 60-65% in
57 days or less under denitrifying conditions, and all compounds but benzene (88-92% remaining) and
o-xylene (17% (Eglin), 48% (Park City) remaining) were biotransformed to undetectable levels.
Generally, the microcosm data obtained in these two mixed electron-acceptor experiments, and in the
others performed in this study (Experiments 4 and 6), were most consistent with the sequential utilization
of 02 and N03: rapid aerobic BTEX biotransformation was typically followed by more-limited
biotransformation under denitrifying conditions over longer time periods. These data indicated therefore
that over relatively-long (i.e., several month) incubation periods, total mass losses were enhanced in the
presence of mixed electron-acceptors relative to losses under microaerophilic and anaerobic, denitrifying
conditions. However, microaerophilic 02 did not facilitate major benzene losses. The apparent enhanced
early BTEX losses observed under mixed electron-acceptor conditions in Experiment (i) of this section
were not observed when the experiment was repeated. This lack of replicability in conjunction with other
laboratory and field evidence obtained in this study suggest that the sparing effect, if present, was not an
important process in this aquifer material.
4.1.4 Experiment 4: Effect of BTEXTMB Concentration Under Microaerophilic Oxygen/Nitrate
Conditions
After the gasoline was spilled at the field site, it became evident that the entire volume of the
treatment cells would be exposed to pure-phase gasoline and/or high dissolved-phase concentrations of
the soluble aromatic hydrocarbons. A microcosm experiment was therefore designed to evaluate the
effect of dissolved concentrations on the extent of biotransformation under conditions similar to the field
system, i.e., gasoline-saturated groundwater plus N03~ and microaerophilic Or This experiment utilized
API 91-01 gasoline (Appendix A) rather than a locally-purchased gasoline, and the larger suite of aromatic
hydrocarbons (i.e., BTEXTMB) was quantified. To simplify the experimental system, particularly the
determination of aromatic-hydrocarbon mass loss, a residual gasoline phase was not present in these
microcosms.
Design 1 microcosms were set up with pristine Borden sand from cores collected near the field site.
A 2-ml air headspace was present in each vial to create microaerophilic conditions. Three treatment
groups were used to evaluate the effect of concentration: high-BTEXTMB concentration (gasoline-
saturated groundwater), low-BTEXTMB concentration (10x dilution of gasoline-saturated groundwater),
and sterile fow-BTEXTMB-concentration controls. For each treatment group multiple replicate microcosms
were sampled on days 0, 42, and 137 for BTEXTMB, dissolved 02 (modified Winkler method), N03, and
N02". The design is summarized in Table 4-8.
This experiment was set up with multiple replicates so that differences between means of treatment
groups could be tested statistically for significance. Because substantial variability in the extent of
biotransformation is often encountered in Borden-sand microcosms, a large number of replicates was
prepared for each treatment group to give the test adequate statistical power. However, concentrations
behaved anomalously in these microcosms, increasing between days 0 and 42, which prevented a
rigorous statistical analysis of the mass-loss data. Nonetheless, the experiment still provided useful
information, as losses relative to controls were observed in the low-BTEXTMB-concentration microcosms,
and N03" data indicated that denitrifiers were inhibited in the high-BTEXTMB microcosms.
Table 4-8. Design Summary, Experiment 4
Treatment (Initial BTEXTMB Concentration)
Replicates
Sampling
Events
Active, 110 mg/L BTEXTMB (gas.-sat. water), 2-ml air headspace, 25 mg/L NO,
8 active
3
-N, N,P
Active, 11 mg/L BTEXTMB (10x dil.), 2-ml air headspace, 25 mg/L NOs'-N, N.P
8 active
3
Sterile, 11 mg/L BTEXTMB (10x dil.), 2-ml air headspace, 25 mg/L NCK-N, N.P
8 sterile
3
46
-------
Nitrate / Nitrite
120
o>
80
40
50 100
Time (days)
1.5
150
o>
E
O
z
0.5
Normalized BTEXTMB
Dissolved Oxygen
50 100
Time.(days)
W 0.4
50 100
Time (days)
- High BTEXTMB ------- Low BTEXTM3
• & - - Control
FIGURE 4-12 Nitrate, NO D.O., and normalized total BTEXTMB concentrations in Experiment 4 microcosms. High
BTEXTMB: gasoline-saturated (110 mg/L); Low BTEXTMB and Control: 10x dilution (11 mg/L). The
detection limit value (0.2 mg/L) plotted for not-detected N02 samples (solid squares). Plotted values are
the mean ± s.d. of three replicate microcosms.
Table 4-9. Percent of Individual Aromatic Hydrocarbons Remaining in Active, Low-BTEXTMB Concentration
Microcosms Relative to Sterile Controls. Microcosms were Amended with an Air Headspace and
NOj, and Contained Pristine Aquifer Material.
Day 137
Compound % Remaining (s.d.)'
Benzene
86
(14)
Toluene
10
(12)
Ethyl benzene
28
(56)
fffl-p-Xylene
24
(10)
o-Xyiene
76
(15)
1,3,5-Trimethylbenzere
72
(38)
1,2,4 Trimethylbenzene
7
(10)
1,2,3-Trimethylbenzene
86
(52)
Naphthalene
43
(59)
(c/c.),
' Percent remaining calculated from ^=^*100 where C0
is the concentration on day 0. (c/c_)
Standard deviation of percent remaining calculated using Equation 5-2.
47
-------
Results and Discussion. BTEXTMB losses were not evident over 137 days of incubation in the
high-BTEXTMB or sterile controls microcosms (Figure 4-12). On the other hand, mass losses were
observed in the low BTEXTMB microcosms (Figure 4-12). In terms of individual compounds, toluene,
ethylbenzene, m+p xylenes, 1,2,4-trimethylbenzene, and naphthalene concentrations all declined relative
to controls by day 137 (Table 4-9). The remaining aromatic hydrocarbons appeared recalcitrant; the
slightly lower concentrations may be attributed to aerobic biotransformation, to greater sorption in the
active microcosms relative to the autoclaved controls, or a combination of the two processes.
By day 42, 02 was depleted to a threshold concentration in the aqueous phases of both active
treatments. Unfortunately, the 02 concentrations in the control microcosms appeared anomalous, possibly
from interferences between the Winkler reagents and the sodium azide. Subsequent controls analyzed
with the dissolved 02 meter indicated that the initial dissolved 02 concentration in microcosms with a 2-ml
air headspace was about 4.5 mg/L, declining asymptotically to below 1 mg/L over a four-month period
(see Section 4.2.1; Figure 4-15). Assuming similar initial concentrations in the active microcosms,
roughly half of the dissolved 02 was consumed within the first several hours of the experiment
(Figure 4-12). As in other laboratory experiments performed for this study, low concentrations
(0.1-0.5 mg/L) of dissolved 02 appeared to persist over the incubation period in the active microcosms.
Nitrate aiso declined in the low-BTEXTMB microcosms relative to the control and high-BTEXTMB
microcosms, but accumulation of detectable concentrations of N02" did not occur (Figure 4-12).
Mineralization reaction stoichiometries suggest that the mass of Oz in the low-BTEXTMB microcosms
was sufficient to account for the observed mass losses. Assuming total BTEXTMB was aerobically
mineralized to C02 and there was no assimilation of C by microbial cells, the mineralization reaction is
ea2H9a + 106.5O2-*82CO2+49H2O 4-2
Based on Equation 4-2, the mass of 02 in the headspace (0.6 mg 0£) would be sufficient to oxidize
0.2 mg BTEXTMB, which is greater than the observed total BTEXTMB mass loss of approximately
0.17 mg on day 137. The continued utilization of N03 after CX depletion in these microcosms indicates
that some of this mass loss may have occurred under denitrifying conditions, although it is conceivable
that natural organic matter or other dissolved constituents in the gasoline-contacted water served as the
carbon source. In the high BTEXTMB microcosms, the mass of 02 was quite low relative to the mass of
dissolved organics. Therefore, although small aerobic losses of aromatics may have occurred, depletion
relative to the day 0 sampling event could not be observed because concentrations increased between
days 0 and 42. Consequently, the effect of microaerophilic 02 was unclear. On the other hand, in
comparison to low-BTEXTMB microcosms, the lack of NO depletion after 02 consumption clearly
indicated that denitrifiers were not active in the high BTEXTMB microcosms.
Despite the unexplained BTEXTMB concentration trends, this experiment provided information useful
for interpreting in situ behavior. Aromatic-hydrocarbon and N03 utilization trends indicated that
biotransformation of labile compounds such as toluene and uptake of NO, were negligible in the presence
of high dissolved phase BTEXTMB concentrations under mixed microaerophilic / N03" conditions. This
experiment indicated therefore that near the source area, aqueous concentrations were high enough to be
inhibitory to an unacclimatized denitrifying population in the Borden aquifer (i.e., N0:J-based
bioremediation would not be effective in this aquifer near the source area). However, as discussed further
in Chapter 5, N03 reduction was observed in the field, suggesting that an acclimatized population did
develop with exposure to gasoline hydrocarbons. Nitrate depletion was also observed in the laboratory
using contaminated core material extracted from the Nitrate Cell (Section 4.2).
4.1.5 Experiment 5: Effect of Oxygen Concentration Under High BTEXTMB Concentration
Conditions
Because the small amount of dissolved O, added to the Experiment 4 microcosms appeared to be
utilized rapidly, biotransformation in the high BlEXTMB microcosms may have been 02 limited. To
address this issue, an additional experiment was performed to determine if the aromatic hydrocarbons
would biotransform at near-source concentrations under fully-aerobic conditions, in both the presence and
absence of N03.
48
-------
Design 1 microcosms were set up with pristine Borden sand. All microcosms in this experiment were
prepared with groundwater saturated with API 91-01 gasoline. Each microcosm contained a 4-ml
headspace purged with pure 02 rather than air to obtain a "high" initial dissolved 02 concentration (and
total 02 mass) to drive.aerobic reactions. Three treatment groups were used to evaluate
biotransformation of high BTEXTMB concentrations under aerobic conditions: aerobic / NO„, aerobic
only, and aerobic sterile controls. Microcosms were incubated at room temperature in a laboratory
cupboard. Five replicates from each treatment group were sampled on days 0, 42, 73, and 163.
Groundwater was analyzed for BTEXTMB, dissolved 0„ (modified Winkler method), and N03 /N02", when
applicable. The experimental design is summarized in Table 4-10.
Results and Discussion. In terms of mean concentrations, total BTEXTMB losses were observed
in both active treatment groups (aerobic / N03" and aerobic only) relative to the sterile controls
(Figure 4-13). Mass losses were slightly more extensive in the mixed electron-acceptor microcosms.
However, concentrations in individual replicates were extremely variable, particularly in the active
microcosms (Figure 4-14). Figure 4-13 also shows that N03 utilization and N02~ production were low, and
that dissolved 02 concentrations declined to about 8 mg/L by day 42 and remained steady for the
remainder of the experiment. In general, there was no preferential utilization of individual compounds
(data not shown); in microcosms that experienced mass loss, ail compounds had roughly the same
proportional loss relative to initial concentrations.
The extremely large variability in the extent of mass loss in the active microcosms was unexpected.
The data suggest either patchy microbial activity and/or biomass, or abiotic losses (e.g., leakage) in
individual replicates, or some combination of the two. The lower variability in control microcosms
(Figure 4-14) supports the former explanation. It is difficult, however, to fully explain these results without
invoking either microcosm leakage or an unknown oxidation reaction. If the observed BTEXTMB mass
loss resulted from aerobic oxidation, then from the stoichiometry of the mineralization reaction given in
Equation 4-2 (106.5 moles 02 per mole BTEXTMB), dissolved 02 should have been completely consumed
in microcosms with extensive mass loss. Complete utilization of 02 was not observed (Figure 4-13);
during this experiment depleted dissolved 02 was observed in only two microcosms. It is conceivable that
the added inorganic nutrients were insufficient to support full consumption of added 02, but corresponding
BTEXTMB mass losses should have been low as well. In fact, if C02 was the major oxidation product,
then from Equation 4-2 the Oa supplied to the microcosms was sufficient to mineralize only 14 % of the
total BTEXTMB. Therefore, unless 02 was leaking into microcosms, or mineralization was incomplete
(which would not be expected with excess 02), aerobic biotransformation cannot adequately explain these
results. It is also possible that a currently undefined oxidation reaction was responsible for the mass loss.
In some microcosms, an orange precipitate was observed on the surface of the aquifer sediment, but
there was no clear correspondence to low BTEXTMB concentrations. The formation of a precipitate was
unusual, but the reaction may have been unrelated to the observed aromatic hydrocarbon depletion
(Millette et al., 1998).
On the final day of the experiment, microcosms were drained of fluid and put aside for additional work
designed to determine whether biotransformation was responsible for the observed aromatic hydrocarbon
losses. An aliquot of aquifer material was removed from nine microcosms for enumeration, and then all
active microcosms were reamended with 25-ml gasoline-saturated groundwater, N03" where required,
nutrients, and sealed with mininert™ valves. Control microcosms received amendments and additional
sodium azide solution. Aquifer material was removed from microcosms and enumerated for aerobic
Table 4-10. Design Summary, Experiment 5
Treatment (Electron-Acceptor Reaime)
Replicates
SamDlina Events
Active, 4-ml pureCfe headspace, 70 mg/L BTEXTMB (gas. sat.), 25 mg/L
5 active
4
NO3--N, N,P
Active, 4-ml pure Op headspace, 70 mg/L BTEXTMB (gas. sat.), N,P
5 active
4
Sterile, 4-ml pureOs headspace, 70 mg/L BTEXTMB (gas. sat.), N,P
5 sterile
4
49
-------
Nitrate / Nitrite
Dissolved Oxygen
150
ra 100
E
6 50
z
0
ii—----J —.....
¦t-i
45
90
BTEXTMB
-D
45 90 135 180
45 90 135 180
Time (days)
-Oxygen Nitrate/Oxygen - - & - - Control
FIGURE 4-13 Nitrate, NO,, D.O., and total BTEXTMB concentrations in Experiment 5 microcosms. The detection limit
value (0.2 mg/L) plotted for not-detected N02 samples (solid squares). Plotted values are mean ± s.d.
of replicate microcosms. For clarity BTEXTMB error bars not plotted.
Nitrate/Oxygen
Oxygen
O)
E
75
" 50
o
O
99 or
X
til
t- 0
CO
o
_~
45
~
~
~
~
~
90
135 180
o 50
cn
2 25
f—
X
LU
co o
75
O)
£ 50
o
0
1 25
£
A
Control
6
A
45 90 135 180
A
A
X
LU
CO
0
45
90
135
180
Time (days)
FIGURE 4-14 BTEXTMB concentrations in individual replicates for each of the treatment groups in Experiment 5. Lines
connect mean values for each sampling event.
50
-------
heterotrophs arid benzene-toluene degraders using methods described in Section 3.5.1. Reamended
microcosms were then sampled periodically over the following 23 days for headspace BTEX. Overall,
these headspace BTEX results agreed fairly well with concentration data obtained on day 163. The
individual microcosms with low aromatic-hydrocarbon concentrations on day 163 underwent mass loss
relative to controls during the 23 day post-experiment incubation period. Similarly, microcosms with high
concentrations on day 163 did not lose mass relative to controls. The enumeration results for the control
and Og-only microcosms also agreed quite well with the mass-loss data; low numbers were present in the
microcosms with high BTEXTMB concentrations on day 163, and little mass loss during the post-
experiment incubation, and vice versa. These data are summarized in Table 4-11. This may be evidence
that the experimental conditions (i.e., concentrations of parent compounds or metabolites) were toxic to
the microbial community in some replicates, but not in others. Enumeration results were more variable for
the N03 /02 treatment group.
The relatively rapid uptake after reamendment suggested that microbial activity was responsible for
the low aromatic-hydrocarbon concentrations observed on day 163. This in turn suggested that there was
potential for aerobic biotransformation of high concentrations of dissolved-phase BTEXTMB in pristine
Borden aquifer material. Because of conflicting dissolved-O., data, however, and variability in the
BTEXTMB concentrations, the possibility of experimental artifact (i.e., abiotic losses) cannot be ruled out.
If microbial activity was responsible for the observed mass loss in these microcosms, the distribution of
populations tolerant of high concentrations of aromatic hydrocarbons was patchy in this pristine material.
This is consistent with the results of Butler et al. (1997) who found that microbial communities were very
localized in the Borden aquifer. Distinct populations with varying metabolic capabilities were observed
within a <5.5 m2 area. In contrast, the large and consistent O demand observed in gasoline-
contaminated aquifer material extracted from the Nitrate Cell (Section 4.2) indicated that an acclimated
population had developed in and immediately below the gasoline source area.
4.2 Microcosm Experiments: Gasoline-Contaminated Borden Sand
4.2.1 Experiment 6: Extent of Biotransformation Under Various Substrate and Mixed Electron-
Acceptor Conditions
The purpose of this experiment was to investigate the effects of in situ gasoline exposure and the
presence of other gasoline hydrocarbons on the ability of the indigenous bacteria to biotransform the
aromatic hydrocarbons. When cores were collected in July, 1997 from the Nitrate Cell, the aquifer
material had been exposed to gasoline (dissolved- and/or residual-phase) for 19 months, and to N03 for
14 months.
Design 1 microcosms were set up with core material extracted from the Nitrate Cell. Core from a
depth interval of approximately 80 to 180 cm bgs was used in the experiment. Before dispensing to
microcosms, most of the residual gasoline phase was removed by saturating the aquifer material with
sterile water, gently stirring the resulting slurry, and then draining the liquid phase. This was done to
remove as much of the existing aqueous aromatics and residual gasoline as possible, with minimal
disturbance to the microbial population. Analysis of the washed material indicated that this procedure was
successful in lowering aqueous concentrations of benzene, toluene, and ethylbenzene, but mg/L
concentrations of the less-soluble aromatics remained. This indicates that the washing procedure did not
completely remove the residual gasoline phase, which contained substantial concentrations of these
constituents at the end of the field experiment. With the exception of one low-substrate-concentration
treatment group, al! microcosms were set up with gasoline-saturated groundwater to reflect field
conditions. Because a residual phase was still present, initial concentrations of the xylene isomers,
trimethylbenzene isomers, and naphthalene were near gasoline-saturated concentrations in the low-
concentration (10x dilution) microcosms; as a result the measured day 1 total BTEXTMB concentration
was 19.5 mg/L (see Figure 4-16c).
Five treatment groups were used to isolate the effects of electron-acceptor regime and BTEXTMB
concentration in contaminated aquifer material: microaerophilic / N03\ aerobic / N03", microaerophilic
/N03- plus low-BTEXTMB (10x dilution of gasoline-saturated water), unamended (no 02 / N03). and a
microaerophilic / NO,; sterile control. All microcosms except in the low-BTEXTMB group were prepared
with gasoline-saturated water and 02 was added to headspaces as described previously. The
51
-------
Table 4-11. Microbial Enumerations and Hydrocarbon Degrading Activity for Nine Replicates Selected for
Reamendment after Day 163.
Replicate
Day 163 Total
BTEXTMB
Concentration
(mg/L)
Heterotrophic
Plate Count
(CFU/g wet wt.)
(s.d.)
Benzene-Toluene
Degraders
(MPM'g wet wt.)
Hydrocarbon-
Degrading Activity
After Reamendment
Control A
65.0
n.d.1
n.d.
(-)
Control B
32.0
5.7 x10s
(3.7 xlO2)
n.d.
(-)
Control E
56.4
n.d.
ad.
(-)
O2 A
0.0
9.0 x107
(1.0 x107)
2.0 x10s
(+)
O2 C
47.9
4.0 x10s
(4.8x10")
23
(-)
O2 E
2.6
2.7x10®
(7.3 x104)
4.3 X102
(+)
O2/NO3" A
41.6
1.3 x 106
(9.5 x104)
2.1 x 1Cf
(+)
Or / NO. C
19.3
1.5x10®
(1.6 x 10s)
1.5x103
(+)
O2/NO3- E
3.6
4.3 x107
(2.6 x 10B)
43
{+)
1 n.d. - none detected. Determination of hydrocarbon degrading activity after reamendment based on observed
compound losses.
Table 4-12. Design Summary, Experiment 6
Treatment (Electron Acceptor Regime, BTEXTMB Concentration)
Replicates
Expected
Sampling
Events
Active, 100 mg/L BTEXTMB {gas. sat.), 2-rn! air headspace, 25 mg/L NO; - 3
N, N,P
Active, 100 mg/L BTEXTMB (gas. sat.), 4-ml pure Os headspace, 25 mg/L 3
NOa-N, N,P
Active, 10 mg/L BTEXTMB (10x dil.), 2-ml air headspace, 25 mg/L NCV-N, 3
N,P
Active, 100 mg/L BTEXTMB gas, sat.), N,P 3
Sterile, 100 mg/L BTEXTMB (gas. sat.), 2-ml air headspace, 25 mg/L NO/ - 3 sterile
N, N,P
Aqueous Sterile (no aquifer material), 90 mg/L BTEXTMB (neat), 4-ml pure 3 sterile
O2 headspace
Aqueous Sterile (no aquifer material), 90 mg/L BTEXTMB (neat), 2-ml air 3 sterile
headspace
4
4
4
4
4
4
4
52
-------
microaerophilic (low 02) microcosms were incubated in the anaerobic chamber, and the aerobic (high-02)
microcosms in a laboratory cupboard. Groundwater was analyzed for BTEXTMB, dissolved 02 (with the
02 meter), NO,, and N02. Extra replicates of each treatment group were also prepared for a
denitrification assay (acetylene block). These microcosms were analyzed for N20 in addition to the other
parameters. The design is summarized in Table 4-12.
In response to rapid dissolved 02 and NO," utilization in the aerobic / N03" microcosms, several
modifications were made to the experimental design. On day 82, N03 was respiked into the aerobic/N03
microcosms to determine if rapid N0„* utilization would continue under anaerobic conditions. Microcosms
were amended and subsequently incubated inside the anaerobic chamber to maintain anoxic conditions.
At the same time, duplicate microcosms from all treatments except the unamended group were spiked
with acetylene and sampled after 15 days of incubation. Finally, on day 154, the headspaces of all
remaining microcosms except the unamended group were flushed with pure 02 to determine if
biotransformation of aromatic hydrocarbons would be stimulated under aerobic conditions. To
accommodate these changes, microcosms were sacrificed and sampled more frequently over a 173-day
incubation period. Microcosms were sampled in triplicate for the first three sampling events, and in
duplicate or singly thereafter.
Aqueous (no aquifer material) control microcosms designed to investigate abiotic losses of dissolved
0? and aromatic hydrocarbons were also prepared (Section 3.2.1) and incubated with other Experiment 6
microcosms (Table 4-12). Microcosms were sampled and analyzed for aromatic hydrocarbons and
dissolved O, (using the 02 meter) periodically between days 0 and 159.
Results and Discussion. The aromatic hydrocarbons in the unamended control microcosms did not
decline relative to the sterile controls; other microbial populations (e.g., Fe or S042 reducers) were
therefore not degrading aromatic-hydrocarbons at a detectable rate. For brevity, unamended microcosm
results are not shown or discussed further.
The aqueous sterile control microcosms were sampled on day 0, several hours after preparation;
measured concentrations showed that the 2-ml air and 4-ml pure Oa headspaces provided 4.9 mg/L and
26.2 mg/L initial dissolved 02, respectively. Diffusive losses of dissolved 0? occurred during storage both
inside (microaerophilic controls) and outside (aerobic controls) the anaerobic chamber, as microcosms
equilibrated with the external atmosphere (Figure 4-15), The rate of loss was, however, much lower than
in active microcosms. The agreement between the microaerophilic controls with and without aquifer
1.5
1.0
8
o
d
O 0.5
0.0
D.O.
t : \o - -
* * *© * ,
" S - »
P
n
... m
• a
n a
~
~
~
40
80
Time (days)
120
160
1.5
O
O 1.0
o
m
UJ 0.5
0.0
BTEXTMB
.--a-'fl'-c
40
80
Time (days)
120
a Microaerophilic Aqueous
~ Microaerophilic Aqueous/Sediment
o Aerobic Aqueous
160
FIGURE 4-15 Normalized D.O. and total BTEXTMB concentrations in sterile microcosms with and without gasoline-
contaminated aquifer material from the Nitrate Celt (Experiment 6). Lines connect single values or
means of duplicate and triplicate replicates.
53
-------
material indicated that the abiotic 02 demand of the contaminated aquifer material was quite low. This is
not surprising because the redox potential in the Nitrate Cell was buffered by the continuous presence of
N03, and consequently, significant quantities of reduced inorganic species such as Fe and Mn that are
capable of reacting abiotically with 02 were not produced. In contrast to 02, diffusive loss of dissolved
BTEXTMB in both aqueous and conventional (i.e., containing aquifer material) sterile controls was low
(Figure 4-15).
Initial BTEXTMB losses relative to controls were observed between preparation and the first
sampling event (24 hours) for all active microcosms, and the extent of consumption was roughly
proportional to the mass of 02 in the microcosm (Figures 4-16a, 4-16b, and 4-16c). For example, the
largest decline during early time occurred in the microcosms amended with pure 02. In the high-
BTEXTMB-concentration microcosms amended with microaerophilic 02 and NO, , which is the best
analogy to the Nitrate Cell, the close agreement between active and sterile-control BTEXTMB
concentrations shows that the microaerophilic 02 had only a minor effect on mass loss (Figure 4-16b).
This resuit is consistent with field observations (Chapter 5).
Once O was depleted, biotransformation losses of the aromatic hydrocarbons were minor.
Percentage losses of individual compounds for the two microaerophilic / NO,- treatment groups (high- and
low-BTEXTMB) are shown on Table 4-13. It should be noted that because the microaerophilic 02 was
essentially consumed by day 1, and losses were calculated relative to day 1 concentrations, depletion of
aromatics between days 1 and 145 occurred primarily under anaerobic conditions, possibly at the expense
of NO,. Under 02-depleted conditions, only toluene, and possibly ethylbenzene declined in low-
BTEXTMB microcosms. After 02 was added on day 154, losses of benzene, toluene, ethylbenzene,
m+p-xylenes, 1,2,4-trimethylbenzene, and to a lesser extent, o-xylene, were observed in the microcosms
amended with a 10x dilution of gasoline-saturated water. In high-BTEXTMB microcosms, on the other
hand, only benzene, toluene, and ethylbenzene concentrations declined after 02 reamendment.
1,3,5-trimethylbenzene, 1,2,3-trimethylbenzene, and naphthalene appeared recalcitrant under these
incubation conditions. It is unclear why the observed trimethylbenzene isomers and naphthalene ratios
are higher on day 173 relative to day 145. One plausible explanation is that the concentrations in the
sterile controls declined more than in the active treatments when the microcosms were opened on
day 154 to replenish headspaces with 02. This most likely resulted from the lack of a residual gasoline
phase in the controls, which acted as a reservoir for the relatively insoluble aromatics in the active
microcosms. Consequently, it is possible that although the relatively insoluble aromatics appeared
recalcitrant, the presence of a residual phase may have obscured minor removal from the aqueous
phase.
In the aerobic / NOr; microcosms, the initial dissolved 02 (ca. 26 mg/L assumed from aqueous sterile
controls) was rapidly consumed (Figure 4-16a). It should be noted that the complete mass of O in these
microcosms was probably not consumed within 24 hours; after the initial rapid consumption of the 02
dissolved in the aqueous phase, the rate of consumption may have been controlled by diffusion across the
water-headspace interface. In both of the microaerophilic treatment groups, the dissolved 02
concentration also dropped rapidly, from an initial concentration of 4.9 mg/L (assumed from sterile control
microcosms) to a threshold concentration within 24 hours. When microcosms were reamended with pure
02 on day 154, rapid BTEXTMB and dissolved 02 depletion were again observed in all active microcosms
(Figures 4-16a, 4-16b, and 4-16c). Overall, these responses showed that an acclimated aerobic
population was capable of producing a large 02 demand in the contaminated aquifer material relative to
the pristine material, and that dissolved 0? was required for biotransformation of the aromatic
hydrocarbons (with possible exceptions of toluene and ethylbenzene).
The utilization of NO.; varied depending on the treatment group. The one similarity was that all three
active groups lost some Isl03 relative to the sterile-control group during the first 24 hours of the
experiment when dissolved 02 was present. The greatest losses occurred in the aerobic microcosms,
where NO, was completely consumed by day 14 (Figure 4-16a). The rapid consumption of NO, in these
microcosms was surprising because such rapid losses had not been observed in the Nitrate Cell. One
possible explanation is that the NO„ was being utilized as an assimilatory nitrogen source during aerobic
biotransformation; because the mass of 0? and available hydrocarbons in these microcosms was large,
the mass of NH4-N may have been insufficient to meet the assimilatory N demand. As a consequence,
54
-------
Nitrate / Nitrite
o
Z
NO,-Ad
Day82>V
30
Dissolved Oxygen
20
E.
o
10
0; Added
Qav 154 -
•* ,
o
45
90
135 180
Nitrate / Nitrite
a a .
90 - 135 180
Dissolved Oxygen
Os Added
Day 154
180
(c)
Nitrate / Nitrite
140
"3s
135 180
Dissolved Oxygen
Q> Added
toy 154
45 90 135 180
120
|? 80
CO
2
mS «o
i—
ao
BTEXTMB
BTEXTMB
<5 o
o 8 8
45 90 135 180
"Time (days)
~ Cbnfroi o Aarobic X Aerobic Acjgcjs Control
45 90 135 180
Time (days)
~ Control a Mcroaerophilie
120
BTEXTMB
c>-» $ ft $ 6
45 90 135 180
Time (days)
o Control o Mcroaerophifc, 10* dil
FIGURE 4-16 Nitrate, N02, D.O., and total BTEXTMB concentrations in active and sterile control microcosms with gasoline-contaminated aquifer
material from the Nitrate Cell (Experiment 6) (a) Aerobic (high 02) treatment (b) Microaerophilic (low-02) treatment (c) Microaerophilic,
10x BTEXTMB dilution treatment. Oxygen added to all microcosms on day 154 except high-02 aqueous control plotted on (a). Nitrite
not detected in sterile controls; for clarity, data not shown on plots. Nitrite in active microcosms plotted as solid symbols. Lines con-
nect single values or means of duplicate and triplicate replicates.
-------
Table 4-13. Percent of Individual Aromatic Hydrocarbons Remaining in Active, Low- and High-BTEXTMB
Concentration Microcosms Relative to Sterile Controls. Microcosms were Amended Initially with an
Air Headspace and NO', and Contained Gasoline-Contaminated Aquifer Material from the Nitrate
Cell.
% Remaining Relative to Sterile Controls1
Compound
10x Dilution of Gasoline-Saturated
Water
(2-ml Air Headspace)
Gasoline-Saturated Water
(2-ml Air Headspace)
Day 145
(Oa depleted)
Day 173
(19 days after pure-
O2 addition)
Day 145
(O2 depleted )
Day 173
(19 days after pure-
O2 addition)
Benzene
83
28
97
80
Toluene
50
0.7
101
55
Ethyl benzene
77
2
106
21
m+p-xylene
101
38
98
109
o-xylene
104
83
96
105
1,3,5-T nmelhyl benzene
106
130
90
127
1,2,4-Trimethyl benzene
105
51
89
124
1,2,3-Trimethyl benzene
104
122
87
120
Naphthalene
98
110
84
125
(c/c,,)
1 1 ^1 ftA
1 Percent remaining for a given sampling event calculated from \ where Ca is mean concentration on day 1.
^ ' 0 ^ Cuniiiii
Table 4-14, Dissolved Nitrous Oxide Concentrations in Selected Microcosms Containing Contaminated Aquifer
Material from the Nitrate Cell. Nitrate was Present in All Microcosms when Analyzed. Acetylene was
Added on Day 82, 15 Days Prior to Analysis.
Individual
Acetylene
Dissolved NzO
Microcosm
Added
(mg/L)
Aerobic, gas.-sat.
(+)
5.45
Aerobic, gas.-sat.
(+)
9.64
Aerobic, gas.-sat.
(-)
<0.45
Microaerophilic, gas.-sat.
(+)
<0.45
Microaerophilic, gas.-sat.
<+)
<0.45
Microaerophilic, gas.-sat
(-)
<0.45
Microaerophilic, lOx dil,
M
0.39
Microaerophilic, 10x dil.
(+)
1.00
Microaerophilic, 10x dil.
(-)
<0.45
Sterile Control
M
<0.45
Sterile Control
W
<0.45
56
-------
inorganic nutrients did not limit the consumption of this large mass of 02. When N03 was replenished on
day 82, additional N03 losses were observed under anaerobic conditions, but the rate was lower
(Figure 4-16a). The production of NaO in acetylene-amended microcosms between days 82 and 97
suggested that denitrification was occurring in these microcosms under anaerobic conditions (Table 4-14).
Nitrate utilization was also observed in the microaerophilic / N03\ high-BTEXTMB microcosms
(Figure 4-16b), the condition most similar to the field treatment cell. A comparison of these results with
those of Experiments 4 and 5 clearly indicates that the extent of N03 utilization in the presence of high
aqueous concentrations of BTEXTMB increased after prolonged exposure to contamination. Although
NO.; declined under Oz-depleted conditions, however, utilization of the substrates that are typically labile
under denitrifying conditions (toluene or ethylbenzene) was not evident. Assuming toluene was
mineralized in a denitrification reaction, and there was no assimilation of C and N by microbial cells, mass
loss is governed by
C7Hs + 7.2 H" + 7.2 NO, -> 3.6 N2 + C02 + 7.6 H20 4-3
Consequently, using an initial toluene concentration of 40 mg/L (2 mg initial mass), consumption of
25 mg/L N03-N (5 mg NO initial mass) would have produced an observable decrease in toluene mass of
about 50%. Therefore, although denitrifying activity in gasoline-contaminated aquifer material was
apparently not inhibited, the labile aromatic hydrocarbons were either recalcitrant or utilized at a very low
rate. This may reflect a preference for other organic substrates such as other gasoline hydrocarbons, as
observed in other studies with hydrocarbon-contaminated aquifer material (Hutchins et al., 1991a).
Alternatively, it is possible that N03 was used to satisfy an N demand for an undefined, anaerobic
reaction (e.g., fermentation reactions).
The specific denitrifying pathway responsible for the NO. utilization in these microcosms is unclear
because although N02 accumulation was evident in the microaerophilic / N03', high BTEXTMB
microcosms after day 82, NzO was not present in the duplicate microcosms amended with acetylene
(Table 4-14). These N03 data yielded a zero-order (linear) depletion rate of 0.43 mg/L/d, which is within
the range of rates measured in the field treatment cell (see Section 5.1.2). In contrast, in the
microaerophilic / N03", low-BTEXTMB group, there was less NO,- utilization, and no observable NO,/
accumulation (Figure 4-16c), but there were minor losses or toluene and ethylbenzene under 02-depleted
conditions (Table 4-13). The reasons for the differences in the extent of N03 utilization in these
microaerophilic microcosms were not evident from the data collected in this experiment.
4.3 Microbial Characterization Results
Microbial enumerations and activity measurements were used as additional indirect lines of evidence
for in situ biotransformation of gasoline hydrocarbons. Patterns of microbial activity and numbers, in
conjunction with other lines of evidence, have been used in other studies to document biotransformation in
contaminated aquifers (Harvey et al., 1984; Song and Bartha, 1990; Madsen et al., 1991). In this study,
we compared numbers and activity in pristine Borden aquifer material with material extracted from the
treatment cells (19 month exposure to dissolved- and /or residual-gasoline). A detailed characterization of
the microbial populations of the Borden aquifer, including changes that occurred in response to
hydrocarbon contamination, is provided by Butler et al. (1997).
4.3.1 Pristine Aquifer Material
Enumerations. Several cores collected near the field site for use in laboratory experiments were
enumerated to determine the numbers of aerobic heterotrophs and denitrifiers in background (pristine)
Borden aquifer material (Table 4-15). In all cases, core material from the shallow saturated zone (depths
less than two meters below the water table) was used. The majority of the cores were collected in 1993
or 1994, near the beginning of this study, and enumerated when used for an experiment. Core 2-2 was
collected in June, 1997, and enumerated in September, 1997. All cores were collected in the northeast
corner of the sand pit, within about 100 m of the treatment cells. Results indicate that the age of the core
had no consistent effect on the numbers of culturable organisms or the capacity to biotransform TEX
under denitrifying conditions (Table 4-15).
57
-------
Table 4-15. Microbial Enumerations of Borden Cores
Core I.D.
Laboratory
Experiment
Heterotrophic
Plate Counts
(CFU/g dry wt)
Denitrifiers
(MPN/g dry
wt)
Approx.
Core Age
(months)1
TEX
Degradation
with NOs"
#7-93
Microcosm 1
6.2 x 104
1.7x10"
6
(+)
#5-93
Microcosm 2
1.1 x 10s
2.8 X102
8
(+) (1 of 3)
#4-93
ETS Activity
1.2 x105
2.5 x103
11
(¦)
#2-94
No rie
1.25x10®
2.7 x105
1
(+)
#6-93
Microcosm 3(i)
n.t.2
n.t.
8
(f
#2-93
None
5.8 x 105
1.2x10®
13
(+)
#1 -94/#3-94
Microcosm 3(ii)
1.1 x 10®
4.9 X104
2
<+)
Mcrocosm 6
' months of storage at 4°C before use.
2 at. not tested,
3 a few replicates (+) after ca. 1 year incubation
Counts of viable, aerobic heterotrophs in pristine aquifer material varied by about two orders of
magnitude {Table 4-15). HPCs ranged from 6x10" CFU/g (colony forming units per gram) to 1 x 10s
CFU/g, Counts from these cores were consistent with other studies of pristine Borden aquifer material
(Butler et al., 1997; Barbara et al., 1994). Using a much larger set of data, Barbara et al. (1994) found
that the numbers of aerobic, culturable microorganisms in the upper 2 m of the aquifer varied by more
than six orders of magnitude, and were correlated strongly with depth and in situ dissolved-02
concentration. Denitrifier numbers were more variable than aerobic heterotrophs in these cores, with a
range from 3 x 102 MPN/g (most probable number per gram) to 1 x 10s MPN/g. As noted by Butler et al.
(1997), the magnitudes of the aerobic heterotroph and denitrifier counts are similar enough to suggest that
a substantial fraction of the population in pristine material has the capacity to denitrify. Overall, the
numbers of culturable microorganisms in Borden aquifer material were consistent with numbers obtained
in other shallow, sandy aquifers (e.g., Beloin et al., 1988).
Activity Assays. The potential for gasoline-saturated groundwater to inhibit the activity of the Borden
microbiota was assessed using the ETS assay, which is based on the reduction of INT to INT-formazan
by microbial activity (Section 3.4.2). The solid:liquid ratio in these systems (10 g:3 ml) differed from that in
the microcosm systems (20 g:50~55 ml). Based on a gasoline-saturated water concentration of 130 mg/L
total BTEX, the BTEX content of the liquid phase in these INT systems was approximately 43 mg/L
(gasoline-saturated water) or 4.3 mg/L (10x dilution of gasoline-saturated water). The experimental
design is summarized in Table 4-16. The endogenous activity of pristine aquifer material is too low to be
detected by the ETS assay (Butler et a!., 1997), so the effect of added gasoline-saturated groundwater
alone, or in the presence of R2A broth, as a nutrient source, was investigated. It should be noted that this
experiment was conducted at natural aquifer temperature (10°C), rather than at room temperature, as was
necessary for most of the other laboratory experiments (Section 3.2.1).
The presence of the 43 mg/L BTEX water slowed, but did not prevent INT-formazan accumulation in
aerobically-incubated, R2A-containing, Borden aquifer material (Figure 4-17). On the other hand, the
more dilute contaminated water had little effect on INT-formazan accumulation, and may have even
stimulated ETS activity in the presence of R2A to a small degree. No ETS activity was detectable over
37 days in the presence of 43 mg/L BTEX alone, but there was a small amount of activity up to day 17 in
the 4.3 mg/L BTEX-amended material. The results in the presence of R2A indicate that the higher level of
water-soluble gasoline components caused some inhibitory effect but did not severely impair aerobic
microbial activity. On this basis, one might expect an eventual onset of metabolic activity in the vials with
43 mg/L BTEX-containing water but no R2A, but this was not observed (Figure 4-17). Such activity may
have been delayed (i.e., lag >37 days), or may have been largely suppressed because of a lack of
58
-------
Table 4-16. Experimental Design: ETS Activity in Pristine Aquifer Material
Treatment1
Hypovlal Contents
Blank
Borden aquifer solids, water
Active
Borden aquifer sol
ds, INT, water
R2A
Borden aquifer sol
ds, INT, R2A, water
Gas
Borden aquifer sol
ds, INT, 1 ml gasoline-saturated water, water
Gas + R2A
Borden aquifer sol
ds, INT, 1 ml gasoline-saturated water, R2A
1/10 gas
Borden aquifer sol
ds, INT, 0.1 ml gasoline-saturated water, water
1/10 gas + R2A
Borden aquifer sol
ds, INT, 0.1 ml gasoline-saturated water, R2A
1 Each treatment was prepared in duplicate. Unleaded gasoline was obtained from a local source.
10
"D
- R
p? 6
c
CO
N
re 4
E *
0
M—
1
z 2
0
10 15 20 25
Time (days)
30
35
-endogenous
- R2A+43 mg/L
-e— R2A
h—4.3 mg/L
40
-x— 43 mg/L
-A— R2A+4.3mg/L
FIGURE 4-17 Effect of gasoline-contaminated groundwater on ETS activity in pristine Borden material. Plotted values
are the mean ± s.d. of two replicates.
59
-------
inorganic nutrients (N, P) in the Borden sand. It is also possible that the 43 mg/L BTEX concentration
was more stressful to the indigenous population when R2A medium was not available. Substrates in the
R2A (e.g., glucose, casamino acids, yeast extract, peptone, starch) are easily metabolized, and protective
components (e.g., pyruvate, peptone, phosphate buffer) are present as well. This could be considerably
more supportive than the fairly high levels of potentially-toxic BTEX substrates. We have readily
developed BTEX-degrading enrichment cultures capable of metabolizing >50 mg/L BTEX from Borden
aquifer material, but this was done in a stepwise manner, which allowed for adaptation to progressively
higher BTEX concentrations.
4.3.2, Gasoline-Contaminated Aquifer Materia!
Enumerations. To compare pristine and gasoline-contaminated aquifer material, aquifer material
from cores extracted aseptically from the treatment cells (July, 1997) was also enumerated for viable,
aerobic heterotrophs and denitrifiers, as well as for aerobic benzene-toluene degraders. Two aquifer
samples were enumerated from each cell: A shallow sample from the 50 to 80 cm bgs interval that
contained a residual gasoline phase, and a deep sample from 155 to 190 cm bgs interval that had been
exposed to high dissolved-phase concentrations but not to gasoline. The results are summarized on
Figure 4-18. For comparison with pristine material, the sample from Core 2 2, which was prepared and
incubated with the contaminated samples, is also plotted on this figure. Viable, aerobic heterotrophs
ranged from 5 x 104 CFU/g (shallow Control Cell) to 107 CFU/g (deep Nitrate Cell), with Core 2-2 falling
within this range. Counts in the pristine sample were greater than in three of the four contaminated
samples (Figure 4-18). However, a response to gasoline contamination was indicated by clearly-
discernable differences in colony types between the pristine- and contaminated-sample plates (data not
shown). Denitrifiers ranged from 4 x l03MPN/g in the pristine sample to >2 x 106 MPN/g in the deep
sample from the Nitrate Cell. Samples from the Nitrate Cell contained the highest numbers of denitrifiers.
Benzene-toluene degraders were less prolific, with numbers ranging from 4 x 101 MPN/g in the pristine
sample to 7.5 x 103 MPN/g in the deep sample from the Nitrate Cell.
Overall the data collected here indicate that the numbers of culturable microorganisms in the samples
extracted from the treatment cells were higher than the pristine sample from Core 2-2. Both denitrifiers
and benzene-toluene degraders were higher by up to two orders of magnitude, but differences in aerobic
heterotrophs were less pronounced. However, it should be noted that the number of samples collected
from the gasoline-contaminated areas was too small to characterize the variability within these areas,
which would be required for a meaningful statistical comparison. If one considers all of the enumeration
results in pristine aquifer material (Table 4-15), numbers of aerobic heterotrophs and denitrifiers are quite
variable, and it is less clear that numbers in contaminated areas are elevated. Similar numbers are
Denitrifiers
Aerobic Heterotrophs
Benzene-Toluene Degraders
ft
03
z
n
2
4
u
2
0
pr ccg cc ncg nc
¦ Sarrple Location
8
6
„D)
u_ 4.
O *
U)
o
_> 2
I r
©
-------
Table 4-17. Experimental Design: ETS Activity in Pristine and Contaminated Aquifer Material
Treatment1
Hypovial Contents
Blank
Sterile
Unamended
Positive
Gasoline-saturated
1/10 gasoline-saturated
Borden aquifer solids, water, N,P
Autoctaved Borden aquifer solids, INT, 0.5 ml sodium azide, water N,P
Borden aquifer solids, INT, water, N,P
Borden aquifer solids, INT, 0.5 ml glucose, water, N,P
Borden aquifer solids, INT, 1 ml gasoline-saturated water, water, N,P
Borden aquifer solids, INT, 1 ml 10x dilution of gasoline-saturated water, water,
N.P
' Each treatment was prepared in duplicate. Gasoline: API 91-01.
consistent with electron-acceptor and organic substrate uptake data which suggested that electron-
acceptor flushing did not stimulate the growth of a large in situ population.
Activity Assays. A second ETS assay was performed to compare aerobic ETS activity in pristine
and contaminated aquifer material. The treatment groups are summarized in Table 4-17. As in the other
ETS assay, because of dilution with other fluid, the actual dissolved hydrocarbon concentrations in the
vials were roughly half the gasoline-contacted concentrations (ca. 50 mg/L and 5 mg/L total aromatics).
For consistency with the enumerations, these vials were incubated at room temperature. Microbial activity
was assayed in samples of pristine aquifer material (Core 2-2), the zones below the gasoline
contamination (155-190 cm bgs) in both treatment cells, and the gasoline-contaminated zone
(80-140 cm bgs) of the Nitrate Cell.
Blank-corrected ETS activity is shown in Figure 4-19. For clarity data are plotted as means of
duplicate determinations, but for some samples there was substantial variability between duplicates.
Trends based on means provided, therefore, only an approximate measure of differences in activity.
Activity varied by both sample location and amendment. The largest accumulation of INT-formazan
occurred in the deep sample from the Nitrate Cell when gasoline-saturated water was supplied as a
carbon source (Figure 4-19). Activity in the deep sample from the Control Cell was similar, except that
formazan accumulation in the gasoline-saturated vials was substantially lower. In both of these samples,
substantial activity also occurred in unamended vials, which probably reflects utilization of existing
gasoline hydrocarbons. In the Control Cell, activity in the positive (glucose-amended) treatment was
similar to the unamended treatment, but in the Nitrate Cell, it was much greater. It is not clear why the
positive controls behaved differently. Activity was generally lower in the pristine sample, but surprisingly,
accumulations over the 28-day incubation in the treatments amended with gasoline-saturated water were
greater than those in the positive control which contained an easily-utilized carbon source. Again, as in
the Control Cell, it is not clear why the activity in the glucose-amended vials was low. The lowest
activities occurred in the sample collected from the gasoline-contaminated zone in the Nitrate Cell.
Therefore, the gasoline phase seemed to inhibit ETS activity somewhat relative to the other contaminated
sample locations included in this experiment.
Overall, this assay indicated that the greatest potential for aerobic biotransformation of gasoline-
saturated water existed in the lower samples collected from the treatment cells, with highest ETS activity
observed in the Nitrate Cell. Activity was relatively low in the presence of a gasoline phase relative to
these lower depths. While activity measured by this assay was low in the presence of a gasoline phase
relative to nearby locations, there was other evidence that the aerobic population, and hence the potential
for aerobic biotransformation of aromatic hydrocarbons, was not completely inhibited within the gasoline
source area (Experiment 6).
61
-------
Nitrate Cell
(Gasoline-Contaminated)
o>
c
ffl
N
ffl
H
Z
CB
Pristine
Nitrate Cell
Control Cell
20 30 0
Time (days)
. pos (glu)
_x gas-sat
,, 1/10 gas-sat
FIGURE 4-19 Accumulation of INT formazan as a measure of ETS activity in pristine and contaminated aquifer mate-
rial. Results shown in bottom two graphs from below gasoline-contaminated zones, con: sterile control;
unam: no carbon source added; pos: glucose added; gas-sat: gasoline-saturated water added; 1/10
gas-sat; 10x dilution of gasoline-saturated water added. Means of two replicates plotted.
4.4 Discussion and Conclusions
These laboratory results indicated that the microbial populations capable of biotransforming high
concentrations of BTEXTMB were patchily distributed in pristine aquifer material. This is not surprising
given the oligotrophic conditions in the Borden aquifer (Butler et al., 1997). It appeared that with in situ
exposure, the population in the Borden aquifer acclimatized to the gasoline phase and associated high
aqueous concentrations of BTEXTMB. Acclimation periods are typically attributed to factors such as
enzyme induction, genetic change, diauxie, or selective enrichment and growth of organisms capable of
degrading hydrocarbon constituents (Chapelle, 1993; Leahy and Coiwell, 1990). We observed distinct
differences in the uptake of 02 and N03, and the utilization of the aromatic hydrocarbons between pristine
and contaminated (19 month exposure) aquifer material. Although the number of replicates was
insufficient for statistical comparisons, the data suggested that numbers of heterotrophs, denitrifiers, and
benzene-toluene degraders in exposed Borden aquifer material did not increase dramatically (<2 orders of
magnitude) relative to pristine material. However, on the basis of the follow-up microcosm study,
aromatic-hydrocarbon degrading activity was much more robust in the exposed aquifer material. There
62
-------
were also indications (e.g., ETS activity) that microbial activity was suppressed somewhat in the presence
of a gasoline phase relative to less-contaminated regions of the treatment cells, but the 02 demand
observed in gasoline-contaminated aquifer material was large, suggesting that stimulating
biotransformation in the residual gasoline source area with 02 would be viable after prolonged exposure.
In the preliminary experiments with low concentrations of benzene, toluene, ethylbenzene, and the
xylene isomers, mass loss under aerobic and denitrifying conditions was generally consistent with
previous studies (Barbara et al., 1992; Major et a!.. 1988; Barker et al., 1987). All of these compounds
were degradable under aerobic conditions, but microbial activity was limited by inorganic nutrients.
Previous studies have also shown that the activity of microorganisms in the Borden aquifer is nitrogen-
limited under aerobic conditions (Barbara et al., 1994). In the presence of 02 m-xylene was typically
utilized first followed by the other aromatic compounds. Under anaerobic, denitrifying conditions, both
toluene and ethylbenzene biodegraded most consistently, while the other aromatic compounds appeared
to be recalcitrant. We observed no nutrient limitations under anaerobic conditions. Under denitrifying
conditions, mass losses were generally small and the minor assimilatory requirement for N may have
been satisfied by NO " eliminating the need for an additional source of supplied N. A concentration of
25 mg/L N03 -N was found to be adequate to support aromatic-hydrocarbon mass loss under denitrifying
conditions over the incubation periods used in this study.
Overall the laboratory experiments indicated that the effect of microaerophilic dissolved 02
concentrations depended primarily on the concentrations of the aromatic hydrocarbons and other carbon
compounds in the system. When the concentration of total aromatic hydrocarbons was low (i.e., on the
order of 10-15 mg/L) and there were no other sources of labile carbon (pristine aquifer material), the mass
of 02 in a microcosm was fairly large relative to the mass of carbon. For example, under these conditions,
the mass of 02 derived from an air headspace was sufficient to mineralize about 50% of the mass of
aromatic hydrocarbons in a microcosm. In these low-carbon systems, we did observe more extensive
mass losses in the presence of microaerophilic 02 relative to losses under anaerobic, N03 -reducing
conditions (e.g., Figures 4-6 and 4-7). Notably, however, benzene losses were minimal in microaerophilic
microcosms.
In gasoline-contaminated microcosms, on the other hand (Experiment 6). less extensive losses of the
aromatics were observed in the presence of microaerophilic 02, even when aqueous hydrocarbon
concentrations were low (10x dilution of gasoline-saturated water). This apparently was the result of 02
consumption by microorganisms growing on other carbon compounds. Abiotic 02 consumption in
laboratory microcosms appeared minor. Other carbon compounds that may have been labile include non-
target dissolved gasoline constituents such as phenolic compounds (see Table 5-2) and various C4
through C7 straight-chained aliphatic compounds which have relatively high aqueous solubilities, or various
insoluble hydrocarbons retained in the gasoline phase. Biosurfactant-producing bacteria capable of
assimilating insoluble hydrocarbons have been isolated in previous studies from petroleum products
(Marin et al., 1996). When the concentrations of dissolved aromatic hydrocarbons were increased to
gasoline-saturated concentrations to reflect field conditions, microaerophilic 0? had no observable effect.
Under these conditions, the mass of 02 was apparently too low to observe any losses even if the aromatic
hydrocarbons were the preferred substrates. Therefore, although there was an observable effect under
favorable conditions (i.e., low substrate concentrations in pristine aquifer material), microaerophilic 0„ was
not effective in enhancing the removal of recalcitrant compounds under conditions similar to those
established in the field. These laboratory data suggest, therefore, that the addition of microaerophilic 02
for enhanced bioremediation may be more effective in locations downgradient of the source. This
conclusion is consistent with observations from the field experiment.
Complete utilization of microaerophilic 02 was not observed in laboratory microcosms. A threshold
dissolved 02 concentration of 0.1-0.5 mg/L persisted in active microcosms incubated in the anaerobic
chamber. It is unclear whether this threshold resulted from positive sampling bias (i.e., from removal of
the microcosms from the anaerobic chamber, or sampling procedures), or was representative of the
microcosm liquid. In contrast to the field cells, where threshold concentrations were also observed
(Chapter 5), laboratory samples were not subjected to negative pressure prior to sampling and steps were
taken to minimize contamination with atmospheric 02 prior to measurement (Appendix C). Nonetheless,
in this study, it was assumed that the threshold represented the lowest measurable concentration, and
63
-------
microcosms with 02 concentrations near this level were considered to be "Oz-depleted". Assuming, on the
other hand, that this residual 02 was not an experimental artifact, the data suggest that microbial
consumption was very slow relative to the lengths of the incubation periods. Persistence could have been
related to slow kinetic uptake at these concentrations (Section 1.3.1). The effect, if any, on the denitrifier
population could not be ascertained. The presence or absence of a threshold concentration, while of
interest for understanding the fate of supplied 02, did not appear to be critical in the assessment of the
effect of microaerophilic 02 in the laboratory experiments.
Although one of the goals of this study was to evaluate nitrate-based biotransformation, the laboratory
experiments clearly showed that aromatic-hydrocarbon mass losses were minor unless 02 was present.
In pristine aquifer material, NO, utilization was observed under anaerobic conditions, but only when
aromatic-hydrocarbon concentrations were low (10x dilution of gasoline-contaminated groundwater).
Under these conditions, mass losses were limited to toluene, ethylbenzene, and less consistently,
m-xylene. When aqueous concentrations were increased to gasoline-saturated levels (Experiment 4),
negligible N03 uptake suggested that denitrifying activity was inhibited in the pristine aquifer material.
Prolonged (19 months) in situ exposure to gasoline hydrocarbons and NO did lead to increased NO„
utilization under anaerobic conditions, but not to increased mass loss of labile aromatic compounds. The
microcosm experiment with gasoline-contaminated aquifer material extracted from the Nitrate Cell
(Experiment 6) suggested that, although N03 -reducing activity was not inhibited, the labile aromatic
compounds were not preferred substrates in this carbon-rich environment. Moreover, the rate and extent
of NOg-utilization was quite variable among the different treatment groups in this experiment. The
inconsistent and often weak NOrreducing activity appears to be characteristic of the Borden aquifer. In
contrast, benzene, toluene, ethylbenzene, m-xylene, p-xyfene, 1,2,4-trimethylbenzene, and naphthalene
were observed to biotransform rapidly at the expense of 02 in the gasoline-contaminated material.
o-Xylene and the other trimethylbenzene isomers may also have been degrading at lower rates, but they
were effectively recalcitrant because of rapid Oy consumption in other reactions.
Nitrate may have been involved in both assimilatory and dissimilatory reactions. Indirect evidence of
assimilatory reduction in gasoline-contaminated material was provided by the rapid utilization of N03 in
the presence of a large mass of 02 (Figure 4-16a) If this N03 utilization had resulted from denitrifying
activity following significant aerobic growth of facultative microorganisms, then rapid utilization of NO.;
should have continued under anaerobic conditions when NO, was replenished. Subsequent rapid uptake
was not observed. Utilization of N03- as a source of N for 02-driven reactions can be considered a
positive aspect of an 0., / N0:J mixture, but under microaerophilic conditions the effect would be limited.
Under 02-depleted conditions, the intermittent accumulation of N02 as well as the production of N20 in
acetylene-blocked microcosms suggested that dissimilatory N03 reduction had been induced, but these
data were variable and the specific nitrate-reducing pathway (e.g., denitrification) remained poorly defined
in some experiments.
Patterns of BTEX, 02, and N03 / N02 concentrations suggested that O and N03" were used
sequentially under mixed electron-acceptor conditions; most BTEX biotransformation occurred early, likely
at the expense of microaerophilic 02; additional TEX losses occurred later under mixed electron-acceptor,
and denitrifying conditions. In contrast to other published studies (e.g., Hutchins, 1991a), our experiments
provided no clear evidence that low levels of O, were facilitating the transformation of recalcitrant
compounds such as benzene under anaerobic, N03-reducing conditions. Our experiments with low
substrate concentrations did show, however, that mass losses of labile compounds in mixed
microaerophilic / N03 microcosms exceeded losses in comparable microaerophilic only and anaerobic,
denitrifying microcosms. This indicated that denitrifying activity commenced after O., depletion with no
apparent lag period, and that the extent of biotransformation was maximized by the presence of two
electron acceptors. Unfortunately, in the Borden aquifer this mixed electron-acceptor-driven oxidation
appeared to be limited to low-carbon conditions. In experiments with gasoline-contaminated aquifer
material, aromatic-hydrocarbon mass loss was either very small or negligible under mixed microaerophilic
/ NOa conditions. In the high-BTEXTMB microcosms, the mass of 02 was apparently too small to
observe (i.e., not detectable above experimental variability) any aromatic-hydrocarbon losses, and the
denitrifying population did not appear to utilize those compounds that are typically labile at lower
concentrations. These results are broadly consistent with those obtained in situ.
64
-------
CHAPTER 5. FIELD EXPERIMENTS
5.1 Overview of Results
5.1.1 Flow Characteristics
All breakthrough curves (BTCs) within the pea gravel layers (60 cm bgs) reached a relative
concentration near C/C =1 within 12 hours of the beginning of the July, 1996 tracer test (Appendix B).
Because the pea-gravel layers extended to the sheet-piling walls, it likely that injected water spread
horizontally to the edges of the cells throughout the flushing experiments, providing a uniform initial
distribution of NO, and dissolved 02, although the apparent rapid consumption of 02 may have restricted
its initial distribution to the region around the injection well.
Breakthrough data from piezometer ports at two depths, 120 and 180 cm bgs, were used to calculate
dispersivities and groundwater velocities (Appendix B). As the tracer front migrated downward through
the ceils, spatial groundwater velocity fluctuations were observed, but tracer broke through all monitored
ports; this indicated that there were no large regions being bypassed by the injected fluid. Velocity
fluctuations in individual flow tubes were caused by spatial variability in the hydraulic gradient (i.e., higher
velocities above the extraction well), in aquifer properties such as hydraulic conductivity (heterogeneities),
and possibly in gasoline content. Average linear groundwater velocities, determined from fitting a one-
dimensional advection dispersion equation to the tracer BTCs, ranged from 0.6 to 1.2 cm/hr in the Control
Cell, and 0.4 to 1.1 cm/hr in the Nitrate Cell (see Table B-2).
The velocities measured from breakthrough data can be compared to the expected velocity under
steady flow conditions. Using a porosity of 0.33, a 200 ml/min target injection rate yields an expected
velocity of 0.9 cm/hr, which is consistent with the breakthrough data. Accordingly, the mean injection flow
rates were used in advective mass-flux calculations to estimate the masses of dissolved constituents
added and removed from the cells under flushing conditions. The good agreement between injection
rates and breakthrough data also indicated that there was no short-circuiting of injected water along the
walls of the piezometers.
5.1.2 Dissolved Oxygen and Nitrate
Dissolved Oxygen, Dissolved 02 depletion was observed in both treatment cells. The mean
injection dissolved 02 concentrations were 2.3±1.8 mg/L (n=80) and 2.3±1.7 mg/L (n=81) for the Nitrate
and Control Cells, respectively, where the variability is expressed as a standard deviation. These means
include the initial seven days of the experiment when fully-oxygenated water was injected into the cells, tf
the measurements from the first week are excluded, mean injection concentrations fall to 1,9±0.9 mg/L
and 2.0±1.0 mg/L for the Nitrate and Control Cells, respectively. The mean concentrations obtained from
the extraction-well sampling ports were 0.30±0.14 mg/L (Nitrate Cell) and 0.27+0.26 mg/L (Control Cell).
Measured concentrations are shown on Figure 5-1, where a mean value is plotted when more than one
sample was collected in a 24 hr period.
Dissolved 02 was depleted rapidly to a non-zero threshold concentration in both cells (Figure 5-2).
Concentrations in samples from 60 cm ports were at the threshold, and additional uptake between 60-
and 180-cm depths was not observed. Therefore, the majority of the dissolved 02 was utilized within
hours of injection. The threshold concentration varied from about 0.4 mg/L to 1 mg/L, depending on the
location, and averaged around 0.75 mg/L at a given depth interval (Figure 5-2). Non-zero threshold
concentrations were also measured during both the static and 24-day flushing periods that followed the
initial flushing experiment (data not shown). It appears, based on rapid rates of uptake in active
laboratory microcosms relative to controls (e.g., Experiment 6), that O, was utilized by microbial activity,
65
-------
Control Cell
—o— Injection Port
—o—Extraction Port
0 30 60 90 120 150 180 210
I ime (days)
CO
E,
o
Q
Nitrate Cel
- Injection Port
—Extraction Port
o,o o
30
60
90 120
Time (days}
150 180 210
FIGURE 5-1 Injection arid extraction D.O. concentrations, Mean values plotted when more than one sample collected
within 24-hr period (solid squares).
Control Cell
—A— 60 cm Fbrts
...O--- 180 cm Fbrts
Mean Inj. Cone.
45
135
180
Time (days)
CD
E.
o
d
45
Nitrate Cell
¦a— 60 cm Ports
o--- 180 cm Ports
Mean Inj. Cone.
- - - »T- - - *
90 ,
Time (days)
135
180
FIGURE 5-2 D.O. concentrations at 60- and 180-cm bgs ports during the 174-day flushing experiment. Plotted values
at each depth are means and standard deviations from five piezometer ports.
66
-------
but given the detection of dissolved Mn in water extracted from the supply well (Table 2-1), it is possible
that in the field some of the 02 was utilized in abiotic reactions as well.
The observed threshold concentrations obtained from the piezometers may not represent in situ
dissolved-02 concentrations. One possible source of positive bias is the sampling procedure. A small
amount of 02 could have been incorporated into the flowing stream of groundwater during sampling under
suction. To further investigate this possibility, similar sampling techniques were used to measure the
dissolved 02 concentrations from bundle piezometers in the monitoring grid used by Barbara et al. (1992),
Mean concentrations of duplicate samples from seven piezometer ports in the anaerobic, landfill leachate
plume ranged from 0.41 mg/L to 0.9 mg/L. This suggested that contamination with atmospheric 02
occurred during sampling. On the other hand, vertical profiles from center piezometers indicated that
dissolved 02 concentrations decreased progressively during the flushing experiments (Figure 5-3).
Concentrations lower than the threshold values discussed above were obtained on the last sampling date
when the cells were static, suggesting that it was possible to measure lower 02 concentrations from these
ports. Because of these inconsistencies, the dissolved 02 data obtained from the multilevel piezometers
are considered semi-quantitative; they showed that dissolved 02 was essentially removed, but could not
be used to define the actual residual concentration.
Nitrate. The mean injection NO," concentration from the initial 174-day flushing experiment was
116133 mg/L (n=25). The mean concentration from the extraction well port was 82±23 mg/L (n=13).
Because there were two target injection concentrations (Figure 5-4), the means and standard deviations
were calculated as described in Section 3.5.3 for injection rates. Nitrate was not detected in the Control-
Cell injection water (n=3).
Control Cell - PZ3E
D.O. (mg/L)
1 2 3
50
E 100
o
Q.
Q 150
200
250
X A
Nlitrate Cell - PZ4E
D.O. (mg/L)
1 2 3
50
E 100
o
Q.
CD
~
150
- - - A—
Day 21
Day 50
Day 141
X —
Day 430 (static)
O
Mean Inj. Cone.
200
250
FIGURE 5-3 Vertical profiles of D.O. at various times during and after the 174-day flushing experiment. Data collected
from center piezometers.
67
-------
200
150
. * xx
X x
D)
E,
CD
CO
100
50
x, v xx
x xx
X X
x Measured
Target
Q I 1 I . IX--J ' ' - ' '—1—1 1 1 ' 1 11,1—I 1—1 L—i 1 t_J i_J I 1_J L_L_
0 30 60 90 120 150 180 210
Time (days)
FIGURE 5-4 Injection NO. concentrations.
Consumption of N03 under flushing conditions was relatively low. Similar to dissolved 02, N03"
depletion during the 174-day experiment appeared to occur rapidly after injection into the cell (Figure 5-5).
Mean concentrations from 60-cm ports were below injection concentrations, but additional losses between
60 and 180 cm depths appeared to be quite small (Figure 5-5). Similar behavior was observed during the
24-day flushing experiment. The addition of the MBH medium during this 24-day period did not result in
an observable increase in the rate of NO.; utilization. This result is consistent with laboratory experiments
(Chapter 4), which show that the sluggish NO.;-reduction typical of the Borden aquifer does not result from
nutrient limitations. The amount of N03 that was utilized during the 174-day flushing experiment is
calculated in Section 5.2.2.
To determine if NO.; utilization would be observed during a longer residence time, a conservative
tracer (bromide) was pumped into the Nitrate Cell during the final week of the 174-day flushing
experiment. Three ports from the 60-cm depth and three from the 180-cm depth were then sampled over
the following 138 days when the cells were static to observe the extent of anaerobic NO.; depletion
relative to Br. However, depletion of both Br and N03 was observed, particularly at the lower depth
where dilution with underlying groundwater may have occurred in response to a rising water table. To
account for depletion of the tracer, Br concentrations from each sampling port were corrected back to the
initial concentration. These correction factors were then applied to the corresponding N0:; value to obtain
a corrected concentration for each sampling event. A mean corrected NO,; concentration was then
calculated from the three ports at each depth for each sampling event (Figure 5-6). While there is
substantial scatter in the data, this figure indicates that there were losses of N03 relative to Br, with
greater losses at the lower depth. If depletion is described with a zero-order (linear) model, depletion
rates of 0.20 mg/L/d (60 cm) and 0.67 mg/L/d (180 cm) are obtained. Given the scatter in the data, a
first-order model could also be fit reasonably well to the data. These fits yielded first-order rate constants
of k=0.002/d (60 cm) and k=0.02/d (180 cm). As discussed further in the following sections, these trends
have been attributed to biological activity.
68
-------
200
x — Injection Port
...o... 60 cm Ports
0 30 60 90 120 150 180 210
Time (days)
200
150
CT>
100
50 ¦ ,
-~•--60 cm Ports
180 cm Ports
30 60 90 120 150
Time (days)
180 210
FIGURE 5-5 Nitrate concentrations iri injection water and 60- arid 180-cm bgs ports during the 174-day flushing
experiment. Plotted values for 60- and 180-cm depths are means and standard deviations from five
piezometer ports.
Elapsed Time - Static Period (days)
FIGURE 5-6 Nitrate concentrations during the static period between flushing experiments. Bromide was used to
correct NO concentrations as discussed in text. Plotted values at each depth are means and standard
deviations from three piezometer ports.
69
-------
5.1.3 Organics
Dissolved-Phase. As shown by effluent data, dissolved aromatic hydrocarbons concentrations were
quite high in both cells (Figure 5-7). By the end of the flushing experiments, concentrations ranged from
about 1 mg/L for benzene to 20 mg/L for toluene. With the exception of benzene and toluene, final
concentrations in effluent water were similar to their respective concentrations in gasoline-saturated water,
as measured in laboratory equilibration experiments with fresh API 91-01 (see Table A-4). Benzene and
toluene concentrations declined to ca. 5% and 50% of initial gasoline-saturated values, respectively,
reflecting rapid changes in the mole fractions of these constituents. In contrast, after about 2 months of
flushing, concentrations of the less soluble compounds (i.e., xylenes, trimethylbenzenes and naphthalene)
reached initial saturated values and remained at these levels throughout the experiment. The reasons for
the relatively slow approach to equilibrium concentrations are not clear; the best explanation may be
substantial preferential flow around the most heavily contaminated regions of the treatment cells during
the early stages of flushing. These trends are shown for the Nitrate Cell on Figure 5-8.
Relatively rapid depletion of the more soluble compounds is consistent with the dissolution of a multi-
component organic liquid. As indicated by these dissolved phase results, the gasoline phase was nearly
depleted in benzene, but not in the less-soluble aromatics, by the end of the experiment (Figures 5-7 and
5-9). The rapid depletion of benzene was expected on the basis of its high solubility, and was consistent
with equilibrium partitioning between the gasoline and mobile groundwater. The vertical profiles shown in
Figure 5-9 suggest that benzene was depleted most rapidly from the top of the gasoline-contaminated
region, and that the depleted zone propagated downward as flushing progressed. Other relatively-soluble
compounds such as toluene showed similar trends. These temporal trends probably represent the
downward propagation of a dissolution front through the gasoline-contaminated region. The relatively
rapid removal of the soluble constituents is also evident from 180-cm BTCs (Figures 5-10 and 5-11).
The temporal and spatial patterns of dissolved-phase concentrations were also quite similar between
cells (Figures 5-7 through 5-11). Based on the 180-cm BTCs, there were no major differences between
the cells in the patterns of dissolved-phased aromatic hydrocarbons (Figures 5-10 and 5-11). For
example, enhanced removal of toluene, the most labile compound under denitrifying conditions, was not
evident in the Nitrate Cell relative to the Control Cell. Similarities between the cells are also apparent in
the extraction-well BTCs (Figure 5-7), which represent volume-averaged concentrations exiting the cells.
The close similarity of the Nitrate- and Control-Cell breakthrough curves is additional evidence that a
common, abiotic transport process such as dissolution was the dominant mass removal mechanism.
Concentrations of aromatic hydrocarbons did not decline appreciably in either cell during the static
period between flushing experiments. This includes compounds such as toluene and ethylbenzene that
are typically labile under anaerobic, denitrifying conditions (Chapter 4). The apparent lack of aromatic-
hydrocarbon utilization in the Nitrate Cell is important because N03 depletion was observed during this
period. Figure 5-12 shows toluene, ethylbenzene, and N03~ from one of the piezometers (PZ4A) that was
sampled during the static period. Concentrations were corrected for declining Br as described in Section
5.1.2. These plots suggest that if N03" was being used by denitrifying bacteria as an electron acceptor,
the labile aromatic hydrocarbons did not serve as electron donors.
Core Extracts. The distribution and quantity of aromatic hydrocarbons remaining in the cells after
the flushing experiments were completed was estimated from core extract samples. Concentrations of
total BTEXTMB ranged from below detection limits to about 6,000 mg/kg of aquifer material (Figures 5-13
and 5-14). Patterns between cells were remarkably similar. Peak concentrations show that the gasoline-
contaminated zones were located within the 50 to 150 cm bgs depth interval. This pattern is consistent
with the dissolved-phase profiles shown in Figures 3-5, 3-6, and 5-9. With the exception of cores 3K and
4L, located near the centers of the cells where the emplaced gasoline was deeper, all cores transected
the entire gasoline-contaminated interval. There was very little difference between cells in the spatial
patterns of BTEXTMB. Highest concentrations of residual aromatics were present in the center of the
cells and in the upper right quadrants (Figures 5-13 and 5-14). As with the dissolved phase,
concentrations of individual compounds in the residual gasoline phase were broadly consistent with trends
expected to result from gasoline dissolution, i.e., the residual gasoline was most depleted in soluble
compounds such as benzene and toluene.
70
-------
50
40 -
O)
o °o _
O -O » 9
30
Control
o 4 4 t + .
40 80 120
Time (days)
160
200
Nitrate
C 20
^—a —- y -x r — x
"~y ^ -4- 4-4- -4- i-f - 4* - 4- i- -+-
40
160 200
•BENZENE
- TOTAL XYLHMES
-O- - TOLUENE
-X — TOTAL TMB
-O ETHYLBENZENE
+ - WPHTHALBC
FIGURE 5-7 Concentrations of dissolved aromatic hydrocarbons in samples collected from extraction-well ports.
Nitrate Cell
25
60 90 120
Time (days)
q Benzene ~ Toluene a Total Xylenes x Btiylbenzene * Total TMB o Naphthalene
FIGURE 5-8 Concentrations of aromatic hydrocarbons in samples collected from the extraction-well expressed as
percentages of concentrations in water equilibrated with fresh API 90-01 gasoline.
71
-------
Control Cell
0
50
100
x
•X
1.
I •
E x.
o,
.c 150 X
S.
CD o-A-.
D XA
200
X5f
250
300
Benzene (mg/L)
20 40 60
A
-o . A
80
BTEXTMB (mg/L)
0 200 400 600
50
100
/
o
150
a
a>
D
200
250
300
'X A
W \
X O A
V; '
x$
1V •,
X ,0 A
J/ '
&'
A
a/
C 25123 mg/L
@ 30 cm
Nitrate Cell
Benzene (mg/L)
0 20 40 60 80
BTEXTMB (mg/L)
150 x
x
200
400 600
Q.
CD
Q
300
50
100
150
200
250
300
X O A -
v\
X O A
Ju'
;x oo
A x6
A
A
¦A - PREPUMPING
-O DAY 141
DAY 50
—X — FOST PUMPING
FIGURE 5-9 Vertical profiles of benzene and total BTEXTMB before, during, and after the flushing experiments. Data
collected from center piezometer.
72
-------
Nitrate Ceil -180 cm Depth
40
_ 30
j
i)
r 20
10
0
Benzene
45 90 135 180
Time (days)
Toluene
45 90 135 180
Time (days)
Ethylbenzene
45 90 135 180
Time (days)
Total TMB
45 90 135 180
Time (days)
35
28
E,
21 -
6
14 -
c
o
O
7 -
I
0 !
0
1.5
1.0
O 0.5
Total Xylenes
JO
x b
45 90 135 180
Time (days)
0.0
Naphthalene
0 45 90 135 180
Time (days)
-PZ4A-6
------
- PZ4B-S —
PZ4C-6 i
X —
-PZ4D-6
— PZ4E-6
FIGURE 5-10 Concentrations of dissolved aromatic hydrocarbons at individual 180-cm bgs ports in the Nitrate Cell
during the 174-day flushing experiment.
73
-------
Control Cell -180 cm Depth
45 90 135 180
Time (days)
80
60
o>
o
U
40
20
Toluene
45 90 135 180
Time (days)
12
O)
o
c
o
O
Ethylbenzene
$
45 90 135 180
Time (days)
35
28
I5 21
Q
g 14
O
7
0
Totaf Xylenes
45 90 135 180
Time (days)
Total TMB
Naphthalene
O)
4
2
0.0
135
180
180
135
Time (days) Time (days)
—X — PZ3D-6 —o—PZ3E-6
FIGURE 5-11 Concentrations of dissolved aromatic hydrocarbons at individual 180-cm bgs ports in the Control Cell
during the 174-day flushing experiment.
74
-------
Piezometer PZ4A
O)
£
c
o
'¦+—»
CD
8
c
o
O
TJ
®
•#—«
O
— Ethylbenzene
0
30
60 90 120 150
Elapsed Time - Static Period (days)
FIGURE 5-12 Corrected toluene, ethylbenzene, and N0a" concentrations from a selected piezometer in the Nitrate Cell
that was sampled during the static period. Upper graph: 60-cm bgs port; Lower graph :180-cm bgs port.
Bromide was used to correct concentrations as discussed in text.
75
-------
BTEXTMB (mg/kg)
0 3000 6000
BTEXTMB (mg/kg)
0 3000 6000
a.
CD
Q
0
50
100
150
200
BTEXTMB (mg/kg)
0 3000 6000
a
s>
D
BTEXTMB {mg/kg)
0 3000 6000
BTEXTMB (mg/kg)
0 3000 6000
CORE 4G
O
CORE 4H
CORE 41
CORE4L
CORE4F
O
CORE 4
a
FIGURE 5-13 Vertical profiles of residual BTEXTMB in the Nitrate Cell. Concentrations obtained from field methanol
extraction of core subsamples. Cores were collected during August, 1997. Surveyed core locations
also shown.
76
-------
0
0 A
50 |
1
JZ
100 &
Q.
ED
O
150 -
200
3D00 600Q
E
3
£
&
Q
BTEXTMB (mg/kg)
0 3000 6000
0 ±~
I
50
100
150
200
s
BTEXTMB (mg/kg)
0 3000 6000
50
100
150
200
%
a
50 -K
100
150
200
BTEXTMB (mg/kg)
0 3000 6000
CORE 3F
o
CORE 3G
o
CORE 3H
CORE3K
o
CORE 3i
CORE 3J
o
CORE 3E
o
CORE 3B
CORE 3C
CORE 3D
o
BTEXTMB (mg/kg)
0 3000 6000
0 j ¦ ¦
50
150
200
BTEXTMB (mg/kg)
0 3000 6000
0 ;
E
50-
3-
.c
100 I
S.
~
150
200
E
.H.
I
o
BTEXTMB (mg/kg)
0 3000 6000
0 ^ , ,
Q.
&
50
100 f
150
6''
200 S
BTEXTMB (mg/kg)
0 3000 6000
0
50
100
150
200
BTEXTMB (mg/kg)
0 3000 6000
50 $>
100 0
150
200
FIGURE 5-14 Vertical profiles of residual BTEXTMB in the Control Cell. Concentrations obtained from field methanol
extraction of core subsamples, Cores were collected during August, 1997. Surveyed core locations
also shown.
77
-------
With the exception of Core 41 which appeared to be depleted in alkanes, the analysis of core extracts
for gasoline component classes did not reveal substantial depletion of any hydrocarbon component group
relative to the fresh gasoline standard and sample spiked with fresh API 91-01 (Table 5-1). This was
consistent with the short duration of the flushing experiments and the low solubilities of the majority of the
gasoline constituents.
5.1.4 Nitrite Production
Nitrite was frequently present in groundwater samples collected from the Nitrate Cell. Nitrite was
never detected in samples of injection water spiked with N03 , but was present in 70% of the groundwater
samples (n=189), collected under both flushing and static conditions, that contained N03 (data not
shown). Concentrations were typically less than 1 mg/L as N02, but there were numerous occurrences
between 1-10 mg/L, and the highest measured concentration was 17.2 mg/L as NO,". In general, N02
was detected more frequently and at higher concentrations in samples collected from lower depth
intervals. The presence of N02 is considered evidence of dissimilatory NO. reduction (Mikesell et al.,
1993). Nitrate can be reduced to N2 abioticaliy in the presence of Fe2, but because the reaction rate is
negligible unless pH is alkaline (i.e., >7), and a catalyst such as Cu is present (Buresh and Moraghan,
1976), it has been discounted as an important mechanism here.
5.1.5 Metabolite Production
Additional evidence of microbial activity was obtained from metabolite production (Table 5-2). It is
recognized that these samples were collected after the flushing experiments were completed, and the
MBH medium had been pumped into the cells, and therefore may not be representative of earlier flushing
conditions when the majority of the electron-acceptor uptake and aromatic-hydrocarbon mass removal
occurred. However, because the response (i.e., changes in electron-acceptor and hydrocarbon utilization)
to the MBH solution was negligible, and concentrations of other redox-sensitive species were similar to
those measured during flushing conditions (Table 5-3), we have assumed that conditions at the time of
metabolite sampling were representative of this earlier period.
The sample of fresh API 91 -01 gasoline did not contain any aromatic or aliphatic acids. The supply
well was sampled on two occasions, and several C:j through C5 aliphatic acids as well as benzoic acid
were present in the supply water. This may indicate that a plume of partially-oxidized hydrocarbon
constituents from the upgradient treatment mound had reached the supply well by the end of the
experiment. Alternatively, short-chained aliphatic acids can be formed from the fermentation of natural
organic matter in low-02 environments (Thurman, 1985; McMahon and Chapelle, 1991). Concentrations
of the constituents detected in supply water were generally higher in the treatment cells, suggesting
additional production within the cells. For example, although benzoic acid was detected at 5 |jg/L in the
supply well, it was present at concentrations up to 41 pg/L (PZ4E-3) in the Nitrate Cell, and 21 pg/L
(PZ3E-3) in the Control Cell.
Table 5-1. Hydrocarbon Component Classes in API 91-01 Gasoline and Gasoline-Contaminated Core Extract
Samples. Core Samples Collected July, 1997. All Results Expressed in Weight Percent.
CON1ROL CELL NITRATE CELL
API 91-01 Gasoline
Gasoline
Spike
Core 3J
Core 3H
Core 3K
Core 41
Core 4L
Core 4K
Class Depth (cm bgs):
80
69
50
70
65
65
Alkanes
46.8
40.7
43.0
44.2
48.0
30.6
46.6
42.8
Aromatics
35.3
44.2
39,9
38.4
34,0
61.1
34.9
40.0
Bicycloalkanes
0.1
<0.1
n.d.'
n.d.
n.d.
n.d.
n.d.
n.d.
Naphthenes
7.9
5.7
7.1
7,3
7.0
3.5
8.0
6.9
Olefins
6.6
5.6
5.4
5.7
5.5
0.7
5.3
5.9
PNA
0.5
0.2
0.5
0.7
0.7
nd
1.1
0.8
Other
2.9
3,6
4.1
3.8
4.8
4.1
4.1
3.7
' n.d. - riot detected.
78
-------
Table 5-2. Detected Organic Acids and Phenols in API 91-01 Gasoline and Groundwater. AH Concentrations in
NITRATE CELL CONTROL CELL
Collection Date1:
NA
6/12/97
7/30/97
6/12/97
7/30/97
6^12/97
7/30/97
6/12/97
7/30/97
8/12/97
7/30/97
Location;
Gasoline Supply Well
Supply Well
PZ4E-2
PZ4E-3
PZ4E-6
P24E-6
PZ3E-2
PZ3E-3
PZ3E-6
PZ3E-6
Compounds Depth (cm bgs):
60
90
180
180
80
90
180
180
Acetic acid
41
51
-
25
3
Propanoic acid
.2
7
6
3
?
"
2-methyi propanoic acid
..3
*
**
32
*
**
50
11
1
*«
3,3-dimethylpropanoic acid
*
**
10
5
5
5
trimetbyiacetic acid
*
14
*
22
*
16
*
26
14
37
butyric acid
2
4
66
**
1
**
10
10
*
"
2-methylbutyric acid
~
3
36
*
**
83
16
*
**
3-methylbutyric acid
"
*
4
48
**
*
¦**
234
34
*
3,3-dimethylbutyric acid
59
26
60
41
42
pentanoic acid
- *
3
75
**
11
**
23
3
80
hexanoic acid
*
*
5
16
7
3
5
*
9
*
2-ethyl hexanoic acid
**
*
**
*
*«
*
*
*
2
**
2-methylhexanoic acid
*
*
*
-
8
10
haptanoic acid
**
*
**
14
**
*
**
*
**
*
**
benzoic acid
*
*
5
22
41
*
41
16
21
*
11
phenyiacetic acid
**
*
71
7
*
12
126
*
*
**
o-methylbenzoic acid
+
*
*
218
8
3
21
*
*
*
*«
m-methylbenzoic acid
*
*
*
83
17
*
5
*
23
w
7
m-tolyacetic acid
*
*
*
*
.*
*
*
*
*
+
*
p-tolyacetic acid
*
*
*
*
6
*
4
*
*
¦*
p-methylbenzoic acid
*
*
*
*
5
*
8
*
*
-*•
**
2,6-dimethylbenzoic acid
•
*
*
26
*
*
*
*
*
*
•
2,5-dimetbylbenzoic acid
*
*
*
409
16
3
11
*
**
*
**
3,5-dimethylbenzoic acid
*
*
29
5
*
**
*
**
*
*
2,4-dimethy [benzoic acid
*
*
37
6
*
4
*
**
*
**
decanoic acid
»
*
**
18
**
*
*
3
*
**
4-ethylbenzoic acid
*
*
*
*
*'*
*
«„
*
**
*
**
2,4,6-trimethylbenzoie acid
*
*
*
99
4
*
**
*
**
*
**
3,4-dimethyibenzoic acid
•*
*
*
14
6
*
5
*
**
*
**
2,4,5-trimethylbenzoic acid
«
*
53/57
7
4
*
Phenol
44,000
**
4
**
4
**
o^cresoi
53,800
**
7
**
**
**
m-cresol
37,600
**
5
**
**
**
p-cresol
19,000
**
**
**
o-ethylphenol
9,600
*
»*
**
**
*
2,6-dimethylphenoi
1,400
*
**
**
*
**
2,5-dimetfiyipheno!
10,700
-
**
*
2,4-dimethylpheno!
5,600
*
**
*
**
*
3,5-dimethylphenol + m-ethylphenol
24,900
*
9
*
**
*
2,3-dimethyl phenol
5,500
*
**
**
**
*
p-ethylphenoi
6,000
*
*
**
*
*
3,4-dirnethylphenol
4,700
*
**
**
*
**
1 Samples collected on 6/12/97 analyzed by National Center for Integrated Bioremediation and Development, University of Michigan, Ann Arbor, Ml.
Samples collected on 7/30/97 and API 91-01 gasoline sample analyzed by National Risk Management Research Laboratory, U.S. EPA, Ada, OK.
2* Not found.
3" Concentration of extract was below lowest calibration standard (3 ug/l).
79
-------
Table 5-3. Concentrations in mg/L of Selected Redox-Sensitivc Constituents in the Experimental Cells and
Injection Water. Concentrations Obtained During the Flushing Periods are Expressed as a Range
of Measured Values. Concentrations from 3/22/97 and 6/12/97 Sampling Dates are from a Single
Piezometer Sample in Each Cell. All Samples in Cells Collected from Various Ports between 60-
and 180-cm Depths.
Injection Water
Nitrate Cell
Control Cell
Sampling Interval:
(5/96- 11/96}
(5/97)
(5/96 - 11/96)
(3/97)
(6/97)
(5/96- 11/96)
(3/97)
(6/97)
# Days:
174
24
174
1
1
174
1
1
flushing
static
static
flushing
static
static
pH
6.84 - 7.48 !3)1
7.56
„ „ 2
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
Alkalinity (CaC03)
252 - 269 S3)
n.a.
248-267®
n.a.
n.a.
252-277®
n.a.
n.a.
Sulfate
2.05- 9.07{7}
4.04
3.13-9.47 n3)
1.70
4.14
2.65 10.6f16:
2.23
4.78
Sulfide
n.d.3
n.a.
n.d.
n.d.
n.d.
n.d.<4)
n.d.
n.d.
Total Iron
n.d. - 0.29 w
2.37
n.d, -0.12(13)
0.28
0.14
n.d. -0,10(16)
0.34
1.30
Methane
1.04
n.a.
n.a.
1.47
0.46
0.24
0.94
0.088
Nitrate
n.d. -175S25M
100 - 123(5M
n.d. -160(1S9)
10.6
88.9
n.d.(13>
n.a.
n.a.
Nitrite
n.d.(26M
n.d.(5M
n.d. -17.2<;89'
0.68
0.7
n.d.(13>
n.a.
n.a.
Dissolved Oxygen
0.2-10.0(161)
3.9 - 5.6
-------
The suite of compounds detected in the Control Cell was somewhat different. With the exception of
m-methylbenzoic acid, samples from the Control Cell did not contain aromatic acids. This may represent
lack of production of these compounds, or relatively rapid turnover to other oxidized constituents (Grbic-
Galic and Vogel, 1987). There were, however, elevated concentrations of short-chained aliphatic acids
(Table 5-2). The origin of these partially-oxidized compounds is less clear because of the numerous
potential parent compounds in the gasoline.
The presence of low molecular weight organic acids derived from gasoline hydrocarbons
demonstrates that, by the end of the flushing experiment, there was microbial activity in the treatment
cells. The number of samples collected in this study was insufficient to determine whether these
compounds resulted from aerobic or anaerobic transformations of gasoline constituents. However,
production and persistence of these compounds has been observed frequently in anaerobic environments
(Cozzarelli et al., 1990 (Fe-reducing and methanogenic); Kuhn et al., 1988 (N03 "-reducing); GrbicSGalic
and Vogel, 1987 (methanogenic)). In the Nitrate Cell, the redox potential of the system was probably
buffered by the continuous presence of NO ' (Kehew and Passero, 1990). This suggests that acid
production was most likely associated with N03"-reducing and/or 02-reducing activity. In the Control Cell,
the dominant terminal electron-accepting process is not known. Nitrate was not present, and periodic
sampling for other electron acceptors did not provide evidence of substantial Fe-reducing, S042 -reducing,
or methanogenic activity. As shown in Table 5-3, ranges of S042~ and total dissolved Fe in the Control
Cell were similar to the supply water, as well as to the Nitrate Cell. Methane was present in both cells, but
was also detected once in the supply-well water. The peat layer encountered at 2.8 m rather than
gasoline hydrocarbons may have been the source of the CH4 in these samples.
5.2 Mass Balance Calculations
Mass balances were performed on both added electron-acceptors and the aromatic hydrocarbons.
To simplify the analysis, it was assumed that flow in the cells was at steady state over the 174-day
flushing period, so that the injection rate could be used to estimate mass fluxes both into and out of the
cells, as required. In addition the volumetric flow rates were assumed constant in space (i.e., the
treatment cells were treated as one-dimensional columns) and time (i.e., mean injection rates were used
in calculations). These assumptions were supported by the continuous injection and constant upper and
lower head conditions maintained during the flushing experiments, as well as the Br tracer-test results.
For all constituents, the mass advected across the horizontal plane located 20 cm above the top of the
extraction well screen (180 cm bgs) was used as the most reliable measure of the mass "extracted" from
the cells. For most of the mass-balance calculations, data from the extraction wells were not used
because concentrations may have been affected by cyclical pumping rates and unquantified dilution with
underlying groundwater. .
5.2.1 Dissolved Oxygen
Because of the difficulties associated with obtaining reliable results from the multilevel piezometers, a
rigorous mass balance was not performed on the dissolved 02. However, the mass injected into each cell
was estimated using a simple mass flux calculation. In this study, mass flux was calculated for the entire
2m x 2m area rather than a unit area; mass flux (QC) therefore had units of [M/T]. The mass injected
into each cell was obtained by integrating the advective mass flux over the 174-day injection period:
M='jFdt-'jQ(t)C(t)dt-QCt 5--|
0 0
where g[z3/7"] and c[m/i3] are the mean volumetric injection rates and dissolved 02 concentrations,
respectively, and M is the mass. The propagated error in this mass estimate was calculated from:
where S2M, *1, and % are the variances of the respective means (Bevington, 1969). Using the mean
injection rates and Dissolved 02 concentrations provided in Sections 3.5.3 and 5.1.2, respectively, the
81
-------
masses of 02 injected into the cells over the 174-day flushing experiment were 143±115 g (Nitrate Cell)
and 136±100 g (Control Cell). To calculate the additional masses pumped into the cells during the 24-day
flushing experiment, mean injection concentrations (4.6+1.3 mg/L (n=5) and 4.5±1.4 mg/L (n=5) for the
Nitrate and Control Cells, respectively) were multiplied by the injection rate (200 ml/min, based on one
measurement in each treatment cell), and the error was calculated from Equation 5-2. Resulting masses
of 02 injected into the cells during this period were 32±9 g (Nitrate Cell) and 31 ±10 g (Control Cell). .
Although the magnitude of the trapped air phase was not measured, it is possible to estimate the size
of this 02 reservoir. Assuming that the volumetric residual air content was in the range of 5-10% (R.W.
Gillham, personal communication), a dewatered volume of 4.4 m3 (2m x 2m x 1.1m height), and an
oxygen gas density of 1.43 g/L (standard temperature and pressure) (Hillel, 1982), the initial mass of 02
could range from 60-120 g per cell. Based on the assumptions given above, the mass of 0? derived from
trapped air was roughly 50-100% of the mean injected 02 masses, and therefore probably constituted a
significant component of the total 02 budget. Considering both sources of 02, the total mass available for
reactions may have been as high as ca. 300 g per cell.
5.2.2 Nitrate
The NOa mass loss during the 174-day flushing experiment was estimated by comparing the N03"
masses injected into and extracted from the Nitrate Cell. Data collected during the 24-day flushing period
were insufficient to perform a mass balance, but qualitatively, the rate of N03 utilization, as indicated by
concentration differences, was similar. Mass loss was obtained by difference from Mt§s = M ed - M gd.
Injected and extracted masses were calculated by integrating the advective mass flux into and out of the
cell, respectively, over the 174 day flushing experiment:
t t _f
M = jFdt = jO(t)C(t)dt = OjC(t)dt 5.3
00 0
where,
F = Mass flux [M/T],
q = Mean injection rate [L3 T],
C = N03 concentration at time t [M/L3],
t = Time [T]
Because the product QC was not available for every sampling event, masses were calculated using
the mean injection rate (Q) over the entire flushing period. The integral Cdt was evaluated numerically
with a FORTRAN code adapted from Bevington (1969). To calculate the mass added to the cell, the
injection N03" concentration vs. time curve was integrated (Figure 5-5), and then multiplied by the mean
injection rate. To calculate the mass advected across the 180 cm bgs plane (extracted mass), the NO.;
BTCs from each of the ports at this depth were integrated with respect to time, and then a mean value for
this integral was calculated. This quantity was then multiplied by the mean injection rate to calculate the
NOs mass. Equation 5-2 was used to calculate the propagated error in the mass estimate. As shown in
Table 5-4, 7,130±390 g were pumped into the cell during the 174-day flushing experiment, and
6,240±1,160 g were removed. By difference, the calculated mass loss for this period was 890±1,220 g,
where the propagated error, expressed as a standard deviation, was calculated from shtoss = s„w + s2MajT
(Bevington, 1969). The large error in the mass-loss estimate arises from the substantial spatial variability
associated with the amount of mass flushed from the cell. During the following 24-day flushing
experiment (mean injection concentration 116.4±10.2 mg/L (n=5)), an additional 802±70 g N03" was
pumped into the cell, for a total injected mass of approximately 7.9 kg N03. Because of sparse data, the
mass extracted from the cell during this period was not calculated, but concentrations at the 180-cm depth
were similar to those measured during the 174-day experiment.
Based on the mass balance there was a small utilization of N03 under flushing conditions. The
estimated mass loss of 890 g represents only 12% of the N03" pumped into the cell over the 174-day
82
-------
flushing experiment. Because there was substantial experimental error, it is possible that the calculated
NOg loss was an experimental artifact. The frequent detection of N02\ however, suggests that at least
some of the observed N03 utilization was real (i.e., due to biological N03" reduction).
5.2.3 BTEXTMB
The goal of the mass balance was to estimate the extent of aromatic-hydrocarbon mass loss in each
treatment cell. BTEXTMB mass losses were calculated by comparing the initial mass of a constituent in
the gasoline to the sum of the mass flushed by advection plus the mass remaining in the gasoline-
contaminated zone. The sorbed mass was negligible relative to the other terms in the mass-balance
equation, and therefore, was not explicitly considered here. Because it was not feasible to core the cells
repeatedly, cores were not collected prior to the flushing experiment to determine the initial mass. This
mass balance therefore represents total losses that were incurred over the entire 19-month period from
the gasoline spill to final core collection. Volatilization losses were minimized by cell construction, but any
losses by this mechanism would be included in the ML0SS term. Total mass loss was obtained by
difference,
Mloss=M,-(Ma+M,) 5-4
where,
MLOss - Mass loss over time period of interest,
M, = Initial mass in gasoline,
Ma = Mass removed from flushing, and
M, = Mass remaining in the gasoline-contaminated zone.
For a given constituent, the initial mass, M^ in 70 liters of fresh gasoline was calculated from weight
fraction of the constituent and the density of the gasoline (see Table A-2). As discussed in Appendix A,
the weight-fraction data were found to be representative of the composition of fresh API 91-01. Initial
masses are shown in Table 5-5. Relative to other components of the mass balance, the uncertainty in the
initial mass was probably small.
Estimates of the mass flushed from the cells were obtained for both the 174-day and the following 24-
day flushing experiments. These masses were then added to obtain total flushed mass, Ma, for each
constituent (Table 5-5). To determine the mass advected across the 180-cm bgs plane during the 174-
day experiment (extracted mass), the approach described above for NO,; was followed. It should be
noted that data points were added to each of the piezometer BTCs to close some existing gaps. Because
samples were not collected at 180 cm until day 16, concentrations from this sampling event were
extrapolated to day 7. An additional data point was also added to Control-Cell BTCs; concentrations from
day 155, the last sampling event in this cell, were extrapolated to day 174. These changes were
Table 5-4. Nitrate Mass Balance Results for 174-Day Flushing Experiment.
Integrated Injection Hate
Breakthrough
Curve Mass
(mg/L*d) (s.d.)1 (L/d) (s.d.) (g) (s.d.)
INJECTED 19791 A2 360 (20.0) 7,130 (390}
EXTRACTED 17.323.43 (3,075.7) 360 (20.0) 6,240 (1,160)
MASS LOSS 890 (1,220)
MASS RECOVERED (%) 88
1 s.d, - standard deviation.
2 injection concentration vs. time record.
3 Mean of integrated breakthrough curves from 180-cm ports.
83
-------
Table 5-5. Aromatic Hydrocarbon Mass Balance
CONTROL CELL
(mass in grams)
Total
Initial Mass Initial Mass
: Initial Mass
Initial
Residual
Flushed Mass
Recovered
Mass
Recovered
Flushed
Remaining in
Mass
Mass
1996
(s.d.)1
1997
(s.d.)
Mass
Loss
(%)
(%)
Gasoline (%)
Benzene
640
80
342
(420)
3
(0.8)
425
216
66
54
13
Toluene
4,000
1,460
1,620
(696)
136
(12)
3,216
784
80
44
37 '
Ethylbenzene
1,760
1,030
385
(41)
35
(1)
1,450
310
82
24
59
m+p - Xylene
3,880
2,330
820
(93)
75
(2)
3,225
655
83
23
60
o-Xylene
1,380
810
339
(39)
30
(2)
1,179
201
85
27
59
135 TMB
570
430
33
(4)
3
(0.1)
466
105
82
6
75
124 TMB
1,760
1,150
111
(15)
11
(0.4)
1,272
488
72
7
65
123 TMB
360
260
32
(4)
3
(0.08)
295
65
82
10
72
Naphthalene
280
240
46
(6)
3
(0.9)
289
-9
103
17
86
Total BTEXTMB
14,630
7,760
3,727
(1,081)
298
(19)
11,785
2,845
81
28
53
NITRATE CELL
(mass in grams)
Total
Initial Mass Initial Mass
Initial Mass
Initial
Residual
Flushed Mass
Recovered
Mass
Recovered
Flushed
Remaining in
Mass
Mass
1996
(s.d.)
1997
(s.d.)
Mass
Loss
(%)
<%)
Gasoline (%)
Benzene
640
80
540
(431)
4
(0.3)
624
16
98
85
13
Toluene
4,000
1,410
2,166
(937)
136
(9)
3,712
288
93
58
35
Ethylbenzene
1,760
940
403
(30)
36
(0.6)
1,379
381
78
25
53
m+p - Xylene
3,880
2,180
883
(88)
80
(1)
3,143
737
81
25
56
o -Xylene
1,380
750
370
(28)
32
(0.2)
1,152
228
83
29
54
135 TMB
570
410
36
(6)
3
(0.2)
449
121
79
7
72
124 TMB
1,760
1,090
117
(10)
11
(0.6)
1,218
542
69
7
62
123 TMB
360
240
33
(3}
3
(0.2)
276
84
77
10
67
Naphthalene
280
210
49
(5)
4
(1)
263
17
94
19
75
Total BTEXTMB
14,630
7,270
4,598 (1,248)
309
(8)
12,177
2,454
83
34
50
1 s.d. - standard deviation.
supported by the tracer test, which showed sharp arrival of fronts at 180 cm within seven days, and the
extraction-welt BTCs, which confirmed that extrapolated concentrations were reasonable. To calculate
the small additional mass removed during the 24-day pumping period, fewer data were available, and the
simpler approach described in Section 5.2.1 for dissolved 02 was followed. For each constituent, a mean
concentration was calculated from extraction-well samples (n=3). This mean was then multiplied by the
injection rate (200 ml/min) to obtain a mass estimate, and the error was again calculated from
Equation 5-2. Proportionally the largest errors were associated with benzene and toluene, which had
large spatial variations in breakthrough behavior.
The mass remaining in the cells, Mr, was estimated from concentrations in core-extract samples. To
obtain a mass estimate for the entire cell from discrete samples, an approach similar in principle to
Freyberg (1986) was followed. As discussed in detail in Freyberg (1986), an estimate of the mass in a
volume of interest can be obtained by integrating the concentration distribution:
84
-------
yi* >
Mr = J jCz(x,y)dxdy 5.5
y* .
where C is obtained from
z
Cz = lC(x,y!z)dz 5-6
Zi
The concentration, Cz, is expressed above in terms of mass per unit volume of porous medium. The
limits of integration in Equation 5-5 correspond to the ceil dimensions (2m by 2m), and in Equation 5-6 to
the vertical length of the core.
The mass of each constituent was determined by first transforming concentrations in mg/kg to g/m3
aquifer material using a wet bulk density of 2.15 g/cm3, calculated from the well-characterized dry bulk
density of Borden sand of 1.82 g/cm3 (Ball et al., 1990), and a fully-saturated porosity of 0.33 (Mackay et
at., 1986). At each core location, concentration profiles (Figures 5-13 and 5-14) were then vertically
integrated by multiplying each concentration by the appropriate depth interval, Az, yielding an integrated
concentration, C2, in g/m2. Kriging was then used to interpolate vertically-integrated data from all of the
core locations within a cell onto a regular grid. A linear model of the observed variogram was used to
determine the weighting factors in the kriging matrix. Concentration contours of the kriged grid data
agreed very well with hand-drawn contours based on linear interpolations between core locations. The
GEOSOFT Mapping and Processing System (Geosoft Inc., 1994) was used to calculate the variogram
and krige the data. An areal integration of the kriged surface (concentration in g/m2 x area in m2) was
then performed with GEOSOFT to obtain the total mass of the constituent in grams.
Results are summarized in the second column of Table 5-5. Several assumptions were made in
arriving at these residual mass estimates. Because core recoveries were less than 100% of the core run,
it was necessary to assign depth intervals to the recovered aquifer material. The calculations in Table 5-5
assume that compaction of the aquifer material inside the core barrel was minimal (reasonable for a sand
deposit), and that, because of increasing frictional resistance inside the core barrel as the run proceeded,
the missing interval was from the bottom of the core run. Consequently, a core from ground surface with
100 cm recovery was taken to represent the 0 to 100 cm depth interval. Second, the magnitude of the
wet bulk density (2.15 g/cm3) was calculated on the assumption that the sub-samples were fully saturated
with water. On the basis of the rapid extraction and capping of the cores, and visual observations while
coring, the assumption of full saturation appeared reasonable. Calculations were also simplified by
treating the bulk density as a spatially-uniform parameter, and ignoring the small effect of residual
gasoline on the magnitude of the bulk density. The value of the bulk density used here may have resulted
in a systematic overestimation or underestimation of the mass estimate. Although the net bias would
affect the absolute magnitude of the mass loss, it would apply to both cells, and relative differences
between cells could still be determined.
In addition, the mass calculation required assigning numerical values to the samples with
concentrations below method detection limits. A sensitivity analysis showed that, with the exception of
benzene and naphthalene, calculations were not sensitive to the value used for the detection limit; mass
estimates varied by less than one percent when concentrations in not-detected samples were varied from
zero to the value of the detection limit. For benzene and naphthalene, which had higher proportions of
samples below detection limits, mass estimates increased by more than 10% when the detection-limit-
value was used in the calculation. However, in the absence of a gasoline phase, the detection-limit value
overestimated the total mass that would be present in the sorbed and dissolved phases. Because most of
the samples were from locations that did not contain gasoline, concentrations of all constituents in
samples below detection limits were assumed to be zero in the calculation of residual mass presented
here. This may have resulted in a small underestimation of the residual mass. Finally, it was necessary
to estimate the lengths of the contaminated intervals of the lowest samples in cores 4L and 3K. These
cores did not straddle the contaminated zone, so the base of the lowest interval was undefined. In both
cells, the lengths were chosen to be consistent with adjacent cores, 4K and 3J, which had similar vertical
concentration profiles (Figures 5-13 and 5-14).
85
-------
These assumptions and process of creating smooth concentration distributions from discrete data
probably lead to substantial uncertainty in the residual mass estimates. It was not possible, however, to
determine the magnitude of this uncertainty quantitatively. Accordingly, the error in the total amount of
recovered mass is not given on Table 5-5, but based on error propagation considerations, its magnitude is
at least as large as the uncertainty in the flushed mass. These mass balance calculations show therefore
that even under highly-controlled experimental conditions, the experimental variability and resulting
uncertainty in the estimates of aromatic-hydrocarbon and electron-acceptor utilization can be difficult to
control.
Nonetheless, with due consideration of the uncertainty associated with these field data, the mass
balance provided reasonable qualitative results. As a percentage of initial mass, the gasoline phase in
both cells was relatively depleted in the most soluble compounds such as benzene and toluene (despite
the high initial mass of toluene in API 91-01), and the majority of relatively insoluble compounds such as
naphthalene remained in the residual mass fraction. Similarly, the percentage of initial mass flushed from
the cells corresponded fairly well to the effective solubility of the compound. However, there were
differences between aqueous concentrations at the end of the experiment and the mass remaining in the
residual phase for most aromatic compounds. For example, as shown on Figure 5-8, by the end of the
experiment, aqueous concentrations of benzene, toluene, ethylbenzene, total xylenes, and total
trimethylbenzenes in the Nitrate Cell had dropped to 1.5%, 39%, 82%, 85%, and 92% of initial gasoline-
saturated values measured in the laboratory, respectively, while masses remaining in the gasoline phase
declined to 13%, 35%, 53%, 56%, and 65% of initial masses. Assuming equilibrium partitioning, which is
likely in this experimental system, the changes in aqueous concentration are directly proportional to
changes in mole fraction, and should be roughly equal to changes in residual mass. The observed
differences may reflect negative bias in the estimation of residual mass. A dissolution model is currently
being developed to evaluate quantitatively the relationship between aqueous concentrations and masses
in the residual gasoline phase.
In general, more mass was flushed from the Nitrate Cell than the Control Cell, and residual masses
were slightly higher in the Control Cell. Total recovered masses ranged from a low of 66% (benzene,
Control Cell) to a high of 103% (naphthalene, Control Cell). In terms of total BTEXTMB, recovered
masses were 81% (Control) and 83% (Nitrate) of initial mass.
It was not possible to conclude that mass loss was enhanced in one cell relative to the other. This
conclusion could be made only if a large difference between cells was observed. For instance, if depletion
of a labile compound such as toluene was much more extensive in the Nitrate Cell than the Control Cell,
then this could reasonably be attributed to biological activity. Moreover, it is important to note that this
approach does not provide a direct estimate of biotransformation mass loss. The mass loss term, ML_pS,
incorporates all mass removal mechanisms, which may include losses from volatilization and cell flooding.
Therefore, even though benzene and toluene mass losses were larger in the Control Cell, the differences
between cells could conceivably have resulted from abiotic processes.
5.3 Discussion and Conclusions
Despite the uncertainty associated with the low-concentration 02 data, Figure 5-1 clearly shows rapid
depletion of the majority of the 02 injected into the treatment cells. After about 15 days in the Nitrate Cell
and 60 days in the Control Cell, groundwater collected from the pea-gravel layer (60-em bgs sampling
ports) was typically at the threshold 02 concentration; depletion therefore occurred within hours of
injection. Downward migration of dissolved 02 probably did not exceed ca. 10 cm. The same pattern was
observed when the injection dissolved-02 concentration was increased to approximately 5 mg/L during the
24-day flushing period the following spring. These results are consistent with the rapid uptake of
dissolved 02 in laboratory microcosms that contained contaminated aquifer material. In those
experiments aromatic hydrocarbon losses were evident under fully-aerobic conditions. The extent of
microaerophilic 02 utilization by other gasoline hydrocarbons or abiotic reactions is not known.
In contrast to rapid dissolved-02 utilization, N03 uptake was relatively low. Most of the NOa" depletion
appeared to occur initially in the presence of dissolved Or possibly to satisfy an assimilatory nitrogen
requirement (Bazylinski and Blakemore, 1983). Previous studies have shown that the activity of
microorganisms in this aquifer is nitrogen-limited under aerobic conditions (Barbara et al., 1994).
86
-------
Additional N03 utilization also occurred under 02-depleted conditions, but rates were too low to observe
much depletion during the nine-day flushing residence time. The concentration profile at 180 cm showed
that about five months were required to deplete the added N03~ under static conditions (i.e., a zero-order
rate of 0.67 mg N03VL/d). An inhibitory effect associated with the gasoline phase may account for the
lower rate (0.2 mg/L/d) observed during this period at the 60-cm depth interval. Toxicity effects in a
microbial community adapted to hydrocarbons have been observed in other studies. Dibble and Bartha
(1979) observed declining microbial activity with increasing concentrations of oil sludge. Their microbial
community was already adapted to the hydrocarbon mixture; toxicity was manifested as a lower rate of
substrate utilization. On the other hand, these rates are similar to the rate observed by Barbara et al.
(1992) during biotransformation of aromatic hydrocarbons and possibly other carbon sources in the
Borden landfill leachate plume, and the laboratory experiments performed in this study in which N03
uptake was observed. On the basis of the lack of response to the addition of MBH medium, it appears
that anaerobic, denitrifying activity was not limited by inorganic nutrients.
Given the low rate of utilization and the nine-day residence time in the cell, NO -reducing reactions in
the absence of dissolved Oa did not appear to be important within the Nitrate Cell. However, to evaluate
the potential of nitrate-based biotransformation in our experimental system, the specific denitrifying
pathway was of some interest. The production of N02 and apparent lack of other terminal electron-
accepting processes (e.g., methanogenesis) in the treatment cell suggested that NO.; was utilized
primarily as an electron-acceptor in a dissimilatory reaction, rather than as a nitrogen source for cell
protein. The specific denitrifying pathway, denitrification to N20 or nitrogen gas or dissimilatory reduction
of N03 to ammonium (NH4*) (DRNA), was not defined in situ. Denitrification is generally assumed to be
the predominant N03-reducing mechanism in carbon-limited environments, while DRNA may be favored
under N03-limited conditions (i.e., in carbon-rich, anoxic environments) (Korom, 1992; Tiedje et al.,
1982). DRNA has been identified as an important mechanism in contaminated aquifers (Bulger et al.,
1989), but there are no reported cases of aromatic hydrocarbon utilization by this pathway.
The major electron-donor was assumed to be carbon, but autotrophic microbial denitrification coupled
to the reduction of sulfide minerals such as pyrite (Aravena and Robertson, 1998) and reduced Fe
(Korom, 1992; Postma et al., 1991) have been reported. Sulfide minerals have not been identified in the
Borden aquifer, but field observations indicate that greater quantities of reduced sulfur are present in
proximity to the peat layer. Assuming carbon was the major donor in the Nitrate Cell, both the laboratory
and field results indicated that other carbon sources were utilized in preference to the aromatic
hydrocarbons under NO "-reducing conditions in this highly-contaminated experimental system. This
conclusion is supported by the laboratory experiments performed in this study, and is consistent with other
experiments using hydrocarbon-contaminated aquifer material (Hutchins et al., 1991a).
The mass balance results and hypothetical reactions were used to estimate the maximum
mineralization that could occur under aerobic conditions. If the observed dissolved 02 threshold
concentration was representative of in situ conditions, then only about 50-75% of the 02 supplied to the
aquifer for microaerophilic reactions was utilized. Even if all of the available 02 was utilized, however, the
mass of 02 added to the cells was probably insufficient to stimulate an observable removal of recalcitrant
compounds such as benzene. For example, if aerobic benzene biotransformation is assumed to proceed
completely to C02 with no assimilation of C by microbial cells, the mineralization reaction is
C6H6 + 7.5 02 6 C02 + 3 H20 5-7
From the stoichiometry of this reaction, 3.1g of 02 are required to oxidize 1g of benzene. Assuming
that all of the injected and residual 0.» (ca. 300 g) in a given cell was used to mineralize benzene, then
97 grams or 15% of the 640 grams of benzene added to the cell would be removed. This is clearly an
overestimate because O would likely be utilized in the oxidation of other carbon compounds as well,
including other aromatic hydrocarbons (Chapter 4). More realistically, if only a fraction (for example, one
quarter) of the available O was utilized for benzene oxidation, then only 24 grams of benzene would be
removed, which is well within the experimental error of the mass balance. Assuming all of the available
Oa was consumed by the aromatic hydrocarbons (i.e., none was consumed in reactions with other
gasoline hydrocarbons), 300 g 02 would oxidize only 95 g total BTEXTMB to C02. Based on these
assumed reactions, the calculated mass loss was very small relative to the initial mass in a given cell.
87
-------
Therefore, if the addition of microaerophilic O, did contribute to a reduction in the flux of soluble
compounds from the source area, the effect would have been very small in relation to the size of the
carbon pool.
The NO, mass balance was used in a similar manner to estimate the maximum mass loss from
mineralization that could occur under NOs-reducing conditions. Although most of the NO,; consumed in
the N03 -amended cell may have been utilized as an N source during aerobic biotransformation, and
aromatic-hydrocarbon uptake under N03-reducing conditions appeared low, some biotransformation of the
more labile compounds may have occurred. These losses would have been difficult to detect in the
experimental system. For example, assuming toluene is mineralized to CO, by a denitrification reaction
and there is no C or N assimilation by microbial cells (Equation 4-3), 446.4 g NO - are required to oxidize
92.1 g toluene. Stoichiometries of the other NO„-reducing reactions (NO, or N.,0 end products) that may
have occurred yield similar molar ratios. From Equation 5-8, the 890 g of N03~ that were utilized over the
174-day flushing experiment {Table 5-4) correspond to about 180 g toluene, or 7.8 % of the flushed mass
(2300 g) and 4.5 % of the total mass (4,000 g) of toluene in the cell. This calculation overestimates
toluene removal because it is based on the assumptions that all of the NOs" was consumed under
anaerobic conditions, with toluene as the sole substrate. Nonetheless, the calculated mass loss is still
small relative to the size of the toluene pool, and would have been within the uncertainty of the toluene
mass balance and difficult to detect.
The detection of metabolites demonstrated that some partial oxidation of gasoline hydrocarbons was
occurring in both cells. A transformation to a partially-oxidized intermediate, such as benzoic acid,
requires less 02 per mole of parent compound oxidized. It is conceivable, therefore, that losses of
aromatic hydrocarbons from biotransformation were more extensive than indicated by the mineralization
reactions. If partial oxidation was occurring, reaction stoichiometries show that mass loss of a given
constituent could have been substantial relative to the total mass in the cell. For example, assuming no
assimilation of C and N by microbial cells, the following reaction controls the partial oxidation of benzene
to phenol (Cozzarelli et al., 1990),
2 C6H6 02 > 2 Cc H60 5 8
Equation 5-8 shows that one gram of 0? would partially oxidize 4.9 g benzene, and if it is assumed
that the maximum benefit was gained from microaerophilic 02 addition (i.e., 100% partial oxidation of
benzene), 300 g of 02 would be sufficient to partially oxidize 1470 g benzene, which exceeds the total
benzene mass in the cell. More realistically, if it is arbitrarily assumed that only 25% of the total O., mass
was consumed for benzene degradation, then 75 g would be available for reaction, which corresponds to
a transformation of 367 g benzene to phenol, or approximately 50 % of the initial benzene mass. Using
similar reasoning, if toluene was transformed to benzoic acid under NO., -reducing conditions (Kuhn et al.,
1988), then partial oxidation of toluene could produce substantial toluene losses and concentrations of
benzoic acid in the mg/L range.
These calculations suggest that if significant partial oxidation had occurred, biotransformation would
clearly have been observable in the treatment cells. However, the low Mg/L concentrations of partially-
oxidized compounds suggest that these hypothetical reactions were only of minor importance.
Alternatively, the metabolites may not have been persistent, in which case additional oxidant would have
been consumed to complete the reaction, and the stoichiometries of the mineralization reactions
discussed above would be predominant. Regardless of the extent of mineralization, the low consumption
of NO? and the presence of multiple carbon sources in combination with a limited mass of Oa indicate that
mass losses of aromatic hydrocarbons from biotransformation were low, and probably much less than the
ca. 2500 g total BTEXTMB mass losses obtained from the mass balances (Table 5-5). These
observations were confirmed by the followup microcosm experiment. The large mass losses from the
mass balance probably resulted from a combination of physical losses and experimental error, with only a
minor contribution from biotransformation.
Although there was evidence from metabolite formation and electron-acceptor uptake that in situ
biotransformation had been stimulated in the gasoline-contaminated zone, the extent of biotransformation
of the aromatic hydrocarbons was too limited to significantly affect the formation of a plume. To
understand the net result of flushing fluid through a NAPL, consideration must be given to the complex
88
-------
interactions between dissolution and chemical reactions. Flushing mixes reactants (in this instance,
electron acceptors) into the source and associated plume, which is required to enhance mass removal,
but also increases the advective flux of dissolved constituents into the aquifer. If the biotransformation
rate is low, soluble compounds may be advected from the residually-saturated zone by flushing before
significant biotransformation occurs (Malone et al., 1993). Conversely, if the biotransformation rate is fast
relative to the dissolution rate, biotransformation can enhance the dissolution rate by increasing the
concentration gradient which is the driving force for dissolution, as well as reduce the mass being
advected into the plume (Seagren et al., 1993, 1994). In our experimental system, gasoline dissolution
was the dominant mass removal mechanism. Concentrations flushed from the cells were consistent with
equilibrium partitioning between the gasoline and water phases. Biotransformation appeared therefore to
be too minimal to appreciably lower the flux of soluble constituents from the gasoline contaminated zone.
89
-------
Intentionally Blank Page
90
-------
6.0 CONCLUSIONS AND IMPLICATIONS
Generally the findings on the degradability of aromatic hydrocarbons in the presence of
microaerophiiic 0, and N03 were consistent between laboratory and field experiments: mass losses were
minor in the Nitrate Cell as well as in both laboratory microcosm experiments in which conditions similar to
the field were established (i.e., BTEXTMB concentrations near gasoline-saturated concentrations and
microaerophiiic / N03 conditions (Experiments 4 and 6)), As discussed in Chapter 5, however,
ascertaining small mass losses in the field was difficult because of the size of the carbon pool and the
uncertainty associated with the mass balance. The apparent lack of extensive mass losses in the field
was also consistent with the activity and biomass assays performed in the laboratory: Despite strong 02
consumption when available, numbers and activity in aquifer materia! extracted from the treatment cells
were not indicative of a large hydrocarbon-degrading population, in both the field and the laboratory
microcosm experiment with pre-exposed aquifer material (Experiment 6), N03 reduction was observed
over time periods on the order of 100 days, but accompanying utilization of labile aromatic compounds
was not apparent. The use of nitrate solely as an electron acceptor was equivocal; utilization of some
N03 as an N source may have occurred in both instances. In both of these experimental systems, the
effect of microaerophiiic 02 was small, although the disparity between static microcosms and dynamic field
conditions makes the comparison tenuous. Perhaps most important for remediation, there was agreement
on the minor benzene losses in the presence of microaerophiiic 02.
Although under certain conditions the extent of mass loss was maximized by the presence of dual
electron acceptors, the bulk of the field and laboratory results indicated that NO " / microaerophiiic-02
based bioremediation was not an effective source-area remedial technology in this aquifer. Rapid uptake
of O,, observed at both laboratory and field scales, demonstrated that aerobic activity was not inhibited
in situ, but mass losses were limited under microaerophiiic conditions by the small quantity of 02 available
for reaction. Based on laboratory results, the addition of dissolved 02 may have led to oxidation of
compounds that otherwise would have been recalcitrant under anaerobic, denitrifying conditions, but the
effect in situ was small relative to the mass of gasoline hydrocarbons in the cells. This suggests that the
partial oxidation of parent compounds by microaerophiiic 02 was a relatively unimportant process.
Consequently, the majority of the mass of recalcitrant compounds {e.g., benzene) was flushed into the
aquifer. Moreover, N03" reduction was slow relative to the residence time in the source area, and did not
appear to be associated with aromatic-hydrocarbon utilization in the presence of other gasoline
constituents. A large denitrifying population capable of rapid aromatic-hydrocarbon biotransformation did
not develop in the treatment cell in response to extended exposure to abundant substrate and NO,.
Consequently, nitrate-based bioremediation was ineffective for source-area remediation in this
experimental system. It is not clear why this population did not develop under the conditions established
in the field. Given these observations and the results of the mass balance, there were no apparent
advantages associated with the microaerophiiic / N03 treatment relative to the unremediated control.
Rapid 02 uptake and aromatic-hydrocarbon utilization in gasoline-contaminated aquifer material
(Experiment 6) suggests that a more conventional injection fluid with NQ3" and air-equilibrated Oz
(7-10 mg/L) would have been more beneficial in this system with a similar level of effort.
These conclusions pertain to specific experimental system evaluated in this study (i.e., a recent
gasoline spill flushed for a relatively short period of time and monitored over a short flow path). It is
conceivable that this approach, particularly with respect to the effects of microaerophiiic 02, would be
more effective during the latter stages of an enhanced bioremediation project when source-area
concentrations were lower. Similarly, although N03 utilization was minor over the flow path evaluated
here, acclimation resulting in the development of a substantial population capable of degrading TEX may
91
-------
have occurred with continued exposure. Existing studies have indicated, however, that N03" uptake rates
are generally slow relative to 02 (Section 1.3.2). This suggests that a longer residence time would be
needed to fully evaluate the effects of NOs addition. Based on existing data for this aquifer (e.g., Barbara
et al., 1992), it is possible that substantial NO.; utilization would have occurred further downgradient
(beyond our experimental system) in the anaerobic core of the plume, providing a benefit to an
engineered or intrinsic remediation strategy.
Finally, the results reported here are specific to the Borden aquifer; generalization of the results to
other sites is inadvisable without verification from site-specific testing. The comparison of aromatic
hydrocarbon mass loss in both pristine and exposed Borden sediment with sediment from other
petroleum-hydrocarbon-contaminated sites revealed a substantial range in the capabilities of the
indigenous denitrifying populations. Similarly, although microaerophiiic 02was not found to be beneficial
in our experimental system, the contribution to in situ mass loss could be significant in other
circumstances (e.g., lower-carbon environments). For example, mixed electron acceptors could
conceivably be effective for downgradient plume control using a reactive wall or other semi-passive
remedial technology.
92
-------
REFERENCES
Anid, P.J., P.J.J. Alvarez, and T.M. Vogel. 1993. Biodegradation of monoaromatic hydrocarbons in
aquifer columns amended with hydrogen peroxide and nitrate. Water Res. 27(4): 685-691.
APHA (American Public Health Association). 1985. Standard Methods for the Examination of Water and
Wastewater. 16th Ed. APHA, Washington, D.C., 1268 p.
Aravena, R. and W.D. Robertson. 1998. Use of multiple isotope tracers to evaluate denitrification in
groundwater: Study of nitrate from a large-flux septic system plume. Ground Water. 36(6):975-982.
Ball, W.P., C.H. Buehler, T.C. Harmon, D.M. Mackay, and P.V. Roberts. 1990. Characterization of a
sandy aquifer material at the grain scale. J. Contam. Hydrol. 5:253-295.
Ball, W.P., G. Xia, D.P. Durfee, R.D. Wilson, M.J. Brown, and D.M Mackay. 1997. Hot methanol
extraction for the analysis of volatile organic chemicals in subsurface core samples from Dover Air
Force Base, Delaware. Ground Water Monit. Rev. 17(1 ):104-121.
Barcelona, M.J., J. Lu, and D.M. Tomczak. 1995. Organic acid derivatization techniques applied to
petroleum hydrocarbon transformations in subsurface environments. Ground Water Monit. Rem.
15(2):114-124.
Barbara, S.E., H. Albrechtsen, B.K. Jensen, G.I. Mayfield, and J.F. Barker. 1994. Relationships between
aquifer properties and microbial populations in the Borden aquifer. Geomicrobiol. J. 12:203-219.
Barbaro, J.R., J.F. Barker, L.A. Lemon, and C.I. Mayfield. 1992. Biotransformation of BTEX under
anaerobic, denitrifying conditions: Field and laboratory observations. J. Contam. Hydrol. 11:245-272.
Barker, J.F., G.C. Patrick, and D. Major. 1987. Natural attenuation of aromatic hydrocarbons in a shallow
sand aquifer. Ground Water Monit. Rev. 7:64-71.
Battermann, G., R. Fried, M. Meier-Lohr, and P. Werner. 1994. Application of nitrate as electron
acceptor at an in situ bioremediation of an abandoned refinery site: pilot study and large-scale
operation. In: R.E. Hinchee et al. (Eds.), Hydrocarbon Remediation. Lewis Publishers, Boca Raton,
FL. pp. 93-99.
Battermann, G, and M. Meier-Lohr. 1995. Nitrate as electron acceptor in in situ abandoned refinery site
bioremediation. In: Hinchee, J.A. Kittel, and H.J. Reisinger. (Eds.), Applied Bioremediation of
Hydrocarbons. Vol. 3(6):155-164, Proceedings of the Third International In Situ and On-Site
Bioreclamation Symposium, Battelle Press, Columbus, OH.
Bazylinski, D.A., and R.P. Blakemore. 1983. Denitrification and assimilatory nitrate reduction in
Aquaspirillum magnetotacticum. Appl. Environ. Microbiol. 46(5): 1118-1124.
Beller, H.R., D. Grbic-Galic, and M. Reinhard. 1992. Microbial degradation of toluene under sulfate-
reducing conditions and the influence of iron on the process. Appl. Environ. Microbiol. 58:786-793.
Beloin, R.M., J.L. Sinclair, and W.C. Ghiorse. 1988. Distribution and activity of microorganisms in
subsurface sediments of a pristine study site in Oklahoma. Microb. Ecol. 16:85-97.
Berry-Spark, K.L., J.F. Barker, D. Major, and C.I. Mayfield. 1986. Remediation of gasoline-contaminated
ground-waters: a controlled experiment. In: Proceedings of Petroleum Hydrocarbons and Organic
Chemicals in Ground Water: Prevention, Detection and Restoration. NWWA/API. Water Well Journal
Publishing, Dublin, OH. pp. 613 623.
Bevington, P.R. 1969. Data Reduction and Error Analysis for the Physical Sciences. McGraw-Hill, Inc.
336 pp.
Bonin, P. and M. Gilewicz. 1991. A direct demonstration of "co respiration" of oxygen and nitrogen
oxides by Pseudomonas nautica: some spectral and kinetic properties of the respiratory components.
FEMS Microbiol. Lett. 80:183-188.
Borden, R.C. and C.M. Kao. 1992. Evaluation of groundwater extraction for remediation of petroleum-
contaminated aquifers. Water Environ. Res. 64(1):28-36.
93
-------
Britton, L, N, 1989. Aerobic denitrification as an innovative method for in-situ biological remediation of
contaminated subsurface sites. Final report ESL-TR-88-40. Submitted to Air Force Engineering
Services Center, Tyndall AFB, FL. 82 pp.
Brookman, G.T., M. Flanagan, and J.O. Kebe. 1985. Laboratory studies on solubilities of petroleum
hydrocarbons in groundwater. TRC Environmental Consultants, project no. 2663-N31-00. Prepared
for: Environmental Affairs Department, American Petroleum Institute, Washington, D.C.
Bulger, P.R., A.E. Kehew, and R.A. Nelson. 1989. Dissimilatory nitrate reduction in a waste-water
contaminated aquifer. Groundwater. 27(5):664-671.
Buresh, R.J. and J.T. Moraghan. 1976. Chemical reduction of nitrate by ferrous iron. J. Environ. Qual.
5(3):320-325.
Butler, B.J., S.E. Barbara, F.H. Crocker, and C.I. Mayfield. 1997. Characterization of microbial
populations of the Borden aquifer. Geomicrobiol. J. 14:253-268.
Carter, J.P., Y.H. Hsiao, S. Spiro, and D.J. Richardson. 1995. Soil and sediment bacteria capable of
aerobic nitrate respiration. Appl. Environ. Microbiol. 61(8):2852-2858.
Chapelle, F.H. 1993. Ground Water Microbiology and Geochemistry. John Wiley and Sons, New York.
424 pp.
Chatzis, I., N. Morrow, and H.T. Lim. 1983. Magnitude and detailed structure of residual oil saturation.
SPEJ Soc. Pet. Eng. J. 23:311-326.
Chiang, C.Y., E.Y. Chai, J.P. Salanitro, and J.D. Cothart. 1987. Effects of dissolved oxygen on the
biodegradation of BTX in a sandy aquifer. Proc. Natl. Water Well Assoc.-Am. Pet. Inst. Conf. on
Petroleum Hydrocarbons and Organic Chemicals in Ground Water-Prevention, Detection, and
Restoration, Houston, Texas, Nov. 17-19, 1987, pp. 451 461.
Cozzarelli, I.M., R.P. Eganhouse, and M.J. Baedecker. 1990. Transformation of monoaromatic
hydrocarbons to organic acids in anoxic groundwater environments. Environ. Geol. Water Sci.
16:135-141.
Cozzarelli, I.M. M.J. Baedecker, R.P. Eganhouse, and D.F. Goerlitz. 1994. The geochenrca! evolution of
low-molecular-weight organic acids derived from the degradation of petroleum contaminants in
groundwater. Geochim. Cosmochim. Acta. 58(2):863 877.
Cozzarelli, I.M., J.S. Herman, M.J. Baedecker. 1995. Fate of microbial metabolites of hydrocarbons in a
coastal plain aquifer: The role of electron acceptors. Environ. Sci. Technol. 29{2):458-469.
Devlin, J.F., and J.F. Barker. 1996. Field investigation of nutrient pulse mixing in an in situ biostimulation
experiment. Water Resour. Res. 32(9):2869-2877.
Dibble, J.T., and R. Bartha. 1979. Effect of environmental parameters on the biodegradation of oil
sludge. Appl. Environ. Microbiol. 37(4):729-739.
Drever.J.I. 1992. The Geochemistry of Natural Waters. Prentice-Hall, New Jersey, 388 pp.
Eganhouse, R.P., T.F. Dorsey, C.S. Phinney. and A.M. Westcott. 1996. Processes affecting the fate of
monoaromatic hydrocarbons in an aquifer contaminated by crude oil. Environ. Sci. Technol.
30(11):3304-3312.
Evans, P.J., W. Ling, B. Goldschmidt, E.R. Ritter, and L.Y. Young. 1992. Metabolites formed during
anaerobic transformation of toluene and o-xylene and their proposed relationship to the initial steps of
toluene mineralization. Appl. Environ. Microbiol. 58:496-501.
Freyberg, D.L. 1986. A natural gradient injection experiment on solute transport in a sand aquifer, 2.
Spatial moments and the advection and dispersion of nonreactive tracers. Water Resour. Res.
22:2031-2046.
Furukawa, K., J.R. Simon, and A.M. Chakrabarty. 1983. Common induction and regulation of biphenyl,
xylene/toluene, and salicylate metabolism in Pseudomonas paucimobilis. J. Bacteriol. 154:1356-
1362.
GeoSoft Inc. 1994. The GEOSOFT Mapping and Processing System. Version 4.00.02. Geosoft, Inc.,
Toronto, Ontario.
Gilbert, T.W., T.D. Behyner, and H.B. Castaneda. 1982. Determination of dissolved oxygen in natural
and wastewaters. Am. Laboratory. March:119-134.
Grbic-Galic, D. and T.M. Vogel. 1987. Transformation of toluene and benzene by mixed methanogenic
cultures. Appl. Environ. Microbiol. 53:254-260.
94
-------
Harvey, R.W., R.L. Smith, and L. George. 1984. Effect of organic contamination upon microbial
distributions and heterotrophic uptake in a Cape Cod, Mass., aquifer. Appl. Environ. Microbiol.
48(6):1197-1202.
Hillel, D. 1982. Introduction to Soil Physics. Academic Press, Orlando, FL. 364 pp.
Hernandez, D. and J.J. Rowe. 1987. Oxygen regulation of nitrate uptake in denitrifying Pseudomonas
aeruginosa. Appl. Environ. Microbiol. 53(4):745-750.
Hutchins, S.R. 1991a. Biodegradation of monoaromatic hydrocarbons by aquifer microorganisms using
oxygen, nitrate, or nitrous oxide as the terminal electron acceptor. Appl. Environ. Microbiol. 57(8);
2403-2407.
Hutchins, S.R. 1991b. Optimizing BTEX biodegradation under denitrifying conditions. Environ. Toxicol.
Chem. 10:1437-1448.
Hutchins, S.R., G.W. Sewell, D.A. Kovacs, and G.A. Smith. 1991a. Biodegradation of aromatic
hydrocarbons by aquifer microorganisms under denitrifying conditions. Environ. Sci. Techno!. 25:68-
76.
Hutchins, S.R., W.C. Downs, J.T. Wilson, G.B. Smith, D.A. Kovacs, D.D. Fine, R.H. Dourlass, and D.J.
Hendrix. 1991b. Effect of nitrate addition on biorestoration of a fuel-contaminated aquifer: Field
demonstration. Groundwater. 29(4):571-580.
Hutchins, S.R. 1992. Inhibition of alkylbenzene biodegradation under denitrifying conditions by using the
acetylene block technique. Appl. Environ. Microbiol. 58(10):3395-3398.
Hutchins, S R., S.W. Moolenaar, and D.E. Rhodes. 1992. Column studies on BTEX biodegradation
under microaerophilic and denitrifying conditions. J. Haz. and Mater. 32:195-214.
Hutchins, S.R. 1993. Biotransformation and mineralization of alkylbenzenes under denitrifying conditions.
Environ. Tox. Chem. 12:1413-1423.
Hutchins, S.R., J.T. Wilson, and D.H. Kambell. 1995. In situ bioremediation of a pipeline spill using
nitrate as the electron acceptor. In: Hinchee, R.E., J.A. Kittel, and H.J. Reisinger, (Eds.), Applied
Bioremediation of Petroleum Hydrocarbons., Vol. 3(6):143-153, Proceedings of the Third International
In Situ and On-Site Bioreclamation Symposium, Battelle Press, Columbus, OH.
Hutchins, S.R., J.A. Bantle, and E.J. Schrock. 1998. Effect of nitrate-based bioremediation on
contaminant distribution and sediment toxicity - column study. Environ. Tox. Chem. 17(3):349-361.
Johnson, M.J. 1967. Aerobic microbial growth at low oxygen concentrations. J. Bacteriol. 94(1):101 -
108.
Kehew, A.E, and R.N. Passero. 1990. pH and redox buffering mechanisms in a glacial drift aquifer
contaminated by landfill leachate. Ground Water. 28(5):728-737.
Kennedy, L.G. and S.R. Hutchins. 1992. Applied geologic, microbiological, and engineering constraints of
in-situ BTEX bioremediation. Remediation. Winter. 1992/93:83-107.
Korom, S.F. 1992. Natural denitrification in the saturated zone: A review. Water Resour. Res.
28(6) :1657-1668.
Krul, J.M. 1976. Dissimiiatory nitrate and nitrite reduction under aerobic conditions by an aerobically and
anaerobicaily grown Alcaligenes sp. and by activated sludge. J. Appl. Bacteriol. 40:245-260.
Kuhn, E.P., J. Zeyer, P. Eicher, and R.P. Schwartzenbach. 1988. Anaerobic degradation of alkylated
benzenes in denitrifying laboratory aquifer columns. Appl. Environ. Microbiol. 54:490-496.
Larson, R.J., G.G. Clinckemailiie, and L. VanBeile. 1981. Effect of temperature and dissolved oxygen on
biodegradation of nitrilotriacetate. Water Res. 15:615-620.
Leahy, J.G. and R.R. Colwell. 1990. Microbial degradation of hydrocarbons in the environment.
Microbiol. Rev. 54:305-315.
Leahy, J.G. and R.H. Olsen. 1997. Kinetics of toluene degradation by toluene-oxidizing bacteria as a
function of oxygen concentration, and the effect of nitrate. FEMS Microbiol. Ecol. 23:23-30.
Lloyd, D., L. Boddy, and K.J.P. Davies. 1987. Persistence of bacterial denitrification capacity under
aerobic conditions: the rule rather than the exception. FEMS Microbiol. Ecol. 45:185-190.
Lloyd, D. 1993. Aerobic denitrification in soils and sediments: From fallacies to facts. Trend. Ecol; Evoi.
8(10):352-356.
Longbottom, J.E., and J.J. Lichtenberg (Eds.). 1982. Appendix A. In: Methods for Organic Chemical
Analysis of Municipal and Industrial Wastewater. EPA-600/4082-057. U.S.EPA/EMSL, Cincinnati, OH.
95
-------
MacFarlane, D.S., J.A. Cherry, R.W. Gillham, and E.A, Sudicky, 1983. Migration of contaminants at a
landfill: A case study, 1. Groundwater flow and contaminant delineation. In; J.A. Cherry (Guest
Editor), Migration of Contaminants in Groundwater at a Landfill: A Case Study. J, Hydrol. 63:1-29
(special issue).
Mackay, D.M., D.L. Freyberg, P.V. Roberts, and J.A. Cherry. 1986. A natural gradient experiment on
solute transport in a sand aquifer, 1. Approach and overview of plume movement. Water Resour.
Res. 26:207-222.
MacQuarrie, K.G.B., E.G. Sudicky and E.G. Find. 1989. Simulation of biodegradable organic
contaminants in groundwater, 1. Numerical formulation and model calibration. Water Resour. Res.
26:207-222.
Madsen, E.L., J.L. Sinclair, and W.C. Ghiorse. 1991. In situ biodegradation: Microbiological patterns in a
contaminated aquifer. Science. 252:830-833.
Major, D.A., C.I. Mayfield, and J.F. Barker. 1988. Biotransformation of benzene by denitrification in
aquifer sand. Groundwater. 26:8-14.
Malone, D.A., C.M. Kao, and R.C. Borden. 1993. Dissolution and biorestoration of nonaqueous phase
hydrocarbons: Model development and laboratory evaluation. Water Resour. Res. 29(7);2203-2213.
Marin, M, A. Pedregosa, and F. Laborda. 1996. Emulsifier production and microscopical study of
emulsions and biofilms formed by the hydrocarbon-utilizing bacteria Acinetobacter calcoaceticus MM5.
Appl. Microbiol. Biotechnol. 44:660 667.
Mayou, J. 1976. MPN - most probable number. In: Speck, M.L. (Ed.). Compendium of Methods for
Microbiological Examination of Foods. American Public Health Association, Washington, D.C. pp.
152-162.
McMahon, P.B. and F.H. Chapelle. 1991. Microbial production of organic acids in aquitard sediments
and its role in aquifer geochemistry. Nature. 349:233-235.
Mikesell, M.D., J.J. Kukor, and R.H. Olsen. 1993. Metabolic diversity of aromatic hydrocarbon-degrading
bacteria from a petroleum-contaminated aquifer. Biodeg. 4:249-259.
Miller, D.E. and S.R. Hutchins. 1995. Petroleum hydrocarbon biodegradation under mixed denitrifying/
microaerophilic conditions. In: Hinchee, R.E., F.J. Brockman, and C.M. Vogel, (Eds.), Microbial
Processes for Bioremediation, Vol. 3(8):130-136, Proceedings of the Third International In Situ and
On-Site Bioreclamation Symposium, Battelle Press, Columbus, OH.
Millette, D., B.J. Butler, E.O. Frind, Y. Comeau. and R. Samson. 1998. Substrate interaction during
aerobic biodegradation of creosote-related compounds in columns of sandy aquifer material. J.
Contam. Hydrol. 29:165-183.
Morgan, P. and R.J. Watkinson. 1992. Factors limiting the supply and efficiency of nutrient and oxygen
supplements for the in situ biotreatment of contaminated soil and groundwater. Water Res. 26:73-78.
Mueller, J.G., S.E. Lantz, B.O. Blattmann, and P.J. Chapman. 1991. Bench-scale evaluation of
alternative biological treatment processes for the remediation of pentachlorophenol- and creosote-
contaminated materials: solid-phase remediation. Environ. Sci. Technol. 25:1045-1055.
Nwankwor, G.I., R.W. Gillham, G. van der Kamp, and F.F. Akindunni. 1992. Unsaturated and saturated
flow in response to pumping of an unconfined aquifer: Field evidence of delayed drainage. Ground
Water. 30(5):690-700.
O'Leary, K.E., J.F. Barker, and R.W. Gillham. 1995. Remediation of dissolved BTEX through surface
application: a prototype field investigation. Ground Water Monit. Rev. 15(4):99-109.
Oliveira, E. 1997. Ethanol flushing of gasoline residuals - Microscale and field scale experiments. Ph.D.
Thesis. University of Waterloo. 291 pp.
Patureau, D., J. Davison, N. Bernet, and R. Moletta. 1994. Denitrification under various aeration
conditions in Comamonas sp., strain SGLY2. FEMS Microbiol. Ecol. 14:71-78.
Postma, D., C. Boesen, H. Kristiansen, and F. Larsen. 1991, Nitrate reduction in an unconfined sandy
aquifer: Water chemistry, reduction processes, and geochemical modeling. Water Resour. Res.
27(8):2027-2045.
Poulson, M., L. Lemon, and J.F. Barker. 1990. Chemical fate and impact of oxygenates in groundwater:
Solubility of BTEX from gasoline-oxygenate mixtures. Draft report for Groundwater Technical Task
Force, API, Washington, D.C., 30 pp.
96
-------
Reasoner, D.J. and E.E. Geldreich. 1985. A new medium for the enumeration and subculture of bacteria
from potable water. Appl. Environ. Microbiol. 49:1-7.
Reinhard, M., G.D, Hopkins, E. Orwin, S. Shang, and C.A. Lebron. 1995. In situ demonstration of
anaerobic BTEX biodegradation through controlled-release experiments. In: Hinchee, J.A. Kittel, and
H.J. Reisinger, (Eds.), Applied Bioremediation of Hydrocarbons, Voi. 3(6):263-270, Proceedings of the
Third International In Situ and On-Site Bioreclamation Symposium, Battelle Press, Columbus, OH.
Robertson, L.A., and J.G. Kuenen. 1984, Aerobic denitrification: a controversy revived. Arch. Microbiol.
139:351-354.
Seagren, E.A., B.R. Rittmann, and A.J. Valocchi. 1993. Quantitative evaluation of flushing and
biodegradation for enhancing in situ dissolution of nonaqueous-phase liquids. J. Contam. Hydrol.
12:103-132.
Seagren, E.A., B.R. Rittmann, and A.J. Valocchi. 1994. Quantitative evaluation of the enhancement of
NAPL-pool dissolution by flushing and biodegradation. Environ. Sci. Technol. 28:833-839.
Shaler, T.A. and G.M. Klecka. 1986. Effects of dissolved oxygen concentration on biodegradation of 2.4-
Dichlorophenoxyacetic acid. Appl. Environ. Microbiol. 51(5):950 955.
Sheehan, P.J., R.W. Schneiter, T.K.G. Mohr. and R.M. Gersberg. 1988. Bioreclamation of gasoline
contaminated groundwater without oxygen addition. In: Proceedings, Second National Outdoor Action
Conference on Aquifer Restoration, Ground Water Monitoring and Geophysical Methods, NWWA,
Dublin, OH. pp. 183-199.
Song, H., and R. Bartha. 1990. Effects of jet fuel spills on the microbial community of soil. Appl. Environ.
Microbiol. 56(3):646-651.
Sorel, D., J.A. Cherry, and S. Lesage. 1998. In situ vertical circulation column: Containment system for
small-scale DNAPL field experiments. Ground Water Monit. Rev. Winter. 18(1 ):114 125.
Su, J.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.
Sudicky, E.A. 1986. A natural gradient experiment on solute transport in a sand aquifer: spatial
variability of hydraulic conductivity and its role in the dispersion process. Water Resour. Res.
22:2069-2082.
Thomas, A., S.R. Hutchins, P.B. Bedient, C.H. Ward, M. Wiesner, J.A. Bantle. and S. Williams. 1995.
Pilot-scale design for nitrate-based bioremediation of jet fuel. In: Hinchee, R.E., J.A. Kittel, and H.J
Reisinger, (Eds.), Applied Bioremediation of Hydrocarbons, Vol. 3(6), Proceedings of the Third
International In Situ and On-Site Bioreclamation Symposium, Battelle Press, Columbus, OH.
Thurman, E.M. 1985. Organic Geochemistry of Natural Waters. Martinus Nijhoff / Dr W. Junk Publ.
Tiedje, J.M. 1982. Denitrification. In Methods of Soil Analysis, Part 2. Page, A.L., R.H. Miller, and D.R.
Keeney, Eds. Agronomy Monograph No. 9, American Society of Agronomy, Madison, Wl, pp. 1011-
1026.
Tiedje, J.M., A.J. Sextone, D.D. Myrold, and J.A. Robinson. 1982. Denitrification: ecological niches,
competition and survival. Antonie van Leeuwenhoek. 48:569-583.
Tiedje, J.M. 1988. Ecology of denitrification and dissimilatory nitrate reduction to ammonium. In Biology
of Anaerobic Microorganisms. Zehnder, A.J.B., (Ed.), John Wiley and Sons, NY. pp. 179-244.
Trevors, J.T., C.I. Mayfield, and W.E. Innis. 1982. Measurement of electron transport system (ETS)
activity in soil. Microbiol. Ecol. 8:163-168.
Vandegrift, S.A., and D.H. Karripbell. 1988. Gas chromatographic determination of aviation gasoline and
JP-4 jet fuel in subsurface core samples. J. Chromat. Sci. 26:566-569.
Van't Riet, J., A.H. Stouthamer, and R.J. Planta. 1968. Regulation of nitrate assimilation and nitrate
respiration in Aerobacter aerogenes. J. Bacteriol. 96(5): 1455-1464.
Wilson, L.P., N.D. Durant, and E.J. Bouwer. 1995. Aromatic hydrocarbon biotransformation under mixed
oxygen/nitrate electron acceptor conditions. In: Hinchee, R.E., F.J. Brockman, and C M. Vogel,
(Eds.), Microbial Processes for Bioremediation, Vol. 3(8);137-144, Proceedings of the Third
International In Situ and On-Site Bioreclamation Symposium, Battelle Press, Columbus, OH.
Wilson, L.P. and E.J. Bouwer. 1997. Biodegradation of aromatic compounds under mixed oxygen/nitrate
conditions: a review. J. Ind. Microbiol. Biotech. 18:116-130.
Zapico, M.M., S. Vales, and J.A. Cherry. 1987. A wireline piston core barrel for sampling cohesionless
sand and gravel below the water table. Ground Water Monit. Rev. 7:74-82.
97
-------
Intentionally Blank Page
98
-------
APPENDIX A: GASOLINE CHARACTERISTICS
99
-------
APPENDIX A. GASOLINE CHARACTERISTICS
American Petroleum institute (API) gasoline was used in most of the experiments in this study. This
unleaded gasoline, designated 91-01, was used for various toxicological research programs.
Consequently, comprehensive analytical work was done by the API in 1991 when the gasoline was
formulated to define its composition and physical properties (raw data not shown). Although gasoline may
contain more than 1,000 compounds (Brookman et at.. 1985), a relatively small group accounts for most
of the weight. Approximately 150 compounds and generic compound classes were identified in
API 91 -01, of which only 25 account for 65% of the gasoline weight (Table A-1). An accurate
characterization of the gasoline composition was required to complete the field mass balance, and to
predict dissolved aromatic hydrocarbon concentrations. Molar fractions for all identified compounds are
compiled in Table A-1. Basic characteristics of the gasoline, including weight fractions of BTEXTMB, are
summarized in Table A-2.
Several analyses were performed over the course of this study to verify the accuracy of the API
weight fraction data. GC analyses (Appendix C) of known volumes of API 91-01 gasoline in methanol
yielded weight fractions of aromatic hydrocarbons that agreed well with the API weight fractions listed in
Tables A-1 and A-2. These analyses are summarized in Table A-3. All of the samples analyzed by the
Organic Geochemistry Laboratory at the University of Waterloo were taken from an aliquot of API 91-01
stored in laboratory. The August, 1997 sample was submitted and analyzed with the core extract samples
collected from the treatment cells. The August, 1997 samples were also analyzed one year later to obtain
additional gasoline composition data. Percent differences were less than approximately 20% for all
compounds, with the exception of naphthalene (up to 40%).
Aqueous concentrations of BTEXTMB in gasoline-saturated water, measured in laboratory batch
equilibrium tests, agreed we'I with concentrations calculated from Raoult's Law (Table A 4; Figure A-1).
Raoult's Law calculations were based on literature solubilities and the mole fractions listed in Table A-1.
' Calculated concentrations were lower than measured by ca. 20%, possibly because of non-ideal behavior
in the gasoline phase (Borden and Kao, 1992). With the exception of naphthalene which behaved
anomalously, the reasonable agreement between measured and calculated concentrations indicated not
only that the API weight-fraction data were representative, but also that Raoult's Law could be used to
describe equilibrium partitioning between the gasoline and groundwater (Eganhouse et al., 1996).
100
-------
TABLE A-1 Mole Fractions of Identified Compounds in API 91-01 Gasoline. Compound Identification and Weight
Fractions Compiled from an Unpublished 1991 Analysis of the Gasoline Performed by the API1.
Compound
Carbon
Hydrogen
Wt.%
C x Wt%
H x Wt%
Mol. Wt.
(g/mole)
wt%/mol. Wt.
(moles/1 OOg)
Mole Fraction
Propane
3
8
0.01
0.03
0.08
44.10
0.0002
0.00021
Isobutane
4
10
0.14
0.56
1.40
58.12
0.0024
0,00227
n-Butane
4
10
4.88
19.52
48.80
58.12
0.0840
0.07896
1,2-Butadiene
4
6
0.03
0.12
0.18
54.09
0.0006
0.00052
3-Methyl-1 -butene
5
10
0.03
0.15
0.30
70.13
0.0004
0,00040
Isopentane
5
12
4.51
22.55
54.12
72.15
0.0625
0.05879
1-Pentene
5
10
0.63
3.15
6.30
70.13
0.0090
0.00845
n-Pentane
5
12
3.61
18.05
43.32
72.15
0.0500
0.04705
trans-2-Pentene
5
10
0.73
3.65
7.30
70.13
0.0104
0.00979
cis-2-Pentene
5
10
0.43
2.15
4.30
70.13
0.0061
0.00577
2-Methyl-2-butene
5
10
1.21
6.05
12.10
70.13
0.0173
0.01622
2,2-dimethylbutane
6
14
0.505
3.03
7.07
86.18
0.0059
0.00551
cyclopentadiene
5
6
0.505
2.53
3,03
66.10
0.0076
0.00718
cyclopentene
5
8
0.17
0.85
1,36
68.12
0.0025
0.00235
4-methyl-1-pentene
6
12
0.06
0.36
0.72
84.16
0.0007
0.00067
3-methyl-1-pentene
e
12
0.08
0.48
0.96
84.16
0.0010
0.00089
Cyclopentane
5
10
0.23
1.15
2.30
70.13
0.0033
0.00308
2,3-dirnethylbutane
6
14
, 1.65
9,90
23.10
86.18
0.0191
0.01801
4-methyl-cis~2-pentene
6
12
0.06
0.36
0.72
84.16
0.0007
0.00067
2-methylpentane
6
14
5.52
33.12
77.28
86.18
0.0641
0.06024
4-methyl-trans-2-pentene
6
12
0.04
0,24
0.48
84.16
0,0005
0.00045
3~methylpentane
6
14
3.12
18.72
43.68
86.18
0.0362
0.03405
2-methyl-1-peritene
6
12
0.28
1.6B
3.36
84.16
0.0033
0.00313
1 -hexene
6
12
0.12
0.72
1.44
84,16
0.0014
0.00134
n-hexane
6
14
2.65
15.90
37.10
86.18
0.0308
0.02892
trans-3-hexene
6
12
0.18
1.08
2.16
84.16
0,0021
0.00201
trans-2-hexene
6
12
0.38
2.28
4.56
84.16
0.0045
0.00425
3-methyl-cis-2-pentene
6
12
0.11
0.66
1.32
84.16
0.0013
0.00123
4-methylcyclopentene
6
10
0.26
1.56
2.60
82.15
0.0032
0.00298
3-methyl-trans-2-pentena
6
12
0.05
0.30
0.60
84.16
0.0006
0,00056
cis-2-hexene
6
12
0.21
1.26
2.52
84.16
0.0025
0.00235
2,2-dimethylperitane
7
16
0.43
3.01
6.88
100.20
0.0043
0.00404
mathylcycloperitane
6
12
1.11
6.66
13.32
84.16
0.0132
0.01240
2,4-dimethylpentane
7
16
0.69
4.83
11.04
100.20
0.0069
0.00648
t -methylcyclopentene
6
10
0.38
2.28
3.80
82.15
0.0046
0.00435
07 olefin
7
14
0.01
0.07
0.14
98.19
0.0001
0,00010
benzene
6
6
1.22
7.32
7.32
78.11
0.0156
0.01469
3,3-dimethylpentane
7
16
0.14
0.98
2.24
100.20
0.0014
0.00131
cyclohexane
6
12
0,28
1.68
3.36
84.16
0.0033
0.00313
C7 cyclo-olefin/diolefin
7
12
0.06
0.42
0.72
96.17
0.0006
0.00059
C7 olefin
7
14
0.12
0.84
1.68
98.19
0.0012
0.00115
2-methylhexane
7
16
1.63
11.41
26.08
100.20
0.0163
0.01530
2,3-dimethylpentane
7
16
1.3
9.10
20.80
100.20
0.0130
0.01220
1,1 -dimethytcyclopentane
7
14
0.04
0.28
0.56
98.19
0.0004
0.00038
3-methylhexane
7
16
1.7
11.90
27.20
100.20
0.0170
0.01595
C7 olefin
7
14
0.04
0.28
0.56
98.19
0.0004
0.00038
trans-1,3-dimethylcyclopentane
7
14
0.32
2.24
4.48
98.19
0.0033
0.00306
cis-1,3~dimethylcyclopentane
7
14
0.28
1.96
3,92
98.19
0.0029
0.00268
3-ethylpentane
7
16
0.235
1.65
3.76
100.20
0.0023
0.00221
07 olefin
7
14
0.235
1.65
3.29
98.19
0.0024
0.00225
trans-1,2-di methy Icycl opentane
7
14
1.76
12.32
24.64
98.19
0.0179
0.01686
2,2,4-trimethylpentane
8
18
0.105
0.84
1.89
114.23
0.0009
0.00086
C7 olefin
7
14
0.105
0.74
1.47
98.19
0.0011
0.00101
C7 olefin
7
14
0.11
0.77
1.54
98.19
0.0011
0.00105
n-heptane
7
16
1.3
9.10
20.80
100.20
0.0130
0.01220
C7 olefin
7
14
0.05
0,35
0.70
98.19
0.0005
0.00048
101
-------
TABLE A-1. Continued.
Compound
Carbon
Hydrogen
Wt.%
C x Wt%
H x Wt%
Mol. Wt.
(g/mole)
wt%/mol. Wt.
(moles/1 OOg)
Mole Fraction
C7 olefin
7
14
0.11
0.77
1.54
98.19
0.0011
0.00105
07 olefin
7
14
0.12
0.84
1.68
98.19
0.0012
0.00115
C7 olefin
7
14
0.06
0.42
0.84
98.19
0.0006
0.00057
C7 olefin
7
14
0.13
0.91
1.82
98.19
0.0013
0,00125
C8 olefin
8
16
0.09
0.72
1.44
112.21
0.0008
0,00075
C8 olefin
8
16
0.09
0.72
1.44
112.21
0.0008
0.00075
C8 olefin
8
16
0.08
0.64
1.28
112.21
0.0007
0,00067
C8 olefin
8
16
0.05
0.40
0.80
112.21
0.0004
0.00042
cis-1,2-dimethylcyclopentane
7
14
0.18
1.26
2.52
98.19
0.0018
0.00172
methytcyclopentane
6
12
0.43
2.58
5.16
84.16
0.0051
0.00480
ethylcyclopentane
7
14
0.46
3.22
6.44
98.19
0,0047
0.00441
2,2,3-tri methylpentane
8
18
0.18
1.44
3,24
114,23
0.0016
0,00148
C8 olefin
8
16
0.19
1.52
3.04
112.21
0.0017
0.00159
2,4-dimethylhexane
8
18
0.18
1.44
3.24
114.23
0.0016
0.00148
C8 olefin
8
16
0.06
0.48
0.96
112.21
0.0005
0.00050
1,2,3-trimethylcyclopentane
8
16
0.05
0.40
0.80
112.21
0.0004
0.00042
C8 cyc!o-olefin/diolefin
8
14
0.05
0.40
0.70
110.20
0.0005
0.00043
2,3,4-trimethylpentane
8
18
0.365
2,92
6.57
114.23
0.0032
0.00300
C8 olefin
8
16
0.365
2.92
5.84
112.21
0.0033
0.00306
toluene
7
8
7.68
53.76
61.44
92.14
0.0834
0.07839
2,3-dimethylhexane
8
18
0.205
1.64
3.69
114.23
0.0018
0.00169
C8 olefin
8
16
0.205
1.64
3.28
112.21
0.0018
0.00172
2-methylhaptane
8
18
0.85
6.80
15.30
114.23
0.0074
0.00700
4-methyl heptane
8
18
0.36
2.88
6.48
114.23
0.0032
0.00296
3-methylheptane
8
18
0.02
0.16
0.36
114.23
0.0002
0.00016
3-ethylhexane
8
18
0.605
4.84
10.89
114.23
0.0053
0.00498
G8 olefin
8
16
0.605
4.84
9.68
112.21
0.0054
0,00507
C8 Naphthene
8
16
0.07
0.56
1.12
112.21
0.0006
0,00059
C8 olefin
8
16
0.07
0.56
1.12
112.21
0.0006
0,00059
C8 Naphthene
8
16
0.08
0.64
1.28
112.21
0.0007
0.00067
C8 olefin
8
16
0.08
0.64
1.28
112.21
0.0007
0.00067
C8 Naphthene
8
16
0.065
0.52
1.04
112.21
0.0006
0.00054
G8 olefin
8
16
0,065
0.52
1.04
112.21
0.0006
0.00054
C8 Naphthene
8
16
0.09
0.72
1.44
112.21
0.0008
0.00075
C8 Naphthene
8
16
0.13
1.04
2.08
112.21
0.0012
0.00109
n-octane
8
18
0.65
5.20
11.70
114.23
0.0057
' 0.00535
trans-1,2-dimethy Icy clohexane
8
16
0.12
0.96
1.92
112.21
0.0011
0.00101
G8 olefin
8
16
0.05
0.40
0.80
112.21
0.0004
0.00042
C9 naphthene
9
18
0.21
1.89
3.78
126.24
0.Q017
0.00156
C9 paraffin
9
20
0.025
0.23
0.50
128.26
0.0002
0,00018
C8 olefin
8
16
0.025
0.20
0.40
112.21 .
0.0002
0.00021
C8 olefin
8
16
0.06
0.48
0.96
112.21
0.0005
0.00050
C9 paraffin
9
20
0.19
1.71
3,80
128,26
0.0015
0.00139
cis-1,2-dimethylcyclohexane
8
16
0.025
0.20
0.40
112.21
0.0002
0.00021
C9 olefin
9
18
0.025
0,23
0,45
126.24
0.0002
0.00019
C9 paraffin
9
20
0.03
0,27
0.60
128.26
0.0002
0.00022
C9 naphthene
9
18
0.03
0,27
0.54
126.24
0.0002
0.00022
Ethylbenzene
8
10
3.37
26.96
33.70
106.17
0.0317
0.02985
m-Xylene
8
10
5.31
42,48
53,10
106.17
0.0500
0.04704
p-Xylene
8
10
2.13
17.04
21.30
106.17
0.0201
0.01887
2-methyloctane
9
20
0.28
2.52
5.60
128,26
0.0022
0.00205
4-methyloctane
9
20
0.32
2.88
6.40
128,26
0.0025
0.00235
3-methyloctane
9
20
0.34
3.06
6.80
128,26
0.0027
0.00249
oXylene
8
10
2.64
21.12
26.40
106.17
0.0249
0.02339
C10 naphthene
10
20
0.04
0,40
0.80
140.27
0.0003
0.00027
n-nonane
9
20
0.24
2.16
4.80
128.26
0.0019
0,00176
C9 naphthene
9
18
0.02
0.18
0.36
126.24
0,0002
0.00015
isopropylbenzene
9
12
0.22
1.98
2.64
120.19
0.0018
0.00172
102
-------
TABLE A-1. Continued.
Compound
Carbon
Hydrogen
Wt.%
C x Wt%
H x Wt%
Mo!. Wt,
(g/mole)
wt%/mol. Wt.
(moles/1 OOg)
Mole Fraction
~C9 naphthene
9
18
0.02
0.18
0.36
126.24
0.0002
0.00015
n-propylbenzene
9
12
0.75
6,75
9.00
120.19
0.0062
0.00587
1 -methyi-3-ethyl-benzene
9
12
2.34
21.06
28.08
120.19
0.0195
0,01831
1 -methyl-4-ethyl-benzene
9
12
1.06
9.54
12.72
120.19
0.0088
0.00829
C10 paraffin
10
22
0.03
0.30
0.66
142.28
0.0002
0.00020
1,3,5-trimethylbenzene
9
12
1.1
9.90
13.20
120.19
0.0092
0.00861
4-methylnonane
10
22
0.11
1.10
2.42
142.28
0.0008
0,00073
2-methylnonane
10
22
0.15
1.50
3.30
142.28
0,0011
0.00099
1-methy!-2-ethylbenzene
9
12
0.72
6.48
8.64
120.19
0.0060
0.00563
1,2,4-trimethylbenzene
9
12
3.37
30.33
40.44
120.19
0,0280
0.02637
1,2,3-trimethylbenzene
9
12
0.68
6.12
8.16
120.19
0,0057
0.00532
Indane
10
26
0,4
4.00
10.40
146.32
0,0027
0.00257
G11 paraffin
11
24
0.05
0.55
1.20
156.31
0.0003
0,00030
C11 paraffin
11
24
0.22
2.42
5.28
156.31
0.0014
0,00132
1 -methyi-3-n-propylbenzene
10
14
0.48
4.80
6.72
134.22
0.0036
0.00336
1 -methyl-4-n-propylbenzene
10
14
0.28
2.80
3.92
134.22
0.0021
0.00196
n-butylfaenzene
10
14
0.14
1.40
1.96
134.22
0.0010
0.00098
1,2-diethylbenzene
10
14
0.44
4.40
8.16
134.22
0.0033
0.00308
1,3-dlmethyl-5-ethylbenzene
10
14
0.02
0.20
0.28
134.22
0.0001
0.00014
1,4-diethylbenzene
10
14
0.02
0,20
0.28
134.22
0.0001
0.00014
C11 paraffin
11
24
0.01
0.11
0.24
156,31
0.0001
0.00006
1,3-dimethyl-4-ethylbenzene
10
14
0.125
1.25
1.75
134.22
0.0009
0.00088
Indane
10
26
0.125
1.25
3.25
146,32
0.0009
0.00080
1,2-dimethyl-4~ethylbenzene
10
14
0.31
3.10
4.34
134.22
0.0023
0.00217
Indane
10
26
0.31
310
8.06
146.32
0.0021
0.00199
1,2-dimethy l-3-ethylbe nze ne
10
14
0.15
1.50
2.10
134.22
0.0011
0.00105
tvundecane
11
24
0.06
0.66
1.44
156,31
0.0004
0.00036
1,2,4,5-tetramethylbenzene
10
14
0.24
2.40
3.36
134,22
0.0018
0.00168
1,2,3,5-tetramethylbenzene
10
14
0.33
3.30
4.62
134.22
0.0025
0.00231
dodecane
12
26
0.05
0.60
1.30
170.34
0.0003
0.00028
Naphthalene
10
8
0.53
5.30
4.24
128.17
0.0041
0.00389
G13 paraffins
13
26
0.19
2.47
5.32
184.36
0,0010
0.00097
C12 aromatics
12
18
1.06
12.72
19,08
162.27
0.0065
0,00614
C12 Indanes
10
26
0.54
5.40
14.04
146.32
0.0037
0,00347
Methylnaphthalenes
11
12
0.69
7.59
8,28
144.22
0.0048
0.00450
Unidentified heavies
12
26
0.99
11.88
25.74
170.34
0.0058
0,00547
unknowns
7
16
4.86
34.02
77.76
100.20
0.0485
0.04561
TOTAL
100.00
94,002
1.0633
1,00000
¦' c8, c9, and c10 naphthene molecular weights taken from Table 2-A Poulsori et.al, (1990) Indanes taken from same source. Assumed
average group molecular weight. Assumed value near gasoline molecular weight lor unknown fraction. Assumed dodecane molecu-
lar weight for unknown heavies.
2 Calculated molecular weight.
103
-------
Table A-2. Characteristics of API 91-01 Gasoline
Molecular Weight
Density1
MTBE
Methanol
(g/mole)
(g/cm3)
(vol %)
(vol %)
94.3
0.747
<0.10
<0.10
HYDROCARBON CLASSES (wt %)
Paraffins 41.01
Naphthenes 6.84
Aromatics 39.05
Olefins 11.74
Unknowns 1.26
AROMATIC HYDROCARBONS
Benzene
Toluene
Ethylbenzene
m-Xylene
p-Xylene
o-Xylene
1,3,5-T rimethylbenzene
1,2,4-T rimethylbenzene
1,2,3-T rimethylbenzene
Naphthalene
(wt % (mole fraction))
1.22
7.68
3.37
5.31
2.13
2.64
1.10
3.37
0.68
0.53
(0.0147)
(0.0787)
(0.0300)
(0.0472)
(0.0189)
(0.0235)
(0.00864)
(0.0265)
(0.00534)
(0.00390)
1 Gasoline density measured at University of Waterloo (Oliveira, 1997). All other
information from an unpublished 1991 analysis of the gasoline performed by API.
Table A-3. Comparison of Measured Concentrations of Aromatic Hydrocarbons in API 91-01 Gasoline. All
Concentrations Expressed as Weight Percent.
Compound
API
(1991)
(August, 1997)
This Study
(August, 1998}' (August, 1998)
(August, 19:
Benzene
1.22
1.28
1.02
0.96
0.97
Toluene
7.68
8.26
8.67
8.31
8.21
Ethylbenzene
3,37
3.48
3.50
3.33
3.30
m+p~Xylene
7.44
7.50
8.06
7.66
7.60
o-Xylene
2.64
2.70
2.75
2.61
2.59
1,3,5 Trimethylbenzene
1.10
1.13
1.18
1.12
1.11
1,2,4-Trimeth y I ben zen e
3.37
3.02
3.18
3.01
297
1,2,3-Trimeth y Iben zen e
0.68
0.67
0.74
0.70
0.69
Naphthalene
0.53
0.66
0.74
0.70
0.69
1 Samples analyzed August, 1998 are means of three injections.
104
-------
Table A-4. Measured and Calculated Concentrations of Aromatic Hydrocarbons in Water Equilibrated with API
91-01 Gasoline at 10°C.
Concentration (mg/L)
Compound
Measured1
Calculated2
% Diff3
Benzene
30.90
26.25
15.1
Toluene
49.72
42.10
15.3
Ethylbenzene
6.41
5.42
15.3
m+p Xylene
13.45
10.59
21,3
o-Xylene
5.79
4.18
27.8
1,3,5-T rimethylbenzcne
0.54
0.42
23.2
1,2,4-T rimethylbenzene
1.89
1.56
17.3
1,2,3-Trimethylbenzene
0.51
0.40
20.7
Naphthalene
0.84
0.13
84.1
1 Measured concentrations are means of four replicates collected from two separatory funnels. Coefficients of variation ranged from 1,3%
(toluene) to 11% (naphthalene).
2 Calculated concentrations are based on Raoult's Law, C.t = XjS, where X is the mole fraction of component i (Table A-1), S is the pure-phase
solubility, and C is the calculated concentration. Solubilities were obtained from published sources: benzene 1780 mg/L; toluene 535 mg/L;
p-xylene 160 mg/L; o-xylene 178 mg/L; 1,3,5-trimethylbenzene 48 mg/L; 1,2,4-trimethylbenzene 59 mg/L; 1,2,3-trimethylbenzene 75 mg/L;
naphthalene 34 mg/L (Source: References cited within Montgomery, j.H, 1996. Groundwater Chemicals Desk Reference, 2nd Ed., CRC-
Lewis Publishers, Boca Raton, Fla.).
3 Percent difference:
NAPH
Measured Concentration (mg/L)
FIGURE A-1 Calculated vs. measured BTEXTMB concentrations in gasoline-saturated water. Concentrations
obtained from equilibration of water with fresh API 91-01 gasoline as described in text.
105
-------
Intentionally Blank Page
106
-------
APPENDIX B: FIELD TRACER TEST
107
-------
APPENDIX B: FIELD TRACER TEST
B.1 Introduction
Tracer tests were performed in both experimental cells during the month of July, 1996 while the
flushing experiment was in progress. The objectives of the test were to confirm that the injected water
was distributed uniformly in the cells, and to calculate linear groundwater velocities and longitudinal
dispersivities at different locations within the cells. Bromide (Br) was used as the conservative tracer.
B.2 Methods
A concentrated Br solution was prepared with KBr salt. The solution was pumped into the injection
flow lines of both cells. A peristaltic pump (Masterflex L'S Series) equipped with an Ismatec multichannel
head was used to feed the stock solution into the flow lines. The pumping rate was calibrated to a rate
that yielded a 100x dilution of the concentrated solution. Brominated water was pumped into the cells
over a 45 hr period to create a slug input. Injection water to each cell was sampled periodically to
determine the mean injection concentrations (C0).
In addition to injection samples, groundwater samples were collected periodically from all of the
60-cm, 120-cm, and 180-cm bgs piezometer ports, and the extraction-well sampling ports. These data
were used to develop relative concentration vs. time breakthrough curves (BTCs) at each sampling
location. A plastic 60 cc syringe was used to collect groundwater samples from the piezometers. Prior to
collecting a sample, about 20-ml water was removed to clear stagnant water from the piezometer tube
and to flush the syringe. Samples from the injection flow lines were collected by holding the sample vial
under the flowing stream of water at the tube outlet. Samples from the extraction flow lines were collected
from the in-line sampling ports (see Figure 3-6). Once sampling began at a given depth, samples were
collected approximately every six hours. A total of 540 samples was collected over a 400 hr period, and
stored in plastic scintillation vials for analysis.
Samples were screened on-site with a conductivity meter. These data were used to monitor the
position of the tracer slug as the test progressed. The goal was to obtain complete breakthrough curves
for each piezometer port and the extraction wells. After the test, samples were returned to the laboratory
and analyzed with a bromide-specific electrode (Coming, Model 476128)." Samples were prepared for
analysis by adding 5 ml ionic strength adjuster (0.2 M KNO, solution) to a 5 ml groundwater sample.
Standards were prepared from a concentrated KBr stock solution, and diluted to the appropriate
concentrations using aerobic Borden groundwater and 5 ml ionic strength adjuster. Each day, standards
were run in duplicate or triplicate to generate a standard curve. Sample quantification was based on a
linear regression of electrode response (mV) vs. natural log of the standard concentration. To check the
accuracy of the Br electrode results, a subset of the groundwater samples analyzed on different days with
the electrode was submitted to the Water Quality Laboratory for analysis by ion chromatography (see
Section 3.6.5). As shown in Table B-1, there was good agreement between methods.
B.3 Results
Injection Br concentrations and extraction well BTCs are shown in Figures B-1 and B-2 for the
Control and Nitrate Cells, respectively. Breakthrough curves from piezometer ports at the 60-, 120- and
180-cm depths of both experimental cells are shown in Figures B-3 through B-8. Based on eight samples,
the average injection concentrations of the Br slug, C0, were 260.1 ± 9 mg/L and 260.1 ± 12.4 mg/L for
the Nitrate and Control Cells, respectively. As shown on the figures, complete BTCs were obtained for
nearly all piezometer ports monitored during the test; in some cases (e.g., PZ3C-4, PZ3D-4), the initial
arrival of the tracer slug was missed. The shapes of the BTCs were quite similar at a given depth, and
there were no indicators (e.g., very early arrival of tracer) of short-circuiting along piezometer casings.
108
-------
Linear groundwater velocities and dispersivities were calculated from breakthrough data at ail 120-cm
and 180-cm piezometer ports, and the extraction-weil sampling ports. These parameters were obtained
by fitting the one-dimensional advection-dispersion equation to the normalized BTCs. Modeling was done
with Wpulsepe (Devlin and Barker, 1996; Sorel et al., 1998). This program uses a simplex optimization
routine to find the best fit of the model solution to the data. The model requires only the length of the
flowpath, the mean injection concentration, and the concentration vs. time breakthrough data as input.
The lengths of the flow paths were calculated assuming the upper head remained at 50-cm bgs during the
test (70 cm and 130 cm flow paths for 120- and 180-cm bgs piezometer ports, respectively, and 175 cm
to the midpoint of the extraction-well screens). Both velocity and dispersivity are used by the model as
fitting parameters. As shown in the figures, the model fit the measured breakthrough data very well at
most sampling locations.
Model-calculated velocities and dispersivities are shown in Table B-2. Velocities ranged from
0.45 cm/hr to 1.2 cm/hr and were consistent with a 200 ml/min injection rate. The velocities from
piezometer BTCs represent the average rate of flow of the tracer in the individual flow tube intersecting
the sample port. If sufficient ports are measured to obtain a representative sample of all of the flow tubes
in the domain, the mean should be equal to the volumetric tracer injection rate. For this test, the expected
linear velocity of 0.91 cm/hr (assumed porosity of 0.33) did fall within the error of the mean velocities
calculated from piezometer breakthrough data in both cells (Table B-2). The velocities at the extraction
wells were slightly higher, possibly reflecting the influence of the early arrival of Br in the faster flow tubes.
Dispersivities were in the millimeter range in the Control Cell, and the millimeter to centimeter range in the
Nitrate Cell (Table B-2). The largest dispersivities were measured at the extraction wells. These large
dispersivities were caused by an integration of variable tracer concentrations along the 50-cm well
screens (Sorel et al., 1998); extraction-well breakthrough curves represent the sum of the relative
concentrations in all of the flow tubes in the domain, as well as clean water from the underlying aquifer.
Table B-1. Comparison of Bromide Concentrations Determined from a Bromide Electrode and Ion
Chromatography (IC). Concentrations in mg/L.
Sample
IC
Electrode
T-24
155.0
154.6
T-41
3.8
4.3
T-50
250.0
246.7
T-84
258.0
260.8
T-108
<0.05
0.6
T-126
7.8
7.8
T-131
44.4
40.9
T-160
222.0
223.5
T-175
0.21
1.1
T-216
289.0
292.8
T-284
1.3
1.5
T-314
62.8
67.1
T-349
293.0
277.8
T-406
87.1
- 88.5
109
-------
Bromide Breakthrough Curves - Nitrate Cell
Injection Well
Co = 260,1 ± 9,0 mg/
r 1 1 t~
0 50 100 150 200 250 300 350 400
time (hours)
Extraction Well
0 50 100 150 200 250 300 350 400
time (hours)
_ model ~ measured
FIGURE B-1 Normalized bromide concentrations in samples of injection and extraction water during bromide tracer
test (Nitrate Cell).
Bromide Breakthrough Curves - Control Cell
njection We
Co = 260,1 +- 12.4 mg/l
Extraction We
T"
0 50 100 150 200 250 300 350 400
time (hours)
s
0 50 100 150 200 250 300 350 400
time (hours)
model ta measured
FIGURE B-2 Normalized bromide concentrations in samples of injection and extraction water during bromide tracer
test (Control Ceil).
110
-------
Bromide Breakthrough Curves - Nitrate Cell, 60 cm Depth
PZ4C-2
tP31
0 50 100 150 200 250 300 350 400
lime (hours)
PZ4D-2
jjEpJ
0.0 fl—po
-l 1 ! 1 r
0 50 100 150 200 250 300 350 400
time (hours)
PZ4E-2
o.o v 1 1 1
0 50 100 150 200 250 300 350 400
time (hours)
PZ4B-2
o
o
0.0 O po—i
0 50 100 150 200 250 300 350 400
time (hours)
PZ4A-2
0.0 p
T T-
0 50 100 150 200 250 300 350 400
time (hours)
! ~ measured
FIGURE B-3 Measured bromide breakthrough curves at 60-em bgs ports in the Nitrate Cell.
111
-------
Bromide Breakthrough Curves - Nitrate Cell, 120 cm Depth
PZ4C-4
PZ4D-4
50 1 00 150 200 250 300
time (hours)
0.0
0 50 100 150 200 250 300
time (hours)
PZ4E-4
Not Available
J^
0 50 100 150 200 250 300 350 400
time (hours)
PZ4B-4
50 100 1 50 200 250 300
time (hours)
PZ4A-4
<=>0.6
50 100 150 200 250 300
time (hours)
model a measured
FIGURE B-4 Measured and calculated bromide breakthrough curves at 120-cm bgs ports in the Nitrate Cell. Samples
could not be collected from the 120 cm port on the center piezometer (PZ4E-4).
112
-------
Bromide Breakthrough Curves - Nitrate Cell, 180 cm Depth
1.2
0.9
Q.8
o
o
0.3
0.0
PZ4C-6
0 50 100 150 200 250 300 350 400
time (hours)
PZ4D-6
0 50 IOC 150 200 250 300 350 400
time (hours)
1.2
0.9
PZ4E-6
.0.6 -
o
o
0.3
0.0
PZ4B-6
0 50 100 150 200 250 300 350 400
time (hours)
1.2
0.9
PZ4A-6
,0.6
P
<3
0.3
0.0
0 50 100 150 200 250 300 350 400
time (hours)
0 50 100 150 200 250 300 350 400
time (hours)
model n measured
FIGURE B-5 Measured and calculated bromide breakthrough curves at 180-em bgs ports in the Nitrate Cell.
113
-------
Bromide Breakthrough Curves - Control Cell. 60 cm Depth
PZ3C-4
0,0 —9°
PZ3D-2
0.0 £
50 1 00 150 200 250 300
time (hours)
0 50 100 1 50 200 250 300
time (hours)
PZ3E-2
50 100 1 50 200 250 300
time (hours)
PZ3A-2
PZ3B-2
100 150 200 250 300
lime (hours)
0 50 1 00 150 200 250 300
time (hours]
~ measured
FIGURE B-6 Measured bromide breakthrough curves at 60-cm bgs ports in the Control Celt.
114
-------
Bromide Breakthrough Curves - Control Cell, 120 cm Depth
PZ3C-4
PZ3D-4
50 1 00 1 50 200 250 300
time (hours)
50 100 1 50 200 250 300
time (hours)
PZ3E-4
PZ3B-4
O
d
0 50 100 150 200 250 300
time (hours)
PZ3A-4
-tr
0 50 1 00 1 50 200 250 300
time (hours)
100 150 200
time (hours)
model ~ measured
FIGURE B-7 Measured and calculated bromide breakthrough curves at 120-cm bgs ports in the Control Cell.
115
-------
Bromide Breakthrough Curves - Control Cell, 180 cm Depth
PZ3C-6
PZ3D-6
100 150 200 250 300
time (hours)
50 100 150 200 250 300
time (hours)
PZ3E-6
PZ3B-6
o.o
0 50
100 150 200 250 300
time (hours)
PZ3A-6
0 50 100 150 200 250 300
time (hours)
50 100 150 200 250 300
lime (hours)
model d measured
FIGURE B-8 Measured and calculated bromide breakthrough curves at 180-cm bgs ports in the Control Cell.
116
-------
Table B-2. Bromide Tracer Test Results. Velocities and Dispersivities Calculated by Fitting Wpulsepe to
Bromide Breakthrough Data.
Vertical
Location Distance1 Velocity Dispersivity
(cm) (cm/hr) (cm)
Control Cell
PZ3A-4
70
1.08
0.17
PZ3B-4
70
0.97
0.31
PZ3C-4
70
1.23
0.35
PZ3D^4
70
1.12
0.33
PZ3E-4
70
0.95
0.17
PZ3A-6
130
0.72
0.32
PZ3B-6
130
0.59
0.30
PZ3C-6
130
0.84
0.32
PZ3D-6
130
0.71
0.81
PZ3E-6
130
0.81
0.13
Extraction Well
1752
1.29
13.45
MEAN3(s.d.)
0.90 (0.2)
0.32 (0.19)
Nitrate Cell
PZ4A-4
70
0.66
0.46
PZ4B-4
70
0.97
0.08
PZ4C-4
70
0.77
0.06
PZ4D-4
70
1.00
0.17
PZ4A-6
130
0.45
0.75
PZ4B-6
130
0.80
0.11
PZ4C-6
130
0.63
2,13
PZ4D-6
130
0.60
1.08
PZ4E-6
130
1.13
0.26
Extraction Well
1752
1.21
5.26
MEAN3fs.d.)
0.78(0.22)
0.57(0.68)
' Vertical distances calculated assuming upper head at 50 cm bgs during test,
2 Distance to center of well screen,
3 Mean velocities and dispersivities do not include extraction well values.
117
-------
Intentionally Blank Page
118
-------
APPENDIX C. ANALYTICAL METHODS
119
-------
APPENDIX C. ANALYTICAL METHODS
C. i Sampling Procedures and Analytical Methods - Laboratory Samples
Aromatic Hydrocarbons. For microcosms with mininert valves, a 400 pL sample of the headspace
gas was collected with a syringe and injected directly onto a Shimadzu GC-9a gas chromatograph
equipped with a flame ionization detector and 60 m Supelcowax-10 capillary column. Each sample was
injected manually with the aid of a sample loop. The GC was run isothermally at a column temperature of
105°C, and an injector temperature of 200°C. The method detection limit (MDL) for benzene and toluene
was 0.001 mg/L headspace; the MDLs for the other aromatics were not determined, but were expected to
be of similar magnitude.
For hypovial microcosms, vials were decrimped and aliquots of liquid quickly transferred with a glass
syringe to either 22-ml autosampler vials or 18-ml hypovials, which were then crimp sealed, and analyzed
immediately. Samples were collected outside the anaerobic chamber. Either the pentane microextraction
technique (described in Section C.2) or headspace GC-PID was used to quantify the aromatic
hydrocarbons, depending on the experiment. GC-PID was used for microcosm Experiment 3
(Section 4.1.3), and the pentane microextraction technique for Experiments 4,5, and 6 (Sections 4.1.4,
4.1.5, and 4.2). The method of analysis was not changed within a given experiment.
For headspace GC-PID, 8 ml of fluid was removed from the 22-ml autosampler vials by glass syringe
just before analysis. Vials were then resealed and placed in an autosampler. Analysis was conducted on
a Hewlett Packard 5890 gas chromatograph equipped with a split injection port, a Varian Genesis
headspace autosampler (platen temperature of 75°C), and a photoionization detector (10.2 eV). A
30m x 0.32mm Stabilwax column (0.5 pm film thickness) was used, operating at 65°C under isothermal
conditions. Injector and detector temperatures were 150°C, and helium was used as the carrier
(3.5 ml/min) and makeup (30 ml/min) gas. Calibration was by the external standard method, using
calibration standards prepared by spiking methanolic stocks into organic-free water contained in the same
autosampler vials. The MDLs for BTEX compounds ranged from 2-15 Mg/L, as determined by the U.S.
EPA method (Longbottom and Lichtenberg, 1982). During analyses, check standards were run
approximately every ten samples.
Dissolved Oxygen. In the laboratory, the dissolved 02 concentration in microcosm liquid was
determined immediately after organic sample collection. Measurements were typically made within
30 seconds of opening a microcosm, and care was taken to minimize the disturbance of the liquid.
Dissolved 02 determinations were made with either the azide-modified Winkler titration method (APHA,
1985), or a D.O. meter (Microelectrodes, Inc., Model Ml 730), depending on the experiment. The
analytical method was not changed within a given experiment.
For the Winkler method, 18-ml glass vials were flushed with argon, filled with microcosm liquid, and
sealed with a Teflon™-lined septum and aluminum crimp seal. Winkler reagents were then injected
directly into the vial using a syringe. The solution in the vial was then titrated with a standardized
0.0025 N sodium thiosulfate solution to determine the dissolved 02 concentration. The MDL for this
sample size was 0.22 mg/L. D.O. values determined with the Winkler method were corrected for the 02
added to samples as a result of reagent addition. It was observed that the sodium azide used to inhibit
microbial activity in control microcosms may have interfered with the analysis, resulting in D.O.
concentrations that appeared to be anomalous. Consequently, the D.O. meter was used in the follow-up
microcosm study (Experiment 6) to obtain data on initial dissolved 02 concentrations and abiotic losses in
the absence of microbial activity (i.e., sterile controls).
120
-------
To obtain sample concentrations with the dissolved-02 meter a daily two-point calibration (0 to 21%
dissolved 02) was used. The instrument was zeroed with a 2% (w/v) sodium sulfite solution.
Measurements were made by inserting the probe directly into the microcosm liquid. To minimize diffusion
of atmospheric 02 into the sample, measurements were made under a stream of argon gas. Although
steps were taken to minimize contamination with atmospheric O, during this sampling procedure, it could
not be determined whether the low concentrations of dissolved 02 that were typically observed after
lengthy incubations in the anaerobic chamber (Chapter 4) resulted from positive sampling bias or were
representative of microcosm liquid.
Headspace Oxygen. The oxygen content of the air-filled headspace of certain microcosms was
measured using a Fisher/Hamilton gas partitioner (Model 29) equipped with columns of 30%
di-2-ethylhexyl-sebacate on 60-80 mesh Chromosorb-P column and molecular sieve 13X In series, and a
thermal conductivity detector. Sample was introduced on-column via a sample loop, and analysis was
conducted at room temperature, using helium (20 ml/min) as the carrier gas. Calibration was by the
external standard method, using commercially-obtained, certified gas mixtures; chromatogram peak
heights of standards and samples were compared to determine unknown concentrations.
Nitrate and Nitrite. After collecting samples for organics and dissolved Oa, samples for NO/ and
NOz analysis were obtained by transferring an additional 15 ml of liquid to 18-ml plastic scintillation vials.
These samples were preserved with 30 pL concentrated sulfuric acid and refrigerated until analysis.
Nitrate and N02" were determined colorimetrically with the automated cadmium reduction method, using a
Technicon Autoanalyzer equipped with a 15 mm tubular flow cell and 550 nm filters. Before analysis,
samples were diluted as required, and the pH adjusted with ammonium hydroxide to between 7 and 9.
The MDLs were 0.2 mg/L and 0.26 mg/L for N02" and N03 , respectively.
Nitrous Oxide. In design 1 microcosms that received acetylene, N20 accumulation was measured
using a GOW-MAC Series 350 GC equipped with a Thermal Conductivity Detector and a 1,8m,
100-120 mesh Poropak Q column. Samples were run isothermaliy at 40°C with a helium carrier
(17 ml/min). Fifteen ml groundwater samples were collected from microcosms in 30 ml glass syringes.
To obtain a gas-phase sample for analysis, syringes received 13-rnl helium, and were sealed, shaken
100 times, and allowed to equilibrate for at least 2 hours. A sample loop was then used to inject a 2-ml
gas sample onto the GC. Triplicate standards (0.1% and 0.5% N O in helium) were run to develop a two-
point calibration curve. Partitioning theory was used to calculate the concentration of N20 in the aqueous
phase. The MDL was 0.45 mg/L dissolved N20. Acetylene was not quantified, but its presence was
confirmed by reviewing chromatograms.
C.2 Analytical Methods - Field Samples
Aromatic Hydrocarbons. Concentrations of aromatic hydrocarbons in all groundwater samples
collected in the field (as well as three laboratory experiments) were determined with a pentane micro-
extraction procedure. Samples were first extracted by adding 1 ml of pentane containing an internal
standard (m-fluorotoluene) to 16.5 ml of groundwater and agitating for twenty minutes. Approximately
0.8 ml pentane was then transferred to an autosampler vial for analysis. Samples were run on a Hewlett
Packard 5890 GC equipped with an HP7673A autosampler and a 30m x 0.25mm I.D. DB-5 column. The
oven temperature program was 35°C for 1 min increasing to 165°C at 13°C/min with a 4 min final hold.
The injector and detector temperatures were 200°C and 250°C, respectively. A calibration curve based on
three standards was prepared for each sample run, and check standards were run approximately every
ten samples. Standards were prepared by weighing known amounts of neat compounds into a known
weight of methanol, and then diluting the stock solution into water to obtain a concentration range that
bounded expected sample concentrations. Standards were analyzed in triplicate. Results for m-xylene
and p-xylene were reported as a sum because these two compounds co-eluted on this column. The
MDLs were 19 Mg/L (benzene), 13 pg/L (toluene), 8 pg/L (ethylbenzene, and o-xylene), 9 Mg/L
(m+p xylene, 135-trimethylbenzene, and 124-trimethylbenzene). 6 pg/L (123-trimethylbenzene), and 120
pg/L (naphthalene).
Concentrations of aromatic hydrocarbons in core-extract samples were determined by a direct
injection of the extractant onto the GC. Samples were run on a Hewlett Packard HP 5890 GC equipped
with an HP7673A autosampler, and a 30m x 0.25mm I.D. DB-5 column. A 3 piL on-column injection of
121
-------
methanol was performed. The oven temperature program was as follows: 35°C for 5 min increasing to
150°C at 10°C/min with a 5 min final hold. The injector and detector temperatures were 200°C and 250°C,
respectively. As for aqueous samples, a calibration curve based on three standards was prepared for
each sample run, and check standards were run approximately every ten samples. Standards were
prepared by weighing known amounts of neat compounds into a known weight of methanol. The
standards were then further diluted in methanol to reach the appropriate concentration range, and
analyzed in triplicate. The mass of aquifer material and methanol in each sample was used to express
concentrations in mg/ml methanol on a wet-weight basis (mg/kg). The average method detection limits,
which varied with methanol and sample mass, were as follows: 2.6 mg/kg for benzene, toluene,
ethylbenzene, and p+m-xylene; 1.3 mg/kg for o-xylene and trimethylbenzene isomers; and 3.9 mg/kg for
naphthalene.
To measure the concentrations of aromatic hydrocarbons in API 91-01 gasoline, samples were
prepared by weighing 15 ml of methanol into pre-weighed 40-ml glass vials with Teflon-faced septa.
Approximately 0.5 ml of gasoline was then weighed into the methanol and mixed. Aliquots were then
transferred to autosampler vials for analysis. Standards were prepared as described in the previous
paragraph. Concentrations were expressed as g analyte/g gasoline or weight percent.
Inorganic Parameters. Inorganic parameters were analyzed by the Water Quality Laboratory at the
University of Waterloo. Anions such as S042, Br, N03, and N02 were analyzed using a Dionex System
2000 Ion Chromatograph equipped with a Dionex AS4A anion exchange column. A daily run of 20 to
50 samples contained 10 to 20 in-house standards. A commercially-prepared standard was run along
with in-house standards to maintain standard quality. Samples were reanalyzed if the commercial
standard did not come within five percent of its stated value. The method detection limit for all
compounds was 0.05 mg/L.
Because the highly-contaminated water samples collected in this study appeared to be damaging the
anion exchange column, most of the N03_ and N02" samples collected after the first month of the flushing
experiment were analyzed colorimetrically using the automated cadmium reduction method. Samples
were run on an Alpkem Perstorp Analytical Environmental Flow Solution system. In-house standards
were run for calibration and quality control as described above. The method detection limits were
0.4 mg/L (NO,,) and 0.2 mg/L (NO./). To verify that these two analytical methods provided consistent
results, several batches of N03~ samples were analyzed by both methods.
Iron was analyzed on a Varian Model 1475 Atomic Absorption Spectrophotometer. All samples are
run in duplicate. As above, samples are run with in-house standards and a commercially-prepared
standard. Commercial standards were analyzed every five samples to monitor instrument drift. The
method detection limit was 0.05 mg/L.
Dissolved Methane. Water samples for CH4 analysis were analyzed on a Hewlett Packard
5840A GC equipped with a flame ionization detector and 30 m megabore GS-Q column. Analyses were
run isothermally at 100°C with a helium carrier gas (12 ml/min). The detector and injector temperatures
were 200°C and 100°C, respectively. To prepare samples for analysis, 15-ml aliquots of groundwater
were withdrawn from the sample bottle into a 30-ml glass syringe. An additional 13 ml of helium was
added, and the syringe was shaken and allowed to equilibrate for 3 hr. A 5 ml sample of the gas phase
was then injected via a 2-ml sample loop. The GC was calibrated in an external standard mode using
several concentrations of a commercial gas mixture (Praxair). Henry's Law, the Ideal Gas Law and CH4
solubility were then used to calculate concentrations in the aqueous phase. The MDL for dissolved CH4
was 0.1 fjg/L.
Metabolites. Descriptions of these analytical methods are provided elsewhere (Barcelona et al.,
1995 (NCIBRD Laboratory); Hutchins et al., 1998 (NRMRL)).
122
-------
APPENDIX D: AQUIFER MONITORING RESULTS
123
-------
Table D-1. Environmental Monitoring Downgradient of Wastewater Treatment Mound and Treatment Cells. All
Concentations in mg/L.
Sample
Date
Depth
Ben
Tol
Eben
m+p-Xyl
o-Xyl
1,3.5-TMB I 1,2.4-TMB
1,2,3-TMB
Naph
l,D.
(m bgs)
Wastewater Treatment
Mound
Biopite Well
E-640
10/2/96
0.4 - 1.9
n.d.1
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
E-641
10/16/96
0.4 - 1.9
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
E-821
5/27/98
0,4 -1.9
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
Bundle Piezometer
BP5-1
E-795
7/9/97
0.5
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP5-1
E-822
5/27/98
0.5
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP5-2
E-796
7/9/97
1.0
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP5-2
E-823
5/27/98
1.0
n.d.
n.d.
n.d.
n.d,
n.d.
n.d.
n.d.
n.d.
n.d.
BP5-3
E-797
7/9/97
1.5
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP5-3
E-824
5/27/98
1.5
n.d.
n.d.
n,d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
Treatment Cells
Bundle Piezometers
BP1-1
E-635
10/2/96
2.9
n.d.
0.010
n.d.
0.010
n.d.
n,d.
n.d.
n.d.
n.d.
BP1-2
E^36
10/2/96
3.2
n.d.
n.d.
n.d.
0.007
n.d.
n.d.
n.d.
n.d.
n.d.
BP 1-3
E-637
10/2/96
3.5
n.d.
n.d.
n.d.
0.007
n.d.
n.d.
n.d.
n.d.
n.d.
BP1-4
E-638
E-639
10/2/96
J0/2/9jf
3.8
r 4.i
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
0.011
0.008
n.d.
n.d.
n.d.
n.d.
n.d.
n.d. ~
n.d.
n.d.
BP1-5
n.d.
n.d.
BP2-1
¦ ~ BP2-2
E-629
10/2/96
2.4
n.d.
0,016
n.d.
0.015
n.d.
n.d.
n.d.
n.d.
n.d.
E-630
10/2/96
2.7
n.d.
0.012
n.d.
0.012
n.d.
n.d.
n.d.
n.d.
n.d.
BP2-3
E-631
10/2/96
3.0
n.d.
0.010
n.d,
0.011
n.d.
n.d.
n.d.
n.d.
n.d.
BP2-4
E-632
10/2/96
3.3
n.d.
0.013
0,016
0.05
0.016
0.013
0.046
0.013
0.074
BP2-4
E-642
10/16/96
3.3
n.d.
0.014
0.009
0.033
0.010
0.009
0.032
0.009
0.026
BP2-5
E-633
10/2/96
3.6
n.d.
0.014
0.010
0.030
0.010
0.015
0.053
0.014
0.027
BP2-6
E-634
10/2/96
_ 3.9
n.d.
0.011
n.d.
(1011
n.d.
n.d.
n.d.
n.d.
n.d.
BP3-4
n.d.
n.d.
E-371
7/2/96
3.6
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP3-4
E-627
10/2/96
3.6
n.d.
0.027
0.008
0.022
0.006
n.d.
n.d.
n.d.
n.d.
BP3-4
E-793
7/9/97
3.6
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
J3P3-4
E-827
5/27 "98
3.6
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP3-5
E-372
7/2/96
3.9
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP3-5
E-628
10/2/96
3.9
n.d.
0.024
0.007
0.019
n.d.
n.d.
n.d.
n.d.
n.d.
BP3-5
E-794
3.9
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP3-5
E-828
5/27/98
3.9
n.d,
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP4-5
E-373
E-625
7/2/96
10/2/96"
3.7
3.7
n.d.
n.d.
n.d.
n.d.
n.d.
0.013
n.d.
n.d.
n.d.
0.007
n.d.
n.d.
n.d.
' n.d.
BP4-5
n.d.
0.06
0.015
0.039
BP4-5
E-642
10/16/96
3.7
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP4-5
E-791
7/9/97
3.7
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP4-5
E-825
5/27/98
3,7
n.d.
n.d.
n.d,
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP4-6
" BP4-6
_E-374_
E-626
7/2/96
10/2/96
4.0
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
4.0
n.d.
0.026
0.02
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP4-6
E-792
7/9/97
4.0
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
BP4-6
E826
5/27/98
4.0
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d. - Not detected. Detection limits provided in Appendix C.
124
-------
TECHNICAL REPORT DATA
NRMKL-ADA-99212
1. REPORT NO.
EPA/SOO/R-99/012
P
0.
E
4. TITLE AND SUBTITLE
BIOTRANSFORMATION OF GASOLINE-CONTAMINATED GOUNDWATER UNDER MIXED
ELECTRON-ACCEPTOR CONDITIONS
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Jeffrey R. Barbaro, Barbara J. Butler, and James F. Barker
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
University of Waterloo
Waterloo, Ontario, CANADA
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
CR-S21887
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. EPA
NATIONAL RISK MANAGEMENT RESEARCH LABORATORY
SUBSURFACE PROTECTION AND REMEDIATION DIVISION
P.O. BOX 1198; ADA, OK 74820
13. TYPE OF REPORT AND PERIOD COVERED
Final Report
14 . SPONSORING AGENCY CODE
EPA/600/15
15. SUPPLEMENTARY NOTES
PROJECT OFFICER: Stephen R. Hutchins, Project Officer, 580-436-8563
16. ABSTRACT
This project represents a cooperative effort between the university of Waterloo and the U.S. Environmental
Protection Agency. This report summarizes research, conducted using both laboratory batch microcosms and field-
scale sheet-piling cells to evaluate whether bioremediation of monoaromatie fuel hydrocarbons can be enhanced
using mixed rather than single electron acceptors. The studies focused on nitrate for anaerobic bioremediation
and oxygen for aerobic bioremediation, and experiments were designed to test the hypothesis that low levels of
oxygen may enhance biodegradation of more recalcitrant compounds {such as benzene) under denitrifying
conditions. The findings from this project are directly applicable to the field-scale remediation of
subsurface environments contaminated by petroleum hydrocarbons.
17. REX WORDS AND DOCUMENT ANALYSIS
A, DESCRIPTORS
B. IDENTIFIERS/OPEN ENDED TERMS
C. COSATI FIELD, GROUP
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS(THIS REPORT)
UNCLASSIFIED
21. NO. OF PAGES 136
20. SECURITY CLASS(THIS PAGE)
UNCLASSIFIED
22. PRICE
EPA FORM 2220-1 (REV.4-77) PREVIOUS EDITION IS OBSOLETE
------- |