Citation
Mass transfer, biodegradation and biostabilization of DNAPL coal tar

Material Information

Title:
Mass transfer, biodegradation and biostabilization of DNAPL coal tar
Creator:
Isleyen, Mehmet
Publication Date:
Language:
English
Physical Description:
xviii, 124 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Coal-tar ( lcsh )
Dense nonaqueous phase liquids ( lcsh )
Mass transfer ( lcsh )
Coal-tar -- Biodegradation ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 118-124).
General Note:
Department of Civil Engineering
Statement of Responsibility:
by Mehmet Isleyen.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
63802864 ( OCLC )
ocm63802864
Classification:
LD1193.E53 2004d I74 ( lcc )

Full Text
MASS TRANSFER, BIODEGRADATION AND BIOSTABILIZATION OF
DNAPL COAL TAR
by
Mehmet Isleyen
B.S., Ataturk University, 1992
MS., University of Colorado at Denver, 1998
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy-
Civil Engineering
2004
&-
i
I
V.,
A
71
t


by Mehmet Isleyen
All rights reserved.


This thesis for the Doctor of Philosophy
degree by
Mehmet Isleyen
has been approved
bv
07 IllfllZQlf
Date
Ram Ramaswami


Isleyen, Mehmet (Ph.D.. Civil Engineering)
Mass Transfer, Biodegradation and Biostabilization of DNAPL Coal Tar
Thesis directed by Associate Professor Anu Ramaswami
ABSTRACT
This thesis consists of three phases and examines mass transfer, biodegradation, and
biostabilization of dense non-aqueous-phase liquid (DNAPL) coal tar. Results
obtained from this thesis may be applied to control the release of Poly Aromatic
Hydrocarbons (PAHs) released from DNAPL pools entrapped in the subsurface.
Phase 1: Multi-substrate biodegradation kinetics, and the resulting overall biogrowth,
are measured and modeled in terms of sole substrate biokinetic parameters for the
case of complex mixtures of more than twenty PAHs released into water from
DNAPL coal tar. A multi-substrate Monod model that incorporated competitive
substrate inhibition among four representative PAHs (styrene, naphthalene, 2-methyl
naphthalene and phenanthrene) was found to be effective in describing the overall
biogrowth and biokinetics of PAH mixtures released into water from both fresh and
depleted coal tar.
IV


Phase 2: The objectives of phase 2 are (1) to experimentally evaluate the aqueous
phase equilibrium concentration of multiple HOCs dissolving simultaneously from a
coal tar pool under fresh and depleted conditions; and (2) to experimentally evaluate
steady state exit concentrations downstream from a DNAPL pool with varied system
parameters pertaining to pool length, aquifer dispersivity, groundwater velocity and
coal tar composition. The equilibrium test results showed that Raoults Law with the
ideal NAPL assumption is adequate to estimate both initial and dynamic changes in
coal tar-water equilibria. Packed bed porous media tests with fresh and depleted coal
tar also showed that an analytical model for the 2-d advective dispersive equation was
effective in describing pool dissolution over the long-term, with depletion of the
DNAPL.
Phase 3: The last phase focused on understanding the role of microorganisms in the
biodegradation and biostabilization of DNAPL coal tar in porous packed bed ID cell.
Not only did this phase demonstrate biodegradation and biostabilization of fresh coal
tar in porous bed cell but also it evaluated biodegradation and biostabilization of
depleted coal tar in cells. Biostabilization started to occur in 11 and 15 days in
depleted and fresh coal tar cells, respectively, and was shown to result in non detect
PAH concentrations downstream from the pool zone over 100 days.
v


This abstract accurately represents the content of the candidate's thesis. I recommend
its publication.
Signed
Anu Ramaswami
vi


ACKNOWLEDGEMENTS
I am very grateful to my advisor, Dr. Anu Ramaswami, for her never ending support,
technical expertise, encouragement, and patience with me during these past 5 years.
I also wish to thank all of my committee members, Joann Silverstein, Angela R.
Bielefeldt, Tissa H. Illangasekare, Timberley Roane, and Ram Ramaswami for their
help with my thesis.
Special thanks for my funding from:
The USEPA a STAR grant award number R 825961-01-0
The Turkish Ministry of Education
The University of Colorado-Denver Faculty Grant
Other people I would like to acknowledge are:
Dr. Tissa Illangasekare from Colorado School of Mines for allowing me to
use his laboratory.
Dr. Kevin Kelly from the Bureau of Reclamation for allowing me to use his
laboratory, IC, and HPLC.
Dr. Timberley Roane for helping me with the bacterial counts and allowing
me to use her biology laboratory.
Ed Moss for continually helping with broken equipment in the laboratory.
Korey J. Kadrmas for helping me with packing the cells and his dissolution
experiment results.
Kendra A. Morrison, Ellen Rubin, and the other students in the laboratory.
My wife, Elvan Isleyen, for always offering the greatest support.
My mother, uncle Dr. Yakup Isleyen, family, and friends for their
encouragement.


TABLE OF CONTENTS
Figures....................................................................xii
Tables.....................................................................xvii
Chapter
1. Phase 1 Biodegradation Kinetics of Complex PAH Mixtures
Released from Coal Tar.................................................1
1.1 Introduction.........................................................1
1.2 Model Formulation...................................................5
1.3 Materials............................................................8
1.3.1 Individual PAH Solutions.............................................8
1.3.2 Coal Tar.............................................................9
1.3.3 Aqueous-Phase PAH Mixtures Released from Fresh and
Aged Coal Tar.........................................................11
1.3 .4 Microbial Population................................................13
1 4 Methodology.........................................................14
1.5 Results.............................................................16
1.5.1 Biodegradation Rates of Individual PAHs.............................16
viii


1.5.2 Biodegradation of PAH Mixture Equilibrated with Fresh and
Aged Coal Tar.........................................................20
2. Phase 2 Sensitivity Analysis of DNAPL Pool Dissolution:
Experimental Evaluation with Multicomponent Coal Tar..................38
2.1 Introduction...........................................................38
2.2 Factors Affecting DNAPL Dissolution....................................42
2.3 Materials and Methods..................................................47
2.3.1 Coal Tar...............................................................47
2.3.2 Equilibrium Tests......................................................48
2.3.3 Packed Bed DNAPL Pool Dissolution Tests................................50
2.3.4 Sample Collection and Analysis.........................................51
2.4 Results................................................................54
2.4.1 Equilibrium Test.......................................................54
2.4.2 Packed Bed DNAPL Pool Dissolution Test with Glass Bead
Particles and Fresh Coal Tar (Experiments of Korey Kadrmas)...........58
2.4.3 Packed Bed DNAPL Pool Dissolution Tests with
Fresh and Depleted Coal Tar...........................................67
2.5 Conclusions............................................................72
IX


3. Phase 3 Experimental Evaluation of Biostabilization of DNAPL
Coal Tar Pools in 1-Dimensional Porous Media Systems...................74
3.1 Introduction............................................................74
3.2 Factors that Affect DNAPL Biostabilization..............................76
3.3 Materials and Method....................................................77
3.3.1 Bacterial Solution......................................................77
3.3.2 Sand....................................................................78
3.3.3 Coal Tar................................................................79
3.3.4 Glass Cells with Embedded DNAPL Reservoir...............................79
3.3.5 Coal Tar Adding.........................................................82
3.3.6 Surface Tension.........................................................84
3.3.7 Aqueous PAH Chemical Analysis...........................................84
3.3.8 Effluent Bacterial Count................................................85
3.3.9 Tracer Tests............................................................85
3.3.10 Direct Measurement of Soil Bacterial Concentration and Porosity........86
3.4 Results and Discussion..................................................86
3.4.1 Aqueous Phase PAH Concentrations........................................86
3.4.2 Aqueous Phase Microbes..................................................96
3.4.3 Surface Tension and Dissolved Oxygen....................................99
3.4.4 Tracer Tests...........................................................102
x


3.4.5 Microbial Concentrations in the Cells................................107
3.5 Conclusions..........................................................116
References..................................................................118
xi


LIST OF FIGURES
Figure
1.1 Coal Tar CSTR.........................................................12
1.2 Biodegradation Kinetics for Individual Substrates Showing
Changes in Substrate and Biomass Concentrations......................17
1.2a Styrene...............................................................17
1.2b Naphthalene...........................................................18
1.2c 2-Methylnaphthalene...................................................18
1.2d Phenanthrene..........................................................19
1.3 Biodegradation of PAHs in Mixture-A........................21
1.3a Styrene...............................................................21
1.3b Naphthalene...........................................................22
1.3c 2-Methylnaphthalene...................................................22
1.3 d Phenanthrene..........................................................23
1.4 Biodegradation of PAHs in Mixture-B...................................24
1.4a Styrene...............................................................24
1.4b Naphthalene...........................................................25
1.4c 2-Methylnaphthalene...................................................25
1. 4d Phenanthrene..........................................................26
xii


1.5 Biogrowth in Mixtures..............................................30
1.5a Biogrowth of PAH-Mixture-A (Fresh Coal Tar)........................30
1.5b Biogrowth of PAH-Mixture-B (Aged Coal Tar).........................31
1.6 Biogrowth Contributions............................................32
1 6a Aqueous PAH-Mixture-A (Equilibrated with Fresh Coal Tar)...........32
1.6b Aqueous PAH-Mixture-B (Equilibrated with Aged Coal Tar)............32
1.7 Biodegradation of Acenaphthylene and Acenaphthene
in Mixture-A and Mixture-B........................................34
1.7a Biodegradation of Acenaphthylene in Mixture-A......................34
1 7b Biodegradation of Acenaphthene in Mixture-A........................35
1.7c Biodegradation of Acenaphthylene in Mixture-B......................35
1 7d Biodegradation of Acenaphthene in Mixture-B........................36
2.1 DNAPL Coal Tar Source Migration at a Manufactured
Gas Plant (MGP) Site..............................................41
2.2 Schematic of the DNAPL Pool Dissolution Model......................44
2.3 Coal Tar CSTR......................................................49
2.4 Packed Bed DNAPL Cell Schematic....................................53
2.5 Equilibrium Test Comparison of Predicted Verses Measured
HOC Concentration.................................................56
2.5a Fresh Coal Tar......................................................56
xiii


2.5b Depleted Coal Tar......................................................57
2.6 Packed Bed DNAPL Cell Used in the Packed Bed DNAPL
Pool Dissolution Test.................................................60
2.7a Comparison of Naphthalene Relative Concentrations at
Different Water Velocities............................................61
2.7b Comparison of Naphthalene Relative Concentrations for Varied
Pool Lengths..........................................................62
2.7c Comparison of Naphthalene Relative Concentrations for Varied
Particle Sizes........................................................63
2 8 Aqueous Phase HOCs for Fresh and Depleted Coal Tar Cells...............70
2.8a Naphthalene............................................................70
2.8b 2-Methylnaphthalene....................................................70
2.8c Styrene................................................................71
2 8d Acenaphthene...........................................................71
3.1 Porous Packed Cells....................................................81
3.2 Average Aqueous Phase Naphthalene Concentrations from the Bioactive
and Control Cells Having Fresh and Depleted Coal Tar as DNAPL
Reservoir.............................................................89
3.2a Fresh Coal Tar Cells...................................................89
3.2b Depleted Coal Tar Cells................................................90
xiv


3.3 Average Aqueous Phase 2 Methylnaphthalene Concentrations from the
Bioactive and Control Cells Having Fresh and Depleted Coal Tar as
DNAPL Reservoir.......................................................92
3.3a Fresh Coal Tar Cells....................................................92
3.3b Depleted Coal Tar Cells.................................................93
3 4 Average Aqueous Phase Styrene Concentrations from the Bioactive
and Control Cells Having Fresh and Depleted Coal Tar as DNAPL
Reservoir.............................................................95
3.4a Fresh Coal Tar Cells....................................................95
3.4b Depleted Coal Tar Cells.................................................96
3.5 Aqueous Phase Bacterial Count...........................................98
3.6 Surface Tension........................................................100
3.7 Tracer Test Results....................................................103
3.7a Bioactive Fresh Coal Tar Cell-A........................................103
3.7b Bioactive Fresh Coal Tar Cell-B........................................104
3.7c Bioactive Depleted Coal Tar Cell-C.....................................104
3.7d Bioactive Depleted Coal Tar Cell-D.....................................105
3.7e Control Fresh Coal Tar Cell-E..........................................105
3.7f Control Fresh Coal Tar Cell-F..........................................106
3.7g Control Depleted Coal Tar Cell-G.......................................106
xv


3.8 Biogrowth in the Cell Sections.........................................108
3.8a Fresh Bioactive Cell A (z=2,y=2, and y=4)...........................108
3.8b Fresh Bioactive Cell A (z=4,y=2, and y=4)...........................108
3.8c Fresh Bioactive Cell B (z=2,y=2, and y=4).............................109
3.8d Fresh Bioactive Cell B (z=4,y=2, and y=4).............................109
3.8e Flushed Bioactive Cell C (z=2,y=2, and y=4).........................110
3.8f Flushed Bioactive Cell C (z=4,y=2, and y=4).........................110
3.8g Flushed Bioactive Cell D (z=2,y=2, and y=4).........................Ill
3.8h Flushed Bioactive Cell D (z=4,y=2, and y=4).........................Ill
3.9 Porosity Profile in the Cell Sections..................................112
3.9a Fresh Bioactive Cell A...............................................112
3.9b Fresh Bioactive Cell B...............................................113
3.9c Flushed Bioactive Cell C.............................................113
3.9d Flushed Bioactive Cell D..............................................114
3.9e Fresh Control Cell E.................................................114
3.9f Fresh Control Cell F.................................................115
3.9g Flushed Control Cell G................................................115
xvi


LIST OF TABLES
Table
1.1 Main Constituents of Coal Tar Used for the Experiments................10
12 Best Fit Monod Model Parameters Describing the Sole Substrate
Degradation Kinetics of the 4 Representative PAHs.....................20
1.3 Microbe-Normalized Initial Degradation Rates Over First Two- Four
Hours in Sole Substrate (Individual) and PAH Mixture Solutions........27
2.1 Chemical Composition of Fresh and Depleted Coal Tar...................54
2 2 Experimental Parameters Varied for the Packed Bed DNAPL Pool
Dissolution Test......................................................59
2.3 Naphthalene Results for Varying Cell Parameters.......................64
2.4 Styrene Results for Varying Reactor Parameters........................66
2.5 2-Methyl Naphthalene Results for Varying Reactor Parameters...........67
2.6 Aqueous Phase HOCs for Fresh and Depleted Coal Tar Cells..............69
3.1 Average Aqueous Phase Naphthalene Concentrations from the
Bioactive and Control Cells Having Fresh and Depleted Coal Tar
as DNAPL Reservoir....................................................88
3.2 Average Aqueous Phase 2-Methylnaphthalene Concentrations from the
Bioactive and Control Cells Having Fresh and Depleted Coal Tar as
DNAPL Reservoir.......................................................91
xvii


3.3 Average Aqueous Phase Styrene Concentrations from the Bioactive and
Control Cells Having Fresh and Depleted Coal Tar as DNAPL Reservoir. 94
3.4 Surface Tension.......................................................99
3.5 Effluent Dissolved Oxygen Concentrations..............................101
xvm


1. Phase 1 Biodegradation Kinetics of Complex PAH
Mixtures Released from Coal Tar
1.1 Introduction
Microbially assisted bioremediation is widely used at land sites and aquifers
contaminated with organic pollutants (Klecka et al., 1990; Wilson and Jones, 1993;
Broholm and Arvin, 2000). Microbial processes provide the potential of transforming
organic pollutants to benign products, including mineralization to carbon dioxide.
Bioremediation has traditionally been applied to control groundwater plumes found at
many contaminated sites such efforts focus on biodegradation of a single
contaminant or a small group of contaminants migrating in water (e g., Borden and
Bedient, 1986, NRC, 1993, Bouwer et al., 1994). Few studies have addressed direct
biological treatment of the contamination source zones that often contain complex
multi component dense non-aqueous phase liquids (DNAPLs) such as coal tar or
Aroclors composed of polychlorinated biphenyl (PCBs). In recent years, biological
stabilization and isolation of the DNAPL source zone has been proposed as a method
of managing the plume emanating from the DNAPL source areas (Ramaswami et al.,
2001). Biostabilization refers to microbial activity in the vicinity of the DNAPL
source zone that can result in plume prevention, and a post-degradation DNAPL
1


residue composed primarily of insoluble compounds. In addition to microbial
degradation of contaminants released from the DNAPL source, biostabilization can
also result from biogrowth and clogging of porous media (Bielefeldt et al., 2002), that
further isolates the DNAPL zone. A complete understanding of DNAPL
biostabilization requires information on the rates of biodegradation of the complex
mixture of chemicals released to water from the multicomponent DNAPL and
quantification of biomass growth arising from such degradation. Aqueous phase
bioprocess would need to be linked with DNAPL zone mass transfer phenomena
(e g., Johansen and Pankow, 1992; Hunt et al., 1998,) and groundwater zone bacterial
and chemical transport processes. The focus of this phase is on examining the
aqueous phase biodegradation kinetics of a complex mixture of polycyclic aromatic
hydrocarbons (PAHs) released from DNAPL coal tar.
Coal tar is a DNAPL produced at former manufactured gas plant (MGP) sites (Luthy
et al., 1994). Coal tar is the principal by-product of coal gasification that was applied
in the early 1900s in the USA to produce combustible gaseous fuel from coal, coke
and oil. During the gasification processes, heavy organic-phase liquid coal tar
residues were produced from coal. The residual coal tar liquids are heavier than water
and have hundreds to thousands of aromatic organic compounds including PAH
compounds. The composition and physical properties of coal tar can vary widely
depending on the operating temperatures, feed stock and the type of gasification
2


process applied in the coal gasification plant (Villaume, 1984). Hazardous waste site
management at coal tar spill sites is complicated by the fact that PAH contaminants in
coal tar typically exist as very complex mixtures, of which no single compound is
predominant. Hence the focus of this phase is on examining the simultaneous
degradation of multiple PAHs released to water from coal tar.
Several studies have quantitatively examined the biodegradation of single aromatic
substrates in water (e g., Bateman et al., 1986; Bestetti et al., 1989; Foght et al., 1990;
Mihelcic and Luthy, 1991, Laha and Luthy, 1991). Some studies have examined the
simultaneous biodegradation of mixtures of aromatic compounds by pure (Machado
and Grady, 1989, Oh et al., 1994; Bielefeldt and Stensel, 1999; Reardon et al., 2000)
and mixed (Klecka and Maier, 1987; Chang et al., 1993, Guha et al, 1999, Knightes
and Peters, 2000) cultures, in all cases employing synthetically created solutions with
no more that three compounds. No study to date has attempted to quantify the
degradation by a mixed consortium of bacteria, of a range of PAH mixtures released
from a complex real-world DNAPL such as coal tar, nor characterize the biogrowth
associated with such degradation. Further, another unique aspect of this work is that
the changes in patterns of biodegradation of aqueous PAH mixtures are studied in the
context of a changing DNAPL composition as would occur in the field upon long-
term flushing and aging.
3


The objective of this study was to quantify the biodegradation kinetics of multiple
PAHs released into water from fresh and aged (depleted) coal tar, and the resulting
overall biogrowth, based upon individual PAH biokinetic parameters. The hypothesis
is that the biodegradation of a complex aqueous phase mixture of PAHs can be
described employing biokinetic parameters measured for a small subset of four
representative aromatic compounds. The representative PAHs chosen were: styrene
for 1 ring aromatic compounds, naphthalene for two-ring non-methylated compounds,
2-methyl naphthalene for methylated 2 ring compounds, and phenanthrene
representing 3 ring and larger PAHs released from coal tar. Biokinetic parameters
were first determined for the four representative PAHs present as single substrate in
solution. The biodegradation kinetics of a mixture of PAHs was measured by
focusing largely on the same four representative PAHs within the mixture. Overall
biogrowth resulting from PAH mixture degradation was also measured. A multi-
substrate model that incorporated competitive substrate inhibition and accrued
microbial growth from all PAH components was tested to describe the biodegradation
behavior of the PAH mixture.
4


1.2 Model Formulation
The multisubstrate model is formulated in Equations 1.1-1.2. Equation 1.1 shows the
general form of the Monod equation for degradation of a single substrate, /, in the
absence of inhibition or competitive effects:
dC,(t)
dt
= -V,
C,(t)
Ks,,+C,(t)
X,{t)
t))_y dC(t) bx
v
dt

dt
Equation 1.1a
Equation 1.1b
In Equation 1.1, C, (/) is the non-inhibitory aqueous concentration of substrate i,
(mass/volume) at any time, t, X(t) is the microbial concentration (mass/volume) at
time, t; KSI the half saturation coefficient (mass/volume), and jumi the maximum
specific microbial growth rate (mg substrate/mg biomass/time) for substrate /. Yi is
the true growth yield (biomass/substrate-mass), which quantifies the relationship
between microbial growth rate and the substrate degradation rate, b is an endogenous
decay constant (1/time). Equations 1.1a and 1.1b together describe the rate of
degradation of a single substrate, i. The Monod parameters for the individual PAHs,
KS I and fjm i, were determined by numerical integration of Equation 1.1a, 1.1b and
fitting the model results with experimental data on single substrate biodegradation
kinetics.
5


When multiple substrates are present in aqueous solution, the degradation rate of
substrate / in a system on n substrates, with co-substrate competitive effects, is
represented as (Chang et al., 1993, Guha et al., 1999):
dt
~ Mm,iCiO)
n-1
Ks,i + C,(f) + ^
J
X(t)
Equation 1.2
Where C,(/) represents the aqueous concentration of substrate i at any time t. The
summation term in the denominator of Equation 1.2 represents inhibition of
degradation of substrate / due to the presence of (n-1) competing substrates, j, at
concentration C;(t).
In the absence of competing substrates, Equation 1.2 reverts to the non-inhibitory
Monod model shown in Equation 1.1. The parameters nm,, Ks i, and KS J represent
individual PAH biokinetic parameters, n is the number of representative chemicals or
target chemicals chosen to describe a complex mixture. Thus, in a system composed
of 4 target chemicals (n=4) biodegradation of any one component is modeled to be
impacted by the remaining (n-1 =3) components. The biodegradation of the entire
complex mixture is assumed to be described effectively by aggregate behavior of the
target (representative) chemicals. Other chemicals interaction can also occur in
mixtures, but, competitive substrate inhibition is typically observed when multiple
6


substrates of similar structure are metabolized using same or similar enzymes
(Reardon et al, 2000; Guha et al., 1999; Shuttleworth and Cerniglia, 1996).
The total biomass growth rate is the summation of biomass growth contributed by all
representative substrates, and is shown as:
Equation 1.2a
at ^ at ^ at
/=! t=l
Numerical integration of Equation 1.2 and 1.2a, employing estimates of umi,K5i,
and determined form single substrate degradation data, is expected to predict the
biodegradation of various components of a PAH mixture, as well as the overall
biogrowth resulting from such biodegradation. Model results obtained from Equation
1.2 and 1.2a, with four (n=4) representative chemicals, were compared with
biodegradation data obtained for a complex mixture of PAHs released from fresh and
aged coal tar. Thus the modified Monod model with competitive substrate inhibition
between a representative set of four component chemicals, was evaluated to
determine efficacy in describing the biokinetics of a complex PAH mixture released
from coal tar.
7


1.3 Materials
1.3.1 Individual PAH Solutions
Individual solutions of the target aromatic compounds were made up by preparing a
saturated aqueous solution in contact with the solid PAHs, naphthalene, 2
methylnaphthalene, and phenanthrene, filtering the saturated solution on a glass filter
to remove any solids, and diluting appropriately with de-ionized (DI) water in order
to attain 14 mg/L, 1.4 mg/L, and 0.3 mg/L, of these chemicals, respectively. Styrene
solutions were prepared by adding an appropriate mass of neat liquid styrene into DI
water to obtain a concentration 0.4 mg/L. The concentrations of the representative
chemicals were chosen to be similar to those observed in water at equilibrium with
fresh coal tar. All the solutions were prepared in tightly closed glass bottles and
continuously stirred with a teflon-coated magnetic stirrer over one day. The initial
concentration of the aqueous solution in the bottles was confirmed by High Pressure
Liquid Chromatography (HPLC) analysis and the individual PAH solutions were then
used in tests to determine individual PAH biokinetic parameters.
8


1.3.2 Coal Tar
Coal tar used in the experiments was obtained from the subsurface (well# AZ-W07-S)
from a contaminated site in Baltimore, MD (MacFarlane et al., 1990). The coal tar
was stored in airtight amber bottles in a refrigerator. The chemical composition of the
coal tar is shown in Table 1.1, and indicates that this DNAPL contains more than
twenty identifiable aromatic compounds. Previous studies have shown that Raoults
Law with the ideal DNAPL assumption is effective in describing the equilibrium
concentration of coal tar constituents in water (Lee et al., 1992, Ramaswami et al
1994, Peters et al., 1999; Ramaswami et al 2001). In this phase we examine the
hypothesis that tracking the biokinetic behavior of a few (four) representative
compounds suffices to describe the biodegradation of the array of PAHs released to
water from coal tar.
9


Table 1.1: Main Constituents of Coal Tar Used for the Experiments
(Provided by Barringer Laboratories, Inc. Golden, CO)
PAHs Mole Fractions (%)
Naphthalene 8.45
2-Methylnaphthalene 4.64
Acenaphthylene b.042
Acenaphthene 0.87
Dibenzofuran 0.3166
Fluorene 0.794
Phenanthrene 1.5536
Anthracene 0.5258
Di-n-butylphthalate 0.3214
Fluoranthene 0.526
Benzo( a)anthracene 0.233
bis(2-Ethylhexyl) phthalate 0.13
Benzo(b) fluoranthene 0.11
Benzo(k)fluoranthene 0.092
Benzo(a) pyrene 0.1435
Indole(l,2,3-cd) pyrene 0.0563
Dibenz(a,h) anthracene jO.0199
Benzo(g,h,I) perylene 0.057
Volatile Organic Compounds!
Styrene 0.158
Ethylbenzene 0.0998
Toluene 0.152
Xylene 0.0767
Benzene 0.428
10


1.3.3 Aqueous-Phase PAH Mixtures Released from
Fresh and Aged Coal Tar
Two types of aqueous PAH mixtures were used for measuring and modeling the
biodegradation kinetics of the PAHs released from coal tar DNAPL: The first mixture
is an aqueous-phase PAH matrix equilibrated with fresh coal tar (PAH-mixture-A).
The second is a mixture containing several PAHs released from depleted coal tar
(PAH-mixture-B). The depleted coal tar was produced by flushing water around a
coal tar globule until approximately half the naphthalene in the DNAPL was depleted,
thus represents an aged DNAPL matrix. Aqueous solutions equilibrated with fresh
and aged coal tar were prepared in a coal tar CSTR shown in Figure 1.1, in which 600
ml of DI water was added to a mixing jar having an elevated teflon-coated magnetic
stirrer above a submerged 4 ml globule of coal tar.
The water in the coal tar CSTR was continuously mixed with the elevated stirrer at
gentle rotational speeds such that the coal tar globule at the bottom of the jar
remained undisturbed. Aqueous phase naphthalene concentrations monitored in this
coal tar CSTR showed that equilibrium concentrations were attained in less than two
days of mixing. The equilibrated aqueous-phase solution, containing a mixture of all
the PAHs released from fresh coal tar (PAH-mixture-A), was withdrawn through a
needle placed in the side arm of the jar and used in biodegradation tests for evaluating
PAH mixture biokinetics in fresh systems The coal tar CSTR was flushed with DI
11


water through stainless steel tubing inserted in the side arm of the jar, until
equilibrium aqueous phase naphthalene concentrations in the CSTR indicated that
approximately 60% of the naphthalene had been depleted from the coal tar. The
solution equilibrated with the depleted (aged) coal tar, termed PAH-mixture-B, was
withdrawn through a needle placed in the jar and used in biodegradation tests for
evaluating PAH mixture biokinetics.
Figure 1.1: Coal Tar CSTR: A 4 mL coal tar globule is placed at the bottom of the
jar containing 600 mL of water. The water is gently mixed by means of an elevated
magnetic stir bar. Coal tar may be sampled by a syringe needle placed on the lid of
the jar; Aqueous phase samples were extracted from a needle inserted into the side
arm of the jar.
12


1.3.4 Microbial Population
Microbes used in this project included a mixed consortiums of aerobic naphthalene
and phenanthrene degrading bacteria obtained from Johns Hopkins University and
Michigan Tech University. The microbes were maintained in a Bio-CSTR
continuously fed with 100 ml of High Biological Oxygen Demand (HBOD) nutrient
solution per day to which was added naphthalene, 2-methlynaphthalene, styrene, and
phenanthrene as carbon source. The HBOD nutrient media used for the Bio-CSTR
consisted of 0.25 mg/L of FeCl3*6H20, 850.0 mg/L of NH4CI, 170.0 mg/L of
KH2PO4, 435.0 mg/L of K2HP04, 22.5 mg/L of MgS04*7H20, 608.0 mg/L of
Na2HP04 and 1100 mg/L of CaCl2. Approximately four drops of coal tar was also
added to the Bio-CSTR every month in order to acclimate the microbes to the range
of PAHs in coal tar Hydraulic and solids retention time in the Bio-CSTR are equal to
seven days. Pure oxygen was replenished in the system every two days. The bacterial
consortium in the Bio-CSTR was identified as predominantly of the Pseudomonas
species, including Stenotrophomonas maltophilia, Pseudomonas fluorescens,
Pseudomonas putida, and Pseudomonas stutzeri (Morrison, 2000). Small quantities
of the bacterial suspension were taken out from the Bio-CSTR and used as inoculum
for the biokinetic experiments.
13


1.4 Methodology
Biokinetic experiments were conducted in duplicate 220 ml glass bottles with Teflon
faced septa caps. Four sets of duplicate bottles were used to measure rates of
biodegradation of the four individual PAHs, naphthalene, 2-methyl naphthalene,
phenanthrene, and styrene. Two additional sets of bottles were used to measure
degradation kinetics of PAH-mixture-A and PAH-mixture-B that had previously
equilibrated with fresh and flushed coal tar, respectively. Each bottle had a total
volume of 220 ml and contained 20 ml of headspace such that sufficient air was
available for aerobic biodegradation. The aqueous solution in each vial was made up
of 20 ml of HBOD nutrient solution, 10 ml of bacterial solution obtained from the
Bio-CSTR, and, 165 ml of substrate solution (individual PAH solution or coal tar-
equilibrated PAH mixture A or B) The bottles were placed in an orbital shaker
(Thermolyne), and gently mixed at 100 rpm.
Kill control bottles were likewise maintained in which the added bacteria were killed
by the addition of 5 ml of 16g/L of mercuric chloride (HgCh) solution. All other
bottles had 5 ml of DI water added for.consistency with the 5 ml of mercuric chloride
solution added to the killed-control bottles. 2 ml samples from individual-PAH and
PAH-mixture bottles were collected periodically in crimp-cap vials over the
experimental period. All samples were spiked with HgCb to stop microbial activity,
and stored at 4C in a refrigerator. The collected samples were placed on HPLC auto-
14


sampler (HP 1100) and analyzed for aqueous phase PAH concentrations. The HPLC
method employed UV (254, 240, and 230 nM wavelengths) detection in a LPAH
column (Supelco), operated at 40 C with 1.5 ml/min of 40:60 acetonitrile:water
solution. The analytical error on the HPLC ranged from 1-3%. Detection limits
ranged from 5-20 ppb for the four target PAHs.
After HPLC analysis, a homogenized 0.9 ml of the aqueous sample was withdrawn
from the sample vials to measure the total biomass concentrations. The 0.9 ml
aqueous solution sample was diluted appropriately and mixed with 0.1 ml of acridine
orange dye, then filtered on black Nucleopore (0.2 pm) filter. The filter was double
rinsed with filtered distilled water and observed under a fluorescent microscope.
Bacterial counts were averaged over 20 fields; standard errors of duplicates were 1.4-
10.6 %. Bacterial counts were converted to mg biomass concentrations in water. In
this manner, biomass concentrations in water were monitored along with PAH
concentrations, enabling PAH biokinetic parameters to be determined.
15


1.5 Results
1.5.1 Biodegradation Rates of Individual PAHs
Time varying aqueous concentrations of the four representative individual PAH
substrates, and associated biogrowth data, are shown in Figure 1.2. The control vials
with inactive microbes showed a small decrease (5.6-13%) in aqueous PAH
concentrations over the one-day test, demonstrating integrity of the system. In
contrast, as seen in Figure 1.2, all the bioactive bottles showed a marked decrease in
individual PAH concentrations with rapid removal of all four PAH compounds to
below detection limit levels in less than 12 hours. A Monod model was fit to describe
the change in substrate concentration in the bioactive vials after correcting for the loss
observed in the control vials. The Monod model was fit to the observed data by
numerical integration of Equations la and lb, (with At = 0.01 hours) using an Excel
Spreadsheet (Laurence et al., 1998).
16


Figure 1.2: Biodegradation Kinetics for Individual Substrates Showing Changes
in Substrate and Biomass Concentrations. Solid () and open (o) symbols
represent observed data from replicate bioactive and control systems, respectively.
Error bars represent the standard deviation between replicates. The bold line (-)
represents the fitted Monod model:
Figure 1.2a Styrene
0 5
S(model)
S(data)
OS (control)
Time(hr)
17


Figure 1.2b Naphthalene
16,
Time(hr)
- S (model)
S(data)
OS (control)
4.5 -
7ime(hr)
- X(model)
X(data)
Figure 1.2c 2-Methylnaphthalene
16 ^
S (model)
S(data)
O S (control)
Time(hr)
0.8
I
0 J---,-,--,--,----:
0 5 10 15 20 25 30
Time(hr)
18


Figure 1.2d Phenanthrene
0.4
- S(model)
S(data)
O S(control) i
Time(hr)
0.3
ra
E,
d
c
o
O
tn
tn
co
E
o
CO
0.2 J
0.1
X(model)
X(data)
0 i----1-r-#---1
0 5 10 15 20 25 30
Time(hr)
The Monod parameters, Ks, (im and Y were determined for each individual PAH
compound by minimizing the sum of the squared error (or differences) between
observations and model predictions. The best-fit parameters were robust and did not
vary significantly upon changing the time step of the integration, At. The best-fit
Monod parameters for the four individual PAHs are summarized in Table 1.2. As
shown in Figures 1.2, the Monod model with these best fit parameters matched the
time-profile of both biomass and aqueous PAH concentrations fairly well. The
patterns of degradation of the individual PAHs were then compared with those
19


observed for the same set of four representative chemicals present in aqueous PAH
mixtures.
Table 1.2: Best Fit Monod Model Parameters Describing the Sole Substrate
Degradation Kinetics of the 4 Representative PAHs
Substrate Monod Model Parameters? Styrene Naphthalene 2 Methyl- Naphthalene Phenanthrene
pm (mg substrate/mg biomass/hr) 0.62 0.83 0.474 0.15
Ks (mg/L) 0.17 0.218 0.1 0.034
Y (mg biomass/mg substrate) 0.0091 0.26 0.319 0.093
1.5.2 Biodegradation of PAH Mixture Equilibrated
with Fresh and Aged Coal Tar
The biodegradation of PAH-mixture-A in water that was isolated after equilibration
with fresh coal tar is shown in Figures 1.3a-1.3d. The biodegradation of the PAH-
mixture-B in water equilibrated with aged (depleted) coal tar is shown in Figures
1.4a-1.4d The same four representative chemicals that were evaluated individually
in Figures 1.2a-1.2d, are examined in the presence of PAH mixtures in Figure 1.3a-
1.3d, 1.4a-1.4d.
20


Figure 1.3: Biodegradation of PAHs in Mixture-A
Biodegradation kinetics for substrate mixtures showing changes in substrate and
biomass concentrations. Solid () and open (o) symbols represent observed data from
replicate bioactive and control systems, respectively. Error bars represent the standard
deviation between replicates. The bold line (-) represents the fitted multisubstrate
Monod model:
Figure 1.3a Styrene
0.4
0.35 -i
- Sty (Model)
Sty (data)
o Sty(control)
0 5 10 15 20 25 30
Time(hr)
21


Figure 1.3b Naphthalene
14 -
Time(hr)
| Nap(model)
i Nap(data)
(o Nap(control)
I______________
Figure 1.3c 2-Methylnaphthalene
1.2
2MNaph( model)
2MNaph(data)
O 2MNaph(control)
71me(hr)
22


Figure 1.3d Phenanthrene
0.16 n
Time(hr)
Phen( Model)
Phen(data)
O Phen(control)
23


Figure 1.4: Biodegradation of PAHs in Mixture-B
Biodegradation kinetics for substrate mixtures showing changes in substrate and
biomass concentrations. Solid () and open (o) symbols represent observed data from
replicate bioactive and control systems, respectively. Error bars represent the standard
deviation between replicates. The bold line (-) represents the fitted multi substrate
Monod model:
Figure 1.4a Styrene
0.09
0.08
. Sty(Model)
Sty(datal)
o Sty(cont)
0 5 10 15 20 25 30
Time(hr)
24


Figure 1.4b Naphthalene
4.5 -
Time(hr)
- Nap(rrodel)
Nap(data)
O Nap(cont)
Figure 1.4c 2-Methylnaphthalene
1.4
2MNaph(model)
2MNaph(datal)
02MNaph(contl)
Time(hr)
25


Figure 1.4d Phenanthrene
As can be seen from the slower initial slopes of the degradation of these chemicals,
co-substrate inhibition appears to be occurring for most of these classes of chemicals.
A quantitative comparison of the initial degradation rates across the PAH-in-a-
mixture system and the individual PAH system offers many useful insights and is
shown in Table 1.3. Since initial microbial concentrations were similar in all the
bioactive bottles, a microbe normalized initial biodegradation rate was computed, as
shown below:
Microbe Normalized Initial Degradation Rate

1
(initial)
dC^
)(initial)
Equation 1.3
26


Table 1.3: Microbe-Normalized Initial Degradation Rates Over First Two-Four
Hours in Sole Substrate (Individual) and PAH Mixture Solutions
Chemical Individual PAH-Mixture-A PAH-Mixture-B
Styrene 0.62 0.050 No Degradation
Naphthalene 0.83 0.874 0.031
2 methylnaphthalene 0.474 0.219 0.012
Phenanthrene 0.15 0.0281 No Degradation
Comparing the microbe-normalized initial degradation rate in mixture versus single
substrate solutions offers direct evidence of competition. The initial normalized
degradation rate should mathematically be approximately equal to the pm for the
individual PAH (Equation 1 with KsC,(t)), as long as any competitive or inhibitory
effects offered by the mixture are absent. From the comparisons shown in Table 1.2,
2methylnaphthalene, phenanthrene, and styrene appear to be distinctly inhibited when
present in PAH-mixture-A compared with the corresponding individual PAH
solutions. Degradation rates of all four target chemicals, styrene, naphthalene, 2-
methylnapthalene, and phenanthrane, appear to be distinctly inhibited when present in
PAH-mixture-B.
A modified Monod model, that incorporates competitive substrate inhibition was used
to model the degradation of aqueous phase PAH mixtures derived from contact with
fresh and depleted coal tar, as well as the biogrowth arising from such simultaneous
degradation of the mixtures of PAHs Data analysis indicated that the biodegradation
27


of naphthalene was competitively inhibited only by the presence of 2-methyl
naphthalene. In PAH-mixture-A, since naphthalene was the dominant chemical at a
concentration of 13 mg/L, little competitive inhibition by 2-methyl naphthalene was
effectively observed. In contrast, competitive inhibition of naphthalene becomes
apparent in PAH-mixture-B (see Table 1.3), when naphthalene and 2-methyl
naphthalene concentrations are comparable. While naphthalene degradation was
assumed to be inhibited only by 2 methyl naphthalene, the remaining target PAHs
were all modeled to be competitively inhibited by each other and naphthalene. Thus,
2-methyl naphthalene degradation was affected by competition from naphthalene,
styrene, and phenanthrene. Phenanthrene was affected by naphthalene, 2-methyl
naphthalene, and styrene, etc., as indicated by Equations 1.2, 1.2a.
The competitive inhibition model was implemented with the above assumptions,
employing the biokinetic parameters (Table 1.2) determined for four individual
representative compounds, among the more than twenty PAHs present in coal tar
(Table 1.1). Numerical integration of Equations 1.2&1.2a yielded model predictions
of the concentration versus time biodegradation profiles of the four target PAHs
present in aqueous mixtures derived from contact with coal tar. Model predictions for
the four target PAHs in both mixtures A and B, matched observation very well, as
indicated by the bold lines in Figures 1.3&1.4. The total biomass growth is the
summation of biomass growth contributed by all representative substrates, as shown
28


in Equation (1.2a). Once again, biomass prediction based on the Monod mixture
model with consideration of four target compounds, matched very well with
biogrowth observed in PAH-mixture-A and PAH-mixture-B, as shown in Figure
1 5a&l 5b, respectively. The good agreement between model and data is remarkable
given the complexity of the aqueous PAH mixtures obtained from coal tar, and the
absence of any new fitting parameters to predict aqueous phase PAH mixture
biodegradation and resulting biogrowth Biogrowth is observed to be greater in PAH-
mixture-A (Figure 1.5a) than PAH-mixture-B (Figure 1.5b) due to higher levels of
naphthalene in mixture-A. Analysis of model simulations at peak conditions indicated
92.7 % and 7.2% of biogrowth in PAH-mixture-A and 70.8 % and 29.1 % of
biogrowth in PAH-mixture-B was contributed by naphthalene and 2methy
naphthalene, respectively (Figure 1.6a, 1.6b). Figure 1.6a demonstrates the relative
dominance of naphthalene and 2 methyl naphthalene in biokinetic modeling of PAHs
mixtures. These results show that individual biokinetic parameters determined for a
small subset of representative PAHs are effective in describing the biodegradation of
a complex mixture of PAHs released from both fresh and depleted coal tar.
29


Figure 1.5: Biogrowth in Mixtures
Biodegradation kinetics for substrate mixtures showing changes in Biomass
concentrations. Solid () symbol represents experimental observation and the bold
line (-) represents the fitted multi substrate Monod model:
Figure 1.5a Biogrowth of PAH-Mixture-A (Fresh Coal Tar)
Time(hr)
30


Figure 1.5b Biogrowth of PAH-Mixture-B (Aged Coal Tar)
2
i
Time(hr)
31


Figure 1.6: Biogrowth Contributions
Biogrowth contribution due to biodegradation of naphthalene and 2-methyl
naphthalene.
Figure 1.6a Aqueous PAH-Mixture-A (Equilibrated with Fresh Coal Tar)
Figure 1.6b Aqueous PAH-Mixture-B (Equilibrated with Aged Coal Tar)
32


An important hypothesis in this test is that the behavior of 3-ring and larger PAHs
would be fairly well represented by the biokinetic parameters and competitive
patterns determined for phenantharene. This hypothesis was tested by measuring
degradation of acenaphthene and acenaphthylene in PAH-mixture-A and PAH-
mixture-B, and comparing observations with Monod mixture model predictions that
employed individual biokinetic parameters determined for phenanthrene. The results
shown in Figure 1.7a-1.7d are in good agreement, justifying the use of phenanthrene
for describing the biodegradation behavior of higher ring PAHs fairly well.
33


Figure 1.7: Biodegradation of Acenaphthylene and Acenaphthene in Mixture-A
and Mixture-B
Biodegradation kinetics for substrate mixtures showing changes in substrate
concentration. Solid () and open (o) symbols represent observed data from replicate
bioactive and control systems, respectively. Error bars represent the standard
deviation between replicates. The bold line (-) represents the fitted multisubstrate
Monod model:
Figure 1.7a Biodegradation of Acenaphthylene in Mixture-A
0.6 -
0.5
- Anapyl(model)
Anapyl(data)
O Anapyl(cont)
0
0 5 10 15 20 25 30
Tlme(hr)
34


Figure 1.7b Biodegradation of Acenaphthene in Mixture-A
0.3
1
E 0 2 RT\
y 1 Anap(model)
O l Anap(data)
s- 0.1 : 1 O Anap(cont)
< U
o -i,-
0 5 10 15 20 25 30
Time(hr)
Figure 1.7c Biodegradation of Acenaphthylene in Mixture-B
0 5
Time(hr)
35


Figure 1.7d Biodegradation of Acenaphthene in Mixture-B
In summary, model simulations of PAH-mixture-A and PAH-mixture-B were found
to predict the observed behavior very well, both in terms of describing the time-
varying concentrations of the PAHs now within a mixture, as well as predicting the
expected growth in biomass due to the simultaneous degradation of multiple PAHs.
Note that no new biokinetic parameters were fitted to describe the behavior of PAHs
in mixture form. The good match between model predictions and experimental
observations indicates that individual biokinetic parameters determined for a small
subset of PAHs may be successfully used in a multisubstrate Monod model to
adequately describe both biomass growth and PAH concentration attenuation in
36


response to microbial degradation. Furthermore, the favorable results obtained with
waters equilibrated with both fresh and depleted coal tar, demonstrate for the first
time, that individual biokinetic parameters determined for a small subset of
representative chemicals may suffice to effectively describe the long term
biodegradation behavior of complex multicomponent DNAPLs such as coal tar.
37


2. Phase 2 Sensitivity Analysis of DNAPL Pool
Dissolution: Experimental Evaluation with
Multicomponent Coal Tar
2.1 Introduction
DNAPLs are dense non-aqueous phase liquids associated with long-term soil and
groundwater contamination when released into the subsurface. Some DNAPLs
commonly found at Superfund sites in the United States include chlorinated solvents
such as trichloroethylene (TCE), coal tar, polychlorinated biphenyl (PCB) soils and
creosote produced from wood-preserving operations (Johansen et al., 1999). Many
DNAPLs such as coal tar are complex multi-component liquids comprised of many
different hydrocarbons spanning a broad spectrum of molecular weights (Lee et al.,
1992, Luthy et al., 1994).
DNAPLs have a specific gravity greater than water that causes DNAPL spills to
migrate vertically in the subsurface through the vadose zone and the saturate
38


groundwater zones, and eventually accumulating as pools on bedrock and on low
permeable lenses encountered during downward migration (see Figure 2.1). On a
micro scale DNAPL ganglia are also trapped within the soil matrix, immobilized by a
balance of viscous, gravity and pressure forces (Hunt et al., 1988). Slow dissolution
of high-molecular weight chemicals from the DNAPL pools and ganglia trapped in
the subsurface causes long-term contamination of groundwater (Johansen et al., 1999;
Ramaswami et al., 2001). Although, many DNAPL constituents are hydrophobic
with low aqueous solubilities, the concentration of these constituents in groundwater
is large enough to pose a risk to human health and the environment (Lane and Loehr,
1992).
Since DNAPL pools contain a large proportion of the mass of DNAPL released to the
subsurface, this research examines the dissolution of contaminant chemicals released
from subsurface DNAPL pools to groundwater. Understanding the dissolution of
DNAPL constituents from DNAPL pools into the aqueous phase is a fundamental
issue surrounding DNAPL contamination of the subsurface. Also, techniques
proposed to stabilize DNAPL pools, e g. by microbial processes (Ramaswami et al.,
2001) also require information on rates of chemical dissolution from DNAPL pools.
A few studies have examined dissolution of a single component from pure DNAPL
pools, e g., trichloroethylene, trichloroethane, toluene, etc, such as Schwille (1988)
Johnson and Pankow (1992), Voudrias and Yeh (1994), and Pearce et al. (1994). No
39


study to date has evaluated the simultaneous dissolution of multiple compounds from
a complex multi-component DNAPL. Furthermore, the impact on dissolution rates of
a changing DNAPL composition due to long-term dissolution in an aquifer, has not
previously been examined. This paper evaluates key parameters that control the
dissolution rate of hydrophobic organic chemicals (HOCs) released from a multi-
component DNAPL. The multi-component DNAPL evaluated in this study is coal
tar.
Coal tar is a by-product from the coal gasification process used in the early 1900s to
produce combustible gaseous fuel from coal, coke, and oil (Luthy et al., 1994). The
gasification processes converted coal into combustible gaseous fuel, and yielded as a
by-product, heavy organic-phase liquid coal tar residuals. These residuals contained
hundreds to thousands of aromatic organic compounds, including polyaromatic
hydrocarbons (PAHs), which are the main constituents in coal tar (Lee et al., 1992,
Peters et al, 1993; Ramaswami et al., 1997). The DNAPL coal tar waste residuals are
now the primary source of soil and groundwater contamination at former
manufactured gas plants (MGP Sites).
40


Figure 2.1: DNAPL Coal Tar Source Migration at a Manufactured Gas Plant
(MGP) Site (Modified from Quling and Weaver, 1991)
41


2.2 Factors Affecting DNAPL Dissolution
DNAPL dissolution is controlled by two main factors (Pfannkuck, 1984, Hunt et al.,
1988): (a) the equilibrium boundary condition at the NAPL-water interface, and (b)
the interaction of the pool boundary with the groundwater flow field.
The concentration boundary condition at the NAPL-water interface is assumed to be
defined by equilibrium conditions described by Raoults Law (Lee et al, 1992,
Schwarzenbach et al., 1993; Peters and Luthy, 1993, Ramaswami et al, 1997) as:
CLq = Xnapl x Ci.sa, x Ynapl Equation 2.1
Where CWieq is the aqueous phase equilibrium concentration of a NAPL-derived
chemical, A; XNAPL is the mole fraction of the compound A in the
multicomponent NAPL; CW!Sat is the solubility of the pure liquid compound A in
water, and Xnapl is the activity coefficient of the compound A in the NAPL, set to
1 for an ideal NAPL (Schwarzenbach et al., 1993) Previous studies have verified the
effectiveness of using Raoults Law with the ideal NAPL assumption for complex
mixtures such as coal tar, gasoline and diesel fuel (Mercer and Cohen, 1990; Lee et
al., 1992, Ramaswami et al., 1997). However, the effectiveness of Raoults law in
determining dynamic changes in equilibrium at the tar-water interface as the
composition of the coal tar changes has yet to be verified for multiple PAHs released
simultaneously from coal tar.
42


The equilibrium concentration boundary condition at the NAPL-water interface
produces primarily a vertical flux of HOC from the NAPL-water surface given by
Ficks Law as:
F = -Dt x
rdC^
<5z j
Pool Surface
Equation 2.2
where F is the mass of solute transferred per unit area per unit time (M/L2T), Dz is the
transverse dispersion coefficient (L2T''), C is the solute concentration (M/L3); and
dC/dz is the vertical concentration gradient (Freeze and Cherry, 1979). The vertical
flux shown in Equation 2.2 is applicable over the pool length, and contributes to the
input of the DNAPL-derived chemical into the 2-d water flow field with Advective-
Dispersive transport described as:
5C(x,z,t)
dt
= -u.
dC(x,z,t)
dx
+ D.
d2C(x,
ox
J)
D,
d2C(x,z,t)
dz2
Equation 2.3
where C (x, z, t) is the solute concentration (M/L3) at location (x, z) at time t; u is the
average groundwater velocity in the x direction, z is the vertical direction, Dx is the
longitudinal dispersion coefficient, and Dz is the vertical transverse dispersion
coefficient (Freeze and Cherry, 1979) defined as:
Dz a, uz + De Equation 2.3a
43


where De is the effective molecular diffusion coefficient, and a, and a, are
longitudinal and transverse (vertical) dispersivity, respectively. The effective
molecular diffusivity for most organic chemicals is of the order of 10"6 cm2/sec (US.
EPA, 1996).
Figure 2.2: Schematic of the DNAPL Pool Dissolution Model
C (x=0.z) = 0
X
The dissolution process can be modeled as a steady-state process, since the time
required for total pool dissolution is extremely long in comparison with contact time
between the pool and the flowing groundwater. Furthermore, when advective
transport in the direction of groundwater flow (x) is much greater than the dispersive
44


transport in the direction of flow longitudinal dispersion can be neglected in Equation
2.3, yielding (Hunt et al., 1988, Seagren et al., 1994):
\ OX J
x, z > 0
Equation 2.4
Hunt et al. (1988) applied Laplace transforms to derive an analytical solution to
Equation 2.4, with the boundary conditions representing a semi-infinite domain in the
vertical dimension, no solute upstream of the pool, and equilibrium boundary
conditions at the NAPL-water interface.
Applying the analytical solution, the average aqueous HOC concentration at the end
of the pool zone was derived as (Hunt et al., 1988):
C(L) \-e
C
eq
0)yl7T
Equation 2.5
where:
co =
Y
2 (Z>,-£/(/,)'"
Equation 2.6
45


Equation 2.5 indicates that the three parameters that primarily determine the steady-
state aqueous chemical concentrations leaving the pool zone are the dispersion
coefficient (Dz), pool length (Lp), and Darcy velocity (ux). These variables were
experimentally varied during the packed bed DNAPL pool dissolution experiments to
test the sensitivity of these parameters in controlling the dissolution of multiple HOCs
from a multicomponent coal tar DNAPL pool. In addition, both fresh coal tar and a
sample of depleted (aged) coal tar were used in experiments to verify the applicability
of the pool dissolution model to simulate long-term dissolution, with varying coal tar
(DNAPL) composition, as occurs in real-world situations.
Thus, the objectives of this research are (1) to experimentally evaluate the aqueous
phase equilibrium concentration of multiple HOCs dissolving simultaneously from a
coal tar pool under fresh and depleted conditions, and (2) to experimental evaluate
steady state exit concentrations downstream from a DNAPL pool with varied system
parameters pertaining to pool length, aquifer dispersivity, groundwater velocity and
coal tar composition. The materials and methods used in the experiments are
described next.
46


2.3 Materials and Methods
2.3.1 Coal Tar
Two types of coal tar, fresh and depleted coal tar, have been used in this research.
The fresh coal tar used in this experiment was obtained from a former MGP site
located in Baltimore, Maryland. The depleted coal tar was produced from the fresh
coal tar. The fresh coal tar in an airtight CSTR was flushed with DI water through
stainless steel tubing inserted in the side arm of the CSTR, until equilibrium aqueous
phase naphthalene concentrations in the CSTR indicated that approximately 50% of
the naphthalene had been depleted from the fresh coal tar. The density of the coal tar
was determined to be approximately 1.05 g/ml. The coal tar composition was
measured by Gas Chromatography-Mass Spectrometry (GC-MS). EPA 8270 analysis
of the coal tar sample determined an abundance of HOCs, including several PAHs,
present in the sample. Based on their abundance, six hydrophobic organic chemicals
(HOCs) (naphthalene, styrene, 2-methyl-naphthalene, acenaphthene, and
acenaphthylene) were chosen as the primary representative chemicals within multi-
component coal tar, representing 1-ring, 2-ring, and 3-ring aromatic compounds.
47


2.3.2 Equilibrium Tests
Batch tests and flow-through tests were conducted in the well-mixed Coal Tar CSTR
shown in Figure 2.3, which contains water and a submerged coal tar DNAPL pool.
Initial equilibrium HOC concentrations in water in contact with the coal tar was
measured in the closed CSTR under no-flow conditions over a 5-day contact period
between coal tar and water. Two ml samples of water from the CSTR were extracted
periodically and assessed for aqueous phase HOC concentration using the HPLC.
When measured aqueous HOC concentrations had increased to a stable value,
equilibrium was considered to have been attained. Measured equilibrium aqueous
phase HOC concentrations were compared with those predicted from Raoults law
employing the ideal NAPL assumption in Equation 2.1.
48


Figure 2.3: Coal Tar CSTR
After initial equilibrium conditions were verified, water was flushed with DI water
through stainless steel tubing inserted in the side arm of the CSTR. Water was
pumped through the CSTR at a rate of 2.3 ml/minute, achieving a residence time of
260.86 minutes for the 600 ml volume of water in the CSTR. Water flow was
maintained at this rate for about 53 days, corresponding to about 50% depletion of
naphthalene relative to the initial coal tar, after which the pump was switched off and
49


the contents of the CSTR was allowed to re-equilibrate. Equilibrium aqueous HOC
concentrations as well as the altered HOC concentration in coal tar were measured
during the re-equilibration period The water flow through the system caused changes
in DNAPL composition and associated equilibrium aqueous phase HOC
concentrations as various HOC components were depleted from coal tar. Measured
HOC equilibria were compared with those predicted by Raoults law, with coal tar
assumed to behave as an ideal DNAPL over the entire range of HOC dissolution and
depletion.
This data provided information on the concentration boundary condition at the coal
tar-water interface.
2.3.3 Packed Bed DNAPL Pool Dissolution Tests
Pool dissolution tests were conducted next in which distilled water was pumped
through a glass cell packed with non-sorbing glass beads and embedded with a
DNAPL pool to simulate a DNAPL-contaminated aquifer. The use of non-sorbing
porous media enabled the test to focus solely on DNAPL pool dissolution processes.
Figure 2.4 provides a schematic of the glass cell fabricated by Allen Scientific Glass
of Boulder, Colorado. The DNAPL pool was fused at the bottom in the center of the
reactor so the inlet and outlet distance would be symmetric. Other features included a
50


glass frit at the entrance to evenly distribute flow, sample ports at the top and bottom
of the cell. The reactor end caps and sampling ports were sealed with septum caps to
avoid leakage and volatilization losses. The inlet and outlet tubing connected to the
reactor consisted of stainless steel tubing to minimize sorption of tar-derived HOCs
present in water. The glass column was packed with sterile, soda lime glass beads
(Jaygo Incorporated, Union, NJ) and pre-baked sand (1.5 mm diameter). The glass
column and glass beads were autoclaved prior to packing to provide a sterile
environment. The cells, added appropriate volumes of fresh and aged coal tar to the
DNAJPL pool, were packed with porous media. After removing any air trapped within
the cells, a dye test was conducted to ensure even flow distribution within the cell. A
tracer (Bromide) test was conducted to determine the longitudinal dispersivity of the
system.
2.3.4 Sample Collection and Analysis
Aqueous samples from the equilibrium test were withdrawn via a glass syringe and
stored in 2 ml glass vials. Effluent samples from the packed bed DNAPL pool
dissolution tests were collected at the outfall of the 1/8-inch metal tubing with
overflow into 2-ml glass auto sampler vials without headspace. All samples were
stored in a refrigerator until analysis. Samples were collected at various time periods
during the experiment. Replicate samples were collected at each sampling event to
51


verify the accuracy of the analysis and in the event one of the samples returned
unreliable results.
The aqueous phase samples collected from the packed bed DNAPL cell were
analyzed by a high-performance liquid chromatography (HPLC, HP-1100 Series)
with a 25-cm-long chromatograph column (LC-PAH, 5 gm, Supelcosil). The samples
were analyzed within two weeks of collection per EPA recommended hold times. The
PAH standards ran on the HPLC along with each sample run
52


Figure 2.4: Packed Bed DNAPL Cell Schematic
0.79 cm
0.79 cm
! 1.5 cm 1.5 cm 1.5 cm
I <
---------------- ---------------
Glass Frit Sampling Ports
PLAN VEIW
53


2.4 Results
2.4.1 Equilibrium Test
The composition of fresh and depleted coal tar was analyzed by Kevin Kelly (the U S.
Bureau of Reclamations, Golden, CO). Sixteen PAHs were measured by GS/MS
since the calibration standard consisted of 16 PAH standards. However, more than
twenty identifiable aromatic compounds were analyzed by Barringer Laboratories
(Golden, CO) in the same coal tar. The chemical composition of fresh and aged coal
tar is given in Table 2.1.
Table 2.1: Chemical Composition of Fresh and Depleted Coal Tar
Fresh Coal Tar Depleted Coal Tar
ComDound Name %(w/w) %(w/w)
Naphthalene 8.7651 4.1144
Acenaphthylene 1.29 0.7210
Acenaphthene 0.7635 0.5344
Fluorene 0.8067 0.7827
Phenanthrene 2.981 2.2802
Anthracene 0.4992 0.4577
Fluoranthene 0.5968 0.0739
Pyrene 0.6754 0.7048
Benzo(a) anthracene 0.2529 0.3175
Chrysene 0.2132 0.2569
Benzo(b) fluoranthene 0.1115 0.1290
Benzo(k) fluoranthene 0.0997 0.0813
Benzo(a) pyrene 0.1298 0.1472
Indenol (1,2,3-cd) pyrene 0.0667 0.0696
Dibenzo (a,h) anthracene 0.0198 0.0184
Benzo(g,h,l) perylene 0.0202 0.0189
2-Methyl Naphthalene 3.0 2.826
Styrene 0.4 0.02
54


The aque.ous phase solution at equilibrium with fresh and depleted coal tar contained
a mixture of Hydrophobic Organic Chemicals (HOCs) released from the coal tar
globule. The solution was analyzed for six (6) different HOCs released from the coal
tar globule. The measured equilibrium concentrations from the CSTR were compared
to the predicted concentrations from Raoults Law. The measured and predicted
equilibrium HOC concentrations of fresh and depleted coal tar are given in Figure
2.5a-2.5b.
55


Figure 2.5: Equilibrium Test Comparison of Predicted Verses Measured HOC
Concentration
Figure 2.5a Fresh Coal Tar
56


Figure 2.5b Depleted Coal Tar
The measured equilibrium aqueous phase concentrations compare well with the
predicted concentrations from Raoults Law for the different HOCs, except for
styrene. Volatilization losses from the aqueous phase solution in the CSTR may
have lead to anomalous styrene results. The results spanned 3 orders of magnitude
for the wide range of chemicals found in coal tar, and indicate that Raoults Law with
the ideal NAPL assumption is an adequate means for estimating equilibrium aqueous-
57


phase concentrations of HOCs released from both fresh and depleted DNAPL coal
tar.
2.4.2 Packed Bed DNAPL Pool Dissolution Test with
Glass Bead Particles and Fresh Coal Tar
(Experiments of Korey Kadrmas)
After dye and tracer tests indicated evenly distributed water flow through the
packed bed DNAPL cell, the experimental runs shown in Table 2.2 were performed
At first, water flow rates were varied over the range shown in Runs A1-A3, and, the
concentration of naphthalene, styrene, and 2-methyl naphthalene in the water exiting
the cell were measured by the HPLC (the concentrations of the rest of HOCs were too
low to measure by HPLC).
The three parameters that affected the rate of dissolution of chemicals from the
DNAPL pool were varied during the tests The three parameters varied were pore
water velocity or average linear groundwater velocity (ux), pool length (Lp), and
particle size (dp). Each parameter was varied independently of the others in separate
experimental runs as shown in Table 2.2, to judge the sensitivity of that parameter on
dissolution rates.
58


Table 2.2: Experimental Parameters Varied for the Packed Bed DNAPL Pool
Dissolution Test
1 1 Exd. Parameters
ux Lp dp
Varying Parameters Run (m/yr) (in.) (mm.)
A. Increasing groundwater velocities A1. 124 1.0 0.5
A2. 432 1.0 0.5
A3. 1843 1.0 0.5
B. Increasing pool lengths B1. 124 1.0 0.5
B2. 124 3.0 0.5
C. Increasing particle sizes C1. 432 1.0 0.5
C2. 515 1.0 2.0
Groundwater velocities were varied from approximately 100 m/yr to 2000 m/yr,
particle sizes of 0.2 mm and 2 mm were considered, and 4-ml and 12-ml volumes of
coal tar were added to the DNAPL reservoirs to create one-inch long and three-inch
long embedded DNAPL pools, respectively, in the packed bed cell. One of the porous
packed cells with one inch-long DNAPL pool is given in Figure 2.6.
59


Figure 2.6: Packed Bed DNAPL Ceil Used in the Packed Bed DNAPL Pool
Dissolution Test
The equilibrium concentration provided the initial concentration boundary condition
at the tar-water interface, which was then employed in Equation 2.5 to estimate the
average steady-state HOC concentration in the water existing in the pool zone.
Predicted average effluent steady state aqueous chemical concentrations were
compared with experimental data obtained from the packed bed DNAPL tests to
evaluate the impact of varying groundwater velocities, particle sizes, and pool
lengths. The experimental data shown in Figure 2.7a, 2.7b, and 4.7c indicate that
average steady-state HOC concentrations in the effluent existing in the pool zone
60


increase as the groundwater velocity decreases (Figure 2.7a), the pool length
increases (Figure 2.7b), and the particle size decreases (Figure 2.7c). The decrease in
groundwater velocity and the increase in pool length increase the contact time
between water and the pool, thereby increasing the average aqueous effluent PAH
concentration. The increase in particle size at the pool water interface is likely to
increase the vertical dispersion coefficient, thereby increasing chemical flux from the
pool into water and increasing aqueous phase HOC concentration existing in the pool
zone.
Figure 2.7a: Comparison of Naphthalene Relative Concentrations at Different
Water Velocities (The Tests were Conducted with Fresh Coal Tar)
0 070
0 060
0.050
o
w
wt
~ 0 040
U
c
o
o
v 0.030
>
5
a 0.020
0.010
o ooo
*
i U, 124 m/yr

& SI % gg U, = 432 m/yr
! 4 A A A UM = 1843 m/yr
12 24 36 48 60 72
Time (hours)
84
Low
Medium
High
96
108 120
61


Relative Cone (C.JC,
Figure 2.7b: Comparison of Naphthalene Relative Concentrations for Varied
Pool Lengths (Tests were Conducted with Fresh Coal Tar)
TIME, min.
62


Figure 2.7c: Comparison of Naphthalene Relative Concentrations for Varied
Particle Sizes
The observed steady state aqueous concentrations were compared against Hunt et
al.( 1988) predicted concentrations for a semi-infinite aquifer with input parameters
that were the same as measured for the packed bed DNAPL pool dissolution tests
(See Table 2.3). Typically, the Hunt predicted values compared well against the
observed measurement, except for Experiment A3 (high flow rate). Although, the
results were still within an order of magnitude the reason for the departure may be
that dispersion tends to dominate over diffusion at higher groundwater flow velocities
and the average percent solubility becomes independent of velocity.
63


Also, Hunt et als statement that molecular diffusion dominated at very low flow
velocities and the average concentrations were observed to decrease with increasing
flow velocities due to a decrease in contract time between the DNAPL pool and
passing flow were observed in the results. The results also show, as expected, that at
typical groundwater flow velocities, the measured HOC concentrations were found to
be far less than the expected solubility limits of these chemicals.
Table 2.3: Naphthalene Results for Varying Cell Parameters
1 1 Exn. Hunt Predicted css Observed C,s Hunt Predicted CSs/Cq Observed Css/Ceq Percent Difference
Varying Parameters Run (mg/L) (mg/L) (%)
A. Increasing groundwater velocities A1 0.69 0.76 0.052 0.057 8.77
A2 0.61 0.50 0.045 0.038 -18.42
A3 0.57 0.30 0.044 0.022 -100
B. Increasing pool lengths B1 0.69 0.76 0.052 0.057 8.77
B2 1.19 1.20 0.088 0.089 1.12
C. Increasing particle sizes C1 0.61 0.50 0.045 0.038 -18.42
C2 0.61 0.82 0.045 0.061 26.23
The prediction model described above used Hunt et al. Equations 2.5 and 2.6 with
known and controlled input parameters from all the experiments. The parameters used
included groundwater flow rates, DNAPL pool lengths, and transverse dispersivity
(a t), assumed as 1/10 particle size used in the columns. The measured a l was
within a factor of two of the media particle sizes. These results lead to the conclusion
that the particle size just at the top of the pool controls in the model (a t) and that the
64


gross dispersivity (a l) within the cell appears to not be affected much by particle
size change. So the model sensitivity is a function of a j and not a l; a t could not
be measured, but was inferred from particle size.
The results of Table 2.3 demonstrate that the assumptions used to yield Equation 2.5
(the two-dimensional form of Equation 2.4) are valid and that advective transport in
the direction of groundwater flow (x) is much greater than the dispersive transport in
the direction of flow and that longitudinal dispersion can be neglected. This was
verified by calculating the Peclet number for each experimental run. In all
experiments the Peclet number was typically much higher than 1 indicating that there
is no relative longitudinal dispersion in the cell compared to the advective flux of the
flow field adjacent to the DNAPL pool. The assumption used to simply the two-
dimensional steady state equations are valid based on the Peclet number results.
The results in Table 2.3 illustrate the sensitivity in the experiments for all the varied
parameters, but the result trends could be predicted based on an understanding of the
DNAPL-water pool dissolution relationship. The naphthalene results were found to
be valid for assessing application of Raoults Law with the ideal NAPL assumption as
an adequate means for estimating initial coal tar-water equilibrium concentrations and
using Hunts average concentration equation (Equation 2.6) for estimating aqueous
65


steady state concentrations from a dissolved multi-component DNAPL pool being
flushed.
The styrene results for the packed bed DNAPL pool dissolution tests for varying
experimental parameters are presented in Table 2.4. The styrene results were deemed
to be acceptable for all three series of experiments and the relationship between
chemical concentrations expected for each experiment were similar to the
naphthalene results. A slight deviation in the observed styrene concentrations was
observed for experiments A2 and Cl. The discrepancy may be a result of the
uncertainty in the predicted equilibrium concentrations for styrene during the
Equilibrium test. The initial coal tar-water equilibrium concentration may have
resulted in volatilization losses from the aqueous phase solution.
Table 2.4: Styrene Results for Varying Reactor Parameters
Varying Parameters Exp. Run Hunt Predicted Css Observed Css Hunt Predicted CSs/Ceq Observed Css/Ceq Percent Difference
(mg/L) (mg/L) (%)
A. Increasing groundwater velocities A1 0.126 0.193 0.051 0.078 34.62
A2 0.112 0.196 0.045 0.079 75.56
A3 0.108 0.162 0.044 0.065 32.31
B. Increasing pool lengths B1 0.126 0.193 0.051 0.078 34.62
B2 0.216 0.252 0.088 0.102 13.73
C. Increasing particle sizes C1 0.112 0.196 0.045 0.079 75.56
C2 0.112 0.141 0.045 0.057 21.05
66


The 2-methyl naphthalene results for the packed bed DNAPL pool dissolution tests
for varying experimental parameters are presented in Table 2.5. The 2-methyl
naphthalene results were deemed to be acceptable for all three series of experiments
and once again similar trends in chemical concentration for each experiment were
similar to the naphthalene results.
Table 2.5: 2-Methyl Naphthalene Results for Varying Reactor Parameters
Varying Parameters Exp. Run Hunt Predicted Css Observed Css Hunt Predicted Css/Ceq Observed Css/Ceq Percent Difference
(mg/L) (mg/L) (%)
A. Increasing groundwater velocities A1 0.065 0.085 0.051 0.066 22.73
A2 0.058 0.059 0.045 0.046 2.17
A3 0.056 0.038 0.044 0.030 -46.67
B. Increasing pool lengths B1 0.065 0.085 0.051 0.066 22.73
B2 0.113 0.090 0.088 0.070 -25.71
C. Increasing particle sizes C1 0.058 0.059 0.045 0.046 2.17
C2 0.058 0.085 0.045 0.066 31.82
2.4.3 Packed Bed DNAPL Pool Dissolution Tests with
Fresh and Depleted Coal Tar
Glass cells containing embedded fresh and depleted coal tar DNAPL pools (of length
1 in, 2.5 cm), were packed with sand grains of median size 1.5 mm diameter. The
cells were operated at water flow rate of 1.0 ml/minute (equivalent to Ux = 484 m/yr)
for 65 days, with periodic flushing with biocides to inhibit microbial growth in the
system.
67


The observed steady state aqueous concentrations of the abiotic cells were compared
against predicted concentrations for the cells, employing equilibrium concentrations
measured for fresh and depleted coal tar in the CSTR tests described previously
(Figure 2.8a, 2.8b, 2.8c, and 2.8d). The observed and predicted steady state aqueous
phase HOC concentrations exiting the depleted and fresh coal tar cells are shown in
Table 2.6, and in Figure 2.8a, 2.8b, 2.8c, and 2.8d. Note, the steady state
concentrations exiting the depleted coal tar cells are much lower than those exiting
the fresh coal tar cells, consistent with the impact of HOC depletion on the
equilibrium boundary condition. The Hunt predicted values compared well against
the observed measurement for fresh and depleted coal tar cells. The results are
adequate enough to validate applying Raoults Law with the ideal NAPL assumption
as an adequate means for estimating coal tar-water equilibrium concentrations and
using Hunts average concentration equation for estimating aqueous steady state
concentrations downstream from a multi-component DNAPL pool being flushed.
(a t) was assumed as 1/10 particle size.
68


Table 2.6: Aqueous Phase HOCs for Fresh and Depleted Coal Tar Cells
Exp. Run Hunt Predicted c Observed Css Observed Css/Ceq Percent Difference
Varying Parameters (mg/L) (mg/L) (%)
Fresh Coal Tar Average Naph. 0.828 0.910 0.063 9.01
Depleted Coal Tar Naph. 0.42 0.557 0.075 24 60
Fresh Coal Tar Average 2Meth. Naph 0.099 0.107 0.062 7.48
Depleted Coal Tar 2Meth. Naph 0.075 0.079 0.060 5.06
Fresh Coal Tar Average Styrene 0.142 0.219 0.089 35.16
Depleted Coal Tar Styrene 0.0047 0 - -
69


Figure 2.8: Aqueous Phase HOCs for Fresh and Depleted Coal Tar Cells
Figure 2.8a Naphthalene
- -o
Contro)-Fresh (Avg)
Di 1.20
y i .oo u
O 0.80
C
| 0.60 4^
| I-
^ 0.40 -jju
z
-j -A- Control-Depleted
i
F

- -o .
O < > HI .
HI ()
-ft
-ft - A
ft A "ft -ft
q. U.4U -ff-
£
ra
S
I 0.00
<
10 15 20 25 30 35 40 45 50 55 60 65 70
Time (day)
Figure 2.8b 2-Methylnaphthalene
70


Figure 2.8c Styrene
Figure 2.8d Acenaphthene
71


2.5 Conclusions
The results of the packed bed DNAPL pool dissolution tests proved that aqueous
steady-state chemical concentrations could be predicted from a multi-component
DNAPL pool using boundary conditions assumed at the NAPL-water interface. The
results verified the modeling assumptions required to simplify the advective-
dispersive equation down to a one-dimensional steady-state equation that could be
used to predict chemical transport within the packed bed DNAPL cell. The chemical
dissolution results demonstrated how the interaction between the DNAPL and
adjacent water were affected by varying groundwater velocity, pool length, and coal
tar type. The affect of altering these parameters was seen to have an affect on the
contact time between the passing groundwater and the DNAPL pool, which affected
the chemical dissolution of coal tar.
Changes in particle size affected the observed steady-state effluent concentrations by
impacting transverse dispersivity of the systems. The observed results showed that an
increase in particle size tends to increase dispersion around the DNAPL pool, which
increases the aqueous chemical concentrations adjacent to the pool. A two-
dimensional cell would be able to measure transverse dispersivity changes directly,
which could not be done in a 1-d system. The sensitivity to particle size changes
could then be predicted by Hunts pool dissolution equations.
72


Overall, the experimental results from the laboratory-confined aquifer compared well
against Hunts equation for determining the average concentration leaving a semi-
infinite, contaminated aquifer. The Hunt et al assumption that the average
concentrations were observed to decrease with increasing flow velocities due to a
decrease in contract time between the pool and passing flow were proved. In
addition, assumptions such as dispersion tending to dominate over diffusion at higher
groundwater flow velocities resulting in the average percent solubility becoming
independent of velocity were helpful in evaluating the results. In evaluation of
Hunts pool dissolution Equations 2.5 and 2.6 sensitivity was observed in varying the
three experimental parameters (Dz Lp, and ux) used to determine the steady-state
aqueous chemical concentrations leaving the pool.
Fresh and depleted coal tar could be modeled in a manner consisten with Raoults
law. The Equilibrium test results proved that Raoults Law with the ideal NAPL
assumption is an adequate means to estimate initial and dynamic changes in coal tar-
water equilibria. The concentration boundary condition at the NAPL-water interface
is assumed to be defined by equilibrium conditions, which were was validated by the
Equilibrium test results. Packed bed porous media tests with fresh and depleted coal
tar also showed that the analytical model was effective in describing pool dissolution
over the long-term, with depletion of the DNAPL.
73


3. Phase 3 Experimental Evaluation of
Biostabilization of DNAPL Coal Tar Pools in
1-Dimensional Porous Media Systems
3.1 Introduction
Polycyclic aromatic hydrocarbons (PAHs) may be released to groundwater from the
hazardous waste sites and spills and leaks during the storage and transportation of
dense non-aqueous-phase liquids (DNAPLs) such as coal tar. Coal tar, produced at
former manufactured gas plant (MGP) sites (Luthy et al, 1994, Villaume 1984), is a
multi-component DNAPL composed of a mixture of several hundred monoaromatic
compounds, chiefly PAHs such as styrene, naphthalene, 2-methylnapthalene,
phenanthrene, and pyrene. DNAPLs have a tendency to sink to the bottom of surface
and ground waters due to its specific gravity. The DNAPL can migrate through the
vadose and saturated zones until contaminants reach the bedrock. Some portion of the
DNAPL, retained in the porous media in the form of NAPL blobs and ganglia, causes
discontinuities in fluid flow of the DNAPL body (Cohen and Mercer, 1993, Hemond
and Fechner, 1994). DNAPL pooling can occur on the bedrock and low permeable
lenses found in aquifer. DNAPL pools and ganglia are the sources of long-term
74


contamination of groundwater as a result of the slow release of low solubility organic
chemicals. The toxicity of these chemicals, and their migration and persistence in
groundwater creates a long-term risk to human health (Cohen and Mercer, 1993,
Ramaswami and Luthy, 1996).
Traditional remediation of DNAPL-contaminated sites is very expensive and often
involves many years of cleanup. Direct biological treatment of the DNAPL source
zone has been addressed in very few studies (Ghoshal et al., 1996; Johansen et al.,
1997; Ramaswami et al., 1994). Recent bench scale screening tests (Ramaswami et
al., 2001) have shown that in situ biodegradation in the area of the DNAPL source
may offer a cost-effective risk management strategy for coal tar contaminated sites
through biological stabilization processes. Biostabiiization is proposed as a means of
reducing risk at DNAPL-contaminated sites through slow dissolution the PAH
contaminants to ground water coupled with a more rapid microbial activity in the
ground water surrounding the DNAPL pool (Ramaswami et al., 1999). Over the long
term, depletion of more soluble and potentially more toxic DNAPL constituents can
stabilize the DNAPL spill zone and result in diminished plume formations and
reduced risk without complete destruction of the entire mass of the DNAPL. However
the simultaneous dissolution and biodegradation of multiple chemicals is a complex
phenomenon, rarely addressed by researchers because of the difficulty in measuring
and modeling the behavior of mixtures in real-world PAH.
75


The focus of this phase is to demonstrate DNAPL biostabilization by coupling
dissolution and biodegradation of polycyclic aromatic hydrocarbons (PAHs) in coal
tar in short-term one dimensional porous bed test in order to manage the risk posed by
groundwater contamination by DNAPL coal tar.
3.2 Factors that Affect DNAPL Biostabilization
Four main processes can affect DNAPL biostabilization:
1) Equilibrium PAH dissolution boundary condition at the DNAL-water
interface (discussed in phase 2)
2) Transport of PAH in the water in the porous matrix (discussed in phase 2)
3) Biodegradation of PAH in the water (discussed in phase 1)
4) Impact of biogrowth on porous media properties affecting PAH transport in
water.
The first three processes have been discussed in detail in Phases 1 and 2. In this
section, a review of literature is presented pertaining to how biogrowth may impact
chemical transport in porous media, through biofilm formation, bioclogging and
production of biosurfactants.
Bacteria is subsurface can be found as attached to soil surface and suspended in
groundwater (Bouwer and McCarty, 1984). Dissolved organic chemicals that enter
76


the subsurface can be transformed by bacteria attached to soil surface (Rittmann,
1993). Bacterial growth can affect the properties of porous media such as hydraulic
conductivity and dispersivity (Bielefeldt et. al, 2002, Taylor and Jaffe, 1990a)
Techniques to evaluate all the above processes were included in the design of the
bench-scale biostabilization tests, as described below.
3.3 Materials and Method
3.3.1 Bacterial Solution
A mixed consortiums of aerobic naphthalene and phenanthrene degrading bacteria
was maintained in a Bio-CSTR continuously fed with 100 ml of High Biological
Oxygen Demand (HBOD) nutrient solution (Ghoshal et al., 1996) per day to which
was added naphthalene, 2-methlynaphthalene, styrene, and phenanthrene as carbon
source (Morrison, 1999). Approximately four drops of coal tar was also added to the
Bio-CSTR every month in order to acclimate the microbes to the range of PAHs in
coal tar. Pure oxygen was replenished in the system every two days. The bacterial
consortium in the Bio-CSTR was identified as predominantly of the Pseudomonas
species, including Stenotrophomonas maltophilia, Pseudomonas fluorescens,
Pseudomonas putida, and Pseudomonas stutzeri (Morrison, 1999/ The bacterial
suspension were taken out from the Bio-CSTR and mixed with the sand used in the
bioactive cells.
77


3.3.2 Sand
Eight Kg of 1.5 mm grain size sand were baked at 400 C for 2 hours. 2.5 Kg of pre-
burned sand was placed into 5L glass beaker and 1.5L of bacterial suspension from
bio-CSTR w'ere added. The sand were mixed with bacterial solution and placed into
an incubator for 1 day. At the end of the time period, soil samples were collected
from different depths and locations in the beakers and assayed to check for uniform
distribution of microbes.
Total number of bacteria in the soil samples was measured by Acridine Orange count
using Fluorescent microscopy. After verifying that the sand has well-mixed bacterial
population, the bioactive cells were packed with the seeded grains whereas the
control cells were packed with pre-baked grains, after the coal tar pool is carefully put
in place in the DNAPL well.
78


3.3.3 Coal Tar
Fresh and depleted coal tar were placed into DNAPL pools at the bottom of the cells.
The fresh coal tar was obtained from the subsurface (well #AZ-W07-S) from a
contaminated site in Baltimore, Maryland (MacFarlane, 1990). The coal tar was stored
in airtight amber bottles in a refrigerator.
Isleyen (2002) gives detailed information on developed procedure of the depleted
coal. In summary, fresh coal tar globule was placed at the bottom of the coal tar
CSTR having an elevated teflon-coated magnetic stirrer. The water in the coal tar
CSTR was continuously mixed with the elevated stirrer at gentle rotational speeds
such that the fresh coal tar globule at the bottom of the jar remained undisturbed. The
depleted coal tar was produced by flushing water around the fresh coal tar globule
until approximately half the naphthalene in the DNAPL was depleted, thus represents
an aged DNAPL coal tar matrix. The composition of the fresh and aged coal tars was
measured by Gas Chromatography-Mass Spectrometry (GC-MS).
3.3.4 Glass Cells with Embedded DNAPL Reservoir
Short-term dissolution and biostabilization experiments were conducted for 9-10
weeks in porous packed glass cells containing an embedded coal tar DNAPL pool at
the base. The glass cells with DNAPL pool lengths of 1.0 inch long were fabricated
(Allen Scientific Glass, Boulder-CO) to demonstrate DNAPL biostabilization by
79


coupling dissolution and biodegradation of PAHs in coal tar in short-term one
dimensional porous bed test. The DNAPL pool was fused at bottom in the center of
the cells. Each cell has three sampling ports at the top and an inset reservoir pool with
one sampling port at the bottom. Evenly disperse flow at the entrances of cells were
achieved by a built-in glass frit. The cell end caps and sampling ports were covered
by teflon tape and tightly sealed with septum caps avoid leakage and volatilization
losses. Stainless steel outlet tubing was connected to cells whereas a constant flow
rate was pumped into cells by inlet tubing made of teflon. Figure 3.1 and Figure 2.4
show the components of the cells.
Two sets of the experiments were conducted, one set with fresh coal tar (4 cells) and
the other one with aged coal tar (3 cells), to assess the impact of PAH depletion on
coal tar biostabilization. Seven cells were packed with the sand grains of median size
1.5 mm diameter, 4 of which were bioactive cells. Three of them were control cells.
80


Figure 3.1: Porous Packed Cells
Four bioactive cells, two cells with fresh and two cells with aged coal tar pools, were
packed with pre-baked sand mixed with microbes. Three control cells, two cells with
fresh coal tar and one with aged coal tar, were packed with pre-baked sterile sand and
maintained corresponding to four bioactive cells. DNAPL wells of the cells were
81


filled with 2 ml coal tar mixed with 5 mm diameter of glass beads to stabilize the
pool. The air volume in the cells was replaced with DI water before applying any
water flow through the cells. The control cells were periodically flushed with 10% of
Sodium Azide to kill any microbial growth. PAH dissolution from the DNAPL pool
was evaluated in the control cells and compared over 9-10 weeks with the bioactive
cells to quantify any biological stabilization occurring in the system. The cells were
operated at the same water flow rate of 1.0 ml of pre-aerated DI water per minute.
In summary, the seven cells used for this phase were:
Two bioactive cells with fresh coal tar: Cell A and B
Two bioactive cells with depleted coal tar (50 % of naphthalene was
depleted): Cell C and D
Two abiotic control cells with fresh coal tar: Cell E and F
One abiotic control cell with depleted coal tar: Cell G
3.3.5 Coal Tar Adding
After adding appropriate volumes of coal tar to the DNAPL pools, glass beads with
diameters of 5 mm were added into the DNAPL pools in order to raise the DNAPL
level to provide a good DNAPL-water interface. The cells were then packed with pre-
baked sand (1.5 mm diameter), and distilled water was added to the cells displacing
trapped air within the cells. All the trapped air bubbles were removed and the cells
82