MICROBIAL DEGRADATION OF CREOSOTE-DERIVED
COMPOUNDS IN NATURAL AND LABORATORY MICROCOSMS
By
E. Michael Godsy
Research Microbiologist
U.S. Geological Survey
345 Middlefield Road
Menlo Park, California 94025
EPA - IAG Identification Number - DW-14934092-0
Final Report
Covering the Period from August 31,1989 to September 1,1992
September 1,1992
-------
1. INTRODUCTION
AMERICAN CREOSOTE WORKS, INC.
In 1983, American Creosote Works, Inc. (ACW) in Pensacola, Florida was selected by
the U.S. Geological Survey (USGS) as one of three national research demonstration areas to
develop an understanding of hazardous waste processes. The criteria governing the selection
of this site included (1) the relatively simple mineralogy of the surficial sand-and-gravel
aquifer, (2) the apparently straightforward flow system within the aquifer, and (3) the
availability of a preliminary data base on the flow system, extent of contamination, and the
water chemistry.
The USGS research at the Pensacola site is a multidisciplinary effort designed to study
the processes that affect the occurrence, transport, and fate of toxic contaminants associated
with wood preservatives in the environment. The research being conducted at the site is
designed to: (1) characterize the geology and hydrology; (2) determine the transport behavior
of the organic compounds in the ground water, (3) describe and quantify the physical,
chemical, and biological processes affecting these compounds; (4) develop a solute-transport
model capable of simulating the various processes that affect the movement of selected
contaminants in ground waters; and (5) determine the impact of selected contaminants on
aquatic organisms in Pensacola Bay.
SITE DESCRIPTION AND HYDROLOGY
The research site is located in Escambia County within the city of Pensacola, Florida, at
and adjacent to the site of an abandoned wood treatment plant (Figure 1-1). The 7.3 hectare
plant site is situated approximately 550 m north of Pensacola Bay and near the entrance to
Bayou Chico.
Pine poles were treated with wood preservatives for nearly 80 years prior to the closing
of the site in December 1981 (Mattraw and Franks, 1986). Prior to 1950, creosote was used
exclusively to treat poles and subsequent to that date, both creosote and pentachlorophenol
(PCP) were used. The plant used approximately 95 m3 of creosote per month and a similar
quantity of PCP prior to closing. It is estimated that approximately 53 m3 of "blowdown" or
residual waste water from the pressure chambers used to treat the poles was discharged to
surface impoundments each month.
-------
The wood treatment process consisted of removing as much of the cellular moisture as
possible from the poles and replacing it with creosote and/or PCP. Prior to 1945, the waste
waters, consisting of extracted moisture from the poles, PCP, creosote, and diesel fuel, from
these processes were discharged to an open pit. Subsequently, the wastes were recirculated
to an unlined impoundment. The recirculation impoundment occupied 7,700 m2 and held
9,400 m3 of waste water. The southern impoundment, constructed in 1954, was used as an
overflow pond for additional storage. It occupied 3,240 m2 and held 4,000 m3 of waste
water. Both impoundments were constructed of clay embankments approximately 1 m high
on all sides. The average depth of both impoundments was 1.2 m.
Prior to 1970 whenever water levels in the impoundments became high, waste water
spilled over and flowed south through a natural depression discharging directly into the
entrance of Bayou Chico and along an abandoned railroad track bed that transects the
research site. Subsequent to enactment of the Clean Water Act in 1970, in an effort to control
water levels in the impoundments, waste waters were periodically drawn off and allowed to
evaporate in designated areas north and south-east of the overflow impoundments. The
depression to the south has since been channelized into a storm water drainage ditch to
convey runoff from the basin north of the plant site.
In October 1983, under the EPA emergency response program of Superfund, waste
waters were drained from the two impoundments and the sludge in the pond was dewatered
using a mixture of lime and fly ash. A clay cap was placed over the entire area once occupied
by the impoundments and the evaporation areas.
Before remedial action, the surface impoundments were unlined and indirect hydraulic
contact with the sand-and-gravel aquifer. The aquifer, the principal source of drinking water
supply in the area, consists of deltaic, fine-to-coarse quartz sand deposits, interrupted by
discontinuous silts and clays, which locally confine the aquifer. Near the site, the aquifer is
about 90 m in total thickness. Only the upper 30 m are affected by wood preserving wastes.
Ground water flow is generally to the south towards Pensacola Bay (Figure 1-2). Darcy
flow velocities are on the order of 0.3 to 1.2 m/day, with lower velocities in the deeper parts
of the aquifer. Pumpage in the study area has fluctuated considerably throughout the last 80
years, largely as a result of varying industrial uses. Withdrawal rates at a nearby chemical
plant have ranged from over 30,000 m3 day in the 1930's to the present day minimum of 1,700
m3/day. Jacob and Cooper (1940) conclude that heavy industrial water use was responsible
for inducing downward gradients in the area. Present vertical gradients, on the contrary, are
-------
generally upward, thus reducing the probability of significant downward transport of the
soluble contaminants.
CONCEPTS OF WASTE TRANSPORT
Effluent from the treatment process consisted of water, cellular debris, diesel fuel,
creosote, and PCP (Goday et al, 1992). The chemically complex, organic-rich mixture was
discharged to the shallow unlined waste disposal ponds and from there, large but unknown
quantities of the waste infiltrated the soil beneath the ponds. The waste mixed with the
ground water, and two distinct phases resulted: a denser than water hydrocarbon phase that
moved vertically downward somewhat perpendicular to the ground water flow, and an
organic rich aqueous phase or water soluble fraction (WSF). The WSF is enriched in organic
acids, phenolic compounds, single and double ring aromatic compounds, and single and
double ring heterocyclic nitrogen, sulfur, and oxygen containing compounds (NSO). As the
contaminated water moves along with the groundwater flow, the dissolved contaminants are
subject to physical, chemical, and biological processes that tend to retard the movement of
these compounds relative to the ground water.
As a result of the periodic overtopping previously mentioned, large but unknown
quantities of creosote were transported to the natural depression south of the ponds and along
the abandoned railroad bed. Preliminary investigations reveal that creosote hydrocarbon fluid
accumulated in the capillary fringe to a height of 0.3 m above the water table with an areal
extent of approximately 2 hectares. This body of creosote hydrocarbons provides a second
source of contamination, separate from the main body of creosote which is located beneath
the storage ponds.
EPA-SPONSORED RESEARCH
The USGS proposes that the EPA help support studies at the Pensacola site, working
towards a more complete understanding of the physical, chemical and microbiological
processes affecting solutes in ground-water systems. The EPA support will be used to (1)
continue and extend the collection and interpretation of the spatial and temporal variations in
the organic geochemistry of the site, (2) expand the laboratory investigations to include the
study of the biodegradation of the WSF compounds and mixtures of WSF compounds under
denitrifying, sulfate reducing, and acetogenic conditions and isolation of single organisms
and/or consortia capable of biodegradation of the WSF compounds, and (3) investigate
-------
strategies and techniques for the remediation of both WSF contaminants and those
contaminants associated with the free hydrocarbon phase.
-------
0.5-
co
I SITE NUMBER
GROUND WATER
LEVEL CONTOUR,
m ABOVE SEA
LEVEL
0 100 m
I i
Figure 1-1. Shallow ground water-level contours and well sites at the
abandoned creosote works.
-------
20-
30
•.•.•.•:•;• •: :•/:,/ ..-. : ,
Appmxr
of Co
unc 'e
Areal Ext,
ilaminalinn
nl
\
-10
-20
0 100m
I I
30
SAND
SANDY CLAY
CLAY
Figure 1-2. Geologic section showing placement of well sites, approximate
area! extent of contamination, and study zone.
-------
2. MATERIALS AND METHODS
SAMPLING SITES
The approximate down gradient distance of the contaminated sampling sites from the
southern most border of the source ponds are as follows (Figure 1-1): Site 3, 6 m; Site 39,53
m; Site 40,99 m; Site 4,122 m; and Site 37,150 m.
WELL DRILLING TECHNIQUE
Wells were drilled using the hollow-stem, continuous-flight auger method described by
Schalf et al. (1981). This method allows boring into soils carrying the cuttings upward along
the auger flights without the use of contaminating drilling fluids. To help prevent sand and
water intrusion into the hollow core while drilling, a knock-out plug was installed in the auger
bit and O-rings were installed between the auger flights.
CORE SAMPLING TECHNIQUE
When the desired depth was reached, a 5 x 60 cm split-spoon core sampling device
(Schalf et al., 1981) was lowered down the hollow stem of the auger flights. The sampler was
then driven through the knock-out plug at the bottom of the bore-hole into the undisturbed
aquifer material below. After the sampler was removed from the borehole, the sampler was
split lengthwise exposing the core material. The first 5 cm of the core was removed with a
sterile spatula exposing an uncontaminated surface. The center of the core was subsampled
for microbial analysis by pushing a sterile 2 x 8 cm brass tube into the core, filling it with the
aquifer material. The aquifer material was then extruded with a sterile 10 cm3 disposable
syringe plunger into a 25 x 142 mm anaerobic isolation roll streak tube (Bellco Glass Inc.,
Vineland, N.J.) filled with 20 mL of prereduced, anaerobically sterilized (PRAS) mineral salts
solution (Holdeman and Moore, 1972). Oxygen-free Ar gas was allowed to flow over the
surface of the mineral salts solution while extruding approximately 20 g of aquifer material
into the sample tube. The tube was then sealed with a butyl rubber stopper with a recessed
top (Thomas Scientific, Philadelphia, Pa.). The recessed top allowed for removal of the liquid
using syringe techniques. The mineral salts solution was composed of the following (per
liter): KH2PO4,0.75 g; K2HPO4,0.89 g; MgCl2-6 H20,0.36 g; NH4C1,0.9 g; trace metal
solution (Zeikus, 1977), 9.0 mL; vitamin solution (Wolin et al., 1963), 5.0 mL. The pH was
adjusted to 7.0, and the medium was then boiled, cooled and dispensed under a stream of
-------
O2-free Ar gas. The medium was then sterilized at 121 *C (1.05 kg/cm2) for 15 min. Just
prior to use, 0.2 mL of a 5 percent solution of Na2S was added as a reducing agent.
After microbial sampling, aquifer material was taken for the determination of porosity and
bulk density as outlined in Methods of Soil Analysis (American Society of Agronomy, 1965).
CORE MICROBIAL PROCESSING
Tubes containing the aquifer material were placed horizontally on a reciprocal shaker for
30 min at 250 reciprocations/min. Samples were removed for acridine orange direct counts
(AODC) using the method described by Wilson el al. (1983). Bacteria capable of rapid
heterotrophic aerobic growth in nutrient liquid media, which were present in the aquifer
material, were determined by a five-tube most probable number (MPN) series using Standard
Methods Broth (BBL Microbiology Systems, Cockeyville, Md.). Denitrifying bacteria were
determined by a MPN series using the medium described by Stanier el al. (1966). Sulfate
reducing bacteria (SRB) were enumerated using a 5-tube MPN procedure using the basal
medium described in the previous section with the addition of 3.0 g/L of Na2SO4,3.0 g/L Na
acetate-2 P^O, and a head space atmosphere of 70 percent H2 - 30 percent CC^. Bacteria
capable of anaerobic fermentative growth were enumerated using a five-tube MPN series with
peptone-yeast extract-glucose broth (PYG) (Holdeman and Moore, 1975). Numbers of
methanogenic bacteria were determined using a MPN method based on the presence of CFfy
in the head space above the medium after 21 days incubation as described by Godsy (1980).
The total biomass concentration on the sediment was determined by total protein on 10.0
g subsamples of the sediment. Samples were treated with 5.0 mL of 0.66 N NaOH allowed to
shake for 14 days in 50 mL DeLong flasks on an orbital shaker at 250 rpm. Total protein
present in the digest was determined using the Coomassie Brilliant Blue method (Appendix A)
as described by Galli (1987) calibrated using Bovine Serum Albumin (BSA).
WATER SAMPLING TECHNIQUES
Immediately after the split-spoon sampler was removed from the auger flights, a sterile
Teflon tube (7.5 mm i.d.) was lowered down to approximately 1 m above the bottom of the
auger flights. A peristaltic pump with sterile silicon rubber tubing was attached to the Teflon
tubing and allowed to pump for approximately 5 min (300 mL/min) before samples were
taken.
-------
Samples for AODC were collected from the pump discharge stream in sterile glass bottles
and determined by the method described by Robbie el al. (1977). Samples for viable bacterial
enumeration were collected as follows: a length of sterile silicon rubber tubing was attached to
the pump discharge line and coiled in a descending loop to eliminate bubbles in the line. One
end of a double-ended Vacutainer Needle (Becton-Dickinson, Rutherford, N.J.) was inserted
into the tubing while the other end of the needle was inserted through the septum of a sterile
evacuated 100 mL serum bottle and allowed to fill with the water sample. Prior to
evacuation, the bottle contained an C^-free Ar atmosphere.
Samples for bacterial enumeration were removed from the serum bottles using the
anaerobic techniques of Hungate (1969) and processed as described for the core samples.
Total biomass was determined on 5.0 mL aliquots using the method described by Ga'lli (1987).
Water samples for organic solutes were then collected by bailer or by means of a
peristaltic pump. A one-L glass bottle was placed between the pump and a length of 6 mm
i.d. sterile Teflon tubing placed down the well. Water was pumped from the well, directly into
the bottle without contacting the pump tubing. Samples were preserved by the addition of 65
mg/L of HgCl2 as a biocide, packed in ice, and shipped by overnight delivery service to the
laboratory for analysis.
GAS CHROMATOGRAPHY-MASS SPECTROMETRY-DATA SYSTEM
Organic solutes were identified using a Finnigan MAT gas chromatography-mass
spectrometer system with a computerized data system (GC/MS). The GC/MS consists of a
Model 4510 quadrupole mass spectrometer interfaced to a Model 9610 gas chromatograph,
both controlled by an Incos data system. Chromatography was performed by splitless
injection on fused silica capillary columns. The helium carrier gas flow rate was set at 2.0
mL/min. The oven temperature was programmed from 50°C at 10°C/min, to the maximum
temperature limit of the column selected. Two columns (30 m x 0.25 mm i.d.), obtained from
J & W Scientific Inc., both having bonded liquid phases 0.25 [im thick, were used. One was a
nonpolar DB-5 column and the other was a polar DB-WAX column with a bonded carbowax
phase. The capillary columns were separately and directly connected to the ion source of the
mass spectrometer which was operated in the electron impact mode.
GAS CHROMATOGRAPHY
A Varian 6000 gas chromatograph equipped with an injection port splitter, fused silica
capillary columns, and a flame ionization detector (FID) was used for the quantitative
10
-------
determinations of phenols, NSO, and PAH. The helium carrier gas flow rate was set at 2.0
mL/min, the hydrogen flow rate was set at 22.0 ml/min, and the air flow rate was 300 mL/min.
The injection port sweep flow rate was set at 4.5 mL/min and the detector make-up flow rate
was set at 13.5 mL/min. The column oven was temperature programmed from 75°C at
10°C/min to the maximum operating temperature for the selected column. The GC was
equipped with columns identical to those described above. A Varian Vista 401 data system
was used to record the chromatograms, peak retention times, and areas, and for computation
of the concentration of the individual components in the sample extract.
ORGANIC ANALYSIS OF GROUND WATER SAMPLES
Water samples were centrifuged at 2000 x g for 10 min to remove particulates and the
supernatant divided into two 5.00 mL subsamples. The bases, neutrals, and phenols were
removed and discarded from the first subsample (pH 6.8) by CH2C12 extraction (except where
specified, all extractions were 1:1 v/v) to facilitate the analysis for organic acids. The organic
acids remaining in the aqueous isolate were determined directly by the HPLC method of
Ehrlich el al. (1981) or acidified to pH 1.5 and the organic acids extracted into ethyl ether and
determined by GC/MS using the DB-WAX column.
Because of the complexity of the mixture of compounds occurring in creosote, a
preliminary separation into three classes of compounds (neutrals, phenols, and bases) was
necessary for the second subsample. The second subsample was made acidic, pH 1.5, and the
neutrals and phenols were extracted into CH2C12- The remaining aqueous layer was
neutralized (pH 6.8) and the bases were extracted into CH2Q2- The phenols were separated
from the neutrals by shaking the mixture with 1.0 mL 5 M KOH solution. The neutrals
remained in the CH2Q2 solvent. The KOH solution was acidified (pH 1.5) and the phenols
were extracted into C^C^- Each of the solutions were analyzed by GC/MS and/or GC/FID.
INORGANIC ANALYSIS OF GROUND WATER SAMPLES
Samples for inorganic constituents were analyzed by the U.S. Geological Survey, Quality
of Water Service Unit, Ocala, Fla., according to the methods described by Skougstad el al.
(1989). In addition, temperature, pH, alkalinity, dissolved O2, CH4, and H2S were measured
in the field immediately upon collection of the water samples (Baedecker and Cozzarelli,
1992).
11
-------
ANAEROBIC LABORATORY MICROCOSMS
Large methanogenic microcosms (Figure 2-1) were constructed in an anaerobic glove
box (Coy Laboratory Products, Ann Arbor, Mich.) containing a 10 H2 - 90% Ar atmosphere.
The microcosms were prepared in 4 L glass bottles, containing approximately 4 kg of
contaminated aquifer material collected from a depth of approximately 6.1 m at Site 39. All
large microcosms were fitted with both liquid and gas sampling ports and a U-tube
manometer. Microcosms were filled with 2.5 L of the contaminated ground water from Site 3
at depth of 6.1 m, and flushed with O2-free Ar. Control microcosm were autoclaved at
121*C (1.05 kg/cm2) for 45 min and used as killed cell controls. A positive control to
determine if organic material sorbed onto the aquifer material was biodegradable was prepared
for each inoculum and contained only the contaminated aquifer material and 2.5 L of PRAS
mineral salts solution. One mL/L of sterile amorphous FeS was aseptically added to all
microcosms as a reducing agent (Brock and O'Dea, 1977). The exact volume of each
microcosm was obtained from the weight of water-filled serum bottle minus the empty
weight. The actual head space volume above the liquid was determined from these weights.
At regular intervals, the microcosms were gently rotated until thorough mixing of liquid
and aquifer material was accomplished. After 30 min, the increase in gas volume was
recorded using a water wetted glass syringe, and corrected for temperature and pressure. The
gas composition was determined by a Baseline 1010A gas chromatograph (GC) fitted with a
thermal conductivity detector. The GC was equipped with both a sample injection port for
small volume injections and a 0.10 mL sample loop. A 10 port switching valve was fitted with
a 1.5 m x 3 mm stainless steel 60/80 Chromosorb 102 (column 1) and a 5 m x 3 mm stainless
steel 60/80 Molecular Sieve 5A (column 2). A sample was injected onto column 1 from the
injection port or the sample loop and held for 40 sec to allow the composite peak of Ar and
CH4 to pass onto column 2, while the CO2 peak was retained. The valve was switched,
reversing the column sequence allowing the CO2 to pass directly to the detector, followed by
the Ar and CH4 peaks that were separated in column 2. The operating conditions were as
follows: injection port, column, and detector temperature, 70°C; and helium carrier gas flow
rate, 25 mL/min.
Working gas standards were prepared from commercially available mixed gas standards
(Scott Specialty Gases, Inc., Plumsteadville, Penn.). A serum bottle was immersed in water
and inverted. The gas from the standard cylinder was allowed to displace the water in the
inverted bottle. The serum bottle was capped with a butyl rubber serum stopper. In this
12
-------
fashion, standards could be prepared that were saturated with water vapor and that were
sampled in the same manner as the microcosms.
Quantities of CPfy and CC>2 in the gas phase of the microcosms were calculated as
follows
i ~TT ^ c T, Pmm 273 /•> 1\
moles CH4 or CO2 = - £..v._sa. ---- (2-1)
4 2 100 * 760 T 22.4 L
where
T = temperature (°K)
Pmm = atmospheric pressure (mm Hg)
Gc = concentration of CH4 or CO2 in the gas phase (%)
Vg = volume of gas phase in microcosm plus the volume of sampled gas (L).
The quantities of dissolved CH4 or CC>2 in the microcosm liquid were calculated as
follows
dissolved moles CH4 or CO, =a.. V, . -£==L. -L^L (2-2)
4 2 100 ' 760 22.4 L
where
a = absorption coefficient (mL gas/mL water)
0.8290 for CO2 and 0.0318 for CH4
Vs = volume of liquid in microcosm (L)
The absorption coefficient (Dean, 1973) is the volume of gas (at standard temperature
and pressure) per volume of water when the partial pressure of the gas is 1 atm.
Dissolved HCOJ was calculated as follows
moles HCOi=Kl'C°'2'Vs (2-3)
H+
where
KI = 4.17 x 10"7 (the log of pKal - the equilibrium constant between between
CC>2 and HCO3 in water, mol/L, Stumm and Morgan, 1981)
CC>2= concentration of carbon dioxide (mol/L)
H = hydrogen ion concentration (mol/L).
13
-------
Small microcosms were prepared as the large microcosms, except that 500 mL serum
bottles were used. The serum bottles contained approximately 650 g of contaminated aquifer
material and approximately 250 mL of PRAS mineral salts containing the organic compound
of interest. Microcosms used for the study of PAH biodegradation were prepared in 500 mL
screw cap bottles capped with Teflon-lined silicone rubber septa and sealed with an open hole
screw cap. Microcosms containing other compounds were capped with sleeve-type red
rubber serum stoppers.
Head space volumes, gas production, and gas composition were determined as for the
large digesters above with the following exception - gas composition was determined using
an injection of the head space gas with 0.10 mL gas tight syringe in lieu of the sample loop.
MICROBIAL CHARACTERIZATION OF THE MICROCOSMS
The total biomass concentration in the microcosms was determined by total protein on
10.0 g subsamples of the sediment in the same manner as described for the aquifer core
material. The total protein present in 5.0 mL subsamples of the liquid in suspension was also
determined by the method of Ga'lli (1987) and added to that present on the aquifer material.
Total aerobic, denitrifying, sulfate reducing, and methane producing bacteria attached onto the
sediment and in suspension were determined as previously described. All microcosms were
examined under phase contrast and epifluorescence (Mink and Dugan, 1977; Doddema and
Vogels, 1978) with an American Optical Fluorstar microscope fitted with a 50-W mercury
vapor lamp and Fluor Cluster no. 2073 (American Optical Co., Buffalo, N.Y.).
NONLINEAR REGRESSION ANALYSIS
The object of mathematical modeling is to construct, from theoretical and empirical
knowledge of a process, a mathematical formulation which can be used to predict the behavior
of the process. Complete understanding of the mechanism of the chemical, physical, or
biological aspects of the process under investigation is not usually possible. However, some
information on the mechanism of the system may be available; therefore, a combination of
empirical and theoretical methods can be used.
The development of mathematical models often requires the implementation of an
experimental program in order to obtain the necessary information for the verification of the
models. The experimental program is originally designed based on the theoretical
considerations coupled with a priori knowledge of the mechanisms and is subsequently
modified based on the results of regression analysis.
14
-------
Regression analysis is the application of mathematical and statistical methods for the
analysis of the experimental data, and the fitting of the mathematical models to these data by
the estimation of the unknown parameters of the models (Bates and Watts, 1988). The series
of statistical tests, which normally accompany regression analysis, serve in model identification
and verification.
The equations comprising the mathematical model should show the effect of the control
variables (e.g., temperature, pressure, and/or the concentration of the components) on the
evolution of the dependent variables and are usually a set of differential equations and/or a set
of algebraic equations (Constantinides, 1987). For example, a set of ordinary differential
equations describing the dynamics of a process may have the general form
where
x = independent variable
y = vector of dependent variables
b = vector of parameters whose values must be determined.
In their integrated form, the above set of equations convert to
y =/W») (2-5)
For regression analysis, mathematical models are classified as linear or nonlinear with
respect to the unknown parameters. For example, the following differential equation
describing microbial growth
f-A* CM)
that is linear with respect to the dependent variable X, is nonlinear with respect to the
parameter ji. This is clearly shown by the integrated form of this equation
X = X0e^ (2-7)
where X is highly nonlinear with respect to (I.
15
-------
Most mathematical models encountered in the sciences are nonlinear in the parameters
(Constantinides, 1987). Attempts at linearizing these models, by rearranging the equations
and regrouping the variables, were common practice in the pre computer era. Such
techniques can lead to sizable errors in the parameters and, as a result, have been replaced by
the implementation of nonlinear regression methods.
Nonlinear regression analysis (NLR) is an extension of the linear regression methods used
iteratively to arrive at the values of the parameters of the nonlinear models. The statistical
analysis of the nonlinear regression results is also an extension of that applied in linear analysis
but does not possess the rigorous theoretical basis of the latter. Consider, for example, the
analysis of a complex chemical reaction such as
(2-8)
E + F
where the rate of formation of each component may be written as
—
^-
dt (2-9)
— = k2B-k3AnCm
dt l 3
^ = k3AnCm
dt 6
This is only one possible formulation of the reaction mechanism; however, the equations
in the given example contain five unknown parameters - £j, k2, k3> n, and m - which must be
calculated by fitting the model to the experimental data, assuming that the concentrations of
A, B, C, and E can be obtained by analytical means. The solution can obtained by the
simultaneous fitting of the rate equations to the data.
Nonlinear regression analysis enables one to fit models consisting of multiple dependent
variables to multiresponse experimental data in order to obtain the values of the parameters of
the model which minimize the overall residual sum of squares (RSS) or
RSS = ZCyobs-;ypred)2 (2-10)
16
-------
where RSS is the sum of the squares of the deviations, y^ represents the observed values of
the dependent variable, and ypred represents the predicted y values (Robinson, 1985).
Values for the parameters that minimize the RSS must be recursively obtained. An initial
set of parameter estimates is determined either by using a linearized form of the chosen
nonlinear model or by guesswork. An initial RSS value is calculated and a new set of
parameter values is calculated in some fashion. The new RSS is compared with the RSS for
the initial parameter estimates, and if the former is less than the RSS for the initial estimates,
then the second set of parameter estimates replaces the first. This process continues until the
RSS reaches a minimum, at which point the best parameter values have been located.
The most common method used to calculate a new set of parameter estimates is the
Gauss-Newton method. The mathematical elements of the Gaussian method are derived
through the application of a truncated Taylor series expansion (Bates and Watts, 1988). This
expansion essentially linearizes the nonlinear RSS function in the neighborhood of the best
parameter estimates. If a one-parameter nonlinear model is considered, a Taylor series
expansion about the best-parameter estimate B given an initial value b is described by
(2-11)
db
where RSS(fl) and RSS(£) define the RSS curve as a function of B and b, respectively. The
term dRSS(b)/db is the first derivative of the RSS function with respect to the estimate of B,
namely b. For models with more than one parameter, the right-hand side of equation 2-1 1
must be modified to include (1) the derivatives for the other parameters and (2) the differences
between the estimates of these additional parameters and the values that define an RSS
minimum. The Gaussian method can be used to estimate the parameters of linear models and
will always converge on the minimum RSS value; however, this is not always the case with
"ill-conditioned" nonlinear models (Bard, 1974; Beck and Arnold, 1977; Bates and Watts,
1988). An ill-conditioned model is one whose sensitivity equations (i.e., the partial
derivatives of the dependent variables with respect to the parameters) are nearly proportional,
and for these models the Gaussian method can fail to find the minimum RSS. This can occur
even when the initial estimate is very close to the best values.
One of the commonly used variations to overcome the shortcomings of the Gaussian
method is the Levenberg-Marquardt modification (Bard, 1974; Beck and Arnold, 1977; Bates
and Watts, 1988). This technique is a combination of the Gaussian method and the method of
17
-------
steepest-descent that alters both the step size and direction taken by the Gaussian technique,
attempting to ensure that the RSS is sequentially reduced.
RSS analysis along with statistical analysis of the regression results allows one to (1)
decide whether the model gives satisfactory fit within the experimental error of the data, (2)
discriminate between competing models, (3) measure the accuracy of the estimation of the
parameters by constructing the confidence region in the parameter space, and (4) measure the
correlation between parameters by examination of the correlation coefficient matrix.
NONLINEAR REGRESSION ANALYSIS OF MICROCOSMS
Only the substrate depletion curves were fitted to the Monod equations using NLR.
Biomass increase was not included in the model fits because of the lack of data points and
the uncertainty associated with the biomass analyses. The method of Marquardt (Bard 1974)
was used for the determination of parameter values that best fit the experimental substrate
depletion data by minimizing the RSS. Because the Monod equations do not have explicit
analytical solutions for substrate and biomass concentrations as a function of time, a
simultaneous solution of both equations was accomplished using a fourth-order Runge-
Kutta numerical-procedure (Constantinides 1987). The statistical basis for these analyses is
presented by Robinson (1985), and requires that the sensitivity of the dependent variable to
changes in each of the parameters be calculable. The partial derivatives of 5 with respect to
V-max' Ks>ancl Y satisfy Ms requirement. These expressions are derived from the integrated
Monod substrate utilization equation by implicit differentiation. Unique determination of the
parameters can best be obtained when S0 (initial substrate concentration) is in the mixed-
order region (0.5 Ks to 2 Ks) and then letting S proceed through the first-order region
(< 0.5 Ks) during the course of the experiment.
COLUMN SORPTION STUDIES
Sorption experiments were done using readily available HPLC (Goerlitz, 1984)
equipment thus allowing automation of the elution and effluent monitoring. The HPLC
equipment consisted of two Isco Model 314,350 mL syringe HPLC pumps; an Isco Model
V4 UV variable wavelength detector; a Waters Model R401 refractive index detector (RI);
and a computer controlled data collection and analysis system (Figure 2-2).
Glass beads (60-80 mesh) were packed into a 25 mm x 500 mm glass column. The
material was packed down firmly with a Teflon tipped tamper after every 3 to 5 mm. After
the column was packed, autoclaved, and the air replaced with sterile CO2, a sterile solution of
18
-------
0.010 M CaCl2 (pH 5.9) was pumped through the column until no gas bubbles were visible.
Similar columns were also packed with uncontaminated aquifer material from Site 1 at the
research site. The aquifer material was air dried, sieved through a 1.0 mm screen,
remoistened to the point of aggregation, and packed in small amounts as was done for the
glass beads.
Initially the 0.010 M CaCl2 eluent was pumped until instrumental stability was achieved.
To start the experiment, the feed solution was abruptly changed for the test solution at the
head of the column. The column characteristics were determined by changing the 0.010 M
CaCl2 eluent to a 0.015 M CaCl2 (pH 5.9) solution. Soluble compounds were tested by
adding the compound of interest to the eluent. Sparingly soluble compounds were introduced
by directing the eluent through a "saturator" column. The saturator column was prepared in a
4 mm x 300 mm commercially available empty stainless steel HPLC column. The column was
packed with 60-80 mesh Chromosorb W coated with a 5 percent by weight coating of the
selected compound. The column material was prepared by first dissolving the correct amount
of organic compound in a volume of 50 ml of Ct^C^- The amount of Chromosorb W for a 5
percent solution was added along with Ct^C^"01^™0 solution to a rotary vacuum
evaporator and rotated until dry. It has been shown that exposure of such a large area, 2-3
m2/g for this substrate, to water brings about saturation with respect to the coating. By the
use of two pumps, solutions at any desired concentration could be prepared as needed.
NONLINEAR REGRESSION ANALYSIS OF COLUMN SORPTION STUDIES
The parameter values (r]/2 and P) for the linear-equilibrium model were estimated using
NLR analysis. The maximum neighborhood method of Marquardt (Bard, 1974) was used to
fit the experimental data points to analytical solution of the one-dimensional
convection-dispersion equation as given by Hashimoto et al., 1964 (equation 6-16) that
minimized the RSS. The computer program HASHPE was used for this portion of the study
(Oravitz, 1984).
The parameter values (P, Rt, p, and co) for the non-equilibrium model were estimated as
above using the computer model CFITIM reported by van Genuchten (1981). This model
also uses the maximum neighborhood method of Marquardt to minimize the RSS when the
experimental data was fitted to the analytical solution presented (equations 6-31 to 6-33).
19
-------
Gas Sampling Port
Gas Volume
Measurement Syringe
Butyl Rubber Stopper
Aquifer Material
U-Tube Manometer
Liquid Sampling Syringe
3 Way Valve
Liquid Sampling Port
Liquid
Figure 2-1. Sample anaerobic microcosm used for biodegradation studies.
20
-------
Mixing
Tee
Saturator
Column
-D
HPLC SYRINGE
PUMPS
Pressure Relief
Valve
Sediment
Column
RI UV
DETECTOR DETECTOR
Figure 2-2. Aquifer column adsorption study apparatus.
21
-------
3. BIODEGRADATION OF WSF COMPOUNDS
RESEARCH TASKS
Laboratory microcosms were establish to simulate the subsurface environment in order to
determine the biodegradability of compounds in the WSF under methanogenic conditions.
The compounds included nitrogen heterocycles (pyridine, indole, quinoline, carbazole,
acridine) sulfur heterocycles (thiophene, benzothiophene, dibenzothiophene) oxygen
heterocycles (furan, benzofuran, dibenzofuran) polynuclear aromatic compounds (indene,
naphthalene, acenaphthene) and pentachlorophenol. The degradation of the WSF compounds
was tested using natural consortia from subsurface materials at the Pensacola, Florida site.
Relevant in-situ biotransformation processes and pathways occurring in the subsurface
environment were be compared to pathways determined in laboratory studies. The USGS has
compiled a data base with over 200 GC/MS/DS analyses of ground-water samples from the
research site from which intermediate compounds were compared.
BACKGROUND
Anaerobic microbial transformations of oxygen substituted aromatic compounds under
methanogenic, denitrifying, and sulfate reducing conditions is a well known fact. The current
knowledge has been summarized in two recent review papers (Berry et al., 1987a; Young,
1984). It is not possible to present a comprehensive review of the literature on this occasion,
but is worth mention that the anaerobic degradation of phenol and 2-, 3-, and
4-methylphenol has been observed not only in laboratory microcosms, but in ground water
aquifers under methanogenic and sulfate reducing conditions (Godsy et al., 1983; Smolenski
and Sulfilta, 1987).
In mixtures of phenolics, phenol was preferentially utilized, followed by 4-methylphenol,
and ultimately 3-methylphenol (Fedorak and Hrudey, 1986); the acclimation lag before the
onset of biodegradation, as well as toxicity of these substrates to anaerobic cultures, also
increase in this order (Fedorak and Hrudey, 1984; Fedorak and Hrudey, 1986; Roberts et al.,
1987).
In 1986, the proof for anaerobic biodegradation of non-oxygenated aromatic
hydrocarbons (e.g., benzene, toluene, and xylenes) was finally obtained (Grbid-Gali<5 and
Vogel, 1987; Vogel and GrbicXJalic", 1986). These processes occur under denitrifying (Kuhn
22
-------
et al, 1988; Zeyer et al, 1986) and under fermentative conditions (Wilson et al, 1987). The
mechanism of the initial anaerobic oxidation of aromatic hydrocarbons is hydroxylation with
water, the same is true for heterocyclic hydrocarbons. Fossil fuel derived heterocyclic
hydrocarbons (e.g., indole, quinoline, benzothiophene, etc.) and homocyclic hydrocarbons
have been shown to degrade under methanogenic, denitrifying, and sulfate reducing conditions
(Bak and Widdel, 1986; Berry et al, 1987a; Godsy et al, 1992; Godsy and Grbid-Galid,
1989; Mihelcic and Luthy, 1988; Pereira et al, 1987).
RESULTS
Compounds Tested for Biodegradability
Microcosms were prepared in triplicate for each of the compounds tested (Figures 3-1
and 3-2) in order to determine: (1) substrate utilization, (2) CH4 and CO2 production, and (3)
if degradation was occurring in an autoclaved control. A single positive control was prepared
from aquifer material and mineral salts solution only to test for the production of CIfy and
CC>2 from any organic compounds that might be present on the aquifer material. A positive
control was prepared for each new inoculum. Physiological types of bacteria and biomass
analyses were not performed in these series of experiments. Table 3-1 contains the
compounds tested, concentration of test compounds, days until the onset of rapid
methanogenesis, duration of methanogenesis, and percent theoretical conversion according to
the equation of Tarvin and Buswell (1934).
" a " "~ (3-1)
Also tested in the same manner were dibenzothiophene, dibenzofuran, and acenaphthene.
These compounds did not show any signs of biodegradation after 294 days of incubation, at
which time the microcosms were discarded.
The autoclaved controls and the replicate microcosms for gas analyses for each of the
tested compounds were allowed to incubate approximately 14 days longer than the viable
microcosms before they were sacrificed for analysis. The autoclaved controls did not show
any detectable loss of the compounds for the duration of the experiment and the positive
control did not produce detectable amounts of Clfy and CO2 after 150 days of incubation.
The times before the rapid onset of methanogenesis varied from as short as 11 days for
indole to as long as 98 days for 2-methylphenol; however, once methanogenesis started, the
23
-------
duration was essentially the same for each microcosm and lasted for approximately 25 days.
The exceptions to this observation were indole, quinoline, and isoquinoline which must be first
oxidized to oxindole, 2(lH)-quinolinone, and l(2H)-isoquinolinone, respectively, before the
conversion was observed of the persistent oxidized intermediate compound to CH4 and CO^.
Augmented Microcosms
A series of microcosms were prepared to determine if the long times before the onset of
methanogenesis from 2-methylphenol, l(2H)-isoquinolinone, and naphthalene could be
shortened by the addition of benzoate (10 mg/L), an intermediate in the degradation of these
compounds, and/or a mixture of volatile fatty acids (VFA) - 45 mg/L acetate, 30 mg/L
propionate, and 15 mg/L butyrate. The microcosms were prepared in 500 mL serum bottles
containing fresh unacclimated aquifer material with just the compound of interest (1), or
augmented with benzoate (2), volatile fatty acid mixture (3), and both benzoate and volatile
fatty acids (4). The three aromatic compounds of interest were added at a concentration of 25
mg/L. An autoclaved sterile control was prepared for each compound. The controls did not
produce any detectable amounts of CH4 and CC>2 after 307 days of incubation. Given in
Table 3-2 are the augmentation mixtures used, the days until the onset of rapid substrate
degradation, and the duration of substrate degradation.
In the case of 2-methylphenol, it appears that the addition of the VFA mixture alone or
with benzoate considerably hastened the onset of substrate degradation. Benzoate alone did
not affect the time to onset of degradation. Benzoate hastened the onset of substrate
degradation for naphthalene, while both amendments slightly decreased the time to onset of
degradation for l(2H)-isoquinolinone.
The addition of potential intermediates to this microcosm generally increased the duration
of substrate degradation with the exception of 2-methylphenol. In this case, the addition of
benzoate slightly decreased the duration of substrate utilization.
Augmented Mixed-Substrate Microcosm
In a separate experiment, 2-methylphenol, 3-methylphenol, 2(lH)-quinolinone, and
l(2H)-isoquinolinone were dissolved in the mineral salts solution and placed in a single 500
ml microcosm containing fresh unacclimated aquifer material. The microcosm was augmented
with the mixture of 45 mg/L acetate, 30 mg/L propionate, and 15 mg/L butyrate. A sterile
control that was similarly prepared did not produce any detectable amounts of CH4 or CC>2
after 100 days of incubation. Given in Table 3-3 are concentrations of individual compounds
24
-------
used, the days until the onset of substrate degradation, and the duration of substrate
degradation. The degradation sequence, 2-methylphenol > 3-methylphenol >
2(lH)-quinolinone > l(2H)-isoquinolinone, is quite a different sequence than was observed in
the large WSF microcosm and during down gradient travel in the aquifer as discussed in
proposal, where, in both instances, these compounds were simultaneously degraded.
DISCUSSION AND CONCLUSIONS
These experiments clearly demonstrate that the WSF compounds, 2-methylphenol and
quinoline, which have been reported to be recalcitrant by a number of workers, are indeed
biodegradable under the methanogenic conditions that exist in the contaminated portion of the
aquifer. The ability of the consortia to degrade these compounds may arise from the long
period of contamination (~ 80 years) and exposure during this time only to creosote-derived
compounds at the Pensacola site. In contrast, sewage sludge derived cultures used by
Chmielowski et al. (1965), Boyd et al. (1983), and Fedorak and Hrudey (1984, 1986) were
continuously supplied with a myriad of readily degradable organic compounds and perhaps
never developed the necessary uptake and enzyme systems required for growth on low
concentrations of homocyclic and heterocyclic compounds.
It would be very difficult to compare the differing times required before the onset of rapid
methanogenesis for those compounds tested in Table 3-1 because these experiments were
conducted over a period of four years using a minimum of four different inocula; however, all
of the inocula were collected form the approximate centroid of the contamination (Site 39) at
a depth of approximately 6.1 m. It is quite possible that these inocula differed in the
concentrations of organisms active towards the tested compounds, for it is unknown at this
time whether one or many different organisms are responsible for the initial attack on the
individual aromatic compounds. Once the ring structure is cleaved, the organisms necessary
for the conversion to CFfy and CO2 are most likely to be similar, if not the same, for each of
the compounds tested.
The mixed substrate-augmented microcosm (MSAM) demonstrated a different
degradation sequence than was observed in the large WSF microcosm discussed in the
proposal. The MSAM microcosm was augmented with only the VFA mixture. The
degradation of 2-methylphenol started after only 16 days, an onset time very similar to the 12
to 18 days required during experiments given in Table 3-2. The first compound to be
degraded was 2-methylphenol, followed by 3-methylphenol, 2(lH)-quinolinone, and
1 (2H)-isoquinolinone. These compounds were all degraded sequentially one at a time in the
25
-------
order mentioned above, unlike the large WSF microcosms and during down gradient travel in
the aquifer, where these compounds were all degraded more or less at the same time.
Recently, Roberts et al. (1986, 1990) may have provided a possible explanation to the
sequential degradation of the methylphenols by demonstrating that 4-methylphenol undergoes
methyl hydroxylation to form 4-hydroxybenzoate when an electron-attracting group is in the
4-position to stabilize the intermediate. Presumably, this reaction sequence would apply to
2-methylphenol with the an electron-attracting group in the 2- position although they did not
investigate this compound. 3-Methylphenol, however is carboxylated to 2-methylbenzoate,
analogous to phenol activation. The different pathways result in the methyl group of
4-methylphenol ending up as CC>2 while the methyl group of 3-methylphenol goes to
Experiments conducted to determine if the addition of benzoate or the VFA mixture was
responsible for decreasing the time to onset of degradation suggest that the addition of the
VFA mixture decreases the onset time for the single ring phenolic compounds to greater
extent than for naphthalene and l(2H)-isoquinolinone. The presence of the VFA mixture in
the MSAM adds credence to this observation. In this microcosm, the phenolic compounds
were the first to be degraded. In contrast, the addition of benzoate hastened the onset of
degradation of both naphthalene and l(2H)-isoquinolinone, since benzoate is presumably an
intermediate compound in the degradation pathway of these compounds. In older published
literature (Evans, 1977; Young, 1984), pathways were suggested for phenol degradation
which proceeded through ring saturation to cyclohexanol, conversion to cyclohexanone, and
subsequent ring cleavage. Recent studies by Haggblom et al. (1990) and Schink et al. (1992),
however, have suggested that phenol is first carboxylated to form hydroxybenzoate indicating
that phenol enters the benzoate degradation pathway. Since the addition of benzoate
apparently did not stimulate the degradation of the phenolic compounds, it is possible that the
long time until the onset of degradation is the time required for the population to prepare
these compounds for introduction into the benzoate pathway. The addition of the
supplemental substrates should supply the ring cleavage organism(s) with benzoate and the
acetogenic and methanogenic populations with volatile fatty acids. This would allow for a
more rapid increase in numbers of active microorganisms resulting in a shorter time to onset
of methanogenesis and/or possibly a differing degradation sequence.
26
-------
OH
Phenol
OH
OH
CHg
2-Methylphenol
OH
CHr
3-Methylphenol
CHg
4—Methylphenol
Figure 3-1. Phenolic compounds tested for biodegradability to CH4 and CC>2.
27
-------
H
Indole
,',
Oxindole
Quinoline
.N
Isoquinoline
Benzothiophene
Benzofuran
Indene
H H
H-l I-H
H
2( lH)-quinolinone
0
1 (2H)-isoquinolinone
Dibenzothiophene
0
Dibenzofuran
Naphthalene
Acenaphthene
Figure 3-2. Homocyclic and heterocyclic aromatic compounds tested for biodegradability to
CH4 and CO2.
28
-------
Table 3-1. Compounds found to act as sole carbon and energy sources in 500 mL
methanogenic microcosms.
Compound
Phenols
Phenol
2-Methylphenol
3-Methylphenol
4-Methylphenol
NSO Compounds
Indole
Oxindole
Quinoline
2(lH)-quinolinone
Isoquinoline
1 (2H)-isoquinolinone
Benzothiophene
Benzofuran
PAH Compounds
Indene
Naphthalene
mg/L
26.5
22.6
18.0
26.1
25.8
31.5
23.3
27.3
22.8
30.6
12.3
7.9
12.0
18.5
Days to onset
of substrate
degradation
65
98
29
84
11
38
20
47
50
75
35
19
18
30
Duration of
substrate
degradation
25
37
21
26
41
26
59
30
69
26
21
26
19
25
% Theoretical
yield of total gas
produced
92.4 ± 6.0
86.4 ±5.6
89.9 ± 5.8
93.516.1
86.8 ± 5.6
92.4 ± 6.0
92.1 ±6.0
89.2 ± 5.8
90.6 ±5.9
85.8 ±5.8
88.9 ±5.8
94.2 ±6.1
91.7 ±6.0
89.9 ± 5.8
29
-------
Table 3-2. Degradation of individual compounds in unamended microcosms (1), or
microcosms augmented with benzoate (2), volatile fatty acids (3), or both mixtures (4).
Augmentation Days to onset of Duration of
Compound mixture substrate substrate
degradation degradation
2-Methylphenol
1 (2H)-isoquinolinone
Naphthalene
(1)
(2)
(3)
(4)
(1)
(2)
(3)
(4)
(1)
(2)
(3)
(4)
Table 3-3. Degradation of a mixture
substrate-augmented microcosm.
Compound
2-Methylphenol
3-Methylphenol
2( lH)-quinolinone
1 (2H)-isoquinolinone
mg/L
42.6
40.2
34.0
26.0
98
94
18
12
276
195
215
178
202
126
168
174
of compounds and VFA in
Days to onset of
substrate
degradation
16
51
67
87
47
35
118
75
55
76
92
64
75
87
92
98
the mixed
Duration of substrate
degradation
34
26
31
32
30
-------
4. DETERMINATION OF THE PATHWAYS OF METHANOGENIC
DEGRADATION FOR THE WSF COMPOUNDS
RESEARCH TASKS
Relevant in-situ biotransformation processes and pathways occurring in the subsurface
environment were be compared to pathways determined in laboratory studies. The USGS has
compiled a data base with over 200 GC/MS/DS analyses of ground-water samples from the
research site from which intermediate compounds were compared.
BACKGROUND
The ultimate fate of homocyclic and heterocyclic aromatic compounds in subsurface
environments is controlled by various transport and transformation processes. Potentially the
most important, but currently one of the least understood, processes affecting ground-water
quality is biotransformation of these pollutants by indigenous microorganisms. Even less is
known about these transformations under methanogenic conditions. Several homocyclic and
heterocyclic aromatic hydrocarbons, which contain nitrogen and sulfur in their ring structure,
have been shown to be susceptible to transformations under methanogenic conditions (Pereira
el al., 1987) and to be biodegradable to CH4 and CO2 (Berry el al., 1987a; Godsy el al.,
1989; and Godsy and Grbic-Galic", 1989); the current knowledge has been summarized in four
recent review papers (Berry el al., 1987a; Grbic-Galic", 1989; and Grbid-Galid, 1990; Schink
el al., 1992). In this section the methanogenic pathways of biodegradation for quinoline and
naphthalene are implied from the analysis of intermediate compounds that appear in the
growth liquor at the onset and during methanogenesis in laboratory microcosms, and
compounds appearing in ground water samples.
RESULTS
Degradation Pathway for Nitrogen-Heterocyclic Compounds
Compounds detected in the growth liquor just before the onset and during
methanogenesis of quinoline are shown in Figure 4-1. Again, compounds that were found in
the autoclaved and organic free controls are not shown. Based on the report of Pereira el al.
(1988), who demonstrated that the first step in the oxidation of quinoline consisted of
oxidation of the heterocyclic ring, the major pathway involves the hydrolytic cleavage of the
oxidized ring. The transient appearance of aniline in the growth liquor suggests that this
31
-------
compound is involved in the major pathway. The minor homocyclic ring oxidizing pathway
observed during the methanogenesis of benzothiophene was also possibly occurring during the
methanogenesis of quinoline. The oxidized intermediate of the minor pathway, 8-quinolinone
<-» quinoline-8-ol, was not detected in the microcosms but was observed in many of the
ground water samples from the study site. After hydrolytic cleavage of the oxidized
homocyclic ring, the remaining compounds were presumably subjected to the similar reactions
as was observed during methanogenesis of benzothiophene. This was corroborated by the
appearance of 2,3-2.
During the methanogenesis of indole in the microcosms, oxindole was the only
intermediate that was observed before complete conversion to CH4 and CC>2. Isatin and
anthranilic acid, however, were detected in ground-water samples from the research site.
These compounds had been shown to be intermediates in the degradation of indole under
denitrifying and, presumably, methanogenic conditions (Madsen and Bollag, 1989).
Degradation Pathway for Homocyclic Compounds
Intermediate compound analyses for indene and naphthalene transformation were
performed in the same manner as for the heterocyclic compounds. The analyses on
microcosms containing indene did not reveal any intermediates, other than the end products
CH4 and CC>2. The aquifer sediments used as the inoculum must have been very active
towards this compound. The analyses for intermediate compounds of naphthalene (Figure
4-2) revealed 2-ethylphenol, a compound that would be in the major degradation pathway
resulting from the hydrolytic cleavage of one of the naphthalene rings. This intermediate
would then be subjected to various reactions similar to the above resulting in CH4 and CC>2 as
the end products.
Ground-water samples from the site contained both l^iaphthalenol and 2-naphthalenol,
possible candidates for the first oxidation step for naphthalene. It is not certain whether these
compounds arise from the degradation of naphthalene or are constituents of creosote itself.
32
-------
Benzofuran was also observed in the growth medium of naphthalene, but the role of this
compound is inconclusive from these studies; however, this compound probably resulted from
a reaction involving the hydroxyl-group of 2-hydroxyphenylacetaldehyde (an oxidation
intermediate of 2-ethylphenol) and the carbonyl carbon of the aldehyde group, followed by
dehydration to yield a furan as proposed by Grbid-Galic (1989).
DISCUSSION AND CONCLUSIONS
Environmental contamination by homocyclic and heterocyclic aromatic compounds is
widespread, but very little information is available on the biotransformation pathways of this
class of compounds under methanogenic conditions and even less on the environmental fate of
these compounds in contaminated ground water. Recent studies on the methanogenic
transformation of aromatic compounds in general have demonstrated three major ring
cleavage transformation pathways are used (Young, 1984; Berry et al, 1987a; Evans and
Fuchs, 1988; Schink et a/., 1992): the benzoic acid, the resorcinol, and the phloroglucinol
pathway. In older reviews, pathways for phenol degradation were suggested which proceeded
through complete ring saturation to cyclohexanol, conversion to cyclohexanone, and
subsequent ring cleavage (Evans, 1977; Young, 1984). These pathways were based on simple
analogy to the benzoate pathway, but no experimental evidence of their existence has ever
been provided. An alternative pathway for phenol was postulated through which the quinoid
phenol tautomer was reduced to form cyclohexanone. This mechanism takes advantage of the
polarizing effect of the hydroxyl group on the 7i-electron distribution in the aromatic ring
(Bakker, 1977; Schink and Tschech, 1988). Recently, evidence has been presented that links
the phenol pathway with the benzoate pathway (Haggblom et al., 1990; Knoll and Winter,
1989; Sarak Genther et al., 1989). The carboxylation of phenol to 4-hydroxybenzoate was
proposed as the first step in the degradation. This step was then followed by reduction of the
hydroxyl group and removal to form benzoate. Because the ground-water consortia used for
this portion of the study were mixed ground water cultures, it is not surprising that the major
transformation pathways for benzothiophene and quinoline intersected segments of both the
benzoate and/or the phenol pathway. It is not possible to determine at this point if phenol or
benzoate is of more importance as key intermediate in the methanogenic transformation of
quinoline and benzothiophene in these microcosms. However, Grbid-Galic" (1989) found that
phenol was the most persistent simple aromatic intermediate of the methanogenic degradation
of naphthalene in a fed-batch reactor. It is unknown at this time whether phenol was then
degraded by the pathway suggested by Schink and Tschech (1988) or carboxylation to
4-hydroxybenzoate and then degradation via the benzoate pathway.
33
-------
As expected, the major pathways for anaerobic transformation of benzothiophene and
quinoline seem to be initiated by the oxidation and then cleavage of the nitrogen- or
sulfur-heterocyclic ring. Delocalization of the 7C-electrons in the heterocyclic ring results in
the heteroatom being relatively electron poor and the carbons in the ring being relatively
electron rich. The carbon atom proximal to the heteroatom, as a result of the electron
delocalization, is then susceptible to nucleophilic attack by water This observation is in line
with the findings of Berry el al. (1987b), who demonstrated that the first step in the
methanogenic transformation of indole, the nitrogen analog of benzothiophene, was the
oxidation of the nitrogen-heterocyclic ring to form oxindole. Similarly, Pereira el al. (1988)
were able to demonstrate the oxidation of the nitrogen-heterocyclic ring of quinoline using
H218O to form 18O-labeled 2(lH)-quinolinone under methanogenic conditions; however,
they were not able to demonstrate the methanogenic fermentation of this compound. This
was probably due to their use of reasonably fresh sewage sludge as the inoculum for the
microcosms. In this work, it was found that the substituent side chains attached to the
homocyclic ring after the cleavage of the heterocyclic ring were subjected to various
transformation reactions, as exemplified by the myriad of single-ring compounds found. The
large number of compounds is, undoubtedly, a result of the activity of a complex
ground-water consortium.
The other, as yet unreported minor pathway for heterocycles involves the initial oxidation
of the homocyclic ring, hydrolytic cleavage of this ring and then attack on the heterocyclic
ring. Several years ago, Vogel and Grbid-Galic (1986) showed that the oxidation of an
unsubstituted homocyclic aromatic ring was possible when they demonstrated the oxidation of
benzene to CH4 and CO2 by a mixed methanogenic culture. The magnitude, importance, and
organisms involved in the minor transformation pathway for benzothiophene and quinoline
and possibly other NSO compounds are unknown at this time; however, the mechanism and
organisms involved in the minor pathway might be similar to those initiating the attack on
indene, naphthalene, and most likely, benzene and toluene. The oxidation of the homocyclic
ring at positions 5 or 8 for quinoline, or at positions 4 or 7 for benzothiophene, are consistent
with observations of organic synthesis reactions, where the carbons of the homocyclic ring
proximal to the heterocyclic ring are most susceptible to attack (Vollhardt, 1987).
The inability to identify intermediate compounds during the methanogenesis for some
compounds (e.g., benzofuran and indene) may be a result of the low starting concentrations of
the compounds in the microcosms. Compounds that could be intermediates of these were
observed in ground water samples. 3(2H)-Benzofuranone and both
34
-------
2,3-dihydro-indene-l-one and 1,3-dihydro-indene-2-one Were observed in ground water
samples and could be intermediates in the degradation of benzofuran and indene, respectively.
35
-------
OXIDATION OF THE
HOMOCYCLIC
RING
8-Quinolinone
OH
8-Quinolinol
OXIDATION OF THE
HETEROCYCLIC
RING
2-Quinolinol
2(lH)-Quinolinone
HYDROLYTIC CLEAVAGE OF
THE OXIDIZED RING
2,3-Dimethylpyridine
\ 9
RING CLEAVAGE
'NHr
Aniline
Benzole Acid
Phenol
\ ACETOGENESIS
\
\
\
Acetic Acid, Formic Acid
I
METHANOGENESIS
CH4 C02 NHg*
® Proposed intermediate - not detected in microcosms but present in ground-water samples.
ft Ammonia analyses were not performed.
Figure 4-1. Proposed pathways of quinoline transformation under methanogenic
conditions in aquifer-derived microcosms. All intermediates shown were detected in the
culture fluid except for those labeled by asterisks.
36
-------
Naphthalene
OXIDATION AND
HYDROLYTIC CLEAVAGE OF
ONE RING
I
CHg
OH
2-Ethylphenol
i
i
i
RING CLEAVAGE
i
i
ACETOGENESIS
Acetate
METHANOGENESIS
CH4 C02
Figure 4-2. Proposed pathway of naphthalene transformation under methanogenic
conditions in ground water aquifer-derived microcosms.
37
-------
5. DETERMINATION OF DEGRADATION KINETICS OF THE WSF
COMPOUNDS
RESEARCH TASKS
Relevant rates of biotransformation processes occurring in the subsurface environment
were determined in laboratory studies. The kinetic constants were determined for a number of
models under conditions that simulated those in the ground-water environment from the
research site.
BACKGROUND
Microbiologists have realized for some time the need for quantitative data in their
research on the movement and fate of organic pollutants in the environment. With the
increased ability of computer simulations to model comprehensively the physical and chemical
factors affecting the fate of pollutants, microbiologists now can describe not only the types of
organisms that inhabit a particular environment but also the rates at which they perform
metabolic functions that affect pollutants. This quantitative approach involves the
determination of parameters in equations chosen to represent the process under study: in this
instance, organic substrate utilization and concomitant bacterial growth. These parameters
can then be incorporated into computer models resulting in improved simulations.
The relationships developed by Monod and others resulted from experiments using
suspensions of pure cultures of bacteria utilizing single organic compounds. It remains to be
determined if these equations, with or without bacterial decay, can describe the degradation of
single compounds or complex mixtures of compounds in the subsurface environment by a
complex mixed microbial community. A major factor affecting microbial growth and decay in
many environments (e.g., aquifer sediments) is the presence of solid surfaces. Surfaces may
affect the bioavailability of organic chemicals, change the concentration of various organic and
inorganic nutrients, or immobilize microbial enzymes or microorganisms. Bacterial cells
attached to subsurface materials may have physiological activities quite different from those of
cells in suspension. It has generally been concluded that the Monod equations represent
reasonable models with which to describe a range of kinetic values describing bacterial growth
in natural environments and are widely used (Lawrence and McCarty, 1970).
38
-------
Substrate utilization curves were fit to a number of kinetic models that describe the
utilization of the major WSF compounds. Thermodynamic analyses were used to help explain
underlying causes of the degradation pathways and sequences observed.
Microbial Growth
The growth of a microbial population is a complex phenomenon composed of a number
of simultaneously occurring events. Relative magnitudes of the respective rates determine the
net effect upon the population. The primary events are the utilization of substrate and the
concurrent growth of the organisms. These two events are closely related because it is only
through the utilization of substrate that energy and carbon are made available for cell growth.
The cells must also use energy for maintenance and if no exogenous energy source is present,
maintenance energy will be provided by energy reserves and the cell mass in the culture will
decline due to death of the cells. From the macroscopic point of view this leads to a decrease
in the total mass of the culture.
Bacteria divide by binary fission and consequently the number of viable cells will increase
in an exponential fashion. The reaction rate for bacterial growth can be expressed as a
first-order equation
(5-1)
where Xa is the concentration of active bacteria (mg/L), and ji is the specific growth rate
(day1). The growth rate constant is referred to as a specific growth rate because it defines
the rate of cell growth in terms of the concentration of cells present.
The growth yield, Y (mg/mg), is defined as the ratio of the rate of cell growth to the rate
of substrate, 5 (mg/L), removal. This can be expressed mathematically as
-* rs 9i
~dt~YXa (5~2)
Consequently it can be seen that the rate of substrate removal is first-order with respect
to the concentration of viable cells.
39
-------
The effect of the concentration of growth limiting nutrient upon the specific growth rate
is a function of the initial substrate concentration and (i increases as the substrate
concentration is increased until it approaches some maximum value (\imax)-
The question of the best mathematical formula to express the relationship of the growth
rate and substrate concentration has been a subject of much debate. No one yet knows
enough about the mechanisms of microbial growth to propose a mechanistic equation which
will characterize growth exactly. Instead, experimenters have observed the effects of various
factors upon growth and have attempted to fit empirical mathematical equations to their
observations. The equation with historical precedence and greatest acceptance is the one
proposed by Monod (1949)
where Ks (mg/L) is the saturation constant. Ks determines how rapidly [I approaches
is defined as the substrate concentration at which (I is equal to half of M-
The above relationships can be combined into an expression that relates substrate
utilization to bacterial growth.
dt Y(KS+S)
The Monod equation was developed from experiments using pure cultures of bacteria
growing on single organic compounds. When growth of bacteria in environmental situations
is considered, many complicating factors enter the picture. Two very important factors are:
(1) the natural environment does not contain single organic compounds and (2) the natural
environment contains complex bacterial communities rather than single species. These
communities are usually in a continuous state of flux with constant changes in relative
magnitudes of the species present. This can have a drastic effect upon observed kinetics so
that the growth "constants" are seldom constant (Kompala et al., 1986).
Many investigators have studied the relationship between \L and S in mixed populations in
order to ascertain whether it can be represented by the Monod equation (e.g. Lawrence and
McCarty, 1970). It has generally been concluded that the Monod equation represents a
40
-------
reasonable model with which to describe a range of kinetic values describing bacterial growth
in natural environments, and it is widely used.
The term representing the loss of cell mass is usually called bacterial decay and can be
expressed as follows
(5-5)
where kj (day-1) is a first-order bacterial decay constant.
These expressions can be combined into one equation that relates bacterial growth and
decay to substrate utilization.
* KS+S
Equations 5^4 and 5-6 are general expressions for a system in which only population
densities and substrate determine the kinetics of degradation. When the inoculum density is
extreme compared to the substrate concentration or, when substrate concentration is very low
or very high with respect to Ks, the Monod equation may be simplified to 5 (Table 2-1)
special forms or Monod family of equations.
When the initial bacterial density is much greater than the number of new bacteria which
would be produced from the substrate at time zero, i.e., X0 » S0, the growth of population
during the course of the experiment is insignificant. Similarly, when the initial substrate
concentration is much greater than Kg, i.e., S0 » Ks, most of the substrate will disappear
while the enzyme uptake systems of the cells are saturated, and Ks + S in the above equation
can be approximated by only 5. Alternatively, the substrate is initially present at much less
than saturating levels (S0 « Ks), the value of Ks + S can be approximated by only Ks. These
simplifications allows the modeling of the variety of shapes of substrate disappearance curves
with only substrate concentration and bacterial density and the parameters of Monod kinetics.
A deterministic 3/2-order kinetic model was proposed by Brunner and Focht (1984) to
describe the mineralization of simple or complex substrates in soils with linear or exponential
bacterial growth. The model is based on a first-order approximation of the Monod substrate
utilization equation with an additional term to describe either linear or exponential growth
41
-------
Linear
dS
—— = k3S+k4St (5-7)
at
Exponential
(5-8)
at H
where
£3 = proportionality constant, day1
£4 = linear growth rate constant, day2
Xp= bacterial biomass multiplied by a proportionality constant p, mg/L
5 = substrate concentration at time t, mg/L
H = exponential growth rate constant, day1
t = time.
These models require no a priori assignment of a lag phase, contain only two
interdependent constants as opposed to four for the Monod equation, and are suitable for
substrate metabolism modeling with or without exponential growth.
The typical groundwater environment has a high surface area to which bacteria can attach
(Holm et al., 1992). This results in higher concentrations of bacteria on the surface of the
aquifer sediments when compared to the surrounding pore water. The low concentration of
the attached bacteria at ACW suggest that the biomass may exist as microcolonies similar to
those described by Harvey et al., (1984) for a contaminated aquifer at Cape Cod, Mass.,
rather than as a continuous biofilm. The low nutrient, low biomass conditions suggests that
the kinetics of substrate removal may best be modeled by static anaerobic microcosms filled
with aquifer material. The apparently slow removal rates observed in the field suggest that the
developing microcolonies in the microcosms would all be exposed to the same concentrations
of organic compound(s) with only an occasional gentle mix. This would allow the substrate
utilization to be treated in the same manner as for dispersed growth using the Monod equation
for a batch microcosm as described by equations 5-4 and 5-6.
The physiology and metabolism of bacterial populations exposed to low substrate
concentrations such as found in the subsurface environment are quite different from those
grown in high nutrient environments (e.g., sewage sludge). Matin and Veldkamp (1978) have
documented many differences in bacterial populations from these environments:
42
-------
• Decreased levels of ribosomal RNA,
• Mixed substrate utilization,
• More efficient enzymes and transport systems,
• Lower maintenance energy,
• Population shifts to more efficient species,
• Increased surface to volume ratio, and
• A tendency to attach to surfaces.
RESULTS
Biodegradation Kinetics of the Complex WSF Mixture
The biodegradation kinetics of the WSF as TOC, calculated from the summation of the
identifiable organic compounds in ground water samples from Site 3 converted to mg/L of
carbon, were determined for the Monod family of equations and a deterministic 3/2-order
model with linear and/or exponential growth (Brunner and Focht, 1984). The model fits are
shown in Figures 5-1 and 5-2 and given in Table 5-1 are the RSS and the LBSSB
(Likelihood Based Sum of Squares Boundaries) values for each determination. Degradation
of organic compounds was not observed in the killed cell (autoclaved) control and,
consequently, is not shown in Figures. In NLR, the assumption is made that the model being
fitted is the correct one, and that the observations deviate from the model in a random fashion.
These values may be used internally to give an estimate of how well the data fit the model
prediction. That family of substrate utilization curves generated from parameter values within
the 95 percent confidence interval parameter that yields a RSS < LBSSB will form
boundaries around the predicted model curve. This is roughly equivalent to a 95 percent
confidence interval for the substrate utilization curves and is based on the F-statistic at the
0.05 level of significance (Bates and Watts, 1988). The Monod first-order equation (&j =
1.01 ± 0.12 x 10'2 day1) gives a reasonable fit to the degradation of the TOC; however, the
3/2-order models provide a better fit, especially with exponential growth. The rate constants
(Chapter 2) for linear growth are £3 = 5.38 ± 1.91 x 10'3 day1, and £4 = 8.65 ± 3.8 x 10'5
day2, while the exponential growth rate constants are £3 = 4.43 ± 280 x 10-4 day1, \i = 8.38
± 22.1 x 10'3 day1, and Xp = 5.77 ± 25.6 x 10'3 mg/L. The range of values for kt must fall in
the range of 1.13 x 10'2 to 0.90 x 10'2 day1 to fall in the LBSSB. This range of values is
essentially the same as the 95 percent confidence determined by the NLR analysis. The other
Monod family of equations either can not be fit to the data at all, or may be dismissed on other
considerations. The fit using the Monod with decay equations require that growth equal
43
-------
decay, an unlikely situation in the microcosms, and the equations for Monod kinetics without
growth are also unlikely since growth does occur when single compounds are tested.
Biodegradation Kinetics of Phenolic Compounds
Monod with decay parameter determinations for phenol, 2-, 3-, and 4-methylphenol
degradation with the 95 percent confidence intervals are given in Table 5-2. The values for
the growth yield (Y) for each compound were determined from protein determinations before
and after substrate utilization in the microcosms. The substrate disappearance curves with the
model predictions are shown in Figures 5-3 to 5-6. The time interval before the onset of
rapid substrate utilization and methanogenesis varied from 28 days for 3-methylphenol to 119
days for 2-methylphenol, even though the inoculum history suggests that the microbial
community in the microcosms had been exposed to all of the phenolic compounds for
considerable length of time (~ 80 years). Degradation of organic compounds was not
observed in the killed cell (autoclaved) controls and, consequently, are not shown in Figures.
The viable-cell organic-free control did not produce any detectable amounts of CH4 and/or
CC>2, and is also not shown.
Figure 5-7 shows the effect of Xao concentration, the fitted starting biomass
concentration, on the substrate utilization curve for the phenol microcosm. Increasing or
decreasing the value of Xao only displaces the curve to the left or right without changing the
shape of the curve. Furthermore, if the value of the ratio of Xao / Y remains constant, the
same substrate utilization curve will be generated regardless of the value of X^ or Y. It is
necessary, therefore, to independently determine the value of Y. Fitting the value of X^
alleviates the problem of arbitrarily picking the point at which a lag or adaptation period ends
and degradation starts. This is justified for this study because the microbial community used
for the inoculum for all of the microcosms have been exposed to these compounds for a
considerable length of time in the aquifer and should not have an adaptation or lag time.
The values for S0, Xa, and Xt (total measured biomass) for each of the microcosms are
given in Table 5-3. The value of Xt in the microcosms only slightly increased during the
course of the experiment and the methanogenic bacteria increased by approximately 1.5 orders
of magnitude; however, the methanogenic bacteria only account for approximately 0.1 percent
of the total bacterial population. Total facultative aerobes, denitrifying, SRB, and methane
producing bacteria bacterial numbers are also given in Tables 5-4 to 5-7, respectively. The
total aerobic population increased by approximately an order of magnitude, while the SRB
and denitrifying bacteria approximately doubled in numbers. The methanogenic bacterial
44
-------
population increased by almost two orders of magnitude; however, they are only a small
percentage of the total population.
Mass balances of phenolic compound utilization with CFLj and CO2 production were
determined and based on the equation proposed by Tarvin & Buswell (1934). The measured
production of CH4 and CC>2 for each of the compounds were compared to the theoretical gas
yields calculated from the equation and are shown in Figures 5-8 to 5-11.
The mass balances on the degradation of the phenolic compounds were corrected for
substrate removed for HPLC analyses. Total gas values were computed from the CHj and
CC>2 values and yielded 95.3 percent of theoretical total gas production for phenol, 85.4
percent for 2-methylphenol, 90.9 percent for 3-methylphenol, and 92.2 percent for
4-methylphenol. These values are well within acceptable and expected ranges.
Given in Table 5-8 are the RSS and the LBSSB values for degradation of each of the
phenolic compounds tested. In NLR, the assumption is made that the model being fitted
(Monod with decay) is the correct one, and that the observations deviate from the model in a
random fashion. These values may be used internally to give an estimate of how well the data
fit the model prediction. The high RSS value for 2-methylphenol reflects the inability of the
model to account for the shoulder at the onset of rapid substrate utilization and
methanogenesis (~ day 105). The RSS value determined as the LBSSB for phenol, 102.8,
would displace the model curve by ± 3-4 days along the time axis. The boundaries for
2-methylphenol would likewise be displaced by ± 6-7 days.
Biodegradation Kinetics of NSO-Compounds
The substrate conversion curves for the oxidation of the nitrogen heterocyclic compounds
indole, quinoline, and isoquinoline to oxindole, 2(lH)-quinolinone, and
l(2H)-isoquinolinone, respectively are shown in Figures 5-12 through 5-14. These
compounds re initially stoichiometrically oxidized. The oxidized intermediates sometimes
persisted for very long (up to 284 days after the parent compound was completely oxidized),
variable, and unreproducible times from the initial oxidation of the parent compound to the
mineralization to CH4 and CO2. The oxidation of these compounds could be fit to only the
Monod-no growth equation (Table 5-9) after an arbitrary lag time was assigned and could
not be fit to Monod equations with or without decay. There are fairly large errors associated
with the individual parameters making it difficult to compare the parameters for each
compound. It is not clear at this time whether the organism(s) that are responsible for the
45
-------
oxidation of these compounds are later involved in the methanogenic degradation of the
oxidized compound.
Growth parameter estimates for the Monod-decay equations for benzothiophene,
oxindole, 2(lH)-quinolinone, and l(2H)-isoquinolinone with decay are given in Table 5-10.
It is of interest to note that the time interval before the onset of rapid methanogenesis varied
considerably as did the time interval for the phenolic compounds and the values for the
individual parameters are almost identical to the parameter values obtained for the phenolic
compounds (Table 5-2).
Substrate utilization curves and Monod-decay model fit for the methanogenic
biodegradation of benzothiophene, oxindole, 2(lH)-quinolinone, and l(2H)-isoquinolinone
are given in Figures 5-15 through 5-18, respectively. Degradation of organic compounds
was not observed in the autoclaved controls and, consequently, are not shown in the Figures.
The viable control with mineral salts solution minus the organic compounds did not produce
detectable amounts of CH4 and/or CO2, and is also not shown. The values for RSS and
LBSSB for the degradation of the heterocyclic compounds are given in Table 8-11. The
initial starting value for S0 and Xao along with the initial and final measured biomass are given
in Table 8-12. The initial and final concentrations for only the methanogenic bacteria are
given in Table 8-13.
The mass balances on the degradation of the above compounds were based on the
equation proposed by Tarvin and Buswell (1934) and are shown in Figures 5-19 through
5-22. The yields were 86.4% of theoretical gas production for benzothiophene, 87.6% for
oxindole, 91.7% for 2(lH)-quinolinone, and 88.5% for l(2H)-isoquinolinone.
DISCUSSION AND CONCLUSIONS
The lack of success of the Monod equations in describing the degradation of TOC in
laboratory microcosms suggests that microbial degradation of complex difficult to degrade
organic mixtures by complex consortia of microorganisms may not be amenable to description
using a simple parameter (e.g., BOD , COD, TOC, etc.). The first-order approximation was
the only Monod family equation that gave a reasonable fit to the data; however, the
assumption of the first-order approximation is that X0 » S0 and S0 « Ks, clearly not the
case in this situation. The Monod-decay equation requires that microbial decay be equal to
growth - a situation that is not realistic for the laboratory microcosms. Other single
compound degradations showed that decay was insignificant in the time frame of the
46
-------
microcosm growth experiments. Monod-no growth equations can be dismissed on the basis
that the individual compounds support growth when converted to Cl-fy and CC^. Monod and
logistic equations require a negative sigmoidal curve and, thus, could not be fit to the TOC
utilization. Logarithmic kinetics could not be fit at all.
The deterministic 3/2-order kinetic models proposed by Brunner and Focht (1984) to
describe the mineralization of simple or complex substrates in soils with linear or exponential
bacterial growth were quite successful in describing the utilization of the TOC. These models
are based on a first-order approximation of the Monod substrate utilization equation with an
additional term to describe either linear or exponential growth. These models require no a
priori assignment of a lag phase, contain only two interdependent constants as opposed to
four for the Monod equation, and are suitable for substrate metabolism modeling with or
without exponential growth.
Laboratory microcosms containing aquifer material used in this study attempt to simulate,
with as little change as possible, the biotic and abiotic interactions that occur in the subsurface
at the Pensacola study site. Most other kinetic studies utilize microbial cultures that have
been adapted to a particular substrate by enrichment or continuous culture techniques. Often
times the microbial populations undergo significant changes during these procedures (Mackey,
1987). Populations that are not required for the degradation of complex organics, but affect
the rate at which the organics are degraded, may be selectively removed from the consortium.
Dills et al. (1980) found evidence for the presence of multiple uptake systems for glucose in
marine microorganisms. Presumably, over several generations of growth at different substrate
concentrations, progeny cells could be enriched for a transport system that is not in use under
natural conditions, again altering the kinetic constants. If the Monod equations are accepted
as viable models, the task is really one of estimating (a.^^ and Ks, as k^ is relatively easy to
evaluate and Yean, generally, be independently determined. Templeton and Grady (1988)
demonstrated that Umax is often overestimated and Ks is often underestimated when bacterial
populations are enriched by continuous culture or fed-batch techniques. The primary
significance of this phenomenon to environmental engineers concerned with the fate of
organic compounds in the environment is the recognition that kinetic analyses will be
influenced by the history of the culture used, and it is for these reasons, that no attempts were
made to enrich or adapt the ground-water cultures to the specific compounds tested.
The sorption of substrate and biomass to the aquifer sediment are important
considerations in determining the ultimate environmental fate of contaminants. This study has
shown that sorption of the biodegradable organic compounds to be negligible. For modeling
47
-------
purposes in this study the biomass attached to the sediment can be treated as if it were
uniformly distributed throughout the liquid volume. The experimental design allows the
determination of the growth kinetics at substrate concentrations similar to the study site and
under conditions as close as possible to natural conditions.
An adaptation period can be considered the time required for the adjustment of a bacterial
population to a new environment. The bacterial cells may be taking in substrate, they may be
synthesizing new enzymes, and they may be undergoing enlargement prior to division, but,
they have not begun an orderly and steady replication. The time required for the adaptation of
the population varies widely and is generally not predictable. The organisms from the centroid
of the contaminated aquifer will have seen low concentrations of the WSF compounds tested
for a considerable length of time and transfer to a microcosm that duplicates the aquifer as
much as possible should not alter the environmental conditions. As a consequence, the time
required for the onset of methanogenesis is presumably the time required for the populations
of the organisms responsible for the initial attack on the WSF compounds to increase to a
level that is sufficient for rapid degradation of these compounds.
Kinetic constants for the anaerobic degradation of the WSF compounds outside of this
work have not been reported in the literature, with the exception of phenol in sewage sludge
derived microcosms or columns. Given in Table 8-14 are the values published by Neufeld et
al. (1980) for unspecified anaerobic conditions and for methanogenic conditions by Suidan et
al. (1989). When anaerobic domestic sewage sludge is used as the inoculum, large quantities
of organic compounds are presumably available for microbial utilization. The energy available
from phenol under methanogenic conditions shown below is small and must be shared by a
number of bacterial populations
) + 4 H2O -> 3.5 CH4 + 2.5 CO2 AG°' = -106 kJ/mol
and would require a relatively high substrate concentration before a microbial population
would initiate the degradation of this compound when other, more energetically favorable and
readily biodegradable, compounds are available. This effect may explain the high Ks value
(686 mg/L) reported by Neufeld et al. (1980). Several other researchers, studying the
methanogenic degradation of phenol with sewage sludge cultures, did not determine kinetic
constants (Healy and Young, 1978; Young and Rivera, 1985); however, the approximate Ks
values can be determined from the substrate utilization progress curves and appear to be
slightly less than 50 mg/L. In this study and the study by Suidan et al. (1989), cultures were
used that had been exposed only to phenol and similar compounds for long periods of time.
48
-------
The low Ks values obtained in these studies (1.33 and 0.03 mg/L, respectively) demonstrate
that the responsible enzyme systems have developed a high affinity for phenol. This is also
generally true for all of the WSF compounds tested.
Low values for Y for the WSF compounds were obtained in this study compared to
theoretical values calculated in Appendix D based on thermodynamic principles (McCarty,
1971). The determined and theoretical values of Y for the compounds tested are given in
Table 5-15. The theoretical values are greater than the value obtained from the protein
increase in the microcosms by approximately an order of magnitude. However, the error
associated with the protein analyses is as great or greater than the difference in starting and
final biomass as determined by protein analyses (Appendix A, Table A-2). Analyses for
biomass on a sediment sample collected at a different time but in the same manner and the
same general area of Site 39, but a depth of 2 m below land surface, had an average biomass
concentration of 57.6 mg/L with a large value for the 95 percent confidence interval of ± 15.6
mg/L. With such a large confidence interval for this determination, it is impossible to
determine if the increases in biomass given in Tables 5-3 and 5-12 are accurate, and
consequently, whether or not the determined values of Y are significantly different from the
theoretical values of Y. However, in the case of phenol, with a starting biomass concentration
of 28 mg/L, a Y of 0.01 mg/mg, and an initial substrate concentration of 41 mg/L phenol, the
biomass would increase by 0.28 mg/L or 1.4 x 109 cells/L (assuming 2 x 10-1° mg/cell) in the
microcosm during the experiment. This cell concentration is in better agreement with all of
the biomass determinations (i.e., AODC, MPN) performed during other portions of the study
on microcosms and aquifer material samples than yields that would increase the biomass by an
order of magnitude. It will remain for a more sensitive technique of biomass determination
(e.g., adenosine triphosphate, muramic acid, or phospholipids) before it can be determined if
the yields are indeed low.
The apparent low biomass yield suggests that the microbial community from this
oligotrophic ground water environment has adapted to these conditions by utilizing a major
portion of the available energy for maintaining cellular integrity in a relatively hostile
environment (Battley, 1987). It seems unlikely that the other possible explanations, i.e.,
inefficiency at capturing the free energy available, wasting of energy, or storing carbon as
intracellular storage products, would account for the low Y values. These Y values are also
consistent with the low biomass on the aquifer sediments and the high dissolved CH4
concentrations (60-70 percent of saturation) throughout the contaminated field site.
49
-------
The bacterial decay term (-Jt^) in the biomass equation (5-6) is apparently not required
to describe substrate utilization and/or biomass increase in the batch growth microcosms. The
values determined by NLR are such that jo.^^ » k^, and as a result k^ can be neglected.
NLR analyses of WSF compound disappearance data using Monod equations without the
decay term resulted in essentially the same kinetic constants being generated. Using the
Monod equations without decay should alleviate the problem of increased uncertainty
associated with fitting three parameters versus two parameters. This is not meant to imply
that in a continuous culture microcosm or in a field situation, the decay term is not important
and/or necessary to describe the kinetics of utilization.
The protein content of bacteria depends on the type of organisms, on their growth state,
and it is assumed that most bacteria have a protein content of 50 percent of their dry weight
(Nester et al., 1983). An assumption is made that during the course of the determination, all
of the biomass on the aquifer sediment or in solution is solubilized for subsequent assay. Both
of these assumptions were verified in Appendix A; however, the error associated with these
determinations may lead to erroneous biomass values. It is for these reasons that only the
substrate utilization curves were used for the kinetic analyses.
The number of methanogenic bacteria appears to be low for a site contaminated with
organic compounds; these numbers are virtually the same as those detected by Belyaev and
Ivanov (1983) in aquifer sediments associated with gas and oil deposits. However, the
microbial community, especially the methanogenic population, appears to be quite active as
demonstrated by the relatively rapid use of WSF compounds in the microcosms once
degradation started. The sulfate reducing bacteria increased by a factor of two in the phenolic
compound microcosms, but it is uncertain at this time what the role of these organisms in this
system is; however, these organisms are most likely growing on volatile fatty acids in a
syntrophic association with the methane producing bacteria (Bryant et al., 1911 \ Puhakka et
al., 1990)
The bacterial substrate utilization for all of the compounds tested was modeled
successfully using the Monod equations with the exception of indole, quinoline, and
isoquinoline. These compounds were all oxidized to intermediate compounds that could
persist for many days after the initial oxidation. This was most obvious during down gradient
transport in the aquifer at the study site where quinoline and isoquinoline were oxidized at the
Site 3, the first observation site, to 2(lH)-<}uinolinone and l(2H)-isoquinolinone. This was
one of the first detectable transformation reactions to take place.
50
-------
The oxidized intermediates actually increased with respect to the tracer indicating that
this reaction was indeed one of the first to take place and continued during down gradient
transport without degradation of the intermediates. These compounds were also the last
compounds to be degraded at the site. In the large WSF microcosms (Gosy et a/., 1992),
quinoline and isoquinoline were rapidly oxidized to 2(lH)-quinolinone and
l(2H)-isoquinolinone, but persisted while other compounds were degraded and were the last
compounds to be degraded to CH4 and CO^ Indole behaved in a similar manner, the parent
compound was first oxidized to oxindole which sometimes persisted for very long and variable
times. However, these compounds were not found at the same concentrations as the other
nitrogen heterocycles studied. In fact, at Site 3, only trace amounts (<, 0.010 mg/L) of indole,
oxindole, and several of the other possible intermediates were identified in ground-water
samples. Indole is a major compound in creosote (Novotony et al., 1981), but apparently is
one of the first compounds to undergo degradation at the site.
51
-------
I
55
O
P
5
w
g
O
O
DAYS
O TOC
--- MONOD DECAY
MONOD NO GROWTH
FIRST ORDER
Figure 5-1. Monod family of substrate disappearance curves that gave the best fit to
the observed TOC degradation.
52
-------
50
O TOC
100 150
DAYS
- LINEAR
200
250
EXPONENTIAL
Figure 5-2. Deterministic three-half-order kinetic model fits to the observed TOC
degradation.
53
-------
50
40
§ 30
g
o
20
10
0
100
DAYS
MODEL O
150
200
PHENOL
Figure 5-3. Phenol utilization compared to the Monod with decay model prediction.
All values are averages of two determinations.
54
-------
50
MODEL
100 150
DAYS
O 2-METHYLPHENOL
200
Figure 5-4. 2-Methylphenol utilization compared to the Monod with decay model
prediction. All values are averages of two determinations.
55
-------
100
DAYS
150
200
MODEL
O 3-METHYLPHENOL
Figure 5-5. 3-Methylphenol utilization compared to the Monod Decay model
prediction. All values are averages of two determinations.
56
-------
I
g~
e
8
§
O
10
100
DAYS
O 4-METHYLPHENOL
Figure 5-6. 4-Methylphenol utilization compared to the Monod Decay model
prediction. All values are averages of two determinations.
57
-------
DAYS
0.0000005 0.00005 0.005
0.000005 —• 0.0005
Figure 5-7. Effect of the starting biomass concentration (mg/L) on the displacement
of the phenol utilization curve based on model predictions.
58
-------
§
p
11
^^ •^c
Q o
,
/
PHENOL
100
DAYS
MODEL
SUBSTRATE
TOTAL GAS
METHANE
CARBON DIOXIDE
150
DATA
TOTAL GAS
METHANE
CARBON DIOXIDE
200
Figure 5-8. Concentrations of methane, carbon dioxide, and total gas producded in the
phenol microcosm compared to theoretical yields.
59
-------
2-METHYLPHENOL
MODEL
SUBSTRATE
TOTAL GAS
METHANE
CARBON DIOXIDE
a
••.
DATA
TOTAL GAS
METHANE
CARBON DIOXIDE
Figure 5-9. Concentrations of methane, carbon dioxide, and total gas producded in the
2-methylphenol microcosm compared to theoretical yields.
60
-------
I
p^ ^^
s Q
UJ
U
§H
O ^
B
0.4
0.3
0
bo
3-METHYLPHENOL
50
MODEL
SUBSTRATE
TOTAL GAS
METHANE
CARBON DIOXIDE
100
DAYS
150
DATA
TOTAL GAS
METHANE
CARBON DIOXIDE
200
Figure 5-10. Concentrations of methane, carbon dioxide, and total gas producded in
the 3-methylphenol microcosm compared to theoretical yields.
61
-------
U.4-
P 0.3-
g 0
§ S 0.2-
O °
0 'o
Q §
o.i-
O
o
()•*
1
4-METHYLPHENOL
"~""""*v^ /
X/*
_J$C
i 50 100 150 2(
DAYS
MODEL DATA
SUBSTRATE
TOTAL GAS ° TOTAL GAS
METHANE D METHANE
<~i A DT)/^\T TMYYVrTYI? A /-( l TITI/-»>T -nif\vrr\Ti
}(}
Figure 5-11. Concentrations of methane, carbon dioxide, and total gas producded in
the 4-methylphenol microcosm compared to theoretical yields.
62
-------
MODEL
O 1NDOLE
40
DAYS
MODEL
60
80
A OXINDOLE
Figure 5-12. The stoichiometric oxidation of indole to oxindole after a lag time of 8
days in laboratory microcosms. Indole oxidation was fit to the Monod without growth
equations. All values are the average of two determinations.
63
-------
MODEL
O QUINOLINE
A 2(1H)-QUINOLINONE
Figure 5-13. The stoichiometric oxidation of quinoline to 2(lH)-quinolinone after a
lag time of 27 days in laboratory microcosms. Quinoline oxidation was fit to the
Monod without growth equations. All values are the average of two determinations.
64
-------
0.4'
0
MODEL
O ISOQUINOLINE
MODEL
A l(2H)-ISOQUINOLINONfE
Figure 5-14. The stoichiometric oxidation of isoquinoline to l(2H)-isoquinolinone
after a lag time of 30 days in laboratory microcoms. Isoquinoline oxidation was fit to
the Monod No Growth equations. All values are the average of two determinations.
65
-------
g
o
o
30
25
20-
15-
50
DAYS
75
100
0 BENZOTHIOPHENE
Figure 5-15. Benzothiophene utilization compared to the Monod with decay model
prediction. All values are averages of two determinations.
66
-------
50 75
DAYS
MODEL O OXINDOLE
Figure 5-16. Oxindole utilization compared to the Monod with decay model
prediction. All values are averages of two determinations.
100
67
-------
s
§
£
o
^
o
o
MODEL
50
DAYS
O 2(1H)-QUINOLINONE
Figure 5-17. 2(l)-Quinolinone utilization compared to the Monod with decay model
prediction. All values are averages of two determinations.
68
-------
MODEL
O 1(2H)-ISOQUINOLINONE
Figure 5-18. l(2H)-Isoquinolinone utilization compared to the Monod with decay
model prediction. All values are averages of two determinations.
69
-------
u.i-
0.08-
1
§ § 0-06-
1 2
Q | 0.04-
0
| 0.02-
0
o
0,
11
1
BENZOTHIOPHENE
^\ //;
VV^X*"
^^^S^
» 10 20 30 40 51
DAYS
MODEL DATA
SUBSTRATE
TOTAL GAS ° TOTAL GAS
METHANE ° METHANE
f~* A T3T>/^XT T\T/^VTT\T7'
Figure 5-19. Concentrations of methane, carbon dioxide, and total gas produced in
the benzothiophene microcosm compared to theoretical yields.
70
-------
DAYS
MODEL
SUBSTRATE
TOTAL GAS
METHANE
CAKBON DIOXIDE
DATA
TOTAL GAS
METHANE
CARBON DIOXIDE
Figure 5-20. Concentrations of methane, carbon dioxide, and total gas produced in the
oxindole microcosm compared to theoretical yields.
71
-------
o
Q
8
2(1H)-QUINOLINONE
40 60
DAYS
MODEL DATA
SUBSTRATE
TOTAL GAS ° TOTAL GAS
METHANE D METHANE
CARBON DIOXIDE A CARBON DIOXIDE
Figure 5-21. Concentrations of methane, carbon dioxide, and total gas produced in
the 2(lH)-quinolinone microcosm compared to theoretical yields.
72
-------
0.12
1(2H)-ISOQUINOLINONE
DAYS
MODEL
SUBSTRATE
TOTAL GAS
METHANE
CARBON DIOXIDE
DATA
0 TOTAL GAS
n METHANE
A CARBON DIOXIDE
Figure 5-22. Concentrations of methane, carbon dioxide, and total gas produced in
the l(2H)-isoquinolinone microcosm compared to theoretical yields.
73
-------
Table 5-1. RSS and the LBSSB of the model fits for the biodegradation of the TOC.
Model
RSS
LBSSB
First-order
3/2-Linear
3/2-Exponential
5954
2170
2040
7179
3242
3048
Table 5-2. Kinetic constants determined for each of the phenolic compounds tested
± 95% confidence intervals using determined values for Y.
Compound
Phenol
2-Methylphenol
3-Methylphenol
4-Methylphenol
V-max (day1)
0.11110.005
0.044 ±0.001
0.103 + 0.078
0.099 ±0.1 10
Mmg/L)
1.33 + 0.07
0.25 ±0.82
0.55 ± 6.67
3.34±1U
*rf(dayl)
0.001+0.012
0.002 ± 0.008
0.000 + 0.019
0.000 + 0.032
rCmg/mg)1
0.013
0.022
0.026
0.025
1 Y values determined from protein determinations before and after substrate utilization
Table 8-3. Initial substrate and biomass concentration and changes in measured biomass
during substrate utilization in the microcosms. All values given as mg/L.
Compound
Phenol
2-Methylphenol
3-Methylphenol
4-Methylphenol
S0
41.0
36.0
34.0
24.5
v 1
Aao
0.00033
0.00190
0.00033
0.00008
Initial biomass
*,„
27.85
33.98
27.97
29.34
Final biomass
X*
28.39
34.78
28.85
29.96
1
Fitted active biomass
74
-------
Table 5-4. Initial and final concentrations of facultative aerobic bacteria in the
microcosms for the phenolic compounds tested. All values given as MPN/L.
Compound Initial Final
Phenol 3.6 x 106 1.2 x 107
2-Methylphenol 4.4 x 106 9.3 x 106
3-Methylphenol 3.6 x 106 9.4 x 106
4-Methylphenol 3.8 x 106 1.1 x 107
Table 5-5. Initial and final concentrations of denitrifying bacteria in the microcosms for
the phenolic compounds tested. All values given as MPN/L.
Compound Initial Final
Phenol 5.5 x 103 6.3 x 103
2-Methylphenol 6.4 x 103 8.7 x 103
3-Methylphenol 5.5 x 103 6.9 x 103
4-Methylphenol 5.6 x 103 9.0 x 103
Table 5-6. Initial and final concentrations of SRB in the microcosms for the phenolic
compounds tested. All values given as MPN/L.
Compound Initial Final
Phenol 3.5x105 1.2X106
2-Methylphenol 4.0 x 105 7.3 x 105
3-Methylphenol 3.4 x 105 9.4 x 105
4-Methylphenol 3.5 x 105 1.1 x 106
75
-------
Table 5-7. Initial and final concentrations of methanogenic bacteria in the microcosmsfor
the phenolic compounds tested. All values given as MPN/L.
Compound Initial Final
Phenol 3.1 x 103 1.6 x 105
2-Methylphenol 3.6 x 103 5.6 x 104
3-Methylphenol 3.1 x 103 2.7 x 105
4-Methylphenol 3.1 x 103 2.4 x 105
Table 5-8. RSS and the LBSSB for the Monod with decay equations for each of the
phenolic compounds tested.
Compound RSS LBSSB
Phenol 62.5 102.8
2-Methylphenol 334.9 459.5
3-Methylphenol 130.9 229.4
4-Methylphenol 109.9 140.2
Table 5-9. Kinetic constants, RSS, and the LBSSB determined for the oxidation of the
nitrogen heterocyclic compounds tested + 95% confidence interval for each parameter.
Compound kn (mg/L-day) Ks(mg/L) RSS LBSSB
Indole 47.2±335 175±1351 8.7 16.9
Quinoline 5.2±0.38 2.011.4 95.7 142.6
Isoquinoline 11.5±59.1 59.9± 355 7/7 10.6
76
-------
Table 5-10. Kinetic constants determined for the methanogenesis of the heterocyclic
compounds tested ± 95% confidence intervals. Values for Y were determined from
protein analyses.
Compound
Benzothiophene
Oxindole
2(lH)-quinolinone
1 (2H)-isoquinolinone
*W(day->>
0.11710.136
0.16010.002
0.08910.058
0.09910.342
Ks(mg/L)
0.80 1 5.06
1.1012.10
11.4110.06
5.00122.00
kd (day*)
0.00010.129
0.00010.012
0.00010.049
0.00010.027
r(mgtag)>
0.025
0.029
0.033
0.035
Y values determined from protein determinations before and after substrate utilization
Table 5-11. RSS and the LBSSB for the Monod with decay equations for each of the
heterocyclic compounds tested.
Compound RSS LBSSB
Benzothiophene 2.0 10.3
Oxindole 19.8 44.0
2(lH)-quinolmone 157.0 226.3
l(2H>-isoquinolinone 8J 13.3
Table 5-12. Initial substrate and biomass concentration and changes in measured biomass
during substrate utilization in the microcosms. All values given as mg/L.
Compound S0 X^l Initial biomass Final biomass Xtj
Benzothiophene
Oxindole
2( lH)-quinolinone
1 (2H)-isoquinolinone
10.0
31.5
18.0
29.0
0.0028
0.0041
0.0033
0.0018
33.74
30.96
35.47
32.58
34.58
31.85
37.67
33.72
Fitted active biomass to account for the delay before the onset of rapid methanogenesis
77
-------
Table 5-13. Initial and final concentrations of methanogenic bacteria for the heterocyclic
compounds tested. All values given as MPN/L.
Compound
Initial
Final
Benzothiophene
Oxindole
2(lH)-quinolinone
l(2H)-isoquinolinone
2.8 x 103
3.5 x 103
3.0 x 103
2.9 x 103
9.8 x 105
7.4 x 104
1.4 x 10s
1.9 x 10s
Table 5-14. Comparison of published kinetic constants for the anaerobic degradation of
phenol in sewage sludge and the parameters obtained in this study.
\ifnax (day1) Ks(mg/L) Y (mg/mg) ^(day1)
Aquifer microcosm,
this study, Methanogenic
Neufeld et al. (1980),
Anaerobic
Suidanetal.(1989),
Methanogenic
0.111
0.08
0.106
1.33
686
0.03
0.013
0.82
0.16
0.001
0.008
0.192
Table 5-15. Determined and theoretical Y (mg cells/mg compound) values for the WSF
compounds tested in microcosms.
Determined Y (mg/mg) value l Theoretical Y (mg/mg) value
Phenol
2-Methylphenol
3-Methylphenol
4-Methylphenol
Benzothiophene
Oxindole
2( lH)-quinolinone
1 (2H)-isoquinoli none
0.013
0.022
0.026
0.025
0.025
0.029
0.033
0.035
0.08
0.12
0.12
0.14
0.14
0.35
0.21
0.21
1 Based on the average of three protein determinations before and after growth.
78
-------
Table 5-16. Gibbs free energy changes during oxidation of NSO compounds at 25*C and
pH7.
Compound AG(aq)
Indole to Oxindole
C8H7N + H2O -> C8H7NO + 2 H+ + 2 e~ 1 68
Quinoline to 2(lH)-quinoUnone
H2O -> C9H7NO + 2 H+ + 2 e" 84
Isoquinoline to I(2H}-isoquinolinone
+
G)H7N + H2Q -> C9H7NO + 2 H + 2 e' 56
a C8H7NO + H2
Quinoline with Hydrogen
CgH'yN + H2O -* C9H7NO + H2
Isoquinoline with Hydrogen
C9H7N + H2O -» C9H7NO + H2
Indole with Methane
C8H7N + 0.5 H2O + 0.25 CO2 -> C8H7NO + 0.25 CH4
Quinoline with Methane
C9H7N + 0.5 H2O + 0.25 CO2 -» C9H7NO + 0.25 CH4
Isoquinoline with Methane
G)H7N + 0.5 H2O + 0.25 CO? -» CQH7NO + 0.25 CH4
146
123
116
113
90
83
(J) calculated using values from Stull et al. (1969) or estimated using the method of Jobak
(Reid et a/., 1985) for compounds not found in the literature.
79
-------
Table 5-18. Gibbs free energy changes during oxidation of phenolic compounds at 25*C
and pH 7.
Compound &G(aa) kJ/mol
Phenol to CH4 and C02
CgHgO + 3 H2O -» 2.5 CO2 + 3.5 CH4 - 106
2-Methylphenol to CH4 and CO2
C7H80 + 4.5 H20 -» 2.75 C02 + 4.25 Cfy - 1 37
3-Methylphenol to CH4 and C02
C7H8O -i- 4.5 H2O -» 2.75 CO2 + 4.25 CH4 - 134
4-Methylphenol to CH4 and CO2
C7HSO + 4.5 H?O -» 2.75 CQ2 + 4.25 CH4 - 146
aq) calculated using values from Stull et al (1969) or estimated using the method of Jobak
(Reid et a/., 1985) for compounds not found in the literature.
Table 5—19. Gibbs free energy changes during oxidation of NSO compounds at 25*C and
pH7.
Compound A(J(aq) kJ/mol
Oxindole to CH4 and CO2
C8H7NO + 6.5 H2O + H+ -» 3.75 CO2 + NHj + 4.25 CH4 - 360
2(lH)-quinolinone to CH4 and CO2
C9H7NO + 7.5 H2O -t- H+ -> 4.25 CO2 + NHj + 4.75 CH4 - 279
l(2H)-isoquinolinone to CH4 and CO 2
C9H7NO + 7.5 H20 + H+ -* 4.25 CO2 + J + 4.75 CH4 - 274
Benzothiophene to CH4 and CO2
CgHeS + 7 H2O -> 3.5 CO2 + 0.5 H2S + 0.5 HS" + 0.5 H+ -f 4.5 CHj - 1 85
Benzofuran to CH4 and C02
+ 6 HO -> 3.5 CO, + 4.5 CH4 _ -246
AG/aq) calculated using values from Stull et al. (1969) or estimated using the method of Jobak
(Reid et a/., 1985) for compounds not found in the literature.
80
-------
Table 5-20. Gibbs free energy changes during oxidation of PAH compounds at 25"C and
pH7.
Compound AG^ kJ/mol
Indene to CH4 and CO2
CgHg + 7 H2O -» 3.5 CO2 + 5.5 CH4 - 1 84
Naphthalene to CH4 and CO2
+ 8 H2O -» 4 C02 + 6 CH4 -154
(aq) calculated using values from Stull el al. (1969) or estimated using the method of Jobak
(Reid et a/., 1985) for compounds not found in the literature.
81
-------
6. DETERMINATION OF THE SORPTION CHARACTERISTICS OF THE WSF
COMPOUNDS
RESEARCH TASKS
Before one can begin to assess the importance and rates of biodegradation during
transport of contaminants, the physical processes occurring during transport must be
determined. The transport processes of concern in ground-water contamination include
advection, dispersion, adsorption, decay, chemical reaction(s), and biodegradation. Recent
literature shows much has been learned about the effects of these processes and numerous
conceptual-mathematical models have been developed in attempts to describe the
one-dimensional transport of chemicals in laboratory columns as well as in field situations.
Breakthrough curves (ETC) for the WSF compounds and CaCl2, determined not to be subject
to detectable anion exclusion in this study, were measured during sterile, saturated, and steady
flow conditions on both 60-80 mesh glass-beads and uncontaminated disturbed aquifer
material from the ACW site. Two conceptual models, the local equilibrium assumption (LEA)
model and the nonequilibrium assumption (NEA) model, were used to simulate the observed
ETC and for the determination of the sorption characteristics of the biodegradable WSF
compounds.
BACKGROUND
Organic contaminants entering a ground-water system from the surface, will tend to
move downward towards the water table. When the contaminated water reaches the water
table, a contaminant will tend to move laterally in the direction of the ground water flow. As
the water moves away from the pollution source, the concentration of contaminants decreases
due to dispersion and may decrease due to adsorption on aquifer particle surfaces, chemical,
and/or biochemical reactions.
The mechanisms of advection and dispersion have an important control on the transport
of organic pollutants in the subsurface environments. Total contaminant flow is composed of
the portion that travels with the average ground-water flow and the portion that deviates from
the average ground-water flow. The latter is responsible for dilution of the contaminant
concentration and a spreading of the contaminated area.
There are two types of dispersion: (1) dispersion that occurs at the pore scale
(microdispersion) and (2) dispersion that occurs at the field scale due to aquifer heterogeneity
82
-------
(macrodispersion). Microdispersion is significant for transport in relatively slow-flowing
ground water and macrodispersion is significant due to the heterogeneity of most aquifers
(e.g., Sudicky et al., 1983). These phenomena make field and laboratory comparisons very
difficult.
Aquifers comprised of deposits where former living organisms are likely to have
accumulated, such as marine terraces, lakes, bogs, etc., tend to have significant organic carbon
content. Studies have shown that at least 0.1 percent carbon content is needed for carbon
adsorption to be significant (e.g., Schwarzenbach and Westall, 1981). Instead of solubility,
the octanol:water partition coefficient (K^) is often used as a measure of the partitioning of
contaminants between water and the organic phases of the aquifer material. An inverse
correlation between log Kow values (1 < log K^, < 6) and log solubility values (-3 < log sol,
mg/L < 5) has been found for non-polar organic compounds (Mackay, 1980).
Useful relationships have been found between the adsorption behavior of a contaminant
and its K^ value and the organic content of an aquifer. Karickoff el al. (1979) demonstrated
that the degree to which a compound is adsorbed, as measured by the partition coefficient
(Kffr, depends on the K^ and the "fraction organic content" (foc) of the aquifer material by
the relation
Kd = 0.63focKow (6-1)
This equation supplies Kd in cm3/g and applies when the contaminant concentration is
less than half of the solubility limit in water and for contaminants with 2 < log K^ < 6 and
0.001
-------
where c is the mass of solute species adsorbed on the material per unit bulk dry mass of the
aquifer, c is the solute concentration, and Kf and n are coefficients that depend on the solute
species, nature of the aquifer material, and other conditions. If the slope of the isotherm, as
represented by the term V/i, is equal to 1, then the isotherm is linear and Kf- Kj. A
comprehensive treatment of adsorption isotherms is presented by Helfferich (1962), who
provides detailed information on many important types of isotherms in addition to the
Freundlich isotherm.
Quantitative relationships have not been well established between sorption and the
controlling factors for aquifers containing less than 0.1 percent organic material. Some
adsorption of non-polar organic compounds was observed in columns containing materials
that do not contain organic matter, such as clean sand, limestone, and montmorillonite clay
(Schwarzenbach and Westall, 1981). Goerlitz et al (1985) demonstrated in laboratory
columns containing sediments from ACW that naphthalene was retarded during movement
through the column and was presumably sorbed onto the aquifer sediments, even though the
foc of the material was less than 0.1 percent. This result suggests that some mechanism other
than adsorption on the organic material may be responsible for the retardation of naphthalene
in the column.
Polar organics appear to be more mobile than non-polar organics, as shown by a study in
an aquifer with significant amounts of organic carbon, because they are poorly retained in the
organic material in the aquifer (Roberts et al., 1982). Goerlitz et al. (1985) found that phenol,
3-methylphenol, and 3,5-dimethylphenol, all polar compounds, were not retarded during
movement through laboratory columns containing aquifer material from ACW while
pentachlorophenol, also a polar compound, was retarded. These results also suggest a
mechanism other than adsorption onto the organic material in the aquifer.
Recent investigations by Westall et al. (1985) on the distribution ratio for the polar
ionizable chlorinated phenols demonstrated that the major influence on the K^ value was the
pH and ionic strength of the aqueous phase. Zacchara et al. (1986) investigated the sorption
isotherms for the hydrophobic ionizable organic base quinoline under differing pH and ionic
strength aqueous solutions. It was also observed that the organic cation was strongly sorbed
when pH was below neutrality while the neutral quinoline molecule was only weakly sorbed
under the same conditions. They concluded that in subsurface materials of low carbon
content, quinoline sorption is controlled by pH, the nature and capacity of the exchange
complex, and the ground water ion composition.
84
-------
It is apparent that methods used to describe the movement of non-polar compounds in
the subsurface are not adequate to describe the movement of hydrophobic ionizable organic
compounds in low organic content aquifer material. The effects of pH and ionic strength are
site specific and must be determined independendy.
The behavior of a conservative tracer which experiences neither adsorption nor chemical
transformation during transport from storage in a one dimensional solution-filled porous
medium (Ogata and Banks, 1961) can be expressed as
*y
9c 9 c Be
where
6 = porosity (-)
t = time of transport (day)
c = solute concentration (mg/L)
q = water flux (cm3/cm2-day)
Dx = longitudinal dispersion coefficient (cm2/day)
x = characteristic length (cm).
Linear sorption on the porous medium can be described as
c = Kdc (6-5)
where
c = concentration of sorbed solute (mg/g)
Kj= partition coefficient (L/g).
Including the storage term on the solids in the transport equation, the result is
where
p = bulk density (g/L).
Partial differentiation of equation 6-5 yields
85
-------
dc -If dc (*1\
— = Ad — (6-7)
3r dt
then combining equation 6-6 with equation 6-7 and simplifying gives
Be B2c dc (6-8)
After dividing through by 0 and defining the dimensionless variable Rt as,
*, = 1+^- (6-9)
0
the familiar form of the solute transport equation is obtained (Hashimoto et a/., 1964)
where
v = average linear pore water velocity (cm/day).
Local Equilibrium Assumption Model
The analytical solution of equation 6-10 for constant flux boundaries (Type 3)
c(jc,0) = 0 (6-11)
D,fH| =vc. («.,»
dx J\X = Q
|^(-,0 = 0 (6-13)
dx
in terms of the dimensionless variables
T = — (6-14)
P = ^~ (6-15)
for values of P £ 40 is given by Hashimoto et al. (1964)
86
-------
where
c = output concentration (mg/L)
c0 = input concentration (mg/L)
erfc = complementary error function
exp = exponent, base e.
Most models that have been used to describe solute transport have been based on the
convection-dispersion transport model, which assumes local physical and/or chemical
equilibrium. Experimental observations, however, have often differed from the sigmoidal
solution predicted by these models. Physical nonequilibrium processes, i.e., water flow in
different domains (van Genuchten and Wierenga, 1976; Crittenden et al., 1986) is one
mechanism thought to be responsible for asymmetric elution profiles.
Nonequilibrium Assumption Model
The convective-dispersive transport of a sorbing solute during one-dimensional transport
in laboratory columns that takes into account mass transfer into immobile water in addition to
advection, dispersion, and linear adsorption (NEA - Nonequilibrium Assumption Model) is
given by van Genuchten and Wierenga (1976)
,-H
= QmDx^-vmQm^ (6-17)
(6-18)
where
cm = solute concentration in the mobile liquid phase (mg/L)
c-m = solute concentration in the immobile liquid phase (mg/L)
Qm = mobile water content (-)
Qim = immobile water content (-)
/ = fraction of sorption sites in mobile region (-)
aj = first-order rate coefficient (day1)
vm = mobile phase pore water velocity (vm = <7/6m) (cm/day).
Equation 6-17 limits convective-dispersive transport to the mobile liquid region and
assumes that diffusion is responsible for the exchange of solute between mobile and immobile
regions. The rate limitations of diffusion and external mass transfer between mobile and
immobile water is approximated by a first-order rate expression given in equation 6-18.
87
-------
n
-D—+vc
Analytical solutions to equations 6-17 and 6-18 exist for several sets of initial and
boundary conditions (van Genuchten and Wierenga, 1976). The initial and boundary
conditions for this study are as follows.
(6-19)
a . - , .-„ (6-20)
dx A^y
—(*,/) = 0 (6-21)
dx
The following dimensionless variables were introduced by van Genuchten and Wierenga
(1976) to facilitate the analytical solution of equations 6-17 and 6-18
T = — = v""™« (6-22)
(6-23)
R,
e«
(6-25)
(6-26)
mzL (6_27)
(6_28)
and with the above dimensionless parameter definitions for the NEA model, equations 6-17
and 6-18 reduce to
(6-30)
-------
Inspection of equations 6-29 and 6-30 shows that the non equilibrium model contains
four independent dimensionless parameters: P, Rt, P, and CO which are fit by the equations to
experimental data by a nonlinear curve fitting routine that minimizes the sum of the squared
residuals. The Peclet number (P) measures the dispersion tendency or the ratio of the rate of
transport by advection to the rate of transport by axial dispersion; the retardation factor (Rt)
represents the influence of sorption on transport; the fraction of instantaneous retardation (P)
represents the influence of distribution of sorption between instantaneous (mobile) and
rate-limited domains (immobile) or the maximum degree of nonequilibrium in the system; and
co represents the ratio between hydrodynamic residence time and characteristic time of
sorption or the rate at which equilibrium is obtained. The values of P and co specify the
degree of nonequilibrium existing in the system, which decreases as either of the two
parameters increase in magnitude. If no adsorption occurs, Rt equals one and P reduces to
<))m. The number of independent parameters is then only three. The original variables can be
obtained using the dimensionless variables and other independently determined measurements
(e.g., column length, volumetric water content, flux rate, etc.).
Equations 6-29 and 6-30 have been solved in a number of ways, but De Smedt and
Wierenga (1979) determined that all solutions can be expressed in the same general format
The general analytical solution for the dimensionless exit concentration, clc0, is given by van
Genuchten (1981). For the initial and boundary conditions of this study, the general solution
is
Co
COT
CO fT
f-f G(T)//(T,T)dT (6-31)
Kt
where
T = dummy integration variable
O = modified Bessel function
1 1 = modified Bessel function
m (6-32)
(1-P)/?,
89
-------
and where G(T) depends on the imposed initial and boundary conditions. For this study
P 2
Px
e\p(P)erfc
(6-33)
RESULTS
Column Characteristics
Shown in Table 6-1 are the operating conditions and column parameters for the glass
bead and the ACW aquifer material adsorption studies. During the course of the experiments
three separate columns were prepared. The effluent line from Column #1 plugged and the
column glass-barrel broke. A pressure relief valve was then added to bypass the flow if the
back pressure reached the breaking point of the glass columns as a result of paniculate
plugging in the effluent line. Column #2 was replaced when the column became contaminated
by microbial growth.
Local Equilibrium Assumption Model Analysis
Table 6-2 contains fi/2, the detention time in minutes until breakthrough of l/2 maximum
concentration; and P, the Peclet number for compounds with log KQ^ < 2.5 using the LEA
model. The 95 percent confidence interval (CI) is given for both t\f2 and P for each
compound. The values that are given for the CaCl2 tracer runs for each of the aquifer
material columns represent only one of many runs. A CaCl2 tracer determination was done
during each run of an organic compound. The elution profiles for CaCl2 for both the glass
beads (Figure 6-1) and aquifer material (Figure 6-2) are symmetrical and the inflection points
(tl/2) occur at one pore volume indicating that this solute is not adsorbed or subject to anion
exclusion on the experimental apparatus or on the aquifer material. The elution profiles for
the compounds listed in Table 6-2 all show adsorption to some degree (Figures 6-3 to 6-13),
but they fit the LEA model predictions remarkably well.
90
-------
Nonequilibrium Assumption Model Analysis
Table 6-3 lists the parameter estimations P, /?,, p, an indication of the maximum degree
of nonequilibrium in the system, either in a physical or chemical sense, and CD, the rate at
which equilibrium was obtained from the initial nonequilibrium, and the 95 percent CI for each
of the compounds with a log K^ > 2.5. Those compounds given in Table 6-3 all show
considerable deviation from the LEA. This model was unsuccessful at predicting the sharper
breakthrough and subsequent tailing as shown in Figures 6-14 to 6-17. LEA has been shown
to be reasonable approximation of the ETC when the value of 00 is larger than approximately
10 (Brusseau et al., 1989), a much larger value than was found for any of compounds with a
log KW > 2.5. The NEA model, however, does a remarkable job of fitting the data and
considerably reduces the RSS as given in Table 6-4.
DISCUSSION AND CONCLUSIONS
The LEA model shows a remarkable fit to the chloride tracer ETC. Both the dispersion
coefficient, Dx, and the time required for half the mass to exit the column (t\/2) were
optimized by reducing the RSS. The pore volume was independently determined for each
column. The dispersion coefficient for the chloride tracer experiments (7.25 to 24.67
cm2/day) were determined using the Peclet number (P), a dimensionless group that measures
the dispersion tendency or the ratio of the rate of transport by advection to the rate of
transport by axial dispersion (P = \LIDX). The term v is the average linear velocity (q/Q) and
L is the column length. Ogata and Banks (1961) show that for P > 100, dispersion can
practically be neglected, whereas for P < 5 the flow regime approaches complete mixing. The
large values of P (or low values of Dx) for the glass bead experiment (Table 6-2) demonstrate
that there is very little dispersion due to the experimental apparatus.
Assuming a linear relationship between the dispersion coefficient Dx and the dispersivity
(a) defined as Dx = ccv, values for a can be determined (Table 6-1). Comparison with
literature values shows that the values in this study for a (0.16 to 0.25 cm) are in good
agreement with experiments on disturbed soil columns in the literature: 0.01 cm < a < 0.5 cm
(Nkedi-Kizza et al., 1983; Khan and Jury, 1990).
Those compounds with a log K^ less than 2.5 can be fit by the LEA model (Table 6-5).
The values obtained for Rt for these compounds correlate with log K^ values obtained from
the literature - the greater the log K^ value the greater the /?,. The possible exception to this
is the value of R{ (1.41) obtained for l(2H)-isoquinolinone. The value of log K^ (0.58)
91
-------
would suggest that Rt would be much less, with a value somewhere near one. The correlation
between the values of Kj, Rt, and log K^ for the other compounds are as expected. The
value given in Hansch and Leo (1979), listed as a personal communication, might be in error.
Zachara et al. (1986,1987) demonstrated that for nitrogen-containing heterocyclic
compounds that may be protonated, sorption is more controlled by pH, the nature and
capacity of the exchange complex, and the ground water ion composition.
Modeling the effects of adsorption on solute transport with LEA assumes that the solute
and the aquifer material react in an instant equilibrium, i.e., no kinetic effects, that the ratio of
the adsorbed solute to the solute dissolved in water is constant, i.e., linear isotherm, and that
adsorption and desorption is a reversible process. Kinetic effects are important when the pore
velocity of the ground water or laboratory column eluent is too fast to allow equilibrium
(Brusseau et al, 1989). Apparently, equilibrium is not reached instantaneously in this system,
even at flow rates that reflect the actual ground water flow rate at ACW. When ETC for
naphthalene were tested at different values of q ( Slow = 23.75 cm/day; Fast = 118.8 cm/day),
the observed tailing behavior was proportional to the average pore velocity much as
Schwarzenbach and Westall (1981) observed in their experiments and Rt was reduced. At
higher velocities, sorption and desorption of naphthalene does not proceed at a rate sufficient
to reach local equilibrium and the apparent value of the Rt decreases (Bahr, 1989; Priddle and
Jackson, 1991).
Diffusion-related nonequilibrium, or physical nonequilibrium assumption (PNEA) is
thought to result from the existence of regions within the porous medium in which there is
minimal or no advecu've flow, called immobile regions (van Genuchten and Wierenga, 1976).
Diffusion of solutes between these mobile and immobile regions results in a source or sink of
the solute during transport and is rate limited by mass transfer between these regions.
Chemical kinetic nonequilibrium (CNEA) is thought to arise from rate-limited interactions
between the solute and specific sorption sites of the sorbent (Brusseau et al., 1989).
Nkedi-Kizza et al. (1984) have demonstrated that these conceptual models are very similar
with respect to the BTC, with deviations too small to be detected experimentally with only
BTC analyses. Thus, differentiation of the processes taking place in the aquifer material
columns cannot be obtained with solutions to Equations 6-29 to 6-30, in fact, Roberts et al.
(1987) clearly demonstrated that independent parameter determinations from theoretical
considerations can distinguish between the various processes occurring within the column;
however, these analyses were not attempted for this study.
92
-------
There is a increasing amount of published literature where nonideal behavior is observed
in disturbed soils; for example, nonsorbing or weakly sorbing solutes exhibit symmetrical
ETC, whereas more strongly sorbing solutes do not (Brusseau el al., 1989). It is suggested
that the nonequilibrium is a result of rate-limited interactions between an organic sorbate and
the organic sorbent on or in the porous media. Further, there is support in the literature
showing organic matter to be the predominant sorbent for organic compounds, with
mineral-surface contributions to sorption of organic compounds being negligible (Ball and
Roberts, 1991a; Brusseau and Rao, 1989). Therefore, it would appear that the observed
nonequilibrium phenomena in this study are due to slow nonaqueous phase diffusion into the
organic matter or resistance to mass transfer across an external fluid boundary layer (Ball and
Roberts, 199 Ib) and subsequent slow mass transfer between liquid and solid phases. If the
assumption that organic matter is the predominant sorbent for hydrophobic organic
compounds is valid, then the paucity of organic material on the ACW sediment (0.07 percent)
would explain why only the more hydrophobic compounds (log K^ > 2.5) exhibit
nonequilibrium behavior.
Table 6-6 gives the values for the Darcy flux (q), the fraction of liquid phase assumed to
be mobile (()>), the partition coefficient (Kd), the first order transfer coefficient (04), the
retardation factor (/?,), and the log K^ for the more hydrophobic compounds. The NEA
models predicts that approximately 90 percent of liquid is mobile, resulting in relatively little
deviation from LEA since only about 10 percent of the liquid is immobile. The BTC for
indene (Figure 6-16) shows the least amount of tailing, as is reflected in the larger value of Oj
exhibited for this compound.
For sorbing solutes, sorption isotherm nonsingularity (i.e., sorption/desorption hysteresis)
and sorption kinetics can contribute to nonequilibrium during solute transport. Further,
asymmetric BTC can be due to sorption isotherm nonlinearity. Sorption isotherm linearity
appears to be a reasonable assumption for low solute concentrations. As demonstrated by
Chiou et al. (1979) and Munz and Roberts (1986) linear isotherms may be expected for
solution concentrations below 56 mmol, or below the solute solubility, whichever is lower.
Sorption linearity can, thus be safely assumed for this study as all of the concentrations used in
this study are several orders of magnitude less than 56 mmol.
The retardation factors for the anaerobically biodegradable compounds tested in this
study are quite low and, apparently, are not the major cause for the rapid decrease of these
compounds during downgradient travel in the aquifer at the study site.
93
-------
§
0.75
o
S
o
0.5
H
5
0.25
0
0
0.5 1 1.5
PORE VOLUMES
O CHLORIDE LEA MODEL
Figure 6-1. Breakthrough curve for calcium chloride using 60-80 mesh glass beads.
94
-------
o
8
§0.75
O
§
I
0.25
0
0
PORE VOLUMES
O CHLORIDE LEA MODEL
Figure 6-2. Breakthrough curve for calcium chloride using aquifer material.
95
-------
PORE VOLUMES
O PHENOL LEA MODEL
Figure 6-3. Breakthrough curve for phenol using aquifer material.
96
-------
0.5 1 1.5
PORE VOLUMES
O 2-METHYLPHENOL LEA MODEL
Figure 6-4. Breakthrough curve for 2-methylphenol using aquifer material.
97
-------
0
PORE VOLUMES
O 3-METHYLPHENOL LEA MODEL
Figure 6-5. Breakthrough curve for 3-methylphenol using aquifer material.
98
-------
PORE VOLUMES
O 4-METHYLPHENOL LEA MODEL
Figure 6-6. Breakthrough curve for 4-methylphenol using aquifer material.
99
-------
PORE VOLUMES
O 3,5-DIMETHYLPHENOL LEA MODEL
Figure 6-7. Breakthrough curve for 3,5-dimethylphenol using aquifer material.
100
-------
i
lzfO.75
o
0.5
1
o
0.25
0
0 0.5 1 1.5
PORE VOLUMES
O INDOLE LEA MODEL
Figure 6-8. Breakthrough curve for indole using aquifer material.
101
-------
1 1.5 2
PORE VOLUMES
O QUINOLINE LEA MODEL
Figure 6-9. Breakthrough curve for quinoline using aquifer material.
102
-------
lr
1 2 3
PORE VOLUMES
O ISOQUINOLINE LEA MODEL
Figure 6-10. Breakthrough curve for isoquinoline using aquifer material.
103
-------
0 0.5 1 1.5
PORE VOLUMES
O OXINDOLE LEA MODEL
Figure 6-11. Breakthrough curve for oxindole using aquifer material.
104
-------
1 1.5 2
PORE VOLUMES
O 2(1H)-QUINOLINONE LEA MODEL
Figure 6-12. Breakthrough curve for 2(lH)-quinolinone using aquifer material.
105
-------
0
0.5
1 1.5 2
PORE VOLUMES
O 1(2H)-ISOQUINOLINONE LEA MODEL
Figure 6-13. Breakthrough curve for l(2H)-isoquinolinone on aquifer material.
106
-------
§ °-75
UJ
O
I
0.5
0.25
0
0 23
PORE VOLUMES
O BENZOTHIOPHENE —• LEA MODEL
NEA MODEL
Figure 6-14. Breakthrough curve for benzothiophene using aquifer material with both
LEA and NEA model predictions.
107
-------
0.5 1.5
PORE VOLUMES
BENZOFURAN -— LEA MODEL
2.5
NEA MODEL
Figure 6-15. Breakthrough curve for benzofuran using aquifer material with both LEA
and NEA model predictions.
108
-------
§
§ 0.75
w 0.5
I
EE
S 0.25
O INDENE
PORE VOLUMES
-— LEA MODEL
NBA MODEL
Figure 6-16. Breakthrough curve for indene using aquifer material with both LEA and
NEA model predictions.
109
-------
0
1.5 2
PORE VOLUMES
O NAPHTHALENE LEA MODEL
NEA MODEL
Figure 6-17. Breakthrough curve for naphthalene using aquifer material with both
LEA and NEA model predictions.
110
-------
Table 6-1. Column elution experimental data for the glass bead and ACW aquifer
material experiments.
60/80 mesh Aquifer Aquifer Aquifer
Porous media glass-beads material, #1 material, #2 material, #3
Column length, cm 35.0 35.0 35.0 35.4
Column diameter, cm 2.5 2.5 2.5 2.5
Column volume, cm3 171.8 171.8 171.8 173.8
Particle density, g/cm3 2.41 2.47 2.47 2.47
Mean particle diameter, cm 0.020 0.038 0.038 0.038
Organic content, % 0 0.07 0.07 0.07
Porosity 0.373 0.381 0.407 0.449
Dispersivity, cm 0.025 0.25 0.43 0.16
Pore volume, cm3 64.08 65.44 69.92 78.02
111
-------
Table 6-2. LEA model prediction data for WSF compounds with a log K^ < 2.5.
Compound (Column #)
CaCl2 (Glass Beads)
CaCl2(l)
CaCl2 (2)
CaCl2 (3)
Phenolic Compounds
Phenol (2)
2-Methylphenol (2)
3-Methylphenol (1)
4-Methylphenol (1)
3,5-Dimethylphenol (1)
NSO Compounds
Indole (2)
Quinoline (2)
Isoquinoline (2)
Oxindole (2)
2(lH)-quinolinone (1)
l(2H)-isoquinolinone (1)
log K^
NA
NA
NA
NA
1.48
1.93
1.98
2.00
2.35
2.10
2.03
2.08
0.85
1.26
0.58
fi/2, day 2
±95%CI
0.588 ±0.001
0.605 ±0.002
0.596 ±0.001
0.563 ±0.001
0.647 ±0.003
0.655 ±0.005
0.647 ±0.002
0.667 ±0.006
0.693 ± 0.002
0.758 ±0.002
0.771 ±0.008
1.088 ±0.007
0.643 ±0.001
0.938 ±0.002
0.848 ±0.003
±95%CI
1393.0 ±127.0
138.3 ± 8.2
27 1.0 ±14.3
178.2 ± 9.8
114.9 ±10.3
130.2 ±20.2
153.1 ±12.5
135.1 ±26.6
136.1 ± 10.4
107.5 ± 6.5
56.2 ±8.6
36.0 ±2.8
254.6 ± 6.4
124.6 ±7.2
87.3 ± 6.5
iRansch and Leo (1979)
2 Detention time until breakthrough of one half of the maximum concentration
3 Peclet number
NA = Not Applicable
112
-------
Table 6-3. NBA model prediction data for WSF compounds with a log K^ > 2.5.
1
Compound (Column #)
NSO Compounds
Benzothiophene (1)
Benzofuran (2)
PAH Compounds
Indene (2)
Naphthalene (2) - Slow
Naphthalene (2) - Fast
LogKw1
3.12
2.67
2.92
3.45
3.45
±95%CI
138.3 ±8.20
101. 7 ±9.67
101.7 + 9.67
101. 7 ±9.67
269.8 ± 89.4
±95%CI
1.74 ±0.01
1.23 ±0.05
1.68 + 0.01
1.48 ±0.06
1.10±0.01
P4
±95%CI
0.90 ± 0.01
0.87 + 0.03
0.85 + 0.02
0.91 ±0.01
0.93 ± 0.01
0)5
±95%CI
0.48 ±0.10
0.32 + 0.15
1.14 + 0.31
0.38 ±0.1 4
0.21+0.06
1 Hansch and Leo (1979)
2 Peclet number
3 Retardation factor
4 Maximum degree of nonequilibrium in the system
5 Rate at which equilibrium was obtained
Table 6-4. RSS data for WSF compounds with a log K^ > 2.5 on aquifer material for
both the Local Equilibrium (LEA) and Non Equilibrium Assumptions (NEA).
Compound (Column #) LEA NEA
NSO Compounds
Benzothiophene (1) 0.035 0.0104
Benzofuran (2) 0.235 0.0894
PAH Compounds
Indene (2) 0.014 0.007
Naphthalene (2) - Slow 0.051 0.009
Naphthalene (2)-Fast 0.145 0.019
113
-------
Table 6-5. Adsorption data for WSF compounds with a log Kow < 2.5 on aquifer
material.
Compound
CaCl2(l)
CaCl2 (2)
CaCl2 (3)
Phenolic Compounds
Phenol
2-Methylphenol
3-Methylphenol
4-Methylphenol
3,5-Dimethylphenol
NSO Compounds
Indole
Quinoline
Isoquinoline
Oxindole
2(lH)-quinolinone
1 (2H)-isoquinolinone
cm/day
22.05
22.83
28.16
22.29
22.29
22.05
22.05
22.05
22.29
23.76
23.91
22.83
22.05
22.05
cm3/g
NA
NA
NA
0.004
0.007
0.018
0.026
0.038
0.052
0.092
0.273
0.022
0.137
0.103
R<
1.00
1.00
1.00
1.01
1.03
1.07
1.10
1.15
1.19
1.33
1.98
1.08
1.55
1.41
log
K 1
NA
NA
NA
1.48
1.93
1.98
2.00
2.35
2.10
2.03
2.08
0.85
1.26
0.58
THansch and Leo (1979)
NA = Not Applicable
114
-------
Table 6-6. Adsorption data for WSF compounds with a log K^ > 2.5 on aquifer
material.
q, 0 Kfa ct,\,
Compound cm/day cm3/g day1
NSO Compounds
Benzothiophene
Benzofuran
PAH Compounds
Indene
Naphthalene - Slow
Naphthalene - Fast
22.05
22.83
22.29
23.75
118.81
0.90
0.87
0.85
0.91
0.93
0.184
0.064
0.186
0.153
0.028
0.30
0.32
0.73
0.38
0.21
1.74
1.23
1.68
1.48
1.10
115
-------
7. SUMMARY AND CONCLUSIONS
The research conducted was a multi-phase study, with the goal of elucidating some of the
microbiological and physical processes (and factors which influence them) affecting separate
groups of the water soluble fraction of creosote and, more specifically, chosen single
compounds representative of the major compound groups. The major conclusions resulting
from this study are summarized below:
1. The major double ring compounds in the water soluble fraction of creosote are
anaerobically degradable as primary substrates to Ctfy and CO2- The compounds tested were
mineralized during downgradient transport in the aquifer and in laboratory microcosms that
simulated the aquifer environment. Examples of three-ring polynuclear aromatic or nitrogen
and sulfur heterocyclic compounds were not degraded under methanogenic conditions in these
microcosms; however, there have been reports of utilization of these compounds in the
literature (Grbid-Galid, 1989).
2. The degradation kinetics of the major individual compounds can be described by
Monod kinetics. The kinetic constants generated are remarkably similar for all of the
compounds tested, suggesting that the degradation rates measured result from the kinetics of
the rate-limiting population. The variable and sometimes quite long onset times for substrate
utilization and methanogenesis suggest that more than one population is responsible for the
initial attack on the aromatic rings. These observations would suggest that the rate-limiting
populations are common to all the microcosms and perhaps different organisms or populations
are responsible for initial ring attacks. This would be a possible explanation as to why the
times to active substrate utilization varied from one microcosm to another while the kinetic
constants for each microcosm are essentially the same.
The inability of the Monod family of equations to describe the degradation of the WSF
compounds as total organic carbon or some other lumped parameter is not surprising since
these equations were developed for pure cultures degrading single organic compounds. Thus,
this investigation reminds us that care must be taken when trying to describe mathematically
the biodegradation of a complex mixture of compounds by a complex consortium of
microorganisms.
3. The consistent observation of phenol and benzole acid as intermediates in the
methanogenic degradation of the WSF compounds strongly suggests that the major
degradation pathway for these compounds intersects the phenol and/or benzoic acid
degradation pathways. Recent studies have questioned the existence of a separate phenol
pathway, suggesting instead the carboxylation of phenol to form a benzoate derivative. If this
116
-------
is the case, the persistence of phenol in many of the microcosms could be due to the
endergonic activation reaction to the benzoate derivative.
A minor pathway based on the abundance of the intermediates was observed for
benzothiophene and quinoline. This pathway is initiated by the attack on the homocyclic ring
and has not been reported in the literature until this study. It is not surprising that this
pathway exists, considering the discovery several years ago of the methanogenic degradation
of benzene and toluene. The minor heterocyclic pathway detected in this study is most likely
the same pathway utilized by the organisms degrading homocyclic compounds. The
importance of this pathway for other homocyclic and heterocyclic compounds is unknown at
this time and will certainly be the topic of further research.
4. The sequence of degradation for the nitrogen heterocycles studied was shown to be
the result of an initial endergonic oxidation to form a stable oxidized compound. It has been
speculated that the initial oxidation of these compounds supplies reducing power for the
reduction of other more readily degradable aromatic compounds. Kompala el al. (1986) have
evidence that the preferential and sequential substrate metabolism is a function of the bacterial
growth rate supported by the individual compound. Those compounds supporting the fastest
growth rate will be preferentially degraded first. The degradation sequences of single
compounds observed in this study could not be attributed to the specific growth on the
individual compounds; the values of n,^ for all of the compounds tested did not vary
significantly from a value of 0.1 day1. Nor could the degradation sequence be explained by
the amount of Gibbs free energy available, as shown in Tables 5-18 to 5-20. Results have led
to speculation that the sequential degradation of these compounds is controlled at the genetic
level and is beyond the scope of this study.
5. Adsorption of organic compounds onto the aquifer sediment was shown not to play a
significant role in the retardation of the major WSF compounds during advective transport
Breakthrough curves for the WSF compounds with log K^ less than 2.5 were readily fitted
by a LEA model at this research site. The values obtained for the retardation factors in this
study correlated with the log K^ values obtained from the literature - the greater the log K^
value, the greater the value of the retardation factor. Breakthrough curve column studies with
WSF compounds with log K^ values greater than 2.5 showed considerable tailing and could
only be fitted with nonequilibrium assumption models; however, the nature of the
nonequilibrium, other than it was chemical in nature, could not be determined with the model
used.
117
-------
This study has demonstrated that integration of both laboratory results and field
observations can provide valuable information regarding the ultimate fate of complex organic
subsurface contaminants. Results of this study indicate that the disproportionate decrease of
selected organic compounds observed during downgradient movement in the aquifer may be
attributed to microbial degradation of selected compounds. The anoxic conditions in the
contaminated area, high concentrations of CH4, and the apparent increase in numbers of
methanogenic bacteria suggest that methanogenic fermentation is the predominant
microbiological process. This observation was corroborated in laboratory microcosm studies.
Furthermore, these processes could be successfully mathematically modeled.
118
-------
8. REFERENCES
American Petroleum Institute (ed.). 1978. Naphthalene. API Monograph Series, Publication
707, American Petroleum Institute Washington, D.C.
American Petroleum Institute (ed.). 1979a. Isoquinoline. API Monograph Series, Publication
712, American Petroleum Institute Washington, D.C.
American Petroleum Institute (ed.). 1979b. Quinoline. API Monograph Series, Publication
711, American Petroleum Institute Washington, D.C.
American Petroleum Institute (ed.). 1980. Indan and indene. API Monograph Series,
Publication 714, American Petroleum Institute Washington, D.C.
American Petroleum Institute (ed.). 1981. Acenaphthylene, acenaphthene, fluorene, and
fluoranthene. API Monograph Series, Publication 715, American Petroleum Institute
Washington, D.C.
American Petroleum Institute (ed.). 1982. Indole. API Monograph Series, Publication 719,
American Petroleum Institute Washington, D.C.
American Petroleum Institute (ed.). 1983. Benzofuran, dibenzofuran, and
benzonaphthofurans. API Monograph Series, Publication 721, American Petroleum
Institute Washington, D.C.
American Society of Agronomy (ed.). 1965. Methods of soil analysis, 2 pts. American Society
of Agronomy, Madison, Wisconsin.
Baedecker, M.J., and I.M. Cozzarelli. 1992. The determination and fate of unstable
constituents in ground water, pp. 425-461, in Lesage, S., and R. Jackson (eds.), Ground
water quality and analysis at hazardous waste sites. Marcel Dekker, New York.
Bak, F., and F. Widdel. 1986. Anaerobic degradation of indolic compounds by
sulfate-reducing enrichment cultures, and description of Desulfobacterium indolicwn
gen. nov. sp. nov. Arch. Microbiol. v. 146, pp. 170-176.
Bakker, G. 1977. Anaerobic degradation of aromatic compounds in the presence of nitrate.
FEMS Letters v. l,pp. 103-108.
Bahr, J.M. 1989. Analysis of nonequilibrium desorption of volatile organics during field test of
aquifer decontamination. J. Contam. Hydrol. v. 4, pp. 205-222.
Ball, W.P., and P.V. Roberts. 1991a. Long-term sorption of halogenated organic chemicals by
aquifer material. 1. Equilibrium. Environ. Sci. Technol. v. 25, pp. 1223-1237.
Ball, W.P., and P.V. Roberts. 1991b. Long-term sorption of halogenated organic chemicals by
aquifer material. 2. Intraparticle diffusion. Environ. Sci. Technol. v. 25, pp. 1237-1249.
Bard, Y. 1974. Nonlinear parameter estimation. Academic Press, Inc., New York.
119
-------
Bates, D.M., and D.G. Watts. 1988. Nonlinear regression analysis and its applications. Wiley
Interscience, New York.
Battley, E.H. 1987. Energetics of microbial growth. Wiley Interscience, New York.
Beck, J.V., and KJ. Arnold. 1977. Parameter estimation in engineering and science. Wiley
Interscience, New York.
Belyaev, S. S., and M.V. Ivanov. 1983. Bacterial methanogenesis in underground waters.
Environ. Biogeochem. v. 35, pp. 273-280.
Berry, D.F., AJ. Francis, and J.M. Bollag. 1987a. Microbial metabolism of homocyclic and
heterocyclic aromatic compounds under anaerobic conditions. Microbiol. Rev. v. 51, pp.
43-59.
Berry, D.F., E.L. Madsen, and J.M. Bollag. 1987b. Conversion of indole to oxindole under
methanogenic conditions. Appl. Environ. Microbiol. v. 53, pp. 180-182.
Boyd, S.A., D.R. Shelton, D. Berry, and J.M. Tiedje. 1983. Anaerobic biodegradation of
phenolic compounds in digested sludge. Appl. Environ. Microbiol. v. 46, pp. 50-54.
Bradford, M.M. 1976. A rapid and sensitive method for the quantitation of microgram
quantities of protein utilizing the principle of protein dye-binding. Analyt. Biochem. v. 72,
pp. 248-254.
Bryant, M.P., L.L. Campbell, and C.A. Reddy. 1977. Growth of Desulfovibrio in lactate or
ethanol media low in sulfate in association with ^-oxidizing methanogenic bacteria.
Appl. Environ. Microbiol. v. 33, pp. 1162-1169.
Brock, T.D., and K. O'Dea. 1977. Amorphous ferrous sulfide as a reducing agent for culture
of anaerobes. Appl. Environ. Microbiol. v. 33, pp. 254-256.
Brunner, W., and D.D. Focht. 1984. Deterministic three-half-order kinetic model for microbial
degradation of added carbon substrates in soil. App. Environ. Microbiol. v. 47, pp. 167-
172.
Brusseau, M.L., R.E. Jessep, and P.S.C. Rao. 1989. Modeling the transport of solutes
influenced by multiprocess nonequilibrium. Wat. Resour. Res. v. 25, pp. 1971-1987.
Brusseau, M.L., and P.S.C. Rao. 1989. The influence of sorbate-organic matter interactions
on sorption nonequilibrium. Chemosphere v. 18, pp. 1691-1706.
Chmielowski, J.A., A. Grossman, and S. Labuzek. 1965. Biochemical degradation of some
phenols during methane fermentation. Zesz. nauk. Politech. Slask., Inz. Sanit. v. 8, pp.
97-122.
Chiou, L.T., L J. Peters, and V.H. Freed. 1979. A physical concept of soil-water-equilibria of
nonionic organic compounds. Science v. 206, pp. 831-832.
Constantinides, A. 1987. Applied numerical methods with personal computers. McGraw-Hill,
New York.
120
-------
Crittenden, J.C., N.J. Hutzler, D.G. Geyer, J.L. Oravitz, and G. Friedman. 1986. Transport of
organic compounds with saturated groundwater flow: model development and parameter
sensitivity. Wat. Resour. Res. v. 22, pp. 271-284.
Dean, J.A. 1973. Solubilities, pp. 10-3 to 10-5 in Lange's Handbook of Chemistry, 13 edition,
McGraw-Hill, New York.
De Smedt, F., and P.J. Wierenga. 1979. A generalized solution for solute flow in soils with
mobile and immobile water. Wat. Resour. Res. v. 15, pp. 1137-1141.
Dills, S.S., A. Apperson, M.R. Schmidt, and M.H. Saier, Jr. 1980. Carbohydrate transport in
bacteria. Microbiol. Rev. v. 44, pp. 385-418.
Doddema, H.J., and G.D. Vogels. 1978. Improved identification of methanogenic bacteria by
fluorescence microscopy. Appl. Environ. Microbiol. v. 36, pp. 752-754.
Ehrlich, G.G., D.F. Goerlitz, J.H. Bourell, G.V. Eisen, and E.M. Godsy. 1981. Liquid
chromatographic procedure for fermentation product analysis in the identification of
anaerobic bacteria. Appl. Environ. Microbiol. v. 42, pp. 878-885.
Evans, W.C. 1977. Biochemistry of the bacterial catabolism of aromatic compounds in
anaerobic environments. Nature v. 270, pp. 17-22.
Evans, W.C., and G. Fuchs. 1988. Anaerobic degradation of aromatic compounds. Ann. Rev.
Microbiol. v. 42, pp. 289-317.
Fedorak, P.M., and S.E. Hrudey. 1984. Batch anaerobic methanogenesis of phenolic coal
conversion wastewater. Wat. Sci. Tech. v. 17, pp. 143-154.
Fedorak, P.M., and S.E. Hrudey. 1986. Anaerobic treatment of phenolic coal conversion
wastewater in semicontinuous cultures. Wat. Res. v. 20, pp. 113-122.
Galli, R. 1987. Biodegradation of dichloromethane in waste water using a fluidized bed
bioreactor. Appl. Microbiol. Biotechnol. v. 27, pp. 206-213.
Godsy, E.M. 1980. Isolation of Methanobacteriwn bryantii from a deep aquifer by using a
novel broth-antibiotic disk method. Appl. Environ. Microbiol. v. 39, pp. 1074-1075.
Godsy, E.M., D.F. Goerlitz, and G.G. Ehrlich. 1983. Methanogenesis of phenolic compounds
by a bacterial consortium from a contaminated aquifer in St. Louis Park, Minnesota. Bull.
Environ. Contam. Toxicol. v. 30, pp. 261-268.
Godsy, E. M., D.F. Goerlitz, and D. Grbid-Galid. 1989. Transport and degradation of water-
soluble creosote-derived compounds, pp. 112-143, in Allen, D.T., (ed.), Intermedia
Pollutant Transport: Modeling and Field Measurements, Plenum Publishing, New York.
Godsy, E. M., D.F. Goerlitz, and D. Grbic'-Galic'. 1992. Methanogenic biodegradation of
creosote contaminants in natural and simulated ground-water ecosystems. Ground Water,
v. 30, pp. 232-242.
121
-------
Godsy, E.M., and D. Grbid-Galid. 1989. Biodegradation pathways for benzothiophene in
methanogenic microcosms, pp. 559-564, in Mallard, G.E., and S.E. Ragone (eds.), U.S.
Geological Survey Toxic Substances Hydrology program-Proceedings of the Technical
Meeting, Phoenix, Arizona, September 26-30,1988, U.S. Geological Survey Water
Resources Investigations Report 88-4220.
Goerlitz, D.F. 1984. A column technique for determining sorption of organic solutes on the
lithologic structure of aquifers. Bull. Environ. Contain. Toxicol. v. 32, pp. 37-44.
Goerlitz, D.F., D.E. Troutman, E.M. Godsy, and BJ. Franks. 1985. Migration of wood-
preserving chemicals in contaminated groundwater in a sand aquifer at Pensacola, Florida.
Environ. Sci. Technol. v. 19, pp. 955-961.
Grbid-Galid, D. 1989. Microbial degradation of homocyclic and heterocyclic aromatic
hydrocarbons under anaerobic conditions. Dev. Ind. Microbiol. v. 30, pp. 237-253.
Grbid-Galic, D. 1990. Anaerobic microbial transformation of nonoxygenated aromatic and
alicyclic compounds in soil, subsurface and freshwater sediments, pp. 117-189, in Bollag,
J.M., and G. Stotzky (eds.), Soil Biochemistry, v. 6, Marcel Dekker, New York.
Grbid-Galid, D., and T.M. Vogel. 1987. Transformation of toluene and benzene by mixed
methanogenic cultures. Appl. Environ. Microbiol. v. 53, pp. 254-260.
Haggblom, M.M., M.D. Rivera, I.D. Bossert, I.E. Rogers, and L.Y. Young. 1990. Anaerobic
biodegradation of para-cresol under three reducing conditions. Microbiol. Ecol. v. 20, pp.
141-150.
Hansch, C., and A.J. Leo. 1979. Substituent constants for correlation analysis in chemistry
and biology. John Wiley and Sons, New York.
Harvey, R.W., R.L. Smith, and L.H. George. 1984. Effect of organic contamination upon
microbial distributions and heterotrophic uptake in a Cape Cod, Mass, aquifer. Appl.
Environ. Microbiol. v. 48, pp. 1197-1202.
Healy, J.B., Jr., and L.Y. Young. 1978. Catechol and phenol degradation by a methanogenic
population of bacteria. Appl. Environ. Microbiol. v. 35, pp. 216-218.
Hashimoto, I., K.B. Deshpande, and H.C. Thomas. 1964. Peclet numbers and retardation
factors for ion exchange columns. Ind. Eng. Chem. Fundamen. v. 3, pp. 213-218.
Helfferich, F. 1962. Ion exchange. McGraw-Hill, New York.
Hobbie, J.E., R.J. Daley, and S. Jasper. 1977. Use of Nucleopore filters for counting bacteria
by fluorescence microscopy. Appl. Environ. Microbiol. v. 33, pp. 1225-1228.
Holm, P.E., P.H. Nielson, H. Albrechtsen, and T.H. Christensen. 1992. Importance of
unattached bacteria and bacteria attached to sediment in determining potentials for
degradation of xenobiotic organic contaminants in an aerobic aquifer. Appl. Environ.
Microbiol. v. 58, pp. 3020-3026.
122
-------
Holdeman, L.V., and W.E.C. Moore. 1972. Anaerobe laboratory manual (2nd. edition).
Virginia Polytechnic Institute and State University, Blacksburg, Virginia.
Hungate, R.E. 1969. A roll tube method for the cultivation of strict anaerobes, pp. 117-132,
in Norris, J.R., and D.W. Ribbons (eds.), Methods in Microbiology, Academic Press,
New York.
Isnard, P., and S. Lambert. 1989. Aqueous solubility and n-octanol/water partition coefficient
correlations. Chemosphere v. 18, pp. 1837-1853.
Jacob, C.E., and H.H. Cooper, Jr. 1940. Report on the ground-water resources of the
Pensacola area in Escambia County, Florida. U.S. Geological Survey Open File Report
FL-40001,89p.
Karickoff, S.W., D.S. Brown, and T.A. Scott. 1979. Sorption of hydrophobic pollutants on
natural sediments. Wat. Res. v. 13, pp. 241-248.
Khan, A.U., and W.A. Jury. 1990. A laboratory study of the dispersion scale effect in column
outflow experiments. J. Contam. Hydrol. v. 5, pp. 119-131.
Knoll, G., and J. Winter. 1989. Degradation of phenol via carboxylation to benzoate by a
defined, obligate syntrophic consortium of anaerobic bacteria. Appl. Microbiol.
Biotechnol. v. 30, pp. 318-324.
Kompala, D.S., D. Ramkrishna, N.B. Jansen, and G.T. Tsao. 1986. Investigation of bacterial
growth on mixed substrates: experimental evaluation of cybernetic models. Biotechnol.
Bioeng. v. 28, pp. 1044-1055.
Kuhn, E.P., J. Zeyer, P. Eicher, and R.P. Schwarzenbach. 1988. Anaerobic degradation of
alkylated benzenes in denitrifying laboratory columns.. Appl. Microbiol. Biotechnol. v. 54,
pp. 490-496.
Lawrence, A.W., and P.L. McCarty. 1970. Unified basis for biological treatment design and
operation. J. Sanitary Eng. Div., ASCE , v. 96, pp. 757-778.
Liley, P.E., R.C. Reid, and E. Buck. 1984. Physical and chemical data, pp. 3-97 to 3-100 in
Perry's Chemical Engineers' Handbook. McGraw-Hill, New York.
Mackay, D. 1980. Solubility, partition coefficients, volatility and evaporation rates. Reactions
and processes, pp. 31-46 in The Handbook of Environmental Chemistry, v. 2, pt. A,
Springer-Verlag, Berlin.
Mackay, D., A. Bora, D.W. Chan, and W.Y. Shiu. 1982. Vapor pressure correlations for low
volatility environmental chemicals. Environ. Sci. Technol., v. 16, pp. 645-649.
Mackey, J.K. 1987. The influence of microbial dynamics on the steady state biodegradation of
2-chlorophenol in continuous culture. MENGR Report, College of Engineering, Clemson
University, Clemson, South Carolina.
123
-------
Madsen, E.L., and J.M. Bollag. 1989. Pathway of indole metabolism by a denitrifying
microbial community. Arch. Microbiol. v. 151, pp. 71-76.
Mahmood, R.J., and R.C. Sims. 1986. Mobility of organics in land treatment systems. J.
Environ. Eng., ASCE , v. 112, pp. 236-245.
Matin, A., and H. Veldkamp. 1978. Physiological basis of the selective advantage of a
Spirillum sp. in a carbon limited environment. J. Gen. Microbiol. v. 105, pp. 187-194.
Mattraw, H.C., Jr., and B.F. Franks. 1986. Description of hazardous-waste research at a
creosote works, Pensacola, Florida, pp. 1-8, in Mattraw, H.C., Jr., and B.F. Franks (eds.),
Movement and fate of creosote wastes in ground water, Pensacola, Florida. U.S.
Geological Survey Toxic Waste-Ground Water Contamination Program. U.S. Geological
Survey Water-Supply Paper 2285.
McCarty, P.L. 1971. Energetics and bacterial growth, pp. 495-531, in Faust, S.D., and J.V.
Hunter (eds.), Organic compounds in aquatic environments, Marcel Dekker, New York.
Mihelcic, J.R., and R.G. Luthy. 1988. Degradation of polycyclic aromatic hydrocarbons
compounds under various redox conditions in soil-water systems. Appl. Environ.
Microbiol. v. 54, pp. 1182-1187.
Mink, R.W., and P.R. Dugan. 1977. Tentative identification of methanogenic bacteria by
fluorescence microscopy. Appl. Environ. Microbiol. v. 33, pp. 713-717.
Monod, J. 1949. The growth of bacterial cultures. Ann. Rev. Microbiol. v. 3, pp. 371-394.
Munz, C., and P.V. Roberts. 1986. Effects of solute concentrations and cosolvents on the
aqueous activity coefficient of halogenated hydrocarbons. Environ. Sci. Technol. v. 20,
pp. 830-836.
Neufeld, R.D., J.D. Mack, and J.P. Strakey. 1980. Anaerobic phenol biokinetics. J. Wat. Poll.
Control Fed. v. 52, pp. 2367-2377.
Nester, E.W., C.E. Roberts, M.E. Lindstrom, N.N. Pearsall, and M.Y. Nester. 1983.
Microbiology, 3rd edition, Saunders, New York
Nkedi-Kizza, P., J.W. Biggar, M.Th. van Genuchten, PJ. Wierenga, H.M. Selim, J.M.
Davidson, and D.R. Nielson. 1983. Modeling tritium and chloride 36 transport through an
aggregated oxisol. Wat. Resour. Res. v. 19, pp. 691-700.
Nkedi-Kizza, P., J.W. Biggar, H.M. Selim, M.Th. van Genuchten, PJ. Wierenga, J.M.
Davidson, and D.R. Nielson. 1984. On the equivalence of two conceptual models for
describing ion exchange during transport through an aggregated oxisol. Wat. Resour.
Res. v. 20, pp. 1123-1130.
Novotony, M., J.W. Strand, S.L. Smith, D. Wiesler, and F.J. Schwende. 1981. Compositional
studies of coal tar by capillary gas chromatography-mass spectrometry. Fuel v. 60, pp.
213-220.
124
-------
Ogata, A., and R.B. Banks. 1961. A solution of the differential equation of longitudinal
dispersion in porous media. U.S. Geological Survey Professional Paper 411-A,
Washington, D.C.
Oravitz, J.L. 1984. Transport of trace organics with one-dimensional saturated flow:
mathematical modeling and parameter sensitivity analysis. M.S.C.E. Thesis, Department
of Civil Engineering, Michigan Technological University, Houghton, Michigan.
Pereira, W.E., C.E. Rostad, D.M. Updegraff, and J.L. Bennett. 1987. Fate and movement of
azaarenes in an aquifer contaminated by wood-treatment chemicals. Environ. Tox. Chem.
v. 6, pp. 6163-6176.
Pereira, W.E., C.E. Rostad, D.M. Updegraff, and J.L. Bennett. 1988. Microbial
transformations of azaarenes in creosote-contaminated soil and ground water: Laboratory
and field studies. International Conference on Water and Wastewater Microbiology,
Newport Beach, California, pp. 5-1 to 5-7.
Priddle, M.W., and R.E. Jackson. 1991. Laboratory column measurement of VOC
retardation factors and comparison with field values. Ground Wat. v. 29, pp. 260-266.
Puhakka, J.A., M. Salkinoja-Salonen, F.F. Ferguson, and M.M. Benjamin. 1990. Carbon
flow in acetotrophic enrichment cultures from pulp mill effluent treatment. Wat. Res. v.
24, pp. 515-519.
Reid, R.C., J.M. Prausnitz, and B.E. Poling. 1985. The properties of gases and liquids.
McGraw-Hill, New York.
Roberts, D.J., P.M. Fedorak, and S.E. Hrudey. 1986. Comparison of the fates of the methyl
carbons of m-cresol andp-cresol in methanogenic consortia. Can. J. Microbiol. v. 33, pp.
335-338.
Roberts, D.J., P.M. Fedorak, and S.E. Hrudey. 1990. CO2 incorporation and 4-hydroxy-2-
methylbenzoic acid formation during anaerobic metabolism of m-cresol by a
methanogenic consortium. Appl. Environ. Microbiol. v. 56, pp. 472-478.
Roberts, P.V., J. Schreiner, and G.D. Hopkins. 1982. Field study of organic water quality
changes during groundwater recharge in the Palo Alto Baylands. Wat. Res. v. 16, pp.
1025-1035.
Roberts, P.V., M.N. Goltz, R.S. Summers, J.C. Crittenden, and P. Nkedi-Kizza. 1987. The
influence of mass transfer on solute transport in column experiments with an aggregated
soil. J. Contam. Hydrol. v. 1, pp. 375-393.
Robinson, J.A. 1985. Nonlinear regression analysis in microbial ecology. Adv. Microbial Ecol.
v. 8, pp. 61-114.
125
-------
Sarak Genther, B.R., G.T. Townsend, and PJ. Chapman. 1989. Anaerobic transformations of
phenol to benzoate via para-carboxylation: use of fluorinated analogs to elucidate the
mechanism of transformation. Biochem. Biophys. Res. Commun. v. 162, pp. 945-951.
Schalf, M.R., J.F. McNabb, W.J. Dunlap, R.L. Crosby, and J.S. Fryberger. 1981. Manual of
ground-water sampling procedures. NWWA-EPA Series: National Water Well
Association, Worthington, Ohio.
Schink, B., A. Brune, and S. Schnell. 1992. Anaerobic degradation of aromatic compounds,
pp. 219-242, in Winkelmann, G. (ed.), Microbial degradation of natural products, Verlag
Chemie, Weinheim, FRG.
Schink, B., and A. Tschech. 1988. Fermentative degradation of aromatic compounds, pp.
213-226, in Hagedorn, S.R., R.S. Hanson, and D.A. Kunz (eds.), Microbial metabolism
and the carbon cycle, Harwood Academic Publishers, New York.
Schwarzenbach, R.P., and J. Westall. 1981. Transport of nonpolar organic pollutants from
surface water to ground water: laboratory sorption studies. Environ. Sci. Technol. v. 15,
pp. 1350-1367.
Skougstad, M.W., M.J. Fishman, L.C. Friedman, D.E. Erdman, and S.S. Duncan. 1989.
Methods for determination of inorganic substances in water and fluvial sediments. U.S.
Geological Survey Techniques of Water-Resources Investigation, Book 5, Chapter Al.
Smolenski, W.J., and J.M. Sulfilta. 1987. Biodegradation of cresol isomers in anoxic aquifers.
Appl. Environ. Microbiol. v 53, pp. 710-716.
Stanier, R.Y., N.J. Palleroni, and M. Doudoroff. 1966. The aerobic pseudomonads: a
toxonomic study. J. Gen. Microbiol. v. 43, pp. 159-271.
Stull, D.R., E.F. Westrum, Jr., and G.C. Sinke. 1969. The chemical thermodynamics of
organic compounds. Marcel Dekker, New York.
Stumm, W., and JJ. Morgan. 1981. Aquatic chemistry. Wiley Interscience, New York.
Sudicky, E.A., J.A. Cherry, and E.O. Frind. 1983. Migration of contaminants in groundwater
at a landfill: a case study, 4. A natural-gradient dispersion test. J. Hydrol. v. 63, pp. 81-
108.
Suidan, M.T., I.N. Najm, J.T. Pfeffer, and Y.T. Wang. 1989. Anaerobic biodegradation of
phenol: inhibition kinetics and system stability. J. Environ. Eng., ASCE v. 114, pp. 1359-
1376.
Tarvin, D., and A.M. Buswell. 1934. The methane fermentation of organic acids and
carbohydrates. J. Amer. Chem. Soc. v. 56, pp. 1751-1755.
Templeton, L.L., and C.P.L. Grady, Jr. 1988. Effects of culture history on the determination
of biodegradation kinetics by batch and fed-batch techniques. J. Wat. Poll. Control Fed. v.
60, pp. 651-658.
126
-------
Thauer, R.K., K. Jungermann, and K. Decker. 1977. Energy conservation in chemotrophic
anaerobic bacteria. Bacteriol. Rev. v. 41, pp. 100-180.
Vogel, T.M., and D. Grbic-Galic". 1986. Incorporation of oxygen from water into toluene and
benzene during anaerobic fermentative transformation. Appl. Environ. Microbiol. v. 52,
pp. 200-202.
van Genuchten, M.Th. 1981. Non-equilibrium transport parameters from miscible
displacement experiments. U.S. Department of Agriculture, U.S. Salinity Laboratory,
Research Report No. 119, 88 p.
van Genuchten, M.Th., and PJ. Wierenga. 1976. Mass transfer studies in sorbing porous
media. I. Analytical solution. Soil Sci. Soc. Amer. Proc. v. 40, pp. 473-480.
Vollhardt, K.P.C. 1987. Organic chemistry, W.H. Freeman and Co., New York, pp. 1179-
1229.
Westall, J.C., C. Leuenberger, and R.P. Schwarzenbach. 1985. Influence of pH and ionic
strength on the aqueous-nonaqueous distribution of chlorinated phenols. Environ. Sci.
Tech. v. 19, pp. 193-198.
Wilson, B.H., B. Bledsoe, and D. Kampbell. 1987. Biological processes occurring at an
aviation gasoline spill site, pp. 125-137, in Everett, R.C., and D.M. McKnight (eds.),
Chemical quality of water and the hydrologic cycle. Lewis Publishers, Inc., Chelsea,
Michigan.
Wilson, J.T., J.F. McNabb, D.L. Balkwill, and W.C. Ghiorse. 1983. Enumeration and
characterization of bacteria indigenous to a shallow water-table aquifer. Ground Wat. v.
21, pp. 134-142.
Wolin, E.A., M.J. Wolin, and R.S. Wolfe. 1963. Formation of methane by bacterial extracts.
J. Biol. Chem. v. 238, pp. 2882-2886.
Young, L.Y. 1984. Anaerobic degradation of aromatic compounds, pp. 487-523, in Gibson,
D.T. (ed.), Microbial degradation of organic compounds. Marcel Dekker, New York.
Young, L.Y., and M.D. Rivera. 1985. Methanogenic degradation of four phenolic
compounds. Wat. Res. v. 19, pp. 1325-1332.
Zachara, J.M., C.C. Ainsworth, C.E. Cowan, and B.L. Thomas. 1987. Sorption of binary
mixtures of aromatic nitrogen heterocyclic compounds on subsurface materials. Environ.
Sci. Technol. v. 21, pp. 397-402.
Zachara, J.M., C.C. Ainsworth, LJ. Felice, and C.T. Resch. 1986. Quinoline sorption to
subsurface materials: role of pH and retention of the organic cation. Environ. Sci.
Technol. v. 20, pp. 620-627.
Zeikus, J.G. 1977. The biology of methanogenic bacteria. Bacteriol. Rev. v. 41, pp. 514-541.
127
-------
Zeyer, J., E.P. Kuhn, and R.P. Schwarzenbach. 1986. Rapid microbial minerization of toluene
and 1,3-dimethylbenzene in the absence of molecular oxygen. Appl. Environ. Microbiol.
v. 52, pp. 944-9447.
128
-------
APPENDIX A
BIOMASS DETERMINATIONS
Calculation OfBiomass From Total Protein Determinations
Biomass determinations were calibrated from total protein determinations of suspensions
of aquifer-derived microorganisms grown in a defined mineral salts solution described by
Stanier et al. (1966) containing 2.0 g sodium acetate-2 r^O per L as the sole carbon and
energy source. Twenty-five mL aliquots of the cell suspensions were dried at 105*C for
dry-weight biomass determinations. An equal number of samples were centrifuged at 5,000 x
g for 10 min to remove the biomass and 25.0 mL of the resulting supernatant was dried at
105'C for determination of the dry weight of the mineral salts. The dry-weight biomass
determinations for the cell suspension are given in Table A-l.
The dry-weight concentration of the biomass may be obtained by subtracting the dry-
weight concentration of the mineral salts from the dry-weight concentration of the cell
biomass plus mineral salts and results in a value of 70 mg/L dry-weight biomass or 3.5 x
101 J cells/L, assuming that the dry weight of a microbial cell is equal to 2 x 10'10 mg/cell.
Total Protein Determination Using Bovine Serum Albumin
Total protein was determined by staining with Coomassie brilliant blue G-250 and
calibrated using bovine serum albumin (BSA) as described by Bradford (1976). The protein
concentration in the standards was determined spectrophotometrically at 595 nm. The
calibration curve for BSA concentration is shown in Figure A-l.
The total protein concentration of the cell suspension were determined by first
centrifuging 25.0 mL aliquots at 5,000 x g for 10 min. The resulting cell pellet was treated
with 5.0 mL of 0.66 N NaOH until lysis of the cell pellet (minimum of 48 hr at
25'C) and analyzed in the same manner as the BSA standards. The total protein for the cells
suspension are given in Table A-2.
The average protein content of the cell suspension was determined to be 28.8 mg/L.
This is in reasonable agreement with the supposition that bacteria are approximately 50
percent protein (Nester et al., 1983) and gives a biomass of 58 mg/L for the cell suspension
as compared to 70 mg/L by dry-weight analysis. No traces of protein were found in samples
of the mineral salts solution.
129
-------
Recovery Of Protein Added To Sediment Samples
The total protein concentration was separately determined for 10.0 g aliquots of the aquifer
sediment and a cell suspension prepared as above. Recovery of protein added to aquifer
sediment was determined by combining a known cell suspension to sediment samples and
analyzing for total protein. Several 25 mL aliquots of liquid cell suspension were centrifuged
at 5,000 x g for 10 min and the cell pellet was treated with 5.0 mL of 0.66 N NaOH. The
pellet was immediately resuspended in the NaOH and several of the suspensions were diluted
with 0.66 N NaOH so that a number of cell concentrations could be added to the aquifer
sediments. Five mL of cell suspensions diluted to 1:2 and 1:5 were added to sediment
samples and allowed to digest on an orbital shaker at 250 rpm for 14 days at 25'C. Protein
analyses of the cell suspension used for the addition of protein to the aquifer sediments are
given in Table A-3.
The results of this portion of the study demonstrate that the protein determination only
slightly underestimates the actual amount of protein added to the protein already present on
the aquifer material. The degree of underestimation may be proportional to the amount of
protein added - the greater the amount of protein added, the greater the amount total protein
is underestimated. The standard error of the y estimate in the protein biomass determinations
is quite large and the differences in protein before and after growth in the microcosms are
quite small. This finding suggests that a higher precision method should be found or
developed and used for these purposes.
130
-------
PQ
CO
§
PQ
200
150
100
50
0
0.1
0.2
OD 595 nm
0.3
0.4
Figure A-l. Optical density of bovine serum albumin at 595 nm. The error bars represent
the data scatter in five replicates. The dotted line represents a linear regression of the
data (y = 522.73 x). Standard error of the Y estimate is 8.60.
131
-------
Table A-l. Dry weight determinations of ground-water derived microbial cultures and
mineral salts dried at 105"C.
Aliquot Pan
number weight,
g
Cell
1
2
3
4
Suspension Plus
1.3692
1.3715
1.3823
1.3530
5 1.3814
Mineral Salts Onlv
6
7
8
9
10
1.3232
1.3411
1.3513
1.3312
1.3616
Pan weight plus
dry weight of 25
mL biomass
and/or mineral
salts solution,
g
Mineral Salts
1.4595
1.4575
1.4684
1.4433
1.4701
1.4118
1.4260
1.4366
1.4175
1.4491
Dry weight of Dry weight Average
25 ml cell of 25 ml of weight of
suspension plus mineral salts suspension
mineral salts solution, and/or mineral
solution, g salts, mg/L
g
0.0903 —
0.0860 —
0.0861 — 3530
0.0903 —
0.0887 —
— 0.0886
— 0.0849
— 0.0853 3461
— 0.0863
— 0.0875
Table A-2. Protein determinations of ground water-derived microbial cultures.
Aliquot
number
1
2
3
4
5
6
7
8
ODat
595 nm
0.312
0.305
0.229
0.269
0.024
0.033
0.024
0.028
Dilution
—
—
—
—
1:10
1:10
1:10
1:10
(J.g protein per
mLNaOH
(dilution
factor
included)
163.1
159.4
119.7
140.6
125.5
172.5
125.5
146.4
\ig protein
in cell
pellet
815.5
797.2
598.5
703.1
627.3
862.5
627.3
731.8
mg
protein/L
32.6
31.9
23.9
28.1
25.1
34.5
25.1
29.3
Average
protein of
cell
suspension
± 95% CI,
mg/L
28.8 ±7.8
132
-------
Table A-3. Protein analyses of cell suspension used for protein recovery studies.
Aliquot number
1
2
3
4
OD at 595 nm
0.154
0.122
0.170
0.148
u,g protein per mL
NaOH
80.5
63.8
88.9
77.4
Average |Xg protein per
mLNaOH
77.6
Table A-4. Analyses of the recovery of protein added to aquifer sediments.
Aliquot
number
1
2
3
4
5
6
7
8
9
Protein
added, Hg/mL
NaOH
—
— •
—
38.8
38.8
38.8
15.5
15.5
15.5
OD at 595
nm
0.165
0.173
0.166
0.199
0.219
0.213
0.179
0.202
0.186
Analyzed
protein
NaOH
86.3
90.4
86.8
104.0
114.5
111.3
93.6
105.6
97.2
Average Protein
protein recovery,
pig/mL %1
NaOH
87.8 —
109.9 86.8
98.8 95.6
Value obtained by dividing the amount of protein obtained after 14 days on the shaker by the
amount of protein initially added plus the protein determined to be present on the aquifer
material.
133
-------
APPENDIX B
FREE ENERGY CALCULATIONS FOR WSF COMPOUNDS
The general electron donor oxidation half reaction that describes the transfer of one mole
of electrons during mineralization under aqueous conditions coupled with the electron
acceptor half reaction may be used to determine the Gibbs free energy of a reaction (McCarty,
1971).
The free energy liberated from a reaction can be estimated by subtracting the free energy
of formation of the reactants from the that of products under standard conditions
(concentrations of 1 mol/L or pressures of 1 atm).
AG° = ISAG - Z8AG°(reactants) (B-l)
where AG° = Gibbs free energy of a reaction under standard conditions, 5 = stoichiometric
coefficients, and AGf = the free energy of formation under standard conditions. For the
balanced half reaction of the hydrolysis of phenol,
^C6H60 +^H2O ->^C02 + H+ + e-
equation B-l at a pH of 7.0 becomes
AG°' = [^ AG° C02 + AG° H+ + AG° e'] - [^ AG° C6H60 + ^ AG° H2o]
where AG° is the Gibbs free energy of a reaction under standard conditions except at pH of
7. By convention, AGf for the e~ is zero whereas that of the proton is - 40.4 kJ/mol for a pH
of 7.0. Since the values for CO2 and I^O are readily available in many handbooks (e.g., Stull
et al., 1969 or Thauer et al, 1977), the problem becomes one of obtaining the AGf ^ for the
creosote-derived compounds under aqueous conditions. The AGf ^, or for that matter,
many of the physical constants for these compounds, have not been determined or are not
readily available in the literature, but may be estimated by a number of different methods.
The AG°/ ) for the ideal gas state for many of the WSF compounds are available in Stull
et al. (1969). For those compounds not listed in Stull et al. (1969), the AGf (g) may be
estimated using the method of Jobak given in Reid et al. (1985).
134
-------
The value for AGf / % may be related to AGf , ^ by the Henry's constant in following
expressions
where Pa{m is the partial pressure of the compound (atm) in the gas phase; H is the Henry's
Law Constant (atm-L/mol); and 5W is the solubility in water (mol/L). The AGf /a£^ of a
particular compound may be computed from the AGf / ^ using the dimensionless Henry's
Constant as follows
CB-3)
where R = 8.314 J/mol-K, Tk = temperature (K), and //c = dimensionless Henry's Constant
tj
equivalent to - where, in this case, R = 0.082054 atm-L/mol-tf.
RTk
Mackay et al. (1982) demonstrated that an estimate of the vapor pressure at a particular
temperature (7^) for a particular compound is related to the melting (rm) and boiling (7^)
points as follows
~
Tk
This method allows the estimation of the Patm for compounds where the value has not
been experimentally determined or is not readily available in the literature.
The solubility in water (Sw, mg/L) of a particular compound where the value has not been
experimentally determined or is not readily available may be estimated from the K^ and Tm
using the relationship proposed by Isnard and Lambent (1989)
log Sw = 5.90 - 1.1 8/s^ - 0.0048(Tm - 293) (B-5)
The values for the appropriate constants to determine the AGf , ^ for the WSF
compounds are given in Table B-l.
135
-------
Table B-l
Phenol
2-Methylphenol
3-Methylphenol
4-Methylphenol
Indole
Oxindole
Quinoline
Isoquinoline
2(lH)-quinolinone
1 (2H)-isoquinolinone
Benzothiophene
Dibenzothiophene
Benzofuran
Dibenzofuran
Indene
Naphthalene
Acenaphthen
Molecular
weight
94.11
108.14
108.14
105.14
117.15
133.15
129.16
129.16
145.16
145.16
134.20
184.26
118.14
168.20
116.60
128.19
154.21
log
KOW
1.48
1.93
1.98
2.00
2.10
0.85
2.03
2.08
1.26
0.58
3.12
4.38
2.67
4.12
2.92
3.45
3.92
AG°(g)
kJ/mol
-50.42 b
-33.07 b
-40.54 b
-30.88 b
237.30 b
231.67
293.50 b
279.70
174.56
174.56
257.54
388.48
103.10 b
182.20 b
233.97 b
224. 10 b
260.20 b
Vapor pressure,
atm
8.19 x 10^ c
1.44 x ID"4 c
1.83 x 10-4 c
9.48 x lO'5 c
1.56 x 10"5 b
6.95 x ID'5 d
7.66 x ID'5 b
l.WxlO^b
4.64 x ID'7 d
3.46 x 10'7 d
2.80 x 10-4 d
2.22 x 10'7 d
1.23 x 10-2 d
1.29b
1.41xlO-3b
1.07 x 10-4 b
3.40 x lO"65
Solubility,
mg/L
82000 a
25000 a
5000 a
18000 a
1862 b
25834 e
6600b
5700 b
3784 e
20796 e
155 e
2.4 e
562 e
10.0 b
88.9 b
31.7 b
39 .Ob
Solubility,
mol/L
871
231
46
166
16
19400
51
44
26
143
1.2
0.013
4.45
0.059
0.762
0.247
0.253
H,
atm-L/mol
9.40 x 10'7
6.23 x 10'7
3.96 x lO'6
5.69 10'7
9.81 x lO'7
3.58 x lO'7
1.50 x 10'6
2.54 x 10'6
1.78 x 10"8
2.42 x 10'9
2.43 x 10-4
1.70 x ID'5
2.59 x 10'3
21.7
1. 83xlO-3
4.33 x 10-4
1.34 x ID'5
He,
(-)
3.84 x 10'8
2.55 x 10'8
1.62xlO-7
2.33 x 10'8
4.01 x 10'8
1.46xlO-g
6.13 x 10'8
1.04X10'7
7.28 x 10- 10
9.88 x 10- u
9.93 x lO'6
6.96 x lO'7
1.06X10-4
8.89 x 10- l
7.57xlO-5
1.77xlO-5
5.50 xlO'7
AGf(aq)
kJ/mol
-92.72
-76.39
-79.28
-74.42
195.10
186.98
252.35
239.86
122.43
117.48
229.00
353.35
80.42
181.91
210.46
196.99
224.49
a=Liley etal. (1984) b=API Monograph Series c= Dean (1973) d=Mackay eial (1982) e=Isnard and Lambert (1989)
-------
APPENDIX C
THEORETICAL ENERGY YIELDS
The AG^ for the oxidation of the compounds studied to Cfy and CO2 are calculated
below. In addition, the energy consumed or liberated during the oxidation of the nitrogen
heterocyclic compounds coupled with H2 and CH^ production is also calculated.
Phenol
Energy half reactions:
donor:
^ CfptO + 55 H20 -» 5Jj CO2 + H+ + e- AG^ = -27.9 kJ/mol e~
acceptor:
I CH4 + \ H2O -> I CO2 + H+ + e- AG(°q} = -24. 1 kJ/mol e~
Adding the two half-reactions and normalizing to phenol gives
+ 4 H20 -» 2.5 C02 + 3.5 CH4 AG° = -106 kJ/mol
2-Methylphenol
Energy half-reactions:
donor:
1 1 ^ *7
C7H80 + H20 -> C02 + H+ + e- AG° = -28. 1 kJ/mol c~
acceptor:
CH4 + H20 -» C02 + H+ + e- AG^ = -24. 1 kJ/mol e'
Adding the two half-reactions and normalizing to 2-methylphenol gives
C7H80 + 4.5 H20 ^ 2.75 CO2 + 4.25 CH4 AG° = -136 kJ/mol
137
-------
3-Methylphenol
Energy half-reactions:
donor:
C7H80 + H20 -» CO2 + H+ + e- AG(£j) = -28. 1 kJ/mol e~
CH4 + H2O -> C02 + H+ + e- AG(°^ = -24. 1 kJ/mol e
Adding the two half-reactions and normalizing to 3-methylphenol gives
C7H8O + 4.5 H2O -> 2.75 CO2 + 4.25 CH4 AG(^) = -136 kJ/mol
4-Methylphenol
Energy half-reactions:
donor:
TT C7H8O + r^ H2O -» 24 CO2 + H+ + e- AG^ = - 28.4 kJ/mol e'
acceptor:
| CH4 + \ H2O -»| CO2 + H+ + e- AG(£j) = - 24.1 kJ/mol e'
Adding the two half-reactions and normalizing to 4-methylphenol gives
C7HgO + 4.5 H2O -+ 2.75 CO2 + 4.25 CH4 AG(aa) = ' 146 kJ/mol
138
-------
Oxindole
Energy half-reactions:
donor:
1 1 C O 1 OO
+
34 C8H7NO + 34 H20 ^ 34 C02 + 34 NH + 34 H + e-
,aq) = '34-7
acceptor:
ICH4 + £ H2O -» | C02 + H+ + e- AG^ = -24.1 kJ/mol e"
Adding the two half-reactions and normalizing to oxindole gives
C8H7NO + 6.5 H2O + H+ -» 3.75 CO2 + 4.25 CH4 +
AGfa,^ = -360 kJ/mol
Va4/
2(JtfH2i/mo/mo/ie
Energy half-reactions:
donor;
^-rHNO + ^HO^-^-ro +^-NH+ + ^H+
W ^9"lNU + ™ H2U -» ,,8 CU2 + ™ MH4 + oo H
xaq) = -31.4 kJ/mol e-
acccptor:
| CH4 + ^ H2O -> ICO2 + H"1" + e" AG(°^ = -24.1 kJ/mol e~
Adding the two half-reactions and normalizing to 2(lH)-quinolinone gives
+ 7.5 H2O + H+ -> 4.25 CO2 + 4.75 CH4 + NH^
°\ = -277kJ/mol
139
-------
1 (2H)-Isoquinolinone
Energy half-reactions:
donor:
acceptor:
ICH4 + ^ H2O -> | CO2 + H+ + e" AG^ = -24.1 kJ/mol e~
Adding the two half-reactions and normalizing to l(2H)-isoquinolinone gives
+ 7.5 H2O + H+ -»4.25 CO2 + 4.75 CH4 + NH*
Jj) = -274 kJ/mol
Benzothiophene
Energy half-reactions:
donor:
_2_r TT c , 32 0 16 . J_H Q . J_HQ- . 73 +
-^ Uorlgo T -« n2vJ ~> -^ LAJ2 + -j« H2o + -,« no + -/» rl +6
/^ l£ /Z /Z // /Z
aa-v = -29.2 kJ/mol e"
g CH4 + ^ H2O -^ | CO2 + H+ + e- AG(°q) = -24.1 kJ/mol c~
Adding the two half-reactions and normalizing to benzothiophene gives
+ 7 H2O -4 3.5 CO2 + 4.5 CH4 + 0.5 H2S + 0.5 HS" + 0.5 H+
140
-------
Benzofuran
Energy half-reactions:
donor:
CgHgO + H2O -> CO2 + H+ + e- AG^ = -30.9 kJ/mol e~
CH4 + H2O -» CO2 + H+ + e- AG(°q} = -24. 1 kJ/mol e'
Adding the two half-reactions and normalizing to benzofuran gives
+ 6 H2O -> 3.5 CO2 + 4.5 CH4 AG(°q} = -245 kJ/mol
Naphthalene
Energy half-reactions:
donor:
C]0H8 + H20 -^ C02 + H+ + e- AG(°^ = -27.3 kJ/mol e"
acceptor:
I CH4 + ^ H2O -^ I C02 + H+ + e- AG(°^ = -24. 1 kJ/mol e'
Adding the two half-reactions and normalizing to naphthalene gives
C10H8 + 8 H2O -» 4 CO2 + 6 CH4 AG(°^ = -154 kJ/mol
141
-------
Indene
Energy half-reactions:
donor:
•ft C9H8 + |J H20 -> £ C02 + H+ + e- AG(a% = -28.3 kJ/mol c~
acceptor:
ICH4 + \ H20 -> IC02 + H+ + e- AG(£^ = -24.1 kJ/mol e'
Adding the two half-reactions and normalizing to indene gives
CgHg + 7 H2O -»3.5 CO2 + 5.5 CH4 AG(°q} = -184 kJ/mol
Indole Oxidation Coupled To Hydrogen
Energy half-reactions:
donor:
\ C8H7N + \ H2O -> | C8H7NO + H+ + e" AG(^ = 32.5 kJ/mol e~
acceptor:
\ H2 -» H+ + e- AG(°q} = -40.5 kJ/mol e"
Adding the two half-reactions and normalizing to indole gives
C8H7N -I- H2O -* C8H7NO + H2 AG(°q) = 146 kJ/mol
142
-------
Indole Oxidation Coupled To Methane
Energy half-reactions:
donor:
\ C8H7N + \ H2O -> \ C8H7NO + H+ + e'
acceptor:
I CH4 + £ H2O -> I CO2 + H+ + e-
Adding the two half-reactions and normalizing to indole gives
C8H7N + 0.5 H2O + 0.25 CO2 -» C8H7NO + 0.25 CH4
Quinoline Oxidation Coupled To Hydrogen
Energy half-reactions:
donor:
C9H7N + H2O -> C9H7NO + H+ + e'
= 32.5 kJ/mol e"
= -24. 1 kJ/mol e~
= 113 kJ/mol
= 21.0 kJ/mol e'
H2 -> H4" + e- AG(°«J) = -40.5 kJ/mol e~
Adding the two half-reactions and normalizing to quinoline gives
+ H2O -^ C9H7NO + H2 AG = 1 23 kJ/mol
143
-------
Quinoline Oxidation Coupled To Methane
Energy half-reactions:
donor:
\ C9H7N + ^ H20 -»^ C9H7NO + H+ + e' AG(^ = 21 .0 kJ/mol e~
2* £2*
acceptor:
CH4 + H20 -» C02 + H+ + e- AG(°q) = -24. 1 kJ/mol e'
Adding the two half-reactions and normalizing to quinoline gives
C9H7N + 0.5 H2O + 0.25 CO2 -» C9H7NO + 0.25 CH4 AG(£q) = 90 kJ/mol
Isoquinoline Oxidation Coupled To Hydrogen
Energy half-reactions:
donor:
C9H7N + H2O -^ C9H7NO + H+ + e' AG^) = 17.4 kJ/mol e'
acceptor:
\ H2 -> H+ + e- AG(°q) = -40.5 kJ/mol e'
X«
Adding the two half-reactions and normalizing to isoquinoline gives
C9H7N + H2O -^ C9H7NO + H2 AG(°^ = 1 16 kJ/mol
144
-------
hoquinoline Oxidation Coupled To Methane
Energy half-reactions:
donor:
^ C9H7N + ^ H2O -> ^ C9H7NO + H+ + e~ AG(°q) = 17.4 kJ/mol e'
acceptor:
CH4 + H2O -> CO2 + H+ + e- AG(°^ = -24. 1 kJ/mol e'
Adding the two half-reactions and normalizing to isoquinoline gives
0.5 H2O + 0.25 CO2 -> C9H7NO + 0.25 CH4 AG(a^ = 83 kJ/mol
Non Standard Conditions
The Gibbs free energy change (AG(J,}) for a particular reaction under non-standard
conditions or environmental conditions is dependent upon the temperature and the
concentrations of the reactants and products as follows:
Concentrations and conditions employed in calculations at approximate environmental
concentrations in phenol microcosms were as follows:
Tk = 292 K
[CO2] = PCQ2 = 0.3 atm
[CH4] = PCH4 = 0.7 atm
[H+] = 10-7 mol/L
[Phenol] = 10-3 mol/L
[Products]
l'
AG(aq) = -106,000 J/mol + (8.314 J/mol-K)(292 K) In
AG(aq) = -99,600 J/mol or -100 kJ/mol
The corrections to AG(a(j) for non-standard conditions are apparently not significant and
the values for the Gibbs free energy under standard conditions may be used.
145
-------
APPENDIX D
THEORETICAL BACTERIAL GROWTH YIELD
Theoretical growth yields for the compounds supporting methanogenesis were computed
using the stoichiometric and thermodynamic model presented by McCarty (1971). Maximum
growth yields were calculated with products and reactants at unit activity and at a pH of 7.0.
Pertinent equations and information are given below
^ + AG +^
m T A\Jr T
v ^ V
A =- ^Gr
AGr = AGd - AGa
AGp = AGd -
Ar- »' , D T i
= AG/-an\ + R Tt In
(aq) k
[Reactants]
Y = thermodynamic growth yield (equivalents cells/equivalents substrate).
A = electron equivalent of substrate converted for energy per electron equivalent of
cells synthesized (-).
AGa = free energy of conversion of an electron equivalent of e" acceptor
(kJ/mol e').
AGC = free energy of conversion of an electron equivalent of intermediate to one
electron equivalent of cells, assumed to be 18.8 kJ/mol e~.
AGd = free energy of conversion of an electron equivalent of e~ donor (kJ/mol e').
AGj = free energy of conversion of an electron equivalent of pyruvate (kJ/mol e").
AGn = free energy per electron equivalent of cells for reduction of nitrogen source to
ammonia, equal to zero if ammonia is the nitrogen source (kJ/mol e").
AGp = free energy of conversion of electron equivalent of cell carbon source to
intermediate pyruvate (kJ/mol e').
AGr = free energy per electron equivalent of substrate converted for energy
(kJ/mol e->.
s
AG(aq) = free energy per electron equivalent.
= free energy per electron equivalent at 25"C at pH = 7.0 (kJ/mol e").
146
-------
Y = efficiency of energy transfer, assumed to be 60 percent.
m = constant, equal to +1 when AGp > 0, and -1 when AGp < 0.
R = universal gas constant (8.314 J/mol-K).
T£ = temperature, K.
PHENOL
Energy Half-Reactions:
donor:
H2O -> CO2 + H+ + e'
acceptor:
CH4 + H2O -> CO2
C amputation Of Free Energy Changes:
AGr = -27.9 - (-24.1) = -3.8 kJ/mol e'
AGp = -27.9 - (-35.8) = 7.9 kJ/mol f
Computation ofY:
A = 19.5
Y = 0.05
e" cells
e" phenol
5.65 mg cells per e'
3.36 mg phenol per e"
Y = 0.08
mg cells
mg phenol
2- AND 3-METHYLPHENOL
Energy half-reactions:
donor:
H20 -»
H
AG(°q) = -27.9 kJ/mol e'
= -24.1kJ/mole-
, = -28.1kJ/mole-
147
-------
acceptor:
|CH4 +^H20 -»|C02 + H+ + e-
Computation Of Free Energy Changes:
AGr = -28.1 - (-24.1) = -4.0 kJ/mol e"
AGp = -28.1 - (-35.8) = 7.7 kJ/mol e~
Computation ofY:
A = 14.0
Y = 0.07
Y = 0.07
Y = 0.12
e" cells
e" methylphenol
5.65 mg cells per e'
3.18 mg methylphenol per e"
mg cells
mg methylphenol
4-METHYLPHENOL
Energy Half-Reactions:
donor:
^CyHgO+lfHap-^
acceptor:
C02 + H
Computation Of Free Energy Changes:
AGr = -28.4 - (-24.1) = -4.3 kJ/mol e'
AGp = -28.4 - (-35.8) = 7.4 kJ/mol e~
= -28.4kJ/mole-
'N = -24.1kJ/mole-
148
-------
Computation ofY
A =12.1
Y = 0.08
Y = 0.08
Y = 0.14
e" methylphenol
5.65 mg cells per e"
3.18 mg methylphenol per e"
mg cells
mg methylphenol
BENZOTHIOPHENE
Energy Half-Reactions:
donor:
AG& = - 29.2 kJ/mol e'
acceptor:
ICH4 +1H2O -» | CO2 + H+ + e- AG(°q} = -24.1 kJ/mol e'
Computation Of Free Energy Changes:
AGr = -29.2 - (-24.1) = -5.1 kJ/mol e'
AGp = -29.2 - (-35.8) = 6.6 kJ/mol e'
Computation ofY:
A = 9.7
Y = 0.09
e" benzothiophene
y = 0 09 5.65 mg cells per e'
3.73 mg benzothiophene per e"
Y = 0.14
mg benzothiophene
149
-------
OXINDOLE
Energy Half-Reactions:
donor:
^ = -34.7 kJ/mol e'
acceptor:
| CH4 + 1 H2O -> I C02 + H+ + e- AG(£j) = -24. 1 kJ/mol e"
C amputation Of Free Energy Changes:
AGr = -34.7 - (-24.1) = -10.6 kJ/mol e~
AGp = -34.7 - (-35.8) = 1.1 kJ/mol e"
Computation ofY:
A = 3.24
e" cells
Y = 0.24
e" oxindole
5.65 mg cells per e"
3.92 mg oxindole per e"
Y = Q35_jngcelk_
mg oxindole
2( 1H)-QUINOLINONE
Energy Half-Reactions:
donor:
1 IT O 1 Q*7
C9H7NO + H2O -> C02 + NH" + H+ + e'
= -31.4 kJ/mol e"
150
-------
acceptor:
I CH4 + ^ H2O -» I CO2 + H+ + e-
Computation Of Free Energy Changes:
AGr = -31.4 - (-24.1) = -7.3 kJ/mol e"
AGp = -31.4 - (-35.8) = 4.4 kJ/mol e~
Computation of If:
A = 6.0
Y = 0.14
Y = 0.14
Y = 0.21
e" cells
e" 2(lH)-quinolinone
5.65 mg cells per e'
3.82 mg 2(lH)-quinolinone pere"
mg cells
mg 2( 1H) - quinol inone
1 (2HHSOQUINOLINONE
Energy Half-Reactions:
donor:
1 1*7
C9H7NO +
O 1
C02 +
"3*7
AG(°q} = -24. 1 kJ/mol e'
acceptor:
| CH4 + ^ H2O -> g CO2 + H+ + e-
Computation Of Free Energy Changes:
AGr = -31.3 - (-24.1) = -7.2 kJ/mol e"
AGp = -31.3 - (-35.8) = 4.5 kJ/mol e~
?n\ = -24.1 kJ/mole'
151
-------
Computation ofY:
A = 6.1
Y-Q.14
Y = 0.14
Y = 0.21
e" l(2H)-isoquinoiinone
5.65 mg cells per e'
3.82mg l(2H)-isoquinolinonepere'
mg cells
mg l(2H)-isoquinolinone
152
------- |