United States
                        Environmental Protection
                        Agency
                                      National Risk Management
                                      Research Laboratory
                                      Ada, OK 74820
                        Research and Development
                                      EPA/600/S-97/003
vvEPA
August 1997
ENVIRONMENTAL
RESEARCH  BRI
               Anaerobic Biodegradation of BTEX in Aquifer Material


             Robert C. Borden*, Melody J. Hunt*, Michael B. Shafer* and Morton A. Barlaz*
Abstract

Laboratory and field experiments were conducted in two
petroleum-contaminated aquifers to examine the anaerobic
biodegradation of benzene, toluene, ethylbenzene and xylene
isomers (BTEX) under ambient conditions. At both sites,
destructive microcosm experiments were conducted
following the EPA protocol for estimation of anaerobic
microbiological transformation rate data (Federal Register,
Vol. 53, No. 115).  Aquifer material was collected from
locations at the source, mid-plume and end-plume at both
sites, incubated under ambient conditions, and monitored
for disappearance of the test compounds. In the mid-plume
location at the second site, in-situ column experiments
were also conducted for comparison with the laboratory
microcosm and field-scale results.
In material from the first site, collected from a location in the
Sleeping Bear Dunes National Lakeshore, Michigan, toluene
biodegraded in microcosms under methanogenic conditions
after a 60- to 246-day lag period. There was no statistically
significant evidence of benzene, ethylbenzene, or xylene
biodegradation in the microcosm study.
At the second site, near Rocky Point, North Carolina, all
BTEX components biodegraded under ambient, anaerobic
conditions.   In the mid-plume microcosms, m-xylene
  Department of Civil Engineering, North Carolina State University,
  Raleigh, North Carolina27695-7908.
                        biodegraded first followed by toluene, oxylene and benzene
                        under iron reducing conditions. None of the compounds
                        biodegraded in the source area microcosms.  In the end-
                        plume microcosms, biodegradation was variable with
                        extensive biodegradation in some microcosms and little or
                        no biodegradaion in others. In all  microcosm sets where
                        biodegradation was measured,'the  compound being
                        investigated degraded to a low but detectable level (5 to
                        30 (ig/Liter), after which biodegradation slowed or stopped.
                        Biodegradation rates for m-xylene and benzene in in-situ
                        columns from the mid-plume location were similar to
                        microcosm rates.
                        Anaerobic biodegradation of individual BTEX components
                        often consisted of three distinct phases: (1) a lag period with
                        little or no biodegradation; (2) a rapid degradation period;
                        and  (3)  an asymptotic  period  where contaminant
                        concentrations remained essentially constant. This pattern
                        of biodegradation cannot be accurately described with a
                        simple first-order decay function. In contrast to the behavior
                        of the individual compounds, the biodegradation of total
                        BTEX appears to  more closely approximate a first-order
                        decay function.

                        Introduction

                        This study focuses on the anaerobic biodegradation of benzene,
                        toluene, ethylbenzene and xylene isomers (BTEX) in aquifer
                        material from two petroleum-contaminated aquifers:  Sleeping
                        Bear Dunes National Lakeshore (SB) in Michigan and a site near
                        Rocky Point, North Carolina (RP). The two sites examined were
                        chosen because of their differences in plume size, contaminant

                                               Gyy  Printed on Recycled Paper

-------
residence time, geochemical environment and geologic setting.
At both sites, previous field monitoring indicated that anaerobic
biodegradation of one or more BTEX components was occurring
(Bordenefa/., 1995; Wilsonefa/., 1994). Twodifferenttechniques
were used to evaluate the ability of indigenous microorganisms
to anaerobically degrade BTEX and to  estimate the rate  of
degradation: (1) destructively sampled laboratory microcosms;
and (2) in-situ test chambers. Both techniques consist of spiking
aquifer sediment with BTEX and monitoring compound
disappearance overtime. Experimental conditionsweredesigned
to mimic ambient conditions in the aquifer to the maximum extent
possible.  Killed controls were monitored at the same time to
differentiate between biological and abiotic losses.

Procedure
Laboratory microcosms were constructed with aquifer material
from source, mid- and end-plume locations at both sites.  The
microcosms were prepared in serum bottles with a high sediment
to water ratio of 1.8 dry g/mLiter. The experimental procedure
was based on the EPA protocol for estimation  of anaerobic
microbiological transformation  rate data (Federal Register,
Vol. 53,No.115). Multiple replicate microcosms were constructed
using blended aquifer sediment and ground water recovered
under anaerobic conditions. The aquifer sediment was collected
aseptically and anaerobically using methods developed  by
U.S. EPA (Dunlap ef al., 1984).  Ground water was collected
anaerobically through a 0.45 micron filter from monitoring wells
adjacent to each core location.  Microcosms were spiked with
approximately 2,000 ng/Liter of benzene, toluene, ethylbenzene,
oxylene  and m-xylene and incubated in anaerobic containers
stored at the ambient ground-water temperature, 16°C.  The
microcosms were constructed in an anaerobic glove box using
aseptic techniques. A reducing agent was added to microcosms
constructed with SB sediment while naturally occurring Fe II (aq)
served as a reducing agent in the RP microcosms. Resazurin
was added to all microcosms and indicated anaerobic conditions
throughout the monitoring period. Triplicate microcosms and
abiotic controls were sacrificed at approximately monthly intervals
for one to two years and analyzed to determine the loss of BTEX
and changes in other electron acceptors and donors.
The in-situ columns were used at one location for comparison
with the  laboratory microcosms and are similar to the system
used  by  Gillham  et al. (1990).  Each  column consists  of a
chamber (1.0 m long) where  sediment and ground water are
isolated from the surrounding aquifer for controlled observation
(Figure 1). Columns were installed by drilling a pilot hole and then
installing a 15 cm-diameter by 3 m-Iong section of  polyvinyl
chloride (PVC) casing. Stainless steel tubing and 3 m of drill rod
were attached to the equipment chamber. Argon was pumped
into the casing to displace any oxygen present. The column was
then pushed into the aquifer. Suction was applied to the stainless
steel feed line to insure that the column was completely filled with
aquifer material. The columns were then filled with anaerobic
ground water containing BTEX.   Two microcosm columns and
one abiotic control were used in each experiment.  The abiotic
control columns were prepared  by adding  an inhibitor to the
injection water (final  concentration  was  either 0.1 N  HCL or
500 mg formaldehyde/Liter).  All columns were monitored  for
BTEX, dissolved iron, sulfate, chloride, pH and dissolved oxygen
(DO). Analytical methods followed in the microcosm and in-situ
column experiments are described elsewhere (Hunt ef al., 1997;
Beckman, 1994).
Effective first-order removal rates for BTEX in the microcosms
and  in-situ  columns were estimated using  the equation
C = Co exp (-Kt), where K"is the apparent first-order decay rate
(day1), f is time and Co is the initial concentration. The biological.
loss rate was calculated as the difference between the loss rates'
in the microcosms and  abiotic controls over the time period in
which biological losses  were observed. A  two-tailed Student's
test,  assuming unequal  variances, was performed to determine
if the microcosm rates were statistically different from the abiotic
control rates.

Results and Discussion Sleeping Bear Site

Site Characteristics
The first site (SB), located at the Sleeping Bear Dunes National
Lakeshore, Michigan, is a highly  transmissive glacial outwash
consisting of  coarse sand and gravel with calcium carbonate
fragments. The soil has a relatively high organic carbon content
(up to 2.2%) that primarily occurs as coatings on sand grains
(>94%)  (West ef al., 1994).  The alkylbenzene plume covers a
relatively short span (25 m) before it is intercepted by the Platte
River.  Contaminant residence time in the aquifer varies from 5
to 53 weeks because of fluctuations in the water table gradient.
Oxygen, nitrate and sulfate are rapidly consumed, and dissolved
iron is produced at the upgradient edge of the hydrocarbon spill.
in the  source area and downgradient contaminated  aquifer,
methane is produced and dissolved BTEX concentrations decline
(Table 1). Field monitoring data indicate that toluene biodegrades
                                Stainless Steel Tubing

                                Ball valve

                                          Ground Surface
                                       15.34 cm (61;)
                                       PVC casing
                                     BW casing
                               End plate with main screens
                                 Test Chamber
 Figure 1. Schematic of in-situ test column.

-------
 TABLE 1. Geochemical Characterization of Ground Water at the SB and RP Sites

 SB site
Parameter
Distance from source (m)
Well Screen (m)
Depth of Core (m)
Oxygen (mg/Liter) b
Nitrate (mg/Liter) b
Fe (aq, mg/Liter) b
Sulfate ( mg/Liter) b
Methane (mg/Liter) b
Eh (mV) c
pH =
BTEX (mg/Liter) c
Background
-23
2.7-4.7
NAa
2.4
67
3.5
20.0
0.08
NMd
NM
<0.001
Source area
0
2.7 - 3.7
2.3 - 4.2
<0.1
2.2
7.7
<0.05
0.10
-294
5.9

Mid-plume
9.1
1.2-2.1
0.5 - 2.4
<0.1
0.3
4.1
6.0
1.7
-180
6.5
3.3
End-plume 	
21.3
0.6-1.5
0.6 - 2.3
0.4
0.1
5.2
<0.05
3.1
-186
6.0
2.0
 RP site
Parameter
Distance from source (m)
Well Screen (m)
Depth of Core (m)
Oxygen (mg/Liter) e
Nitrate (mg/Liter) e
Fe (aq, mg/Liter) e
Sulfate ( mg/Liter) e
Methane (mg/Liter) e
Eh (mV) o
pH*
Benzene (mg/Liter) e
Toluene (mg/Liter) e
Ethylbenzene (mg/Liter) e
o-Xylene (mg/Liter) e
m-, p-Xylene ( mg/Liter) e
Background
-55
0.9-3.9
NA
3.1
19.6
0.2
18.9
<0.001
196
4.6
<0.005
<0.005
<0.005
<0.005
<0.005
Source area
0
1 .0-4.0
4.3-5.0
0.6
<1.5
24.6
34.8
0.096
-132
5.8
1.33
10.4
1.82
3.88
8.18
Mid-plume
183
4.9-6.4
3.8-4.8
0.4
<1.5
52.0
4.0
0.096
-118
6.1
0.62
0.083
1.92
0.042
2.75
End-plume
327
5.0-6.5
3.8-4.6
1.4
<1.5
1.8
12.8
0.40
-187
7.1
0.62
0.028
0.042
0.010
0.34
        bData are average of 2 samples taken over a 4-month period as reported by Wilson etal., 1994.
        'Measuredat time of core collection.
        dNM - not measured.
        "Data are average of 4 to 7 measurements taken over an 18-month period as reported by Gomez, 1993.
most rapidly with slower biodegradation of ethylbenzene and the
xylene isomers (Wilson etal., 1994). There was no evidence of
anaerobic benzene degradation.

Microcosm Results

The rate and pattern of biodegradation were similar at all three
locations at the SB site.  Toluene biodegraded  under
methanpgenic conditions at all three locations after a lag period
that varied from 60 to 246 days.  First-order decay rates during
the period  of active biodegradation  were 0.042,  0.023 and
0.032 d~1  for  the  source, mid-  and end-plume  microcosms,
respectively.  Figure 2 shows the variation  in dissolved toluene
and methane in microcosms from the mid-plume location.  The
increase in methane greatly exceeded the amount that could be
expected from the measured BTEX loss, indicating that other
undefined substrates were being biotransformed.  In addition,
most  of the methane was produced  during the period when
toluene was not biodegrading (Figure 2).  Similar trends were
measured in the  source and  end-plume microcosms.  The
aquifer material used in the microcosms contained ~0.6% organic
carbon while West etal. (1994) found up to 2.2% organic carbon
in aquifer sands 21 m upgradient of the plume. Acetate had also
been detected in ground water from the site (Wilson etal., 1994).
Thus, background organic carbon was the major electron donor
in the SB sediment from all three locations.  The initiation of
methane production prior to toluene biodegradation and the long
lag times prior to the onset of toluene biodegradation may be due
to the presence of more readily degradable substrates.  In a

-------
                                200
                               Days
Figure 2.    Average concentration of toluene (a) and methane (b) in
microcosms (solid) and abiotic controls (open) from the mid-plume location
at the SB site.  Data are the average of three destructively sampled
microcosms at each time point. Error bars represent one standard deviation.


study of  methanogenic enrichment cultures derived from
contaminated aquifer sediments, Edwards and Grbic-Galic (1994)
found the presence of other organic substrates such as acetate,
amino acids and propionate inhibited anaerobic degradation of
toluene and o-xylene.
There was no evidence of  o-xylene, m-xylene or benzene
biodegradation at any location or ethylbenzene biodegradation in
the source and mid-plume microcosms.  However, for the end-
plume microcosms, the ethylbenzene results are ambiguous. The
first-order loss rate for ethylbenzene in the microcosms was not
significantly different from the abiotic control over the entire test
period; however, there was evidence of ethylbenzene degradation
in selected replicates.
Wilson ef a/.  (1994)  reported  biodegradation  rates for  BTEX
components based on field monitoring of the SB plume coupled
with estimates of contaminant retention time prior to discharge to
the Platte River. Given the uncertainty in the calculation of both
in-situ and  microcosm biodegradation rates, quantitative
comparison of these laboratory results and the field rate of Wilson
ef a/, is inappropriate. Qualitative comparison of the field and
laboratory data shows consistency with respect to: (1) the absence
of benzene biodegradation;  (2)  the production of methane; and
(3) the preferential  biodegradation  of  toluene.  While  the
laboratory data  do not support or  preclude  low rates of
ethylbenzene biodegradation,  Wilson  ef a/. (1994) reported
biodegradation rates for ethylbenzene that were 10% of the
rates reported for toluene.

Rocky Point Site

Site Characteristics
The second site (RP) is located on the coastal plain near Rocky
Point, North Carolina, and is of marine origin. The site geology
consists of dark gray and green micaceous fine sand, overlain
by 1.5 to 4.5 m of silts, clays and clayey sands. The sediment
is over 90%  quartz sand with minor amounts of pyrite  and
muscovite flakes in a clay matrix. X-ray diffraction has shown
that most of the iron in the aquifer solids is present as glauconite,
an iron-rich clay mineral, with smaller  amounts of  the clays
berthierine and possibly iron-rich illite (Becker, 1992).
Field monitoring  (Table 1) indicates that toluene and o-xylene
decline rapidly during transport through the first 100 m of the
300 m-long contaminant plume (Borden ef a/., 1995). Benzene
and m-, p-xylene decline more slowly before discharging  to a
small drainage ditch. Anaerobic ethylbenzene biodegradation
was not evident from the field  data. At the upgradient edge of
the plume, oxygen  is rapidly depleted. Sulfate  is depleted
immediately  downgradient of the  contaminant source  and
dissolved  Fe (II)  increases in  the mid-plume area, suggesting
that biodegradation occurs using both sulfate and ferric iron as
electron acceptors.  The absence of significant methane in the
monitoring wells indicates methanogenic fermentation is not a
major process at this site.
The average ground-water velocity in the aquifer is 30 m/yr,
resulting in a contaminant residence time in excess of 12 years
(Borden ef a/., 1995).

Microcosm Results
At the RP site, anaerobic biodegradation of BTEX was a
function of location within the contaminant plume. In the source
area microcosms, none of the  BTEX components degraded
during 388 days of incubation. In the mid-plume microcosms,
m-xylene biodegradation began with no lag, followed by toluene,
o-xylene and benzene  (Figure 3a). By day 140, both toluene
and o-xylene had degraded to between 3 and 11 u,g/Liter and
then remained constant for the duration of the experiment.  With
the onset of toluene  degradation  after day 22, the rate of
m-xylene  loss declined and did not increase until toluene and o-
xylenewerebelow22 ug/l_iterat120days. Benzene degradation
began after day 180, and by day 403 the benzene concentration
was between 8  and 12 ug/Liter in each replicate (Figure 3b).
Two isolated replicates in which BTEX degraded also exhibited
ethylbenzene biodegradation during the last  100 days of
incubation; however, the average decay rate for ethylbenzene
was minimal over the sampling period.
 In the mid-plume location, the dominant electron acceptor was
ferric  iron present  as an  iron-rich  clay mineral,  glauconite.
 Methane was not produced in any microcosm and manganese
 concentrations were below detection.  While small amounts of
 sulfate were present initially, the dissolved sulfate concentration
 remained constant (Figure 3d) and was not sufficient to account
 for the observed BTEX loss.  Dissolved iron concentrations
 varied but appeared to increase with time (Figure 3e). However,
 this increase was relatively small in proportion to the increase

-------
          Abiotic Benzene
          Benzene
          Toluene
                                    V   Ethylbenzene
                                        m-Xylene
                                        o-Xylene
                 O  Benzene - Abiotic
                 •  Benzene - Live
          10  Ni- Live
              :~H~ Abiotic
              I- - - Abiotic 1st Order Rate
              	 Live 1st Order Rate
      3

      1
      "o
      CO
      CO
.15


.10


.05


.00
      £

      1
       o
       01
       to
             0      100     200     300     400
                              Days


Figures. Resultsfrommid-plumeRPmicrocosms: (a)BTEXcomponents,
(b) benzene, (c) total BTEX, (d) dissolved SO4'2, and (e) dissolved Fe. Data
in figures a, c, d and e are the average of three replicates destructively
sampled at each time point. Error bars are ± one standard deviation. Abiotic
benzene concentrations in Fgure 3a represent the behavior of the other
BTEX components. Benzenedatain FigureSbareforindividual microcosms.
 in solid-phase Fe  (II).  At the end of the experiment, the
 measured increase in total Fe (II) was 200% of the amount
 predicted based  on the measured  loss of BTEX, assuming
 conversion to CO2. This indicates that significant quantities of
 non-BTEX organic carbon also biodegraded under iron reducing
 conditions.
 In the mid-plume microcosms, biodegradation of toluene,  o-
 xylene and benzene consisted of three distinct phases: (1) a lag
 period with little or no biodegradation;  (2) a rapid degradation
 period; and  (3)  an  asymptotic  period where contaminant
 concentrations remained essentially  constant.  This pattern of
 biodegradation cannot be accurately described with a simple
 first-order decay curve. However, when total  BTEX is plotted
 versus time (Figure 3c), the lag and asymptotic periods disappear
 and  the experimental results closely approximate a first-order
 decay process.  The first-order  decay rates for individual
 compounds reported  in Table 2  are for the  period  of  rapid
 biodegradation and are not representative of the entire incubation
 period.  Use of these rates in a simple first-order decay function
 will  substantially underestimate  the time required for
 biodegradation.

 Biodegradation in the  end-plume RP  microcosms was variable.
 Some microcosms exhibited clear evidence of biodegradation by
 day  106 while other microcosms did not show any evidence of
 biodegradation after 327 days (Figure 4). Toluene and o-xylene
 degradation appeared to proceed concurrently starting at 106
 days (Figures 4b and 4c). On days 184 and 253, there was clear
 evidence for toluene and o-xylene degradation in all microcosms
 sampled.  Yet on the last two sampling dates, the evidence was
 ambiguous as toluene had degraded in three of six microcosms
 sampled and o-xylene  had degraded in only two of six. Benzene,
 m-xylene and ethylbenzene also degraded in selected microcosms
 but usually only when  toluene and o-xylene  were also depleted
 (Figure 4).   Estimation of first-order decay rates was not
 appropriate at this location, since the calculated rate would be
 more a function of the random order of microcosm sampling than
 the actual rate of biodegradation. It was not possible to conclusively
 identify the electron acceptor in  the end-plume  microcosms
 because of the presence of high concentrations of  dissolved
 sulfate and interferences in the solid-phase iron analyses.
 Field (Borden etal., 1995) and laboratory results indicate that at
 the RP site, the  various BTEX components biodegrade in a
 sequential process: toluene and o-xylene are usually depleted
 first,  followed by benzene and/or  m-xylene  and  finally
 ethylbenzene. In the mid-plume microcosms, this pattern varied
 somewhat. The m-xylene biodegradation began without a lag
 but then slowed once toluene and o-xylene degradation began.
 At this location in the aquifer, toluene ando-xylene concentrations
 are low (~5 ug/Liter), but significant quantities of m-, p-xylene still
 remain (>1000 ug/Liter). These results suggest that m-xylene
 degradation was occurring/n-s/fuat the time of sediment removal.
 However,  toluene and  o-xylene were preferentially degraded
when elevated  levels  of these compounds  were added to the
 microcosms.  Benzene biodegradation did not begin until  after
 180 days.  By this time, toluene and the xylene isomers were
depleted to less than 10 ug/Liter. In the end-plume microcosms,
the added compounds also degraded  sequentially; however,
toluene and o-xylene degradation  did not occur until after 106
days (Figure 4). At this time, the various factors that control the
 order of biodegradation are not fully understood . Clearly, there
are some interactions  between the  degradation of the different
compounds. As discussed above, the presence of other organic
substrates, including acetate, amino acids and propionate, has

-------
TABLE 2. First-order Decay Rates from the RP Site

Compound
Benzene
Toluene
Ethylbenzene
o-Xylene
m-, p-Xylene
Total BTEX
Mid-Plume
Decay rate"
(d-1)
0.024
0.045
NSC
0.020d
0.0066
Microcosms
Decay Period
(days)
184-403
22-120
37-120
0-184
0-403
Mid-Plume In-Situ Columns
Decay rate3
(d-1)
Test 1 : 0.0049
Test 2: 0.023
NS°
Ns°
NSC
0.0143
0.0029
Decay Period
(days)
Test 1: 155-251
Test 2: 41 -181

121-251
121-251
Field
Decay rate"
(d-1)
0.0002
0.0021
0.0021
0.0013
0.0011
* Rates represent the difference between the microcosms and control abiotic loss rate over the period of decay.
"Reid decay rates were calculated over the entire length of the plume.
" Microcosms and control abiotic loss rates were not significantly different at the 99% level.
* Only m-xylene was present in the microcosms.
been reported to inhibit anaerobic degradation of toluene and
o-xylene. While only BTEX concentrations were quantified in the
present work, the gas chromatograms from the  mid-plume
microcosmsshowthatalargenumberof unidentified compounds
were present at low concentrations at the start of the experiment.
Over the course  of the  experiment, these compounds  also
disappeared.
The lag period prior to biodegradation varied significantly among
sampling  locations at the  RP  site and among individual
microcosms.  The exact cause of this variability is unknown but
is believed to  be due to differences in prior adaptation to BTEX,
electron acceptor availability and aquifer geochemistry. In the
source area,  field monitoring data  indicated that concurrent
BTEX  biodegradation and  sulfate reduction were occurring
immediately downgradient of the gasoline spill (Borden et al.,
1995). Consequently, sediment and ground water were collected
from an area with significant concentrations of BTEX (26 mg/Liter)
and sulfate (-35 mg/Liter) in the ground water. In retrospect, it
appears that this was an area where active sulfate reduction was
not occurring, since significant concentrations of sulfate were still
present in the ground water.  In the end-plume location, there
was tremendous variability in aquifer geochemistry in the
immediate  area where the  sediment was collected.   Prior  to
construction of the microcosms, the sediment from each location
was blended  and  passed  through a No. 8 sieve to homogenize
the sediment. The highly variable response in the destructively
sampled microcosms from the end-plume location indicates that
this procedure was not sufficient to eliminate differences among
replicate microcosms.

In-Situ Column Results
Two sets of column experiments were performed at the mid-
plume location at the RP site for comparison with the microcosm
results.  In each  set of experiments, two microcosm columns
were operated in parallel with one abiotic control. All microcosm
and abiotic control  columns exhibited an initial concentration
decrease of several hundred ng/Liter because of sorption to the
aquifer sediment (Beckman, 1994).  The concentrations  of
hydrocarbons in  the  abiotic columns remained constant  or
declined slowly after the initial drop, indicating that biological
activity or short-circuiting did not occur in the control columns.
The DO concentration remained low (<0.2 jag/Liter) throughout
all experiments.  Biological  loss rates were calculated as the
difference between the rates in the microcosms and the abiotic
controls, when these  rates were statistically different.  The
microcosm decay rate was estimated by pooling the result from
the two  microcosm  columns over  the  period of active
biodegradation.
The first set of experiments was performed using ground water
from a nearby well that was depleted in toluene and o-xylene but
contained higher concentrations of benzene, ethylbenzene, m-,
p-xylene, pseudocumene (1,2,4-trimethylbenzene) and
mesitylene (1,3,5-trimethylbenzene). Samples were collected
monthly for approximately 250 days when the experiment was
terminated because of limited sample volume. The m-, p-xylene,
benzene and pseudocumene biodegraded after initial lag periods
that varied from 85 to 121 days. By day 251, m-, p-xylene had
decreased by over 90% (Figure 5), benzene had decreased by
50% (Figure 6a) and pseudocumene had decreased by 75% to
90%.  In the control column, m-, p-xylene had decreased toy over
48%,  benzene had decreased by 22% and pseudocumene had
decreased by 35%. There was no evidence of toluene, o-xylene,
ethyl-benzene or mesitylene biodegradation in either microcosm
column. The absence of toluene and o-xylene biodegradation
was likely due to the low initial concentration of these compounds
(<50 ng/Liter). In the microcosm columns, DO remained below
detection (<0.2  mg/Liter), sulfate  declined  slightly from
1.3 mg/Literto the detection limit (-0.3 mg/Liter) andpH remained
constant at 6.3.  In the first microcosm column, dissolved iron
increased from 111  to over 200 mg/Liter; while in the second
microcosm  column, dissolved  iron  increased  from  99 to
140 mg/Liter.
In the second set of experiments, the  columns were reloaded
with ground waterthat contained higher concentrations of benzene
(1,000 to 1,300 ng/Liter) and very  low concentrations of TEX
(-20  ng/Liter) to determine if benzene biodegradation would
continue.  Benzene biodegradation began  after a 41-day lag
period, and  by day 334  benzene had declined from over

-------
           100
        a  10
        CD
        c
         I ' I •! I 'I '
2 f S
|io|
CD
N
c
CD
-^ 1
LU
n 1
!""! • s §.
•™


t t i i I i i i i I i i i i I i i i r I i i r i ^< i t i 1 i i i <
              0    50   100  150  200  250  300  350
                              Days
Figure 4.  Variation in (a) benzene, (b) toluene, (c) o-xylene, (d) m-xylene
and (e) ethyl-benzene in microcosms constructed with material from the RP
end-plume location. Microcosm (•) and control (o) results are from individual
destructively sampled microcosms. Three microcosms and threecontrols
were sampled at each time point, except the last time point, when no control
was sampled.
 1,000 ug/Liter to  11  jj.g/Liter in column 1 and to 8  ng/Liter in
 column 2 (Figure 6b). The ground water used to reload the in-situ
 column was obtained from a nearby multilevel sampler. Previous
 monitoring had indicated that the sulfate concentration of this
 ground water was very low (~1  mg/Liter).   However,  after
 reloading the columns,  sulfate concentrations in  microcosm
 columns 1 and 2 were 85 and 65 mg/Liter, respectively. Additional
 monitoring confirmed that a  pulse of high sulfate ground water
 had migrated past the multilevel sampler  intake at the time
 ground water was collected for injection into the in-situ columns.
 Over the course of this experiment, sulfate remained constant in
 column 2 but declined by 25% in column  1, and dissolved iron
 remained constant in both columns.  Data are insufficient to
 positively identify the electron  acceptor  used for benzene
 degradation in the in-situ column experiments.

 First-order Biodegradation Rates
 First-order  biodegradation rates  from  the  in-situ  column
 experiments are compared to the biodegradation rates from the
 laboratory microcosms  and field measurements in Table 2.
 Biodegradation rates for m-, p-xylene in the in-situ columns and
 m-xylene in the laboratory microcosms were similar.  However,
 the lag period prior to the start of biodegradation was longer in
 the in-situ columns. The lag period prior to benzene biodegradation
 in the laboratory  microcosms  and  the first in-situ  column
 experiment was similar.   However, the  rate of benzene
 biodegradation in the first in-situ column experiment was a factor
 of five lower than the laboratory microcosm rate. During this
 experiment, benzene biodegraded concurrently with m-, p-xylene.
 However, in the microcosms, m-xylene was completely degraded
 before the start of benzene biodegradation. In the second in-situ
 experiment, only benzene  was present and the  benzene
 biodegradation rate was similar to the  laboratory microcosms.
 In most cases, the measured biodegradation rates for individual
 compounds are comparable in both columns and microcosms
 but are one  or  two orders of  magnitude  higher than  rates
 estimated from field investigations (Borden  et a/., 1994).  One
 likely cause of this difference is the procedure used to calculate
 biodegradation rates. In the  laboratory microcosms and the in-
situ columns, there is a definite lag  period prior to the start of
biodegradation. The laboratory and in-situ column rates were
calculated during the period of active biodegradation after the lag
                                                                  10
                                           0     50    100   150    200   250   300
                                                             Days
                                 Figure 5.  Variation in m-, p-xylene in the first set of in-situ test columns.
                                 Open symbols are abiotic controls; solid symbols are microcosm columns.

-------
      10°
   a
               -o-
                           -©
                                •Q	O
      •JO  I •. t  i I I i i i I I I  I I I I I I I I  I I I I I l I I I I  1 I I I
           0    50   100  150  200   250  300  350
                            Days
Figure 6.  Variation in benzene in in-situ columns: (a) initial experiment;
and (b) after reloading with benzene only. Open symbols are abiotic
controls; solid symbols are microcosm columns.

period had ended. In the field, it is usually not possible to identify
the zones where biodegradation is most active, and the reported
degradation rates are for the travel time over the entire plume.
The  large differences between laboratory microcosm, in-situ
column  and field-scale biodegradation rates for individual
compounds could be reduced by using a grouped parameter
such as total BTEX.  In the laboratory microcosms, one or more
BTEX components were degrading throughout the experiment,
so a simple first-order decay model  closely matched the total
BTEX results for the entire experiment with no observable  lag in
total BTEX biodegradation (Rgure 3c). In the field, one or more
BTEX components are  biodegrading at any location, and
consequently a first-order decay function more closely matches
the field data throughout the entire length of the plume. When
laboratory microcosm, in-situ column and field rates are compared,
biodegradation rates for total BTEX are much more consistent
than the rates for individual compounds.  For example, the
laboratory biodegradation rate for benzene was 120 times the
field rate, while the laboratory rate for total BTEX was only 6 times
the field rate. The highest benzene degradation rate in the in-situ
columns  was  115 times the field rate, while  the total BTEX
degradation rate in the columns was only 2.6 times the field rate.
A second potential cause for the observed differences between
laboratory microcosm, in-situ column and field biodegradation
rates is spatial variations in biological activity. The rates reported
in Table 2 were  calculated  from column and  laboratory
measurements at a single location (mid-plume). At a second
location  (source area)  there was no evidence  of BTEX
biodegradation, and in a third location (end-plume) the results
were variable.  In  contrast, the field  degradation rates were
estimated from monitoring well data collected along the length of
the plume  and should represent the large scale,  spatially-
averaged rate.

Conclusions and Recommendations

Anaerobic Biodegradation of BTEX
1.  Benzene, toluene, ethylbenzene and the xylene isomers are
    anaerobically  biodegradable under  ambient subsurface
    conditions using ferric iron, sulfate and/or carbon dioxide as
    terminal electron acceptors.
2.  A distinct order of biodegradation is  often observed,  with
    toluene being  the  most rapidly biodegraded compound.
    However, this order may vary from site to site. For example,
    at the Rocky Point site, o-xylene was rapidly biodegraded;
    while at the Sleeping Bear site, no significant biodegradation
    of o-xylene was observed.
3.  The more easily biodegradable compounds (toluene, o-
    xylene, m-xylene) appear to anaerobically biodegrade to a
    low but detectable concentration (10 to 30 ug/Liter) after
    which biodegradation slows or stops.  It is not clear whether
    biodegradation of these compounds will continue once the
    more difficult to degrade compounds are depleted.
4.  Use of a simple first-order decay model does not adequately
    describe the  anaerobic  biodegradation of individual
    compounds in laboratory microcosms  or in-situ columns. To
    accurately simulate the anaerobic biodegradation of individual
    compounds, a model that includes  two variables will be
    required: (1) the lag period prior to biodegradation; and (2)
    the rate of biodegradation.
5.  The lag period prior to  the start  of anaerobic BTEX
    biodegradation varies from compound to  compound and
    from site to site. Until the source of this variability is better
    understood, it will not be possible  to  use laboratory
    microcosms or in-situ columns to accurately predict the time
    required for the field biodegradation of individual compounds.
6.  Destructive microcosms  yield only one concentration
    measurementfor each independent experiment (microcosm)
    and are poorly suited to generating the data required to fit a
    two-parameter model.  Consequently, use  of destructive
    microcosms, as specified in the EPA protocol for estimation
    of anaerobic microbiological transformation rate data, is not
    appropriate when  the lag  period  prior  to the  start of
    biodegradation is significant.
7.  Anaerobic biodegradation of total  BTEX  more closely
    approximates  a first-order  decay curve  than  the
    biodegradation of  the individual  compounds.  First-order
    decay rates for total  BTEX estimated from  laboratory
    microcosms, in-situ columns and field monitoring data also
    appear to be more consistent than the rates for individual
    BTEX components.

 Use of Laboratory Microcosms and In-situ Columns
 to Evaluate Natural Attenuation of  BTEX

 1.  At this time, biodegradation rates for individual compounds
    derived  from  laboratory microcosms and in-situ columns
    cannot be reliably used to estimate the time required for
     complete biodegradation in the field.

-------
2.  Laboratory microcosms are useful for: (1) demonstrating
    that  a compound of regulatory concern  can and does
    biodegrade under ambient subsurface conditions; and (2)
    evaluating the effect of different environmental variables on
    the rate and extent of biodegradation.
3.  In-situ columns may also be used to evaluate compound
    biodegradation under ambient conditions and the effect of
    different amendments. In-situ columns are relatively simple
    to install and operate, closely replicate ambient conditions,
    and result in minimal disturbance of the aquifer material.
4.  Total  BTEX may be a  more appropriate parameter for
    describing natural attenuation than the concentration of
    individual compounds. Biodegradation of total BTEX more
    closely matches a first-order decay curve than the individual
    compounds.  The degradation rate  for total BTEX also
    appears to be more consistent than biodegradation rates for
    individual compounds. However in some cases, it may be
    necessary to model individual compounds to accurately
    assess the risk to human health and the environment.
5.  Estimation of field biodegradation rates from limited point
    measurements (laboratory microcosms or in-situ columns)
    will be difficult because  of spatial  variations in  biological
    activity.

Disclaimer
The U.S.  Environmental Protection Agency through its Office of
Research and Development partially funded and collaborated in
the research described here under Cooperative Agreement
No. CR-819630 to North Carolina State University.  It has been
subjected to the Agency's peer and administrative review and
has been approved for publication as an EPA document. Mention
of trade names or commercial products does not constitute
endorsement or recommendation for use.

Quality Assurance Statement
All research projects making conclusions or recommendations
based on environmentally related measurements and funded by
the Environmental Protection Agency are required to participate
in the Agency  Quality Assurance Program.  This project was
conducted under an approved Quality Assurance Program Plan.
The procedures specified in the plan were used without exception.
Information on the  plan  and documentation  of the quality
assurance activities and results are available from the Principal
Investigator.

References
Becker, M. T. 1992. Iron Reduction in a Gasoline-Contaminated
Aquifer Containing Glauconite. M.S. Thesis, North Carolina
State University, Raleigh, North Carolina. 97 pages.
Beckman, M.  1994.  In-Situ  Measurement  of  Intrinsic
Bioremediation Rates, M.S. Thesis,  North  Carolina  State
University, Raleigh, North Carolina. 121 pages.
Borden. R. C., C. A. Gomez, and M. T. Becker. 1994. Natural
Bioremediation of a Gasoline Spill, pp. 290-295. In R. E. Hinchee,
B. C. Alleman, R. E. Hoeppel, and R. N. Miller (ed.).Hydrocarbon
Bioremediation, CRC Press, Boca Raton,  Florida.
Borden. R. C., C. A. Gomez, and M. T. Becker. 1995. Geochemical
Indicators  of Natural  Bioremediation. Ground  Water 33:180-
189.
Dunlap, W. J., J. F. McNabb, M. R. Scalf, and R. L. Cosby. 1984.
Sampling  for Organic Chemicals and Microorganisms in the
Subsurface.  EPA/600/2-77/176, U.S. Environmental Protection
Agency, Ada, Oklahoma.
Edwards, E. A., and D. Grbic'-Galic. 1994. Anaerobic Degradation
of Toluene and o-Xylene by a Methanogenic Consortiu " Appl.
Environ. Microbiol. 60:313-322.
Gillham, R. W., R. C. Starr, and D. J. Miller. 1990. A Device for
In-situ Determination of Geochemical Transport Parameters; 2
Biochemical  Reactions. Ground Water 82: 858-862.
Gomez, C. A. 1993. Characterization of a Dissolved Hydrocarbon
Plume, M.S. Thesis, North Carolina State University, Raleigh,
North Carolina. 107 pages.
Hunt, M. J., M. B. Shafer, M. A. Barlaz, and R. C. Borden. 1997.
Anaerobic Biodegradation of  Alkylbenzenes in  Laboratory
Microcosms  Representing Ambient Conditions. Bioremediation
Journal 1(1):53-64.
West, C. C., W. G. Lyon, D. L. Ross, and L. K. Pennington. 1994.
Investigation of the Vertical  Distribution and Morphology of
Indigenous Organic Matter at the Sleeping Bear Site, Michigan.
Environ. Geol. 24:176-187.
Wilson, J.T., D. H. Kampbell, and J. Armstrong.  1994. Natural
Bioreclamation of Alkylbenzenes (BTEX) from a Gasoline Spill in
Methanogenic Groundwater, pp. 201-218. In R. E. Hinchee, B.
C. Alleman,  R. E. Hoeppel, R. N.  Miller (ed.),  Hydrocarbon
Bioremediation. CRC  Press, Inc., Boca Raton, Florida.

-------

-------

-------
m
-o

5
o
o
en
o
o
CO
                   o o m c
o m Q.
  ^ en

  •O 5"


  I"
  CO. CO

  (D
      CO

      CD
si
iU
051


n
en 3
oo CD
3
<


3




CD
                      CD
                      Q.
                          CD
                          CO
                   O

                   oT

                 _ a

                 £§'

                 j] >
                 CD CQ
                 CO CD
                 CD 3
                 £U O
                      O
                     m
                     33
                         T>

                         O
                     O



                     &

-------