-------
acceptable quality are used, the calculated Wilcoxon one-
tailed probability is 0.314. This number is well above the
significance  threshold  of 0.1.  Therefore,  the  null
hypothesis (EL) is not rejected and there is no tendency for
the predemonstration data set to contain larger or smaller
values than the postdemonstration data set.

Coal Tar Recovery-

Based on the measured recovery  of free-phase coal tar,
1.504 gallons of coal tar was removed from the aquifer
during the demonstration (Johnson and Fahy 1997). This
result indicates that  the CROW  process is capable  of
removing coal tar from the aquifer;  however,  since no
reliable estimate of the initial amount of coal tar present in
the aquifer is available, there is no way  to determine
removal efficiency of the technology using this evaluation
test.

Summary

The results  of the three evaluation techniques for primary
objective P-l  demonstrate that the  CROW process is
capable  of removing oily wastes from the subsurface.
However,  the recover}'  efficiency may be low.   The
presence of  free-phase  coal tar in  wells  after  the
demonstration indicates that the CROW process did not
reduce the concentration of coal tar to residual immobile
levels.   Furthermore, no  measurable change in  TRPH
concentration  was   recorded  before   and after  the
demonstration, suggesting that the CROW process did not
significantly reduce the concentration of coal tar in  the
treatment zone.

4.2.2              P-2:
                              of

The goal of this objective was to  determine whether the
injection of hot water would increase  the migration  of
volatile contaminants into the vadose zone. The following
sections discuss the objective, methods for evaluating the
objective, and the results.

4.2.2.1 Discussion of Objective

The  upward  migration  of contaminants  was  to  be
evaluated by measuring the concentration of BTEX in soil
gas before,  during, and after the demonstration.  Soil gas
samples were to be collected 2 days before system startup,
at 30-day intervals during the demonstration, and within
10 days of system shutdown.  In addition, one soil gas
sample was to be collected approximately 3 weeks priorto
the demonstration to test the applicability of the analytical
method. The resulting data were to be plotted as a function
of time to  assess the change  in BTEX concentrations.
However, construction activities during  replumbing and
sealing of the injection wells  destroyed all the soil gas
sampling probes. Only the predemonstration samples and
one round of demonstration samples were collected before
the probes were destroyed.

4.2.2.2

This section describes the methods and procedures used to
collect and analyze samples for the SITE demonstration of
the CROW process technology. The field and analytical
methods  and procedures used to collect and analyze
samples were in accordance with the CROW Process DP/
QAPP (PRC  1994).   A detailed  description of the
demonstration procedures is also provided in the  final
CROW Process TER (TtEMI 1997).

Predemonstration

Between April  12 and  28,  1994, soil gas probes were
installed 4 feet above the water table in eight (CB1, CB2,
CB3,  CB4, CBS, CB6,  CB7, and CB9)  of the nine soil
borings drilled inside the treatment area.  No probe was
installed in soil boring CBS  because the  water table was
within 2  feet of ground surface and the soil  gas probe
would have been  too  close  to the surface  to  ensure
collection of representative samples.  The soil  boring
locations are presented in Figure 4-4. A  test sample was
collected from  boring CB1 on August  1,  1994.   This
sample was analyzed to assess whether  the proposed
analytical  methodology  for  the   investigation  was
applicable to site conditions.  No analytical problems were
noted. The soil gas probe in boring CB2 was destroyed
before  predemonstration  samples  were   collected.
Predemonstration   samples  were  collected  from the
remaining soil  gas probes  using Summa canisters on
August 23,  1994. The Summa canisters for CB6 and CB7
malfunctioned and the soil gas  probes were resamplcd on
September 9,1994. All soil gas samples were analyzed for
BTEX.

Demonstration

Soil  gas within the treatment  area was  sampled on
November  23,  1994 to evaluate whether the  CROW
                                                   41

-------
process was causing volatile contaminants to migrate from
groundwater into the vadose zone.  During November
1994, the water heater operated intermittently.  By early
December, it became clear that the water heater would not
be operational for an extended period of time. Soil gas
sampling was suspended until the CROW process started
to inject hot water.   Unfortunately, site  construction
activities related to problems with the injection wells
destroyed all the soil gas sampling stations before the
CROW  process   started  steady-state  operation.
Reinstallation of the soil gas probes was not possible due
to access restrictions, and no additional soil gas samples
were collected.

4.2.2.3

The only injection wells  capable of injecting water  in
November 1994 were IW-1, IW-2 and IW-6. The soil gas
probes in close proximity to active injection wells were
CB1, CB3, CB4, and CB9. The total BTEX concentrations
in soil  gas  for the  demonstration were   an  order  of
magnitude larger than predemonstration concentrations in
samples from probes CB1 and CBS. The concentration of
total  BTEX in samples  from  probes  CB4 and  CB9
increased slightly.  Soil gas probes CB6 and CB7 were
located in an area where no water was injected.  The total
BTEX concentrations in samples from probes CB6 and
CB7 were lower in the demonstration samples than  in
predemonstration samples. These data suggest that total
BTEX concentrations were higher in the area influenced
by active injection, while  BTEX concentrations were
lower in the area with no active injection. The destruction
of the  sampling  probes prevented  the  collection  of
additional data and evaluation of long-term contaminant
concentrations  in  the  vadose  zone.    The  limited
availability of data prevents the complete  evaluation  of
this demonstration objective and no definitive conclusions
about the upward migration of contaminants can be made.

4.2.3              P-3:
                                 of       Tar

The goal of this objective was to determine whether the
injection of hot water would decrease the viscosity of the
coal tar and allow it to infiltrate into the lower fine grained
confining  unit.  The following  sections discuss the
objective, methods for evaluating the objective, and the
results.
4.2.3.1 Discussion of Objective

The downward migration of coal tar was evaluated by
measuring  the  TRPH concentration in  soil samples
collected from the silty sand lithologic unit directly below
the treatment zone in nine soil borings drilled before the
demonstration and in nine soil borings drilled after the
demonstration.   The  soil  borings  drilled  after  the
demonstration were within 3 feet of the soil borings drilled
before the demonstration. One sample was collected from
each borehole at a depth of approximately I foot into the
silty sand unit.  These borings were also used to collect
samples from the overlying stream gravel unit (P-l). The
locations of the nine soil borings (CB1, CB2, CBS, CB4,
CBS, CB6, CB7, CBS, and CB9) are presented in Figure 4-
4. The TRPH concentration in samples collected before
the  demonstration   was  compared  to  the  TRPH
concentration in the adjacent sample collected after the
demonstration.   A  paired sample  t-test  was used  to
determine if there  was a significant increase  in  the
concentration  of  TRPH  over the  course  of  the
demonstration.

4.2.3.2 Methods

Predemonstration

Soil samples were collected to measure the amount of coal
tar present in the aquifer before the CROW process was
implemented. A hollow-stem auger drill rig was used to
install the boreholes.  Soil samples were collected between
April  12 and 28, 1994 using a 3-inch-diameter split-spoon
sampler.  Soil from the sampler was logged and transferred
to the sample container. Nine  soil borings were drilled
within the treatment area.  Nine samples were collected
from  the silty sand  unit below the treatment area and
analyzed for TRPH.  The  presence or absence of free-
phase coal tar was not evaluated using the soil samples.

Postdemonstration

Soil samples were collected to measure the amount of coal
tar present in the aquifer after the CROW demonstration
was complete. A hollow-stem auger drill rig was used to
install the boreholes.  Soil samples were collected between
August  12  and 15,  1997 using a 3-inch-diameter split-
spoon sampler. Soil from the  sampler was logged and
transferred to the sample container. Nine soil borings were
drilled within the treatment  area.  Nine samples were
                                                   42

-------
collected from the silty sand unit below the treatment area
and analyzed for TRPH.

Postdemonstration sampling was designed to collect soil
samples  as near  as  possible to  the  locations where
predemonstration soil samples were collected.  The areal
positions  of  the  postdemonstration  boreholes  were
established  by  measurement from monitoring wells
present both before  and after the demonstration.  The
postdemonstration samples were also collected from the
top of the silty sand unit.

4.2.3.3 Results

The demonstration plan proposed that a paired sample li-
test be used to compare the predemonstration TRPH
sample data to the postdemonstration TRPH sample data
and determine if there was a significant increase in the
TRPH concentration.  The paired sample t-test requires
that the  data be normally or log-normally distributed
(Gilbert 1987) and that the differences between the paired
data must be normally distributed.  The Shapiro-Wilk w-
test (Gilbert  1987)  indicated that  the  data sets for the
predemonstration and postdemonstration samples were
log normally distributed and that the difference between
the paired data were normally distributed. Therefore, the
t-test is appropriate (Gilbert 1987).

The paired sample t-test was applied to the TRPH data
using a  significance  level of 0.1.   The t-test statistic
calculations are summarized in Table 4-4. The data used
for the statistical analysis are the average of all acceptable
Table 4-4. Statistical Tests for the Silty Sand Unit Below the Treatment Area
Soil Boring Depth
(feet)
CB1 (31 .5-32)
CB2 (25.5-26)
CBS (31 .5-32)
CB4 (25.5-26)
CB5 (30-30.5)
CB6 (27.5-28)
CB7 (30-30.5)
CBS (22.5-23)
CB9 (24.5-25)
Mean
Samples, n
Distribution (a)
Difference Distribution (a)
Significance threshold, a
Paired sample t-test statistic (b)
One-tailed probability
Conclusion 
-------
quality data for each sampling point.   The analytical
results for  predemonstration samples CB-6 (27.5-28).
CB-7 (30-30.5),  and CB-8 (22.5-23) were  qualified
estimated nondetect due to method blank contamination.
These results were included in the calculation because
even though the results are biased high two of the three
were  lower than the postdemonstration samples.   A
thorough evaluation of analytical data was conducted and
conclusions regarding the validity of sample results are
presented in  Section 4.2.10.  When all the data with
acceptable quality are used, the calculated t-test one-tailed
probability  is 0.832.  This is well above the significance
threshold of 0.1. Therefore, the null hypothesis (H0) is not
rejected and there is no tendency for the predemonstration
data set to contain larger or smaller values than the
postdemonstration data set.

Qualitative  evaluation of the  predemonstration  and
postdemonstration data suggests that postdemonstration
TRPH concentrations are higher than predemonstration
TRPH concentrations. The mean of the postdemonstration
TRPH results (584 milligrams per kilogram [mg/kg]) is
approximately twice  the mean  of the predemonstration
results (236 mg/kg).  At  seven  out of nine  sampling
locations, the postdemonstration results are higher than
the predemonstration results. These results suggest that
the CROW  process caused contamination to migrate from
the stream gravel unit into the underlying silty sand unit.

4.2.4              P-4:
                        of       Tar

The goal of this objective was to determine whether the
CROW process allowed contamination to migrate outside
of the treatment area.  The following sections discuss of
the objective, methods for evaluating the objective, and
the results.

4.2.4.1 Discussion of Objective

The areal containment of coal tar was  evaluated in two
way s. The first way was to measure the TRPH, BETX, and
PAH  concentration in groundwater samples  collected
from  five  monitoring wells located  outside  of the
treatment area (Figure 4-2).  One well was located
upgradient  (CDW-1) and  four  wells were  located
downgradient  (CDW-2,  CDW-3, CDW-4, and S-1A).
Monitoring well S-1A was installed by AES during  a
supplemental  investigation, and monitoring wells CDW-
1, CDW-2, CDW-3, and CDW-4 were installed pnorto the
demonstration.   The wells were screened completely
through the upper aquifer.  Samples collected from the
upgradient well  were used to determine if upgradient
contaminants migrated into the treatment area. Samples
collected  from  the  downgradient wells were  used to
determine if the CROW process flushed contaminants
downgradient.   The resulting data  were plotted  as  a
function of time to determine the change in TRPH, BTEX,
and PAH concentrations.

The areal  containment was also evaluated by measuring
the concentration of TRPH in soil samples collected from
below7 the water table in four soil  borings drilled outside
the treatment area before the demonstration and four soil
borings drilled after the demonstration. The soil borings
drilled after the demonstration were within 3  feet of the
soil borings drilled before the demonstration. The samples
from soil  borings drilled after the  demonstration  were
collected at the same depth as samples collected prior to
the demonstration.   Samples were collected at 5-foot
intervals starting 1 foot below the water table. Since the
aquifer is  15 to 25  feet thick outside the  treatment area,
three to five samples were collected from each boring. The
locations of the four soil borings (CB10, CB11, CB12, and
CB13) are presented in Figure 4-4.  TRPH concentration in
samples collected before the demonstration was compared
to the  TRPH  concentration in  the  adjacent  sample
collected after the demonstration.  A paired sample  t-test
was used to determine if there was a significant change in
the TRPH concentrations.

4.2.4.2

Predemonstration

Soil samples were collected to measure the amount of coal
tar present in the aquifer before the CROW demonstration
began. A hollow-stem auger drill rig was used to install
the boreholes. Soil samples were collected between April
12 and 28,  1994 using  a 3-inch-diameter split-spoon
sampler. Soil from the sampler was logged and transferred
to the sample container.  Four soil borings were drilled
outside the treatment area and 11 soil samples and four
duplicates were collected from the stream gravel unit and
analyzed for  TRPH.  The split-spoon sampler did not
collect samples representative of the stream gravel unit
due to the abundance of large diameter materials.  The
split-spoon sampler collected the fine-grained portion of
the unit; the analytical  results are therefore expected to
overestimate the  amount of coal tar present in the stream
                                                    44
-------
gravel unit. The presence or absence of free-phase coal tar
was not evaluated using the soil samples.

Groundwatcr samples were  collected  to  measure the
distribution   of  contaminants  before  the   CROW
demonstration began.    Groundwater  samples  were
collected from monitoring wells CDW-1, CDW-2, CDW-
3, CDW-4, and S-1A on August 23, 1994. All samples
w7ere analyzed for TRPH, BTEX, and PAHs. Monitoring
well CDW-1 was located upgradient of the treatment area,
monitoring well CDW-2 was located 15 feet downgradient
of the  treatment area, and  monitoring wells  CDW-3,
CDW-4, and S-1A were located in an arc approximately
100 feet downgradient of the treatment area. Monitoring
wells CDW-1 through CDW-4 were installed  in April
1994.

Demonstration

Groundwater samples were  collected  to  measure the
distribution of contaminants during implementation of the
CROW process.  Groundwater samples were collected
from monitoring wells CDW-1, CDW-2, CDW-3, CDW-
4, and  S-1A.  Monitoring well CDW-1  was sampled to
establish the concentration of contaminants flowing into
the treatment zone.   Monitoring wells CDW-2,  CDW-3,
CDW-4, and S-1A were sampled to evaluate whether the
CROW process was flushing contaminants downgradient.

All monitoring wells were sampled once even'  2 weeks
from November 21, 1994 to January 30,  1995. One set of
groundwater samples was collected on March  2,  1995
while the  CROW process was  shut down to provide
baseline information when the system resumed operation.
Once the CROW process was operated at steady state, the
monitoring wells were sampled on July 12, August 16,
October 4, and November 21, 1995.  All samples  were
analyzed for TRPH, BTEX, and PAHs. Groundwater was
not  sampled again until postdcmonstration sampling in
August 1997.

Postdemonstration

Soil samples were collected to measure the amount of coal
tar present in the aquifer after the CROW demonstration
was complete. A hollow-stem auger drill rig was used to
install the boreholes.  Soil samples were collected between
August 12 and 15,  1997 using a 3-inch-diameter  split-
spoon sampler. Soil from the sampler  was logged and
transferred to the sample container. Four soil borings were
drilled outside the treatment area.  Eleven samples and
four duplicates were collected from the stream gravel unit
outside the treatment area and analyzed for TRPH.

Postdemonstration sampling was designed to  collect soil
samples as  near  as  possible to the locations where
predemonstration soil samples were collected. The areal
positions of  the postdcmonstration boreholes  were
established  by measurement  from  monitoring wells
present both before  and after the demonstration.   The
addition of fill  during construction of the CROW process
made the collection of samples relative to ground surface
inappropriate. To determine the correct depth interval, the
soil samples were therefore collected relative to the water
table and the top of the silty sand unit. Measurement of the
depth to water in the monitoring wells installed during the
predemonstration indicated that the water table elevation
was  within 0.5 foot during the predemonstration and
Postdemonstration soil sampling activities.

Groundwater samples were collected to  measure the
distribution of contaminants afterthe CROW demonstration
was complete.  Groundwater samples were collected from
monitoring wells CDW-1, CDW-2, CDW-3, CDW-4, and
S-1A on August 13, 1997. All samples were analyzed for
TRPH, BTEX, and PAHs.

4.2.4.3

Coal Tar Concentrations in Soil

The demonstration plan proposed that a paired sample t-
test be used to compare the  predemonstration TRPH
sample data to  the postdcmonstration TRPH sample data
and determine  if there was a significant increase in the
TRPH concentration.  The paired sample t-test requires
that the data be normally or log-normally  distributed
(Gilbert 1987) and mat the differences between the paired
data must be normally distributed. The Shapiro-Wilk w-
test (Gilbert 1987) indicated that the data sets for the
predemonstration and postdemonstration  samples  were
log normally distributed and that the differences between
the paired data were normally distributed.  Therefore, the
t-test is appropriate (Gilbert 1987).

The paired  sample t-test was applied to the TRPH data
using a significance  level of 0.1.   The  t-test statistic
calculations are summarized in Table 4-5.  The data used
for the statistical analysis are the average of all acceptable
quality data for each sampl ing point. Analytical data were
                                                   45
-------
Table 4-5. Statistical Tests for the Stream Gravel Unit Outside the Treatment Area
Location Depth
(feet)
CB10 (6.5-7)
CB10 (11.5-12)
CB11 (8-8.5)
CB11 (13.5-14)
CB12 (11.5-12)
CB12 (18.5-19)
CB12 (20.5-21)
CB13 (13.5-14)
CB13 (15.5-16)
CB13 (20.5-21)
CB13 (24.5-25)
Mean
Samples, n
Distribution (a)
Difference Distribution (a)
Significance threshold, a
Paired sample t-test statistic (b)
One-tailed probability
Conclusion 
-------
predemonstration concentrations.   The exception  is
monitoring  well  CDW-2,  where  postdemonstration
samples had higher concentrations of BTEX and PAH and
a lower concentration of TKPH than the predemonstration
samples.

These groundwater data indicate that the CROW process
may  cause   short-term  increases  in  contaminant
concentrations downgradient of the treatment area.

4.2.5              S-1:


The goal of this  objective was to obtain data on CROW
process performance. The following sections discuss the
objective, methods for  evaluating  the objective, and the
results.

4.2.5.1 Discussion of Objective

The data documenting  the operational parameters were
collected by the data acquisition and control system.  The
operational parameters  that were measured consisted of
temperature of injected water and recovered groundwater;
injection pressure; recovery well pressure; and the water
flow rate for recovery wells, injection wells,  recycled
water, and treated water. These data were used to evaluate
the  pore  volume  flushing  rates  and the  thermal
equilibration rate in the aquifer.

4.2.5.2

Beginning  in July  1995.  hot   water  injection  was
established and was nearly continuous until shutdown in
June 1996. During this period, 9.5 x 106 gallons of water
at  an average heater outlet  temperature of 147°F  was
injected.   Figure 4-5  shows the hot water injection and
extraction rates and the  cumulative pore volumes injected
and extracted over the 366-day-test. The average  injection
and extraction rates for the hot-water injection period were
19.6 and 24.0 gpm, respectively.  The  groundwater
extraction rate exceeded the total water  injection by
approximately 4 gpm  throughout the  test to provide
hydraulic containment  and recovery of the injected hot
water.  The affected pore volume was considered the area
heated to 150°F, with a thickness of 20 feet and a porosity
of 35 percent. The pore volume was estimated at 455,000
gallons. Overthe 366-day period, 20.8 pore volumes were
injected into the  treatment area and a total of 25.5 pore
volumes were extracted from the treatment area.
Initial thermal equilibration in the aquifer was reached
approximately  30  days after continuous  hot  water
injection was established. Profiles of total water injection
rate, average injected water temperature, and average
extracted water temperature are shown in Figure  4-5.
Aquifer temperatures  in  extraction well water reached
120°F  after 30 days  and 130°F  after about 50 days.
Normal extraction water temperatures ranged from 130 to
140°F during the remainder of the test (see Figure 4-6).

4.2.6


The  goal of this  objective  was to determine the costs
incurred while installing, operating, and decommissioning
the CROW system. The following sections discuss the
objective, methods for evaluating the objective, and the
results.

4.2.6.1 Discussion of Objective

The cost to implement the CROW process at the Brodhead
Creek Superfund  site  was determined by assessing the
following 12 cost categories.

1.     Site preparation
2.     Permitting and regulatory requirements
3.     Mobilization and startup
4.     Equipment
5.     Labor
6.     Supplies
7.     Utilities
8.     Effluent treatment and disposal
9.     Residual waste shipping and handling
10.    Analytical services
11.    Maintenance
12.    Site demobilization



The actual costs associated with the implementation of the
CROW Process SITE demonstration  at  the  Brodhead
Creek  Superfund  site  are  presented  and analyzed in
Section 3.  The demonstration costs are grouped into 12
cost categories, and a breakdown of these costs under the
12 cost categories is presented in Table 3-1 and Figure 3-
                                                    47
-------
                                                                                                            0
                    0   20   40   60   80   100  120  140  160  180   200  220  240  260  280  300  320  340
                                                          Day of Test
                              Total Injection Rate
                             - Total Extraction Rate
                       	Total Injected Pore Volume
                       	Total Extracted Pore Volume
Figure 4-5. Pore volume flushing rates.
            1/1
            01
            £!
            O)
            Ol
            Q
            &
            i
180
160-
140
120
100
 80
 60
                40
                   0   2   4   6   8  10  12  14  16  18 20  22  24  26  28
                                           Day of Test
—r 200
'"<- 180
 - 160
 -- 140
 - 120
 '- 100
 - 80
 -- 60
 - 40
^-- 20
 - 0
 30
	Average Injection
        Temperature (degrees F)
	Average Extraction
        Temperature (degrees F)
	Injection Flow Rate (gpm)
Figure 4-6. Thermal response at hot water flushing startup.
                                                             48
-------
4.2.7  Objective S-3: Assess Potential
        Fractionation of Coal Tar

The goal of this objective was to determine whether the
CROW process  preferentially  removed more  mobile
contaminants.  The following sections discuss of the
objective, methods for evaluating the objective, and the
results.

4.2.7.1  Discussion of Objective

The soil samples required to evaluate coal tar fractionation
were collected and analyzed with the soil samples required
to assess the primary objectives. Five soil samples were
collected  from  below  the  water   table  before  the
demonstration and four  adjacent soil  samples were
collected afterthe demonstration. A fifth postdemonstration
sample  was mistakenly collected at the  wrong depth
interval. Before the demonstration, the soil samples were
collected from the nine soil borings (Figure 4-4) installed
to evaluate the primary objectives. The exact location was
determined by the site geologist after a sufficient volume
of sample was collected. Afterthe demonstration, samples
were collected adj acent to the samples collected before the
demonstration.

The  analytical suite for these samples was  limited to
BTEX and PAHs.  To  evaluate whether fractionation
occurred, the ratio of total BTEX to total PAHs was
calculated.  In addition, the ratio of the total concentration
of two-  and three-ring PAHs to the total concentration of
four- and five-ring PAHs was calculated.  It was presumed
that heating of contaminated soil areas as a result of the
CROW process might cause lower density contaminants
(BTEX and 2- and 3-ring PAHs) to physically separate or
fractionate from heavier constituents. Physical separation
of these lower viscosity fluids might allow them to more
freely  move  to groundwater  extraction  points.  The
expected analytical variability from adjacent samples was
estimated at approximately 20 percent, and it was assumed
that changes in mean ratios greater than 30 percent would
indicate that the CROW process fractionates the coal tar.

4.2.7.2 Results

The results of the soil boring analyses are shown in Table
4-6 and indicate  no consistent pattern of changes  in
contaminant ratios.  Large changes in the ratios  of total
BTEX to total PAHs concentrations  were indicated,
ranging from a 90 percent decrease to a 1,348  percent
increase; however, it is not  clear what phenomenon is
responsible for the changes.   Increases in ratios at soil
borings CB3  and CB6  were primarily  the  result  of
increases in BTEX concentrations. Decreases in ratios at
CB7 and CB9 were the result of both increases in PAH
concentrations and decreases in BTEX concentrations.  In
addition, large changes were  indicated  at CB7 and CB9,
where  little or no  effective hot water injection occurred.
Although ratios of total 2- and 3-ring PAHs to total 4- and
5-ring  PAHs  at the four  sampling  locations were more
comparable to each other than the total BTEX to total PAH
ratios, percent changes at CB6, CB7, and CB9 were within
the 20 percent variability considered expected.  The 30
percent decrease in ratios of total 2- and 3-ring PAHs to
total  4-  and  5-ring  PAHs  at   CB3  may  indicate
fractionation; however, these  results are not corroborated
by the results at the other sampling locations. The results
Table 4-6. Predemonstration and Postdemonstration Contaminant Ratios
Soil
Boring
CBS
CB6
CB7
CB9
Ratio of Total BTEX to Total PAHs
Pre- Post- Percent
demonstration demonstration Change
0.02299
0.02392
0.01062
0.05764
0.33538
0.13194
0.00018
0.00550
+1 ,359%
+452%
-98%
-90%
Ratio of 2- and 3-Rina PAHs to 4- and 5-Rina PAHs
Pre-
demonstration
3.5872
2.6273
1 .9400
1 .3376
Post-
demonstration
2.5146
2.6497
1 .7284
1.1452
Percent
Change
-30%
+1%
-11%
-14%
                                                    49
-------
of the data analysis are inconclusive regarding coal tar
fractionation due to the CROW process.

4.2.8             S-4:


The goal of this objective was to determine whether the
water  treatment   system  effectively  removed  the
contamination prior to discharge. The following sections
discuss  the objective,  methods  for evaluating  the
objective, and the results.

4.2.8.1 Discussion of Objective

Effluent water samples were collected at the treatment
system  discharge  and were  chemically  analyzed  to
determine   compliance   with  PADER  discharge
requirements.   These  monitored parameters included
BTEX, PAHs, pH, biochemical oxygen demand (BOD).
chemical oxygen demand (COD), oil and grease (O&G),
total suspended solids (TSS), total phenols,  and TOC.
Process water was sampled at three locations.  Figure 4-3
provides  a schematic  diagram of the water treatment
system.  The first sampling location.  SP22, was located
downstream of the oil separator tank and represented
pretreatment conditions.  The second sampling location,
SP23,  was  located  downstream  of the GAC-FBR
treatment unit  and upstream of the  carbon  adsorption
units. The third sampling location. SP29. was located after
the carbon adsorption treatment  unit  and represented
discharge water quality. These three sampling locations,
but specifically SP23 and SP29, were used to monitor the
ability of the  water treatment system  to conform  to
PADER discharge requirements. When the system started
operation,  samples  were  collected  once per  week.
Samples were collected once per 2 weeks after the system
started steady-state operation in July 1995.



Process   water  samples  were  collected  from  three
locations:  SP22, SP23, and SP29. SP22 was located after
tank 5 and before the water was heated and  reinjected.
SP23 was located after the GAC-FBR unit and before the
carbon adsorption units. SP29 was located just before the
treated water was discharged to Brodhead Creek.

Samples  were  collected  from SP22 to  measure  the
concentrations  of TRPH,  BTEX.  and PAHs that were
being reinjected into the  treatment area.  The CROW
process  started operation on November 9, 1994.  From
November 10, 1994 through February 16,  1995. samples
were collected from SP22 once or twice per week. In late
February through mid-March 1995, the CROW process
was shut down to replumb the tank farm, redevelop the
injection wells, and clean the process piping. The system
again started to pump water in late March 1995. Samples
were collected from SP22 once every 2 weeks until early
July 1995, when the heater was turned on.  Samples were
collected from  SP22 once per week or once even' other
week from July 12 through August 16, 1995. Samples
were collected once  every 2 to 3 weeks from October 4
through  December 6, 1995. No demonstration samples
were collected after December 6, 1995.

Samples were collected from SP23 and SP29 to assess the
performance of the water treatment system. Samples were
collected at both locations once per week from November
9, 1994  through February 16. 1995.  Samples were not
collected again until  after the CROW process had started
steady-state operation in early July  1995.  Samples were
collected on July 27 and August 16, 1995 and once every
2 weeks from October 4 through December 6, 1995.  No
samples were  collected from  SP23 and  SP29 after
December 6, 1995.

4.2.8.3

Figures  4-7 and 4-8  present comparisons of total BTEX
and total PAH sampling analytical results, respectively, at
the three sampling  locations.   As illustrated by these
figures,  the water treatment  system  was successful at
reducing contaminant concentrations throughout most of
the demonstration. On average, total BTEX concentrations
were reduced by more than 98 percent by the GAC-FBR
system and by more than 99.9 percent before discharge.
Total PAH concentrations were reduced by an average of
more than 96 percent by the  GAC-FBR system  and by
more than 98 percent before discharge.  Some elevated
concentrations of PAHs were detected in the treated water
discharge during system startup and in the later part of the
test.

Throughout the demonstration, measured BTEX and PAH
concentrations at SP29 were compared with the federal
MCLs.  All BTEX concentrations were below the MCL.
The only PAH with a promulgated MCL is bcnzo(a)pyrcnc
(BaP). As shown in Figure 4-9, detectable concentrations
of BaP above the MCL of 0.2 microgram per liter (mg/L)
were measured  during the demonstration in 10 of 22
                                                  50
-------
                                                                            QSP22 Pre-Treatment
                                                                            D SP23 After GAC/FBR
                                                                            • SP29 After Carbon Polishing Units
                        X
                        m
                        00
                        1
                                                         Date
Figure 4-7. Water treatment system BTEX sampling results.
                                                                               QSP22 Pre-Treatment
                                                                               D SP23 After GAC/FBR
                                                                               • SP29 After Carbon Polishing Units
                                                    Date
Figure 4-8. Water treatment system PAH sampling results.
                                                              51
-------


100 -
CJ
centration (ug
3 _>.
n ->• o
fi D 9
0.1 -
n n-i .

p ( \p
MPT fCt ° im/T "l •' v
	 Reporting Limit .' I
.*, A J "--
• » * * . _ i \. •
Jk. * * * ^k \ **
, ~ » -^A_ A ^ . A- _ A A A -A A*

¥
o
r^
                o
§

ok
(N
                        Q
                            Q
                            
-------
Table 4-7. PADER Effluent Limits for the CROW Process Demonstration
            Parameter
Daily Maximum
SP29-November 17,
       1994
SP29-October 18,
      1995
Benzene
Toluene
Ethylbenzene
Acenaphthene
Acenaphthylene
Anthracene
Benzo(a)anthracene
Benzo(a)pyrene
Benzo(g,h,l)perylene
Chrysene
Dibenzo(a,h)anthracene
Fluoranthene
Fluorene
lndeno(1 ,2,3-cd)pyrene
Naphthalene
Phenanthrene
Pyrene
Total Phenol
BOD/COD
TSS
pH 6-9
Notes:
U Compound not detected at
J Estimated concentration.
10.0
20.0
10.0
20.0
20.0
2.00
20.0
0.24
0.18
20.0
20.0
2.20
20.0
2.00
6.00
2.80
1.96
20.0
50,000
16,000
at all times

concentration shown.

5.0 (U)
5.0 (U)
5.0 (U)
9.3 (J)
10.0
5.0
9.2
10.0
2.3 (J)
8.6
5.0 (U)
9.1
6.5
2.8 (J)
6.1
13.0
13.0
5.0 (U)
5.0 (U)
5.0 (U)
7.55



3.5
0.5 (U)
0.5 (U)
7.8
130(E)
47.0
18.0
160(E)
35.0
60.0
13.0
5.5
18.0
34.0
0.5 (U)
13.0
150(E)
64.6
148,800
95,800
6.37



E Concentration exceeded calibration range.
                                                 53
-------
injected water.   The  following  sections discuss the
objective,  methods for evaluating the objective, and the
results.

4.2.9.1 Discussion of Objective

The objectives of groundwater modeling at the Brodhead
Creek Supcrfund site were to determine (1) the extent of
capture by the on-site extraction wells (PW-1 and PW-2),
and (2) whether water that is reinjected using on-site
injection wells (IW-1, IW-2, IW-3, IW-4, IW-5, and IW-
6) is completely captured by the on-site extraction wells.

4.2.9.2

A conceptual  model of  the  site  hydrogeology was
formulated before a computer code was  selected to
simulate capture zones and reinjection water flow paths. A
conceptual model describes the  components  of the
groundwater flow system and is developed from regional.
local,  and site-specific  data.  Flow system components
include parameters such as groundwater flow direction
and gradient, aquifer thickness, and  water transmitting
properties.   Development of a conceptual  model was
necessary before constructing a computerized groundwater
flow model.

The conceptual model was formulated to organize existing
field data so that the groundwater flow system could be
analyzed more readily.   The conceptual model was
simplified  as  much  as  possible;  however,  enough
complexity was retained to simulate groundwater system
behavior  adequately  for  the  intended purposes  of
modeling   (Anderson  and  Woessner  1992).    The
conceptual model for the  site was developed using  all
existing data and information.  Model assumptions that
were applied are listed below.

Assumptions Required for Use in Analytical Models:

 *  The aquifer is homogeneous and isotropic

 •  Groundwater flow is horizontal, unidirectional, and at
    a steady state

Assumptions Based on Available Field Data:

 •  The hydraulic conductivity is equal to 1.29.6 feet per
    day (ft/d)
 •  The saturated thickness of the aquifer is equal to 10
    feet

 *  The transmissivity equals 1296 square feet per day
    (fWd)

 •  The magnitude of the hydraulic gradient is equal to
    0.0036 ft/ft

 «  The levee core to the east of the site treatment area can
    be modeled as a no-flow boundary condition

 •  The hydraulic gradient direction  is primarily to the
    south, and parallel to the levee core

 •  The aquifer porosity is equal to 0.30 (unitless)

 *  The groundwater seepage velocity is  1.56  ft/d, or
    about 569 feet per year. This estimate is based on an
    average  hydraulic  conductivity  of 129.6  ft/d.  a
    hydraulic gradient of 0.0036 ft/ft, and an effective
    porosity of 0.30, using a variation  of Darcy's  Law
    (Fetter 1980).

Hie calculation used to determine these parameters is as
follows:

 Q  =(KxI)/ne

where:

 Q  =   seepage velocity, or pore water velocity (ft/d)
 K  =   hydraulic conductivity (ft/d)
 I  =   hydraulic gradient (ft/ft)
 ne  =   effective porosity (unitless)

The   Well  Head Protection Area   (WHPA)   model
(Blandford and Huyakorn 1991) was selected to simulate
capture zones and reinjection water flow paths associated
with site activities. The WHPA model is a semianalytical
program based on superposition of mathematical solutions
for  groundwater movement  that would  result  from
pumping extraction or injection wells in the presence of a
regional hydraulic gradient. The WHPA model delineates
capture zones associated with discharging extraction wells
and flow paths associated with reinjection water using a
particle  tracking technique.  A particle is viewed as an
individual molecule of water or molecule of a conservative
tracer that moves through the aquifer coincident with the
bulk  movement  of  groundwater flow.  Time-related
                                                    54
-------
capture zones are obtained by tracing the pathlines formed
by a series of particles placed around the well bore of the
pumping well.   These particles  are either  forward- or
reverse-tracked with time.  The WHPA model is EPA-
approved and is widely used in the public domain.  It is
distributed and supported by the International Groundwater
Modeling Center (IGWMC) in Golden, Colorado.

4.2.9.3

Capture zones  and reinjection water  flow  paths  were
analyzed for this pumping and reinjection scenario using
the WHPA model code. The flow rate data from the data
acquisition system was evaluated and average flow rates
for the extraction and injection wells were calculated. For
this scenario, extraction wells PW-1 and PW-2 discharge
at 5.9 gpm and 24.3 gpm. respectively. Water is rcinjcctcd
into the  aquifer at the  following rates:  IW-1 (8.7 gpm);
IW-2 (6.4  gpm); IW-3 (0.0 gpm); IW-4 (0.0  gpm); IW-5
(0.0 gpm); and IW-6 (6.0 gpm). Simulated capture zones
for the  two  discharging  wells  and flowpaths  for the
reinjected water from the three active reinjection wells are
provided in  Figure  4-10.   This figure  indicates  that all
reinjected water is captured by the two  on-site extraction
wells.

4.2.10  Quality Control

A data quality assessment was conducted to  evaluate the
field and laboratory QC results, evaluate the impact of all
QC measures on the overall data quality, and remove all
unusable values from the investigation data  set.   The
results of this assessment were used to produce the known,
defensible information employed to define the investigation
findings and to draw  conclusions.  The QC data were
evaluated with respect to the QA objectives defined in the
CROW  Process DP/QAPP (PRC 1994).

The analytical data for groundwatcr, process water, and
soil  samples  collected  during  the  CROW  Process
demonstration were reviewed to ensure  that all laboratory
data generated and processed are scientifically valid,
defensible, and comparable.   Data  verification  was
conducted using both field QC samples and laboratory QC
samples.  The  field QC samples included equipment
blanks, field blanks, trip blanks, matrix spike/matrix spike
duplicates (MS/MSD), and sample duplicates. Laboratory
QC samples included blank spike/blank spike duplicates
(BS/BSD), and  laboratory  control  sample/laboratory
control  sample  duplicates  (LCS/LCSD).   Initial  and
continuing calibration results were also analyzed to assure
the quality of the data and that proper procedures were
used. The results of these samples were used to calculate
the precision, accuracy, completeness, representativeness,
and comparability of the data.

The  following  items  were  evaluated during the data
review:

 •  Sample chain-of-custody condition and holding times

 «  Instrument performance  check  (gas chromatograph/
    matrix spike [GC/MS]  volatile and  semivolatile
    analysis)

 •  Initial and continuing calibrations

 *  Surrogate spike  recoveries  (GC/MS  volatile  and
    semivolatile  analysis)

 «  Blanks (trip, field, and laboratory)

 •  BS/BSD recoveries and precision

 «  MS/MSD recoveries and precision

 •  LCS/LCSD recoveries and precision

 •  Sample/sample duplicate precision

The  following subsections summarize the limitations of
analytical data based on the evaluation of QA/QC samples
and discuss whether data quality objectives were met. For
the critical parameters of interest, analytical data was
investigated and conclusions regarding  the  validity of
sample results are presented below. Review of the overall
data packages indicates  that the data are useful for the
purpose of evaluating the technology. Table 4-8 presents
the percentage of useable data based on a review of the QC
results.  Although there were QC issues in the analytical
results for the predemonstration soils, enough data were of
acceptable   quality  to  allow  comparison   with  the
postdcmonstration soils results and to evaluate changes in
contaminant concentrations.

Soil

Soil samples were analyzed for TRPH to evaluate primary
objectives P-l, P-3, and  P-4. A select number were also
analyzed for SVOCs  and VOCs to evaluate secondary
-------
                                                                                                                                  WETLANDS
                                                                                                                        LEGEND
                                                                                                                          d    RECOVERY WELL LOCATION
                                                                                                                          A    INJECTION WELL LOCATION
                                                                                                                        - —  GROUNDWATER CONTOUR LINE
                                                                                                                        —^—  INJECTED WATER FLOW PATH

                                                                                                                             ^ 90-DAY CAPTURE ZONE FOR PUMPING WELL
                                                                                   20'      0      20'	40'
                                                                                        SCALE: 1" = 40'
Figure 4-10. Capture zone and flow line analyses.
-------
Table 4-8. Percentage of Useable Data
Matrix
Soil
Ground water


Process water


Critical
Parameters
TRPH
TRPH
BTEX
PAHs
TRPH
BTEX
PAHs
Predemonstration a
91 %b
100%
100%
100%
Not Conducted
Not Conducted
Not Conducted
Demonstration a
Not Conducted
100%
100%
100%
100%
100%
100%
Postdemonstration a
100%
100%
100%
100%
Not Conducted
Not Conducted
Not Conducted
    Notes:

    a The completeness goal for the demonstration was 90 percent.
    b Data were rejected (9%) due to method blank contamination - rejected data were less than five
     times the method blank concentration.
objective S-3.  The predemonstration soil samples were
analyzed by Versar Laboratories in Springfield, Virginia
and Radian Analytical Services in Austin, Texas.  The
samples were analyzed for the critical parameter TRPH by
EPA method 418.1. Versar analyzed the samples in five
sample data groups (SDG).  Radian analyzed the samples
in one SDG.  Table 4-9 presents a summary of the Versar
and  Radian  data  quality information.   There  are no
apparent QA/QC problems with the  Radian data.   All
blank, calibration, LCS, duplicate, and MS data are within
acceptable limits.

Some data quality issues were associated with the Versar
data.  SDG  15 was extracted on April 20, 1994, and
analyzed on May 10, 1994.  SDG 15 includes the samples
collected from borings CB2, CBS, CB4, CB5, CB9, CB12,
and CB13. The data from borings CB2, CBS, CB4, CB5,
and CB9 were used to evaluate objectives P-l and P-3.
The  data from borings CB12 and CB13 were used to
evaluate objective P-4.  Due to low recoveries  for the
initial calibration verification  standard and MS samples,
the   instrument was  recalibrated  and  the SDG  was
reanalyzed.   The  reanalysis  also exhibited  low MS
recoveries. Since the calibration, LCS, and BS recoveries
were acceptable, the low MS recoveries were likely due to
matrix effects. The low  MS recoveries suggest that the
data from these sample runs are biased low. Splits of 10
samples were analyzed in SDG 15 and by Radian.  For
seven of the 10 samples, the TRPH concentration reported
by Radian was lower than the concentration reported by
Versar.  Comparison to the Radian data suggests that the
SDG 15 data are not biased low.

SDG 16 was extracted on April 22, 1994, and analyzed on
May 16, 1994.  SDG 16 includes the samples collected
from boring CBS,  and the data were used to evaluate
objectives P-l and P-3.   The method blank contained
TRPH ata concentration of 58.9 mg/kg and MS recoveries
were relatively  low (59 to 62.2 percent).   Since the
calibration, LCS, and BS recoveries were acceptable, the
low MS recoveries are likely due to matrix effects.  The
low MS recoveries suggest that the data from these sample
runs are biased low. However, the contamination in the
method blank suggest that the data could be biased high.

The sample and duplicate results for  sample CB815516
were qualified estimated nondetect since the reported
concentrations were less than 5 times the amount of blank
contamination. Splits of three samples were analyzed in
SDG  16 by  Radian.  All three TRPH concentration
reported by Radian were lower than the concentration
reported by Versar.  Comparison to the Radian data
suggests that the SDG 16 data are not biased low.
                                                  57
-------
Table 4-9.  TRPH Analytical Quality Assurance Data
Laboratory Sample Data
Group
Versar 1 5


15
Reanalysis

15
Reextraction
reanalysis
16

16a
Reextraction
reanalysis
17

17a
Reextraction
reanalysis
18

18a
Reextraction
reanalysis
21

21
Reextraction
reanalysis
Radian NR

Columbia 9708000220
9709000100

Holding
Times
Met


Met


Exceeded
by >15 days

Met

Exceeded
by >1 4 days

Met

Exceeded
by >1 4 days

Met

Exceeded
by >1 4 days

Met

Exceeded
by >15 days

Met

Met
5 samples -
4 days out
ICV and CCV
Recovery
(percent)
87 to 101


110to112


105 to 109


109to112

105 to 106


104 to 105

104 to 106


99 to 99

103 to 106


94 to 95

100 to 106


NAb

NA
NA

Method Blank
TRPH
Concentration
(mg/kg)
<5


<5


25.3 to 28.3


59.8

99 to 103


37.4

99 to 103


39.8

99 to 103


<5

10.8


<5

<33
<33

LCS
Recovery
(percent)
105


90


113


119

160


102

160


101

160


133

99


11010127

NA
NA

Duplicate RPDs
(percent)
Laboratory duplicate 9.9
to 35.5; MSD 7.9 to 23.2;
BSD 50
Laboratory duplicate 4.2
to 29.7; MSD 5.3 to 23.1;
BSD 50
Laboratory duplicate 0.2
to 1.7; MSD 0.9 to 61 .7

Laboratory Duplicate
18.2; MSD 3.1; BSD 16.7
Laboratory Duplicate
34.4; MSD 15

Laboratory Duplicate 5.1;
MSD 5.2; BSD 1.4
Laboratory Duplicate
37.9; MSD 1.1

Laboratory Duplicate
12.4; MSD 1.7; BSD 5.6
Laboratory Duplicate 3.4;
MSD 15.4

Laboratory Duplicate
16.8; MSD 24.4 BSD 0.3
Laboratory Duplicate
23.3; MSD 3.2

Laboratory Duplicates
1.4 to 2.0
Laboratory Duplicate 12
Laboratory Duplicates 7
to 15
MS
Recovery
(percent)
-48 to -0.6


-47 to 2.8


-209 to 374


59 to 62

88.7 to 112


1.5 to 1.7

83 to 90


1.3 to 2.2

131 to 260


-1.3 to 2.5

535 to 61 8


105 to 112

55
112 to 293

BS
Recovery
(percent)
106 to 176


122 to 127


NA


95 to 1 1 3

NA


111 to 113

NA


99 to 104

NA


112 to 112

NA


NA

96 to 104
105 to 112

       Notes:

       a The reextraction and reanalysis of sample data groups 16,17,
       " Calibration data for standards were acceptable.
and 18 were completed at the same time.
-------
SDG 17 was extracted on April 28, 1994, and analyzed on
May 17,  1994.  SDG 17 includes the samples collected
from borings CB6 and CB7. The data from these borings
were used to evaluate objectives P-l and P-3. The method
blank contained 37.4 mg/kg and MS recoveries were low
(1.5 to 1.7 percent).  Since the calibration, LCS, and BS
recoveries were acceptable, the low MS recoveries are
likely due to matrix effects.   The low MS  recoveries
suggest that the data from these sample runs are biased
low. However, the contamination in the method blanks
suggests that the data could be biased high. The results for
samples CB612513,CB627528,CB716517,andCB730305
were  qualified estimated nondetect since  the  reported
concentrations were less than 5 times the amount of blank
contamination.  Splits of two  samples were analyzed in
SDG  17  and by Radian.  In  one sample,  the TRPH
concentration reported  by Radian was lower than the
concentration reported by Versar. Comparison to the
Radian data suggests that the SDG 17 data are not biased
low.

SDG 18 was extracted on April 30, 1994 and analyzed on
May 18,  1994.  SDG 18 includes the samples collected
from borings CB1, CB10,  and CB11.   The data  from
boring CB1 were used to evaluate objectives P-l and P-3.
The data from borings  CB10 and CB11 were used to
evaluate objective P-4. The method blank contained 39.8
mg/kg and MS recoveries were  low (1.3 to 2.2 percent).
Since the calibration, LCS.  and BS  recoveries  were
acceptable, the low MS recoveries arc likely due to matrix
effects. The low MS recoveries suggest that the data from
                              oo
these  sample  runs  are biased low.    However,  the
contamination in the method blanks suggests that the data
are biased high.  No sample data were qualified due to the
method blank contamination. Splits of four samples  were
analyzed  in SDG 18 and by Radian. In three samples, the
TRPH concentrations reported by Radian were lower than
the concentrations reported by Versar. Comparison to the
Radian data suggests that the SDG 18 data are not biased
low.

SDG 21 was extracted on May 17, 1994. and analyzed on
May 20,  1994.  SDG 21 includes one  sample collected
from boring CB11  and  the data were  used to  evaluate
objective  P-4.  The MS  recoveries were low (-1.3 to 2.5
percent).  Since the calibration,  LCS, and BS recoveries
were acceptable, the low MS recoveries  are likely due to
matrix effects. The low MS recoveries suggest that the
data from these sample runs are biased low. A split of one
sample was analyzed in SDG 21  and by Radian.  The
TRPH concentration reported  by Radian was  lower than
the concentration reported by Versar.  Comparison to the
Radian data suggests that the SDG 21  data are not biased
low.

The postdemonstration  soil samples  were analyzed  by
Columbia Analytical Sendees in two SDGs. The samples
were analyzed for TRPH by EPA method 418.1.  The
sample cooler contents for both SDGs were received
above the acceptable temperature range by the laboratory.
All samples were analyzed within the specified holding
time with the exception  of five samples  in one SDG that
were analyzed 4 days past holding time.  Elevated cooler
temperatures and exceedance of the  holding time may
indicate a potential bias low. The laboratory blanks were
nondetect for each SDG and the duplicate relative percent
differences (RPD) were acceptable. The  BS recoveries
were also acceptable however, for SDG 9708000220 the
MS recoveries were acceptable (112 percent) to high (293
percent) and for SDG 9709000100 the MS recovery was
low (55 percent).  The results for SDG 9708000220 may be
biased high. The results  for samples in SDG 9709000100
are likely biased low.

Five  predemonstration  soil   samples   and  five
postdemonstration soil samples were analyzed for BTEX
and SVOCs to evaluate secondary objective S-3.  The
predemonstration samples were analyzed by  Versar in
SDGs 15 and 17. For SDG 15, all QA  parameters were
within acceptable levels, except that  the  BS recoveries
were  generally  high  while the  MS  recoveries for
fluoranthene were low.  The original analysis of SDG 17
did not include a BS/BSD and the surrogates were diluted
out of all the samples. SDG 17 was reanalyzed to include
a BS/BSD and surrogates at higher concentrations.  All
QA parameters were within acceptable levels except that
the recovery of benzo(g,h,i)perylene in the MS/MSD and
BS/BSD were low.  The relatively low MS recoveries for
SDGs 15 and 17 suggest that the SVOC data are biased
low7.  There is no apparent bias for the  VOC data.

The postdemonstration  SVOC and VOC  samples  were
analyzed by Columbia Analytical Sendees. The samples
were analyzed in two SDGs. The temperatures inside the
coolers when they arrived at the laboratory were elevated
for both SDGs.   Elevated  cooler  temperatures  may
indicate a potential bias low. All other QA parameters
were  acceptable  for  SDG 9708000220.   There is  no
apparent bias for the SVOC or  VOC  data.   For  SDG
9709000100, two VOC and SVOC samples were analyzed
outside of holding time  (between  11 and 17 days). The
VOC  data mav be biased low due to analysis after the
                                                  59
-------
holding time.  The 2-fluorobiphenyl surrogate recovery
and the acenaphthalene, fluoranthene. and pyrene MS
were high.  The  SVOC data are therefore likely biased
high.

Groundwater

Groundwatcr  samples  were  collected  to  evaluate
demonstration objective P-4.  The predemonstration and
demonstration  groundwater samples were analyzed by
Columbia Analytical Services.  The postdemonstration
groundwater samples were also analyzed by Columbia
Analytical Sendees. One predemonstration SDG and 13
demonstration SDGs were analyzed by GTC.  Columbia
Analytical  Sendees  analyzed one postdemonstration
SDG.

All predemonstration groundwater samples were analyzed
for the critical parameters TKPH by EPA method 418.1,
BTEX by SW-846 method  8260, and PAHs by SW-846
method 8270A. The sample cooler contents were received
within the acceptable temperature range by the laboratory.
All samples were extracted and  analyzed  within the
specified holding  times.  The initial and  continuing
calibrations were acceptable for all samples, and all tuning
criteria   for   bromofluorobenzene   (BFB)   and
decafluorotriphenylphosphine   (DFTPP)  were  within
limits.   All  field  and  laboratory  blanks  for the
predemonstration groundwater samples were  acceptable.
Accuracy was  evaluated  by  calculating the  percent
recovery  of the BS, MS,  LCS, and  surrogates. Precision
was determined through  the use of BS/BSD pairs, MS/
MSB pairs, and sample/sample duplicate pairs. Accuracy
and precision were within QC limits with the exception of
low  terphenyl-d!4  (20 to  30  percent)  and  high
nitrobenzene-d5  (131  to   167 percent)  semivolatile
surrogate recoveries.  The outliers,  however did not
significantly affect data  quality, and all of the data are
considered  adequate for the intended use and without
significant bias.

All demonstration groundwater samples were  analyzed
for TRPH by  EPA method  418.1,        by SW-846
method 8260, and PAHs by SW-846 method 8270A. Two
SDGs were not analyzed for BTEX. The sample cooler
contents for all SDGs were received within the acceptable
temperature range by the laboratory.  All samples were
extracted and analyzed within the specified holding times,
except one PAH  sample from well  CDW-2. collected on
November 21,1994, and the BETX  samples from wells S-
1A and CDW-1, collected on January 16,  1995, were
analyzed 1 day past holding time.  Missing the holding
time by 1 day should not significantly bias the data.  The
initial and continuing calibrations were acceptable for all
samples, and all tuning criteria for BFB and DFTPP were
within limits.  All laboratory blanks for the demonstration
groundwater samples were acceptable. The trip blanks for
two SDGs were contaminated with traces of toluene and
xylcnc, but there was no significant effect on data quality.
Terphenol-dl4 surrogate  recovery was  low7 (31 to 32
percent) for the sample from well  CDW-1 collected on
December 5. 1994 and for the sample collected from well
CDW-3 collected on December 19. 1994.  Nitrobenzene-
d5 surrogate recover}? was high (123 to 150 percent) for the
sample from well CDW-4 collected on July 12, 1995 and
for the sample collected from well CDW-3 collected on
August  16, 1995.   The SVOC MS/MSD recoveries for
several compounds were high for the samples collected on
December 5,  1994 and  July 12,  1995. The SVOC MS/
MSD recoveries for several compounds were low for the
samples collected on December 5, 1994; January 4, 1995;
January 16, 1995; and July 12, 1995. The VOC MS/MSD
recoveries for benzene was low (40 to 60 percent) for the
samples collected on January  4, 1995 and January 16,
1995. Although there were QA data outliers, data quality
was  not significantly affected, and all of the data are
considered adequate for the intended use and without
significant bias.

All  postdemonstration  groundwater   samples  were
analyzed for TRPH by EPA method 418.1,       by SW-
846 method 8260, and PAHs by SW-846 method 8270B.
The  sample cooler contents were received  within the
acceptable  temperature range by  the laboratory.  All
samples were  extracted and analyzed within the specified
holding times.  The initial  and continuing calibrations
were acceptable for all samples, and all tuning criteria for
BFB  and  DFTPP  were within limits.   All field and
laboratory blanks for the postdemonstration groundwater
samples were acceptable.  All  QA  data were within QC
limits with the exception of acenaphthene MS recover}? (8
percent). The low MS  recovery is likely the result of a
matrix effect.   The data from the SDG with the low MS
recover}' arc comparable to the results from other data
groups.  The outlier however, did not significantly affect
data quality,  and  all of the data  are adequate for the
intended use and without significant bias.

           Process Water

Samples of the process water were collected after tank 5.
prior  to biological treatment  (Figure 4-3) to evaluate
                                                   60
-------
objective P-l.  They were to be used for calculating the
amount of dissolved coal tar removed from and reinjected
into the aquifer.  Unfortunately, the major gaps in data
collection prevent a meaningful utilization of these data.
The  untreated  process  water  samples  (SP22) were
collected only during the demonstration phase of the
project  and were  analyzed  by  Columbia  Analytical
Services.  Forty-five  SDGs containing results from the
analysis  of untreated  process water  samples w7ere
submitted.  All samples  were analyzed for the critical
parameters TRPH by EPA method 418.1, BTEX by SW-
846 method 8260, and PAHs by SW-846 method 8270A.

With only one exception (sample collected on April 19,
1995),  the  sample  cooler contents for all SDGs were
received within the acceptable temperature range by the
laboratory.  The temperature of the cooler did not result in
the disqualification of any data.  The only sample not
extracted and analyzed within the specified holding times
was for the BTEX analysis of the sample collected on
January 16, 1995. The initial and continuing calibrations
were acceptable for all samples, and all tuning criteria for
BFB and DFTPP were within limits. All laboratory blanks
for the  untreated process water samples were acceptable.
The trip blanks for seven SDGs were contaminated with
traces of toluene and xylene, but with no significant effect
on data quality. The recovery of surrogate nitrobenzene-
d5 was high for the samples collected on November 10,
1994; November 17,  1994; June 14,  1995; and July 12,
1995. The recover}' of surrogate tcrphcnyl-dl4 was  low
for the samples collected on November 21, 1994; January
6, 1995; and  July 27,  1995.   The  SVOC MS/MSD
recoveries for several compounds  were high  for the
samples collected on November 17,  1994 and July 18,
1995.   The  SVOC  MS/MSD  recoveries for several
compounds were low  for the samples collected  on
November 17,  1994; December 22, 1994;  July 18, 1995;
and October 18, 1.995. The VOC MS/M.SD recovery for
benzene  was  low (40 to 60  percent) for the sample
collected on November 17, 1994. Although  there were
accuracy and precision outliers, data quality  was  not
significantly affected, and all of the data were  considered
uscablc and without significant bias.

Treated Process Water

A portion of the process water that was not heated  and
reinjected into the subsurface was treated using a GAC-
FBR and with carbon adsorption units before discharge to
Brodhead Creek. The treated process  water was sampled
at two  locations in  the  treatment  process  to ensure
conformance  with   PADER.   effluent  limitations
(demonstration objective S-4).  The process water was
sampled after the GAC-FBR (SP23 on Figure 4-3) and
after the  carbon  adsorption units (SP29).  The treated
process water samples  were collected only  during the
demonstration phase of the project and were analyzed by
Columbia Analytical Sendees.

The treated process water samples (SP23 and SP29) were
generally collected at the  same time as the untreated
process  water sample  (SP22).   These  samples  were
analyzed  together  with the  analytical  results  being
submitted in the same SDGs. The treated process water
however, was not sampled as frequently as the untreated
process water.

Twenty-three SDGs containing results from the analysis
of treated process water samples were submitted.  All
samples were analyzed for BTEX by EPA method 524.2
and PAHs by EPA method  525.  The exceptions include
two SDGs (November 17, 1994 and July 27, 1995) where
the samples were analyzed for BTEX by SW-846 method
8260 and PAHs by SW-846 method 8270A.  All of the
treated process water samples, except those in one SDG,
were  also  analyzed  for   the   following  noncritical
parameters:  (1) pH by  SW-846 method 1110A or EPA
method 150.1, (2) BOD by EPA method 405.1, (3)  COD
by EPA method 410.4, (4) O&G by EPA method 413.1, (5)
TSS by EPA method  160.2, (6) total phenols  by  EPA
method 420.2, and (7) TOC by EPA method 9060A.

The sample cooler contents for all SDGs were received
within the acceptable temperature range by the laboratory.
The majority of the samples were extracted and analyzed
within the specified holding times. Both BOD samples in
one SDG, both TSS samples in a separate SDG, and both
pH samples in three separate SDGs were analyzed outside
of the holding time.  Additionally, both PAH samples in
one SDG were extracted outside of the holding time. The
initial and continuing calibrations were acceptable for all
samples, and all tuning criteria for BFB and DFTPP were
within limits.

All laboratory blanks for the treated process water samples
were acceptable.  The trip blanks for five SDGs  were
contaminated with traces of toluene and xylene, but with
no significant effect  on data quality.  Accuracy was
evaluated by calculating the percent recovery of the BS,
MS, LCS, and  surrogates.   Precision was determined
through the use of BS/BSD pairs, M'S/MSD pairs,  LCS/
LCSD pairs, and sample/sample duplicate pairs. Although
                                                   61
-------
there were accuracy and precision outliers, data quality
was not significantly affected, and 100 percent of the data
were considered useable.

Audit  Findings

As a vital part of the QA program, two field audits  and
three laboratory audits were conducted by EPA to ensure
that measurements associated with sampling and analysis
were in conformance with the CROW Process DP/QAPP
(PRC 1994).  The  field audit of the soil sampling was
conducted on April  13,1994 and the audit of groundwater
and soil gas sampling was completed on August 23, 1994.
No concerns were found during either review. The audit of
the Versar Laboratory  was completed on April 28, 1994.
Although the auditors noted some  concerns  during the
audit of Versar Laboratories that could affect data quality,
the  impact was  minimized  by the  laboratory taking
immediate and proper corrective actions.  The auditors"
concerns did not result in the disqualification of any data.
The audit of the Radian Laboratory was completed on
September 8,  1994. Four minor concerns on the TO-14
analysis of the soil gas  samples were noted. The concerns
were not anticipated to affect data quality. The audit of the
GTC laboratory  was  completed on August 31,  1994.
Several minor concerns that were not anticipated to affect
data quality were noted.

4.3                  Conclusions

The primary demonstration objectives were to determine
whether the CROW process removed coal tar from the
subsurface or flushed the coal tar outside of the treatment
area. The CROW process was successful in removing coal
tar from the subsurface (1,504 gallons recovered), but it
was unable to reduce coal tar concentrations to residual
immobile levels since free-phase coal tar was present after
the   demonstration.    Site  characterization  activities
demonstrated that  the stream  gravel unit contained
interlayered lenses  of fine- and coarse-grained material.
The  water  injected by the  CROW process probably
preferentially flowed through the coarse-grained layers,
leaving the fine-grained  intervals  relatively untreated.
Since the free-phase coal tar is perched on or in the finer
grained layers, much of the coal tar was not hydraulically
available for removal by the CROW process.

Groundwater samples collected downgradient of the site,
and the groundwater flow and capture zone model, both
suggest that the CROW process was able to recover the
injected water during steady-state operation.  IT—vever,
the groundwater samples also show that duriL0 initial
startup the changes in the ambient groundwater  flow
system resulted in spikes of contamination being released
downgradient. Measurements of the concentration of coal
tar in the soil outside of the treatment area before and after
the demonstration did not show a significant change.  This
result suggests that the CROW process did not flush large
amounts of contamination outside of the treatment area.
Measurements of the  amount of coal tar in the lower
confining layer under the treatment area before and after
the demonstration suggest that some coal tar was pushed
down into the lower confining unit. This result is likely
due to the increase in temperature in the treatment area that
reduced the viscosity of the coal tar and allowed the coal
tar to migrate into the finer grained sediments of the lower
confining unit.
                                                    62
-------
                                                          5
The CROW process may be used to remove NAPL from
the subsurface.  At the demonstration  site the CROW
process failed to remove NAPL to residual, immobile
levels. Waste types that may be recovered with the CROW
process include  coal  tar.  creosote,  fuel oils, or  other
SVOCs.  The  CROW process has been installed at the
Brodhead  Creek   Superfund  site  in  Stroudsburg,
Pennsylvania and the Bell Pole  site  in New Brighton,
Minnesota.  At  the Brodhead Creek site, the CROW
process was used to recover coal tar at an abandoned MGP.
At the Bell Pole site, the  CROW process was used to
remove creosote and pentachlorophenol  in  a  fuel  oil
carrier at a wood treating facility.

The  equipment  and  materials  necessary  to  install the
CROW process are readily available. Prior to installation,
the  subsurface  lithology,  waste  distribution,  waste
characteristics,  and  groundwater  chemistry  must  be
characterized.  To complete the design, a treatability test
should  be  conducted  to  optimize  the   extraction
temperature, pumping rates, and water treatment system.
Treatability testing can be completed within 2  months.
Once the treatment design is completed, installation of the
treatment system can take from 1 to 6 months depending
on regulatory requirements, the number of injection and
recovery wells, and the complexity of the water treatment
svstem.
                                                    63
-------
                                                       6
Anderson M.P., and W.W. Woessner.   1992.  Applied
   Groundwater  Modeling,  Simulation of Flow and
   Advective Transport. Academic Press. San Diego,
   California.

Atlantic Environmental Services. Inc.  (AES).   1993.
   "Draft Brodhead Creek NPL Site Supplemental Field
   Investigation Report." July.

Blandford, N.T., and P.S. Huyakorn. 1991. "A Modular
   Semi-analytical  Model  for  the   Delineation  of
   Wellhead Protection Area, Version 2.0." International
   Ground  Water  Modeling  Center.   Indianapolis,
   Indiana.

Carswell, L.D. and O.B. Lloyd.   1979.  Geology and
   Ground Water Resources of Monroe County, PA:
   Water Re sources Report 4 7. Pennsylvania Geological
   Survey.  Harrisburg. Pennsylvania.

Environmental Resources Management. Inc.  (ERM).
   1990.    "Brodhead  Creek  Site  Final  Remedial
   Investigation Report.'"   Prepared for Pennsylvania
   Power and Light (PP&L). September 25.

Evans, G.  1990.  "Estimating Innovative Technology
   Costs for the  SITE Program." Journal of Air and
   Waste Management Assessment. Volume 40, Number
   7.  Pages 1047 through 1051.

Fahy, L.J., L.A. Johnson, Jr., D.V. Sola, S.G. Horn, and
   J.L. Chnstofferson.  1992. "Bell Pole CROW Pilot
   Test Results and Evaluation."   Colorado Hazardous
   Waste Management Society.   Annual Conference.
   Denver, Colorado.  October 1992.

Fetter, C.W. 1980. Applied Hydrogeology. Charles E.
   Merrill Publishing Company. Columbus, Ohio.
Gilbert,   R.O.    1987.     Statistical  Methods for
   Environmental Pollution Monitoring. Van Nostrand
   Reinhold. New York.

Gmber,  W.   1996.   "A  Fluidized Bed Enhances
   Biotreatment."  Environmental Engineering  World.
   March-April. Pages 27-28.

Johnson,  L.A., Jr., and L.J. Fahy.   1997.  "In Situ
   Treatment of Soils Contaminated with Manufactured
   Gas Plant  Wastes,  Demonstration Program Final
   Report." Prepared for the U.S. Department of Energy
   (DOE), Federal Energy Technology Center (FETC)
   under Cooperative Agreement No.
   DE-FC21-93MC30127. March.

Johnson, L.A., Jr., and F.D. Guffey.  1990.  "Contained
   Recovery of Oily Wastes (CROW), Final
   Report."   Prepared for  the U.S.  Environmental
   Protection Agency (EPA), Risk Reduction Engineering
   Laboratory (RREL) under Assistance Agreement No.
   CR-815333. August.

PRC Environmental  Management, Inc.  (PRC).   1994.
   "Final Demonstration Plan and Quality Assurance
   Project Plan  for the Demonstration of Contained
   Recovery of Oily Wastes Process." Prepared for EPA,
   Office of Research and Development (ORD) under
   Contract No. 68-CO-0047. June 6.

Remediation Technologies, Inc. (ReTeC). 1993. "Final
   Design for  Brodhead Creek Site."  July.

ReTeC.   1998.   E-Mail Message Regarding  CROW
   Process Demonstration  Cost  Data for the Brodhead
   Creek   Superfund   Site.    From  Jason Gerrish,
   Environmental Engineer, ReTeC. To Chris Reynolds,
   Project Manager, TtEMI.  November.
                                                 64
-------
Tetra Tech EM Inc. (TtEMI). 1997. "Final Technology
   Evaluation Report for the Demonstration of
   Contained  Recovery  of Oily  Wastes  Process."'
   Prepared for EPA, ORD under Contract No.
   67-1534. December 30.

U.S. Environmental Protection Agency (EPA).  1987.
   'Test Methods for Evaluating Solid Waste."
   SW-846.   Volumes IA-IC:   Laboratory  Manual,
   Physical/Chemical Methods; and Volume II:  Field
   Manual, Physical/Chemical Methods. Office of Solid
   Waste  and   Emergency   Response  (OSWER).
   Washington, D.C. Third Edition, Revision 0.

EPA.  1988.  "Guidance  for  Conducting  Remedial
   Investigations  and  Feasibility  Studies  Under
   CERCLA."  EPA/540/G-89/004.  October.

EPA. 1989.  "Control of Air Emissions from Superfund
   Air  Strippers  at  Superfund Groundwater  Sites.
   OSWER Directive 9355.0-28. June.

EPA. 1991a. Record  of Decision Transmittal Memo.
   Region III. March.

EPA. 199 Ib. "Ground Water Issue, Dense Nonaqueous
   Phase Liquids." ORD. EPA/540/4-91-002.

Western  Research Institute  (WRI).  1998.  E-Mail
   Message Regarding  CROW Process  Remediation
   Cost Data for the Bell Lumber and Pole Company Site.
   From Lyle Johnson, Project Manager, WRI. To Chris
   Reynolds, Project Manager, TtEMI. November 16.
                                                65
-------
                                            Appendix A
Nonaqueous-phase liquids have contaminated groundwater
at a variety of locations in the  United  States.   Dense
organic liquids represent a  special waste  management
problem. When these liquids are denser than water and
immiscible with water, waste discharges onto the ground
surface  result in  downward permeation  through the
saturated groundwater environment.   This downward
permeation  continues  until the waste  penetration  is
blocked by an impermeable barrier.  At these locations,
high waste accumulations remain as long-term sources of
contamination to  local aquifers.  For this reason, the
Western Research Institute (WRI) has developed a new in
situ process to contain and recover organic wastes.

A.1

WRI's Contained Recovery of Oily  Waste (CROW™)
process is  designed to address  both  lighter-than-water
nonaqueous-phase  liquids  (LNAPL)  and  denser-than-
water nonaqueous-phase liquids (DNAPL).  The CROW
process is an in situ remediation process for  light and
dense  organic  liquids such  as  coal  tars, chlorinated
hydrocarbons,  and   petroleum  products  that  have
contaminated groundwater at numerous industrial sites.

Lateral containment of organic waste accumulations docs
not stop the contamination of aquifers. Containments such
as slurry walls can temporarily isolate organic wastes from
lateral groundwater transport into adjacent surface waters.
High  waste  concentrations  also  inhibit  microbial
degradation and extend the threat of lateral  transport
beyond the effective lifetime of the containment.  During
the period of containment, the organic wastes can also
penetrate deeper through fractures or discontinuities in
natural impermeable barriers.  These deeper penetrations
of organic wastes can contaminate underlying aquifers and
can result in lateral transport of organic wastes underneath
the surface containment.
Excavation can only remove organic waste accumulations
above natural barriers.  Any deeper penetrations of high
waste concentrations  remain  as long-term sources of
groundwater contamination. During excavation, workers
and adjacent  residents are exposed to vapor emissions
from the saturated organic wastes.  Exposing the bedrock
may also result in long-term vapor emissions from deeper
waste accumulation. In this case, removal of the natural
cover can  actually hinder  subsequent recover}' of the
deeper waste accumulations.

In situ recover}? and treatment of organic wastes restores
both the subsurface soils  and bedrock  to the original
condition before contamination.  In this two-step process,
organic  wastes are first immobilized by  reducing waste
concentrations  to  residual  saturation,  and  then the
immobile wastes are degraded microbially.  During the
initial recover}' of organic wastes, lateral containment of
the site prevents contamination of adjacent surface waters.
However, long-term maintenance of the lateral containment
is avoided by immobilizing the organic wastes in the first
step of the  restoration. WRI has developed the CROW
process  to  accomplish this  crucial  first step in the
restoration of organic waste sites.

The initial  developmental work for the CROW process
was completed under the U.S. Environmental Protection
Agency (EPA) Superfund Emerging Technology Program.
The development consisted of  several one- and  three-
dimensional physical simulations of the process. Based on
the laboratory  performance of the process,  the  EPA
advanced  the  process  to  the Superfund Innovative
Technology Evaluation (SITE) Demonstration Program,
and a field-scale  demonstration  was  conducted (the
subject of this report). Further development of the process
has been funded by the  U.S. Department of Energy (DOE)
as two separate projects. The first project was a study to
determine the enhancement of the process performance by
                                                    66
-------
the addition of a small volume of biodegradable chemical
to the injected  hot water.   The second  project is  the
application of the process  to  LNAPL and chlorinated
hydrocarbon contamination.

A.2   Technology

The general technical description describes the process as
it applies to denser-than-water organic liquid, which will
be the hardest to handle. However, the CROW process is
applicable to a wider range of organic liquid densities as
covered in the patent for the process.

CROW uses hot-water or steam displacement with or
without chemical addition to reduce concentrations of
organic wastes in  subsurface  soils  and underlying
bedrock.  In the  CROW process, downward penetration of
dense organic liquids is reversed by controlled heating of
the subsurface to suspend organic wastes in the w7ater. The
buoyant wastes are displaced  to  extraction  wells by
sweeping the subsurface with hot water. Waste flotation
and vapor emissions are controlled by maintaining both
temperature and concentration gradients near the ground
surface.   Reducing waste  concentrations to residual
saturation immobilizes  the organic wastes and promotes
complete  restoration  of  the   site  using  microbial
degradation.

Wastewater treatment is minimized in the CROW process
by reinjecting water that is  recovered with the  organic
wastes. The hot-water or steam displacement produces a
mixture of organic waste and water. After separation, the
recovered water  is heated and  reinjected  into  the
subsurface above  any aquifers.  This reinjected water is
contained laterally, and cannot permeate downward when
there  is a rising flow of hot water from steam injection
beneath the  recovery operation.  During the recover}7
operation, the net wastewater production corresponds only
to steam condensation and groundwater influx.  Only this
net production of wastewater is treated for discharge or
boiler feed.

Vapor emissions are controlled in the CROW process by
maintaining both temperature and concentration gradients
in the injected w7ater near the ground surface.   At some
waste sites, subsurface  heating may result in appreciable
vapor pressures  of phenols or other volatile components in
the organic wastes.  The concentrations of water-soluble
volatiles  can also  increase  when recovered water is
reinjected without  treatment.   If  the  cooler  water
temperatures near the ground surface do not lower vapor
pressures  sufficiently,  steam-stripping can reduce  the
concentration of volatiles in the water that is reinjected at
the top level of the recovery process. In this case, the water
barrier to oil flotation also serves as an absorber to reduce
vapor emissions during recovery operations.

A.3                   of the CROW
       Technology

The  CROW process  has many  benefits over other
technologies that remediate contaminated aquifers.  The
benefits of the CROW process are listed below.

 •  Organic recovery reduces waste management costs in
    comparison with either conventional containment or
    excavation. In some cases, the recovered  organic is
    reusable and offsets a portion of the processing costs.

 *  By lowering waste concentrations  in the subsurface,
    the process promotes natural  restoration and reduces
    the duration of long-term containment.  The injection
    and production wells may also be used to inject air or
    oxygen  and  nutrients for  accelerating  microbial
    degradation of residual organic wastes.  During this
    restoration, groundwater contamination is avoided
    because the residual wastes have been leached and are
    no longer mobile.

 •  Controls  groundwater contamination  and  vapor
    emissions by thermal gradients and well placement.

 •  Not limited by depth, deep contaminations even in the
    bedrock  can  be   treated   as  well   as  shallow
    contaminations.

 *  Because it is an in situ process, there is minimization
    of the exposure of contaminated material  at  the
    surface or to personnel.

 *  Should be considerably faster for remediation of a site
    than most other processes.

 •  The  process  can  also  remediate an  area  where
    buildings or active  facilities exist without  disrupting
    the on going work or removing the buildings.

 •  No specialized equipment is required.
                                                    61
-------
A.4   Treatment

The  treatment  systems  for  the  CROW  process are
designed as modular units.  Each unit of the system is a
commercially available  unit.   By using  commercially
available units, lower equipment costs can be realized over
noncommercial  units.

A.5
       Applications

General considerations for the  CROW process are (1) the
contaminating organic should be free-phase in the aquifer
or liquid saturated zone,  (2) the hydraulic conductivity of
the zone must be sufficient to allow sustained injection and
production from  multiple wells,  and  (3)  the  physical
characteristics and depth of the contaminate area. For the
best  performance,  the  organic  phase should  be  a
continuous, free-fluid phase overthe treatment area.  If the
contaminate is in  dispersed pockets throughout the area,
the CROW process would remove some of the organics,
but may  spread the remaining organics over the  sw7ept
area. The spreading of the organics may not be a drawback
because the  organic will be immobile and at the lower
saturation it will be more amenable to natural or induced
biorcmcdiation.

Chlorinated hydrocarbons, coal tars, and heavy petroleum
products  are common examples of dense organic liquids.
Coal tars have been produced  as byproducts in the
manufacture of town gas and in coking operations by the
steel industry.   Creosote derived from  coal  tar and
pentachlorophenol mixed with diesel oils have been used
as wood treating preservatives. The petroleum  refining,
storage,  and transportation  industries  have produced
assorted petroleum based contaminants. At a variety of
sites throughout the United States, these complex mixtures
of dense  organic liquids have  leaked from tanks, ponds.
ditches or other impoundments and have accumulated as
organic wastes in the saturated groundwater environment.

If further remediation of a site is required, the use of
bioremediation  following the  CROW process has been
evaluated by  Remediation   Technologies  Inc.,  now7
ThermoRetec   (ReTeC).   ReTeC was  successful in
evaluating in situ bioremediation processes for treatment
of CROW conditioned  soils.   In addition, biological
treatment of  CROW  process   product  water   was
demonstrated. Biological treatment of the CROW product
water was incorporated into the Stroudsburg, Pennsylvania
project.

A.6

The cost for application of the CROW technology is highly-
dependent upon the site characteristics and size, and the
extent of the process monitoring required.  Generally, the
larger the site, the lower the treatment cost per cubic yard
of contaminated soil. To give an idea of the cost range, two
sites have projected costs of $34 and $350 per cubic yard
of contaminated  soil for a  2.6  and  0.2  acre  areas,
respectively.   Both sites have a 20-  to 30-foot-thick
contaminated zone within  a  highly permeable aquifer.
The use of the CROW process for a given site should be
highly  competitive with other  processes, if not more cost
effective.

A.7

Full-scale remediations  of a wood treatment  site  is
presently being conducted with the remediation of two
manufactured gas plant (MGP) sites completed.  One of
the   MGP  sites  is  the  Superfund  Site  located  in
Stroudsburg, Pennsylvania and is the subject of this report.
The other MGP site is located  in Columbia, Pennsylvania
and consisted of the remediation of a cement capped, 60-
foot-diameter by 27-foot-deep former gas relief holder.
The project objective was the removal of a portion of the
coal tars from the debris filled holder  followed by the
stabilization of the holder with  grout. Closure of the site is
presently in front of the EPA.

The wood treatment project is  located at the Bell Lumber
and  Pole  facilities  in  New  Brighton,  Minnesota.
Construction and operation of the facility is being done by
the site owner. WRI designed the treatment scheme, and
monitors and evaluates the operations.   The full-scale
remediation is being carried out as a staged remediation
using three five-spot patterns with each pattern flushed in
a sequential order. Prior to implementing the full-scale
project, a pilot test was operated to demonstrate the
containment and organic removal features of the CROW
process. The pilot test was a success in both areas. As of
November  1998,  over  55.000  gallons  of  creosote,
pcntachlorophcnol  (PCP), and a petroleum carrier fluid
have been extracted from the first pattern.  Half of the
recovered organics have been reused in the ongoing wood
treatment operations.
                                                    68
-------
Also, WRI has conducted treatability studies for wood
treatment MGP, Brownfield, and chemical waste sites in
several states. Included in the laboratory studies have been
preliminary testing of chlorinated hydrocarbon remediation
with excellent results.  Laboratory tests of the process
conducted for the  U.S.  EPA and private clients using
materials from MGP and wood treatment sites indicated
that 60 to 70 percent of the MGP contaminant and 84 to 94
percent  of the  wood  treatment contaminant can be
recovered at  the optimum  water flushing temperature.
Additionally, removals of 90 percent or greater  can be
achieved for MGP materials if surfactants, at 1 percent or
less by volume, are incorporated into the flush water.
                                                    69
-------
                                                            B
                                        of
Western  Research Institute (WRT) tested  the  CROW
process  effectiveness  in  the  laboratory-  using one-
dimensional and three-dimensional displacement tests.
The  reactor system  used for  the  one-dimensional
displacement tests was the tube reactor shown in Figure
Bl-1.  The disposable  chlorinated poly vinyl chloride
reactor tube was uniformly packed with contaminated soil
from the  Brodhead Creek site and was vertically oriented
within  a series of insulated shield heaters.  Auxiliary
equipment  included inlet water injection and metering
devices, a water heater, product collection equipment, and
a gas chromatograph. The entire system was connected to
a data acquisition computer that recorded temperatures,
pressures, and flow rates (WRI 1990).

WTater was  metered into the bottom of the reactor by a
positive displacement  pump. The injected water passed
through a heater to generate steam or hot water.  Product
water  samples  were  collected  from  the  automatic
sampling valve system.  Product gas was collected from
the sample  vessels, and gas  composition was analyzed as
needed by an on-line gas chromatograph.  After each test,
the treated sample was extruded  from the tube  for
sampling (WRT 990).  The  experimental apparatus used
for the three-dimensional tests  was a large, thick-walled
vessel  into which an  encapsulated sample  was placed
(Figure Bl-2).  The vessel was sealed using screw-on
domed ends.  The fluid-handling system consisted of an
injection system,  a  product-collection  and sampling
system,  and a product  gas-collection  and  analyzing
system.  The reactor instrumentation and control system
consisted of flow  controllers  and  meters;  pressure
regulators and transmitters; microprocessors; recorders;
thermocouples;   gas-analysis  equipment;   and  a
minicomputer for data collection, storage, and analysis.
The injection array for the sample consisted of multiple
injection points to simulate the entire CROW concept.
Process-water samples were routinely collected to provide
information on organic production.
The samples forthe three-dimensional tests were placed in
the reactor from bottom to top in the following order: (1)
a base of grout to simulate an impermeable barrier, (2) a
layer of resaturated site material to represent an oily waste
accumulation, (3) a layer of material from the Brodhead
Creek site, and (4) clean, dry sand to represent the vadose
zone.  Three wells  were inserted  in  the sample, a
production well in the center of the sample and an injection
well at each end of the sample. The injection wells were
equipped for cold-water injection in the upper 4 inches and
hot-water injection in the bottom 8 inches. The cold-water
layer was intended to prevent the heated contaminants
from rising to the vadose region (WRI 1990).

The results of the one-dimensional tests indicated that the
percent reduction in oily saturation measured in the treated
soil  increased up to 61.8 percent  as the  temperature
increased to approximately 156 °F.  Results  of the  one-
dimensional tests with additions  of surfactants indicated
oily  saturation reduction of 87 percent at 156 °F.  The
three-dimensional  test without addition  of surfactants
produced a 60 percent reduction  in the oily saturation at
140°F (WRI 1990).

References

    Western   Research   Institute,   1990.    Contained
    Recover of Oily Wastes (CROW) Final  Report.
    August.
                                                    70
-------
    SOURCE: Modified from WR11990
                                                                                    PRESSURE
                                                                                      GAUGE
                                                                 PRESSURE   i
                                                                TRANSDUCER Cl>	
                                         INSULATED SHIELD
                                             HEATERS
                         WATER FEED
                            TANK
                                 PRESSURE
                                TRANSDUCERS
                                     HEATER
                   METERING
                 WATER PUMP
                                                                       PACKED
                                                                       COLUMN
                                                                       (4"X36")
SAMPLING •
 VALVES
                   SAMPLE
                 CONTAINERS
                                                                     THERMOCOUPLES
Figure Bl-1. Bench-scale tube reactor.
-------
           SOURCE: Modified from WR11990
                             GAS

                           STEAM
to
SHELL
                                               PRESSURE SHELL
VENT
             N2 GAS PRESSURIZATION
                                                 REACTION BOX
                                                    SAMPLE
                                                          PRODUCTION
                                                             LINES
                                                                              KNOCKOUT
                                                                               POTS (4)
                                                                         I
                                                                 LIQUID
                                                              COLLECTION
                                                                                                              STACK GAS
       Figure Bl-2. Bench-scale experimental unit.
-------
                   c
at the               in
              73
-------
                                                 WRI-97-R010
CROW™ FIELD DEMONSTRATION WITH
BELL LUMBER AND POLE


Topical Report
By
L. John Fahy
Lyle A. Johnson Jr.
April 1997
Work Performed as Jointly Sponsored Research
Under Cooperative Agreement
DE-FC21-93MC30127 Task 13
For
Bell Lumber and Pole
New Brighton, Minnesota
and
U.S. Department of Energy
Office of Fossil Energy
Federal Energy Technology Center
Morgantown, West Virginia
By
Western Research Institute
Laramie, Wyoming
                            74
-------
                     ACKNOWLEDGMENT AND DISCLAIMER

       This report was prepared with the support of the U.S. Department of Energy (DOE), Federal
Energy Technology  Center,  under Cooperative Agreement Number DE-FC21-93MC30127.
However, any opinions, findings, conclusions, or recommendations expressed herein are those of the
authors and do not necessarily reflect the views of the DOE.

       This report was prepared as an account of work sponsored by an agency of the United States
Government.  Neither the United States Government nor any agencies thereof, nor any of its
employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility
for the accuracy,  completeness, or usefulness of any information, apparatus, product, or process
disclosed, or represents that its use would not infringe on privately owned rights. Reference herein
to any specific commercial product, process, or service by trade name, trademark, manufacturer, or
otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by
the United States Government or any agency thereof. The views and opinions of authors expressed
herein do not necessarily state or reflect those of the United States Government or any agency
thereof.
                                           75
-------
                        TABLE OF CONTENTS







                                                              Page




LIST OF TABLES AND FIGURES  	   iv




EXECUTIVE SUMMARY	   v




INTRODUCTION	   1




BELL POLE PROJECT CHRONOLOGY 	   1




SITE CHARACTERIZATION	   3




TREAT ABILITY TESTS	   4




PILOT TEST OBJECTIVES	   7




PILOT TEST DESCRIPTION 	   7




PILOT TEST RESULTS	   8




PILOT TEST CONCLUSIONS AND RECOMMENDATIONS	   9




CROW TEST PROCEDURES AND DESIGN	   10




WELL NETWORK DESIGN	   10




SURFACE TREATMENT SYSTEM DESIGN	   11




CONTROL AND DATA ACQUISITION SYSTEM	   11




PROJECT OPERATION	   12




DISCUSSION	   15




REFERENCES	   20
                                76
-------
                          LIST OF TABLES AND FIGURES

Table                                                                          Page

  1.       Process Simulations for Bell Pole  	   6

  2.       Pilot Test Operating Conditions and Results  	   8

  3.       Water Disposal PAH Concentration  	   13

  4.       Process Train Hydrocarbon Sampling, March 13, 1996	   14

  5.       Bell Pole CROW Test Summary,
          January 3, 1995 through February 25, 1996	   15

  6.       Bell Pole CROW Test Summary,
          February 26, 1995 through December 31, 1996	   16



Figure                                                                          Page

  1.       Phase 1 Well Pattern	   5

  2.       Treatment System Conceptual Design	   12

  3.       Aquifer Areal Temperature Contour Map 	   18

  4.       Monitor Well BP27 Aquifer Temperature Profile  	   19
                                         77
-------
                                EXECUTIVE SUMMARY

       In 1990 efforts were initiated to implement an in situ remediation project for the contaminated
aquifer at the Bell Lumber and Pole Company (Bell Pole) Site in New Brighton, Minnesota.  The
remediation project involves the application of the Contained Recovery of Oily Waste (CROW™)
process, which consists of hot-water injection to displace and recover the nonaqueous phase liquids
(NAPL).

       While  reviewing the site  evaluation information, it  became  apparent that  better site
characterization would enhance the outcome of the project. Additional coring indicated that the areal
extent  of the contaminated  soils was approximately eight times greater than initially believed.
Because of these uncertainties, a pilot test was conducted, which provided containment and organic
recovery information that assisted in the design of the full-scale CROW process demonstration.

       Based on the results from the pilot test the following conclusions were made:

1. The  pilot test provided sufficient  hydraulic information to  design the full-scale  CROW
   remediation system.  The pumping test portion of the pilot test  indicated uniform aquifer
   properties.  The  entire thickness of the aquifer reached  the target  temperature  range, and
   containment of the injected hot water was achieved.

2. Pre-test injection  and extraction rate predictions were achieved.

3. The post-test soil boring data indicated hot-water injection displaced more than 80% of the
   NAPL near the injection well. The data indicates that a NAPL saturation of approximately  19%
   (pore volume basis) and a 500-fold decrease in pentachlorophenol (PCP) concentration can be
   achieved with 20  pore volumes of flushing.

4. The produced water treatment system used during the pilot test was effective in reducing  PCP
   and polynuclear  aromatic hydrocarbon (PAH) compounds to concentrations acceptable for
   sanitary sewer discharge.

5. The  microbial assay of  the  post-test samples  found  an  encouraging increase in microbial
   population compared to data collected before the pilot test.

       Based on the results from the pilot test, conditions and procedures  were developed for
implementing a full-scale CROW process demonstration to remediate the remaining contaminated soil
at the Bell Pole site.
                                            78
-------
       After reviewing the  cost  ramifications  of implementing the  full-scale CROW field
demonstration, Bell Pole approached Western Research Institute (WRI) with a request for a staged,
sequential site remediation. Bell Pole's request for the change in the project scope was prompted by
budgetary constraints.  Bell Pole felt that even though a longer project might be more costly, by
extending the length of the project, the yearly cost burden would be more  manageable.

       After considering several options, WRI recommended implementing a phased approach to
remediate the contaminated area. Phase 1 involves a CROW process demonstration to remediate the
upgradient one-third of the contaminated area, which is believed to contain the largest amount of free
organic material.

       The Bell Pole Phase 1 CROW demonstration is operating satisfactorily.  However, due to
equipment problems, the system  is operating at less than the design conditions and is unable to
operate continuously for extended periods of time. Only two pore volumes of hot water and two
pore volumes of cold water were injected during 1996. By the end of 1996, over 20,000 gallons of
oil had been transferred to the oil storage tank. Bell Pole has also used about 6000 gallons of the
produced oil in its pole treating operation.
                                           79
-------
                                   INTRODUCTION

       Beginning in 1990, efforts were initiated for Western Research Institute (WRI) to implement
an in situ remediation project for the contaminated aquifer at the Bell Lumber and Pole Company
(Bell Pole) Site in New Brighton, Minnesota.  The remediation project involves the application of the
Contained Recovery  of Oily Waste (CROW™) process, which consists of hot-water injection to
displace and recover the non-aqueous phase liquids (NAPL) (Johnson and Sudduth 1989).

       Wood treating activities began at the Bell Pole Site in 1923 and have included the use of
creosote and  pentachlorophenol (PCP) in a  fuel oil carrier.  Creosote was used as  a wood
preservative from  1923 to 1958.  Provalene 4-A, a non-sludging fuel-oil-type carrier for PCP, was
used from 1952 until  it was no longer commercially available in 1968.  A 5—6% mixture of PCP in
fuel oil has been used as a wood preservative since 1952, and a fuel-oil-type carrier, P-9, has been
used since 1968.

       While  reviewing the site evaluation  information, it became  apparent  that better site
characterization would enhance the outcome of the project.  Additional coring indicated that the areal
extent  of the contaminated  soils was approximately eight times greater than initially believed.
Because of these uncertainties, a pilot test was conducted, which provided containment and organic
recovery information  that assisted in the design of the full-scale CROW process demonstration.
1979

1983
       BELL POLE PROJECT CHRONOLOGY

Five monitoring wells were installed by Bell Pole and MacGillis-Gibbs Company.

The Bell Pole New Brighton site was placed on the EPA National Priorities List.
Bell Pole signed a consent order and agreed to voluntary site remediation and
began site cleanup and removal of disposal areas.
September 1985   The  groundwater purge  well, PW-1 was installed and pumping tests  were
                 conducted. Bell Pole subsequently pumped approximately 2000 gallons of free
                 organic product over the next few years.
April 1986


February 1989
Conestoga-Rovers & Associates  Limited (CRA) completed the "Remedial
Investigation Phase One Report" for Bell Pole.

Bell Pole constructed a rotary kiln incinerator and completed soil incineration
operations at the Bell Pole site east yard.
                                            80
-------
December 1989  Western Research Institute and Bell Pole submitted a proposal to the Department
                of Energy (DOE) and was awarded funding for a Jointly Sponsored Research
                (JSR) project to apply CROW process technology to remediate the Bell Pole New
                Brighton site.
March 1990
August 1990
February 1991


April 1991

April 1991


June 1991
CRA completed a site soil boring study indicating the contaminated area was
about two acres.

CRA and WRI completed for Bell Pole an Interim Response Action Work Plan
which was submitted to the Minnesota Pollution Control Agency (MPCA) and the
Minnesota Department of Health (MDH). This document proposed conducting
a 30-day,  two-well pilot test to demonstrate the feasibility of using the CROW
process to remediate the Bell Pole site.

CRA submitted for Bell Pole an Interim Response Action Work Plan for process
area soil removal.

Bell Pole began operations in its new process plant.

At the request of the MPCA, CRA and Bell Pole submitted an Application for
Variance Interim Response Action for the CROW process source remediation.

Approval to conduct the two-well pilot test of the CROW process was granted
by the MPCA and MDH.
September 1991  The two-well pilot test was initiated.

November 1991  The two-well pilot test was completed and the system equipment dismantled.

June 1992       CRA and WRI submitted the Bell Pole CROW 30-Day Pilot Test Report to the
                MPCA.

July 1992        The Bell Pole CROW 30-Day Pilot Test Report was found acceptable by the
                MPCA staff.

August 1992     WRI submitted the Bell Pole Pilot Test Evaluation report to DOE.

August 1992     Bell Pole completed the incineration of the process area contaminated soil located
                above the water table.

June 1993       CRA submitted for Bell Pole a draft of the final design report of CROW and a
                plan for a phased implementation of the CROW process.

July 1993        CRA and WRI submitted for Bell Pole the final design report of CROW to the
                MPCA.
                                          81
-------
August 1993     Bell Pole submitted a permit application to construct the CROW/maintenance
                 building.

November 1993   WRI and CRA completed the drilling and installation of six injection and three
                 monitoring wells at the Bell Pole site.

February 1994    CRA and Bell Pole submitted an application for a variance interim response action
                 to extend the previous variance.

August 1994     WRI completed fabrication of a data acquisition and control system for use by
                 Bell Pole during the CROW field demonstration.
March 1995


May 1995

July 1995


February 1996



March 1996

July 1996
Construction for the CROW process system was completed, and groundwater
extraction was initiated on a limited basis.

Hot-water injection was initiated.

Continuous injection/extraction was terminated because sewer discharge criteria
were not being met.

A hydrogen  peroxide injection system was added to the water cleanup system,
which  resulted  in meeting  discharge criteria.  Groundwater extraction  was
restarted.

Hot-water injection was restarted.

Heat exchanger failure occurred. Cold-water injection and extraction continued.
November 1996   Injection and extraction were terminated because of emulsion problems in the
                 oil/water treatment system.
                             SITE CHARACTERIZATION

       Site characterization of the contaminated area at the Bell Pole site has been conducted for
several years by Conestoga-Rovers & Associates Limited (CRA) and other consultants.   The
contaminated soil is contained in the New Brighton Formation (Stone  1966). It has been described
as a relatively uniform silty fine-medium grain sand, 23 to 47 feet  thick (CRA  1986).   The
contaminated soil is underlain by the Twin Cities Formation, which is  a silty to sandy clay till.  The
New Brighton Formation is highly permeable, with hydraulic conductivities  in the range of 3.1x103
to 9.5xlO"3 cm/sec.  Conversely, the underlying Twin Cities Formation has low permeability, with a
conductivity on the order of l.OxlO"7 cm/sec (CRA 1986). The underlying clay till has provided an
                                           82
-------
effective lower boundary to fluid migration and has been responsible for limiting the downward
migration of the organic material.

       A continuous aquifer  lies  at  a  depth of 10 to 20 feet below  ground surface (BGS).
Groundwater flows radially from a pond, located to the northeast, at a velocity of 0.1 to 0.6 ft/day.
Across the Bell Pole site the groundwater gradient is 0.004 ft/ft toward the southwest, where the
water appears to discharge into a drainage ditch.

       In early 1990, 22 boreholes were drilled to define the extent of the contamination.  Later, in
preparation for the two-well pilot test, one new injection well and three  monitor wells were also
drilled and cored. Based on the evaluation of the coring data, it appears that the contaminated or
saturated interval has an elongated teardrop shape which dips toward the northeast (Figure 1). The
maximum thickness in the center  of  the zone is approximately  25 feet, while the edge of the
contaminated zone is only a foot or two thick.
                                TREATABILITY TESTS

       While the coring operations were being conducted, two large samples of contaminated soil
were collected.  These samples were used to conduct laboratory treatability tests. These flushing
tests were necessary to appraise the effectiveness of the CROW process at this site and to determine
operating conditions.

       For each flushing test, approximately 30 Ib of the contaminated site material was packed into
a 3.75-in. diameter by 36-in. long reactor tube. The reactor tube was then placed vertically within
the reactor shell.  During the packing of each reactor tube, a composite sample of the packed material
was prepared for  organic loading determination. Each test was conducted by establishing water flow
at the desired flow rate through the bottom of the tube with the flush water produced from the top
of the tube.

       Two tests were conducted, one each at a nominal 120°F and 140TF.  The operating conditions
and results for the two flushing tests are listed in Table 1. The reduction in the organic saturation was
essentially the same, 0.53 and 0.54 wt %, even with the variance in the weight percent oil for the pre-
test samples ranges, 2.87 to 7.44%.

       The initial and post-test samples submitted for PCP analyses show that the decrease in PCP
concentration during  the flushing tests was  higher than the decrease  in  the  total  oil phase
concentration.
                                            83
-------
                                                                                                                                                     iESEMQ




                                                                                                                                                     MONITORING WELL LOCATION




                                                                                                                                                     PURGE WELL LOCATION




                                                                                                                                                     INJECTION WELL LOCATION




                                                                                                                                           	2O	ESTIMATED NAPL  THICKNESS




                                                                                                                                                O    PROPOSED MONITORING WELL LOCATION




                                                                                                                                                     PROPOSED PURGE WELL LOCATION




                                                                                                                                                *    PROPOSED INJECTION WELL LOCATION
I532{L)-JUNE «, »3-REV.O-(P-01)CMN)
                                                                  Figure 1.  Phase I Well Pattern
-------
                       Table 1. Process Simulations for Bell Pole
Test Number	103	104

Water Injection
        Temperature, °F                           140                   120
        Flux, cmVmin             -               107                   118
        Velocity, cm/min                         2.5xlO'2               2.8xlO'2

Porosity, %                                       35.5                   33.6

Initial Oil Saturation of Mobile Oil Zone
       % Pore Volume                            42.2                   16.2
       wt %                                       7.44                  2.87

Residual Oil Saturation
       % Pore Volume                            10.0                   10.0
       wt %                                       0.54                  0.53

Removal of Oil, wt %                              93.5                   84.3

PCP Concentration, ppm
       Initial Material                              3200                   1500
       Flushed Material                            2.3                  BDLa
       % Reduction                               99.9                   99.8b

a      BDL = Below detection limit
b      Value based on the flushed material for test 103
                                           85
-------
                              PILOT TEST OBJECTIVES

       An Interim Response Action (IRA) work plan was prepared in 1990 by CRA and WRI. The
IRA detailed how the CROW process would be implemented at the Bell Pole Site (CRA and WRI
1990). Based on the IRA and after the granting of variances by the Minnesota Pollution Control
Agency and the Minnesota Department of Health, a two-well pilot test of the CROW process was
conducted. The test consisted of injecting hot, potable water into the NAPL-saturated area of the
aquifer, producing groundwater (and NAPL) from an existing extraction well, PW1, and treating the
produced water for sanitary sewer discharge.

       The objectives of the pilot test were to:

1. Compare predicted injection and extraction rates with actual field data;

2. Demonstrate the ability to heat the aquifer to the 120°F to 140°F range;

3. Demonstrate the ability  to  hydraulically control  the  injected water to  prevent spreading
   contamination;

4. Confirm treatment system effectiveness in reducing PCP and polynuclear aromatic hydrocarbons
   (PAHs) prior to sanitary sewer discharge; and

5. Predict anticipated operating conditions for full-scale CROW application.


                             PILOT TEST DESCRIPTION

       The pilot-test location was selected from the site characterization mapping and the location
of the existing extraction well, PW1. One new injection well, IW1, was drilled 50 feet upgradient to
the northeast from well PW1. Both the injection and extraction wells were located in an area that
contained high organic accumulations (Figure 1).

       The pilot test began on September 24, 1991. The first step of the test involved pumping the
extraction well, PW1. Treatment of water began on September 26, day 3 of the test. Hot-water
injection started on day 7 at an initial injection temperature of 147°F. On day 9,  the injection
temperature was increased to 203 °F.  Injection was terminated on October 31, day 37 of the pilot
test. Pumping continued at PW1 until day 41 when the test ended. Water treatment continued until
day 45, and the treatment system was subsequently dismantled.
                                           86
-------
                               PILOT TEST RESULTS

       Flow rates and injection pressures were recorded by the data acquisition system.  The
pumping rate at PW1 was started at 5 gpm and stepped up to 9 gpm during the seven days prior to
injection startup. During the remainder of the test, PW1 averaged 6.5 gpm (Table 2).
                  Table 2.  Pilot Test Operating Conditions and Results


Total Hot-Water Injection Time                                           30 days

Average Hot-Water Injection  Rate                                         4.5 gpm

Steady-State Hot-Water Injection                                         200°F
   Wellhead Temperature

Total Water Injected                                                     193,000 gallons

Total Water and NAPL Production Time                                   41 days

Average Fluid Production Rate                                            6.5 gpm
   During Hot-Water Injection Phase

First Pumping Test Production Rate                                       5.0 gpm

Second Pumping Test Production Rate                                     9.0 gpm

Total Fluids Produced                                                    390,000 gallons

Total NAPL Production                                                  2000 gallons

Areal Extent of Injected Water                                            3285 ft2

Time to NAPL Production Response                                      14 days
   From Start Of Injection

Time to Breakthrough from                                               20 days
   Start of Hot-Water Injection

Average Hot-Water Injection                                              2.5 ft/day
   Front Velocity, ft/day
                                          87
-------
3.    The post-test soil boring data indicated hot-water injection displaced more than 80% of the
      NAPL near the injection well. The data indicates that a NAPL saturation of approximately
      19% (pore volume basis) and a 500-fold decrease in PCP concentration can be achieved with
      20 pore volumes of flushing.

4.    The produced water treatment system used during the pilot test was effective in reducing PCP
      and PAH compounds to concentrations acceptable for sanitary sewer discharge.

5.    The microbial assay of the post-test samples found an  encouraging increase in microbial
      population compared to earlier data collected before the pilot test.


                       CROW TEST PROCEDURE AND DESIGN

       Based  on. the results from the pilot test, conditions and procedures were developed for
implementing a full-scale CROW process demonstration to remediate the remaining contaminated soil
at the Bell Pole site.

       After  reviewing the  cost ramifications of implementing the full-scale CROW field
demonstration, Bell Pole approached WRI and the MPCA with a request for a staged, sequential site
remediation.  Bell Pole's request for the change in the project scope was prompted by budgetary
constraints. Bell Pole felt that even though a longer project might be more costly, by extending the
length of the project, the yearly cost burden would be more manageable.

       After considering several options, WRI recommended implementing a phased approach to
remediate the contaminated area.  Phase 1 involves a CROW process demonstration to remediate the
upgradient one-third of the contaminated area, which is believed to contain the largest amount of free
organic material. The phased approach to remediating the site is not expected to cause any adverse
effects except for extending the time required to complete the entire project.


                              WELL NETWORK DESIGN

       During 1993, WRI drilled four, Phase 1 injection wells and three monitoring wells, plus two
Phase 2 injection wells, which are being used as downgradient monitoring wells during the first phase.

       By using the existing extraction well, PW1, and the new injection wells, an inverted five-spot
pattern was installed (Figure 1). Due to its pre-existing location, PW1 is closer to the downgradient
injection wells than to  the upgradient injection wells, which is anticipated to enhance the overall
-------
capture efficiency of the system. Injection-to-extraction well spacings are approximately 100 feet,
which is about twice the spacing utilized during the pilot test.


                      SURFACE TREATMENT SYSTEM DESIGN

       Based on results from the pilot test, plus bench-scale tests conducted by Bell Pole and various
vendors, a produced fluid treatment system was designed and installed.

       During the pilot test it was observed that a significant amount of oil/water separation was
occurring in the 40,000-gallon tank into which all produced oil and water was being pumped. To
capitalize on this occurrence, all produced water and oil is pumped into a 40,000 gallon process tank
after sulfuric acid has been added to lower the pH to approximately 3.5. Oil is skimmed from the top
of the tank and pumped off of the bottom of the tank and then routed to an oil storage tank.  This is
a batch operation that is performed daily.

       Water is continuously pumped from the 40,000 gallon process tank to an air flotation unit
where the oily water is aerated and most of the remaining oil and grease, PCP, and organic carbon
are removed and recycled back to the 40,000 gallon process tank.

       The treated water leaving the air flotation unit is treated with sodium hydroxide, then pumped
to a 10,000-gallon equalization tank. From this tank, part of the water, 5 to 10 gpm, is pumped to
an ozonation unit, which removes the PCP.  The water is then treated with hydrogen peroxide to
break down the remaining PAH compounds and is disposed of in the sewer. The water that is not
pumped to the ozonation unit is recycled through a boiler/heat exchanger system where it is heated
and reinjected. The conceptual design of the water treatment system  is shown in Figure 2.

       Prior to installing the CROW process system, Bell Pole installed a two-well pump and
treat system.  The water produced from the pump and treat wells enters the 10,000-gallon
equalization tank and is either treated for disposal or reinjected.


                   CONTROL AND DATA ACQUISITION SYSTEM

       For the Bell Pole Phase 1 CROW demonstration, WRI developed and installed a control and
data acquisition system (CDAS). This system collects all temperature, pressure, flow, and pH data
generated by the process.  From this data, the CDAS  determines what type of control should be
exerted on the process. If required, the CDAS will turn a  pump, valve, or alarm on or off as specified
by the control logic.
                                           89
-------
  PW1
Extraction
 40,000
 Gallon
Process
  Tank
pH = 3-4
 H2SO4
                     40,000
                     Gallon
                       Oil
                     Storage
                      Tank
  Air
Flotation
  Unit
                              NaOH
                                                            NaOH
                                                                    H2O2
  10,000
  Gallon
Equalization
   Tank
 pH  = 5-6
                                     To Injection
                                        Wells
Ozonation
   Unit
 pH = 6-8
                                                 To
                                               Sanitary
                                                Sewer
                     Figure 2. Treatment System Conceptual Design
       In addition to controlling the physical process, the CDAS also displays the status of the

various parameters on the computer monitor through the use of several computer screens. The

system also records the status of these parameters to computer files, which are routinely downloaded

via the modem system for analyses and archiving.


       From the beginning, the CDAS  system operated basically  as designed.   However, the

computer had a tendency to "hang up" occasionally. In October, 1995, an upgrade of the control

system and Windows 95.0 were installed. These upgrades have eliminated the previous problems,

and the system has been operating trouble free.
                                PROJECT OPERATION


       By early 1995, all of the equipment, except for the hydrogen peroxide system, had been

 installed. Water extraction began March 1995, and the system was operated intermittently through

 April 1995. On May 16, 1995, continuous operation of the CROW system began. Continuous hot-

 water injection was terminated on June 29, and continuous extraction and disposal of excess water

 was terminated July 12, 1995, because of failure to meet the sewer discharge criteria.
                                            90
-------
       The ozonation  unit was  originally designed for removal of PCP and  has  functioned
satisfactorily.  However, high concentrations of PAHs, particularly naphthalene and phenanthrene,
exceeding the discharge criteria  were occurring.  After several attempts to reduce the PAH
concentration  in  the discharge water,  the hydrogen peroxide injection system was installed
downstream of the ozonation unit. Hydrogen peroxide injection brought the PAH concentrations
down to acceptable discharge limits (Table 3).
                   Table 3.  Water Disposal PAH Concentration, mg/L
PAH Compound
Naphthalene
Acenaphthene
Fluorene
Pentachlorophenol
Phenanthrene
Anthracene
Fluoranthene
Pyrene 1400
Benzo-a-anthracene
Chrysene
Benzo-b-fluoranthene
Benzo-k-fluoranthene
Benzo-a-pyrene
2-Methylnapthalene
Before Hydrogen
Peroxide Injection
3900
530
'BDL
BDL
5200
340
1100
370
260
340
BDL
BDL
BDL
1400
After Hydrogen
Peroxide Injection
630
40
340
280
780
37
360
3000
57
110
23
60
14
290
Discharge
Limits
3000
3000
3000
3000
3000
3000
3000

3000
3000
3000
3000
3000
3000
Total PAH Concentration
14470
3391
10000
'Below Detection Limits
                                          91
-------
       Once it was demonstrated that the discharge criteria could routinely be met, the water
contained in the water treatment system was treated and disposed of, and extraction from PW1 was
restarted.  Continuous groundwater extraction was established February 26, 1996, and continuous
hot-water injection began a week later on March 4, 1996.

       On March  12, 1996, the entire water treatment system was analyzed for oil and grease
concentration and partially analyzed for PCP and total PAH concentration (Table 4).  The extraction
well, PW1,  oil and grease concentration was uncharacteristically low, suggesting groundwater
pumping prior to hot-water response has lowered the oil concentration in the immediate area. PW1
oil and grease concentrations had typically been in the 1000 to 3000 mg/L range when sampled. The
oil and grease concentration after the air flotation unit was reduced significantly compared to earlier
results and this is attributed to operating at a lower pH.
                     Table 4. Process Train Hydrocarbon Sampling
                                    March 13,1996


PW1 Effluent
       Oil and Grease Concentration, mg/L                    300

After Air Flotation Unit
       Oil and Grease Concentration, mg/L                     71

Injection Water
       Oil and Grease Concentration, mg/L                     96
       PCP Concentration, mg/L       10

Discharge Water
       PCP Concentration, mg/L       <1
       Total PAH Concentration, mg/L                        <4
       Continuous hot-water injection was terminated on July 15, 1996, following a heat exchanger
failure.  At that time, aquifer temperatures were approaching 120°F, and 70 °F water was being
produced at the extraction well. Cold-water injection and groundwater extraction continued while
efforts were made to replace the heat exchanger.  The entire system was shut down November 8,
1996, because of problems caused by oil/water emulsion in the water treatment system.
                                          92
-------
       During this shutdown period, the 40,000-gallon process tank is being heated, and the oil/water
emulsion is slowly being broken. The oil is being transferred to the oil storage tank and the produced
water treated and sent to the sewer. A new heat exchanger is being procured and should be available
soon, at which time the CROW system will be restarted.
                                     DISCUSSION

       The Bell Pole Phase 1 CROW demonstration is operating satisfactorily.  However, due to
equipment problems, the system is operating at less than the design conditions and is unable to
operate continuously for extended periods.  When the replacement heat exchanger is brought online,
efforts will be made to increase the injection temperature to the 195-200°F range, which will improve
the aquifer temperature response.

       Only two pore volumes of hot water and two pore volumes of cold water were injected during
1996.  Actual injection and extraction fluid rates have been 11 and 14 gpm, respectively. These
conditions are about half the original designed operating conditions.  At the current rates, it will take
another 30 months to complete 20 pore volumes of injection. The Bell Pole CROW test summaries
are shown in Tables 5 and 6.
                        Table 5. Bell Pole CROW Test Summary
                       January 3,1995 through February 25,1996
Total Water Injected, gal                222,811

Total Fluid Extracted, gal               642,138

Total Water Disposed, gal              543,315
       (Includes Off Pattern
       Pump and Treat Production)

Total Water Inventory in Tanks, gal      25,890
                                           93
-------
                        Table 6. Bell Pole CROW Test Summary
                     February 26,1996 through December 31,1996
Continuous Extraction Time, days                             257
Hot-Water Injection Time, days                               134
Cold-Water Injection Time, days                              116

Average Hot-Water Injection Temperatures
    Heater Temperature, °F                                  172
    Injection Manifold Temp, °F                              171
    IW4 Injection Line Temp, °F                             165
    Maximum PW1 Aquifer Temp, °F                          78
    Injection Well Aquifer Temp Range
       (measured 7/11/96), °F                               166-175
Total Hot-Water Injected, gallons                            4,103,856

Total Fluid Extracted, gallons                                5,288,544

Total Water Disposed, gallons
    (Including Off-Pattern Pump and Treat Production)         1,670,883


Average Pattern Water Injection Rate, gpm                      11.2

Average Pattern Water Extraction Rate, gpm                     14.1

Average Water Disposal Rate, gpm
    (Including Off Pattern Pump and Treat Production)           4.3

Individual Injection Well Flow Rates (Normalized Values), gpm
    IW2                                                    2.8
    IW4                                                    1.9
    IW10                                                   3.7
    IW12                                                   2.8

Cumulative Product Recovery Estimate, gal
    (excluding oil in Process Tank)  .                          20,000
                                         94
-------
       Based on the aquifer temperature measurements, an area! temperature contour map was
prepared (Figure 3). The high temperature front was arbitrarily defined by the 75°F temperature
contour. While there are a number of monitoring wells within the pattern area, the data are limited,
making the contours somewhat interpretive.

       However, the data does suggest some important trends. First, the hot-water injection period
has not progressed long enough to establish an interconnected hot-water front or fronts. Second, the
majority of the high temperature measurements in the pattern appear to be influenced by injection into
IW10.  However, the relatively low temperature response at PW1 indicates the extraction well was
mainly influenced by the injection at IW2.  Third, the more downgradient wells, IW4 and IW12, will
require a longer time and more injected pore volumes before they noticeably  affect the extraction
well, PW1. Fourth, the aquifer temperature data confirms that the injected water is contained within
the pattern area.

       Monitor well BP27, which is located on a line between wells IW10 and PW1, experienced the
greatest temperature response. Figure 4 shows the aquifer temperature profile at different times
before and after termination of hot-water injection. As expected, the aquifer is returning to ambient
temperature without the injection of hot water.

       Oil production has been estimated daily from the transfer of oil from the process tank to the
oil storage tank.  An actual daily rate has been difficult to determine because the oil remaining in the
production tank after oil transfer can only be estimated.  By the end of 1996, more than 20,000
gallons of oil had been transferred to the oil storage tank. Bell Pole has used about 6000 gallons of
the produced oil in its pole treating operation.
                                            95
-------
                                      BP7U .
                                     BP7L •
0BP31
 .BPZ!
                                                                                                                      .BP30
                             8P32,
                                   Figure 3. Aquifer Areal Temperature Contour Map
-------
   900 r-
                     Groundwater Elevation
   895
         13/19/96
   890
g
LU 885
   880
                                                             8/13/96
                                                9/10/96
   875.
          I     I    I    I    I    I    I    I    I    I    I    I    I    I	I    I
      50      60      70
80      90      100
 Temperature,  °F
110     120     130
             Figure 4. Monitor Well BP27 Aquifer Temperature Profile
                                    97
-------
                                    REFERENCES

Conestoga-Rovers and Associates Limited, 1986, Remedial Investigation Phase One Report Bell
   Lumber and Pole Company Site New Brighton, Minnesota, Unpublished Report, 58 pp.

Conestoga-Rovers and Associates Limited and Western Research Institute, 1990, Interim Response
   Action Work Plan Bell Lumber and Pole Company Site, New Brighton, MN, Unpublished Report,
   40 pp.

Johnson, L.A., Jr., and B.C. Sudduth, 1989, Contained Recovery of Oily Waste, United States Patent
   No. 4,848,460, 12 pp.

Stone, J.E., 1966, Surficial Geology of the New Brighton Quadrangle, Minnesota, University of
   Minnesota Press, Minneapolis, MN, 39  pp.
                                          98
-------