EPA/600/2-91/009
March 1991
NITRATE FOR BIORESTORATION OF AN
AQUIFER CONTAMINATED WITH JET FUEL
by
S.R. Hutchins, W.C. Downs, G.B. Smith, and J.T. Wilson,
R.S. Kerr Environmental Research Lab, US EPA, Ada, OK 74820
D.J. Hendrix, Solar Universal Technologies, Inc., Ground Water
Remediation Division, Traverse City, MI 49684
D.D. Fine, D.A. Kovacs, NSI Technology Services, Corporation
R.S. Kerr Environmental Research Lab, US EPA, Ada, OK 74820
R.H. Douglass, The Traverse Group, Inc., 2525 Aero Park Drive,
Traverse City, MI 49684
F.A. Blaha, U.S. Coast Guard, Shore Maintenance Detachment,
1240 East 9th Street, Cleveland, OH 44199
DW69933299
Project Officer
John T. Wilson
Robert S. Kerr Environmental Research Laboratory
P.O. Box 1198, Ada Oklahoma, 74820
A product of the Biosystems Research Program
This study was conducted in cooperation with
The United States Coast Guard,
Shore Maintainace Detachment, 9th District
and prepared for
The United States Air Force
Engineering Services Center,
Tyndall AFB, F1 32403
ROBERT S. KERR ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ADA, OKLAHOMA 74820
-------
DISCLAIMER
The information in this document has been funded wholly or in
part by the U.S. Environmental Protection Agency through the
Biosystems Technology Development Program (IAG DW69933299 from
RSKERL to the 9th District, U.S. Coast Guard), through the United
States Air Force (MIPR N-89-44 from HQ AFESC/RDXP, Tyndall AFB, FL
to RSKERL), and through an RSKERL in-house program
(Y105/B/02/01/3012) . It has been subjected to the Agency's peer
and administrative review, and it has been approved for publication
as an EPA document. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
ii
-------
Foreword
EPA is charged by Congress to protect the Nation's land, air and water
systems. Under a mandate of national environmental laws focused on air and
water quality, solid waste management and the control of toxic substances,
pesticides, noise and radiation, the Agency strives to formulate and implement
actions which lead to a compatible balance between human activities and the
ability of natural systems to support and nurture life.
The Robert S. Kerr Environmental Research Laboratory is the Agency's
center of expertise for investigation of the soil and subsurface environment.
Personnel at the Laboratory are responsible for management of research
programs to: (a) determine the fate, transport and transformation rates of
pollutants in the soil, the unsaturated zone and the saturated zones of the
subsurface environment; (b) define the processes to be used in characterizing
the soil and subsurface environment as a receptor of pollutants; (c) develop
techniques for predicting the effect of pollutants on ground water, soil and
indigenous organisms; and (d) define and demonstrate the applicability and
limitation of using natural processes, indigenous to the soil and subsurface
environment, for the protection of this resource.
There is little information available in the open literature on the
performance of bioremediation at field scale. This report documents the rate
and extent of treatment of a fuel spill in a drinking-water aquifer, using
nitrate as the primary electron acceptor for microbial respiration of the
contaminant hydrocarbons. Nitrate has theoretical advantages over the more
traditional electron acceptors used in the United States. It is much more
soluble than oxygen, and less costly and less toxic than hydrogen peroxide.
Clinton W. Hall
4"' 7^ /,¦ //-«(/
f » / . w (i- C- ^ (
Director
Robert S. Kerr Environmental
Research Laboratory
11 i
-------
EXECUTIVE SUMMARY
A field demonstration project on Nitrate-mediated biorestoration of a
fuel-contaminated aquifer was conducted at a U.S. Coast Guard facility in
Traverse City, Michigan. Several leaks from an underground storage'facility
containing JP-4 jet fuel have resulted in contamination of ground water at the
site. The focus of the field demonstration project is a 10 m x 10 m
infiltration area located within the larger area contaminated by the JP-4
spill. An infiltration gallery was installed above the study area; it is part
of a closed-loop system designed to perfuse the study area with ground water
supplemented with nitrate and nutrients. The 10 m x 10 m section of the site
was instrumented with monitoring wells and piezometers. A series of
recirculation wells was installed down gradient to intercept contaminants,
nutrients, and nitrate and provide hydraulic recirculation back through the
infiltration gallery. In addition, four purge wells are in place to provide a
net discharge from the site and prevent escape of nitrate or contaminants to
regional flow in the aquifer.
The design of the system was facilitated by hydraulic modeling to
evaluate the infiltration rate necessary to raise the piezometric surface
above the contaminated zone, the withdrawal rates necessary to retain the
contaminants and nutrients on-site, and the nutrient contact time important to
biological treatment. A tracer study was conducted to confirm estimated
breakthrough times and give a preliminary evaluation of the performance of the
in-situ bioreactor.
The effects of recirculation, purging to waste, and biodegradation on
the decrease in solution concentration of BTEX compounds within the treatment
zone were examined. The aquifer was cored and analyzed for total petroleum
hydrocarbons and for the quantity of selected fuel hydrocarbons. Water was
recirculated through the system for 41 days to bring the system to hydraulic
and chemical equilibrium. Then nitrate and mineral nutrients were added for
an additional 160 days. Biological processes supported on the ambient
concentrations of oxygen and nitrate removed Benzene from the fuel spill and
the recirculated water before nitrate was added. Benzene concentrations were
below 0.1 ug/1. After addition of nitrate, Toluene was rapidly removed in the
fuel spill, but not in the recirculated water. Ethvlbenzene and m+p-Xylene
were also removed during denitrification; however, there was little evidence
for biodegradation of o-Xylene until the end of the demonstration. As
expected, minor amounts of the alkane fraction were removed.
The technology produced excellent results. Concentrations of BTEX in
monitoring wells were below the Drinking Water Standard within 165 days. Unit
costs for the remediation were calculated by dividing the cost for
construction, labor, chemicals and electrical service by (1) the volume of JP-
4 beneath the infiltration gallery, (2) the volume of aquifer material
contaminated with JP-4 under the infiltration gallery, and (3) the volume of
aquifer between the infiltration gallery and the confining unit beneath the
aqjuifer. The unit costs for the remediation were $22 per li^er JP-4, $200 per
m of aquifer material contaminated with JP-4, and $17 per m of aquifer
materia] down to the confining layer.
-------
CONTENTS
DISCLAIMER ii
FOREWORD iii
EXECUTIVE SUMMARY iv
CONTENTS V
INTRODUCTION 1
SITE DESCRIPTION 2
FIELD DEMONSTRATION PROJECT 3
RESULTS 9
TRACER ANALYSIS 9
WATER QUALITY IN THE INFILTRATION GALLERY 19
WATER QUALITY IN THE RECIRCULATION SYSTEM 3 0
CORE ANALYSIS 3 3
DISCUSSION 47
QUALITY ASSURANCE AND QUALITY CONTROL 51
Extraction of Subcores 51
Fuel Carbon Analysis for JP-4 51
Mass Spectral Analysis of Core Extracts 5?
Analysis of Water Samples for BTEX 52
Dissolved Oxygen Analysis of Water Samples 52
Analysis of Water Samples for Inorganic Ions.... 53
COSTS 53
CONCLUSIONS 55
REFERENCES 57
v
-------
INTRODUCTION
Leaking underground storage tanks are a major source of ground-water
contamination by petroleum hydrocarbons. There are approximately 1.4 million
underground tanks storing gasoline in the U.S., and some petroleum experts
estimate that 75,000 to 100,000 of these tanks are leaking (Feliciano, 1984).
Gasoline and other fuels contain Benzene, Toluene, Ethylbenzene, and Xylenes
(BTEX) which are hazardous compounds regulated by the U.S. Environmental
Protection Agency. Although these aromatic hydrocarbons are relatively water-
soluble, they are contained in the immiscible bulk fuel phase which serves as
a slow-release mechanism for sustained ground-water contamination. Pump-and-
treat technology alone is economically impractical for renovating aquifers
contaminated with bulk fuel, because the dynamics of immiscible fluid flow
result in prohibitively long time periods for complete removal of the organic
phase (Wilson and Conrad, 1984; Bouchard £t al. , 1989). Biorestoration has
been recommended as a viable treatment alternative and involves enhancing the
activity of the native subsurface bacteria to degrade fuel hydrocarbons
through addition of nutrients and other compounds. Aerobic biorestqration has
been shown to be effective for many fuel spills (Thomas et al., 1987; Lee et
al., 1988). However, success is often limited by the inability to provide
sufficient oxygen to the contaminated intervals due to the low solubility of
oxygen (Wilson ejt al., 1986).
Nitrate can also serve as an electron acceptor in place of oxygen;
this results in anaerobic biodegradation of organic compounds via the
processes of nitrate reduction and denitrification (Tiedje, 1988). Because
nitrate is less expensive and more soluble than oxygen, it may be more
economical to restore fuel-contaminated aquifers using nitrate rather than
oxygen. Several investigators have observed biodegradation of aromatic fuel
hydrocarbons under denitrifying conditions (Kuhn et al., 1988; Major et al.,
1988; Mihelcic and Luthy, 1988; Lemon et al., 1989). However, the process is
still not well understood at field scale where several other processes,
including aerobic biodegradation, can proceed concomitantly. There have been
few definitive field studies ascertaining the extent and efficacy of nitrate-
mediated biorestoration of fuel-contaminated aquifers.
Our demonstration repeated and validated a remediation conducted in
the Upper Rhine Valley more than 10 years ago (Battermann, 1986). A spill of
fuel oil into a shallow water-table aquifer threatened municipal water wells.
An estimated 20 to 30 metric tons were releasee} to the subsurface. The
hydraulic conductivity of the aquifer (5 x 10 cm/s) was very close to the
conditions at Traverse City, Michigan. The depth to the water table was about
5 m, the saturated thickness was 4 to 5 m. Water table fluctuations spread
the fuel over an interval of 3 to 4 meters, beginning at the water. The area
at residual saturation was about 200 m long x 100 m wide, somewhat larger than
the spill at Traverse City. The ground-water temperature was 12 to 13 C,
almost the same temperature as the groundwater at Traverse City (11 to 12 °C).
Injected water required 20 to 30 days to move from injection to recirculation
wells, compared to 7 to 12 days in the demonstration at Traverse City. The
injected water was well oxygenated, and contained up to 113 mg/1 Nitrate-
nitrogen. About 45 mg/1 of Nitrate-nitrogen was removed in one pass.
Battermann also added small amounts of ammonia^and orthophosphate. The
remedial water was recirculated at about 408 m /day. Over a two-year period
1
-------
they added 22.6 metric tons of Nitrate-nitrogen. Benzene was completely
removed from the recirculation water in 8 months. The Xylenes were more
recalcitrant; but after two years, their concentrations were greatly reduced.
The European experience was met with skepticism in the United States;
and despite its promise, it is rarely used. To the authors' knowledge, there
has been only one report of a remediation of a fuel spill using Nitrate in
North America. Sheehan et: al. (1988) recirculated nitrate and mineral
nutrients through a spill of gasoline from a drainage pipe. The spill area
was apparently about 10 m x 10 m. The geology of the site is complex; the
spill was in deep soil overlying weathered and fractured bedrock, and
ground-water flow is along fracture and weathering zones of unknown
orientation. The ground water was recirculated at 3.8 to 6.5 m /day. It was
amended with 32 kg of soluble reactive phosphorus (sic), 108 kg of nitrate-
nitrogen and 39 kg of ammonia nitrogen. Total purgeable alkylbenzenes were
reduced from initial concentrations of 19 to 44 mg/1 to concentrations less
than 5 mg/1. In a second phase of remediation, total BTX was reduced to
0.9 mg/1.
SITE DESCRIPTION
In February of 1985, four large underground storage tanks in a fuel
farm at a U.S. Coast Guard facility in Traverse City, Michigan, were found to
be leaking JP-4 jet fuel from loosely connected piping into a shallow
underlying aquifer. By the time the leaks were discovered and the tanks and
contaminated soil were excavated, several thousand gallons of JP-4 jet fuel
had been lost to the subsurface. One monitoring well contained over 130 cm of
free product. A series of purge wells was installed to contain the
contaminated ground water. Product recovery pumps were installed in the purge
wells to recover the fuel overlying the water table (Figure 1). Although this
successfully retained the plume within the facility's boundaries, a large area
predominantly underlying a concrete apron and runway had been contaminated.
The aquifer is composed of thick glacial deposits with the upper
portions being lacustrine in origin (Twenter et al., 1985). These lacustrine
glacial deposits consist of an upper sand and gravel unit and an underlying
clay unit. The thickness of the sand ranges from 15 to 18 m in the immediate
vicinity of the plume and consists of fine- to medium-grained sand in the
upper 4.6 to 6.1 m and coarse to very coarse sajjd with gravel in the lower
section. The hydraulic conductivity is 4 x 10 cm/s; the seepage velocity in
the regional aquifer is 1.5 m/day. The water table varies seasonally and
ranges from 3.7 to 5.5 m below the land surface. This has effectively smeared
the contamination over a 1.5- to 2.5-m depth interval. There is JP-4 at
residual saturation in the unsaturated zone as well as free product at the
water table. Extensive subsurface coring was carried out to better define the
three-dimensional extent of contamination and to provide samples for chemical
analyses. Core samples were obtained using aseptic sampling techniques under
non-oxidizing conditions to preserve the microbial community structure and
chemical integrity of the cores (Leach et al., 1988).
2
-------
FIELD DEMONSTRATION PROJECT
As part of a settlement with the State of Michigan, the U.S. Coast
Guard had agreed to install and operate a purge well system that was down
gradient from the former tank farm. Their pre-existing line of purge wells,
PP-5 through PP-8 in Figure 1, had been installed in an unpaved area as close
as possible to the spill. The field demonstration was conducted in a 10 m x
10 m area located within the larger area contaminated by the JP-4 spill
(Figure 1). The study area was (1) adjacent to the former tank farm, (2)
unpaved, (3) in the flow path from the former tank farm to the purge veils,
and (4) in an area that was undisturbed by excavation and backfilling.
An infiltration gallery was used as part of a recirculation system
designed to immerse the study area in ground water supplemented with nitrate
and nutrients. Computer modeling was used to design the infiltration system.
The U.S. Geological Survey Modular Three-Dimensional Finite-Difference
Ground-Water Flow Model, MODFLOV (McDonald and Harbaugh, 1988), was used in
the simulations. A variable-spaced grid system of 70 rows by 70 columns was
designed with smaller grids in the center area of interest. Steady-state
conditions were simulated at various injection and withdrawal rates. Results
of the model indicated an infiltration rate of 1090 m /day was required to
create a water table mound encompassing the contaminated zone within the study
area. This would enable the contaminated unsaturated zone to be completely
saturated during the study and allow even distribution of nitrate and
nutrients. The mounding created by the injection of nitrate and nutrients
would be completely contained by the line of recirculation and purge wells so
that contaminants and nitrate would be retained on-site.
The infiltration gallery consists of a primary 10-cm PVC pipe equipped
with 5-cm PVC pipe arms. Each arm contains 144 1-cm holes spaced 2.5 cm apart
to ensure even hydraulic distribution. The gallery is contained within a 1-m
gravel pack buried 1.5 m below land surface to prevent freezing during winter
months. The bottom of the gravel pack is sloped downward towards the center
to concentrate the mound at the center of the study area. The study area
contains five well clusters to monitor ground water quality (Figures 1 and 2),
and 11 piezometers to monitor water levels and head loss across the
contaminated zone (not shown). Each well cluster was constructed of 6-mm
stainless steel tubing and contains six discrete wells with 40 x 40 mesh
stainless steel screens. The wells in each cluster well are spaced 60 cm
apart with interspersed bentonite seals to prevent channeling of ground-water
recharge. The wells in each cluster were located both within and beneath the
contaminated zone to allow monitoring of biodegradation and contaminant
transport within the different regions (Figure 4). The well clusters in the
infiltration area were arranged in a five- spot pattern (Figure 3). The wells
are numbered from 1 to 6, with 1 being the most shallow.
The lines for the well clusters were routed below land surface to a
sample building and terminated in separate stainless steel traps where water
was collected for chemical analysis. The sample building also contained the
equipment required for recirculating the ground water obtained by the pumping
wells, as well as the chemical addition facilitie^. Nitrate and nutrients are
batch mixed in separate polyethylene tanks (4.2 m each). The chemical feed
pumps are automatically controlled with flow rates adjusted weekly to
3
-------
Hangar / Administration
MW-21
Direction of
ground water flow
• Monitoring well
¦ Purge well
O Recirculation well - ground water
used for Infiltration recharge ^
PP-5A:
Sample
Building
9 meters
Contaminated Area
PP-7A
PP-8
PP-8A
PP-9A
!=~
• MW-34
Figure 1. Disposition of a JF-4 spill under the apron and taxi way of the .
Hanger Administration Building o:: the U.S. Coast Guard Air Station at Traverse
City Michigan.
-------
Hooaow hydraulic
surface
O 0-0
* aPPr„prJ
l0* be'«*«n «„
-------
Hangar ! Administration
/
N
9 Cluster well
¦ Purge well
O Recirculation well - ground water
used for infiltration recharge
9 meters.
PP-5
PP-6
PP-5 A
' PP-6A
Infiltration
feed line
Direction of
ground water flow
Sample
Building
PP-9A
. Infiltration
study area
Figure 3. Positions of the infiltration gallery, well clusters, recirculation
wells and purge wells.
6
-------
compensate for changes in nitrate and nutrient composition in the recirculated
water.
3
The purge wells pump at 82 m /day each; the effluents discharge to a
carbon treatment system. These wells are screened throughout the aquifer and
include the contaminated zone (Figure 3). In addition, five new wells. PP-5A
through PP-9A, were installed next to the purge well to recirculate-ground
water back to the infiltration gallery. Because they were installed next to
the existing purge wells, the recirculation wells did not shift the cone of
depression from the purge well field, which would abrogate the agreement with
the State of Michigan. Each of these wells is equipped with a 3-hp
submersible pump. They were screened below the contaminated zone to avoid
influx of free„product. The combined flows from the six recirculation wells
totaled 1090 m /day. They were monitored and amended with nitrate and
nutrients as required.
Performance of the pilot project was assessed by monitoring BTEX and
inorganic chemicals throughout the system of wells and by extensive coring of
the contaminated zone in the study area. Core samples have been exajnined for
total fuel content and BTEX concentrations. Coring is required to verify
monitoring well data because the slow release of BTEX from residual saturation
in the contaminated material might lead to an overestimate of field
performance based of monitoring well data alone. Core samples were taken from
two locations within the study area in discrete 10~cm segments throughout the
vertical extent of the contaminated zone. These analyses provided background
information on fuel distribution prior to infiltration (Summarized in
Figure 4).
Infiltration, using recirculated ground water without amendments, was
begun and continued for six weeks until hydraulic and chemical equilibrium
were attained. During this time, and throughout the remainder of the project,
water samples were obtained from the cluster wells on a weekly basis and
analyzed for dissolved oxygen and BTEX to ascertain movement of contaminants
through the system. Once hydraulic equilibrium was attained, the study area
was cored again. This was done to evaluate the effects of infiltration alone
on BTEX removal from the contaminated zone without the influence of nitrate or
nutrient amendments. Nitrate and nutrients were then applied at design
concentrations of 62 mg/1 sodium nitrate (10 mg/1 as N), 10 mg/1 monobasic
potassium phosphate, 10 mg/1 disodium phosphate, and 20 mg/1 ammonium chloride
for 109 days, at which time a third coring was done.
A tracer study was conducted to define the horizontal and vertical
flow characteristics of the aquifer, the contact time between injected
amendments and the contaminated materials beneath the infiltration gallery,
the travel time of water to the recirculation wells, and the volume of water
in the recirculation loop.
7
-------
Recirculation wells
(1090 m3/day)
Purge wells
(330 m3/day to carbon)
Elev
(meters)
188-r
184
CE CD'
CC CA
CF
CG
Contaminated
interval
Water Table
176 —
CH
PP
PPA
172
Clay Confining Layer
20
10
30
40
0
Trar sect (meters)
»
Figure 4. Cross section shoving elevations of water table, JP-4 contaminated
interval, the: confining layer, infiltration gallery, sampling points in well
clusters, recirculation wells, and purge veils.
-------
RESULTS
TRACER ANALYSIS
The area under the infiltration gallery and along the flow path to the
recirculation wells was instrumented with a series of multilevel sampling
wells. The wells produce point samples of the groundwater from the three-
dimensional space of the aquifer. Well clusters CA through CD were installed
under the infiltration gallery. Well clusters CI, CF, CG, and CH were
installed in a line between the infiltration gallery and the the recirculation
wells. Figure 5 models the the flow paths for water between the injection
gallery, the regional flow of the aquifer, and the pairs of purge and
recirculation wells. Clusters CI, CF, CG, and CH sampled water along the
shortest flow path from the injection gallery to the recirculation wells
(Compare Figures 5 and 3).
The state of Michigan gave permission to the Coast Guard to use
ammonium chloride as a nutrient. Because background concentrations in the
aquifer are low, the chloride made a convenient tracer. To start a tracer
test, a pulse of ammonium chloride was added the injection water for 8 hours
at a concentration of 230 to 260 mg/1 chloride. The pulse was then followed
vertically through the injection gallery and contaminated zone, then
horizontally to the pumping wells. Continuously-operating peristaltic pumps
were connected to the three most shallow wells in two of the clusters.
Samples were taken every 15 minutes for chloride analysis. After the chloride
concentration peaked in a particular well, the pump was transferred to the
next available lower well. Three separate chloride pulses were required to
test the vertical conductivity in the clusters within the infiltration
gallery. The first pulse was followed down gradient past clusters CI, CF, CG,
and CH and into the recirculation wells. Sampling times were adjusted to
provide approximately ten points on the rising limb of each breakthrough
curve.
If flow is vertical, chloride should arrive at the shallow veils first
because the flow path is shorter. If flow is horizontal, chloride should
arrive at all the wells at the same time. Chloride breakthrough of four of
the five well clusters beneath the infiltration gallery are presented in
Figures 6 and 7 (See Figure 3 for the position of the well clusters).
Vertical infiltration was very uniform; the arrival time of the pulse was
uniformly related to the depth of the well. Breakthrough of chloride was
sharp, with little evidence of dispersion or channelling. The average time
required for the chloride pulse (C/Co = 0.5) to move past the six wells was
11.5 hr. Chloride breakthrough averaged vertically through the the
infiltration gallery in four well clusters with six wells indicating a
vertical pore velocity of 0.01 cm/sec. The fluid contact time in the segment
from the bottom of the infiltration gallery to the bottom of the contaminated
interval was 8.3 hr.
Analysis of head differences in a nest of piezometers positioned at
sucessive 60 cm intervals through the injection gallery and contaminated
interval confirmed saturation through the contaminated interval and a unit
9
-------
Hanger Administration
Building
¥//////////
Figure 5. Hydraulic model shoving capture zones of the pumping veils. The
model combined the influence of adjacent purge veils (PP-) and recirculation
veils (PP-A).
-------
CLUSTER HELL CA
240
220
200 -
180 -
160 -
01
E
140 -
120 -
g
£
CA—1
CA-2
CA-3
CA-4
CA-5
CA-6
00-
2.5
5.0
0.0
7.5
10.0
12.5
15.0
17.5
TIKE thrs)
CLUSTER WELL CB
220 -
200 -
180 -
183 -
-j
140 -
LLJ
o
»—1
cc
CD
120 -
CB-1
CB-2
CB-3
C8-4
CB-5
CB-6
5:
u
'«•—-
100 -
60-
0.000.00 2.50
5.00
10.00
7.50
12.50
TIKE (hrs!
Figure 6. Breakthrough of a chloride tracer in well clusters CA and CB. See
Figure 3 for the location of the clusters. See Figure 4 for the depth of the
wells. The wells are numbered from shallow to deep. The depth between wells
is 60 cm.
11
-------
CLUSTER HELL CC
200 -
180
160-
5
o?
JE
140 •
Qj
a
•—<
120-
i
100 -
280 -
260
240
2?0
200
160
160
140
» H
or
120
u
100
eo-
60-
40-
20-
10.0 12.5 15.0
TIKE (hrsj
CLUSTER WELL CE
25.0
TIME Ihrs)
Figure 7. Breakthrough of a chloride tracer in well clusters CC and CE. See
Figure 3 for the location of the clusters. See Figure 4 for the depth of the
wells. The wells are numbered from shallow to deep. The depth between wells
is 90 cm.
12
-------
hydraulic gradient down to the level of the regional water table. A unit
gradient means the water table declines 60 cm with each 60 cm increase in
depth of the piezometer. Under a unit gradient, flow is entirely vertical.
Well cluster CI is located 1.8 m downgradient of the injection
gallery. The nine breakthrough curves at cluster CI (Figure 8) are evenly
spaced and indicate a vertical component of flow of 0.022 cm/sec. The
chloride pulse reached all the way to the level 9 at the bottom of the
aquifer. The strong vertical flow reflects the influence of the local
piezometric surface as well as the the influence of the increased head of the
infiltration gallery. Because the vertical velocity is uniform between wells
in the cluster and because the flow is less than the vertical flow induced by
a unit hydraulic gradient under the infiltration gallery, there is no evidence
of geohydraulic anomalies such as clay lenses that might produce significant
changes in the hydraulic conductivity.
A slight downward hydraulic gradient is still evident at cluster CF,
9.1 m from the infiltration gallery (Figure 9). The level 1 well 4.3 m below
the land surface was dry, and levels 2 and 3 appeared to be stagnant. There
was no chloride breakthrough. At cluster CG, 15.2 m from the infiltration
gallery, the flow is mostly horizontal (Figure 10). The pulse arrived at the
deeper wells first, indicating that the hydraulic conductivity at level 6, 7.3
m below land surface, is slightly greater than the more shallow layers.
Sharp breakthrough of the chloride tracer was even observed at well
cluster CH (Figure 11). Veil cluster CH was located 30.5 m down gradient of
the infiltration gallery, 1.8 m from the purge well PP-7, and 2.4 m from the
recirculation well PP-7A. The arrival of the tracer seems to have been
influenced by the position of the pumps in the wells. The pump in PP-7 was
located 9.1 m below land surface, and the pump in PP-7A was located 14.9 m
below land surface near the clay aquitard (the pumps are represented by black
rectangles in Figure 4). The tracer reached level 4 first, which was screened
at 8.8 m below land surface, followed by level 5, then 3, then 6. The tracer
then broke through at level 9, close to the depth of the pump in PP-7A. The
last well in cluster CH to break through was level 7, intermediate in depth
between the pumps.
Arrival of the chloride tracer at the pumping wells is presented in
Figure 12. The flow model depicted in Figure 5 suggests that PP-5 and PP-5A
should recruit water from the regional flow of the aquifer, PP-6 and PP-6A
should recruit water from both the regional flow and the infiltration gallery,
and the other wells should largely recruit water from the the infiltration
gallery. There is little evidence of chloride breakthrough in PP-5 and PP-5A,
weak breakthrough in PP-6 and PP-6A, and strong breakthrough in PP-7, PP-7A,
PP 8, and PP-8A. Breakthrough in PP-9A was weaker than might be expected from
the model.
Chloride broke through at the injected concentrations in the passive
monitoring wells in cluster CH, just in front of the pumped wells PP-7 and PP-
7A. Chloride broke through in the pumped wells as a peak, instead of a step
increase. The peak concentration would correspond to the midpoint of the
breakthrough of a step increase. The peak concentration in PP-8A was 45£ of
the injected concentration; and the peak concentration in PP-8 was 55%,
indicating good recovery of the tracer.
13
-------
CLUSTER WELL CI
o>
K
a
»—<
re
CD
rn
CJ
>
50 60 7C
TIME tors)
CM(
CI-2i
CI-3 j
CI-41
CI-5 {
CI-6!
CI-7S
CI—8 j
ci-gj
— I-' 1
110 120
Figure 8. Breakthrough of a chloride tracer in veil cluster CI. See Figure 3
for the location of the cluster. See Figure 4 for the depth of the veils.
The veils are numbered from shallov to deep. The depth betveen veils is
180 cm.
14
-------
CLUSTER HELL CF
CTJ
F=
UJ
CD
»H
cc
/V
HF-4
Cr-5
CF-6
20 0 22.5 25.0 27.5 30.0 32.5 35.0
TIME (hrs)
37.5
40.0
Figure 9. Breakthrough of a chloride tracer in well cluster CF. See Figure 3
for the location of the cluster. See Figure A for the depth of the veils.
The veils are numbered from shallov to deep. The depth betveen veils is
90 cm.
15
-------
CLUSTER HELL CG
O)
E
I i_I
a
i—i
nc
o
nr
C-3
260 j
240 -
220 -
200 -j
180 -j
J
150 -j
140 J
120
J
100 -i
4
BO -!
60
40 H
20
j/j
W —-+
, r—j~ . —r —-T --r 1—I r- T"'~T ' I r ~r" T"T"
15 20 25 30 35 40 45 50 55 60 65
TIME Ihrs)
CG—4 ;
cg-5 ;
C&-6 |
—r—* j
70 75
Figure 10. Breakthrough of a chloride tracer in veil cluster CG. See Figure 3
for the location of the cluster. See Figure A for the depth of the wells.
The wells are numbered from shallow to deep. The depth between wells is
90 cm.
16
-------
CLUSTER WELL CH
01
UJ
cc
CD
U
H
155 175 195
TIME (hrs)
-i
235
Figure 11. Breakthrough of a chloride tracer in well cluster CH. See Figure 3
for the location of the cluster. See Figure A for the depth of the wells.
The wells are numbered from shallow to deep. The depth between wells is
180 cm.
17
-------
C71
RECIRCULATION
PP-6A •
PP-7A i
PP-flA i
PP-SA i
-j 1 , . , -r—, . , > | r—1 ¦"
100 120 140 160 180 200 220 240 260 290 300
THE Chrs]
PURGE
I
Ol
£
o
if:
u
/
*W~\
~l—•—J—•—I *»
-** i - '-*-1—•—i—¦—i—•—j—•—r~"•—i—¦—r
100 120 140 1G0 100 200 220 240 2G0
TIKE Ihrs)
PP-5
PM !
PP-7 I
PP-8 j
200 ^300
Figure 12. Breakthrough of a chloride tracer in recirculation and purge veils
In?. A ?' ' locflon ot the See Figure 4 £„r the depth
interval that is screened.
18
-------
The tracer test conducted between the infiltration gallery and the
recirculation veils indicated a breakthrough of chloride after 7 to 12 days,
depending on the distance to each well and its pumping rate. A detailed
analysis of the arrival times of the chloride front along the line of cluster
wells was conducted to refine the estimate of travel times for the pumping
wells. The volume of water recirculated within the remediation zone was
estimated as the flow-weighted average: the pumping rate of each well was
multiplied by the time required for breakthrough of chloride, and the sum of
all the wells was divided by the total rate of pumping. ^The flow-weighted
average was 239 hr, or approximately 10 days. At 1090 m /day, this
corresponds to 10,900 m of water recirculated through the treatment zone.
The tracer data was cross-checked by measuring in Figure 5 the surface
area of the region where flow lines from the gallery converge on the wells,
multiplying by the saturated thickness of the aquifer to estimate the total
volume of the aquifer in the treatment zone, and correcting for water filled
pgrosity (Figure 5). The surface area of the predicted flow field is 6,000
m , the saturated thickness is 11 m (Figure 4), and the effective porosity is
0.15 (Twenter et al., 1985), giving a second estimate of the volume of
recirculated water at 9,900 m .
WATER QUALITY IN THE INFILTRATION GALLERY
Samples were collected weekly for the well clusters under the
infiltation gallery. The following discussion will focus on water quality at
level 2 (within the contaminated zone, above the original water table), level
4 (just beneath the contaminated zone), and level 6 (deeper below the
uncontaminated zone). The relationship between the wells and the fuel
contaminated interval is presented graphically in Figure 13. Data were
averaged for each of the five wells at a given level.
The dissolved oxygen concentrations in the infiltration water and the
wells are presented in Figure 14. Infiltration began on Day 3 and nitrate and
nutrient addition began on Day 44. Samples were collected from the well
points at Day 0 to provide background information. Because the level 2 wells
were in the unsaturated zone, data from level 3 were substituted at Day 0.
Oxygen concentrations were initially high beneath the contaminated
zone; they were near 3 mg/1 all the way to the bottom of the aquifer.
However, oxygen concentrations dropped rapidly as water recirculated through
the c^ntaminated^zone, and stabilized at 0.4 to 1.0 mg/1 after Day 7. If
330 m of 1090 m recirculated through the pipeline to the infiltration
gallery is oxygenated water recruited from the regional flow of the aquifer,
then the expected oxygen concentration in the blended water would be 0.9 mg/1.
Nitrate was added to the recirculated water 41 days after start-up.
Nitrite began to accumulate immediately upon the addition of nitrate (Figure
15 and 16). Around 60 days after start-up, nitrate removals during transit of
the water through the fuel-contaminated interval increased abruptly
(Figure 15), and nitrate concentrations declined to 0.1 to 0.5 mg/1 as N
(Figure 16). Comparison of the concentration of nitrate in the injected water
to the concentration that survived transit through the contaminated interval
19
-------
C - Stainless steel cluster wells
Elev
(meters)
188 -
187 -
186
185
184
183
182
181
180
CE CD
CC
CA CB
rnf*r»r *r*p i >
***********
(i ('
illtllflll
Core
Interval
I* :;i;t
j#©Wm
4 2
¦
4
5
6
Cluster
well lines
Infiltration
feed line
Contaminated
zone
Figure 13. Relationship between the infiltration gallery, the monitoring well
clusters, and thu fuel-contaminated interval in the aquifer.
-------
Dissolved Oxygen
6.0
Injection Water
Level 2
Level 4
Level 6
o> 4.0"
o
Begin Nitrate
Injection
CO
LSD
CD
£ 2.0
o
O
1.0-
0.0
40
80
0
20
60
100
120
Time (days)
Figure 14. Concentration of dissolved oxygen in the recirculation water, and
at various levels in the well clusters. See Figure 4 for the depth of the
wells. The wells are numbered from shallow to deep. Level 2 is in the JP-4
contaminated interval, level 4 is just under the contaminated interval and
level 6 is deeper under the contaminated interval. LSD is the least
significant difference between means at the 90X confidence level.
21
-------
14
12
— 10
OJ
E
-
o
£ 6
c
a>
o
§ 4
O
Nitrate-Nitrogen
—a-- injection Water
—o— Level 2
LSD
r \
/ \
_._.n4r._. Level 4
/ \
/ t.
+ Level 6
\
N
i \
' \
/
V ' A
V / \
Begin Nitrate
if \
/ \
Injection*
13 \\\
13 V-
/ D
m
i/ \\b^
if V.
1/ v
iij t
jB
2-
0 20 40 60 80
Time (days)
100
120
Figure 15. Concentration of nitrate-nitrogen in the recirculation water and
at various levels in the veil clusters. See Figure A for the depth of the
wells. The wells are numbered from shallow to deep. Level 2 is in the JP-4
contaminated interval, level 4 is just under the contaminated interval and
level 6 is deeper under the contaminated interval. LSD is the least
significant difference between means at the 95£ confidence level.
22
-------
Nitrite-Nitrogen
3.0
Injection Water
Level 2
Level 4
Level 6
2.5-
I5 2.0"
o
LSD
(U
Begin Nitrate
Injection v
cz
a>
o
cz
o
O
0.5-
80
0
20
40
60
100
120
Time (days)
Figure 16. Concentration of nitrite-nitrogen in the recirculation water, and
at various levels in the well clusters. See Figure 4 for the depth of the
wells. The wells are numbered from shallow to deep. Level 2 is in the JP-4
contaminated interval, level 4 is just under the contaminated interval and
ievel 6 is deeper under the contaminated interval. LSD is the least
significant difference between means at the 95% confidence level.
23
-------
(Figure 15) indicates that 6 to 7 mg/1 of nitrate-nitrogen was removed during
passage through the contaminated interval, and that this level of removal
continued over a 50 day period.
There was essentially no nitrate removal between levels 4 and 6, that
is, after the infiltration water had passed beyond the fuel contaminated zone.
This is not surprising; after complete acclimation to nitrate on Day 70, the
total BTEX concentration at level 4 never exceeded 0.3 mg/1, which would have
a theoretical nitrate-nitrogen demand of only 0.3 mg/1.
Ammonia was added to the infiltration water as a nutrient. There was
little evidence of ammonia removal during passage across the fuel contaminated
interval, and little evidence of ammonia accumulation from nitrate reduction
(Figure 17). If ammonia was the end product of anaerobic nitrate metabolism
in the infiltration gallery, the concentrations of ammonia should have
doubled.
Benzene removal from the infiltration water was rapid and complete,
even prior to nitrate addition (Figure 18). By Day 14, Benzene concentrations
were typically less than 1 ug/1 throughout the contaminated zone. Significant
concentrations of Benzene were present in the recirculated infiltrataion water
up to Day 35, but removal of Benzene occurred before the water reached wells
at level 2. Removal of Benzene in the recirculation water is not related to
the addition of nitrate on Day 44. The concentration of Benzene in
recirculation well PP-7A was 161 ug/1 on Day 34; on Day 36 it was less than
0.1 ug/1 (extracted from data presented in Figure 23). Benzene disappeared in
recirculation well PP-8A by Day 21 and in PP-9A by Day 35 (data not shown).
Toluene removal was more complex (Figure 19). By Day 21, the system
had equilibrated, and Toluene concentrations were similar throughout. From
Day 21 to Day 50, about 90% of the Toluene in the infiltration water was
removed by the time the water reached level 2. After Day 50, after nitrate
addition, Toluene concentrations dropped below 1 ug/1 even though much higher
concentrations were delivered in the infiltration water.
Ethylbenzene and m+p-Xylene concentrations followed a very complex
pattern. Data is shown for m+p-Xylene (Figure 20). These Xylenes were
removed between the injection water and level 2, but concentrations actually
increased as the infiltration water passed deeper into the fuel contaminated
interval. The effect may be caused by leaching of the Xylenes from
unweathered material deeper in the uncontaminated zone. There is no evidence
of removal of Ethylbenzene or m+p-Xylenes before the addition of Nitrate.
After addition of Nitrate, about 90% was removed during passage from the
infiltration gallery to level 2.
The profile of o-Xylene was unique (Figure 21). There was no removal
of this compound during passage through the contaminated interval. There was
a gradual decline in concentrations due to removal from the recirculation cell
by the purge wells. This effect will be described in detail in the next
section.
24
-------
14
12 i
g> 10
Ammonia-Nitrogen
—
—i—
8"
Injection Water
Level 2
Level 4
Level 6
LSD I
Begin Nitrate
Injection
T
60
Time (days)
100 120
Figure 17. Concentration of ammonia-nitrogen in the recirculation water, and
at various levels in the veil clusters. See Figure A for the depth of the
wells. The wells are numbered from shallow to deep. Level 2 is in the JP-4
contaminated interval, level A is just under the contaminated interval and
level 6 is deeper under the contaminated interval. LSD is the least
significant difference between means at the 95X confidence level.
25
-------
Benzene
10
—¦— Injection Water
—¦— Level 2
A-— Level 4
•— Level 6
CD
LSD
tz
01 -
Begin Nitrate
/ Injection
.001
40
60
o
20
100
120
Time (days)
Figure 18. Concentration of Benzene in the recirculation water, and at
iileV6ls in,the wfU clusters. See Figure 4 for the depth of the wells,
ine veils are numbered from shallow to deep. Level 2 is in the JP-4
contaminated interval, level 4 is just under the contaminated interval and
level 6 is deeper under the contaminated interval. LSD is the least
significant difference between means at the 95* confidence level, appropriate
only for the first 40 days.
26
-------
Toluene
Injection Water
Level 2
Level 4
Level 6
Begin Nitrate
Injection
o .01
40 60 80
Time (days)
Figure 19. Concentration of Toluene in the recirculation water, and at
various levels in the well clusters. See Figure 4 for the depth of the wells.
The veils are numbered from shallow to deep. Level 2 is in the JP-4
contaminated interval, level 4 is just under the contaminated interval and
level 6 is deeper under the contaminated interval. LSD is the least
significant difference between means at the 952 confidence level, appropriate
only to the first 55 days.
27
-------
O)
J=
c:
o
2
ci
CD
O
c:
o
O
.001
m,p-Xylene
Begin Nitrate
Injection
Injection Water
Level 2
Level 4
Level 6
40 60 80
Time (days)
Figure 20. Concentration of m+p-Xylene in the recirculation water, and at
various levels in the well clusters. See Figure 4 for the depth of the wells
ihe wells are numbered from shallow to deep. Level 2 is in the JP-4
contaminated interval, level 4 is just under the contaminated interval and
level 6 is deeper under the contaminated interval. LSD is the least
significant difference between means at the 95X confidence level.
28
-------
o-Xylene
Begin Nitrate
Injection
Injection Water
Level 2
Level 4
Level 6
CD
a
LSD
.01 -
.001
40
60
80
1 00
20
0
1 20
Time (days)
Figure 21. Concentration of o-Xylene in the recirculation water, and at
various levels in the well clusters. See Figure A for the depth of the wells.
The wells are numbered from shallow to deep. Level 2 is in the JP-4
contaminated interval, level 4 is just under the contaminated interval and
level 6 is deeper under the contaminated interval. LSD is the least
significant difference between means at the 95X confidence level.
29
-------
WATER QUALITY IN THE RECIRCULATION SYSTEM.
The effect of dilution of the BETX compounds in the recirculated water
was evaluated by estimating the partitioning of the individual BETX compounds
between the fuel and water in the treatment zone. The initial concentrations
of BTEX compounds expected in the recirculated groundwater were estimated from
the core data and literature partitioning information (Smith et al., 1981),
assuming equilibrium conditions in the recirculation zone.
The level of water in the piezometric wells indicated that the
infiltration gallery effectively saturated an area of 129 m , several meters
beyond the boundaries of the 10 m x 10 m area beneath the arms of the
infiltration gallery. As will be discussed in the next section, the
contaminated interval is 2.14 ^ deep (Table 2), giving a volume of JP-4
contaminated material of 276 m . The bulk density of the aquifer material is
1530 kg/m ; multiplying by the volume gives a mass of JP-4 contaminated
aquifer material of 420,000 kg. Two boreholes were cored continuously
through the contaminated interval before remediation began. JP-4 in- borehole
51AF averaged 7,100 mg/kg dry wt; borehole 51AG averaged 3,100 mg/kg. The
grand average of the two is 5,100 mg/kg averaged over a 2.14 m interval.
Multiplying the average concentration in the aquifer material by the mass of
contaminated aquifer material gives an estimate of the amount of JP-4 in the
contaminated interval under the infiltration gallery^of 2,140 kg. Dividing by
the specific gravity of JP-4 gives a volume of 2.7 m .
The volume of recirculated water was estimated in the previous section
to be 10,900 m , and , wate is available from the literature. At
equilibrium, the distriButlon of each BTEX compound between the fuel and the
recirculated water can be estimated as:
Volume JP-4 mass in JP-4
K,,. = K, . ^ X Volume water = mass in water
~(compound) fuel-water
£
The fraction of each compound in the fuel = d(compound) —_
a(compound)
The fraction in the water = 1 - the fraction in the fuel
The numerical values of these calculations are summarized in Table 1.
The values of from Smith e£ al. (1981) were confirmed for these conditions
by comparing the solution concentrations predicted for the BTEX compounds
based on the BTEX content of the fuel prior to start up, to the actual
solution concentrations measured in the fuel contaminated interval prior to
hydraulic equilibrium. They were found to compare very favorably.
The equilibrium concentration of each compound in the recirculation
water was predicted by multiplying the concentration of the compound in the
contaminated interval (as determined from core data) by the total mass of
contaminated aquifer material (420,000 kg) to get the mass of each compound,
-------
3
then dividing by the volume of water in the treatment cell (10,900 m ), then
multiplying by its estimated fraction in the water (from Table 1). The
results are summarized in Table 1.
Table 1. PARTITIONING VALUES FOR THREE BTEX COMPOUNDS IN THE RECIRCULATION
ZONE.
Parameter Benzene Toluene o-Xylene
K. . ^ 2454 2754 7079
K,u vater 0.607 0.68 1.75
Fraction in fuel 0.38 0.40 0.64
Fraction in water 0.62 0.60 0.36
Cone, in fuel originally ,
(mg/kg dry wt.) 11 a 38 29
Mass in fuel originally
(mg) 3,600 16,000 12,000
Concentration in water
at equilibrium
(mg/m = ug/1) 262 900 400
Rate of dilution by purge wells
(per day) 0.019 0.018 0.011
^Borehole 51AZ, near cluster CG, as discussed in Hutchins et al.(1989).
Weighted average of boreholes 51AF and 51AG, Table 1.
After it comes to hydraulic and chemical equilibrium, the
recirculation zone^can be yodelled as a completely mixed reactor. The reactor
volume is 10,900 m , 330 m /day is discharged by the purge wells, and 330 m
of clean water must be recruited from the regional flow of the aquifer. The
rate of dilution of the water in the recirculation zone due to the purge wells
is equal to the discharge divided by the volume, or 0.03/day. The rate ul
dilution of a BTEX compound would equal the rate of dilution of water
(0.03/day) multiplied by the fraction of that compound that is in the water
(From Table 1). Results are presented in Table 1.
This analysis assumes that the rate of recirculation was constant.
Figure 22 plots recirculation against days of operation. Except for a brief
interval during start-up, and two periods when the wells were shut down to
allow coring through the infiltration gallery, the assumption is reasonable.
Dilution in a completely mixed reactor follows first order kinetics.
The decrease in aqueous concentration of BTEX compounds is expected to be
related to the initial concentration of each compound by the relationship:
C
o
31
-------
1600-
1400-
1200-
1000-
3 800-
400-
200-
20
40
60
80
140
Time (days)
Figure 22. Rate of recirculation of water to the infiltration gallery.
-------
where C = the solution concentration at time (t),
C = the calculated equilibrium concentration in the
recirculated water
k = the rate of dilution for each compound due to the
clean water recruited by the purge wells, and
t = time
Plots of the predicted aqueous concentrations over a 200-day period
incorporating the predicted initial concentrations and the effects of dilution
by the purge wells are shown in Figures 23-25. The results of the daily
monitoring of Benzene and Toluene are shown in Figures 23 and 24. The
concentrations of these compounds at recirculation veil PP-7A were seen to
rise as hydraulic injection mobilized the dissolved fraction. Chemical
equilibrium appears to have been reached after about 21 or 22 days. Benzene
disappeared in well PP-7A about Day 35, and stayed below 0.1 ug/1 for the
remainder of the demonstration. Toluene concentrations were reduced only
about an order of magnitude after Day 30.
This difference in behavior is difficult to rationalize. There is no
a priori reason for Benzene and Toluene to behave in such radically "different
ways. Some of the flow paths through the contaminated interval under the
infiltration gallery are skewed and pass through a much greater pathlength of
JP-4. These flow lines would tend to leach alkylbenzenes and support higher
concentrations of the compounds if there is no further degradation of the
alkylbenzene after the water left the fuel-contaminated interval. The
simplest explanation is that Benzene degraded in the aquifer below the fuel
contaminated interval, and Toluene did not. This is contrary to the
experience of Lemon et al. (1989). When they injected a plume of nitrate,
Benzene, Toluene, EtTiylTien^ene, and Xylenes into a plume of anaerobic landfill
leachate at Canadian Forces Borden, Borden, Ontario; Toluene disappeared after
60 days, but the other compounds persisted up to 200 days.
Contrary to the experience under the infiltration gallery, the
disappearance of Ethylbenzene and the Xylenes in the recirculation water
followed the removal expected from dilution. Figure 25 presents data for o-
Xylene. Only toward the end of the demonstration did removal exceed that
expected from the physical removal processes.
CORE ANALYSES
Cores were obtained through a hollow stem auger with a sliding-piston
core barrel, using the protocol of Leach el al., 1989. The location of the
boreholes is presented in Figure 26. Core samples were obtained from the
study area 15 days prior to hydraulic loading (51AF and 51AG), 33 days after
start-up but prior to nitrate and nutrient addition (51AH and 51AI), and after
109 days of operation (51AJ and 51AK). Both duplicate boreholes at each
sampling event were continuously cored from 182.18 to 185.17 m above mean sea
level for each sampling event. Each core was 45 to 90 cm long and provided
several 7.5 to 10-cm subcores. These subcores were collected anaerobically
into sterile 250-ml Mason jars. Sterile 10-ml tuberculin syringes, with the
tips removed, were then used to obtain duplicate mini-cores from each jar;
this provided replicate composites for chemical analyses. The mini-cores were
33
-------
Daily Benzene Levels
4.0
3.5-
Start Nitrate
3.0-
_i
Stop
Nitrate
2.5-
d
€Z
2.0-
o
O
c:
CD
M
1.5-
£Z
CD
m
0.5"
o
O)
o
0.0 i
Predicted
PP7A
0 20 40 60 80 100 120 140 160 180 200
Days Since Startup
Figure 23. Concentration of Benzene in recirculation well PP-7A compared to
water!" d concentration based on dilution and wasting of the recirculati
34
-------
Daily Toluene Levels
4.0
3.5-
Start Nitrate
3.0-
_i
Stop
Nitrate
2.5-
o
cr
2.0"
o
O
CD
Q)
JD
O
o
0.5-
CD
O
_l
0.0
Predicted
PP7A
-0.5"
20 40 60 80 100 120 140 160 180 200
0
Days Since Startup
Figure 24. Concentration of Toluene in recirculation well PP-7A compared to
the predicted concentration based on dilution and wasting of the recirculation
water•
35
-------
4.U-
3.5-
3.0"
_l
O)
3.
2.5-
d
tz
o
2.0-
O
-
CD
tz
1.5-
¦
X
1.0-
o
"o
0.5-
CD
o
_J
o
o
-0.5-
-1.0-1
0
Daily o-Xylene Levels
Start Nitrate
Stop
Nitrate
Predicted
-~— PP7A
20 40 60 80 100 120 140 160 180 200
Days Since Startup
Figure 25. Concentration of o-Xylene in recirculation veil PP-7A compared to
the predicted concentration based on dilution and wasting of the recirculation
water.
36
-------
01 23456789
meters I—I—I—I—I—I—I—I—I—I
INFILTRATION
STUDY AREA
51A
•
51/v
J •
F
51A
•
[
•
51£
H •
51A
•
51
G
K
CHEMICAL FEED LINE.
Figure 26. Location of boreholes used to collect continuous cores through the
contaminated interval under the infiltration gallery before flooding, after
flooding but before nitrate addition, and after nitrate addition.
37
-------
transferred to sterile, nitrogen-filled VOA bottles inside the glovebox and
sealed with Teflon-lined silicone septa. These samples were analyzed for fuel
carbon and JP-4 content according to the method developed by Vandegrift and
Kampbell (1988). In addition, the extracts were analyzed by GC/MS to quantify
the alkylbenzenes and to provide semi-quantitative information on the major
classes of hydrocarbons present in the cores.
The distribution of JP-4 within the contaminated zone at various times
during the project is shown in Figure 27. The contaminated zone covers a
depth of about 2.1 m; concentrations vary considerably within narrow depth
intervals. In the background core, the highest concentrations of JP-4 are
located at the water table. After 1 month of flooding to achieve hydraulic
equilibrium, there was still a significant amount of fuel in the subsurface
even though mobilization and partitioning appears to have shifted the profile
downward slightly (Figure 27). After an additional 2 months of flooding using
nitrate and nutrients, concentrations of JP-4 were further reduced.
The results of each set of replicate cores were extrapolated over the
entire study area, and the total mass of JP-4 was calculated for the^
contaminated zone outlined by the study area boundaries. The mass of JP-4 was
reduced from a background of 2200 + 1300 kg to 2000 + 240 kg after 1 month of
flooding, and then to 1400 + 330 kg after 2 months of nitrate addition (means
and standard deviations). This was expected; the technology was designed to
remove the BTEX fraction of the JP-4.
Table 2. DEPLETION OF ALKYLBENZENES AND TOTAL PETROLEUM HYDROCARBONS IN
CORE MATERIAL FROM UNDERNEATH THE INFILTRATION GALLERY.
Borehole
Benzene Toluene Ethylbenzene m+p-Xylene o-Xylene
mg/kg dry weight*
JP-4
4/6/89 prior to
the demonstration
51AF 1.3 48 26
51AG 0.27 16 9.9
5/24/89, 34 days
after start-up
51AH 0.026 0.15 0.41
51AI 0.024 0.14 0.38
8/9/89, 110 days
after start-up,
69 days after
Nitrate addition
51AJ 0.012 0.034 0.019
51AK 0.020 0.036 0.018
80
32
7.2
6.6
35
15
3.0
2.8
0.059
0.055
0.13
0.40
7,100
3,100
4,600
4,000
2,700
3,700
* Weighted averages of at least 15 determinations on contiguous cores
from a 2.1 m interval that spanned the contaminated interval in the
aquifer. Alkylbenzenes were determined by GC/MS. Total JP-4 was
deteremined by GC.
38
-------
Background
Core
(Core 51AF)
18
184
183-S
182
20000
0
10000
15000
25000
5000
30000
184
182
After One Month Flooding,
Before Nitrate Addition
(Core 51AI)
-+-
snoo
10000
1 5000
2 00 00
25000
innnn
18 7"
After Two Months
Nitrate Addition
(Core 5iAJ)
182
0
10000
15000
20000
5000
25000
30000
Concentration JF-4 (mq/kg dry weight)
Figure 27. Vertical distribution of JP-4 during bioremediation
demonst rati on. Land surface is at 187.5 m AM5T.. and original water
table is shown at 182.9 ir; AMSL.
39
-------
The average concentrations of total JP-4 and individual Alkylbenzenes
across the replicate boreholes at the various sampling times are presented in
Table 2. There was a large difference in JP-4 concentrations and
concentrations of alkylbenzenes in the first set of cores (See Table 3).
Alkylbenzene concentrations in borehole 51AF were more typical of prior cores
taken to map the spill. For example, a 2.1 m interval from borehole 51AZ,
(situated close to well cluster CG) averaged 4,100 mg/kg JP-4 and
8.6 mg/kg Benzene as determined by GC analysis (Hutchins et al., 1989).
Extensive depletion of alkylbenzenes occurred after hydraulic
equilibrium, particularly for Benzene and Toluene. After 2 months of nitrate
and nutrients, the removal of all the Alkylbenzenes vere extensive. The
averages presented in Table 2 are broken out by depth interval in Tables 3, 4,
and 5. Although concentration of total JP-4 and individual alkylbenzenes
varied widely with depth, the proportions of fuel components were remarkably
consistent at a given sampling time, indicating that the residual fuel was in
chemical equilibrium with the infiltrating water.
40
-------
Table 3. CONCENTRATIONS OF ALKYLBENZENES AND JP-4 IN CORE MATERIAL PRIOR TO
APPLICATION OF WATER OR NITRATE. BOREHOLE 51AF.
Elevation Benzene Toluene Ethylbenzene m+p-Xylene o-Xylene 'JP-4
(meters AMSL) mg/kg dry weight
184.34-184.43
0.0007
0.007
0.0016
0.018
0.01
114
184.22-184.34
0.077
2.3
1.08
9.6
3.7
5,360
184.13-184.22
0.226
9.0
6.8
31
10.0
5,800
184.04-184.13
0.46
22
16
57
17.3
.2,320
183.98-184.04
1.1
49
28
114
33
6,700
183.89-183.98
1.76
45
25
81
23
6,800
183.76-183.89
0.82
26
11.5
38
11.6
8,900
183.67-183.76
0.89
28
13.3
41
12.1
4,100
183.58-183.67
2.4
78
36
115
34
9,500
183.46-183.58
1.15
45
21
68
20
7,600
183.37-183.46
0.91
28
12.8
41
12.1
2,700
183.28-183.37
1.45
43
20
65
18.3
4,200
183.15-183.28
0.84
26
11.8
38
10.6
3,800
183.06-183.15
1.67
52
24
71
19.7
6,100
182.91-183.06
1.54
94
35
102
32
14,000
182.83-182.91
3.4
260
94
270
88
22,000
182.70-182.83
3.5
175
67
210
59
29,000
182.61-182.70
0.158
4.8
1.69
7.0
2.1
990
182.51-182.61
0.065
0.75
0.26
0.53
0.183
46
41
-------
Table 4. CONCENTRATIONS OF ALKYLBENZENES AND JP-4 IN CORE MATERIAL AFTER
AFTER HYDRAULIC EQUILIBRIUM BUT BEFORE NITRATE ADDITION. BOREHOLE
51AH-
Elevation Benzene Toluene Ethylbenzene m+p-Xylene o-Xylene JP-4
(meters AMSL) mg/kg dry weight
184.28-184.40
184.19-184.28
184.10-184.19
183.98-184.10
183.89-183.98
183.79-183.89
183.67-183.79
183.58-183.67
183.49-183.58
183.37-183.49
183.28-183.37
183.18-183.28
183.06-183.18
182.97-183.06
182.88-182.97
182.76-182.88
182.64-182.76
182.51-182.64
182.42-182.51
0.0043 0.017
0.0077 0.045
0.0065 0.038
0.0086 0.027
0.00] 2. 0.0075
0.0089 0.034
0.0058 0.028
0.0070 0.030
0.0085 0.0092
0.010 0.051
0.014 0.026
0.026 0.038
0.018 0.060
0.060 0.125
0.099 0.25
0.182 0.67
0.081 0.33
0.049 0.22
0.052 0.050
0.064 0.79
0.071 2.4
0.044 0.18
0.047 2.2
0.0052 0.71
0.048 2.4
0.037 1.8
0.027 2.2
0.010 2.3
0.055 2.1
0.074 3.8
0.069 6.1
0.081 5.5
0.35 13.7
0.73 18.8
1.53 34
1.02 18.3
0.95 12.2
0.015 0.106
0.38
360
0.069
2,500
0.98
-2,400
1.26
2,300
0.35
620
1.34
3,100
0.99
2,300
1.23
3,000
1.31
2,900
0.98
3,200
1.59
3,600
2.5
3,900
2.6
3,400
6.6
5,900
8.9
15,300
15.6
18,300
7.6
7,600
5.1
6,600
0.052
33
42
-------
Table 5. CONCENTRATIONS OF ALKYLBENZENES AND JP-4 AFTER TWO MONTHS OF
NITRATE ADDITION. BOREHOLE 51AK.
Elevation Benzene Toluene Ethylbenzene m+p-Xylene o-Xylene JP-4
(meters AMSL) mg/kg dry weight
184.34-184.40
184.25-184.34
184.16-184.25
184.04-184.16
183.95-184.04
183.86-183.95
183.76-183.86
183.70-183.76
183.55-183.70
183.46-183.55
183.34-183.46
183.25-183.34
183.15-183.25
183.03-183.15
182.94-183.03
182.88-182.94
182.85-182.88
182.64-182.86
182.54-182.64
182.42-182.54
182.33-182.42
0.003
0.046
0.003
0.010
0.008
0.078
0.003
0.010
0.008
0.076
0.010
0.022
0.011
0.078
0.005
0.089
0.005
0.017
0.008
0.019
0.014
0.020
0.016
0.019
0.014
0.014
0.020
0.017
0.032
0.035
0.142
0.083
0.160
0.05
0.002
0.019
NF
0.011
0.001
0.011
0.001
0.010
0.011
0.038
NF
0.005
0.053
0.118
0.007
0.020
0.014
0.013
0.013
0.034
0.020
0.063
0.014
0.050
0.016
0.059
0.018
0.053
0.022
0.065
0.013
0.032
0.013
0.056
0.020
0.057
0.024
0.077
0.047
0.154
0.02
0.03
0.002
0.002
0.007
0.035
0.011
0.063
0.013
0.062
0.032
410
0.069
1,100
0.22
1,900
0.18
-1,300
0.12
2,400
0.15
2,600
0.16
3,100
0.14
2,600
0.13
3,000
0.17
3,900
0.23
4,700
0.31
5,800
0.25
4,300
0.54
7,300
0.88
11,100
3.9
26,000
1.26
8,300
0.063
390
0.073
580
0.025
4
0.029
5
-------
Table 6. CONCENTRATIONS OF BENZENE AND JP-4 IN CORES PRIOR TO APPLICATION OF
WATER OR NITRATE, AND AFTER NITRATE ADDITION, AND PREDICTED
CONCENTRATIONS OF BENZENE IN THE POREVATER OF THE CORES.
After (Borehole 51AK)
Elevation Benzene JP-4 Benzene
(meters AMSL) (mg/kg dry wt.) (ug/L)
184.34-184.43
0.0007
114
2.5
184.34-184.40
0.003
410
2.9
184.22-184.34
0.077
5,360
5.6
184.25-184.34
0.003
1,100
1.1
184.13-184.22
0.226
5,800
15
184.16-184.25
0.008
1,900
1.7
184.04-184.13
0.46
2,320
78
184.04-184.16
0.003
1,300
1.1
183.98-184.04
1.1
6,700
65
183.95-184.04
0.008
2,400
1.3
183.89-183.98
1.76
6,800
102
183.86-183.95
0.010
2,600
1.5
183.76-183.89
0.82
8,900
36
183.76-183.86
0.011
3,100
1.4
183.67-183.76
0.89
4,100
85
183.70-183.76
0.005
2,600
0.8
183.58-183.67
2.4
9,500
99
183.55-183.70
0.005
3,000
0.7
183.46-183.58
1.15
7,600
59
183.46-183.55
0.008
3,900
0.8
183.37-183.46
0.91
2,700
132
183.34-183.46
0.014
4,700
1.2
183.28-183.37
1.45
4,200
137
183.25-183.34
0.016
5,800
1.1
183.15-183.28
0.84
3,800
90
183.15-183.25
0.014
4,300
1.3
183.06-183.15
1.67
6,100
111
183.03-183.15
0.020
7,300
1.1
182.91-183.06
1.54
14,000
45
182.94-183.03
0.032
11,100
1.1
182.83-182.91
3.4
22,000
63
182.88-182.94
0.142
26,000
2.1
182.70-182.83
3.5
29,000
49
182.85-182.88
0.160
8,300
7.6
182.61-182.70
0.158
990
65
182.64-182.86
0.002
390
2.0
Before (Borehole 51AF)
Elevation Benzene JP-4 Benzene
(meters AMSL) (mg\kg dry wt.) (ug/L)
-------
Table 6 extracts data from Tables 3 and 5, showing the Benzene content
and total JP-4 content before and after remediation. In general, Benzene
concentrations dropped two to three orders of magnitude. The fuel to water
partition coefficients (k_ . ) of Smith et al. (1981) were used to
« ruel-vater' — —
estimate the expected equilibrium solution concentration of Benzene in water
in contact with the core. The concentration of Benzene in the fuel was
determined by dividing the concentration of Benzene in the aquifer by the
concentration of JP-4 in the aquifer. The concentration of Benzene in the
water was estimated by dividing the concentration of Benzene in the fuel by
kf , . These estimates should indicate whether the fuel spill was
really remediated, or if there is a significant chance that unacceptable
concentrations of Benzene could return after the remediation was halted. With
one exception at 182.7 m, the estimated concentrations of Benzene were at or
just slightly above the Federal drinking water standard.
Selected cores were further analyzed to ascertain changes in
composition of the residual fuel hydrocarbon during the course of the study.
Samples were selected from five depth intervals: 1) unsaturated, clean zone;
2) unsaturated, contaminated zone; 3) contaminated zone at the water table; 4)
saturated, contaminated zone; and 5) in the relatively clean saturated zone
immediately beneath the contamination. The reconstructed ion chromatograms
were evaluated using selective ion searches for the four major classes of
hydrocarbons in JP-4. These classes are Alkanes (61X v/v), Cycloalkanes (29%
v/v), Alkylbenzenes (8X v/v), and Indans (IX v/v) (Smith et al., 1981).
Concentrations of these classes of hydrocarbons, along viTR TReir relative
proportions, are shown in Table 7.
Prior to hydraulic loading, the distribution was similar to that of
JP-4 in the unsaturated zone with the proportion of alkylbenzenes increasing
in the saturated zone. This would be expected if the more water-soluble
alkylbenzenes were leached from other regions of the site and were in
equilibrium with the residual fuel hydrocarbon. After flooding began and
before nitrate addition, much of the alkylbenzenes were removed and the alkane
fraction increased proportionally (Table 7). Estimates of total fuel by GC
analysis for JP-4 and GC/MS analysis for total class agreed reasonably well,
and indicated that the bulk mass decreased in the upper regions of the
contaminated zone and increased in the lower region, probably in response to
mobilization of the fuel. After two months of nitrate addition, there was
generally a further decline in concentrations of each class, although the
proportions did not change appreciably (Table 7). Review of the reconstructed
ion chromatograms reveals that most of the lower molecular weight compounds of
each class have been removed by this time.
The BTEX data shown in Table 7 are consistent with the alkylbenzene
data shown in Table 3 for the initial sampling event prior to hydraulic
loading; the target aromatics comprise 20-40% of the total alkylbenzene
fraction and show similar changes in concentration with depth. However, after
flooding and before nitrate addition, the target aromatics comprise only 2-20X
of the total alkylbenzene fraction, and most of this contribution is due to
1,2,4-Trimethylbenzene (pseudocumene)- This again indicates that the lower
molecular weight aromatics are being preferentially leached and/or degraded.
Toluene and Ethylbenzene are completely removed from most of the cores by this
time, but the corresponding alkane and cycloalkane are still present in
relatively high concentrations (Table 7). After 2 months of nitrate addition,
45
-------
most of the BTEX is at or below the detection limits in all of the sample
cores, although a significant amount of the higher molecular weight
alkylbenzenes still remain (cf. Table 7 with Table 5).
Table 7. CHANGES IN CONCENTRATIONS AND PROPORTIONS OF THE MAJOR CLASSES OF
HYDROCARBONS IN SELECTED CORES. DATA ARE IN MG/KG DRY WEIGHT,
MEAN OF TWO REPLICATE SUBSAMPLES OF SINGLE CORES.
Core Elevation
(m AMSL) and
Description
Class
TREATMENT
Prior to After Flood- After Two
Hydraulic ing, Before Months of
Loading Nitrate Nitrate
185.0-
185.3
Unsat'd,
Clean
Alkylbenzenes
Indans
Alkanes
Cycloalkanes
0.7
1.5
0.9
0.6
0.5
Total of above
JP-4
0.0
3.1
5.3
1.1
5.1
;
—
-
Alkylbenzenes
809
(11%)
175
( 4%)
17
( 1%)
183.8-
Indans
153
( 2%)
63
( 2%)
12
( 1%)
184.1
Alkanes
5460
(74%)
3270
(80%)
1500
(83%)
Unsat'd,
Cycloalkanes
978
(13%)
586
(14%)
266
(15%)
Contam
Total of above
JP-4
7400
6720
4090
1610
1800
1310
Alkylbenzenes
2010
(15%)
124
( 1%)
258
( 3%)
182.9-
Indans
236
( 2X)
34
( 0%)
101
( 1%)
183.2
Alkanes
9170
(67%)
8550
(85%)
7080
(80%)
Water
Cycloalkanes
2130
(16%)
1380
(14%)
1450
(16%)
Table
Total of above
JP-4
13500
22100
10100
6060
8890
8300
Alkylbenzenes
6.5
(28%)
230
( 4%)
17
( *%)
182.6-
Indans
1.6
( ?%)
89
( 2%)
14
( 4%)
182.9
Alkanes
11.8
(50%)
4110
(79%)
281
(73%)
Sat'd,
Cycloalkanes
3.6
(15%)
808
(15%)
73
(19%)
Contam
Total of above
JP-4
23.5
46.1
5230
8224
385
386
Alkylbenzenes
7.1
3.2
0.9
182.0-
Indans
1.0
1.9
-
182.3
Alkanes
1.6
124
2.1
Sat'd,
Cycloalkanes
-
21.2
1.5
Clean
Total of above
JP-4
9.7
10.2
150
91.1
4.5
18.0
46
-------
DISCUSSION
The reduction in BTEX compounds in the aquifer, as a result of the
demonstration, is depicted in Table 8. With the exception of one of the five
monitoring wells, Benzene was removed to concentrations much lower than the
Federal Drinking Water Standard. Toluene was removed in all the monitoring
wells. Removal of Ethylbenzene and the Xylenes followed an interesting
pattern. Ethylbenzene and the different Xylene isomers were persistent at low
concentrations in different wells. This suggests that there is no one fixed
pattern of BTEX utilization, even at the same site.
Table 8. BTEX CONCENTRATIONS IN THE AQUIFER UNDER THE INFILTRATION GALLERY
AT THE END OF THE DEMONSTRATION.
After Nitrate Addition
165 Days
Compound CA-6 CB-6 CC-6 CD-6 CE-6
(ug/1)
Benzene
<0.1
10.2
<0.1
<0.1
<0.1
Toluene
<0.1
<0.1
<0.1
<0.1
<0.1
Ethylbenzene
<0.1
2.3
<0.1
4.1
4.6
£-Xylene
<0.1
<0.1
32
<0.4
12
m-Xylene
<0.1
<0.1
56
<0.1
<0.1
o-Xylene
<0.1
<0.1
<0.1
24
<0.1
The concentration of BTEX in the monitoring wells was consistent with
the concentrations that would be expected from the core analysis. The
concentration of each BTEX compound and the concentration of JP-4 in each core
was used to calculate the concentration of each BTEX compound in the JP-4 for
each core. Then partition coefficients for BTEX between JP-4 and water were
used to estimate the equilibrium concentrations of the BTEX compounds in the
pore water (Smith et^ al., 1981). The estimates from all the cores collected
at a particular time period were averaged to predict the concentration of each
BTEX compound in the monitoring wells underneath the infiltration gallery.
These estimates were weighted for the depth interval represented by the core.
The estimates and the monitoring well data agree reasonably well,
although actual concentrations in the monitoring wells prior to recirculation
of water were much higher than predicted from the core samples (Table 9).
There is little chance that the reduction in BTEX concentration in the
monitoring wells was caused by a mass-transfer limitation between the JP-4 and
the infiltrating water, and that unacceptable concentrations of Benzene will
be re-established in the monitoring wells once pumping has stopped.
47
-------
Table 9. COMPARSION OF ACTUAL ALKYLBENZENES CONCENTRATIONS IN MONITORING
WELLS WITH PREDICTED VALUES AT EQUILIBRIUM, BASED ON PARTITION THEORY
OF INDIVIDUAL ALKYLBENZENES BETWEEN THE RESIDUAL JP-4 AND THE
INFILTRATING WATER.
Treatment
Compound Value Prior to Hydraulic After Flooding After Two Months
Loading Before Nitrate Nitrate
ug/li ter
Benzene
Actual
760
<1
<1
Predicted
58
5
2
Toluene
Actual
4500
17
<1
Predicted
1400
12
15
Ethyl-
Actual
840
44
6
benzene
Predicted
450
9
6
m,p-Xylene
Actual
2600
490
* 23
Predicted
1300
200
27
o-Xylene
Actual
1380
260
37
Predicted
480
70
18
Actual concentrations are averages of samples from each of the five cluster
monitoring wells under the infiltration gallery at level 2 (or level 3 prior
to hydraulic loading) and level 4.
The disappearance of Benzene and Toluene in the presence of significant
oxygen prior to the addition of nitrate on Day 40 suggests that the aerobic
degradation of these compounds may have occurred. While very little dissolved
oxygen was measured in the recirculation water (on the order of 1.0 ppm), the
high flow rates carried a significant mass of oxygen into the injection
gallery (Table 10). A calculation of the cumulative mass of dissolved oxygen
transported to the injection gallery reveals 140 k.g of dissolved oxygen to
have been transported into the injection gallery by Day 109 when the last set
of cores was acquired. Calculation of the theoretical oxygen demand of the
BTEX compounds in the fuel before circulation of water is presented in Table
10. When the theoretical oxygen demand of BTEX compounds is calculated (Table
10) and compared to the cumulative supply of dissolved oxygen, it appears that
the theoretical oxygen demand of Benzene and Toluene in the original spill
could have been met by the supply of oxygen delivered in the recirculated
water after 12 days, but this implies that Benzene and Toluene were the only
components demanding oxygen. The theoretical nitrate demand of the remaining
alkylbenzenes could easily have been met by the nitrate supplied.
The amount of nitrate removed during passage of the infiltration water
through the contaminated area (499 kg nitrate-nitrogen) is almost an order of
magnitude greater than the theoretical nitrate demand of Benzene, Toluene,
Ethylbenzene. Xylenes, and Trimethylbenzene (Table 11). Obviously, other
compounds in the fuel are exhibiting a nitrate demand, and an analysis of BTEX
cannot be used to estimate nitrate requirements. Regardless of the particular
electron acceptor used to respire a particular alkylbenzene, oxygen provided
less than 4% of the electron accepting capacity (Table 11).
48
-------
Table 10. CALCULATED MASS OF NUTRIENTS AND CONTAMINANTS IN INFILTRATION
WATER BEFORE AND AFTER NITRATE ADDITION, AND CHANGES IN MASS AFTER
PASSAGE OF THE WATER THROUGH THE CONTAMINATED ZONE.
Before Nitrate Addition
(Day = 1 to 41)
After Nitrate Addition
(Day = 41 to 109)
Parameter
Total Mass
Supplied in
Recharge
Mass Removed
from Recharge
by Level 6
Total Mass
Supplied in
Recharge
Mass Removed
from Recharge
by Level 6
-kg-
Dissolved-Oxygen
92
55
48
4
Ni t ra te-Nit rogen
33
25
690
499
Nitri te-Ni trogen
<2
<2
66
26
Ammonia-Ni trogen
<2
<2
301
18
Phosphate-Phosphorus
<2
<2
97
20
Dissolved-Oxygen 6.9
Nitrate+Nitrite-Nitrogen <0.7
-kilomoles electrons accepted-
0.5
178
Table 11. THEORETICAL OXEGEN AND NITRATE DEMAND OF BTEX COMPOUNDS.
Benzene
Toluene
Ethylbenzene
m+p Xylene
o-Xylene
Pseudocumene
Total Alkylbenzenes
Total JP-4
Actually supplied
00 Demand
" (Kg)
19.6
48.4
43.4
131.1
60.9
85.5
799
8,072
140
NO^-N Demand
(Kg)
6.8
16.5
13-7
41.4
13.6
27.0
252.0
499
49
-------
QUALITY ASSURANCE AND QUALITY CONTROL
Extraction of Subcores
Subcores were taken from field samples collected in Mason jars by
pushing a plastic syringe with the end removed the entire length of the core.
The subcores were then extruded with the plunger into a 40 mg VOA bottle
containing 10 ml of acidified water. The samples were capped with a Teflon-
faced silicone septa. The samples were stored on ice or at 4 C until
extraction. Methylene chloride was injected through the septum with a
syringe, the bottle was shaken at high speed with a wrist-action shaker for
15 minutes. The extract was dried, then stored in a crimp-cap autosampler
vial at 4 C until analysis.
Fuel Carbon Analysis for JP-4
GC analysis was carried out following RSKS0P-72, Revision No. 0.
One ul of extract was injected into a HP5880 Gas Chromatograph with an HP7673A
Autosampler. The column was a J&V Scientific 15 m x 0.53 mm i.d. DB-5, 1.5
uM. Injector and FID temperature was 300 C, solvent purge at 0.7 min.
Quantification of individual components used a 4-point standard curve
(1, 10, 100, 1000 ug/ml) as calculated by the HP5880 GC terminal.
Quantification of total JP-4 was based on five standard concentrations
(0.0369, 0.0738, 0./38, and 73.8 ug/ul) of JP-4 free product from the Traverse
City spill, using linear regression software (Microsoft EXCEL) and a Macintosh
computer. See Table 12 for precision and accuracy estimates.
able 12. QUALITY ASSURANCE INDICATORS FOR JP-4 ANALYSIS.
Component
Precision
(Coeff. Var.)
[100 ppm]
Accuracy
(Z Difference)
[low] [high]
Detection Limit
(ug/ml extract) (mg/kg dry wt)
Benzene
0.03
1.91
3.79
0.072
0.017
Toluene
0.03
0.72
2.31
0.069
0.016
Ethylbenzene
0.03
0.24
1.73
0.036
0.008
m+p-Xylene
0.03
0.90
2.32
0.039
0.009
O-Xylene
0.08
nd
nd
0.045
0.010
JP-4 Standard
0.04
nd
nd
0.014
0.003
Precision calculated as the Coefficient of Variation (Standard
Deviation/Mean).
Accuracy calculated as Percent Difference (Analyzed-True/True)*100.
Detection limit calculated as three times the standard deviation of the
lowest calibration standard.
50
-------
Cores from equivalent depths in duplicate boreholes 1.3 m apart were
compared to estimate varibility due to natural heterogeneity. The coefficient
of variation at field scale for Total JP-4 was 0.13.
Mass Spectral Analysis of Core Extracts
The Methylene chloride extracts of the subcores were analyzed for
Benzene, Toluene, Ethylbenzene, m+p-Xylene and o-Xylene using a Finnigan 4500
GC/MS system. For samples which contained compounds in the concentration
range 1 to 200 ppm, MS scanning parameters were set to scan for 35 to 110 m/z
in 0.2 s. For the concentration range 10 to 10,000 ppb, scan parameters were
set to accumulate peak areas at 78, 91, and 96 m/z for 0.05 s each with a
total scan time of 0.20 s. The electron energy of the MS ion source was set
at 70 eV. A 100 ul aliquot of each core extract (neat or diluted) was spiked
with 10 ul of the internal standard Fluorobenzene to achieve 1.0 ppm of 50 ppm
depending on the sample concentration range. In either case, a 30 meter, 0.25
mm i.d. J&V DB-5 fused silica capillary column with a 1.0 um film thickness
was held at 5 C for 0.5 min^ temperature programmed to 30 C at 25 . C / min
and finally to 200 C at 10 C/min. A 2.0 ul aliquot of each spiked standard
or sample was injected splitless for 30 s before the temperature program
began.
Quality assurance was maintained during the sample analysis by
injecting check standards of the original calibrating solution and EPA quality
assurance standards. These were run at ppm and ppb concentration ranges. The
coefficient of variation for BTEX at the 50 ppb level was estimated to be
0.09. The method detection limit at 99% confidence is estimated to be 13 ppb.
This corresponds to a concentration of 0.003 mg BTEX compound/kg dry wt. core
sample.
Analysis of Water Samples for BTEX
All BTEX samples were run with the headspace method on a HP5890 Gas
Chromatograph equiped with a PID and an FID in series. A packed column of 5%
SP-1200/1.15% Bentone 34 on 100/120 Supelcoport and an isothermal program of
80 C was used to resolve the largest compounds. A standard curve containing
three points (8 to 4300 ppb) was constructed monthly. Standard solutions were
kept in the freezer when not in use. Each day, two or three standards were
checked against the curve.
The detection limit for BTEX was 0.1 ug/1. The accuracy (Z recovery
of spikes) for Benzene and Toluene were 97.6 and 92.7% respectively. The
coefficient of variation of dupicate lab samples was 0.05 for Benzene and
Toluene. The coefficient of variation of replicate samples from the five
wells at each depth in the study area was 1.06 for Benzene and 0.40 for
Toluene. The pooled estimate of accuracy for the Xylenes was 90.2%, the
coefficient of variation of lab samples was 0.05, and the coefficient of
variation at field scale was 0.27 for p-Xylene, 0.45 for m-Xylene, and 0.22
for o-Xylene.
Dissolved Oxygen Analysis of Water Samples
51
-------
Dissolved oxygen was determined by titration using the azide
modification of the Winkler method (Methods for Chemical Analysis of Water and
Waste EPA - 600/4-79-020, Revised March 1983) using 0.06 M KI as a standard.
One standard and one blank were run with each sample set. The detection limit
was 0.15 mg/1. The coefficient at field scale was 0.13.
Analysis of Water Samples for Inorganic Ions
A Technicon AutoAnalyzer II was used to determine concentrations of
nitrate, nitrite, ammonia, ortho-phosphate, and chloride. The instument was
calibrated each time it was used. American Chemistry Society grade chemicals
were used for inorganic standards. Fresh working standards were prepared
before each analysis. All standards were kept in the refrigerator, with the
exception of chloride, which was kept at room temperature. A standard curve
containing no less than four points was calculated with the new working
standards before each run. A water blank and a duplicate sample were
analyzed after every ten samples. A standard was analyzed at the end of the
run to verify the stability of the initial instrument calibration. One spike
per run was analyzed.
For nitrate, nitrate, and ammonium ion, the accuracy (% recovery of
spikes) was 105.8, 107.8 and 98.9% respectively. The coefficient of variation
of lab samples was 0.019, 0.015, and 0.013 respectively, and the coefficient
of variation at field scale was 0.78, 0.13, and 0.21 respectively.
Most of the chloride analyses for the tracer tests were titrated by
hand, using the mercuric nitrate method for chloride (EPA Method 325.3 in
Methods for Chemical Analysis of Water and Wastes, EPA-600/4-79 020, revised
March 1983). The accuracy (% recovery of spikes) was 100.8% and the
coefficient of variation of lab samples was 0.009.
COSTS
The costs associated with the demonstration were broken down into
capital and operating costs. This evaluation does not include analytical
costs, because the investment in analytical effort was at least ten times
greater than ordinary practice in the industry. The infrastructure still
exists and will be used by the U.S. Coast Guard to complete the remediation.
The cost evaluation prorates the construction costs over a 5-year period, the
time the Coast Guard expects to be required to remediate the Air Station.
Unit costs for the remediation were calculated by dividing the cost
for construction, labor, chemicals and electrical service by (1) the volume of
JP-4 underneath the infiltration gallery, (2) the volume of aquifer material
contaminated with JP-4 under the infiltration gallery, and (3) the volume of
aquifer between the infiltration gallery and the confining unit beneath the
aquifer. The unit costs for the remediation were $84 per gallon^JP^, $200
per m of aquifer material contaminated with JP-4, and $17 per m of aquifer
material down to the confining layer. Costs are summarized in Table 13.
52
-------
Table 13. CAPITAL AND OPERATING COSTS.
Category Unit Cost Consumed Monthly Monthly Average Total Project
$/lb lb $ $
Construction 138,816
Construction prorated to five-year service life 27,764
Labor (10 mo) 2,256 22,560
Chemicals (8 Mo)
Sodium 0.165 3708 612 3,671
Ni trate
Ammonium 0.285 450 129 769
Chloride
Monopotassium 1.095 225 247 1,478
Phosphate
Disodium 0.81 225 182 1,093
Phosphate
Electricity 340 2,730
TOTAL 171,126
TOTAL PRORATED TO 5-YEAR SERVICE LIFE 60,000
53
-------
CONCLUSIONS
The technology worked well, even though the concentrations of nitrate
injected into the aquifer were held at or below the Drinking Water Standard
for nitate. Monitoring wells under the study site were below appropriate
State of Michigan and Federal Drinking Water Standards within 165 days. The
reduction in BTEX concentrations were extensive, on the order of three to four
orders of magnitude.
Computer simulation was very useful in determining reasonable
injection and withdrawal rates of water before construction of the gallery.
It allowed a prediction of the areal extent of the hydraulically-affected
zones and an estimate of the vertical mounding of the water surface at the
injection gallery and drawdown at the withdrawal wells. This information was
also required to size pumping and piping apparatus, to estimate electrical
power requirements, to size mixing tanks, and to estimate bulk chemical
purchases.
Proper hydraulic control of an infiltration gallery to maintain
saturated conditions throughout the contaminated zone is absolutely necessary
before any quantitative evaluation of contaminant removal can be made.
In this study, the calculated effect of dilution using partitioning
theory on the equilibrium solution concentration of BTEX compounds does not
explain their continued disappearance. Including the effect of steady purging
of a certain percentage of the recirculated water better explains the trends
in the observed data, but does not explain the sharp drops seen in the
solution concentrations of Benzene, Toluene, and m-Xylene. After taking these
physical processes into account, biological processes provide a reasonable
explanation for Benzene, Toluene, and m+p-Xylene removal. The decrease in
solution concentrations of o-Xylene was observed to follow that predicted by
dilution and wasting, indicating that it may be less sensitive to degradation
processes. Concentrations decreased significantly in the monitoring wells only
towards the end of the demonstration when the other compounds were depleted
The laboratory work has shown that aromatic hydrocarbons, with the
possible exception of Benzene, can be degraded under strictly anaerobic
conditions by native subsurface bacteria using nitrate as the terminal
electron acceptor (Hutchins et^. al, 1989). Toluene is most easily degraded
with the Xylenes being more recalcitrant. The field demonstration project has
demonstrated, through extensive core analyses, that simple hydraulic flooding
will remove a significant amount of the lower molecular weight aromatic
hydrocarbons in far greater proportion than the residual fuel hydrocarbons.
Addition of nitrate and nutrients results in denitrification occurring
within the contaminated zone, as shown by decreases in nitrate concentrations
through the infiltration gallery along with transient nitrite production.
Core analyses revealed that BTEX was removed to very low concentrations after
two months of nitrate addition. There was a general decrease in all
constituents of the jet fuel, although detectable concentrations of the higher
molecular weight alkylbenzenes still remain.
54
-------
It was impossible to determine the extent to which a particular BTEX
compound was removed through denitrification. Aerobic biodegradation was
feasible in the system. It is certain, however, that the nitrate serves to
satisfy a portion of the oxygen demand which would otherwise be exerted.
The actual amount of nitrate consumed was ten times greater than the
theoretical nitrate demand for oxidation of the BTEX compounds alone.
55
-------
REFERENCES
Battermann, G. 1986. Decontamination of polluted aquifers by biodegradation.
In; 1985 International TNO Conference on Contaminated Soil, J.W. Assink and
W.J. van der Brink, Nijhoff, Dordrecht, pp 711-722.
Bouchard, D.C., C.G. Enfield, and M.D. Piwoni. 1989. Transport processes
involving organic chemicals. In: Reactions and Movement of Organic Chemicals
in Soil, B.L. Sawhney and K. Brown. Soil Science Society of America and
American Society of Agronomy, SSSA Special Publication Number 22. pp 349-371.
Downs, W.C., S.R. Hutchins, J.T. Wilson, R.H. Douglass, and D.J. Hendrix.
1989. Pilot project on biorestoration of fuel-contaminated aquifer using
nitrate: Part I - Field design and ground water modeling. In: Proceedings,
Petroleum Hydrocarbons and Organic Chemicals in Ground Water: Prevention,
Detection, and Restoration. Houston, Texas, November 15 - 17, 1989, pp. 219-
233.
Feliciano, D. 1984. Leaking undergound storage tanks: a potential*
environmental problem. Congressional Research Service Report. U.S. Library
of Congress. Washington, D.C.
Hutchins, S.R., J.T. Wilson, R.H. Douglass, and D.J. Hendrix. 1989. Field and
laboratory evaluation of the use of nitrate to remove BTX from a fuel spill.
*n: Proceedings, Environmental Research Conference on Groundwater Quality and
Waste PisposaT7 EPA/EPRI (in press).
Kuhn, E.P. J. Zeyer, P. Eicher, and R.P. Schwarzenbach, 1988. Anaerobic
degradation of alkylated benzenes in denitrifying laboratory columns. AddI.
Environ. Microbiol. 54:490-496.
Leach, L.E., F.P. Beck, J.T. Wilson, and D.H. Kampbell. 1988. Aseptic
subsurface sampling techniques for hollow-stem auger drilling. In:
Proceedings, Second National Outdoor Action Conference on Aquifer Restoration.
Ground Water Monitoring and Geophysical Methods. NWWA , Las Vegas, Nevada,
May 23-26, 1988. Vol 1:31-51.
Lee, M.D., J.M. Thomas, R.C. Borden, P.B. Bedient, J.T. Wilson, and C.H. Ward.
1988. Biorestoration of aquifers contaminated with organic compounds. Crit.
Rev. Environ. Control. 18:29-89.
Lemon, L.A., J.R. Barbarok, and J.F. Barker. 1989. Biotransformation of BTEX
Under Anaerobic Denitrifying Conditions: Evaluation of Field Observations. In:
Proceedings of the FOCUS Conference on Eastern Regional Ground Water Issues.
Oct. 17-19, 1989, Kitchener, Ontario, Canada. National Water Well
Association, Dublin, Ohio.
Major, D.W., C.I. Mayfield, and J.F. Barker. 1988. Biotransformation of
benzene by denitrification in aquifer sand. Ground Water. 26:8-14.
McDonald, M.G., and A.W. Harbaugh. 1988. A modular three-dimensional
finite-difference ground-water flow model. U.S. Geological Survey Open-File
Report 83-875. 528 pp.
56
-------
Mihelcic, J.R., and R.G- Luthy. 1988. Microbial degradation of acenaphthene
and naphthalene under denitrification conditions in soil-water systems. Appl.
Environ. Microbiol. 54:1188-1198.
Sheehan, P.J., R.V. Schneiter, T.K.G. Mohr and R.M. Gersberg. 1988.
Bioreclamation of gasoline contamination without oxygen addition. In:
Proceedings of the Second National Outdoor Action Conference on Aquifer
Restoration, Ground Water Monitoring and Geophysical Methods, Volume I,
National Water Well Association, Las Vegas, Nevada, May 23-26, 1988. 183-199.
Smith, J.H., J.C. Harper, and H. Jaber. 1981. Analysis and environmental
fate of Air Force distillate and high density fuels. Engineering and Services
Laboratory, Air Force Engineering and Services Center, Tyndall Air Force
Base. Report ESL-TR-81-54.
Thomas, J.M., M.D. Lee, P.B. Bedient, R.C. Borden, L.W. Canter, and C.H. Ward.
1987. Leaking underground storage tanks: remediation with emphasis on in situ
biorestoration. EPA 600/2-87/008.
Tiedje, J.M. 1988. Ecology of denitrification and dissimilatory nitfate
reduction to ammonium. In: Biology of Anaerobic Microorganisms, A.J.B.
Zehnder, John Wiley and Sons, New York. pp. 179-244.
Tventer, F.R., T.R. Cummings, and N.G. Grannemann. 1985. Groundwater
contamination in East Bay Township, Michigan. U.S. Geological Survey Water
Resources Investigation Report. 85-4064.
U.S. Environmental Protection Agency. EPA. 1979. Methods for chemical
analysis of water and wastes. EPA-600/4-79-020.
Vandegrift, S.A. and D.H. Kampbell. 1988. Gas chromatographic determination
of aviation gasoline and JP-4 jet fuel in subsurface core samples. J.
Chromatogr. Sci. 26:566-569.
Wilson, J.L. and S.H. Conrad. 1984. Is physical displacement of residual
hydrocarbons a realistic possibility in aquifer restoration? In:
Proceedings, Conference on Petroleum Hydrocarbons and Organic Chemicals in
Ground Water: Prevention, Detection, and Restoration. NWWA, API, Houston, TX.
pp. 274-298.
Wilson, J.T., L.E. Leach, M. Henson, and J.N. Jones. 1986. In situ
biorestoration as a ground water remediation technique. Ground Water
Monitoring Review. 6:56-64.
57
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completr
1. REPORT NO. 2.
KPA/60U/2-91/U0y
3.
4. TITLE AND SUBTITLE
NITRATE FOR BIORESTORATION OF AN AQUIFER CONTAMINATED'
WITH JET FUEL . .
5. REPORT DATE
March 1991
6. PERFORMING ORGANIZATION CODE
71AUTHOR(S) p o
iif KS5,nS D-J- "e"dr" D-D- r,no '"-H. Douglass
S!f! SNSS™ 5-A- K°"aCS 5'-»-
8. PERFORMING ORGANIZATION REPORT NO
9-PERFORMING ORGANIZATION NAME AND ADDRESS
p'n'ERni'*i!mq'~ 2Solar Universal 3NSI Technologies, Inc.
a Traverse Citv, MI RSKERL
Ada, OK 74820 " A(ja, 0K 74B2q
ravcp50 Group "U.S. COAST GUARD
Traverse City, MI Cleveland. OH 44199
10. PROGRAM ELEMENT NO.
TEKY1A
1 1. CONTRACT/GRANT NO.
DK69933299
12. SPONSORING AGENCY NAME AND ADDRESS
Robert S. Kerr Environmental Research T
------- |