United States Office of
Environmental Protection Research and Development
Agency Cincinnati, OH 45268
EPA/540/R-94/510a
August 1995
&EPA
SITE Technology Capsule
In Situ Steam Enhanced
Recovery Process
Abstract
The SERP technology is designed to treat soils contaminated
with VOCs and SVOCs in situ. Steam injection and vacuum
extraction are used to remove the organic compounds from
the soil and concentrate them for disposal or recycling.
A full-scale demonstration of SERP was conducted at the
Rainbow Disposal site in Huntington Beach, CA. The technol-
ogy removed some of the diesel fuel hydrocarbons from the
site soil, but did not meet the site cleanup objective of 1,000
mg/kg (ppm) total petroleum hydrocarbons (TPH). Operational
problems during treatment may have impacted the treatment
effectiveness. A detailed economic analysis of the application
of this technology was performed.
The SERP technology was evaluated based on seven criteria
used for decision-making in the Superfund Feasibility Study
process. Results of the evaluation are summarized in Table 1.
Introduction
This Technology Capsule provides information on the U.S.
Environmental Protection Agency (EPA) Superfund Innovative
Technology Evaluation (SITE) Program Demonstration of the
Steam Enhanced Recovery Process (SERP). This technology
is designed to remove volatile and semivolatile organic com-
pounds (VOCs and SVOCs) from soils in situ and was evalu-
ated during a full-scale remediation of diesel-contaminated
soils in Huntington Beach, CA. The SERP treatment occurred
between August 1991 and August 1993, with post-treatment
soil sampling completed under the SITE Program in Septem-
ber 1993. Information in this capsule describes specific site
characteristics and results of the SITE Demonstration. This
capsule contains the following information:
Abstract
Technology Description
Technology Applicability and Site Requirements
Technology Limitations
Performance Data
Process Residuals
Economics
Technology Status
SITE Program
Source for Further Information.
Technology Description
SERP is an in situ technology that uses steam injection and
vacuum extraction to remove VOCs and SVOCs from soil.
Treatment is performed using injection wells and extraction
wells that are constructed in the soil.
The steam is introduced into the soil through injection wells
and is drawn through the soil towards the extraction wells, A
vacuum pump is used to draw a vacuum on the extraction
wells, while liquid lift pumps in each well are used to remove
the accumulated liquid. The steam heats the soil matrix,
causing the more volatile compounds to vaporize. The pres-
sure gradient moves the steam, pore water, and contami-
nants (in liquid and vapor form) toward the extraction wells.
The application of steam to the contaminated soil provides
several potential mechanisms for removal of contaminants.The
heat increases the vapor pressure of the less volatile com-
pounds and decreases their viscosity, thus making them easier
to desorb from soil particles [1]. The condensation of steam
on soil particles provides energy to release adsorbed con-
taminant molecules. The addition of water in the form of
steam dilutes the existing contaminated pore water, causing a
Printed on Recycled Paper
SUPERFUND INNOVATIVE
TECHNOLOGY EVALUATION
-------
Table 1 Comparison of SERF with Seven of the Feasibility Study Criteria
CRITERIA
Overall Protection
of Human Health
and the Environment
SERF reduced soil
concentrations
without excavation.
Technology may
reduce the mobility
of contamination into
groundwater after
treatment, due to
reduced concentrat-
ions and free product.
Process did not
appear to cause
lateral or down-
ward migration of
contaminants.
Compliance
with ARARs
Process did not
meet soil clean-
up criterion, on
the average, in
this application.
Less soil is
excavated, thus
less soil requires
disposal, which
avoids land
ban restrictions.
Permits for drilling,
operating, and air
and water dis-
charges are
required.
Long- Term Effec-
tiveness and
Permanence
A portion of con-
taminants is
permanently re-
moved from the soil.
Removed con-
taminants can
be incinerated
or recycled.
Lower levels of
residual contami-
nation present
reduced risk.
Reduction of
Toxicity, Mobility,
or Volume
Through
Treatment
Treated soil had
lower concentrations
overall; some areas
were cleaned to
well below the
cleanup criterion.
Remaining contam-
inants may be less
mobile since less
tightly adhered con-
taminants were
removed.
Lower soil concen-
trations are
amenable to
natural or enhanced
biodegradation.
Volume of technol-
ogy residuals is
small compared to
the treated soil
volume.
Short- Term
Effectiveness
Soil is treated
below ground
so potential air
emissions are
minimized.
Other activities
can continue at
surface of treat-
ment area with
minor disruption.
Drilling and treat-
ment may cause
noise and dust
emissions, which
can be mitigated.
Implementability
Technology uses
widely available
construction and
process equipment.
Most regulatory
permits are
readily acquired
for fuel-related
cleanups.
Treatment of sites
with other chem-
icals may pose
additional problems.
Operational
problems that
may delay the
remediation can
occur.
The technology
may not be able
to meet stringent
cleanup require-
ments, necessitat-
ing post-process-
ing such as assist-
ed biodegradation.
Cost
Total cost
ranged from
$29 to
fora large
site, depending
on the on-line
factor.
Capital equip-
ment, start-up,
and utilities
costs are high.
Remediation
time is the
major factor in
the costs.
Because the
process oper-
ates in situ,
off-site disposal
costs are min-
imized.
soil flushing action that carries dissolved contamination to the
extraction wells. [1]
At the Rainbow Disposal site, water used in the treatment
process was extracted from an on-site water well. An average
of more than 20,000 gal of water was used each day during the
24-hr cycle of operation. The water was softened, chemically
conditioned, and preheated in a heat exchanger that simulta-
neously cooled the liquids removed from the extraction wells.
This water was collected in a boiler feed tank and passed to
the boiler. High quality steam was then injected into the soil
through the injection wells. A schematic of the process is
shown in Figure 1.
Thirty-five injection wells and 38 extraction wells were used at
the Rainbow Disposal site. The wells were arranged in an
overlapping pattern of alternating injection wells and extraction
wells. Over the treatment area of approximately 2.3 acres, the
spacing between wells was about 45 ft for wells of the opposite
type and 60 ft for wells of the same type. The site layout is
shown in Figure 2.
Contaminant vapor and contaminated water were removed
from the extraction wells to the above-ground treatment system
for further treatment. There, the liquid contaminants were gravi-
metrically removed from the water by an oil/water separator.
The recovered diesel was stored in a large collection tank. The
aqueous phase from the separator was passed through a
series of 5-micron filters and carbon adsorbers to remove the
residual organic contamination. The cleaned water was dis-
charged to an underground storm sewer through a pipe. This
treated water stream was sampled and tested every two weeks
for TPH and for benzene, toluene, ethylbenzene, and xylenes
(BTEX) to fulfill the requirements of a National Pollutant Dis-
charge Elimination System (NPDES) permit.
The vapor phase removed from the extraction wells was passed
through a knock-out drum to remove entrained liquids and
solids. After steam breakthrough occurred in most of the ex-
traction wells, a simple copper coil condenser was added to
the system to lower the temperature of the vapor and increase
the condensation of water and contaminant. Most of the re-
moved diesel compounds remained in the vapor phase during
treatment. The vapor stream was treated in a thermal oxidizer
unit (TOU) that used electrical heating to destroy the organic
compounds. The treated gas stream from the TOU was re-
leased through a stack to the atmosphere.
The growth of the steam zone in the soil was monitored
through use of temperature wells equipped with thermocouples.
By using thermocouples to detect the rise in temperature that
precedes the steam front, the progress of the steam zone in
the soil can be determined. Temperature wells inside the treat-
ment area were used to ensure that the flow of steam and heat
was under control. Temperature wells outside the treatment
area were monitored to determine whether off-site migration of
-------
f
4 (VW i,w//f m i ju* Uniting ljn"c t'aftxw
, ic;
F/gure 1 Schematic of the SERF process.
• Sampling Location
Perimeter of
Contatrmattpn
F/gure 2. Tfte /ayouf of SERF wells and soil sampling locations at the Rainbow Disposal site.
-------
the steam was occurring. If an increase in perimeter soil tem-
perature had been detected, process conditions could have
been changed to regain control of the steam zone.
The process configuration at any site must be determined by
site-specific characteristics. The equipment used for SERP
treatment consists of standard, available components.
Technology Applicability and Site
Requirements
Because SERP operates on soils in situ, specific contaminant
and site characteristics are required for the technology to be
effective. The following are site requirements for the use of
SERP:
. The contamination must consist of VOCs, SVOCs, or a
combination of both.
. Maximum concentrations of 200 to 1,000 mg/kg are typical
limits for compounds with a liquid phase that is heavier than
water, based on the characteristics of the contaminant. If
higher concentrations are present, there is a potential for
downward migration when these compounds are concen-
trated by the process [2].
. The area of soil to be treated must have a confining layer
below the treatment zone to prevent downward contaminant
migration. This confining layer could be a watertable, a zone
with permeabilities two or more orders of magnitude less than
the overlying treatment zone, bedrock, or an artificial bound-
ary. It is possible that a much more permeable zone can act
as a confining layer if a flow of steam can be maintained in this
zone [2].
. If the treatment zone is close to the soil surface, a confining
layer above the area of soil to be treated is required to prevent
short-circuiting of the steam to the surface. If the site geology
does not have such a layer, a temporary concrete or asphalt
cap can be installed before treatment is started.
The SERP technology is most economical and technically
effective on a large volume of moderately contaminated soil.
Soils with low levels of contamination (e.g., less than twice the
cleanup level) would require a high expenditure of time and
energy for a small gain. Extremely high levels of compounds
that form denser-than-water liquid phases might increase the
potential for downward contaminant migration.Small volumes
of soil might be better treated with an ex situ technology. For
very large or stratified sites, separate applications of the tech-
nology could be made, either concurrently or sequentially to
provide better process control.
SERP can effectively treat in both the vadose (unsaturated)
zone and saturated soils. Greater amounts of energy would be
required to heat saturated soil, and water recovery would be
much higher. However, treatment effectiveness is expected to
be as good as or better than that for unsaturated soils.
The technology is capable of treating soils to a significant
depth, up to 100 ft or more. Use of SERP is especially practical
and cost-effective for deep contamination because the wells
permit deep access to the treatment area at low cost. Other
site requirements for use of SERP include utilities (water, fuel,
and electricity), accessible roads for personnel and equipment,
and room to install wells and processing equipment.
The remediation at the Rainbow Disposal site utilized one of
the more unusual aspects of the SERP technology-the claimed
ability to treat soil under and around existing structures. This
site is an active trash transfer facility consisting of a covered
transfer bay, a truck service shop, and a truck wash facility.
Activities at the transfer facility could not be halted during
remediation. The technology was used to treat soil underneath
the structures, as well as around some empty underground
tanks, underground utilities, and other underground and
aboveground obstructions.
Technology Limitations
The SERP technology is not applicable for treatment in non-
permeable or low permeability strata such as rock or thick clay
layers. Fractured rock formations and geological structures
with high permeability "tunnels" within them would cause pref-
erential steam flow and would not allow most areas of soil to
be appropriately treated by the technology.
The less volatile the contaminants, the more difficult they are to
remove from the soil. This condition would lead to longer
treatment times, less complete removal efficiencies, and higher
costs overall for treatment.
Shallow contamination, or very narrow depth intervals of con-
tamination, would not be appropriate to treat due to the diffi-
culty of controlling the steam zone to a narrow range and the
high costs per cubic yard based on the area of the site. Capital
and mobilization costs for the technology are high enough that
only large volumes of soil are economical for treatment.
Performance Data
The SITE Demonstration at the Rainbow Disposal site focused
on the condition of the soil after treatment with SERP. The
primary objective of the Demonstration was to determine whether
the technology could meet the remediation goals for the site.
The California Regional Water Quality Control Board had set a
limit of 1,000 mg/kg of TPH [diesel fraction by the California
Leaking Underground Fuel Tank (LUFT) method] in soil.
Pre-treatment soil data were collected by the technology devel-
oper and the site owner with oversight by EPA. However, this
was not conducted under a SITE Program Quality Assurance
Project Plan (QAPP) and, therefore, these data cannot be
directly compared with the post-treatment data as a critical
objective. These data have been used as an informational
comparison of pre- and post-treatment soil conditions. During
pre-treatment sampling, 12 boreholes were drilled and a total
of 23 soil samples were collected from discrete depths in these
boreholes. The pre-treatment samples were analyzed for TPH
and BTEX. All samples collected before treatment were from
inside the treatment area (perimeter of contamination). Post-
treatment soil data were collected from both inside and outside
the treatment area by EPA according to a site-specific QAPP.
To characterize the site soil after treatment, a total of 72 soil
samples were collected from 24 boreholes. All of the samples
were analyzed for TPH, Total Recoverable Petroleum Hydro-
carbons (TRPH), and BTEX. (TRPH analysis provides informa-
tion similar to TPH analysis, but is performed using an
EPA-approved method. Although the TPH method is widely
used, it is not an EPA-approved method.) Twelve of the bore-
hole locations were selected to correspond with the pre-treat-
ment boreholes (approximately four feet away), and samples
were collected from the same depths in these boreholes as
were collected before treatment to improve comparability be-
tween the pre- and post-treatment data sets. Six of the post-
treatment boreholes were drilled and sampled outside the
-------
treatment area in order to determine whether treatment caused
any lateral migration to occur. The locations for pre- and post-
treatment soil sampling are shown on Figure 2. Pre- and post-
treatment sampling boreholes refer to the same numbered
location.
Table 2 presents the post-treatment soil sampling results for
TPH and TRPH. Based on analysis of the post-treatment TPH
and TRPH data, removal of contamination by the SERP tech-
nology was less complete than expected. Forty-five percent of
the post-treatment soil sample results inside the treatment area
were above the cleanup criterion of 1,000 mg/kg TPH. Seven
percent of samples had TPH levels in excess of 10,000 mg/kg.
BTEX was not detected in any of the post-treatment samples.
The analytical detection limit was 6 ng/kg (ppb). This may be
an indication that the SERP technology was effective in remov-
ing these compounds because BTEX compounds were found
at low mg/kg levels in a few pre-treatment soil samples. How-
ever, this finding is not conclusive. There were no cleanup
criteria established for BTEX compounds in soil.
As a part of the post-treatment demonstration sampling, tripli-
cate samples of soil were collected at pre-determined sampling
locations to evaluate in situ soil contaminant variability. Three
samples of soil were collected vertically from the same 18-in.
split-spoon sampler in separate 3-in. brass sleeves and were
analyzed separately at the laboratory. Results of the analysis
of triplicate samples were highly variable, showing that the site
contamination was heterogeneous, even over small vertical
distances (approximately 3 in.). Table 3 presents the triplicate
sample results and associated statistics.
A geostatistical analysis of the post-treatment soil data was
conducted using a "nearest neighbor" approach on a comput-
erized model to assess the spatial variability of soil contamina-
tion and to determine a weighted average of the soil results.
The use of this geostatistical approach results in the calcula-
tion of a more "unbiased" estimate of the true average level of
contamination for a site, based on a subset of data for all soil
parcels within a site and where there is no pattern of spatial
correlation (i.e., high spatial variability) for soil contamination,
as was the case at the Rainbow Disposal site [3]. Based on the
geostatistical analysis, the post-treatment weighted average
soil TPH concentration is 2,290 mg/kg, with a standard error of
784 mg/kg. Based on an approximate normal distribution for
the weighted average, the 90 percent confidence interval for
TPH concentration is 996 mg/kg to 3,570 mg/kg. This large
interval is because of the heterogeneity of the in situ soil
contamination as seen by the variability of site soil sampling
results; analytical variability was within established quality con-
trol limits and contributed little to overall data variability. Ac-
cording to this analysis, at 90 percent confidence, the true
average may be below the cleanup criterion of 1,000 mg/kg,
but this represents a small probability. Therefore, with almost
90 percent confidence, the average concentration of the site
soil after treatment with SERP is above the cleanup criterion.
The geostatistical analysis results for TRPH yielded a weighted
average concentration of 1,680 mg/kg, with a standard error of
608 mg/kg. The 90 percent confidence interval for the weighted
average for TRPH is 676 to 2,680 mg/kg. The calculated
weighted average and confidence interval for TRPH are lower
in magnitude than for TPH. No TRPH cleanup criteria were set
for the Rainbow Disposal site.
The treatment area was based on a perimeter of contamination
that had been determined using the available site characteriza-
tion data at the time the SERP system was installed. The
perimeter encompassed all soil on the site with TPH levels
higher than 1,000 mg/kg. After treatment, two samples in each
of six boreholes were collected outside of this perimeter in
areas known to be clean or to have levels of contamination
less than 200 mg/kg of TPH, based on site characterization
data, to assess whether the treatment was causing any lateral
migration of the contamination. These samples were analyzed
for TPH, TRPH, and BTEX. Only one of the 12 samples
collected had TPH levels higher than 200 mg/kg, the limit set to
determine whether lateral migration had occurred. Since this
one sample was less than twice the limit and the variability
found in samples from the site was so great, this result is not
felt to indicate that any significant lateral migration had oc-
curred. Additionally, of the remaining perimeter samples, only
two contained greater than 10 mg/kg TPH, and many con-
tained levels less than 5 mg/kg, which would have been unde-
tectable at the detection limit achieved in the original site
survey. TRPH results were similar to TPH results.
A secondary (non-critical) objective of the demonstration was
to determine a removal efficiency (or percent removal) by
comparing pre- and post-treatment sample analysis data. Per-
cent removal was calculated for TPH only, since no pre-treat-
ment TRPH data exist and only three of the pre-treatment
samples contained detectable amounts of BTEX.
Due to the high spatial variability of the in situ soil contamina-
tion at this site, the pre- and post-treatment data sets were
pooled to compare the two. To accomplish this, weighted
average concentrations of TPH in the soil before and after
treatment were compared. A weighted average TPH concen-
tration in the soil before treatment was calculated using
geostatistical modeling (nearest neighbor approach), as was
done for the post-treatment data. The weighted average pre-
treatment concentration was calculated to be 3,790 mg/kg with
a standard error of 2,340 mg/kg. Since the distribution of the
pre-treatment weighted average did not conform to a normal
distribution, the confidence interval on this average was calcu-
lated using a computerized "resampling" technique. This tech-
nique is often used to more accurately estimate confidence
intervals for statistics with non-standard and non-normal distri-
butions [4]. At 90 percent confidence, the calculated interval on
this weighted average is 1,390 to 7,290 mglkg. This large
range is due to the small number of pre-treatment samples
collected and to the variability in the data set.
Comparing the pre-treatment soil TPH weighted average to the
post-treatment soil TPH weighted average, the overall removal
efficiency was calculated to be about 40 percent. Using the
resampling technique to calculate the confidence interval, at 90
percent confidence, the percent removal could be significantly
higher or lower. This large removal efficiency confidence inter-
val is due primarily to the lack of sufficient pre-treatment sample
measurements, and the resultant data set variability. According
to process data, however, it is known that some diesel con-
tamination was removed from the soil during treatment.
The amount of diesel recovered in the storage tank during
treatment was measured and totalled approximately 700 gal at
the end of the project. This is much less than the amount
anticipated when the system was designed and installed, be-
cause a larger percentage of the recovered diesel remained in
the vapor phase and was oxidized in the TOU. This was also
due to inadequate vapor stream condenser design. More die-
sel was removed through the vapor treatment system and
oxidized in the TOU than was collected in the storage tank.
Vapor concentration measurements taken at the inlet of the
-------
Table 2. Stimmdry ol Post-JreattiitintAneifytK.}}! /teu/fs to; Samite* litwtfe tha I'wattnent At an
f< mny
1
1
i
1
1A
tA
(A
1A
V
r
i
j
4
4
,'t
5
fi
6
it
7
/
H
!',*»
;?«
35
,T
^«'>
A?
>'ty
3M
.W
,', 100
»?,./{W
I37i)*
flfff
,79W
'I,'-1
C,80ij'
1.300
11
5,^?CW
7 (700
1,10fl
i,?oo"
03
,'<,,WO
3,400
tit
nPH(ingkg)
(if!
?i',0t)0
1,400
l.lftt/
-L'O
1 1C
5,600
4,t
20
JO
21
?1
21
23
?
3?
,"{(>
37
40
w
,'15
SK
3i*
,7S
,'»
«
41
30
;1S
40
;f/
37
47
20
,'jf)
3fl
TPHftng/kq}
2S
x?4
,'t^
179*
1,400
3.MU
1,900
810
41
350
fan
190
A'.Wtf
j?/J
n.ooo
y, /(/<<
r.aoo
f4
w
d'30
5,S
f, fOO*
440
nmttmg,
4U
<:V
•.•&)
-i*f/
760
,-Pf?
<'/;,•(/
f40
2W
Roy
230
1 ^W
1, 700
,'W
*->S,OQO
!l/0(l
'JhO
160
yfi
S9Q
•*i'Q
ft, too"
//n
" ,4w3f,i|}p ol TnphMto Results
Table 3. Post-Treatment Triplicate Sample Analytical Results
Sorehoto Oeptf> (tfj and
Number Designation
Z 32 primary
duplicafs
triplicate
4 30 primary
duplicate
triplicate
6 40 primary
duplicalB
triplicate
10 30 primary
duplicate
triplicate
13 32 primary
duplicate
triplicate
16 35 primary
duplicate
triplicate
23 30 primary
duplicate
tnptic&te
TPH Results
670
120
310
12,000
240
7,700
590
3,700
860
82
5,6
120
3,3
14
5,9
5.3
9,3
4,5
5, tOO
5,300
8,000
TRPH Results
(mg/kg)
80
160
320
1,400
260
1,100
2500
640
<2Q
<20
130
«;2Q
S?
20
<20
85
<2B
5,800
8,200
Standard
Deviation of TPH
Results (mg/kg)
230
5,000
1,400
47
4.6
2,1
1,300
Standard
Deviation of
TRPH
(mg/kg)
100
2,400
770
>so"
>29*
>30"
190
< Not detected at tfaa detection limit shown.
" Calculated using nan-detect results at the detection limit; actual standard deviations may be slightly higher.
-------
TOU over the course of treatment by the flame ionization
detector (and the LEL meter before the FID was on line), along
with the flow rate and inlet temperature, were used to estimate
the amount of diesel which was removed in vapor form. Based
on these data, it was calculated that approximately 15,400 gal
of diesel were treated by the TOU. Therefore, a combined total
of approximately 16,000 gal of diesel were removed during
treatment with SERP. This volume, compared with the initial
estimate of the amount of fuel released (70,000 to 135,000 gal
[5,6]), less 4,000 gal recovered prior to treatment with SERP,
suggests that between 12 and 24 percent of the original spill
volume was removed from the soil and treated above-ground
by the SERP technology. This removal efficiency is based on
diesel recovered and treated, and although lower than the
removal efficiency based on the soil data, is within the percent
removal confidence interval for the soil data. It should be noted
that vapor stream system measurements were not critical mea-
surements for this demonstration.
Operational problems were encountered during treatment with
the SERP technology. Most of the major problems were due to
malfunctions of major above-ground process equipment, espe-
cially the boilers and the TOU. These malfunctions were partly
the result of process system design deficiencies and intermit-
tent operation.
Due to the innovative and experimental nature of the treatment
process, adjustments were made frequently. Since the system
had been designed for a single year of operation, and opera-
tion occurred over a two-year period, some of the process
equipment had to be replaced including all the downhole piping
for the extraction wells.
It should be noted that this was one of the first full-scale
applications of this technology. Operational problems were,
therefore, partly caused by the learning process common to
any new technology, and probably greater than typically ex-
pected.
Process Residuals
One of the major advantages to an in situ technology is the
reduction in the amount of waste that needs to be disposed of
after treatment. Material that is not removed from the site is not
subject to any land disposal restrictions and can be left in place
at the end of treatment. Residuals from the SERP treatment
process include soil removed from boreholes during well con-
struction or sampling (drill cuttings); recovered contaminant
product; contaminated water; contaminated vapor; sludge from
the oil/water separator; contaminated process equipment, in-
cluding filters and carbon; and disposable equipment and cloth-
ing that has become contaminated through contact with the
waste material.
At the Rainbow Disposal site, the contaminated water was
treated using activated carbon and discharged to an under-
ground sewer. The contaminant-laden vapor from the extrac-
tion wells was treated (oxidized) in the TOU. Therefore, neither
of these process streams required off-site treatment or dis-
posal.
Soil removed from boreholes on the site was placed into 55-gal
drums. Samples from each drum were analyzed for petroleum
hydrocarbons. Drums containing contaminated soil were dis-
posed of off-site at a hazardous waste landfill, while soil from
the drums containing nondetectable levels of petroleum hydro-
carbons was placed back on the site as fill. Approximately 230
drums of drill cuttings, approximately 40 percent of which were
contaminated, were generated during the mobilization, treat-
ment, and post-treatment sampling.
Sludge was collected from the bottom of the oil/water separa-
tor. This sludge was composed of heavier-than-water compo-
nents of the wastewater and residue from the extraction wells.
This material was drummed for off-site disposal as hazardous
waste. Only a few drums of this material were generated and
disposed of, mostly in the later stages of treatment. During
treatment, filters and spent activated carbon were periodically
removed from the system and sent off-site for regeneration
and/or disposal as hazardous waste. At the end of the project,
the recovered diesel was removed from the site for recycling.
Economics
A detailed cost analysis was performed for the SERP
remediation at the Rainbow Disposal site. The total cost for
remediation at the site was approximately $4,400,000. This
value does not include legal fees, which were significant in this
case. The total amount of soil that was considered to have
undergone treatment was 95,000 yd3, which includes the vol-
ume of soil within the treatment perimeter between the depths
of 20 and 40 ft. This yields a cost per cubic yard of approxi-
mately $43.
Since this was the first commercial application of SERP, and
because of the frequency and nature of equipment problems
experienced, the cost calculated above is higher than expected
for other applications. Based on operating logs, a 50 percent
on-line factor was calculated for this process at the Rainbow
Disposal site.
Two additional cost estimates were prepared based on ideal
conditions (i.e., no downtime or a 100 percent on-line factor)
and on conditions expected to be typical of a treatment pro-
cess of this type (using an on-line factor of 75 percent). Total
costs for the ideal and typical cases were estimated to be
$2,800,000 and $3,400,000 for costs of$29/yd3 and $36/yd3,
respectively.
The most significant cost factor in all three cases was the labor
to operate and maintain the process equipment, especially the
boilers. Labor was approximately 30 percent of the total costs
estimated in this economic analysis. Other significant costs
were mobilization, which include designing and installing the
process equipment, and utility costs. Natural gas usage was
the greatest utility cost, and comprised more than 10 percent of
the total costs.
Costs for use of SERP are highly site-specific, since the equip-
ment and operating techniques are tailored to the individual
characteristics of the site. Factors that would tend to increase
treatment time, such as less volatile contamination, larger site
area, or less permeable soils, would have the most significant
effect on the costs for use of the technology.
Technology Status
Technologies similar in concept to in situ SERP have been
investigated on a field-scale level at other contaminated sites.
Most notably, a portion of a site contaminated by gasoline to a
depth of about 135 ft was recently remediated with Dynamic
(Steam) Stripping at the Lawrence Livermore National Labora-
tory in Livermore, CA. The technology was successful in re-
moving and recovering a significant portion of the gasoline
-------
contamination from more permeable unsaturated and satu-
rated soil in the test area. Innovative techniques were applied
to monitor the steam zone and control the process.
At this time, other tests of steam injection technologies are
planned at sites contaminated with dense non-aqueous phase
liquids such as trichloroethene. The ability of the technology to
remediate sites contaminated with these less volatile com-
pounds, without causing downward or off-site migration, will be
key evaluation objectives for these tests.
The SERP technology can be designed, within engineering
constraints, to treat a large area of a site to significant depths.
Based on results from the full-scale application of SERP tech-
nology at the Rainbow Disposal site, the critical factor in scale-
up from pilot- or field-scale to full-scale site remediation seems
to be maintaining control of the in situ process. The poor
results at this site appear to be partly due to inadequate
monitoring capability over the large treatment area (2.3 acres),
resulting in poor process control.
Hughes Environmental Systems operated the SERP technol-
ogy at the Rainbow Disposal site; however, they are not vend-
ing the technology for use at other sites. Since SERP uses
commonly available process equipment, the technology can be
designed and operated by consultants knowledgeable of the
process. Developers of similar in situ steam technologies may
have a patented process or monitoring equipment available
only through them.
SITE Program
In 1980, the U.S. Congress passed the Comprehensive Envi-
ronmental Response, Compensation and Liability Act (CERCLA),
also known as Super-fund, committed to protecting human
health and the environment from uncontrolled hazardous waste
sites. CERCLA was amended by the Super-fund Amendments
and Reauthorization Act (SARA) in 1986 to include amend-
ments that emphasize the achievement of long-term effective-
ness and permanence of remedies at Superfund sites. SARA
mandates implementing permanent solutions and using alter-
nate treatment technologies or resource recovery technologies,
to the maximum extent possible, to clean up hazardous waste
sites.
State and federal agencies, as well as private parties, are now
exploring a growing number of innovative technologies for
treating hazardous wastes. The sites on the National Priorities
List total over 1,700 and comprise a broad spectrum of physi-
cal, chemical, and environmental conditions requiring varying
types of remediation. The EPA has focused on policy, techni-
cal, and informational issues related to exploring and applying
new remediation technologies applicable to Super-fund sites.
One such initiative is EPA's SITE Program, which was estab-
lished to accelerate the development, demonstration, and use
of innovative technologies for site cleanups. EPA SITE Tech-
nology Capsules summarize the latest information available on
selected innovative treatment and site remediation technolo-
gies and related issues. These capsules are designed to help
EPA remedial project managers, EPA on-scene coordinators,
contractors, and other site cleanup managers understand the
types of data and site characteristics needed to effectively
evaluate a technology's applicability for cleaning up Superfund
sites.
Disclaimer
While the technology conclusions presented in this report may
not change, the data have not been reviewed by EPA's Quality
Assurance/Quality Control office.
Source for Further Information
EPA Contact:
U.S. EPA Project Manager
Paul de Percin
U.S. Environmental Protection Agency
National Risk Management Research Laboratory
26 West Martin Luther King Drive
Cincinnati, OH 45268
Telephone No: (513) 569-7797
Fax No: (513)567-7620
E-Mail: dePercin.Paul § EPAMAILEPA.GOV
References
1. Udell, K.S., and L.D. Stewart. Combined Steam Injec-
tion and Vacuum Extraction for Aquifer Cleanup. Key-
note Paper for Conference of the International
Association of Hydrogeologists, Calgary, AB, Canada,
1990.
2. Udell Technologies, Inc. Field Design for In Situ Re-
covery of Hazardous Wastes by Combined Steam
Injection and Vacuum Extraction. Prepared for Naval
Civil Engineering Laboratory. Jan. 1991.
3. Isaaks, E.H., & Srivastava, R.M. An Introduction to
Applied Geostatistics. New York: Oxford University
Press, 1989.
4. Efron, B., & Tibshirani, R. Bootstrap methods for stan-
dard errors, confidence intervals, and other measures
of statistical accuracy. Statistical Science, Volume 1,
1986, pp. 54-76.
5. Converse Environmental Consultants California. Site
Characterization, Conducted for Rainbow Disposal,
17121 Nichols Street, Huntington Beach, CA. Costa
Mesa, CA, Aug. 11, 1988
6. Telephone communication with John Dablow, formerly
of Hydrofluent, Inc. Jan. 26, 1994.
------- |