EPA/600/A-92/224
CONTAMINATED LAND
TREATMENT TECHNOLOGIES
Edited h\
JOHN F. REES
Cehic Technologies Ltd,
Cardiff. UK

-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before compiev 	\
1. REPORT NO. 2.
EPA/600/A-92/224
3.
4. TITLE AND SUBTITLE
CRITICAL EVALUATION OF TREATMENT TECHNOLOGIES WITH
PARTICULAR REFERENCE TO PUMP-AND-TREAT SYSTEMS
5 REPORT DATE
6. PERFORMING ORGANIZATION CODE
7 AUT.HORISI
Stephen G. Schmelling, 2Jack W. Keeley, and
1 Carl G. Enfield
8. PERFORMING ORGANIZATION REPORT NO.
9-, PERFORMING ORGANIZATION NAME ANO ADDRESS
R.S. Kerr Environmental Research Lab., USEPA
P.O. Box 1198, Ada, OK 74820
2Dynamac, Inc., R.S. Kerr Environmental Research Lab,
P.O. Box 1198, Ada, OK 74820
10. PROGRAM ELEMENT NO
TD1Y1A
11 CONTRACT/GRANT NO
In-House
12. SPONSORING AGENCY NAME ANO ADDRESS
Robert S. Kerr Environmental Research Laboratory
U.S. Environmental Research Laboratory
P.O. Box 1198
Ada; OK 74820
13. TYPE OF REPORT AND PERIOD COVERED
Symposium Paper
14 SPONSORING AGENCY COOE
EPA/600/15
15. SUPPLEMENTARY NOTES
Published: ELSEVIER APPLIED SCIENCE, London and New York, for SCI WATER AND ENVIRONMENTAL
GROUP. Proceedings of Contaminated Land Treatment Technologies, pp 220-234.
16. ABSTRACT
W
Ground-water extraction and treatment, or pump and treat, is the most commonly used
technology for remediating contaminated ground water at hazardous waste sites in the United
States. There are major limitations to using this technology for restoration of aquifers
to drinking-water quality in a reasonable time frame. The major limitations to pump-and-
treat technology, which are connected with the difficulty in extracting of contaminants
from the subsurface, can be explained in terms of the basic processes controlling sub-
surface, can be explained in terms of the basic processes controlling subsurface
contaminant transport and fate. The same processes that limit the effectiveness of pump
and treat limit most other aquifer remediation technologies, as well. It is important to
understand and account for these processes when designing aquifer remediation projects.
Research is being carried out by the United States Environmetnal Protection Agency and
other organizations to reduce some of the limitations and improve the efficiency of
pump-and-treat
\
17. KEY WORDS AND DOCUMENT ANALYSIS
i. DESCRIPTORS
b.IDENTIFIERS/OPEN ENOED TERMS
c. COSATI Field,Croup
Pump-and-Treat
Ground-water remediation
Ground Water
Subsurface Processes.


18. DISTRI flUTION STATEMENT
19. SECURITY CLASS i This Report)
UNCLASSIFIED
21 NO- op PAGES
19
20 SECURITY CLASS (This pjfc '
UNCLASSIFIED
22 PRICE
tPA Form 2220-1 (R«». 4-77) phcviou* edition is obsolete

-------
IX
Contents
fnrp\n:.,r,in,iii.»n	I •
K J. Potter
On Site and In Situ Treatment or Contaminated Sites	 30
B Ellis
Containment and Remediation or Contaminated Sites by Extraction of
Vapour or Groundwater	 47
G A. M van Mews
Remedial Barriers and Containment	 .	58
S A.JeiU't m
Sotsen: Extraction for the Treatment of Contaminated Soil . . .	S3
H. Eccies and C P Hotroy d
Theory and Practice of Clean Cover Reclamation	 9?
T. Cairncy
Principles and Application of Physical Particle Separation for Treatment of
Contaminated Land .	.. . 		113
K P. Williams
Treatment of Acidic Industrial Lai:	. . . .	129
W. Davison
Vitnfication of Contaminated Soils	143
K R. McSeiil and S. Marine

-------
X
Organic Stabilization/Solidification: Theon and Practice	160
R Soundararajan
The Role of Planning Authorities in Influencing Design of Land Treatment 180
R. Turner
A Thermal Method for Cleaning Contaminated Soil 	195
K. Bohm
Critical Evaluation of Treatment Technologies with Particular Reference to
Pump-and-Trca; S} stems	 . . 	220
S. G Schmelling, J W A'eeley and C. G. Enfield
Experience Acquired with the Oecotec High-Pressure Soil Washing Plant
2000 in Cleaning Contaminated Soil	235
H.-J. Heimhard
Land Reclamation and Rede\e!cpmen: ,r. Blekholn^torget	252
H. Ki unberg
Bioremediation of Contaminated Ground	270
R J F Bewtey
Technology to Application in Recycling Contaminated Land	285
S, T. Johnson and F. M. Jar dine
Index of Contributors	309
t^»v-

-------
220
CRITICAL EVALUATION OF TREATMENT TECHNOLOGIES WITH
PARTICULAR REFERENCE TO PUMP-AND-TREAT SYSTEMS
STEPHEN G, SCHMELLING", JACK W. KEELEY2, AND CARL G. ENFIELD)
1) United States Environmental Protection Agency, 2) Dynamac, Inc.
Robert S, Kerr Environmental Research Laboratory
P. O. Box 1198, Ada, Oklahoma, 74g20, USA
ABSTRACT
Ground-water extraction and treatment, or pump and treat, is the most commonly used
technology for remediating contaminated ground water at hazardous waste sites in the United
States. There are major limitations to using this technology for restoration of aquifers to
drinking-water quality in a reasonable time frame. The major limitations to pump-and-treat
technology, which are connected with the difficulty in extracting of contaminants from the
subsurface, can be explained in terms of the basic processes controlling subsurface
contaminant transport and fate. The same processes that limit the effectiveness of pump and
treat limit most other aquifer remediation technologies, as well. It is important to understand
and account for these processes when designing aquifer remediation projects. Research is
being carried out by the United States Environmental Protection Agency and other
organizations to reduce some of the limitations and improve the efficiency of pump-and-treat.
INTRODUCTION
Slightly over a dccuvi- there were relative!) fov> -c:i\e efforts to rectify known cases
of ground-water contamination. Indeed, it was only at this rime that early efforts were being
made to estimate the extent and magnitude of ground-water contamination problems. By now,
the number of hazardous waste sites on the United States Environmental Protection Agency
(EPA) National Priorities List (NPL) exceeds 1,000, and estimates have been made that the
number could grow to 2,000 The United States Congressional Office of Technology
Assessment (OTA) estimated that the list could reach 10,000, requiring remediation activities
well into the 21st century (1). The most common area of concern at sites on the EPA
Superfund list is ground-water contamination [2],
During the early days of ground-water remediation, pump-and-treat systems were the
leading, if not the only, technology available. Pump and treat was based on a common-sense
idea. If, as a result of aqueous samples from monitoring wells, the ground water was known
to be contaminated, all that had to be done was to pump the contaminated water from the

-------
221
aquifer and treat it However, the results of the typical pump-arid-treat remediation project
were often discouraging. The system would initially remove a large mass of contaminant with
a concomitant decrease in the concentration of the contaminant However, after a relatively
shot time, the concentration of the contaminant leveled off at a value well above the design
goal of the system. The system kept pumping and treating, but the contamination lingered,
even with continued pumping. It was much like Lady Macbeth trying to wash the blood from
her hands - "Out, damned spot! Out, I say!" [3]. Pumping at a higher rate caused the
concentration of the contaminant in the extracted water to diminish, but, if the pumps were
turned off far some period of time, the water being brought to the surface would once again
have increased levels of contaminants.
As a result of such field experience, and concurrent laboratory studies, scientists and
engineers began to question the effectiveness and efficiency of such systems. During the last
several years They also began to examine the causes of the problems and to look for ways to
improve the situation.
The major difficulties with pump and treat are related to the "pump", or extraction,
portion of the process. There are fundamental reasons v. hv it is difficult to extract
contaminants from the subsurface. These fundamental reasons, which arc discussed in detail
below, also limit the efficiency of many other aquifer remediation technologies. Surface
treatment of the extracted water, the "treat" portion of pump and treat, has its own set of
engineering problems, but in general, surface treatment is a much more mature technology and
is not the limiting factor in the successful use of pump and treat
The purpose of this paper is to briefly discuss: (1) some of the reasons why the removal
of ground water for remediation does not always lead to the desired results, (2) the contaminant-
transport processes responsible for this behavior, and (3) possible schemes to help overcome
some of these problems. There is also a brief discussion of ground-water extraction systems in
conren with other technologies.
Keely [4] identified four factors which affect the efficiency by which ground-water
extraction removes contaminants from the saturated subsurface. These are:
1)	diffusion of contaminants into low permeability sediments,
2)	hydrodynamic isolation (dead spots) within well fields,
3)	desorption of contaminants from sediment surfaces, and
4)	liquid-liquid partitioning of an immiscible contaminant as a result of the presence of a
separate non-aqueous phase liquid (NAPL)
The first two factors might be combined under the general heading of hydraulic inefficiencies.
Equally important to the success of any aquifer remediation is the need to adequately
characterize the site and the contaminant characteristics. Without an adequate site
characterization it will be highly unlikely that the remedial action will be well designed, and it
will be almost impossible to determine whether or not it is actually cleaning up the site.

-------
222
In order to better understand the influence of these factors in pump-and-treat systems, it
may be helpful to define a simple hydrologic scenario against which each can be tested This
scenario, depicted in Figure 1, consists of a 4 hectare area of ground-water contamination with
a saturated thickness of I? meters and an effective porosity of 30 percent of the aquifer
volume. The example includes the assumptions that the aquifer has sufficient hydraulic
conductivity to allow pumping at any reasonable rate without adjusting the hydraulic gradient,
the aquifer boundaries are constant, and there is an infinite source of water.
226 m
Extraction Wei!
Land Surface
-Water Table
Aquifer
Impermeable
Aquttard
Figure 1. Idealized remediation scsr.a:
Under these ideal conditions, and with an ideaJ extracuon well system, it ought to be
possible, pumping at 400 L/min, to exchange the water in this 4-hectare plume in about one
year and remove the contamination from a soluble contaminant (e g chloride) at low
concentration. In reality, however, it may be necessary to pump for two or three or more years
to reach an acceptable contaminant concentration due to the "tailing" effect often observed
under even these simple conditions.
HYDRAULIC INEFFICIENCIES
As shown in Figure 2, tailing is the slow, almost asymptotic decrease in contaminant
concentration as contaminated water is extracted from the aquifer and fresh water takes its
place. When compared to the theoretical or more ideal concepts of removal, tailing can require

-------
223
significantly longer pumping times even when dealing with the simple conditions outlined in
this scenario.
c e
© o
a a
<_ u
C B
If
UU
> =
X. at
— 3
-E
W
Removal with Tailing
Theoretical Removal
Water Filled Aquifer Volumes
Figure 2. Example of tailing dur;n^ pi.n:p-and-irca: icniccianor.
Subsurface heterogeneity, as shown in Figure 3, is one cause of tailing. In the simplest
case of a highly soluble contaminant such as chloride, tailing results when the contaminants
Sand .
KM.*..*
' Clav
Sand
STlt
Vertical Section
Through Aquifer
Average
Velocity
Velocity/Hydraulic
Conductivity
Convection
DQImsIojl
Convection
Diffusion
Convection
Diffusion and
Convection
"OflTuMon"
Dominant
Flow Process
Figure 3. Example of subsurface heterogeneity
diffuse into low permeability sediments over a long period of contamination and slowly diffuse
back out during a pump-and-treat remediation. Under a uniform hydraulic gradient, the velocity
of water varies directly with the hydraulic conductivity of individual layers, and field
measurements have demonstrated that, even in relatively homogeneous aquifers, the hydraulic

-------
224
conductivity can vary by a factor of 10, or more, over distances on the order of tens of
centimeters. When water is pumped through the aquifer, contaminants in the regions of high
permeability are removed relatively efficiently. Contaminants in the zones of low permeability
are only slowly removed by diffusion to the bulk of the flowing water in the zones of high
permeability or by advection at much slower velocities through the zones of low permeability.
A rule of thumb suggests that the longer a site has been contaminated and the more layered the
geologic material, the longer the effects of tailing will be present. Even when these zones of
low permeability are stressed by pumping wells, the time required to purge contaminants is
likely to be extensive.
A simple example can demonstrate the effect of the variability of hydraulic conductivity
oo the removal of a soluble contaminant, such as chloride. The example assumes an aquifer
with two equally thick layers — one layer having 10 times the hydraulic conductivity of the
other, and both having an effective porosity of 30 percent If it is assumed that the initial
chloride concentration is the same in both layers, and that there is no diffusion between the
layers, the time of remediation usinc	free aojrd water c.m be estimated.
Figure 4a shows a curve of relative concentration (concentration relative to the initial
concentration) at the extraction point versus time for an assumed flow rate. The relative
"n
n s.
1
0.8


e
0.6


0
c
0.4-
a 0.2 ¦:
0.0-
10
Time
15
20
Figure 4a. Effect of variability of hydraulic conductivity on removal of a miscible contaminant
concentration remains at unity for a short period of time then drops off and reestablishes a new
concentration where it remains for a considerable period of time. The shape of the curve is
typical of field observations and has also been demonstrated in the laboratory.
If the pumping rate is increased by 50 percent (Figure 4b), the concentration drops off
more rapidly and stabilizes at the same plateau. The net result of the higher pumping rate is to
reduce the overall time required for remediation while the plateau concentration is dependent on
the relative hydraulic conductivities of the sediment layers and the portion each of the layers
contributes to flow through the system However, for a surficial aquifer.increasing the

-------
225
pumping rate will change the shape of the free water surface, often isolating 8 portion of the
contamination from the flow path. When the pumps are turned off and the aquifer is allowed to
stabilize, the concentration of contaminants in the ground water may return to the original level.
1.0-
! 0J-
s
1 0.6
I
X 0.4
"5
" 0.2
0.0
10
Time
IS
20
Figure 4b. Effect of variability <•;' hydraulic conductivity at higher flow- rate
Zones of hydrodvnamic isolation (dead spots) within well fields both down gradient of
extraction wells and up gradient of injection wells [4], are an inevitable consequence of the
natural hydraulic gradient and the gradients created by the pumping and injection wells. Like
zones of low permeability, the movement of contaminants from these zones occurs primarily
by diffusion, and the removal of contaminants from these zones will be inefficient during
remediation using pump-and-treat systems
SORPTION
As discussed earlier, highly soluble contaminants such as chloride are dissolved in the
liquid phase and, at low concentrations, are essentially transported along with the moving
ground water in most soils. Most other contaminants, however, tend to partition, or be
distributed, between the liquid, solid, and vapor phases which comprise the subsurface matrix.
In the ground-water zone of the subsurface, the relationship between the concentration of the
contaminant in solution and the mass sorbed on the aquifer solids depends on the chemical
characteristics of the contaminant, the chemical properties of the ground water, and the
properties of the geologic matrix. In many cases, the relationship between the mass of
contaminant sorbed and the concentration in solution is approximately linear. In such cases the
extent of sorption can be describe in terms of a partition coefficient which is the ratio between
the amount on the solid phase and the amount in the liquid phase. As a general rule, highly
hydrophobic organic compounds are much more strongly sorbed than more soluble organic

-------
226
compounds. Sorption of metals and other ionizable contaminants is a more complex
phenomenon and is dependent on pH and other system parameters that may be highly variable
[5].
Sorption must be accounted for when estimating the contaminant mass to be removed.
Because of sorption, the mass of the contaminant adsorbed onto die aquifer solids may be as
great or greater than the mass in solution. When there is significant sorption, estimates of
contaminant mass based only on water samples will seriously underestimate the mass 10 be
removed. When dealing with highly insoluble compounds like DDT, PCBs, or dioxin, almost
all of the material will normally be associated with the solid phase and very little will be
associated with the water phase. The best way to estimate the mass of sorbed contaminants is
by collecting cores of the aquifer material and measuring a partition coefficient
Sorption reduces the efficiency of pump and treat. Sorbed contaminants can only be
extracted if they are desorbed and in solution. When contaminants are strongly sorbed. only 3
small fraction of the total contaminant mass is removed with each pore volume pumped.
Alternatively, the average velocity of the sorbed contaminant may be viewed as retarded relative
to the average velocity of the ground water or highly soluble contaminants. In either case,
additional pore volumes will have to be pumped to remove the sorbed contaminants. The
increase can be from a factor of slighdy greater than one to 10 or more. This can increase the
time required to remediate the aquifer from a few years to tens of years.
Research has shown that the release of many contaminants from the solid phase can be
exceedingly slow [6], This slow desorptive release acts much like diffusion from zone of low
permeability. Acting together, these two processes greatly accentuate the tailing effect.
Depending upon the nature of the compounds, uie slow, deborption o! contaminants from
sediment surfaces often results in a profound tailing effect when attempting to remediate
s
1.0
'•5
0.8-
L.
c

ij
'w
0.6
C

w
0.4


c5
0.2



0.0 J
0	5	10	15	20 400 500 600
Time
Figure 5. The effect of sorpnon. in addition to subsurface heterogeneity, or the removal of u
soluble contaminant

-------
227
ground-water by pumping. The more a contaminant preferentially associates with the solid
phase, the more difficult that contaminant is to remove.
The partitioning of a contaminant between water and aquifer material is commonly
proportional to the amount of organic material present in die aquifer sediments. Frequently
there is more organic carbon and therefore greater partitioning to the finer sediments. If the
example discussed in Figure 4a were altered to include a partition coefficient of unity in the
layer of high hydraulic conductivity (the amount associated with the solid phase is equal to that
associated with the liquid phase), and a partition coefficient of 5 in the layer of low hydraulic
conductivity, it would take longer before the initial decline in relative concentration (Figure 5)
and much longer before complete remediation could be realized The relative concentration ai
the plateau remains the same as in the example of Figure 4a. It can be seen thai the magnitude
of this plateau or oil is independent of the partition coefficients of the individual layers and thai
the effect of partitioning is to increase the amount of time required to achieve remediation goals
IMMISCIBLE FLUID PHASES
When an immiscible fluid, or non-aqueous phase liquid (NAPL) is released into the
unsaturated zone a fraction of it will volatilize and be released to the atmosphere, and the
remainder will start to move downward toward the water table under the influence of the force
of gravity as shown in Figure 6. If the original source of contamination is removed, capillary
forces will immobilize part of the separate phase liquid as discontinuous blobs trapped within
the pore spaces as shown in Figure 7. This immobile material, which can occupy from five to
Leaking Underground Storage Tank
unsaturated Zone
Vapor Phase
.//>Vapor Phase
*1 V I
Capillary Fringe
Continuous Fluid
^Dissolved Contaminants
Water Table
Saturated Zone
Figure 6 Release of a NAPL from an underground storage tank

-------
228
40 percent of the pore space [6], is referred to as residual saturation" or "residual phase" and
cannot be removed in any substantial quantity by pumping.
	-			6-1
Unsaturated Zone
Vapor Phase

Residual Saturation;
Dissolved
ontaminants
,W.gmg-MjSEfc
Figure 7. Residual saturation remaining after removal of original source of contamination'
Gasoline is an example of an immiscible fluid with a specific gravity less than one,
commonly referred to as a light non-aqueous phase liquid or (LNAPL). Because they are less
dense than water, LNAPLs remain near the water table. The soluble components of the
LNAPL will partition into the moving ground water and contaminate it This partitioning
appears to be greatly enhanced if the water table fluctuates. LNAPLS, primarily gasoline from
leaking underground storage tanks, are a leading cause of ground water contamination in the
United States.
Of even greater concern, from a remediation standpoint, are dense nonaqueous phase
liquids (DNAPLs) such as coal Lars or chlonnated solvents which have specific gravities
greater than one. If enough DNAPL volume is released to the subsurface to overcome capillars
forces, the DNAPL will continue to migrate vertically through the saturated zone until it
encounters a relatively impermeable layer where it may form a perched DNAPL pool. In
addition to accumulating in DNAPL pools, residual DNAPL will be trapped in pore spaces
within the saturated zone as well as in the vadose zone, dissolve into passing ground water,
and be transported in the direction of ground-water flow.
The rate of contaminant partitioning from NAPLs into ground water, and the eventual
concentration reached is dependent on the characteristics of the contaminant and the location of
the residual phase with respect to the flowing ground water. If the residual phase is a complex
fluid such as gasoline, the rate of contaminant partitioning from the NAPL into water will be
different for the individual constituents of the non-aqueous fluid As a result the composition
of the non-aqueous fluid will change with time

-------
229
It is apparent that if remediation is dependent on allowing the DNAPL to dissolve into
ground water and pumping the aqueous phase contamination to the surface, the remediation
process will require a very, very long time. To remove one liter of a common and relatively
soluble contaminant such as trichloroethyfcne (TCE) from the ground water would require
pumping 1,400 liters of ground water contaminated at 1,100 mg/1, the aqueous solubility level
of TCE.
If the immiscible phase can be located it should, as a rule, be pumped directly and
removed. However, the relative importance of forces that control the rate, direction, and
ultimate fate of DNAPLs is different than for the distribution of dissolved phase plumes.
DNAPL movement is more strongly controlled by both large and small scale geologic
heterogeneities than by the movement of ground water. As a result, the distribution and
movement of DNAPLs is difficult to determine, even at sites of relatively homogeneous
geology with a well understood DNAPL source. Even if the free-phase DNAPL can be located
and pumped, residual saturation will remain and continue to contaminate ground water as the
contaminants partition into it. The kinetics of this partitioning, whether from free product
DNAPL or from residual saturation. C3n prove deceptive when attempting io remove
contaminants using pump and treat.
The contaminant cannot be removed faster than it is released from residual saturation or
pools of immiscible fluids, or than it can diffuse from regions of immobile water into the
passing ground water. The result is similar to that described earlier for the kinetics of diffusion
from zones of low permeability or desorption. The concentration may initially appear to be
reduced, or even eliminated, by dilution when bringing larger amounts of u neon laminated
water into play, or by dropping the water table below the source of contamination, or both.
However, if the pumps are stopped for a period of time, the contaminants will again partition
or diffuse into the moving ground water and their concentrations will return to their previous
leveis. Without an understanding of iiiciC processes pamp-and-rre3t system1; will probably be
poorly designed and will likely contaminate more fresh water than would be the case if no
pump-and-treat remediation were attempted.
Nonetheless, pump and treat can be of considerable use in reducing the extent of the
dissolved phase plume, and this action can be of considerable value in the overall plan of site
remediation. Control of the aqueous phase plume by pumping can prevent further
contamination of down-gradient ground water. It can also reduce the mass of contaminant to
be dealt with in further remedial activities. It should also be noted that there is a great deal of
research underway to enhance the effectiveness of pump and treat in dealing with DNAPLs.
Some insight into the effect of N'APLs on remediation time can be gained by returning to
the scenario developed for the removal of chloride in Figure 1. If/axher than chloride, the
constituent is toluene dissolved in the residual saturation of a gasoline plume, it would be
necessary to pump at a rate of 400 L/min for about 1,500 years to reduce the initial amount of

-------
230
toluene by 80 percent if the residual saturation amounted to only 10 percent of the aquifer voids
and no other processes of transformation were occurring. Even if the configuration of the
plume and the hydrology of the aquifer would allow increasing the pumping rate to 4,000
L/min without de watering the contaminant plume, it would still require 150 years to reduce
toluene by 80 percent Almost certainly a system such as this would result in the contamination
of a great deal of fresh ground water by moving it through the area of contamination. Even if
the aquifer would allow unlimited pumping, it is important to realize that there is a point
beyond which the removal of contaminants would not increase because of the limitations
imposed by partitioning and diffusion rates.
IMPROVING THE SITUATION
The acknowledged problems with pump and treat, along with the recognition that it is
often the only available aquifer remediation technology for many situations, have led to
considerable discussion and research on ways to make it more effective. Moreover, since
many of the problems of pump and treat also linut ine efieaiveness of other aquifer remediation
technologies, steps taken to improve pump and treat will have benefits for other technologies as
well. There are steps that can be taken to use existing knowledge to improve pump and treat,
and steps for which research and new knowledge are required.
Site Characterization
There is general agreement that one of the most important parts of any pump-and-treat
remedial action is a good site characterization. The subsurface can be an extremely complex
environment whose characteristics change dramatically over small horizontal and vertical
distances. The ability to design remediation systems and make an estimate of their
effectiveness is proportional to the amour: of information available about a number of factors
including:
1)	the location and distribution of contaminants,
2)	the identity and quantitative values of hydrogeologic parameters, flow paths, and
other influences such as pumping wells and streams,
3)	quantitative measurements of contaminant partitioning between liquid, solid, vapor,
and NAPL phases,
4)	the effect of remediation activities on flow paths and interactions between
contaminants and subsurface solids.
Obviously, the more complex a site, the more intense the effort required to obtain these
types of information. In some cases the technology required to obtain the information is
available, in other cases it is theoretically available but its use is not routine, even for relatively
homogeneous sites, and in other cases the technology is still the subject of research. Among
the tools which are available and being used by experienced ground-water scientists are depth-
specific clusters of monitoring wells which can be used to locate areas of water miscible

-------
231
contaminants and can sometimes be used to suggest the locations of immiscible contaminants
Extensive core sampling can (a) assist in the determination of permeability distributions and
therefore predominant flow paths; (b) provide information concerning the location,
distribution, and total amount of contaminants; and (c) allow estimates to be made of
contaminant partitioning between the liquid, solid, vapor, and immiscible phases. Standard
hydrologic tests are required to provide information on the hydrogeology of the system, while
geophysical techniques can sometimes be used to reduce the number of samples required
It is not always apparent where to locate sampling points or how many points will be
required to adequately characterize a site. As the geology and the parameters discussed above
become more complex, the more sample points are required for characterization, and the more
difficult the interpretation of information becomes. As the hydrology of a site and the nature of
the contaminants becomes more complex, the confidence in any remediation technology is
decreased. In fractured and karst media, which underlie many hazardous waste sites, the
technology available to characterize water and contaminant movement is largely undeveloped.
Slurry Walls
The use of slurry' of cut-off u alk in conjunction with pump and treat can be used to
improve the effectiveness of pump and treat. A slurry wall placed in from of an advancing
plume can greatly reduce the amount of water extracted and requiring treatment This will also
reduce the amount of uncontaminaied water that would otherwise become contaminated using
an extraction system alone. If a slurry wall surrounds a contaminant source and plume, ground-
water extraction could maintain a negative head at the site. The slurry wall will reduce the
amount of fresh water being contaminated by the remediation, and reduce the amount of water
requiring treatment to the leakage rate of the barrier wall, plus that due to infiltration and any
water applied at the site to accelerate the remediation.
Research
The US EPA and other orgar.i;ui:ions x-e conducting research to improve the
effectiveness of pump and treat. One effort is the EPA's Subsurface Cleanup and Mobilization
Processes (SCAMP) research program which is looking at ways to enhance the effectiveness
of pump and treat and for sites contaminated by DNAPLs. SCAMP research is focusing on
two major applied research areas: (1) improved site characterization, and development and (2)
evaluation of means to enhance the effectiveness of pump and treat SCAMP is also funding
some of the necessary basic research on the fundamental transport and fate processes needed to
support the two applied areas.
Research on site characterization that is being carried out under SCAMP is presently
composed of two activities. The first is to develop a manual for use by practitioners for
characterizing sites contaminated by DNAPLs. This document is expected to: (a)summarize the
current state-of-the-art for characterizing sites suspected of DNAPL contamination; (b)
summarize likely near-term improvements; and guide further EPA research on site

-------
232
characterization. SCAMP research is also investigating methods which combine field
measurements and modeling to characterize fractured rock sites, keeping in mind a realistic
estimate of the resources available for site characterization at a typical hazardous waste site.
Future work will be aimed at direct methods for the detection and quantification of DNAPLs
and measurement of multiphase parameters such as relative permeabilities.
SCAMP research on the development and evaluation of technologies to improve pump
and treat is aimed at overcoming the problems identified earlier A modeling study, using the
opportunity for more realistic simulations offered by supercomputers, is investigating the
benefits to be expected from novel pumping schemes. These include vertical pumping to force
water through layers of low permeability, and pulse pumping to potentially reduce the amount
of water to be treated. In theory pulse pumping can be useful in breaking up zones of
stagnation dismissed earlier and changing the direction of flow to improve the efficiency of
removal when contaminants must enter ground water through slow partitioning or diffusion.
However, there has been little field uo-k or other evidence to evaluate the effects of pulse
pumping.
Other work under the SCAMP research initiative is investigating the use of chemical
additives in an attempt to improve pump and treat. Chemical additives such as surfactants and
solvents are among the most promising short-term approaches for enhancing the effectiveness
of pump and treat. Surfactants have the potential to improve the effectiveness of pump and
treat in two ways. The first is by increasing the solubility of hydrophobic contaminants in
ground water. Increasing the solubility will generally reduce the extent of sorption to the
aquifer solids, and has the potential to reduce the number of pore volumes, and time, required
to remove sorbed contaminants The effect of surfactants on sorption kinetics is less clear and
is pan of the research. Increasing th. v/.-b;li:y coj!J also increp.se the extern of dissolution of
residual or free-phase NAPL. The second way ihat the use of surfactants can potentially
enhance the effectiveness of pump and treat for NAPLs is by reducing the interfacial tension
between the NAPL and water, making it possible to mobilize the residual saturation.
SCAMP research on the use of surfactants for increasing the effectiveness of pump and
treat is being carried out at the laboratory-bench scale and through the use of a large physical
model aquifer at RSKERL. The large physical model will be used to simulate an enhanced
pump-and-treat scenario to remediate a DNAPL spill. In addition to investigating the
effectiveness of surfactants for enhancing pump and treat, the research is also paying close
attention to the characteristics of surfactants that will make them acceptable to the public for
injection into potential or actual sources of public drinking water.
Extensive laboratory work at RSKERL, and other institutions, has shown that miscible
solvents such as ethanol can increase the solubility of hydrophobic contaminants in water [9].
Ethanol has relatively low toxicity and probably would be acceptable for addition to the
subsurface Work is currently underlay on a small field project to test the effectiveness of

-------
233
ethanol in improving the efficiency of extracting a mixture of aviation gasoline and
tetrachlorethane (PCE) from a shallow aquifer in the state of Michigan.
Additional research, sponsored by the EPA and other US federal agencies, is
concentrating on the use of physical agents such as steam or hot water to increase the
effectiveness of DNAFL removal. This technology, like the surfactant and solvent work, is an
attempt to adapt techniques used for enhanced oil recovery to the remediation of hazardous
waste sites. Differences in the objectives and conditions under which the work is carried out
ma Iff this adaptation doo-trivial.
It is often possible, indeed desirable, to use pump-and-treat remediation systems in
concert with other technologies. For example, to be effective, pump and treat of contaminated
ground water must also be combined with efforts to remove contaminants from the vadose
zone. Bioremediaoon of contaminated ground water may be thought of as a pump-and-treat
system in which ground water is extracted, supplemented with necessary nutrients to stimulate
biodegradanoo, and reinjected into the aquifer. Efforts to improve pump and treat, particularly
with regard to site characterization and subsurface fluid movement, will almost certainly benefit
bioremediation as well
CONCLUSION
It is important to understand the processes that limit the effectiveness of pump-and-treat
technology in order to develop more efficient and effective remediation projects. Research has
the potential to improve the technology available to characterize the system, to control the
movement of fluids in the subsurface, and to influence interactions between the various
contaminant phases and the subsurface matrix. It is also important to have realistic
expectations about what can, or cannot, be accomplished with pump and treat. With research
and additional experience in building and operating pump-and-treat systems, the situation will
improve, but it will be many yearb. nowever, beiore pump-and-L'cui technology is at a le\u'.
where effective and efficient remediauon systems can be routinely designed and implement for
even moderately complex subsurface problems.
REFERENCES
1.	Train, R.E., Big question facing the clean up of ground water. EPA Journal. 1987, 13(1),
8 - 11.
2.	Ward, B., Groundwater - Another trend, fad ... or a legitimate Washington issue?
Pollution Engineering. 1983,15(10), 14-16
3.
Shakespeare, W„ Macbeth. Act V, Sc. I, line 30

-------
234
4.	Keely, J., Performance evaluation of pump-and-treat remediations USEPA/540/4-89-005,
1989, 19 pp.
5.	Weber, W.J., Jr. McGinley, P.M., and Katz, L.E., Sorption processes and their effects on
contaminant fate and transport in subsurface systems, Water Research. 1991,25,499-
528.
6.	Brusseau, M.L. and Rao, P.S.C , Sorption nonideality during organic contaminant
transport in porous media. Critical Reviews in Environmental Control. 1989.19, 33-99.
7.	Wilson, J.L., Conrad, SJi, Mason, W.R., Peplinski, W„ and Hagan, E., Laboratory
investigation of residual liquid organics. U SEP A/600/6-90/004,1990,267 pp.
8.	Robert S. Ken- Environmental Research Laboratory, Dense nonaqueous phase liquid? -- a
workshop summary, Dallas Texas. April 16-18. 1991, (10 be published 1992),
US EPA/600/, 81 pp.
9. Rao, P.S. C, Lee, L.S., and Wood, A.L., Solubility, sorption, and transport of
hydrophobic organic chemicals in complex mixtures, EPA/60CVM-91/16,1991.

-------