United States
                            Environmental Protection
                        National Risk Management
                        Research Laboratory
                        Ada, OK 74820
                            Research and Development    EPA/600/S-96/002
                                                       August  1996
               Surfactant-Enhanced DNAPL Remediation:  Surfactant Selection,
                           Hydraulic Efficiency, and Economic Factors

                      David A. Sabatini1-3, Robert C. Knoxu, and Jeffrey H. Harwell2-3

Chlorinated hydrocarbons are ubiquitous ground water
contaminants due to their widespread use as organic
solvents  and cleaners/degreasers.  The immiscibility of
chlorinated organics with ground water causes them to
exist as nonaqueous phase liquids (NAPLs); this results in
their occurrence in the subsurface as residual  and free
phases. Having a density greater than water, they are often
referred to as dense nonaqueous phase liquids (DNAPLs).
Water solubilities of these chlorinated  hydrocarbons are
frequently several orders of magnitude above their drinking
water standards, yet low enough to limit dissolution during
pump-and-treat remediation. Remediation of residual DNAPL
contamination can require hundreds to thousands of
flushings of the ground  water (pore volumes) using
conventional pump-and-treat methods.  Strongly sorbing
(hydrophobic) compounds will experience a similar fate
(e.g., PAHs, PCBs). The inefficiency of conventional pump-
and-treat methods for these contaminants has recently
been addressed, with surfactants being mentioned as a
promising technology for enhancing conventional
approaches (Haley, et  al. 1991; Palmer and Fish, 1992).

Two obstacles to widespread implementation of surfactant-
enhanced subsurface remediation are (1) gaining regulatory
1School ofCivil Engineering and Environmental Science
2School of Chemical Engineering and Materials Science
3The Institute for Applied Surfactant Research
The University of Oklahoma, Norman, OK 73019
                     approval for the injection  of surfactants, and  (2) the
                     economics as impacted by  surfactant costs (suggesting
                     minimization of hydraulic and physicochemical surfactant
                     losses). The objective of this brief is to present research
                     results addressing these two issues, as further delineated
                     below. Of course, proper surfactant selection is imperative
                     to successful enhancement  of DNAPL extraction; factors
                     affecting surfactant selection arediscussed inthisdocument.
                     Gaining regulatory approval  is an obstacle common to all
                     chemical amendments being considered for subsurface
                     remediation.   Surfactants with U.S. Food and Drug
                     Administration direct food additive status and which are
                     commonly used in consumer products are the focus of this
                     research.  The economics of surfactant-enhanced
                     remediation processes will potentially  be limited by
                     surfactant losses in the subsurface, including hydraulic
                     losses to uncontaminated  portions of the aquifer and
                     physicochemical losses (e.g., sorption, precipitation).
                     Hydraulic approaches for minimizing surfactant losses are
                     presented, including discussion of vertical circulation wells.
                     Finally, results of a preliminary economic analysis for
                     surfactant-enhanced DNAPL remediation are presented,
                     along with suggestions for optimizing the economics of this


                     Surfactants (surface-active-agents) accumulate at surfaces or
                     interfaces, a  result of the dual nature of surfactant molecules.
                     Surfactant molecules have hydrophobic (water disliking) and
                     hydrophilic (water liking) moieties.  Surfactants are commonly
                     utilized in detergents and food products to alter the surface

chemistry of the system in a desirable manner.  Above a certain
concentration surfactant molecules self-assemble into aqueous-
phase spherical aggregates with the hydrophobic portions of the
molecule  in the interior of the  aggregate and the hydrophilic
portions at the exterior. This aggregate is referred to as a micelle
and the surfactant concentration above which  micelles form is
referred to as the critical micelle concentration (CMC).

Surfactants can improve subsurface remediation by solubilization
and mobilization. Solubilization significantly increasesthe aqueous
concentration of the contaminant via micellar partitioning, thereby
reducing the number of pore volumes which must  be pumped to
extract the DNAPL.   Optimal  mobilization utilizes ultra-low
interfacial tensions (evidenced  concomitant with formation of
middle-phase microemulsions) to significantly reducethe capillary
forces trapping the DNAPL, thereby allowing the oil to be readily
extracted  with the water.

Solubilization enhancementfor neutral organic compounds results
from the partitioning of the contaminant into the hydrophobic
core  (oil-like center) of the  micelle.  Two  parameters that
describe this process are the molar solubilization ratio (MSR) and
the micellar-water partition coefficient (KJ.  Given a graph of
aqueous  contaminant  concentration  versus surfactant
concentration, the molar solubilization ratio (moles of contaminant
per mole of surfactant) is the slope of the straight line portion of
the plot above the CMC.  The  micellar-water partition coefficient
(K  ) is the molar ratio of the contaminant in the micellar phase
                  divided by the molar ratio of the contaminant in the aqueous
                  phase, andean be determined given values of MSR, contaminant
                  water solubility, and the molar concentration of water (Shiau et
                  al., 1994; Edwards et al., 1991).

                  Optimal mobilization  requires alteration of the surfactant system
                  to produce a middle-phase microemulsion; these systems have
                  the potential to elute the residual saturation  in several pore
                  volumes due to significant  reductions in interfacial  tensions.
                  Micellar  systems transition  from normal  to swollen  micelles
                  (WinsorType I), to middle-phase systems (WinsorType III), and
                  finally to reverse micelles (Winsor Type II system, surfactants
                  reside in the oil phase) as properties of the surfactant system are
                  varied. As the surfactant goes from the water phase (hydrophilic)
                  to the oil phase  (lipophilic), the  hydrophilic-lipophilic balance
                  (HLB) decreases. Thus, surfactants with high values of HLB are
                  highly water soluble.  As shown in Figure 1, fora very hydrophilic
                  surfactant system (right side of figure), the surfactant resides in
                  the water phase as  micelles, and a portion  of the oil  phase
                  partitions into the micellar phase. For a very lipophilic surfactant
                  system (the left side  of the figure), the surfactant resides in the
                  oil  phase as reverse micelles.   Intermediate between  these
                  extremes, a third phase appears which consists of water, oil and
                  surfactant; this is the middle-phase system (so designated by
                  virtue of its intermediate density). As shown in Figure 1, these
                  phase transitions  can be  achieved for ionic  surfactants by
                  adjusting the salinity or hardness of the aqueous phase (commonly
                  utilized in surfactant-enhanced  oil  recovery  —  Bourrel  and
                                                           tnll ''lU-Ha
                                                          -'T'p M.siP
                                                             HLO M
Schechter,  1988).   However,  introduction  of high salt
concentrations is not desirable in aquifer restoration as remediation
of brine contamination is also a difficult problem. In this research,
middle-phase systems are achieved by altering the  HLB of a
binary surfactant system.

Surfactant-enhanced  environmental remediation research to
date has evaluated a wide spectrum of issues (e.g., Valsaraj and
Thibodeaux, 1989;  Vignon and Rubin, 1989; Edwards  et al.,
1991; Smith and Jaffe, 1991; Jafvert and Heath, 1991; Peters et
al.,  1992; West,  1992; West and  Harwell, 1992;  Abdul  et al.,
1992;  Fountain, 1992; Pennel  et al., 1993; Baran et  al.,  1994;
Sabatini et al.,  1995).  The objective of this document is to
discuss the technical  feasibility and  limitations of using food
grade surfactants for solubilization and mobilization of chlorinated
solvents,  describe physicochemical and hydraulic approaches
that minimize  surfactant losses while  not  compromising
remediation enhancement, and present a preliminary evaluation
of the economics  of implementing  surfactant-enhanced
subsurface remediation.   This document  emphasizes the
importance of proper design  of  the  surfactant  system and
hydraulic regime for the technical and economical feasibility of
surfactant-enhanced subsurface remediation.


The chlorinated organics evaluated  in  this research were
tetrachloroethylene (PCE), trichloroethylene (TCE), and  trans-
1,2-dichloroethylene (1,2-DCE).   Table I  summarizes
characteristic parameters for these  contaminants.  These
compounds were selected  based on their ubiquitous occurrence
as subsurface contaminants and their range in hydrophobicity
(as  noted by their relative K^ values). The surfactants evaluated
in this research were selected based on their status as FDA direct
food additive compounds and their relative HLB values. The food
grade surfactants are largely combinations of fatty acids and
sugars.   The S-MAZ surfactants  and T-MAZ surfactants are
sorbitan esters and ethyloxylated sorbitan esters, respectively
(with ethylene oxide  groups ranging from 0 to 80).   Table  II
summarizes properties of select surfactants evaluated  in this
research.  Surfactant solubilization, mobilization, precipitation
and sorption  studies were conducted according to  standard
procedures (as  documented  in Shiau et al., 1994,  1995a,b;
1996a,b; Rouse etal., 1993; 1995; 1996). Sorption assays were
conducted utilizing the Canadian River alluvium (CRA) material
which consists of 72% sand, and 27% silt and clay, and has an
organic carbon content of 0.07%.


Solubilization of chlorocarbons was evaluated  using  single
surfactant systems  (SDS and three of the T-MAZ surfactants).
Solubilization  curves  for the three  chlorinated organics  (PCE,
TCE, 1,2-DCE) with T-MAZ 60 are  presented in Figure 2. The
aqueous  phase concentration of the chlorinated  organics
increases linearly above the CMC. However, at higher surfactant
concentrations (exceeding the surfactant solubility limit) negative
deviations from this linear trend are observed  (Shiau  et al.,
1994).  These results  illustrate several points:   (1) below the
CMC, surfactant addition has little to no effect on the solubility of
these contaminants, and (2) the higher the surfactant
concentration is above the CMC the greater is the  chlorocarbon
solubility enhancement due to the increased number of micelles
(i.e.,  operate  well  above the  CMC to achieve  maximum
enhancement,  but below the surfactant's solubility limit). From
Figure 2 the 1,2-DCE solubility enhancement is approximately
two-fold at 10  mM T-MAZ 60 and approximately eight-fold at
50 mM T-MAZ  60; the enhancement approaches two orders of
magnitude for PCE. The salinity of these systems was less than
30 mg/L and the hardness was 16 mg/L (this is true throughout
this section unless noted otherwise).

Solubilization parameters (MSR and Km,  on a molar basis) are
summarized in Table III for SDS, T-MAZ 28, T-MAZ 20, and T-
MAZ 60. The data in Table III demonstrate that more hydrophobic
compounds (e.g.,  PCE) will realize the greatest enhancement
(higher Km indicates greater contaminant partitioning into micelles).
Km  is observed to vary more between  contaminants for any
surfactant  than between surfactants for a given contaminant.
Thus, for the solubilization mechanism, surfactant selection is
relatively independent of the contaminant(s) and will most likely
be made based on factors such as cost, susceptibility to losses,
toxicity.  At the same time, when estimating (extrapolating) Km
values or when modeling the solubilization process, it should be
recognized that micellar solubilization (and thus Km) varies as a
function of contaminant type (nonpolar versus polar/ionic) and
aqueous contaminant concentrations belowtheir water solubility
(Nayyaret al., 1994; Rouse etal., 1995).  Aqueous contaminant
concentrations below water solubility may be experienced due to
nonequilibrium solubilization, mixed NAPL phases, etc. (Pennell
et al., 1993; Rouse et al., 1995;  Pennel  et al.,  1994; Soerens,
•"• 350
0 300
O 250
K 200 -
2 100 .


^ ',
•' « *
...:: - "
                     15   20   25    30

                       T-MAZ 60 CONC. (mM)
                         * TCE
Figure 2: Solubilization of DNAPLs (PCE, TCE, 1,2-DCE) in T-MAZ 60 @
        15°C, After Shiau etal., 1994a.  T MAZ 60 CMC = 0.02 mM.
        Molecular Weight: PCE, 166mg/mM; TCE, 131 mg/mM;DCE,
        97 mg/mM; and T-MAZ 60,  1310 mg/mM.  Data Shown is
        Average of Replicates, Errors within Size of Symbols.

 Table I: Contaminant Properties
Chemical Molecular
Tetrachloroethylene(PCE) C2CI4
Trichloroethylene(TCE) C2HCI3
Trans-1 ,2-Dichloroethylene (DCE) C2H2CI2
Solubility (mg/L)
log K«a
aK = octanol-water partition coefficient
After Shiauetal. (1994)
Table II: Surfactant Properties
Sodium Dodecyl Sulfate (SDS)
POE (80)e sorbitan monolaurate (T-MAZ 28)
POE (20)e sorbitan monolaurate (T-MAZ 20)
POE (20)e sorbitan monostearate (T-MAZ 60)
Aersol OT^AOT)
POE (20)e sorbitan tristearate
Sodium stearoyl-2-lactylate
Sorbitan monooleate
Sorbitan trioleate (S-MAZ85K)
   *MW= molecular weight
   bA = anionic, N = nonionic
   cHLB=hydrophilic-lipophilic balance
   dSMDNS = Sodium mono and dimethyl naphthalene sulfonate
   e() = number ofethylene oxides
   'Bis-2-ethylhexyl sodium sulfosuccinate
   gNA = not available, but included in table in order of HLB (i.e., between 14.9 and 10.5)
   After Shiauetal., 1994
Table III: Solubilization Parameters for Chlorinated Organics
         Chlorinated Organic
MSR    Log Km




After Shiauetal., 1994
T-MAZ 28
T-MAZ 20
T-MAZ 60
T-MAZ 28
T-MAZ 20
T-MAZ 60
T-MAZ 28
T-MAZ 20
T-MAZ 60

Initial efforts  to  achieve  middle-phase  systems, without
consideration of surfactant structure, were unsuccessful.  The
HLB of the surfactant systems was varied  from 2.1  to  40.
Although phase inversion was observed in this HLB range (Type
II to Type  I), a clear middle-phase  was not achieved in  the
transition (instead a mesophase  was  realized).  Using  the
branched surfactant Aerosol OT (AOT) and sodium mono- and
di-methyl naphthalene sulfonate  (SMDNS),  a middle-phase
microemulsion was realized in the  transition region. Thus,  it is
observed that surfactant selection (structure) is critical to achieving
middle-phase systems.

Mobilization with AOT and SMDNS was achieved by varying the
SMDNS  concentration  while  holding the AOT concentration
constant.  Figure  1  shows a phase diagram for 1,2-DCE using
AOT and SMDNS as the surfactant system.   At  low SMDNS
concentrations  a  Type II system  is  realized  (surfactant  has
partitioned  into the oil  phase).    Increasing  the  SMDNS
concentration diminishes the potential for AOT to  partition  into
the oil, thereby allowing it to accumulate at the  interface and
produce  a middle-phase system.   At yet  higher  SMDNS
concentrations, the system is over-optimized and the surfactants
reside  in the  water phase (Type I system).  Middle-phase
systems were achieved via this approach for PCE, TCE, and 1,2-
DCE individually (Shiau et al.,  1994) and in binary and ternary
mixtures  of these  chlorocarbons (Shiau et al., 1996a).

Salinity and hardness can affect the formation of middle-phase
systems. Potential impacts  of naturally occurring  hardness on
middle-phase systems are illustrated in Figure 3 (TCE middle-
phase systems at two  levels  of calcium (hardness)).   Higher
hardness levels required more SMDNS to maintain the surfactant
balance and achieve a middle-phase system (the optimal SMDNS
concentration is higher). The  increased calcium concentration
will tend to drive the ionic AOT into the oil phase, thus requiring
additional SMDNS to achieve the middle-phase microemulsion.
Similarly,  lower  temperatures  increased  the  SMDNS
concentrations necessary to obtain middle-phase systems (Shiau
et al.,  1994).  As expected, surfactant systems  capable  of
achieving middle-phase systems are a function of the DNAPL
(forsingle component) or DNAPLcomposition (for multicomponent
oils). For example, the optimal SMDNS concentration for 0.5%
AOT with PCE, TCE, and  1,2-DCE is 1.40, 2.43 and 2.19 wt%,
respectively (Shiau et al.,  1994).

For multicomponent residual phases,  the optimal surfactant
concentration is a function of the mole fraction of each phase in
the mixed waste and the optimal SMDNS concentration for each
individual phase.  Figure 4 illustrates this by considering a ternary
DNAPL system with a constant mole fraction of 1,2-DCE (0.5050)
and a variable mole fraction  of PCE and TCE (the mole fraction
of PCE is shown on the ordinate and the mole fraction  of TCE is
equal to 0.4950 minus the PCE mole fraction). Thus, an ordinate
value of 0 corresponds to a binary system of TCE and 1,2-DCE
while an ordinate value of 0.4950 corresponds to a binary system
of PCE and 1,2-DCE. The abscissa corresponds to the average
equivalent weight of the surfactant system ([SMDNS]* plus AOT,
where AOT is constant at 0.5 wt%; these units are utilized  to
facilitate the  regular solution analysis).  The  ideal solution
analysis is based on ideal mixing while the regular solution plot
accounts for nonidealities (parameters for the ternary regular
solution predictions were determined independently in three
binary systems).  As can  be seen  in Figure 4, regular solution
theory  is better able to predict phase behavior in  the ternary
system than ideal solution theory (Shiau et al., 1996a).  It is
especially encouraging that, given analyses from binary systems,
one can better predict results for  ternary systems of variable
DNAPL composition.
                    2.3      2.4      2.5

                     WT% of SMDNS ADDED

    0.45 -

     0.4 -

    0.35 -

     0.3 -

    0-26 -

     0.2 -

    0.16 -

     0-1 -

    0.05 -

                                                                              AVE. EQUIL WT. FOR PCE/TCE/DCE MIXTURE
                                                                              AOT = 0.5 WT%, W/0 = 1, OIL = PCE/TCE/DCE
                                                                                              MOLE FRACTION
                                                                                             OF DCE(Xd) = 05050
                                                                                 282     284     286    268
                                                                               AVE. SURFACT. EQUIL. WT. FOR MIXTURE
   n [Ca] = 16 mg/l (gw)
                                [Ca] = 478mg/l
                                                                       • REGULAR SOLUTION
                                                                                         - IDEAL SOLUTION  #  EXPERIMENTAL
Figure 3: Phase Diagram for TCE at 15°C for AOT and SMDNS for Two
        Calcium ValuesfSMDNS* Optimal Concentration at Each Calcium
        Level), Adapted from Shiau etal., 1994.
Figure 4: Phase Diagram for Ternary DNAPL System for Constant Mole
        Fraction of 1,2-DCE and Varying Mole Ratio of PCE and TCE.

These results illustrate the sensitivity of middle-phase systems
to aquifer conditions (temperature, hardness, etc.), contaminant
composition, and surfactant structure, and should alert users to
potential problems  if proper surfactant selection and  design
efforts are not  utilized when implementing mobilization via
middle-phase microemulsions. Also, vertical migration of released
residual may be realized depending on aquifer conditions and the
hydraulics of the extraction system.  This again illustrates the
care that must be taken in utilizing mobilization. Despite these
limitations, mobilization has the potential to be significantly more
efficient  than  solubilization  and  should not  be  prematurely
dismissed as a viable technology.  In  fact, vertical  circulation
wells can potentially mitigate and even take advantage of any
tendencies for  vertical  migration.  Both  of these issues are
discussed below in more detail.

In comparing the efficiency of the solubilization and mobilization
mechanisms, Table IV documents the enhancements resulting
from these two mechanisms for a similar weight percent of
surfactant (via solubilization with T-MAZ 60 and mobilization with
Aerosol  OT and SMDNS).   As observed in Table IV, the
enhancement is two orders of magnitude for PCE via solubilization
and  three and  one-half orders of magnitude via mobilization
(relative  to water alone).  For  1,2-DCE the enhancement  by
solubilization is  approximately one order of magnitude, while
being two orders of magnitude for mobilization.  The increase in
efficiency via mobilization versus solubilization is dramatic. The
data in Table  IV also demonstrate that surfactant-enhanced
subsurface remediation will be of the greatest benefit for more
hydrophobic compounds.

In Figure 5 solubilization and  mobilization  results  from one-
dimensional column studies are presented (Shiau et al., 1995b).
Mobilization  (AOT/SMDNS) achieved  higher contaminant
concentrations and eluted the PCE more quickly than solubilization
(T-MAZ-60) (> 97% extracted in ca. 3 pore volumes).  The tail on
the solubilization curve indicates the reduced extraction rate and
thus slow approach to complete PCE elution for solubilization
(likely due to interfacial area constraints); while ca. 85% of the
PCE is eluted within 10 pore volumes, less than 90% has been
eluted by 30 pore volumes. Again, the potential advantages of
mobilization over solubilization are apparent.  From Figure 5, the
aqueous PCE concentration prior to surfactant introduction is
observed to be 80 mg/L.  Based  on  this initial concentration,
Table IV: Comparison of Solubilization and Mobilization
                                                                1000000 r
  Chlorinated    GW>
   Organic     Solubility
   Solubilization       Mobilization11
(6.5wt%T-MAZ60)   (5.0 wt% AOTand
     (mg/L)         SMDNS)(mg/L)
    PCE       80.6

    TCE       990

   1,2-DCE      5,340




 *GW= Ground Water
 ''Surfactant concentration is based on initial aqueous volume;
 contaminant concentration is in middle phase system.
 After Shiau etal., 1994
                                                            10     15      20      25

                                                            RELATIVE PORE VOLUME
                                                 *  T-MAZ-60
Figure 5: Column Results for PCE Elution via Solubilization (T-MAZ 60)
        and Mobilization (AOT/SMDNS) Systems with CRA Medium.

water dissolution would require in excess of 400 pore volumes to
dissolve the residual PCE.  Considering  reduced mass transfer
as the interfacial area decreases, the actual number of pore
volumes for water dissolution  is expected to  be much greater
than 400 pore volumes.  The relative advantages of surfactant-
enhanced  remediation over water dissolution are apparent.  It
should be  noted  that the concentrations in Figure 5 are still
several orders of magnitude above the drinking water standard.
While surfactanttechniques significantly expedite mass removal,
protracted amounts of time  may be required to achieve drinking
water standards (although much less than with water alone).

It is  observed that  the food grade surfactants  are technically
viable for use in subsurface remediation activities. This may be
advantageous for obtaining  regulatory approval for utilizing
surfactants in  subsurface remediation.  However,  edible
surfactants may be more susceptible to losses (e.g., precipitation,
sorption;  Shiau et al., 1995a),  which can cause system failure
(e.g., pore clogging  due to  precipitation) and will result in
increased costs of the surfactant (thereby hindering the economic
viability of the process).  For this reason,  research in our
laboratories has evaluated methods for minimizing precipitation
and sorption of food grade surfactants (e.g.,  use of hydrotropes
or cosolvents to decrease the sorption and precipitation of these
surfactants — Krebbs-Yuill, 1995).  Also,  in other research
surfactants with indirect food additive status have been evaluated
that are less susceptible to precipitation and sorption (e.g., alkyl
diphenyloxide disulfonates  - Rouse et  al., 1993; ethoxylated
alkylsulfates - Rouse et al., 1996).  Space limitations prevent a
detailed discussion  of techniques for limiting surfactant losses;
the interested reader is directed to the  above references and
Shiau et al.  (1995b) for more  detailed discussions.   Thus,
surfactant selection is  not  only critical to  the technical

implementation of surfactant-enhanced subsurface remediation,
but is  also imperative  to  minimizing subsurface losses and
improving the economics of implementing the technology.  This
will be further discussed in a later section focusing on economic
The  technical and  economic feasibility of surfactant-based
remediation  processes will  depend on  the ability to achieve
hydraulic control in  the subsurface while also attempting to
maximize  hydraulic efficiency.  Hydraulic  efficiency can be
increased by: (1) minimizing the volume of injected surfactant
solution; (2) minimizing the volume of fluid to be pumped to the
surface (reducing treatment costs);  (3) targeting injected
chemicals to  the contaminated zones of the  aquifer (thereby
preventing the movement of injected fluids towards clean portions
of the aquifer); and (4) maximizing capture of resulting water-
surfactant-contaminant mixtures.

One  method of improving the hydraulic efficiency of surfactant-
enhanced subsurface remediation is the strategic placement of
impermeable and/or hydraulic barriers.  Impermeable physical
barriers (e.g., grout curtains, slurry walls, sheet piling) can be
used to deflect flows into or  away from contaminated zones by
creating zones of low permeability.   Hydraulic barriers  (e.g.,
injection wells, infiltration galleries) can be used to deflect flows
into or away from contaminated  zones by creating zones of
increased hydraulic potential (head). A recent numerical modeling
study assessed  the  relative effectiveness of hydraulic and
impermeable  barriers  for improving the efficiency of DNAPL
remediation processes, both with and without surfactants (Gupta,
1993). The overall conclusion drawn from these results was that
mass transfer of the contaminant from the residual phase to the
fluidmov'mg through the contaminated zone should be maximized,
regardless of the fluid used (e.g., air, water, chemical solution).
Simple upgradient  injection of  surfactants  followed  by
downgradient extraction is tremendously inefficient due to loss of
surfactant to uncontaminated  zones.  Injection  of surfactant
solutions inside partially encircling  impermeable barriers with
downgradient deflector wells was found to be the most efficient
of the systems evaluated for the  surfactant-based processes.
The impermeable barrier cuts off upgradient water (eliminates
dilution of surfactant solution) and prevents migration of surfactant
solutions into uncontaminated areas.  Hydraulic barriers (deflector
wells) provide increased gradient in addition to directional control
while also having the advantage of being a temporary measure.
The volumes (mass) of surfactant solution required to exceed the
CMC  in the contaminated zone decreased significantly (up  to
65%) with barriers over simple injection/extraction (Gupta, 1993).

The use of operational measures  has also been suggested for
improving  pump-and-treat  efficiency; such measures  include
cyclic(pulsed) pumping, push-pull pumping, and variable injection/
extraction ratios (Keely,  1989).   Pulsed  pumping was first
proposed by petroleum engineers  to improve  recovery from
hydrocarbon reservoirs (Aguilera, 1980).  Disadvantages
associated with pulsed pumping identified in laboratory and field
studies include increased remediation times (due to decreased
concentration gradients during resting  phase),  operation and
maintenance issues,  and lack of necessary hydraulic control
(Stallard  and Anderson, 1992; Armstrong et al.,  1994; and
Voudrias and Yeh, 1994).

Simultaneous injection to and  extraction from a common vertical
borehole creates a circulating flow  pattern  (Figure 6) within a
sphere or ellipsoid around the borehole (referred to as vertical
circulation wells — VCWs), a concept previously evaluated  in
petroleum production.   A myriad of potential benefits of VCWs
can be delineated, including:   (1) reduced costs  over systems
involving multiple wells; (2) effective hydraulic control achieved
                     Injection  Interval
                                                                                   vertical Migration
                                                                                   of Mobilized DNAPL
                                                                               Diffusion Limited Zone

                                                                                Limiting Streamline
Figure 6: Vertical Circulation Well (VCW) Flow Pattern.

over limited volumes of the formation; (3)  ability to capture
DNAPLs that might sink when mobilized; (4)  applicable to both
light NAPLs (LNAPLs or floaters)  and DNAPLs (sinkers); (5)
minimizes surfactant losses; (6) minimizes  volumes of fluids
produced at the surface requiring  treatment; and (7) induced
mounding can remediate portions of the contaminated vadose
zone around the well. Obviously, these and  alternate systems
must be evaluated on a case-by-case basis to determine optimal

Two-dimensional steady state flow induced by the VCW system
in an aquifer with a regional gradient can be described using the
complex potential,  Q,
                   = O
where, Q is the complex potential,   is the hydraulic potential,
and *¥ is the stream function. Lines of constant  O are called
equipotentials and they describe the head distribution within the
aquifer. Lines of constant *¥ are  called streamlines and  they
describe the flow paths of ground water within the aquifer.

The two screened intervals behave as a line source and a line
sink, respectively.  By superposition, the  complex potential for
the line source  and line sink can  be combined,  along with the
complex potential for a regional gradient (lateral flow), to produce
the overall complex potential forthe aquifer. Using the equations
for line sources/sinks and regional gradient developed by Strack
(1989), the complex potential becomes:
Figure 7 shows equipotentials and streamlines for a VCW in an
aquifer having a local  regional gradient,  where the  bottom
screened interval of the VCW is extracting at a rate higher than
the upper screened interval is injecting (Qout >Qin).  By having Qout
> Qin, it is possible to approach complete capture of the injected
solution; however, the extracted solution is  significantly diluted
by fresh ground water (this effect increases as the Qou/Qin ratio
increases). This will result in diluted extraction concentrations for
both the injected surfactant  solution  and  the solubilized  (or
mobilized)  contaminant.  It is  important to recognize  that the
performance of the  surfactant solution  relative to extraction of
the contaminant is  masked in the VCW system effluent by
dilution due to the fresh ground water that is extracted.  It is also
important to note that the dilution occurs in the borehole and not
in the formation; thus, surfactant concentrations in the aquiferwill
not drop below the CMC due to this phenomenon.

The relative performance oftheVCWsystem versus thetraditional
injection/extraction (two-well) system was assessed using a two-
dimensional (sand tank) model packed with glass beads.   The
tank, constructed of aluminum, is 36 inches wide by 18 inches
high, with a 2 inch depth. The tank has adjustable constant-head
end  reservoirs, a glass  front  plate for visual observation,  and
piezometers for sampling and head measurement distributed
horizontally and vertically throughout the tank. A known mass of
DNAPL was gravity fed from the surface in each test  prior to
removal via surfactant solutions. Aqueous DNAPL concentrations
were quantified throughout each test.  Enhanced solubilization
and  mobilization (microemulsification) were evaluated for each
hydraulic configuration (in both cases the injection and extraction
pumping  rates were  100 ml/min with no  regional  gradient).
                                       a  .
       Z-l)  ln| Z-l] -  Z + l] ln(Z+l|]
            D-  z.  -z.
        D = total length of vertical circulation well (both screens
        and  spacer)
        a  = separation distance between screens (spacer)
        z = x +  iy
        Z = X + iY
        Qo = lateral Darcy velocity
        o  = strength of injection/extraction interval
        z1 =  bottom of lower screen
        z4 =  top of upper screen
                 Z =
The variable transformation from z to Z is simply a change from
the global (x,y) coordinate system to a local (X,Y) system based
on the geometry of the vertical circulation well.
                                 Horizontal Direction

       Figure 7: Cross Section Showing Streamlines for a VCW System in an
               Aquifer with a Regional Gradient. Injection in Upper Zone,
               Extraction in Lower Zone, Regional Gradient from Right to Left.

Preliminary studies of the VCW system and two-well systems
have  also  been conducted  in three-dimensional  sand tanks
(Roberts et al.,  1993; Chen,  1995).

Plotted in Figure 8 are the mass recovery curves for the two-well
and VCW systems using the  surfactant-enhanced solubilization
mechanism for residual saturation of PCE. A4.56 weight percent
(wt%) solution of T-MAZ 20 was used during the solubilization
studies.  The interfacial tension of this solution is 5 dynes/cm.
The curves show that the VCW system achieves higher effluent
contaminant concentrations (increased mass recovery  at early
extraction  volumes)  due primarily  to  reduced  dilution  by
uncontaminatedgroundwater(i.e., improved hydraulicefficiency).
The surfactant recovery is likewise expected to be greater in the
VCW as surfactant lost to the  uncontaminated portions of the
aquifer is minimized.  Decreasing efficiency in the VCW system
is observed at ca. 5 liters. If the initial efficiency of the VCW had
been  maintained it appears  that the remediation  would have
been  complete  within  10 liters; however, in reality it required
greater than 30 liters to  approach  100% recovery  of the PCE.
Thus, the tailing of the VCW mass removal curve  negates the
advantage of the initially higher concentrations

Plotted in Figure 9 are the mass recovery curves for the two-well
and VCW systems via surfactant-enhanced mobilization (again
for residual saturation of PCE).  The surfactant system used was
1.3 wt%AOT and 3.6 wt%SMDNS, which produced an interfacial
tension of 5 x 10"3 dynes/cm.  It can be seen in  Figure 9 that the
two-well system approaches  a  plateau at ca.  70% recovery of
the PCE; visual observations indicated that some ofthe mobilized
residual was lost from  the flowlines due to vertical migration.  It
is very encouraging to note that this did not occur in the VCW
system and that removal efficiencies approached 100% for PCE
in  the VCW  system.   However,  a significant decline in the
extraction efficiency ofthe VCW system occurred at ca.  5 liters.
It is interesting to note that, if not for this decline in efficiency, the
                        VCW system could have remediated the PCE before significant
                        PCE extraction in the two-well system was observed.

                        The change in efficiency for mobilization using the VCW system
                        (ca.  5  liters in Figure  9) was due  to the accumulation of a
                        diffusion-limited mass of contaminant atthe fresh water-surfactant
                        solution interface (i.e.,  the  outermost streamline) of the VCW
                        (see Figure 6). Because the fluid outside the streamline (i.e., the
                        ground water) contains no surfactant, the surfactant system at
                        this interface is diluted and  no longer exists as a middle-phase
                        system. At this interface, slow diffusion-limited dissolution of this
                        contaminant mass  reduces the efficiency of the extraction
                        process, as  reflected in the long tailing ofthe mobilization mass
                        recovery above 5 liters.  It is suggested that this  phenomenon
                        can be addressed through relatively simple operational variations
                        in the VCW system; i.e., reversing injection-extraction direction,
                        operating with three screened intervals with alternating injection
                        and extraction points, etc.

                        In summary, the VCW system has potential advantages overthe
                        two-well system relative to  hydraulic capture of the mobilized
                        residual. The efficiency ofthe VCW system will be maximized
                        when operational variations prevent the limiting effects of dilution
                        at the outermost  streamline.  The VCW system  also has the
                        distinct advantage of higher surfactant recovery compared to the
                        two-well system and potentially lower volumes of water to be
                        treated at the  surface.
                        In conventional pump-and-treat remediation of a site containing
                        residual saturation of a DNAPL, it is not uncommon for contaminant
                        concentrations in the ground waterto be well belowtheirsolubility
                        20       30      40

                           Fluid Extracted (liters)
                                                3D     40     50

                                                 Fluid Extracted (liters)
                   - Two-Well System
VCW System
                                                                               - Two-Well System
                                                                                                  VCW System
Figured: Effluent  Mass Recovery Curves  for Surfactant enhanced
                        Figure 9: Effluent Mass Recovery Curves for Surfactant-enhanced

values. For a given mass of residual DNAPL and a given ground-
water extraction rate, the time required to remediate an individual
DNAPL should be proportional to its aqueous solubility.  This is
plotted in Figure 10, where R is the ratio of the total mass of the
particular DNAPL in  the  soil to the recovery  water flow  rate
(pounds per gallon per minute) and the aqueous concentrations
of the chlorocarbons  are assumed to be 10% of their solubility
limit(e.g.,duetononequilibriumeffects, dilution).  Thus, increasing
values of R  generally indicate either more DNAPL mass for a
given volume of ground water extracted, or lower ground-water
extraction rates for a given DNAPL mass; both of these will result
in greater remediation times. Also shown in Figure 10 is the
remediation  time required for surfactant-enhanced pump-and-
treat remediation; surfactant introduction significantly  reduces
the  remediation time for the more hydrophobic DNAPLs (i.e.,
aqueous solubilities less than about 3000  mg/L).

The curves for surfactant-amended pump-and-treat remediation
in Figure 10 were generated assuming that the  recovered water
has a sodium lauryl sulfate surfactant concentration of 1.7 wt. %
or 10 times its CMC (CMC = 0.006 M) (Rosen, 1989), and the
concentration of DNAPL in the surfactant micelles corresponds
to an aqueous DNAPL concentration  of 10% of its aqueous
solubility.  Molar solubilization ratio (MSR) values utilized were
0.275 for tetrachloroethylene (PCE, aq. solubility = 200 mg/L),
0.20 for trichloroethylene (TCE, aq. solubility =1100  mg/L)
(Sabatini and Knox, 1992), 0.15 for 1,1,2,2-tetrachloroethane
(aq. solubility = 2900  mg/L),  and 0.10 for 1,2-dichloroethane
(1,2-DCA, aq. solubility = 8690 mg/L).  The relationships shown
in Figure 10 provide reasonable estimates of relative costs for
sodium lauryl sulfate as evaluated in this preliminary economic
analysis.  It is recognized that incorporating changes in mass
recovery over the course  of  remediation (as documented in
Figure 5) will more accurately capture the absolute times (and
costs) of remediation. However, the baseline case of conventional
pump-and-treat will be  similarly impacted, minimizing the impact
on the comparative observations presented below.

Another major concern with  surfactant-enhanced  pump-and-
treat technologies is the initial cost of surfactant required to fill the
aquifer.  This cost can be greaterthan $20 million fora very large
aquifer or less than $1 million when  targeted to the residual
DNAPL zone.  Figure 11 shows surfactant capital cost for PCE
remediation with varying aquifer volumes and surfactant prices.
Two different R  values (ratio of PCE  in the aquifer to water
removal rate) are shown. The surfactant concentration is set to
theoretically allow complete remediation of PCE in seven years.
An aquifer volume of 50 million gallons corresponds roughly to 25
acres cross-sectional area by 20 feet thickness  at 30 percent
porosity; 10 million gallons corresponds to 5 acres with the same
thickness and porosity. This range is not unusual for dissolved
contaminant plumes; however,  residual saturation zones are
           10000 =
            1000 =
                           R=25,000; No Surf.

                           R=1 5,000, No Surf.

                           R=5,000; No Surf.
                           R=25,000; w/Surf
                           R=1 5,000; w/Surf
                100           1000           10000
                       Aqueous Solubilty, mg/L
 R=15,000; 18xCMC

""; 6.25xCMC
                                                                                                   $1.00 lib.
                                                                         0    10     20    30    40    50    60
                                                                          Aquifer Pore Volume, Million Gallons
 Figure 10: The Remediation Time Required for Very Insoluble Contaminants
         with 1.7wt percent Surfactant and Water (R = ratio of mass of
         contaminant in aquifer to water removal rate — Ib/gpm).         Figure 11: Initial Surfactant Cost as a Function of Aquifer Pore Volume.

typically only a fraction of this size.  These results show that
surfactant  remediation  should be  targeted to the  residual
saturation zone (where it provides the greatest benefit) as well
as the importance of minimizing surfactant losses (i.e., sorption
and precipitation).

Based  on the capital costs  of the surfactant and  given  that
multiple pore volumes of injection will be utilized, economics
dictate surfactant recovery and reuse.   In a typical hydraulic
regime,  recovery wells will likely produce more water than  is
injected to maintain hydraulic control, which  can result  in  a
significant bleed stream.  Figure 12 demonstrates the cost of
surfactant losses without a recovery step from the bleed stream.
The cost of surfactant recovery from the bleed stream is typically
only a small fraction of the cost of lost surfactant.

In order to identify the most economic surfactant recovery
process, a  base case was studied.  In the base case, DNAPL
contaminated water is extracted at the rate of 500 gpm, the bleed
stream is 150 gpm, and 350  gpm of treated water is reinjected
along with surfactant. It is assumed that DNAPL concentrations
in the reinjected water should be at or below the 0.5 ppb level.
The  treatment  process will  be  based on  chemical/physical
differences between the DNAPL, surfactant and ground water.

The initial separation can be based on differences in  micellar/
aqueous properties, volatility, surfactant/solution properties, or
organic/aqueous properties. Potential unit operations to achieve
the recommended separation are listed in Table V. Two process
Table V: Unit Operations Considered
               Surfactant @ $2.25/lb
                                          ry Cost
         0     50    100   150   200   250   300

                 Bleed, Gallons per Minute

Figure 12: The Costof Surfactant Losseswithouta Recovery Step fromthe
        Bleed Stream.
        Steam Strip
        Vacuum Steam Strip
        Pervapo ration
        Organic Extraction

        Foam Fractionation
        Al, Ca Precipitation; Cation Exchange
        Micellar Enhanced Ultrafi It ration (MEUF)
        MEUF combined with Foam Fractionation

        Carbon Adsorption
schemes are recommended based on economics and proven
ability to remove DNAPL and surfactant to the desired levels.  If
the recovered DNAPL has no value, standard air stripping of the
process feed will remove the DNAPL and leave the surfactant for
reinjection. Catalytic incineration of the DNAPL can be utilized,
followed by caustic scrubbing to remove HCI.  The surfactant is
recovered  from the bleed  stream using a  combination of
ultrafiltration  and foam fractionation.   Ultrafiltration  is  most
efficient down to the CMC level and foam fractionation is very
efficient below the CMC level.  If the DNAPL has sufficient value
to a reclaimer  or can  be  recycled to an adjacent operation,
vacuum-steam  distillation  becomes attractive.   Surfactant  is
recovered from the bleed stream as in the previous case.

The initial design and cost estimates in  this preliminary analysis
assume that surfactants have only minor impacts on stripping
efficiency.  Recent research  has  demonstrated  that micellar
solubilization  decreases the overall efficiency of the stripping
process by reducing the aqueous contaminant  activity and thus
the driving force for volatilization (Lipe et al.,  1996).  Foaming
may also result, which can create significant operating problems
(Lipe et al., 1996); vacuum stripping has been shown to minimize
this effect (Choori,  1994).   Also,  for  nonvolatile compounds
liquid-liquid extraction has demonstrated potential given that the
reduced activity due to micellar  solubilization is considered
(Hasegawa et al.,  1996).   Incorporation of such effects was
beyond the scope of this analysis.

The hypothetical case  studied represents  an  extremely  large
volume of residual  DNAPL contamination; such  a case was
selected to highlightthe important factors in designing surfactant-
enhanced subsurface remediation.  It was assumed that 5 acres
of contaminated soil, 20 feet in depth, had an average porosity
of 0.30, and  the void  fraction contained 13  volume percent
DNAPL. The DNAPLs present were assumed to consist of equal
volumes of PCE, TCE and 1,2-DCA. This equates to 5.74 million
pounds of PCE, 5.16 million pounds of TCE and 4.46 million
pounds of DCA.  The surfactant was assumed to cost $1.00 per
pound and was added in sufficient quantities to achieve 15 times
its CMC in the recovered water. In this analysis, 20 percent of

the surfactant was assumed to be lost due to adsorption in the
first two pore volumes. This surfactant loss is consistent with
surfactant adsorption values of 10 mg/g or greater (Rouse et al.,
1993),  and illustrates the economic importance  of minimizing
surfactant losses. The recovery wells were assumed to operate
at 500 gpm and the bleed stream necessary was assumed to be
150gpm. The recovery stream was assumed to contain 330 mg/
L TCE, 400 mg/L PCE and 950 mg/L 1,2-DCA (based on MSR
values above and assuming aqueous concentrations at 10 percent
of solubility limits and no competitive effects of solubilizates),
giving a theoretical remediation time of seven years.

A total of 68 injection and recovery  wells were assumed along
with piping to the treatment facility. The DUALL Division of MET-
PRO Corporation recommends three air strippers in  series to
treat 500 gpm process water containing PCE, TCE and 1,2-DCA.
The columns are 9 feet in  diameter with 28 feet of packing.
Global Environmental estimates that  a catalytic  incinerator
operating at 1000° F with 30 ft3 of Englehard catalyst will achieve
99 percent destruction efficiency. (Allied also manufactures a
catalyst that can be used with high chloride streams.)  The hot
gas stream must be quenched to about 180°F. A 5-foot diameter
caustic scrubber is  recommended for  removing  HCI from the
cooled gas stream with a design efficiency of 99.5 percent.  The
cost of ultrafiltration and foam fractionation will depend on the
specific surfactant used. Conservative costs were utilized based
on previous work at the  University  of Oklahoma (Dunn et al.,
1985).  Based on this analysis,  the  approximate capital cost is
$3,570,000 and the annual O&M cost is $940,000 (see Krebbs-
Yuill et al., 1995 for more  details).

As an  alternative, costs for the vacuum-steam stripping case
were estimated. AWD, a  division of DOW, estimates installed
costs of $3 million for a 5-foot diameter column, package boiler
(1500 to 2000 Ib/hr steam), and instrumentation.  The overall
capital cost forthis approach is $5,070,000, with the annual O&M
costs at $765,000 (see KrebbsYuill et al., 1995 for more details).
Costs will increase rapidly  if the  water requires significant
pretreatment for inorganics and  if operating  problems  are
experienced due to the presence of the surfactant.  Again, the
actual tower dimensions may increase due to micellar effects
(Lipeetal., 1995).

For a given contamination site, there is  a cost tradeoff between
the number of wells (assuming  the  optimum pumping  rate per
well is known and fixed),  the time required for remediation, the
size of the treatment process, and the initial surfactant cost.  This
is shown for our base case in Figure  13 where the above-ground
process is air stripping-incineration  with surfactant recovery of
the bleed.   The negative net present value,  discounted at
10 percent, is plotted against the gpm of well water recovered.
Again, it is emphasized that these numbers should be viewed in
a relative sense (remembering the simplifying assumptions
made in the analysis).  The surfactant concentration is varied to
give a theoretical remediation time of seven years; therefore, the
initial surfactant cost decreases with increasing gpm.  The cost
of wells and the above ground processing costs increase with
increasing flow rate. For the base case, the optimum flow  rate,
corresponding to the minimum negative net present value, is
about 500 gpm. Initial surfactant costs due to adsorption losses
were not included since they were assumed not to vary between
    Q.  10
                        Total Cost
                   V^       Optimum

Initial Surfactant
  / Cost
                 200     400     600    800
                  Water Recovery Rate, gpm
 Figure 13: Cost Optimization fora Five Acre Surfactant Flood.
The cost benefits of a small, well defined area of contamination
can be shown in the following  example using pump-and-treat
technology to  remediate residual DNAPL from only 1/4  acre
covering a depth of 20 feet and treating 35 gpm using air stripping
and incineration.  The contaminated soil contains approximately
125,000 pounds of residual PCE and estimates indicate it will
take  another 41  years to  remediate.  The operating cost is
$125,000  per year and the net negative present value of this
operating  cost, discounted at  10 percent,  is $1.225 Million.
Adding $350,000 of surfactant (@ $1.00/lb) and  $120,000
capital for surfactant recovery, the remediation time is decreased
to about 3 years. If the additional operating cost is about $30,000
per year, the negative net present value is decreased to $545,000.
Not only is the cost to the company reduced,  but also a  very
undesirable environmental  liability will be eliminated  sooner.

Part of the initial surfactant  cost may be recovered at the end of
the remediation project. This would be accomplished by continued
water flooding at a reduced rate without surfactant addition.  Only
the surfactant recovery portion  of the above-ground  process
would be operated.  The surfactant should retain a fair fraction
of its original value, especially to another surfactant-enhanced
remediation  project (assuming  regulatory  approval of  used
surfactant).  As an alternate, a site may be  compartmentalized
with a smaller initial volume of surfactant shifted from compartment
to compartment (resulting in decreased capital costs of surfactant

but increased remediation time).   Maximizing  recovery of
surfactant from the subsurface will also be necessary if there is
environmental concern regarding residual surfactant in the aquifer.

In  summary, surfactant-enhanced pump-and-treat  remediation
is  effective  for DNAPL chemicals with relatively low aqueous
solubilities,  but requires surfactant reuse to  be  economical.
From  this preliminary  analysis,  the  best surfactant recovery
process is probably  a combination of microfiltration and foam
fractionation.   Total project  costs can  be minimized by (1)
defining the area of residual saturation, thus  minimizing the
volume of soil requiring surfactant flooding, and (2) balancing the
number of wells and  volume of water to be treated with the cost
of  surfactant and surfactant concentration required to achieve
remediation.  For relatively small areas  of residual saturation
(i.e., acres or less which is common even for dissolved plumes
having dimensions  in miles), surfactant-enhanced pump-and-
treat via solubilization can be less expensive, on a present value
basis, than pump-and-treat alone. The economics for mobilization
will be even more  favorable than solubilization,  especially  if
surfactant slugs  of  less than a  pore volume  can be utilized
followed by polymer drives, etc.; future research should evaluate
such methods  for environmental applications.   When external
costs are incorporated into the analysis (e.g., regulatory or public
mandates, liability, property transactions,  etc.),  it is obvious that
surfactant-enhanced subsurface  remediation has the potential
for widespread implementation.


This work was funded by the U.S. Environmental Protection
Agency under Cooperative Agreement CR818553-01-0 with the
Robert S. Kerr Environmental  Research Center, Ada, OK.  It has
been subjected to Agency review and approved for publication.
Mention  of trade names or commercial  products  does not
constitute endorsement or recommendation for use.  The authors
acknowledge  input and guidance from  Dr. Candida  West of
RSKERC. The many  students who worked faithfully and diligently
on this project (including Bor Jier Shiau, Yili Chen,  and Barbara
Krebbs-Yuill) are gratefully acknowledged by the authors.  The
authors also acknowledge the helpful input of the  anonymous

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

Abdul,  A.  S., Gibson, T. L,  Ang,  C. C.,  Smith, J. C. and
Sobczynski, R. E. "In Situ Surfactant Washing of Plychlorinated
Biphenyls and Oils from a Contaminated Site."  Ground Water,
30(2), 1992,  219-231.
Aguilera, R.,Naturally Fractured Reservoirs, Petroleum Publishing
Company, Tulsa, OK, 1980.

Armstrong, J.E., Frind, E.O. and McClellan, R.D., "Nonequilibrium
Mass Transfer between the Vapor, Aqueous and Solid Phases
in Unsaturated Soils during Vapor Extraction." Water Resources
Research, 1994, Vol. 30, No. 2, pp. 355-368.

Baran, J. R., Pope, G. A., Wade, W. H., Weerasooriyaa, V. and
Yapa,  A., "Microemulsion  Formation with Chlorinated
Hydrocarbons of Differing Polarity." Environmental Science and
Technology. 28(7), 1994, 1361-1366.

Bourrel,  M. and Schechter,  R. S.  Microemulsions and Related
Systems:  Formulations. Solvency,  and  Physical  Properties.
Surfactant Science Series,  Vol. 30,  Marcel Dekker, Inc., New
York, 1988.

Chen, L. Y. "Three-Dimensional Sand TankTesting of Surfactant
Enhanced  Subsurface Remediation."   Ph.D. Dissertation,
University of Oklahoma, Norman,  1995.

Choori, U. N.  "Removal of Volatile Organic Compounds from
Surfactant Solutions by Flash Vacuum Stripping in a Packed
Column."  MS  Thesis,  Chemical  Engineering, University  of
Oklahoma,  1994.

Dunn, R. O. Jr., Scamehorn, J. F.  and Christian, S.  D. , "Use of
Micellar-Enhanced Ultrafiltration  to  Remove  Organics from
Aqueous Streams," Separation Science and Technology, 20(4),
pp.  257-284, 1985.

Edwards,  D. A.,  Luthy, R.  G. and  Liu,  Z.,  "Solubilization  of
Polycyclic Aromatic Hydrocarbons in MicellarNonionic Surfactant
Solutions."EnvironmentalScience  and Technology. 25(1), 1991,

Fountain, J. C.  in  Transport and Remediation of Subsurface
Contaminants: Colloidal, Interfacial and Surfactant Phenomena.
Sabatini, D. A. and Knox, R. C., eds., ACS Symposium Series
491, American Chemical Society,  Washington, DC, 1992, 182-

Gupta, H.S.,"Modeling the Effectiveness of Barriers for Improving
the  Hydraulic Efficiency of a Pump and Treat System Utilizing
Chemical Agents." Master's  thesis,  University of Oklahoma,
Norman, OK,  1993.

Haley, J. L., Hanson, B., Enfield, C., and Glass, J., "Evaluating
the  Effectiveness of Ground Water Extraction Systems." Ground
Water Monitoring Review.   Winter 1991,  119-124.

Hasegawa, M., Sabatini, D.  A. and Harwell, J. H. "Liquid-Liquid
Extraction for Surfactant-Contaminated Separation." Accepted
by  Journal of Environmental Engineering Division -  ASCE,
Revised and resubmitted May 10, 1996.

Jafvert, C. T. and Heath, J. K., "Sediment and Saturated - Soil
Associated reactions Involving  an  Anionic  Surfactant
(Dodecylsulfate).  1. Precipitation  and  Micelle  Formation."
Environmental Science and Technology.   25(6),  1991,  1031-

Keely, J.F., "Performance Evaluation of Pump-and-Treat
Remediations", U.S. Environmental Protection  Agency,  EPA/
540/4-89/005, Ada, OK, 1989.

Krebbs-Yuill, B. "Surfactant Enhanced Remediation of Residual
Organic Contaminants."  Ph.D. Dissertation, The University of
Oklahoma,  1995.

Krebbs-Yuill, B., Harwell, J. H., Sabatini, D. A. and Knox, R. C.
"Economic Considerations in Surfactant enhanced Pump-and-
Treat." in  Surfactant Enhanced  Subsurface  Remediation:
Emerging Technologies.  D. A. Sabatini, R. C. Knox and J. H.
Harwell, eds.  ACS Symposium Series 594, American Chemical
Society, Washingtion, DC, 1995, 265-278.

Lipe,  M., Sabatini, D.  A., Hasegawa,  M.,  and  Harwell,  J. H.
"Micellar-EnhancedUltrafiltration and Air Stripping for Surfactant-
Contaminant Separation and Surfactant Reuse."  Ground Water
Monitoring and Remediation, 16(1), Winter 1996, 85-92.

Nayyar, S.  P., Sabatini, D. A. and Harwell, J. H.  "Surfactant
Adsolubilization and  Modified Admicellar Sorption of Nonpolar,
Polar and  lonizable  Organic  Contaminants,"    Environmental
Science and Technology. 28(11), 1994, 1874-1881.

Palmer, C. D. and Fish, W.  "Chemical Enhancements to Pump
and  Treat Remediation," USEPA, EPA/540/S-92/001, 1992,

Pennell, K.  D., Abriola, L. M. and Weber, W. J. Jr., "Surfactant
enhanced Solubilization of Residual Dodecane in Soil Columns.
2.  Mathematical  Modeling."  Environmental  Science  and
Technology. 27(12), 1993,2341-2351.

Pennell, K.  D., Minquan,  J., Abriola, L. M. and Pope, G. A.
"Surfactant Enhanced Remediation of Soil Columns Contaminated
by Residual Tetrachloroethylene."  Journal of Contaminant
Hydrology.  16, 1994, 35-53.

Peters, R. W., Montemagno, C. D., Shem, L.,  and  Lewis, B. A.
"Surfactant Screening of Diesel-Contaminanted Soil." Hazardous
Waste and Hazardous Materials.   9(2), 1992,  113-136.

Roberts, B., Harwell, J. H., Sabatini, D. A., and  Knox, R. C.
"Sandtank Testing of Surfactant Enhanced  Remediation  of
Subsurface Contamination by Chlorinated Solvents," Completion
report submitted to Air Force Civil Engineering Support Agency,
Civil  Engineering Laboratory,  Tyndall Air Force  Base, Florida,
April  1993.

Rosen, M.J., Surfactants and Interfacial Phenomena. Second
Edition, John Wiley and Sons, 1989.

Rouse, J. D., Sabatini, D.  A. and  Harwell, J. H., "Minimizing
Surfactant Losses Using Twin-Head  Anionic  Surfactants  in
Subsurface  Remediation."   Environmental Science  and
Technology. 27(10),  1993,  2072-2078.

Rouse,  J.  D., Sabatini,  D. A. and Harwell,  J. H.   "Micellar
Solubilization of Unsaturated  Hydrocarbon Concentrations as
Evaluated by Semiequilibrium  Dialysis." Environmental Science
and  Technology,  29(10),  1995, 2484-2489.
Rouse, J. D., Sabatini, D. A. and Harwell, J. H.  "Evaluation of
Ethoxylated Alkylsulfate Surfactants for use in Subsurface
Remediation."   Water Environmental Research.  68(2), 1996,

Sabatini, David A. and  Robert C. Knox,  "Transport  and
Remediation of Subsurface Contaminants",  Sabatini, D. A. and
Knox, R. C.,  eds., ACS  Symposium Series 491,  American
Chemical Society, Washington, DC,  1992.

Sabatini, D. A., R. C.  Knox and J. H. Harwell, eds.  Surfactant
Enhanced Subsurface Remediation:  Emerging Technologies.
ACS Symposium Series  594,  American  Chemical Society,
Washington, DC, 1995.

Shiau, B. J., Sabatini, D. A. and Harwell, J. H., "Solubilization and
Microemulsification of Chlorinated Solvents Using Direct Food
Additive (Edible) Surfactants." GroundWater.  32(4), 1994, 561-

Shiau,  B. J., Sabatini, D. A. and Harwell, J. H, and Vu,  D.
"Microemulsion of Mixed Chlorinated Solvents Using Food Grade
(Edible) Surfactants,"  Environmental Science  and Technology,
30(1), 1996a, 97-103.

Shiau,  B. J., Sabatini, D. A. and Harwell, J.  H.  "Food Grade
Surfactants for Subsurface Remediation: Column Studies."  In
Review, Journal of Environmental Engineering - ASCE, 1996b.

Shiau, B. J., Sabatini, D. A. and Harwell, J. H.  "Properties of Food
Grade (Edible) Surfactants Affecting Subsurface Remediation of
Chlorinated Solvents." In  Press,  Environmental Science and
Technology. 29(12), 1995a, 2929-2935.

Shiau,  B. J., Rouse, J. D., Sabatini, D. A.  and Harwell, J.  H.
"Surfactant Selection for  Optimizing  Surfactant  Enhanced
Subsurface Remediation."  in Surfactant Enhanced Subsurface
Remediation:   Emerging Technologies.   D. A. Sabatini, R.  C.
Knox and J. H. Harwell, eds.   ACS Symposium Series 594,
American Chemical Society, Washingtion, DC, 1995b,  65-79.

Smith, J. A. and Jaffe, P. R., "Comparison ofTetrachloromethane
Sorption to  an  Alkylammonium-Clay and an Alkyldiammonium
and Clay." Environmental Science and  Technology.  25,1991,

Soerens, T. S.   "Dissolution of Nonaqueous  Phase  Liquids in
Porous  Media:   Mass Transfer in  Aqueous and  Surfactant
Systems." PhD Dissertation, The University of Oklahoma, 1995.

Stallard, M. and Anderson, E., Aquifer Restoration: Pump-and-
Treat and the Alternatives, AGWSE  Educational Program, Las
Vegas, NV, 1992.

Strack, O.D.L,  Groundwater Mechanics. Prentice Hall, 1989.

Valsaraj, K. T., and Thibodeaux,  L J., "Relationships between
Micelle-Water  and  Octanol-Water Partition Constants  for
Hydrophobic Organics of Environmental Interest."   Water
Research. 23(2), 1989, 183-189.

Vignon, B. W. and Rubin, A. J., "Practical Considerations in the
Surfactant-Aided Mobilization of Contaminants in Aquifers."
Journal of the Water Pollution Control Federation.  61(7), 1989,

Voudrias, E.A. and Yeh, M., "Dissolution of a Toluene Pool under
Constant and Variable Hydraulic Gradients with Implications for
Aquifer Remediation." Ground Water, 1994, Vol. 32, No. 2, pp.

West, C. C.,  in  Transport and Remediation  of Subsurface
Contaminants: Colloidal, Interfacial and Surfactant Phenomena.
Sabatini,  D. A. and Knox,  R. C., eds., ACS Symposium Series
491, American Chemical Society, Washington, DC, 1992, 149-

West, C. C.  and Harwell, J.  H.,  "Surfactants and  Subsurface
Remediation."  Environmental Science and Technology.  26,
1992, 2324-2330.