United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/R-99/095C
September 1999
An In-Situ Permeable
Reactive Barrier for the
Treatment of Hexavalent
Chromium and
Trichloroethylene in
Ground Water:
Volume 3
Multicomponent Reactive
Transport Modeling
7 L£
0 0.5
i
Cr(VI)
8.4xlO'05
y.Oxio"05
5.6xlO-°5
4.2xlO-°5
2.8xlO'05
1.4xlO-°5
O.OxlO+0(
[mol I"1]
0 0.5 1
1.5
distance [m]
2.5
3.5
-------
EPA/600/R-99/095C
September 1999
An In-Situ Permeable Reactive Barrier for
the Treatment of Hexavalent Chromium and
Trichloroethylene in Ground Water:
Volume 3
Multicomponent Reactive Transport Modeling
David W. Blowes
K. Ulrich Mayer
Department of Earth Sciences
University of Waterloo
Waterloo, Ontario, Canada
Cooperative Agreement No. CR-823017
Project Officer
Robert W. Puls
Subsurface Protection and Remediation Division
National Risk Management Research Laboratory
Ada, OK 74820
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
Notice
The U. S. Environmental Protection Agency through its Office of Research and
Development partially funded and collaborated in the research described here
under Cooperative Agreement No. CR-823017 to University of Waterloo. It has
been subjected to the Agency's peer and administrative review and has been
approved for publication as an EPA document. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
All research projects making conclusions or recommendations based on
environmentally related measurements and funded by the Environmental Protec-
tion Agency are required to participate in the Agency Quality Assurance Program.
This project was conducted under an approved Quality Assurance Project Plan.
The procedures specified in this plan were used without exception. Information on
the plan and documentation of the quality assurance activities and results are
available from the Principal Investigator.
-------
Foreword
The U.S. Environmental Protection Agency is charged by Congress with
protecting the Nation's land, air, and water resources. Under a mandate of national
environmental laws, the Agency strives to formulate and implement actions leading
to a compatible balance between human activities and the ability of natural systems
to support and nurture life. To meet these mandates, EPA's research program is
providing data and technical support for solving environmental problems today and
building a science knowledge base necessary to manage our ecological resources
wisely, understand how pollutants affect our health, and prevent or reduce environ-
mental risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's
center for investigation of technological and management approaches for reducing
risks from threats to human health and the environment. The focus of the
Laboratory's research program is on methods for the prevention and control of
pollution to air, land, water, and subsurface resources; protection of water quality in
public water systems; remediation of contaminated sites and ground water; and
prevention and control of indoor air pollution. The goal of this research effort is to
catalyze development and implementation of innovative, cost-effective environ-
mental technologies; develop scientific and engineering information needed by
EPA to support regulatory and policy decisions; and provide technical support and
information transfer to ensure effective implementation of environmental regula-
tions and strategies.
Environmental scientists are generally familiar with the concept of barriers for
restricting the movement of contaminant plumes in ground water. Such barriers are
typically constructed of highly impermeable emplacements of materials such as
grouts, slurries, or sheet pilings to form a subsurface "wall." The goal of such
installations is to eliminate the possibility that a contaminant plume can move
toward and endanger sensitive receptors such as drinking water wells or discharge
into surface waters. Permeable reactive barrier walls reverse this concept of
subsurface barriers. Rather than serving to constrain plume migration, permeable
reactive barriers (PRBs) are designed as preferential conduits for the contaminated
ground water flow. A permeable reactive subsurface barrier is an emplacement of
reactive materials where a contaminant plume must move through it as it flows,
typically under natural gradient, and treated water exits on the other side. The
purpose of this document is to provide detailed design, installation and perfor-
mance monitoring data on a full-scale PRB application which successfully remediated
a mixed waste (chromate and chlorinated organic compounds) ground-water
plume. It was also the first full-scale installation of this technology to use a trencher
to install a continuous reactive wall to intercept a contaminant plume. The informa-
tion will be of use to stakeholders such as implementors, state and federal
regulators, Native American tribes, consultants, contractors, and all other inter-
ested parties. There currently is no other site which has used this innovative
technology and reported on its performance to the extent detailed in this report. It
is hoped that this will prove to be a very valuable technical resource for all parties
with interest in the implementation of this innovative, passive, remedial technology.
Clinton W. Hall, Director
Subsurface Protection and Remediation Division
National Risk Management Research Laboratory
-------
IV
-------
Abstract
Reactive transport modeling has been conducted to describe the performance
of the permeable reactive barrier at the U.S. Coast Guard Support Center near
Elizabeth City, N.C. The reactive barrier was installed to treat groundwater con-
taminated by hexavalent chromium and chlorinated organic solvents. The concep-
tual model of the Elizabeth City site described in Volumes 1 and 2 of this document
series (Blowes et al., 2000) provide the basis for the modeling study. The
multicomponent reactive transport model MIN3P was used forthe simulations. The
essential reactions contained in the conceptual model are aqueous complexation
reactions, combined reduction-corrosion reactions between the treatment material
zero-valent iron and the contaminants or other electron acceptors dissolved in the
ambient groundwater and the precipitation of secondary minerals within the reac-
tive barrier. The simulations have been carried out along a cross-section through
the barrier that corresponds to a transect of the monitoring network. One- and two-
dimensional simulations were conducted. The one-dimensional simulations were
carried out along a zone of preferential flow, which conveys the most pronounced
Cr(VI)-contamination. The model has been calibrated using field data, laboratory
data and reaction rates reported in the literature. The two-dimensional simulations
were conducted based on hydraulic conductivities determined from slug-tests.
These simulations allow an evaluation of the impact of preferential flow on the
treatment of the contaminants and secondary reactions. The model results provide
estimates of the potential effects of the consumption of zero-valent iron and the
precipitation of secondary minerals on the long-term efficiency of the treatment
system.
-------
VI
-------
Contents
Foreword iii
Abstract v
Tables viii
Figures ix
Introduction 1
Description of MIN3P 1
Conceptual Model 5
Definition of Reaction Network 5
Complexation Reactions 5
Reduction-corrosion Reactions 5
Formation of Secondary Minerals in Treatment Zone 6
pH and Eh-buffering Downgradient of Barrier 6
pH-buffering 6
Eh-buffering 7
Solution Domain and Model Parameters 8
Spatial Discretization 8
Physical Parameters and Hydraulic Conductivity Distribution 8
Mineralogical Parameters 9
Boundary and Initial Chemical Composition of Ground Water 9
Calibrated Rate Constants 9
Results and Discussion 10
One-dimensional Simulations 11
Removal of Contaminants 11
Reduction of Electron Acceptors 11
Selected Cation Concentrations 11
pHand Eh 12
Discussion of Reaction Mechanisms 12
Corrosion of Zero-valent Iron 13
Precipitation of Secondary Minerals 13
Long Term Efficiency 13
Two-dimemsional Simulations 14
Ground-water Flow 14
Removal of Contaminants 14
pHand Eh 14
Sulfate Reduction 15
Conclusions 15
References 16
VII
-------
Tables
Table 1. Complexation reactions and equilibrium constants 19
Table 2. Reaction stoichiometries of reduction-corrosion reactions 21
Table 3. Secondary minerals in reactive barrier and corresponding equilibrium constants 22
Table 4. Physical parameters for aquifer and reactive barrier material 22
Table 5. Initial mineral volume fractions in reactive barrier and aquifer 22
Table 6. Reactive surface area estimates for zero iron-valent (field installation) 22
Table 7. Reactive surface area estimates for Eh buffer minerals 23
Table 8. Input concentrations at boundary located upgradient of reactive barrier,
Transect 2, 21-1 -21-4 23
Table 9. Input concentrations at boundary located upgradient of reactive barrier,
Transect 2, 21-5-21-7 24
Table 10. Reaction processes affecting component concentrations 25
Table 11. Rate constants for reduction-corrosion reactions 26
Table 12. Calibrated effective rate constants for secondary mineral formation 26
Table 13. Estimated rate constants for reductive dissolution reactions 26
VIM
-------
Figures
Figure 1. Configuration of reactive barrier and approximate location of
chromium plume, from Bennett [1997] 28
Figure 2. Monitoring network, from Benneff[1997] 28
Figure 3. Conceptual model for reactive barriers comprised of
zero-valent iron, from Benneff [1997] 29
Figure 4. Solution domain including location of barrier and monitoring
points along Transect 2 29
Figure 5. Spatial discretization of two-dimensional solution domain 30
Figure 6. Hydraulic conductivity distribution in two-dimensional solution domain,
modified from Bennett [1997] 30
Figure 7. Contaminant concentrations after t = 240 days: a) chromium,
b) organics - one-dimensional simulation 31
Figure 8. Redox couple concentrations after t = 240 days: a) nitrate/ammonia,
b) sulfate/sulfide - one-dimensional simulation 31
Figure 9. Selected cation concentrations after t = 240 days - one-dimensional simulation 32
Figure 10.pH and Eh aftert = 240 days - one-dimensional simulation 32
Figure 11. Iron corrosion rates in reactive barrier after t = 240 days 33
Figure 12.Secondary mineral volume fractions in reactive barrier after t= 240 days 33
Figure 13. Long term effect of iron corrosion and secondary mineral formation 34
Figure 14.Streamlines in two-dimensional solution domain 34
Figure 15.Hexavalent and trivalent chromium concentrations and CrOH3(am)
volume fractions after t = 2 years 35
Figure 16.TCE, cis-1,2 DCE and VC concentrations after t = 2 years 36
Figure 17. pH and Eh distribution after t = 2 years 37
Figure 18.Sulfate and sulfide concentrations and mackinawite volume fractions
aftert = 2 years 38
-------
-------
Introduction
This report describes the application of the multicomponent reactive solute transport model MIN3P [Mayer, 1999], using
one- and two-dimensional reactive transport simulations to model treatment of contaminated ground water by an in-situ
permeable reactive barrier. The modeling study is based on the conceptual model developed by Bennett [1997] for the
barrier installation at the U.S. Coast Guard Support Center near Elizabeth City, North Carolina. The reactive barrier was
designed to treat the ambient ground water, which is contaminated with hexavalent chromium, and trichloroethylene
(TCE) and its degradation products. The technology for this barrier was developed and is atented by the University of
Waterloo. The chromium contamination originates from a plating facility located in Hangar 79 (Figure 1). The
contaminated ground water is moving north, and is intercepted by the reactive barrier before entering the Pasquotank
River. Chromium concentrations upgradient of the reactive barrier exceed 1 mg L1 and locally reach up to 5 mg L1
[Benneff,1997; Blowes et al., 2000]. A larger plume containing chlorinated organic compounds emanating from the
vicinity of Hangar 79 and is also treated by the barrier [Benneff,1997; Blowes et al., 2000]. The barrier is comprised of
granular iron and remediates the ground water by reduction of hexavalent chromium and subsequent precipitation in the
form of chromium containing hydroxides [Bennett, 1997; Blowes et a/.,1999]. Trichloroethylene and its major degradation
products cis-1,2 dichloroethylene (cis-1,2 DCE) and vinyl-chloride (VC) are converted to non-toxic hydrocarbons by
reductive elimination and hydrogenolysis [Bennett, 1997; Blowes etal, 1997]. An extensive field monitoring program was
initiated at the site [Benneff,1997; Blowes et al., 2000]. Multi-level sampling wells facilitate a detailed description of the
hydrogeology of the aquifer and the geochemical conditions upgradient, within and downgradient of the treatment system
along three transects (Figure 2). Only limited data regarding the mineralogical composition of the aquifer and the
treatment material were available when conducting the simulations. Assumptions made with respect to the mineralogy
are compared to the results of the mineralogical analyses carried out by Pa/mer[1999].
In the following section the formulation of the numerical reactive transport model MIN3P is introduced. The conceptual
model developed by Bennett [1997], which describes the controlling transport and reaction processes in the treatment
system and forms the basis for this modeling study, is presented. The reaction network for the numerical analysis is
defined based on this conceptual model. One- and two-dimensional reactive transport modeling is conducted to describe
the geochemical evolution of ground water along Transect 2 (Figure 2). Changes of the geochemical composition of the
reactive mixture and the aquifer material downgradient of the barrier are also addressed. An investigation of processes
potentially affecting the long-term performance of the reactive barrier is carried out. The modeling results are discussed
with a focus on the expected efficiency and longevity of the treatment system. The effect of preferential flow on the quality
of the treatment for the various contaminants is investigated.
Description of MIN3P
The model MIN3P was developed by Mayer [1999] as a general purpose multicomponent reactive transport model
facilitating the simulation of reactive solute transport in variably-saturated media. The governing equations presented
here focus on the formulation for saturated porous media only, because the current modeling study does not include
unsaturated conditions. In this case, the mass conservation equations for reactive transport consist of relationships
describing advective-dispersive transport of aqueous species under the influence of homogeneous and heterogeneous
geochemical reactions. A global implicit formulation [Steefel and Lasaga,1994] is used, in which case the geochemical
reaction expressions are directly substituted into the transport equations. Reaction processes considered are aqueous
complexation, oxidation-reduction, ion-exchange, and dissolution-precipitation reactions. The formulation of the model is
based on a partial equilibrium approach [Lichtner, 1985; Sevougian et al., 1993; Steefel and Lasaga, 1994], which
contains geochemical equilibrium reactions as well as homogeneous and heterogeneous kinetic reactions. The global
mass conservation equation for the components Acare defined by:
-------
where t defines time and is porosity. The concentrations of the dissolved species are defined by Cyc (concentrations of
components as species in solution) and (^(concentrations of complexed species). C* defines the concentrations of the
ion-exchanged species. Nx defines the number of aqueous complexes, and Ns identifies the number of ion-exchanged
species in the system, v^and v* are the stoichiometric coefficients of the components in the complexed species and ion-
exchanged species. The source-sink terms Qyam and Qyaa define the gain or loss of mass for a component due to
kinetically-controlled dissolution-precipitation reactions, and kinetically-controlled intra-aqueous reactions, respectively.
va defines the Darcy velocity.
The hydrodynamic dispersion coefficient Da is given by [Bear, 1972; Linger et al., 1995]
+oc,
where k and /define the spatial coordinates, a, is the longitudinal and at is the transverse dispersivity of the porous
medium, |vj is the magnitude of the Darcy flux in the aqueous phase, D* is an averaged free liquid diffusion coefficient
used for all dissolved species [m2 s~1] and ra is the tortuosity of the medium [-]. The formulation of the dispersion tensor
is based on the implementation by Linger et a/.[1995] and allows for different transverse dispersivities in horizontal and
vertical directions, which can be defined as ath and a^. 5W defines the Kronecker delta.
The governing equation 1 for reactive transport can be simplified by expressing the concentrations of all aqueous
species in terms of total aqueous component concentrations Ta [mol L1 water] [Kirkner and Reeves, 1988; Steefel and
Lasaga, 1994; Lichtner, 1996]:
(3)
A corresponding relationship can be defined forthe adsorbed species in terms of total sorbed component concentrations
P [mol L1 bulk]:
'
(4)
Equation 1 can then be written as:
-\ 3j"
-e-e=o J = I,NC <5)
To complete the system of governing equations, an additional set of mass conservation equations has to be defined,
which describes the change of mineral quantities overtime [Steefel and Lasaga, 1994]:
^L = \Q-^V,mRm i = \,N
dt ' ' m (6)
where
-------
where Ad and 6dare constants, aJsthe ion size parameter, bisan ion specific parameter that accounts for the decrease
in solvent concentration in concentrated solutions. If ai is available, but not bf equation 7 is used with £>. = 0. The Davies
equation is used as an approximation when the parameter ai cannot be provided:
log rf=-4Z? --0.247
(8)
i_ _i
The activity coefficients for all neutral species excluding water are calculated according to:
log 7* =0.1/
(9)
while the activity correction for water is defined by:
r^n =1-0.017
A stoichiometric relationship for the dissociation of the aqueous complex A* into components as species in solution can
be formulated as:
where v;x is the stoichiometric coefficient of the j ft component in the / th aqueous complex A*. Equilibrium complexation
reactions can be described by the law of mass action. The set of algebraic equations, describing the dissociation of
aqueous complex A* into components as species in solution can be written as:
j ] ]' x (12)
where K* is the equilibrium constant for the dissociation of the /ft aqueous complex into components as species in
solution, Tfisthe activity coefficient for the Ith aqueous complex and -fr is the activity coefficient for they'*1 component as
species in solution. Equation 12 can also be used for equilibrium-controlled homogeneous oxidation-reduction reactions.
Complexation and oxidation-reduction reactions can also be described as kinetically-controlled reactions. The reaction
stoichiometry for a kinetically-controlled intra-aqueous reaction can be expressed in terms of components as species in
solution [Lichtner, 1996b]
Nc
^mj tf J' a (1 3)
7 = 1
where vra are the stoichiometric coefficients of the species participating in the reaction, and Na defines the number of
kinetically-controlled intra-aqueous reactions.
The model includes a general formulation for kinetically-controlled reactions. The reaction rate for an intra-aqueous
reaction can be expressed as a function of the forward and backward rate constants kkaf and k*", the activities of the
components as species in solution and aqueous complexes ayc and a*, total aqueous component concentrations and the
equilibrium constant of the reaction. A rate expression in symbolic form for the kth kinetically-controlled intra-aqueous
reaction can be written as
where kakf and /c^faare the rate constants for the forward and backward reactions respectively. Mass loss or gain fora
particular component enters the global mass conservation equation through the source-sink term Qyaa, which is defined
by:
-------
J = l>*c (15)
1=1
where the contributions of the various reactions rates were scaled to express the source-sink term in the units of the
global mass conservation equations [mol L1 bulk s~1].
The model allows the consideration of ion-exchange reactions in terms of the Gapon- and Gaines-Thomas conventions
[Appelo and Postma, 1993]. A general stoichiometric relationship based on the Gaines Thomas model can be written as
]A] i = l,Na (16)
7=1
where v* is the stoichiometric coefficient of the ion-exchanged species A* in the Ith ion exchange reaction, which is
defined by the ratio of the charge of the two participating cations, v* are the stoichiometric coefficients of the components
and A/s defines the number of the ion-exchanged species. Based on the stoichiometry defined in reaction equation 16, the
law of mass action can be used to obtain a relationship that defines the activities of the ion-exchanged species Af in
terms of equivalent fractions
' = » i* (17)
where £ and &k are the activities of the ion-exchanged species Af and A* [meq meq-1] and Kf is the selectivity coefficient
for the Ith ion-exchange reaction. The actual concentrations of the ion-exchanged species can be obtained by applying
the conversion:
' 100 " " (18)
where Cf is the concentration of the ion-exchanged species Af [mol L1 bulk], pb is the dry bulk density of the porous
medium [g solid cnrr3 bulk] and CEC defines the cation exchange capacity [meq (1 OOg)-1 solid]. The factor 100 provides
the conversion [g (100g)-1].
A general stoichiometric relationship, which describes the congruent dissolution or precipitation of a mineral in terms of
components as species in solution can be written as:
l~ ' m (19)
where A defines the /th mineral, v are the stoichiometric coefficients of the components as species in solution
comprising the mineral Am, and Nm defines the number of minerals actively participating in dissolution-precipitation
reactions.
General rate expressions for surface- and transport-controlled dissolution-precipitation reactions can be written as a
function of the dissolution and precipitation rate constants (kkmdand kkmp), the diffusion coefficients of the primary reactant
species through a protective surface layer (Dkm), geometric parameters such as the representative mineral particle radius
(r/), the thickness of a protective surface layer (rfi, and the reactive mineral surface area (S). The rate expression will
depend in addition on the activities of the components as species in solution and aqueous complexes, total aqueous
component concentrations, and the equilibrium constant of the reaction (Kkm). The rate expression for the dissolution of
the k* mineral phase can be expressed in symbolic form as:
7~)ffl v*P vr C rtc nx rFa lfm\
The formulation also allows the consideration of parallel reaction pathways. The source-sink term describing the
production or consumption of components due to dissolution-precipitation reactions can be written as:
mRm i = \N (2D
-------
Conceptual Model
The conceptual model developed through field studies at the Elizabeth City site [Bennett, 1997; Blowes etal., 2000] was
subdivided into three zones: upgradient aquifer, the reactive barrier, and the aquifer downgradient of the barrier
(Figure 3). The zone located upgradient of the treatment system was termed the "contaminated zone." The contaminated
ground water in this zone has been in contact with aquifer material for a considerable amount of time, and dissolved
species are either in equilibrium with the aquifer material, or reactions are kinetically limited. Dissolved species are
transported more or less conservatively through this zone by advective and dispersive transport processes [Bennett,
1997].
The second zone is located within the reactive barrier and was termed the "treatment zone." In this zone, the
contaminants are transformed by reduction and immobilized by subsequent precipitation [Bennett, 1997; Blowes etal.,
2000]. A number of secondary reactions occur simultaneously in the treatment zone. Other oxidized species, such as
dissolved oxygen, nitrate, sulfate, dissolved inorganic carbon, dissolved organic carbon and water are reduced either
directly by zero-valent iron or by dissolved reduced species; for example, hydrogen gas [Bennett, 1997]. Isotopic data
indicates that sulfate reduction may be microbially-mediated. These reduction reactions lead to the corrosion of zero-
valent iron, which serves as the ultimate electron donor. Major reaction products of these corrosion reactions are ferrous
or ferric iron and dissolved gases. The combined effect of the reduction and corrosion reactions leads to a significant
increase of pH- values and causes the Eh of the water passing through the barrier to decrease [Bennett, 1997; Blowes
et al., 2000]. It appears likely that high pH conditions promote the precipitation of a number of secondary minerals
throughout the treatment zone. These reactions consume alkalinity and act to buffer further increases in pH [Bennett,
1997; Blowes et al., 2000]. The exsolution of dissolved gases may act to buffer the redox potential of the pore water.
The ground water leaving the treatment zone is characterized by low dissolved species concentrations and exhibits high
pH- and low En-conditions [Bennett, 1997]. However, pH and Eh are restored to near background values downgradient
of the reactive barrier, indicating that reactions with the native aquifer material are buffering the infiltrating high pH, low
Eh water. This zone is termed "buffer zone," since interactions with the native aquifer minerals will tend to re-establish a
new equilibrium condition. Important processes for pH- buffering may be the dissolution of aluminosilicate and clay
minerals and the sorption of silicic acid combined with the release of protons [Powell et al., 1995]. The desorption of
hydrogen ions from oxide and clay mineral surfaces may be an additional, possibly more important, pH- buffering
process. The reductive dissolution of oxides and oxy-hydroxides and possibly degassing may be responsible for the
observed En-increase.
The conceptual model is applicable for the zones located upgradient and within any zero-valent iron reactive barrier.
However, geochemical processes controlling pH- and En-buffering downgradient of the reactive barrier depend on the
site-specific mineralogy and, therefore, apply only to the Elizabeth City site.
Definition of Reaction Network
This section introduces the reaction stoichiometries, rate expressions and equilibrium constants of the reactions
considered in this study based on the conceptual model by Bennett [1997]. In general, the equilibrium constants were
taken from the database of MINTEQA2 [Allison et al., 1991], unless otherwise noted. All reactions considered can be
expressed in terms of the following 27 components: AI3+, Ca2+, Cr, CH4(aq), CO32-, CrO42-, Cr(OH)2+, DOC, Fe2+, Fe3+, H+,
H2(aq), H4Si04, HS-, K+, Mg2+, Mn2+, Na+, NH4+, NO3', O2(aq), SO42-, TCE, cis-1,2 DCE, VC, ethane and H2O.
Complexation Reactions
Table 1 lists all 79 aqueous complexes considered in this study along with the corresponding equilibrium constants
reported as dissociation constants.
Reduction-corrosion Reactions
The conceptual model considers the reduction of hexavalent chromium, of the chlorinated organic compounds, and of
other oxidized species dissolved in the ambient ground water. Due to its extreme reduction capacity, zero-valent iron
ultimately leads to the reduction of all electron acceptors [Bennett, 1997; Blowes etal., 2000] including dissolved oxygen
[MacKenzie etal., 1997], Mn(IV), Fe(lll), nitrate [Rahman and Agrawal., 1997; Cneng et al., 1997], sulfate, DIG, DOC
[Weathers et al., 1995, Orth and Gillham, 1996] and the solvent water itself [Reardon, 1995]. In the present study, it was
assumed that dissolved manganese occurs exclusively as Mn(ll); the reduction of Mn(IV) was, therefore, neglected.
Furthermore, it was assumed that ferric iron can occur as a reaction product of iron corrosion, if the half reaction for the
electron acceptor considered has a higher standard potential than the half reaction for the Fe2+/Fe3+ redox couple. For
example, Powell et al. [1995] reported that the corrosion of zero-valent iron in the presence of oxygen produces ferric
iron. Ferric iron reduction was excluded in this study.
-------
The reaction stoichiometries of all reduction-corrosion reactions were normalized with respect to zero-valent iron. The
degradation of the organic compounds by reductive elimination and hydrogenolysis was considered. It was assumed that
hydrogenolysis leads to the sequential degradation of TCE (C2HCI3) to cis-1 ,2 DCE (C2H2CI2), VC (C2H3CI), and ethane
(C2H6) [Bennett, 1997; Blowes et al., 2000]. Laboratory experiments [O'Hannesin et al., 1995] conducted using water
from the Elizabeth City site and zero-valent iron indicated that only 7% of TCE was degraded by hydrogenolysis to
cis-1 ,2 DCE. When determining the reaction stoichiometry of TCE-degradation, it was assumed that the remaining 93%
is directly degraded to ethane.
All reduction-corrosion reactions are assumed to be irreversible. The reaction rate of hexavalent chromium reduction by
zero-valent iron is characterized by a square root dependence on Cr(VI) and H+ and is proportional to the reactive
surface area of zero-valent iron [Gould, 1982]. For all other electron acceptors except water, it was assumed that the
reaction rate is first order with respect to the electron acceptor and proportional to iron surface area. Iron corrosion by
waterwas described by a rate expression with a first order dependence on iron surface area [Reardon, 1995]. Since this
reaction is not dependent on the concentration of the electron acceptor, it was assumed that the reaction rate
approaches zero, when equilibrium conditions are approached. An equilibrium constant of log ^Fe\s)HO = -1 1 .78 was
calculated based on data from Reardon [1995] and Stumm and Morgan [1996].
Secondary reactions between reduced reaction products (e.g., hydrogen gas, hydrogen sulfide, methane or ammonia)
and oxidized species (e.g., hexavalent chromium, dissolved oxygen, nitrate and sulfate) may lead to inhibitive or
competitive effects influencing the reaction progress of a particular reduction-corrosion reaction. For example, Siantaretal.
[1995] observed that the presence of oxygen or nitrite affected the degradation of pesticide by zero-valent iron, indicating
that inhibition or competition may play a role. Inhibitive or competitive effects are neglected here and reduction-corrosion
reactions are assumed to occur as parallel reactions.
Table 2 summarizes the reaction stoichiometries of the reduction-corrosion reactions considered. The total iron
corrosion rate can be estimated as the sum of all reduction-corrosion rates.
Formation of Secondary Minerals in Treatment Zone
The conceptual model also accounts for the precipitation of secondary minerals within the treatment system. The
reaction stoichiometries of the reduction-corrosion reactions in Table 2 imply a net pH- increase creating conditions
favorable for the precipitation of carbonate minerals and hydroxide mineral phases [Bennett, 1997; Blowes etal., 2000].
In addition supersaturated conditions were observed locally with respect to iron sulfide minerals [Bennett, 1997;
Blowes etal., 2000].
Examination of the core samples by scanning electron microscopy (SEM) and transmission electron microscopy (TEM)
identified the possible occurrence of amakinite (Fe(OH)2), goethite (a-FeOOH) orakagenite (B-FeOOH) [Palmer, 1999].
It was determined that Fe(OH)2 is the mineral phase most likely present as a secondary precipitate [Palmer, 1999].
Secondary carbonate and sulfide mineral phases could not be identified; however, this may be due to the low volume
fractions or the amorphous nature of these mineral phases.
The precipitation of secondary mineral phases Am is described by rate expressions based on transition state theory
[Lasaga, 1998]:
(22)
where k is an effective rate constant forthe dissolution of the mineral phase A, IAP!" is the ion activity product and K
defines the corresponding equilibrium constant. Reactions describing the formation of secondary mineral phases within
the treatment zone are summarized in Table 3. The equilibrium constant for Fe(OH)2(am) was taken from the database
of EQ3/EQ6 [Wolery et al., 1990].
pH- and En-buffering Downgradient of Barrier
pH -buffering
The water exiting the reactive barrier is characterized by alkaline conditions with pH- values ranging from 9 to 1 1
[Bennett, 1997, Blowes etal., 1999]. This strongly alkaline pore water is likely to interact with the native aquifer material
downgradient of the barrier. According to Puls et al. [1992] and Pa/mer[1999], the mineralogy of the aquifer consists
-------
primarily of aluminosilicate minerals. The primary minerals present in the aquifer sediments are quartz > albite >
sanidine > muscovite and kaolinite [Puts et al., 1992]. Powell et al. [1995] reported that the aquifer material from the
Elizabeth City site is capable of buffering the pH towards neutral conditions by the dissolution of aluminosilicate and clay
minerals. An additional source of acidity may be the sorption of silicic acid, originating from Si-mineral dissolution, onto
iron-oxide/hydroxide surfaces combined with the release of hydrogen ions [Powell et al., 1995].
In preliminary simulations, it was assumed that pH- buffering is due to the dissolution of kaolinite. The pH dependent rate
expression given by Carroll and l/l/a/tf?er[1990] was used. The field-observed pH- buffering could only be reproduced by
disallowing the precipitation of gibbsite and AI(OH)3(am) and by setting the reactive surface area for kaolinite to values
of 107m2 mineral surface perm3 bulk porous medium. The field data [Bennett, 1997; Blowesetal., 1999] is characterized
by low dissolved Al-concentrations downgradient of the barrier indicating that aluminosilicate dissolution is not as
pronounced or that gibbsite or AI(OH)3(am) precipitation occurs. The effect of sorption of silicic acid and subsequent
release of protons was not investigated here. However, a similar kaolinite reactive surface area would be necessary to
reproduce the observed pH- buffering, because sorbing silica is originating from the dissolution of the aluminosilicates.
Sequential extraction tests conducted by Palmer [1999] revealed that much of the extractable Al immediately
downgradient of the reactive barrier is present in amorphous form. This is not the case for the undisturbed aquifer
material upgradient of the barrier, where most of the extractable Al is crystalline. This indicates that weathering of Al-Si
mineral phases with the subsequent precipitation of an amorphous Al-bearing phase, possibly AI(OH)3(am), is taking
place.
It is likely that other processes also contribute to pH- buffering. Another possible explanation for the observed pH-
buffering is the desorption of hydrogen ions from oxide and clay mineral surfaces, which have been in contact with the
slightly acidic ambient ground water prior to installation of the reactive barrier. The desorption of hydrogen ions can be
described as [Stumm and Morgan, 1996]:
S-OH2+OH~ ^S-OH + H2O (23)
S-OH + OH- ^S-O~+H2O (24)
A surface complexation model can be used to describe this pH- buffering process. However, such a model is presently
not included in MIN3P. This process is, therefore, approximated as a release of hydrogen ions from a limited reservoir,
until the pH reaches circumneutral values. It should be noted that Pa/mer[1999] determined from extraction tests
significantly higher concentrations of exchangeable cations downgradient of the barrier than upgradient of the barrier
(= 1 meq/100 g soil). This may be caused by the desorption of H+-ions followed by the sorption of cations from solution.
Eh-buffering
The treated pore water exiting from the reactive barrier is extremely reducing due to the presence of reduced gaseous
species such as dissolved hydrogen gas, H2S(aq), ammonia and methane. Bennett [1997] and Blowes et al., [2000]
observed Eh-values locally lower than -500 mV. Reduced gaseous species may either degas or react with Mn- and Fe-
oxides and oxy-hydroxides contained in the aquifer material. Iron oxy-hydroxides are abundant in the native aquifer
mineral, while manganese oxides occur at lower concentrations [Puls, personal communication, 1998; Palmer, 1999].
Degassing was observed at the Elizabeth City site within the reactive barrier, but may not be an important process
downgradient of the treatment system. Appelo and Postma [1993] and Stumm and Morgan [1996] give redox half
reactions for the oxidation of the relevant dissolved gas species and the reduction of Mn- and Fe-mineral phases, which
can be combined to describe Eh-buffering reactions possibly occurring downgradient of a reactive barrier. The field data
[Bennett, 1997; Blowes etal., 2000] does not indicate the oxidation of H2S(aq) and methane, since sulfate concentrations
at the monitoring well located 1 m downgradient of the barrier are negligible, and methane concentrations remain high.
Data for ammonia was not available. These data indicate that these reactions are not taking place. Eh-buffering reactions
involving the consumption of H2S(aq), ammonia and methane are, therefore, neglected. However, Fe-and Mn-oxides
may undergo reductive dissolution when in contact with dissolved hydrogen gas emanating from the barrier. Assuming
that iron and manganese oxides can be represented by goethite and pyrolusite, overall reactions for the oxidation of
dissolved hydrogen gas combined with the reductive dissolution of the mineral phases can be written as:
FeOOH(s] + 2H+ + \H2(aq) ->Fe2+ + 2H2O (25)
MnO2 0) + 2H+ + H2 (aq) -> Mn2+ + 2H2O (26)
The equilibrium constant for the reductive dissolution of pyrolusite combined with the oxidation of H2(aq) (log
K= -43.9640) was calculated based on redox half reactions as reported by Stumm and Morgan [1996]. The appropriate
equilibrium constant for goethite (log K= -13.5940) was obtained from equilibrium constants in the MINTEQA2-database
-------
[Allison etal., 1991] and from Stumm and Morgan [1996]. These reactions are described by a transition state theory rate
expression of the form [Lasaga, 1998]:
( TAPm^\
(27)
which includes reactive surface area and allows the update of mineral reactivity with progressing dissolution.
The reductive dissolution of iron and manganese oxides according to equations 25 and 26 consumes acidity. This
additional pH- increase may enhance the depletion of the pH- buffer capacity of the aquifer. The reductive dissolution of
oxides and oxy-hydroxides will also lead to an increase in dissolved ferrous iron and manganese concentrations. It is
possible that these concentration increases will be controlled by the precipitation of siderite and rhodochrosite,
respectively. Hydroxides such as Fe(OH)2(am) and amorphous pyrocroite (Mn(OH)2(am)) precipitate only in alkaline
waters, and are, therefore, incapable of controlling dissolved iron and manganese concentrations downgradient of the
barrier.
Solution Domain and Model Parameters
One-dimensional and two-dimensional reactive transport analyses were conducted. The two-dimensional solution
domain is a vertical cross section through the center of the reactive barrier, as illustrated in Figure 4, and is aligned along
Transect 2 (Figure 2). The solution domain extends 4 m in the horizontal direction and 3.2 m in the vertical direction and
ground-water flow takes place from the left to the right. The domain contains the multi-level monitoring wells 21-25, which
have been installed upgradient, within and downgradient of the reactive barrier (Figure 2, Figure 4), [Bennett, 1997;
Blowes et al., 1999]. Each well contains 7 monitoring points providing a detailed description of the geochemical
composition of the ground water entering the solution domain. The approximate location of the 0.6 m thick reactive
barrier is indicated by the dashed vertical lines. One-dimensional simulations were carried out along the flowline carrying
the highest chromium concentrations towards the barrier. The field data indicates that this flowline follows a zone of
preferential flow and passes through monitoring points 21-5, 22-3, 23-3, 24-3 and 25-5 [Bennett, 1997; Blowes et al.,
2000], as indicated in Figure 4.
Spatial Discretization
The discretized solution domain for the two-dimensional simulation is shown in Figure 5. A discretization interval of 10 cm
in both vertical and horizontal directions was used. The discretization was refined to 5 cm in the horizontal direction within
and in the direct vicinity of the reactive barrier to facilitate a more accurate representation of rapid geochemical changes
within and downgradient of the treatment system. This discretization leads to 50 grid points in the horizontal direction and
33 grid points in the vertical direction (Figure 5). For the one-dimensional simulations, a discretization interval of 2.5 cm
was used within and in the vicinity of the treatment zone leading to a total number of 68 grid points.
Physical Parameters and Hydraulic Conductivity Distribution
Field measurements indicated a hydraulic gradient varying between 0.0011 and 0.0033 during the sampling intervals
[Bennett, 1997; Blowes et al., 1999]. For the modeling study it was assumed that a hydraulic gradient of 0.0022 can be
used to represent average flow conditions. The hydraulic head loss along the vertical cross section is negligible and
recharge is insignificant, because the ground surface is paved above the treatment system [Bennett, 1997]. The flow
system is modeled as a fully saturated system with no flow boundaries at the top and the bottom of the domain and first
type boundaries at the upgradient and downgradient boundaries. The upper portion of the aquifer was found to be less
hydraulically conductive than the underlying layers [Bennett, 1997; Blowes et al., 2000]. The field observations of
Puls et al. [1995] and Bennett [1997] indicate the presence of a highly conductive layer, which is located roughly 4.5-6.5
m below ground surface. Significant hydraulic conductivity variations lead to large differences in ground-water flow
velocities with depth [Bennett, 1997]. Constant head and rising head response tests showed hydraulic conductivity
values ranging from 1.2 x 10'5 to 1.9 x 10~4 m s~1 within the aquifer and ranged from to 1.2 x 10~7 to 2.3 x 10~3 m s~1 within
and in the vicinity of the reactive barrier [Bennett, 1997; Blowes et al., 2000]. Hydraulic conductivities have been
assigned to the two-dimensional solution domain based on the hydraulic response and the modeling analysis conducted
by Bennett [1997] (Figure 6 and Table 4). Locally isotropic conditions were assumed in this context.
For the one-dimensional simulations the hydraulic conductivity in the aquifer was estimated to be K = 8.1 x 10'5 m s~1,
while a hydraulic conductivity of K = 1.2 x 10~3 m s~1 was assigned to the treatment zone. High hydraulic conductivities in
the reactive barrier can be explained by a relatively loose packing of the treatment material as a result of the installation
procedure [Bennett, 1997]. Bennett [1997] estimated a porosity within the reactive barrier of approximately 0.43-0.62. It
-------
was assumed that the porosity of the barrier is § = 0.5 within the high hydraulic conductivity zones. A porosity of § = 0.38
was assumed to be representative for the aquifer and the low hydraulic conductivity zones in the barrier. Dispersive and
diffusive transport processes were neglected, since the problem can be characterized as advection-reaction dominated.
This simplification eliminates adverse effects of artificial dispersion at the upgradient side of the reactive barrier on the
simulation results.
Mineralogical Parameters
The volume fraction for zero-valent iron is ^oW = 0.5 and ^oW = 0.62 for porosities of fy = 0.5 and fy = 0.38,
respectively. Only qualitative information is available regarding the volume fractions of the minerals contained in the
aquifer [Puts et al., 1992]. The bulk of the aquifer material was assumed to be non-reactive. The dissolution of kaolinite
and other aluminosilicate minerals was excluded, because the preliminary simulations showed that these reactions are
not likely to be important for controlling pH and the geochemical composition of the pore water. The dissolution reactions
considered include pyrolusite and goethite, which are important for Eh-buffer reactions downgradient of the reactive
barrier. The volume fractions of these minerals have been estimated to allow at least a generic description of the
specified buffer reactions. The estimated mineral volume fractions are summarized in Table 5. Mineralogical analyses of
the aquifer material [Palmer, 1999] indicate that the goethite content was estimated correctly (measured: 5.5x10~4-1.4
x 10~3 m3 mineral nrv3 bulk), while the MnO2-content was overestimated (measured: 2.2 - 6.5 x 10~6 m3 mineral nrv3 bulk).
When conducting the simulations, there was no information available regarding the degree of surface protonation in the
sediment prior to installation of the reactive barrier. It was assumed that 50 mol H+ per m3 bulk porous medium are
available for desorption when in contact with the infiltrating high pH waters.
The largest uncertainty with respect to determining reaction rates can be attributed to reactive surface area estimates.
Reactive surface areas for the treatment material used at the field site are presented in Table 6 based on data reported
by Bennett [1997]. It is apparent that the specific reactive surface area is more than 500 times larger than the geometric
surface area calculated based on the average grain size dgo. Differences between geometric and reactive surface areas
may be explained by the large intragranular porosity of zero-valent iron. The density of the treatment material can be
calculated based on the bulk density from laboratory studies and is y = 4.77 g crrr3, which is much lower than average
literature values for native iron (y = 7.3-7.9 g cnrr3, [Klein and Hurlbut, 1993]. This deviation indicates that the treatment
material is characterized by a pronounced secondary porosity.
The reactive surface areas of goethite and pyrolusite were estimated, since no site-specific information was available
and are summarized in Table 7. All other minerals considered in this study are secondary minerals and effective rate
constants were used, which implicitly include reactive surface areas. The reactive surface areas of zero-valent iron,
goethite and pyrolusite are updated, as the minerals become depleted.
Boundary and Initial Chemical Composition of Ground Water
Geochemical data from Bennett[1997] was used to define the boundary and initial condition in the solution domain. The
chemical composition of water samples from monitoring well 21, which is located upgradient of the reactive barrier (see
Figures 2 and 4), was analyzed on temporal variability. It was found that the general geochemical composition of the
ground water remained constant over time. The data from November 1996 was used to describe the source
concentrations upgradient of the treatment system.
pH and Eh were taken from field measurements. The Eh was slightly increased for the sampling points 21-3, 21-5 and
21-7 to allow the determination of ammonia from nitrate based on the assumption of equilibrium for the NO3VNH4+ redox
couple. Total dissolved carbonate concentrations were obtained from field measured alkalinity [Bennett, 1997] using
MINTEQA2 [Allison etal., 1991]. Field measured total concentrations were used forCa2+, Cr, K+, H4SiO4, Mg2+, Mn2+, Na+,
NO3~, O2(aq) and SO42~ [Bennett, 1997]. The data for dissolved oxygen was not available for November 1996 and data
from June 1997 was used instead. Field measured concentrations from Bennett [1997] were also used forthe chlorinated
organic compounds TCE, cis-1,2 DCE, VC, ethane and methane. The field data for total dissolved organic carbon,
expressed in terms of CH2O [Bennett, 1997] was corrected forthe chlorinated organic compounds, which are considered
separately. Dissolved hydrogen gas concentrations were calculated based on pH and Eh. Puts etal. [1992] reported that
more than 98% of dissolved chromium is present as Cr(VI) in the contaminated ground water, since the reduction
capacity of the aquifer material with respect to hexavalent chromium is low. It, therefore, was assumed that total
dissolved chromium can be used to represent hexavalent chromium. This assumption is consistent with field-measured
hexavalent chromium concentrations, which coincide well with analytically determined total chromium concentrations
[Bennett, 1997]. Trivalent chromium concentrations were determined based on the assumption of equilibrium with
amorphous chromium hydroxide. Total ferrous and ferric iron concentrations were close to or below detection in the
upgradient portion of the aquifer and were determined based on the assumption of equilibrium with goethite and
equilibrium conditions for the Fe2+/Fe3+ redox couple. Total NH4+and HS-concentrations were not analyzed for and were
-------
calculated assuming equilibrium for the NO3VNH4+ and SO42VHS- redox couples, respectively. Dissolved aluminum
concentrations were determined by equilibrating the ground water with kaolinite.
The initial condition in the solution domain affects the simulation results only at early time. The chemical composition of
the water samples taken from monitoring point 21-7 was, therefore, used to describe the initial condition in the entire
aquifer.
Calibrated Rate Constants
Some of the reaction rates used in this study were determined in laboratory experiments using ground water from the
field site [O'Hannesin et al., 1995; Bennett, 1997; Blowes et al. 2000]. The rate constants, therefore, include influences
due to the interactions with other dissolved species and can be applied directly in the modeling study, provided that
laboratory conditions are representative for conditions encountered in the field. In other cases, laboratory-derived rate
constants are based on ideal, single component experiments. It may not be possible to use these rate constants, since
they do not account for complex interactions between reduced and oxidized species. Instead, a calibration procedure
has to be used to determine effective reaction rates. The calibration of these rate constants was conducted with the
objective to reproduce concentrations of dissolved reactant and product species similar to those observed in the field.
This approach is justified, because zero-valent iron acts as the ultimate electron donor in the system. However, the
application of the method is limited, because calibrated rate constants represent apparent rate constants and may vary
significantly depending on the specific geochemical composition of the ground water in contact with zero-valent iron.
These variations were neglected in the present study.
Preliminary simulations were conducted to approximately calibrate the model with respect to aqueous concentrations
observed in the field [Bennett, 1997; Blowes etal., 2000]. The calibration was carried outforthe one-dimensional solution
domain. Effects due to preferential flow or varying water chemistry were, therefore, not taken into account. The calibrated
rate constants were used for the one- and two- dimensional simulations presented here.
The simulated data was compared to field observations from February 1997 (240 days after completion of the barrier
installation). The calibration involved adjusting the rate constants for reduction-corrosion reactions, the reactive surface
area of zero-valent iron, and effective rate constants for the precipitation of secondary minerals. Table 10 shows a list of
the components and source and sink terms affecting the component concentrations. This table identifies the interactions
between the components and the effect of dissolution-precipitation reactions and serves as a basis for the calibration
procedure.
It was assumed that the laboratory-derived rate constants for reductive dechlorination of TCE, cis-1,2 DCE and vinyl
chloride [O'Hannesin et al, 1995; Bennett, 1997] are representative for the conditions at the site. These rate constants
appear to be most reliable, since the experiments were conducted with the treatment material used at the field site and
with the Elizabeth City ground water. To approximately match the field data, it was necessary to decrease the reactive
surface area for zero-valent iron listed in Table 6 by one order of magnitude. This difference may be attributed to scaling
from laboratory to field conditions. Possible reasons for this scaling include locally higher flow velocities in the field or
mixing of the treatment material with the native aquifer material during installation. The reactive surface area used in the
simulations was 3.88 x 105 [m2 mineral rrr3 mineral]. The rate constant for hexavalent chromium reduction was taken
directly from the laboratory study by Gould [1982]. The remaining rate constants were obtained by calibration in an
attempt to match the concentrations of the various electron acceptors and reaction products (Table 11). The table also
includes laboratory-derived rate constants for nitrate reduction and iron corrosion by water for comparative purposes.
Discrepancies between the calibrated and measured reaction rates will be addressed in a later section along with the
discussion of the simulation results.
Effective rate constants for the precipitation of secondary mineral phases were adjusted to approximately reproduce
dissolved Cr3+, Ca2+, Mg2+, Mn2+, Fe2+, Fe3+, HS- and CO32- concentrations and pH. The resulting rate constants are listed
in Table 12. Rate constants forthe reductive dissolution of goethite and pyrocroite were estimated. Since these minerals
are relatively insoluble, comparable slow reaction kinetics have been assumed. The rate constants are summarized in
Table 13.
Results and Discussion
In the following sections, the results of the one- and two-dimensional simulations are presented and discussed. The
simulations represent quasi-steady state conditions with respect to the dissolved contaminant concentrations. Quasi-
steady state conditions prevail because the volume fraction of the treatment material is large in comparison with the
contaminant and other electron acceptor concentrations entering the treatment system.
10
-------
One-dimensional Simulations
The simulations presented here are based on a simplified one-dimensional flow field. The effect of the complex flow
conditions characteristic for the site cannot be reproduced using this approach. On the other hand, two-dimensional
simulations enhance the complexity and make an interpretation of the geochemical data more difficult. The results of the
one-dimensional simulations conducted here are compared directly to field observations from February 1997, corre-
sponding to 240 days of barrier operation [Bennett, 1997; Blowes et al., 2000].
Removal of Contaminants
Figure 7a illustrates the removal of hexavalent chromium by the treatment system. The simulations show very rapid
reduction of hexavalent chromium. The results are not sensitive to the chromium reduction rate, since the time scale of
chromium removal is much shorter than the time scale of advective transport through the barrier. Trivalent chromium
concentrations never increase significantly, since the conditions in the barrier favor the precipitation of amorphous
chromium hydroxide [Bennett, 1997; Blowes et al., 2000]. The combination of the reduction and precipitation reactions
leads to low hexavalent and trivalent chromium concentrations within and downgradient of the reactive barrier, and are
in agreement with field measured total chromium concentrations, which are below the detection limit of 0.01 mg L1
[Bennett, 1997; Blowes etal., 2000].
The simulated and measured concentrations of the chlorinated organic compounds and ethane are depicted in Figure
7b. The model results show that concentrations of TCE, cis-1,2 DCE and VC decline by 1 -3 orders of magnitude across
the barrier, while ethane concentrations increase to values above the detection limit. These results are generally in
agreement with the field observations by Bennett [1997]. Within the barrier system, TCE is partly reduced to ethane and
cis-1,2 DCE. The simulated vinyl chloride concentrations increase temporarily due to the degradation of cis-1,2 DCE. As
indicated by the rate constants shown in Table 11, the reduction of the chlorinated organic compounds is not as rapid as
the reduction of chromium and the transformation remains incomplete. Primarily affected are the degradation products
cis-1,2-DCE and VC. The results agree reasonably well with the field data [Bennett, 1997; Blowes et al., 2000]. TCE-
concentrations downgradient of the barrier are below detection limit of 1 u,g/l, while cis-1,2-DCE and VC persist in low, but
measurable concentrations. The simulation significantly overpredicts ethane concentrations, suggesting that ethane is
not the degradation product, or that ethane is further degraded, possibly to inorganic carbon.
Reduction of Electron Acceptors
Figure 8 compares the computed concentrations for nitrate and sulfate to the field-measured concentrations. The trend
of the model results is in agreement with the field data. The concentration of the reaction products sulfide and ammonia
are also shown. Nitrate reduction is rapid and field measured concentrations fall below the detection limit in the entry
portion of the barrier [Bennett, 1997]. Field measured ammonia concentrations are not available; a mass balance
between nitrate and ammonia is, therefore, not possible. Sulfate concentrations decrease also by 1-2 orders of
magnitude. The reaction product hydrogen sulfide does not reach significant concentrations indicating that the
precipitation of mackinawite is taking place (see below), even though sulfide mineral phases could not be identified in
mineralogical analyses [Palmer, 1999].
Selected Cation Concentrations
Figure 9 compares computed concentrations for selected cations to analytical concentrations from Bennett [1997]. The
simulated results are consistent with the field observations, which show declining dissolved cation concentrations while
the pore water is flowing through the barrier. The decrease of cation concentrations indicates that secondary minerals
precipitate in the treatment system, as was discussed by Bennett[1997] and Blowes etal. [2000]. The simulated results
for calcium and magnesium compare well with the field data, except for the buffer zone downgradient of the barrier,
where magnesium concentrations are overpredicted. The lower field-measured concentrations may be due to surface
complexation reactions, involving the desorption of protons as a result of the infiltrating high pH water in conjunction with
the sorption of Mg. This hypothesis is supported by the elevated exchangeable Mg-concentrations observed downgradient
of the barrier [Palmer, 1999]. The solubility of manganese was assumed to be controlled by rhodochrosite. However, the
model results overpredict dissolved manganese concentrations in the treatment zone and downgradient of the barrier.
This indicates that Mn does not precipitate as rhodochrosite, but may co-precipitate with other carbonate minerals. The
pore water also becomes slightly supersaturated with respect to amorphous pyrocroite [Mn(OH)2(am), S/max = 0.5], which
may be an additional sink for dissolved manganese. Dissolved Mn-concentrations may also be limited by sorption or
ion-exchange reactions [Palmer, 1999]. The model results overpredicted dissolved iron concentrations slightly and
underpredicted carbonate concentrations (not shown). It is possible that Mg does not precipitate as a carbonate mineral
phase, but rather as brucite [Mg(OH)2, S/max = 1.2]. The pore water also reached supersaturated conditions with respect
to artinite [Mg2CO3(OH)2, S/max = 0.2]. If this is the case, dissolved carbonate becomes less depleted, while the
11
-------
precipitation of siderite may be more extensive. On the other hand, the precipitation of Ca-hydroxides, such as
portlandite (Ca(OH)2), is unlikely (S/max = -5.2), since the pH in the pore water is not sufficiently alkaline.
pH and En
Reduction-corrosion reactions taking place in the treatment zone lead to a pronounced pH- increase and a decrease of
the redox-potential of the ground water passing through the barrier [Bennett, 1997]. Figure 10 shows the results for pH
and Ehforthe one-dimensional simulation. The pH of the ground water upgradient of the reactive barrier is approximately
6.3 and rises in the wall to values up to 9-11. Within the treatment zone, the simulation results agree very well with field-
measured pH- values. The field data from Bennett [1997] show that pH- values downgradient of the barrier drop rapidly
and approach close-to-background values. It was assumed that pH- buffering is due to deprotonation from mineral
surfaces. The observed strong pH- buffering could only be reproduced by specifying a large reservoir of sorbed
hydrogen ions (50 mol rrr3 porous medium). Assuming that the observed increase in exchangeable cations downgradient
of the reactive barrier [Palmer, 1999] is due to desorption of H+ followed by sorption of dissolved cations, the amount of
hydrogen ions desorbed can be calculated as approximately 17 mol per m3 bulk porous medium (0= 0.38, pb= 1643 kg nr3).
This value agrees relatively well with the estimated H+ available for desorption and further supports that deprotonation
indeed may be an important pH- buffer process downgradient of the barrier.
It is likely that, if hydrogen gas is present in sufficient quantities, Eh measurements reflect the state of the H2(aq)/H+redox
couple. Therefore, the redox potential was computed based on dissolved hydrogen gas concentrations. The pore water
entering the barrier is characterized by Eh-values of approximately 400-500 mV. The simulation results show a rapid
decrease of Eh within the reactive barrier to values below -500 mV. Although the field-measurements used for this
comparison are less reducing than calculated by the model, Eh-values lower than -550 mV have been measured locally
in the treatment zone at other locations and sampling times [Bennett, 1997; Blowes etal., 1999]. Eh-values increase up
to close to background-values of 200 - 400 mV downgradient of the reactive barrier [Bennett, 1997] due to Eh-buffer
reactions. The reductive dissolution of goethite and pyrolusite overpredicts En-buffering in the downgradient zone
indicating that these reactions are characterized by slow reaction kinetics. Unlike pH- buffering, Eh-buffer reactions do
not require a large reservoir of buffer agents. This is due to the comparably low H2(aq)-concentrations leaving the
treatment zone and the apparent persistence of other reduced dissolved gases, such as methane and H2S(aq) (not
shown).
Discussion of Reaction Mechanisms
In order to approximately match the field data, the rate constant for nitrate reduction had to be increased by one order of
magnitude, while the iron corrosion rate by water had to be decreased by four orders of magnitude in comparison to
laboratory-derived rate constants (Table 11). Large uncertainties exist with respect to the applicability of the rate
constant for nitrate reduction [Rahman and Agrawal, 1997] to the reactive barrier at the Elizabeth City site, since the
tested treatment material was different, and the reported rate constant had to be corrected for reactive surface area. The
discrepancies with respect to the rate constant for nitrate reduction were, therefore, considered within the uncertainty of
the rate constant derived from the work of Rahman and Agrawal [1997].
The differences between the measured rate constant for iron corrosion by water [Reardon, 1995] and the corresponding
calibrated rate constant is more significant (Table 11). Modeled Eh-values (based on the H2(aq)/H+ redox couple) still fall
notably below the field measured values, despite the decreased rate constant for iron corrosion by water. The laboratory-
derived reaction rate from Reardon [1995] (0.7 mmol kg-1 Fe day1) was normalized with respect to reactive surface area.
In this context, it was assumed that the reactive surface area reported by Reardon [1995] is representative for the
investigated material, which was put into question in the original reference. The differences may be explained by the
significant rate of sulfate reduction at the field site in Elizabeth City (Figure 8). Sulfate reduction at the Elizabeth City site
may be microbially mediated, as proposed by Bennett [1997]. In this case, dissolved hydrogen gas may be used as the
electron donor. In the present modeling study, sulfate reduction was described as a heterogeneous reaction between
sulfate and zero-valent iron. However, the following reaction sequence may better describe the reaction mechanisms
controlling the geochemical conditions at the field-site:
4Fe° 0) + 8H2O^ 4Fe2+ + 8OH~ + 4H2 (aq) (28)
SO2' + 4H2 (aq) -> HS~ + OH~ + 3H2O (29)
4Fe2+ + HS~ + 1OH~ -> FeS(am) + 3Fe(OH)2 (am) + H2O (3°)
This reaction sequence leads to a net pH- increase, the removal of sulfate and dissolved hydrogen gas, and the
immobilization of the reaction products iron and sulfide. Nevertheless, it is also possible that sulfate is simultaneously
reduced by the direct interaction with zero-valent iron (Table 2). Hydrogen gas may also be consumed as an electron
12
-------
donor in other reduction reactions (nitrate, DIG and DOC, Lovleyand Goodwin, 1988). These reactions may explain the
apparent iron corrosion rates by water. Microbially-mediated reduction reactions in porous media composed of zero-
valent iron and involving hydrogen gas as the electron donorwere previously reported by Weathers etal. [1995]. It is not
possible to uniquely describe the contributions of the various reaction processes.
Corrosion of Zero-valent Iron
The simulation results indicate that iron corrosion is most significant in the entry area of the barrier and decreases along
the flow path, when the electron acceptors become depleted. The reaction rates of selected reduction-corrosion
reactions are shown in Figure 11. Nitrate and sulfate are the most important electron acceptors in the system, which is
in agreement with the interpretation of Bennett [1997]. The reduction of hexavalent chromium also contributes to iron
corrosion. The remaining electron acceptors are lumped together, since their contributions are comparably small.
Figure 11 shows that chromium reduction only takes place in the entry area of the reactive barrier, since the reaction rate
is fast in comparison to ground water velocities. The reduction of nitrate to ammonia is slower, and persists deeper into
the barrier. The reduction of sulfate is characterized by even slower reaction kinetics. Reduction rates decrease
continuously throughout the barrier, but are still discernable before leaving the treatment system into the downgradient
portion of the aquifer. Figure 11 illustrates that the major contribution to iron corrosion is ultimately due to the reduction
of sulfate. Iron corrosion due to the reduction of the chlorinated organic compounds is not explicitly presented in
Figure 11, since the corrosion rates are negligible on the scale of the graph.
Precipitation of Secondary Minerals
Figure 12 shows that the precipitation of carbonate minerals, such as calcite and siderite, takes place close to the
upgradient end of the barrier. Siderite is the dominant carbonate phase, because the corrosion of zero-valent iron
liberates large amounts of ferrous iron. Small amounts of rhodocrosite also precipitate in the entry area of the barrier (not
shown). It was assumed that the precipitation of a Ca-Mg-carbonate mineral phase controls aqueous Mg-concentrations.
The formation of this mineral phase takes place throughout the barrier due to slow reaction kinetics. The real system will
likely be characterized by the precipitation of a complex carbonate solid solution containing ferrous iron, calcium,
magnesium and manganese rather than by the formation of distinct mineral phases [Reardon, 1995]. Ferric iron
produced from iron corrosion also precipitates rapidly in the entry zone. On the other hand, minerals such as
mackinawite and ferrous hydroxide precipitate throughout the treatment system. The formation of these mineral phases
is controlled by the availability of the reactants, which are produced by sulfate reduction and iron corrosion by water.
These results are consistent with the conceptual model of Bennett [1997]. Mineralogical analyses support the formation
of secondary Fe(OH)2 and possibly a-FeOOH [Palmer, 1999]. Carbonate and sulfide mineral phases were not found by
Pa/mer[1999]. This may be due to the small volume fractions of these secondary precipitates (core samples were taken
only two years after barrier installation), or due to the amorphous nature of these phases. However, mass balance
considerations based on field-measured aqueous concentrations [Bennett, 1997; Blowes etal., 2000] suggest that these
mineral phases do form.
Long Term Efficiency
The tendency to locally precipitate relatively large amounts of secondary minerals overextended time periods may have
an impact on the long term efficiency of a reactive barrier [MacKenzie etal., 1997; Bennett, 1997]. Significant amounts
of the treatment material may also be consumed due to the combined effect of sulfate reduction and iron-corrosion by
water. Treatment material depletion in conjunction with secondary mineral formation may also affect the porosity of the
treatment zone [MacKenzie etal., 1997; Bennett, 1997].
Within the reactive barrier, the bulk porous medium consists at any time of porosity (void space), the volume fraction of
the treatment material and the sum of the volume fractions of all secondary mineral phases. Figure 13 illustrates the
potential for depletion of the treatment material and the precipitation of secondary minerals along the barrier after 5, 10
and 20 years of operation. Secondary minerals are not present initially and the porosity is 0= 0.5 which is equal to the
volume fraction of zero-valent iron. After 20 years the volume fraction of zero-valent iron has decreased from 0.5 to
approximately 0.42 in the entry zone of the treatment system. However, the zone of significant treatment material
depletion is limited to the first 10 cm of the barrier. At the same time, it can be observed that the porosity decreased over
the 20 year simulation period from 0 = 0.5 to approximately 0 = 0.36 in the entry area of the barrier, indicating that a
significant amount of secondary minerals has precipitated. The total volume fraction of secondary minerals can be
estimated from the decrease of porosity and the depletion of the treatment material and amounts to a maximum value of
(f> = 0.22 in the entry area of the reactive barrier. These results are comparable with porosity loss calculations performed
by Bennett [1997] and Blowes et al. [2000].
13
-------
The model results indicate that over a long period of time porosity may decrease significantly, which will almost certainly
affect the hydraulic properties of the treatment system. More importantly, the reactivity of the treatment material may
decline overtime. In addition to the consumption of Fe°, the accumulation of secondary mineral phases may compromise
the reactivity of the remaining treatment material. It can be hypothesized that the reduction of the reactivity of the
treatment material in the entry zone due to the precipitation of secondary minerals will allow the contaminants and other
electron acceptors to pass this less reactive zone more or less unaffected. Reduction reactions will still occur in areas
located deeper into the barrier, that have been less affected by the precipitation of secondary minerals. However, this
process will decrease the effective thickness of the barrier and, therefore, the contact time of the contaminants with the
treatment material. This may lead to the incomplete treatment of contaminants that require a long residence time.
Although the potential for mass loss and secondary mineral formation has been identified previously [e.g., MacKenzie et al.,
1997; Bennett, 1997], the simulation results illustrate the distribution of iron-consumption and secondary mineral
precipitation in a semi-quantitative way. In a real system, it can be expected that the depletion of the treatment material
and the precipitation of secondary minerals is less concentrated in the inflow area than predicted in Figure 13, because
decreasing iron reactivity was not accounted for in this study.
Two-dimensional Simulations
Two-dimensional simulations were carried out with input concentrations defined in Tables 8 and 9. The remaining
geochemical parameters were as used in the one-dimensional simlations. These simulations illustrate the effect of
preferential flow on the treatment of the contaminants. All results of the following reactive transport simulation represent
conditions after 2 years of barrier operation.
Ground-water Flow
Figure 14 shows the streamlines constructed from the steady-state velocity field through the two-dimensional solution
domain, which was calculated based on the hydraulic conductivity distribution presented in Figure 6 and is based on
earlier work from Bennett [1997]. The model results indicate a heterogeneous flow field in the aquifer and through the
reactive barrier. A zone of preferential flow exists in the reactive barrier at a depth of approximately 6 m below ground
surface, which is consistent with the findings of Puts et al. [1995] and Bennett [1997].
Removal of Contaminants
The two-dimensional simulation clarifies the effect of the heterogeneous ground-water flow field on the treatment of
contaminants. Figure 15 illustrates that the remediation of hexavalent chromium appears to be unaffected by the zones
of preferential flow within the aquifer and the reactive barrier. The rapid reduction of hexavalent chromium ensures a
successful treatment, even in zones of high flow velocities. Amorphous chromium hydroxide precipitates in a narrow
fringe in the entry area of the treatment system.
Figure 16 shows the concentration distributions of the chlorinated organic compounds for the two-dimensional
simulation. The concentration distribution upgradient of the reactive barrier does not coincide well with the concentration
distribution of most inorganic compounds (e.g., chromium). This deviation may be due to the infiltration of TCE as a free
phase product and subsequent dissolution in the source area [Bennett, 1997]. The simulation results indicate that zones
of preferential flow may have a significant impact on the treatment of the organic compounds. The removal of the
organics in areas of lower hydraulic conductivity are more pronounced, because of the longer contact time, cis-1,2 DCE
and VC are treated to below MCL values (2u,g/l VC and 7 u,g/l cisl ,2 DCE) in the upper portion of the solution domain,
while the reductive dechlorination of TCE (MCL 5u,g/l) in the lower portion of the domain leads to the production of these
degradation products, which are only partially removed. The simulation results for chromium and the chlorinated
organics agree reasonably well with field measured concentrations [not shown, see Bennett, 1997 or Blowes et al.,
2000].
pH and Eh
Figure 17 presents the results for pH and Eh. The effect of the heterogeneous flow field on these parameters can be
clearly observed. pH- values increase more slowly within the reactive barrier in the zone of preferential flow. The pH-
distribution is also affected by chemical heterogeneities. The concentrations of the electron acceptors entering the
reactive barrier vary with depth, and higher pH- values can be correlated to high infiltrating electron acceptor
concentrations. For example, sulfate concentrations are low in the upper and lower portion of the solution domain
(Figure 18), resulting in a less pronounced pH- increase. This behavior may be an artifact of the decreased rate
constants for iron corrosion by water, which were calibrated for conditions where sulfate is present, and may not be valid
for conditions where sulfate is depleted. Elevated pH- values downgradient of the barrier can be observed in areas of
high flow velocities, where the pH- buffer capacity of the aquifer is exhausted.
14
-------
The redox potential decreases rapidly within the reactive barrier. The specified reaction network does not allow a
consistent description of Eh-values, since hydrogen gas concentrations are assumed to be independent of the
concentrations of electron acceptors such as nitrate and sulfate, which may be reduced with dissolved hydrogen gas as
the electron donor. In areas where these electron acceptors are abundant, dissolved hydrogen gas concentrations may
be lower than predicted by the simulations, while conditions may be more reducing in zones where the electron
acceptors have been depleted. The Eh-increases downgradient of the barrier due to reductive dissolution of goethite and
pyrolusite to close-to-background values.
Sulfate Reduction
Figure 18 illustrates the reduction of sulfate by the treatment system. Sulfate enters the reactive barrier primarily through
the area of preferential flow, where it is reduced to dissolved hydrogen sulfide. However, hydrogen sulfide concentrations
remain low, because mackinawite controls the solubility of HS-. Low concentrations of HS- are observed in selected
zones located downgradient of the barrier. The model results indicate that the precipitation of mackinawite is focused on
the zone of preferential flow. Secondary mineral precipitation may affect the hydraulic properties of this high permeability
zone and may alter the flow distribution in the long term.
Conclusions
The conceptual model developed by Bennett [1997] was implemented into the numerical model MIN3P to describe the
interactions between reaction and transport processes taking place at the reactive barrier in Elizabeth City, North
Carolina. The simulation results agree well with the field observations. Processes downgradient of the barrier could only
be investigated in a qualitative way, since data was not available and surface complexation reactions are not included in
the present model.
The model allows visualization of important processes occurring in a reactive barrier composed of zero-valent iron.
Processes such as the corrosion-reduction reactions and secondary mineral precipitation can be evaluated semi-
quantitatively. The simulations provide estimates of iron corrosion rates from the sum of the reduction-corrosion
reactions. The total accumulation and distribution of secondary precipitates can be estimated. Two-dimensional
simulations highlight possibly adverse effects of preferential flow on contaminant treatment.
Discrepancies between the simulated Eh conditions within the barrier and the field observations by Bennett [1997]
illustrate that secondary reactions between the reduced reaction products, such as hydrogen gas, and the electron
acceptors, which enter the barrier, may be important. The hypothesized reduction of sulfate by hydrogen gas may
explain observed relatively low hydrogen gas concentrations. Degassing and the formation of hydrogen gas bubbles,
which may affect the permeability and reactivity of the barrier, may be inhibited by this process.
More work is warranted to investigate reactive barriers for ground-water remediation and to study the long term
performance of passive treatment systems. Possible model enhancements include the implementation of a more
complete reaction network including secondary redox reactions, inhibition effects, surface complexation and effects of
secondary mineral precipitation on the permeability and reactivity of the treatment material.
15
-------
References
Allison, J. D., Brown, D. S. and Novo-Gradac, K. J. 1991. MINTEQA2/PRODEFA2, A geochemical assessment model
for environmental systems: Version 3.0 User's Manual. U.S. Environmental Protection Agency, EPA/600/3-91/021.
Appelo, C. A. J., and Postma, D. 1993. Geochemistry, groundwater and pollution, A. A. Balkema, Rotterdam,
Netherlands.
Bear, J. 1972. Dynamics of fluids in porous media, Elsevier Science, New York.
Bennett, T. A. 1997. An in-situ reactive barrier for the treatment of hexavalent chromium and trichloroethylene in
groundwater, M.Sc. thesis, University of Waterloo, Waterloo, Ontario, Canada.
Blowes, D. W., Ptacek, C. J. and Jambor, J. L. 1997. In-Situ remediation of Cr(VI)-contaminated groundwater using
permeable reactive walls: Laboratory studies, Environ. Sci. Technol., 31(12), 3348-3357.
Blowes, D. W., Gillham, R. W. and Ptacek, C. J., Puls, R. W. 2000. An in-situ permeable reactive barrier for the treatment
of hexavalent chromium and trichloroethylene in groundwater: Vol. 2. Performance monitoring, January, EPA/600/R-
99/095b.
Carroll, S. A., and Walther, J. V. 1990. Kaolinite dissolution at 25°, 60°, and 80°, Am. J. Sci., 290, 797-810.
Cheng, I. F., Muftikian, R., Fernando, Q. and Korte, N. 1997. Reduction of nitrate to ammonia by zero-valent iron, in
Proc. 213th ACS National Meeting, Vol. 37, No. 1,165-166, San Francisco, CA, American Chemical Society, Division
of Environmental Chemistry.
Gould, J. P. 1982. The kinetics of hexavalent chromium reduction by metallic iron, Water Res., 16, 871-877.
Kirkner, D. J., and Reeves, H. 1988. Multicomponent mass transport with homogeneous and heterogeneous chemical
reactions: Effect of the chemistry on the choice of the numerical algorithm 1. Theory, Water Resour. Res., 24(10),
1719-1729.
Klein, C., and Hurlbut Jr., C. S. 1993. Manual of Mineralogy (21 ed.), John Wiley & Sons, New York.
Lasaga, A. C. 1998. Kinetic Theory in the Earth Sciences, Princeton University Press, Princeton, New Jersey.
Lichtner, P. C. 1985. Continuum model for simultaneous chemical reactions and mass transport in hydrothermal
systems, Geochim. Cosmochim. Acta, 49, 779-800.
Lichtner, P. C. 1996a. Continuum formulation of multicomponent-multiphase reactive transport, in Reactive Transport in
Porous Media, edited by P. C. Lichtner, C. I. Steefel, and E. H. Oelkers, Vol. 34 of Reviews in Mineralogy,
Mineralogical Society of America, Washington, DC.
Lichtner, P. C. 1996b. Modeling of reactive flow and transport in natural systems, in Proceedings of the Rome Seminar
on Environmental Geochemistry, 5-72, Castelnuovo di Porto, May 22-26.
Lovley, D. R., and Goodwin, S. 1988. Hydrogen concentrations as an indicator of the predominant terminal electron-
accepting reactions in aquatic sediments, Geochim. Cosmochim. Acta, 52, 2993-3003.
MacKenzie, P. D., Sivavec, T. M. and Horney, D. P. 1997. Mineral precipitation and porosity losses in iron treatment
zones, in Proc. 213th ACS National Meeting, Vol. 37, No. 1, 154-157, San Francisco, CA, American Chemical
Society, Division of Environmental Chemistry.
Mayer, K.U., 1999. A numerical model for multicomponent reactive transport in variably saturated porous media. Ph.D.
thesis, Department of Earth Sciences, University of Waterloo, Waterloo, Ontario.
O'Hannesin, S. F., Hanton-Fong, C. J., Blowes, D. W., Gillham, R. W., and Ptacek, C.J. 1995. Remediation of
groundwater contaminated with chromium and TCE using reactive barriers: Laboratory batch and column testing.
Progress Report II for EPA - R.S. Kerr Environmental Research Lab, Waterloo Centre for Groundwater Research,
University of Waterloo, Waterloo, Ontario, June.
Orth, W. S., and Gillham, R. W. 1996. Dechlorination of trichloroethene in aqueous solution using Fe°, Environ. Sci.
Technol., 30(1), 66-71.
Palmer, C. D. 1999. Groundwater remediation using reactive barriers, Elizabeth City, NCsite, Final Report to University
of Waterloo, Department of Geology, Portland State University.
Powell, R. M., Puls, R. W., Hightower, S. K. and Sabatini, D. A. 1995. Coupled iron corrosion and chromate reduction:
Mechanisms for subsurface remediation, Environ. Sci. Technol., 29, 1913-1922.
16
-------
Puls, R. W., Clark, D., Paul, C. J., and Vardy, J. 1992. Characterization of chromium contaminated soils and ground
water in the vicinity of a chrome plating operation, in Subsurface Restoration Conference. Third International
Conference on Ground Water Quality Research, 180-182, Dallas, Texas.
Puls, R. W., Powell, R. M., and Paul, C. J. 1995. In situ remediation of ground water contaminated with chromate and
chlorinated solvents using zero-valent iron: A field study, in Proc. 209th ACS National Meeting, Vol. 35, No. 1, 788-
791, Anaheim, CA, American Chemical Society, Division of Environmental Chemistry.
Puls, R. W., Rees, T. F., Lindley, P. M., and Paul, C. J. 1992. Surface analysis of chromium contaminated soils using
scanning electron microscopy with energy dispersive X-ray (SEM-EDS) and secondary ion mass spectroscopy
(SIMS), in Subsurface Restoration Conference. Third International Conference on Ground Water Quality Research,
125-127, Dallas, Texas.
Rahman, A., and Agrawal, A. 1997. Reduction of nitrate and nitrite by iron metal: Implications for groundwater
remediation, in Proc. 213th ACS National Meeting, Vol. 37, No. 1,157-159, San Francisco, CA, American Chemical
Society, Division of Environmental Chemistry.
Reardon, E. J. 1995. Anaerobic corrosion of granular iron: Measurement and interpretation of hydrogen evolution rates,
Environ. Sci. Techno!., 29(12), 2936-2945.
Sevougian, S. D., Schechter, R. S., and Lake, L. W. 1993. Effect of partial local equilibrium on the propagation of
precipitation/dissolution waves, Ind. Eng. Chem. Res., 32, 2281-2304.
Siantar, D. P., Schreier, C. G., and Reinhard, M. 1995. Transformation of the pesticide 1,2-dibromo-3-chloropropane
(DBCP) and nitrate by iron powder and by H2/Pd/AI2O3, in Proc. 209th ACS National Meeting, Vol. 35, No. 1, 745-748,
Anaheim, CA, American Chemical Society, Division of Environmental Chemistry.
Steefel, C. I., and Lasaga, A. C. 1994. A coupled model for transport of multiple chemical species and kinetic
precipitation/dissolution reactions with application to reactive flow in single phase hydrothermal systems, Am. J. Sci.,
294, 529-592.
Stumm, W., and Morgan, J. J. 1996. Aquatic chemistry, John Wiley & Sons, New York.
Linger, A. J. A., Sudicky, E. A., and Forsyth, P. A. 1995. Mechanisms controlling vacuum extraction coupled with air
sparging for remediation of heterogeneous formations contaminated by dense non-aqueous phase liquids, Water
Resour. Res., 31(8), 1913-1925.
Weathers, L. J., Parkin, G. F., Nivak, P. J., and Alvarez, P. J. J. 1995. Methanogens couple anaerobic Fe(0) oxidation
and CHCI3 reduction, in Proc. 209th ACS National Meeting, Vol. 35, No. 1, 829-831, Anaheim, CA, American
Chemical Society, Division of Environmental Chemistry.
Wolery, T. J., Jackson, K. J., Bourcier, W. L., Bruton, C. J., Viani, B. E., Knauss, K. G., and Delany, J. M. 1990. Current
status of the EQ3/6 software package for geochemical modelling, in Chemical Modeling of Aqueous Systems II,
edited by D. Melchior, and R. Bassett, Chapters, 105-116. Washington, D.C., USA, American Chemical Society.
17
-------
Tables
18
-------
Table 1. Complexation Reactions and Equilibrium
Reaction
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
(21)
(22)
(23)
(24)
(25)
(26)
(27)
(28)
(29)
(30)
(31)
(32)
(33)
OH-
H3SiOj
H2SiOl-
NH3(aq)
NH4SOJ
MgOH+
MgC03(aq)
MgHCO+
MgS04(aq)
CaOH+
CaHCOj
CaCO3(aq)
CaSO4(aq)
CaHSOj
NaCOj
NaHCO3(aq)
NaSOj
KSOJ
A10H2+
A1(OH)+
A1(OH)4
A1SOJ
A1HS01+
A1(S04)2
Al(OH)3(aq)
FeOH+
Fe(OH)^
FeSO4(aq)
FeHSOj
FeHCOj
FeCO3(aq)
Fe(OH)2(aq)
FeOH2+
Constants
^ H2O - H+
^ H4SiO4 - H+
^ H4SiO4 - 2H+
^ NHJ - H+
^ NHJ + SOl"
^ Mg2+ + H2O -
^ Mg2+ + coi-
^ Mg2+ + C0§-
^ Mg2+ + SOI"
^ Ca2+ + H2O -
^ Ca2+ + CO2" -
-^ Ta2+ -i- rn2"
T v^a T ^^3
-^ rn2+ -i- c:n2^
T v^a T ovj4
^H+
+ H+
H+
f H+
^ Ca2+ + SOI" + H+
^ Na+ + CO^-
^ Na+ + CO3- 4
^ Na+ + SOl~
^ K+ + SOl"
^ A13+ + H2O -
^ A13+ + 2H2O -
^ A13+ + 4H2O -
^ A13+ + SOl"
^ A13+ + SOI" 4
^ A13+ + 2SO1-
^ A13+ + 3H2O -
^ Fe2+ + H2O -
^ Fe2+ + 3H2O -
^ Fe2+ + SOl~
^H+
H+
^2H+
^4H+
~H+
^3H+
H+
^3H+
^ Fe2+ + SOI" + H+
^ Fe2+ + COfr H
^ Fe2+ + 01"
^ Fe2+ + 2H2O -
^ Fe3+ + H2O -
hH+
^2H+
H+
log Kf
-13.9980
-9.8300
-23.0000
-9.2520
1.1100
-11.4400
2.9800
11.4000
2.3700
-12.7800
11.4400
3.2200
2.3090
3.0680
1.2680
10.0800
0.7000
0.8500
-4.9900
-10.1000
-22.7000
3.5000
2.4480
5.0000
-16.9000
-9.5000
-31.0000
2.2500
3.0680
12.3300
4.3800
-20.5700
-2.1900
19
-------
Table 1. Complexation Reactions and Equilibrium Constants - continued
Reaction
log Kf
(34)
(35)
(36)
(37)
(38)
(39)
(40)
(41)
(42)
(43)
(44)
(45)
(46)
(47)
(48)
(49)
(50)
(51)
(52)
(53)
(54)
(55)
(56)
(57)
(58)
(59)
(60)
(61)
(62)
(63)
(64)
(65)
(66)
FeSO|
FeHSOl+
FeCl2+
FeCl+
FeClJ
FeCl3(aq)
FeOHj
Fe(OH)3(aq)
FefOH)^
Fe(S04)2-
Fe2(OH)f
Fe3(OH)f
MnCl+
MnCl2(aq)
MnClJ
MnOH+
Mn(OH)3-
MnCO3(aq)
MnSO4(aq)
Mn(N03)2(aq)
MnHCOj
HCO-^
H2CO3(aq)
HSOJ
H2S(aq)
s2-
Cr3+
CrOH2+
Cr(OH)3(aq)
Cr(OH)j
Cr02-
CrCl2+
CrClJ
^ Fe3+ + SOl~
^ Fe3+ + SOl" + H+
^ Fe3+ + Cl"
^ Fe2+ + Cl"
^ Fe3+ + 2C1-
^ Fe3+ + 3C1"
^ Fe3+ + 2H2O - 2H+
^ Fe3+ + 3H2O - 3H+
^ Fe3+ + 4H2O - 4H+
^ Fe3+ + SOl"
^ Fe3+ + 2H2O - 2H+
^ Fe3+ + 4H2O - 4H+
Mn2+ + Cl"
^ Mn2+ + 2C1"
Mn2+ + 3C1"
Mn2+ + H2O - H+
^ Mn2+ + 3H2O - 3H+
^ Mn2+ + CO2-
^ Mn2+ + SOl"
Mn2+ + 2NO^
^ Mn2+ + CO2,- + H+
^ H+ + CO^-
^ 2H+ + CO3-
^ H+ + SOl"
^ HS" + H+
^ HS- - H+
^ CrfOH)^ + 2H+ - 2H2O
^ Cr(OH)2f + H+ - H2O
^ CrfOHJof - H+ + H2O
^ Cr(OH)2f - 2H+ + 2H2O
^ CrfOH)^ - 2H+
^ CrfOH)^ + Cl" + 2H+ - 2H2O
^ Cr(OH)2f + 2C1" + 2H+ - 2H2O
4.0400
4.4780
1.4800
0.1400
2.1300
1.1300
-5.6700
-12.5600
-21.6000
5.3800
-2.9500
-6.3000
0.6070
0.2500
-.3050
-10.5900
-34.8000
4.9000
2.2600
0.6000
12.2800
10.3300
16.6810
1.9870
6.9940
-12.9180
9.6200
5.6200
-7.1300
-18.1500
-17.7456
9.3683
8.6580
20
-------
Table 1.
Complexation Reactions and Equilibrium Constants - continued
Reaction
log Kf
(67)
(68)
(69)
(70)
(71)
(72)
(73)
(74)
(75)
(76)
(77)
(78)
(79)
CrOHCl2
CrNOi+
CrSOj
CrOHSO
(aq)
4(aq)
Cr2(OH)2SOl+
Cr2(OH)2(S04)2(aq)
HCrO4-
H2Cr04(aq)
Cr-jO?"
CrO3Cl-
Cr03SOl
NaCrOj
KCrOj
^ Cr(OH)2f
^ Cr(OH)2f
^ Cr(OH)2f
^ Cr(OH)2f
^ 2Cr(OH)
^ 2Cr(OH)
^ CrOl" +
^ CrOl" +
^ 2CrO4~ -
- + 2C1-
- + N03-
- + sol~
+ H+^H
2o
+ 2H+ - 2H2O
+ 2H+^
2H2O
+ SOl~ + H+ - H2O
J + sol
^ + 280
H+
2H+
f2H+^
- + 2H+ -
1" + 2H+
H2O
^2H20
^2H20
^ CrOl" + Cl- + 2H+ - H2O
_
^ CrOl- +
so2- +
2H+ - H2
O
^ Na+ + CrOl"
^ K+ + CrOl-
2,
8,
10,
8,
16,
17,
6,
5.
14.
7.
8.
0.
0.
.9627
.2094
.9654
.2754
.1550
.9288
.5089
,6513
,5571
,3086
,9937
,6963
,7990
Table 2.
Reaction Stoichiometries of Reduction-corrosion Reactions
Oxidant
Reaction
Cr(VI)
TCE
Fe
Fe
°i
°i
(s)
(s)
+ CrOl" + 6H+ -
+ 0.3025 C2HC13
-> Fe3+ + ^(OH)^ + 2H2O
+ 1.2325H+ ->
Fe2+ + 0.07 C2H2C12 + 0.2325 C2H6 + 0.7675 Cl"
cis-1,2 DCE
VC
oxygen
nitrate
sulfate
DOC
DIG
water
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe
°i
°i
°i
°i
°i
°i
°i
°i
(s)
(s)
(s)
(s)
(s)
(s)
(s)
(s)
+ C2H2C12
2 2 3
+ |02(aq) -
+ f NO;- +
+ \sol~ +
+ |CH2O -\
+ |CO3" +
+ H+
+ §H-
+ 3H+
45H+
|H+^
^2H+
^|H+
+ 2H+ -> Fe2+ +
^Fe2+4
f -)> Fe2+
-)> Fe3+ -
^Fe3+ +
^Fe2+ +
^Fe2+4
^Fe2+ +
H2(aq)
- C2H3C1 + C
_i 1 /-I TT _i_ 1
+ 2^2tt6 + 2
f |H20 + H2
3NH| + |H
i
rCl"
'(aq)
2o
ITTQ , TT f~\
|Jib + n2U
-|CH4(aq) +
|CH4(aq) +
|H2O
|H20
21
-------
Table 3.
Table 4.
Secondary Minerals in Reactive Barrier and Corresponding Equilibrium Constants
Reaction
logK
Fe(OH)2(am)
Fe(OH)3(am)
Cr(OH)3(am)
CaCO3(s)
CaMg(C03)2(s)
FeC03(s)
MnCO3(s)
FeS(am)
^ Fe2+ 4
=i Fe3+ 4
^ Cr3+ 4
^ Ca2+ -
^ Ca2+ -
^ Fe2+ 4
^ Mn2+
^ Fe2+ 4
- 2H20 - 2H+
- 3H2O - 3H+
- 3H2O - 3H+
hCO2-
h Mg2+ + 2CO23-
-COi"
+ G023-
- HS- - H+
-13.9045
-4.8910
0.7500
8.4750
17.0900
10.4500
10.4100
4.6480
Physical Parameters for Aquifer and Reactive Barrier Material
parameter
unit
hydraulic conductivity (1-D)
hydraulic conductivity (2-D)
hydraulic gradient (average)
porosity (aquifer)
porosity (reactive barrier)
[m s l]
[m s^1]
H
H
H
8.1
1.2-
10"5 - 1.2
10"6 - 1.2
2.2 -1Q-3
0.38
0.5
ID"3
ID"3
Table 5.
Initial Mineral Volume Fractions in Reactive Barrier and Aquifer
Mineral volume fraction Reference
Reactive barrier
Aquifer
Fe°(s)
goethite
pyrolusite
non-reactive
0.5 - 0.62
1.0 -10-3
1.0 -10-3
0.698
Bennett [1997]
estimated
estimated
estimated
Table 6.
Reactive Surface Area Estimates for Zero Valent-iron (Field Installation)
surface area [m2 mineral m 3 mineral]
geometric surface area 7.50 103
(calculated from dso = 0.4 mm)
specific reactive surface area 3.88 106
(calculated from laboratory experiment)
22
-------
Table 7. Reactive Surface Area Estimates for Eh-buffer Minerals
mineral reactive surface area
goethite
pyrolusite
i.o.
i.o.
IO2
102
Table 8. Input Concentrations at Boundary Located Upgradient of Reactive Barrier, Transect 2, 21-1 - 21-4
21-1 21-2 21-3 21-4
depth
pH
EH
co2-
ci-
so%-
HS-
N03-
NH|
Na+
K+
Ca2+
Mg2+
Fe2+
Fe3+
CrO2-
Cr(OH)+
Mn2+
A13+
H4SiO4
02(aq)
H2(aq)
TCE
cis-1,2 DCE
VC
ethane
CH2O
CH4(aq)
[m]
H
[mV]
[mol I"1]
[mol I-1]
[mol I"1]
[mol I"1]
[mol H]
[mol I"1]
[mol I"1]
[mol H]
[mol I"1]
[mol r1]
[mol I"1]
[mol I"1]
[mol I"1]
[mol I"1]
[mol r1]
[moi r1]
[moi r1]
[moi r1]
[moi r1]
[moi r1]
[mol I"1]
[mol I"1]
[mol H]
[mol I"1]
[mol H]
7
7,
'.0
.36
377
1.
4,
1.
5,
1.
3,
8,
2,
2,
2,
1.
1.
5,
2,
1.
3,
9,
9,
2,
2,
5,
8,
1.
1.
4,
.13.
.23.
.91-
.54.
.18.
.72.
.96.
.05.
.77.
.69.
.69.
.20.
.78.
.16.
.77.
.88.
.43.
.13.
.57.
.04.
.16.
.00.
.66.
.37.
.80.
. 1Q-03
. 1Q-04
. 1Q-04
. 1Q-88
. 1Q-05
lo-11
. 1Q-04
. 1Q-05
. 1Q-04
. 1Q-04
ID"15
io-12
. 1Q-08
. 1Q-08
. 1Q-06
. 1Q-08
. 1Q-06
. 1Q-05
io-31
. 1Q-05
. 1Q-09
. 1Q-09
. 10-07
. 1Q-04
. 1Q-06
1
5
3
8.
1
1
1
4
2
2
5
1
1
1
1
1
9
3
1
3
5
8
1
6
1
6
7.
.5
84
441
.23-
.87-
.28-
00-
.16-
.32-
.54-
.35-
.22-
.06-
.56-
.03-
.93-
.60-
.57-
.15-
.50-
.72-
.93-
.27-
.16-
.00-
.66-
.99-
.75-
1Q-03
1Q-04
1Q-04
10-ioo
1Q-05
io-24
1Q-03
1Q-05
1Q-04
1Q-04
io-18
io-12
1Q-05
1Q-08
1Q-06
10-or
1Q-06
1Q-05
io-34
1Q-06
1Q-09
1Q-09
10-or
1Q-06
1Q-06
6
7.
.0
97
356
1.22-
8.44-
5.30-
3.80-
1.08-
2.72-
2.19-
2.05-
1.96-
1.81-
6.84-
1.03-
3.71-
1.53-
2.04-
1.55-
9.54-
6.25-
8.63-
2.82-
5.16-
8.00-
1.66-
4.84-
3.86-
1Q-03
1Q-04
1Q-04
io-90
1Q-05
io-14
1Q-03
1Q-05
1Q-04
1Q-04
io-17
io-12
1Q-05
1Q-08
1Q-06
1Q-07
1Q-06
1Q-06
io-32
10-07
1Q-09
1Q-09
1Q-07
1Q-05
1Q-06
5
6.
.5
40
437
1.85-
2.33-
9.38-
2.52-
4.61-
4.44-
3.64-
4.86-
3.92-
4.20-
1.42-
3.81-
4.75-
1.11-
5.97-
8.07-
9.40-
1.56-
2.00-
2.59-
9.29-
8.00-
1.66-
3.55-
3.37-
1Q-03
1Q-03
1Q-04
io-86
1Q-05
1Q-09
1Q-03
1Q-05
1Q-04
1Q-04
io-13
io-12
1Q-05
1Q-07
1Q-06
1Q-09
1Q-06
1Q-05
io-31
10-07
1Q-09
1Q-09
1Q-07
1Q-05
1Q-06
23
-------
Table 9. Input Concentrations at Boundary Located Upgradient of Reactive Barrier, Transect 2, 21-5 - 21-7
21-5 21-6 21-7
depth
PH
EH
co2-
cr
sor
HS-
NOJ
NHJ
Na+
K+
Ca2+
Mg2+
Fe2+
Fe3+
CrOl-
Cr(OH)+
Mn2+
A13+
H4SiO4
02(aq)
H2(aq)
TCE
cis-1,2 DCE
VC
ethane
CH2O
CH4(aq)
[m]
H
[mV]
[mol I"1]
[mol r1]
[mol T1]
[mol T1]
[mol T1]
[mol T1]
[mol T1]
[mol H]
[mol I"1]
[mol H]
[mol I"1]
[mol I"1]
[mol H]
[mol I"1]
[mol H]
[mol I"1]
[mol I"1]
[mol H]
[mol I"1]
[mol H]
[mol I"1]
[mol I"1]
[mol r1]
[moi r1]
[moi r1]
5.0
6.29
444
1.
4.
1.
5.
8.
1.
6.
6.
5.
5.
2.
4.
9.
1.
7.
8.
9.
8.
1.
1.
1.
1.
2.
7.
1.
.95-
.04-
.44-
.29-
.83-
.35-
.01-
.65-
.32-
.97-
.53-
.71-
.84-
.52-
.88-
.17-
.40-
.13-
.98-
.98-
.37-
.73-
.99-
.78-
.37-
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
-03
-03
-03
-86
-05
-08
-03
-05
-04
-04
-13
-12
-05
-07
-06
-09
-06
-06
-31
-06
-06
-07
-06
-05
-06
2.
1.
3.
4.
2.
1.
1.
1.
6.
6.
6.
3.
8.
8.
2.
8.
9.
1.
1.
2.
2.
1.
1.
1.
9.
4.5
6.48
441
,98'
,98.
,24.
,03.
,69.
,18.
,78.
,25.
,86.
,38.
,91
,30.
,93.
,93.
,13.
,38.
,40.
,06.
,01
,29.
,95.
,05.
,66.
,01.
,29.
. iQ-03
. iQ-03
. 1Q-04
. 1Q-88
. 1Q-05
. 10-10
. 1Q-03
. 1Q-04
. 1Q-04
. 1Q-04
. 1Q-14
io-12
. 1Q-06
. 1Q-08
. 1Q-05
. 1Q-09
. 1Q-06
. 1Q-05
io-31
. 1Q-06
. 1Q-06
. 1Q-06
. 1Q-07
. 1Q-04
. 1Q-06
2.
7.
8.
1.
3.
1.
7.
1.
5.
4.
2.
1.
2.
3.
4.
1.
9.
2.
9.
4.
5.
1.
1.
4.
1.
4.0
6.94
385
,40.
,36.
,40.
,86.
,69.
,77.
,92.
,48.
,47.
,36.
,48.
,71
,89.
,64.
,37.
,62.
,40.
,47.
,91
,72.
,73.
,23.
,66.
,98.
,35.
. 1Q-03
. 1Q-04
. 1Q-05
. 1Q-85
. 1Q-06
. 1Q-08
. 1Q-04
. 1Q-04
. 1Q-04
. 1Q-04
. 1Q-14
io-12
. 1Q-08
. 1Q-08
. 1Q-05
. 1Q-08
. 1Q-06
. 1Q-05
io-31
. io-or
. io-or
. iQ-06
. iQ-07
. iQ-05
. 1Q-05
24
-------
Table 10. Reaction Processes Affecting Component Concentrations
component
sources
sinks
cor
ci-
so2-
NHj
Na+
K+
Ca2+
Mg2+
Fe2+
Fe3+
CrO2-
Cr(OH)+
Mn2+
A13+
H4SiO4
02(aq)
H2(aq)
TCE
cis-1,2 DCE
VC
ethane
CH2O
CH4(aq)
pH
sulfate reduction
nitrate reduction
iron corrosion
goethite
iron corrosion
Cr(VI)-reduction
pyrolusite
iron corrosion
carbonate minerals
DIC-reduction
sulfate reduction
mackinawite
nitrate reduction
calcite, Ca-Mg-carbonate
Ca-Mg-carbonate
siderite, Fe(OH)2,
mackinawite
Fe(OH)3
Cr(VI)-reduction
Cr(OH)3
rhodocrosite
Oo-reduction
TCE-reduction
cis-1,2 DCE-reduction
VC-reduction
TCE-reduction
cis-1,2 DCE-reduction
TCE- and VC-reduction
DOC-reduction
DIG- and DOC-reduction
reduction-corrosion reactions
secondary mineral formation
deprotonation
reduction-corrosion reactions
secondary mineral formation
25
-------
Table 11. Rate Constants for Reduction-corrosion Reactions
oxidant
Cr(VI)
TCE
cis-1,2 DCE
VC
oxygen
nitrate
sulfate
DOC
DIG
water
i)
2)
log k
log A;
lab
calibrated data
2.895
-2.207
-3.179
-2.383
2.5
-1.589
-2.5
-3.8
-3.8
-10.331
calculated
calculated
2.895
-2.207
-3.179
-2.383
-
-2.589
-
-
-
-6.331
unit
[1 m-2 d-1]
[1 m-2 d-1]
[1 m-2 d-1]
[1 m-2 d-1]
[1 m-2 d-1]
[1 m-2 d-1]
[1 m-2 d-1]
[1 m-2 d-1]
[1 m-2 d-1]
[mol m^2 d^1]
based on surface area estimated
reference
Gould [1982]
O'Hannesin et al [1995]
O'Hannesin et al. [1995]
O'Hannesin et al. [1995]
-
Rahman and Agrawal [1997] ^
-
-
-
Reardon [1995]2)
from grain size
based on measured surface area
Table 12. Calibrated Effective Rate Constants for Secondary Mineral Formation
log keS
mineral [mol m^3 d^1]
Fe(OH)2(am)
Fe(OH)3(am)
Cr(OH)3(am)
CaCO3(s)
CaMg(C03)2(s)
FeC03(s)
MnCO3(s)
FeS(am)
-1.000
-1.000
-2.000
-1.200
-2.500
-2.000
-2.500
-2.000
Table 13. Estimated Rate Constants for Reductive Dissolution Reactions
mineral
[mol
log k
FeOOH(s) -5.000
MnO2(s) -5.000
26
-------
Figures
27
-------
Pasquotank River
oo
p-
10
0.05 mg/1
20
meters
lmg/1
N
A
Hangar 79
Plating
Shop
Figure 1. Configuration of reactive barrier and approximate location of chromium plume, from Bennett [1997].
Groundwater
flow
Multilevel
sampling wells
Figure 2. Monitoring network, from Bennett [1997].
28
-------
Contaminated
Zone
Advective-dispersive
transport of
contaminants
and other
dissolved species
upgradient
Treatment
Zone
Removal of contaminants
by reduction and precipitation
Reduction of other
Corrosion of
zero-valent iron
pR-increase
Eh-decrease
Precipitation of secondary
minerals
Exsolution of dissolved gases
pH and Eh buffering
reactive barrier
Buffer
Zone
Dissolution of
clay minerals
Desorption of hydrogen ions
pH - decrease
Reductive dissolution of
oxides and oxy-hydroxides
Exsolution of dissolved
gases
Eh-increase
downgradient
Figure 3. Conceptual model for reactive barriers comprised of zero-valent iron, from Bennett [1997].
Monitoring points (chemistry of infiltrating groundwater)
D Monitoring points (barrier and downstream of barrier)
21
22 23 24
25
>
21-5
P
b
b
1 3.2m p
>
> :
22-3 b
b
h
D
D
D
D
23-3
D
D
D
q
c|
1 Location
I of barrier
d
CJ 24-3
id
d
D
D
25-5 D
D
D
D
D
4.0m
Figure 4. Solution domain including location of barrier and monitoring points along Transect 2.
29
-------
4
4.5
, 5
j
5.5
6.5
7
0
0.5
1.5 2 2.5
distance [m]
3.5
Figure 5. Spatial discretization of two-dimensional solution domain.
0.5
1.5 2 2.5
distance [m]
3.5
K
1.2x10
3.3x10
9.2x10
2.6x10
7.3x10
2.1x10''
5.8x10
[m
-04
-05
-05
-06
-07
Figure 6. Hydraulic conductivity distribution in two-dimensional solution domain, modified from Bennett [1997].
30
-------
a) 10
1U
1(T4
10'5
7
10"
io-9
r>-10
Cr(VI)
Cr(III)
Crtot - field data
detection
limit
ID'
10-
10-
0.5 1.0 1.5 2.0 2.5
distance [m]
3.0 3.5
4.0
TCE - field data
* cisl,2-DCE- field data
VC - field data
± ethane - field data
ethane
(r..
I^"X A
T v*'-. ^.x (-ic 1 0 FiPF *
| \^*
1
£
detection
limit
0.0 0.5 1.0 1.5 2.0 2.5
distance [m]
3.0 3.5
4.0
Figure 7. Contaminant concentrations after t = 240 days: a) chromium, b) organics - one-dimensional simulation.
' IV
ID'4
10'5
:: io"6
£lO'7
H"
10
io-9
IO,OQ
NO;
NO; - field data
NH;
0 0.5 1.0 1.5
di
ID'3
ID'4
TL, 10
| 10"6
£j
H-10
10'8
ID'9
in-10
2-
SO2; - field data
f
\
\
\
\
\
\.
1
detection
limit
2.0 2.5 3.0 3.5 4.0
stance [m]
^
^S.
,\^
s
t
I
i
i:
HS"
detection
limit
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
distance [m]
4.0
Figure 8. Redox couple concentrations after t = 240 days: a) nitrate/ammonia, b) sulfate/sulfide - one-dimensional simulation.
31
-------
1U
4 Ca2+ - field data
» Mg2+ - field data
t~ Mn2+ - field data
"o in'5
SI \J t '.
/ \ ''
/ V*
/ A ^
i * \
i * \ -
i * \
i * \
i * \ ~
'X
H \ i
i V i
V _ -
^
%
\
x-..
1 :
l
2.0 2.5 3.0 3.5 4
500
400
300
200
100
0
-100
-200
-300
-400
-500
-600
0-700
[election
limit
>
53
distance [m]
Figure 10. pH and Eh after t = 240 days - one-dimensional simulation.
32
-------
T3
13
a
t«
o
o
o
Cr
others
overall
-0.0010
2.0 2.2
distance [m]
Figure 11. Iron corrosion rates in reactive barrier after t = 240 days.
0.005
0.004
0.003
0.002
0.001
calcite
Ca-Mg-carbonate
siderite
Fe(OH)2(s)
ferrihydrite
mackinawite
2.4
2.6
1.6
1.
2.0 2.2
distance [m]
Figure 12. Secondary mineral volume fractions in reactive barrier after t = 240 days.
2.4
2.6
33
-------
l.U
0.9
0.8
0.7
0.6
0.4
0.3
0.2
0.1
n n
-
-
s
x.
" * »
*
-
-
porosity _:
-
-
».. volume fraction of secondary minerals ~
-**I3 :
cpFe - T = 20 years ~
volume fraction of a. T <- :
zero-valent iron \ _ n/r
(f) - T = 10 years
(|) - T = 20 years :
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
i n
0.0
0.1
0.2
0.3
distance [m]
0.4
0.5
0.6
Figure 13. Long term effect of iron corrosion and secondary mineral formation.
4.5
i5-5
6.5
7
0
0.5
1.5 2 2.5
distance [m]
3.5
Figure 14. Streamlines in two-dimensional solution domain.
34
-------
0
0.5
1.5 2 2.5
distance [m]
3.5
4
Cr(VI)
8.4x10
7.0x10
5.6x10
4.2x10
2.8x10
1.4x10
0.0x10
[mol I'1]
-05
-05
-05
-05
-05
-05
+00
0
0.5
1.5 2 2.5
distance [m]
3.5
4
Cr(III)
1.4x10
1.2x10
9.7x10
7.5x10
5.3x10
3.2x10
1.0x10
[mol I'1]
-07
-07
-08
-08
"
I
I
1
1 " "
1
Cr(OH)3(am)
1.2xlO"°3
1.6xlO-°4
2.4xlO'05
|3.4xlO-°6
4.9xlO-°7
7.0xlO"°8
l.OxlO"08
0
0.5
1.5 2 2.5
distance [m]
3.5
4
Figure 15. Hexavalent and trivalent chromium concentrations and CrOH,(am) volume fractions after t = 2 years.
35
-------
TCE
0.5
1.5 2 2.5
distance [m]
3.5
-05
2.0x10
6.3x10"
2.0x10
6.3x10
2.0x10
6.3x10
2.0x10
[mol I'1]
-06
-07
-07
-08
-08
cis-l,2DCE
0.5
1.5 2 2.5
distance [m]
3.5
3.0x10
1.4x10
6.5x10
3.0x10
1.4x10
6.5x10
3.0x10
[mol I'1]
-06
-07
-07
-07
-08
-08
4
4.5
5
B
£5.5
(~\
i H
CD
^ 6
6.5
7
ta .
P . . . .
vc
0
0.5
1.5 2 2.5
distance [m]
3.5
2.0x10
9.3x10
4.3x10
2.0x10
9.3x10
4.3x10
2.0x10
[mol I'1]
-06
-07
-07
-07
-08
-08
-08
Figure 16. TCE, cis-1,2 DCE and VC concentrations after t = 2 years.
36
-------
0
0.5
1.5 2 2.5
distance [m]
4
4.5
, 5
£5.5
6
6.5
7
0 0.5 1 1.5 2 2.5 3 3.5
distance [m]
4
550.0
375.0
200.0
25.0
-150.0
-325.0
-500.0
[mV]
Figure 17. pH and En-distribution after t = 2 years.
37
-------
0.5
1.5 2 2.5
distance [m]
3.5
4
1.4x10
8.1x10
4.7x10
2.8x10
1.6x10
9.5x10
5.5x10
[mol I'1]
-04
-04
-04
-04
-05
-05
4
4.5
, , 5
I5'5
G
^ 6
6.5
7
(
) 0.5 1 1.5
>'
.^1
2
t*
* . . . .
2.5 3 3.5 A
HS'
distance [m]
distance [m]
6.0x10
3.5x10
2.0x10
1.2x10
7.0x10
4.1x10
2.4x10
[mol I'1]
-08
-09
-09
-09
4
4.5
i i 5
I5'5
^ 6
6.5
7
(
i
i
) 0.5 1 1.5
. . .
r
. . .
2
.
-
P . . . .
2.5 3 3.5 A
mackinawite
2.0x10
5.6x10
1.6x10
4.5x10
1.3x10
3.5x10
1.0x10
-03
-04
-04
-05
-05
-06
-06
Figure 18. Sulfate and sulfide concentrations and mackinawite volume fractions after t = 2 years.
38
-------
39
------- |