&EPA
                          United States
                          Environmental Protection
                          Agency
                          Research and Development
                                     Environmental Research
                                     Laboratory
                                     Athens GA 30605-2720
                                     EPA/600/S-93/002  April 1993
ENVIRONMENTAL
RESEARCH   BRIEF
          Oxidation-Reduction Mechanisms in Iron-Bearing Phyllosilicates
                          Joseph W. Stucki1, George W. Bailey2, AND Huamin Gan1
 Abstract
 Oxidation-reduction reactions of structural iron in smectite clay
 minerals were studied with respect to reduction reaction mecha-
 nisms and effects on physical-chemical properties of the clay.
 The purpose of these studies was to assess the implications of
 redox chemistry at mineral surfaces on transport, transforma-
 tion, and bioavailability of redox-sensitive metal pollutants. Re-
 ducing agents selected  were dithionite (S2O42-),  sutfide (S2-),
 thiosulfate (S2O 2'), hydrazine (NLH4), ascorbic acid (C Hpj,
 hydroquinone (C6H6O2),  and sodium oxalate (Na2C2O4). Clay
 samples were prepared as aqueous suspensions  of <2-u.m
 particle-size fractions of Na-saturated, freeze-dried ferruginous
 smectite. The reductive strength of each  reducing agent was
 determined by measuring the resultant level of Fe(ll) in the clay
 crystal using either a photo-colorimetric method or Mossbauer
 spectroscopy. Heats of reaction were used with S2O42, S2-, and
 S O32- to further determine the extent of  redox reaction. The
 role of free radicals in the reduction process was measured
 using electron spin resonance (ESR) spectroscopy.

 Results revealed that the order of reduction potential for the
 various agents was S2O42 > S22' > C6H8O6> S2O32 > C HO  »
 C2H2O4. Free radical activity was found only in the reaction with
 dithionite and was assumed to be due to the sulphoxylate (SO2
 •) free radical. The free radical was labile in pure  solution, but
 its lifetime increased at least two-fold in the presence of the
 clay. The signal from the clay-S O42 suspension may, in fact,
 be only partially attributable to SO,,-. Reduction by this highly
 reactive agent may induce  electron hopping within  the clay
'crystal itself, which could also be reflected in the persistent
 signal from the clay. When free radicals or unpaired electrons
1  Department of Agronomy, University of Illinois, Urbana IL 61801
2  Environmental Research Laboratory, U.S. Environmental Protection Agency,
 Athens GA 30605-2720
                         are involved in the reducing processes, the following steps
                         may occur: a) the active free radical (e.g., SO,,-) approaches
                         the clay surface and transfers an electron to structural Fe(lll),
                         reducing it to Fe(ll); b) because of this initial reduction, the
                         crystalline structure is energetically destabilized by an excess
                         negative charge,  causing  partial dehydroxylation,  which, in
                         turn, energetically activates point defects, such as tetrahedral
                         aluminum (Al) sites, within the clay crystal; c) excess electrons
                         at point defects may pass to structural Fe(lll), reducing it.
                         These processes continue until all structural Fe(lll) is reduced.

                         At least two types of reducing agent were identified based on
                         their reducing mechanism,  namely, those with and those with-
                         out free radical activity. The number of steps involved in the
                         reduction  process depends on the reduction potential of the
                         reducing agent. Measurements of rheological characteristics of
                         oxidized and reduced clay suspensions indicated that struc-
                         tural Fe(ll) increases the attractive bond energies between clay
                         particles. The type of bonding between  particles is uncertain;
                         hydrogen  bonding may  make an important contribution. The
                         effects of microbial reduction and interactions with clay miner-
                         als were reviewed.


                         Background
                         The  risk that metal cations pose to the global ecosystem
                         depends largely on their activity in porous media, which can be
                         calculated only if the true  exposure  levels of the cations to
                         biota are  known.  Exposure is determined  by the transport,
                         transformation, and bioavailability characteristics of the metal
                         cation, which in turn depends on speciation. The fate of redox-
                         active pollutants in porous media are governed by solubility
                         and adsorption processes, but modifications also occur due to
                         redox reactions in the solid-liquid interface. These effects are
                         governed by the thermodynamic energy  of redox couples and

                                                  < X'fx Printed on Recycled Papei

-------
by the rates at which reactions proceed. Solid surfaces in the
media create a unique chemical environment that influences
both the energetics and the kinetics of reactions among chemi-
cal components.

Knowledge of the solution concentration or activity of metals
alone is insufficient to provide a complete  model of exposure
because, in this interfacial region, two phenomena occur that
are absent  in solution.  First, the solid surfaces  become  an
active participant,  as both reactant and product, in the chemi-
cal reactions. Second, chemical constituents in solution come
under the influence of van der Waals, electrostatic, hydration,
and possibly other forces that alter their total potential energy
or reactivity. Accurate definition  of speciation thus requires an
understanding of the processes and forces that are operating
in the interface, in terms of both the participants and the rates
of reaction. When this understanding is achieved, the capability
to model and predict more successfully the behavior of perco-
lating ions in porous media will be greatly enhanced.

The purpose of this study was  to characterize more  fully the
physical-chemical properties  and  processes  that  occur when
iron-rich clay minerals undergo reduction and reoxidation. Re-
dox reaction mechanisms in  phyllosilicate  clay minerals were
studied with respect to the free radical activity, heat of reaction,
and effect  on clay interlayer  forces. Several reducing agents
representing a  wide range of  reduction  potentials and free
radical activities were compared, and  the redox interactions
between clays and microorganisms were reviewed. The project
also produced a publication containing a detailed  review of the
thermodynamic basis of redox reactions in  clays (Stucki et ai,
1992).


Methods
The <2-u,m particle-size fraction  of ferruginous smectite SWa-1
(Source Clays Repository of  The Clay Minerals  Society, Co-
lumbia, Missouri)  was Na-saturated, washed free of excess
salts to approximately  10/3 M Na, and freeze  dried.  Twenty-
five- to 30-mg  portions  of the freeze-dried clay  were then
resuspended for subsequent reduction treatments generally by
mixing with 37.5 ml of high-purity H2O (18  Mohm crrr1 resistiv-
ity) and  2.5 ml of  citrate-bicarbonate buffer  (1  part 0.3 M
Na2C6O5-6H2O and 8 parts 1 M NaHCO3). The  mixture was
shaken gently overnight. The citrate-bicarbonate (CB) buffer
was used  in order to maintain near-neutral pH  and thereby
minimize acid dissolution of  the clay  during  treatment with
Na2S2O4. In some experiments,  however, only high purity H2O
was used.  Structural Fe  in  the clay was reduced  at  room
temperature (nominally 25°C), for time periods  ranging from 1
to 57 hr, by reagent-grade reducing agents. Those used were
sodium dithionite  (Na^OJ,  sodium  sulfide (Na.S*9H2O), so-
dium  thiosulfate (Na2S2O3-5H2O), hydrazine (N2H4),  ascorbic
acid (C6H8O6),  hydroquinone (C6H6O2), and sodium  oxalate
(NeLC2O4).  Iron(ll) and total Fe  were measured quantitatively
by the photo-colorimetric method of Komadel and Stucki (1988),
or semi-quantitatively by Mossbauer spectroscopy.

Electron spin resonance (ESR) spectra were obtained at room
temperature using a Bruker ESP 300 (X-band) ESR spectrom-
eter equipped with  both liquid  and solid sample cells.  Each
reducing agent was analyzed  in  the solid phase and in solution
with either  high purity HO or  CB buffer solution.  ESR spectra
of the clays also were obtained after resuspension, but in the
absence of reducing  agent.  Each reducing agent then was
added to the suspended clay  to bring the Na concentration in
suspension to 0.01 M, and the mixture was  immediately loaded
into the ESR liquid cell. ESR spectra were  obtained after
various time intervals up to 57 hr.
The  resonance  peak position  in  each ESR  spectrum was
expressed  in terms of the value  of g, which was calculated
from the magnetic field intensity (H) using the relation
                           (M
                                                      (1)
where  h is Planck's constant, v is  the microwave frequency
with which the sample was irradiated while the magnetic field
was varied (for an X-band ESR spectrometer, the value of vis
about 9 GHz), and 8 is the  Bohr magneton (= 9.2741-1021 erg
gauss'1).

Line intensity is  an important characteristic of an ESR spec-
trum and can be a qualitative indicator of spin concentrations if
relaxation and  saturation  effects are  absent.  According to
Vedrine (1980),  the spin populations directly determine the
magnetic susceptibility,  \, to  which the  line intensity, 7, is
proportional according to the relationship
                      3kT
                                                       (2)
where  L,  S, and  J (=  L + S) are, respectively, the orbital,
spin,  and total angular momenta  of the electron; N  is the
number of unpaired spins; k is Boltzmann's constant; / is the
absolute temperature; and  g is the "so called" electron free-
spin g- factor. In the present study,  all of the terms on the right
side of  Equation 4, except  Np, were assumed to be constant.
The intensity of the  ESR  signal  thus  was assumed  to  be
directly  proportional to No and was obtained by integrating the
area under the pre-derivative spectrum.

For rheological measurements,  clay suspensions  were pre-
pared as above but treated only with Na2S2O4, N^SO  (sodium
suit ate), or NaCI (sodium chloride).  Shear stress vs. shear rate
for  different treatments were made  using  a  Fann  rotational
viscometer (Model 35, Fann Instrument Corp.,  Houston,  TX) at
25°C. During these measurements, the R1-B1  rotor-bob com-
bination was used with a spring having a constant of 0.2  or 0.5.
When the  oxidation of  the structural iron was a factor, the
measurement was made inside an inert-atmosphere glove box.

Enthalpy, or  heat  of reaction, was measured  using a  Calvet
microcalorimeter in the laboratory of Dr. Philip F. Low at Purdue
University,  using the techniques described by Gan (1990) and
by Gan  and Low (1992).


Results and Discussion

Comparison of Reducing Agents
The reduction of structural iron in ferruginous smectites  by
different reducing  agents  has important implications for the
behavior of redox-sensitive  metals  in the environment. Reduc-
ing agents used were dithionite  (S O42-), sulfide (S2-), thiosul-
fate (S2O2"), stannous chloride (SnCl2), hydrazine (N  H4),  ascor-
bic acid (C H8O6), oxalate (C2H2O4), phenol (C6H6O), and hyd-
roquinone (C6H6O2). Experiments were carried  out at different
temperatures and  with  different contact times between clay
and reducing agent. Survey results revealed that great differ-
ences in reducing  power exist among these reducing agents.
Results from three of these are  compared in Figure  1.  In the
inorganic group, S2O42 demonstrated very strong reducing abil-
ity, whereas S2O32 was the weakest. In the organic group, the
reducing power was in the order C6H8O6 > C?H6O2 > C2H2O4 «
C6H6O.  The  reduction process sometimes involves  ancillary

-------
    0.6  -
 5"
 fi
    0.2
             0.57
                            0.11
                                          0.042
Figure 1.
            Dithionite        Sulfide       Jhiosulfate

                        Reducing Agent
Reduction of structural Fe in Na smectite SWa-1 by freshly
prepared 0.01 M solutions o1Na2S3O, NaaS, and Na2SaO3 for
24hrat25°C.
reactions in the mineral structure, depending on the reducing
power of the reductant.
                                                   sisted for up to 9 hr of contact between clay and Na2S2O4,
                                                   which was more than double its lifetime in pure Na S n  solu-
                                                   tion. Hence, the concentration of free radicals from the Na2S,O4
                                                   solution was preserved and enhanced if added to the clay.
                                                   Apparently, the  SWa-1  initially reacted with SO/- which do-
                                                   nated an electron to the clay; but as the reaction proceeded,
                                                   unpaired electrons were produced within the clay crystal struc-
                                                   ture, giving rise  to the persistent  signal in  Figure 3B. This
                                                   explanation is  consistent with earlier studies  (Stuck! et  a/.,
                                                   1984b;  Lear and Stucki, 1985; Stuck! and Lear,  1989) that
                                                   strongly Indicated, based on layer charge measurements, that
                                                   some of the Fe is reduced by a source of electrons (Z) within
                                                   the clay structure. Aluminum-substituted tetrahedral sites may
                                                   provide such a source  of high-potential electrons due to the
                                                   lower-valent tetrahedral cation. This step would have occurred
                                                   only after sufficient  Fe(lll) was  reduced  by SO/-  to either
                                                   activate  or catalyze  the  movement  of internal electrons  to
                                                   structural Fe(lll). This process would be similar to the reaction
                                                   proposed by Stucki and Lear (1989):
                                                                                               ^)s + (m-a) Z1'1
                                                                                                          (3)


                                                                                                          (4)


                                                                                                          (5)
                                                             where  subscripts x and s denote  clay and solution phases,
                                                             respectively; Z is an unidentified electron donor located within
                                                             the clay crystal, which may be the Al-substituted tetrahedral
Free Radical Activity
Further comparisons were made among the inorganic group of
reducing agents. Figure 1 reveals the relative reducing abilities
of S2O42', S2-,  and SLO32  in ferruginous smectite (SWa-1) at
room temperature (25°C). Hydrazine's effect was similar to that
of S2\ Notice that S2O42- is peculiar and demonstrates a much
higher ability to reduce structural Fe(lll) than any of the other
compounds.

The standard  electrode  reduction potential, g**,  of S2O42  is
approximately -1.12 V (Vanysek, 1992). But N H4 (or its conju-
gate acid N H * if  hydrated), which generally reduces  only
about 10% of the structural Fe  in these same clays (Stucki et
a/., 1976; Rozenson and Heller-Kallai, 1976a,  1976b; Stucki
and Roth, 1977; Russell era/.,  1979; Stucki, 1988; Stucki and
Lear, 1989), has an  jf*  of -0.94 V (Douglas et a/., 1983).
Obviously, the difference  between these two reducing agents
must be due to something other than their standard electrode
potentials.

The ESR spectra of these same reducing agents revealed no
signal from N2H4, S2O32, or S2-,  and a strong signal from S O42
(Figure 2). These results clearly indicate the presence of un-
paired electrons in S2O42' by the resonance signal centered at
about g = 2.0091 for the solid phase and shifted to g = 2.0113
for the solution  phase. The signals were most intense in the
freshly prepared S2O42- solutions, then decreased in intensity
over time, disappearing completely after about 57 hr in the 1.0
Absolution, and after less than 4 hr in the 0.01  M solution.

When S2O42' was added  to the clay suspension, making the
final Na concentration 0.01 M, a moderately strong ESR signal
was evident initially (Figure 3A), then after 4  hr the signal
became even stronger than that of the pure Na2SX>4 solution of
the same concentration (0.01 M) (Figure 3B).  The signal per-
                                                          3.44
 3.48

Field (kG)
3.52
                                                     Figure 2.  ESR spectra of solid (A) and 1.0M aqueous solution (B) on
                                                                       Na,S,04.

-------
                                        g = 2.01077
 3.45
             3.46
 3.47

Field (kG)
                                        3.48
                           3.49
Figure 3.  ESP spectra of 0.01 M Na2S2O4 in smectitle SWa-1 suspen-
         sion: A) fresh, B) after 4 hr.
sites as suggested above; e represents the reducing agent in
the solution surrounding the clay crystal, and is believed to be
the SO,;- free radical when Na^G^ is the reducing agent; and
m,  r,  and a are stoichiometry coefficients. According to this
hypothesis, structural  Fe is reduced partially by internal  and
partially  by  external reducing agents, and  the reduction is
accompanied by dehydroxylation and reprotonation reactions.
Lear and Stucki (1985) determined that two  of the stoichiom-
etry coefficients are linearly related,  viz., r = 0.32 m, but the
value of a has yet to be determined.

The point of contact between  the clay and the reducing agent
is still unknown. The same arguments of coulombic repulsion
against the S2O42  anion  approaching the  negatively charged
basal surfaces also apply to the  SO2-   free  radical anion,
except perhaps the high  reactivity of the free  radical would
overcome the coulombic repulsion energy barrier. Like-charge
collisions where one is a free  radical are not unusual, however
(Neta et  al., 1988). Alternatively, the  free radical may react at
the edges  of  clay layers,  setting  up a conduction  pathway
through the layer to internal Fe ions. Such a process is con-
ceivable  in Fe-rich smectite, and may explain why Rozenson
and Heller-Kallai (1976a, 1976b) observed incomplete reduc-
tion of Fe-poor  montmorillonite by Na2S2O4.

The conductivity of electrons  from the layer edge through the
octahedral sheet would depend heavily on the presence of the
transition metal. But Lear and Stucki (1987), based on mea-
surements of magnetic exchange interactions and  of inter/a-
                                   lence electron transfer, concluded that structural Fe is reduced
                                   nearly randomly  within  the  octahedral sheet.  Reduction only
                                   from layer edges would likely create a reducing front passing
                                   through  the  layer,  creating homogeneous domains of  Fe(ll)
                                   and Fe(lll) with a rather constant number of Fe(ll)-Fe(lll) pairs
                                   at the reduced-oxidized interface.  Results of Lear and Stucki
                                   (1987) clearly reject this possibility in ferruginous smectite.


                                   Heats of Reaction
                                   The driving force for any chemical reaction is reflected totally in
                                   the partial molar Gibbs free energy, AG, of the reaction^which
                                   is given  by  the  sum  of  the changes  in  enthalpy, AR, and
                                   entropy, AS,  viz.,
                                                                               AG=AH-TAS
                                                                                         (6)
where  T is the absolute temperature. A more detailed thermo-
dynamic treatment is provided by Stucki et al. (1992).

Enthalpy changes were  measured for the reaction of ferrugi-
nous smectite with reducing agents S2O42-, S2, and S2O32.  As
noted above, the  strength of these three reducing agents is in
the order S2O42 > S2 > S2O 2\ The value _pf AG", therefore,
should follow the same trend. The value of AR is known as the
heat of reaction and can be measured experimentally with a
microcalorimetejr.  If AS is constant  in all three reactions,  AR
should reflect AG.

The heat  of reduction reaction was found to follow the order
S2O42 > S2O 2 > S2. Compared to the levels of Fe(ll) achieved,
the results for S2O32 and  S2'  are reversed, suggesting that
entropy changes  are greater in the S2' treatment.  Further ex-
planation of this phenomenon will require more  detailed under-
standing of the S2-clay  reaction. These results also indicate
that, in addition to considering  AR, one must not overlook the
AS term.

Recall  that the reducing power of S2O42- deteriorates with the
age of the solution because of the lability  of the sulphoxylate
(SO?-») free radical, but S2- and S2O 2- solutions are unaffected
by time. Measurements of  heats of reduction reaction using
variably aged solutions  revealed  that AR  decreased with  in-
creasing age  only with the S2O42- solutions.  These results,
therefore, are consistent with the free radical studies that found
that reducing power of S2O42 is labile.

Microbial Reduction of Clay Fe
Microorganisms apparently  also  reduce structural  Fe in clay
minerals (Komadel et al., 1987; Stucki et al., 1987; Wu et  al.,
1988).  The mechanisms for microbial reduction have yet to be
identified, and many questions arise as to the precise role of
microorganisms in Fe redox reactions. For instance, does the
reduction occur because of a membrane-bound  process requir-
ing intimate contact between clay mineral and organism, or is it
due to an extra-cellular or exudate compound from the organ-
ism? Is it an aerobic or anaerobic process? What are the
metabolic sequences responsible for reduction? Which organ-
isms are most efficient? How does microbial  reduction affect
physical-chemical properties of the minerals and their effect on
metal fate, transport, and bioavailability characteristics? Stud-
ies to  answer these questions are currently underway (e.g.,
Gates era/., 1991, 1992).


Reduction Process Schematic
The reduction mechanism(s) for Fe(lll) smectites depend  on
the nature of the  reducing agent and medium in which reduc-

-------
tion takes place. Laboratory methods may or may not replicate
actual redox processes that occur in nature but should provide
insights as to the potential range of reactions that can occur.
Standardization of  techniques,  natural  variations among
samples, and handling of air-sensitive clays all are factors that
must be considered before valid comparisons between studies
can be made. In Figure 4 is a conceptual model illustrating how
the reduction potential or energy of the external reducing elec-
tron may determine the  extent of Fe(ll) produced and the type
of ancillary processes that occur, such as those represented in
Equations  3  through 5. A small  amount of  Fe(lll) can be
reduced to Fe(ll) in the clay crystal by a number of reducing
agents having only modest reductive capability  (having energy
» E,), as indicated by the small energy barrier A. The principal
effect of these agents is simply the reduction of Fe(lll) to Fe(ll),
with  a concomitant increase in layer charge and probably a
decrease in the crystal lattice stabilization energy because the
dioctahedral structure naturally  prefers trivalent  octahedral cat-
ions.

Further reduction by agents having greater reductive capabili-
ties (having energies > E, or E3), such as the SO2'»free radical,
invokes a dehydroxylation  process as  indicated by  energy
barrier B. Dehydroxylation was reported  by Roth and Tullock
(1973) and Stuck! and Roth (1976), based on infrared spectral
information, and by  Lear  and Stucki (1985) based on tritium
exchange between structural OH  in the  clay and H2O  in the
surrounding solution.  These changes in the clay crystal, com-
bined with the high electron energy of the free radical, eventu-
ally surmount  energy barrier C, which mobilizes electrons al-
ready present in the clay crystal into molecular or metallic-like
orbitals or into  semiconductor-like conductivity bands. The de-
  UJ
  c
  UJ
  I
                                  Electron
                                  Delocalization
Figure 4.
       Progression of Reduction Process

Conceptual model of relationships among energy of reducing
electrons, structural Fe(ll) content, and progression of the
reduction process in ferruginous smectites.
                                                     localized electrons then are captured by Fe(lll), thus effecting
                                                     further Fe reduction.

                                                     Because these latter electrons were initially present in the clay,
                                                     the change in Fe(ll) content at this point is not reflected in the
                                                     total  layer charge of the clay, which explains the discrepancy
                                                     between the predicted and  the  observed  layer charge de-
                                                     scribed by Stucki et al. (1984b), Lear and Stucki (1985), and
                                                     Khaled  and Stucki (1991).  The energy barrier configuration
                                                     may  vary depending on the total  Fe content  of the clay be-
                                                     cause a low-Fe smectite, such as montmorillonite, may have a
                                                     much lower metallic character than a high-Fe  smectite, which
                                                     would  diminish the probability for the electron delocalization
                                                     process represented by barrier C (Figure 4).

                                                     Experimental results presented herein may  explain some as-
                                                     pects of why Na2S2O4 reduces more Fe(lll) in the clay structure
                                                     than  other reducing agents that have similar electrode reduc-
                                                     tion potentials.  When  unpaired electrons are  involved  in the
                                                     reducing processes, the following steps  may take place: a) the
                                                     active free radicals  (e.g., SO2-) approach the clay surface and
                                                     transfer electrons to structural Fe(lll), reducing it to Fe(ll); b)
                                                     because of this initial  reduction,  the  crystalline structure  is
                                                     energetically destabilized by an excess negative charge, caus-
                                                     ing partial dehydroxylation,  which,  in turn,  energetically acti-
                                                     vates point defects, such as tetrahedral Al sites, within the clay
                                                     crystal; c) excess electrons at point defects may pass to struc-
                                                     tural  Fe(lll), reducing it. The processes continue until all struc-
                                                     tural  Fe(lll) is reduced.  At least two types of reducing agent are
                                                     identified  based on their reducing  mechanism, namely, those
                                                     with and those without  free radical activity.


                                                     Effects of Reduction on Interlayer Forces
                                                     Studies by Stucki era/.  (1984c), Chen et al. (1987), Khaled and
                                                     Stucki  (1991), and Stucki and Tessier (1991) indicated that Fe
                                                     reduction increased attractive forces between  clay layers, but
                                                     the exact nature of these  is still  unknown. Shear stress  mea-
                                                     surements in the present study indicate that  hydrogen bonding
                                                     may  be an  important factor. Oxidized  and reduced clay sus-
                                                     pensions  were  placed  in a rotational viscometer, which  mea-
                                                     sured the shear stress (s) as a function  of the shear  rate (o).
                                                     Results were then  plotted according to the relation
                                                                                  s = rjpio + 9
                                                                                                             (7)
                                                     where the slope, TJ , is the plastic viscosity of the system, and
                                                     the y-intercept, 6, is the Bingham yield stress. 6 is a measure
                                                     of the number of interacting clay particles and the strength of
                                                     bonds  or links that occur between them, as  given  by the
                                                     relation
                                                                                                                      (8)
                                                              where EA is  the energy  of the  interparticle  bond, N is the
                                                              number of particles, and K1 is  a constant.

                                                              Flow curves of reduced and unaltered  1%  suspensions  of
                                                              ferruginous smectite SWa-1 are shown in Figure 5. Notice that
                                                              the slope (TJ )  and y-axis intercept  (9) are higher for the re-
                                                              duced than for the unaltered sample. According to Stucki and
                                                              Tessier (1991), the particles are larger in reduced compared to
                                                              oxidized smectites, so for the same mass the number of par-
                                                              ticles,  N, must be less in the reduced clay. Hence, by Equation
(8), the particle interaction  energy,
duced clay.
                                                                                                 E  is  stronger in the re-

-------
             — 6-  Reduced

                   Oxidized
    ^  4
 Figure 5.
                         400             800
                         Shear Rate (1/sec)
Flow curves of oxidized and reduced 1% suspensions ofNa-
nontronite (SWa-1) at 1Q3 M A/a* concentration
strong flocculation decreases the particle  interaction energy.
(4)  Reduction of structural  Fe(lll) increases the interparticle
forces as indicated by  increases  in 6.  And, (5) the possible
forces controlling the interaction energy between particles are
hydrogen bonding and cation bridging.


Environmental Implications
Soil  serves many functions in the environment. It  acts  as a
geologic filter scavenging undesirable solutes from a percolat-
ing  solution, or  it may  release previously  sorbed substances
back into the  surrounding  media  depending on climatic and
other conditions. The surfaces of soil minerals provide unique
chemical environments  that often  facilitate the transformation
of some chemical species  to very  different forms, some of
which  may be less desirable and  more mobile, or vice-versa.
Changing the surface properties of the minerals through oxida-
tion or reduction of structural Fe will  likely have a great impact
on the capability of the clay  interlayers to trap or sequester
metal cations, which in  turn  will affect the bioavailability of the
metal.

The  chemical  nature of mineral surfaces is widely recognized
as one of the most significant factors contributing to the fate
and  behavior of  contaminants, but these properties are gener-
ally considered to be  unchanging  during the time of contami-
nant exposure and reaction. The aforementioned studies clearly
demonstrate, however, that mineral surface  chemistry is greatly
modified by changes in  the oxidation state of Fe in the  mineral
crystal structure.

The  redox activity of mineral surfaces also  affect the oxidation
state, and consequently the  speciation and  chemical behavior,
Figure 6 shows the flow behavior of  reduced and  unaltered
SWa-1 clay in  3% suspension, and compares the effects of
three different anions,  namely, S2O42- (reducing agent), SO42-,
and  Cr.  The Na* concentration was  the same in all three
samples. Notice that both S2O42 and SO 2 have greater effects
than Ch  The difference between the S2O42 and SO42 may be
attributed to structural Fe reduction, but the difference between
the sulfate and  Cr must be due to different anion effects.  The
ability of Cr to break hydrogen bonds is well known, so the
smaller interaction energy between particles in the presence of
Cr  may  be due  to  fewer or weaker  H  bonds between the
particles.

The  increase of particle  interaction indicated by increasing  6
and  T) clearly reveals  that the  reduction of structural  Fe(lll)
increases the attractive forces between particles. This, in turn,
likely alters the particle arrangement in the system. This sup-
ports the  hypothesis that reduction  of  structural  Fe(lll) in-
creases the proportion of  collapsed layers, decreases the swell-
ing (Stuck! et al., 1984a; Stuck! and  Lear, 1989; Wu  et al.,
1989; Gates et al., 1991), and changes the hydraulic conduc-
tivity of the system (Shen et al., 1992).

These and other results from this study help us understand the
colloidal  properties of the oxidized and reduced samples,  and
lead to the following general  conclusions: (1) The  interaction
between clay particles in  aqueous suspension increases as the
solid concentration increases; the sol to gel transition is found
to occur  at a solid to liquid concentration of 2%.  (2) Stronger
interaction  was  detected  for the intermediate particle-size frac-
tion  <2 and  >0.5 u,m. (3)  Interaction between  particles in-
creases as the  Na* concentration increases up to 0.01 M, then
                                                          50   -
                                                            0 '•
                                                                             400              800
                                                                             Shear Rate (1/sec)
                                                              Figure 6.   Flow curves of oxidized and reduced 3% suspensions of Na-
                                                                        nontronite (SWa-1) at 0.5 M A/a* concentration.

-------
of redox-sensitive metal ions in the surrounding solution. The
application of such insight can be seen in the case of Cr(VI), a
major environmental pollutant. Chromium may exist as Cr(VI)
anion in solution (CrO42-, Cr,O72-)- A reduced Fe clay may react
with the Cr(VI)  species, reducing it to Cr(lll), which will either
be cationic or precipitated as the oxide. By this process, Cr will
be  less mobile and  less hazardous in  the soil environment.
Geochemical models, e.g.,  MINTED (Allison et al., 1991), for
predicting the fate and behavior of metals in the vadose zone,
therefore, must account for  redox transformations of the metal
ion at mineral surfaces.

The exposure  of soil minerals  to  redox-sensitive sorbates,
solvents,  humic materials,  varying  redox environments,  and
microorganisms may invoke surface chemical changes in situ
over short time  periods and  thereby produce vital differences in
the speciation and reactivity of all components in the soil-water
system.


Acknowledgments
Financial  support of this project by the U.S. Environmental
Protection Agency, under Cooperative Agreement CR816780-
01-2, is gratefully acknowledged.  Mention of trade names or
commercial products does  not constitute endorsement or rec-
ommendation by the  U.S.  Environmental Protection Agency.
The authors also thank Dr. Lelia Coyne for her interest and
numerous suggestions regarding the reduction mechanism of
structural Fe.


References Cited
  Allison,  J. D., D.  S.  Brown, and K. J. Novo-Gradac. 1990.
    MINTEQA2/PRODEFA2,  A Geochemical Assessment
    Model for Environmental Systems. U.S. Environmental Pro-
    tection Agency, Athens, GA. EPA/600/3-91/021.
  Chen,  S.  Z., P. F.  Low, and C.  B.  Roth. 1987. Relation
    between potassium fixation and the oxidation state of
    octahedral  iron. Soil Sci. Soc. Am. J. 51:82-86.
  Douglas, B.,  D.  H. McDaniel, and J. J.  Alexander. 1983.
    Concepts and Models of Inorganic Chemistry. John Wiley
    and Sons, New York.
  Gan, H. 1990. Factors affecting the interparticle bond energy
    of Na-montmorillonite. Ph.D. Thesis,  Purdue University.
  Gan, H., and  P. F. Low. 1992. Heats of immersion of silicon
    and aluminum oxides  and clays. Clays Clay  Miner. (In
    Preparation).
  Gan, H., J. W. Stucki, and G. W. Bailey. 1992. Free radical
    reduction of structural iron in ferruginous  smectite. Clays
    Clay Miner, (submitted)
  Gates, W. P., H.  T. Wilkinson, and  J.  W. Stucki. 1991.
    Effects of microbial reduction on clay swelling.  Agronomy
    Abstracts 1991:365.
  Gates, W. P., H.  T. Wilkinson,  and  J.  W. Stucki. 1992.
    Effects of microbial reduction on clay swelling. Clays Clay
    Miner. (In Preparation).
  Khaled,  E. M., and J. W.  Stucki. 1991. Effects of iron oxida-
    tion state on cation fixation in smectites. Soil Sci. Soc. Am.
    J. 55:550-554.
  Komadel, P.,  and J. W.  Stucki. 1988. The quantitative assay
    of Fe2* and Fe3* using  1, 10-phenanthroline:  III. A rapid
    photochemical method:  Clays Clay Miner. 36:379-381.
  Komadel, P.,  J. W.  Stucki,  and H. T.  Wilkinson.  1987.  Re-
    duction of structural iron in smectites by microorganisms.
    p. 322-324. In E.  Gal£n, J. L.  Perez-Rodriguez,  and J.
    Cornejo  (eds.)  Proc. Sixth Meeting of the European  Clay
    Groups,  Seville, 1987.  Sociedad Espanola  de Arcillas,
    Sevilla.
 Lear, P. R.,  and J. W. Stucki. 1985. The role of structural
   hydrogen  in  the  reduction  and  reoxidatton of iron  in
   nontronrte. Clays Clay Miner. 33:539-545.
 Lear, P. R.,  and J. W. Stucki. 1987. Intervalence electron
   transfer  and magnetic exchange interactions in reduced
   nontronrte. Clays Clay Miner. 35:373-378.
 Lear, P. R., and J. W.  Stucki. 1989. Effects of iron oxidation
   state on  the specific  surface area of nontronite. Clays Clay
   Miner. 37:547-552.
 Neta, P., R. E. Huie, and A. B. Rose. 1988. Rate constants
   for reactions of inorganic radicals and aqueous solution. J.
   Phys. Chem. Ref. Data 17:1031-1032, 1128-1141.
 Roth,  C. B.,  and  R.  J.  Tullock. 1973.  Deprotonation  of
   nontronrte  resulting from chemical reduction of structural
   ferric iron. p.  107-114. In J. M Serratosa  (ed.), Proc. Int.
   Clay  Conf., Madrid, 1972 (publ. 1973), C.S.I.C., Madrid.
 Rozenson, I., and  L.  Heller-Kallai.  1976a.  Reduction and
   oxidation of Fe3* in  dioctahedral smectite—1: Reduction
   with hydrazine and dithionite. Clays Clay Miner. 24:271-
   282.
 Rozenson, I., and  L.  Heller-Kallai.  1976b.  Reduction and
   oxidation of Fe3* in  dioctahedral smectite—2: Reduction
   wrth sodium sulphide solution. Clays Clay Miner. 24:283-
   288.
 Russell, J. D.,  B. A.  Goodman, and A.  R. Fraser.  1979.
   Infrared  and Mossbauer studies of reduced nontronites.
   Clays Clay Miner. 27:63-71.
 Stucki, J. W. 1988. Structural iron in smectites, p. 625-675.
   In J.  W. Stucki,  B.  A. Goodman, and  U. Schwertmann
   (eds.) Iron in Soils and Clay Minerals. D. Reidel, Dordrecht.
 Stucki, J. W., and P. R. Lear. 1989. Variable oxidation states
   of iron in the  crystal structure of smectrte clay minerals.
   Chapter  17. In L. M. Coyne, D. Blake, and S. McKeever
   (ed.)  Structures and  Active Sites  of Minerals. American
   Chemical Society, Washington, D.C.
 Stucki, J. W., and C. B. Roth. 1976. Interpretation of infrared
   spectra of oxidized  and  reduced  nontronrte. Clays Clay
   Miner. 24:293-296.
 Stucki,  J. W., and C. B. Roth.  1977 Oxidation-reduction
   mechanism for structural iron in nontronite. Soil Sci. Soc.
   Am. J. 41:808-814.
 Stucki, J. W., and D. Tessier. 1991. Effects of iron oxidation
   state  on  the texture and structural order of Na-nontronite.
   Clays Clay Miner. 39:137-143.
 Stucki, J. W., C.  B. Roth, and W. E. Bartinger. 1976. Analysis
   of iron-bearing clay minerals by electron spectroscopy for
   chemical analysis (ESCA). Clays Clay Miner. 24:289-292.
 Stucki, J. W., Golden,  D. C., and Roth, C.  B. 1984a. The
   preparation  and handling of drthionrte-reduced smectrte
   suspensions. Clays Clay Miner. 32:191-197.
 Stucki, J. W., D. C. Golden, and C. B. Roth. 1984b. Effects
   of reduction and reoxidation of  structural iron on the sur-
   face charge and dissolution of dioctahedral smectrtes. Clays
   Clay Miner. 32:350-356.
 Stucki, J. W., P. F.  Low, C. B.  Roth, and  D. C.  Golden.
   1984c. Effects of oxidation state of octahedral iron on clay
   swelling.  Clays Clay Miner. 32:357-362.
 Stucki, J. W., P.  Komadel, and H. T. Wilkinson. 1987. Micro-
   bial reduction  of structural iron(lll)  in smectrtes. Soil Sci.
   Soc. Am. J. 51:1663-1665.
 Stucki,  J. W., G.  W.  Bailey,  and  H. Gan.  1992.  Redox
   reactions in phyllosilicates: thermodynamic basis and im-
   plications for metal behavior. In H. E. Allen, G. W. Bailey,
   A. R.  Bowers, A. W. Garrison, and C.  P. Huang  (eds.)
   Metal Speciation and Contamination of  Soil. Lewis  Pub-
   lishers, Inc., Chelsea, Michigan. (In Preparation)
Vanysek, P. 1992. Electrochemical Series, p. 8-16 to 8-23.
   In Lide, D.R. (ed.) CRC Handbook of Chemistry and Phys-
   ics, 71st ed. CRC Press, Boca Raton, FL
                                                                                 *U.S. Government Printing Office: 1993 — 750-071/60226

-------
Vedrine, J. C.  1980. General theory and experimental as-      Wu, J., P. F. Low, and C. B. Roth. 1989. Effects of octahe-
  pects of electron spin resonance: In J. W. Stucki and W. L.        dral-iron reduction and swelling pressure on interlayer dis-
  Banwart (eds.) Advanced Chemical Methods for Soil and        tances in Na-nontronite. Clays Clay Miner. 37:211-218.
  Clay Minerals Research, D. Reidel, Dordrecht, 331-389.        Zhang, Z. Z., and P. F. Low. 1989. Relation between the
Wu, J., P.  F. Low, and C. B. Roth. 1988. Biological reduction        heat of immersion and the initial water content of Li-, Na-,
  of structural iron in sodium-nontronite. Soil Sci. Soc.  Am.        and K-montmorillonite. J. Colloid  Interface Sci.  133:461-
  J. 52:295-296.                                               472.
United States
Environmental Protection Agency
                                                                                               FPA
Center for Environmental Research Information                                           DCDUIT M  o oc
_.   .    .. _. , ..____                                                                    rbHMI I NO. Ca-35
Cincinnati, OH 45268
      BULK RATE
POSTAGE & FEES PAID
Official Business
Penalty for Private Use
$300

EPA/600/S-93/002

-------