Mechanisms and Products of Surface-Mediated Reductive Dehalogenation of Carbon Tetrachloride by Fe(II) on Goethite

Article abstract of DOI:10.1021/es034741m

Aliphatic chlorinated hydrocarbons, including CCl₄, are widespread groundwater contaminants. Mechanisms and product formation of CCl₄ reduction by Fe(II) sorbed to goethite, which may lead to completely dehalogenated products or to form chloroform, a toxic product that is fairly persistent under anoxic conditions, were studied. A simultaneous transfer of two electrons and cleavage of two C-Cl bonds of CCl₄ would completely circumvent chloroform production. Product formation pathways did not primarily depend on the competition between an initial one- and two-electron transfer, but on the presence of different radical scavengers and the properties of the mineral surface with respect to stabilization of reaction intermediates. Specific adsorption of major anions or pH effects could modify the capability of the goethite surface to stabilize short-lived radical intermediates.

Full text of DOI:10.1021/es034741m

Environ. Sci. Technol. 2004, 38, 2058-2066
natural organic matter, enhances formation of toxic
CHCl3.
Mechanisms and Products of
Surface-Mediated Reductive
Dehalogenation of Carbon
Introduction
Aliphatic chlorinated hydrocarbons, including carbon tet-
rachloride (CCl4), are widespread groundwater contaminants
(1-4). While such compounds are rather persistent under
oxic conditions, they may undergo reductive dehalogenation
under reducing conditions in the subsurface. Depending on
the predominating environmental conditions, dehalogena-
tion of CCl4 can lead to harmless end products, but can also
lead to more persistent compounds. The majority of studies
on abiotic reductive dehalogenation of CCl4 reported hy-
drogenolysis to chloroform (CHCl3), which is toxic and more
recalcitrant than CCl4 under anoxic conditions. Most studies,
however, also report incomplete mass balances (5,6) or partial
conversion of CCl4 to completely dehalogenated products
that are indicative of alternative dehalogenation pathways
of CCl4 (7-14). Formate (HCOO-), carbon monoxide (CO),
and carbon dioxide (CO2) were found during electrolytic
dehalogenation of CCl4 at a silver electrode (7); formate in
the reaction of CCl4 with goethite/ Fe(II) (8); and CO and CO2
in the reaction with pyrite or with sulfide in the presence of
layer silicates (9,10) as well as in photochemical transforma-
tion of CCl4 in water (ice) films (11). Transformation of CCl4
by Pseudomonas stutzeri KC is the only case of biotrans-
formation in which readily degradable (thio)phosgene
(SdCCl2/ OdCCl2) and CO2 accounted for most of the mass
balance, and only traces of CHCl3 were found (12,13). Other
microorganisms studied produced CHCl3 as a major product
(14). Therefore, hydrogenolysis and complete dehalogenation
are concurrent and, thus, competing pathways in the
dehalogenation of CCl4 by various reductants, and knowledge
about the conditions that favor either pathway is of great
interest, both for evaluation of natural attentuation and for
the design of remediation schemes.
Tetrachloride by Fe(II) on Goethite
M A R T I N E L S N E R , †
S T E F A N B . H A D E R L E I N , ‡
T H O M A S K E L L E R H A L S , S A M U E L L U Z I ,
L U C Z W A N K , W E R N E R A N G S T , A N D
R E N EÄ P . S C H W A R Z E N B A C H *
Swiss Federal Institute of Environmental Science and
Technology (EAWAG), Ueberlandstrasse 133,
CH-8600 Duebendorf
Natural attenuation processes of chlorinated solvents in
soils and groundwaters are increasingly considered as
options to manage contaminated sites. Under anoxic
conditions, reactions with ferrous iron sorbed at iron(hyro)-
xides may dominate the overall transformation of carbon
tetrachloride (CCl4) and other chlorinated aliphatic
hydrocarbons. We investigated mechanisms and product
formation of CCl4 reduction by Fe(II) sorbed to goethite, which
may lead to completely dehalogenated products or to
chloroform (CHCl3), a toxic product which is fairly persistent
under anoxic conditions. A simultaneous transfer of two
electrons and cleavage of two C-Cl bonds of CCl4 would
completely circumvent chloroform production. To distinguish
between initial one- or two-bond cleavage, 13C-isotope
fractionation of CCl4 was studied for reactions with Fe(II)/
goethite (isotopic enrichment factor ꢀ ) -26.5‰) and
with model systems for one C-Cl bond cleavage and either
single-electron transfer (Fe(II) porphyrin, ꢀ ) -26.1‰) or
partial two-electron transfer (polysulfide, ꢀ ) -22.2‰). These
ꢀvalues differ significantly from calculations for simultaneous
cleavage of two C-Cl bonds (ꢀ ≈ -50‰), indicating
that only one C-Cl bond is broken in the critical first step
of the reaction. At pH 7, reduction of CCl4 by Fe(II)/
goethite produced ∼33% CHCl3, 20% carbon monoxide
(CO), and up to 40% formate (HCOO-). Addition of 2-propanol-
d8 resulted in 33% CDCl3 and only 4% CO, indicating that
both products were generated from trichloromethyl radicals
(•CCl3), chloroform by reaction with hydrogen radical
donors and CO by an alternative pathway likely to involve
surface-bound intermediates. Hydrolysis of CO to HCOO-
was surface-catalyzed by goethite but was too slow to account
for the measured formate concentrations. Chloroform
yields slightly increased with pH at constant Fe(II) sorption
density, suggesting that pH-dependent surface processes
direct product branching ratios. Surface-stabilized
Reduction of CCl4 is believed to start in most cases with
a single electron transfer and simultaneous cleavage of one
C-Cl bond, which leads to short-lived trichloromethyl
radicals (single-electron-transfer, Scheme 1).
SCHEME 1
In such a case, product formation is determined by radical
•
reactions of CCl3 with system components. Presence of
molecular oxygen, for example, causes reoxidation to phos-
gene/ CO2 (11,15); sulfide creates sulfur adducts (9,10,16);
transfer of a second electron causes further reduction; and
H• radical donors, such as organic compounds, favor CHCl3
formation (13,16,17). Alternatively, a concerted transfer of
two electrons with simultaneous cleavage of one C-Cl bond
is believed to take place in a nucleophilic attack of strong
reductants at the halogen center (X-philic reactions, Scheme
2) (16,18), thus leading directly to short-lived trichloromethyl
anion intermediates.
intermediates may thus facilitate abiotic mineralization of
CCl4, whereas the presence of H radical donors, such as
* Corresponding author phone: (++41 1) 823 5109; fax: (++41
1) 823 5471, e-mail: schwaba@eawag.ch.
† Present address: Department of Geology, University of Toronto,
22 Russell Street, Toronto, Ontario M5S 3B1.
In such a case, the product formation is governed by the
rate of H+ transfer relative to that of self-decomposition of
CCl3- (19). Proton transfer leads to chloroform. Loss of Cl-
‡ Present address: Center for Applied Geosciences (ZAG), Eber-
hard-Karls University, Wilhelmstr. 56, D-72074 Tuebingen.
9
2 0 5 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004
10.1021/es034741m CCC: $27.50
2004 Am erican Chem ical Society
Published on Web 03/03/2004

and to investigate how environmental conditions (pH, Fe(II)
surface site coverage, presence of •H donors) influence
product formation (i.e., chloroform, formate, and potentially
other products); and (ii) to gain more insight into the
mechanism of product formation on the basis of information
from product studies and trapping experiments, as well as
from a complementary isotope fractionation study.
SCHEME 2
from CCl3- yields dichlorocarbene, which rapidly hydrolyses
to formyl chloride (HCOCl) (20) and, eventually, to CO (21).
Because both Scheme 1 and Scheme 2 can lead to CHCl3, it
is intrinsically impossible to circumvent chloroform produc-
tion, except for a theoretical third case in which two electrons
are transferred and simultaneously, two C-Cl bonds are
broken in the initial reaction step so that dichlorocarbene
is formed directly (8) (case 3, Scheme 3).
To distinguish between initial one- and two-bond-
cleavage processes, 13C-isotope fractionation during deha-
logenation of CCl4 was studied with various electron donors.
Completion of mass balances was attempted by compre-
hensive analyses of reaction products at pH 7, including CO,
•
CHCl3, and formate. To trap intermediate CCl3 radicals, an
efficient D• radical donor (perdeuterated 2-propanol) was
used, or alternatively, great care was taken to exclude possible
H• radical donors from the experimental system such as
organic buffers (e.g., MOPS, HEPES). To distinguish the
influence of pH from that of Fe(II) surface site coverage,
experiments were conducted at different pHs, but constant
Fe(II) sorption site density.
SCHEME 3
Simultaneous cleavage of two C-Cl bonds, however, is
highly endergonic, whereas the postulate of transition state
theory requires that the reaction coordinate proceeds along
a path of lowest activation energy. In solution, that is, without
stabilization of the respective intermediate(s), generally only
one C-Cl bond will thus be initially broken; the situation,
however, may be different on surfaces, where a transition
state of the type dS- - -Cl- - -CCl2- - -Cl- - -Sd can be imag-
ined that may reduce the activation energy for simultaneous
cleavage of two bonds (where dS denotes surface groups).
Materials and Methods
Great care was taken in experiments with goethite/ Fe(II) to
minimize organic residues which could serve as possible H
radical donors in the experimental system. Prior to the
experiments, the goethite was repeatedly washed with
deionized water to remove organic acids known to be present
at the surface (30,31). Moreover, organic (“Goods”) buffers
were avoided, since they are known to transfer H• atoms to
•CCl3 radicals and, thus, generate chloroform [see Scheme 1
(13)]. Instead, pH control was ensured by the intrinsic buffer
effect of dS-O-Fe(II)-OH groups at the goethite surface.
This buffer capacity was examined for representative ex-
perimental conditions (50 m2/ L goethite in equilibrium with
1 mM dissolved Fe(II)aq) by acid/ base titrations between pH
7 and 8. The determined buffer intensity was 250-300 µM
per pH unit, meaning that 250-300 µM acid/ base had to be
added in order to change the pH value by one unit.
Calculations showed that this was sufficient to ensure pH
control during our experiments. (For titration data, calcula-
tion of buffer intensities, and estimates of corresponding pH
changes during reaction of CCl4, see Supporting Information).
Goethite. Goethite (R-FeOOH) (17 ( 1 m2/ g BET-surface,
determined by N2 adsorption with Sorptomatic 1990, Fisons
Instruments) was purchased from Bayer (Bayferrox 910,
Standard 86). Tetrachloromethane (Fluka), dichloromethane
(Fluka), chloroform (Aldrich), and trichloroethene (Aldrich)
were of the greatest available purity (g99%) and were used
as received. Methanol-free CCl4 spike solutions were prepared
by adding 3.5 µL of pure liquid CCl4 (>99.5%, Fluka) to 110
mL anoxic water (3 h purged with Ar) and stirred for at least
24 h prior to transfer into an anoxic glovebox.
Because of the great importance and ubiquitous occur-
rence of ferrous iron sorbed at natural mineral surfaces, we
focus in this study on elucidating the pathways of reductive
dehalogenation of CCl4 by reactive surface-bound Fe(II)
species. McCormick et al. (22) reported that reductive
dehalogenation of CCl4 in the presence of iron-reducing
bacteria was mainly due to abiotic reaction with Fe(II) at the
surface of biogenic magnetite and only to a small extent to
direct enzymatic dehalogenation. In many other studies, Fe-
(II) sorbed to mineral surfaces was found to be by orders of
magnitude more reactive than Fe(II) in solution (23-26). It
has been found that for some contaminants including CCl4,
reactions with Fe(II) sorbed at iron(hyro)xides may be the
dominating removal process under anoxic conditions (27,28).
Surface-bound Fe(II) is also thought to play a role in metal
iron reactive walls that are increasingly used to clean up
contaminated aquifers (29).
Pecher et al. (8) identified chloroform and formate as
major products of the reaction of CCl4 with Fe(II) sorbed to
goethite at circumneutral pH and found increasing yields of
CHCl3 with higher pH. Amonette et al. (6) obtained in similar
systems under acidic conditions a rate law that was second-
order with respect to sorbed Fe(II) and first-order with respect
to dissolved CCl4. They postulated a termolecular initial two-
electron reaction step in which each of two adjacent Fe(II)
surface sites conveyed simultaneously one electron to a CCl4
molecule.
Fe(II) Stock Solutions. Fe(II) stock solutions (0.5 M, pH
5) were prepared by adding 3.63 g (0.065 mol) iron powder
(Merck) to 100 mL of a 1.00 M HCl solution (prepared from
32% HCl p.a., Merck) that had been purged with Ar for 1 h.
The mixture was heated to 80 °C under slow stirring for 2 h,
until most of the powder had dissolved and visible hydrogen
evolution ceased. The mixture was subsequently transferred
into a glovebox and was filtered through a 0.2-µm PTFE filter
to remove excess iron powder. The exact Fe(II) concentration
was determined photometrically after complexation with
phenanthroline (32).
To date, the reaction mechanisms and conditions that
cause formation of toxic chloroform vs completely dehalo-
genated products are not well-understood for this reaction.
In particular, it remains unclear whether and under which
conditions a simultaneous two-electron transfer associated
with a two-bond-cleavage process can occur and whether
formation of chloroform may be attributable to an initial
single-electron transfer or, alternatively, generation of for-
mate to an initial two-electron transfer. The objectives of
this study were, therefore, (i) to complete the mass balance
of the reductive dehalogenation of CCl4 by goethite/ Fe(II)
Washing Procedure. To remove traces of organic residues,
the glassware used in experiments for formate analysis was
(1) soaked in ethanol for several hours, (2) rinsed with
deionized water, (3) soaked in 1.5 M HNO3 for several hours,
and (4) rinsed six times with deionized water immediately
before use.
9
VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 0 5 9

Preparation of Goethite/Fe(II) Suspensions. An amount
of goethite corresponding to 50 m2/ L was suspended three
times in deionized water in order to remove adsorbed ions
from the surface. The resulting suspension was sparged with
argon for 4 h and subsequently transferred into an anoxic
glovebox with external regenerator (Vaccum Atmospheres
Corp.) at an oxygen level of e0.5 ppm O2 as monitored by
an oxygen sensor (PBI Dansensor, Module ISM-3). There,
the pH (initially ∼5) was adjusted with 0.1 M NaOH (Titrisol,
Merck) to pH ) 7.0. FeCl2 solution was added, and the pH
was adjusted again until the desired conditions were obtained
(generally 1 mM Fe(II) in solution at pH 7). If necessary, pH
and Fe(II) were readjusted after an equilibration time of 48
h. (Sorption of Fe(II) to glassware was found to be negligible.)
Two additional sets of experiments were conducted at
different conditions: One assay contained 3% (vol) of
perdeuterated 2-propanol (Aldrich); the other, only 0.45 mM
Fe(II) total (all sorbed to goethite) at a pH of 8.
added to the remaining suspension. Extraction took place
on a reciprocating shaker for 10 min, and 3 × 0.8-mL aliquots
of extracts were taken for analysis of chlorinated hydrocar-
bons.
In isotope fractionation experiments, 10 mL of diethyl
ether containing 9 µM benzene (>99.5%, Fluka) as internal
standard was added into the headspace of the reaction vials,
followed by extraction for 10 min in a reciprocating shaker.
Then the cannula of a 1-mL Hamilton glass syringe without
piston was bored through the Viton stopper. The ether phase
(∼1 mL remained after equilibration) was forced into the
glass tube of the syringe by injecting water into the reaction
vial through a second cannula; 3 × 200 µL of the ether extract
was taken for analysis by GC-C-IRMS.
Analytical Methods. For all analytical methods, external
standards were prepared in exactly the same way as the
samples. Standards for chlorinated alkanes were set up in
serum vials with goethite suspensions (but without Fe(II)),
and they were extracted according to the same procedures.
Formate standards were set up in goethite/ Fe(II) suspensions,
which were filtered after addition of NaOH. They had to be
prepared daily, because nearly complete disappearance of
formate was observed in the goethite/ Fe(II) standards at room
temperature within a week (see discussion below). CO
standards were prepared by mixing varying volumes of a 900
ppm standard with air.
Iron(II) Porphyrin. Iron(II) porphyrin (meso-tetrakis(N-
Methylpyridyl)iron(II) porphyrin) was prepared from meso-
tetrakis(N-methylpyridyl) iron(III) porphyrin kindly provided
by Buschmann (16) according to the following method from
Wade and Castro (33): 13 g of iron powder (>99.5%, Merck)
was treated with 1 N HCl for 1 h, transferred into the glovebox,
and washed three times with anoxic water. The iron was
then added to 0.25 L of 0.5 mM iron(III) porphyrin, and the
mixture was shaken for 15 min and subsequently filtered
through a 0.2 µM PTFE filter (BGB Analytik). During reduction,
the color of the solution changed from green to red, as verified
by UV-vis spectroscopy (see Figure S1 in the Supporting
Information). In isotope fractionation experiments, the iron-
(II) porphyrin concentration was adjusted to 150 µM, and
200 µM 4-morpholinopropanesulfonic acid (MOPS, Fluka)
was added as buffer at pH 7.
CO Analysis. CO was determined on a gas chromatograph
equipped with a thermal conductivity detector (GC-TCD,
Shimadzu GC-8A) and a packed column (molecular sieve
8 Å 80/ 100, Brechbu¨ hler, Switzerland). Helium (99.999%)
was used as carrier gas at 25 mL/ min, the oven temperature
was isothermal at 80 °C, injector and detector tempera-
tures were at 150 °C, and the detector current was set
at 140 mA. The detection limit of the method was ∼150
ppm.
Polysulfide Solutions. Polysulfide solutions were prepared
by mixing 49 g (0.2 mol) of Na2S‚H2O (p.a., 32-38% S, Fluka)
with 50 g of sulfur powder in 1 L of deionized anoxic water
for 2 months (pH 9) (34). In isotope fractionation experiments,
the polysulfide concentration was adjusted to 40 mM total
S(-II), at pH 8.3.
Formate Analysis. Formate was quantified after enzymatic
reaction by measuring the production of NADH photo-
metrically at 339 nm according to a procedure modified from
Schaller and Triebig (35). A mixture of 0.5 mL of a â-NAD
solution (20 mM), 1 mL of phosphate buffer (0.15 M, pH 7.5),
and 1.5 mL of sample solution was prepared in a polystyrene
vial (Semadeni, all chemicals from Fluka). The photometric
absorption of the solution was measured at 339 nm in a
quartz precision cuvette (Suprasil, Hellma, 5-cm length).
The solution was subsequently transferred back into the
polystyrene vial, where 10 µL of formate dehydrogenase
solution was added (formate dehydrogenase Pseudomonas
spec., recombinant mutant to 79900 from E. coli; ∼175 U/ mL,
Fluka Biochemicals, 75274). After an incubation time of
30-60 min at room temperature, the mixture was trans-
ferred to the quartz cuvette, which had been cleaned by
rinsing twice with deionized water, and the absorption
was measured again. The difference between absorption
before and after incubation with formate dehydrogenase
was used for quantification. The detection limit of this
method was ∼2 µM formate (sample concentration in
experiment).
Transform ation Experim ents of CCl4. Experiments were
set up in the glovebox. For each experiment, 18 replicates
were prepared in parallel by transferring 49 g (96 g) of the
stirred suspension into 57-mL (115 mL) serum vials. (Values
in parentheses describe the setup in isotope fractionation
experiments.) Then freshly prepared CCl4 spike solution was
added to yield an initial concentration of 40-50 µM in the
product study and 10 µM in the isotope fractionation
experiments. The vials were closed with Viton stoppers, taken
out of the glovebox, and agitated on a horizontal shaker at
25 °C/ 140 rpm in the dark until they were sacrificed for
analysis. Experiments to study the conversion of CO to
formate were set up at pH 7 in 57-mL vials, with (a) 10 mL
of goethite/ Fe(II) suspension and (b) 10 mL water. Outside
the glovebox, the headspace was filled with pure CO gas via
a cannula through the Viton septum, corresponding to a
concentration of CO in solution of ∼1 mM.
For sampling in the product study, first two 1-mL aliquots
of gas were taken from the headspace of the vials by piercing
the Viton stopper with a gastight syringe (A2, 1 mL, 0.29 ×
0.12 ×2 in., VICI AG) with side-port taper, under simultaneous
introduction of 2 × 1 mL of water with a second syringe.
These gas samples were subsequently analyzed for CO. Then
7 mL of liquid was withdrawn, to which 300 µL of 0.1 M
NaOH was added in order to precipitate Fe(II). This sample
was filtered with a 0.2-µm PTFE filter and stored at -20 °C
until analysis for formate. Finally, 8 mL diethyl ether (Merck)
containing 10 µM trichloroethene as internal standard was
Quantification of Volatile Halogenated Compounds. CCl4
and CHCl3, as well as CDCl3, were quantified on a GC/ MS
(GC Fisons 8000 Series, autosampler Fisons AS 800, quad-
rupole MS Fisons MD 800) equipped with a 60 m × 0.32 mm
Stabilwax fused-silica column (film thickness 1 µm) and a 8
m
× 0.53 mm deactivated guard column. On-column
injection was used to introduce 1 µL of sample into the
column. The temperature program was 40 °C (2 min), ramp
8 °C/ min to 130 °C (0 min), ramp 30 °C/ min to 200 °C (5
min). Quantification was carried out in the single ion
monitoring mode, for CCl4 at m/ z ) 84, 123, for CHCl3 at m/ z
) 47, 87, and for CDCl3 at m/ z ) 88. The detection limit of
this method was ∼0.1 µM.
9
2 0 6 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004

the Rayleigh equation (36),
(R-1)
(1 + R0)
(1 + R)
R
R0
ln
) ln f
[
]
(1 + R0)
(1 + R)
) ln f
ꢀ ≈ ln f ꢀ ) ꢀ ln f (1)
[
]
with
δ13CCCl
δ13C0,CCl
R
R0
ln ) ln
4 + 1 /
4 + 1
(2)
(
) (
)
]
[
1000
1000
where R is the isotope ratio 13C/ 12C in CCl4 at time t, δ13C the
isotopic enrichment (in ‰ ) with respect to the international
VPDB standard, and f is the fraction of substrate remaining
at time t, where the subscript index “0” denotes values at
time 0 (start of experiment). Because CCl4 contains only one
C atom, all carbon isotopes of the molecule take part in the
reaction, and the fractionation factor R obtained by a
Rayleigh-type analysis of experimental data is directly
equivalent to the inverse kinetic isotope effect at the reactive
site
FIGURE 1. Theoretical dependence of the kinetic carbon isotope
effect, 12k/13k, on the extent of bond cleavage in the transition state,
for one (solid line) and two broken bonds (dashed line). Small
arrow diagrams: activation energies of isotopic bonds in a reactant-
like transition state (left), a symmetric transition state (center) and
a product-like transition state (right). Note that the difference
between the activation energies for isotopic bonds (arrows) becomes
larger for later transition states.
R ) (12k/ 13k)-1
(3)
where 12k and 13k are the rate constants for C-Cl bond
cleavage involving 12C and 13C carbon isotopes, respectively.
Such kinetic isotope effects are independent of the
absolute value of the activation energy, but strongly influ-
enced by changes in bond strength affecting the vibrational
zero-point energy levels in the transition state (TS) (38).
Reductive dehalogenation of CCl4 proceeds via a dissociative
electron transfer, that is, electron transfer and C-Cl bond
cleavage occur simultaneously (39). In such a simple bond-
breaking process, the kinetic isotope effect, KIE, increases
steadily from unity for a very early () reactant-like) TS to a
maximum for a very late () product-like) TS [see Figure 1
and, e.g., refs 40,41]. The approximate magnitude of such
maximum values can be estimated by calculating Streitwieser
limits (40), which take into account the contribution of zero-
point energy differences to the overall kinetic isotope effect.
GC-C-IRMS Analysis. Concentrations and isotope ratios
in isotope fractionation experiments were determined on a
GC-C-IRMS (gas chromatograph with combustion unit and
isotope ratio mass spectrometer, Finnigan MAT deltaplusXL),
with a Combi Pal autosampler (CTC Analytics) and a 60 m
× 0.32 mm Rtx-VMS column (Restek, Bellefonte, PA). On-
column injection was used to introduce 1 µL of sample into
the column. The temperature program was 40 °C (2 min), 8
°C/ min, 100 °C (5 min), 20 °C/ min, 210 °C (8 min). All
measurements were run in triplicate (three samples from
the extract of the same reaction vial). δ13C values for CCl4
were automatically determined relative to a CO2 reference
gas. Concentrations of CCl4 were determined by the peak
area ratio of CCl4 and the internal standard benzene using
IRMS data. This yielded calibrations with R2 of 0.998 (five
calibration standards, Fe(II)porphyrin, and polysulfide) and
0.996 (four calibration standards, goethite/ Fe(II)), respec-
tively. The goethite used contained trace contaminations of
ethyl acetate, which could not be completely baseline-
separated from the analyte peak of CCl4. Their influence on
δ13C values of CCl4 was corrected for by an empirical function
obtained from the calibration standards
The calculated value for cleavage of a C-Cl bond (750 cm-1
)
is ∼(12k/ 13k)max ) 1.057 (40). This means that the KIE in
reductive cleavage of a single C-Cl bond increases from 1.000
for a completely reactant-like TS to ∼1.057 for a completely
product-like TS and will in most cases adopt intermediate
values (see Figure 1). If, however, two C-Cl bonds are broken
at the same time, then isotopic energy differences will be
manifested in two weakened bonds in the transition state,
and the KIE can be expected to be about twice as large (1.0572
) 1.117 for the maximum KIE and correspondingly lower
values for earlier transition states, see Figure 1). A vibrational
analysis confirms this conclusion: The reactant CCl4 (point
group Td) possesses a triply degenerate asymmetric stretching
mode (T2) at 776 cm-1 (42), whereas the product :CCl2 (point
group C2v) has a nondegenerate asymmetric stretching mode
(B2) at ∼750 cm-1 (43). The two asymmetric stretching modes
that are present in the reactant, but not in the product,
correspond to the two vibrations that will be weakened and
become nonperiodic reaction motions in the transition state
if two C-Cl bonds are cleaved simultaneously.
1
2
δ13Cmeasured - δ13Ctrue
)
(-2.3153 ln(I44) + 18.439)
where I44 is the intensity of the ion beam at the mass 44. The
uncertainty of this correction is reflected in a larger error in
the goethite/ Fe(II) system compared to experiments with
Fe(II) porphyrin and polysulfide. Likewise, impurities of
tetrahydrofuran in the diethyl ether coeluted with CHCl3.
δ13C data for this compound could therefore not be obtained.
Kinetic isotope effects were calculated from concentrations
and δ13C values of the substrate CCl4 according to the Rayleigh
equation [(36), see Figure 3 and S3 in the Supporting
Information], with progression of experimental errors in
the x and y directions using the software Lin2d (37). The
total experimental error, however, is likely to be overestimated
by this method so that the errors reported for the result-
ing enrichment factors describe a confidence interval of
>95%.
Results and Discussion
Characterization of the Goethite/Fe(II) System with Respect
to Fe(II) Sorption. At pH 7.0 the experimentally determined
sorption capacity of Fe(II) on goethite was 7-12 µmol/ m2
(or 4.2-7.2 sites/ nm2) for suspensions containing 12.5-100
m2/ L goethite in the presence of 1 mM dissolved Fe(II) (Figure
Theoretical Background. Kinetic isotope fractiona-
tion can be expressed by the fractionation factor R or
the enrichment factor (ꢀ ) R - 1) evaluated according to
9
VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 0 6 1

FIGURE 2. (a) Fe(II) sorption isotherms (left) and corresponding “proton release” isotherms (right) at pH 7, calculated from changes in
pH values upon Fe(II) addition. b, 12.5 m2/L; +, 25 m2/L; 0, 50 m2/L; 2, 100 m2/L mineral loading.
2a). The observed sorption isotherms of Fe(II) revealed several
peculiarities:
was nonlinear in terms of sorbed Fe(II). This elevated
reactivity could not be explained by a higher number of
isolated S-OFe(II)OH species alone, because otherwise,
reaction rates would have depended on their concentration
per volume solution, not, as observed, on the sorption density
per surface area. Amonette et al. therefore suggested that a
simultaneous transfer of two single electrons took place from
two adjacent Fe(II) surface sites, thereby leading to a rate
law that was second-order with respect to sorbed Fe(II) and
first-order with respect to CCl4 (6). This explanation, however,
must be discarded in our systems for which we present
evidence for an initial single-electron transfer (see below).
We suggest that an elevated Fe(II) sorption density is
instrumental, primarily, in creating a higher concentration
of reduced species at the goethite surface. This accumulation
may then facilitate the initial transfer of one single electron
to CCl4. [An analogous case (single-electron transfer/ second-
order kinetics) has also been observed in the reduction of
CCl4 by Cr(II) (46).] Our evidence for surface-Fe(II)OH
complexes rather than surface-Fe(II)+ and, thus, the absence
of repulsive electrostatic effects corroborates the build-up
of such high Fe(II) densities. This accumulation of reduction
equivalents could finally correspond to the slow rearrange-
ment of surface bound Fe(II) postulated by Pecher et al. (8)
and be a first step in the generation of surface precipitates
observed by these authors at higher pH.
13C Isotope Fractionation Study To Evaluate the Num ber
of Bonds Broken in the Initial Reaction Step. The carbon
kinetic isotope effect for the reductive dehalogenation of
CCl4 by goethite/ Fe(II) was investigated and compared to
the KIE observed for two other model reductants, Fe(II)
porphyrin (single-electron transfer, one C-Cl bond broken)
and polysulfide (single- and two-electron transfer, one C-Cl
bond broken). Evaluation of experimental data according to
the Rayleigh equation (Figure 3 and S3) gave kinetic isotope
effects of 1.027 for the Fe(II) porphyrin and Fe(II)/ goethite
systems and 1.023 for the reductant polysulfide, indicative
of cleavage of a single C-Cl bond with ∼50 and 40% bond
weakening in the transition state (see Table 1 and Figure 3c).
Although the difference between both values is statistically
significant, they are remarkably similar if the wide range of
isotope effects, 12k/ 13k, is considered that may be expected
for this reaction (see Figure 3). Specifically, if two C-Cl bonds
were cleaved simultaneously, the kinetic isotope effect would
be expected to be twice as pronounced, around 1.046-1.054
(see Figures 1 and 3c), or if both processes took place in
parallel, a weighted average would be observed. From the
KIE of 1.027 for goethite/ Fe(II) it can thus be concluded that
generally only one C-Cl bond was broken in the initial step
of the reaction with goethite/ Fe(II) and that a simultaneous
(1) Effect of Mineral Loading. When increasing amounts
of goethite were investigated in equilibrium with a constant
concentration of 1 mM dissolved Fe(II), the amount of sorbed
Fe(II)/ m2 surface area was found to decrease at higher mineral
loading. For instance, at a 4 times higher goethite loading,
the amount of adsorbed Fe(II) increased only by a factor of
2.3 (Figure 2a). Such a phenomenon has previously been
ascribed to particle coagulation and, thus, a decrease of
accessible surface area [(6); for a more detailed discussion,
see Supporting Information].
(2) Continued Sorption at High Fe(II) Loadings. After an
apparent Fe(II)ads saturation of the goethite surface at 500
µM Fe(II) in solution, uptake of dissolved Fe(II) by goethite
continued at higher concentrations of 1 mM Fe(II)aq (Figure
2a). In a previous study, this effect has been ascribed to the
onset of surface precipitation processes [(8); for a more
detailed discussion, see Supporting Information].
(3) Stoichiometry of Fe(II) Sorption and Proton Release.
From the amount of base that was necessary to adjust the
pH after addition and consecutive sorption of Fe(II) to
goethite, it was possible to quantify the proton release during
the sorption process (see Figure 2b). These calculations made
use of the buffer intensity determined beforehand for the
goethite/ Fe(II) suspensions of this study (see Materials and
Methods as well as Supporting Information). Calculated
values of R ) (H+ released)/ (Fe2+ sorbed) were consistently
around 2 (Figure 2a/ b), which provides strong experimental
evidence for an Fe(II) sorption reaction of the type
surface-O-H + Fe2aq+ + H2O f
surface-O-Fe-OH + 2 H+ (4)
The same surface speciation and sorption stoichiometry
has been postulated in previous studies on the basis of surface
complexation models [(25,44,45); for a more detailed discus-
sion, see Supporting Information]. The buffer effect observed
in goethite/ Fe(II) suspensions can thus be rationalized in
terms of ad- and desorption of Fe(II) at the surface according
to eq 4. Moreover, the formation of the neutral surface
complexes “surface-O-Fe-OH” suggests that further ad-
sorption of Fe(II) from solution is not prevented by elec-
trostatic repulsion. Thus, high sorption densities may be
expected, and possibly even surface precipitation at already
slightly alkaline pH and higher concentrations of Fe(II)aq.
Effect of Fe(II) Sorption on Dehalogenation Reactivity.
Both Amonette et al. (6) and Pecher et al. (8) found that
higher sorption densities of Fe(II) led to an enormous
reactivity increase with respect to CCl4 dehalogenation, which
9
2 0 6 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004

TABLE 1. Measured and Expected Isotope Effects in Reductive Dehalogenation of CCl4
E (‰)a
r ) 1 + Ea
r-1 12k/13k
)
iron(II) porphyrin (single electron transfer, one C-Cl bond broken)
polysulfide (partial two electron transfer, one C-Cl bond broken)
goethite/Fe(II)
-26.10 ( 1.25
-22.22 ( 1.73
-26.48 ( 2.79
0.973 90 ( 0.001 25
0.977 78 ( 0.001 73
0.973 52 ( 0.002 79
1.0268
1.0227
1.0272
expected range for two broken C-Cl bonds
1.046-1.054
a
Uncertainties are given as 95% confidence intervals.
two-electron transfer/ two-bond cleavage process must be
discarded.
Products in the Dehalogenation of CCl4. Dehalogenation
of CCl4 by goethite/ Fe(II) at pH 7 produced 33% CHCl3, 20%
CO, and up to 40% HCOO- (see Figure 4a) Other volatile
halogenated compounds, such as CH2Cl2, C2Cl6, or C2Cl4,
were not detected. In a second experiment, 3 vol % per-
deuterated 2-propanol (2-propanol-d8) was added, which
has a weak tertiary C-D bond that facilitates abstraction of
•
D atoms by •CCl3 radicals (16). [Note that CCl3 radicals are
difficult to detect otherwise, as most alternative radical traps,
such as nitroxyl radicals, would rapidly be reduced in our
system, and analysis by EPR (electron paramegnatic reso-
nance spectroscopy) is not sensitive enough.] The formation
of 33% CDCl3 and only 11% CHCl3 (Figure 4b) in the presence
of 2-propanol-d8 indicates (a) that the total yield of chloroform
(CHCl3 and CDCl3) was considerably higher due to addition
•
of the D donor, and (b) that a strong competition between
•H and •D donors took place in the system, leading to a much
lower yield of CHCl3. The results provide strong evidence
•
that CCl3 radicals were intermediates in the reaction and
that, thus, an initial single-electron-transfer took place
(Scheme 1).
Formation of CHCl3. Significant fractions of CHCl3 were
formed at all experimental conditions studied, despite our
attempts to remove organic residues and, thus, possible
alternative H• donors from the experimental setup (see
Materials and Methods). A possible pathway to explain this
result involves the transfer of a second single electron to an
•
initially formed CCl3 radical, leading to a trichloromethyl
carbanion (:CCl3-) (see Scheme 1). In analogy to Scheme 2,
:CCl3- can either be protonated, thus explaining formation
of chloroform, or decompose to dichlorocarbene (19), which
reacts further to CO (Scheme 3). Results by Pecher et al. (8),
however, rule out the protonation of :CCl3-, because no CDCl3
was formed in experiments performed in deuterated water
(D2O). This leaves hydrogen (H•) transfer to trichloromethyl
radicals (Scheme 1) as the most likely process, despite the
enforced depletion in organic H• donors. Traces of organic
matter at the surface of goethite or present in the deionized
water (foc e 0.1 mg C/ L) supposedly reacted as H• donor with
•CCl3 radicals. This indicates that minute amounts of organic
matter may have affected surface-catalized reductive deha-
logenation of CCl4, shifting product distribution toward
chloroform.
Formation of CO. Addition of perdeuterated 2-propanol
strongly decreased the yield of carbon monoxide (Figure 4a/
•
b). The pathway to CO must, therefore, also involve CCl3
radicals as intermediates, since otherwise, CO formation
would not have been affected by the presence of the radical
scavenger. The most plausible pathway from •CCl3 to CO
entails transfer of a second electron, leading to :CCl3-; then
:CCl2 (19); HCOCl (20); and finally, CO (21) (see Schemes 1
and 2). Free trichloromethyl anions (:CCl3-) in solution,
however, must again be discarded on grounds of experiments
by Pecher et al. (8) with D2O (see above). This indicates that
FIGURE 3. (a) Changes in concentration and isotopic signature of
CCl4 during reduction by goethite/Fe(II); (b) Evaluation of E according
to the Rayleigh equation; (c) Measured and expected isotope effects
in reductive dehalogenation of CCl4 with different reductants
investigated in this study.
-
:CCl3 may be stabilized in some way, most probably as a
short-lived surface complex. Evidence for such a stabilization
by complex formation with Fe centers comes from experi-
ments with microsomal cytochrome P-450, where carbenes
9
VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 0 6 3

FIGURE 4. Time course and product yield in the reaction of CCl4 with goethite/Fe(II) (50 m2/L mineral, 23 °C), (a) at pH 7, (b) at pH 7 with
added perdeuterated 2-propanol and (c) at pH 8. (, CCl4; 4, CHCl3; 3, CDCl3; 9, HCOO-; and b, CO. Crosses indicate molar balance. Lines
are exponential fits with pseudo-first-order rate constants (a) 0.0076, (b) 0.0092, and (c) 0.0145 h-1 and product yields (a) 33% CHCl3/18%
CO, (b) 33% CDCl3/13% CHCl3/4% CO, and (c) 40% CHCl3.
and radicals were found to form iron(II) porphyrin complexes
(47) (see also ref 48), as opposed to free radicals, carbanions,
and carbenes in solution. One possibility could therefore be
the formation of a surface complex of the type
balance observed is attributable primarily to microbial
oxidation of formate to CO2, because the carbon-based mass
balance was essentially complete in cases where HCOO- was
found (see Figure 4a), and other volatile halogenated
compounds were not detected. This assumption is cor-
roborated by recent findings that the iron reducing micro-
organism Geobacter metalireducens (GS 15) is able to grow
on formate present at low concentrations (49). Production
of formate cannot be rationalized by hydrolysis of formyl
chloride (HCOCl, see Scheme 2) because decomposition of
this intermediate leads to CO at circumneutral pH (21).
Generation of formate by hydrolysis of CO at pH 7 is
thermodynamically favorable (∆Gr° ) -73.9 kJ/ mol, see
Supporting Information), but was found to be very slow in
homogeneous solution, with a half-life of 275 years at our
experimental conditions (k ) 8 × 10-4 L mol-1 s-1, (50) where
k ) d[HCOO-]/ dt ) [OH-]-1[CO]-1 and species in brackets
denote aqueous concentrations). To test for surface catalysis
by goethite, CO hydrolysis was measured at pH 7 in pure
water and in the presence of goethite/ Fe(II). In both
experiments, the headspace contained pure carbon monoxide
so that the aqueous concentration of CO corresponded to a
saturated solution [17.7 ppm or 9.86 × 10-4 M (51)]. Formate
could be detected neither in batches with pure water nor in
controls of goethite without CO, whereas 3.2-4.8 µM formate
was formed in the presence of goethite/ Fe(II) after 36-67
days (see Figure S4 in the Supporting Information), indicating
surface catalysis of CO hydrolysis by goethite/ Fe(II). A rough
-
+ e , - Cl
[S-Fe(II) •CCl3 T S-Fe(III) - CCl3]
-8
[S-Fe(II) ) CCl2]
as a short-lived intermediate in the reaction, where S-Fe(II)
denotes surface entities. Formation of CO would thus be
-
possible without transient occurrence of free :CCl3 in
solution.
Formation of HCOO-. Formate was detected only in some
of the reaction vials (see Figure 4), which is consistent with
earlier experiments at lower CCl4 concentrations (data not
shown). Contamination can be ruled out, because formate
was not found in blanks prepared according to the general
strict washing protocol. On the contrary, HCOO- often
disappeared in calibration standards (identical in composi-
tion to the reaction vials) with a half-life of 3-4 days. It is
therefore likely that formate was produced in all vials to the
same extent, but disappeared from some of them in a
subsequent reaction. Since all vials were set up identically
with regard to their chemical content, but were not sterile
(owing to their preparation in a glovebox and the fact that
mineral suspensions cannot be filter-sterilized), microbial
degradation of formate to CO2 is a likely cause of the
disappearance. We hypothesize that a deficit in the mass
9
2 0 6 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004

and (2) transfer of one electron. This initial reaction was
found to generate partly trichloromethyl radicals and to lead
partly to formate, possibly via surface-bound intermediates.
Consequently, unwanted chloroform was found to be
produced mainly in consecutive radical reactions with H•
donors, whereas surface stabilization of short-lived inter-
mediates may play a key role in the alternative formation of
benign products. Product formation pathways, therefore, do
not primarily depend on the competition between an initial
one- and two-electron transfer, but rather on the presence
of different radical scavengers and the properties of the
mineral surface with respect to stabilization of reaction
intermediates. The presence of oxygen and sulfur species
can thus lead to completely dehalogenated products in radical
reactions (11,16), whereas even trace amounts of organic
matter may result in chloroform formation, as indicated by
the results of this study. In addition, specific adsorption of
major anions or pH effects may modify the capability of the
goethite surface to stabilize short-lived radical intermediates.
The key to predicting product formation in reductive
dehalogenation of CCl4 by Fe(II) is, therefore, a profound
understanding of the factors that may determine the
stabilization of radical intermediates at reactive Fe(II) surface
sites. Further research is needed to address this topic,
including the effects of coadsorbates, such as inorganic ions
and natural organic matter.
FIGURE 5. Summary of mechanistic evidence obtained for the
reaction of CCl4 with goethite/Fe(II). Compounds in boxes were
detected during the reaction. Bold arrows and ticks designate
reaction paths for which experimental evidence (written in italic
letters) could be obtained, whereas dashed arrows and crosses
indicate pathways that must be excluded. Question marks point out
hypothesized pathways and intermediates.
estimate of the second-order rate constant (k ) d[HCOO-]/
dt ) [OH-]-1[CO]-1 ) 1 × 10-2 L mol-1 s-1) based on these
results reveals that the reaction was accelerated by the
presence of goethite by a factor of ∼10, which is, however,
by far not sufficient to explain the amount of formate observed
in the dehalogenation experiments of CCl4. The pathway CO
f HCOO- can therefore be discarded, in accordance with
the observation that the radical scavenger 2-propanol-d8
influenced CO formation but had no effect on the missing
mass balance () product fraction attributable to formate)
(see Figure 4b). Formate is therefore likely to be formed by
an alternative pathway that does not involve free trichlo-
romethyl radicals as intermediates, possibly because of
specific stabilization of intermediates at the surface.
Influence of pH on Product Distribution. Pecher et al. (8)
observed that higher Fe(II)-surface-site densities at constant
pH decreased the fraction of CHCl3 formed, whereas an
increase of both sorption site density and pH led to increased
chloroform production. To investigate the influence of pH
alone, experiments were conducted at pH 7 and 8, with similar
sorption site densities of Fe(II) on goethite. The chloroform
yield increased slightly from 33% at pH 7 to ∼40% at pH 8
(see Figure 4c). Direct involvement of H+ in the reaction of
CCl4 to CHCl3 is very unlikely, because H2O can be expected
to be the most important proton donor at circumneutral pH.
As discussed above, chloroform is formed by hydrogen (H•)
transfer to trichloromethyl radicals rather than by H+ transfer
to trichloromethyl anions. An increase of pH, however,
changes the surface chemistry of goethite, leading, for
example, to deprotonation of surface hydroxyl groups accord-
ing to dS-OH + OH- f dS-O- + H2O. Such processes may
affect the formation of surface stabilized intermediates and
thus discriminate against alternative pathways in favor of
chloroform production. Product formation can, therefore,
be expected to be primarily influenced by changes in surface
chemistry and only indirectly by pH.
Acknowledgments
We thank Lynn Roberts, Klaus Pecher, and Thomas Hofstetter
and two anonymous reviewers for helpful comments on the
manuscript. This work was funded in part by the Swiss
National Science Foundation (Grant No. 21-57171.99), the
DOW Foundation Supporting Public Health and Environ-
mental Research Effects (SPHERE) program, and by the Swiss
Federal Office for Education and Science (Grant No. BBW
99.0396) as part of the European research project “Protection
of Groundwater Resources at Industrially Contaminated
Sites” (PURE, EVK1-CT-1999-00030).
Supporting Information Available
Four figures showing (1) UV-vis spectra of Fe(III) and Fe(II)
porphyrin, (2) buffer intensity in goethite/ Fe(II) suspensions,
(3) changes in concentration and isotopic signature of CCl4
during reduction by iron(II) porphyrin and polysulfide
(Rayleigh plots), and (4) reaction of CO to formate in
suspensions of goethite/ Fe(II). In addition, information about
the dependence of Fe(II) sorption on mineral loading,
characterization of Fe(II) surface complexes, and discussion
of possible reactive species; thermodynamic and kinetic data
on the conversion of CO to formate. This material is available
free of charge via the Internet at http:/ pubs.acs.org.
Literature Cited
(1) Arneth, J.-D.; Milde, H.; Kerndorff, H.; Schleyer, R. In The Landfill,
Lecture Notes in Earth Sciences; Baccini, P., Ed.; Springer-
Verlag: Berlin, 1989; Vol. 20 P, p 399.
(2) Christensen, T. H.; Kjeldsen, P.; Albrechtsen, H.-J.; Heron, G.;
Nielsen, P. H.; Bjerg, P. L.; Holm, P. E. Crit. Rev. Environ. Sci.
Technol. 1994, 24, 119-202.
(3) Moniel, A.; Rauzy, S. Toxicol. Environ. Chem. 1985, 10, 315-
320.
(4) Plumb, R. H. Ground Water Mon. Rev. 1991, 11, 157-164.
(5) Matheson, L. J.; Tratnyek, P. G. Environ. Sci. Technol. 1994, 28,
2045-2053.
Environmental Significance. By applying various comple-
mentary experimental techniques, this study provides new
mechanistic information for a better understanding of
surface-mediated reductions of chlorinated aliphatic con-
taminants by Fe(II) (see Figure 5). Our results indicate that
the first step in reductive dehalogenation of CCl4 by surface-
bound Fe(II) involves (1) cleavage of only one C-Cl bond
(6) Amonette, J.; Workman, D. J.; Kennedy, D. W.; Fruchter, J. S.;
Gorbi, Y. A. Environ. Sci. Technol. 2000, 34, 4606-4613.
(7) Criddle, C. S.; McCarty, P. L. Environ. Sci. Technol. 1991, 25,
973-978.
(8) Pecher, K.; Haderlein, S. B.; Schwarzenbach, R. P. Environ. Sci.
Technol. 2002, 36, 1734-1741.
(9) Kriegman-King, M. R.; Reinhard, M. Environ. Sci. Technol. 1992,
26, 2198-2206.
9
VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 0 6 5

(10) Kriegman-King, M. R.; Reinhard, M. Environ. Sci. Technol. 1994,
28, 692-700.
(11) Wagner, A. J.; Vecitis, C.; Fairbrother, D. H. J. Phys. Chem. B
2002, 106, 4432-4440.
(12) Criddle, C. S.; DeWitt, J. T.; Grbic-Galic, D.; McCarty, P. L. Appl.
Environm. Microbiol. 1990, 56, 3240-3246.
(13) Lewis, T. A.; Crawford, R. L. In Novel Approaches for Bioreme-
diation of Organic Pollution; Fass, Ed.; Kluwer Academic: New
York, 1999; pp 1-11.
(14) Holliger, C.; Schraa, G. Microbiol. Rev. 1994, 15, 297-305.
(15) Monig, J.; Bahnemann, D.; Asmus, K.-D. Chem.-Biol. Interact.
1983, 47, 15.
(31) Kaiser, K.; Guggenberger, G. Eur. J. Soil Sci. 2003, 54, 219-236.
(32) Fadrus, H.; Maly, J. Analyst 1975, 100, 549-554.
(33) Wade, R. S.; Castro, C. E. J. Am. Chem. Soc. 1973, 95, 226-230.
(34) Perlinger, J.; Angst, W.; Schwarzenbach, R. P. Environ. Sci.
Technol. 1996, 30, 3408-3417.
(35) Schaller, K.-H.; Triebig, G. In Methods of Enzymatic Analysis;
Bergmeier, H. U., Ed.; Verlag Chemie: Weinheim, 1984; Vol. 6,
pp 668-672.
(36) Mariotti, A.; Germon, J. C.; Hubert, P.; Kaiser, P.; Letolle, T.;
Tardeux, A.; Tardieux, P. Plant Soil 1981, 62, 413-430.
(37) Graf, T. Ph.D. Dissertation, ETH Zu¨ rich, Zu¨ rich, 1988; p 134.
(38) Melander, L.; Saunders: W. H. Reaction Rates of Isotopic
Molecules, 2nd ed.; John Wiley & Sons: New York, 1980.
(39) Costentin, C.; Robert, M.; Saveant, J.-M. J. Phys. Chem. A 2000,
104, 7492-7501.
(40) Huskey, W. P. In Enzyme Mechanism from Isotope Effects; Cook,
P. F., Ed.; CRC Press: Boca Raton, Ann Arbour, Boston, London,
1991.
(16) Buschmann, J.; Angst, W.; Schwarzenbach, R. P. Environ. Sci.
Technol. 1999, 33, 1015-1020.
(17) Balko, B. A.; Tratnyek, P. G. J. Phys. Chem. B 1998, 102, 1459-
1465.
(18) Zefirov, N. S.; Makhon’kov, D. I. Chem. Rev. 1982, 82, 615-624.
(19) Hine, J.; Ehrenson, S. J. J. Am. Chem. Soc. 1958, 80, 824-830.
(20) Pliego, J. R.; De Almeida, W. B. J. Phys. Chem. 1996, 100, 12410-
12413.
(21) Dowideit, P.; Mertens, R.; von Sonntag, C. J. Am. Chem. Soc.
1996, 118, 11288-11292.
(41) Cook, P. F. Isot. Environ. Health Stud. 1998, 34, 3-17.
(42) Lide, D. R. CRC Handbook of Chemistry and Physics, 76th ed.;
CRC Press: Boca Raton, Ann Arbour, Boston, London, 1995.
(43) Cheong, B.-S.; Cho, H.-G. J. Phys. Chem. A 1997, 101, 7901-
7906.
(44) Coughlin, B. R.; Stone, A. T. Environ. Sci. Technol. 1995, 29,
2445-2455.
(22) McCormick, M. L.; Bouwer, E. J.; Adriaens, P. Environ. Sci.
Technol. 2002, 36, 403-410.
(23) Buerge, I. J.; Hug, S. J. Environ. Sci. Technol. 1999, 33, 4285-
4291.
(24) Klausen, J.; Tro¨ ber, S. P.; Haderlein, S. B.; Schwarzenbach, R.
P. Environ. Sci. Technol. 1995, 29, 2396-2404.
(25) Liger, E.; Charlet, L.; Van Capellen, P. Geochim. Cosmochim.
Acta 1999, 63, 2939-2955.
(26) Elsner, M.; Schwarzenbach, R. P.; Haderlein, S. B. Environ. Sci.
Technol. 2004, 38, 799-807.
(45) Charlet, L.; Silvester, E.; Liger, E. Chem. Geol. 1998, 151, 85-93.
(46) Castro, C. E.; Kray, W. C. J. Am. Chem. Soc. 1966, 88, 4447.
(47) Ahr, H. J.; King, L. J.; Nastainczyk, W.; Ullrich, V. Biochem.
Pharmacol. 1980, 29, 2855-2861.
(48) Ziegler, C. J.; Suslick, K. S. J. Am. Chem. Soc. 1996, 118, 5306-
5307.
(27) Ru¨ gge, K.; Hofstetter, T.; Haderlein, S. B.; Bjerg, P. L.; Knudsen,
S.; Zraunig, C.; Mosbaek, H.; Christensen, T. Environ. Sci.
Technol. 1998, 32, 23-31.
(49) Straub, K. University of Konstanz, Germany; Personal com-
munication.
(50) Robinson, E. A. J. Chem. Soc. 1961, 1663-1671.
(51) Rettich, T. R.; Battino, R.; Wilhelm, E. Ber. Bunsen-Ges. Phys.
Chem. 1982, 86, 1128-1132.
(28) Kenneke, J. F.; Weber, E. J. Environ. Sci. Technol. 2003, 37, 713-
720.
(29) Scherer, M. M.; Balko, B. A.; Tratnyek, P. G. In Kinetics and
Mechanisms of Reactions at the Mineral/Water Interface; Sparks,
D. L., Grundl, T., Eds.; ACS Symposium Series; American
Chemical Society: Washington, DC, 1999; Vol. 715, Chapter 15,
pp 301-320.
Received for review July 10, 2003. Revised manuscript re-
ceived December 12, 2003. Accepted December 19, 2003.
(30) Evanko, C. R.; Dzombak, D. A. J. Colloid Interface Sci. 1999, 214,
189-206.
ES034741M
9
2 0 6 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004