Pd-catalyzed TCE dechlorination in water: Effect of [H2](aq) and H2-utilizing competitive solutes on the TCE dechlorination rate and product distribution

Article abstract of DOI:10.1021/es001623f

The aqueous-phase H₂ concentration ([H₂](aq)) and the presence of H₂-utilizing competitive solutes affect trichloroethylene (TCE) dechlorination efficiency in Pd-based in-well treatment reactors. Batch kinetic studies in m

Full text of DOI:10.1021/es001623f

Environ. Sci. Technol. 2001, 35, 696-702
water, Pd catalysts rapidly dehalogenate a variety of com-
Pd-Catalyzed TCE Dechlorination in
monly encountered halogenated groundwater contaminants
including chlorinated ethylenes (PCE, TCE, DCE, vinyl
chloride), chlorinated methanes (carbon tetrachloride, chlo-
roform), chlorinated aromatic compounds (1,2-dichloroben-
zene), chlorinated biphenyls (4-chlorobiphenyl), chloroflu-
orocarbons (Freon-113), and halogenated pesticides (1,2-
dibromo-3-chloropropane (DBCP) and lindane) without
forming substantial amounts of chlorinated intermediate
compounds (1-4). Common groundwater solutes such as
bicarbonate, sulfate, and chloride have little effect on catalyst
activity during the aqueous-phase destruction of TCE.
Laboratory-scale column reactor experiments indicate that
some groundwater solutes reduce catalyst activity, such as
sulfite and bisulfide, but the adverse effects are reversible
and catalyst activity is sustainable under groundwater
treatment conditions (5). An in-well Pd-based pilot-scale
fixed-bed reactor has been used to treat halogenated
hydrocarbon contaminated groundwater for more than 1
year at Lawrence Livermore National Laboratory without
significant loss of catalyst activity (6). All the research efforts
mentioned above were performed under hydrogen-rich
conditions, and no work has been published regarding the
effects of the aqueous hydrogen concentration ([H2](aq)) on
the contaminant transformation rate and product distribu-
tion or the conditions under which H2 deficiencies might
occur, such as the presence of additional H2-utilizing
substrates.
Only trace quantities of undesirable halogenated inter-
mediate compounds form under H2-rich conditions (1), but
production may increase under hydrogen-limited conditions
because the first step of the hydrodehalogenation reaction,
removal of the halogen, is reversible (7). Trace amounts of
butane and hexane have been observed during TCE trans-
formation using palladized iron (8), and ethane and propane/
propene have been observed during transformation of carbon
tetrachloride under H2-rich conditions (1), indicating that a
free radical mechanism exists. Under H2-limited conditions,
isomerization and polymerization side reactions may increase
significantly and potentially lead to catalyst deactivation if
these less reactive unsaturated C4-C6 hydrocarbon byprod-
ucts remain on the catalyst surface, effectively blocking active
catalyst sites.
Water: Effect of [H ](aq) and
2
H -Utilizing Competitive Solutes on
2
the TCE Dechlorination Rate and
Product Distribution
G R E G O R Y V . L O W R Y A N D
M A R T I N R E I N H A R D *
Department of Civil and Environmental Engineering,
Stanford University, Stanford, California 94305-4020
The aqueous-phase H concentration ([H ](aq)) and the
2
2
presence of H -utilizing competitive solutes affect TCE
2
dechlorination efficiency in Pd-based in-well treatment
reactors. The effect of [H ](aq) and H -utilizing competing
2
2
solutes (cis-DCE, trans-DCE, 1,1-DCE, dissolved oxygen
(DO), nitrite, nitrate) on the TCE transformation rate and
product distribution were evaluated using 100 mg/L of a
powdered Pd-on-Al2O catalysts in batch reactors or 1.0 g
3
of a 1.6-mm Pd-on-γ-Al2O catalyst in column reactors.
3
The TCE dechlorination rate constant decreased by 55%
from 0.034 ( 0.006 to 0.015 ( 0.001 min-1 when the [H ](aq)
2
decreased from 1000 to 100 µM and decreased sharply
to 0.0007 ( 0.0003 min-1 when the [H ](aq) decreased from
2
100 to 10 µM. Production of reactive chlorinated
intermediates and C4-C6 radical coupling products increased
with decreasing [H ](aq). At an [H ](aq) of 10 µM (P/Po
2
2
) 0.01), DCE isomers and vinyl chloride accounted for as
much as 9.8% of the TCE transformed at their maximum
but disappeared thereafter, and C4-C6 radical coupling
products accounted for as much as 18% of TCE transformed.
The TCE transformation rate was unaffected by the
presence of cis-DCE (202 µM), trans-DCE (89 µM), and 1,1-
DCE (91 µM), indicating that these compounds do not
compete with TCE for catalyst active sites. DO is twice as
reactive as TCE but had no effect on TCE conversion in
the column below a concentration of 370 µM (11.8 mg/L),
indicating that DO and TCE will not compete for active
catalyst sites at typical groundwater DO concentrations.
TCE conversion in the column was reduced by as much as
a factor of 10 at influent DO levels greater than 450 mM
Groundwater contamination by halogenated hydrocar-
bons often involves multiple contaminants at one site, as in
ref 6, rather than a single contaminant, and Pd catalysts
readily transform many of these contaminants (1-4). Dis-
solved oxygen (DO) is also a constituent of most surface water
and groundwater, and palladium catalyzes the rapid reaction
between oxygen and hydrogen to form water (eq 1) (9):
(14.3 mg/L) because the [H ](aq) fell below 100 µM due to
H utilized in DO conversion. Nitrite reacts 2-5 times
2
2
2H2 + O2
9
Pd8 2H2O
(1)
slower than TCE and reduced TCE conversion by less than
4% at a concentration of 6630 µM (305 mg/L). Nitrate
was not reactive and did not effect TCE conversion at a
concentration of 1290 µM (80 mg/L).
Similarly, Pd/ Cu and Pd catalysts in the presence of dissolved
H2 catalyze the aqueous-phase transformation of nitrate and
nitrite to N2 (eqs 2 and 3) (10, 11):
2NO3- + 5H2 Pd/ Cu8 N2 + 2OH- + 4H2O
Pd8 N2 + 2H2O + 2OH-
(2)
(3)
Introduction
2NO2- + 3H2
9
Palladium (Pd)-based catalytic fixed-bed reactors show
promise as a treatment technology for in situ destruction of
halogenated groundwater contaminants. In deionized (DI)
The presence of multiple contaminants, DO, nitrate, or nitrite
may affect the catalyst’s ability to transform TCE by compet-
ing with TCE for active catalyst sites if they utilize the same
active sites as TCE for transformation or by depleting the
* Corresponding author phone: (650)723-0308; fax: (650)723-7058;
e-mail: reinhard@ce.stanford.edu.
9
6 9 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 4, 2001
10.1021/es001623f CCC: $20.00
2001 Am erican Chem ical Society
Published on Web 01/11/2001

TABLE 1. Headspace Reactor Operating Conditions
TABLE 2. H Partial Pressures and Corresponding [H ](aq) and
2 2
Observed TCE Reaction Rate Constants and Half-Lifetimes in
Headspace Reactors
tem p (K)
total pressure (atm )
vol (L)
295 (0.2
1.3
2.0
(P/Po)
[H ](aq) (µM)
2
kobs (min-1
)
t1/2 (min)
aq/vap
agitator speed (rpm )
[Cat] (g/L)
1.5/0.5
1000
0.1
1.0
0.4
0.104
0.04
0.01
1000
400
100
40
0.034 ( 0.006a
0.025 ( 0.004
0.015 ( 0.001
0.0037 ( 0.0005
0.0007 ( 0.0003
20
28
46
187
990
[TCE]o (µM)
140-180
10
a
reactor of H2 if their concentration and reactivity is significant.
Diminished [H2](aq) in the reactor or competitive sorption
onto catalyst active sites may ultimately lead to a lower TCE
transformation rate and increase production of unwanted
halogenated reaction intermediates. For example, TCE
conversion decreased significantly when the reactor influent
H2 concentrations decreased (6), and DO was suggested to
compete with TCE for H2 on the catalyst surface (12).
In this study, batch kinetic studies in model systems are
used to (i) determine the effect of [H2](aq) on the TCE
transformation rate, (ii) assess the potential for production
of halogenated intermediate compounds and C4-C6 radical
coupling products under H2-limited conditions, (iii) evaluate
the role of C4-C6 radical coupling products in catalyst
deactivation, and (iv) determine if the presence of cis-DCE,
trans-DCE, and 1,1-DCE affects the TCE dehalogenation rate.
Column reactors are used to quantify the H2 requirements
for DO, nitrate, and nitrite and to evaluate the effect of these
solutes on the catalyst’s ability to transform TCE to ethane.
Batch studies indicate that the TCE transformation rate
decreases sharply at [H2](aq) less than 100 µM (P/ Po<0.1)
and production of halogenated intermediates and radical
coupling products increases as the [H2](aq) decreases. The
presence of cis-DCE, trans-DCE, and 1,1-DCE did not effect
the transformation rate of TCE. Column reactor studies show
that DO and nitrite are reactive solutes, utilize H2, and can
sometimes interfere with the catalyst’s ability to transform
TCE to ethane.
Errors represent 95% confidence intervals.
FIGURE 1. Column experiment apparatus. V , n, and θ are the
column empty-bed volume, porosity, and fluid retention time,
respectively.
EB
experimental procedure for the headspace reactor are
described in ref 1, but a summary of the experimental
procedure is as follows. The reactor was filled with DI water
and sparged with pure H2 gas or a H2/ N2 gas mixture for 15
min to remove dissolved oxygen and provide a specific
[H2](aq). Pure H2 gas (P/ Po ) 1) and H2/ N2 gas mixes of 40%,
10.4%, 4%, and 1% (P/ Po )0.4, 0.1, 0.04, and 0.01, respectively)
were used to provide five [H2](aq) ranging from 10 to 1000
µM as calculated using a Henry’s constant of 5.34 × 107 Torr
(13). The H2 partial pressures and corresponding calculated
[H2](aq) for the reactor total pressure are given in Table 2.
After sparging, a saturated aqueous TCE (or TCE and DCE
isomers) solution was quickly added to provide an initial
reactant concentration of approximately 160 µM (21 mg/ L).
Pentane (nonreactive) was also added as an internal standard.
The reactants and pentane were agitated and equilibrated
between the vapor and aqueous phase for 1 h before 0.1 g
of powdered Pd-on-Al2O3 was injected to begin the reaction.
The H2 partial pressure was held constant throughout the
experiment to maintain a constant [H2](aq). Headspace
samples (200 µL) were taken from the reactor periodically
during the experiment and analyzed for TCE as well as volatile
reaction intermediates and products as described in ref 1.
Colum n Reactors. The reactors used in this study are
identical to those in ref 5, except that an inlet is inserted
ahead of the column for easy addition of the nitrate or nitrite
solution or dissolved oxygen (DO) in the column influent
(Figure 1). Complete details of the reactor experimental setup
and sampling protocol can be found in ref 5, but a brief
description follows. The feed tank contains DI water amended
with TCE (4 mg/ L) under hydrogen gas (1.24-1.27 atm). A
positive displacement pump drives flow through the column
in an upflow configuration at 1 mL/ min. Concentrated nitrate
or nitrite solutions are added using a 100-mL syringe and
syringe pump (0.01 mL/ min) to give the desired nitrate/
nitrite concentration in the column influent. A separate feed
tank stores DI water amended with TCE (4 mg/ L) under pure
oxygen (1.5 atm). DO concentrations in the feed tank are
Experimental Section
Chem icals. All chemicals were reagent grade and 99+% pure
unless noted. TCE was supplied by Fisher Scientific. 1,1-
Dichloroethylene (DCE), cis-DCE (97%), and trans-DCE (98%)
were supplied by Aldrich. Sodium nitrate (anhydrous) and
sodium nitrite (anhydrous) were supplied by J. T. Baker. Pure
hydrogen gas (zero grade), oxygen gas, and a 100 ppmv H2
in N2 gas standard were supplied by Praxair. Gas standards
for vinyl chloride, 1-butene, n-butane, cis-2-butene, trans-
2-butene, and 2-hexene were supplied by Scott Specialty
Gases (Plumsteadville, PA).
Catalysts. Two prereduced commercially available 1 wt
% Pd-on-Al2O3 catalysts were used. A powdered (38-70 µm)
catalyst supplied by Aldrich was used in batch experiments.
A spherical 1.6-mm edge-loaded (∼70 µm thickness) pal-
ladium-on-γ-Al2O3 catalyst supplied by PMC (Precious Metals
Corporation, Sevierville, TN) was used in the column
experiments. The 1 wt % Pd content specified by the
manufacturer was assumed to be accurate, and no special
precautions were taken to avoid catalyst exposure to air prior
to column experiments. Catalyst physical properties such as
BET surface area, pore volume, and pore size distribution
for the powdered and spherical catalysts are reported in refs
1 and 5, respectively.
Headspace Reactors. The headspace reactor was a gas-
tight 2.0-L glass and stainless steel batch reactor (Bioengi-
neering AG, Wald, Switzerland) equipped with a magnetically
driven mixer, temperature probe, heat exchanger, and
pressure gauge as described in ref 1. Batch reactor operating
conditions are given in Table 1. Complete details of the
9
VOL. 35, NO. 4, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 6 9 7

approximately 63 mg/ L as calculated using a Henry’s constant
of 3.3 × 107 Torr (13). DO is added to the column influent
using a separate feed pump, and DO concentrations at the
column influent are set by adjusting the flow rates on each
feed pump (TCE/ H2 and TCE/ O2) to maintain a total flow of
1 mL/ min.
The reactors are stainless steel tubular fixed-bed reactors
(1.27 cm × 9.8 cm, VEB ) 10.5 mL). A catalyst loading of 1.0
g was used to give approximately 50% steady-state TCE
conversion. The moderate TCE conversion provides maxi-
mum sensitivity to small changes in TCE reactivity. Inert 1.6
mm diameter borosilicate glass beads were packed above
and below the catalyst to improve the flow characteristics in
the column by minimizing end and wall effects. Ideal plug
flow is assumed. Column porosity was 0.44 ( 0.02, giving a
fluid residence time of 4.4 ( 0.2 min in the column. Column
influent and effluent samples were collected and analyzed
for TCE, nitrate/ nitrite, DO, and dissolved H2.
Analytical Methods. Headspace reactor samples (200 µL)
were analyzed for TCE and volatile reaction intermediates
and products using GC/ FID as described in ref 1. The
instrument was calibrated daily using standards prepared
from volumetric dilutions of a stock solution. Halogenated
hydrocarbon (except vinyl chloride) response factors were
calculated based on the total amount of compound added
to the bottle (25% headspace/ 75% water). Calibrations for
C2-C6 nonhalogenated reaction products and vinyl chloride
were determined using certified gas standards (Scott Specialty
Gases).
Column reactor influent and effluent samples (0.5 mL)
were taken using a 1/ 4-28 threaded nose 2.5-mL gas-tight
syringe (Unimetrics) and extracted into 1 mL of hexane
containing 1.3 mg/ kg PCE as an internal standard. One-
microliter samples of the extract were analyzed for TCE using
a HP 5890 series II GC equipped with an ECD as described
in ref 1. Nonhalogenated reaction products were not
measured using this method.
electrode was calibrated using a 2-point calibration with N2-
sparged water as DO ) 0 mg/ L and air-saturated water as
8.8 mg/ L. It was assumed that the electrode response was
linear.
Dissolved hydrogen concentrations were analyzed using
a KAPPA3 model E-001 reduction gas analyzer (Trace
Analytical, Menlo Park, CA). A 2.0-mL column influent/
effluent sample was collected in a 1/ 4-28 threaded nose 5.0-
mL gas-tight syringe, a 2.0-mL headspace was created, and
then the syringe was capped. The syringe was shaken to
equilibrate the vapor and aqueous phases. A 250-µL head-
space sample was removed and diluted (1:100) in a 25-mL
vial capped with a Mininert valve. A 50-µL aliquot of the
diluted headspace sample was then analyzed on the KAPPA3
reduction gas analyzer. The instrument was calibrated daily
using 50-200-µL samples of a 100 ppmv H2 in N2 gas standard
supplied by Praxair (San Carlos, CA)
Evaluation of Reaction Kinetics. A Langmuir-Hinshel-
wood (L-H) kinetic formulation was used to model batch
reactor experimental data that assumes noncompetitive
equilibrium Langmuir type adsorption of TCE and adsorption
of H2 (molecular or dissociative) and that a surface reaction
between adsorbed species is the rate-controlling step:
KnH CnH
dC
TCE ) krxn
dt ccat
KTCECTCE
2
2
-
(4)
n
n
1 + KTCECTCE
(
)(
)
1 + KH CH
2
2
In eq 4, krxn is the apparent reaction rate constant; ccat is the
catalyst concentration; KTCE and KH are the equilibrium
2
adsorption constants for TCE and H2, respectively; CTCE and
CH are the aqueous concentration of TCE and H2, respec-
2
tively; and n is an integer exponent equal to 1/ 2 for
dissociative adsorption of H2 and equal to 1 for molecular
adsorption of H2. A similar approach was successfully used
to describe the kinetics of catalytic liquid-phase hydrogena-
tion of aqueous nitrate solutions using a bimetallic Pd/ Cu
catalyst (14). TCE disappearance was first order in each
experiment at fixed [H2](aq) (Co(TCE) ) 180 µM) and has been
shown to transform via first-order kinetics for initial TCE
concentrations as high as 250 µM (1). This implies low surface
coverage for TCE and that KTCECTCE , 1. For a fixed [H2](aq),
the last term in parentheses in eq 4 is constant. Using these
simplifications and consolidating fixed constants yields a
form useful for determining the influence of the [H2](aq) on
the observed TCE transformation rate constant, kobs (eqs 5
and 6):
Column reactor influent and effluent samples were also
periodically analyzed for halogenated reaction intermediates
(cis-DCE, trans-DCE, 1,1-DCE, and vinyl chloride) and
nonhalogenated reaction products (ethane and ethylene) to
evaluate a carbon mass balance for the TCE transformation
and to screen for the formation of undesired reaction
intermediates. Aqueous 2.5-mL column influent and effluent
samples were collected in a 5.0-mL gas-tight 1/ 4-28 threaded
nose syringe (Unimetrics). A 2.5-mL headspace was created
in the syringe and then capped and shaken to equilibrate the
aqueous and gas phases. A 200-µL headspace sample was
withdrawn from the syringe and analyzed using GC/ FID as
described in ref 1.
dC
-
TCE ) kobsCTCE
(5)
dt ccat
Column influent and effluent pH measurements were
made using a Ross semi-microcombination pH electrode
and EA940 ion analyzer (Orion Research, MA). The instrument
was calibrated prior to each set of measurements using pH
buffers of 4.01, 7.00, and 10.01. Column influent and effluent
samples were collected in 4-mL glass vials, capped, and stored
no longer than 15 days prior to analysis.
Nitrate and nitrite analysis was performed with a Dionex
series 4000I ion chromatograph (Dionex, CA) equipped with
a Dionex IonPac AS4A column (4 × 250 mm), an AG4A guard
column (4 × 50 mm), and a conductivity detector. A 0.75
mM sodium bicarbonate/ 2.2 mM sodium carbonate solution
was used as an eluent. The instrument was calibrated prior
to analysis using volumetric dilutions of sodium nitrate and
sodium nitrite stock solutions.
where
KnH CnH
2
2
kobs ) krxnKTCE
(6)
n
n
(
)
1 + KH CH
2
2
A kinetic modeling software package (Micromath Scientist
v. 2.01), which minimizes the sum of the squared residuals,
was used to model kinetic data obtained in the batch
experiments. Reported errors are 95% confidence intervals.
In column experiments, substrate conversion through the
column was calculated using
(Cin - Ceff)
% conversion )
× 100
(7)
Dissolved oxygen measurements were made with a
Hansatech DW1 oxygen electrode (Hansatech Instruments
Ltd, Norfolk, England). A 2-mL column influent/ effluent
sample was collected in a 2.5-mL gas-tight syringe and
carefully injected into the electrode cell for analysis. The
Cin
where Cin and Ceff are the measured influent and effluent
substrate concentrations, respectively. Transformation rate
constants were calculated using eq 8 assuming pseudo-first-
9
6 9 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 4, 2001

FIGURE 3. Temporal concentration profile of total chlorinated
intermediates for an [H ](aq) of 1000 and 40 µM (P/Po ) 1 and 0.04).
2
FIGURE 2. Effect of [H ](aq) on TCE transformation rate constant
determined in batch reactor. Solid line represents L-H model fit
using eq 6. The value of n is fixed at 1/2.
2
Cmax occurs at different times.
order reaction kinetics and ideal plug flow as in ref 5:
Ceff
-ln
(
)
Cin
kobs
)
(8)
θ
where kobs is the observed pseudo-first-order TCE reaction
rate constant and θ is the fluid residence time in the column:
nVEB
θ )
(9)
Q
In eq 9, n is the column porosity, VEB is the column empty-
bed volume, and Q is the fluid flow rate through the column.
The column porosity was calculated from gravimetric
measurements. Details of the column porosity calculations
are given in ref 5. It should be noted that θ is calculated for
the entire column rather than just the catalyst bed. The
catalyst bed is approximately 30% of the total bed volume
(Figure 1), so the actual fluid residence time in the catalyst
bed is approximately 30% of the reported reactor residence
time.
FIGURE 4. Maximum amounts of chlorinated intermediates formed
in batch reactors. Total bar height represents sum of chlorinated
intermediates formed, Cmax, as a percentage of the TCE mass
transformed at the time Cmax occurred.
dissociative adsorption of H2 onto active catalyst sites is
expected and n ) 1/ 2 was successfully employed in ref 14
using similar catalysts and hydrogen concentrations during
the aqueous-phase reduction of nitrate solutions. The best
fit was obtained by excluding data obtained at the lowest
hydrogen concentrations as these rate constants are sig-
nificantly lower than predicted based on the L-H model.
This is consistent with catalyst deactivation occurring at H2
concentrations less than 100 µM. Figure 2 also shows that
the reaction order with respect to the [H2](aq) approaches
zero order very slowly, indicating that significant increases
in the [H2](aq) above 1000 µM (P/ Po ) 1) would result in only
moderate increases in the TCE reaction rate. For instance,
doubling the [H2](aq) from 1000 to 2000 µM would increase
kobs by approximately 20%.
In previous batch studies under H2-rich conditions
([H2](aq) ) 1000 µM), transient production of chlorinated
intermediates occurred (DCE isomers and vinyl chloride)
but accounted for less than 3% of the TCE transformed at
their maximum (1). Results reported here support this
conclusion; however, the production of chlorinated inter-
mediate compounds increases with decreasing [H2](aq). The
time at which the total chlorinated intermediate concentra-
tion reached a maximum changed depending on the amount
of H2 available in each experiment (Figure 3). Figure 4 shows
that chlorinated intermediates account for only 2.7% of the
TCE transformed at the maximum under H2-rich conditions
([H2](aq) ) 1000 µM) but can account for as much as 9.8%
of the TCE transformed under H2-poor conditions ([H2](aq)
) 10 µM). The values reported in Figure 4 are the maximum
Results and Discussion
Effect of [H2](aq) on TCE Transform ation Rate. TCE
transformation rates (t < 100 min) were measured in batch
experiments at five different aqueous hydrogen concentra-
tions ranging from 10 to 1000 µM (P/ Po ) 0.01-1). The
observed TCE transformation rate constant and half-lifetime
for each [H2](aq) is given in Table 2. The TCE transformation
rate constant only decreased by only a factor of 2 for [H2](aq)
decreasing from 1000 to 100 µM but decreased sharply for
[H2](aq) less than 100 µM. At an [H2](aq) of 40 and 10 µM,
the TCE transformation rate constant is 10% and 2% of the
value at [H2](aq) ) 1000 µM, respectively. The TCE trans-
formation rate appeared to be slowing in the lowest [H2](aq)
experiment ([H2](aq) ) 10 µM), and in fact, C/ Co decreased
to only 0.8 after 1100 min (data not shown). On the basis of
the observed reaction rate constant for t < 100 min (kobs
)
0.0007 ( 0.0003 min-1), C/ Co should reach 0.8 at some time
between 220 and 560 min. This indicates that the reaction
slows considerably (kobs < 0.0002 min-1) after the first 100
min of reaction. This is likely due to the buildup of
unsaturated hydrocarbons such as ethylene and C4-C6
radical coupling products on the catalyst surface, effectively
blocking catalyst active sites for TCE transformation.
The effect of aqueous hydrogen concentration on the TCE
transformation rate and the best fit of the data using eq 6 is
shown in Figure 2. The value of n was fixed at 1/ 2 because
9
VOL. 35, NO. 4, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 6 9 9

FIGURE 5. Temporal concentration profile of individual DCE isomers
and vinyl chloride for an [H ](aq) of 400 µM(P/Po ) 0.4). Cmax occurs
2
FIGURE 6. Maximum amounts of C4-C6 radical coupling products
(RCPs) formed in batch reactors. Maximum RCP production occurred
at the end of each experiment, so total bar height represents the
sum of RCPs formed as a percentage of the total TCE mass
transformed in each experiment.
at the same time for all chlorinated intermediates.
amount of chlorinated intermediates normalized by the total
TCE transformed at that time such that the results from each
experiment are directly comparable. Although the [H2](aq)
affects the total amount of chlorinated intermediates formed
as a fraction of the TCE transformed, close examination of
Figure 4 reveals that [H2](aq) has little effect on the distribu-
tion of halogenated intermediates formed, with cis-DCE and
1,1-DCE accounting for approximately 80% of the halo-
genated intermediates formed, regardless of the [H2](aq).
Figure 5 shows the concentration profile of all three DCE
isomers and vinyl chloride for an [H2](aq) of 400 µM (P/ Po
) 0.4). It is apparent that all three DCE isomers and vinyl
chloride reach their maximum concentrations simulta-
neously, which is consistent with published results on the
production of chlorinated intermediates using pure Pd metal
as a catalyst (1). Experiments at other [H2](aq) showed
identical behavior (data not shown). The simultaneous
peaking of DCE isomers and vinyl chloride indicate that their
production is not due to sequential dechlorination (TCE f
DCE f vinyl chloride) but rather a random process in which
TCE has one (yielding a DCE isomer) or two (yielding vinyl
chloride) chlorine molecules replaced by hydrogen before
desorbing from the catalyst surface. The dominant trans-
formation pathway however is the removal of all three
chlorine atoms, as chlorinated intermediates account for only
9.8% of the TCE removed, even at the lowest [H2](aq).
Attempts to model production of DCE isomers and vinyl
chloride assuming sequential or parallel first-order produc-
tion and decay were unsuccessful, which is additional
evidence that formation of the DCE isomers and vinyl chloride
is a transient process. Chlorinated intermediates were also
never detected in laboratory-scale column studies (5) or a
pilot-scale field experiment (6) under normal operation,
indicating that production of chlorinated intermediates
during TCE transformation may indeed be a transient
phenomena, as suggested in ref 1. Ethane is the primary
product in the Pd-catalyzed aqueous-phase reduction of TCE,
but C4 and C6 radical coupling products are also observed.
These include 1-butene, n-butane, cis-2-butene, trans-2-
butene, and 2-hexene. Figure 6 shows that production of
C4-C6 radical coupling products increases with decreasing
[H2](aq) in the reactor and accounts for as much as 18% of
the TCE converted at the lowest [H2](aq). The fully saturated
C4 product, n-butane, is the primary radical coupling product
at [H2](aq) greater than 100 µM (P/ Po ) 0.1), but the product
distribution shifts to undersaturated coupling products (1-
butene, cis-2-butene, trans-2-butene, and 2-hexene) at
[H2](aq) less than 100 µM. The undersaturated products are
reactive but less reactive than TCE. The catalyst deactivation
observed at [H2](aq) less than 100 µM may be due to the
presence of these undersaturated products if they sorb
TABLE 3. Influent/Effluent [H ](aq) Concentrations and TCE and
Oxygen Conversion through 2the Column
TCE
O
2
[TCE]in [O ]
(µM)
conv
ktce
conv
koxy
[H ]
(µM)
[H ]
2 out
(µM)
2 in
2 in
(µM) (%) (min-1
)
(%) (min-1
)
32.5
31.0
21.6
29.9
34.1
26.2
46.0 0.138
250 45.8 0.137
370 42.3 0.123
450 13.4 0.032
670
700
790
530
220
220
40
60
490
79
64
67
31
67
0.35
0.23
0.25
0.08
0.25
550
600
4.4 0.010
420
540 44.8 0.133
1150a
a
Influent [H2](aq) increased by increasing total H2 pressure on feed
tank to 2.36 atm .
appreciably to catalyst reactive sites and effectively block
access of TCE or H2 to those sites.
Effect of Com peting Halogenated Organic Com pounds.
The TCE transformation rate was determined in batch
reactors in the presence of competing halogenated organic
substrates under H2-rich conditions ([H2](aq) ) 1000 µM).
TCE (66 µM), cis-DCE (202 µM), trans-DCE (89 µM), and
1,1-DCE (91 µM) were simultaneously present in the reactor.
The pseudo-first-order transformation rate constants for TCE,
1,1-DCE, cis-DCE, and trans-DCE were 0.040 ( 0.006, 0.042
( 0.003, 0.052 ( 0.002, and 0.050 ( 0.004, respectively. The
trends in reactivity (TCE ) 1,1-DCE < cis-DCE ) trans-DCE)
are consistent with previously reported rates (1). The
transformation rate of TCE was identical to that measured
in the absence of competing halogenated substrates at
[H2](aq) ) 1000 µM (Table 2) indicating that cis-DCE, trans-
DCE, and 1,1-DCE do not compete with TCE for catalyst
active sites. All substrates exhibited pseudo-first-order
transformation behavior even though they were present at
relatively high concentrations. This indicates that substrate
coverage was low (K Ci , 1 for all substrates) and may explain
i
the absence of competition for active catalyst sites.
Effect of Dissolved Oxygen as a Com petitive Solute. The
effects of DO in the column influent are summarized in Table
3. DO transforms in the column and utilizes hydrogen
approximately in the expected stoichiometric ratio of 2:1 (eq
1). Without H2 limitations, DO conversion in the column
typically ranges from 64% to 79%, which is greater than the
TCE conversion (46%). The reaction rate constant for DO is
approximately 2-2.5 times higher than that of TCE, indicating
that the undesired reaction between H2 and O2 (eq 1) is faster
than the desired TCE conversion. Despite this, DO in the
9
7 0 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 4, 2001

or reaction of TCE. Additionally, H2 consumption by nitrate
is not a concern because nitrate is not reactive in the system.
Although nitrate transformation requires a bimetallic
catalyst, nitrite has been shown to transform to N2 (eq 3)
using a Pd-on-Al2O3 catalyst (10, 11). The presence of nitrite
in the column influent decreases TCE conversion in the
column, but the effect is minor considering that the nitrite
concentration is over 2 orders of magnitude higher than TCE
at the highest influent nitrite concentration (Table 4). Influent
nitrite concentrations ranging from 1565 (72 mg/ L) to 6630
µM (305 mg/ L) decreased kTCE approximately 10% to 13%,
TABLE 4. TCE and Nitrate/Nitrite Conversion through the
Column
TCE
solute
conv
(%)
[solute]in [TCE]in conv
ktce
ksolute
solute
(µM)
(µM)
(%)
(min-1
)
(min-1
)
none
nitrate
0
371
1290
1565
2609
6630
26.6
30.1
31.0
23.1
30.3
25.1
48.2
48.0
47.4
44.0
43.3
44.4
0.147
0.146
0.143
0.129
0.127
0.131
0
2
0
23
10
0
0.005
0
0.058
0.024
nitrite
from 0.147 (value without nitrite present) to 0.13 min-1
.
Steady-state nitrite conversion (eq 3) through the column
ranged from 10% to 25%, with the lowest nitrite conversion
occurring at highest influent nitrite concentration. Addition-
ally, the column effluent pH increased from the normally
measured value of 8.8 to 10.2 when nitrite was present in the
influent, with no change in the column influent pH. The
increase in effluent pH is consistent with the production of
OH- during the reduction of nitrite to N2 (eq 3) and is further
evidence that nitrite reduction occurs. The lower percent
conversion of nitrite at a higher nitrite concentration suggests
that the nitrite transformation rate is approaching zero order
with respect to nitrite, as might be expected at the high
influent concentrations used. The minimal observed effect
of nitrite on TCE transformation is therefore near the
maximum expected as nitrite transformation is operating at
or near saturation kinetics. The percent nitrite converted
(10-25%) was always less than the percent TCE converted
(43%), indicating that TCE is more reactive than nitrite in the
system (kTCE is approximately 2-5 times higher than knitrite).
Increased reactor effluent pH, measured nitrite loss across
the column, and a move toward zero-order nitrite transfor-
mation kinetics at high nitrite concentrations indicate that
nitrite interacts significantly with the catalyst surface.
However, the minor influence of nitrite on the observed TCE
transformation rate constant (<13%) suggests that nitrite
does not substantially compete with TCE for the same active
catalyst sites. The lower reactivity of nitrite relative to TCE
and the fact that nitrite concentrations in groundwater are
typically very low indicate that nitrite will not lead to H2
depletion in the reactor even though some H2 is utilized for
the reductive transformation of nitrite to N2.
In conclusion, the effect of [H2](aq) on the TCE dehalo-
genation rate and product distribution is substantial, but
complications such as production of chlorinated intermediate
compounds and catalyst deactivation due to adsorbed radical
coupling products can be avoided by maintaining [H2](aq)
at levels achieved by saturation at near ambient pressures
(approximately 1000 µM). The presence of dissolved oxygen
expected in groundwater, even up to the saturation value of
approximately 8.4 mg/ L, will have little influence on the
catalyst’s ability to transform TCE as long as this [H2](aq) is
maintained above approximately 100 µM. Similarly, nitrate
and nitrite at concentrations expected under groundwater
remediation conditions (<100 mg/ L) will have little or no
effect on the catalyst’s ability to transform TCE.
column influent has little or no effect (<10%) on the TCE
transformation rate constant up to values of 370 µM (11.8
mg/ L), which is approximately an order of magnitude greater
than the TCE concentration (∼30 µM). The TCE transforma-
tion rate constant is reduced however at DO concentrations
above 370 µM. The TCE transformation rate constant
decreases by a factor of 4 at a DO concentration of 450 µM
(14.3 mg/ L) and by a factor of 10 at a DO concentration of
600 µM (19.2 mg/ L). Additionally, the DO transformation
rate constant, koxy, decreases by approximately a factor of 3
at the highest influent concentration (600 µM, 19.2 mg/ L).
Although the presence of DO reduced TCE transformation
in the column in some cases, it has no apparent effect on the
product distribution. Ethane and trace amounts of ethene
(<1%) are the only reaction products observed, and carbon
mass balances of greater than 86% were obtained. This is
consistent with previous results in the absence of DO (5).
In cases where the TCE transformation rate constant is
reduced however, effluent [H2](aq) are 60 µM or lower (Table
3). Therefore, the reduced TCE conversion (and oxygen
conversion) may be due to H2 depletion in the system rather
than oxygen out-competing TCE for active catalyst sites as
this is consistent with results from the batch studies where
krxn decreased sharply at [H2](aq) less than 100 µM. This
hypothesis was tested by increasing the column influent
[H2](aq) to 1150 µM but maintaining the DO concentration
at 540 µM (17.3 mg/ L). By providing additional dissolved H2,
the TCE conversion increased to 44.8%, indicating that DO
and TCE do not compete for the same catalyst active sites,
but rather that reduced TCE conversion in the presence of
DO occurs as a result of H2 limitations in the reactor. Batch
data indicated that production of DCE isomers, vinyl chloride,
and C4-C6 radical coupling products increases under H2-
limited conditions, but none of these products were detected
in the reactor effluent.
Effects of Nitrate and Nitrite as Com petitive Solutes.
The effects of moderate to high concentrations of nitrate
and nitrite on TCE conversion in the column are summarized
in Table 4. The presence of nitrate at moderate (371 µM, 23
mg/ L) or high (1290 µM, 80 mg/ L) influent nitrate concen-
trations had no effect on TCE conversion in the column and
conversion of nitrate through the column was essentially
zero. Minor losses of nitrate (<5%) were observed the first
day using the lowest nitrate concentration, but these losses
ceased by the second day and are most likely due to
adsorption of nitrate onto the catalyst support. The non-
reactive behavior of nitrate using Pd-on-Al2O3 catalysts is
consistent with previous studies which reported that bi-
metallic catalysts such as Pd/ Cu or Pd/ Sn were required for
the conversion of nitrate (10, 11, 14). Nitrate and TCE do not
appear to compete for catalyst active sites because high
concentrations of nitrate did not affect TCE conversion even
though significant adsorption of nitrate to the catalyst surface
is expected (14). Also, the presence of nitrate in the system
did not lead to production of detectable chlorinated inter-
mediates (DCE isomers and vinyl chloride) that may have
occurred if adsorbed nitrate interfered with the adsorption
Acknowledgments
Although the research described in this paper has been
funded in part by the U.S. Environmental Protection Agency
(R825421) and the Environmental Security Technology
Certification Program (ESTCP), it has not been subject to the
Agency’s required peer and policy review and therefore does
not necessarily reflect the views of the Agency, and no official
endorsement should be inferred.
Literature Cited
(1) Lowry, G. V.; Reinhard, M. Environ. Sci. Technol. 1999, 33 (11),
1905.
9
VOL. 35, NO. 4, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 7 0 1

(2) Schu¨ th, C.; Reinhard, M. Appl. Catal. B: Environ. 1998, 18, 215.
(3) Schreier, C. G.; Reinhard, M. Chemosphere 1994, 29 (8), 1743.
(4) Siantar, D.; Schreier, C.; Chou, C.; Reinhard, M. Water Res. 1996,
30 (10), 2315.
(5) Lowry, G. V.; Reinhard, M. Environ. Sci. Technol. 2000, 34 (15),
3217.
(11) Daub, K.; Emig, G.; Chollier, M.-J.; Callant, M.; Dittmeyer, R.
Chem. Eng. Sci. 1999, 54 (10), 1577.
(12) McNab, W.; Ruiz, R. Chemosphere 1998, 37 (5), 925.
(13) Atkins, P. W. Physical Chemistry, 3rd ed.; W. H. Freeman and
Company: New York, 1986.
(14) Pintar, A.; Batista, J.; Levec, J.; Kajiuchi, T. App. Catal. B: Environ.
1996, 11 (1), 81.
(6) McNab, W.; Ruiz, R.; Reinhard, M. Environ. Sci. Technol. 2000,
34 (1), 149.
(7) Rylander, P. N. Hydrogenation Methods; Academic Press: New
York, 1985.
(8) Muftikian, R.; Fernando, Q.; Korte, N. Water Res. 1995, 29 (10),
2434.
Received for review August 28, 2000. Revised manuscript
received November 22, 2000. Accepted November 28, 2000.
(9) Fogelberg, J.; Petersson, L. Surf. Sci. 1996, 350 (1-3), 91.
(10) Horold, S.; Vorlop, K.-D.; Tacke, T.; Sell, M. Catal. Today 1993,
17 (1-2), 21.
ES001623F
9
7 0 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 4, 2001