Heterogeneous catalytic reduction of perchlorate in water with Re-Pd/C catalysts derived from an oxorhenium(V) molecular precursor

Article abstract of DOI:10.1021/ic102158a

The molecular Re(V) complex, chlorobis(2-(2′-hydroxyphenyl)-2- oxazoline)-oxorhenium(V), Re(O)(hoz)₂Cl, has been investigated as a suitable precursor, when combined with activated carbon powder containing 5 wt % Pd, to provide a heterogeneous catalyst for the reduction of aqueous perchlorate by hydrogen. Two general methods for catalyst preparation have been adopted: first, by a standard "incipient wetness" impregnation of the carbon powder with handling under largely aerobic conditions for convenience and, second, by a completely anaerobic procedure maintaining a hydrogen atmosphere during adsorption of the complex in water onto the powder. Both types of catalyst were efficient for the complete reduction of perchlorate to chloride within a few hours at room temperature over a range of initial concentrations (2-200 ppm) under 1 atm of H₂ and acidic conditions (pH 2.7-3.7). The perchlorate reduction profiles displayed pseudo-first-order kinetics, and the rates were insensitive to excess chloride. Complete reduction of perchlorate was observed even at pH 5.9 in a phosphate buffer over the course of two weeks. Under comparable conditions, chlorate reduction proceeded ca. 10 times more quickly than perchlorate reduction. The impregnated catalyst was examined by STEM/EDS, which revealed a wide distribution in Pd nanoparticle sizes and also suggested that the Re complex did not aggregate preferentially on or near the Pd particles. XPS of this material provided evidence for reduced Pd after the reaction, but only a +7 oxidation state was seen for the Re sites both pre- and postreduction. Elemental analyses of the catalyst materials taken pre- and postreduction showed variable amounts of Re loss (0-50%) but relatively unchanged amounts of nitrogen. These results show the need to maintain a reducing atmosphere during the preparation and operation of the catalyst in order to achieve optimum activity and stability.

Full text of DOI:10.1021/ic102158a

1534 Inorg. Chem. 2011, 50, 1534–1543
DOI: 10.1021/ic102158a
Heterogeneous Catalytic Reduction of Perchlorate in Water with Re-Pd/C
Catalysts Derived from an Oxorhenium(V) Molecular Precursor
Yunxia Zhang,†,‡ Keith D. Hurley,‡,§ and John R. Shapley*,‡
†Key Laboratory of Pollution Processes and Environmental Criteria, Ministry of Education,
College of Environmental Science and Engineering, Nankai University, Tianjin 300071, P. R. China, and
‡Department of Chemistry and Center of Advanced Materials for the Purification of Water with Systems,
University of Illinois at Urbana-Champaign, Urbana, Illinois 61801, United States.
§ Current address: Department of Chemical and Biochemical Engineering, University of Maine,
5737 Jenness Hall, Orono, Maine 04469
Received October 26, 2010
The molecular Re(V) complex, chlorobis(2-(20-hydroxyphenyl)-2-oxazoline)-oxorhenium(V), Re(O)(hoz)2Cl, has
been investigated as a suitable precursor, when combined with activated carbon powder containing 5 wt % Pd, to
provide a heterogeneous catalyst for the reduction of aqueous perchlorate by hydrogen. Two general methods for
catalyst preparation have been adopted: first, by a standard “incipient wetness” impregnation of the carbon powder
with handling under largely aerobic conditions for convenience and, second, by a completely anaerobic procedure
maintaining a hydrogen atmosphere during adsorption of the complex in water onto the powder. Both types of catalyst
were efficient for the complete reduction of perchlorate to chloride within a few hours at room temperature over a range
of initial concentrations (2-200 ppm) under 1 atm of H2 and acidic conditions (pH 2.7-3.7). The perchlorate reduction
profiles displayed pseudo-first-order kinetics, and the rates were insensitive to excess chloride. Complete reduction of
perchlorate was observed even at pH 5.9 in a phosphate buffer over the course of two weeks. Under comparable
conditions, chlorate reduction proceeded ca. 10 times more quickly than perchlorate reduction. The impregnated
catalyst was examined by STEM/EDS, which revealed a wide distribution in Pd nanoparticle sizes and also suggested
that the Re complex did not aggregate preferentially on or near the Pd particles. XPS of this material provided evidence
for reduced Pd after the reaction, but only a þ7 oxidation state was seen for the Re sites both pre- and postreduction.
Elemental analyses of the catalyst materials taken pre- and postreduction showed variable amounts of Re loss
(0-50%) but relatively unchanged amounts of nitrogen. These results show the need to maintain a reducing
atmosphere during the preparation and operation of the catalyst in order to achieve optimum activity and stability.
Introduction
environmental systems.4-6 The public health concerns sur-
rounding perchlorate contamination are based on its known
interference with the uptake of iodide by the thyroid, which is
especially significant due to the enhanced role played by this
gland in regulating neural development in the early stages of
human life.7-9 The U.S. EPA is currently conducting an
intensive assessment of the environmental risk posed by per-
chlorate in advance of a decision whether to issue a national
regulation for drinking water.8,9 The interim advisory level is
The perchlorate ion, ClO4-, is established as a widespread
contaminant of ground and surface waters, as well as soil,
throughout the United States, with corresponding carryover
into some foodstuffs.1-3 The aqueous chemistry of perchlo-
rate is dominated by its high kinetic barrier to reduction,
which makes it essentially inert to reaction with typical
nucleophilic reducing agents, and by its low basicity, which
leads to weak surface adsorption and ease of transport in
*To whom correspondence should be addressed. E-mail: shapley@
illinois.edu.
(1) Emerging Contaminant;Perchlorate, EPA 505-F-09-005; U.S. EPA:
Washington, DC, Sept 2008.
(5) Perchlorate: Overview of Issues, Status, and Remedial Options; Inter-
state Technology and Regulatory Council: Washington, DC, 2005.
(6) Seyfferth, A. L.; Sturchio, N. C.; Parker, D. R. Environ. Sci. Technol.
2008, 42, 9437–9442.
(2) Brandhuber, P.; Clark, S.; Morley, K. A. AWWA J. 2009, 101(11),
63–73.
(7) Dohan, O.; Portulano, C.; Basquin, C.; Reyna-Neyra, A.; Amzel,
L. M.; Carrasco, N. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 20250–20255.
(8) Interim Drinking Water Health Advisory for Perchlorate, EPA 822-R-
08-025; U.S. EPA: Washington, DC, Dec 2008.
(9) Office of Inspector General Scientific Analysis of Perchlorate,
20100419-10-P0101; U.S. EPA: Washington, DC, Apr 2010.
(3) Trumbo, P. R. Nutr. Rev. 2010, 68, 62–66.
(4) Brown, G. M.; Gu, B. The Chemistry of Perchlorate in the Environ-
ment. In Perchlorate. Environmental Occurrence, Interactions and Treatment;
Gu, B., Coates, J. D., Eds.; Springer Science: New York, 2006; pp 17-47.
r
pubs.acs.org/IC
Published on Web 01/12/2011
2011 American Chemical Society

Article
Inorganic Chemistry, Vol. 50, No. 4, 2011 1535
15 ppm,8 but in the absence of a federal mandate, some states
have set lower legal limits, e.g., California at 6 ppm.10
Effective remediation strategies for perchlorate are limited
currently to selective ion exchange or bioremediation.11,12
The biological approach completely destroys perchlorate, but
doubts remain about public acceptance of its direct applica-
tion to drinking water. Ion exchange effectively separates
perchlorate from drinking water, but postseparation treat-
ment (e.g., incineration) of the loaded resin or of a concen-
trated regeneration brine is necessary for complete destruction.
The chemical or electrochemical reduction of perchlorate
by a heterogeneous catalyst is an alternative approach, but
typical cases are slow and inefficient under ambient con-
ditions.13-16 Nevertheless, in previous work, we showed that
efficient reduction of aqueous perchlorate to chloride can be
achieved with a bimetallic heterogeneous catalyst prepared by
dispersing a Re(VII) complex precursor, e.g., ammonium
perrhenate, onto a commercial 5 wt % Pd/activated carbon
powder.17 Hydrogen provides the reducing equivalents, and
no intermediates or products other than chloride are ob-
served (see eq 1).
an oxygen atom transfer (OAT) cycle.20 In succeeding
investigations, Abu-Omar and co-workers demonstratedthat
the complex chlorobis(2-(20-hydroxyphenyl)-2-oxazoline)-
oxorhenium(V), Re(O)(hoz)2Cl (see II), was an effective
perchlorate reduction catalyst in aqueous acetonitrile, by
using organic sulfides as oxygen acceptors from an inter-
mediate Re(VII) dioxo complex.21-23 Oxazoline ligands are
found in naturally occurring metalloenzymes, and they are
attractive as ligands for metal catalysts in water, since they
are typically less sensitive to hydrolysis than more commonly
used chelating imine ligands.24 We have found that using
compound II as a precursor provides an active and robust
catalyst system for perchlorate (and chlorate) reduction.
Significant aspects of the preparation, characterization, and
use of this new Re-Pd/C heterogeneous catalyst are reported
here.
Experimental Section
Chemicals and Compounds. Catalytic reduction experiments
used 18.2 MΩ cm of water (TOC < 10 ppb) obtained from a
3
Barnstead E-pure four-cartidge 240 V deionization system,
Model D4642-33. Tanks of hydrogen (>99.9%) were supplied
by Linde Gas. All other chemicals and solvents were reagent
grade or better and were supplied by Sigma Aldrich or Fisher
Scientific. The complex ReO(hoz)2Cl, the precursor compound
Re(O)Cl3(OPPh3)(SMe2), and the ligand Hhoz were prepared
according to literature methods.25
-
ClO4 þ 4H2
Cl- þ 4H2O
ð1Þ
f
Re-Pd=C
More recently, we reported that adding a substituted pyridine
ligand with the perrhenate or starting with a preformed Re(V)
complex, e.g., [ReO2(py-X)4]þ (see I), generates a catalyst
with significantly improved activity as well as greater stability
toward pH changes.18
Palladium, as 5 wt % Pd on activated carbon (wet, Degussa-
type E101 NO/W), was a damp powder as received from Sigma
Aldrich. It was first calcined at 110 °C for 1 h under flowing air,
then reduced at 250 °C for 1 h under flowing hydrogen. The
material was allowed to cool to room temperature under the
hydrogen atmosphere, then exposed to air until the highly
exothermic formation of surface oxide was complete. Since this
surface oxide of palladium is easily reduced by hydrogen,26,27
the material was stored in the air until needed. The 5 wt % Pd on
γ-alumina (Sigma Aldrich) and 5 wt % Pd on silica (Strem)
materials were received as dry powders, and no pretreatment
was applied.
Analytical Methods. Measurements of pH were conducted
with a ThermoOrion Model 420 m and a standard Ag/AgCl pH
electrode. Elemental analyses were performed in the Microana-
lysis Laboratory of the School of Chemical Sciences (UIUC);
analysis for metal content was conducted by ICP-MS. Perchlo-
rate, chlorate, chloride, and perrhenate concentrations were
determined by ion chromatography with a Dionex ICS-1000
system (AS 40 autosampler, 25 μL injection loop, 35 mM NaOH
eluent, 1 mL/min flow rate, 4 mm AS-16 analytical and guard
columns, 30 °C). The hydroxide eluent was maintained under an
argon atmosphere to limit carbonate formation.
Our initial work was inspired by seminal observations pub-
lished in 1995 by Abu-Omar and Espenson,19 who showed
that the compound methyltrioxorhenium(VII) in strongly
acidic aqueous solution could catalyze the reduction of
perchlorate with hypophosphorous acid (H3PO2) through
X-ray photoelectron spectra were obtained on a Kratos Axis
ULTRA spectrometer by using monochromatized Al KR radia-
tion with a hemispherical mirror analyzer. For the survey scans,
(10) Perchlorate in Drinking Water, R-16-04; California Department of
Health Services: Sacramento, CA, Oct 2007.
(11) Remediation Technologies for Perchlorate Contamination in Water
and Soil; Interstate Technology and Regulatory Council: Washington, DC, 2008.
(12) Srinivasan, R.; Sorial, G. M. Sep. Purif. Technol. 2009, 69, 7–21.
(13) Lang, G. G.; Horanyi, G. J. Electroanal. Chem. 2003, 552, 197–211.
(14) Cao, J. S.; Elliott, D.; Zhang, W. X. J. Nanopart. Res. 2005, 7(4-5),
499–506.
(15) Wang, D. M.; Shah, S. I.; Chen, J. G.; Huang, C. P. Sep. Purif.
Technol. 2008, 60(1), 14–21.
(16) Mahmudov, R.; Shu, Y.; Rykov, S.; Chen, J.; Huang, C. P. Appl.
Catal., B: Environ. 2008, 81, 78–87.
(17) Hurley, K. D.; Shapley, J. R. Environ. Sci. Technol. 2007, 41, 2044–
2049.
(18) Hurley, K. D.; Zhang, Y.; Shapley, J. R. J. Am. Chem. Soc. 2009, 131,
14172–14173.
(19) Abu-Omar, M. M.; Espenson, J. H. Inorg. Chem. 1995, 34, 6239–
6240.
(20) Abu-Omar, M. M.; Appelman, E. H.; Espenson, J. H. Inorg. Chem.
1996, 35, 7751–7757.
(21) Abu-Omar, M. M.; McPherson, L. D.; Arias, J.; Bereau, V. M.
Angew. Chem., Int. Ed. 2000, 39, 4310–4313.
(22) Arias, J.; Newlands, C. R.; Abu-Omar, M. M. Inorg. Chem. 2001, 40,
2185–2192.
(23) McPherson, L. D.; Drees, M.; Khan, S. I.; Strassner, T.; Abu-Omar,
M. M. Inorg. Chem. 2004, 43, 4036–4050.
(24) Shuter, E.; Hoveyda, H. R.; Karunaratne, V.; Rettig, S. J.; Orvig, C.
Inorg. Chem. 1996, 35, 368–372.
(25) McPherson, L. D.; BeReau, V. M.; Abu-Omar, M. M. Inorg. Synth.
2004, 34, 54–59.
(26) Otto, K.; Hubbard, C. P.; Weber, W. H.; Graham, G. W. Appl.
Catal., B: Environ. 1992, 1, 317–327.
(27) Su, S. C.; Carstens, J. N.; Bell, A. T. J. Catal. 1998, 176, 125–135.

1536 Inorganic Chemistry, Vol. 50, No. 4, 2011
Zhang et al.
a 15 keV accelerating voltage was used with an emission current
of 8 mA and a pass energy of 160 eV. For the high resolution
scans, the pass energy was lowered to 40 eV. The spectra were
referenced to the C 1s binding energy peak at 284.5 eV and were
fit by using the curve-fitting function in CasaXPS.
Dark-field scanning transmission electron microscopy (STEM)
and energy-dispersive X-ray spectroscopy (EDS) were per-
formed on a JEOL 2010F field emission TEM with an accel-
erating voltage of 200 keV and an emission current of 150 μA.
The HREM resolution was 0.24 nm, and the probe size for spot
analysis by EDS was ca. 1 nm.
The BET surface area of the activated carbon powder was
determined on a Quantachrome Nova 2200e analyzer. Palla-
dium metal surface areas were determined by CO pulse chemi-
sorption on a Micromeritics Pulse ChemiSorb 2705 system. The
specific quantities of CO adsorbed (mmol CO/g catalyst) were
0.077 for 5 wt % Pd/C, 0.119 for 5 wt % Pd/alumina, and 0.054
for 5 wt % Pd/silica.
Figure 1. Perchlorate reduction and chloride release profiles at two
different pH values. Solid lines for perchlorate are data fits; dotted lines
for chloride are added for clarity. Conditions: 2 mM HClO4, 0.5 g/L
catalyst, 5.5 wt % Re; pH 3.8 obtained by adding NaOH.
Effect of Hydrogen on pH of Pd/C Powder Suspended in
Water. A set of two experiments were performed with the dry
Pd/C powder in the absence of any added acid, base, or Re com-
plex. Two samples of equal mass were dispersed into two flasks,
each filled with an equal volume of deionized water (loading =
0.5 g/L). One of the flasks was exposed to a flowing H2 atmo-
sphere, whereas the other was exposed to flowing N2. The pH
was measured prior to introducing the gas, again after 1 h of gas
flow, and finally after 5 h of gas flow. The solution in the flask
under nitrogen displayed a pH of 6.4, which was unchanged
after either 1 h or 5 h. In contrast, the solution in the flask under
hydrogen had dropped to a pH of 3.7 after 1 h and remained at
this value at 5 h.
Catalyst Preparation from Perrhenate and Hhoz. A 100 mM
Hhoz stock solution was prepared with HCl at a pH of ca. 3. The
Hhoz solution and a perrhenate solution were mixed to form a
2:1 molar ratio (Hhoz/Re) solution with [Re] = 0.2 mM, and the
pH was readjusted to 3 with HCl. A weighed amount of Pd/C
powder (loading 0.35 g/L) was dispersed into this solution with
stirring, and the flask was purged with hydrogen for 6 h. At this
point, perrhenate was nolonger detectedinsolution(see Figure9).
In the absence of added Hhoz, more than 12 h was required for
complete adsorption of perrhenate. At the end of the adsorption
period, a concentrated, deoxygenated perchlorate solution was
added, and sampling for IC analysis began immediately.
Catalyst Preparation by Incipient Wetness Impregnation. The
Re(O)(hoz)2Cl complex was dissolved in a minimum of dichloro-
methane, and the solution was added in several small portions to
the dry Pd/C powder, taking care not to exceed the pore volume
and wet the surface. Once the incipient wetness point was
reached, the damp material was dried in vacuo at 70-80 °C for
10-15 min. This process was repeated, if necessary, until the
desired Re loading was reached. Finally, the fully loaded
material was dried in vacuo at 70-80 °C for 1 h. The same pro-
cedure was used to prepare catalysts with 5 wt % Pd/γ-alumina
and 5 wt % Pd/silica.
Buffer Systems. A NaH2PO4 H2O (0.01 M)/H3PO4 (x M)
3
buffer system was employed for the pH range of 2.6-4.5. A a pH
of 5.9, buffer was prepared by combining 10 mM NaH2PO4 and
1.1 mM Na2HPO4. The catalyst was stirred in the buffer solu-
tion overnight under flowing hydrogen to establish a stable pH
value before ammonium perchlorate (2 mM) was added.
Results
Perchlorate Reduction by Impregnated Re-Pd/C Cat-
alysts. The material prepared from Re(O)(hoz)2Cl by
standard incipient wetness impregnation of Pd/C powder
showed significant catalytic activity for perchlorate reduc-
tion under conditions analogous to those used in our
previous studies.17,18 As shown in Figure 1, a reaction
conducted with an initial concentration of 2 mM HClO4
(200 ppm ClO4-, pH 2.70) proceeded to completion in a
few hours under a hydrogen atmosphere. Only chloride
was observed as a product, with good overall material
balance. The perchlorate concentration profile was fit
well by a pseudo-first-order kinetic model with k(obs) =
0.76 h-1. As also shown in Figure 1, raising the solution
pH to 3.8 caused the reaction rate to decrease (k(obs) =
0.31 h-1), but the reduction continued nevertheless to
below detection limits (<1 ppm). At pH 2.7, our previous
perrhenate-based system showed a reduction rate ap-
proximately an order of magnitude smaller, and at pH
3.7 the reduction was not only slower in initial rate but
also incomplete overall due to catalyst instability.17 How-
ever, the properties of this catalyst compare well with the
best performance observed for perrhenate catalysts pro-
moted by pyridine ligands.18
Reactor System and Sampling. Catalytic reduction experi-
ments were performed in a batch reactor situated in a water bath
at 23 ( 1 °C on top of a magnetic stirrer. The reactor consisted of
a three-neck round-bottom flask containing 200 mL of aqueous
solution, the powdered catalyst (typically 100 mg), and a Teflon-
coated magnetic stir bar. With continuous stirring of the suspen-
sion, hydrogen was introduced via the submerged tip of a glass
pipet seated in a rubber O-ring sealed adapter in one neck of the
flask. A second neck was sealed with a rubber septum and fit
with a 25 gauge needle to allow for gas outflow. The final neck
was sealed with a glass stopper or a rubber septum. Samples
were withdrawn periodically by syringe, filtered through a 0.2 μm
nylon syringe filter, diluted if necessary, and stored in glass vials
for later analysis by ion chromatography. Measurements of pH
and/or temperature were conducted periodically. The used
catalyst was recovered by filtration in the air and was dried at
ca. 110 °C before subsequent microanalysis.
Catalyst Preparation by in Situ Adsorption of Re(O)(hoz)2Cl.
A weighed amount of Pd/C powder was dispersed into nanopure
water with stirring, and the flask was purged with hydrogen for
ca. 2 h until the pH was steady (vide supra). Then, solid Re(O)-
(hoz)2Cl (an amount sufficient to achieve ca. 7-8 wt % Re in the
final Re-Pd/C catalyst) was added to the flask against the
hydrogen flow, and the mixture was maintained under a hydro-
gen atmosphere for 6 h. Deoxygenated concentrated ammo-
nium perchlorate solution was added at the end of the adsorp-
tion period in order to begin the reaction.
A higher pH for this system was also produced by
starting with a lower concentration of perchloric acid.
Thus, 20 ppm HClO4 provided a reaction pH of 3.5, and

Article
Inorganic Chemistry, Vol. 50, No. 4, 2011 1537
Figure 2. Reduction of 0.2 mM ClO4- (added as HClO4), pH 3.5, 0.5 g/
L catalyst. Solid line for perchlorate is data fit; dotted line for chloride
connects successive points.
Figure 4. Reduction of perchlorate and release of chloride in pH 5.9
phosphate buffer solution. Solid line for perchlorate is data fit; dotted line
for chloride connects successive points.
-
Figure 3. Reduction of 0.02 mM ClO4 (NH4ClO4), pH 2.7 (HCl),
0.5 g/L catalyst. Solid line is data fit for perchlorate.
Figure 5. Correlation of observed reduction rate constant with buffered
reaction pH.
reduction proceeded to completion, as shown in Figure 2,
with k(obs) = 0.15 h-1. At this concentration level, the
excess chloride released from the catalyst material (vide
infra) became visible.
found to be -1.0(0.1). It should be noted that the Re nor-
malized rate at a pH of 2.7 was smaller in the phosphate
buffered system, suggesting that competitive inhibition
by (dihydrogen) phosphate was occurring.
The catalyst was also effective at a much lower con-
centration of perchlorate (∼2 ppm) added as ammonium
perchlorate with HCl added to achieve a pH of 2.7. This
resulted in a 100-fold excess of chloride over perchlorate.
As shown in Figure 3, perchlorate reduction neverthe-
less proceeded to completion in a matter of hours, with
k(obs) = 0.83 h-1. This value is quite comparable with
the reduction rate observed at pH 2.7 with 200 ppm
HClO4 when excess chloride was not present.
Desorption Studies of Impregnated Catalyst. A sample
of impregnated catalyst was suspended in water under
a hydrogen atmosphere (no perchlorate added). Under
these conditions, the pH of the solution was ca. 3.6 due to
acid released from the Pd/C support material under H2
(see the Experimental Section). As shown in Figure 6,
significant amounts of both chloride and perrhenate were
released. For chloride, an initial amount of 0.18 mM rose
slowly to a value of 0.34 mM after 24 h. From the amount
of Re(O)(hoz)2Cl complex delivered with the powdered
support, ca. 0.19 mM would have been expected. There-
fore, it appears that the Pd/C powder contained the
equivalent of ca. 0.15 mM chloride and the remaining
0.19 mM was released from the complex. For perrhenate,
the amount in solution rose rapidly to ca. 0.09 mM and
then slowly declined to zero after 24 h. The maximum
amount of perrhenate released corresponded to almost
half of the Re present on the support from the prepara-
tion.
An attempt was made to reduce 200 ppm perchlorate
(2 mM) at an initial reaction pH of 5.6, created by adding
NaOH to the standard mixture. The reaction proceeded
to completion, but the pH at the end of the reaction was
determined to be 3.6, apparently due to acid produced by
the Pd/C material under hydrogen. Separate experiments
verified that the Pd/C material alone in water produced
acid when exposed to hydrogen (ca. pH 3.7 at 0.5 g/L).
However, a phosphate buffer could be used to maintain
consistent pH values in contact with the Pd/C material
under a hydrogen atmosphere. As shown in Figure 4,
perchlorate could be completely reduced to chloride at a
pH of 5.9 over several days. The observed reduction rate
Perchlorate Reduction with in Situ Catalyst Using Re-
(O)(hoz)2Cl As Precursor. The Re-Pd/C catalyst pre-
pared from Re(O)(hoz)2Cl by adsorption onto Pd/C
powder under the protection of hydrogen showed high
activity for perchlorate reduction. As shown in Figure 7,
reductions run at initial pH values of 2.7 or 3.7 proceeded
constant was 0.28 d-1
.
A plot of log reduction rate versus different (buffered)
reaction pH values is shown in Figure 5. The slope was

1538 Inorganic Chemistry, Vol. 50, No. 4, 2011
Zhang et al.
Figure 8. Perchlorate reduction rate constants observed for different
concentrations of catalyst prepared by adsorption: pH 3 (HCl), 0.2 mM
ClO4-, 8.05 wt % Re.
Figure 6. Perrhenate and chloride desorption from incipient wetness
catalyst (0.5 g/L). The dotted and dashed lines are added to aid the eye.
The solid middle line represents the concentration of Re(O)(hoz)2Cl as
added with the catalyst sample.
Figure 7. In situ catalyst (0.5 g/L) for perchlorate reduction: pH 2.7
(HCl), 0.2 mM ClO4- (NH4ClO4); pH 3.7, 0.1 mM ClO4- (HClO4). Solid
lines are data fits.
to completion in a few hours, and the profiles were well fit
by a pseudo-first-order kinetic model. Little or no per-
rhenate was detected in solution during these reactions, in
contrast to the significant amounts of perrhenate observed
during reactions conducted with the impregnated cata-
lyst. Note that the reduction rates depicted in Figure 7 are
effectively double those shown for the impregnated cat-
alyst in Figure 1.
A set of experiments conducted with different amounts
of catalyst prepared by adsorption directly in the reactor
demonstrated a good linear correlation of observed re-
duction rate with catalyst concentration, as shown in
Figure 8. The slope of this plot can be normalized by
the amount of rhenium in the reactor to give 84.8 h-1
(g Re/L)-1, the value of which is ca. 20 times greater than
that derived from a similar plot reported for the first
generation catalyst with 9.4 wt % Re at a pH of 2.7.17 This
linear plot also suggests that mass transfer of hydrogen to
the surface of the powdered catalyst is not a limiting
factor affecting the reaction rate.
Figure 9. (a)Top:EffectofHhoz onperrhenateadsorption, pH3 (HCl),
0.35 g/L Pd/C; lines shown connect successive points. (b) Bottom:
comparison of perchlorate reduction profiles; solid lines show data fits.
Perchlorate was added following complete precursor adsorption.
displayed a significantly better capability for reducing
perchlorate.
Physical Characterization of the Impregnated Re-Pd/C
Catalyst. Since the impregnated Re-Pd/C catalyst ma-
terial was prepared and handled in the air, no special
precautions were taken before it was examined by several
physical techniques. The total (BET) surface area of the
5 wt % Pd on activated carbon material employed in this
work was found to be 744 m2/g by N2 physisorption. This
result compares well with literature values for other
activated carbon catalysts.28 From CO chemisorption
data, the active metal surface area and dispersion were
-
In Situ Preparation of a Re-Pd/C Catalyst from ReO4
and the Hhoz Ligand. The effects of added Hhoz on the
adsorption of perrhenate under hydrogen as well as on the
activity of the resulting catalysts were briefly examined.
As shown in Figure 9, adding two equivalents of Hhoz
caused a significantly faster and more complete loss of
perrhenate from the solution, and the resulting material
(28) Amorim, C.; Keane, M. A. J. Colloid Interface Sci. 2008, 322(1), 196–
208.

Article
Inorganic Chemistry, Vol. 50, No. 4, 2011 1539
calculated, by assuming a Pd/CO binding ratio of 2:1,29,30
as 7.54 m2/g Pd, corresponding to a dispersion (% surface
Pd) of 28%. Following the addition of the Re(O)(hoz)2Cl
complex to the Pd/C powder, a lowered CO chemisorp-
tion was measured, corresponding to an active palladium
surface area of 3.8 m2/g Pd and a dispersion of 18%.
Elemental analyses were obtained for a variety of catalyst
samples both before and after use for perchlorate reduc-
tion. Postreduction values for catalysts filtered in the air
showed an amount of Re lost that corresponded with ca.
30-50% of that initially present, but the observed wt %
N was relatively constant and close to that expected from
the initial impregnation. Thus, Re was lost to solution
from the catalytic material without a corresponding loss
of N-containing ligand or ligand fragment.
Scanning transmission electron microscopy (STEM)
provided images of the catalyst material that showed
carbon particles with diameters of tens of micrometers
decorated with Pd nanoparticles (see Figure S-1, Support-
ing Information). There was a wide variation in the size of
the metal particles with the average Pd particle size calcul-
ated to be ca. 4 nm, on the basis of the analysis of 111
nanoparticles. This value is consistent with the CO che-
misorption data. The addition of the Re(O)(hoz)2Cl
complex to the carbon material had no apparent effect
on the Pd nanoparticle size distribution by visual exam-
ination of STEM images. Energy dispersive X-ray spec-
troscopy (EDX) spot analysis of a bright palladium
particle showed the expected large Pd peak as well as
small Re peaks. When the X-ray beam was aimed away
from any Pd particles, there was the expected reduction in
Pd peak height, but small Re peaks remained and Cu
appeared from the carbon-coated TEM grid (see Figure
S-2, Supporting Information). These results suggest that
the Re species were distributed rather evenly across the
entire surface of the material and that they did not aggre-
gate preferentially near the Pd particles.
X-ray photoelectron spectroscopy (XPS) data were
obtained on a sample of Re(O)(hoz)2Cl/Pd/C both before
and after its use as a catalyst. Scans of the Pd 3d5/2 region
(see Figure 10) for the as-prepared catalyst were consis-
tent with only one Pd oxidation state, with a binding
energy of 337.3 eV indicating PdO.31 Following exposure
to the reducing environment, but re-exposure to air with
handling, the material exhibited two distinct binding
energies of 335.5 and 337.7 eV, due to Pd(0) and Pd(II)
oxidation states, respectively,31 in the ratio of ca. 1.5:1.
Figure 11 shows high resolution XPS scans over the Re
4f7/2 region. The binding energy for both the pre- and
postreaction sample was 45.6 eV, which is consistent with
a þ7 oxidation state for rhenium.31,32 Since XPS has a
relatively shallow sampling depth, these results indicate
that facile oxidation of the Re(V) complex took place on
Figure 10. High-resolution XPS scans (and fit curves) of Pd 3d5/2 region
for Re(O)Cl(hoz)2/Pd/C catalysts handled in the air: (top) prereduction,
(bottom) postreduction.
the surface of the carbon support as the material was
handled under ambient conditions.
Silica and Alumina as Supports. We briefly examined
catalysts prepared by the addition of a comparable
amount of Re(O)(hoz)2Cl to both 5 wt % Pd/silica and
5 wt % Pd/γ-alumina. These two oxide support materials
do yield catalysts for perchlorate reduction (see Figure
S-3, Supporting Information), but the rate constants are
more than an order of magnitude slower at pH 2.7 than
the value found for the activated carbon system. The
catalyst stability toward a unit increase in pH is signifi-
cantly poorer. The characteristics of the Re-Pd based
catalyst system are strongly tied to the nature of the
support surface, and activated carbon is currently the
support of choice.
(29) Canton, P.; Fagherazzi, G.; Battagliarin, M.; Menegazzo, F.; Pinna,
F.; Pernicone, N. Langmuir 2002, 18(17), 6530–6535.
(30) Fagherazzi, G.; Canton, P.; Riello, P.; Pernicone, N.; Pinna, F.;
Battagliarin, M. Langmuir 2000, 16(10), 4539–4546.
(31) X-Ray Photoelectron Spectroscopy Database. Available on the Internet
from NIST at http://srdata.nist.gov/xps/Default.aspx (accessed June 2008).
(32) Moulder, J.; Bomben, K. D.; Sobol, P. E.; Stickle, W. F. Handbook of
X-Ray Photoelectron Spectroscopy; Physical Electronics, Inc.: Chanhassen,
MN, 1995.
Chlorate Reduction. Chlorate is reduced by metal com-
plexes typically much more rapidly than perchlorate.19,23
Figure 12 provides representative concentration profiles

1540 Inorganic Chemistry, Vol. 50, No. 4, 2011
Zhang et al.
Figure 13. Dependence of chlorate initial reduction rate on catalyst
loading: 5.95% Re; [ClO3-] = 1.2 mM; pH 2.7 (HCl).
The concentration profiles for chlorate were not fit well
by an exponential function, since they typically were
nearly linear for more than half the reaction, with curva-
ture appearing only near the end. Attempts were made to
fit the entire profile with a two-parameter pre-equilibrium
kinetic model (see Discussion); however, the two fitting
parameters were highly correlated and not individually
statistically reliable. Therefore, initial rates, determined
from the first 50% or less of the reaction, were adopted as
a means of comparison among different experiments.
As shown in Figure 13, a plot of the initial chlorate
reduction rate vs catalyst loading is reasonably linear
within the scatter of the data. This confirms that delivery
of hydrogen to the catalysts particles is not a limiting
factor in the reduction of chlorate over the loading range
examined. The slope of this plot provides a normalized
initial rate of 0.868 mM ClO3-/min g cat, which is
3
14.59 mM ClO3-/min g Re or 875 mM ClO3-/h g Re.
3
3
Figure 11. High-resolution XPS scans (and fit curves) of Re 4f7/2 region
for Re(O)Cl(hoz)2/Pd/C catalysts handled in the air: (top) prereduction,
(bottom) postreduction.
From the slope of the similar plot for perchlorate shown
in Figure 8, an initial reduction rate for 100 ppm perchlorate
(1.0 mM/L) would be 85 mM ClO4-/h g Re. Thus, at
3
comparable concentrations of chlorate and perchlorate,
the turnover rate at each active rhenium center is ca. 10
times faster for chlorate than for perchlorate under these
conditions.
Discussion
Characteristics of the Re(O)(hoz)2Cl-Derived Re-Pd/C
System. The process of “heterogenizing” a molecular
complex by interacting it with a solid support necessarily
raises the question of what changes ensue at the molecular
center after the adsorption process is complete. The Pd/C
“support” used in this work is a typical commercial high
surface area activated carbon material, consisting of
micrometer-sized carbon particles decorated with nano-
sized Pd particles. The latter have a distribution of sizes
centered at 4 nm, with CO chemisorption measurements
indicating ca. 28% of the total Pd atoms accessible before
adsorption of ReO(hoz)2Cl and 18% thereafter. Although
the decrease in metal surface area might indicate sub-
stantial selective adsorption onto Pd nanoparticles, this
conclusion is not supported by the EDS data, which show
no preferential linkage between Pd and Re concentrations
in the material. The measured loss of metal surface area
was more likely due to blockage of micropores contain-
ing some Pd nanoparticles by the adsorbed molecular
complex.
Figure 12. Representative profiles for the reduction of chlorate by
Re(O)Cl(hoz)2/Pd/C with different catalyst loadings: 5.95 wt % Re,
1.2 mM ClO3-, pH 2.7 (HCl). Lines shown connect successive points to
aid the eye.
for chlorate reduction with the Re-Pd/C catalyst. The
reaction proceeded to completion in minutes rather than
hours, and it did so with 10% or less of the catalyst
loading used for perchlorate. Under comparable conditions,
the Pd/C powder alone also catalyzed chlorate reduction,
but the rates were 10-20 times slower in the absence
of added rhenium. No evidence for perchlorate reduction
by Pd/C alone was observed, even under more severe
conditions.

Article
The specific ligand set surrounding the Re center dur-
Inorganic Chemistry, Vol. 50, No. 4, 2011 1541
required for maximum activity was found in our previous
report on promotion of the perrhenate-based Re-Pd/C
catalyst by adding N,N-dimethylaminopyridine.18
Mechanism for the ReO(hoz)2Cl-Derived Re-Pd/C
System. The specific properties displayed by this new
Re-Pd/C catalyst are as follows: (i) There is a pseudo-
first-order reduction of perchlorate; separate experiments
showed that chlorate, presumably the first intermediate,
is reduced ca. 10 times faster than perchlorate. (ii) Chlo-
ride is the only observed product, but a 100-fold excess of
chloride does not affect the reduction rate of perchlorate.
(iii) The reduction rate is significantly dependent on pH,
decreasing as the pH rises. These are the same basic
features exhibited by our original Re-Pd/C system using
perrhenate as a precursor17 as well as by our more recently
developed system in which perrhenate is enhanced by sub-
stituted pyridine ligands.18
ing and after catalysis is subject to some uncertainty. The
release of an equivalent of chloride to solution was clearly
observed for the impregnated material (see Figure 6). This
result is consistent with the facile formation of the cat-
ionic complex [ReO(hoz)2(S)]þ (S = H2O, CH3CN) in
aqueous acetonitrile solution that has been described by
Abu-Omar and co-workers.21,22 For the complex ad-
sorbed onto Pd/C, the chloride release may be promoted
by coordination of a surface functional group (e.g.,
hydroxide, phenoxide, or carboxylate) to the Re center.
Pre- and postreduction elemental analyses of the catalytic
material showed relatively constant %N values, suggest-
ing that free Hhoz was not released to solution. In con-
trast, the %Re values of catalyst material recovered by
filtration frequently showed significant decreases, consis-
tent with the significant amounts of perrhenate observed
in solution for the air-handled catalysts.
The dominant presence of one or more compounds
with rhenium in a Re(VII) oxidation state on the surface
of the impregnated material was revealed by XPS anal-
ysis. However, these data do not distinguish among
various possible Re(VII) compounds that might be pre-
sent. Of foremost interest is the dioxo Re(VII) species
[ReO2(hoz)2]þ that has been observed in solution as an
intermediate stage in OAT reactions.21-23 This type of
species might well exhibit a longer lifetime and higher
stability when dispersed on the support surface in the
absence of easily oxidizable acceptors. Recently, the
direct formation of oxorhenium(VII) complexes from
rhenium(V) complexes reacting with dioxygen has been
reported.33 Second, partial hydrolysis of [ReO2(hoz)2]þ
has been reported to form the mono-hoz complex ReO3-
(hoz).21 This Re(VII) complex would very likely form
ReO4- upon further contact with water, especially in the
presence of acid, so that it might represent an intermedi-
ate stage in the formation of the perrhenate ion actually
observed to be released into solution from the air-handled
catalyst.
Reductive adsorption of perrhenate onto Pd/C powder
under H2 is relatively slow,34 and we observed that adding
two equivalents of Hhoz not only enhanced the rate and
extent of perrhenate adsorption but also significantly
improved the catalytic activity of the resulting material
for perchlorate reduction (see Figure 9). However, the
observed catalytic activity was still markedly less than
that achieved by starting with the preformed ReO-
(hoz)2Cl complex. Overall, maximal activity was ob-
tained when the catalyst preparation was conducted by
in situ adsorption of the Re(O)(hoz)2Cl precursor under
the protection of hydrogen, in which case no perrhenate
was detected in solution during the reduction reaction.
This result presumably indicates that for the Re center
two hoz ligands are better than one, although it does not
establish that the hoz ligands are bound to Re in exactly
the same manner (chelating, O-deprotonated) as in the pre-
cursor ReO(hoz)2Cl complex.21-23 A strong indica-
tion that two N-donor ligands per rhenium center are
These properties are embodied in the following set of
eqs 2-4 that describe the catalytic cycle for the reduction
of perchlorate (x = 4) or chlorate (x = 3).
k1
-
ReV þ ClOx þ Hþ s ReV- O- ClOx - 1- H ðIÞ
ð2Þ
ð3Þ
ð4Þ
F
s
R
k - 1
k2
s
I ðslowÞ
ReVII¼O þ ClOx - 1- þ Hþ
f
k3
f
ReVII¼O þ 2HðPdÞ ðfastÞ
s
ReV þ H O
2
Equation 4 indicates that the role of Pd is to activate H2
and thereby provide the two hydrogen atom equivalents
needed to reduce Re(VII)dO to Re(V)(OH2) for a com-
plete catalytic cycle. Hydrogen spillover is a well-docu-
mented phenomenon for supported metal particles,35-40
and it has been observed directly at room temperature for
Pd/C.39,40 In the present case, we observed Hþ production
when the aqueous suspension of Pd/C was exposed to H2.
This is very likely due to hydrogen spillover that results in
partial hydrogenation of oxygenated functional groups
on the activated carbon41 (e.g., aldehyde/ketone to alcohol
or quinone to hydroquinone) to produce more acidic
groups. Activation of dihydrogen on a Pd surface is
known to be very fast at room temperature,42,43 and
limiting the rates of hydrogen transfer reactions on Pd
nanoparticle sites requires very low H2 partial pressures.44
We suggest that overall hydrogen transport (Pd activa-
tion þ spillover) is not rate limiting in our system for
perchlorate (or chlorate) reduction. Varying the loading
of the catalyst in the reactor gave a linear response in
observed reduction rate. These data expressed in terms of
(35) Conner, W. C.; Falconer, J. L. Chem. Rev. 1995, 95, 759–788.
(36) Pajonk, G. M. Appl. Catal., A 2000, 202, 157–169.
(37) Wang, L.; Yang, R. T. J. Am. Chem. Soc. 2009, 131, 4224–4226.
(38) Psofogiannakis, G. M.; Froudakis, G. E. J. Am. Chem. Soc. 2009,
131, 15133–15135.
(39) Amorim, C.; Keane, M. A.. J. Coll. Interface Sci. 2008, 322, 196–208.
(40) Contescu, C. I.; Brown, C. M.; Liu, Y.; Bhat, V. V.; Gallego, N. C. J.
Phys. Chem. C 2009, 113, 5886–5890.
(41) Amorim, C.; Keane, M. A. J. Chem. Technol. Biotechnol. 2008, 83,
662–672.
(42) Conrad, H.; Ertl, G.; Latta, E. E. Surf. Sci. 1974, 41, 435–446.
(43) Shin, E. W.; Cho, S. I.; Kang, J. H.; Kim, W. J.; Park, J. D.; Moon,
S. H. Korean J. Chem Eng. 2000, 17, 468–472.
(44) Lowry, G. V.; Reinhard, M. Environ. Sci. Technol. 2001, 35(4), 696–
702.
(33) Lippert, C. A.; Arnstein, S. A.; Sherrill, C. D.; Soper, J. D. J. Am.
Chem. Soc. 2010, 132, 3879–3892.
(34) Choe, J. K.; Shapley, J. R.; Strathmann, T. J.; Werth, C. J. Environ.
Sci. Technol. 2010, 44, 4716–4721.

1542 Inorganic Chemistry, Vol. 50, No. 4, 2011
Zhang et al.
the concentration of Re in the reactor are shown in
Figure 10 (and Figure 13). The linear plot and its inter-
section at the origin strongly suggest that the observed
rates are directly determined by the availability and
activity of the Re sites in the catalyst and that the hydro-
gen reducing equivalents originating at Pd are provided in
a continuous flux completely sufficient to meet the needs
of the perchlorate (or chlorate) reduction reaction. We
note reports of oxo-rhenium compounds activating hy-
drogen directly in reactions with epoxides45 and alkynes;46
however, the conditions used (80-150 °C, 5-40 atm)
were much more severe than those needed for perchlorate
reduction, which further supports the idea of separate
roles for Pd and Re in the Re-Pd/C catalyst system.
Equation 3 depicts the rate-determining step as the
OAT reaction occurring within the (protonated) perchlorate-
Re(V) complex (I) to generate an oxo-rhenium(VII)
species with release of chlorate and a proton. Since this
step involves oxidation of the rhenium center, its rate
should depend on the electron-releasing capability of ligands
attached to rhenium. Our results with substituted pyridine
ligands showed a direct correlation between the donor ability
of the ligand and the rate of perchlorate reduction.18
Table 1. Composite Rate Parameters for Perchlorate Reduction with Various Re
Catalysts
k2K0
(L/M s)
aqueous
medium
catalyst
ReO2CH3
ReO(hoz)2Cl
ref
7.3
1.0 M CF3SO3H
90% CH3CN
2.0 mM HCl
1.0 mM HCl
19
21
17
18
0.45
0.21
1.6
[NH4][ReO4] þ Pd/C
[NH4][ReO4] þ
2 DMAP þ Pd/C
ReO(hoz)2Cl þ Pd/C
4.4
1.0 mM HCl
This work
rate expression emerges (eq 8).
-
d=dt½ClOx ꢀ ¼ k2ðReÞT
ð8Þ
The kinetic behavior observed for chlorate reduction
appears to approach the second case, in that an initially
linear concentration profile was typical (see Figure 12).
However, the onset of curvature late in the reaction
suggested that the strict condition for case 2 was not true
throughout the reaction. We examined using the more
complete eq 6 for numerical fits to the entire reduction
profiles. Unfortunately, the two fitting parameters were
found to be highly correlated. A value of ca. 3.4 L/mM for
K0 allowed a reasonable fit in nearly all cases, but the
values for “best” fit (minimizing RMS deviations) in each
case varied from ca. 1 to 10 L/mM, with a corresponding
large variation in values of k2. Nevertheless, with a value
of K0 in this range, the denominator in eq 6 will have
Finally, eq 2 explicitly includes the role of a proton,
which is to promote a secondary hydrogen bonding inter-
action between one of the perchlorate oxygens and the
surface in order facilitate perchlorate coordination. This
interaction is not available for chloride, which therefore
does not compete strongly with perchlorate for coordina-
tion to rhenium and does not act to inhibit the OAT
reaction, at least at the levels we have examined.
-
values of [ClO3 ] greater than 1/K0 through at least
the early part of the reaction. Thus, the observed initial
rate for chlorate reduction will approximate k2 (Re)T, and
an estimate of k2 is provided from the slope of a plot of
initial rates against total Re as shown in Figure 13.
Kinetics Model. Applying the steady-state assumption
for the intermediate complex (I) shown in eqs 2 and 3
leads to the following rate expression (eq 5):
Our observed perchlorate reduction profiles were pseu-
do-first-order, which conforms to the first case. This is
understandable, since K0 is expected to be very small for
this weakly binding oxyanion, so [ClO4-] will always be
,1/K0. As shown in eq 7, the observed pseudo-first-order
rate parameter is the product of k2, the intrinsic rate for
the OAT reaction; K0, the “complexation constant” for
coordination of perchlorate to the Re center under the
reaction conditions; and (Re)T, the total amount of Re
available in the reaction system. A plot of kobs vs catalyst
loading is indeed linear, as shown in Figure 8, and the
slope provides, after normalizing for Re content in the
catalyst, a value of the product k2K0 = 4.4 s-1 L/M Re.
Furthermore, since K0 as defined is proportional to [Hþ],
kobs is predicted to decrease significantly as the pH
increases. This is qualitatively consistent with the ob-
served effect of a unit increase in pH on the reduction rate,
as shown in Figures 1 and 7. In an attempt to overcome
the buffering effect of the Pd/C catalyst material in water
when exposed to H2, we used a phosphate buffer to access
a larger range of pH values under relatively consistent
conditions. As shown by the data plotted in Figure 5, the
dependence of log kobs on pH under these conditions has
the predicted slope of -1, although confidence in the
precise value of this result must be tempered by the
-
k2ðReÞT½ClOx
ꢀ
-
Rate ¼ d=dt½ClOx ꢀ ¼
ð5Þ
k - 1 þ k2
þ ½ClOx
k1½Hþꢀ
-
ꢀ
where (Re)T = [ReV] þ [ReV-O-ClOx-1-H] and
[ReVII=O] is assumed negligible at the steady state. Since
we assume k2 , k-1, the former parameter can be
neglected in the denominator, anþd if we define a pre-equi-
librium “constant” K0 = k1[H ]/k-1, the result is ex-
pressed in eq 6:
-
k2ðReÞT½ClOx
1=K0 þ ½ClOx
ꢀ
-
d=dt½ClOx ꢀ ¼
ð6Þ
-
ꢀ
From this rate expression, two limiting cases can be
considered. In the first, 1/K0 . [ClOx-], and the rate
expression collapses to a pseudo-first-order dependence
on [ClOx-] (eq 7):
-
-
d=dt½ClOx ꢀ ¼ k2K0ðReÞT½ClOx
ꢀ
-
¼ kobs½ClOx
ꢀ
ð7Þ
where the observed rate constant kobs = k2K0(Re)T. In the
second case, [ClOx ] . 1/K0, and a pseudo-zero-order
-
-
-
expectation of competition by H2PO4 with ClO4 for
binding to the Re centers.
As is apparent from examination of eq 7, the product is
effectively a second-order rate coefficient for the direct
(45) Ziegler, J. E.; Zdilla, M. J.; Evans, A. J.; Abu-Omar, M. M. Inorg.
Chem. 2009, 48, 9998–10000.
(46) Reis, P. M.; Costa, P. J.; Romao, C. C.; Fernandes, J. A.; Calhorda,
M. J.; Royo, B. Dalton Trans. 2008, 1727–1733.

Article
Inorganic Chemistry, Vol. 50, No. 4, 2011 1543
interaction of a Re center with the oxyanion substrate.
This composite parameter provides a measure of the
catalytic efficiency for OAT from perchlorate to the Re
centers in various catalysts. A similar pre-equilibrium
kinetics model was developed by Espenson and Abu-Omar
in the previous studies of the homogeneous MDO19 and
Re(O)(hoz)2Cl21 systems reacting with perchlorate. Va-
lues of the “second-order” rate parameter for these sys-
tems are compared with analogous values derived for the
heterogeneous Pd/C-supported systems in Table 1. It
should be noted that the solution media were quite varied,
ranging from aqueous triflic acid at pH 1 for MDO to
10% H2O/90% NCCH3 with no added acid for the
homogeneous ReO(hoz)2Cl system. Nevertheless, it is
interesting that the “heterogenized” ReO(hoz)2Cl cata-
lyst reported here is an order of magnitude more
efficient for perchlorate reduction than the homoge-
neous system, under the specific conditions used in each
case.
that complete reduction of perchlorate to chloride could be
achieved after a reaction period of ca. 15 days.
The most likely use of a Re-Pd/C catalyst for perchlorate
remediation would be in a two-stage system involving direct
removal from drinking water by ion exchange followed by
batch purification of the brine solution used for regeneration
of the ion exchange column.17,34,47 Any needed adjustment of
pH in the regenerant brine for optimal activity in perchlorate
reduction would not be prohibitive for this relatively modest
volume of water. The significantly higher normalized activity
for the catalyst derived from ReO(hoz)2Cl compared with
that from ReO4- would allow much shorter treatment times
or much lower catalyst loadings. However, in both cases,
-
leaching of ReO4 from the catalyst into solution due to
oxygen exposure serves to limit the catalyst activity, to reduce
its capacity for reuse, and also to release a new contaminant
into solution. Thus, careful and continuous maintenance of a
reducing atmosphere for both the preparation and the use of
a ReO(hoz)2Cl-derived catalyst would be necessary in order
to take advantage of its superior properties.
Conclusion and Prospects
Acknowledgment. This work was supported by the
National Science Foundation Division of Chemistry, Grant
Number CHE 07-18078; by WaterCAMPWS, a Science
and Technology Center of Advanced Materials for the
Purification of Water with Systems, under NSF Agree-
ment Number CTS-0120978; and by NSF Grant CBET
0730050. Y.Z. acknowledges the China Scholarship
Council (CSC, No. 2008620016) for partial financial
support at UIUC. We thank Dr. Rick Haasch (XPS) and
Dr. Changui Lei (STEM, EDS) for assistance with instru-
mentation in the Frederick Seitz Materials Research Lab-
oratory Central Facilities, University of Illinois, which are
partially supported by the U.S. DOE grants DE-FG02-
07ER46453 and DE-FG02-07ER46471. Prof. K. Suslick
and Prof. R. Masel kindly provided access to instrumenta-
tion in their laboratories for surface area and CO chemi-
sorption measurements. We also wish to thank J. Daleiden
forhisworkonthesynthesisoftheRe(O)(hoz)2Cl complex.
TheReO(hoz)2Cl-Pd/C catalyst, when prepared and main-
tained under a hydrogen atmosphere, exhibited a Re nor-
malized composite rate constant, k2K0, that was over 20 times
faster than that observed for the perrhenate-based catalyst.17
The activity improvement of the new system toward per-
chlorate reduction is clearly due to the presence of the hoz
ligands around the Re center. This ligand promotion effect
has also been demonstrated with a series of substituted
pyridine ligands, most dramatically with N-dimethylamino-
pyridine (DMAP).18 The effect on rate enhancement is
attributed to promotion of the rate-determining OAT step
in the catalytic cycle, since this step results in a decrease in
electron density at the Re center as the rhenium to oxygen
double bond is formed (Re(V) oxidized to Re(VII)dO), and
the donating ligand interaction with the Re center stabilizes
the developing positive character. Furthermore, in both
cases, the ligand-modified Re reaction site is stabilized toward
increases in pH relative to the perrhenate-only system. In
particular, we showed here that the ReO(hoz)2Cl-derived
catalyst is sufficiently stable in a pH 5.9 phosphate buffer
Supporting Information Available: STEM images of 5 wt %
Pd on activated carbon powder, histogram of Pd nanoparticle
size, EDS spot analyses of Re(O)(hoz)2Cl-Pd/C catalyst, and
perchlorate reduction profiles for catalysts with 5 wt % Pd on
silica and γ-alumina. This material is available free of charge via
the Internet at http://pubs.acs.org.
(47) Gu, B.; Brown, G. M. Recent Advances in Ion Exchange for
Perchlorate Treatment, Recovery and Destruction. In Perchlorate. Environ-
mental Occurrence, Interactions and Treatment; Gu, B., Coates, J. D., Eds.;
Springer Science: New York, 2006; pp 209-251.