Glycerol hydrogenolysis on carbon-supported PtRu and AuRu bimetallic catalysts

Article abstract of DOI:10.1016/j.jcat.2007.08.007

Bimetallic PtRu and AuRu catalysts were prepared by a surface redox method in which Pt or Au was deposited onto the surface of carbon-supported Ru nanoparticles with an average diameter of 2-3 nm. Characterization by H₂ chemisorption, analytical TEM, and X-ray absorption spectroscopy at the Ru K-edge, Pt L_III-edge, and Au L_III-edge confirmed that Pt and Au were successfully deposited onto Ru without disrupting the Ru particles. Depression of the ethane hydrogenolysis rate over Ru after addition of Au provided further evidence of successful deposition. The bimetallic particles were subsequently evaluated in the aqueous-phase hydrogenolysis of glycerol at 473 K and 40 bar H₂ at neutral and elevated pH. Although monometallic Pt and Ru exhibited different activities and selectivities to products, the bimetallic PtRu catalyst functioned more like Ru. A similar result was obtained for the AuRu bimetallic catalyst. The PtRu catalyst appeared to be stable under the aqueous-phase reaction conditions, whereas the AuRu catalyst was altered by the harsh conditions. Gold appeared to migrate off the Ru and agglomerate on the carbon during the reaction in liquid water.

Full text of DOI:10.1016/j.jcat.2007.08.007

Journal of Catalysis 251 (2007) 281–294
www.elsevier.com/locate/jcat
Glycerol hydrogenolysis on carbon-supported PtRu and AuRu
bimetallic catalysts
a
a
b
a,∗
Erin P. Maris , William C. Ketchie , Mitsuhiro Murayama , Robert J. Davis
a
Department of Chemical Engineering, University of Virginia, 102 Engineers Way, Charlottesville, VA 22904-4741, USA
b
Department of Materials Science and Engineering, University of Virginia, 116 Engineers Way, Charlottesville, VA 22904-4745, USA
Received 9 May 2007; revised 19 July 2007; accepted 11 August 2007
Abstract
Bimetallic PtRu and AuRu catalysts were prepared by a surface redox method in which Pt or Au was deposited onto the surface of carbon-
supported Ru nanoparticles with an average diameter of 2–3 nm. Characterization by H chemisorption, analytical TEM, and X-ray absorption
2
spectroscopy at the Ru K-edge, Pt L -edge, and Au L -edge confirmed that Pt and Au were successfully deposited onto Ru without disrupting
III
III
the Ru particles. Depression of the ethane hydrogenolysis rate over Ru after addition of Au provided further evidence of successful deposition.
The bimetallic particles were subsequently evaluated in the aqueous-phase hydrogenolysis of glycerol at 473 K and 40 bar H at neutral and
2
elevated pH. Although monometallic Pt and Ru exhibited different activities and selectivities to products, the bimetallic PtRu catalyst functioned
more like Ru. A similar result was obtained for the AuRu bimetallic catalyst. The PtRu catalyst appeared to be stable under the aqueous-phase
reaction conditions, whereas the AuRu catalyst was altered by the harsh conditions. Gold appeared to migrate off the Ru and agglomerate on the
carbon during the reaction in liquid water.
©
2007 Elsevier Inc. All rights reserved.
Keywords: Glycerol; Hydrogenolysis; Bimetallic catalysts; X-ray absorption spectroscopy; Transmission electron microscopy
1
. Introduction
molecular weight glycols and acids [9–11]. Recent work in
our laboratory has shown that Ru/C and Pt/C are catalysts for
the hydrogenolysis of glycerol, where the selectivities to eth-
ylene glycol and propylene glycol are a function of metal type
and solution pH [12]. Other studies have shown that addition
of a modifier, such as sulfur, can alter the activity and selec-
tivity of glycerol hydrogenolysis catalysts [13,14]. Montassier
et al. demonstrated that increasing coverage of sulfur on a
Ru/C catalyst decreased the activity for glycerol hydrogenol-
ysis under neutral pH but increased selectivity to propylene
glycol [13]. These authors also observed a maximum in the
ethylene glycol selectivity at a sulfur coverage of 25%. Lahr
and Shanks [14] also investigated the effects of sulfur addi-
tion on glycerol hydrogenolysis over Ru/C at high pH, and
found that higher sulfur loadings increased the selectivity to
propylene glycol without altering the selectivity to ethylene gly-
col.
The primary feedstocks of renewable organic fuels, chemi-
cals, and materials are and will be derived from biomass [1–4].
In particular, glycerol has been identified as a promising al-
ternative to petroleum and natural gas for the production of
commodity chemicals and materials [5]. Because glycerol is
also the major byproduct of biodiesel production by transes-
terification of vegetable oil [6], using the growing supply of
glycerol is a logical step in moving toward a more sustainable
economy. One industrially relevant route for the conversion of
glycerol to oxygenated chemicals involves hydrogenolysis to
ethylene glycol, propylene glycol, and lactic acid, as shown in
Scheme 1 [7,8].
Previous studies have demonstrated the effectiveness of het-
erogeneous catalysts in the hydrogenolysis of polyols to lower
Hydrogenolysis reactions are well known to be structure
sensitive; that is, the activity and selectivity depend on cata-
lyst surface structure or particle size. However, it is not clear
*
Corresponding author. Fax: +1 (434) 982 2658.
E-mail address: rjd4f@virginia.edu (R.J. Davis).
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2007.08.007

282
E.P. Maris et al. / Journal of Catalysis 251 (2007) 281–294
Scheme 1. Proposed mechanism of glycerol hydrogenolysis. Adapted from Ref. [12].
whether the role of sulfur in the study of glycerol hydrogenoly-
sis is simple site-blockage of the Ru surface atoms that form an
active ensemble or some other modification. Adsorbed sulfur
has been suggested to electronically modify the Ru surface [13]
and possibly inhibit a low-energy catalytic pathway to propy-
lene glycol in favor of a higher energy pathway [14]; thus, the
role of sulfur may involve both ensemble effects and electronic
fore, evaluation of Ru–Au bimetallic catalysts for glycerol hy-
drogenolysis is a logical test for ensemble effects.
In this work, we compare carbon-supported Pt–Ru and Au–
Ru bimetallic catalysts to their monometallic analogues for
the hydrogenolysis of glycerol under neutral conditions and at
high pH. The catalysts were characterized by H2 chemisorp-
tion, transmission electron microscopy, in situ X-ray absorption
spectroscopy, and ethane hydrogenolysis so that catalyst struc-
ture could be related directly to activity and selectivity.
(
or ligand) effects.
Bimetallic catalysts have been used to alter the activity
and selectivity of various reactions, including hydrogenolysis
reactions [15–19]. Moreover, Pt–Ru bimetallic catalysts are
widely studied for potential use as anodes in fuel cell appli-
cations [19,20]. Although Ru and Pt are partially miscible in
the bulk [21], they need not form an alloy to display synergis-
tic catalytic activity [20]. As mentioned earlier, previous work
in our lab has demonstrated clear differences between carbon-
supported Ru and Pt catalysts for glycerol hydrogenolysis [12].
Because bimetallic catalysts can exhibit catalytic activities that
surpass those of their monometallic analogues [18], we consid-
ered the evaluation of Pt–Ru bimetallic particles as a reasonable
next step. Addition of Au has also been shown to alter the activ-
ity of Ru for ethane and propane hydrogenolysis [22]. Because
Ru and Au are not miscible in the bulk [21], particular attention
must be paid to catalyst synthesis and characterization of this
bimetallic system. In previous studies of alkane hydrogenoly-
sis on bimetallic Au–Ru catalysts, the depression of catalytic
activity of Ru by addition of Au was considered primarily an
ensemble effect instead of an electronic effect [22–24]; there-
2. Experimental
2.1. Catalyst preparation
Activated carbon-supported Ru and Pt monometallic cata-
lysts were obtained from Acros Organics and Sigma–Aldrich,
respectively. The Ru/C (5 wt% Ru, 50% w/w water) catalyst
was dried in air at 493 K overnight before use; Pt/C (3 wt%
Pt) was used as received. Gold was supported on activated car-
bon by cation exchange. The activated carbon was first oxidized
with aqueous sodium hypochlorite (NaClO) solution according
to the method reported by Gallezot et al. [25]. The [Au(en) ]Cl
2
3
(en = ethylenediamine) needed as a precursor for ion exchange
was prepared according to the method of Block and Bailar [26].
3
−1
To prepare the Au/C, 20 cm g of the oxidized carbon support
was slurried with 1 M ammonium hydroxide (Alfa Aesar, 50%
v/v), while bubbling N through the mixture at ambient condi-
2
tions. The appropriate amount of [Au(en)2]Cl3 was dissolved in

E.P. Maris et al. / Journal of Catalysis 251 (2007) 281–294
283
1
M ammonium hydroxide to form a 0.01 M solution and this
lysts to compare the amount of hydrogen adsorbed per gram of
catalyst with that on the parent Ru/C catalyst.
solution was added dropwise to the carbon slurry. The resulting
suspension was stirred at ambient conditions in the presence of
bubbling N for 24 h. The suspension was filtered and washed
with distilled deionized water until no Cl could be detected
2.3. Transmission electron microscopy
2
−
with a 0.1 M AgNO3 test. The catalyst was dried at ambient
conditions overnight, followed by drying in air at 383 K for 2 h.
The Ru/C catalyst also was used as a precursor for the
preparation of bimetallic Pt–Ru and Au–Ru catalysts by sur-
face redox reactions [27–29]. For the preparation of the PtRu/C
bimetallic catalyst, the parent Ru/C catalyst was slurried with
The catalyst particle size and morphology were examined
by bright-field and dark-field transmission electron microscopy
(TEM) using a JEOL 2010F transmission electron microscope
operating at 200 kV and equipped with a Gatan imaging fil-
ter (GIF). Samples were prepared by suspending the catalyst in
ethanol and agitating in an ultrasonic bath for 2 h. The suspen-
sion was allowed to sit overnight so that larger particles could
settle out. Two drops of the suspended catalyst were applied to a
copper mesh grid with lacy carbon film, with the ethanol evapo-
rating in between drops. Images were recorded on film and on a
slow-scan CCD camera that permitted analysis with the Gatan
Digital Micrograph software package. Energy dispersive X-ray
spectroscopy (EDS) was carried out using an Oxford Instru-
ments model 6498 detector and the NIST Desktop Spectrum
Analyzer software [30]. For analysis of individual particles,
the electron beam was focused to a diameter that was slightly
less than the diameter of the particle being investigated. Metal
particle size distributions were determined using ImageJ soft-
ware [31].
3
−1
2
0 cm g 0.05 M HCl (Acros Organics) in a sealed glass re-
actor equipped with a magnetic stirrer. To remove air in the
system, N2 (GT&S, 99.998%) was bubbled through the slurry
at room temperature for 1 h. To remove N2 and reduce the Ru
surface, H2 (GT&S, 99.995%) was bubbled through the slurry
for 2 h at 353 K. The appropriate amount of H2PtCl6·xH2O
3
−1
(
Acros Organics, 38.75 wt% Pt) was dissolved in 20 cm g
of 0.05 M HCl and degassed by flowing N2 through the so-
lution for 1 h. The Pt precursor solution was added dropwise
through a gas-tight syringe to the Ru/C slurry. Once the Pt pre-
cursor was added, the slurry was stirred for 1 h under bubbling
H at 353 K, then cooled to room temperature under H , after
2
2
which the reactor was purged with N2 before being exposed to
air. The catalyst was filtered and washed with 6 L of distilled
−
2.4. X-ray absorption spectroscopy
deionized water to remove residual Cl . The final filtrate tested
−
negative for Cl using a 0.1 M AgNO3 solution. The catalyst
The X-ray absorption spectroscopy (XAS) was performed at
the National Synchrotron Light Source, Brookhaven National
Laboratory. The storage ring operated at 2.8 GeV with a ring
current of 150–300 mA. Data were collected on beamline X-
18B operating in transmission mode at the Ru K-edge (22.118
keV) with a beam spot size of 0.5 mm × 10 mm, the Pt LIII-
edge (11.564 keV), and the Au LIII-edge (11.919 keV) with a
spot size of 1 mm × 10 mm.
was allowed to dry in air at 393 K overnight. A bimetallic Au–
Ru catalyst also was prepared in this manner, where HAuCl4
(
Aldrich) was used as the Au precursor. The catalysts are des-
ignated PtRu/C and AuRu/C in this work.
After preparation, the catalysts were heated at 2 K min in
H2 (GT&S, 99.999%, purified using a Matheson 8371V palla-
dium hydrogen purifier) flowing at 100 cm min to 523 K and
held at that temperature for 4 h. The reduced catalysts were al-
lowed to cool under H2. Once room temperature was reached,
a stream of He was passed over the catalyst bed, followed by
slow exposure to air. The Ru, Au, and Pt weight loadings were
determined by ICP analysis (Galbraith Laboratories, Knoxville,
TN).
−
1
3
−1
Samples were initially pressed into a copper holder and an-
alyzed in a sample cell that allowed for heating to 473 K at
−
1
4
K min under flowing H2 (g). After 2 h at 473 K, the sam-
ples were cooled to 298 K under H2 (g) flow. For each sample,
a minimum of 3 scans was collected at 298 K over the appropri-
ate energy range. A second in situ sample cell with its attached
flow system (described in Ref. [32]) was used to analyze various
catalysts during aqueous treatment. The samples were placed
into a 1.5 mm × 20 mm × 6 (or 3) mm (X-ray path length)
sample holder that was then sealed with poly-ether-ether-ketone
disk windows on both sides to allow for the transmission of
photons. The cell enabled data collection while an aqueous so-
2
.2. H chemisorption
2
The metal dispersion and particle size of the Ru/C and Pt/C
catalysts were evaluated by H chemisorption performed on a
2
Micromeritics ASAP 2020 chemisorption analyzer at 308 K
in the pressure range of 10–450 Torr. Before chemisorption,
the catalysts were heated from ambient conditions to 523 K at
lution saturated with 40 bar H was pumped through a packed
2
catalyst bed heated to 473 K. The flow system comprised a ves-
−1
3
4
K min in flowing H (GT&S 99.999%). The catalysts were
sel filled with 60 cm of 0.4 M NaOH aqueous solution that
2
reduced in H2 at 523 K for 90 min, followed by evacuation at
23 K for 4 h. After cooling to 308 K, the catalysts were evac-
uated again for 2 h followed by analysis at 308 K. Surface Ru
was first purged by N , then pressurized with 40 bar H . The
2
2
5
2
H -saturated solution (40 bar) was then pumped through the
3
−1
packed bed of catalyst at a rate of 1.0 cm min while being
heated to 473 K. The X-ray scans were collected while the so-
lution flowed through the sample at 473 K.
and Pt were evaluated by the total amount of H adsorbed at
2
3
03 K extrapolated to zero pressure, assuming a stoichiometry
H/Msurf, M = Ru, Pt) equal to unity. Dihydrogen adsorption
also was performed on the bimetallic PtRu/C and AuRu/C cata-
(
The extended X-ray absorption fine structure (EXAFS) find-
ings were analyzed using the WinXAS software package [33].

284
E.P. Maris et al. / Journal of Catalysis 251 (2007) 281–294
Between 3 and 6 scans were energy-referenced to a Ru pow-
der, a Pt foil, or an Au foil placed between the transmission
and reference ion chambers, then averaged before analysis. The
first-shell Ru–Ru amplitude function and phase shift were ex-
tracted from a Ru/Al2O3 (17.6 wt% Ru) sample. The Ru ref-
erence was assumed to have the bulk crystallographic values
of the Ru hcp structure with a first-shell coordination num-
ber (CN) of 12 and nearest-neighbor distance of 2.68 Å. For
analysis of the EXAFS above the Pt LIII-edge, the energy refer-
ence and the first-shell Pt–Pt amplitude function and phase shift
were extracted from a 4-µm Pt foil (Goodfellow, 99.95%) that
was assumed to have the bulk crystallographic values of the Pt
fcc structure with a first-shell CN of 12 and nearest-neighbor
distance of 2.78 Å. For analysis of the EXAFS above the Au
LIII-edge, the energy reference and the first-shell Au–Au am-
plitude function and phase shift were extracted from a 5-µm
Au foil (Goodfellow, 99.9%). The foil was assumed to have the
bulk crystallographic values of the Au fcc structure with a first-
shell CN of 12 and nearest-neighbor distance of 2.88 Å. The
experimental data were used to calibrate the theoretical photo-
electron scattering amplitude and phase-shift functions for Ru,
Pt, and Au calculated using FEFF8.2 [34]. The amplitude and
phase-shift functions for the heterogeneous absorber-scatterer
pairs (Ru–Pt and Ru–Au) also were generated using FEFF8.2.
equipped with an electronic temperature controller, a mechani-
cal stirrer, a catalyst addition device, and a dip tube for periodic
sampling of the liquid phase. In a typical reaction, glycerol
(Acros Organics, 99%) was diluted with distilled deionized wa-
ter to form a 1 wt% solution. Then 150 ml of this solution was
loaded into the reactor along with the appropriate amount of
NaOH (Mallinckrodt) when necessary. When the monometal-
lic catalysts were used, the appropriate amount of either Ru/C
or Pt/C was loaded into the catalyst addition device to main-
tain a substrate/surface metal (S/Msurf) ratio of approximately
7
00. In studies of the bimetallic catalysts, the same mass of ei-
ther PtRu/C or AuRu/C was used relative to the monometallic
Ru/C case, to maintain comparable Ru loadings. Glycerol hy-
drogenolysis also was investigated with physical mixtures of
the monometallic catalysts. In the case of Ru/C and Pt/C, the
same mass of Ru/C was used as in the monometallic case; how-
ever, the appropriate amount of Pt/C was added to equal the
amount of Pt present in the PtRu/C on a surface Pt atom ba-
sis. The number of surface Pt atoms present in the Pt/C catalyst
was determined by H chemisorption and it was assumed that
2
all Pt atoms in the bimetallic catalyst were on the surface. In
the case of Ru/C mixed with Au/C, the same mass of Ru/C was
used as in the monometallic case, and sufficient Au/C was used
to maintain a total Au loading equivalent to that in the AuRu/C
catalyst.
−
1
For samples evaluated at 298 K, data in the k-range of 3–15 Å
were used for analysis, whereas for the samples examined under
After the catalysts were loaded into the catalyst addition
device, the reactor was sealed and flushed with flowing N2
−
1
aqueous conditions at 473 K, data in the k-range of 3–14 Å
were used. Curve fitting was performed in both R-space and
(
GT&S, 99.998%) at 1 bar for 10 min to remove air in the
1
3
k-space with k and k weighting, and the average of these four
values is reported herein. The fitting procedure results in an es-
timated accuracy of ± 10% for the first-shell CN, ± 0.01 Å for
interatomic distances, and ± 10% for the Debye–Waller factor
headspace. To remove the N2, the reactor was subsequently
flushed with flowing H2 (GT&S, 99.995%) at 2 bar for 20 min.
The reactor was then pressurized to 5 bar with H2 and heated
under moderate agitation (100 rpm) to a final reaction temper-
ature of 473 K. Once this temperature was reached, an initial
liquid sample was removed to mark the start of the reaction.
Catalyst was subsequently introduced into the reaction medium
through the catalyst addition device, the pressure was increased
[
35]. The physical constraints necessary for bimetallic fitting
set forth by Via et al. [36] were maintained during the fitting
process.
2
.5. Ethane hydrogenolysis
to 40 bar with H , and the rate of agitation was increased to
2
4
75 rpm. The reaction was allowed to proceed under these
The hydrogenolysis of ethane was conducted using a fixed-
conditions for 5 h while liquid samples were periodically re-
moved. The reactor was back-filled with H after each sample
bed quartz tubular reactor flow system operating at 473 K
and atmospheric pressure. All data were collected at a total
2
3
−1
to maintain constant pressure. Liquid samples were analyzed by
high performance liquid chromatography (HPLC) using a Ther-
moSeparations Products (TSP) AS1000 autosampler equipped
with a TSP P2000 pump, an Aminex HPX-87H (Bio Rad) col-
umn, a Waters R-401 refractive index detector, and Millennium
data acquisition software. The HPLC column was maintained
flow rate of approximately 45 cm min , where the catalyst
amounts were adjusted to maintain ethane conversions <20%.
Gas flow rates were maintained using mass flow controllers
(
Brooks Instruments, series 5850E) to deliver a stream contain-
ing 10 mol% ethane (GT&S, >99.95%), 33 mol% H2 (GT&S,
9.999%), and He (GT&S, 99.9995%) as balance. The reactor
9
at 333 K, with a mobile phase of 5 mM H SO flowing at
effluent was analyzed by gas chromatography (GC) using an
HP 5890 gas chromatograph equipped with an Alltech CTR1
column and a thermal conductivity detector (TCD). Before col-
lection of kinetic data, the catalysts were reduced in situ by
2
4
−
1
0
.7 ml min . The major liquid-phase products observed were
ethylene glycol, propylene glycol, lactate, and formate; other
byproducts detected in trace amounts by HPLC included glyc-
eraldehyde, methanol, and ethanol.
3
−1
flowing 15 cm min H2 at 473 K for 1 h.
At the conclusion of the 5 h reaction, the reactor was allowed
to cool to room temperature. Gas samples in the headspace were
2
.6. Glycerol hydrogenolysis
removed by a gas-tight syringe and analyzed for CO and CH
2
4
Glycerol hydrogenolysis reactions were conducted semi-
batchwise in a 300-ml stainless steel reactor (Parr Instruments)
by GC, using an HP 5890 gas chromatograph equipped with an
Alltech CTR1 column and TCD.

E.P. Maris et al. / Journal of Catalysis 251 (2007) 281–294
285
Table 1
Results from elemental analysis, H chemisorption, and electron microscopy
2
b
Catalyst
Metal loading
Atomic ratio
(M/Ru)
H
/gcat
H/M
H chem. particle
2
diameter (nm)
TEM particle
diameter (nm)
Surface averaged
diameter (nm)
ads
a
c
d
(wt%)
e
Ru/C
Pt/C
Au/C
5.0
3.0
5.3
–
–
–
21 ± 0.78
6.6 ± 1.1
–
0.43
0.43
–
2.3
2.3
–
2.6
2.5
–
3.1
2.8
–
e
PtRu/C
AuRu/C
1.9 (Pt), 4.7 (Ru)
0.85 (Au), 5.0 (Ru)
0.21
0.087
14 ± 0.86
0.30
0.17
3.3
–
3.3
2.9
3.6
5.8
8.2 ± 0.62
a
M denotes either Pt or Au.
M denotes either Ru or Pt.
b
c
d
e
Calculated as the inverse of the metal dispersion as determined by H chemisorption.
2
ꢀ
ꢀ
3
2
Determined by TEM as ( d )/( d ).
Metal loading provided by manufacturer.
3
. Results and discussion
had a broader PSD, as shown in Fig. 2, resulting in a larger
difference between the mean and the surface-averaged particle
diameter than that reported in Table 1. The particle density was
highest in the 1.5–3.0 nm particles; EDS over an average area of
these small particles revealed primarily Ru, as shown in Fig. 2a.
As shown in Fig. 2b, mid-range particles (5–10 nm) contained
both Au and Ru, and larger particles (>10 nm) contained pri-
marily Au.
These results indicate that although some Au was deposited
onto the Ru particles, significant deposition and growth of Au
monometallic particles occurred on the carbon support. In sum-
mary, electron microscopy of the bimetallic catalysts confirmed
that Pt and Au were successfully deposited onto Ru, with a
greater degree of interaction for Pt.
Table 1 summarizes the elemental composition and hydro-
gen adsorption capacity of the various samples. The hydrogen
adsorption capacity per gram of catalyst (Hads/gcat) for each
sample was determined by extrapolation of the linear portion
of the isotherm to zero pressure. The monometallic Ru/C sam-
ple exhibited the highest hydrogen capacity, followed, in order,
by PtRu/C, AuRu/C, and Pt/C. Note that Pt/C had a lower ca-
pacity than Ru/C because of its lower metal loading; on a per
metal atom basis, the two samples had comparable hydrogen
uptakes. The similar dispersion of the monometallic Ru/C and
Pt/C samples (43%) corresponds to an average metal particle
diameter of about 2.3 nm. Dihydrogen chemisorption was not
attempted on the monometallic Au/C, because it is well known
that Au does not dissociatively adsorb H2 [37]. It is interesting
that the PtRu/C sample had a lower uptake than the monometal-
lic Ru/C sample. Moreover, the addition of Au to Ru/C resulted
in a 2.5-fold decrease in hydrogen uptake. If every Au atom
were deposited on the Ru surface, then the ratio of surface Au
to surface Ru would be 0.20. The decrease in hydrogen ca-
pacity on the AuRu/C sample should be at most only 20%,
yet we observed a more substantial suppression of adsorption.
Consequently, we evaluated the particle sizes for the bimetallic
samples by TEM.
Table 1 reports the mean particle diameter and surface av-
erage diameter for several of the samples as determined by
electron microscopy. Particle size distributions (PSDs) were ob-
tained by measuring at least 200 particles over multiple areas
for each sample. Examination of Table 1 shows good agree-
ment between the particle sizes for the monometallic samples
determined by H2 chemisorption and TEM. Representative mi-
crographs are given for the PtRu/C and AuRu/C bimetallic sam-
ples, along with their corresponding EDS spectra and PSDs in
Figs. 1 and 2. For the PtRu/C bimetallic sample, the particle
size distribution was fairly narrow, in good agreement with the
mean particle diameter and surface-averaged particle diameter
data reported in Table 1. The addition of Pt to Ru/C resulted
in a slight increase in the mean particle diameter, likely due to
the deposition of Pt onto the Ru surface. The EDS of more than
X-ray absorption spectroscopy also was used to character-
ize the catalyst samples. Inspection of the X-ray absorption
near-edge spectra at the Ru K-edge, Pt LIII-edge, and Au LIII-
edge revealed that after sample preparation, Au remained in
a reduced metallic state, whereas Ru and Pt both were in an
oxidized state. However, treatments under both H2 (g) and H2-
saturated aqueous solutions at 473 K were able to reduce the Ru
and Pt to a metallic state.
3
Representative χ data with k weighting are presented in
3
Fig. 3 for the PtRu/C sample. The Fourier-transformed k -
weighted χ functions for the PtRu/C sample are also presented
in Fig. 4. The EXAFS fitting results for all of the samples are
summarized in Table 2. Treatment of the monometallic sam-
ples in H2 (g) resulted in first-shell Ru–Ru and Pt–Pt CNs of
8
.8 and 8.6, respectively, corresponding to a particle size of
approximately 2 nm [38]. This is in good agreement with the
particle sizes determined by H2 chemisorption and TEM. The
Ru–Ru first-shell CN remained essentially constant (8.8 ± 0.4)
for all of the bimetallic samples, regardless of pretreatment. We
previously reported that Ru particles are stable on an activated
carbon support under aqueous-phase conditions [32,39]. Analy-
sis of the EXAFS for the Pt/C sample revealed that treatment
under aqueous-phase conditions increased the Pt–Pt first-shell
CN from 8.6 to 10.4, which is consistent with growth of the
Pt particles. Microscopy of Pt/C recovered after use in the hy-
drogenolysis of glycerol in the presence of 0.8 M NaOH at
473 K and 40 bar H2 confirmed an increase in the average
particle diameter from 2.5 to 3.9 nm and an increase in the
1
0 individual particles revealed the presence of both Ru and Pt
in each case; a representative EDS spectrum obtained from an
individual particle is provided in Fig. 1a. The AuRu/C sample

286
E.P. Maris et al. / Journal of Catalysis 251 (2007) 281–294
Fig. 1. TEM micrographs and corresponding PSDs and EDS spectra for PtRu/C (a) fresh and (b) recovered after use for glycerol hydrogenolysis in 0.8 M NaOH at
73 K and 40 bar H .
4
2
surface-averaged particle diameter from 2.8 to 4.6 nm as a re-
sult of aqueous-phase treatment.
increase in the average particle size was observed by TEM of
the fresh PtRu/C and of the catalyst recovered after use in the
hydrogenolysis of glycerol in 0.8 M NaOH at 473 K and 40 bar
H2. Fig. 1b illustrates a broader PSD for PtRu/C recovered after
catalysis, with the mean particle diameter increasing from 3.3
to 3.9 nm, and the surface-averaged particle diameter increas-
ing from 3.6 to 4.4 nm after treatment in the aqueous phase.
The TEM and EXAFS results both show that sintering of metal
particles due to aqueous-phase processing was not as severe on
the bimetallic PtRu/C sample as the monometallic Pt/C sample.
Apparently, the Ru particles were able to stabilize the Pt against
coalescence and sintering.
The Fourier transforms presented in Figs. 4a and 4b illustrate
the Pt and Ru interaction in the bimetallic sample after reduc-
tion in H2 (g). The structure around Ru did not appear to be
significantly altered from that of the Ru bulk reference (Fig. 4a).
Because the Ru–Ru first-shell CN reported in Table 2 did not
vary from Ru/C to PtRu/C, the Ru particles were not altered
by the deposition of Pt; however, comparison of the Fourier
transform of the PtRu/C sample with the Pt foil (Fig. 4b) re-
vealed interaction of Pt with Ru. The Pt–Ru first-shell CN was
2
.2 out of a total Pt first-shell CN of 7.8 (from Pt–Pt and Pt–
Ru), indicating a substantial presence of Ru in the average first
coordination sphere of Pt. It is interesting to note that the total
first-shell Pt CN did not change significantly when the PtRu/C
sample was treated under aqueous-phase conditions. A slight
The Fourier transforms presented in Figs. 5a and 5b illustrate
the extent of Au and Ru interaction for the AuRu/C bimetal-
lic sample after reduction in H2. Similar to the PtRu/C sample,
the Ru particles maintained a structure corresponding to the

E.P. Maris et al. / Journal of Catalysis 251 (2007) 281–294
287
Fig. 2. TEM micrographs of AuRu/C with corresponding PSD and EDS spectra. (a) Particles in the range of 1.5–3 nm consist primarily of monometallic Ru;
the EDS spectrum was taken from a ∼80 nm diameter circular area and is shown to the right of the micrograph. (b) Particles in the 5–10 nm range contain both Au
and Ru; the EDS spectra of five individual particles are presented to the right of the micrograph.
Ru bulk reference (Fig. 5a). The Ru–Ru first-shell CN was
.9 in comparison to the Ru–Au first-shell CN of 0.1, indi-
modified by the addition of Pt or Au. The hydrogenolysis of
ethane to methane is a well-known structure-sensitive reaction
on Ru [15]. Fig. 6 summarizes the activity results of the various
catalysts in the hydrogenolysis of ethane at 473 K. Equivalent
masses of the Ru/C and the bimetallic PtRu/C and AuRu/C
catalysts were used in these studies to maintain a constant Ru
loading. In the studies of the monometallic Pt/C and the physi-
cal mixture of the Ru/C and Pt/C catalysts, the amount of Pt/C
used was determined by matching the amount of surface Pt to
that present in the bimetallic PtRu/C catalyst (assuming that all
Pt was present on the surface). The amount of Au/C used in
the studies of the monometallic Au/C catalyst and the phys-
ical mixture of Au/C and Ru/C was determined by matching
the total metal loading of Au present in the AuRu/C catalyst.
A turnover frequency (TOF) for ethane hydrogenolysis of 0.049
8
cating very poor interaction of Au with Ru, on average. The
Fourier transform of the Au EXAFS for the AuRu/C sample
in Fig. 5b showed a slight modification of the Au structure on
the bimetallic catalyst compared with Au foil. Fitting results
for the catalyst resulted in a Au–Au first-shell CN of 6.9 and a
Au–Ru first-shell CN of 1.0, indicating poor interaction of Au
with Ru. Moreover, these EXAFS results are consistent with
those from TEM and EDS showing that the sample contained
bimetallic particles of Au and Ru in addition to monometallic
particles of Ru and Au. Inspection of Fig. 5d and the structural
parameters reported for the AuRu/C sample after treatment B
in Table 2 showed further structural changes of the catalyst due
to aqueous-phase processing.
−
1
A structure-sensitive probe reaction also can be used to as-
certain whether or not the surface of our Ru catalyst has been
± 0.0055 s was obtained on Ru/C, which is comparable to a
rate of ∼0.03 s extrapolated from the work of Doi et al., who
−
1

288
E.P. Maris et al. / Journal of Catalysis 251 (2007) 281–294
3
Fig. 3. Experimental χ data with k weighting for sample PtRu/C; (a, c) Ru K-edge and (b, d) Pt LIII-edge. Treatments, (a, b) 1 atm H at 298 K, following reduction
2
at 473 K; (c, d) aqueous hydrogenolysis conditions, 0.4 M NaOH, 40 atm H , at 473 K.
2
3
Fig. 4. Magnitude of the Fourier transformed k -weighted χ function (not corrected for phase shifts) with non-linear least squares fits for sample PtRu/C; (a, c) Ru
K-edge (Ru–Ru and Ru–Pt shells) and (b, d) Pt LIII-edge (Pt–Pt and Pt–Ru shells). Treatments: (a, b) 1 atm H at 298 K, following reduction at 473 K; (c, d) aqueous
2
hydrogenolysis conditions, 0.4 M NaOH, 40 atm H , at 473 K.
2

E.P. Maris et al. / Journal of Catalysis 251 (2007) 281–294
289
Table 2
Structural parameters for monometallic and bimetallic catalysts derived from EXAFS
a
2
2
Catalyst
Treatment
Absorber-scatterer
1st-shell CN
Interatomic distance (Å)
ꢀσ (Å )
ꢀE (eV)
0
Ru/C
Pt/C
Pt/C
A
A
B
A
Ru–Ru
Pt–Pt
Pt–Pt
8.8
8.6
10.4
2.65
2.74
2.75
2.66
2.70
2.70
2.73
2.66
2.69
2.69
2.72
2.66
2.72
2.72
2.84
2.65
–
0.0079
0.0077
0.0082
0.0066
0.0037
0.0037
0.0060
0.0084
0.0044
0.0044
0.0090
0.0068
0.013
0.013
0.0091
0.0077
–
1.9
3.4
4.4
3.7
3.7
4.3
4.3
2.6
2.6
3.1
3.1
3.2
3.2
−1.0
−1.0
2.0
–
PtRu/C
Ru–Ru
Ru–Pt
Pt–Ru
Pt–Pt
Ru–Ru
Ru–Pt
Pt–Ru
Pt–Pt
Ru–Ru
Ru–Au
Au–Ru
Au–Au
Ru–Ru
Ru–Au
Au–Ru
Au–Au
8.4
0.5
2.2
5.4
9.0
0.4
2.0
5.8
8.9
0.1
1.0
6.9
9.1
0
PtRu/C
AuRu/C
AuRu/C
B
A
B
0
9.6
–
2.83
–
–
0.014
−2.4
a
(A) 1 bar H at 298 K, following reduction at 473 K; (B) aqueous-phase treatment: 0.4 M NaOH, 40 bar H , at 473 K.
2
2
studied the hydrogenolysis of ethane over a Ru/SiO2 catalyst at
comparable reaction conditions [40].
ethane occurred on Ru, whereas H2 adsorption occurred on Pt
adjacent to Ru.
The activity of Pt/C could not be determined under our hy-
drogenolysis conditions, presumably because of the low load-
ing of Pt in the system. Under similar reaction conditions,
a Pt/SiO2 catalyst was found to be approximately three orders
of magnitude less active than Ru/SiO2 for ethane hydrogenol-
ysis [40,41]. Therefore, given the low loadings of Pt/C, any
conversion of ethane to methane likely fell below the detection
limits of our system.
In agreement with previous studies [22], we found that the
Au/C catalyst was inactive for ethane hydrogenolysis. A phys-
ical mixture of the Ru/C and Au/C yielded a TOF comparable
to that of the Ru/C monometallic catalyst, again where the TOF
of the mixture was based solely on the Ru contribution, due to
the inactivity of Au. However, the bimetallic AuRu/C displayed
a TOF of 0.011 ± 0.0020 s , indicating a more than fourfold
decrease in activity compared with the Ru/C catalyst. These ac-
tivity results corroborated those from H2 chemisorption, TEM,
and EXAFS, all of which demonstrated that Au modified the
surface of Ru.
We then evaluated the activity of the bimetallic PtRu/C and
AuRu/C catalysts in the aqueous-phase hydrogenolysis of glyc-
erol both under neutral conditions and in the presence of base.
The results of these studies are summarized in Table 3. The
overall turnover frequency (TOFoverall) is reported as the rate of
glycerol converted at approximately 20% conversion per sur-
face metal atom as determined by H2 chemisorption. For runs
involving a physical mixture of the monometallic catalysts, the
total number of surface metal atoms was determined from the
individual contributions of Ru/C and Pt/C to the mixture. The
carbon balance is given as the percentage of carbon accounted
for in the system (both liquid and gas phases) at the end of 5 h.
The amount of missing carbon in the system can be explained
by a loss of gas-phase products during sampling of the reac-
tion mixture. Because the reactor system was not equipped to
separate and analyze the gas and liquid phase products simul-
taneously, a small amount of gas-phase products inevitably was
lost during liquid sampling. Thus, larger errors in the mater-
ial balance would be expected for runs with larger gas-phase
product inventories. The product selectivities of the reactions
are summarized in Table 4. The selectivities are reported at ap-
proximately 20% conversion of glycerol and after the reaction
−
1
A physical mixture of the monometallic Ru/C and Pt/C cata-
lysts exhibited a TOF similar to the Ru/C, however, the TOF of
the physical mixture was based solely on the Ru contribution,
because Pt was effectively inactive for ethane hydrogenolysis.
Therefore, in studies involving a physical mixture of the cata-
lysts, the TOF was determined by the rate of methane formation
per surface Ru atom as determined by H2 chemisorption. Inter-
−
1
estingly, the PtRu/C catalyst gave a TOF of 0.039 ± 0.0047 s
,
which was not significantly lower than the value for the Ru/C
catalyst or the physical mixture of the two monometallic cata-
lysts. Because the TEM and EDS analyses detected bimetallic
Pt–Ru particles, and EXAFS indicated an interaction between
the Ru and Pt, we expected the Pt to substantially depress the
activity of the PtRu/C catalyst. One possible explanation for
the observed activity of the PtRu/C catalyst is that the Pt and
Ru were acting cooperatively. It has been suggested that in
bimetallic systems, each metal can separately activate a dif-
ferent reactant, and, thereafter, surface diffusion allows for mi-
gration and reaction of the activated species to form the prod-
uct [42]. Ethane hydrogenolysis on the monometallic catalysts
revealed that Ru is much more efficient than Pt for C–C cleav-
age; however, the H2 chemisorption indicated that both metals
had equivalent capacity for dissociative adsorption of H2. It is
possible that for the PtRu bimetallic system, C–C cleavage of

290
E.P. Maris et al. / Journal of Catalysis 251 (2007) 281–294
3
Fig. 5. Magnitude of the Fourier transformed k -weighted χ function (not corrected for phase shifts) with non-linear least squares fits for sample AuRu/C; (a, c) Ru
K-edge and (b, d) Au LIII-edge. Treatments: (a, b) 1 bar H at 298 K, following reduction at 473 K; (c, d) aqueous hydrogenolysis conditions, 0.4 M NaOH, 40 bar
2
H , at 473 K.
2
The conversion of glycerol under neutral conditions on the
Ru-, Pt-, and Au-containing catalysts is shown in Fig. 7. Com-
parison of the TOFs reported in runs 1, 2, and 3 of Table 3
together with the conversion profiles shown in Fig. 7, demon-
strate that under neutral conditions, Ru/C is more active than
Pt/C and Au/C is completely inactive for the hydrogenolysis of
glycerol. Because TEM and EXAFS analyses confirmed that
aqueous-phase processing caused sintering of the Pt particles,
the TOF based on surface Pt determined by H chemisorption
2
on a fresh catalyst may be underestimated. For this reason, the
reaction rates normalized per total metal atoms are also reported
in Table 3.
On a total metal basis, Ru/C was still superior to Pt/C and
Au/C in glycerol hydrogenolysis under neutral conditions. This
activity trend is the same as that for ethane hydrogenolysis.
From the product selectivities reported in Table 4, Ru appar-
ently favored the production of ethylene glycol, whereas Pt
favored the production of propylene glycol under neutral con-
ditions. This can be explained by the mechanism of glycerol
hydrogenolysis presented in Scheme 1 [12]. In this mechanism,
both metals catalyze the initial dehydrogenation of glycerol
to glyceraldehyde; however, a metal-catalyzed C–C cleavage
pathway leading to ethylene glycol formation is available on
Ru that does not occur significantly on Pt, and thus, a base-
catalyzed retro-aldol reaction accounts for C–C cleavage prod-
ucts in the presence of Pt. Over both metals, the scission of
Fig. 6. Turnover frequencies for the hydrogenolysis of ethane at 473 K over
monometallic, bimetallic, and physical mixtures of the catalysts. Error bars rep-
resent 95% confidence limits.
was stopped at 5 h. The selectivity of each product was based
on the carbon selectivity, where
moles of carbon in specific product
selectivity =
.
moles of carbon in all observed products

E.P. Maris et al. / Journal of Catalysis 251 (2007) 281–294
291
Table 3
a
Rates of glycerol hydrogenolysis over monometallic and bimetallic catalysts
b
c
−1 d
3
Run #
Catalyst
Base
Conversion (%)
Carbon balance
TOF
(s
)
Rate × 10
overall.
−
1
−1 e
(
mol s mol metal
)
1
2
3
4
5
6
7
8
9
0
1
2
3
Ru/C
Pt/C
Au/C
None
None
None
None
None
None
None
NaOH
NaOH
NaOH
NaOH
NaOH
NaOH
40
13
0
44
25
84
97
100
81
95
78
92
85
83
90
91
88
90
0.02 ± 0.003
9.9
2.5
–
7.8
4.6
12
5.1
77
141
102
69
0.006 ± 0.002
–
f
g
Ru/C + Pt/C
0.02 ± 0.004
h
Ru/C + Au/C
PtRu/C
AuRu/C
Ru/C
0.01 ± 0.003
0.05 ± 0.02
0.03 ± 0.009
0.2 ± 0.03
0.3 ± 0.05
0.2 ± 0.04
0.3 ± 0.04
0.2 ± 0.07
0.5 ± 0.05
42
25
100
92
100
100
100
100
Pt/C
f
1
1
1
1
Ru/C + Pt/C
PtRu/C
h
Ru/C + Au/C
74
80
AuRu/C
a
Reaction conditions: 1 wt% glycerol solution, 0.8 M NaOH (when used), T = 473 K, PH2 = 40 bar.
b
c
d
e
f
Conversion determined after 5 h.
Percentage of carbon accounted for after 5 h of reaction.
Overall rate determined at ∼20% conversion of glycerol. Error represents 95% confidence limits.
Rate normalized by the total moles of metal in the catalyst.
Physical mixture of the monometallic Ru/C and Pt/C, where the number of surface Pt atoms equals those present in the bimetallic PtRu/C (assuming all Pt was
deposited on the surface of the PtRu/C catalyst).
g
Rates were normalized by the appropriate contribution from Ru/C and Pt/C in the mixture.
Physical mixture of the monometallic Ru/C and Au/C, where the total Au loading equals that present in the bimetallic AuRu/C.
h
Table 4
a
Product selectivities during glycerol hydrogenolysis
b
c
Run #
Catalyst
Base
Conversion (%)
17
Carbon selectivity
d
d
EG
PG
LA
FA
CH
–
CO
–
4
2
1
2
4
5
6
7
8
9
Ru/C
None
0.68
0.47
–
0.32
0.26
–
0
0
40
0
–
0
0
0.26
–
0.01
–
Pt/C
None
–
13
0.17
0.59
0.41
0.50
0.44
0.70
0.49
0.65
0.49
0.13
0.01
0.02
0.02
0.11
0.002
0.15
0.02
0.14
0.01
0.10
0.001
0.79
0.39
0.39
0.50
0.46
0.30
0.24
0.35
0.30
0.36
0.19
0.30
0.47
0.49
0.18
0.37
0.18
0.41
0.20
0.25
0.12
0
0
0
0
Ru/C + Pt/Ce
Ru/C + Au/Cf
PtRu/C
None
19
0
0
–
–
44
0
0
0
0
0.18
–
0.01
–
None
19
25
0
0
0
0
0.09
–
0.01
–
None
23
42
0
0
0
0
0
0
0
0.43
0
0
0
0.36
0
0.27
–
0.21
–
0.02
–
0
–
0.02
–
0.01
–
0.02
–
Trace
AuRu/C
Ru/C
None
18
–
0
–
25
NaOH
NaOH
NaOH
NaOH
NaOH
NaOH
25
00
25
0.47
0.34
0.62
0.49
0.37
0.37
0.41
0.37
0.41
0.35
0.60
0.38
g
1
0.01
Pt/C
–
0
–
92
10
11
12
13
Ru/C + Pt/Ce
PtRu/C
23
00
22
00
24
00
21
00
g
1
1
1
1
0.05
–
0.02
g
0.37
0
0.39
0
Ru/C + Au/Cf
AuRu/C
–
0.02
g
–
0.03
g
0.44
0.01
a
Reaction conditions: 1 wt% glycerol solution, 0.8 M NaOH (when used), T = 473 K, PH2 = 40 bar. EG: ethylene glycol; PG: propylene glycol; LA: lactate;
FA: formate.
b
Run numbers are the same as those listed in Table 3.
c
d
e
Carbon selectivities were determined both at ∼20% conversion of glycerol and at the final conversion attained after 5 h of reaction.
The gas phase was only sampled after the completion of the reaction. Therefore gas phase compositions at 20% conversion of glycerol are not available.
Physical mixture of the monometallic Ru/C and Pt/C, where the number of surface Pt atoms equals those present in the bimetallic PtRu/C (assuming all Pt was
deposited on the surface of the PtRu/C catalyst).
f
Physical mixture of the monometallic Ru/C and Au/C, where the total Au loading equals that present in the bimetallic AuRu/C.
In the form of carbonate.
g

292
E.P. Maris et al. / Journal of Catalysis 251 (2007) 281–294
Fig. 7. Conversion of glycerol under neutral conditions at 473 K and 40 bar H on (a) Ru- and Pt-containing catalysts and (b) Ru- and Au-containing catalysts. Error
2
bars represent 95% confidence limits.
C–O bonds leading to propylene glycol formation is attributed
to a base-catalyzed dehydration reaction.
a bimodal PSD of the AuRu/C catalyst. The as-prepared cat-
alyst had ∼82% of the particles in the 1.5–3 nm range; in
contrast, the AuRu/C recovered after use in the hydrogenoly-
sis of glycerol under neutral conditions had only 48% of the
particles in that size range. Similarly, the as-prepared AuRu/C
catalyst contained only ∼8% of 6–10 nm particles, compared
with 42% in the AuRu/C catalyst recovered after reaction. Fur-
thermore, the density of >10-nm particles also increased after
reaction, with these particles containing primarily Au. Appar-
ently, the aqueous-phase reaction conditions promoted sintering
of the Au particles. The TOF for run 7 based on characteri-
zation of a fresh catalyst is perhaps an overestimation; the Au
initially present on Ru apparently moved off the surface and ag-
gregated, thus exposing additional Ru to the reaction medium.
Normalizing the rate per total metal loading in the catalyst re-
veals a closer agreement between the rate of hydrogenolysis on
the physical mixture of Ru/C and Au/C (run 5) and the bimetal-
lic AuRu/C catalyst (run 7). The product selectivities reported
in Table 4 for AuRu/C and Ru/C were also quite similar; both
favored ethylene glycol production over propylene glycol pro-
duction, with significant production of methane. Interestingly,
the physical mixture of Ru/C and Au/C, although less active
than Ru/C alone, had a slightly greater selectivity to propylene
glycol with decreased formation of methane.
The physical mixture of Ru/C and Pt/C (run 4 in Tables 3
and 4) had initial activity and selectivity comparable to that
of the Ru/C monometallic catalyst, which is expected because
Pt/C was significantly less active than Ru/C under neutral con-
ditions. Interestingly, the bimetallic PtRu/C catalyst also exhib-
ited activity and selectivity similar to that of the monometallic
Ru/C catalyst. Again, these results are similar to those of our
earlier ethane hydrogenolysis study, in which we found that al-
though Pt/C was relatively inactive for hydrogenolysis, PtRu/C
and Ru/C displayed similar activity. As in the case of ethane
hydrogenolysis, we propose that Pt and Ru catalyze separate
steps of the reaction. In a study of the liquid-phase hydrogena-
tion of 2-butanone on carbon-supported Pt–Ru bimetallic cat-
alysts, Breen et al. observed that the bimetallic catalysts were
more than twice as active as the sum of the activities of the
monometallic catalysts [18]. They suggested that the presence
of Pt clusters on the Ru surface could enhance the dissociative
adsorption of H , which, if competitive adsorption between H
2
2
and 2-butanone were a rate-determining process, could enhance
the overall kinetics of the reaction. Applying similar reasoning
to our system, it is possible that the presence of Pt provided
sites for dissociative adsorption of H , whereas Ru provided
2
sites for glycerol adsorption and C–C cleavage. Although Pt/C
itself was less active for glycerol hydrogenolysis, Pt-modified
Ru could be as active as a monometallic Ru/C catalyst.
Inspection of Fig. 7b and runs 5 and 7 in Table 3 reveals
that the addition of Au decreased the hydrogenolysis activity of
Ru/C. However, modifying the Ru surface with Au was almost
the same as a physical mixture of the two monometallic cata-
lysts. Note that the TOF reported for run 7 in Table 3 was deter-
Fig. 8 presents the glycerol reaction profile over the vari-
ous catalysts in 0.8 M NaOH. The presence of base is known
to significantly enhance the rate of glycerol hydrogenolysis
[43–45]. We previously reported that the extent of enhancement
is greater for Pt/C than Ru/C [12], as shown in runs 1, 2, 8, and
9 in Table 3. Whereas direct comparison of the TOFs for these
reactions is somewhat misleading due to the sintering of Pt par-
ticles on the carbon support, the normalized rates still show that
the initial activity of Pt was enhanced by base to a greater ex-
tent than Ru. Inspection of the product selectivities for Ru/C
and Pt/C reported in Table 4 shows that in the presence of base
(entries 8 and 9), lactate was the major product formed during
mined using the metal dispersion obtained from H chemisorp-
2
tion. Nonetheless, TEM of the AuRu/C recovered after use in
this reaction showed a change in the particle morphology as a
result of aqueous-phase processing. After reaction, there was

E.P. Maris et al. / Journal of Catalysis 251 (2007) 281–294
293
Fig. 8. Conversion of glycerol in the presence of 0.8 M NaOH at 473 K and 40 bar H over (a) Ru-, Pt-containing catalysts and (b) Ru- and Au-containing catalysts.
2
Error bars represent 95% confidence limits.
glycerol hydrogenolysis. Propylene glycol was the next most
abundant product formed on both catalysts at low conversions
glycerol hydrogenolysis in the presence of 0.8 M NaOH, the
AuRu/C bimetallic catalyst as well as a physical mixture of the
monometallic Ru/C and Au/C catalysts exhibited comparable
rates. In addition, the product selectivities also were similar,
as reported in Table 4. Because the modification of Ru with
Au resulted in decreased activity for ethane hydrogenolysis, we
anticipated a corresponding decrease in rate for glycerol hy-
drogenolysis. However, the glycerol hydrogenolysis rate was
not decreased by Au, suggesting that the Au may have migrated
off the Ru surface during reaction in the aqueous phase. Indeed,
TEM and EDS of the AuRu/C catalyst recovered after use in the
hydrogenolysis of glycerol in 0.8 M NaOH revealed segregated
regions of monometallic Ru and Au. The small particles ob-
served by TEM were composed of Ru, and the larger particles
were exclusively Au. The change in first-shell CNs reported in
Table 2 also supports this hypothesis. Whereas treatment of the
(
25%). Although ethylene glycol was no longer a major product
formed on Ru/C, its selectivity at high pH was still higher than
that on Pt/C (runs 8 and 9 in Table 4). Furthermore, the prod-
ucts formate, methane, and carbonate were produced over Ru/C
at high glycerol conversion but were not formed over Pt/C.
The initial activity of PtRu/C (run 11 of Table 3) for glycerol
hydrogenolysis was similar to that of the Ru/C in the presence
of base, as depicted in Fig. 8a. Inspection of runs 10 and 11
of Table 4 also shows that both the physical mixture of Ru/C
and Pt/C and the bimetallic PtRu/C exhibited similar product
selectivities to the monometallic Ru/C (run 8) in the presence
of base. One possible explanation for these similar product se-
lectivities is that the Pt and Ru particles segregated under the
aqueous-phase processing conditions. To evaluate this hypothe-
sis, TEM was performed on the PtRu/C catalyst recovered after
use in the hydrogenolysis of glycerol in the presence of 0.8 M
NaOH, as shown in Fig. 1b. Because EDS analysis of more than
sample in H (g) resulted in a Ru–Au first-shell CN of 0.1 and
2
an Au–Ru first-shell CN of 1.0, exposure of the sample to a H2-
saturated basic solution resulted in complete loss of the Ru–Au
interaction detected by EXAFS. Therefore, during glycerol hy-
drogenolysis, Au appeared to move from the Ru surface to the
carbon support. Because the remaining Ru particles were still
highly dispersed, the activity and selectivity of AuRu/C were
similar to those of the parent monometallic Ru/C.
1
0 individual particles confirmed the presence of both Ru and
Pt, the 2 metals apparently did not separate on the carbon sup-
port during aqueous-phase processing. In situ EXAFS of the
PtRu/C sample provided further evidence that the Pt and Ru
particles remained intact as bimetallic clusters during aqueous-
phase processing. Inspection of the Fourier transforms at the Pt
LIII-edge shown in Fig. 4d indicates that even in the aqueous
phase, under basic conditions, there was still significant inter-
action of Pt with Ru. Furthermore, the first-shell CNs for Ru–Pt
and Pt–Ru reported in Table 2 remained relatively unchanged
as a function of pretreatment. These results, which agree with
those from TEM, confirm that Pt and Ru particles did not mi-
grate substantially during glycerol hydrogenolysis. Evidently,
modification of the Ru surface with Pt did not alter the activity
or selectivity for glycerol hydrogenolysis, a conclusion similar
to that derived from ethane hydrogenolysis.
4. Conclusion
Activated carbon-supported bimetallic PtRu and AuRu cat-
alysts were prepared by surface redox reactions for use in
the aqueous-phase hydrogenolysis of glycerol. Results from
characterization by H2 chemisorption, analytical electron mi-
croscopy, and X-ray absorption spectroscopy confirmed that Pt
or Au was successfully deposited on supported Ru particles.
In the case of PtRu/C, TEM and EDS showed the presence of
relatively small bimetallic Pt–Ru particles with a narrow PSD
(mean diameter ∼3.3 nm). Both TEM of the PtRu/C catalyst
recovered after use in the hydrogenolysis of glycerol in the pres-
Comparing the activity results for Ru and Au catalysts (runs
8
, 12, and 13 of Table 3 and inspection of Fig. 8b) used for

294
E.P. Maris et al. / Journal of Catalysis 251 (2007) 281–294
ence of 0.8 M NaOH and in situ XAS of PtRu/C under an H2-
saturated basic solution revealed that the Pt and Ru remained in
close contact, indicating that segregation of the metals did not
occur during reaction. Modification of the Ru surface with Pt
did not alter the catalytic activity of PtRu/C for ethane or glyc-
erol hydrogenolysis compared with the monometallic Ru/C.
TEM and EDS showed that the as-prepared AuRu/C cata-
lyst had a broader PSD than the PtRu/C; however, individual
particles containing both Ru and Au were observed. These re-
sults were corroborated by X-ray absorption spectroscopy. The
hydrogenolysis of ethane, a structure-sensitive probe reaction,
confirmed that Au modified a fraction of the Ru surface atoms;
however, the activity and selectivity of AuRu/C in the hy-
drogenolysis of glycerol was similar to the monometallic Ru/C.
The activity results, coupled with postreaction characterization,
strongly suggest that Au migrates off of Ru during glycerol hy-
drogenolysis and agglomerates on the support.
[7] ADM plans polyols unit using renewable feedstock (Archer Daniels Mid-
land Co. to build renewable plant) (Brief article), Chem. Week 167 (1)
(2005) 23.
[
[
8] E. Chynoweth, ICIS Chem. Bus. (Weekly) 1 (1) (2006) 39.
9] I.T. Clark, Ind. Eng. Chem. 50 (1958) 1125.
[
10] C. Montassier, D. Giraud, J. Barbier, Polyol Conversion by Liquid Phase
Heterogeneous Catalysis over Metals, Elsevier, Amsterdam, 1988, p. 165.
11] D.K. Sohounloue, C. Montassier, J. Barbier, React. Kinet. Catal. Lett. 22
[
(1983) 391.
[
[
12] E.P. Maris, R.J. Davis, J. Catal. 249 (2007) 328.
13] C. Montassier, J.C. Menezo, L.C. Hoang, C. Renaud, J. Barbier, J. Mol.
Catal. 70 (1991) 99.
[14] D.G. Lahr, B.H. Shanks, J. Catal. 232 (2005) 386.
15] J.H. Sinfelt, Bimetallic Catalysts: Discoveries, Concepts, and Applica-
tions, Wiley, New York, 1983.
16] C. Montassier, J.C. Menezo, J. Moukolo, J. Naja, L.C. Hoang, J. Barbier,
J.P. Boitiaux, J. Mol. Catal. 70 (1991) 65.
[17] C. Montassier, J.C. Menezo, J. Naja, P. Granger, J. Barbier, P. Sarrazin,
B. Didillon, J. Mol. Catal. 91 (1994) 119.
18] J.P. Breen, R. Burch, K. Griffin, C. Hardacre, M. Hayes, X. Huang, S.D.
O’Brien, J. Catal. 236 (2005) 270.
[
[
[
Both bimetallic catalysts, PtRu/C and AuRu/C, favored the
formation of ethylene glycol over propylene glycol during glyc-
erol hydrogenolysis in neutral pH conditions. In the presence of
base, the bimetallic catalysts favored the production of lactate
and propylene glycol over ethylene glycol. The product selec-
tivities over the bimetallic catalysts at both high pH and neutral
conditions were quantitatively similar to those observed over
the monometallic Ru catalyst.
[
19] Y.J. Zhang, A. Maroto-Valiente, I. Rodriguez-Ramos, Q. Xin, A.
Guerrero-Ruiz, Catal. Today 93–95 (2004) 619.
[20] E. Antolini, L. Giorgi, F. Cardellini, E. Passalacqua, J. Solid State Elec-
trochem. 5 (2001) 131.
[
21] H. Okamoto, P.R. Subramanian, L. Kacprzak, Binary Alloy Phase Dia-
grams, in: T.B. Massalaski (Ed.), ASM International, USA, 1990.
22] S. Galvagno, J. Schwank, G. Parravano, F. Garbassi, A. Marzi, G.R.
Tauszik, J. Catal. 69 (1981) 283.
[
[23] A.G. Shastri, J. Schwank, J. Catal. 95 (1985) 271.
[24] A.G. Shastri, J. Schwank, J. Catal. 95 (1985) 284.
[
25] P. Gallezot, N. Nicolaus, G. Fleche, P. Fuertes, A. Perrard, J. Catal. 180
Acknowledgments
(1998) 51.
[26] B.P. Block, J.C. Bailar, J. Am. Chem. Soc. 73 (1951) 4722.
This work was supported by the National Science Foun-
dation (Grants CTS-0313484 and CTS-0624608). Research
was carried out on beamline X-18B (operated by the Syn-
chrotron Catalysis Consortium, funded by US Department of
Energy [DOE] Grant DE-FG02-05ER15688) at the National
Synchrotron Light Source, Brookhaven National Laboratory,
which is supported by the US DOE, Division of Materials
Sciences and Division of Chemical Sciences, under Contract
DE-AC02-98CH10886. Partial support also was provided by
US DOE, Office of Basic Energy Sciences Grant DE-FG02-
[27] J. Barbier, P. Marecot, G. Delangel, P. Bosch, J.P. Boitiaux, B. Didillon,
J.M. Dominguez, I. Schifter, G. Espinosa, Appl. Catal. A 116 (1994) 179.
[28] C.L. Pieck, P. Marecot, J. Barbier, Appl. Catal. A 143 (1996) 283.
[29] C.L. Pieck, P. Marecot, J. Barbier, Appl. Catal. A 134 (1996) 319.
[30] C.E. Fiori, C.R. Swyt, R.L. Myklebust, Desktop Spectrum Analyzer (Pro-
gram), National Institute of Standards and Technology, Gaithersburg, MD,
1992.
[
[
[
31] W. Rasband, ImageJ Software v. 1.37 (program), National Institute of
Health, http://rsb.info.nih.gov/ij/.
32] E.P. Maris, W.C. Ketchie, V. Oleshko, R.J. Davis, J. Phys. Chem. B 110
(2006) 7869.
33] T. Ressler, WinXAS v. 2.33 (program), Berlin, 2001.
9
5ER14549.
[34] A.L. Aunkudinov, C. Bouldin, J.J. Rehr, J. Sims, H. Hung, Phys. Rev. B 65
2002) 104107.
(
[
[
[
35] D.C. Koningsberger, B.L. Mojet, G.E. van Dorssen, D.E. Ramaker, Top.
Catal. 10 (2000) 143.
36] G.H. Via, K.F. Drake, G. Meitzner, F.W. Lytle, J.H. Sinfelt, Catal. Lett. 5
References
[1] L.R. Lynd, C.E. Wyman, T.U. Gerngross, Biotechnol. Progr. 15 (1999)
(1990) 25.
777.
37] X. Wu, B.C. Gerstein, T.S. King, J. Catal. 123 (1990) 43.
[
[
[
2] B.E. Dale, J. Chem. Technol. Biotechnol. 78 (2003) 1093.
3] B. Kamm, M. Kamm, Appl. Microbiol. Biotechnol. 64 (2004) 137.
4] A.J. Ragauskas, C.K. Williams, B.H. Davison, G. Britovsek, J. Cairney,
C.A. Eckert, W.J. Frederick, J.P. Hallett, D.J. Leak, C.L. Liotta, J.R. Mie-
lenz, R. Murphy, R. Templer, T. Tschaplinski, Science 311 (2006) 484.
5] T. Werpy, G. Petersen, US DOE Report: Top Value Added Chemicals from
Biomass, Vol. 1: Results of Screening for Potential Candidates from Sug-
ars and Synthesis Gas, 2004.
[38] A. Jentys, Phys. Chem. Chem. Phys. 1 (1999) 4059.
[39] W.C. Ketchie, E.P. Maris, R.J. Davis, Chem. Mater. 19 (2007) 3406.
[40] Y. Doi, H. Miyake, K. Soga, J. Mol. Catal. 48 (1988) 123.
[41] R.D. Cortright, R.M. Watwe, J.A. Dumesic, J. Mol. Catal. A 163 (2000)
91.
[42] R.D. Adams, J. Organomet. Chem. 600 (2000) 1.
[43] B. Casale, A.M. Gomez, US Patent 5 276 181 (1994).
[44] L. Schuster, M. Eggersdorfer, US Patent 5 616 817 (1997).
[45] D.G. Lahr, B.H. Shanks, Ind. Eng. Chem. Res. 42 (2003) 5467.
[
[6] C. Chiu, Ind. Eng. Chem. Res. 45 (2006) 791.