Top Catal
DOI 10.1007/s11244-013-0115-1
ORIGINAL PAPER
Aqueous Phase Glycerol Reforming with Pt and PtMo Bimetallic
Nanoparticle Catalysts: The Role of the Mo Promoter
Paul J. Dietrich • Tianpin Wu • Aslihan Sumer •
James A. Dumesic • Julius Jellinek • W. Nicholas Delgass
Fabio H. Ribeiro • Jeffrey T. Miller
Ó Springer Science+Business Media New York 2013
Abstract The turnover rate (TOR, normalized to sites
measured by CO chemisorption before reaction) and
selectivity for the aqueous phase reforming of glycerol
have been determined for Pt/C and PtMo/C catalysts.
While the TOR of PtMo/C is higher than that of Pt/C by
about 4 times at comparable conversion, the selectivity to
C–O bond cleavage is higher, thus reducing the H2 yield at
high conversion. Under reaction conditions on Pt/C, CO is
observed as the most abundant Pt surface species with a
fractional coverage of about 0.6 using operando X-ray
absorption spectroscopy. Since there is little CO in the
effluent (CO2:CO ratios [ 100:1, when CO is detected), it
is thought that surface CO is converted to H2 and CO2 by
the water gas shift reaction. DFT calculations suggest that
the role of metallic Mo is to alter the electronic properties
of Pt lowering the binding energy of CO and reducing the
activation energies of dehydrogenation and C–O bond
cleavage. Because the activation energy for C–O cleavage
is lowered more than for dehydrogenation, the selectivity
for C–O bond cleavage is increased, ultimately lowering
the H2 yield compared to Pt/C.
P. J. Dietrich  W. N. Delgass  F. H. Ribeiro (&)
School of Chemical Engineering, Purdue University,
W. Lafayette, IN 47907, USA
e-mail: fabio@purdue.edu
T. Wu  A. Sumer  J. Jellinek  J. T. Miller (&)
Argonne National Laboratory, Chemical Science and
Engineering, Argonne, IL 60439, USA
e-mail: millerjt@anl.gov
J. A. Dumesic
Department of Chemical and Biological Engineering, University
of Wisconsin-Madison, Madison, WI 53706, USA
•
Keywords Glycerol reforming  Aqueous reforming 
PtMo bimetallic catalyst  Delta-XANES (DXANES) 
Operando X-ray absorption spectroscopy  Mo promoter
of Pt  Hydrogen production from biomass
1 Introduction
Biomass has been proposed as an alternative feedstock for
fuels and chemicals in response to calls to lower depen-
dence on petroleum. However, petroleum contains low
oxygen levels, in contrast to the high oxygen content of
biomass. Even conversion by acid hydrolysis or pyrolysis
generates oxygen rich bio-oils, which then must be treated
in extensive downstream processing [1, 2]. To turn bio-oils
or biomass derived carbohydrates into usable products,
e.g., energy dense fuels with similar properties to petro-
leum derived fuels that can be used as drop-in replace-
ments, a large amount of hydrogen must be added to
remove the oxygen. To maintain biomass conversion pro-
cesses as carbon neutral, hydrogen generation must utilize
renewable feed stocks.
Catalytic conversion of biomass via aqueous phase
reforming has been demonstrated as a technically feasible
process for generating hydrogen [3]. These liquid phase
reactions are thermodynamically favorable [4] and operate
at mild temperatures and pressures compared to traditional
hydrocarbon steam reforming. Product distributions can be
tuned by catalyst formulation or conditions to make either
gas phase products (H2, syngas) [3, 5, 6] or liquid phase
oxygenates that can be further upgraded into transportation
fuels [7]. Carbon or metal oxide supported Pt catalysts have
been studied in the past, often with a second metal to act as
a rate promoter [8]. Carbon supports have been demon-
strated to be more stable and have high activity under harsh
123

aqueous phase conditions compared to metal oxide sup-
ports [9].
Previously, a PtMo/C bimetallic reforming catalyst was
characterized using X-ray absorption spectroscopy, trans-
mission electron microscopy, X-ray photoelectron spec-
troscopy, density functional theory, and reaction kinetics to
understand the active phase of the catalyst and how Mo
changes particle morphology [10]. In this work, we com-
pare this previously characterized catalyst to its monome-
tallic analog, a Pt/Norit C catalyst, to understand the
effect of Mo on the catalytic activity. In particular, we
investigate the effect of the secondary metal on rates and
product selectivity, and combine with operando spectros-
copy and DFT calculations to understand how the elec-
tronic and morphological properties of the catalyst
contribute to the differences in activity and selectivity.
Reaction kinetics and operando experiments were con-
ducted in a fixed-bed, plug flow system, using aqueous
glycerol as a model compound feed with a 1:1 C:O stoi-
chiometry typical of biomass derived carbohydrates. The
results indicate that the addition of Mo increases the rate of
reaction, but has an undesired side effect of shifting
product selectivity towards products displaying significant
deoxygenation. Spectroscopy and DFT results indicate that
surface Mo plays a key role in the undesired side reaction,
while still having a promotional effect on the overall rate.
2 Experimental
2.1 Catalyst Preparation
The carbon supported Pt catalyst was prepared by incipient
wetness impregnation of a commercially available carbon
black (Norit-SX1G, 1,000 m2 g-1) with an aqueous solu-
tion of H2PtCl66H2O (Strem Chemicals) to yield a catalyst
with 5 wt% Pt loading. For the bimetallic PtMo catalyst,
Mo was added by incipient wetness impregnation of a
Mo(CO)6 in THF solution on the dried and unreduced
5 wt% Pt on carbon catalyst (atomic Pt:Mo ratio of 1:1).
The catalysts were dried in air for 12 h at 25 °C. After
drying, each catalyst was reduced for 2 h in flowing H2 at
200 sccm by heating to 450 °C at 0.5 °C min-1. The cat-
alyst was cooled in He and passivated in flowing 2 % O2 in
He at 25 °C. Prior to reaction, the passivated catalyst was
re-reduced for 2 h under 50 sccm of 4 % H2 in Ar at
300 °C.
2.2 Glycerol Reforming Experiments
Glycerol reforming experiments were performed in a spe-
cially designed operando plug flow reactor system that
allows for the simultaneous measurement of both kinetics
123
Top Catal
and X-ray absorption spectroscopy [11]. Quartz tubes
(6.35 mm/3.124 mm OD/ID, ChemGlass) or Sigradur
G-tube glassy carbon tubes (10 mm/4 mm OD/ID, Hoch-
temperatur-Werkstoffe GmbH) connected to standard
Swagelok fittings with graphite ferrules (Chromalytic
Technology, Pty Ltd.) were used as the reactors. The glassy
carbon tubes are a unique material that allow for excellent
X-ray transparency and operation at high temperature and
pressure [12]. Catalyst loadings varied between 20 and
250 mg, and the catalyst bed was held in place by stainless
steel mesh frits with plugs of quartz wool between the frits
and catalyst bed. A type K thermocouple (Omega Engi-
neering) was inserted into the top of the catalyst bed and
used to continuously monitor and control reactor temper-
ature via a PID controller (Omega Engineering) and an
aluminum heating block. A syringe pump (Teledyne ISCO)
was used to control the feed rate of a 30 wt% glycerol in
water solution between 0.005 and 0.5 mL min-1, corre-
sponding to a weight hourly space velocity (WHSV, g feed
g cat-1 h-1) range of 0.5 to 100 h-1. System pressure was
controlled by a back pressure regulator (GO Regulators),
and 27 mL min-1 Ar was used as a sweep gas through the
gas/liquid phase separator (Jergusen) to remove gas phase
products, which were analyzed online by 6890 and 7890
gas chromatographs (Agilent Technologies) equipped with
packed columns (Carboxen 1000) and thermal conductivity
detectors. Liquid effluent was periodically collected from
the phase separator and analyzed offline in a 7890 GC
equipped with a flame ionization detector and capillary
column (DB-WAX). System mass balances generally
closed to 100 ± 2 %.
2.3 Transmission Electron Microscopy
A suspension of catalyst powder (1–2 mg) in pure ethanol
(1–3 cc) was dispersed using an ultrasonic bath for about
5 min. Several drops of the resulting solution were spread
over a 200 mesh copper TEM grid coated with a lacey
carbon film. HRTEM images of the samples were taken on
an FEI Titan apparatus operating at 300 kV equipped with
a Gatan Imaging Filter (GIF). The images were captured
digitally on a CCD camera (1024 9 1024 pixel) and
recorded using the Gatan DigitalMicrograph software.
Metal particle sizes and distributions were determined for
fresh and used catalysts using the ImageJ software pack-
age. The total number of particles counted was 79 and 71,
respectively.
2.4 X-Ray Absorption Spectroscopy (XAS)
Pt L3-edge (11.564 keV) X-ray absorption measurements
were conducted on the insertion device beamline of the
Materials Research Collaborative Access Team (MRCAT,
Top Catal
10-ID) at the Advanced Photon Source (APS) at Argonne
National Laboratory. Ionization chambers were optimized
for the maximum current with linear response (ca. 1010
photons detected s-1). A third detector in the series
simultaneously collected a Pt foil reference spectrum with
each measurement for energy calibration. The X-ray beam
was 0.7 mm 9 1.0 mm, and data was collected in trans-
mission mode with 3 min quick scans.
For ex situ measurements, Pt/C catalysts were reduced in
a continuous-flow reactor, which consisted of a quartz tube
(100 OD, 1000 length) with Kapton windows sealed by two
Ultra-Torr fittings. Two ball valves welded to the Ultra-Torr
fittings served as gas inlet and outlet. The catalyst was gently
pressed into a cylindrical sample holder consisting of 6 wells,
forming a self-supporting wafer. The catalyst amount was
calculated to give an absorbance (lx) of approximately 2.0.
The catalysts were reduced in 3.5 % H2/He (100 mL min-1)
at 300 °C (measured by an internal thermocouple at the
position of the samples), purged with He for 10 min, and then
cooled to RT. Traces of oxygen and moisture were removed
by means of a gas purifier (Matheson PUR-Gas Triple Purifier
Cartridge). XAS spectra were collected on the reduced sam-
ples in helium at RT. Following spectrum collection on the
reduced catalyst, the catalyst was exposed to 3.5 % H2/He or
1 % CO/He at room temperature. The catalysts were then
purged with He at RT before XAS spectrum collection. To
remove adsorbed gases, the samples were heated to 300 °C
under He for 30 min, then cooled to RT for subsequent
adsorption. Spectra with adsorbed water were captured in the
operando reactor after reduction by filling the reactor with DI
H2O purged with N2, and scanned at ambient conditions,
20 °C and 26 bar, and 230 °C and 26 bar (all liquid phase
water) prior to introducing the glycerol feed.
Operando XAS measurements were performed under
glycerol reforming conditions identical to those of the reaction
studies using the reactor described above. Reactions at WHSV
of 16 and 0.5 h-1, characteristic of low and high conversion,
were conducted. Multiple XAS spectra were collected and
averaged to obtain adequate signal-to-noise ratio. The Debye–
Waller factor (DWF, Dr2) is temperature dependent, and the
value at reaction temperature was determined by obtaining the
XAS spectrum first at the reaction temperature and again after
cooling to room temperature and assuming the same coordi-
nation number at both temperatures. The gas reaction products
were continuously monitored to determine when the system
had reached steady state. Liquid products were collected after
the catalyst had been at steady state for several hours and were
analyzed off-line.
2.4.1 XAS Data Analysis
The normalized energy calibrated Pt L3 edge X-ray absorp-
tion near edge spectroscopy (XANES) spectra were obtained
by standard methods for XANES on the reduced catalyst and
catalyst after gas adsorption. The edge energy was determined
from the position of the maximum in the first peak of the 1st
derivative of the XANES. After adsorption of different gases,
the Pt L3 adsorption edge shifts and the DXANES spectra were
obtained by subtracting the XANES spectrum of the reduced
catalyst scanned in He from the spectrum after saturation by
adsorbing gas. For extended X-ray absorption fine structure
(EXAFS) analysis, experimental phase shift and backscat-
tering amplitudes were obtained from the Pt foil for Pt–Pt
(coordination number 12 at bond distance 2.77 Å) scattering.
Standard procedures using WINXAS 3.1 software were
employed to fit the XAS data. The EXAFS parameters were
obtained by a least square fit in R-space of the k2-weighted
Fourier transform (FT) data from 2.7 to 12.2 Å-1 and fit from
1.6 to 3.2 Å. The fit parameters were determined by fitting
both the real and imaginary parts of the Fourier transform in
R-space of the k2-weighted EXAFS spectra.
2.5 Computational Details
The computations were performed within the DFT framework
with the PW91 exchange–correlation functional [13, 14] and
the Stuttgart pseudo-potentials/basis sets for Pt and Mo [15] as
implemented in the NWChem package [16]. This selection is
based on extensive tests of various alternative choices, which
were performed on Pt and Mo atoms and dimers [the details
will be given elsewhere (J. Jellinek, A. Sumer, in prepara-
tion)]. For the C, O, and H atoms, the corresponding all-
electron 6-311G(2df,2pd) basis sets were utilized [15]. The
structural optimizations were carried out using gradient-based
techniques. Spin-unrestricted formalism was used to allow for
optimization of not only the interatomic distances, but also the
spin-defined electronic states.
For the CO adsorption studies, we consider the case of
55-atom clusters. Optimized icosahedral (ico, Ih), cubocta-
hedral (cubo, Ih), and Cs-symmetry conformations, which are
about 1 nm in diameter, were chosen as paradigmatic repre-
sentatives of Pt55 and its Mo-doped derivatives. The Cs con-
formation is generated by reflecting half of the ico structure
with respect to a plane that contains its central atom and is
perpendicular to its fivefold axis. The cubo conformation is
consistent with the FCC lattice structure of bulk Pt, but neither
the ico, the cubo, nor the Cs conformation is necessarily the
energetically most preferred form of Pt55 (the imposed sym-
metries result in a significant speed-up of the computations).
The average energy of CO adsorption on the clusters
was computed as
EAds ¼ ½ECluster þ pECO  ETotal =p
where ECluster, ECO and ETotal are the equilibrium energies
of the bare cluster, the CO molecule, and the CO-covered
cluster, and p is the number of the adsorbed CO molecules.
123
 Top Catal

To analyze the energetics of the catalytic reactions of Table 1 Turnover rate (normalized to Pt surface area determined by
interest (removal of a hydrogen atom or a hydroxyl group) CO chemisorption) comparisons for Pt/C and PtMo/C catalysts
including glycerol consumption and gas phase production rates
from glycerol or its fragments, we considered Pt7, Pt5Mo2
and Pt3Mo4, each in its corresponding lowest energy con- Conversion (%) TOR/10-3 s-1 CO2/CH4
figuration (J. Jellinek, A. Sumer, in preparation). All pos- Glycerol H2 CO2 CH4
sible adsorption conformations were obtained on each of
these nano-catalysts, and the complete minimum energy Pt/C 0.4 9.3 6.7 3.7 0.5 7.4
paths for various possible dehydrogenation and dehydr- 6 5.8 17 5.9 – –
oxylation reactions were mapped out for the three lowest 12 3.7 9.9 5.0 1.6 3.0
energy catalyst–reactant complexes. This was done by 14 2.3 5.2 5.6 1.5 3.7
performing constrained energy minimizations on grids of PtMo/C 0.3 17 6.1 6.6 0.7 9.7
C–H, O–H, and C–OH distances. Further refinement of the 5 37 47 41 2.3 –
transition states was obtained by using the structures of the 8 41 64 31 4.4 8.7
highest energy configuration along the grid-based mini- 21 17 25 26 4.6 5.6
mum energy paths as inputs for the so-called mode-fol- 26 5.1 4.6 8.3 1.3 6.6
lowing procedure as implemented in NWChem [16]. The Data collected at 230 °C and 30 bar with varying WHSV
transition states were finally verified by normal mode
analysis and through their connectivity to the reactants and
products of interest. Although the size of the nano-catalysts times higher rate than the Pt/C at a conversion of 5 %.
considered in computations was smaller than those Below 1 % conversion, the rates of hydrogen generation
explored experimentally, it allowed for elucidation of the are similar within the range of the error of measurement. At
global and generic qualitative effects of the nano-catalyst conversions above 1 %, PtMo/C had higher rates than Pt/C.
composition (pure Pt vs. mixed Pt–Mo). PtMo/C also displayed an overall increase in glycerol
conversion rate at all studied conversions. At conversions
up to 40 %, CO2 was preferentially produced over CH4
3 Results (Fig. 1), with CO2/CH4 ratios (a measure of pathway
selectivity) between 3 and 10 (Table 1), indicating a gen-
3.1 Kinetic Characterization eral selectivity to the hydrogen formation pathway. At
conversions below 5 %, selectivity to CO2 and CH4 were
Conversion of glycerol in the aqueous phase was studied nearly equal between Pt/C and PtMo/C. At conversions
over both PtMo/C and Pt/C catalysts. Weight hourly space higher than 5 %, Pt/C showed higher selectivity to gas
velocity (WHSV) was varied to investigate the product phase molecules compared to PtMo/C (5 % higher CO2
distributions as a function of conversion and to probe major
reaction pathways. As in a previous study of the PtMo/C
catalyst [10], the products observed were consistent with
both C–OH and C–C bond cleavage reactions. Carbon–
carbon cleavage reactions of oxygenated hydrocarbons
result in the formation of CO2 and H2, the desired products
for this reaction, whereas carbon–oxygen bond cleavage
results in the formation of liquid phase species with satu-
rated hydrocarbon fragments (i.e. R-CH3 type compounds)
which require hydrogen to form. Thus, the formation of
products resulting from C–C cleavage reactions leads to the
generation of H2, whereas products resulting from C–OH
cleavage reduce hydrogen yields and represent an unde-
sirable side reaction ultimately leading to the formation of
light hydrocarbons.
The major gas phase products identified were H2, CO2,
and CH4. CO was not a major product and was generally
only observed in trace quantities; when it was present,
Fig. 1 Gas phase carbon selectivity to carbon dioxide (diamond) and
CO2:CO ratios exceeded 100:1. A comparison of turnover
methane (square) as a function of conversion comparing the Pt/C
rates for hydrogen at similar conversions (Table 1) indi- catalyst (blue, hollow markers, solid line) to the PtMo/C catalyst (red,
cates that the PtMo/C catalyst produces hydrogen at a 4 filled markers, dashed line)

123
Top Catal
Table 2 Carbon distribution of [C–C] and [C–O] products in the
liquid phase effluent as a function of conversion
Conversion (%) [C–C] (%) [C–O] (%)
Pt/C 0.4 34 66
6 21 79
12 18 82
14 19 81
40 25 75
PtMo/C 0.3 9 91
5 18 82
8 18 82
21 8 92
26 8 92
Data collected at 230 °C and 30 bar with varying WHSV
selectivity and 2 % higher CH4 selectivity). Despite the
differences in selectivity, both Pt/C and PtMo/C had sim-
ilar CO2/CH4 ratios at the reported conversions (Table 1).
A significant portion of reacted carbon was contained in
the liquid phase effluent, from 90 % at low conversions to
50 % at intermediate conversions (the remaining carbon
was converted to CO2 and CH4). Products identified as
liquid phase intermediates include propylene glycol, eth-
ylene glycol, hydroxyacetone, ethanol, 1- and 2-propanol,
methanol, propionaldehyde, and acetaldehyde. Liquid
phase products were grouped by the type of bond cleavage.
Those that undergo solely carbon–carbon bond cleavage
(ethylene glycol, methanol) were classified as C–C scission
intermediates ([C–C]). Products resulting from a mixed
pathway of C–OH and C–C bond cleavage, or products
that have been deoxygenated (propylene glycol, hydrox-
yacetone, 1- and 2-propanol, ethanol, acetaldehyde, pro-
pionaldehyde), were designated as C–OH scission
intermediates ([C–O]). The products in the liquid phase
were generally dominated by [C–O] products for both Pt
(60–80 %) and PtMo (80–90 %) catalysts (Table 2). There
are also differences in the liquid selectivity profiles for the
Pt vs. PtMo catalyst. As Fig. 2 demonstrates, the Pt catalyst
had lower selectivity to [C–O] scission products and gen-
erally higher selectivity to [C–C] scission products when
compared with the PtMo catalyst at all tested conversions.
This indicates that the addition of Mo shifts selectivity
towards deoxygenation.
3.2 Catalyst Characterization
3.2.1 TEM
Transmission electron micrographs for the Pt/C catalyst
were obtained before and after 6 days under reaction
(Fig. 3). Analysis of the images indicates the average size
Fig. 2 Overall carbon selectivity to liquid phase products comparing
selectivity to [C-C] products (squares) and [C-O] products (dia-
monds) as a function of conversion for Pt/C (blue, hollow markers,
solid line) and PtMo/C (red, filled markers, dashed line) catalysts
of the fresh catalyst was 1.8 ± 0.6 nm, with 11 % of the
particles larger than 3 nm. After 6 days on stream under
glycerol APR conditions, there is an increase in size of the
nanoparticles (Fig. 3c, d and in the histogram Fig. 4). The
average size and distribution of the used catalyst is much
larger than in the fresh sample (5.2 ± 4.2 nm), indicating
that the nanoparticles sinter under the reaction conditions.
These particles are of similar average size to those that
were observed in the characterization of the PtMo/C cata-
lyst (2.0 ± 0.5 nm fresh, 5.1 ± 3.1 nm used [10]).
3.2.2 XAS of Reduced Catalyst
The Pt L3 XANES of Pt foil was compared to that of the Pt/C
catalyst. Prior to reduction, the catalyst had small amounts of
oxidized Pt, as expected from a passivated Pt particle. Pt
reduces completely to the metallic state (Pt0) upon H2
reduction at 300 °C. The edge energy, which is determined
from the maximum in the first peak of the 1st derivative of the
XANES, suggests no significant edge position shift between
Pt/C catalyst and Pt foil. The white line intensity above the
edge of XANES on Pt/C is lower than that of Pt foil (Fig. 5),
indicating complete reduction of the Pt/C catalyst and nano-
particles smaller than about 3 nm [17, 18].
Additional information on the Pt particle sizes were
determined by EXAFS spectroscopy at room temperature
after the catalysts were reduced in 3.5 % H2/He at 300 °C for
30 min. The fit of the Pt–Pt first-shell EXAFS spectra were
obtained by a Fourier transform of the k2-weighted data and
are summarized in Table 3. Figure 6 shows the magnitude of
the Fourier transform (uncorrected phase and amplitude) for
representative data of the Pt/C catalyst and Pt foil. The Pt/C
catalyst has a lower peak intensity in the first shell Pt–Pt
123
 Top Catal

Fig. 3 Transmission electron

micrographs of the Pt/C catalyst
for samples of fresh (a, b) and
used for 6 days on stream (c,
d) catalysts, and representative
images of the PtMo/C catalyst
fresh (e) and used (f)

distance (between 1.7 and 3.1 Å) compared to the foil. The
FT peak size is proportional to the coordination number [17]
and can be used as a measure of particle size, thus allowing
for the observation of relative particle size during reaction. In
addition, there is a shift to lower R indicating a decrease in
the Pt–Pt bond distance for the Pt/C catalyst, a characteristic
of nanoparticles less than about 3 nm [18]. The fit results
show that for the reduced catalyst there is a contraction in the
Pt bond distances (2.71 Å) compared to bulk Pt (2.77 Å),
with a Pt–Pt coordination number of 6.1. Assuming spherical
nanoparticles [19, 20], the average particle size was esti-
mated as 1.7 nm for the Pt/C catalyst, in reasonable agree-
ment with TEM results. These estimates were based on
previous correlations of the NPt–Pt with dispersion [17].

123
Top Catal
Fig. 4 Histogram showing changes in particle size distribution for Pt/
C catalysts before and after reaction; fresh (1.8 ± 0.6 nm, red) and
used (5.2 ± 4.2 nm, black, hatched)
Fig. 5 Pt L3 absorption near edge spectra from 11.54 to 11.59 keV
comparing the Pt nanoparticles on the reduced Pt/C catalyst (red,
dashed line) to a Pt foil (black)
3.2.3 XANES of Catalysts After Adsorption of CO, H2
and H2O
Reactant and product chemisorption cause changes in the
Pt L3 XANES spectra and can be used to determine
adsorbed surface species under reaction conditions [21].
Because the L3 XANES is a 2p to 5d (valence level)
transition, the edge energy and shape are sensitive to
changes in the Pt d-orbital electron density due to bond
formation. More specifically, chemisorption of gases, e.g.
CO, H2 and H2O, result in different changes to the shape of
the Pt L3 XANES [22]. Thus, changes induced by adsorbed
species can be used to determine their identity. In addition,
the intensity of these changes are linearly dependent on the
amount of the adsorbed species [21]. The advantage of
XANES spectroscopy is that it is specific only to the
amount of gas adsorbed on Pt and is independent of any
adsorbates on the substrate or other components in the
catalyst. Thus, the relative coverage of gases (i.e. CO and
H2) on the Pt surface under reaction conditions can be
determined from the XANES spectra.
The Pt L3 edge XANES spectra of Pt/C with and without
CO and H2 adsorption (ambient conditions) and water
adsorption (230 °C and 26 bar) are shown in Fig. 7. Using
CO adsorption as a representative example, the white line
intensity of the XANES increased after adsorption, and the
edge shifts to higher energy (*0.5 eV shift). If one sub-
tracts the XANES in He from that with adsorbed CO, the
difference (or DXANES) isolates changes in shape and
intensity of the edge due to CO adsorption. Figure 8 shows
the DXANES spectra for each of the adsorbate standards
on Pt/C. Each of the adsorbed species has a distinct shift in
the edge position and XANES shape compared to the
spectra in He.
3.2.4 XAS of Catalysts Under Reaction Condition
(Operando XAS)
Operando XAS studies were carried out to study the
structure of the catalyst during aqueous phase glycerol
reforming at 26 bar and 230 °C at different space velocities
in order to achieve low and high glycerol conversions.
Figure 6 compares the magnitude of the Fourier transform
of the Pt L3 edge EXAFS spectra for Pt/C catalyst during
glycerol reforming at high and low conversion. The mag-
nitudes of the FT are smaller than a Pt foil and similar in
size to that of the reduced catalyst. However, the former
were obtained at reaction temperature which lowers the
intensity of the magnitude of the Fourier transform. The fit
parameters are summarized in Table 3. The fits suggest Pt–
Pt contributions with coordination numbers around 8.2 and
an interatomic distance of 2.72 Å for both high and low
glycerol conversion. The particle size is estimated at about
3 nm [17], indicating a slight increase in particle size
during the first few hours under reaction.
The Pt L3 XANES spectra of the Pt/C catalyst reduced
and under glycerol reforming reaction at high and low
conversion are shown in Fig. 9. The XANES under reac-
tion shares very similar spectra shape and edge position for
both high and low conversions, and both show a positive
*0.2 eV edge position shift relative to the spectra taken in
helium after reduction. The XANES are characteristic of
metallic Pt under reaction conditions. Figure 10 shows the
DXANES of Pt/C catalyst under glycerol reforming con-
ditions at high and low conversion along with that for CO
saturation at RT (shown in Fig. 8). The DXANES spectra
123
 Top Catal

Table 3 EXAFS fit parameters for Pt/C catalysts measured ex situ and in operando
Sample TreatmentScan conditions Edge energy (keV) Pt–Pt N Pt–Pt R (Å) Dr2 (Å2) E0 (eV)

Pt foil As is 11.5640 12 2.77 0.000 -1.1

Pt/C Reduced in H2, RT in He 11.5640 6.1 2.71 0.002 -3.4
Pt/C Reduced in H2, glycerol reforming at 230 °C
High conversion 11.5642 8.2 2.72 0.005 -2.5
Low conversion 8.2 2.73 0.005 -2.9

Fig. 6 Pt L3 Magnitude of the k2-weighted, uncorrected phase and
amplitude Fourier transform (Dk = 2.7–12.2 Å-1) of the EXAFS for
the Pt/C catalyst under different conditions. Pt foil (black, solid line),
Pt/C (red, dashed line), Pt/C with high glycerol conversion (green,
dotted line), Pt/C with low glycerol conversion (blue, dash-dot line)
Fig. 8 DXANES from 11.54 to 11.59 keV showing the isolated
changes to the XANES caused by CO (red, dashed line), H2 (blue,
dotted line) and H2O (green, dash-dot line, at 230 °C, 26 bar)
adsorption; all spectra recorded at ambient conditions, except where
noted

at both glycerol conversions are very similar in position

and shape to that for Pt/C saturated with CO at RT for Pt/C
(Fig. 8), and distinct from that of either H2 or H2O. Thus,
although CO is observed in trace amounts in the gas phase
effluent, the XANES spectra indicate it is the most abun-
dant surface species under reforming conditions.

3.2.5 DFT Computational Results

Density functional theory calculations were performed on

Fig. 7 Pt L3 absorption near edge spectra from 11.54 to 11.59 keV
showing the effects of gas adsorption on the XANES for Pt/C reduced
at 300 °C. Pt/C in He (black, solid line) and with adsorbed CO (red,
dashed line), H2 (blue, dotted line), and H2O (green, dot-dash line, at
230 °C, 26 bar); all spectra recorded ambient conditions, except
where noted
the Pt/C and PtMo/C catalyst. The latter was described in
previous work [10] and demonstrated the effects of alloy-
ing on the electronic properties of the nanoparticles. For a
monometallic Pt system, three conformations of Pt55 are
shown in Fig. 11. Their computed binding energies are
245.66, 245.35, and 244.40 eV with preferred spin states of
6, 6, and 5 for the Cs, Ih and Oh conformations. Figure 12
shows three possible topologically different arrangements
for adsorption of two CO molecules on the energetically
most stable (of the three considered) Cs Pt55; the two CO
molecules are added in a manner that preserves the Cs
symmetry. The computations indicate that atop adsorption

123
Top Catal
Fig. 9 Pt L3 XANES spectra from 11.54 to 11.59 keV of reduced Pt/
C catalyst in He (black, solid line) under low (green, dotted line) and
high (blue, dash-dot line) conversions compared to the Pt/C catalyst
Fig. 10 Pt L3 DXANES spectra from 11.55 to 11.59 keV for the Pt/C
catalysts under operando conditions at low (green, dotted line) and
high (blue, dash-dot line) conversions; CO saturated at RT (red,
dashed line)
is the energetically most favored configuration (Ead-
s = 2.62 eV), followed by over an edge (Eads = 2.20 eV),
and then over a face (Eads = 1.92 eV). The same trend was
observed for Pt clusters of other sizes and structures (J.
Jellinek, A. Sumer, in preparation). The atop arrangement
has been found experimentally to be the strongest for CO
adsorption on small Pt clusters [23] and the (111) face of
the bulk Pt [24].
The strength of CO adsorption was also used as a
descriptor to analyze the effects of admixing Mo to the Pt in
metallic nanoparticles. The studies were performed for
various sizes, structures and composition of pure Pt and
mixed Pt/Mo clusters (J. Jellinek, A. Sumer, in preparation).
Figure 13 displays the results for the cases of the ico and
cubo conformations of the pure Pt55 and mixed Pt43Mo12.
The stoichiometry of the mixed clusters was chosen to
allow for preserving the Ih and Oh symmetries of the cor-
responding parent Pt55. Both symmetries allow for two
types of replacements of 12 Pt atoms by Mo atoms, which
result in different so-called homotopic forms; the latter
differ by the occupation of the sites of a given geometric
structure by different types of atoms at a fixed elemental and
percentile composition [10, 25–27]. The first places the Mo
atoms in the interior of the clusters (Mo forms a 12-atom
shell around the central Pt atom), and the second places the
12 Mo atoms in symmetry-consistent sites on the surface of
the cluster. The homotops with Mo in the interior are more
stable than their counterparts with Mo in the surface [10] (J.
Jellinek, A. Sumer, in preparation). This apparent incon-
sistency with our experimental results was discussed in
detail in our previous work [10].
In each case, 30 CO molecules have been added to the
same 30 surface sites of the ico and cubo conformations.
Inspection of the CO adsorption energies (Fig. 13) indi-
cates that the presence of Mo in the 55-atom Pt-based
nanoparticles leads to a significant reduction in the strength
of adsorption of CO to the Pt sites, regardless of the
location of the Mo. However, the degree of reduction is
larger in the case of homotops with Mo in the interior. The
same qualitative conclusions follow from the results
obtained for Pt and PtMo clusters of other sizes, structures,
and percentile compositions (J. Jellinek, A. Sumer, in
preparation). Our analysis reveals that the fundamental
reason for the general reduction in the CO binding energy
is lowering of the density of electronic states in the vicinity
of the ‘‘Fermi level’’ (i.e. the HOMO) caused by admixing
Mo to Pt [the details will be discussed elsewhere (J. Jel-
linek, A. Sumer, in preparation)].
The computations also clarify the effects of Mo on the
reaction energetics of dehydrogenation versus dehydroxy-
lation, with dehydrogenation being a representative reaction
for hydrogen generation, as it has been shown that signifi-
cant dehydrogenation is required for the desired C–C bond
cleavage to become favorable [28]. Figure 14 displays the
relative energies of the transition states (activation barriers)
and the products for the dehydrogenation and dehydroxy-
lation reactions of C3O3H7 on Pt7, Pt5Mo2 and Pt3Mo4
nanocatalysts (C3O3H7 was chosen as a prototypical inter-
mediate in glycerol reforming); in each case, the energy of
the reactant C3O3H7-nanocatalyst complex is used as the
reference zero level. As the Mo content of the catalyst
increases, the barriers for both dehydrogenation and dehy-
droxylation decrease. The degree of the barrier reduction
for dehydroxylation is larger than that for dehydrogenation.
These results help to explain the mechanisms through which
Mo changes the activity of Pt nanoparticles by both
123

Fig. 11 Three symmetric
conformations and binding
energies (BE) of Pt55 (see the
text for details)
BE (eV/atom): 4.47
Fig. 12 Configurations and
energies of atop, over bridge,
and over face CO adsorption on
the Cs conformation of Pt55
(orange circle Pt, green circle C
and red circle O)
EAds (eV/CO): 2.62
mitigation of the CO poisoning and decrease in the acti-
vation energy for dehydrogenation and dehydroxylation.
They also indicate that one may expect lower H2 selectivity
on the Mo containing catalysts, which is a consequence of
the larger degree of reduction in the barrier for dehydr-
oxylation than for dehydrogenation.
As a final remark, the computational results presented
above are obtained for gas phase nanoparticles, and
therefore they do not incorporate support or solvent effects.
The analysis of the consequences of Mo-doping is, how-
ever, general and robust, and it helps explain the experi-
mentally observed improvement in catalytic activity of
Mo-doped metallic Pt nanoparticles compared to their
mono-metallic Pt.
4 Discussion
Spectra of the Pt/C sample before and after adsorption of
CO (Fig. 7) shows that the presence of CO influences the
Pt L3 XANES features. Since Pt L3 XANES intensity and
position are dependent on the d orbital occupancy, these
changes indicate that CO is chemically bonded. The white
line intensity increase can be explained as a change in
partial density of the d state and the d-band due to a
charge transfer event, i.e. the partial removal of d electron
density from the metal to the p anti-bonding orbital of
CO [29].
123
Top Catal
Cs Ih Oh
4.46 4.44
Atop Over Edge Over Face
2.20 1.92
During the first few hours under reforming conditions,
the EXAFS indicates the Pt nanoparticles remains metal-
lic, but increases in size from 1.7 nm (ca. 0.6 dispersion)
to about 3 nm (0.3 dispersion) Pt particles. To determine
the likely adsorbed species on the catalyst metal particles
under reaction conditions, the DXANES spectra of the
sample was compared with the DXANES spectra of the
same sample with adsorbed CO, H2, or H2O. Comparing
the DXANES of the Pt/C catalyst under glycerol reform-
ing reaction with that of CO adsorption at RT (Fig. 9), it
is observed that the shapes of the DXANES are very
similar, indicating that the major surface species during
reaction is CO, likely a result of the hydrogen generation
pathway described in our previous work [10]. The
DXANES under reaction is about 0.3 that of CO adsorp-
tion at RT, and nearly identical for both conversions.
Including the DXANES of H2 and H2O in the fits did not
improve the quality of the fit, indicating the coverage of
these species are too low to be reliably estimated. Since
the Pt under reaction is a different size than that of the
fresh catalyst, in order to determine the true CO surface
coverage, one must account for the difference in the
number of surface atoms. With the increase in size alone,
the DXANES would be lower than that in the fresh cat-
alyst. Taking the ratio of the dispersions (0.6/0.3), the
number of surface atoms in the reduced catalyst is twice
that under glycerol reforming. Thus, the true CO surface
coverage is about 0.6.
Top Catal

Fig. 13 Configurations and Ico Cubo

energies of CO adsorption on 30
identical surface Pt sites for,
respectively, ico and cubo
conformations of Pt55,
Pt43Mo12(core), and
Pt43Mo12(surface) (orange circle Pt 55
Pt, blue circle Mo, green circle
C and red circle O)

EAds (eV/CO): 2.14 2.11

Pt 43Mo12(core)

EAds (eV/CO): 1.52 1.72

Pt 43 Mo12(surface)

EAds (eV/CO): 1.72 1.80

Fig. 14 Energetics of the most

facile H and OH removal from
C3O3H7 on Pt7, Pt5Mo2 and
Pt3Mo4. R, TS and P stand for
reactant, transition state and
product, respectively; the
energies of the reactant
C3O3H7-nanocatalyst
complexes are chosen as the
reference zero level of energy

While the DXANES spectrum of Pt/C indicates that CO that any CO present was below the detection limit. This
is the predominant surface species, similar analysis on a may be explained by the structure of the PtMo/C catalyst in
PtMo/C catalyst during reforming [10] did not show a which the fractional surface coverage of Pt is about 0.25
measurable contribution from CO adsorption, suggesting versus a similar sized monometallic Pt particle. In addition,

123

the 2 nm bimetallic PtMo nanoparticles rapidly sinter to
about 5 nm under reforming conditions, further reducing
the Pt dispersion. Density functional theory also indicates
that CO is more weakly bound to Pt in the bimetallic PtMo/
C catalyst. Thus, because of the lower dispersion, low Pt
surface composition, and weaker CO binding (leading to
higher rates of water gas shift, see below), changes to the
XANES spectrum by CO adsorption on PtMo catalysts
were not observed.
The addition of Mo to a Pt catalyst causes changes in
both rate and selectivity of the glycerol reforming reaction.
In the PtMo particles, the Mo alloys with the Pt, changing
both the structure and electronic properties of the nano-
particles. In particular, DFT calculations indicate that the
addition of Mo to a Pt system decreases the binding energy
for CO on the surface Pt of the catalyst. These differences
in CO binding energy help explain the observed activity
differences between the two catalysts. First, a lower CO
binding energy suggests that it is more difficult for the CO
to stick on the PtMo surface, freeing up reaction sites.
However, perhaps more importantly, the gas phase effluent
from both catalysts contains negligible amounts of CO
(CO2/CO [ 100, when observable). Thus, it is unlikely that
the decrease in binding energy is simply causing additional
CO to desorb from the PtMo surface and exit the reactor in
the gas phase, leading to the negligible coverage observed.
Rather, we note that CO binding energy is often used as a
rate descriptor for the water gas shift (WGS) reaction [30].
Thus, a decrease in CO binding energy leads to an increase
in the rate of the water gas shift reaction that converts the
surface CO to CO2, removing rate inhibiting CO and
freeing up surface sites for reaction. In-house water gas
shift experiments on supported Pt and PtMo catalysts
confirm that the PtMo has a higher gas phase WGS turn-
over rate than Pt [31].
While the PtMo/C displayed increased glycerol and H2
TORs compared to Pt/C, the selectivity to intermediate
liquid products must also be considered. While the selec-
tivity ratios in the gas phase, particularly CO2/CH4 ratios,
were similar between Pt and PtMo at the reported con-
versions, the liquid phase carbon distribution was different.
The Pt/C catalyst had a higher selectivity to [C–C] type
cleavage products, whereas the PtMo/C catalyst had a
higher selectivity to [C–O] cleavage products. As these
products are completely converted into gas phase products
as conversion approaches 100 %, the PtMo/C catalyst will
have a higher selectivity to light alkane products, which
will lower the overall hydrogen yield. In addition, the
propensity of the PtMo/C catalyst to catalyze deoxygen-
ation reactions suggests that at higher conversions the
differences in liquid phase selectivity will increase.
The two catalysts were compared at glycerol conversion
below 1 % (0.3 % for PtMo and 0.4 % for Pt) to determine
123
Top Catal
the initial reaction products and study the relative selec-
tivity of C–C versus C–OH bond cleavage (Fig. 15). At this
conversion level, significant differences between the Pt and
PtMo catalysts are observed. Considering only the 5 most
abundant species (which at this conversion make up 92 and
96 % of the reacted carbon), the PtMo has a significantly
higher selectivity to [C–O] cleavage compared to Pt/C (79
vs. 57 %), whereas the Pt/C catalyst has a much higher
selectivity to the [C–C] cleavage compounds (29 vs. 8 %).
The gas phase products observed below 1 % conversion are
predominantly CO2 (and H2). Under these conditions, the
selectivity to CO2 and CH4 are equal (CO2: 12 vs. 12 %;
CH4: 2 vs. 1 %; for PtMo compared to Pt), resulting in
comparable CO2/CH4 ratios. At this conversion level, the
differences between the two catalysts are not apparent by
only measuring the gas products. However, from an anal-
ysis of the liquid products, the differences in catalyst
selectivity are clearly distinguishable. As the conversion
increases, the selectivity to gas products on Pt/C becomes
higher than for PtMo/C, while maintaining CO2/CH4 ratios
between 3 and 6, which suggests that the Pt/C catalyst
better catalyzes reactions that produce H2 (i.e. C–C bond
cleavage) versus those that yield light alkanes (i.e. C–O
bond cleavage).
The underlying cause for this increase in [C–O] cleav-
age rates is the result of adding Mo to the Pt nanoparticles.
There are two proposed mechanisms for this cleavage
reaction. The first is via an acid catalyzed dehydration of a
C–OH group by MoOx clusters on the carbon support (as
proposed in [10], followed by subsequent hydrogenation.
Supported MoO3 has been demonstrated to have weakly
acidic functionality [32, 33], but experiments on MoOx/C
have shown no activity for loss of C–O bonds of biomass
compounds [34, 35]. Koso et al. [35] have reported activity
for C–O bond cleavage on Rh–Mo, and proposed that Mo
sits on the surface of a Rh particle. It was also suggested
that the Mo and Rh have a synergistic effect in which the
oxygenated group adsorbs to the Mo, and the hydrogen-
olysis occurs with a nearby Rh. The selectivity for the
PtMo catalyst described here are similar to those on RhMo
catalysts with similar surface configurations, suggesting
that the reaction mechanisms for deoxygenation may be
similar. It has also been reported that oxygen binding
energy on the PtMo species is significantly higher than on
the pure Pt species [34], which is consistent with the
mechanism proposed by Koso. Following C–O cleavage,
the deoxygenated species could then be either desorbed
(likely as the aldehyde or ketone, as keto-enol tautomer-
ization generally favors the keto form, i.e. hydroxyacetone)
or be hydrogenated by the Pt to the corresponding alcohol
(i.e. propylene glycol). Selectivity results indicate that as
hydrogen partial pressures in the reactor increase, partic-
ularly at higher conversions, the favored products are the
Top Catal
Fig. 15 Comparison of carbon selectivity for various classes of
compounds for Pt/C (blue, solid bars) and PtMo/C (red, hatched bars)
catalysts at 230 °C, 30 bar with 30 wt% glycerol at conversions of
conversions of 0.4 ± 0.1 % (Pt/C) and 0.3 ± 0.02 % (PtMo/C)
hydrogenated species. The keto-species made up a signif-
icant portion of the liquid products at conversion levels
below 1 % (24 and 37 % selectivity for Pt and PtMo), but
decreased significantly as conversion increased (2 %
selectivity at conversion above 20 % for both Pt and
PtMo).
An alternative explanation for the increased C–O
cleavage selectivity is the Mo modifies the electronic
properties of the Pt active site. While gas and liquid
product distribution of PtMo/C differs slightly from that of
Pt/C, the catalysts operate at similar reaction temperature,
and have comparable TORs and selectivity. While the TOR
and selectivity of Mo is not known, it would be expected to
be very different from that of Pt. Thus, in the bimetallic
PtMo where Mo occupies 0.75 of the surface, if the Mo
TOR were near that of Pt, the gas and liquid products
would be expected to be significantly different from that of
Pt. The experimental results indicate that the CO2(H2):CH4
ratio, liquid products and TOR are more similar than fun-
damentally different. The changes in rate and selectivity
suggest that Pt is the active site and the Mo induces
changes in the Pt in the bimetallic PtMo/C catalyst. It is
suggested that the role of the Mo, therefore, is to modify
the energy of the Pt-adsorbate bond energy and activation
energies, altering its chemistry rather than to directly par-
ticipate in the reaction. Specifically, Mo appears to pref-
erentially lower the C–O bond cleavage pathway leading to
lower H2 yield (and higher CH4 yield).
The addition of Mo has both positive and negative
effects for hydrogen generation catalysts. The DFT results
presented here and in previous work [10] provide a blue-
print for improving future catalysts. The DFT results
indicate that the addition of Mo to the particle, even when
it preferentially occupies surface sites, helps to drive down
the Pt–CO binding energy increasing the catalytic activity.
DFT calculations also show that for a PtMo morphology
where the Mo segregates to the center of the particle giving
a bimetallic particle with surface Pt only, the Pt–CO
binding energy decreases even further. With the increase in
the fraction of surface Pt and the much lower Pt–CO
binding energy, this would be expected to significantly
increase the rate over these catalysts. The effect of this
alternate configuration on the selectivity is not known. If
the Mo acts as an oxophilic center as proposed by Koso,
then non-surface Mo should promote both the rate and
display selectivities similar to that of monometallic Pt.
However, if the Mo is an electronic Pt promoter, then
subsurface Mo might still display high C–O bond cleavage
selectivity and low H2 yield.
5 Conclusions
The TORs and gas and liquid product selectivity for
aqueous phase reforming of glycerol on Pt/C and PtMo/C
have been determined. The TORs of PtMo/C are approxi-
mately 4 times higher than that of Pt/C at comparable
conversions. While at low conversions the selectivity for
H2/CO2 and CH4 are similar for both Pt and PtMo catalysts,
the partially converted liquid products indicate a higher
selectivity to C–O bond cleavage products, i.e. those with
saturated hydrocarbon fragments, on PtMo/C catalysts. At
higher conversions these fragments form methane, sug-
gesting that PtMo/C will have lower H2 selectivity as
conversion approaches 100 %.
Operando EXAFS spectroscopy indicates that the 2 nm
Pt/C nanoparticles are metallic under reaction conditions,
and changes in the XANES indicate that CO is the most
abundant surface species despite CO2:CO ratios above 100
in the reaction effluent. The CO surface coverage for Pt/C
is estimated to be about 0.6.
DFT calculations indicate that metallic Mo lowers the
Pt–CO binding energy and also lowers the activation
energies for dehydrogenation and C–O bond cleavage. The
latter is more strongly affected leading to an increase in the
C–O bond cleavage selectivity and ultimately lower H2
yields, consistent with experimental results.
Finally, the product selectivity and DFT calculations
suggest that the surface Pt atoms are the active site and the
role of the Mo is to modify the electronic properties of the
Pt-adsorbate bond energies, which lead to the observed
changes in TOR and selectivity.
Acknowledgments This material is based upon work supported as
part of the Institute for Atom-efficient Chemical Transformations
(IACT), an Energy Frontier Research Center funded by the U.S.
Department of Energy, Office of Science, Office of Basic Energy
123

Sciences. Use of the Advanced Photon Source is supported by the
U. S. Department of Energy, Office of Science, and Office of Basic
Energy Sciences, under Contract DE-AC02-06CH11357. MRCAT
operations are supported by the Department of Energy and the
MRCAT member institutions. J.J. was also supported by the Office of
Basic Energy Sciences, Division of Chemical Sciences, Geosciences
and Biosciences, U.S. Department of Energy under Contract DE-
AC02-06CH11357. This research used the resources of the National
Energy Research Scientific Computing Center (NERSC), which is
supported by the Office of Science of the U.S. Department of Energy
under Contract DE-AC02-05CH11231 and of the Laboratory Com-
puting Resource Center (Fusion/LCRC) at Argonne National Labo-
ratory. The authors would like to thank M. Cem Akatay for his
assistance with the TEM images.
References
1. Huber GW, Iborra S, Corma A (2006) Synthesis of transportation
fuels from biomass: chemistry. Catal Eng Chem Rev 106:4044–4098
2. Bridgwater AV, Peacocke GVC (2000) Fast pyrolysis processes
for biomass. Renew Sustain Energy Rev 4:1–73
3. Cortright RD, Davda RR, Dumesic JA (2002) Hydrogen from
catalytic reforming of biomass-derived hydrocarbons in liquid
water. Nature 418:964–967
4. Luo N, Zhao X, Cao F, Xiao T, Fang D (2007) Thermodynamic
study on hydrogen generation from different glycerol reforming
processes. Energy Fuels 21:3505–3512
5. Kunkes EL, Simonetti DA, Dumesic JA, Pyrz WD, Murillo LE,
Chen JG, Buttrey DJ (2008) The role of rhenium in the conver-
sion of glycerol to synthesis gas over carbon supported platinum-
rhenium catalysts. J Catal 260:164–177
6. Kunkes EL, Soares RR, Simonetti DA, Dumesic JA (2009) An
integrated catalytic approach for the production of hydrogen by
glycerol reforming coupled with water-gas shift. Appl Catal B
90:693–698
7. Kunkes EL, Simonetti DA, West RM, Serrano-Ruiz JC, Gärtner
CA, Dumesic JA (2008) Catalytic conversion of biomass to
monofunctional hydrocarbons and targeted liquid-fuel classes.
Science 322:417–421
8. Huber GW, Shabaker JW, Evans ST, Dumesic JA (2006)
Aqueous-phase reforming of ethylene glycol over supported Pt
and Pd bimetallic catalysts. Appl Catal B 62:226–235
9. Soares RR, Simonetti DA, Dumesic JA (2006) Glycerol as a
source for fuels and chemicals by low-temperature catalytic
processing. Angew Chem Int Ed 45:3982–3985
10. Dietrich P, Lobo-Lapidus R, Wu T, Sumer A, Akatay M, Fin-
gland B, Guo N, Dumesic J, Marshall C, Stach E, Jellinek J,
Delgass W, Ribeiro F, Miller J (2012) Aqueous phase glycerol
reforming by PtMo bimetallic nano-particle catalyst: product
selectivity and structural characterization. Top Catal 55:53–69
11. Fingland B, Ribeiro F, Miller J (2009) Simultaneous measure-
ment of X-ray absorption spectra and kinetics: a fixed-bed, plug-
flow operando reactor. Catal Lett 131:1–6
12. Kispersky VF, Kropf AJ, Ribeiro FH, Miller JT (2012) Low
absorption vitreous carbon reactors for operando XAS: a case
study on Cu/Zeolites for selective catalytic reduction of NOx by
NH3. Phys Chem Chem Phys 14:2229–2238
13. Perdew JP, Wang Y (1992) Accurate and simple analytic repre-
sentation of the electron–gas correlation energy. Phys Rev B
45:13244
14. Perdew JP, Chevary JA, Vosko SH, Jackson KA, Pederson MR,
Singh DJ, Fiolhais C (1992) Atoms, molecules, solids, and sur-
faces: applications of the generalized gradient approximation for
exchange and correlation. Phys Rev B 46:6671
123
Top Catal
15. Schuchardt KL, Didier BT, Elsethagen T, Sun L, Gurumoorthi V,
Chase J, Li J, Windus TL (2007) Basis set exchange: a com-
munity database for computational sciences. J Chem Inf Model
47:1045–1052
16. Valiev M, Bylaska EJ, Govind N, Kowalski K, Straatsma TP,
Van Dam HJJ, Wang D, Nieplocha J, Apra E, Windus TL, de
Jong WA (2010) NWChem: a comprehensive and scalable open-
source solution for large scale molecular simulations. Comput
Phys Commun 181:1477–1489
17. Miller JT, Kropf AJ, Zha Y, Regalbuto JR, Delannoy L, Louis C,
Bus E, van Bokhoven JA (2006) The effect of gold particle size
on AuAu bond length and reactivity toward oxygen in supported
catalysts. J Catal 240:222–234
18. Lei Y, Jelic J, Nitsche L, Meyer R, Miller J (2011) The effect of
particle size and adsorbates on L3, L2 and L1 X-ray absorption
near-edge spectra of supported Pt nanoparticles. Top Catal 54:
334–348
19. de Graaf J, van Dillen AJ, de Jong KP, Koningsberger DC (2001)
Preparation of highly dispersed Pt particles in zeolite Y with a
narrow particle size distribution: characterization by hydrogen
chemisorption, TEM, EXAFS spectroscopy, and particle model-
ing. J Catal 203:307–321
20. Frenkel AI, Hills CW, Nuzzo RG (2001) A View from the Inside:
complexity in the atomic scale ordering of supported metal
nanoparticles. J Phys Chem B 105:12689–12703
21. Guo N, Fingland BR, Williams WD, Kispersky VF, Jelic J,
Delgass WN, Ribeiro FH, Meyer RJ, Miller JT (2010) Determi-
nation of CO, H2O and H2 coverage by XANES and EXAFS on
Pt and Au during water gas shift reaction. Phys Chem Chem Phys
12:5678–5693
22. Stakheev AY, Zhang Y, Ivanov AV, Baeva GN, Ramaker DE,
Koningsberger DC (2007) Separation of geometric and electronic
effects of the support on the CO and H2 chemisorption properties
of supported Pt particles: the effect of ionicity in modified alu-
mina supports. J Phys Chem C 111:3938–3948
23. Gruene P, Fielicke A, Meijer G, Rayner DM (2008) The
adsorption of CO on group 10 (Ni, Pd, Pt) transition-metal
clusters. Phys Chem Chem Phys 10:6144–6149
24. Rupprechter G, Dellwig T, Unterhalt H, Freund HJ (2001) High-
pressure carbon monoxide adsorption on Pt(111) revisited: a sum
frequency generation study. J Phys Chem B 105:3797–3802
25. Jellinek J, Krissinel EB (1996) NinAlm alloy clusters: analysis of
structural forms and their energy ordering. Chem Phys Lett 258:
283–292
26. Ferrando R, Jellinek J, Johnston RL (2008) Nanoalloys: from
theory to applications of alloy clusters and nanoparticles. Chem
Rev 108:845–910
27. Jellinek J (2008) Nanoalloys: tuning properties and characteris-
tics through size and composition. Faraday Discuss 138:11–35
28. Liu B, Greeley J (2011) Decomposition pathways of glycerol via
C–H, O–H, and C–C bond scission on Pt(111): a density func-
tional theory study. J Phys Chem C 115:19702–19709
29. Shriver D, Atkins P, Overton T, Rourke J (2009) Inorganic
chemistry. W. H. Freeman, New York
30. Schumacher N, Boisen A, Dahl S, Gokhale AA, Kandoi S,
Grabow LC, Dumesic JA, Mavrikakis M, Chorkendorff I (2005)
Trends in low-temperature water–gas shift reactivity on transition
metals. J Catal 229:265–275
31. Williams WD, Bollmann L, Miller JT, Delgass WN, Ribeiro FH
(2012) Effect of molybdenum addition on supported platinum cat-
alysts for the water–gas shift reaction. Appl Catal B 125:206–214
32. Nair H, Baertsch CD (2008) Method for quantifying redox site
densities in metal oxide catalysts: application to the comparison
of turnover frequencies for ethanol oxidative dehydrogenation
over alumina-supported VOx, MoOx, and WOx catalysts. J Catal
258:1–4
Top Catal
33. Nair H, Gatt JE, Miller JT, Baertsch CD (2011) Mechanistic
insights into the formation of acetaldehyde and diethyl ether from
ethanol over supported VOx, MoOx, and WOx catalysts. J Catal
279:144–154
34. Chia M, Pagán-Torres YJ, Hibbitts D, Tan Q, Pham HN, Datye
AK, Neurock M, Davis RJ, Dumesic JA (2011) Selective
hydrogenolysis of polyols and cyclic ethers over bifunctional
surface sites on rhodium–rhenium catalysts. J Am Chem Soc
133:12675–12689
35. Koso S, Ueda N, Shinmi Y, Okumura K, Kizuka T, Tomishige K
(2009) Promoting effect of Mo on the hydrogenolysis of tetra-
hydrofurfuryl alcohol to 1,5-pentanediol over Rh/SiO2. J Catal
267:89–92
123