Glycerol hydrogenolysis to propylene glycol and ethylene glycol on zirconia supported noble metal catalysts

Article abstract of DOI:10.1021/cs400486z

Monoclinic zirconia (m-ZrO₂) supported Ru, Rh, Pt, and Pd nanoparticles with controlled sizes were prepared and examined in glycerol hydrogenolysis to propylene glycol and ethylene glycol at similar conversions in the kinetic regime. Their activity (normalized per exposed surface metal atom, i.e., turnover rate) and selectivity depend sensitively on the nature of the noble metals and their particle size. At a similar size (ca. 2 nm), Ru exhibited a greater turnover rate than Rh, Pt, and Pd, and the rate decreased in the sequence Ru Rh > Pt > Pd by a factor of about 25 (from 0.035 to 0.0014 mol glycerol (mol surface metal·s)^-1) at 473 K and 6.0 MPa H ₂. Following such activity sequence, Ru was more prone to catalyze excessive cleavage of C-C bonds, leading to the formation of ethylene glycol and methane, while Pd exhibited the highest selectivity to cleavage of C-O bonds to propylene glycol. Similarly, larger Ru particles possessed higher glycerol hydrogenolysis activity concurrently with higher selectivities to ethylene glycol and especially methane at the expense of propylene glycol in the range of 1.8-4.5 nm. Analysis of kinetics and thermodynamics for the proposed elementary steps involving kinetically relevant glycerol dehydrogenation to glyceraldehyde leads to expressions of glycerol hydrogenolysis rate and selectivity to cleavage of C-O bonds relative to C-C bonds. Together with different effects of reaction temperature and atmosphere of H₂ and N₂ on the activity and selectivity for Ru/m-ZrO₂ and Pt/m-ZrO₂, these results suggest that the observed difference for different noble metals and particle sizes can be attributed to the difference in the strength of adsorption of glycerol and glyceraldehyde, derived from their different availability of unoccupied d orbitals.

Full text of DOI:10.1021/cs400486z

Research Article
pubs.acs.org/acscatalysis
Glycerol Hydrogenolysis to Propylene Glycol and Ethylene Glycol on
Zirconia Supported Noble Metal Catalysts
Shuai Wang, Kehua Yin, Yichi Zhang, and Haichao Liu*
Beijing National Laboratory for Molecular Sciences, State Key Laboratory for Structural Chemistry of Unstable and Stable Species,
College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China
*
S Supporting Information
ABSTRACT: Monoclinic zirconia (m-ZrO ) supported Ru, Rh, Pt, and
2
Pd nanoparticles with controlled sizes were prepared and examined in
glycerol hydrogenolysis to propylene glycol and ethylene glycol at similar
conversions in the kinetic regime. Their activity (normalized per exposed
surface metal atom, i.e., turnover rate) and selectivity depend sensitively
on the nature of the noble metals and their particle size. At a similar size
(
ca. 2 nm), Ru exhibited a greater turnover rate than Rh, Pt, and Pd, and
the rate decreased in the sequence Ru ≫ Rh > Pt > Pd by a factor of
−1
about 25 (from 0.035 to 0.0014 mol glycerol (mol surface metal·s) ) at
73 K and 6.0 MPa H . Following such activity sequence, Ru was more
4
2
prone to catalyze excessive cleavage of C−C bonds, leading to the formation of ethylene glycol and methane, while Pd exhibited
the highest selectivity to cleavage of C−O bonds to propylene glycol. Similarly, larger Ru particles possessed higher glycerol
hydrogenolysis activity concurrently with higher selectivities to ethylene glycol and especially methane at the expense of
propylene glycol in the range of 1.8−4.5 nm. Analysis of kinetics and thermodynamics for the proposed elementary steps
involving kinetically relevant glycerol dehydrogenation to glyceraldehyde leads to expressions of glycerol hydrogenolysis rate and
selectivity to cleavage of C−O bonds relative to C−C bonds. Together with different effects of reaction temperature and
atmosphere of H and N on the activity and selectivity for Ru/m-ZrO and Pt/m-ZrO , these results suggest that the observed
2
2
2
2
difference for different noble metals and particle sizes can be attributed to the difference in the strength of adsorption of glycerol
and glyceraldehyde, derived from their different availability of unoccupied d orbitals.
KEYWORDS: glycerol, selective hydrogenolysis, propylene glycol, ethylene glycol, noble metal, size effect
1
. INTRODUCTION
carbon nanotubes decreased by two-thirds when the particle
size increased from 3 to 10 nm, whereas the selectivities to
propylene glycol and ethylene glycol increased significantly
Selective hydrogenolysis of glycerol to propylene glycol and
ethylene glycol provides a viable route to utilize the large
surplus of glycerol, the principal byproduct from the biodiesel
production. This reaction also provides an appropriate probe
for studying the selective cleavage of C−O and C−C bonds in
the catalytic hydrodeoxygenation of biomass-derived poly-
(
from 32% to 58% and 10% to 21%, respectively) at 473 K and
27
1
4.0 MPa H2. Different from such inconsistent size effects, Ru
generally exhibits higher activity, but also higher selectivity to
cleavage of C−C bonds over C−O bonds than Rh, Pt, and Pd
28
2
−4
in glycerol hydrogenolysis especially under neutral conditions.
ols,
which is regarded as an important process for future
5
,6
For instance, Maris and Davis found that the turnover rate of
Ru on active carbon was about 2-fold greater than that of Pt
with the same metal particle size of 2.3 nm at 473 K and 4.0
biorefineries. As a consequence, glycerol hydrogenolysis has
been intensively studied in the past decade,
metal-based catalysts have been accordingly explored, including
Ru,
Specifically, the noble metals, that is, Ru, Rh, Pt, and Pd, have
received great attention because of their superior catalytic
activity
genolysis.
Glycerol hydrogenolysis on the noble metals is structure
sensitive, depending on their surface structure and particle
size. Montassier et al. primarily reported that the turnover
rate per surface Ru atom on Ru/C increased sharply with the
Ru particle size in the range of 1.2−4.2 nm, while the
selectivities to propylene glycol and ethylene glycol changed
1
−3
and various
7
−10
11
12
13
14
15
16
17−23
24
MPa H , while the selectivity ratios of propylene glycol to
2
Rh, Pt, Pd, Ir, Ni, Co, Cu,
and Ag.
ethylene glycol on Ru/C and Pt/C were 0.5 and 4.6,
8
respectively. Recently, van Ryneveld et al. also reported a
1
,2
6
hydrogenolysis activity order of Ru/Al O > Pd/Al O > Pt/
and hydrothermal stability in glycerol hydro-
2
3
2
3
Al O with 1 nm metal particles at 453 K and 8.0 MPa H , and
2
3
2
the corresponding cleavage ratios of C−O/C−C bonds showed
29
an opposite trend.
2
5
To improve the activity and selectivity on the noble metal
catalysts, recent studies have focused on the promoting effects
Received: June 28, 2013
Revised: July 30, 2013
26
slightly at 483 K and 6 MPa. However, Wang et al. found that
the turnover rate of the Ru particles supported on multiwalled
©
XXXX American Chemical Society
2112
dx.doi.org/10.1021/cs400486z | ACS Catal. 2013, 3, 2112−2121

ACS Catalysis
Research Article
Table 1. Metal Particle Sizes, Turnover Rates, and Selectivities in Glycerol Hydrogenolysis for m-ZrO -Supported Ru, Rh, Pt,
2
a
and Pd and γ-Al O Supported Ru with 2 wt % Metal Loading
2
3
c
selectivity (%)
b
e
d
catalyst
dH2 (nm)
conversion (%)
turnover rate
propylene glycol
ethylene glycol
CH4
ethanol
propanol
Ru/m-ZrO2
2.3
2.1
2.2
2.4
2.1
22.9
18.7
22.8
20.4
21.0
34.9
22.1
9.1
45.7
14.3
65.6
85.7
90.5
21.0
34.5
12.4
7.4
25.0
44.3
5.4
3.7
3.8
0.8
0.5
2.9
3.2
2.5
6.1
3.5
Ru/γ-Al O3
2
Rh/m-ZrO2
Pt/m-ZrO2
Pd/m-ZrO2
15.7
5.6
trace
trace
1.4
5.5
a
3
Reaction conditions: 50 cm 10 wt % glycerol aqueous solution, 473 K, 6.0 MPa H , 3 h, about 20% glycerol conversion obtained by varying the
2
b
c
catalyst amount. Metal particle sizes measured by H -chemisorption. Trace amounts of CO and methanol were detected in the products and not
shown here. Including 1-propanol and 2-propanol. 10 mol glycerol (mol surface metal·s) .
2
2
d
e
−3
−1
3
of second metal components and acid-basic functions in
A.R., Beijing Chemical) were dissolved in 80 cm deionized
3
,28
4+
−4
−3
glycerol hydrogenolysis.
Addition of non-noble metal such
water with a Zr concentration of 4.0 × 10 mol cm and a
as Cu and Fe into Ru concurrently improved the C−O cleavage
urea/Zr4+ molar ratio of 10/1. The solution was then placed in
3
to propylene glycol and suppressed the C−C cleavage to
a Teflon-lined stainless-steel autoclave (100 cm ) and held at
1
0,30
ethylene glycol.
Such effects were also observed for Pd−Fe,
433 K under autogenous pressure for 20 h. Afterward, the
resulting precipitates were washed with deionized water until
the filtrates were neutral and then treated subsequently in
ambient air at 383 K overnight and at 673 K for 4 h. The as-
31
Pd−Zn, and Pd−Ni bimetallic catalysts. Tomishige and co-
workers increased the glycerol hydrogenolysis rate by about
three fold and the propylene glycerol selectivity from 26% to
32
5
5% by mixing an acidic ion-exchange resin with Ru/C.
synthesized m-ZrO
Brunauer−Emmett−Teller (BET) surface area of 131 m /g.
Ru, Rh, Pt, and Pd nanoparticles supported on m-ZrO were
prepared by incipient wetness impregnation, including Ru/m-
ZrO with Ru loadings of 0.51, 1.0, 2.1, 2.9, 4.0, and 4.8 wt %,
2.0 wt % Rh/m-ZrO , 2.0 wt % Pt/m-ZrO , and 2.1 wt % Pd/
m-ZrO . RuCl ·nH O (G.R., Sinopharm Chemical), RhCl
nH O (A.R., Alfa Aesar), H PtCl ·6H O (A.R., Beijing
Chemical), and PdCl (A.R., Shenyang Research Institute of
2
support was measured to have a
2
Alhanash et al. obtained about 96% propylene glycerol
selectivity at 21% glycerol conversion at 423 K and 0.5 MPa
H on an acidic Cs H [PW O ]-supported Ru catalyst,
2
2
2.5 0.5
12 40
although the selectivity decreased to 88% when the conversion
2
3
3
increased to 31%. Yuan et al. found that the activity of
glycerol hydrogenolysis and selectivity to propylene glycol on
2
2
·
3
2
3
2
12
Pt-based catalysts increased with their support basicity. Feng
et al. reported that addition of LiOH significantly improved the
propylene glycol selectivity from 48% to 87% with the
2
2
6
2
2
Nonferrous Metals) were used as the metal precursors
dissolved in deionized water. After impregnation, these catalysts
were treated at 383 K overnight in dry air, and then reduced in
flowing 20% H /N (>99.99%, Beijing Huayuan) by heating to
73 K at a ramping rate of 0.083 K s and holding at 673 K for
h. Similarly, 2.2 wt % Ru/γ-Al O was also prepared, and a
commercial γ-Al O (A.R., Alfa Aesar) with a BET surface area
of 195 m /g was used as the support.
.2. Catalyst Characterization. Crystalline structures of
the catalysts were characterized by X-ray diffraction (XRD) on
a Rigaku D/MAX-2400 diffractometer using Cu Kα1 radiation
λ = 1.5406 Å) operated at 40 kV and 100 mA. The noble
metal particle sizes of the catalysts were measured by
transmission electron microscopy (TEM) and H -chemisorp-
tion. TEM images of fresh samples were obtained at an
acceleration voltage of 300 kV on Philips Tecnai F30
FEGTEM. H -chemisorption for the supported Ru, Rh, and
Pt samples was performed at 313 K and 10−50 kPa H using a
Quantachrome Autosorb-1 analyzer after the samples were
concurrent increase in the glycerol conversion from 66% to
9
9
0% on Ru/TiO at 443 K and 3 MPa H . Noticeably, acid
2
2
sites on sulfated zirconia, WO , and ReO can cooperate with
2
2
3
x
−1
6
the noble metals such as Pt and Ir, leading to selective
hydrogenolysis of the secondary hydroxyl group of glycerol to
3
2
3
1
4,34,35
form 1,3-propanediol.
For example, a high 1,3-propane-
2
3
2
diol yield of about 56% was achieved on Pt/sulfated ZrO at
2
34
2
4
43 K and 7.3 MPa H2. Clearly, these previous efforts and
advances have provided a basis for further improving the
efficiency of the noble metals, which will be beneficial to this
study toward a fundamental understanding that is still unclear
on the intrinsic catalytic activity and selectivity of the noble
metals and effects of their particle size in glycerol hydro-
genolysis.
(
2
In this work, we prepare supported Ru, Rh, Pt, and Pd
catalysts with similar metal dispersions on hydrothermally
stable monoclinic ZrO (m-ZrO ), and compare their intrinsic
2
2
2
2
activities and selectivities in glycerol hydrogenolysis at similar
glycerol conversions within the kinetic regime. We also examine
the effects of reaction temperature, atmosphere (H and N ),
treated in flowing 5% H /N (>99.99%, Beijing Huayuan) at
23 K for 2 h and then evacuated at the same temperature for
2 h. For 2 wt % Pd/m-ZrO , H -chemisorption was
performed at 343 K and 0.4−1.5 kPa H to avoid the formation
of the β hydride phase. Metal particle sizes for the samples were
accordingly calculated from the H -chemisorption results by
assuming the chemisorption stoichiometry of H/metal molar
2
2
6
1
2
2
37
and metal particle size, in attempts to unveil the reaction
mechanism of glycerol hydrogenolysis and the factors dictating
the cleavage of C−C and C−O bonds on the noble metal
catalysts.
2
2
2
2
ratio of 1.0 and that the metal particles were spherical.
2
. EXPERIMENTAL SECTION
.1. Preparation of Catalysts. Pure m-ZrO2 was
synthesized using a facile hydrothermal method reported
2
.3. Catalytic Reaction. Glycerol hydrogenolysis reactions
2
were carried out in a Teflon-lined stainless steel autoclave (100
mL) at a stirring speed of 800 rpm. In a typical run, 50 g of 10
wt % glycerol (A.R., Beijing Chemical) aqueous solution and
0.35 g of metal catalyst were placed in the autoclave. After fully
3
6
previously. Certain amounts of zirconyl nitrate (ZrO-
NO ) ·2H O, A.R., Beijing Chemical) and urea (CO(NH ) ,
(
3
2
2
2 2
2
113
dx.doi.org/10.1021/cs400486z | ACS Catal. 2013, 3, 2112−2121

ACS Catalysis
Research Article
purging the autoclave with H (>99.99%, Beijing Huayuan), the
effects of the metal particle size prevalently found for many
reactions.
2
reactor was pressurized with H to 6.0 MPa and then heated to
2
the reaction temperatures (453−513 K). Liquid products were
separated by centrifugation and analyzed by gas chromatog-
raphy (Agilent 7890A GC) using a capillary column (AT-
Aquawax: 30 m × 0.25 mm × 0.25 m) connected to a flame
ionization detector. 1-Butanol and 1,4-butanediol were used as
external standards for volatile products (e.g., methanol, ethanol,
The glycerol hydrogenolysis activities in Table 1 were
normalized by the number of exposed metal atoms estimated
from the results of H -chemisorption, which represent the
2
turnover rates and reflect the intrinsic activity of the accessible
metal atoms. Detected products included propylene glycol, 1-
propanol, and 2-propanol, which were formed via cleavage of
C−O bonds of glycerol, and ethylene glycol, ethanol, methanol,
1
-propanol, and acetal) and products of high boiling points
(
e.g., ethylene glycol, propylene glycol, and glycerol),
methane, and CO , which were formed via cleavage of C−C
2
7
respectively. Gas products, mainly methane and CO , were
bonds, as reported previously. The turnover rates and
2
analyzed by gas chromatography (Shimadzu 2010 GC,
Porapak-Q column) with a thermal conductivity detector.
The carbon balance of the analysis was generally within 100 ±
selectivities were compared at similar glycerol conversions of
approximately 20% within the kinetic regime. As shown in
Table 1, Ru/m-ZrO showed a higher turnover rate than Ru/γ-
2
5
%. Reaction activities were reported by the molar glycerol
Al O with a similar Ru particle size of 2 nm (0.035 vs 0.022
2
3
−1
conversion rates per mole of exposed metal atoms (estimated
mol glycerol (mol surface Ru·s) ). Selectivities to propylene
glycol, ethylene glycol, and methane on Ru/m-ZrO were
from the H -chemisorption data described above), that is,
2
2
−1
turnover rates (mol glycerol (mol surface metal·s) ) and
product selectivities were reported on the carbon basis, as
employed by Miyazawa et al.
45.7%, 21.0%, and 25.0%, respectively, at ∼20% glycerol
conversion. On Ru/γ-Al O , the selectivity to propylene glycol
2
3
7
decreased to 14.3%, whereas the selectivity to ethylene glycol
increased to 34.5% with a large amount of methane (44.3%
selectivity). Such difference in the hydrogenolysis activity and
selectivity between the two Ru catalysts is clearly not due to the
difference in their metal particle sizes, but due to the strong
3
. RESULTS AND DISCUSSION
.1. Activity and Selectivity of Supported Noble-
3
Metal Catalysts in Glycerol Hydrogenolysis. Table 1
shows glycerol hydrogenolysis activities and selectivities on Ru/
m-ZrO , Rh/m-ZrO , Pt/m-ZrO , and Pd/m-ZrO , and also for
support effect. It is known that m-ZrO surface possesses
2
38
stronger basicity than γ-Al O , facilitating the kinetically
2
2
2
2
2
3
comparison on Ru/γ-Al O with 2 wt % metal loading at 473 K
relevant cleavage of the α-C-H bond of glycerol, and the
selective formation of propylene glycol on the Ru surface, as
discussed below.
2
3
and 6.0 MPa H . The five catalysts present merely the XRD
2
patterns of the corresponding support of m-ZrO or γ-Al O
2
2
3
(Figure 1). The absence of the characteristic diffraction peaks
Glycerol hydrogenolysis is sensitive to the property of the
noble metals as well. Compared with Ru/m-ZrO , the turnover
2
rates on Rh/m-ZrO , Pt/m-ZrO , and Pd/m-ZrO decreased
2
2
2
from 0.035 to 0.0091, 0.0056, and 0.0014 mol glycerol (mol
−
1
surface metal·s) , respectively (Table 1). Specifically, the
catalytic activity of Ru/m-ZrO was about 25 times greater than
2
that of Pd/m-ZrO . The selectivities to the products via the
2
cleavage of C−O and C−C bonds presented the opposite
trends with the variation of the noble metals (Table 1). At
∼
20% glycerol conversion, Ru/m-ZrO exhibited the lowest
selectivity to propylene glycol (45.7%), and the highest
selectivities to ethylene glycol (21.0%) and methane (25.0%)
2
among the four m-ZrO supported noble metal catalysts. The
2
selectivity to propylene glycol increased to 65.6%, and the
selectivities to ethylene glycol and methane decreased to 12.4%
and 15.7%, respectively, on Rh/m-ZrO . On Pd/m-ZrO , the
2
2
selectivity to propylene glycol was as high as 90.5% while the
selectivities to ethylene glycol decreased to 5.5% with the
negligible formation of methane. Such strong effects of the
noble metals, as discussed below, are closely relevant to their
effects on the reaction pathways of glycerol hydrogenolysis.
Two reaction routes, that is, dehydration and dehydrogen-
ation routes, for glycerol hydrogenolysis to propylene glycol
Figure 1. X-ray diffraction patterns for m-ZrO -supported Ru, Rh, Pt,
2
and Pd with 2 wt % metal loading, γ-Al O -supported Ru with 2 wt %
2
3
metal loading, and the corresponding supports of m-ZrO and γ-Al2O3
2
2
,3
have been intensively discussed in the literature. In the
dehydration route, glycerol directly dehydrates to acetol on acid
sites, which then hydrogenates to propylene glycol on metal
as references.
7
of the noble metals indicates that the metals were highly
sites. In the dehydrogenation route, glycerol first dehydrogen-
dispersed on the supports. TEM and H -chemisorption
ates to glyceraldehyde on metal sites, followed by its
dehydration and subsequent hydrogenation to acetol and
propylene glycol. Our control experiments showed that pure
2
characterization results confirm the XRD results, showing that
the metal particles of these catalysts possessed similar average
sizes (∼2 nm) and narrow size distributions, irrespective of the
identity of the metal and support (Figure 2 and Table 1). Such
catalyst properties enable us to compare their intrinsic activity
and selectivity in glycerol hydrogenolysis, free of the known
8
m-ZrO and γ-Al O did not give detectable glycerol conversion
2
2
3
in the absence of the noble metals under the identical
conditions used in this study, suggesting that the metal sites
are prerequisite for the initial activation and reaction of
2
114
dx.doi.org/10.1021/cs400486z | ACS Catal. 2013, 3, 2112−2121

ACS Catalysis
Research Article
Figure 2. Transmission electron microscopy micrographs and histograms of metal particle size distribution for m-ZrO -supported Ru, Rh, and Pt
2
and γ-Al O -supported Ru with 2 wt % metal loading (scale bar = 10 nm). (a) 2 wt % Ru/m-ZrO , (b) 2 wt % Rh/m-ZrO , (c) 2 wt % Pt/m-ZrO ,
2
3
2
2
2
and (d) 2 wt % Ru/γ-Al O .
2
3
glycerol. Such inertness of the two oxide supports for glycerol
hydrogenolysis were also observed by Vasiliadou et al., who
proposed that metal sites correlated with the dehydration rate
of glycerol. Recently, Anueau et al. reported that the
dehydrogenation route is kinetically more favorable than the
dehydration route on a Rh surface based on their experimental
11
and theoretical results. Consistent with this study, glycer-
aldehyde from the dehydrogenation of glycerol has also been
proposed as an intermediate for ethylene glycol via its C−C
3
9
bond cleavage and for C products via its decarbonylation on
1
2
115
dx.doi.org/10.1021/cs400486z | ACS Catal. 2013, 3, 2112−2121

ACS Catalysis
Research Article
Scheme 1. Proposed Pathways for Cleavage of C−O and C−C Bonds in Glycerol Hydrogenolysis on Noble Metal Catalysts
4
0
noble metal surfaces or via retro-aldol condensation on basic
relative rates of glyceraldehyde dehydration to direct decarbon-
ylation and retro-aldol condensation of glyceraldehyde
(Scheme 1). The turnover rates on Ru/m-ZrO , Rh/m-ZrO ,
4
1
sites. Therefore, we assume that the dehydrogenation route is
dominant in glycerol hydrogenolysis on the supported noble
metal catalysts in our reaction conditions.
2
2
Pt/m-ZrO , and Pd/m-ZrO at similar metal dispersions and
2
2
8,40,42
Together with the studies in the literature,
the reaction
kinetic regimes indicate that the intrinsic dehydrogenation
ability of noble metals follows the sequence Ru ≫ Rh > Pt >
Pd, which is consistent with the results for 2-propanol
dehydrogenation on the noble metals reported by Ukisu and
pathways of glycerol hydrogenolysis are proposed in Scheme 1.
Glycerol dehydrogenates to glyceraldehyde on the noble metal
surface. Then the glyceraldehyde intermediate undergoes
dehydration to form 2-hydroxyacrylaldehyde, which hydro-
genates sequentially to acetol and propylene glycol. Alter-
natively, glyceraldehyde converts to ethylene glycerol and CO
via its decarbonylation on the noble metal surfaces or to the 2-
hydroxyacetaldehyde and formaldehyde intermediates via retro-
aldol condensation on the basic sites, which further hydro-
genate to ethylene glycol and methanol, respectively. The
formation of methane is mainly contributed from hydro-
genolysis of methanol and hydrogenation of CO on noble
metals. Moreover, CO may convert to CO via water gas shift
reaction. Such proposed reaction pathways on the noble metals
are similar to those on Cu-based catalysts, except for the
insignificant cleavage of the C−C bonds because of the known
low catalytic activity of the Cu surface for the decarbonylation
reaction. Instead, glyceraldehyde tends to degrade through the
43
Miyadera. Specifically, they suggested the cleavage of α-C-H
bond of glycerol in its dehydrogenation step is kinetically
43
relevant. Maris and Davis have demonstrated that both
dehydration and retro-aldol condensation of the glyceraldehyde
intermediate are base-catalyzed reactions in glycerol hydro-
8
genolysis. Accordingly, the observed variation of the ethylene
glycol selectivity with the noble metals on m-ZrO (Table 1)
2
indicates that the C−C bond cleavage in glycerol hydro-
genolysis mostly occurs via glyceraldehyde decarbonylation on
the metal surfaces rather than via retro-aldol condensation.
Otherwise, the selectivities to the aforementioned products via
cleavage of C−O and C−C bonds should be determined by the
4
2
2
18
basicity of the m-ZrO
the noble metals.
support, irrespective of the identity of
2
Scheme 2 lists a proposed sequence of elementary steps,
which are related to rGly and rC−O/rC−C, and the associated
kinetic and thermodynamic constants. Here rGly and rC−O/rC−C
18
retro-Claisen route to form ethylene glycol on the Cu surface.
In addition, the primary hydrogenolysis products, propylene
glycerol and ethylene glycerol, undergo further hydrogenolysis
reactions to form 1-propanol and ethanol, and so on (not
shown in Scheme 1 for clarity).
Scheme 2. . Proposed Sequence of Elementary Steps and
Associated Kinetic or Thermodynamic Constants in
Glycerol Hydrogenolysis Catalyzed by Noble Metals
The undetectable amount of glyceraldehyde in the liquid
products indicates that dehydrogenation of glycerol to
glyceraldehyde on the metal surfaces was irreversible and far
from the equilibrium. Consistently, the dehydrogenation step
has been suggested to be kinetically relevant to the glycerol
hydrogenolysis in neutral and basic conditions as evidenced
8
,11
18,19
both on noble metals and Cu-based catalysts.
According
to the reaction pathways discussed above, the reaction rates of
glycerol hydrogenolysis (r ) are determined by the step of
Gly
glycerol dehydrogenation to glyceraldehyde on metal surfaces,
while the rate ratios of the primary hydrogenolysis of C−O
bonds (leading to the formation of propylene glycol) and C−C
bonds (leading to the formation of ethylene glycol and the C1
products, etc.), denoted as rC−O/rC−C, are determined by the
2
116
dx.doi.org/10.1021/cs400486z | ACS Catal. 2013, 3, 2112−2121

ACS Catalysis
Research Article
are expressed as eqs 1 and 2 by applying the quasi-equilibrium
assumption for Steps 1, 2, 4, and 5, in which kDH and k
are
CO
the respective kinetic constants for the cleavage of the α-C-H
bond of glycerol and glyceraldehyde decarbonylation on metal
surfaces; kDOH is the kinetic constant for glyceraldehyde
dehydration catalyzed by basic sites on m-ZrO in liquid
2
phase; K , K , and K are the respective adsorption
Gly
GA
H2
constants for glycerol, glyceraldehyde, and H , and K
is
2
Alkoxide
the dissociation constant of glycerol to the alkoxide species on
metal surfaces; cGly represents the molar concentration of
glycerol in liquid phase; and θ refers to the concentration of
*
vacant sites on metal surfaces. The details of the derivation of
eqs 1 and 2 are included in the Supporting Information.
kDHKGlyKAlkoxide cGly
2
rGly
=
θ
KH21/2
PH2
1/2 *
(1)
r
kDOH
1
C−O
= k
KGA
θ
2
r
(2)
C−C
CO
*
Based on transition-state theory (eq 3) and the relationship
between the equilibrium constant, enthalpy, and entropy of a
reaction (eq 4), eq 2 is converted to its thermodynamic form,
eq 5, in which Ea,DOH and E
are the activation energies for
a,CO
glyceraldehyde dehydration and decarbonylation, respectively;
⧧
⧧
ΔS
and ΔS
are the entropy changes from the
DOH
CO
reactant to the transition state for glyceraldehyde dehydration
and decarbonylation, respectively; ΔH and ΔS are the
GA
GA
adsorption enthalpy and entropy for glyceraldehyde on metal
surfaces. It is clear from eqs 1 and 5 that rGly and rC−O/rC−C are
both functions of the available vacant sites on metal surfaces.
Therefore, varying the vacancy concentration of the metal
surfaces, for example, by changing the reaction temperature or
hydrogenolysis atmosphere from H to N , can provide insight
2
2
into the difference of the catalytic activity and selectivity among
the different noble metals examined in this work, as shown in
the next section.
k T
h
⧧
B
(ΔS +1)/R −E /RT
a
k(T) =
e
e
(3)
(4)
Figure 3. Effects of reaction temperature on turnover rates and
selectivities for glycerol hydrogenolysis on Ru/m-ZrO (a) and Pt/m-
2
K(T) = eΔS/R
e
−ΔH/RT
3
ZrO (b) with 2 wt % metal loading. Reaction conditions: 50 cm 10
2
wt % glycerol aqueous solution, 6.0 MPa H , 3 h, about 20% glycerol
2
r
C−O
⧧
− ΔS⧧
=
exp ⎡⎣ ΔS
− ΔSGA)/R⎤⎦
conversion obtained by varying the catalyst amount.
(
DOH
CO
r
C−C
exp ⎡⎣ −(Ea,DOH − Ea,CO − ΔHGA)/RT ⎤⎦ 1
2
*
increasing the reaction temperature from 453 to 483 K, and
then decreased to 51.1% when the temperature further
increased to 513 K (Figure 3a). The selectivities to ethylene
glycol via C−C cleavage and 1-propanol via hydrogenolysis of
propylene glycol decreased from 30.0% to 17.2% and from
2.8% to 1.9%, respectively, with increasing reaction temperature
in the whole range of 453−513 K. Accompanying the above
changes, the selectivity to methane increased monotonically
from 16.8% to 25.3% with the reaction temperature (453−513
K). It is obvious that high temperatures favor the deep
hydrogenolysis of glycerol to methane. Miyazawa et al. has
reported that propylene glycol and 1-propanol are less reactive
θ
(5)
3
.2. Effects of Reaction Temperature and Atmosphere
on Glycerol Hydrogenolysis. Figure 3 shows the turnover
rates and selectivities for glycerol hydrogenolysis as a function
of reaction temperature on the two representative catalysts, 2
wt % Ru/m-ZrO and 2 wt % Pt/m-ZrO . The turnover rate of
2
2
Ru/m-ZrO increased exponentially from 0.019 to 0.133 mol
2
−1
glycerol (mol surface Ru·s) with increasing reaction temper-
ature from 453 to 513 K, while the turnover rate of Pt/m-ZrO2
increased from 0.0026 to 0.0082 mol glycerol (mol surface Pt·
−1
s) in the temperature range of 453−483 K. The calculated
7
apparent activation energies for Ru/m-ZrO and Pt/m-ZrO
than ethylene glycol in hydrogenolysis reactions, indicating
2
2
from Figure 3 were 72 and 84 kJ/mol, respectively, which are
consistent with the above statement that the alkanol
dehydrogenation ability of Ru is superior to that of Pt.
that methane is mainly formed from the C−C cleavage of
glycerol or further hydrogenolysis of ethylene glycol, but not
from propylene glycol or 1-propanol. Noticeably, the total
selectivities to propylene glycol, 1-propanol, ethylene glycol,
and methane were always kept above 95% in this work. Such
The selectivity to propylene glycol on Ru/m-ZrO increased
2
from 47.2% to 56.5% at ∼20% glycerol conversion when
2
117
dx.doi.org/10.1021/cs400486z | ACS Catal. 2013, 3, 2112−2121

ACS Catalysis
Research Article
excellent carbon balance enables us to compare the primary
hydrogenolysis selectivities to the cleavage of C−O and C−C
bonds by the selectivity ratios of the sum of propylene glycol
and 1-propanol to the sum of ethylene glycol and methane. It is
readily seen that the ratios of the primary hydrogenolysis
selectivity of C−O bonds to that of the C−C bonds on Ru/m-
m-ZrO support, which means E
should be constant for all
a,DOH
2
the m-ZrO -supported noble metal catalysts. In contrast,
2
decarbonylation of carbonyl compounds on the noble metal
surfaces involves a cleavage of C−C bonds and usually occurs at
high temperature because of its large activation barrier. Wang
et al. reported that the cleavage of C−C bonds on the noble
metals showed similar activation barriers. For instance, the
44
45
ZrO possessed a similar volcano-type dependence on the
2
reaction temperature, and reached the maximum value at 483 K
activation barriers for the conversion of CH CH * surface
3
2
(Figure 4), as observed for the propylene glycol selectivity
species to CH * and CH * on Ru(211) and Pt(211) were 1.3
3 2
and 1.4 eV, respectively, whereas a high value of 2.0 eV was
4
5
found on the Cu(211) surface. As a consequence, (Ea,DOH
−
E
) in eq 5 is assumed to be negative and identical for Ru/
a,CO
m-ZrO and Pt/m-ZrO . When |ΔH | is large enough and
2
2
GA
(
E
− E
− ΔH ) > 0, the second term of eq 5 will
a,DOH
a,CO
GA
increase with the reaction temperature and lead to a volcano-
type dependence of the rC−O/rC−C ratios on the reaction
temperature, as indeed is observed with Ru/m-ZrO (Figure 4).
2
When |ΔH | is small and (E
− E
− ΔH ) < 0, the
GA
a,DOH
a,CO
GA
second term of eq 5 will decrease with the reaction
temperature, and thus rC−O/rC−C will decrease with the reaction
temperature monotonically, as indeed observed with Pt/m-
ZrO (Figure 4). Such analysis suggests that the adsorption of
2
glyceraldehyde on Ru is much stronger than on Pt. Similarly,
Abild-Pedersen and Andersson reported that the CO
adsorption energy on Ru(0001) and Pt(111) measured at
low coverage (0.25 ML) was −1.49 ± 0.22 and −1.37 ± 0.13
eV, respectively, implying a general trend of carbonyl group
46
adsorption on the different noble metal surfaces.
The assumed stronger adsorption of glyceraldehyde on Ru
than on Pt accords with the different atmosphere effects of H2
and N in glycerol hydrogenolysis on 2 wt % Ru/m-ZrO and 2
Figure 4. Primary C−O/C−C cleavage ratio for glycerol hydro-
genolysis on Ru/m-ZrO and Pt/m-ZrO with 2 wt % metal loading.
2
2
2
2
3
Reaction conditions: 50 cm 10 wt % glycerol aqueous solution, 6.0
wt % Pt/m-ZrO . As shown in Figure 5, the catalytic activity of
2
MPa H , 3 h, ∼20% glycerol conversion obtained by varying the
2
Ru/m-ZrO under H was higher than that under N , whereas
2
2
2
catalyst amount.
glycerol hydrogenolysis occurred faster under N than H on
2
2
Pt/m-ZrO . To make a rigorous comparison, the initial glycerol
2
(
Figure 3a). Different temperature effects on the selectivities
hydrogenolysis rates were calculated by extracting the slopes of
18
were found for Pt/m-ZrO (Figure 3b). No methane was
formed in the reaction temperature range 453−513 K at about
the fitted conversion profiles at zero reaction time. The initial
turnover rates on Ru/m-ZrO under H and N were 0.166 and
0.016 mol glycerol (mol surface Ru·s) , respectively, showing
the much more favorable glycerol reaction under H , relative to
. For Pt/m-ZrO , the reverse order was observed, and the
initial rate under H was nearly three times smaller than that
under N (0.027 vs 0.072 mol glycerol (mol surface Pt·s) ).
The expression of rGly in eq 1 shows clearly that the presence of
inhibits glycerol hydrogenolysis, provided that the vacant
metal sites are readily available, which is consistent with the
atmospheric effect of H and N on Pt/m-ZrO . The observed
opposite effect on Ru/m-ZrO indicates that the concentration
2
2
2
2
−
1
2
0% glycerol conversion. The selectivity to propylene glycol
monotonically decreased from 94.9% to 82.8% with concurrent
increase in the selectivities to ethylene glycol and 1-propanol
from 1.0% to 7.4% and 2.1% to 7.7%, respectively, as the
temperature ramped from 453 to 513 K. Therefore, the primary
2
N
2
2
2
−1
2
C−O/C−C cleavage ratio on Pt/m-ZrO decreased with
2
increasing reaction temperature (Figure 4). Taken together,
these results show that an increase of reaction temperature can
improve the selective hydrogenolysis of C−O bonds in glycerol
to propylene glycol on Ru/m-ZrO , but not on Pt/m-ZrO .
H
2
2
2
2
2
2
2
The dependence of the selectivity to the cleavage of C−O
and C−C bonds is derived from eq 5. It is known that enthalpy
and entropy apparently do not change with temperature.
Therefore, the first term of eq 5 which has the contribution of
of vacant sites on the Ru surfaces play a dominant role in
determining the hydrogenolysis turnover rate, most likely
because the Ru sites strongly adsorbed glyceraldehyde and
other unsaturated intermediates in glycerol hydrogenolysis.
Such strong adsorbates are required to be hydrogenated by H2
to consequently liberate the vacant Ru sites. Thus the glycerol
entropy is nearly independent of the reaction temperature,
2
*
whereas the third term (1/θ ) decreases with the reaction
temperature because the adsorption of glycerol and hydro-
hydrogenolysis rate was much higher under H than that under
2
genolysis products on metal surfaces is exothermic (ΔH
N on Ru/m-ZrO . These results can also account for the
adsorption
2
2
<
0). Thus, the source of the observed difference between Ru/
observed larger selectivities to the cleavage of C−C bonds and
the deep hydrogenolysis to methane on Ru, compared with Rh,
Pt, and Pd, which are derived from its stronger adsorption of
glyceraldehyde and other unsaturated intermediates in glycerol
hydrogenolysis. Mavrikakis and Barteau have pointed out that
the adsorption of carbonyl compounds on Group VIII metals is
m-ZrO and Pt/m-ZrO as it concerns the temperature effect
2
2
on rC−O/rC−C lies in the second term of eq 5. Glyceraldehyde
dehydration appears to occur readily, considering the
thermodynamically favored formation of the π-conjugated
18
dehydration product (2-hydroxyacrylaldehyde). Moreover,
the dehydration reaction is catalyzed by the basic sites on the
2
47
via a η (C, O) configuration. Compared with Rh, Pt, and Pd,
2
118
dx.doi.org/10.1021/cs400486z | ACS Catal. 2013, 3, 2112−2121

ACS Catalysis
Research Article
wt % to 5 wt %. The H -chemisorption results showed that
2
these samples possessed Ru particle sizes of 1.8, 2.0, 2.3, 3.5,
4
.1, and 4.5 nm. As shown in Figure 6, the turnover rate
Figure 6. Effects of Ru particle size on turnover rates and selectivities
for glycerol hydrogenolysis on monoclinic zirconia supported Ru
catalysts. Reaction conditions: 50 cm3 10 wt % glycerol aqueous
solution, 473 K, 6.0 MPa H , 3 h, about 20% glycerol conversion
2
obtained by varying the catalyst amount.
(
normalized per surface Ru atom) increased by nearly three
times monotonically from 0.026 to 0.071 mol glycerol (mol
surface Ru·s)− when increasing the Ru particle size from 1.8 to
1
4
.5 nm at ∼20% glycerol conversion. A similar trend was
observed by Montassier et al. in glycerol hydrogenolysis on Ru/
C catalysts with Ru particles of diameters in the range of 1.2−
24
4
.2 nm. Ru particle size also significantly influenced the
product selectivities (Figure 6). At ∼20% glycerol conversion,
the increase of the Ru particle size from 1.8 to 4.5 nm led to a
decrease of the propylene glycol selectivity from 55.8% to
30.6%, and a concurrent increase of the selectivities to ethylene
glycol, 1-propanol, and methane from 19.1% to 27.3%, 1.9% to
2.8%, and 18.3% to 36.1%, respectively. Clearly, the larger Ru
particles favor the cleavage of C−C bonds and deep
hydrogenolysis reactions, in line with a recent study on the
ring-opening of cyclopentane showing that turnover rates of the
C−C bond cleavage on Pt/Al O increased monotonically with
Figure 5. Comparison of glycerol hydrogenolysis rates under H and
2
N on monoclinic zirconia (m-ZrO ) supported Ru and Pt catalysts
2
2
with 2 wt % metal loading. (a) 2 wt % Ru/m-ZrO ; (b) 2 wt % Pt/m-
2
3
ZrO ; reaction conditions: 50 cm 10 wt % glycerol aqueous solution,
2
4
73 K, 6.0 MPa H or N , 0.35 g 2 wt % Ru/m-ZrO and 1.0 g 2 wt %
2
2
2
Pt/m-ZrO for the reactions under H , 2.0 g 2 wt % Ru/m-ZrO and
2
2
2
1
.0 g 2 wt % Pt/m-ZrO for the reactions under N .
2 2
Ru possesses more available unoccupied d orbitals, and thus it
can accept more π electrons from the carbonyl intermediates
and adsorb them more strongly. This Ru property, according to
the Sabatier principle, also leads to its stronger interaction with
glycerol to lower the activation barrier of the cleavage of α-C-H
bonds during glycerol dehydrogenation to glyceraldehyde, the
kinetically relevant step. As a consequence, Ru exhibits a higher
catalytic activity than Rh, Pt, and Pd in glycerol hydrogenolysis,
as shown in Table 1. Similar effects have been reported for
many reactions on the noble metals, for instance, the activity
2
3
49
increasing Pt particle size (1−15 nm). Such strong size effects
on the catalytic activity and selectivity demonstrate that glycerol
hydrogenolysis is a typical structure-sensitive reaction.
It is known that the fraction of edge sites on the surface of
metal particles below 5 nm decreases sharply with increase of
their particle size while the fraction of terrace sites increases
5
0
accordingly. The edge sites are more coordinatively
unsaturated than the terrace sites. Accordingly, the increase
of the Ru particle size weakens the strength of chemisorption of
reactants and products on the Ru surface and favors the release
of the vacant sites during glycerol hydrogenolysis. Moreover,
the weaker interaction between glycerol and the Ru surface
leads to the higher activation barrier, that is, the lower kinetic
constant, for the cleavage of α-C-H bonds in the kinetically
relevant glycerol dehydrogenation step. As shown in eq 1, the
turnover rate of glycerol hydrogenolysis is proportional to both
the concentration of the vacant metal sites and the kinetic
for N dissociation changes in the order of Ru > Rh > Pt,
2
48
consistent with their d-band unoccupancy.
3
.3. Size Effects of Ru Particles on Glycerol Hydro-
genolysis. To further understand the activity and selectivity of
the Ru catalysts in glycerol hydrogenolysis, we examined the
effects of the Ru nanoparticle size. Six Ru/m-ZrO samples
2
were prepared with controllable Ru particle size of diameters in
the range of 1.8−4.5 nm by changing the Ru loadings from 0.5
2
119
dx.doi.org/10.1021/cs400486z | ACS Catal. 2013, 3, 2112−2121

ACS Catalysis
Research Article
constant for the cleavage of the α-C-H bond of glycerol.
Therefore, the observed increase in the turnover rate with
increasing the Ru particle size is related to the increase in the
concentration of the available vacant Ru sites, as a result of the
aforementioned weaker adsorption of the reactants and
products in glycerol hydrogenolysis.
Rate ratios of the cleavage of the C−O bonds to the C−C
bonds reflect the relative activity between the competitive
dehydration and decarbonylation reactions of glyceraldehyde,
according to eq 5. The selectivity dependence on the Ru
particle size in glycerol hydrogenolysis is derived from the site-
requirement for cleavage of C−C bonds in the step of
glyceraldehyde decarbonylation on the metal surface, consid-
ering the dehydration of glyceraldehyde to cleavage of C−O
bonds is solely catalyzed by the basic sites on the m-ZrO2
ASSOCIATED CONTENT
Supporting Information
Derivation details of eqs 1 and 2. This material is available free
of charge via the Internet at http://pubs.acs.org.
■
*
S
AUTHOR INFORMATION
Corresponding Author
E-mail: hcliu@pku.edu.cn.
■
*
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by National Natural Science
Foundation of China (20825310, 21173008, and 51121091)
and National Basic Research Project of China (2011CB201400
and 2011CB808700).
support, as discussed above. Thus, rC−O/rC−C is determined by
⧧
E
and ΔS
GA
of glyceraldehyde decarbonylation, ΔH
a,CO
CO
GA
and ΔS of glyceraldehyde adsorption, and the concentration
of the vacant metal sites on the catalyst surface (eq 5). Here
REFERENCES
■
190.
(
(
2
(
4
(
⧧
ΔS and ΔS
should change slightly with the Ru particle
(1) Nakagawa, Y.; Tomishige, K. Catal. Sci. Technol. 2011, 1, 179−
GA
CO
size because glyceraldehyde and the transition state are both
2) ten Dam, J.; Hanefeld, U. ChemSusChem 2011, 4, 1017−1034.
3) Ruppert, A. M.; Weinberg, K.; Palkovits, R. Angew. Chem., Int. Ed.
012, 51, 2564−2601.
4) Shen, Y.; Wang, S.; Luo, C.; Liu, H. Prog. Chem. 2007, 19, 431−
36.
5) Luo, C.; Wang, S.; Liu, H. Angew. Chem., Int. Ed. 2007, 46, 7636−
7639.
(6) Sun, J.; Liu, H. Green Chem. 2011, 13, 135−142.
(7) Miyazawa, T.; Kusunoki, Y.; Kunimori, K.; Tomishige, K. J. Catal.
tightly bound to the Ru surface. As discussed above, |ΔH |
GA
decreases with the increase of Ru particle size, whereas E
a,CO
shows an opposite change. These dual effects will lead to the
increase in the second term of eq 5 with increasing Ru particle
size. Consequently, the observed higher selectivity to the
cleavage of C−C bonds at larger Ru particles in Figure 5 is
probably caused by the increase of the available vacant Ru sites,
because rC−O/rC−C is inversely dependent on the concentration
of the Ru vacant sites, as shown in eq 5. Such analysis suggests
that the concentration of the vacant sites on the Ru surface is
2
(
(
2
006, 240, 213−221.
8) Maris, E. P.; Davis, R. J. J. Catal. 2007, 249, 328−337.
9) Feng, J.; Wang, J.; Zhou, Y.; Fu, H.; Chen, H.; Li, J. Chem. Lett.
more critical to rC−O/rC−C than ΔH and E
, which is
GA
a,CO
007, 36, 1274−1275.
10) Li, B.; Wang, J.; Yuan, Y.; Ariga, H.; Takakusagi, S.; Asakura, K.
consistent with the fact that C−C bond dissociation usually
(
51
requires at least two adjacent vacant sites. Taken together, the
size effect unveils the determining role of the vacant sites for
the catalytic activity and selectivity in glycerol hydrogenolysis
on Ru, reflecting the nature of the strong interaction between
ACS Catal. 2011, 1, 1521−1528.
(11) Auneau, F.; Michel, C.; Delbecq, F.; Pinel, C.; Sautet, P.
Chem.Eur. J. 2011, 17, 14288−14299.
(
12) Yuan, Z.; Wu, P.; Gao, J.; Lu, X.; Hou, Z.; Zheng, X. Catal. Lett.
52
2009, 130, 261−265.
13) Musolino, M. G.; Scarpino, L. A.; Mauriello, F.; Pietropaolo, R.
Green Chem. 2009, 11, 1511−1513.
14) Nakagawa, Y.; Ning, X.; Amada, Y.; Tomishige, K. Appl. Catal.,
A 2012, 433, 128−134.
15) van Ryneveld, E.; Mahomed, A. S.; van Heerden, P. S.; Green,
M. J.; Friedrich, H. B. Green Chem. 2011, 13, 1819−1827.
16) Guo, X.; Li, Y.; Song, W.; Shen, W. Catal. Lett. 2011, 141,
the Ru particles and oxygenates.
(
4
. CONCLUSIONS
(
Ru, Rh, Pt, and Pd nanoparticles supported on m-ZrO catalyze
2
(
glycerol hydrogenolysis to propylene glycol and ethylene glycol,
which largely depends on the nature of the metals and their
particle size. Their turnover rate decreases in the order of Ru ≫
Rh > Pt > Pd with a similar particle size (∼2 nm), whereas the
selectivity ratio of the cleavage of C−O bonds to C−C bonds,
leading to propylene glycol and ethylene glycol, respectively,
decreases in the reverse order. Both the turnover rate and
selectivity to the cleavage of C−C bonds on Ru/m-ZrO2
increase monotonically with the Ru particle size in the range
of 1.8−4.5 nm. Such difference in the rate and selectivity for
different noble metals and different particle sizes, based on the
analysis of the elementary steps in glycerol hydrogenolysis
together with the observed effects of reaction temperature and
atmosphere of H and N , can be tentatively ascribed to their
(
1
458−1463.
(17) Wang, S.; Liu, H. Catal. Lett. 2007, 117, 62−67.
(18) Wang, S.; Zhang, Y.; Liu, H. Chem.Asian J. 2010, 5, 1100−
1
(
1
(
111.
19) Wang, S.; Li, Y.; Liu, H. Acta. Chim. Sinica 2012, 70, 1897−
903.
20) Huang, Z.; Cui, F.; Kang, H.; Chen, J.; Zhang, X.; Xia, C. Chem.
Mater. 2008, 20, 5090−5099.
21) Huang, Z.; Cui, F.; Xue, J.; Zuo, J.; Chen, J.; Xia, C. Catal. Today
012, 183, 42−51.
(22) Xia, S.; Yuan, Z.; Wang, L.; Chen, P.; Hou, Z. Appl. Catal., A
2011, 403, 173−182.
23) Xia, S.; Nie, R.; Lu, X.; Wang, L.; Chen, P.; Hou, Z. J. Catal.
(
2
(
2
(
2
(
2
2
2
012, 296, 1−11.
24) Zhou, J.; Zhang, J.; Guo, X.; Mao, J.; Zhang, S. Green Chem.
012, 14, 156−163.
25) Maris, E. P.; Ketchie, W. C.; Murayama, M.; Davis, R. J. J. Catal.
007, 251, 281−294.
(26) Montassier, C.; Men
Barbier, J. J. Mol. Catal. 1991, 70, 99−110.
difference in the adsorption strength of the glycerol and
glyceraldehyde intermediate, and so on, as a result of the
different availability of their unoccupied d orbitals. These
results and their understanding provide directions toward
design of more effective metal catalysts for selective hydro-
genolysis of glycerol and also other biomass-derived polyols.
́ ́
ezo, J. C.; Hoang, L. C.; Renaud, C.;
2
120
dx.doi.org/10.1021/cs400486z | ACS Catal. 2013, 3, 2112−2121

ACS Catalysis
Research Article
(
27) Wang, J.; Shen, S.; Li, B.; Lin, H.; Yuan, Y. Chem. Lett. 2009, 38,
72−573.
28) Alonso, D. M.; Wettstein, S. G.; Dumesic, J. A. Chem. Soc. Rev.
012, 41, 8075−8098.
29) van Ryneveld, E.; Mahomed, A. S.; van Heerden, P. S.;
Friedrich, H. B. Catal. Lett. 2011, 141, 958−967.
30) Jiang, T.; Zhou, Y.; Liang, S.; Liu, H.; Han, B. Green Chem.
009, 11, 1000−1006.
31) Musolino, M. G.; Scarpino, L. A.; Mauriello, F.; Pietropaolo, R.
ChemSusChem 2011, 4, 1143−1150.
32) Miyazawa, T.; Koso, S.; Kunimori, K.; Tomishige, K. Appl.
Catal., A 2007, 329, 30−35.
33) Alhanash, A.; Kozhevnikova, E. F.; Kozhevnikov, I. V. Catal.
5
(
2
(
(
2
(
(
(
Lett. 2008, 120, 307−311.
(
(
34) Oh, J.; Dash, S.; Lee, H. Green Chem. 2011, 13, 2004−2007.
35) Liu, L.; Zhang, Y.; Wang, A.; Zhang, T. Chin. J. Catal. 2012, 33,
1
(
2
(
257−1261.
36) Li, W.; Huang, H.; Li, H.; Zhang, W.; Liu, H. Langmuir 2008,
4, 8358−8366.
37) Choi, M.; Wu, Z.; Iglesia, E. J. Am. Chem. Soc. 2010, 132, 9129−
137.
9
(
(
38) Martin, D.; Duprez, D. J. Phys. Chem. 1996, 100, 9429−9438.
39) Vasiliadou, E. S.; Heracleous, E.; Vasalos, I. A.; Lemonidou, A.
A. Appl. Catal., B 2009, 92, 90−99.
40) Deutsch, K. L.; Lahr, D. G.; Shanks, B. H. Green Chem. 2012, 14,
635−1642.
41) Wang, K.; Hawley, M. C.; Furney, T. D. Ind. Eng. Chem. Res.
995, 34, 3766−3770.
42) Roy, D.; Subramaniam, B.; Chaudhari, R. V. Catal. Today 2010,
56, 31−37.
43) Ukisu, Y.; Miyadera, T. React. Kinet. Catal. Lett. 2004, 81, 305−
11.
44) Pham, T. T.; Lobban, L. L.; Resasco, D. E.; Mallinson, R. G. J.
Catal. 2009, 266, 9−14.
45) (a) Wang, S.; Temel, B.; Shen, J.; Jones, G.; Grabow, L. C.;
(
1
(
1
(
1
(
3
(
(
Studt, F.; Bligaard, T.; Abild-Pedersen, F.; Christensen, C. H.;
Nørskov, J. K. Catal. Lett. 2011, 141, 370−373. (b) CatApp; Suncat,
SLAC: Menlo Park, CA, 2011; http://suncat.slac.stanford.edu/catapp/
.
(
1
(
1
46) Abild-Pedersen, F.; Andersson, M. P. Surf. Sci. 2007, 601, 1747−
753.
47) Mavrikakis, M.; Barteau, M. A. J. Mol. Catal. A 1998, 131, 135−
47.
48) Ozaki, A.; Aika, K. Catalytic Activation of Dinitrogen. In
(
Catalysis: Science and Technology; Anderson, R. H., Boudart, M., Eds.;
Springer Verlag: Berlin, Germany, 1981; Vol. 1, pp 87−158.
(
́
49) Shi, H.; Gutierrez, O. Y.; Yang, H.; Browning, N. D.; Haller, G.
L.; Lercher, J. A. ACS Catal. 2013, 3, 328−338.
(
(
(
50) Murzin, D. Y. J. Catal. 2010, 276, 85−91.
51) van Santen, R. A. Acc. Chem. Res. 2009, 42, 57−66.
52) Peereboom, L.; Jackson, J. E.; Miller, D. J. Green Chem. 2009,
1
1, 1979−1986.
2
121
dx.doi.org/10.1021/cs400486z | ACS Catal. 2013, 3, 2112−2121