Aqueous-Phase Hydrogenation of Saturated and Unsaturated Ketones and Aldehydes over Supported Platinum-Rhenium Catalysts

Article abstract of DOI:10.1002/cctc.201501293

The hydrogenation rates of C=C and C=O bonds in methyl vinyl ketone and crotonaldehyde were measured over a series of silica-supported Pt-Re catalysts (1:1 atomic ratio of Pt/Re) in liquid water with H₂ (15 psig) at 333 K. The hydrogenation of methyl vinyl ketone did not produce any unsaturated alcohol because of the rapid hydrogenation of C=C relative to that of C=O. The addition of Re to Pt impacted the rate of C=C hydrogenation negatively in methyl vinyl ketone and crotonaldehyde, but promoted the selectivity of C=O hydrogenation in crotonaldehyde in which the unsaturated alcohol increased from 5 % on Pt to 21 % on Pt-Re. The addition of Re to Pt also promoted the rate of C=O hydrogenation in 2-butanone, whereas little effect was observed during the hydrogenation of butanal. The results of electron microscopy and H₂ chemisorption on the Pt-Re catalysts showed the increasing interaction between Pt and Re with the increasing metal weight loading, and results from rate measurements suggest that oxophilic Re can be used to promote the Pt-catalyzed hydrogenation of carbonyl groups in multifunctional molecules.

Full text of DOI:10.1002/cctc.201501293

DOI: 10.1002/cctc.201501293
Full Papers
Aqueous-Phase Hydrogenation of Saturated and
Unsaturated Ketones and Aldehydes over Supported
Platinum–Rhenium Catalysts
Derek D. Falcone, John H. Hack, and Robert J. Davis*[a]
The hydrogenation rates of C=C and C=O bonds in methyl
vinyl ketone and crotonaldehyde were measured over a series
of silica-supported Pt-Re catalysts (1:1 atomic ratio of Pt/Re) in
liquid water with H2 (15 psig) at 333 K. The hydrogenation of
methyl vinyl ketone did not produce any unsaturated alcohol
because of the rapid hydrogenation of C=C relative to that of
C=O. The addition of Re to Pt impacted the rate of C=C hydro-
genation negatively in methyl vinyl ketone and crotonalde-
hyde, but promoted the selectivity of C=O hydrogenation in
crotonaldehyde in which the unsaturated alcohol increased
from 5% on Pt to 21% on Pt-Re. The addition of Re to Pt also
promoted the rate of C=O hydrogenation in 2-butanone,
whereas little effect was observed during the hydrogenation of
butanal. The results of electron microscopy and H2 chemisorp-
tion on the Pt-Re catalysts showed the increasing interaction
between Pt and Re with the increasing metal weight loading,
and results from rate measurements suggest that oxophilic Re
can be used to promote the Pt-catalyzed hydrogenation of car-
bonyl groups in multifunctional molecules.
Introduction
The selective hydrogenation of C=C and C=O bonds in a,b-un-
saturated ketones and aldehydes, such as methyl vinyl ketone
(MVK) and crotonaldehyde, has been studied widely for appli-
cations in the pharmaceutical, cosmetic, and food indus-
alysts has been observed to have a similar effect as Sn addi-
tion.[10,11]
The promotional effects of metallic additives such as Sn and
Fe are explained in two ways. If Sn is added to Pt, PtdÀSnd+ en-
tries.[1–5] Selective hydrogenation of the C=O bond over the C= sembles are formed,[6,7,12] and the polarity between the two
C bond to obtain unsaturated alcohols is preferred, but the hy-
drogenation of aliphatic C=C bonds is favored thermodynami-
cally, and typically favored kinetically, over that of C=O bonds.
Therefore, catalysts must be designed to kinetically control the
hydrogenation of C=O bonds over C=C bonds.[2,5]
metals increases the binding energy of C=O to allow the C=O
bond to be hydrogenated selectively. Concepción et al. used IR
spectroscopy to suggest that a unique CO adsorption mode is
present on Pt impregnated into Sn-Beta, which may facilitate
the adsorption of aldehydes onto a bimetallic ensemble to
allow the bond to be more easily activated by hydrogen.[6] Ad-
ditionally, the results of X-ray photoelectron spectroscopy
(XPS) revealed that surface Sn was still oxidized even after a re-
duction at 623 K, which supports the existence of Sn with
a partial positive charge.[7] A second explanation for the pro-
motional effect is that the addition of Sn dilutes Pt ensembles
that are highly active for C=C hydrogenation.[3,7] Passos et al.
demonstrated with results from microcalorimetry experiments
that the integral heats of adsorption of H2 and CO on Pt-Sn/
Al2O3 were approximately the same as that on Pt/Al2O3,[13]
which suggests that the heat of adsorption of aldehyde
groups may be the same on Pt-Sn catalysts as on Pt.[3] Even
though the addition of Sn to Pt and Ru has been shown to in-
crease the selectivity of unsaturated aldehyde hydrogenation
to unsaturated alcohol, Sn addition does not appear to in-
crease the selectivity of unsaturated ketone hydrogenation to
unsaturated alcohols.[2,12]
Monometallic transition-metal catalysts have a very high ac-
tivity for hydrogenation reactions, but have a very high selec-
tivity to the undesired saturated aldehydes and ketones.[2] Sev-
eral bimetallic formulations have increased the selectivity to
the desired unsaturated alcohol.[3,6–9] Tin is one of the most
common transition-metal promoters for hydrogenation reac-
tions and has been shown to increase the selectivity and activi-
ty for the hydrogenation of several a,b-unsaturated aldehydes
on Pt- and Ru-based catalysts. For example, the addition of Sn
to Pt increased the turnover frequency (TOF) of crotonalde-
hyde hydrogenation by over 12 times and increased the selec-
tivity of crotyl alcohol from close to zero on a Pt catalyst to
over 40% on a Pt-Sn catalyst.[7] Galvagno et al. demonstrated
that a positive correlation existed between the Sn/Ru ratio of
the carbon-supported bimetallic nanoparticles and the rate of
cinnamaldehyde hydrogenation.[9] The addition of Fe to Pt cat-
[a] Dr. D. D. Falcone, J. H. Hack, Prof. R. J. Davis
Department of Chemical Engineering
University of Virginia
An increase of the selectivity of unsaturated aldehyde hydro-
genation to unsaturated alcohols has also been achieved by
supporting late-transition metals on reducible supports. Dan-
dekar and Vannice demonstrated that titania-supported Pt pro-
moted the rate and the selectivity of crotonaldehyde hydroge-
102 Engineers’ Way, PO Box 400741
Charlottesville, VA 22904 (USA)
E-mail: rjd4f@virginia.edu
ChemCatChem 2016, 8, 1074 – 1083
1074
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
nation only if the catalyst was reduced above 773 K.[14] It is
well known that if reducible supports such as titania are re-
duced at high temperatures, TiOx species can migrate onto the
late-transition-metal nanoparticles and interact more strongly
with the Pt-group metals.[15] Dandekar and Vannice attributed
the increase in the rate and selectivity of crotonaldehyde hy-
drogenation to Pt metal-interfacial titania sites. The coordina-
tively unsaturated Ti cations can interact more strongly with
the C=O bond in crotonaldehyde, which increases the rate of
reaction and selectivity to crotyl alcohol.[14]
light on the kinetics of polyol hydrogenolysis over bimetallic
catalysts that consist of an early- and a late-transition metal.
In this work, the rate of the hydrogenation of molecules that
contain C=O and sometimes C=C bonds was investigated. The
hydrogenation rate was normalized to H uptake, measured by
H2 chemisorption, the surface average particle size, obtained
by using TEM, and total metal atoms to better quantify the
effect of Re on Pt-catalyzed rates. In addition, several experi-
ments were performed to evaluate the significance of mass
transfer limitations on the measured conversion to ensure that
rate measurements were obtained in the kinetically controlled
regime.
Recently, Tamura et al. reported that the addition of Re to an
Ir-based catalyst increased the conversion of crotonaldehyde
hydrogenation and the resulting selectivity to crotyl alcohol to
above 90%.[16] The monometallic Re catalyst was essentially in-
active at 303 K and 8 MPa of H2, and the selectivity to crotyl al- Results and Discussion
cohol over monometallic Ir was approximately 60%. They as-
Catalyst characterization
cribe the increased conversion and selectivity to crotyl alcohol
over Ir-Re to the ability of the catalysts to perform the hetero-
lytic cleavage of H2.[16] It has also been hypothesized that the
heterolytic cleavage of H2 explains the high rate and selectivity
of polyol (CÀO bond) hydrogenolysis over catalysts that con-
tain Re and a Pt-group metal.[17,18] The Ir-Re,[19,20] Rh-Re,[17,21,22]
and Pt-Re[23] bimetallic systems exhibit high rates of polyol hy-
drogenolysis and high selectivity for the activation of secon-
dary CÀO bonds. Interestingly, the Ir-Re, Rh-Re, Pt-Re, Pd-Re,
and Ru-Re systems were evaluated for the hydrogenation of
crotonaldehyde by Tamura et al. and only the Ir-Re system
showed a higher conversion and selectivity than the monome-
tallic catalysts.[16]
The Pt and Re loadings, H2 chemisorption results, and particle
size analysis from TEM for the silica-supported catalysts used in
this work are provided in Table 1. The 8 wt% Pt-Re catalyst
was characterized extensively and tested for a number of reac-
tions in our previous work.[24] The H/Pt ratio given in Table 1 is
a measure of the exposed Pt atoms per total Pt atoms. For
supported monometallic Pt catalysts, the particle size [nm] can
be estimated from the inverse of H/Pt to yield an average par-
ticle size of 2.2 and 3.3 nm for the 4 and 8 wt% Pt catalysts, re-
spectively.[25] A relationship between H/Pt and the average par-
ticle size is not well known for the Pt-Re catalyst, however, be-
cause the average number of adsorbed hydrogen atoms to
surface Re atoms in these samples is unknown. Daniel et al.
observed that there was no hydrogen uptake for a monometal-
lic Re catalyst supported on Norit carbon in which the average
Re oxidation state was measured to be 7+ by using X-ray ab-
sorption spectroscopy.[23] Characterization of the Pt-Re catalyst
revealed that the average Re oxidation state for the silica-sup-
ported Pt-Re catalyst is between 2+ and 4+, and it is unknown
whether hydrogen spillover from surface Pt to surface Re can
occur on these materials.[24] Nevertheless, for the Pt-Re cata-
lysts listed in Table 1, the H2 chemisorption results give an esti-
mation of the exposed Pt.
Our previous work on the hydrogenolysis of polyols over Pt-
Re suggests that the high rates and selectivities of glycerol hy-
drogenolysis are explained by a bifunctional mechanism in
which the protolytic activation of alcohols by a Brønsted acid
site associated with the Re component of the catalyst is fol-
lowed by the rapid hydrogenation of the unsaturated inter-
mediate, which contains either a C=C or C=O bond, on the Pt
component of the catalyst.[24] For glycerol hydrogenolysis over
Pt-Re, the order in H2 was 0.5 at high glycerol concentrations.
In classical hydrogenation kinetics on late-transition metals,
a half order dependence in H2 corresponds to the equilibrated
adsorption of H2 on a surface that catalyzes hydrogen addition
reactions. Thus, an understanding of the kinetics of C=C and
C=O bond hydrogenation on Pt-Re is required to shed further
Particle size analysis by using TEM revealed that the average
particle size and the standard deviation for the particle size dis-
tribution for the Pt-Re catalysts did not change as a function
Table 1. Metal weight loadings, results from H2 chemisorption, and TEM analysis of silica-supported Pt, Re, and Pt-Re catalysts.
Catalyst
Pt[a]
[wt%]
Re[a]
[wt%]
H uptake[b]
[mmolgÀ1
]
H/Pt[b]
Average diameter[c]
[nm]
Surface average
diameter[d] [nm]
1 wt% Pt
4 wt% Pt
8 wt% Pt
1 wt% Pt-Re
2 wt% Pt-Re
4 wt% Pt-Re
8 wt% Pt-Re
1.0
4.0
8.0
1.0
2.0
4.0
8.0
–
–
–
0.9
1.9
3.6
7.6
41
92
122
41
75
0.84
0.45
0.30
0.80
0.73
0.49
0.42
1.7Æ0.75
2Æ1.0
2.5
3.2
8.0
2.3
3.1
2.5
2.1
4Æ2.3
1.9Æ0.56
2.3Æ0.85
1.8Æ0.76
1.6Æ0.59
100
172
[a] Nominal weight loadings [b] Obtained from H2 chemisorption at 308 K after the catalysts were reduced in situ at 473 K [c] Obtained from TEM [d] Ob-
tained from TEM by Sd3/Sd2.
ChemCatChem 2016, 8, 1074 – 1083
www.chemcatchem.org
1075
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
not be obtained. The characterization of a silica-supported
monometallic Re catalyst synthesized by the incipient wetness
impregnation of an aqueous metal precursor suggests that Re
is held weakly and is highly dispersed on silica.[28] In this work,
the resolution of the TEM was not sufficient to observe atomi-
cally dispersed Re, which explains the low number of particles
observed.
Hydrogenation of C=C bonds in MVK and crotonaldehyde
The potential influence of H2 mass transfer limitations in the
batch reactor were first evaluated by studying the hydrogena-
tion of MVK over monometallic Pt, which consumed H2 at the
fastest rate for any reaction in this work. The rates of MVK hy-
drogenation normalized to H uptake sites (TOFH), rates normal-
ized to the amount of total surface metal atoms calculated
from the inverse of the surface average particle size from TEM
(TOFSM), and the global rate per total amount of metal are
listed in Table 2 for the monometallic Pt catalysts at 333 K. Reli-
able rates can be obtained at the high levels of conversion
listed in Table 2 because the hydrogenation of MVK is typically
zero order in the substrate over transition-metal catalysts.[2]
The reaction profiles of MVK hydrogenation over the Pt and
Pt-Re catalysts are shown in Figure 2. The rate of reaction did
not change as the reaction approached 100% conversion in
MVK, which confirms the zero-order behavior. The rates of
Figure 1. Transmission electron micrographs of a) 1 wt% Pt, b) 8 wt% Pt,
c) 1 wt% Pt-Re, and d) 8 wt% Pt-Re.
of the metal loading. The local density of metal particles
simply increased for higher loadings of Pt and Re (Figure 1c
and d). The 8 wt% Pt catalyst, however, had a significantly
higher average particle size and standard deviation
than the other catalysts studied in this work. Some
Table 2. Testing the mass transfer limitations of MVK hydrogenation over monometal-
lic Pt at 333 K.
particles as large as 20 nm were observed on the
8 wt% Pt sample (not shown), whereas the largest
particles observed in the other samples studied were
only as large as ~7 nm. The large particle size for
8 wt% Pt was a result of the high metal loading,
which increases the likelihood of sintering during cat-
alyst pretreatment. The 8 wt% Pt-Re catalyst had
a smaller average particle size than monometallic
8 wt% Pt as the addition of Re increases the stability
of Pt and mitigates sintering.[26] Deng et al. synthe-
sized a number of Pt-Re catalysts with a molar ratio
of unity supported on carbon nanotubes that had
loadings of Pt from 1 to 30 wt%.[27] Similar to our re-
sults, the average particle size measured by using
TEM increased from 1.5 nm on a 1 wt% Pt-Re catalyst
[a]
[b]
Catalyst Catalyst mass Conversion TOFH TOFSM Rate
Selectivity
[%]
[mg]
[%]
[sÀ1
]
[sÀ1
]
[mmolgmetalÀ1 sÀ1
]
SK SA UA
1 wt% Pt 10
21
22
18
57
0.54
0.52
0.49
0.43
1.1
0.75
1.2
2300
1200
750
100
99
99
0
1
1
3
0
0
0
0
4 wt% Pt 10
8 wt% Pt 10
8 wt% Pt 20
1.0
660
97
Reaction conditions: 333 K, 15 psig H2, 45 cm3 of 0.2m aqueous MVK solution. In situ
pretreatment: 393 K, 15 psig H2 for 1 h. [a] Normalized by H chemisorption uptake.
[b] Normalized by surface metal atoms (Pt+Re) evaluated from the inverse of the
average particle size obtained from TEM.
to only 2.3 nm on a 5 wt% Pt-Re catalyst.[27] The 30 wt% cata-
lyst, however, had an average particle size of 4.9 nm.[27] Inter-
estingly, the XPS characterization revealed that the surface Pt
to Re ratio was approximately unity for the low-weight-loaded
catalysts, but Pt became more enriched at the surface for
higher weight loadings.[27]
An electron micrograph of a monometallic Re catalyst did
not contain a narrow distribution of particle sizes as observed
with the Pt and Pt-Re catalysts (not shown). Instead, TEM
images revealed a bimodal distribution that consisted of ex-
tremely large particles, as large as 50 nm in diameter, and
smaller particles roughly 4–6 nm in diameter. In addition, very
few particles with sufficient contrast were observed for the
4 wt% Re catalyst, and a reliable average particle size could
Figure 2. Conversion of MVK hydrogenation over monometallic Pt and Pt-Re
catalysts as a function of time.
ChemCatChem 2016, 8, 1074 – 1083
www.chemcatchem.org
1076
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
MVK hydrogenation given in Table 2 were approximately the
same if 10 mg of 1, 4, or 8 wt% Pt was used. In addition,
20 mg of 8 wt% Pt, which converted 56.6% of the MVK after
30 min of reaction, had the same normalized rate of MVK hy-
drogenation as 10 mg of 8 wt% Pt, which confirms that exter-
nal mass transfer effects were not present in the system. The
small change in rate (approximately 30%) as the amount of
metal in the reactor varied by a factor of approximately 16 in-
dicates that external and internal mass transfer limitations did
not affect the measured rates.[29]
C=O and C=C bonds with four Ru atoms. After coordination
through this planar h4 mode, hydrogen can then add to either
carbon in the adsorbed C=C moiety in addition to the carbon
or oxygen atom in the ketone group. The results of the quan-
tum chemical calculations presented by Ide et al. showed that
addition of hydrogen to the carbon atom of the carbonyl had
the lowest barrier of each of the four possible atoms. The sec-
ondary carbon atom in the C=C bond had the next lowest bar-
rier for the initial hydrogen addition. The lower barrier for the
carbonyl carbon atom was ascribed to the enhanced stabiliza-
tion of the transition state by the strong RuÀO bond. The addi-
tion of a second hydrogen atom can then occur on the C=C
bond to create the intermediate for the SK or the C=O bond
to form the UA. The activation barrier to add the initial hydro-
gen to the C=O bond was 10 kJmolÀ1 lower than that of the
addition to the C=C bond, but this is not reflected in the selec-
tivity trends over Ru or Pt, in which over 95% of the product is
the SK because the desorption barrier for the intermediate
leading to the UA (113 kJmolÀ1) is significantly higher than the
desorption barrier for the SK intermediate (28 kJmolÀ1) over Ru
(0001).
The selectivity of MVK hydrogenation for the monometallic
Pt catalyst also did not change significantly for different Pt
loadings (Table 2). Approximately 99% of the product was the
saturated ketone (SK) formed by the hydrogenation of the C=C
bond. The remaining product was the saturated alcohol (SA),
which is a result of the sequential hydrogenation of the SK as
depicted in Scheme 1. The production of the unsaturated alco-
The work by Ide et al. highlighted the difficulty in the selec-
tive hydrogenation of unsaturated ketones to the unsaturated
alcohols with monometallic catalysts.[2] There are few reports
of metal catalysts that can hydrogenate MVK to the UA selec-
tively.[30] The production of UA from MVK, however, has been
demonstrated with homogeneous borohydrides[4,31] and over
MgO.[4,32] In the case of MgO, hydrogen was not supplied with
H2 but with a hydrogen donor molecule, such as 2-propanol.
Magnesia was able to form the UA with high selectivity (69%
at 78% conversion) as the barriers to adsorb C=C bond to
MgO are very high.[32] To the best of our knowledge, the Pt-Re
catalyst has not yet been tested for this reaction, and it is pos-
sible that the presence of oxophilic Re on the surface may
preferentially adsorb the C=O bond.
Scheme 1. Reaction network for MVK hydrogenation.
hol (UA) was below the detectable limit for MVK hydrogena-
tion over Pt. Ide et al. studied the hydrogenation of MVK over
a carbon-supported Pt catalyst in liquid water and also did not
observe the production of the UA and the selectivities to SK
and SA that they observed agree well with the results given in
Table 2.[2] The rates for MVK hydrogenation in liquid water pre-
sented in Table 2 are also similar to the TOF (normalized by H2
chemisorption) of 0.94 sÀ1 reported by Ide et al.[2] The differ-
ence between the TOFH and TOFSM given in Table 2 can be as-
cribed to a discrepancy between the techniques used to mea-
sure active sites. It is possible that atomically dispersed Pt or
metal particles less than 0.4 nm may be present on the silica-
supported catalysts, which would not be detected by TEM
analysis because of the resolution limits of the technique.
Thus, the true surface average particle size may be smaller
than that measured by using TEM, which would decrease the
TOFSM given in Table 2 and make it more comparable to the
TOFH.
The results of MVK hydrogenation over the Pt-Re catalysts
are presented in Table 3. The TOFSM of MVK hydrogenation
over 1 wt% Pt-Re agrees well with the TOF values reported in
Table 2 for the monometallic Pt catalyst. As the metal loadings
increased, the TOFH, TOFSM, and the rate per total amount of
Table 3. Reaction results for the hydrogenation of MVK over Pt-Re cata-
lysts with different weight loadings of metal and a Pt to Re molar ratio of
unity.
[a]
[b]
Catalyst
Conversion TOFH TOFSM Rate
Selectivity
[%]
[%]
[sÀ1
]
[sÀ1
]
[mmolgmetalÀ1 sÀ1
]
SK SA UA
1 wt% Pt-Re 30
2 wt% Pt-Re 19
4 wt% Pt-Re 19
8 wt% Pt-Re 30
1.2
1.2
2700
1300
660
99
99
99
98
1
1
1
2
0
0
0
0
0.70
0.50
0.37
0.80
0.31
0.16
Ide et al. studied the hydrogenation of MVK over carbon-
supported Pd, Pt, Ru, and Au catalysts and did not report the
formation of UA regardless of the metal used. This result was
explained by using quantum chemical calculations for the reac-
tion that occurred on a Ru (0001) surface.[2] For the most
stable adsorption mode of MVK on Ru (0001), the molecule is
parallel to the surface so bonding can occur through both the
410
Reaction conditions: 333 K, 15 psig H2, 45 cm3 of 0.2m aqueous MVK solu-
tion. In situ pretreatment: 393 K, 15 psig H2 for 1 h. [a] Normalized by H
chemisorption uptake. [b] Normalized by surface metal atoms (Pt+Re)
evaluated from the inverse of the average particle size obtained from
TEM.
ChemCatChem 2016, 8, 1074 – 1083
www.chemcatchem.org
1077
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
metal for MVK hydrogenation appeared to decrease over the
Pt-Re catalysts. Nanoparticles on the low-weight-loaded Pt-Re
catalyst sparsely populate the silica support (Figure 1). As the
average particle size of 1 wt% Pt-Re and 8 wt% Pt-Re are es-
sentially the same, it follows logically that the nanoparticles on
8 wt% Pt-Re would be very densely populated as confirmed
by the results shown in Figure 1d. It is well known that Re can
be highly dispersed on metal oxide supports such as silica,[28]
and it is possible that a larger proportion of the Re is more
widely dispersed on the silica support in 1 wt% Pt-Re than in
8 wt% Pt-Re and not associated with Pt nanoparticles. In other
words, Re is more likely to cover Pt hydrogenation sites on
8 wt% Pt-Re than 1 wt% Pt-Re as the nanoparticles are more
densely packed at high weight loadings. Consistent with this
idea, H2 chemisorption measured a higher fraction of exposed
Pt on 1 wt% Pt-Re than 8 wt% Pt-Re and in general, as the Pt-
Re weight loading increased the fraction of Pt exposed de-
creased (Table 1). Thus, the lower rate of MVK hydrogenation
over the high-weight-loaded Pt-Re catalysts can be ascribed to
the blocking of the Pt hydrogenation sites by Re. The addition
of Re does not appear to change the selectivity of MVK hydro-
genation (Table 3).
Table 4. Reaction results for the hydrogenation of crotonaldehyde after
30 min of reaction over 4 wt% Pt and the Pt-Re catalysts with a Pt to Re
molar ratio of unity for the bimetallic catalysts.
[a]
[b]
Catalyst
Conversion TOFH TOFSM Rate
Selectivity
[%]
[%]
[sÀ1
]
[sÀ1
]
[mmolgmetalÀ1 sÀ1
]
SAL SA UA
4 wt% Pt
2.5
0.082 0.12
0.13 0.12
0.075 0.085
0.092 0.056
0.081 0.036
190
270
140
120
89
92
82
86
3
8
7
5
10
7
1 wt% Pt-Re 1.4
2 wt% Pt-Re 2.3
4 wt% Pt-Re 2.7
8 wt% Pt-Re 4.6
69 11 20
74 21
5
Reaction conditions: 333 K, 15 psig H2, 45 cm3 of 0.2m aqueous crotonal-
dehyde solution. In situ pretreatment: 393 K, 15 psig H2 for 1 h. [a] Nor-
malized by H chemisorption uptake. [b] Normalized by surface metal
atoms (Pt+Re) evaluated from the inverse of the average particle size ob-
tained from TEM.
the Pt-Re metal weight loading increased. Similar to MVK hy-
drogenation, the overall rate of crotonaldehyde hydrogenation
over the Pt-Re catalyst was comparable to the rate over the
monometallic Pt catalyst, and the TOFSM and rate per total
amount of metal for crotonaldehyde hydrogenation appeared
to decrease as the Pt-Re weight loading increased. The slower
rate of crotonaldehyde hydrogenation compared to MVK hy-
drogenation over Pt is ascribed to the steric hindrance of the
C=C bond in crotonaldehyde by the methyl group.[2,4,33] Never-
theless, the TOFH for 4 wt% Pt/SiO2 agrees closely with the
TOF of 0.13 sÀ1 under similar conditions reported by Ide et al.
Tamura et al. explained the increased selectivity to UA
during crotonaldehyde hydrogenation over an Ir-Re catalyst by
speculating that the presence of Re causes the heterolytic
cleavage of H2, which leads to the selective activation of polar
C=O bonds. The addition of hydride to C=O bonds by homo-
geneous metal hydride complexes has been reported in the lit-
erature.[34–37] The addition of Re or Mo to late-transition metals
was also claimed to lower the barrier to heterolytically activate
H2 to form a hydride that attacks polar bonds selectively, such
as the CÀO bond in glycerol.[17,38] Alternatively, the increased
rate of glycerol (CÀO bond) hydrogenolysis upon the addition
of Re to a Pt catalyst was discussed in terms of a bifunctional
mechanism in which a Brønsted acid associated with the Re
component of the catalyst protolytically activates the alcohol
of glycerol.[24] Nevertheless, hydride addition catalyzed by the
Re component of the catalyst may explain the selective hydro-
genation of C=O bonds.
An appreciable selectivity to the UA can be obtained more
easily during the selective hydrogenation of unsaturated alde-
hydes such as crotonaldehyde (Scheme 2). Marinelli et al. were
Scheme 2. Reaction network for crotonaldehyde hydrogenation.
able to obtain 13% selectivity to the unsaturated alcohol in
the gas phase over a Pt/SiO2 catalyst,[33] and Ide et al. demon-
strated that it is possible to obtain a selectivity as high as 37%
to crotyl alcohol in liquid water over an Au/C catalyst.[2]
Tamura et al. used a silica-supported Ir catalyst to achieve
63.7% selectivity to the unsaturated alcohol at 5.4% conver-
sion of crotonaldehyde.[16] The methyl group next to the C=O
bond in MVK apparently inhibits the selective hydrogenation
of MVK to the UA, but the low steric hindrance near the C=O
bond in crotonaldehyde evidently improves the selectivity of
crotyl alcohol from crotonaldehyde hydrogenation.[2,4]
Reyes et al. observed that the selectivity during cinnamalde-
hyde hydrogenation to cinnamyl alcohol increased from 70%
over Rh/SiO2 to almost 100% over a Rh-Mo/SiO2 catalyst.[39]
The results of the XPS characterization of the Rh-Mo catalysts
revealed several Mo oxide species and a Rhd+ species that was
only observed in the presence of Mo. The promotion of Rh by
Mo was thought to result in a Rhd+-MoOx species, which was
responsible for the increased selectivity to the UA.[39] As men-
tioned earlier, the presence of coordinatively unsaturated Ti
cations has also been invoked to explain the increased selectiv-
ity to the UA during crotonaldehyde hydrogenation over Pt/
TiO2 and Ni/TiO2 that were reduced at adequately high temper-
Tamura et al. observed that if a silica-supported Ir catalyst
was promoted by Re, the selectivity of crotonaldehyde hydro-
genation to the UA was as high as 92.0% at 43.0% conver-
sion.[16] If Re was used to promote silica-supported Pt, however,
only a moderate increase in the selectivity to the UA was ach-
ieved, and the highest selectivity to UA reached 21% (Table 4).
In general, the selectivity to the UA appeared to increase as
ChemCatChem 2016, 8, 1074 – 1083
www.chemcatchem.org
1078
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
atures to initiate the migration of Ti species onto the Pt or Ni
particles.[14] Our characterization of the Pt-Re catalysts by using
XPS indicated that Pt was essentially all metallic, although
there was a measurable shift for the core levels of surface Pt to
higher binding energy in the presence of Re.[24]
or 40 mg of 1 wt% Pt-Re was approximately the same, which
suggests that the poison did not affect the rate of reaction sig-
nificantly at short reaction times.
As the product selectivity consisted predominately of the SK
and the SAL for the hydrogenation of MVK and crotonalde-
hyde, respectively, the overall hydrogenation rates presented
in Tables 3 and 4 are mainly associated with the rate of C=C
bond hydrogenation. To understand the kinetics of glycerol hy-
drogenolysis, the hydrogenation of C=O bonds is likely to be
more relevant because a hypothetical intermediate that may
be formed after the dehydration of glycerol is most likely
a ketone (or aldehyde) as the equilibrium associated with
keto–enol tautomerization in liquid water is shifted away from
the enol.[41–43] Bohne et al. measured the [keto]/[enol] equilibri-
um constant of propanal and butanal in liquid water at 328 K
to be 1.3”105 and 1.8”105, respectively.[42] In addition, the
rate of tautomerization to acetone at 328 K in liquid water re-
ported by Chaing et al.[41] was approximately 70 sÀ1, which is
significantly faster than the hydrogenation of both C=C and
C=O and indicates that upon the formation of the proposed
enol intermediate during glycerol hydrogenolysis, it will rapidly
tautomerize to the keto (or aldehyde) form.
Even though a significant level of electron sharing was not
detected in Pt-Re nanoparticles by using XPS, we ascribe the
increase in the UA selectivity for crotonaldehyde hydrogena-
tion to the preferential adsorption of the C=O bond onto the
oxophilic Re component of the Pt-Re bimetallic catalyst. The
MVK hydrogenation results indicate that Re blocks Pt sites
more extensively on the high-weight-loaded Pt-Re catalysts,
which is also consistent with the selectivity trends in crotonal-
dehyde hydrogenation. The high-weight-loaded Pt-Re catalysts
exhibited a higher selectivity to UA than the low-weight-
loaded Pt-Re catalysts.
The results summarized in Table 4 were obtained after
30 min of reaction because a significant deactivation of the
catalysts was observed during the hydrogenation of crotonal-
dehyde over Pt and Pt-Re. The conversion of crotonaldehyde
during hydrogenation as a function of time over 10 mg of the
Pt-Re catalysts and 10 mg of the 4 wt% Pt catalyst is shown in
Figure 3. Deactivation, as indicated by the decrease in the
Hydrogenation of C=O bonds in 2-butanone and butanal
To study the rate of C=O bond hydrogenation in the absence
of C=C bonds, the hydrogenation of 2-butanone to 2-butanol
and butanal to 1-butanol was evaluated over Pt and Pt-Re. The
results for 2-butanone hydrogenation are presented in Table 5.
As seen for both MVK and crotonaldehyde hydrogenation, the
TOFH and TOFSM of 2-butanone hydrogenation over the low-
weight-loaded Pt-Re catalyst was similar to that of 2-butanone
hydrogenation over a monometallic Pt catalyst. In the case of
2-butantone, however, upon increasing the weight loading of
Pt-Re, which likely increased the Re/Pt ratio at the surface of
the nanoparticles, the TOFH of 2-butanone hydrogenation in-
creased by a factor of approximately five (Table 5). Although
a higher Re content on the surface appears to decrease the
rate of C=C bond hydrogenation (as shown with MVK and cro-
tonaldehyde), the rate of 2-butanone hydrogenation and the
Figure 3. Conversion during crotonaldehyde hydrogenation over monome-
tallic Pt and bimetallic Pt-Re catalysts. The mass of catalyst was 10 mg
unless otherwise stated.
Table 5. Reaction results for the hydrogenation of 2-butanone over
4 wt% Pt and the Pt-Re catalysts with a Pt to Re molar ratio of unity for
the bimetallic catalysts.
slope of the curve, was significant over 4 wt% Pt and the low-
weight-loaded Pt-Re catalysts. The 8 wt% Pt-Re catalyst did
not appear to deactivate as quickly. A significant level of deac-
tivation was not observed during the hydrogenation of MVK
over these catalysts (Figure 2). Englisch et al. also observed de-
activation during the hydrogenation of crotonaldehyde in
liquid water over Pt/SiO2.[40] They found that the initial rate of
hydrogenation was restored by the treatment of the used cata-
lyst in air.[40] If 40 mg of 1 wt% Pt-Re was used as the catalyst,
however, the deactivation behavior was similar to that of the
8 wt% catalyst. Apparently, the deactivation rate is related to
the catalyst/substrate ratio, which suggests that a poisoning
species may be present in the crotonaldehyde feed. Neverthe-
less, the rate of hydrogenation after 30 min of reaction for 10
[a]
[b]
Catalyst
Conversion
[%]
TOFH
[sÀ1
TOFSM
[sÀ1
Rate
[mmolgmetalÀ1 sÀ1
]
]
]
4 wt% Pt
29
0.0036
0.0053
0.0081
0.014
0.0052
0.0050
0.0092
0.0088
0.013
8.3
11
16
18
32
1 wt% Pt-Re
2 wt% Pt-Re
4 wt% Pt-Re
8 wt% Pt-Re
0.7
1.6
3.8
10
0.029
Reaction conditions: 333 K, 15 psig H2, 45 cm3 of 0.2m aqueous 2-buta-
none solution. In situ pretreatment: 393 K, 15 psig H2 for 1 h. [a] Normal-
ized by H chemisorption uptake. [b] Normalized by surface metal atoms
(Pt+Re) evaluated from the inverse of the average particle size obtained
from TEM. The product selectivity was 100% to the saturated alcohol.
ChemCatChem 2016, 8, 1074 – 1083
www.chemcatchem.org
1079
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
selectivity trends for crotonaldehyde hydrogenation in Table 4
suggest that more Re on the surface promotes the rate of C=O
bond hydrogenation.
The results from the hydrogenation of butanal to 1-butanol
are presented after 1 h of reaction in Table 6. The results are
presented after the same reaction time because deactivation
was observed, similar to that found in crotonaldehyde hydro-
genation (Figure 3). The results of butanal hydrogenation over
Figure 4. Schematic representation of the Pt-Re catalysts used for hydroge-
nation reactions.
Table 6. Reaction results for the hydrogenation of butanal after 60 min
of reaction over 4 wt% Pt and the Pt-Re catalysts with a Pt to Re molar
ratio of unity for the bimetallic catalysts.
packed on the silica support to result evidently in the more in-
timate mixing of Pt and Re, which created bimetallic ensem-
bles. The decoration of Re on the Pt nanoparticles blocked Pt
sites that catalyzed C=C bond hydrogenation to result in lower
rates of MVK and crotonaldehyde hydrogenation over the
high-weight-loaded Pt-Re catalysts (Tables 3 and 4) relative to
monometallic Pt. The presence of oxophilic Re in close contact
with Pt on the high-weight-loaded catalysts enhanced the rate
of C=O hydrogenation in 2-butanone. The Re may have facili-
tated the preferential adsorption of C=O in 2-butanone to lead
to an increased rate of hydrogenation by hydrogen spillover
from the Pt component of the catalyst. The increased selectivi-
ty to the UA during crotonaldehyde hydrogenation appeared
to be the result of the selective inhibition of C=C hydrogena-
tion by Re without the inhibition of C=O hydrogenation. For
the 8 wt% Pt-Re catalyst, the hydrogenation TOFH for crotonal-
dehyde (which was a measure of mostly C=C bond hydrogena-
tion) and butanal (which was a measure C=O bond hydrogena-
tion) was approximately the same, 0.036 and 0.032 sÀ1, respec-
tively, which explains the enhanced selectivity of crotonalde-
hyde hydrogenation to the unsaturated alcohol upon the addi-
tion of Re.
[a]
[b]
Catalyst
Conversion
[%]
TOFH
[sÀ1
TOFSM
[sÀ1
Rate
[mmolgmetalÀ1 sÀ1
]
]
]
4 wt% Pt
4 wt% Pt-Re
8 wt% Pt-Re
2.2
3.4
3.7
0.016
0.038
0.032
0.023
0.024
0.014
37
50
35
Reaction conditions: 333 K, 15 psig H2, 45 cm3 of 0.2m aqueous butanal
solution. In situ pretreatment: 393 K, 15 psig H2 for 1 h. [a] Normalized by
H chemisorption uptake. [b] Normalized by surface metal atoms (Pt+Re)
evaluated from the inverse of the average particle size obtained from
TEM. The product selectivity was 100% to the saturated alcohol.
the low-weight-loaded catalysts are not reported because relia-
ble levels of conversion were not obtained after 1 h of reac-
tion. A comparison of the rates of 2-butanone hydrogenation
(Table 5) and butanal hydrogenation (Table 6) over monometal-
lic Pt indicates that the hydrogenation of the aldehyde is sub-
stantially faster than that of the ketone. The TOFH of butanal
hydrogenation over the high-weight-loaded Pt-Re catalysts
was slightly higher than that over the monometallic Pt catalyst,
which agrees with results from 2-butanone hydrogenation as
reported in Table 5. Unlike the case with 2-butanone, no dis-
cernable trend in TOFSM or global rate with the weight loading
of metals was observed with butanal. The rate of butanal hy-
drogenation did not seem to be affected by the interaction be-
tween Pt and Re as significantly as that of the other molecules
studied (MVK, crotonaldehyde, and 2-butanone). Thus, the in-
teraction of Pt and Re enhances the rate of C=O bond hydro-
genation more strongly if the C=O bond is a sterically hindered
ketone than if the C=O bond is an aldehyde. It is possible that
the oxophilic Re increases the preferential adsorption of the
ketone, but the expected increased adsorption strength of the
aldehyde, which is not sterically hindered, is apparently less
pronounced.
Implications for glycerol hydrogenolysis over Pt-Re
The rates and selectivities during glycerol hydrogenolysis at
393 K were also measured over the Pt-Re catalysts with varying
metal weight loadings (Table 7), which confirmed that all of
the Pt-Re catalysts were active for glycerol hydrogenolysis. A
monometallic Pt catalyst was essentially inactive under the
Table 7. Rate and selectivity of glycerol hydrogenolysis over the Pt-Re
catalysts with a Pt to Re molar ratio of unity.
A summary of the C=C and C=O bond hydrogenation results
observed for MVK, crotonaldehyde, and 2-butanone over differ-
ent weight-loaded Pt-Re catalysts is depicted in Figure 4. Over
the low-weight-loaded catalyst, the rates of C=C and C=O
bond hydrogenation were similar to those over monometallic
Pt. The low-weight-loaded Pt-Re catalyst had a higher fraction
of exposed silica support that causes Re to reside preferentially
on the silica instead of on the Pt nanoparticles. Thus, the unde-
corated Pt particles exhibited a catalytic performance similar to
that of monometallic Pt. The Pt-Re nanoparticles on the high-
weight-loaded Pt-Re catalysts, however, were more densely
[a]
[b]
Catalyst
Conversion TOFH TOFSM Rate Secondary/
[%] ] primary
[sÀ1 [sÀ1 [mmolgmetalÀ1 sÀ1
]
]
1 wt% Pt-Re 5.4
2 wt% Pt-Re 8.6
4 wt% Pt-Re 6.9
8 wt% Pt-Re 6.9
0.0029 0.0027 6.2
0.0049 0.0056 9.4
0.0066 0.0035 7.5
0.0083 0.0030 7.5
0.44
0.57
0.47
0.57
Reaction conditions: 393 K, 4 MPa H2, 90 cm3 of 110 mm aqueous glycerol
solution, Gly/Re=150. In situ pretreatment: 393 K, 1.4 MPa H2 for 1 h.
[a] Normalized by H chemisorption uptake. [b] Normalized by surface
metal atoms (Pt+Re) evaluated from the inverse of the average particle
size obtained from TEM.
ChemCatChem 2016, 8, 1074 – 1083
www.chemcatchem.org
1080
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
conditions used. The results given in Table 7 reveal that selec-
tivity (indicated by the Sec/Prim ratio, which is measure of se-
lectivity explained in the Experimental Section) was not a func-
tion of the Pt-Re metal loading. As discussed earlier, the TOFSM
and the global rate of C=C bond hydrogenation appeared to
decrease with increasing Pt-Re weight loading (Tables 3 and 4),
which differs from the trend in the glycerol hydrogenolysis
rates reported in Table 7. The TOFH for glycerol hydrogenolysis
follows the same trend as the TOFH for C=O bond hydrogena-
tion in 2-butanone. In other words, the highest TOFH for glyc-
erol hydrogenolysis and 2-butanone hydrogenation were both
observed on the 8 wt% Pt-Re catalyst.
Experimental Section
Catalyst preparation
All catalysts were prepared by the incipient wetness impregnation
of metal precursors followed by a thermal treatment. The Pt-Re bi-
metallic catalysts were synthesized by the successive impregnation
of the Pt precursor followed by the Re precursor. The Pt precursor
was Pt(NH3)4(NO3)2 (Sigma Aldrich), and the Re precursor was
NH4ReO4 (Sigma Aldrich). Typically, the Pt precursor was dissolved
in a predetermined amount of distilled, deionized water to cause
the silica gel support (Fuji Silysia G6) to reach incipient wetness.
After the Pt solution was added slowly and mixed thoroughly with
the support, the catalyst was dried overnight at 393 K in air. The
Re precursor was then dissolved in water, and the second impreg-
nation was performed in the same manner. The resulting catalyst
was then dried overnight in air at 393 K.
The rates for 2-butanone and butanal hydrogenation at
333 K over 8 wt% Pt-Re were substantially greater than the
rate of glycerol hydrogenolysis at 393 K. This comparison sug-
gests that the rates of C=O (and C=C) bond hydrogenation
under the conditions for glycerol hydrogenolysis would be
much higher than the rate of hydrogenolysis as the activation
energy for C=O bond hydrogenation has been measured to be
greater than 40 kJmolÀ1 over transition-metal catalysts.[44,45] For
an activation energy of 40 kJmolÀ1, an increase of the temper-
ature of hydrogenation from 333 to 393 K would result in
a ninefold increase in the rate. The rapid rate of hydrogenation
relative to hydrogenolysis is consistent with a mechanism of
hydrogenolysis presented previously,[24] in which glycerol dehy-
dration, catalyzed by a Brønsted acid site, is a kinetically signifi-
cant step and that subsequent hydrogenation is fast.[24]
Before any measurements or reactions were performed, the bimet-
allic and monometallic catalysts were heated to 723 K and reduced
for 3 h in 100 cm3 minÀ1 of flowing H2 (GTS Welco, 99.999%). A pas-
sivation gas (GTS Welco, 1.03% O2 balance N2) was introduced to
the catalysts for 30 min after they were cooled to RT. Following the
gas treatment, the catalysts were stored under air at RT in a closed
vial.
Catalyst characterization
The total number of hydrogen adsorption sites was determined by
H2 chemisorption by using a Micromeritics ASAP 2020 automated
adsorption analyzer. Samples were first heated at 4 KminÀ1 and re-
duced at 473 K in flowing H2 for 1.5 h. Next, the samples were
evacuated for 2 h at 473 K and then for another 2 h at 308 K. After
the evacuation, analysis was performed at 308 K in the pressure
range of 75 to 450 Torr. The total number of hydrogen adsorption
sites was calculated by extrapolating the total amount of H2 ad-
sorbed in the saturated region to zero pressure to remove contri-
butions from physisorbed hydrogen. TEM images of the catalysts
were acquired by using an FEI Titan operated at 300 kV and
equipped with a Gatan 794 Multi-scan Camera and an energy dis-
persive spectrometer for elemental X-ray analysis. Samples for mi-
croscopy were prepared by dipping TEM grids into a mixture of
approximately 50 mg of catalyst and 7 cm3 of cyclohexane, which
was sonicated for 15 min. To evaluate the particle size of the
sample, the diameter of over 300 individual particles was measured
for each sample.
Conclusions
The rates of C=C and C=O bond hydrogenation in methyl vinyl
ketone (MVK), crotonaldehyde, 2-butanone, and butanal in
liquid water were measured over Pt and Pt-Re catalysts sup-
ported on silica in a kinetically controlled regime. The normal-
ized rate of MVK and crotonaldehyde hydrogenation, which in-
volves primarily C=C bond hydrogenation, decreased as the
metal weight loading of a Pt-Re catalyst with a 1:1 nominal Pt/
Re ratio increased. Results from H2 chemisorption and TEM
suggest that the high-weight-loaded Pt-Re catalysts had a rela-
tively higher ratio of Re to Pt at the surface compared to the
low-weight-loaded Pt-Re catalysts. Consistent with this obser-
vation, the rates of MVK and crotonaldehyde hydrogenation
over monometallic Pt were similar to those over the low-
weight-loaded Pt-Re catalyst. In contrast, the rate of C=O bond
hydrogenation in 2-butanone was higher for the high-weight-
loaded Pt-Re catalysts than that of monometallic Pt. For croto-
naldehyde hydrogenation, the addition of Re adjusted the rela-
tive rates of C=C and C=O bond hydrogenation, which thereby
increased the selectivity to unsaturated aldehyde from 5%
over Pt to 21% over Pt-Re. The same Pt-Re catalysts were also
tested for glycerol hydrogenolysis in liquid water. The relative
rates of C=C and C=O bond hydrogenation over Pt-Re were
significantly faster than that of CÀO bond hydrogenolysis,
which supports the hypothesis that a hydrogenation step in
the glycerol hydrogenolysis mechanism would be kinetically ir-
relevant.
Hydrogenation reactions
Hydrogenation reactions were performed by using a 100 cm3 Parr
autoclave equipped with a glass liner, PID-controlled heater,
Teflon-lined magnetically-driven stir bar, and Teflon-lined dip tube
for periodic sampling. The reaction solution was prepared by di-
luting MVK (Acros Organics, 95%), 2-butanone (Sigma Aldrich,
99.7%), crotonaldehyde (Aldrich, 99%), or butanal (Fluka, 99.0%)
with distilled, deionized water and an internal standard. The inter-
nal standards 1-propanol and 1-pentanol were used for the ke-
tones and aldehydes, respectively. Typically, 10 mg of catalyst
(unless otherwise stated) was loaded into the autoclave, and the
solution was loaded into a separate reactant vessel. After the con-
tents were sealed in their respective containers, H2 was used to
purge the reactor and the reactant solution vessel for 10 min at
atmospheric pressure. To re-reduce the catalyst after exposure to
air, an in situ reduction was performed at 393 K and 15 psig of H2
ChemCatChem 2016, 8, 1074 – 1083
www.chemcatchem.org
1081
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[1] P. Mäki-Arvela, J. Hµjek, T. Salmi, D. Y. Murzin, Appl. Catal. A 2005, 292,
1–49.
[2] M. S. Ide, B. Hao, M. Neurock, R. J. Davis, ACS Catal. 2012, 2, 671–683.
[3] P. Claus, Top. Catal. 1998, 5, 51–62.
for 1 h. After reduction the reactor was cooled to 333 K, the solu-
tion was then charged into the reactor, the reactor was pressur-
ized to 15 psig with H2, and the magnetically-driven stir bar was
set to 650 rpm. An initial sample of the reaction medium was
taken through the dip tube while the reactor was kept sealed. The
desired pressure of H2 was maintained throughout the reaction by
feeding H2 constantly. Samples were analyzed by using a gas chro-
matograph (Hewlett Packard 5890 Series II) equipped with a flame
ionization detector and a 50 m long HP-1 cross-linked methyl sili-
cone gum capillary column. The conversion was calculated as the
number of moles of products formed divided by the number of
moles of substrate loaded into the reactor initially. Response fac-
tors were obtained using known concentrations of reactants and
products. All reactions had a carbon balance between 90 and
105%.
[4] V. Ponec, Appl. Catal. A 1997, 149, 27–48.
[5] D. Loffreda, F. Delbecq, F. VignØ, P. Sautet, Angew. Chem. Int. Ed. 2005,
44, 5279–5282; Angew. Chem. 2005, 117, 5413–5416.
[6] P. Concepción, Y. PØrez, J. C. Hernµndez-Garrido, M. Fajardo, J. J. Calvino,
A. Corma, Phys. Chem. Chem. Phys. 2013, 15, 12048–12055.
[7] F. Coloma, A. Sepulveda-Escribano, J. L. G. Fierro, F. Rodriguez-Reinoso,
Appl. Catal. A 1996, 136, 231–248.
[8] A. Hammoudeh, S. Mahmoud, J. Mol. Catal. A 2003, 203, 231–239.
[9] S. Galvagno, A. Donato, G. Neri, R. Pietropaolo, G. Capannelli, J. Mol.
Catal. 1993, 78, 227–236.
[10] R. Hirschl, F. Delbecq, P. Sautet, J. Hafner, J. Catal. 2003, 217, 354–366.
[11] R. Hirschl, A. Eichler, J. Hafner, J. Catal. 2004, 226, 273–282.
[12] J. L. Margitfalvi, I. Borbµth, M. Hegedus, A. Tompos, Appl. Catal. A 2002,
229, 35–49.
[13] F. B. Passos, M. Schmal, M. A. Vannice, J. Catal. 1996, 160, 106–117.
[14] A. Dandekar, M. A. Vannice, J. Catal. 1999, 183, 344–354.
[15] S. J. Tauster, S. C. Fung, R. L. Garten, J. Am. Chem. Soc. 1978, 100, 170–
175.
[16] M. Tamura, K. Tokonami, Y. Nakagawa, K. Tomishige, Chem. Commun.
2013, 49, 7034–7036.
[17] S. Koso, Y. Nakagawa, K. Tomishige, J. Catal. 2011, 280, 221–229.
[18] Y. Amada, Y. Shinmi, S. Koso, T. Kubota, Y. Nakagawa, K. Tomishige, Appl.
Catal. B 2011, 105, 117–127.
[19] Y. Nakagawa, Y. Shinmi, S. Koso, K. Tomishige, J. Catal. 2010, 272, 191–
194.
[20] K. Chen, K. Mori, H. Watanabe, Y. Nakagawa, K. Tomishige, J. Catal.
2012, 294, 171–183.
[21] M. Chia, Y. J. Pagµn-Torres, D. Hibbitts, Q. Tan, H. N. Pham, A. K. Datye,
M. Neurock, R. J. Davis, J. A. Dumesic, J. Am. Chem. Soc. 2011, 133,
12675–12689.
[22] M. Chia, B. J. O’Neill, R. Alamillo, P. J. Dietrich, F. H. Ribeiro, J. T. Miller,
J. A. Dumesic, J. Catal. 2013, 308, 226–236.
[23] O. M. Daniel, A. DeLaRiva, E. L. Kunkes, A. K. Datye, J. A. Dumesic, R. J.
Davis, ChemCatChem 2010, 2, 1107–1114.
Glycerol hydrogenolysis reactions
The reactor set-up and procedure for glycerol hydrogenolysis were
similar to those used for the hydrogenation reactions. The reaction
solution was prepared by diluting glycerol (Acros Organics, ACS
Grade) with distilled, deionized water. The appropriate amount of
catalyst was loaded into the empty reactor to give a ratio of moles
of glycerol to total moles of Re (Gly/Re) equal to 150. After the in
situ reduction at 393 K, the reactant vessel was then heated to the
reaction temperature of 393 K. When the reactant vessel reached
the desired temperature, the solution was charged into the reactor,
and the reactor was pressurized to the desired pressure with H2.
Samples were taken periodically throughout the course of the re-
action and were analyzed by using an Alliance high-performance
liquid chromatograph (Waters e2695) equipped with a refractive
index detector (Waters 2414) and an Aminex ion exclusion column
(Bio-Rad, HPX-87H) operated at 318 K with a 0.5m H2SO4 mobile
phase. Calibration curves were obtained using known concentra-
tions of reactants and products.
[24] D. D. Falcone, J. H. Hack, A. Y. Klyushin, A. Knop-Gericke, R. Schlçgl, R. J.
Davis, ACS Catal. 2015, 5, 5679–5695.
The conversion was calculated as the total number of moles of
products divided by the initial number of moles of glycerol. The
only products observed during glycerol hydrogenolysis at 393 K
were 1,2-propanediol, 1,3-propanediol, 1-propanol, and 2-propanol.
All glycerol hydrogenolysis reactions performed in this work had
a carbon balance between 90 and 105%. For simplicity, the ratio of
CÀO cleavages at secondary carbon atoms versus that at primary
carbon atoms was calculated as the Sec/Prim ratio to compare se-
lectivities for reactions performed with different catalysts. This
number was calculated using the concentrations of the final prod-
ucts as shown below [Eq. (1)].
[25] M. Boudart, G. DjØga-Mariadassou, Kinetics of Heterogeneous Catalytic
Reactions, Princeton University Press, Princeton, NJ, 1984.
[26] E. Kunkes, D. Simonetti, J. Dumesic, W. Pyrz, L. Murillo, J. Chen, D. But-
trey, J. Catal. 2008, 260, 164–177.
[27] C. Deng, X. Duan, J. Zhou, D. Chen, X. Zhou, W. Yuan, Catal. Today
2014, 234, 208–214.
[28] B. C. Vicente, R. C. Nelson, A. W. Moses, S. Chattopadhyay, S. L. Scott, J.
Phys. Chem. C 2011, 115, 9012–9024.
[29] R. J. Madon, M. Boudart, Ind. Eng. Chem. Fundam. 1982, 21, 438–447.
[30] C. Milone, R. Ingoglia, A. Pistone, G. Neri, F. Frusteri, S. Galvagno, J.
Catal. 2004, 222, 348–356.
[31] A. L. Gemal, J. L. Luche, J. Am. Chem. Soc. 1981, 103, 5454–5459.
[32] J. Kaspar, A. Trovarelli, F. Zamoner, E. Farnetti, M. Graziani, Stud. Surf. Sci.
Catal. 1991, 59, 253–261.
[33] T. B. L. W. Marinelli, S. Nabuurs, V. Ponec, J. Catal. 1995, 151, 431–438.
[34] S. E. Clapham, A. Hadzovic, R. H. Morris, Coord. Chem. Rev. 2004, 248,
2201–2237.
Sec
½1; 3 À PDOŠ þ ½1 À PropŠ
¼
Prim
½1; 2 À PDOŠ þ 2 ½2 À PropŠ þ ½1 À PropŠ
*
[35] J. X. Chen, J. F. Daeuble, D. M. Brestensky, J. M. Stryker, Tetrahedron
2000, 56, 2153–2166.
[36] T. Ohkuma, H. Ooka, T. Ikariya, R. Noyori, J. Am. Chem. Soc. 1995, 117,
10417–10418.
[37] R. Grey, G. P. Pez, A. Wallo, J. Am. Chem. Soc. 1981, 103, 7536–7542.
[38] S. Koso, N. Ueda, Y. Shinmi, K. Okumura, T. Kizuka, K. Tomishige, J. Catal.
2009, 267, 89–92.
Acknowledgements
This material was based primarily upon the work supported by
the National Science Foundation under award number CBET-
1157829.
[39] P. Reyes, C. Rodriguez, G. Pecchi, J. L. G. Fierro, Catal. Lett. 2000, 69, 27–
32.
[40] M. Englisch, V. S. Ranade, J. A. Lercher, Appl. Catal. A 1997, 163, 111 –
122.
Keywords: aldehydes · hydrogenation · platinum · rhenium ·
supported catalysts
[41] Y. Chiang, A. J. Krege, Y. S. Tang, J. Wirz, J. Am. Chem. Soc. 1984, 106,
460–462.
ChemCatChem 2016, 8, 1074 – 1083
www.chemcatchem.org
1082
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[42] C. Bohne, I. D. MacDonald, H. B. Dunford, J. Am. Chem. Soc. 1986, 108,
[45] S. H. Pang, C. A. Schoenbaum, D. K. Schwartz, J. W. Medlin, ACS Catal.
2014, 4, 3123–3131.
7867–7868.
[43] Y. Chiang, A. J. Kresge, N. P. Schepp, J. Am. Chem. Soc. 1989, 111, 3977–
3980.
Received: November 23, 2015
[44] C. J. Kliewer, M. Bieri, G. A. Somorjai, J. Am. Chem. Soc. 2009, 131, 9958–
9966.
Published online on February 17, 2016
ChemCatChem 2016, 8, 1074 – 1083
www.chemcatchem.org
1083
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim