Evidence for the Bifunctional Nature of Pt-Re Catalysts for Selective Glycerol Hydrogenolysis

Article abstract of DOI:10.1021/acscatal.5b01371

Rhenium substantially promotes the rate of Pt-catalyzed glycerol hydrogenolysis to propanediols and shifts the product selectivity from 1,2-propanediol to a mixture of 1,2 and 1,3-propanediols. This work presents experimental evidence for a tandem dehydration-hydrogenation mechanism that occurs over a bifunctional Pt-Re catalyst. Infrared spectroscopy of adsorbed pyridine and the rate of aqueous-phase hydrolysis of propyl acetate were used to identify and quantify Bronsted acid sites associated with the Re component. Near-ambient-pressure XPS revealed a range of Re oxidation states on the Pt-Re catalysts after reduction in H₂ at 393 and 493 K, which accounts for the presence of Bronsted acidity. A mechanism involving acid-catalyzed dehydration followed by Pt-catalyzed hydrogenation was consistent with the negative influence of added base, a primary kinetic isotope effect with deuterated glycerol, an inverse isotope effect with dideuterium gas, and the observed orders of reaction.

Full text of DOI:10.1021/acscatal.5b01371

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
Evidence for the Bifunctional Nature of Pt-
Re Catalysts for Selective Polyol Hydrogenolysis
Derek D. Falcone, John H Hack, Alexander Yu. Klyushin,
Axel Knop-Gericke, Robert Schlögl, and Robert J Davis
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.5b01371 • Publication Date (Web): 18 Aug 2015
Downloaded from http://pubs.acs.org on August 20, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 65
ACS Catalysis
Page 1 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
Evidence for the Bifunctional Nature of
Pt-Re Catalysts for Selective Glycerol
Hydrogenolysis
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
1
2
2
Derek D. Falcone , John H. Hack , Alexander Yu. Klyushin , Axel KnopꢀGericke , Robert
2
1
Schlögl , and Robert J. Davis *
1
University of Virginia, Department of Chemical Engineering,
102 Engineers’ Way, PO Box 400741, Charlottesville, VA 22904ꢀ4741, United States
2
FritzꢀHaber Institut der MaxꢀPlanck Gesellschaft, Department of Inorganic Chemistry,
4–6 Faradayweg, 14195, Berlin, Germany
*
Author to whom correspondence is addressed
Phone: 1 ꢀ 434 ꢀ 924 ꢀ 6284
Email: rjd4f@virginia.edu
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 65
Page 2 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT
Rhenium substantially promotes the rate of Ptꢀcatalyzed glycerol hydrogenolysis to
propanediols and shifts the product selectivity from 1,2ꢀpropanediol to a mixture of 1,2 and 1,3ꢀ
propanediols. This work presents experimental evidence for a tandem dehydrationꢀhydrogenation
mechanism that occurs over a bifunctional PtꢀRe catalyst. Infrared spectroscopy of adsorbed
pyridine and the rate of aqueousꢀphase hydrolysis of propyl acetate were used to identify and
quantify Brønsted acid sites associated with the Re component. Near ambient pressure XPS
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
revealed a range of Re oxidation states on the PtꢀRe catalysts after reduction in H at 393 and
2
493 K, which accounts for the presence of Brønsted acidity. A mechanism involving acidꢀ
catalyzed dehydration followed by Ptꢀcatalyzed hydrogenation was consistent with the negative
influence of added base, a primary kinetic isotope effect with deuterated glycerol, an inverse
isotope effect with dideuterium gas, and the observed orders of reaction.
KEYWORDS: heterogeneous catalysis, bimetallic, biomass, platinum, rhenium, Brønsted
acidity, Xꢀray photoelectron spectroscopy
ACS Paragon Plus Environment

Page 3 of 65
ACS Catalysis
Page 3 of 64
. INTRODUCTION
Interest in bimetallic catalysts consisting of a late transition metal such as Pt together
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
with an oxophilic metal like Re has increased in recent years because of their potential
application in several reactions relevant to the transformation of biomassꢀderived molecules to
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
–14
fuels and chemicals.
Although aluminaꢀsupported sulfided PtꢀRe bimetallic catalysts have
1
5–21
been used industrially for decades in naphtha reforming,
the unꢀsulfided version of this
catalyst exhibits very interesting catalytic performance in liquid water.
In the early 2000’s, Dumesic and coꢀworkers reported that platinum group metals
2
2
catalyze the aqueous phase reforming (APR) of biomassꢀderived molecules. Soon thereafter,
several monometallic and bimetallic catalysts based on transition group metals were developed
2
3–25
to reform glycerol, a waste product of fatty acid transesterification to produce biodiesel.
Promotion of carbonꢀsupported Pt nanoparticles with a second oxophilic metal increased the rate
of glycerol reforming to H , CO, CO , and light alkanes in liquid water by an order of magnitude
2
2
1–3
compared to monometallic Pt.
Supported platinum group nanoparticles are also effective catalysts for the
hydrogenolysis of glycerol. The hydrogenolysis of glycerol is often carried out at temperatures
2
6–29
lower than that for APR to maximize the selectivity to propanediols.
Daniel et al. reported
that the addition of Re to Pt increased the rate of glycerol hydrogenolysis substantially and
shifted the product selectivity from almost exclusively terminal hydrogenolysis, which produces
1
,2ꢀpropanediol (1,2ꢀPDO), toward internal hydrogenolysis, which formed 1,3ꢀpropanediol (1,3ꢀ
4
PDO). At temperatures as low as 393 K, monometallic Pt, Rh, or Ir catalysts are nearly inactive
for glycerol hydrogenolysis, but the addition of Re increases the rate of the reaction in liquid
7
,30,31
water and increases the selectivity of internal CꢀO bond hydrogenolysis.
These types of
bimetallic catalysts also demonstrate high activity and selectivity for the hydrogenolysis of other
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 65
Page 4 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
α,βꢀoxygen substituted substrates like 2ꢀ(hydroxymethyl)tetrahydropyran and tetrahydrofurfuryl
7
,32,33
alcohol.
While a significant amount of work has documented the performance of bimetallic
catalysts consisting of Re together with a platinum group metal, the mechanism of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5–8
hydrogenolysis in liquid water and the promotional role of Re are still unclear.
The Tomishige group has studied the hydrogenolysis of different molecules and proposes
the role of Re is to act as a Lewis acid. For glycerol hydrogenolysis, they speculate glycerol
anchors to the surface through the terminal alcohol group followed by the activation of the
5
,6
internal CꢀO bond of glycerol by a hydride produced from the heterolytic cleavage of H2.
A different mechanism for hydrogenolysis of oxygenates was proposed by Chia et al. In
7
that work, researchers used results from reaction studies, NH temperature programed
3
desorption, and density functional theory (DFT) calculations to suggest that a Brønsted acid site
is associated with the Re component of the bimetallic catalyst in liquid water. Supported by
results from quantum chemical calculations, Chia et al. suggested that protolytic activation
preferentially occurred at the internal carbon of glycerol instead of the terminal carbon because
7
secondary carbenium ions are more stable than primary carbenium ions. Following the acidꢀ
catalyzed dehydration of glycerol on the speculated acid site associated with Re, the unsaturated
reaction intermediate could then be hydrogenated by the late transition metal component of the
catalysts (Rh) to form the final product (1,3ꢀPDO). Consistent with the mechanism involving a
Brønsted acid site, the ion formation energies of several other oxocarbenium ions showed a
7
strong correlation with the rate of hydrogenolysis of the corresponding oxygenate.
Other works have also suggested that Brønsted acid sites associated with Re are
responsible for the increased rate of glycerol APR when Re is added to Pt based
12,13,34,35
catalysts.
In addition, FTIR spectroscopy of adsorbed pyridine showed that Brønsted acid
ACS Paragon Plus Environment

Page 5 of 65
ACS Catalysis
Page 5 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
12,34
sites were present on PtꢀRe bimetallic catalysts supported on silica.
It is unclear from these
studies, however, whether or not Brønsted acid sites are ascribed to the PtꢀRe bimetallic
3
6
nanoparticles or to weaklyꢀheld perrhenate species on the silica surface. Ciftci et al. also used
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
evidence from CO stripping voltammetry to show that Re is a strong binder of oxygen species,
1
2
which could facilitate water activation to produce OH species on the surface.
In this work, the bifunctional nature of supported PtꢀRe bimetallic catalysts is
investigated. The PtꢀRe system was chosen because theoretical calculations suggest that PtꢀRe
bimetallic nanoparticles form the strongest Brønsted acid sites (based on deprotonation energy)
compared to other common bimetallic systems containing a reducible metal and an oxophilic
7
,37
metal. Glycerol hydrogenolysis in the presence of NaOH, diffuse reflectance infrared Fourier
transform spectroscopy (DRIFTS) of adsorbed pyridine, and propyl acetate hydrolysis were used
to gain critical insights into the role of the Re component in bimetallic PtꢀRe nanoparticles for
glycerol hydrogenolysis. In addition, the kinetics of glycerol hydrogenolysis were investigated
by varying glycerol concentration and H pressure as well as measuring the kinetic isotope effect
2
using d ꢀglycerol(OH) and D .
5
3
2
The oxidation state of Re in the working state of the bimetallic catalysts is still unknown
even though the oxidation state of Re in PtꢀRe bimetallic catalysts has been under investigation
3
8
for several decades. One of the reasons for the ambiguity regarding the oxidation state of Re is
7
+
its dependence on the conditions to which the Re is subjected. The reduction of Re or
,
39
perrhenate (the most stable oxidation state), usually does not begin until approximately 573 K,
4
but can occur as low as 373 K in the presence of Pt because of hydrogen spillover. Since Re is
7
+
oxophilic, the partiallyꢀreduced Re in the PtꢀRe bimetallic catalyst is oxidized to Re upon
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 65
Page 6 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
exposure to air. Therefore, the oxidation state of Re depends on the pretreatment of the catalyst
3
5
and the conditions under which the catalyst is probed.
Recently, in situ Xꢀray absorption spectroscopy (XAS) has been used to study PtꢀRe
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
40
41
particles supported on Norit carbon as well as RhꢀRe and IrꢀRe particles supported on silica.
The average oxidation state of Re in those works was determined to be approximately 2+ after
reduction in flowing H at temperatures between 373 and 473 K. Another study by Chia et al.
2
8
concluded that Re in RhꢀRe nanoparticles supported on Vulcan carbon was metallic.
Interestingly, that work showed that the oxidation state of Re was the same in gaseous H as well
2
as under reaction conditions with H ꢀsaturated water.
2
In the current study, Xꢀray photoelectron spectroscopy was used to study the oxidation
state of Re in PtꢀRe bimetallic catalysts supported on both Vulcan carbon and silica. The
advantages of using XPS are that the various oxidation states can be discerned instead of an
overall bulk average, and the technique is more sensitive to the elemental composition at the
surface, which is most applicable to catalysis. In addition to using conventional XPS equipped
with an Al Kα radiation source, experiments were performed using a near ambient pressure Xꢀ
ray photoelectron spectrometer (NAPꢀXPS), which allowed for measurements to take place in
the presence of H and water vapor at 393 and 473 K.
2
2
. RESULTS
2.1 Catalyst Characterization
The properties of the supported bimetallic catalysts used in this study are summarized in
Table 1. Silica gel and Vulcan carbon were used as supports to permit interrogation of the PtꢀRe
nanoparticles by infrared spectroscopy and NAPꢀXPS, respectively. All of the catalyst studied in
this work had a 1:1 nominal molar ratio of Pt to Re and the naming convention used only
ACS Paragon Plus Environment

Page 7 of 65
ACS Catalysis
Page 7 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
describes the weight loading of Pt. For example, as shown in Table 1 the catalyst named 8 wt%
PtꢀRe/SiO had a Pt weight loading of 8% and a Re weight loading of 7.6%, which corresponds
2
to a 1:1 molar ratio of Pt and Re.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Dihydrogen chemisorption was used to estimate the number of exposed Pt sites on the
bimetallic catalysts. Previous studies revealed that the uptake of chemisorbed hydrogen on
4
supported monometallic Re catalysts is below the detection limit. The results in Table 1 show
the bimetallic catalysts chemisorbed more hydrogen compared the monometallic Pt catalysts
indicating a greater fraction of the Pt was exposed. This is partly attributed to the fact that the Ptꢀ
Re catalysts had smaller particle sizes compared to the monometallic Pt catalysts, as revealed by
TEM. This trend was especially evident on the Vulcanꢀsupported catalysts. The bimetallic 8 wt%
PtꢀRe/C catalyst had a significantly higher H/Pt ratio than the 8 wt% Pt/C catalyst and the
number average particle size measured by TEM was about half the size of the monometallic
catalyst. Moreover, the monometallic Pt/C catalyst also had a much broader particle size
distribution that included some particles greater than 20 nm. In fact, the bimetallic catalysts had
narrower particle size distributions compared to the corresponding monometallic catalysts,
regardless of support. The particle size distributions for the catalysts in Table 1 are depicted in
Figure S1. The consistently smaller particle sizes of the PtꢀRe bimetallic catalysts indicates a
2
,42
stabilization of Pt against sintering at elevated temperatures by the presence of Re.
Table 1 also summarizes the properties of 8 wt% PtꢀRe/SiO ꢀW, which is catalyst that
2
has been “washed” to remove weaklyꢀheld, waterꢀsoluble perrhenate that was not associated with
the Pt on 8 wt% PtꢀRe/SiO . Vicente et al. report that Re catalysts synthesized via incipient
2
wetness impregnation of aqueous perrhenate results in wellꢀdispersed, but weaklyꢀheld species
36
on silica supports. Therefore, it is likely that some weaklyꢀheld perrhenate ions were present
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 65
Page 8 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
on the silicaꢀbased PtꢀRe bimetallic catalysts used in this work and others reported previously. In
1
2
addition, silicaꢀsupported perrhenate has been shown to expose Brønsted acid sites. The
washing procedure is described in Section 5.1.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2.2 Polyol Hydrogenolysis
2.2.1 Glycerol hydrogenolysis
The rates and selectivities of glycerol hydrogenolysis at 393 K in a 110 mM aqueous
glycerol solution with 4 MPa of H over 8 wt% PtꢀRe/SiO 8 wt% PtꢀRe/SiO ꢀW, and 8 wt% Ptꢀ
2
2,
2
Re/C are presented in Table 2. Two measures of activity are presented in Table 2, one
normalized to the mass of the catalyst, and the other normalized to the hydrogen uptake
measured via H chemisorption to give a turnover frequency (TOF). The TOF is one way to
2
report a rate that properly accounts for differences in exposed metal attributed to differences in
particle size. Unfortunately, the different affinity of Re and Pt for hydrogen means that the TOF
expressed here is still based on an approximation of exposed active sites. Nevertheless, we
observed that the PtꢀRe nanoparticles supported on silica had a higher activity per mass of
catalyst as well as a higher TOF compared to the Vulcanꢀsupported bimetallic catalyst. While the
rate was higher on the silicaꢀsupported catalyst, the Sec:Prim ratio (a ratio of internal
hydrogenolysis products to terminal hydrogenolysis products defined in Section 5.1) was lower
on the silicaꢀsupported catalyst when compared to carbonꢀsupported catalyst, indicating that the
carbonꢀsupported catalyst was more selective to internal hydrogenolysis reactions. In general, at
393 K in a 110 mM aqueous glycerol solution with 4 MPa of H , we observed the Sec:Prim ratio
2
on silicaꢀsupported bimetallic catalysts to range between 0.5 and 0.6 whereas the ratio on the
carbonꢀsupported bimetallic catalysts was between 0.65 and 0.8. The Sec:Prim ratio was
relatively independent of the glycerol conversion, up to 55%. Even though the addition of Re
ACS Paragon Plus Environment

Page 9 of 65
ACS Catalysis
Page 9 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
shifts the product selectivity to 1,3ꢀPDO, corresponding to an increase in the Sec:Prim ratio,
00% selectivity to products resulting from the activation of secondary CꢀO bonds was never
1
obtained and primary CꢀO bond activation over PtꢀRe was always observed. As explained by
Chia et al., a possible intermediate for the hydrogenolysis of glycerol was found to be an
oxiraneꢀtype carbocation formed via dehydration and ring closure. Ring opening of the oxirane
carbocation can lead to either terminal alcohol (1,3ꢀPDO) or secondary alcohol (1,2ꢀPDO)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
7
depending on which bond is activated. We suspect this is why very high selectivity to 1,3ꢀPDO
is never achieved with glycerol. A time profile of glycerol hydrogenolysis on 8 wt% PtꢀRe/SiO2
is presented in the Supporting Information (Figure S2).
As seen in Table 2, the rate of glycerol hydrogenolysis per mass of catalyst of the washed
catalyst (8 wt% PtꢀRe/SiO ꢀW) was approximately 60% of the rate of the unwashed catalyst (8
2
wt% PtꢀRe/SiO ), while the Sec:Prim ratio was slightly higher on the washed catalyst. To wash
2
the catalyst, liquid water, which was purged with N and saturated with H , flowed over the
2
2
3
ꢀ1
catalyst at a rate of 2 cm min for approximately 30 min. In Table 1, ICPꢀOES results show that
the Re/Pt ratio decreased by approximately 11%, indicating that some Re was removed from the
catalyst as a result of washing. In addition, the hydrogen uptake measured via H chemisorption
2
decreased slightly. The particle size analysis, however, showed that the metal particle size did
not change significantly during the washing procedure. In light of the reactivity results in Table
2, we conclude that the washing procedure did not significantly alter the nature of the active
sites. As discussed later, the weaklyꢀheld, soluble perrhenate that is removed during the washing
treatment is inactive for glycerol hydrogenolysis. A small decrease in glycerol hydrogenolysis
activity following the washing treatment suggests that some active Re associated with the Pt is
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 65
Page 10 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
removed during washing. The main goal of the washing procedure, however, was to ensure that
all soluble perrhenate was removed from the catalyst.
The monometallic Pt/SiO catalyst demonstrated a glycerol conversion of only 0.3% after
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
34 h of reaction at 393 K (utilizing three times the catalyst loadings used for results in Table 2)
ꢀ
6 ꢀ1
which corresponds to a TOF over monometallic Pt of 4.2 x 10 s . Conversion of glycerol was
not observed on 8 wt% Pt/C at 393 K because of its very low dispersion of Pt. Daniel et al.
showed that at 473 K, a temperature at which monometallic Pt supported on Norit activated
carbon had some measurable activity, the TOF of their PtꢀRe/C catalyst was over an order of
magnitude higher than Pt/C. That study also showed that their carbonꢀsupported monometallic
4
Re catalyst was inactive for glycerol hydrogenolysis at 473 K. Clearly, the promotional effect of
adding Re to supported Pt catalysts was observed, regardless of the support.
Switching the Pt precursor used to make 8 wt% PtꢀRe/SiO from Pt(NH ) (NO ) to
2
3 4
3 2
H PtCl decreased the TOF of hydrogenolysis by almost 50%, but the Sec:Prim ratio only
2
6
changed from 0.47 to 0.39 at 7% conversion. Interestingly, the change in Pt precursor only
slightly decreased the rate of hydrogenolysis and the Sec:Prim ratio when Vulcan carbon was
used as the support, as shown in the Supporting Information (Table S1).
Li and coꢀauthors have studied glycerol hydrogenolysis extensively using bifunctional
4
3–46
catalysts comprised of Pt and different heteropolyacids (HPA) supported on zirconia
. They
observed a very strong correlation between the number of Brønsted acid sites, measured by the
FTIR spectroscopy of adsorbed pyridine, and the yield of 1,3ꢀPDO from glycerol hydrogenolysis
at 453 K. Their work suggests that the Brønsted acid sites on the HPA’s contribute to the
production of 1,3ꢀPDO, presumably by acidꢀcatalyzed dehydration followed by Ptꢀcatalyzed
hydrogenation.
ACS Paragon Plus Environment

Page 11 of 65
ACS Catalysis
Page 11 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
2.2.2 Glycerol hydrogenolysis using heterogeneous Pt and homogeneous acid
An experiment was performed to test whether or not it was sufficient to have the acid in
solution together with supported Pt particles to produce 1,3ꢀPDO during glycerol
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
hydrogenolysis. An amount of HCl that was equivalent to the moles of Re on the 8 wt% Ptꢀ
Re/SiO was added to the reaction medium with the monometallic 8 wt% Pt/SiO catalyst. The
2
2
ꢀ5 ꢀ1
reaction with HCl and 8 wt% Pt/SiO had a TOF of only 5.5 x 10 s at very low conversion
2
ꢀ1
compared to a TOF of 0.0105 s observed on the bimetallic catalyst. Since the TOF on 8 wt%
ꢀ6 ꢀ1
Pt/SiO in the absence of HCl was 4.2 x 10 s , the addition of HCl increased the rate by an
2
order of magnitude, but it was still several orders of magnitude slower than 8 wt% PtꢀRe/SiO2.
Another control reaction was conducted in which soluble HReO was added to the
4
reaction medium together with 8 wt% Pt/SiO . In this experiment, the molar ratio of Re and Pt
2
was also chosen to be 1:1. The TOF (based on exposed Pt) for the reaction with the HReO and 8
4
ꢀ4 ꢀ1
wt% Pt/SiO was 1.6 x 10 s , which was nearly 300% higher than the TOF for 8 wt% Pt/SiO
2
2
with HCl. The increase in rate with the addition of Re suggests that perrhenate in solution may
reduce onto the Pt nanoparticles during the course of the reaction and form the active sites that
are present on the PtꢀRe bimetallic catalyst. It also must be noted that the experiments with HCl
and HReO added to the reaction with 8 wt% Pt/SiO demonstrated Sec:Prim ratios of 1.0 and
4
2
0.9, respectively, indicating internal CꢀO bond activation is more favored in acidic solutions.
Moreover, both Pt and acid are required for high activity since conversion of glycerol was not
observed in the presence of only acid. The results from experiments with the homogenous acids
4
3–46
and supported Pt, considered together with the work conducted by Li and coꢀauthors,
suggest
that intimate contact of the Brønsted acid site and a Pt metal site is required to produce 1,3ꢀPDO
at a high rate from glycerol.
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 65
Page 12 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
.2.3 Kinetics of glycerol hydrogenolysis on 8 wt% Pt-Re/SiO2
Figure 1 shows rate of glycerol hydrogenolysis measured over a wide range of glycerol
concentrations (15 to 550 mM) after 1 h of reaction. The conversions and selectivities for these
experiments are presented in the Supporting Information (Table S2). As illustrated in Figure 1,
the reaction order in glycerol approached zero at high concentrations suggesting that the catalytic
surface was nearly saturated with a glycerolꢀderived species at concentrations above 550 mM. In
contrast, the rate of hydrogenolysis was approximately first order in glycerol at low
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
concentrations. Chia et al. studied tetrahydrofurfuryl alcohol hydrogenolysis on a carbon
supported RhꢀRe catalyst and also reported a first order dependence on tetrahydrofurfuryl
7
alcohol concentration in the range of 94 to 439 mM. Tomishige and coꢀauthors, however,
observed the rate of tetrahydrofurfuryl alcohol hydrogenolysis on both silicaꢀsupported RhꢀRe
47
48
and IrꢀRe catalysts, and the rate of glycerol hydrogenolysis on silicaꢀsupported RhꢀRe and Irꢀ
6
Re catalysts to be zero order in reactant. Those studies of cyclic ether and polyol hydrogenolysis
in liquid water by Tomishige et al. involved substrate concentrations of approximately 2 M or
higher. Thus, the kinetics were most likely measured in the surface saturation regime for both
tetrahydrofurfuryl alcohol and glycerol.
In the current work, an interesting trend between the Sec:Prim and the concentration of
glycerol was observed (Table S2). As the concentration of glycerol increased from 15 to 550 mM
the Sec:Prim ratio increased from 0.55 to 0.67 indicating that the catalyst was more selective to
internal hydrogenolysis reactions at higher concentrations of glycerol.
The order of reaction in H was also investigated for glycerol hydrogenolysis. The H
2
2
order was measured at three different glycerol concentrations (15, 110, and 550 mM) on 8 wt%
PtꢀRe/SiO Those results are tabulated in Table S3 and summarized in Figure 1. At the lowest
2.
ACS Paragon Plus Environment

Page 13 of 65
ACS Catalysis
Page 13 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
concentration of glycerol studied (15 mM) the order of H was essentially zero since the rate of
2
ꢀ1
ꢀ1
glycerol hydrogenolysis was approximately 20 ꢁmol g min when the H pressure was varied
2
from 2 to 8 MPa. The observed zero order behavior suggests that saturation coverage of
hydrogen is achieved on the PtꢀRe catalyst under very low glycerol concentrations. The order of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
reaction in H at higher concentrations of glycerol (110 and 550 mM), however, was quite close
2
to 0.5, which is a result typical of equilibrated dissociative adsorption of H during
2
7
hydrogenation reactions. Interestingly, the study by Chia et al. and the works by Tomishige and
4
7,48
coꢀauthors
reported first order kinetics in H pressure for cyclic ether and glycerol
2
hydrogenolysis on carbonꢀsupported RhꢀRe catalysts and silicaꢀsupported RhꢀRe and IrꢀRe
catalysts, respectively. First order behavior in H was not observed during glycerol
2
hydrogenolysis over 8 wt% PtꢀRe/SiO under conditions investigated in this work.
2
A reaction was also performed with N in the head space instead of H with a 110 mM
2
2
glycerol solution at 393 K. For this experiment, the normal in situ reduction of the catalyst was
performed, but prior to adding the reactant solution, the reactor was purged four times with
N and then 1 MPa of N was added and maintained throughout the course of the reaction. In the
2
2
in the absence of added H , the typical glycerol hydrogenolysis products were still observed (1,2ꢀ
2
PDO, 1,3ꢀPDO, 1ꢀProp, and 2ꢀProp), even after the catalytic site turned over several hundred
times. The reaction with N in the head space, however, also yielded several additional products
2
that were not observed in the presence of H , which include acetone, propanal, acetic acid,
2
propanoic acid, and ethanol (observed by GCꢀMS). The appearance of oxygenates containing
carbonyls may be the result of glycerol dehydration. The Brønsted acidꢀcatalyzed dehydration of
an alcohol group in glycerol produces an enol, which is rapidly converted in liquid water to a
49–51
ketone or an aldehyde via ketoꢀenol tautomerization.
We suspect that these products are not
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 65
Page 14 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
observed during reactions in high pressure H because of the rapid hydrogenation of these
2
intermediates. Alternatively, glycerol may act both as the substrate and as the source of
hydrogen. Other works observed that secondary alcohols can act as the hydrogen source to
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
52–54
reduce polyols instead of H following a MeerweinꢀPonndorfꢀVerley reaction mechanism.
2
Small amounts of hydrogen may also be formed by APR of glycerol over metal catalysts in
3
,13,54
liquid water.
For example, King et al. observed the evolution of 1,2ꢀPDO, 1ꢀProp and 2ꢀ
3
Prop in the liquid phase during the APR of glycerol over supported PtꢀRe particles at 498 K.
The mechanism of glycerol hydrogenolysis was further investigated by probing a kinetic
isotope effect with deuterated glycerol, which will be referred to as d5ꢀglycerol(OH) . In
3
particular, we measured the rate of hydrogenolysis of glycerol in which all CꢀH groups were
replaced with CꢀD groups. Using a 15 mM glycerol solution with 4 MPa of H , a primary kinetic
2
isotope effect of 2.0 ± 0.1 was observed at 393 K. A kinetic isotope effect (rate /rate ) of
H
D
approximately 2 was also measured in experiments with 4 MPa of H where the glycerol
2
concentration was 110 and 550 mM. As a control experiment, hydrogenolysis of glycerol (both
labeled and unlabeled) was also tested using D O as the solvent. No influence of using D O
2
2
instead of H O as the solvent was observed. The kinetic isotope effect of D was also measured
2
2
using a 15 mM aqueous solution of normal glycerol and 1 MPa of H (D ) at 393 K.
2
2
Interestingly, a kinetic isotope effect (rate /rate ) of approximately 0.7 was measured when D
2
H
D
was used as the reactant gas, indicating that the rate of glycerol hydrogenolysis was faster with
D .
2
2.2.4 Polyol hydrogenolysis on 8 wt% Pt-Re/SiO in an alkane solvent
2
To better understand the role of water in the reaction, hydrogenolysis experiments were
also performed using alkane (nꢀheptane) as the solvent. Since the solubility of glycerol in nꢀ
ACS Paragon Plus Environment

Page 15 of 65
ACS Catalysis
Page 15 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
heptane is extremely low, 1,2ꢀhexanediol (1,2ꢀHDO) was used as the model polyol substrate
because its solubility is relatively high in both water as and nꢀheptane. The 1,2ꢀHDO
hydrogenolysis results over 8 wt% PtꢀRe/SiO using water or nꢀheptane as the solvent are
2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀ1
ꢀ1
presented in Table 3. The observed rate of 1,2ꢀHDO hydrogenolysis was 23 and 48 ꢁmol g min
in nꢀheptane and water, respectively, indicating water is not required to perform hydrogenolysis
of 1,2ꢀHDO. The rate of 2ꢀhexanol production (2ꢀHex), which is produced by activating a
primary CꢀO bond in 1,2ꢀHDO, was the same regardless of the solvent used. The rate of 1ꢀ
hexanol production, on the other hand, was four times higher when water was used as the solvent
compared to heptane. The monometallic 8 wt% Pt/SiO catalyst was inactive for 1,2ꢀHDO
2
hydrogenolysis at 393 K in both water and nꢀheptane indicating that the Re component of the
catalyst is crucial for catalysis and the mechanism of 1,2ꢀHDO hydrogenolysis on PtꢀRe
bimetallic nanoparticles is similar to that of glycerol hydrogenolysis. The results from
hydrogenolysis of 1,2ꢀHDO over 8 wt% PtꢀRe/SiO also agree very well with work that was
2
previously performed by Nakagawa et al. who studied the hydrogenolysis of 1,2ꢀHDO in nꢀ
55
heptane on IrꢀRe/SiO2.
2.3 X-ray Photoelectron Spectroscopy of Pt-Re Catalysts
2.3.1 Conventional X-ray photoelectron spectroscopy
Xꢀray photoelectron spectroscopy was used to ascertain the oxidation states of Pt and Re
after the supported nanoparticles were reduced at the moderate temperatures used in the reaction
conditions of this study. Before measurements of the catalysts were performed, conventional
XPS of Re standards was conducted using Al Kα radiation, the spectra of which are shown in
Figure S3. All of the Re standards were characterized “as is” from the manufacturer with no
further pretreatment. Multiple Re oxidation states were observed in all of the spectra because the
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 65
Page 16 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
surfaces of the samples readily oxidize in air. In fact, the most thermodynamicallyꢀstable Re
3
9
oxide is Re O . Since Al Kα radiation at 1486.6 eV was used to measure photoemission spectra
2
7
of the Re standards, the inelastic mean free path of the electrons was approximately 1.6 nm,
which means photoelectrons from the bulk material will also contribute to the overall signals.
The majority of the photoemission response in the conventional XP spectra (Figure S3) of Re
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
56
0
4+
6+
metal (Re ), ReO (Re ), and ReO (Re ) standards can be ascribed to the bulk material. The Re
2
3
0
4+
6+
4f7/2 binding energies for Re , Re and Re were 40.5, 42.6, and 43.5 eV, respectively, and
,
3
8,39,57
+7
compare well with reported values.
To estimate the binding energy of Re , NH ReO was
4 4
used as the standard since this particular Re salt was used to prepare the catalysts. Two Re
species were observed in this sample, which had 4f7/2 peaks at 45.7 and 47.4 eV. The 47.4 eV
4f7/2 peak has been observed by others for NH ReO , while a 45.7 eV 4f peak is most likely
4
4
7/2
39,58,59
attributed to Re O .
In addition to features associated with the stable bulk oxides of Re, a
2
7
4f7/2 peak with a binding energy of 41.5 eV was also observed in the spectrum of Re metal
+
2 39,57
(Figure S3), which has been ascribed to ReO (Re ).
It should be noted that ReO is not stable
as a bulk material, but it has been observed at the interface of higher oxidation state oxides and
3
9,57,60
Re metal.
Once the binding energies for all of the standard Re oxide peaks were determined, the
photoemission spectra of PtꢀRe catalysts were then analyzed using the CasaXPS fitting software
with constraints that are described in Section 5.5. The Pt 4f and Re 4f conventional XP spectra of
8
wt% PtꢀRe/C and 8 wt% PtꢀRe/SiO , exposed to air without in situ reduction, are shown in
2
Figure 2. Even without in situ reduction, the Pt 4f spectra in Figure 2 reveal binding energies for
the 4f and 4f peaks of approximately 71.1 and 74.4 eV, respectively, which correspond to
7
/2
5/2
metallic Pt. In fact, the spectra for all of the catalysts studied in this work revealed that Pt was
ACS Paragon Plus Environment

Page 17 of 65
ACS Catalysis
Page 17 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
always metallic. The results for the Re binding energies, however, depended strongly on the
pretreatment conditions.
Figure 2 also shows the 4f photoemission spectra of the Re component of the catalysts
with results from curve fitting tabulated in Table 4. In contrast to Pt, the Re component of the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
bimetallic catalysts oxidizes readily when exposed to air. In addition, the degree of Re oxidation
by air exposure was different on the two catalysts studied. The peak fitting results show that
+
7
most of the Re in 8 wt% PtꢀRe/C had an oxidation state of Re , which is consistent with the
previously reported XANES of PtꢀRe bimetallic catalysts supported on Norit Activated Carbon.
The conventional XPS results from this work also revealed that a very small portion of the Re
4
0
+4
was present as Re metal (Re ) and ReO (Re ). For the silicaꢀsupported catalyst, the Re
2
component of the catalyst was more evenly distributed among the various oxidations states
0
2+
4+
7+
corresponding to Re , Re , Re , and Re . In addition, two peaks with binding energies of 47.4
and 49.8 eV were observed, which are close to the binding energies of peaks observed in the
spectra of NH ReO (Figure S3). Because the NH ReO precursor should have decomposed
4
4
4
4
during the 723 K reduction treatment used in the catalyst synthesis procedure, we ascribe these
7
+
38,59
36
peaks to Re ꢀlike species,
which are well dispersed, but weaklyꢀheld on the silica support.
Peaks with binding energies of 47.4 and 49.8 eV were not observed in the 8 wt% PtꢀRe/C
catalyst, even though the same reduction conditions were used for the 8 wt% PtꢀRe/SiO catalyst
2
during synthesis. Evidently, the weaklyꢀheld perrhenate species was not as stable on the carbon
support after the reduction treatment as it was on silica.
The carbonꢀsupported and silicaꢀsupported bimetallic catalysts were also studied after
reduction in H at 393 and 473 K in a cell attached to the conventional XPS system. Great care
2
was taken to ensure that significant oxidation of the samples did not occur while they were being
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 65
Page 18 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
transferred from the reduction cell to the analysis chamber. To test the level of oxidation that
might occur during the transfer, an experiment was performed during which the samples were
ꢀ
6
left in the preparation chamber at 10 mbar for 90 min after reduction. These samples were
compared to those that were transferred as quickly as possible through the preparation chamber,
which was approximately 5 min. The percent contribution of each Re oxidation state from the
curve fitting results of the Re 4f XP spectra for these experiments is summarized in the
Supporting Information (Table S4). Even though the pressure in the preparation chamber was
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ6
maintained at 10 mbar, some oxidation of the rhenium was observed after 90 min. This result
shows that the rate of Re oxidation is extremely fast and great care must be taken to keep the Re
reduced during reaction and characterization experiments. To determine the significance of the
oxidation that occurred after 5 min, the percent contributions of the Re oxidation states at the 5
and 90 min time points were extrapolated to time zero, assuming a linear fit, which are also
shown in the Supporting Information (Table S4) in italicized text. This extrapolation suggests
that the oxidation occurring after 5 min in the preparation chamber was minimal and that the
state of the catalysts after 5 min in the chamber closely resembles the state of the catalyst under
reduced conditions.
Conventional XPS results from the curve fitting of spectra for the catalysts reduced at
393 and 473 K in H are also summarized in Figure 3 and Table 4. As mentioned previously, the
2
Pt 4f spectra revealed that Pt was always in the metallic state on the samples, regardless of how
the catalysts were pretreated prior to XPS. In contrast, a large portion of the Re remained in an
oxidized state even after reduction in H at 393 and 473 K. Furthermore, the distribution of Re
2
oxidation states was influenced by reduction over this narrow temperature range. After a 1 h
0
2+
4+
reduction at 393 K, Re , Re , and Re were observed on both carbonꢀsupported and silicaꢀ
ACS Paragon Plus Environment

Page 19 of 65
ACS Catalysis
Page 19 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
+7
supported catalysts while there was also a small percentage of Re that was observed on the
carbonꢀsupported sample. The silicaꢀsupported catalyst appeared to be slightly more oxidized
2
+
0
than the carbon supported catalyst since the ratio of Re /Re in the silicaꢀsupported catalyst was
higher. The same general behavior was observed after the catalysts were reduced at 473 K – the
silicaꢀsupported catalyst appeared to be slightly more oxidized than the carbonꢀsupported
catalyst. As expected, the distribution of Re oxidation states was shifted towards Re metal when
the catalysts were reduced at 473 K compared to 393 K, but there was still a significant amount
of Re in the +2 and +4 oxidation states on both catalysts under all of the conditions studied.
The Re/Pt ratio of the catalysts as evaluated by XPS was also affected by the support and
reduction temperature. In the right most column of Table 4, the areas under the photoemission
response curves from Pt and Re were used to calculate the atomic ratio of Re/Pt associated with
the bimetallic nanoparticles. The Re/Pt ratio for the carbonꢀsupported catalyst did not appear to
change as the catalyst was reduced at higher temperatures and the surface appeared to be slightly
enriched in Re. The silicaꢀsupported catalyst, however, showed different behavior. As the
reduction temperature was increased, the Re/Pt ratio indicates that the surface became enriched
in Pt, while Re segregated to the core of the catalytic particles.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.3.2 Near ambient pressure X-ray photoelectron spectroscopy
Additional characterization of the Re in 8 wt% PtꢀRe/C exposed to various environments
was obtained using a synchrotronꢀbased near ambient pressure Xꢀray photoelectron spectroscopy
NAPꢀXPS) system. The two major advantages of this system were 1) the ability to tune the
(
photon energy of the Xꢀray source, which allowed a range of sample depths to be probed and 2)
the ability to perform measurements in the presence of reactant gases up to pressures of 0.3
mbar. Unfortunately, the silicaꢀsupported catalyst experienced significant charging effects during
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 65
Page 20 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
irradiation and reliable spectra could not be acquired because this system was not amenable to
utilizing charge compensation devices.
The NAPꢀXPS measurements were performed on 8 wt% PtꢀRe/C at three different kinetic
energies under several different environmental conditions. The kinetic energy was changed by
tuning the incident photon energy for each element of interest so that the photon energy minus
the binding energy of the element equaled the desired kinetic energy of the photoelectrons. By
recording measurements at kinetic energies of 150, 500, and 1000 eV, the oxidation states of Pt
and Re at different distances from the surface of the nanoparticles could be investigated. Table
S5 shows the inelastic mean free path of the electrons from both Pt and Re at the three kinetic
energies studied using the synchrotronꢀbased technique as well as the inelastic mean free path of
the conventional XPS experiments using Al Kα radiation. Since the photon energy could not be
tuned in the conventional XPS experiments, the kinetic energy of photoelectrons differed for Pt
and Re by 30 eV, which is the difference in binding energy for the two elements.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4 shows the NAPꢀXPS of 8 wt% PtꢀRe/C at three different kinetic energies at 393
and 473 K in 0.3 mbar of H . At both of these temperatures, the curveꢀfitting results show that as
2
the kinetic energy of the photoelectrons increased, which corresponded to an increase in the
probing depth, the oxidation state of the Re shifted towards Re metal. This result indicates that
under reducing conditions at the temperatures studied, the Re closest to the surface of the
nanoparticles was present in a range of oxidation states, and Re in the core of the nanoparticle
had a lower oxidation state that was most likely metallic.
The NAPꢀXPS results of the Pt component of the catalyst, presented in Figure S4,
revealed that the Pt was also in the metallic state. Since only one set of 4f and 4f peaks was
7
/2
5/2
observed in the spectra, we conclude that Pt resides primarily in a single state in the bimetallic
ACS Paragon Plus Environment

Page 21 of 65
ACS Catalysis
Page 21 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
samples. The binding energy of the Pt 4f7/2 peak was approximately 71.1 eV for the
measurements obtained with kinetic energies of 500 and 1000 eV, which has been ascribed to
61
metallic Pt. The measurements obtained at a kinetic energy of 150 eV, however, revealed a
shift in the binding energy of Pt of approximately 0.6 eV to around 71.7 eV. The Pt 4f7/2 binding
energies for the NAPꢀXPS measurements are tabulated in Table S6. Duke et al. also observed a
binding energy shift of approximately 0.55 eV in the Pt 4f peak upon the physical addition of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
7
/2
5
7
Re to a Pt(111) single crystal using NAPꢀXPS at a photon energy of 545 eV. Interestingly,
when the Re was removed via oxidation at 500 K, the Pt 4f binding energy returned to that
7
/2
associated with the clean crystal at 71.0 eV. In addition, the Pt L XANES of aluminaꢀ
III
6
2
4
supported and Norit carbonꢀsupported PtꢀRe nanoparticles were shifted to slightly higher
energies than supported monometallic Pt nanoparticles. The fact that the white line intensity of
the monometallic Pt was the same as that in the PtꢀRe bimetallic sample indicates the Pt
4
remained fully reduced in the presence of Re. We ascribe the observed shift in the Pt 4f7/2
binding energy in the current work to an electronic interaction with Re on the surface of the
nanoparticles. As mentioned above, the shift in binding energy was only evident at a kinetic
energy of 150 eV, which suggests the affected Pt is mainly at the surface of the nanoparticles.
The quantitative results of the curve fitting of the NAPꢀXPS experiments in 0.3 mbar of
H and the conventional XPS measurements, after 8 wt% PtꢀRe/C was treated in a separate
2
reduction cell, are shown in Table 5. Trends in the distribution of Re oxidation states in Table 5
show good agreement between the two different XPS techniques. Conventional XPS and NAPꢀ
XPS also show that the distribution of Re oxidation states shifts towards Re metal as the
reduction temperature increased from 393 to 473 K, as indicated by the weighted average
binding energy of the Re photoelectron response, which is tabulated in the table. Table 5 also
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 65
Page 22 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
presents conventional XPS and NAPꢀXPS results of the catalyst measured at room temperature
in UHV. Again, these two techniques show general agreement in the distribution of Re oxidation
states. Both techniques confirm that the Re on the surface of the catalyst was highly oxidized
after it was exposed to air prior to any in situ reduction.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Since glycerol hydrogenolysis takes place in liquid water saturated with H , the influence
2
of water on the distribution of Re oxidation states and the Re/Pt ratio was studied using NAPꢀ
XPS. The curve fitting results of spectra taken at 150 eV, which are the most surface sensitive
measurements, are tabulated in Table 6 and the photoemission spectra are shown in Figure 5.
Interestingly, when 0.15 mbar of H O was introduced with 0.15 mbar of H , little effect on the
2
2
distribution of Re oxidation states, the binding energy of the Re species, and the Re/Pt ratio at
393 and 473 K was observed. Even when the H was removed and 0.3 mbar H O was allowed to
2
2
react with 8 wt% PtꢀRe/C no significant changes in the XPS response of the Re component of the
catalyst were detected.
A separate experiment was then performed to test whether oxidation could be observed
when a strong oxidant was introduced at a pressure of 0.3 mbar. The catalyst was first reduced at
393 K in situ and an XPS measurement was recorded. The catalyst was then cooled to 333 K, 0.3
mbar of O gas was introduced, and after 20 min a NAPꢀXPS measurement was obtained. The
2
results of this experiment are summarized in Figure 6 and Table 6. Severe oxidation of Re
occurred at 333 K in the presence of O compared to the reduced catalyst at 393 K. This control
2
experiment indicates that oxidation of the Re did take place with an oxidant at the pressures used
for the NAPꢀXPS technique.
2.4 IR Spectroscopy of Adsorbed Pyridine
ACS Paragon Plus Environment

Page 23 of 65
ACS Catalysis
Page 23 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
To characterize any Brønsted acid sites associated with the PtꢀRe bimetallic
nanoparticles, DRIFTS of adsorbed pyridine was used. Pyridine is a strong base that is
commonly used to characterize acid sites on solid surfaces using IR spectroscopy because the IR
signature of adsorbed pyridine is different when it is associated with a hydrogen bond, a
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
4,63–69
Brønsted acid site, or a Lewis acid site.
Table S7 shows the IR absorption frequencies of
the three vibrational modes (8a, 19a, and 19b) that are typically used to study the interaction of
pyridine with solid surfaces. In this work, the silicaꢀsupported catalysts were used in the DRIFTS
experiments because carbonꢀsupported catalysts absorbed the IR radiation.
Spectra of pyridine adsorbed on 8 wt% PtꢀRe/SiO ꢀW and 8 wt% Pt/SiO , collected at
2
2
3
73 K after in situ reduction in H at 473 K, are presented in Figure 7. The major peaks at 1597
2
ꢀ1
and 1446 cm are associated with physically adsorbed pyridine, presumably on the silica
63–69
ꢀ1
support.
The peak located at 1541 cm in Figure 7a is associated with the protonated
pyridinium ion and therefore confirms the presence of Brønsted acid sites on 8 wt% PtꢀRe/SiO2ꢀ
W. A spectrum of the unwashed catalyst showed the same features that can also be ascribed to
pyridine bound to a Brønsted acid site. Figure 7b confirmed that acid sites were not present on 8
wt% Pt/SiO2.
Zhang et al. and Ciftci et al. also observed Brønsted acid sites on silicaꢀsupported PtꢀRe
1
2,34
bimetallic catalysts revealed by IR spectroscopy of adsorbed pyridine.
In those works,
however, the authors reported that the number of Brønsted acid sites increased when the catalysts
were exposed to water vapor prior to pyridine. To investigate the role of water vapor, a DRIFTS
experiment was conducted with 8 wt% Re/SiO whereby water vapor was introduced at 473 K
2
before the introduction of pyridine. Figure 8 compares the resulting spectrum from this
experiment to a spectrum of 8 wt% Re/SiO that was not pretreated with water.
2
ACS Paragon Plus Environment

ACS Catalysis
Page 24 of 65
Page 24 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The spectra in Figure 8 suggest that, after reduction in H at 473 K, Re supported on
2
silica mainly functions as a Lewis acid for pyridine. The result is corroborated by the work of
Kawai et al. who observed higher reduction temperatures of Re/Al O lowered the ratio of
2
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
70
Brønsted acid sites to Lewis acid sites while maintaining the total amount of adsorbed pyridine.
The spectra in Figure 8 also suggest that the addition of water vapor hydrates the supported Re
species and converts the Lewis acid sites to Brønsted acid sites. It is possible that catalysts
pretreated with water vapor in the works by Zhang et al. and Ciftci et al. converted weaklyꢀheld
perrhenate species on the silica to Brønsted acid sites rather than converting bimetallic PtꢀRe
sites, which is why we probed a washed catalyst with soluble Re species removed from the
surface. Nevertheless, DRIFTS of adsorbed pyridine confirms that Brønsted acid sites are
associated with PtꢀRe bimetallic particles and the addition of water is not required to form active
sites on catalysts reduced at 473 K.
2.5 Propyl Acetate Hydrolysis
The hydrolysis of propyl acetate, a wellꢀknown Brønsted acidꢀcatalyzed reaction, was
used to obtain an estimate of the number of acid sites present on the PtꢀRe bimetallic catalysts in
liquid water. The hydrolysis of aqueous esters has been commonly used as a probe reaction to
7
1–73
characterize other solid acids.
The conversion of propyl acetate was below the detectable limit at 333 K in 2.1 MPa of
N after 24 h of reaction in the absence of catalyst. The silica support and the Pt/SiO catalyst did
2
2
show a low hydrolysis activity, most likely from the few weak acid sites on SiO . For example,
2
the conversion of propyl acetate after 24 h of reaction of the silica support was 0.2%, while the
conversion on 8 wt% Pt/SiO was slightly lower. The conversion of propyl acetate on 8 wt% Ptꢀ
2
Re/SiO was 4.3% after 24 h, which was significantly higher than the silica support. The rate of
2
ACS Paragon Plus Environment

Page 25 of 65
ACS Catalysis
Page 25 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
propyl acetate hydrolysis remained constant on the bimetallic PtꢀRe catalyst over the course of
2
8 h. Table 7 shows the conversions and rates of propyl acetate hydrolysis on 8 wt% PtꢀRe/SiO2
and 8 wt% PtꢀRe/SiO ꢀW measured after 8 h of reaction. A plot of the conversion as a function
2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
of time is shown in the Supporting Information (Figure S5). The 8 wt% PtꢀRe/SiO ꢀW catalyst
2
exhibited approximately 60% of the rate of the 8 wt% PtꢀRe/SiO catalyst. Interestingly, the
2
same fractional decrease in the rate of glycerol hydrogenolysis was also observed on the washed
catalyst compared to the unwashed catalyst (Table 2). Evidently, the washing procedure removed
about 30 to 40% of the Re sites that are active for both propyl acetate hydrolysis and glycerol
hydrogenolysis. The higher propyl acetate hydrolysis rate of the Reꢀpromoted Pt catalyst
compared to Pt/SiO supports the hypothesis that PtꢀRe bimetallic particles contain acid sites
2
after reduction in H at 393 K, which participate in hydrolysis of propyl acetate at 333 K.
2
Propyl acetate hydrolysis was also performed with 8 wt% Pt/C and 8 wt% PtꢀRe/C (not
shown), but the bare Vulcan support had a very high activity, presumably due to surface
carboxylic acid groups. Nevertheless, 8 wt% PtꢀRe/C still had a higher rate than both the bare
Vulcan support and the 8 wt% Pt/C catalyst.
To obtain an estimation of the number of Re atoms in the bimetallic catalyst that form
acid sites, the rate of hydrolysis on the PtꢀRe catalysts was compared to the rate of hydrolysis in
aqueous HCl. Thus, a reaction was performed in which the amount of HCl used as a catalyst was
equal to the nominal loading of Re used in a typical hydrolysis experiment with PtꢀRe bimetallic
catalyst. In Table 7, the ratio of the propyl acetate conversion on the PtꢀRe catalysts to the
+
conversion in aqueous HCl was then used to infer the number of H equivalents on the PtꢀRe
catalyst. This measure suggests that approximately 10% of the Re on the unwashed catalyst
forms Brønsted acid sites that are active for propyl acetate hydrolysis. Since the acid sites that
ACS Paragon Plus Environment

ACS Catalysis
Page 26 of 65
Page 26 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
catalyze hydrolysis on the unwashed catalyst can be attributed to the PtꢀRe nanoparticles as well
+
as adsorbed perrhenate, it may be more appropriate to conclude that the H equivalents measured
using 8 wt% PtꢀRe/SiO ꢀW was a better estimate of acid sites associated with the bimetallic
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
clusters. Our results suggest that at least 6% of the Re on the catalyst forms Brønsted acid sites.
2.6 Selective Poisoning of Pt-Re Catalyst
To determine whether the Brønsted acid sites associated with the Re component of the Ptꢀ
Re bimetallic catalyst played a role in glycerol hydrogenolysis, the rate of the hydrogenolysis
reaction was measured on a 4 wt% PtꢀRe/C catalyst in the presence of NaOH. The physical
characteristics of this particular catalyst are summarized in Table 1. A carbonꢀsupported
bimetallic catalyst was chosen for this set of experiments because silica is known to dissolve in
7
4
aqueous solutions at high pH. The results from glycerol hydrogenolysis in the presence of
different concentrations of NaOH after 2 h of reaction are shown in Table 8.
Very small amounts of NaOH decreased the rate of glycerol hydrogenolysis significantly
over 4 wt% PtꢀRe/C. When a molar ratio of NaOH/Re=0.5 was used, the rate of reaction
decreased by 37%. The rate decreased even further when the NaOH/Re molar ratio increased to
5. At 100 mM NaOH, the rate actually increased while the selectivity to 1,3ꢀPDO as well as the
Sec:Prim ratio were very low. This increase in rate was attributed to a change in reaction
mechanism because of the added base. Other studies have demonstrated baseꢀcatalyzed
hydrogenolysis of the terminal CꢀO bonds of glycerol takes place in the presence of Pt at high
4,26,27
pH.
In 100 mM NaOH, 17% of the product was actually lactate after 2 h of reaction, which
also supports the change in reaction path. Clearly, addition of NaOH neutralized the acid sites
responsible for activation of glycerol at the secondary carbon and increased the rate of terminal
CꢀO bond activation, thus lowering the Sec:Prim ratio. Nevertheless, a significant decrease in the
ACS Paragon Plus Environment

Page 27 of 65
ACS Catalysis
Page 27 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
rate of glycerol hydrogenolysis was observed as the molar ratio of NaOH to Re was increased
from 0 to 50, before a significant change in the reaction mechanism was reflected in the rate of
reaction or the Sec:Prim ratio.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Additional experiments were performed in which NaCl was added to the reaction mixture
+
in an attempt to decouple the possible influence of Na on activity and selectivity. As shown in
Table 8, the reaction with 10 mM NaCl revealed a rate and product selectivity that was
approximately the same as the reaction without NaCl, indicating that the catalyst was not
negatively affected by the addition of NaCl salt at these conditions. This result also suggests that
the 63% decrease in rate observed in the presence of 10 mM NaOH can be attributed to the
ꢀ
added presence of OH . Interestingly the rate of glycerol hydrogenolysis decreased from 6.7 to
ꢀ1
ꢀ1
3.5 ꢁmol g min when 100 mM NaCl was added to the reaction medium. It is possible that some
of the sodium ions at high concentration exchanged with the acidic protons associated with the
Brønsted acid sites created by the PtꢀRe bimetallic nanoparticles or that the higher concentration
ꢀ
of Cl may have negatively affected the Pt component of the catalyst, since it is well known that
ꢀ
the presence of Cl can coordinate to platinum group metals and negatively affect the catalytic
2
6,75,76
activity.
high in the presence of NaCl.
In summary, results from the addition of NaOH and NaCl to the reaction medium are
Regardless of the effect on absolute activity, the Sec:Prim ratio remained very
consistent with the hypothesis that a Brønsted acid site associated with the Re component of the
catalyst is critical in the activation of secondary CꢀO bonds during glycerol hydrogenolysis on
PtꢀRe bimetallic catalysts.
3
. DISCUSSION
3
.1 Oxidation State of Re
ACS Paragon Plus Environment

ACS Catalysis
Page 28 of 65
Page 28 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The extensive characterization of PtꢀRe bimetallic nanoparticles revealed that Re exists in
a variety of oxidation states when the PtꢀRe samples were reduced in H at 473 K or lower.
2
Daniel et al. used Xꢀray absorption spectroscopy to study PtꢀRe bimetallic catalyst supported on
Norit Carbon and reported that the Re L edge energy shift of the catalyst reduced in H at 473
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
III
2
K corresponded to an average Re oxidation state of approximately +2, which is in general
4
agreement with our XPS results. The Re L edge energy shift was also utilized by Koso et al.
III
and Nakagawa et al. to characterize RhꢀRe and IrꢀRe bimetallic catalysts supported on silica,
respectively, and they also reported that the average Re oxidation state was approximately +2
40,55
after reduction in H at 393 K.
2
The Re/Pt atomic ratio, derived from XPS, of the silicaꢀsupported catalyst PtꢀRe
bimetallic catalyst decreased as the catalyst was reduced, indicating that Re segregated from the
surface of the catalyst to the core of the catalyst under reducing conditions. Somorjai and coꢀ
authors have recently studied the restructuring of the PtPd and RhPd bimetallic systems under
different reaction conditions using atomic ratios derived from a near ambient pressure
77,78
synchrotron based XPS technique similar to that used in this work.
The phenomena of Re
segregation to the core with reduction at elevated temperatures is consistent with the model
8
proposed by Chia et al. for RhꢀRe nanoparticles. In that work, EXAFS results indicated that
bimetallic particles with a Reꢀrich core were formed at high reduction temperatures. Concurrent
results from the catalytic hydrogenolysis of 2ꢀ(hydroxymethyl)tetrahydropyran showed that the
rate decreased with increasing reduction temperature (up to 723 K) indicating that the number of
active sites, which most likely require Re, decreased at the surface. The authors used CO
chemisorption to rule out metal sintering as a cause for the decreased rate. Rhenium segregation
to the core of PtꢀRe bimetallic alloys is also consistent with calculations performed by Greely
ACS Paragon Plus Environment

Page 29 of 65
ACS Catalysis
Page 29 of 64
ACS Catalysis
1
2
3
4
5
6
7
8
9
and Mavrikakis, which suggest that as Re/Pt alloys are heated to higher temperatures the most
7
9
stable configuration consists of subsurface Re with a Pt overlayer.
.2 The Nature of the Brønsted Acid Site
Results from this work confirm that Brønsted acid sites are present on the PtꢀRe
3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
bimetallic catalysts and are most likely associated with the oxophilic Re component, which is
consistent with NH ꢀTPD results suggesting that acid sites are also present on carbonꢀsupported
3
7
,8
RhꢀRe bimetallic catalysts. The DRIFTS spectra in Figure 7 indicate that very few Lewis acid
sites were present under the conditions of the experiment and nearly all of the acid sites on the
silicaꢀsupported PtꢀRe nanoparticles could be attributed to Brønsted acid sites. These Brønsted
acid sites were catalytically active for propyl acetate hydrolysis in liquid water. Since a catalyst
that was washed to remove soluble perrhenate species weaklyꢀheld to the silica support was still
active for propyl acetate hydrolysis, we suggest the acid site is attributed to a partially oxidized
Re species (Re oxidation state < 7+) associated with Pt. Consistent with the idea that the acid site
remains associated with the solid catalyst in liquid water, Chia et al. reported that RhꢀRe
bimetallic catalysts were active for the acidꢀcatalyzed of hydrolysis of fructose to
8
hydroxymethylfurfural in a continuous flow reactor for 250 h.
3.3 The Role of Solvent in Hydrogenolysis on Pt-Re Catalysts
The results from this work also shed light on the role of water in glycerol hydrogenolysis.
Pyridine DRIFTS showed that the addition of water is not required for Brønsted acid sites to be
present on PtꢀRe bimetallic nanoparticles. In addition, the NAPꢀXPS revealed that the addition of
0.3 mbar of water vapor does not change the oxidation state or binding energy of Re in the
reduced catalyst indicating that water does not react with or electronically alter the Re
component of the catalyst significantly after the catalyst is reduced in H at temperatures of 473
2
ACS Paragon Plus Environment