Selective Hydrogenolysis of Glycerol to Propylene Glycol on Supported Pd Catalysts: Promoting Effects of ZnO and Mechanistic Assessment of Active PdZn Alloy Surfaces

Article abstract of DOI:10.1021/acscatal.7b00995

Pd catalysts have received increasing attention for selective hydrogenolysis of glycerol to propylene glycol because of their good hydrothermal stability and high selectivity for cleavage of C-O bonds over C-C bonds. Addition of Zn can facilitate glycerol hydrogenolysis to propylene glycol on Pd surface, but the promoting role of Zn, stability of the resulting active PdZn alloys and reaction mechanism remain largely unexplored. Here, we synthesized monoclinic zirconia-supported PdZn (PdZn/m-ZrO₂) catalysts via an incipient wetness impregnation method. Glycerol hydrogenolysis turnover rates (normalized per surface Pd atom measured by H₂ chemisorption) and propylene glycol selectivity on these PdZn/m-ZrO₂ catalysts depended sensitively on their Zn/Pd molar ratios, and Zn leaching from the PdZn alloy phases led to deactivation of PdZn/m-ZrO₂. Such deactivation was efficiently inhibited when physical mixtures of Pd/m-ZrO₂ and ZnO were directly used in glycerol hydrogenolysis, leading to in situ formation of PdZn alloy layers on Pd surfaces with excellent stability and recyclability. Dependence of turnover rates on glycerol and H₂ concentrations, combined with the primary kinetic isotope effects (k_H/k_D = 2.6 at 493 K), reveals the kinetically relevant step of glycerol hydrogenolysis involving the α-C-H cleavage in 2,3-dihydroxypropanoxide intermediate to glyceraldehyde on PdZn alloys and Pd. Measured rate constants show that the transition state of α-C-H cleavage is more stable because of the stronger oxophilicity of Zn on PdZn alloys than on Pd, which thus facilitates α-C-H cleavage of the Zn-bound intermediate by adjacent Pd on PdZn alloys. Such synergy between Zn and Pd sites accounts for the observed superiority of PdZn alloys to Pd in glycerol hydrogenolysis. (Chemical Equation Presented).

Full text of DOI:10.1021/acscatal.7b00995

Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Article
Selective hydrogenolysis of glycerol to propylene glycol
on supported Pd catalysts: Promoting effects of ZnO and
mechanistic assessment of active PdZn alloy surfaces
Qianhui Sun, Shuai Wang, and Haichao Liu
ACS Catal., Just Accepted Manuscript • Publication Date (Web): 16 May 2017
Downloaded from http://pubs.acs.org on May 16, 2017
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 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
Selective hydrogenolysis of glycerol to propylene glycol
on supported Pd catalysts: Promoting effects of ZnO and
mechanistic assessment of active PdZn alloy surfaces
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
Qianhui Sun,† Shuai Wang,† and Haichao Liu*
Beijing National Laboratory for Molecular Sciences, State Key Laboratory for Structural
Chemistry of Unstable and Stable Species, College of Chemistry and Molecular Engineering,
Peking University, Beijing 100871, China
ABSTRACT:
Pd catalysts receive increasing attention for selective hydrogenolysis of glycerol to propylene
glycol, because of their good hydrothermal stability and high selectivity for cleavage of CꢀO
bonds over CꢀC bonds. Addition of Zn can facilitate glycerol hydrogenolysis to propylene glycol
on Pd surface, but the promoting role of Zn, stability of the resulting active PdZn alloys and
reaction mechanism remain largely unexplored. Here, we synthesized monoclinic zirconiaꢀ
supported PdZn (PdZn/mꢀZrO2) catalysts via incipient wetness impregnation method. Glycerol
hydrogenolysis turnover rates (normalized per surface Pd atom measured by H2 chemisorption)
and propylene glycol selectivity on these PdZn/mꢀZrO2 catalysts depended sensitively on their
Zn/Pd molar ratios, and Zn leaching from the PdZn alloy phases led to deactivation of PdZn/mꢀ
ZrO2. Such deactivation was efficiently inhibited when physical mixtures of Pd/mꢀZrO2 and ZnO
were directly used in glycerol hydrogenolysis, leading to in situ formation of PdZn alloy layers
on Pd surfaces with excellent stability and recyclability. Dependence of turnover rates on
ACS Paragon Plus Environment
1

ACS Catalysis
Page 2 of 32
1
2
3
4
5
6
7
8
9
glycerol and H2 concentrations, combined with the primary kinetic isotope effects (kH/kD = 2.6 at
493 K), unveils the kineticallyꢀrelevant step of glycerol hydrogenolysis involving the αꢀCꢀH
cleavage in 2,3ꢀdihydroxypropanoxide intermediate to glyceraldehyde on PdZn alloys and Pd.
Measured rate constants show that the transition state of αꢀCꢀH cleavage is more stable, due to
the stronger oxophilicity of Zn, on PdZn alloys than on Pd, which thus facilitates αꢀCꢀH cleavage
of the Znꢀbound intermediate by adjacent Pd on PdZn alloys. Such synergy between Zn and Pd
sites accounts for the observed superiority of PdZn alloys to Pd in 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
KEYWORDS: selective hydrogenolysis, glycerol, propylene glycol, PdZn alloy, kinetics, H/D
isotope effects
1. INTRODUCTION
Selective hydrogenolysis of glycerol to propylene glycol provides an economically viable
route for conversion of glycerol byꢀproducts, in large surplus from biodiesel production, to
valueꢀadded products.1ꢀ6 A wide range of transition metals (e.g. Ru,7ꢀ9 Rh,10 Pt,11 Pd,12 Cu,13ꢀ18
Co,19,20 and Ni21) can catalyze this reaction, and have been intensively studied to date.22,23
Previous studies reported that the selectivity to CꢀO cleavage over CꢀC cleavage, required for
synthesis of propylene glycol in glycerol hydrogenolysis, follows an order of Ru < Rh < Pt < Pd
< Cu, while their hydrogenolysis rates show a reverse order.14, 24ꢀ26 In particular, Pd is selective
to propylene glycol, similar to Cu, but with superior activity and hydrothermal stability,12,26
which renders Pdꢀbased catalysts potentially more favorable for the selective glycerol
hydrogenolysis to propylene glycol.
ACS Paragon Plus Environment
2

Page 3 of 32
ACS Catalysis
1
2
Introduction of additional metals (e.g. Co,12,27 Fe,12,27 and Zn28,29) to monometallic Pd
catalysts can modify the property of Pd surface and consequently improve its glycerol
hydrogenolysis rate and selectivity to propylene glycol. Especially, great attention has been paid
to bimetallic PdꢀZn catalysts, because of the readily formation of uniform PdZn alloys and their
wide applications, not only in glycerol hydrogenolysis,28,29 but also in other different reactions,
such as methanol reforming 30ꢀ32 and waterꢀgas shift reactions.33,34 In particular, a high propylene
glycol selectivity of 92.3% was obtained at 80.4% glycerol conversion for glycerol
hydrogenolysis on a PdZn/ZnO@Al2O3 catalyst at 503 K and 3 MPa H2.29 The PdZn catalysts
were mainly prepared via the reduction of dispersed Pd precursors on ZnOꢀbased supports.
However, the use of ZnO encounters two problems: (a) the Zn/Pd ratio of the PdZn alloys cannot
be mediated, which is critical to the catalytic activity of the PdZn alloys;34,35 (b) ZnO support is
unstable in glycerol hydrogenolysis reactions because of its poor hydrothermal stability,28 and
thus needs to be stabilized by structural promoters (e.g. Al2O3).29
3
4
5
6
7
8
9
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
Glycerol hydrogenolysis to propylene glycol on metal surfaces in neutral and basic aqueous
solutions involves glycerol dehydrogenation to glyceraldehyde, glyceraldehyde dehydration to 2ꢀ
hydroxyacrylaldehyde, and sequential hydrogenations of 2ꢀhydroxyacrylaldehyde to acetol and
propylene glycol (Scheme 1).2,14 The first dehydrogenation step has been widely proposed as a
kineticallyꢀrelevant step on different metals, such as Ru, Rh, Pt, Pd, and Cu, based on the good
correlations between their glycerol dehydrogenation activity and hydrogenolysis turnover
rates,8,26 and also on the independence of initial glycerol hydrogenolysis rates on H2 pressures as
observed with Cuꢀbased catalysts.14 However, kinetic and isotopic evidences for the glycerol
hydrogenolysis mechanism on these metal catalysts are rare, and thus the kinetic relevance of the
involved elementary steps remains still ambiguous. Such mechanistic details are clearly required
ACS Paragon Plus Environment
3

ACS Catalysis
Page 4 of 32
1
2
3
4
5
6
to further improve the efficacy of the metal catalysts, and for the specific case of the PdZn alloys,
to elucidate the promoting effect of Zn on the Pd activity in glycerol hydrogenolysis.
7
8
9
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
Scheme 1. Reaction route of glycerol hydrogenolysis to propylene glycol on metal catalysts in
neutral and basic aqueous solutions.
In an attempt to address the aforementioned problem with the stability of the PdZn alloy
phases, we herein prepare PdZn catalysts on hydrothermallyꢀstable monoclinic ZrO2 (mꢀZrO2)
support via two different methods, i.e. incipient wetness impregnation of Pd/mꢀZrO2 with zinc
acetate and physical mixing of Pd/mꢀZrO2 with ZnO. On these catalysts, reaction kinetic and
isotopic studies are carried out to elucidate the reaction mechanism for glycerol hydrogenolysis.
We find that for PdZn/mꢀZrO2 catalysts, the glycerol hydrogenolysis rate and propylene glycol
selectivity sensitively depend on their Zn/Pd ratios. Active PdZn alloy phases can be readily
formed in situ on mꢀZrO2 directly from physical mixtures of Pd/mꢀZrO2 and ZnO with excellent
recyclability in glycerol conversion to propylene glycerol. Rigorous kinetic and isotopic
assessment of the reaction pathways and mechanism shows that the αꢀCꢀH bond cleavage of
glycerol on Pd/mꢀZrO2 and PdZn/mꢀZrO2 is the kineticallyꢀrelevant step in glycerol
hydrogenolysis, which is facilitated by synergism between adjacent Pd and Zn sites on the PdZn
alloys.
2. EXPERIMENTAL SECTION
2.1. Catalyst preparation. 1.0 wt% and 10 wt% Pd nanoparticles supported on monoclinic
ZrO2 (mꢀZrO2, 99%, Alfa Aesar) were prepared by incipient wetness impregnation, and were
denoted as Pd/mZrO2ꢀ1.0 and Pd/mꢀZrO2ꢀ10, respectively. A diluted HCl solution of PdCl2
ACS Paragon Plus Environment
4

Page 5 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
(A.R., Shenyang Research Institute of Nonferrous Metals) was added dropwise to mꢀZrO2. The
resulting solids were treated in ambient air at 383 K overnight, heated to 823 K at 5 K minꢀ1, and
then held at 823 K for 4 h in flowing air (40 mL minꢀ1). These samples (denoted as PdO/mꢀZrO2)
were subsequently treated in flowing 20% H2/N2 (>99.99%, Beijing Huayuan, 40 mL minꢀ1) at
673 K (5 K minꢀ1) for 3 h. Pd dispersions for Pd/mꢀZrO2ꢀ1.0 and Pd/mꢀZrO2ꢀ10, estimated from
H2 chemisorption (described in Section 2.2), were 19.6% and 14.5%, respectively.
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
PdZn/mꢀZrO2 catalysts (denoted as PdZn/mꢀZrO2ꢀx(y), where x represents the weight
percentages for Pd loadings, and y represents the Zn/Pd molar ratio, 0 ≤ y ≤ 8.0 when x = 1.0, and
0 ≤ y ≤ 2.0 when x = 10) were prepared via incipient wetness impregnation of PdO/mꢀZrO2
samples with an aqueous solution of zinc acetate (Zn(AcO)2ꢁ2H2O, 98%, Alfa Aesar). These
samples were treated in ambient air at 383 K overnight, and then in flowing 20% H2/N2 (40 mL
minꢀ1) at 773 K (5 K minꢀ1) for 3 h for the formation of PdZn alloys.30 The compositions of these
catalysts were measured by inductively coupled plasmaꢀatomic emission spectrometry (ICPꢀAES,
Teledyne Leeman Labs). H2ꢀchemisorption showed that Pd dispersions for the PdZn/mꢀZrO2
samples were similar to those of the Pd/mꢀZrO2 samples with the same Pd loadings.
2.2. Catalyst characterization. XRD patterns were recorded on a Rigaku D/MAXꢀ2400
diffractometer in the 2θ range of 30ꢀ60°, using Cu Kα1 radiation (λ = 1.5406 Å) operated at 40
kV and 100 mA. Xꢀray photoelectron spectra (XPS) were collected on an Axis Ultra
spectrometer (Kratos Analytical Ltd.) using monochromatic Al Kα radiation (hv = 1486.6 eV)
operated at 150 W; samples were treated in flowing 20% H2/N2 (40 mL minꢀ1) at 493 K for 2 h in
a process chamber before the XPS characterization, in order to eliminate slow oxidation of the
metal catalysts caused by the unavoidable exposure to ambient air. The binding energies were
calibrated by referring to the Zr 3d5/2 signal of mꢀZrO2 (182.1 eV).36 H2ꢀchemisorption
ACS Paragon Plus Environment
5

ACS Catalysis
Page 6 of 32
1
2
3
4
5
6
7
8
9
(Quantachrome Autosorbꢀ1) was used to measure the dispersion of Pd for Pd/mꢀZrO2 and
PdZn/mꢀZrO2 catalysts, assuming the chemisorption stoichiometry of 1/1 for H/Pd. Samples
were treated in flowing 5% H2/N2 (>99.99%, Beijing Huayuan; 10 mL minꢀ1) at 673 K (5 K minꢀ
1) for 2 h and evacuated at the same temperature for 12 h; H2ꢀchemisorption was performed at
373 K and within 10ꢀ50 kPa H2.
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. Glycerol hydrogenolysis reactions. Catalytic reactions were carried out in a 100 mL
Teflonꢀlined stainless steel autoclave with vigorous stirring at a speed of 700 rpm. A certain
amount of glycerol (A.R., Beijing Chemical) was first added into the autoclave and diluted with
deionized water to 50 g to prepare the glycerol solution with a certain mass concentration (e.g. 5
g glycerol and 45 g water for a 10 wt% glycerol solution). Pdꢀbased catalysts (i.e. Pd/mꢀZrO2
and PdZn/mꢀZrO2), together with additives (e.g. ZnO (A.R.), MgO (A.R.), or Zn(AcO)2ꢁ2H2O
(98%), all purchased from Alfa Aesar and used without further thermal treatments) in cases
mentioned below, were then added to the glycerol aqueous solution. The autoclave was
pressurized with H2 (>99.99%, Beijing Huayuan) after fully purging out the air, and heated to the
reaction temperature (463ꢀ513 K) monitored by a thermocouple inserted in the autoclave. The
concentrations of glycerol and hydrogenolysis products were analyzed by highꢀperformance
liquid chromatography (HPLC; Shimadzu LCꢀ20A) using a RID detector and an Alltech OAꢀ
1000 organic acid column (0.005 mol Lꢀ1 H2SO4 mobile phase, 0.6 mL minꢀ1 flow rate, and 343
K oven temperature). Turnover rates (molglycerol (molsurfaceꢀPdꢁks)−1) were reported as molar
glycerol conversion rates per mole of exposed Pd atoms estimated from the H2ꢀchemisorption
data, as described in Section 2.2; product selectivities were reported on a carbon basis. For
recycling tests, the used catalysts were washed three times with deionized water and three times
with acetone, and then dried in a vacuum oven at ambient temperature before the next cycle.
ACS Paragon Plus Environment
6

Page 7 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
Kinetic isotope effects for the glycerol hydrogenolysis on Pdꢀbased catalysts were examined
with deuterated glycerolꢀ1,1,2,3,3ꢀd5 reactants (98 atom% D, SigmaꢀAldrich), in which all Hꢀ
atoms in the CꢀH bonds of glycerol were replaced by Dꢀatoms. Glycerol hydrogenolysis rates
and selectivities for glycerolꢀ1,1,2,3,3ꢀd5 were measured using the protocols described above for
the protiumꢀform glycerol reactants.
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
3. RESULTS AND DISCUSSION
3.1. Turnover rates, selectivity, and stability of Pd/mꢀZrO2 and PdZn/mꢀZrO2 catalysts.
Hydrogenolysis of glycerol on Pd/mꢀZrO2 and PdZn/mꢀZrO2 catalysts dominantly forms C3 (e.g.
propylene glycol, acetol and 1ꢀpropanol) and C2 (e.g. ethylene glycol) products via cleavage of
CꢀO bonds and CꢀC bonds, respectively (Scheme 2). Examined for comparison, glycerol did not
show detectable conversion on mꢀZrO2 support under the identical conditions, as reported
previously,26 consistent with the requirement of metal function for dehydrogenation of glycerol
to glyceraldehyde as an initial step involved in glycerol hydrogenolysis in neutral aqueous
solutions (Scheme 1). The selectivities of propylene glycol, acetol and 1ꢀpropanol on Pd/mꢀ
ZrO2ꢀ10 (with 10 wt% Pd loading and Pd nanoparticles of 6.9 nm) at 493 K and 6.0 MPa H2,
were 78.5%, 7.8%, and 4.3%, respectively (Table 1), while the selectivity of ethylene glycol was
7.4% at ~20% glycerol conversion. The combined selectivity of the C3 products reached 90.6%,
reflecting the preferential cleavage of CꢀO bond over CꢀC bond for glycerol hydrogenolysis on
Pdꢀbased catalysts.24 The high selectivities to the C3 products (95.0%) were also obtained on
Pd/mꢀZrO2ꢀ1.0 with a lower Pd loading (1.0 wt%) and smaller Pd nanoparticles (5.1 nm). The
turnover rate on Pd/mꢀZrO2ꢀ1.0 was slightly lower than that on Pd/mꢀZrO2ꢀ10 (7.4 vs. 11.1
ACS Paragon Plus Environment
7

ACS Catalysis
Page 8 of 32
1
2
molglycerol (molsurfaceꢀPdꢁks)−1, Table 1), indicating that larger Pd particles are more active in
glycerol hydrogenolysis, as also observed for Ru/mꢀZrO2 catalysts.26
3
4
5
6
7
8
9
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
Scheme 2. Products of glycerol hydrogenolysis on Pdꢀbased catalysts.
Table 1. Turnover rates and selectivities of glycerol hydrogenolysis on Pd/mꢀZrO2 and PdZn/mꢀ
ZrO2 catalysts.a
Turnover
rate
Selectivity (on a carbon basis, %)
Catalyst
(molglycerol
(molsurfaceꢀ
Pdꢁks)−1)
Propylene
glycol
Ethylene
Acetol
glycol
1ꢀPropanol
Pd/mꢀZrO2ꢀ10b
11.1
7.4
78.5
77.4
91.5
93.2
94.1
80.7
7.4
2.6
2.8
2.1
2.5
8.8
7.8
3.0
0.3
0.3
0.8
4.9
4.3
14.6
1.8
Pd/mꢀZrO2ꢀ1.0b
PdZn/mꢀZrO2ꢀ10(1.2)c
PdZn/mꢀZrO2ꢀ1.0(5.0)c
Pd/mꢀZrO2ꢀ1.0+ZnOc
Pd/mꢀZrO2ꢀ1.0+MgOb
119.5
90.9
90.2
7.0
1.2
2.5
1.5
aReaction conditions: 493 K, 6.0 MPa H2, 10 wt% glycerol in water, 1 g catalyst with a 1 wt% Pd
loading or 0.1 g catalyst with a 10 wt% Pd loading; additive: 0.04 g ZnO or 0.2 g MgO. 20 h,
~20% glycerol conversion. c4 h, ~40% glycerol conversion.
b
Upon formation of PdZn alloy phases on the Pd/mꢀZrO2 catalysts, as described in Section
2.1, via treating them with impregnated Zn(AcO)2 in flowing H2 at 773 K, the glycerol
hydrogenolysis rate and propylene glycol selectivity significantly increased, depending on Zn/Pd
molar ratios. Figure 1 shows that the turnover rates on PdZn/mꢀZrO2ꢀ10(y) (0 ≤ y (Zn/Pd molar
ACS Paragon Plus Environment
8

Page 9 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
ratio) ≤ 2.0; 493 K, 6.0 MPa H2, 10 wt% glycerol aqueous solution henceforth) increased from
11.1 to 119.5 molglycerol (molsurfaceꢀPdꢁks)−1 with increasing the Zn/Pd ratio from 0 to 1.2, and then
decreased to 50.9 molglycerol (molsurfaceꢀPdꢁks)−1 with a further increase of the Zn/Pd ratio to 2.0.
Concurrently, the selectivity to propylene glycol increased from 78.5% to a maximum value of
91.5% with increasing the Zn/Pd ratio to 1.2, at the expense of the selectivities to acetol (7.8%
vs. 0.3%), 1ꢀpropanol (4.3% vs. 1.8%), and ethylene glycol (7.4% vs. 2.8%) (Table 1). Similar
effects of Zn/Pd ratios were also observed for PdZn/mꢀZrO2ꢀ1.0 in the range of 0ꢀ8.0 (Table 1
and Fig. S1 in Supporting Information (SI)). At the optimal Zn/Pd ratio of 5.0, the turnover rate
and propylene glycol selectivity were increased to 90.9 molglycerol (molsurfaceꢀPdꢁks)−1 and 93.2%,
respectively, showing the superiority of the PdZn alloys to monometallic Pd phase for the
selective hydrogenolysis of glycerol to propylene glycol, irrespective of the Pd loadings (Table
1).
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
Figure 1. Effects of Zn/Pd molar ratio on turnover rates and selectivities in glycerol
hydrogenolysis on PdZn/mꢀZrO2ꢀ10(y) (0 ≤ y (Zn/Pd molar ratio) ≤ 2.0, 0.1 g catalyst, 493 K,
6.0 MPa H2, 4 h, 10 wt% aqueous glycerol solution). Acetol selectivities on all catalysts were
below 2% and were not shown here.
ACS Paragon Plus Environment
9

ACS Catalysis
Page 10 of 32
1
2
3
4
5
6
7
8
9
The stability of the active PdZn alloys under glycerol hydrogenolysis conditions was
examined using the PdZn/mꢀZrO2ꢀ10 sample with the optimal Zn/Pd molar ratio of 1.2. As
shown in Figure 2, the hydrogenolysis turnover rates decreased from 119.5 to 35.3 molglycerol
(molsurfaceꢀPdꢁks)−1 over seven consecutive runs (493 K, 6.0 MPa H2, 4 h per run), in concomitance
with a slight decrease of propylene glycol selectivity from 94.8% to 90.5%. Characterization of
the catalyst by XRD and XPS showed no detectable change in the Pd particle size or oxidation
state after the seventh run (as discussed below in Section 3.3). However, elemental analysis
showed that the PdZn/mꢀZrO2ꢀ10 catalyst possessed a Zn/Pd molar ratio of only 0.68 after the
seventh run, corresponding to a loss of 43% Zn during the recycling. The loss of Zn apparently
led to the decrease of the turnover rate and propylene glycol selectivity, in line with the observed
effects of Zn/Pd ratio on the activity and selectivity of PdZn alloys in glycerol hydrogenolysis
(Fig. 1).
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
Figure 2. Turnover rates and selectivities to propylene glycol for seven consecutive cycles of
glycerol hydrogenolysis on PdZn/mꢀZrO2ꢀ10(1.2) (0.1 g catalyst, 493 K, 6.0 MPa H2, 4 h, 10 wt%
aqueous glycerol solution).
ACS Paragon Plus Environment
10

Page 11 of 32
ACS Catalysis
1
2
We surmise that metallic Zn is readily oxidized by H2O to form Zn2+ ions at hydrothermal
conditions, and the dissolution of Zn2+ ions in the aqueous solution leads to the loss of Zn from
the active PdZn alloys during glycerol hydrogenolysis reaction. This unstable nature of metallic
Zn renders the PdZn alloys, even supported on stable mꢀZrO2 surface, impractical for reactions at
such hydrothermal conditions of glycerol hydrogenolysis, in spite of their higher initial turnover
rate and propylene glycol selectivity. Considering such nature of Zn and the consequent
instability of PdZn alloys, in the next section, we report that direct use of physical mixtures of
Pd/mꢀZrO2 and ZnO (denoted as Pd/mꢀZrO2+ZnO) leads to the in situ formation of active PdZn
alloys on Pd surfaces under the glycerol hydrogenolysis conditions, and the presence of ZnO can
provide Zn2+ ions to prevent the PdZn alloys from Zn leakage and consequently render them
stable.
3
4
5
6
7
8
9
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
3.2. Turnover rates, selectivity, and stability of in situ formed PdZn alloys. ZnO was
chosen here as a Zn2+ donor for PdZn alloys, because ZnO can partially hydrate to Zn(OH)2,
which then dissolves in hot water to from Zn2+ and OHꢀ ions. The concentration of OHꢀ ions
derived from ZnO was limited by its solubility (0.3 mg mLꢀ1 at 493 K, measured in glycerol
hydrogenolysis conditions, Table S1, SI), and did not affect the glycerol hydrogenolysis turnover
rate or propylene glycol selectivity, as evidenced by a control experiment using MgO (Table 1),
which is more soluble than ZnO in H2O.37 Similar glycerol hydrogenolysis turnover rates (7.4 v.s
7.0 molglycerol (molsurfaceꢀPdꢁks)−1) and selectivities of propylene glycol (77.4% vs. 80.7%) were
obtained on Pd/mꢀZrO2ꢀ1.0 and on its physical mixture with MgO (Table 1), although the use of
MgO provided a higher concentration of OHꢀ ions than the case of ZnO. Therefore, ZnO can be
regarded only as a source of Zn2+ ions in glycerol hydrogenolysis.
ACS Paragon Plus Environment
11

ACS Catalysis
Page 12 of 32
1
2
3
4
5
6
7
8
9
Figure 3 shows the glycerol hydrogenolysis turnover rates and selectivities as a function of
ZnO amount on Pd/mꢀZrO2ꢀ1.0+ZnO. As the amount of ZnO increased from 0 to 0.04 g, the
rates increased from 7.4 to 90.2 molglycerol (molsurfaceꢀPdꢁks)−1 at 493 K (Fig. 3), and then gradually
decreased to 79.9 molglycerol (molsurfaceꢀPdꢁks)−1 as the ZnO amount further increased to 0.30 g. The
selectivity to propylene glycol monotonically increased to 94.1%, which then remained almost
constant upon addition of ZnO larger than 0.04 g (Fig. 3). Similar promoting effects were also
observed in the control experiments when ZnO was replaced by soluble Zn(AcO)2 salt (Fig. S2,
SI), confirming that for Pd/mꢀZrO2+ZnO, it is the Zn2+ ions, derived from the dissolution of ZnO,
to play the actual role in improving the turnover rate and selectivity of Pd/mꢀZrO2.
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
Figure 3. Turnover rates and selectivities in glycerol hydrogenolysis as a function of ZnO
amount on Pd/mꢀZrO2ꢀ1.0 physically mixed with different amounts of ZnO (1 g Pd/mꢀZrO2ꢀ1.0,
493 K, 6.0 MPa H2, 4 h, 10 wt% aqueous glycerol solution). Acetol selectivities on all catalysts
were below 2% and were not shown here.
Such promoting effects of ZnO and Zn2+ ions resemble the observed effects of the Zn/Pd
ratios for the PdZn/mꢀZrO2 catalysts (Figs. 1 and S1). For example, the maximum glycerol
ACS Paragon Plus Environment
12

Page 13 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
hydrogenolysis turnover rate and the corresponding propylene glycol selectivity on Pd/mꢀZrO2ꢀ
1.0+ZnO (0.04 g ZnO, Fig. 3) were close to those obtained from the optimal PdZn/mꢀZrO2ꢀ
1.0(5.0) catalyst (Fig. S1) (90.2 vs. 90.9 molglycerol (molsurfaceꢀPdꢁks)−1, 94.1% vs. 93.2%, Table 1).
Moreover, on the two catalysts, the apparent activation energies for glycerol hydrogenolysis at
463ꢀ513 K, estimated using the Arrhenius equation, were essentially same (123 ± 5 vs. 122 ± 3
kJ molꢀ1, Fig. 4), which were lower than the value for Pd/mꢀZrO2ꢀ1.0 (135 ± 2 kJ molꢀ1). These
similarities between Pd/mꢀZrO2+ZnO and PdZn/mꢀZrO2 indicate the in situ formation of PdZn
alloys on Pd/mꢀZrO2 in the presence of Zn2+ ions for Pd/mꢀZrO2+ZnO during the glycerol
hydrogenolysis, as indeed confirmed by XRD and XPS in in the next section.
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
Figure 4. Effects of temperature on turnover rates in glycerol hydrogenolysis on Pd/mꢀZrO2ꢀ
1.0+ZnO, PdZn/mꢀZrO2ꢀ1.0(5.0) and Pd/mꢀZrO2ꢀ1.0 (1 g catalyst, 6.0 MPa H2, 10 wt% aqueous
glycerol solution, ~20% glycerol conversion obtained by varying reaction time). Solid lines
represent exponential regression fits.
It is noted that the physical mixtures of Pd/mꢀZrO2 and ZnO are stable and recyclable in
glycerol hydrogenolysis. Representative results are shown in Figure 5. The turnover rate and
propylene glycol selectivity on Pd/mꢀZrO2ꢀ1.0+ZnO remained constant for five consecutive runs
ACS Paragon Plus Environment
13

ACS Catalysis
Page 14 of 32
1
2
at 493 K and 6.0 MPa H2, being 82.0ꢀ85.9 molglycerol (molsurfaceꢀPdꢁks)−1 and 94.1ꢀ93.7%,
respectively, within the kineticꢀcontrolled regime. Analysis of the filtrate after each run by ICPꢀ
AES showed no detectable leaching of Pd, while Zn2+ ions were detected in a concentration of
0.3 mg mLꢀ1, reflecting the solubility of ZnO at the reaction conditions. These constant rates and
selectivities on Pd/mꢀZrO2+ZnO thus suggest that the presence of Zn2+ ions derived from ZnO
stabilizes the active PdZn alloys during glycerol hydrogenolysis, which would otherwise undergo
severe deactivation, as observed with PdZn/mꢀZrO2 (Fig. 2).
3
4
5
6
7
8
9
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
Such high stability of the physical mixtures of Pd/mꢀZrO2 and ZnO implies their potential as
a facile source of the active PdZn alloys for practical application to the selective synthesis of
propylene glycol from glycerol hydrogenolysis. Furthermore, even at 100% glycerol conversion,
the propylene glycol selectivity remained as high as 91.8%, corresponding to a yield of 91.8%,
on Pd/mꢀZrO2ꢀ1.0+ZnO (493 K, 6.0 MPa H2; Fig. S3, SI).
Figure 5. Turnover rates and selectivities to propylene glycol for five consecutive cycles of
glycerol hydrogenolysis on Pd/mꢀZrO2ꢀ1.0+ZnO (1 g Pd/mꢀZrO2ꢀ1.0, 0.2 g ZnO, 493 K, 6.0
MPa H2, 4 h, 10 wt% aqueous glycerol solution).
ACS Paragon Plus Environment
14

Page 15 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
3.3. Structural characterization of Pd and PdZn catalysts before and after glycerol
hydrogenolysis reaction. The monometallic Pd and PdZn alloy catalysts examined in this work
were characterized by XRD and XPS, and the representative results are shown in Figures 6 and
7. Pd/mꢀZrO2ꢀ10 exhibits diffraction peaks at 40.1° and 46.7°, characteristic of metallic Pd, in
addition to those of mꢀZrO2 support (Figs. 6a and b). For PdZn/mꢀZrO2ꢀ10(1.2), its diffraction
peaks appeared at 41.2° and 44.1° (Fig. 6c), assigning to the (111) and (200) planes of PdZn
alloy phases, respectively;38 no diffractions for metallic Pd or any phases related to metallic Zn
(e.g. 36.3° and 43.2°)39 and ZnO (e.g. 31.8° and 36.3°)40 were detected, indicating that PdZn
alloys were solely present on mꢀZrO2 surface. XPS spectra confirmed the formation of PdZn
alloys on mꢀZrO2. Relative to the binding energies of the Pd 3d3/2 and 3d5/2 signals for Pd/mꢀ
ZrO2ꢀ10 (340.6 and 335.3 eV, Fig. 7a), these values for PdZn/mꢀZrO2ꢀ10(1.2) shifted by 0.7 eV
to 341.3 and 336.0 eV, respectively (Fig. 7b), which reflects an electronic transfer from Pd to Zn
in the PdZn alloys.30 Similar XRD and XPS results (Figs. 6d and 7c) were obtained for PdZn/mꢀ
ZrO2 after the glycerol hydrogenolysis reaction (493 K, 6.0 MPa H2, 4h), suggesting that this
catalyst still contained only PdZn alloys on mꢀZrO2, in spite of the detected leaching of the Zn
component.
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
To confirm the formation of the PdZn alloys from the physical mixtures of Pd/mꢀZrO2ꢀ10
and ZnO during the glycerol hydrogenolysis reaction, XRD patterns were collected for Pd/mꢀ
ZrO2ꢀ10 in the presence of ~0.60 mg mLꢀ1 Zn2+ ions (using Zn(AcO)2ꢁ2H2O to mimic ZnO for
simplicity) after the reaction. The pattern (Fig. 6e) shows only characteristic diffraction peaks of
metallic Pd (at 2θ = 40.1° and 46.7°) with no any detectable peaks of PdZn alloys (at 2θ = 41.2°
and 44.1°), similar to the pattern before the reaction (Fig. 6a), thus indicating that the bulk
structure of Pd nanoparticles was not affected by Zn2+ ions during the reaction. Differently, the
ACS Paragon Plus Environment
15

ACS Catalysis
Page 16 of 32
1
2
3
4
5
6
7
8
9
corresponding XPS spectrum after the reaction (Fig. 7d) shows the binding energies of Pd 3d3/2
and 3d5/2 at 341.3 and 336.0 eV, respectively, identical to those for the PdZn alloys on PdZn/mꢀ
ZrO2ꢀ10 (Fig. 7b). Taken together, these results suggest that the PdZn alloy layers are formed,
most likely in the form of thin layers, via the reduction of Zn2+ ions on Pd surfaces during
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
Figure 6. Xꢀray diffraction patterns of (a) mꢀZrO2, (b) Pd/mꢀZrO2ꢀ10 and (c) PdZn/mꢀZrO2ꢀ
10(1.2) before glycerol hydrogenolysis reaction, and (d) PdZn/mꢀZrO2ꢀ10(1.2) and (e) a physical
mixture of 0.1 g Pd/mꢀZrO2ꢀ10 and 0.1 g Zn(AcO)2ꢁ2H2O after glycerol hydrogenolysis reaction
(493K, 6.0 MPa H2, 4h, 10 wt% glycerol aqueous solution).
ACS Paragon Plus Environment
16

Page 17 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
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
Figure 7. XPS spectra of Pd 3d and Zr 3p signals for (a) Pd/mꢀZrO2ꢀ10 and (b) PdZn/mꢀZrO2ꢀ
10(1.2) before glycerol hydrogenolysis reaction, and (c) PdZn/mꢀZrO2ꢀ10(1.2) and (d) a physical
mixture of 0.1 g Pd/mꢀZrO2ꢀ10 and 0.1 g Zn(AcO)2ꢁ2H2O after glycerol hydrogenolysis reaction
(493K, 6.0 MPa H2, 4h, 10 wt% glycerol aqueous solution).
3.4. Effects of glycerol and H2 concentrations on turnover rates and the glycerol
hydrogenolysis mechanism on PdZn alloy surfaces. The effects of concentrations of glycerol
and H2 on glycerol hydrogenolysis turnover rates and selectivities were examined at 493 K on
PdZn alloys within kineticꢀcontrolled regime (at ~20% glycerol conversions). The PdZn alloys
were formed in situ on mꢀZrO2 directly from physical mixtures of Pd/mꢀZrO2 and ZnO under the
glycerol hydrogenolysis conditions. The aforementioned high stability of the PdZn alloys allows
rigorous kinetic and isotopic assessment of the reaction mechanism and the role of Zn in glycerol
hydrogenolysis using Pd/mꢀZrO2+ZnO. As shown in Figure 8a, the glycerol hydrogenolysis rates
increased almost linearly from 88.0 to 211 molglycerol (molsurfaceꢀPdꢁks)−1 with increasing glycerol
concentration from 0.5 to 2.1 mol Lꢀ1 (i.e. 5ꢀ20 wt% aqueous solutions), and then approached a
ACS Paragon Plus Environment
17

ACS Catalysis
Page 18 of 32
1
2
constant value of 270 molglycerol (molsurfaceꢀPdꢁks)−1 at higher concentrations above 4.4 mol Lꢀ1 (i.e.
40 wt%). This rate dependence on the glycerol concentration indicates that the PdZn alloy
surfaces are densely covered with the strongly adsorbed glycerol or alkoxide intermediates
derived from the dissociation of hydroxyls in glycerol at high glycerol concentrations. However,
propylene glycol selectivity was essentially unaffected by glycerol concentrations, which kept
between 93.1ꢀ96.0% (Fig. S4a, SI).
3
4
5
6
7
8
9
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
As H2 partial pressure increased from 1.0 to 7.0 MPa, the glycerol hydrogenolysis rates
monotonically decreased from 298 to 138 molglycerol (molsurfaceꢀPdꢁks)−1 (Fig. 8b). The propylene
glycol selectivity slightly varied in the range of 91.0ꢀ93.7% (Fig. S4b), except at the low H2
pressure of 1.0 MPa, which led to a high acetol selectivity (13.5%) at the expense of propylene
glycol (56.8%) because of the inefficient hydrogenation of acetol to propylene glycol. The strong
inhibition of H2 pressure on glycerol hydrogenolysis in such a broad range of H2 pressures (1.0ꢀ
7.0 MPa), taken together with weak adsorption of H2 on PdZn alloys (shown below), indicates
that glycerol hydrogenolysis on PdZn alloys is kinetically limited by glycerol dehydrogenation,
which is thermodynamically unfavorable at high H2 pressures, consistent with the kinetic isotope
effects and reaction mechanism for glycerol hydrogenolysis, as discussed below.
The hydrogenolysis turnover rates of deuterated glycerol reactants (i.e. CD2OHꢀCDOHꢀ
CD2OH) were 1.6 times lower than those for the protiumꢀform reactants (i.e. CH2OHꢀCHOHꢀ
CH2OH) at 493 K (Table 2), while their selectivities to propylene glycol (89.2% vs. 90.5%) and
ethylene glycol (5.8% vs. 7.7%) remained essentially identical. This primary kinetic isotope
effect (kH/kD = 2.6) on glycerol hydrogenolysis rates indicates that cleavage of CꢀH bonds is
involved in the kineticallyꢀrelevant steps, which is consistent with regression fittings of the
glycerol hydrogenolysis rates with mechanismꢀbased rate equations, as discussed below.
ACS Paragon Plus Environment
18

Page 19 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
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
Figure 8. Effects of (a) glycerol concentration and (b) H2 pressure on turnover rates in glycerol
hydrogenolysis on Pd/mꢀZrO2ꢀ1.0+ZnO (1 g Pd/mꢀZrO2ꢀ1.0, 0.2 g ZnO, 493 K, ~20% glycerol
conversion obtained by varying reaction time; (a) 6.0 MPa H2; (b) 10 wt% aqueous glycerol
solution). Dashed curves represent regressed fits to the functional form of Equation 1.
Table 2. Turnover rates and selectivities for glycerol and deuterated glycerol in hydrogenolysis
reactions on Pd/mꢀZrO2ꢀ1.0+ZnOa
Selectivity (%)
Propylene glycol
Turnover rate
(molglycerol(molsurfaceꢀPd·ks)ꢀ1)
Reactant
Ethylene glycol
HOCH2CH(OH)CH2OH
HOCD2CD(OH)CD2OH
127
90.5
89.2
7.7
5.8
48.7
aReaction conditions: 0.3 g Pd/mꢀZrO2ꢀ1.0, 0.1 g ZnO, 493 K, 6.0 MPa H2, 0.5 h, 0.5 g glycerol
in 40 g water.
Glycerol hydrogenolysis on metal surfaces (such as Ru, Rh, Pt, Pd, and Cu) in neutral or
basic aqueous solutions has been proved to occur via a dehydrogenationꢀdehydrationꢀ
hydrogenation pathway.14,26 A glycerol molecule adsorbed on metal surface first dissociates to
form a bound alkoxide species (i.e. 2,3ꢀdihydroxypropanoxide) and a H atom on two vicinal
metal sites (Step 2, Scheme 3), which readily occurs even at modest temperatures.41 An αꢀH
atom in the bound alkoxide species is then abstracted by its adjacent vacant metal site to form
ACS Paragon Plus Environment
19

ACS Catalysis
Page 20 of 32
1
2
3
4
5
6
7
8
9
glyceraldehyde (Step 3, Scheme 3). Glyceraldehyde can desorb from the metal surfaces into
aqueous solution (Step 5, Scheme 3) and dehydrate to 2ꢀhydroxyacrylaldehyde with the aid of
OHꢀ ions for the cleavage of a terminal CꢀO bond (Step 6, Scheme 3), which subsequently
hydrogenates to acetol and propylene glycol. Alternatively, glyceraldehyde can undergo
decarbonylation to ethylene glycol and CO (Step 7, Scheme 3). The CO molecule ultimately
converts to CO2 via waterꢀgas shift reactions or to CH4 with hydrogenolysis reactions.26 In
particular, αꢀCꢀH cleavage in 2,3ꢀdihydroxypropanoxide to glyceraldehyde (Step 2, Scheme 3)
has been proposed to be the kineticallyꢀrelevant step on different metal surfaces (i.e. Ru, Rh, Pt,
Pd, and Cu),14, 26 and the subsequent cleavage of CꢀO and CꢀC bonds (Steps 6 and 7, Scheme 3)
determines the relative selectivities of propylene glycol and ethylene glycol.
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. Adsorption of glycerol on metal surface to form adsorbed glycerol species
2. Dissociation of glycerol* to form a bound alkoxide species (2,3ꢀdihydroxypropanoxide) and a
bound H atom
3. αꢀCꢀH cleavage of 2,3ꢀdihydroxypropanoxide to form absorbed glyceraldehyde species
ACS Paragon Plus Environment
20

Page 21 of 32
ACS Catalysis
1
2
3
4
4. H2 desorption
5
6
7
8
9
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
5. Glyceraldehyde desorption
HO H
C H
HO
K5
C
OH
H
H
C
+
O
HO
O
M
M
M
M
M
M
6. Dehydration of glyceraldehyde to form 2ꢀhydroxyacrylaldehyde in H2O
7. Decarbonylation of absorbed glyceraldehyde to form ethylene glycol and CO
Scheme 3. Proposed sequence of elementary steps and associated kinetic constants and
equilibrium constants for glycerol hydrogenolysis on metal surface.26
The kinetic relevance of the αꢀCꢀH cleavage in 2,3ꢀdihydroxypropanoxide (Step 3, Scheme
3), taken together with the assumption of the quasiꢀequilibrated nature for Steps 1, 2, 4, and 5 in
the proposed elementary steps (Scheme 3), gives the glycerol hydrogenolysis rate equation as
ꢋ
ꢊ
ꢄ
ꢈꢄ ꢈꢂ
ꢀꢁꢂꢃ ꢅꢆꢇ ꢉꢊ
r =
(1)
ꢊ
ꢋ
ꢊ
ꢋ
ꢊ
ꢋ
ꢊ
ꢄ
ꢈ
ꢄ
ꢈꢄ ꢈꢂ
ꢌ1 + ꢍꢋ ꢅꢆꢇ + ꢍꢎꢏꢐ ꢅꢆꢇ ꢉꢊ
ꢄ
ꢈ
+ ꢍꢑ ꢉꢊ
ꢒ
ACS Paragon Plus Environment
21

ACS Catalysis
Page 22 of 32
1
2
3
4
Here, [Gly] is the activity of glycerol in water, [H2] is the partial pressure of H2 in gas phase; ꢀꢓ
5
6
7
and ꢍ are the respective kinetic constant and equilibrium constant for Step j in Scheme 3; ꢀꢁꢂꢃ
ꢓ
8
9
is the apparent kinetic constant for glycerol hydrogenolysis in aqueous phase:
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
ꢕ
ꢂ
ꢖ
ꢀꢁꢂꢃ = ꢀꢔꢍꢋꢍꢊꢍꢑ
(2)
and ꢍꢎꢏꢐ is the apparent equilibrium constant for the formation of bound 2,3ꢀ
dihydroxypropanoxide from glycerol on metal surface:
ꢕ
ꢂ
ꢖ
ꢍꢎꢏꢐ = ꢍꢋꢍꢊꢍꢑ
(3)
Equation 1 accurately described the rates measured in a broad range of reactant
concentrations (5ꢀ50 wt% glycerol aqueous solutions, and 1.0ꢀ7.0 MPa H2 in Fig. 8). A parity
plot for the measured and predicted rates was shown in Figure 9a, and the corresponding
regressed parameters of Equation 1 for the PdZn alloys were listed in Table 3. These excellent
regression fits indicate that the proposed elementary steps in Scheme 3 reflect the mechanism of
glycerol hydrogenolysis on PdZn alloys, and the glycerol hydrogenolysis rates are mediated by
the αꢀCꢀH cleavage in glycerol to form glyceraldehyde intermediates, as also found on supported
monometallic catalysts.26 The Zn component in the PdZn alloys appears to stabilize the αꢀCꢀH
cleavage transition states, leading to higher glycerol hydrogenolysis turnover rates. The
regressed values for the relevant equilibrium constants in Equation 1 show that the
ꢕ
ꢖ
ꢄ
ꢈꢄ ꢈꢂ
ꢍꢎꢏꢐ ꢅꢆꢇ ꢉꢊ
term, which reflects the coverage of 2,3ꢀdihydroxypropanoxide on metal
surface, dominates the denominator (Table 3). This suggests that 2,3ꢀdihydroxypropanoxide is
the most abundant surface species (MASI) on the PdZn alloys in glycerol hydrogenolysis.
ACS Paragon Plus Environment
22

Page 23 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
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
Figure 9. Parity plots for the measured and predicted rates of glycerol hydrogenolysis (Eq. (1))
on (a) a physical mixture of Pd/mꢀZrO2ꢀ1.0 and ZnO and (b) Pd/mꢀZrO2ꢀ1.0 with the regressionꢀ
fitted parameters shown in Table 3.
Table 3. Regressed kinetic parameters for glycerol hydrogenolysis on Pd/mꢀZrO2ꢀ1.0+ZnO and
Pd/mꢀZrO2ꢀ1.0 catalysts at 493 K.
ꢋ
ꢊ
ꢀꢁꢂꢃ
((mol/L)ꢀ1MPa1/2sꢀ1)
980 ± 70
ꢍꢋ
ꢍꢎꢏꢐ
ꢍꢑ
Catalyst
(mol/L)ꢀ1
< 0.01
< 0.01
((mol/L)ꢀ1 MPa1/2)
(MPa) ꢀ1/2
0.15 ± 0.02
< 0.01
Pd/mꢀZrO2ꢀ1.0+ZnO
Pd/mꢀZrO2ꢀ1.0
0.65 ± 0.03
15 ± 2
0.15 ± 0.05
The dependence of glycerol hydrogenolysis rates on the concentrations of glycerol and H2
was also examined for Pd/mꢀZrO2, in order to unveil the promoting effects of Zn on the Pdꢀ
catalyzed glycerol hydrogenolysis reactions. Figure 10 shows that the effects of the reactant
concentrations on the glycerol hydrogenolysis rates resemble the trends observed on the PdZn
alloys (Fig. 8). The glycerol hydrogenolysis rates on Pd/mꢀZrO2, however, were more sensitive
to glycerol concentration and less sensitive to H2 pressure, relative to those on the PdZn alloys,
under the identical reaction conditions. As discussed for the PdZn alloys, the hydrogenolysis
ACS Paragon Plus Environment
23

ACS Catalysis
Page 24 of 32
1
2
3
4
5
6
7
8
9
rates on Pd/mꢀZrO2 were also accurately described by Equation 1 (Fig. 10), and the predicted
rates agree excellently with the measured rates (Fig. 9b; the regressed parameters are shown in
Table 3). These results indicate that the glycerol hydrogenolysis rates on Pd/mꢀZrO2 are also
kinetically limited by the αꢀCꢀH cleavage in glycerol (Step 3, Scheme 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
Figure 10. Effects of (a) glycerol concentration and (b) H2 pressure on turnover rates in glycerol
hydrogenolysis on Pd/mꢀZrO2ꢀ1.0 (493 K, ~20% glycerol conversion obtained by varying
catalyst amount and reaction time; (a) 6.0 MPa H2; (b) 10 wt% aqueous glycerol solution).
Dashed curves represent regressed fits to the functional form of Equation 1.
The regressed value for ꢀꢁꢂꢃ on the PdZn alloys (i.e. Pd/mꢀZrO2+ZnO) is over 60 times
larger than that for Pd (i.e. Pd/mꢀZrO2) (980 ± 70 vs. 15 ± 2 (mol/L)ꢀ1 MPa1/2 sꢀ1, Table 3),
consistent with the much higher hydrogenolysis turnover rates on the PdZn alloys (Figs. 1 and 3).
As shown in Scheme 4, the ꢀꢁꢂꢃ parameter reflects the difference (ꢂG‡CꢀH) between the Gibbs
free energy of the αꢀCꢀH cleavage transition state (G‡CꢀH) and the sum of free energies of a
glycerol molecule in aqueous solution (GGly) and of a bare metal surface (GM):
ꢗꢅ‡
= ꢅ‡ꢘꢂꢙ ꢚ ꢅꢛꢜꢝ ꢚ ꢅꢞ
(4)
ꢘꢂꢙ
ACS Paragon Plus Environment
24

Page 25 of 32
ACS Catalysis
1
2
The ꢂG‡CꢀH value for the PdZn alloys is lower than the value for Pd by 17 kJ molꢀ1 (i.e. ꢂꢂG‡
3
4
CꢀH
5
6
7
8
in Scheme 4, 493 K), reflecting that the presence of Zn stabilizes the αꢀCꢀH cleavage transition
states on the Pd surfaces. Such stabilization conferred by the PdZn alloys is also found for the
adsorbed 2,3ꢀdihydroxypropanoxide species, the precursor of the αꢀCꢀH cleavage transition state,
as evidenced by the larger equilibrium constant for the formation of 2,3ꢀdihydroxypropanoxide
on PdZn than on Pd (ꢍꢎꢏꢐ: 0.65 ± 0.03 vs 0.15 ± 0.03 (mol/L)ꢀ1 MPa1/2, 493 K, Table 3).
9
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
Scheme 4. Schematic reaction coordinate diagram of Gibbs free energy in glycerol
hydrogenolysis on Pd and PdZn alloy surfaces. ꢂG‡CꢀH represents the experimentally measurable
free energy barrier for glycerol hydrogenolysis.
The formation of PdZn alloys from the reduction of Zn2+ ions on the Pd surfaces can lead to a
change in the electronic properties of the Pd sites (i.e. the known electronic effects of bimetallic
catalysis24) and also to a collaboration between vicinal Pd and Zn sites in the kineticallyꢀrelevant
αꢀCꢀH cleavage (i.e. the known synergistic effects of bimetallic catalysis24). Theoretical studies
show that Pd in PdZn alloys exhibits a notably lower density of states at the Fermi level than
ACS Paragon Plus Environment
25

ACS Catalysis
Page 26 of 32
1
2
pure Pd, and resembles that for Cu.42 Pure Cu catalysts, however, are much less reactive than Pd
in glycerol hydrogenolysis.14,26 Such comparison suggests that it is not the electronic effects
induced by Zn on the PdZn alloy surfaces to contribute to the observed promoting effects on the
hydrogenolysis turnover rates (Figs. 1 and 3). On the other hand, metallic Zn is more oxophilic
than Pd,43 and thus Zn is the preferred site to bind 2,3ꢀdihydroxypropanoxide via the interaction
between the Zn site and the terminal O atom in 2,3ꢀdihydroxypropanoxide on the PdZn alloy
surfaces, consistent with the higher measured ꢍꢎꢏꢐ values on PdZn than on Pd (Table 3). The αꢀ
H atom in the 2,3ꢀdihydroxypropanoxide species bound to a Zn site can be abstracted by a
vicinal Pd site to form glyceraldehyde (Scheme 4). This synergy between Zn and Pd sites on the
alloy surfaces would lead to stronger stability of the αꢀCꢀH cleavage transition state compared to
pure Pd surfaces. The lower ꢂG‡CꢀH values measured on the PdZn alloy surfaces than on Pd thus
indicate that the synergistic effects between the Zn and Pd sites account for the increased
glycerol hydrogenolysis turnover rates on the PdZn alloy surfaces, as also found for other
bimetallic catalysts (e.g. supported PtRe,44 IrRe,45 RhRe,46 and RhMo47) in selective
hydrogenolysis of other polyols and cyclic ethers, important reactions for upgrading of biomassꢀ
derived oxygenates into useful products.
3
4
5
6
7
8
9
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. CONCLUSIONS
Glycerol hydrogenolysis turnover rate and propylene glycol selectivity on PdZn/mꢀZrO2 are
mediated by the Zn/Pd ratio of the PdZn alloy phases, both showing maximum values at
moderate Zn/Pd ratios. These PdZn/mꢀZrO2 catalysts, however, tend to deactivate, because of the
dissolution ability of metallic Zn and its consequent loss in the form of Zn2+ ions from the PdZn
alloys in glycerol hydrogenolysis reactions. Such deactivation can be inhibited by addition of
ACS Paragon Plus Environment
26

Page 27 of 32
ACS Catalysis
1
2
Zn2+ donors (e.g. ZnO). The presence of the Zn2+ donors also leads to the in situ formation of
active PdZn alloy layers on Pd/mꢀZrO2, which renders direct use of physical mixtures of Pd/mꢀ
ZrO2 and ZnO as efficient and stable catalysts for glycerol hydrogenolysis to propylene glycol.
Rigorous kinetic assessment, together with primary kinetic isotope effects, suggests that the
cleavage of αꢀCꢀH bond in 2,3ꢀdihydroxypropanoxide intermediate to form glyceraldehyde is the
kineticallyꢀrelevant step in glycerol hydrogenolysis on the Pd and PdZn surfaces. The higher
oxophilicity of Zn than Pd makes 2,3ꢀdihydroxypropanoxide more stable on PdZn alloys than on
Pd surfaces, and thus facilitates the abstraction of αꢀH atoms synergistically by adjacent Pd sites,
leading to the observed superiority of the PdZn alloys to Pd.
3
4
5
6
7
8
9
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
ASSOCIATED CONTENT
Supporting Information
Effects of Zn/Pd ratio on rates and selectivities for PdZn/mꢀZrO2ꢀ1.0; Zn2+ concentration in
filtrate after reaction as a function of the amount of ZnO additive; promoting effects of
Zn(AcO)2ꢁ2H2O on rates and selectivities for Pd/m-ZrO2ꢀ1.0; rates and selectivities as a function
of glycerol conversion on Pd/mꢀZrO2+ZnO; and effects of glycerol and H2 concentrations on the
selectivity of glycerol hydrogenolysis on Pd/mꢀZrO2+ZnO
AUTHOR INFORMATION
Corresponding Author
*Eꢀmail: hcliu@pku.edu.cn.
Author Contributions
†Q. Sun and S. Wang contributed equally to this work.
ACS Paragon Plus Environment
27

ACS Catalysis
Page 28 of 32
1
2
3
4
Notes
5
6
The authors declare no competing financial interest.
7
8
9
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
ACKNOWLEDGMENT
This work was supported by the National Natural Science Foundation of China (Grant Nos.
21690081, 21433001, 21373019) and the National Key Research and Development Program of
China (Grant No.2016YFB0701100).
REFERENCES
(1) Gilkey, M. J.; Xu, B. ACS Catal. 2016, 6, 1420ꢀ1436.
(2) Wang, Y. L.; Zhou, J. X.; Guo, X. W. RSC Adv. 2015, 5, 74611ꢀ74628.
(3) Corma, A.; Iborra, S.; Velty, A. Chem. Rev. 2007, 107, 2411ꢀ2502.
(4) Katryniok, B.; Paul, S.; Dumeignil, F. ACS Catal. 2013, 3, 1819ꢀ1834.
(5) Nakagawa, Y.; Tomishige, K. Catal. Sci. Technol. 2011, 1, 179ꢀ190.
(6) Hejna, A.; Kosmela, P.; Formela, K.; Piszczyk, Ł.; Haponiuk, J. T. Renewable Sustainable
Energy Rev. 2016, 66, 449ꢀ475.
(7) Miyazawa, T.; Kusunoki, Y.; Kunimori, K.; Tomishige, K. J. Catal. 2006, 240, 213ꢀ221.
(8) Maris, E. P.; Davis, R. J. J. Catal. 2007, 249, 328ꢀ337.
(9) Feng, J.; Wang, J. B.; Zhou, Y. F.; Fu, H. Y.; Chen, H.; Li, X. J. Chem. Lett. 2007, 36,
1274ꢀ1275.
(10) Shinmi, Y.; Koso, S.; Kubota, T.; Nakagawa, Y.; Tomishige, K. Appl. Catal., B 2010, 94,
318ꢀ326.
ACS Paragon Plus Environment
28

Page 29 of 32
ACS Catalysis
1
2
3
4
5
6
(11) Priya, S. S.; Kumar, V. P.; Kantam, M. L.; Bhargava, S. K.; Chary, K. V. R. Catal. Lett.
2014, 144, 2129ꢀ2143.
7
8
9
(12) Musolino, M. G.; Scarpino, L. A.; Mauriello, F.; Pietropaolo, R. ChemSusChem 2011, 4,
1143ꢀ1150.
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
(13) Wang, S.; Liu, H. C. Catal. Lett. 2007, 117, 62ꢀ67.
(14) Wang, S; Zhang, Y. C.; Liu, H. C. Chem. Asian J. 2010, 5, 1100ꢀ1111.
(15) Huang, Z. W.; Cui, F.; Kang, H. X.; Chen, J.; Zhang, X. Z.; Xia, C. G. Chem. Mater. 2008,
20, 5090ꢀ5099.
(16) Huang, Z. W.; Cui, F.; Xue, J. J.; Zuo, J. L.; Chen, J.; Xia, C. G. Catal. Today 2012, 183,
42ꢀ51.
(17) Xia, S. X.; Zheng, L. P.; Wang, L. N.; Chen, P.; Hou, Z. Y. RSC Adv. 2013, 3, 16569ꢀ
16576.
(18) Zheng, L. P.; Li, X. W.; Du, W. C.; Shi, D. W.; Ning, W. S.; Lu, X. Y.; Hou, Z. Y. Appl.
Catal., B 2017, 203, 146ꢀ153.
(19) Guo, X. H.; Li, Y.; Shi, R. J.; Liu, Q. Y.; Zhan, E. S.; Shen, W. J. Appl. Catal., A 2009,
371, 108ꢀ113.
(20) Guo, X. H.; Li, Y.; Song, W.; Shen, W. J. Catal. Lett. 2011, 141, 1458ꢀ1463.
(21) van Ryneveld, E.; Mahomed, A. S.; van Heerden, P. S.; Green, M. J.; Friedrich, H. B.
Green Chem. 2011, 13, 1819ꢀ1827.
(22) Sun, D.; Yamada, Y.; Sato, S.; Ueda, W. Appl. Catal., B 2016, 193, 75ꢀ92.
(23) Nanda, M. R.; Yuan, Z. S.; Qin, W. S.; Xu, C. B. Catal. Rev.: Sci. Eng. 2016, 58, 309ꢀ336.
(24) Alonso, D. M.; Wettstein, S. G.; Dumesic, J. A. Chem. Soc. Rev. 2012, 41, 8075ꢀ8098.
ACS Paragon Plus Environment
29