Performance, Structure, and Mechanism of ReOx-Pd/CeO2 Catalyst for Simultaneous Removal of Vicinal OH Groups with H2

Article abstract of DOI:10.1021/acscatal.6b00491

The ceria-supported rhenium catalyst modified with palladium (ReO_x-Pd/CeO₂ (Re = 2 wt %, Pd/Re = 0.25)) is still the best catalyst for simultaneous hydrodeoxygenation. Higher Re loading amount decreased the activity. The simultaneous hydrodeoxygenation of cyclic vicinal diols occurs with high cis-stereoselectivity. ReO_x-Pd/CeO₂ catalysts were characterized by means of XRD, TEM, H₂-TPR, XAFS, XPS, Raman, and DFT calculations. The Re species on ReO_x-Pd/CeO₂ (Re = 2 wt %, Pd/Re = 0.25) catalyst after reduction and after stoichiometric reaction of 1,2-hexanediol to 1-hexene were Re^IV and Re^VI, and the Re^IV species were converted to Re^VI through the stoichiometric reaction. The Re species on ReO_x-Pd/CeO₂ are proposed to be randomly located on the CeO₂ surface, and probably only monomeric Re species have catalytic activity for simultaneous hydrodeoxygenation. This model can explain the higher activity of Re = 2 wt % catalyst than those of higher Re loading catalysts. The reaction is proposed to proceed by the tetra/hexavalent redox cycle of the Re center in the catalysis followed by hydrogenation.

Full text of DOI:10.1021/acscatal.6b00491

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES
Article
x
2
Performance, Structure and Mechanism of ReO–Pd/CeO
2
Catalyst for Simultaneous Removal of Vicinal OH Groups with H
Nobuhiko Ota, Masazumi Tamura, Yoshinao Nakagawa, Kazu Okumura, and Keiichi Tomishige
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.6b00491 • Publication Date (Web): 05 Apr 2016
Downloaded from http://pubs.acs.org on April 12, 2016
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 56
ACS Catalysis
1
2
3
4
5
6
7
8
9
Performance, Structure and Mechanism of ReO
of Vicinal OH Groups with H
x 2
–Pd/CeO Catalyst for Simultaneous Removal
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
[
a]
[a, c]
[a, c]
[b]
[a, c]
Nobuhiko Ota, Masazumi Tamura, Yoshinao Nakagawa,*
Kazu Okumura, Keiichi Tomishige*
[
a]
Department of Applied Chemistry, School of Engineering, Tohoku University,
6ꢀ6ꢀ07, Aoba, Aramaki, Aobaꢀku, Sendai 980ꢀ8579, Japan
[b]
Department of Applied Chemistry, Faculty of Engineering, Kogakuin University,
2
665ꢀ1 Nakanoꢀmachi, Hachioji, Tokyo 192ꢀ0015, Japan
Research Center for Rare Metal and Green Innovation, Tohoku University,
68ꢀ1, Aoba, Aramaki, Aobaꢀku, Sendai 980ꢀ0845, Japan
[
c]
4
*
Corresponding author: Keiichi Tomishige
School of Engineering, Tohoku University,
6
ꢀ6ꢀ07, Aoba, Aramaki, Aobaꢀku, Sendai, 980ꢀ8579, Japan
Eꢀmail: tomi@erec.che.tohoku.ac.jp
*
Corresponding author: Yoshinao Nakagawa
School of Engineering, Tohoku University,
ꢀ6ꢀ07, Aoba, Aramaki, Aobaꢀku, Sendai, 980ꢀ8579, Japan
Eꢀmail: yoshinao@erec.che.tohoku.ac.jp
6
ACS Paragon Plus Environment
1

ACS Catalysis
Page 2 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
Abstract
Ceriaꢀsupported rhenium catalyst modified with palladium (ReO
x
2
–Pd/CeO (Re = 2 wt%, Pd/Re
= 0.25)) is still the best catalyst for simultaneous hydrodeoxygenation. Higher Re loading amount
decreased the activity. The simultaneous hydrodeoxygenation of cyclic vicinal diols occur with high
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
cisꢀstereoselectivity. ReO
XAFS, XPS, Raman, and DFT calculations. The Re species on ReO
0.25) catalyst after reduction and after stoichiometric reaction of 1,2ꢀhexanediol to 1ꢀhexene were
x
–Pd/CeO
2
catalysts were characterized by means of XRD, TEM, H
2
ꢀTPR,
x
–Pd/CeO (Re = 2 wt%, Pd/Re =
2
IV
VI
IV
VI
Re and Re , and the Re species converted to Re through the stoichiometric reaction. The Re
species on ReO –Pd/CeO are proposed to be randomly located on CeO surface, and probably only
x
2
2
1
1
1
monomeric Re species have catalytic activity for simultaneous hydrodeoxygenation. This model can
explain the higher activity of Re = 2 wt% catalyst than those of higher Re loading catalysts. The
reaction is proposed to proceed by the tetra/hexavalent redox cycle of Re center in the catalysis
followed by hydrogenation.
13
14
15
Keywords
16
17
18
19
heterogeneous catalysis, rhenium oxide, palladium, hydrodeoxygenation, deoxydehydration.
Introduction
In recent years, because of the increasing need for energy and material sources, biomassꢀderived
feedstocks have attracted attention as alternative renewable resources for chemicals and fuels [1–7].
Based on this background, biomassꢀderived chemicals, such as sugars and sugar alcohols, are
attractive starting materials to substitute for unsustainable fossil resources. Since oxygen is rich in
these biomassꢀderived materials compared with most valueꢀadded chemicals, hydrodeoxygenation,
20
21
22
23
ACS Paragon Plus Environment
2

Page 3 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
especially partial hydrodeoxygenation, is a key reaction in the conversion of biomass to valueꢀadded
chemicals. For substrates with a few OH groups such as glycerol, a number of partial
hydrodeoxygenation systems that selectively remove one OH group have been reported [8–12].
However, it is much more difficult to selectively remove OH groups with these catalysts from
substrates with four or more OH groups such as erythritol, xylitol, and sorbitol [13, 14]. To develop
the conversion processes of these polyols, more effective and selective catalytic systems for
hydrodeoxygenation may be necessary. One of such systems is deoxydehydration (DODH; or
didehydroxylation), which converts vicinal two OH groups to one C=C double bond [15, 16].
Previously, Re [17, 18], V [19], and Mo [20–22] homogeneous and heterogeneous catalysts,
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
1
1
1
especially Re, have been reported to be active in DODH. DODH systems typically use nonꢀH
reductants such as PPh [23, 24], sulfite [25, 26], metal [27], hydroaromatics [28, 29] and alcohols
[29–37], and lower yields of the DODH products are obtained when H is used as a reductant [24, 29,
2
3
2
13
35, 38–40]. Existing data and proposals from the prior stoichiometric or catalytic DODH studies
point to three basic reactions that are likely involved in the DODH catalytic cycle: (1) twoꢀelectron
14
VII
15
reduction of high valent species such as Re , (2) the formation of diolate by the coordination of diol
16
17
18
19
with the reduced Re species, and (3) elimination of produced alkene from the lowerꢀvalentꢀRe diolate
and regeneration of the catalyst. Steps (1) and (2) may be interchangeable: (1)’ Formation of diolate
with high valent species and (2)’ twoꢀelectron reduction of Re diolate.
As a similar reaction to the DODH, we reported simultaneous hydrodeoxygenation over ReO
x
–
20
Pd/CeO catalyst (Scheme 1) [41]. In this simultaneous hydrodeoxygenation process, two vicinal OH
2
21
22
23
groups are transformed to H atoms, and it is particularly effective in the selective conversion of
polyols. This catalyst system has excellent stability: the catalyst can be reused without loss of
activity and selectivity when the very small amount of organic deposit on the used catalyst is
ACS Paragon Plus Environment
3

ACS Catalysis
Page 4 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
removed by calcination. Because the substrate, main product, and active element of simultaneous
hydrodeoxygenation are similar to those of DODH, our system was supposed to be a combination of
DODH + hydrogenation. It is characteristic that ReO
activity (TOF) than all previously reported homogeneous and heterogeneous catalysts for DODH
[41]. Therefore, even though the reaction mechanism of ReO –Pd/CeO catalyst is similar to that of
x 2
–Pd/CeO catalyst showed much higher
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
x
2
homogeneous catalysts, there can be some differences such as the structure of active sites and the
activation mechanism of the reductant.
It is well known that Re species can have various valences (0~+7), and the structure and/or
valence of Re species is much influenced by conditions. For example, Re species on reduced
1
1
1
Ir−ReO
are threeꢀdimensional ReO
metal particles covered with those ReO
which is catalytically active in hydrogenation of carboxylic acids, contains Pd , Re (HCP), Re (FCC)
x
/SiO
2
, which is catalytically active in hydrogenolysis of CꢀO bonds in ethers and alcohols,
subꢀnanometer order clusters with average Re valence of +2~3, and Ir
clusters [42]. Reduced Re–Pd/SiO (14 wt% Re, 1 wt% Pd),
x
x
2
0
0
0
13
n+
0
0
14
and Re species [43]. Re (HCP) and Re (FCC) are present as metal particles with nanometer size,
n+
15
x 2
and they are suggested to be covered with Re species. Thus, in this report, ReO –Pd/CeO catalysts
16
17
18
19
were characterized by means of various characterization methods for each state in the catalytic cycle.
On the basis of the spectroscopic analyses, investigation of reactivity of various substrates, kinetic
studies, and DFT calculations, we confirmed that the reaction proceeded via DODH + hydrogenation.
IV
VI
2
The active site was proposed to be monomeric Re /Re species stabilized on the CeO support.
20
21
22
23
Results and discussion
Screening of catalyst
First, 1,4ꢀanhydroerythritol (1,4ꢀAHERY) was used as a model substrate. In the preliminary report
ACS Paragon Plus Environment
4

Page 5 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
[41], we tested various combination of group 5–7 metal (MO
x
) + Pd and ReO + noble metal using
x
silica support, and the other combinations than ReO
x
–Pd showed lower activity and/or lower
–Pd/Support was conducted,
selectivity to tetrahydrofuran. Besides, screening of support using ReO
x
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
and ceria support dramatically increases the catalytic activity. However, the most effective catalyst
might be changed by employing ceria support. Therefore, initially, various combination of group 5–7
metal (MO
hydrogen activation and/or hydrogenation activity) (Figure S2) were tested again using ceria support.
Nevertheless, ReO was the most effective as an active species for simultaneous hydrodeoxygenation,
and other metal oxides were almost inactive (little (<3%) or no conversion was observed), similarly
to the case of silica support (Figure S1) [41]. Among additive metals for ReO /CeO , Pd is still the
best metal in view of both activity and selectivity (Figure S2). The addition of Ru, Rh, Ir and Pt to
ReO /CeO also enhanced the catalytic activity of simultaneous hydrodeoxygenation. The activity of
the ReO –Ru/CeO , ReO –Rh/CeO , ReO –Pt/CeO and ReO –Ir/CeO catalysts was almost equal,
and lower than that of ReO –Pd/CeO . ReO –Pd/CeO and ReO –Rh/CeO showed excellent
selectivity (>99%). Over ReO –Ru/CeO or ReO –Pt/CeO , a small amount of 1ꢀbutanol was
observed. The main product over ReO –Ir/CeO was dihydrofuran which is a precursor of
x x
) + Pd (Figure S1) and ReO + additive metal (selected from the viewpoint of the
x
1
1
1
x
2
x
2
13
x
2
x
2
x
2
x
2
14
x
2
x
2
x
2
15
x
2
x
2
16
17
18
19
x
2
tetrahydrofuran. The addition of base metals (Co, Ni and Cu) did not promote this reaction, and
decreased the selectivity to tetrahydrofuran significantly.
Figure 1 shows the dependence of Pd amount on the simultaneous hydrodeoxygenation of
20
1,4ꢀAHERY over ReO
Although the catalytic activity hardly depended on the amount of Pd, the conversion was maximum
at Pd/Re=0.25 and the conversion became 38 times higher than that on ReO /CeO . In the range of
Pd/Re ≥ 0.5, in addition to the slight decrease in activity, isomerization of 1,4ꢀAHERY to
x 2
–Pd/CeO . The loading amount of Re was set constant (Re = 2 wt%).
21
22
23
x
2
ACS Paragon Plus Environment
5

ACS Catalysis
Page 6 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
1,4ꢀanhydrothreitol (1,4ꢀAHTHR), which might be Pdꢀcatalyzed dehydrogenationꢀhydrogenation
process, increased gradually with increasing Pd amount.
The effect of loading amount of Re on the simultaneous hydrodeoxygenation of 1,4ꢀAHERY was
also investigated (Figure 2). The Pd/Re molar ratio was set constant (Pd/Re = 0.25). The selectivity
to THF was very high (>99%) except in the case of very large loading amount of Re (10 wt%, 97%
selectivity). The catalytic activity increased up to 2 wt% of loading amount, and then decreased
monotonically. In spite of the much higher loading amount of Re, the activity of 10 wt% Re catalyst
was even less than half of that of the best catalyst (2 wt% Re), which represents less than 1/10 of
activity per Re. The relation between Re loading amount and the activity will be discussed in later
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
1
1
1
x 2
sections based on the structure model. After all, it was indicated again that the ReO –Pd/CeO (Re =
2 wt%, Pd/Re = 0.25), which was chosen in our previous report [41], is the optimum catalyst for this
reaction.
13
14
Effect of reaction conditions for the reaction of 1,4-AHERY over ReO
Figure 3 shows solvent effects on the simultaneous hydrodeoxygenation of 1,4ꢀAHERY over
ReO –Pd/CeO . The reaction rate was much lower in water solvent. Ether and alkane solvents such
x 2
–Pd/CeO
15
16
17
18
19
x
2
as 1,4ꢀdioxane, 1,2ꢀdimethoxyethane and dodecane, were good solvents, and 1,4ꢀdioxane was the
best solvent in terms of the activity and selectivity among the tested solvents. In the case of alcohol
solvents such as 1ꢀpentanol and 3ꢀpentanol, the reaction proceeds slightly slower than the case of
ether solvents. These lower reaction rates are probably due to the competing coordination of solvent
molecules with the substrate at the active species. The difference in the activity in the solvents such
as ether, alkanes and alcohols was not large except the case of water. It should be noted that the
simultaneous hydrodeoxygenation of 1,4ꢀAHERY was catalyzed under the neat condition in high
20
21
22
23
ACS Paragon Plus Environment
6

Page 7 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
selectivity.
The effects of H
2
pressure and 1,4ꢀAHERY concentration in the simultaneous hydrodeoxygenation
x 2 2
of 1,4ꢀAHERY over ReO –Pd/CeO was investigated. The reaction order with respect to H pressure
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
around the standard conditions was estimated to be 0 (Figure 4, the detailed data are shown in Figure
S4), which suggests that the reaction steps associated with hydrogen are fast, including reduction of
the catalyst and hydrogenation. The reaction rate was saturated with increasing concentration of
1,4ꢀAHERY, and the reaction order with respect to the concentration of 1,4ꢀAHERY around 15–20
wt% in the solvent was also estimated to be 0 (Figure 5, the detailed data are shown in Figure S5).
This behavior indicates that the reaction rate can decrease with increasing the conversion. In the case
of the standard reaction conditions (1,4ꢀAHERY 1 g, 1,4ꢀdioxane 4 g, the concentration of
1,4ꢀAHERY 20 wt%), the reaction rate can decrease with increasing the conversion and decreasing
the 1,4ꢀAHERY concentration. According to the time course of 1,4ꢀAHERY reaction in the previous
report [41], the conversion was increased by ~20% at first 4 h, and because of decrease of the
reaction rate at higher conversion, the conversion reached >95% at 48 h. This behavior can be
explained to some extent by the effect of the 1,4ꢀAHERY concentration.
1
1
1
13
14
15
16
17
18
19
Because alcohols such as 3ꢀpentanol have been used as a reductant in the DODH reaction over
homogeneous Reꢀcatalysts [29–37], 3ꢀpentanol was tested instead of H
of 1,4ꢀAHERY over ReO –Pd/CeO catalyst (Table S1). Here, the reaction was performed in similar
conditions to standard ones except using 3ꢀpentanol as the reductant and Ar gas to pressurize. The
conversion rate in the reaction using 3ꢀpentanol was much lower than that in the reaction using H
2
as a reductant in the reaction
x
2
20
2
21
22
23
(2% vs. 38% conversion). These results suggest that the Re species on CeO
2 2
can be reduced with H ,
not with alcohols.
ACS Paragon Plus Environment
7

ACS Catalysis
Page 8 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
Reactivities of various substrates
The ReO –Pd/CeO catalyst was applied to various substrates, and the reactivities were compared
x
2
at low conversion level (Table 1). The reactivity of 1,4ꢀAHTHR, 3ꢀhydroxytetrahydrofuran, and THF
was much lower than that of 1,4ꢀAHERY (Table 1, entries 1–4). cisꢀ1,2ꢀCyclopentanediol and
cisꢀ1,2ꢀcyclohexanediol were converted to corresponding cycloalkane at moderate conversion and
high selectivity, however, the reactivity of transꢀ1,2ꢀcyclopentanediol and transꢀ1,2ꢀcyclohexanediol
was much lower than that of cisꢀ isomers (Table 1, entries 5, 6, 8, 9). These low reactivities of
transꢀdiols and monoꢀols were also reported in Reꢀcatalyzed DODH systems [17, 18]. These results
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
supported DODH
+
hydrogenation mechanism of ReO
x
–Pd/CeO
2
system. From
1
1
1
transꢀ1,2ꢀcyclopentanediol and transꢀ1,2ꢀcyclohexanediol, the corresponding cycloalkanes
(cyclopentane and cyclohexane) were slightly formed probably via Pdꢀcatalyzed isomerization, i.e.,
via dehydrogenation–hydrogenation process, to cisꢀisomer. In the case of transꢀ1,2ꢀcyclopentanediol,
the cisꢀisomer, which was formed via Pdꢀcatalyzed isomerization, was not detected. However, in the
case of transꢀ1,2ꢀcyclohexanediol, the cisꢀisomer was detected in high selectivity with respect to the
conversion. These results suggest the difference in the presence of the transꢀisomer on the reaction of
cisꢀisomer. For example, it is expected that adsorption of cisꢀ1,2ꢀcyclopentanediol to the Re center
was preferred for that of transꢀ1,2ꢀcyclopentanediol, but there is no clear difference in the adsorption
between cisꢀ and transꢀ1,2ꢀcyclohexanediols to the Re center. In order to discuss the effect of
coexistence of the cisꢀdiol and the transꢀdiol, the reactions of mixture of cisꢀ and transꢀdiols (c : t =
1 : 1) were conducted (Table 1, entries 7 and 10). Conversion of transꢀdiols in these mixtures was
negligible (lower than experimental error), and the cycloalkane could be regarded as the product
from cisꢀdiols. The yield of cyclopentane from mixture of cisꢀ and transꢀ1,2ꢀcyclopentanediols, was
comparable to that from pure cisꢀ1,2ꢀcyclopentanediol (Table 1, entries 5 and 7). However, In the
13
14
15
16
17
18
19
20
21
22
23
ACS Paragon Plus Environment
8

Page 9 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
case of the reaction using mixture of cisꢀ and transꢀ1,2ꢀcyclohexanediols, the reactivity of
cisꢀ1,2ꢀcyclohexanediol was decreased to less than half of that in the absence of transꢀisomer (Table
1, entries 8 and 10), and the reaction of cisꢀ1,2ꢀcyclohexanediol is suppressed by the presence of
transꢀ1,2ꢀcyclohexanediol. Considering that the dihedral angle (O–C–C–O) of the vicinal OH groups
of cisꢀ1,2ꢀcyclopentanediol (~40°; determined by DFT calculation) is smaller than that in
transꢀ1,2ꢀcyclopentanediol (~70°), transꢀ1,2ꢀcyclopentanediol probably cannot form diolate with Re
center. On the other hand, transꢀ1,2ꢀcyclohexanediol, cisꢀ1,2ꢀcyclohexanediol and 1,2ꢀalkanediols
have similar dihedral angle of the vicinal OH groups (~60°), and the presence of
transꢀ1,2ꢀcyclohexanediol can decrease the reactivity of the coexistent cisꢀ1,2ꢀcyclohexanediol by
the competitive formation of diolates. The reactivity of 1,2ꢀalkanediols such as 1,2ꢀbutanediol (Table
1, entry 11) and 1,2ꢀhexanediol (Table 1, entry 12) was almost the same as that of 1,4ꢀAHERY, and
simultaneous hydrodeoxygenation products such as nꢀbutane (from 1,2ꢀbutanediol) and nꢀhexane
(from 1,2ꢀhexanediol) were obtained in high selectivity. The reactivity of 2,3ꢀbutanediol (entry 13)
was about half as compared to that of 1,2ꢀalkanediols (Table 1, entries 11, 12). From 1,3ꢀbutanediol
substrate, 2ꢀbutanol was mainly produced and small amount of ethanol was formed (Table 1, entry
14), probably through a dehydrogenation–dehydration–hydrogenation process (Scheme 2) [8, 44]
and retroꢀaldol reaction following dehydrogenation of primary OH group [8, 44–47], respectively.
1,4ꢀButanediol was almost inert (Table 1, entry 15). Erythritol was mainly converted to
1,2ꢀbutanediol with similar reactivity to 1,2ꢀhexanediol (Table 1, entry 16). The order of the
reactivity of the substrates was cisꢀvicinal diols on fiveꢀmembered ring > 1,2ꢀalkanediol > cisꢀvicinal
diols on sixꢀmembered ring ≈ vicinal alkanediols which have two secondary OH groups. The higher
reactivity of cisꢀvicinal diols on fiveꢀmembered ring is probably due to the smaller dihedral angle of
vicinal OH groups of cisꢀvicinal diols. The order of the reactivity of substrates except cisꢀvicinal
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
1
1
1
13
14
15
16
17
18
19
20
21
22
23
ACS Paragon Plus Environment
9

ACS Catalysis
Page 10 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
diols on fiveꢀmembered ring is probably determined by the steric hindrance around the OH groups.
Characterizations and reaction mechanism
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
Because of the similarity of reactivity order of various substrates such as cis/transꢀdiols between
ReO
ReO
x
x
–Pd/CeO
–Pd/CeO
2
2
and homogeneous Reꢀcatalyzed DODH systems [16, 17], the reaction pathway over
is probably DODH + hydrogenation. As shown in the Introduction section, the
reaction mechanism of DODH with homogeneous Re catalysts is believed to be a twoꢀelectron redox
cycle of high valent Re and there are following three elementary reaction steps (1), (2), and (3): (1)
reduction of Re species, (2) diol coordination with Re species to form a diolate (steps (1) and (2)
may be interchangeable), and (3) elimination of produced alkene and regeneration of the catalyst [16,
17]. In order to investigate the reaction mechanism in detail, characterization by various techniques
1
1
1
x 2
was conducted for the ReO –Pd/CeO catalysts before and after each elementary reaction step.
13
14
XRD and TEM
15
In the previous report, XRD measurements of ReO
calcination and catalytic use were conducted, and the results indicate that both Re and Pd species on
ReO –Pd/CeO were highly dispersed and/or amorphous [41]. We further measured the XRD pattern
of ReO –Pd/CeO (Re = 2 wt%, Pd/Re = 0.25) after reduction at 423 K and those of ReO –Pd/CeO
(Re = 10 wt%, Pd/Re = 0.25) after calcination, reduction at 423 K and catalytic use (Figure S5). All
samples showed similar patterns; only peaks of CeO support were observed, and these were no
peaks for ReO or Pd species. The TEM image of ReO –Pd/CeO (Re = 2 wt%) after calcination is
shown in Figure S6. No particle due to ReO or Pd species was observed. These XRD and TEM
results indicate that both ReO and Pd species are highly dispersed on both ReO –Pd/CeO catalysts.
x 2
–Pd/CeO (Re = 2 wt%, Pd/Re = 0.25) after
16
17
18
19
x
2
x
2
x
2
20
2
21
22
23
x
x
2
x
x
x
2
ACS Paragon Plus Environment
10

Page 11 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
Effect of gasꢀphase reduction preꢀtreatment
The temperature programmed reduction (TPR) profiles of ReO
Pd/Re = 0.25) and ReO /CeO (Re = 2 wt%) after the calcination at 773 K are shown in Figure 5. All
patterns have a broad signal above 900 K with almost constant area, suggesting that this broad signal
above 900 K is not due to reduction of Re species, and is probably due to reduction of CeO support.
In the case of ReO /CeO , a sharp peak at 710 K was observed with a total H consumption in the
range of 440–770 K of 4.8 equivalents of Re in addition to a broad signal above 900 K (Figure 5c).
The H consumption value in the range of 440–770 K was significantly larger than that for the
reduction of Re (H /Re ≤ 3.5), indicating that some amount of Ce was also reduced in this
temperature range. In the case of ReO –Pd/CeO (Re = 2 wt%, Figure 5a), a signal with two sharp
x
2
–Pd/CeO (Re = 2 or 10 wt%,
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
x
2
2
x
2
2
2
1
1
1
2
x
2
peaks at 490 and 530 K was observed in addition to the broad signal above 900 K. The presence of
Pd lowered the reduction temperature of Re. The range of these peaks did not overlap with
13
14
ReO
consumption amount of low temperature signal (H
for the reduction of Re. As reported previously, Re species on ReO
partially reduced at 773 K to the valence of +4 [41]. For Re = 10 wt% catalyst, the lowꢀtemperature
signal had one peak at 550 K and the ratio of H consumption amount to Re was 2.9 (Figure 5b). The
x
/CeO
2
at all, indicating that all the Re species was affected by Pd in the reducibility. The H
/Re = 5.6; 450–560 K) was also larger than that
–Pd/CeO (Re = 2 wt%) was
2
15
2
16
17
18
19
x
2
2
reduction degree of Ce can vary in different Re amount and temperature, and thus the average
valence of Re could not be determined from TPR measurement alone. The determination of Re
valence by XANES and XPS will be discussed in later sections.
20
21
22
23
Next, the effect of reduction preꢀtreatment was investigated and the effect was compared with TPR
x 2
results. ReO –Pd/CeO (Re = 2 wt%, Pd/Re = 0.25) catalyst was reduced in the gas phase at various
ACS Paragon Plus Environment
11

ACS Catalysis
Page 12 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
temperatures and then it was applied to the activity test at 413 K (Figure 6). We have reported that
the catalyst after reduction at 773 K shows similar conversion and selectivity to those of fresh
catalyst; however the reduced catalyst was exposed to air after reduction in that case [41]. In this
study, the reduced catalysts were loaded to the reactor without exposure to air. The catalytic activity
was dramatically decreased by reducing at higher temperature. In the cases of reduction at 673 K and
773 K, which temperatures were higher than the TPR peaks (Figure 5a), the reaction hardly
proceeded. Even in the case of reduction at 423 K, which temperature was in the left shoulder of the
first TPR peak (Figure 5a), the activity was significantly lower than that of nonꢀreduced catalyst.
Considering that reduction of Re is involved in the catalytic cycle of DODH, this decrease 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
IV
III
1
1
1
catalytic activity is probably caused by reduction of Ce to Ce . Exposure to air can rapidly oxidize
III
IV
the Ce to Ce , and the catalytic activity was restored to similar level of fresh catalyst as described
in the previous report [41].
13
14
Stoichiometric reactions using gasꢀphase reduced catalysts
15
In order to understand the mechanism of this process, stoichiometric reactions of 1,2ꢀhexanediol
using gasꢀphase preꢀreduced catalysts were performed in a test tube under an Ar atmosphere (Table
2). In these tests, to suppress reduction of Re species by hydrogen species via dehydrogenation of
16
17
18
19
x 2
substrate, the reaction temperature was set at 353 K. ReO –Pd/CeO (Re = 0.5, 2, 4, and 10 wt%,
Pd/Re = 0.25) catalysts were reduced in gasꢀphase at 423 K in order to sufficiently reduce Re species
IV
20
with minimum reduction of Ce . In fact, the catalytic reaction with the catalyst without
21
22
23
preꢀreduction hardly proceeded at too low temperature (Table S2) such as 393 K, probably because
reduction of Re did not proceed as shown in the TPR result (Figure 6). 1ꢀHexene, the DODH product,
x 2
was formed by the reaction with reduced ReO –Pd/CeO . In the case of Re = 2 wt% catalyst,
ACS Paragon Plus Environment
12

Page 13 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
turnover number (TON) per Re atom achieved about 0.4 at 2 h (Table 2, entry 2) and it was almost
constant even at 4 h (Table 2, entry 3). It is suggested that about 40% of total Re is catalytically
active. Thus, the active Re species can be characterized using various spectroscopic methods using
this 423 Kꢀreduced catalyst. On the other hand, TON per Re atom was decreased with increasing Re
loading amount. In the case of Re = 10 wt%, TON per Re atom achieved only 0.02 at 2 h (Table 2,
entry 5), and therefore the amount of the active Re species was very small in Re = 10 wt% catalyst.
This dependence of these stoichiometric reactions on the Re loading amount is consistent with the
results of the catalytic use under the standard conditions (Figure 2). The amount ratio of catalytically
active Re to total Re will be discussed below based on further characterizations and the structure
model.
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
1
1
1
XPS and Re L
3
ꢀedge XANES analyses
2
–Pd/CeO (before and after stoichiometric reaction, and after catalytic
13
The Re 4f XPS data of ReO
x
14
use) are shown in Figure 7 and Table 3. The peak positions of the functions with lower binding
IV
15
energy were within the range of Re (42.3 and 43.6 eV [48–54]), and those with higher binding
VI
16
17
18
19
energy were within the range of Re (44.3 and 44.9 eV [50–52, 55, 56]). Therefore the samples
before and after stoichiometric reaction and after catalytic use under the standard conditions
IV
VI
contained both Re and Re species. The Pd 3d XPS data of the same samples and summary of
XPS data are shown in Figure S7 and Table 3, and these results indicated that all samples contained
0
20
only Pd . Both Re and Pd species were highly dispersed because Pd/Re molar ratio calculated from
21
22
23
XPS was nearly equal to the loaded Pd/Re ratio. The absence of XRD peaks for Pd and Re species
also supports the high dispersion of both species. Thus, all the Re and Pd species in the ReO
x
–
Pd/CeO catalyst can be detected by XPS measurement.
2
ACS Paragon Plus Environment
13

ACS Catalysis
Page 14 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
Average valences of Re calculated from XPS on ReO
x
–Pd/CeO
2
(Re = 2 wt%, Pd/Re = 0.25)
IV
before and after stoichiometric reaction were +4.2 and +5.3, respectively (Table 3). The ratio of Re
IV
VI
to (Re + Re ) was determined to be 91% and 37% for the sample of before and after
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
IV
stoichiometric reaction, respectively (Table 3). The decreasing amount of Re and increasing
VI
amount of Re species through the stoichiometric reaction fairly agreed with that about half of the
Re species reacted with 1,2ꢀdiol (TON = 0.42, Table 2, entry 3). The reaction mechanism of this
reaction is suggested to be the tetra/hexavalent redox cycle of Re center. The XPS result after
catalytic use was similar to that after stoichiometric reaction. The stoichiometric reaction can
proceed at lower temperature (353 K) than the reduction of Re (≥393 K), and therefore the active
reduced Re species after the catalytic use can react with the remaining substrate during cooling. In
1
1
1
2 x
addition, in order to investigate the effect of CeO support, Re 4f XPS measurement of ReO –
Pd/SiO after catalytic use under the standard conditions was conducted, and comparison was made
2
13
x 2
between ceria support and silica support (Figure S8). Some Re species on ReO –Pd/SiO after
0
14
catalytic use was Re , and the valence of Re of ReO
x
–Pd/SiO
support suppressed the reduction of Re species.
ꢀXANES spectra of ReO –Pd/CeO and reference compounds. The first
ꢀedge XANES is called as a white line, and the white line area in the L ꢀedge XANES
2
from XPS was lower than that of
15
ReO
Figure 8 shows the Re L
peak in the L
x 2 2
–Pd/CeO . The results indicated that CeO
16
17
18
19
3
x
2
3
3
is known to be an informative indication of the electronic state. The larger white line area is due to
greater electron vacancy in dꢀorbital. As reported previously, a relative electron deficiency and
valence can be determined on the basis of the white line area [57–60]. Regarding the reference
compounds, valence of Re species is almost proportional to the white line area. Therefore, the
average value of Re species can be estimated by examining the white line area in the XANES spectra
20
21
22
23
x 2
[61, 62]. The valence of Re species on ReO –Pd/CeO (Re = 2 wt%, Pd/Re = 0.25) can be
ACS Paragon Plus Environment
14

Page 15 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
determined by the relation between the valence and the white line area of the reference compounds
x 2
(Figure S9). The average valence of the Re species on ReO –Pd/CeO before and after stoichiometric
reaction was determined to be +4.2 and +5.3, respectively (Table 3), which was almost the same as
those calculated from XPS. The consistent results support tetra/hexavalent redox cycle of Re center
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
x 2
in the catalysis. In the case of ReO –Pd/CeO after catalytic use, the average valence was determined
to be +5.0, which was also consistent with XPS.
Re L
In order to characterize the structure of Re species precisely and the interaction of Re species with
Pd metal clusters, Re L ꢀedge and Pd Kꢀedge EXAFS analyses were carried out. Figure 9 shows the
Re L ꢀedge EXAFS spectra of ReO –Pd/CeO (Re = 2 wt%, Pd/Re = 0.25) after calcination, before
and after stoichiometric reaction, and after catalytic use under the standard conditions. Table 4
3
ꢀedge EXAFS analyses
1
1
1
3
3
x
2
13
summarizes the curve fitting results of Re L
3
ꢀedge EXAFS. In the case of the calcined ReO
x
–
3
14
Pd/CeO
2
, the Fourier transform of k ꢀweighted spectrum was similar to that of NH
4
ReO (Figures 9d
4
15
and e), and the curve fitting results indicate the presence of Re–O bond with coordination number
–
16
17
18
19
(CN) of 4.0±0.3 (Table 4). It is suggested that the Re species were present as ReO
ReO –Pd/CeO
4
. The spectrum of
x
2
was drastically changed by the reduction and the reaction. The curve fitting analysis
indicates the presence of Re=O, Re–O and Re–Pd bonds. The absence of Re–Re bond indicates that
almost all the Re species were in oxidized state, and this is also supported by XPS (Figure 7).
20
The length of Re–O bond (0.193–0.194 nm) on ReO
reaction and after the catalytic use was clearly longer than that (0.173 nm) on NH
Pd/CeO after the calcination, and the longer Re–O bond can be assigned to a single bond. Although
the coordination number was low (0.4), the presence of Re–Pd bond demonstrates the direct
x
–Pd/CeO
2
before and after stoichiometric
21
22
23
4
ReO and ReO –
4
x
2
ACS Paragon Plus Environment
15

ACS Catalysis
Page 16 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
interaction between Re species and Pd metal clusters. The bond length of the Re–Pd bond (0.266–
0.269 nm) on the catalyst after the reduction and the reaction corresponded with the direct metal
bond (cf. Re–Re in Re powder: 0.275 nm, Pd–Pd in Pd foil: 0.275 nm). Assuming that Re atoms
were adsorbed at threeꢀfold hollow site of Pd (1 1 1) surface [42, 63], the coordination number (CN)
of Re–Pd should be 3 per one Re atom on Pd (1 1 1). In this case, the obtained CN (0.4) suggests that
about 10% of Re species interacted with Pd metal surface and about 90% of Re did not interact with
Pd metal. If the Pdꢀbound Re atom is incorporated in the Pd particles rather than located on the
surface of Pd, the CN of Re–Pd per one Re should be higher than 3, and the amount of Pdꢀbound Re
species will be further smaller. The small amount of Pdꢀbound Re species can be connected to small
dependency of activity on the Pd amount (Figure 1), indicating that the main active site is not the
small amount of Pdꢀbound Re species. Rather, the small decrease in the activity with increasing the
Pd amount above Pd/Re = 0.25 suggests that the Pdꢀbound Re species is inactive. The CN of the Re–
O single bond was increased by the stoichiometric reaction and the catalytic use under the standard
conditions (3.0→3.5 and 4.2, respectively), while the CN of the Re=O double bond and Re–Pd bond
are maintained during the reactions.
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
1
1
1
13
14
15
16
17
18
19
x 2
The results of Pd Kꢀedge EXAFS analysis of ReO –Pd/CeO (Re = 2 wt%, Pd/Re = 0.25) after
the catalytic use are shown in Figure S10 and Table S3, and the Pd–Pd bond (CN = 3.1) and Pd–Re
bond (CN = 1.6) were detected. It was also verified from Pd Kꢀedge EXAFS that Pd species was
x
present as highly dispersed clusters from the small CN, and some ReO species were attached to Pd
20
clusters.
21
22
23
In our previous reports, the valence of Re can be estimated approximately from the CN of Re=O
and Re−O (valence of Re = 2 × CNRe=O + 1 × CNRe−O) by assuming that all bonds of Re and O
contribute to valence of Re [42, 43, 64]. However, the calculated values of the Re valence from
ACS Paragon Plus Environment
16

Page 17 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
EXAFS analyses were clearly higher (mono–divalent) than those from Re L
ꢀedge XANES and Re
3
x 2
4f XPS (Table 3 and Figure S9). In fact, in the case of ReO –Pd/CeO (Re = 2 wt%, Pd/Re = 0.25)
before stoichiometric reaction, the valence of Re estimated approximately from the CN of Re=O and
Re−O was +6.0 (about divalent higher than that from XANES and XPS (Table 3)). This difference
suggests that some bonds existed which did not contribute to the valence of Re, such as Re←O
coordination bonds.
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
Density functional theory (DFT) calculations and proposed structure model
As discussed in the previous sections, the catalyst contains two types of Re sites in addition to
small amount of Pdꢀbound Re species: catalytically active ones and inactive ones. Further analysis of
1
1
1
Re L
DFT calculations for the determination of Re species on CeO
assumed to be occupied with the (1 1 1) surface [65–68]. The structures of monomeric Re species
3
ꢀedge EXAFS requires additional information for the structure of each Re species. We used
2
surface. Most of CeO surface is
2
13
14
were calculated with or without coordinated 1,4ꢀAHERY molecule using Ce70
O140 model cluster. The
IV
15
structure of the model cluster is shown in Figure S11. The Re O
2
species on CeO
2
(1 1 1) surface
16
17
18
19
have two Re=O bonds and two Re←O coordination bonds (Figure 10a). The O→Re←O plane leans
out of the perpendicular, and the Re atom was almost located above an O atom in the 2nd layer. The
VI
Re O
3
species on CeO
2
surface have three Re=O bonds and one Re←O bond (Figure 10b). The
IV
Re←O bond also slightly leans out of the perpendicular, toward an O atom in the 2nd layer. The Re
VI
20
and Re species after coordination of 1,4ꢀAHERY have two Re–OC single bonds, less number of
21
22
23
Re=O bonds by one than that before the coordination, and the same number of Re←O bonds to that
before the coordination (Figures 10c and 10d). The structure of transition state (TS) of DODH is
shown in Figure 10e, and the energy diagram is shown in Figure S12. The two C–ORe bonds are
ACS Paragon Plus Environment
17

ACS Catalysis
Page 18 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
dissociated simultaneously, and the Re–OC single bonds are converted to double bonds, similarly to
the reported transition state for DODH [16, 17]. An important point is that only one Re←O bond is
VI
present in this calculated TS structure, and the coordination mode is more similar to Re O
3
/CeO
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
IV
than to Re O
2
2
/CeO . The difference in the Reꢀsupport bonds between initial and transition states
suggests that polymeric Re species with many rigid Re–O bonds have lower activity.
Not all Re species can be monomeric ones because of the limited surface area of CeO
2
. Here we
pay attention to the surface density of Re species. Figure 11 shows CeO
2
(1 1 1) surface, and the
IV
circles of dotted line in Figure 11 represent sites that can be occupied with Re species (Re site), i.e.
2
ꢀ1
2
the sites above the O atoms in 2nd layer. Considering the surface area of CeO (87 m g ), the total
1
1
1
number of possible Re sites (circles of dotted line) corresponded to about 20 wt% loading of Re. In
the hexagon in Figure 11, three Re atoms can be present at a maximum. To produce monomeric Re
species, only one Re atom can be present in the hexagon. If adjacent Re sites are occupied, i.e., more
than one Re atoms are present in the hexagon in Figure 11, Re–O–Re bonds are formed. Now, we
13
14
x 2
consider two structure models of ReO –Pd/CeO , and discuss the correlation between the amount of
15
isolated Re species and the results of activity tests in both reactions under the standard conditions
and stoichiometric reactions (Figure 12).
16
17
18
19
In the model (I), it is assumed that the Re species are located monomerically as much as possible,
i.e., one Re species is located at the center of the hexagon in Figure 11 and no other Re species are
present in the surrounding six sites (repulsivelyꢀdispersed model). In this case, amount of isolated Re
species should increase linearly up to about 7 wt% of loading amount, where all the surface is filled
with hexagon with one Re atom (Figure 12, solid line). In this repulsivelyꢀdispersed model, all the
Re sites are equivalent and active when Re loading amount is lower than 7 wt%. However, the
activity tests revealed that the number of active sites did not linearly increase and rather decreased at
20
21
22
23
ACS Paragon Plus Environment
18

Page 19 of 56
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
1
2
3
4
5
6
7
Re > 2 wt%.
In the model (II), it is assumed that Re species are randomly located among all the possible sites in
Figure 12 (randomlyꢀlocated model). Here, the amount of isolated Re species is calculated based on
the following equation (1). The part in parentheses in the equation (1) represents the probability of
absence of Re species in one Re site, and the six power of the part in parentheses represents the
number of Re sites around one certain Re site.
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

Re amount
[
mol
]

8
Amount of isolated Re species
[
mol
]
= 1−
 × Re amount
[
mol
]


Total amount of Re sites
[mol
]


9
0
1
2
(1)
1
1
1
In this case, amount of isolated Re species is maximized at about 2 wt% of Re, and decrease
monotonically after that (Figure 12, broken curve). The curve well agrees with the curves of the
catalytic activities vs. Re loading amount. This agreement suggests that the model (II) is plausible
and only isolated Re species is catalytically active. When loading amount of Re is 2 wt%, about 50%
of total Re amount are isolated and will have catalytic activity. Figure 13 shows the detailed structure
13
14
15
16
17
x 2
model of Re species in ReO –Pd/CeO (Re = 2 wt%, Pd/Re = 0.25) after reduction (before
stoichiometric reaction) and after stoichiometric reaction, based on randomlyꢀlocated model and
DFT calculations. The amount of Pdꢀbound Re species (~10%) was neglected in this model. The
18
IV
19
DFT calculation (Figure 10) showed that the isolated Re species have two Re←O and two Re=O
20
bonds in the absence of substrate, giving CN of 2 and 2 for Re–O and Re=O shells, respectively, in
VI
2
1
2
the EXAFS analysis. After reaction, the presence of substrate produces substrateꢀcoordinated Re
2
species which have one Re←O, two Re–O and two Re=O bonds, giving CN of 3 and 2 for Re–O and
ACS Paragon Plus Environment
19

ACS Catalysis
Page 20 of 56
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
IV
1
2
3
4
5
6
7
8
9
0
1
2
Re=O shells, respectively. According to XPS results, the inactive Re species have Re valent and the
valence is not changed during the DODH. As discussed above, the inactive Re species are probably
IV
polymeric ones. Considering the coordination mode of substrateꢀbound Re species, we assume the
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
IV
coordination mode of polymeric Re species as two Re←O, two Re–O, and one Re=O bonds, which
give CN of 4 and 1 for Re–O and Re=O shells, respectively. The 1 : 1 molar ratio of isolated species
and polymeric species gives the overall CN of 3 and 1.5 for Re–O and Re=O shells, respectively, for
x 2
reduced ReO –Pd/CeO (Re = 2 wt%). For the catalyst after stoichiometric reaction, isolated species
and polymeric species gives the overall CN of 3.5 and 1.5 for Re–O and Re=O shells, respectively.
These coordination numbers based on this randomlyꢀlocated model (II) and DFT calculations well
1
1
1
3
agreed with the results of Re L ꢀedge EXAFS analyses (Table 4).
In order to detect the formation of polymeric Re species in the catalysts with high Re loading, we
carried out Raman measurements. Unfortunately, the Raman spectra could not be measured
successfully for reduced and used catalysts because of the gray color of the catalysts. The Raman
13
ꢀ1
14
spectra of calcined catalysts are shown in Figure S13. The Re=O bond (960–1000 cm [69–71]) was
15
observed for all samples containing Re. For ReO
x
–Pd/CeO
2
(Re = 4, 10 wt%) samples, broad band
ꢀ1
ꢀ1
16
17
18
19
around 900 cm was observed with much larger intensity than the 920 cm band of Pd/CeO
2
. The
intense band can be assigned to Re–O–Re [69–71]. The presence of Re–O–Re in ReO
= 4 wt%) sample also supports the randomlyꢀlocated model.
x
–Pd/CeO (Re
2
20
Reaction mechanism
21
22
23
The reaction mechanism proposed by this study is summarized in Scheme 3. As previously
indicated, the reaction mechanism of simultaneous hydrodeoxygenation can be regarded as DODH +
hydrogenation. This reaction mechanism has following elementary reaction steps (1), (2), (3), and
ACS Paragon Plus Environment
20

Page 21 of 56
ACS Catalysis
1
2
3
VI
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
1
2
3
4
5
6
7
8
9
0
1
2
(4): (1) the isolated Re species is reduced by H
as a reductant with promotion by Pd, and becomes
2
IV
IV
the active Re species; (2) the substrate with vicinal OH groups is coordinated to the reduced Re
species with the two OH groups as diolate; (3) alkene is eliminated with the regeneration of Re
VI
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
center (IV→VI); (4) Pdꢀcatalyzed hydrogenation of the produced alkene gives the simultaneous
IV
hydrodeoxygenation product. Additionally, the substrate is coordinated to not only the reduced Re
VI
species but also the oxidized Re species even at room temperature. As discussed in the previous
paper, the role of CeO
interaction between Re species and the CeO
role of CeO support is to stabilize the monomeric Re species. According to the literature, polymeric
2
support is probably to prevent the overreduction of Re species by the
2
surface [41]. In addition to this role, another important
2
1
1
1
Re species are easily formed over supported noble metal–Re catalysts by the reduction [42, 72].
V
Even in the homogeneous MeReO
3
catalyst for DODH, formation of oligomeric Re ꢀglycolate
species has been reported by the reduction [26]. The reasons for the much higher activity of ReO
x
–
13
Pd/CeO than other DODH catalyst can include the high rate of reduction of Re species assisted by
2
14
Pd and large number of isolated Re sites, which are probably the active sites.
15
16
17
18
19
Conclusions
ReO
hydrodeoxygenation. The simultaneous hydrodeoxygenation over ReO
for cisꢀ or synꢀelimination, and 1,2ꢀelimination rather than 2,3ꢀelimination. This regioꢀ and
stereoꢀselectivity are similar to DODH systems such as CH ReO ꢀcatalyzed one. Reduction of CeO
–Pd/CeO (Re = 2 wt%, Pd/Re=0.25) catalyst
x
–Pd/CeO
2
(Re = 2 wt%, Pd/Re = 0.25) is still the best catalyst for simultaneous
x
–Pd/CeO catalyst is favored
2
20
3
3
2
21
22
23
support decreased the activity. The Re species on ReO
x
2
IV
after reduction was mainly Re , and that after stoichiometric reaction of 1,2ꢀhexanediol to 1ꢀhexene
IV
VI
IV
VI
were Re and Re . The amount of converted Re to Re corresponded with that of produced
ACS Paragon Plus Environment
21

ACS Catalysis
Page 22 of 56
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
IV
1
2
3
4
5
6
7
8
9
0
1
2
1ꢀhexene. There are different types of Re ; active ones and inactive ones. The structure model in
which Re species are randomly located on CeO (1 1 1) surface is proposed, and only isolated Re
species are suggested to be catalytically active. The roles of Pd and CeO support are to promote the
2
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
IV
IV
reduction to active Re centers and to stabilize the isolated Re species under pressurized H
2
,
respectively
.
The reaction is proposed to proceed by the reduction of the coordinated substrate by the
IV
Re center followed by hydrogenation.
Experimental section
General
1
1
1
Suppliers of the reagents are listed in Supporting Information (Table S4).
Catalyst preparation
1
1
1
The M O
with an aqueous solution of NH
(NH )NbO(C O, (NH
x
/CeO
2
(M = V, Cr, Mn, Nb, Mo, W, Re) catalysts were prepared by impregnating M O
VO , Cr(NO ·9H O, Mn(NO ·6H O, (NH Mo
O or NH ReO . CeO
x
13
4
3
3
)
3
2
3
)
2
2
4
)
6
7 2
O24·4H O,
14
4
2
O
4
)
2
·
x
H
2
4
)
6
H
2
W
12
O
40·nH
2
4
4
2
(Daiichi Kigenso Co., Ltd., HS,
2
ꢀ1
1
15
BET surface area: 84 m g , after calcination at 873 K for 3 h) was used. The M O
x
–Pd/CeO
2
1
16
17
18
19
catalysts were prepared by impregnating M O
solution of Pd(NO . The ReO –Pd/SiO catalyst was prepared in a similar way to the case of ReO
Pd/CeO except using SiO (Fuji Silysia Chemical Ltd., Gꢀ6, BET surface area 535 m /g) as a
support.
The ReO
ReO /CeO
Cu(NO ·3H
amount of M was 2 wt%, and M /M molar ratio was 0.25.
x 2
/CeO after drying at 383 K for 12 h with aqueous
3
)
2
x
2
x
–
2
2
2
2
2
20
x
–M /CeO
after drying at 383 K for 12 h with aqueous solution of Co(NO
O, Ru(NO)(NO , RhCl ·3H O, H IrCl or H PtCl
2
(M =Co, Ni, Cu, Ru, Rh, Ir, Pt) catalysts were prepared by impregnating
·6H O, Ni(NO ·6H O,
·6H O. The loading
21
22
23
x
2
3
)
2
2
3
)
2
2
3
)
2
2
3
)
3ꢀx(OH)
x
3
2
2
6
2
6
2
1
2
1
ACS Paragon Plus Environment
22

Page 23 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
These were calcined in air at 773 K for 3 h. The standard loading amount of Re and Pd was 2 and
0.3 wt%, respectively (Pd/Re molar ratio was 0.25). All catalysts were used in powdery form.
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
Activity test
Activity tests were performed in a 190 mL stainless steel autoclave. 1,4ꢀAnhydroerythritol
(1,4ꢀAHERY) and 1,4ꢀdioxane was placed into the autoclaves with a spinner and appropriate amount
of catalyst. After sealing the reactors, their air content was purged by flushing thrice with hydrogen.
Autoclaves were then heated to the reaction temperature. The temperature was monitored using a
thermocouple inserted in the autoclave. During the experiment, the stirring rate was fixed at 250 rpm
(magnetic stirring). After an appropriate reaction time, the reactors were cooled down. The gases
were collected in a gas bag. The autoclave contents were transformed to vials, and the catalysts were
separated by centrifugation and filtration. The standard conditions for the reaction of
1,4ꢀanhydroerythritol were as follows: 1 g 1,4ꢀanhydroerythritol, 4 g 1,4ꢀdioxane, 413 K reaction
temperature, 8.0 MPa initial hydrogen pressure (at reaction temperature), 4 h reaction time, and 150
mg catalyst. The parameters were changed appropriately in order to investigate the effect of reaction
conditions. Details of the reaction conditions are described in each result. The hydrodeoxygenation
of 1,4ꢀanhydrothreitol (1,4ꢀAHTHR), cisꢀ1,2ꢀcyclopentanediol, transꢀ1,2ꢀcyclopentanediol,
cisꢀ1,2ꢀcyclohexanediol, transꢀ1,2ꢀcyclohexanediol, 2,3ꢀbutanediol, 3ꢀhydroxytetrahydrofuran,
tetrahydrofuran, 1,2ꢀbutanediol, 1,4ꢀbutanediol, 1,3ꢀbutanediol, 1,2ꢀhexanediol and erythritol were
tested in a similar way to the case of 1,4ꢀAHERY. In addition, the reaction tests using
1,2ꢀdimethoxyethane, dodecane, 1ꢀpentanol, 3ꢀpentanol, and water as a solvent were also carried out
in a similar way to the case of 1,4ꢀdioxane. Appropriate reaction conditions of these substrates are
described in each result.
1
1
1
13
14
15
16
17
18
19
20
21
22
23
ACS Paragon Plus Environment
23

ACS Catalysis
Page 24 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
The experimental method of the activity tests of the reduced catalysts is described below. The
2
calcined catalysts were reduced in H flow (30 mL/min) at appropriate temperature (423–773 K) for
1 h (at a heating rate 10 K/min) by using a glass tube. The reduced catalysts were introduced to an
autoclave under high grade Ar atmosphere in order to avoid exposing the catalyst to air. For
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
comparison, we also evaluated the catalyst passivated after the reduction in H
2
flow [31]. The
(10
catalyst was reduced at 773 K for 1 h in H flow and then passivated 2% O diluted with N
2
2
2
ml/min) for 15 min before the catalyst transfer. The reaction conditions for the reaction using
reduced catalysts were as follows: 1 g 1,4ꢀanhydroerythritol, 4 g 1,4ꢀdioxane, 413 K reaction
temperature, 8.0 MPa initial hydrogen pressure (at reaction temperature), 4 h reaction time, and 150
mg catalyst.
1
1
1
The stoichiometric reactions were performed in a 20 mL glass test tube, and were carried out in
the following procedures (1), (2), (3) and (4). (1) 1,2ꢀHexanediol and 1,4ꢀdioxane was placed in the
test tube. The mixture was frozen using acetone sherbet and degasified (freezeꢀpumpꢀthaw cycles) to
13
14
2
remove dissolved oxygen. (2) The catalysts were externally reduced with H (30 cc/min) at 423 K for
15
1 h (at a heating rate 10 K/min) by using a glass tube. (3) The reduced catalyst and a spinner were
added to the test tube containing the degasified solution in a gas bag filled with high grade Ar, and
the test tubes were sealed using a rubber stopper in the gas bag. (4) The test tubes were then heated to
the reaction temperature (353 K). The temperature was monitored using a thermocouple. During the
experiment, the stirring rate was fixed at 250 rpm (magnetic stirring).
16
17
18
19
20
21
22
23
Analysis of products
The products in both gas and liquid phases were analyzed using a gas chromatograph (Shimadzu
GCꢀ2025) equipped with flame ionization detector (FID). A TCꢀWAX capillary column (diameter
ACS Paragon Plus Environment
24

Page 25 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
0.25 mm
φ
, 30 m) was used for the separation. Products were also identified using GCꢀMS (QP5050,
Shimadzu). In the case of the reaction of erythritol, in addition to the analysis using FIDꢀGC, HPLC
(Shimadzu LCꢀ10A) equipped with a refractive index detector was used for quantitative analysis of
erythritol. An Aminex HPXꢀ87C column (diameter 7.8 mm, 300 mm) were used HPLC. The
conversion and the selectivity were calculated on the molarꢀbasis. The mass balance was also
confirmed in each result and the difference in mass balance was always in the range of the
experimental error. The agreement in terms of the mass balance indicated that polymeric byꢀproducts
were not formed (±10%).
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
1
1
1
Catalyst characterization
The surface areas of the catalysts were measured using BET method (N
apparatus (Micromeritics Gemini). Xꢀray diffraction (XRD) patterns were recorded by a
diffractometer (Rigaku MiniFlex600). Cu K radiation was used as an Xꢀray source. The samples
2
adsorption) with an
13
α
14
after the reaction or reduction were transferred to an atmosphere separator (Air sensitive sample
holder; Rigaku Corporation) using a gas bag filled with high grade Ar to avoid exposure to air.
Transmission electron microscope (TEM) images were measured by Hitachi HFꢀ2000EDX. The
samples were dispersed in ethanol with supersonic waves and placed on Mo or Cu grids under air
15
16
17
18
19
atmosphere. Temperatureꢀprogrammed reduction (H
2
ꢀTPR) was carried out in a fixedꢀbed reactor
diluted with Ar (30 ml/min). The
equipped with a thermal conductivity detector (TCD) using 5% H
2
20
amount of catalyst was 50 mg, and temperature was increased from room temperature to 1273 K at a
heating rate of 10 K/min. Raman spectra were obtained with JASCO NRSꢀ5100 spectrometer. The
powder sample was sandwiched between a slide glass and a cover glass under air atmosphere. The
excitation was conducted using a 532 nm laser and ×20 objective lens. The spectra were normalized
21
22
23
ACS Paragon Plus Environment
25

ACS Catalysis
Page 26 of 56
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
ꢀ
1
1
2
3
4
5
6
7
8
9
0
1
2
2
with the CeO ꢀderived 464 cm peak.
The extended Xꢀray absorption fine structure (EXAFS and XANES) spectra were measured at the
BL01B1 station at SPringꢀ8 with the approval of the Japan Synchrotron Radiation Research Institute
(JASRI; Proposal No. 2014A1119, No. 2014B1248). The storage ring was operated at 8 GeV, and a
Si (1 1 1) single crystal were used to obtain a monochromatic Xꢀray beam. The detailed conditions
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
for measurement of (i) reference compounds (Re powder, ReO
(ii) catalyst samples for Re ꢀedge measurements, and (iii) catalyst samples for Pd
measurements were as follows (i), (ii), and (iii). (i) Ion chambers for were filled with 85% N
15% Ar and 100% Ar, respectively, for Re ꢀedge and Pd ꢀedge measurement. Ion chambers for
were filled with 50% N + 50% Ar and 70% Ar + 30% Kr, respectively, for Re
ꢀedge measurement. The EXAFS and XANES data were collected in a transmission mode. The
2
, ReO
3
, NH
4
ReO
4
, Pd foil, and PdO),
L
3
K
ꢀedge
I
0
2
+
L
3
K
I
1
1
1
2
3
L ꢀedge and Pd
K
edge jumps were in the range of 0.6–1.5. (ii) We prepared the sample after calcination and after the
reduction by pressing catalyst powder to disk. In the case of sample after the reduction, the sample
13
14
2
disk was transferred to the inꢀsitu cell and the sample was heated to 423 K in H flow (30 mL/min)
15
under atmospheric pressure. The samples after the reaction were prepared by transferring the used
catalyst powder to a measurement cell after reaction and cooling in a gas bag filled with high grade
Ar to avoid exposure to air. The EXAFS and XANES data were collected in a fluorescence mode
using a 19ꢀelement Ge solidꢀstate detector (19ꢀSSD), and in the case of sample after calcination a
16
17
18
19
Lytle detector was used. Two ion chambers for
Kr, respectively, for Re ꢀedge measurement. The edge jumps were in the range of 0.0017–0.013,
and in the case of sample after calcination the edge jump was about 0.07. (iii) The same sample for
Re ꢀedge measurement was used. Two ion chambers for and were filled with 100% Ar and
100% Kr, respectively for Pd ꢀedge measurement. The EXAFS and XANES data were collected in
0 2
I and I were filled with 85% N + 15% Ar and 100%
20
L
3
21
22
23
L
3
I
0
I
K
ACS Paragon Plus Environment
26

Page 27 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
a fluorescence mode using a Lytle detector. The edge jump was about 0.3.
For EXAFS analysis, the oscillation was first extracted from the EXAFS data using a spline
3
smoothing method [73]. Fourier transformation of the
space to the
k
ꢀweighted EXAFS oscillation from the
k
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
r
space was performed to obtain a radial distribution function. The inversely Fourier
filtered data were analyzed using a usual curve fitting method [74, 75]. For curve fitting analysis, the
empirical phase shift and amplitude functions for the Re–O, Re–Re and Pd–Pd bonds were extracted
4 4
from data for NH ReO , Re metal and Pd metal, respectively. Theoretical functions for the Pd–Re
bond were calculated using the FEFF8.2 program [76]. Analyses of EXAFS data were performed
using a computer program REX2000, ver. 2.6 (Rigaku Corp.). Error bars for each parameter were
estimated by stepping each parameter, while optimizing the others parameter, until the residual factor
becomes two times as its minimum value [77].
1
1
1
In the analysis of XANES spectra, the normalized spectra were obtained by subtracting the
preꢀedge background from the raw data with a modified Victoreen equation and normalizing them by
the edge height [57–60].
13
14
15
Xꢀray photoelectron spectra (XPS) were measured using an AXISꢀULTRA DLD (Shimadzu Co.,
16
17
18
19
Ltd.) using a monochromatic Al Kα Xꢀray radiation (hν = 1486.6 eV) operated at 20 mA and 15 kV
at room temperature under 10−8 Pa. The spectra in energy were referenced to the C 1s peak at 284.6
eV binding energy (BE). The catalysts after the reaction or reduction were transported to the
measurement chamber in high grade Ar atmosphere to avoid any exposure to air. Peak fitting for the
Re 4f and Pd 3d spectra were done using the computer program CasaXPS, ver. 2.3.15 (Casa software
Ltd.) employing a peaks with a 30% Lorentzian and 70% Gaussian peak share. Re 4f7/2 and Re 4f5/2
are separated by 2.43 eV and have relative peak area with a fixed ratio of 4:3 and the same full width
half maximum values (FWHM). Pd 3d5/2 and Pd 3d3/2 are separated by 5.26 eV and have relative
20
21
22
23
ACS Paragon Plus Environment
27

ACS Catalysis
Page 28 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
peak area with a fixed ratio of 4:3 and the same FWHM.
DFT (density functional theory) calculation
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
3
The calculations were conducted with Materials Studio DMol software [78–80]. The used functional
was PW91, and the basis sets were builtꢀin double numerical functions plus ꢀfunctions for
d
nonꢀhydrogen atoms (DND3.5; comparable to 6ꢀ31G(d)). Core electrons were represented by DSPP
(DFT semiꢀcore pseudopotentials) [81] built in the software. The orbitals were unrestricted. The
2
CeO support was modeled with Ce70O140 cluster which is sculpted from the 5×5×5 supercell of the
fluorite structure (Figure S11). In this model cluster, there is no oxygen atom bonded to only one
atom, and there is no cerium atom bonded to less than 5 atoms, which minimizes the distortion of
1
1
1
2
structure from infinite CeO crystal. The geometry optimization was conducted without any
constraint. For small systems (MTO catalyst), transition states were optimized using initial Hessian
obtained by numerical differentiation. Frequency calculation for the optimized structure confirmed
the existence of one imaginary frequency mode, in which the two ReO–C bonds are being
13
14
15
simultaneously broken. For large systems (Ce70O140 cluster system), because calculation of Hessian
16
17
18
19
by numerical differentiation is practically difficult, the transition state was searched by fixing and
stepping the lengths (by 0.01 Å) of ReO…C bonds which are most changed near the transition state.
The maximum gradient of energy to Cartesian coordinates in the located transition state was below
<0.0015 Hartree/Å. For both small and large systems, the geometry optimizations using an initial
structure with slight shift from the obtained transition state gave either intermediate structure.
20
3
21
22
23
The functional (PW91) was selected because DMol program is designed for pure DFT and the
calculation with a more common hybrid functional, B3LYP, is much slower with this program. PW91
3
is one of major pure density functionals with GGA and it is the default setting of DMol program.
ACS Paragon Plus Environment
28

Page 29 of 56
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
1
2
3
4
5
6
7
8
9
0
1
2
The calculation level was compared with the literature study [82] using B3LYP functional for
methyltrioxorhenium(MTO)ꢀcatalyzed DODH of ethylene glycol (Table S6). The difference in the
bond lengths between the calculation methods is <0.03 Å for stable bonds, <0.08 Å for breaking
bonds in transition states (ReO…C), and <0.11 Å for coordination bonds of weaklyꢀbound ligand
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
(ReꢀOH ). The activation barriers calculated in this study were ~30% lower than those in the
literature study. These differences agree with the known trends of pure GGAꢀDFT which stabilizes
the breaking bonds in comparison with hybrid DFT.
Associated Content
1
1
1
Supporting Information
1
2
1
2
x 2
Screening of M and M in M O –M /CeO , detailed data for kinetics, effect of reductant, effect of
reaction temperature, additional XRD, TEM, XPS, EXAFS and Raman data, model structure and
energy diagram in DFT calculation, and supplier of reagents. This material is available free of charge
via the Internet at http://pubs.acs.org.
13
14
15
16
17
18
19
Acknowledgement
This work is in part supported by “ALCA” program of Japan Science and Technology Agency. We
thank Ms. M. Nemoto (Instrumental Analysis Group, Graduate School of Engineering, Tohoku
University) for Raman measurements. We also thank Technical Division, School of Engineering,
Tohoku University for XPS and TEM measurements.
20
21
22
23
References
[1] Schlaf, M. Dalton Trans. 2006, 4645–4653.
ACS Paragon Plus Environment
29