Hydrogenolysis of Glycidol as an Alternative Route to Obtain 1,3-Propanediol Selectively Using MOx-Modified Nickel-Copper Catalysts Supported on Acid Mesoporous Saponite

Article abstract of DOI:10.1002/cctc.201700536

Ni and Cu mono- and bimetallic catalysts modified with various types of acid oxides MO_x (M=Mo, V, W, and Re) were tested for the hydrogenolysis of glycidol as an alternative route to the hydrogenolysis of glycerol to obtain 1,3-propanediol (1,3-PrD). Characterization results revealed that the presence of modifiers affected the dispersion and reducibility of the NiO particles and the strength and amount of acid sites. Among the modifiers tested, Re led to the highest activity, a high propanediols selectivity, and the highest 1,3-PrD/1,2-propanediol (1,2-PrD) ratio. The Ni-Cu/Re ratio was optimized to improve the catalytic activity. The best catalytic result, with a 46 % 1,3-PrD yield and a 1,3-PrD/1,2-PrD ratio of 1.24, was obtained if the monometallic Ni catalyst at 40 wt % loading and modified with 7 wt % Re was used at 393 K and 5 MPa H₂ pressure after 4 h of reaction. The overall 1,3-PrD yield starting from glycerol and assuming a two-step synthesis (glycerol→glycidol→1,3-PrD) and a yield of 78 % for the first step would be 36 %. This 1,3-PD yield is the highest for a reaction catalyzed by a non-noble metal and is comparable to the direct hydrogenolysis of glycerol using noble metal catalysts at a longer time and a high H₂ pressure.

Full text of DOI:10.1002/cctc.201700536

Accepted Article
Title: Hydrogenolysis of glycidol as alternative route to obtain
selectively 1,3-propanediol using MOx modified Ni-Cu catalysts
supported on acid mesoporous saponite
Authors: Fiseha Bogale Gebretsadik, Jordi Llorca, Yolanda Cesteros,
and PILAR SALAGRE
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201700536
Link to VoR: http://dx.doi.org/10.1002/cctc.201700536
A Journal of
www.chemcatchem.org

10.1002/cctc.201700536
ChemCatChem
FULL PAPER
Hydrogenolysis of glycidol as alternative route to obtain
selectively 1,3-propanediol using MOx modified Ni-Cu catalysts
supported on acid mesoporous saponite
Fiseha B. Gebrestadik a Jordi Llorca b, Pilar Salagre,*, a Yolanda Cesteros a
Ni and Cu mono- and bi-metallic catalysts modified with various types
of acid oxides MOx (M = Mo, V, W and Re) were tested for
hydrogenolysis of glycidol, as alternative route to hydrogenolysis of
glycerol to obtain 1,3-propanediol. Characterization results revealed
that the presence of modifiers affected dispersion and reducibility of
the NiO particles, and the strength and amount of the acid sites.
Among the modifiers tested, Re showed the highest activity, high
propanediols selectivity and the highest 1,3-PrD/ 1,2-PrD ratio. Ni-
Cu/Re ratio was optimized to improve the catalytic activity. The best
catalytic result, with 46 % 1,3-PrD yield and 1,3-PrD/1,2-PrD ratio of
1.24, was obtained when monometallic Ni catalyst at 40 wt % loading
and modified with 7 wt % Re was used at 393 K, 5 MPa H2 pressure
after 4 h of reaction. The overall 1,3-PrD yield starting from glycerol
and assuming a two-step synthesis (Glycerol ‰ Glycidol ‰ 1,3-PrD)
and a yield of 78% for the first step, would be 36 %. This 1,3-PD yield
is the highest for non-noble metal catalyzed reaction and is
comparable to direct hydrogenolysis of glycerol using noble metal
catalysts, at longer time and high H2 pressure.
to 1,3-PrD is economically more attractive because 1,3-PrD is an
important monomer in the synthesis of polypropylene
terephthalate (PPT),
a
polymer that displays excellent
properties.[7] Additionally, the current industrial production of 1,3-
PrD is costly and involves the use of petrochemical based
hazardous starting materials.[8] In a recent review by Nakagawa
et al., tungsten-based materials or acid oxides, such as ReOx or
MoOx, which are capable of activating glycerol were used as co-
catalysts of several supported-noble metal catalysts in order to
favor the formation of 1,3-PrD.[5b]
With respect to the use of W compounds, the presence of
H2WO4, phosphotungstic acid (H3PW12O40), silicotungstic acid
(H4SiW12O40) or WO3 in the catalysts favored the selectivity to 1,3-
PrD.[9] This has been related to an increase of the amount of
Brønsted acid sites in these catalysts.[10] The best result was
obtained with a Pt-WOx/Al2O3 catalyst that resulted in a selectivity
to 1,3-PrD of 66% for a 64 % of conversion at 433 K and 5 MPa.[5b]
The role of ReOx in improving the selectivity to 1,3-PrD was
associated to the acidity of the -OH groups present on its surface,
[12]
which are capable of activating glycerol.[11] Ir–ReOx/SiO2
and
Rh–ReOx/SiO2 [13] catalysts were tested for the hydrogenolysis of
glycerol at 393 K in the presence of small amounts of mineral or
solid acid additives. In the former catalyst, 1,3-PrD was obtained
in 46% selectivity at 81% of conversion at 393 K, 8 MPa after 35
h while the later was equally active, but its selectivity to 1,3-PrD
was much lower (14%).
In general, the selective hydrogenolysis of glycerol to 1,3-
PrD requires expensive noble metal-based catalysts, high
temperature and/ or pressure and long reaction times. Even at
these reaction conditions, the yield of 1,3-PrD was low or
moderate.[5b] Hence, we propose the use of glycidol as an
alternative to glycerol to obtain propanediols by catalytic
hydrogenolysis, due to the high reactivity of glycidol and the fact
that it could be readily obtained from glycerol.[14]
Introduction
Biomass is a renewable source of organic carbon that is
becoming important for the production of sustainable fuels and
chemicals.[1] One of the most abundant biomass-related
renewable materials is glycerol, which is generated during
biodiesel production by transesterification of vegetable oils.[2]
Biomass has higher O/C ratio than most valuable chemicals.[3]
Hence, it is desirable to develop catalytic hydrogenolysis reaction
to transform it into useful chemical products.[4] Hydrogenolysis
catalysts should have the ability to activate hydrogen. Hydrogen
can be activated on a metal surface. Typical metals used in
hydrogenolysis include, nickel, copper and noble metals.[4c, 5]
Hydrogenolysis of glycerol has been widely studied. With
most of the catalytic systems, the main product obtained was 1,2-
propanediol (1,2-PrD).[6] However, the hydrogenolysis of glycerol
In the literature, the hydrogenolysis of tetrahydrofurfuryl
alcohol (THFA), a cyclic ether with similar substitution pattern as
[15]
glycidol, has been reported using Ir–ReOx/SiO2
, Rh–
ReOx/SiO2 [5b, 16], and Rh–MoOx/SiO2 [17] catalysts. These catalysts
led to 1,5-pentanediol (1,5-PD) with very high selectivity (> 90%).
The mechanism of THFA hydrogenolysis involved adsorption of
THFA, at the -CH2OH group, on the protonated ReOx or MoOx
surface. H2 activated on the metal attacks C–O bond
neighbouring the -CH2OH substituent and releases 1,5-PD.[15, 17a]
In our previous work on the hydrogenolysis of glycidol using
mesoporous acid saponite supported Ni catalysts, it was
observed that the presence of strong Brønsted acid sites,
improved the selectivity to 1,3-PrD.[18] The highest 1,3-PrD/1,2-
PrD ratio was 0.97 at total conversion at 453 K using a catalyst
with 40 wt % of Ni loading. In preliminary studies about the use
[a]
Dr. F. B. Gebretsadik, Dr. P. Salarge, Dr. Y. Cesteros
Dpt. Quimica Fisica i Inorgànica
Universitat Rovira i Virgili
C/Marcel·lí Domingo s/n
43007 Tarragona, Spain
E-mail: pilar.salagre@urv.cat
[b]
Dr. J. Llorca
Institut de Tècniques Energètiques
Universitat Politècnica de Catalunya
Avda. Diagonal, 647
08028 Barcelona (Spain)
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700536
ChemCatChem
FULL PAPER
of bimetallic Ni-Cu catalysts we found that the Ni-Cu bimetallic
catalyst at Ni: Cu atomic ratio 6:1 gave a better 1,3-PrD selectivity.
In this work, we modified several mesoporous acid saponite
supported Ni, Cu mono- and bimetallic catalysts using various
modifiers, MOx (M = Mo, V, W and Re) in order to investigate their
activity for the catalytic hydrogenolysis of glycidol. The effect of
metal to modifier ratio was studied for ReOx.
observed at 43.3 and 50.4º, 2θ, which were attributed to metallic
copper crystalline phase. In the other samples, Ni phase was
mainly observed and the Ni crystallite size increased when
increasing the Ni-Cu wt % loading. For the catalysts prepared
with lower Ni-Cu loading (30 and 20 wt%) (Ni-Cu*Re(7)/MS and
Ni-Cu**Re(7)/MS), the peaks appeared broader and a shoulder
was observed at about 43º, 2θ indicating the presence of residual
amounts of NiO (Figure 2(b) and (c), respectively). The Ni
crystallite sizes of these two samples were 6 and 3 nm,
respectively, in agreement with their lower metal loading. The
value increased to 11 nm for Ni-CuRe(7)/MS. The other samples
had comparable crystallite sizes (12-14 nm).
Results and Discussion
Characterization of the catalyst precursors and catalysts
XRD patterns of the Ni-Cu and modified Ni-Cu catalyst samples,
after reduction at 623 K for 6 h, are displayed in Figure 1. All
catalysts presented peaks at 19.4, 35.6 and 60.5º, 2θ,
corresponding to the (110, 020), (201) and (060) reflections of the
saponite support and at 44.5 and 51.9º, 2θ, related to the
presence of metallic nickel. Broad or shoulder peaks observed in
samples, Ni-CuMo(7)/MS, Ni-CuV(7)/MS and Ni-CuW(7)/MS
(Figure 1 (b), (c) and (d), respectively) at 43.2 and 62.8º, 2θ, were
identified as NiO phase). Peaks corresponding to the modifiers or
Cu were not observed in any sample, suggesting their small
crystallite size or low crystallinity. In addition, there was no
appreciable shift in the cell parameter of Ni in bimetallic catalysts
and the formation of a Ni-Cu alloy was not clear from the XRD
pattern. The presence of modifier seems to favor the dispersion
of Ni, since the Ni crystallite size decreased from 27 nm to 11-21
nm in the presence of the modifiers (Table 1). The acidity of the
modifiers and the basicity of the NiO phase and the subsequent
acid-base interaction could justify this improved dispersion.
Figure 2. XRD patterns of (a) NiRe(7)/MS (b) Ni-CuRe**(7)/MS (c) Ni-
Cu*Re(7)/MS (d) Ni-CuRe(3)/MS (e) Ni-CuRe(15)/MS (f) CuRe(7)/MS catalysts.
Figure 1. XRD patterns of Ni-Cu/MS samples with 40 wt % Ni-Cu loading (a)
without modifier (b) Mo (c) V (d) W (e) Re modifier at 7 wt % loading.
Figure 2 shows the XRD patterns of supported monometallic Ni
and Cu catalysts modified with Re and Ni-Cu bimetallic catalysts
with different wt % of Ni-Cu loading and different modifier amount.
For the monometallic modified Cu catalyst, sharp peaks were
Figure 3: TEM images of A) Ni-Cu/MS B)Ni-CuMo(7)/MS C) Ni-CuRe(7)/MS
and D) NiRe (7)/MS all at *100 K.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700536
ChemCatChem
FULL PAPER
TEM micrographies of the catalysts Ni-Cu/MS; Ni-CuMo(7)/MS;
Ni-CuRe(7)/MS and Ni-CuRe(7)/MS (Fig 3A, 3B, 3C and 3D
respectively) allowed us to observe that the nickel particles sizes
had similar values than the crystallite sizes obtained by XRD for
these samples.
modifiers. Another broader minimum, which appeared at an
inflection point of about 750 K in these 3 modified catalyst
precursors could be due to additional stronger acid sites
associated to the presence of the modifiers. In contrast, for PNi-
CuRe(7)/MS, mainly one minimum at 650 K was observed with
only a small shoulder at around 600 K. It can be argued that Re,
Ni and Cu oxides in PNi-CuRe(7)/MS are well dispersed, covering
the main part of the support external surface, favouring the
interaction of the H+ of the support with the ReOx, resulting in
mainly one type of acid sites with stronger acidity than that of the
support without modifiers. The stronger acid sites detected in the
catalyst precursors in the presence of modifiers could be
[15]
associated with –OH groups
modifiers.
or with Lewis acid sites of the
The TGA analysis results of cyclohexylamine treated Ni-Cu
catalysts precursors at different wt % Ni-Cu loading and modified
with different amounts of Re are displayed in figure 5. For catalyst
precursors with 40 wt % Ni-Cu loading, the small minimum
corresponding to the loss of CHA adsorbed on the acid sites of
the support (600 K), appeared more defined at 3 and 15 wt % Re
loading than at 7 wt %. When the Re loading was increased to 15
wt %, the position of the minimum related to the CHA adsorbed
on the acid sites generated by the presence of modifiers shifted
to higher temperature (700 K). This minimum also appeared at
about 700
K for PNi-Cu*Re(7)/MS and PNi-Cu**Re(7)/MS.
Figure 4. TGA results of cyclohexylamine treated samples of the modified
catalyst precursors.
However, for these two samples, the mass loss due to desorption
of CHA adsorbed on the acid sites corresponding to the support
increased when decreasing the Ni-Cu loading. These results are
consistent with the lower coverage of the support, when the Ni-
Cu loading was lower.
In order to explain the XRD results, it is necessary to have
information about the interactions between NiO and CuO phases
with the support and the modifiers, present in the precursors.
Additionally, the acidity values of the catalytic precursors could be
also interesting in order to have comparative information
regarding the acidity of the catalysts. For these reasons, acidity
and reducibility studies of the catalyst precursors were performed.
Figure 4 displays the thermogravimetric analysis results of
cyclohexylamine treated modified Ni-Cu catalyst precursors, used
for the determination of the total surface acidity of the samples
before reduction. The minima corresponding to the first derivative
of the TGA curve can be associated to different processes related
to the mass loss. Thus, those centred at about 350 and 1100 K,
can be attributed to the water loss due to the dehydration of
interlayer cations and dehydroxylation of the saponite structure,
respectively. The other minima correspond to the mass loss
associated with cyclohexylamine desorption. The number of
these minima can be attributed to different acid sites while their
temperature can be related to their acidity strength. Thus, the
minima appeared at higher temperature correspond to stronger
acid sites.
The total surface acidity of all the modified catalyst
precursors was higher (1.45-2.03 mEq. CHA/g) than that of the
unmodified Ni-Cu catalyst precursor (0.85 mEq. CHA/g) (Table 1).
This confirms that modifiers provide an additional acidity. The
values of total surface acidity increased with an increase in the
amount of modifier or a decrease in the wt % of Ni-Cu loading.
NH3-TPD experiments were performed in selected catalysts
precursor samples (Table 1) in order to supplement the results
obtained from the CHA-TGA experiments. It was observed that
the results obtained from the two methodologies are quite similar
with the values slightly higher in the case of the NH3 TPD. These
results are in agreement with the mesoporosity of the saponite
support and suggest that there was not any accessibility problem
to the surface acid sites.
The temperature-programmed reduction profiles of the
unmodified Ni-Cu catalyst precursors and those modified with Mo,
V, W and Re are displayed in figure 6. The temperature marked
in the figure corresponds to the NiO reduction. However, the
contribution of some modifier reduction cannot be discarded. The
difference in the position of the NiO reduction peak should be
consequence of its interaction with the support and/ or with the
modifiers, considering that NiO should be located between the
support and modifiers taking into account the order of
impregnation.
For PNi-CuMo(7)/MS¸ PNi-CuV(7)/MS and PNi-CuW(7)/MS
catalyst precursors, two types of acid sites were observed. The
minimum observed for all samples at about 600 K was similar to
that of the unmodified catalyst precursor. This can be related to
the acid sites of the support that were not influenced by the
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700536
ChemCatChem
FULL PAPER
Table 1. Characterization results of the catalyst precursors and catalysts.
BET
Av. pore
radius (Å)
Acidity [a]
(mEq./g)
Acidity [b]
(mEq./g)
Catalysts
Ni (NiO) [c]
Catalyst
Precursors
(m2/g)
crystallite
size (nm)
27
21 (3)
21 (6)
24 (4)
11
1.02
1.85
1.75
1.77
__
__
__
2.06
2.13
__
220
179
156
157
340
347
268
362
429
331
27
25
26
23
31
31
34
30
33
33
0.96
1.82
1.69
1.74
1.88
1.45
1.92
1.95
2.03
1.71
PNi-Cu/MS
Ni-Cu/MS
PNi-CuV(7)/MS
PNi-CuMo(7)/MS
PNi-CuW(7)/MS
PNi-CuRe(7)/MS
PNi-CuRe(3)/MS
PNi-CuRe(15)/MS
PNi-Cu*Re(7)/MS
PNi-Cu**Re(7)/MS
PNiRe(7)/MS
Ni-CuV(7)/MS
Ni-CuMo(7)/MS
Ni-CuW(7)/MS
Ni-CuRe(7)/MS
Ni-CuRe(3)/MS
Ni-CuRe(15)/MS
Ni-Cu*Re(7)/MS
Ni-Cu**Re(7)/MS
NiRe(7)/MS
12
14
6
3
13
Acidity in mEq. CHA/g of sample of the catalyst precursors from [a] TGA of cyclohexylamine (CHA) treated samples
and [b] NH3 TPD and [c] average crystallite size from XRD measurement.
lower reduction temperature (530 K) that could be correlated with
more effective NiO-Re2O7 interactions in some particles.
For Ni-CuW(7)/MS, the main reduction peak was shifted to
lower temperature by about 9 K. In our opinion, this should be
more related to a certain decrease of the NiO particle size respect
to the unmodified catalyst (from 27 to 24 nm) than to differences
in interactions, since the degree of NiO dispersion was probably
worse when the modifier was WO3. The higher size of the Ni
particles in this catalyst when compared with that of the other
modified catalysts is in agreement with this fact. This also agrees
with the fact that this modifier is the oxide with less acidity.
Figure 5. TGA of cyclohexylamine treated catalysts precursors with different
Re/Ni+Cu ratio.
For all the modified catalyst precursors, except for PNi-
CuW(7)/MS, the main reduction peak was shifted to higher
Figure 6. TPR profiles of Ni-Cu/MS catalysts precursors with 40 wt % of Ni-Cu
loading (a) without modifier (b) Mo (c) V (d) W and (e) Re modifier at 7 wt %
loading.
temperatures with respect to that
corresponding to the
unmodified one. This shift in the reduction temperature could be
related to smaller NiO particles, interacting strongly with the
support. This makes the reduction of the NiO more difficult.
However, the shift for the sample with Re was very small (7 K).
This lower shift is difficult to explain, but probably an opposite
effect could take place in this sample favoring the reduction; this
could be due to a decrease of the electronic density on the Ni(II)
by interaction with the Re2O7 phase, as will be discussed later. In
this sample, additionally a well-defined peak was observed at
Figure 7 shows the reduction profiles of some catalyst precursors
with different metal to Re ratio. The reduction of the modified
monometallic Ni catalyst precursor, PNiRe(7)/MS, occurred in a
broad interval (570-750 K) with two maxima, one more intense at
639 K, and other at 703 K. Comparing this graph with the curve
corresponding to
the modified bimetallic catalyst PNi-
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700536
ChemCatChem
FULL PAPER
CuRe(7)/MS of Fig 6(e), some part of NiO was easily reduced but
other was more difficult to reduce. These differences in reducibility
could be related to differences in NiO and Re2O7 interactions. As
previously commented, an effective NiO-Re2O7 interaction should
favor NiO reduction.
support (600 m2/g) [19] indicating that some of the pore volume was
occupied by the metal oxide particles. Comparing the BET areas
of the unmodified catalyst precursor with the modified ones, they
were lower for the V, Mo and W modified catalyst precursors. In
contrast, for those modified with Re, the BET area was much
higher (340 m2/g). This result suggested that Re can favor the
dispersion of the metal oxide particles to a greater extent than the
other modifiers. The smaller metal crystallite size of the
corresponding reduced catalysts, determined by XRD, could
support this fact, as will be discussed later.
At 3 wt % Re loading, the BET area was slightly higher (347
m2/g). However, when the Re loading was increased to 15 %, the
BET area decreased to 268 m2/g. The BET area of the modified
monometallic catalyst precursor, PNiRe(7)/MS (331 m2/g) was
slightly lower than that of the bimetallic modified catalyst
precursor, PNi-CuRe(7)/MS. This might indicate that the
presence of Cu could assist the dispersion of Ni. The BET area
values increased to 362 m2/g and 429 m2/g, respectively, when
the amount of Ni-Cu loading decreased to 30 and 20 wt %,
respectively. The average pore radius of all the catalyst
precursors was in the mesoporous range (Table 1).
Catalytic Activity
Figure 7. TPR profiles of the catalyst precursors (a) PNi(40)Re(7)/MS (b) PNi-
Cu*Re(7)/MS (c) PNi-Cu**Re(7)/MS.
Influence of the type of modifier
The catalytic activity involves a complex interplay of competitive
and/or cooperative metal and acid sites, as discussed later. For
this reason, it was not possible to report catalytic rates in terms of
turnover frequencies (TOF). Hence, in this work comparative
catalytic activity results are presented based on conversion and
selectivity.
Table 2 shows the catalytic activity results of Ni-Cu catalysts
supported on a mesoporous acid saponite, modified with various
types of acidic metal oxides. The activity should be a result of the
metal and acid sites contribution, responsible for mainly
hydrogenolysis and condensation reactions, respectively. These
are competitive reactions. However, a cooperative effect of metal
and acid sites could be expected in the hydrogenolysis reaction
resulting in an increase of the selectivity to 1,3-PrD. This could
take place by activation of glycidol, by H+ interacting with the
oxirane ring [11] or by interactions of acidic metal oxides MOx with
the alcohol functionality of the glycidol.[15]
For the other Ni-Cu catalyst precursors prepared with 20 and 30
wt % total metal loading, the reduction peaks shifted to slightly
higher temperatures. Two different reduction peaks were
observed for both samples, which might stand for NiO particles
with different interactions. The reducibility was more difficult for
the precursor that after reduction presented the smallest Ni
crystallite size (3 nm), probably due to a stronger interaction with
the support. The reduction peaks at lower temperature, at 642 and
659 K for PNi-Cu*Re(7)/MS and PNi-Cu**Re(7)/MS, respectively,
could be attributed to efficient NiO interactions with Re2O7 that
can favor its reduction. The lower reducibility of these samples
could justify the presence of residual amounts of NiO.
The N2-physisorption analysis results of the catalyst
precursors are presented in table 1. The specific BET areas of all
the catalysts precursors were lower than the BET area of the
Table 2. Catalytic activity test results.
Selectivity (%)
1,3-PrD/
1,2-PrD
1,3-PrD
yield (%)
Catalyst
Conversion
Propanol
Condensation products
1,2-PrD
1,3-PrD
(dioxane/ dioxalane)
CuRe(7)/MS
Ni-Cu/MS
40
73
73
88
95
98
98
98
96
98
98
9
8
9
5
7
5
5
8
8
6
8
7
7
59
21
28
22
2
0.89
0.33
0.33
0.45
0.24
0.76
0.59
0.83
0.82
0.89
1.24
3.2
47
48
43
66
45
49
42
38
35
38
15
16
19
16
34
29
35
31
31
47
11.0
11.7
16.7
15.2
31.6
28.4
34.3
31.7
33.3
46.1
Ni-CuMo(7)/MS
Ni-CuV(7)/MS
Ni-CuW(7)/MS
Ni-CuRe(7)/MS
Ni-CuRe(3)/MS
Ni-CuRe(15)/MS
Ni-Cu*Re(7)/MS
Ni-Cu**Re(7)/MS
NiRe(7)/MS
8
14
13
11
13
6
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700536
ChemCatChem
FULL PAPER
Re(7)/MS
47
n.d
n.d
__
59
__
__
The total Ni-Cu loading in * 30 wt% and ** 20 wt % in the other catalysts the total metal loading is 40 wt%. The Ni:Cu molar ratio
in Ni-Cu bimetallic catalysts is 6:1.
All the modified samples showed comparable or higher activity
than that of the unmodified catalyst. This might indicate that the
modifiers could have a catalytic role either by acid activation of
glycidol in the metal and H+ cooperative route or as acid catalysts.
The diol selectivity of these catalysts increased when W or Re
were used as modifiers. This could be explained by the higher
amount of metal phase, as a consequence of the higher
reducibility of the corresponding catalyst precursors. For Mo and
V modified catalysts, the propanediols selectivity was lower, in
agreement with their lower NiO reducibility and the presence of
residual amounts of NiO in the fresh catalysts.
All catalysts displayed comparable propanol selectivity
suggesting that the formation of propanol was not related to high
metal activity and was not the result of over hydrogenolysis of
propanediols. The presence of propanol can be justified by the
capacity of Cu to deoxygenate oxirane rings, as previously
reported in the literature.[20] Regarding the condensation products
selectivity (dioxane/ dioxolane), the unmodified catalyst and those
modified with Mo and V gave comparable selectivity (21-28%).
The selectivity towards these products was lower for Re and W
modified catalysts. These results agree with the higher metal
activity of the later samples.
All the catalysts gave comparable propanol selectivity (6-8%)
suggesting that its formation was neither related to acid sites nor
to the hydrogenolysis activity of the metal. The selectivity to
condensation products was higher at higher and lower Re loading.
At higher Re/M molar ratio, this can be related to the higher and
stronger acidity (Figure 3) while at lower Re loading, it could be
associated with lower hydrogenolysis metal activity due to its
lower NiO reducibility, as discussed earlier.
When Ni-Cu loading decreased (40, 30 and 20 wt %), the
molar ratio Re/Ni-Cu increased, resulting in different catalytic
behaviour. The propanediols selectivity slightly decreased and
the 1,3-PrD/1,2-PrD ratio increased. The selectivity to
condensation products increased when the Ni-Cu loading was
lower. The reducibility of NiO was difficult and acidity was higher
for samples with lower metal loading. Hence, the lower diol
selectivity and higher condensation products selectivity could be
explained by lower metal and higher acid activity, respectively.
However, this increase of acidity had a positive effect to improve
the selective formation of 1,3-PrD.
Regarding the 1,3-PrD/1,2-PrD ratio, the unmodified catalyst,
Ni-Cu/MS, showed a value of 0.33. The value decreased in the
case of W modified catalyst, was comparable in Mo modified
catalyst, slightly higher for V modified catalyst and higher for Re
modified catalyst (Table 2). The lower 1,3-PrD/1,2-PrD ratio
(0.24) of Ni-CuW(7)/MS could be associated to Ni activity without
acid collaboration, due to, probably a worse dispersion of WO3 on
the metal particles. For V and Re modified catalysts, the improved
1,3-PrD/1,2-PrD ratio might reflect the cooperation between acid
and metal sites. This was especially the case for Re modified
catalyst. The highest dispersion of the metal and the O-H groups
of the ReOx phase makes possible this collaborative action. In
general, among all the modified catalysts tested, the highest 1,3-
PrD yield (31.6%) was obtained with Ni-Cu-Re (7)/MS catalyst.
Scheme 1. The major reaction pathways of acid and/or metal catalyzed
reactions of glycidol.
Monometallic Re modified Ni and Cu and the saponite support
modified with Re without metal phase catalysts were prepared
(NiRe(7)/MS, CuRe(7)/MS and Re(7)/MS) and their catalytic
activity was tested for comparison (Table 3). Re modified Ni
catalyst showed the highest selectivity to propanediols (85%) and
1,3-PrD (47 %) with the highest 1,3-PrD/1,2-PrD ratio (1.24) and
low condensation products selectivity. The high 1,3-PrD
selectivity could be related to an effective collaborative action
between metal and acid sites. On the other hand, the modified Cu
catalyst afforded the lowest propanediol selectivity, higher
propanol selectivity and the highest condensation products
selectivity. This might reflect the poor hydrogenolysis activity of
Cu at the conditions used, and the high activity of the acid sites.
The propanol, as discussed earlier, is the result of Cu catalysed
deoxygenation reaction of the epoxide ring followed by
hydrogenation of the resulting double bond. Re modified support
Effect of Re to metal loading ratio
The effect of the amount of Re loading was studied in two ways;
in one way, the amount of Re loading was varied (3, 7, 15 wt %)
at constant Ni-Cu loading (40 wt %) and in the other, the Ni-Cu
wt % loading was varied (20, 30 and 40 %) at constant Re loading
(7 wt %).
The conversion was similar for all catalysts and was not
significantly affected by the Re to metal ratio. The selectivity
towards each of the products varied depending on the Re/Ni-Cu
ratio. When the amount of Re loading increased at constant Ni-
Cu loading, the 1,2-PrD selectivity decreased, the 1,3-PrD
selectivity slightly increased, and hence the 1,3-PrD/1,2-PrD ratio
increased. This might confirm that ReOx could be involved in the
activation of glycidol leading to the formation of 1,3-PrD.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700536
ChemCatChem
FULL PAPER
showed comparable moderate activity as that of the modified Cu
catalyst with identical condensation products selectivity. This
confirms the previous argument that condensation products can
be related to acid sites. The major reaction pathways of the acid
and metal catalysed reactions, discussed above, are summarized
in scheme 1.
In summary, among all the modified catalysts tested, Re
modified catalysts showed higher activity, propanediols selectivity
and 1,3-PrD/1,2-PrD ratio. The experiments about Re/M ratio
revealed that the optimum modifier and active metal loading was
obtained for the monometallic Ni catalyst at 40 wt % loading and
when the Re content was 7 wt %. The activity test at this optimum
composition, catalyst NiRe(7)/MS, at 393 K gave 1,3-PrD in
46.1 %yield with a 1,3-PrD/1,2-PrD ratio of 1.24 at 98 % of
conversion after 4 h of reaction.
was also performed for comparison. The binding energies of the
characteristics XPS peaks are summarized in table 3.
The surface composition of the fresh and spent samples was
compared to that of the theoretical composition of the catalysts
(Table 3). Regarding the ratio between Ni and Cu, the nominal
bulk values of the catalyst was Ni:Cu=6:1 by weight, which
corresponds to an atomic ratio of Cu/Ni=0.16. The fresh sample
Ni-CuRe(7)/MS exhibited a surface atomic ratio of Cu/Ni close to
this value, 0.15, which means that Ni and Cu are both equally
dispersed on the surface. The spent sample showed a slightly
higher Cu/Ni ratio of 0.19, indicating a slight enrichment of Cu at
the surface during reaction. In the H2 activated sample, the Cu/Ni
surface atomic ratio increased considerably up to 0.28, indicating
segregation of Cu towards the surface. The theoretical
Re/(Ni+Cu) atomic ratio was approximately 0.052 for samples
NiRe(7)/MS and Ni-CuRe(7)/MS. The values of the Re/(Ni+Cu)
atomic ratios at the surface measured by XPS were 0.031 in
NiRe(7)/MS and 0.047 in Ni-CuRe(7)/MS, in the fresh samples.
This means that the dispersion of Re was good especially when
Cu was present in the catalyst. For NiRe(7)/MS catalyst, it could
be possible that Re was decorated by Ni. Interestingly, the spent
samples showed comparable Re/(Ni+Cu) atomic ratios than the
fresh ones, 0.039 and 0.045. In contrast, the samples reduced in-
situ at 623 K had lower Re/metal atomic ratios, 0.020 Re/ Ni and
0.029 Re/(Ni+Cu). The decoration of ReOx by the Ni particles
could explain a best collaborative effect between Ni and ReOx in
the reaction, which is considered responsible for the observed
1,3-PD production with NiRe(7)/MS.
XPS study: Ni-CuRe(7)/MS vs NiRe(7)/MS catalysts
In order to better understand the differences in the catalytic
behaviour of NiRe(7)/MS and Ni-CuRe(7)/MS catalysts, three
XPS experiments were designed. Samples were prepared and
analyzed as described in the experimental section. The first group
of samples were named as fresh catalysts, since they were the
catalysts obtained just after reduction and before reaction.
However, in these samples some degree of oxidation could be
expected during catalyst handling for the XPS analysis. The
second group of samples were the catalysts, which were activated
by reduction in-situ with H2 at 623 K for 30 min, and the third group
of samples were the spent catalysts, which were obtained after
being used in the catalytic test and without any treatment. The
XPS of the support modified with ReOX without metal, Re/MS,
Figure 8. Re 4d spectra of A) fresh and H2 activated Re(7)/MS and fresh, H2 activated and spent catalysts, from left to right,
of B) NiRe(7)/MS and C) Ni-CuRe(7)/MS. The red line indicates the position of Re in a low oxidation state whereas the blue
one indicates the position of oxidized Re, both corresponding to 4d5/2 photoelectrons. The 4d3/2 peaks are overlapped to the
C 1s photoelectrons.
peaks, since the 4d3/2 peaks are superposed to the C 1s
photoelectrons. All the spectra showed two components
corresponding to two different oxidation states for Re. The red line
indicates the position of the Re 4d5/2 photoelectrons
The Re 4d spectra of the different catalysts as prepared (fresh),
after in-situ H2 activation at 623 K, and after reaction (spent) are
displayed in figure 8. The spectra of Re correspond to the 4d5/2
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700536
ChemCatChem
FULL PAPER
corresponding to the lower oxidation state whereas the blue one
indicates the position of the Re 4d5/2 photoelectrons
corresponding to the higher oxidation state. In the fresh samples,
Re was mainly in the higher oxidation state for NiRe(7)/MS) and
Ni-CuRe(7)/MS) catalysts, whereas in the Re(7)/MS sample the
contribution of the less oxidized form was more important. After in
situ H2 activation at 623 K, Re became less oxidized, but there
was always a component of highly oxidized Re (probably VI or
VII). It is interesting to note that the reduction degree of Re was
similar for samples Re and NiRe(7)/MS) (88-93%), but in the
sample Ni-CuRe(7)/MS, Re was much more difficult to reduce
(after reduction there was about 75% Re reduced to lower
oxidation state). Probably, the Cu-Re interaction favors the
electronic transfer from Re to Cu, increasing the oxidation state of
Re. The spent samples exhibited both Re species, in similar
amounts as in the corresponding fresh samples.
Figure 9 shows the Ni 2p spectra for the fresh, activated and
spent (NiRe(7)/MS and Ni-CuRe(7)/MS) samples. The fresh
samples exhibited similar Ni spectra, with Ni mainly oxidized, but
with the presence of a minor component of reduced Ni as well.
After in-situ H2 activation at 623 K, Ni was reduced considerably,
but not totally. It is interesting to highlight that a significant fraction
of Ni was still oxidized after the in-situ reduction in the Ni-
CuRe(7)/MS) sample (about 75% of oxidized Ni). In contrast, the
degree of Ni reduction was much higher in the sample without Cu,
the NiRe(7)/MS sample (about 57% of oxidized Ni). This fact,
together with the difficulty of reducing Re in the Ni-CuRe(7)/MS
sample during the in-situ reduction treatment, as pointed out
above, suggests that the presence of Cu strongly interacting with
both Re and Ni and that Cu prevents both Re and Ni reduction.
This is likely an electronic donation from Re and Ni towards Cu.
Finally, the Ni spectra of NiRe(7)/MS and Ni-CuRe(7)/MS spent
samples were virtually identical and similar to the corresponding
fresh samples. These experiments showed that in the catalyst
without Cu, NiRe(7)/MS, there was a higher amount of reduced
Ni, which in turn can be correlated with its higher propanediol
selectivity.
Considering the real catalysts, from the commented XPS
results, it is possible to argue that there is a higher amount of Re
with higher oxidation state in Ni-CuRe(7)/MS catalyst compared
to NiRe(7)/MS catalyst. This might result in a higher acid activity
of the catalyst, Ni-CuRe(7)/MS, according to the observed
products distribution.
Figure 9. Ni 2p spectra of the fresh, H2 activated and spent catalysts, from left to right of A) NiRe(7)/MS and B) Ni-
CuRe(7)/MS. The red line indicates the position of reduced Ni whereas the blue one indicates the position of oxidized Ni
(2p3/2 photoelectrons).
Table 3. XPS analysis results of selected fresh, H2 activated and spent catalysts.
Catalyst
Binding energies (eV) of characteristic peaks
Ni Cu Re
Ox 4d5/2-265.1
Re/
Ni-Cu
Ni/Cu
Ni(Red)/
Ni (Ox)
Cu(Red)/
Cu(Ox)
Re/MS-fresh
__
__
__
__
__
__
__
__
__
__
Red 4d5/2-260.8
Ox 4d5/2-264.8
Red 4d5/2-260.6
Ox 4d5/2-263.9
Red 4d5/2-260.6
Ox 4d5/2-263.5
Red 4d5/2-260.6
Ox 4d5/2-263.8
Red4d5/2-259.5
Ox 4d5/2-263.4
Re/MS- H2 act.
__
__
Ni-CuRe(7)/MS-fresh
Ni-CuRe(7)/MS- H2 act.
Ni-CuRe(7)/MS-spent
NiRe(7)/MS-fresh
Ni0 2p3/2-852.4
Ni2+ 2p3/2-854.4
Ni0 2p3/2-852.5
Ni2+ 2p3/2-855.8
Ni0 2p3/2-852.5
Ni2+ 2p3/2-854.5
Ni0 2p3/2-852.6
Cu2+ 2p3/2-933.1
Cu0,+ 2p3/2-931.1
Cu2+ 2p3/2-928.5
Cu0,+ 2p3/2-931.7
Cu2+ 2p3/2-932.7
Cu0,+ 2p3/2-932.2
__
0.047
0.029
0.045
0.031
0.15
0.28
0.19
__
0.14
0.15
0.45
0.15
1.14
1.76
3.60
__
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700536
ChemCatChem
FULL PAPER
Ni2+ 2p3/2-854.7
Ni0 2p3/2-852.4
Ni2+ 2p3/2-855.9
Ni0 2p3/2-852.3
Ni2+ 2p3/2-854.6
Red4d5/2-260.0
Ox 4d5/2-263.9
Red 4d5/2-261.3
Ox 4d5/2-265.0
Red 4d5/2-261.0
NiRe(7)/MS- H2 act.
NiRe(7)/MS-spent
__
__
0.020
0.039
__
__
0.13
0.96
__
__
The total Ni-Cu loading in * 30 wt% and ** 20 wt % in the other catalysts the total metal loading is 40 wt%. The Ni:Cu molar ratio in Ni-
Cu bimetallic catalysts is 6:1.
Figure 10 shows the Cu 2p spectra for the sample Ni-
the components may correspond to metallic Cu and the other one
can be assigned to a Cu-M interaction. In fact, the electron
donation from Re and Ni towards Cu in the Ni-CuRe(7)/MS)
sample, discussed above, agrees with the observed strongly
reduced component in the Cu 2p spectrum, which is compatible
with the existence of a type of Cu that acts as an electron acceptor.
Finally, for the spent sample the spectrum was similar to that of
the fresh one, although the lower intensity of the satellite lines in
the 2p spectrum indicates a slightly higher reduction character in
the spent sample.
CuRe(7)/MS) as well as the Cu LMM Auger lines. In the fresh
sample both reduced and oxidized Cu can be observed. However,
after in situ H2 activation at 623 K the spectrum changed
dramatically and Cu was strongly reduced (about 78%). This was
also clearly observed in the Auger lines, where the peaks at 334.9
and 332.0 eV clearly showed the presence of metallic Cu. It is
outstanding the absence of satellite lines and the appearance of
two different reduced Cu species after the in situ reduction
treatment in the Cu 2p spectra. At this point it is difficult to
unambiguously identify the nature of these two species. One of
Figure 10. Cu 2p (A) and Auger Cu LMM (B) spectra of fresh, H2 activated and spent Ni-CuRe(7)/MS catalyst. The red line
indicates the position of reduced Cu whereas the blue one indicates the position of oxidized Cu (2p3/2 photoelectrons). In addition,
Cu(II) species are accompanied by satellite lines (sat).
with a decrease in the amount of Ni-Cu loading. The dispersive
effect was higher in the presence of Re modifier.
The catalytic activity of all the modified catalysts was
comparable or higher than the unmodified catalysts. The
presence of V and Re modifier increased the 1,3-PrD and 1,2-PrD
ratio to 0.45 and 0.7, respectively when compared to the
unmodified catalyst (0.33) or to those modified with W (0.24) or
Mo (0.33). The results confirmed that Re plays an important role
in improving the 1,3-PrD production. The W modified catalyst
gave the highest selectivity to propanediol (82%), the lowest
selectivity to condensation products (2%), and the lowest 1,3-
PrD/1,2-PrD ratio, in agreement with its high NiO reducibility and
hence metal activity.
Conclusions
Ni-Cu catalysts prepared at 40 wt % total metal loading were
modified with 7 wt % Mo, V, W and Re loading. Some catalysts at
different Ni-Cu/Re ratio and Re modified monometallic Ni and Cu
catalysts, were also prepared for comparison. The
characterization results revealed that the presence of vanadium
and molybdenum modifiers made the reduction of NiO more
difficult. The reducibility of NiO got harder when the amount of
metal loading was lower. Total surface acidity and acidity strength
increased in the presence of modifiers, with an increase in the
amount of modifier and with a decrease in the amount of metal
loading. BET area and XRD analysis confirmed that the
dispersion of the NiO and the Ni in the catalyst precursors and
catalysts, respectively, increased in the presence of modifiers and
The important role of Re modifier was confirmed by the
increase of 1,3-PrD/1,2-PrD ratio when Re/ Ni + Cu ratio
increased by increasing the Re content or decreasing the Ni-Cu
loading. However, at lower metal loading the diol selectivity was
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700536
ChemCatChem
FULL PAPER
respectively. The (Ni + Cu)/Re molar ratios for these catalysts were 13 and
9. Monometallic Ni and Cu catalysts were prepared at 40 wt % metal
loading and modified with 7 wt % Re, for comparison.
Table 4. Nomenclature, type of modifier, amount of metal and modifier
loading of the catalysts.
Catalysts
Total Ni-Cu
wt % loading
Modifier
M (MOx)
Modifier
wt %
All catalysts were reduced in a tubular reactor under hydrogen flow (75
mL/ min) at 623 K for 6 h. The preparation conditions and nomenclature of
the catalysts are summarized in table 1. The catalyst precursors were
labelled with P at the beginning of the name.
Ni-Cu/MS
40
40
40
40
40
40
40
30
20
40
40
__
V
Mo
W
Re
Re
Re
Re
Re
Re
Re
__
7
7
7
7
3
15
7
7
7
Ni-CuV(7)/MS
Ni-CuMo(7)/MS
Ni-CuW(7)/MS
Ni-CuRe(7)/MS
Ni-CuRe(3)MS
Ni-CuRe(15)/MS
Ni-Cu*Re(7)/MS
Ni-Cu**Re(7)/MS
NiRe(7)/MS
Catalysts characterization
X-ray diffractogram (XRD) patterns of powdered catalyst samples were
recorded on a Siemens D5000 diffractometer equipped with a CuKα
radiation (λ =1.54 Å) source. Measurements were done in 2θ diffraction
angle between 5 and 70°, with an angular step of 0.05° and at rate of 3 s
per step. Crystalline phases were identified by cross comparison of the
diffractogram of the sample with a reference data from international centre
for diffraction data (JCPDS files). Integral breadth estimated by fitting
intense reflection peaks of metal or metal oxide particles and by applying
TOPAS 4.1, was used to calculate the crystallite sizes using the Scherrer
equation.
N2 adsorption-desorption analysis was performed on Quadrasorb
sorptometer at liquid nitrogen temperature (77 K) to determine the specific
BET area and porosity. Before measurement, samples (~0.1 g) were
degassed overnight at 383 K. Specific surface area was determined by
applying the BET method and the pore size distribution was estimated by
BJH method.
TEM characterization was done using JEOL 1011 transmission
electron microscope operating at an accelerating voltage of 100 kV and
magnification of 200 k. Sample (0.1 mg) was dispersed in ethanol (50 µL)
with the aid of an ultrasounds. Then, it was deposited on a carbon coated
copper grid and air dried.
Temperature-programmed reduction of the catalyst precursors was
measured by an Autochem AC2920 Micrometric apparatus. The catalyst
precursor sample, calcined beforehand (0.1 g), was placed in a tubular
reactor and heated between room temperature and 1173 K at a rate of 10
K/min under a flow of 5 vol.% H2 in argon (50 mL/min). The hydrogen
consumption was controlled by a TCD detector.
CuRe(7)/MS
7
lower and the selectivity to the condensation products was higher
as would be expected from the low NiO reducibility and high
acidity of these samples. The monometallic Cu catalysts gave the
lowest propanediol and the highest condensation products
selectivity comparable to that of the Re modified support without
an active metal. This confirmed the lower hydrogenolysis capacity
of Cu in the catalytic system. The Re modified Ni catalysts
showed high 1,3-PrD/1,2-PrD ratio (1.24), high propanediol
selectivity and the highest 1,3-PrD yield (46.1%) of all the
catalysts studied. These results can be explained by some
decoration of the Ni particles by the ReOx, as can be deduced
from XPS results,that favor the collaborative action between the
metal sites and ReOx responsible for the 1,3-PrD production. The
yield of 1,3-PrD obtained in this work was comparable or higher
than those previously reported in the literature for the
hydrogenolysis of glycerol, using noble metal catalysts, at harsh
reaction conditions and longer reaction times.
Experimental Section
Total surface acidity of the catalyst precursors was determined by
thermogravimetric analysis (TGA) of cyclohexylamine treated samples
(~0.1 g) following the protocol described by Mokaya et al.[21] The analysis
was conducted in a Labsys Setaram TGA microbalance equipped with a
programmable temperature furnace. Each sample was heated under N2
flow (80 cm3/ min) from 303 to 1173 K at a rate of 5 K/ min. The total acidity
was equivalent to the amount of cyclohexylamine (CHA) desorbed at the
Preparation of supported catalysts
Mesoporous saponite in the H+ form (MS) (surface acidity, 1.02 mEq H+/g)
synthesized as previously reported by our research group [19], was used as
support for catalyst preparation. MS was stable until 1035 K, (temperature
of clay dehydroxylation), as determined by TGA.[19] Ni-Cu/MS catalysts
with 20, 30 and 40 wt % total metal loading, and Ni: Cu wt % ratio 6:1 were
prepared by impregnation of 1 g of the support with a calculated volume of
0.5 M Ni-Cu in ethanol solution containing appropriate amounts of
Ni(NO3)2 and Cu(NO3)2 salts. After removing the solvent by rotary
evaporation and drying at 383 K overnight, the samples were calcined at
723 K for 5 h.
Several Ni-Cu-MOx/MS (M = W, Mo, V and Re) catalysts were
prepared by impregnating the Ni-Cu/MS (prepared at 40 wt % Ni-Cu
loading) with an aqueous solution of the modifier metal precursors,
(NH4)10W12O41·4H2O, (NH4)6Mo7O24·4H2O, NH4VO3 and NH4ReO4. The
amount of MOx modifier expressed per mass of W, Mo and V respect to
unmodified catalyst was 7 wt %. The amount of Re loading was varied at
3, 7 and 15 wt %, resulting in catalysts, Ni-CuRe(3)/MS, Ni-CuRe(7)/MS
and Ni-CuRe(15)/MS, with molar ratio values (M/Re) of 42, 18 and 8
respectively.
temperature range of 523-923
K in which the sample without
cyclohexylamine treatment did not show any significant mass loss.[19]
Several NH3-TPD experiments were performed on selected catalyst
precursors to complement the acidity results obtained by CHA-TGA
analysis, and to compare the relative contribution of external and internal
acid sites. The analysis was conducted on AC2920 apparatus. In a typical
experiment, about 0.2 g, sieved catalyst precursor sample was loaded in
U-shaped quartz tube and submitted to surface pre-treatment under a flow
of He at 623 K for 1 h. It was then cooled down to 373 K and saturated
with pure NH3 for 30 min. The physically adsorbed NH3 was removed by
purging the sample with pure He for 30 min and the sample was heated
from 373 to 973 K at a heating rate of 10 K/min and the NH3 desorption
was monitored with a TCD detector. The peak area of detected TPD peak
was correlated with the amount of desorbed NH3 on the basis of pulsed
NH3 injection experiment and the total surface acidity was predicted from
the amount of ammonia desorbed.
The effect of Ni-Cu/Re wt % ratio was also studied by decreasing the
metal loading at 30 and 20 wt % Ni-Cu wt % and maintaining the Re
loading as in the Ni-CuRe(7)/MS catalyst. This was performed in order to
increase the acid contribution of the support by decreasing the support
acid site coverage. The resulting catalysts were labelled as Ni-
Cu*Re(7)/MS and Ni-Cu**Re(7)/MS, for 30 and 20% of metal loading,
Surface characterization of some fresh, spent and H2 in-situ activated
catalysts was performed by X-ray photoelectron spectroscopy (XPS) on a
SPECS system equipped with an Al anode XR50 source operating at 150
mW and a Phoibos MCD-9 detector. The pressure in the analysis chamber
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700536
ChemCatChem
FULL PAPER
was always kept below 10-7 Pa. The area analyzed was about 2 mm x 2
mm. The pass energy of the hemispherical analyzer was set at 25 eV and
the energy step was fixed at 0.1 eV. Powdered samples were pressed to
self-supported pellets. In-situ experiments were performed under dynamic
conditions in an adjacent chamber at atmospheric pressure equipped with
a mass spectrometer and an IR lamp to heat the sample. The sample was
transferred under ultra-high vacuum between the in-situ chamber and the
analysis chamber. Gases were accurately dosed into the in-situ chamber
by using mass flow controllers, and the temperature was measured by
using a K-type thermocouple in contact with the sample holder. Data
processing was performed with the CasaXPS program (Casa Software
Ltd., UK). The binding energy (BE) values were referred to the C 1s peak
at 284.8 eV. Atomic fractions were calculated using peak areas normalized
on the basis of acquisition parameters after background subtraction,
experimental sensitivity factors and transmission factors provided by the
manufacturer.
[1]
a) M. Schlaf, Dalton Trans. 2006, 4645-4653. b) A. Corma, S. Iborra, A.
Velty, Chem. Rev. 2007, 1072411-2502. c)P. C. A. Bruijnincx, Ramản-
Leshkov, , Catal. Sci. Technol. 2014, 4, 2180-2181.
[2]
[3]
C.-H. Zhou, J. N. Beltramini, Y.-X. Fan, G. Q. Lu, Chem. Soc. Rev., 2008,
37, 527–549.
a) R.M. West, E.L. Kunkes, D.A. Simonetti, J.A. Dumesic, Catal. Today,
2009, 147, 115-125. b) A. Yamaguchi, N. Hiyoshi, O. Sato, K.K. Bando,
M. Shirai, Green Chem. 2009, 11, 48-52.
[4]
a) Y. Nakagawa, K. Tomishige, Catal. Sci. Technol. 2011, 1, 179-190. b).
S. Sitthisa, T. Sooknoi, Y. Ma, P.B. Balbuena, D.E. Resasco, J. Catal.
2011, 277, 1-13. c). K.L. Deutsch, B.H. Shanks, J. Catal. 2012, 285, 235-
241.
[5]
[6]
a) P. Mäki-Arvela, J. Hájek, T. Salmi, D. Y. Murzin, Appl. Catal., A, 2005,
292, 1–49. b) Y. Nakagawa, M. Tamura, K. Tomishige, J. Mater. Chem.
A, 2014, 2, 6688-6702.
a) M. A. Dasari, P.-P. Kiatsimkul, W. R. Sutterlin, G. J. Suppes, Appl.
Catal., A, 2005, 281, 225–231. b) Y. Kusunoki, T. Miyazawa, K. Kunimori,
K. Tomishige, Catal. Commun. 2005, 6, 645–649. c) T. Miyazawa, S.
Koso, K. Kunimori, K. Tomishige, Appl. Catal., A, 2007, 329, 30–35. d) J.
Feng, J. Wang, Y. Zhou, H. Fu, H. Chen, X. Li, Chem. Lett., 2007, 36,
1274–1275. e) T. Sánchez, P. Salagre, Y. Cesteros, A. B. López, Chem.
Eng. J. 179 (2012) 302– 311.
Catalytic activity test
Activity test was performed in a 50 mL stainless steel autoclave equipped
with an electronic temperature controller and a mechanical stirrer. Glycidol
(3.77 mL) was dissolved in sulfolane (30 mL) and purged with nitrogen and
1 g of freshly reduced catalyst was carefully transferred into the reactor
under a positive nitrogen pressure to avoid contact with atmosphere. The
reactor was sealed, purged three times with 4 MPa N2 pressure and three
times with 1 MPa H2 pressure. After performing a leakage test, the
temperature of the reaction was set at 393 K. The reaction mixture was
vigorously stirred at 600 rpm. This together with the use of powdered
catalysts and the mesoporous nature of the support used for catalyst
preparation avoids internal and external mass and heat transfer limitations
and diffusion problems. When the temperature in the electronic control
read the final temperature of the reaction, the hydrogen pressure was
adjusted at 5 MPa and the reaction time was started (to). The reaction was
run for 4 h while feeding hydrogen on demand. At the end of the reaction
time (tf), the mixture was cooled down and the reaction products were
separated from the catalyst by filtration. Liquid products were analysed on
Shimadzu GC-2010 chromatography using SupraWAX-280 capillary
column, 1-butanol as internal standard and FID detector. Some liquid
samples were submitted for GC-MS analysis to identify unknown products.
Conversion and selectivity were calculated according to the following
equations:
[7]
a) P. N. Caley, R. C. Everett, Patent US 3350871, 1967, DuPont. b) D.
Zimmerman, R. B. Isaacson, Patent US 3814725, 1974, Celanese Corp.
A. Perosa, P. Tundo, Ind. Eng. Chem. Res. 2005, 44, 8535-8537.
a) T. Kurosaka, H. Maruyama, I. Naribayashi, Y. Sasaki, Catal. Commun.
2008, 9, 1360–1363. b) J. Chaminand, L. Djakovitch, P. Gallezot, P.
Marion, C. Pinel, C. Rosier, Green Chem. 2004, 6, 359–361. c) N. Suzuki,
Y. Yoshikawa, M. Takahashi, M. Tamura, World Pat. 2007129560, 2007,
KAO Corporation. d) E. I. Ross-Medgaarden, W. V. Knowles, T. Kim, M.
S. Wong, W. Zhou, C. J. Kiely, I. E. Wachs, J. Catal. 2008,256, 108–125.
e) L. Liu, Y. Zhang, A. Wang, T. Zhang, Chin. J. Catal. 2012, 33, 1257–
1261. f) J. ten Dam, K. Djanashvili, F. Kapteijn, U. Hanefeld,
ChemCatChem. 2013, 5, 497–505. g) R. Arundhathi, T. Mizugaki, T.
Mitsudome, K. Jitsukawa, K. Kaneda, ChemSusChem. 2013, 6, 1345–
1347. h) S. Zhu, X. Gao, Y. Zhu, Y. Li, J. Mol. Catal. A: Chem. 2015, 398,
391-398. i) S. Zhu, Y. Qiu, Y. Zhu, S. Hao, H. Zheng, Y. Li, Catal. Today
2013, 212, 120-126. j) S. Zhu, X. Gao, Y. Zhu, J. Cui, H. Zheng, Y. Li,
Appl. Catal., B, 2014, 158-159, 391-399.
[8]
[9]
[10] L. Gong, Y. Lu, Y. Ding, R. Lin, J. Li, W. Dong, T. Wang, W. Chen, Appl.
Catal., A, 2010, 390, 119–126.
[11] M. Chia, Y. J. Pagán-Torres, D. Hibbitts, Q. Tan, H. N. Pham, A. K. Datye,
M. Neurock, R. J. Davis, J. A. Dumesic, J. Am. Chem. Soc. 2011, 133,
12675–12689.
ꢗ ꢊꢋꢌꢍꢎ ꢋꢏ ꢐꢌꢑꢒꢓꢔꢋꢌ @ ꢕ
ꢀꢁꢂꢃꢄꢅꢆꢇꢁꢂ ꢈ ꢉꢊꢋꢌꢍꢎ ꢋꢏ ꢐꢌꢑꢒꢓꢔꢋꢌ @ ꢕ
ꢙ ∗ 100%
ꢖ
ꢘ
ꢊꢋꢌꢍꢎ ꢋꢏ ꢐꢌꢑꢒꢓꢔꢋꢌ ꢕ
ꢖ
[12] Y. Nakagawa, Y. Shinmi, S. Koso, K. Tomishige, J. Catal. 2010, 272,
191–194.
ꢠꢁꢛꢄꢆ ꢁꢡ ꢟ ∗ # ꢜꢢꢅꢣꢁꢂ ꢇꢂ ꢟ
% ꢚꢄꢛꢄꢜꢝꢇꢃꢇꢝꢞ ꢟ ꢈ
∗ 100%
ꢤꢠꢁꢛꢄ ꢥꢛꢞꢜꢇꢦꢁꢛ @ ꢝꢋ ꢧ ꢠꢁꢛꢄ ꢥꢛꢞꢜꢇꢦꢁꢛ @ ꢝꢏ ꢨ ∗ 3
[13] K. Chen, S. Koso, T. Kubota, Y. Nakagawa, K. Tomishige,
ChemCatChem. 2010, 2, 547–555.
Where to and tf are the initial and the final times, respectively.
[14] a) W. Yoo, Z. Mouloungui, A. Gaset, USP 6316641, 2001. b) R. Bai, H.
Zhang, F. Mei, S. Wang, T. Li, Y. Gu and G. Li, Green Chem. 2013, 15,
2929-2934.
[15] K. Chen, K. Mori, H. Watanabe, Y. Nakagawa, K. Tomishige, J. Catal.
2012, 294, 171–183.
Acknowledgements
[16] S. Liu, Y. Amada, M. Tamura, Y. Nakagawa, K. Tomishige, Green Chem.
2014, 16, 617–626.
The authors are grateful for the financial support of the Ministerio
de Economia y Competividad of Spain and FEDER funds
(CTQ2011-24610). J. Llorca is Serra Húnter Fellow and is grateful
to ICREA Academia program.
[17] a) S. Koso, N. Ueda, Y. Shinmi, K. Okumura, T. Kizuka, K. Tomishige, J.
Catal. 2009, 267, 89–92. b) S. Koso, H. Watanabe, K. Okumura, Y.
Nakagawa, K. Tomishige, Appl. Catal., B. 2012, 111–112, 27–37.
[18] F. B. Gebretsadik, J. Ruiz-Martinez, P. Salagre, Y. Cesteros, Appl. Catal.,
A, DOI: 10.1016/j.apcata.2017.03.018.
Keywords: glycidol • hydrogenolysis • nickel • metal oxide
[19] F. B. Gebretsadik, D. Mance, M. Baldus, P. Salagre, Y. Cesteros, Appl.
Clay Sci. 2015, 114, 20–30.
modifier • propanediol • bifunctional catalyst
[20] M. Bartók1, A. Fási, F. Notheisz, J. Catal. 1998, 175, 40–47
[21] R. Mokaya, W. Jones, S. Moreno, G. Poncelet, Catal. Lett. 1998, 49, 87-
94.
This article is protected by copyright. All rights reserved.