Supported Co–Re Bimetallic Catalysts with Different Structures as Efficient Catalysts for Hydrogenation of Citral

Article abstract of DOI:10.1002/cssc.201802744

Bimetallic Co–Re/TiO₂ catalysts were developed for efficient citral hydrogenation. Bimetallic catalysts were prepared by co-impregnation (CI), successive-impregnation (SI), and surface redox method (SR). The arrangement between the Co and Re species on these systems was fully characterized using several techniques (TEM–energy-dispersive X-ray spectroscopy, H₂ temperature-programmed reduction, temperature-programmed desorption, XRD, CO FTIR spectroscopy, model reaction of cyclohexane dehydrogenation), and their catalytic performances were evaluated for the selective hydrogenation of citral towards unsaturated alcohols. The Re and Co species are completely isolated in the CI sample, presenting a very limited Co–Re interaction. In SI samples, the metals coexist in a Janus-type structure with a concentration of Re around Co. Decoration/core–shell structures are observed for SR samples resulting from the redox exchange between the metallic surface of the parent Co/TiO₂ catalyst and the Re⁷⁺ species of the modifier precursor salt. The contact degree between the two metals gradually increases as follows: Isolated structure (CI)a disadvantage for the hydrogenation reaction. For SR samples, the increase of Re loading contributes to the electron transfer from Re to Co that is consistent with a change of structure from decoration to core–shell. The lack of directly accessible Co atoms for SR catalysts with fully coated structure decreases the efficiency of Re reduction. The presence of Co–Re interaction resulting from the close contact between metals plays a dominant role in the hydrogenation of citral. Nevertheless, an excessively high contact degree is unnecessary for citral hydrogenation once Co–Re interaction has formed.

Full text of DOI:10.1002/cssc.201802744

Accepted Article
Title: Supported Co-Re bimetallic catalysts with different structures as
efficient catalysts for hydrogenation of citral
Authors: Gwendoline Lafaye, Xin Di, Catherine Especel, Florence
Epron, Ji Qi, and Changhai Liang
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: ChemSusChem 10.1002/cssc.201802744
Link to VoR: http://dx.doi.org/10.1002/cssc.201802744
A Journal of
www.chemsuschem.org

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
Supported Co-Re bimetallic catalysts with different structures as
efficient catalysts for hydrogenation of citral
Xin Di [
a], [b]
, Gwendoline Lafaye * , Catherine Especel , Florence Epron , Ji Qi , Chuang Li ,
[b]
[b]
[b]
[a]
[a]
Changhai Liang *[a]
Abstract: Bimetallic Co-Re/TiO
2
catalysts were developed for
pesticides and foods. Citral can be easily produced in large
quantities without complicated synthesis steps, that makes the
hydrogenation of citral an interesting economical route to obtain
unsaturated alcohols [ . Furthermore, this reaction is attractive to
develop new catalytic formulations for application in the
hydrogenation of bio-based molecules, since citral can be
efficient citral hydrogenation. Bimetallic catalysts were prepared by
co-impregnation (CI), successive-impregnation (SI) and surface redox
method (SR). The arrangement between Co and Re species on these
systems was fully characterized by several techniques (TEM-EDX,
1]
2
H -TPR, TPD, XRD, CO-FTIR, model reaction of cyclohexane
dehydrogenation) and their catalytic performances were evaluated for
the selective hydrogenation of citral towards unsaturated alcohols.
The Re and Co species are totally isolated in CI sample, presenting a
very limited Co-Re interaction. In SI samples, metals coexist in a
considered as a model molecule with three potentially
hydrogenable double bonds, an isolated C=C bond in addition to
conjugated C=C and C=O bonds, of which the selective
hydrogenation is very sensitive to the nature and the arrangement
of the active phases.
Janus-type structure with
a concentration of Re around Co.
Decoration/Core-shell structures are observed for SR samples
Noble metal-based monometallic catalysts are very active for
C=C hydrogenation, but they frequently tend to exhibit over-
hydrogenation activities associated with limited C=O
resulting from the redox exchange between the metallic surface of the
7+
2
parent Co/TiO catalyst and the Re species of the modifier precursor
[2]
salt. The contact degree between the two metals gradually increases
as follows: Isolated structure (CI) < Janus-type structure (SI) <
Decoration/Core-shell structure (SR). The unchanging structure of all
SI samples whatever the Re loading leads to similar electron transfer
and the increase in Re content results in agglomeration of rhenium,
thus decreasing the catalytic activity. Density of state (DOS)
calculation proves that high valence of Re is a disadvantage for
hydrogenation reaction. For SR samples, the increase of Re loading
contributes to the electron transfer from Re to Co, that is consistent
with a change of structure from decoration to core-shell. The lack of
directly accessible Co atoms for SR catalysts with fully coating
structure decreases the reduction efficiency of Re. The presence of
Co-Re interaction issued from close contact between metals plays a
dominant role in the hydrogenation of citral. Nevertheless, excessive
contact degree is unnecessary for citral hydrogenation once Co-Re
interaction has formed.
hydrogenation properties . Their modification by a non-noble
promoter is a common way to increase their selectivity, and many
bimetallic supported systems such as Pt-Ge, Pt-Co, Rh-Ge or Pd-
Sn have demonstrated high catalytic performances for citral
hydrogenation [3]. While rhenium can be an interesting promoter
for this reaction, since it can effectively reduce C=O bond, to the
best of our knowledge it has never been used as active phase for
this reaction. As the activity of Re-based catalysts is significantly
influenced by Re oxidation state, it is the most often associated
with noble metals in bimetallic M-Re nanoparticles in order to
improve catalytic activity [4]
.
In bimetallic catalysts, the first approach consists in the finding of
optimal metal combination, but the different structures of the
catalyst, and especially the contact degree between the two
metals, can also play a crucial role in the synergetic effect [5] . For
M-M’ based bimetallic catalysts associating a metal M able to
activate hydrogen to another one M’ difficult to reduce, the
arrangement between the two metals can be roughly divided into
two categories: (1) M is in close contact with M’ and thus the
dissociated hydrogen can directly transfer from M to M’ [5a, 6], thus
Introduction
increasing the reduction efficiency of this latter. This direct
bimetallic contact contributes to the possibility of geometrical
effect and electron transfer between the two metal species; (2) M
is isolated from M’, and the dissociated hydrogen will transfer from
M to M’ via the support [5a, 6] and the possibility of electron transfer
α, β-unsaturated aldehydes, such as citral, a naturally produced
terpenoid abundantly found in the oil of lemon grass, can be
selectively hydrogenated into unsaturated alcohols (UA), valuable
intermediates widely used in perfumes, pharmaceuticals,
will be relatively limited. The type of contact between the two
metals in bimetallic nanoparticles is complex, with the possible
formation of alloy, core-shell or other shapes. Thus, more in tune
[
[
a]
b]
X. Di, J. Qi, C. Li, C. Liang.
Laboratory of Advanced Materials and Catalytic Engineering,
Dalian University of Technology
No.2 Linggong Road, Dalian City, Liaoning Province, P.R.China,
E-mail: changhai@dlut.edu.cn
G. Lafaye, C. Especel, F. Epron
with the reality, the reduction efficiency and electronic transfer will
depend on the different structural properties. For example [
,
5c]
Pd(shell)–Au(core)/TiO
2
,
Au(shell)–Pd(core)/TiO
2
and
PdAu(alloy)/TiO catalysts were synthesized and the sample with
2
Institut de Chimie des Milieux & Matériaux de Poitiers (IC2MP)
Université de Poitiers
Pd(shell) and Au(core) morphology exhibits higher activity for the
total oxidation of VOCs than the other ones. Besides, some
studies devoted to Pt-Co catalysts demonstrated that an intimate
physical and/or chemical bimetallic interaction is not necessarily
required to obtain a beneficial catalytic effect, since a simple
4
rue Michel Brunet, 86073 Poitiers, France
E-mail: gwendoline.lafaye@univ-poitiers.fr
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
mechanical mixture of Pt/Al
improve the reducibility of Co O
2
O
3
and Co/Al
2 3
O was proven to
species [5d]. It follows that the
3 4
influence of different contact types on the structure and catalytic
activity of nanoparticles depends strongly on the two metals and
cannot be generalized.
Therefore, the purpose of this work was to investigate the
influence of different types of contact in M-Re bimetallic catalysts
on catalytic hydrogenation performances. With the aim of
replacing costly and rare noble metals which are usually
associated with Re [4a], this study will report the use for the first
2
time of bimetallic Co-Re/TiO catalysts for citral hydrogenation,
with various degrees of metal-metal interactions depending on the
preparation method. Impregnation methods and a surface redox
method [7] were chosen in order to prepare a series of Co-Re/TiO
samples with different contact degrees between the two metals.
2
Results and Discussion
Morphology and structure of CI and SI catalysts.
Firstly, the structures of the bimetallic Co-Re/TiO
2
samples
prepared with CI and SI methods were analyzed. Figure 1(A)-(E)
compares the TEM-EDX results obtained for both Re and Co
monometallic samples and the SI and CI bimetallic catalysts.
Figure 1(A) shows that 5Re/TiO
2
consists of very small particles
(
mean diameter = 0.8 nm) well-dispersed on the titania surface.
The lattice spacing (2.10 Å) observed on HRTEM images clearly
points to hexagonal rhenium [101] lattice plane. According to
Figure 1(B), 5Co/TiO
2
catalyst shows a high dispersion of Co
particles on TiO although the particle size is rather big (~ 25 nm).
2
The large accumulation of Co particles makes lattice spacing
difficult to be observed by HRTEM in Figure 1(B).
Figure 1. TEM-EDX and elemental mapping of (A) 5Re/TiO
SI-4Co-1Re/TiO , (D) SI-4Co-5Re/TiO and (E) CI-4Co-1Re/TiO
and SI samples. (G) Schemes of CI and SI catalysts
2
, (B) 5Co/TiO
2
, (C)
2
2
2
. (F) Particle
distribution of 5Re/TiO
structure.
2
Figure. 1(C) and Figure. 1(D) show the TEM and corresponding
elemental mapping for SI-4Co-1Re/TiO and SI-4Co-5Re/TiO
samples, respectively. For SI-4Co-1Re/TiO , the lattice spacing
.02 Å, 1.91 Å and 2.08 Å can be attributed to the Co [111], Co
101] and Re [101] lattice planes, respectively. Similar lattice
2
2
2
2
[
spacing 2.01 Å, 1.93 Å and 2.11 Å assigned to Co [111], Co [101]
and Re [101] lattice planes, respectively, were observed on SI-
4
Co-5Re/TiO
2
. The little shift in lattice spacing compared with
is caused by the Co-Re interaction and the
monometallic Re/TiO
2
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
influence of the support. According to elemental mappings
obtained on both SI samples, Re and Co do not exist as isolated
species thus indicating the presence of bimetallic entities. Based
on TEM results, Co-Re bimetallic catalysts prepared by SI method
can be described mainly by a Janus-type structure (Figure 1(G))
with particles displaying two size populations as illustrated on
Figure 1(F). Indeed, the metallic Co species pre-existing on the
the value of H
performed on the ex situ reduced Co/TiO
(result not shown) leads to a very small H
result reveals that the cobalt species are practically not oxidized
during the storage of the catalyst. The reduced cobalt species are
then stable on this sample. On the other hand, Figure 2 shows
2
consumption. Besides, the TPR experiment
2
monometallic catalyst
consumption. This
2
that monometallic Co/TiO
2
can be oxidized by water since the
5Co/TiO
2
catalyst can contribute to the reduction of NH
4
ReO
4
4Co-H O/TiO sample consumes an important amount of
2
2
precursor during the second impregnation step, which will be also
in favor of concentrating Re particles around Co (these Re
particles exhibit relatively small average sizes, i.e. 1.1 and 2.3 nm
hydrogen during TPR analysis. The reduced monometallic Co
catalyst is thus oxidized after immersion in pure water [8], but the
reduced state can be recovered by a subsequent reduction
2 2
for SI-4Co-1Re/TiO and SI-4Co-5Re/TiO samples, respectively). treatment.
Further, agglomeration of rhenium is gradually significant with the
increasing of Re content according to average particle size.
Besides, these Re species as deposited on the catalyst surface
contribute to the re-dispersion of Co species in return, leading to
a decrease of the particle size of cobalt. Indeed, when Re loading
increases from 1 wt.% to 5 wt.% on the SI catalysts, the average
particle size of Co decreases from 18.8 to 8.8 nm, which is in all
cases lower than that observed for the Co/TiO
2
parent sample.
Concerning the CI-4Co-1Re/TiO sample, rhenium particles are
2
not found in Figure 1(E), which may be caused by the low loading
and highly dispersion of Re. Furthermore, the average particle
size of Co is similar to that of monometallic Co catalyst. It seems
that Co and Re species exist as isolated monometallic particles
and that there is almost no interaction between them in this
sample (Figure 1(G)). The difference in structure between CI-
2 2
4Co-1Re/TiO and SI-4Co-1Re/TiO catalysts can be explained
by the different preparation methods used to synthesis these
samples. Especially, in the case of SI sample, the Co species are
pre-reduced before impregnation of Re salt what undeniably
contributes to generate Co-Re interactions.
TPR analysis was used to characterize the reducibility of catalysts.
This technique shows the way in which different species in the
catalyst interact with each other. The results of the temperature
programmed reduction experiments and the reduction degree of
Re species are shown in Figure 2 for Co and Re monometallic
samples, as well as for CI and SI bimetallic samples, all analyzed
directly after the impregnation step of the Co and/or Re salts. The
Figure 2.
H
2
-TPR results of SI-4Co-XNH
4
ReO
4
/TiO
2
,
CI-4CoCl
2
-
1
NH ReO /TiO
4
4
2
n
, 5NH ReO /TiO , 4CoCl /TiO2, 4Co-H
4
4
2
2
2
O/TiO
2
and bare TiO
2
profile for the bare TiO
around 300 °C which originates from the surface reduction of TiO
however the resulting H consumption is very small which could
2
support presents a reduction peak at
support. In Re , n indicates the oxidation state at the end of the TPR. (4Co-
O/TiO : 4 wt.% Co/TiO was ex situ reduced, treated by water for 12 h, then
2
,
H
2
2
2
2
dried under 90°C for 24 h after evaporation).
be ignored compared with the consumptions obtained on other
samples.
Since the reduction degree of Re could play a key role in the
structure and performance of bimetallic catalysts, this one was
determined from TPR results obtained for the SI and CI bimetallic
catalysts according to the (1) and (2) following formulas:
The main reduction temperatures for 4CoCl
2
/TiO
2
at 441 and 526
precursor.
/TiO
catalyst is obtained at the end of the TPR procedure according to
o
C can be attributed to the gradual reduction of CoCl
2
The complete reduction of the Co2+ species of the CoCl
2
2
(
1)Reduction degree (RD %) of Re for SI-4Co-XNH
4
ReO
4
/TiO
ReO
Expected hydrogen consumption of Re precursor for SI-4Co-XNH
2
catalysts
/TiO - Hydrogen consumption of 4Co-H
ReO /TiO
Hydrogen consumption of SI-4Co-XNH
4
4
2
2 2
O/TiO
=
×100
×100
4
4
2
(
2) Reduction degree (RD %) of Re for CI-4Co-1NH
4
ReO
4
/TiO
2
Hydrogen consumption of CI-4CoCl
2
-1NH
4
ReO
4
/TiO
2
- Hydrogen consumption of 4CoCl
-1NH ReO /TiO
2 2
/TiO
=
Expected hydrogen consumption of Re precursor for CI-4CoCl
2
4
4
2
(
3) Average oxidation state of Re after reduction of the SI or CI sample = +[7 × (1 - RD(%) / 100)] with RD ≤ 100%
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
In each case, the expected hydrogen consumptions were
calculated by assuming the fully reduction of Re7+ precursor.
From the reduction degree values, an average valence state was
then estimated for the Re species of each Re-based catalyst
existence of a contact between Re and Co species in SI samples
thus modifying their reduction properties.
Morphology and structure of SR catalysts.
(
equation (3)).
The TPR results of 5NH
4
ReO
4
/TiO
2
sample show that Re
The differences in metal arrangement and in Co-Re interactions
precursor salt impregnated alone on the titania support cannot be
fully reduced in the studied temperature range, and the hydrogen
consumption value indicates that the Re species are
observed between CI-4Co-1Re/TiO
catalysts result from the different methods used to prepare these
samples. For SI samples, the Co species are pre-reduced before
2
and SI-4Co-1Re/TiO
2
stoichiometrically reduced until an average oxidation state of +2.4. impregnation of Re salt. In that case, we can assume that a redox
So we can make the hypothesis that Re is mainly present as a
mixture of Re and high valence Re , such as Re , Re and/or
reaction occurs between the reduced Co species and the Re salt.
To verify this hypothesis, a series of Co-Re bimetallic catalysts
was prepared by enhancing the surface redox reaction between
the pre-reduced parent Co and the oxidized form of Re. The
preparation protocol is given in the catalyst preparation section
0
n+
3+
4+
7
+
Re . The reduction peak for bimetallic SI-4Co-XNH
catalysts is located around 350 °C, i.e. between the temperature
ranges observed on monometallic 5NH ReO /TiO sample (thin
peak at 298 °C) and on the 4Co-H O/TiO sample (large massif
4 4 2
ReO /TiO
4
4
2
2
2
and the resulting catalysts are referred as SR-XCo-YRe/TiO
2
(X
until 500 °C). Therefore, this result can be attributed to the co-
reduction of Re and Co species. The significantly higher reduction
degree of Re for SI bimetallic catalysts compared to the
monometallic Re sample signifies an improvement in the rhenium
reducibility. The Co-Re interaction coming from Janus-type
structure of SI bimetallic samples contributes to the reduction of
and Y represent the respective expected weight metal loadings).
4 4
NH ReO precursor, which in turn is in favor of concentrating Re
particles around Co. In that case, the reduction efficiency of Re
should gradually decrease with increasing the Re loading,
2
because of the agglomeration of rhenium clusters. The H -TPR
results effectively prove this conclusion since the average
oxidation state of Re after reduction increases from 0 to +1.1 with
Re loading increasing from 1 wt.% to 5 wt.%. The SI-4Co-
4 4 2
1NH ReO /TiO catalyst exhibits even a value of reduction degree
above 100 %, so a very superior reduction efficiency than other
catalysts of the series. It also means that not only the surface
redox process but also the hydrogen spillover is involved in the
Re reduction mechanism, beneficial for Co-Re interaction.
However, hydrogen spillover has a limited influence on the
calculation of reduction degree, as it was demonstrated in a
previous study [7a]. The
NH ReO /TiO catalyst can be understood as the addition of the
profiles of 4CoCl /TiO and 5NH ReO /TiO monometallic
samples. Although the reduction range of CoCl species appears
at slightly lower temperature on the CI bimetallic catalyst than on
the monometallic 4CoCl /TiO sample, the reduction degree and
average oxidation state of Re (+2.7) are comparable to the value
obtained on monometallic 5NH ReO /TiO sample. This result
demonstrates that a Co-Re interaction is not clearly present in CI-
Co-1Re/TiO sample, as previously proved by TEM results. To
2 2
H -TPR profile of CI-4CoCl -
1
4
4
2
2
2
4
4
2
2
2
2
4
4
2
4
2
achieve significantly improvement in Re reducibility, the intimate
contact between Co and Re is then necessary in Co-Re bimetallic
samples. The Re and Co isolated particles in CI sample are
disadvantageous for increasing Re reduction efficiency compared
with Janus-type structure in SI samples.
2
This series of catalysts was also characterized by H -TPR after
their preparation and activation by reduction. For that purpose,
the samples were pre-oxidized at 300 °C for 1 h prior to the TPR
experiment. The results and specific analysis are shown in
Supporting Information (Figure S1). It leads to a same conclusion
as the one emerging from the discussion above proving the
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
of SR samples, SR-5Co-5Re/TiO
2
catalyst was analyzed by
HAADF-STEM (Figure 3(D)). Unsurprisingly, the bimetallic
particles are highly dispersed on TiO support and the linear
2
scanning of single particle exhibits that Co is fully covered by Re.
Besides, the EDX results at different positions in linear scanning
show that only Re is present on the edge of the observed particle
(
① and ③) and both Co and Re exist in the middle (②), which
further confirms the core-shell structure of SR-5Co-5Re/TiO
This analysis was carried out on several randomly selected
2
.
2
particles on SR-5Co-5Re/TiO (Figure S2) leading to the same
conclusion.
Figure S3 shows the H
preparation, reduction and re-oxidation before TPR experiment.
TPR profiles of SR-5Co-XRe/TiO catalysts show that the addition
of Re to Co/TiO has a substantial effect on the reduction
2
-TPR results of SR samples after their
2
2
behavior: the reduction peaks of the SR bimetallic sample display
a maximum around 320 °C i.e. between the reduction ranges of
both monometallic catalysts, and the bimetallic catalysts exhibit
significantly higher hydrogen consumption ratios (HCR,
Supporting Information) than monometallic catalysts, evidencing
close contact between Re and Co species. An increase of Re
2
Figure 3. TEM-EDX and elemental mapping of (A) SR-5Co-1Re/TiO and (B)
SR-5Co-5Re/TiO . (C) Particles distribution of SR samples. (D) HAADF-STEM,
2
content causes a decrease of the H
2
/Re value and the SR-5Co-
2
Re/TiO sample exhibits the highest HCR value of the SR
2
linear scanning and associated EDX of SR-5Co-5Re/TiO
catalysts structure.
2
. (E) Schemes of SR
bimetallic series. The SR method is based on the surface redox
reaction between pre-existing Co metallic particles and Re
oxidized species of the precursor salt, leading to decoration and
core-shell structures. The deposition of a high Re content would
protect Co surface of catalysts, leading to fully Re coated surface
(Figure 3(E)). The lack of accessible Co atoms for SR-5Co-
Figure 3(A) and Figure 3(B) illustrate the TEM-EDX results
obtained for SR-5Co-1Re/TiO and SR-5Co-5Re/TiO catalysts
2
2
prepared by surface redox reaction. The existence of a single
population of particle size and the EDX analysis prove that Re is
tightly bonded with Co generating bimetallic particles with quite
homogeneous sizes. The elemental mappings show that the
addition of Re induces a significant decrease of the Co particle
size, by comparison to the monometallic Co catalyst, which is in
accordance with the preparation method principle. The decrease
in average particle size intensifies with increasing Re loading, with
average values evolving from 3.3 to 1.2 nm (Figure 3(C)).
Moreover, the distribution of Re becomes much denser for SR-
2
5Re/TiO catalyst may limit the hydrogen transfer and then
decrease the reduction efficiency of ReO
compared with other SR samples.
x
to a certain extent
Further structural analysis of the different catalysts.
The catalysts prepared in this work present different metal
arrangement and the contact degree between Co and Re would
gradually increase in this order: CI sample (Isolated structure) <
SI sample (Janus-type structure) < SR sample (Decoration and
core-shell structures). In this part, the influence of catalyst
structures on the interaction between metals will be analyzed and
compared.
2 2
5Co-5Re/TiO than that for SR-5Co-1Re/TiO (Figure 3(B)
compared to Figure 3(A)). These results suggest that the structure
of SR catalysts gradually changes from decoration structure to
core-shell structure when Re loading is high (Figure 3(E)). The
SR preparation method, that is a redox process between Co metal
7
+
and Re ions of the Re precursor salt, can indeed promote the
formation of such structures. For further insight into the structure
Table 1. Physicochemical properties of monometallic and bimetallic catalysts.
Metal loading (wt.%)
S
(
BET
m2 g-1)
V
pore
-3 -1
Average pore
size (nm)
Average
particle size
(nm) [
Acid sites
Samples
TiO
(moL g cat) [b]
-1
(cm g )
/
2
a]
Co
-
Re
-
TiO
Co
CI-4Co-1Re
2
65.4
52.8
61.5
0.34
0.32
0.37
20.9
24.2
24.2
-
48
33
-
5
5.3
4.1
-
25
30
1.0
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
SI-4Co-1Re
SI-4Co-2Re
SI-4Co-3Re
SI-4Co-5Re
SR-5Co-1Re
SR-5Co-2Re
SR-5Co-5Re
4.2
4.0
4.0
4.2
4.0
3.9
2.2
-
1.0
1.9
2.8
4.5
1.1
1.5
4.8
4.5
55.6
58.8
53.7
54.7
58.6
55.2
51.8
66.1
0.38
0.37
0.34
0.32
0.34
0.35
0.31
0.33
27.2
25.0
25.0
23.2
23.1
25.6
23.9
19.7
1.1/18.8
33
-
-
-
-
2.3/8.8
3.3
-
60
40
-
1.2
0.8
68
42
5Re
[
3
a] Determined by TEM analysis. [b] Determined by NH -TPD analysis.
The main characteristics of catalysts synthesized by different
methods gathered in Table 1 show that the preparation of the Co-
Re bimetallic catalysts results in a slight decrease in surface area
2
of lone TiO , in a lower extend in the case of CI-4Co-1Re sample
compared to the others. The pore volume and average pore size,
around 0.35 cm-3 g-1 and 25 nm, respectively, are rather similar
whatever the bimetallic sample.
The ICP-OES analysis of the catalysts reveals that the actual
metal loadings are in agreement with the expected ones.
Moreover, it must be pointed out that for SR samples the Co
content gradually decreases when increasing Re loading,
because of the oxidation of Co metal atoms by Re ions and
elimination of these oxidized species by filtration of the solution at
2
the end of the SR protocol. A BLANK-5Co/TiO sample was also
prepared in order to evaluate the influence of the preparation
conditions used during the SR preparation procedure on the
2
monometallic Co/TiO catalyst, and notably the acidity of the
medium (pH 1). For that purpose, the parent 5Co catalyst was
treated in the same way as for preparing a SR bimetallic catalyst
but without introduction of the Re salt solution (replaced by a
solution of water acidified to pH 1). Acid solution leads to a partial
leaching of the Co phase since a decrease of the Co load is
Figure 4. XRD patterns of Co-Re/TiO2 catalysts prepared by SI (A) and SR (B)
methods. BLANK-5Co/TiO : 5 wt.%Co/TiO treated by HCl solution (pH 1) under
atmosphere.
2
2
N
2
2
observed from 5.3 (5Co/TiO sample) to 3.6 wt.% (BLANK-
5
Co/TiO ). As the loss of Co content is lower on the SR-5Co-1Re
2
The XRD patterns of Co and Re monometallic catalysts, and of SI
and SR bimetallic samples gathered in Figure 4, exhibit mainly
peaks attributed to the TiO support phases (anatase and rutile).
2
No significant peak attributed to Re phase is observed for
monometallic and bimetallic catalysts. It indicates that Re
2
particles are very small or highly dispersed on TiO support. A
very small diffraction peak is observed on monometallic Co
catalysts (zoom in in the spectrum of the 5Co and Blank-5Co
samples).
and SR-5Co-2Re samples (Table 1), this suggests that during the
SR procedure H ions would be in priority involved in the reduction
process of Re species instead of dissolving pre-existing Co atoms,
according to the following possible surface redox reactions (E0
values given for pH equal to 0):
Re species reduction:
ReO + 8H + 7e →Re + 4H O
ReO + 2H + e → ReO + H O
+
-
+
-
0
0
E = 0.37 eV/SHE
4
2
-
4
+
-
0
E = 0.77 eV/SHE
3
2
Co species oxidation:
Co → Co + 2e
0
2+
-
0
E = -0.28 eV/SHE
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
Figure 5. Infrared spectra of the adsorbed CO species over Re and Co-Re
catalysts prepared by (A) CI, SI and (B) SR methods.
Figure 5 shows the in situ IR spectra of CO species adsorbed on
monometallic Re and bimetallic Co-Re catalysts. CO adsorption
on Co/TiO
1
2
results in a spectrum without any peak in the range
850-2150 cm-1 (not shown). The spectrum of CO adsorbed on
-1
Re/TiO
2
reveals a band at 2027 cm , assigned to linearly
adsorbed CO on Re [9]. Linearly adsorbed CO was observed at
2
032 cm-1 in the case of SI-4Co-1Re/TiO
catalysts that is at higher wavenumber compared to Re/TiO
2
and SI-4Co-5Re/TiO
2
2
Figure 6. H
2
-TPD profiles of monometallic Co, Re and bimetallic Co-Re
samples.
(
Figure 5(A)). This shift towards higher wavenumber indicates an
[
9a, 9b,
electron transfer from Re to Co caused by Co-Re interaction
10].
Additionally, the vCO value is the same for SI-4Co-1Re/TiO
2
2
A H -TPD analysis was also performed in order to study the
2
and SI-4Co-5Re/TiO , implying similar structure for the various SI
hydrogen adsorption ability of the different samples and to relate
the information to their structures (Figure 6 and Table 2). After
being activated by metal, the adsorbed hydrogen can dissociate
from catalysts in the course of the TPD analysis program. The
hydrogen desorption amount of 5Co/TiO catalyst was used as
2
standard for quantitative analysis and the relative amount of
hydrogen desorption was calculated with the following equation:
samples, which is consistent with TEM images indicating a Janus-
type structure for all SI-4Co-XRe/TiO catalysts. Besides, the
wavenumber of CO adsorption peak for CI-4Co-1Re/TiO is much
closer to that obtained for the monometallic Re/TiO proving the
2
2
2
structure of isolated particles on that catalyst and limited Co-Re
interaction. Figure 5(B) represents the FT-IR spectra of SR
samples. The shift towards higher wavenumber for vCO band is
also observed for these samples, and moreover the wavenumber
(4) Relative amount of hydrogen desorption
-1
gradually increases from 2031 to 2035 cm with the increasing of
Re content. The different frequencies indicate the presence of
different adsorption sites, that is, CO is adsorbed on Re atoms
with different electron density. The arrangement of the metallic
particles changes from decoration to core-shell structure for SR
catalysts with the increasing of Re content, resulting in more
significantly electron transfer. On the basis of these results, SR-
Peak area of 5Co, 5Re, SI-4Co-XRe or SR-5Co-XRe
=
Peak area of 5Co
The hydrogen adsorption ability (HAA) was calculated by defining
the values corresponding to each monometallic sample as equal
to one, in order to evaluate the influence of Co-Re interaction on
the hydrogen adsorption ability. To calculate the HAA value, the
5
Co-5Re/TiO
2
sample shows a stronger Co-Re interaction than
and SR-5Co-2Re/TiO catalysts. To
areas of the H -TPD peaks of the monometallic catalysts are
2
the SR-5Co-1Re/TiO
2
2
weighted to take into account the differences of metallic content
(of Co and of Re) between the monometallic and bimetallic
samples. The HAA value was calculated according to the
following equation:
conclude, FT-IR analyses give clear evidence about the
difference of structure and adsorption sites for Co-Re catalysts
according to the preparation method. It also indicates that Re
species in a metallic state are present on the samples after a
reduction at 350 °C.
(
5) Hydrogen adsorption ability (HAA) of bimetallic A sample
Peak area of A sample
=
Peak area of 5Co
Co content of 5Co
Peak area of 5Re
Re content of 5Re
Co content of A sample ×
+ Re content of A sample ×
with A = CI-4Co-1Re, SI-4Co-XRe or SR-5Co-XRe catalyst
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
ability, especially for SR-5Co-2Re/TiO
2
sample displaying the
highest HAA value of the series (1.9), which is consistent with H
2
-
Table 2. H
samples.
2
-TPD analysis of monometallic Co, Re and bimetallic Co-Re
TPR results. Besides, the Co-Re interaction on SR catalysts also
contributes to hydrogen desorption at lower temperature (Peak III).
It is noteworthy that dissociated hydrogen on SR samples
desorbed at around 700 °C (Peak IV) in a similar way than on
Relative desorption
quantity
Hydrogen adsorption ability
(HAA)
Samples/TiO
2
2
5Re/TiO (Peak II). It indicates the existence of similar desorption
5
Co
Re
1.00
1.54
1.26
1.94
2.76
2.86
1.58
2.40
2.92
1.0
1.0
1.1
1.7
1.6
1.2
1.4
1.9
1.4
sites, causing the consistent desorption behavior. The
morphology of the SR catalysts changes from decoration to core-
shell structure when increasing Re content, leading to the
increase of Re coverage. The full coverage of Co by Re on SR-
5
CI-4Co-1Re
SI-4Co-1Re
SI-4Co-3Re
SI-4Co-5Re
SR-5Co-1Re
SR-5Co-2Re
SR-5Co-5Re
5
Co-5Re/TiO
2
catalyst results in similar desorption behavior than
. The results of H -TPD in turn prove
that of monometallic Re/TiO
2
2
the different structures of SR samples depending on the Re
loading.
Catalytic results.
Dehydrogenation of cyclohexane was used as a probe reaction to
evaluate the accessibility and the nature of active surface metal
sites as this model reaction performed in gas phase is considered
as structure insensitive [12]. The results of cyclohexane
dehydrogenation (Table 3) show that the total conversion
obtained with SI bimetallic catalysts is higher than the one
obtained on monometallic Co and Re catalysts and the value
tends to increase with the increase of Re content. The
The thermal desorption curves of hydrogen for monometallic
samples can be divided into two main peaks. Peak I ranging from
00 to 500 °C can be attributed to desorption of atomic hydrogen
directly bonded to active metal Co and Re. The hydrogen
adsorbed on the metal surface can migrate in the form of atomic
hydrogen onto the neighboring support via surface diffusion, and
1
2
monometallic Re/TiO exhibits a low conversion, with similar
percentages of dehydrogenated product (benzene) and by-
products resulting from C-C cleavages (C1-C5 products). In
addition, the percentage of by-products also gradually becomes
majority for SI bimetallic catalysts when increasing Re loading.
the migrated atomic hydrogen desorbed at higher temperature
Peak II) [
6a, 11]
(
. The Co-Re interaction on SI samples results in the
3
According to NH -TPD results (Table 1), the amount of acid sites
advanced desorption of a small amount of dissociated hydrogen
at relative low temperature (Peak III) by weakening the connection
between atomic hydrogen and active metal. Further, the Co-Re
interaction has also a positive influence on the hydrogen
adsorption ability considering that the HAA value of SI samples is
significantly higher than monometallic and CI samples. As already
is very small and comparable on all studied catalysts indicating
that the C-C cleavages don’t originate from cracking process on
acid sites but rather from hydrogenolysis on metallic sites. It
suggests that the presence of reduced Re species on SI samples
caused by Co-Re interaction contributes to the hydrogenolysis of
cyclohexane to C1-C5 products. The low conversion obtained for
monometallic Re catalyst can be related to the high oxidation
state of Re on this sample, which is not favorable neither for the
dehydrogenation nor for the hydrogenolysis of cyclohexane. Thus,
it indicates that the nature of the active surface of these catalysts
2
discussed in H -TPR section, the Co-Re interaction would
increase the reduction efficiency of Re which is then beneficial to
hydrogen adsorption. The increase of Re content on SI samples
leads to the agglomeration of particles and reduces the hydrogen
adsorption ability, manifesting as the gradually decrease of HAA
value with the increasing of Re loading for SI-4Co-XRe/TiO
catalysts.
is in accordance with the results obtained by TEM and H
addition, the conversion to by-products with the CI-4Co-1Re/TiO
catalyst, which exhibits a low Re reduction efficiency, is lower than
that of SI-4Co-1Re/TiO
2
-TPR. In
2
2
By comparison with their SI counterparts, the SR-5Co-XRe/TiO
catalysts (with X = 2 and 5) exhibit superior hydrogen adsorption
2
2
.
Table 3. Catalytic performances for cyclohexane dehydrogenation and citral hydrogenation of monometallic Co, Re and bimetallic
SI-4Co-XRe/TiO and CI-4Co-1Re/TiO catalysts.
2
2
Cyclohexane dehydrogenation [a]
Citral hydrogenation [b]
Samples/TiO
2
Total
conversion (%)
Benzene (%)
-
By-products
(%) [
Conversion at
30 min (%)
UA selectivity
(%)
Maximum UA
yield (%)
c]
[d]
[e]
TiO
2
-
-
9.5
18.0
9.5
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
5
Co
2.5
4.6
5.2
7.3
7.9
8.9
3.8
1.1
0.4
1.6
2.8
3.4
4.4
4.0
3.9
3.0
0.9
0.2
0.9
1.8
1.8
2.9
3.9
5.0
0.8
0.2
0.2
9.0
28.0
41.5
45.0
46.4
46.1
45.0
42.0
37.2
20.0
5.5
SI-4Co-0.5Re
SI-4Co-1Re
SI-4Co-2Re
SI-4Co-3Re
SI-4Co-5Re
CI-4Co-1Re
46.0
72.0
57.0
93.0
99.3
13.8
8.8
47.0
44.0
48.0
38.0
40.0
19.0
12.1
12.2
4
Co + 1Re
Re
5
22.0
[
a] Reaction conditions: 300°C, 80 mg catalyst, 100 mL min-1
H
2
, 0.03 mL min-1 cyclohexane. [b] Reaction conditions: 70°C, 0.4 g
catalyst, 100 mL solution (3 mL citral), 75 bar H . [c] By-products: C1-C5 products issued from hydrogenolysis reactions. [d] UA
2
selectivity determined at 40 % conversion. [e] Maximum yield reached during the 180 min reaction time.
The product distribution of citral hydrogenation over SI series
samples is summarized in Figure
7 and Table 3. The
hydrogenation of citral, an unsaturated aldehyde, is a consecutive
reaction. Scheme 1 summarizes the main reaction pathways that
can occur during citral hydrogenation. The observed products are
(
i) unsaturated products, i.e. citronellol, citronellal, nerol and
geraniol (UA: unsaturated alcohols), and (ii) the saturated product
,7-dimethyloctanol.
3
Scheme 1. Main reaction pathways of citral hydrogenation over Co-Re catalysts.
Figure 7. Citral hydrogenation on monometallic Co, Re and SI bimetallic
catalysts. (Reaction conditions: 70°C, 0.4 g catalyst, 100 mL solution (3 mL
citral), 75 bar H )
2
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
transformed citral (mol)
reaction time (min) × metal content (mol)
(
6) EOM =
Although SI bimetallic samples with high Re loadings (3 and 5
wt.%) show superior conversion, their EOM value relative to the
actual total metal loadings seems quite comparable with that of
the SI-4Co-1Re catalyst, less loaded in Re (Figure 9(A)).
Moreover, when considering Re as major contributor in the
transformation of citral to calculate the EOM values (Figure 9(B)),
the SI-4Co-1Re sample finally appears as the most efficient of the
SI catalyst series. The progressive decrease in the efficiency of
SI-4Co-XRe catalysts as function of the increasing Re content can
be attributed to the partial agglomeration and lower reduction of
Re particles.
Figure 8. Total density of states (TDOS) and partial density of states (d-PDOS)
of Re (A) and Re O (B).
2 7
According to the results reported in Table 3, SI bimetallic catalysts
exhibit significantly higher activity than monometallic Co and Re
catalysts. For example, the conversion of citral after 30 min
reaction time is 72 % for SI-4Co-1Re/TiO
2
, largely higher than
2
(22 %) under same
those of 5Co/TiO (9 %) and 5Re/TiO
2
reaction conditions. The citral conversion tends to increase with
the Re content introduced, as previously observed for the
cyclohexane dehydrogenation model reaction (Figure 7(H)). This
confirms that the presence of accessible active sites increases
with the addition of Re by SI method so as to accelerate citral
transformation. This synergetic effect on activity for citral
hydrogenation observed for SI samples can result from the
influence of Co-Re interaction on the reaction and in particular
from the changes in the oxidation state of the active species. The
Janus-type structure of SI catalysts can contribute to the reduction
of Re species, increasing consequently the activity. In order to
further investigate this effect, we calculated the total density of
states (TDOS) and partial density of states related to d-state (d-
Figure 9. The efficiency of metal (EOM) for hydrogenation of citral on SI
bimetallic catalysts series.
Concerning the products distribution obtained during citral
hydrogenation, a similar selectivity to UA, around 45% for a 40%
conversion, is observed for all SI bimetallic catalysts, which is
consistent with the presence of the same Janus-type structure
and the same nature of active sites for all these systems.
PDOS) of a surface constituted of Re0 or Re
O atoms,
2 7
The influence of the Co-Re bimetallic interaction on catalytic
corresponding to the extreme valence states of Re species
participating in the adsorption process during the catalytic
reaction. The results in Figure 8 show that the adsorption capacity
and thus catalytic activity of Re-based components are decided
by d-state of Re element, because it makes major contribution to
TDOS. As far as we know, the TDOS and d-PDOS at Fermi level
performances can be evaluated by comparing the 4Co-1Re/TiO
systems prepared either by successive (SI) or co-impregnation
2
(
(
CI), or by a physical mixture of 4Co/TiO
2 2
and 1Re/TiO catalysts
Table 3). The CI-4Co-1Re/TiO sample displays poor
2
performances for the citral hydrogenation and cyclohexane
dehydrogenation, resulting from the presence of isolated Co and
Re entities and the absence of Co-Re interaction. During the
preparation of this sample, the absence of Co0 species at the
surface of the catalyst prevents the deposition of Re precursors
in contact with Co particles and thus the formation of Co-Re
interaction favoring the reduction of Re species. To evaluate the
influence of Co/Re ratio on contact degree in CI samples, the citral
F
(E ) would influence the adsorption and activation of active
[
13]
0
components on active metal . Re exhibits higher DOS at E
compared with Re for which states density is zero,
demonstrating that catalytic activity of Re is weakened by the
2 7
presence of oxygen. The significant band gap nearby E for Re O
F
2
O
7
2
O
7
F
is disadvantage for citral hydrogenation reaction when it is
regarded as main active sites. It further proves the attenuation of
catalytic activity for catalysts with high valence Re element.
hydrogenation was also investigated on CI-4Co-3Re/TiO
2
and CI-
0
4Co-5Re/TiO (Table S1). These catalysts exhibit significantly
2
Besides, the lack of electron density of Re shown previously by
lower activity than their SI counterparts, demonstrating that Co
and Re mainly exist in isolated structure on CI samples whatever
the Re content. The physical mixture of 4Co/TiO and 1Re/TiO
2 2
FT-IR of adsorbed CO on SI samples may be beneficial to the
adsorption of electron rich oxygen of aldehyde functions, which
would also enhance the activity.
catalysts exhibits even lower activities, demonstrating that close
contact between the metals is necessary to achieve a catalytic
promotion effect. Therefore, the SR catalysts with higher contact
degree between metals than SI samples were evaluated for the
hydrogenation of citral.
To effectively compare the activity of the catalysts of different
metal contents for citral hydrogenation, the efficiency of active
metal (EOM, with metal corresponding to Co and Re, or only Re)
at citral isoconversion (20 %) was calculated using the following
equation:
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
Table
4
compares the catalytic results of cyclohexane
dehydrogenation of BLANK-5Co/TiO
2
compared to the parent
dehydrogenation and citral hydrogenation obtained with the SR
bimetallic series to those displayed by the Co and Re
monometallic samples. The lower activity in cyclohexane
5Co/TiO results from the decrease of active sites caused by acid
treatment (partial loss of Co by leaching as observed by ICP-OES,
from 5.3 wt.% to 3.6 wt.%).
2
Table 4. Catalytic performances for cyclohexane dehydrogenation and citral hydrogenation of monometallic Co, Re and bimetallic
SR-5Co-XRe/TiO
2
catalysts.
Cyclohexane dehydrogenation [a]
Citral hydrogenation [b]
Samples/TiO
Co
2
Total
conversion (%)
Benzene (%)
By-products
(%) [
Conversion at
30 min (%)
UA selectivity
(%)
Maximum UA
yield (%)
c]
[d]
[e]
5
2.5
1.3
4.4
5.0
8.4
0.4
1.6
0.8
3.4
3.6
3.8
0.2
0.9
0.5
1.0
1.4
4.6
0.2
9.0
28.0
35.1
48.0
44.6
52.0
20.0
5.5
BLANK-5Co [f]
SR-5Co-1Re
SR-5Co-2Re
SR-5Co-5Re
11.4
16.2
65.1
84.0
22.0
7.9
12.1
44.6
51.0
12.2
5Re
[
a] Reaction conditions: 300°C, 80 mg catalyst, 100 mL min-1
catalyst, 100 mL solution (3 mL citral), 75 bar H . [c] By-products: C1-C5 products issued from hydrogenolysis reactions. [d] UA
selectivity determined at 40 % conversion. [e] Maximum yield reached during the 180 min reaction time. [f] BLANK-5Co/TiO : 5
wt.%Co/TiO treated by HCl solution (pH 1) under N atmosphere
H
2
, 0.03 mL min-1 cyclohexane. [b] Reaction conditions: 70°C, 0.4 g
2
2
2
2
2
of these C1-C5 products. Especially for SR-5Co-5Re/TiO with a
core(Co)-shell(Re) structure exhibits highest selectivity to C1-C5
products, which confirm the influence of reduced Re on
cyclohexane dehydrogenation. Furthermore, the SR catalysts
exhibit a noticeable promotion effect on the citral conversion
compared with monometallic Co and Re samples (Table 4 and
Figure 10). The increase of Re content has a positive influence
on the amount of accessible active sites and then is beneficial to
the citral conversion as illustrated by Figure 10(D). The FT-IR
spectra of adsorbed CO over SR samples have proven the
strengthening of the Co-Re interaction with the increasing of Re
2
content (Figure 5). As anticipated, the SR-5Co-5Re/TiO catalyst
exhibits higher EOM than other SR samples when considering
both Re and Co contributions (Figure 11(A)). But as highlighted
by the H
SR-5Co-5Re/TiO
2
-TPR analysis, the core-shell structure displayed by the
catalyst reduces the reduction efficiency of Re
2
species compared to other SR bimetallic catalysts exhibiting a
decoration structure. This phenomenon is consistent with the
EOM values related to the lone Re contribution (Figure 11(B))
showing a superior efficiency for the SR-5Co-2Re/TiO
2
sample
Figure 10. Citral hydrogenation on SR bimetallic catalysts. (Reaction
conditions: 70°C, 0.4 g catalyst, 100 mL solution (3 mL citral), 75 bar H
compared to SR-5Co-5Re/TiO . The enrichment of Re on the
2
2
)
surface of catalyst, as the main contributor to citral hydrogenation,
is an advantage for the production of UA since a small amount of
rise in UA selectivity is observed for the SR-5Co-5Re/TiO
catalyst (Table 4).
2
As with the SI bimetallic series, the percentage of by-products
formed during the cyclohexane dehydrogenation also gradually
becomes predominant for SR bimetallic catalysts when increasing
Re loading. In the same way as in the SI catalysts, the reduced
Re species in SR samples play an important role in the production
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
SR-5Co-1Re
SR-5Co-2Re
SR-5Co-5Re
1.4
21
35
1.1
20
22
3
2
6
3
32
90
1 2
As shown in Table 5, the hydrogenation rate of citral (k +k ) for
bimetallic catalysts is up to 40-100 times higher than for
monometallic catalysts and the CI sample, proving that Co-Re
interaction can obviously improve the catalytic activity. The
Figure 11. The efficiency of metal (EOM) for hydrogenation of citral on SR
bimetallic catalysts series.
4
reaction rate for Step 4 (k ) is significantly higher than for the other
steps indicating that citronellal can be quickly transformed to
citronellol on Co-Re bimetallic catalysts. Besides, the
A kinetic study was performed by fitting catalytic data and using a
first order model in order to determine the hydrogenation kinetic
constant values to understand reaction mechanism (Figure S4).
In accordance with the results described in the present
manuscript, the reaction pathway was shown in Scheme 1. The
yield in saturated product (Step 5) is very small; thus the kinetics
of citral hydrogenation can be represented by the following
equations:
hydrogenation rate to citronellol from UA products (k
than that from citronellal (k ). It leads to the conclusion that
citronellol is mainly produced from citronellal at initial stage of
reaction. Considering the limited reaction rate of k , the yield of
UA is mainly decided by the ratio of k /k . The k /k ratio for most
3
) is slower
4
3
1
2
1
2
bimetallic catalysts is nearly equal to 1, which explains the fact
that the yield of UA is around 50 % for all bimetallic catalysts in
our study.
dCꢀꢁꢂ
(
(
(
(
7) rCAL = −
= k × ꢃ
+ k × ꢃ
CAL 2 CAL
1
dt
dCꢄꢁ
dt
8) rUA
=
= k × ꢃ
− k × ꢃ
CAL 3 UA
1
dCꢀꢅꢁꢂ
9) rCNAL
=
= k × ꢃ
2
CAL
− k × ꢃ
4
CNAL
dt
dCꢀꢅꢆꢂ
= k × ꢃ + k × ꢃ
3 UA 4 CNAL
dt
10) rCNOL
=
The differential equations were solved by using the Runge–Kutta–
Merson method. The model parameter estimation was performed
by non-linear regression using a Universal Global Optimization
*
2
method, which minimizes the function: Q=∑(ꢃ -ꢃ ) , where C
i,j i,j
and C* are the experimental and calculated concentrations, i is
the chemical compound, and j is the reaction time. In any case,
the kinetic parameters were statistically significant with
confidence above 98 %.
a
Figure 12. Reaction mechanism over bimetallic entities with various Co-Re
interactions.
Table 5. Kinetic constant values of Co-Re catalysts for citral and products
formation/hydrogenation.
-
3
-3
-3
-3
k
(
1
×10
k
2
×10
k
3
×10
k
4
×10
Figure 12 proposes reaction mechanism of Co-Re bimetallic
catalysts with different degree of interaction. For CI samples, the
citral hydrogenation reaction is carried out on Co species or Re
species separately. The very limited Co-Re interaction on CI
samples not only weakens the electronic transfer between Co and
Re, but also decreases the reduction efficiency of Re species
resulting in a low reduction degree of Re. As a result, CI samples
exhibit low hydrogen activation ability and catalytic activity. The
Janus-type structure of SI samples leads to the adsorption of citral
on both Co and Re species. FT-IR analysis indicates the
electronic transfer from Re to Co, which is beneficial to the
adsorption of electron rich oxygen of aldehyde function on Re
Samples/TiO
2
-
1
-1
-1
-1
min )
(min )
(min )
(min )
5
Co
Re
0.3
0.5
1.4
10
1
0
7
5
4.4
1.2
11
23
48
53
0
15
10
30
80
80
130
CI-4Co-1Re
0
SI-4Co-0.5Re
SI-4Co-1Re
SI-4Co-3Re
SI-4Co-5Re
0
28
9
47
10
12
52
species. H
2
-TPD results show that reduced Re species is more
favorable to the activation of hydrogen than Co. Therefore,
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
hydrogen is mainly activated on reduced Re species on SI
samples and then participates to the hydrogenation of aldehyde
group. Meanwhile, a small amount of hydrogen is also activated
on Co species, and consequently it will assist Re in citral
hydrogenation. As described above, the accessible active sites
are Co and Re for both CI and SI samples. On the contrary, the
main accessible active sites are Re species on SR samples,
especially at high Re loading with a change of structure from
decoration to core(Co)-shell(Re). FT-IR results demonstrate
obvious electronic transfer on SR samples. For this reason, the
adsorption of aldehyde groups mainly happens on Re species and
hydrogen is also mainly activated on reduced Re. In conclusion,
the reaction mechanism is different for CI, SI and SR samples and
significantly influenced by the Co-Re contact degree.
(SI) and surface redox reaction (SR). The resulting samples are
distinguished from each other by their metal arrangement, which
is directly related to the preparation method used. The basic
difference for these samples is the contact degree between Co
and Re clearly observed by TEM-EDX analysis. The CI sample
presents a very limited bimetallic interaction since the particles of
Re and Co are completely separated on the surface of the catalyst.
The SI samples exist in a Janus-type structure because of the pre-
existing metallic Co species that can contribute to the
concentration of Re around Co, resulting in direct contact or at
least close contact between Re and Co. Decoration/Core-shell
structures are observed for SR samples due to redox reaction
0
between Co atoms of the pre-reduced Co parent catalyst and
Re7+ ions of the modifier precursor salt. Thus, the degree of
contact between the two metals gradually increases in the
following order: CI sample < SI samples < SR samples.
Table 6. Citral hydrogenation over different supported noble metal-based
catalysts.
Concerning the hydrogenation of citral, SR and SI samples show
higher activity than CI sample and a noticeable synergetic effect
compared to Co and Re monometallic catalysts. This implies that
the presence of Co-Re interaction plays a predominant role in
citral transformation. In addition, the bimetallic interaction allows
to increase the reduction efficiency of Re precursors (as revealed
Reaction conditions
Metallic phase
Main product
of the catalysts
Temperature
Pressure
Time
6 h
Pd [14]
Citronellal
85 % yield)
by H -TPR analysis), which is beneficial for the hydrogenation
80-120 °C
0.6 MPa
2
(
reaction according to DOS calculation. Moreover, this bimetallic
Citronellal
80 % yield)
interaction in the SR and SI catalysts results in an electronic
transfer from Re to Co, as highlighted by in situ CO-FTIR analysis,
which is beneficial for the adsorption of electron rich oxygen of
aldehyde functions and thus beneficial for the selective
hydrogenation of citral. The in situ CO-FTIR also proves that this
electron transfer between Co and Re is more significant on SR
samples than on SI ones, in agreement with the higher degree of
Pd-Ni [3e]
100 °C
130 °C
1.0 MPa
7.0 MPa
3 h
(
UA
Pd-Sn [3a]
1 h
(80 % selectivity
at 30 % conv.)
UA
Pt [15]
70 °C
70 °C
70 °C
70 °C
1.0 MPa
1.0 MPa
0.1 MPa
7.0 MPa
12 h
(60 % yield)
2
metal interaction for SR catalysts. Furthermore, based on H -TPD
Pt-Co [16]
UA
65 % yield)
6 h
results, Co-Re bimetallic catalysts prepared by SI and SR
methods possess more prominent hydrogen adsorption ability
(HAA) than monometallic and CI catalysts. It should also be noted
(
Pt-Fe/In/Ga [3d]
0.1-
UA
20 h
(90 % selectivity)
that for similar metal contents, the HAA of SR-5Co-2Re/TiO
2
(1.9)
UA
Rh-Ge [3b,17]
is slightly higher than that of SI-4Co-1Re/TiO (1.7).
2
2 h
(70 % yield)
Finally, although SR samples have more significant degree of
contact between the two metals and greater bimetallic interaction
than SI samples, quite similar performances in citral
hydrogenation are observed on both catalysts series. For
example, the citral conversion (at 30 min) is 72.0 % for SI-4Co-
1Re/TiO and 65.1 % for SR-5Co-2Re/TiO , and the max yield of
2 2
UA is around 44.0 % for both of them. Moreover, once Co-Re
interaction has formed, it seems that excessive contact degree is
The objective of this study was to replace noble metals for
selective hydrogenation reactions. To the best of our knowledge,
it is the first time that bimetallic catalysts without noble metal have
been studied for citral hydrogenation. A very interesting synergic
effect between Co and Re has been highlighted. Nevertheless,
compared to noble metal-based catalysts, reported in Table 6, for
which up to 90 % of UA selectivity can be obtained, Co-Re
systems should be further optimized by varying other parameters
such as Co particle size, reduction temperature and so on.
However, these systems are stable in the reaction medium as
demonstrated by recyclability test (Figure S5).
unnecessary for citral hydrogenation. The similar values of k
in kinetic study also support this conclusion. In fact, an
excessive contact degree such as in SR-5Co-5Re/TiO catalyst
1
and
k
2
2
limits the reduction efficiency of Re due to the lack of directly
accessible Co atoms, and then leads to a limited EOM (efficiency
of metal) of Re during citral hydrogenation. Likewise, according to
2
the H -TPR analysis, the SI-4Co-5Re sample exhibits lower
reduction efficiency than other SI catalysts due to an aggregation
of Re particles, which results in a decrease in EOM of Re. This
undesirable effect of an excessive Re loading on the SI and SR
General Discussion and Conclusions
2
series is also highlighted by H -TPD results, which show that the
2
In this study, Co-Re/TiO bimetallic catalysts were prepared by
three methods: co-impregnation (CI), successive impregnation
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
HAA is compromised for the bimetallic catalysts displaying the
higher Re contents.
thermal conductivity detector (TCD). The procedure for TPR analysis was
as followed: the sample was heated in 10 % H
2
/Ar at a flow rate of 30 mL
min-1 from room temperature to 700 °C with a temperature ramping rate of
-
1
5
°C min , and then maintained at 700 °C for 30 min.
Experimental Section
Temperature-programmed desorption (TPD) was carried out on
AutoChem 2910 instrument (Quantachrome, USA). The samples were
loaded into a fritted quartz tube, pretreated at 500 °C for 120 min under
Catalysts preparation.
1
0 % H
atmosphere. H
with a flow rate of 100 mL. After H
with N for 90 min to remove physical adsorbed species. The H
2
/Ar atmosphere and afterwards cooled down to 50 °C under N
adsorption was performed at 50 °C in 10 % H /Ar mixture
adsorption, the samples were purged
desorption
atmosphere with a temperature ramp of 5 °C
2
All the catalysts were prepared by using TiO
2
(from Alfa Aesar) as support,
2
2
and CoCl and NH ReO (Alfa Aesar) as precursor salts.
2
4
4
2
2
2
was then conducted under N
min-1 from 50 to 800 °C.
2
Monometallic catalysts
The monometallic catalysts were prepared by impregnation in excess of
The Fourier transform infrared (FT-IR) spectra of CO adsorbed over Co-
Re catalysts were collected on a Bruker EQUINOX55 spectrometer. The
sample was pressed into a self-supporting disk and placed in an IR cell
attached to a closed glass-circulation system. Prior to CO adsorption, the
solution of CoCl
dried at 90 °C for 24 h and reduced in H
2 4 4 2
or NH ReO on TiO . The impregnated support was then
at 500 °C for 2 h (2 °C min-1
2
heating rate). These Co or Re monometallic catalysts are referred as 5Co
TiO or 5Re/TiO (with 5 wt.% as expected weight metal loading).
/
2
2
2
sample disk was in situ pre-treated at 350 °C for 2 h under H and then
cooled down to 30 °C. After evacuation of hydrogen by vacuum pump, the
sample disk was exposed to CO for 30 min. The IR spectrum of the
chemisorbed CO was recorded after physically adsorbed CO was
evacuated.
Bimetallic catalysts
Firstly,
impregnation (SI) method. The NH
various concentrations on monometallic 4Co/TiO
dried at 90 °C for 24 h. Finally, they were reduced in H
2 °C min-1 heating rate). These Co-Re bimetallic catalysts are referred as
SI-4Co-YRe/TiO (Y represents the Re expected weight loading). A Co-Re
bimetallic sample was also prepared by co-impregnation (CI) method. In
this case, both CoCl and NH ReO precursor salts were impregnated
simultaneously on TiO , and subsequent steps were the same as
described above. The corresponding catalyst is referred as CI-4Co-
Re/TiO (4 wt.% and 1 wt.% being the respective expected weight metal
loadings).
a
series of Co-Re catalysts was prepared by successive
ReO salt was impregnated with
. The catalysts were then
at 500 °C for 2 h
4
4
2
Experiments of temperature programmed desorption of ammonia (NH
3
-
2
TPD) were performed on samples outgassed in He at 200 °C for 1 h, and
(
-
1
then saturated at 50 °C in a 5.62 % NH
After removing most weakly physisorbed NH
the chemisorbed NH quantity was analyzed with TCD detector by heating
3
/He stream (30 mL min ) for 1 h.
2
-
1
3
by flowing He (30 mL min ),
3
2
4
4
at 10 °C min-1 up to 800 °C and by maintaining this temperature for 30 min
under the same flow of He.
2
1
2
Transmission electron microscopy (TEM) images were performed on a FEI
Technai F20 electron microscope. The elemental mapping was achieved
with energy dispersive X-ray spectroscopy (EDX) using a DX4 analyzer
system. The reduced catalysts were ultrasonically dispersed in ethanol
and the suspension was brought onto a copper grid with carbon film. The
average diameters were calculated by measuring over 100 particles from
various TEM images.
Another series of bimetallic catalysts was also prepared by a surface redox
SR) method. This method consists of a direct surface redox reaction
between the Re precursor salt and Co particles of the parent monometallic
catalyst. The parent 5Co/TiO sample was placed into a fixed-bed reactor
and re-activated at 500 °C under H . Then, the NH ReO aqueous solution
acidified to pH 1 with HCl), previously degassed under nitrogen flow, was
(
2
2
4
4
(
The model reaction of cyclohexane dehydrogenation was performed in a
fixed-bed reactor at atmospheric pressure. A mass of 80 mg of catalyst
was reactivated by reduction at 500 °C for 2 h and the temperature was
decreased to 300 °C to perform the reaction. Cyclohexane was injected
2 2
poured onto the Co/TiO catalyst. After 1 h reaction time under N bubbling
at room temperature, the solution was filtered out and the catalyst was
dried overnight at 100 °C. Finally, the bimetallic catalysts were reduced
flow at 500 °C for 2 h (2 °C min-1 heating rate). These Co-Re
(Y represents the Re
under H
2
-
1
using a calibrated motor-driven syringe (0.03 mL min ) in a hydrogen flow
bimetallic catalysts are referred as SR-5Co-YRe/TiO
2
-
1
(
100 mL min ). Cyclohexane and the products were analyzed every 10
expected weight metal loadings).
min by GC.
Catalysts characterization.
Citral hydrogenation.
The metallic loadings of monometallic and bimetallic catalysts were
analyzed by inductively coupled plasma optical emission spectroscopy
The hydrogenation of citral was carried out in a 250 mL autoclave with a
magnetic stirrer and temperature controller unit. The catalysts (400 mg)
(ICP-OES) equipment (Perkin Elmer Optima 2000 DV) after dissolution of
were previously treated with H
catalysts were immersed in 90 mL of isopropanol used as solvent and
transferred in situ into the autoclave under the protection of N atmosphere
by using a glove box. The batch reactor was purged first with N and then
with H prior to raise the temperature to 70 °C. Then, a mixture of 3 mL
citral and 10 mL of isopropanol was loaded into the reactor through a
cylinder under 75 bar H pressure. Zero time was taken at this moment.
2
for 2 h at 500 °C. After reduction, the
the solid samples.
2
Powder X-ray diffraction (XRD) was used to probe metal interaction in the
bimetallic catalysts and to provide information regarding crystal phases.
Powder diffraction was performed on a Bruker D5005 X-ray diffractometer
with a CoKα radiation source (λ = 1.54178 Å) and a scan speed of 0.1 °
2
2
-
1
2
min .
During the catalytic test, the reaction was performed under constant
pressure by using a pressure control system. At different reaction times,
Temperature-programmed reduction (TPR) analyses were performed on
an AutoChem 2910 instrument (Micromeritics, USA) equipped with a
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
liquid samples were manually collected and analyzed by gas
chromatograph to determine the conversion and selectivity values.
[5]
a) T. N. Phaahlamohlaka, D. O. Kumi, M. W. Dlamini, R. Forbes, L. L.
Jewell, D. G. Billing, N. J. Coville, ACS Catal. 2017, 7, 1568-1578; b) S.
K. Beaumont, S. Alayoglu, C. Specht, N. Kruse, G. A. Somorjai, Nano
Lett. 2014, 14, 4792-4796; c) M. Hosseini, T. Barakat, R. Cousin, A.
Aboukaïs, B. L. Su, G. De Weireld, S. Siffert, Appl. Catal., B 2012, 111-
112, 218-224; d) D. Nabaho, J. W. Niemantsverdriet, M. Claeys, E. van
Steen, Catal. Today 2016, 261, 17-27; e) S. Alayoglu, B. Eichhorn, J. Am.
Chem. Soc. 2008, 130, 17479-17486.
Acknowledgements
[
[
6]
7]
a) R. Prins, Chem. Rev. 2012, 112, 2714-2738; b) V. V. Rozanov, O. V.
Krylov, Russ. Chem. Rev. 1997, 66, 107-119.
We gratefully acknowledge the financial support provided by the
National Natural Science Foundation of China (21573031),
Program for Excellent Talents in Dalian City (2016RD09), the
European communities (FEDER) and the “Région Nouvelle
Aquitaine”, and the Project Partenariats Hubert Curien (PHC) Cai
Yuanpei supported by the French Embassy in China and the
China Scholarship Council.
a) B. K. Ly, B. Tapin, M. Aouine, P. Delichere, F. Epron, C. Pinel, C.
Especel, M. Besson, ChemCatChem 2015, 7, 2161-2178; b) L. Corbel-
Demailly, B. K. Ly, D. P. Minh, B. Tapin, C. Especel, F. Epron, A. Cabiac,
E. Guillon, M. Besson, C. Pinel, ChemSusChem 2013, 6, 2388-2395.
a) S. Storsater, O. Borg, E. Blekkan, A. Holmen, J. Catal. 2005, 231, 405-
419; b) A. M. Hilmen, D. Schanke, K. F. Hanssen, A. Holmen, Appl. Catal.,
A 1999, 186, 169-188.
[
[
8]
9]
a) H. Iida, A. Igarashi, Appl. Catal., A 2006, 303, 48-55; b) C. Bolivar, H.
Charcosset, R. Frety, M. Primet, L. Tournayan, C. Betizeau, G. Leclercq,
R. Maurel, J. Catal. 1976, 45, 163-178; c) Z. Wei, A. Karim, Y. Li, Y.
Wang, ACS Catal. 2015, 5, 7312-7320.
Keywords: Co-Re bimetallic catalysts • Janus-type structure •
Decoration/Core-shell structure • Citral hydrogenation
[
10] a) Y. Sato, Y. Soma, T. Miyao, S. Naito, Appl. Catal., A 2006, 304, 78-
85; b) W. Daniell, J. Mol. Catal. A: Chem. 2003, 204-205, 519-526.
[
[
1]
2]
E. Bailón-García, F. Carrasco-Marín, A. F. Pérez-Cadenas, F. J.
Maldonado-Hódar, Appl. Catal., A 2016, 512, 63-73.
[11] a) W. Rachmady, M. A. Vannice, J. Catal. 2000, 192, 322-334; b) D. S.
Park, D. Yun, T. Y. Kim, J. Baek, Y. S. Yun, J. Yi, ChemSusChem 2013,
6, 2281-2289.
a) P. Meric, K. Yu, A. Kong, S. Tsang, J. Catal. 2006, 237, 330-336; b)
U. K. Singh, M. A. Vannice, J. Catal. 2001, 199, 73-84; c) M. S. Ide, B.
Hao, M. Neurock, R. J. Davis, ACS Catal. 2012, 2, 671-683.
[12] N. Hérault, L. Olivet, L. Pirault-Roy, C. Especel, M. A. Vicerich, C. L.
Pieck, F. Epron, Appl. Catal., A 2016, 517, 81-90.
[
[
3]
4]
a) A. Vicente, G. Lafaye, C. Especel, P. Marécot, C. T. Williams, J. Catal.
2011, 283, 133-142; b) T. Ekou, A. Vicente, G. Lafaye, C. Especel, P.
[13] a) Y. Ma, X. Wu, G. Zhang, Appl. Catal., B 2017, 205, 262-270; b) O. R.
Inderwildi, S. J. Jenkins, D. A. King, Surf. Sci. 2007, 601, 103-108.
[14] S. Yilmaz, S. Ucar, L. Artok, H. Gulec, Appl. Catal., A 2005, 287, 261-
266.
Marecot, Appl. Catal., A 2006, 314, 64-72; c) C. Especel, D. Duprez, F.
Epron, C. R. Chim. 2014, 17, 790-800; d) J. P. Stassi, V. I. Rodríguez,
M. J. Yañez, S. R. de Miguel, P. D. Zgolicz, Catal. Lett. 2017, 147, 1903-
1921; e) C. Liu, C. Nan, G. Fan, L. Yang, F. Li, Mol. Catal. 2017, 436,
[15] X. Wang, W. Hu, B. Deng, X. Liang, J. Nanopart. Res. 2017, 19, 153.
[16] N. M. Bertero, A. F. Trasarti, B. Moraweck, A. Borgna, A. J. Marchi, Appl.
Catal., A 2009, 358, 32-41.
237-247.
a) D. D. Falcone, J. H. Hack, R. J. Davis, ChemCatChem 2016, 8, 1074-
1
083; b) Z. Wei, A. M. Karim, Y. Li, D. L. King, Y. Wang, J. Catal. 2015,
[17] G. Lafaye, T. Ekou, C. Micheaud-Especel, C. Montassier, P. Marecot,
Appl. Catal., A 2004, 257, 107-117.
322, 49-59; c) X. Di, C. Li, B. Zhang, J. Qi, W. Li, D. Su, C. Liang, Ind.
Eng. Chem. Res. 2017, 56, 4672-4683; d) X. Di, C. Li, G. Lafaye, C.
Especel, F. Epron, C. Liang, Catal Sci Technol 2017, 7, 5212-5223.
This article is protected by copyright. All rights reserved.

ChemSusChem
10.1002/cssc.201802744
FULL PAPER
Entry for the Table of Contents
FULL PAPER
Xin Di, Gwendoline Lafaye *, Catherine
Especel, Florence Epron, Ji Qi, Chuang
Li, Changhai Liang*
Page No. – Page No.
Supported Co-Re bimetallic catalysts
with different structures as efficient
catalysts for hydrogenation of citral
The influence of contact degree of Co and Re on Co-Re interaction and its
application in hydrogenation of citral was investigated.
This article is protected by copyright. All rights reserved.