Catalysts based on Rh(III)-hexamolybdate/γ-Al2O 3 and their application in the selective hydrogenation of cinnamaldehyde to hydrocinnamaldehyde

Article abstract of DOI:10.1016/j.molcata.2012.09.013

Several catalysts based on [RhMo₆O₂₄H ₆]^3- (RhMo₆) heteropolyanion supported on γ-Al₂O₃ with different textural properties were prepared, characterized and tested in the liquid-phase hydrogenation of cinnamaldehyde. The characterization of pure and supported systems was carried out using several techniques including XRD, SEM-EDS microscopy, Raman microprobe, X-ray photoelectron spectroscopies and temperature programmed reduction (TPR). The catalytic performance was monitored by conversion of the starting cinnamaldehyde as a function of time: the initial activities represented as the turnover frequency (TOF), were measured considering the surface Rh atoms. The planar heteropolyanion RhMo₆/(EI) based systems showed enhanced catalytic activity over the RhMo conventional catalyst obtained by successive impregnation of both transition metal ions (Rh(III) chloride and heptamolybdate). The selectivity for the RhMo₆ systems was mainly toward HCAL unlike the conventional catalyst which showed selectivity toward CA. This study also showed a synergetic effect between Rh and Mo through which Rh promoted Mo reducibility.

Full text of DOI:10.1016/j.molcata.2012.09.013

Journal of Molecular Catalysis A: Chemical 366 (2013) 109–115
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Catalysts based on Rh(III)-hexamolybdate/␥-Al2O3 and their application in the
selective hydrogenation of cinnamaldehyde to hydrocinnamaldehyde
Guillermo R. Bertolinia, Carmen I. Cabelloa,b,∗, Mercedes Mun˜oza, Mónica Casellaa, Delia Gazzolic,
Ida Pettitic, Giovanni Ferrarisc
a “Centro de Investigación y Desarrollo en Ciencias Aplicadas Dr. J. Ronco” (CINDECA-CCT CONICET La Plata – Universidad Nacional de La Plata), Calle 47 No. 257, 1900 La Plata,
Argentina
b Facultad de Ingeniería UNLP and Member CICPBA, Argentina
c “Dipartimento di Chimica” La Sapienza Università di Roma, Piazzale Aldo Moro 5, I-00185 Roma, Italy
a r t i c l e i n f o
a b s t r a c t
Several catalysts based on [RhMo6O24H6]3− (RhMo6) heteropolyanion supported on ␥-Al2O3 with dif-
ferent textural properties were prepared, characterized and tested in the liquid-phase hydrogenation of
cinnamaldehyde. The characterization of pure and supported systems was carried out using several tech-
niques including XRD, SEM-EDS microscopy, Raman microprobe, X-ray photoelectron spectroscopies and
temperature programmed reduction (TPR). The catalytic performance was monitored by conversion of
the starting cinnamaldehyde as a function of time: the initial activities represented as the turnover fre-
quency (TOF), were measured considering the surface Rh atoms. The planar heteropolyanion RhMo6/(EI)
based systems showed enhanced catalytic activity over the RhMo conventional catalyst obtained by suc-
cessive impregnation of both transition metal ions (Rh(III) chloride and heptamolybdate). The selectivity
for the RhMo6 systems was mainly toward HCAL unlike the conventional catalyst which showed selectiv-
ity toward CA. This study also showed a synergetic effect between Rh and Mo through which Rh promoted
Mo reducibility.
Article history:
Received 14 December 2011
Received in revised form
13 September 2012
Accepted 14 September 2012
Available online 25 September 2012
Keywords:
Heteropolyoxomolybdates
Anderson phases of Rh(III)
RhMo6/␥-Al2O3
Cinnamaldehyde
Selective hydrogenation
Hydrocinnamaldehyde
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
ketones. An example is the selective hydrogenation of cinnamalde-
hyde (CAL) [9–12]. This species is a vinylic derived of benzaldehyde
The selective hydrogenation of ␣,␤-unsaturated aldehydes
assisted by supported metal catalysts is a key stage in the prepa-
ration of pharmaceutical products, flavors and fragrances [1–3].
The selective hydrogenation of cinnamaldehyde (CAL) is a good
example of this commercially important reaction that allows the
attainment of cinnamic alcohol (CA), hydrocinnamaldehyde (HCAL)
and 3-phenyl propanol (PP). The cinnamic alcohol is a valuable
product in the perfume industry for its aroma as well as for its fixa-
tion properties. Also, it is used in the pharmaceutical industry in the
synthesis of chloromycetin antibiotic [4,5]. More recently, impor-
tant applications were found for the hydrocinnamaldehyde used as
intermediate in the synthesis of medicines for the treatment of HIV
[6–8].
and the aldehyde group behaves with catalyst as if it were an aro-
matic aldehyde. The carbonyl group is reduced easily and competes
with saturation of the C C double bond [13]. The hydrogenation of
CAL catalyzed by supported noble metals leads usually to a mix-
ture of HCAL, CA and PP. The selectivity to CA is highly dependent
on the nature of the precious metal used as catalyst [14], and the
noble metals selectivity can be classified in the following sequence:
Pd < Rh < Ru < Pt < Ir. Pd and Rh generally display high activity but
rather poor selectivity toward cinnamyl alcohol and during the past
few years, several examples of synthesis and properties of mono
or bi metallic Pd catalysts for selective hydrogenation of CAL to
HCAL have been reported [15–17]. Little has been published con-
cerning the use of rhodium for the selective hydrogenation of CAL
to HCAL. Additionally, it was found that the catalysts deactivation
rate in monometallic systems is significant and hence a low stability
derives [18]. Therefore, in the hydrogenation field, bimetallic cata-
lysts are often used in order to improve selectivity and stability of a
single component catalyst [19,20]. In general, supported bimetal-
lic catalysts are very interesting materials because one metal can
tune and/or modify the catalytic properties of the other metal as
a result of both electronic and structural effects [20,21]. In this
The hydrogenation process is a complex one when the molecule
to be hydrogenated has two or more functional groups of compara-
ble reactivity. This situation is present in molecules that have C
C
and C O conjugated groups such as ␣,␤-unsaturated aldehydes or
∗
Corresponding author: Tel.: +54 221 4210711; fax: +54 221 4210711.
E-mail address: ccabello@quimica.unlp.edu.ar (C.I. Cabello).
1381-1169/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2012.09.013

110
G.R. Bertolini et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 109–115
sense, bi- or trimetallic catalytic systems based on heteropolyacids
or polymetalates (HPOMs), both as bulk and supported systems,
have started to replace a large number of conventional oxide pre-
cursors, improving the efficiency and environmental conditions of
several catalytic processes [22].
2.3. Characterization of catalysts
X-ray powder diffraction patterns were recorded in the 2ꢀ angu-
lar range from 5◦ to 70◦ by a diffractometer Philips PW 1714 using
Cu K␣ radiation (Ni filter).
Our recent studies have shown that Co, Ni or Rh containing
heteropolymolybdates with Anderson-type structure are inter-
esting precursors in heterogeneous catalysts for hydrotreatment
processes [23,24]. Such studies have shown that the planar struc-
ture of the heteropolyoxoanion as well as its redox and solubility
properties are relevant factors in the heteropolyanion-support
interaction, producing an active surface with an ordered distribu-
tion and uniform deposition of the metallic elements, which favors
the synergic effect.
In continuation of our studies of these phases as bimetallic cat-
alytic precursors, in the present paper we have studied the selective
hydrogenation of CAL by using catalysts based on Anderson-type
heteropolyoxometalates with formula (NH4)3[RhMo6O24H6]·7H2O
supported on ␥-Al2O3. The RhMo6/␥-Al2O3 systems resulted active
and selective for the hydrogenation of CAL to HCAL and their perfor-
mance were compared with those of a conventional RhMo/␥-Al2O3
catalyst, prepared by successive impregnation.
Scanning electron microscopy (SEM) with EDS analysis was per-
formed by a Microscope Philips SEM 505 with dispersive energy
system for microanalysis, EDAX 9100.
Raman spectra were collected on powder samples at room tem-
perature in the back-scattering geometry with an in Via Renishaw
spectrometer equipped with an air-cooled CCD detector and a
super-Notch filter. The emission line at 514.5 nm from an Ar+ ion
laser was focused on the sample under a Leica DLML microscope
using a 20× objective. Five 20-s accumulations were attained for
each sample with an incident beam power of about 5 mW. The spec-
tral resolution was 2 cm−1 and the spectra were calibrated using the
520.5 cm−1 line of a silicon wafer.
XPS spectra were measured on a Leybold–Heraeus LHS10 spec-
trometer in FAT mode (50 eV pass energy) with Mg K␣ radiation
(12 kV, 20 mA) at a pressure lower than 10−9 Torr. The samples,
manually pressed onto a gold plate, were analyzed as prepared
and after treatment in H2 outside the spectrometer. The Rh3d,
Mo3d, O1s, C1s and Al2p regions were processed by computer.
Electron binding energies were referenced to the C1s peak at
285.0 eV. The intensity ratios, IMe/IAl, (Me = Rh or Mo) were obtained
from peak area determination by integration of the appropri-
ate peak after smoothing and nonlinear Shirley-type background
subtraction by Esca Tools 4.2 software (Surface Interface, Inc.,
Mountain View, CA). To assess a quantitative relationship between
the XPS peak intensity ratio and surface composition, the exper-
imental results were compared with the values predicted by the
sensitivity factor, S, approach firstly proposed by Wagner et al.
[26].
Surface area measurements, Brunauer–Emmett–Teller (BET)
[27] multipoint method, and textural analysis were obtained using
a Micromeritics ASAP 2020 equipment. The samples were pre-
treated under vacuum in two stages of 1 h at 373 and 573 K. The
pore distribution was determined by the Barret–Joyner–Halenda
(BJH) method [28] from the analysis of micropore isotherm by the
t-test [29] taking the curve of Harkins and Jura [30]; the total pore
volume was determined by the rule of Gurvitsch [31].
The temperature programmed reduction (TPR) diagrams were
obtained in a flow system with a mixture of 10% volume H2 and
90% N2 volume (20 cm3/min), by heating from room temperature
up 1123 K at 10 K/min with a Quantasorb Jr. (Quantachrome) equip-
ment.
Measurements of Rh dispersion were made following the
Boudart method: Rh dispersion, defined as the ratio of the exposed
Rh atoms to the number of total Rh atoms, was determined by H2/O2
titration [32] using a volumetric apparatus. The fresh samples (typ-
ically 0.5–1 g, depending on Rh content) were pretreated in flowing
N2 at 423 K for 30 min, reduced in H2 (10% in He) at 623 K for 1 h and
then purged with N2 at 623 K for 1 h before cooling down to RT in
N2. Samples were then exposed to O2 (10% in N2) at RT for 30 min,
and flushed with N2 for 30 min at RT. After careful evacuation at RT
down to 1 Pa, the irreversibly adsorbed H2 at RT, Hirr, was deter-
mined. In particular, total H2 uptakes were determined by zero
pressure extrapolation of the adsorption isotherm, usually in the
hydrogen pressure range 6.6–26.6 kPa. Samples were subsequently
evacuated at RT for 0.5 h and a second isotherm due to the reversible
part of the adsorption was determined. The difference between the
total (first isotherm) and the reversible (second isotherm) uptakes
in the parallel and linear part of the isotherms gave the irreversible
part of the adsorption (double isotherm method). The exposed
Rh atoms, Rhs, were then calculated from Hirr (irr = irreversible),
assuming a Hirr/Rhs ratio of 1 and by taking into account the
2. Experimental
2.1. Catalyst preparation
The
synthesis
of
the
precursor
with
formula
(NH4)3[RhMo6O24H6]·7H2O (RhMo6) was carried out from
the aqueous solution reaction of ammonium heptamolybdate and
RhCl3·6H2O in stoichiometric proportions, as previously reported
[23,24].
In order to analyze the support effect, we selected an alu-
mina support thermally treated at 773 K in air for 24 h (hereafter
B) and two types of commercial ␥-Al2O3 spheralite, hereafter EI
and EII from YPF Oil Company (Yacimientos Petrolíferos Fiscales
Argentina). The procedure adopted for the catalyst preparation was
the equilibrium adsorption impregnation using heteropolymolyb-
date aqueous solution prepared as previously reported [24,25]. The
RhMo6 catalysts were obtained by adding 10 mg Mo/mL of RhMo6
aqueous solution to ␥-Al2O3 (EI and EII types) and Al2O3 (B) sup-
ports to obtain catalysts with ca. Mo 6 wt% and Rh 1 wt%, and
drying at 353 K. The resulting samples were labeled RhMo6/E(I),
RhMo6/E(II) and RhMo6/(B), respectively.
A conventional RhMo catalyst was prepared by successive
impregnation of type (B) alumina, in excess of pore volume. Accord-
ing to a literature method [21] this catalyst was obtained by first
adding the RhCl3·6H2O solution to the support and drying at 353 K.
The solid was then impregnated with a (NH4)6[Mo7O24]·4H2O
(AHM) solution and dried at 353 K. The resulting sample, containing
6% of Mo and 1% of Rh, was labeled RhMo/(B).
Before the catalytic tests, all catalysts were pretreated in a flow
of H2, at 623 K, for 2 h.
2.2. Chemical analysis
The Mo and Rh content was determined by atomic absorption
spectrometry (AAS), using an IL-457 spectrometer, on initial (Ci)
and final (Cf) solutions. The Mo and Rh adsorbed concentrations
(CaMo or CaRh) were calculated from experimental Ci (initial) and Cf
(final) data, taking into account the volume of the impregnated
solution and the support mass (mass balance) according to the
expression:
ꢀ
ꢁ
(Ci − Cf × V)
Ca
=
× 100.
m

G.R. Bertolini et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 109–115
111
overall stoichiometry: Rhs O + 3/2 H2 = Rhs Hirr + H2O. Finally the
Rh dispersion, D, was calculated as Hirr/Rhtot
.
2.4. Catalytic test
All catalysts were pretreated in H2 atmosphere at 623 K for 2 h
to achieve the reduction of metallic ions.
The hydrogenation reaction in liquid phase was performed in
an Autoclave Engineers reactor provided with a stirrer (500 rpm).
The reaction was performed under kinetic regime at 358 K with
a cinnamaldehyde concentration of 0.1 mol/L and a H2 pressure
of 10 bar, using 0.2 g of pretreated catalyst. The cinnamaldehyde
was introduced in the reactor dissolved in 60 mL of toluene. The
reaction progress was monitored by taking a series of micro sam-
ples, which were analyzed chromatographically by Varian 3400 CX
equipment provided with a capillary column Megabore DB-WAX
and a flame ionization detector. At the same time, complementary
measurements were made with a spectrometer CG/EM Q-mass 910.
The % CAL conversion was calculated from the difference
between initial CAL moles and CAL moles not consumed in a time
t/initial CAL moles.
The selectivity (%) was calculated based on the ratio product
moles/total moles of obtained products.
TOFRhsurf (turnover frequencies) numbers were calculated by
the [CAL] ␮mol s−1/surface Rh atoms equation from initial rate of
CAL conversion vs. time (from t = 0 to 50 min of reaction). Data of
Rh dispersion as Hirr/Rhtot were previously obtained by the above-
described Boudart method [32].
3. Results and discussion
3.1. Precursor and catalyst characterization
SEM-EDS microscopy showed that the synthesized crystals of
(NH4)3[RhMo6O24H6]·7H2O (RhMo6) (pale orange) were similar to
other Anderson phases widely studied in our laboratory such as
CoMo6, CrMo6 and AlMo6 [24]. The XRD pattern on powder samples
also demonstrated that the Rh phase was isomorphous to phases
containing a trivalent metal as heteroatom [23–25]. The polyhedral
Fig. 1. Representation of structures corresponding to: (a) planar heteropolyan-
structure of Rh(III) hexamolybdate of formula [RhMo6O24H6]3−
,
ion Rh(III) hexamolybdate [RhMo6O24H6]3− and (b) isopolyanion heptamolybdate
6−
with planar D3d symmetry, is constituted by Rh(III) as heteroatom
in the center of an octahedron whose corners are (OH) groups
shared with six Mo(VI)O6 octahedrons (Fig. 1(a)). The structure of
the heptamolybdate isopolyanion, formed by seven octahedrons
Mo(VI)O6 in anion planar configuration, is shown comparatively in
Fig. 1(b).
All catalysts prepared by both equilibrium adsorption and suc-
cessive impregnation methods contained about 6 wt% Mo and
1 wt% Rh. As demonstrated in our previous work, the adsorp-
tion process of RhMo6 responded to a Langmuir-type adsorption
isotherm as for all XMo6 phases [24]. We have previously
shown that on ␥-Al2O3 spheralite (EI) [24,25] the heteropolyanion
adsorption depends strongly on the heteroatom. The adsorp-
[Mo7O24
]
.
The XRD analysis of pretreated RhMo6 catalysts supported
on spheralite-type alumina (EI and EII) showed the presence of
␥-Al2O3 (ASTM: 89-7717), Fig. 2, curve a. The XRD pattern of
both pretreated RhMo6 and RhMo catalysts on support B showed
lines corresponding to ␣-AlO(OH) (boehmite, ASTM-83-2384) and
emerging peaks from ␥-Al2O3 (Fig. 2, curve b). However, all the pat-
terns of supported catalysts showed no lines other than those of the
supports, indicating spreading of different structures containing Rh
and Mo.
The textural analysis of the supports, Table 1, indicated that
support B had lower surface area and pore volume values than
E(I) and E(II) supports. The pore size distribution of the supports,
Fig. 3, reveals that the higher surface area of E(II) with respect to
those measured for E(I) and ␥-Al2O3 B (319 m2 g−1 vs. 255 and
tion constant, Kad
, varies according to the following order:
RhMo6 > CrMo6 > CoMo6 > TeMo6–AlMo6 > NiMo6 ꢀ AHM [24,25].
The Kad resulted several orders of magnitude higher than those of
AHM, and this was explained not only by the planar heteropolyan-
ion structure but also by the heteroatom (X) type. All these factors
indicated that in the first impregnation step, a series of homo-
geneous and heterogeneous processes in equilibrium occurred
involving the phase deposition and some secondary reactions.
These reactions, including the Al dissolution of the support, counter
diffusion and exchange of cations, have already been characterized
by spectroscopic and thermal methods [24].
Table 1
Textural parameters of different alumina supports.
Support type
SBET (m2 g−1
)
Pore volume (cm3 g−1
)
(B) Al2O3 (T = 500 ◦C)
(EI) ␥-Al2O3 Spheralite I
(EII) ␥-Al2O3 Spheralite II
149
255
319
0.23
0.65
0.36

112
G.R. Bertolini et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 109–115
(b)
(a)
10
20
30
40
50
60
2θ (degree)
Fig. 2. XRD pattern of pretreated supported RhMo6 catalysts: (a) RhMo6/␥-Al2O3
Fig. 4. Raman spectra: (a) (NH4)3[RhMo6O24H6]·7H2O (RhMo6) and (b) RhMo6/E(I).
spheralite E(I) or E(II); (b) RhMo or RhMo6/Al2O3(B).
149 m2 g−1, respectively), may be assigned to the higher fraction of
narrow mesopores and/or micropores (i.e. pores around and below
30 A in average diameter) present in this support.
and 852 cm−1 (ꢁas), merged into an unresolved feature. A different
behavior was observed for the fresh catalysts prepared by succes-
sive impregnation of the support with rhodium chloride and AHM
aqueous solutions.
˚
The identification of molecular structures, such as those of
iso and heteropolyoxomolybdates characterized by Mo
Mo Mo bridges and terminal Mo O bonds, can easily be accom-
plished by Raman spectroscopy [33].
O X and
The comparison of the Raman features of AHM (Fig. 5, curve a)
with those of the RhMo/(B) catalyst (Fig. 5, curve b) shows that
the symmetric stretching mode MoO2t at 934 cm−1 of the hep-
tamolybdate species broadened and shifted to higher frequencies
(947 cm−1), which is indicative of a change in the structure of the
supported isopolyanion. It is well known that during impregnation
of the alumina support with an aqueous AHM solution, precipi-
tation of the hexamolybdoaluminate of Anderson-type structure
(AlMo6) occurs due to the extraction of Al atoms from the alu-
O
By comparing the Raman spectrum of the RhMo6 precursor
(Fig. 4, curve a) with that of RhMo6/␥-Al2O3 [E(I)] fresh catalyst
(Fig. 4, curve b), it is evident that the interaction of the Ander-
son phase with the support yielded band broadening and shift to
lower frequencies in agreement with previously published results
[33]. In fact, the sharp bands, typical of Anderson structure, asso-
ciated with the symmetric stretching mode of terminal Mo O2t
bonds at 949 cm−1 (ꢁs) and the antisymmetric stretching at 889
mina surface [34]. The presence of a broad band around 540 cm−1
ascribed to Mo Al bonds typical of Anderson structures, corre-
,
O
sponds to Mo Ob modes observed in the spectrum of analog RhMo6
phase [Fig. 4(a)] (ꢁMo Ob at 549 cm−1). In addition, the presence
of ␣-AlO(OH) in the support as a major component is confirmed by
the band centered around 333 cm−1, in agreement with the Raman
pattern of crystalline ␣-AlO(OH) [35].
On the other hand, in the system prepared by successive impreg-
nation, RhMo/(B), no bands corresponding to the stretching modes
of Rh O (560 cm−1) were found. It is also suggested that in this
Fig. 3. Pore size distributions of the pure supports Spheralite E(I), E(II) and ␥-
Al2O3(B).
Fig. 5. Raman spectra of (a) (NH4)6Mo7O24·4H2O (AHM) and (b) RhMo/(B).

G.R. Bertolini et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 109–115
113
Table 2
645
(f) RhMo/(B)
876
Binding energies (eV) and surface composition of pure RhMo6 and different sup-
ported catalysts.
516
Sample
Rh3d5/2 (eV)
Mo3d5/2 (eV)
Sample composition
(nRh/nMo)a
(nRh/nMo)b
482
RhMo6
RhMo6/(EI)
RhMo/(B)
310.8
309.4
309.3
233.0
232.7
232.6
0.17
0.23
1.22
0.167
0.223
0.229
639
660
604
(e) RhMo6/(B)
856
480
a
XPS derived values.
Analytical data.
b
(d) RhMo6/(EII)
(c) RhMo6/(EI)
484
system, Rh
assigned to the Mo
O
Al bands may overlap with the broad feature
Al stretching modes [36,37].
O
XPS analysis provided some insight into the chemical state
and dispersion degree of the surface species both for the
unsupported RhMo6 compound and for the supported catalysts:
RhMo6/Al2O3[E(I)] and RhMo/Al2O3(B), Table 2.
580
550
Data for the pure RhMo6 phase revealed a Rh3d5/2 signal at
310.8 eV and a Mo3d5/2 signal at 233.0 eV, corresponding to Rh(III)
and Mo(VI), respectively. Because of the overlapping of N1s with the
Mo3p3/2 component, the state and amount of N could not be deter-
mined. The surface atomic composition nRh/nMo = 0.17, derived
from the Rh/Mo intensity ratio [26], is in excellent agreement with
the bulk composition, nRh/nMo = 0.167.
643
(b) RhMo6
980
1034
(a) AHM
Both
RhMo6/Al2O3[E(I)]
and
conventional
system
703
RhMo/Al2O3(B) samples showed practically constant Rh3d5/2
(309.2 0.2 eV) and Mo3d5/2 (232.4 0.2 eV) binding energies that
are characteristic of Rh(III) and Mo(VI), respectively (see Table 2).
As for H2 pretreated samples, although care was taken to avoid
surface reoxidation during insertion into the XPS instrument,
no reduced Mo and Rh species were detected. However, surface
reoxidation is not expected to affect surface composition. Similar
results, within experimental error, were obtained both for fresh
and pretreated samples. For the RhMo6/Al2O3 [E(I)] sample, the
agreement between bulk and surface nRh/nMo atomic ratios was
indicative of a uniform distribution of the RhMo6 phase on the
support surface. Conversely, the conventional catalyst prepared
by successive impregnation of alumina B, exhibits Rh surface
enrichment, as inferred from the increase of the nRh/nMo surface
ratio.
The TPR technique can be an interesting method for studying the
RhMo6 heteropolyanion-support interaction and the influence of
the Rh heteroatom on Mo reducibility [24–26,38–41]. Fig. 6 shows
comparative TPR diagrams of the AHM and RhMo6 bulk phases and
RhMo6/E(I)/(B) and RhMo/(B) fresh catalysts, whereas Table 3 col-
lects the TPR data corresponding to the Rh and Mo reduction steps
observed for the various samples.
400
600
800
1000
T [K]
Fig. 6. TPR patterns of bulk phases and supported catalysts: (a) AHM and (b) RhMo6;
(c) RhMo6/E(I); (d) RhMo6/E(II); (e) RhMo6/(B) and (f) RhMo/(B).
The profile for RhMo6 species displayed an intense peak at 550 K
and low intensity H2 consumption peaks at 643 and 980 K (Fig. 6,
curve b). The first peak can be related to the reduction of Rh(III)
to Rh0 and the two low intensity peaks can be assigned to the
molybdenum reduction steps. The RhMo6/E(I) catalyst exhibited H2
consumption peaks at 484 and 580 K and a broad feature at about
1000 K (Fig. 6, curve c), whereas both RhMo/(B) and RhMo6/(B) cat-
alysts showed complex TPR profiles with a new contribution in the
range 600–900 K (Fig. 6, curves e and f).
Dispersion of the RhMo6 phase, either on E(I) or E(II) support,
resulted in a promoting effect on the reduction behavior of both Rh
and Mo species. The reduction temperature shifted to lower val-
ues and yielded a partial overlapping of both the contributions due
to Rh(III) → Rh0 and Mo reduction steps. Likewise, the decrease in
temperature of the first TPR peak of the RhMo6 catalysts, if com-
pared with the same TPR peak of the bulk phase, indicates that there
was an enrichment of Rh on the surface of the catalysts, especially
for support B, as demonstrated by XPS analysis.
The TPR profile of AHM (Fig. 6, curve a) consisted of a low inten-
sity peak at about 703 K and a high intensity one at about 1033 K,
(1)
(2)
corresponding to two reduction steps: Mo(VI)−→ Mo(IV)−→ Mo0.
For the [RhMo6/(B)] and [RhMo/(B)] catalysts, the peaks at
lower temperature are assigned to Rh reduction and the other can
be assigned to Mo reduction steps. Their patterns (Fig. 6, curves
e and f), are more complex than the patterns detected for the
RhMo6–E(I)/E(II) samples (Fig. 6, curves c and d). The TPR results
confirm that the successive adsorption of Rh(III) and Mo(VI) on alu-
mina produced the formation of X(Rh/Al)Mo6 Anderson species and
other oxides such as MoO3 or MoO3−x whose reduction tempera-
tures lie around 900 K [24,39–41].
The high RhMo6 reducibility can be associated with the reduc-
tion potential of the heteroatom [E0 Rh(III)–Rh0 = 0.44 V], which
promotes H2 activation. In fact, Rh(III) affects the stability of Mo(VI),
whose reduction starts at lower temperature, as evidenced by high
H2 consumptions peaks centered at 550 K in the pure phase and at
Table 3
TPR data corresponding to the Rh and Mo reduction steps observed for the various
bulk and supported catalysts.
Sample
TPR signal (◦K)
Rh(III)–Rh0
Mo(VI)–Mo(IV)
Mo(IV)–Mo0
AHM
RhMo6
RhMo6/(EI) or
(EII)
RhMo6/(B)
RhMo/(B)
–
703 vw
643 w
580 (660) w
1034 vs
980 vw
–
550 vs
484 (480) vs
482 vs
516 m
604–639 w
645 m
856 w
876 m
Ref.: m: medium, s: strong, w: weak, v: very.

114
G.R. Bertolini et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 109–115
100
80
60
40
20
0
RhMo6/(B)
RhMo/(B)
RhMo6/E(II)
RhMo6/E(I)
0
50
100
150
time (min)
Scheme 1.
Fig. 7. Cinnamaldehyde conversion as a function of time for RhMo6/E(I); E(II); (B)
and RhMo/(B) catalysts.
484 K in the supported ones. According to our recent investigations,
this behavior is comparable to that found in other heteropolyanions
of Anderson type XMo6 [X = Cu(II), Co(III), Ni(II), Al(III)] supported
on alumina [24,25,39–41]. The comparison between different
phases allowed us to conclude that, in general, the reduction tem-
peratures for the two processes Mo(VI)−→ Mo(IV)−→ Mo0 depend
mainly on the reducibility of the heteroatom. In general, when the
heteroatom was less reducible, Mo(VI) became more stable.
It was also evident that the heteroatom influence on Mo
reducibility depends on its redox potential according to the fol-
lowing order: E0 Rh(III) > E0 Al(III) > E0 Mo(VI) [24,39].
(1)
(2)
3.2. Catalytic evaluation
Possible paths for the hydrogenation of cinnamaldehyde (CAL),
involving the formation of cinnamic alcohol (CA, formed by hydro-
genation of the C O bond), hydrocinnamaldehyde (HCAL, formed
by hydrogenation of C C bond) and 3-phenylpropanol (PP, formed
through the hydrogenation of both C O and C C bonds) can be
observed in Scheme 1.
The chromatographic analysis indicated that cinnamaldehyde
hydrogenation on our studied catalysts essentially formed HCAL,
CA and PP. Catalyst composition, conversion and selectivity values
after 75 min of hydrogenation of cinnamaldehyde (CAL) are listed
in Table 4.
The evolution of cinnamaldehyde conversion as a function of
time is shown in Fig. 7 for all the catalysts. The conversion attained
with RhMo6/E(I) and RhMo6/E(II) catalysts was significantly higher
than the corresponding values for RhMo/(B) and RhMo6/(B). After
ca. 60 min reaction time, more than 60% CAL had been converted in
RhMo6/E(I) and RhMo6/E(II) catalysts.
On the other hand, B-supported catalysts showed a conversion
profile corresponding to a catalyst of very low activity. Initial CAL
conversion rates (r0, ␮mol s−1) were calculated by extrapolation
to zero time and gathered in Table 5 with the Rh atoms number,
Rh dispersion as H/Rh and TOF values [42,43]. The r0 and TOF for
Fig. 8. Selectivity (%) at 55% conversion for the cinnamaldehyde hydrogenation
reaction for RhMo6/(EI, EII and B) and RhMo/(B) catalysts.
RhMo6/E(I) are about one order of magnitude higher than those of
the rest of the catalysts studied. Although RhMo6/E(II) showed the
highest activity at 75 min, its values of TOF and dispersion of Rh
were comparable to those of catalysts supported on alumina (B).
The product selectivity at 15% conversion of CAL is listed in
Table 4 (except for RhMo/(B) measured at 10% conversion). The
main reaction product with the four catalysts was HCAL (as seen in
Fig. 8), thus indicating that they selectively catalyze the hydrogena-
tion of the C C bond. To explain the behavior related to selectivity,
we must consider that the proximity between the Rh and Mo, in
addition to the stoichiometry and structure of the heteropolyan-
ion, could lead to the oxidative stabilization of small aggregates
of Rh. As a result, the interaction of Rh with carbonyl from cin-
namaldehyde, favors the formation of HCAL. These results are in
agreement with those reported by the extensive work of Lowen-
thal et al. about chemisorption of H and CO for various systems of
Rh and Mo supported on both silica and alumina [21].
These studies have showed that a close proximity of Mo and Rh
lead to both oxidative and kinetic stabilization of small aggregates
of Rh and the coexistence of species Rhı+ and MoOx. Nevertheless,
Table 4
Composition (adsorbed Mo and Rh, wt%); conversion (%) and selectivity (%) for all
catalysts on the hydrogenation of cinnamaldehyde (CAL) to hydrocinnamaldehyde
(HCAL), cinnamic alcohol (CA) and 3-phenyl propanol (PP).
Table 5
Rh number atoms, initial rate of CAL conversion, Rh dispersion as H/Rh and TOF
calculated from [CAL] ␮moles s−1/Rh surface atoms.
Catalyst
CaMo (%)
CaRh (%)
Conversiona (%)
Selectivity (%)b
HCAL
CA
0.0
0.0
0.0
PP
Catalyst
Rh atomsa
Rh dispersion
(H/Rh)
Initial rate
TOF
(×1019
)
(CAL ␮mol s−1
)
(×10−3
)
RhMo6/(EI)
RhMo6/(EII)
RhMo6/(B)
RhMo/(B)
6.0
6.7
5.7
6.0
1.5
1.5
1.4
1.0
79.0
59.0
16.0
11.0
100.0
100.0
94.0
0.0
0.0
6.0
0.0
RhMo6/E(I)
RhMo6/E(II)
RhMo6/(B)
RhMo/(B)
1.75
1.75
1.64
1.17
0.50
0.35
0.34
0.30
2.45
168
49
40
0.502
0.370
0.256
90.0
10.0
44
a
Conversion at 75 min reaction.
Selectivity at 15% conversion, except for RhMo/(B) calculated at 10% conversion.
b
a
In 0.200 g of sample.

G.R. Bertolini et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 109–115
115
with RhMo/(B) catalyst a certain level of CA was obtained (10%).
In general, systems based on RhMo6 showed a different catalytic
performance in the hydrogenation of cinnamaldehyde if they are
compared to the RhMo/(B) catalyst based on Rh and AHM.
Taking into account that XRD did not indicate the presence of
structures containing Rh and/or Mo in the supported catalysts, it is
possible to suggest that Rh species are present as small metallic par-
ticles that cannot be observed by XRD. In addition, TPR results for
RhMo6 systems showed the Rh reduction at low temperature. This
effect may also indicate that reduced species are dispersed as small
particles weakly interacting with the support and therefore, the
selectivity results can be explained in terms of a particle size effect.
According to the classical paper of Richard et al. [44], large
particles lead to the unsaturated alcohol preferentially, whereas
small particles lead to the saturated aldehyde. Besides, steric factors
related to the molecule to be reduced cannot be discarded, and thus,
the stiffness of the cinnamaldehyde molecule (where the phenyl
ring is conjugated with C C and C O bonds) favors its adsorption
via the C C bond on small particles and the C O bond on large
particles [10].
The temperature programmed reduction (TPR) technique
revealed that Rh species induce Mo(VI) reducibility at lower tem-
peratures in the catalysts as well as in the pure phase, confirming
the coexistence of Rhı+ and MoOx, species, and thus the oxida-
tive stabilization of small Rh aggregates for the heteropolyanion
structure.
This study confirms that the heteropolyanion-support interac-
tion produces an active surface with an ordered distribution of Rh
and Mo in the support causing a synergic effect that favors the
catalytic activity.
References
[1] P. Gallezot, D. Richard, Catal. Rev. Sci. Eng. 40 (1998) 81.
[2] J. Kijenski, P. Winiarek, Appl. Catal. A: Gen. 193 (2000) L1–L4.
[3] F. Delbecq, P. Sautet, J. Catal. 152 (1995) 217.
[4] I. Kroschwitz (Ed.), Kirk-Othmer, En. of Chem. Tech., vol. 6, 4th ed., Wiley, New
York, 1992, p. 349.
[5] S. Narayanan, Bull. Catal. Soc. India 2 (2003) 107.
[6] A.M.C.F. Castelijns, J.M. Hogeweg, S.P.J.M. van Nispen, PCT Int. Appl. WO
96/11898 Al (April 25, 1996) 14 pp.
[7] A.M.C.F. Castelijns, J.M. Hogeweg, S.P.J.M. van Nispen, PCT Int. Appl., US Patent
5,811,588 (September 22, 1998) pp. 6.
Then, the selectivity to CA obtained with the RhMo/(B) catalyst
may be explained as a consequence of the effect of both Mo(VI) and
Rh(III). In fact, the RhMo/(B) (Fig. 6(f)) TPR showed that the reduc-
tion peaks appear at higher temperature than those of the other
catalysts and thus the amount of non-reduced Rh species at the
temperature employed in the reduction pretreatment of our cata-
lysts (623 K) could be higher than for the other catalysts. Besides,
XPS data revealed a surface enrichment in Rhı+ species, which could
be responsible for C O bond polarization. In addition, the effect of
Mo(VI) may have caused the activation of the C O bond, facilitating
the hydrogen transfer from adjacent Rh sites.
[8] A. Muller, J. Bowers, WO Patent WO 99/08989 (February 25, 1999) to First
Chemical Corporation.
[9] G.F. Santori, M.L. Casella, O.A. Ferretti, J. Mol. Catal. A: Chem. 186 (2002) 223.
[10] M. Lashdaf, A.O.I. Krause, M. Lindblad, M. Tiitta, T. Venäläinen, Appl. Catal. A:
Gen. 241 (2003) 65.
[11] J. Breen, R. Burch, J. Gomez-Lopez, K. Griffin, M. Hayes, Appl. Catal. A: Gen. 268
(2004) 267.
[12] P. Reyes, C. Rodriguez, G. Pecchi, J.L. Fierro, Catal. Lett. 69 (2000) 27.
[13] P.N. Rylander, Catalytic Hydrogenation in Organic Syntheses, Academic Press,
New York/San Francisco/London, 1979, p. 78.
[14] A. Giroir-Fendler, D. Richard, P. Gallezot, Stud. Surf. Sci. Catal. 41 (1988) 171.
[15] Y. Zhang, S. Liao, Y. Xu, D. Yu, Appl. Catal. A: Gen. 192 (2000) 247.
[16] J.-P. Tessonnier, L. Pesant, G. Ehret, M.J. Ledoux, C. Pham-Huu, Appl. Catal. A:
Gen. 288 (2005) 203.
[17] Y. Kume, K. Qiao, D. Tomida, C. Yokoyama, Catal. Commun.
369–375.
9 (2008)
On the other hand, the comparison between the three cata-
lysts based on supported Anderson phase, RhMo6/E(I)/E(II) and/(B),
indicates that textural differences of the support seem to cause a
significant effect on their catalytic behavior. In fact, BJH analysis of
[18] V. Ponec, Appl. Catal. A: Gen. 222 (2001) 31.
[19] B.M. Reddya, G.M. Kumara, I. Ganesh, A. Khan, J. Mol. Catal. A: Chem. 247 (2006)
80.
[20] B. Coq, F. Figueras, J. Mol. Catal. A: Chem. 173 (2001) 117.
[21] E.E. Lowenthal, L.F. Allard, M. Te, H.C. Foley, J. Mol. Catal. A: Chem. 100 (1995)
129–145.
adsorption data on the three supports shows that the cumulati−v1e
2
˚
pore areas for mesopores (below 30 A) are 150, 50 and 20 m g
for E(II), E(I) and alumina (B), respectively. Maybe this high fraction
(about 50%) of the BET surface area of the support E(II) is difficult
to be reached by Anderson phase during the preparation, whereas
alumina E(I) exhibits a more favorable pore size distribution with
only 20% of the total BET surface area entrapped in narrow meso-
pores (see Fig. 3). The observed differences in both the dispersion
of the metallic phase and the TOF values may be assigned to the
different textural characteristics of these two supports.
[22] N. Mizuno, M. Misono, Chem. Rev. 98 (1998) 199.
[23] C.I. Cabello, I.L. Botto, H.J. Thomas, Appl. Catal. A 197 (2000) 79.
[24] C.I. Cabello, I.L. Botto, F. Cabrerizo, M. González, H. Thomas, Adsorpt. Sci. Tech-
nol. 18 (7) (2000) 591.
[25] C.I. Cabello, I.L. Botto, M. Mun˜oz, H. Thomas, Stud. Surf. Sci. Catal. 143 (2002)
565–573.
[26] C.D. Wagner, L.E. Davis, M.V. Zeller, J.A. Taylor, R.H. Raymond, L.H. Gale, Surf.
Interface Anal. 3 (1981) 211.
[27] Brunauer, Emmett and Teller (BET) method, in: J.R. Anderson, K.C. Pratt (Eds.),
Introduction to Characterization and Testing of Catalysts, Academic Press,
Australia, 1985.
[28] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Ceram. Soc. 73 (1951) 373.
[29] B.J. Lippens, J.H. de Boer, J. Catal. 4 (1965) 319.
[30] W.D. Harkins, G. Jura, J. Am. Ceram. Soc. 66 (1944) 1366.
[31] L. Gurvitsch, J. Phys. Chem. Soc. Russ. 47 (1915) 805.
[32] J.E. Benson, M. Boudart, J. Catal. 4 (1965) 704.
[33] C.I. Cabello, M. Mun˜oz, I.L. Botto, E. Payen, Thermochim. Acta 447 (2006) 22.
[34] L. Le Bihan, P. Blanchard, M. Fournier, J. Grimblot, E. Payen, J. Chem. Soc., Faraday
Trans. 94 (7) (1998) 937.
Finally, it is possible to suggest that the preferentially C
C
hydrogenation of catalysts based on RhMo6 species depends mainly
on the nature of the heteropolyanion that provokes a reduction
in the ability of both metals needed for generating the selective
hydrogenation of CAL. An additional advantage in the use of the
heteropolyanion as catalyst precursor is the smallest amount of Rh
and the avoidance of sintering.
[35] C.J. Doss, Phys. Rev. B 48 (1993) 15626.
ˇ
[36] S. Music´, A. Saric´, S. Popovic´, M. Ivanda, J. Mol. Struct. 221 (2009) 924–926.
[37] C.T. Williams, E.K.-Y. Chen, C.G. Takoudis, M.J. Weaver, Phys. Chem. J. B102
(1998) 4785.
4. Conclusions
[38] C.I. Cabello, I.L. Botto, H.J. Thomas, Thermochim. Acta 232 (2) (1994) 183.
[39] I.L. Botto, C.I. Cabello, G. Minelli, M. Occhiuzzi, Mater. Chem. Phys. 39 (1) (1994)
21–28.
[40] I.L. Botto, C.I. Cabello, H.J. Thomas, Mater. Chem. Phys. 47 (1) (1997) 37.
[41] I.L. Botto, C.I. Cabello, H.J. Thomas, D. Cordischi, G. Minelli, P. Porta, Mater. Chem.
Phys. 62 (3) (2000) 254–262.
[42] A. Merlo, B. Machado, V. Vetere, J.L. Faria, M. Casella, Appl. Catal. A: Gen. 383
(2010) 43.
[43] L. Mercadante, G. Neri, C. Milone, A. Donato, S. Galvagno, J. Mol. Catal. A: Chem.
105 (1996) 93.
In general, systems based on RhMo6 show a better cat-
alytic performance in the hydrogenation of cinnamaldehyde to
hydrocinnamaldehyde than the catalyst prepared by successive
impregnation.
Raman and XPS spectroscopies detected typical features cor-
responding to the presence of Anderson type structures on the
surfaceofall catalysts, includingthecatalystpreparedbysuccessive
impregnation of Rh(III) and Mo(VI) salts.
[44] A. Giroir-Fendler, D. Richard, P. Gallezot, Catal. Lett. 5 (1990) 175.