Hydrogenation of arenes over silica-supported catalysts that combine a grafted rhodium complex and palladium nanoparticles: Evidence for substrate activation on Rhsingle-site-Pdmetal moieties

Article abstract of DOI:10.1021/ja060235w

The complex Rh(cod)(sulfos) (Rh¹; sulfos = ^-O ₃S(C₆H₄)CH₂C(CH₂PPh ₂)₃; cod = cycloocta-1,5-diene), either free or supported on silica, does not catalyze the

Full text of DOI:10.1021/ja060235w

Published on Web 05/05/2006
Hydrogenation of Arenes over Silica-Supported Catalysts That
Combine a Grafted Rhodium Complex and Palladium
Nanoparticles: Evidence for Substrate Activation on
Rhsingle-site-Pdmetal Moieties
†
,†
‡
†
Pierluigi Barbaro, Claudio Bianchini,* Vladimiro Dal Santo, Andrea Meli,
†
,‡
†
‡
Simonetta Moneti, Rinaldo Psaro,* Adriana Scaffidi, Laura Sordelli, and
Francesco Vizza†
Contribution from ICCOM-CNR, Area di Ricerca CNR di Firenze, Via Madonna del Piano 10,
5
0019 Sesto Fiorentino, Firenze, Italy, and ISTM-CNR, Via C. Golgi 19, 20133 Milano, Italy
Received January 12, 2006; E-mail: claudio.bianchini@iccom.cnr.it
Abstract: The complex Rh(cod)(sulfos) (Rh ; sulfos ) -O
I
3 6 4 2 2 2
S(C H )CH C(CH PPh )
3
; cod ) cycloocta-1,5-
diene), either free or supported on silica, does not catalyze the hydrogenation of benzene in either
homogeneous or heterogeneous phase. However, when silica contains supported Pd metal nanoparticles
0
I
0
(
Pd /SiO
2
), a hybrid catalyst (Rh -Pd /SiO
2
) is formed that hydrogenates benzene 4 times faster than does
0
Pd /SiO
2
alone. EXAFS and DRIFT measurements of in situ and ex situ prepared samples, batch catalytic
reactions under different conditions, deuterium labeling experiments, and model organometallic studies,
taken together, have shown that the rhodium single sites and the palladium nanoparticles cooperate with
each other in promoting the hydrogenation of benzene through the formation of a unique entity throughout
the catalytic cycle. Besides decreasing the extent of cyclohexa-1,3-diene disproportionation at palladium,
the combined action of the two metals activates the arene so as to allow the rhodium sites to enter the
catalytic cycle and speed up the overall hydrogenation process by rapidly reducing benzene to cyclohexa-
1,3-diene.
Introduction
Scheme 1
The capability of the sulfonate group to form robust H-bonds
to terminal silanols has been recently exploited to immobilize
transition metal complexes over silica.1-3 This grafting proce-
dure has been successfully employed to heterogenize the rho-
I
-
dium(I) complex Rh(cod)(sulfos) (Rh ; sulfos ) O3S(C6H4)-
CH2C(CH2PPh2)3; cod ) cycloocta-1,5-diene), yielding the
supported hydrogen-bonded catalyst Rh(cod)(sulfos)/SiO2
I
1d
(
Rh /SiO2, Scheme 1a).
If silica already holds supported palladium particles
0
I
0
(
Pd /SiO2), the hybrid system Rh -Pd /SiO2 is formed, contain-
ing both rhodium single sites and palladium particles (Scheme
1
b), which exhibits unusual properties as a catalyst for the
4
hydrogenation of arenes. In particular, the combined single-
†
ICCOM-CNR, Sesto Fiorentino, Firenze.
ISTM-CNR, Milano.
‡
(
1) (a) Bianchini, C.; Dal Santo, V.; Meli, A.; Moneti, S.; Moreno, M.;
Oberhauser, W.; Psaro, R.; Soredelli, L.; Vizza, F. J. Catal. 2003, 213, 47.
(b) Bianchini, C.; Barbaro, P.; Dal Santo, V.; Gobetto, R.; Meli, A.;
I
0
site/dispersed-metal catalyst Rh -Pd /SiO2 is 4 times more
Oberhauser, W.; Psaro, R.; Vizza, F. AdV. Synth. Catal. 2001, 343, 41. (c)
Bianchini, C.; Dal Santo, V.; Meli, A.; Oberhauser, W.; Psaro, R.; Vizza,
F. Organometallics 2000, 19, 2433. (d) Bianchini, C.; Burnaby, D. G.;
Evans, J.; Frediani, P.; Meli, A.; Oberhauser, W.; Psaro, R.; Sordelli, L.;
Vizza, F. J. Am. Chem. Soc. 1999, 121, 5961.
0
active than Pd /SiO2 for the hydrogenation of benzene or
I
toluene, while Rh /SiO2 is totally inactive. Independent experi-
ments with cyclohexa-1,3-dienes and cyclohexenes showed that
cyclohexa-1,3-dienes are more rapidly reduced to cyclohexenes
(
2) De Rege, F. M.; Morita, D. K.; Ott, K. C.; Tumas, W.; Broene, R. D.
Chem. Commun. 2000, 1797.
(
3) (a) Bianchini, C.; Meli, A.; Vizza, F. In The Handbook of Homogeneous
Hydrogenation; de Vries, J. G., Ed.; Wiley-VCH: Weinheim, Germany,
(4) Bianchini, C.; Dal Santo, V.; Meli, A.; Moneti, S.; Moreno, M.; Oberhauser,
W.; Psaro, R.; Sordelli, L.; Vizza, F. Angew. Chem., Int. Ed. 2003, 42,
2636.
2
006. (b) Thomas, J. M.; Raja, R.; Lewis, D. W. Angew. Chem., Int. Ed.
005, 44, 6456.
2
10.1021/ja060235w CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 7065-7076
9
7065

A R T I C L E S
Scheme 2
Barbaro et al.
which we feel has substantially contributed to shed light on the
intimate mechanism of the palladium-rhodium interaction.
Experimental Section
General Information. All reactions and manipulations were rou-
tinely performed under a nitrogen or argon atmosphere by using
standard Schlenk techniques, unless stated otherwise. CH
2 2
Cl and
n-pentane were distilled under nitrogen from CaH and LiAlH ,
2
4
respectively. Benzene and toluene employed in the catalytic reactions
were distilled over Na prior to use. The rhodium complex Rh(cod)-
I
11
(
sulfos) (Rh ) was prepared as previously described. All the other
at rhodium, while cyclohexenes are predominantly reduced at
palladium. It was also found that the cyclohexa-1,3-diene
disproportionation, occurring on palladium, is inhibited by the
grafted rhodium complex. On the basis of these pieces of
information as well as a number of experiments, including the
isolation of relevant intermediates, the authors concluded that
reagents and chemicals were reagent grade and were used as received
from commercial suppliers. The Davicat (Grace) silica employed in
this work was a high-surface-area hydrophilic mesoporous nonordered
3
material. The support was ground, washed with 1 M HNO and distilled
water to neutrality, and dried overnight in an oven at 100 °C.
Porosimetry and surface area were determined by nitrogen adsorption.
Nitrogen adsorption/desorption isotherms at liquid nitrogen temperature
were measured on a Micromeritics ASAP 2010 instrument. All silica
samples were routinely pre-outgassed at 300 °C. Average pore radius
I
0
the enhanced activity of the Rh -Pd /SiO2 catalyst is due not
5
to hydrogen spillover, as suggested by other authors for arene
hydrogenation by tethered complexes on supported metals
3
6
(9.70 nm) and specific pore volume (1.43 cm /g) were calculated
(
TCSM), but to the fact that the rate-limiting hydrogenation
12
according to the Barret-Joyner-Halenda (BJH) theory. The specific
of benzenes to cyclohexa-1,3-dienes is assisted by both pal-
ladium and rhodium. The concerted action of the two metals,
besides preventing the competitive diene disproportionation to
benzene and cyclohexene, apparently speeds up the reduction
of the first double bond (Scheme 2). However, it was concluded
that the intimate mechanism of the Pd-Rh interaction in the
2
surface area (295 m /g) was obtained using the Brunauer-Emmett-
Teller (BET) equation.13 The rhodium contents in the tethered catalysts
were determined by inductively coupled plasma atomic emission
spectroscopy (ICP-AES) with a Jobin Yvon (series JY24) instrument
at a sensitivity level of 500 ppb. Each sample (20-50 mg) was treated
in a microwave-heated digestion bomb (Milestone, MLS-200) with
4
first hydrogenation step is still rather obscure. Addressing this
concentrated HNO
and a pellet (0.4 g) of a digestion aid reagent (0.1% Se in K
3
(1.5 mL), 98% H
2
SO
4
(2 mL), 37% HCl (0.5 mL),
SO ).
question would be of importance, as it would open the door to
the rational design of a new generation of tailored heterogeneous
catalysts for diverse applications through appropriate combina-
tions of supported metals and grafted molecular complexes.
Indeed, several TCSM catalysts, with either covalent or H-bond
grafting, are actually available with proven effectiveness in a
2
4
After the silica particles were filtered off, the solutions were analyzed.
The addition of selenium was necessary to get an effective digestion
of the phosphine ligand, which was hardly achievable by usual acid
dissolution procedures. The same digestion method was employed to
determine the metal contents in the products recovered after catalysis
and in the organic solutions. The palladium contents in the heteroge-
neous catalysts were determined by flame atomic absorption spectros-
copy (AAS) with a Perkin-Elmer 400 atomic absorption spectropho-
tometer. Each sample (ca. 50 mg) was treated with aqua regia (ca. 12
mL) and then heated to the boiling temperature for 1-2 h, during which
time two 3-mL portions of HCl were added to the mixture. Batch
reactions under a controlled pressure of gas were performed with a
stainless steel Parr 4565 reactor (100 mL) equipped with a Parr 4842
temperature and pressure controller and a paddle stirrer. GC analyses
of the solutions were performed on a Shimadzu GC-14 A gas
chromatograph equipped with a flame ionization detector and a 30 m
5,6
variety of processes such as arene hydrogenation, enantiose-
1
b,2
lective hydrogenation of alkenes,
hydrodefluorination of
7
1d,8
fluoroarenes, alkene hydroformylation, hydrodechlorination
of chlorophenols, and hydrogenolysis of benzylic functions.10
In an attempt at elucidating the synergic action exerted by
rhodium and palladium on arene hydrogenation, an extended
X-ray absorption fine structure (EXAFS) study has been carried
out on several catalytically relevant or model compounds,
5,9
I
0
including the starting precursor Rh -Pd /SiO2, before and after
hydrogenation, and the termination metal product of a batch
hydrogenation of benzene. Parallel to the EXAFS investigation,
all relevant materials were characterized by high-resolution
transmission electron microscopy (HRTEM), and the hydroge-
nation of benzene and toluene was achieved under different
catalytic conditions to draw out experimental kinetic informa-
tion. In this paper is reported an account of this multiform study,
(
0.25 mm i.d., 0.25 µm film thickness) SPB-1 Supelco fused silica
capillary column. GC/MS analyses were performed on a Shimadzu QP
000 apparatus equipped with an identical capillary column. The
stereochemistry of the benzene hydrogenation product was investigated
5
13
1
13
1
by C{ H} NMR spectroscopy using 99.9% deuterated C D . C{ H}
6
6
NMR spectra were recorded on a Bruker Avance DRX-400 spectrom-
eter operating at 100.613 MHz and equipped with a temperature
controller accurate to (0.1 °C. A total of 2048 scans of size 64K
covering the full range (15 432.1 Hz) were acquired with a relaxation
delay of 5 s. Chemical shifts are relative to external TMS, with
downfield values reported as positive. Experimental 13C{ H} NMR
(
5) Abu-Reziq, R.; Avnir, D.; Miloslavski, I.; Schumann, H.; Blum, J. J. Mol.
Catal. A: Chem. 2002, 185, 179.
(
6) (a) Stanger, K. J.; Tang, Y.; Anderegg, J.; Angelici, R. J. J. Mol. Catal. A:
Chem. 2003, 202, 147. (b) Yang, H.; Gao, H.; Angelici, R. J. Organome-
tallics 2000, 19, 622. (c) Gao, H.; Angelici, R. J. Organometallics 1999,
1
14
spectra were computer simulated using the gNMR program. Diffuse
1
8
1
8, 989. (d) Gao, H.; Angelici, R. J. J. Mol. Catal. A: Chem. 1999, 145,
3. (e) Perera, M. A. D. N.; Angelici, R. J. J. Mol. Catal. A: Chem. 1999,
49, 99. (f) Gao, H.; Angelici, R. J. New J. Chem. 1999, 23, 633. (g) Gao,
reflectance infrared Fourier transform (DRIFT) spectra were recorded
on a Digilab FTS-60 instrument, equipped with a KBr beam-splitter,
H.; Angelici, R. J. J. Am. Chem. Soc. 1997, 119, 6937.
(
(
(
7) Yang, H.; Gao, H.; Angelici, R. J. Organometallics 1999, 18, 2285.
8) Gao, H.; Angelici, R. J. Organometallics 1998, 17, 3063.
9) (a) Bovkun, T. T.; Sasson, Y.; Blum, J. J. Mol. Catal. A: Chem. 2005,
(11) Bianchini, C.; Frediani, P.; Sernau, V. Organometallics 1995, 14, 5458.
(12) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. J. Am. Chem. Soc. 1951, 73,
373.
(13) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309.
(14) Budzelaar, P. H. M. gNMR V4.0; Cherwell Scientific Publishing: Oxford,
1995-1997 (Ivory Soft).
2
42, 68. (b) Ghattas, A.; Abu-Reziq, R.; Avnir, D.; Blum, J. Green Chem.
003, 5, 40.
2
(
10) Abu-Reziq, R.; Avnir, D.; Blum, J. J. Mol. Catal. A: Chem. 2002, 187,
2
77.
7066 J. AM. CHEM. SOC.
9
VOL. 128, NO. 21, 2006

Benzene Hydrogenation over RhI
−
Pd /SiO
0
Catalysts
A R T I C L E S
2
using a homemade reaction chamber with ZnSe windows and an Harrick
DRA-2C1 accessory.15 HRTEM micrographs were recorded on a JEOL
0
2
00EX instrument. Pd /SiO
2
, usually 1 g batches (Pd content ranging
4
I
from 1.99 to 9.85 wt %), and Rh /SiO
.7 wt %) were prepared following previously reported procedures.
2
(rhodium content from 0.3 to
1
d
1
Details of syntheses and acquisition of DRIFT spectra and HRTEM
micrographs are given as Supporting Information.
I
0
2
Preparation of Rh -Pd /SiO . The samples of this bimetallic
catalyst (0.26-1.35 wt % Rh and 1.99-9.85 wt % Pd) were prepared
following a procedure analogous to that reported above for the silica-
I
tethered rhodium(I) complex Rh /SiO
of the pretreated silica, Pd /SiO
2
using as support material, instead
(Pd content ranging from 1.99 to 9.85
wt %), previously passivated under O flow (40 mL/min) for 30 min
followed by flushing in argon flow (40 mL/min) for 30 min. The mild
0
2
Figure 1. Pd K-edge Fourier spectra (not phase-corrected) of (bold s)
0
0
I
0
2
Pd /SiO2, 1.99 wt % Pd; (s) Pd /SiO2, 9.85 wt % Pd; (bold ‚‚‚) Rh -Pd /
4
I
0
SiO2, 0.54-1.99 wt % Rh-Pd; (‚‚‚) Rh -Pd /SiO2, 0.46-9.85 wt %
Rh-Pd.
0
passivation of Pd /SiO
2
is a necessary step to allow for the grafting
of Rh(cod)(sulfos), as it takes place in the potentially reactive solvent
1
6
2 2
CH Cl . However, this does not represent a limitation to catalysis, as
the thin oxide layer is rapidly and completely removed under
hydrogenation conditions (in situ EXAFS evidence).
0
Preparation of Pd /NaY. Zeolite (NaY)-supported palladium nano-
0
particles (Pd /NaY), usually 1 g batches (ca. 2 wt % Pd), were obtained
by ionic exchange of NaY with Pd(NH
calcination at 250 °C for 2 h in O flow (40 mL/min) at a heating rate
of 1 °C/min, followed by cooling in argon flow (40 mL/min) to room
temperature; (2) reduction at 300 °C for 1 h in H flow (40 mL/min)
3 4 3 2
) (NO ) followed by (1)
2
2
at a heating rate of 10 °C/min followed by cooling in argon flow (40
mL/min) to room temperature. Ionic exchange was performed as
follows: 5 g of NaY was treated with 100 mL of a 0.01 M solution of
I
Figure 2. Rh K-edge Fourier spectra (not phase-corrected) of (‚‚‚) Rh ;
I
I
0
(
bold ‚‚‚) Rh /SiO2, 0.94 wt % Rh; (bold s) Rh -Pd /SiO2, 0.54-1.99 wt
Pd(NH )
3 4
3
(NO )
2
and stirred overnight at 70 °C. The suspension was
% Rh-Pd; (s) RhI-Pd0/SiO , 0.46-9.85 wt % Rh-Pd.
2
filtered, thoroughly washed, and dried overnight at 100 °C.
I
0
Preparation of Rh -Pd / NaY. This bimetallic catalyst was
prepared following a procedure analogous to that reported above for
I
0
0
Rh -Pd /SiO
2
, using as support material Pd /NaY (ca. 2 wt % Pd),
flow (40 mL/min) for 30 min followed
previously passivated under O
2
by flushing in argon flow (40 mL/min) for 30 min.
EXAFS Experiments. EXAFS experiments were conducted at the
EXAFS XAS13 beamline of the DCI synchrotron of LURE at Orsay,
with a double-crystal Ge(400) monochromator. The spectra were
recorded in transmission at the Pd K-edge and in fluorescence at the
Rh K-edge. The samples were loaded inside the sample holders under
a nitrogen atmosphere and measured at room temperature over a range
of 1000 eV after the edge, with a sampling step of 2 eV and 2-4 s
integration per point. Each spectrum was repeated three or four times
for the signal-to-noise ratio optimization and for the error bars
evaluation. Pd and Rh foils were employed as reference standards for
the extraction of phase and amplitude experimental functions. The single
scattering data analysis, used in the signal modeling at the Pd K-edge,
was performed with the software package of A. Michalowicz. Experi-
mental ø(k) functions were extracted from the absorption spectra with
a standard procedure and Fourier transformed over a k range of 3-15
Figure 3. Rh K-edge Fourier spectra (not phase-corrected) of (bold s)
I
Rh /SiO2, 0.94 wt % Rh, after in situ reduction with H2 flow at 130 °C;
I
0
(bold ‚‚‚) Rh -Pd /SiO2, 0.46-9.85 wt % Rh-Pd, after in situ reduction
I
with H2 flow at 130 °C; (‚‚‚) Rh /SiO2, 0.94 wt % Rh, after in situ reduction
I
0
with H2 flow at 300 °C; (s) Rh -Pd /SiO2, 0.46-9.85 wt % Rh-Pd, at
the end of a batch reaction of benzene hydrogenation (30 bar, 40 °C).
signal was detectable in the FFT of any samples beyond r ) 4 Å. Paths
with relative amplitude lower than 6% were neglected. The Rh and Pd
K-edge Fourier transform spectra (not phase corrected) are reported in
Figures 1-4. The best-fit values determined for the structural parameters
of each shell of every sample are reported in Tables 1-3.
-1
Å
. The main peaks of the Fourier transformed modulus were filtered
and analyzed with a nonlinear least-squares fit program which provides
the atomic species, the number N of nearest neighbors, their distance
R from the absorber atom, and the disorder factor σDW for each shell.
The analysis at the Rh K-edge was performed also by including the
simulation of the multiple scattering contributions present in the signals.
For this analysis the FEFF7 program package was used17 which
calculates, starting from a model structure, all the theoretical signals
produced by all the single and multiple scattering paths.
I
0
Heterogeneous Hydrogenation Reactions with Rh -Pd /SiO
2
as
Catalyst. The autoclave was charged with the appropriate amount of
I
0
a selected sample of Rh -Pd /SiO
toluene, and n-pentane (complement to 30 mL total volume). The
ensemble was pressurized with H to the desired pressure, heated to
0 °C, and then stirred at 1500 rpm. After 2 h, the autoclave was cooled
2
, the desired amount of benzene or
2
4
to ambient temperature and depressurized. The liquid contents were
analyzed by GC and GC/MS. Above 1500 rpm, the rates were
independent of the agitation speed at all the temperatures studied, thus
indicating the absence of mass transport resistance. For comparative
purposes, some hydrogenation reactions were also carried out using
3
The fits were performed on the unfiltered k -weighted spectra.
Scattering paths were included up to the distance of 4.5 Å, because no
(
15) Dal Santo, V.; Dossi, C.; Fusi, A.; Psaro, R.; Mondelli, C.; Recchia, S.
Talanta 2005, 66, 674.
(
16) Solymosi, F.; Rasko, J. J. Catal. 1995, 155, 74.
0
2 2
either Pd /SiO or Rh /SiO or Rh -Pd /NaY as catalysts. The stability
I
I
0
(
17) (a) Zabinsky, S. I.; Rehr, J. J.; Ankuninov, A.; Albers, R. C.; Eller, M. J.
Phys. ReV. B 1995, 52, 2995. (b) Mustre de Leon, J.; Rehr, J. J.; Zabinsky,
S. I.; Albers, R. C. Phys. ReV. B 1991, 44, 4146.
of the immobilized catalyst against leaching from the support was tested
by recycling experiments (see Supporting Information).
J. AM. CHEM. SOC.
9
VOL. 128, NO. 21, 2006 7067

A R T I C L E S
Barbaro et al.
with 12 nearest neighbors at 2.75 Å, on the assumption of
particles growth with the same packing and cubooctahedric
geometry, a mean diameter of about 10 Å was calculated for
the palladium particles.
0
In Pd /SiO2 (9.85 wt % Pd), the metal particles were still
small enough to exhibit a contribution from the surface oxygen
(
(
NPd-O ) 1.3), but with a larger Pd-Pd nearest-neighbors value
NPd-Pd ) 7.8) as compared to the 1.99 wt % sample, cor-
responding to an estimated mean diameter of 13 Å (Table 1).
I
0
An EXAFS analysis of Rh -Pd /SiO2 (0.46-9.85 wt %
Rh-Pd) at the Rh K-edge (Table 2, Figure 2) revealed that all
rhodium atoms on the support maintain the coordination
Figure 4. Rh K-edge Fourier spectra (not phase-corrected) of (‚‚‚)
I
0
I
Rh -Pd /SiO2, 0.46-9.85 wt % Rh-Pd, exposed to 1 bar CO after reduction
environment as in the free complex Rh (Table 3, Figure 2),
I
0
in H2 flow at 130 °C; (s) Rh -Pd /SiO2, 0.46-9.85 wt % Rh-Pd, exposed
to 5 bar CO after hydrogenation in an autoclave (30 bar, 40°C).
the four C atoms from cod and the three P atoms from the
triphosphine ligand being well resolved. An identical coordina-
tion geometry about rhodium was observed by EXAFS for the
Table 1. Best-Fit EXAFS Data at the Pd K-Edge
I
0
mixed-metal system Rh -Pd /SiO2 (0.54-1.99 wt % Rh-Pd)
r
σ
DW
10 Å)
-
2
I
sample
shell
N
(Å)
(×
(Table 2, Figure 2) as well as for Rh /SiO2 (0.94 wt % Rh)
0
(Table 3, Figure 2).
Pd /SiO2 (1.99 wt % Pd)
O
2.5 ( 0.5 1.98 ( 0.01 9.7 ( 0.4
Pd 5.8 ( 0.6 2.70 ( 0.006 8.8 ( 0.5
1.3 ( 0.2 1.97 ( 0.01 9.2 ( 0.2
Pd 7.8 ( 0.3 2.74 ( 0.002 7.9 ( 0.1
I
0
Irrespective of the metal loading in Rh -Pd /SiO2, contribu-
tions due to Pd-Rh couples were not observed at the Rh K-edge
0
Pd /SiO2 (9.85 wt % Pd)
O
(
Table 2, Figure 2), while rinsing the hybrid complex with neat
MeOH-d4 gave a solution of Rh (NMR evidence), which is
consistent with the integrity of the grafted molecular rhodium
I
0
I
Rh -Pd /SiO2
O
2.5 ( 0.4 1.98 ( 0.01 9.9 ( 0.3
(
I
0.54-1.99 wt % Rh-Pd) Pd 4.8 ( 0.8 2.70 ( 0.01 9.9 ( 0.1
0
Rh -Pd /SiO2
Pd 9.0 ( 0.1 2.70 ( 0.003 8.8 ( 0.1
1d
complex.
(
0.46-9.85 wt % Rh-Pd)
On the basis of the EXAFS data, one may therefore conclude
that Rh is anchored intact on the silica surface irrespective of
I
the palladium loading with no evidence of Pd-Rh interaction.
Preparation of Samples for EXAFS Measurements. Hydrogena-
tion Reactions in an Autoclave. (A) With No Substrate. A 100 mL
0
The Pd metal particle mean diameter measured by HRTEM
I
0
0
Parr autoclave was charged with 47 mg of Rh -Pd /SiO
wt % Rh-Pd) and n-pentane (30 mL). The ensemble was pressurized
with H to 30 bar, heated to 40 °C, and then stirred (1500 rpm). After
h, the autoclave was cooled to ambient temperature and depressurized.
2
(0.46-9.85
(Table 4) on Pd /SiO (1.99 wt % Pd) was 18.3 Å. Metal parti-
2
cles were visible both on the surface and on the edges of the
support grains that were totally amorphous. The particle distri-
bution function was sharp and symmetric, but the distribution
through the different support grains was not homogeneous. The
morphology and dispersion of the palladium particles, with a
2
2
The solid residue was separated by filtration over a sintered glass from
the liquid phase and washed with n-pentane (3 × 20 mL) under a
hydrogen pressure.
0
mean diameter of 20.5 Å, were almost unchanged in Pd /SiO2
(B) With No Substrate Followed by CO Pressurization. Following
(9.85 wt % Pd) as compared to the 1.99 wt % sample, while
a procedure analogous to that reported above, after the hydrogenation
phase, the autoclave was cooled to ambient temperature, depressurized,
and then pressurized to 5 bar with CO. After 30 min, the autoclave
was depressurized, and the solid residue was filtered off and washed
with n-pentane (3 × 20 mL) under CO.
the total number of particles on the surface was apparently
higher.
A large-spot (50 nm) energy-dispersive spectrometry (EDS)
I
0
analysis on Rh -Pd /SiO2 (1.08-1.99 wt % Rh-Pd) showed
the emission peaks of rhodium and palladium to be consistent
with a Pd/Rh molar ratio of 2 over the areas where the palladium
particles were visible by TEM (in perfect agreement with the
overall metal loading). When a region without visible metal
aggregates was shot, the EDS still detected both metals but in
a Pd/Rh molar ratio <1. Taking into account the molar weight
of rhodium, palladium, and SiO2, Rh-Pd loadings of 1.08-
1.99 wt % correspond to molar fractions of about 0.005 and
0.010, respectively, which is at the limit of the EDS sensitivity.
However, when small probe areas were selected around the
palladium particles, the molar fraction of supported Pd atoms
per volume unit could be increased so as to detect the
contribution from either metal. In this way, operating with an
electron probe spot of 5 nm centered on a single metal particle
(17 Å diameter), a molar ratio could be estimated. In particular,
since in fcc packing a palladium particle with a diameter of 17
Å has a surface atoms/total atoms ratio of 70%, a Pd/Rh value
of 10 corresponds to an average of one rhodium atom every
seven surface palladium atoms.
(
C) With Benzene as Substrate. The autoclave was charged with
I
0
4
7 mg of Rh -Pd /SiO
of 0.43 mL (4.81 mmol) of benzene in n-pentane (complement to 30
mL total volume). The ensemble was pressurized with H to 30 bar,
2
(0.46-9.85 wt % Rh-Pd) and a solution
2
heated to 40 °C, and then stirred (1500 rpm). After 2 h, the autoclave
was cooled to ambient temperature and depressurized. The liquid
contents were analyzed by GC and GC/MS. The grafted rhodium/
palladium product was separated by filtration over a sintered glass from
the liquid phase under nitrogen and washed with n-pentane (3 ×
2
0 mL).
Results and Discussion
EXAFS and HRTEM Characterization of Pd /SiO2,
0
I
I
0
0
Rh /SiO2, and Rh -Pd /SiO2. An EXAFS analysis of Pd /SiO2
1.99 wt % Pd) at the Pd K-edge showed a Pd-Pd nearest-
(
neighbors value (NPd-Pd) of 5.8 as well as a peak trend for the
next shells, consistent with the formation of small metal particles
dispersed on the support (Table 1, Figure 1). An appreciable
contribution from the surface oxygen atoms (NPd-O ) 2.5) was
also observed. As the palladium foil has a fcc crystal frame
7068 J. AM. CHEM. SOC.
9
VOL. 128, NO. 21, 2006

Benzene Hydrogenation over RhI
−
Pd /SiO
0
Catalysts
A R T I C L E S
2
Table 2. Best-Fit EXAFS Data at the Rh K-Edge for Bimetallic Samples
r
σ
DW
(×10 Å)
-
2
sample
shell
N
(Å)
I
0
Rh -Pd /SiO2 (0.54-1.99 wt % Rh-Pd)
C
P
4.0 ( 0.2
3.0 ( 0.2
2.19 ( 0.02
2.40 ( 0.01
8.1 ( 0.5
9.3 ( 0.5
I
0
Rh -Pd /SiO2 (0.46-9.85 wt % Rh-Pd)
C
P
4.0 ( 0.2
3.0 ( 0.2
2.19 ( 0.03
2.41 ( 0.02
8.1 ( 0.5
9.3 ( 0.5
I
0
Rh(H)x-Pd/SiO2 obtained by in situ reduction of Rh -Pd /SiO2
0.46-9.85 wt % Rh-Pd) with H2 flow at 130 °C
P
Pd
3.0 ( 0.2
3.4 ( 0.2
2.32 ( 0.02
2.77 ( 0.03
9.3 ( 0.5
9.8 ( 0.5
(
Rh(H)x(CO)y-Pd(CO)x/SiO2 obtained by in situ reduction of
C
P
Ob
Ol
Pd
1.4 ( 0.2
2.0 ( 0.2
0.9 ( 0.2
0.6 ( 0.2
2.1 ( 0.2
1.91 ( 0.01
2.34 ( 0.01
2.62 ( 0.03
2.92 ( 0.03
2.77 ( 0.02
9.9 ( 0.5
9.3 ( 0.5
5.8 ( 0.5
4.2 ( 0.5
9.9 ( 0.5
I
0
Rh -Pd /SiO2 (0.46-9.85 wt % Rh-Pd) with H2 flow
at 130 °C, followed by CO addition
I
0
Rh(H)x(CO)y-Pd(CO)x/SiO2 obtained by reduction of Rh -Pd /
SiO2 (0.46-9.85 wt % Rh-Pd) with H2 (30 bar) at 40 °C
in an autoclave followed by CO (5 bar)
C
P
Ob
Ol
Pd
1.4 ( 0.2
2.1 ( 0.2
1.0 ( 0.2
0.4 ( 0.2
1.8 ( 0.2
1.90 ( 0.04
2.31 ( 0.04
2.61 ( 0.02
2.93 ( 0.03
2.77 ( 0.03
9.9 ( 0.5
9.9 ( 0.5
6.2 ( 0.5
5.1 ( 0.5
9.1 ( 0.5
solid product obtained after catalytic hydrogenation (30 bar H2)
of benzene with Rh -Pd /SiO2 (0.46-9.85 wt % Rh-Pd)
at 40 °C in an autoclave
C
P
Pd
4.0 ( 0.2
1.6 ( 0.2
2.6 ( 0.2
2.08 ( 0.02
2.38 ( 0.02
2.76 ( 0.03
11.0 ( 0.5
9.7 ( 0.5
8.9 ( 0.5
I
0
Table 3. Best-Fit EXAFS Data at the Rh K-Edge for Monometallic
Samples
130 °C directly in the EXAFS cell (Table 3, Figure 3). An
EXAFS analysis showed unambiguously the rhodium atoms to
be surrounded by three phosphorus atoms with Rh-P distances
slightly shorter than those in the unsupported and supported
2
10- Å)
sample
shell
N
r (Å)
σ
DW
(×
RhI
C
P
C
P
P
4.1 ( 0.4 2.19 ( 0.01 9.1 ( 0.6
3.2 ( 0.3 2.49 ( 0.01 9.0 ( 0.8
4.0 ( 0.4 2.20 ( 0.01 6.2 ( 0.5
2.8 ( 0.1 2.43 ( 0.01 9.3 ( 0.9
3.0 ( 0.2 2.27 ( 0.01 9.9 ( 0.6
I
Rh complex (Table 3). Since no other atom was seen at a
I
Rh /SiO2 (0.94 wt % Rh)
bonding distance, one may conclude that, besides P atoms, the
rhodium single sites are very likely surrounded by hydrogen
atoms to complete their coordination sphere. In line with a
Rh(H)x/SiO2 obtained by
in situ reduction of
Rh /SiO2 (0.94 wt % Rh)
I
1d
previous study, the rhodium complex was removed from the
with H2 flow at 130 °C
Rh(H)x/SiO2 obtained by
in situ reduction of
silica surface by rinsing with MeOH-d4, and the product was
P
3.0 ( 0.2 2.27 ( 0.01 9.9 ( 0.6
III
authenticated by NMR as the binuclear Rh tetrahydride [RhH-
I
Rh /SiO2 (0.94 wt % Rh)
(
µ-H)(sulfos)]2. It has been also reported that the homogeneous
with H2 flow at 300 °C
+
hydrogenation of [Rh(triphos)(cod)] (triphos ) CH3C-
(
tetrahydride [Rh(H)2(H2)(triphos)] , which is unstable in both
the solid state and solution unless protected by a hydrogen
atmosphere, in the absence of which it spontaneously loses H2,
converting to the tetrahydride [RhH(µ-H)(triphos)]2
therefore possible that the hydrogenation of Rh /SiO2 follows
an identical path, leading to the formation of a silica-anchored
III
CH2PPh2)3) leads to the formation of the nonclassical Rh
0
Table 4. Mean Diameters of the Pd Particles by HRTEM and
+
Calculated by EXAFS
dmean (Å)
2
+ 18
sample
HRTEM
EXAFS
.
It is
I
0
Pd /SiO2 (1.99 wt %)
18.3 ( 1.9
20.5 ( 1.5
15.5 ( 1.1
13.7 ( 2.5
20.6 ( 2.0
11.8 ( 0.7
16.5 ( 1.3
10
13
8
0
Pd /SiO2 (9.85 wt %)
I
0
Rh -Pd SiO2 (0.54-1.99 wt % Rh-Pd)
III
Rh polyhydride, hereafter referred to as Rh(H)x/SiO2, which
I
0
Rh -Pd /SiO2 (1.08-1.99 wt % Rh-Pd)
I
0
may contain both classical and nonclassical hydride ligands
(Scheme 3).
Rh -Pd /SiO2 (0.46-9.85 wt % Rh-Pd)
18
0
Pd /NaY (2.01 wt %)
I
0
Rh -Pd NaY (0.15-2.01 wt % Rh-Pd)
I
0
In the case of the hybrid system Rh -Pd /SiO2 (0.46-9.85
wt % Rh-Pd), after reduction with H2 at 130 °C in the EXAFS
cell, besides appreciable contraction of the average Rh-P
distance, a relevant contribution from a palladium shell was also
detected (NRh-Pd ) 3.4). Apparently, this contribution origi-
nates from surface atoms of the palladium particles dispersed
on the support (Table 2, Figure 3). In principle, it is not pos-
sible to discriminate palladium from rhodium on the basis of
photoelectron scatterer, but the attribution of the Rh metal
shell to palladium neighbors was unambiguously supported
I
0
In Rh -Pd /SiO2 (0.54-1.99 wt % Rh-Pd), the particle size
distribution was sharp, with a mean diameter of 15.5 Å (Table
4
). No rhodium contribution was detected in the EDS signal,
even when the electron beam (5 nm spot) was focused on a
single Pd particle, which means that the rhodium molar fraction
was everywhere lower than the sensitivity limit.
Provided one considers that EXAFS tends to overestimate
the contribution of the smallest particles which, in contrast, are
generally underestimated by the TEM images, a good agreement
has been found between the mean dimensions estimated from
the EXAFS Pd-Pd nearest-neighbors numbers and the HRTEM
micrographs analysis.
I
by the evidence that Rh /SiO did not exhibit any Rh-Rh
2
contribution, even upon treatment with H up to 300 °C (Table
2
3, Figure 3).
The Rh-Pd distance of 2.77 Å, only slightly longer than that
of a Rh-Pd alloy (2.72 Å), and the number of the Rh-Pd
EXAFS Analysis of the Products Obtained by Hydroge-
I
I
0
nation of Rh /SiO2 and Rh -Pd /SiO2. Solid-Gas Condi-
(
18) Bakhmutov, V. I.; Bianchini, C.; Peruzzini, M.; Vizza, F.; Vorontsov, E.
I
tions. Rh /SiO2 (0.94 wt % Rh) was treated with a H2 flow at
V. Inorg. Chem. 2000, 38, 1655.
J. AM. CHEM. SOC.
9
VOL. 128, NO. 21, 2006 7069

A R T I C L E S
Scheme 3
Barbaro et al.
Scheme 4
either for Rh(H)x/SiO2 (see above) or for Rh(CO)2(sulfos)/
1d,4
SiO2. Actually, most of rhodium remained grafted to the silica
surface, and extraction with MeOH gave only a tiny amount of
an intractable material, containing palladium metal, phosphine
oxide, and unknown rhodium-sulfos species (NMR, IR, ICP-
AES). We therefore concluded that the hydrogenation of
I
0
Rh -Pd /SiO2 in n-pentane gives a product in which the
Rh-Pd interaction is strong enough to persist after treatment
with a polar solvent such as MeOH that it is known to break
the H-bonding interaction between sulfos and silica.
In an attempt at stabilizing the hydrogenation product of
Rh -Pd /SiO2, a sample with 0.46-9.85 wt % Rh-Pd was
hydrogenated in n-pentane in an autoclave (30 bar H2,
neighbors (3.4) point to rhodium and palladium atoms in close
proximity. Metal-metal bonds are frequently encountered in
organometallic and coordination compounds, where hydride
ligands bridge metal centers. Unfortunately, we did not find
evidence in the literature for bimetallic complexes containing
Rh(µ-H)xPd moieties, yet several compounds are known where
the Rh(triphos) fragment is connected to other metal frag-
ments (M ) Fe, Co, Ni, Rh) by double and triple hydride bridges
to give binuclear complexes containing short metal-metal
I
0
4
0 °C). CO (5 bar) was then admitted into the reactor. The
isolated product was studied by EXAFS. For comparative
purposes, an EXAFS study was carried out also on the
carbonylation product of Rh(H)x-Pd/SiO2, prepared in solid-
gas conditions in the EXAFS cell as previously described (Table
). Multiple scattering was employed for the carbonyl groups,
with a bond angle of 180° between the carbon and oxygen
atoms.
The spectra of the two products were extremely similar to
each other, which indicates an identical reaction path in either
phase system. The corresponding fits (Table 2, Figure 4) show
that, in either system, the rhodium atom is coordinated by two
phosphorus atoms and two palladium atoms (RRh-P ) 2.34 Å;
RRh-Pd ) 2.77 Å). Each rhodium atom is also bonded to an
average of 1.4 CO groups (RRh-C ) 1.91 Å, RRh-Ob ) 2.62 Å,
RRh-Ol ) 2.92 Å), yet the multiple scattering path contribution
from linear carbonyls corresponds to about 40% and 25% of
the total carbonyls in the molecule, respectively.
2
1
9
bonds. On the basis of the EXAFS measurements and of
previous reports on the homogeneous hydrogenation of
Rh-triphos and Rh-sulfos complexes, showing the formation of
2
0
Rh(µ-H)xRh binuclear species with metal-metal bonds, we
I
0
suggest that Rh -Pd /SiO2 reacts with H2 to give new surface
species where the rhodium single sites and the palladium
particles are connected by Rh-Pd bonds as well as bridging
hydrides (Scheme 4). This product is hereafter denoted as
Rh(H)x-Pd/SiO2.
Solid-Liquid Conditions. Since the catalytic hydrogenation
of arenes investigated in this work was carried out in solid-
liquid conditions, attempts were made at studying the product
I
0
The fact that only two phosphorus atoms are detected in the
rhodium coordination sphere suggests that, on the average, one
of the sulfos arm is free, which has several precedents in the
arising from the hydrogenation of Rh -Pd /SiO2 in n-pentane
in an autoclave (30 bar H2, 40 °C). Unfortunately, all our efforts
to intercept and characterize Rh(H)x-Pd/SiO2 from a solid-
liquid reaction were unsuccessful due to the extreme sensitivity
to oxygen of the isolated product. Moreover, treatment of the
latter with MeOH, even under a H2 atmosphere, did not allow
us to extract a well-defined rhodium-sulfos complex as occurs
2
0f,21,22
literature for tripodal triphosphine ligands.
Moreover,
the presence of only 1.4 carbonyl groups and the contribution
still relevant from the palladium shell at 2.77 Å show that
the interaction between rhodium and palladium, which origi-
nated during the hydrogenation process, is not removed by
carbonylation (in contrast to what happens to the cod ligand,
(
19) (a) Bianchini, C.; Laschi, F.; Masi, M.; Mealli, C.; Meli, A.; Ottaviani, M.
F.; Proserpio, D. M.; Sabat, M.; Zanello, P. Inorg. Chem. 1989, 28, 2552.
I
which is completely removed by CO from Rh /SiO2 at room
(b) Bianchini, C.; Meli, A.; Laschi, F.; Ramirez, J. A.; Zanello, P.; Vacca,
1
d
A. Inorg. Chem. 1988, 27, 4429. (c) Bianchini, C.; Meli, A.; Zanello, P.
Chem. Commun. 1986, 628. (d) Bianchini, C.; Mealli, C.; Meli, A.; Sabat,
M. Chem. Commun. 1986, 777.
temperature).
Some possible structures for the hydrogenation/carbonylation
(
20) (a) Barbaro, P.; Bianchini, C.; Meli, A.; Moreno, M.; Vizza, F. Organo-
metallics 2002, 21, 1430. (b) Bianchini, C.; Barbaro, P.; Macchi, M.; Meli,
A.; Vizza, F. HelV. Chim. Acta 2001, 84, 2895. (c) Bianchini, C.; Meli,
A.; Moneti, S.; Oberhauser, W.; Vizza, F.; Herrera, V.; Fuentes, A.;
Sanchez-Delgado R. A. J. Am. Chem. Soc. 1999, 121, 7071. (d) Bianchini,
C.; Meli, A.; Moneti, S.; Vizza, F. Organometallics 1998, 17, 2636. (e)
Bianchini, C.; Meli, A.; Patinec, V.; Sernau, V.; Vizza, F. J. Am. Chem.
Soc. 1997, 119, 4945. (f) Bianchini, C.; Meli, A.; Peruzzini, M.; Vizza, F.;
Frediani, P.; Ramirez, J. A. Organometallics 1990, 9, 226.
I
0
product of Rh -Pd /SiO2 (see also the IR study reported in the
following section), hereafter denoted as Rh(H)x(CO)y-Pd(CO)x/
SiO2, are reported in Scheme 5.
(21) Boxwell, C. J.; Dyson, P. J.; Ellis, D. J.; Welton, T. J. Am. Chem. Soc.
2002, 124, 9334.
(22) Kiss, G.; Horvath, I. Organometallics 1991, 10, 3798.
7070 J. AM. CHEM. SOC.
9
VOL. 128, NO. 21, 2006

Benzene Hydrogenation over RhI
−
Pd /SiO
0
Catalysts
A R T I C L E S
2
Scheme 5
IR Characterization of the Carbonylated Products. Since
CO is an excellent IR probe molecule, a DRIFT study has been
carried out on significant carbonylation products.
respectively, measured as area CObridging/area COlinear), one may
readily conclude that the palladium particles in the two samples
exhibit similar dispersion, which is consistent with the HRTEM
data.
0
Pd /SiO2 (1.99 and 9.85 wt % Pd), prepared directly in the
I
0
DRIFT cell by in situ hydrogenation of ex situ calcinated PdCl2/
SiO2, was reacted at room temperature with a flow of 1 bar CO
for 30 min. The DRIFT spectra, acquired after the cell was
purged with He, are shown in Figure 5.
Samples of Rh -Pd /SiO2 (0.54-1.99 wt % Rh-Pd and
0.46-9.85 wt % Rh-Pd) were treated in the DRIFT cell with
H2 at 120 °C and then reacted at room temperature with 1 bar
CO. DRIFT spectra, taken after the cell was purged with He,
showed a complex system of bands located in the region
-1
between 2150 and 1900 cm (Figure 6a,b) that may be ascribed
to CO bonded both to palladium particles and rhodium single
sites. Traces a and b in Figure 6 were well fitted by a
superposition of mixed Gaussian/Lorentzian bands, and contri-
butions from different species were put in evidence (fitting bands
not showed).
0
Figure 5. DRIFT spectra of Pd /SiO2 (9.85 wt % Pd, trace a; 1.99 wt %
Pd, trace b), after hydrogenation at 120 °C, followed by exposure to 1 bar
CO flow for 30 min and purging in He flow for 1 min.
Both samples showed bands located between 2150 and 1900
cm- due to CO adsorbed on Pd particles. The 1.99 wt %
1
0
-
1
sample showed a band located at 2096 cm which is a
convolution of three components with maxima at 2108, 2099,
and 2080 cm . These bands can be attributed to CO adsorbed
-
1
Figure 6. DRIFT spectra of Rh(H)x(CO)y-Pd(CO)x/SiO2 after hydrogena-
0
I
0
linearly on different sites of Pd crystallites, while a broad band
tion of Rh -Pd /SiO2 at 120 °C, followed by exposure at room tempera-
-1
23
ture to 1 bar CO flow for 30 min and purging in He flow for 1 min: (a)
located at 1977 cm may be assigned to bridging CO groups.
A band at 2096 cm-1 was also present in the spectrum of the
.85 wt % sample, but the three previous components were less
evident, while the broad band attributable to bridging CO was
shifted to 1991 cm . The intensity ratio of the bands due to
bridging and terminal CO has been reported to depend on the
0
.54-1.99 wt % Rh-Pd, (b) 0.46-9.85 wt % Rh-Pd. For comparison,
the IR spectrum of the sample obtained after hydrogenation/carbonylation
of Rh /SiO (0.94 wt % Rh) is reported (c).
I
9
2
-
1
-1
The bands at 2100 and 1950 cm (0.54-1.99 wt %
-
1
Rh-Pd) and at 2100 and 1968 cm
(0.46-9.85 wt %
2
4
metal dispersion. Since quite similar ratios have been found
for the samples with 1.99 and 9.85 wt % Pd (2.62 and 2.51,
Rh-Pd) are attributable to CO bonded to palladium particles
in linear and bridging fashion, respectively. However, the for-
(
23) Liotta, L. F.; Martin, G. A.; Deganello, G. J. Catal. 1996, 164, 322 and
(24) Sheu, L. L.; Sachtler, W. M. H J. Mol. Catal. 1993, 81, 267 and references
references therein.
therein.
J. AM. CHEM. SOC.
9
VOL. 128, NO. 21, 2006 7071

A R T I C L E S
Barbaro et al.
mation of Rh(µ-CO)Pd moieties cannot be excluded. The bands
Scheme 6
-
1
located at 1997 and 2063 cm are readily attributable to gem-
I
dicarbonyl Rh species, as they are present also in the spectra
1
1
1d
of unsupported and silica-supported Rh(CO)2(sulfos).
The band at ca. 2030 cm- has been never observed on freshly
hydrogenated/carbonylated samples of Rh /SiO2 (Figure 6c),
where only the two bands attributable to gem-dicarbonyl Rh
species are present. Since the EXAFS analysis and the catalytic
data led us to exclude the decomposition of the rhodium
complex to metallic rhodium particles, this band is not attribut-
1
I
I
11
0
bonded to rhodium may be either the C4 portion of a benzene
molecule or an intermediate species along the reduction of
benzene to cyclohexa-1,3-diene. In the former case, it is very
likely that the C2 portion of benzene is bonded to palladium;
however, that cannot be proved by EXAFS at the Pd K-edge.
On the other hand, Rh , either free or immobilized on silica,
and its cationic derivative, [Rh(triphos)(cod)] , are unable to
coordinate benzene in any fashion, nor are they able to catalyze
benzene hydrogenation to any extent. As it will be discussed
below, benzene hydrogenation by late transition metal com-
plexes stabilized by phosphines generally requires the activation
of the arene to the η coordination mode, which is apparently
an inaccessible activation path to the [Rh(triphos)] fragment.
Moreover, hydride(s) transfer from metal to η -benzene has been
never reported to proceed via a C4 ligand, allyl-type and ene
ligands (η ) being invariably observed.
reasoning, which will be further developed in a following
section, we propose that the product intercepted at the end of
the batch catalytic reaction has the structure shown in Scheme
able to CO linearly bonded to Rh sites and therefore can be
ascribed to a new Rh-COterminal species. An in-depth inspection
of the silanols’ O-H stretching bands did not reveal any
appreciable variation between spectra recorded under He before
and after the treatment with CO.
2
5
I
In conclusion, the DRIFT spectra confirm that the carbony-
+
lation of Rh(H)x-Pd/SiO2 results in a variety of CO bonding
I
modes and also that the hydrogenation of Rh -Pd/SiO2 leads
4
to the formation of surface species where rhodium single sites
and palladium particles are interacting with each other.
I
0
Zeolite as Support for Rh and Pd . Palladium nanoparticles
were supported on a zeolite NaY material which would prevent
the diffusion of the large rhodium complex inside its channels,
where the palladium particles are preferentially located. The
4
3a
+
25
4
0
HRTEM images taken on Pd /NaY (2.01 wt %) showed a very
3
25,26
On the basis of this
narrow particle size distribution function, with a mean value of
1
1.8 Å (Table 4). All support grains are cubic (with well-
resolved channel diffraction fringes of 13 Å) and exhibit a
perfectly uniform dispersion of the metal particles. No palladium
particles were visible at the external surface of the zeolite, while
some palladium particles were visible along the external edge
of the zeolite grains.
6
, where a benzene molecule is simultaneously activated by
palladium and rhodium.
Notably, a Pd K-edge EXAFS analysis on Rh(C4)(H)x-
Pd(C2)/SiO2 showed a Pd-Pd coordination number (N ) 9)
coincident with that of the catalyst precursor Rh -Pd /SiO2 (see
Table 1), which confirms the stability of the palladium particles
under hydrogenation conditions.
0
I
After impregnation of Pd /NaY with Rh , HRTEM analysis
I
0
I
0
of the resulting hybrid product Rh -Pd /NaY (0.15-2.01 wt
Rh-Pd) showed a slightly larger size distribution of the Pd
%
particles (average diameter 16.5 Å) as compared to the sample
with no tethered rhodium complex (Table 4). No rhodium
contribution was visible in the EDS signal by focusing the
electron beam on single Pd particles inside the zeolite, while
the emission signal of rhodium was detected on the external
surface of the support. This experimental observation is clear
evidence that the rhodium complex is preferentially located at
the external surface of the zeolite.
Benzene and Toluene Hydrogenation Catalyzed by
Rh -Pd /SiO2 and Rh -Pd /NaY. All previously reported
I
0
I
0
studies on the hydrogenation of benzene and toluene catalyzed
I
0
by Rh -Pd /SiO2 were carried out with only one type of catalyst
0.56-9.86 wt %) at a H2 pressure of 30 bar and fixed
(
4
concentrations of substrate and catalyst amount. To complement
the previous data and, most importantly, to obtain information
on the reaction mechanism, especially as regards the synergistic
cooperation between the rhodium single sites and the palladium
particles, we decided to perform a more systematic study where
EXAFS Analysis after a Batch Hydrogenation of Benzene
I
0
Catalyzed by Rh -Pd /SiO2. A Rh K-edge EXAFS analysis
on the product obtained at the end of a batch reaction of benzene
H pressure, substrate concentration, rhodium-to-palladium ratio,
and overall catalyst amount are systematically varied.
2
I
0
hydrogenation catalyzed by Rh -Pd /SiO2 (0.46-9.85 wt %)
in n-pentane (Table 2, Figure 3) showed the presence of four
carbon atoms at a distance of 2.08 Å from the rhodium center,
which is also connected to 1.6 phosphorus atoms at a distance
of 2.38 Å and to 2.6 palladium atoms at a distance of 2.76 Å.
This means that the solid product which separates from the
catalytic mixture contains rhodium atoms coordinated by a
dihapto sulfos ligand and by a C4 ligand. As previously
mentioned, tripodal triphosphines, such as triphos, can act as
bidentate ligands, with an unfastened P donor atom, in conjunc-
Tables 5-7 report the data obtained for the hydrogenation
I
0
of benzene to cyclohexane in the presence of Rh -Pd /SiO ,
2
while Table 8 summarizes the results for toluene. Partial
hydrogenation products were not detected at any stage of the
reactions. All reactions were performed in n-pentane at 40 °C.
The catalytic activity, in terms of dependence of substrate con-
(
25) Bianchini, C.; Caulton, K. G.; Folting, K.; Meli, A.; Peruzzini, M.; Polo,
A., Vizza, F. J. Am. Chem. Soc. 1992, 114, 7290.
(26) (a) Bleeke, J. R.; Muetterties, E. L. J. Am. Chem. Soc. 1981, 103, 556. (b)
Stuhl, L. S.; Rakowski Du Bois, M. C.; Hirsekorn, F. J.; Bleeke, J. R.;
Stevens, A. E.; Muetterties, E. L. J. Am. Chem. Soc. 1978, 100, 2405. (c)
Rakowski, M. C.; Hirsekorn, F. J.; Stuhl, L. S.; Muetterties, E. L. Inorg.
Chem. 1976, 15, 2379. (d) Muetterties, E. L.; Hirsekorn, F. J. J. Am. Chem.
Soc. 1974, 96, 4063. (e) Muetterties, E. L.; Hirsekorn, F. J. J. Am. Chem.
Soc. 1974, 96, 7920.
tion with Group 8 and 9 metals, when the coligands occupy
more than three coordination positions.2
1,22
Since the GC
analysis of the liquid phase after catalysis showed the exclusive
presence of unreacted benzene and cyclohexane, the C4 fragment
7072 J. AM. CHEM. SOC.
9
VOL. 128, NO. 21, 2006

Benzene Hydrogenation over RhI
−
Pd /SiO
0
Catalysts
A R T I C L E S
2
I
0
Table 5. Hydrogenation of Benzene to Cyclohexane Catalyzed by
Rh -Pd /SiO2 were examined by NMR spectroscopy and
I
0
Rh -Pd /SiO2 (0.46-9.85 wt %): Dependence on Hydrogen
13
1
GC/MS. The C{ H} NMR spectrum of the product obtained
with Pd /SiO2 (Figure 11a) showed the formation of several
Pressurea
0
H
2
pressure
(bar)
yield (%);b
TOF (
10)c
regioisomers C6H6+xD6-x, which is consistent with the fact
that palladium particles, besides hydrogenating benzene to
cyclohexane, disproportionate cyclohexa-1,3-diene to ben-
zene and cyclohexene. This means that also C-H/C-D
exchange is achieved (GC/MS evidence). The major product
was the cis-trans regioisomer, followed by the all-cis regio-
entry
×
1
2
3
4
5
15
30
45
40; 9.62
42; 10.10
40; 9.62
41; 10.10
2
7
a
-3
Experimental conditions: catalyst precursor, 47 mg (Rh, 2.1 × 10
1
3
1
-3
isomer (Chart 1). The C{ H} NMR spectrum of the product
mmol; Pd, 43.5 × 10 mmol); benzene, 0.43 mL (4.81 mmol); benzene/M
ratio, 111 (Pd)-2290 (Rh); n-pentane, complement to 30 mL total volume;
temperature, 40 °C; time, 2 h; stirring rate, 1500 rpm. Average values
I
0
obtained with Rh -Pd /SiO2 was quite similar except for a
higher proportion of the all-cis isotopomer (Figure 11b).
The regiochemistry of benzene-d6 hydrogenation by hetero-
genized single-site catalysts has been previously studied for
organoactinides and organozirconium complexes.28 In either
case, mixtures of all-cis and cis-trans isotopomers were
obtained, with a prevalence of the latter. In contrast, the
homogeneous hydrogenation of benzene catalyzed by (η -C3H5)-
Co(PR3)3 (PR3 ) phosphite, phosphine) has been found to give
exclusively the all-cis product. The complexed isotopomer
distribution obtained with Rh -Pd /SiO2 can be accounted for
by the concomitant presence of the palladium nanoparticles, yet
the increased proportion of the all-cis product suggests that the
rhodium single sites may have a their own role, similar to that
of the cobalt complex, along the hydrogenation path down to
cyclohexane, once benzene is activated by palladium (see
below).
Proposed Catalytic Mechanism. A number of key points
have been unambiguously established in the course of the
experimental studies reported in this paper, which can be
summarized as follows:
b
c
-1
over at least three runs. In mmol of product h
.
version on H2 pressure, arene concentration, and overall metal
I
0
amount in the catalyst, was evaluated by using a Rh -Pd /SiO2
sample containing ca. 0.5 wt % Rh and ca. 10 wt % Pd. The
dependence of substrate conversion on either rhodium or
palladium amount in the catalyst precursor was evaluated with
3
I
0
Rh -Pd /SiO2 samples containing from 0.26 to 1.35 wt % Rh
and from 1.99 to 4.87 wt % Pd.
2
9
I
0
From a perusal of the data obtained by varying the H2 pressure
(Table 5), one may readily infer that the hydrogenation of
benzene is zero-order in hydrogen concentration in the pressure
range from 5 to 45 bar. A plot of the turnover frequency (TOF)
vs substrate concentration showed instead a linear correlation
for benzene concentrations lower than 1 M (Table 6, Figure 7).
For higher substrate concentrations, a saturation effect apparently
occurred, indicated by a plateau in activity.
The TOF tended to a plateau also at overall catalyst amounts
-
3
higher than 30 × 10 Pd + Rh mmol (Table 6, Figure 8).
However, a linear correlation was observed at lower catalyst
amounts.
I
0
(i) The much faster hydrogenation rate of Rh -Pd /SiO2 as
0
compared to Pd /SiO2 is due not to a hydrogen spill-over effect,
Finally, the catalytic activity was evaluated as a function of
the relative amount of either rhodium or palladium in the
catalysts, keeping H2 pressure (30 bar) and benzene concentra-
tion (0.16 M) constant (Table 7). As is straightforwardly shown
in Figures 9 and 10, a linear correlation between TOF and metal
content was found for either metal, with a major slope exhibited
by the rhodium graph. Quite similar results have been obtained
in the hydrogenation of toluene, the only difference being that
6
b,g
as suggested by some authors, but to the concomitant action
of the two metals on the substrate. In particular, the hydrogena-
tion of either metal and the transfer of hydrogen atoms to the
substrates is not rate limiting, while both the substrate concen-
tration and the amount of either catalyst determine the reaction
rate.
(ii) The rhodium single sites must be in close contact with
the palladium particles to exert a beneficial effect on the
hydrogenation rate; in the absence of this requirement, as occurs
it was reduced at a slower rate than benzene, most likely due
to steric effects (Table 8).4
Interestingly, when the zeolite-supported catalyst Rh -Pd /
NaY (0.15-2.01 wt % Rh-Pd) was substituted for Rh -Pd /
SiO2 in a standard hydrogenation reaction of benzene (30 bar
H2, 40 °C, 47 mg of catalyst, 0.43 mL of benzene, 2 h,
n-pentane), the activity (4% conversion, TOF ) 11) was
I
0
for Rh -Pd /NaY, no rate increase is observed. In particular,
there is evidence that rhodium and palladium atoms are linked
to each other by a direct metal-metal bond throughout the
catalytic cycle. Indirect evidence has also been obtained for the
presence of hydride ligands bridging rhodium and palladium.
(iii) The ligand sulfos maintains a high degree of flexibility,
even when grafted to the support, allowing the fastening/
unfastening of a phosphine arm, which is certainly important
for the coordination of a multi-hapto substrate.
I
0
I
0
0
identical to that obtained with a pure sample of Pd /NaY under
comparable experimental conditions, with no appreciable effect
of the rhodium single sites. No positive effect on the conversion
of benzene was similarly reported for hydrogenation reactions
0
(iv) The termination metal product after a catalysis run
contains rhodium centers bonded to four carbon atoms, besides
phosphorus and palladium atoms. Whether this species has the
catalyzed by Rh(CO)2(sulfos)-Pd /SiO2, which differs from
I
0
Rh -Pd /SiO2 in having two CO co-ligands in the place of
cod.1d,4
This evidence confirms that the tethered rhodium
complex must bear an easily reducible co-ligand to generate a
better catalyst than Pd /SiO2 and also rules out that the simple
(
27) (a) Bunjes, A.; Eilks, I.; Pahlke, M.; Ralle, B. J. Chem. Educ. 1997, 74,
0
1323. (b) Cerveny, L.; Kluson, P.; Paseka, I. React. Kinet. Catal. Lett. 1991,
43, 533. (c) Cerveny, L.; Rehurkova, S. Collect. Czech. Chem. Commun.
1987, 52, 2909.
presence of phosphine ligands on the support may increase the
0
catalytic activity of the Pd particles.
(28) (a) Ahn, H.; Nicholas, C. P.; Marks, T. J. Organometallics 2002, 21 1788.
(b) Eisen, M. S.; Marks, T. J. J. Am. Chem. Soc. 1992, 114, 10358.
Benzene Hydrogenation Regiochemistry. The reduction
(29) Muetterties, E. L.; Rakowski, M. C.; Hirsekorn, F. J.; Larson, W. D.; Bauss,
0
products of C6D6 with H2 in the presence of either Pd /SiO2 or
V. J.; Anet, F. A. L. J. Am. Chem. Soc. 1975, 97, 1266.
J. AM. CHEM. SOC.
9
VOL. 128, NO. 21, 2006 7073

A R T I C L E S
Barbaro et al.
I
0
Table 6. Hydrogenation of Benzene to Cyclohexane Catalyzed by Rh -Pd /SiO2 (0.39-9.85 wt %): Dependence on the Catalyst Amount
a
and the Substrate Concentration
catalyst
Pd
103 mmol)
Rh
103 mmol)
benzene/Pd
ratio
benzene/Rh
ratio
benzene
yield (%);b
TOF (
10)c
entry
amount (mg)
(
×
(
×
(mL/mmol)
×
1
2
3
4
5
6
7
8
12
23
47
93
23
23
23
23
11.1
21.3
43.5
86.1
21.3
21.3
21.3
21.3
0.5
0.9
1.8
3.5
0.9
0.9
0.9
0.9
2015
1051
514
260
263
515
1997
3995
49081
25607
12531
6333
2.00/22.38
2.00/22.38
2.00/22.38
2.00/22.38
0.50/5.59
0.98/10.97
3.80/42.52
7.60/85.04
11; 12.31
18; 20.14
27; 30.21
42; 47.00
31; 8.67
27; 14.80
12; 25.51
6; 25.51
6402
12548
48654
97308
a
Experimental conditions: n-pentane, complement to 30 mL total volume; H2 pressure, 30 bar; temperature, 40 °C; time, 2 h; stirring rate, 1500 rpm.
b
c
-1
Average values over at least three runs. In mmol of product h
.
I
0
a
Table 7. Hydrogenation of Benzene to Cyclohexane Catalyzed by Rh -Pd /SiO2: Dependence on Rhodium and Palladium Contents
Pd content
(wt %)
Rh content
(wt %)
Pd
103 mmol)
Rh
103 mmol)
benzene/Pd
ratio
benzene/Rh
ratio
yield (%);b
TOF (
10)c
entry
catalyst
(
×
(
×
×
0
1
2
3
4
5
6
7
8
9
Pd /SiO2
1.99
8.8
547
3; 0.72
0; 0.00
7; 1.68
15; 3.61
18; 4.33
25; 6.01
33; 7.94
10; 2.41
13; 3.13
16; 3.85
I
Rh /SiO2
0.53
0.26
0.54
0.64
1.08
1.35
0.26
0.26
0.26
2.4
1.2
2.5
2.9
4.9
6.2
1.2
1.2
1.2
1688
4052
1951
1646
975
I
0
Rh -Pd /SiO2
1.99
1.99
1.99
1.99
1.99
2.83
4.10
4.87
8.8
8.8
8.8
8.8
8.8
12.5
18.1
21.5
547
547
547
547
547
385
266
224
I
0
Rh -Pd /SiO2
I
0
Rh -Pd /SiO2
I
0
Rh -Pd /SiO2
I
0
Rh -Pd /SiO2
780
I
0
Rh -Pd /SiO2
4052
4052
4052
I
0
Rh -Pd /SiO2
I
0
10
Rh -Pd /SiO2
a
Experimental conditions: catalyst precursor, 47 mg; benzene, 0.43 mL (4.81 mmol); n-pentane, complement to 30 mL total volume; H2 pressure, 30 bar;
temperature, 40 °C; time, 2 h; stirring rate, 1500 rpm. Average values over at least three runs. c In mmol of product h
b
-1
.
I
0
a
Table 8. Hydrogenation of Toluene to Methylcyclohexane Catalyzed by Rh -Pd /SiO2: Dependence on Rhodium and Palladium Contents
Pd content
(wt %)
Rh content
(wt %)
Pd
103 mmol)
Rh
103 mmol)
toluene/Pd
ratio
toluene/Rh
ratio
yield (%)b
TOF (
10)c
entry
catalyst
(
×
(
×
×
0
1
2
3
4
5
6
7
8
9
Pd /SiO2
1.99
8.8
534
1; 0.23
0; 0.00
5; 1.17
10; 2.35
13; 3.05
17; 3.99
23; 5.40
8; 1.88
I
Rh /SiO2
0.53
0.26
0.54
0.64
1.08
1.35
0.26
0.26
0.26
2.4
1.2
2.5
2.9
4.9
6.2
1.2
1.2
1.2
1688
3953
1903
1606
952
I
0
Rh -Pd /SiO2
1.99
1.99
1.99
1.99
1.99
2.83
4.10
4.87
8.8
8.8
8.8
8.8
8.8
12.5
18.1
21.5
534
534
534
534
534
376
259
218
I
0
Rh -Pd /SiO2
I
0
Rh -Pd /SiO2
I
0
Rh -Pd /SiO2
I
0
Rh -Pd /SiO2
761
I
0
Rh -Pd /SiO2
3953
3953
3953
I
0
Rh -Pd /SiO2
11; 2.58
13; 3.05
I
0
10
Rh -Pd /SiO2
a
Experimental conditions: catalyst precursor, 47 mg; toluene, 0.5 mL (4.69 mmol); n-pentane, complement to 30 mL total volume; H2 pressure, 30 bar;
temperature, 40 °C; time, 2 h; stirring rate, 1500 rpm. Average values over at least three runs. c In mmol of product h
b
-1
.
I
0
Figure 7. Hydrogenation reactions of benzene catalyzed by Rh -Pd /SiO2
Table 6, entries 2, 5-8). Dependence of the rate on benzene concen-
I
0
Figure 8. Hydrogenation reactions of benzene catalyzed by Rh -Pd /SiO2
(
(
Table 6, entries 1-4). Dependence of the rate on catalyst amount.
tration.
4
2
4
µ(η Rh-η Pd) benzene structure proposed in Scheme 6 is a
diene reduction at the rhodium single site. In this respect, the
increased proportion of all-cis cyclohexane in the reduction of
matter of troublesome assessment. However, this is quite likely
as, if it were a Rh-η -cylohexa-1,3-diene product, it would be
4
I
0
0
C6D6 with H2 on Rh -Pd /SiO2, as compared to Pd /SiO2, is
very difficult to intercept due to the fast kinetics of cyclohexa-
an indirect proof that, once benzene is concomitantly activated
7074 J. AM. CHEM. SOC.
9
VOL. 128, NO. 21, 2006

Benzene Hydrogenation over RhI
−
Pd /SiO
0
Catalysts
A R T I C L E S
2
Scheme 7
Figure 9. Hydrogenation reactions of benzene (a) and toluene (b) catalyzed
I
0
by Rh -Pd /SiO2 (Tables 7 and 8, entries 3-7). Dependence of the rate on
rhodium content.
ties.26 This author was able to show that the true catalyst is a
Co hydride with two phosphine ligands, which coordinates
Figure 10. Hydrogenation reactions of benzene (a) and toluene (b)
catalyzed by Rh -Pd /SiO2 (Tables 7 and 8, entries 3, 8-10). Dependence
of the rate on palladium content.
I
0
4
4
benzene in the η -fashion, to give a transient CoH(η -C6H6)-
PR3)2 complex, followed by intramolecular hydride transfer to
(
3
give the allylic intermediate Co(η -C3H7)(PR3)2. Hydrogen
addition has been proposed to generate an η -cyclohexa-1,3-
4
diene complex that ultimately releases cyclohexane via H2
addition/hydride migration steps (Scheme 7a).
The mechanism of benzene hydrogenation demonstrated for
I
the cobalt complex is apparently inaccessible to the “Rh (sulfos)”
fragment, either free or supported, due to its inability to activate
4
benzene in the η -fashion, at least within the temperature range
that precedes its decomposition. Only when the benzene ligand
4
is artificially created around the metal center, as in [Ir(η -C6H6)-
1
3
1
Figure 11. C{ H} NMR spectrum (21 °C, C6D6, 100.613 MHz) in the
+
(
triphos)] , obtained by cyclotrimerization of acetylene, has the
cyclohexane region of the mixture of products obtained by hydrogenation
4
-
η -coordination of benzene been observed for a 16e metal
fragment. However, this Ir complex is not a catalyst for
0
I
0
of C6D6 in the presence of (a) Pd /SiO2 and (b) Rh -Pd /SiO2. Representa-
tive resonances attributed to all-cis, cis-trans products and C-H/C-D
exchange products are marked with #, *, and °, respectively.
2
5
benzene hydrogenation. Only through separate multi-additions
-
+
4
of either H or H could the benzene ligand in [Ir(η -C6H6)-
(
Chart 1
+
triphos)] be selectively reduced to cyclohexene (Scheme 7b).
Nevertheless, it is interesting and thought-provoking to notice
that the Ir and Co catalytic cycles share most of the same
intermediates, which suggests an analogous mechanism for
I
grafted “Rh (sulfos)”, which is isostructural and isoelectronic
I
to “Ir (triphos)”, provided that palladium activates the substrate
by the two metals, the rhodium single sites can then bring about
and the Rh fragment.
by themselves the stepwise reduction by stereoselective H2 cis
From a comparison of the reaction sequences for Co and Ir,
one may also realize that either system promotes the hydrogena-
tion of benzene through intermediates that bear a single hydride
ligand. This commonality suggests that, in the benzene hydro-
genation catalyzed by Rh -Pd /SiO2, the Pd nanoparticles may
also control the number and nature of the hydride ligands at
the Rh site, possibly generating more reactive monohydride
additions.2
0b,26,29
On the other hand, a relevant contribution to
the catalysis would be still made by palladium, especially as
4
regards the hydrogenation of cyclohexene.
I
0
At this stage, it is worth commenting on a well-documented
3
case of homogeneous hydrogenation of benzene by (η -C3H5)-
Co(PR3)3 (PR3 ) phosphite, phosphine) described by Muetter-
J. AM. CHEM. SOC.
9
VOL. 128, NO. 21, 2006 7075

A R T I C L E S
Scheme 8
Barbaro et al.
species. This reaction opportunity is well known in homoge-
neous hydrogenation reactions, where the dihydride systems
Conclusions
This paper reports an in-depth study of benzene hydrogenation
over a silica-supported catalyst that combines a grafted rhodium
complex and palladium nanoparticles. In situ and ex situ EXAFS
and DRIFT measurements, batch catalytic reactions under
different conditions, deuterium labeling experiments, and model
organometallic studies, taken together, have provided valuable
information on the mechanism by which rhodium single sites
and palladium nanoparticles cooperate with each other in
promoting the hydrogenation of arenes.
The rhodium complex and the palladium particles form a
unique, stable entity throughout the catalytic cycle. Besides
decreasing the extent of cyclohexa-1,3-diene disproportionation
at palladium, the combined action of the two metals activates
the arene so as to allow the rhodium sites to enter the catalytic
cycle and speed up the overall hydrogenation process by rapidly
reducing benzene to cyclohexa-1,3-diene.
+
2+
[
Rh(H)2(sulfos)], [Rh(H)2(triphos)] , and [Ru(H)2(triphos)]
can be converted to the more active monohydride systems
-
+
[
RhH(sulfos)] , [RhH(triphos)], and [RuH(triphos)] , respec-
1a,20,30
tively, by deprotonation with strong bases.
In fact, the
hydride migratory addition to C-C double bonds is generally
easier for Rh monohydrides than for Rh dihydrides, as the latter
can undergo competitive H2 reductive elimination.
On the basis of the new experimental data described in this
work, incorporated with the data previously reported and with
relevant literature reports, a full mechanism for the hydrogena-
I
0
tion of benzene by Rh -Pd /SiO2 can therefore be proposed
(Scheme 8).
The overall mechanism comprises three distinct reaction paths
that reflect the concomitant presence on the support material
I
of chemically interacting Rh single sites and palladium nano-
particles (a), isolated palladium nanoparticles (b), and isolated
Acknowledgment. The LURE synchrotron laboratory in
Orsay (Paris, France) and the staff of the XAS-13 beamline are
gratefully acknowledged for the support and technical assistance.
We also thank the Ministry of University and Research of Italy
I
Rh single sites (c). The catalytic action of the latter two is well
0
established: Pd is able to reduce benzene to cyclohexane with
a mechanism that involves disproportionation of the cyclohexa-
I
1
,3-diene product and fast cyclohexene hydrogenation; Rh is
(MIUR) for financial support (FISR project: Nanotecnologie
0
faster than Pd in reducing cyclohexa-1,3-diene, yet slower in
the conversion of cyclohexene to cyclohexane. It is our opinion
molecolari per l’immagazzinamento e la trasmissione delle
informazioni) and the European Commission for financing the
Network of Excellence IDECAT (contract nï. NMP3-CT-2005-
4
that the combined action of the two catalysts compensates for
I
0
each individual weak point: upon hydrogenation of Rh -Pd /
SiO2, the Rh single site and the Pd particle form a unique, stable
heterobimetallic catalyst in which Pd adsorbs and activates ben-
0
11730).
0
Supporting Information Available: Syntheses of Pd /SiO2
I
4
and Rh /SiO2, recycling procedures of the heterogeneous
catalysts, and details of acquisition of DRIFT spectra and
HRTEM micrographs. This material is available free of charge
via the Internet at http://pubs.acs.org.
zene so as to allow its η -coordination to Rh (favored by un-
fastening of a phosphine arm), followed by reduction to cyclo-
hexa-1,3-diene. Fast hydrogenation of cyclohexa-1,3-diene
(prevalently at the Rh site) would also prevent the undesired
disproportionation at Pd, and the resulting cyclohexene would
be hydrogenated to cyclohexane by either metal (prevalently
on Pd).
JA060235W
(
30) Bianchini, C.; Casares, J. A.; Meli, A.; Sernau, V.; Vizza, F., Sanchez-
Delgado R. A. Polyhedron 1997, 16, 3099.
7076 J. AM. CHEM. SOC.
9
VOL. 128, NO. 21, 2006