Synthesis and reactivity of heterometallic RhO-M (M = Si, Ti) complexes: Memory effect in their catalytic performance in CO hydrogenation

Article abstract of DOI:10.1016/j.molcata.2005.11.032

The titanium salycilate complex [Cp*Ti(sal)(salH)] 1 (salH ₂ = salycilic acid) was synthesised by reacting Cp*TiMe ₃ and salycilic acid. The molecular structure of 1 studied by X-ray shows it to be dinuclear with a carboxylate group acting as a bridge ligand. [Rh(μ-OH)COD]₂ reacts with 1 to yield the bimetallic complex 2, or with the siloxanodiol (HOSiPh)₂O to yield the bimetalic bis-siloxide Rh complex 3. Carbon monoxide displaces COD in 2 and 3 to give the carbonylic complexes 4 and 5 respectively. Complex 4 can also be prepared from 2 and [Rh(acac)(CO)₂]. Compounds 2, 3, 4 and 5 can be envisaged as models of Rh complexes anchored on titania or silica. The catalytic performance of heterometallic complexes is studied in the CO hydrogenation reaction and compared to rhodium-based (Rh-based) catalysts prepared by impregnation from Rh(NO₃)₃ or [Rh(acac)(CO)₂] showing higher yields to oxygenated products.

Full text of DOI:10.1016/j.molcata.2005.11.032

Journal of Molecular Catalysis A: Chemical 247 (2006) 44–51
Synthesis and reactivity of heterometallic RhO–M (M = Si, Ti)
complexes Memory effect in their catalytic
performance in CO hydrogenation
M. Ojedaa, R. Fandosb,1, J.L.G. Fierroa, A. Oteroc, C. Pastore,
d
b
a,∗
´
A. Rodrıguez , M.J. Ruiz , P. Terreros
a Instituto de Cata´lisis y Petroleoqu´ımica, CSIC, C/Marie Curie 2, 28049 Madrid, Spain
b Facultad de Ciencias del Medio Ambiente, UCLM, Avda. de Carlos III, s/n, 45061 Toledo, Spain
c Facultad de Ciencias Qu´ımicas, UCLM, Campus Ciudad Real, 13071 Ciudad Real, Spain
d ETS Ingenieros Industriales, UCLM, Campus Ciudad Real, 13071 Ciudad Real, Spain
e Servicio Interdepartamental de Apoyo a la Investigacio´n, Facultad de Ciencias, Universidad Auto´noma de Madrid, Madrid, Spain
Received 30 September 2005; received in revised form 14 November 2005; accepted 16 November 2005
Available online 22 December 2005
Abstract
The titanium salycilate complex [Cp*Ti(sal)(salH)] 1 (salH2 = salycilic acid) was synthesised by reacting Cp*TiMe3 and salycilic acid. The
molecular structure of 1 studied by X-ray shows it to be dinuclear with a carboxylate group acting as a bridge ligand. [Rh(␮-OH)COD]2 reacts
with 1 to yield the bimetallic complex 2, or with the siloxanodiol (HOSiPh)2O to yield the bimetalic bis-siloxide Rh complex 3. Carbon monoxide
displaces COD in 2 and 3 to give the carbonylic complexes 4 and 5 respectively. Complex 4 can also be prepared from 2 and [Rh(acac)(CO)2].
Compounds 2, 3, 4 and 5 can be envisaged as models of Rh complexes anchored on titania or silica. The catalytic performance of heterometallic
complexes is studied in the CO hydrogenation reaction and compared to rhodium-based (Rh-based) catalysts prepared by impregnation from
Rh(NO3)3 or [Rh(acac)(CO)2] showing higher yields to oxygenated products.
© 2005 Elsevier B.V. All rights reserved.
Keywords: Early–late heterometallic complexes; CO hydrogenation; Rhodium; O ligands; Titanium; Silicon
1. Introduction
the H2 reduction is carried out at high temperature (≥773 K).
The catalytic performance is then severely affected [1].
The nature of the metal-support interaction can be modulated
by many factors, i.e., the preparation method, metal precursor
and the transition metal compounds used in the catalyst prepa-
ration [2,3]. The ligands in the coordination sphere plus the
interaction with the support will account for the nature of the
metal grafted species.
The chemistry of heteronuclear complexes has been exten-
sively studied due to its importance in different areas. Those
compounds containing an oxygen atom as bridging ligand
between the noble metal and the other component can be envis-
aged as models of noble metals deposited on inorganic oxides
(SiO2, TiO2, etc.). Knowing in detail, at molecular level, the
properties of the active centre is of a great importance. The oxide
used as support is also a key feature in this interaction. When tita-
nia or other reducible oxides are used, the phenomenon known as
strong metal support interaction (SMSI) can be produced when
In this sense, we have carried on with the study of transition-
metal alkoxides [4] to other O containing ligands.
We report here the synthesis of a salicylate titanium com-
pound and we study its reactivity with [Rh(␮-OH)(COD)]2
which leads to the formation of the corresponding early–late het-
erometallic complex that can be considered as molecular model
for a late metal catalysts supported on titania. On the other hand,
we report the synthesis and characterization of a Rh (I) siloxide
complex that constitute a good model for rhodium complexes
supported on a silica surface.
∗
Corresponding author. Tel.: +34 915854768; fax: +34 915854760.
E-mail addresses: rfandos@amb-to.uclm.es (R. Fandos),
pterreros@icp.csic.es (P. Terreros).
1
Tel.: +34 925265727; fax: +34 925268841.
1381-1169/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2005.11.032

M. Ojeda et al. / Journal of Molecular Catalysis A: Chemical 247 (2006) 44–51
45
Besides, we have studied their reactivity towards CO and the
catalytic behaviour in the CO hydrogenation (Fischer–Tropsch
synthesis). The interest of this reaction has continuously grown
up in the last years as demonstrated by some reviews published
recently [5–7]. The synthesis of hydrocarbons from CO hydro-
genation over transition metals is perhaps the most promising
source of chemicals and fuels from non-petroleum-based supply
such as coal and natural gas. The reaction selectivity depends
strongly on the metal used as a catalyst. When Fe and/or Co
is used, the reaction products consist in a mixture of paraf-
fins and olefins, while oxygenates compounds are obtained with
rhodium-based catalysts. In this work, we compare the catalytic
performance of conventional Rh-based catalysts prepared from
the usual precursors (rhodium chloride and nitrate) or closely
related complexes ([Rh(acac)(CO)2]), with that corresponding
to the Rh complexes used as models. The nature of the active
centre plus the support interaction in the big scale FT reaction
is of a great importance to modulate the selectivity towards the
desired products [8,9].
1399 (m), 1380 (m), 1317 (w), 1247 (m), 1142 (w), 1030 (w),
879 (w), 825 (w), 762 (m), 704 (w), 669 (w). 1H NMR (CDCl3,
rt, 200 MHz): 1.87 (s, 15 H, Cp*), 2.11 (s, 15 H, Cp*), 6.48–8.5
(m, 16 H, Ar), 10.19 (br, 1 H, OH), 12.8 (br, 1 H, OH). 13C{1H}
NMR: 12,5 (s, Cp*), 13,4 (s, Cp*), 114.3 (s, Aripso), 114.8 (s,
Aripso), 116.2 (s, Aripso). 119.5 (s, Cp*), 117.7, 118.5, 119.2,
119.6, 120.6, 120.7, 131.0, 131.1, 131.9, 131.9, 132.4, 134.6,
135.3, 135.5, 135.7, 135.8 (s, Ar), 161.7, 167.4, 168.1, 170.9
(s, Aripso), 174.3, 174.6, 175.5, 182.4 (s, COO). Anal. Cald. for
C24H24O6Ti: C, 63.17; H, 5.30. Found: C, 63.66; H, 5.51.
2.2.2. [TiCp*(sal)2Rh(COD)] (2)
To a solution of complex 1 in CH2Cl2 (0.142 g, 0.31 mmol)
was added [Rh(␮-OH)(COD)]2 (0.071 g, 0.15 mmol) and the
mixture was stirred at room temperature for 30 min. After that,
the solvent was removed under vacuum and the residue washed
with pentane to yield an orange solid, which was characterized
as complex 2 (0.134 g, 65%). IR (KBr, cm−1): 1598 (s), 1567
(s), 1516 (vs), 1457 (vs), 1382 (m), 1315 (w), 1247 (m), 1141
(m), 1029 (w), 893 (m), 833 (m), 754 (m), 701 (w), 676 (w),
636 (w), 596 (w). 1H NMR (CD2Cl2, rt, 200 MHz): 1.66 (m, 4
H, COD), 2.12 (s, 15 H, Cp*), 2.40 (m, 4 H, COD), 4.10 (m,
4 H, COD), 6.78 (m, 2 H, Ar), 6.84 (m, 2 H, Ar), 7.39 (m, 2
H, Ar), 7.84 (m, 2 H, Ar). 13C{1H} NMR: 12,2 (s, Cp*), 31,0
(s, COD), 77.7 (br, COD), 119.2, 120.1, 132.3, 134.9 (s, Ar),
131.6 (s, Cp*), 166.7 (s, Aripso), 172.8 (s, COO). Anal. Cald.
for C32H35O6TiRh: C, 57.67; H, 5.29. Found: C, 57.58; H, 5.53.
2. Experimental
2.1. General procedures
The preparation and handling of the described compounds
was performed with rigorous exclusion of air and moisture under
nitrogen atmosphere using standard vacuum line and Schlenk
techniques. All solvents were dried and distilled under a nitrogen
atmosphere.
2.2.3. [{Rh(COD)}2{µ-(OSiPh2)2O}] (3)
The following reagents were prepared according to literature
procedures: [TiCp*(Me)3] [10], [Rh(␮-OH)(COD)]2 [11,12],
and Cp*Ti(O2Bz)2Rh(COD) [4].
To mixture of [Rh((-OH)(COD)]2 (0.090 g, 0.20 mmol) and
(HOSiPh2)2O (0.082 g, 20 mmol) was added toluene (8 mL) at
room temperature and the solution was stirred for 1 h. After
that, the solvent was partially evaporated under vacuum. Slow
diffusion of pentane into the toluene solution afforded yellow
crystals of 3. (0.101 g, 61%). IR (KBr, cm−1): 1428 (m), 1115
(s), 1001 (m), 977 (vs), 964 (s), 914 (s), 871 (m), 743 (w), 713
(s), 699 (s), 527 (vs). 1H NMR (C6 D6, rt, 200 MHz): 1.24 (m, 4
H, COD), 1.58 (m, 4 H, COD), 1.83 (m, 4 H, COD), 2.57 (m, 4 H,
COD), 3.59 (m, 4 H, COD), 4.34 (m, 4 H, COD), 7.20 (m, 12 H,
Ar), 8.13 (m, 8 H, Ar). 13C{1H} NMR: 30,7 (s, COD), 31,2 (s,
COD), 76.5 (d, J = 14.5 Hz, COD), 77.7 (d, J = 14.5 Hz, COD),
127.8, 129.7, 135.3 (s, Ar), 138.9 (s, Aripso). Anal. Cald. for
C40H44O3Si2Rh2: C, 57.55; H, 5.31. Found: C, 57.72; H, 5.09.
The commercially available compounds, salycilic acid and
LiMe in diethyl ether were used as received from Aldrich.
1H and 13C NMR spectra were recorded on a 200 Mercury
Varian Fourier Transform spectrometer. Trace amounts of pro-
tonated solvents were used as references, and chemical shifts are
reported in units of parts per million (ppm) relative to SiMe4.
Transmission infrared spectra were obtained on a Nicolet
ZDX Fourier transform IR spectrophotometer connected to a
Nicolet 680 Spectral Workstation equipped with a DTGS detec-
tor.
Thermal analyses of the fresh catalysts were performed on
a Mettler Toledo TGA/SDTA 851 apparatus. Typically, about
10 mg of the sample was heated from 298 to 1223 K at a rate of
1 K/min under H2 flow.
2.2.4. [TiCp*(sal)2(Rh(CO)2)] (4)
CO was bubbled through a suspension of 2 in hexane for
30 min in an ice bath. The red solid was filtered off and washed
with cold hexane. Yield: 80%. IR (KBr, cm−1): 2083 (vs), 2008
(vs), 1600 (s), 1567 (s), 1497 (vs), 1457 (vs), 1391 (m), 1316 (w),
1246 (m), 1142 (w), 1029 (w), 895 (m), 835 (m), 757 (m), 702
(w), 677 (w), 635 (w), 604 (w); Anal. Cald. for C26H23O8TiRh:
C, 50.75; H, 3.77. Found: C, 51.35; H, 3.89.
2.2. Preparation of compounds
2.2.1. [TiCp*(sal)(salH)] (1)
To a solution of [TiCp*(Me)3] (0.234 g, 1.03 mmol) in
5 mL of toluene, at 233 K, was added salycilic acid (0.283 g,
2.05 mmol). The mixture was allowed to reach the room tem-
perature and then was stirred for 1 h. After filtration, the solvent
was removed under vacuum and the residue washed with 5 mL
of pentane to yield complex 1 as a dark red solid. Yield: 0.357 g,
76%. IR (KBr, cm−1): 1603 (vs), 1570 (s), 1530 (s), 1457 (vs),
2.2.5. [{Rh(CO)2}2{µ-(OSiPh2)2O}] (5)
CO was bubbled through a suspension of 3 in hexane for
20 min in an ice bath. The yellow pale solid was filtered off and
washed with hexane. Yield: 82%. IR (KBr, cm−1): 2077 (vs),

46
M. Ojeda et al. / Journal of Molecular Catalysis A: Chemical 247 (2006) 44–51
2017 (vs), 1429 (w), 1116 (s), 1091 (s), 919 (s), 964 (s), 741
2.5. X-ray data collection
1
(w), 718 (m), 698 (s), 568 (m), 529 (m). H NMR (C6 D6, rt,
200 MHz): 7.23 (m, 12 H, Ar), 8.10 (m, 8 H, Ar). 13C{1H}
NMR: 128.1, 130.7, 135.7 (s, Ar), 136.0 (s, Aripso), 180.8 (d,
1JRh-C = 74.7 Hz, CO). Anal. Cald. for C28H20O7Si2Rh2: C,
46.04; H, 2.76. Found: C, 46.37; H, 3.19.
A
red single crystal of approximate dimensions
0.15 × 0.10 × 0.04 mm with prismatic shape of 1 was mounted
on a glass fibber and transferred to a Bruker SMART 6 K CCD
area-detector three-circle diffractometer with a MAC Science
˚
Rotating Anode (Cu K␣ radiation, λ = 1.54178 A) generator
2.3. Catalysts preparation
equipped with Goebel mirrors at settings of 45 kV and 100 mA.
X-ray data were collected at 100 K, with a combination of six
runs at different ϕ and 2θ angles, 3600 frames. The data were
collected using 0.3◦ wide ω scans (2 s/frame at 2θ = 40◦ and
6 s/frame at 2θ = 100◦), crystal-to-detector distance of 5.0 cm.
The substantial redundancy in data allows empirical absorp-
tion corrections (SADABS) [15] to be applied using multiple
measurements of symmetry-equivalent reflections. The soft-
ware package SHELXTL [16–18] version 6.10 was used for
space group determination, structure solution and refinement.
The structure was solved by direct methods (SHELXS-97)
[17], completed with difference Fourier syntheses, and refined
with full-matrix least-squares using SHELXL-97 [18] mini-
The Rh/␥-Al2O3 sample (reference) containing 3 wt.%
of Rh was prepared by impregnation of ␥-Al2O3 (Condea
Puralox Nwa–155, SBET = 150 m2/g) with an aqueous solution
of Rh(NO3)3·xH2O). Samples 7 and 8 were prepared by using
the same support but [Rh(acac)(CO)2] as metal precursor. For
the preparation of samples named as 9 and 10, silica (PQ Cor-
poration, MS-3030, SBET = 300 m2/g) was used as a support and
[Rh(acac)(CO)2] as rhodium precursor. In all cases, the prepa-
ration procedure was similar, and the rhodium metal loading
was ∼3 wt.%. The support and the impregnating solution were
stirred for several hours and then, water was removed by con-
trolled evaporation in a rotary evaporator. Following, the sample
was dried at 383 K for 6 h and finally calcined at 773 K for 12 h
in air. A detailed charaterisation study of these samples has been
reported elsewhere [13,14].
mizing ω(F02Fc2)2. All non-hydrogen atoms were refined with
anisotropic displacement parameters. The hydrogen atom posi-
tions were calculated geometrically and were allowed to ride on
their parent carbon atoms with fixed isotropic U.
The compound has a molecule of solvent disordered that
seems to be pentane. This disorder was impossible to model.
In the solvent and in the empirical formula were not included
the hydrogen atoms.
2.4. Catalytic measurements
CO hydrogenation reactions were performed using a high-
pressure fixed-bed microreactor (stainless steel 316, 150 mm
length and 9 mm internal diameter). The samples (0.25–0.30 mm
particle size) were directly diluted with SiC (0.25–0.30 mm par-
ticle size) to avoid hot spots. The weight of sample and SiC
were changed for each experiment in order to keep constant
the amount of Rh and the volume of the catalytic bed. In some
experiments, the catalysts were previously reduced in situ at
533 K (heating rate of 10 K/min) for 1 h in a flow of H2/N2 (1:9)
at a rate of 100 mL/min. The reactor was then cooled down and
the gas flow was switched to 50 mL/min of syngas (molar ratio
H2/CO = 2). The system was pressurized at 2.03 MPa and the
reactortemperature was increased ataheating ramp of 10 K/min.
Temperature was measured with a type-K thermocouple buried
in the catalytic bed. Flow rates were controlled using a Brooks
5850TRSeriesmassflowcontrollers. Allexperimentalvariables
were carefully controlled to ensure identical reaction conditions
when testing catalysts. Once the catalyst reached the steady-state
(about 3-4 h), the catalytic performance was measured for about
48 h, showing no deactivation and high catalytic stability.
Product analysis was performed on-line with a gas chro-
matograph (HP 5890 Series II) equipped with TCD and FID
detectors and two in-series fused silica capillary columns:
SPB-5 (60 m × 0.53 mm) and Supel-Q Plot (30 m × 0.53 mm).
It was possible to analyze inorganic gases (H2, CO and
CO2), C1–C20 hydrocarbons, C1–C10 alcohols, and other oxy-
genated compounds. The identification of reaction products and
gas chromatograph calibration were accomplished using gas
chromatography-mass spectrometry (GC–MS) and quantitative
standard solutions from AccuStandard.
A single crystal of a yellow block of 3 was placed on
a NONIUS-MACH3 diffractometer equipped with a graphite
˚
monochromated Mo K␣ radiation (λ = 0.71073 A). Intensity
data were collected by using an ω/2θ scan technique. Exam-
ination of two standard reflections, monitored after 60 min,
showed no signal of crystal deterioration. Data were corrected
for Lorentz and polarization effects, and semi-empirical absorp-
tion correction (Psi-scans) was carried out [19]. The structures
were solved by the direct methods using the SIR92 com-
puter software [20], completed by subsequent difference Fourier
syntheses, and refined by full matrix least-squares procedures
(SHELXL97) [18] on F2. All non-hydrogen atoms were refined
with anisotropic thermal parameters. The hydrogen atoms were
placed using a “riding model” and included in the refinement at
calculated positions.
3. Results and discussion
3.1. Rh–Ti, Si heterometallic complexes
The chemistry of carboxylate [21–28], and in particular of
salicylate complexes, has attracted a great deal of interest as a
consequence of its importance in many areas such as catalysis
or molecular architecture construction [29–35].
The titanium complex [TiCp*(Me)3] reacts with salycilic
acid in a 1:2 molar ratio to yield complex 1 (Scheme 1), which is
isolated as a dark red solid. It is rather soluble in THF or toluene
and less soluble in pentane or Et2O.

M. Ojeda et al. / Journal of Molecular Catalysis A: Chemical 247 (2006) 44–51
47
Scheme 1.
Complex 1 has been characterized by NMR and IR spec-
aryloxide group. On the other hand, the Ti2 atom is bonded to
one of the two salicylate groups in a similar fashion to that in
Ti1, while the second one is bonded to the metal solely through
the carboxylic moiety in a bidentate manner. Besides, Ti1 and
Ti2 are bonded together through the carboxylate group of one
of the salicylate ligands in Ti1 that is acting as bridging group.
It is interesting to note that a single molecule displays all the
coordination modes of carboxylic ligands, i.e., monodentate,
bidentate and bridge.
It is possible that, in solution of toluene, complex 1 is present
as mixture of the two isomers 1a and 1b. But in highly concen-
trated solutions, both isomers would assemble to yield the dark
red crystals, which have been characterized (see Scheme 1).
It is also worth to point out that complex 1 still contains two
hydroxylgroups. Thischaracteristicmakesitaninterestingstart-
ing material in progressive molecular architectures construction.
Complex 1 reacts with [Rh(␮-OH)(COD]2 to yield the
hetero-bimetallic complex 2 (see Scheme 2) through a conden-
sation reaction. It is scarcely soluble in most of the commonly
used organic solvents.
troscopy, as well as by elemental analysis and X-ray diffraction.
1
The H NMR spectrum shows two singlet signals at 1.87 and
2.11 ppm which are assigned to the Cp* ligand in each tita-
nium centre along with several multiplet signals corresponding
to the aromatic protons. The hydroxylic protons appear as broad
absorptions at 10.19 and 12.80 ppm. The 13C{1H} NMR spec-
trum shows that all the four salycilate groups in the complex
are in a different chemical environment. It can be obserseved
16 signals corresponding to the CH moieties and eight differ-
ent peaks corresponding to the ipso carbon atoms. On the other
hand, the IR spectrum, although not conclusively, indicates that
the carboxylic moieties are coordinated to the metallic centers
in different ways, monodentate, bidentate and as a bridge lig-
and. Given the multiple coordination modes of the salicylate
ligand [36–38], we have carried out its characterization by X-ray
diffraction methods in order to establish accurately the structure
of this complex.
A diagram for the structure of 1 is shown in Fig. 1, and
some selected bond distances and angles are summarized in the
Appendix A.
The molecular structure shows it to be a dimetallic compound
in which each one of the two metal centers, Ti1 and Ti2, are
coordinated to one Cp* ligand and to two salicylate groups that
exhibit different coordination modes. The Ti1 metal centre is
bonded to both salicylate groups in a chelate fashion through
one of the two oxygen atoms of the carboxylate moiety and the
Complex 2 has been characterized by some spectroscopic
techniques as well as by elemental analysis. The NMR spectra
show that it is a rather symmetric molecule. Both salicylate moi-
eties are equivalent and give rise to four multiplet signal in the
1H NMR spectrum. On the other hand, the Cp* ligand gives rise
to a singlet signal at 2.12 ppm while the COD group shows a
broad signal corresponding to the olefinic protons.
The rhodium derivative [Rh(␮-OH)(COD)]2 reacts with
(HOSiPh2)2O at room temperature in toluene to yield the cor-
responding rhodium siloxide 3 (Scheme 3). Complex 3 can
be considered as a model of heterogeneous rhodium catalyst
supported on silica. It has been characterized by some spectro-
scopic and analytical techniques as well as by X-ray diffraction
methods. According to the NMR spectra, compound 3 is rather
symmetric, the COD ligand as well as the siloxide ligand are
Fig. 1. An ORTEP representation of the structure of complex [Cp*Ti
(sal)(salH)]2 (1) showing the atom labelling scheme.
Scheme 2.

48
M. Ojeda et al. / Journal of Molecular Catalysis A: Chemical 247 (2006) 44–51
Scheme 3.
˚
(2.765(2) A), and could be considered indicative of some
metal–metal bonding interaction. Nevertheless, it could be also
related to the small size of the bridging atom that forces the metal
centres to be closer to each other [39]. The Rh–O bond lengths
˚
(Rh(1)–O(1), 2.121(5) and Rh(1)–O(2), 2.131(5) A) are similar
to those observed for other rhodium derivatives with bridging
siloxide groups.
The Rh C bond distances fall in the range 2.062(10)–
˚
2.114(8) A, which are normal values for Rh(I) complexes con-
taining COD ligands.
The Ti–Rh heterometallic derivative 2 reacts with an excess
of carbon monoxide at room temperature under atmospheric
pressure to yield the corresponding dicarbonyl complex 4 (see
Scheme 4). Thus, the COD is easily replaced by carbon monox-
ide in complex 2 to give the orange heterometallic complex
[Ti Cp*(sal)2Rh(CO)2] (4) with a good yield (61%). Complex
4 was isolated as a crystalline solid. It is highly insoluble in
most common organic solvents and has been characterized by
IR spectroscopy and by elemental analysis. The carbonyl deriva-
tive 4 can be synthesized as well by reaction of complex 1 and
[Rh(acac)(CO)2].
Fig. 2. An ORTEP representation of the structure of complex [Rh(COD)]2[␮-
(OSiPh2)2O] (3) showing the atom labelling scheme.
bisected by a mirror plane. On the other hand, all the phenyl
groups are in an equivalent chemical environment.
In an analogous way, the rhodium siloxide compound 3 reacts
with CO at room temperature and atmospheric pressure to render
the dicarbonyl derivative 5. It has been isolated as a pale yellow
complex and has been characterized by the normal spectroscopic
techniques as well as by elemental analysis.
An X-ray diffraction study was carried out in order to estab-
lish the molecular structure of complex 3. Fig. 2 shows the
ORTEP diagram of the molecule. Furthermore, the most impor-
tant bond distances and angles are listed in the Appendix A.
The crystal structure shows that coordination around each
rhodium atom is approximately square-planar. Two bridging
oxygen atoms and a chelating cyclooctadiene are bonded to
each metal atom. The Rh(1)–Rh(2) distance is rather short
Complexes 4 and 5 can be envisaged as gem-Rh+(CO)2
species grafted on titania and silica respectively. The ν(CO)
bands for both compounds 4 and 5 are located at 2084, 2009
and 2077, 2017 cm−1, respectively. The pattern and the rel-
Scheme 4.

M. Ojeda et al. / Journal of Molecular Catalysis A: Chemical 247 (2006) 44–51
49
Table 1
Catalysts nomenclature
Catalyst
Rh source
Support
Rh (wt.%)
Pretreatment
Reference
7
8
9
10
3
6
2
4
Rh(NO3)3
[Rh(acac)(CO)2]
[Rh(acac)(CO)2]
[Rh(acac)(CO)2]
␥-Al2O3
␥-Al2O3
␥-Al2O3
SiO2
3
3
3
3
3
H2/N2, 533 K, 1 h
H2/CO
H2/N2, 533K, 1 h
H2/CO
H2/N2, 533K, 1 h
H2/CO
H2/CO
[Rh(acac)(CO)2]
SiO2
[{Rh(COD)}2{␮-OSiPh2O}]
[Ti Cp* (O2Bz)2Rh(COD)]
[Ti Cp*(sal)2Rh(COD)]
[Ti Cp*(sal)2Rh(CO)2]
–
–
–
–
H2/CO
H2/CO
ative intensity of absorptions is the expected for a cis dicar-
bonyl rhodium (I) complex [40,41]. They are comparable
to the characteristic doublet of the νsym, νasym of the gem-
dicarbonylRh+(CO)2 speciespreparedfromdifferentRhprecur-
sors ([Rh(␮-Cl)(CO)2]2, [Rh(acac)(CO)2], [Rh(allyl)(CO)2])
on different oxide supports (SiO2, Al2O3, TiO2, zeolites)
[42–48]. Also, the Rh0(CO)2 groups formed by adsorption of
CO on Rh crystallites can react with the hydroxo groups of SiO2
and Al2O3 to yield the Rh+(CO)2 species [49,50].
Compounds 3 and 5 can be considered as “a silica–silanol
group” with an anchored transition metal complex [51–54]. The
siloxanodiol can be compared with the vicinal diols present in
the silica.
The catalytic activity of 2, 3 and other closely related com-
plexes [Cp*Ti(O2Bz)Rh(COD)] (6) was studied in the CO
hydrogenation reaction in order to determine if, despite the pre-
treatment conditions, the activity and selectivity reflects some-
how the initial complex structure. The results are compared
with those obtained using “standard” catalysts prepared from
other rhodium precursors and oxide supports (silica, alumina).
Catalyst nomenclature, rhodium precursors, supports and pre-
treatment conditions are summarized in Table 1.
The experimental reaction conditions were the following:
533 K, 2.03 MPa, 50 mL/min and H2/CO = 2. In some cases, the
catalysts were previously reduced in a H2/N2 flow at 533 K for
1 h. The reaction products were namely composed by hydrocar-
bons (C1–C10) and oxygenated compounds (1-alcohols).
With respect to the influence of the pre-treatment step, cat-
alysts named as 7 and 8 should be compared. In this way,
it is possible to investigate if the catalyst reduction in a
H2/N2 flow affects its catalytic behaviour. From Table 2, it is
clear that the CO conversion rate decreases after the hydro-
gen treatment. Simultaneously, oxygenates formation decreases
and more hydrocarbons are formed. This could be tentatively
explained assuming that some rhodium atoms have suffered sin-
tering. As it has been published recently [55,56], oxygenates
formation is favoured as the rhodium particle size diminishes.
However, by comparing 9 and 10, it can be observed that
the activity remains constant after hydrogen treatment. This
leads us to conclude that the rhodium particles are more stable
against sintering where they are deposited over silica support
instead of over alumina but the selectivity towards oxygenates
increases.
The nature of the oxide used as a support also plays an impor-
tant role. The effect of the supporting material can be observed
by comparing samples 7 and 9 [(Rh(acac)(CO)2]/␥-Al2O3 ver-
sus [Rh(acac)(CO)2]/SiO2). The catalytic activity is moderately
affected. However, it can be observed that reaction selectivity
is strongly influenced by the support. The alumina-supported
sample shows higher selectivity to the oxygenated compounds.
This is probably due to the interaction between the rhodium
particles and the alumina support. The electrons of the rhodium
atoms are attracted by the oxygen atoms of the support, leading
to some partially positively charged Rh␦+ atoms [57]. Previous
works have suggested that the typical Fischer–Tropsch catalysts
3.2. Catalytic activity
Table 2
Table 2 shows the results of CO hydrogenation with all the
tested samples. First of all, it can be observed that all the Rh-
based catalysts are active in the preparation of hydrocarbons and
oxygenates from synthesis gas. However, the catalytic behaviour
depends strongly on the catalyst precursor, pre-treatment and the
nature of the oxide used as a support.
By comparing the catalysts named as reference and 8
(Rh(NO3)3/␥-Al2O3 versus [Rh(acac)(CO)2]/␥-Al2O3), it can
be seen that the appropriate election of the Rh precursor is very
important. Thus, the catalyst made from rhodium nitrate pre-
cursor shows a higher CO conversion rate (about four times).
Nevertheless, changes in reaction selectivity are not so clear.
Both catalysts display about 50% selectivity to oxygenates.
Catalytic results for the CO hydrogenation over Rh-based catalysts
Catalyst
CO conv. (%) mmol CO/h/
mmol Rh
C1
C2–C4 C5+
Coxyg
Reference
7
8
9
10
3
6
2
4
7.3
4.2
2.1
26.8
15.4
7.7
31.7 12.0
35.1 8.8
7.3
2.3
1.9
49.1
53.5
46.0
23.6 28.5
9.0 50.0
17.4 22.2
27.5 43.1
23.8 27.4
21.6 17.1
21.6 14.6
2.4
2.4
8.1
13.7
56.0
8.8
11.1
11.1
25.0
41.4
195.9
161.1
17.2 23.8
4.8
4.0
5.4
5.4
4.8
55.6
25.4
43.4
56.0
59.0
Reaction conditions: 533 K, 2.03 MPa, H2/CO = 2.

50
M. Ojeda et al. / Journal of Molecular Catalysis A: Chemical 247 (2006) 44–51
Table 3
Oxygenated products distribution
Cat.
mmolCO/h/mmol Rh
Coxyg
C2OHa
C2Ob
C2OOMec
C2OOEtd
C2OOHe
C
3+Of
Reference
7
8
9
10
3
6
2
4
26.8
15.4
7.7
11.1
11.1
25.0
41.4
195.9
161.1
49.1
53.5
46.0
23.8
55.6
25.4
43.4
56.0
59.0
20.8
17.2
17.9
6.5
19.1
18.0
23.6
24.9
24.8
7.4
7.0
0.0
14.7
35.4
3.8
0.0
1.4
7.7
4.8
6.8
8.8
0.0
0.0
0.0
4.6
3.8
7.3
13.5
14.2
13.6
0.0
0.0
0.0
9.6
18.4
13.1
0.3
6.7
3.9
0.0
0.0
0.0
2.4
5.6
4.3
2.3
1.6
1.8
2.6
0.9
3.6
3.2
1.9
1.8
Reaction conditions: 533 K, 2.03 MPa, H2/CO = 2.
a
C2OH: ethanol.
C2O: acetaldehyde (+ methanol traces).
C2OOMe: methylacetate.
C2OOEt: ethyl acetate.
C2OOH: acetic acid.
C3+O.
b
c
d
e
f
can produce oxygenates when they are in ␦+ oxidation states
[58,59].
formance of complex 2, which displays nor only the highest
CO conversion, but also a high selectivity to the desired oxy-
genated compounds (Table 3). In this Table, it can be seen that
most of the oxygenated compounds contain or come from oxy-
genates with two carbon atoms. Only trace amounts of methanol
were detected. This observation shows up the ability of rhodium-
based catalysts to insert CO into hydrocarbons chains, especially
into CHx units, which leads to C2-oxygenates. The formation
of two carbon atoms oxygenates has a great interest because
these compounds are very important feedstocks for the chemi-
cal industry.
Several rhodium-based complexes have been also tested in
the CO hydrogenation reaction (see catalysts named as 2, 3, 4
and 6). The high activity of these complexes is not due to the
formation of hot spots in the catalytic bed because the samples
are highly diluted in SiC.
The thermogravimetric analysis of complexes 2 and 3 under
H2 shows an initial exothermic weight increase of 0.9 and 0.4%
respectively, which is indicative of an uptake of H2 or hydride
formation. Compound 3 is very stable thermically, loosing the
organic ligands between 483 and 573 K (48% yield). The the-
oretical yield for “Rh2O3Si” should be 37%, lower that the
41% obtained at 873 K. Compound 2 has a gradual weight loss
between 120 and 873 K. The theoretical yield for “RhTiO2”
(27.4%) is also higher than the 31% yield at 873 K (Fig. 3).
The Rh–Ti bimetallic compounds (2, 4, 6) display a higher
activity that the Rh–Si complex 3. Moreover, the selectivity
towards oxygenated products is much higher. Compounds 2 and
4, containing salycilate ligands, are more active and selective
than the alkoxide complex 6. It is especially remarkable the per-
4. Conclusions
Although it is difficult to know the exact nature of the
active centres in the reaction conditions, it can be seen that all
complexes are more active than the conventional Rh-supported
catalysts. The Ti–O–Rh bimetallic complexes exhibit a better
performance that the Rh–O–Si complex being remarkable the
effect of the salycilate ligand compare to the alkoxide.
5. Supplementary material
Crystallographic data for the structural analysis of 1 and
3 have been deposited with the Cambridge Crystallographic
Data Centre, CCDC-284704 & 284705. Copies of this infor-
mation may be obtained free of charge from The Direc-
tor, CCDC, 12 Union Road, Cambridge CB2 1EZ, UK
(fax: +44 1223 336033; E-mail: deposit@ccdc.cam.ac.uk or
http://www.ccdc.cam.ac.uk).
Acknowledgements
This work was supported by the Ministerio de Ciencia y
´
Tecnologıa, Spain (Grant No. BQU2002-04638-C02-02 and
MAT2001-2215-C03-01) and the Junta de Comunidades de
Castilla-La Mancha (Grant No. PAC-02-003 and GC-02-010).
Fig. 3. TGA profiles of 2 and 3 under a flow of H2, with a heating rate of
1 K min−1
.

M. Ojeda et al. / Journal of Molecular Catalysis A: Chemical 247 (2006) 44–51
51
Appendix A
[25] U. M
o¨rhring, M. Scha¨fer, F. Kukla, M. Schlaf, H. Werner, J. Mol. Catal.
A: Chem. 99 (1995) 53.
[26] F. Kukla, H. Werner, Inorg. Chim. Acta 235 (1995) 253.
[27] H. Werner, M. Scha¨fer, O. Nu¨rnberg, J. Wolf, Chem. Ber. 127 (1994)
27.
[28] F. Nicolo`, G. Bruno, S. Lo Schiavo, M.S. Sinicropi, P. Piraino, Inorg.
Chim. Acta 223 (1994) 145.
[29] V.G. Kessler, Chem. Commun. (1985) 1213.
[30] C.N.R. Rao, S. Natarajan, R. Vaidhyanathan, Angew. Chem. Int. Ed. 43
(2004) 1466.
[31] D.E. Edwards, M.F. Mahon, T.J. Paget, N.W. Summerhill, Transition
Metal Chem. 26 (2001) 116.
[32] S.L. Castro, W.E. Streib, J.C. Huffmann, G. Christou, Chem. Commun.
(1996) 2177.
[33] K. Gigant, A. Rammal, M. Henry, J. Am. Chem. Soc. 123 (2001) 11632.
[34] D.H. Gibson, Y. Ding, R.L. Miller, B.A. Sleadd, M.S. Mashuta, J.F.
Richardson, Polyhedron 18 (1999) 1189.
˚
Bond lengths [A] for 1: Ti(1)-[O(2)/1.897(2), O(5)/1.912(2),
O(4)/2.015(2), O(1)/2.019(2)], Ti(2)-[O(7)/1.921(2), O(8)/
1.998(2), O(6)/2.085(2), O(10)/2.095(2), O(11)/2.216(2)],
C(1)-Ti(1)/2.379(4), Ti(2)-C42/2.521(4). Angles [◦]: C(17)-O
(1)-Ti(1)/135.7(2), C(11)-O(2)-Ti(1)/136.6(2), C(18)-O(4)-
Ti(1)/134.4(2), C(24)-O(5)-Ti(1)/131.7(2), C(18)-O(6)-Ti(2)/
142.9(2), C(41)-O(7)-Ti(2)/130.8(2), C(35)-O(8)-Ti(2)/132.2
(2), C(42)-O(10)-Ti(2)/93.32(2), C(42)-O(11)-Ti(2)/88.2(2).
˚
Bond lengths [A] for 3: Rh(1)-[Rh(2)/2.7646(14),
O(1)/2.121(5), O(2)/2.131(5)]; Rh(2)-[O(2)/2.102(5), O(1)/
2.133(4)]. Angles [◦]: Si(1)-O(1)-Rh(1)/115.2(3), Si(1)-O(1)-
Rh(2)/118.6(3), Si(2)-O(2)-Rh(2)/ 114.7(3), Si(1)-O(3)-Si(2)/
130.1(3).
[35] E. Mieczyn´ska, A.M. Trzeciak, J.J. Zio´łkowski, T. Lis, Polyhedron 13
(1994) 655.
[36] M.G. Meirin, E.W. Neuse, M. Rhemtula, S. Schmitt, H.H. Brintzinger,
Trans. Met. Chem. 13 (1988) 272.
References
[37] N.J. Brownless, D.A. Edwards, M.F. Mahon, Inorg. Chim. Acta 287
(1999) 89.
[1] S.J. Tauster, Acc. Chem. Res. 20 (1987) 389.
[2] J.F. Lambert, M. Che, J. Mol. Catal. A: Chem. 162 (2000) 5.
[3] J. Guzman, B.C. Gates, Dalton Trans. (2003) 3303.
[4] R. Fandos, C. Herna´ndez, A. Otero, A. Rodr´ıguez, M.J. Ruiz, P. Terreros,
Chem. Eur. J. 9 (2003) 671.
[38] G.B. Deacon, R.J. Phillips, Coord. Chem. Rev. 33 (1980) 227.
[39] A. Vizi-Orosz, R. Ugo, R. Psaro, A. Sironi, M. Moret, C. Zucchi, F.
Ghelfi, G. P
a´lyi, Inorg. Chem. 33 (1994) 4600.
[40] S. Rojas, J.L. Garc
´ıa-Fierro, R. Fandos, A. Rodr´ıguez, P. Terreros, J.
[5] T.H. Fleisch, R.A. Sills, M.D. Briscoe, J. Nat. Gas Chem. 11 (2002) 1.
[6] J. Zhang, J. Chen, Y. Li, Y. Sun, J. Nat. Gas Chem. 11 (2002) 99.
[7] A.A. Adesina, Appl. Catal. A: Gen. 138 (1996) 345.
[8] P.M. Maitlis, J. Mol. Catal. A: Chem. 204–205 (2003) 54.
[9] P.M. Maitlis, J. Organomet. Chem. 689 (2004) 4366.
[10] M. Mena, M.A. Pellinghelli, P. Royo, R. Serrano, A. Tiripicchio, J.
Chem. Soc., Chem. Commun. (1986) 1118.
Chem. Soc., Dalton Trans. (2001) 2316.
[41] M.A. Ciriano, F. Viguri, J.J. P
e´rez-Torrente, F.J. Lahoz, L.A. Oro,
A. Tiripicchio, M. Tiripicchio-Camellini, J. Chem. Soc., Dalton Trans.
(1989) 25.
[42] J. Evans, B.E. Hayden, M.A. Newton, Surf. Sci. 462 (2000) 169.
[43] B.E. Hayden, A. King, M.A. Newton, N. Yoshikawa, J. Mol. Catal. A:
Chem. 167 (2001) 33.
[11] R. Uso´n, L.A. Oro, J.A. Cabeza, Inorg. Synth. 23 (1985) 126.
[12] D. Selent, M. Ramm, J. Organomet. Chem. 485 (1995) 135.
[13] M. Ojeda, M. Lo´pez Granados, S. Rojas, P. Terreros, J.L.G. Fierro, J.
Mol. Catal. A: Chem. 202 (2003) 179.
[14] M. Ojeda, M. Lo´pez Granados, S. Rojas, P. Terreros, F.J. Garc´ıa-Garc´ıa,
J.L.G. Fierro, Appl. Catal. A: Gen. 261 (2004) 47.
[15] G.M. Sheldrick, SADABS version 2.03, a Program for Empirical
Absorption Correction; Universita¨t Go¨ttingen, 1997–2001.
[16] Bruker AXS SHELXTL version 6.10, Structure Determination Package,
Bruker AXS 2000. Madison, WI.
[17] G.M. Sheldrick, SHELXS-97 Program for Structure Solution, Acta Crys-
tallogr. Sect. A 46 (1990) 467.
[44] B.E. Hayden, A. King, M.A. Newton, Surf. Sci. 397 (1998) 306.
[45] J.M. Andersen, Platinum Met. Rev. 41 (1997) 132.
[46] J.A.M. Andersen, A.W. Currie, Chem. Commun. (1996) 1543.
[47] J. Evans, B. Hayden, F. Mosselmans, A. Murray, Surf. Sci. 301 (1994)
61.
[48] S.C.C. Chuang, G. Srinivas, A. Mukherjee, J. Catal. 139 (1993) 490,
and references therein.
[49] P. Basu, D. Panayotov, J.T. Yates Jr., J. Am. Chem. Soc. 110 (1998)
2074.
[50] A.M. Thayer, T.M. Duncan, J. Phys. Chem. 93 (1989) 6763.
[51] B. Marciniec, H. Maciejewski, Coord. Chem. Rev. 223 (2001) 301, and
references therein.
[18] G.M. Sheldrick, SHELXL-97 Program for Crystal Structure Refinement;
Universita¨t Go¨ttingen, 1997.
[19] A.C.T. North, D.C. Phillips, F.S. Mathews, Acta Crystallogr. Sect A 24
(1968) 351.
[52] K. Merz, S. Block, R. Schoenen, M. Driess, Dalton Trans. (2003) 3365.
[53] M. Nishiura, Z. Hou, Y. Wakatsuki, Organometallics 23 (2004)
1359.
[54] J. Jarupatrakorn, T.D. Tilley, Dalton Trans. (2004) 2808.
[20] A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi, J. Appl.
Crystallogr. 26 (1993) 343.
[55] M. Ojeda, S. Rojas, M. Boutonnet, F.J. P
e´rez-Alonso, F.J. Garc´ıa-Garc´ıa,
J.L.G. Fierro, Appl. Catal. A: Gen. 274 (2004) 33.
[21] E. Mieczyn´ska, A.M. Trzeciak, J.J. Zio´łkowski, T. Lis, J. Chem. Soc.
Dalton Trans. (1995) 105.
[56] M. Ojeda, S. Rojas, F.J. Garc
´ıa-Garc´ıa, M. Lo´pez Granados, P. Terreros,
J.L.G. Fierro, Catal. Commun. 5 (2004) 703.
[22] L. Carlton, M.P. Belciug, G. Pattrick, Polyhedron 11 (1992) 1501.
[23] E. Mieczyn´ska, A.M. Trzeciak, J.J. Zio´łkowski, J. Mol. Catal. 80 (1993)
189.
[57] J. Rask
o´, J. Bontovics, Catal. Lett. 58 (1999) 27.
[58] H. Hayashi, L.Z. Chen, T. Tago, M. Kishida, K. Wakabayashi, Appl.
Catal. A: Gen. 231 (2002) 81.
[24] H. Werner, S. Poelsma, M.E. Schneider, B. Windmu¨ller, D. Barth, Chem.
Ber. 129 (1996) 647.
[59] J.P. Hindermann, G.J. Hutchings, A. Kinnenmann, Catal. Rev. Sci. Eng.
35 (1993) 1.