Rhodium-diphosphine complex bound to activated carbon: An effective catalyst for the hydroformylation of 1-octene

Article abstract of DOI:10.1016/j.molcata.2003.12.015

A Rh(I) diphosphine complex was anchored on an activated carbon by means of a covalent bond and tested as catalyst for the hydroformylation of 1-octene. The heterogenization procedure included the functionalization of the carbon surface to create acid chloride groups where an esterification reaction with the hydroxyl end of a diphosphine ligand (HONP) could take place. Two strategies were followed, i.e., route 1, ligand anchorage followed by complex formation (using the [Rh(μ-Cl)(COD)]₂ complex as precursor) and route II, synthesis of the rhodium diphosphine complex [Rh(HONP)(COD)]Cl followed by anchorage on the functionalized carbon support. The two synthetic routes rendered catalysts that were active, selective, and stable enough to be used in consecutive catalytic runs. The anchoring procedure that first synthesized the Rh diphosphine catalyst and then bonded it to the carbon surface, was particularly attractive: the resulting catalyst is fully active in four consecutive catalytic cycles with a conversion to the linear aldehyde that increased with reuse.

Full text of DOI:10.1016/j.molcata.2003.12.015

Journal of Molecular Catalysis A: Chemical 213 (2004) 177–182
Rhodium-diphosphine complex bound to activated carbon
An effective catalyst for the hydroformylation of 1-octene
a,∗
a,b
´
´
´
M. Carmen Román-Martınez , José A. Dıaz-Auñón
,
a
b
Concepción Salinas-Martınez de Lecea , Howard Alper
a
Departamento de Qu´ımica Inorgánica, Universidad de Alicante, Apartado 99, 03080 Alicante, Spain
b
Department of Chemistry, Centre for Catalysis Research and Innovation, University of Ottawa,
10 Marie Curie, Ottawa, Ont., Canada K1N 6N5
Received 9 September 2003; received in revised form 9 September 2003; accepted 12 December 2003
Abstract
A rhodium(I) diphosphine complex has been anchored on an activated carbon by means of a covalent bond and then it has been tested as
catalyst for the hydroformylation of 1-octene. The heterogenization procedure includes the functionalization of the carbon surface to create
acid chloride groups where an esterification reaction with the hydroxyl end of a diphosphine ligand (HONP) can take place. Two strategies
have been followed: route I: ligand anchorage followed by complex formation (using the [Rh(␮-Cl)(COD)]2 complex as precursor) and route
II: synthesis of the rhodium diphosphine complex [Rh(HONP)(COD)]Cl followed by anchorage on the functionalized carbon support. The
solid samples, functionalized activated carbon and anchored complexes, have been characterized by XPS and 31P NMR. The two synthetic
routes render catalysts that are active, selective and stable enough to be used in consecutive catalytic runs. The catalyst prepared by route II
shows an outstanding behaviour, being fully active and with almost constant selectivity to the linear aldehyde in four consecutive catalytic
runs.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Immobilization; Covalent bond; Activated carbon; Hydroformylation; Rh complex
1. Introduction
The main problems related with heterogenisation on solid
supports are the leaching of the metal complex under reac-
tion conditions and the decrease in catalytic activity with re-
spect to the homogeneous process. Thus, for example, Balué
and Bayón [12] reported that a rhodium thiolate binuclear
catalyst supported on a cationic resin is much slower than
the homogeneous catalyst with the same molecular struc-
ture. And more recently, Bryant and Kilner [17] found that
the overall yield in the hydroformylation of 1-hexene is re-
duced by supporting rhodium catalysts in nafion, and many
of the systems investigated suffer an important leaching
process [17].
It must be remembered that leaching under hydroformy-
lation conditions usually takes place by substitution of the
␲-bonding ligands (like PR3, dienes, etc.) by CO, with the
formation of metal carbonyls that wash out easily from the
support [1]. Because of that, a robust anchorage is required
for supported complexes used as hydroformylation catalysts.
In a previous work [13], the complex [Rh(␮-Cl)(COD)]2
was heterogenized on activated carbons by ion exchange.
The catalysts showed catalytic properties similar to those
The hydroformylation of olefins (known as the oxo pro-
cess) is one of the most important industrial processes using
transition metal complexes as homogeneous catalysts. To
overcome the disadvantages of the homogeneous process,
mainly related with the recovery of the metal complex from
the reaction media, some alternatives to support the metal
complexes are in use or under investigation [1–18]. With
basis on the work of Arhancet et al. [5], dealing with sup-
ported aqueous-phase catalysts (SAPCs), two-phase catal-
ysis has already achieved industrial-scale importance in
hydroformylation [4]. However, to avoid solubility limita-
tions and to exploit the advantages of heterogeneous catal-
ysis, there is great interest in the development of so-called
immobilized metal complex catalysts. These are transition
metal complexes heterogenised on various supports by
physical adsorption or chemical anchoring.
∗
Corresponding author. Tel.: +34-965903975; fax: +34-965903454.
E-mail address: mcroman@ua.es (M.C. Roma´n-Mart´ınez).
1381-1169/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2003.12.015

178
M.C. Roma´n-Mart´ınez et al. / Journal of Molecular Catalysis A: Chemical 213 (2004) 177–182
of the homogeneous complex in the first run, but leaching
was close to 50%. With an oxidized support, the resistance
against leaching was slightly improved but the induction
time was higher. This anchorage was considered not effective
enough because leaching of the complex was high, and the
catalytic activity decreased dramatically in consecutive runs.
A new strategy is reported in this paper: the development
of a system with the Rh complex anchored to the support
by a covalent bond [14]. To achieve this goal, the carbon
surface must be functionalized and a ligand with a suitable
functional group has to be used. The procedure described in
this work includes the formation of carboxylic acid groups
followed by their transformation into acid chlorides. These
are good nucleophilic groups where ligands containing the
hydroxyl functionality can be fixed through an esterification
reaction. This heterogenisation procedure resembles the one
used to anchor complexes on the SiO2 surface [19] and it is
also related to the one reported by Pugin [20] and more re-
cently by Stanger et al. [21]. In the case of activated carbon,
Silva et al. [22] have functionalized the support with acid
chloride groups and immobilised amine nickel complexes
via an amide bond. The characterisation of these catalysts
shows that the anchoring was successful, but no activity tests
were reported [22].
Plate 1. Schematic representation of some species involved in this work.
SOCl2. About 1 g of carbon GF-N was then introduced in
a Schlenk tube with 0.5 ml of SOCl2 in 20 ml of degassed
CH2Cl2 and treated at 80 ◦C under N2 for 2 h. The sample
was washed three times with degassed CH2Cl2 and dried
under vacuum at 40 ◦C. The chlorinated activated carbon is
named GF-Cl (1, see Plate 1).
2.1.2. Synthesis and characterization of the HONP ligand
and the Rh diphosphine complex
In the present work, the heterogenised catalysts have been
characterized by XPS and solid state 31P NMR and, what is
essential in this kind of investigation, they have been tested
under reaction conditions, specifically for the hydroformy-
lation of 1-octene. The study of the catalytic properties in-
cludes the determination of total conversion and selectivity,
and the analysis of the reutilization in consecutive catalytic
runs.
The synthesis of the ligand HONP (2, Plate 1) was per-
formed in a Schlenk system, with basis on the procedure
described in the literature [6]. A mixture of formaldehyde
(37 wt.% in water, 0.6 ml, 8.0 mmol) and diphenylphosphane
(1.2 ml in 5 ml methanol, 6.8 mmol) was heated at 70 ◦C
for 30 min under inert atmosphere. The reaction mixture
was cooled to room temperature and treated with 0.25 ml of
HO(CH2)2NH2 (3.3 mmol). After 30 min, the solution was
heated at 70 ◦C overnight, and then it was cooled down to
room temperature. The product was concentrated and dried
in vacuum for 6 h, resulting in a pale yellow oil-like sub-
stance. It was characterised by 1H, 13C and 31P NMR [6,19].
2. Experimental
2.1. Preparation of catalysts
31P (300 MHz, CDCl3): −27.2 ppm. H NMR (300 MHz,
1
2.1.1. Functionalization of the carbon support
CDCl3): 1.58 (q, 2H, methylene, JH–H = 5.9 Hz), 3.04 (t,
2H, –CH2–NR2, JH–H = 6.0 Hz), 3.52 (t, 2H, –CH2OH,
JH–H = 5.9 Hz), 3.67 (m, 4H, RN–(CH2–PR3)2), 7.33 (m,
12H, phenyl), 7.45 (m, 8H, phenyl). 13C (300 MHz, CDCl3):
A commercial activated carbon, GF-45 from NORIT,
was used as support. This activated carbon, in pelletised
form (about 0.5 mm ∅ and 5 mm length), has a surface area
of 1700 m2/g and the following pore-volume distribution:
micropore (∅ < 0.7 nm) = 0.34 cm3/g, supermicropore
(0.7 nm < ∅ < 2.0 nm) = 0.50 cm3/g, mesopore (2.0 nm <
∅ < 50 nm) = 0.45 cm3/g, and macropore (50 nm < ∅)
= 0.40 cm3/g.
1
28.4 (C–CH2–C), 54.0 (N–CH2–P, JP−C = 19.5 Hz),
3
58.4 (CH2–N, JP−C = 6.5 Hz), 61.1 (CH2–OH), 128.2
2
3
(o-phenyl, JP−C = 7.0 Hz), 132.6 (m-phenyl, JP−C
=
=
1
3.9 Hz), 132.9 (p-phenyl), 137.2 (P–C (phenyl), JP−C
12.2 Hz). The NMR spectra also showed a small amount
of HOCH2P(Ph)2 (31P peak: −10.5 ppm) and monophos-
phinated amine HO(CH2)3NH(CH2P(C6H5)2) (31P peak:
−24.2 ppm).
The support functionalization includes two steps: oxida-
tion and chlorination. Oxidation was carried out by treat-
ment with a solution of HNO3 35 wt.% (5 ml/g) at 80 ◦C
until dryness. The oxidised carbon was extensively washed
with deionised water until the total elimination of nitrates,
and dried in oven at 100 ◦C overnight. With this treatment,
a significant proportion of carboxylic groups are formed
on the carbon surface [23]. The oxidised activated carbon
is named GF-N. The chlorination step was performed with
The rhodium complex [Rh(HONP)(COD)]Cl (4) was syn-
thesized inside the Schlenk apparatus as follows: 11.8 mg of
complex [Rh(␮-Cl)(COD)]2 (3) (abbreviated as Rh(COD))
(24 ␮mol) was added to a solution of 32.9 mg of the syn-
thesized HONP ligand (2) in 20 ml MeOH (molar ratio
1:2). The mixture was stirred in N2 at room temperature

M.C. Roma´n-Mart´ınez et al. / Journal of Molecular Catalysis A: Chemical 213 (2004) 177–182
179
Scheme 1. Synthetic routes to obtain the Rh complex covalently bounded to the activated carbon surface.
for 2 h. A dark-red solution, different from the pale-yellow
solution of Rh(COD) (3), was obtained. The synthesized
complex was characterized by 31P and 1H NMR [6,19]:
31P (300 MHz, CDCl3): 8.3 ppm. 1H NMR (300 MHz,
CDCl3): 1.35 (2H, methylene), 2.12 (8H, methylene COD),
2.98 (2H, –CH2–NR2), 3.18 (2H, –CH2OH), 3.63 (4H,
similar to the one described above. In this case, a solution
of the complex, in degassed methanol, with the appro-
priated concentration to obtain a catalyst with 1 wt.% Rh
was employed. The catalyst prepared by route II is named
Rh(ONP)/GF-Cl (6II).
2.2. Characterisation by XPS and solid state 31P NMR
=
RN–(CH2–PR3)2), 7.06–7.70 (24H, phenyl and H–C C
COD).
The XPS spectra were obtained with a VG-Microtech
Multilab electron spectrometer, using Mg K␣ (1253.6 eV)
radiation. The binding energy was adjusted by setting the C
1s transition at 284.6 eV. The accuracy of BE and KE values
was 0.2 and 0.3 eV, respectively.
2.1.3. Anchorage on the carbon support
Heterogenised catalysts were prepared by the formation
of a covalent bond, via an esterification reaction, between
the acid chloride groups of the carbon surface and the –OH
functional group of ligand HONP (2). Scheme 1 shows the
two different anchoring strategies followed:
Analysis by solid state 31P NMR was carried out
in a Bruker 200 MHz equipment. Samples GF-ONP(5),
Rh(COD)/GF-ONP (6I) and Rh(ONP)/GF-Cl (6II) were
analysed by this technique.
(i) Route I: The ligand HONP (2) is bonded to the sup-
port GF-Cl (1) and later the diphosphine Rh complex is
formed through reaction with Rh(COD) (3).
(ii) Route II: The rhodium diphosphine complex [Rh-
(HONP)(COD)]Cl (4), previously synthesized, is
bonded to the support GF-Cl (1) through the HONP
ligand (2).
2.3. Determination of Rh content
The rhodium content in fresh and used catalysts was de-
termined using the method described in the literature [12]. It
consists of the extraction of Rh from the catalyst with a hot
concentrated solution of HCl and HNO3. The filtrate is fur-
ther heated to dryness, treated with concentrated HCl, and
heated again to dryness. The residue is dissolved in water
and treated with a solution of SnCl2 in HCl 2 M. The result-
ing solution is heated until a dark red complex is formed,
and finally diluted with HCl 2 M to a total volume of 25 ml.
The determination of the concentration of rhodium in the
solution was performed by UV-Vis absorption at 475 nm.
The anchorage of ligand HONP on the carbon support
(route I in Scheme 1) was carried out inside a Schlenk
tube by adding 0.118 g of HONP ligand dissolved in de-
gassed CH2Cl2 to 1 g of sample GF-Cl (1). The slurry was
stirred and heated to reflux (at 70 ◦C) under N2 for 1 h. Af-
terwards, it was cooled to room temperature and dried at
70 ◦C for 4 h under vacuum. The ligand grafted-support is
named GF-ONP (5). To synthesize the supported Rh com-
plex, about 1 g of the dried GF-ONP (5) support was added
to a solution of Rh(COD) (3) in methanol of the appro-
priated concentration to obtain a catalyst with 1 wt.% Rh
(97.2 ␮mol/g). The slurry was stirred under inert atmosphere
at room temperature for 2 h. Then, the solution was removed
with a syringe and the catalyst was dried inside the Schlenk
tube (T ∼ 50 ◦C). The catalyst thus prepared is named
Rh(COD)/GF-ONP (6I).
2.4. Catalytic activity measurements
Catalytic activity experiments were run in a stainless steel
stirred tank reactor. The hydroformylation of 1-octene was
carried out at 80 ◦C and a total pressure of 4 MPa (with
a H2:CO ratio of 1:1) using a solution (10 ml) of 5 vol.%
1-octene in hexane (3 mmol) and 50 mg of the heterogeneous
catalyst (4.9 ␮molRh for catalysts with 1 wt.% Rh). For the
homogeneous reference test, the metal complex Rh(COD)
(3) in a similar amount (in mole Rh) to that present in the
The procedure used to anchor the complex [Rh-
(HONP)(COD)]Cl (4) on GF-Cl (1) (route II, Scheme 1) is

180
M.C. Roma´n-Mart´ınez et al. / Journal of Molecular Catalysis A: Chemical 213 (2004) 177–182
heterogeneous catalysts was dissolved in the liquid phase.
The reaction was carried out for 7 h. Then, the reactor was
depressurized and allowed to cool to room temperature. With
the purpose of testing the reutilization of the catalyst, it was
recovered from the reaction media by removing the solution
with the aid of a syringe and washed three times with hexane
(also by removing the solution with a syringe). Afterwards,
a new catalytic run was performed. Catalytic activity and
selectivity were determined by 1H NMR spectroscopy of the
resultant concentrate solution.
As a consequence of the oxidising treatment with nitric
acid (sample GF-N), organic nitrogen and chloride anions
are removed, but inorganic nitrogen (as nitrate) is intro-
duced. As shown in Table 2, phosphorous has been also
partially removed by treatment with HNO3, and oxygen
groups have been extensively formed. The binding energies
observed for oxygen (Table 1) shift to slightly higher val-
ues probably because new oxygen groups like carboxylic
acids, esters, anhydrides or lactones have been formed
[24].
Treatment with SOCl2 (sample GF-Cl (1)) results in a
substantial increase of organic chlorine (Tables 1 and 2). A
decrease in the O/C ratio confirms the substitution of OH
groups originally in GF-N by Cl in GF-Cl.
3. Results and discussion
3.1. XPS analysis
The analysis of sample GF-ONP (5) shows the expected
increase in P/C and N/C ratios due to the ligand bonded
to the carbon surface. The decrease in the Cl/C ratio is a
consequence of the esterification reaction. It must be noted
that the XPS analysis of sample GF-ONP (5) was conducted
in an air exposed sample, and for that reason, the O/C ra-
tio could be affected by the reaction of chlorine groups
with air.
In the case of the catalyst Rh(COD)/GF-ONP (6I), the
XPS results (Table 1) reveal that rhodium is Rh(I) but with
a binding energy higher than in the pure Rh(COD) com-
plex [13] (308.4 eV). According to the literature [25], it
has been assigned to Rh bonded to phosphine, indicating
that the expected anchored complex was formed. Chloride
Table 1 shows the binding energies of Rh 3d5/2, P 2p,
N 1s, Cl 2p and O 1s in samples GF-45, GF-N, GF-Cl (1),
GF-ONP (5), Rh(COD)/GF-ONP (6I) and Rh(ONP)/GF-Cl
(6II). Table 2 includes the Rh, P, Cl, N and O atomic ratios
referred to C.
The original activated carbon GF-45 contains organic
and inorganic phosphorous. They are probably related
with the carbon precursor material and the procedure used
for the activation (with phosphoric acid). It contains, as
well, organic nitrogen, chloride anions and oxygen, mainly
=
as OH and C O (signals (1) and (2), respectively, in
Table 1).
Table 1
XPS data. Binding energies (in eV)
GF-45
–
GF-N
–
GF-Cl
–
GF-ONP
–
Rh(COD)/GF-ONP
309.0
Rh(ONP)/GF-Cl
309.0
Rh
P
Org
Inorg
133.1
134.7
132.9
134.0
132.8
134.1
132.4
133.6
132.4
133.6
132.1
133.6
N
Org
Inorg
398.6/400.6
–
–
400.0
405.7
399.7
405.8
399.6/401.3
405.8
400.0
405.8
405.8
Cl
O
Inorg
Org
197.8
–
–
–
–
–
198.0
199.9
197.7
199.8
199.8
199.9
(1)
(2)
530.9
532.4
531.5
533.0
531.7
533.4
531.4
533.1
531.5
533.2
531.6
533.2
Table 2
XPS data. Atomic ratio
GF-45
GF-N
–
0.0031
–
0.0204
0.2432
GF-Cl
GF-ONP
Rh(COD)/GF-ONP
Rh(ONP)/GF-Cl
Rh/C
P/C
Cl/C
N/C
O/C
–
–
–
0.0048
0.0116
0.0210
0.0254
0.2427
0.0026
0.0043
0.0100
0.0232
0.2456
0.0118
0.0025
0.0050
0.1604
0.0033
0.0213
0.0258
0.1999
0.0101
0.0103
0.0263
0.2361

M.C. Roma´n-Mart´ınez et al. / Journal of Molecular Catalysis A: Chemical 213 (2004) 177–182
181
introduced with the Rh(COD) complex is also detected (sig-
nal at about 198 eV) and, as a consequence, the ratio Cl/C
increases (Table 2). This chloride must be adsorbed on the
carbon surface.
in four consecutive catalytic runs. For comparative purposes,
data corresponding to the Rh(COD) complex (3) in homo-
geneous phase have been also included. Results of the total
conversion of 1-octene, the conversion to aldehydes (nonanal
and 2-methyloctanal) and the conversion to the linear alde-
hyde (nonanal) are presented. The difference between the
total conversion and conversion to aldehydes corresponds to
the isomerisation products, mainly 2-octene.
In the first run, total conversion is very high for the three
catalysts investigated. However, while the homogeneous
Rh(COD) catalyst produces only the aldehydes, with the
heterogenised catalysts a noticeable amount of 2-octene is
obtained. Considering that the phosphine ligands are usually
employed to decrease isomerization [1], it was not expected
that complex [Rh(HONP)(COD)]Cl (4) would be more
active than Rh(COD) for the isomerization reaction. This
suggests that the anchorage of the complex could have some
influence in the selectivity. The conversion to nonanal with
catalysts Rh(COD) (homogeneous) and Rh(COD)/GF-ONP
(6I) and Rh(ONP)/GF-Cl (6II) is 47.4, 33.9 and 44.0%,
respectively (with L/B ratio: 0.9, 2.4 and 2.0, respectively).
It means that although the heterogenised catalysts are less
selective to aldehydes in front of isomerization, the conver-
sion to the linear aldehyde is comparable to that obtained
with the homogeneous catalyst.
For catalyst Rh(ONP)/GF-Cl (6II), very similar features
to those indicated above have been found. The ratio Rh/C is
lower in this case. This can be due in part to a lower amount
of complex (0.72 wt.% Rh in sample Rh(ONP)/GF-Cl com-
pared to 0.86 wt.% Rh in sample Rh(COD)/GF-ONP) and
it can also be related to a deeper location of the complex
in the carbon porosity. The higher value of the P/C ratio
in sample Rh(COD)/GF-ONP is also expected because the
amount of diphosphine ligand present in carbon GF-ONP
is higher than the corresponding to the stoichiometry of the
metal complex. In relation to the location of the Rh complex,
it could be also considered that the presence of an excess of
phosphine ligands on the carbon surface might hinder the
penetration of the complex in the carbon porosity.
3.2. Solid state 31P NMR analysis
The 31P NMR spectra of sample GF-ONP (5) and also
of catalyst Rh(COD)/GF-ONP (6I) reveal partial oxidation
of the phosphine (δ = 31 ppm) [18]. A second analysis
carried on a GF-ONP sample, handled with care to avoid
exposition to air and sealed under inert atmosphere for the
analysis, reveals as well the oxidized form of the phosphine,
meaning that the phosphine group of the ligand anchored
to the support is not stable. It seems that oxidation is due
to the interaction with the carbon support, maybe related
to its extensive surface oxidation. On the other hand, in
catalyst Rh(ONP)/GF-Cl (6II), the state of P (δ = 25 ppm)
corresponds to a phosphine bonded to the metal [18]. That
is, when the ligand HONP forms the complex with rhodium,
the phosphine group is stabilised against oxidation.
The complex leaching expressed as percentage of Rh
removed from the catalyst, after the four catalytic runs,
for catalysts Rh(COD)/GF-ONP and Rh(ONP)/GF-Cl is
59 and 43%, respectively. After the first catalytic run, the
resulting solution shows a red wine colour. Analysis by 31
P
NMR of this solution reveals the presence of the diphos-
phine ligands in the solution. This confirms that part of the
complex is in solution. After subsequent catalytic runs, the
resulting solution is colourless and 31P NMR analysis does
not reveal the presence of phosphine. This observation can
be interpreted considering that part of the complex is more
loosely bound to the support and is leached in the first run.
The rest of the complex is effectively anchored. The gradual
loss of conversion found for catalyst Rh(COD)/GF-ONP
(6I) in the four consecutive catalytic runs may be, then,
related to a progressive deactivation of the supported
complex.
3.3. Catalytic activity
Fig. 1 shows the catalytic activity results obtained with
catalysts Rh(COD)/GF-ONP (6I) and Rh(ONP)/GF-Cl (6II)
The differences in catalytic activity between the two het-
erogenised catalysts can be related to the different nature of
the –ONP ligand, since as revealed by 31P NMR, phosphine
is partially oxidised in sample Rh(COD)/GF-ONP (6I).
The most relevant results are those corresponding to
runs 2–4, that is, on the reutilization of the heterogenised
catalysts. Both catalysts show an acceptable conversion
level and selectivity in four catalytic runs. With cata-
lyst Rh(COD)/GF-ONP (6I), the conversion to aldehydes
and the L/B ratio remain constant. The case of catalyst
Rh(ONP)/GF-Cl (6II) is intriguing since it not only remains
fully active, but the selectivity to aldehydes increases. The
TO (turnover, molaldehyde/molRh corresponding to the to-
tal reaction time) of this catalyst in the consecutive runs
Fig. 1. Catalytic activity of catalysts: homogeneous (᭜) Rh(COD) (2),
(᭡) Rh(COD)/GF-ONP (6I) and (᭹) Rh(ONP)/GF-Cl (6II).

182
M.C. Roma´n-Mart´ınez et al. / Journal of Molecular Catalysis A: Chemical 213 (2004) 177–182
are: 566, 583, 694 and 677 (calculated without considering
leaching) and the L/B ratio remains close to 2.
Acknowledgements
A comparison with literature on supported systems re-
veals that the catalytic performance of the catalysts studied
in this work is similar or better than that shown in previously
published papers. Thus, for example, Bourque et al. [18],
using Rh-complexed dendrimers also obtained for the hy-
droformylation of octene a L/B = 2 ratio, but the total con-
version (of 2 mmol substrate) is achieved after 22 h. Nozaki
et al. [11] reported the behaviour of polymer immobilized
chiral phosphine–phosphite–Rh(I) complexes for the hydro-
formylation of styrene and they found an interesting conver-
sion in the first run (about 80%), but it is strongly reduced in
the second one. Besides the L/B ratio is <1. More recently,
Wrzyszcz et al. [16] prepared Rh complexes supported on a
zinc aluminate spinel and found in the hydroformylation of
1-hexene a high conversion to aldehydes (96%) with L/R =
3. No leaching was detected although the catalysts were not
tested in a second run. Finally, Fierro et al. [15a] in their
work using Rh thiolate complexes anchored on phosphi-
nated silica for the hydroformylation of 1-heptene, obtain a
100% conversion (TO (as defined above) of about 300) and
a regioselectivity higher than 60% with L/B close to 2.5.
Furthermore, the catalysts do show activity after seven con-
secutive cycles. This brief revision on the behaviour of sup-
ported Rh complexes for olefine hydroformylation allows to
state that the results of the present work show a promising
perspective for the use of carbon materials as support for
the anchorage of metal complexes.
The authors thank project PPQ2002-01025 of the Spanish
Ministry of Science and Technology for financial support,
and J.A.D.A. thanks the same organism for the Ph.D. grant.
References
[1] V.A. Likholobov, B.L. Moroz, in: G. Ertl, H. Knözinger, J. Weitkamp
(Eds.), Handbook of Heterogeneous Catalysis, vol. 5, Wiley-VCH
Verlag, Weinheim, Germany, 1997, p. 2231.
[2] M. Lenarda, L. Storaro, R. Ganzerla, J. Mol. Catal. A: Chem. 111
(1996) 203.
[3] E. Lindner, T. Schneller, F. Auer, H.A. Mayer, Angew. Chem. Int.
Ed. 38 (1999) 2154.
[4] B. Cornils, W.A. Herrmann, R.W. Eckl, J. Mol. Catal. A: Chem. 116
(1997) 27.
[5] J.P. Arhancet, M.E. Davis, J.S. Merola, B. Hanson, Nature 339 (1989)
454.
[6] M.T. Reetz, G. Lohmer, R. Schwickardi, Angew. Chem. Int. Ed.
Engl. 36 (1997) 1526.
[7] A.N. Ajjou, H. Alper, J. Am. Chem. Soc. 120 (1998) 1466.
[8] K.K. Bando, K. Asakura, H. Arakawa, K. Isobe, Y. Iwasawa, J. Phys.
Chem. 100 (1996) 13636.
[9] V.L.K. Valli, H. Alper, Chem. Mater. 7 (1995) 359.
[10] Y. Zhang, H.-B. Zhang, G.-D. Lin, P. Chen, Y.-Z. Yuan, K.R. Tsai,
Appl. Catal. A 187 (1999) 213.
[11] K. Nozaki, F. Shibahara, Y. Itoi, E. Shirakawa, T. Ohta, H. Takaya,
T. Hiyama, Bull. Chem. Soc. Jpn. 72 (1999) 1911.
[12] J. Balúe, J.C. Bayón, J. Mol. Catal. A 137 (1999) 193.
[13] J.A. D´ıaz-Auñón, M.C. Román-Mart´ınez, C. Salinas-Mart´ınez de
Lecea, J. Mol. Catal. A 170 (2001) 81.
[14] J.A. D´ıaz Auñón, M.C. Román Mart´ınez, P. L’Argentiére, C. Salinas
Mart´ınez de Lecea, Stud. Surf. Sci. Catal. 130 (2000) 2075.
[15] (a) J.L.G. Fierro, M.D. Merchán, S. Rojas, P. Terreros, J. Mol. Catal.
A 166 (2001) 255;
4. Conclusions
(b) S. Rojas, P. Terreros, J.L.G. Fierro, J. Mol. Catal. A 184 (2002)
19.
[16] J. Wrzyszcz, M. Zawadzki, A.M. Trzeciak, J.J. Ziólokowski, J. Mol.
Catal. A: Chem. 189 (2002) 203.
[17] D.E. Bryant, M. Kilner, J. Mol. Catal. A 193 (2003) 83.
[18] S.C. Bourque, F. Maltais, W.J. Xiao, O. Tardiff, H. Alper, P. Arya,
L. Manzer, J. Am. Chem. Soc. 121 (1999) 3035.
[19] C.M. Crudden, D. Allen, M.D. Mikoluk, J. Sun, Chem. Commun.
(2001) 1154.
[20] B. Pugin, J. Mol. Catal. A: Chem. 107 (1996) 273.
[21] K.J. Stanger, J.W. Wiench, M. Pruski, R.J. Angelici, J. Mol. Catal.
A: Chem. 195 (2003) 63.
[22] A.R. Silva, M. Martins, M.M.M. Freitas, A. Valente, C. Freiere,
B. de Castro, J.L. Figueiredo, Microporous Mesoporous Mater. 55
(2002) 275.
[23] M.C. Román Mart´ınez, D. Cazorla Amorós, L. Linares Solano, C.
Salinas Mart´ınez de Lecea, Carbon 31 (1993) 895.
[24] H.P. Boehm, High Temp. High Pressures 22 (1990) 275.
[25] D. Briggs, M.P. Seah (Eds.), Practical Surface Analysis, vol. 1, John
Wiley and Sons, Chichester, UK, 1993.
A rhodium diphosphine complex was effectively an-
chored by covalent bonding on an activated carbon. The
complex is bound to the support via an esterification re-
action between a hydroxyl group in the ligand and the
acid-chloride-functionalised activated carbon surface. The
two synthetic routes developed for the heterogenisation (I:
ligand anchorage and then, complex synthesis and II: com-
plex synthesis followed by anchorage) give catalysts with
interesting catalytic properties, with the noteworthy feature
of being usable in consecutive catalytic cycles. However,
it must be pointed out that the anchoring procedure that
first synthesizes the Rh diphosphine catalyst (4) and then
bonds it to the carbon surface, is particularly attractive: the
resulting catalyst is fully active in four consecutive cat-
alytic cycles with a conversion to the linear aldehyde that
increases with reuse.