Rhodium phosphine complexes immobilized on silica as active catalysts for 1-hexene hydroformylation and arene hydrogenation

Article abstract of DOI:10.1016/j.molcata.2003.09.012

Rhodium phosphine complexes immobilized on silica as active catalysts were successfully tested with respect to their catalytic action on 1-hexene hydroformylation as well as benzene and toluene hydrogenation. The reaction outcome, i.e., the formation of aldehydes vs. isomerization, depended strongly on the presence and concentration of a phosphine co-catalyst. During immobilization, Hacac was completely removed after protonation by silica proton. In the presence of an excess of phosphinated silica P-{SiO₂}, the yield of aldehyde increased with concomitant suppression of the isomerization. The pre-catalyst [Rh(O-{SiO₂}(HO-{SiO₂})(CO)(P-{SiO ₂})] (1a) was more active both in hydroformylation and in isomerization of 1-hexene compared with [Rh(O-{SiO₂})(HO-{SiO₂})(P-{SiO ₂})₂]. Additionally, 1a catalyzed hydrogenation of aromatic hydrocarbons (benzene and toluene).

Full text of DOI:10.1016/j.molcata.2003.09.012

Journal of Molecular Catalysis A: Chemical 210 (2004) 179–187
Rhodium phosphine complexes immobilized on silica as active catalysts
for 1-hexene hydroformylation and arene hydrogenation
Christina M. Standfest-Hausera, Thomas Lummerstorfera, Roland Schmida,
Helmuth Hoffmanna, Karl Kirchnera,1, Michael Puchbergerb, Anna M. Trzeciakc,∗
,
c
d
a
´
Ewa Mieczynska , Włodzimierz Tylus , Józef J. Ziółkowski
a
Institute of Applied Synthetic Chemistry, Vienna University of Technology, Getreidemarkt 9, A-1060 Vienna, Austria
b
Institute of Material Chemistry, Vienna University of Technology, Getreidemarkt 9, A-1060 Vienna, Austria
c
Faculty of Chemistry, University of Wrocław, 14 F. Joliot-Curie Street, 50-383 Wrocław, Poland
d
Institute of Inorganic Technology, Wrocław University of Technology, Wybrzez˙e Wyspian´skiego 27, 50-370 Wrocław, Poland
Received 28 April 2003; received in revised form 5 September 2003; accepted 19 September 2003
Abstract
In immobilizing the rhodium complexes [Rh(acac)(CO)(P)] (1) and [Rh(acac)(P)2] (2) (P = Ph2PCH2CH2Si(OMe)3) onto SiO2, acetylace-
tone is found to be released through protonation of the acac ligand by the acidic silica-OH groups. The resulting complexes [Rh(O-{SiO2}(HO-
{SiO2})(CO)(P-{SiO2})](1a)and[Rh(O-{SiO2})(HO-{SiO2})(P-{SiO2})2](2a)weresuccessfullytestedwithrespecttotheircatalyticaction
on 1-hexene hydroformylation as well as benzene and toluene hydrogenation. The reaction outcome, viz. the formation of aldehydes versus
isomerization, depends strongly on the presence and concentration of a phosphine co-catalyst. Thus, while 1a gave only a 17% yield of
aldehyde in the absence of phosphines, the yield is increased to 54% in the presence of phosphinated silica P-{SiO2} or even 94% if PPh3 is
added to the solution. Without extra added phosphine, both 1a and 2a effect mainly the isomerization of 1-hexene to 2-hexene. Pre-catalyst
1a catalyzes also the hydrogenation of benzene at 10.5 atm H2 and 90 ◦C to give cyclohexane with a TOF of 608 h−1
.
© 2003 Elsevier B.V. All rights reserved.
Keywords: Rhodium complexes; Surface confined catalysts; Hydroformylation; Hydrogenation; Alkylsiloxanes
1. Introduction
ered catalyst activity [2] may partly be balanced with longer
lifetimes and easier regeneration. Among the available sup-
ports for homogeneous catalysts, silica is most frequently
selected as the carrier for rhodium catalysts used for the hy-
droformylation and/or hydrogenation of olefins. This choice
is appealing in view of the low price of silica, its chemical
neutrality, and a relatively easy immobilization procedure,
often leading to a stable heterogenized catalyst.
Several catalysis experiments with immobilized rhodium
complexes have been described in the literature. (i) Rhodium
carbonyl thiolate complexes supported on silica exhibit high
catalytic activity for the hydroformylation of 1-octene un-
der mild conditions (60 ◦C, 1 atm CO/H2) in the presence
of an excess of a phosphorus ligand [3–5]. Both highest
rate and regioselectivity have been reported in the case of
taking P(OPh)3. It should be mentioned that the chemose-
lectivity for aldehydes is rather low because of competing
isomerization of 1-octene [5]. (ii) Zwitterionic Rh(I) com-
plexes have been heterogenized on partially dehydroxylated
The separation of reaction products from the catalyst
is an important drawback in homogeneous metal com-
plex catalysis. Improvements may be sought in at least
two directions. On the one hand, one can switch over to
water-soluble catalysts as for instance accomplished in the
Ruhr–Chemie/Rhone–Poulenc industrial hydroformylation
process by rhodium [1]. However, this scheme is successful
for the hydroformylation of propylene only, whereas the
problem remains with the conversions of higher olefins.
Another way to circumvent the separation problem can be
the application of immobilized catalysts. An usually low-
∗
Corresponding author. Tel.: +48-71-3757-253;
fax: +48-71-3757-356.
E-mail addresses: kkirch@mail.zserv.tuwien.ac.at (K. Kirchner),
ania@wchuwr.chem.uni.wroc.pl (A.M. Trzeciak).
1
Co-corresponding author. Fax: +43-1-58801-15499.
1381-1169/$ – see front matter © 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2003.09.012

180
C.M. Standfest-Hauser et al. / Journal of Molecular Catalysis A: Chemical 210 (2004) 179–187
silica with the formation of hydrogen bonds between the
sulfonate ligand coordinated to rhodium and the OH group
of silica [6]. This arrangement has been shown to catalyze
the hydrogenation of ethene and propene under solid–gas
phase conditions, as well as the hydrogenation of styrene
and hydroformylation of 1-hexene under solid–liquid
phase conditions [6]. (iii) A rhodium complex containing
xanthene-based diphosphine, immobilized on polysilicate,
is capable of catalyzing the hydroformylation of 1-octene
to nonanal and its subsequent hydrogenation to 1-nonanol
under a pressure of 50 atm [7]. Variations observed in the
selectivity with the method of catalyst preparation appears
to be related to the variable acidic properties of the sil-
ica support [8]. (iv) Rhodium complexes incorporated into
a polysiloxane matrix by the sol–gel method have been
applied to the hydroformylation of 1-hexene at 20 atm of
droformylation (Low Pressure Oxo (LPO) process), in the
presence of an excess of free PPh3. The bisphosphite com-
plex Rh(acac)[P(OPh)3]2 has been used in the hydroformy-
lation of unsaturated alcohols, dienes, and olefins as well as
the hydrogenation of arenes [13–15].
2. Results and discussion
2.1. Preparation and characterization of immobilized
rhodium complexes
The rhodium complexes were supported on silica in a
two-step synthesis. In the first step, [Rh(acac)(CO)2] and
[Rh(acac)(C2H4)2] were reacted with Ph2PCH2CH2Si-
(OMe)3 (henceforth abbreviated to P) giving 1 and 2,
respectively, in 83 and 83% isolated yield. It should be
mentioned that it is not feasible to replace both carbonyls in
[Rh(acac)(CO)2] with phosphorus ligands owing to the high
␲-acceptor strength of CO [16,17]. Therefore, for 2 to be
obtained, the more labile complex [Rh(acac)(C2H4)2] was
used as the starting material [18]. Both 1 and 2 were char-
acterized by solution and solid state NMR spectroscopy,
IR spectroscopy and elemental analysis. The ν(CO) posi-
tion in the IR spectrum of 1 appears at 1975 cm−1, while
that in [Rh(acac)(CO)(PPh3)] is found at 1982 cm−1. This
points to a slightly increased ␴-donor property of P relative
to PPh3.
In the second step, 1 and 2 were grafted onto the silica
surface by the method described in Section 4. In a similar
way, P was tethered to silica (giving P-{SiO2}) [19,20] for
subsequent use as a modifying ligand in the hydroformyla-
tion reactions. However, things are not as simple as initially
believed and presented by the complexes [Rh(acac)(CO)(P-
{SiO2})] (1-{SiO2}) and [Rh(acac)(P-{SiO2})2] (2-{SiO2})
shown by the first steps in Schemes 1 and 2. On the ba-
sis of the shift of the 31P-CP/MAS signal from 52.4 to
38.1 ppm for 1 before and after immobilization a change
in the rhodium coordination sphere is implicated. In fact,
the GC–MS analysis of the toluene supernatants after im-
mobilization revealed the presence of free acetylacetone
(Hacac), obviously formed through protonation of the acac
CO/H2 with an average TOF of 164 molsub mol−ca1t −1. It is
h
worthy noting that the n/iso ratio increases in the presence
of PPh3 taken in excess [9]. (v) Cationic Rh(I) complexes
containing modified 1,3-bis(diphenylphosphinylpropane)
attached to polysiloxane-based ligands with a spacer of six
methylene units have been prepared as stationary phases
for chemistry in interphases and applied for hydrogena-
tion of 1-hexene at 10 atm of H2 giving TOF in toluene
of up to 1394 molsub mol−ca1t h−1 [10]. (vi) The activity of
Wilkinson-type catalysts immobilized on silica used for
the hydrogenation of dodecene, cyclohexanone, and bro-
mostyrene under H2 (1.1 bar) has been found to be even
comparable to homogeneous systems [11]. In addition,
the catalysts can be easily recovered and recycled many
times before gradual loss of activity. (vii) Rhodium com-
plexes with diphenylphosphine-functionalized carbosilane
dendrimers have been used as catalysts of 1-octene hydro-
formylation at 10 atm CO/H2, noting that dendrimers with
bidentate end groups gave less active catalysts relative to
the monodentate cognates [12].
In the present paper, we describe our experiments of im-
mobilizing two rhodium complexes, Rh(acac)(CO)(Ph2PC-
H2CH2Si(OMe)3) and Rh(acac)(Ph2PCH2CH2Si(OMe)3)2,
which may be treated as analogs of industrially employed ho-
mogeneous catalysts. [Rh(acac)(CO)(PPh3)] has been used
by Union Carbide since nearly 30 years for propylene hy-
Scheme 1.

C.M. Standfest-Hauser et al. / Journal of Molecular Catalysis A: Chemical 210 (2004) 179–187
181
Scheme 2.
ligand by the acidic silica OH groups. Quantitative anal-
ysis (GC–MS) shown that amount of Hacac evolved per
one immobilized Rh varied from 0.97 to 1.05, confirming
practically total removal of acac ligand during immobi-
lization. Therefore, 1- and 2-{SiO2} are better formulated
as [Rh(O-{SiO2})(HO-{SiO2})(CO)(P-{SiO2})] (1a) and
[Rh(O-{SiO2})(HO-{SiO2})(P-{SiO2})2] (2a) (second
steps in Schemes 1 and 2).
Hacac was also found in toluene solutions in which
rhodium complexes were reacted with P-{SiO2} and even
with neat silica, i.e. in solutions containing [Rh(acac)(CO)2]
+ P-{SiO2}, [Rh(acac)(CO)2] + SiO2, and [Rh(acac)(CO)-
(PPh3)] + SiO2. Thus, the acac anion has been replaced
with a Si-O− anion and a Si-OH group. A similar Hacac
elimination has been observed during the sol–gel impreg-
nation of [Rh(acac)(CO)(siloxantphos)] on silica [8] and
protonation of the allyl ligand in [Rh(␩3-C3H5)3] by silica
[21]. Also, during the immobilization of [Rh(acac)(CO)2]
on ZnAl2O4 spinel Hacac displacement became evident
[22]. The rhodium content of the heterogenized complexes
as determined by inductively coupled plasma (ICP) (after
treatment with HCl) was 1.65% for 1a and 0.48% for 2a.
From single pulse, 31P solid state NMR spectroscopy 2.5%
of Rh for 1a was determined.
spectroscopy. The two ν(CO) frequencies of similar inten-
sities found at 2004 and 2078 cm−1 may be assigned to a
silica-attached dicarbonyl rhodium species best described
by the formula [Rh(CO)2(O-{SiO2})(HO-{SiO2})] (3). It
structurally resembles the [Rh(CO)2]+ cation immobilized
on ZnAl2O4 spinel reported previously [22] with corre-
sponding IR signals appearing at 2015 and 2088 cm−1. The
same complex (based on IR data) was formed on treatment
of [Rh(acac)(CO)2] with neat SiO2. From this comparison it
can be concluded that silica oxygens are more basic than the
spinel oxygens. It is worth noting that the phosphine moiety
in P-{SiO2} does not displace CO in [Rh(acac)(CO)2] in
contrast to free P.
2.2. XPS studies 1 and 2 before and after immobilization
The X-ray photoelectron spectroscopy (XPS) method
has been used for quantitative and qualitative characteri-
zations of the obtained catalysts together with reference
samples. The binding energies for Rh 3d5/2, Rh 3d3/2
,
Si 2p and P 2p as well as atomic ratios of Rh/P and
Rh/Si on the surface of the samples are given in Table 1.
The numerical binding energies confirm the presence of
rhodium in the +1 oxidation state [3]. The slightly vary-
ing binding energies observed for Rh(+1) complexes un-
der study may be used for their identification (Fig. 1).
Rhodium/phosphorus atomic ratios calculated for 1a and
2a correspond very well to the stoichiometric ratios from
We employed also another method for tethering the
rhodium complex onto the silica surface by reacting
[Rh(acac)(CO)2] with P-{SiO2}. The product was iso-
lated, washed with toluene, dried, and characterized by IR
Table 1
XPS analysis of pre-catalysts 1a and 2a and the reference complexes Rh(acac)(CO)PPh3 and Rh(acac)(PPh3)2
Sample
Binding energy (eV)
Atomic ratio
Rh/P
Rh
Si
P
Rh/Si
3d5/2
3d3/2
2p
2p
1a
2a
309.38
309.33
308.60
309.8
309.34
309.33
314.12
314.07
313.64
314.48
314.08
314.07
103.4
103.37
132.55
132.2
131.64
1.04
0.51
0.717
2.78 × 10−2
1.84 × 10−2
Rh(acac)(CO)PPh3
3
Rh(acac)(PPh3)2
2
103.4
0.89 × 10−2
9.12 × 10−2
131.52
131.75
132.2
0.5
0.41
102.2
103.5
P-{SiO2}

182
C.M. Standfest-Hauser et al. / Journal of Molecular Catalysis A: Chemical 210 (2004) 179–187
the formulas [Rh(O-{SiO2}{HO-{SiO2})(CO)(P-{SiO2})]
and [Rh(O-{SiO2}{HO-{SiO2})(P-{SiO2})2].
Binding energies for Si 2p on the level 103.4–103.5 eV
characterize the SiO2 support. These figures were addition-
ally used for the calibration. The binding energy of Si 2p of
102.2 eV for [Rh(acac)(Ph2PCH2CH2Si(OMe)3)2] is in ac-
cordance with the silicones character of the Si(OMe)3 group.
The symmetric shape of the P 2p spectra suggest only one
form of phosphorus present in each system (PPh2 or PPh3).
Fig. 2 shows the survey scan of modified silica and a narrow
scan of P 2p peak. The determination of binding energies
of P 2p were quite difficult due to the small concentration
of P ([P]:[Si] = 2.2 × 10−2) and partial peak overlap with
the satellites from Si 2p. Nevertheless, we were able to fit a
value of ca. 132.2 eV.
2.3. Hydroformylation of 1-hexene in the presence of 1a
Generally, the products obtained are aldehydes and the
isomerization product 2-hexene according to Scheme 3 with
the product ratio varying dramatically with the kind and
amount of phosphine present as a co-catalyst. The results are
shown in Table 2 valid for standard 1-hexene hydroformy-
lation reaction conditions and toluene as the solvent. As is
seen, in the absence of extra phosphine (entry 2) the conver-
sion was almost quantitative alas with mainly isomerization
having taken place. This should be compared with the re-
sults of the blank reaction with P-{SiO2} in the absence of a
rhodium complex at otherwise the same conditions in which
case only 5% 2-hexene are obtained (entry 1). When the re-
cycled catalyst (after removal of liquids by vacuum transfer)
was employed (entries 3 and 4), the products yield remained
Fig. 1. XPS Rh 3d spectra of samples: (a) 3, (b) 1, (c) 1a, (d) 2a, (e)
(acac)Rh(PPh3)2, (f) (acac)Rh(CO)(PPh3).
Fig. 2. XPS survey scan of phosphine supported on silica (P-{SiO2}) and P 2p core-level spectra.

C.M. Standfest-Hauser et al. / Journal of Molecular Catalysis A: Chemical 210 (2004) 179–187
183
best result (entry 8) was obtained after pre-treatment of
1a with H2/CO. Under this condition 1a might have been
transformed to a hydrido-carbonyl species, which is likely
the active form of the catalyst. According to IR measure-
ments, 1a was transformed to a new species after only 1 h
treatment under H2/CO (10 atm, 80 ◦C).
It is not unexpected that compared to P-{SiO2} the pres-
ence of free PPh3 is more effective with the results strongly
varying with the phosphine concentration (entries 10–14).
Thus the yield of aldehydes increased tremendously when
the [PPh3]:[Rh] ratio changed from 5 to 19. Likewise, the
n/i ratio increased in the same direction. Such a great effect
of the phosphine concentration is worth emphasizing. At the
lowest concentration used, [PPh3]:[Rh] = 5, the product ra-
tio was practically identical with that when no phosphine
was present, with 2-hexene being the main product (82%
yield). Doubling the concentration of phosphine increased
the aldehyde yield from 14 to 86%. Further doubling re-
sulted in a 94% yield.
Scheme 3.
An excellent result (95% aldehydes) was obtained for cat-
alyst 3, in the presence of an excess of free PPh3. Without
adding extra PPh3, only 2-hexene was obtained in ca. 85%
yield.
Since ICP analysis of rhodium in 1a after hydroformy-
lation showed a decrease from 1.65 to 1.44–1.45% (entries
11 and 12), or 1.1% (entry 13) depending on separation
procedure, some leaching of rhodium during the reaction
and participation of a soluble rhodium complex cannot
be excluded in the system containing PPh3 co-catalyst. In
similar to this one in the first reaction (entry 2) (2-hexene:
82–84%, aldehydes: 13–15%). When switching over to neat
1-hexene, without toluene, the conversion decreased to 14%
aldehydes and 71% 2-hexene (entry 6). Furthermore, lower-
ing the temperature to 60 ◦C made the yield drop to the low
value of 12% 2-hexene (entry 5).
In the presence of an excess of phosphinated silica
P-{SiO2} (entries 7 and 9), the yield of aldehydes increased
with concomitant suppression of the isomerization. The
Table 2
Hydroformylation of 1-hexene with 1aa
Entry
Additive
[P]:[Rh] ratio
Conversion (%)
Aldehydes yield (%)
2-Hexeneb yield (%)
n/i
Rh (%)
c
1
2
3
4
5
6
7
8
9
10
11
12
13
14
P-{SiO2}
0
0
0
0
0
0
5
98
95
97
12
85
92
70
86
96
99
96
98
98
–
17
13
15
–
14
46
54
53
14
61
49
86
94
5
81
82
82
12
71
46
16
33
82
38
47
12
4
–
–
1.65
–
2.0
2.1
2.0
2.1
2.2
2.6
2.6
2.7
2.6
2.4
2.9
2.7
3.6
d
–
d
–
e
–
f
–
P-{SiO2}
P-{SiO2}
P-{SiO2}
PPh3
PPh3
PPh3
1.3
1.3g
2.6
5h
7.0h
7.0h,i
9.5g
19.0g
1.63j
1.65j
1.44j
1.45j
1.1k
PPh3
PPh3
a
Reaction conditions: 80 ◦C, 10 atm H2/CO, 5 h, 1.5 cm3 of toluene, 1.5 cm3 of 1-hexene, [1-hexene]:[Rh] = 2400, P-{SiO2} = Ph2PCH2CH2Si-{SiO2}.
The mixture of cis/trans 2-hexenes/3-hexene was not found.
Blank experiment without 1a pre-catalyst.
Catalyst recovered from the previous reaction and reused.
Rection at 60 ◦C.
b
c
d
e
f
Reaction in neat 1-hexene.
g
h
i
Pre-catalyst 1a activated for 3 h in toluene at 80 ◦C under 10 atm of H2/CO before introduction of substrates.
Reaction time = 2.5 h.
Catalyst recovered and reused with the new portion of PPh3.
Rhodium content (determined by ICP) in pre-catalyst separated after reaction by ultrafiltration.
Rhodium content (determined by ICP) in pre-catalyst separated after reaction by filtration.
j
k

184
C.M. Standfest-Hauser et al. / Journal of Molecular Catalysis A: Chemical 210 (2004) 179–187
contrast, the rhodium content in 1a recovered after reaction
performed with P-{SiO2} was practically unchanged (entry
9). We observed, that at higher excess of PPh3 ([PPh3]:[Rh]
> 7) the density of solution increased and separation of cat-
alyst by normal filtration was not complete. Some amounts
of catalyst still remained in solution forming very fine sus-
pension. This was confirmed by two experiments in which
post-reaction solution was used for the next hydroformy-
lation reaction. The solution obtained by normal filtration
(from entry 13) gave 39% of aldehydes and only 13% of
aldehydes were obtained with ultra-centrifugated solution
from entry 11.
Scheme 4.
A slightly better result (7% cyclohexane) was obtained
with a catalyst that was recovered after 1-hexene hydro-
formylation. It can thus be concluded that CO acts as an
inhibitor of the hydrogenation reaction, obviously due to
strong bonding to rhodium. IR studies showed the presence
of rhodium carbonyl clusters in a sample of 1a reacted with
CO.
2.4. Hydroformylation of 1-hexene with 2a
2.6. IR studies of reactions of 1a with H2, CO,
and H2/CO
Catalytic tests with 2a were performed under the same
conditions as for 1a. The reaction in neat 1-hexene pro-
duced only 29% of 2-hexene and traces of aldehydes. A
better result (65% 2-hexene, 5% aldehydes) was obtained
with toluene as the solvent (Table 3). The recycled catalyst
was less active and produced mainly 2-hexene (90%) and
only 7% of aldehydes. However, the product distribution re-
mained unchanged upon using the recovered catalyst for the
third time. The fact that 2-hexene was the main product in-
stead of aldehydes appears to be related to the high ability
of 2a to catalyze the double bond shift noting the low n/i
ratio (ca. 1.4) of the aldehydes formed. When the reaction
was performed in the presence of PPh3 very low conversion
(4%) of 1-hexene to 2-hexene was observed.
IR spectroscopy was used for the preliminary identifica-
tion of the immobilized complexes formed upon treatment of
1a with both H2 and CO as well as with an H2/CO mixture
in the absence of 1-hexene. All experiments were performed
in toluene (80 ◦C, 10 atm, 2 h). The IR spectra of the isolated
complexes are shown in Fig. 3. Spectrum (A) belongs to 1a
heated under CO indicating the formation of a new species.
The ν(CO) at 1975 cm−1 of the starting complex disappeared
while three new bands at 1802 cm−1 (bridging CO), 2009,
and 2076 cm−1 (both terminal CO) formed which may be
attributed to a Rh(0) carbonyl species (Table 5). We suspect
that the reduction of rhodium in the presence of CO is facil-
itated at the silica surface, since under the same conditions
from soluble [Rh(acac)(CO)(PPh3)] only small amounts of
[Rh(acac)(CO)2] are formed but no Rh(0) species. The par-
ticipation of silica in the reduction process has been reported
also for other supported rhodium complexes [23]. Likewise,
the formation of rhodium carbonyls containing bridging CO
groups was observed when 1a reacted with CO in the pres-
ence of P-{SiO2}. In contrast, no noticeable reaction with
CO took place in the presence of PPh3 as indicated by the
ν(CO) band at 1976 cm−1 being essentially identical with
that of 1a.
2.5. Hydrogenation of benzene and toluene with
pre-catalyst 1a
Pre-catalyst 1a used alone (i.e. without a phosphine
co-catalyst) catalyzed also the hydrogenation of aromatic
hydrocarbons such as benzene and toluene with 14–17%
yields in 3 h (Scheme 4). The results are presented in
Table 4. The yield increased to 31% when the catalyst
was pretreated by heating at 90 ◦C under an atmosphere
of H2 (10.5 atm) for 3 h. The completely opposite effect,
i.e. inhibition of hydrogenation, occurred upon treatment
of 1a with H2/CO before adding benzene in which case no
cyclohexane was formed.
The reaction of 1a with H2 leads to an unstable hydrido
species characterized by a strong absorption at 2015 cm−1
(ν(Rh–H)) together with lower intensity bands at 2040 and
Table 3
Hydroformylation of 1-hexene with 2aa
Entry
Catalyst (g)
Solvent
Conversion (%)
2-Hexene (%)
Aldehydes (n + i) (%)
n/i
15
16
17
18
0.08
–
30
70
97
96
29
65
90
89
1
5
7
7
1.3
1.9
1.4
1.4
0.08
0.08b
0.07c
Toluene
Toluene
Toluene
a
Reaction conditions: 80 ◦C, 10 atm H2/CO, 5 h, 1.5 cm3 of toluene, 1.5 cm3 of 1-hexene or 3 cm3 of 1-hexene; [1-hexene]:[Rh] = 3200.
b
c
Second run with recycled catalyst.
Third run with recycled catalyst.

C.M. Standfest-Hauser et al. / Journal of Molecular Catalysis A: Chemical 210 (2004) 179–187
185
Table 4
Hydrogenation of benzene and toluene with 1aa
Entry
Substrate
Pre-treatment a catalyst in the absence of substrate
Product
Yield (%)
TOF (h−1
)
19
20
21
22
23
24
Benzene
Benzene
Benzene
Benzene
Benzene
Toluene
Cyclohexane
Cyclohexane
Cyclohexane
–
Cyclohexane
Methyl-cyclohexane
14
17
31
–
7
17
274
494
608
Catalyst recovered after benzene hydrogenationb
3 h, 10.5 atm of H2
3 h, 10.5 atm of H2/CO
Catalyst recovered after hydroformylation (5 h)
137
278
a
Reaction conditions: 10.5 atm of H2, 90 ◦C, 4 cm3 of substrate (without solvent), mcat 0.03 g (1.65% Rh), 5 h.
mcat 0.02 g.
b
2080 cm−1 (B). A very weak ν(CO) band of a bridging CO
ligand was observed at 1844 cm−1. Similar spectra were
obtained for the reaction products of 1a heated with an
H2/CO mixture as well as for the catalyst isolated after the
hydroformylation reaction test (C). The IR spectra of fresh
samples taken out of the autoclave showed bands only in the
2000–2068 cm−1 range, which can be assigned to terminal
ν(CO) frequencies and ν(Rh–H). In the region of bridging
CO groups (1800–1850 cm−1) merely low intensity bands
were observed. However, after a few minutes of evacuation
an intensive absorption at 1800 cm−1 appeared confirming
Table 5
IR data (in KBr) of 1a exposed to CO, H2 and CO/H2 (80 ◦C, 10 atm, 2 h)
Reaction of 1a with
1a (neat)
ν(CO) and ν(Rh–H)
1975
CO
1800 m, 2009 m, 2076 s
1800 m, 2003 m, 2073 s
1976
CO, P-{SiO2}
CO, PPh3
H2
2015 s, 2040 m, 2074 s
H2/CO
2022 s, 2048 s, 2075 s
H2/CO, evacuation
H2/CO, PPh3
1800 m, 2021 m, 2045 m, 2073 vs
1942 m, 1970 vs, 2005 s, 2050 w
the formation of a Rh(0) carbonyl complex with bridging
CO. Also, in this case, the formation of an Rh–CO–Rh core
is prevented by added PPh3 and, therefore, only terminal
carbonyls are detected in the IR spectrum (Table 5). The
catalyst isolated after a reaction performed with the addition
of PPh3 shows an IR spectrum with three bands at 1946,
1974, and 2050 cm−1 which may be assigned to ν(CO) of
terminal carbonyls and possibly to ν(Rh–H).
3. Conclusions
1. Rhodium complexes [Rh(acac)(CO)(Ph2PCH2CH2Si-
(OMe)3)] (1) and [Rh(acac)(Ph2PCH2CH2Si(OMe)3)2]
(2) were found as convenient precursors for immobiliza-
tion at silica and designing of new pre-catalysts.
2. It was proved, that during immobilization Hacac is com-
pletely removed after protonation by silica proton.
3. Immobilized complexes, [Rh(O-{SiO2}(HO-{SiO2})
(CO)(P-{SiO2})] (1a) and [Rh(O-{SiO2})(HO-{SiO2})
(P-{SiO2})2] (2a), were characterized with 31P-CP/MAS,
XPS and IR methods. The Rh/P ratios on the surface
were determined as equal 1.04 for 1a and 0.51 for 2a,
confirming complexes composition.
4. The pre-catalyst 1a was found more active both in hydro-
formylation and in isomerization of 1-hexene compared
with 2a. Additionally 1a catalyze hydrogenation of aro-
matic hydrocarbons (benzene and toluene).
5. P-{SiO2} as well as PPh3 can be used as co-catalysts
with 1a, however, high concentration of PPh3 may lead to
Fig. 3. IR spectra of complex 1a after reaction with: (A) CO, (B) H2,
(C) H2/CO (1:1) at 10 atm and 80 ◦C for 2 h.

186
C.M. Standfest-Hauser et al. / Journal of Molecular Catalysis A: Chemical 210 (2004) 179–187
some rhodium leaching as well as creates some problems
with separation of catalyst from the solution.
to the analysis chamber. Stability of Rh spectra during the
XPS measurements was proved by a series of successive
measurements. The spectra were collected and processed by
SpecsLab software. The data analysis procedure involved the
following steps: data smoothing, background calculation.
4. Experimental
4.1. General remarks
4.2. Synthesis of complexes 1 and 2
All manipulations were performed under an inert atmo-
sphere of nitrogen or argon by using Schlenk techniques
and a glove box. All chemicals were standard reagent grade
and used without further purification. The solvents were
purified according to standard procedures. The deuterated
solvents were dried over 4 Å molecular sieves. All experi-
ments were carried out with Merck silica (specific surface
area 388 2 m2 g−1). The silica has been treated in an
UV-cleaning chamber prior to use to remove adsorbed
organic impurities but has not been dried at elevated tem-
peratures. [Rh(acac)(CO)2] and [Rh(acac)(C2H4)2] were
obtained according to literature methods [24,25].
4.2.1. [Rh(acac)(CO)(Ph2PCH2CH2Si(OMe)3)] 1
[Rh(acac)(CO)2] (100 mg, 0.387 mmol) and Ph2PCH2-
CH2Si(OMe)3 (130 mg, 0.387 mmol) dissolved in 5 ml
CH2Cl2 were stirred for 4 h at room temperature. The
color changed from orange to dark green. After evaporation
of the solvent the product was washed twice with Et2O
(10 ml) and separated as a dark green oil. Yield: 182 mg
(83%). Analytically calculated for C23H30O6PRhSi (MG =
564.46 g mol−1): C, 48.94; H, 5.34. Found: C, 49.09; H,
5.22. 1H NMR (acetone-d6, 20 ◦C): δ = 7.61–7.47 (m,
10H, PPh2), 5.51 (s, 1H, MeCOCHCOMe), 3.49 (s, 9H,
Si(OMe)3), 2.01 (s, 3H, MeCOCHCOMe), 1.81 (s, 3H,
MeCOCHCOMe), 2.57–2.50 (m, 2H, PCH2CH2), 0.85–0.78
(m, 2H, PCH2CH2). 13C{1H} NMR (acetone-d6, 20 ◦C):
δ = 190.4 (dd, JCRh = 76.1 Hz, JCP = 25.6 Hz, CO), 188.0
(s, MeCOCHCOMe), 185.4 (s, MeCOCHCOMe), 134.1
(d, JCP = 48.4 Hz, Ph1), 133.5 (d, JCP = 10.4 Hz, Ph2,6),
130.7 (d, JCP = 2.0 Hz, Ph4), 128.7 (d, JCP = 9.7 Hz,
Ph3,5), 100.5 (s, MeCOCHCOMe), 50.2 (s, Si(OMe)3),
27.1 (s, MeCOCHCOMe), 26.4 (s, MeCOCHCOMe), 20.8
1
Solution H, 13C{1H}, and 31P{1H} NMR spectra were
recorded on a Bruker Avance-250 spectrometer operating
at 250.13, 62.86, and 101.26 MHz, respectively, and were
referenced to SiMe4 and H3PO4 (85%). Infrared spec-
tra were recorded on Mattson RS-10000, Nicolet Impact
400 and Perkin-Elmer 16PC FT-IR spectrometers. All
solid-state 31P-NMR spectra were recorded on a Bruker
Avance 300 spectrometer (standard bore), equipped with a
4 mm broad-band MAS probe-head and ZrO2 rotors. The
rotational speed for all experiments was 12 kHz. Phospho-
rous spectra were referenced with respect to 85% H3PO4(aq)
by setting the 31P NMR peak of solid PPh3 to −7.0 ppm.
31P-CP/MAS spectra (cross polarization with magic angle
spinning) were measured with a contact time of 3 ms and
a relaxation delay of 6 s. Quantitative measurements were
done by 31P single pulse experiments with a relaxation de-
lay time of 120 s. The reaction products obtained from hy-
droformylation and hydrogenation reactions were analyzed
with GC–MS Hewlett-Packard 5890 II + Hewlett-Packard
5971 A instruments.
XPS spectra were recorded on SPECS UHV/XPS/AES
system using Mg-K␣ source for excitation and equipped
with a hemispherical analyzer operating in fixed analyzer
transmission (FAT) mode with pass energy 10 or 20 eV. A
power setting of 10 kV and 120 W was applied. The working
pressure in an analyzing chamber was <5 × 10−10 mbar.
The binding energy (BE) scale was calibrated by taking the
Au 4f7/2 peak at 84 eV. Correction of the energy shift due to
the static charging of the samples was accomplished using as
reference the C 1s peak at 284.8 eV and additionally where
it was possible the Si 2p peak taken at 103.4 eV. The accu-
racy of the reported binding energies was 0.1 eV. For the
XPS analysis, all sample specimens were obtained by press-
ing sample powders into thin disks which were mounted
on sample holders and placed in a pre-chamber, outgassed
to <10−8 mbar at room temperature, and then transferred
1
2
(d, JCP = 29.1 Hz, PCH2CH2), 4.3 (d, JCP = 2.8 Hz,
PCH2CH2). 31P{1H} NMR (acetone-d6, 20 ◦C): δ = 50.0
1
(d, JPRh
=
173.7 Hz, PPh2). IR ν(CO) 1975 cm−1
.
31P-CP/MAS spectrum: δ = 52.4 ppm.
4.2.2. [Rh(acac)(Ph2PCH2CH2Si(OMe)3)2] 2
[Rh(acac)(C2H4)2] (100 mg, 0.387 mmol) and Ph2-
PCH2CH2Si(OMe)3 (258 mg, 0.772 mmol) dissolved in
5 ml CH2Cl2 were stirred for 4 h at room temperature. The
color changed from orange to red. After evaporation of
the solvent the product was washed twice with petroleum
ether (10 ml) and separated as a yellow solid. Yield: 278 mg
(83%). Analytically calculated for C39H53O8P2RhSi2
(MG = 870.02 g mol−1): C, 53.79; H, 6.13. Found: C,
53.97; H, 6.40. 1H NMR (acetone-d6, 20 ◦C): δ = 8.11–7.93
(m, 4H, PPh2), 7.65–7.43 (m, 4H, PPh2), 7.40–7.20
(m, 8H, PPh2), 6.83–6.70 (m, 4H, PPh2), 5.36 (s, 1H,
MeCOCHCOMe), 3.43 (s, 18H, Si(OMe)3), 3.25–3.12
(m, 2H, PCH2CH2), 3.04–2.83 (m, 2H, PCH2CH2), 1.97
(s, 6H, MeCOCHCOMe), 0.29–0.14 (m, 2H, PCH2CH2),
0.12–0.05 (m, 2H, PCH2CH2). 31P{1H} NMR (acetone-d6,
20 ◦C): δ = 33.9 (d, JPRh = 132.7 Hz, PPh2). 13C{1H}
1
NMR (acetone-d6, 20 ◦C): δ = 184.4 (s, MeCOCHCOMe),
135.0 (t, J = 4.60 Hz, Ph), 133.4 (t, J = 3.45 Hz, Ph),
131.1 (t, J = 1.41 Hz, Ph), 131.1 (d, J = 9.12 Hz,
Ph), 130.5 (t, J = 1.44 Hz, Ph), 128.9 (d, J = 11.5 Hz,
Ph), 128.5 (t, J = 4.56 Hz, Ph), 128.3 (t, J = 5.04 Hz,
Ph), 98.6 (s, MeCOCHCOMe), 50.2 (s, Si(OMe)3), 27.5

C.M. Standfest-Hauser et al. / Journal of Molecular Catalysis A: Chemical 210 (2004) 179–187
187
(dt, J1 = 4.56 Hz, J2 = 0.64 Hz, PCH2CH2), 22.7
(s, MeCOCHCOMe), 3.3 (t, J = 5.04 Hz, PCH2CH2).
31P-CP/MAS spectrum: δ = 35.8 ppm.
KBN 15/T09/99/01D (A.M. Trzeciak and J.J. Ziółkowski).
Financial support by the “Fonds zur Förderung der wis-
senschaftlichen Forschung” with the Projects No. P14681-
CHE and P12769-CHE is also gratefully acknowledged (K.
Kirchner and H. Hoffman).
4.3. Immobilization of Ph2PCH2CH2Si(OMe)3, 1, and 2
The phosphine ligand Ph2PCH2CH2Si(OMe)3 as well as
complexes 1 and 2 were immobilized by rigorously stirring
a suspension of SiO2 and an appropriate amount of ligand
or complex (typically 0.280 mmol of substrate per 250 mg
of SiO2) in 18 ml of toluene under an atmosphere of dry
nitrogen for 24 h at room temperature. The solid was sep-
arated by centrifuging. Upon removal of the supernatant,
the silica was washed several times with dry toluene and
CH2Cl2 and again centrifuged. The samples were dried un-
der vacuum at room temperature for 24 h and characterized
by means of 31P-CP/MAS NMR spectroscopy exhibiting
signals at −14.7, 35.3, and 37.2 ppm, respectively, for im-
mobilized Ph2PCH2CH2Si(OMe)3, 1, and 2.
References
[1] B. Cornils, E. Wiebus, Chemtech. 1999, 33.
[2] B. Cornils, W.A. Herrmann, in: B. Cornils, W.A. Herrmann (Eds.),
Applied Homogeneous Catalysis with Organometallic Compounds,
Wiley–VCH, New York, 1996.
[3] S. Rojas, P. Terreros, J.L.G. Fierro, J. Mol. Catal. A: Chem. 184
(2002) 19.
[4] H. Gao, R.J. Angelici, J. Mol. Catal. A: Chem. 145 (1999) 83.
[5] H. Gao, R.J. Angelici, Organometallics 17 (1998) 3063.
[6] C. Bianchini, D.G. Burnaby, J. Evans, P. Frediani, A. Meli, W.
Oberhauser, R. Psaro, L. Sordelli, F. Vizza, J. Am. Chem. Soc. 121
(1999) 5961.
[7] P.W.N.M. van Leeuwen, A.J. Sandee, J.N.H. Reek, P.C.J. Kamer, J.
Mol. Catal. A: Chem. 182–183 (2002) 107.
[8] A.J. Sandee, J.N.H. Reek, P.C.J. Kamer, P.W.N.M. van Leeuwen, J.
Am. Chem. Soc. 123 (2001) 8468.
[9] E. Lindner, F. Auer, A. Baumann, P. Wegner, H.A. Mayer, H. Bertag-
nollo, U. Reinohl, T.S. Ertel, A. Weber, J. Mol. Catal. A: Chem.
157 (2000) 97.
[10] E. Lindner, S. Brugger, S. Steinbrecher, E. Plies, M. Seiler, H.
Bertagnollo, P. Wegner, H.A. Mayer, Inorg. Chim. Acta 327 (2002)
54.
The surface coverage was determined by means of 31P
MAS NMR spectroscopy integrating the 31P signals of mix-
tures of the surface-immobilized compounds together with a
weighted amount of [NBun4]PF6 as a standard utilizing single
pulse experiments with a pulse delay of 120 s. The integral of
the 31P signal of PF6 counterion in 1 has been set to unity.
−
Values for the surface coverage for Ph2PCH2CH2Si(OMe)3
and 1 are 52 molecules/100 nm2 (8.63·10−4 mmol m−2) and
44 molecules/100 nm2 (7.3·10−4 mmol m−2 = 2.5 mol%
Rh).
[11] C. Merckle, S. Haubrich, J. Blumel, J. Organomet. Chem. 627 (2001)
44.
[12] D. de Groot, P.G. Emmerink, C. Caucke, J.N.H. Reck, P.C.J. Kamer,
P.W.N.M. van Leeuwen, Inorg. Chem. Commun. 3 (2000) 711.
[13] D. Pie˛ta, A.M. Trzeciak, J.J. Ziółkowski, J. Mol. Catal. 75 (1982)
302.
4.4. Hydroformylation and hydrogenation reactions
[14] A.M. Trzeciak, J.J. Ziółkowski, J. Mol. Catal. 34 (1986) 213.
[15] A.M. Trzeciak, J.J. Ziółkowski, J. Mol. Catal. 43 (1988) 335.
[16] A.M. Trzeciak, T. Głowiak, R. Grzybek, J.J. Ziółkowski, J. Chem.
Soc. Dalton Trans. (1997) 1831.
[17] Yu.S. Varshavsky, M.R. Galding, T.G. Cherkasova, I.S. Podkorytov,
A.B. Nikolskii, A.M. Trzeciak, Z. Olejnik, T. Lis, J.J. Ziółkowski,
J. Organomet. Chem. 628 (2001) 195.
[18] H. Suzuki, S. Matsuura, Y. Moro-Oka, T. Ikawa, J. Organomet.
Chem. 286 (1985) 247.
[19] C.M. Standfest-Hauser, T. Lummerstorfer, R. Schmid, K. Kirchner,
H. Hoffmann, M. Puchberger, Monatsh. Chem. 134 (2003) 1167.
[20] A. Carmona, A. Corma, M. Iglesias, A. San José, F. Sánchez, J.
Organomet. Chem. 492 (1995) 11.
Hydroformylation reactions have been performed in a
steel autoclave (50 cm3) with magnetic stirring at 80 ◦C and
10 atm H2/CO = 1. A suitable amount of catalyst in small
Teflon vessel, 1.5 cm3 of toluene and 1.5 cm3 of 1-hexene
were introduced to the autoclave in N2 atmosphere. The au-
toclave was closed, filled with H2 (5 atm) and with CO up
to 10 atm. After the reaction the autoclave was cooled down
and the liquid sample was analyzed by GC–MS. Hydrogena-
tion reactions have been performed in a similar manner in
150 cm3 steel autoclave at 90 ◦C and 10.5 atm of H2.
[21] J.M. Basset, F. Lefebrre, C. Santini, Coord. Chem. Rev. 178–180
(1998) 1703.
[22] J. Wrzyszcz, M. Zawadzki, A.M. Trzeciak, J.J. Ziółkowski, J. Mol.
Catal. A: Chem. 189 (2002) 203.
Acknowledgements
[23] M. Niwa, J.H. Lunsford, J. Catal. 75 (1982) 302.
[24] Y.S. Varshavsky, T.G. Cherkasova, Zh. Neorg. Khim. 12 (1967) 1709.
[25] R. Cramer, Inorg. Synth. 15 (1976) 14.
This work was supported by the State Committee for
Scientific Research KBN (Poland) with the Grant PBZ