A facile method for promoting activities of ordered mesoporous silica-anchored Rh-P complex catalysts in 1-octene hydroformylation

Article abstract of DOI:10.1039/b900591a

This work deals with the promotion of immobilized Rh catalyst activities in olefin hydroformylation by lengthening the alkyl spacers and choosing an active Rh precursor. The flexibility of long chain alkyls was used to free the motion of the anchored Rh c

Full text of DOI:10.1039/b900591a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/greenchem | Green Chemistry
A facile method for promoting activities of ordered mesoporous
silica-anchored Rh–P complex catalysts in 1-octene hydroformylation†
Wei Zhou and Dehua He*
Received 12th January 2009, Accepted 30th April 2009
First published as an Advance Article on the web 14th May 2009
DOI: 10.1039/b900591a
This work deals with the promotion of immobilized Rh catalyst activities in olefin
hydroformylation by lengthening the alkyl spacers and choosing an active Rh precursor. The
flexibility of long chain alkyls was used to free the motion of the anchored Rh complex in
confining porous spaces so as to increase the substrate activation efficiency. RhCl(CO)(PPh3)2 was
anchored to mesoporous silicas SBA-15 and MCM-41 through n-alkyl (carbon number n = 1, 3,
5, 8, 11) as spacer. The catalysts were characterized by X-ray powder diffraction (XRD),
isothermal N2 sorption analysis and inductive coupling plasma-atomic emission spectroscopy
(ICP-AES). The multi-step-assembly of the complex into surface-organic modified mesoporous
silica was monitored and evidenced by infrared spectroscopy (IR). The prepared Rh-immobilized
catalysts were applied in 1-octene hydroformylation to nonyl aldehydes. Rh leaching in the
reaction solutions was determined by ICP-MS, which was found to be satisfyingly low along with
successful catalyst cycles. With the increase of the alkyl spacer length, the specific activity (shown
as turnover number, TON) and stability of the immobilized catalyst were both increased. The
catalyst prepared by the longest alkyl spacer (n = 11, C11) revealed activity comparable to the
homogeneous counterpart and resistance to the by-product poisoning. RhCl(CO)(PPh3)2-
supported catalysts were also prepared by impregnating RhCl(CO)(PPh3)2 into SBA-15 and
MCM-41 without the presence of diphenylphosphinoalkyl, the prepared catalysts showed a sharp
activity decrease within only two catalyst cycles, owing to the significant active species leaching.
is exactly the origin of the concept of a molecular sieve.
1. Introduction
Consequently, the ordered mesopores have advantages over
Zeolite chemists take for granted that the invention1 of or-
amorphous silica, nano SiO2 and active carbon, since they offer
dered mesoporous silica (hereafter OMS) is consequent to the
a uniform confining function and activity sites thereby form.14
requirement of processing large molecules, either as active com-
Currently, as the largest industrial homogeneous process,
positions or reactants. These materials are synthesized by self-
hydroformylation of olefins15 with syngas (equimolar CO
assembled surfactants as template,2,3 which can be removed by
and H2) produces aldehydes with one more carbon than
calcination or solvent-extraction. With hexagonal symmetry and
the initially employed olefin. The products, including al-
honeycomb-like pore channels, MCM-414 and SBA-155,6 are ap-
cohols by further hydrogenation of aldehydes, are widely
parently the most widely studied OMS referring to applications
employed in the production of detergents, solvents, soften-
such as separation, chemical sensing7 and catalysis.8 OMSs are
ers and plasticizers with great market profits. Hydroformy-
promising supports for hosting catalytic species varying from
lation can be catalyzed by Rh–P complexes, for example
monometallic complexes9,10 to nanoparticles,11,12 benefiting from
oil-soluble RhCl(PPh3)3, RhCl(CO)(PPh3)2 or water-soluble
their high specific surface area, high adsorption capacity, easy
RhCl(CO)(TPPTS)2 (TPPTS = triphenylphosphine-3,3¢,3¢¢-
surface modification and certain hydro/solvothermal stabilities.
trisulfonic acid trisodium salt). The reaction goes under mild
In addition, the confining effects13 of porous materials enhance
conditions with a great turnover number (TON) of substrate.
the intermolecular interactions inside the nanospaces, which
Heterogenization of the Rh complex to the support has po-
tential industrially, since it aims at easy catalyst separation
and recycling, atom-economy and low metal loss, and in this
Innovative Catalysis Program, Key Lab of Organoelectronics &
Molecular Engineering of Ministry of Education, Department of
Chemistry, Tsinghua University, Beijing, 100084, China.
E-mail: hedeh@mail.tsinghua.edu.cn; Fax: +8610 6277 3346;
Tel: +8610 6277 3346
† Electronic supplementary information (ESI) available: GC-MS anal-
ysis of the 18-carbon a,b-unsaturated aldehyde produced by the
dehydration of the aldol consendation adduct of nonyl aldehydes in the
1-octene hydroformylation catalyzed by S–C3–PPh2–[RhCl(CO)(PPh3)2]
(3rd catalyst cycle) in the present study. See DOI: 10.1039/b900591a
aspect OMS is a good support candidate. Immobilization of Rh
complexes to MCM-41, MCM-48 and SBA-15 has been previ-
ously reported.16–18 Although the immobilized catalysts are easily
recyclable, the catalytic application of the OMS-anchored Rh
complex encounters two main problems. First, the leaching of
active species is sometimes serious over successful catalyst cycles,
since the complex is not stably immobilized. A general solution is
to apply better ligands for stable coordination. Li et al.16 recently
1146 | Green Chem., 2009, 11, 1146–1154
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
reported a dendritic SBA-15 supported Wilkinson’s catalyst for
the styrene hydroformylation and Rh leaching was minimized.
The other significant problem is the decrease of catalyst activity
compared to homogeneous counterparts, which is related to
mass transfer. More effective ligands can be then employed19
to promote the heterogenized catalyst activities. This can be
achieved by corresponding homogeneous catalyst study ad-
vances. Peng et al.18 employed two triphenylphosphine analogs
as additional ligands to MCM-41 or MCM-48 anchored Rh–N
complex catalysts. The immobilized catalysts showed activity
and linear product selectivity comparable to corresponding
homogeneous counterparts in the 1-hexene hydroformylation.
To chemically bond the catalytic organometallic complex
to the inorganic support, some organic medium is required
as spacer. Hence, organic functionalization20,21 of the OMS
to produce a hybrid material is generally performed before
further possible modification. Previous researches have been
focusing on altering the ligand, either the ligand connected to
spacer or the ligand additionally added. In our recent work,22
RhCl3 as precursor was anchored to SBA-15 modified with
diphenylphosphinoalkyl (–(CH2)nPPh2, n = 1, 3, 5, 8, 11). The
catalysts were employed in the 1-octene hydroformylation and
it was found that the TON of 1-octene was increased with the
lengthening of the alkyl spacer. The catalyst prepared with the
longest alkyl spacer (C11) showed comparable activity to the ho-
mogeneous precursor RhCl3. The selectivity however was higher,
probably because the by-products were restricted in the pores.
Rh leaching in the reaction solution was found to be satisfyingly
low, owing to the stable complexity of the phenyl phosphine
ligand.23 However, these immobilized catalysts SBA-15–Cn–
PPh2–[RhCl3] showed lower activity than the homogeneous
RhCl(CO)(PPh3)2, probably owing to the less efficient precursor
that was used. It was found22,24 by choosing a more active Rh
precursor that the heterogenized catalyst activity was promoted.
In the present work, RhCl(CO)(PPh3)2, the famous catalyst for
hydroformylation developed by UCC, will be employed as Rh
precursor to prepare OMS-anchored Rh–P complex catalysts.
The OMS support involves the structurally similar SBA-15 and
MCM-41. 1-Octene hydroformylation to nonyl aldehydes will
be still applied as a probe reaction. We will show that the alkyl
spacer is a significant function in promoting OMS-anchored
Rh–P complex activities in the 1-octene hydroformylation. For
a comparison, RhCl(CO)(PPh3)2-supported catalysts (without
the diphenylphosphinoalkyl-modification of the support) will
be also prepared and applied.
Scheme 1 Preparation of Rh-immobilized catalysts.
and washed until the filtrate was neutral. The as-synthesized
SBA-15 was calcined at 773 K for 6 h with a temperature-ramp
of 1 K min-1.
MCM-41 was synthesized according to literature
methods.4,12,24 In a typical synthesis, ammonia (aq) and
cetyltrimethylammonium bromide (CTABr) were dissolved in
deionized water under vigorous stirring. TEOS was then added
dropwise. After stirring for 2 h, a gel with a molar composition
of 0.12 CTABr : 112 H2O : 8 NH4OH : 1.00 TEOS was obtained,
transferred into a Teflon-lined autoclave and then heated at
383 K for 72 h. The solid product was recovered by filtration
and washed with deionized water until the filtrate was neutral.
The as-synthesized material was calcined at 823 K for 6 h in air
with a temperature-programmed rate of 1 K min-1.
2.1.2 Surface modification: haloalkylation by silane. 1.00 g
OMS was suspended in toluene (A.R.) and 3.0 mL trimethoxysi-
lane X(CH2)nSi(OCH3)3 was added (silanes with alkyl spacer n =
1, 3, 11 were purchased from Gelest; silanes with spacer n = 5
and 8 were synthesized according to literature methods,22,25,26
X = Br in the case of n = 11, for the rest X = Cl). The above
mixture was refluxed for 24 h. The resulting solid was recovered
by filtration and washed with dichloromethane in a Soxhlet’s
extractor for 12 h to remove the organic species physisorbed.
Haloalkyl-modified OMSs are denoted as “S(or M)–Cn–X”.
S indicates SBA-15 and M indicates MCM-41. Alkyl carbon
number n = 1, 3, 5, 8, 11. X is the haloatom (Cl or Br).
2.1.3 Diphenylphosphine ligand connection. The connec-
tion of diphenyl phophine ligands (–PPh2) was carried out
under Schlenk operation conditions. THF (A.R.) was in advance
refluxed with sodium under nitrogen for dehydration and
deoxidation. 20.8 g triphenyl phosphine (Fluka) was dissolved in
200 mL THF. 1.1 g lithium turnings (A.R., Sinopharm Chemical
Reagent Co., Ltd) were then added. The mixture was stirred
at room temperature and in dry nitrogen atmosphere for 3 h.
After removing the surplus metal, 9.0 mL tert-butyl chloride
(C.P., Sinopharm Chemical Reagent Co., Ltd) was added at
273 K to decompose the phenyllithium. The electromagnetic
stirrer was held at 273 K for 1 h and then at room temperature
for 2 h. S(or M)–Cn–X was afterwards mixed with the red
2. Experimental
2.1 Preparation of the catalysts
2.1.1 Synthesis of SBA-15 and MCM-41 (OMS). Prepara-
tion of the OMS-immobilized Rh catalysts is shown in Scheme 1.
SBA-15 was synthesized in a three-necked flask.6,22 Tri-block
copolymer P123 (PEO20PPO70PEO20, Mn = 5800, Aldrich) was
dissolved in HCl (2 mol L-1). Tetraethyl orthosilicate (TEOS,
Acro¨s) was added dropwise. The mixture was stirred for 20 h
under 308 K. The flask was then sealed and kept under 353 K
for 24 h. Molar ratio of the initial gel was 0.02 P123 : 40
H2O : 4.78 HCl : 1.00 TEOS. The resulting solid was recovered
This journal is
The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 1146–1154 | 1147
©

View Article Online
lithium diphenylphosphide solution and the continuous stirring
was held for 16 h at room temperature. The resulting solid
was recovered and washed with dichloromethane in nitrogen
atmosphere for 12 h. Diphenylphosphinoalkyl-modified OMS
is denoted as S(or M)–Cn–PPh2.
for 2.5 h. At the end of the reaction, the autoclave was cooled
to lower than 278 K and the pressure was released. 1.0 mL
isopropyl alcohol (A.R.) was added as internal standard. The
resulting mixture was analyzed by a GC-HP4890 equipped with
a HP-5 capillary column (0.25 mm ¥ 30 m). In the catalyst
recycling experiments, the separated catalysts were above all
washed by toluene under ultrasonic washing12 to remove residual
substances.
2.1.4 Immobilization of RhCl(CO)(PPh3)2 complex. The
immobilization of RhCl(CO)(PPh3)2 (Strem) was finished by
mixing S(or M)–Cn–PPh2 in dichloromethane with the rhodium
precursor (10 wt%), which was stirred at room temperature and
in nitrogen atmosphere for 36 h.24,27,28 The final solid catalyst
was recovered and ultrasonic-washed to remove uncoordinated
ions.12 The catalyst is denoted as S(or M)–Cn–PPh2–Rh.
Seven OMS-anchored Rh–P complex catalysts were prepared,
including S–Cn–PPh2–Rh (n = 1, 3, 5, 8, 11) and M–Cn–PPh2–Rh
(n = 3, 5).
2.1.5 Preparation of RhCl(CO)(PPh3)2-supported catalysts.
For comparison, RhCl(CO)(PPh3)2-supported catalysts were
prepared by the impregnation method. Support OMS (SBA-15
or MCM-41) was mixed with RhCl(CO)(PPh3)2 (10 wt%) in
dichloromethane. The mixture was stirred at room tempera-
ture and in nitrogen atmosphere for 36 h. The catalyst was
recovered and washed with dichloromethane until the filtrate
was colorless. RhCl(CO)(PPh3)2-supported catalyst is denoted
as RhCl(CO)(PPh3)2/S or RhCl(CO)(PPh3)2/M.
3. Results and discussion
3.1 Catalyst characterization results
3.1.1 X-ray diffraction (XRD). Fig. 1 shows the XRD
patterns of catalysts S–Cn–PPh2–Rh ((A), n = 1, 3, 5, 8, 11)
and M–Cn–PPh2–Rh ((B), n = 3, 5). All the XRD patterns
revealed ordered hexagonal symmetry with a well-resolved (100)
reflection.1,5 The preparative procedure of the catalysts did not
ruin the support structure, revealing the good solvothermal
stabilities of SBA-15 and MCM-41. The (100) reflection is
associated with the unit cell parameter a0 of the hexagonal
symmetry by the equation,4
2
a0 =
d100
(1)
3
where a0 equals the pore size plus one wall thickness and d100
is the interplanar crystal spacing. By the Bragg’s equation it is
known,
2.2 Characterization of catalysts
Powder X-ray diffraction (XRD) data at low angles were col-
lected on a Rigaku D/max-RB diffractometer at a scan solution
of 2◦ min-1 with CuKa radiation (l = 0.1542 nm). XRD charac-
terization was powered at 40 kV and 100 mA. Fourier-transform
infrared spectrometry (FT-IR) in the 4000 cm-1 to 450 cm-1
range was operated on a Perkin SpectrumOne spectrometer by
KBr wafers (1 wt% of sample). Isothermal nitrogen sorption
analysis at 77 K was performed on a Micromeritics ASAP
2010 analysis system; the BET (Brunauer–Emmett–Teller) and
BJH (Barrett–Joyner–Halenda) methods were applied for the
specific surface area and pore size distribution calculations,
respectively. Rh contents of the immobilized and supported
catalysts were obtained by an IRIS Intrepid II Inductive
Coupling Plasma-Atomic Emission Spectroscopy (ICP-AES)
spectrometer. Leached Rh contents in reaction solutions were
analyzed by an X series In 20000CPS/ppb (1.5%RSD) ICP-MS
spectrometer. 2 mL reaction solution was calcined at 573 K to
evaporate all liquids. 0.5 mL H2SO4 (98%) was added to the
residue and heated to eliminate organic residues. Then, 1.0 mL
HNO3 (63%) was added. The resulting acidic liquid mixture was
then diluted to 25 mL for ICP-MS analysis. Determination of
the by-product was finished by a Shimadzu GCMS-QP2010S
system.
nl = 2d100 ¥ sinq
(2)
where l is the radiation wavelength. By eqn (1) and (2), the (100)
reflection located at lower 2Theta for larger pores. In Fig. 1,
SBA-15-anchored catalysts (A) gave the (100) reflection at ca.
2.3 1-Octene hydroformylation. 1-Octene hydroformyla-
tion was carried out as follows. 5.0 mL 1-octene (Acro¨s, 99%)
was mixed with 15.0 mL toluene (A.R.) in a 100 mL autoclave
and the catalyst was added. The autoclave was then firmly sealed
and syngas (equimolar H2 and CO) was injected at a temperature
lower than 278 K to replace the air inside the autoclave (3¥). The
autoclave was heated to 393 K and the syngas was injected to
5.0 MPa. The electromagnetic stirring was started and lasted
Fig. 1 XRD patterns of (A) S–Cn–PPh2–Rh and (B) M–Cn–PPh2–Rh.
1148 | Green Chem., 2009, 11, 1146–1154
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
2Theta = 1.00◦, and MCM-41-anchored catalysts (B) show the
reflection at ca. 2Theta = 2.30◦.
3.1.2 Infrared spectroscopy (IR). To clarify the immobi-
lized catalyst structure formed inside the OMS pores still
remains difficult. However, as it has been developed in our
experiments,22,24 IR spectroscopy has been shown to be powerful
for monitoring the multi-step-assembly of the complex into the
OMS if critical IR absorptions are carefully assigned.29 Since the
organic modified OMSs were washed during the preparation of
catalyst, the physisorption was beyond consideration. Two ex-
amples are presented in Fig. 2 with quite similar regularities and
conclusions. Please refer to Scheme 1 for sample correspondence.
When RhCl(CO)(PPh3)2 was anchored to SBA-15 through
spacer C3 (Fig. 2(A)), nas(C–H) at 2933 cm-1, ns(C–H) at
2878 cm-1, and d(C–H) at 1457, 1375, 1350 and 1296 cm-1
in P123-bearing SBA-15 (a) vanished after calcination (b),
indicating the successful removal of the polymer template.
These peaks reappeared after chloropropyl-modification (c) by
3-chloropropyltrimethoxyl silane. The absorption at 955 cm-1 (b)
was significantly weakened after chloropropylation (from (b) to
(c)), revealing that the amount of modifiable silanols30 decreased
owing to silane modification. The successful chloropropyl-
modification was also reflected in the observation of n(C–Si)
at 696 cm-1 and n(C–Cl) at 647 cm-1 (c). n(C–Si) was preserved
after connection with –PPh2 ligands (d) and the immobilization
of RhCl(CO)(PPh3)2 (e), indicating that the grafted silanes were
not lost. n(C–Cl) disappeared after the C–P bond formation (d),
showing n(C–P) at 748 cm-1. S–C3–PPh2 (d) showed an intense
and broad aromatic skeleton stretching at 1508 cm-1 owing to
the conjugation of phosphorous p elections with p electrons of
phenyls. Aromatic d(C–H) was observed at 866 cm-1. A weak
absorption at 3073 cm-1 arose due to the overtone contributed by
=
n(Ar C C) and n(Ar C–H) of single-substituted benzene.
=
n(Ar C C) was weakened after Rh complex immobilization
(from (d) to (e)), probably because the cation withdrew electrons
from the ligands. This also showed the interaction of the
–(CH2)3PPh2 ligand with Rh, implying the successful immo-
-1
=
bilization of the complex. n(C O) was observed at 1979 cm
(see insert). The stretching of the Si–O–Si bond31 in the saturated
range of 1290 to 996 cm-1 and d(Si–O–Si) at 467 cm-1 were
observed through the multi-step modification, indicating that
the SBA-15 framework was not affected, which is in accordance
with the XRD characterization.
Similar regularities were found if IR was performed
to monitor the immobilization of RhCl(CO)(PPh3)2 to
diphenylphosphinopentyl-modified MCM-41 (Fig. 2(B)).
CTABr-bearing MCM-41 (a) showed intense nas(C–H) at
2922 cm-1, ns(C–H) at 2852 cm-1 and long chain alkyl rocking
at neighboring 1488, 1481, 1473, and 1465 cm-1. n(N+–H) was
observed at 1397 cm-1; d(N+–H) was observed at 838, 720, and
697 cm-1 (a). These absorptions disappeared after calcination
with the terminal silanol stretching found at 3747 cm-1 ((b)
and insert). The signal characteristic of the modifiable degree
of MCM-41 showed at 966 cm-1 (b), and it was weakened
after silane grafting (from (b) to (c)). After the grafting of
silane (c), nas(C–H) at 2958 cm-1, ns(C–H) at 2851 cm-1,
n(C–Si) at 693 cm-1, and n(C–Cl) at 647 cm-1 were observed. The
connection of –PPh2 ligands (d) produced n(C–P) at 741 cm-1
Fig. 2 FT-IR spectra of (A): P123-bearing SBA-15 (a), calcined SBA-
15 (b), S–C3–Cl (c), S–C3–PPh2 (d), S–C3–Rh (e). Insert: an enlarged area
of the S–C3–Rh catalyst showing the carbonyl stretching. (B): CTABr-
bearing MCM-41 (a), calcined MCM-41 (b), M–C5–Cl (c), M–C5–PPh2
(d), M–C5–Rh. Insert: an enlarged area of calcined MCM-41 showing
the stretching of the terminal silanol.
-1
=
and a broad n(Ar C C) band at 1508 cm . The overtone shown
=
by n(Ar C C) and n(Ar C–H) of single-substituted benzene
was similarly found at ca. 3070 cm-1. n(Si–O–Si) stretching
appeared in the 1250 to 1000 cm-1 range and was preserved. The
corresponding IR absorption of organic species seems stronger
This journal is
The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 1146–1154 | 1149
©

View Article Online
Table 1 Textual properties and Rh contents of the catalysts
SBETa/m2 g-1 DBJHa/nm VPa/cm3 g-1 Rhb(wt%)
than that of the SBA-15-based materials, since MCM-41 has a
thinner SiO2 framework.
3.1.3 N2 sorption analysis. Fig. 3 shows the isotherms
of parent SBA-15 and S–Cn–PPh2–Rh catalysts (A) and the
corresponding pore size distribution of the samples (B). All
the isotherms revealed to be type IV according to the IUPAC
classification,32 characteristic of typical ordered mesoporous
materials. S–Cn–PPh2–Rh catalysts gave type H1 hysteresis
loop, revealing the uniform pore size distribution and highly
ordered mesostructure, which was consistent with the XRD
characterization. Some textual properties of the catalysts are
summarized in Table 1. The immobilization of Rh complex into
mesoporous SBA-15 decreased the BET specific surface area
of the material, probably because the inner surface had been
tailored with polar species and thus the adsorption of nitrogen
molecules was more difficult. With the increase of alkyl spacer
length, the hysteresis occurred at lower P/P0, indicating the
decrease of pore size.
SBA-15
673
440
375
355
385
429
755
693
642
—
6.2
4.8
3.9
3.8
3.6
3.5
2.75
2.52
2.28
0.75
0.53
0.35
0.38
0.39
0.45
0.52
0.43
0.38
—
0.22
0.35
0.32
0.29
0.28
S–C1–PPh2–Rh
S–C3–PPh2–Rh
S–C5–PPh2–Rh
S–C8–PPh2–Rh
S–C11–PPh2–Rh
MCM-41c
—
M–C3–PPh2–Rhc
M–C5–PPh2–Rh
RhCl(CO)(PPh3)2/S
RhCl(CO)(PPh3)2/M
1.50
1.20
0.38
0.50
—
—
—
—
—
a SBET: BET specific surface area; DBJH: BJH primary pore diameter;
VP: single point total pore volume. b Immobilized-Rh contents were
determined by ICP-AES analysis. c Data from ref. 24.
MCM-41 and M–Cn–PPh2–Rh catalysts (Fig. 4) also re-
vealed type IV isotherms with H1 hysteresis loop.24 However,
the hysteresis loops revealed by MCM-41-anchored catalysts
were smaller and the isotherms were more likely approaching
langmuir-type (Fig. 4(A)). This is probably because the pore size
Fig. 3 (A) Isotherms of parent SBA-15 and S–Cn–PPh2–Rh catalysts
and (B) DBJH pore size distribution of SBA-15 and S–Cn–PPh2–Rh
catalysts.
On the other hand, the assemble of organometallic species
also increased the bulky density of the materials. Noticeably,
although the DBJH somewhat decreased with the increase of
spacer length (Table 1), their differences were not consistent with
theoretical calculation with C–C bond length, probably owing
to the flexibility of the alkyl spacer.18
Fig. 4 (A) Isotherms of parent MCM-41 and M–Cn–PPh2–Rh catalysts
and (B) DBJH pore size distribution of MCM-41 and S–Cn–PPh2–Rh
catalysts. Data of M–C3–PPh2–Rh are from ref. 24.
1150 | Green Chem., 2009, 11, 1146–1154
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
of MCM-41-anchored catalysts are just a little bigger than 2 nm
(Table 1). Textual properties of MCM-41 and M–Cn–PPh2–Rh
are summarized in Table 1. M–Cn–PPh2–Rh catalysts had higher
SBET than S–Cn–PPh2–Rh catalysts because of the thinner SiO2
framework.
According to Fig. 5, 1-octene TON was increased with the
increasing length of alkyl spacer. Catalysts prepared with the
shortest spacer (C1, the red lines in (A) and (B)) revealed
the lowest 1-octene TON. Catalysts prepared with the longest
alkyl spacer (C11, the royal blue lines in (A) and (B)) revealed
an activity comparable to the homogeneous counterpart (the
purple dot lines in (A) and (B)). Obviously, the catalyst prepared
with RhCl(CO)(PPh3)2 as precursor (B) revealed a much higher
activity than the corresponding analogue prepared with RhCl3
as precursor (A), indicating that the catalyst activity could be
further promoted by choosing a more active precursor, similar
to the strategy of synthesizing more efficient ligands to promote
the immobilized catalyst activity.18 The high turnover capacity
revealed by the OMS-anchored Rh catalyst through the longer
alkyl as spacer might result from the flexibility of the alkyl.
A longer alkyl spacer brings the anchored catalyst far from
the solid surface and the motion of the anchored catalyst is
thus relatively free. According to (B) and (C), corresponding
S–Cn–PPh2–Rh and M–Cn–PPh2–Rh with the same alkyl spacer
(C3 and C5, respectively) revealed a similar 1-octene TON. This
indicates that the OMS-anchored Rh complex catalyst activity
depends greatly on the spacer length, as long as inner mass
transfer is eliminated.
In contrast, RhCl(CO)(PPh3)2-supported catalysts (black
dash–dot lines showing 1-octene TON in Fig. 5(B) and (C))
showed a very quick activity loss when the catalysts were
recycled. Both RhCl(CO)(PPh3)2/S and RhCl(CO)(PPh3)2/M
revealed a sharp activity decrease within only two catalyst
cycles. This is because without the diphenylphosphinoalkyl that
covalently bonds the complex and the support, the Rh species
leaches significantly during catalyst cycles.18 Noticeably, fresh
RhCl(CO)(PPh3)2-supported catalyst revealed an activity as
high as the homogeneous RhCl(CO)(PPh3)2-catalyzed process.
3.2 1-Octene hydroformylation results
3.2.1 Reaction operation conditions and specific activity defi-
nition. 1-Octene hydroformylation conditions were the same
as our previous work22,24 so that a comparison could be
made. Initial syngas pressure (CO/H2) was 5.0 MPa; reac-
tion temperature was 393 K, and reaction time was 2.5 h.
1-Octene hydroformylation was also performed by correspond-
ing homogeneous counterparts under the same reaction con-
ditions. Turnover number (TON) is adopted for disclosing the
specific catalyst activity and efficiency, which is defined as,
converted 1-octene
TON = molar
(3)
immobilized Rh
3.2.2 Catalyst turnover capacity. The results of the
1-octene hydroformylation catalyzed by S(or M)–Cn–PPh2–Rh
catalysts are shown in Fig. 5(B) and (C). Previous work22 is also
quoted for comparison (Fig. 5(A)). Fig. 5(A), (B) and (C) are
plotted with the same scale of TON (vertical axis on the left)
and selectivity of nonyl aldehydes (vertical axis on the right).
(A) and (B) show the results from Rh-immobilized catalysts with
the same support (SBA-15) but different Rh precursors, while
(B) and (C) show the results from the catalysts with the same
precursor (RhCl(CO)(PPh3)2) but different OMSs as support.
Two purple dot lines are respectively plotted in (A) and (B),
showing the results catalyzed by corresponding homogeneous
catalysts under the same reaction conditions.
Fig. 5 1-Octene hydroformylation results over the catalysts. (A) SBA-15–Cn–PPh2–[RhCl3], by using RhCl3 as precursor, data from ref. 22;
(B) SBA-15–Cn–PPh2–[RhCl(CO)(PPh3)2] viz. S–Cn–PPh2–Rh defined in the present work, by using RhCl(CO)(PPh3)2 as precursor; (C) M–Cn–
PPh2–Rh, by using MCM-41 as support and RhCl(CO)(PPh3)2 as precursor, M–C3–PPh2–Rh data from ref. 24. (ꢀ) 1-Octene TON, (᭺) selectivity
of nonyl aldehydes. A specific spacer is given a specific color, (red) n = 1, (orange) n = 3, (dark yellow) n = 5, (olive) n = 8, (royal blue)
n = 11. Homogeneous processes are in purple dot lines, RhCl(CO)(PPh3)2-supported catalysts results in black dash–dot lines. Reaction temperature
393 K, initial syngas pressure 5.0 MPa, reaction time 2.5 h. Initial Rh/1-octene ratio (¥10-4): 1.2 (S–C1–PPh2–Rh), 1.9 (S–C3–PPh2–Rh), 1.7 (S–C5–
PPh2–Rh), 1.6 (S–C8–PPh2–Rh), 1.5 (S–C11–PPh2–Rh), 2.6 (homogeneous RhCl(CO)(PPh3)2), 1.8 (RhCl(CO)(PPh3)2/S), 4.5 (M–C3–PPh2–Rh), 3.3
(M–C5–PPh2–Rh), 2.3 (RhCl(CO)(PPh3)2/M).
This journal is
The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 1146–1154 | 1151
©

View Article Online
Table 2 Ratio of linear and branched nonyl aldehydes shown by Rh-immobilized and RhCl(CO)(PPh3)2-supported catalysts along with catalyst
cycles
Ratio of linear/branched nonyl aldehydes (n/i)
Catalyst cycle
Catalyst
1
2
3
4
5
6
7
8
S–C1–PPh2–Rh
S–C3–PPh2–Rh
S–C5–PPh2–Rh
S–C8–PPh2–Rh
S–C11–PPh2–Rh
M–C3–PPh2–Rha
M–C5–PPh2–Rh
RhCl(CO)(PPh3)2/S
RhCl(CO)(PPh3)2/M
1.8
1.6
1.6
1.6
1.3
1.5
1.6
1.6
0.6
1.6
1.7
1.6
1.4
1.4
1.4
1.5
1.8
1.5
1.6
1.8
1.5
1.3
1.3
1.6
1.7
—
—
—
—
—
—
1.7
1.7
1.5
1.2
1.5
1.8
—
1.4
1.6
1.3
1.5
1.7
1.4
—
1.9
1.8
1.3
1.4
1.8
1.5
—
1.7
1.4
1.4
1.5
—
—
—
1.5
1.5
1.3
1.3
—
—
—
—
—
—
—
—
—
a Data from ref. 24.
Table 3 Rh leaching from S(or M)–Cn–PPh2–Rh catalystsa
This is because supported Rh species leached into the reaction
solution and thus a homogeneous process took place. After the
first cycle of RhCl(CO)(PPh3)2/M, 82% of the supported Rh
had been detected to have leached into the liquid.
Average Rh
leaching per cat.
cycle times leaching (%) cycle (%)
Catalyst
Total Rh
Catalyst
S–C1–PPh2–[RhCl3]b
S–C3–PPh2–[RhCl3]b
S–C5–PPh2–[RhCl3]b
S–C8–PPh2–[RhCl3]b
3
10
10
10
1.1
2.3
2.5
1.4
1.3
4.0
3.8
3.3
3.0
2.9
0.29
0.22
0.37
0.23
0.25
0.14
0.13
1.30
0.48
0.41
0.38
0.36
0.05
0.04
3.2.3 n/i ratio of nonyl aldehydes. Table 2 shows the ratio
of linear and branched products of nonyl aldehydes along with
catalyst cycles. The n/i ratio varied in the range of 1.0 to
2.0 along with the catalyst cycles, which seemed lower since
hydroformylation is accompanied by isomerization.33 It appears
that catalysts prepared with longer alkyl spacers revealed a lower
n/i ratio. These catalysts prepared with RhCl(CO)(PPh3)2 as
precursor showed higher selectivity towards linear products than
the catalyst analogs22 prepared with RhCl3 as precursor, as a
result of abundant triphenyl phosphine ligands.
S–C11–PPh2–[RhCl3]b 10
S–C1–PPh2–Rh
S–C3–PPh2–Rh
S–C5–PPh2–Rh
S–C8–PPh2–Rh
S–C11–PPh2–Rh
M–C3–PPh2–Rhc
M–C5–PPh2–Rh
3
8
8
8
8
6
6
a Determined by ICP-MS. b Prepared with RhCl3 as precursor, data from
ref. 22. c Data from ref. 24, S = SBA-15, M = MCM-41. Rh = using
RhCl(CO)(PPh3)2 as precursor.
3.2.4 Catalyst stability and Rh leaching. In Fig. 5, the
OMS-anchored Rh catalysts could be employed along with
several cycles in 1-octene hydroformylation, but the catalysts
revealed quite different stabilities. The catalysts prepared with
methyl (C1) as spacer (red lines in Fig. 5(A) and (B)) revealed
an obvious activity decrease within three catalyst cycles. On
the contrary, the catalysts prepared with longer spacers show
relative stability. The catalysts prepared with the longest alkyl
spacer (C11, royal blue lines in (A) and (B)) could be successfully
employed along with several cycles with no obvious activity
decrease observed. With the increasing length of alkyl spacer,
the stability of the prepared catalysts was increased as the slope
of the plotted TON line becomes lower (Fig. 5(A) and (B)).
The stability of M–Cn–PPh2–Rh catalysts (Fig. 5(C)) will be
discussed later.
In order to investigate whether Rh leaching is the reason for
the catalyst activity decrease, reaction solutions were analyzed
for leached Rh content. Table 3 summarizes the total Rh
leaching percentage from the sum of catalyst cycles for every
immobilized catalyst. According to the data presented, the Rh
leaching is satisfyingly low, comparing with related work.34
MCM-41-based catalysts revealed even lower active species
leaching, probably resulting from the higher specific area (see
Table 1). Both the good coordination capacity of the phosphine
ligand23 and the confining effect originating from the larger
surface area of OMS-anchored Rh–P catalyst contribute to the
lower Rh leaching.
Additional work was carried out by pre-treating the fresh
S–C3–PPh2–Rh catalyst (which showed a moderate catalyst
activity decrease along with catalyst cycles among the analogues)
with toluene at 393 K and 5.0 MPa syngas pressure for 2.5 h
(exactly the hydroformylation conditions). Following the hydro-
formylation operations in section 2.3 and reaction conditions
described in section 3.2.1, the catalyst was stirred in fresh
toluene. The mother liquid was collected, to which 1-octene was
added. The mixed liquids were allowed to carry out the hydro-
formylation under the same reaction conditions, and the result
was a 1-octene TON = 53. The treated catalyst was again treated
with toluene and the collected solvent was mixed with 1-octene.
Hydroformylation of the mixed liquids then resulted in a 1-
octene TON of 8. The two-time-treated catalyst S–C3–PPh2–Rh
was employed in 1-octene hydroformylation, and the substrate
TON was 566, with a 97% activity compared to the fresh catalyst
without pre-treating with toluene. By pre-treating the catalyst,
the active species immobilized to the external surface18 of
SBA-15 were most likely to be removed. Although washing
assisted by an ultra-sonic device can remove the uncoordinated
1152 | Green Chem., 2009, 11, 1146–1154
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
species when synthesizing the catalyst,12 it appears that some
active species will only leach under the reaction conditions.
Both the determination of the leached Rh in the reaction
solution and the pre-treating fresh catalyst experiment indicate
the good stability of the Rh–P complex. Thus, Rh leaching won’t
account for the catalyst activity decrease along with catalyst
cycles observed in Fig. 5(A) and (B).
3.2.5 Surface properties of recycled catalysts. Infrared
spectroscopic analysis has been performed on recycled catalysts,
as is shown in Fig. 6. The IR spectra of recycled catalysts were
quite different from those of the fresh catalysts. The stretching
Scheme 2 The dehydrated product of the aldol condensation adduct
coordinates with Rh so as to poison it. (a) Longer spacers on larger
pores with lower surface curvature; (b) shorter spacers on larger pores;
(c) shorter spacers on smaller pores with higher curvature.
=
vibration of saturated C–H, aromatic C C stretching, C–P
as well as C–Si stretchings were greatly enhanced. Aromatic
C–H stretching can be observed as the shoulder peak of alkyl
C–H stretching. The rocking of the long chain –CH2–, as a
fingerprint of polymeric alkane,35 was clearly observed as a
series of neighboring peaks in the range of 1500 to 1300 cm-1.
A new peak was observed at ca. 1700 cm-1, which could be
22,29,36
=
assigned to the n(C O) in the a,b-unsaturated species.
The stretching vibrations were significantly enhanced, which
implies that a compound with conjunction structure exists in
the recycled catalysts.
Fig. 7 BJH pore size of S–Cn–PPh2–Rh catalysts (a) fresh and (b)
recycled.
BJH pore size of the recycled catalysts was slightly decreased
compared to the corresponding fresh catalyst, which evidences
that the unsaturated aldehyde exists inside the pores.
The stably coordinated a,b-unsaturated aldehyde was difficult
to release during the catalytic process, and thus the catalysts
were irreversibly poisoned. The poisoning of catalysts in hydro-
formylations has been also found in homogeneous processes.37,38
When a longer spacer is employed, the pore size of the catalyst
is smaller. The terminal active sites (Scheme 2(a)) can stretch
into a smaller central circle, which disfavors the linear aldehyde
coordination owing to the steric hindrance. Catalysts with the
spacer C11 (royal blue lines in (A) and (B), Fig. 5) showed no
obvious activity decrease during the catalyst recycles. However,
when a shorter spacer was employed (Scheme 2(b)), the terminal
active sites stretched into a larger circle, not far from the support
surface. The immobilized Rh–P complex maintains a relatively
longer distance with each other, which favors the by-product
coordination and such complex is inert towards hydroformyla-
tion. When MCM-41 was employed as support, the catalysts
had smaller pores than the SBA-15-anchored Rh–P complex.
Smaller pores had higher surface curvature, which allowed
shorter spacers to stretch into smaller circles and to approach
each other near (Scheme 2(c)), providing similar resistance to
the by-product coordination. This might be the reason why the
M–Cn–PPh2–Rh (n = 3, 5) catalysts showed a relatively more
stable capacity of substrate turnover than the corresponding
S–Cn–PPh2–Rh (n = 3, 5) analogues (Fig. 5, compare (B)
Fig. 6 FT-IR spectra of recycled S–Cn–PPh2–Rh catalysts.
The aldol condensation of the aldehyde product has been
regarded as the side reaction of olefin hydroformylation.15 Owing
to the stability by conjunction, the aldol condensation adduct of
the aldehyde will dehydrate, forming an acyclic a,b-unsaturated
compound. On the basis of the literature15 and the IR spectra,
we assume that an 18-carbon a,b-unsaturated aldehyde, which
is the dehydrated product of the aldol condensation adduct
of nonyl aldehyde, was produced and coordinated with Rh
(Scheme 2) owing to the confining effect and high surface
area of porous materials. This product has been determined
by the GC-MS analysis of the reaction solution (see ESI).†
The phosphorous p electrons, phenyl p electrons, p electrons
=
=
from C C and C O in the a,b-unsaturated aldehyde, and
the rhodium d electrons together lead to a strong conjunction
and thus stabilizes the coordination of the a,b-unsaturated
aldehyde with Rh. The conjunction enhanced all the stretching
IR vibrations as mentioned in Fig. 6. Shown in Fig. 7, the
This journal is
The Royal Society of Chemistry 2009
Green Chem., 2009, 11, 1146–1154 | 1153
©

View Article Online
and (C)), as no obvious activity decrease was observed when
M–Cn–PPh2–Rh catalyst was recycled (Fig. 5(C)).
On the other hand, the by-product was confined in the
pores of the catalysts so it was in lower concentration in the
reaction solution. Consequently, the solid catalysts showed
higher selectivity than homogeneous processes (Fig. 5(A)
and (B)).
5 D. Zhao, J. Feng, Q. Huo, N. Melosh, G. H. Fredrickson, B. F.
Chmelka and G. D. Stucky, Science, 1998, 279, 548.
6 D. Zhao, Q. Huo, J. Feng, B. F. Chmelka and G. D. Stucky, J. Am.
Chem. Soc., 1998, 120, 6024.
7 S. Tao, G. Li and H. Zhu, J. Mater. Chem., 2006, 16, 4521.
8 A. Taguchi and F. Schu¨th, Microporous Mesoporous Mater., 2005,
77, 1.
9 D. E. De Vos, M. Dams, B. F. Sels and P. A. Jacobs, Chem. Rev.,
2002, 102, 3615.
10 A. Corma and H. Garcia, Adv. Synth. Catal., 2006, 348, 1391.
11 X.-Y. Hao, W. Zhou, J.-W. Wang, Y.-Q. Zhang and S. Liu, Chem.
Lett., 2005, 34, 1000.
4. Conclusions
12 X.-Y. Hao, Y.-Q. Zhang, J.-W. Wang, W. Zhou, C. Zhang and S. Liu,
Microporous Mesoporous Mater., 2005, 88, 38.
13 K. Ariga, A. Vinu, Q. Ji, O. Ohmori, J. P. Hill, S. Acharya, J. Koike and
S. Shiratori, Angew. Chem., Int. Ed., 2008, 47, 7254 and literatures
therein cited.
14 L. Mercier and T. J. Pinnavaia, Adv. Mater., 1997, 9, 500.
15 G. van Koten and P. W. N. M. van Leeuwen, in Catalysis: an
Integrated Approach, ed. R. A. van Santen, P. W. N. M. van Leeuwen,
J. A. Moulijn and B. A. Averill, Elsevier, Amsterdam, 2nd edn, 1999,
ch. 6.
16 P. Li and S. Kawi, Catal. Today, 2008, 231, 61.
17 L. Huang, Y. He and S. Kawi, J. Mol. Catal. A: Chem., 2004, 213,
241.
The present work is a continuing study22 on alkyl spacer effects
of ordered mesoporous silica-anchored Rh–P complex catalysts
in olefin hydroformylation. In conclusion, the present work
introduces a facile method to promote ordered mesoporous
silica-anchored Rh–P complex catalyst activities in olefin hy-
droformylation by simply lengthening the alkyl spacers that
covalently connect the support and the anchored complex and
choosing a more efficient Rh precursor. The length of the
alkyl spacer has a drastic effect on the specific activity (shown
as turnover number of substrate, TON) of the immobilized
catalysts. The catalyst prepared with the longest alkyl spacer
(C11) revealed a specific activity comparable to the homogeneous
counterpart. The immobilized catalysts reveal a remarkable
low Rh leaching along with successful catalyst cycles. The
dehydrated product of the aldol condensation adduct of nonyl
aldehydes might be the reason for the catalyst activity decrease.
A longer spacer on SBA-15 helps the Rh-immobilized catalysts
to resist the by-product poisoning coordination owing to steric
hindrance. Similarly, when MCM-41 is employed as support, it
has a higher surface curvature; shorter spacers can provide the
steric hindrance function to resist by-product coordination.
18 Q. Peng, Y. Yang and Y. Yuan, J. Mol. Catal. A: Chem., 2004, 219,
175.
19 A. J. Sandee, L. A. van der Veen, J. N. H. Reek, P. C. J. Kamer,
M. Lutz, A. L. Spek and P. W. N. M. van Leeuwen, Angew. Chem.,
Int. Ed., 1999, 38, 3231.
20 C. Zhang, W. Zhou and S. Liu, J. Phys. Chem. B, 2005, 109, 24319.
21 F. Hoffmann, M. Cornelius, J. Morell and M. Fro¨ba, Angew. Chem.,
Int. Ed., 2006, 45, 3216.
22 W. Zhou and D. He, Chem. Commun., 2008, 5839.
23 M. T. Honaker, B. J. Sandefur, J. L. Hargett, A. L. McDaniel and
R. N. Salvatore, Tetrahedron Lett., 2003, 44, 8373.
24 W. Zhou and D. He, Catal. Lett., 2009, 127, 437.
25 C. S. Marvel and W. E. Garrison, Jr, J. Am. Chem. Soc., 1959, 81,
4737.
26 R. A. Benkeser, S. D. Smith and J. L. Noe, J. Org. Chem., 1968, 33,
597.
Acknowledgements
27 S.-G. Shyu, S.-W. Cheng and D.-L. Tzou, Chem. Commun., 1999,
2337.
This work is financially supported by the National Natural
Science Foundation of China (No. 20673064), Specialized
Research Fund for the Doctoral Program of Higher Education
of Ministry of Education of China (2007003108) and Fund for
Analysis of Tsinghua University. Wei Zhou would like to thank
Mr Qingwei Zheng for his help in GC-MS analysis of the by-
product.
28 J. Kramer, E. No¨llen, W. Buijs, W. L. Driessen and J. Reedijk, React.
Funct. Polym., 2003, 57, 1.
29 J. Dean, in Analytical Chemistry Handbook, McGraw-Hill Book Co.,
Singapore, 1995, Section 6.
30 B. Tian, X. Liu, C. Yu, F. Gao, W. Luo, S. Xie, B. Tu and D. Zhao,
Chem. Commun., 2002, 1186.
31 L. Zhao, S. Wang, Y. Wu, Q. Hou, Y. Wang and S. Jiang, J. Phys.
Chem. C, 2007, 111, 18387.
32 K. S. W. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierotti,
J. Rouque´rol and T. Siemieniewska, Pure Appl. Chem., 1985, 57, 603.
33 A. C. J. Koeken, M. C. A. van Vliet, L. J. P. van den Broeke, B.-J.
Deelman and J. T. F. Keurentjes, Adv. Synth. Catal., 2008, 350, 179.
34 B. Tan, J. Jiang, Y. Wang, L. Wei, D. Chen and Z. Lin, Appl.
Organomet. Chem., 2008, 22, 620.
Notes and references
1 C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli and J. S.
Beck, Nature, 1992, 359, 710.
35 N.-C. Chia and R. Mendelsob, J. Phys. Chem., 1992, 96, 10534.
36 IR spectrum of acrolein is available from Aldrich website,
see http://www.sigmaaldrich.com/catalog/search/ProductDetail/
ALDRICH/110221.
37 G. Liu and M. Gardland, Organometallics, 1999, 18, 3459.
38 Y. Liu, D. He, C. Li, T. Wang, J. Liu and Q. Zhu, Chin. J. Mol. Catal.,
2000, 14, 337.
2 G. J. de, A. A. Soler-Illia, C. Sanchez, B. Lebeau and J. Patarin, Chem.
Rev., 2002, 102, 4093.
3 Y. Wan and D. Zhao, Chem. Rev., 2007, 107, 2821.
4 J. S. Beck, J. C. Vartuli, W. J. Roth, M. E. Leonowicz, C. T. Kresge,
K. T. Schmitt, C. T. W. Chu, D. H. Olson, E. W. Sheppard, S. B.
McCullen, J. B. Higgins and J. L. Schlenker, J. Am. Chem. Soc.,
1992, 114, 10834.
1154 | Green Chem., 2009, 11, 1146–1154
This journal is
The Royal Society of Chemistry 2009
©