Chiral catalysts confined in porous hosts 2. Catalysis

Article abstract of DOI:10.1016/S0021-9517(02)00155-0

The catalytic activities of the dirhodium carboxamide catalysts immobilized in the pores of MCM-41 and on the surface of Aerosil 200 were studied. The catalysts were tested in the Si-H insertion reaction of dimethylphenylsilane with methyl phenyl diazoace

Full text of DOI:10.1016/S0021-9517(02)00155-0

Journal of Catalysis 217 (2003) 275–283
www.elsevier.com/locate/jcat
Chiral catalysts confined in porous hosts
2
. Catalysis
a,b
a,b
b
a
Heidi M. Hultman, Martine de Lang, Isabel W.C.E. Arends, Ulf Hanefeld,
b
a,∗
Roger A. Sheldon, and Thomas Maschmeyer
a
Applied Organic Chemistry and Catalysis, Delft University of Technology, Julianalaan 136, 2628 BL, Delft, The Netherlands
Biocatalysis and Organic Chemistry, Delft University of Technology, Julianalaan 136, 2628 BL, Delft, The Netherlands
b
Received 11 July 2002; revised 8 November 2002; accepted 12 November 2002
Abstract
The catalytic activities of the dirhodium carboxamide catalysts immobilised in the pores of MCM-41 and on the surface of Aerosil 200
see preceeding paper) have been investigated. The catalysts were tested in the Si–H insertion reaction of dimethylphenylsilane with methyl
(
phenyl diazoacetate and in the cyclopropanation reaction of styrene with ethyl and tert-butyl diazoacetate. Significant improvements in
enantioselectivity (Si–H insertion) and in regioselectivity (cyclopropanation) were induced due to the spatial confinement by the carrier.

2003 Elsevier Science (USA). All rights reserved.
Keywords: Cyclopropanation; Si–H insertion; Dirhodium; Immobilisation; Catalysis
1
. Introduction
and framework of the chiral catalysts has a strong influence
on the orientation of the approaching alkene relative to the
carbenic centre. Bulkier ester groups give higher trans/cis
ratios and better enantioselectivities; e.g., tert-butyl diazo-
acetate produces a significant enhancement compared to
ethyl diazoacetate in the cyclopropanation of styrene catal-
ysed by Rh2(5S-MEPY)4. This effect was observed not only
for the chiral dirhodium catalysts, but also for the homoge-
neous Pfaltz and Aratani catalysts [5,6]. Due to confinement
effects the immobilised dirhodium catalysts described here
might be expected to display this effect in an even more pro-
nounced fashion.
Cyclopropanes are widely used as starting compounds
and intermediates in organic synthesis. Several natural and
synthetic cyclopropanes display interesting activities [7].
They are essential building blocks for pyrethroid insec-
ticides [8,9] and also in pharmaceuticals a wide variety
of active compounds containing cyclopropane rings are
known [10–12], e.g., the antidepressant milnacipran [13–15].
In addition to their application in agrochemicals and medic-
inal chemistry, cyclopropanes are versatile intermediates in
organic synthesis. Due to their ring-strain they are readily
converted into a large range of interesting compounds [16].
Thus, the enantio- and diastereoselective synthesis of cyclo-
propanes has attracted considerable attention [17].
Chiral dirhodium carboxamide complexes have attracted
considerable interest as enantioselective catalysts for, among
other reactions, cyclopropanations and Si–H insertions [1].
In order not only to combine the advantages of homogeneous
and heterogeneous catalysts in general, but also to improve
selectivity (via synergistic interactions [2] between the cat-
alytic centre and its porous surroundings), we immobilised
them on silica (Aerosil 200) surfaces and inside the pores of
siliceous MCM-41 [3]. Here, the catalytic behaviour of these
immobilised catalysts is presented and a comparison is made
with their homogeneous counterparts.
As model reactions the Si–H insertion of methyl phenyl-
diazoacetate (1) with dimethylphenylsilane (2) yielding
methyl 2-(dimethylphenylsilyl)phenylacetate (3) (Fig. 1, re-
action (1)) as well as the cyclopropanation of styrene (4)
with ethyl diazoacetate (5a) or tert-butyl diazoacetate (5b)
yielding cis and trans ethyl or tert-butyl 2-phenylcyclopro-
pane carboxylate (6a/b) (Fig. 1, reaction (2)) were investi-
gated. Earlier studies [4] showed that in cyclopropanation
reactions the interaction of the ester alkyl group with the lig-
*
Corresponding author.
E-mail address: t.maschmeyer@tnw.tudelft.nl (T. Maschmeyer).
0021-9517/03/$ – see front matter  2003 Elsevier Science (USA). All rights reserved.
doi:10.1016/S0021-9517(02)00155-0

276
H.M. Hultman et al. / Journal of Catalysis 217 (2003) 275–283
Fig. 1. Catalytic test reactions.
Although α-silyl carbonyl compounds are widely recog-
gether so that they react in an intramolecular fashion while
the noncatalytic reaction is intermolecular. The selectivities
are, therefore, much higher than in the normal intermolecu-
lar reactions. Rh2(MEPY)4 and Rh2(BNOX)4 are normally
not used as catalysts for carbene insertions into the Si–H
bonds of organosilanes, because they are considered to be
less selective, making this a good test case for the immo-
bilised catalysts.
From previous work [22,23], it might be expected that
spatial constraints induced by the carrier, and especially by
the pores of MCM-41, should increase the influence of the
chiral ligands. These studies demonstrated that enantiose-
lective conversions catalysed by a palladium complex im-
mobilised inside the pores of MCM-41 can give a signifi-
cant increase in ee compared to the homogeneous palladium
complex (ee’s increased from 43 to 96% [23] and from 6 to
17% [22]).
nised for their utility in organic synthesis [18], there are few
general methods for their preparation. Since silanes are ex-
cellent scavengers for free carbenes [19], carbene insertion
into the Si–H bond of organosilanes is an attractive method
for the synthesis of these α-silyl carbonyl compounds [18].
Chiral dirhodium catalysts have proven to be effective cata-
lysts for this transformation. The enantiomerically enriched
products are versatile building blocks for the synthesis of a
variety of natural products.
Dirhodium tetrakis-methyl-2-pyrrolidone-5-carboxylate
Rh2(MEPY)4) and dirhodium tetrakis-4-benzyl-2-oxazoli-
(
dinone (Rh2(BNOX)4) (Fig. 2) were used as model catalysts.
Rh2(MEPY)4 is a very selective catalyst in intramolecular
cyclopropanation reactions [20,21], but is not as selective in
intermolecular cyclopropanation. With the immobilisation it
was aimed to bring the reagents and the catalyst in such close
proximity to each other that an intramolecular reaction anal-
ogous to the situation in enzyme catalysis would be mim-
icked: in enzyme catalysis, reagents are brought close to-
However, this need not always be the case, as can be
seen from the following example. Copper bis(oxazoline)
catalysts with an alkyl tether were immobilised inside
Fig. 2. Homogeneous catalysts and heterogeneous catalysts.

H.M. Hultman et al. / Journal of Catalysis 217 (2003) 275–283
277
MCM-41 and MCM-48 [24]. In this way highly active
catalysts for the cyclopropanation of styrene with ethyl
diazoacetate were produced; however, they did not show
improved selectivities.
with dimethoxydimethylsilane modified the polarity of the
carrier surface and reduced the pore size [3].
Two previous examples of the immobilisation of the
Doyle catalyst are known. In one case [25] oligomer-bound
MEPY was prepared from polyethylene oligomers that were
esterified with 2-pyrrolidone-5(S)-carboxylic acid. By li-
gand exchange the immobilised catalyst was formed. It dis-
played behaviour similar to the homogeneous catalysts, but
the enantioselectivity and activity for some applications did
decrease. More recently, NovaSyn Tentagel and the Merri-
field resin were used by the same group for immobilisation
of this catalyst [26]. In this case similar or higher selectivi-
ties were obtained.
In the second case [27] Rh2(5S-MEPY)4 was immo-
bilised on graphite. The electrochemical behaviour of this
catalyst was tested in the presence and absence of DNA, but
no catalytic reactions were performed. In both cases no ob-
vious steric constraints were introduced.
The dirhodium carboxamide catalysts in our study were
immobilised by the use of direct covalent anchoring of the
catalyst to the carrier surface [3]. Immobilisation took place
via ligand exchange of surface-anchored carboxylic acid
groups with approximately one ligand per homogeneous chi-
ral catalyst. The steric constraint induced by immobilisation
might be expected to partially compensate for the loss of the
ligand, which has been replaced by a stabilising carboxylate
group.
Next to the steric constraint exercised by the porous car-
riers, the method of attachment can influence the selectivi-
ty and activity of the immobilised catalytic species. To in-
vestigate whether there is a significant difference between
flexible linkers or more rigid ones, three different nonchi-
ral carboxylate spacer groups were used (–(CH2)2COOH,
2
. Experimental section
All reactions and manipulations were performed under
an atmosphere of dry nitrogen using standard Schlenk-
type techniques. Dry solvents were purchased from Aldrich
and used without further purification. Styrene was obtained
from Aldrich and was vacuum-distilled before use. Methyl
phenyldiazoacetate [28] and t-butyl diazoacetate [29] were
prepared according to literature procedures. All other rea-
gents were purchased from Aldrich, Acros, or Baker and
used without further purification. For column chromatog-
raphy silica gel 60 (particle size 0.063–0.2 mm, Fluka)
was used. NMR spectra were obtained on a Varian Inova
3
00 MHz or a Varian VXR 400s spectrometer, relative to
TMS. Trans/cis ratios for 6a/b were determined using GC
analysis on a Varian Star 3400 CX GC with a CP wax 52
CB column (l = 50 m, od = 0.70 mm, df = 2.0 µm) and
on-column injection (retention times: 6a: 41.5 min (cis),
4
2.6 min (trans), 6b: 43.6 min (cis), 44.3 min (trans). Chi-
ral GC analysis for cis and trans 6a/b was performed us-
ing a B-DA [30] (1 = 40 m, Ø = 0.25 mm, split injec-
◦
tion) (methylesters of 6a, 110 C, 129.4 (1R, 2S), 133.0
(
(
1S, 2R) min (cis), 157.6 (1S, 2S), 159.1 (1R, 2R) min
trans)) or B-PH (1 = 40 m, Ø = 0.25 mm, split injec-
◦
tion) (6b, 100 C, 85.4, 87.7 min (cis), 119.7, 122.6 min
(
trans)) column. HPLC analysis for Si–H insertions was per-
formed using a literature procedure [31]. Rhodium elemental
analysis was performed using ICP-OES on a Perkin–Elmer
Optima 4300DV after the solid samples were dissolved in
1
% v/v HF and 1.3% v/v H2SO4 in water. The Rhodium
leaching was determined by analysing the reaction filtrates
with graphite AAS on a Perkin–Elmer 4100ZL. All yields
are isolated yields.
–
(CH2)3COOH, and –p-C6H4COOH groups, Fig. 3). These
tethers were chosen to generate defined distances between
the catalysts and the carrier in order to prevent blocking
of the active site by the surface. The –(CH2)2COOH and
–
(CH2)3COOH spacer groups are flexible and the catalysts
2.1. Typical Si–H insertion procedure
can, therefore, have interactions with the surface of the car-
rier. The different types of surfaces (with protected or unpro-
tected surface silanol groups) [3] could, therefore, be probed
as to whether or not they have an influence on the per-
formance of the catalysts. With the –p-C6H4COOH spacer
group these through-space type interactions should be re-
duced significantly. Protecting the surface silanol groups
Dichloromethane, 1 ml, was added to the catalyst (SiO –
2
−6
C H COO–Rh (4R-BNOX) , 0.176 g, 8.41×10 mol Rh)
6
4
2
3
and the mixture was stirred. Subsequently methyl phenyldia-
zoacetate (0.188 g, 1.07 mmol) in 0.5 ml dichloromethane
and dimethylphenylsilane (0.196 g, 1.45 mmol) in 1 ml
dichloromethane were added. The resulting mixture was re-
fluxed overnight. After evaporation of the solvent in vacuo,
the reaction mixture was chromatographed on silica gel us-
ing 19/1 light petroleum/ethyl acetate. Yield: 67%, ee: 28%
(
1
R-major product). 13C NMR (300 MHz, CDCl3, δ (ppm)):
73.1 (C=O), 135.9, 135.5, 134.0 (2 C’s), 129.6, 128.3 (2
C’s), 128.0 (2 C’s), 127.7 (2 C’s), 125.7 (aromatic C’s), 51.2
(
COOCH3), 46.0 (C–SiPhMe2), −4.1 (Si–CH3), −4.5 (Si–
1
CH3). H NMR (300 MHz, CDCl3, δ (ppm)): 7.4–7.1 (m, 10
H, Ar–H ), 3.602 (s, 1 H, C–H ), 3.532, (s, 3 H, COOCH3),
Fig. 3. Spacer groups.

278
H.M. Hultman et al. / Journal of Catalysis 217 (2003) 275–283
0
(
.348 (s, 3 H, Si–CH3), 0.319 (s, 3 H, Si–CH3). 29Si NMR
300 MHz, CDCl3, δ (ppm)): 0.157 (R2CHSi).
60.7 (COOCH2CH3), 26.2 (Ph–CHCH2CH), 24.2 (Ph–
CHCH2CH), 17.0 (Ph–CHCH2CH), 14.3 (COOCH2CH3).
1
H NMR (300 MHz, CDCl3, δ (ppm)): 7.3–7.0 (m, 5H, Ar–
3
2
.2. Reuse following typical Si–H insertion procedure
H ), 4.16 (q, J = 7.2 Hz, 2 H, COOCH2CH3), 2.5 (ddd, 1
3
3
H, JH1–H2 = 4.2 Hz, JH1–H3 = 6.6, 9.2 Hz CH2CH COOEt),
3
3
The Si–H insertion reaction was performed following
1.90 (ddd, 1 H, JH2–H1 = 4.2 Hz, JH2–H3 = 5.3, 8.4 Hz,
3 3
the typical procedure described above, using SiO2–(CH2)2
PhCH CH2), 1.60 (ddd, 1 H, JH2–H3 = 9.2 Hz, JH3–H2 =
−6
3
COO–Rh2(4S-BNOX)3 (0.324 g, 1.78 × 10
mol Rh),
5.3 Hz, JH3–H3 = 4.7 Hz, PhCHCH2CH), 1.35 (m, 1 H,
3
methyl phenyldiazoacetate (0.242 g, 1.38 mmol), dimethyl-
phenylsilane (0.205 g, 1.52 mmol, and 1,2-dichlorobenzene
PhCHCH2CH), 1.27 (t, J = 7.2 Hz, 3 H, COOCH2CH3).
(
0.2229 g) as the internal standard). After refluxing over-
2.3.2. Cis 6a
1
3
night, the catalyst was allowed to settle. The solution was
transferred to another vial through a syringe filter to re-
move traces of immobilised catalyst. In this solution 2.9 ×
C NMR (300 MHz, CDCl3, δ (ppm)): 171.0 (C=O),
136.6, 129.3 (2 C’s), 127.9 (2 C’s), 126.6, (aromatic C’s),
60.2 (COOCH2CH3), 25.5 (Ph–CHCH2CH), 21.8 (Ph–
CHCH2CH), 14.0 (Ph–CHCH2CH), 11.1 (COOCH2CH3).
−7
1
0
mol rhodium was present (determined by AAS). The
1
ee of the product and the conversion were determined
by chiral HPLC. The remaining solid was washed with
dichloromethane and dried. It was then used again follow-
ing the same procedure. After three cycles, the rhodium con-
tent of the catalyst was determined by ICP-OES. After the
third cycle, the catalyst was again separated from the so-
lution and to this solution 1,2-dichlorobenzene (0.181 g),
dimethylphenylsilane (0.166 g, 1.23 mmol), and methyl
phenyldiazoacetate (0.199 g, 1.13 mmol) were added. Sam-
ples were taken after 4 min and 20 h and were analysed by
chiral HPLC to determine the activity of the filtrate. In the
first cycle 79% conversion (after 20 h) was reached, while
after the third cycle only 19% conversion (after 20 h) was
observed. No catalytic activity was observed in the liquid
phase that was removed after the third cycle. The ee’s were
approximately 35% (the S product was the major product)
in the first two cycles. In the third cycle the ee decreased
to 11%. The rhodium content before reaction was 0.0055
mmol/g, while after three cycles no rhodium could be de-
tected anymore on the carrier.
H NMR (300 MHz, CDCl3, δ (ppm)): 7.3–7.1 (m, 5 H, Ar–
3
H ), 3.8 (q, J = 7.2 Hz, 2 H, COOCH2CH3), 2.6 (ddd, 1 H,
3
3
JH1–H2 = 7.7 Hz, JH1–H3 = 9.0, 7.7 Hz CH2CH COOEt),
3
3
2.10 (ddd, 1 H, JH2–H1 = 7.7 Hz, JH2–H3 = 5.5, 9.2 Hz,
3
3
PhCH CH2), 1.70 (ddd, 1 H, JH3–H1 = 7.7 Hz, JH3–H2 =
3
5.5 Hz, JH3–H3 = 5.1 Hz, PhCHCH2CH), 1.3 (m, 1 H,
3
PhCHCH2CH), 0.95 (t, J = 7.2 Hz, 3 H, COOCH2CH3).
If tert-butyl diazoacetate (0.089 g, 0.62 mmol), styrene
(0.476 g, 4.57 mmol), chlorobenzene (0.490 g) and MCM-
−6
41-(CH2)2COO–Rh2(4R-BNOX)3 (0.018 g, 3.21 × 10
mol Rh) were used, the reaction was performed under reflux.
The following results were obtained: yield: 51%, trans/cis:
72/28, eecis: 21%, eetrans: 14%.
2.3.3. Trans 6b
1
3
C NMR (400 MHz, CDCl3, δ (ppm)): 172.5 (C=O,
trans), 140.5, 128.4 (2 C’s), 126.3, 126.1 (2 C’s), (aro-
matic C’s), 80.5 (C(CH3)3), 28.2 (C(CH3)3), 25.7 (Ph–
CHCH2CH), 25.3 (Ph–CHCH2CH), 17.0 (Ph–CHCH2CH).
1
H NMR (400 MHz, CDCl3, δ (ppm)): 7.3–7.0 (m, 5H,
3
3
Ar–H ), 2.43 (ddd, 1 H, JH1–H2 = 4.2 Hz, JH1–H3 = 6.3,
3
2
.3. Typical cyclopropanation procedure
9.2 Hz, CH2CH COOtert-Bu), 1.83 (ddd, 1 H, JH2–H1 =
3
4
.2 Hz, JH2–H3 = 5.3, 8.4 Hz, PhCH CH2), 1.53 (m, 1 H,
Dichloromethane, 3 ml, styrene (0.490 g, 4.70 mmol),
PhCHCH2CH), 1.46 (s, 9H, COOC(CH3)3), 1.22 (m, 1 H,
PhCHCH2CH).
and chlorobenzene (0.2435 g) as the internal standard
were added to the catalyst (SiO2–(CH2)2COO–Rh2(4R-
−
6
BNOX)3, 0.077 g, 4.36 × 10 mol Rh) and the mixture
was stirred at room temperature. Over a period of 3–5 h
a solution of ethyl diazoacetate (0.052 g, 0.453 mmol) in
2.4. Leaching test
The cyclopropanation reaction was performed following
the typical procedure described above for ethyl diazoacetate,
using SiO2–(CH2)3COO–Rh2 (4R-BNOX)3 (0.325 g, 4.23×
3
ml dichloromethane was added. After stirring overnight
at room temperature, the solvent was evaporated in vacuo
and the residue was chromatographed on silica gel using
−6
10 mol Rh), ethyl diazoacetate (0.211 g, 1.85 mmol),
styrene (1.999 g, 0.0192 mol) and chlorobenzene (0.4484 g).
Two minutes after the addition of ethyl diazoacetate (5a) was
complete, no diazo compound could be detected by GC any-
more. After stirring overnight at room temperature, the solid
catalyst was allowed to settle. The solution was transferred
to another vial through a syringe filter to remove traces of
immobilised catalyst. After 30 min a GC sample was taken
of this solution as a baseline sample. Then ethyl diazoac-
etate (0.220 g, 1.92 mmol) was added and two minutes later
9
/1 hexane/ethyl acetate. Yield: 84%, trans/cis: 60/40,
eecis: 33% (1S, 2R), eetrans: 35% (1S, 2S). Before chiral GC
analysis, the products were converted into the corresponding
methyl esters by reaction with 0.1 molar solution of NaOH
in MeOH [30].
2
.3.1. Trans 6a
1
3
C NMR (300 MHz, CDCl3, δ (ppm)): 173.4 (C=O),
1
40.1, 128.5 (2 C’s), 126.5, 126.1 (2 C’s), (aromatic C’s),

H.M. Hultman et al. / Journal of Catalysis 217 (2003) 275–283
279
a GC sample was taken. Hardly any changes were observed.
The mixture was left to stir. Samples were taken after 16 h
and 87.5 h. From GC analysis it was shown that 66% of 5a
was consumed after 16 h. After 87.5 h, 11% of 5a were still
present. This is approximately equivalent to 0.5% of the ac-
tivity of the immobilised catalyst. From this we can conclude
that only a very small amount of catalytically active material
leached during the reaction.
MEPY system displays chiral induction. The catalysts im-
mobilised on Aerosil 200 showed similar activity; moreover
the BNOX systems (entries 2–4, 6) showed more than a
10-fold increase in enantioselectivity. This is clear evidence
that, although one chiral ligand was lost due to the method
of immobilisation, the spatial confinement can lead to a sig-
nificant improvement of enantioselectivity. In contrast to the
catalysts immobilised on silica, none of the catalysts immo-
bilised inside the pores of MCM-41 showed significant ac-
tivity. Even after refluxing overnight, large amounts of un-
modified diazo starting compound remained. Possibly there
is not enough space inside the pores of MCM-41 for the reac-
2
.5. Reuse following typical cyclopropanation procedure
The cyclopropanation reaction was performed follow-
3
ing the typical procedure described above, using SiO2–
tion to take place. Indeed, the average pore diameter (19 Å)
−5
C6H4COO–Rh2(5S-MEPY)3 (0.185 g, 1.67 × 10 mol),
is only slightly larger than the catalyst size (between 19 and
13 Å). A transition state requiring a space-demanding con-
formation of the catalyst might, therefore, be impossible un-
der these circumstances.
ethyl diazoacetate (0.111 g, 0.968 mmol), styrene (1.017 g,
9
.76 mmol), and chlorobenzene (0.3071 g). After stirring
overnight at room temperature, the solid catalyst was al-
lowed to settle. The solution was transferred to another vial
through a syringe filter to remove traces of immobilised cat-
alyst. The conversion was determined by GC. The solution
We also tested if the immobilised catalyst could be re-
cycled in the Si–H insertion reaction (Fig. 1, reaction 1).
SiO –(CH ) COO–Rh (4S-BNOX) was used as the cata-
2
2 2
2
3
−8
contained 3.20 × 10 mol rhodium (determined by AAS).
The remaining solid was washed with dichloromethane and
dried. It was then used again following the same procedure.
After three cycles, the rhodium content of the catalyst was
determined by ICP OES. In all three cycles complete conver-
sion was observed after 16 h. Before reaction the rhodium
content was 0.090 mmol/g, and after three cycles 0.041
mmol/g was left on the carrier.
lyst. Since all catalysts were immobilised in the same man-
ner it can be assumed that these results are representative
for all the different immobilised catalysts described here.
These experiments show significant deactivation of the im-
mobilised catalysts. In the first cycle 79% conversion was
reached, while after the third cycle only 19% conversion was
observed. After the third cycle, the liquid phase was removed
from the catalyst and to this solution an additional amount
of dimethylphenylsilane and methyl phenyldiazoacetate was
added. No catalytic activity was observed in the resulting
mixture. The ee’s were approximately 35% (the S product
was the major product) in the first two cycles. In the third
cycle the ee decreased to 11%.
3
. Results and discussion
3
.1. Si–H insertion
In the Si–H insertion reaction (Fig. 1, reaction 1) very
3.2. Cyclopropanation
clear differences can be observed between the different ho-
mogeneous and immobilised catalysts (Table 1). The ho-
mogeneous catalysts (entries 1, 7) are active, but only the
For the cyclopropanation reaction (Fig. 1, reaction 2),
the immobilisation of the catalyst gave an increase of the
Table 1
Comparison of different rhodium catalysts used in the Si–H insertion reaction immobilised on different supports via different tethers
a
b
Entry
Catalyst
c
Yield (%)
ee (%)
1
2
3
4
5
6
7
8
9
Rh (4R-BNOX)
73
88
61
67
2
20
26
28
–
33
37
2
2
4
SiO –(CH ) COO–Rh (4R-BNOX)
2
2 2
2
3
3
SiO –(CH ) COO–Rh (4R-BNOX)
2
2 3
2
SiO –C H COO–Rh (4R-BNOX)
2
6
4
2
3
MCM-41-(CH ) COO–Rh (4R-BNOX)
Only traces of product detected
2
2
2
3
d
SiO –(CH ) COO–Rh (4S-BNOX)
79
70
78
65
2
2 2
2
3
e
Rh (5S-MEPY)
2
4
SiO –(CH ) COO–Rh (5S-MEPY)
2
2 2
2
3
SiO –C H COO–Rh (5S-MEPY)
3
1
2
6
4
2
10
MCM-41-(CH ) COO–Rh (5S-MEPY)
3
Only traces of product detected
–
2
2
2
a
All reported yields are isolated yields.
b
Analysis was performed according to Ref. [31], where the second eluted product was determined to be the S-product by correlation to methyl-(S)-(+)-
mandelate.
c
The reactions using R-BNOX as the ligand gave the R product as the major isomer.
The reactions using S-BNOX as the ligand gave the S product as the major isomer.
The reactions using S-MEPY as the ligand gave the S product as the major isomer.
d
e

280
H.M. Hultman et al. / Journal of Catalysis 217 (2003) 275–283
Table 2
Comparison of EDA (5a) and TBDA (5b) in the cyclopropanation reaction (Fig. 1, reaction 2)
Entry
Catalyst
Diazo
compound
Yield
(%)
trans/cis
ratio
eecis
(%)
eetrans
(%)
b
c
d
d
a
1
2
3
4
5
6
7
8
9
Rh (5S-MEPY)
5a
5a
5a
5b
5b
5b
5a
5a
5a
5b
5b
5b
59
73
65
50
62
50
79
84
51
64
53
51
56/44
59/41
60/40
60/40
71/29
74/26
46/54
60/40
70/30
59/41
66/34
72/28
33
29
19
66
30
55
2
33
29
34
19
21
58
35
22
14
27
14
17
35
36
9
2
4
SiO –(CH ) COO–Rh (5S-MEPY)
2
2 2
2
3
MCM-41-(CH ) COO–Rh (5S-MEPY)
2
2
2
3
Rh (5S-MEPY)
2
4
SiO –(CH ) COO–Rh (5S-MEPY)
2
2 2
2
3
MCM-41-(CH ) COO–Rh (5S-MEPY)
2
2
2
3
Rh (4R-BNOX)
2
4
SiO –(CH ) COO–Rh (4R-BNOX)
2
2 2
2
3
MCM-41-(CH ) COO–Rh (4R-BNOX)
2
2
2
3
1
1
1
0
1
2
Rh (4R-BNOX)
2 4
SiO –(CH ) COO–Rh (4R-BNOX)
3
9
14
2
2 2
2
MCM-41-(CH ) COO–Rh (4R-BNOX)
2
2
2
3
a
Literature value [20].
All reported yields are isolated yields.
Determined by GC.
b
c
d
The ee’s for the ethyl esters were determined by chiral GC after conversion to the corresponding methyl esters [30], the ee’s for tert-butyl esters were
determined by chiral GC. The main enantiomer for cis 6a using S-MEPY or R-BNOX as the ligand is 1S,2R; the main enantiomer for trans 6a using S-MEPY
or R-BNOX as the ligand is 1S,2S [30]. For 6b, the same enantioselectivity is assumed.
trans/cis ratios when ethyl diazoacetate (5a) was used
as the diazo compound, and even more (from 56/44 to
actions described in Table 3, dichloromethane was the sol-
vent. In most reactions the noncoordinating chlorobenzene
was added as the internal standard. In some cases this was
replaced by toluene, which then served as both the internal
standard and the inhibitor. The catalysts described in Table 3
were immobilised on carriers with free silanol groups. The
outer surface of the MCM-41 carriers was protected in order
to make sure that the catalyst was immobilised inside the
pore. Selectivities are expected to be higher at lower conver-
sions. Indeed, while the yields decrease significantly in the
presence of toluene (Table 3), the trans/cis ratios and ee’s
(especially of the cis product) are higher in its presence.
The above results show that, as predicted, modification
of the regioselectivity of the immobilised catalysts can
be observed for the cyclopropanation reaction (Fig. 1,
reaction 2). In this reaction the ratio between the trans and
cis product formed is determined by the steric repulsion
between the phenyl group of the styrene and the ester group
of the carbene. As can be seen in Figs. 4a and 4b, in
the homogeneous reaction, the ligand does not significantly
restrict the incoming styrene, and, thus, the difference in
7
6/24) for tert-butyl diazoacetate (5b) (Table 2). While the
trans/cis ratios were modified significantly due to the spatial
confinement, some enantioselectivity was lost, probably due
to the spatial confinement in this case not fully compensating
for the loss of one of the chiral ligands per complex.
All immobilised catalysts listed in Table 2 have protected
surfaces [3].
A method to enhance the effects of the immobilisation is
to add a weak inhibitor. In the presence of an inhibitor, the
speed of the reaction decreases and differences in selectivity
become more evident. Because of the coordinative unsatu-
ration of the active metal catalyst, Lewis bases that can as-
sociate with the metal inhibit diazo decomposition. Amines,
sulphides, and nitriles are generally effective inhibitors for
dirhodium(II)-catalysed diazo decomposition, but alkenes
and alkylbenzenes can also play this role. Halogenated hy-
drocarbons are known not to coordinate with dirhodium(II)
complexes and, therefore, they serve as useful solvents for
the reactions catalysed by these complexes [32]. In the re-
Table 3
Influence inhibitor in the reaction of ethyl diazoacetate (5a) and styrene (Fig. 1, reaction 2)
a
b
c
c
Catalyst
Toluene
Yield (%)
trans/cis
eetrans (%)
eecis (%)
1
2
3
4
5
6
. Rh (4S-BNOX)
yes
no
yes
no
yes
no
32
85
14
30
10
58
45/55
49/51
66/34
60/40
65/35
59/41
43
34
26
27
37
39
19
8
24
16
36
27
2
4
. Rh (4S-BNOX)
2
4
. MCM-41-C H COO–Rh (4S-BNOX)
6
4
2
3
3
. MCM-41-C H COO–Rh (4S-BNOX)
6
4
2
. SiO –(CH ) COO–Rh (4S-BNOX)
2
2 3
2
3
. SiO –(CH ) COO–Rh (4S-BNOX)
2
2 3
2
3
a
b
c
All reported yields are isolated yields.
Determined by GC.
The ee’s for the ethyl esters were determined by chiral GC after conversion to the corresponding methyl esters [30]. The main enantiomer for cis 6a using
S-MEPY or R-BNOX as the ligand is 1S,2R; the main enantiomer for cis 6a using R-MEPY or S-BNOX as the ligand is 1R,2S; the main enantiomer for
trans 6a using R-MEPY or S-BNOX as the ligand is 1R,2R; the main enantiomer for trans 6a using S-MEPY or R-BNOX as the ligand: 1S,2S [30].

H.M. Hultman et al. / Journal of Catalysis 217 (2003) 275–283
281
that the increased spatial restrictions, due to the surface pro-
tection, are causing these improvements. Additionally, polar
interactions can also have a positive effect on the catalyst, as
can be concluded from the fact that in the case of catalysts
immobilised on unprotected carriers, the observed ee’s are
moderately higher than with the protected surfaces. Molec-
ular modelling studies are under way to probe these effects
more deeply.
The influence of the spacer group in the reaction of
styrene with ethyl diazoacetate does not show a clear pattern
(
Fig. 1, reaction 2). Contrary to expectation, its influence
seems rather small (Table 5). Although entries 4, 5, and
indicate that shorter and, especially, rigid spacer groups
improve the trans/cis ratios significantly, entries 9, 10, and
1 do not support this trend. Even if it is taken into account
6
1
that some of the surfaces were protected and others not, there
is no clear pattern.
We also immobilised Rh2(4R-BNOX)4 on MCM-41
with free internal and external silanol groups and on fully
protected MCM-41 (protected with dimethoxydimethylsi-
lane) [3]. The immobilisation on unprotected MCM-41 gave
a blue solid after Soxhlet extraction. This solid was active
in the cyclopropanation of styrene (4) with ethyl diazoac-
etate (5a) (Table 6, entry 2). The results were similar to those
for the homogeneous catalyst (Table 6, entry 1). This clearly
demonstrates that the fine tuning of the catalyst in the pore
is essential. Without a tether and surface protection, the pore
is larger and, therefore, does not induce any additional se-
lectivity. Hölderich and co-workers [33] reported similar re-
sults when they immobilised complexes inside the pores of
MCM-41 via ionic interactions. The immobilisation on pro-
tected MCM-41 was not successful. The resulting solid was
white and was not active in the cyclopropanation of styrene
(4) with ethyl diazoacetate (5a). We also compared the selec-
tivity of Rh2(4R-BNOX)4 immobilised on protected MCM-
41-(CH2)3COOH in dichloromethane (Table 6, entry 5) and
of Rh2(4R-BNOX)4 immobilised on protected MCM-41-
(CH2)3CN in toluene (Table 6, entry 4). The catalyst immo-
bilised via the CN tether behaved as the homogeneous cata-
lyst, while the catalyst immobilised via the COO tether gave
the increased selectivity that was observed for all the other
examples in this paper. From these results we assume that the
CN tether immobilises the catalyst via the axial position and,
therefore, the catalyst behaves as if it was homogeneous.
Fig. 4. Transition state of the homogeneous reaction (a trans orientation, b
cis orientation) and heterogeneous reaction (c).
the formation of the trans and cis products is not large. In
Fig. 4c, the transition state of the heterogeneous reaction
is shown. In this case, steric hindrance by the bulky carrier
surface forces the carbene upwards, away from the surface.
The carbene then directs, due to the bulk of the ester
group, the incoming styrene in such a way that more trans
compound is formed than in the homogeneous reaction. If
the more bulky tert-butyl diazoacetate is used, both effects
are even more pronounced. If the catalysts are immobilised
on Aerosil 200, the effect is smaller than if the catalysts are
immobilised inside MCM-41. This can be explained by the
predicted steric confinement within the pores, which restricts
the incoming styrene even more.
The influence of the polarity of the surface was inves-
tigated in the cyclopropanation of styrene with tert-butyl
diazoacetate (Fig. 1, reaction 2). The trans/cis ratios (vide
supra) increase if the catalysts are immobilised; however, if
the surfaces are protected, the increases are even higher for
both Aerosil 200 and MCM-41 immobilised catalysts (Ta-
ble 4). The results indicate that, if no polar interactions are
possible between the surface and the catalyst, this has a pos-
itive influence on the selectivity. Alternatively, it is possible
Table 4
Influence surface polarity on the cyclopropanation reaction of styrene with tert-butyl diazoacetate (5b) (Fig. 1, reaction 2)
a
b
c
c
Entry
Catalyst
Protected
Yield (%)
trans/cis
eetrans (%)
ee (%)
cis
1
2
3
4
5
Rh (4R-BNOX)
–
yes
no
yes
no
64
51
45
53
61
59/41
72/28
56/44
66/34
64/36
9
14
18
9
34
21
27
19
21
2
4
MCM-41-(CH ) COO–Rh (4R-BNOX)
2
2
2
3
3
MCM-41-(CH ) COO–Rh (4R-BNOX)
2
2
2
SiO –(CH ) COO–Rh (4R-BNOX)
2
2 2
2
3
3
SiO –(CH ) COO–Rh (4R-BNOX)
16
2
2 2
2
a
All reported yields are isolated yields.
Determined by GC.
ee’s for tert-butyl esters were determined by chiral GC.
b
c

282
H.M. Hultman et al. / Journal of Catalysis 217 (2003) 275–283
Table 5
Influence spacer length on the reaction of styrene with ethyl diazoacetate (5a) (Fig. 1, reaction 2)
b
c
d
d
Catalyst
Protected
Yield (%)
trans/cis
eecis (%)
eetrans (%)
a
1
2
3
4
5
6
. Rh (5S-MEPY)
–
59
27
65
51
73
53
56/44
52/48
60/40
94/6
59/41
53/47
33
24
19
15
29
21
58
20
22
37
35
29
2
4
. MCM-41-C H COO–Rh (5S-MEPY)
no
yes
no
yes
no
6
4
2
3
. MCM-41-(CH ) COO–Rh (5S-MEPY)
2
2
2
3
. SiO –C H COO–Rh (5S-MEPY)
. SiO –(CH ) COO–Rh (5S-MEPY)
. SiO –(CH ) COO–Rh (5S-MEPY)
2
6
4
2
3
2
2 2
2
3
3
2
2 3
2
7
8
9
1
1
. Rh (4S-BNOX)
–
85
30
50
71
58
49/51
60/40
66/34
82/18
59/41
8
16
17
35
27
34
27
18
40
39
2
4
. MCM-41-C H COO–Rh (4S-BNOX)
no
no
yes
no
6
4
2
3
. SiO –C H COO–Rh (4S-BNOX)
2
6
4
2
3
0. SiO –(CH ) COO–Rh (4S-BNOX)
2
2 2
2
3
1. SiO –(CH ) COO–Rh (4S-BNOX)
2
2 3
2
3
1
1
1
1
1
1
2. Rh (4R-BNOX)
–
no
yes
no
yes
yes
79
70
51
75
84
63
46/54
48/52
70/30
40/60
60/40
56/44
2
16
29
15
33
38
17
39
36
36
35
45
2
4
3. MCM-41-C H COO–Rh (4R-BNOX)
6
4
2
3
4. MCM-41-(CH ) COO–Rh (4R-BNOX)
2
2
2
3
5. SiO –C H COO–Rh (4R-BNOX)
2
6
4
2
3
6. SiO –(CH ) COO–Rh (4R-BNOX)
2
2 2
2
3
3
7. SiO –(CH ) COO–Rh (4R-BNOX)
2
2 3
2
a
b
c
d
Literature value [20].
All reported yields are isolated yields.
Determined by GC.
The ee’s for the ethyl esters were determined by chiral GC after conversion to the corresponding methyl esters [30]. The main enantiomer for cis 6a using
S-MEPY or R-BNOX as the ligand is 1S,2R; the main enantiomer for cis 6a using R-MEPY or S-BNOX as the ligand is 1R,2S; the main enantiomer for
trans 6a using R-MEPY or S-BNOX as the ligand is 1R,2R; the main enantiomer for trans 6a using S-MEPY or R-BNOX is the ligand: 1S,2S [30].
3
.3. Leaching test
dichloromethanewith chlorobenzene as an internal standard,
with SiO2–C6H4COO–Rh2(5S-MEPY)3 as the catalyst at
room temperature (Fig. 1, reaction 2). The conversions were
determined by GC. In the second cycle a decrease in activity
by 20% was observed, which remained stable in the third cy-
cle. In all cases complete conversionwas observed after 16 h.
In order to evaluate whether any of the rhodium leaches
into solution during the reaction, the activity of the filtrate
of a cyclopropanation reaction of styrene (4) with ethyl
diazoacetate (5a) catalysed by SiO2–(CH2)2COO–Rh2(4R-
BNOX)3 was determined. One minute after the addition of
ethyl diazoacetate (5a) was complete, no diazo compound
could be detected by GC anymore; i.e., 100% had been
converted. To the filtrate of this reaction 5a was added. From
GC analysis it was shown that 66% of 5a was consumed after
4
. Conclusion
By immobilising the chiral dirhodium complexes on
1
6 h. After a further three days, 11% of 5a was still present.
Aerosil 200 and inside the pores of MCM-41, we demon-
strate that it is possible to achieve better selectivities using
this strategy:
In the Si–H insertion reactions a significant increase of
enantioselectivity was obtained with the BNOX catalysts
immobilised on silica. Earlier work in this area described
an increase in ee from 6 to 17% for a reduction and from 43
This behaviour corresponds to a conversion of, after 2 min,
less than 0.5%, whereas in the homogeneous case after 2
min full conversion was reached. Therefore, only traces of
catalyst leach during the reaction.
Recycling experiments were performed for the cyclo-
propanation of styrene (4) with ethyl diazoacetate (5a) in
Table 6
Variation of immobilisation mechanisms in the cyclopropanation of styrene and ethyl diazoacetate (5a) (Fig. 1, reaction 2)
a
b
c
c
Catalyst
Protected
Yield (%)
trans/cis
eecis (%)
eetrans (%)
1
2
3
4
5
. Rh (4R-BNOX)
. MCM-41-OH–Rh (4R-BNOX)
. MCM-41-SiMe2–Rh (4R-BNOX)
–
no
yes
yes
yes
79
74
–
63
65
46/54
47/53
–
47/53
63/37
2
7
–
13
37
17
32
–
30
37
2
4
2
4
2
4
. MCM-41-(CH ) CN–Rh (4R-BNOX)
2
3
2
4
. MCM-41-(CH ) COO–Rh (4R-BNOX)
2
3
2
3
a
b
c
All reported yields are isolated yields.
Determined by GC.
The ee’s for the ethyl esters were determined by chiral GC after conversion to the corresponding methyl esters [30]. The main enantiomer for cis 6a using
R-BNOX as the ligand is 1S,2R; the main enantiomer for trans 6a using R-BNOX is the ligand: 1S,2S [30].

H.M. Hultman et al. / Journal of Catalysis 217 (2003) 275–283
283
to 96% for an allylic amination [22,23]. This is to say, the
selectivity of the reactions was improved. Our conversion of
an essentially unselective reaction (ee 2%) into a selective
reaction (ee 28%) in the case of the Si–H insertion reaction
is, thus, another step forward.
1982–1994, in: Transition Metal Organometallics in Organic Synthe-
sis, Vol. 12, Elsevier/Pergamonn, Oxford, 1995.
2] J.M. Thomas, T. Maschmeyer, B.F.G. Johnson, D.S. Shephard, J. Mol.
Catal. A Chem. 141 (1999) 139–144.
3] H.M. Hultman, M. de Lang, M. Nowotny, I.W.C.E. Arends, U. Hane-
feld, R.A. Sheldon, T. Maschmeyer, J. Catal., in press.
[
[
[
In the case of the cyclopropanation reactions, we ob-
served a significant improvement of the trans/cis ratios.
This is a general result for both catalysts, immobilised
on a variety of different supports. However, we have to
note that in this case the enantioselectivity of the reaction
did not improve. Leaching experiments showed that only
traces of catalytically active material leached during the
cyclopropanation reaction. We also proved that with pro-
tected carrier surfaces, the catalyst was immobilised via the
tether. With unprotected carrier surfaces, the immobilisation
can proceed via ligand exchange with the tether or via ad-
sorption to the surface silanol groups. In this case the influ-
ence of the spatial confinement is less pronounced.
4] M.P. Doyle, B.D. Brandes, A.P. Kazala, R.J. Pieters, M.B. Jarstfer,
L.M. Watkins, C.T. Eagle, Tetrahedron Lett. 31 (1990) 6613–6616.
[5] H. Fritschi, U. Leutenegger, A. Pfaltz, Helv. Chim. Acta 71 (1988)
553.
1
[
[
[
[
6] T. Aratani, Y. Yoneyoshi, T. Nagase, Tetrahedron Lett. (1977) 2599.
7] C.J. Suckling, Angew. Chem. Int. Ed. Engl. 27 (1988) 537–552.
8] G. Perger, D. Szadkowski, Ann. Agr. Environ. Med. 1 (1994) 11–17.
9] M.G. Banwell, G.S. Forman, J. Chem. Soc. Perkin Trans. I (1996)
2565.
[10] C. Duquenne, S. Goumain, P. Jubault, C. Feasson, J-C. Quirion, Org.
Lett. 2 (2000) 453–455.
11] F.C. Shu, Q.-L. Zhou, Synth. Commun. 29 (4) (1999) 567–572.
12] R. Csuk, M.J. Schabel, Y. von Scholz, Tetrahedron Asymmetry 7 (12)
[
[
(1996) 3505–3512.
[
13] M.P. Doyle, W. Hu, Adv. Synth. Catal. 343 (2001) 299–302.
With the above experiments, we have shown that our
synthesis method gives us the opportunity to fine-tune the
pore size in such a way that we can increase the selectivity
of the test reactions. By changing the pore size of the
carrier materials (using silica or MCM-41) we can increase
the selectivity for the cyclopropanation reaction, while by
immobilising the catalysts on Aerosil 200 we can do the
same for the Si–H insertion reaction.
[
14] I. Hindmarch, U. Rigney, N. Stanley, M. Briley, Br. J. Clin. Pharm. 49
(2000) 118–125.
[15] H.J. Moller, J. Clin. Psychiatry 61 (2000) 24–28.
[16] H.N.C. Wong, M.-Y. Hon, C.-W. Tse, Y.-C. Yip, J. Tanko, T. Hudlicky,
Chem. Rev. 89 (1989) 165–198.
[17] A.B. Charette, H. Lebel, in: E.N. Jacobsen, A. Pfaltz, H. Yamamoto
Eds.), Comprehensive Asymmetric Catalysis, Springer, Berlin/Hei-
(
delberg/New York, 1999, p. 581.
[18] V. Bagheri, M.P. Doyle, J. Taunton, E. Claxton, J. Org. Chem. 53
The deactivation that occurrs during recycling of the
immobilised catalysts is caused by rhodium leaching and
probably also by clogging of the pores. The deactivation
is higher at a higher temperature. Tests of the filtrate
demonstrated that the rhodium that leaches is not active.
Despite the leaching, we were still able to obtain full
conversion in the cyclopropanation reaction. In the Si–H
insertion reactions a loss of activity was observed; however,
the ee’s remained.
In short, we have shown that immobilisation of the ho-
mogeneous dirhodium catalysts on Aerosil 200 surfaces and
inside MCM-41 affords a significant improvement in regio-
selectivity (cyclopropanation reaction) and enantioselectiv-
ity (Si–H insertion). The improvement is attributed to the
confinement resulting from immobilisation.
(
1988) 6158–6160, and references therein.
[19] M.P. Doyle, J. Taunton, S.-M. Oon, M.T.H. Liu, N. Soundararajan,
M.S. Platz, J.E. Jackson, Tetrahedron Lett. 29 (1988) 5863–5866.
20] P. Müller, C. Baud, D. Ene, S. Motallebi, M.P. Doyle, B.D. Brandes,
A.B. Dyatkin, M.M. See, Helv. Chim. Acta 78 (1995) 459–470.
21] M.P. Doyle, W.R. Winchester, J.A.A. Hoorn, V. Lynch, S.H. Simon-
sen, R. Ghosh, J. Am. Chem. Soc. 115 (1993) 9968–9978.
22] S.A. Raynor, J.M. Thomas, R. Raja, B.F.G. Johnson, R.G. Bell, M.D.
Mantle, Chem. Commun. (2000) 1925–1926.
[23] B.F.G. Johnson, S.A. Raynor, D.S. Shephard, T. Maschmeyer, J.M.
Thomas, G. Sankar, S. Bromley, R. Oldroyd, L. Gladden, M.D.
Mantle, Chem. Commun. 13 (1999) 1167–1168.
[
[
[
[
[
24] R.J. Clarke, I.J. Shannon, Chem. Commun. (2001) 1936–1937.
25] M.P. Doyle, M.Y. Eismont, D.E. Berzbreiter, H.N. Gray, J. Org.
Chem. 57 (1992) 6103–6105.
[
26] M.P. Doyle, D.J. Timmons, J.S. Tumonis, H-M. Gau, E.C. Blossey,
Organometallics 21 (2002) 1747–1749.
[
[
27] E.S. Gil, L.T. Kubota, Bioelectrochemistry 51 (2000) 145–149.
28] W.A.J. Starmans, L. Thijs, B. Zwanenburg, Tetrahedron 54 (1998)
6
29–636.
Acknowledgments
[
29] M. Regitz, J. Hocker, A. Liedhegener, in: Organic Syntheses, Collec-
tive Volume V, a Revised Edition of Annual Volumes 40–49, 1973,
pp. 179–183.
30] H. Nishiyama, Y. Itoh, Y. Sugawara, H. Matsumoto, K. Aoki, K. Itoh,
Bull. Chem. Soc. Jpn. 68 (1995) 1247–1262.
The Royal Netherlands Academy of Arts and Sciences
(
U.H.) and the Netherlands Research School Combination
[
Catalysis (H.M.H.) are acknowledged for financial support.
[31] L.A. Dakin, S.E. Schaus, E.N. Jacobsen, J.S. Panek, Tetrahedron
Lett. 39 (1998) 8947–8950.
References
[32] M.P. Doyle, T. Ren, in: K.D. Karlin (Ed.), Progress in Inorganic
Chemistry, Vol. 49, Wiley, New York, 2001, and references therein.
[33] H.H. Wagner, H. Hausmann, W.F. Hölderich, J. Catal. 203 (2001) 150–
156, and references cited therein.
[1] M.P. Doyle, in: E.W. Abel, F.G.A. Stone, G. Wilkinson (Eds.), Com-
prehensive Organometallic Chemistry II. A Review of the Literature