Modified silica-heterogenised catalysts for use in aqueous enantioselective hydrogenations

Article abstract of DOI:10.1016/S0022-328X(00)00925-6

Modified mesoporous silicas have been prepared, homogeneous organometallic catalysts incorporated and the resulting heterogeneous systems used as catalysts in aqueous enantioselective hydrogenation reactions. A series of catalysts in which the organometallic species was incorporated during gel synthesis generally gave good conversions but low ee values which were usually less than 25%. Preformed silicas with a narrow range of pore size (26 and 37 ?) and an amorphous silica (average pore size 68 ?) were modified by external surface deactivation and/or derivatization of the internal surface for potential tethering of the catalytic species. In general, excellent conversions (80-100%) but modest enantioselectivities (40-50%) were obtained when using 37 and 68 ? silicas, similar to those values obtained using unmodified silicas. Only low conversions were obtained using the modified 26 ? silicas. Pore volume measurements suggest that this is due to very restricted access to the pores when the tether is present. The results indicate that for aqueous hydrogenation the van der Waals interactions of the catalyst with the porous matrix are sufficiently strong to retain the catalyst, thus allowing for its reuse. No major advantage then appears to be gained by the use of potential internal tethers.

Full text of DOI:10.1016/S0022-328X(00)00925-6

Journal of Organometallic Chemistry 627 (2001) 37–43
www.elsevier.nl/locate/jorganchem
Modified silica-heterogenised catalysts for use in aqueous
enantioselective hydrogenations
Jim Jamis, John R. Anderson, Ron S. Dickson *, Eva M. Campi, W. Roy Jackson
Department of Chemistry, PO Box 23, Monash Uni6ersity, Vic. 3800, Australia
Received 25 September 2000; received in revised form 26 November 2000; accepted 26 November 2000
Abstract
Modified mesoporous silicas have been prepared, homogeneous organometallic catalysts incorporated and the resulting
heterogeneous systems used as catalysts in aqueous enantioselective hydrogenation reactions. A series of catalysts in which the
organometallic species was incorporated during gel synthesis generally gave good conversions but low ee values which were
,
usually less than 25%. Preformed silicas with a narrow range of pore size (26 and 37 A) and an amorphous silica (average pore
,
size 68 A) were modified by external surface deactivation and/or derivatization of the internal surface for potential tethering of
the catalytic species. In general, excellent conversions (80–100%) but modest enantioselectivities (40–50%) were obtained when
,
using 37 and 68 A silicas, similar to those values obtained using unmodified silicas. Only low conversions were obtained using the
,
modified 26 A silicas. Pore volume measurements suggest that this is due to very restricted access to the pores when the tether
is present. The results indicate that for aqueous hydrogenation the van der Waals interactions of the catalyst with the porous
matrix are sufficiently strong to retain the catalyst, thus allowing for its reuse. No major advantage then appears to be gained by
the use of potential internal tethers. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Rhodium; Ruthenium; Asymmetric hydrogenation; Aqueous solution; Heterogenised catalyst; Silica
1. Introduction
Other workers have modified preformed mesoporous
silicas by attaching potential tethers, e.g. by reaction
with (MeO)3SiCH2CH2CH2NH2 (3AMPS) [3]. Restric-
tion of reaction to within the pores of the silicas has
been achieved by blocking the surface hydroxyl groups
by reaction with Ph2SiCl2 followed by attachment of
potential tethers to internal surfaces, e.g. by reactions
with 3AMPS [4] or (MeO)3SiCH2CH2CH2CN (CETS)
(subsquently hydrolysed to the carboxylic acid) [5].
In this paper we report experiments designed to
compare the efficiency of our previous methodology (in
which the catalysts were simply incorporated into silicas
of appropriate pore sizes) with that using modified
silicas containing tethers with potential ligands for
binding the metal atoms.
Incorporation of homogeneous organometallic cata-
lysts into hexagonal mesoporous silicas (HMS) and
their use for enantioselective hydrogenation reactions in
aqueous media have been described by us [1]. Recently,
other workers have incorporated similar organometallic
catalysts into silicas by adding the catalyst to a sol–gel
matrix based on tetramethyl orthosilicate prior to dry-
ing [2]. Use of these materials as heterogeneous cata-
lysts for the enantioselective hydrogenation of, e.g.
itaconic acid gave quantitative conversion to materials
with enantioselectivities of up to 78%. The catalysts
were recoverable and reusable. Catalyst activity was
enhanced by a hot aqueous bicarbonate wash which
was reported to open blocked pores and create new
ones [2].
2. Experimental
2.1. Instrumentation
* Corresponding author. Tel.: +61-3-99054560; fax: +61-3-
99059129.
E-mail address: dickson@smtp.monash.edu.au (R.S. Dickson).
1H-NMR spectra were recorded on a Bruker DRX-
400 spectrometer (400 MHz). The 1H-NMR spectra
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(00)00925-6

38
J. Jamis et al. / Journal of Organometallic Chemistry 627 (2001) 37–43
were run in D2O solution with TMSP as the internal
standard or in CD3OD with TMS as the internal stan-
dard with chemical shifts measured in ppm. Solid state
31P-NMR spectra were recorded on an INOVA-300
spectrometer using magic-angle spinning at 121.5 MHz
and were referenced to P2O5. The optical yields were
determined by polarimetry using a Perkin Elmer 1600
polarimeter by comparison with the literature values for
the optically pure compounds: N-acetyl-(R)-phenylala-
nine, [h]2D6 −51.8° (C=1, EtOH) [6]; (R)-2-methylsuc-
cinic acid, [h]2D0 +16.88° (C=22.2, EtOH) [7]. Ru and
Rh analyses were carried out using a Varian Liberty
200 inductively coupled plasma emission spectrometer
(ICP-AES) with diisobutylketone or water as solvent.
dimer, dichloro[(R)-(+)-2,2%-bis(di-p-tolylphosphino)-
1,1%-binaphthyl]ruthenium(III) dimer triethylamine ad-
duct, [(Ru–BINAP)2] were obtained from commercial
sources. Sodium a-acetamidocinnamate (1a) was pre-
pared by stirring a-acetamidocinnamic acid (1b) (2.03 g,
9.9 mmol) in an aqueous solution of sodium hydroxide
(10 ml) containing 9.8 mmol of NaOH. Excess acid was
removed by filtration and the sodium salt isolated by
the removal of water under vacuum at ambient temper-
ature.
2.3. Silicas and modified silicas
,
Ordered mesoporous silica (HMS) of 26 A average
pore size was synthesised in the presence of dodecyl-
amine in water with ethanol as a co-solvent by follow-
ing procedures similar to those reported previously [8].
Before impregnation, the hexagonal mesoporous silica
(HMS) was calcined in air at 550°C for 4 h to remove
the structurally incorporated template and fully hy-
drated by saturation with water vapour at ambient
temperature in a desiccator over saturated aqueous
NH4Cl for 48 h. The water content determined by TGA
2.2. Materials and reagents
Benzene was dried over phosphorus pentoxide, de-
canted, distilled from fresh phosphorus pentoxide and
stored over sodium wire. Tetrahydrofuran (THF) was
stirred over KOH pellets, decanted and distilled under
N2 from sodium and benzophenone prior to use. Dis-
tilled water was used for all aqueous manipulations.
Methanol was freshly distilled before use. All manipula-
tions were carried out under N2 and solvents degassed
before use.
,
was 21%. The average pore size was confirmed as 26 A
by nitrogen adsorption at −196°C using a standard
,
procedure [9]. An HMS of 37 A average pore size was
Itaconic acid (3); a-acetamidocinnamic acid (1b);
bis(cycloocta-1,5-diene)dichlorodirhodium (I), [Rh-
(COD)Cl]2; (R)-(+)-2,2%-bis(diphenylphosphino)-1,1%-
binaphthyl, (+)-(BINAP); (2S,4S)-N-(tert-butoxycar-
bonyl) - 4 - (diphenylphosphino) - 2 - [(diphenylphos-
phino)methyl]pyrrolidine, (BPPM); [(R)-(+)-2,2%-bis-
(diphenylphosphino)-1,1%-binaphthyl]chloro(p-cymene)-
ruthenium chloride, [Ru-BINAP(cymene)] and the
prepared following the published procedure [10] and
the average pore size confirmed using the standard
procedure [9] (see Fig. 1).
An amorphous mesoporous silica was obtained by
first preparing a solution of triethanolamine (3 ml),
N,N-dimethylformamide (2 ml) and phosphoric acid
(10 ml, 85% w/w) in water (125 ml). A portion of this
solution (10 ml) was added with rapid stirring to a
Fig. 1. Pore size distributions for unmodified silicas, [A, Si (37); B, CO2HSi (37); C, NH2Si (37)].

J. Jamis et al. / Journal of Organometallic Chemistry 627 (2001) 37–43
39
solution (40 ml) made up by adding a solution of
sodium silicate (density 1.37 g ml−1) (5 ml) and te-
tramethylammonium bromide (3 g) in water (125 ml).
The resulting gel was left to stand overnight, washed
with water, filtered and dried overnight at ambient
temperature. Calcination at 550°C for 72 h gave a white
powder shown to be amorphous by XRD and charac-
CH2Cl2 and dried under vacuum to constant weight.
The rhodium content of the washings was ca. 5% of the
Rh used (as estimated by ICP-AES analysis).
A commercial Ru–BINAP(cymene) catalyst (20.41
mg, 0.022 mmol) was also incorporated using the above
procedure. The washings contained ca. 6% of the Ru
used.
,
terised as above as having an average pore size of 68 A.
Selected silicas were modified using standard litera-
Two other ruthenium containing gels were prepared
in which solutions of either (Ru–BINAP)2 (16.92 mg,
0.01 mmol) or Ru–BINAP(cymene) (12.06 mg, 0.013
mmol) in THF (2.5 ml) were added simultaneously but
separately with the NH4OH (0.44 ml) to a pre-stirred
mixture of HCl (1.2 ml of a HCl solution, pH 1.96),
(EtO)4Si (3.8 ml), Ph2SiCl2 (0.84 ml) and MeOH (3.5
ml). The mixture was stirred for 5 min at ambient
temperature. Addition of 3AMPS (1 ml) gave a gel. The
gel was dried, crushed, washed thoroughly with boiling
CH2Cl2, sonicated for 30 min in the same solvent and
dried under vacuum until constant weight was
achieved. The ruthenium content of the washings was
negligible (ICP-AES).
ture procedures [4b] to prepare:
The silica with average pore size 26 A and the
,
,
amorphous silica (68 A) were treated with Ph2SiCl2
to give PhSi (26) and PhSi (68), respectively.
,
Silicas (26 and 37 A) treated with 3AMPS gave
NH2Si (26) and NH2Si (37).
,
Silicas (26 and 37 A) treated with Ph2SiCl2 followed
by 3AMPS gave Ph/NH2Si (26) and Ph/NH2Si (37).
,
The silica (37 A) was reacted with CETS [4b] and
hydrolysed using H2SO4 50% v/v to the acidic form
to give CO2HSi (37) [5b,c].
2.4. Incorporation of a ruthenium complex into
prepared silicas
2.6. Hydrogenation procedure
Commercial Ru–BINAP(cymene) (4.64 mg, 0.005
mmol) was dissolved in dry, deoxygenated benzene (5
ml) by stirring for 30 min and the resulting solution
added, using a syringe, to a Schlenk flask containing
the pretreated HMS (100 mg). The mixture was stirred
for 16 h at ambient temperature, filtered under dry N2
and washed three times with degassed benzene (3×5
ml). The filtrate and washings were collected and evap-
orated in vacuo in order to determine the amount of
catalyst absorbed. The impregnated support was dried
under vacuum at ambient temperature for 4 h and then
used immediately. In this manner the following incor-
porated Ru catalysts were prepared:- Ru/PhSi (26 and
68); Ru/NH2Si (26 and 37); Ru/Ph/NH2Si (26 and 37);
and Ru/CO2HSi (37).
The procedure was essentially identical to that de-
scribed previously [1]. A solution of the substrate (200
mg) in degassed distilled water or MeOH.H2O (1:1, v/v)
(10 ml) and the impregnated silica (100 mg) were
hydrogenated under the reported hydrogen pressure
with stirring.
Product isolation from reactions of itaconic acid (3)
involved washing the silica with distilled water (3×5
ml) and evaporating combined filtrate and washings.
1H-NMR spectroscopy was used to determine the ex-
tent of reaction by integration of the CH2 signals at l
2.6 and 2.7 (D2O) in the product and l 3.4 (D2O) in
itaconic acid (3).
In the case of sodium a-acetamidocinnamate (1a), the
silica was washed with distilled water (3×15 ml). The
combined filtrate and washing were evaporated to ca.
10 ml residue and acidified with 1 M HCl. The free acid
was extracted with ethyl acetate (3×15 ml) and the
extract was dried and evaporated. The conversion was
determined by integration of the CH3CON peaks at l
1.92 (CD3OD) for N-acetylphenylalanine (2) and l 2.1
(CD3OD) for a-acetamidocinnamic acid (1b).
The product from reactions of the free acid (1b) was
isolated by washing the silica with MeOH–H2O (1:1,
v/v) (3×15 ml). The combined washing and filtrate
were evaporated and the residue quickly dissolved in
0.5 M NaOH (5 ml) and filtered through a Celite pad.
The flask and Celite pad were washed with water
(2×20 ml). The combined filtrate and washings were
reduced to ca. 10 ml under vacuum and extracted with
ethyl acetate (2×20 ml). The aqueous layer was aci-
dified and the product isolated as above.
2.5. Preparation of sol–gel entrapped catalyst species
A modification of a two step procedure described
previously [2] was used. A rhodium catalyst solution
was prepared by stirring [Rh(COD)Cl]2 (10.85 mg,
0.022 mmol) with the ligand BPPM (24.92 mg, 0.044
mmol) in dry, deoxygenated THF (2.5 ml) for 30 min.
Aqueous NH4OH (0.44 ml of 0.1 M) and the prepared
catalyst solution were added simultaneously but sepa-
rately to a Schlenk flask containing a pre-stirred mix-
ture (15 min) of HCl (1.2 ml of a HCl solution, pH
1.96), (EtO)4Si (3.8 ml) and MeOH (3.5 ml). The mix-
ture was stirred for 5 min at ambient temperature.
Addition of Et3N (0.5 ml) resulted in gel formation in a
few minutes. The gel was dried, crushed, washed thor-
oughly with boiling CH2Cl2, sonicated for 30 min in

40
J. Jamis et al. / Journal of Organometallic Chemistry 627 (2001) 37–43
Table 1
Heterogeneous hydrogenation of sodium a-acetamidocinnamate (1a) in water a,b
c
Entry
Catalyst
Run
Temp. (°C)
Time (h)
Conv. (%)
Ee (%), (R)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Ru/NH2Si (26)
Ru/NH2Si (37)
Ru/PhSi (26)
Ru/PhSi (26)
Ru/PhSi (26)
Ru/PhSi (68)
Ru/PhSi (68)
Ru/PhSi (68)
Ru/PhSi (68)
0
0
0
1
2
0
1
2
0
1
2
0
0
1
2
0
1
2
0
1
2
0
1
2
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
70
60
60
60
60
60
60
72
72
72
72
72
72
72
72
22
22
22
72
72
72
72
22
22
22
72
72
72
72
72
72
21
100
99
99
99
99
99
99
97
95
87
17
100
97
97
89
80
67
49
44
41
40
43
42
39
42
42
40
Ru/PhSi (68)
Ru/PhSi (68)
Ru/Ph/NH2Si (26)
Ru/Ph/NH2Si (37)
Ru/Ph/NH2Si (37)
Ru/Ph/NH2Si (37)
Ru/Ph/NH2Si (37)
Ru/Ph/NH2Si (37)
Ru/Ph/NH2Si (37)
Ru/CO2HSi (37)
Ru/CO2HSi (37)
Ru/CO2HSi (37)
Ru/CO2NaSi (37) d
Ru/CO2NaSi (37) d
Ru/CO2NaSi (37) d
51
48
44
50
47
45
41
37
37
46
45
45
91
85
80
100
100
98
a Hydrogenation using 2.76 Mpa H2.
b All solutions were examined for metal content by ICP-AES and leaching was found to be negligible.
c Number of previous runs with the catalyst (reuse).
d CO2HSi (37) treated with NaHCO3 prior to incorporation of [Ru].
3. Results and discussion
a shorter time (22 h) gave incomplete conversion. Re-
covery of the catalyst and reuse under the shorter
reaction time conditions (entries 9–11) showed a small
decrease in conversion but no significant change in
enantioselectivity. External surface deactivation with
Ph2SiCl2 followed by derivatisation of the internal sur-
3.1. Reactions in6ol6ing catalysts prepared from
modified silicas
The results of hydrogenations of sodium acetami-
docinnamate (1a) in aqueous solutions using a ruthe-
nium catalyst incorporated into modified silicas are
summarised in Table 1 (Scheme 1).
The catalyst formed by incorporation of Ru–BI-
NAP(cymene) into the silica with average pore size 26
,
face with 3AMPS of the silicas 26 A, Ru/Ph/NH2Si
,
(26), and 37 A, Ru/Ph/NH2Si (37) gave, as expected,
results almost identical to those obtained where the
treatment with Ph2SiCl2 had been omitted (compare
entry 12 with entry 1 and entries 13–15 with entry 2).
Reactions using the Ru/Ph/NH2Si (37) catalyst system
were repeated using the shorter reaction time (22 h).
Again recovery and reuse led to some loss of conver-
sion but little change in enantioselectivity. The possibil-
ity that the amino function may have been protonated
and thus become unlikely to coordinate to a cationic
,
A which had been treated with 3AMPS (Ru/NH2Si
(26)) gave only a low conversion and no ee (entry 1). In
,
contrast similar modification of a 37 A silica, Ru/
NH2Si (37) led to complete conversion and a 49% ee
(entry 2) comparable with values using the same cata-
lyst and unmodified silica [1]. A reaction in water
carried out using a mixture of the unmodified silica Si
(37) together with the Ru–BINAP(cymene) catalyst
gave only 1% conversion, confirming the lack of cata-
lytic activity of non-incorporated ruthenium species.
,
External surface deactivation of both the 26 A, Ru/
,
PhSi (26) and 68 A, Ru/PhSi (68) silicas by treatment
with Ph2SiCl2 [4] led to systems which again gave
complete conversions, even after two recoveries and
reuse with ee’s all 39–44% (entries 3–8). Reactions for
Scheme 1.

J. Jamis et al. / Journal of Organometallic Chemistry 627 (2001) 37–43
41
3.3. Solid state 31P-NMR measurements
Table 2
Pore size and pore volumes of modified silicas
Solid state 31P-NMR spectra of Ru–BINAP(cymene)
and Ru/Si (37), where this catalyst had been incorpo-
rated into Si (37) gave spectra with identical major
peaks at 31 and 44 ppm. A further spectrum of the
Ru/Si (37) after it had been used and recovered from a
hydrogenation was recorded and gave three major
peaks at 31, 57 and 71 ppm. No peaks were observed in
the region around 0 ppm where free phosphine would
be expected to resonate. One of these peaks could be
due to adventitious BINAP oxide but its chemical shift
is almost coincident with the peak at 31 ppm assigned
to metal coordinated phosphorus. In addition IR could
not be used to establish the presence of phosphine
oxides due to the overlap of P–O and Si–O absorption
regions.
a
,
Entry Silica
Pore Size (A)
Pore volume (cm3
g−1
)
25
26
27
28
29
30
Si (37)
37
31
32
26
23
32
0.99
0.45
0.41
0.86
0.28
0.79
NH2Si (37)
Ph/NH2Si (37)
Si (26)
NH2Si (26)
CO2HSi (37)
a Indicates most frequent pore size in the distribution.
ruthenium species was considered. Thus alternative
tethers containing either a carboxylic acid function or
its carboxylate analogue were prepared, Ru/CO2HSi
(37) and Ru/CO2NaSi (37). Reactions using these cata-
lyst systems at a slightly lower temperature (60°C) all
gave conversions between 80–100 and ee’s 37–46%
even after two recoveries and reuse (entries 19–24).
3.4. Hydrogenations using catalyst systems with Ru or
Rh incorporated into silica gels
3.2. Pore 6olume measurements
A series of catalysts was prepared in which a ruthe-
nium or rhodium compound was added to te-
traethoxysilane and the resulting mixture converted
into a gel using the two step procedure described by
Gelman et al. [2]. It was found necessary to add some
triethylamine to induce gel formation in this system
based on (EtO)4Si in contrast to the literature prepara-
tion based on (MeO)4Si [2]. The resulting catalysts were
used for hydrogenations of aqueous solutions of
sodium acetamidocinnamate (1a) and itaconic acid (3)
and the results summarised in Table 3.
The amount added was such that the mol ratio of
metal:substrate was the same as in the experiments
described above. Similar reaction conditions were also
used to facilitate comparisons. Use of the catalyst based
on Ru–BINAP(cymene) initially gave a lower conver-
sion (69%) but a comparable ee (41%) (entry 31) to
those reported by us using preformed silicas as de-
scribed above and previously [1]. In contrast to these
results, recovery and reuse led to a decrease in conver-
sion and ee (entries 32 and 33). Two gels were prepared
in which 3AMPS was used in place of Et3N and some
Ph2SiCl2 was also added. It has been reported that use
of Ph2SiCl2 leads to the formation of gels in which the
external surfaces are capped [11]. Both catalysts how-
ever were virtually inactive (entries 34 and 35). Hydro-
genation of an aqueous solution of itaconic acid (3)
using the sol–gel system used in entries 31–33 gave
conversion (100%) and ee (15%) virtually identical to
that reported by us previously using a catalyst based on
preformed mesoporous silica [1]. A gel containing Rh–
BPPM was prepared as for the Ru–BINAP gel and
used as a catalyst for the hydrogenation of aqueous
solutions of sodium acetamidocinnamate (1a). The ini-
tial reaction of the cinnamate (1a) gave a low conver-
,
The pore volumes of the 26 and 37 A silicas were
measured [9] and compared with similar measurements
for some of the modified silicas and the results sum-
marised in Table 2.
Modification of Si (37) either by treatment with
3AMPS or by prior treatment with Ph2SiCl2 followed
by 3AMPS gave materials NH2Si (37) and Ph/NH2Si
(37) with pore volumes 0.45 and 0.41 cm3 g−1 (entries
25–27). Fig. 1 shows the distribution of pore size for Si
(37) and the modified materials NH2Si (37) and
CO2HSi (37). The peaks have been broadened and
shifted to a smaller pore size. Treatment of Si (26) with
3AMPS gave a material NH2Si (26) with a pore volume
of 0.28 cm3 g−1 (entries 28 and 29). The lack of
reactivity when NH2Si (26) was used as the matrix for
attempted hydrogenations involving Ru-BINAP
(cymene) as the catalyst system (see Table 1, entries 1
and 12) could therefore be attributed to difficulties in
incorporating a molecule of the catalyst and a molecule
of the substrate in the same cavity.
Introduction of the tether containing a carboxylic
function gave a material CO2HSi (37) with a pore
volume of 0.79 cm3 g−1 (entry 24) smaller than that of
Si (37) but significantly larger than that of both NH2Si
(37) and Ph/NH2Si (37). This difference could be due to
incorporation/retention of fewer tether groups. There
was considerable tailing in the pore size distribution
which could possibly be attributed to the hydrolysis
process. However, using this material as Ru/CO2HSi
(37) gave an almost identical conversion and ee to
reactions involving RuSi (37) and its amino-modified
analogues (see Table 1, entries 2 and 13–15).

42
J. Jamis et al. / Journal of Organometallic Chemistry 627 (2001) 37–43
Table 3
Heterogeneous hydrogenation of sodium a-acetamidocinnamate (1a) and itaconic acid (3) in water using sol–gel systems a
b
Entry
Substrate
Catalyst
Run
Time (h)
Temp. (°C)
Pressure (MPa)
Conv. (%)
Ee (%)
31
32
33
34
35
36
37
38
39
40
41
42
1a
1a
1a
1a
1a
3
1a
1a
1a
3
Ru–BINAP-(cymene) c
Ru–BINAP–(cymene) c
Ru–BINAP–(cymene) c
Ru–BINAP–(cymene) d
0
1
2
0
0
0
0
1
2
0
1
2
92
92
92
72
72
46
46
46
46
46
46
46
60
60
60
70
70
60
50
50
50
50
50
50
2.76
2.76
2.76
2.76
2.76
1
0.2
0.2
0.2
0.3
69
50
39
5
4
100
37
88
100
100
75
41
15
d
(Ru–BINAP)2
Ru–BINAP–(cymene) c
Rh–BPPM c
15
21
Rh–BPPM c
Rh–BPPM c
Rh–BPPM c
24
15
3
3
Rh–BPPM c
0.3
0.3
Rh–BPPM c
100
a All solutions were examined for metal content by ICP-AES and leaching was found to be negligible.
b Number of previous runs with the catalyst (reuse).
c Et3N used to form gel.
d 3AMPS used to form gel in presence of Ph2SiCl2.
catalyst activity could be improved by treatment with
hot aqueous sodium bicarbonate solution but this treat-
ment proved ineffective in our case (entry 46).
sion (37%) and ee (21%) (entry 37) compared to quanti-
tative conversion and ee (49%) when preformed silicas
were used [1]. Recovery and reuse led to increased
conversions (88 and 100%) but gave a racemic product
(entries 38 and 39), consistent with formation of highly
active but non-enantioselective catalyst which could be
small rhodium particles. Solid state 31P-NMR spectra
of recovered material (l 21.8 ppm) indicated that not
all of the Rh–P bonds had broken (l 24.8 ppm before
reaction). However it is not possible to establish
whether these Rh–P containing species are the active
species leading to racemic products. Hydrogenation of
itaconic acid (3) using this catalyst system initially gave
a quantitative conversion with 25% ee (entry 40), identi-
cal to the result obtained using a preformed silica. High
conversions were obtained for recovered and reused
catalyst but with an eventual total loss of enantioselec-
tivity (entries 41 and 42). A closely related reaction has
been reported to give a similar high conversion but
improved ee (78%) [2].
Some hydrogenations of a-acetamidocinnamic acid
(1b) in a methanol/water, 1/1 solvent system were car-
ried out using the Rh–BPPM containing gel and the
results summarised in Table 4. The initial reaction gave
a quantitative conversion with a reasonable ee (59%)
(entry 43), comparable with that obtained using a cata-
lyst involving apreformed silica (100% conversion, 80%
ee) [1]. However, use of the gel-based catalyst led to a
major loss of Rh (41%) and the recovered catalysts
showed decreasing reactivity and enantioselectivity (en-
tries 44 and 45).
3.5. Pore 6olumes of gel-based catalysts
The pore volume of the Rh–BPPM gel recovered
after use from the hydrogenation of sodium acetami-
docinnamate (1a) (Table 3, entry 37) was found to have
increased to a value of 0.37 cm3 g−1 from an initial
value of 0.21 cm3 g−1 presumably by a hydrolytic
process. These values are consistent with an increase in
conversion (compare entries 37 and 38). This, coupled
with the total loss of enantioselectivity, could be ex-
plained by the improved access of the substrate to
internal pores containing active catalyst species or pos-
sibly finely divided rhodium. The pore volume of the
gel incorporating Ru–BINAP(cymene) involving Et3N
was 0.45 cm3 g−1. In contrast, the gels obtained when
Et3N was replaced by 3AMPS for preparations involv-
ing both Ru–BINAP(cymene) and (Ru–BINAP)2 were
Table 4
Heterogeneous hydrogenation of a-acetamidocinnamic acid (1b) in
MeOH–H2O (1:1 v/v) using a sol–gel system a
b
Entry
Run
Conv. (%)
Ee (%)
Rh leached (%) c
43
44
45
46
0
1
100
79
57
8
59
47
26
41
Neg
Neg
Neg
2
3 d
a Hydrogenation at 50°C for 20 h using 0.2 MPa H2 and sol–gel
system containing Rh–BPPM, prepared using Et3N to form gel.
b Number of previous runs with the catalyst (reuse).
c Measured by ICP-AES.
These results contrast with the reported use of a
similar material for hydrogenations of itaconic acid (3)
with methanol/benzene, 3/1 as solvent where the ee’s
were maintained at 67% but the conversions decreased
with repeated use [2]. These authors also reported that
d The recovered and reused catalyst was treated with boiling 2%
aqueous NaHCO3 solution.

J. Jamis et al. / Journal of Organometallic Chemistry 627 (2001) 37–43
43
essentially non-porous. These results are consistent with
the lack of catalytic reactivity of the latter systems
(compare entires 34 and 35 with entry 31, Table 3).
Chemistry, Monash University (ICP-AES measure-
ments).
References
4. Conclusion
[1] J. Jamis, J.R. Anderson, R.S. Dickson, E.M. Campi, W.R.
Jackson, J. Organomet. Chem. 603 (2000) 80.
[2] F. Gelman, D. Avnir, H. Schumann, J. Blum, J. Mol. Catal. A:
Chem. 146 (1999) 123.
[3] C.-J. Liu, S.-G. Li, W.-Q. Pang, Chi.-Ming Che, Chem. Com-
mun. (1997) 65.
The hydrogenation catalysts obtained by incorporat-
ing Ru or Rh species into mesoporous silicas which
have been modified by external surface deactivation
and/or internal derivatisation show no improvement
over catalysts prepared using the non-modified meso-
porous silicas when there is an effective match between
pore size and catalyst dimension. The results vindicate
the approach to developing recoverable and reusable
heterogeneous catalysts by incorporating commercially
available homogeneous catalysts in non-modified
silicas.
[4] (a) A. Bleloch, B.F.G. Johnson, S.V. Ley, A.J. Price, D.S.
Shephard, A.W. Thomas, Chem. Commun. (1999) 1907. (b) D.S.
Shephard, W. Zhou, T. Maschmeyer, J.M. Matters, C.L. Roper,
S. Parsons, B.F. Johnson, M.J. Duer, Angew. Chem. Int. Ed.
Engl. 37 (1998) 2719. (c) J. Liu, X. Feng. G.E. Fryzell, L.-Q.
Wang, A.Y. Kim, M. Gong, Adv. Mater. 10 (1998) 161.
[5] (a) J. Macquarrie, D.B. Jackson, J.E.G. Mdoe, J.H. Clark, New
J. Chem. 23 (1999) 539. (b) J.A. Elings, R. Ait-Meddour, J.H.
Clark, D.J. Macquarrie, Chem. Commun. (1998) 2707. (c) J.H.
Clark, A.J. Butterworth, S.J. Tavener, A.J. Teasdale, S.J. Bar-
low, T.W. Bastock, K. Martin, J. Chem. Tech. Biotechnol. 68
(1997) 367.
[6] F. Knoop, J.G. Blanco, Hoppe-Seyler’s Z. Phys. Chem. 146
(1925) 272.
Acknowledgements
[7] (a) J.P. Genet, C. Pinel, V. Ratovelomana-Vidal, S. Mallart., X.
Pfister, L. Bischoft, M.C. Cano De Andrade, S. Darses, C.
Galopin, J. Laffittee, Tetrahedron: Asymmetry 5 (1994) 675. (b)
E. Berner, R. Leonardsen, Liebigs Ann. Chem. 538 (1939) 1.
[8] P.T. Tanev, T.J. Pinnavaia, Science 267 (1995) 865.
[9] P.A. Webb, C. Orr, Analytical Methods in Fine Particle Tech-
nology, Micrometrics Instrument Corporation, Norcross, GA,
USA, 1997, p. 53.
This project has been supported by funds from the
Australian Research Council and we thank Johnson
Matthey Pty Ltd. for loan of precious metals. We
also thank Dr F. Separovic, School of Chemistry,
University of Melbourne (solid state NMR spectra),
John Barnard, Department of Chemical Engineering,
[10] R. Mokaya, W. Zhou, W. Jones, J. Mater. Chem. 10 (2000)
1139.
Monash
University
(pore
size/pore
volume
measurements) and Dr R. Hodges, School of Applied
[11] D.J. Macquarrie, Green Chem. 1 (1999) 195.
.