Solvent-free, low-temperature, selective hydrogenation of polyenes using a bimetallic nanoparticle Ru-Sn catalyst

Article abstract of DOI:10.1002/1521-3773(20010401)40:7<1211::AID-ANIE1211>3.0.CO;2-P

The point of attachment of bimetallic Ru₆Sn particles which are anchored to the pore walls of a highly dispersed high-area mesoporous silica is found to be the tin atom, as indicated by in situ and ex situ measurements. This catalyst displays high activity for the low-temperature, selective hydrogenation of cyclic polyenes under solvent-free conditions (see scheme).

Full text of DOI:10.1002/1521-3773(20010401)40:7<1211::AID-ANIE1211>3.0.CO;2-P

COMMUNICATIONS
[14] D. A. Huse, S. Leibler, J. Phys. 1988, 49, 605.
[15] J. Allgaier, A. Poppe, L. Willner, D. Richter, Macromolecules 1997, 30,
1582.
[16] A. C. Finnefrock, G. E. S. Toombes, S. M. Gruner, R. Ulrich, U.
Wiesner, unpublished results.
[17] S. Brunauer, L. S. Deming, W. S. Deming, E. Teller, J. Am. Chem. Soc.
1940, 62, 1723.
[18] E. P. Barrett, L. G. Joyner, P. P. Halenda, J. Am. Chem. Soc. 1951, 73,
373.
[19] S. J. Gregg, K. S. W. Sing, Adsorption, Surface and Porosity, 2nd ed.,
Academic Press, 1982.
[20] J. M. Thomas, Angew. Chem. 1999, 111, 3800 ± 3843; Angew. Chem.
Int. Ed. 1999, 38, 3588 ± 3628.
[21] P. Ström, D. M. Anderson, Langmuir 1992, 8, 691.
[22] M. W. Tate, S. M. Gruner, E. F. Eikenberry, Rev. Sci. Instrum. 1997, 68,
47.
the silica host. Their atomic structure is straightforwardly
established,[1, 6±8] from in situ X-ray absorption and FTIR
measurements.
Tin has been widely used as a ªpromoterº in heterogeneous
catalysis, and has been shown to increase dramatically the
selectivity of ruthenium catalysts in a variety of chemical
transformations. These usually involve the selective hydro-
genation of a carbonyl group in the vicinity of a conjugated or
isolated double bond, such as the hydrogenation of benzoic
acid and fatty esters to their corresponding alcohols[9, 10] or the
hydrogenation of acrolein and its derivatives.[11] It has been
suggested that the effect of tin is, besides others, to deactivate
selectively certain catalytic sites on the surface thereby
impeding undesirable side reactions.[10, 12] Other explanations,
such as alloying and its consequential electronic influences,
have also been proposed.[11]
The chemical conversions upon which we focus herein
consist of the hydrogenation of cyclic polyenes: 1,5,9-cyclo-
dodecatriene, 1,5-cyclooctadiene, and 2,5-norbornadiene. The
monoenes of all three polyenes are used extensively as
intermediates in the synthesis of bicarboxylic aliphatic acids,
ketones, cyclic alcohols, lactones, and other intermediates.
The selective hydrogenation of 1,5,9-cyclododecatriene to
cyclododecane and cyclododecene is industrially important in
the synthesis of valuable organic and polymer intermediates,
such as 12-laurolactam and dodecanedioic acid,[13] which are
important monomers for nylon 12, nylon 612, copolyamides,
polyesters, and coating applications.
A wide variety of homogeneous and heterogeneous hydro-
genation catalysts such as Raney nickel,[14] palladium,[15]
platinum,[16] cobalt,[17] and mixed transition-metal complexes
have been previously used for the above-mentioned hydro-
genations.[18, 19] But all the reactions entailed the use of
organic solvents (such as n-heptane, benzonitrile, and so
forth),[13, 20±22] and some required utilization of efficient
hydrogen donors such as 9,10-dihydroanthracene, often at
temperatures in excess of 3008C, to achieve the desired
selectivities.[23] Recently, Reetz et al.[24] have shown that
entrapment of palladium clusters in micro/mesoporous hydro-
phobic sol ± gel matrices prevents undesired agglomeration of
the clusters, to result in active catalysts for the hydrogenation
of 1,5-cyclooctadiene.
Solvent-Free, Low-Temperature, Selective
Hydrogenation of Polyenes using a Bimetallic
Nanoparticle Ru ± Sn Catalyst**
Sophie Hermans, Robert Raja, John M. Thomas,*
Brian F. G. Johnson,* Gopinathan Sankar, and
David Gleeson
Progress in making solvent-free chemical conversions much
more feasible than they are at present awaits the development
of highly active (and selective) heterogeneous catalysts. Not
only must the sites at which turnover occurs be of high
intrinsic activity, their concentration (per unit mass) must also
be large. Moreover, diffusion of reactant species to, and of
products away from, such sites must also be facile.[1]
Herein we demonstrate how freely accessible active sites on
bimetallic nanoparticles, finely dispersed and firmly anchored
1
along the interior surfaces of high area (around 800 m2 g )
mesoporous silica function as highly effective catalysts for a
number of selective hydrogenations that are of considerable
significance for chemical and fine-chemical production. Such
heterogeneous nanocatalysts may be readily prepared[2±5]
from mixed-metal carbonylates and introduced in a spatially
uniform fashion along the pores (around 30 Š diameter) of
Previously, we have described the catalytic performance of
other, finely dispersed bimetallic catalysts (Ag ± Ru,[2] Cu ±
Ru,[3] and Pd ± Ru,[4]), some of which possessed remarkable
properties. Herein we report the catalytic performance of a
supported carbidic Ru6Sn nanoparticle and compare it with
that of the above-mentioned bimetallic nanocatalysts. In
particular, we highlight the changing selectivity (in the
hydrogenation of 1,5,9-cyclododecatriene and other polyenes)
as a function of operating temperature and degree of
conversion.
The compound (PPN)[Ru6C(CO)16SnCl3] (PPN: bis(triphe-
nylphosphane)iminium cation) is obtained in a two step
reaction from the known carbido ± hexaruthenium cluster
[Ru6C(CO)17]. Details of the preparation along with
the single-crystal X-ray structure of the anion
[Ru6C(CO)16SnCl3] (1), have been given elsewhere,[25] as
have the corresponding details for the neutral species
[*] Prof. Sir J. M. Thomas, Dr. G. Sankar, Dr. D. Gleeson
The Royal Institution of Great Britain
Davy Faraday Research Laboratory
21 Albemarle Street, London W1X 4BS (UK)
Fax : (‡44)0207-670-2988
E-mail: dawn@ri.ac.uk
Prof. B. F. G. Johnson, S. Hermans, Dr. R. Raja
Department of Chemistry
University of Cambridge
Lensfield Road, Cambridge CB2 1EW (UK)
Fax : (‡44)1223-339016
E-mail: jpt25@cam.ac.uk
[**] We thank Dr. R. G. Bell for assistance with the computer graphics,
Drs. P. A. Midgeley, V. Keast, and M. Weyland for help with STEM,
and gratefully acknowledge the support (via a rolling grant to J.M.T.
and an award to B.F.G.J.) of EPSRC and the award of a research
fellowship (for G.S.) from the Leverhulme Foundation and a Marie
Curie Fellowship within the TMR Programme of the European
Commission (for S.H.).
Angew. Chem. Int. Ed. 2001, 40, No. 7
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4007-1211 $ 17.50+.50/0
1211

COMMUNICATIONS
Figure 1. Molecular structures of [Ru6C(CO)16SnCl3] 1 and [Ru6C(CO)16SnCl2] 2, as determined by X-ray crystallography.
[Ru6C(CO)16SnCl2] (2; see Figure 1). This latter compound
can be obtained by chloride abstraction from 1 or directly by
reacting [Ru6C(CO)17] with SnCl2. The mesoporous silica was
loaded with cluster 1 by making a slurry in ether/dichloro-
methane, stirring for 48 h under nitrogen, and filtering.[3] The
pink solid obtained was then washed with ether and dried
under vacuum. The same procedure was used for the
incorporation of cluster 2 within the mesopores. However, a
small amount of nonadsorbed compound was lost in the
filtrate, demonstrating the less favorable interaction of this
neutral species with the surface hydroxyl groups of MCM-41,
when compared to its charged analog 1.
The catalytic precursor material 1/MCM-41, which was
characterized by FTIR (using a purpose-built evacuable
cell[7]), revealed that the carbonyl cluster was still intact
within the pores. This was confirmed by extended x-ray
absorption fine structure (EXAFS) analysis, which showed
that the structure of the pure, unsupported carbonyl cluster is
exactly identical to that of the species adsorbed into the
mesoporous silica.
The frequency characteristics of the terminal carbonyl
groups disappeared totally at a heat-treatment temperature of
2008C. During this activation procedureÐand concomitant
with the loss of all carbonyl ligandsÐthe metallic core of the
cluster becomes firmly anchored to the siliceous surface
(Figure 2).
The EXAFS data, recorded in situ,[2, 3] showed significant
changes in the coordination environment of both the tin and
ruthenium centers (Figure 3). The best fit for the ruthenium
K-EXAFS data yielded a structure for the denuded bimetallic
catalyst consisting of mainly ruthenium neighbours. To obtain
a better fit for the low k data, it was necessary to include a
carbon neighbour. The Ru ± Ru and Ru ± C distances and the
coordination number of 2.67 and 2.07 Š, and 3 and 1,
Figure 2. Computer graphic model of Cu4Ru12C2 nanocatalyst clusters
anchored through pendant silanol groups to the inner walls of mesoporous
silica (MCM-41). This nanocatalyst is akin to Ru6Sn and other clusters
described in the text. A typical reactant (1,5,9-cyclododecatriene) and H2
are also shown in the mesopore channel (diameter around 30 Š).
respectively, suggest that the arrangement of Ru, C, and Sn
are quite similar to that of the starting material. Upon
removal of the carbonyl ligands the Ru ± Ru distances
decreased from an average value of around 2.8 to 2.67 Š. It
is not directly feasible to detect the Ru ± Sn interaction, since
only one of the six ruthenium centers has a close contact with
the tin atom. Similarly, from the Sn K-edge, two of the
chlorine neighbors are completely replaced by the oxygen
atoms, and the best fit is obtained by taking two Sn ± O, one
Sn ± Cl, and one Sn ± Ru interaction,[26] which shows that the
tin atom is the anchoring point of the bimetallic, nanoparticle
1212
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4007-1212 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2001, 40, No. 7

COMMUNICATIONS
Figure 3. Fourier transform of the EXAFS a) at the Ru K-edge and b) Sn
K-edge (of the [Ru6C(CO)16SnCl3] cluster anchored on MCM-41. c) The
single crystal-derived structure is broadly consistent with the experimental
EXAFS data. Fourier transforms of the d) Ru K-edge and e) Sn K-edge
EXAFS of the decarbonylated (by mild thermolysis at 1808C in vacuum for
1 h) catalysts. f) Local structure of the activated catalyst, derived from the
analysis of the EXAFS data. Solid and dashed lines in (a), (b), (d) and (e)
represent the experimental and calculated data, respectively.
Figure 4. High-angle, annular dark field electron micrographs of the
Ru6Sn/MCM-41 catalyst a) prior to and b) after catalytic use.
particles or breakdown of the MCM-41 matrix) occurs during
catalysis. This demonstrates both the robust nature of the
nanoparticle catalyst and its effective anchoring to the walls of
the high-area silica. The overall composition and the invar-
iance of the ratio of the two metals were deduced from
energy-dispersive X-ray analysis on both the as-prepared and
used specimens of catalyst. Consistent with the fact that no
Sn ± Sn bond were detected by EXAFS, this indicates that
each nanoparticle is derived from an intact cluster core made
of six ruthenium atoms, one carbon atom and a single atom of
tin.
The marked dependence of temperature upon selectivity in
the catalytic hydrogenation of 1,5,9-cyclododecatriene is
shown in Figure 5 , from which we note that, even as low as
808C, this solvent-free hydrogenation to cyclododecene is
70% selective and well in excess of 90% at 1008C. This
catalyst on the silica surface. Taking these results into account,
we propose a model for the active denuded cluster that is
shown in Figure 3 f. We found that cluster 2 could not be freed
completely from its residual chlorine ligands; in fact, the Ru
K-edge EXAFS data could be refined only on the assumption
that a residual chlorine atom was attached to the ruthenium
centers. Furthermore, this denuded material derived from
cluster 2 is not catalytically active.[27]
High-angle, annular dark-field (HAADF) scanning trans-
mission electron microscopy (STEM),[1, 5] shows (Figure 4)
how well the nanoparticle catalysts are distributed within the
mesopores of the silica support both prior to and after
catalytic testing (see below). No dramatic change in the
structure of the material (such as sintering of the nano-
Angew. Chem. Int. Ed. 2001, 40, No. 7
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4007-1213 $ 17.50+.50/0
1213

COMMUNICATIONS
or norbornane. It is also evident from
Table 1 that bimetallic nanoparticle cata-
lysts are far superior in performance to
their monometallic analogs and more im-
portantly yield a higher selectivity for
hydrogenated products, which shows
a
synergism between the two components of
the bimetallic nanoparticle. We established
in separate experiments that monometallic
Ru6 and Pd anchored on MCM-41 were
inactive for the above-mentioned reactions.
Considerable scope clearly exists, with the
principles that we have outlined here and
previously to develop other, solvent-free
systems for the selective conversion of
organic compounds.
We have demonstrated how to regulate
the selectivity of nanocatalysts for solvent-
free hydrogenations by changing the nature
of the constituents in its active phase.
&
Figure 5. The effect of temperature on the hydrogenation conversion ( ) of the prodcut
~
!
*
selectivity is marked in terms of singly ( ), doubly ( ), and fully ( ) hydrogenated products of
1,5,9-cyclododecatriene with the Ru6Sn catalyst.
particular selective hydrogenation falls off in efficiency with
further increase in temperature, and conversion to the fully
hydrogenated cyclododecane is progressively favored. It is
noteworthy that conversion to the cyclododecadiene on the
other hand is favored at the lowest test temperature (808C).
By contrast, the Pd6Ru6 catalyst reported earlier[4] displays
very little selectivity for the 1,5-cyclododecadiene or cyclo-
dodecene, even at temperatures as low as 408C. By compar-
ison with three other bimetallic catalysts (Cu4Ru12, Ag4Ru12
and Pd6Ru6), we note that precursor 1 yields the bimetallic
catalyst (Ru6Sn) with the best performance, so far as selective
hydrogenation of one of the two double bonds in the
hydrogenation of 1,5-cyclooctadiene is concerned (Figure 6).
Experimental Section
Electron Microscopy: The images were obtained on a VG HB501 field-
emission STEM microscope. A small amount of activated catalyst was
ground and suspended in hexane by ultrasonic treatment. A holey carbon
film supported on a copper grid was dipped into the solution and allowed to
dry in the air before introduction to the microscope. The sample dispersed
on the grid was further dried inside the preparation chamber of the
microscope by heating at ca. 1008C under a light bulb and then was left
under high vacuum overnight before acquiring the images.
EXAFS: Ru and Sn K-edge EXAFS data were collected at station 9.2 at
the Daresbury Synchrotron Laboratory, which operates at 2 GeV and a
typical current in the range of 150 to 250 mA. This station is equipped with
a quick scanning Si(220) double-crystal monochromator, ion chambers for
measuring incident (I0), transmitted (It) beam intensities, and a 13 element
Canberra fluorescence detector. In a typical experiment 150 mg of the
sample was pressed into a self-supporting wafer and loaded in to the in situ
cell.[28] The cell was evacuated to around 4 Â 10 3 Torr and the catalyst was
1
heated to 1808C at 5Kmin and retained at this temperature for around
1 h before cooling it to room temperature. All the EXAFS experiments
were performed in the fluorescence mode and 10 scans were taken to
improve the signal to noise ratio. The data processing was done using the
suite of programs available at Daresbury, namely, EXCALIB, EXBROOK,
and EXCURV98. The summed data were analysed without Fourier
filtering and a full multiple scattering procedure was employed to analyse
all the data sets, in particular when carbonyl species and second silicon
neighbors were included in the analysis.
Catalysis: The catalytic testing was carried out in a high-pressure stainless
steel reactor (Cambridge Reactor Design) lined with PEEK (polyether-
etherketone). Nanoparticle bimetallic catalyst (20 mg) was activated
(473 K, 2 h, in the presence of hydrogen (0.5 MPa) prior to the hydro-
genation reaction. The reactor was then depressurised and cooled to room
temperature, before introducing the substrate (50 g;see Table 1) and
internal standard (mesitylene, 2.5 g). The vessel was pressurised with
hydrogen (30 bar) and heated to the desired temperature with continued
stirring (400 min 1). During the reaction, small aliquots were removed,
using a mini-robot autosampler, to enable the kinetics to be studied. The
products of the reaction were analysed with gas chromatograpy (GC,
Varian, Model 3400 CX) employing a HP-1 capillary column (25 m Â
0.32 mm) and flame ionisation detector. The identity of the products was
confirmed by LC-MS (Shimadzu, QP 8000).
Figure 6. Comparison of catalytic performance and degree of selectivity
for the hydrogenation of 1,5-cyclooctadiene, after 24 h, at 353 K. (See
Table 1 for reaction conditions).
Further scope for fine-tuning both the conversion and
selectivity is obtained by varying the contact times, for a range
of polyenes (Table 1). It is particularly interesting that, in the
case of 2,5-norbornadiene, the conditions (and catalyst) may
be so chosen as to produce a high yield of either norbornene
Both the Ru6Sn and Pd6Ru6 bimetallic nanoparticle anchored catalysts
were reused six or seven times for the hydrogenation of 1,5-cyclooctadiene
and 1,5,9-cyclododecatriene without appreciable loss in catalytic activity or
1214
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4007-1214 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2001, 40, No. 7

COMMUNICATIONS
Table 1. Hydrogenation of polyenesÐcomparison of catalysts.
1
Catalyst
Substrate
t [h]
T [K]
Conv. [mol%]
TOF[a] [h
]
Prod. dist. [mol%]
A
B
Ru6Sn/MCM-41
Cu4Ru12/MCM-41
Ag4Ru12/MCM-41
Pd6Ru6/MCM-41
1,5-cyclooctadiene
1,5-cyclooctadiene
1,5-cyclooctadiene
1,5-cyclooctadiene
8
24
8
24
8
24
8
24
24
24
353
353
353
353
11.7
72.5
11.5
42.7
9.0
24.4
36.9
91.4
2.5
1980
4090
690
1422
465
1316
2012
4965
141
100
86.9
±
12.8
29.2
31.0
42.5
49.6
84.5
96.5
36.2
70.4
68.8
57.3
50
15.7
3.6
±
Ru6/MCM-41
Pd/MCM-41
1,5-cyclooctadiene
1,5-cyclooctadiene
353
353
7.3
212
±
100
C
D
E
Ru6Sn/MCM-41
Ru6Sn/MCM-41
Ru6Sn/MCM-41
Ru6Sn/MCM-41
Pd6Ru6/MCM-41
Pd6Ru6/MCM-41
1,5,9-cyclododecatriene
1,5,9-cyclododecatriene
1,5,9-cyclododecatriene
1,5,9-cyclododecatriene
1,5,9-cyclododecatriene
1,5,9-cyclododecatriene
8
24
8
24
8
24
8
24
8
353
373
393
4.7
13.4
17.2
28.9
28.2
41.7
39.1
54.5
47.2
62.0
64.9
86.2
530
503
84.0
32.5
17.2
5.0
6.8
±
±
±
±
±
15.8
±
±
±
3.5
68.2
82.4
91.3
87.0
86.7
80.2
60.5
48.2
36.2
11.7
6.2
1940
1130
3280
1785
4400
2170
3320
1975
5350
3200
6.0
13.5
19.5
39.3
51.5
63.6
88.5
95.7
413
±
353
24
8
24
373
±
±
F
G
Ru6Sn/MCM-41
Pd6Ru6/MCM-41
2,5-norbornadiene
2,5-norbornadiene
8
24
8
333
333
51.4
89.3
76.4
98.7
10210
5960
11176
6855
88.6
46.8
24.7
2.5
11.3
53.3
75.1
97.6
24
[a] TOF ˆ [(molsubstr)(molcluster
)
1h 1]. Reaction conditions: substrate ꢀ 50 g; catalyst ˆ 25 mg; H2 pressure ˆ 30 bar; A ˆ cyclooctene; B ˆ cyclooctane; C ˆ
1,9-cyclododecadiene; D ˆ cyclododecene; E ˆ cyclododecane; F ˆ norbornene; G ˆ norbornane; .
selectivity. In addition, the following experiments were performed to rule
out the possibility of leaching:
[11] B. Coq, F. Figueras, C. Moreau, P. Moreau, M. Warawdekar, Catal.
Lett. 1993, 22, 189.
*
In a typical experiment (see Table 1 and Figure 6), the solid catalyst was
[12] G. Neri, R. Pietropaolo, S. Galvagno, C. Milone, J. Schwank, J. Chem.
Soc. Faraday Trans. 1994, 90, 2803.
[13] J. Hanika, I. Svoboda, V. Ruzicka, Coll. Czech. Chem. Commun. 1981,
46, 1039.
[14] Y. Kurokawa, UPE Industries Ltd, Jap. Patent 6,274,938, 1987.
[15] G. Strukul, G. Carturan, Inorg. Chim. Acta 1979, 35, 99.
[16] C. A. Brown, H. C. Brown, J. Org. Chem. 1966, 31, 3989.
[17] R. E. Harmon, S. K. Gupta, D. J. Brown, Chem. Rev. 1973, 73, 21.
[18] BASF, Ger. Patent 1,230,790.
filtered from the reaction mixture (when hot) after 8 h, and the reaction
was continued with the resulting filtrate for a further 16 h. No significant
change in conversion or product selectivity was observed, indicating that
the metal ions leached out (if any) are not responsible for the observed
activity and selectivity.
The resulting filtrate (at the end of the reaction, after 24 h) was
independently analyzed by ICP and AAS for free or dissolved metal (Sn,
Pd, Ru) ions, and only trace amounts (<3 ppb) were detected.
*
[19] BASF, UK Patent 826,832.
[20] C. Ruchardt, M. Gerst, M. Nolke, Angew. Chem. 1992, 104, 1516;
Angew. Chem. Int. Ed. Engl. 1992, 31, 1523.
Received: November 20, 2000 [Z16135]
[21] C. Hofler, C. Ruchardt, Liebigs Ann. 1995, 183.
[22] M. Gerst, C. Ruchardt, Tetrahedron Lett. 1993, 34, 7733.
[23] J. Morgenthaler, C. Ruchardt, Liebigs Ann. 1996, 1529.
[24] M. T. Reetz, M. Dugal, Catal. Lett. 1999, 58, 207.
[25] S. Hermans, B. F. G. Johnson, Chem. Commun. 2000, 1955.
[26] We endeavoured, but failed, to detect this chlorine atom(present in
trace amounts) by electron-stimulated X-ray emission, using an
energy-dispersive analyzer inside our scanning electron microscope,
due to overlap with strong peaks corresponding to ruthenium.
[27] A more detailed structure analysis of this material, which is not
discussed further here, will be published elsewhere.
[28] P. A. Barrett, G. Sankar, C. R. A. Catlow, J. M. Thomas, J. Phys.
Chem. A 1996, 100, 8977.
[1] J. M. Thomas, Angew. Chem. 1999, 111, 3800; Angew. Chem. Int. Ed.
1999, 38, 3589.
[2] D. S. Shephard, T. Maschmeyer, B. F. G. Johnson, J. M. Thomas, G.
Sankar, D. Ozkaya, W. Z. Zhou, R. D. Oldroyd, R. G. Bell, Angew.
Chem. 1997, 109, 2337; Angew. Chem. Int. Ed. Engl. 1997, 36, 2242.
[3] D. S. Shephard, T. Maschmeyer, G. Sankar, J. M. Thomas, D. Ozkaya,
B. F. G. Johnson, R. Raja, R. D. Oldroyd, R. G. Bell, Chem. Eur. J.
1998, 4, 1214.
[4] R. Raja, S. Hermans, D. S. Shephard, S. Bromley, J. M. Thomas,
B. F. G. Johnson, T. Maschmeyer, Chem. Commun. 1999, 2131.
[5] D. Ozkaya, W. Z. Zhou, J. M. Thomas, P. Midgeley, V. J. Keast, S.
Hermans, Catal. Lett. 1999, 60, 113.
[6] J. M. Thomas, Chem. Eur. J. 1997, 3, 1557.
[7] G. Sankar, J. M. Thomas, Top. Catal. 1999, 8, 1.
[8] J. M. Thomas, G. Sankar, J. Synchrotron Radiat. 2000, in press.
[9] K. Ishii, F. Mizukami, S. Niwa, R. Kutsuzawa, M. Toba, F. Fujii, Catal.
Lett. 1998, 52, 49.
[10] V. M. Deshpande, K. Ramnarayan, C. S. Narasimhan, J. Catal. 1990,
121, 174.
Angew. Chem. Int. Ed. 2001, 40, No. 7
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4007-1215 $ 17.50+.50/0
1215