Hybrid bidentate ligand for functional recognition: An application to regioselective C=C double bond hydrogenation

Article abstract of DOI:10.1039/b600127k

Regioselectivity increases in C=C double bond hydrogenation could be obtained for Lewis basic substrates on a Lewis acidic support by using a rhodium complex supported on a mesoporous solid. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b600127k

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Hybrid bidentate ligand for functional recognition: an application to
regioselective CLC double bond hydrogenation
Fre´de´ric Goettmann,a Pascal Le Floch*b and Cle´ment Sanchez*a
Received (in Cambridge, UK) 5th January 2006, Accepted 20th March 2006
First published as an Advance Article on the web 6th April 2006
DOI: 10.1039/b600127k
Regioselectivity increases in CLC double bond hydrogenation
could be obtained for Lewis basic substrates on a Lewis acidic
support by using a rhodium complex supported on a
mesoporous solid.
Natural enzymes are more active and selective than any man made
system.1 One of the most fascinating properties of enzymes is their
ability to orientate their substrate in order to discriminate similar
active functions. For instance, the well-known Cytochrome
P-450CAM is able to very selectively hydroxylate the 5-exo position
of camphor. Here, regioselectivity relies on the binding of camphor
to the OH group of tyrosine 96.2 To transpose this concept
to homogeneous catalysis, chemists have used van der Waals
forces,3–5 hydrogen bonding6,7 or Lewis acid/base interactions,8,9
Fig. 1 Schematic representation of the concept of functional recognition,
showing that designing artificial mimics of enzyme sites requires
fine molecular engineering. In the case of heterogeneous catalysts
(with the intention of overcoming the usual separation problems of
homogeneous systems10), such a strategy usually requires multi-
step syntheses.11,12 Herein, we wish to report on the use of a hybrid
meso-organized catalytic system, based on a rhodium complex
grafted onto mesoporous mixed zirconia/silica powders, in the
catalytic hydrogenation of CLC double bonds. This material
proved to be more selective than its homogenous counterpart. The
Lewis acidic sites due to the zirconium atoms could positively
interact with Lewis basic substrates, thus allowing discrimination
between the double bonds in the substrate skeleton.
based on the support’s wall properties. The substrate binds anywhere on
the support, then migrates and meets the active site, where one of its
reactive functions is better placed to react.
rhodium(I) precursor, to be very active in the hydrogenation and
hydroformylation of olefins.
We supposed that this system could be an ideal support for
testing our hypothesis. Indeed, as previously shown, the position of
the grafted rhodium complex near the wall of the support is very
well controlled. On the other hand, the Lewis acidic properties of
zirconia are well known, and the naked ZS20C material has
recently been proved to act as an efficient hydroformylation
catalyst, partly through its Lewis acidic binding sites.15 We
therefore investigated the use of this HBL in the regioselective
hydrogenation of olefins. Note that the structure of the complex
employed has been studied previously.13 One could suggest the
formation of a Zr–O–Rh bond as an alternative to the structure
proposed in Scheme 1. However, solid state 31P NMR data
supported the formation of the depicted 5-membered metallacycle.
Moreover, the presence of one available Lewis acidic site near the
rhodium center was established during this study; the phosphorus
The main problem in the synthesis of a heterogeneous
equivalent to enzymatic active sites is controlling the active and
the binding site positions. To circumvent this obstacle, we propose
to use the whole support surface as the binding function; the
substrate bearing a specific function presenting a good affinity for
the wall. Indeed, one could imagine that after grafting onto the
surface, the substrate would be able to migrate until it meets a
catalytic site, providing that available neighbouring sites exist.
(Fig. 1)
We recently described hybrid catalytic systems called hybrid
bidentate ligands (HBL),13 in which the active complex is grafted
as closely as possible to the surface.14 A phosphanorbornadiene
phosphonic acid derivative (1-phospha-4,5-dimethyl-3,6-diphenyl-
norbornadienyl phosphonic acid PNBDP) grafted onto a zirconia-
rich mesoporous material (ZS20C) proved, once complexed to a
aLaboratoire de Chimie de la Matie`re Condense´e de Paris, UMR CNRS
7574, Universite´ Pierre et Marie Curie, 4 place Jussieu, 75005 Paris,
France. E-mail: clems@ccr.jussieu.fr; Fax: (+33) 1-44-27-47-69
bLaboratoire He´te´roe´le´ments et Coordination, UMR CNRS 7653
(DCPH), De´partement de Chimie, Ecole Polytechnique,
91128 Palaiseau cedex, France. E-mail: lefloch@poly.polytechnique.fr;
Fax: (+33) 1-69-33-39-90
Scheme 1 Structure of the catalytic site and possible binding of a
substrate.
2036 | Chem. Commun., 2006, 2036–2038
This journal is ß The Royal Society of Chemistry 2006

View Article Online
Fig. 2 Structure of the substrates. The double bonds are labelled A or B
depending one the distance to the Lewis basic function. The mono-
hydrogenation products are hereafter labelled as A or B products when
they are hydrogenated in the A or B positions, respectively.
Fig. 3 Representation of the two possible paths for the coordination of
geraniol to the rhodium center in the presence of methanol. In the left
diagram, a methoxy moiety is coordinated to the zirconium center so that
bond-B of geraniol can enter the rhodium coordination sphere.
atom of the phosphine moiety being coordinated through its lone
pair to a zirconium atom prior to coordination to the [Rh(CO)2]
fragment. The presence of this neighboring anchoring site could
allow a substrate bearing a Lewis basic moiety to anchor itself near
the rhodium complexes and thus control the regioselectivity, as
depicted in Scheme 1 for geraniol.
that the binding of the ester moiety to the surface efficiently
protects the double bond in the A-position, preventing its
coordination to rhodium. One could of course suggest that
the observed selectivity doesn’t rely on the coordination of the
substrate to the zirconium oxide surface but instead on the
increased steric hindrance around the metallic centre due to
the grafting of the complex. Indeed, in substrate 1, double bond B
is the least hindered and the most hydrogenated. However, the
case of geraniol (substrate 2) contradicts this assumption. We
found that the heterogeneous catalyst is much more active than the
homogeneous one (5% conversion, Table 1, entry 6),{ while
attempts to increase regioselectivities by enhancing the steric
hindrance usually resulted in activity losses. Moreover, the
observed regioselectivities strongly depend on the experimental
conditions. In methanol, the reaction is totally selective towards
the A-product at 60% conversion. Increasing the reaction time
results in hydrogenation of the second double bond (B-product)
but no traces of the B-compound could be detected. This indicates
that hydrogenation of bond A is much faster than that of bond B,
even if neither of them is clearly less sterically hindered. An
explanation of this behaviour could be that geraniol and methanol
compete to complex the zirconium surface, since both are alcohols.
It would therefore be possible for a geraniol molecule to enter the
coordination sphere of the rhodium centre without being bonded
to the surface, while a methoxy moiety would occupy the
molecular recognition position on the zirconium centre (Fig. 3).
However, hydrogenation of a geraniol molecule that is not bonded
to the surface seems to be far less favourable, as at 60%
The catalyst, ([Rh(PNBDP)(CO)2]@ZS20C), was prepared as
previously described.13 The mixed oxide was synthesized by spray
drying an ethanolic sol of ZrCl4 and SiCl4 (molar ratio of 4 : 1)
with cetyltrimethylammonium bromide (CTAB) as surfactant.
After surfactant elimination, the powder exhibited a 230 m2 g21
specific surface area and an average pore diameter of about 20 s.
PNBDP was grafted under argon in toluene at 90 uC for 12 h,
resulting in a 20 mass% organic charge. Solid state 31P CP-MAS
NMR and FT-IR confirmed the proposed structure of the
catalyst.13 Two model substrates, bearing a Lewis basic function
like an ester or an alcohol, were chosen: methyl bicyclo[2.2.2]octa-
2,5-diene-2-carboxylate (1) and geraniol (2) (Fig. 2).
Hydrogenations were carried out in methanol at 50 uC under
10 bar of H2 using 0.5 mol% of rhodium. The products were
analyzed by 1H NMR, and the results are summarized in Table 1.
When required, the corresponding homogeneous catalyst,
[Rh(PNBDP)(CO)2], was used for comparison.
As can be seen, the hydrogenation reaction is totally
regioselective in the case of compound 1 (Table 1, entry 1),
whereas the homogeneous catalyst only yields 80% of the same
product (Table 1, entry 2). Indeed, bicyclodienes like barreledienes
or norbornadiene are known to act as bidentate ligands towards
rhodium(I) fragments.16,17 Therefore, both double bonds can be
hydrogenated. With the heterogeneous catalyst, one can propose
Table 1 Catalytic hydrogenation of methyl bicyclo[2.2.2]octa-2,5-diene-2-carboxylate (1) and geraniol (2)a
Products (%)
Monohydrogenation
Entry
Substrate
Catalyst
Reaction time/h
Conversion (%) A product
B product
Dihydrogenation
1
2
3
4
5
6
a
1
1
2
Heterogeneous
Homogeneous
Heterogeneous
Heterogeneous
Heterogeneous
Homogeneous
24
24
24
72
72
72
100
100
60
100
100
5
0
10
100
30
70
50
100
80
0
0
0
0
10
0
70
30
50
2
2b
2
0
The catalysts used were: [Rh(PNBDP)(CO)2]@ZS20C (referred to as heterogeneous) and [Rh(PNBDP)(CO)2] (referred to as homogeneous).
For monohydrogenations, product A refers to that in which the double bond located nearest to the Lewis basic group was hydrogenated and
B to the other one. The solvent used was toluene.
b
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 2036–2038 | 2037

View Article Online
conversion, only A-product is detected. It is only when almost all
the A-bonds have reacted that hydrogenation of the B-bonds
starts, yielding dihydrogenation products. In order to confirm this
hypothesis, we decided to employ a solvent that has a much lower
affinity for Lewis acids than alcohols, such as toluene. As expected,
a much higher selectivity towards the A-product (70%) was
obtained at 100% conversion (Table 1, entry 5). As geraniol is in
large excess compared with the number of the Lewis acidic binding
sites, it is a likely assumption that all of these sites are occupied
either by a geraniol molecule or a hydrogenated molecule. This
strongly hinders a non-grafted molecule from entering the rhodium
coordination sphere, even at 100% conversion; thereby accounting
for the observed regioselectivity.
The CNRS, the Universite´ Pierre et Marie Curie, Ecole
Polytechnique and Saint-Gobain Recherche are gratefully
acknowledged for financial support.
Notes and references
{ The activity of the homogeneous complex in the case of geraniol is
remarkably low. This feature is however consistent with previous reports
on the poor activity of Rh–PNBDP complexes in the hydrogenation of
protic substrates.19
1 A. Corma, Catal. Rev. Sci. Eng., 2004, 46, 369.
2 B. Meunier, S. P. de Visser and S. Shaik, Chem. Rev., 2004, 104, 3947.
3 R. Breslow and B. Zhang, J. Am. Chem. Soc., 1994, 116, 7893.
4 R. Breslow, Acc. Chem. Res., 1995, 28, 146.
This work shows that the use of a HBL in rhodium-catalyzed
CLC double bond hydrogenation results in an increased regios-
electivity compared with the corresponding homogeneous equiva-
lent. This increase in selectivity, together with the increased
activity, could rely on the complexation of the Lewis basic
substrate onto the Lewis acidic zirconia wall. To fully validate the
proposed mechanism, a full in situ FT-IR study might allow the
detection of the binding of the substrate to the support. A further
step would be to find substrates where the bond that is ideally
placed for reacting is, simultaneously, the most difficult to reduce,
in order to favour unexpected regioselectivities. Direct conversion
of citral to citronellol would exemplify this new concept. We also
believe that this simple concept could be extended to other
catalytic processes of synthetic relevance. Importantly, applying
this strategy to synthesizing molecular imprinted devices18 would
probably strengthen the interaction between the substrates and the
cavities, improving the level of molecular control.
5 R. Breslow and S. D. Dong, Chem. Rev., 1998, 98, 1997.
6 R. H. Crabtree, J. A. Loch, K. Gruet, D.-H. Lee and C. Borgmann,
J. Organomet. Chem., 2000, 600, 7.
7 G. Rivera and R. H. Crabtree, J. Mol. Catal. A: Chem., 2004, 222, 59.
8 G. J. Rowlands, Tetrahedron, 2001, 57, 1865.
9 Y. Murakami, J.-I. Kikuchi, Y. Hisaeda and O. Hayashida, Chem. Rev.,
1996, 96, 721.
10 D. J. Cole-Hamilton, Science, 2003, 299, 1702.
11 V. Dufaud and M. E. Davis, J. Am. Chem. Soc., 2003, 125, 9403.
12 A. Katz and M. E. Davis, Nature, 2000, 403, 286.
13 F. Goettmann, C. Boissie`re, D. Grosso, F. Mercier, P. Le Floch and
C. Sanchez, Chem.–Eur. J., 2005, 11, 7416.
14 F. Goettmann, D. Grosso, F. Mercier, F. Mathey and C. Sanchez,
Chem. Commun., 2004, 1240.
15 F. Goettmann, P. Le Floch and C. Sanchez, Chem. Commun., 2006, 180.
16 Y. Otomaru, K. Okamoto, R. Shintani and T. Hayashi, J. Org. Chem.,
2005, 70, 2503.
17 R. Shintani, K. Ueyama, I. Yamada and T. Hayashi, Org. Lett., 2004, 6,
3425.
18 M. E. Davis, A. Katz and W. R. Ahmad, Chem. Mater., 1996, 8, 1820.
19 S. Lelievre, F. Mercier and F. Mathey, J. Org. Chem., 1996, 61, 3531.
2038 | Chem. Commun., 2006, 2036–2038
This journal is ß The Royal Society of Chemistry 2006