Carbohydrates as building blocks of “privilegiad” ligands
for multifasic asymmetric catalysis
Vincenzo Benessere, Antonella De Roma and Francesco Ruffo
Università di Napoli “Federico II”, Italy (e-mail: [email protected])
Homogeneous enantioselective catalysis is a fundamental technology for the modern production of
fine chemicals.1 Two aspects are important for its extension to large scale production: (i) accurate
design of the precious metal catalyst and (ii) its efficient re-cycle.
The first demand is generally fulfilled through appropriate choice of both the active metal and the
chiral ligand. The latter one is usually selected among those acknowledged as privileged, an
attribute coined by Jacobsen for indicating chiral structures of large applicability, widely used for
the synthesis of chiral molecules and for discovering new enantioselective reactions.2 The second
aspect is of extreme importance, because difficult catalyst re-cycle is often the major hinder to the
industrial application of homogeneous catalysis.
Nowadays, the methodology of choice for solving this problem is the multiphase
homogeneous catalysis.3 This strategy requires heterogenisation of the catalyst,
P
through either its anchorage to a solid support or its selective immobilisation in a
C
liquid phase immiscible with the products phase (Figure 1). In these conditions, the
Fig.1
catalyst is easily recycled at the end of the reaction by simple phase separation. Of
course, this approach requires suitable tagging of the catalyst, because precise
physical properties are necessary for its efficient heterogeneisation. Within
this challenging frame, our laboratory has recently developed a strategy
O
O
NH
HN
aimed to extend the use of privileged structures based on transcyclohexanediamine (e.g. Trost ligand in Figure 2) to multiphase
homogeneous catalysis.4
P
Ph2
P
Ph2
Fig.2
The approach involves introduction of the coordinating functions in the skeleton of D-glucose, and
is inspired by the clear structural analogy between trans-cyclohexanediamine and 2,3-Dglucodiamine (Figure 3).5
1
Asymmetric Catalysis on Industrial Scale, 1st ed.; Blaser, H.U., Schmidt, E.,Eds.; Wiley-VCH: Weinheim, Germany,
2004.
2
Yoon, P.; Jacobsen, E.N. Science 2003, 299, 1691.
3
Multiphase Homogeneous Catalysis, 1st ed.; Cornils, B.; Herrmann, W.A., Horvarth, I.T.; Leitner, W.; Mecking, S.;
Olivier-Borbigou, H.; Vogt, D.; Eds.; Wiley-VCH: Weinheim, Germany, 2005
4
(a)Trost, B.M.; Van Vranken, D.L. Angew.Chem. Int. Ed. Engl. 1992, 31, 228. (b) Trost, B.M.; Van Kranken, D.L.;
Bingel, C. J. Am. Chem. Soc. 1992, 114, 9327. (c) Trost, B.M.; Van Vranken, D.L. J. Am. Chem. Soc. 1993
5
(a) Steinborn, D.; Junicke, H. Chem. Rev. 2000, 100, 4283. (b) Dièguez, M.; Pàmies, O.; Ruiz, A.; Diaz, Y.; Castillòn,
S.; Claver, C. Coord. Chem. Rev. 2004, 248, 2165. (c) Dièguez, M.; Pàmies, O.; Claver, C. Chem. Rev. 2004, 104,
3189. (d ) Diaz, Y.; Castillòn, S.; Claver, C. Chem. Soc. Rev. 2005, 34, 702.
RO
6
OR
5
RO
O
4
1
3
H2N
O
OR RO
H2N
2
H2N
NH2
NH2
H2N
OR
NH2
NH2
trans-cyclohexanediamine
2,3-glucodiamine
Fig.3
In both cases, the ligand arms are in adjacent equatorial positions of a six-member ring. The extra
advantage of the sugar moiety is the presence of additional functional sites, useful for tagging the
ligands in view of their application in multiphase catalysis. The essential coordination features are
identical to those of the Trost ligand, because tagging affects other portions of the molecules. Thus,
it is expected that the performance of the corresponding catalyst is similar in the different versions,
which offers the brilliant possibility to select the catalyst-phase more suited for a given reaction.
Preliminary studies dealt with the synthesis of the bis(phosphinoamide) (1) and the corresponding
bis(phosphinoester) (2) prepared by condensation of the appropriate D-glucose precursors with 2diphenylphosphinobenzoic acid (Figure 4).6
O
O
O
Ph
O
CH2Ph
O
O
NH
O
Ph
O
HN
P
CH2Ph
O
O
P
Ph Ph Ph Ph
O
O
P
P
Ph Ph Ph Ph
1
2
Fig.4
The ligands were examined in the Pd-catalysed asymmetric desymmetrisation of meso-2cyclopenten-1,4-diol biscarbamate in traditional conditions (Scheme 1).
Ts
TsNCO
HO
OH
NH
O
O
Ts
HN
O
O
Pd (dba)2
Ligand
-TsNH2
-CO2
Ts
H
N
O
O
H
-(3R,6S)-I
Scheme 1
This intramolecular allylic substitution affords the key precursors of mannostatines, and is also a
standard test for the assessment of the stereo-orienting properties of new ligands.
Bis(phosphinoamide) yielded the product high ee’s (up to 97%), though the multi-step synthesis
6
F. Ruffo, R. Del Litto, A. De Roma, A. D’Errico, S. Magnolia, Tetrahedron: Asymmetry, 17 (2006) 2265; R. Del Litto, A. De
Roma, F. Ruffo, Inorganic Chemistry Communications, 10 (2007) 618.
does not encourage its use. On the other hand, the bis(phosphinoester) is immediately available
from commercial sources. Unfortunately, its activity is less satisfying, because the ee of the product
did not exceed 82%.
On these grounds and with the intent of combining both synthetic convenience and high catalytic
performance, we conveyed our attention towards the synthesis of mixed (phoshinoesterphosphinoamide) ligands (3a, 3b, 3c)(Figure 5).
HO
O
O
MeO
O
O
OCH2Ph
O
O
O
P
Ph Ph
O
O
OCH2Ph
O
O
O
HN
P
Ph Ph
O
P
Ph Ph
O
HN
P
Ph Ph
P
Ph Ph
3b
3a
OCH2Ph
O
HO
O
HN
P
Ph Ph
3c
Fig.5
The new ligands plainly fulfil our expectation:
- the synthesis is very convenient, and requires only four simple steps from inexpensive Nacetylglucosamine.
- in traditional catalytic conditions the corresponding Pd complexes are as active as the one
containing the analogous bis(phosphinoamide) (1).
- the functionalised ligand 3c reveals to be suited for multiphase catalysis.
More precisely, 3c was used both in ionic liquids and as a supported catalyst, provided its
heterogeneisation to a suitable solid.
In line with our assumption, the catalytic performance in traditional conditions was as high as that
of the original Trost molecule, and promising results were achieved in multiphase homogeneous
conditions.