A concise synthesis of a rigid isomannide-based diphosphine ligand and structural characterisation of an alkoxyphosphonium intermediate

Article abstract of DOI:10.1039/b401301h

The synthesis of the novel C₂-symmetric diphosphine 1,4:3,6-dianhydro-2,5-bis(diphenylphosphino)-D-mannitol (ddppm) from D-isomannide is reported and its performance in asymmetric hydrogenations discussed.

Full text of DOI:10.1039/b401301h

A concise synthesis of a rigid isomannide-based diphosphine ligand and
structural characterisation of an alkoxyphosphonium intermediate†
Cristina Carcedo, Athanasia Dervisi,* Ian A. Fallis, Liling Ooi and K. M. Abdul Malik
Department of Chemistry, Cardiff University, PO Box 912, Cardiff UK CF10 3TB.
E-mail: dervisia@cardiff.ac.uk; Fax: 029 20874030; Tel: 029 20874081
Received (in Cambridge, Uk) 29th January 2004, Accepted 6th April 2004
First published as an Advance Article on the web 27th April 2004
The synthesis of the novel C2-symmetric diphosphine 1,4:3,6-di-
anhydro-2,5-bis(diphenylphosphino)- -mannitol (ddppm)
from -isomannide is reported and its performance in asym-
metric hydrogenations discussed.
reaction carried out at room temperature affords only the bis-
alkoxyphosphonium salt 2 as a stable intermediate. Compound 2 is
the first structurally characterised alkoxyphosphonium salt of this
type. An ORTEP drawing of 2 shows both – OPPh3 substituents
adopting pseudoaxial positions in the solid state (Fig. 1).‡
D
D
Although
D
-isomannide (1) was first described in 1882 by
The last step in the ddppm synthesis involves an SN2
displacement of the bromide of 3 by diphenyl phosphide as a
nucleophile. It is known that the exo groups of 1,4:3,6 dianhy-
drohexitols are reluctant to undergo SN2 displacements due to steric
constraints, in many cases leading to elimination and racemisation
by-products.2,6 Although no elimination by-products were ob-
served in the last step of the ddppm synthesis, as was the case for
the endo diimine isomannide derivative,2 the epimer 1,4:3,6-dia-
Fauconnier, reports on its synthetic applications as a commercially
available chiral auxiliary have been so far limited.1 This reflects the
synthetic difficulties encountered in forming the hindered endo
derivatives (the endo and exo prefixes refer to the cavity formed by
the two cis fused tetrahydrofuran rings). A report on an endo-
diimine isomannide derivative, used in catalytic asymmetric Diels–
Alder reactions, is the only example to our knowledge of a di-endo
isomannide-derived ligand.2 In the case of phosphine isomannide
derivatives only the di-exo-diphosphine 1,4:3,6-dianhydro-2,5-di-
nhydro-2,5-dideoxy-2,5-bis(diphenylphosphino)- -glucitol
D
(ddppg) was also isolated as shown in Scheme 1. The two isomers
were separated as colourless, air stable crystals by fractional
crystallisation from ethanol. The crystal structure of ddppm shows
the two phosphine groups occupying pseudoequatorial positions
pointing away from the central cavity (Fig. 2).‡ We can speculate
that the formation of the minor isomer arises from the hydro-
phosphination of elimination intermediates such as I (where X =
PPh2) by HPPh2 leading to ddppg as the more thermodynamically
stable product. Alternatively, ddppg may be formed by attack of
Ph2P2 at a bridged phosphonium salt, intermediate II. A secondary
amine analogue of II has been reported previously.6 Ether was the
solvent of choice in the formation of ddppm. In thf the di-endo
ligand was not detected and ddppg was a minor product. The major
products formed were the mono- and di-exo phosphines. This
preferential formation of exo phosphine derivatives favours a
competing mechanism via phosphide-promoted elimination to
yield intermediate I and HPPh2 followed by hydrophosphina-
tion.7
deoxy-2,5-bis(diphenyl-phosphino)- -iditol (ddppi) is known.
L
ddppi is a non-chelating ligand and has been used as a catalyst in
various asymmetric reactions, including hydrogenations.3 Herein
we wish to describe the first reported synthesis of ddppm, a di-
endo phosphine derivative of isomannide. Our interest in develop-
ing isomannide phosphine derivatives and their metal complexes
stems from their potential uses in catalytic applications. An
efficient chirality transfer from ddppm was anticipated because of
the rigid backbone conformation, its chelating nature and the
presence of a C2 axis. In addition, large bite angles were expected
due to the backbone steric requirements that may be a useful feature
in catalytic reactions which involve a reductive elimination step.
ddppm was synthesised in just two successive steps from -
D
isomannide according to the reaction sequence illustrated in
Scheme 1. Precursors to ddppm with chloride and mesylate leaving
groups were also studied but with inferior results to that of
dibromide 3. After several unsuccessful attempts, the key di-
bromide 3 was prepared using a protocol for the conversion of alkyl
alcohols to the corresponding halides.4 This method constitutes a
significant improvement on the reported two step procedure for the
preparation of 3 from
D
-isomannide involving difficult separations
from other epimers and low yields.5 Reaction of
D-isomannide with
two equivalents each of bromine, triphenyphosphine and imidazole
in refluxing acetonitrile afforded 3 in quantitative yields. The same
The chelating ability of ddppm is demonstrated with the
synthesis of the rhodium complex [Rh(ddppm)(CH3CN)2]BF4 (4),
Scheme 1. The crystal structure of 4 (Fig. 2) shows the two PPh2
groups of ddppm adopting pseudoaxial positions necessary for
chelation to the metal.‡ As a result the P–P distance changes from
5.9085(24) Å in the free ligand to 3.4043(1) Å on co-ordination. A
distorted tetrahedral arrangement is observed around the phospho-
rus atoms imposed from the steric requirements of the isommanide
backbone, with tetrahedral angles ranging from 98.97° to 127.01°.
ddppm has a large bite angle of 98.61(3)° accommodated in the
square co-ordination plane by the smaller P–Rh–N angles at
85.84(8)° and 87.21(8)° and N–Rh–N angle at 88.39(11)°. For
comparison the bidentate angle of BINAP, another seven-mem-
bered chelate ligand, in an analogous complex is 91.8(1)°.8
In order to assess the efficiency of ddppm in asymmetric
catalysis we tested ddppm in the well established reaction of olefin
hydrogenation (Table 1).9 Using a rhodium catalyst itaconic acid is
hydrogenated quantitatively under an atmospheric pressure of H2
with a good enantioselectivity (run 1). An increase in the hydrogen
pressure results in decreased yields and selectivities (run 2). Similar
Scheme 1 (a) 2 eq. Br2/PPh3/imidazole, CH3CN, reflux, 3d; (b) 2 eq.
LiPPh2, Et2O, 3h; (c) [Rh(COD)2]BF4/ddppm, CH3CN.
† Electronic Supplementary Information (ESI) available: characterisation
data for 2, 4, ddppm and ddppg (including structural data for ddppg),
experimental for hydrogenation reactions.
1236
C h e m . C o m m u n . , 2 0 0 4 , 1 2 3 6 – 1 2 3 7
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4

Table 1 Asymmetric hydrogenation of olefins
H2
(atm)
%
Run R1
R2
R3
Catalyst Yield
% ee
1
2
3
4
5
6
7
H
H
Me
Me
H
CH2CO2H
CH2CO2H
CH2CO2Me
CH2CO2Me
NHCOCH3 Ph
NHCOCH3 Ph
NHCOCH3 Ph
H
H
H
H
1
3.5
1
20
1
A
A
A
A
A
A
B
> 99
89
> 99
15
> 99
85
50
64 (S)
59 (S)
41 (S)
31 (S)
30 (R)
36 (R)
73 (R)
H
H
3.5
3.5
Reaction conditions: 0.01 mmol catalyst, 1 mmol substrate in MeOH (15
mL) at room temperature. Catalyst A = [Rh(COD)2]BF4/ddppm, B =
[Ru(p-cymene)Cl2]2/ddppm. %ee values were measured by chiral HPLC
using a Chiralcel OD column and hexane/IPA as the eluent.
Fig. 1 ORTEP ellipsoid plot at 30% probability of the molecular structure
of 2. The bromide counter ions and solvent molecules have been omitted for
clarity. Selected distances (Å): P(1)–O(1) 1.577(4), P(2)–O(2) 1.572(4),
O(1)–C(1) 1.460(6), O(2)–C(6) 1.462(6).
To conclude, we have prepared a new chiral chelating diph-
osphine ligand with a rigid backbone conformation in two steps
from readily available starting materials. Compared to other known
chiral diphosphines such as BINAP, ddppm synthesis involves a
simpler synthetic methodology that does not require a resolution
step.11 Further work that explores possible catalytic applications of
ddppm and its derivatives is currently underway.
A.D. thanks Synetix for lectureship funding, Johnson Matthey
for a precious metals loan and the EPSRC mass spectrometry
service.
Notes and references
‡ Crystal data for ddppm: C30H28O2P2, M
= 482.46, orthorombic,
P212121, a = 11.5522(9), b = 14.0034(11), c = 15.4518(14) Å3, V =
2499.6(4) Å3, Z = 4, Dc = 1.282 g cm23, m(Mo–Ka) = 0.71073 Å, T =
150(2) K, 24363 reflections collected, 4278 independent reflections [R(int)
= 0.2142], F2 refinement, R1 = 0.0731, wR2 = 0.1400 for [I > 2s(I)], 308
parameters. Flack parameter = [0.14(19)]. For 2·3CH4O: C45H50O7P2Br2,
M = 924.61, Monoclinic, P21, a = 9.3604(2), b = 23.4811(5), c =
9.7168(2) Å3, b = 95.4883(7)°, V = 2125.89(8) Å3, Z = 2, Dc = 1.444 g
cm23, m(Mo–Ka) = 0.71073 Å, T = 150(2) K, 16969 reflections collected,
9299 independent reflections [R(int)
= =
0.0766], F2 refinement, R1
0.0609, wR2 = 0.1466 for [I > 2s(I)], 508 parameters. The absolute
structure was correctly indicated by the Flack parameter being zero within
experimental error [0.020(9)]. For 4·C2H3N: C36H37BF4N3O2P2Rh, M =
795.35, Monoclinic, P21, a = 10.2793(2), b = 15.1367(4), c = 11.5954(4)
Å3, b = 100.7005(12)°, V = 1772.81(8) Å3, Z = 2, Dc = 1.490 g cm23
,
m(Mo–Ka) = 0.71073 Å, T = 150(2) K, 16148 reflections collected, 6563
independent reflections [R(int) = 0.0477], F2 refinement, R1 = 0.0312,
wR2 = 0.0698 for [I > 2s(I)], 445 parameters. The absolute structure was
correctly indicated by the Flack parameter being zero within experimental
error [20.04(2)].CCDC 230131–230134. See http://www.rsc.org/suppdata/
cc/b4/b401301h/ for crystallographic data in .cif or other electronic
format.
Fig. 2 ORTEP ellipsoid plot at 30% probability of the molecular structure
of ddppm (top) and 4 (bottom). Solvent molecules and the BF4 anion in 4
have been omitted for clarity. Selected distances (Å) for 4: Rh(1)–N(1)
2.068(3), Rh(1)–N(2) 2.069(3), Rh(1)–P(2) 2.2398(9), Rh(1)–P(1)
2.2503(9).
1 A. C. R. Fauconnier, Acad. Sci., 1882, 95, 991.
2 G. De Coster, K. Vandyck, E. Van der Eycken, J. Van der Eycken, M.
Elseviers and H. Roper, Tetrahedron: Asymmetry, 2002, 13, 1673.
3 (a) J. Bakos, B. Heil and L. Markó, J. Organomet. Chem., 1983, 253(2),
249; (b) J.-C. Yuan and S.-J. Lu, Organometallics, 2001, 20, 2697.
4 J. G. Calzada and J. Hooz, Org. Synth. Coll., 1988, Vol VI, 634.
5 V. Cere, C. Paolucci, S. Pollicino, E. Sandri and A. Fava, Tetrahedron
Lett., 1989, 30(48), 6737.
6 A. C. Cope and T. V. Shen, J. Am. Chem. Soc., 1956, 78, 3177.
7 T. Bunlaksananusorn and P. Knochel, Tetrahedron Lett., 2002, 43,
5817.
8 K. Toriumi, T. Ito, H. Takaya, T. Souchi and R. Noyori, Acta
Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1982, B38,
807.
9 W. Tang and H. Zhang, Chem. Rev., 2003, 103, 3029.
10 J. Wu, C. C. Pai, W. H. Kwok, R. W. Guo, T. T. L. Au-Yeung, C. H.
Yeung and A. S. C. Chan, Tetrahedron: Asymmetry, 2003, 14, 987.
11 H. Takaya, K. Mashima, K. Koyano, M. Yagi, H. Kumobayashi, T.
Taketomi, S. Akutagawa and R. Noyori, J. Org. Chem., 1986, 51,
629.
observations have been reported previously by other researchers.10
The corresponding methyl ester affords methyl succinate quantita-
tively but with lower enantioselectivity than itaconic acid (compare
runs 1 and 3). (Z)-N-acetamido cinnamic acid is again hydro-
genated quantitatively but with moderate enantioselectivity (30%,
run 5). For comparison, we also generated in situ a ruthenium
catalyst from [Ru(p-cymene)Cl2]2 and ddppm (run 7).10 Here
hydrogenation of (Z)-N-acetamido cinnamic acid gives the best
enantioselectivity (73%) observed from this set of experiments.
These unoptimised results show that ddppm forms active catalysts
although currently not with the high enantioselectivities observed
with other hydrogenation catalysts.9
C h e m . C o m m u n . , 2 0 0 4 , 1 2 3 6 – 1 2 3 7
1237