Angewandte
Chemie
Table 1: Rhodium-catalyzed asymmetric hydrogenation of methyl
2-hydroxymethylacrylate (10a).[a]
Entry
L
L’ ee [%]
R, S Effect
Heterocomplex [%][b]
1
2
3
4
5
6
7
8
9
1
2
1
2
1
1
1
1
1
1
2
3
3
4
5
6
7
8
31
13
S
S
R
S
R
R
R
R
R
–
–
–
–
91
85
94
97
86
70
>99
94 (94)[c]
34
electronic
electronic
electronic
electronic
steric
94
94
94
95
steric
H bond
>99
[a] Ratio L/L’/[Rh(cod)2]BF4/substrate=1.1:1:1:100; solvent: CH2Cl2.
Reaction performed at 10 bar H2 pressure at 298 K for 16h. Full
conversions were obtained in all cases. [b] The amount of heterocomplex
present in solution was determined by integration of the phosphine
signals in the 31P NMR spectrum (20 mm in CD2Cl2, 298 K). [c] Deter-
mined in the presence of phenylurea (1 equiv with respect to L’).
Figure 1. Supramolecular bidentate complexes [Rh(cod)(1)(8)]BF4 (cod
omitted for clarity) and DFT calculations: a) Hydrogen bonds between
the ester and urea units. Relative energy=+7.5 kcalmolÀ1, d(H-bond)
=
phosphine ligands have, on the other hand, a dramatic effect
on heterocomplex formation, as is evident from experiments
with 6 and 7 (Table 1, 31P NMR spectra are shown in the
Supporting Information). Importantly, the complex that
1.9 and 2.6 ꢀ. b) Single hydrogen bond between the NH group of the
phosphoramidite unit and the urea unit. Relative energy=0 kcalmolÀ1
,
d(H-bond) =2.0 ꢀ.
forms
a hydrogen bond between the ligands, [Rh-
(cod)(1)(8)BF4], is the only complex that was formed in
more than 99% purity (see the Supporting Information).
Consistent with this observation, the combination of ligands 8
and 2 did not lead to pure heterocomplex formation, but a
mixture of different species which are difficult to assign.
We next studied the performance of these complexes in
the asymmetric hydrogenation of methyl 2-hydroxymethyl-
acrylate 10a. Ligand 2 was used for comparison with 1. Under
mild conditions (catalyst (1 mol%), H2 (10 bar), 298 K, 16 h),
full conversion was obtained in all experiments. Both
homocomplexes [Rh(cod)(1)2]BF4 and [Rh(cod)(2)2]BF4
gave low selectivities of 31% and 13%, respectively
(Table 1 entries 1 and 2). An excellent enantioselectivity
(94% ee) was obtained with 1 in combination with PPh3 (3;
Table 1 entry 3) while 2 with PPh3 afforded only a moderate
ee values of 34% (Table 1 entry 4). Contrary to our expect-
ations, the amount of heterocomplex present in solution
hardly affected the enantiopurity of the product that is
formed; products were obtained in 94-95% ee in all cases
where this mixed ligand approach was used with 1 (Table 1,
entries 5–8). This result suggests that the heterocomplexes are
much more active than the unselective homocomplexes.
Although the formation of heterocomplexes can be dramat-
ically enhanced (50% for a statistical mixture, 97% for the
combination of ligands 1 and 5) by fine-tuning the electronic
and steric properties of a series of ligands, it does not translate
to higher selectivity in the reaction studied here. In contrast to
these experiments, the supramolecular complex [Rh-
(cod)(1)(8)]BF4 did convert the substrate with the highest
selectivity reported to date (Table 1 entry 9).[11] In a control
experiment in which phenylurea was used as an additive for
the complex [Rh(cod)(1)(3)]BF4 , the selectivity did not
change (Table 1 entry 3), which indicates that the urea
group of the ligand 8 plays a crucial role in the selectivity of
the reaction. Since it is unlikely that the purity of the complex
1 (Figure 1a). Alternatively, a hydrogen bond may be formed
between the NH group of the phosphoramidite 1, which is
known to be a good hydrogen-bond donor[12] and the urea
carbonyl group (Figure 1b). The structures (calculated by
using DFT, BLYP) show that the single hydrogen-bond
interaction (d(H-bond) = 2.0 ꢀ) is more favorable (7.5 kcal
molÀ1) than the double hydrogen bond between the urea–
NH group and the ester carbonyl group. Some other
structures have been calculated in which no hydrogen bonds
were formed; these were all higher in energy (see the
Supporting Information). IR studies on the [Rh-
(cod)(1)(8)]BF4 complex (cod = cyclooctadiene) also con-
firmed the formation of the hydrogen bond between the PNH
unit of 1 and the urea carbonyl group (see below); the effect
of this hydrogen bond on the selectivity of heterocomplex
formation was next studied.
We first studied the complexes that were formed by
mixing [Rh(cod)2]BF4 , 1, and one of the phosphines 3–8 in a
1:1:1 ratio. Interestingly, heterocomplex [Rh(cod)(1)(3)]BF4
was formed in 91% yield, according to the 31P NMR spectrum
of the mixture in CD2Cl2. This value is far above the
statistically expected value and the remaining signals in the
NMR spectrum correspond to the homocombinations. A
similar experiment with the archetypical phosphoramidite
ligand (S)-(+)-(3,5-dioxa-4-phosphacyclohepta[2,1-a;3,4-a’]-
dinaphthalen-4-yl)dimethylamine
((S)-MonoPhos,
2)
showed that only 85% of the heterocomplex [Rh-
(cod)(2)(3)]BF4 was formed. By varying the electronic
properties of the aromatic phosphines, the formation of the
heterocomplex occurred in up to 97% yield for [Rh-
(cod)(1)(5)]BF4 (Table 1, entries 5 and 6). This result is
interesting in itself as it provides a simple tool to make
relatively pure heterocomplexes without using an excess of
one of the ligands. Small changes in the size of the aromatic
Angew. Chem. Int. Ed. 2009, 48, 2162 –2165
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2163