.
Angewandte
Communications
We found that diphosphine ligands, (rac)-2,2’-bis(diphe-
nylphosphino)-1,1’-binaphthyl (binap) and MeO-biphep (5;
see Table 1 for structure), when used with 3 also promoted the
1,2-addition, but a prolonged reaction time was required
(Table 1, entries 2 and 3). Moreover, no significant enantio-
selectivity was observed with chiral ligands, because the
reaction was much faster with PPh3 as the ligand (Table 1,
entry 3). Therefore, we decided to change the catalytic
system. When diphosphine 5 was used in combination with
the CuCl/NaOtBu system, the reaction did not give the
desired product 4a (Table 1, entry 4). When we used the P–N
ligand 6, a 12% yield was obtained, but we observed a modest
increase in enantioselectivity up to 20%. The use of the
sterically hindered ligand DTBM-Segphos (7) considerably
improved the enantioselectivity (Table 1, entry 6). Finally, the
addition of methanol led to significant improvement of both
the yield and enantioselectivity, which suggests that methanol
plays a significant role in regenerating the catalyst (Table 1,
entry 7). The optimized conditions were found to be: CuCl
(5 mol%), NaOtBu (5 mol%), MeOH (4 equiv), and (R)-
DTBM-Segphos (10 mol%) at room temperature in THF.
The desired product was obtained in 46% yield and 97% ee.
However, we found that these conditions gave fairly irrepro-
ducible results, both for yields and enantioselectivities, with
significant drops (down to 65% from 97%) in the enantio-
selection.
Scheme 1. Synthesis of new diphosphine copper(I) bifluoride com-
plexes.
À139.6 ppm for Cu-F-H-F and d = À192.4 ppm for Cu-F-H-
F). This was confirmed by crystallization in toluene/n-hexane
which afforded crystals of X-ray quality for analysis. The
structure confirmed the formation of a neutral complex
bearing a s-bonded H2F3À anion. Finally, we were delighted to
find that those complexes are very air and moisture stable,
both in the solid state and in solution, and display excellent
reactivity in our model reaction.
Recently, our group developed a new pathway to unpre-
cedented N-heterocyclic carbene (NHC) copper(I) bifluoride
complexes, which proved to be excellent catalysts for
nucleophilic transfer to electrophilic double bonds, aldehydes,
and chiral imines.[19] Furthermore, hydrogen bifluoride
(FHFÀ) is a unique anion, as it features the strongest known
hydrogen bond.[18] This anion is also an anhydrous source of
These new complexes 9a,b were tested in the copper-
catalyzed 1,2-addition of a silicon nucleophile to aldehydes.
Reaction of 1a with 2 and copper complex 9a (5 mol%)
under the previously established optimized conditions, was
completed in 16 h to afford the (R)-a-hydroxysilane (R)-4a in
> 99% ee and 87% yield of isolated product. The use of
copper complex 9b afforded the (S)-a-hydroxysilane (S)-4a
in > 99% ee and 82% yield of isolated product (Table 2,
entries 1 and 2). The absolute configuration of compound 4a
is based on literature data.[8d]
The optimized reaction conditions were then applied to
various aromatic and aliphatic aldehydes 1a–q (Table 2).
Regarding aromatic aldehydes, both electron-donating as well
as electron-withdrawing groups were well tolerated. Yields of
isolated products were good to high and the corresponding a-
hydroxysilanes were obtained with excellent enantioselectiv-
ity. The electronic substitutions on the aromatic ring did not
affect the enantioselectivity. However, aromatic aldehydes
with functional groups at the ortho-position proved to be less
reactive and gave the product in moderate yield (Table 2,
entries 3 and 16), as a consequence of the steric hindrance at
the ortho position. To our delight, we found that aliphatic
aldehydes reacted well (Table 2, entries 9, 12, and 13).
However, propionaldehyde gave the desired product in only
87% ee (Table 2, entry 12). This result suggests that the
aldehyde is not bulky enough to obtain good selectivity.
Neither aromatic aldehydes with reactive functional groups in
the para position (NO2 and OH groups) nor pyridine-4-
carboxaldehyde yielded the desired a-hydroxysilane.
À
fluoride that can activate the Si B reagent 2, thus avoiding
the need for co-catalysts such as alkoxides. These observa-
tions prompted us to design a new family of well-defined
diphosphine copper(I) complexes bearing a bifluoride coun-
teranion, which might lead to better control of the formation
of the copper–nucleophile active species, and thus reprodu-
cible experiments. The original procedure was then modified
and we found that the reaction of copper(I) iodide with
silver(I) hydrogen fluoride in acetonitrile gave, after the
removal of silver iodide by filtration, the tetrakis(acetonitrile)
copper(I) hydrogen bifluoride intermediate (8) in situ. The
complex was identified by a sharp singlet at d = À166 ppm by
19F NMR analysis (in CD3CN) and an acidic proton at d =
1
14.1 ppm by H NMR analysis, but could not be isolated by
crystallization. However, the addition of ligands 7a,b gave,
after removal of the solvent, the new cationic complexes 9a,b
(Scheme 1) in excellent yield. We found that, depending on
the quality of the starting commercial silver bifluoride, the
bifluoride anion could be contaminated with dihydrogen
trifluoride anion (see the Supporting Information for details),
albeit without any detrimental effect on their catalytic
activity. Furthermore, 19F NMR analysis suggested that
those complexes can exist either in cationic form in coordi-
nating solvents such as acetonitrile (one sharp peak around
d = À165.5 ppm) or as neutral copper-fluoride-bound com-
plexes in non-coordinating solvents such as benzene (d =
Finally, we decided to check the reactivity of more
challenging substrates, such as a,b-unsaturated aldehydes,
for which competition between 1,2- and 1,4- addition could be
expected. Citral 10[20] reacted with Me2PhSiBpin in the
2
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!