Communications
(para), À130.8 ppm (ortho-C6F5)] suggest that neutral boron
Accordingly, the absolute configuration of the anion in 9
was assigned by single-crystal X-ray analysis as (1R, 2S, 3S,
4S; Figure 2).[16] Again, a parallel orientation of the phenyl
group and the C6F5 ring is observed, but there is a change in
the orientation of the B–H bond, indicating the possibility of
an altered chiral induction of the two isomers in an
asymmetric hydrogenation reaction. In the solid state, 9 and
centers are present.[15]
As separation of the two diastereomers 6 and 7 was not
possible at this point, partitioning of the salts formed from the
FLPs after the hydrogen splitting was investigated. Treatment
of an n-pentane solution of the borane mixture 6 and 7 with
hydrogen at 258C in the presence of tri-tert-butylphosphine
(tBu3P; 8) resulted in the precipitation of a colorless solid in
53% yield (Scheme 1). Multinuclear NMR spectroscopy
corroborated the product as a mixture of the activated FLP
salts 9 and 10 after the hydrogen splitting. Furthermore,
recrystallization in dichloromethane and n-pentane produced
single crystals consisting of 9 and 10 in a 1:1 ratio as confirmed
by X-ray analysis (see the Supporting Information),[16] and
precluded separation at this stage. However, a more detailed
investigation revealed that the FLP 6/8 led to a faster
hydrogen splitting reaction than the corresponding FLP 7/8.
This observation enabled isolation of the diastereomerically
pure compounds 9 and 10 through kinetically controlled
product formation. In the 31P NMR spectrum of compound 10
a doublet at d = 59.8 ppm with a JP-H coupling of 431 Hz is
consistent with the presence of the tri-tert-butylphosphonium
[tBu3PH]+ cation. In the 1H NMR spectrum a broad multiplet
at d = 2.87 ppm and a doublet in the 11B NMR spectrum at
d = À18.8 ppm (JB-H = 88 Hz) support the existence of a
hydridoborate anion. The 19F NMR spectrum reveals two
sets of typical C6F5 signals [d = À132.3 (ortho), À132.6
(ortho), À166.4 (para), À167.5 (para), À167.9 (meta),
À168.5 ppm (meta-C6F5)], which can be attributed to the
presence of two diastereotopic C6F5 rings. Comparable
spectral data were observed for compound 9. Single crystals
of the salt 10, suitable for X-ray structure determination, were
grown from a dichloromethane/n-pentane solution, and one
of two molecules in the asymmetric unit is shown in
Figure 1.[16] The absolute configuration of the anion in the
salt 10 was determined as bis(perfluorophenyl)-
((1R,2R,3R,4S)-4,7,7-trimethyl-3-phenylbicyclo[2.2.1]heptan-
2-yl)hydroborate. Interestingly, the phenyl ring in the chiral
backbone is oriented parallel to one of the C6F5 rings and is
separated by a distance of around 350 pm, thus providing the
basis for a controlled conformation which should be impor-
tant for subsequent effective catalytic applications.[17]
À
10 have multiple C H···F hydrogen-bonding interactions that
connect the phosphonium and hydridoborate moieties.[5b]
Figure 2. Crystal structure of 9. Hydrogen atoms and solvent mole-
cules were omitted for clarity—except for the hydrogen atoms bonded
to boron and phosphorus. Thermal ellipsoids are set at 50% proba-
bility.
With the chiral compounds 9 and 10 in hand, the
envisioned catalytic hydrogenation of prochiral imines was
investigated (Table 1). In the presence of 5 mol% catalyst
(1:1 mixture of 9 and 10) at 658C and 25 bar hydrogen, imine
N-(1-phenylethylidene)aniline (11a) was transformed into
the secondary amine 12a with an enantioselectivity of 20% ee
(S enantiomer; Table 1, entry 1). Using the diastereomerical-
ly pure salts (9 and 10) as catalysts for the hydrogenation of
11a gave more encouraging results. In the case of hydro-
genation using 9, full conversion into the S product was
achieved in 48% ee (Table 1, entry 2). Salt 10 led to the
R enantiomer with an even higher enantioselectivity of
79% ee (Table 1, entry 3). In addition to this, a comparison
of the measured enantioselectivities obtained with pure 9, a
1:1 mixture of 9 and 10, and pure 10 supported the assumption
that catalyst 9 was more active in the catalytic hydrogenation
than 10. Correlating this observation to the fact that hydrogen
splitting was also faster with the Lewis pair 6/8 (precursor of
9) gives significant information with respect to the rate-
determining step in the reactions using the two diastereomers.
To assess the substrate scope, a variety of substituted
imine derivatives were hydrogenated using diastereomerical-
ly pure 10 as the catalyst. Upon increasing the steric
hindrance of the substrate, the yield of the corresponding
amine decreased significantly (Table 1, entries 4 and 5). For
Figure 1. Crystal structure of 10. Hydrogen atoms and solvent mole-
cules were omitted for clarity—except for the hydrogen atoms bonded
to boron and phosphorus. Thermal ellipsoids are set at 50% proba-
bility.
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 9475 –9478