Communication
(entry 1). With the bulky 1,3-bis(2,4,6-trimethylphenyl)imidazo-
lin-2-ylidene ligand (IMes) as the NHC in our catalyst series, we
found the reactivity to be strongly linked to the size of the
partner phosphine ligand (entries 2–6). The large rigid catalyst
3b, in which the phosphine is triphenylphosphine, delivered
only 27% conversion (entry 2). Utilising more flexible catalysts
bearing tribenzylphosphine (3c) and tri-n-butylphosphine[7] li-
gands (3d) resulted in a large increase in activity, giving near
quantitative conversion (Table 1, entries 3–4). However, the
best results were obtained with catalysts bearing smaller li-
gands, such as triethylphosphine (3e) and dimethylphenyl-
phosphine (3 f) (entries 5–6). Having established that catalysts
bearing small phosphine ligands gave increased activity, we
sought to further improve activity with less encumbered, N-
alkyl-substituted NHCs. However, each catalyst of this type
(3g–i; entries 7–9) failed to deliver any hydrogenated product
2a.
It was proposed that these complexes 3g–i exhibited poor
activity due to a strong substrate–catalyst binding that limits
the recycling of the activated catalyst. In contrast, we have
shown that more encumbered IMes/phosphine catalysts paired
with a less coordinating counter ion (BArF) have increased ac-
tivity at lower catalyst loading and an appreciably enhanced
range of applicable reaction solvents in HIE processes.[8d] Ac-
cordingly, using the success of catalyst 3 f as a foundation, we
synthesised BArF complex 3j by a recently developed proce-
dure circumventing difficult inert atmosphere filtration meth-
ods (see the Supporting Information, Section 7).[8d] As shown
for entry 10, this new complex (3j) gave complete conversion
in the hydrogenation of 1a to 2a; furthermore, the hydroge-
nation process was shown to proceed more rapidly with the
BArF complex than with the equivalent PF6 species (see the
Supporting Information, Section 10).
a two-level, three-factor, full factorial design of experiments
(see the Supporting Information, Section 11). The three factors
chosen for observation were catalyst loading, reaction concen-
tration and reaction time. The study showed, perhaps unsur-
prisingly, that increasing catalyst loading and reaction time
both strongly enhanced the reaction efficiency. More interest-
ingly, the study also revealed that overly increasing the con-
centration was detrimental to the reaction, plausibly indicating
that the substrate complexation and subsequent product de-
complexation is inhibiting catalyst turnover,[2] in accordance
with our observations on the inactivity of catalysts 3g–i.
Following on from this experimental design process, we ap-
plied the optimised conditions (0.5 mol% 3j, 2 h, 0.1m in
CH2Cl2), to a broad range of unsaturated substrates (Table 2).
After the initial success in the reduction of 1a, further enone
substrates 1b–d all performed well, with no hindrance to the
reduction by para-, meta- or ortho-substitution of the aromatic
ring. Increasing the steric bulk adjacent to the donor group
also resulted in full conversion (1e). Pleasingly, alkyl-substitut-
ed enones 1 f and g also readily underwent hydrogenation;
however, the increased steric bulk in 1g required moderately
increased catalyst loading and extended reaction time
(1 mol% and 16 h) for complete conversion. In contrast, the
standard optimised conditions proved to be effective in the
hydrogenation of the chalcone derivative 1h. More challeng-
ing a-substituted enones 1i and j required both higher cata-
lyst loading and longer reaction times (1 mol% and 16 h), but,
notably, complete conversion was still achieved at 1 atm of H2
pressure. Furthermore, b-disubstituted enone 1k initially
proved to be problematic under the optimised conditions, but
a modest increase in temperature, along with catalyst loading
and reaction time (2 mol%, 358C, 40 h), gave quantitative con-
version to the reduced product.
With complex 3j chosen for further study due to its superior
performance, we turned our attention to understanding the
factors affecting this overall process. To this end, we utilised
Following the selective reduction of a range of ketones, we
next investigated a range of alternative directing groups. Nota-
bly, the sensitive carbonate group in 1l remained intact under
Table 2. Substrate scope and chemoselectivity.
[a] 1 (0.4 mmol), 3j (0.002 mmol, 0.5 mol%), CH2Cl2 (4 mL), H2 (1 atm). [b] Conversion calculated from 1H NMR analysis of the crude product. [c] 3j
(0.004 mmol, 1.0 mol%) for 16 h. [d] 3j (0.008 mmol, 2.0 mol%) at 358C for 40 h. [e] 3j (0.004 mmol, 1.0 mol%).
Chem. Eur. J. 2016, 22, 4738 – 4742
4739
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim