and 38% yield of the aldehyde 3a. By-products of benzylic
formate could not be detected by GC–MS; no reaction was
observed without addition of a base (Table 1, entry 4). Base
screening showed that the organic base NEt3 (Table 1,
entry 6) was ineffective while a stronger base (NaOtBu)
gave the best reactivity with a 74% yield of the ester, as
determined by GC; the aldehyde could not be observed
(Table 1, entry 8). Solvent effects were also tested: using
MeOH as solvent afforded the best result (Table 1, entries 8–
10). No benzylic formate was detected even in other solvents
(DMSO, THF; Table 1, entries 9–10).
desired ester 2c was isolated, no aldehyde was left (Scheme 2
bottom, (c)). Consistent with these results, the following
pathway was proposed (Scheme 3) and a possible mechanism
was outlined (see the Supporting Information).
The conditions of entry 8 in Table 1 were then applied to
investigate the substrate scope. Unfortunately, the reaction
was very sensitive to the substituent on the phenyl ring of the
benzylic alcohol. Unsubstituted benzyl alcohol gave the
desired product 2a smoothly and no aldehyde 3a was
observed (Scheme 2, (a)), but for strongly electron-donating
Scheme 3. Possible pathway for the oxidative esterification reaction.
Firstly, benzylic alcohol coordinates to the PdII center
followed by b-hydride elimination to generate benzaldehyde
(step I), which directly reacts with another alcohol to form the
corresponding hemiacetal. The desired ester is a result of the
palladium-catalyzed b-hydride elimination of the hemiacetal
(step II). In these two reaction steps, the nature of the
substituent R has a significant influence on the reactivity of
the intermediates. An electron-donating group such as p-
À
OMe makes the C H bonds of CH2 more electron-rich so that
the b hydride can be easily eliminated and it thus becomes
easier to form the corresponding aldehyde (Scheme 2, (b)).
However, a strongly electron-withdrawing group such as p-
À
NO2 makes the C H bonds more electron-deficient so that
the b hydride cannot be easily eliminated; thus, it is relatively
difficult to form the corresponding aldehyde (Scheme 2, (c)).
For the aldehyde intermediate product, the carbonyl group is
also strongly influenced by the electronic properties of the
substitutent on the phenyl ring. The electron-donating p-OMe
group passivates the carbonyl group and thus hinders the
formation of the corresponding hemiacetal, so that the
reaction will prefer the aldehyde step (Scheme 2 bottom,
(b)). However, the electron-withdrawing group p-NO2 acti-
vates the carbonyl group and makes it more electron-
deficient, thus the carbonyl group easily reacts with another
alcohol to form the corresponding hemiacetal followed by
palladium-catalyzed b-hydride elimination to result in the
desired product (Scheme 2 bottom, (c)).
According to this hypothesis, a more electron-deficient
palladium center is required, as it can promote the harder to
accomplish b-hydride elimination of the electron-deficient
benzylic alcohols and can coordinate to the corresponding
aldehyde of the electron-rich benzylic alcohols to ease the
formation of the hemiacetal. To make the catalyst more
electron-deficient, a silver salt (AgBF4) was added to remove
the anionic ligand ClÀ from the palladium catalyst precursor.
Good results were achieved with 61% and 75% yield of the
desired methyl 4-methoxybenzoate and methyl 4-nitroben-
zoate (Table 2, entries 1 and 7), respectively.
Scheme 2. Substitution effect on the oxidative esterification reaction.
groups (EDGs), such as the methoxy group, no 4-methox-
ybenzyl alcohol could be found while 4-methoxybenzalde-
hyde 3b was observed in 76% yield by NMR spectroscopy; in
addition, 20% of the desired methyl ester 2b was detected
(Scheme 2, (b)). However, for the strongly electron-with-
drawing groups (EWGs), such as p-NO2, conversion of the
substrate 4-nitrobenzyl alcohol 1c was incomplete (54%
unreacted); no formation of the aldehyde was observed and
36% of the desired ester 2c was detected (Scheme 2, (c)). In
addition, the esterifications of three typical aldehydes with
different electronic properties were carried out under the
same conditions. The benzaldehyde gave the corresponding
ester in high selectivity and good yield (Scheme 2 bottom,
(a)). The benzaldehyde substituted with the strongly electron-
donating group p-OMe gave the desired product 2b in only
14% yield in NMR spectroscopy, with 68% of aldehyde 3b
remaining unreacted (Scheme 2 bottom, (b)). For the benzal-
dehyde substituted with the strongly electron-withdrawing
group p-NO2, the reaction proceeded smoothly; 88% of the
The optimized conditions were then used for a variety of
substituted benzylic alcohols and it could be shown that the
substitution sensitivity had been overcome (Table 2). Various
substituted benzylic alcohols could be employed in this
aerobic oxidative esterification reaction. The desired ester
was isolated in 61% yield when the benzylic alcohol
Angew. Chem. Int. Ed. 2011, 50, 5144 –5148
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5145