FULL PAPER
Reduction
amides using tetramethyldisil-
AHCTUNGERTGoNNUN xane (TMDS): It is well
of
secondary
known that catalytic hydroge-
nations of secondary amides
are more difficult than the re-
duction of tertiary amides.[19]
This reactivity pattern is sim-
ilar for hydrosilylations. Never-
theless, recently Nagashima
and co-workers[4,12] were able
Scheme 1. Selective reduction of amides in the presence of other reducible functional groups.
industry for the production of organofunctional coupling
agents, low-temperature-vulcanizing silicone rubbers and
elastomers, as well as specialty monomers, we looked for
safer alternatives. Among the various silanes tested for the
Zn-catalyzed reduction of N,N-dimethylbenzamide, phenyl
silane, diphenyl silane, and methyldiethoxy silane worked
excellently at 658C. Due to the cheaper price, and with re-
spect to its chemoselectivity, we thought it best to use
to achieve this challenging goal by using ruthenium and
À
platinum catalysts in the presence of silanes with dual Si H
moieties. Their elegant work encouraged us to reconsider
our zinc-catalyzed procedure by using other silanes. Initially,
we investigated N-benzylbenzamide as model substrate for
the reduction of secondary amides (Table 2). To our delight,
moderate activity (40% yield) was observed by applying the
more reactive zinc triflate and 1,1,3,3-tetramethyldisiloxane
(3 equiv) at 658C. Increasing the temperature gave im-
proved yields. Hence, at 1008C, 85% of dibenzyl amine was
obtained with full conversion of the starting material. Nota-
bly, monosilanes such as phenylsilane and diphenylsilane
were significantly less active (10% yield). Methyldiethoxysi-
lane, triethoxysilane, and triethylsilane showed only 2%
yield of the desired product, and in all of the cases we re-
covered only starting material. In contrast to the work of
Nagashimaꢀs group, 1,1,1,3,3-pentamethyldisiloxane and 1,2-
bis(dimethylsilyl)benzene also afforded the product in low
yield (Table 2, entries 10 and 11). This demonstrates that the
sterics and electronics of the disilane are crucial in enforcing
the reduction of secondary amides. It is worth noting that
the optimized catalytic system for the reduction of tertiary
amides was completely unsuccessful for secondary amides
even at higher temperature (1008C). By applying commer-
cially available 1,1,3,3-tetramethyldisiloxane, we investigated
the effect of various metal triflates and zinc sources for the
reduction of N-benzyl benzamide. Clearly, the reaction did
not proceed without any catalyst. Other metal triflates such
methylACHTUNGTRENNUNGdiethoxysilane as the silane of choice for the reduc-
tion of tertiary amides. In fact, by applying a threefold
excess amount of methyldiethoxysilane in the presence of
catalytic amounts of inexpensive zinc acetate in THF, the re-
duction of a plethora of tertiary amides took place with ex-
cellent chemoselectivity (Table 1). A variety of amides, in-
cluding aromatic and heteroaromatic amides, were hydrosi-
lylated smoothly with high yields up to 93%. Electron-with-
drawing substituents (Table 1, entries 3, 5, and 8) at the para
position gave better yields than electron-donating substitu-
ents (entry 10). Steric hindrance on the amine part of the
amide bond had a significant effect and reduced the rate of
reduction. On the other hand, the reaction proceeded well
with aliphatic groups and N-piperidinyl as well as N-cyclo-
pentyl moieties. However, no reaction was observed with
anilides. Heteroaromatic amides were also reduced in excel-
lent yields (entries 15 and 16). Here furan-based amides
were reduced in higher yields than the corresponding thio-
phene derivatives. This is in contrast to our previous reduc-
tion by applying triethoxysilane. Using halide-substituted
benzamides the reactions proceeded with excellent chemo-
selectivity and no other reductive dehalogenation was ob-
served (entries 5–7). Various other functional groups such as
double bonds, nitro, azo, nitrile, ester, and ether were satis-
factorily accepted under our reaction conditions, too (en-
tries 3, 12, 13, 17, 19, and 20). In none of these cases a re-
duction of the respective functional group except the ester
substituent was observed. Notably, the reaction also tolerat-
ed hydroxy groups as well, although it is well known that al-
cohols might react with the silane (entry 18).[6b] On the
other hand, a methylthio substituent and a keto group were
not tolerated under these conditions. In the latter case re-
duction of the ketone took place. Gratifyingly, scale up of
N,N-dimethylbenzamide reaction to 20 mmol-scale resulted
in no problem and produced 90% of the corresponding
amine.
as scandium triflate, iron(II) triflate, ironACHTNUTRGNEUNG(III) triflate, yt-
terbium triflate, and zirconium triflate also did not show any
activity, even at 1008C (Table 3).
However, in the presence of zinc halides (chloride, bro-
mide, and iodide) some activity with up to 20% yield of the
product was observed. On the other hand, zinc fluoride, ace-
tate, and nitrate gave no conversion at all. In all these cases,
we observed only starting material after the reaction. At-
tempted lower catalyst loading of zinc triflate (5–10 mol%)
led to lower yields of the corresponding amine. Appropriate
selection of the solvent is also important in this reductive
hydrosilylation. For example, the reaction becomes sluggish
in the presence of n-butyl ether or 1,4-dioxane, and only
50% of the secondary amine was obtained. To check the in-
fluence of metal contaminants[20] in the precursor, we used
zinc triflate from different suppliers (ABCR, Sigma Aldrich,
Acros, Alfa Aesar), which all provided similar yields.
Chem. Eur. J. 2011, 17, 12186 – 12192
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
12187