Further reactions of phenyldimethylsilyllithium with
N,N-dimethylamides
Ian Fleming* and Matthew G. Russell
Department of Chemistry, Lensfield Road, Cambridge, UK CB2 1EW. E-mail: if10000@cam.ac.uk; Fax: 44
1223 336362; Tel: 44 1223 336372
Received (in Cambridge, UK) 29th October 2002, Accepted 9th December 2002
First published as an Advance Article on the web 17th December 2002
Phenyldimethylsilyllithium reacts with several N,N-dime-
thylamides, and the intermediates, formulated here as
successively a carbene and an a-silyllithium species, may be
trapped with nucleophiles and electrophiles, respectively,
although not always with the nucleophile or electrophile of
your choice.
We report in this and the following communications other
higher-yielding or informative results from more recent in-
vestigations. Given that our earlier work was full of unexpected
reactions, we have not been too surprised to find that almost
every change we have made has given us more.
One limitation that we particularly wanted to address was
how to trap the lithium reagent with an electrophile when the
lithium reagent was not stabilised by an adjacent phenyl group.
The problem is that the electrophile must not react with the
silyllithium reagent that is used to initiate the sequence of
reactions, which severely limits the possibilities. The only
success that we had earlier was with an isolated terminal double
bond built into the starting amide, but the yield of cyclic product
was low. To try to improve on this result, we replaced the
double bond with a triple bond 11, in the hope that the
intermediate lithium reagent 12 would lead to the cyclopentane
13. In the event, cyclisation took place in even lower yield, and
the minor products 14 and 15 showed that propargylic proton
transfer had taken place in competition with nucleophilic attack
on the triple bond (Scheme 3). In the absence of the silyl group
at the acetylenic terminus, the acetylenic C–H quenched the
lithium reagent.
We reported earlier in three preliminary communications that
the phenyldimethylsilyllithium reagent reacts with the N,N-
dimethylamide 1 to give one of three different products 2, 5 or
8, depending upon the conditions used, but each in high yield
(Scheme 1).1–3 If one equivalent of the silyllithium reagent is
used at 278 °C, and the mixture is quenched at 278 °C, the
product is the acylsilane 2. If the same reaction mixture is first
warmed to 220 °C before quenching, the product is the
enediamine 5. If two equivalents of the silyllithium reagent are
used and the mixture allowed to come to 220 °C before
quenching, the product is the a-silylamine 8. We provided
evidence that the initial reaction at 278 °C gives the tetrahedral
intermediate 3, and that when this is warmed Brook rearrange-
ment and elimination of silanoxide gives a species which we
draw for simplicity as a carbene 6, but about which we have no
detailed information.4 The same Brook rearrangement creates a
nucleophile which can react with the carbene to give the
intermediate 4, leading to the enediamine 5. If, instead, there is
a second equivalent of the silyllithium reagent, it reacts with the
carbene to give the a-aminolithium species 7, which is
quenched by a proton source to give the amine 8.
Similar reactions take place with other N,N-dialkylamides,
the carbene or carbenoid intermediates can be trapped by other
nucleophiles such as alkyl- and aryllithium reagents, and the a-
aminolithium species can be trapped with other electrophiles
than a proton when a phenyl group is present to stabilise them,
as in the acylation of the intermediate lithium reagent 9 giving
the phenylglycine 10 (Scheme 2), which we now add to the
alkylation reactions reported earlier. The silyl group in this new
reaction is removed during the aqueous workup, because it is
adjacent to the carbonyl group.
Proton transfer became the only pathway in the reaction with
a terminal alkene and one extra methylene unit (Scheme 4). The
amide 16 gave the a-silylamines 19 and 20, with the latter a
mixture of cis and trans isomers. All three of the products could
have come by proton transfer from the intermediate 17,
followed by protonation of the allyllithium species 18 on work
up.
Although alkyl halides react with silyllithium reagents,5 there
was some hope that both the amide group in the amide 21, and
the carbene derived from it, would react more rapidly than the
chloride. Cyclisation of the intermediate lithium reagent 22
Scheme 2 Reagents and conditions: i, PhMe2SiLi, 278 °C; ii, warm to 220
°C; iii, ClCO2Et.
Scheme 1 Reagents and conditions: i, PhMe2SiLi, 278 °C; ii, NH4Cl, H2O;
iii, warm to 220 °C.
Scheme 3 Reagents and conditions: i, 2.4 equiv. PhMe2SiLi, THF, 278 °C
? 220 °C; ii, NaHCO3, H2O.
198
CHEM. COMMUN., 2003, 198–199
This journal is © The Royal Society of Chemistry 2003