These experiments allow the conclusion that the ortho-
lithiated species 5 generated under these conditions is formed
efficiently, does not undergo self-condensation with starting
4a to form a benzophenone product as is the case for the
ortho-lithiated N,N-dialkyl benzamide12 and, significantly,
is stable to room temperature and does not equilibrate with
7 (k2 > k3).
Scheme 4
In contrast, when LDA deprotonation of 4b was carried
out at -20 °C,13 followed by addition of MeI or TMSCl at
this temperature, 10a and 10b were not formed and only
fluorenone (3) was isolated. This experiment suggests a slow
remote deprotonation step occurs to give 7, which has a short
lifetime and undergoes cyclization to 3 in a fast and
irreversible step. To corroborate this proposal and determine
the behavior of anion 7 directly, N,N-diisopropyl 2-(2′-
bromophenyl)benzamide (11) was subjected to metal-halogen
exchange (1.1 equiv of n-BuLi, THF, -78 °C, 5 min)
followed by CD3OD quench. Formation of 2′-deuterated 10c
was not observed and undeuterated fluorenone (3) was
obtained exclusively (89% yield).14 Combined, the above
results show that (1) the conversion of 5 to 7 does not occur
under the RLi/TMEDA and in situ LDA/TMSCl conditions
and (2) the LDA-mediated cyclization of independently
generated 7 to 3 is faster than the isomerization of 7 to 5 (k4
> k-3).
To demonstrate that the kinetically deprotonated anion 5
may be converted into the thermodynamically favored
carbinolamine oxide 8 through proton transfer involving DIA/
LDA, we adopted the poignant experiments of Seebach.15
Thus, biarylamide 4a was treated first with s-BuLi/TMEDA
(1.1 equiv) at -78 °C to form the stable ortho-lithiated
species 5 (Vide supra). Stirring at -78 °C was maintained
for 1 h, and 10 mol % of DIA was then added. After warming
to room temperature and normal workup, 3 was isolated in
30% yield with 70% recovery of starting material. This result
is consistent with reprotonation of 5 by DIA affording 4a,
which in turn can be deprotonated by the generated residual
LDA and give, either directly or via the postulated complex
C, remote metalated species 7. The subsequent cyclization
of 7 leading to 8 is fast and irreversible, and the LDA-
mediated equilibrium 5 h 7 is shifted toward the formation
of 7.
of alkyllithium bases to the generated 3.16 If, however, only
8 is present in solution, expectation that peri-metalated
species 12 will be formed with excess base is inferred from
our previous observations of the formation of the analogous
trianion 15a arising from 2-biphenyl carboxylic acid (2b)
using an excess of n-BuLi/t-BuOK6 to give products 16 and,
from the results of Demeter, for the closer related carbino-
lamine oxide 15b resulting from sequential reaction of 3 with
lithio N-methylpiperazide and n-BuLi followed by quench
with B(OBun)3 to afford 17 (Scheme 4).17 In the event,
reaction of 4a with 3 equiv of LDA at 0 °C followed by
addition of D2O provided the undeuterated fluorenone 3 in
an improved yield (92%) compared with that obtained in
the original experiments.4 Furthermore, when 2-4 equiv of
n-BuLi, s-BuLi, or s-BuLi/TMEDA were added to the
reaction mixture following the treatment with LDA, undeu-
terated fluorenone was also obtained in >90% yields.
The modest DMG power of the carbinolamine oxide
C(OLi)(NR2),18 as suggested by the vigorous conditions
(refluxing benzene) necessary to generate the analogous
species 15b, was considered to be the reason for the lack of
formation of peri-deuterated fluorenone 13 under our condi-
tions (3 equiv of LDA/0 °C/D2O quench). To provide further
evidence for the intermediacy of 15b and thus indirect
evidence for 12, the metalation conditions of Demeter were
reproduced, but the reaction was quenched with D2O.
1-Deutero-9H-fluoren-9-one (13) was obtained in 77% yield
and >95% d1-incorporation.14b Although conclusive evidence
has not been obtained here, we note that corroborative
support for carbinol amine intermediates in these DreM
We now sought to address the nature of the organometallic
species in the reaction mixture prior to hydrolysis. The
carbinolamine oxide 8 is potentially present in equilibrium
with fluorenone (3) by addition-elimination of LiNEt2
(Scheme 4). If this is the case, the formation of 9-butyl-9H-
fluoren-9-ol (14) is expected to some extent by the addition
(12) Beak, P.; Brown, R. A. J. Org. Chem. 1977, 42, 1823.
(13) Tilly, D. Ph.D. thesis, Universite´ du Maine, 2004.
(14) (a) Additionally, metalation of N,N-diethyl 3-deuterio-2-biphenyl
carboxamide (d0:d1 1:99) with LDA (3 equiv) at 0 °C followed by quench
with 2 M HCl furnished, in addition to unreacted starting material (8%,
d0:d1 1:99), 1-d1-fluorenone (d0:d1 78:22, 83% yield), indicative of some
erosion of d1-content due to a KIE. 4a-3d1 was prepared by amidation of
3-deutero-2-biphenyl carboxylic acid.6b (b) Isotope ratios were determined
(16) In a separate experiment, treatment of 3 with n-BuLi (1.1 equiv)
in THF at rt afforded 14 (51% yield) as expected.
(17) Demeter, A.; Tima´ri, G.; Kotschy, A.; Be´rces, T. Tetrahedron Lett.
1997, 38, 5219.
1
by H NMR and FIMS. The error is approximated to be (5%.
(18) (a) Comins, D. L. Synlett 1992, 615. (b) For the complexity of
lithio carbinolamine oxide solid-state structures, see: Clegg, W.; Liddle,
S. T.; Snaith, R.; Wheatley, A. E. H. New J. Chem. 1998, 1323, and
references therein.
(15) (a) Laube, T.; Dunitz, J. D.; Seebach, D. HelV. Chim. Acta 1985,
68, 1373. (b) For a review, see: Seebach, D. Angew. Chem., Int. Ed. Engl.
1988, 27, 1624.
70
Org. Lett., Vol. 12, No. 1, 2010