LETTER
Reversibility in Lewis-acid Promoted Reactions of N-Arylcinnamamides
899
compounds 2a and 5 were inseparable, as were 6 and 7. rial’ was in fact 1a. As expected, there was no evidence of
Presumably compound 7 is formed by dearylation of 6 cyclisation via the aromatic ring of the benzyl group. In no
followed by addition of a second molecule of chloro- instance was there any chlorine incorporation into any of
benzene. This clearly implies the reversible Friedel– these products, so that a reversible addition/elimination of
Crafts addition of chlorobenzene to compound 1.
solvent followed by cyclisation can be excluded.
In the search for milder reaction conditions, we conducted The microwave reactions were carried out for 15 minutes
these reactions in a microwave reactor. Complete conver- in order to to measure the conversion. The observed rate
sion was observed in only 15 minutes, with a similar ratio of reaction was H > Ph ca. = Me > Bn and the rate of
of intermolecular to intramolecular reaction being ob- cyclisation Ph ca. = Me > H ca. = Bn. The similar reac-
served to that shown above. However, in this case tivity of 1b and 1c is not entirely surprising, as the appar-
compound 7 was notable by its absence. Reaction of the ent size of Ph and Me at the nitrogen are not dissimilar.
mixture of 6 and 7 with excess aluminium chloride gave We attribute the lower reactivity of 1d to the increased
only a complex mixture.
steric hindrance to the approach of Lewis acid to the
amide oxygen. Cyclisation must precede debenzylation
since no intermolecular reaction with solvent is observed.
Compound 1a is the exception, in that it is locked into
conformer B. Complexation of aluminium chloride to
compounds 1 is expected to increase the rotational barrier
further (structure C, Scheme 3). However, this may be an
oversimplification, as shown by the varied structures of
amide aluminium complexes.10
With bismuth chloride, the reaction was much slower,
taking 4 days to reach completion, but gave only 2a (47%
isolated yield). The lack of intermolecular reaction in this
case may be due to the reduced ability of bismuth chloride
to form a complex with chlorobenzene prior to the
Friedel–Crafts reaction.8 Microwave reactions with bis-
muth chloride in chlorobenzene were not attempted since
this compound is not readily soluble, and we were con-
cerned about the safety of the procedure. No conversion
was observed when bismuth chloride was used in THF
under similar conditions.
Table 1 Preparation and Cyclisation of Compounds 1b–d under a
Range of Conditions
Substrate
1.75 AlCl3, PhCl, 2 BiCl3, PhCl, 2 AlCl3, PhCl, mi-
We anticipated that these reactions might be sensitive to
conformational effects, and that introduction of bulky
groups onto nitrogen might favour conformation A, which
would be able to cyclise more readily than B (Scheme 3).
This aspect has not been addressed by previous studies,9
so compounds 1b–d were prepared and cyclised under a
range of conditions. The results of these reactions are
shown in Table 1.
90 °C, 2 h
4 d, 90 °C
crowave, 15 min
2b, 85%
2b,
Ph
Ph
53% conversion
N
O
N
O
O
O
CH3
CH3
1b
1c
1d
2b, 81%
2c, 78%
2d, 41%
2c, 59%
2c,
Ph
Ph
60% conversion
Ph
Ph
Ph
N
N
O
R
R
N
O
N
OAlCl3
Ph
Ph
N
R
O
2a, 45%
2d, 31%
2d,
Ph
Ph
33% conversion
A
B
C
N
O
N
Scheme 3
Bn
Bn
The first notable point is the lack of intermolecular
Friedel–Crafts addition with these substrates. In fact, a
trace of this product was observed in the NMR spectrum
of the crude reaction mixture for the reaction of 1b with
aluminium chloride. Cyclisation of 1b and 1c was clean
and high yielding, with no significant by-products being
formed. Reaction of 1d was more complex. With alumin-
ium chloride, clean cyclisation was observed, although
the isolated yield was not high. With bismuth chloride, a
significant amount of debenzylated product 2a was ob-
tained, with the desired compound being the minor con-
stituent of the mixture. Conversion under these conditions
was incomplete (ca. 90%), but the residual ‘starting mate-
With three equivalents of aluminium chloride (PhCl, 90
ºC, 2 h), these three substrates gave compounds 3b (67%),
3c (61%) and 3d (62%) cleanly.
Since there is a relatively narrow window between no re-
action (up to 1.5 equiv) and dearylation of the initial cy-
clisation products (3 equiv), the reactions with aluminium
chloride are quite capricious. Poor quality aluminium
chloride naturally gives poor conversions, while carrying
out the reactions on a small scale requires care to exclude
all traces of water for reproducible results to be obtained.
In contrast, bismuth chloride is more reliable, so that de-
spite the longer reaction times, we recommend its use.
Synlett 2004, No. 5, 898–900 © Thieme Stuttgart · New York