7316
D. Astley et al. / Tetrahedron Letters 45(2004) 7315–7317
2. Results and discussion
reactivity of 1c with butyllithium, which has been
reported to afford 2,6-disubstituted aryloxazolines
after orthometallation and combination with an
electrophile.15
The reactions of the aryl chloride 1c with lithium rea-
gents and Grignard reagents were carried out in reflux-
ing diethyl ether using an excess of the organometallic
reagent (Eq. 1). Conditions and yields are given in
Table 1.
A direct comparison of this uncatalysed approach with
the Mn-catalysed method of Cahiez et al. can only be
made for PhMgBr and the conditions for the two meth-
ods are given in entries 5 and 6. For this combination of
aryl halide and Grignard reagent, a yield of 63% was
reported using the catalytic method. This can be deemed
more favourable than our yield of 40%, although both
methods required a considerable excess of reagent and
a lengthy reaction time. However, the conditions
employed by us for the lithium reagents are superior
to those used for the Grignard reagents with or without
catalysis. As indicated above, extensive use has been
made of compound 1b and derivatives in synthesis.5–7
Examples of relevance to this work include the reactions
of tolyl16 and anisyl17,18 Grignard reagents with 1b as
key steps in the synthesis of aryl substituted tetrahydro-
isoquinoline derivatives, which were required for screen-
ing as dopamine antagonists17 or angiotensin II receptor
antagonists.16,18 It is, therefore, extremely notable
that 1c, which is considerably more favourable than 1b
on availability and economic grounds, couples directly
with lithium reagents under surprisingly convenient
conditions.
ð1Þ
In all cases the coupled products were obtained in fair to
reasonable yields, although the reactions with the Grig-
nard reagents required a larger excess of reagent and
longer reaction times. No evidence was obtained for
either metal–halogen exchange or orthometallation
processes. If metal–halogen exchange had occurred,
one would have expected the unsubstituted derivative,
2-phenyl-4,4-dimethyl-2-oxazoline, to have formed
upon aqueous work up by overall displacement of Cl
from 1c with H. However, it was determined by TLC
monitoring of experiments that none of this unsubsti-
tuted product was formed. This is in contrast with the
chemistry of the bromo derivative (1d) and the iodo ana-
logue, which have recently been observed to undergo
metal–halogen exchange with Ph2CuLi.14 In experi-
ments, which were carried out to determine whether
orthometallation had occurred, benzoyl chloride was
added to the reaction mixture, subsequent to the addi-
tion of the phenyllithium reagent. In all experiments,
substitution of Cl with Ph was determined to be the pri-
mary product. The only other product formed to a sig-
nificant extent was benzophenone, which arised from
coupling of benzoyl chloride with excess phenyllithium.
This presents an interesting contrast to the observed
A comparison of this work with the reported reactivity
of the 3-chloro isomer of 1c, is also warranted. This
meta substituted aryl chloride reacts with PhLi to afford
the corresponding meta-substituted biphenyloxazoline
(by overall displacement of Cl with Ph) but also affords
the isomeric ortho-substituted biphenyloxazoline that
was formed in this study (Table 1, entry 1) in a reaction,
which occurs via a benzyne intermediate.19 By carrying
out the reaction at ꢀ78°C in pentane, product yields
of 24% for the meta-substituted product and 48% for
the ortho-substituted product were obtained as deter-
mined by gas chromatography.20
As seen in Table 1, for both lithium and Grignard rea-
gents, slightly better yields were observed for the ortho-
and meta-tolyl reagents than for the para-substituted
and unsubstituted examples. The Grignard reagents
seem to be more sensitive to this effect, with the best
yield we obtained for coupling to the chloro compound
1c, being for 3-CH3C6H4MgBr (78% entry 8). For com-
parative purposes, we also carried out the reactions of
the fluoro compound 1a with organolithium reagents
under the same conditions. As expected, due to the
greater reactivity of Fꢀ as a leaving group in the SnAr
reaction, it was observed that the fluoro derivative does
generally give slightly better yields than were obtained
for the chloro derivative. However, again it was noticed
that under these conditions, better yields were observed
with the 2-tolyl and 3-tolyl reagents. We suspect that the
reasons for this preference are likely to be connected
with the orientation of coordination that is expected to
occur between the organometallic reagent and the oxa-
zoline group prior to coupling.4 Thus, it may be sug-
gested that the different positions of the methyl group
Table 1. Product yields and conditions for the formation of biphenyl
derivatives according to Eq. 1
Entry Ar M
X
Ratioa Conditionb Yield
1
PhLi
Cl 1:4
Cl 1:4
Cl 1:4
Cl 1:4
Cl 1:8
Cl 1:6
1h
1h
54
56
61
40
40
63c
64
78
37
55
86
71
59
2
2-CH3C6H4Li
3-CH3C6H4Li
4-CH3C6H4Li
PhMgBr
3
1h
1h
4
5
18h
6c
7
PhMgBr
24hc
18h
2-CH3C6H4MgBr Cl 1:8
3-CH3C6H4MgBr Cl 1:8
4-CH3C6H4MgBr Cl 1:8
8
18h
18h
9
10
11
12
13
PhLi
F
F
F
F
1:4
1:4
1:4
1:4
15min
15min
15min
15min
2-CH3C6H4Li
3-CH3C6H4Li
4-CH3C6H4Li
a Ratio of 1c:ArM.
b All reactions carried out in refluxing diethyl ether.
c Entry 6 represents results given in Ref. 13 using Mn catalysis.