S. G. Ouellet et al. / Tetrahedron Letters 50 (2009) 3776–3779
3777
non-protic solvent (Table 1, entries 6 and 7), however the selectiv-
ity was significantly lower. The best selectivities were obtained in
etheral solvents with dioxane providing the highest regioselectivi-
ty (38:1) and acceptable rates.
To extend the scope of this transformation, we next evaluated
the influence of various alcohols on the regioselectivity of the SNAr
reaction, again using 2,4-difluorobromobenzene as a model aryl
fluoride. Based on the results of Table 1, we selected t-BuOK as
the base of choice and subsequent bench scale reactions were con-
ducted in dioxane at room temperature.8 As shown in Table 2, a
wide variety of alcohols were added with excellent chemical yields
and with very high selectivities.
In general, regioselectivity increases with the sterics of the alk-
oxide used. When primary alcohols were used, selectivities of
about 50:1 were observed (entries 1, 2, and 7–9). The selectivity in-
creased to approximately 75:1 when secondary alkoxides were
used (entries 3 and 10) and finally selectivities exceeding 125: 1
were observed for sterically hindered nucleophiles such as t-BuOK
(entry 4). We also observed that the nucleophilicity/basicity of the
alkoxide impacted both the rate and the selectivity of this SNAr dis-
placement. This trend becomes obvious when we compare the en-
tries 2, 5, and 6. In the case of trifluoroethanol, a significantly
slower reaction was observed. Hexafluoro-2-propanol was found
to be unreactive, even under forcing conditions (entry 6, 100 °C,
for 2 days). Finally, we also demonstrated that high level of che-
moselectivity in the nucleophile could be obtained as shown in
entry 11 where an unprotected piperidine alcohol was added in
excellent chemical yield.9 These reactions are operationally simple
and are typically very clean. No byproducts were observed arising
from displacement of the bromide10 and we never observed more
than 1% of the double addition product under these conditions.
For preparative scale purposes, we decided to use THF as sol-
vent instead of dioxane and commercial availability of KOMe obvi-
ated the use of t-BuOK. However, the selectivity in THF was
somewhat lower than that observed in dioxane (Table 1, entry
2). To overcome this setback, we studied the effect of the reaction
time, temperature and the stoichiometry of the alkoxide used. We
observed a significant improvement in selectivity as we increased
the amount of alcohol used (Eq. 1 vs 2). The improved selectivity
arises from the favored consumption of the undesired isomer (B)
which reacts faster in a second addition than the desired isomer
(A). The bis-addition product formed could easily be removed by
fractional distillation to provide a highly enriched mixture of iso-
mer A.11
Table 2
Substrate scope for the selective SNAr reaction of 2,4-difluorobromobenzene with a
variety of alkoxides
ROH
F
F
RO
Br
F
F
OR
t-BuOK
dioxane
Br
Br
A
B
Entry
1
Product
Yield (%)
83
Ratioa (A:B)
54:1
MeO
F
F
Br
EtO
2
3
85
47:1
72:1
128:1
77:1
—
Br
i-PrO
F
82
Br
F
F
MeO
F
F
OMe
KOMe (2 equiv)
t-BuO
F
THF, 65 ºC
ð1Þ
ð2Þ
Br
Br
Br
(18 : 1)
< 0.5% bis-OMe
4
99
2 h
Br
A
B
F3C
O
O
F
F
F
F
5
60b
N.R.
87
F
F
MeO
Br
F
F
OMe
KOMe (5 equiv)
Br
THF, 65 ºC
18 h
Br
Br
(65 : 1)
9% bis-OMe
F3C
F3C
A
B
6
Br
Table 3 demonstrates the wide range of 1-substituted 2,4-diflu-
O
orobenzenes which can be successfully used to prepare 3-fluoro-
anisole derivatives in good chemical yields and with high levels
of regioselectivity. Both the aryl chloride and bromide were found
to undergo selective SNAr reaction using KOMe with excellent
selectivity (Table 3, entries 1 and 2). The reaction of 1-iodo-2,4-
difluorobenzene was slower and required more forcing conditions,
resulting in decreased selectivity (entry 3). Substrates bearing elec-
tron-withdrawing groups such as nitrile, ester, ketone, sulfone, and
nitro provided the desired anisole derivatives with good to excel-
lent regioselectivity.
Not surprisingly, the electronic characteristics of the substitu-
ent on the aryl ring had a profound influence on the reactivity. This
was confirmed in the case of methyl and phenyl-substituted
difluorobenzene substrates where no reactions were observed (en-
tries 9 and 10). We also found that more vigorous conditions
(100 °C) were necessary in order to promote the addition of KOMe
onto 2,4-difluorothioanisole (entry 7).
7
59:1
52:1
55:1
79:1
30:1
Br
O
8
98
Br
Ph
O
F
9
96
Br
O
F
10
11
98
N
Boc
HN
Br
O
F
92c
Br
Finally, we set out to evaluate the possibility of preparing differ-
entially substituted resorcinol derivatives by sequential addition of
alkoxides. The synthesis of fully differentially substituted resor-
cinol is still a nontrivial task. While interesting methods have been
Typical conditions: 3 equiv of alcohol, 2.9 equiv of t-BuOK in dioxane at rt for
6–24 h.
a
Determined by GC.
Reaction carried out at 45 °C with 4 equiv of alcohol for 24 h.
Reaction carried out in THF, HPLC assay yield.
b
c