Communication
by the addition of Cs2CO3; changing the base to K3PO4 also
gave an almost complete conversion (Table 1, entries 2 and
3).[7] On the other hand, using K2CO3 led to a good yield and
LiOtBu gave only fair results (Table 1, entries 4 and 5). Hydrox-
ide bases were then tested, LiOH resulted in a better conver-
sion than that of both NaOH and CsOH (Table 1, entry 6 versus
entries 7 and 8).[8] No reaction occurred when an organic base
(pyridine) was used or when the reaction was conducted in
toluene, the starting material being quantitatively recovered in
both cases (Table 1, entries 9 and 10). A low yield was obtained
when the reaction was performed in acetonitrile under reflux,
whereas on changing the solvent to DMSO a good yield was
achieved (Table 1, entries 11 and 12). Cs2CO3 was chosen for
further exemplification studies to allow a comparison of our re-
sults with those in the literature. Indeed, Cs2CO3 (in association
with DMF) is very frequently used in related metal-catalyzed
CÀO coupling reactions.
For these two entries, conducting the reactions at 908C also
greatly reduced the required reaction times (Table 2, entries 5’
and 6’). On using 1-iodo-2-nitrobenzene, a satisfying yield of 4
was obtained at 908C (Table 2, entry 7).[11] Importantly, no trace
of the desired product was detected on reaction of 1-iodo-3-
nitrobenzene with 3,5-dimethylphenol at this temperature,
demonstrating that an appropriate substitution pattern (ortho
and para versus meta), typical of addition–elimination SNAr re-
actions of aryl halides, is crucial for the reaction to proceed.
In addition, aryl halides substituted by the cyano group
were also tested. The reactions of 3,5-dimethylphenol with
4-iodo-, 4-bromo-, and 4-chlorobenzonitrile furnished diaryl
ether 5 in good yields at 1208C (Table 2, entries 8–10). The dif-
ference in reactivity between the aryl halides was clearly high-
lighted when the temperature was lowered to 908C; 4-iodo-
benzonitrile gave
a significantly lower conversion than
4-bromo- and 4-chlorobenzonitrile (Table 2, entries 8’–10’). Phe-
nols substituted by various electron-withdrawing and electron-
donating groups could be coupled with 4-iodobenzonitrile to
give diaryl ethers 6–9 in yields ranging from 53 to 97%, dem-
onstrating the generality of this method (Table 3, entries 11–
14). Moreover, when 2-iodobenzonitrile was used, an excellent
97% yield of diaryl ether 10 was obtained (Table 2, entry 15).
The reactions of 4-haloacetophenones with 3,5-dimethylphe-
nol were also investigated. It was observed that an elevated
temperature (1458C) is necessary to furnish the desired diaryl
ether 11 in moderate to excellent yields (Table 2, entries 16–
18). When the reactions were repeated, at 1208C, with the
three aryl halides, the best conversion was obtained in the
case of 4-bromoacetophenone, whereas 4-iodo- and 4-chloro-
acetophenone gave modest results (Table 2, entries 16’–18’).
Although these results seem to confirm preliminary reports
on SNAr reactions of 1-iodo-4-nitrobenzene,[6] we controlled
the purity of the reagents to detect the possibility of catalysis
by trace metal contaminants.[9] Cs2CO3 was purchased from
Sigma–Aldrich (99.995% trace metals basis) and Alfa Aesar
(99.994% trace metals basis). For the latter, a certificate of anal-
ysis was available and indicated that metals known to catalyze
such reactions (mainly palladium, copper, and rhodium) were
not detected in measurable amounts by inductively coupled
plasma (ICP) analysis.[10] In addition, both reaction partners
(3,5-dimethylphenol and 1-iodo-4-nitrobenzene) were carefully
purified by column chromatography, and reactions were con-
ducted in new glassware by using magnetic stirrers treated
with aqua regia (nitro-hydrochloric acid) prior to use. Even
with these precautions, the reactivity remained essentially the
same for each individual reaction.
The trifluoromethyl group,
a weak electron-withdrawing
group, was also tested. The reactions of 4-halobenzotrifluorides
with 3,5-dimethylphenol gave the corresponding product 12
in yields ranging from 60 to 91% at a surprisingly “low” tem-
perature (1458C), despite the weak activation from the trifluor-
omethyl group (Table 2, entries 19–21). Even more surprising
was the fact that the relative reactivity of the aryl halides de-
creased in the order IꢀBr>Cl,[12] which is not in perfect agree-
ment with the literature.[1,13] Moreover, the tendency was the
same when the reactions were conducted at 1208C (Table 2,
entries 19’–21’). Finally, a 65% yield of diaryl ether 13 was ob-
tained by the reaction of 3,5-dimethylphenol with 2-iodoben-
zotrifluoride (Table 2, entry 22). It is noteworthy that we never
observed reduction products of the aryl iodides under these
conditions, with the exception of the reaction using 1-iodo-2-
nitrobenzene.
Next, we attempted to extend the scope of this method to
various phenols and activated aryl iodides. It is noteworthy
that except one previous report,[6b] aryl iodides with a substitu-
ent other than NO2 have never been used in SNAr reactions.
For each type of substituted aryl iodide, a comparison of its re-
activity with the corresponding bromide and chloride was also
undertaken to verify the expected order of reactivity of aryl
halides; I<BrꢀCl according to the literature. Because of the
high reactivity of nitro-substituted aryl halides, reactions of
these substrates were carried out at room temperature. We
first observed that the reaction of 1-iodo-4-nitrobenzene with
3,5-dimethylphenol proceeded to completion at room temper-
ature, although the reaction time was very long (Table 2,
entry 1). As previously observed, increasing the temperature to
908C greatly reduced the required reaction time (Table 2,
entry 1’). Reactions of 3,5-dimethylphenol with 1-bromo-,
1-chloro-, and 1-fluoro-4-nitrobenzene also furnished 1 in ex-
cellent yields at room temperature and in acceptable reaction
times for 1-chloro, and 1-fluoro-4-nitrobenzene (Table 2, en-
tries 2–4). Furthermore, the order of reactivity of the four aryl
halides is in good agreement with the literature for SNAr reac-
tions. The reactions of phenol and 4-fluorophenol with 1-iodo-
4-nitrobenzene under similar conditions also provided diaryl
ethers 2 and 3 (Table 2, entries 5 and 6) in quantitative yields.
This methodology was next extended to nitrogen nucleo-
philes (Table 3). We tested the reactivity of azoles, with pyra-
zole being chosen as the model substrate. The reaction of pyr-
azole with 1-iodo-4-nitrobenzene at 1208C provided N-aryl
azole 14 in excellent yield (Table 3, entry 1). It was also possible
to generate product 15 in very good yields from the corre-
sponding 4-halobenzonitriles at 1458C (Table 3, entries 2–4),
and product 16 in fair to good yields by using 4-halobenzotri-
fluorides (Table 3, entries 5–7). For the latter, 4-chlorobenzotri-
fluoride again proved to be less reactive than its iodo- and
Chem. Eur. J. 2014, 20, 5231 – 5236
5232
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim