Organic Letters
Letter
The 1-aryl- and 1-alkylketones used in this study were readily
accessed from commercially available phenylacetic acids. For
example, the synthesis of 1-arylketone 3a, used to optimize the
one-pot processes, was prepared in two steps (Scheme 2). 3-
O-arylation step (e.g., FeCl3 with 99.9% purity contains 31.1
ppm of Cu).8 However, this presented an opportunity to
develop a one-pot process using a single iron complex in which
the residual ppm levels of copper could be used to catalyze the
cyclization step. The one-pot process was repeated using only
FeCl3, with either 97 or 99.9% purity, which gave benzo[b]furan
4a in 48 and 44% yields, respectively (Table 1, entries 2 and 3).
To verify that ppm loading of copper could catalyze the
heterocyclization of 1-(2-haloaryl)ketones, we performed the
single-step cyclization with the iodide of 3a using CuI (0.001
mol %, 14 ppm).9 This gave benzo[b]furan 4a in 55% yield.
Thus, in the one-pot process, we believe that whereas iron(III)
performs the iodination step, the ppm loading of copper is
responsible for the intramolecular O-arylation step. To improve
the yield of this one-pot process, catalyst loading was next
investigated. Using FeCl3 (99.9% purity) at 5 mol % loading
gave 4a in 60% yield (entry 4). Increasing the catalyst loading
further (entry 5) did not lead to substantially higher yields, and
so 5 mol % was deemed the optimal amount for subsequent
studies.
To determine whether iron(III) could catalyze both
halogenation and cyclization steps, the one-pot process was
performed using an ultrapure iron(III) salt. As ultrapure
iron(III) chloride was not commercially available, iron(III)
nitrate nonahydrate (99.999% purity), which contains no
copper, was investigated. Initial trials demonstrated that in
combination with [BMIM]NTf2, Fe(NO3)3·9H2O was an
effective Lewis acid for regioselective halogenation, with
complete conversion to the iodide intermediate after 5 h.
Then, using only this complex (5 mol %) for the entire one-pot
process gave benzo[b]furan 4a in 55% yield (Table 1, entry 6).
To verify that iron(III) was responsible for catalysis of the
cyclization step, control reactions were performed. For example,
a reaction in which the iodide of 3a was treated with DMEDA
(10 mol %), Cs2CO3 in a mixture of toluene and water, under
standard cyclization conditions (130 °C) resulted in no
conversion to benzo[b]furan 4a. This result confirmed that
the C−O bond forming step (entry 6) is catalyzed by iron(III)
and not due to the introduction of ppm amounts of copper with
the addition of reagents and solvents (e.g., water) during the
second step.
a
Scheme 2. Synthesis of 1-Arylketone 3a
a
Isolated yields.
Methoxyphenylacetic acid (1) was converted to Weinreb amide
2 under standard conditions, and subsequent reaction with 4-
chlorophenylmagnesium bromide allowed the efficient, scalable
synthesis of 3a.
Initial development of a one-pot synthesis of benzo[b]furans
focused on iron(III)-catalyzed halogenation and copper(I)-
catalyzed cyclization of 1-arylketone 3a (Table 1). Iron
Table 1. Optimization Studies for the Iron- and Copper-
a
Catalyzed Synthesis of Benzo[b]furan 4a
catalyst loading
(mol %)
catalyst purity
(%)
yield
b
entry
catalyst
(%)
c
1
FeCl3 + CuI
FeCl3
FeCl3
FeCl3
FeCl3
2.5 + 10
97
97
99.9
99.9
99.9
99.999
59
48
44
60
63
55
2
3
4
5
2.5
2.5
5
10
5
Having developed three different one-pot processes involving
various copper loadings (zero, ppm, and 10 mol %), the scope of
these with a range of 1-arylketones was next studied. Initially, the
one-process involving Fe(NO3)3·9H2O that uses iron(III) to
catalyze both steps was performed with electron-deficient (p-
CF3Ph, 3e) and electron-rich (p-MeOPh, 3j) 1-arylketone
substrates (Scheme 3). This gave benzo[b]furan 4e and the
natural product, corsifuran C (4j) (from Corsinia coriandrina),10
in 55 and 41% yields, respectively. These results confirm that a
single iron complex can be used to catalyze both steps and access
various benzo[b]furans. However, due to the moderate yields
(also, 55% for 4a), the requirement of an ultrapure metal salt,
and the avoidance of copper contaminants, we felt that the other
one-pot processes that use standard grades of iron(III) chloride
would be more synthetically useful.
d
6
Fe(NO3)3·9H2O
a
All reactions used [BMIM]NTf2 (3 times the amount of Fe catalyst).
b
c
Isolated yields. Conducted using CuI (10 mol %) and DMEDA (20
d
mol %). Reaction times for each step was 5 and 20 h.
triflimide (2.5 mol %), generated from iron(III) chloride and
the ionic liquid, [BMIM]NTf2, was used for activation of N-
iodosuccinimide (NIS) and subsequent iodination of 3a (entry
1). 1H NMR studies confirmed that despite the meta-
arylethanone substituent, iron(III)-catalyzed iodination pro-
ceeded exclusively at the position para to the methoxy activating
group. The second stage of the one-pot process was conducted
using copper(I) iodide (10 mol %) and DMEDA (20 mol %),
which gave benzo[b]furan 4a in 59% yield.
The one-pot process involving FeCl3 (99.9% purity) to
catalyze the iodination step and then residual copper (31.1 ppm)
to catalyze the O-heterocyclization was next investigated for the
preparation of biologically important 2-arylbenzo[b]furans
(Scheme 3). 1-Arylketone substrates bearing electron-deficient
and electron-rich aryl groups, as well as various o-, m-, and p-
substituents were found to be effective substrates, allowing the
As Bolm and co-workers reported that iron(III) could be used
for the O-cyclization of 1-(2-bromoaryl)ketones,5f we next
investigated whether a single iron salt could be used to perform
both iodination and cyclization steps during a one-pot process.
This goal became more challenging when we were made aware
that most grades of commercial FeCl3 contain various levels of
metal contaminants, including sufficient copper to catalyze the
B
Org. Lett. XXXX, XXX, XXX−XXX