Fang and Li
TABLE 1. Optimization of Reaction Conditions in the Synthesis of
5a
Buchwald et al. reported the first examples of copper-catalyzed
synthesis of diaryl ethers from the reaction of aryl bromides or
iodides with phenols using (CuOTf)2‚PhH as the catalyst and
Cs2CO3 as the base.5 With the help of an appropriate ligand,
the reaction could be conducted under milder conditions.6 This
was then nicely applied into the total synthesis of the natural
product antitumor agent K-13 as reported by Ma et al.7 In the
meantime, O-arylation of aliphatic alcohols could also be
performed with high efficiency by the catalysis of Cu(I).8 More
recently, this methodology was extended to the synthesis of
substituted benzoxazoles9 and benzo[b]furans10 via intramo-
lecular O-arylation of amides or ketones.
Of particular interest to us was the use of ketone enolates as
nucleophiles in O-arylation, which significantly broadened its
scope of application, although so far only the five-membered
ring closure leading to benzo[b]furans has been successful.10
We recently reported the first examples of copper-catalyzed
intramolecular O-vinylation of carbonyl compounds with vinyl
bromides, among which five-, six- and even seven-membered
cyclic alkenyl ethers could be efficiently synthesized.3f We
envisioned that our success in O-vinylation might be extended
to O-arylation, which would offer a convenient route to the
synthesis of 4H-1-benzopyran derivatives via the six-membered
ring closure. 4H-1-Benzopyrans are an important class of
naturally occurring compounds with biological activities.11
However, unlike 2H-1-benzopyrans, no general method has so
far been reported for their synthesis.11,12
copper
source
entry
base
ligandb
yield (%)c
1
2
3
4
5
6
7
8
9
10
11
12
13
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
K2CO3
A
A
A
A
A
A
A
A
A
B
C
D
0
Cu
trace
30
10
99
40
0
0
0
<5
15
<5
26
CuCl
CuCl2
CuI
CuI
CuI
CuI
CuI
CuI
CuI
CuI
CuI
DABCO
NaOtBu
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
a Reaction conditions: 2a (0.1 M in THF), copper catalyst (10 mol %),
ligand (20 mol %), reflux, 1 h. b A: N,N′-dimethylethylenediamine
(DMEDA). B: 1,10-phenanthroline. C: L-proline, D: PPh3. c Isolated yield
based on 2a.
factors could play an important role in the two types of
cyclizations, the above different reactivities might also be
attributed to the much easier enolization of the deoxybenzoins
in the five-membered ring closure cases. We envisioned that,
by changing the substrates from simple ketones to the more
readily enolizable 1,3-dicarbonyls, such as 2a or 3a, the
formation of 4H-1-benzopyrans might be possible as a similar
situation was observed in our O-vinylation cases.3f Thus, â-keto
ester 2a was subjected to the treatment of CuI (0.1 equiv),
DMEDA (0.2 equiv), and K2CO3 (2 equiv) in THF at refluxing
temperature for 1 h. To our delight, the corresponding cycliza-
tion product 4H-1-benzopyran 5a was isolated in 40% yield.
We then used â-keto ester 2a as the model compound to
optimize the reaction conditions (eq 1), and the results are
summarized in Table 1.
Results and Discussion
We chose ketones 1a,b and â-keto esters 2a and 3a as the
model substrates to explore the possible formation of 4H-1-
benzopyrans. These four compounds were readily prepared from
the common starting material 2-bromobenzyl bromide 4 ac-
cording to the conventional methods.13 We first tested ketones
1a and 1b according to Chen’s procedure.10a However, no
reaction occurred. Changing the base from K3PO4 to the much
stronger NaOtBu did not show any improvement. We then
screened various bases (Na2CO3, Cs2CO3, NaOtBu, Et3N),
ligands (N,N′-dimethylethylenediamine (DMEDA), L-proline, 1,-
10-phenanthroline, and PPh3), and copper sources (Cu, CuI,
CuCl, CuOAc) in different solvents (THF, dioxane, toluene, and
DMF) at refluxing temperature. To our disappointment, under
all the experimental conditions screened, no expected cyclization
products could be obtained while, in most cases, the starting
materials simply remained unchanged.
As can be seen in Table 1, the careful choice of copper source
was critical for the high yield of product in a short time. As
expected, no coupling took place when the reaction was carried
out in the absence of the metal catalyst (entry 1, Table 1).
Among the readily available copper compounds screened,
including Cu, CuCl, CuCl2, and CuI, the air-stable CuI had the
best performance (entries 2-5, Table 1). In addition to a copper
source, bases were also found to have a dramatic influence on
the reaction outcome. For example, the reaction was much faster
with Cs2CO3 as the base (entry 5, Table 1) than with K2CO3
(entry 6, Table 1). An organic base, such as DABCO or NaOt-
Bu, completely inhibited the reaction (entries 7 and 8, Table
1), while no reaction occurred in the absence of a base (entry
9, Table 1). A brief study of the effect of the ligand was also
carried out. Dramatic differences in the yield of 5a were
observed when DMEDA (99%), 1,10-phenanthroline (<5%),
L-proline (15%), PPh3 (<5%), or no ligand (26%) was used
(entries 5 and 10-13, Table 1).
Apparently the six-membered ring closure was much more
difficult than the five-membered ring closure. Although steric
(11) Ellis, G. P., Ed. Chromenes, Chromanones and Chromones;
Wiley: New York, 1977.
(12) For recent examples on the synthesis of 4H-1-benzopyran deriva-
tives, see: (a) Yavari, I.; Anary-Abbasinejad, M.; Alizadeh, A.; Hossaini,
Z. Tetrahedron 2003, 59, 1289. (b) Yavari, I.; Djahaniani, H.; Nasiri, F.
Tetrahedron 2003, 59, 9409. (c) Miao, H.; Yang, Z. Org. Lett. 2000, 2,
1765. (d) Kumar, P.; Bodas, M. S. Org. Lett. 2000, 2, 3821.
(13) (a) Cooke, M. P., Jr.; Windener, R. K. J. Org. Chem. 1987, 52,
1381. (b) Trost, B. M.; Chen, D. W. C. J. Am. Chem. Soc. 1996, 118, 12541.
(c) Huckin, S. N.; Weiler, L. J. Am. Chem. Soc. 1974, 96, 1082.
On the basis of the above results, we concluded that the
optimized conditions were CuI (10 mol %) as the catalyst,
DMEDA (20 mol %) as the ligand, Cs2CO3 (2 equiv) as the
base, and THF as the solvent. We then prepared a variety of
6428 J. Org. Chem., Vol. 71, No. 17, 2006