The Kostanecki reaction involves the formation of coumarins or chromones from readily available 2-hydroxyarylketones and an aliphatic anhydride in the presence of the corresponding sodium salt via O
-acylation and aldol condensation. An approach for the synthesis of 4-arylcoumarins via the Kostanecki reaction has already reported, however, that it suffers from limitations including the use of a large excess of reagents and harsh reaction conditions resulting in low yields [20
]. Indeed, 4-arylcoumarin 3a
was prepared from 2-hydroxybenzophenone 1a
and excess acetic anhydride as the solvent and reagent with sodium acetate by heating for 72 h at 190 °C (Table 1
, entry 2). The choice of base is of great importance in such aldol condensations; however, it has been sparsely investigated to date [21
]. Thus, we undertook a survey of organic bases to examine the feasibility of the Kostanecki reaction of 1a
. The initial attempt with piperidine was useless and only prone to liberate the acetate 2a
(entry 3). It is interesting to note that coumarin derivatives have been prepared by Knoevenagel condensation of 2-hydroxybenzaldehyde and benzophenone with active methylene compounds in the presence of piperidine [21
] and DBU [23
], respectively. After the investigation of various bases for the ring annulations, we found amidine bases were effective and DBU in acetonitrile was the best choice for the synthesis of 3a
] affording a 86% yield at 90 °C and 82% yield at ambient temperature, respectively, only after employing 3 equiv. of acetic anhydride (entries 6 and 7). There was no significant difference in yield regarding the reaction temperature. Using the optimized conditions, we next examined the substrate scope for this reaction. The other 2-hydroxybenzophenone derivatives 1b–1d
were readily prepared in 81%–
95% yields through the Fries rearrangement of the corresponding phenyl benzoates in the presence of aluminum chloride according to the known procedures [24
]. It was found that the substrates tested were uniformly converted to the corresponding 4-arylcoumarin derivatives (entries 8–
10). The spectral data of 3b
were in agreement with those reported in the literature [14
]. This result suggests that utilization of o
-hydroxybenzophenones as acyclic precursors would be an attractive route to access this type of molecules, with respect to ready availability and easy manipulation.
We then evaluated the same strategy using several 2-acyloxybenzophenone derivatives, aiming to synthesize 3,4-difunctionalized coumarins. It is interesting to note that the synthesis of 3,4-difunctionalized coumarins has rarely been investigated up to date, except for few examples [26
]. They frequently require expensive catalysts, laborious multi-step procedures, and long reaction times. The substrates, 2-acyloxyaryl derivatives 4a–4l
, could be easily prepared in 84%–94% yields by the treatment of the corresponding 2-hydroxyarylketones and acid anhydrides or acyl chlorides with Et3
N and a catalytic amount of DMAP in chloroform at ambient temperature for 2 h. Indeed, we achieved a very convenient synthesis of 3-ethyl-4-phenylcoumarin 5a
in 47% yield by simply mixing 4a
using 2 equiv. of DBU in acetonitrile (Table 2
, entry 1). Several 3,4-difunctionalized coumarins shown in Table 2
were successfully synthesized employing the established DBU-mediated reaction of 2-acyloxybenzophenones. Although the yields were moderate, a range of substrates containing electronically and sterically demanding groups on both the aryl rings proved to be compatible with this operation.
It is noteworthy that the substrates having an electron-withdrawing group such as fluorine smoothly provided the corresponding 4-arylcoumarins, which were previously difficult to obtain using the Pechmann protocol. Thus we provided a facile route to 3,4-difunctionalized coumarins via the intermolecular cyclization of 2-acyloxybenzophenone derivatives employing DBU. It appears that this methodology could be quite useful for the synthesis of highly functionalized 4-arylcoumarin materials with structural diversity.
Further, we thought that a one-pot preparation avoiding isolation of acyl intermediates would improve the utility of this method. Indeed, we observed the one-pot preparation of 3,4-diarylcoumarin 5m
in 76% yield, when 2-hydroxybenzophenone 1d
and phenylacetyl chloride were simply reacted in the presence of DBU, as shown in Scheme 1
One-pot preparation of 3,4-diarylcoumarin 5m.