ied.[11b] This reaction may also occur through a photoinduced
electron-transfer mechanism[11c] and would result in the ra-
ACHTUNGTRENNUNGceACHTUNGTRENNUNGmic saturated ketones, which would ultimately give rac-3.
red. However, 1a was smoothly reacted to the desired 2H-
chromen-2-ols 2 when the photocyclization was performed
in toluene, setting the stage for exploring the photo assisted
Brønsted acid catalyzed transfer hydrogenation. Initial reac-
tions were carried out with different dihydropyridines 5a–e
and BINOL-derived phosphoric acid diester 4a, which had
proven to be a good catalyst in organocatalytic hydrogena-
tions of N-heterocycles. To our delight, not only was the
product 3a formed, but enantioselectivities of up to 18%
enantiomeric excess (ee) were obtained if dihydropyridine
5d was applied (Table 1, entry 4).
Furthermore, the chroman-2-ol could also be reduced by the
photoactivated Hantzsch ester yielding rac-3; 3) the photo-
oxidation and photocatalytic production of hydrogen from
Hantzsch esters 5 has been well established and could lead
to the undesired decomposition of the hydride donor;
4) next to their use as antioxidants, benzopyrylium cations
have been shown to be excellent electron acceptors, which
are involved in charge-transfer reactions. They can also act
as a source of free radicals, which could lead to undesired
side reactions.[11d]
In order to increase the enantioselectivity, several BINOL
derived phosphoric acids and the corresponding triflyl-
Given the above considerations we performed several test
reactions (Table 1). First, we started with the Brønsted acid
catalyzed reduction of chalcone 1a (R and R’=Ph) employ-
ing the Hantzsch ester 5 as the hydride source. Pleasingly,
no reduction was observed when the reaction was per-
formed in aromatic solvents at room temperature. Addition-
ally, no direct acid-catalyzed cyclization of 1a to 2a occur-
ACHTUNGTRENaNUGN mides were evaluated. From all catalysts tested, 4 f was the
best and yielded 97% of the desired chiral 4H-chromene 3a
with 73% ee (Table 1, entry 10). Interestingly, performing
the reaction with 2,4-diphenyl-2H-chromen-2-ol (2a) result-
ed in reduced enantiomeric excess (Table 1, entry 11). All
N-triflylphosphoramides resulted in lower enantioselection.
In order to ascertain that the light-driven reduction by pho-
toexcited dihydropyridines had not occurred we performed
a reaction without catalyst under otherwise identical condi-
tions. Fortunately, no considerable product formation was
observed (Table 1, entry 13). However, at higher light inten-
sity and with prolonged reaction times this reaction occurs
leading to considerable loss in enantioselectivity. The pres-
ence of oxygen results in the photodecomposition of the di-
hydropyridine. In order to further improve the enantioselec-
tivity, the reaction was performed at lower temperatures.
Pleasingly, decreasing the reaction temperature to ꢀ208C
improved the enantioselectivity significantly and the product
was isolated in 69% yield and with an excellent enantiomer-
ic excess of 94% ee (Table 1, entry 14).
Table 1. Optimization of the Brønsted acid catalyzed photocyclization re-
duction reaction.[a]
Further optimization, including the addition of molecular
sieves for trapping any water produced, performing the re-
action in different solvents[12] or using lower temperatures
did not improve the results. With the optimal reaction con-
ditions in hand the scope of the asymmetric Brønsted acid
catalyzed photocyclization–reduction sequence was explored
(Table 2).
In general a range of chalcone derivatives 1 with various
substitution patterns could be applied and the 4H-chromene
derivatives 3a–o were isolated in good to excellent yields
and enantioselectivities (Table 2, entries 1–15). It is worth
mentioning that the newly developed protocol is not only
the first example of a catalytic asymmetric reduction of
pyrylium ions, but also the first example in which 4H-chro-
menes are obtained in organocatalytic enantioselective fash-
ion.
In order to shed light on the reaction mechanism of this
new light driven Brønsted acid catalyzed hydrogenation of
benzopyrylium ions, we performed a series of experiments.
In addition to the overall transformation (1!3), we also ex-
amined the individual cyclization (1!2) and hydrogenation
(2!3) steps of the reaction sequence as our experiments in-
dicated a more complex mechanistic scenario (Figure 1).
Under the optimal reaction conditions 1) the phosphoric
T
[8C]
4
5
Yield
[%][d]
ee
[%][e]
1
2
3
4
5
6
7
8
9
10
11[b]
12
13
14
15[b]
16[c]
17
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
RT
ꢀ20
ꢀ20
ꢀ20
ꢀ45
4a
4a
4a
4a
4a
4b
4c
4d
4e
4 f
4g
4 f
–
5a
5b
5c
5d
5e
5d
5d
5d
5d
5d
5d
5d
5d
5d
5d
5d
5d
57
55
41
70
67
50
83
77
87
97
91
92
3
9
17
8
18
racemic
8
racemic
racemic
22
73
21
34
–
94
55
80
94
4 f
4 f
4 f
4 f
69
84
89
42
[a] Reactions were performed with chalcone 1a at 0.06m concentration,
dihydropyridines 5 (1.3 equiv), and 4 (5 mol%). The solution was irradi-
ated for 1 h at room temperature or for 12 h at ꢀ208C. Irradiation was
carried out in a Rayonet reactor at 300 nm or with a TQ 150 high pres-
sure mercury lamp, lꢁ300 nm. [b] Reactions performed with preformed
2,4-diphenyl-2H-chromen-2-ol. [c] Addition of 3 ꢂ MS. [d] Yield of iso-
lated product after column chromatography. [e] Enantiomeric excess was
determined by HPLC on a chiral stationary phase.
&
2
&
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!