Communication
Table 1. Screening and optimization.[a]
Table 2. Control reactions.
Entry
Conditions
Conversion [%][a] Yield [%][b] d.r.[a]
1
2
3
4
5
6
Standard conditions I
No light
No photocatalyst
33W CFL (72 h)
TfOH[c]
TsOH[d]
> 95
<5
<5
> 95
> 95
> 95
93
–
–
83
mixture
mixture
>19:1
–
–
>19:1
–
–
Entry[b] Catalyst Solvents Conversion [%][c] Yield[d] [%] d.r.[c]
[a] Determined by 1H NMR spectroscopy; [b] yields of compounds isolat-
ed by chromatographic purification; [c] 0.025 equivalent of TfOH;
[d] 0.025 equivalent of TsOH.
1
2
3
4
5
6
7
8
9[e]
10
11
3a
3a
3a
3a
3a
3a
3a
3b
3b
3c
[Ir][f]
[Ru][g]
3b
DCE
> 95
> 95
–
> 95
<5
> 95
> 95
> 95
> 95
> 95
<5
21
24
9
31
–
79
83
93
93
83
–
2:1
3:1
2:1
DCM
toluene
CH3CN
DMF
TFE
HFIP
HFIP
HFIP
HFIP
HFIP
HFIP
HFIP
8:1
--
9:1
version but a complicated reaction mixture with only minor
desired adduct detected[19] (Table 2, entries 5 and 6). Taken to-
gether, these results indicated that the current reaction was
through a photocatalytic process but not promoted by HFIP
acting as an acid.
>19:1
>19:1
>19:1
>19:1
--
12
13[h]
<5
>95
–
47
--
>19:1
Having optimized a viable catalyst system, we next investi-
gated the newly developed radical protocol in a series of cas-
cade cyclizations (Table 3). Various electron-withdrawing sub-
stituents on the phenol ring were well tolerated (Table 3, en-
tries 1–6). Moreover, electron-rich arenes, such as methoxy-sub-
stituted phenol, could also afford the desired product (Table 3,
entry 7). In addition, when applying this cyclization strategy to
extended ring systems, tricyclization products were furnished
in moderate yields (Table 3, entries 8 and 9). In accordance
with our strategy, simple alkyl hydroxy groups could also serve
as suitable terminators to give the cyclized adducts under
these mild conditions (Table 3, entries 11 and 12). The reaction
with a terminal olefin without methyl substitution (1j) was
also examined and ‘incomplete’ polycyclization product was
observed (Table 3, entry 10). It seemed that unsubstituted ter-
minal vinyl moiety was not a viable initiator and the reaction
tended to initialize with 1,2-disubstituted C=C bond, probably
due to the higher oxidation potential of the monosubstituted
olefin.[20] To our delight, the catalyst system performed well
when the reaction was scaled up to 1.0 mmol scale (Table 3,
entry 1).
[a] Unless otherwise noted, the reactions were performed with 1a
(0.05 mmol) in the presence of photocatalysts 3 (5 mol%) in solvent
(0.5 mL) at room temperature under argon atmosphere; [b] entries 1–7
and 10 were irradiated by Blue LEDs; entries 8–9 and 13 were irradiated
by Green LEDs; entries 11–12 were irradiated by 33 W CFL; [c] determined
by 1H NMR spectroscopy; [d] yields of compounds isolated by chromato-
graphic purification; [e] the loading of 3b was 2.5 mol%; [f] [Ir]=tris(2-
phenylpyridinato)iridium(III); [g] [Ru]=tris(2,2’-bipyridyl)ruthenium chlo-
ride hexahydrate; [h] the reaction was open to air.
tion, and 5 mol% 9-mesityl-10-methylacridinium perchlorate
(3a)[18] as a photoredox catalyst at room temperature (Table 1,
entry 1), affording product 2a in 21% yield and 2:1 d.r. In
screening a variety of solvents, a dramatic solvent effect was
noted (Table 1, entries 1–7). The use of polar protic solvent led
to not only clean reaction, but also significantly improved dia-
stereoselectivity. Hexafluoro-2-propanol (HFIP) was identified
as the optimal solvent (Table 1, entry 7). When examining
other photocatalysts, we found the yield of product was raised
to 93%, alongside >19:1 d.r., with eosin Y (EY; 3b) as the pho-
tocatalyst (Table 1, entries 8 and 9). Organic photocatalyst 3c
also worked comparably well (entry 10). In contrast, inorganic
[Ir]- or [Ru]-based photosensitizers were ineffective (Table 1, en-
tries 11 and 12). In addition, the reaction was more favorably
conducted under inert gas conditions, and the presence of
excess O2 was found to decrease the product yield (Table 1,
entry 13) mainly due to the overoxidation of substrate.
We further examined the reactions with polyenes bearing
1,3-ketocarbonyls as terminators based on their known ten-
dency to equilibrate into the enol forms. However, the initial
experiments with 1,3-diketone 5a under the standard condi-
tions led to no reaction at all. To our delight, when a weak
Lewis acid LiBr was added, the reaction proceeded smoothly
to give the desired cyclic adduct 6a in high yield and good se-
lectivity (Table 4A, entry 1). Further control reactions showed
the use of LiBr only could not promote the reaction, indicating
no acid catalysis was involved (Table 4A, entry 3) and that the
reaction did not proceed without light irradiation, further veri-
fying the photocatalytic nature (Table 4A, entry 4). Other 1,3-
ketocarbonyls were also examined. Both aromatic and aliphatic
Further control reactions confirmed that both light and
eosin Y were essential for significant conversion to the cyclic
product (Table 2, entries 2 and 3). The reaction worked less ef-
fectively under compact fluorescent lamp (CFL) irradiation
(Table 2, entry 4). The use of acid catalysts such as TfOH or p-
toluene sulfonic acid was also examined, giving complete con-
Chem. Eur. J. 2015, 21, 14723 – 14727
14724
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim