2778
E. Hasegawa et al. / Tetrahedron 70 (2014) 2776e2783
rise to formation of spirocyclic ketone 3 (entries 1e3 in Table 1).
While 2a and 2b are similarly effective (entries 1 and 3), the use of
the silicone-activated amine Et2NCH2SiMe3 (TMSA)18 (5.0 equiv) or
1,4-diazabicyclo[2.2.2]octane (DABCO) (5.0 equiv) in place of 2
(1.2 equiv) in these processes results in either a low yield or none of
3, respectively (entries 4 and 5). These observations are consistent
these substances form 4c more selectively (entries 2 and 3).19 While
the use of DMF in place of MeCN as solvent brings about the same
results (entry 4), reactions in DMSO and MeOH are less efficient
(entries 5 and 6) and reaction in DMSO is complicated by a low
recovered mass balance. On the other hand, the reaction is signif-
icantly decelerated in MeOH although the total yields of 4c and 5c
based on the conversion of 1c are above 70%. Similar deceleration
occurred when other protic media, such as aqueous MeCN, were
employed in previous studies of PET reactions of 1c with 2a.6j An-
other important observation is that the concentration of 2a mark-
edly affects the 4ce5c product ratio (entries 7e10). Use of a lower
concentration of 2a brings about formation of a higher yield 4c
(compare 7 and 9 with 1) while a higher concentration of this
substance leads to an increased amount of 5c (compare 8 and 10
with 1). Finally, Ir(ppy)2(dmpy)PF6 (0.02 equiv) also serves as
a photosensitizer for the PET reaction of 1c with 2a (entry 11). Al-
though the absorption profile of the Ir complex17 requires the use of
with
the
electron-donating
abilities
of
the
amines
(2>TMSA>DABCO as judged by their oxidation potentials), which
should determine the efficiencies of SET-promoted reductive
quenching of the excited state of Ru(bpy)3Cl2.19
Table 1
Ru(bpy)3Cl2 and DMBIH or amine photosensitized reactions of 1aa
shorter wavelength of light (l>360 nm), this complex promotes
a reaction that takes place in a comparable yield and selectivity to
those observed using Ru(bpy)3Cl2 (compare with entry 1).
ox
Entry
Amine
(equiv vs 1a)
E
Solvent
Conv of
1a (%)
Yield of
1=2
We next examined reactions of various halomethyl ketone
substrates 1 with 2a under Ru(bpy)3Cl2-sensitized conditions
(Table 3). The results show that reaction of indanone 1d affords 4d
as a major product along with a small amount of naphthol 6, and 5d
(entry 2). In the reaction of benzosuberone 1e, 4e is produced as the
major product although the yield is only moderate (45% based on
the conversion of 1e) (entry 3). In spite of the low yield, this result is
encouraging because direct irradiation of 1e with 2a gave only
a trace quantity of 4e.6j Notably, use of the sensitization protocol
described above leads to cleavage of carbonechlorine bond in 1f
(entry 4) and reaction of the corresponding iodide 1g produces the
expected products 4c and 5c in good yields (entry 5).
The both Ru(bpy)3Cl2 and Ir(ppy)2(dmpy)PF6-based photosen-
sitization methods were applied to promote reaction of the alkyl
ketone substrate 1h (Table 4).22 Using Ru(bpy)3Cl2 and the same
conditions employed for reaction of 1c (entry 1 in Table 2), except
for a longer irradiation time (4 h), 1h is transformed to the expected
ring-expansion product 4h albeit in low yield (recovered 1h, 23%).
Isolation of 4h could not be achieved by using chromatography and,
therefore, the yield was determined by using 1H NMR analysis of
the inseparable product mixture. Notably, use of Ir(ppy)2(dmpy)PF6
as the photosensitizer for this reaction enabled complete con-
(V vs SCE)
3b (%)
1
2a (1.2)
2a (1.2)
0.34
0.34
0.32
0.49f
0.69h
MeCN
DMF
MeCN
MeCN
MeCN
44
28
8c
22
5
2
19c
3d
4
2b (1.2)
34
10
TMSAe (5.0)
DABCOg (5.0)
5
No reaction
a
1a (0.40 mmol), Ru(bpy)3Cl2 (0.02 equiv vs 1a), solvent (4 mL),
l
>390 nm, 1 h.
b
c
Isolated yields.
Determined by 1H NMR.
Average of two independent experiments.
Et2NCH2SiMe3.
d
e
f
Ref. 6h.
g
h
1,4-Diazabicyclo[2.2.2]octane.
Ref. 20.
In contrast, reaction of bromomethyl ketone 1b, which does not
possess an alkenyl tether, under the same conditions as used for the
reaction in entry 2 of Table 1, produces debrominated ketone 5b as
the major product along with a small amount of the ring-expanded
ketone 4b (Scheme 3). The observation suggests that, in contrast to
the primary alkyl radical formed from 1a, which efficiently un-
dergoes 5-exo hexenyl radical cyclization, the one derived from 1b
abstracts a hydrogen atom preferentially.
Scheme 3. Ru(bpy)3Cl2 and DMBIH photosensitized reaction of 1b.
The above findings encouraged us to investigate metal complex
sensitized photoreactions of bromomethyl ketones 1c and 2a
(Table 2). As expected, photoreaction of 1c in MeCN sensitized by
Ru(bpy)3Cl2 and 2a proceeds smoothly to yield the ring-expanded
ketone 4c and debrominated ketone 5c (entry 1). In the absence
of 2a or Ru(bpy)3Cl2, and without irradiation, no reaction takes
place (not shown in Table 2). It should be noted that the yield and
the selectivity for formation of 4c (see the ratio 4c/5c) are high
compared with those of previously studied photochemical and free
radical reactions.21 Here again, both TMSA and DABCO are less
effective co-photocatalysts even though processes promoted by
sumption of 1h and gave 4h in a significantly improved yield. In
both reactions, the formation of debrominated ketone 5h is
negligible.
On the basis of the above observations as well as the results of
previous studies,6 it is possible to propose the plausible mecha-
nistic pathways shown in Scheme 4 for reactions of 1ae1c. The
route is initiated by reductive quenching of the excited state of
Ru(bpy)3Cl2 by DMBIH produces Ru(bpy)3Cl,6h,23,24 which then
donates an electron to 1 giving the radical anion of 1 (1ꢀꢁ, ketyl
form).25,26 Subsequent intramolecular SET to the CbeBr bond gives
a primary alkyl radical 7, as discussed in the context of previous