Angewandte
Communications
Chemie
ineffective (Runs 15 and 16). It is well known that copper
complexes with PMDETA efficiently generate alkyl radical
species from alkyl halides in ATRP or ATRA chemistry;[17]
however, the addition of an acid did not affect the coordina-
tion of PMDETA to copper. This reaction required the use of
DABCO, a strong base, with 17% yield of 3aa obtained in the
absence of DABCO (Run 12). A dramatically lower yield of
3aa was obtained when K2CO3, i-Pr2EtN, or 4-dimethylami-
nopyridine (DMAP) was used instead of DABCO (for
optimization details, see Supporting Information).
Scheme 1. This work: Dual catalyst system.
triamine), DABCO (1 equiv, 1,4-diaza bicyclo[2.2.2]octane),
and molecular sieves (MS 3A) in hexane at 1008C for 20 h
(Table 1). At this stage, we expected that generating the
enamine from 1a and the tertiary alkyl radical derived from
2a is critical in this catalyst system. Indeed, 3aa was not
obtained in the absence of a Cu catalyst and pyrrolidine, and
the yield of 3aa depended on the amount of pyrrolidine (Runs
1–4). These results revealed that enamine formation affects
the yield of 3aa. Various acids were examined with the aim of
accelerating the formation of the enamine. p-TsOH and
CF3CO2H gave moderate yields of 3aa (Runs 5 and 6), while
the reaction involving AcOH resulted in 64% yield of 3aa
(Run 7). Although we examined some acetic acid derivatives,
no improvement in yield was observed. On the other hand,
phenol derivatives were effective (Runs 8–14), with 80%
yield of 3aa obtained when methyl 4-hydroxybenzoate was
used (Run 11). Other acids with pKa values of around 9 were
The reactivities of various ketones 1 were next examined
under the optimized reaction conditions (Table 2). Although
the alkylation mechanism is discussed later (vide infra), the
formation of the enamine is an important step in the catalytic
cycle. Pyrrolidine reacted smoothly with electron-poor and
sterically less-hindered ketones 1 to produce the enamine
intermediate, which reacted with the tert-alkyl radical gen-
erated from the reaction of the copper catalyst with the a-
bromocarbonyl compound 2. Therefore, relatively electron-
rich ketones 1 (as in 1b, 1c, 1r, and 1s) resulted in lower yields
of 3, while the reactions proceeded smoothly with relatively
electron-poor ketones 1 bearing (pseudo) halogens (i.e., 1d,
1m, and 1q) and esters (i.e., 1 f–1j) to produce 3 in good
yields. Ketones bearing aminocarbonyl groups (as in 1k and
1l), however, resulted in moderate yields. Ketones bearing
electron-poor aryl and phenyl groups (as in 1n, 1o, and 1p)
also provided 3 in good yields. We examined various solvents
with the aim of increasing the yield, with hexane/CPME
(cyclopentyl methyl ether) found to be the better solvent in
the case of 1b. Sterically hindered ketone 1e gave a low yield
of 3ea, but similarly hindered ketone 1r gave a higher yield of
3ra. Higher yields of 3 were obtained when 50 mol%
pyrrolidine was included in the reactions of 1c and 1r.
Acetophenones bearing thiophene or pyridine (1t, 1u), which
can poison the catalyst, afforded 3ta or 3ua in 38% or 52%
isolated yield, respectively. In the case of 1u, increasing
amount of pyrrolidine was effective to obtain high yield.
We next examined the reactivities of a-bromocarbonyl
compounds 2 (Table 3). To check functional group compat-
ibility, 2b–2j bearing alkyl, substituted benzyl, ester, bromine,
alcohol, and amino functional groups were examined, with
our alkylation reaction exhibiting good functional group
compatibility. We also checked the reactivities of substrates
2k–2o bearing alkyl chains of various length, including
cyclobutyl, cyclohexyl, ethyl, propyl, butyl, and other longer
chains at the a-position, which revealed that this radical
reaction is basically insensitive to steric bulk at the reaction
site. Therefore, even 2o substituted with both n-octyl and n-
hexyl groups afforded 3jo in 48% yield. Increasing the
amount of pyrrolidine was also effective in the reactions of 2n
and 2o, to give 3jn and 3jm in yields of 62% and 60%,
respectively. 2-Bromolactone 2p gave 53% yield of 3jp. We
found that 2-bromocarboxamide 2q was a limitation of this
reaction. Indeed, we examined various 2-bromocarboxa-
mides, with the corresponding reduction, HBr elimination,
and 1,4-HAT-like (hydrogen-atom-transfer-like) side-reac-
tions observed. Although our radical methodology was not
applicable to primary- and secondary-alkyl halide but sec-
Table 1: Additive effect.[a]
Run
Additive
Yield of 3aa [%][b]
1
none
none
none
none
26
0
59
81
2[c]
3[d]
4[e]
5
p-TsOH
47
6
CF3CO2H
43
7
AcOH
64
8
9
p-MeC6H4OH
phenol
51
55
10
11
12
13
14
15
16
C6F5OH
49
p-MeO2CC6H4OH
p-MeO2CC6H4OH
m-MeO2CC6H4OH
o-MeO2CC6H4OH
2-(benzo[d]oxazol-2-yl)phenol
(S)-binol
80 (70)
17[f]
66
32
48
71
[a] All reactions were conducted with 1a (3.0 equiv), 2a (1.0 equiv),
pyrrolidine (30 mol%), CuI (5 mol%), PMDETA (5 mol%), DABCO
(1.0 equiv), additive (30 mol%) and MS 3ꢀ in hexane for 20 h at 1008C.
[b] Yield determined by 1H NMR spectroscopy, 1,1,2,2-tetrachloroethane
used as internal standard; yield of isolated product shown in paren-
thesis. [c] Without pyrrolidine. [d] 50 mol% of pyrrolidine. [e] 100 mol%
of pyrrolidine. [f] Without DABCO.
Angew. Chem. Int. Ed. 2021, 60, 10620 –10625
ꢀ 2021 Wiley-VCH GmbH