Angewandte
Communications
Chemie
formation currently not possible using either mode of
Table 1: Reaction optimization.[a]
catalysis alone (Figure 1b).
This a-amine functionalization reaction[8] generates a-
amino ketones in a single step from simple and inexpensive
starting materials. Classic approaches to the synthesis of this
valuable motif require multistep processes and prefunction-
3
alized reagents.[9,10] An a-amino C(sp ) H carbonylation has
À
Entry[a]
Reaction conditions
Base
Yield [%][a]
been previously reported by Murai and co-workers utilizing
a rhodium catalyst, CO, and ethylene. However, this reaction
requires a 2-pyridyl directing group on the amine substrate,
temperatures exceeding 1008C, and only generates ethyl
ketones.[11] The metallaphotoredox strategy described herein
1
2
3
4
5
6
7
8
9
10
as shown
as shown
as shown
as shown
as shown
as shown
NiCl2·glyme as Ni catalyst
no photocatalyst
no Ni catalyst
no light
no base
Cs2CO3
NaOC(O)Et
DABCO
DBU
quinuclidine
quinuclidine
quinuclidine
quinuclidine
quinuclidine
65
30
82
37
57
83
70
0
utilizes instead an N-aryl moiety for the photoredox poten-
3
À
tial-gated mechanism of C(sp ) H activation. This key
mechanistic feature allows the reaction to take place under
exceptionally mild reaction conditions compared to most
0
0
3
À
transition metal catalyzed C(sp ) H functionalization reac-
tions, thus enabling late-stage coupling of complex and
[a] Yield determined by 1H NMR spectroscopy using 1,1,2,2-tetrachloro-
ethane as an external standard. DABCO=1,4-diazabicyclo[2.2.2]-octane,
DBU=1,8-diazabicyclo[5.4.0]undec-7-ene, DMF=N,N-dimethylform-
amide, dtbbpy=4,4’-di-tert-butyl-2,2’-dipyridyl, cod=1,5-cyclooctadiene,
ppy=2-phenylpyridine.
biologically relevant partners.
3
À
For the C(sp ) H acylation reaction, we envisioned
a catalytic cycle (Figure 1c) initiated by oxidative insertion
of the nickel(0) catalyst A into the acyl-X B to afford the
nickel(II)-acyl oxidative adduct C.[12] Concurrently, irradia-
tion of the iridium photocatalyst, [Ir(ppy)2(dtbbpy)]PF6 (D)
(ppy = 2-phenylpyridine,
dtbbpy = 4,4’-di-tert-butyl-2,2’-
dipyridyl), produces the long-lived excited state complex E
(t = 557 ns).[13] The complex E (E1/2red[*IrIII/IrII] =+ 0.70 V vs.
SCE in MeCN)[13] can undergo SETwith N-phenylpyrrolidine
(1; E1/2red =+ 0.70 V),[14] which, upon deprotonation, gener-
ates the a-amino radical F and reduced iridium(II) species G.
Interception of F by C likely affords the nickel(III) complex
(entry 8), a nickel catalyst (entry 9), and light (entry 10)
result in no conversion to product.
With optimized reaction conditions in hand, we inves-
tigated the scope with respect to the amine coupling partner
(Table 2). A variety of symmetric (4,7) and nonsymmetric
(5,6) acyclic N-aryl amines are tolerated, with ketone
À
À
H. Subsequent reductive elimination forges the C C bond,
formation occurring selectively at the less hindered C H
thus providing J with concomitant generation of the nickel(I)
intermediate I. We expect that both catalytic cycles are closed
by SET from the highly reducing G (E1/2red[IrIII/IrII] = À1.5 V
vs. SCE in MeCN)[13] to I (E1/2red[NiII/Ni0] = À1.2 V vs. SCE in
DMF),[15] thereby reconstituting the nickel(0) catalyst and the
iridium(III) photocatalyst.[16]
bond. Five-, six-, and seven-membered cyclic amine substrates
afford a-amino ketone products in high yields (3, 8–9).
Consistent with a sterically driven deprotonation, the gen-
eration of the quaternary ketone product is not observed with
2-methyl-N-phenylpyrrolidine. Instead, the tertiary ketone
product 11 is formed in 78% yield as a 1:1 mixture of
diastereomers. We were pleased to find that the acylation of
pharmacologically important heterocycles, such as indolines
(12), tetrahydroquinolines (13), and morpholines (10) occurs
smoothly. Furthermore, examination of the substituents on
the N-aryl group revealed that a para-methoxyphenyl (PMP)
group (14), which can be removed under oxidative conditions
to unmask the free amine, is competent under slightly
modified reaction conditions. Additionally, electron-with-
drawing (15, 16), and ortho-substituents (17) are well
tolerated.
Our initial investigations began with the coupling of N-
phenylpyrrolidine (1) and commercially available propionic
anhydride (2a; Table 1). Using [Ni(cod)2], dtbbpy, and [Ir-
(ppy)2(dtbbpy)]PF6, it was discovered that the identity of the
base was crucial for obtaining high yield of the a-amino
ketone product 3.[17] In the absence of exogenous base, 3 was
obtained in 65% yield (entry 1). Addition of sodium propi-
onate afforded the desired ketone product in an improved
82% yield (entry 3). However, employing a carboxylate base
that was not matched to the anhydride, such as sodium
acetate, provided a mixture of methyl and ethyl ketone
products, presumably due to the in situ generation of a mixed
anhydride. Given the impracticality of using unique carbox-
ylate bases for each anhydride partner, we sought a solution
that could be universally applicable. Our attention turned to
amine bases. While 3 was obtained in modest yield using
DABCO and DBU (entries 4 and 5), switching to quinucli-
dine provided the coupled product in 83% yield (entry 6).[18]
The reaction performs with marginally lower efficiency using
NiCl2·glyme, an air-stable nickel precatalyst (entry 7). Nota-
bly, control reactions in the absence of a photocatalyst
We next investigated the scope with respect to the
symmetric anhydride electrophile (Table 3). While acetic
anhydride provides the methyl ketone 18 in only moderate
yield, the reaction performs well with linear, a-branched (19),
and b-branched (20,21) alkyl substituents.[19] Cyclopropyl (22)
and cyclohexyl (23) ketones are also synthesized in high
3
À
yields. The C(sp ) H acylation is sufficiently mild that
a primary alkyl chloride (24) remains unaltered. In addition,
an anhydride containing a trisubstituted double bond fur-
nishes the corresponding a,b-unsaturated ketone 25 in good
yield.[20]
Angew. Chem. Int. Ed. 2016, 55, 4040 –4043
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim