Journal of the American Chemical Society
Communication
paradigms that are often complementary in nature. One such
example is the development of cyclization reactions of
unsaturated carbonyl compounds and alkynes to generate
cyclopentenones using Co/Rh and Ni catalysts, respectively
(Scheme 1a, right).18,19
reaction. Presumably, this decreased reactivity is due to
hampered alkenyl coordination to the Pd species, which is
required for the intramolecular cyclization.23
The yield of the desired product increased proportionally to
the steric bulk of the trialkylsilane (Table 1, entry 5, and see
SI). An inexpensive bulky silane, triisopropylsilane (iPr3SiH),
proved optimal. Alternative hydride sources, such as formate
salts, failed to deliver the product (Table 1, entry 6). We
performed an experiment using 1a as a limiting reagent, as well
as a separate experiment using equimolar amounts of all
reagents, to demonstrate the reaction’s efficiency. Both
reactions afforded 3a in good yields of 84% and 78%,
respectively (Table 1, entries 7 and 8). This result highlights
the high efficiency with respect to the CO surrogate.24 The
catalyst system was found to be highly active, with lower
catalyst loadings of 1.0, 0.5, and 0.2 mol % giving the desired
product in 95%, 82%, and 53% yield, respectively (see SI). In
the absence of either the ligand or Pd precatalyst, the
formation of 3a was not observed (Table 1, entry 9). Other
transition metal catalysts such as Ni(cod)2, Co2(CO)8,
[Rh(cod)Cl]2, or [Ir(cod)Cl]2 did not afford 3a (Table 1,
entry 11, and see SI for details).
We hypothesized that we could utilize the previously
described carboformylation process as a distinct manifold for
the intermolecular creation of carbocycles with a unique
substitution pattern (Scheme 1b).20,21 Specifically, we
questioned if a proposed key intermediate of the carboformy-
lation reaction, I, could be diverted to undergo an intra-
molecular cyclization. Such a framework for cyclopentenone
synthesis could, in principle, offer several complementary
features when compared to the state-of-the-art: (1) an
alternative de novo reaction design where a wide range of
alkene substrates can be tolerated; (2) safe and practical
application due to the absence of pressurized CO as a reagent;
(3) facile creation of a quaternary carbon center at the 5-
position, a rare substitution pattern. Herein, we report the
successful realization of an intermolecular PK-type reaction of
internal alkynes using an α,β-unsaturated acid chloride as an
alkenyl and CO source (Scheme 1c).
We next explored the substrate scope of this transformation
(Table 2). β-Dialkyl, β-diarylsubstituted and β-monosubsti-
tuted enoyl chlorides are well tolerated, giving a single
regioisomer in each instance. We were able to confirm the
structure of compound 3e by X-ray crystallography. Notably,
we could construct a variety of challenging quaternary carbon
centers demonstrated by several novel examples (3b−3f),
including spirocycles. Previously, this rare substitution pattern
has only ever been formed in Pauson−Khand cycloaddition
reactions as a minor regioisomer with highly strained
alkenes.25,26 Sterically hindered β-aryl enoyl chlorides (3k−
3o) provided excellent yields, while other aryl-substituted acid
chlorides resulted in moderate to good yields, regardless of
electronic properties. A large variety of functional groups
including ethers (3b, 3l, 3q, 3r, 3ac), alkenes (3d), carbamates
(3f), nitriles (3z, 3an), esters (3m, 3w), aldehydes (3x),
sulfones (3y), halogens (F: 3u, 3ad, 3af, Cl: 3o, 3s, Br: 3t, and
even I: 3ab), and nitro groups (3aa) were compatible under
the reaction conditions. It is noteworthy that the reductively
labile groups (e.g., esters, halides) survived the reductive
reaction conditions. Acid chlorides bearing heterocycles such
as thiophenes (3ah, 3aj) and furans (3ag, 3ai) are also effective
reaction partners.
As a model reaction for our initial investigations, an α,β-
unsaturated acid chloride (1a) and an alkyne (2a) were
combined with a Pd precatalyst and a range of phosphine
ligands that have been shown to mediate the reversible
decarbonylation of acid chlorides.4,22 Gratifyingly, we identi-
fied several suitable phosphine ligands for this reaction.
Importantly, the carboformylation product was not observed
to any detectable extent, effectively validating our hypothesis
that the carboformylation process can indeed be interrupted to
deliver cyclopentenones (see SI). A systematic investigation
led to the best yields of 3a being obtained with an electron-
deficient monophosphine ligand, MP1 (Table 1 and see SI).
While triarylphosphines other than MP1 resulted in moderated
yields of 3a, the results show a dependence on both the ratio of
P to Pd and the electronics of the phosphine ligands (Table 1
and see SI). An increased P to Pd ratio, particularly with more
basic or sterically less hindered phosphines, impeded the
Table 1. Reaction Optimization
Conveniently, the reaction can be performed directly from
the α,β-unsaturated carboxylic acid, when combined with
Ghosez’s reagent, for in situ formation of the acid chloride (3d,
3f, 3x).27 In this manner, geranic acid was smoothly converted
to the corresponding cyclopentenone (3d).
a
entry
variations from above conditions
3a (%)
1
2
3
4
5
6
7
8
none
99
79
37
68
30
<3
84
78
n.d.
8
Remarkably, acryloyl chloride could be used as both an
ethene and a CO surrogate (3ak), addressing a classical
limitation of the conventional PKRs.28 Further, the reaction
also works with both α-substituted and α,β-disubstituted enoyl
chlorides bearing methyl (3al), fluoro (3am),29 cyano (3an),
and phenyl (3ao) groups, to furnish a single diastereomer in
these cases bearing an additional β-substituent in moderate to
low yield. In the reaction of 2,3-diphenylacryloyl chloride, an
additional product (3ao-2) arising from isomerization of the
desired product was also observed. Likewise, the reaction with
sorbic acid chloride bearing a conjugated diene functionality
also resulted in isomerization of the distant alkene to give 3ar
(E/Z = 82:18, separable E/Z isomers).17d To further test
MP2 (5.0 mol %) instead of MP1
MP2 (10.0 mol %) instead of MP1
MP3 (5.0 mol %) instead of MP1
Et3SiH instead of iPr3SiH
HCO2M (M = K or Na) instead of iPr3SiH
1a/2a/iPr3SiH equivalent ratio = 1.0/1.5/1.0
1a/2a/iPr3SiH equivalent ratio = 1.0/1.0/1.0
no MP1 or no [Pd(allyl)Cl]2
Pd(PPh3)4 (5.0 mol %) instead of [Pd(allyl)Cl]2/MP1
Ni, Co, Rh, or Ir instead of [Pd(allyl)Cl]2
9
10
11
n.d.
a
The GC yields are based on the moles of a limiting reagent versus n-
dodecane, an internal standard.
B
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX