Full Paper
tives.[22] Inspired by this work, we conceived the possibility of
introducing electron-deficient olefins to react with azomethine
imines to furnish [3+2] cycloaddition reactions under phos-
phine catalysis conditions. Nucleophilic phosphine catalysis has
emerged as an efficient tool to prepare carbo- and heterocy-
cles.[23] Generally, in most of the phosphine-catalyzed cycload-
ditions, activated allenes, alkynes, Morita–Baylis–Hillman (MBH)
carbonates and acetates acted as electrophiles to be attacked
by phosphine to give active zwitterions, which then react with
electrophilic coupling partners, leading to various annulation
reactions.[23] The electron-deficient alkenes had often been
used as electrophilic coupling reagents to accomplish the cy-
cloaddition reactions,[23] and seldom employed as an electro-
phile to react with phosphine to form active zwitterions for
conducting further annulation reactions. Herein, we report the
first example of phosphine-catalyzed [3+2] cycloaddition reac-
tions of various azomethine imines with electron-deficient phe-
nylsulfonyl alkenes.
nucleophilicity were screened (entries 2–9). Using 20 mol% of
PPh3, the azomethine imine 1a was treated with 1.5 equiva-
lents of 2a for 48 h; delightfully, the [3+2] product 3a could
be obtained in 76% yield (entry 2). Unfortunately, the product
is a mixture of several diastereomers, which could not be sepa-
rated by flash column and could also not completely be sepa-
rated on several commercially available chiral columns by
HPLC analysis. Moreover, the NMR spectrum of the product 3a
looks like the data of a pure compound. Therefore, the diaste-
reomeric ratio of the cycloaddition reactions could not be de-
termined. In Table 1, only the isolated yields of the product 3a
as a diastereomeric mixture were reported. Next, screening re-
vealed that several other phosphines could also catalyze the
cycloaddition reaction, giving the corresponding product in
moderate to good yields (entries 3–9). Among these phos-
phines, MePPh2 gave the best yield of 84% (entry 5). Some ter-
tiary amines, such as Et3N, 1,4-diazobicyclo[2.2.2]octane
(DABCO), 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU), and 4-di-
methylaminopyridine (DMAP), have also been examined for im-
proving yields (entries 10–13). These amine catalysts could also
enable the reaction to give the corresponding products, albeit
in moderate yields. To achieve asymmetric cycloaddition, sever-
al commercially available chiral phosphines, such as 2,2’-bis-
(diphenylphosphino)-1,1’-binaphthyl (BINAP) and (+)-1,2-
bis[(2R,5R)-2,5-diethylphospholano]ethane (BPE), have been
tried in the reaction of azomethine imine 1a with 2a, but the
racemic product was obtained. These results have been includ-
ed in the Supporting Information.
Results and Discussion
Initial optimization experiments were conducted with N,N’-
cyclic azomethine imine 1a and commercially available (Z)-1,2-
bis(phenylsulfonyl)ethylene (2a) as the model substrates. Since
numerous cycloadditions of azomethine imines with electron-
deficient alkenes in the absence of metal catalyst or organoca-
talyst are known,[24] the potential background reaction from
direct cycloaddition of 1a with 2a was first investigated. The
reaction of 1a and 2a was carried out in 1,2-dichloroethane at
808C in the absence of catalyst for 48 h. TLC monitoring re-
vealed no new spot was generated and the substrate 1a was
almost not consumed (Table 1, entry 1), therefore the back-
ground reaction at room temperature could be excluded.
Then, several commercially available phosphines with different
With the optimized conditions established (1 (1 equiv), 2a
(1.5 equiv), MePPh2 (20 mol% in CH2Cl2)), the scope of the sub-
strate 1 was examined (Table 2). Various N,N’-cyclic azomethine
imines bearing different substituents on the benzene rings car-
ried out the [3+2] cycloaddition, giving the tetrahydro-1H,5H-
pyrazolo[1,2-a]pyrazol-1-one derivatives in 69–93% yields
(Table 2, entries 1–21). Azomethine imines bearing electron-do-
nating groups on the benzene ring afforded excellent yields of
the cycloadducts (entries 2–4). Azomethine imines bearing
electron-withdrawing groups on the benzene ring also worked
smoothly to provide the corresponding products in high yields
(entries 5–18), but those azomethine imines with strong elec-
tron-withdrawing groups (NO2, CN, CF3) on the benzene ring
resulted in slightly lower yields (entries 14–18) relative to other
azomethine imines. Interestingly, for those azomethine imines
with halo substituents on the benzene ring, azomethine
imines containing meta-substituted benzene rings provided
the corresponding cycloadducts in a somewhat lower yield
than those azomethine imines bearing ortho- or para-substitut-
ed benzene rings (entries 5–13). The azomethine imines bear-
ing 1-naphthyl, 2-naphthyl, and 2-furyl groups, were also suita-
ble substrates, readily giving the corresponding cycloadducts
in 70–84% yields (entries 19–21). Very satisfactorily, when alkyl
imine was employed, the yield of the cycloadduct reached an
excellent 90% (entry 22). Disappointingly, in most cases, the
diastereomeric mixture of the products could not be separated
by chiral HPLC analysis. The diastereomeric mixtures of only
3e, 3 f, 3p, 3s, and 3t were barely separated. In these cases,
without exception, all diastereomeric ratios were unanimously
Table 1. Optimization of reaction conditions for [3+2] cycloaddition reac-
tions of N,N’-cyclic azomethine imines with diphenylsulfonyl alkenes.[a]
Entry
Catalyst
T [8C]
Yield [%][b]
1[c]
2
3
4
5
6
7
8
–
PPh3
Bu3P
80
25
25
25
25
25
25
25
25
25
25
25
25
0
76
72
80
84
72
74
50
67
61
72
63
35
Me2PPh
MePPh2
n-PrPPh2
iPrPPh2
tBuPPh2
CyPPh2
Et3N
DABCO
DBU
DMAP
9
10
11
12
13
[a] 1.5 equiv of 2a was used. [b] Isolated yield. [c] Without catalyst and
with 1,2-dichloroethane as the solvent.
Chem. Eur. J. 2014, 20, 1731 – 1736
1732
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim