Organic Letters
Letter
1c). Herein we report the results of our study, which establish
microwave-promoted 5-exo-trig iminyl radical cyclizations as
convenient and user-friendly reactions that forge pyrrolines
endowed with diverse functionality. The broad scope of this
process can be attributed to the catalyst- and base-free
conditions as well as the absence of redox cycles. The
reactions are also fast, easy to perform, and in some cases
stereoselective.
We began by probing the microwave-promoted cyclization
of O-phenyloxime 1 in the presence of allylsulfone 2a15 (Table
1). This radical trap permitted convenient measurement of
Table 1. Optimization of Cyclization Conditions
entry
solvent
temp (°C)
time (min)
yield of 3a (%)
a
1
2
3
4
5
6
PhCF3
PhCF3
PhCF3
CH3OH
CH3CN
PhCF3
100
120
130
110
120
120
60
45
45
45
120
120
20
a
35
a
30
a
30
b
41
b
72
Figure 1. Scope of radical traps in cyclizations of 1. Conditions were
PhCF3, 120 °C (μW), and 1−2 h unless otherwise specified.
aIrradiated at 110 °C for 5 h. bIrradiated at 120 °C for 3 h. The major
detected product was an adduct where the cyclic radical was trapped
a
b
1
Calculated from H NMR spectra of reaction mixtures. Isolated
yield.
c
with PhCF3. Irradiated at 110 °C for 2 h.
1
reaction yields via H NMR spectroscopy. Performing the
cyclization at 100 °C in PhCF3 as solvent afforded a low yield
of pyrroline 3a (entry 1). Elevating the temperature to 120 °C
delivered better results (entry 2), but a further increase was not
beneficial (entry 3). Switching to a more polar solvent did not
significantly improve the yield (entries 4 and 5). Finally, we
were pleased to discover that extending the reaction time to 2
h furnished 3a in a satisfactory 72% isolated yield (entry 6).
We then evaluated several other radical traps in the
microwave-promoted cyclization of 1 (Figure 1). A host of
different reagents were viable, affording pyrrolines 3 in
generally good yields. For example, C−O bond formation
could be accomplished by trapping the cyclic radical
intermediate with TEMPO (entry 1). C−X bonds were forged
by employing CCl4,16a CBr4,17 or 2-iodopropane.18 (entries 2−
4). C−N and C−S bonds were constructed by using sulfonyl
azide 2f19 and xanthate 2g,20 respectively (entries 5 and 6).
Finally, C−C bond formation was achieved by trapping with
benziodoxolone-based hypervalent iodine reagent 2i21 (entry
8). The ability to install a diverse range of functional groups is
clearly a hallmark of this radical process that does not require
SET.
traps in these iminyl radical cyclizations must be able to
outcompete the solvent for the cyclic radical intermediate.
Substrates that can undergo β-elimination of a thiyl radical
after cyclization22 provide an attractive alternative to using
radical traps, as the resulting alkene can be elaborated to
introduce numerous functional groups. Accordingly, we
performed the cyclization of allylic sulfide 4 (Scheme 2).
Gratifyingly, this substrate reacted smoothly under microwave
irradiation to produce alkene-containing pyrroline 5 in good
yield.
Scheme 2. Cyclization−Thiyl Radical β-Elimination
Upon establishing the wide scope of the iminyl radical
cyclization with respect to radical traps, we subsequently
demonstrated the viability of various O-phenyloximes 6 in
iminyl radical cyclizations with TEMPO trapping (Scheme 3).
These substrates were readily obtained by condensation of the
corresponding ketones with O-phenylhydroxylamine hydro-
chloride (PhONH2·HCl). Replacement of the phenyl sub-
stituent in 1 with an alkyl group was permitted, albeit with a
somewhat lower cyclization yield (7a; cf. Figure 1, entry 1).
Alkyl substitution of the alkene acceptor at the distal (7b) or
proximal (7c) positions was also tolerated. The use of a cyclic
alkene substrate afforded cis-fused bicycle 7d as a 12:1 mixture
of C−O epimers with TEMPO trapping favored from the
convex face of the radical intermediate. A geminal dimethyl-
Unfortunately, use of Selectfluor16 (2h) as a radical trap
yielded only trace amounts of the desired fluorinated adduct
3h (entry 7). The major product (ca. 10−15%) was an adduct
of the cyclic radical intermediate with PhCF3. Apparently, the
rate of radical trapping by Selectfluor was slower than the rate
of trapping by the solvent. The poor solubility of Selectfluor in
PhCF3 was likely responsible for this problem. However, other
solvents such as CH3CN or CH3OH did not afford detectable
amounts of the desired product. Microwave irradiation of a
solution of 1 in PhCF3 in the absence of radical traps resulted
in slow formation of the PhCF3 adduct. Thus, practical radical
3971
Org. Lett. 2021, 23, 3970−3974