Y. Zhang et al. / Tetrahedron Letters 56 (2015) 6499–6502
6501
CN
EtOOC COOEt
CN
CN
EtOOC
COOEt
Condition A
+
Ph
COOEt
CN
CN
Br Br
Z-4
2 (equiv.)
1a
2a
2 (equiv.)
3aa
53%
TEMPO (1 equiv.), 3 h, 66%
TEMPO (5 equiv.), 3 h, 61%
BHT (1 equiv.), 3 h, 60%
BHT (5 equiv.), 3 h, 48%
O2 purged, 4 h, 82%
EtOOC COOEt
EtOOC
COOEt
COOEt
CN
Condition A
H
COOEt
CN
Br Br
1a
Ph
Ph
E-4
trans-5
2 (equiv.)
Scheme 5. Effect of radical scavenger.
59%
EtOOC COOEt
Br CN
EtOOC COOEt
Br
COOEt
i-Pr2NEt
hν
i-Pr2NEt
hν
Ph
COOEt
Ph
favored
CN
E2
E2
disfavored
* i-Pr2NEt
* i-Pr2NEt
i-Pr2NEt
PET
i-Pr2NEt
E1 E1
Ar
EtOOC COOEt
EtOOC COOEt
2
PET
E1 E1
Br Br
1
E1 E1
E1 E1
E2
E2
Ar
H
COOEt
CN
H
CN
COOEt
Br
Ph
Ph
Br
Br
3
9
trans-5
6
cis-5
Scheme 3. Photocyclopropanation of two stereoisomeric alkenes Z-4 and E-4.
Scheme 6. Carbene involved mechanism.
obtained for weak electron-withdrawing group, like fluorine (entry
6, Table 2) and chlorine (entry 7, Table 2), substituted substrates.
Alkenes with weak electron-donating groups, like alkyl (entries
9–13, Table 2) and phenyl (entry 14, Table 2), gave good yields,
too. As for strong electron-donating group, like alkoxy (entries
15–18, Table 2) and acetoxy (entry 19, Table 2), the yields
decreased slightly. Then other dibromomalonates 1b and 1c were
applied in this photocyclopropanation. The desired products were
formed in good yields (entries 20 and 21, Table 2).
Next, alkenes 4 with different configurations were employed to
gain stereochemistry information of this photocyclopropanation.
As shown in Scheme 3, reactions under Condition A with either
Z-4 or E-4 gave the same product trans-5 in similar yields. No
cis-5 was formed in either case. These results might indicate a free
bond rotation during the reaction process (vide infra), which led
to a favored configuration of trans-5 in both cases. For further con-
sideration of possible isomerization of Z-4 or E-4 before cyclization,
the unconsumed alkenes Z-4 and E-4 were recovered and no iso-
merization of the double bond was obtained by 1H NMR analysis.
This observation further proved that the initial isomerization of
alkene under UV irradiation did not take place at all and the change
of the configuration (Z-4 to trans-5) should only occur in the cycliza-
tion step. This also negated the possibility that the excitation of the
alkene was the initial step for the whole procedure, since such exci-
tation must lead to the isomerization of Z-4 into E-4.11
alkene 4 was detected (Scheme 3) proved that the excitation of
alkene 2 was not the key step during this transformation. On the
other hand, the observation that no cyclopropane was formed
without Hünig base (entry 17, Table 2) suggested that the direct
excitation of dibromomalonate 1 could only lead to decomposition.
Thus, the electron transfer starting from Hünig base under the
excitation of UV light10 seemed to be the most reasonable. The
Photon-induced Electron Transfer (PET) from Hünig base to dibro-
momalonate could result in several intermediates which were
listed below.
Firstly, a free radical mechanism was proposed as shown in
Scheme 4. The PET from excited Hünig base to 1 led to a free radical
6 along with the release of Brꢀ. The radical addition of 6 with 2
afforded radical intermediate 7 which underwent another PET to
give a di-radical 8. The final radical coupling yielded cyclopropane 3.
To check the possibility of the above mechanism, reactions with
radical scavengers were carried out under Condition A (Scheme 5).
The addition of 1 equiv of TEMPO resulted in a decreased yield.
Further increasing the amount of TEMPO to 5 equiv led to a similar
result. Next, BHT was used instead of TEMPO. The reaction showed
similar efficiency and yield with 1 equiv of BHT. Lower yield was
observed with 5 equiv of BHT compared with reactions employing
TEMPO. Oxygen was also employed as radical scavenger and led to
82% yield. Considering that reactions with radical scavengers led to
a decreased yield, radical might participate in the reaction path-
way. On the other hand, neither 5 equiv of radical scavengers nor
O2 could completely inhibit this transformation, which might indi-
cate that radical species were not the predominant intermediate
during this reaction process. Additionally, a dark reaction after
5 min photo irradiation was also carried out yielding only 19% of
With these results in hand, we further studied the mechanism.
The ultraviolet–visible spectra of 1a, 2a, and Hünig base (see Fig.S1
in Supplementary data) showed clearly that all these three compo-
nents absorb light at k = 185 nm and 254 nm. Nevertheless, as dis-
cussed above, the fact that no isomerization of the recovered
i-Pr2NEt
hν
i-Pr2NEt
hν
E2
E2
* i-Pr2NEt
i-Pr2NEt
PET
* i-Pr2NEt
i-Pr2NEt
E1 E1
E2
E2
Ar
E1
E2
PET
Br
E1
E1
E2
2
E1
Br
E1 E1
Br Br
1
E1 E1
E2
E2
Ar
Ar
Ar
Br
3
Br
7
6
8
Scheme 4. Free radical involved mechanism.