Photochemical C-C Bond Cleavage of 1,2-Diarylcyclopropanes Bearing an Acetylphenyl Group. Generation and Observation of Triplet 1,3-Biradicals

Article abstract of DOI:10.1021/jo971111s

The photochemical properties of 1,2-diarylcyclopropanes bearing an acetylphenyl group were studied by product analysis and laser flash photolysis. All cyclopropanes showed efficient cis-trans photoisomerization followed by inefficient isomerization to 1,3-diarylpropenes. All the products were unquenched by the addition of triplet quencher, 2-methyl-1,3-butadiene (E_T ? 60 kcal mol^-1), whereas molecular dioxygen gave oxygenated products. Triplet 1,3-biradicals generated from a short-lived acetophenone-like triplet (?1 ns) were observed as intermediates in these reactions through nanosecond laser flash photolysis. Polar substituent effects on the lifetime of the 1,3-biradicals were small, except for a heavy atom effect in the case of Br. Spin-orbit coupling calculations on model 1,3-biradicals show a negligible effect on polar substituent and thus predict a negligible effect on the intersystem crossing rate. Conjugated biradicals in general now seem unlikely to show the "ionic character effect" on intersystem crossing suggested by the work of Salem and Rowland.

Full text of DOI:10.1021/jo971111s

3176
J . Org. Chem. 1998, 63, 3176-3184
P h otoch em ica l C-C Bon d Clea va ge of 1,2-Dia r ylcyclop r op a n es
Bea r in g a n Acetylp h en yl Gr ou p . Gen er a tion a n d Obser va tion of
Tr ip let 1,3-Bir a d ica ls
Nobuyuki Ichinose,† Kazuhiko Mizuno,* and Yoshio Otsuji
Department of Applied Chemistry, College of Engineering, Osaka Prefecture University,
Sakai, Osaka 599-8531, J apan
Richard A. Caldwell* and Anna M. Helms
Department of Chemistry, University of Texas at Dallas, Richardson, Texas 75080
Received J une 19, 1997
The photochemical properties of 1,2-diarylcyclopropanes bearing an acetylphenyl group were studied
by product analysis and laser flash photolysis. All cyclopropanes showed efficient cis-trans
photoisomerization followed by inefficient isomerization to 1,3-diarylpropenes. All the products
were unquenched by the addition of triplet quencher, 2-methyl-1,3-butadiene (ET = 60 kcal mol-1),
whereas molecular dioxygen gave oxygenated products. Triplet 1,3-biradicals generated from a
short-lived acetophenone-like triplet (j1 ns) were observed as intermediates in these reactions
through nanosecond laser flash photolysis. Polar substituent effects on the lifetime of the 1,3-
biradicals were small, except for a heavy atom effect in the case of Br. Spin-orbit coupling
calculations on model 1,3-biradicals show a negligible effect on polar substituent and thus predict
a negligible effect on the intersystem crossing rate. Conjugated biradicals in general now seem
unlikely to show the “ionic character effect” on intersystem crossing suggested by the work of Salem
and Rowland.
In tr od u ction
orientation; and (3) ionic character in the S wave func-
tion.1 For short-chain biradicals, the SOC mechanism
will dominate because of the closeness of the termini. For
hexamethylenes and longer chain biradicals, however,
hyperfine coupling (HFC) dominates in isc for extended
conformations,5-8 since SOC decreases exponentially with
an increase in the inter terminal distance and HFC
interaction is distance independent.
After the first flash spectroscopic detection of 1,4-
biradical intermediates in the Norrish type II reaction
by Scaiano in 1977,9,10 several biradicals, including longer
chain species produced by the Norrish type I reaction,
have been detected by laser flash photolysis.2,3 The
lifetimes of 1,n-biradicals have been compared as a
function of either n,11-15 or the distance between the
termini.16,17 These results have qualitatively supported
the distance dependence predicted by Salem and Row-
Biradicals are among the most important intermedi-
ates in solution phase organic chemistry. The chemical
behavior of biradicals (i.e., products and respective ratios)
is strongly dependent upon spin multiplicity. Thermally
produced singlet biradicals decay to ground state prod-
ucts too rapidly to be detected with spectroscopic tech-
niques. Triplet biradicals, produced from excited triplet
state precursors, cannot decay directly to the ground
state since this process would be spin-forbidden. They
must convert to singlet biradicals via intersystem cross-
ing (isc) before yielding final products. The lifetime, τ,
of a triplet biradical is normally expressed as
-1
τ ) (kisc
)
where kisc is the rate of intersystem crossing and is
determined by the magnitude of spin-orbit coupling
(SOC) between the singlet (S) and the triplet (T) states.1
The lifetimes of typical triplet biradicals, measured either
by laser flash photolysis2,3 or through kinetic studies,4
are of nanosecond order (1-900 ns). Spin-orbit coupling
is increased by (1) minimized distance between the
termini; (2) perpendicular yet overlapping p orbital
(5) Zimmt, M. B.; Doubleday, C.; Turro, N. J . J . Am. Chem. Soc.
1985, 107, 6726.
(6) Wang, J .; Doubleday, C.; Turro, N. J . J . Phys. Chem. 1989, 93,
4780.
(7) Wang, J .; Doubleday, C.; Turro, N. J . J . Am. Chem. Soc. 1989,
111, 3962.
(8) Doubleday, C.; Turro, N. J .; Wang, J . Acc. Chem. Res. 1989, 22,
199.
(9) Scaiano, J . C. J . Phys. Chem. 1977, 81, 828.
(10) Small, R. D.; Scaiano, J . C. J . Photochem. 1977, 6, 453.
(11) Caldwell, R. A.; Sakuragi, H.; Majima, T. J . Am. Chem. Soc.
1984, 106, 2471.
(12) Barton, D. H. R.; Charpiot, B.; Ingold, K. U.; J ohnston, L. J .;
Motherwell, W. B.; Scaiano, J . C.; Staforth, S. J . Am. Chem. Soc. 1985,
107, 3607.
(13) Zimmt, M. B.; Doubleday, C.; Gould, I.; Turro, N. J . J . Am.
Chem. Soc. 1985, 107, 6724.
(14) Zimmt, M. B.; Doubleday, C.; Turro, N. J . J . Am. Chem. Soc.
1986, 108, 3618.
† Present address: Kansai Research Establishment, J apan Atomic
Energy Research Institute, 25-1 Mii-minamimachi, Neyagawa, Osaka
572-0019, J apan.
(1) Salem, L.; Rowland, C. Angew. Chem., Int. Ed. Engl. 1972, 11,
92.
(2) J ohnston, L. J .; Scaiano, J . C. Chem. Rev. 1989, 89, 521.
(3) Caldwell, R. A. In Kinetics and Spectroscopy of Carbenes and
Biradicals; Platz, M. S., Ed.; Plenum Publishing Corp.: New York,
1990; pp 77-116.
(4) Adam, W.; Grabowski, S.; Wilson, R. M. Acc. Chem. Res. 1990,
23, 165.
(15) Zimmt, M. B.; Doubleday, C.; Turro, N. J . Chem. Phys. Lett.
1987, 134, 549.
(16) Scaiano, J . C. Tetrahedron 1982, 38, 819.
S0022-3263(97)01111-0 CCC: $15.00 © 1998 American Chemical Society
Published on Web 04/23/1998

Generation of Triplet 1,3-Biradicals
J . Org. Chem., Vol. 63, No. 10, 1998 3177
Sch em e 1
Ta ble 1. Qu a n tu m Yield s for th e F or m a tion of
P h otop r od u cts in th e P h otoisom er iza tion in Ben zen e a t
18 °Ca
reaction
1a f 2a
2a f 1a
1a or 2a f 3a + 4a
Φb
0.45 ( 0.02
0.11 ( 0.02
0.005
a
b
[1a ] ) [2a ] ) 1.0 × 10-2 mol L-1
.
Values for 313 nm light.
Determined by HPLC.
F igu r e 1. The 1,2-diarylcyclopropanes investigated in this
work.
Ta ble 2. Qu a n tu m Yield s for th e F or m a tion of 2a in th e
P h otoisom er iza tion of 1a a t Va r iou s Con cen tr a tion s of
1a in Ben zen e a t 18 °C
land. Trimethylene, a proposed intermediate in the
isomerization of cyclopropane derivatives,18-25 has long
been a model for theoretical studies of intersystem
crossing16,26,27 since, with the exception of carbenes and
excited alkenes, trimethylene is the smallest biradical.
Although lifetimes for 1,3-cyclopentanediyls28,29 were
known, lifetimes of flexible 1,3-biradicals in room-tem-
perature solution, though desired,30,31 were not available
until 1985, when we first reported the measurement of
1,3-biradical lifetimes.32
a
concn, mol L-1
Φ1a f2a
1.0 × 10-3
5.0 × 10-3
1.0 × 10-2
2.5 × 10-2
0.51 ( 0.04
0.46 ( 0.02
0.45 ( 0.02
0.47 ( 0.02
a
Values for 313 nm light. Determined by HPLC.
biradicals, thus avoiding the operational difficulty of an
external sensitizer.
In this paper, we report substituent and conforma-
tional effects on the photochemistry of 1,2-diarylcyclo-
propanes bearing acetylphenyl groups and related com-
pounds (1a -k ). These compounds are shown in Figure
1. We show the intermediacy of triplet 1,3-biradicals in
their geometric photoisomerization and determine the
lifetimes of the corresponding 1,3-biradicals by laser flash
photolysis. Since excitation of the acetylphenyl group
will produce the excited triplet state by rapid intersystem
crossing from the excited singlet state, we expected
intramolecular sensitization by the acetophenone chro-
mophore, leading to efficient generation of triplet 1,3-
Resu lts a n d Discu ssion
Sta tion a r y Sta te P h otoch em istr y of 1,2-Dia r yl-
cyclop r op a n es Bea r in g a n Acetylp h en yl Gr ou p .
Irradiation of a benzene solution containing trans-1,2-
bis(4-acetylphenyl)cyclopropane (1a ) with 313 nm light
caused efficient cis-trans isomerization of the cyclopro-
pane and gave a mixture of cis- and trans-cyclopropanes
in a ratio of 55:45 ((1) within 1 h (Scheme 1). On
prolonged irradiation (54 h) the cyclopropanes gradually
isomerized to (E)- and (Z)-1,3-bis(4-acetylphenyl)propene
(E-3a and Z-4a , E/Z ) 2/1) which were isolated in 42%
yield. 1H NMR and HPLC analyses of this photoreaction
revealed that 3a and 4a were primary products which
were formed even in the early stages of the reaction;
however, the efficiency for formation of 3a and 4a was
quite low. Similarly, 2a also isomerized to give a mixture
of 1a , 3a , and 4a upon irradiation. Irradiation of a
mixture of 3a and 4a did not yield 1a and 2a . Irradiation
of 1a or 2a in methanol gave their isomers 1a , 2a , 3a ,
and 4a . No polar addition products were obtained.33,34
This result excludes the intermediacy of ionic species in
the isomerization. Moreover, irradiation in the presence
of 2-propanol or in the presence of 1.0 M ethanethiol in
benzene did not cause the reduction of the acetyl group
to a 1-hydroxyethyl group. Quantum yields for these
photoreactions in benzene were measured in both the
absence and presence of additives (Tables 1-3). The
difference between Φcft and Φtfc may be explained by
energy wastage. Differential yields of transients are
inconsistent with results from transient spectra which
(17) Caldwell, R. A.; Dhawan, S. N.; Majima, T. J . Am. Chem. Soc.
1984, 106, 6454.
(18) Berson, J . In Rearrangements in Ground and Excited States;
de Mayo, P., Ed.; Academic Press: New York, 1980.
(19) Zimmerman, H. E. In Rearrangements in Ground and Excited
States; de Mayo, P., Ed.; Academic Press: New York, 1980.
(20) Hammond, G. S.; Wyatt, P.; DeBoer, C. D.; Turro, N. J . J . Am.
Chem. Soc. 1964, 86, 2532.
(21) Griffin, G. W.; Covell, J .; Petterson, R. C.; Dodson, R. M.; Klose,
G. J . Am. Chem. Soc. 1965, 87, 1410.
(22) Valyocsik, E. W.; Sigal, P. J . Org. Chem. 1971, 36, 66.
(23) Hixson, S. S.; Boyer, J .; Gallucci, C. J . Chem. Soc., Chem.
Commun. 1974, 540.
(24) Wong, P. C.; Arnold, D. R. Tetrahedron Lett. 1979, 23, 2101.
(25) Roth, H. D.; Manion Schilling, M. L. J . Am Chem. Soc. 1981,
103, 7210.
(26) Doubleday, C.; McIver, J . W.; Page, M. J . Phys. Chem. 1988,
92, 4367.
(27) Carlacci, L.; Doubleday, C. E., J r.; Furlani, T. R.; King, H. F.;
McIver, J . W., J r. J . Am. Chem. Soc. 1987, 109, 5323.
(28) Adam, W.; Hossel, P.; Hummer, W.; Platsch, H.; Wilson, R. M.
J . Am. Chem. Soc. 1987, 109, 7570.
(29) Adam, W.; Grabowski, S.; Platsch, H.; Hannemann, K.; Wirz,
J .; Wilson, R. M. J . Am. Chem. Soc. 1989, 111, 751.
(30) Nishino, H.; Toki, S.; Takamuku, S. J . Am. Chem. Soc. 1986,
108, 5030.
(31) Nakabayashi, K.; Nishino, H.; Toki, S.; Takamuku, S. Radiat.
Phys. Chem. 1989, 34, 809.
(32) Mizuno, K.; Ichinose, N.; Otsuji, Y.; Caldwell, R. A. J . Am.
Chem. Soc. 1985, 107, 5797.
(33) Irving, C. S.; Petterson, R. C.; Sarker, I.; Kristinson, H.; Aaron,
C. S.; Griffin, G. W.; Boudreaux, G. J . J . Am Chem. Soc. 1966, 88,
5675.
(34) Hixson, S. S. J . Am. Chem. Soc. 1974, 96, 4866.

3178 J . Org. Chem., Vol. 63, No. 10, 1998
Ichinose et al.
Ta ble 3. Qu a n tu m Yield s for th e F or m a tion of 2a in th e
Ta ble 4. Qu a n tu m Yield for th e F or m a tion of 2a in th e
Ar om a tic Keton e-Sen sitized P h otoisom er iza tion of 1a in
Ben zen ea
P r esen ce of Tr ip let Qu en ch er in Ben zen e a t 18 °Ca
b
quencher
concn, mol L-1
Φ1a f2a
aromatic ketone
2-methyl-1,3-butadiene
0.05
0.42
2.82
saturated
0.49 ( 0.04
0.57 (0.05
0.45 (0.06
0.53 ( 0.04
b
(ET, kcal mol-1
)
concn/mol L-1
Φ1a f2a
acetophenone (73.7)
0.036
0.32
0.41 ( 0.02
0.44 ( 0.02
0.39 ( 0.02
0.39 ( 0.02
0.07 ( 0.03
oxygen
1.01c
0.40c
0.04c
b
a
b
[1a ] ) 1.0 × 10-2 mol L-1
.
Values for 313 nm light.
benzophenone (68.6)
benzil (53.4)
Determined by HPLC.
a
[1a ] ) 1.0 × 10-2 mol L-1
.
Values for 313 nm light. c >90%
Sch em e 2
of 313 nm light was absorbed by the sensitizer.
Ta ble 5. Qu a n tu m Yield s for th e F or m a tion of 2b-d in
th e P h otoisom er iza tion of 1b-d in Ben zen e in th e
Absen ce a n d P r esen ce of 2-Meth yl-1,3-bu ta d ien e
a
compound
[2-methyl-1,3-butadiene], mol L-1
Φ1f2
1bb
0
0.38 ( 0.03
0.38 ( 0.03
0.33 ( 0.03
0.20 ( 0.02
0.23 ( 0.02
0.21 ( 0.02
0.24 ( 0.02
0.45 ( 0.03
0.39 ( 0.03
0.39 ( 0.03
0.31 ( 0.02
0.515
0.784
0
0.140
0.235
0.456
0
0.226
0.434
0.754
1cc
1d d
a
show identical transient intensity from matched optical
densities of 90:10 cis/trans mixtures. Quantum yield
discrepancies are under further investigation. The quan-
tum yield for the cis-trans isomerization was not de-
pendent on the concentration of 1a . The addition of a
triplet quencher, 2-methyl-1,3-butadiene (isoprene), did
not affect the isomerization. The above behavior may
suggest the incorporation of singlet intermediates in
these photoreactions; however, irradiation of 1a in ben-
zene with vigorous bubbling of oxygen afforded a mix-
ture of 3,5-bis(4-acetylphenyl)-1,2-dioxolane (5a), 4-acetyl-
benzaldehyde (6a ), and 1,4-diacetylbenzene (7a ) in a 1:1:1
ratio (Scheme 2). The structure of 5a was assigned by
b
Values for 313 nm light. Determined by GLC. [1b] ) 4.78
× 10-3 mol L-1
.
c [1c] ) 5.19 × 10-3 mol L-1
.
[1d ] ) 4.25 × 10-3
d
mol L-1
.
Ta ble 6. P r op er ties of Tr a n sien t Sp ecies
compound
solvent
MeOH
Heptane
MeOH
Heptane
Benzene
EtOH
lifetime, ns λmax, nm ref
1
the H NMR spectrum and the chemical conversions.35
1a
13.4 ( 1
14.7 ( 1
13.0 ( 1
15.1 ( 1
3000
327
320
327
311
c
c
c
Either irradiation or heating to 150 °C of a mixture of
5a , 6a , and 7a quantitatively afforded 6a and 7a in a
1:1 mixture. Whereas direct irradiation of 1,2-bis(4-
methoxyphenyl)cyclopropane did not yield oxygenation
products, but rather caused the cis-trans isomerization
although 9,10-dicyanoanthracene-sensitized photoxygen-
ation of this compound efficiently gave 3,5-bis(4-meth-
oxyphenyl)-1,2-dioxolanes.35 These results suggest the
intermediacy of a triplet 1,3-biradical in the photoisomer-
ization of 1a and 2a , although irradiation of 1a in the
presence of an alkene, such as methyl acrylate or butyl
vinyl ether (0.5 mol L-1), did not produce any products
other than 2a , 3a , and 4a .
2a
1b
c
acetophenone
38
38
39
c
41
41
37
39
38
140
270
340
320
4-methylacetophenone cyclohexane
3a /4a a
MeOH
29 ( 1
34 ( 1
33
(E)-AnCHdCHCH2Phb MeOH
(E)-PhCHdCDCH2Ph MeOH
PhCH2
EPAd
cyclohexane
EtOH
319
370
330
PhC(OH)CH3
a
b
(3a /4a ) 2/1). An ) anisyl ) 4-MeOC6H4. c This work.
d
Ether-isopentane-EtOH glass at 77 K.
Similarly trans-1-(2-acetylphenyl)-2-phenylcyclopropane
(1c) and trans-1-(3-acetylphenyl)-2-phenylcyclopropane
(1d ) isomerized and afforded photostationary mixtures
of cis and trans isomers in ratios of 62:38 and 44:56 ((1),
respectively. The addition of 2-methyl-1,3-butadiene did
not affect the photoisomerization of 1c to 2c, but did
slightly affect that of 1d . Quenching by 2-methyl-1,3-
butadiene may suggest the incorporation of a locally
excited acetophenone-like triplet precursor in the pho-
toisomerization. The quantum yields for the photoiso-
merization of 1b-d to 2b-d in the absence and presence
of 2-methyl-1,3-butadiene are summarized in Table 5.
Efficient cis-trans isomerization of 1-(4-acetylphenyl)-
2-(4-substituted phenyl)cyclopropanes (1e-i) was also
observed by irradiation (313 nm) in benzene with no
substantial substituent effect observed (Table 7).
Various sensitizers were employed to effect the triplet
photosensitization of the cis-trans isomerization of 1a
to 2a . The isomerization sensitized by acetophenone,
which is the same chromophore contained in 1a , was as
efficient as that of direct irradiation. However, when
sensitzers with lower triplet energy were employed, the
quantum yield for the isomerization decreased (Table 4).
These results strongly suggest that the cis-trans isomer-
ization proceeds from an acetophenone-like triplet (ET =
72 kcal mol-1).
The irradiation (313 nm) of trans-1-(4-acetylphenyl)-
2-phenylcyclopropane (1b) in benzene rapidly afforded
the cis isomer 2b. The ratio of 1b to 2b reached a
photostationary state of 45:55 ((1). They afforded a
mixture of four isomeric 1,3-diarylpropenes upon pro-
longed irradiation. These alkenes were also formed as
primary products. The photoreaction was not quenched
by the addition of 2-methyl-1,3-butadiene (0.5 mol L-1).
La ser F la sh Sp ectr oscop y. Laser flash photolysis
of a heptane solution of 1a (fourth harmonic of Nd:YAG
laser, 266 nm, fwhm 1.75 ns, unfocused ca. 20 mJ /pulse)
produced a transient with an absorption maximum, λmax
at 320 nm and a lifetime of 14.7 ( 1 ns. A similar
,
(35) Mizuno, K.; Kamiyama, N.; Ichinose, N.; Otsuji, Y. Tetrahedron
1985, 41, 2207.

Generation of Triplet 1,3-Biradicals
J . Org. Chem., Vol. 63, No. 10, 1998 3179
Ta ble 7. Su bstitu en t Effects on th e P h otoisom er iza tion
of tr a n s-1-(4-Acetylp h en yl)-2-(4-su bstitu ted
p h en yl)cyclop r op a n es (1a -b, 1e-i) a n d th e Lifetim e of
th e Cor r esp on d in g 1,3-Bir a d ica ls
a
b
compound
X
Φ1f2
λmax
lifetime/ns
1a
1b
1e
1f
1g
1h
1i
CH3CO
H
CH3O
CH3
Cl
0.45
0.38
0.19
0.51
0.58
0.57
0.40
320
311
315
310
310
14.7 ( 1.0
15.1 (1.0
14.5 ( 0.5
14.0 ( 0.2
13.1 ( 0.6
5.2 ( 0.4
Br
CN
305
16.0 ( 0.5
a
Values for 313 nm light. Determined by GLC. [1] ) 5 × 10-3
mol L-1
.
Values in heptane.
b
transient absorption (λmax ) 320 nm, τ ) 13.4 (1 ns) was
observed for a methanol solution of 1a . These signals
were unaffected by the addition of 2-methyl-1,3-butadiene
F igu r e 2. Transient absorption spectrum of compound 1j in
heptane: 266 nm excitation, 12 mJ /P, λmax ) 320 nm.
(1 mol L-1) in either their intensity, lifetime, or λmax
.
Oxygen saturation of an acetonitrile solution of 1a
shortened the lifetime from 15.7 ( 0.2 to 7.1 ( 0.5 ns.
The transient spectroscopic study of 2a also showed
strong signals identical in lifetime, absorption maxima,
and changes in optical density, ∆OD, to those observed
for 1a . These transient signals were assigned to a triplet
1,3-biradical (3BR) for the following reasons: (1) Due to
the very fast intersystem crossing rate observed for the
acetophenone chromophore (1010-1011 s-1),17 the spin
multiplicity of these signals were assigned as triplet. (2)
The absence of quenching by 2-methyl-1,3-butadiene
3
3
rules out the spectroscopic states 1a * and 2a * which,
like triplet acetophenone, should be quenched at a
diffusion-controlled rate.36 (3) The absorption maxima
are in accord with that of a benzyl radical,37 not with that
of a ketyl radical.38,39 (4) The fact that excitation of 1a
and 2a produces identical transients indicates that the
transient is a stereorandomized species. (5) The tran-
F igu r e 3. Transient decay curve of compound 1j in heptane:
266 nm excitation, 12 mJ /P, monitored wavelength ) 330 nm,
τ ) 12.9 ns.
3
3
sient is not the triplet of 3a or 4a , 3a * and 4a *, since
the transient observed by the irradiation of a mixture of
3a and 4a in methanol had both a different lifetime, τ )
and related compounds are listed in Table 6. The lifetime
and λmax of these species is in good agreement with those
reported for the triplet 1,3-diphenyltrimethylene biradi-
cal.42 The spectral broadness exhibited by these com-
pounds was also observed for 1,3-diaryl-1,3-cyclopen-
tanediyl triplet biradicals.43,44 Figures 2 and 3 illustrate
typical spectra and kinetic decay curves for these com-
pounds.
29 ( 1 ns, and a different absorption maximum, λmax
)
310 nm. The lifetime agrees well with values reported
for 1,3-diarylpropene triplets in methanol.40,41 (6) The
lifetime of the transient generated from 1a did not show
any appreciable temperature effect (Ea ) 0.14 ( 0.1 kcal
mol-1). The absence of the acetophenone-like triplet
3
3
3
precursors of the BR, 1a *, and 2a * in the laser flash
photolysis study can be explained by their extremely
short lifetimes. Similar results were obtained for trans-
1-(4-acetylphenyl)-2-phenylcyclopropane (1b). The laser
flash photolysis of 1b produced a transient which was
also assigned to the corresponding 1,3-biradical (τ ) 15.1
( 1 ns, λmax ) 311 nm in heptane). This transient was
also unaffected by the addition of 2-methyl-1,3-butadiene
(1 mol L-1). trans-1,2-Bis(4-methoxyphenyl)cyclopro-
pane, which isomerizes to the cis isomer upon direct
excitation, afforded only a modest fluorescence and no
observable absorption. The lifetimes and λmax of the
transient species generated from the excitation of 1a , 2a ,
An increase in isc rate generated by an increase in the
ionic character of the termini within a biradical was
predicted by Salem and Rowland.1 To examine this
proposed effect, a laser flash photolysis study of 1-(4-
acetylphenyl)-2-(4-substituted phenyl)cyclopropanes (1e-
i) was conducted. These compounds also showed tran-
sients which were assigned to the corresponding 1,3-
biradicals. With the exception of 1-(4-acetylphenyl)-2-
(4-bromophenyl)cyclopropane (1h ), which produced a
short-lived transient due most likely to the heavy atom
effect,15,28,29,45 these cyclopropanes produced transients
with similar lifetimes and absorption maxima. These
results are summarized in Table 7 which also includes
quantum yields measured for the trans-cis photoisomer-
ization.
(36) Turro, N. J . Modern Molecular Photochemistry; Benjamin:
Menlo Park, CA, 1978.
(37) Porter, G.; Strachan, E. Trans. Faraday Soc. 1958, 54, 1595.
(38) Lutz, H.; Lindqvist, L. J . Chem. Soc., Chem. Commun. 1971,
493.
(39) Lutz, H.; Breheret, E.; Lindqvist, L. J . Phys. Chem. 1973, 77,
1758.
(42) Karki, S. B.; Dinnocenzo, J . P.; Farid, S.; Goodman, J . L.; Gould,
I. R.; Zona, T. A. J . Am. Chem. Soc. 1997, 119, 431.
(43) Kita, F.; Nau, W. M.; Adam, W.; Wirz, J . J . Am. Chem. Soc.
1995, 117, 8670.
(40) Caldwell, R. A.; Cao, C. V. J . Am. Chem. Soc. 1981, 103, 3594.
(41) Caldwell, R. A.; Cao, C. V. J . Am. Chem. Soc. 1982, 104, 6174.
(44) Nau, W. M. Personal Communication.
(45) Caldwell, R. A. Pure Appl. Chem. 1984, 56, 1167.

3180 J . Org. Chem., Vol. 63, No. 10, 1998
Ichinose et al.
Sch em e 3
Sch em e 4
F igu r e 4. Arrhenius plot for thermal isomerization of 2a to
1a .
Conformational effects on the lifetime of 1,3-biradicals
were examined through the use of a series of sterically
restricted precursors, 1j-k . Ring opening of 1j-k will
yield 1,3-biradicals where the intramolecular rotation of
one radical terminal is limited so that the two p orbitals
at the termini will have a nearly perpendicular config-
uration. Intramolecular rotation will be allowed for only
one terminal; the other terminal will remain fixed. This
arrangement will allow conformational effects to be
probed.1 Upon laser flash photolysis, these compounds
also showed efficient generation of transients assigned
as 1,3-biradicals. Compounds 1j-k did not show any
considerable photoisomerization, probably due to the
instability of the cis isomers. The lifetimes were slightly
shorter (14 and 12 ns for 1j and 1k , respectively) than
that generated from 1a or 1b.
Th er m a l Isom er iza tion of 1,2-Bis(4-a cetylp h en yl)-
cyclop r op a n e. As described above, the photoisomer-
ization of 1,2-diarylcyclopropanes bearing an acetylphen-
yl group proceeds via the corresponding triplet biradicals.
It has been proposed that thermal isomerization of
cyclopropane derivatives also proceeds via a biradical
mechanism. In this section, the thermal isomerization
of 2a to 1a was studied to evaluate the energy surface of
the reaction via the singlet biradical.
Mech a n ism . Acetophenone derivatives typically have
triplet energies of 72-74 kcal mol-1. Because the com-
pounds, 1a -k and 2a , all have either one or two
acetylphenyl groups at vicinal positions, the triplet
energies for these compounds are expected to be within
this range. The excited triplet of acetophenone is
quenched by conjugated dienes such as 2-methyl-1,3-
butadiene (ET ) 60 kcal mol-1) which in benzene quenches
the acetophenone triplet at an almost diffusion controlled
rate (kq ) 5 × 109 mol-1 s-1).47
A plausible mechanism for the photoisomerization of
1 and 2 is shown in Schemes 3 and 4. The first step is
1
excitation of the acetylphenyl group, giving 1* (or 12*)
most probably in the 1n,π* state, followed by rapid
intersystem crossing (1010-1011 s-1) and internal conver-
3
sion to the lowest triplet of the acetylphenyl group, 1*.
The thermal isomerization of 2a to 1a was observed
by heating a diethylene glycol solution containing 2a at
193 °C for 2 h. The photoisomerization products, 3a and
4a , were not observed. From the first-order plot of the
isomerization reaction, the rate constant for isomeriza-
tion was estimated as 4.33 × 10-5 s-1. Rate constants
were measured at ambient temperature. The activation
parameters for the thermal isomerization of 2a to 1a
were obtained from an Arrhenius plot of the rate con-
stants (Figure 4) where ∆H ) 33.4 ( 1 kcal mol-1 and
log A ) 11.2. These values are in accord with the
reported values for 1,2-diarylcyclopropanes.46
The absence of quenching of this triplet state by 2-meth-
yl-1,3-butadiene and nonphosphorescent nature of 1 (or
3
2) are due to the extremely short lifetime of 1*. In the
case of 1d , the quenching coefficient, kqτ, was ca. 0.5,
3
giving the lifetime of 1d * as 0.1 ns. The other cyclo-
propanes did not show considerable quenching, and
therefore the lifetime of 31* will be less than 0.1 ns. This
value is too short to detect with our present system for
laser flash photolysis. The cyclopropane ring of 31*
rapidly opens to give the 1,3-biradical (BR). In the
triplet-sensitized photoreaction of 1a , a decrease in the
(47) Scaiano, J . C.; Leigh, W. J .; Meador, M. A.; Wagner, P. J . J .
Am. Chem. Soc. 1985, 107, 5806.
(46) Rodewald, L. B.; DePuy, C. H. Tetrahedron Lett. 1964, 2951.

Generation of Triplet 1,3-Biradicals
J . Org. Chem., Vol. 63, No. 10, 1998 3181
triplet energy of the sensitizer lowered the efficiency of
isomerization. This indicates that the ring opening of
1a requires excitation of the acetylphenyl group and
illustrates the absence of nonvertical excitation to the
biradical state. The spin multiplicity of the biradical was
assigned as triplet (3BR) on the basis of the results of
dioxygen trapping and from the generation process.
importance of the closeness between the two termini in
intersystem crossing. Since the orientational factor for
mutual interaction between termini in 1,3-biradicals is
determined only by two rotational angles, through-bond
interaction must be a dominant factor in the isc mech-
anism rather than through-space interaction. We also
have observed a lack of conformational effect in the isc
rate between sterically restricted and unrestricted 1,4-
biradicals.3,17,45,52,53 This observation will be discussed in
a separate paper.
En er gy Cor r ela tion . The C-C bond cleavage of
3
4-alkylacetophenones from the π,π* state is symmetri-
3
cally allowed, but forbidden from the n,π* state because
a spin on the n orbital in the 3n,π* state will be conserved
through the bond cleavage leading to a very high energy
species in the latter case.48 Even if the reaction origi-
nated in the lowest 3n,π* state of an acetophenone
chromophore, the C-C bond stretching will couple best
to the 3π,π* state. The reaction would then simply
involve a surface crossing to the latter, which will be
lower energy in the transition state region in any case.
Cleavage of the C-C bond (normal bonds ca. 85 kcal
mol-1) will be facilitated by the generation of two benzylic
radical sites (12.5 × 2 kcal mol-1) and by the strain relief
resulting from opening the cyclopropane ring (27 kcal
mol-1). The estimated triplet energy of these compounds
is 72 kcal mol-1. Thus, the energy change for the bond
cleavage from the excited triplet state will be negative:
(85 - 72 - 12.5 × 2 - 27) ) -39 kcal mol-1. We neglect
the small effect of the two aryl substituents. The
resultant radical sites have been predicted to contain no
strain energy.49
F a te of BR. After intersystem crossing from BR to
1BR, ring closure occurs, with the biradical isomerizing
to 1 or 2 and 3 or 4. For simple trimethylene derivatives,
this process tends to lead to the formation of cyclopro-
pane, yielding less olefin.50 In the present work, the
quantum yield for the formation of 3 and 4 is also much
smaller than that of cis-trans isomerization, though
chemical yield is considerably high. This can be ex-
plained by repetitive excitation of 1a and 2a ; the triplet
olefin has no path for hydrogen migration as that which
is known for the excited singlet 1,3-diphenylpropene
derivatives.51 The absence of 3a or 4a upon thermolysis
of 2a also supports the conclusion that the energy
surfaces of the present 1,3-diaryl-substituted 1,3-biradi-
cals are similar to those of the simpler trimethylene.
Com p u ta tion of Sp in -Or bit Cou p lin g Con sta n ts
for Mod el Tr im eth ylen es. In view of the virtual
absence of a polar substituent effect on the isc rate of
the triplet biradicals, we calculated spin-orbit coupling
constants (SOCC) for 1-amino-3-cyanotrimethylene and,
for comparison, unsubstituted trimethylene. Spin-orbit
coupling in biradicals is expected1 to depend on the “ionic
character in the singlet wave function,” by which is meant
the fractional contribution of the VB hole-pair configu-
ration.54 If the anticipated effect of ionic character in the
singlet biradical wave function were ever to be mani-
fested in a study of polar substituents, surely 1-amino-
3-cyanotrimethylene would be among the best examples.
The absence of a polar substituent effect for the isc rate
of the diaryl-1,3-propanediyls here reported is in complete
accord with the outcome of our computations. We also
performed analogous computations for (aminomethyl)-
borane, H2NCH2BH2. The singlet in this case will be
essentially a pure hole-pair state, the “hole” on the boron
and the lone pair of electrons on the nitrogen, since that
form is in fact the one with zero formal charge on all
atoms. The triplet would necessarily be of the “dot-dot”
type (one electron mainly on B and one on N). Clearly
the “ionic character in the singlet wave function” will be
maximized for such a case, completely analogously to the
perpendicular aminoborane case described by Michl and
Bonacic-Koutecky,54 and this system will show the maxi-
mum SOCC to be expected for a trimethylene analogue
in the absence of heavy atom effects.
We used the 2-in-2 CAS-MCSCF protocol and the
spin-orbit coupling program of Furlani and King55 with
the 3-21G basis set as we have previously described.56,57
For 1-amino-3-cyanotrimethylene and trimethylene, the
C1-C2 and C2-C3 distances were taken as 1.502 Å, the
sp2 C-H bond distances were 1.07 Å, the sp3 C-H bond
distances were 1.09 Å, the bond angles at C1 and C3 were
taken as 120°, and the bond angles at C2 were taken as
tetrahedral (109.5°). For the amino group, the C1-N
distance was 1.36 Å, the N-H distances were 0.96 Å, and
planarity was maintained about the C-N bond. The
cyano group was kept linear, with a C3-C distance of
1.31 Å and a C-N triple bond distance of 1.16 Å. For
(aminomethyl)borane, geometries were kept identical to
those for trimethylene; the atomic numbers of the 1 and
3 carbons were changed to 7 and 5, respectively. The
only degrees of freedom studied in either case were the
torsions around the 1-2 and 2-3 bonds.
3
3
A few examples of substituent effect on the lifetime of
1,4-biradicals have been reported for Norrish II biradicals
and photoenols. A substituent effect for 1,3-biradicals
bearing a 4-substituted phenyl group, with the exception
of a heavy atom effect by the 4-Br group, was not
observed in this work. This suggests that any polar effect
induced by the substituents does not affect the intersys-
tem crossing rates of the present 1,3-biradicals; rather,
the interaction between the two termini may be dominant
due to their relative closeness. The heavy atom effect
showed that considerable spin density is present at the
4-position of the phenyl ring.
The lack of any observed conformational effect in the
1,3-biradicals obtained from 1j-k also illustrates the
(52) Caldwell, R. A.; Gupta, S. C. J . Am. Chem. Soc. 1989, 111, 740.
(53) Caldwell, R. A.; Sink, R. M.; Ichinose, N.; Mizuno, K.; Otsuji,
Y. To be published.
(54) Michl, J .; Bonacic-Koutecky, V. In Electronic Aspects of Organic
Photochemistry; J ohn Wiley & Sons: New York, 1990; pp 179-200.
(55) Furlani, T. R.; King, H. F. J . Chem. Phys. 1985, 82, 5577.
(56) Caldwell, R. A.; Carlacci, L.; Doubleday, C. E., J r.; Furlani, T.
R.; King, H. F.; McIver, J . W., J r. J . Am. Chem. Soc. 1988, 110, 6901.
(57) Caldwell, R. A.; J acobs, L. D.; Furlani, T. R.; Nalley, E. A.;
Laboy, J . J . Am. Chem. Soc. 1992, 114, 1623.
(48) Ichinose, N.; Mizuno, K.; Otsuji, Y.; Tachikawa, H. Tetrahedron
Lett. 1994, 35, 587.
(49) Ruchardt, C.; Beckhaus, H. D. Angew. Chem., Int. Ed. Engl.
1980, 19, 429.
(50) Doubleday, C.; McIver, J . W.; Page, M. J . Am. Chem. Soc. 1982,
104, 6533.
(51) Griffin, G. W.; Marcantonio, A. F.; Kristinsson, H.; Petterson,
R. C.; Irving, C. S. Tetrahedron Lett. 1965, 2951.

3182 J . Org. Chem., Vol. 63, No. 10, 1998
Ichinose et al.
HMO spin density on the carbon bearing the substituent.
Since the spin density is 1 at a simple alkyl radical
terminus but only 1/7 at a para position of a benzyl
radical, we predict ca. 1.5(1/7) ) 1.06 for the acceleration
of 1-(p-aminophenyl)-3-(p-cyanophenyl)-1,3-propanediyl
relative to 1,3-diphenyl-1,3-propanediyl. This is about
the precision of our measurements in the present series.
The much larger values of SOCC and thus slope were
expected for (aminomethyl)borane. By themselves they
would predict much shorter lifetimes for analogous triplet
biradicals, but we find large (ca. 30-40 kcal/mol) T1-S0
gaps which would tend to increase lifetime1,3,45,54,59,60 and
thus counteract the effect.
The cases we examined were intended to probe when
polar substituents can accentuate the “hole-pair” terms
enough to cause observable effects on isc rates. We
conclude that such will occur only in extreme cases and
that in an ordinary Hammett substituent series the effect
will not be manifested. A substituent series for Norrish
II biradicals, which are 1,4-butanediyl derivatives, simi-
larly showed negligible polar substituent effects.45 In
light of the present calculations, we believe that observa-
tion was to be expected. A surprisingly long lifetime
assigned to a Norrish II biradical analogue, the triplet
species derived from â-dimethylaminopropiophenone,60
for which the contribution of the hole-pair state to the
singlet must be substantial, deserves reexamination. It
is, however, possible that the long lifetime is caused by
a large T1-S0 gap as pointed out above.
Is the lifetime of ca. 14 ns reasonable for the 1,3-diaryl-
1,3-trimethylene biradical here described? Indeed so.
SOC values of ca. 1 cm-1 are attainable at several low
energy geometries (Figure 5 and supplementary table).
By contrast, 1,4-diaryl-1,4-tetramethylene shows SOC
values typically below 0.1 cm-1 for low energy geom-
etries.61 Since 1,4-diaryl-1,4-tetramethylenes show trip-
let lifetimes in the 200 ns range3, the value for trimeth-
ylenes would be expected to be much shorter than 200
ns. The variation of SOC with conformation of these
highly flexible biradicals prevents any more quantitative
comparison without further studies. Recent reports have
reached similar conclusions both theoretically and ex-
perimentally. Kita et al. studied 1,3-diaryl-1,3-cyclopen-
tanediyl biradicals which avoid the conformational flex-
ibility inherent in the present work.43 Michl has developed
a more rigorous version of the Salem-Rowland rules
which highlights inter alia the present point and the
crucial role of through bond coupling of SOC.62 Bock-
mann and co-workers have developed a formatism for
calculating SOC effects in organic molecules based on the
Rumer spin eigenfunction and the second quantization.63
Zimmerman has also studied SOC in R,ω-alkanediyls.64
F igu r e 5. Spin-orbit coupling constants (SOCC) for H2BCH2-
NH2 (b) and 1-amino-3-cyano-1,3-propanediyl (9) plotted
against SOCC for trimethylene at the same geometry. Trend
lines: y ) 2.10x + 0.85 (H2BCH2NH2) and 1.19x - 0.015 (1-
amino-3-cyano-1,3-propanediyl).
Values for trimethylene agreed in essence with those
previously reported for slightly different bond lengths and
angles.27 Figure 5 shows the plot of SOCC for 1-amino-
3-cyanotrimethylene and (aminomethyl)borane against
that for trimethylene at identical torsional angles
(geometry details and a tabulation of the results are
available as Supporting Information). Up to four geom-
etries for 1-amino-3-cyanotrimethylene and two for (ami-
nomethyl)borane correspond to a single geometry of
trimethylene due to the asymmetries introduced by the
substituents.
We note that the basic trend in either case is rough
proportionality of SOCC to that for the trimethylene
analogue. While there are conformational issues embed-
ded in Figure 5, there are no experimental results which
make their further examination profitable. We take the
Franck-Condon component of the Golden Rule equation
as constant. If the isc is assumed to arise from all
conformations indiscriminately, the square of the slope
of either plot then affords the predicted ratio of isc rate
to that of trimethylene.58 The ratios are 1.41 for 1-amino-
3-cyanotrimethylene and 2.1 for (aminomethyl)borane.
If we assume that isc arises only at the conformation
affording the SOCC maximum, the ratios are ca. 1.5 and
5.0, respectively. As crude as the assumptions we make
are, the difference between them is insignificant. Pre-
sumably substituent effects on isc rates would follow
Exp er im en ta l Section
Gen er a l. General experimental information was described
in a previous paper.65
(59) Caldwell, R. A.; Singh, M. J . Am. Chem. Soc. 1983, 105, 5139.
(60) Encinas, M. V.; Scaiano, J . C. J . Am. Chem. Soc. 1979, 101,
2146.
(61) Caldwell, R. A.; J acobs, L. D. Unpublished results.
(62) Michl, J . J . Am. Chem. Soc. 1996, 118, 3568.
(63) Bockmann, M.; Klessinger, M.; Zerner, M. J . Phys. Chem. 1996,
100, 10570.
(64) Zimmerman, H. E.; Kutaleladze, A. G. J . Am. Chem. Soc. 1996,
118, 249.
(58) The singlet-triplet gap is expected to affect the isc rate.
Unfortunately the calculations we performed are not able to give
reliable estimates for S-T gaps. However, we note that there is little
difference between the S-T gaps calculated for trimethylene and
1-amino-3-cyanotrimethylene at any geometry. Our calculations thus
provide no evidence that differences in the S-T gap with substituent
contribute to substituent effects on isc for substituted trimethylenes.
(65) Mizuno, K.; Ichinose, N.; Otsuji, Y. J . Org. Chem. 1992, 57,
1855.

Generation of Triplet 1,3-Biradicals
J . Org. Chem., Vol. 63, No. 10, 1998 3183
Syn th esis of 1,2-Dia r ylcyclop r op a n es Bea r in g Acetyl-
p h en yl Gr ou p s. The introduction of an acetyl group to the
phenyl ring was carried out by two methods: (1) through a
Friedel-Crafts acylation (method A) and (2) by methylation
of a cyano group through the use of a Grignard reagent
(method B).
the residue was distilled under reduced pressure using a glass
tube oven (9.3 Pa; 120 °C). 1-(2-Acetylphenyl)-2-phenylcyclo-
propane was obtained as a mixture of trans and cis isomers
(1c and 2c) in 67% yield. Pure 1c and 2c were obtained by
1
careful distillation of the mixture. 1c: oil; H NMR (CDCl3)
δ 1.27 (dd, 2H, J ) 6 and 9 Hz), 1.83-2.83 (m, 2H), 2.36 (s,
3H), 6.73-7.53 (m, 4H), 7.07 (s, 5H); MS (70 eV) m/z 236 (M+).
Anal. Calcd for C17H16O: C, 86.40; H, 6.83. Found: C, 86.20;
Syn th esis of tr a n s- a n d cis-1,2-Bis(4-a cetylp h en yl)-
cyclop r op a n es (1a a n d 2a ). Gen er a l P r oced u r e for
Meth od A. Small portions of AlCl3 were added to a mixture
containing acetyl chloride and 1,2-diphenylcyclopropanes66 in
CS2 within 30 min at room temperature. After 2 h of stirring,
the reaction mixture was poured into a concentrated HCl/ice
mixture and extracted with benzene:ether (1:1). The organic
layer was then dried over Na2SO4. The solvent was evaporated
and the residue chromatographed on silica gel. The unreacted
1,2-diphenylcyclopropanes were recovered from the hexanes
eluent. From the hexane:benzene (1:1) eluent, 10-20% of 1-(4-
acetylphenyl)-2-phenylcyclopropanes (1b and 2b) were ob-
tained, and from the benzene:ethyl acetate (98:2) eluent, 30-
60% of 1a was obtained. Repeated recrystallization of 1a from
acetone-methanol yielded pure 1a : mp 171-172 °C; 1H NMR
(CDCl3) δ 1.65 (dd, 2H, J ) 6 and 9 Hz), 2.30 (dd, 2H), 2.52 (s,
6H), 7.50 (ABq, 8H, ∆ν ) 42 Hz, J ) 9 Hz); 13C NMR (CDCl3)
δ 197.59, 147.70, 135.13, 128.70, 125.73, 29.05, 19.50; IR (KBr)
1646, 1578, 1330, 1248, 1170 cm-1; MS (70 eV) m/z 278 (M+).
Anal. Calcd for C19H18O2: C, 81.98; H, 6.52. Found: C, 82.02;
H, 6.58.
2a was obtained by the photoreaction of 1a (see text).
Repeated recrystallization of the reaction mixture from hexane
gave pure 2a : mp 108.8-109.8 °C; 1H NMR (CDCl3) δ 1.52
(dd, 2H, J ) 6 and 9 Hz), 2.46 (s, 6H), 2.55 (dd, 2H), 7.36 (ABq,
8H, ∆ν ) 42 Hz, J ) 9 Hz); 13C NMR (CDCl3) δ 197.77, 143.75,
134.93, 128.94, 127.97, 26.48, 25.11, 11.97; IR (KBr) 1650,
1590, 1346, 1260, 1172, 832 cm-1; MS (70 eV) m/z 278 (M+).
Anal. Calcd for C19H18O2: C, 81.98; H, 6.52. Found: C, 81.90;
H, 6.43.
1
H, 6.48. 2c: oil; H NMR (CDCl3) δ 1.26 (dd, 2H, J ) 6 and
9H), 1.88 (s, 3H), 2.53-2.94 (m, 2H), 6.40-7.53 (m, 9H); MS
(70 eV) m/z 236 (M+).
Other cyclopropanes were prepared by similar methods.
P h otoisom er iza tion of 1a a n d 2a . Gen er a l P r oced u r e.
A benzene solution of 1a (0.01 mol L-1, 8 mL) was degassed
by repeated freeze-pump-thaw cycles (four times) and then
sealed in a Pyrex tube (Φ 10 mm). The sample was irradiated
at 313 nm after which the reaction mixture was analyzed by
HPLC (Daiseru Nucleosil Φ 4.6 mm × 150 mm, MeOH-water
(7:3)). 1H NMR analysis revealed the formation of 2a , 3a , and
4a . 2a was irradiated in a similar manner. Quantum yields
were measured by the method of Parker.68 The quantum
yields obtained from argon-saturated solutions agreed with
those obtained from degassed solutions.
Sen sitized P h otor ea ction of 1a . Benzene solutions of 1a
(0.01 mol L-1) containing triplet sensitizers such as acetophe-
none, benzophenone, and benzil were irradiated at the same
time using a merry-go-round apparatus. The concentration
of the sensitizer, determined by its extinction coefficient at 313
nm, was selected to absorb 90% of the light (OD ) 20, ꢀ1a (313
nm) ) 100). The quantum yields for benzophenone or benzil
sensitized reactions were determined by the yields of 2a
relative to the sample containing acetophenone (Φ ) 0.40).
The results are given in Table 3.
P r ep a r a tive P h otor ea ction of 1a . A benzene solution
(50 mL) of 1a (1.7 mmol) was irradiated under an argon
atmosphere for 5 h, after which the solvent was evaporated.
1a was removed by recrystallization of the residue from
hexane. Repeated recrystallization of the mother liquor from
hexane yielded pure 2a . A similar irradiation of 1a (0.4 mmol)
in benzene was carried out under an argon atmosphere for 54
h. 1a and 2a were removed by recrystallization from hexane.
The mother liquor was evaporated and the residue dissolved
in dimethoxyethane (0.5 mL). A mixture of 3a and 4a (2:1)
was obtained by preparative HPLC separation in 42% yield.
(E)-1,3-Bis(4-acetylphenyl)propene (3a ): 1H NMR (CDCl3) δ
2.60 (s, 6H), 3.73 (d, 2H, J ) 7.3 Hz), 5.96 (dt, 1H, J ) 7.3 and
11.3 Hz), 6.67 (d, 1H, J ) 11.3 Hz), 7.28-7.95 (m, 8H); MS
(70 eV) m/z 278 (M+). (Z)-1,3-Bis(4-acetylphenyl)propene
(4a ): 1H NMR (CDCl3) δ 2.60 (s, 6H), 3.64 (d, 2H, J ) 4.9 Hz),
6.49 (m, 2H), 7.28-7.95 (m, 8H); MS (70 eV) m/z 278 (M+).
P h otor ea ction of 3a a n d 4a . A benzene solution of a
mixture of 3a and 4a (3a /4a ) 2/1) (0.015 mol L-1, 4 mL) was
irradiated in a Pyrex tube for 4 days under an argon atmo-
sphere. The reaction mixture was analyzed by GC, HPLC, and
1H NMR. A mixture of 3a and 4a was quantitatively recov-
ered.
trans-1-(4-Acetylphenyl)-2-phenylcyclopropane (1b) was ob-
tained by recrystallization of a mixture of 1b and 2b from
1
hexane. 1b: mp 50.3-51.1 °C; H NMR (CDCl3) δ 1.43 (dd,
2H, J ) 6 and 9 Hz), 2.13 (dd, 2H), 2.45 (s, 3H), 7.08 (s, 5H),
7.42 (ABq, 4H, ∆ν ) 42 Hz, J ) 9 Hz); 13C NMR (CDCl3) δ
197.63, 148.58, 141.74, 134.85, 128.62, 126.07, 125.79, 125.63,
29.05, 28.17, 26.54, 18.92; IR (KBr) 1650, 1580, 1340, 1260,
1175, 800, 735, 680 cm-1; MS (70 eV) m/z 236 (M+). Anal.
Calcd for C17H16O: C, 86.40; H, 6.83. Found: C, 86.30; H, 6.55.
cis-1-(4-Acetylphenyl)-2-phenylcyclopropane (2b) was iso-
lated by through preparative HPLC. 2b: oil; 1H NMR (CDCl3)
δ 1.42 (dd, 2H, J ) 6 and 9 Hz), 2.40 (dd, 2H), 2.42 (s, 3H),
6.98 (s, 5H), 7.36 (ABq, 4H, ∆ν ) 42 Hz, J ) 9 Hz); MS (70
eV) m/z 236 (M+).
Syn th esis of tr a n s-1-(2-Acetylp h en yl)-2-p h en ylcyclo-
p r op a n e (1c). Gen er a l P r oced u r e for Meth od B. 1-(2-
Bromophenyl)-2-phenylcyclopropane was synthesized from
2-bromobenzaldehyde and acetophenone following a similar
method as described in the literature.66 1c: 1H NMR (CDCl3)
δ 1.30-1.35 (m, 2H), 2.40-2.60 (m, 2H), 6.67-7.90 (m, 9H);
MS (70 eV) m/z 274, 272 (M+). This compound (9.7 mmol) was
refluxed with CuCN (15.1 mmol) in N-methyl-2-pyrrolidone
(20 mL) for 3 h. The reaction mixture was then extracted with
benzene-ether and washed with water, and the organic layer
was dried over Na2SO4. After removal of the solvent, the
residue was chromatographed on silica gel. From hexane
eluent, 1-(2-cyanophenyl)-2-phenylcyclopropane was obtained
in a 94% yield: 1H NMR (CDCl3) δ 1.40-1.60 (m, 2H), 2.25-
2.65 (m, 2H), 6.70-7.60 (m, 9H); MS (70 eV) m/z 221 (M+). To
a solution of CH3MgI (10 mmol) in ether (20 mL) was added
1-(2-cyanophenyl)-2-phenyl-cyclopropane (2.7 mmol) in ether
(10 mL) at room temperature. Dry benzene (10 mL) was added
to the reaction mixture which was refluxed for 5 h and then
stirred for 12 h. Acetic acid was added, and the mixture was
heated to reflux for 4 h. The organic layer was washed with
water and dried over Na2SO4. After the removal of the solvent,
P h otor ea ction of 1a in th e P r esen ce of Oxygen .
A
benzene solution of 1a (0.015 mol L-1, 4 mL) was irradiated
in a Pyrex tube at 313 nm with vigorous oxygen bubbling. After
the consumption of 1a and 2a within 1 h, the reaction mixture
was analyzed by 1H NMR and GC-MS. From the 1H NMR
spectrum, the formation of 3,5-bis(4-acetylphenyl)-1,2-diox-
olane (5a ), 4-acetylbenzaldehyde (6a ), and 1,4-diacetylbenzene
(7a ) was confirmed. 5a was formed as a mixture of cis and
trans isomers (cis/trans ) 7/3). The structure of 5a was
determined by the typical chemical shifts of 1,2-dioxolanes and
their coupling pattern in the 1H NMR spectrum in comparison
with those of 3,5-diaryl-1,2-dioxolanes, which we have reported
previously.35 The formation of 6a and 7a was also confirmed
by GC-MS. 3,5-Bis(4-acetylphenyl)-1,2-dioxolane (5a ): 1H
NMR (CDCl3) (trans) δ 3.13 (t, 2H, J ) 6.5 Hz), 5.48 (t, 2H);
(67) Foyt, D. C. Comput. Chem. 1981, 5, 49.
(68) Hatchard, C. G.; Parker, C. A. Proc. R. Soc. London 1956, A235,
518.
(66) Hixson, S. S.; Garrett, D. W. J . Am. Chem. Soc. 1974, 96, 4872.

3184 J . Org. Chem., Vol. 63, No. 10, 1998
Ichinose et al.
(cis) 3.60 (dt, 1H, J ) 7.6 and 12.2 Hz), 5.50 (t, 2H); other
peaks were not assigned due to overlapping signals. 4-Acetyl-
benzaldehyde (6a ): 1H NMR (CDCl3) δ 2.62 (s, 3H), 7.96 (s,
4H), 10.10 (s, 1H); MS (20 eV) m/z 148 (M+). 1,4-Dia cetyl-
ben zen e (7a ). To a solution of CH3MgI (40 mmol) in dry ether
(20 mL) was added 1,4-dicyanobenzene (10 mmol) in dry
benzene (50 mL), and the mixture was then refluxed for 4 h.
The solution was hydrolyzed with 50% AcOH and refluxed for
4 h. The organic layer was dried over Na2SO4 and the solvent
evaporated to give 1,4-diacetylbenzene: 1H NMR (CDCl3) δ
2.64 (s, 6H), 7.99 (s, 4H); MS (20 eV) m/z 162 (M+).
La ser F la sh P h otolysis. The laser flash photolysis ex-
periments were carried out by well-known procedures.3,67 The
transients from compounds 1a -b, 1e-k , 2a , and 3a /4a were
monitored at 310 or 315 nm. The concentration of the
compound was (1-4) × 10-4 mol L-1. These experiments were
performed under nitrogen atmosphere.
Th er m a l Isom er iza tion of 2a . A diethylene glycol solu-
tion of 2a (1.29 × 10-3 mol L-1, 0.3 mL) was placed in a Φ 5
mm NMR tube and saturated with argon. The sample was
then heated to 174 °C ((1 °C) in an oil bath. The reaction
mixture was analyzed by HPLC, and the yield of 1a was
determined. The logarithm of the concentration of 2a , log[2a ],
vs reaction time was plotted. The rate constant for the thermal
isomerization of 2a to 1a was estimated from the slope of the
plot in the initial stage of the reaction, where the reverse
reaction of 1a to 2a can be neglected. Rate constants at
various temperatures (4 points) were obtained in the same
manner. These rate constants were plotted as log k vs
reciprocal of absolute temperature, T-1, to generate an Arrhe-
nius plot. From the slope and intercept of the plot, the
activation energy, Ea, and the preexponential factor, A, were
obtained.
Ack n ow led gm en t. This work was supported in part
by National Science Foundation Grants CHE-8820268,
9121313, and 9422824. Flash photolysis experiments
were performed at the Center for Fast Kinetics Research
at The University of Texas at Austin (supported by NIH
Grant RR-00886 from the Biotechnology Branch of the
Division of Research Resources and by The University
of Texas at Austin) and at The University of Texas at
Dallas. K.M. is grateful for a Grant-in-Aid for Scientific
Research from the Ministry of Education, Science,
Sports and Culture of J apan.
Su p p or tin g In for m a tion Ava ila ble: Tables showing
SOC values between S and T states of the three biradicals
studied, with structure and geometry variations included.
Spectral properties for 1d -k (8 pages). This material is
contained in libraries on microfiche, immediately follows this
article in the microfilm version of the journal, and can be
ordered from the ACS; see any current masthead page for
ordering information.
J O971111S