Influence of the Geminal Biscyclopropyl Group in the Reactions of Trimethylenemethane: Reversible Intersystem Crossing

Article abstract of DOI:10.1021/jo9513727

The azo compound precursor to Berson's trimethylenemethane (TMM) biradical system, 2-methylene-1,3-cyclopentanediyl, was substituted with two cyclopropyl groups on the exomethylene.Generation of the TMM species by pyrolysis, direct photolysis, and benzophenone-sensitized photolysis resulted in formation of both cyclopropane ring-opened monomer and cyclopropane ring-unopened dimers whose ratio depended directly on temperature and extent of dilution.The ground state appears to be triplet on the basis of ESR spectroscopy.The results contrast with previous work (Waldemar Adam) showing that monocyclopropyl substitution on Berson's TMM species gives both ring-opened monomer and ring-unopened dimers upon pyrolysis and gives mostly ring-unopened dimer upon direct photolysis at 0-20 deg C in what appear to be concentration-independent reactions.The observations with the biscyclopropyl-substituted species are consistent with steric effects retardind dimerization of the triplet allowing reversible intersystem crossing back to singlet and perhaps with reduction in the singlet-triplet gap as well.

Full text of DOI:10.1021/jo9513727

J . Org. Chem. 1996, 61, 1399-1404
1399
In flu en ce of th e Gem in a l Biscyclop r op yl Gr ou p in th e Rea ction s of
Tr im eth ylen em eth a n e: Rever sible In ter system Cr ossin g
Gitendra C. Paul* and J oseph J . Gajewski*
Department of Chemistry, Indiana University, Bloomington, Indiana 47405
Received J uly 26, 1995X
The azo compound precursor to Berson’s trimethylenemethane (TMM) biradical system, 2-meth-
ylene-1,3-cyclopentanediyl, was substituted with two cyclopropyl groups on the exomethylene.
Generation of the TMM species by pyrolysis, direct photolysis, and benzophenone-sensitized
photolysis resulted in formation of both cyclopropane ring-opened monomer and cyclopropane ring-
unopened dimers whose ratio depended directly on temperature and extent of dilution. The ground
state appears to be triplet on the basis of ESR spectrometry. The results contrast with previous
work (Waldemar Adam) showing that monocyclopropyl substitution on Berson’s TMM species gives
both ring-opened monomer and ring-unopened dimers upon pyrolysis and gives mostly ring-
unopened dimer upon direct photolysis at 0-20 °C in what appear to be concentration-independent
reactions. The observations with the biscyclopropyl-substituted species are consistent with steric
effects retarding dimerization of the triplet allowing reversible intersystem crossing back to singlet
and perhaps with reduction in the singlet-triplet gap as well.
Sch em e 1
The determination of electron multiplicity in cross-
conjugated organic biradical systems is an important
pursuit and is one which is made more difficult if the
species cannot be generated under matrix isolation
conditions and studied by electron spin or optical spec-
troscopy. The tetramethyl-m-xylylene generated by isobu-
tylidene addition to dimethylfulvene1 is one such case.
Our recent work2 has shown that addition of dioxygen
suppresses formation of the dimeric product while not
inhibiting generation and reaction of isobutylidene.
Unfortunately, both singlet and triplet biradicals are
known to react with oxygen so a distinction cannot be
made between spin states on the basis of reactivity with
oxygen except by comparison with some other reaction
that clearly indicates the presence of at least two
intermediates.
Sch em e 2
Waldemar Adam recently introduced the criterion of
ring opening of cyclopropyl-substituted cross-conjugated
biradicals as evidence for involvement of singlet states.3
Thus, thermal and direct photolysis of a cyclopropyl
derivative of Berson’s azo compound precursor4 to TMM,
1, gives monomeric product resulting from cyclopropyl-
carbinyl radical rearrangement and closure along with
cyclopropane ring-unopened dimerssall as a function of
temperaturesbut triplet sensitized photolysis gives al-
most exclusively dimeric products with no cyclopropane
ring opening (Scheme 1).
It was recognized (a) that the ring-opened singlet
biradical would be roughly 6 kcal/mol more stable than
the most stable singlet TMM species which has an
orthogonal π system, (b) that the ring-opened triplet
biradical would have nearly the same energy as the
singlet since the radical sites are not coupled directly,
and (c) that the planar triplet TMM has been calculated
to be 14 kcal/mol more stable than the perpendicular
singlet.3 It was then suggested that the activation energy
for cyclopropylcarbinyl radical ring opening (ca. 5-8 kcal/
mol, see below) must be added to the difference in energy
between the triplet TMM and the ring-opened triplet so
that the overall preference for ring opening from singlet
was greater than that from triplet TMM (Scheme 2). It
should be noted here that the relative activation energies
for ring opening in the two states may not be as relevant
as the rate of ring opening vs intersystem crossing in the
X Abstract published in Advance ACS Abstracts, February 1, 1996.
(1) Gajewski, J . J .; Chang, M. J .; Stang, P. J .: Fisk, T. E. J . Am.
Chem. Soc. 1980, 102, 2096.
(2) Gajewski, J . J .; Paul, G. C.; Chang, M. J .; Gortva, A. M. J . Am.
Chem. Soc. 1994, 116, 5150.
(3) Adam, W.; Finzel, R. J . Am. Chem. Soc. 1992, 114, 4563.
(4) (a) Berson, J . A.; Bushby, R. J .; McBride, J . M.; Tremelling, M.
J . Am. Chem. Soc. 1971, 93, 1544. (b) Corwin, L. R.; McDaniel, D. M.;
Bushby, R. J .; Berson, J . A. J . Am. Chem. Soc. 1980, 102, 276. See
also: Berson, J . A. In Diradicals; Borden, W. T., Ed.; Wiley: New York,
1982; pp 151-194. (c) Berson, J . A. In The Chemistry of Quinonoid
Couponds, Vol. II; Patai, S., Rappoport, Z., Eds.; J ohn Wiley & Sons
Ltd: New York, 1988; pp 481-485. (d) Rule, M.; Mondo, J . A.; Berson,
J . A. J . Am. Chem. Soc. 1982, 104, 2209. (e) Mazur, M. R.; Berson, J .
A. J . Am. Chem. Soc. 1982, 104, 2217. (f) Closs, G. L. J . Am. Chem.
Soc. 1971, 93, 1546.
0022-3263/96/1961-1399$12.00/0 © 1996 American Chemical Society

1400 J . Org. Chem., Vol. 61, No. 4, 1996
Paul and Gajewski
Sch em e 3
Sch em e 4
omeric methylenecyclopropane product. With such con-
siderations in mind, an effort was made to examine the
response of a TMM biradical system to more sterically
demanding substitution like that derived from 3.
singlet state and the relative rate of ring opening vs
dimerization in the triplet state; thus the relative rates
of ring-opening in either state may be irrelevant.
Resu lts
There can be little doubt that both singlet and triplet
trimethylenemethane biradicals are involved in the dea-
zetization of 1, and in the corresponding geminallyl-
substituted dimethyl case Berson observed differential
trapping of at least two different intermediates by
fumarate as a function of concentration.4b And one of
these cannot be the ring-closed hydrocarbon, 5-isopropyli-
denebicyclo[2.1.0]pentane, 2, since the trapping reaction
of this species does not occur directly but via an inter-
mediate which is most likely one of the TMM biradical
species.4e Further, the ring-closed hydrocarbon gives
dimers at -55 °C (the same dimers as found in the
deazetizations) via a first-order reaction which has a
preexponential term consistent with reversible ring open-
ing and rate-determining intersystem crossing.4d Finally
the formation of dimers is accompanied by CIDNP signals
which almost certainly require a triplet precursor.4f The
reaction pathways of Scheme 3 involving an irreversible
cascade of a singlet to a triplet TMM species are
consistent with the observations and kinetic data.
It should be noted, however, that at higher tempera-
tures closure of the singlet biradical to the ring-closed
hydrocarbon is fast compared with intersystem crossing
but at low temperatures ring closure is slow compared
with intersystem crossing. This occurs because the
intersystem crossing rate is relatively temperature in-
dependent, but ring closure apparently has a finite
activation energy barrier. Finally, it must be recognized
that the intersystem crossing described in Scheme 3 must
be irreversible or effectively so. If this were not the case
then the ratio of product derived from singlet TMM
(cycloadducts 1 from Scheme 3 or ring-opened monomer
of Scheme 2) and product derived from triplet TMM
(dimers or cycloadducts 2 from Scheme 3) would neces-
sarily change with initial concentration. In trimethyl-
enemethane chemistry, reversible intersystem crossing
has not be observed, although it was suggested by Dowd5
with the parent case where the major product from low-
temperature photolysis of 4-methylenepyrazoline is me-
thylenecyclopropane under conditions where the sample
gives a triplet ESR spectrum. His suggestion therefore
is that the triplet must undergo intersystem crossing
back to singlet before closing, provided, of course, that
the ESR-active species is actually responsible for mon-
Syn th esis of Dia zen e 3. Diazene 3 was synthesized
by standard procedures3 following Scheme 4. Diethyl
azodicarboxylate (DEAD) adds in a Diels-Alder fashion
to the fulvene to give the adduct(s). Reduction of the
endocyclic double bond by catalytic hydrogenation re-
sulted in the formation of the biscarbamate. Hydrolysis
with KOH in refluxing i-PrOH and oxidation with CuCl2/
NH3 afforded diazene 3. Compound 3 is labile and slowly
decomposes at room temperature.
Th er m olysis a n d P h otolysis of 3. Thermolysis of
diazene 3 in benzene solvent in a vacuum degassed sealed
tube (boiling water bath) resulted in the formation of only
monomeric rearranged hydrocarbon 2-cyclopropylbicyclo-
[4.3.0]nona-1(9), 2-diene, 5 (Table 1 and Scheme 5). No
dimers or any other products were formed in detectable
amounts. Furthermore, an NMR sample left in chloro-
form-d at room temperature gave only ring-opened
monomer. Photolysis of 3 in a benzene solution at 0-5
°C, however, gives both monomeric product 5 and C24H32
dimers in roughly a 1:1 ratio. Direct photolysis starting
with a sample twice as dilute in benzene gives relatively
more ring-opened monomer (ratio monomer to dimer )
2.85:1). No attempt was made to determine the ratio as
a function of percent reaction since it is clear that this
should vary over the course of the reaction whereas the
total product provides an integration of this ratio.
Triplet-sensitized photolysis (benzophenone irradiation
from an argon ion laser at λ ) 364 nm to avoid direct
irradiation of 3) generates dimers and ring-opened
monomer. NMR spectrometry also revealed that the
dimers possessed intact cyclopropane rings since the
vinyl proton to secondary cyclopropane ring proton ratio
is roughly 15, which is characteristic of a mixed bridged-
fused dimer or a 1:1 mixture of fused and bridged dimer
or an indeterminant mixture of both possibilities by
analogy to previous work.4a Further, direct photolysis
at -78 °C in CFCl3 gives only ring-unopened dimers.
Finally direct photolysis of the diazene in cyclohexane
which was saturated with oxygen gas gave no hydrocar-
bon product; however, the residue of the reaction was
chromatographed to give a material whose NMR spec-
trum is consistent with 4,4-di(cyclopropyl-2,3-dioxabicyclo-
[3.3.0]oct-5-ene, 6. An unidentifiable material with a
higher Rf is also present in smaller amounts, but it could
not be isolated. However, its proton resonances in the
reaction mixture occur at δ 2.6, 2.55, 1.1, 0.6, and 0.3.
Thus, it is not a ring-opened monomeric peroxide like
that observed as a minor product by Adam in the
thermolysis of his methylcyclopropyl azo compound in the
presence of oxygen. Table 1 lists the results of experi-
(5) Dowd, P.; Chow, M. J . Am. Chem. Soc. 1977, 99, 6438. It is of
some concern here that both the activation parameters for the loss of
ESR signal intensity are low, suggesting that intersystem crossing
occurs, but the activation energy is much too low relative to that
expected on the basis of high level calculations (14 kcal/mol: see ref
7). It may be that the loss of ESR intensity does not give the major
product which itself can result from closure of the initially formed
singlet trimethylenemethane species.

Reactions of Trimethylenemethane
J . Org. Chem., Vol. 61, No. 4, 1996 1401
Ta ble 1. P r od u ct Distr ibu tion fr om Ber son ’s Azo Com p ou n d TMM P r ecu r sor w ith Meth ylcyclop r op yl a n d
Biscyclop r op yl Su bstitu tion
methylcyclopropylTMMd
biscyclopropyl TMMe
conditions
fraction monomera
fraction dimerb
fraction monomera
fraction dimerb
unrearr. peroxide
80-100 °C
hν, 0 °C
2× dilution
hν, 20 °C, O2
hν, 0 °C, O2
hν, 0 °C, Ph2CO
hν, -78 °C
0.59
0.09
0.41
0.86
1.0.
0.5
0.74
0.0
0.5
0.26
0.15
0.0
0.85
major prod.
0.0
ca. 0.55
0.0
0.0
ca. 0.45
1.0
0.02
0.98
0.85
0.02 + housanec
a
b
Cyclopropane ring-opened monomer. Intact cyclopropane rings in dimers. c The housane goes exclusively to dimers upon warming.
From Adam’s compound, ref 3. e Compound 3 this work.
d
Sch em e 5
F igu r e 1. Curie plot from the ESR spectrum of the photolysis
of 3 in a propylene glycol matrix.
and in 1- and 2-propanol mixtures, the triplet spectrum
collapsed to a more narrow line with hyperfine splitting
ments reported here that can be compared to those
reported by Adam on the monocyclopropyl azo com-
pound.6
at the position expected for a doublet, and this spectrum
disappeared upon warming above 128 K. In a propylene
glycol matrix, the triplet spectrum did not collapse until
higher temperatures. The ESR intensity increased at
lower temperatures, and indeed, a linear Curie plot was
obtained from 167 to 96 K (see Figure 1). This behavior
is characteristic of a ground state triplet or of a singlet-
triplet equilibrium where it can be calculated that the
singlet is no lower than ca. 0.3 kcal/mol below the
triplet.4c Upon warming to room temperature, only ring-
unopened dimers were found by NMR spectrometry.
ESR Stu d ies of th e Biscyclop r op yl TMM Sp ecies
Der ived fr om 3. Concern about the multiplicity of the
ground state of the biscyclopropyl TMM species being
generated prompted examination of the ESR spectrum
of the photolysate of 3 at low temperatures in viscous
matrices. At 77 K in methyltetrahydrofuran, an ESR
spectrum was obtained that was characteristic of a
randomly oriented, rotationally frozen triplet species
looking identical to that first observed by Berson with
the dimethyl derivative.4a For the triplet derived from
3, |D|/hc equals 0.0232 cm-1 and |E|/hc equals 0.00234
Discu ssion
cm-1
.
In addition, a half-field absorption was also
observed. At higher temperatures in both methyl-THF
The results shown in Table 1 clearly indicate a
concentration dependence in the product distribution
when reactions are conducted at the same temperature
using the same energy source. However, comparison of
the relative amounts of ring-opened product and dimer
under one set of reaction conditions with another is
difficult. The dimerization rate of the biradicals depends
on the concentration of the biradicals, and this changes
not only with temperature and the energy source but
necessarily during the course of the reaction. So pho-
tolysis at 0 °C can generate a higher concentration of
biradicals than pyrolysis at 80 °C from precursor 3 at
any given time during a reaction. Nonetheless, the
results described in Table 1 make it clear that 6,6-
dicyclopropyl substitution on the TMM species has a
profound effect on the energy surface of the TMM system
since the product distribution (cyclopropane rearranged
(6) Attempts were made to trap the TMM species from 3 with
dimethyl fumarate under the photolysis conditions (see ref 4b), but
there was rapid conversion of fumarate to maleate in photolyses
conducted in acetonitrile solvent even upon irradiation through Pyrex
in a reaction reminiscent of that reported by Albone (Albone, E. S. J .
Am. Chem. Soc. 1968, 90, 4663). In contrast, pyrolysis of 3 in the
presence of dimethyl fumarate (two experiments with a ca. 20-fold
difference in concentration) gave both ring-opened monomer and at
least two fumarate adducts (HRMS for C18H24O4: calcd 304.1674; found
304.1675). From the NMR in the vinyl proton region, there appear to
be two fused adducts in a 1:1 ratio with the ratio of monomer to adducts
being approximately 11-fold higher in the more dilute case. However,
the GC reveals the presence of two peaks with the appropriate masses
for adducts in a 1:1.75 ratio consistent with the presence of a bridged
adduct coeluting with one of the fused adducts. Despite the incomplete
characterization of these adducts, it is important to note that their
relative proportions did not change with concentration of starting
material and fumarate, unlike that reported in ref 4b.
(7) Feller, D.; Tanaka, K.; Davidson, E. R.; Borden, W. T. J . Am.
Chem. Soc. 1982, 104, 967. Davidson, E. R., private communication.

1402 J . Org. Chem., Vol. 61, No. 4, 1996
Paul and Gajewski
Sch em e 6
Sch em e 7
to the systems studied by Berson and by Adam. It is also
possible that the one-bond-broken diazene biradical
might be involved, and it could undergo the cyclopropyl-
carbinyl radical cleavage before loss of nitrogen and give
rearranged monomer after nitrogen loss. It might also
be responsible for formation of a triplet TMM species and
at the same time react with the TMM triplet to give
dimer precursors. However, this behavior must have
resulted from the bis-geminal cyclopropyl substitution
again in contrast to the Berson and Adam systems for
reasons that are not obvious. Finally, it is possible that
the Adam supposition of a higher activation energy for
ring opening of the triplet TMM species than for the
singlet TMM species is not correct and that ring opening
of triplet TMM with an activation energy of roughly 7
kcal/mol cannot compete with dimerization except in the
case of substitution with larger groups. This is formally
equivalent to Scheme 7 with the substitution of rapid
interconversion of singlet and triplet ring-opened biradi-
cals for rapid intersystem crossing in the TMM biradicals.
monomer to dimers) varies with the initial concentration
of azo compound precursor.
A number of hypotheses can be entertained to account
for the difference resulting from biscyclopropyl substi-
tution: (A) steric retardation of the dimerization which
would either allow reversible intersystem crossing to
form singlet which ring opens (variation A1, Scheme 6)
or allow ring opening out of triplet (variation A2, Scheme
7)8 or (B) reduction in the size of the singlet-triplet gap
by destabilizing the planar triplet9 which would permit
reversible intersystem crossing relative to dimerization,
thus forming a singlet which undergoes ring opening.
It might also be argued that biscyclopropylcarbinyl
radicals might undergo increased rates of ring-opening
relative to monocyclopropylcarbinyl radicals, and this
would rationalize the increased amount of ring-opened
monomeric product with 3 relative to the methylcyclo-
propyl case studied by Adam;3 however, the concentration
dependence could not be rationalized by this. Further,
studies by Beckwith reveal little difference in the rate of
ring opening of cyclopropylcarbinyl radical in the mono-
or dicyclopropyl-substituted parent systems.10,11
If the last possibility were correct, then none of the
results reported here bear on the question of intersystem
crossing in the TMM biradical; they merely refer to the
perturbation introduced by the bis-geminal cyclopropyl
substitution. That this possibility is not viable can be
deduced from consideration of the trans isomer of the
ring-opened biradical. Given Adam’s scheme (Scheme 2)
the ring opening of the singlet TMM species can occur to
give both a cis and a trans allylcarbinyl radical, but it is
only the cis one which can ring-close to the monomeric
product. Therefore the trans one must return to its
singlet TMM precursor (Eact ) 11-14 kcal/mol, see
Scheme 2) to be recycled back to the cis allylcarbinyl ring-
opened biradical. However, intersystem crossing of the
trans ring-opened singlet to triplet should be fast, and it
can close to the triplet TMM species. Since high-
temperature reactions result in mostly or exclusively
ring-opened monomer in the mono- and bis cyclopropyl
cases, respectively, with little or no dimer being formed,
the transition state energy for closure of the trans triplet
biradical (to give triplet TMM and ultimately dimer)
must be higher in energy than the transition state for
ring-opening of the singlet TMM species to the cis
allylcarbinyl species (Scheme 8). This not only places the
transition state for the ring-opening of the triplet TMM
above the transition state for ring-opening of the singlet
but also higher than that for intersystem crossing at
lower temperatures.12 It is therefore true that the only
possible scheme consistent with all the facts at lower
temperatures is that described in Scheme 6 which
involves reversible intersystem crossing. Apparently, it
is the reduction in the rate of dimerization of the bis-
geminal cyclopropyl substituted species which allows this.
Other mechanistic alternatives include formation of the
ring-closed biscyclopropyl-substituted 5-isopropylidene-
bicyclo[2.1.0]pentane which for no obvious reason under-
goes both unimolecular cyclopropane ring-opening and
rearrangement and bimolecular dimerization in contrast
(8) A variation on this hypothesis is reaction exclusively out of the
singlet state. This would appear to require either that the singlet be
much more stable than the triplet or that all reactions are faster with
the singlet than the triplet. The ESR argues against a singlet ground
state of substantially greater stability than the triplet, and diradical
trapping by oxygen and by dimerization appear to be diffusion-
controlled processes with triplet biradicals but not with singlet
biradicals or biradicaloids (Reynolds, J . H.; Berson, J . A.; Kumashiro,
K. K.; Duchamp, J . C.; Zilm, K. W.; Scaiano, J . C.; Berinstain, A. B.;
Rubello, A.; Vogel, P. J . Am. Chem. Soc. 1993, 115, 8073) so that
exclusive singlet reactivity is hard to justify.
(9) Salinaro, R. F.; Berson, J . A. Tetrahedron Lett. 1982, 1447, 1451.
(10) Beckwith, A. L. J .; Bowry, V. W. J . Am. Chem. Soc. 1994, 116,
2710. Ring opening of the biscyclopropylcarbinyl radical has log k (s-1
)
) 13.9-8600/RT ln 10, and ring opening of methylcyclopropylcarbinyl
radical has log k (s-1) ) 13.1-7700/RT ln 10. At 0 °C both radicals
open with the same rate. Over the temperature range studied here,
the largest rate difference is a factor of 2. See also: Lemieux, R. P.;
Beak, P. J . Org. Chem. 1990, 55, 5454, where a dicyclopropylcarbinyl
radical ring opens a factor of 4 slower than a monocyclopropyl radical
where both are spiro to a 3-indanyl radical.
(11) It should be noted that Adam suggests that ring-opening must
have an activation energy of only 5 kcal/mol to make it competitive
with intersystem crossing in the methylcyclopropyl-substituted case.

Reactions of Trimethylenemethane
J . Org. Chem., Vol. 61, No. 4, 1996 1403
the mixture was treated with 110 mL of water. The organic
layer was separated, and the aqueous layer was extracted
(CH2Cl2, 50 mL × 3). The combined organic layer was washed
with water until washings were neutral and dried over
anhydrous sodium sulfate. The methylene chloride was
removed under aspirator vacuum. Fractional distillation at
∼15-8 Torr gave a yellow orange fraction (3.8 g, bp 100-120
°C) having 90% of the desired fulvene contaminated with
starting ketone (∼8.5%). A pure sample of the dicyclopropyl-
fulvene was obtained upon silica gel flash chromatography (5%
ether in pentane). 1H NMR: δ 6.69 (m, 2 H), 6.46 (m, 2 H),
1.83 (m, 2 H), 1.05-0.84 (m, 8 H).
Sch em e 8
2,3-Dica r b et h oxy-7-(2,2-d icyclop r op ylm et h ylid en e)-
2,3-d ia za bicyclo[2.2.1]h ep ta n e (4). Following the general
procedures for the synthesis of cyclopropyl methyl-substituted
diazene by Adam and Finzel,3 4 was synthesized from dicy-
clopropylfulvene (0.57 g, 3.61 mmol), diethyl azodicarboxylate
(0.636 g, 3.65 mmol), and tert-butyl methyl ether (10 mL) at
20 °C followed by reduction of the endocyclic double bond by
catalytic hydrogenation over palladium-carbon (5%, 40 mg).
Chromatography (70% EtOAc, 30% pentane) over silica gel
gave a viscous liquid (∼1.0 g) as a moderately pure sample.
1H NMR: δ 5.05 (bd, 2 H), 4.20 (bs, 4 H), 1.9 (bs, 2 H), 1.26
(m, 10 H), 0.70-0.40 (m, 8 H). MS: 334 (M+, 100), 261 (22),
J ust why the triplet energy surface is as described above
is not clear, nor is it obvious that the dimerization would
be so dramatically inhibited by the presence of a second
cyclopropyl group unless there are specific geometric
requirements for the reactions on the triplet surface.
Lastly, there should be some concern about possible
steric effects on the singlet-triplet gap in the bis cyclo-
propyl-substituted TMM species derived from 3. Modi-
fied INDO calculations by Lahti have revealed a 2-3
kcal/mol contraction in the gap presumably due to steric
destabilization of the planar triplet.13 This, however,
does not dramatically alter any of the considerations
above.
234 (33), 189 (25), 159 (70), 117 (46), 91(28). HRMS for C18H26
N2O4: calcd 334.1892, found 334.1902.
-
7-(2,2-Dicyclopropylmethylidene)-2,3-diazabicyclo[2.2.1]-
h ep t-2-en e (3). Compound 4 (∼1.0 g) was hydrolyzed with
KOH (1.55 g) in 2-propanol (16 mL) followed by oxidation via
a copper complex as is described for monocyclopropyl-substi-
tuted diazene3 to give the desired diazene 3. This was purified
by flash chromatography (ether/pentane, 1:1) over silica gel.
1H NMR: δ 5.58 (t, J ) 1.6 Hz, 2 H, bridgehead H), 1.63 (m,
2 H, Hexo), 1.20 (tt, J ) 8.4 and 5.6 Hz, 2 H, cyclopropyl H),
1.08 (m, simulated J ) 10, 5.2 and 3.8 Hz, 2 H, Hendo), 0.62
(m, 4 H), 0.49 (m, 4 H). 13C NMR: 128.37, 73.96, 21.23, 12.62,
4.81, 4.79. UV (PhH) λmax 340 nm (log ꢀ ) 2.13); ꢀ = 4 at 360-
365 nm. CIHRMS for C12H16N2: calcd for M+ + 1 189.1391,
found 189.1396.
Con clu sion
Th er m olysis of Dia zen e 3. A small amount of diazene
(8-10 mg) taken in a glass tube containing benzene (1.5 mL)
was degassed (freeze-pump-thaw cycle (3 times)) and sealed
under vacuum. After the sample was heated in a boiling water
bath for 20 min, the contents of the tube were analyzed by
GC and GC-MS, showing basically one compound. Removal
of the solvent by rotary evaporation left a colorless liquid which
was spectroscopically characterized as 2-cyclopropylbicyclo-
[4.3.0]nona-1(9),2-diene (5). 1H NMR: δ 5.78 (bm, 1 H), 5.44
(bm, 1 H), 2.66-2.54 (m, 1 H), 2.4-2.30 (m, 2 H), 2.16 (m, 3
H), 2.00-1.90 (m, 1 H), 1.51-1.30 (m,2H), 1.28-1.12 (m, 1 H,
cyclopropyl H), 0.70-0.54 (m, 2 H), 0.49-0.40 (m, 1 H), 0.39-
0.30 (m, 1 H). 13C NMR: 144.67, 135.37, 123.10, 121.23, 44.03,
32.13, 31.55, 30.32, 26.42, 12.69, 5.84, 4.14. GC-MS: 160 (M+,
52), 145 (11), 131 (52), 117 (100), 104 (22), 91 (91), 77 (23), 65
An important observation is the fact that bis-geminal
cyclopropyl substitution on the trimethylenemethane
biradical reduces the extent of dimerization and increases
the extent of cyclopropane ring-cleavage relative to
monocyclopropyl substitution. The observations are best
rationalized by steric retardation of dimerization by
triplet and perhaps reduction in the singlet-triplet gap.
The retardation of dimerization of moderately stable
triplet states then allows reversion to the singlet TMM
which undergoes ring opening.
Exp er im en ta l Section
(20), 41 (28), 39 (33), 28 (100). HRMS for C12H16
160.1252, found 160.1252.
: calcd
Proton (1H) and carbon (13C) NMR spectra were recorded
on a Varian VXR-400 MHz instrument. All chemical shifts
are reported in parts per million (ppm) downfield from
tetramethylsilane and were taken in chloroform-d solution.
High-resolution mass spectra (HRMS) were recorded on a
KRATOS MS-80/RFAQ spectrometer.
6,6-Dicyclop r op ylfu lven e. 6,6-Dicyclopropylfulvene was
synthesized from sodium metal (1.8 g, 0.08 mol), absolute
ethanol (40 mL), dicyclopropyl ketone (0.08 mol), and freshly
distilled cyclopentadiene (6.7 mL) according to literature
procedures.14 After stirring at room temperature (overnight),
An NMR sample of diazene in benzene-d6 was degassed and
sealed under vacuum. After the tube was heated in a water
1
bath (100 °C) for 20 min, an H NMR spectrum was recorded.
Only hydrocarbon 5 was formed; dimers (see below) were not
detected. An NMR sample of diazene 3 in deuteriochloroform
was allowed to stand at room temperature for a week, and
the NMR spectrum revealed the presence of only hydrocarbon
5.
P h otolysis of Dia zen e 3 a t -78 °C. An NMR tube
containing diazene (4 mg) and CFCl3 (∼0.8 mL) (degassed and
sealed) was photolyzed at -78 °C (dry ice/acetone) for 2.5 h
with a Hanovia high-pressure mercury arc lamp using a Pyrex
(12) Relative energies are difficult to compare here since intersystem
crossing in TMM species is slow and results from a temperature-
independent small transmission factor in transition state theory while
cyclopropylcarbinyl ring-opening reactions have 7-8 kcal/mol enthalpic
barriers so that the relative utilization of pathways is a strong function
of temperature. Nonetheless, consideration of the fate of the trans
allylcarbinyl species results in recognition that the transition state
for ring opening of triplet TMM is higher than than for ring opening
of singlet TMM.
1
filter. An H NMR spectrum was recorded in a precooled NMR
probe (-80 °C) followed by a gradual increase in temperature
up to -10 °C. The spectrum reveals few if any monomer
resonances and consists mostly of dimers. MS of dimers: 320
(M+, 3), 160 (47), 131 (17), 117 (28), 91 (24), 84 (100). HRMS
(13) We thank Professor Paul Lahti for these calculations.
(14) Kerber, R. C.; Linde, H. G., J r. J . Org. Chem. 1966, 31, 4321.

1404 J . Org. Chem., Vol. 61, No. 4, 1996
Paul and Gajewski
for C24H32: calcd 320.2504, found 320.2492. 1H NMR (rel areas
given): δ 5.70 (bm, 1.1); 5.64 (bm, 1.67); 5.56 (q, J ) 4 Hz,
0.84); 3.0 (three multiplets, 6.41); 2.8 (bm, 2.18); 2.4-1.95 (m,
10.0); 1.95-1.65 (m, 5.70); 1.65-1.1 (m, 33.35); 1.0 (m) and
0.85 (m) ca. 1:1 (6.43 total rel area); 0.8 to -0.15 (m, 57.91).
Dir ect P h otolysis of Dia zen e 3. Diazene 3 (8.6 mg) taken
in an NMR tube containing benzene-d6 (∼0.6 mL) was de-
gassed and sealed by a torch. The NMR sample was photo-
lyzed at 0-5 °C (ice/H2O) for an hour by a Hanovia high-
pressure mercury arc lamp using a Pyrex filter. In addition
to monomer 5, fused C24H32 dimers could be detected in the
NMR judging by the presence of vinyl protons δ 5.70 (bm), δ
5.64 (bm), and δ 5.56 (quartet) in the ratio of 4:6:3. The ratio
of monomer to dimers is ∼1:0.8. Further photolysis did not
change the product compositions.
Effect of Con cen tr a tion in th e Dir ect P h otolysis of 3.
Two sealed tubes containing 5.0 and 5.7 mg of 3 and 0.7 mL
and 1.5 mL of benzene, respectively, were photolyzed as
described above. After photolysis the solvents were removed
under aspirator vacuum, and the residues were analyzed by
NMR. The ratios of monomer to dimer are 1:1 and 76:24,
respectively, and are accurate to plus and minus 5%.
Sen sitized P h otolysis of Dia zen e 3 w ith a n Ar gon Ion
La ser . The contents of an NMR tube, 5 mg of diazene 3, 2.5
mg of benzophenone, and 0.6 mL of benzene-d6 (degassed and
sealed), were photolyzed at 0-5 °C with the 364 nm line of an
argon ion laser (CW) with the reaction being followed by NMR.
After ∼3.5 h of photolysis almost all of the diazene decom-
posed, forming 5 and dimers mixture as was observed above
(monomer : dimers ) ∼1:0.8).
Electr on Sp in Reson a n ce Exp er im en ts. Diazene 3 (∼20
mg) was placed in a 4 mm o.d. quartz ESR tube to which was
added 100-130 µL of the appropriate solvent (2-methyltet-
rahydrofuran, 1- and 2-propanol mixture (2:3), propylene
glycol). All solvents were distilled prior to use and were stored
over molecular sieves. After degassing (three freeze-pump-
thaw cycles) the tube was sealed. ESR spectra were obtained
in a Bruker ESP-300 multiscan X-band spectrometer at 9.45
GHz field modulation frequency, 8.0 G modulation amplitude
and 63-20 µW microwave power (for half-field absorptions
higher power was used). ESR measurement at 77 K were
carried out in frozen 2-methyl-THF solution with liquid
nitrogen in a Suprasil finger dewar. The samples were
photolyzed with a Photo Technology International 200 W Hg-
Xe arc lamp in a housing with a parabolic reflector and a Pyrex
water filter. The light was focused into the ESR cavity using
a fiber optic light pipe. Typical photolysis times were 10-20
min. Variable temperature ESR measurements for photolysis
were carried out in the ESP-300 spectrometer cavity equipped
with a Bruker ER 4111 VT liquid nitrogen cryostat. Temper-
atures at the sample were calibrated with a carbon-glass
thermocouple. ESR peak intensities were obtained as a
function of temperature under identical scanning and micro-
wave power conditions and then doubly integrated using
Bruker ESP-300 software with the same integration endpoints
for all specta of the sample obtained at various temperatures.
P h otolysis of 3 in th e P r esen ce of Oxygen . A Pyrex
tube containing diazene (5 mg) and cyclohexane (3.0 mL) at
∼0 °C was saturated with oxygen by bubbling oxygen gas
through the solution by a 18 gauge needle. The solution was
then photolyzed at 0-5 °C (ice/water) for 1.75 h. Removal of
the solvent by rotary evaporation gave a residue which was
identified by 1H NMR as the 4,4-dicyclopropyl-2,3-dioxabicyclo-
[3.3.0]oct-5-ene, 6, (Rf ) 0.5) along with a minor product (Rf
) 0.8, see Results). Running the reaction on a larger scale
followed by silica gel chromatography (CH2Cl2, eluent) did not
1
afford any identifiable product other than the peroxide, 6. H
NMR: δ 5.52 (td (5 lines), J ) 3.6 and 1.6 Hz, 1H, 6-H), 5.30-
5.25 (m (14 lines), 1H, 1-H), 2.94-2.82 (m, 1H, 7-H), 2.80-
2.70 (m, 1H, 7-H), 2.17-2.10 (m, 1H, 8-H), 1.84-1.73 (m, 1H,
8-H), 1.19 (tt, J ) 8.0 and 5.6 Hz, 1H, cyclopropyl H), 0.99 (tt,
J ) 8.0 and 5.6 Hz, 1H, cyclopropyl H), 0.70-0.22 (m, 8H,
cyclopropyl H). HRMS for C12H16O2: calcd 192.1150, found
192.1145
Ack n ow led gm en t. We thank the Department of
Energy for support of this work. We also thank Professor
G. M. Hieftje for use of his argon ion laser source.
J O9513727