Access to the naphthylcarbene rearrangement manifold via isomeric benzodiazocycloheptatrienes

Article abstract of DOI:10.1021/jo020304z

Irradiation (λ = 670 or >613 nm) of 4,5-benzodiazocycloheptatriene (15), matrix isolated in argon at 10 K, produces primarily 2,3-benzobicyclo[4.1.0]hepta-2,4,6-triene (9) accompanied by small amounts of triplet 4,5-benzocycloheptatrienylidene (2) and 2-naphthylcarbene (10). A reversible photoequilibrium is established in which 9 is converted to 10 at λ = 290 nm and then regenerated at λ = 360 nm. Similarly, matrix-isolated 2,3-benzodiazocycloheptatriene (16) produces 4,5-benzobicyclo[4.1.0]hepta-2,4,6-triene (11) at λ = 670 or >613 nm, but without detection of 2,3-benzocycloheptatrienylidene (4). Irradiation of 11 at λ = 290 nm induces ring opening to triplet 1-naphthylcarbene (12), which, in turn, cyclizes back to 11 at λ = 342 or >497 nm. The diazo compounds and photoproducts are characterized by IR, UV/visible, and ESR spectroscopy, where appropriate, and by comparison of the experimental and B3LYP/6-31G* calculated IR spectra for each species. Alternate rearrangement products such as allenes 6, 7, and 8 are not detected in the photolysis of either diazo compound.

Full text of DOI:10.1021/jo020304z

Access to th e Na p h th ylca r ben e Rea r r a n gem en t Ma n ifold via
Isom er ic Ben zod ia zocycloh ep ta tr ien es
Paul A. Bonvallet, Eric M. Todd, Yong Seol Kim, and Robert J . McMahon*
Department of Chemistry, University of Wisconsin, 1101 University Avenue,
Madison, Wisconsin 53706-1396
mcmahon@chem.wisc.edu
Received April 30, 2002
Irradiation (λ ) 670 or >613 nm) of 4,5-benzodiazocycloheptatriene (15), matrix isolated in argon
at 10 K, produces primarily 2,3-benzobicyclo[4.1.0]hepta-2,4,6-triene (9) accompanied by small
amounts of triplet 4,5-benzocycloheptatrienylidene (2) and 2-naphthylcarbene (10). A reversible
photoequilibrium is established in which 9 is converted to 10 at λ ) 290 nm and then regenerated
at λ ) 360 nm. Similarly, matrix-isolated 2,3-benzodiazocycloheptatriene (16) produces 4,5-
benzobicyclo[4.1.0]hepta-2,4,6-triene (11) at λ ) 670 or >613 nm, but without detection of 2,3-
benzocycloheptatrienylidene (4). Irradiation of 11 at λ ) 290 nm induces ring opening to triplet
1-naphthylcarbene (12), which, in turn, cyclizes back to 11 at λ ) 342 or >497 nm. The diazo
compounds and photoproducts are characterized by IR, UV/visible, and ESR spectroscopy, where
appropriate, and by comparison of the experimental and B3LYP/6-31G* calculated IR spectra for
each species. Alternate rearrangement products such as allenes 6, 7, and 8 are not detected in the
photolysis of either diazo compound.
SCHEME 1
In tr od u ction
Investigations of the structure and reactivity of organic
carbenes bear a long and vibrant history.1 Arylcarbenes
have attracted attention due to their complex thermal
and photochemical transformations, as illustrated in
Scheme 1 for the case of naphthylcarbenes.2-16 The
(1) For reviews and leading references on carbenes, see: (a) To-
mioka, H. Bull. Chem. Soc. J pn. 1998, 71, 1501-1524. (b) Platz, M.
S.; Modarelli, D. A.; Morgan, S.; White, W. R.; Mullins, M.; C¸ elebi, S.;
Toscano, J . P. Prog. React. Kinet. 1994, 19, 93-137. (c) Sander, W.;
Bucher, G.; Wierlacher, S. Chem. Rev. 1993, 93, 1583-1671. (d) Platz,
M. S. Kinetics and Spectroscopy of Carbenes and Biradicals; Plenum:
New York, 1990. (e) Kirmse, W. Carbene Chemistry, 2nd ed.; Academic
Press: New York, 1971. (f) Zeller, K.-P.; Gugel, H. In Methoden der
Organischen Chemie (Houben-Weyl); Regitz, M., Ed.; G. Thieme:
Stuttgart, 1989; Vol. E19b, pp 165-336.
(2) J ones, W. M. Angew. Chem., Int. Ed. Engl 1972, 11, 325-326.
(3) Mitsuhashi, T.; J ones, W. M. J . Am. Chem. Soc. 1972, 94, 677-
679.
(4) Krajca, K. E.; Mitsuhashi, T.; J ones, W. M. J . Am. Chem. Soc.
1972, 94, 3661-3662.
general features of arylcarbene rearrangements have
long been established on the basis of inter- and intra-
molecular trapping studies, but a detailed understanding
of the various reactive intermediates involved in these
rearrangement mechanisms remains a challenge to both
experimentalists and theorists. In a broader context, the
chemistry of arylcarbenes, which are formed upon reac-
tion of atomic carbon with aromatic compounds,17 is
relevant to an understanding of a variety of carbon-rich
environments, including carbon arcs, discharges, inter-
(5) J ones, W. M.; J oines, R. C.; Myers, J . A.; Mitsuhashi, T.; Krajca,
K. E.; Waali, E. E.; Davis, T. L.; Turner, A. B. J . Am. Chem. Soc. 1973,
95, 826-835.
(6) Tyner, R. L.; J ones, W. M.; Ohrn, Y.; Sabin, J . R. J . Am. Chem.
Soc. 1974, 96, 3765-3769.
(7) Coburn, T. T.; J ones, W. M. J . Am. Chem. Soc. 1974, 96, 5218-
5227.
(8) J ones, W. M. Acc. Chem. Res. 1977, 10, 353-359.
(9) Waali, E. E.; Lewis, J . M.; Lee, D. E.; Allen, E. W. I.; Chappell,
A. K. J . Org. Chem. 1977, 42, 3460-3462.
(10) Becker, J .; Wentrup, C. J . Chem. Soc., Chem. Commun. 1980,
190-191.
(11) West, P. R.; Mooring, A. M.; McMahon, R. J .; Chapman, O. L.
J . Org. Chem. 1986, 51, 1316-1320.
(12) Chateauneuf, J . E.; Horn, K. A.; Savino, T. G. J . Am. Chem.
Soc. 1988, 110, 539-545.
(15) Bonvallet, P. A.; McMahon, R. J . J . Am. Chem. Soc. 1999, 121,
10496-10503.
(13) Albrecht, S. W.; McMahon, R. J . J . Am. Chem. Soc. 1993, 115,
855-859.
(16) Bonvallet, P. A.; McMahon, R. J . J . Am. Chem. Soc. 2000, 122,
9332-9333.
(14) Xie, Y.; Schreiner, P. R.; Schleyer, P. v. R.; Schaefer, H. F. J .
Am. Chem. Soc. 1997, 119, 1370-1377.
(17) Zheng, F.; McKee, M. L.; Shevlin, P. B. J . Am. Chem. Soc. 1999,
121, 11237-11238.
10.1021/jo020304z CCC: $22.00 © 2002 American Chemical Society
Published on Web 11/13/2002
J . Org. Chem. 2002, 67, 9031-9042
9031

Bonvallet et al.
SCHEME 2a
SCHEME 3
triplet.25-27 A precise description of the planar carbene
is complex, in the sense that several electronic configura-
tions lie close in energy (A2 and B2 for the triplet, A2 and
B2 for the open shell singlet, and A1 for the closed shell
singlet).24 In the benzo series, the position of benzanne-
lation further modulates the relative stabilities of the
various isomers.14,28 At the B3LYP/6-311+G* level of
theory, the planar singlet benzocycloheptatrienylidene
structures 2 and 4 are found to be a second-order saddle
point and a transition state, respectively.14 Imaginary
frequencies notwithstanding, singlets 2 and 4 (1A1 state)
are calculated to be higher in energy than the corre-
sponding planar triplet benzocycloheptatrienylidene and
twisted singlet allene by approximately 5 and 10 kcal/
mol, respectively.
Chemical trapping studies of the thermal or photo-
chemical decomposition of the tosylhydrazone sodium salt
of 4,5-benzotropone (19) suggest a sequence of rearrange-
ments involving 4,5-benzocycloheptatrienylidene (2), 4,5-
benzobicyclo[4.1.0]hepta-2,4,6-triene (9), and 2-naphth-
ylcarbene (10) (Scheme 3).8 Thermal decomposition of the
tosylhydrazone salt of the 2,3-isomer (23) affords a
trapping product of 1-naphthylcarbene (12)29 (Scheme 4).
Kirmse and Sluma interpret their trapping studies in
terms of distinct carbenic vs allenic reactivity, supporting
the concept of benzocycloheptatrienylidene and benzo-
cycloheptatetraene as equilibrating intermediates.30,31
Isolation of a formal [2+2] dimer of a simple derivative
of 4,5-benzocyclohepta-1,2,4,6-tetraene supports the in-
termediacy of the strained allene during the dehydro-
halogenation of benzo-chlorocycloheptatrienes.9
a
B3LYP energies (kcal/mol) given in italics (singlet state +
b
ZPVE; reference 4) relative to cyclobuta[de]naphthalene. Two
imaginary frequencies. c One imaginary frequency.
stellar clouds, and circumstellar envelopes. Recent specu-
lation concerning the build-up of polycyclic aromatic
hydrocarbons in interstellar clouds has been discussed
in the context of carbon atom chemistry and the inter-
mediates involved in arylcarbene interconversions.18
In earlier studies of naphthylcarbene rearrangements,
we investigated the chemistry and spectroscopy of naph-
thylcarbene, benzobicyclo[4.1.0]hepta-2,4,6-triene, and
4,5-benzocyclohepta-1,2,4,6-tetraene isomers.11,13,15,16 In
the current study, we describe experimental and com-
putational studies of benzocycloheptatrienylidene and
benzobicyclo[4.1.0]hepta-2,4-dien-7-ylidene isomers.
Ba ck gr ou n d
The relationship between the benzocycloheptatrien-
ylidenes 2-4 and the benzocyclohepta-1,2,4,6-tetraenes
6-8 is perhaps even more subtle and complex than the
relationship between the parent species, cycloheptatrie-
nylidene (1) and cycloheptatetraene (5) (Scheme 2). In
the parent case, early qualitative arguments described
cycloheptatrienylidene (1) as a planar aromatic singlet
carbene. Experiment19,20 and theory,21-24 however, estab-
lished that a nonplanar twisted allene structure, corre-
sponding to cycloheptatetraene (5), lies ca. 25 kcal/mol
lower in energy. The planar singlet cycloheptatrien-
ylidene structure represents the transition state for
enantiomerization of the C2-symmetric allene, and the
ground state of cycloheptatrienylidene (1) is, in fact, a
Of the three planar and three nonplanar C11H8 struc-
tures shown in Scheme 2, only two have been observed
experimentally. Chateauneuf, Horn, and Savino reported
the transient electronic absorption spectrum of triplet
4,5-benzocycloheptatrienylidene (2) in fluid solution, as
(18) Kaiser, R. I.; Asvany, O.; Lee, Y. T. Planet. Space Sci. 2000,
48, 483-492.
(19) (a) West, P. R.; Chapman, O. L.; LeRoux, J .-P. J . Am. Chem.
Soc. 1982, 104, 1779-1782. (b) McMahon, R. J .; Abelt, C. J .; Chapman,
O. L.; J ohnson, J . W.; Kreil, C. L.; LeRoux, J .-P.; Mooring, A. M.; West,
P. R. J . Am. Chem. Soc. 1987, 109, 2456-2459. (c) Matzinger, S.; Bally,
T. J . Phys. Chem. A 2000, 104, 3544-3552.
(20) Matzinger, S.; Bally, T. J . Phys. Chem. A 2000, 104, 3544-
3552.
(21) Radom, L.; Schaefer, H. F.; Vincent, M. A. Nouv. J . Chim. 1980,
4, 411-415.
(22) Matzinger, S.; Bally, T.; Patterson, E. V.; McMahon, R. J . J .
Am. Chem. Soc. 1996, 118, 1535-1542.
(23) Wong, M. W.; Wentrup, C. J . Org. Chem. 1996, 61, 7022-7029.
(24) Schreiner, P. R.; Karney, W. L.; Schleyer, P. v. R.; Borden, W.
T.; Hamilton, T. P.; Schaefer, H. F. J . Org. Chem. 1996, 61, 7030-
7039.
(25) Kuzaj, M.; Lu¨erssen, H.; Wentrup, C. Angew. Chem., Int. Ed.
Engl. 1986, 25, 480-482.
(26) McMahon, R. J .; Chapman, O. L. J . Am. Chem. Soc. 1986, 108,
1713-1714.
(27) Evans, R. A.; Wong, M. W.; Wentrup, C. J . Am. Chem. Soc.
1996, 118, 4009-4017, footnote 26.
(28) Balci, M.; Winchester, W. R.; J ones, W. M. J . Org. Chem. 1982,
47, 5180-5186.
(29) Reference 8, footnote 19.
(30) Kirmse, W.; Loosen, K.; Sluma, H.-D. J . Am. Chem. Soc. 1981,
103, 5935-5937.
(31) Kirmse, W.; Sluma, H.-D. J . Org. Chem. 1988, 53, 763-767.
9032 J . Org. Chem., Vol. 67, No. 25, 2002

Access to the Naphthylcarbene Rearrangement Manifold
SCHEME 4
SCHEME 6
SCHEME 7
SCHEME 5a
a
B3LYP/6-31G* energy (kcal/mol, ZPVE corrected) relative to
15 in italics.
well as the triplet ESR spectrum of 2 in a frozen glass.12
We recently reported the IR and UV/visible absorption
spectra of the strained allene, 4,5-benzocycloheptatet-
raene (7), under matrix isolation conditions.16
Resu lts a n d Discu ssion
P r ep a r a tion of Ca r ben e P r ecu r sor s. Diazirines
generally serve as convenient and robust carbene precur-
sors. A typical diazirine preparation involves oxidation
of the corresponding diaziridine, which, in turn, is
prepared by treatment of an activated imine derivative
with ammonia.32 Despite our considerable efforts, the
attempted syntheses of diazirines 13 and 14 (Scheme 5)
proved to be unfruitful, either because of the inability to
prepare the activated imine derivative (due to the
unusual electronic character of tropone derivatives33) or
because the activated imine derivative reacts readily via
the Beckmann rearrangement or other decomposition
pathway.34
With simple routes toward the diazirines thwarted, we
turned our attention to the preparation of diazo com-
pounds 15 and 16. One synthetic approach involves the
oxidation of the corresponding benzotropone hydrazones
20 or 24 (Schemes 6 and 7). An earlier report12 describes
the rapid oxidation (5 min) of 4,5-benzotropone hydrazone
at 0 °C with nickel peroxide.35 In our hands, the hydra-
zone appears unreactive under these conditions, even
upon extended reaction times. (The same batch of nickel
peroxide readily converts benzyl alcohol to benzoic acid
in a control experiment). In the presence of silver oxide
at 0 °C, however, both benzotropone hydrazones are
satisfactorily oxidized to afford the crude diazo compound
as a dark red-orange solid. Although purified samples of
diazo compounds 15 and 16 could not be isolated by
laboratory-scale sublimation, small quantities of diazo
compound could be volatilized from the crude mixture
and co-deposited with argon at 30 K. The matrix-isolated
sample of 4,5-benzodiazocycloheptatriene (15) generated
by this method is substantially more pure than that
obtained in the thermolysis of tosylhydrazone salt 19
(vide infra). The matrix-isolated sample of 2,3-benzodia-
zocycloheptatriene (16), on the other hand, is often
impure and is obtained in extremely small amounts.
An alternate route to diazo compounds 15 and 16
involves the preparation of tosylhydrazones 18 and 22
(Schemes 6 and 7).5,8 It has long been known that
thermolysis of the corresponding tosylhydrazone salts 19
and 23 does not afford isolable quantities of diazo
compound. Thermolysis of 19, with direct co-deposition
of the pyrolysate with argon at 30 K, however, affords a
small amount of 4,5-benzodiazocycloheptatriene (15)
accompanied by several unidentified impurities. Simi-
larly, impure diazo compound 16 can be obtained upon
pyrolysis of tosylhydrazone salt 23. Although the direct
oxidation of 4,5-benzotropone hydrazone (20) is the
preferred method of preparing 4,5-benzodiazocyclohep-
(32) (a) Chemistry of Diazirines; Liu, M. H. T., Ed.; CRC Press: Boca
Raton, FL, 1987. (b) Schmitz, E. Adv. Heterocycl. Chem. 1979, 24, 63-
107.
(33) For leading references, see: (a) Skancke, A.; Hosmane, R. S.;
Liebman, J . F. Acta Chem. Scand. 1998, 52, 967-974. (b) J ug, K. J .
Org. Chem. 1983, 48, 1344-1348. (c) Nozoe, T.; Mukai, T.; Takase, K.
Chem. Abstr. 1957, 51, 7316.
(34) Information available as Supporting Information.
(35) Nakagawa, K.; Konaka, R.; Nakata, T. J . Org. Chem. 1962, 27,
1597-1601.
J . Org. Chem, Vol. 67, No. 25, 2002 9033

Bonvallet et al.
tatriene (15), pyrolysis of tosylhydrazone salt 23 is a more
reliable method for the generation of 2,3-benzodiazo-
cycloheptatriene (16). 2,3-Benzodiazocycloheptatriene
(16), which has not been reported previously, appears to
be more susceptible to decomposition than the isomeric
diazo compound 15.
Computational studies of diazirines 13 and 14 and
diazo compounds 15 and 16 (Scheme 5) reveal that
diazirines 13 and 14 are predictably higher in energy
than the corresponding diazo compounds 15 and 16. The
energy difference between the two isomeric diazo com-
pounds (3.3 kcal/mol), as well as the isomeric diazirines
(1.3 kcal/mol), is relatively small. The experimental
difficulty in preparing and handling 2,3-benzodiazo-
cycloheptatriene (16) cannot be explained in terms of the
compound possessing a significantly higher energy than
the isomeric 4,5-benzodiazocycloheptatriene (15). The
calculated geometry of diazo compound 16, however, is
noticeably distorted from planarity. This structural effect
may play a role in the thermal instability of 16 in the
laboratory.
Although the matrix-isolated samples of diazo com-
pounds 15 and 16 inevitably contain (matrix-isolated)
impurities, the combination of photochemical and spec-
troscopic data permits meaningful conclusions to be
drawn from the experiments. IR difference spectra reli-
ably reveal the major photochemical processes that occur
in the matrix. UV/visible spectra, however, must be
interpreted with caution, as minor impurities may have
large extinction coefficients. ESR spectra of triplet car-
benes are readily interpreted because of their unique
characteristics.
Ma tr ix P h otoch em istr y of 4,5-Ben zod ia zocyclo-
h ep ta tr ien e (15). Irradiation (λ ) 670 or >613 nm) of
matrix-isolated 4,5-benzodiazocycloheptatriene (15), gen-
erated either from thermolysis of 19 or oxidation of 20,
causes complete disappearance of diazo compound ac-
companied by growth of 2,3-benzobicyclo[4.1.0]hepta-
2,4,6-triene (9) and trace amounts of triplet 2-naphthyl-
carbene (10) and 4,5-benzocycloheptatrienylidene (2)
(Figures 1-3, Scheme 8). The detailed justification for
the structural assignments is discussed below. Cyclopro-
pene 9 is detected by IR and UV/visible spectroscopy,
2-naphthylcarbene (10) is detected by UV/visible and
ESR spectroscopy, and 4,5-benzocycloheptatrienylidene
(2) is detected only by ESR spectroscopy. As a result of
its low concentration, 2 is not detected by other spectro-
scopic methods. Carbene 2 appears to exist in a low
steady-state concentration under these photochemical
conditions, as the intensity of the new ESR absorptions
remains constant throughout the photolysis. Even after
prolonged irradiation of the matrix (λ ) 670 nm, 37 h),
the ESR signal for 2 does not change, although 2-naph-
thylcarbene (10) gradually begins to appear (Figure 3).
In a result reminiscent of an earlier study,13 band-pass
irradiation (λ ) 290 nm) converts cyclopropene 9 to
carbene 10 with a concomitant loss of carbene 2. At this
point, 2-naphthylcarbene (10) is observable by IR spec-
troscopy. Reversion of 2-naphthylcarbene (10) to cyclo-
propene 9 is achieved by band-pass irradiation (λ ) 360
F IGURE 1. (a) Top: Calculated IR spectrum of 2,3-benzobi-
cyclo[4.1.0]hepta-2,4,6-triene (9). Middle: IR difference spec-
trum (Ar, 10 K) showing complete conversion of diazo com-
pound 15 to cyclopropene 9 upon irradiation at λ > 613 nm,
18 h. Bottom: Calculated IR spectrum of 4,5-benzodiazocy-
cloheptatriene (15). (b) Top: Calculated IR spectrum of triplet
s-E 2-naphthylcarbene (10). Middle: IR difference spectrum
showing disappearance of cyclopropene 9 and emergence of
carbene 10 at λ ) 290 nm, 4 h. Bottom: Calculated IR
spectrum of 9. (c) Top: Calculated IR spectrum of 9. Middle:
IR difference spectrum showing complete conversion of carbene
10 to cyclopropene 9 upon irradiation at λ ) 360 nm, 4 h.
Bottom: Calculated IR spectrum of 10. All B3LYP/6-31G*
calculated frequencies are scaled by 0.98.
nm), thus establishing a clean, reversible photoequilib-
rium between 9 and 10.
Cyclopropene 9 is characterized by IR and UV/visible
spectroscopy: IR (Ar, 10 K) 1758 w, 1747 w, 1474 w, 1118
w, 1009 w, 959 w, 885 m, 795 s, 781 s, 748 m, 715 m, 662
m, 604 w, 525 w cm-1; UV/visible (Ar, 10 K) λmax 217 nm.
Triplet 2-naphthylcarbene (10) is characterized by IR,
UV/visible, and ESR spectroscopy: IR (Ar, 10 K) 837 s,
809 m, 801 w, 741 s, 617 w cm-1; UV/visible (Ar, 10 K)
(36) West, P. R. Ph.D. Dissertation, University of California, Los
Angeles, 1981.
9034 J . Org. Chem., Vol. 67, No. 25, 2002

Access to the Naphthylcarbene Rearrangement Manifold
F IGURE 3. (a) ESR spectrum (Ar, 14 K) of 4,5-benzocyclo-
heptatrienylidene (2) upon irradiation of diazo compound 15
at λ ) 670 nm, 12 h. (b) Further growth of 2 and emergence of
2-naphthylcarbene (10) at λ ) 670 nm, 37 h. (c) Rapid
disappearance of 2 and growth of 10 upon irradiation at λ >
261 nm, 5 min. (d) Complete disappearance of 2 and continued
growth of 10 at λ > 261 nm, 5 h. In each spectrum, the signal
at ca. 3300 G is the g ) 2 (radical) impurity.
F IGURE 2. (a) UV/visible spectrum (Ar, 10 K) of 4,5-
benzodiazocycloheptatriene (15). (b) Mixture of 2,3-benzobicyclo-
[4.1.0]hepta-2,4,6-triene (9) and 2-naphthylcarbene (10) after
irradiation at λ > 613 nm, 17 h. (c) Growth of carbene 10 at λ
) 290 nm, 24 h. (d) Disappearance of 10 and emergence of 9
upon irradiation at λ ) 360 nm, 23 h.
J . Org. Chem, Vol. 67, No. 25, 2002 9035

Bonvallet et al.
TABLE 1. Zer o-F ield Sp littin g P a r a m eter s for Tr ip let
Ca r ben es
SCHEME 8
λmax 360, 341, 326, 279, 270, 240, 232, 218 nm; ESR (Ar,
14 K) s-E isomer Z1 ) 1692 G, X2 ) 4841 G, Y2 ) 5891
G, Z2 ) 8533 G, |D/hc| ) 0.4785 cm-1, |E/hc| ) 0.0258
cm-1; s-Z isomer Z1 ) 1909 G, X2 ) 4980 G, Y2 ) 5891
G, Z2 ) 8752 G, |D/hc| ) 0.4985 cm-1, |E/hc| ) 0.0222
cm-1; microwave frequency ) 9.530 GHz. Both of these
structural assignments are confirmed by excellent agree-
ment with previously reported spectroscopic data,11,13,36,37
and by comparison with B3LYP/6-31G* calculated IR
spectra (Figure 1). Other potential photoproducts, such
as allenes 6-8, can be ruled out on the basis of their
calculated IR spectra.34
Three new ESR absorptions arising from the photolysis
of 15 are ascribed to triplet 4,5-benzocycloheptatrien-
ylidene (2): (Ar, 14 K) Z1 ) 507 G, X2 ) 4651 G, Y2 )
5317 G, |D/hc| ) 0.368 cm-1, |E/hc| ) 0.0172 cm-1
;
a
b
Reference 25. See also refs 26 and 27. Reference 12. c Ref-
microwave frequency ) 9.529 GHz. On the basis of the
zero-field splitting parameters and the microwave fre-
quency,38 a weak Z2 transition is anticipated at ca. 7300
G. This absorption is likely unobserved due to the low
concentration of 2. The new species is thermally stable
in the dark (14 K, 11 h). As would be expected for an
endocyclic carbene, the signals are not mirrored by a
second set of absorptions arising from rotational isomer-
ism. The ESR transitions are consistent with a triplet
carbene, yet they differ substantially from the transi-
tions exhibited by the triplet naphthylcarbenes 10 and
12.11,13,15,37 Given that identical ESR signals are obtained
regardless of the method of diazo compound preparation,
the absorptions are unlikely to arise from a thermally
rearranged diazo compound or impurity within the
matrix.
erence 39. This work. e Reference 41. f Reference 13. g Reference
37.
d
radiation conditions employed in the earlier study. Thus,
the ESR signals previously attributed to triplet 4,5-
benzocycloheptatrienylidene (2) may actually belong to
triplet 2-naphthylcarbene (10).
The zero-field splitting parameters that we assign to
triplet 4,5-benzocycloheptatrienylidene (2) are in accord
with those of structurally similar triplet carbenes 1 and
25-27 (Table 1), but they differ considerably from those
reported earlier for 2.12,39 Actually, the previously re-
ported zero-field splitting parameters of 4,5-benzocyclo-
heptatrienylidene (2)12,39 display better agreement with
triplet 2-naphthylcarbene (10) than with triplet cyclo-
heptatrienylidene derivatives (Table 1). In a simplistic
analysis, benzannelation of cycloheptatrienylidene (1)
might be expected to increase the average distance
between the interacting π and σ electrons in triplet 4,5-
benzocycloheptatrienylidene (2), thereby decreasing the
value of D.40 Although this relationship is borne out in
the comparison of cycloheptatrienylidene (1) (D ) 0.425
cm-1) vs 4,5-benzocycloheptatrienylidene (2) (D ) 0.368
cm-1), Table 1 reveals that this relationship is not
consistently manifested among a larger series of benzan-
nelated cycloheptatrienylidenes. A simple structure-
In an earlier ESR study,12 a triplet species assigned
as 4,5-benzocycloheptatrienylidene (2) was obtained when
a hydrocarbon glass containing diazo compound 15 was
irradiated briefly with a 1000-W Xe-Hg arc lamp. In our
experiments, irradiation of carbene 2 under comparable
conditions (λ > 261 nm, 5 min, 300-W Xe arc lamp)
produces a sudden decrease in the ESR signal of 2 and
growth of the ESR signal of triplet 2-naphthylcarbene
(10) (Figure 3). Further irradiation results in the com-
plete disappearance of 2, while triplet 10 persists. This
control experiment suggests that triplet 4,5-benzocyclo-
heptatrienylidene (2) would not survive under the ir-
(39) Using the original data of Horn and Platz, we recalculated the
zero-field splitting parameters using an analytical mathematical
solution, rather than a graphical solution. The analytical solution
(37) Trozzolo, A. M.; Wasserman, E.; Yager, W. A. J . Am. Chem.
Soc. 1965, 87, 129-130.
(38) Wasserman, E.; Snyder, L. C.; Yager, W. A. J . Chem. Phys.
1964, 41, 1763-1772.
affords a somewhat smaller D value of 0.489 cm-1
.
(40) Carrington, A.; McLachlan, A. D. Introduction to Magnetic
Resonance; Harper & Row: New York, 1967.
9036 J . Org. Chem., Vol. 67, No. 25, 2002

Access to the Naphthylcarbene Rearrangement Manifold
SCHEME 9
property relationship among the species depicted in Table
1 is not readily apparent. The anomalously large D value
(0.52 cm-1) reported previously for triplet 4,5-benzocy-
cloheptatrienylidene (2) was rationalized in terms of a
nonplanar “boat” conformation of 2, conferred to the
carbene by its diazo precursor 15 in a rigid matrix.12,39
A
similar argument regarding steric hindrance to planarity
was invoked for carbene 27.41 In the case of 4,5-benzo-
cycloheptatrienylidene (2), this explanation appears to
be inconsistent with the computational predictions of
planar geometries for both diazo compound 1534 and
triplet carbene 2.14
Ma tr ix P h otoch em istr y of 2,3-Ben zod ia zocyclo-
h ep ta tr ien e (16). Irradiation (λ ) 670 or >613 nm) of
matrix-isolated 2,3-benzodiazocycloheptatriene (16), gen-
erated from thermolysis of tosylhydrazone salt 23,42
results in the disappearance of diazo compound ac-
companied by the appearance of 4,5-benzobicyclo[4.1.0]-
hepta-2,4,6-triene (11) as the sole product (Scheme 9) as
determined by IR and UV/visible spectroscopy (Figures
4 and 5). The basis for structural assignments is dis-
cussed below. Neither triplet 1-naphthylcarbene (12) nor
2,3-benzocycloheptatrienylidene (4) is detected by ESR
spectroscopy (Figure 6). Consistent with earlier work,15
cyclopropene 11 can be completely converted to 1-naph-
thylcarbene (12) by irradiation at λ ) 290 nm. Carbene
12 is characterized by IR, UV/visible, and ESR spectros-
copy, although only the s-E rotational isomer is detected.
As with the isomeric system (vide supra), a reversible
photoequilibrium exists between the naphthylcarbene
and its cyclopropene derivative. Band-pass or broadband
irradiation (λ ) 342 or >497 nm) induces complete
disappearance of 12 and reappearance of 11. Interest-
ingly, conditions for the regeneration of 1-naphthylcar-
bene (λ ) 290 nm) once again apparently produce only
the s-E isomer of 12.
F IGURE 4. (a) IR difference spectrum (Ar, 10 K) showing
independent IR spectra of 2,3-benzobicyclo[4.1.0]hepta-2,4,6-
triene (11) and triplet 1-naphthylcarbene (12), as generated
upon irradiation of 1-naphthyldiazomethane (see ref 15). (b)
Top: Calculated IR spectrum of 2,3-benzobicyclo[4.1.0]hepta-
2,4,6-triene (11). Middle: IR difference spectrum showing
conversion of 2,3-benzodiazocycloheptatriene (16) to cyclopro-
pene 11 after irradiation at λ ) 670 nm, 21 h. Bottom:
Calculated IR spectrum of 2,3-benzodiazocycloheptatriene (16).
All B3LYP/6-31G* calculated frequencies are scaled by 0.98.
copy: IR (Ar, 10 K) 3062 m, 3046 m, 1494 w, 1213 w,
1095 m, 1020 m, 785 s, 764 s, 719 w, 540 m cm-1; UV/
visible (Ar, 10 K) λmax 341, 328, 320, 307 nm; ESR (Ar,
14 K) s-E isomer Z1 ) 1538 G, X2 ) 4895 G, Y2 ) 5740
G, Z2 ) 8376 G, |D/hc| ) 0.4636 cm-1, |E/hc| ) 0.0209
cm-1; microwave frequency ) 9.530 GHz. The structural
assignment of each species is confirmed by excellent
agreement with previously reported spectroscopic
data11,13,15,36,37 and by comparison with B3LYP/6-31G*
calculated IR spectra (Figure 4). Other potential photo-
products, such as allenes 6-8, are ruled out based on
their calculated IR spectra.34
The failure to observe the minor (s-Z) rotational isomer
of 1-naphthylcarbene (12) is a puzzling result. Given the
reasonably intense ESR signals for 12 (Figure 6), it seems
unlikely that the s-Z isomer is present below the thres-
hold of instrumental detection. Previous investigations
indicate that the ratio of the integrated Z1 transitions
for the s-E and s-Z isomers of 1-naphthylcarbene (gener-
ated from 1-naphthyldiazomethane or 1-naphthyldiazir-
ine) is approximately 4:1.15 The carbene precursor (gen-
eration of 12 indirectly from 2,3-benzodiazocyclo-
heptatriene (16) versus directly from 1-naphthyldiazo-
methane) may impart a conformational bias through a
Cyclopropene 11 is characterized by IR and UV/visible
spectroscopy: IR (Ar, 10 K) 3020 m, 2981 m, 1763 m,
1751 m, 1468 s, 1450 s, 1102 w, 1033 w, 1003 w, 943 w,
930 w, 777 s, 752 m, 697 s, 669 s, 586 m, 568 m, 504 w,
489 w cm-1; UV/visible (Ar, 10 K) λmax 220 nm. Carbene
12 is characterized by IR, UV/visible, and ESR spectros-
(41) Moritani, I.; Murahashi, S.-I.; Nishino, M.; Yamamoto, Y.; Itoh,
K.; Mataga, N. J . Am. Chem. Soc. 1967, 89, 1259-1260.
(42) Qualitatively identical results are obtained when 16 is prepared
from the low-temperature oxidation of 24, although the material is
less pure and present in smaller amounts.
J . Org. Chem, Vol. 67, No. 25, 2002 9037

Bonvallet et al.
F IGURE 5. (a) UV/visible spectrum (Ar, 10 K) of 2,3-
benzodiazocycloheptatriene (16). (b) Irradiation at λ > 613 nm
for 7 h produces 4,5-benzobicyclo[4.1.0]hepta-2,4,6-triene (11).
(c) Cyclopropene 11 is converted to 1-naphthylcarbene (12) at
λ ) 290 nm, 1 h. (d) Broadband irradiation (λ > 497 nm, 3 h)
induces disappearance of 12 as 11 returns. Because the
extinction coefficients of the absorptions for 12 are much
smaller than that of the other species, spectra (a) and (b) were
obtained from a thin matrix, while spectra (c) and (d) were
obtained from a separate experiment employing a thicker
matrix.
F IGURE 6. (a) ESR spectrum (Ar, 14 K) showing no detect-
able triplet signal upon irradiation of 2,3-benzodiazocyclohep-
tatriene (16) at λ ) 670 nm, 37 h. (b) Subsequent irradiation
at λ > 613 nm for 13 h similarly fails to produce any triplet
species. (c) Spectrum of triplet s-E-1-naphthylcarbene (12)
after irradiation at λ ) 290 nm, 14 h. (d) ESR spectrum (Ar,
14 K) displaying both rotational isomers of triplet 1-naphth-
ylcarbene (12) generated upon irradiation (λ > 300 nm, 1 h)
of 1-naphthyldiazomethane. (Shown for the purpose of com-
parison to spectrum (c).) In each spectrum, the signal at ca.
3300 G is the g ) 2 (radical) impurity.
9038 J . Org. Chem., Vol. 67, No. 25, 2002

Access to the Naphthylcarbene Rearrangement Manifold
SCHEME 10a
SCHEME 11a
a
B3LYP/6-31G* energy (kcal/mol, ZPVE corrected), relative to
cyclobuta[de]naphthalene, in italics.
SCHEME 12a
a
B3LYP/6-31G* energy (kcal/mol, ZPVE corrected), relative to
cyclobuta[de]naphthalene, in italics.
a
B3LYP/6-31G* energy (kcal/mol, ZPVE corrected), relative to
cyclobuta[de]naphthalene, in italics. b Two imaginary frequencies.
c One imaginary frequency.
frequencies follow distinctly different modes of atomic
motion. One of these (237i cm-1), in which both hydrogens
flanking the carbene twist in opposite directions, leads
to either enantiomer of nonplanar allene 6. The other
imaginary frequency (297i cm-1) corresponds to an
upward out-of-plane deformation of the carbene carbon
as the flanking hydrogens both stretch downward. The
resulting structure converges to singlet Cs symmetric
carbene 28, which possesses zero imaginary frequencies
(Scheme 11). This behavior has been independently
discovered by Wentrup and co-workers.48 Comparison of
the calculated IR frequencies of “puckered” carbene 28
to the experimental data rule out its presence in the
matrix-isolation studies.34 Several unsuccessful attempts
were made to discover a transition state (TS2) connecting
singlet carbene 28 to cyclopropene 9 (Scheme 11). In each
case a transition structure was found, but intrinsic
reaction coordinate (IRC) calculations revealed that the
species was actually TS1, which connects cyclopropene
9 to allene 6 rather than to carbene 28. If TS2 does exist,
its structural similarity to TS1 complicates characteriza-
tion by computational study.
Elucidation of the structure of carbene 28 spurred us
to explore the relationship of these “benzocycloheptatrie-
nylidene” structures with the corresponding “benzonor-
caradienylidene” structures (Scheme 12). The “benzonor-
caradienylidene” (benzobicyclo[4.1.0]hepta-2,4-dien-7-
ylidene) structures may be pertinent to the mechanism
of addition of a carbon atom to naphthalene. Attempted
geometry optimization of singlet 29, whether the initial
structure was Cs or C1 symmetric, resulted in conver-
gence to the puckered 4,5-benzocycloheptatrienylidene 28
reported above. Triplet 29, on the other hand, is both a
stationary point and a minimum. The isomeric singlet
cyclopropylidene 30 is a minimum, in agreement with a
previous report,14 as is triplet 30. Triplet 30 possesses
an elongated carbon-carbon bond,34 but is otherwise
unremarkable. The third isomeric benzonorcaradien-
ylidene 31 is also considered. Triplet 31 displays an
abnormally long (1.94 Å)34 and partially occupied49 bond
cavity effect in the solid matrix. This hypothesis must
be considered with due caution, however, as the isomeric
2-naphthylcarbene system (vide supra) fails to display
this effect and gives both rotational isomers of 12 in their
“normal” proportion.
The ESR spectrum of carbene 12 obtained in this series
of experiments differs slightly from those obtained previ-
ously.15,37,43,44 The Z1 and X2 transitions are broad and
display some fine structure, and the Z2 transition is
broad. These characteristics, which may arise from either
matrix site effects or poor matrix isolation, could poten-
tially obfuscate a weak signal corresponding to s-Z 12.
Surprisingly, even when photochemically converted to
cyclopropene 11 and then regenerated, carbene 12 still
exhibits a single rotational isomer.34 This result differs
from the photochemical ring opening of cyclopropene 11
in earlier experiments,15 in which both rotational isomers
of carbene 12 are clearly obtained. Professor Wentrup
informed us of related observations regarding differences
in ESR spectra of arylcarbenes and nitrenes depending
on the precursor and method of generation (thermal or
photochemical).45,46 These intriguing phenomena are
simply not understood at the current time.
Com p u ta tion a l Resu lts. Although significant por-
tions of the C11H8 potential energy surface have already
been explored computationally,14 several issues regarding
the benzocycloheptatrienylidenes remain to be described.
The calculation of 3,4-benzocycloheptatrienylidene (3)
completes the comparison drawn in Scheme 2.47 Like its
related isomers at the B3LYP level of theory,14 planar
triplet 3 is a minimum while planar singlet 3 is a
transition structure for the enantiomerization of the
twisted allene 7. Singlet carbene 3 lies higher in energy
than singlet allene 7 and triplet carbene 3 by 29.6 and
11.6 kcal/mol, respectively (Scheme 10).
In singlet 4,5-benzocycloheptatrienylidene (2), calcu-
lated to be a second-order saddle point, both imaginary
(43) Senthilnathan, V. P.; Platz, M. S. J . Am. Chem. Soc. 1981, 103,
5503-5511.
(44) Roth, H. D.; Hutton, R. S. Tetrahedron 1985, 41, 1567-1578.
(45) Kuzaj, M. Ph.D. Thesis, University of Marburg, Germany, 1985.
(46) Lu¨erssen, H. Ph.D. Thesis, University of Marburg, 1985.
(47) This species is also reported in the Supporting Information of
ref 16.
(48) Kuhn, A.; Vosswinkel, M.; Wentrup, C. J . Org. Chem. 2002,
67, 9023-9030.
(49) Glendening, E. D.; Badenhoop, J . K.; Reed, A. E.; Carpenter,
J . E.; Weinhold, F. NBO 4.0; Theoretical Chemistry Institute, Uni-
versity of Wisconsin: Madison, 1996.
J . Org. Chem, Vol. 67, No. 25, 2002 9039

Bonvallet et al.
from the quaternary carbon to the carbene. Attempted
geometry optimizations of singlet cyclopropylidene 31
generally converged to the ring-opened allene 8. Employ-
ing the rather unusual optimized geometry of triplet 31
as a starting point in the search for singlet 31 resulted
in convergence to the isomeric singlet benzonorcaradi-
enylidene 30. Due to their extremely high energy and
failure to match with any experimentally observed IR
absorptions,34 none of the benzobicyclo[4.1.0]hepta-2,4-
dien-7-ylidenes are likely to be involved in the photo-
chemical interconversions of C11H8 isomers encountered
in the current study.
nonplanar structure of diazo compound 16, which leads
to an enhanced participation by vibrationally hot carbene
or excited-state diazo compound.
Su m m a r y
Matrix-isolated 4,5-benzodiazocycloheptatriene (15) is
photochemically converted to a mixture of triplet 4,5-
benzocycloheptatrienylidene (2), 2,3-benzobicyclo[4.1.0]-
hepta-2,4,6-triene (9), and triplet 2-naphthylcarbene (10).
Similarly, 2,3-benzodiazocycloheptatriene (16) produces
2,3-benzobicyclo[4.1.0]hepta-2,4,6-triene (11) and, even-
tually, the s-E rotational isomer of triplet 1-naphthyl-
carbene (12) without detection of triplet 2,3-benzocyclo-
heptatrienylidene (4). Under the appropriate photo-
equilibrating conditions, each cyclopropene and its re-
spective naphthylcarbene can be reversibly intercon-
verted. Benzocycloheptatetraenes 6-8 are not detected
under these experimental conditions.
Mech a n istic Im p lica tion s. The extensive rearrange-
ments observed upon gentle visible irradiation (λ ) 670
nm or >613 nm) of 4,5-benzodiazocycloheptatriene (15)
raise interesting mechanistic questions. The “expected”
triplet 4,5-benzocycloheptatrienylidene (2) is formed only
in a small amount. Possible explanations for the forma-
tion of cyclopropene 9 include (i) photosensitivity of
carbene 2 under the irradiation conditions, (ii) rearrange-
ment of vibrationally hot carbene 2, and (iii) rearrange-
ment in an excited state of diazo compound 15, bypassing
carbene 2 altogether.50 Control experiments rule out the
thermal conversion of carbene 2 to cyclopropene 9 at 14
K. Explanations involving vibrationally hot carbene or
excited-state diazo compound, while accounting for the
partitioning of diazo compound between carbene and
cyclopropene formation, do not rationalize the fact that
the absolute amount of carbene cannot be increased by
carrying the photolysis to higher conversion. Carbene
concentration is independent of diazo compound conver-
sion, which implies that carbene 2 is photosensitive under
the irradiation conditions. (This circumstance, of course,
does not preclude the involvement of vibrationally hot
carbene or excited diazo compound.) The observation of
triplet 2-naphthylcarbene (10) by ESR spectroscopy
under these irradiation conditions is an intriguing result.
Control experiments rule out the photochemical conver-
sion of cyclopropene 9 to carbene 10 under the irradiation
conditions (λ ) 670 nm or >613 nm). Possible explana-
tions for the formation of 2-naphthylcarbene (10) include
(i) photochemical rearrangement of carbene 2 under the
irradiation conditions, (ii) thermal rearrangement of
vibrationally hot carbene 2 or cyclopropene 9, and (iii)
rearrangement via another undetected intermediate
present in low, steady-state concentration.
The foregoing experiments establish the photosensitiv-
ity of carbene 2 at visible wavelengths. Other experi-
ments establish the photosensitivity of carbene 2 in the
UV. Irradiation (λ ) 290 or >261 nm) of a matrix
containing carbene 2 rapidly destroys 2 and causes triplet
2-naphthylcarbene (10) to appear (Figure 3), presumably
via cyclopropene 9.
Given our ability to prepare 2,3-benzodiazocyclohep-
tatriene (16), for the first time, the inability to generate
the corresponding triplet 2,3-benzocycloheptatrienyli-
dene (4) is disappointing. We can only speculate that this
circumstance reflects an even greater photosensitivity of
carbene 4 to the irradiation conditions (compared to the
case of the 4,5-isomer), or that it is a consequence of the
Meth od s
Com p u ta tion a l Meth od s. Calculations were performed
with the Becke three-parameter gradient-corrected exchange
functional51 with the correlation functional of Lee, Yang, and
Parr52 (B3LYP) as implemented in the Gaussian 98 package.53
Triplet states were computed using an unrestricted wave
function (UB3LYP) and closed-shell singlet states were com-
puted using a restricted wave function (RB3LYP). All singlet
states reported in this study are closed-shell; open-shell singlet
states were not investigated. The 6-31G* basis set was
employed for geometry optimizations as well as for single-point
energy and harmonic vibrational frequency calculations. Har-
monic vibrational frequency calculations established the num-
ber of imaginary frequencies for each species, and afforded the
zero-point vibrational energy correction and computed infrared
spectrum. The location of bonds and lone electron pairs was
verified with Natural Bond Orbital (NBO) calculations where
appropriate.49
Ma tr ix-Isola tion Sp ectr oscop y a n d P h otoch em istr y.
Techniques and apparatus for low-temperature matrix isola-
tion spectroscopy are described elsewhere.54,55 The deposition
temperature and argon flow rate remained constant regardless
of the particular instrumental technique employed, although
the matrix thickness (i.e. the amount of sample) was generally
greater for ESR and IR experiments than for a UV/visible
experiment. Irradiation was carried out with a 300-W high-
pressure xenon arc lamp, and wavelength selection was
achieved with cutoff filters or a high-energy monochromator.
Cutoff filters display <0.1% transmittance at the specified
wavelength: Corning 2-58 (>613 nm), Corning 3-69 (>497
(51) Becke, A. D. J . Chem. Phys. 1993, 98, 5648-5652.
(52) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B-Condens
Matter 1988, 37, 785-789.
(53) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; Montgomery, J . A.,
J r.; Stratmann, R. E.; Burant, J . C.; Dapprich, S.; Millam, J . M.;
Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J .;
Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo,
C.; Clifford, S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J . B.; Cioslowski, J .; Ortiz, J . V.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.;
Fox, D. J .; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; J ohnson, B. G.; Chen,
W.; Wong, M. W.; Andres, J . L.; Head-Gordon, M.; Replogle, E. S.;
Pople, J . A. Gaussian 98, revision A.6; Gaussian, Inc.: Pittsburgh, PA,
1998.
(54) McMahon, R. J .; Chapman, O. L.; Hayes, R. A.; Hess, T. C.;
Krimmer, H.-P. J . Am. Chem. Soc. 1985, 107, 7597-7606.
(55) Seburg, R. A.; McMahon, R. J . J . Am. Chem. Soc. 1992, 114,
7183-7189.
(50) For a recent review and leading references, see: Platz, M. S.
In Advances in Carbene Chemistry II; Brinker, U., Ed.; J AI Press: Boca
Raton, FL, 1998; Vol. 2, p 133.
9040 J . Org. Chem., Vol. 67, No. 25, 2002

Access to the Naphthylcarbene Rearrangement Manifold
nm), Corning 0-53 (>261 nm). Wavelengths selected with the
monochromator are (10 nm. Photolysis times often varied
depending on the instrumental technique employed. IR spectra
were recorded on an FT-IR spectrometer equipped with a
MCT-B detector. UV/visible spectra were acquired using a 2
nm slit width. ESR experiments were conducted at X-band.
Zero-field splitting parameters were assigned by a best fit of
the observed spectrum to the spin Hamiltonian (assuming gx
) gy ) gz ) ge).38
solution was cooled and filtered, and the filtrate was diluted
with 300 mL of water. The solution was extracted with 3 ×
250 mL of CH2Cl2 and the combined organic extracts were
washed successively with 500 mL of water and 500 mL of
brine. The extracts were dried over MgSO4, filtered, and
concentrated by rotary evaporation to a dark brown oil. The
products were isolated by column chromatography (silica gel,
30% ethyl acetate/hexane), 2,3-benzotropone (21) as a red-
orange oil (534 mg, 6%, Rf 0.39) and 4,5-benzotropone (17) as
a red-brown solid (1.03 g, 12%, Rf 0.21); 2,3-benzotropone (21)
1H NMR δ 6.71 (ddd, J ) 12, 8, 1 Hz, 1H), 6.96 (d, J ) 12 Hz,
1H), 7.07 (ddd, J ) 12, 8, 1 Hz, 1H), 7.33 (d, J ) 11 Hz, 1H),
7.68 (m, 3H), 8.57 (dd, J ) 8, 2 Hz, 1H); 13C NMR δ 126.5,
130.3, 130.7, 132.2, 133.7, 135.3, 135.7, 136.1, 138.7, 139.3,
188.2; MS m/z (rel intensity) 156 (M+, 85), 131 (13), 128 (100),
127 (27), 64 (21), 51 (26); HRMS calcd for C11H8O 156.0575,
found 156.0582; 4,5-benzotropone (17) mp 59-62 °C (lit.57 mp
Gen er a l E xp er im en t a l Met h od s. Carbon tetrachloride
was dried over molecular sieves, and solvent grade dichlo-
romethane and ether were distilled from CaH2. All reactions
are carried out under a nitrogen atmosphere. Compounds from
1
commercial sources gave acceptable H NMR spectra and were
used without further purification. Uncorrected melting points
1
were measured in open capillaries. H and 13C NMR spectra
were obtained in CDCl3 on a 300-MHz spectrometer, with
chemical shifts reported in parts per million downfield from
Me4Si. Mass spectra were obtained employing electron impact
ionization unless otherwise specified.
1
63-64 °C), H NMR δ 6.82 (d, J ) 12 Hz, 2H), 7.47 (d, J ) 12
Hz, 2H), 7.60 (m, 2H), 7.68 (m, 2H); 13C NMR δ 130.4, 134.0,
134.9, 136.0, 141.5, 188.3; MS m/z (rel intensity) 156 (M+, 38),
133 (5), 128 (100), 64 (15), 51 (19); HRMS calcd for C11H8O
156.0575, found 156.0577.
Ben zosu ber en e.56 Sodium borohydride (6.03 g, 159 mmol)
was added portionwise to a solution of 1-benzosuberone (20.05
g, 125 mmol) in 200 mL of methanol and 0.5 mL of water. After
the vigorous reaction subsided, the solution was stirred for 16
h. The solution was then refluxed briefly (10 min), cooled,
diluted with 200 mL of water, and extracted with 3 × 200 mL
of ether. The combined organic extracts were dried with MgSO4
and filtered, and the solvent was removed by rotary evapora-
tion to afford 1-hydroxybenzosuberane in quantitative yield
as a white solid: 1H NMR δ 1.75-2.09 (m, 6H), 2.71 (ddd, J
)14, 10, 1 Hz, 1H), 2.91 (dd, J ) 14, 8 Hz, 1H), 3.46 (d, J ) 4
Hz, 1H), 4.92 (d, J ) 6 Hz, 1H), 7.07-7.23 (m, 3H), 7.43 (d, J
) 7 Hz, 1H); 13C NMR δ 27.6, 27.8, 35.7, 36.5, 73.9, 124.5,
126.1, 126.9, 129.4, 140.8, 144.2; MS m/z (rel intensity) 162
(M+, 13), 144 (100), 133 (61), 129 (64), 116 (24), 105 (31), 91
(60), 77 (17), 65 (11).
4,5-Ben zotr op on e Tosylh yd r a zon e (18).5 A solution of
4,5-benzotropone 17 (625 mg, 4 mmol) and p-toluenesulfonyl-
hydrazide (745 mg, 4 mmol) in 8 mL of absolute ethanol was
stirred for 24 h. The resulting precipitate was isolated by
vacuum filtration and washed with cold ethanol to afford the
tosylhydrazone as a yellow solid (340 mg, 24%): mp 180-182
1
°C dec; H NMR δ 2.41 (s, 3H), 6.43 (m, 2H), 6.64 (d, J ) 12
Hz, 1H), 6.81 (d, J ) 12 Hz, 1H), 7.30 (m, 6H), 7.69 (br s, 1H),
7.88 (d, J ) 8 Hz, 2H); 13C NMR δ 21.6, 119.1, 128.1, 128.9,
129.5, 130.2, 132.1, 132.4, 133.5, 134.4, 134.8, 135.3, 136.6,
138.8, 144.0, 152.6; MS m/z (rel intensity) 192 (7), 169 (74),
139 (100), 91 (99); electrospray ionization MS m/z 325 (MH+);
HRMS (electrospray ionization) calcd for C18H17N2O2S 325.1011
(MH+), found 325.1017.
The white solid was dissolved in a solution of 600 mL of
benzene and p-toluenesulfonic acid monohydrate (1.12 g, 6
mmol). Water was azeotropically removed (Dean-Stark trap)
as the solution was refluxed for 4 h. The solution was cooled,
washed with 400 mL of water, dried with MgSO4, filtered, and
concentrated to a pale yellow oil (16.79 g, 93% from 1-benzo-
suberone). The benzosuberene requires no further purifica-
tion: 1H NMR δ 1.92-2.00 (m, 2H), 2.39-2.45 (m, 2H), 2.83-
2.86 (m, 2H), 5.90 (dt, J ) 12, 4 Hz, 1H), 6.40 (dt, J ) 12, 2
Hz, 1H), 7.08-7.16 (m, 4H); 13C NMR δ 26.9, 32.5, 36.1, 125.9,
126.6, 129.0, 129.8, 130.8, 132.2, 136.3, 141.7; MS m/z (rel
intensity) 144 (M+, 56), 129 (100), 128 (37), 115 (20); HRMS
calcd for C11H12 144.0939, found 144.0937.
1,2-Ben zo-1,3,5-cycloh ep ta tr ien e.57 A solution of benzo-
suberene (15.74 g, 109 mmol), NBS (19.44 g, 109 mmol), and
benzoyl peroxide (72 mg, 0.3 mmol) in 110 mL of dry CCl4 was
refluxed for 2 h and irradiated with a 200-W tungsten lamp.
The solution was cooled and the resulting solid succinimide
was removed by vacuum filtration. The filtrate was washed
with 40 mL of 5% NaHCO3 followed by 80 mL of water, dried
over MgSO4, filtered, and concentrated by rotary evaporation
to an orange oil. The product was isolated by vacuum distil-
lation (64-68 °C, 0.2 mmHg) as a colorless oil (10.52 g, 68%):
1H NMR δ 3.03 (d, J ) 7 Hz, 2H), 5.77 (m, 1H), 6.07 (dd, J )
10, 6 Hz, 1H), 6.46 (dd, J ) 12, 6 Hz, 1H), 7.07 (d, J ) 11 Hz,
1H), 7.18 (m, 2H), 7.31 (m, 2H); 13C NMR δ 34.4, 125.4, 125.8,
127.0, 127.4, 127.8, 128.5, 128.6, 133.4, 136.0, 136.3; MS m/z
(rel intensity) 142 (M+, 73), 141 (100), 129 (17), 115 (33); HRMS
calcd for C11H10 142.0783, found 142.0768.
2,3-Ben zotr op on e Tosylh yd r a zon e (22). A mixture of
2,3-benzotropone 21 (460 mg, 3.0 mmol) and p-toluenesul-
fonylhydrazide (490 mg, 2.6 mmol) in 5 mL of absolute ethanol
was refluxed under nitrogen for 1 h and then stirred at room
temperature for an additional 48 h. A light colored solid (91
mg), presumably a double tosylhydrazide adduct, was removed
by vacuum filtration. The filtrate was concentrated and
chromatographed (silica gel, 1:1 ether/hexanes) to afford
unreacted 2,3-benzotropone (69 mg, 17%, Rf 0.29) and the
tosylhydrazone as a bright yellow solid (263 mg, 31%, Rf
0.16): mp 132-134 °C (lit.31 mp 134-135 °C); 1H NMR δ 2.39
(s, 3H), 6.32 (m, 2H), 6.50 (ddd, J ) 12, 7, 1 Hz, 1H), 6.92 (d,
J ) 12 Hz, 1H), 7.28 (d, J ) 8 Hz, 2H), 7.33 (dd, J ) 8, 2 Hz,
1H), 7.43 (td, J ) 7, 2 Hz, 1H), 7.50 (td, J ) 7, 2 Hz, 1H), 7.62
(br s, 1H), 7.80 (dd, J ) 8, 2 Hz, 1H), 7.87 (d, J ) 8 Hz, 2H);
13C NMR δ 21.5, 121.1, 126.0, 128.1, 128.5, 129.0, 129.5, 130.2,
130.6, 131.8, 134.8, 135.4, 135.7, 136.5, 144.0, 152.6; MS m/z
(rel intensity) 277 (11), 217 (100), 185 (53), 140 (100), 94 (44),
69 (62); HRMS (electrospray ionization) calcd for C18H17N2O2S
325.1011 (MH+), found 325.0994.
Ben zotr op on e Tosylh yd r a zon e Sod iu m Sa lts 19 a n d
23. Sodium hydride (60% dispersion in oil, 20 mg, 0.5 mmol)
was slowly added to a solution of 4,5-benzotropone tosylhy-
drazone 18 or 2,3-benzotropone tosylhydrazone 22 (100 mg,
0.3 mmol) in 5 mL of dry CH2Cl2. The orange slurry was stirred
under argon for 60 min and poured into 100 mL of hexane.
The resulting bright orange precipitate was isolated by vacuum
filtration and dried in vacuo (room temperature, e10-6 mmHg)
before use. The salts were sensitive to moisture and thus were
dried immediately.
4,5-Ben zotr op on e (17) a n d 2,3-Ben zotr op on e (21).57 1,2-
Benzo-1,3,5-cycloheptatriene (7.77 g, 54.6 mmol), SeO2 (7.77
g, 70.0 mmol), and KH2PO4 (2.33 g, 17.1 mmol) in 70 mL of
dioxane and 10 mL of water were refluxed for 16 h. The
4,5-Ben zotr op on e h yd r a zon e (20).12 A solution of 4,5-
benzotropone 17 (300 mg, 1.9 mmol) in 1.2 mL of absolute
ethanol was added over 5 min to a solution of hydrazine
monohydrate in 4.8 mL of ethanol at room temperature. The
mixture was refluxed for 1 h and gradually darkened to a red-
brown. After cooling and dilution with 35 mL of CHCl3, the
(56) Rona, P. J . Chem. Soc. 1962, 3629-3635.
(57) Srivastava, K. C.; Dev, S. Tetrahedron 1972, 28, 1083-1091.
J . Org. Chem, Vol. 67, No. 25, 2002 9041

Bonvallet et al.
solution was washed with 3 × 15 mL of water. The organic
phase was dried over Na2SO4, filtered, and concentrated to an
orange solid. The hydrazone was purified by column chroma-
tography (basic alumina, activity III, CHCl3) as a bright red-
orange solid (90 mg, 28%, Rf 0.09): 1H NMR δ 5.26 (br s, 2H),
6.28-6.41 (m, 3H), 6.60 (d, J ) 12.6 Hz, 1H), 7.14-7.26 (m,
4H); 13C NMR δ 119.0, 128.2, 129.4, 130.8, 131.9, 132.9, 133.7,
135.2, 135.9, 137.0, 148.7.
2,3-Ben zotr op on e Hyd r a zon e (24). A solution of 2,3-
benzotropone 21 (320 mg, 2.1 mmol) in 1.5 mL of absolute
ethanol was added over 5 min to a solution of hydrazine
monohydrate (290 mg, 6 mmol) in 5 mL of ethanol at room
temperature. The mixture was refluxed for 20 h and gradually
darkened to a red-brown. After cooling and dilution with 35
mL of CHCl3, the solution was washed with 3 × 15 mL of
water. The organic phase was dried over Na2SO4, filtered, and
concentrated to an orange oil. The hydrazone was purified by
column chromatography (basic alumina, activity III, CHCl3)
as a bright red-orange solid (90 mg, 28%, Rf 0.17): 1H NMR δ
5.41 (br s, 2H), 6.17-6.23 (m, 1H), 6.28-6.31 (m, 2H), 6.79 (d,
J ) 12 Hz, 1H), 7.20-7.23 (m, 1H), 7.28 (td, J ) 8, 1 Hz, 1H),
7.37 (td, J ) 7, 1 Hz, 1H), 7.58 (d, J ) 8 Hz, 1H); 13C NMR δ
122.4, 126.3, 127.8, 128.0, 129.8, 103.0, 103.2, 134.5, 135.8,
137.5, 148.8.
P r ep a r a tion a n d Ma tr ix Isola tion of Ben zod ia zocy-
cloh ep ta tr ien es 15 a n d 16. (A) P yr olytic m eth od : ben-
zotropone tosylhydrazone sodium salt 19 or 23 was transferred
to a round-bottom flask that was affixed directly to the matrix
isolation apparatus. The salt was heated in an oil bath (125
°C for 19, 90 °C for 23) at 10-7 mmHg and the pyrolysate co-
deposited with argon onto a spectroscopic window at 30 K. (B)
Oxid a tive m eth od :12 a suspension of silver oxide58 (230 mg,
1 mmol) and benzotropone hydrazone 20 or 24 (90 mg, 0.5
mmol) in 25 mL of dry ether was stirred in the dark at 0 °C
for 1 h. The solution was quickly filtered and transferred to a
precooled round-bottom flask, and the solvent was removed
by rotary evaporation in the dark at 0 °C. The dark orange
solid was extracted with 5 × 2 mL of distilled hexane and the
extracts were combined in a matrix-isolation deposition tube.
Crude diazo compound was isolated as an orange solid after
the hexane was removed by evaporation (0 °C at 0.02 mmHg).
The diazo compound was matrix isolated by subliming the
crude material (15 °C at 10-7 mmHg for both 15 and 16) and
co-depositing with argon at 30 K. 4,5-Benzodiazocyclohep-
tatriene (15): IR (Ar, 10 K) 2045 vs, 1981 w, 1650 m, 1608 m,
1569 w, 1450 m, 1444 m, 1332 w, 1190 w, 1054 w, 891 m, 880
m, 775 s, 560 w cm-1; UV/visible (Ar, 10 K) λmax 365, 351, 338,
268, 260 nm. 2,3-Benzodiazocycloheptatriene (16): IR (Ar, 10
K) 3050 m, 3030 m, 2049 vs, 2040 vs, 1646 m, 1563 s, 1490 m,
1433 m, 1420 w, 1408 w, 1371 w, 1339 m, 1246 m, 1221 s,
1180 w, 1071 w, 1072 w, 1049 m, 953 m, 939 m, 882 s, 795 m,
750 s, 745 s, 686 s, 626 m, 598 w cm-1; UV/visible (Ar, 10K)
λmax 376, 348, 332, 299, 259, 252 nm. In either method the
matrix was cooled to 10 K prior to irradiation.
Thermolysis of 4,5-benzotropone tosylhydrazone sodium salt
19 (130 °C, 0.05 mmHg) in the dark in a sublimation apparatus
afforded a dark orange solid (-78 °C coldfinger) that was
nonvolatile and thus could not be co-deposited with argon to
form a matrix. Proton NMR at room temperature indicated
that the solid was a complex mixture of unknown composition.
Ack n ow led gm en t. We thank Prof. Curt Wentrup
(University Queensland) for stimulating discussions and
for sharing his results in advance of publication. We
thank Dr. Keith Horn (Corning) and Prof. Matthew
Platz (Ohio State University) for providing a copy of
their triplet ESR spectrum and for fruitful discussions.
We gratefully acknowledge the National Science Foun-
dation for financial support of this project (CHE-
9800716; CHE-0110769) as well as departmental com-
putationalfacilities(CHE-0091916)andtheESRspectrometer
(CHE-9013030).
Su p p or tin g In for m a tion Ava ila ble: Alternate synthesis
of 2,3-benzotropone (21); attempted syntheses of diazirines 13
and 14; additional spectra from matrix experiments; calculated
energy, IR spectrum, and Cartesian coordinates of selected
species. This material is available free of charge via the
Internet at http://pubs.acs.org.
(58) Broadus, K. M.; Kass, S. R. J . Org. Chem. 2000, 65, 6566-
6571.
J O020304Z
9042 J . Org. Chem., Vol. 67, No. 25, 2002