The elusive benzocyclobutenylidene: A combined computational and experimental attempt

Article abstract of DOI:10.1021/ja0039482

Ab initio and density functional theory calculations predict that benzocyclobutenylidene (1) has a singlet ground state in contrast to the parent phenylcarbene and many other simply substituted arylcarbenes. Calculations also predict that 1 should lie in a relatively deep potential well, while its triplet state is 14.5 kcal mol^-1 higher in energy. However, attempts to observe 1 directly by photolysis of two different nitrogenous precursors were not successful. Irradiation of diazobenzocyclobutene (7) (λ > 534 nm or λ > 300 nm) or azibenzocyclobutene (10) (λ > 328 nm) in Ar matrixes at 10 K leads to the formation of the strained cycloalkyne 7-methylenecyclohepta-3,5-dien-1-yne (3). ¹³C-Labeled 3 was also prepared in a similar manner. There is very good agreement between experimental IR spectra and computationally derived harmonic vibrational frequencies for 3 and [¹³C]-3 and excellent agreement between observed and calculated isotopic shifts. Prolonged short-wavelength irradiation converts 3 into benzocyclobutadiene (5). Phenylacetylene (6) and benzocyclobutadiene dimer (11) were identified as products arising from flash vacuum pyrolysis of diazirine 10 at 500 °C.

Full text of DOI:10.1021/ja0039482

2870
J. Am. Chem. Soc. 2001, 123, 2870-2876
The Elusive Benzocyclobutenylidene: A Combined Computational
and Experimental Attempt
Athanassios Nicolaides,*,†,‡ Takeshi Matsushita,‡ Kohichi Yonezawa,‡ Shinji Sawai,‡
Hideo Tomioka,*,‡ Louise L. Stracener,§ Jonathan A. Hodges,§ and Robert J. McMahon*,§
Contribution from the Chemistry Department for Materials, Faculty of Engineering, Mie UniVersity,
Tsu, Mie 514-8507 Japan, and the Department of Chemistry, UniVersity of Wisconsin,
1101 UniVersity AVenue, Madison, Wisconsin 53706-1396
ReceiVed NoVember 13, 2000
Abstract: Ab initio and density functional theory calculations predict that benzocyclobutenylidene (1) has a
singlet ground state in contrast to the parent phenylcarbene and many other simply substituted arylcarbenes.
Calculations also predict that 1 should lie in a relatively deep potential well, while its triplet state is 14.5 kcal
mol-1 higher in energy. However, attempts to observe 1 directly by photolysis of two different nitrogenous
precursors were not successful. Irradiation of diazobenzocyclobutene (7) (λ > 534 nm or λ > 300 nm) or
azibenzocyclobutene (10) (λ > 328 nm) in Ar matrixes at 10 K leads to the formation of the strained cycloalkyne
7-methylenecyclohepta-3,5-dien-1-yne (3). 13C-Labeled 3 was also prepared in a similar manner. There is
very good agreement between experimental IR spectra and computationally derived harmonic vibrational
frequencies for 3 and [13C]-3 and excellent agreement between observed and calculated isotopic shifts. Prolonged
short-wavelength irradiation converts 3 into benzocyclobutadiene (5). Phenylacetylene (6) and benzocyclo-
butadiene dimer (11) were identified as products arising from flash vacuum pyrolysis of diazirine 10 at 500
°C.
Introduction
Phenylcarbene, the prototypical arylcarbene,1 has a triplet
ground state.2 However, it is through its low-lying singlet state
that phenylcarbene usually reacts.3 Thus, factors that affect the
S-T gap are important in understanding the reactivity of
arylcarbenes.1 In terms of geometric factors, a small bond angle
at a carbenic center energetically favors the singlet state relative
to the triplet. Thus, benzocyclobutenylidene (1) provides an
attractive model for exploring the interdependence of carbene
geometry, substitution, and multiplicity. The bond angle at the
divalent carbon in 1 is significantly smaller than that of
phenylcarbene (90° vs 135-155°),2 although in other respects
the electronic perturbation introduced by the small ring is
minimal.
Several groups attempted to study carbene 1 or its simple
derivatives by photolysis or thermolysis of various precursors.4-6
Product analyses revealed a complex and intriguing reactivity
for 1 but did not provide a clear conclusion as to the multiplicity
of the ground state. Semiempirical calculations suggested a
triplet ground state.5 An equilibrium between singlet and triplet
states was invoked to explain the experimental data. Photolysis
of the sodium salt of benzocyclobutenone p-toluenesulfonyl-
hydrazone in solution afforded the (formal) dimer and trimer
of 1 as major products.6 Although all of the earlier reactions
† Current address: Department of Chemistry, University of Cyprus,
Nicosia, Cyprus.
‡ Mie University.
§ University of Wisconsin.
(1) Kirmse, W. Carbene Chemistry, 2nd ed.; Academic Press: New York,
1971. Jones, M., Moss, R. A., Eds. Carbenes; R. E. Krieger: Malabar, FL,
1983; Vol. I. Moss, R. A., Jones, M., Eds. Carbenes; R. E. Krieger:
Malabar, FL, 1983; Vol. II. Wentrup, C. ReactiVe Molecules; John Wiley:
New York, 1984. Regitz, M. In Methoden der Organischen Chemie
(Houben-Weyl); G. Thieme Verlag: Stuttgart, 1989; Vol. E19b. Platz, M.
S., Ed. Kinetics and Spectroscopy of Carbenes and Biradicals; Plenum:
New York, 1990. Sander, W.; Bucher, G.; Wierlacher, S. Chem. ReV. 1993,
93, 1583-1621.
(2) Trozzolo, A. M.; Murray, R. W.; Wasserman, E. J. Am. Chem. Soc.
1962, 84, 4990-4991. Higuchi, J. J. Chem. Phys. 1963, 38, 1237-1245.
Wasserman, E.; Trozzolo, A. M.; Yager, W. A.; Murray, R. W. J. Chem.
Phys. 1964, 40, 2408-2410. West, P. R.; Chapman, O. L.; LeRoux, J.-P.
J. Am. Chem. Soc. 1982, 104, 1779-1782. McMahon, R. J.; Abelt, C. J.;
Chapman, O. L.; Johnson, J. W.; Kreil, C. L.; LeRoux, J.-P.; Mooring, A.
M.; West, P. R. J. Am. Chem. Soc. 1987, 109, 2456-2469. Haider, K. W.;
Platz, M. S.; Despres, A.; Migirdicyan, E. Chem. Phys. Lett. 1989, 164,
443-448. Matzinger, S.; Bally, T.; Patterson, E. V.; McMahon, R. J. J.
Am. Chem. Soc. 1996, 118, 1535-1542. Wong, M. W.; Wentrup, C. J.
Org. Chem. 1996, 61, 7022-7029. 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.
gave products consistent with those expected from carbene 1,
the carbene has not been observed spectroscopically and its
involvement in the aforementioned chemistry has not been
(3) Savino, T. G.; Kanakarajan, K.; Platz, M. S. J. Org. Chem. 1986,
51, 1305-1309.
10.1021/ja0039482 CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/03/2001

The ElusiVe Benzocyclobutenylidene
J. Am. Chem. Soc., Vol. 123, No. 12, 2001 2871
Scheme 1a
Scheme 2
a G2(MP2,SVP) relative energies (kcal mol-1).
Table 1. Relative Energiesa of Various C8H6 Isomers (∆H, kcal
mol-1) and Barriers of Reactionsa (∆H‡, kcal mol-1
)
∆Hb
∆Hc
∆H‡,c
11
31
2
3
4
0.0
7.8
13.3
-10.5
-27.8
-37.5
-59.2
0.0
14.5e
14.5
-9.2
-27.4
-38.5
TS(11-11)d
6.8b-7.0c
TS(1-2)
TS(1-3) f
TS(1-4)
TS(1-5)
15.1
21.7
41.0
28.4
mated in two ways. The value of -7.8 kcal mol-1 was
determined directly from the BLYP relative energies, while the
value of -14.5 kcal mol-1 was determined using the experi-
mental singlet-triplet gap of methylene (∆ES-T = +9.0 kcal
5
6
a At 0 K. b BLYP/6-31G(d) + ZPVE. c G2(MP2,SVP) relative
energies using B3LYP/6-31G(d) optimized structures. d Transition state
for conformational ring inversion in 11. e Determined using isodesmic
reaction (see text). f Via intermediacy of 2.
mol-1 10
) and the G2(MP2,SVP) energy of the isodesmic reaction
1
1
3
31 + CH2 f 1 + CH2.
The intramolecular rearrangement pathways of singlet ben-
zocyclobutenylidene (11) were examined computationally (Scheme
1): (a) 1,2 hydrogen shift (the most common rearrangement
path for carbenes bearing R-hydrogens),11 (b) 1,2 carbon shift
(this path can become dominant in sterically strained carbenes),12
and (c) ring opening of the fused ring (forming vinylidene 2)
followed by a C-C insertion leading to strained cycloalkyne
3. Path c may seem unusual, but it is likely to be the path
followed in the isomerization of cyclobutenylidene.7,8 From the
calculated enthalpic barriers for the three processes (Table 1),
it is found that while path c is preferred, its barrier of 21.7 kcal
mol-1 is still large. Therefore, it appears that benzocyclobuten-
ylidene 1 lies in a rather deep potential well and should be
observable under suitable experimental conditions.
unequivocally established. We therefore decided to investigate
this interesting carbene by attempting to observe it directly in
an inert matrix at low temperature and studying it at modern
levels of theory.
Results and Discussion
The G2(MP2,SVP) relative energies for several C8H6 isomers
(1-6), computed using B3LYP/6-31G(d) geometries, are given
in Scheme 1 and Table 1. According to our calculations, the
ground state of benzocyclobutenylidene 1 is a singlet and its
geometry is nonplanar (Table S1). Singlet 1 undergoes confor-
mational ring inversion through a planar transition state with a
computed barrier of ca. 7 kcal mol-1. Triplet 1 is found to be
planar and to lie 8-15 kcal mol-1 higher in energy than the
singlet ground state. This situation is similar to the simple
analogue of 1, cyclobutenylidene, which displays a pronounced
deviation from planarity in its singlet state.7-9 However, in the
case of 1, the presence of the rather rigid benzene ring does
not allow for much out-of-plane deformation. The degree of
stabilization afforded the singlet state by virtue of the geometric
distortion is reflected in the planarization barrier, which is
calculated to be 7 kcal mol-1 for benzocyclobutenylidene (1)
but 21 kcal mol-1 for cyclobutenylidene.8,9 The singlet-triplet
energy gap (∆ES-T) of benzocyclobutenylidene (1) was esti-
An obvious photochemical precursor for 1 is diazobenzocy-
clobutene (7), but this species was found to be too unstable for
isolation and purification under normal preparative conditions.
Therefore, we sought to generate 7 in the matrix by two different
(7) Dyer, S. F.; Kammula, S.; Shevlin, P. B. J. Am. Chem. Soc. 1977,
99, 8104-8106. Kollmar, H.; Carrion, F.; Dewar, M. J. S.; Bingham, R.
C. J. Am. Chem. Soc. 1981, 103, 5292-5303.
(8) Nicolaides, A.; Matsushita, T.; Tomioka, H. J. Org. Chem. 1999,
64, 3299-3305.
(9) Stracener, L. L. Ph.D. Thesis, University of Wisconsin, Madison,
WI, 1998.
(10) McKellar, A. R. W.; Bunker, P. R.; Sears, T. J.; Evenson, K. M.;
Saykally, R. J.; Langhoff, S. R. J. Chem. Phys. 1983, 79, 5251-5264.
(11) Liu, M. T. H. Acc. Chem. Res. 1994, 27, 287-294. Evanseck, J.
D.; Houk, K. N. J. Am. Chem. Soc. 1990, 112, 9148-9156. Evanseck, J.
D.; Houk, K. N. J. Phys. Chem. 1990, 94, 5518-5523.
(12) Wiberg, K. B.; Burgmaier, G. J.; Warner, P. J. Am. Chem. Soc.
1971, 93, 246-247. Shevlin, P. B.; McKee, M. L. J. Am. Chem. Soc. 1989,
111, 519-524. Sulzbach, H. M.; Platz, M. S.; Schaefer, H. F.; Hadad, C.
M. J. Am. Chem. Soc. 1997, 119, 5682-5689. Pezacki, J. P.; Pole, D. L.;
Warkentin, J.; Chen, T. Q.; Ford, F.; Toscano, J. P.; Fell, J.; Platz, M. S.
J. Am. Chem. Soc. 1997, 119, 3191-3192.
(4) Blomquist, A. T.; Heins, C. F. J. Org. Chem. 1969, 34, 2906-2908.
O’Leary, M. A.; Wege, D. Tetrahedron Lett. 1978, 2811-2814. O’Leary,
M. A.; Richardson, G. W.; Wege, D. Tetrahedron 1981, 37, 813-823.
(5) Du¨rr, H.; Nickels, H.; Pacala, L. A.; Jones, M., Jr. J. Org. Chem.
1980, 45, 973-980.
(6) Frimer, A. A.; Weiss, J.; Rosental, Z. J. Org. Chem. 1994, 59, 2516-
2522.

2872 J. Am. Chem. Soc., Vol. 123, No. 12, 2001
Nicolaides et al.
Table 2. Observed and Calculateda Frequencies (ν, cm-1), Relative Intensities (%I) and 13C Isotopic Shifts (∆ν, cm-1) for Cycloalkyne 3
observed
calculated
[1-13C]-3
[12C]-3
[13C]-3
[12C]-3
[2-13C]-3
ν
%I
ν
%I
∆ν
ν
%I
ν
%I
∆ν
ν
%I
∆ν
608
616
732
848
55.6
59.3
107.0
66.7
603
616
732
848
39.1
48.4
122.0
57.8
5
0
0
0
587
614
728
30.3
19.0
34.5
61.5
34.1
36.2
100.0
35.8
586
613
728
29.2
18.6
34.5
61.4
34.2
36.5
100.0
32.5
1
1
0
0
0
0
0
41
583
613
728
28.5
20.0
35.0
62.7
34.6
36.8
100.0
38.4
4
1
0
0
0
0
1
40
813
813
813
1497b
1569b
1626
2147
1497b
1569b
1626
2106
1497b
1569b
1625
2107
1564b
1612
2098
29.6
100.0
22.2
1564b
1606
2062
29.7
100.0
18.8
0
6
36
a At the B3LYP/6-31G(d) level of theory. Only diagnostically useful frequencies are shown. For a complete list of the vibrational analysis data
for 3 and its isomers, see Tables S2 and S3. b The observed peak is quite broad, suggesting that there are two absorptions very close to each other.
Apparently this level of theory overestimates the relative distance of the two peaks.
Scheme 3
between the experimental IR spectrum of A and the computed
spectrum for singlet benzocyclobutenylidene (11) is very poor
(Table S2). These pieces of evidence are inconsistent with a
benzocyclobutenylidene intermediate (11 or 31), suggesting that
intermediate A is perhaps a rearrangement product of 1.
Comparison between the experimental IR spectrum of A and
the IR spectra computed for other C8H6 isomers (Scheme 1)
enables the assignment of photoproduct A as 7-methylenecy-
clohepta-3,5-dien-1-yne (3) (Figure 1).14 The main discriminat-
ing factor in the computed IR spectra of vinylidene 2 and
cycloalkyne 3 is the presence of the CtC bond absorption in
the latter, which experimentally is observed at 2098 cm-1 15
.
While this absorption is weak, it is well separated from the rest
of the absorptions and this suggests 13C isotopic labeling as a
means of differentiating between the two possible structures.
Calculations predict an isotopic shift of approximately 10 cm-1
for the CdC of vinylidene 2, and a much larger shift of
approximately 40 cm-1 for either of the two 13C isotopomers
of the CtC of cycloalkyne 3 (Tables 2 and S3). Labeled 7 (with
the 13C-label at the diazo carbon) was synthesized, and the
photolysis was carried out under conditions identical to those
described previously (Scheme 3). A small isotopic shift of 6
cm-1 (calculated: 4 cm-1) was observed for the CdNdN
stretch of the labeled diazo compound. The subsequent photo-
product (13C-A, Figure 1a) showed an absorption at 2062 cm-1
Figure 1. Comparison of experimental (Ar, 10 K) and theoretical IR
spectra: (a) Experimental IR spectrum of the photoproducts formed
by irradiation of unlabeled N-aziridinylimine 8. Bands due to photo-
product A are marked with • and those due to styrene with ×. (b)
Calculated IR spectrum (B3LYP/6-31G(d) scaled by 0.961) for
unlabeled cycloalkyne 3. The insets in both a and b show the major
13C isotope effects on the spectrum (see Table 2).
methods (Scheme 2). In the first case, N-aziridinylimine 8 was
prepared in a pure form and deposited into the Ar matrix.
Irradiation of 8 (λ > 300 nm) gave rise to diazo compound 7
and styrene. In the second case, the trisopropylbenzenesulfo-
nylhydrazone sodium salt 9 was thermolyzed and the volatile
diazo compound 7 trapped in the Ar matrix. The structure of
7 was assigned on the basis of the strong, characteristic
CdNdN absorption at 2042 cm-1 and the good agreement of
the experimental and computed IR spectra.
Photolysis of diazo compound 7 (using either λ > 300 nm
or λ > 534 nm) gave rise to new peaks assigned to A (•, Figure
1a). When the photolysis was carried out in an Ar matrix doped
with 0.3% oxygen, the same final photoproduct was observed.13
Furthermore, annealing the matrix containing the products at
30-35 K did not result in appreciable changes in the spectrum.
An ESR study performed under similar experimental conditions
failed to detect the presence of any triplet species. The agreement
(13) In the presence of 24% oxygen, different products were obtained.
The major product was benzocyclobutenone.
(14) Cycloocta-1,3,5-trien-7-yne can be excluded as the structure for
intermediate A because the alkyne stretch (scaled frequency ) 2083 cm-1
)
is computed to have a vanishingly small intensity (0.026 km mol-1).
Cycloocta-1,3,5-triene-7-yne is computed to be 15 kcal mol-1 higher in
energy than cycloalkyne 3.
(15) By comparison, the alkyne stretch in 3,3,7,7-tetramethylcycloheptyne
is 2180 cm-1. Krebs, A.; Kimling, H. Angew. Chem., Int. Ed. Engl. 1971,
10, 509-510.

The ElusiVe Benzocyclobutenylidene
J. Am. Chem. Soc., Vol. 123, No. 12, 2001 2873
Figure 3. (a) Computed IR spectrum of 7-methylenecyclohepta-3,5-
dien-1-yne (3) (BLYP/6-31G(d) scaled by 0.9945). (b) Difference IR
spectrum (Ar, 10 K) showing formation of alkyne 3 and disappearance
of diazobenzocyclobutene (7) upon irradiation (λ > 534 nm, 18 h).
Small amounts of benzocyclobutenone (O), an unknown impurity (?),
and CO2 are also formed under these conditions. (c) Computed IR
spectrum of benzocyclobutadiene (5) (BLYP/6-31G(d), scaled by
0.9945). (d) Difference IR spectrum (Ar, 10 K) showing formation of
benzocyclobutadiene (5) and disappearance of alkyne 3 upon irradiation
(λ > 261 nm, 48 h). The spectrum also shows the disappearance of
residual diazo compound 7 (D) and the growth of CO2.
position of the 13C label in cycloalkyne 3 is believed to be at
C2, although this assignment is not established unambiguously.
The small isotopic shift of 5 cm-1 observed experimentally at
608/603 cm-1 agrees better with the shift computed for [2-13C]-
3 (4 cm-1) than for [1-13C]-3 (1 cm-1) (Table 2). The thermal
barrier for the ring expansion of [13C]-2 to [2-13C]-3 (21.7 kcal/
mol) is computed to be 2.1 kcal/mol lower than the barrier for
the rearrangement of [13C]-2 to [1-13C]-3 (23.8 kcal/mol)
(Scheme 3).
A disadvantage of generating diazo compound 7 from 8 or 9
is that additional organic species are introduced into the matrix,
in the form of either styrene or volatile impurities. Thus, we
utilized azibenzocyclobutene (10),16 which can be obtained in
pure form, as an independent precursor to benzocyclobuten-
ylidene (1). The “clean” diazirine enabled us to monitor the
matrix photochemistry by UV/vis spectroscopy as well as IR
spectroscopy. Broadband photolysis (λ > 328 nm) of matrix-
isolated diazirine 10 produced diazo compound 7 and alkyne
3. The diazo compound was photosensitive under the irradiation
conditions and was converted to alkyne 3 upon continued
irradiation. Band-pass photolysis (λ ) 378 ( 10 nm) of 10
displayed greater selectivity. At short irradiation times, diazirine
10 isomerized cleanly to diazo compound 7 (Scheme 2; Figure
2). Conversion of diazo compound 7 to alkyne 3 occurred only
Figure 2. UV/vis spectra (Ar, 10 K): (a) azibenzocyclobutene (10)
before photolysis; (b) diazobenzocyclobutene (7) formed upon irradia-
tion (λ ) 378 ( 10 nm, 2.5 h) of diazirine 10; (c) alkyne 3 formed
upon subsequent irradiation (λ > 534 nm, 42.5 h) of diazo compound
7; (d) benzocyclobutadiene (5) formed upon subsequent irradiation (λ
> 261 nm, 82.9 h) of alkyne 3. Peaks labeled # and * represent distinct,
unidentified photoproducts.
corresponding to an isotopic shift of 36 cm-1 in excellent
agreement with the predicted values (Table 2). Thus, there is
little doubt that species A is the strained cycloalkyne 3. The
(16) Broadus, K. M.; Kass, S. R. J. Org. Chem. 2000, 65, 6566-6571.

2874 J. Am. Chem. Soc., Vol. 123, No. 12, 2001
Nicolaides et al.
Scheme 4
Scheme 5
upon prolonged irradiation. Although it is clear that diazirine
10 serves merely as a source of diazo compound 7 under band-
pass irradiation, we cannot exclude the possibility that 10 serves
as a direct precursor to alkyne 3 (or carbene 1) under other
irradiation conditions. As in the earlier experiments, long-
wavelength irradiation (λ > 534 nm) of diazo compound 7
afforded alkyne 3. With the electronic absorption spectrum of
alkyne 3 in hand, we investigated a variety of irradiation
conditions in an attempt to photoisomerize alkyne 3 to the
desired benzocyclobutenylidene (1). Although alkyne 3 did
indeed exhibit photochemistry (see below), efforts to observe
benzocyclobutenylidene (1) were unsuccessful.
Short-wavelength irradiation (λ > 261 nm) of alkyne 3
afforded benzocyclobutadiene (5) (Figure 3). Although the
efficiency of this photochemical conversion varied from experi-
ment to experiment,17 there is little doubt about the identity of
product 5. The IR spectrum displayed good agreement with both
the experimental spectrum previously reported by Chapman18
and the computed IR spectrum (Table S3). To obtain indepen-
dent evidence for the formation of 5, diazirine 10 was subjected
to flash vacuum pyrolysis (500 °C, unpacked quartz tube, 0.01
mmHg). A mixture containing phenylacetylene (6) and benzo-
cyclobutadiene dimer (11; 7:1 ratio) was isolated in low overall
(1), but this species may help explain data from both matrix
and condensed phase experiments.22 Under matrix isolation
conditions, vinylidene 2 might be expected to undergo reversible
photoequilibration with cycloalkyne 3, a process analogous to
the equilibration of cyclopentadienylidenecarbene and o-benzyne
(Scheme 6).23 The formation of benzocyclobutadiene (5) upon
short-wavelength photolysis of cycloalkyne 3 may be considered
to proceed by ring contraction to vinylidene 2. Again, the
intermediacy of carbene 1 remains uncertain. Vinylidene 2 may
undergo direct insertion into a C-H bond to afford benzocy-
clobutadiene (5) or it may undergo electrocyclization to carbene
1 followed by 1,2-hydrogen migration (Scheme 6). It is clear
that further experiments will be required in order to elucidate
the intimate details of these photochemical transformations.
In summary, photolysis of either diazo compound 7 or
diazirine 10 provided the cycloalkyne 7-methylenecyclohepta-
3,5-dien-1-yne (3). Alkyne 3 photochemically rearranged to
benzocyclobutadiene (5). Benzocyclobutadiene (5), along with
phenylacetylene (6), was formed upon flash vacuum pyrolysis
of diazirine 10. Vinylidene 2 is a possible intermediate that has
1
yield (ca. 10%) and characterized by H NMR spectroscopy
(Scheme 4). This result is consistent with previously reported
observations.18,19 Chapman’s experiments suggest that benzo-
cyclobutadiene (5) is stable in the gas phase at 230 °C,18 while
Kla¨rner’s experiments suggest that benzomethylenecyclopropene
(4), a higher-energy C8H6 isomer, rearranges to phenylacetylene
(6) at 400 °C.19
Our results establish that cycloalkyne 3 is the photoproduct
obtained upon irradiation of the precursors 7 and 10. However,
no evidence for the formation of carbene 1 was obtained. If
carbene 1 is indeed formed as the photoproduct from diazo
compound 7, it must either be formed with sufficient vibrational
energy during the decomposition of the excited-state precursor
to overcome the barrier to isomerization to alkyne 3 or it must
be photosensitive under the irradiation conditions (λ > 534 nm).
Alternatively, there are several plausible pathways by which
alkyne 3 could be formed from diazo compound 7 without
intervention of carbene 1.20 By analogy to the known photo-
chemical ring opening of benzocyclobutenes to o-quino-
dimethanes (Scheme 5),21 photochemical ring opening of diazo
compound 7, with either concerted or stepwise loss of nitrogen,
offers a route to alkyne 3 that circumvents carbene 1 (Scheme
5). The possible involvement of vinylidene 2 had not been
postulated in earlier investigations of benzocyclobutenylidene
(20) It has been amply demonstrated that the products of the photochemi-
cal decomposition of diazo compounds derive in substantial measure from
rearrangements of the nitrogen-containing precursors themselves. Tomioka,
H.; Kitagawa, H.; Izawa, Y. J. Org. Chem. 1979, 44, 3072-3075. See
also: C¸ elebi, S.; Leyva, S.; Modarelli, D. A.; Platz, M. S. J. Am. Chem.
Soc. 1993, 115, 8613-8620 and references therein.
(21) Chapman, O. L.; Chang, C. C.; Kolc, J.; Rosenquist, N. R.; Tomioka,
H. J. Am. Chem. Soc. 1975, 97, 6586-6588.
(22) One may ponder whether vinylidene 2 or cycloalkyne 3 play any
role in the condensed phase experiments.4-6 Vinylidene 2, if formed, may
be expected to undergo rapid thermal isomerization to benzocyclobuten-
1
ylidene 1sthe intermediate that has been presumed to be responsible for
the solution phase chemistry. It is also intriguing to note that vinylidene 2
is computed to be isoenergetic with triplet benzocyclobutenylidene 31. The
complex product mixture derived from solution chemistry does not appear
to be derived from cycloalkyne 3, and this result is perhaps not surprising,
given the large thermal barrier of 21 kcal mol-1 separating 1 and 3 as well
as the expected high reactivity of 1 and/or 2. The matrix chemistry is
dominated by photochemistry, while the solution chemistry, even if initiated
photochemically, is likely to be dominated by thermal chemistry.
(23) Armstrong, R. J.; Brown, R. F. C.; Eastwood, F. W.; Romyn, M.
E. Aust. J. Chem. 1979, 32, 1767-1774. Simon, J. G. G.; Muenzel, N.;
Schweig, A. Chem. Phys. Lett. 1990, 170, 187-192. Cioslowski, J.; Piskorz,
P.; Moncrieff, D. J. Am. Chem. Soc. 1998, 120, 1695-1700.
(17) The variability in conversion efficiency of alkyne 3 to benzocy-
clobutadiene (5) may involve several interrelated factors, including (i)
impurities in the matrix, (ii) internal filtering by strongly absorbing
impurities and/or photoproducts, (iii) matrix thickness, and (iv) intensity
of photolysis source.
(18) Chapman, O. L.; Chang, C. C.; Rosenquist, N. R. J. Am. Chem.
Soc. 1976, 98, 261-262.
(19) Kla¨rner, F. G.; Dogan, B. M. J.; Weider, R.; Ginsburg, D.; Vogel,
E. Angew. Chem., Int. Ed. Engl. 1986, 25, 346-348.

The ElusiVe Benzocyclobutenylidene
J. Am. Chem. Soc., Vol. 123, No. 12, 2001 2875
overnight, the solvent was removed on a rotary evaporator with a bath
at room temperature. The resulting crude hydrazone was purified with
gel permeation chromatography with CHCl3 to give a sticky liquid in
56% yield: 1H NMR (270 MHz, CDCl3) δ 2.39-2.46 (m, 1 H), 2.56-
2.60 (m, 1 H), 3.09-3.13 (m, 1 H), 3.85 (s, 2 H), 7.21-7.39 (m, 9 H);
IR (Ar, 10 K) 3068 vw, 2932 w, 1687 m, 1608 w, 1586 w, 1496 m,
Scheme 6
1461w, 1160 vw, 983 m, 935 m, 757 vs, 743 s, 703 m, 695 vs cm-1
.
Diazobenzocyclobutene (7). Sodium benzocyclobutenone 2,4,6-
triisopropylbenzenesulfonhydrazonate (9; 0.5 g) was placed in a 100
mL flask, which was connected to the matrix isolation chamber by a
glass tube. The entire system was evacuated to 10-6 mmHg; upon
cooling the CsI window to 30 K, the flask was heated very slowly to
135 °C. Heating the salt too quickly resulted in rapid decomposition.
Thermolysis of the salt afforded diazobenzocyclobutene (7), which was
co-deposited with argon over a period of 45 min. The matrix was cooled
to 10 K and photolysis was started: IR (Ar, 10 K) 3087 w, 3076 w,
3066 w, 2967 w, 2937 m, 2069 w, 2047 vs, 1599 m, 1593 m, 1464 m,
1443 s, 1434 s, 1410 w, 1309 w, 1296 w, 1289 w, 1163 m, 1137 m,
1000 w, 989 w, 740 s, 705 m, 701 m, 463 w cm-1 (Figure 3). A small
amount of benzocyclobutenone was observed, indicated by the presence
of IR absorptions at 1796, 1131, 954, and 764 cm-1. UV/vis (Ar, 10
K): λmax 225, 233, 248, 277, 283, 297, 302, 305, 309, 312, 316, 320
nm (Figure S2).
not been previously considered in the thermal and photochemical
transformations of nitrogenous precursors of benzocyclobuten-
ylidene (1). The elusive benzocyclobutenylidene (1), which is
computed to display a singlet ground state, was not observed
in any of our experiments.
The UV/vis spectrum of diazo compound 7 in argon at 10 K displays
subtle differences depending on whether the diazo compound is
generated by thermolysis of salt 9 or by photoisomerization of diazirine
10. These differences are attributed to impurities that co-deposit with
diazo compound 7 during the thermolysis of 9. Spectroscopic data for
diazo compound 7, as generated by photoisomerization of 10: UV/vis
(Ar, 10 K) λmax 226, 233, 252, 258, 268, 274, 280, 282, 302, 305, 309,
311, 315, 319, 351 (broad), 482 (broad) nm (Figure S2).
Methods Section
Computational Methods. Optimized geometries and harmonic
vibrational frequencies were obtained at the BLYP/6-31G(d) and
B3LYP/6-31G(d) levels of theory using the Gaussian 94 package of
programs.24 Vibrational frequencies were scaled by 0.9945 (BLYP) or
0.961 (B3LYP), and zero point vibrational energies were scaled by
1.0126 (BLYP) or 0.981 (B3LYP), as recommended by Scott and
Radom.25 Relative energies for the B3LYP/6-31G(d) structures are
reported at the G2(MP2,SVP) level.26 A full description of the
geometries and vibrational frequencies computed for these species is
provided in the Supporting Information.
Matrix Isolation Spectroscopy. Details of experimental procedures
for matrix isolation spectroscopy have been described previously.27 At
Tsu, photolyses were carried out using a Wacom 500-W xenon high-
pressure arc lamp. For broadband irradiation, Toshiba cutoff filters were
used (50% transmittance at the wavelength specified.) At Wisconsin,
photolyses were carried out using an ILC Technology 300 W xenon
arc lamp. Wavelength selection was achieved with cutoff filters (λ >
534 nm, Corning 3-67; λ > 328 nm, Schott WG 345; λ > 300 nm,
Schott WG 320; λ > 261 nm, Corning 0-53) or a Spectral Energy GM
252 monochromator (bandwidth 20 nm).
Azibenzocyclobutene (10). Diazirine 10 was prepared by the method
of Broadus and Kass.16 Sublimation of the diazirine at -29 °C into
argon provided matrix-isolated material: IR (Ar, 10 K) 3093 w, 3056
w, 2937 m, 2837 w, 1951 w, 1905 w, 1793 w, 1683 w, 1636 w, 1626
m, 1618 m, 1601 s, 1555 m, 1529 w, 1447 m, 1427 w, 1351 w, 1220
w, 1212 w, 1172 w, 1118 w, 1002 w, 783 w, 748 s, 712 m, 603 w, 505
w, 405 m cm-1 (Figure S1); UV/vis (Ar, 10 K) λmax 221, 223, 249,
260, 265, 267, 271, 274, 280, 324, 334, 340, 352, 358, 365, 371, 378
nm (Figure 2).
Flash Vacuum Pyrolysis of Azibenzocyclobutene (10). To one end
of an unpacked 60 cm quartz tube was affixed a 50 mL round-bottom
flask containing 80 mg (0.6 mmol) of diazirine 10. To the other end
were affixed a vacuum adapter and a cold trap. Diazirine 10 was cooled
to -78 °C, and the apparatus was evacuated to 0.01 mmHg while the
quartz tube was heated to 500 °C in a furnace. After cooling the cold
trap to 77 K, the temperature of diazirine 10 was raised to 0 °C. After
1 h, no 10 remained in the round-bottom flask and the system was
vented to N2. During the pyrolysis, the section of quartz tube near the
entrance to the furnace became heavily coated with charred material.
Approximately 5-10 mg of pyrolysate was recovered from the cold
trap and analyzed by 1H NMR spectroscopy (CDCl3). An intense singlet
at δ 3.07 revealed phenylacetylene (6)30 as the primary pyrolysis
product, and a series of alkenyl absorptions [δ 6.28 (d, J ) 9.9 Hz, 1
H), 6.12 (dd, J ) 9.9, 4.8 Hz, 1 H), 4.82 (d, J ) 6.9 Hz, 1 H), 4.43 (m,
1 H)] were consistent with those reported for the benzocyclobutadiene
dimer (11).31 According to the NMR integrations, the ratio of
phenylacetylene (6) to benzocyclobutadiene dimer (11) was about 7:1.
Preparation of N-(2-Phenylaziridyl)imine of Benzocyclobutenone
(8). To a solution of 118 mg (1 mmol) of benzocyclobutenone28 in 5
mL of benzene was added in one portion 132 mg (1 mmol) of 1-amino-
2-phenylaziridine,29 and the mixture was stirred at room temperature.
The progress of the reaction was followed by TLC. After stirring
(24) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M.; Johnson,
B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G. A.;
Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V.
G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; DeFrees, D. J.; Baker, J.; Stewart, J. J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. GAUSSIAN 94; Gaussian Inc.:
Pittsburgh, PA, 1995.
(25) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502-16513.
(26) Curtiss, L. A.; Redfern, P. C.; Smith, B. J.; Radom, L. J. Chem.
Phys. 1996, 104, 5148-5152.
(27) McMahon, R. J.; Chapman, O. L.; Hayes, R. A.; Hess, T. C.;
Krimmer, H.-P. J. Am. Chem. Soc. 1985, 107, 7597-7606. Tomioka, H.;
Ichikawa, N.; Komatsu, K. J. Am. Chem. Soc. 1992, 114, 8045-8053.
Seburg, R. A.; McMahon, R. J. J. Am. Chem. Soc. 1992, 114, 7183-7189.
(28) Abou-Teim, O.; Goodland, M. C.; McOmie, J. F. W. J. Chem. Soc.,
Perkin Trans. 1 1983, 2659-2662.
(29) Felix, D.; Mu¨ller, R. K.; Horn, U.; Joos, R.; Schreiber, J.;
Eschenmoser, A. HelV. Chim. Acta 1972, 55, 1276-1319. Felix, D.;
Wintner, C.; Eschenmoser, A. Org. Synth. 1976, 55, 52-56. Mu¨ller, R.
K.; Joos, R.; Felix, D.; Schreiber, J.; Wintner, C.; Eschenmoser, A. Org.
Synth. 1976, 55, 114-121.
Acknowledgment. We are indebted to Katherine M. Broadus
and Prof. Steven R. Kass (University of Minnesota) for
(30) Phenylacetylene (6): 1H NMR (CDCl3, 300 MHz) δ 3.07 (s, 1H),
7.28-7.42 (m, 3H), 7.46-7.62 (m, 2H), reported by Katritzky, A. R.; Wang,
J.; Karodia, N.; Li, J. Q. J. Org. Chem. 1997, 62, 4142-4147. IR (argon,
10 K): 3339, 1606, 1490, 1445, 1218, 1071, 1028, 914, 758 cm-1, reported
by Jørgensen, T.; Pedersen, C. T.; Flammang, R.; Wentrup, C. J. Chem.
Soc., Perkin Trans. 2 1997, 173-177.
(31) Trahanovsky, W. S.; Arvidson, K. B. J. Org. Chem. 1996, 61, 9528-
9533. Hada, C.; Roussel, J.-C.; Beccat, P.; Boulet, R.; Banciu, M. D. ReV.
Roum. Chim. 1997, 42, 777-785.

2876 J. Am. Chem. Soc., Vol. 123, No. 12, 2001
Nicolaides et al.
providing a sample of benzocyclobutenyl diaziridine and the
procedure for converting it to azibenzocyclobutene (10). We
thank Prof. Curt Wentrup (University of Queensland) and Prof.
William L. Karney (University of San Francisco) for insightful
discussions. We acknowledge the Computer Center of IMS for
a generous allocation of time and JSPS and NSF for an award
of a JSPS Fellowship (No. P96140) to A.N. This work was
supported by a Grant-in-Aid for Scientific Research from the
Ministry of Education, Science, Sports and Culture of Japan
and by the U.S. National Science Foundation (CHE-9800716).
Supporting Information Available: Experimental details
for the syntheses of 13C-labeled hydrazone 8, benzocyclobuten-
one triisopropylbenzenesulfonhydrazone 9, and diazirine 10.
Various experimental IR and UV/vis spectra of the matrix
isolation photochemistry; computed structures, energies, and IR
spectra for various C8H6 and C8H6N2 isomers. Archive entries
for 1-5. This material is available free of charge via the Internet
at http://pubs.acs.org.
JA0039482