Photoreactions of 3-Diazo-3H-benzofuran-2-one; Dimerization and Hydrolysis of Its Primary Photoproduct, A Quinonoid Cumulenone: A Study by Time-Resolved Optical and Infrared Spectroscopy

Article abstract of DOI:10.1021/ja0365476

Light-induced deazotization of 3-diazo-3H-benzofuran-2-one (1) in solution is accompanied by facile (CO)-O bond cleavage yielding 6-(oxoethenylidene)-2,4-cyclohexadien-1-one (3), which appears with a rise time of 28 ps. The expected Wolff-rearrangement product, 7-oxabicyclo[4.2.0]octa-1,3,5-trien-8-ylidenemethanone (4), is not formed. The efficient light-induced formation of the quinonoid cumulenone 3 opens the way to determine the reactivity of a cumulenone in solution. The reaction kinetics of 3 were monitored by nanosecond flash photolysis with optical (λ _max ≈ 460 nm) as well as Raman (1526 cm^-1) and IR detection (2050 cm^-1). Remarkably, the reactivity of 3 is that expected from its valence isomer, the cyclic carbene 3H-benzofuran-2-one-3-ylidene, 2. In aqueous solution, acid-catalyzed addition of water forms the lactone 3-hydroxy-3H-benzofuran-2-one (5). The reaction is initiated by protonation of the cumulenone on its β-carbon atom. In hexane, cumulenone 3 dimerizes to isoxindigo ((E)-[3,3′ ]bibenzofuranylidene-2,2′-dione, 7), coumestan (6H-benzofuro[3,2-c][1]benzopyran-6-one, 8), and a small amount of dibenzonaphthyrone ([1]benzopyrano[4,3-][1]benzopyran-5,11-dione, 9) at a nearly diffusion-controlled rate. Ab initio calculations (G3) are consistent with the observed data. Carbene 2 is predicted to have a singlet ground state, which undergoes very facile, strongly exothermic (irreversible) ring opening to the cumulenone 3. The calculated barrier to formation of 4 (Wolff-rearrangement) is prohibitive. DFT calculations indicate that protonation of 3 on the β-carbon is accompanied by cyclization to the protonated carbene 2H ⁺, and that dimerization of 3 to 7 and 9 takes place in a single step with negligible activation energy.

Full text of DOI:10.1021/ja0365476

Published on Web 09/27/2003
Photoreactions of 3-Diazo-3H-benzofuran-2-one; Dimerization
and Hydrolysis of Its Primary Photoproduct, A Quinonoid
Cumulenone: A Study by Time-Resolved Optical and Infrared
Spectroscopy
Yvonne Chiang,† Martin Gaplovsky,‡ A. Jerry Kresge,*,† King Hung Leung,§
Christian Ley,‡ Marek Mac,‡ Gabriele Persy,‡ David L. Phillips,§ Vladimir V. Popik,†,|
Christoph Ro¨dig,‡ Jakob Wirz,*,‡ and Yu Zhu†
Contribution from The Department of Chemistry, UniVersity of Toronto, Toronto, Ontario M5S
3H6, Canada, Departement Chemie, UniVersita¨t Basel, Klingelbergstrasse 80, CH-4056 Basel,
Switzerland, and the Department of Chemistry, UniVersity of Hong Kong, Pokfulham Road,
Hong Kong S.A.R., P.R China
Received June 6, 2003
Abstract: Light-induced deazotization of 3-diazo-3H-benzofuran-2-one (1) in solution is accompanied by
facile (CO)-O bond cleavage yielding 6-(oxoethenylidene)-2,4-cyclohexadien-1-one (3), which appears
with a rise time of 28 ps. The expected Wolff-rearrangement product, 7-oxabicyclo[4.2.0]octa-1,3,5-trien-
8-ylidenemethanone (4), is not formed. The efficient light-induced formation of the quinonoid cumulenone
3 opens the way to determine the reactivity of a cumulenone in solution. The reaction kinetics of 3 were
monitored by nanosecond flash photolysis with optical (λmax ≈ 460 nm) as well as Raman (1526 cm-1) and
IR detection (2050 cm-1). Remarkably, the reactivity of 3 is that expected from its valence isomer, the
cyclic carbene 3H-benzofuran-2-one-3-ylidene, 2. In aqueous solution, acid-catalyzed addition of water
forms the lactone 3-hydroxy-3H-benzofuran-2-one (5). The reaction is initiated by protonation of the
cumulenone on its â-carbon atom. In hexane, cumulenone 3 dimerizes to isoxindigo ((E)-[3,3′]bibenzo-
furanylidene-2,2′-dione, 7), coumestan (6H-benzofuro[3,2-c][1]benzopyran-6-one, 8), and a small amount
of dibenzonaphthyrone ([1]benzopyrano[4,3-][1]benzopyran-5,11-dione, 9) at a nearly diffusion-controlled
rate. Ab initio calculations (G3) are consistent with the observed data. Carbene 2 is predicted to have a
singlet ground state, which undergoes very facile, strongly exothermic (irreversible) ring opening to the
cumulenone 3. The calculated barrier to formation of 4 (Wolff-rearrangement) is prohibitive. DFT calculations
indicate that protonation of 3 on the â-carbon is accompanied by cyclization to the protonated carbene
2H+, and that dimerization of 3 to 7 and 9 takes place in a single step with negligible activation energy.
Introduction
ab initio calculations have predicted substantial (10-75 kJ
mol-1) barriers to Wolff rearrangement of the singlet carbonyl
In 1902, Wolff1 discovered the rearrangement of R-diazo-
carbonyl compounds, which at the time were thought to have a
cyclic 1,2,3-oxadiazole structure, to ketenes. In recent years,
several R-carbonylcarbenes have been identified as transient
intermediates in the photoinduced Wolff rearrangement,2-4 and
carbenes showing some correlation with the exothermicity of
ketene formation.5 In cases where the migratory aptitude of the
potentially migrating group is poor, the Wolff rearrangement
has been found to give way, in aqueous solution, to conjugate
addition of water, forming â-hydroxyenol products.6
The R-carbonylcarbene 2 formed by dediazotization of
3-diazo-3H-benzofuran-2-one (1) has two options for further
intramolecular reaction: (CO)-O ring cleavage to the quinonoid
cumulenone 3, and Wolff-rearrangement to the cyclic ketene 4
† University of Toronto. E-mail: AKresge@chem.utoronto.ca.
‡ University of Basel. E-mail: J.Wirz@unibas.ch.
§ University of Hong Kong. E-mail: Phillips@hkucc.hku.hk.
| Present address: Department of Chemistry, Bowling Green State
University, Bowling Green, OH. E-mail: VPopik@bgnet.bgsu.edu.
(1) Wolff, L. Justus Liebigs Ann. Chem. 1902, 325, 129-195.
(2) Bucher, G.; Scaiano, J. C.; Platz, M. S. Kinetics of Carbene Reactions in
Solution, Landolt-Bo¨rnstein, Group II, Vol. 18, Subvolume E2; Springer:
Germany, 1998; p 141. Toscano, J. P. In AdVances in Carbene Chemistry;
Brinker, U. H., Ed.; JAI Press: Greenwich, CT, 1998; Vol. 2, p 215.
(3) Toscano, J. P.; Platz, M. S.; Nikolaev, V.; Cao, Y.; Zimmt, M. B. J. Am.
Chem. Soc. 1996, 118, 3527-3528. Zhu, Z.; Bally, T.; Stracener, L. L.;
McMahon, R. J. J. Am. Chem. Soc. 1999, 121, 2863-2874. Wang, Y.;
Hadad, C. M.; Toscano, J. P. J. Am. Chem. Soc. 1999, 121, 2875-2882.
Wang, Y.; Toscano, J. P. J. Am. Chem. Soc. 2000, 122, 4512-4513. Wang,
Y.; Hadad, C. M.; Toscano, J. P. J. Am. Chem. Soc. 2002, 124, 1761-
1767.
(5) Scott, A. P.; Platz, M. S.; Radom, L. J. Am. Chem. Soc. 2001, 123, 6069-
6076.
(6) (a) Chiang, Y.; Kresge, A. J.; Popik, V. V.; Schepp, N. P. J. Am. Chem.
Soc. 1997, 119, 10 203-10 212. (b) Chiang, Y.; Jefferson, E. A.; Kresge,
A. J.; Popik, V. V.; Xie, R.-Q. J. Phys. Org. Chem. 1998, 11, 610-613.
(c) Chiang, Y.; Eustace, S. J.; Jefferson, E. A.; Kresge, A. J.; Popik, V.
V.; Xie, R.-Q. J. Phys. Org. Chem. 2000, 13, 461-467. (d) Chiang, Y.;
Kresge, A. J.; Schepp, N. P.; Xie, R.-Q. J. Org. Chem. 2000, 65, 1175-
1180. (e) Bakulev, V. A.; Chiang, Y.; Kresge, A. J.; Meng, Q.; Morzherin,
Y. Y.; Popik, V. V. J. Am. Chem. Soc. 2001, 123, 2681-2682. (f) Chiang,
Y.; Guo, H.-X.; Kresge, A. J.; Richard, J. P.; Toth, K. J. Am. Chem. Soc.
2003, 125, 187-194.
(4) Likhotvorik, I.; Zhu, Z.; Tae, E. L.; Tippmann, E.; Hill, B. E.; Platz, M. S.
J. Am. Chem. Soc. 2001, 123, 6061-6068.
9
12872
J. AM. CHEM. SOC. 2003, 125, 12872-12880
10.1021/ja0365476 CCC: $25.00 © 2003 American Chemical Society

Photoreactions of 3-Diazo-3H-benzofuran-2-one
A R T I C L E S
Scheme 1. Photoreactions of 1 in an Argon Matrix as Proposed
were collected on a Bruker IFS 66 v/S Fourier transform instrument.
An IR-band-pass filter (2257-1129 cm-1) was inserted between the
probe and the interferometer. The signals from a photovoltaic MCT
detector (Kolmar 100-1-B) were processed by a 20-MHz dc-coupled
preamplifier (Kolmar KA020-E6/MU), and further by a home-built
amplifier with a rise time of 20 ns at 200-fold amplification. This setup
had a rise time of 50 ns. To reduce noise, the signals were usually
passed through an electronic low-pass filter, which increased the rise
time to 330 ns. The signals were fed to a 50-Ohm load on the transient
recorder (Spectrum, PAD 1232, maximum sampling rate 40 MHz).
Time-Resolved Raman Spectroscopy. The experimental setup has
been described previously13 so only a brief description is given here.
Solutions of 1 were prepared in either pure acetonitrile or mixed water/
acetonitrile (50%/50 vol %) solvent. The fourth harmonic (266 nm)
and the first anti-Stokes hydrogen Raman shifted laser line (436 nm)
of the second harmonic from a Nd:YAG laser provided the pump and
probe excitation wavelengths, respectively. An optical delay of about
10 ns between the pump and probe pulses was used in the experiments.
The pump and probe beams were lightly focused onto a flowing liquid
stream of sample solution using a near collinear geometry. The Raman
scattering was acquired using a backscattering geometry and reflective
optics and imaged through a depolarizer mounted on the entrance slit
of a 0.5-meter spectrograph. The Raman light was then dispersed onto
a CCD detector and accumulated for about 300 to 600 s before being
read out to an interfaced PC computer and 10 to 20 read outs were
added together to obtain a resonance Raman spectrum. Pump only,
probe only, and pump-probe Raman spectra as well as a background
scan were acquired. The known Raman bands of the acetonitrile and
water/acetonitrile solvents were used to calibrate the Raman shifts of
the spectra. The solvent and precursor 1 Raman bands were removed
from the pump-probe transient resonance Raman spectrum by
subtracting a probe only Raman spectrum. The pump only spectrum
and a background scan were also subtracted from the pump-probe
spectrum so as to obtain the transient resonance Raman spectrum.
Laser Flash Photolysis (LFP). Measurements with acidic aqueous
solutions were made using an excimer laser nanosecond flash photolysis
system (Toronto) that provided a 20-ns, 100-mJ, 248-nm pulse. The
sample temperature was controlled at 25.0 ( 0.1 °C. Transient decays
conformed to the first-order rate law well, and observed first-order rate
constants were obtained by nonlinear least-squares fitting of an
exponential function. LFP of 1 in aprotic solvents was done with a
similar setup (Basel).
Product Analyses. Product compositions formed by irradiation of
1 in acidic aqueous solution were determined by HPLC using a Varian
Vista 5500 instrument with a NovoPak C18 reverse phase column and
methanol-water (70:30) as the eluent. Reaction solutions containing
the photolysis substrate at the same concentration as used for flash
photolysis (10-4 M) were subjected to three flashes from a microsecond
flash photolysis system.14 Control experiments showed that the pho-
tolysis products formed by the first flash did not undergo further
photoreactions in the subsequent two flashes. Products were identified
by comparing retention times and UV spectra with those of authentic
samples.
Calculations. All quantum chemical calculations were done with
the Gaussian 98 package of programs.15 Geometries of stationary points
(intermediates and transition states) were optimized using either B3LYP
density functional theory or MP2 perturbation theory with the 6-31G-
(d) basis set. Frequency calculations were done for all stationary points
and the connection of the transition states with the reactants 2-4 was
established by intrinsic reaction path (IRC) calculations. The standard
G3(MP2)16a composite procedure as well as its variant G3(MP2)//
B3LYP5,16b were used to calculate the energies of selected species.
by Chapman et al.7
(Scheme 1). In 1975, Chapman and co-workers7 studied the
photoreactions of 3-diazo-3H-benzofuran-2-one (1) under matrix
isolation conditions and identified both 3 and 4 as photochemical
products. The presumed carbene intermediate 2 was not
observed. Both 3 and 4 were stable at 12 K, but interconverted
upon further irradiation, with short wavelength (254 nm)
favoring 3, and long wavelengths (>350 nm) favoring 4. Soon
after, Voigt and Meier8 reported formation of 2-(2-hydroxyphen-
yl)-2-methoxyacetic acid methyl ester upon irradiation of 1 in
methanol solution. These authors considered several reaction
paths and concluded that the reaction proceeds either by direct
O-H insertion of the carbene 2 into methanol or via the cyclic
ketene 4.
We have studied the photoreaction of 1 in solution by time-
resolved optical, infrared and Raman spectroscopy. We identify
cumulenone 3 as the sole primary photoproduct of 1 in solution,
and report on its reaction kinetics in aqueous acid and in aprotic
solvents. A kinetic investigation of the thermal, acid-catalyzed
hydrolysis of 1 to 3-hydroxy-3H-benzofuran-2-one (5), and
further hydrolysis of 5 to 2-hydroxymandelic acid (6), was
reported in a separate paper.9
Experimental Section
Materials. 3-Diazo-3H-benzofuran-2-one (1) was synthesized from
isatin (2,3-indolinedione, Aldrich) by the route described10 with some
modification of the experimental details (cf. Supporting Information).11
3-Hydroxy-3H-benzofuran-2-one (5) and 2-hydroxymandelic acid (6)
were samples that had been prepared for another purpose.9 Solvents
were of spectroscopic grade, where available. Genetron 113 (1,1,2-
trichlorotrifluoroethane) was purchased from Fluka.
Step-Scan IR. The design of the step-scan apparatus was that
described in detail by Siebert and co-workers.12 Solutions of 1 in
mixtures of CD3CN and D2O were pumped through a sample cell (CaF2)
with either a peristaltic pump (Ismatec MS-2/12-160) or a graphite
cog-wheel pump (Ismatec Reglo-Z). The temperature in the sample
chamber of the instrument was about 30 °C. Solute concentrations were
adjusted to an absorbance of 0.5 (200 µm path length) at the excitation
wavelength. A Quantel Brillant W Q-switched Nd:YAG laser operated
at 10 Hz was used for excitation. The frequency-quadrupled (266 nm)
pulses of 4.3 ns duration (fwhm) were reduced to about 6 mJ per pulse
and the pulse diameter was widened to 8 mm at the sample. The spectra
(7) Chapman, O. L.; Chang, C.-C.; Kole, J.; Rosenquist, N. R.; Tomioka, H.
J. Am. Chem. Soc. 1975, 97, 6586-6588.
(8) Voigt, E.; Meier, H. Chem. Ber. 1977, 110, 2242-2248.
(9) Chiang, Y.; Kresge, A. J.; Meng, Q. Can. J. Chem. 2002, 80, 82-88.
(10) (a) Horspool, W. M.; Khandelwal, G. D. J. Chem. Soc. (C) 1971, 3328-
3331. (b) Huntress, E. H.; Hearon, W. A. J. Am. Chem. Soc. 1941, 63,
2762-2766.
(11) Supporting Information: see paragraph at the end of this paper regarding
availability.
(12) Uhmann, W.; Becker, A.; Taran, C.; Siebert, F. Appl. Spectrosc. 1991, 45,
390-397. Ro¨dig, C.; Siebert, F. Appl. Spectrosc. 1999, 53, 893-901.
(13) Zhu, P.; Ong, S. Y.; Chan, P. Y.; Leung, K. H.; Phillips, D. L. J. Am.
Chem. Soc. 2001, 123, 2645-2649.
(14) Chiang, Y.; Hojatti, M.; Keeffe, J. R.; Kresge, A. J.; Schepp, N. P.; Wirz,
J. J. Am. Chem. Soc. 1987, 109, 4000-4009.
9
J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003 12873

A R T I C L E S
Chiang et al.
Chart 1. Relative Calculated Energiesa in kJ mol-1 for the
Different Structures of Protonated Cumulenone 3
Figure 1. G3(MP2)//B3LYP zero-point energies (relative to syn-cumule-
none 3, Table 1) of the stationary points connecting 2-4. The structure
shown on top is that calculated for the singlet carbene 2.
a Zero-point energies calculated by DFT (B3LYP/6-31G(d)). Frequencies
were scaled by a factor of 0.9804. Absolute energies and structures are
given in Table S4 of the Supporting Information.11
Table 1. Calculated Zero-Point Energies (relative to cumulenone
3) of Reaction Intermediates 2-4 and the Transition States
Connecting Them
∆E/(kJ mol-1
)
shell species. On the other hand, the energies calculated for 3
and 4, and for the transition states connecting them, showed
satisfactory agreement between the four methods (Table 1). The
cumulenone moiety of 3 is kinked, as are the parent cumule-
nones,17 and the “syn”-isomer was predicted to be about 13 kJ
mol-1 more stable than its “anti”-counterpart.
stationary points
B3LYP/6-31G(d)
MP2/6-31G(d)
G3MP2
G3MP2/B3LYP
12
61.4
42.8
0
12.2
45.9
60.6b
155.6
12.7
74.8
162.0
0
16.7
43.0
73.3b
154.8
22.7
54.2
60.0
0
54.5
59.4
0
13.0
42.6
52.8
157.1
15.2
32
syn-3
anti-3
4
41.5
12 f syn-3a
syn-3 f 4a
syn-3 f anti-3a
Ring cleavage of 12 to the cumulenone 3 encounters a
negligible barrier. A stationary point with one imaginary
1
frequency lying just a few kJ mol-1 above 2 was located by
Vibrational corrections were obtained from the corresponding frequency
calculations, which were scaled by standard factors of 0.9804 for B3LYP,
0.9676 for MP2, 0.8929 for G3(MP2), and 0.96 for G3(MP2)//B3LYP.16
both the MP2 and the B3LYP models. However, after correction
for the zero-point energies, the energy of this point was slightly
a Transition state. b The transition state energy was below that of 2 after
1
1
below that of 2. On the other hand, a large barrier of about
zero-point energy correction.
156 kJ mol-1 separates cumulenone 3 from ketene 4 and no
other transition state could be located for the Wolff-rearrange-
ment, 2 f 4.
Results
1
Calculations. The energies of the stationary points for
intermediates 2-4 and the transition states connecting these
species are depicted in Figure 1. The structures were optimized
by the B3LYP/6-31G(d) method and the energies were recal-
culated by the G3(MP2)//B3LYP model. In a recent theoretical
study,5 this high-level theory was found to give reliable results
for singlet-triplet energy gaps of R-carbonylcarbenes and the
energy barriers to Wolff-rearrangement. Compounds 2-4 were
also calculated with the standard G3(MP2) model, which gave
very similar results (Table 1). Both of these methods predict a
singlet ground state for carbene 2 and a singlet-triplet energy
gap of about 5 kJ mol-1. The energy of the singlet carbene, 12,
is calculated to be 54 kJ mol-1 above that of 3. However, relative
energies predicted for 12 and 32 by the B3LYP and MP2
methods differed quite substantially. Not surprisingly, high-level
methods are required for an adequate treatment of these open-
The dimerization of cumulenone 3 to give either isoxindigo
(7) or dibenzonaphthyrone (8) was studied by DFT calculations
using the B3LYP functional and the 6-31G(d) basis set. Both
reactions are highly exothermic, ∆H (298 K) ) -573 and -618
kJ mol-1, respectively, and they also face a negligible barrier:
A stationary point with one imaginary frequency (for the
incipient bond between the â-sp carbon atoms on each cumu-
lenone) was located which, however, had a zero-point energy
9 kJ mol-1 lower than that of two separate molecules 3. This
may in part be attributed to the effect of basis set superposition.
The calculated free energy of this stationary point (298 K) lies
46 kJ mol-1 above that of two separate molecules due to the
entropic contributions. Attempts to locate a stationary minimum
for a diradical intermediate with a single bond connecting the
two cumulenones failed. Optimization of such trial structures
always led to one of the dimers 7 or 8.
Finally, gas-phase DFT calculations (B3LYP) were done for
the addition of a proton to the three conceivable protonation
sites of cumulenone 3. The results are shown in Chart 1. A
stable minimum for protonation at the â-carbon of the cumu-
lenone moiety, structure c, was found only when the calculation
was started with the proton situated in plane and endo to the
â-carbon atom. When the proton was initially placed exo to the
â-carbon, optimization directly proceeded to the cyclic structure
d, which corresponds to the protonated carbene 2, and is the
most stable of the carbocation structures a-d.
(15) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.; Stratmann,
R. E., Jr.; 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.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-
Gordon, M.; Replogle, E. S.; Pople, J. A. GAUSSIAN 98, Revision A.5;
Gaussian, Inc.: Pittsburgh, PA, 1998.
(16) (a) Curtiss, L. A.; Raghavachari, K.; Redfern, P. C.; Rassolov, V.; Pople,
J. A. J. Chem. Phys. 1998, 109, 7764-7776. (b) Baboul, A. G.; Curtiss, L.
A.; Redfern, P. C.; Raghavachari, K. J. Chem. Phys. 1999, 110, 7650-
7657.
(17) Scott, A. P.; Radom, L. J. Mol. Struct. 2000, 556, 253-261.
9
12874 J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003

Photoreactions of 3-Diazo-3H-benzofuran-2-one
A R T I C L E S
Table 2. Experimental Raman and IR Bands (cm-1) with the
Highest Intensities of Cumulenone 3, and Calculated Bands
(B3LYP/6-31G(d), Scaled with a Factor of 0.9613) for 2-4
compd
12
32
3
4
IR exp
2050
2150a
IR calc
1417 (112) 1725 (380) 2083 (1291)
2146 (967)
1598 (336) 1567 (85)
1838 (212)
2071
Raman exp
1618
1526
1402
1241
Raman calcd 3107 (210) 1725 (174) 3100 (267)
3102 (106) 3086 (123) 3080 (119)
3106 (192)
3096 (118)
3083 (141)
1590 (91)
3089 (94)
3071 (92)
1838 (138)
3103 (246) 2084 (242)
1621 (78)
1533.94 (411) 1418 (149)
1409 (81)
1238 (80)
1164 (92)
Calculated relative intensities are given in brackets. a From ref 7.
Scheme 2
Figure 2. Picosecond pump-probe spectra of 1 in hexane. The kinks
around 496 nm are artifacts due to the probe laser beam. The inset shows
the loading coefficients of the second eigenvector determined by component
analysis of 38 pump-probe spectra and a monoexponential fit function.
spectra.18 The second fraction (6 mg) contained mostly coumestan
(6H-benzofuro[3,2-c][1]benzopyran-6-one, 8),19 and a small
amount of dibenzonaphthyrone ([1]benzopyrano[4,3-c][1]-
benzopyran-5,11-dione, 9).18 The minor product 9 was identified
in the second fraction by its characteristic absorption and
fluorescence excitation and emission spectra,18 GC-MS, and the
H NMR peaks at δ ) 9.26 ppm (dd, J ) 1.5 and 8.4 Hz).18
Coumestan 8 was purified by sublimation at 130 °C in vacuo
and identified by comparison with the H NMR, UV/Vis and
mass spectral data given in the literature.19
Picosecond Pump-Probe Spectroscopy. Excitation of 1 in
hexane at 248 nm with a subpicosecond pump pulse produced
a structureless absorbance that decayed with a lifetime of 28 ps
leaving a broad Gaussian absorption band, λmax ) 460 nm
(Figure 2), which did not change further up to the maximum
delay of 1.8 ns. Chapman et al.7 had observed the 460-nm band
in their matrix work and attributed it to the quinonoid cumu-
lenone 3. This assignment will be corroborated below.
Frequencies and IR- and Raman-intensities of the strongest
vibrational transitions for the transient intermediates 3 and 4
and for the singlet and triplet states of carbene 2 calculated at
the B3LYP/6-31G(d) level of theory are given in Table 2.
Product Analysis, Aqueous Solutions. HPLC analysis of
photolyzed reaction mixtures obtained by flashing 10-3 and 10-5
M aqueous perchloric acid solutions of 3-diazo-3H-benzofuran-
2-one (1), performed one minute after the flash, showed the
presence of one major product, identified as 3-hydroxy-3H-
benzofuran-2-one (5), plus minor amounts of 2-hydroxymandelic
acid (6) and an additional unidentified substance. Subsequent
analyses performed at longer times after flashing indicated that
a further nonphotochemical reaction was converting 5 into 6 at
a velocity consistent with the known rate of acid-catalyzed
hydrolysis of this furanone to the substituted mandelic acid,
Scheme 2.9 This subsequent reaction serves to reinforce
identification of lactone 5 as the principal product formed by
photolysis of 1 in acidic aqueous solution.
Nanosecond Laser Flash Photolysis (LFP). The transient
absorption by 3, λmax ) 270 and 460 nm, appeared within the
duration of the laser pulse upon nanosecond LFP of 1 in
Genetron 113 (1,1,2-trichlorotrifluoroethane) or in acetonitrile
(Figure 3). It decayed by second-order kinetics and left persistent
absorption at 460 nm. The addition of water (10% by vol) had
little effect on the first half-life of 3 in acetonitrile, τ1/2 ≈ 0.7
ms, but the decay was accelerated, approaching first-order
kinetics and leaving no end absorbance, upon addition of
aqueous acid (upper inset of Figure 3). Because dimeric products
(7-9) are formed by irradiation of 1 in aprotic solvents, the
observed second-order rate law may be attributed to dimerization
Product analyses were also performed on photolysis reaction
mixtures formed by flashing 3-diazo-3H-benzofuran-2-one (1)
in less acidic aqueous solutions, all the way up to 5 × 10-4
M
sodium hydroxide. As these solutions became more basic, the
yields of 3-hydroxy-3H-benzofuranon-2-one (5) and its hy-
drolysis product 6 decreased, and other unidentified products
appeared. This increasing complexity discouraged us from
examining the photochemisty of 1 in these more basic solutions
in more detail.
2
of 3, -dc3/dt ) 2kc3 , where c3(t) ) A450(t)/[ꢀ450(3)d] is the
(unknown) concentration of transient 3 at time t after the laser
pulse and A450(t) is the transient absorbance measured at 450
nm. Integration of this rate law gives τ1/2 ) 1/[2kc3(0)]. The
optical path length d of the sample cell was 4.2 cm. Assuming
Product Analysis, Hexane. Irradiation of 1 (16 mg) in
hexane (50 mL) at 313 nm (medium-pressure mercury arc with
band-pass filter) gave three major products that were separated
by column chromatography using methylene chloride/hexane
1:1 as an eluent. The first, orange-red fraction was isoxindigo
((E)-[3,3′]bibenzofuranylidene-2,2′-dione, 7, 5 mg), which was
identified by its characteristic 1H NMR, UV/vis and mass
(18) Becker, H.-D.; Lingnert, H. J. Org. Chem. 1982, 47, 1095-1101.
(19) Kraus, G. A.; Zhang, N. J. Org. Chem. 2000, 65, 6544-5646. Govindachari,
T. R.; Nagarajan, K.; Parthasarathy, P. C. J. Chem. Soc. 1957, 548-551.
9
J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003 12875

A R T I C L E S
Chiang et al.
) (7.45 ( 0.08) × 108 M-1 s-1 and the isotope effect kH /kD
+
+
) 1.74 ( 0.03.
Rates of transient decay were also measured in acetic acid
buffer solutions. These measurements were made in series of
solutions of varying buffer concentration but constant buffer
ratio and constant ionic strength (0.10 M), and therefore constant
hydronium ion concentration. The data so obtained are sum-
marized in Table S2.11
Within each of these buffer solution series, observed first-
order rate constants increased linearly with increasing buffer
concentration. The data were therefore analyzed by least-squares
fitting of the simple buffer dilution expression shown in eq 1.
The buffer catalytic coefficients
kobs ) kuc + kbuff[buffer]
(1)
so obtained, kbuff, were separated into their general acid, kHA
and general base, kB, components with the aid of eq 2, in which
fA is the fraction of buffer present in the acidic form
,
Figure 3. Spectrographic trace obtained by LFP of 1 in wet acetonitrile
(10% water). The spectrum was recorded with a delay of 100 ns relative to
the excitation pulse at 308 nm. The kinetic traces in the insets show the
decay of transient absorbance at 450 nm in the same solvent mixture without
(lower) and with (upper) addition of 1 × 10-3 M HClO4.
kbuff ) kB + (kHA - kB)fA
(2)
The data (4 points) conformed to this expression well. Linear
regression gave kHA ) (3.08 ( 0.05) × 107 M-1 s-1 and kB )
(3.3 ( 4.4) × 105 M-1 s-1, indicating that the buffer catalysis
is wholly of the general acid kind (Figure S1).11
Step-Scan IR Measurements. The IR spectrum of 3-diazo-
3H-benzofuran-2-one (1) exhibits very strong bands at 2127,
2107, and 1767 cm-1, and weaker bands at 1467, 1401, and
1328 cm-1 in CD3CN. Some of these bands are slightly shifted
in Genetron 113 (1,1,2-trichlorotrifluoroethane): 2127, 2099,
and 1799 cm-1 (strong) and 1464, 1401, and 1328 cm-1 (weak).
Time-resolved IR difference spectra were generated by pulsed
laser irradiation of 1 at 266 nm in CD3CN, in wet CD3CN
(10% D2O), and in Genetron 113. These spectra exhibited strong
negative bands at the positions given above due to depletion of
1 by the laser flash. The negative bands were formed within
the time resolution (50 ns) of the instrument and remained
constant thereafter. In addition, a strong positive absorption
band at 2050 cm-1 was formed within 50 ns. The decay of
this band obeyed second-order kinetics, and the initial half-
lives were hardly dependent on solvent or water concentration,
τ1/2 ≈ 50 ( 10 µs. In wet, but not in dry acetonitrile, the growth
of a very weak absorption band at ca. 1816 cm-1 with kinetics
matching that of the decay at 2050 cm-1 was observed. In the
presence of acid (CD3CN/10% D2O, 1.0 × 10-4 M DClO4) the
decay of the 2050-cm-1 band was accelerated and approached
a first-order rate law, k ≈ 1.8 × 104 s-1. The simultaneous
growth of a band at 1818 cm-1 was now much more pronounced
(Figure 5).
The photoproduct absorbing at 1818 cm-1 was still present
after completion of the time-resolved experiment (30 min). It
slowly disappeared within 3 d (τ1/2 ≈ 1 d) at ambient
temperature in the dark. At the same time, new absorption bands
appeared at 1712 and 3540 cm-1. The rate of this slow reaction
is similar to that reported for the hydrolysis of 3-hydroxy-3H-
benzofuran-2-one (5) to 2-hydroxymandelic acid (6), Scheme
2, in weakly acidic aqueous solutions.9 Thus, 5 may be identified
as a major product formed by irradiation of 1 under the
conditions of the step-scan experiment. The transient absorption
at 2050 cm-1 is attributed to cumulenone 3.7 For comparison,
Figure 4. Observed decay rate constants of cumulenone 3 in H2O (O) and
D2O (4) solutions at 25 °C as a function of perchloric acid concentration
(data in Table S1).11
a molar absorbance coefficient ꢀ450(3) ≈ 5000 M-1 cm-1 20
, we
estimate a value of k ≈ 3 × 108 M-1 s-1 for the second-order
rate constant of dimerization.
LFP of 3-diazo-3H-benzofuran-2-one (1) in acidic aqueous
solution also produced intermediate 3, λmax ) 270 and 460 nm,
within the duration of the flash (20 ns). The transient decays
were in the microsecond time range and obeyed the first-order
rate law. Rates of decay were monitored at both λmax ) 270
and 460 nm and concordant results were obtained at the two
wavelengths. These rates were measured in both H2O and D2O
solutions of perchloric acid over the concentration range 0.001-
0.025 M. The ionic strength of these solutions was maintained
at 0.10 M by adding sodium perchlorate as required. The data
so obtained are summarized in Table S111 and are displayed in
Figure 4.
It may be seen that observed first-order rate constants
increase linearly with increasing acid concentration. Linear
+
regression produced the hydronium ion catalytic coefficient kH
(20) This estimate is based on the assumption that the persistent end absorbance
(lower inset of Figure 3) is due to the formation of 25% isoxindigo (7),
ꢀ450 ≈ 1.2 × 104 M-1 cm-1 18
. Similar values have been reported for related
1,2-quinone methides.
9
12876 J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003

Photoreactions of 3-Diazo-3H-benzofuran-2-one
A R T I C L E S
Figure 5. Temporal evolution of the IR difference spectra of 1 in CD3CN
with 10% (by vol.) D2O and 1.0 × 10-4 M DClO4 after excitation at 266
nm. The concentration of DClO4 refers to the total volume.
optical LFP of 1 (excitation at 308 nm) was also done in the
solvent mixture used for the step-scan measurements (CD3CN/
10% D2O, 1.0 × 10-4 M DClO4). This gave a clean first-order
decay of the 450-nm transient, k ) 5 × 103 s-1, about three
times lower than the rate constant determined by the step-scan
experiments. The difference may be attributed to a difference
in temperature (about 22 °C in LFP, 35 °C in step-scan) and to
a second-order contribution in the IR kinetics, where the
transient is produced in much higher concentration (path length
0.2 mm versus 1 cm in LFP). Indeed the first half-lives of the
decays in the absence and presence of 1.0 × 10-4 M DClO4
were comparable, τ1/2 ≈ 50 and 38 µs, respectively.
Figure 6. Transient resonance Raman spectra obtained for cumulenone 3
in acetonitrile (top) and water/acetonitrile (bottom) solvents. The numbers
label the Raman shifts (in cm-1) of the resonance Raman bands of 3. The
asterisks mark solvent/parent band subtraction artifacts and the daggers
represent ambient/stray light features.
Transient Raman Spectroscopy. Figure 6 presents transient
resonance Raman spectra obtained for 3 in pure acetonitrile and
mixed water/acetonitrile (50%/50 vol %) solvents at ∼10 ns
after 266-nm photoexcitation of 1. Similar spectra were obtained
in both solvents (indicating the same species is produced in both
solvents) with strong fundamental bands observed at 1400, 1526,
and 1618 cm-1 and weaker bands at 1150, 1241, and 2089 cm-1
in the water/acetonitrile solvent. The strong fundamentals at
1526 and 1618 cm-1 also exhibit noticeable intensity in their
overtones at 3033 and 3229 cm-1. This indicates these modes
are strongly resonantly enhanced and the resonance Raman
spectra in Figure 6 are associated with the transient absorption
band observed at ∼460 nm in the laser flash photolysis
experiments. The vibrational frequencies observed in the
transient resonance Raman spectra of Figure 6 are in good
agreement with those computed by the DFT calculations for
the cumulenone 3 (see Table 1). The DFT computed Raman
intensities are also in reasonable agreement with those observed
experimentally for most modes considering that the experimental
spectra are resonantly enhanced, whereas the computations are
an estimate for a nonresonant Raman spectrum. The differences
in resonance enhancement for the Raman modes could probably
account for the experimental 2089-cm-1 cumulenone carbonyl
band being noticeably smaller than that predicted from the DFT
calculations. The transient resonance Raman spectra in Figure
6 are consistent with results from the time-resolved IR experi-
ments and further confirm the assignment of the cumulenone 3
to the transient absorption spectra observed at 460 nm following
ultraviolet excitation of 1.
It is interesting to note that the cumulenone carbonyl Raman
band upshifts noticeably from 2071 cm-1 in acetonitrile solvent
to 2089 cm-1 in the water/acetonitrile solvent, whereas there
are much smaller changes observed for the other Raman bands.
This suggests the cumulenone carbonyl mode interacts more
strongly with the aqueous solvent (presumably through some
hydrogen-bonding interaction). Similar upshifts of vibrational
frequencies upon changing the solvent from acetonitrile to water
have been observed for the para-benzosemiquinone radical anion
and attributed to hydrogen bonding interactions.21 This stronger
solvent interaction for the cumulenone carbonyl mode may
be a consequence of this moiety being highly reactive and
susceptible to acid-catalyzed hydrolysis (Scheme 4).
Discussion
Chapman, in his study of the photolysis of 3-diazo-3H-
benzofuran-2-one (1) under low-temperature matrix isolation
conditions, was able to identify the cumulenone 3 as one of the
primary products formed, both by its characteristic long-
wavelength ortho-quinone methide-type absorption band at 460
nm and by its equally characteristic ketene carbonyl group-type
(21) Zhan, C.-G.; Chipman, D. M. J. Phys. Chem. A 1998, 102, 1230-1235.
9
J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003 12877

A R T I C L E S
Chiang et al.
Scheme 3. Conceivable (but dismissed) Mechanism for
IR band at 2040 cm-1.7 Association of the 2040-cm-1 IR band
and the electronic absorption at 460 nm with the same species
was established by conversion of 3 to the cyclic ketene 4 (ν˜ )
2150 cm-1) upon irradiation at long wavelengths. The frequen-
cies obtained here by DFT calculations (2083 cm-1 for 3 and
2146 cm-1 for 4, Table 1) are in good agreement with these
assignments.
Acid-Catalyzed Hydrolysis of 3
Primary Photoreaction. Pump-probe measurements (Figure
2) show formation of transient absorbance, λmax ) 460 nm, with
a rate constant of 3.5 × 1010 s-1 after subpicosecond excitation
of 3-diazo-3H-benzofuran-2-one (1). The 460-nm band agrees
in shape and position with that reported for cumulenone 3 by
Chapman et al., which serves to identify the present transient
species as 3 and shows this substance to be a primary
photoproduct of 1 in solution as well as in low-temperature
matrixes. The same transient, λmax ) 460 nm (Figure 3), is
observed by nanosecond LFP of 1 in various solvents. The
assignment of this transient is reinforced by the observation of
strong transient IR absorbance at ν˜ ) 2050 cm-1 upon LFP of
1 in acetonitrile solution (Figure 5). The other IR and the Raman
bands also agree well with the bands calculated for cumulenone
3 by DFT methods (Table 1).
Scheme 4. Acid-Catalyzed Hydrolysis of 3 in Aqueous Solution
The short-lived precursor of 3, τ ) 28 ps, which exhibits
diffuse absorption throughout the visible region (Figure 2), is
attributed to the excited singlet state of the diazo compound,
11*. If a singlet carbene, 12, is formed by elimination of nitrogen
from 11*, then its lifetime must be very short, τ < 28 ps. Indeed,
ion thus formed by water.23 In the present case, this would give
the vinyl cation 10, the reaction of which with water would
give the hydroxyketene 11, Scheme 3. This hydroxyketene could
then undergo intramolecular cyclization to enol 12, and keton-
ization of that would give the observed 3-hydroxy-3H-benzo-
furan-2-one product, 5.
1
G3 calculations indicate that ring opening of 2 to the cumu-
lenone 3 encounters virtually no barrier. More likely, vibrational
relaxation following the highly exothermic release of N2 from
11* affords 3 directly. On the other hand, Wolff-rearrangement
of 11* to the cyclic ketene 4 would require an activation energy
of about 155 kJ mol-1 (Figure 1). Yet, Chapman et al.7 observed
both 3 and 4 upon photolysis of 1 in an argon matrix at 12 K.
As these authors have shown that 3 is converted to 4 by
irradiation in the visible (Scheme 1), the formation of 4 was
presumably due to secondary irradiation of the primary product
3 during photolysis of 1.
Acyclic R-carboalkoxy carbenes have triplet ground states
with planar geometries and the corresponding singlet carbenes
normally adopt a near-orthogonal geometry between the car-
bonyl and the carbene plane, allowing for overlap between the
oxygen lone pair and the formally vacant p orbital at the carbene
center.5 Given the fact that geometrical constraints force carbene
2 to be planar (Figure 1), it is remarkable that the singlet state
is predicted to be the ground state of 2 at the G3 level of theory.
Presumably, the singlet is stabilized relative to the triplet state
due to the reduced bond angle at the carbene center.22
Decay Kinetics of 3 in Aqueous Solution. Our product study
indicates that cumulenone 3, once formed in aqueous solution,
undergoes hydration to 3-hydroxy-3H-benzofuran-2-one, 5. The
data displayed in Figure 4 show, moreover, that this hydration
is an acid-catalyzed process. Both of the functional groups of
cumulenone 3, the quinone methide group and the ketene group,
are known to undergo acid-catalyzed hydration reactions. Acid-
catalyzed hydration of quinone methides occurs by rapid pre-
equilibrium protonation of the quinone methide carbonyl oxygen
atom followed by rate-determining capture of the carbenium
+
/
The solvent isotope effect observed for this reaction, kH
+
kD ) 1.74, however, argues against this mechanism. The
preequilibrium proton transfer that initiates the process of
Scheme 3 converts a hydronium ion, whose O-H bonds are
relatively loose, into a water molecule, whose O-H bonds are
substantially tighter. This produces a tightening up of the
isotopically substituted reactant bonds, and that generates an
inverse (kH/kD < 1) isotope effect. There is both theoretical and
experimental evidence for such an inverse isotope effect.24 Of
special relevance to the present situation are the inverse isotope
effects recently found for the hydration of quinone methides,
e.g., kH /kD ) 0.42 for ortho-quinone methide25 and kH /kD
+
+
+
+
) 0.41 for para-quinone methide.26 The presently determined
isotope effect is, of course, not inverse, and that rules out the
mechanism of Scheme 3 initiated by reaction of the cumule-
none’s quinone methide functional group. Consistently, we failed
to detect any transient absorbance attributable to the enediol
12, which would be expected to show similar absorbance (λmax
≈ 280 nm) and ketonization kinetics (τ ≈ 1 ms at pH 3) as the
enol of mandelic acid.6a
This leaves an acid-catalyzed reaction of the cumulenone’s
ketene functional group. Acid-catalyzed hydration of ketenes
is known to take place by rate-determining protonation of the
(23) Chiang, Y.; Kresge, A. J.; Zhu, Y. Pure Appl. Chem. 2000, 72, 2299-
2308.
(24) Kresge, A. J.; More O’Ferrall, R. A.; Powell, M. F. In Isotopes in Organic
Chemistry; Buncel, E., Lee, C. C., Eds.; Elsevier: New York, 1987; Chapter
4.
(25) Chiang, Y.; Kresge, A. J.; Zhu, Y. J. Am. Chem. Soc. 2001, 123, 8089-
8094.
(22) Nicolaides, A.; Matsushita, T.; Tomioka, H. J. Org. Chem. 1999, 64, 3299-
(26) Chiang, Y.; Kresge, A. J.; Zhu, Y. J. Am. Chem. Soc. 2002, 124, 6349-
3305.
6356.
9
12878 J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003

Photoreactions of 3-Diazo-3H-benzofuran-2-one
A R T I C L E S
Scheme 5. Dimerization Reactions of 3 in Aprotic Solvents
ketene on its â-carbon atom.27 In the present case, this would
produce the acylium ion 13, Scheme 4, and intramolecular
capture of that by the quinone methide carbonyl oxygen atom
would give the carbenium ion 14. The DFT calculations
presented above for gas-phase protonation of 3, however,
indicate that structure 13 is not a metastable intermediate; exo
addition of a proton to 3 at the â-carbon proceeds directly to
14, which is identical with structure d of Chart 1, and is by far
the most stable isomer of protonated 3. Hydration of this
carbenium ion would then provide the observed 3-hydroxy-3H-
benzofuran-2-one product (5).
anism. The reaction mechanism of Scheme 3, on the other hand,
might give a semblance of general acid catalysis, in a process
where the buffer acid protonates cumulenone 3 on its carbonyl
oxygen atom in a rapid preequilibrium step, with the buffer base
then serving as a nucleophile to capture the carbenium ion 10
thus formed. In this event, the buffer base would react with the
unprotonated cumulenone as well, giving a process that would
appear as general base catalysis. However, no general base
catalysis was observed (eq 2, Figure S1, Table S2),11 and the
buffer data are therefore not consistent with the reaction
mechanism of Scheme 3.
Because proton transfer to the cumulenone â-carbon atom is
rate-determining in this reaction scheme, the hydronium ion
isotope effect would contain a primary component and conse-
quently be in the normal (kH/kD > 1) direction. Such isotope
effects, however, also contain an inverse secondary component
produced by tightening up of the “non-reacting” O-H bonds
of the hydronium ion as they are being converted into a water
molecule, which serves to reduce the magnitude of this isotope
effect.28 A further reduction can be expected on the basis of
The second-order rate constant for protonation of cumulenone
3 at the â-carbon, kH ) 7.5 × 108 M-1 s-1, is, to our
+
knowledge, the first of its kind. Its high value, 4-5 orders of
magnitude higher than those of simple ketenes,27 manifests the
higher intrinsic reactivity of cumulenones as compared to
ketenes. It should be mentioned, however, that a concerted
protonation-cyclization process forming the carbenium ion 14
in a single step may be driven partly by aromatization of the
quinone methide moiety.
the very fast nature of the present reaction (kH ) 7.5 × 108
+
Reactions of 3 in Aprotic Solvents. Irradiation of 3-diazo-
3H-benzofuran-2-one (1) in acetonitrile or hexane yields a
mixture of dimers (Scheme 5). Although isoxindigo (7) may
formally be regarded as a dimer of carbene 2, this is not the
case for dibenzonaphthyrone (8) and coumestan (9). Moreover,
the cumulenone 3 is formed within 50 ps of excitation of 1 in
solution, and nanosecond LFP showed that its decay obeys a
second-order rate law in the absence of acid (Figure 3). These
findings leave no doubt that cumulenone 3, rather than carbene
12, is the precursor of 7-9. The facile addition-cyclization-
(elimination) reactions of 3, k ≈ 3 × 108 M-1 s-1, involve
formation of a CdC double bond and two C-O bonds. No
intermediates of this reaction were detected by flash photolysis
or found as stationary points in the DFT calculations (B3LYP/
6-31G(d)). All attempts to optimize the structures of the primary
biradical intermediates shown in Scheme 5 led to one of the
dimers 7 or 8, even in the triplet state.
M-1 s-1), which will give rise to an unsymmetrical, reactant-
like transition state: primary isotope effects are known to vary
in magnitude with transition state symmetry, reaching maximum
values for symmetrical transition states and falling off to smaller
values for unsymmetrical, reactant-like or product-like transition
states.29 The observed isotope effect, kH /kD ) 1.74, is in fact
a very reasonable value for a reaction with the present velocity
occurring by the reaction mechanism of Scheme 4.
+
+
Reactions that occur by rate-determining proton transfer to
the substrate, such as that in Scheme 4, should also show general
acid catalysis, and such catalysis has in fact been observed for
the hydration of ketenes in acidic buffer solutions.27b Its
occurrence here in acetic acid buffers (see Figure S1)11
consequently provides further support for this reaction mech-
(27) (a) Tidwell, T. T. Ketenes; Wiley-Interscience: New York, 1995; pp 585-
587. (b) Andraos, J.; Kresge, A. J. J. Photochem. Photobiol. A: Chem.
1991, 57, 165-173. (c) Andraos, J.; Kresge, A. J.; Schepp, N. P. Can. J.
Chem. 1995, 73, 539-543.
The high, carbene-like reactivity of cumulenone 3, both in
aqueous acid (carbon protonation, Scheme 4) and in aprotic
solvents (dimerization, Scheme 5) is remarkable. Several
previous publications have surmised equilibration of quinonoid
(28) Kresge, A. J.; Sagatys, D. S.; Chen, H. L. J. Am. Chem. Soc. 1977, 99,
7228-7233.
(29) Kresge, A. J. In Isotope Effects on Enzyme-Catalyzed Reactions; Cleland,
W. W., O’Leary, M. H., Northup, D. B., Eds.; University Park Press:
Baltimore, 1977; pp 37-63.
9
J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003 12879

A R T I C L E S
Chiang et al.
ketenes with a cyclic carbene isomer in an attempt to explain
the formation of reaction products expected from a carbene,
but Mosandl and Wentrup pointed out that the formation of such
products should not be taken as unequivocal evidence for
reaction via a carbene intermediate.30 In the present case,
simultaneously. Hydrolysis of 3 in aqueous acid yielding the
lactone 5 is initiated by rate-determining protonation of the
cumulenone at its â-carbon atom. The rate constant of proton
addition, kH ≈ 7.5 × 10 M-1 s-1, is 4-5 orders of magnitude
larger than those observed for related ketenes, revealing the high
intrinsic reactivity of the C3O moiety. Thus, the chemical
reactivity of cumulenone 3 mimics that expected from the singlet
carbene, 12. This finding is fully consistent with ab initio
calculations. The short lifetime of 12 precludes any bimolecular
trapping reactions, and the high energy of the singlet carbene
relative to 3 excludes its participation as a pre-equilibrating
intermediate in the reactions of 3.
8
+
1
reaction of 3 by preequilibration with 2 is quite unlikely in
1
view of the 54-kJ mol-1 energy difference between 3 and 2
calculated at the G3 level of theory. The stage at which CO is
eliminated in the formation of coumestan (8) has not been
established. Neither of the isolated dimers 7 and 9 eliminated
CO upon further irradiation.
Conclusion
Acknowledgment. This work is part of Projects Nos. 20-
68087.02 and 2160-064525.01 of the Swiss National Science
Foundation. Financial support by Ciba Specialty Chemicals and
the Natural Sciences and Engineering Research Council of
Canada is gratefully acknowledged.
Cumulenone 3 is the sole primary photoproduct of 3-diazo-
2-oxo-2,3-dihydrobenzofuran (1) in solution. High-level ab initio
calculations indeed predict that ring-opening of an initially
formed singlet carbene 2 is strongly favored over Wolff-
rearrangement to form the cyclic ketene 4. Isoxindigo (7) and
coumestan (8) are the major products isolated after irradiation
of 1 in aprotic solvents. They are formed by a remarkably facile
addition-cyclization reaction of 3, k ≈ 3 × 108 M-1 s-1, in
which a CdC bond and two C-O bonds appear to be formed
Supporting Information Available: Synthesis of 1, rate data
for the decay of 3 in aqueous acid and buffer solutions, and
details of quantum chemical calculations. This material is
available free of charge via the Internet at http://pubs.acs.org.
(30) Mosandl, T.; Wentrup, C. J. Org. Chem. 1993, 58, 747-749 and references
therein.
JA0365476
9
12880 J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003