Photochemistry of p-benzoquinone diazide carboxylic acids: Formation of 2,4-didehydrophenols

Article abstract of DOI:10.1021/ja971731a

The photochemistry of p-benzoquinone diazide carboxylic acids (7) was studied using matrix isolation spectroscopy, product analysis, and high-level ab initio molecular orbital theory. The general photochemical pathway observed includes primary carbene formation, followed by secondary photodecarboxylation to yield derivatives of 2,4-didehydrophenol 9. CCSD(T) calculations on the parent 2,4-didehydrophenol (9a) lead to an infrared spectrum which is in excellent agreement with the experiment alone. Furthermore; calculations predict 9a to be characterized by a distorted benzene ring with the hydroxy group pointing toward the radical center in ortho position. The heat of formation of 9a is predicted to be 85 kcal/mol. Its formation from 1-oxo-2,5-cyclohexadien-4-ylidene-2-carboxylic acid (8a) by decarboxylation is exothermic by 30 kcal/mol where strong H-bonding in 8a can be considered to facilitate the formation of 9a.

Full text of DOI:10.1021/ja971731a

10660
J. Am. Chem. Soc. 1997, 119, 10660-10672
Photochemistry of p-Benzoquinone Diazide Carboxylic Acids:
Formation of 2,4-Didehydrophenols
Wolfram Sander,*,† Go1tz Bucher,† Holger Wandel,† Elfi Kraka,*,‡
Dieter Cremer,‡ and William S. Sheldrick§
Contribution from the Lehrstuhl fu¨r Organische Chemie der Ruhr-UniVersita¨t,
D-44780 Bochum, Germany, and Department of Theoretical Chemistry, UniVersity of Go¨teborg,
Kemigården 3, S-41296 Go¨teborg, Sweden
ReceiVed May 27, 1997X
Abstract: The photochemistry of p-benzoquinone diazide carboxylic acids (7) was studied using matrix isolation
spectroscopy, product analysis, and high-level ab initio molecular orbital theory. The general photochemical pathway
observed includes primary carbene formation, followed by secondary photodecarboxylation to yield derivatives of
2,4-didehydrophenol 9. CCSD(T) calculations on the parent 2,4-didehydrophenol (9a) lead to an infrared spectrum
which is in excellent agreement with the experimental one. Furthermore, calculations predict 9a to be characterized
by a distorted benzene ring with the hydroxy group pointing toward the radical center in ortho position. The heat
of formation of 9a is predicted to be 85 kcal/mol. Its formation from 1-oxo-2,5-cyclohexadien-4-ylidene-2-carboxylic
acid (8a) by decarboxylation is exothermic by 30 kcal/mol where strong H-bonding in 8a can be considered to
facilitate the formation of 9a.
Introduction
Marquardt, Sander, and Kraka reported on the first matrix
isolation and spectroscopic characterization of m-benzyne (1).7
This aryne is more stable than p-benzyne and, accordingly, can
be better handled in experiment. Because of this, 1 and its
derivatives are of considerable interest for synthetic chemists,
spectroscopists, structural chemists, and theoreticians.
In this contribution, we present results of our studies dealing
with m-arynes. In particular, we concentrate on substituted
m-arynes bearing hydroxy groups, i.e., 2,4-didehydrophenols.
These reactive intermediates were generated by sequential
photodediazotation and photodecarboxylation of appropriate
p-benzoquinone diazide carboxylic acids and investigated by
matrix isolation spectroscopy in connection with product studies.
m-Benzyne can be depicted as either an aromatic σ-biradical
1 or as bicyclo[3.1.0]hexatriene (2). The literature provides
evidence for both structures: Washburn et al. investigated the
dehydrobromination of dibromobicyclo[3.1.0]hexene (3) and
The chemistry of aromatic σ-biradicals (arynes) has been the
subject of considerable interest throughout this century. Re-
search on o-arynes started early with the pioneering work by
Wittig,1 and o-benzyne can now be considered a well-character-
ized molecule, which has been investigated extensively both
with matrix isolation spectroscopy and with high-level ab initio
calculations.2 Recently, the focus of many research groups has
shifted toward p-arynes and related diradicals, as these com-
pounds are considered to be the critical intermediates in the
DNA-cleaving reaction of enediyne antibiotics/cytostatica.3,4 The
natural enediyne cytostatica are optimized to fulfill under certain
conditions a specific biologic task and not necessarily that of a
useful anticancer drug. Therefore, ongoing research focuses on
replacing the natural “warhead” 3 of the enediyne cytostatica,
namely p-benzyne, by other highly reactive diradicals which
are more suitable for an anticancer drug.4
The heats of formation of o-, m-, and p-benzyne have been
determined by Wenthold and Squires5 from collision-induced
dissociation threshold energies to 106.6 ( 3.0, 122.0 ( 3.1,
and 137.3 ( 3.3 kcal/mol, respectively, and are in excellent
agreement with high level ab initio calculations.6,26 Recently
(10) Bucher, G.; Sander, W.; Kraka, E.; Cremer, D. Angew. Chem. 1992,
104, 1225-1228; Angew. Chem., Int. Ed. Engl. 1992, 31, 1230-1233.
(11) Bucher, G.; Sander, W. Chem. Ber. 1992, 125, 1851-1859.
(12) Sander, W. Angew. Chem. 1990, 102, 362-72; Angew. Chem., Int.
Ed. Engl. 1990, 29, 344-354.
(13) Bucher, G. Studien auf der C6H4O-Hyperfla¨che. Dissertation, TU
Braunschweig, 1992.
† Lehrstuhl fu¨r Organische Chemie, Ruhr-Universita¨t Bochum.
‡ University of Go¨teborg.
(14) Sander, W.; Bucher, G.; Reichel, F.; Cremer, D. J. Am. Chem. Soc.
1991, 113, 5311-5322.
§ Lehrstuhl fu¨r Analytische Chemie, Ruhr-Universita¨t Bochum.
X Abstract published in AdVance ACS Abstracts, October 1, 1997.
(1) Wittig, G. Naturwissenschaften 1942, 696-703.
(2) Radziszewski, J. G.; Hess, B. A. J.; Zahradnik, R. J. Am. Chem. Soc.
1992, 114, 52-57.
(15) Bucher, G.; Sander, W. J. Org. Chem. 1992, 57, 1346-1351.
(16) Sander, W.; Bucher, G.; Komnick, P.; Morawietz, J.; Bubenitschek,
P.; Jones, P. G.; Chrapkowski, A. Chem. Ber. 1993, 126, 2101-2109.
(17) Komnick, P.; Sander, W. Liebigs Ann. 1996, 114, 7-9.
(18) Sues, O. Liebigs Ann. Chem. 1944, 556, 65-84.
(19) Tsuda, M.; Oikawa, S. J. Photopolym. Sci. Technol. 1989, 2, 325-
339.
(3) Nicolaou, K. C.; Dai, W. M. Angew. Chem. 1991, 103, 1453-81;
Angew. Chem., Int. Ed. Engl. 1991, 30, 1387-1415.
(4) Chen, P. Angew. Chem. 1996, 108, 1584-1586; Angew. Chem., Int.
Ed. Engl. 1996, 35, 1478-1480.
(5) Wenthold, P. G.; Squires, R. R. J. Am. Chem. Soc. 1994, 116, 6401-
6412.
(6) Cramer, C. J.; Nash, J. J.; Squires, R. R., Chem. Phys. Lett., in press.
(7) Marquardt, R.; Sander, W.; Kraka, E. Angew. Chem. 1996, 35, 746-
748; Angew. Chem., Int. Ed. Engl. 1996, 35, 746-748.
(8) (a) Washburn, W. N.; Zahler, R. J. Am. Chem. Soc. 1976, 98, 7827-
7828. (b) Washburn, W. N.; Zahler, R. J. Am. Chem. Soc. 1976, 98, 7828-
7830. (c) Washburn, W. N. J. Am. Chem. Soc. 1975, 97, 1615-1616.
(9) Billups, W. E.; Buynak, J. D.; Butler, D. J. Org. Chem. 1979, 44,
4218-4219.
(20) Reiser, A.; Shih, H.-Y.; Yeh, T.-F.; Huang, J.-P. Angew. Chem. 1996,
108, 2611-2622; Angew. Chem., Int. Ed. Engl. 1996, 35, 2428-2440.
(21) Becke, A. J. Chem. Phys. 1993, 98, 5648-5652.
(22) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J.
Phys. Chem. 1993, 98, 11623.
(23) Hariharan, P. C.; Pople, J. A. Chem. Phys. Lett. 1972, 66, 217-
219.
(24) Raghavachari, K.; Trucks, G. W.; Pople, J. A.; Head-Gordon, M.
Chem. Phys. Lett. 1989, 157, 479-483.
(25) Cox, J. D.; Pilcher, G. Thermochemistry of Organic; Organometallic
Compounds; Academic Press: New York, 1970.
(26) Kraka, E.; Cremer, D. Chem. Phys. Lett. 1993, 216, 333-340.
S0002-7863(97)01731-9 CCC: $14.00 © 1997 American Chemical Society

Formation of 2,4-Didehydrophenols
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10661
Scheme 1
obtained trapping evidence for 2,8 while Billups et al., on the
other hand, obtained trapping products from the m-dehy-
dronaphthalene 4, although a synthesis of the corresponding
bicyclo[3.1.0]hexatriene system had been attempted.9
Thus, the question whether structure 1 or 2 represents the
true formula of m-benzyne remained unanswered until recently,
when the spectroscopic characterization of m-benzyne by
Marquardt, Sander, and Kraka revealed that this molecule is in
fact best described as the aromatic σ-biradical 1. Aryne 1 was
generated in argon matrix at 10 K from two independent
precursors: both flash vacuum pyrolysis of isophthaloyl di-
acetyldiperoxide 5 and photolysis of matrix-isolated cyclophane
6 yielded, among other products, benzyne 1. The structural
assignment could be confirmed by correlation of the experi-
mental infrared spectrum with a spectrum obtained by CCSD-
(T) ab initio calculations.
diazides 7, this procedure might provide a viable route to study
the solution chemistry of m-benzynes.
The goal of this study is to gain further insights into the
mechanism of the photodecarboxylation of carbene 8a, to
investigate the electronic structure of 9a by high-level ab initio
theory, to characterize some of its substituted derivatives 9
spectroscopically, and to investigate its solution reactivity.
Results
4-Diazo-1-oxocyclohexa-2,5-diene-2-carboxylic Acid (7a).
The photochemistry of 7a has already been briefly described
in a preliminary communication.10 Matrix isolation of 7a could
be achieved by subliming the diazo compound in an electrically
heated oven at 136 °C and 10-5 mbar. Even under optimized
conditions, however, the sublimation of the highly polar 7a (and,
in analogy, all other quinone diazide carboxylic acids investi-
gated in the present study) remained difficult, and only modest
amounts could be brought into the gas phase.
In initial experiments, 7a was irradiated using broad-band
visible light (λ > 475 nm). Under these conditions, 7a
decomposes rapidly and a new product (A) with a characteristic
IR absorption at 3612 cm-1, indicating a hydroxy functionality,
appears together with an intense CO2 band. This new com-
pound, however, does not show a prominent band in the
carbonyl region. Under more selective conditions (monochro-
matic irradiation, λ ) 435 nm), another product (B) is formed,
which still exhibits a band characteristic of a carboxylic acid
(ν ) 1758 cm-1). This compound decarboxylates photochemi-
cally with orange or red light (complete conversion in less than
1 min with λ > 475 nm, slow conversion with λ > 780 nm)
and yields the same hydroxy compound (A) found in the initial
experiments (Figure 1, Scheme 1)). In separate experiments
using oxygen-doped argon matrices (typically 1% O2 in Ar),
product B could be identified as triplet carbene 8a by oxygen
trapping.11 The formation of carbonyl O-oxides by the thermal
reaction of carbenes with 3O2 is highly characteristic and
frequently used to identify matrix-isolated carbenes.12
According to the calculations, m-benzyne (1) is best described
as a planar distorted hexagon, in which the intra-annular C1-
C3 distance is somewhat shortened, giving evidence for
considerable interaction between the radical centers. In a recent
communication, we also described the matrix isolation of 2,4-
didehydrophenol (9a), starting from p-benzoquinone diazide
carboxylic acid (7a) via carbene 8a. Its geometry, as calculated
on the GVB 6-31G(d) level, is very similar to that of the parent
m-benzyne and the bicyclic isomer 10a was excluded.10 In the
minimum energy conformation, the hydroxy group points toward
the ortho radical center, indicating a stabilizing interaction. Thus,
the substituted m-benzyne 9 represents an interesting model
system for the study of the interaction of aromatic singlet
biradicals with substituents. The corresponding quinone diaz-
ides with a variety of substituents are readily accessible
precursors of substituted dehydrophenols 9. In addition, since
the benzynes 9 are formed on visible light irradiation of quinone
When a sample of 8a, matrix isolated in argon doped with
1% O2, is briefly warmed to 40 K, the matrix becomes intensely
yellow. The assignment of the primary oxidation product as
carbonyl O-oxide 13a can be justified by the intense color (λmax
) 440 nm, Ar, 10 K) and the strong IR absorption at 1082
cm-1, characteristic of the O-O stretching vibration of a

10662 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Sander et al.
Figure 1. IR difference spectra showing the photochemistry of 7a in argon at 10 K. (a) Bottom: bands of quinone diazide 7a disappearing on
irradiation (λ ) 435 nm, 20 min). Top: bands of carbene 8a and small amounts of 9a and CO2 appearing. (b) Bottom: bands of 8a disappearing
on irradiation (λ > 475 nm, 30 min). Top: bands of didehydrophenol 9a and CO2 appearing. (c) Bottom: bands of 9a disappearing on irradiation
(λ > 260 nm, 1 h). Top: bands appearing, assigned to a ketene of unknown structure.
Figure 2. IR difference spectra showing the oxygenation of carbene 8a in a 2% O2-doped argon matrix and subsequent photochemistry. (a)
Bottom: bands of 8a disappearing on annealing at 35-45 K. Top: bands of carbonyl oxide 13a appearing. (b) Bottom: bands of 13a disappearing
on irradiation (λ > 515 nm, 1 min). Top: bands of dioxirane 14a appearing.
quinone O-oxide.11 When 18O2 was used in these experiments
the absorption was red-shifted to 1034 cm-1. Irradiation of
quinone O-oxide 13a with orange light (λ > 515 nm) results in
the rapid conversion to dioxirane 14a (Figure 2). The dioxirane
14a, in turn, is also photolabile and converted to a mixture of
lactones 15a and 16a13 by irradiation with blue light (λ > 400
nm).
two structures remain to be discussed: 2,4-didehydrophenol (9a)
and bicyclo[3.1.0]hexatriene-2-ol (10a).
A discrimination between these two possibilities was only
possible by ab initio molecular orbital theory, see below. These
studies gave unequivocal evidence that 2,4-didehydrophenol (9a)
was indeed the intermediate observed by us (Figure 3).
Upon UV irradiation (λ > 260 nm), biradical 9a proved to
be photolabile. However, it was not possible to characterize
the product formed, as a large peak at 2138 cm-1, very close to
the absorption of matrix-isolated carbon monoxide, is the only
strong feature in the IR spectrum. (Further weak bands: 1419,
1047, and 963 cm-1.) Thus, the photoproduct of 9a is likely
to be a ketene, but its structure remains unclear.
Product A is formed from carbene 8a by photodecarboxyla-
tion. Simultaneously the hydrogen shift from the carboxy group
to the keto functionality of carbene 8a occurs. Additional
evidence for this rearrangement comes from the investigation
of the partially deuterated quinone diazide 8a-d1. In this case,
the hydroxy stretching frequency in product A is shifted to 2666
cm-1
.
The photochemistry of 7a in an argon matrix was also studied
by UV-vis detection. Carbene 8a shows a weak and very broad
absorption starting from ca. 400 nm tailing out into the red, a
medium-intensity, relatively sharp maximum at 390 nm (this is
The infrared spectrum of product A offers no indication of
the presence of a ketene, alkyne, or allene functionality, which
would be the product of a ring opening reaction. Given this,

Formation of 2,4-Didehydrophenols
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10663
Figure 3. IR spectrum of 2,4-didehydrophenol 9a. (a) Spectrum of 9a calculated with CCSD(T). The band positions are scaled with 0.96. (b)
Difference spectrum showing the λ > 260 nm photochemistry of 9a. Bands pointing up are assigned to 9a. x: unknown impurities.
teristic of diazonium salts, which indicates that 7a must be at
least partially protonated in 1,1,1-trifluoroethanol.
6-Alkyl-Substituted Derivatives 7. To characterize some
derivatives of m-benzyne 9a and to improve the solubility of
the corresponding diazo compounds in common solvents such
as MTHF, we synthesized a series of p-benzoquinone diazide
carboxylic acids (7) with alkyl substituents in 6-position. While
the solubility of the 6-methyl-substituted diazo compound 7b
in MTHF was still unsatisfactory, detailed product studies were
possible with the 6-isobutyl-substituted diazo compound 7c.
Figure 4. UV-vis spectrum of didehydropenol 9a in argon at 10 K.
a feature characteristic of 4-oxocyclohexadienylidenes),14,15 and
strong bands at 356, 344, 316, and 308 nm. The UV-vis
spectrum of biradical 9a is similar: bands are found with λmax
) 356 (sh), 344, 318, 304 (sh), 256, and 238 nm (Figure 4),
but there is no band discernible at λ > 400 nm, although the
compound undergoes a photochemical transformation when
irradiated with these wavelengths.
Thus, when 7c is irradiated in MTHF solution at room
temperature or glass at 77 K using a mercury high-pressure arc
lamp, the main product is 3-isobutylsalicylic acid (which is the
product formed via hydrogen abstraction by the carbene), while
308 nm pulsed laser irradiation of 7c yields a significant amount
(up to 30%, depending on the reaction conditions) of 2-iso-
butylphenol, a product derived from m-aryne 9c. Irradiation
in ethanol at room temperature or 77 K exclusively yields
trapping products of carbene 8c (OH insertion and H abstraction)
and no decarboxylation.
As with the parent diazo compound 7a, the alkyl-substituted
derivatives 7b,c were also studied by matrix isolation spectros-
copy. Because the melting points of 7b,c are lower than that
of parent diazo compound 7a, we had initially hoped for a higher
volatility of 7b,c as compared to 7a. However, concurrently
with the melting point, also the temperature of decomposition
was lowered upon alkyl substitution, so that the amounts of
Product studies with quinone diazide 7a were difficult because
of the extremely low solubility of 7a in common solvents. One
of the few solvents which allowed for convenient concentrations
of 7a is 1,1,1-trifluoroethanol. In this case, the main product
is 5-(2,2,2-trifluoroethoxy)salicylic acid, derived from the
insertion of the intermediate singlet carbene into the O-H bond.
Interestingly, significant amounts (ca. 10%) of 5-fluorosalicylic
acid are also formed. Insertion into carbon-fluorine bonds is
a very uncommon reaction, even for highly reactive singlet
carbenes. In a control experiment, 4-oxocyclohexadienylidene
(12a) was generated by irradiation of quinone diazide 11a in
1,1,1-trifluoroethanol. While the yield of 4-fluorophenol is less
than 1% in this case, the photolysis of 4-hydroxybenzene-
diazonium chloride under otherwise identical conditions results
in a 1:1 mixture of 4-fluorophenol and 4-trifluoroethoxyphenol.
Obviously, reactivity toward C-F bonds is a property charac-

10664 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Sander et al.
Table 1. Selected Interatomic Distances (Å) in 7a16 and 7c
(Standard Deviation).
7a
7c
C(1)-C(2)
C(2)-C(3)
C(3)-C(4)
C(4)-C(5)
C(5)-C(6)
C(6)-C(1)
C(4)-O(4)
C(3)-C(31)
C(31)-O(31)
C(31)-O(32)
C(1)-N(1)
N(1)-N(2)
O(4)-O(32)
1.396(4)
1.362(4)
1.450(3)
1.426(4)
1.356(4)
1.416(4)
1.285(3)
1.496(4)
1.208(4)
1.332(4)
1.368(4)
1.102(4)
2.475
1.390(7)
1.397(7)
1.431(7)
1.455(7)
1.357(6)
1.405(7)
1.293(6)
1.495(7)
1.229(6)
1.321(6)
1.391(7)
1.109(6)
2.442
Table 2. Selected Bond Angles (deg) in 7a16 and 7c (Standard
Deviation)
7a
7c
Figure 5. Structure of 7c as determined by X-ray crystallography.
C(1)-C(2)-C(3)
C(2)-C(3)-C(4)
C(3)-C(4)-C(5)
C(4)-C(5)-C(6)
C(5)-C(6)-C(1)
C(6)-C(1)-C(2)
O(4)-C(4)-C(3)
C(4)-C(3)-C(31)
C(3)-C(31)-O(31)
O(31)-C(31)-O(32)
C(6)-C1)-N(1)
C(1)-N(1)-N(2)
118.9(3)
120.7(2)
117.7(2)
121.7(2)
118.3(2)
122.6(2)
120.9(2)
120.5(2)
123.3(2)
121.8(2)
119.4(2)
178.8(3)
116.4(5)
121.2(2)
119.8(5)
117.7(5)
121.0(5)
123.9(5)
119.7(5)
119.9(5)
121.7(5)
122.0(5)
118.6(5)
177.8(6)
diazo compounds 7b,c that could be sublimed were even smaller
than that of 7a.
The photochemical behavior of diazo compounds 7b,c in Ar
matrix is analogous to that of 7a. Monochromatic photolysis
(λ ) 435 nm) of the diazo compounds yields the corresponding
carbene carboxylic acids 8b,c (Scheme 1), which could be
photochemically decarboxylated with orange or red light. The
resulting m-arynes 9b,c show prominent νO-H bands at 3612
(9b) or 3613 cm-1 (9c). In the case of 9b-d1, νO-D is found at
ν ) 2666 cm-1sat exactly the same position as in the parent
compound 9a-d1. The reactivity of carbenes 8b,c toward
molecular oxygen is analogous to the reactivity of 8a; the
corresponding quinone O-oxides 13b,c are obtained by annealing
argon matrices containing the carbenes and 1-2% oxygen at
ca. 40 K. Irradiation (λ > 515 nm) of the intensely yellow
13b (λmax(Ar, 10 K) ) 464 nm) produces dioxirane 14b, which
in turn is photochemically converted into a mixture of the
lactones 15b and 16b.13 Infrared data for these compounds are
listed in the Experimental Section.
Diazo compound 7c was characterized by X-ray crystal-
lography. The structure clearly reveals the presence of an
intramolecular hydrogen bridge, as in the case of parent
compound 7a.16 The distance between the keto oxygen atom
(O4 in Figure 5) and the hydroxy oxygen atom in the carboxy
functionality (O32) is only 2.442 Å and thus even shorter than
the corresponding OO distance in 7a (2.475 Å). The ketocar-
bonyl bond length C4-O4 is unusually long (1.293 Å),
indicating a bond order of significantly less than 2. The NtN
bond length is in a range typical for p-benzoquinone diazides
(7c, 1.109 Å; 7a, 1.102 Å; p-benzoquinone diazide 11a, 1.115
Å).16 Diazo compound 7c has an essentially planar structure,
but unlike in the case of 7a, which can crystallize in a very
dense layered structure (and is thus very hard to dissolve in
common solvents, like MTHF),16 the isobutyl substituents
somewhat reduce the density of crystal packing, so that 7c is
sufficiently soluble in MTHF. Tables 1 and 2 give an overview
of selected geometric parameters in 7c.
frequency observed indicates a bond order of ca. 1.5), which
indicates that the contribution of the 4-naphthyl-1-naphthoxy
biradical mesomeric structure must be significant. In the UV-
vis regime, the characteristic feature of carbene 8d is a weak
and broad absorption between ca. 425 and 600 nm, in addition
to two strong maxima at 342 and 358 nm, similar to the UV-
vis spectrum of the corresponding carbene without carboxy
function in 2-position.15 If 8d is converted to 9d, the broad
band in the visible portion of the spectrum disappears, while
the remainder of the spectrum remains virtually unchanged.
Upon warming an oxygen-doped Ar matrix to 40 K, carbene
3
8d reacts readily with O2 to yield the corresponding deep red
carbonyl O-oxide 13d (characteristic broad absorption between
400 and 600 nm), which confirms our assignment for 8d.
Biradical 9d proved to be photolabile. After photolysis (λ
) 248 nm), several infrared bands appeared that can be assigned
to a mixture of indeneketene and indenylidene (Scheme 3). The
same set of bands were observed upon photolysis of 1,4-
naphthoquinone diazide 11b in an Ar matrix,17 and therefore,
it is tempting to postulate the rearrangement of 9d to carbene
12b as the primary step of the short-wavelength UV photolysis.
Carbene 12b is not stable under these conditions17 and thus not
observed.
Photolysis (350-450 nm) of 7d in ethanol or EPA at room
temperature resulted in the formation of 1-hydroxy-2-naphthoic
acid and 4-ethoxy-1-hydroxy-2-naphthoic acid (GC analysis of
the methyl esters obtained by treatment of the photoproducts
with diazomethane). Thus, as with the alkyl-substituted deriva-
tives 7b,c, decarboxylation is slow compared to the trapping of
the carbene and only products derived from carbene 8dsand
not from diradical 9dsare formed.
1,4-Naphthoquinone Diazide 2-Carboxylic Acid (7d).
When 7d, matrix-isolated in Ar at 10 K, is irradiated with
monochromatic light (λ ) 435 ( 5 nm), carbene 8d (νCdO,carboxy
) 1753 cm-1, νCdO,keto ) 1429 cm-1) was obtained in
quantitative yield, which in turn could be converted into m-aryne
9d by yellow light irradiation (λ > 475 nm) (Figure 6, Scheme
2). Biradical 9d shows a characteristic νO-H at 3608 cm-1 (νO-D
) 2663 cm-1), thus slightly shifted relative to dehydrophenol
9a. Interesting, as with carbene 8a, is the low frequency of the
CdO stretch vibration of the cyclohexadienone chromophor (the
1,2-Naphthoquinone 1-Diazide 3-Carboxylic Acid (18).
When o-quinone diazides are photolyzed, they usually undergo
a rapid Wolff rearrangement to yield indene ketenes. This
reaction is called the Su¨s reaction18 and is basis of the chemistry
of o-naphthoquinone diazide photoresists.19,20 We reasoned that

Formation of 2,4-Didehydrophenols
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10665
Figure 6. IR difference spectra showing the photochemistry of diazo compound 7d in argon at 10 K. (a) Bottom: bands of quinone diazide 7d
disappearing on irradiation (λ ) 435 nm, 20 min). Top: bands of carbene 8d and small amounts of 9d and CO2 appearing. (b) Bottom: bands of
8d disappearing on irradiation (λ > 475 nm, 30 min). Top: bands of didehydrophenol 9d and CO2 appearing.
Scheme 2
K. Fortunately, 18 was much easier to sublime than, for
instance, the corresponding p-derivative 7d, which is likely due
to its reduced dipole moment. However, we did not obtain
evidence for photodecarboxylation yielding biradical 17. The
exclusive product observed was tentatively assigned to indene
ketene 19 (typical IR bands: 3570, 2142, 1758 cm-1, Figure
7), and carbon dioxide was not formed.
While the absence of a detectable νO-H makes is likely that
an intramolecular hydrogen bond is indeed present in quinone
diazide carboxylic acid 18 (and thus also in the resulting singlet
ketocarbene), it is evident that this hydrogen bond is not strong
enough to sufficiently slow the Wolff rearrangement. Even
under high-energy pulsed laser (KrF Excimer Laser, 248 nm,
250 mJ/pulse) conditions, no photodecarboxylation was ob-
served.
Scheme 3
Quantum Chemical Calculations
For 1, 2, 9a, and 10a, density functional theory (DFT) and
coupled cluster (CC) theory was used to get reasonable
geometries and to clarify the question whether the biradical or
bicyclic forms are more stable in these cases. At the UB3LYP/
6-31G(d,p)21-23 level of theory (unrestricted DFT with a hybrid
functional) a stationary point was found for each of the four
molecules. However, at the more reliable CCSD(T)/6-31G(d,p)
level of theory (CC calculations with all single and double
excitations and a perturbative treatment of the triple excita-
tions24), no minimum was found for the bicyclic forms 2 and
10a and only the σ-biradicals 1 and 9a turned out to be stable.
The CCSD(T) geometrical parameters are shown in Figure 8,
while the corresponding UB3LYP/6-31G(d,p) values are listed
in Table 3. (The numbering of the ring atoms in 1 and phenol
is adjusted to that in 9a.)
Hydroxyl substitution leads to little changes in the geometry
of 1. The ring keeps the same distorted form of a hexagon,
and the bond lengths of 1 and 9a are almost identical (Figure
8). The OH group takes a position in the ring plane with the H
atom directed toward the radical center because steric repulsion
is minimized in this way. Rotation of the OH group is hindered
by a barrier of 4.6 kcal/mol (UB3LYP/6-31G(d,p) geometry
optimizations) which corresponds to a CCOH rotational angle
of 90°. The confirmation with the H(O) atom directed toward
C6H is 3.4 kcal/mol higher in energy than the equilibrium form
the Wolff rearrangement might be retarded if the carbonyl group
of the intermediary singlet ketocarbene was involved in an
intramolecular hydrogen bond. In the hope to obtain evidence
for photodecarboxylation also in 2-oxocyclohexadienylidene-
3-carboxylic acids, we synthesized the o-naphthoquinone diazide
18 and investigated its photochemistry in an Ar matrix at 10

10666 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Sander et al.
Figure 7. IR difference spectrum showing the photochemistry of 18 in argon at 10 K. Bottom: bands of 18 disappearing on irradiation (λ > 420
nm, 50 min). Top: bands appearing, tentatively assigned to indeneketene 19.
Table 3. Adiabatic Frequencies for Phenol, 2,4-Dehydrophenol
(9a), and m-Benzyne (1)a
phenol
9a
r, R, τ
A. Stretching Modes
1
r, R, τ
ω
ω
r, R, τ
ω
C1C2
C2C3
C3C4
C4C5
C5C6
C6C1
C1O7
O7H
1.399 1350
1.395 1367
1.395 1371
1.397 1361
1.393 1379
1.399 1353
1.368 1219
0.966 3815
1.385 1349
1.376 1303
1.377 1255
1.374 1379
1.400 1350
1.405 1333
1.370 1212
0.967 3809
1.375 1381
1.375 1218
1.375 1218
1.375 1381
1.403 1335
1.403 1335
1.086 3193
Figure 8. CCSD(T)/6-31G(d,p) geometry of m-benzyne (1) and 2,4-
dehydrophenol (9a). Distances in angstroms and angles in degrees.
B. Bending Modes
C6C1C2
C1C2C3
C2C3C4
C3C4C5
C4C5C6
C5C6C1
C2C1O7
C1O7H
C2C3H
C6C5H
C1C6H
120.0
119.8
120.5
119.3
120.8
119.6
122.6
109.0
119.3
119.2
118.9
945 118.2
920 131.5
921 103.9
932 133.0
922 117.5
925 115.7
707 122.7
1274 108.1
1333 127.9
1335 121.5
1299 121.0
874 117.3
826
404
314
404
826
704
1279
583 134.0
448 101.9
498 134.0
832 117.3
789 115.5
670 121.1
of 9a. At the same level of theory, the rotational barrier of the
OH group in phenol is calculated to be 4.1 kcal/mol.
The stability of 9a can be determined with the help of the
formal reactions i-iii:
1269
1 + phenol f 9a + benzene
9a + H2 f phenol
(i)
1271 129.0
1280 121.6
1282 122.3
1270
1279
1308
(ii)
C. Torsional Modes
9a + CH3CH3 f phenol + CH2dCH2
(iii)
C6C1C2C3
C1C2C3C4
C2C3C4C5
C3C4C5C6
C4C5C6C1
C5C6C1C2
C2C1O7H
0
0
0
0
0
0
0
554
581
593
592
584
557
373
0
0
0
0
0
0
0
526
491
489
540
544
524
383
0
0
0
0
0
0
573
531
531
573
561
561
For this purpose, B3LYP/6-31G(d,p) reaction energies have
been corrected by differences in the appropriate zero point
energies (ZPE) and checked by comparing with CCSD(T)
reaction energies. By combining the reaction enthalpies ob-
tained in this way with known reaction enthalpies of benzene
(19.8),25 phenol (-23.0),25 m-benzyne (122.8),26 ethane (-20.2),25
and ethene (12.4 kcal/mol),25 one gets from reaction i a ∆Hf°
value of 83.4 kcal/mol for 9a, from reaction ii 86.0, and from
reaction iii 85.5 kcal/mol, respectively. In view of the relatively
large uncertainty of ∆Hf° for 1, we suggest a value of 85 ( 1
kcal/mol.
a Distances in Å, angles in deg, frequencies in cm-1. For the
numbering of atoms, see Figure 8.
explains the somewhat lower stabilization effect of the OH group
in the case of dehydrophenol 9a.
Substituting 1 by an OH group as in 9a is 3.4 kcal/mol less
stabilizing than substituting benzene by an OH group (total
stabilization energy: 14.4 kcal/mol at B3LYP/6-31G(d,p).
While both m-benzynes 1 and 9a are stabilized by through-
space and through-bond interaction as well as reduced steric
interactions between the CH groups, π-delocalization and
through-bond interactions are opposing each other, which

Formation of 2,4-Didehydrophenols
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10667
Table 4. Calculated and Measured Vibrational Frequencies ω and Infrared Intensities I of m-Benzyne (1) and 2,4-Dehydrophenol (9A)
m-Benzyne (1)
2,4-Dehydrophenol (9a)
CCSD(T) B3LYP
CCSD(T)
B3LYP
exptl
exptl
no.
sym
ω
I
ω
I
ω
I
no.
aym
ω
I
ω
I
ω
I
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
b1
a1
b1
a2
b2
b1
b1
a2
a1
b1
b2
a1
a1
b2
b2
b2
b2
a1
b2
a1
a1
b2
a1
a1
367
386
479
497
545
743
818
824
849
4.6
5.7
0
387
233
583
505
567
768
837
845
856
4.1
1.3
10.6
0
74.7
46.3
5.5
0
0.2
0.8
27.9
0.7
0.9
0.1
0.0
0.0
0.0
16.1
8.0
3.9
21.6
8.7
7.3
3.4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
a′′
a′′
a′
a′′
a′
a′′
a′
a′′
a′
a′′
a′′
a′′
a′
a′
a′
a′
a′
a′
a′
a′
a′
a′
a′
251
320
332
356
436
445
516
530
735
769
817
882
889
1008
1134
1195
1275
1317
1336
1426
1475
1589
1703
3232
3254
3295
3880
0.0
133.3
7.7
5.3
1.3
2.4
65.4
0.3
58.0
37.4
7.7
259
358
300
389
420
504
541
632
745
796
835
929
910
1015
1122
1178
1266
1310
1320
1418
1448
1564
1678
3180
3200
3228
3817
0.0
126.7
20.8
4.0
2.0
4.2
32.4
2.4
63.2
36.1
2.6
0
104.1
56.4
10.7
0
0.8
0.0
29.1
0.5
0.4
0.2
0.5
547
751
824
100
45
20
518.8
s
s
641.2
694.7
918
975
972
982
m
936
25
1052
1118
1152
1284
1336
1420
1454
1544
1706
3210
3251
3256
3289
1052
1107
1141
1268
1320
1410
1432
1531
1687
3159
3195
3200
3225
0.7
0.4
35.6
11.4
2.5
116.9
97.7
42.0
4.8
64.9
15.4
118.5
15.4
11.2
8.3
16.6
16.9
4.8
134.3
84.5
28.6
6.1
42.5
15.1
108.2
27.5
10.4
6.1
877.0
971.0
w
w
vw
m
m
m
w
m
w
s
1128.6
1157.6
1209.1
1254.9
1290.1
1368.2
1429.0
1516.3
0.4
0.0
14.9
18.6
3.9
22.8
12.6
7.4
1402
1486
15
15
4.9
3037
5
a′
a′
a′
a′
3.9
51.7
2.7
51.4
3612.0
m
a
Frequencies ω in cm-1, intensities I in km/mol. In the case of the experimental data relative intensities are given. The experimental spectra
were measured only above 500 cm-1
.
In Figure 3, the infrared spectrum of 9a, calculated at the
CCSD(T)/6-31G(d,p) level of theory, is compared to the
corresponding measured infrared spectrum. In Table 4, CCSD-
(T) frequency and intensity values are given for all vibrational
modes of 1 and 9a. A characterization of these modes including
phenol as a suitable reference molecule is given in Table 5,
where the adiabatic mode analysis of Konkoli and Cremer27
has been used to carry out correlation and characterization of
calculated normal modes. Since a CCSD(T) calculation of the
vibrational spectrum of phenol is not feasible, the analysis is
based on the B3LYP/6-31G(d,p) data. Table 4 reveals that this
is justified considering the similarity of the CCSD(T) and
B3LYP description of the vibrational modes.
For both 1 and 9a there is convincing agreement between
measured and calculated (CCSD(T) and B3LYP) infrared
spectra, while HF, MP2, and GVB lead to infrared spectra that
strongly differ from the CCSD(T) and the experimental spectra.
Reliable calculations on m-benzynes are problematic because
of their multireference character. Accordingly, HF and low-
order MP perturbation theory fail in this case while a two-
configuration approach, e.g., GVB leads to a reasonable
description of the m-benzyne wave function but, however, lacks
needed dynamic electron correlation. Both CCSD(T) and DFT
cover high-order dynamic electron correlation effects and, by
this, also some of the multireference effects. Because of this
and because of the reduced biradical character of m-benzynes
(20-40% according to different ways of calculating the biradical
character),26 both CCSD(T) and B3LYP lead to reliable descrip-
tions of energetic, structural, and spectroscopic properties of
m-benzynes.
5). Since this motion implies opposing rehybridization effects
at the radical centers, it is sensitive to a correct description of
electron correlation. For example, GVB seriously overestimates
the harmonic frequency and underestimates the intensity of this
mode. The other infrared bands measured correspond to b1-
symmetrical H out-of-plane bendings (CH wagging motions no.
6 and no. 7 at 751 and 824 cm-1), a b2-symmetrical CC
stretching (and partially HCC bending) motion at 936 cm-1 (no.
11), and a1- and b2-symmetrical modes at 1402 (no. 18) and
1486 cm-1 (no. 19) with character similar to that of mode no.
11 (Table 5).
Hence, 1 can best be recognized by the very strong ring
deformation band at 547 cm-1 and the half as intense CH
wagging band at 751 cm-1. These two bands can be used to
distinguish benzene, o-benzene, 1, and p-benzyne by their
infrared spectra. For benzene, the CH wagging band is very
strong (671 cm-1),28,29 while the ring deformation mode has
zero intensity because of symmetry reasons. For o-benzyne,
the intensities for the two vibrational modes are similar as found
for 1, but the ring deformation mode is found below 500 at
472 cm-1 30,2
CCSD(T) and DFT calculations for p-benzyne
.
suggest that the ring deformation band is only of medium and
the CH wagging motion again of very strong intensity.31
OH substitution results in splitting of the ring deformation
mode at 547 cm-1 into two less intense bands at 519 and 641
cm-1. The CH wagging band in 9a (at 695 cm-1) is again of
lower intensity than the ring deformation modes. The calcula-
tions suggest that the OH torsion band (no. 2, not measured),
the COH bending band (no. 16, 1158 cm-1) and the CC
stretching/HCC bending band corresponding to band no. 19 of
The most intense infrared band of 1 (at 547 cm-1) corresponds
to a b2-symmetrical ring deformation mode that causes CCC
bending at C1 and C4 and (out of phase) at C2 and C5 (Table
(28) Varsanyi, G. Vibrational Spectra of Benzene DeriVatiVes; Academic
Press: New York; London, 1969.
(29) Brown, K. G.; Person, W. B. Spectrochim. Acta 1978, 34A, 117-
122.
(30) Simon, J. G. G .; Mu¨nzel, N.; Schweig, A. Chem. Phys. Lett. 1990,
170, 187-192.
(31) Bochum and Go¨teborg et al. To be published.
(27) (a) Konkoli, Z; Cremer, D. Int. J. Quantum Chem., submitted. (b)
Konkoli, Z.; Larsson, A.; Cremer, D. Int. J. Quantum Chem., submitted.

10668 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Sander et al.
Table 5. Characterization and Correlation of the Normal Modes of Phenol, 2,4-Dehydrophenol (9a), and m-Benzyne (1)
phenol, Cs symmetry
9a, Cs symmetry
1, C2V symmetry
mode sym
character
mode
character
mode sym
character
1
2
3
4
5
a′′
a′′
a′
a′′
a′′
79% ring (boat)
96% OH torsion
69% CCO bend
72% ring (twistboat)
29% O oop bend, 43% ring
torsion
4
3
5
6
8
66% ring torsion
73% OH torsion
47% COH bend, 24% CCC bend
73% ring torsion (twistboat)
25% O oop bend, 53% ring
torsion
2
b1
91% ring torsion (boat)
13
3
8
a1
a2
a2
49% HCC bend, 17% CC stretch
60% ring torsion (twistboat)
93% H oop bend
6
7
8
9
10
11
a′
a′
a′′
a′′
a′′
a′
70% CCC bend
92% CCC bend
74% ring (chair)
36% H oop, 11% O oop bend
88.5 H oop bend
2
7
1
86% CCC bend
85% CCC bend
84% ring torsion (halfchair)
1
4
5
a1
b2
b1
89% CCC bend
92% CCC bend
90% ring torsion (chair)
53% CCC bend, 18% CO, 20%
9
59% CCC bend, 15% CO stretch
9
a1
86% CCC bend
CC stretch
77% H oop bend
84% H oop bend
88% H oop bend
80% CCC bend
50% CC stretch
32% CCH bend, 29% CC stretch
75% HCC bend
78% HCC bend
12
13
14
15
16
17
18
19
20
a′′
a′′
a′′
a′
a′
a′
a′
a′
a′
13
10
11
12
14
15
82% H oop bend
84% H oop bend
87% H oop bend
44% CCC bend, 38% CC stretch
71% CC stretch, 12% HCC bend
65% HCC bend, 15% CC stretch
10
6
7
12
11
14
b1
b1
b1
a1
b2
b2
83% H oop bend
81% H oop bend
85% H oop bend
66% CC stretch, 24% HCC bend
66% CC stretch, 10% HCC bend
75 HCC bend
50% COH bend, 13% CC stretch
16
18
21
17
41% COH bend, 22 CC stretch,
14% HCC
37% CO stretch, 21% HCC, 14%
CC stretch
41% CC stretch, 22% COH, 12%
HCC
21
22
23
a′
a′
a′
47% CO stretch
22
17
15
b2
b2
b2
99% CH stretch
73% CC stretch
75% HCC bend
42% HCC bend, 37% CC stretch
45% HCC bend
56% HCC bend, 16% CO stretch,
13% COH
24
25
a′
a′
47% HCC bend, 24% CC stretch
43% HCC bend, 29% CC stretch
19
20
57% CC stretch, 25% HCC bend
55% CC stretch, 20% HCC, 22%
COH
16
18
b2
a1
42% CC stretch, 41% HCC bend
51% HCC bend, 29% CC stretch
26
27
28
29
30
31
32
33
a′
a′
a′
a′
a′
a′
a′
a′
42% CC stretch
59% CC stretch
90% CH stretch
91% CH stretch
86% CH stretch
97% CH stretch
96% CH stretch
100% OH stretch
23
22
78% CC stretch
41% CC stretch, 19% HCC
20
19
a1
b2
85% CC stretch
34% CC stretch, 29% HCC bend
25
26
99% CH stretch
99% CH stretch
23
24
a1
a1
90% CH stretch
99% CH stretch
24
27
99% CH stretch
100% OH stretch
21
a1
90% CH stretch
a
According to B3LYP/6-31G(d,p) calculations.
1 (no. 22, 1516 cm -1) should be of higher intensity, which is
only partially confirmed by the experiment. The best agreement
between experiment and theory is obtained when scaling factors
of 0.979 (CCSD(T)) and 0.970 (B3LYP) are used in the case
of 1, and 0.955 (0.954), in the case of 9a. As can be seen from
Figure 3, the agreement between CCSD(T) and experimental
infrared spectra clearly confirms that 9a has been isolated in
the matrix and that the geometry is described best as σ-biradical
(Figure 8).
Although the bicyclic isomers 2 and 10a are not observed
under the conditions of matrix isolation, it is of general interest
to calculate the infrared spectra of the bicyclic forms. 2 and
10a are stationary points at the B3LYP/6-31G(d,p) but not at
the CCSD(T)/6-31G(d,p) level of theory. The best way of
identifying the bicyclic isomer of a m-benzyne is to look for
the CC stretching motion of the three-membered ring, which
should occur close to 1800 cm-1 (B3LYP, 1854; scaled with
0.96, 1780 cm-1). This band is weak but should be detectable,
since there are no other overlapping absorptions in this region.
Apart from this, the infrared spectra of the closed-shell structures
2 and 10a differ significantly from those of the diradicals 1
and 9a.
this information is the adiabatic mode analysis of Konkoli and
Cremer,27 which is based on localized adiabatic internal modes
that are associated with the internal coordinates used to define
the geometry of a molecule. Once the normal modes of a
molecule are known, the calculation of the corresponding
adiabatic modes is straightforward.
In Table 3 for selected geometrical parameters the corre-
sponding B3LYP adiabatic frequencies are given. In most cases,
adiabatic frequencies of stretching modes are inversely propor-
tional to the corresponding lengths, i.e., shorter bond lengths
lead to larger adiabatic frequencies and vice versa.32 For
example, the calculated CC bond lengths of phenol (Table 3)
correlate linearly with the corresponding adiabatic frequencies.
However, adiabatic stretching frequencies depend not only on
the bond length but also on the environment of a bond as, e.g.,
the type of atoms, groups or substituents attached to the bond
in question, or the steric arrangement of neighboring bonds.
Thus, the adiabatic stretching frequencies are more appropriate
than bond lengths to reflect the strength of a bond, which
depends on the electron density between the bonded atoms as
well as on environmental effects.
The adiabatic stretching frequencies of 1 and 9a reveal that
despite the CC bond length shortening the ring bonds are not
necessarily more stable. On the contrary, in some cases, the
adiabatic frequencies are smaller than in the reference molecule
In principle, the vibrational frequencies and intensities of a
molecule contain important information on the electronic
structure. However, the vibrational normal modes of a molecule
are delocalized modes and the electronic structure information
is encoded in a complicated way. A powerful way to unravel
(32) Larsson, A.; Cremer, D. To be published.

Formation of 2,4-Didehydrophenols
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10669
phenol. Also, they indicate a strengthening of bonds C1C2 and
C4C5 relative to the other ring bonds, which confirms the
importance of through-bond stabilization effects. However, in
total, the carbon ring is less stiff than in phenol, which is
particularly reflected by the much smaller adiabatic CCC
bending and CCCC torsional frequencies. Both in-plane and
out-of-plane distortions of the six-membered ring should easier
occur in the case of 1 and 9a than either benzene or phenol.
Scheme 4
Discussion
It has been shown that the photodecarboxylation of 4-oxo-
cyclohexadienylidene-3-carboxylic acids (8), which is a special
type of the decarboxylation reaction of â-keto acids, yields 2,4-
didehydrophenols (9) smoothly and in high quantum yield. The
formation of 9 from 7 is a biphotonic process with loss of N2
with carbene 8 as the primary intermediate followed by
decarboxylation on λ > 470 nm irradiation as the second step.
Interestingly, the highly photolabile carbenes 8 can be cleanly
prepared under the conditions of matrix isolation by monochro-
matic irradiation of 7 with blue light.
vibrations that can be easily recognized by matrix isolation
spectroscopy (cf. ketene 19). The hydrogen bridge results in a
significant lowering of the carbonyl bond order. Consequently,
the CO stretching frequency in 7a is red-shifted by 29 cm-1
compared to quinone diazide 11a (Scheme 4).16,14 Concurrently,
the NN stretching vibration of 7a at 2114 cm-1 is red-shifted
by almost 38 cm-1 compared to 11a. These infrared shifts,
which are similarly also observed in 7b-d, indicate that the
introduction of an intramolecular hydrogen bridge augments the
weight of the diazonium phenolate mesomeric structure of 7.
While the tools of X-ray structure analysis and proton NMR
spectroscopy are not available for the determination of the
structures of the corresponding carbenes 8, the IR spectra clearly
indicate very strong hydrogen bridges, which might even
exaggerate that of 7. Again, the intensity of the OH stretching
vibrations in 8 are very low and not to be detected in the spectra.
The carbonyl stretching vibration of carbene 8a is found at 1415
The thermodynamic driving force for this reaction is the
formation of carbon dioxide: the heat of formation of 9a is 85
kcal/mol (see previous section) so that the sums of the heats of
formation of 9a and CO2 is -9 kcal/mol. The heat of formation
of triplet carbene 8a can be estimated from Benson’s group
increments33 to be 9 kcal/mol. Hence, the reaction 8a f 9a +
CO2 is exothermic by 18 kcal mol-1
.
Nevertheless, it has to be discussed why this reaction proceeds
so readily. Usually, the hydroxy protons of carboxylic acids
are very firmly bound and hard to abstract. The observation
that intramolecular hydrogen abstraction takes place by the keto
oxygen atom, which is a radical site that bears only partial
phenoxy character (and should thus not have a particularly high
free radical reactivity), indicates that the O-H bond of the
carboxylic acid moiety of carbene 8a has to be already
significantly weakened. The clue to the problem comes from
the X-ray structure analysis of diazo compounds 7a,c.16 The
structures clearly reveal the presence of an intramolecular
hydrogen bridge. Of particular interest are the extremely short
distances between the keto- and OH-carboxy oxygen atoms with
2.475 Å in 7a and even less (2.442 Å) in 7c. The O-O
distances in O-H-O hydrogen bridges have been correlated
with the strength of the hydrogen bridge;34 thus, the hydrogen
bridges observed in 7a,c can be counted among the strongest
hydrogen bridges known.
This is confirmed by the proton NMR spectra of 7, where
the O-H protons give sharp resonances between δ ) 11 and
17, and by the 13C NMR spectra. The chemical shift of the
diazo carbon (C4) of 7a is found at δ ) 87.0 (in DMSO-d6),
while the corresponding shift of p-benzoquinone diazide 11a
is δ ) 73.6 (in CDCl3). Since the value for δ (C1) of the
benzene diazonium ion is around δ ) 109.7 (as chloride),35 it
is clear that the quinone diazide carboxylic acids (7) range
approximately halfway between typical quinone diazides and
highly polar intramolecular diazonium-phenolate ion pairs.
cm-1 and thus red-shifted by 82 cm-1 compared to 12a (Scheme
4). This clearly demonstrates the importance of the diradicaloid
and zwitterionic resonance structures B and C, respectively, to
describe the wave function of 8a. The zwitterionic resonance
structure C is also in line with the phenyl cation type reactivity
in slightly acidic solvents such as trifluoroethanol.
It is not clear from our experiments if the decarboxylation
of carbenes 8 is a photochemical or a hot ground-state reaction
and at which point along the reaction coordinate the intersystem
crossing from the triplet (carbene) to the singlet surface (m-
benzyne) occurs. A strong hydrogen bridge was observed in
all diazo compounds 7, carbenes 8, carbonyl oxides 13, and
dioxiranes 14 with quite different substituents at the six-
membered ring. Even if in a reactive excited state of 8 the
charge at the carbonyl oxygen atom might be reduced compared
to the triplet ground state, it is reasonable to assume that there
is still a strong hydrogen bond. In any case, the transfer of the
hydrogen atom is facilitated by the very strong O-H-O
hydrogen bridge. The decarboxylation might occur concerted
with the hydrogen migration or in a second step. In an acidic
solvent, where the carbonyl oxygen atom is already protonated
Further evidence for strong intramolecular hydrogen bonding
can be found in the infrared spectra of 7. Due to the strong
O-H-O bridges, the intensities of the OH stretching vibrations
are very low and generally not to be detected in the spectra,
while free carboxylic acids exhibit strong OH stretching
(33) Benson, S. W.; Cruickshank, F. R.; Golden, D. M.; Haugen, G. R.;
OÅNeal, H. E.; Rodgers, A. S.; Shaw, R.; Walsh, R. Chem. ReV. 1969, 69,
279-324.
(34) Ichikawa, M. Acta Crystallogr. 1978, B34, 2074-2080.
(35) Duthaler, R. O.; Foerster, H. G.; Roberts, J. D. J. Am. Chem. Soc.
1978, 100, 4974-4979.

10670 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Sander et al.
acetone/ water (1:1): yellow needles; mp 182 °C (dec); 1H NMR
(DMSO-d6) δ 15.24 (s, br, 1H), 9.28 (s, 1H), 8.42 (d, 1H, J ) 8.6
Scheme 5
3
Hz), 7.88 (m, 2H), 7.67 (m, 1H); 13C NMR (DMSO-d6) δ 81.0, 114.1,
121.8, 126.6, 127.1, 128.5, 129.4, 132.8, 142.5, 165.2, 180.0; IR (KBr)
3432 (w), 3087 (w), 2159 (s, NdN), 1717 (s), 1613 (m), 1578 (s),
1540 (s), 1506 (s), 1476 (s), 1408 (m), 1363 (w), 1319 (w), 1291 (m),
1266 (w), 1198 (s), 1168 (s), 1050 (w), 941 (w), 810 (vw), 763 (s)
cm-1; HRMS calcd for C11H6N2O3 214.0378, found 214.0377; MS m/z
(relative intensity) 214 (4, M+), 186 (14), 170 (27), 158 (15), 142 (7),
114 (39), 102 (10), 88 (9), 71 (7), 63 (12), 57 (7), 50 (5), 44 (27), 39
(7), 28 (100). Anal. Calcd for C11H6N2O3 C, 61.73; H, 2.82, found
C, 61.82; H, 2.78.
4-Amino-3-hydroxy-2-naphthoic Acid was synthesized by azo
coupling of benzenediazonium chloride with 3-hydroxy-2-naphthoic
acid followed by reduction of the azo compound with Na2S2O4.
1,2-Naphthoquinone 1-Diazide 3-Carboxylic Acid (18). 4-Amino-
3-hydroxy-2-naphthoic acid was dissolved in ethanol, saturated with
gaseous HCl. After being cooled to 5 °C, the mixture was treated with
a small excess of isoamylnitrite. The reaction mixture was stirred for
1 h, poured into diethyl ether, and cooled. The diazide precipitated
after a while and was recrystallized from acetone/water (1:1): yellow
and thus proton transfer from the carboxylic group cannot occur,
no decarboxylation is observed. This clearly demonstrates that
proton transfer and decarboxylation are linked together.
In summary, quinone diazide carbocylic acids (7) provide
interesting precursors of 2,4-didehydrophenols (9), a novel class
of m-benzynes. Since the diazo compounds are readily syn-
thesized via standard procedures, a variety of substitution
patterns are easily accessible. Due to the highly polar character
and strong packing forces in the crystals of 7a, the solubility
of this diazo compound in common solvents is very low.
However, by introducing alkyl substituents, the solubility
increases and the solution chemistry can be conveniently studied.
Carbenes 8 as well as the m-benzynes 9 were unequivocally
characterized using matrix isolation spectroscopy and by trap-
ping experiments in solution. Detailed insight into the spec-
troscopic properties and electronic structure of 9a is provided
by ab initio calculations at the CCSD(T) level of theory in
combination with the infrared spectra.
1
needles; mp 182 °C (182 °C); H NMR (DMSO-d6) δ 10.6 (s, broad,
1H), 8.8 (s, 1H), 8.16-8.06 (m, 1H), 7.82-7.76 (m, 2H), 7.6-7.1 (m,
1H); 13C NMR (DMSO-d6) δ 178.2, 165.0, 145.9, 132.9, 132.4, 129.1,
125.3, 122.9, 121.5, 120.6, 83.2; IR (KBr) 3034 (m), 2140 (vs), 1717
(vs), 1612 (vs), 1581 (vs), 1552 (vs), 1491 (vs), 1459 (vs), 1384 (vs),
1352 (vs), 1276 (s), 1215 (vs), 1166 (s), 1087 (m), 883 (m), 799 (m),
763 (m) cm-1; MS m/z (relative intensity) 214 (M+, 47), 186 (84), 169
(23), 142 (52), 130 (20), 114 (54), 102 (100), 87 (25), 76 (17), 71
(12), 63 (47), 57 (11), 51 (14), 39 (20), 28 (38);
Matrix Isolation. Matrix isolation experiments were performed by
standard techniques with an APD CSW-202 Displex closed-cycle
helium cryostat.39 Matrices were produced by deposition of argon
(Linde, 99.9999%) on top of CsI (IR) or sapphire (UV-vis) window
with a rate of approximately 0.15 mmol/min. Infrared spectra were
recorded by using a Bruker IFS66 FTIR spectrometer with a standard
resolution of 1 cm-1 in the range 400-4000 cm-1. UV-vis spectra
were recorded on a Hewlett-Packard 8452A diode array spectropho-
tometer with a resolution of 2 nm. Irradiations were carried out with
use of Ushio HBO 500 W mercury high-pressure arc lamps in Oriel
housings equipped with quartz optics. IR irradiation from the lamps
was absorbed by a 10 cm path of water. For broad-band irradiation,
Schott cutoff filters were used (50% transmission at the wavelength
specified), and for narrow-band irradiation, interference filters in
combination with dichroic mirrors (“cold mirrors”) and cutoff filters
were used.
Experimental Section
Materials and General Methods. The NMR spectra were recorded
with a Bruker AM-400 spectrometer, and chemical shifts are reported
in parts per million (δ) relative to TMS unless otherwise indicated.
Gas chromatography (GC) was performed by the use of a Siemens
Sichromat equipped with glass capillary columns (39 m, OV17, 160
°C). High-pressure liquid chromatography with LCD (Milton Roy)
chromatographs and refractometric detection. Mass spectra were
obtained on Varian MAT CH5 instrument (70 eV). Melting points
were determined on a Kofler hot-stage apparatus and are uncorrected.
Acetonitrile and tetrachloromethane (for UV spectroscopy) were
obtained from Acros Chimica, while MTHF was distilled immediately
prior to use. For product analysis, dilute degassed solutions of the
diazo compounds were irradiated.
3-Diazo-6-oxo-1,4-cyclohexadiene-1-carboxylic acid (8a)10,36 and its
3-alkylated derivatives 7b-d were synthesized by diazotization of the
corresponding 5-aminosalicylic acids according to the procedure of Ried
and Appel (Scheme 5).37 Purification of the diazo compounds was
accomplished by recrystallization from water/acetone (1:1).
4-Diazo-1-oxo-2,5-cyclohexadiene-2-carboxylic Acid (7a): IR (Ar,
10 K) 2113.6 (s), 2092.9 (w), 2057.2 (vw), 1740.9 (w), 1595.8 (vs),
1564.0 (vw), 1535.1 (w), 1481.1 (m), 1473.3 (s), 1458.4 (w), 1379.3
(w), 1335.0 (vw), 1260.3 (vw), 1204.3 (vw), 1164.8 (s), 1127.2 (w),
838.9 (vw), 528.9 (vw) cm-1(relative intensity); temperature of deposi-
tion 136 °C.
1-Oxo-2,5-Cyclohexadien-4-ylidene-2-carboxylic Acid (8a): IR
(Ar, 10 K) 1758.3 (vs, 0.994), 1750.1 (m, -), 1523.5 (w, -), 1495.0 (m,
0.998), 1458.9 (w, 0.999), 1422.2 (s, -), 1415.0 (vs, 1.000), 1383.7
(w, 1.000), 1291.6 (w, 1.001), 1228.9 (w, -), 1053.9 (vw, -), 982.6
(vw, -), 826.3 (w, -), 565.0 (vw, 1.000), 542.9 (vw, -) cm-1 (relative
2
intensity, ratio of H/1H isotopic frequencies νi/ν).
2,4-Didehydrophenol (9a): IR (Ar, 10 K) 3612.0 (s, 0.738), 1516.3
(s, 0.997), 1429.0 (w, 0.982), 1368.2 (m, 0.987), 1290.1 (w, 0.979),
1254.9 (m, -), 1209.1 (m, -), 1157.6 (m, 0.814), 1128.6 (vw, -),
971.0 (w, 1.008), 877.0 (w, 0.997), 694.7 (m, 0.981), 641.2 (s, 1.000),
518.8 (s, 0.984) cm-1 (relative intensity, ratio of 2H/ 1H isotopic
frequencies νi/ν).
Diazo compounds 7 were deuterated at the carboxy position by
briefly heating the compounds to ∼70 °C in a 1:1 mixture of acetonitrile
and D2O. The degree of deuteration obtained was never in excess of
80%. Attempts to drive the exchange reaction to completion by longer
reaction times or higher temperatures failed due to competing decom-
position of the diazo compounds.
1,4-Benzoquinone-2-carboxylic Acid 4-O-Oxide (13a): IR (Ar,
10K) 1774.2 (m, 0.998), 1764.5 (vs, 1.000), 1753.3 (s, 1.000), 1607.9
(vs, 1.000), 1577.0 (s, 0.999), 1573.1 (s, 0.998), 1416.5 (vs, 1.000),
1408.3 (s, 1.001), 1382.2 (w, 0.999), 1351.4 (w, 1.001), 1343.7 (w,
0.999), 1235.2 (m, 0.999), 1222.6 (s, 0.998), 1218.8 (m, -), 1176.4
(m, 0.996), 1101.6 (s, 0.984), 1069.3 (m, 0.968), 1066.0 (m, 0.960),
1003.8 (m, 0.988), 943.0 (m, 0.999), 932.4 (vw, -), 890.5 (w, 1.001),
886.1 (w, 0.998), 880.3 (w, 1.001), 823.5 (vw, 1.001), 819.6 (vw,
1,4-Naphthoquinone Diazide 2-Carboxylic Acid (7d). 4-Amino-
1-hydroxy-2-naphthoic acid38 was dissolved in ethanol, saturated with
gaseous HCl. After being cooled to 5 °C, the mixture was treated with
a small excess of isoamylnitrite. The reaction mixture was then stirred
for 1 h and subsequently poured into diethyl ether and cooled to -10
°C. The diazide precipitated after a while and was recrystallized from
(36) Rugeddi, A. Gazz. Chim. Ital. 1929, 13.
(37) Ried, W.; Appel, H. Liebigs Ann. Chem. 1961, 646, 82-95.
(38) Grandmougin, E. Chem. Ber. 1906, 39, 3609-3611.
(39) Sander, W. W. J. Org. Chem. 1989, 54, 333-339.

Formation of 2,4-Didehydrophenols
J. Am. Chem. Soc., Vol. 119, No. 44, 1997 10671
0.999), 813.3 (w, 0.999), 723.2 (w, 1.001), 708.2 (w, 0.990), 655.2
(w, -), 594.4 (w, -), 585.3 (w, 1.001) cm-1 (relative intensity, ratio
of 18O/16O isotopic frequencies νi/ν).
(m), 1211.6 (m), 1196.4 (vs), 1166.3 (m), 762.4 (m), 759.0 (w), 612.8
(w), 495.2 (vw) cm-1 (relative intensity); temperature of deposition
155 °C.
1,4-Dihydro-1-oxonaphthalen-4-ylidene-2-carboxylic Acid (8d):
IR (Ar, 10K) 1753.0 (s), 1744.6 (m), 1716.9 (w), 1561.2 (m), 1558.3
(m), 1540.2 (m), 1470.9 (m), 1441 (m), 1429.0 (vs), 1405.8 (m), 1397.3
(s), 1325.4 (w), 1305.7 (vw), 1273.8 (w), 1140.9 (w), 901.0(vw), 737.6
(vw), 724.5 (vw), 640.5 (w), 476.3 (w) cm-1 (relative intensity). 2,4-
Didehydro-1-naphthol (9d): IR (Ar, 10K) 3608 (s), 1604.6 (m), 1582.5
(w), 1546.0 (vw), 1485.1 (vw), 1453.9 (vw), 1392.1 (w), 1343.0 (w),
1336.2 (w), 1334.4 (w), 1236.1 (w), 1233.2 (w), 1216.4 (vw), 1212.3
(vw), 1167.8 (m), 1011.0 (vw), 981.5 (m), 976.8 (m), 788.8 (m), 758.5
(m), 659.9 (w), 645.7 (vw), 641.6 (vw), 639.0 (vw), 631.0 (vw), 603
(vw) cm-1 (relative intensity).
1,2-Naphthoquinone 2-Diazide 3-Carboxylic Acid (18): IR (Ar,
10K) 2120 (79), 2117 (5), 2103 (19), 1768 (1), 1767 (69), 1762 (14),
1670 (5), 1620 (100), 1611 (13), 1606 (2), 1600 (22), 1566 (88), 1487
(18), 1462 (55), 1443 (2), 1438 (8), 1377 (15), 1375 (4), 1299 (6),
1295 (2), 1289 (4), 1277 (19), 1215 (19), 1209 (76), 1203 (15), 1193
(5), 1163 (15), 1093 (6), 1017 (2), 967 (4), 857 (11), 800 (19), 757
(5), 749 (16), 701 (4), 665 (9), 590 (8), 492 (6), 454 (6) cm-1 (relative
intensity); temperature of deposition 140 °C.
1-Carbonyl-1H-indene-2-carboxylic Acid (19): IR (Ar, 10 K) 3578
(2), 3570 (10), 2165 (2), 2142 (100), 2128 (4), 1758 (21), 1540 (6),
1416 (2), 1350 (3), 1264 (2), 1216 (4), 1149 (10), 1143 (2), 1139 (5),
1103 (4), 1068 (3), 767 (2), 746 (2), 735 (2), 653 (2), 568 (4), 533 (2),
505 (2) cm-1 (relative intensity).
Computational Methods. CCSD(T)24 geometry optimizations based
on analytical energy gradients have been carried out for 1, 9a, 2, and
10a using the standard 6-31G(d,p) basis set.23 Vibrational frequencies
and infrared intensities of 2a have been obtained by combining
analytical and numerical derivative techniques. Calculations have been
repeated at the DFT level with Becke’s three-parameter functional
B3LYP21 for which analytical first and second derivatives are available.
6-Oxo-1,2-Dioxaspiro[2,5]octa-4,7-diene-5-carboxylic
Acid
(14a): IR (Ar, 10K) 1778.0 (vs, 1.000), 1674.9 (s, 0.998), 1657.0 (s,
0.995), 1633.4 (w, 1.001), 1405.9 (vs, 0.996), 1395.2 (vs, 1.000), 1391.9
(vs, -), 1321.5 (m, 0.998), 1271.3 (vw, 0.994), 1105.0 (m, 1.000),
1039.0 (m, -), 1034.1 (m, 0.995), 989.3 (vw, 0.992), 953.1 (vw, 0.995),
864.0 (w, 0.995), 853.3 (m, 1.001), 847.6 (m, 0.999), 804.2 (w, 0.990),
793.1 (w, 0.998), 654.2 (m, -) cms1 (relative intensity, ratio of 18O/
16O isotopic frequencies νi/ν).
6-Methyl-4-diazo-1-oxo-2,5-cyclohexadiene-2-carboxylic Acid
(7b): IR (Ar, 10K) 2181.7 (11), 2213 (14), 2175.8 (15), 2129 (41),
2121.6 (38), 2116.6 (44), 2103.8 (100), 2098.8 (82), 2080.5 (12), 2073.6
(16), 1742.1 (48), 1728.4 (28), 1618.1 (26), 1605.3 (72), 1599.4 (64),
1564.6 (78), 1536.3 (15), 1492.8 (39), 1478.2 (50), 1459.5 (44), 1433.8
(29), 1382.2 (14), 1358.8 (42), 1240.9 (23), 1227.1 (71), 1182.8 (30),
1176.8 (71), 1106.9 (55), 1010.4 (11), 952.3 (10), 926.2 (4), 899.1
(-), 666.2 (-), 536.6 (-) cm-1 (relative intensity). Temperature of
deposition 140 °C.
6-Methyl-1-oxo-2,5-cyclohexadien-4-ylidene-2-carboxylic Acid
(8b): IR (Ar, 10K) 1764.6 (73, 0.999), 1757.4 (100, 0.999), 1520.7
(13, 1.000), 1504.8 (33, 0.996), 1424.9 (78, 0.996), 1417.6 (50, 0.990),
1399.1 (49, 0.996), 1374.5 (21, 1.000), 1326.5 (vw, -), 1288.7 (34,
0.988), 1136.2 (26, 0.954), 1023.8 (w, 0.999), 851.5 (w, 1.004), 790.5
(w, -), 724.0 (w, -), 598.0 (w, -) cm-1 (relative intensity, ratio of
2H/1H isotopic frequencies νi/ν).
6-Methyl-2,4-didehydrophenol (9b): IR (Ar, 10K) 3612.3 (100,
0.738), 1500.9 (37, 0.996), 1377.0 (34, -), 1262.6 (12, 0.993), 1249.2
(38, 0.996), 1242.9 (30, 0.742), 1204.7 (22, 0.993), 1063.0 (11, 0.995),
1039.3 (2, 0.999), 1015.3 (16, 1.000), 944.1 (18, 0.999), 731.7 (14,
0.977), 662.1 (24, 1.004), 641.3 (22, 0.999), 620.9 (4, 1.000), 509.8
(w, 1.000) cm-1 (relative intensity, ratio of 2H/1H isotopic frequencies
νi/ν).
Calculated vibrational modes were investigated using the adiabatic
mode analysis of Konkoli and Cremer.27 This approach is based on a
decomposition of normal modes in terms of adiabatically relaxed
internal parameter modes that are not contaminated by any other mode
of the molecule. As has been shown previously, the adiabatic mode
analysis is superior to the potential energy distribution (PED) analysis
and provides reliable internal frequencies that can directly be assigned
to the internal parameters of a molecule.27 For all molecules considered,
zero point energy (ZPE) and thermal corrections have been determined
to evaluate reaction enthalpies at 298 K. Calculations were carried
out with the COLOGNE96,40 ACES,41 and GAUSSIAN94 ab initio
packages.42
X-ray Crystal Structure Determination of 7c: Siemens P4
diffractometer, graphite monochromator, Mo KR radiation (λ )
0.710 73 Å), SHELXS-86 program for structure solution by direct
methods and SHELXL-93 program for refinement by full-matrix least-
squares techniques against F.243, 44
Crystal Data: C11H12N2O3, M ) 219.2, monoclinic, space group
P21/c (no. 14) with a ) 9.154(2), b ) 10.377(2), c ) 11.362(2) Å, â
) 93.69(3)°, V ) 1077.1(4) Å,3 Z ) 4, F(000) ) 460, Dcalc ) 1.352
g‚cm-3. Intensity data were collected using ω scans at T ) 218 K on
a Siemens P4 diffractometer for a yellow crystal of dimensions 0.24
× 0.36 × 0.78 mm with Mo KR radiation (λ ) 0.7173 Å, µ ) 0.100
mm-1) in the range 2.66 e 0 e 22.56° (-9 e h e 9, -11 e k e 11,
6-Methyl-1,4-benzoquinone-2-carboxylic Acid 4-O-Oxide (13b):
IR (Ar, 10 K) 1774.2 (m, 0.998), 1764.5 (vs, 1.000), 1753.3 (s, 1.000),
1607.9 (vs, 1.000), 1577.0 (s, 0.999), 1573.1 (s, 0.998), 1416.5 (vs,
1.000), 1408.3 (s, 1.001), 1382.2 (w, 0.999), 1351.4 (w, 1.001), 1343.7
(w, 0.999), 1235.2 (m, 0.999), 1222.6 (s, 0.998), 1218.8 (m, -), 1176.4
(m, 0.996), 1101.6 (s, 0.984), 1069.3 (m, 0.968), 1066.0 (m, 0.960),
1003.8 (m, 0.988), 943.0 (m, 0.999), 932.4 (vw, -), 890.5 (w, 1.001),
886.1 (w, 0.998), 880.3 (w, 1.000), 823.5 (vw, 1.001), 819.6 (vw,
0.999), 813.3 (w, 0.999), 723.2 (w, 1.001), 708.2 (w, 0.990), 655.2
(w, -), 594.4 (w, -), 585.3 (w, 1.001) cm-1 (relative intensity, ratio
of 18O/16O isotopic frequencies νi/ν).
1,2-Dioxa-7-methyl-6-oxospiro[2,5]octa-4,7-diene-5-carboxylic Acid
(14b): IR (Ar, 10 K) 1783.8 (s, -), 1779.0 (vs, 1.001), 1769.8 (m, -),
1674.4 (m, 1.001), 1669.1 (m, -), 1657.5 (m, 1.000), 1645.5 (m, 0.999),
1448.3 (w, 1.000), 1394.3 (vs, 0.999), 1387.0 (s, -), 1358.1 (vw, -),
1354.7 (vw, 0.995), 1317.6 (m, 0.999), 1294.5 (w, 1.001), 1249.2 (vw,
1.000), 1241.9 (vw, 1.000), 1120.0 (vw, 1.000), 1039.4 (vw, 1.000),
1031.2 (vw, -), 1014.9 (w, 0.999), 950.3 (vw, 0.994), 926.2 (vw,
0.999), 872.6 (vw, -), 810.4 (vw, -), 798.9 (m, 0.998), 770.9 (vw,
1.000), 731.4 (vw, 1.000), 690.9 (vw, 0.997), 674.5 (vw, 0.993), 660.5
(w, 1.000), 642.2 (w, 1.000), 510.1 (vw, 0.998) cm-1 (relative intensity,
ratio of 18O/16O isotopic frequencies νi/ν).
6-Isobutyl-4-diazo-1-oxo-2,5-cyclohexadiene-2-carboxylic Acid
(7c): IR (Ar, 10K) 2105.6 (m), 1742.4 (m), 1607.9 (m), 1561.4 (w),
1479.0 (w), 1461.0 (vw), 1289.1 (vw), 1176.6 (w), 1126 (vw), 1113
(vw) cm-1 (relative intensity); temperature of deposition 112 °C.
(40) Kraka, E.; Gauss, J.; Reichel, F.; He, Zhi; Olsson, L.; Konkoli, Z.;
Cremer, D. COLOGNE 96; Go¨teborg, Sweden, 1996.
(41) Stanton, J. F.; Gauss, J.; Watts, J. D.; Lauderdale, W. J.; Bartlett,
R. J. ACES II, Quantum Theory Project; ACES II: University of Florida,
FL, 1992.
(42) Frisch, M. J.; Head-Gordon, M.; Trucks, G. W.; Foresman, J. B.;
Schlegel, H. B.; Raghavachari, K.; Robb, M. A.; Binkley, J. S.; Gonzalez,
C.; Defrees, D. J.; ox, D. J.; Whiteside, R. A.; Seeger, R.; Melius, C. F.;
Baker, J.; Martin, R. L.; Kahn, L. R.; Stewart, J. J. P.; Topiol, S.; Pople, J.
A. Gaussian 94; Gaussian Inc.: Pittsburgh, PA, 1994.
6-Isobutyl-1-oxo-2,5-cyclohexadien-4-ylidene-2-carboxylic Acid
(8c): IR (Ar, 10K) 1760.5 (s), 1717.4 (m), 1699 (vw), 1695.7 (w),
1684.6 (w), 1652.5 (w), 1616.4 (w), 1576.0 (m), 1495.7 (vw), 1436.2
(vw), 1419.1 (vw) cm-1 (relative intensity).
6-Isobutyl-2,4-didehydrophenol (9c): IR (Ar, 10K) 3612.5 (vs),
1595 (w), 1493 (m), 1289.0 (m), 1246.4 (s), 1092 (m), 667.4 (m), 659
(m), 641 (m) cm-1 (relative intensity).
(43) (a) Sheldrick, G. M. SHELXS 86, A Program for Structure
Determination, Go¨ttingen, 1986. (b) Sheldrick, G. M. SHELXS 93, A
Program for Structure Refinement, Go¨ttingen, 1993.
1,4-Naphthoquinone 4-Diazide 2-Carboxylic Acid (7d): IR (Ar,
10K) 2134.1 (m), 2105.1 (s), 2086.5 (s), 2073.3 (w), 1762 (s), 1731.9
(m), 1616.8 (m), 1608.2 (s), 1585.9 (vs), 1551.6 (s), 1543.7 (w), 1477.9
(m), 1462.2 (m), 1456.8 (m), 1452.3 (s), 1448.4 (s), 1295.7 (s), 1266.6
(44) Atom coordinates (× 104) with equivalent isotropic displacement
parameters (Å2 × 103). Ueq is defined as one-third of the trace of the
orthogonalized Uij tensor.

10672 J. Am. Chem. Soc., Vol. 119, No. 44, 1997
Sander et al.
0 e l e 12). Three control reflections monitored at regular intervals
during data collection displayed no significant deviations in intensity.
Semiempirical absorption corrections were performed using ψ scan
data. The structure was solved by direct methods using SHELX-8643a
and refined against F2 with SHELX-9343b for 1401 independent
reflections (from 1498 measured, Rint ) 0.051) to R ) 0.061 [I > 2σ-
(I)], wR2 ) 0.160 [all data]44 and S ) 0.907 [goodness-of-fit].
Hydrogen atoms were included at calculated positions with group
isotropic temperature factors. All other atoms were refined anisotro-
pically. The largest difference synthesis peak and hole ∆F were
Go¨teborg this work was supported by the Swedish Natural
Science Research Council (NFR). Calculations were done on a
CRAY YMP/464 of the Nationellt Superdatorcentrum (NSC),
Linko¨ping, Sweden. E.K. and D.C. thank the NSC for a generous
allotment of computer time.
Supporting Information Available: Tables of crystal-
lographic data, atomic coordinates, anisotropic and isotropic
displacement parameters, the package diagram of quinone
diazide 7c, and the calculated IR spectra of 2 and 10a (6 pages).
See any current masthead page for ordering and Internet access
instructions.
respectively 0.200 and -0.193 Å-3
.
Acknowledgment. Dedicated to Prof. Dr. Orville L. Chap-
man on the occasion of his 65th birthday. This work was
financially supported by the Deutsche Forschungsgemeinschaft
and the Fonds der Chemischen Industrie. At the University of
JA971731A