Structure determination of triplet diphenylcarbenes by in situ X-ray crystallographic analysis

Article abstract of DOI:10.1021/ja067306b

Crystalline-state photoreactions of the following diphenyldiazomethanes were investigated by in situ X-ray crystallography, spectroscopy, and theoretical calculations: bis(2,4,6-trichlorophenyl)diazomethane (1-N ₂), bis(2,4,6-tribromophenyl)dia

Full text of DOI:10.1021/ja067306b

Published on Web 01/31/2007
Structure Determination of Triplet Diphenylcarbenes by in
Situ X-ray Crystallographic Analysis
Masaki Kawano,*,†, Katsuyuki Hirai,‡ Hideo Tomioka,*,‡,§ and Yuji Ohashi*,#
Contribution from the Department of Chemistry and Materials Science, Tokyo Institute of
Technology, Tokyo 152-8551, Japan, CREST, Japan Science and Technology Corporation,
Kawaguchi, Saitama 332-0012, Japan, and Chemistry Department for Materials, Faculty of
Engineering, and Instrumental Analysis Facilities, Life Science Research Center, Mie UniVersity,
Tsu, Mie 514-8507, Japan
Received October 17, 2006; E-mail: mkawano@appchem.t.u-tokyo.ac.jp
Abstract: Crystalline-state photoreactions of the following diphenyldiazomethanes were investigated by in
situ X-ray crystallography, spectroscopy, and theoretical calculations: bis(2,4,6-trichlorophenyl)diazomethane
(1-N2), bis(2,4,6-tribromophenyl)diazomethane (2-N2), bis(2,6-dibromo-4-methylphenyl)diazomethane (3-
N2), bis(2,6-dibromo-4-tert-butylphenyl)diazomethane (4-N2), (2,4,6-tribromophenyl)-(2,6-dimethyl-4-tert-
butylphenyl)diazomethane (5-N2), bis(4-bromophenyl)diazomethane(6-N2), and diazofluorene (7-N2). Crystal
structures of photoinduced triplet diphenylcarbenes (DPCs) of 1, 2, and 4 were determined. We found
remarkable differences between their structural information obtained in the crystalline state and that
previously obtained spectroscopically in a glass matrix. Although the triplet DPCs of 1, 2, and 4 have
significantly different stabilities in solution, only subtle differences in their structural parameters, except for
their C(:)-Ar bond lengths, are observed. It is noteworthy that the average bond length of C(:)-Ar for 4
(1.374 Å) is considerably shorter than those for 31 and 32 (1.430 and 1.428 Å, respectively), provided that
the two C(:)-Ar bonds being compared were chemically equivalent. The most likely explanations for the
small and large differences in bond lengths in 1, 2, and 4 may be derived from the packing effect. The
packing patterns of 1 and 2 are identical, but that of 4 is totally different from those of 1 and 2. Moreover,
these results are interpreted as indicating that triplet DPCs undergo relaxation upon softening of the
environments. Theoretical calculations indicate that the potential energy surface of triplet DPCs in terms of
the carbene angle is extremely flat and changes in the angles have little effect on the energies. Triplet
DPCs with a sterically congested carbene center are trapped in a structure dictated by the precursor structure
in a rigid matrix, even if this is not the thermodynamically most stable geometry, but undergo geometrical
relaxation upon softening the matrix to relieve steric compression. ESR studies indicate that the interplanar
angles are more flexible than the bond angles.
Introduction
chers’ understanding of the properties of these highly elusive
species has been deepened tremendously. This is especially impor-
The past two decades have witnessed tremendous advances in
techniques to observe highly elusive species with only fleeting
existence at room temperature.1 Generation of extremely reactive
molecules under low-temperature matrix isolation conditions has
provided considerable information on these otherwise elusive
molecules.2 Time-resolved laser experiments have permitted the
spectroscopic detection of ephemeral species with lifetimes as
short as picoseconds.3 Moreover, rapid advances in computation-
al chemistry have made it possible to predict the structures and
properties of reactive intermediates fairly precisely.4 Thus, resear-
tant since one of the central objectives of the study of chemistry
is sufficient knowledge to permit the forecasting of the chemical
and physical properties of a substance directly from its structure
and since it is the properties of these reactive intermediates that
often control the outcome of chemical reactions.
Other approaches to reactive intermediates have also been
investigated. Organic chemists have attempted to isolate oth-
(2) (a) Sander, W.; Bettinger, H. F. In AdVances in Carbene Chemistry; Brinker,
U., Ed.; JAI Press: Greenwich, CT, 2001; Vol. 3, pp 160-203. (b) Sander,
W.; Kirschfeld, A. In AdVances in Strain and Organic Chemistry; Halton,
B., Ed.; JAI Press: Greenwich, CT, 1995; Vol. 4, pp 1-80. (c) Tomioka,
H. Bull. Chem. Soc. Jpn. 1998, 71, 1501. (d) Sander, W. In Carbene
Chemistry; Bertrand, Ed.; Fontis Media SA: Lausanne, 2002; pp 1-25.
(e) Maier, G.; Reisenauer, H. P. In AdVances in Carbene Chemistry; Brinker,
U., Ed.; JAI Press: Greenwich, CT, 2001; Vol. 3, pp 116-157.
# Tokyo Institute of Technology.
† CREST, Japan Science and Technology Corporation.
‡ Mie University.
(3) See, for reviews: (a) Eisental, K. B. In Ultrashort Light Pulse; Shapiro,
S., Ed.; Spinger-Verlag: Berlin, 1977; Chapter 5. (b) Scaiano, J. C. In
ReactiVe Intermediate Chemistry; Moss, A. M., Platz, M. S., Jones, M.,
Jr., Eds.; Wiley: New York, 2004; pp 847-871. (c) Hilinski, E. In ReactiVe
Intermediate Chemistry; Moss, A. M., Platz, M. S., Jones, M., Jr., Eds.;
Wiley: New York, 2004; pp 873-897.
Present address: Departure of Applied Chemistry, School of Engineer-
ing, The University of Tokyo, 7-3-1 Hongo Bunkyo-ku, Tokyo 113-8656,
Japan.
§ Present address: Department of Applied Chemistry, Aichi Institute of
Technology, Toyota, Aichi 470-0392, Japan.
(1) ReactiVe Intermediate Chemistry; Moss, A. M., Platz, M. S., Jones, M.,
Jr., Eds.; Wiley: New York, 2004.
(4) Borden, W. T. In ReactiVe Intermediate Chemistry; Moss, A. M., Platz,
M. S., Jones, M., Jr., Eds.; Wiley: New York, 2004; pp 961-1004.
9
10.1021/ja067306b CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 2383-2391
2383

A R T I C L E S
Kawano et al.
erwise transient species by chemical modification. Many of those
species have been stabilized, some of which have been isolated
under ambient conditions5 and even structurally fully character-
ized by X-ray crystallographic techniques. Such “direct”
observations, reinforced by theoretical predictions, have revealed
many previously unknown aspects of those species.
are still not stable enough to be isolated under ambient
conditions. Thus, it is still not possible to obtain a crystal of
stable triplet carbene for X-ray crystallographic analysis.
A very powerful technique to characterize the crystal structure
of highly elusive species emerges. This technique enables us
to observe in situ the molecular structure of unstable species
generated photochemically in a single crystal of an appropriate
precursor molecule to the extent that the crystallinity of the
sample is retained. Starting with single-crystal-to-single-crystal
reaction of various organic and organometallic compounds, the
in situ method has been successfully used to characterize the
molecular structures of very unstable species such as radical
pairs from hexaarylbiimidazole derivatives,13 triplet nitrenes,14
the photoinduced metastable state of a transition-metal nitrosyl
complex,15 and triplet excited states of [Pt2(H2P2O5)4]4- ion16
and [Rh2(1,3-diisocyanopropane)4]2+ ion.17
Carbenes became very attractive target molecules. Carbenes
are neutral, divalent derivatives of carbon. The carbene atom
has two electrons, not involved in bonding, that can be spin-
paired (singlet state) or unpaired (triplet state). Thus, the species
presents many challenging issues, not present in other reactive
species, such as the effect of structure on singlet-triplet energy
gap, chemical reactivities of each state, intersystem-crossing
efficiency, and so on.6 A great deal of spectroscopical7 as well
as theoretical works8 have been devoted to these issues and have
revealed the nature of the species in considerable detail.
Simultaneously, efforts to stabilize and isolate those highly
elusive species have also been made.9 Singlet carbenes undergo
thermodynamic stabilization more easily than the triplet states
and are stabilized with substituents such as R2N and R2P. Some
of these species are thermodynamically more stable than their
alkene dimerization products. The structures of those carbenes
are fully characterized by X-ray crystallographic analysis.10,11
Stable forms of triplet carbenes are more difficult to obtain.9
The dimerization reactions of triplet methylene, phenylcarbene,
and vinylcarbene are exothermic by 728, 628, and 610 kJ/mol,
respectively, indicating that conjugation with π systems will
not lead to thermodynamically stable triplet carbenes.9g Hence,
kinetic stabilization is a more promising approach to persistent
triplet carbenes.
In light of the fact that even the structures of very persistent
triplet diphenylcarbenes (DPCs) have been characterized only
by “indirect” spectroscopic methods,18 it is very tempting to
use this in situ method to characterize the structures of triplet
carbenes. The fact that carbenes can be efficiently and cleanly
generated by photolysis of precursor diazo compounds19 makes
the idea feasible. Actually, we have shown that triplet bis(2,4,6-
trichlorophenyl)carbene (1), photolytically generated in a single
crystal of the corresponding precursory diazo compound, could
be structurally characterized by using this method. In this way,
the molecular structure of triplet diphenylcarbene was revealed
for the first time.20
As an extension of this approach, we have carried out the in
situ observation of photoinduced DPCs in single crystals of a
series of diphenyldiazomethanes (DDMs) at low temperatures.
The diazo precursors were chosen so as to generate triplet DPCs
with different stabilities (Chart 1). Those chosen are bis(2,4,6-
tribromophenyl)diazomethane (2-N2),21 bis(2,6-dibromo-4-meth-
ylphenyl)diazomethane (3-N2),21 bis(2,6-dibromo-4-tert-bu-
tylphenyl)diazomethane (4-N2),21 and (2,4,6-tribromophenyl)-
(2,6-dimethyl-4-tert-butylphenyl)diazomethane (5-N2),21 in
addition to bis(2,4,6-trichlorophenyl)diazomethane (1-N2),22 all
A great effort to stabilize triplet carbenes has been made, by
which fairly stable ones have been realized.12 However, they
(5) See, for instance: (a)West, R.; Fink, M. J.; Michl, J. Science 1981, 214,
1343. (b) Okazaki, R.; West, R. AdV. Organomet. Chem. 1996, 39, 231.
(c) Kira, M.; Iwamoto, T. J. Organomet. Chem. 2000, 611, 236. (d)
Weidenbruch, M. In The Chemistry of Organic Silicon Compounds;
Rapporport, Z., Apeloig, Y., Eds.; John Wiley & Sons: Chichester, 2001;
Vol 3, p 391. (e) Sekiguch, A.; Lee, V. Y. Chem. ReV. 2003, 103, 1429.
(6) For reviews of general reactions of carbenes, see: (a) Kirmse, W. Carbene
Chemistry, 2nd ed.; Academic Press: New York, 1971. (b) Carbenes, Vols.
1 and 2; Moss, R. A., Jones, M., Jr., Eds.; Wiley: New York, 1973 and
1975. (c) Carbene(oide), Carbine; Regitz, M. Ed.; Thieme: Stuttgart, 1989.
(7) For reviews of laser flash photolysis study of carbenes, see: (a) Moss, R.
S.; Turro, N. J. In Kinetics and Spectroscopy of Carbenes and Biradicals;
Platz, M. S., Ed.; Plenum Press: New York, 1990; pp 213-238. (b) Platz,
M. S.; Maloney, V. M. In Kinetics and Spectroscopy of Carbenes and
Biradicals; Platz, M. S., Ed.; Plenum Press: New York, 1990; pp 239-
352. (c) Moss, R. A. In AdVances in Carbene Chemistry; Brinker, U. H.,
Ed.; JAI Press: Greenwich, 1994; pp 59-88. (d) Jackson, J. E.; Platz, M.
S. In AdVances in Carbene Chemistry; Brinker, U. H., Ed.; JAI Press:
Greenwich, 1994; pp 89-160.
(8) For an excellent review on how to deal computationally with open-shell
species, see: Bally, T.; Borden, W. T. In ReViews in Computational
Chemistry; Lipowitz, K. B., Boyd, D. B., Eds.; Wiley: New York, 1998.
(9) (a) Regitz, M. Angew. Chem., Int. Ed. Engl. 1991, 30, 674. (b) Dagani, R.
Chem. Eng. News 1991, Jan 28, 19; 1994, May 2, 20. (c) Heinemann, C.;
Mu¨ller, T.; Apeloig, Y.; Schwartz, H. J. Am. Chem. Soc. 1996, 118, 2023.
(d) Beohme, C.; Frenking, G. J. Am. Chem. Soc. 1996, 118, 2039. (e)
Wentrup, C. Science 2001, 292, 1846. (f) Roth, H. D. Nature 2001, 412,
598. (g) Kirmse, W. Angew. Chem., Int. Ed. 2003, 42, 2117.
(10) Igau, A.; Gru¨tzmacher, H.; Baceiredo, A.; Bertrand, M. J. Am. Chem. Soc.
1988, 110, 6463. See, for review: (a) Bourissou, D.; Guerret, O.; Gabbai,
F. P.; Bertrand, G. Chem. ReV. 2000, 100, 39. (b) Bertrand, G. In ReactiVe
Intermediate Chemistry; Moss, A. M., Platz, M. S., Jones, M., Jr., Eds.;
Wiley: New York, 2004; pp 329-373.
(11) Arduengo, A. J., III; Harlow, R. L.; Kline, M. J. Am. Chem. Soc. 1991,
113, 361. See also: Arduengo, A. J., III. Acc. Chem. Res. 1999, 32, 913.
(12) (a) Tomioka, H. Acc. Chem. Res. 1997, 30, 315. (b) Tomioka, H. In
AdVances in Carbene Chemistry; Brinker, U., Ed.; JAI Press: Greenwich,
CT, 1998, Vol. 2, pp 175-214. (c) Tomioka, H. In AdVances in Strained
and Interesting Organic Molecules; Halton B., Ed.; JAI Press: Greenwich,
CT, 2000; Vol. 8, pp 83-112. (d) Tomioka, H. In Carbene Chemistry;
Bertrand G., Ed.; Fontis Media SA: Lausanne, 2002; pp 103-152.
(13) (a) Kawano, M.; Sano, T.; Abe, J.; Ohashi, Y. J. Am. Chem. Soc. 1999,
121, 8106. (b) Kawano, M.; Sano, T.; Abe, J.; Ohashi, Y. Chem. Lett. 2000,
1372. (c) Kawano, M.; Ozawa, Y.; Matsubara, K.; Imabayashi, H.; Mitsumi,
M.; Toriumi, K.; Ohashi, Y. Chem. Lett. 2002, 1130.
(14) Kawano, M.; Takayama, T.; Uekusa, H.; Ohashi, Y.; Ozawa, Y.; Matsubara,
K.; Imabayashi, H.; Mitsumi, M.; Toriumi, K. Chem. Lett. 2003, 922.
(15) (a) Coppens, P.; Novozhilova, I.; Kovalevsky, A. Chem. ReV. 2002, 102,
861. (b) Kawano, M.; Ishikawa, A.; Morioka, Y.; Tomizawa, H.; Miki, E.;
Ohashi, Y. J. Chem. Soc., Dalton Trans. 2000, 2425.
(16) (a) Yasuda, N.; Kanazawa, M.; Uekusa, H.; Ohashi, Y. Chem. Lett. 2002,
31, 1132-1133. (b) Ozawa, Y.; Terashima, M.; Mitsumi, M.; Toriumi,
K.; Yasuda, N.; Uekusa, H.; Ohashi, Y. Chem. Lett. 2003, 32, 62. (c) Kim,
C. D.; Pillet, S.; Wu, G.; Fullagar, W. K.; Coppens, P. Acta Crystallogr.
2002, A58, 133. (d) Novozhilova, I. V.; Volkov, A. V.; Coppens, P. J. Am.
Chem. Soc. 2003, 125, 1079. (e) Yasuda, N.; Uekusa, H.; Ohashi, Y. Bull.
Chem. Soc. Jpn. 2004, 77, 933.
(17) Novozhilova, I. V.; Volkov, A. V.; Coppens, P. Inorg. Chem. 2004, 43,
2299.
(18) For a general aspect of the reactions of diarylcarbenes, see: (a) Moss, R.
A.; Jones, M., Jr. Carbenes; John Wiley and Sons, Inc.: New York, 1973;
Vol. 1, pp 73-95. (b) Wentrup, C. In Methoden der Organischen Chemic
(Houben-Weyl); Regitz, M., Ed.; Thieme: Stuttgart, 1989; Vol. E19, pp
824-1021.
(19) Regitz, M.; Maas, G. Diazo Compounds. Properties and Syntheses;
Academic Press: Orlando, FL, 1986.
(20) Kawano, M.; Hirai, K.; Tomioka, H.; Ohashi, Y. J. Am. Chem. Soc. 2001,
123, 6904.
(21) (a) Tomioka, H.; Watanabe, T.; Hirai, K.; Furukawa, K.; Takui, T.; Itoh,
K. J. Am. Chem. Soc. 1995, 117, 6376. (b) Tomioka, H.; Hattori, M.; Hirai,
K.; Murata, S. J. Am. Chem. Soc. 1996, 118, 8723. (c) Tomioka, H.;
Watanabe, T.; Hattori, M.; Nomura, N.; Hirai, K. J. Am. Chem. Soc. 2002,
124, 474. (d) Hirai, K.; Iikubo, T.; Tomioka, H. Chem. Lett. 2002, 1226.
9
2384 J. AM. CHEM. SOC. VOL. 129, NO. 8, 2007

Structure Determination of Triplet Diphenylcarbenes
A R T I C L E S
Chart 1
of which can generate persistent triplet DPCs, but their lifetimes
span from several tens of milliseconds to several tens of seconds
in solution at room temperature. Thus, one will be able to obtain
direct information on how the stability of DPCs is affected by
the substituents if one compares the structural parameters
obtained from X-ray crystallographic analysis. We also used
diaryldiazomethanes having no ortho substituents, i.e., bis(4-
bromophenyl)diazomethane (6-N2)23a and diazofluorene (7-
N2),23b which are not expected to generate persistent triplet
carbenes, in order to check the versatility of the method and, it
is hoped, to obtain the molecular structures of triplet DPCs not
perturbed by kinetic protectors. We found remarkable differ-
ences between the structural information obtained in the
crystalline state in this work and that obtained previously,
spectroscopically, in a glass matrix.
calculations can reproduce the experimental structural param-
eters determined by X-ray crystallography. The selected bond
lengths and angles are displayed in Table 1. Comparison of the
calculated and experimental data indicates that the calculated
values for the diazo carbon C-C-C angle (θ) and the C-C
distances (d1, d2) between aromatic and diazo carbons and the
distances (d3, d4) between C-N1 and N1-N2 are in fair
agreement with the observed ones (<(1.4° for θ and (1% for
d values). On the other hand, the difference between the
calculated and observed values for the interplanar angle (ω) is
slightly larger, being <(3% for DDMs having ortho substituents
and increasing to 10% for DDM having no ortho substituent
(6-N2). This can be interpreted as indicating that this angle is
sensitive to the crystal environment, presumably due to a shallow
energy surface along the rotation of the phenyl rings along the
C(dN2)sAr bond. These results indicate that DFT methods can
reproduce the diazo carbon C-C-C angle (θ) and the C-C
distances (d1-d4) in the crystalline state reasonably well.
Characteristic differences in the crystal packing among these
DDMs should also be noted here in connection with the
photochemical reactivity of the crystal sample and the molecular
structures of the triplet carbenes to be generated within the
crystal (vide infra). Generally speaking, the molecules are three-
dimensionally connected with two kinds of contacts, i.e., π-π
stacking interaction between phenyl rings and dipolar interaction
between positively charged nitrogen atoms of the diazo group
and negatively charged halogen atoms (only for halogenated
DDMs).
Results and Discussion
X-ray Crystal Structures of Precursor Diazo Compounds.
In order to analyze the crystal structures of triplet DPCs
generated in a crystal of DDMs, the crystal structures of the
DDMs themselves must be clarified in connection with the
photochemical reactivity of the crystal sample and the molecular
structures of the triplet carbenes to be generated within the
crystal.
Crystals suitable for X-ray analysis were obtained for all
seven DDMs shown in Chart 1. Crystallographic information
on the DDMs studied in this work is provided in the Supporting
Information (Chart 2).24
We carried out DFT calculations (RB3LYP/6-31G*) for all
DDMs employed in order to examine how well theoretical
The interplanar distances (due to π-π stacking interaction)
are very similar for 1-N2, 2-N2, 3-N2, and 6-N2 (3.5-3.6 Å),
where the stacking is observed. However, for DDMs having
tert-butyl groups, i.e., 4-N2 and 5-N2, π-π stacking interaction
between phenyl rings is not seen, probably due to the bulkiness
of the butyl group. In diazofluorene 7-N2, the molecules are
tightly packed without π-π stacking. On the other hand, the
distances N‚‚‚Br (due to dipolar interaction) range from 3.1 Å
(22) (a) Zimmerman, H. E.; Paskovich, D. H. J. Am. Chem. Soc. 1964, 86, 2149.
(b) Tomioka, H.; Hirai, K.; Fujii, C. Acta Chem. Scand. 1993, 46, 680. (c)
Tomioka, H.; Hirai, K.; Nakayama, T. J. Am. Chem. Soc. 1993, 115, 1285.
(23) (a) Staudinger, H.; Goldstein, J. Chem. Ber. 1916, 49, 1923. (b) Wright,
B. B.; Platz, M. S. J. Am. Chem. Soc. 1984, 106, 4175.
(24) See the following reference as well: Iikubo, T.; Itoh, T.; Hirai, K.;
Takahashi, Y.; Kawano, M.; Ohashi, Y.; Tomioka, H. Eur. J. Org. Chem.
2004, 3004.
9
J. AM. CHEM. SOC. VOL. 129, NO. 8, 2007 2385

A R T I C L E S
Kawano et al.
d4, Å
a
Table 1. Selected Bond Lengths and Angles for 1-N2 to 7-N2
θ
, deg
Ar
dihedral angle
Ar‚‚‚Ar
ω, deg
d1, Å
Ar
d2, Å
Ar
d3, Å
Ar−
C
−
−
C
C
−
Cd
N2
NdN
1-N2
2-N2
3-N2
4-N2
5-N2
127.1(1)
126.3
87.09 (2)
88.93
1.480(1)
1.480
1.477(1)
1.480
1.326(1)
1.313
1.141(1)
1.140
127.6(1)
126.3
86.45(4)
85.55
1.470(2)
1.480
1.479(2)
1.480
1.329(2)
1.311
1.135(2)
1.140
125.6(1)
126.2
82.46(5)
85.41
1.480(2)
1.481
1.479(2)
1.481
1.319(2)
1.310
1.137(2)
1.141
127.0(1)
126.2
87.74(2)
84.69
1.480(1)
1.481
1.481(2)b
1.476(2) b
1.479 b
1.479(1)
1.481
1.491(2)c
1.490(2)c
1.491
1.324(1)
1.310
1.137(1)
1.141
molecule 1
molecule 2
125.8(1)
125.4(1)
126.9
85.36(5)
82.95(4)
88.77
1.314(2)
1.318(2)
1.310
1.143(2)
1.140(2)
1.145
6-N2
7-N2
127.9(1)
126.5
46.61(3)
52.32
1.464(2)
1.477
1.484(2)
1.477
1.314(2)
1.311
1.146(2)
1.146
molecule 1
molecule 2
109.5(2)
109.4(2)
109.0
1.36(13)
1.12(10)
0.00
1.460(3)
1.455(3)
1.461
1.460(3)
1.451(3)
1.461
1.317(3)
1.322(3)
1.303
1.134(3)
1.133(3)
1.145
a The values shown in italics are those calculated at the RB3LYP/6-31G* level of theory. b Ar(PhBr3)-C bond distance. c Ar(PhMe2-t-Bu)-C bond
distance.
Chart 2
was possible only for DDMs 1-N2, 2-N2, 4-N2, and 5-N2.
Reliable crystal data were obtained for the corresponding DPCs,
1, 2 and 4, while the crystallinity of the irradiated sample of
5-N2 was not good enough to characterize the structure of 5.
X-ray Crystal Structures of DPCs. The ORTEP views of
DPCs 2 and 4 are shown in Figures 1 and 2, respectively, and
crystallographic data are presented in Table 2 (the crystal
structure of 1 was reported in ref 20). Selected bond angles
and bond lengths for the main framework of DPCs are listed in
Table 3, where θ is the carbene carbon C-C-C angle (°) and
ω the interplanar angle (°) between the two phenyl rings, while
d1 and d2 are the C-C distances (Å) between the aromatic and
carbene carbons, respectively. Table 3 includes the data for 1
for comparison.
(for 2-N2) to 3.4 Å (for 5-N2). In 3-N2, there is no special short
contact between N and Br (>3.5 Å).
The reaction cavity volume about a diazo group surrounded
by the spheres whose centers and radii are the neighboring
atomic positions and the van der Waals radii plus 1.2 Å,
respectively, is the smallest (1.2 Å3) for 3-N2, followed by 6-N2
and 7-N2, while 4-N2 has the largest volume (2.9 Å3). DDMs
having no protecting group around the diazo carbon, i.e., 6-N2
and 7-N2, have a more planar structure and hence are tightly
packed.
In order to confirm that those carbenes generated in the
crystals of precursor diazo compounds are indeed in the triplet
In Situ Observation of DPCs Induced by Irradiation of
Single Crystals of DDMs at Low Temperature. Photolyses
of single crystals of DDMs (1-N2 to 7-N2) were performed with
slow rotation about the φ axis at 80 K using a high-pressure
mercury lamp. Low temperature, close to the boiling point of
liquid nitrogen, is essential to suppress thermal motion of
dinitrogen molecules trapped in a crystal. Photolysis with
unfiltered light from the lamp resulted in a gradual decay of
the crystal, partly because photoirradiation with all emission
lines from a high-pressure mercury lamp induced dimerization
of carbenes, which caused large geometrical changes in the
crystal. Therefore, the wavelength of the irradiating light was
carefully selected by using appropriate band path filters to
suppress crystal decay (see the Experimental Section).
The crystallinity was retained in the case of irradiation of
DDMs 1-N2, 2-N2, 4-N2, and 5-N2 with a band path filter, while
3-N2 and 6-N2 showed deterioration upon irradiation. Interest-
ingly, no decomposition was observed for 7-N2, even upon
prolonged irradiation. Therefore, in situ crystallographic analysis
Figure 1. Thermal ellipsoid (50%) plots: (A) a disordered structure of
bis(2,4,6-tribromophenyl)carbene 2 (open line) and photoinduced carbene
(solid line), and (B) without the initial structure of 2-N2.
9
2386 J. AM. CHEM. SOC. VOL. 129, NO. 8, 2007

Structure Determination of Triplet Diphenylcarbenes
A R T I C L E S
Table 3. Selected Geometrical Parameters for Triplet Carbenes
31, 32, and 34, Obtained by X-ray Analysis and Theoretical
Calculationsa
θ
, deg
dihedral angle
ω, deg
d1, Å
Ar
d2, Å
Ar
Ar
−
C
−Ar
Ar‚‚‚Ar
−
C
C
−
31
32
34
142(2)
160.0
140.5
86.4(4)
80.79
81.33
1.423(16)
1.375
1.431
1.437(15)
1.375
1.431
141(2)
157.1
141.4
87.2(4)
68.75
83.93
1.431(18)
1.379
1.430
1.425(18)
1.379
1.430
138(1)
154.7
142.0
89.9(3)
69.46
84.00
1.375(17)
1.383
1.422
1.373(17)
1.382
1.422
a The values listed in the first row for each carbenes are those obtained
by X-ray analysis, while those in the second and third rows are estimated
by DFT UB3LYP/6-31G* and CASSCF(6,8)/6-31G* theoretical calcula-
tions, respectively.
Table 4. ESR Zero-Field Splitting Parameters (D and E)a and
Kinetic Datab Observed for Triplet Diphenylcarbenes
c
d
e
D
(cm-
E
(cm-
t1
kCHD
kO
2
/
2
1
1
1
1
)
)
E/D
(s)
(M-1 s-
)
(M-1 s-
)
31
22
34
0.371
(0.360) (0.011)
0.389
0.013
0.036
(0.029)
0.026
0.018 3.5 × 103 7.4 × 107
Figure 2. Thermal ellipsoid (50%) plots: (A) a disordered structure of
bis(2,6-dibromo-4-tert-butylphenyl)carbene 4 (open line) and photoinduced
carbene (solid line), and (B) without the initial structure of 4-N2.
0.010
0.396
0.030
0.075
1
7.4 × 102 1.1 × 107
5.3 × 102 2.3 × 107
(0.368) (0.0003) (0.0007)
0.369
0.015
0.042
3
3
Table 2. Crystallographic Data for Carbenes 2 and 4
0.397
(0.442)
0.423
0.031
(∼0)
0.078
(∼0)
16
32
34
after 16 min irrad
of 2-N2
after 10 min irrad
of 4-N2
0.030
0.070
Co39B filter
Co39B filter
a See text for the relationship between the ZFS parameters and the
structures of the triplet carbenes. In 2-methyltetrahydrofuran at 77 K. The
values in parentheses and in italics refer to those observed in 3-methyl-
pentane at 77 K and in the crystalline state at 20 K, respectively. b In
degassed benzene at room temperature. c Half-life. d The rate constant for
H abstraction of carbene from 1,4-cyclohexadiene. e The rate constant for
quenching carbenes by oxygen.
chemical formula
Mr
crystal system
space group
temp, K
a, Å
C13H4N2Br6
667.64
monoclinic
P21/n
C21H22N2Br4
622.05
monoclinic
P21/n
80(2)
80(2)
9.9776(3)
13.0960(4)
12.6356(4)
91.409(1)
1650.55(9)
4
13.9645(3)
10.1039(3)
17.5479(4)
112.625(1)
2285.4(1)
4
1.808
7.053
0.38 × 0.30 × 0.30
223 835
102
b, Å
c, Å
antiferromagnetic interactions. However, we also observed weak
signals ascribable to a triplet species with large D values (Figures
S7-S9). Zero-field splitting (ZFS) parameters estimated from
those signals reported in Table 4 are similar to those observed
in 2-methyltetrahydrofuran at 77 K for the corresponding triplet
carbenes.21,22 Thus, it is highly likely that carbenes generated
in the single crystals of DDMs are in the triplet ground states.
Theoretical calculations support this assignment (vide infra).
Inspection of the data in Table 3 reveals two interesting
features. First, there is only a subtle difference not only in the
angles (θ and ω) but also in the distances (d1 and d2) between
hexachlorinated (31) and hexabrominated DPCs (32), although
there are significant difference in bulkiness between their ortho
substituents (vide infra). Second, 34 exhibits slightly but clearly
different parameters for both angles and distances in comparison
to the other two DPCs, especially 32, which has the same ortho
substituents. Thus, on going from 32 and 31 to 34, θ and ω values
decrease from 141(2)°/87.2(4)° and 142(2)°/86.4(4)° to 138-
(1)°/89.9(3)°, and distances d1 and d2 also decrease from 1.431-
(18)/1.425(18) Å and 1.423(16)/1.437(15) Å to 1.375(17)/
1.373(17) Å, respectively. It is noteworthy that the average bond
â, deg
V, Å 3
Z
d
calcd, mg/m3
2.687
14.587
abs coeff, mm-1
crystal dimens, mm
no. of reflns measd
2θmax, deg
0.40 × 0.22 × 0.22
58 051
80
10 005
0.0609
0.0506
no. of symm-unique reflns
merging R(I)
merging R(sigm)
no. of retained reflns with
I > 2σ(I)
no. of variables
no. of restraints
R1(F) (I > 2σ(I))
wR2(F2)a (all data)
GOF
25 177
0.1080
0.0663
13 350
6758
247
13
0.0408
0.0848
1.054
298
46
0.0473
0.1082
0.991
residual electron density
max, e Å-3
1.452
-1.580
1.452
-2.189
min, e Å-3
2
2
2
a w ) 1/[σ2(Fo ) + (aP)2 + bP], where P ) [2Fc + max(Fo ,0)]/3.
ground state, we carried out ESR measurements. Irradiation of
single crystals of DDMs 1-N2, 2-N2, and 4-N2 gave rise to a
weak, broad signal around 334 mT. This has been occasionally
observed in crystalline samples of triplet carbenes, probably
because of strong exchange coupling of the triplet states and/or
3
length C(:)-Ar of 4 (1.374 Å) is considerably shorter than
those of 31 and 32 (1.430 and 1.428 Å, respectively), provided
that the two lengths of C(:)-Ar for each carbene were
chemically equivalent.
9
J. AM. CHEM. SOC. VOL. 129, NO. 8, 2007 2387

A R T I C L E S
Kawano et al.
Kinetics Data and ESR Parameters Observed for Triplet
DPCs. The structural and kinetic data concerning triplet DPCs,
31, 32, and 34, generated from diazo precursors are summarized
in Table 4.21,22 They are kinetic data measured by laser flash
photolysis (LFP) in a degassed benzene solution at room
temperature and zero-field splitting (ZFS) parameters measured
by electron spin resonance (ESR) spectroscopy in an organic
matrix at low temperatures.
The lifetime (or half-life) of triplet DPCs obtained provides
direct information about their stability in solution. The data in
Table 4 show two prominent trends in the effect of structure
on the stability of triplet DPCs.
3
First, 1 is the least stable among the three DPCs. The half-
life (t1/2) of 31 in benzene at room temperature is 0.018 s, 2 and
3 orders of magnitude smaller than those of 32 (t1/2 ) 1 s) and
34 (t1/2 ) 16 s), respectively. But half-life is just a measure of
lifetime and cannot be regarded as a quantitative scale for
reactivity. In this respect, the rate constant of the triplet carbene,
measured with a typical triplet quencher, can be employed as a
more quantitative scale of the reactivity. It is well documented
that carbenes with triplet ground states are readily trapped with
oxygen or a good hydrogen donor such as 1,4-cyclohexadiene
(CHD).25 Therefore, the rate constants of the trapping reactions
Figure 3. Potential energy surface of bis(2,4,6-tribromophenyl)carbene 2
as a function of the carbene bond angle (θ) in the range of 115-165°,
calculated at the UB3LYP/6-31G* level of theory.
The structures were then optimized for the triplet ground
states of 31, 32, and 34 at the UB3LYP/6-31G* levels of theory,
and selected bond lengths and angles are also given in Table 3.
UB3LYP methods give essentially identical C-Ar bond lengths
d (1.38 Å) for all three triplet carbenes. This length is smaller,
3
3
when compared with the observed values, for 1 and 2 (ca.
1.43(2) Å) but very close for 34 (1.37(2) Å). On the other hand,
the carbene angles (θ) are not identical for the three, decreasing
from 160° for 31 to 155° for 34. Those values are significantly
larger (by 17-18°) than the observed ones (142(2)-138(1)°),
although the trend of decreasing order is similar to the observed
one. The interplanar angles (ω) are also not identical for the
three and are much larger (by 5-21°) than the observed ones
(86(1)-90(1)°). The increasing of ω determined by UB3LYP
methods on going from o-chlorinated to o-brominated DPCs,
by O2 and CHD, i.e., kO and kCHD, respectively, are used as a
2
more quantitative scale to estimate the reactivities of the triplet
carbenes. These values again show that 31 is more reactive than
3
32 and 4. This is interpreted in terms of the difference in van
der Waals radius between chlorine and bromine groups. In other
words, the carbene center in 32 and 34 is shielded more tightly
3
than that in 1.
Second, among o-brominated DPCs, i.e., 32 and 34, the half-
life is significantly increased once the bromine groups at the
para position are replaced with tert-butyl groups. However, the
3
3
3
i.e., from 1 to 2 and 4, is reversed from the observed trend.
3
3
rate constants for quenching of 2 and 4 by O2 and CHD are
not affected by this change. Similarly, there is little difference
in the ESR structural parameters between 32 and 34 (vide infra).
Therefore, this difference can be ascribed not to structural
changes in the carbene itself but to the effect of the para
substituents on the decay pathway of the two DPCs in benzene.
The main decay pathway of those DPCs in benzene is
dimerization at the carbene center, but, in o-brominated DPCs,
dimerization at the para positions becomes significant due to
severe steric hindrance at the carbene center.26 Therefore, a
significant increase in the lifetime upon substitution of a tert-
butyl group in benzene (in the absence of a trapping reagent) is
interpreted in terms of steric hindrance of the dimerization at
the para position.27
We also calculated a potential energy surface of 2 as a
function of θ for both singlet and triplet states in the range of
115-165° in order to explore further the assignment of the spin
state of 2 in the crystal. The results (Figure 3) clearly indicate
3
1
that 2 is always more stable than 2. Especially important is
the fact that the singlet-triplet energy gap at an angle of 140°
is 47.7 kJ/mol, which is large enough to ensure that the spin
state could be the triplet state. It is also interesting to note here
that both 12 and 32 have very shallow potential energy surfaces.
Since DFT is not able to treat multiconfigurational problems
properly,8 an improved method to take some type of configu-
rational interaction (CI) into account is required in order to
optimize the structures of these open-shell species more
adequately. Thus, the structures are also optimized at CASSCF/
3
3
3
ZFS parameters D and E give information on the molecular
and electronic structures of triplet carbenes.28,29 The D value is
related to the separation between the unpaired electrons. The E
6-31G* levels of theory for the ground state of 1, 2, and 4,
and selected bond lengths and angles are given in Table 3.
Interestingly, the CASSCF(6,8) method gives identical C-Ar
lengths of 1.43 Å and angles of 141° for all three carbenes,
and these values are in better agreement with the observed ones
than those predicted by DFT, although one should perform the
CASSCF calculations with a larger basis set for more accurate
comparison. The interplanar angles (ω) are not identical for the
three and are smaller than the observed ones (86-90°), but the
discrepancy (by 6-20°) is smaller than that predicted by DFT.
(25) Tomioka, H. In ReactiVe Intermediate Chemistry; Moss, A. M., Platz, M.
S., Jones, M., Jr., Eds.; Wiley: New York, 2004; pp 375-461.
(26) Oligomerization from the aromatic ring in the reaction of sterically
congested diphenylcarbenes usually resulted in complex mixtures, but the
formation of trimer and tetramer as a result of oligomerization from aromatic
ring is noted for bis(9-anthryl)carbene: (a) Takahashi, Y.; Tomura, M.;
Yoshida, K.; Murata, S.; Tomioka, H. Angew. Chem., Int. Ed. 2000, 39,
3478. (b) Yoshida, K.; Iiba, E.; Nozaki, Y.; Hirai, K.; Takahashi, Y.; Tomi-
oka, H.; Lin, C-T.; Gaspar, P. P. Bull. Chem. Soc. Jpn. 2004, 77, 1509.
(27) The lifetime of bis(9-anthryl)carbenes is greatly increased by introducing
a bulky group at the 9 position, where oligomerization takes place: (a)
Tomioka, H.; Iwamoto, E.; Itakura, H.; Hirai, K. Nature 2001, 412, 626.
(b) Iwamoto, E.; Hirai, K.; Tomioka, H. J. Am. Chem. Soc. 2003, 125,
14664.
Before discussing the structural differences in terms of the
stability, we would like to summarize what has been observed
for three carbenes by “indirect” methods.
9
2388 J. AM. CHEM. SOC. VOL. 129, NO. 8, 2007

Structure Determination of Triplet Diphenylcarbenes
A R T I C L E S
value, on the other hand, when weighted by D, is a measure of
the deviation from axial symmetry. For diarylcarbenes, this value
will thus depend on the magnitude of the central C-C-C angle.
More plainly, the more the two electrons are delocalized in
carbenes with a conjugated π-system, the smaller the value of
the repulsive interaction D will be. On the other hand, increasing
the bond angle at the carbene center leads to a higher π-orbital
contribution and a smaller value for E. Although the values D
and E depend on the electronic distribution, there is a good
correlation between the E/D ratio and the bond angle at the
divalent carbon atom.
significant difference in stability in solution at room temperature
(t1/2 ) 0.018 s for 31 and 1 s for 32) and the ESR parameters in
3
a matrix at low temperatures (E/D ) 0.0133/0.371 for 1 and
3
0.0297/0.396 for 2) between two DPCs.
Bromine has a larger van der Waals radius (1.85 Å) than
chlorine (1.70 Å), but the bond distance of C-Br (1.95 Å) is
longer than that of C-Cl (1.80 Å).32 This may mean that o-Br
groups can hang over the carbene center better than o-Cl groups
without interacting with each other. However, the extent of
geometrical relaxation of triplet DPC upon softening of the
3
3
matrix is larger for 2 than for 1, as judged by comparing the
extent of change in E/D upon warming between the two DPCs.
This means that steric repulsion between o-bromine groups in
32 is more severe than that between o-chlorine groups in 31. A
significant decrease in the reactivity in solution is clearly noted
ZFS parameters in 2-methyltetrahydrofuran at 77 K show that
3
the D values of 31 are slightly smaller than those of 2 and 34,
3
3
while the E/D values of 1 are much smaller than those of 2
3
3
and 4. This suggests that 1 has a less bent structure where
3
3
3
unpaired electrons are more delocalized. In other words, 1 is
on going from 1 to 2 and is interpreted as reflecting this
difference. Thus, it is rather surprising to note that the structural
parameters between the two carbenes are unexpectedly small.
Second, significantly smaller values for both angles and
distances are observed for 34 as opposed to those observed for
thermodynamically more stable than 32 and 34. This prediction,
however, is not in accord with the kinetic data obtained in
solution at room temperature.
ZFS parameters of sterically congested DPCs generated in a
rigid matrix at low temperature are known to change when the
matrix is warmed, as observed for the three DPCs in Table 4.
These changes are interpreted in terms of geometrical relaxation
of triplet DPC as the matrix is softened.30,31 Thus, when a
carbene is generated in rigid matrices at low temperatures, its
initial geometry should be dictated by that of the precursor. Even
if the thermodynamically most stable geometry of the carbene
is different from that at birth, the rigidity of the matrices prevents
the carbene from achieving its minimum-energy geometry. But
when the matrix is softened by warming, the carbene is allowed
to relax to its preferred geometry, probably to gain relief from
steric compression.
3
the other two DPCs, especially for 2. A decrease in the bond
3
3
3
3
length in 4 as opposed to 2 (and 1) may indicate that 4 is
3
3
thermodynamically more stabilized than 2 (and 1).
3
3
Since the only difference in structure between 4 and 2 is
the substituents at the para position, one needs to examine the
difference in effect of p-bromine vs p-tert-butyl groups on the
stability of triplet DPCs.
The dominant interaction of the unpaired electrons in a
π-radical is with the electrons paired in the π-bonds. Such
interactions are characterized by the delocalization of the spin
throughout the π-system. To estimate the relative abilities of
substituents to delocalize the spin, sigma-dot substituent con-
stants (σ‚) have been proposed.33
Attempts to correlate the D values with σa‚ have been carried
out for bis(4-X-2,6-dimethylphenyl)carbene (8),34 where a series
of X substituents with different σ‚ values are employed. The D
values observed for carbenes 8 in their minimum-energy
geometries, attained upon annealing, are found to correlate
relatively well (F ) 0.589, r ) 0.9) with σa‚.35
The data in Table 4 indicate that this change is larger for 32
and 34 than for 31. Thus, 32 and 34 undergo relaxation to nearly
linear geometry, as judged from their nearly zero E/D values,
3
while 1 has still a bent structure in its relaxed geometry. In
other words, 32 and 34 are considered to be thermodynamically
3
more stable than 1 in their relaxed geometries. Since it is
probable that those DPCs react in their relaxed geometry in
solution at room temperature, ZFS parameters obtained in a
softened matrix correlate with kinetic data in solution better
than those obtained in a rigid matrix.
Structural Parameters and Reactivity. How can one
correlate those data from ESR and LFP with the structural
features noted by in situ crystallographic observation of three
triplet DPCs?
However, the ability of the para substituents to delocalize
3
3
unpaired electrons is essentially identical for 2 and 4 (σC‚ )
0.13 for both 4-Br and 4-t-Bu).36 Thus, one cannot explain the
difference in the geometry observed between the two in terms
of delocalization of unpaired electrons.
The difference in bulkiness between bromine and tert-butyl
groups is significant. However, it is unlikely that the tert-butyl
group is large enough to effect steric congestion at the remote
carbene center.
First, the rather subtle difference in structural parameters
between 31 and 32 in Table 3 is not in accord with the observed
(28) A large number of triplet carbenes have been characterized by ESR
spectroscopy since the pioneering work of Wasserman et al.: Wasserman,
E.; Hutton, R. S. Acc. Chem. Res. 1997, 10, 27. Wasserman, E.; Snyder,
L. C.; Yager, W. A. J. Chem. Phys. 1964, 41, 1763.
(29) For reviews of the EPR spectra of triplet carbenes, see: (a) Sander, W.;
Bucher, G.; Wierlacher, S. Chem. ReV. 1993, 93, 1583. (b) Trozzolo, A.
M.; Wasserman, E. In Carbenes; Jones, M., Jr., Moss, R. A., Eds.; Wiley:
New York, 1975; Vol. 2, pp 185-206.
(30) (a) Tukada, H.; Sugawara, T.; Murata, S.; Iwamura, H. Tetrahedron Lett.
1986, 27, 235. (b) Nazran, A. S.; Gabe, F. J.; LePage, Y.; Northcott, D. J.;
Park, J. M.; Griller, D. J. Am. Chem. Soc. 1983, 105, 2912. (c) Nazran, A.
S.; Griller, D. J. Chem. Soc., Chem. Commun. 1983, 850. (d) Gilbert, B.
C.; Griller, D.; Nazran, A. S. J. Org. Chem. 1985, 50, 4738. (e) Nazran,
A. S.; Lee, F. L; LePage, Y.; Northcott, D. J.; Park, J. M.; Griller, D. J.
Phys. Chem. 1984, 88, 5251.
(31) For a review, see: Tomioka, H. In AdVances in Strained and Interesting
Organic Molecules; Halton, B., Ed.; JAI Press: Greenwich, CT, 2000; Vol.
8, pp 83-112.
ZFS parameters and rate constants for quenching by triplet
carbene trapping reagents suggest that the difference in half-
life between 2 and 4 in benzene (in the absence of trapping
3
3
(32) Rowland, R. S.; Taylor, R. J. Phys. Chem. 1996, 100, 7384.
(33) For a review, see: Jiang, X.-K. Acc. Chem. Res. 1997, 30, 283.
(34) (a) Dust, J. M.; Arnold, D. R. J. Am. Chem. Soc. 1983, 105, 1221 and
6531. (b) Wayner, D. D. M.; Arnold, D. R. Can. J. Chem. 1984, 62, 1164
and 1985, 63, 2378.
(35) (a) Tomioka, H.; Hu, Y.; Ishikawa, Y.; Hirai, K. Bull. Chem. Soc. Jpn.
2001, 74, 2207. (b) Hu, Y.; Hirai, K.; Tomioka, H. J. Phys. Chem. A 1999,
103, 9280.
(36) Creary’s σC‚ scale is also useful, since it is based on the thermal
rearrangement that is devoid of polar character in the transition state: (a)
Creary, X. J. Org. Chem. 1980, 45, 280. (b) Creary, X. Mehrsheik-
Mohammadi, M. E.; McDonald, S. J. Org. Chem. 1987, 52, 3254.
9
J. AM. CHEM. SOC. VOL. 129, NO. 8, 2007 2389

A R T I C L E S
Kawano et al.
reagents) is explained in terms of steric inhibition of the decay
pathway from the aromatic ring rather than an inherent increase
in kinetic protection of the carbene center (vide supra).
It is interesting to compare ZFS parameters observed in an
organic glass (both before and after warming) and in the crystal-
line state (Table 4). The most prominent difference appears in
the E values, especially for brominated DPCs, 32 and 34. The E
values observed for both carbenes in the crystalline state are
significant (0.015 for 2 and 0.030 for 4), while they reduced
from significant (0.030 for 32 and 0.031 for 34) in a rigid matrix
to nearly zero (0.0003 for 32 and ∼0 for 34) in a softened matrix.
The significant E values in the crystalline state suggest bent
structures for the three triplet DPCs, in accord with the structures
directly observed by the present method. On the other hand,
triplet DPCs undergo relaxation to less bent and probably more
stable geometry in soft, unconstrained environments like in a
solution phase before they react. In other words, the structures
of triplet DPCs observed in the crystalline state are not exactly
the same as the ones reacting in solution.
Among DDMs that undergo photodissociation of nitrogen in
the crystalline state, 3-N2 and 6-N2 decomposed without keeping
their crystallinity, while other DDMs retained their crystallinity
at least to an extent that allowed us to observe the corresponding
carbene. It will not be easy to find a reliable factor to control
this difference, since this will be affected by differences in the
crystal form between the precursor diazo compound and the
product carbene and also in the relative easiness of subsequent
reaction of the carbene generated in the crystal.
3
3
Concluding Remarks
The structural parameters of three persistent triplet diphen-
ylcarbenes with significantly different stabilities are obtained
by in situ X-ray crystallographic observations. Surprisingly
subtle differences in structural parameters (except for carbene-
phenyl group bond length) are observed for those three triplet
DPCs, the half-lives of which span from 0.02 to 18 s in solution
at room temperature. Noteworthy is that the average bond length
3
of carbene-phenyl group for 4 is considerably shorter than
those for 31 and 32. Such geometrical parameters can be obtained
only by crystallographic analysis. At this moment, the most
likely explanation for the small and large differences in bond
lengths in 1, 2, and 4 may be derived from the different packing
patterns of these DPCs. The packing patterns of 1 and 2 are
identical, but that of 4 is totally different. Therefore, their
intermolecular interactions should be quite different. In order
to determine the crucial factor in the differences, we should
investigate these DPCs under various conditions. Moreover,
these rather unexpected results are interpreted as indicating that
triplet DPCs undergo relaxation upon softening of the environ-
ment, since the carbenic centers are sterically congested before
they decay in solution.
Theoretical calculations provide a clue toward understanding
the situations in the structure observed by the present in situ
method and those presumed by the traditional means. The
potential energy surface of the triplet DPC in terms of the
carbene angle is extremely flat, and changes in the angles have
little effect on the energies. Only 8.1 kJ/mol is required to
linearize the triplet DPC from its 142° equilibrium geometry.41
This will be much lower for DPCs having sterically congested
substituents at their ortho positions. In fact, a potential energy
surface calculation for 31 and 32 indicates that the energy
difference between 140° and 160° is only 4.2 kJ/mol.20 This
suggests that experimental θ values will be readily affected by
the rigidity of the environment. This is reasonable in the light
of the fact that triplet DPCs with a sterically congested carbene
center are trapped in a structure dictated by the precursor
structure in a rigid matrix, even if this is not the thermodynam-
ically most stable geometry, but undergo geometrical relaxation
upon softening of the matrix to gain relief from steric compres-
sion.
Origin of the Difference in Crystalline-State Photorea-
civity of DDMs. The observation that diazofluorene 7-N2 did
not decompose upon irradiation in the crystalline state is rather
surprising in light of the fact that most diazo compounds undergo
smooth and efficient photodissociation of molecular nitrogen
not only in solution at room temperature but also in a rigid
matrix at very low temperature. Diazofluorene, for instance,
efficiently generates the corresponding carbene when irradiated
in an argon matrix at temperature as low as 10 K.37 It is very
important to examine the origin of the sluggishness of 7-N2 to
photodissociation in the crystalline state.
The Ar-C distances for 7-N2 are shorter than those of the
other DDMs, but the CdN2 and NdN bond lengths in the diazo
functional groups are comparable with other ones. Thus, it is
unlikely that the inherent reactivity of the diazo group is forced
to change in the crystalline state. The reaction cavity volume38
about a diazo group in the crystal of 7-N2 is not smallest among
the DDMs studied and is slightly larger than that of the smallest
one, 3-N2. Since 3-N2 undergoes smooth photodissociation of
nitrogen under these conditions, the cavity seems to not be a
crucial factor controlling the photoreactivity of DDMs. On the
other hand, the molecules are most tightly packed in the 7-N2
crystal among the DDMs employed, as judged by the smallest
non-hydrogen atomic volume. This volume is pretty large for
3-N2, which has the smallest reaction cavity but undergoes
photodissociation. Thus, it seems that nitrogen cannot easily
leave the procarbenic carbon in the crystal of 7-N2. Moreover,
since fluorenylidene 7 is incorporated in a five-membered ring,
its singlet-triplet energy gap is much smaller than that of the
diphenylcarbene derivative of 7,39 so that nascent singlet 7
generated upon photodissociation of nitrogen undergoes relax-
ation to the corresponding triplet state less reluctantly than
singlet diphenylcarbene derivatives do, especially in the crystal-
line state. It is then not unreasonable to assume that, even if 7
is generated upon irradiation of 7-N2 in a crystal, the nascent
singlet state of 7 is immediately trapped by nitrogen, which is
tightly trapped near the carbene center, to reproduce 7-N2.40
(40) It has been shown that some carbenes react with nitrogen to form the
corresponding diazomethanes: (a) Moore, C. B.; Pimentel, G. C. J. Phys.
Chem. 1964, 41, 3504. (b) O’Gara, J. E.; Dailey, W. P. J. Am. Chem. Soc.
1992, 114, 3582.
(41) We recalculated the potential energy surface of a triplet diphenyl carbene
for carbene angle in the range of 115-179.9° at the UB3LYP/6-31G* level
of theory. The carbene angle in the optimized structure is 142.3°. See also
the Supporting Information and the following references: (a) Hoffman,
R.; Zeiss, G. D.; van Dine, G. W. J. Am. Chem. Soc. 1968, 90, 1485. (b)
Metcalf, J.; Halevi, E. A. J. Chem. Soc., Perkin Trans. 2 1977, 634. (c)
Dannenberg, J. J.; Vinson, L. K.; Moreno, M.; Bertran, J. J. Org. Chem.
1989, 54, 5487. (d) Sulzbach, H. M.; Bolton, E.; Lenoir, D.; Schleyer, P.
v. R.; Schaefer, H. F., III. J. Am. Chem. Soc. 1996, 118, 9908.
(37) Bell, G. A.; Dunkin, I. R.; Shields, C. J. Spectrochim. Acta 1985, 41A,
1221.
(38) Ohashi, Y.; Yanagi, K.; Kurihara, T.; Sasada, Y.; Ohgo, Y. J. Am. Chem.
Soc. 1981, 103, 5805.
(39) Schuster, G. B. AdV. Phys. Org. Chem. 1986, 22, 311.
9
2390 J. AM. CHEM. SOC. VOL. 129, NO. 8, 2007

Structure Determination of Triplet Diphenylcarbenes
A R T I C L E S
corrected with the SADABS program. Because polybrominated diphen-
yldiazomethanes react with Mo KR radiation to emit X-ray fluorescence
and yield carbene, the minimum exposure time should be used for data
collection.
It is important to point out here that recent ESR studies
revealed that, upon warming the matrix containing 1, the D
3
value decreases first, which is followed by a decrease in the E
value.42 This suggests that the phenyl ring rotates along the
C(:)-Ar bond at the initial stage, while the matrix is still not
highly fluid, and then the bond angle on the carbene center starts
to open at the latter stage, when the matrix become more fluid.
This means that the interplanar angles are more flexible than
the bond angle.
(c) Structure Refinement. All structures were solved by use of the
direct method program XS, which is part of the SHELXTL program
package.44 Structure refinements were carried out with SHELXL-97.43
2
All least-squares refinements minimized the function ∑w(|Fo|
-
k|Fc| )2, where w ) 1/[σ2(Fo ) + (aP)2 + bP] and P ) [2Fc2 + max-
2
2
2
(Fo ,0)]/3. Crystallographic data before and after irradiation are
In situ X-ray crystallographic analysis can be successfully
employed to determine the molecular structure of triplet DPCs,
which is otherwise impossible to analyze. However, the
structural parameters obtained by this method are not in accord
with those of the triplet DPCs that are involved in the reaction,
mainly because triplet DPCs have very flat potential energy
surfaces, and hence solvation effects cannot be ignored. This
mobile complex nature of triplet carbenes reflects on their unique
reactivity.
summarized in Tables S1 and 2.
The initial structure was treated as a rigid group and allowed to
translate and rotate during the least-squares refinement, in which its
thermal parameters were refined anisotropically. The structure was
solved as a disordered structure. Chemical restraints were applied to
the phenyl rings of the divalent species because of their severe
overlapping with the diazo species.
Computational Procedures. Ab initio molecular orbital calcula-
tions45 were carried out using the GAUSSIAN 98 program.46 Optimized
geometries were obtained at the (U)B3LYP/6-31G*46,47 and CASSCF/
6-31G* levels of theory. Vibrational frequencies obtained at the (U)-
B3LYP level of theory were scaled by 0.97. For the DFT calculations,
UB3LYP wave functions were used to describe the open-shell triplet
carbenes. For triplet carbenes, the active space used in the CASSCF
calculations comprised six electrons in eight MOs (three π and four
π* MOs plus a σ one), denoted as CASSCF(6,8). The carbene structures
were optimized in C2 and Cs symmetry, both of which gave identical
structures. CASSCF calculations were performed using several com-
binations of the active space for the π MOs of the phenyl groups, except
for two electrons assigned to the carbene center. CASSCF(6,8)
calculations using each set of the aromatic π MOs resulted in similar
bond lengths between the carbene center and phenyl rings, but the C-C
bond distances in the phenyl rings depended considerably on the
combination of the π MOs of the phenyl groups. This is not surprising,
because we had to treat small active spaces for the phenyl rings because
of the limitation of computer resources. In this paper, therefore, we
discuss only the geometries about the carbene center.
Experimental Section
General Considerations. Compounds 1-N2, 6-N2, and 7-N2 were
prepared by the method described in refs 2, 22a, and 22b, respectively.
A series of brominated diphenyldiazomethanes (2-N2-5-N2) were
prepared by the method described in ref 20. Samples were irradiated
with a 350-W high-pressure mercury lamp (SAN-EI UVF-35S) in
combination with a band-path filter (Toshiba Co39B or UV-D36B).43
UV-vis spectra were recorded on a JASCO CT-560 spectrophotometer
in 2-methyltetrahydrofuran (2-MTHF) using an Oxford Instruments
variable-temperature liquid nitrogen cryostat (DN 2704) equipped with
a quartz outer window and a sapphire inner window and were also
recorded on a Shimadzu UV-3100PC spectrophotometer in a KBr pellet.
Fourier-transform infrared spectra were recorded on a JASCO FT/IR-
5300 instrument. A sample dispersed in a KBr pellet (several DPC
crystals were ground in a mortar) was attached to the coldfinger of a
helium cryogenic refrigerator system (Daikin PS24SS) equipped with
KBr or quartz windows. The presence of several micrometer-size
microcrystals in a KBr pellet was confirmed with an optical microscope.
ESR measurements were made on a JEOL TE-200 instrument equipped
with an Oxford Instruments ESR910 temperature controller.
Single-Crystal X-ray Diffraction Analysis. (a) Sample Prepara-
tion. Single crystals of 1-N2, 6-N2, and 7-N2 were obtained by slow
evaporation of benzene, that of 2-N2 from dichloromethane, and those
of 3-N2-5-N2 from acetone. Crystals were mounted on the tips of glass
fibers with epoxy or Paratone oil and cooled to low temperatures (80 K),
controlled by a Rigaku cryostat system equipped with a N2 generator.
The photoreaction and crystal quality were monitored on the basis of
total intensities of selected frames and shapes of Bragg diffractions.
Different crystals were used for the structure determination of the initial
molecules and photoproducts.
Acknowledgment. This work was supported by CREST from
JST and a Grant-in-Aid for Scientific Research for Specially
Promoted Research (No. 12002007) from the Ministry of
Education, Culture, Sports, Science and Technology of Japan.
We also thank Dr. Akira Miyazaki of Tokyo Institute of
Technology for ESR measurements.
Supporting Information Available: Crystallographic infor-
mation for DDMs (2-N2-7-N2), ESR spectra for 1, 2, and 4
measured in the crystalline state at 20 K (Figures S7-S9),
Cartesian coordinates, absolute energies for all compounds
calculated at the B3LYP/6-31G* level of theory, and complete
ref 46 (PDF); X-ray crystallographic files (CIF; also deposited
as CCDC 619889-619896). This material is available free of
charge via the Internet at http://pubs.acs.org.
(b) Data Collection and Reduction. Data were collected on a
Siemens SMART CCD X-ray diffractometry system controlled by a
Pentium-based PC running the SMART software package. Graphite-
monochromatized Mo KR (λ ) 0.71073 Å) radiation was used with a
Rigaku rotating anode generator (50 kV, 250 mA). A modification of
the data collection strategy described in ref 20 was utilized. Frame
data were integrated by using SAINT version 5, and intensity data were
JA067306B
(44) Sheldrick, G. M. SHELXL-97, Program for Crystal Structure Analysis,
UNIX version, Release 97-2; 1998.
(45) Hehre, W.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. Ab Initio Molecular
Orbital Theory; Wiley: New York, 1986.
(42) Makarov, B. P.; Tomioka, H. Org. Biomol. Chem. 2004, 2, 1834.
(43) The Co39B and UV-D36B filters gave identical spectroscopic results. A
high-pressure mercury lamp equipped with the Co39B filter provides λ )
365, 406, and 436 nm emission lines.
(46) Frisch, M. J.; et al. GAUSSIAN 98; Gaussian, Inc.: Pittsburgh, PA, 1998.
(47) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785. (b) Miehlich,
B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989, 157, 200. (c)
Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
9
J. AM. CHEM. SOC. VOL. 129, NO. 8, 2007 2391