Photoinduced C-N bond cleavage in 2-azido-1,3-diphenyl-propan-1-one derivatives: Photorelease of hydrazoic acid

Article abstract of DOI:10.1021/jo070558q

(Chemical Equation Presented) Photolysis of 3-azido-1,3-diphenyl-propan-1- one (1a) in toluene yields 1,3-diphenyl-propen-1-one (2), whereas irradiation of 3-azido-2,2-dimethyl-1,3-diphenyl-propan-1-one (1b) results in the formation of mainly 2,2-dimethyl-1,3-diphenyl-propan-1-one. Laser flash photolysis (308 nm) of 1a,b in acetonitrile reveals a transient absorption (λ_max = ～310 nm) due to the formation of radicals 4a and 4b, respectively, which have lifetimes of ～14 μs at ambient temperature. TD-DFT calculations (B3LYP/6-31+G(d)) reveal that the first and second excited states of the triplet ketone (T_1K (n,π*) and T_2K (π,π*)) in azide 1a are almost degenerate, at ～74 and 76 kcal/mol above the ground state (S₀), respectively. Similarly, azide 1b has T_1K and T₂K 75 and 82 kcal/mol above S₀, respectively. The calculated transition state for cleaving the C-N bond is located 71 and 74 kcal/mol above S₀ in azides 1a and 1b, respectively. The calculated bond dissociation energies for breaking the C-N bond are 55 and 58 kcal/mol for azides 1a and 1b, respectively, making C-N bond breakage accessible from T _1K in azides 1 at ambient temperature. In comparison, the irradiation of azides 1 in argon matrices at 14 K lead to the formation of the corresponding triplet alkyl nitrenes (1-n), via intramolecular energy transfer from T_2K- The characterization of 1-n was supported by isotope labeling, IR spectroscopy, and molecular modeling.

Full text of DOI:10.1021/jo070558q

Photoinduced C-N Bond Cleavage in
2-Azido-1,3-diphenyl-propan-1-one Derivatives: Photorelease of
Hydrazoic Acid
Rodney F. Klima, Abhijit V. Jadhav, Pradeep N. D. Singh, Mingxin Chang,
Christine Vanos, Jagadis Sankaranarayanan, Mai Vu, Nazarin Ibrahim, Elaine Ross,
Shaun McCloskey, Rajesh S. Murthy, Jeanette A. Krause, Bruce S. Ault, and
Anna D. Gudmundsdo´ttir*
Department of Chemistry, UniVersity of Cincinnati, Cincinnati, Ohio 45221-0172
Anna.Gudmundsdottir@uc.edu
ReceiVed March 17, 2007
Photolysis of 3-azido-1,3-diphenyl-propan-1-one (1a) in toluene yields 1,3-diphenyl-propen-1-one (2),
whereas irradiation of 3-azido-2,2-dimethyl-1,3-diphenyl-propan-1-one (1b) results in the formation of
mainly 2,2-dimethyl-1,3-diphenyl-propan-1-one. Laser flash photolysis (308 nm) of 1a,b in acetonitrile
reveals a transient absorption (λmax ) ∼310 nm) due to the formation of radicals 4a and 4b, respectively,
which have lifetimes of ∼14 µs at ambient temperature. TD-DFT calculations (B3LYP/6-31+G(d))
reveal that the first and second excited states of the triplet ketone (T1K (n,π*) and T2K (π,π*)) in azide
1a are almost degenerate, at ∼74 and 76 kcal/mol above the ground state (S0), respectively. Similarly,
azide 1b has T1K and T2K 75 and 82 kcal/mol above S0, respectively. The calculated transition state for
cleaving the C-N bond is located 71 and 74 kcal/mol above S0 in azides 1a and 1b, respectively. The
calculated bond dissociation energies for breaking the C-N bond are 55 and 58 kcal/mol for azides 1a
and 1b, respectively, making C-N bond breakage accessible from T1K in azides 1 at ambient temperature.
In comparison, the irradiation of azides 1 in argon matrices at 14 K lead to the formation of the
corresponding triplet alkyl nitrenes (1-n), via intramolecular energy transfer from T2K. The characterization
of 1-n was supported by isotope labeling, IR spectroscopy, and molecular modeling.
1. Introduction
it possible to generate hydrazoic acid under benign conditions,
without any heating. It should be noted that hydrazoic acid is
toxic and explosive in high concentrations, making it a treacher-
ous compound with which to work.3 Traditionally, hydrazoic
acid is prepared by heating sodium azide in excess acid.4
Photoprotecting or photocaging groups are used in a variety
of applications, including the slow release of fragrances in
household goods, synthesis of organic molecules, drug delivery,
and as probes of biological processes.1,2 Since photoprotecting
groups are so widely used, there is a need for new photore-
movable protecting groups that can be tailored to specific
applications.1 We have designed the photoprotecting group
2-azido-1,3-diphenyl-propan-1-one 1a and its derivative 1b,
which slowly release hydrazoic acid upon irradiation, making
(2) Some recent publications on photoremoveable protecting groups: (a)
Literak, J.; Dostalova, A.; Klan, P. J. Org. Chem. 2006, 71, 713. (b) Zhu,
Y.; Pavlos, C. M.; Toscano, J. P.; Dore, T. M. J. Am. Chem. Soc. 2006,
128, 4267. (c) Sundararajan, C.; Falvey, D. E. J. Am. Chem. Soc. 2005,
127, 8000. (d) Lage, R. J.; Bochet, C. G. Org. Lett. 2005, 7, 3545. (e)
Konosonoks, A.; Wright, P. J.; Tsao, M.-L.; Pika, J.; Novak, K.; Mandel,
S. M.; Krause Bauer, J. A.; Bohne, C.; Gudmundsdottir, A. D. J. Org. Chem.
2005, 70, 2763. (f) Ma, C.; Kwok, W. M.; Chan, W. S; Du, Y.; Kan, J. T.
W.; Toy, P. H.; Phillips, D. L. J. Am. Chem. Soc. 2006, 128, 2558. (g)
Pika, J.; Konosonoks, A.; Robinson, R. M.; Singh, P. N. D.; Gudmundsdottir,
A. D. J. Org. Chem. 2003, 68, 1964.
* Corresponding author. Tel.: (513) 556-3380; fax: (513) 556-9239.
(1) (a) Mayer, G.; Heckel, A. Ang. Chem., Int. Ed. 2006, 45, 4900. (b)
Pelliccioli A. P.; Wirz, J. Photochem. Photobiol. Sci. 2002, 1, 441. (c)
Goeldner, M.; Givens, R. Dynamic Studies in Biology; Wiley-VCH:
Weinheim, Germany, 2005.
(3) Wiss, J.; Fleury, C.; Onken, U. Org. Process Res. DeV. 2006, 10, 349.
10.1021/jo070558q CCC: $37.00 © 2007 American Chemical Society
Published on Web 07/27/2007
6372
J. Org. Chem. 2007, 72, 6372-6381

2-Azido-1,3-diphenyl-propan-1-one DeriVatiVes
SCHEME 1
The photochemistry and spectroscopy of hydrazoic acid have
been studied in detail.5,6 Recently, however, the atmospheric
chemistry of hydrazoic acid has become a focus of attention.7
This renewed interest in hydrazoic acid stems from the fact that
sodium azide is now heavily used in the automotive industry
for air bag inflators, thus increasing the risk of release of
hydrazoic acid into the environment.8 Hydrazoic acid has a pKa
of 4.7 and is therefore easily generated from aqueous sodium
azide.9 Furthermore, hydrazoic acid is very volatile and
evaporates effectively under normal atmospheric pressure.
Effective studies of the reactivity of hydrazoic acid in the
environment will require diverse methods for generating hy-
drazoic acid.
In this study, we used laser flash photolysis and product
studies at ambient temperature to demonstrate that azides 1a,b
undergo photoinduced C-N bond cleavage, forming azido
radicals and the corresponding benzyl radical derivatives. The
azido radical abstracts a H-atom to form hydrazoic acid, as
shown by IR spectroscopy. Molecular modeling reveals that the
bond dissociation energies of the C-N bonds in azides 1a,b
are less than 59 kcal/mol, considerably lower than the energy
of their triplet ketones (TK), and thus, the cleavage is possible.
In comparison, photolysis of azides 1 in argon matrices at
14 K yields the triplet alkyl nitrene intermediates, which we
identified with the aid of IR spectroscopy and isotope labeling.
Azides 1 have a built-in acetophenone moiety as a triplet
sensitizer, and acetophenone absorbs above 300 nm. More
specifically, the first excited state of acetophenone, S1 (n,π*),
has a maximum absorbance around 320 nm (ꢀ ) 50).10 S1 in
acetophenone undergoes efficient intersystem crossing, 1011 s-1
to T1K, with a quantum efficiency approximating unity.11 In
comparison, alkyl azides have maximum absorbance around
280 nm,5 and thus, by irradiating 1 above 300 nm, one can
ensure that the ketone chromophore absorbs the light and that
it will undergo intersystem crossing to form the first excited
state of the singlet ketone, which intersystem crosses to form
T1K. On the basis of the similarity of azides 1 with the analogous
1,3-diphenyl-propanone derivatives, azides 1 are expected to
have almost degenerate T1K and T2K, which are in thermal
equilibrium with each other.12 We propose that in azides 1a,b,
TK with the (n,π*) configuration results in cleavage of the C-N
bond, whereas the (π,π*) state undergoes intramolecular energy
transfer to form nitrene intermediates.
2. Results
2.1. Product Studies. Photolyzing azide 1a in an argon-
saturated solution via a Pyrex filter yielded product 2 and a
small amount of 3 (Scheme 1). We hypothesize that 2 is derived
from breaking the C-N bond to form an azido radical and 4a.
The azido radical abstracts a H-atom from 4a to form 2. Dimers
3a and 3b are formed from secondary photolysis of 2. The
dimerization of the trans isomer of 2 in solution has been studied
extensively and shown to yield mainly 3a, which arises from
head-to-head dimerization of trans-2.13 We established the
stereochemistry of dimer 3b by X-ray structure analysis, which
showed that it must be formed by head-to-head dimerization of
cis- and trans-2. Photolyzing azide 1a in the presence of a good
H-atom donor, (CH3Si)3SiH, makes it possible to trap 4a to
form 6.
(4) Krakow, B.; Lord, R. C.; Neely, G. O. J. Mol. Spectrosc. 1968, 27,
148.
(5) Reiser, A.; Wagner, H. M. In Chemistry of the Azido Group; Patai,
S., Ed.; Wiley Interscience: London, 1971; pp 441-501.
(6) (a) Beckman, A. O.; Dickinson, R. G. J. Am. Chem. Soc. 1928, 50,
1870. (b) Shapira, D.; Treinin, A. J. Phys. Chem. 1973, 77, 1195. (c) Piper,
L. G.; Krech, R. H.; Taylor, R. L. J. Chem. Phys. 1980, 73, 791. (d) Kodama,
S. Bull. Chem. Soc. Jpn. 1983, 56, 2348. (e) Fang, W.-H. J. Phys. Chem.
A 2000, 104, 4045. (f) Okabe, H. J. Chem. Phys. 1968, 49, 2726. (g) Paur,
R. J.; Bair, E. J. Int. J. Chem. Kinet. 1976, 8, 139. (h) Welge, K. H. J.
Chem. Phys. 1966, 45, 166. (i) McDonald, J. R.; Miller, R. G.; Baronavski,
A. V. Chem. Phys. Lett. 1977, 51, 57. (j) Gericke, K.-H.; Lock, M.; Fasold,
R.; Comes, F. J. J. Chem. Phys. 1992, 96, 422. (k) Bohn, B.; Stuhl, F. J.
Phys. Chem. 1993, 97, 4891. (l) Barnes, R. J.; Gross, A.; Lock, M.; Sinha,
A. J. Phys. Chem. A 1997, 101, 6133.
(7) (a) Li, S.; Zhang, Q.; Wang, W. Chem. Phys. Lett. 2006, 428, 262.
(b) Fang, W.-H. J. Phys. Chem. A 2000, 104, 4045. (c) Cook, P. A.;
Langford, S. R.; Ashfold, M. N. R. Phys. Chem. Chem. Phys. 1999, 1, 45.
(d) Hewett, K. B.; Setser, D. W. J. Phys. Chem. A 1998, 102, 6274.
(8) (a) Li, S.; Zhang, Q.; Wang, W. Chem. Phys. Lett. 2006, 428, 262.
(b) Cui, Y.; Barford, J. P.; Renneberg, R. Anal. Sci. 2006, 22, 1279. (c)
Orlando, J. J.; Tyndall, G. S. EnViron. Sci. Technol. 2005, 39, 1632. (d)
Betterton, E. A. Crit. ReV. EnViron. Sci. Technol. 2003, 33, 423.
(9) Betterton, E. A.; Robinson, J. L. J. Air Waste Manage. Assoc. 1997,
47, 1216.
(10) Braude, E. A. Ultraviolet and Visible Light Absorption in Deter-
mination of Organic Structures by Physical Methods; Academic Press: New
York, 1955; Vol. 4, pp 131-94.
(11) (a) McGarry, P. F.; Doubleday, C. E., Jr.; Wu, C.-H.; Stabb, H. A.;
Turro, N. J. J. Photochem. Photobiol., A 1994, 77, 109. (b) Lamola, A. A.;
Hammond, G. S. J. Chem. Phys. 1965, 43, 2129-2139. (c) Borkman, R.
F.; Kearns, D. R. Chem. Commun. 1966, 14, 446.
J. Org. Chem, Vol. 72, No. 17, 2007 6373

Klima et al.
SCHEME 2
To rule out the possibility that 2 is being formed by direct
excitation of the azido chromophore in azide 1a, we irradiated
1a through a quartz filter, resulting in the formation of dimer
3a and amine 5. Direct irradiation of the azido chromophore in
1a is expected to yield imine 5i, either in a concerted pathway
from the singlet excited azide or via singlet nitrene intermedi-
ates.14 Since amine 5 is a tautomer of imine 5i, this verifies
that photolysis of 1a via a Pyrex filter must involve the triplet
ketone in 1a. Photolysis via quartz also produces 3a, as expected,
since the ketone chromophore in 1a absorbs light in competition
with the azido chromophore.
Time-dependent density functional theory (TD-DFT) cal-
culations18 of azide 1a-A showed that the first and second
excited states of the triplet ketone (T1K and T2K, respectively)
were located 75 and 76 kcal/mol above its ground state (S0).
Inspection of the molecular orbitals shows that T1K and T2K
have (n,π*) and (π,π*) configurations, respectively. Solvation19
of 1a was calculated using the self-consistent reaction field
(SCRF) method with the integral equation formalism polariza-
tion continuum model (IEFPCM). In methanol and acetonitrile,
the configurations of the triplet ketone are interchangeable, such
that T1K has (π,π*) and T2K (n,π*) configurations. The
calculated energies of T1K and T2K in azide 1a-A are listed in
Table 1. Similar results were obtained for conformers B and C
of azide 1a.
In comparison, T1K (n,π*) and T2K (π,π*) in azide 1b are
located 75 and 82 kcal/mol, respectively, above S0. Solvation
calculations in methanol and acetonitrile do not affect the
energies of T1K and T2K significantly for azide 1b.
We also optimized the structure of T1K for azides 1a,b at the
UB3LYP level of theory. The optimized structures of T1K for
azides 1a,b have carbonyl groups with respective bond lengths
of 1.32 and 1.33 Å, and the respective vibration stretches of
the carbonyl bond are at 1149 and 1200 cm-1 (Figure 2). The
Photolysis of azide 1b in an argon-saturated toluene solution
at ambient temperature resulted in products 7-9 (Scheme 2).
We propose that azide 1b also undergoes â-cleavage to form
radical 4b and an azido radical. Radical 4b can abstract a H-atom
from the solvent to form 7, whereas products 8 and 9 can be
attributed to the formation of triplet alkyl nitrene 1b-n, rather
than to the azido radical reacting with 4b. Since radical 4b does
not have any easily abstractable H-atom as does radical 4a, we
do not observe any product from 1b corresponding to 2.
2.2. Quantum Yields. We measured the quantum yield for
the formation of 2 from 1a using p-methoxy valerophenone as
the actinometer and a 313 nm light source.15 The conversion
was kept below 20%, and we did not observe a significant
amount of dimers 3. The quantum yield for forming 2 is 0.05.
The quantum yields are low, presumably, because the azido
radical and 4 can recombine to regenerate the starting material.
2.3. Molecular Modeling. All molecular modeling was
performed using Gaussian 0316 at the B3LYP level of theory
and using 6-31+G(d) as the basis set.17 We optimized the
structures of azides 1 and radicals 4a,b and calculated their IR
spectra. Azides 1a,b have several minimal energy conformations
that are within a few kilocalories per mole of each other. The
three lowest energy conformers of each azide are depicted in
Figure 1.
(16) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K.
N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,
X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.; Jaramillo, J.;
Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.;
Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.;
Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels,
A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.;
Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,
P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson,
B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03,
revision A.1; Gaussian, Inc.: Pittsburgh, PA, 2003.
(17) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Lee, C.; Yang,
W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(18) (a) Labanowski, J. Density Functional Methods in Chemistry;
Springer-Verlag: Berlin, 1991. (b) Parr, R. G.; Weitao, Y. Density
Functional Theory in Atoms and Molecules; Oxford University Press: New
York, 1989. (c) Bauernschmitt, R. Chem. Phys. Lett. 1996, 256, 454. (d)
Stratmann, R. J. Chem. Phys. 1998, 109, 8218. (e) Foresman, J. J. Phys.
Chem. 1992, 96, 135.
(12) Samanta, S.; Mishra, B. K.; Pace, T. C. S.; Sathyamurthy, N.; Bohne,
C.; Moorthy, J. N. J. Org. Chem. 2006, 71, 4453.
(13) (a) Bergmann, E. D.; Solomonovici, A. Tetrahedron 1971, 27, 2675.
(b) Stobbe, H.; Hansel A. Ber. Dtsch. Chem. Ges. B 1926, 59, 2254.
(14) Scriven, E. F. V. Azides and Nitrenes. ReactiVity and Utility;
Academic Press: New York, 1984.
(19) (a) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. ReV. 2005, 105,
2999. (b) Mennucci, B.; Cance`s, E.; Tomasi, J. J. Phys. Chem. B 1997,
101, 10506. (c) Cance`s, E.; Mennucci, B. J. Chem. Phys. 2001, 114, 4744.
(d) Tomasi, J.; Persico, M. Chem. ReV. 1994, 94, 2027. (e) Cramer, C. J.;
Truhlar, D. J. Chem. ReV. 1999, 99, 2161.
(15) Wagner, P. J.; Kochevar, I. E.; Kemppainen, A. E. J. Am. Chem.
Soc. 1972, 94, 7489.
6374 J. Org. Chem., Vol. 72, No. 17, 2007

2-Azido-1,3-diphenyl-propan-1-one DeriVatiVes
FIGURE 1. Optimized structures of azides 1a and 1b.
TABLE 1. Comparison of Energies of Triplet Excited States of
0.1 kcal/mol above the T1A of azides 1. Intrinsic reaction
coordinate (IRC) calculations24 were performed to correlate these
transition states with T1A and 1-n. Thus, as expected, 1-n can
easily be formed from T1A in azides 1.
Azides 1 Using TD-DFT Calculations
azide 1a
azide 1b
gas phase CH3CN MeOH gas phase CH3CN MeOH
We calculated the triplet transition states for cleavage of the
azido group in azides 1. These transitions states are located 71
and 74 kcal/mol above S0 in azides 1a,b, respectively. IRC
calculations correlate these transition states not with the T1K of
azide 1 but with a triplet excited state that has elongated
C-N-N bonds (see Supporting Information) and the azido
radical and 4. This is similar to R-cleavage in triplet ketones,
where the characteristic state for an R-cleavage has a (σ,σ*)
nature, whereas the excitation is into the triplet (n,π*) state of
the ketone chromophore.25 More specifically, we theorize
that the energy surface for T1K in azides 1 has an avoided
crossing with the energy surface of 4 and the azido radical. This
is therefore comparable to what has been proposed for R-cleav-
age of triplet ketones.26
(n,π*)
(π,π*)
74
76
77
76
78
76
75
82
76
83
76
83
band at 1200 cm-1 in T1K of azide 1b is coupled with the
bending of the methyl substituents. The bond length of the
carbonyl bonds and their vibration stretches fit with the triplet
ketones having an (n,π*) configuration.20,21 Furthermore, visu-
alization of the molecular orbitals of T1K in 1a,b further support
this assignment (see Supporting Information). However, the
calculated energies of the optimized T1K are 69 and 67 kcal/
mol above S0 in azides 1a,b, respectively, which is somewhat
lower than those estimated from the TD-DFT calculations.
However, we have previously shown that the B3LYP/6-31G+-
(d) calculation underestimates the energies of triplet ketones
with the (n,π*) configuration.22
Optimization of the triplet configuration of the azide moieties
(T1A) in 1a,b shows that the N1-N2 and N2-N3 bonds are
both lengthened to 1.42 and 1.188 Å, respectively, as compared
to S0 of 1 (Figure 2). The N1-N2-N3 bonds have an angle of
120°. The vibration stretches of the azide bands in triplet azides
1a,b are observed at 1708 and 1724 cm-1, respectively. The
energies of T1A of azides 1a,b are 44 and 45 kcal/mol above
S0. The triplet energies and optimized structures of the triplet
azides are similar to what we have previously estimated for
triplet alkyl azides.22,23
In Figure 3, we have plotted the stationary points on the triplet
surface of azides 1a,b. For comparison, we have added
stationary points of the triplet surface of azide 1c, which does
not have â-phenyl substituents (we have previously reported
some of these calculations for azide 1c).23 T1K and T2K in azides
1a,c have very similar energies, whereas the energy gap between
T1K and T2K in azide 1b is somewhat larger, due to the R-methyl
substituents. The energy of T1A for azides 1a-c is very similar,
and the transition state for forming the corresponding triplet
alkyl nitrenes is located less than 1 kcal/mol above T1A. In
comparison, the energies of the transition states for cleaving
the C-N bond in azides 1 are different. In azides 1a,b, this
transition state is less than 74 kcal/mol above S0, whereas it is
82 kcal/mol above S0 for azide 1c. Thus, a radical-stabilizing
substituent in the â-position facilitates cleavage of the C-N
bond.
We calculated the transition states for forming the triplet alkyl
nitrenes (1-n) from T1A and found that they were located
(20) (a) He, H.-Y.; Fang, W.-H.; Phillips, D. L. J. Phys. Chem. A 2004,
108, 5386. (b) Fang, W.-H.; Phillips, D. L. Chem. Phys. Chem. 2002, 3,
889. (c) Fang, W.-H.; Phillips, D. L. J. Theor. Comput. Chem. 2003, 2, 23.
(21) Srivastava, S.; Yourd, E.; Toscano, J. P. J. Am. Chem. Soc. 1998,
120, 6173.
(22) Muthukrishnan, S.; Mandel, S. M.; Hackett, J. C.; Singh, P. N. D.;
Hadad, C. M.; Krause, J. A.; Gudmundsdo´ttir, A. D. J. Org. Chem. 2007,
72, 2757.
(23) Singh, P. N. D.; Mandel, S. M.; Sankaranarayanan, J.; Muthukrish-
nan, S.; Chang, M.; Robinson, R. M.; Ault, B. S.; Gudmundsdo´ttir, A. D.,
unpublished work.
(24) (a) Gonzalez, C.; Schlegel, H. B. J. Chem. Phys. 1989, 90, 2154.
(b) Gonzalez, C.; Schlegel, H. B. J. Phys. Chem. 1990, 94, 5523.
(25) (a) Yadav, J. S.; Goddard, J. D. J. Chem. Phys. 1986, 84, 2682. (b)
Haas, Y. Photochem. Photobiol. Sci. 2004, 3, 6.
(26) Klessinger, M.; Michl, J. Excited States and Photochemistry of
Organic Molecules; VCH Publishers, Inc.; New York, 1995; Ch. 6.
J. Org. Chem, Vol. 72, No. 17, 2007 6375

Klima et al.
FIGURE 2. Optimized structure of (A) T1K and (B) T1A of azide 1a. Bond lengths are in angstroms.
FIGURE 3. Stationary points on triplet energy surface of azides 1a-c. Calculated energies are in kilocalories per mole. The energies of T1K and
T2K were obtained from their optimized structures and from TD-DFT calculations, respectively.
TD-DFT calculations were used to estimate the absorption
spectra of 4a,b, as shown in Table 2. The effect of solvation
was calculated with acetonitrile as the solvent.
We calculated the bond dissociation energy (BDE) for the
cleavage of the C-N bond in azides 1. BDE has been defined
as the enthalpy change of a reaction where a bond is broken
and each of the products retains one of the electrons from the
broken bond.27,28
The BDE can be calculated with eq 1, where ∆Hf(XY) is
the enthalpy change for the dissociation of X-Y, Σ∆Hf(products)
is the sum of the heats of formation of all the products in the
reaction, and Σ∆Hf(reactant) is the sum of the heat of formation
of each reactant.
BDE(XY) ) ∆Hr(XY) ) Σ∆Hf(products) - Σ∆Hf(reactant) (1)
The calculated BDEs for azides 1a,b are 55 and 59 kcal/
mol, respectively. Since the BDEs for breaking the C-N bond
X-Y f X• + Y•
6376 J. Org. Chem., Vol. 72, No. 17, 2007

2-Azido-1,3-diphenyl-propan-1-one DeriVatiVes
TABLE 2. Major Absorption Spectral Features (nm, Oscillator Strength) for 4a and 4ba
electronic transition above 300 nm
1a
1b
gas phase
CH3CN
gas phase
CH3CN
524 (0.0104)
525 (0.0159)
506 (0.0087)
510 (0.0109)
369 (0.0016)
371 (0.0011)
369 (0.009)
368 (0.0013)
331 (0.0127)
368 (0.0025)
331 (0.0033)
329 (0.0288)
326 (0.0324)
327 (0.0389)
330 (0.022)
321(0.0027)
323 (0.0040)
326 (0.0184)
316 (0.0021)
316 (0.0151)
317 (0.011)
308 (0.0152)
a Numbers in parentheses are the calculated oscillator strengths (f) for the electronic transition, in nanometers.
FIGURE 4. Laser flash photolysis of (A) azide 1a, (B) azide 1b, and (C) decay of radical 4a at 310 nm.
in azides 1 are considerably lower than the energies of their
1K, it is feasible to break their C-N bond photochemically to
into a π* orbital. Furthermore, the transient spectra of radical
4a is similar to the transient spectra of the benzyl radical, which
has λmax at ∼320 nm and a lifetime of 20 µs.30 The rate of
decay of the transient absorption of 4a at 310 nm was 7 ×
104 s-1, whereas in oxygen-saturated acetonitrile, the transient
absorption was quenched.
T
form an azido radical and the corresponding alkyl radical. In
comparison, the calculated BDE for the C-N bond in azide 1c
is 71 kcal/mol and thus comparable to the energy of its T1K
and T2K, making cleavage of the C-N in azide 1c less likely.
2.4. Laser Flash Photolysis. Laser flash photolysis (excimer
laser 308 nm)29 of azide 1a in acetonitrile and in methanol
produced transient spectra with λmax at ∼310 nm (Figure 4),
which we assigned to 4a. We based this assignment on the
similarity of this transient to the calculated absorption spectrum
of 4a. TD-B3LYP calculations predicted that 4a has absorption
bands in acetonitrile at 316 and 327 nm. The band at 327 nm
is mainly due to a mixed transition out of the lone pair on the
oxygen atom into the half-full p orbital on the â-carbon atom
and a π* orbital, whereas the transition at 316 nm is mainly
due to a transition out of the half-full p orbital on the â-carbon
Laser flash photolysis of azide 1b resulted in the formation
of a similar transient spectrum, with λmax at 310 nm. As before,
we used TD-DFT to assign this transient absorption to radical
4b (Table 2). The rate of decay of radicals 4b was 6.5 ×
104 s-1
.
2.5. Matrix Isolation. We deposited 1b into argon matrices
in several different experiments, entraining the vapor pressure
over a sample of pure azide in flowing argon and depositing
the resulting sample on a 14 K cryogenic surface. The infrared
spectrum of such a sample showed a strong band at 2108 cm-1
and a band at 1250 cm-1 (Figure 5). We assign the band at
1250 cm-1 to C-H bending that is coupled to the C-CO-C
stretching, which is calculated to be at 1275 cm-1 (Table 3).
As expected, the bands at 2108 and 1250 cm-1 were destroyed
upon irradiation, and two new major bands were observed at
(27) (a) McMillen, D. F.; Golden, D. M. Annu. ReV. Phys. Chem. 1982,
33, 493. (b) Brocks, J. J.; Beckhaus, H.-D.; Beckwith, A. L. J.; Ruchardt,
C. J. Org. Chem. 1998, 63, 1935. (c) CRC Handbook of Chemistry and
Physics; CRC Press: Boca Raton, FL, 1989; Vol. 1, pp 369-403.
(28) Campos, L. M.; Dang, H.; Ng, D.; Yang, Z.; Martinez, H. L.; Garcia-
Garibay, M. A. J. Org. Chem. 2002, 67, 3749.
(30) Maillard, B.; Ingold, K. U.; Scaiano, J. C. J. Am. Chem. Soc. 1983,
105, 5095.
(29) Cosa, G.; Scaiano, J. C. Photochem. Photobio. 2004, 80, 159.
J. Org. Chem, Vol. 72, No. 17, 2007 6377

Klima et al.
FIGURE 5. IR Spectra of 1b in argon matrices (A) before irradiation (blue) and after irradiation (red) and (B) IR spectra after irradiation of 1b
(red) and 1b-d (blue).
1286 cm-1, which fits well with the calculated band of the
C-CO-C stretching for 1b-n-d.
TABLE 3. Calculated and Observed C-H Bands in Azides 1b and
Nitrene 1b-n
C-H
C-D
The IR spectra of azide 1a in argon matrices show a strong
azido band at 2106 cm-1 and two less intense bands at 2128
and 2136 cm-1 (Figure 6). Upon irradiation, the bands at 2106
and 2136 cm-1 were depleted, along with medium intensity
bands at 1350, 1251, and 1211 cm-1, which can all be assigned
to coupled bending of the C-H bond on the â-carbon atom.
The major new bands that were formed are at 1217 and
1207 cm-1, and we assign them to the calculated C-H bending
calcd
obsd
calcd
obsd
1b
1b-n
1275
1276
1251
1250
1323
1275
1280
1274
1252
1286
1323 and 1274 cm-1. In contrast, the remaining major bands
were not shifted significantly upon irradiation. We assign this
new photoproduct in matrices to nitrene 1b-n, based on the
following observations. The calculated IR spectrum of nitrene
2b has vibrational bands at 1276 and 1251 cm-1, which are
due to stretching of the C-CO-C bands coupled with the C-H
bending. We assign the observed bands at 1323 and 1274 cm-1
to those calculated bands. To support this assignment further,
we prepared azide 1b with a deuterium atom in the â-position.
The calculated IR spectrum of nitrene 1b-n-d has only one IR
band for the stretching of the C-CO-C bonds at 1280 cm-1
because the C-D band does not couple to the C-CO-C stretch
as in 1b-n. Thus, if the assignment of the bands at 1276 and
1251 is correct, only one major band will be observed in this
region of the spectrum of nitrene 1b-n-d. Deposition of azide
1b-d into argon matrices resulted in an IR spectrum that has an
azido band at 2108 cm-1 as before, whereas the C-H bend
that is coupled with the stretching of the azide band is located
at 1252 cm-1, which fits reasonably well with the calculated
isotope shift of 5 cm-1. Upon irradiation of 1b-d, we observed
depletion of the azido band and the band at 1252 cm-1. As
expected, the photoproduct had only one major band at
that is coupled with C-CO-C bending, at 1251 and 1210 cm-1
.
Thus, matrix photolysis of azide 1a,b does not form azido
radical and 4 since we do not observe the formation of either
azido radical or hydrazoic acid; rather, photolysis of azides 1
at cryogenic temperatures results in the formation of nitrenes
1-n.
2.6. Trapping of Hydrazoic Acid. To verify that the azido
radical formed by irradiating azides 1 abstracts a H-atom to
form hydrazoic acid, we placed crystals of 1a in a Pyrex segment
inserted into the argon deposition line. The glass segment of
the line was irradiated for 24 h while argon flowed over the
crystals. The flowing argon was condensed on the matrix cold
window at 14 K and monitored by IR spectroscopy. After
24 h, IR bands were observed at 2135 and 1146 cm-1, matching
the two most intense bands reported for hydrazoic acid in argon
matrices. Moreover, their relative intensities were consistent with
the published spectrum.31 1H NMR spectra of the remaining
yellow solid in the glass tube showed unreacted 1a and the
(31) Himmel, H.-J.; Junker, M.; Schnockel, H. J. Chem. Phys. 2002,
117, 3321.
6378 J. Org. Chem., Vol. 72, No. 17, 2007

2-Azido-1,3-diphenyl-propan-1-one DeriVatiVes
FIGURE 6. IR spectra of azide 1a before (black) and after (red) irradiation in argon matrices.
FIGURE 7. X-ray structure of azide 1a.
formation of mainly trans-2 and dimer 3a. The X-ray structure
of azide 1a shows that it crystallizes in a conformer that
is similar to the calculated minimal energy structure, with
the carbonyl and azide groups in a pseudo-syn alignment
(Figure 7). The azide moiety in 1a forms non-traditional H-atom
bonding with the â-hydrogen atom (H9â) on an adjacent
molecule, and the N-H distance is 2.54 Å. The distances from
the N1 to the H-atoms H8R1 and H8R2 are 2.64 and 3.35 Å,
respectively (Figure 6). Solid state irradiation of 1a breaks the
C-N bond to form 4a and an azido radical. The azido radical
abstracts H1 to form hydrazoic acid and 2. We observe mainly
the trans isomers of 2, presumably because the azido radical
abstracts HR1. Examination of the crystal lattice shows that the
shortest intermolecular distances between CR and Câ atoms are
5.8 and 5.7 Å, respectively. These distances are considerably
longer than the intermolecular distance observed for topochemi-
cally allowed dimerization in the solid state, which generally
has an upper limit of 4.3 Å.32 However, it should be noted that
the intermolecular distances between adjacent CR and Câ will
be different after the formation of 2.
tions place T2K in azide 1b ∼6 kcal/mol higher in energy than
1K. However, the energies obtained from the TD-DFT
calculations are based on the assumption that the geometry of
1K and T2K are similar to that of S0 (Frank-Condon excitation);
thus, we cannot accurately determine the energy gap between
these triplet excited states. Besides, they are close in energy
and thus in equilibrium with each other. This is in agreement
with what has been reported for valerophenone, R,R-dimeth-
ylvalerophenone, and R,R-dimethylacetophenone, which have
their T1K and T2K values within a few kilocalories per mole of
each other.33
We propose that the cleavage of the azido group in azides
1a,b comes from TK with the (n,π*) configuration, whereas the
energy transfer to form triplet alkyl nitrenes 1a-n and 1b-n must
take place from TK with a (π,π*) configuration. This is similar
to what we have observed for R-azido acetophenone derivatives,
where the TK of the ketone with the (n,π*) configuration cleaves
to form an azido methyl and benzoyl radical pair, whereas TK
with the (π,π*) configuration undergoes intramolecular energy
transfer to form R-nitrenoacetophenone derivatives.22,34 In azide
1b, only cleavage of the C-N bond is observed from T1K with
T
T
3. Discussion
(33) Wagner, P. J.; May, M. J.; Haug, A.; Graber, D. R. J. Am. Chem.
Soc. 1970, 92, 5269.
(34) (a) Mandel, S. M.; Singh, P. N. D.; Muthukrishnan, S.; Chang, M.;
Krause, J. A.; Gudmundsdottir, A. D. Org. Lett. 2006, 8, 4207. (b) Singh,
P. N. D.; Mandel, S. M.; Zhu, Z.; Franz, R.; Ault, B. S.; Gudmundsdo´ttir,
A. D. J. Org. Chem. 2003, 68, 7951. (c) Mandel, S. M.; Krause Bauer, J.
A.; Gudmundsdo´ttir, A. D. Org. Lett. 2001, 3, 523.
The TD-DFT calculations show that T1K and T2K in azide
1a are almost degenerate. In comparison, the TD-DFT calcula-
(32) (a) Schmidt, M. J. Pure Appl. Chem. 1971, 27, 647. (b) Ramamurthy,
V.; Venkatesan, K. Chem. ReV. 1987, 87, 433.
J. Org. Chem, Vol. 72, No. 17, 2007 6379

Klima et al.
FIGURE 8. Transition state for C-N bond cleavage in azides 1.
the (n,π*) configuration. In comparison, R-azidoacetophenone
underwent both R-cleavage and energy transfer from T1K (n,π*)
and T2K (π,π*), respectively. However, this is reasonable since
the transition state for R-cleavage in azidoacetophenone is
located 77 kcal/mol above its S0, whereas the transition state
for C-N cleavage in azide 1b is only 72 kcal/mol above its
S0.22
geometry for â-quenching.12 It is likely that â-quenching occurs
when one of the lone pairs on the oxygen atom in the benzoyl
moiety is oriented toward the center of the â-phenyl group. Thus,
we theorize that â-quenching in azide 1b is less efficient than
in azide 1, due to the R-methyl groups. Moreover, steric
repulsion in azide 1b makes both â-quenching and C-N bond
cleavage less efficient and thus results in nitrene 1b-n being
We reported previously that intramolecular triplet sensitization
of â-azido-1-phenyl-propan-1-one, 1c, yields mainly â-nitreno-
1-phenyl-propan-1-one, 1c-n, via intramolecular energy transfer
from the triplet ketone to the azido moiety, rather than cleavage
of the azido group.23 Since the calculated transition state for
cleaving the C-N bond in azide 1c is 82 kcal/mol higher in
energy than S0, it is reasonable that 1c does not undergo C-N
bond cleavage efficiently. In comparison, the calculated transi-
tion states for cleaving the C-N bond in 1a,b are 71 and
74 kcal/mol above their S0, respectively. Thus, C-N cleavage
of azides 1a,b is facilitated by the resonance stabilization
provided by the phenyl group. The calculated transition state
for the cleavage of the C-N bond in azide 1b is approximately
3 kcal/mol higher in energy than for azide 1a, presumably
because in the transition states the â-C radical is conjugated
with the â-phenyl ring and there is more steric repulsion between
the R-methyl groups and the â-phenyl group in azide 1b than
the R-H-atoms and the â-phenyl group in azide 1a (Figure 8).
This explains why products 8 and 9, which are attributable to
the formation of triplet alkyl nitrene 1b-n, are observed in
solution. The energy transfer from T2K to form a nitrene can
compete more efficiently with C-N cleavage from T1K in azide
1b than in azide 1a. However, we were not able to detect nitrene
formation directly using laser flash photolysis, presumably
because the quantum yields for forming the nitrene intermediates
are low.
Propiophenone derivatives with phenyl substituents in the
â-position undergo so-called â-quenching, which is an efficient
deactivation of their triplet ketones with an (n,π*) configura-
tion.12,35 The mechanism of this process is not yet fully
characterized, but it is thought to involve a charge transfer from
the â-phenyl ring to the electronic-deficient triplet ketone.36 It
is possible that some of the T1K (n,π*) in azides 1 is quenched
by the â-phenyl ring, which would also explain the low quantum
yields for the cleavage of the C-N bond. Bohne and co-workers
have shown that propiophenone derivatives with an R-methyl
substituent undergo â-quenching less efficiently than propiophe-
none due to the steric demands of the methyl groups on the
formed from T2K.
At cryogenic temperatures, irradiation of azides 1a,b yields
triplet alkyl nitrenes 1-n, rather than the formation of azido
radicals and 4. Presumably, the thermal population of T2K at
low temperatures is more favorable than the thermal activation
for cleaving the C-N bond from T1K in matrices. Furthermore,
the formation of the azido radical and 4 in matrices will only
lead to the regeneration of 1 since the radical pair cannot diffuse
apart.
Irradiation of azide 1a selectively cleaves the C-N bond to
form 4a and the azido radical, and the latter abstracts an H-atom
resulting in the release of hydrazoic acid. Interestingly, pho-
tolysis of azide 1a both in solution and in the solid state at
ambient temperature can be used to release hydrazoic acid. In
comparison, photolysis of azide 1b forms nitrene 1b-n in
competition with cleavage to form the azido radical.
Conclusion
We have shown that photolysis of â-azido-1,3 diphenyl-
propiophenone derivatives 1 results in C-N bond cleavage to
form an azido radical and 4. Irradiation of azides 1a,b populates
their T1K (n,π*), which results in cleavage of the C-N bond.
The radical stabilization effect of the â-phenyl group makes
the C-N bond cleavage in azides 1a,b feasible at room
temperature. Both solution and solid state irradiation of azide
1a results in the formation of azido radicals that abstract a
H-atom to form hydrazoic acid, and thus, azides 1 can be used
as photoremoveable protecting groups of hydrazoic acid.
4. Experimental Section
4.1. Molecular Modeling. All geometries were optimized as
implemented in the Gaussian 03 programs, at the B3LYP level of
theory and with the 6-31+G(d) basis set.16,17 All transition states
were located at the UB3LYP level of theory, and each transition
state was confirmed to have one imaginary vibrational frequency.
IRC calculations were used to verify that the located transition states
corresponded to the attributed reactant and product.24 UV absorption
spectra were calculated at the TD-B3LYP level with the 6-31+G-
(d) basis set.18 The effect of solvation was calculated using the
SCRF method with IEFPCM19 with toluene as the solvent.
4.2. Laser Flash Photolysis. Laser flash photolysis was per-
formed with an excimer laser (308 nm, 17 ns). The system has
been described in detail elsewhere.29 A stock solution of azides
1a,b in acetonitrile was prepared with spectroscopic grade aceto-
nitrile, such that the solutions had an absorption between 0.6 and
(35) (a) Boch, R.; Bohne, C.; Scaiano, J. C. J. Org. Chem. 1996, 61,
1423. (b) Ng, D.; Yang, Z.; Garcia-Garibay, M. A. Tetrahedron Lett. 2001,
42, 9113. (c) Netto-Ferreira, J. C.; Leigh, W. J.; Scaiano, J. C. J. Am. Chem.
Soc. 1985, 107, 2617. (d) Wismontski-Knittel, T.; Kilp, T. J. Phys. Chem.
1984, 88, 110. (e) Scaiano, J. C.; Perkins, M. J.; Sheppard, J. W.; Platz, M.
S.; Barcus, R. L. J. Photochem. 1983, 21, 137.
(36) Leigh, W. J.; Banisch, J. H.; Workentin, M. S. J. Chem. Soc., Chem.
Commun. 1993, 988.
6380 J. Org. Chem., Vol. 72, No. 17, 2007

2-Azido-1,3-diphenyl-propan-1-one DeriVatiVes
0.8 at 308 nm. Typically, ∼1 mL of the stock solution was placed
in a 10 mm × 10 mm wide, 48 mm long quartz cuvette and was
purged with nitrogen for 5 min. The rates were obtained by fitting
an average of three to five kinetic traces.
NMR (60 MHz, CDCl3) 199.52, 141.9, 136.0, 133.9, 129.3, 129.2,
128.3, 127.8, 127.7, 48.4, 48.1 ppm. GC/MS (EI): m/z 416 (M+),
311, 207, 105.
3b:39 1H NMR (250 MHz, CDCl3): δ 8.68 (d, 7 Hz, 2H), 7.51
(d, 7 Hz, 2H), 7.3-6.8 (m, 16H), 5.76 (t, 9 Hz, 1H), 5.01 (t, 9 Hz,
1H), 4.51 (t, 9 Hz, 1H), 3.93 (t, 9 Hz, 1H) ppm. 13C NMR
(60 MHz, CDCl3): δ 197.2, 136.3, 135.9, 133.8, 133.1, 130.5,
129.9, 129.4, 128.7, 128.4, 128.4, 128.1, 127.0, 126.4, 48.5, 48.2,
43.6, 42.5 ppm. GC/MS (EI): m/z 416 (M+), 311, 296, 207, 105.
4.9. Quartz Photolysis of Azide 1a. Azide 1a (50 mg,
0.20 mmol) was placed in a quartz test tube, and toluene was added
(1.5 mL) with a septum fixed by copper wire. The resulting solution
was bubbled in argon for 15 min, sealed by Parafilm, and then
photolyzed for 5 h via a quartz filter. Product 5 was isolated from
the reaction mixture using HPLC (8 mg, 0.036 mmol, 18% yield).
The progress of the reaction was followed by HPLC chromatog-
raphy, which showed that 3 and 5 formed in the ratio of ∼4:1,
respectively. After 5 h of irradiation, HPLC showed some remaining
starting material (10%) and that 3 and 5 (90%) formed in the ratio
of 4:1
4.3. Matrix Isolation. Matrix isolation studies were performed
using conventional equipment.37
4.4. Photolysis of Azide 1b. A solution of 1b (663.0 mg,
2.38 mmol) in toluene (125 mL) was purged with argon and
photolyzed via a Pyrex filter for 18 h. GC analysis of the reaction
mixture showed the formation of photoproducts 7-9 in the ratio
48:32:20 and no remaining starting material. Column chromatog-
raphy (silica eluted with hexane/ethyl acetate) of the reaction
mixture yielded photoproducts 8 (166 mg, 0.66 mmol, 28% yield)
and 7 (189 mg, 0.85 mmol, 36%). Product 9 degraded on the
column, so it was characterized by injection of authentic sample
1
on the GC chromatography and H NMR spectra of the reaction
mixture.
4.5. 2,2-Dimethyl-1,3-diphenyl-propan-1-one (7). IR (CHCl3)
ν
max: 1675 cm-1. 1H NMR (250 MHz, CDCl3): δ 7.02-7.52 (m,
12 H) 3.07 (s, 2 H), 1.30 (s, 6H), ppm; GC/MS (EI): m/z 238
(M+), 223, 131, 132, 117, 105, 91, 77.
1
5: IR (CHCl3): 3356, 1601, 1526 cm-1. H NMR (250 MHz,
4.6. 4,4-Dimethyl-3,5-diphenyl-isoxazolidine (8). IR (CHCl3)
CDCl3): δ 10.43 (b s, N-H), 7.94-7.97 (m, 2H), 7.63-7.66 (m,
2H), 7.43-7.50 (m, 6H), 6.12 (broad s, N-H) ppm. 13C NMR
(250 MHz, CDCl3): δ 190.2, 162.9, 140.3, 137.6, 131.1, 130.7,
129.1, 128.3, 127.2, 126.4, 91.9 ppm. HRMS: calcd for C15H13-
ON (M + 1): 224.1075. Found: 224.1069.
4.10. Photolysis of 1a in the Presence of (SiMe3)3SiH. Azide
1a (0.01 g, 0.04 mmol) was placed in a Pyrex test tube, and toluene
(1.5 mL) and (SiMe3)3SiH (0.031 g, 0.08 mmol) were added. The
resulting solution was bubbled in argon for 15 min and photolyzed
for 6 h via a Pyrex filter. HPLC and GC-MS analyses of the reaction
mixture showed the formation of 6 (95% yield), which was
characterized by the injection of authentic samples on the GC and
the HPLC with a commercially available standard.
1
ν
max: 2958, 2925, 2854, 1453, 909 cm-1; H NMR (250 MHz,
CDCl3): δ 1.58 (d, 6 H, 2.5 Hz), 5.94 (s, 1 H), 7.33-7.92 (m,
10 H) ppm; 13C NMR (60 MHz, CDCl3): δ 178.2, 142.6, 133.6,
130.8, 129.2, 129.1, 129.0, 128.9, 128.8, 128.7, 128.5, 127.3, 127.2,
86.8, 70.1, 27.8, 26.6 ppm; GC/MS (EI): m/z 252 (M+ + 1), 236,
194, 167, 146 (100), 131, 130, 119, 105, 104, 91, 77.
4.7. Synthesis of Azide 1a. 3-Bromo-1,3-diphenyl-propenone
(1.7 g, 6 mmol) was dissolved in ethanol (60 mL) and placed in a
100 mL flask. To this solution was added sodium azides (1.5 g,
24 mmol) dissolved in a minimum amount of water. The resulting
mixture was stirred for 12 h. The reaction mixture was extracted
with diethyl ether, and the solvent was removed under vacuum to
yield oil, which was crystallized from ethanol to yield white crystals
of 3-azido-1,3-diphenyl-propenone, 1a (1.1 g, 4.5 mmol, 75% yield).
The solid was recrystallized 3 times from ethanol to yield white
crystals of 1a.
4.11. Solid State Photolysis of 1a. Crystals of azide 1a in a
Pyrex glass tube were irradiated for 24 h while argon flowed over
the crystals. 1H NMR spectra of the irradiated crystals showed the
remaining starting material and the formation of trans-2 and dimer
3a in the ratio of 4:1.
mp: 77-78 °C. IR (CHCl3): 1680, 2094, 2123 cm-1. 1H NMR
(250 MHz, CDCl3): δ 7.95 (d, 6 Hz, 2H), 7.56-7.59 (m, 1H),
7.40-7.50 (m, 7H), 5.27 (m, 1H), 3.51-3.55 (m, 1H), 3.29-3.31
(m, 1H) ppm. 13C NMR (CDCl3, 60 MHz): δ 196.4, 139.3, 136.5,
133.5, 128.9, 128.7, 128.4, 128.1, 126.8, 61.5, 45.2 ppm. GC/MS
(EI): m/z 249 (M+), 221, 105.
Acknowledgment. We thank the National Science Founda-
tion and the Ohio Supercomputer Center for support. C.V.
thanks the NSF for support via the REU program at the
University of Cincinnati. We are grateful to Prof. J. C. Scaiano
at the University of Ottawa for the use of his laser flash systems.
We also thank Dr. J. Pika at Firmenich, Inc. for her assistance.
4.8. Photolysis of Azide 1a. Azide 1a (153 mg, 0.61 mmol)
was dissolved in toluene, argon was bubbled through the solution
for 15 min, and the solution was irradiated via a Pyrex filter for
10 h. The progress of the reaction was followed with HPLC, which
showed the formation of only 2 initially, but upon prolonged
irradiation, more was formed of 3, at the expense of 2. The
photolysis was stopped when HPLC showed the formation of cis-
and trans-2 and 3 in a ratio of ∼1:4 and ∼20% remaining starting
material. Dimers 3 were isolated using HPLC and yielded mainly
3a (54 mg, 0.12 mmol, 40% yield) and a small amount of 3b
(6 mg, 0.015 mmol, 5% yield).
Supporting Information Available: Cartesian coordinates,
energies, vibrational frequencies, and X-ray structures of 1a and
3b. NMR spectra of 1b-d, 5, and 8. This material is available free
of charge via the Internet at http://pubs.acs.org.
JO070558Q
3a:38 1H NMR (250 MHz, CDCl3): δ 7.80 (d, 7 Hz, 4H), 7.5-
(38) Toda, F.; Tanka, K.; Kato, M. J. Chem. Soc., Perkin Trans. 1 1998,
7, 1315.
7.2 (m, 16H), 4.60 (d, 8.7 Hz, 2H), 3.98 (d, 8.7 Hz, 2H) ppm. 13
C
(39) Stoermer, R.; Starck, H.; Anker, H. E. Ber. Dtsch. Chem. Ges. 1937,
70, 483.
(37) Ault, B. S. J. Am. Chem. Soc. 1978, 100, 2426.
J. Org. Chem, Vol. 72, No. 17, 2007 6381