The photochemistry of 4-azidopyridine-1-oxide

Article abstract of DOI:10.1021/jo061259o

(Chemical Equation Presented) Laser flash photolysis of 4-azidopyridine-1-oxide at 266 or 308 nm yields triplet 4-nitrenopyridine-1- oxide as the dominant reactive intermediate species, with k_1SC of approximately 2 × 10⁷ s^-1. No evidence of products arising from the singlet nitrene was observed, indicating a slow rate of cyclization to the benzazirine and didehydroazepine species. The slow rate of cyclization is postulated to be due to the aminoxyl-like electronic configuration of this species, which withdraws spin density from sites for potential cyclization.

Full text of DOI:10.1021/jo061259o

The Photochemistry of 4-Azidopyridine-1-oxide
Katherine J. Hostetler, Kyle N. Crabtree, and James S. Poole*
Department of Chemistry, Ball State UniVersity, Muncie, Indiana 47306
jspoole@bsu.edu
ReceiVed June 19, 2006
Laser flash photolysis of 4-azidopyridine-1-oxide at 266 or 308 nm yields triplet 4-nitrenopyridine-1-
oxide as the dominant reactive intermediate species, with kISC of approximately 2 × 107 s-1. No evidence
of products arising from the singlet nitrene was observed, indicating a slow rate of cyclization to the
benzazirine and didehydroazepine species. The slow rate of cyclization is postulated to be due to the
aminoxyl-like electronic configuration of this species, which withdraws spin density from sites for potential
cyclization.
Introduction
symmetry afforded us with the simplest of these types of
systems. If photolysis of the azido group proceeds in the same
manner as other aryl azides,5,6 we would expect the reactions
shown in Scheme 1.
Alternative reaction paths involving the photoactive N-oxide
group are also possible, but it is known that photolysis of
aromatic azides proceeds with reasonable quantum yield.7
The chemistry of azidoheterocyclic-N-oxides has been in-
vestigated in the literature due to the potential of these
compounds as antitumor agents.1,2 More recently, the potential
of these reagents as sources of nitrene species was studied by
product analysis following thermolysis or photolysis of the
azides in the presence of nucleophiles.3 In general terms, it was
found that the product distributions arising from thermolysis
and photolysis of both 3- and 4-azidopyridine-1-oxides were
complex, and not readily resolved. The thermo- and photo-
chemical reactions of the 2-azido analogues of these compounds
follow significantly different reaction paths to generate 2-cy-
anopyrroles.4
We have chosen to study the chemistry of azidoheterocyclic-
N-oxides using nanosecond laser flash photolysis (LFP) and
time-resolved infrared (TRIR) techniques in order to elucidate
the nature of the nitrene intermediates and perhaps clarify the
product analysis studies of prior workers.3 In this contribution,
we report the results of our studies on an archetypal system,
4-azidopyridine-1-oxide (1). 4-Azidopyridine-1-oxide was cho-
sen for our initial study due to the fact that its molecular
Experimental Methods
Compound 1 was prepared from the corresponding hydrazine,8
using the method of Itai and Kamiya,1 and recrystallized from
acetone (see Supporting Information). All solvents used in this study
(including pyridine) were obtained in spectroscopic purity from
commercial sources and were used without additional purification.
Diethylamine was purified by distillation, stored over KOH, and
used as soon after distillation as possible. The instruments used
(5) Platz, M. S. In ReactiVe Intermediate Chemistry; Moss, R. A., Platz,
M. S, Jones, M. J., Eds.; John Wiley & Sons Inc.: Hoboken, NJ, 2004; pp
522-545. (b) Borden, W. T.; Gritsan, N. P.; Hadad, C. M.; Karney, W. L.;
Kemnitz, C. R.; Platz, M. S. Acc. Chem. Res. 2000, 33, 765. (c) Gritsan,
N. P.; Zhu, Z.; Hadad, C. M.; Platz, M. S. J. Am. Chem. Soc. 1999, 121,
1202.
(1) Itai, T.; Kamiya, S. Chem. Pharm. Bull. 1961, 9, 87.
(2) (a) Itai, T.; Kamiya, S. Chem. Pharm. Bull. 1963, 11, 348. (b) Itai,
T.; Kamiya, S. Chem. Pharm. Bull. 1963, 11, 1059.
(3) Abramovitch, R. A.; Bachowska, B.; Tomasik, P. Pol. J. Chem. 1984,
58, 805.
(4) Abramovitch, R. A.; Cue, B. W., Jr. J. Org. Chem. 1973, 38, 173.
(b) Abramovitch, R. A.; Cue, B. W., Jr. J. Am. Chem. Soc. 1976, 98, 1478.
(6) Gritsan, N. P.; Likhotvoric, I.; Tsao, M.-L.; Celibi, N.; Platz, M. S.;
Karney, W. L.; Kemnitz, C. R.; Borden, W. T. J. Am. Chem. Soc. 2001,
123, 1425. (b) Gritsan, N. P.; Tigelaar, D.; Platz, M. S. J. Phys. Chem. A
1999, 103, 4465.
(7) Leyshon, L. J.; Reiser, A. J. Chem. Soc., Faraday Trans. 2, 1972,
1918. (b) Go, C. L.; Waddell, W. H. J. Org. Chem. 1983, 48, 2897.
(8) Katritzky, A. R. J. Chem. Soc. Abstr. 1956, 2404.
10.1021/jo061259o CCC: $33.50 © 2006 American Chemical Society
Published on Web 10/21/2006
J. Org. Chem. 2006, 71, 9023-9029
9023

Hostetler et al.
SCHEME 1. Potential Reactions Following Photolysis of
4-Azidopyridine-1-oxide (1)
for characterization of organic compounds are described in the
Supporting Information.
The Laser flash photolysis system utilized at The Ohio State
University has been described previously.9 A solution of 1 (2-3
mL), in an appropriate solvent, was deoxygenated using a thin
stream of bubbling argon over a period of at least 10 min. The
solvents used in this component of the study were 3-methylpentane,
dichloromethane, acetonitrile, and pyridine. The concentration of
solution for most experiments was chosen to give an optical density
of approximately 1.5-2.0 at 308 nm, except in the cases of
experiments where the concentration of 1 had to be varied in order
to yield kinetic information: in such cases the optical density was
varied between 0.5 and 2.5. Samples were irradiated using a Xe-
Cl excimer laser (308 nm, nominal pulse width ) 30 ns, 50 mJ/
pulse), and the transient absorbance signals captured by CCD (for
transient spectra) or fast photomultiplier (for kinetic traces).
The sub-microsecond time-resolved infrared (TRIR) laser flash
photolysis system used at The Ohio State University has also been
described previously.10 In the TRIR experiment, a solution of 1
(3-10 mM in a 6-10 mL reservoir) in h3- or d3-acetonitrile was
pumped through a 0.5 mm path length BaF2 flow cell. The sample
was irradiated at 266 nm, using an Nd:YAG laser (50 Hz repetition
rate, 0.5 mJ/pulse). Signal collection was achieved using a dispersive
infrared spectrometer (broadband MoSi2 source, resolution ap-
proximately 16 cm-1).
FIGURE 1. Transient spectra obtained following 308 nm laser flash
photolysis of 1 (0.07 mM) in 3-methylpentane at 298 K. Each spectrum
is obtained over a 10 ns time interval.
tions and energies were performed at the B3LYP/6-31G* level
of theory for all species,12 and vibrational frequencies were
calculated at the same level of theory, corrected by factors of
0.9614 (frequencies), 0.9806 (zero-point energies), and 0.9989
(enthalpic corrections to 298 K) as recommended by Scott and
Radom.13 Transition states for nitrene and azide rearrangements
were found to have a single imaginary frequency, corresponding
to the appropriate reaction coordinate. Natural population
analysis (NPA)14 was performed on critical intermediates. The
relative energetics of 2 and 3 were further refined using
CASSCF calculations up to the CAS(10,9)/6-31G* level of
theory.15
Results
Laser Flash Photolysis. Room temperature irradiation of 1
in 3-methylpentane at 308 nm yielded the transient spectra
shown in Figure 1. The spectra show two temporally distinct
featuressa partially structured band with maxima at 523 and
557 nm, and a broader feature with an absorbance maximum at
435 nm. The 523 and 557 nm signals decay at the same rate
(within experimental uncertainty, Figure 2), leading us to believe
that the same transient species is responsible for both signals.
The bands at 520-560 nm are assigned to the triplet nitrene
(3). Such an assignment is consistent with other triplet arylni-
trene spectra, although somewhat more intense than usual for
simple aryl nitrenes,5,6 but not necessarily for more complex
heteroarylnitrenes.16 The growth rate of the signal at 557 nm
could also be measured at small timescales (Figure 2 inset).
Theoretical Calculations
Density functional theory (DFT) calculations were performed
using the GAUSSIAN 03 program suite.11 Geometry optimiza-
(9) For example, see Martin, C. B.; Shi, X.; Tsao, M.-L.; Karweik, D.;
Brooke, J.; Hadad, C. M.; Platz, M. S.; J. Phys. Chem. B 2002, 106, 10263.
Temperature control was achieved by utilizing the quartz cryostat/
thermostabilized nitrogen stream system described in ref 5c.
(10) For example, see: Martin, C. B.; Tsao, M-L.; Hadad, C. M.; Platz,
M. J. J. Am. Chem. Soc. 2002, 125, 7226.
(11) 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 C.02; Gaussian, Inc.: Wallingford, CT, 2004.
(12) (a) Becke, A. D. Phys. ReV. A 1988, 38, 3098. (b) Becke, A. D. J.
Chem. Phys. 1993, 98, 5648. (c) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV.
B 1988, 37, 785.
(13) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502.
(14) (a) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys.
1985, 83, 735. (b) Reed, A. E.; Weinhold, F.; Curtiss, J. A. Chem. ReV.
1988, 88, 899.
(15) (a) Karney, W. L.; Borden, W. T. J. Am. Chem. Soc. 1997, 119,
1378. (b) Johnson, W. T. G.; Sullivan, M. B.; Cramer, C. J. Int. J. Quantum
Chem. 2001, 85, 492.
(16) Cerro-Lopez, M.; Gritsan, N. P.; Zhu, Z.; Platz, M. S. J. Phys. Chem.
A 2000, 104, 9681.
9024 J. Org. Chem., Vol. 71, No. 24, 2006

Photochemistry of 4-Azidopyridine-1-oxide
FIGURE 2. Decay of signal at 557 nm following 308 nm LFP of 1
(0.07 mM) in 3-methylpentane at 298 K. Inset: Short time scale growth
of signal at 557 nm following LFP of 1 in 3-methylpentane. Signals
were averaged over four acquisitions each
FIGURE 3. Dependence of the rate of decay of triplet nitrene signal
(3, 552 nm) and/or rate of growth of 6 (413 nm) on the solution
concentration of 1 (pseudo-first-order kinetic plot).
The growth rate was found to be independent of the starting
azide concentration and some 2 orders of magnitude greater
than the decay rate for the signal at 557 nm. Therefore, to
reasonable approximation, the growth and decay rates for this
signal were treated as being temporally separable. It is further
assumed that, like other arylnitrenes, the triplet species is
generated by intersystem crossing from the nitrene singlet state,
rather than from an azide triplet state.17 Thus, the observed rate
coefficient for intersystem crossing (kISC) from the singlet nitrene
was determined to be (2.14 ( 0.08) × 107 s-1, a value that lies
in the range of values reported for substituted phenylnitrenes.5,6
The feature at 435 nm has a growth rate which was
demonstrated to be dependent on the starting concentration of
azide and the same (within experimental uncertainty) as the
decay rate for the triplet (Figures 2 and 3). In addition, this
signal did not decay within the longest observable timebase of
the instrument. On this basis, we have assigned the signal at
435 nm to the compound 4,4′-azo-bis(pyridine-1-oxide), 6. It
is noteworthy that the spectra obtained in 3-methylpentane
remain unchanged with the addition of 0.5 M diethylamine to
solution, further evidence against the formation of 5, which
would be rapidly trapped to form the azepine product.18
Photolysis of 4-azido-pyridine-1-oxide in dichloromethane at
298 K yielded transient spectra that are effectively the same as
those shown in Figure 1, except that the absorbances of the
azo-compound are shifted (Table 1). The significant solvato-
chromic behavior of compound 6 has been reported previously
and is consistent with our observations.19 An interesting
difference between the two experiments is the presence of a
very weak and short-lived transient signal observable at 493
nm in dichloromethane. The rate of decay for the signal at 493
nm is (1.9 ( 0.3) × 107 s-1, in reasonable agreement with the
growth rate measured for the signal at 554 nm, (2.41 ( 0.08)
× 107 s-1. On this basis, we tentatively assign the signal at
493 nm to the singlet nitrene, 2.
In order to further confirm our assignment of spectral peaks
to the triplet (3) and azo-dimer (6), we measured the growth
rate of the signal at 554 nm in dichloromethane as a function
of temperature. The LFP signal was obtained over a temperature
range of 179-298 K, and the value of kISC varied by less than
10% over this range, effectively temperature independent, as
would be expected for an intersystem crossing process.5,6 In
addition, the observed value for kISC at 298 K is not strongly
dependent on solvent polarity (Table 1).
The dependence of decay kinetics of the 554 nm signal was
also measured as a function of azide concentration (Figure 3)
and, as expected, exhibited nonlinear behavior at low concentra-
tions. Regression analysis of the linear (pseudo-first-order)
region of the curve yielded a rate coefficient for the triplet
nitrene-azide coupling reaction of kC ) (1.96 ( 0.13) × 109
M-1 s-1 in dichloromethane at 298 K. This compares with the
slightly higher value obtained in 3-methylpentane at 298 K
(Table 1).
Attempts to observe quenching of the triplet signal in the
presence of oxygen were unsuccessful. LFP of a solution of 1
(0.13 mM) in dichloromethane in the presence and absence of
oxygen showed only a marginal increase in the observed decay
rate of the 554 nm signal in the presence of oxygen (Table 1),
but since both derived second-order rate coefficients lie within
the experimental uncertainty of the rate coefficient for dimer-
ization over a range of concentrations of 1 (0.02-0.14 mM), it
seems unlikely that oxygen trapping of 3 competes with
dimerization to any significant extent.
Attempts to observe the pyridine-nitrene ylide 7 were
(17) Marcinek, A.; Leyva, E.; Whitt, D.; Platz, M. S. J. Am. Chem. Soc.
1993, 115, 8609.
(18) (a) DeGraff, B. A.; Gillespie, D. W.; Sundberg, R. J. J. Am. Chem.
Soc. 1974, 96, 7491. (b) Li, Y.-Z.; Kirby, J. P.; George, M. W.; Poliakoff,
M.; Schuster, G. B. J. Am. Chem. Soc. 1998, 110, 8092.
(19) Muniz-Miranda, M.; Pergolese, B.; Sbrana, G.; Bigotto, A. J. Mol.
Struct. 2005, 744-747, 339.
unsuccessful - photolysis of 1 in neat pyridine at 298 K
J. Org. Chem, Vol. 71, No. 24, 2006 9025

Hostetler et al.
TABLE 1. Kinetic Data for Photochemistry (Laser Flash Photolysis) of 4-Nitrenopyridine-1-oxide
max(2) max(3) max(6)
λ
λ
λ
kISC
kC
solvent
(nm)
(nm)
(nm)
(107 s-1
)
(109 M-1 s-1
)
3-methylpentane
not obs.
523, 557
435
2.40 ( 0.07
5.6 ( 0.4a
5.0 ( 0.5b
3-methylpentane + 0.5 M HNEt2
dichloromethane
523, 559
522, 554
438
417
493
2.23 ( 0.04
1.96 ( 0.13c
1.86 ( 0.01d
1.96 ( 0.01d
dichloromethane + O2
dichloromethane + 0.5 M HNEt2
acetonitrile
521, 555
517, 547
526, 557
420
415
427
not obs.
not obs.
2.49 ( 0.05
2.3 ( 0.1e
pyridine
a Measured from signal decay at 557 nm. b Measured from signal growth at 435 nm. c Determined from combined data for decay at 554 nm and growth
at 417 nm. d Determined from decay at 417 nm at a single concentration of 1. e Estimated from growth at 415 nm and decay at 547 nm at a single concentration
of 1 in acetonitrile
4 were determined computationally at the B3LYP/6-31G* level
of theory for the isolated species (i.e., no solvent).
On the basis of the spectra in Figure 4, the kinetics of the
signals at 1370 (assigned to 3) and 1600 (assigned to 6) cm-1
were investigated (Figure 5). Attempts to obtain kinetic data
for the signals at 1329 (6) and 1575 (3) cm-1 were unsuccess-
ful: the former due to spectral interference arising from
bleaching of the azide at 1305 cm-1, the latter due to spectral
interference due to partial overlap of the feature at 1600 cm-1
.
The TRIR signal at 1370 cm-1 shows a rapid growth,
followed by a considerably slower decay. The growth kinetics
for this signal cannot be reliably determined due to the time
scale constraints of the instrumental method, but it is estimated
to be of the order 107 s-1sbroadly consistent with ISC rate
coefficients measured by LFP. The observed first-order rate
constant for the decay of the signal was determined to be (7.6
( 1.1) × 106 s-1, which based on a concentration for 1 of about
4-5 mM (and assuming pseudo-first-order kinetics) gives a
value for the second-order rate coefficient of reaction of 3 with
1 which is consistent with the value obtained from UV-visible
LFP studies (1.5 × 109 M-1 s-1). What is most important to
note is that the rate coefficient for the growth of signal at 1600
cm-1 (corresponding to the formation of 6 from 1 and 3) has
the same value, within the limits of experimental uncertainty.
An additional study of the region 1900-1700 cm-1 following
laser excitation of 1 in h3-acetonitrile solution showed no
discernible signals. Calculated vibrational spectra of the dide-
hydroazepine (or ketenimine) 5 indicate that this species should
have a diagnostic absorbance at 1881 cm-1 and that the
benzazirine 4 should have an absorbance at 1744 cm-1. The
absence of these signals indicates that singlet nitrene chemistry
plays a relatively minor role in the overall photochemistry of
1.
Product Analysis. HPLC analysis of solutions of 1 photo-
lyzed under argon show only two significant peaks, one
corresponding to the starting material 1 and the other to 6. No
other peaks were observed to any significant degree. Similarly,
1H NMR spectra of the d3-acetonitrile solution shows only two
compounds to any significant extent. The primary photoproduct
was isolated, and confirmed to be the azo-compound 6.
Theoretical Calculations. Calculations were performed at
the B3LYP/6-31G* level of theory for 4-nitrenopyridine-1-
oxide, 4-pyridylnitrene, and phenylnitrene (as a comparator) to
determine the relative energies of the species 2, 3, 4, and 5 (and
their analogues), as well as the transition states leading to the
formation of these species where applicable. In addition,
calculations for species arising from N-oxide chemistry were
also performed, specifically those arising from cyclization of
FIGURE 4. Top half: simulated differential IR spectra for disappear-
ance of 1 and appearance of 3 and 6, based on calculated vibrational
frequencies. Bottom half: transient differential IR spectra (time slices)
following 298 K irradiation of a d3-acetonitrile solution of 1 (5 mM),
at 266 nm. Experimental spectra have been smoothed for reasons of
clarity.
generated spectra that were qualitatively similar to those
generated in 3-methylpentane and dichloromethane, the only
quantitative change in absorbance maxima for 3 and 6 due to
solvatochromic shifts (Table 1).
Time-Resolved Infrared Spectroscopy. Irradiation of a
solution of 1 in d3-acetonitrile yielded the transient differential
spectra shown in Figure 4. The differential spectra observed
are consistent with the simulated spectra for the bleaching of 1
and the formation of 3 and 6. The simulated spectra in Figure
9026 J. Org. Chem., Vol. 71, No. 24, 2006

Photochemistry of 4-Azidopyridine-1-oxide
FIGURE 5. Time-resolved behavior of TRIR signals at (a) 1600 cm-1 and (b) 1370 cm-1, following laser flash photolysis of 1 (5 mM) in d3-
acetonitrile at 298 K. The 1370 cm-1 signal corresponds to the triplet nitrene 3, and the 1600 cm-1 signal corresponds to the azo-compound 6.
TABLE 2. Summary of Calculated Relative Thermochemical Data (B3LYP/6-31G*, kcal/mol)
4-nitrenopyridine-1-oxide
4-pyridylnitrene
phenylnitrene
B3LY P/6-31G*
CASP T2/6-31G*a
0 Kb 298 Kc
0.0 0.0
B3LYP/6-31G*
B3LYP/6-31G*
species
0 Kb
298 Kc
0 Kb
298 Kc
0 Kb
triplet state (3A2)d
0.0
9.8
0.0
9.9
0.0
16.7
33.1e
33.1
19.7
21.9
12.7
0.0
16.7
33.4e
32.8
19.7
21.7
12.7
0.0
19.1
37.3
27.7
20.7
25.9
17.8
open-shell singlet (1A2)
closed-shell singlet (1A1)
singlet-benzazirine TS
benzazirine
15.7
47.2
31.7
19.9
23.6
14.7
15.8
47.5
31.4
19.8
23.4
14.8
34.4
34.0
29.7
36.7
24.3
0.0
34.3
33.8
29.7
36.5
24.2
0.0
benzazirine-ketenimine TS
ketenimine
azide singlet state (S0)d,f
azide triplet state (T1)f
N-oxide cyclization TSf
cyclized N-oxidef
44.6
58.2
39.4
45.3
58.1
39.4
1
a Ref 15. b Includes scaled ZPEs. c Includes scaled ∆Hvib
azide S0 state.
.
d Defined as zero. e This species has C1 symmetry, therefore A state. f Calculated relative to
oxygen onto the aromatic ring, a previously reported photo-
chemical reaction for aromatic N-oxides.20 The energetics of
the open-shell singlet states of the nitrenes listed above were
estimated according to the method of Johnson et al.,15b whereby
a UDFT optimization based on the RDFT nitrene wavefunction
with promotion of a single electron into the LUMO yielded a
mixed singlet state with an <S2> expectation value of ap-
proximately 1, presumed to consist of a 50:50 mixture of open-
shell singlet and triplet states. The energy of the open-shell
singlet state was then estimated by the sum method (eq 1).
at least a qualitative (if not semiquantitative) description of the
three systems.
All of the nitrene species in this study exhibit C2V symmetry
with the sole exception of the closed-shell singlet (S2) state of
4-pyridylnitrene, which exhibits a single imaginary frequency
under such symmetry constraints. The level of theory employed
in this study indicates that this species is slightly distorted (C1
symmetry) from a truly planar formsthe potential energy well
for this state is quite shallow. The other noteworthy feature of
this species is that the transition state for cyclization to the
benzazirine species is quite early compared to that for phe-
nylnitrene cyclization. B3LYP/6-31G* calculations indicate that
only a small deviation from planarity is required to reach the
transition state. By and large, however, the energetic profiles
calculated for 4-pyridylnitrene and phenylnitrene at this level
of theory are quite similar. It should be noted, however, that
the transition state for nitrene cyclization will retain a certain
amount of biradical character that is not readily modeled by
DFT methods of this type.
E(OSS) ) 2E(50:50 mix) - E(T)
(1)
The results obtained for the three nitrene species are sum-
marized in tabular form in Table 2 and in diagrammatic form
in Figure 6. All energies are computed relative to the triplet
ground state for each nitrene species. The thermochemical data
calculated for phenylnitrene at this level of theory compares
reasonably well with data generated at the CASPT2/6-31G*
level of theory,15 and so this approach should provide us with
Consideration of N-oxide cyclization in 1 was considered in
terms of the relative energies of the azide S0 state and those of
its cyclized intermediates (Table 2). The data indicate that such
a reaction is enthalpically unfavorable to a significant degree
(20) (a) Albini, A.; Alpegiani, M. Chem. ReV. 1984, 84, 43. (b) Shi, X.;
Poole, J. S.; Emenike, I.; Burdzinski, G.; Platz, M. S. J. Phys. Chem. A
2005, 109, 1491.
J. Org. Chem, Vol. 71, No. 24, 2006 9027

Hostetler et al.
FIGURE 6. Relative energies calculated at 0 K (including ZPEs) at the B3LYP/6-31G* level of theory for the unimolecular chemistries of
phenylnitrene (blue, dashed), 4-pyridylnitrene (green, dotted), and 4-nitrenopyridine-1-oxide (red, solid). All energies shown relative to the respective
3
triplet A2 states.
SCHEME 2. Electronic Structures of
1
3
4-Nitrenopyridine-1-oxide A2 and A2 States
and is unlikely to occur under normal thermal conditions. Such
an observation does not necessarily preclude this reaction,
particularly within the photochemical realm, but it does provide
additional support for the observed product distributions in this
study.
FIGURE 7. Natural population analysis of spin and charge density
distribution in the 3A2 state of 3, B3LYP/6-31G* level of theory. Charge
density shown in parentheses.
Given the known stability of aminoxyl radicals,22 it is
reasonable to assume that a significant amount of spin density
in the ring lies on the N1 and O atoms. A natural population
Discussion
On the basis of both the observed data and previous product
distribution studies, the cyclization of 2 to yield ketenimine 5
is not an energetically favorable process. In previously studied
systems, this may be ascribed to steric hindrance to cyclization
(e.g., o,o′-disubstituted phenylnitrenes).21 In this particular case,
no such effect is present, and so therefore we must consider
the effect of electron configuration.
3
analysis (NPA) was performed on the A2 state (3) to obtain a
measure of the spin density distribution, and the results are
shown in Figure 7. The data indicate that indeed there is a
significant degree of spin character on O. Such results are
consistent with those previously reported by other workers.23
On this basis, the diminution of spin density at sites o- to the
nitreno group relative to the analogous phenylnitrene system
will make cyclization a less favored process. No meaningful
NPA data may be determined for the 1A2 state, due to the high
degree of spin contamination in the computed wavefunction,
It is well-established that the lowest singlet state of phenylni-
trene (1A2 state) is perhaps more appropriately considered as a
iminyl-cyclohexadienyl biradical system5,15 rather than a “for-
mal” singlet nitrene. In the case of the 4-nitrenopyridine-1-oxide
system, the 1A2 state for 2 may be considered as an iminylami-
noxyl biradical (Scheme 2), as in principle can the 3A2 state of
3.
(22) (a) Lebvedev, Y. A.; Rosantzev, E. G.; Kalashnikova, L. A.;
Lebvedev, V. P.; Neiman, M. B.; Apin, A. Y. Dokl. Akad. Nauk SSSR (Engl.
Trans.) 1966, 40, 460. (b) Lebvedev, Y. A.; Rosantzev, E. G.; Neiman, M.
B.; Apin, A. Y. Dokl. Akad. Nauk SSSR (Engl. Trans.) 1966, 40, 460. (c)
Mahoney, L. R.; Mendenhall, G. D; Ingold, K. U. J. Am. Chem. Soc. 1973,
95, 8610.
(21) Gritsan, N. P.; Gudmundsdottir, A. D.; Tigelar, D.; Platz, M. S. J.
Phys. Chem. A 1999, 103, 3458.
(23) Pietrzycki, W.; Tomasik, P. Pol. J. Chem. 1998, 72, 893.
9028 J. Org. Chem., Vol. 71, No. 24, 2006

Photochemistry of 4-Azidopyridine-1-oxide
but it is reasonable to assume that the spin densities in the 1A2
and A2 would be at least qualitatively similar.
-5.2 kcal/mol for ∆Hr respectively). Once the spin density in
1
3
the A2 state is annihilated, these three systems share similar
reactivity.
At the B3LYP/6-31G* level of theory used in this study, the
singlet-triplet splitting, EST, for 2/3 was found to be 9.9 kcal/
mol. This is some 6-7 kcal/mol smaller than those calculated
for phenylnitrene and 4-pyridylnitrene. Higher level calculations
at the CASSCF(10,9)/6-31G* level of theory predict EST values
as low as 5.6-6.6 kcal/mol, depending on the nature of the
orbitals chosen for the active space (see Supporting Information).
The much smaller singlet-triplet energy gap would result in
an enhanced rate of intersystem crossing observed for this
species relative to phenyl nitrene, and this is observed experi-
mentally.24
Formation of 6 from the triplet 3A2 state of 4-nitrenopyridine-
1-oxide (3) and parent azide (1) appears to be a diffusion-limited
process, and so we would expect most of the complex product
mixtures obtained in previous studies3 to be the result of
secondary thermal or photochemical reactions of 6, although
such product distributions may well be observable under
“infinite dilution” conditions. Studies of the photochemistry of
3-azidopyridine-1-oxide and its corresponding nitrene are
underway. On the basis of the arguments above, it is anticipated
that this species should react more like phenylnitrene, as the
positioning of the N-O group on the aromatic ring is less
amenable to spin delocalization in the form of an iminyl-
aminoxyl resonance contributor.
Upon cyclization, the stabilization afforded by additional
delocalization of spin density to oxygen is lost as spin density
is annihilated. The loss of this stabilizing influence upon
cyclization may go part of the way to explaining why this
process is much more energetically disfavored for 2 than for
analogous forms: the barrier to cyclization is calculated to be
24.2 kcal/mol (∆Hr ) 19.9 kcal/mol) for 2, but only 16 kcal/
mol (∆Hr ) 3.0 and 4.2 kcal/mol) for the other systems. This
level of theory tends to significantly overestimate the height of
the barrier, as it both overestimates the stability of the nitrene
and underestimates the stability of the transition state, which
may retain significant biradical character not readily modeled
by DFT methods. The CASPT2 value for the energy barrier
for nitrene cyclization is 8.6 kcal/mol (the experimental value
is 5.6 kcal/mol).5c The variation between EST calculated by DFT
and CASPT2 methods is relatively systematic, and so we may
conclude that the barrier to cyclization of 2 is significantly larger.
It is also worth noting that the transition state for cyclization of
2 is also significantly later as might be expected from applying
the Hammond postulate to Figure 6. The C-C-N angle
defining the azirene ring lies between 68 and 72 for all of the
benzazirine products, but is 95° in the transition state for
cyclization of 2, 105° for cyclization of phenylnitrene, and 126°
for 4-pyridylnitrene.
Conclusions
The photochemistry of 4-azidopyridine-N-oxide is dominated
by fragmentation of the azide moietysno products consistent
with N-oxide cyclization photochemistry were observed, and
calculations indicate that this process is thermodynamically
disfavored. The singlet nitrene (1A2) formed undergoes rapid
ISC to generate the triplet (3A2) species, which dimerizes with
the azide at a diffusion-limited rate to generate the diazo-
compound 6. No singlet nitrene chemistry was observed, and
calculations indicate that this is due to the significant barrier to
nitrene cyclization. The barrier is significantly larger than those
calculated for similar arylnitrene systems and is due to the
1
3
stabilization of the A2 and A2 states by spin delocalization
due to resonance contributors with iminyl-aminoxyl biradical
character.
Acknowledgment. The authors thank Professor Matthew
Platz of The Ohio State University for his gracious provision
of instrument time as well as Dr. Sarah Mandel for her
assistance. The authors also thank Professors Chris Hadad and
William Karney for helpful discussions regarding computational
methods. Calculations were made possible by use of the Beowulf
cluster at the Center for Computational Nanoscience at Ball State
University. Financial support in the form of a Ball State
University Internal Grant and the Lilly II endowment are also
gratefully acknowledged.
The energetics for subsequent steps (ring-opening from
benzazirine to ketenimine) is much the same for the three
systems: the barriers to ring expansion vary between 2.2-7.0
kcal/mol for all three systems, and enthalpies of reaction vary
between -5.2 and -7.0 kcal/mol. CASPT2 and DFT energies
for the phenylnitrene system show good agreement (5.2 and
3.6 kcal/mol for the activation barrier to expansion, -2.9 and
Supporting Information Available: Synthetic methodology for
the starting material, additional spectroscopic and kinetic data
obtained in this study, and computational output in the form of
optimized geometries, energies, and frequencies for species relevant
to this study. This material is available free of charge at
http://pubs.acs.org.
(24) kISC for phenylnitrene is (3.2 ( 0.3) × 106 s-1 (ref 5b), close to an
order of magnitude lower than the value for 4-nitrenopyridine-1-oxide. It
is noteworthy that kISC may be increased three orders of magnitude with
appropriate π-donor substituents (ref 6b).
JO061259O
J. Org. Chem, Vol. 71, No. 24, 2006 9029