Photochemistry of 2-naphthoyl azide. An ultrafast time-resolved UV-Vis and IR spectroscopic and computational study

Article abstract of DOI:10.1021/ja109098w

The photochemistry of 2-naphthoyl azide was studied in various solvents by femtosecond time-resolved transient absorption spectroscopy with IR and UV-vis detection. The experimental findings were interpreted with the aid of computational studies. Using polar and nonpolar solvents, the formation and decay of the first singlet excited state (S₁) was observed by both time-resolved techniques. Three processes are involved in the decay of the S₁ excited state of 2-naphthoyl azide: intersystem crossing, singlet nitrene formation, and isocyanate formation. The lifetime of the S₁ state decreases significantly as the solvent polarity increases. In all solvents studied, isocyanate formation correlates with the decay of the azide S ₁ state. Nitrene formation correlates with the decay of the relaxed S₁ state only upon 350 nm excitation (S₀ → S ₁ excitation). When S_n (n ≥ 2) states are populated upon excitation (λ_ex = 270 nm), most nitrene formation takes place within a few picoseconds through the hot S₁ and higher singlet excited states (S_n) of 2-naphthoyl azide. The data correlate with the results of electron density difference calculations that predict nitrene formation from the higher-energy singlet excited states, in addition to the S₁ state. For all of these experiments, no recovery of the ground state was observed up to 3 ns after photolysis, which indicates that both internal conversion and fluorescence have very low efficiencies.

Full text of DOI:10.1021/ja109098w

ARTICLE
pubs.acs.org/JACS
Photochemistry of 2-Naphthoyl Azide. An Ultrafast Time-Resolved
UVꢀVis and IR Spectroscopic and Computational Study
Jacek Kubicki,† Yunlong Zhang,‡ Shubham Vyas,‡ Gotard Burdzinski,† Hoi Ling Luk,‡ Jin Wang,‡,#
Jiadan Xue,‡ Huo-Lei Peng,‡ Elena A. Pritchina,^ Michel Sliwa,§ Guy Buntinx,§ Nina P. Gritsan,^
Christopher M. Hadad,*,‡ and Matthew S. Platz*,‡
†Quantum Electronics Laboratory, Faculty of Physics, Adam Mickiewicz University, Umultowska 85, 61-614 Poznaꢀn, Poland
‡Department of Chemistry, The Ohio State University, 100 West 18th Avenue, Columbus, Ohio, 43210, United States
§Laboratoire de Spectrochimie Infrarouge et Raman (LASIR), UMR 8516/CNRS ꢀUniversitꢀe Lille 1 Sciences et Technologies,
59655 Villeneuve d’Ascq, France
^Institute of Chemical Kinetics and Combustion, Siberian Branch of Russian Academy of Science and Novosibirsk State University,
630090 Novosibirsk, Russia
S Supporting Information
b
ABSTRACT: The photochemistry of 2-naphthoyl azide was studied in
various solvents by femtosecond time-resolved transient absorption spec-
troscopy with IR and UVꢀvis detection. The experimental findings were
interpreted with the aid of computational studies. Using polar and nonpolar
solvents, the formation and decay of the first singlet excited state (S1) was
observed by both time-resolved techniques. Three processes are involved in
the decay of the S1 excited state of 2-naphthoyl azide: intersystem crossing,
singlet nitrene formation, and isocyanate formation. The lifetime of the S1
state decreases significantly as the solvent polarity increases. In all solvents
studied, isocyanate formation correlates with the decay of the azide S1 state.
Nitrene formation correlates with the decay of the relaxed S1 state only
upon 350 nm excitation (S0 f S1 excitation). When Sn (n g 2) states are
populated upon excitation (λex = 270 nm), most nitrene formation takes
place within a few picoseconds through the hot S1 and higher singlet excited
states (Sn) of 2-naphthoyl azide. The data correlate with the results of electron density difference calculations that predict nitrene
formation from the higher-energy singlet excited states, in addition to the S1 state. For all of these experiments, no recovery of the
ground state was observed up to 3 ns after photolysis, which indicates that both internal conversion and fluorescence have very low
efficiencies.
trigonal planar geometry and cyclic oxazirine.5,6,9 Nevertheless,
1. INTRODUCTION
singlet carbonyl nitrenes do follow traditional nitrene reactions,
The Curtius rearrangement can be initiated with either light or
such as insertion into CꢀH bonds and addition to CdC bonds.1,2,9
heat.1,2 Irradiation of acyl azides in solution produces two prima-
The N O bonding interaction explains why the closed-shell
3 3 3
ry photoproducts: isocyanate (photo-Curtius rearrangement) and
nitrene.1,2
singlet states of aroylnitrenes are more stable than the corre-
sponding triplet isomers.5,6,9 Similar effects have been noted for
phosphorylnitrenes.10 As isocyanate formation competes with
nitrene formation, the efficiency of acyl azides as photoaffinity
labels is reduced. On the other hand, isocyanates are important
monomers in polyurethane synthesis and are easily produced by
Curtius rearrangement.11 Recently, it was shown that a cyanate
product is also formed upon carbonyl azide decomposition but in
very low yield relative to the isocyanate isomer (∼1%).12
As singlet nitrenes are very reactive species, azides are often used
as photoaffinity labels.3 It is well-known that alkyl- and arylnitrenes
have triplet ground states,2,4 although aroylnitrenes have singlet
ground states.1,2,5ꢀ9 Theory predicts that the CON moiety in the
singlet ground state of formyl-, acetyl-, benzoyl-, and naphthoyl-
nitrenes have geometries that are intermediate between an idealized
Received: October 10, 2010
Published: May 09, 2011
r
2011 American Chemical Society
9751
dx.doi.org/10.1021/ja109098w J. Am. Chem. Soc. 2011, 133, 9751–9761
|

Journal of the American Chemical Society
ARTICLE
Thermal Curtius rearrangement of various acyl azides are well-
known,13ꢀ15 and this process is a concerted reaction of azides, which
bypasses nitrene intermediates.1,9,15,16 Thermolysis of pivaloyl azide
in cyclohexane does not generate a trappable nitrene in more than
minute quantities and instead produces isocyanate in nearly quanti-
tative yield.15 On the contrary, photolysis of pivaloyl azide gives
isocyanate only in 40% yield as well as products derived from
interception of a nitrene intermediate.13,14 Photolysis of benzoyl
azide and a series of its derivatives afforded isocyanates in similar
yields (40ꢀ50%) along with adducts derived from nitrenes.17,18
Therefore, carbonylnitrenes are clearly produced upon photolysis of
carbonyl azides, in contrast to their thermolysis.
The photochemical Curtius rearrangement is a concerted
reaction to form the isocyanate product and does not involve
either thermally relaxed nitrenes5 or hot nitrenes.19 We view this
as an example of a “Rearrangement in the Excited State” (an
RIES20) mechanism.
The photochemistry of benzoyl azide was previously studied
by nanosecond time-resolved infrared (ns TRIR) spectroscopy.5
In this seminal study, the formation of isocyanate and nitrene
products could not be resolved with the experimental time
resolution available (50 ns) at that time. These experiments
did demonstrate, however, that benzoylnitrene, in the ground
state, does not isomerize to isocyanate in solution under condi-
tions typical of laser flash photolysis studies.
Figure 1. (a) TRIR spectra produced upon 350 nm excitation of
2-NpCON3 (c ∼ 4.0 ꢁ 10ꢀ3 M) in chloroform. The normalized
steady-state IR absorption spectrum (FTIR) of 2-NpCON3 in chloro-
form is shown as a dotted line. (b) The decay of the S1 state recorded at
2096 cmꢀ1, and the kinetics of the signal at 2143 cmꢀ1 (the depopula-
tion band of the ground state).
Xanthone-sensitized ns TRIR experiments21 demonstrated that
an excited singlet state of benzoyl azide must be the precursor of
phenyl isocyanate, consistent with the earlier findings of Schuster
and co-workers with substituted aroyl azides containing internal
sensitizers.7 Similar conclusions were drawn by Toscano et al. in a
related study of the photochemistry of N-benzoyl and N-acetyl
dibenzothiophene sulfilimines with ns TRIR techniques.21,22
Triplet-sensitized photolysis of 2-naphthoyl azide (2-NpCON3)
in cyclohexane produces only a very small amount of isocyanate
(5%) as compared to direct excitation of 2-NpCON3 (∼58%).8 It
was concluded that the isocyanate formed upon triplet sensitization
was produced by inadvertent direct irradiation of 2-NpCON3.
Thus, an excited singlet state of the acyl azide is the likely pre-
cursor of the isocyanate. However, this hypothesis has never been
confirmed by direct correlation of the decay of the excited singlet
state of the acyl azide with the corresponding growth of the
isocyanate. To test this hypothesis, we sought unique vibrational
marker bands for all of the relevant species using femtosecond
TRIR spectroscopy. Marker bands in the IR region can allow
independent monitoring of the dynamics of different species in
cases where their electronic spectra may significantly overlap.
Fortunately, the NCO vibration of the isocyanate moiety is IR
active and its intensity is large. The singlet excited states of acyl
azides have prominent IR bands as well.19 Recently, we reported
observation of the decay of the S1 state of an aryldiazirine and the
formation of its isomeric diazo compound and related singlet
carbene using this approach.23
and Schuster’s findings (<2 ns).8 2-NpCON3 has a convenient
chromophore for UVꢀvis studies, and intense IR markers. Thus,
we were encouraged to study the photochemistry of 2-NpCON3 in
different solvents using both the fs TRIR and fs UVꢀvis transient
absorption techniques, as well as with modern quantum chemical
calculations.
2. EXPERIMENTAL RESULTS
2.1. Ultrafast IR and UVꢀVis Studies of the Photochem-
istry of 2-Naphthoyl Azide with 350 nm Excitation. Ultrafast
IR Studies. Figure 1 and Figure S2 (see Supporting Information)
present the IR transient absorption spectra produced upon
350 nm excitation of 2-NpCON3 in selected solvents. Analysis
of the electronic absorption spectrum of 2-NpCON3 (Figure S3,
Supporting Information) and time-dependent B3LYP (TD-
B3LYP) calculations predict that 350 nm light directly populates
the S1 state (vide infra, Table S7, Supporting Information). In all
solvents employed, photoexcitation of 2-NpCON3 generates a
transient 2100 cmꢀ1 band, which is formed within the laser pulse
(300 fs, Figure S4, Supporting Information). This band was
assigned previously to the S1 state of 2-NpCON3.19 This assign-
ment is supported by theory (vide infra). Indeed, calculations
predict that the vibrational frequency of the N3 stretch in the S1
state of 2-NpCON3 is lower by 72 cmꢀ1 and its intensity is higher
by about a factor of 3, relative to the S0 state (Figure S5, Table S2,
Supporting Information).
The photochemistry of 2-NpCON3 has been studied exten-
sively,7,8 but the dynamics of isocyanate and nitrene formation were
not previously determined. In our preliminary report, the photo-
chemistry of benzoyl azide (PhCON3), 2-naphthoyl azide (2-
NpCON3), and pivalolyl azide (t-BuCON3) was examined in
chloroform at ambient temperature solely by means of femtose-
cond (fs) TRIR spectroscopy.19 We reported that the first singlet
excited states of the acyl azides are the precursors of the corre-
sponding isocyanate products.19 The lifetime of the S1 state of
2-NpCON3 was found to be about 120 ps,19 consistent with Autrey
The decay of the 2100 cmꢀ1 band can be adequately fit to a
monoexponential decay upon 350 nm excitation in all solvents
used in this study (Figure 1 and Figure S2). The lifetimes of the
S1 state of 2-NpCON3, determined using the fs TRIR transient
absorption technique, are presented in Table 1. This table also
9752
dx.doi.org/10.1021/ja109098w |J. Am. Chem. Soc. 2011, 133, 9751–9761

Journal of the American Chemical Society
ARTICLE
Table 1. Solvent Properties and Time Constants (τ) Describing the Decay of the Excited S1 State of 2-NpCON3 Obtained from fs
TRIR and fs UVꢀVis Spectroscopic Detection
decay of the S1 state
solvent
f(ε,n)a
initially excited state
τ (ps) IR detection 2100 cmꢀ1
τ (ps) UVꢀvis detection 350ꢀ650 nm
cyclohexane
ꢀ0.0007
Sn
14.0 ( 1.6,c 580 ( 60 d
800 ( 120 d,e
ꢀ
b
e
S1
ꢀ
b
carbon tetrachloride
chloroform
0.0116
0.1550
Sn
8.6 ( 0.8,c 570 ( 50 d
690 ( 60 d,e
ꢀ
S1
425 ( 12 d,e
6 ( 2, c 130 ( 20 d
ꢀ
e
Sn
6.8 ( 1.3,c,f 120 ( 30 d,f
133 ( 6 d,e
b
e
S1
b
dichloromethane
acetonitrile
0.2305
0.3303
Sn
6.5 ( 1.1,c 70 ( 6 d
3.0 ( 0.7,c 19 ( 4 d
28 ( 5 d,e
ꢀ
Sn
3 ( 1, c 23 ( 4 d
26 ( 4 d,e
b
e
S1
methanol
0.3327
Sn
3.6 ( 0.9,c 19 ( 3 d
2.0 ( 0.2,c 24 ( 3 d
b
a f(ε,n) = (ε ꢀ 1)/(2ε þ 1) ꢀ (n2 ꢀ 1)/(2n2 þ 1): n = refractive index, ε = dielectric constant.24 b 270 nm excitation was used to populate Sn (n > 1)
excited states. With this excitation wavelength, biexponential decays were observed. c Vibrational cooling (VC) and Curtius rearrangement of the hot S1.
d Decay of vibrationally relaxed S1 state. e When 340 or 350 nm excitation was used, a monoexponential decay of S1 was observed. f τ in chloroform was
taken from ref 19.
Table 2. The Efficiency of Isocyanate (ΦNCO), Nitrene (Φnitrene), and Triplet Azide (ΦT) Formation and the Corresponding Rate
Constants in Three Selected Solvents
kNCOb (ꢁ 109 sꢀ1
)
ΦT
kISCf (ꢁ 109 sꢀ1
)
Φnitrene
knitrened (ꢁ 109 sꢀ1
)
a
e
c
solvent
ΦNCO
CH2Cl2
CHCl3
CCl4
0.67 ( 0.11
0.62 ( 0.09
0.52 ( 0.09
9.6 ( 2.4
5.2 ( 0.9
0.020 ( 0.005
0.04 ( 0.01
0.29 ( 0.06
0.29 ( 0.10
0.31 ( 0.09
0.51 ( 0.15
0.31 ( 0.06
0.34 ( 0.05
0.19 ( 0.03
4.4 ( 1.2
2.8 ( 0.5
0.75 ( 0.20
0.28 ( 0.07
a ΦNCO = |ANCO/AN | ꢁ (εN3/εNCO). b kNCO = ΦNCO ꢁ τꢀ1, τ the lifetime of the S1 state taken from Table 1. c Φnitrene = 1 ꢀ (ΦNCO þ ΦT).
3
d knitrene = Φnitrene ꢁ τꢀ1. e Taken from Table S1. f kISC = ΦT ꢁ τꢀ1
.
presents the time constants measured using fs UVꢀvis transient
absorption spectroscopy. In the latter case, the decay of the S1
absorption band of 2-NpCON3 was analyzed (vide infra). In
addition, Table 1 includes the time constants obtained when
270 nm excitation was used, and thus higher excited states were
initially populated (vide infra).
A small shift of the maximum of the IR band of the 2-NpCON3 S1
state to lower frequency was observed during the first 5 ps in
chloroform. This effect was attributed to the dynamics of solvation25
(Figure S6, Supporting Information). This interpretation is in
agreement with the computational prediction that the dipole mo-
ment of 2-NpCON3 increases significantly upon excitation, from 3.8
D in the S0 state to 11.3 D in the S1 (π,π*) state.
To better understand the influence of solvent, the efficiency of
isocyanate formation was estimated in solvents with different
polarities (CCl4, CHCl3, CH2Cl2). These solvents are relatively
transparent to IR light (1 mm layer of solution) in the spectral
range where an isocyanate has a strong IR marker. In these
experiments, the S1 state was directly populated using 340 nm
excitation. The experimental IR difference spectra were recorded,
and the amplitudes of the isocyanate (ANCO) and azide (AN3
)
bands were compared in different solvents (Figure S7, Support-
ing Information). The extinction coefficients for the NCO
vibration (εNCO) in these solvents were determined with an
authentic sample of 2-NpNCO and the extinction coefficients for
the N3 vibration (εN ) for 2-NpCON3 were determined as well
3
(Table S3, Supporting Information). Using values of εNCO and
εN we estimated the efficiency (ΦNCO) and the rate constant
(kN3 CO = ΦNCO ꢁ τꢀ1) of isocyanate formation in different
solvents (Table 2). In the most polar solvent employed, CH2Cl2,
the rate constant and efficiency of isocyanate formation are the
largest among the three solvents used. The ratios of the rate
constants of isocyanate formation were estimated to be about
1:7:13 in CCl4, CHCl3, and CH2Cl2, respectively.
As the S1 state is very polar, one might reasonably expect, apply-
ing ground-state reasoning, that it would be kinetically stabilized in
polar solvents, yet the trend in lifetime is in the opposite direction.
Of course, the solvent effect is difficult to interpret because the decay
of the S1 state is controlled by several solvent-dependent processes
including isocyanate formation, nitrene formation, and triplet for-
mation and the rate of each process will respond differently to
solvent variation.
Having estimated the efficiency of isocyanate formation, we
attempted next to estimate the efficiency of triplet formation.
The tripletꢀtriplet UVꢀvis absorption spectrum of 2-NpCON3
in cyclohexane has been observed previously in conventional
laser flash photolysis experiments.8 The lifetime of the triplet
azide is on the order of microseconds8 in a deoxygenated solvent.
Thus, in principle, the 2-NpCON3 triplet state should be
detectable by means of fs IR techniques as the residual spectrum
The amplitude of the N3 bleach of 2-NpCON3 at about
2140 cmꢀ1 does not change over 3 ns (Figures 1 and S2). Thus,
in all of the solvents utilized in this study, there is no recovery of
the ground state on this time scale. This means that both internal
conversion (IC, S1fS0) and fluorescence have small quantum
yields. In our opinion, the large energy gap between the S1 and S0
states (∼85 kcal/mol, see Table S7) plays a key role in lowering
the rate constant and thus, the efficiency of IC.
9753
dx.doi.org/10.1021/ja109098w |J. Am. Chem. Soc. 2011, 133, 9751–9761

Journal of the American Chemical Society
ARTICLE
recorded 3 ns after the laser excitation. According to calculations
at the UB3LYP/TZVP level of theory, the N3 stretch of triplet
2-NpCON3 is located between the vibrations of the ground and
S1 states (Figure S5, Table S2). On first inspection of the
transient spectrum, there is no evidence of a positive band that
can be assigned to the triplet azide. However, in two solvents
(cyclohexane and carbon tetrachloride), the shape of the tran-
sient spectrum recorded at 3 ns does not match the normalized
FTIR spectrum (Figure S2a,e), although the concentration of
the S1 state is very low at that time. This difference may be due
to triplet formation in these two solvents. In acetonitrile and
chloroform, the IR transient spectrum recorded after completion
of the S1 decay has the same shape as the normalized FTIR
spectrum (Figures 1 and S2c). If this interpretation is correct,
then in acetonitrile and chloroform, the triplet azide must be
formed with much lower efficiency than in cyclohexane and
carbon tetrachloride.
To estimate the quantum yield of triplet formation for 2-NpCON3
in different solvents (ΦT, Table S1, Supporting Information,
Table 2), the relative actinometrymethod was applied using transient
UVꢀvis absorption spectroscopy (see Supporting Information for
details), and benzophenone was used as the reference compound. In
all solvents employed, the transient UVꢀvis spectra (with a max-
imum at ∼450 nm) were obtained 3 ns after 345 nm laser excitation
of 2-NpCON3 (Figure S1, Supporting Information). These spectra
are similar to the spectrum assigned previously8 to the tripletꢀtriplet
absorption of 2-NpCON3. The intensity of these spectra was strongly
dependent on the solvent polarity and drops significantly upon
changing from the nonpolar cyclohexane and CCl4 solvents to
more polar solvents. It was assumed that the extinction coeffi-
cient at the maximum of the 2-NpCON3 tripletꢀtriplet absorp-
tion spectrum in all solvents is the same and is equal to the value
of 10500 L molꢀ1 cmꢀ1 reported for 2-NpCOCH3.26 The
Figure 2. Isocyanate formation from 2-NpCON3 (c ∼ 4.0 ꢁ 10ꢀ3 M)
after 350 nm excitation in chloroform. (a) TRIR spectra produced by
ultrafast photolysis of 2-NpCON3 in chloroform with 350 nm excitation.
The normalized steady-state IR absorption spectrum (FTIR) of
2-NpNCO is also shown as a dotted line. (b) The kinetics of isocyanate
formation recorded at 2269 cmꢀ1
.
solvent. The rate constant of the intersystem crossing is almost
independent of the solvent polarity. The efficiency of nitrene
formation depends slightly on polarity. The efficiency of the
isocyanate formation increases significantly as polarity increases,
and the opposite trend is observed for the efficiency of triplet
formation.
The formation of the isocyanate IR band was studied following
350 nm excitation of 2-NpCON3. The transient spectra recorded
in chloroform are shown in Figure 2a. There is a correlation
between the rate of decay of the S1 band and formation of the
isocyanate. Both kinetic processes can be adequately described as
monoexponential functions. The time constant of the latter
process was found to be 130 ( 14 ps (Figure 2b). A global fit
of the decay kinetics for the S1 band and formation kinetics for
the isocyanate gave a time constant of 141 ( 6 ps. These results
clearly show that the S1 state of the acyl azide is the precursor of
the isocyanate.
A vibrational band of singlet benzoylnitrene has been observed as
a doublet at 1727 and 1765 cmꢀ1 by matrix isolation spectroscopy
(Figure S9, Supporting Information) and in solution at 1760 cmꢀ1
by ns IR spectroscopy5 and by fs IR spectroscopy.19 On the basis of
Autrey and Schuster’s8 product studies, it is clear that nitrene is
formed when the S1 state of 2-NpCON3 is populated by direct
excitation with 350 nm light.
Thus, fs TRIR experiments were performed with 340 nm
excitation in order to observe nitrene formation in CCl4. The
characteristic (cyclic CNO) nitrene band isveryweak, but the long
decay time constant of the putative precursor S1 state allows its
formation to be monitored. The S1 lifetime of 2-NpCON3 in CCl4
is about 690 ( 60 ps (Table 1). The transient spectra obtained in
CCl4 are shown in Figures 3 and S10, Supporting Information.
The intensity of the nitrene band at 1740 cmꢀ1 rises with a time
3
3
quantum yield of triplet 2-NpCON3 formation is thus estimated
to be very low in polar solvents (about 1% in acetonitrile) and
relatively high in nonpolar solvents (about 25%, Table S1).
Therefore, the concentration of triplet 2-NpCON3 is too low
to be detected by fs TRIR spectroscopy in polar solvents.
The absolute rate constants of intersystem crossing (kISC
=
ΦT ꢁ τꢀ1) in different solvents can thereby be roughly estimated
using values of the triplet quantum yield (ΦT, Table S1) and the
time constants of the S1 state decay (τ, Table 1). These
estimations demonstrate that, within experimental accuracy
(∼20ꢀ30%), the ISC rate constant is independent of the nature
of solvent (kISC = (3.3 ( 0.7) ꢁ 108, (5.1 ( 1.5) ꢁ 108, (3.1 (
0.9) ꢁ 108, (2.9 ( 1.0) ꢁ 108, and (6 ( 2) ꢁ 108 sꢀ1 in
cyclohexane, CCl4, CHCl3, CH2Cl2, and CH3CN, respectively).
As the nitrene is not a stable (minutes, hours) photoproduct in
solution at room temperature, it was impossible to estimate the
efficiency of nitrene formation in different solvents by measuring
experimental IR difference spectra. However, the efficiency of nitrene
formation (Φnitrene) was calculated after assuming that (Φnitrene
þ
ΦNCO þ ΦT) = 1. The rate constants of the nitrene formation could
then be estimated (knitrene = Φnitrene ꢁ τꢀ1, Table 2).
Interestingly, the dependence of the logarithm of the S1 life-
time on a traditional function of polarity, f(ε,n) = (ε ꢀ 1)/
(2ε þ 1) ꢀ (n2 ꢀ 1)/(2n2 þ 1),24 is linear (Figure S8, Supporting
Information). Figure S8 also presents the dependence of the
efficiency of isocyanate, nitrene, and triplet formation and the rate
constants of these processes versus a function of polarity (data
taken from Table 2). Figure S8 clearly shows that the rate
constants of isocyanate and nitrene formation increase in polar
9754
dx.doi.org/10.1021/ja109098w |J. Am. Chem. Soc. 2011, 133, 9751–9761

Journal of the American Chemical Society
ARTICLE
Figure 4. (a) Transient UVꢀvis spectra produced by fs laser excitation
of 2-NpCON3 (c ∼ 4.0 ꢁ 10ꢀ3 M) in acetonitrile (λex = 350 nm). The
vertical bars indicate the positions and oscillator strengths (f) of the
electronic transitions calculated for 2-NpCON3 in the excited singlet
(π,π*) state at the MS-CASPT2//CASSCF(14,13) level. (b) Mono-
exponential fit to the transient absorption kinetics at 500 nm.
Figure 3. Nitrene formation upon excitation of 2-NpCON3 in carbon
tetrachloride (c ∼ 4.0 ꢁ 10ꢀ3 M, λexc = 340 nm). (a) TRIR spectra
produced upon ultrafast photolysis of 2-NpCON3 in CCl4. (b) The
kinetics of nitrene formation recorded at 1737 cmꢀ1
.
constant of 790 ( 140 ps. This is in fairly good agreement with the
time constant of the S1 state decay in this solvent (Table 1).
Unfortunately, the accuracy of the time constant measurement is
low, as the intensity of the nitrene band is very weak (<10ꢀ4 OD).
Nevertheless, this experiment clearly demonstrates that the acyl
azide S1 state is the precursor of nitrene, consistent with the
product studies.8
A similar pattern of results regarding nitrene formation was
observed in CH2Cl2. In this solvent, the S1 state lifetime is equal to
70 ( 6 ps (Table 1). The rise time of the nitrene band in this
solvent was found to be 88 ( 14 ps (Figure S11, Supporting
Information). These results are significant because in our previous
report19 we were unable to obtain direct kinetic evidence that the
S1 state of an acyl azide is the precursor of the corresponding
nitrene. However, in our previous studies, the higher electronic
states of 2-NpCON3 were populated using 270 nm excitation
(vide infra),19 and the yield of S1 formation from the initially
populated Sn (ng2) states was low.
Transient UVꢀVis Detection. Transient UVꢀvis absorption
spectra were also recorded upon 350 nm excitation of 2-NpCON3
in carbon tetrachloride and in acetonitrile using the pumpꢀprobe
absorption technique. Excitation of 2-NpCON3 in CCl4 produces
a UVꢀvis transient absorption whose decay is well described by a
monoexponential function with a time constant of 425 ( 12 ps
(Figure S12, Supporting Information). This time constant is in
fair agreement with that measured using the fs TRIR technique
(690 ( 60 ps, Table 1). The transient spectrum recorded at ∼3 ns
is very similar to the tripletꢀtriplet absorption spectrum of Autrey
and Schuster as recorded by ns LFP (Figure S13, Supporting
Information).8
constant is in good agreement with that of the S1 state of 2-NpCON3
obtained by the TRIR technique (28 (5 ps, Table 1). In this solvent,
the tripletꢀtriplet band was also observed (Figure S14, Supporting
Information), but its intensity was very low, which is consistent with
the low efficiency of triplet formation in polar solvents (Table 2).
The UVꢀvis spectrum of 2-NpCON3 in the first singlet excited
(π,π*) state has been calculated at the MS-CASPT2//CASSCF
level (Figure 4a, Table S4). Calculations predict that this species
exhibits strong absorption bands in the visible region. As seen in
Figure 4a, the calculated electronic absorption spectrum agrees well
with experiment. Note that the experimental transient spectrum
may be contaminated in the range of 375ꢀ425 nm, due to
stimulated emission (Figure S15, Supporting Information). There-
fore, calculations support assignment of the transient absorption
spectra recorded on the ps time scale (Figures 4 and S12) mainly to
2-NpCON3 in the S1 (π,π*) state.
The results of our ultrafast time-resolved studies with 350 nm
excitation light demonstrate that the S1 state of 2-NpCON3
is the precursor of isocyanate and nitrene in all of the solvents
employed (Scheme 1). In nonpolar solvents, the triplet state of
2-NpCON3 is also formed in high yield upon direct excitation of
the S1 state.
These findings are consistent with prior conclusions based on
ns LFP, product studies, and triplet-sensitized experiments, but
this is the first direct experimental observation that the decay of
the S1 state correlates with the formation of the nitrene. Our
results clearly show that internal conversion (S1fS0) and fluo-
rescence are minor processes, as the recovery of the ground state
of 2-NpCON3 was not observed up to 3 ns postexcitation. The
lifetime of S1 state of 2-NpCON3 is influenced by solvent because
the rate constants of the nitrene and isocyanate formation grow
significantly as the solvent polarity increases.
As in the case of CCl4, excitation of 2-NpCON3 in acetoni-
trile yields a UVꢀvis transient absorption spectrum which decays
exponentially with a time constant of 26 ( 4 ps (Figure 4). This time
9755
dx.doi.org/10.1021/ja109098w |J. Am. Chem. Soc. 2011, 133, 9751–9761

Journal of the American Chemical Society
ARTICLE
Scheme 1
2.2. Ultrafast IR Studies of the 270 nm Photochemistry of
2-NpCON3. In our communication,19 the 270 nm photochemistry
of 2-NpCON3 in chloroform was studied. Femtosecond time-
resolvedresults obtained with350 nmexcitation for2-NpCON3 in
chloroform can be profitably compared with the data obtained
using 270 nm excitation.19 Similar features were observed in
experiments with both excitation wavelengths: two positive bands
at 2100 cmꢀ1 (S1 state of azide) and at 2265 cmꢀ1 (isocyanate),
and one negative band at 2140 cmꢀ1 (bleaching of azide in the
ground state). As the 350 nm light pumps the S1 state directly, the
2100 cmꢀ1 band, recorded just after excitation, is much more
narrow than that produced with 270 nm light. Clearly, a “hot” S1
state with an initially much broader band is produced with short
wavelength light. The time constant of decay of the hot S1 state in
chloroform is 6.8 ps (Table 1). Thus, one predicts that the TRIR
spectra recorded with 270 nm light, at time delays longer than
∼30 ps, should have the same width as the spectra recorded with
350 nm light. Indeed, the transient spectrum recorded at 33 ps
using 270 nm excitation has the same shape as the transient
spectrum observed at 11 ps with 350 nm excitation (Figure S16,
Supporting Information).
Figure 5. Transient IR spectra produced upon excitation (λexc
=
270 nm) of 2-NpCON3 (c ∼ 1.4 ꢁ 10ꢀ3 M) in acetonitrile and in
carbon tetrachloride. The normalized steady state IR absorption spectra
(FTIR) of 2-NpCON3 are also shown as dotted lines.
vibrational cooling (VC) in the time window of tens of picose-
conds. In nonpolar solvents, this shift is apparently larger.
However, this is due to the fact that the low frequency part of
this band decays much faster relative to the band of the thermally
relaxed S1 state.
As in the case of 350 nm excitation, there is no recovery of the
N3 IR stretch (∼2140 cmꢀ1) in the ground state of 2-NpCON3
over 3 ns (Figure S19, Supporting Information). There is only a
slight increase of the negative signal at 2140 cmꢀ1 due to the
overlap of this negative band with the wide, positive 2100 cmꢀ1
band (Figures 5, S17, S18, and S19). Recovery of the CdO stretch
(∼1690 cmꢀ1) in the ground state of 2-NpCON3 was not ob-
served either. Recall that with 350 nm excitation, the weak triplet
IR band was detected in CCl4 and cyclohexane (Figure S2a,e).
This was also the case when 270 nm excitation was utilized. In
these two solvents, the shape of the transient spectrum recorded at
3 ns is clearly different than the appropriately normalized FTIR
spectrum (Figures 5 and S17).
To better understand the photochemistry of 2-NpCON3 in
higher excited states, experiments in some other solvents were
performed. Figures 5 and S17, Supporting Information, present
the IR transient absorption spectra produced upon excitation of
2-NpCON3 at 270 nm in selected solvents. With this excitation
wavelength, higher electronic states of 2-NpCON3 are immedi-
ately populated (vide infra, Table S7). In all solvents employed,
270 nm photoexcitation of 2-NpCON3 yields the 2100 cmꢀ1
band, which decays biexponentially. The time constants of this
biexponential decay are presented in Table 1.
In CCl4, the decay of the 2100 cmꢀ1 band is slower than that in
chloroform,19 and this allowed us to conveniently monitor the
formation of isocyanate band at 2270 cmꢀ1 (Figure S20, Support-
ing Information). The rate of formation of isocyanate is the same
as the rate of decay of the 2-NpCON3 in the S1 state (Figure S21,
Supporting Information, Table S5), and approximately equal
amounts of 2-naphthyl isocyanate are formed in the fast and slow
processes.
As in our preliminary report,19 the fast component (Table 1) was
assigned to the hot S1 state of 2-NpCON3 undergoing vibrational
relaxation27,28 and Curtius rearrangement. The slow component was
assigned to the decay of the thermally relaxed S1 state of 2-NpCON3.
Indeed, the time constants of the slow components obtained with
270 nm excitation are consistent with the time constants obtained
when the S1 state was excited directly with 350 nm light.
A similar ratio of the amplitudes of the fast and slow processes
was observed in chloroform.19 These results confirm that both
#
vibrationally hot (S1 ) and cool (S1), first singlet excited states
of 2-NpCON3 are precursors of isocyanate. Note, that the iso-
cyanate band recorded immediately after the 270 nm laser
excitation in chloroform is shifted to lower frequency relative
to the band recorded at longer delays (>30 ps). This is not the
case after 350 nm excitation (Figure 2). Thus, formation of the
hot isocyanate and its VC were not as prominent with 350 nm
irradiation as with 270 nm excitation (Figure S22, Supporting
Information).
Although the IR experiments suffer from artifacts29 in the first
picosecond time window, it is very likely that the formation of the
2100 cmꢀ1 band is instantaneous (within the laser pulse) in all
solvents employed (Figure S18, Supporting Information). In-
stantaneous growth was also recently observed for benzoyl and
pivalolyl azides in chloroform.19 In all solvents, the maximum of
the 2100 cmꢀ1 band of 2-NpCON3 shifts to higher frequency by
about 5 cmꢀ1 (Figure 5, Figure S17). This result is typical of
The fs TRIR results clearly show that the S1 state is a precursor of
nitrene when the S1 state is populated directly (350 nm). However,
9756
dx.doi.org/10.1021/ja109098w |J. Am. Chem. Soc. 2011, 133, 9751–9761

Journal of the American Chemical Society
ARTICLE
problem of resolving the minor component of nitrene formation
is made more difficult by the intrinsic low amplitude of the
nitrene IR band (Figures S10, S11, S23, and S24). As will be
shown later, theory supports this explanation (vide infra).
The experiments with 270 and 350 nm excitation were per-
formed using different experimental conditions. To correlate the
relative intensities of the S1 band with the excitation wavelength
(λexc), normalization of the transient spectra was performed. The
spectra were normalized to the same intensity of the 2140 cmꢀ1
bleach of precursor to compare the intensities of the S1 bands
produced with different λexc. Figure S26, Supporting Information,
demonstrates that the amplitude of the S1 band recorded with
350 nm light in chloroform is much larger than that recorded with
270 nm light. Note that only ∼33% of the signal detected with
270 nm light could be assigned to the thermally relaxed S1 state. As
a result, the 2100 cmꢀ1 signal detected with 350 nm excitation is
about 6 times larger than that for thermally relaxed S1 state with
270 nm excitation. Furthermore, if we compare the overall
intensity of the 2100 cmꢀ1 band in the two experiments,
350 nm excitation leads to about twice the signal strength than
production of the corresponding band with 270 nm excitation
(Figure S26). This means that only about half of the 2-NpCON3
molecules in the Sn state (n g 2) produce the S1 state through IC.
Similar results were obtained in CCl4 and acetonitrile (Figures
S27 and S28, Supporting Information). These results are well
explained by the formation of singlet nitrene (2-NpCON) in the
upper excited singlet states (Sn, n g 2, Scheme 2).
2.3. Ultrafast UVꢀVis Studies of the 270 nm Photochemistry
of 2-NpCON3. As 2-NpCON3 has a convenient UVꢀvis chromo-
phore, fs time-resolved UVꢀvis transient absorption experiments
were also performed in selected solvents: chloroform, acetonitrile,
and methanol. In all of the solvents studied, transient spectra were
recorded between 350 ꢀ 650 nm when 270 nm excitation was
utilized. Figure 7 presents the transient absorption spectra detected
in chloroform over a ꢀ2 to 550 ps time window. A very wide
transient absorption spectrum with a maximum at about 470 nm is
formed within a laser pulse. This transient absorption decays on a ps
time scale. Analysis (Figures S29 and S30, Supporting Information)
demonstrates that this decay is best described by a biexponential
function with time constants equal to 6 ( 2 and 130 ( 20 ps. These
time constants are in very good agreement with the time constants
of the azide singlet excited state (S1) decay monitored by fs TRIR
techniques (6.8 ( 1.3 and 120 ( 30 ps, Table 1).
Figure 6. Integrated signal of nitrene band formed upon photolysis of
2-NpCON3 (c ∼ 1.4 ꢁ 10ꢀ3 M) in CCl4 (λexc = 270 nm).
we were unable to obtain direct evidence for this conclusion using fs
TRIR spectroscopy with 270 nm excitation.19 As the intensity of
the nitrene IR band is very low (see Figures 1, S1, and S2 from ref
19 and Figures 3, S10, and S11 in this work) and as VC of the
nascent nitrene is relatively slow (30 ps time constant), we could
not resolve the dynamics of nitrene formation and could not assign
the precursor of singlet 2-NpCON in chloroform19 when higher
electronic states are excited initially with 270 nm light.
Thus, we studied nitrene formation with 2-NpCON3 in CCl4
(λexc = 270 nm) where the decay of the S1 state is the slowest among
the solvents employed and is in fact much slower than that of a typical
VC process. The broad band at 1710ꢀ1780 cmꢀ1 assigned to the
singlet nitrene again is formed within the laser pulse. The band
sharpens and shifts to higher frequency in about 100 ps (Figure S23,
Supporting Information). Note, that there is a difference between the
shape of nitrene bands obtained for 2-NpCON3 in CCl4 at 10 ps with
two different excitation wavelengths (270 and 340 nm, Figure S24,
Supporting Information). With 270 nm excitation, the transient
spectrum is more intense and shifted to the lower frequency and thus
can be assigned to a hot nitrene. On the other hand, the spectra
recorded at 2 ns with both excitations have similar shapes.
The intensity of the nitrene band integrated over the spectral
range 1670ꢀ1780 cmꢀ1 can be fitted to a monoexponential
decay curve with a time constant of 7.0 ( 2.0 ps (Figure 6).
There is no evidence of a rising component with a time constant
close to the lifetime of the thermally equilibrated S1 state of
2-NpCON3 (570 ps in this solvent). Therefore, there is no
evidence available from time-resolved IR experiments that some
nitrene is formed from the relaxed S1 state of 2-NpCON3, when
the precursor is excited at 270 nm in contrast to experiments
performed with 340 nm excitation.
The transient spectrum detected 550 ps after the laser pulse
has very low intensity (ΔA ∼ 0.001). Nevertheless, it is clear that
its shape in the range of 400ꢀ520 nm is close to that of the
tripletꢀtriplet transient absorption spectrum recorded by Autrey
and Schuster8 in cyclohexane at room temperature (Figure S31,
Supporting Information).
Unsurprisingly, the same pattern of results was observed in
acetonitrile, a solvent in which the decay of the S1 state is even
faster than in chloroform. Yet again, there is no experimental
evidence that the formation of the nitrene band correlates
(Figure S25, Supporting Information) with the decay of S1
produced indirectly with 270 nm excitation.
To rationalize this conflict, we postulate that with 270 nm
excitation, the efficiency of nitrene formation from the upper
electronic states (Sn, n g 2) is large, and that the relative yield of
nitrene formation from the vibrationally cooled S1 state is small.
Thus, formation of the nitrene produced from the relaxed S1 state
cannot be resolved from the major pathway because of a low yield
superimposed on a larger yielding and faster process. The
Note, for singlet benzoylnitrene (PhCON), the transient ab-
sorptionspectrum has a UVbandnear 300 nm5,9 in agreement with
TD-B3LYP calculations.6 A similar spectrum, slightly shifted to the
red (∼320 nm), was predicted previously6 for singlet 2-naphtho-
ylnitrene (2-NpCON). Therefore, the singlet nitrene does not
contribute noticeably to the experimental transient absorption
spectra detected in the range of 380ꢀ650 nm (Figure 7).
To determine if excited isocyanate has a contribution to the
transient spectra recorded upon excitation of 2-NpCON3, we
performed a fs time-resolved UVꢀvis control study of the isocya-
nate (2-NpNCO). After excitation of 2-NpNCO with 270 nm light,
a wide band was observed, which decays almost to baseline with a
time constant of 2.4 ps (Figure S32, Supporting Information).
9757
dx.doi.org/10.1021/ja109098w |J. Am. Chem. Soc. 2011, 133, 9751–9761

Journal of the American Chemical Society
ARTICLE
Scheme 2
equilibrated S1 state with a time constant in the range of 2 to 14
ps depending on the solvent (Table 1). Excitation of 2-NpCON3
at 340 and 350 nm results in the population of the S1 state with
only a small excess of vibrational energy, which manifests itself by
the absence of the fast decay component. Moreover, in all
solvents of this work, the transient absorption UVꢀvis spectra
detected at 3 ns in the range 400ꢀ520 nm are in fair agreement
with the absorption spectrum detected previously in cyclohexane
using conventional LFP techniques and assigned to the tripletꢀ
triplet absorption of 2-NpCON3.8 Note that isocyanate and
singlet nitrene have no noticeable absorption in the
350ꢀ650 nm spectral region because these species do not absorb
light in this spectral range.
Figure 7. Transient UVꢀvis absorption spectra obtained upon excita-
tion of 2-NpCON3 (c ∼ 1.4 ꢁ 10ꢀ3 M) in chloroform at ambient
temperature recorded at different time delays (λexc = 270 nm).
3. COMPUTATIONAL RESULTS
Recently, wereported anexperimentaland computationalstudy
of the primary processes in the photochemistry of aryl azides.30ꢀ33
The lifetimes of the singlet excited states of 2-naphthyl azide and
1-chloro-2-naphthyl azide were found to be shorter than 1 ps.30 A
very short lifetime (less than 1 ps) was also observed for excited
states of phenyl azide,32 p-biphenyl, o-biphenyl, and 1-naphthyl
azides.31,33 According to reported calculations, the detected life-
times of the excited aryl azides wereassigned to the S2 states, rather
than the S1 states,30ꢀ33 as the S1 states of the studied aryl azides are
dissociative in nature and thus were not detected.
To better understand and support the experimental results
obtained in this study, we similarly performed density functional
theory calculations on some singlet excited states of 2-NpCON3.
In the past, we have shown that the excited state frequencies
calculated at the TD-B3LYP/TZVP levels of theory are in very
good agreement with the fs TRIR experiments.34
Initially, we computed the vertical excitations (Tables S7) for
2-NpCON3 using the optimized ground-state geometry (Tables
S8), and then we rendered the corresponding difference density
plots in order to identify the character of the singlet excited
states (Figures S38, Supporting Information). In order to
compute the vibrational spectra of the singlet and triplet excited
states, we further optimized the geometries of the excited states
(Tables S9, and S10, Supporting Information) using either TD-
B3LYP (singlet states) or UB3LYP (triplet states).
A dissociative singlet excited state of the azide unit produces
the corresponding singlet nitrene and molecular nitrogen. As was
shown for the S1 state of aryl azides, this excitation corresponds
to promotion of an electron from the π-orbital of the aryl moiety
to the in-plane π* orbital on the azide group. However, as shown
Thus, neither the nitrene nor the excited isocyanate contribute
to the transient absorption spectra recorded upon excitation of
2-NpCON3. The ground-state isocyanate steady-state absorption
spectrum lies below 340 nm (Figure S33, Supporting Information).
A similar pattern of results was observed with 2-NpCON3 in
acetonitrile. A very broad transient absorption spectrum is formed
within the laser pulse (∼300 fs), and then it decays biexponentially
(Figures S34, S35, and S36, Supporting Information). Kinetic fits
at individual wavelengths (Figure S35) and global analysis (Figure
S36) indicate that this decay is best described by a biexponential
dependence with time constants equal to 3 ( 1 and 23 ( 4 ps.
These time constants are in good agreement with time constants
obtained in this solvent by the fs TRIR technique (Table 1). The
spectrum detected on a time scale of hundreds of picoseconds has
a shape similar (Figure S37, Supporting Information) to the shape
of the tripletꢀtriplet absorption of 2-NpCON3 recorded pre-
viously in cyclohexane.8
Note, in both solvents, the decay-associated spectra (DAS,
Figures S30 and S36) corresponding to the fast decay (3 ps
in acetonitrile and 6 ps in chloroform) are much wider than the
DAS associated with the slow decay (23 and 130 ps, respectively).
This confirms that the fast component detected by fs UVꢀvis
spectroscopy is influenced by the band-narrowing effect that is
typical of VC.
Thus, in all solvents under study, the same species were
detected in both ultrafast IR and UVꢀvis pumpꢀprobe absorp-
tion measurements. Excitation of 2-NpCON3 at 270 nm yields
#
a hot first singlet excited state (S1 ) which decays to a thermally
9758
dx.doi.org/10.1021/ja109098w |J. Am. Chem. Soc. 2011, 133, 9751–9761

Journal of the American Chemical Society
ARTICLE
transient IR band at 2100 cmꢀ1. The carrier of the 2100 cmꢀ1
transient was assigned to the S1 state of 2-NpCON3, and this
assignment was supported by calculations. Calculations predict
that the S1 state is (π,π*) in nature with the excitation localized
mainly on the carbonyl group. However, the frequency of the N3
stretch in the S1 state is shifted to lower energy in agreement with
experiments. This explains why the S1 states of carbonyl azides
are longer lived and more readily observed than the correspond-
ing S1 states of aryl azide where the excitation is localized on the
azide moiety.
As isocyanates have very strong, characteristic IR bands at
about 2260 cmꢀ1, their formation was monitored using the fs IR
transient absorption spectroscopy. It was demonstrated that
formation of isocyanate (photo-Curtius rearrangement) corre-
lates with the decay of the S1 singlet excited state of 2-NpCON3
(Scheme 1). Moreover, the photo-Curtius rearrangement pro-
ceeds in both the hot and vibrationally cooled S1 states upon
270 nm excitation (Scheme 2). The hot S1 state produces
isocyanate with a rate constant larger than that of the thermally
relaxed S1 state. When the S1 state was populated directly using
350 nm excitation, only monoexponential decay of the thermally
relaxed S1 state was observed and the concentration of isocyanate
rises monoexponentially as well (Scheme 1). Thus, our experi-
mental results unambiguously demonstrate that isocyanate is
produced in the lowest singlet excited state of 2-NpCON3.
However, more advanced quantum chemical calculations are
required to understand in detail the mechanism of photo-Curtius
rearrangement.
Figure 8. Geometries of the ground (left) and lowest singlet excited
(right) states of 2-NpCON3. The ground and S1 states were optimized at
the B3LYP/TZVP and TD-B3LYP/TZVP levels of theory, respectively
(bond lengths are in angstroms).
in Figure S38, the S1 state of 2-NpCON3 is a delocalized (π,π*)
excited state, and it does not have significant dissociative charac-
ter. Nevertheless, it has some electron density accumulation in
the π* orbital of the distal NꢀN bond on the departing N2
fragment. This aspect would be a significant requirement for a
(prompt) dissociative state for nitrene formation from excited
carbonyl azides. Indeed, it appears that the “dissociative” state for
2-NpCON3, the S4 state, is located higher in the singlet manifold.
Femtosecond TRIR experiments demonstrate that upon 270 nm
excitation, the nitrene is partly produced by the higher excited
singlet states (Sn, n g 2, Scheme 2). This is in agreement with our
computational results, which predict that the S4 singlet excited state
is “dissociative”.
It was unambiguously demonstrated that nitrene formation
correlates with the S1 state decay of 2-NpCON3 when the latter
is populated by direct excitation with 350 nm light (Scheme 1). This
is in accord with the results of product analysis. The singlet nitrene
band at 1760 cmꢀ1 was also observed upon 270 nm excitation;
however, in this case, most of the nitrene is formed instantaneously.
Thus, higher singlet excited states of 2-NpCON3 also produce
singlet nitrenes (Scheme 2). As the IR markers of the nitrenes near
1760 cmꢀ1 are very weak, conclusions based on the IR data cannot
be made with certainty. Because 270 nm excitation populates highly
excited singlet states (Sn, n > 1), it is possible that nitrene is
produced from three different azide excited states (Scheme 2): Sn
(n > 1, very fast process), hot S1 (fast process), and relaxed S1 (slow
process). According to the calculations, the typical azide dissociative
state, similar to those of aryl azides, lies above the S1 state for
2-NpCON3. This supports our conclusion that the singlet nitrenes
are also generated from the highly excited singlet states.
Thus, according to theory (see Table S7), the 270 nm laser pulse
excites 2-NpCON3 to a “dissociative”excited state where two distinct,
concurrent channels were considered: (1) production of the singlet
nitrene, and (2) internal conversion to the S1 state (Scheme 2). This
explains why a 270 nm laser pulse simultaneously produces the
vibrational band of the nitrene at 1720 cmꢀ1 and the 2100 cmꢀ1 band
belonging to the S1 excited state of 2-NpCON3. Formation of
nitrenes takes place also in the S1 state, but more slowly and in very
low yield, which is evident from the experiments with 350 nm
excitation.
In order to support the assignment of the 2100 cmꢀ1 experi-
mental band to the S1 state, we optimized the geometry of the
lowest singlet excited state of 2-NpCON3. The optimized geome-
tries of the S0 and S1 states of 2-NpCON3 are shown in Figure 8.
It is clear from the difference density plots (Figures S38),
2-NpCON3 have an S1 state that is (π,π*) in nature, and the
electronic reorganization is delocalized over the entire molecule.
Thus the S1 state geometries are characterized by changes in the
aromatic CdC bond lengths and by only a moderate elongation in
the CdO bond length for 2-NpCON3.
We also found that the bleaching of the IR bands of the ground-
state 2-NpCON3 does not recover on the time scale of our
experiment (∼3 ns). We conclude that internal conversion and
fluorescence leading to the ground state of 2-NpCON3 is negli-
gible compared to other deactivation processes in the S1 state (ISC
and isocyanate and nitrene formation).
The computed frequency of the azide stretch is ∼2200 cmꢀ1 for
the S1 excited state, which is approximately 100 cmꢀ1 higher than
the experimental observations. If one applies a scaling factor similar
to that used with ground state vibrations, the calculated 2210 cmꢀ1
band of the S1 state of 2-NpCON3 will move to about 2090 cmꢀ1 in
perfect agreement with the experimentally observed value. Conse-
quently, our calculations support the assignment of the 2100 cmꢀ1
band to the S1 state.
The graphical scheme of Figure 9 summarizes results of our
study and shows different deactivation channels detected for the
lowest and upper singlet excited states of 2-NpCON3 using the
ultrafast UVꢀvis and IR spectroscopies.
In contrast to aryl azides, the lifetime of the thermally relaxed S1
excited state of 2-NpCON3 is relatively long because excitation is
localized on the aromatic ring and not directly on the azide group.
Moreover, this lifetime depends on the solvent polarity, being shorter
in more polar solvents. It was clearly demonstrated that both reactions
(isocyanate and nitrene formation) are accelerated by the polar
solvent while, as expected, the intersystem crossing rate constant is
4. CONCLUSIONS
The photochemistry of 2-NpCON3 was studied using ultrafast
experimental techniques and computational methods. Excita-
tion of 2-NpCON3 in solvents of different polarity produced a
9759
dx.doi.org/10.1021/ja109098w |J. Am. Chem. Soc. 2011, 133, 9751–9761

Journal of the American Chemical Society
ARTICLE
computed difference electronic density plots (between S0 and S1ꢀS4
states) as described in previous reports.44ꢀ46 The first singlet excited
state of 2-NpCON3 was optimized at the TD-B3LYP/TZVP level of
theory. The stationary point obtained for the singlet excited state was
confirmed to be a minimum by calculating the second derivatives
numerically, utilizing the NumForce module in Turbomole, thereby
also predicting the infrared (IR) spectrum of the excited state at the TD-
B3LYP/TZVP level of theory.
Vertical excitation energies of 2-NpCON3 in the first excited singlet
state were calculated at the TD-B3LYP/TZVP geometry (Table S9) by
the MS-CASPT2//CASSCF procedure47 with the ANO-S basis set of
Pierloot et al.48 using the MOLCAS program49 under limiting condi-
tions of Cs symmetry. In order to arrive at a satisfactory description of all
excited states at the CASPT2 level (i.e., to remove intruder states), the
imaginary level-shifting technique was used.50 The active space used in
these calculations included 14 electrons with 13 orbitals (Figure S39,
Supporting Information).
Figure 9. Graphical presentation of the experimental findings.
independent of solvent nature. However, any conclusive mechanistic
interpretation of these results awaits more advanced excited-state
calculations. To explain the effect of the solvent, extremely rare and
computationally expensive calculations need to be carried out to out-
line the deactivation pathways of the S1 state in the presence of solvent
molecules.
’ ASSOCIATED CONTENT
S
Supporting Information. Spectra and kinetics of ultra-
b
fast studies, the quantum yields of triplet formation (ΦT) for
2-NpCON3 in different solvents, and details of the quantum
chemical calculations. This material is available free of charge via
the Internet at http://pubs.acs.org.
5 . EXPERIMENTAL SECTION
5.1. Materials and Experimental Techniques. Ultrafast IR and
UVꢀvis pumpꢀprobe absorption measurements were performed using
home-built spectrometers at The Ohio State University. The time resolu-
tion is about 300 fs for both the UVꢀvis and TRIR instruments. They are
described in more detail elsewhere.33,35 Some femtosecond TRIR experi-
ments with 350 nm excitation (time resolution 300 fs) were performed in
Lille, France.23 The absorbance of the sample solutions was about 1.0 in a
1 mm cell at the excitation wavelength. Sample solutions were excited in a
stainless steel flow cell equipped with 2 mm thick BaF2 windows in the case
of the IR instrument and CaF2 windows (1 mm thick front window and
2 mm thick back window) in the case of UVꢀvis instrument. After passing
through the sample, the reference and probe beam were spectrally dispersed
with a polychromator and independently imaged on a liquid-nitrogen
cooled HgCdTe detector (2 ꢁ 32 pixels) (TRIR setup) and thermoelec-
trically cooled CCD camera (UVꢀvis setup). The pump pulse energy was
e4 μJ at the sample position, and the pump beam diameter (fwhm) was
equal to about 250 μm. The entire set of pumpꢀprobe delay positions
(cycle) is repeated at least three times, to observe data reproducibility from
cycle to cycle. To avoid rotational diffusion effects, the angle between
polarization of the pump beam and the probe beam was set to the magic
angle (54.7ꢀ). Kinetic traces were analyzed by fitting to a sum of exponential
terms. All experiments were performed at room temperature. The relative
actinometry method was applied to estimate the quantum yield of triplet
formation ΦT for 2-NpCON3 (Supporting Information).
Acetonitrile, chloroform, and methanol (Burdick and Jackson, spec-
trometric grade) were used as received. Unless otherwise noted, other
materials were obtained from Sigma Aldrich Chemical Co. and used
without further purification. The 2-NpCON3 was synthesized following
the procedure described for benzoyl azide by Barrett and Porter.9,36 The
2-naphthyl isocyanate (2-NpNCO) was bought from Aldrich and used
as received. Its purity is 97%.
5.2. Computational Methodology. All calculations were per-
formed using the Turbomole-5.91 suite of programs except complete
active space (CAS) calculations.37ꢀ39 The ground-state geometry of
2-NpCON3 was optimized at the B3LYP level40,41 with triple-ζ valence
polarized basis sets (TZVP),42 using C1 symmetry. Vertical excitations
were computed using the time-dependent (TD) B3LYP43 level of theory
at these ground-state geometries. To characterize the excited states, we
’ AUTHOR INFORMATION
Corresponding Author
hadad.1@osu.edu; platz.1@osu.edu
Present Addresses
#Department of Chemistry, 101 South Rd., The University of
North Carolina, Chapel Hill, NC 27514.
’ ACKNOWLEDGMENT
This work was performed at The Ohio State University Center
for Chemical and Biophysical Dynamics. Support by the National
Science Foundation is gratefully acknowledged along with generous
allocations of computational resources by the Ohio Supercom-
puter Center. M.S. acknowledge funding from the French Centre
Nationale de la Recherche Scientifique (CNRS), the Ministꢁere de
l’Enseignement Supꢀerieur et de la Recherche and the Rꢀegion Nord-
Pas de Calais. S.V. gratefully acknowledges an Ohio State University
Presidential Fellowship.
’ REFERENCES
(1) Platz, M. S. Nitrenes in Reactive Intermediate Chemistry; Moss,
R. A.; Platz, M. S.; Jons, M. J., Eds.; John Wiley & Sons, Inc.: Hoboken,
NJ, 2005; pp 501ꢀ559.
(2) Gritsan, N. P. Russ. Chem. Rev. 2007, 76, 1139.
(3) Bayley, H.; Staros, J. Azides and Nitrenes; Reactivity and Utility;
Scriven, E.F.V., Ed.; Academic Press: New York, 1984.
(4) Gritsan, N. P.; Platz, M. S. Chem. Rev. 2006, 106, 3844.
(5) Pritchina, E. A.; Gritsan, N. P.; Maltsev, A.; Bally, T.; Autrey, T.;
Liu, Y.; Wang, Y.; Toscano, J. P. Phys. Chem. Chem. Phys. 2003, 5, 1010.
(6) Gritsan, N. P.; Pritchina, E. A. Mendeleev Commun. 2001, 11, 94.
(7) Sigman, M. E.; Autrey, T.; Schuster, G. B. J. Am. Chem. Soc. 1988,
110, 4297.
(8) Autrey, T.; Schuster, G. B. J. Am. Chem. Soc. 1987, 109, 5814.
(9) Liu, J.; Mandel, S.; Hadad, C. M.; Platz, M. S. J. Org. Chem. 2004,
69, 8583.
9760
dx.doi.org/10.1021/ja109098w |J. Am. Chem. Soc. 2011, 133, 9751–9761

Journal of the American Chemical Society
ARTICLE
ꢀ
(10) McCulla, R. D.; Gohar, G. A.; Hadad, C. M.; Platz, M. S. J. Org.
Chem. 2007, 72, 9426.
(11) Caraculacu, A. A.; Coseri, S. Prog. Polym. Sci. 2001, 26, 799.
(12) Wentrup, C.; Bornemann, H. Eur. J. Org. Chem. 2005, 2005
4521.
(49) Karlstr€om, G.; Lindh, R.; Malmqvist, P.-L.; Roos, B. O.; Ryde,
U.; Veryazov, V.; Widmark, P.-O.; Cossi, M.; Schimmelpfennig, B.;
Neogrady, P.; Seijo, L. Comput. Mater. Sci. 2003, 28, 222.
ꢀ
(50) Forsberg, N.; Malmqvist, P.-L. Chem. Phys. Lett. 1997, 274, 196.
(13) Linke, S.; Tisue, G. T.; Lwowski, W. J. Am. Chem. Soc. 1967, 89
6308.
(14) Tisue, G. T.; Linke, S.; Lwowski, W. J. Am. Chem. Soc. 1967, 89
6303.
(15) Lwowski, W.; Tisue, G. T. J. Am. Chem. Soc. 1965, 87, 4022.
(16) Zabalov, M.; Tiger, R. Russ. Chem. Bull. 2005, 54, 2270.
(17) Eibler, E.; Sauer, J. Tetrahedron Lett. 1974, 15, 2569.
(18) Semenov, V. P.; Studenikov, A. N.; Bespalov, A. D.; Ogloblin,
K. A. Zh. Organ. Khim. (Russ.) 1977, 13, 2202.
(19) Kubicki, J.; Zhang, Y.; Wang, J.; Luk, H. L.; Peng, H.-L.; Vyas,
S.; Platz, M. S. J. Am. Chem. Soc. 2009, 131, 4212.
(20) Bonneau, R.; Liu, M. T. H.; Kim, K. C.; Goodman, J. L. J. Am.
Chem. Soc. 1996, 118, 3829.
(21) Desikan, V.; Liu, Y.; Toscano, J. P.; Jenks, W. S. J. Org. Chem.
2007, 72, 6848.
(22) Desikan, V.; Liu, Y.; Toscano, J. P.; Jenks, W. S. J. Org. Chem.
2008, 73, 4398.
(23) Zhang, Y.; Burdzinski, G.; Kubicki, J.; Vyas, S.; Hadad, C. M.;
Sliwa, M.; Poizat, O.; Buntinx, G.; Platz, M. S. J. Am. Chem. Soc. 2009,
131, 13784.
(24) Suppan, P.; Ghoneim, N. Solvatochromism; The Royal Society
of Chemistry, Cambridge, 1997.
(25) Nibbering, E. T. J.; Elsaesser, T. Appl. Phys. B: Lasers Opt. 2000,
71, 439.
(26) Carmichael, I.; Hug, G. L. J. Phys. Chem. Ref. Data 1986, 15, 1.
(27) Klessinger, M.; Michl, J. In Excited States and Photochemistry of
Organic Molecules; VCH Publishers, Inc.: New York, 1995; Chapter 6.
Photochemical Reaction Model. 309ꢀ360.
(28) Rabinovitch, B. S.; Rynbrandt, J. D. J. Phys. Chem. 1971, 75, 2164.
(29) Hamm, P. Chem. Phys. 1995, 200, 415.
(30) Wang, J.; Kubicki, J.; Burdzinski, G.; Hackett, J. C.; Gustafson,
T. L.; Hadad, C. M.; Platz, M. S. J. Org. Chem. 2007, 72, 7581.
(31) Wang, J.; Burdzinski, G.; Zhu, Z.; Platz, M. S.; Carra, C.; Bally,
T. J. Am. Chem. Soc. 2007, 129, 8380.
(32) Wang, J.; Kubicki, J.; Platz, M. S. Org. Lett. 2007, 9, 3973.
(33) Burdzinski, G.; Hackett, J. C.; Wang, J.; Gustafson, T. L.;
Hadad, C. M.; Platz, M. S. J. Am. Chem. Soc. 2006, 128, 13402.
(34) Li, G.; Parimal, K.; Vyas, S.; Hadad, C. M.; Flood, A. H.; Glusac,
K. D. J. Am. Chem. Soc. 2009, 131, 11656.
(35) Wang, J.; Burdzinski, G.; Kubicki, J.; Platz, M. S. J. Am. Chem.
Soc. 2008, 130, 11195.
(36) Barrett, E. W.; Porter, C. W. J. Am. Chem. Soc. 1941, 63, 3434.
(37) Treutler, O.; Ahlrichs, R. J. Chem. Phys. 1995, 102, 346.
(38) Ahlrichs, R.; B€ar, M.; H€aser, M.; Horn, H.; K€olmel, C. Chem.
Phys. Lett. 1989, 162, 165.
(39) For the current version of TURBOMOLE, see http://www.
turbomole.de.
(40) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(41) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
(42) Weigend, F.; H€aser, M.; Patzelt, H.; Ahlrichs, R. Chem. Phys.
Lett. 1998, 294, 143.
(43) Olivucci, M. Computational Photochemistry; Elsevier: Amsterdam,
2005.
(44) Vyas, S.; Onchoke, K. K.; Rajesh, C. S.; Hadad, C. M.; Dutta,
P. K. J. Phys. Chem. A 2009, 113, 12558.
(45) Vyas, S.; Hadad, C. M.; Modarelli, D. A. J. Phys. Chem. A 2008,
112, 6533.
(46) Burdzinski, G. T.; Gustafson, T. L.; Hackett, J. C.; Hadad,
C. M.; Platz, M. S. J. Am. Chem. Soc. 2005, 127, 13764.
ꢀ
(47) Finley, J.; Malmqvist, P.-L.; Roos, B. O.; Serrano-Andrꢀes, L.
Chem. Phys. Lett. 1998, 288, 299.
(48) Pierloot, K.; Dumez, B.; Widmark, P.-O.; Roos, B. O. Theor.
Chim. Acta 1995, 90, 87.
9761
dx.doi.org/10.1021/ja109098w |J. Am. Chem. Soc. 2011, 133, 9751–9761