NO product yield excitation spectrum of the S0→ S2 transition of nitrosobenzene in a supersonic jet

Article abstract of DOI:10.1016/S0009-2614(98)00448-5

The absorption spectrum of the S₀→ S₂ transition of ultracold nitrosobenzene in a supersonic jet was measured indirectly by monitoring the product yield of the NO fragment. The vibrational structure and relative line intensities are in good agreement with the corresponding spectrum in an argon matrix. It is concluded that the yield of photofragmentation is constant for all vibronic S₂-states up to an excess energy of 3500 cm^-1. The electronic origin line has a Lorentzian lineshape with a homogeneous width of (90 ± 5) cm^-1. This corresponds to a lifetime of τ(S₂) = 60±3 fs.

Full text of DOI:10.1016/S0009-2614(98)00448-5

19 June 1998
Ž
.
Chemical Physics Letters 289 1998 516–520
NO product yield excitation spectrum of the S0 ™ S2 transition
of nitrosobenzene in a supersonic jet
Angela Keßler, Uwe Kensy, Bernhard Dick
Institut fur Physikalische und Theoretische Chemie, UniÕersitat Regensburg, 93053 Regensburg, Federal Republic of Germany
¨
¨
Received 17 March 1998; in final form 14 April 1998
Abstract
The absorption spectrum of the S0™ S2 transition of ultracold nitrosobenzene in a supersonic jet was measured
indirectly by monitoring the product yield of the NO fragment. The vibrational structure and relative line intensities are in
good agreement with the corresponding spectrum in an argon matrix. It is concluded that the yield of photofragmentation is
constant for all vibronic S2-states up to an excess energy of 3500 cmy1. The electronic origin line has a Lorentzian lineshape
y
1
Ž
.
Ž
.
with a homogeneous width of 90 " 5 cm . This corresponds to a lifetime of t S2 s 60"3 fs. q 1998 Elsevier
Science B.V. All rights reserved.
y5
w
x
1. Introduction
6–8 . An upper limit of FF F10
has been esti-
mated for the fluorescence quantum yield in heptane
Nitrosobenzene has several rather unusual photo-
physical and photochemical properties. One of these
is the large energy gap between the first excited
singlet state S1 and the second excited singlet state
S2. The S0 ™ S1 transition is very weak and occurs
in the deep red spectral region. The origin was
solution by using the spontaneous Raman scattering
w x
of the solvent as an internal standard 7 . In the gas
phase no signal above the noise level could be
detected, leading to FF F5=10y7 7 . When this
w x
was combined with the rate constant for the radiative
decay estimated from the integrated absorption spec-
observed at 11504 cmy1 in the gas phase 1 and at
trum kr S2 f10 s , a lifetime of t S2 f10 fs
8
y1
w x
Ž
Ž
.
.
Ž
.
11442 cmy1 in an argon matrix 2 . The origin of the
was estimated 7 .
w x
w x
S0 ™ S2 transition has so far only been localized in
Nitrosobenzene efficiently fragments into an NO
radical and a phenyl radical following excitation to
the S2 state. This reaction occurs even at tempera-
tures as low as 10 K in an argon matrix. The phenyl
radical in the argon matrix has been identified by IR
an argon matrix at 30260 cmy1 2 . The energy gap
w x
between these states D E12 is larger than 18800
cmy1. For the molecule azulene, the origin of the
S1-state was found in the supersonic jet at 14285
y1
cmy1 3,4 and that of the S2-state at 28757 cm
w x
9,10 and UV 11 absorption spectroscopy.
The observation of a well resolved S0 ™ S2
w
x
w
x
w x
y1
5 . In this case the energy gap D E12 s14473 cm
is large enough that fluorescence from the S2-state
can be observed. However, in spite of the larger
energy gap in nitrosobenzene, no fluorescence from
the S2 state could be identified in this compound
origin band in the argon matrix in combination with
the photochemical activity opened the possibility of
a measurement of the homogeneous linewidth with
spectral hole-burning. This should yield a much more
0009-2614r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S0009-2614 98 00448-5

(
)
A. Keßler et al.rChemical Physics Letters 289 1998 516–520
517
precise value of the lifetime of the excited state.
Unfortunately, it turned out that bleaching of the
absorption in the origin band was reversible within a
mated, or the radiative lifetime overestimated, in
applying the Strickler-Berg relation.
Hence it seemed interesting to measure the ab-
sorption spectrum of the isolated molecule in the gas
phase near the electronic origin of the S0 ™ S2
transition. Under these conditions – preferably in a
supersonic jet – broadening due to matrix sites and
crystal defects are completely absent. This experi-
ment should also answer the question whether ni-
trosobenzene has a low-frequency mode with a
wavenumber of ca. 40 cmy1 in the S2-state or not.
Due to the complete lack of detectable fluores-
cence from nitrosobenzene in the S2-state a fluores-
cence excitation spectrum could not be measured.
We therefore decided to measure the product yield
spectrum of NO as a function of the wavenumber of
excitation. The principle of the experiment is appar-
ent from Fig. 1. Cold nitrosobenzene is produced in
a supersonic expansion and excited to the S2-state by
a tunable photolysis laser with wavenumber nD. The
excited nitrosobenzene will dissociate with the quan-
tum yield of dissociation FD, forming a phenyl
radical and an NO radical. The latter is excited by
the probe laser tuned to a characteristic transition of
NO at wavenumber nP . The fluorescence following
this excitation is monitored and normalized to the
pulse energies of the two laser pulses. As long as the
excitation steps are not saturated, the resulting spec-
trum will be the product of the absorption spectrum
and the quantum yield of photodissociation of ni-
trosobenzene into this particular product state of NO.
w x
few minutes 2 . This was attributed either to a
recombination reaction within the matrix cage, or to
a barrier for the photoreaction on the S2 potential
energy surface. In this case the temporal bleaching of
the absorption could be explained by the assumption
that the excited molecules relax to a metastable
matrix site. At higher excitation energies the photo-
reaction was irreversible, but no spectrally homoge-
neous subensemble could be selected in this case.
A closer examination of the origin band in the
matrix spectrum revealed two shoulders. The red
edge could be fitted by a Lorentzian lineshape. In
fact, the whole band could be fitted by a superposi-
tion of three Lorentzian lines with the same width of
67 cmy1. The substructure was attributed to the
existence of three matrix sites, although a low-
frequency vibration with a wavenumber of f40
cmy1 could not be ruled out as an alternative expla-
nation. If residual inhomogeneous broadening and
pure dephasing can be neglected, this linewidth cor-
Ž
.
responds to a lifetime of t S2 f80 fs. This value is
certainly a lower limit to the lifetime in the argon
matrix. Apparently there is a discrepancy between
Ž
.
this lifetime and the estimate of t S2 f10 fs ob-
tained from the integrated absorption spectrum and
the upper limit of the fluorescence quantum yield in
the gas phase. Several reasons could account for this.
The lifetime might strongly depend on the vibra-
tional level of the S2-state. The lineshape analysis
gives a value only for the electronic origin, whereas
the other method yields a value which is averaged
over the whole absorption band. The matrix mea-
surement was performed at 15 K whereas the mea-
w x
surements in Ref.
7 were carried out at room
temperature. A strong increase in the decay rate with
increasing temperature could be due to the existence
of a thermally activated decay channel. Alterna-
tively, the lifetime could strongly depend on the
environment. The fluorescence quantum yield esti-
mated for the heptane solution is compatible with a
lifetime of f100 fs. One could imagine that the
rigid matrix cage exerts additional forces upon the
molecule which inhibit photodissociation in the vi-
brational ground state of S2. Finally, the fluores-
cence quantum yield could have been underesti-
Fig. 1. Schematic representation of the experiment. nD: frequency
of the tunable photolysis laser. nP : frequency of the fixed probe
laser.

(
)
518
A. Keßler et al.rChemical Physics Letters 289 1998 516–520
This quantum yield might depend on the excess
energy. A barrier to photodissociation would become
apparent by a strong increase in this quantum yield
above a certain threshold. In this case it might be
possible to observe the fluorescence from ni-
trosobenzene excited below the threshold.
2. Experimental
The jet apparatus consists of a single vacuum
chamber of 271 mm diameter and 515 mm length
which is pumped by the combination of an oil
Ž
.
diffusion pump Leybold DI 3000 , a roots pump
Ž
.
Leybold, RUVAC WA 251 and a rotary pump
Leybold Trivac D 65B . Nitrosobenzene was pur-
Ž
.
Fig. 2. A NO photoproduct-yield spectrum of nitrosobenzene in
Ž .
Ž
.
the region of the S0 ™ S2 transition.
tion spectrum of nitrosobenzene in an argon matrix, shifted by 988
Ž .
C the same
B Superimposed absorp-
chased from Aldrich and was used, with the same
results, either directly or purified by sublimation.
The sample was heated to 508C, seeded into 1.5–2
bar of He and expanded through a pulsed valve
cmy1 to lower wavenumbers see axis on top .
Ž
.
matrix spectrum shifted by 0.2 units of intensity for clarity. The
jet spectrum was normalized for optimum overlap with the matrix
spectrum.
Ž
.
General valve No. 9 . The photolysis laser and the
probe laser are frequency-doubled dye lasers
.
published results . The matrix spectrum was shifted
Ž
.
Lambda Physics LPD 3002 and FL 3002 both
by 988 cmy1 to higher wavenumbers in order to
obtain the best coincidence of the electronic origin
bands. The lower abscissa refers to the matrix spec-
trum, the upper abscissa to the jet spectrum. Appar-
ently both spectra match perfectly. All details of the
vibrational structure seen in the matrix spectrum
appear also in the jet spectrum. In order to empha-
size the similarity of the spectra, which might be
obscured by the perfect overlap, the matrix spectrum
is displayed a second time shifted by 0.2 units of
)
Ž
pumped by the same XeCl -excimer laser Lambda
Physics Lextra 200 . The laser beams were aligned
.
to overlap in a counterpropagating geometry. Stray
light was eliminated by home-built light skimmers
w
x
12 . Fluorescence from NO was collected with a
Ž
.
Ž
lens fs25 mm and a spherical mirror Rs60
.
Ž
mm and detected with a photomultiplier Thorn Emi
9789QA . The laser pulse energies were measured
.
with photodiodes. All signals were integrated with
Ž
boxcar averagers Stanford Research Systems
Ž
.
intensity curve C .
.
Ž
.
SR250 , digitized SR245 and processed by a PC,
which also controlled the trigger of the laser and the
valve as well as the dye laser scan.
As long as the two lasers do not saturate the
transitions to which they are tuned, the product-
yield spectrum should be proportional to
Ž
. Ž
.
Ž
.
Ž
.
s nD F nD Pj NO , where s nD is the absorption
Ž
.
cross section of nitrosobenzene, F nD the quantum
3. Results and discussion
Ž
.
yield of photodissociation and Pj NO the fraction of
NO molecules produced in the particular state moni-
tored by the probe laser. Since, apart from a shift of
the spectrum, the absorption spectrum in argon ma-
trix and the photofragment yield spectrum in the jet
are proportional to each other, the factor
The photodissociation yield was monitored on the
2
X
2
XX
11
12.5
qŽ
.
Ž
.
Q
line of the A S Õ s0 §X P1r2 Õ s0
transition of NO. The product-yield excitation spec-
trum of nitrosobenzene obtained in this way in the
spectral region between 30400 and 34400 cmy1 is
shown in Fig. 2 as curve A. It is overlayed, after
proper scaling of the vertical axis, with the absorp-
tion spectrum of the S0 ™ S2 transition of ni-
Ž
.
Ž
.
F nD Pj NO must be constant over the energy
range investigated. We have scanned the excitation
spectrum of NO for several fixed values of nD in the
range between 31250 and 34500 cmy1. We found
Ž
trosobenzene in an argon matrix J.M. Engert, un-

(
)
A. Keßler et al.rChemical Physics Letters 289 1998 516–520
519
that the population of the rotational states of NO was
similar in all cases. Hence we conclude that the
corresponds to the 95 % probability range, i.e. two
standard deviations if a Gaussian distribution of the
Ž
.
w x
quantum yield of photodissociation F nD is con-
stant over the whole energy range investigated.
parameter values is assumed 13 . If further sources
of broadening existed, the contribution of the life-
time to the linewidth could only be smaller than 90
Under the same expansion conditions we found
sharp spectra with narrow lines for molecules with
long-lived excited states like NO or tetracene. For
NO the relative intensities of the few lines observed,
correspond to a rotational temperature of 1 K. In the
case of tetracene, the width of the rotational enve-
lope was ca. 1.6 cmy1 and the best fit was obtained
for a simulation with a rotational temperature of 8 K.
No vibrational hot bands could be found. Hence we
believe that the spectrum in Fig. 2 is the spectrum of
cold nitrosobenzene. However, all lines and in par-
ticular the 000-line are much broader than the ex-
pected rotational envelope. If this linewidth is deter-
mined by a short lifetime of the excited state, the
shape of the line should be a Lorentzian. A fit of a
Lorentzian to the 000-line, shown in Fig. 3, is indeed
in good agreement with the data. The agreement is
far better than with a Gaussian lineshape as a trial
function. Since the line is apparently homogeneous
and since in the absence of collisions and at low
temperature the only mechanism for linebroadening
is the decay of the excited state, we interpret the
cmy1 and hence a lifetime of t S2 s60 fs is a
lower limit.
Ž
.
If the lifetime were as small as 10 fs, we would
have seen a line width of at least 530 cmy1. This is
inconsistent with the presence of the resolved vibra-
tional finestructure in the absorption spectrum. Ap-
parently the estimate of the lifetime with the help of
w x
the Strickler–Berg relation given in Ref. 7 is too
small. One source of error could be that the sensitiv-
ity of the apparatus was overestimated and hence the
value for the quantum yield of fluorescence in the
gas phase FF -5=10y7 was underestimated. A
radiative lifetime of 10 ns and a real lifetime of 60 fs
should lead to a quantum yield of fluorescence of
FF s6=10y6 which is certainly difficult to detect.
However, integration of the absorption spectrum can
yield a value for the radiative lifetime which is too
small if the intensity of the band is mainly borrowed
through vibronic coupling from another state. It has
been argued that the Strickler–Berg relation can not
w
x
be applied in such cases 14 . According to our
semiempirical CNDOrS calculations the S2-state of
nitrosobenzene originates from a doubly excited nn
Ž
.
width of the Lorentzian FWHM of Dn0 s90.2"
4.6 cmy1 as the manifestation of an excited state
lifetime of t S2 s60"3 fs. The error estimate
™p )p ) configuration 2 . The intrinsic oscillator
Ž
.
w x
strength is hence small.
The difference between the lifetimes obtained
Ž
.
from the linewidth in the jet 60 fs and in the matrix
Ž
.
80 fs is significantly larger than the error estimated
from the fitting procedure for each value. The fit
yields slightly different results when a linear back-
ground is added to the fitting function, but this is not
enough to explain the difference. We are, hence, lead
to the conclusion that the matrix lengthens slightly
the lifetime of the S2-state. If the origin line in the
jet spectrum were the superposition of three
Lorentzian lines, as is the case in the argon matrix,
fits of the red edge of the line and of the whole line
would yields quite different values for the linewidth.
This is, however, not the case. Hence we conclude
that the substructure seen in the argon matrix is due
to three matrix sites and not the manifestation of a
low-frequency mode.
Fig. 3. Expanded view of the product-yield spectrum of Fig. 2 in
the region of the electronic origin line. The heavy line is a fit of a
Ž
.
Ž
.
Lorentzian lineshape with a width FWHM of Dn0 s 90.2"4.6
y1
cmy1 centered at n0 s 31247.9"0.6 cm
.
Ž
.

(
)
520
A. Keßler et al.rChemical Physics Letters 289 1998 516–520
4. Conclusions
potential energy surface of the S2-state. Fast internal
conversion to the electronic ground state followed by
dissociation on the S0 potential surface could be an
alternative mechanism. In order to clarify this ques-
tion we have measured the product state distribution
of NO for several photolysis energies. The results
will be presented in a forthcoming paper.
These observations allow us to draw the following
conclusions:
1. The electronic origin of the S0 ™ S2 transition of
isolated nitrosobenzene in the gas phase occurs at
31248 cmy1. In an argon matrix, the only other
medium in which a vibrational structure of this
transition could be resolved so far, the origin is
observed at 30260 cmy1
.
Acknowledgements
2. Since the relative intensities of the vibronic lines
in the photodissociation product-yield spectrum of
this work perfectly agree with those from the
absorption spectrum in an argon matrix, the quan-
tum yield of photodissociation must be constant,
at least over the energy range from the electronic
origin of the S0 ™ S2 transition up to an excess
This work has been supported by the Deutsche
Forschungsgemeinschaft and the Fonds der Chemis-
chen Industrie.
References
energy of ca. 3500 cmy1
.
w x
1
Ž
.
M.R.S. McCoustra, J. Pfab, Chem. Phys. Lett. 122 1985
395.
3. The statement above is in particular true for the
electronic origin line. There is no evidence for a
barrier to photodissociation on the potential en-
ergy surface of the S2-state. The fact that the
photobleaching after excitation into the electronic
origin band of nitrosobenzene in an argon matrix
appears to be reversible must then be due to
efficient recombination in the matrix cage.
w x
2
J.M. Engert, A. Slenczka, U. Kensy, B. Dick, J. Phys. Chem.
Ž
.
100 1996 11883.
w x
Ž .
3
w x
4
A. Amirav, J. Jortner, J. Chem. Phys. 81 1984 4200.
Ž
.
T. Suzuki, M. Ito, J. Phys. Chem. 91 1987 3537.
w x
5
M. Fujii, T. Ebata, N. Mikami, M. Ito, Chem. Phys. 77
Ž
.
1983 191.
w x
Ž
.
6
w x
7
V.V. Bhujle, Spectry Letters 10 1977 587.
D.A. Chernoff, R.M. Hochstrasser, Chem. Phys. Letters 70
Ž
.
4. The analysis of the lineshape of the electronic
origin line yields a homogeneous linewidth of 90
cmy1 corresponding to a lower limit of the ex-
1980 213.
w x
8
D.A. Condirston, A.R. Knight, R.P. Steer, J. Photochemistry
Ž
.
14 1980 257.
w x
9
W.G. Hatton, N.P. Hacker, P.H. Kasai, J. Chem. Soc., Chem.
Ž
.
cited state lifetime of t S2 s60 fs.
Ž
.
Commun. 1990 227.
5. The substructure of the three lines with similar
intensities and a spacing of ca. 40 cmy1 seen in
the argon matrix is not due to a low-frequency
mode of nitrosobenzene in the S2-state but most
likely results from three matrix sites.
w
x
10 J.G. Radziszewski, M.R. Nimlos, P.R. Winter, G.B. Ellison,
Ž
.
J. Am. Chem. Soc. 118 1996 7400.
w
w
w
x
Ž
.
11 J.M. Engert, B. Dick, Appl. Phys. B 63 1996 531.
x
Ž
.
12 W.R. Gentry, C.F. Giese, Rev. Sci. Instr. 46 1975 104.
x
13 W.H. Press, S.A. Teukolsky, W.T. Vetterling, B.P. Flannery,
Numerical Recipes in FORTRAN, 2nd ed., Cambridge Uni-
versity Press 1992, p. 684 ff.
From the data presented here it can not be con-
cluded that the photodissociation takes place on the
w
x
Ž
.
14 B. Nickel, J. Hertzberg, Chem. Phys. 132 1989 219.