Laser ionization measurements of the photodissociation kinetics of jet-cooled benzaldehyde

Article abstract of DOI:10.1021/jp9707551

The kinetics of benzaldehyde dissociation into benzene and carbon monoxide following excitation to the S₂ state has been investigated under jet-cooled conditions by two-color laser ionization mass spectrometry using a pump-probe technique. A kinetic model has been proposed to explain the observed biexponential decay of the benzaldehyde signal and the single-exponential growth of the benzene product signal in terms of a sequential decay of two excited states of benzaldehyde, one of which leads to formation of benzene molecules in its lowest triplet state. Reactant disappearance and product appearance rates were determined for a number of vibronic transitions in the first 2000 cm^-1 energy region of the S₂ state. At the S₂ origin the decay rate of the short-lived state of benzaldehyde was found to be 1.2 × 10⁶ s^-1.

Full text of DOI:10.1021/jp9707551

7934
J. Phys. Chem. A 1997, 101, 7934-7942
Laser Ionization Measurements of the Photodissociation Kinetics of Jet-Cooled
Benzaldehyde
†
Cristina R. Silva and James P. Reilly*
Department of Chemistry, Indiana UniVersity, Bloomington, Indiana 47405
X
ReceiVed: February 27, 1997; In Final Form: August 4, 1997
The kinetics of benzaldehyde dissociation into benzene and carbon monoxide following excitation to the S
2
state has been investigated under jet-cooled conditions by two-color laser ionization mass spectrometry using
a pump-probe technique. A kinetic model has been proposed to explain the observed biexponential decay
of the benzaldehyde signal and the single-exponential growth of the benzene product signal in terms of a
sequential decay of two excited states of benzaldehyde, one of which leads to formation of benzene molecules
in its lowest triplet state. Reactant disappearance and product appearance rates were determined for a number
-
1
2 2
of vibronic transitions in the first 2000 cm energy region of the S state. At the S origin the decay rate
6
-1
of the short-lived state of benzaldehyde was found to be 1.2 × 10 s .
Introduction
pressures, collisional deactivation is faster than dissociation, and
the result is that the emitting state is populated. The shape of
the wavelength-resolved phosphorescence spectrum was found
to be independent of pressure with the emission originating from
Benzaldehyde, the simplest aromatic carbonyl compound, is
an interesting molecular system due to the presence of close-
lying ππ* and nπ* states. Intersystem crossing (ISC) is favored
by spin-orbit coupling, and excitation to the singlet states is
followed by phosphorescence, not by fluorescence. A reactive
channel leading to dissociation into benzene and carbon
monoxide is observed following activation to S2(ππ*). We have
recently reported the excitation spectrum of jet-cooled benzal-
dehyde in the S2 r S0 absorption band.1 In the present work
we employ the same experimental apparatus to investigate the
kinetics of dissociation of jet-cooled benzaldehyde following
excitation of different vibronic levels in the S2 state and to
explore the relaxation dynamics of its excited states under
isolated conditions.
2
,6
the vibrationless level of T(nπ*). The vapor-phase emission
of benzaldehyde has been observed under high resolution in a
1
8-24
number of studies.
fluorescence,
Apart from some weak E-type delayed
6
,22
emission occurs from the vibrationless level
of T(nπ*), with the CO stretching vibration being the dominant
accepting mode. These results are in agreement with molecular
1
orbital calculations that predict that the largest geometric change
on nπ* excitation corresponds to a substantial lengthening of
the CO bond.
Laser-induced multiphoton ionization mass spectrometry
(
MPI/MS) has also been used to investigate the photodissocia-
9
-17
tion of benzaldehyde.
The general trend observed is that
The photodissociation of benzaldehyde has been investigated
2,3
2-8
extensive benzene ion formation occurs following excitation into
S2(ππ*), while mainly benzaldehyde ions are observed following
excitation to S1(nπ**). The presence of benzene ions in the
MPI mass spectrum of benzaldehyde can be explained either
by ionization of neutral benzene formed as a product in the
photodissociation of benzaldehyde or by fragmentation of
by photolysis, phosphorescence emission, and multiphoton
ionization mass spectrometry9
-17
2
experiments. Berger et al.
reported that following irradiation into the S2 band of benz-
aldehyde vapor at 276 nm, benzene and carbon monoxide were
detected by gas chromatography. The dissociation quantum
yield was found to be pressure dependent, tending toward unity
1
4,15
at very low pressures. Bruhlmann et al.3 observed that
benzaldehyde ions. El-Sayed and co-workers
studied the
ionization-fragmentation mass pattern of benzaldehyde using
nanosecond laser pulses at 355 and 266 nm and concluded that
both processes occurred, the predominance of either one being
a function of excitation wavelength and laser intensity.14 Using
photolysis of a mixture of equal parts of benzaldehyde-h6 and
benzaldehyde-d6 at 276 nm produced exclusively benzene-h6
and benzene-d6, and they proposed a concerted reaction mech-
anism. Excitation to S1(nπ*) with wavelengths of 328 and 365
2
66 nm picosecond laser pulses, only C6H5CO radical cations
nm, on the other hand, did not lead to benzene or carbon
monoxide formation.2
were detected, indicating that further photon absorption occurred
before dissociation of benzaldehyde into neutral benzene took
A number of groups have investigated the wavelength and
pressure dependence of the phosphorescence emission of
1
5
place following the excitation step. Since large amounts of
benzene ions were observed using nanosecond duration laser
pulses, they concluded that the dissociation lifetime of benz-
aldehyde following excitation to S2 at this wavelength should
lie in the picosecond to nanosecond time range.
2
-8
5,6
benzaldehyde.
As reported by Baba and co-workers, the
relative phosphorescence quantum yield is nearly constant
throughout the S1 and S2 regions at higher pressures (600
mTorr), while at low pressures (2 mTorr) it decreases abruptly
at the onset of the S2 absorption band. These results were
interpreted in terms of an increase in the efficiency of a
nonradiative relaxation pathway in S2 leading to dissociation.
At low pressures, collisional deactivation is slow and dissocia-
tion occurs with consequent suppression of emission. At higher
Photoelectron spectroscopy has been used to identify the
benzene ion precursor as neutral ground-state benzene. Reilly
13
and co-workers reported that, when benzaldehyde is irradiated
at 258.9 nm, the observed laser photoelectron spectrum is
identical with that of benzene at this wavelength. This has been
interpreted as resonant stepwise ionization of benzene molecules
formed in the ground state following dissociation of benzalde-
hyde.
†
Permanent address: Instituto de Qu ´ı mica, Departamento de F ´ı sico-
Qu ´ı mica, Universidade Federal do Rio de Janeiro, Cidade Universit a´ ria CT
Bloco A, Rio de Janeiro, RJ, 21949-900, Brazil.
X
Polevoy et al.16,17 employed two-color laser ionization pump-
Abstract published in AdVance ACS Abstracts, September 15, 1997.
S1089-5639(97)00755-X CCC: $14.00 © 1997 American Chemical Society

Jet-Cooled Benzaldehyde
J. Phys. Chem. A, Vol. 101, No. 43, 1997 7935
probe experiments to investigate the dissociation time scale of
benzaldehyde over a range of excitation wavelengths in the S2
absorption band. They observed the kinetics of disappearance
of the reactant and appearance of the benzene product by
following the production of ions of these species as a function
of the time delay between the excitation and the ionization steps.
The rate of formation of benzene ions was found to be related
to the rate of disappearance of benzaldehyde ions, which, in
turn, was observed to increase exponentially with excitation
energy.
A pulse generator was used to fire the pulsed valve and trigger
a delay generator that provided the delay time for cold
benzaldehyde molecules to reach the ionization region. The
delay generator simultaneously triggered the pump laser and a
boxcar averager (EG&G Princeton Applied Research Model
162). The latter was used as a variable delay generator to trigger
the probe laser. Time intervals of 2.17 ( 0.02 and 5.35 ( 0.02
µs were scanned at an approximate resolution of 10 ns/point,
each point corresponding to 20 laser shots. Due to the relatively
short operating life of the F2 laser gas mixture, the laser pulse
energy dropped by 10-20% during the 4-9 min required to
scan these delay ranges at a repetition rate of 20 Hz. This decay
could not be neglected, and a procedure to correct the ion signal
for laser energy fluctuations is described below.
All of the above studies involved room-temperature gas-phase
molecules. The absorption and excitation spectra of benzalde-
2,4,25-27
hyde in the S2 r S0 region
are structured but highly
congested under these conditions. Apart from our recent
1
excitation spectrum, no other studies have probed the S2 state
The ion signal was detected by a microchannel plate detector,
recorded by a waveform digitizer (Biomation 6500) and
accumulated in a microcomputer. The digitizer was always
triggered by the probe laser so ions formed by two-color
ionization have the same flight time independent of the delay
between the lasers. Ions of the carrier gas produced by
under jet-cooled conditions. The nπ* states, in contrast, have
received more attention. In the past decade a number of studies
have addressed the spectroscopy and dynamics of these states
under free-jet conditions.2
8-31
In the present work we use a
supersonic expansion to obtain rotationally and vibrationally
cold benzaldehyde molecules that allow us to prepare a well-
defined single vibronic level of the excited state. Furthermore,
the molecules travel under collision-free conditions, and the
dissociation kinetics can be observed without the interference
of collisional processes. A pump-probe technique similar to
that employed by Polevoy et al. is used. The pump laser excites
the molecule to a resonant (single vibronic) level of the S2 state,
and the probe laser, delayed by a variable time, interrogates
the population of the transient states.
32
photoemission electron impact ionization induced by the VUV
radiation (hνioniz ) 7.9 eV) hitting the interior walls of the mass
1
spectrometer were used to monitor the F2 laser energy. The
+
Ar ion current was assumed to be proportional to the energy
of the probe laser, and it was used as a normalizing factor to
correct the benzaldehyde and benzene ion signals for variations
in the pulse energy of the probe laser. The pump laser pulse
energy remained constant, as indicated by measurements
performed before and after each scan.
To verify the precision of the normalization procedure, the
pump-probe delay was scanned forward and backward. In the
forward scan, as already described, the pump laser was fired at
a fixed time after the pulsed valse and the probe laser was
delayed relative to the pump pulse. In the backward scan, the
pump-probe delay was decreased by delaying the pump laser
trigger signal while the probe laser was fired at a fixed time
relative to the pulsed valve. These two arrangements excited
and probed slightly different parts of the molecular beam.
However, since the molecular beam pulse displayed a fwhm of
more than 100 µs, the molecular density did not vary signifi-
cantly over a 5 µs time interval (see Figure 8, ref 1, for
molecular beam shape). Decay rates obtained with the two
arrangements were the same within 10-20%.
Experimental Section
The experimental apparatus employed to record two-color
photoionization mass spectra has been recently described.
Briefly, the frequency-doubled output of a XeCl excimer-
pumped dye laser provided 80 µJ/pulse of exciting (pump)
1
-
1
radiation with a bandwidth of less than 0.4 cm , a pulse
duration of 3-5 ns, and a cross section of about 3 × 5 mm at
the ionization region. A VUV F2 laser connected to the mass
spectrometer through an evacuated glass tube provided 6 ns
pulses of 157 nm ionizing radiation. This was collimated by a
set of four apertures to produce a beam having a cross section
of about 4 × 6 mm, as measured by the light emitted in the
visible red. The apertures could be moved along the beam path
in order to maximize ion signal. Due to its VUV nature, the
probe laser pulse energy was not actually measured, but from
the fraction of the beam cross section that reached the ionization
region, it was estimated to be less than 0.3 mJ. The two laser
beams counterpropagated coaxially intercepting the molecular
beam orthogonally in the interaction region of our time-of-flight
mass spectrometer 15 cm downstream from the supersonic
nozzle.
Results
Mass Spectra. In an initial experiment to investigate the S2
state dynamics of benzaldehyde we employed an ArF laser as
the source of the probe radiation. Extensive benzene ion
formation was observed when benzaldehyde was irradiated with
this 193 nm light alone. Since the absorption maximum of the
S4 r S0 band, the strongest band in the optical spectrum of
33,34
The supersonic molecular beam consisted of a mixture of
approximately 1 Torr of benzaldehyde vapor seeded in 1.5 atm
of argon. The gas mixture was expanded through a 0.75 mm
diameter pulsed nozzle into the mass spectrometer chamber
benzaldehyde, lies around 194-195 nm,
this was interpreted
as the result of fast dissociation of S4 benzaldehyde molecules
followed by ionization of the nascent benzene molecules. The
F2 laser radiation, on the other hand, provided an efficient way
of ionizing electronically excited benzaldehyde without exciting
ground-state benzaldehyde molecules. We have successfully
used this radiation to record the two-color ionization excitation
spectrum of jet-cooled benzaldehyde in this region.1
-
5
where an average pressure of 2 × 10 Torr was maintained.
Under these expansion conditions the molecular beam traveled
at a speed of 600 m/s as determined by measurements of the
shift in the time of arrival of benzaldehyde molecules at the
interaction region caused by changes in the nozzle to interaction
region distance. On the basis of this measurement and on the
shapes of the two overlapping beams, we estimate that molecules
pass through the light interaction region in about 4 µs. Any
process that occurs on a time scale much longer than this cannot
be studied with this apparatus.
Under the light conditions employed, benzaldehyde ions were
observed only with two-color ionization. but ions of the carrier
+
2+
gas (Ar and Ar ) were produced whenever the F2 laser was
employed (see Figure 6, ref 1, for a comparison of one- and
two-color ionization mass spectra). The mechanism responsible
for the formation of the latter ions, electron impact ionization,

7936 J. Phys. Chem. A, Vol. 101, No. 43, 1997
Silva and Reilly
Figure 1. Two-color photoionization mass spectra of jet-cooled
benzaldehyde as a function of excess energy above the S origin. Peaks
at m/e ) 78 and 106 are assigned to benzene and benzaldehyde ions,
respectively. Left column: spectra scaled so that the tallest peak is the
same height. Right column: spectra normalized to light intensity (peaks
are proportional to absolute ion yields).
2
Figure 2. Photoionization excitation spectra of jet-cooled benzaldehyde
1
(
A) and benzene (B) in the region of the 6 S
1
r S
0
transition of
0
benzene: (A) two-color ion signal at m/e ) 78; (B) one-color ion signal
at m/e ) 78.
Wavelength Dependence. The S2 r S0 absorption band of
did not lead to any other ions because all other species were
present at much lower abundance than argon. As for the
dissociation reaction products, CO ions were not detected with
any combination of the two lasers, whereas benzene ions were
produced not only by two-color ionization but, at excitation
wavelengths corresponding to allowed transitions in the S1 r
S0 spectrum of benzene, by one-color ionization with the pump
laser.
benzaldehyde that extends from 284 nm to approximately 250
2
nm is featureless below 270 nm under room-temperature and
1
jet-cooled conditions. While acquiring the two-color excitation
1
spectrum of benzaldehyde, we noted that at some wavelengths
there was a significant enhancement of the benzene ion signal
even though the benzaldehyde ion signal displayed no spectral
dependence on the excitation step. This occurred at wavelengths
1
0
0
1
1
1
corresponding to the 6 and 6 transitions of the B2u r A1g
Figure 1 displays a series of two-color ionization mass spectra
of jet-cooled benzaldehyde taken at different excitation energies
within the S2 band. The probe laser was fired 10-20 ns after
the pump laser in order to allow some buildup of the benzene
population. In the left-hand column, each spectrum has been
scaled so that the tallest peak has the same height in order to
compare the relative amounts of benzene and benzaldehyde ions
produced at different excitation frequencies. The same data are
displayed in the right-hand column but with peak heights being
proportional to absolute ion yields. A trend clearly identified
band of benzene. As mentioned above, enhancement of the
benzene signal at these wavelengths did not depend on the
presence of the probe radiation; it could be observed when
benzaldehyde was irradiated with the pump laser alone.
The wavelength dependence of the benzene ion signal in the
259 nm region is shown in Figure 2. Part A of this figure shows
the wavelength dependence of the ion current observed at mass
channel 78 when jet-cooled benzaldehyde is irradiated in this
wavelength region. The signal displayed in part B is that
obtained at the same mass channel when jet-cooled benzene is
irradiated. The two bands are extremely similar, with the former
being slightly wider. On the basis of the two observed rotational
+
in this figure is that the C6H6+/C6H5CHO ratio increases (left
column) while the absolute benzaldehyde ion yield decreases
(
right column) with excess energy. These results are understood
contours, we estimate rotational temperatures of 10 and 5 K.
0
in terms of an increase in the dissociation rate with excitation
energy, a point that is confirmed by the time-dependent
measurements to be presented.
Irradiation of jet-cooled benzaldehyde in the region of the 6
1
S r S transition of benzene below 267 nm led to a similar
1
0
benzene-like wavelength dependence for the mass 78 ion signal.
These results clearly show that the observed signal comes from
the 1+1 one-color ionization of ground-state benzene molecules
with either 0 or 1 quantum of vibrational excitation in mode 6.
What is not absolutely clear is whether these molecules were
already present in the benzaldehyde sample as an impurity or
if they were formed in the photodissociation of benzaldehyde.
Although we did not purify the benzaldehyde sample prior to
its use, the latter hypothesis is consistent with previous mass
spectrometry and photoelectron spectroscopy results obtained
-
1
Note that the excitation frequency (3415 cm above the
-1
3
5 191 cm S2 origin) of the two-color ionization mass spectra
shown at the bottom of Figure 1 corresponds to the frequency
of the 6 S1 r S0 transition of benzene.
1
0
35,36
At this excitation
frequency, benzene ions can be observed with the pump laser
alone. They are produced through a resonant process in which
benzene molecules in the vibrationless ground electronic state
are electronically excited and then ionized. The resonance
enhancement that occurs at this particular excitation frequency
is discussed in the next section.
1
3
by this group. Long et al. observed that, when benzaldehyde

Jet-Cooled Benzaldehyde
J. Phys. Chem. A, Vol. 101, No. 43, 1997 7937
is irradiated at 258.9 nm, the strongest peak in the mass
spectrum, by 2 orders of magnitude, is at mass 78. In that
experiment care was taken to remove any benzene present in
the benzaldehyde sample by using a gas chromatographic inlet
with the mass spectrometer. Since the benzaldehyde mass
spectrum was unaffected by this separation and, in addition to
that, the photoelectron spectrum of benzaldehyde at 258.9 nm
was identical with that of benzene at this wavelength, the mass
decay is approximately that expected for the translational motion
of the molecular beam, we conclude that the decay of the
benzaldehyde and the benzene ion currents observed in the long
time range is due to the escape of excited molecules from the
laser interaction region. A third curve displayed in this part of
the figure corresponds to the logarithm of the sum of benzal-
dehyde and benzene ion yields. It illustrates an important aspect
related to mass conservation to be discussed later. The increase
in the dissociation rate with excitation energy is evident in both
columns of Figure 3. The solid lines represent the results of a
curve-fitting procedure following a kinetic model described in
the next section.
78 peak was attributed to ground-state benzene molecules
formed in the photodissociation of benzaldehyde. This assign-
ment, however, leaves an excess energy larger than 4 eV to be
distributed among translational motion of the reaction products
and the vibrational and rotational degrees of freedom of CO.
Such a product energy distribution seems unlikely under isolated
molecule conditions.
Analysis and Discussion
The above results characterize the photodissociation of
benzaldehyde as a predissociation process. Direct dissociation
can be ruled out on the basis of the time dependence of the
benzene ion signal observed in Figure 3 and the structured nature
of the absorption and excitation spectra of benzaldehyde in this
region. Figure 4 shows the energy level diagram for the
photodissociation of benzaldehyde into benzene and carbon
monoxide along with the ionization scheme and the kinetic
model proposed for the reaction. For benzaldehyde, T(nπ*),
S1(nπ*), and S2(ππ*) lie 3.1, 3.3, and 4.4 eV above the
An estimate of the relative population of benzene molecules
in V ) 1 and V ) 0 following excitation of benzaldehyde has
been made on the basis of our present measurements of the
0
1
intensity ratio of the 6 to the 6 bands. The value thus
1
0
obtained, 0.06 ( 0.02, is close to that reported by Atkinson
3
5
and Parmenter (0.056) for benzene under room-temperature
conditions, but it is somewhat larger than that reported by Rice
3
6
and co-workers (0.023) for a jet-cooled benzene expansion.
0
1
1
0
While, on one hand, the 6 to 6 band intensity ratio that we
measured is roughly the same as that for a room temperature
1,29
vibrationless level of S0, respectively. The triplet ππ* states
0
1
1
0
benzene sample, the contours of the 6 and 6 bands cor-
have not been observed in the gas phase, but according to
respond to a sample with a rotational temperature about twice
that of jet-cooled benzene. These limited results thus prevent
us from drawing firm conclusions as to the relative amounts of
benzene impurity and photoproduct.
In the analysis of the results of the pump-probe experiments
presented in the next section it is shown that the product entrance
channel interrogated by the probe radiation corresponds to
electronically excited benzene. From our estimates of the
dissociation quantum yield this is indeed the dominant entrance
channel, but we are unable to establish whether this is the only
channel of benzene formation. The benzaldehyde photodisso-
ciation product energy distribution certainly requires further
investigation.
1,37
molecular orbital calculations, the lowest excited triplet state
is not the triplet nπ* state observed in emission but a ππ* state.
1
According to our ab initio calculations, there should be at least
three triplet states below the origin of S2 ordered as T1(ππ*),
T (nπ*), and T (ππ*).
2
3
In the reaction C6H5CHO f C6H6 + CO, two bonds are
broken and two others are formed, resulting in a positive but
38
-1
small heat of reaction (∆H°298 ) 0.1 eV mol ). Since the
3
1
lowest excited triplet and singlet states of CO, Π and Π, are
1
+
39
located 6.0 and 8.1 eV above its Σ ground electronic state,
the energy available to the reaction products, Eavl ) hνexc -
U°298, is insufficient to produce electronically excited CO
∆
following excitation to S2 in benzaldehyde. For this reason the
Dissociation Kinetics. The temporal evolution of the pho-
todissociation of benzaldehyde following excitation to five single
vibronic levels of the S2(ππ*) r S0 band is summarized in
Figure 3. The left-hand column shows the benzene and
benzaldehyde ion currents as a function of the time delay
between pump and probe lasers. The ion signal observed for
m/e in the 102-108 amu range, represented by crosses in this
figure, is attributed to benzaldehyde parent ions or to fragments
electronic levels of CO are not included in the diagram. The
ground state of benzene is positioned 0.1 eV above S in
0
3
benzaldehyde, based on the reaction energy. As its T ( B ),
1
1u
3
1
T ( E ), and S ( B ) excited states are located 3.7, 4.5, and
2
1u
1
2u
36,40,41
4.7 eV above the ground state,
2 1
neither T nor S of benzene
is accessible following excitation to the lowest levels of S in
2
42
benzaldehyde. Ionization potentials of benzaldehyde, ben-
4
3
44
zene, and CO are 9.51, 9.25, and 14.01 eV, respectively.
Two-color ionization results are analyzed on the basis of the
premise that the F2 laser radiation ionizes only electronically
excited molecules. Ionization of benzaldehyde from the vibra-
tionless level of S0 requires the absorption of two photons of
the 7.9 eV radiation, and such a process is much less probable
than direct ionization and is not observed at the modest light
intensities of the present experiment. Benzaldehyde molecules
in vibrationally excited levels of S0, if produced by internal
conversion in the relaxation process, do not participate in the
ionization scheme due to unfavorable Franck-Condon factors
with the ionization continuum. Stepwise ionization of ground-
state benzene molecules by the probe laser should not be
appreciable since the 157 nm radiation of the F2 laser falls
+
of the type C6HxCO . The ion signal in the 74-80 amu range,
represented by empty circles, is attributed to benzene or its
+
fragment ions C6Hx . While the benzaldehyde ion signal decays
continuously with time, the benzene ion signal starts at zero,
increases to a maximum at some intermediate time, and then
decreases at long times. Note that at these five excitation
wavelengths neither benzaldehyde nor benzene ions are detected
when the probe laser is fired before the pump laser (t < 0).
The right-hand column of Figure 3 displays the same data
plotted in semilogarithmic coordinates. This representation is
particularly convenient as the two straight lines observed for
the benzaldehyde ion yield reflect the biexponential decay of
the excited states of benzaldehyde. The benzene ion signal, on
the other hand, decays according to a single exponential with a
rate that is independent of excitation energy. One aspect that
is clearly perceived at high excitation energies is that benzene
and benzaldehyde are observed to decay at approximately the
same rate in the long time range. Since the time scale of this
1
between the S3( E1u) and the Rydberg absorption regions of
4
5
benzene.
We have verified that irradiation of jet-cooled
benzene with the probe laser produces negligible amounts of
benzene ions. The ion signals displayed in Figure 3 therefore
reflect the temporal evolution of the electronically excited states

7938 J. Phys. Chem. A, Vol. 101, No. 43, 1997
Silva and Reilly
Figure 3. Effect of excess energy above the S
2
origin (35 191 cm-1) on the dissociation rate of benzaldehyde into benzene and CO. Left column:
benzaldehyde (×) and benzene (O) ion signals as a function of pump-probe delay time. Right column: logarithm of the benzaldehyde (×),
benzene (O), and compounded benzaldehyde and benzene ion signals as a function of pump-probe delay time. The solid lines derive from the
nonlinear least-squares fit described in the text.
of benzaldehyde and benzene. The excess energy available to
benzaldehyde and benzene ions produced with this experimental
The existence of two excited states that lead to ionization of
benzaldehyde is inferred from the biexponential decay of the
benzaldehyde ion signal. One of the states is short-lived (τ <
arrangement is large enough to ensure that large Franck-
Condon factors should be accessible4
6-48
and that ionization
10 s), and its lifetime depends strongly on the energy of the
-6
-
6
efficiencies for all intermediate states involved in the two-color
ionization scheme should be comparable.
pumped transition. The other state is long-lived (τ > 10 s),
but its decay rate does not reflect its lifetime since it is strongly

Jet-Cooled Benzaldehyde
J. Phys. Chem. A, Vol. 101, No. 43, 1997 7939
emission that has been observed following excitation to S2 under
2,6
low-pressure conditions. Under high-pressure conditions, the
origin of T(nπ*) is the end point of the relaxation cascade in
the electronically excited surface, as demonstrated by the strong
2
-8,22-24
emission observed from its vibrationless level.
The relaxation paths and rate constants considered in the
kinetic model are illustrated in Figure 4. The relaxation cascade
starts with excited benzaldehyde molecules leaving the optically
prepared S2 state with rate constant kST and populating the short-
†
lived predissociating state T . Energy redistribution following
the excitation step is expected to occur during the nanosecond
laser pulse since kST is estimated to be greater than 1012
s
-1
on
the basis of the lifetime of the S2 state mentioned above.
†
Relaxation proceeds from T through multiple paths: dissocia-
tion into benzene and CO, identified by the rate constant kd;
internal conversion to T(nπ*), represented by the rate constant
†
kT T; and relaxation to the ground electronic state, represented
†
by the rate constant kT 0. The latter relaxation path, which
includes radiative and nonradiative transitions to S0, is formally
considered for all states involved in the ionization scheme,
although it should be slower than other relaxation processes
under consideration. Assuming this rate constant to be on the
†
same order as the equivalent relaxation rate in T(nπ*), kT 0 is
3
4
-1
31
expected to have a value in the 10 -10 s range. The total
†
†
†
†
decay rate for state T is given by kT ) kd + kT T + kT 0.
Figure 4. Energy level diagram for the photodissociation of benzal-
The long-lived T(nπ*) state is supposed to be populated by
internal conversion from the predissociating state T and not
dehyde into benzene and CO showing the excited states involved, the
†
†
pump-probe scheme, and the important relaxation pathways. T is a
triplet state from which dissociation occurs. See text for rate constant
definitions.
by direct intersystem crossing from S2 at the beginning of the
relaxation cascade. (This supposition is based on the emission
behavior of benzaldehyde under high pressures that suggests
affected by sample translation. The benzene ion signal increases
at short times when the benzaldehyde ion signal is rapidly
decaying. Since the rate of benzene appearance and the rate of
benzaldehyde disappearance in the short time range depend
similarly on the vibronic level that is excited, the short-lived
state of benzaldehyde is identified as the predissociating state.
The decay of the benzene and benzaldehyde ion signals observed
at long times is attributed to translation of reactant and product
molecules out of the light interaction region. Since there are
no signs of further buildup of the benzene ion signal in the long
time range, we conclude that the long-lived state of benzalde-
hyde does not lead to benzene product formation.
2,6
the emitting state is populated by the predissociating state.
This is also consistent with the rapid decay of the dissociating
state of benzaldehyde evidenced in Figure 3.) Independent of
which state feeds the long-lived state, the equations to be derived
will have the same biexponential form. Decay routes considered
for the long-lived state of benzaldehyde include radiative
(
phosphorescence) and nonradiative (ISC) transitions to the
ground state with a combined rate constant kT0. No reactive
decay channels are expected for this state since the observed
channel of benzene formation does not show any signs of a
second entrance channel in the long time range.
Benzene molecules detected by two-color ionization in the
The kinetic model depicted in Figure 4 is similar to those
3
present experiment are assumed to be formed in T1( B1u) since
5,6
2
proposed by Baba and co-workers and by Berger et al. They
interpreted the pressure-dependent behavior of the phosphores-
cence emission of benzaldehyde in terms of a competition
between dissociation and collisional deactivation of a short-
lived predissociating state with consequent population of a long-
lived emitting state. As pointed out by them, the lifetimes of
these states are both too long for singlet states. Based on
radiative lifetime values determined from oscillator strength data
and on fluorescence quantum yields measured to be less than
the probe radiation is not able to ionize ground-state benzene
molecules and this is the only electronically excited state
accessible in benzene following excitation to low levels of the
S2 state of benzaldehyde. Possible relaxation routes from this
state include radiative and nonradiative transitions to S0
represented by the rate constant k′T0.
According to this kinetic model, the time dependence of the
population of the excited states that lead to benzaldehyde and
benzene ions is given by
-
4
1
0 , the lifetimes of S1 and S2 of benzaldehyde have been
-10
-12
2,6,49
estimated to be less than 4 × 10 and 10 s, respectively.
A preliminary inspection of Figure 3 indicates that the lifetime
of the short-lived state of benzaldehyde that we observe, on
the order of hundreds of nanoseconds, is much larger than either
of these numbers. This short-lived predissociating state must
therefore be a triplet state. From the evidence gathered so far,
it could either be a vibrationally excited level of T(nπ*) or one
of the T(ππ*) states of benzaldehyde. Because this could not
short-lived state:
†
d[T ]
†
) - k [T ]
T†
(1)
(2)
dt
long-lived state:
d[T(nπ*)]
†
be determined, the predissociating state is labeled T in Figure
4
, and its origin has been arbitrarily positioned in this diagram.
The long-lived state, on the other hand, is identified with T(nπ*)
since the latter state is the origin of the weak phosphorescence
†
)
k
T†T
[T ] - k [T(nπ*)]
T0
dt

7940 J. Phys. Chem. A, Vol. 101, No. 43, 1997
Silva and Reilly
benzene:
the data analysis treated translation as a first-order process with
an effective rate constant of ktr. The resulting detection
probability function decays exponentially with time according
3
d[ B ]
1u
†
3
)
k [T ] - k′ [ B ]
(3)
-k t
d
T0
1u
to P(t) ) e tr . This model preserved the functional forms of
dt
eqs 4 and 5 that predict the observed double-exponential decay
of the benzaldehyde signal and the single-exponential decay of
the benzene signal.
Solution of the coupled differential equations gives the following
expressions for the benzaldehyde (N106) and the benzene (N78)
excited-state populations:
The experimental benzaldehyde and benzene ion signals were
fitted to eqs 6 and 7, respectively, using a nonlinear least-squares
†
-kT†t
-kT0t
-kT†t
N106 ∝ [T ] + [T(nπ*)] ) A e
+ AIIe
≈ AIe
+
AII (4)
50
I
routine based on a modified Marquardt algorithm. From the
exponential factors in both equations the decay rate of the short-
lived state of benzaldehyde and the benzene product growth
rate were independently determined along with the preexpo-
nential factors. The set of parameters representing beam sizes
that resulted in the best fit were also assessed.
3
-k′T0t
- e-kT†t) ≈ B(1 - e-kT†t) (5)
N ∝ [ B ] ) B(e
7
8
1u
where AI, AII, and B are coefficients that depend on the
population of excited molecules at time t ) 0 and on the
individual rate constants kd, kT T, kT 0, kT0, and k′T0. Note that,
since kST is fast compared to the laser pulse duration, the
All models tested were able to provide good fits as long as
laser beam parameters were allowed to vary from experiment
to experiment. They all yielded rather similar values for the
decay rate of the short-lived state of benzaldehyde but not for
the benzene growth rate. While benzaldehyde parameters were
determined with an uncertainty of less than 20%, benzene
parameters turned out to be very sensitive to the shape of the
function P(t), and the resulting values were subject to an
uncertainty on the order of 30-50%. This is understood in
terms of the functional forms of eqs 6 and 7 as the benzene ion
signal is given by a difference of two terms, one representing
product appearance and the other product disappearance.
Changes in one term affect the value of the other term more
drastically than in the case of the benzaldehyde ion signal, in
which two terms representing disappearance rates are added
together. This effect is worse at low excitation energies since
appearance and disappearance rates differ only by 1 order of
magnitude under these conditions.
†
†
†
population of T at the end of the pump laser pulse is expected
to be equal to the total population of excited molecules at time
t ) 0. The population of all other states at this time is assumed
to be zero. The approximations indicated in eqs 4 and 5 are
valid if kT0 and k′T0 are at least 2 orders of magnitude smaller
†
than kT . Furthermore, with the latter rate constant being in
6
7
-1
the 10 -10
s
range and the rate of escape of excited
molecules out of the light interaction region being on the order
5
-1
of 10 s , any photophysical relaxation process occurring at a
5
-1
rate slower than 10 s would be masked by this sample
translation effect. According to the data presented in Figure 3,
the latter effect is indeed dominant since the benzaldehyde and
the benzene ion signals in the long time range appear to decay
at the same rate independent of excitation energy.
To establish the relation between the excited-state populations
(N106 and N78) and the observed ion yields (Y106 and Y78), we
The escape rate due to sample translation was determined to
multiplied the approximate forms of eqs 4 and 5 by two factors.
To account for possibly different absorption cross sections of
the excited species that participate in the ionization scheme,
the populations of T , T(nπ*), and B1u are multiplied by a factor
corresponding to the ionization efficiency of each specific
5
-1
be in the 2.5-5.0 × 10 s range using the ktr parameter and
5
-1
in the 1.5-3.0 × 10 s range with the a parameter. The
spread in the values of these parameters appeared to be due to
laser beam variations in different experiments since no wave-
length dependence was observed. The cross section of the probe
laser beam was not necessarily the same from run to run since
the apertures disposed along the beam path were moved in order
to maximize the overlap between the two light beams and, at
the same time, minimize the light scattered onto the walls of
the instrument. At higher excitation energies, however, all
models performed equally well and the escape parameters
displayed a smaller spread. This is because small variations in
beam size do not affect the shape of the detection probability
curve in the short time range as much as they do at longer times.
†
3
†
state: ηT , ηT, and η78, respectively. The other factor is a
detection probability function P(t) that is used to account for
sample translation. It represents the probability of finding an
excited molecule in the laser interaction volume at a given time
after the excitation step. The expressions thus obtained for the
observed ion yields are
AIe- + η A )
kT†t
(6)
(7)
Y106 ) P(t)(ηT
†
T
II
-
kT†t
Y ) P(t)η B(1 - e
)
7
8
78
Table 1 lists the decay rate of the short-lived predissociating
†
state of benzaldehyde (kT ) and the growth rate of the benzene
In treating the effect of sample translation, we examined a few
different forms of the detection probability function. A time-
delayed cross-correlation of the pump volume with the probe
volume should result in an adequate description of this function,
but it does require information concerning laser beam shapes.
Due to the uncertainty concerning the spatial profiles of the
light reported in the Experimental Section, cross-correlation
functions were calculated considering two possible types of
beam shapes, rectangular and Gaussian. Each type of cross-
correlation function was calculated over a range of possible
values. One function that proved to be very useful in the data
analysis was that resulting from two light beams with rectangular
profiles of equal sizes. In this case the detection probability
function is a simple analytic expression that depends on a single
parameter: P(t) ) 1 - at. Another simple function used in
†
product (k′T ) as a function of excess energy within the S2 band.
Values listed correspond to the average of the rate constants
obtained from the best fits at each excitation wavelength. As
seen in Figure 3, agreement between model and experiment is
indeed very good. The fitted curves displayed in that figure
are those obtained with the linear form of the detection
probability function. The measured decay rate of the predis-
sociating state shown in Table 1 is strongly dependent on
excitation frequency, and its value increases by an order of
-1
magnitude within the investigated range (∼1800 cm ). Figure
5 shows the dependence of this rate constant on excitation
energy. The increase is exponential, as indicated by the dotted
curve, and there are no indications of mode specific effects.
This behavior is consistent with that of a statistical reaction, as

Jet-Cooled Benzaldehyde
J. Phys. Chem. A, Vol. 101, No. 43, 1997 7941
TABLE 1: Rate Constants and Quantum Yields for the
Dissociation of Benzaldehyde as a Function of Excess Energy
energy distributions on the measured energy-dependent rate
constants kT (E). Under the jet-cooled conditions of our
†
in S
2
experiment, the energy distribution of the excited molecules is
much narrower than it would be for a room-temperature sample
so that incoherent averaging caused by simultaneous dissociation
of several dissociation channels is minimized. Yet another
difference concerning our results and theirs is related to the
behavior of the benzene signal in the long time range as a
function of excitation energy. They reported an increase in the
benzene decay rate with excess energy and suggested it might
be due to changes in the relaxation rate k′T0. However, as long
T† × 10
-
6a
k' †
T
× 10
-6b
excess energy
k
-
1
-1
-1
φBdisc
φAdisd
(cm
)
(s
)
(s )
0
1.2
2.5
3.0
4.0
4.9
6.4
7.8
10.3
12.0
1.1
1.6
2.8
3.0
3.5
6.0
7.8
9.5
12.5
0.72
0.71
0.69
0.60
0.76
0.68
0.80
0.79
0.72
1.00
0.99
0.96
0.98
0.98
0.96
0.98
0.95
0.97
4
5
8
9
00
50
00
55
200
359
502
759
1
1
1
1
5
-1
as this rate constant is significantly smaller than 10 s , any
changes in its value with excitation energy are not detected with
our experimental arrangement. It seems unlikely that it would
be observable with their apparatus.
a
Decay rate of the short-lived state of benzaldehyde. b Benzene
c
appearance rate. Dissociation quantum yield defined in terms of
benzene’s ion signal: φBdis ) (η106 /η78) × B /(A
Y
Y
+ AYII). Disso-
d
I
The benzaldehyde decay rate at the S2 origin reported here
can also be compared with those determined from quenching
ciation quantum yield defined in terms of benzaldehyde’s ion signal:
A
dis ) AY
Y
Y
φ
I
/(A
I
+ A II).
5
-1 6
6 -1 8
measurements at 284 nm: 3.2 × 10 s
and 1.14 × 10 s .
6 -1
The latter agrees well with our value of (1.2 ( 0.2) × 10 s .
The dissociation quantum yield was estimated using the
preexponential factors of eqs 6 and 7. According to the kinetic
model presented, the efficiency of formation of triplet benzene
†
can be written as kd/kT in terms of the rate constants or as B/(AI
+
AII) in terms of the population coefficients. If we consider
that there is only one reaction channel open, the one leading to
3
B1u benzene, then the efficiency of formation of triplet benzene
corresponds to the dissociation quantum yield, φdis. Further-
more, if dissociation into this channel and internal conversion
†
to T(nπ*) are the main relaxation routes for molecules in T ,
†
†
i.e. kT ≈ kd + kT T as previously assumed, then the dissociation
quantum yield can also be expressed as AI/(AI + AII). The
parameters experimentally assessed, the preexponential factors
Y
Y
Y
of eqs 6 and 7, A I, A II, and B , are related to the population
Y
†
Y
Y
coefficients above by A I ) AIηT , A II ) AIIηT and B ) Bη78.
Although ionization efficiencies have not been determined, it
seems reasonable to expect that T and T(nπ*) benzaldehyde
†
molecules will have similar ionization cross sections at 157
nm: the two states have the same multiplicity, they may even
have the same overall symmetry, and their origins are not far
apart. By assuming a single value for the ionization efficiency
of both states of benzaldehyde and naming it η106, the two
expressions for the dissociation quantum yields derived above
Figure 5. Dependence of the rate constant for the loss of benzaldehyde,
T†, on vibrational energy in S
. The dotted line corresponds to an
exponential fit of the decay rate kT†
k
2
.
complete randomization of the deposited energy takes place prior
to the dissociation step.
B
Y
Y
Y
A
can be written as φ dis ) (η106 /η78)B /(A I + A II) and φ dis )
Y
Y
Y
A I/(A I + A II).
According to the kinetic model presented the benzene product
is expected to appear at the same rate as the predissociating
The last two columns of Table 1 list our best estimates of
the dissociation quantum yield as a function of excess energy
according to the latter expressions. The values are subject to
†
state T decays. As seen in Table 1, six of the nine pairs of
rate constants are in good agreement. Due to the uncertainties
related to the benzene data fit mentioned above, it is unlikely
that differences between reactant disappearance and product
appearance rates shown in Table 1 represent anything more than
experimental error. To derive more accurate values for the
product appearance rate, the effect of molecular translation
should be eliminated. An experimental setup in which the pump
beam propagates coaxially with the molecular beam not only
should lead to more precise values of the benzene growth rate
but also should confirm the assumptions made here concerning
the slow photophysical relaxation rates of product and reactant.
Our results are found to be in rough agreement with those of
B
uncertainties on the order of 40-70% for φ dis and 20-30%
A
B
for φ dis. The fluctuations observed in the values of φ dis are
understood in terms of the large uncertainties involved in the
benzene data fit. Although uncertainties are large enough to
hide any wavelength dependence that might occur, two points
A
are noteworthy. First, φ dis is found to be greater than 0.9 at
all excitation energies. That indicates that deactivation of T†
through internal conversion to T(nπ*) occurs with an efficiency
†
of less than 0.1 if ionization cross sections of T and T(nπ*)
benzaldehyde are taken to be equal. Second, there is a
systematic difference between the values of the dissociation
quantum yield estimated using the benzene population coef-
1
7
†
Polevoy et al. Their reported value for the rate constant kT
6
-1
6
-1
B
at the S2 origin, 1.6 × 10 s , is similar to the 1.5 × 10 s
ficient (φ dis ) 0.6-0.8) and that estimated using only benzal-
A
value that we find when signal decay due to sample translation
dehyde coefficients (φ dis > 0.9) at each excitation energy.
†
is incorporated in the rate constant kT . Our results, however,
show a monotonic increase of the rate constant kT with excess
While this difference may be due to the large uncertainties
†
B
involved in deriving φ dis, we note that this result is consistent
energy, while theirs show strong fluctuations. This is under-
stood in terms of the effect of different thermal vibrational
with the observed behavior of the total ion yield: as displayed
in the right-hand column of Figure 3, the sum of the benzene

7942 J. Phys. Chem. A, Vol. 101, No. 43, 1997
Silva and Reilly
and benzaldehyde signals drops at a faster rate in the time range
that benzaldehyde dissociates. Altogether, these observations
indicate either that the ionization efficiency of benzene is indeed
different from that of benzaldehyde or that there are other routes
References and Notes
(
1) Silva, C. R.; Reilly, J. P. J. Phys. Chem. 1996, 100, 17111.
(2) Berger, M.; Goldblatt, I. L.; Steel, C. J. Am. Chem. Soc. 1973, 95,
717.
3) Bruhlmann, U.; Nonella, M.; Russegger, P.; Huber, J. R. Chem.
1
(
†
of relaxation available for molecules in T that lead to
Phys. 1983, 81, 439.
nonionizable species. The wavelength dependence of the
benzene ion signal shown in Figure 2 suggests that at least some
benzene molecules are formed in the ground electronic state
when benzaldehyde is irradiated at 258.9 nm, and it is reasonable
to expect this product entrance channel to be open at other
excitation wavelengths as well. Although we cannot be sure if
part of the signal recorded in this case is not due to a benzene
impurity in the benzaldehyde sample, this observation indicates
the need for further experiments capable of probing the
population of ground-state benzene molecules following excita-
tion of benzaldehyde in order to establish the benzene state
distribution as a function of excitation wavelength. Together
with the determination of excited-state ionization efficiencies
such measurements should provide further insight concerning
the product distribution in the photodissociation of benzalde-
hyde.
(4) Metcalfe, J.; Brown, R. G.; Phillips, D. J. Chem. Soc., Faraday
Trans. 2 1975, 71, 409.
(
(
5) Itoh, T.; Takemura, T.; Baba, H. Chem. Phys. Lett. 1976, 40, 481.
6) Itoh, T.; Baba, H.; Takemura, T. Bull. Chem. Soc. Jpn. 1978, 51,
2841.
(
(
(
7) Bruhlmann, U.; Huber, J. R. Chem. Phys. Lett. 1979, 66, 353.
8) Hirata, Y.; Lim, E. C. J. Chem. Phys. 1980, 72, 5505.
9) Antonov, V. S.; Knyazev, I. N.; Letokhov, V. S.; Matyuk, V. M.;
Movshev, V. G.; Potapov; V. K. Opt. Lett. 1978, 3, 37.
10) Antonov, V. S.; Letokhov, V. S.; Shibanov, A. N. Appl. Phys. 1980,
2, 293.
(
2
(11) Antonov, V. S.; Letokhov, V. S. Appl. Phys. 1981, 24, 89.
(12) Antonov, V. S. Opt. Spectrosc. (USSR) 1982, 52, 5.
(13) Long, S. R.; Meek, J. T.; Harrington, P. J.; Reilly, J. P. J. Chem.
Phys. 1983, 78, 3341.
14) Yang, J. J.; Gobeli, D. A.; Pandolfl, R. S.; El-Sayed, M. A. J. Phys.
(
Chem. 1983, 87, 2255.
(15) Yang, J. J.; Gobeli, D. A.; El-Sayed, M. A. J. Phys. Chem. 1985,
8
9, 3426.
16) Polevoy, A. V.; Matyuk, V. M.; Grigor′eva, G. A.; Potapov, V. K.
Khim. Vys. Energ. 1984, 18, 195.
17) Polevoy, A. V.; Matyuk, V. M.; Potapov, V. K. SoV. J. Chem. Phys.
Engl. Transl.) 1990, 6, 1157; Khim. Fiz. 1987, 6, 620.
(
(
(
(
(
(
18) Prileshajewa, N. Acta Physichochim. URSS 1935, 3, 195.
19) Sch u¨ ler, H.; Reinebeck, L. Z. Naturforsch. 1950, 5A, 448.
20) Asundi, R. K.; Battacharya, B.; Appalanarasimham, N. Curr. Sci.
Conclusions
A pump-probe mass spectrometry technique has been used
to investigate the photodissociation of benzaldehyde into
benzene and CO following excitation to the S2 state, revealing
a number of aspects concerning the kinetics and dynamics of
this dissociation process. From the observed biexponential
decay of the benzaldehyde signal it has been inferred that two
metastable states are populated in the relaxation process. The
time dependence of the reactant and benzene product signal
indicates that dissociation occurs from the short-lived state of
benzaldehyde on a time scale of hundreds of nanoseconds. The
dissociation rate increases with vibrational energy, and there is
no suggestion of mode selective behavior. According to the
analysis presented, benzene molecules are produced in its lowest
triplet state. The threshold for dissociation along this channel
is located somewhere below the origin of the S2 state of
benzaldehyde since it is possible to detect benzene molecules
following excitation to the S2 vibrationless level. The efficiency
of formation of triplet benzene molecules as a function of
excitation energy has been estimated, but the values obtained
depend on the unknown ionization efficiencies of the excited
species involved in the ionization scheme. The existence of a
secondary entrance channel leading to benzene molecules in
the ground electronic state has been suggested since enhance-
ment of the benzene signal at excitation frequencies correspond-
ing to allowed transitions of the S1r S0 spectrum of benzene
has been observed, but this product entrance channel needs to
be further investigated and characterized.
(India) 1952, 21, 273.
(
(
(
21) Robinson, G. W. J. Chem. Phys. 1954, 22, 1384.
22) Stockburger, M. Z. Phys. Chem. (Frankfurt) 1962, 31, 350.
23) Hollas, J. M.; Thakur, S. N. Chem. Phys. 1973, 1, 385.
(24) Koyanagi, M.; Goodman L. Chem. Phys. 1979, 39, 237.
(
(
(
25) Almasy, F. J. Chim. Phys. 1933, 30, 528, 634.
26) Garg, S. N. J. Sci. Res. Banaras Hindu UniV. 1951-52, 2, 153.
27) Imanishi, S.; Semba, K.; Ito, M.; Anno, T. Bull. Chem. Soc. Jpn.
1952, 25, 150.
(28) Abe, H.; Kamei, S.; Mikami, N.; Ito, M. Chem. Phys. Lett. 1984,
1
09, 217.
(29) Ohmori, N.; Suzuki, T.; Ito, M. J. Phys. Chem. 1988, 92, 1086.
(30) Villa, E.; Amirav, A.; Chen, W.; Lim, E. C. Chem. Phys. Lett. 1988,
47, 43.
1
(
(
31) Sneh, O.; Cheshnovsky, O. J. Phys. Chem. 1991, 95, 7154.
32) Colby, S. M.; Reilly, J. P. Int. J. Mass Spectrom. Ion Processes
1994, 131, 125.
(
(
(
33) Walsh, A. D. Trans. Faraday Soc. 1946, 42, 62.
34) Kimura, K.; Nagakura, S. Theor. Chim. Acta 1965, 3, 164.
35) Atkinson, G. H.; Parmenter, C. S. J. Mol. Spectrosc. 1978, 73, 20.
(36) Stephenson, T. A.; Radloff, P. L.; Rice, S. A. J. Chem. Phys. 1984,
1, 1060.
8
(
(
37) Inuzuka, K.; Yokota, T. Bull. Chem. Soc. Jpn. 1965, 38, 1055.
38) Pedley, J. B.; Naylor, R. D.; Kirby, S. P. Thermochemical Data of
Organic Compounds; Chapman and Hall: New York, 1986.
39) Herzberg, G. Molecular Spectra & Molecular Structure Vol. I:
Spectra of Diatomic Molecules; Kriger: Malabar, 1989.
40) Parmenter, C. S. AdV. Chem. Phys. 1972, 22, 365.
(41) Sharpe, S.; Johnson, P. J. Chem. Phys. 1984, 81, 4176.
42) Rosenstock, H. M.; Draxl, K.; Steiner, B. W.; Herron, J. T. J. Phys.
Chem. Ref. Data 1977, 6, Suppl. 1.
43) Grubb, S. G.; Whetten, R. L.; Albrecht, A. C.; Grant, E. R. Chem.
Phys. Lett. 1984, 108, 420.
44) Handbook of Chemistry and Physics; Lide, D. R., Ed.; CRC: New
York, 1996.
(
(
(
(
(
(
(
45) Hiraya, A.; Shobatake, K. J. Chem. Phys. 1991, 94, 7700.
46) Dietz, T. G.; Duncan, M. A.; Smalley, R. E. J. Chem. Phys. 1982,
7
6, 1227.
(
47) Johnson, P. M.; Otis, C. E. Annu. ReV. Phys. Chem. 1981, 32, 139.
Acknowledgment. C.R.S. thanks Brazilian CNPq for finan-
cial support and Universidade Federal do Rio de Janeiro for a
leave of absence. This work has been supported by the U.S.
National Science Foundation.
(
48) Otis, C. E.; Knee, J. L.; Johnson, P. M. J. Phys. Chem. 1983, 87,
2
232.
(49) Kanda, Y.; Kaseda, H.; Matumura, T. Spectrochim. Acta 1964, 20,
387.
50) Numerical Recipes in Pascal: The Art of Scientific Computing;
Cambridge University: Cambridge, 1989.
1
(
Registry No. Benzaldehyde, 100-52-7.