Three-body photodissociation of 1,3,5-triazine

Article abstract of DOI:10.1021/jp9609592

The three-body dissociation of 1,3,5-triazine (H₃C₃N₃ → 3HCN) has been studied by molecular beam photofragment translational spectroscopy following excitation at 308, 295, 285, 275, 248, and 193 nm. The analysis of the measured translational energy distributions of the HCN photofragments has shown that the dissociation mechanism is not, as previously suggested, a symmetric (or synchronous) three-body process but rather is a two-step concerted process, where concerted refers to correlation between the asymptotic velocity vectors of the three ejected HCN fragments. Furthermore, there is evidence that the initially excited electronic state (¹E″, ¹A₂) is deactivated by radiationless processes (internal conversion/intersystem crossing) to a lower electronic and finally dissociative state which is, according to symmetry correlation, the electronic ground state (electronic predissociation).

Full text of DOI:10.1021/jp9609592

J. Phys. Chem. 1996, 100, 13941-13949
13941
Three-Body Photodissociation of 1,3,5-Triazine
T. Gejo, J. A. Harrison,† and J. Robert Huber*
Physikalisch-Chemisches Institut der UniVersita¨t Zu¨rich, Winterthurerstrasse 190, CH-8057 Zu¨rich, Switzerland
ReceiVed: April 2, 1996; In Final Form: May 28, 1996X
The three-body dissociation of 1,3,5-triazine (H3C3N3 f 3HCN) has been studied by molecular beam
photofragment translational spectroscopy following excitation at 308, 295, 285, 275, 248, and 193 nm. The
analysis of the measured translational energy distributions of the HCN photofragments has shown that the
dissociation mechanism is not, as previously suggested, a symmetric (or synchronous) three-body process
but rather is a two-step concerted process, where concerted refers to correlation between the asymptotic velocity
vectors of the three ejected HCN fragments. Furthermore, there is evidence that the initially excited electronic
state (1E′′, 1A2) is deactivated by radiationless processes (internal conversion/intersystem crossing) to a lower
electronic and finally dissociative state which is, according to symmetry correlation, the electronic ground
state (electronic predissociation).
1. Introduction
symmetry would be retained during the dissociation, they found
that the HCN molecules each have an average kinetic energy
The three-body dissociation process of a polyatomic molecule
has attracted recent interest due to the mechanism which
involves a simultaneous breaking of multiple bonds.1-19 Fol-
lowing absorption of a photon, a variety of molecules have
shown a decay behavior which appears consistent with a
synchronous or concerted three-body dissociation. Of particular
interest in such studies is the energy disposal among the
emerging fragments, as this information reveals features of the
potential energy surfaces involved and the ensuing dynamics
on these surfaces during dissociation. Among the molecules
recently studied to gain insight in a three-body decay are acetyl
iodide,2 trifluoroacetyl iodide,4 1,3,5 - triazine,8 s-tetrazine,9
C3O2,10 acetone,13,14,17 CF2I2,15,16 and azomethane.18,19 In most
cases these systems have been investigated at only a single
photolysis wavelength, because of the difficult and lengthy
nature of experiments that may involve up to three polyatomic
fragments produced with varying amounts of internal excitation.
One technique that has found widespread application is photo-
fragment translational spectroscopy.20,21 This technique, though
not fragment quantum state specific, provides sufficient versatil-
ity to allow examination of the recoil energy of all the fragments
produced with reasonable resolution. Most of the molecules
mentioned above have been studied by this method with 1,3,5-
triazine (s-triazine or H3C3N3) showing particularly intriguing
features.
of 10 kcal/mol at 248 nm but only 2 kcal/mol at 193 nm. Not
only was the average translational energy much lower at the
shorter wavelength but the translational energy distribution was
also much narrower, indicating a considerable difference in the
dissociation dynamics. Two different electronic dissociation
channels cannot account for this feature since symmetry
correlation shows that s-triazine must dissociate on the lowest
potential energy surface in order to produce three ground state
HCN molecules.8 At these wavelengths it is energetically
impossible to populate the A electronic state of HCN. This
intriguing result was rationalized as originating from a non-
statistical distribution on the ground state due to rapid dissocia-
tion caused by the large amount of available energy. In an
accompanying paper, Goates, Chu, and Flynn24 reported that
HCN produced from the 193 nm photolysis of s-triazine has a
large amount of excitation in the bending mode. This finding
appeared to be consistent with that of Ondrey and Bersohn
because, on the basis of their idea, the difference of available
energy is accounted for by the internal energy of HCN, more
specifically, the increase of available energy allows the dis-
sociation to proceed more rapidly since the HCN molecules do
not have to relax as much from their initial bent geometry
following photolysis at 193 nm compared to photolysis at 248
nm. The result of an ab initio calculation25 also supports Ondrey
and Bersohn’s conclusion that all three HCN molecules possess
equivalent translational energies as a consequence of the 3-fold
symmetry of the predicted transition state.
It is well established that s-triazine yields three HCN
fragments following photolysis:8
As also observed,8 the comparative result between these two
photolysis wavelengths appears to show a remarkable and
unexpected behavior for the dissociation dynamics. If, indeed,
a rapid dissociation causes a less statistical translational energy
distribution in the HCN products, a decrease in the average
translational energy at other, intermediate photolysis wave-
lengths should also be observed. The question of whether there
is a gradual change in dynamics as the internal energy is
increased or whether the change is due to a “memory” of the
initial excited state which can be suddenly switched on upon
accessing certain quantum states is, therefore, an open one at
present. There is also some doubt about the symmetric three-
body nature of the dissociation of s-triazine; apparently, the
photon energy at 193 nm (148 kcal/mol) is not sufficient to
break the three CN bonds simultaneously without a concomitant
N
N
+ hν
3HCN
(1)
N
The HCN dimer is not expected to have a significant binding
energy,22 and it thus appears that it will not play an important
role in the dissociation. Using photofragment translational
spectroscopy Ondrey and Bersohn8 measured the translational
energies of the HCN fragments produced at an excitation
wavelength of 248 and 193 nm, which corresponds to a nπ*
type transition (from 230 to 320 nm) and to ππ* and Rydberg
transitions below 230 nm.23 By arguing that the initial 3-fold
† Permanent address: Department of Chemistry, Massey University,
Albany, Private Bag 102-904, NSMC, Auckland, New Zealand.
X Abstract published in AdVance ACS Abstracts, July 15, 1996.
S0022-3654(96)00959-8 CCC: $12.00 © 1996 American Chemical Society

13942 J. Phys. Chem., Vol. 100, No. 33, 1996
Gejo et al.
change of geometry from the 120° HCN angle that exists in
s-triazine.24,26,27 Due to the amount of relaxation of energy and
geometry required before fragmentation can occur, one might
reasonably expect that not all bonds will be broken simulta-
neously.
The present work deals with a reinvestigation of the photo-
dissociation of s-triazine by photofragment translational spec-
troscopy at a variety of photolysis wavelengths. The use of an
excimer laser and a dye laser pumped by a Nd:YAG laser has
allowed us to measure the photofragment translational energy
distributions over photolysis wavelengths ranging from 193-
308 nm, which corresponds to a range in available fragment
energy of between 105 and 53 kcal/mol. From the high-
resolution TOF spectra obtained with our apparatus the dis-
sociation of s-triazine is clearly shown to occur via a nonsym-
metric route. On the basis of these results, we have simulated
successfully this three-body dissociation by consideration of the
expected angular distribution of photofragment recoil vectors.
The simulations also provide information on the translational
energy distributions due to primary and secondary dissociation.
2. Experimental Section
The experiments were carried out with the high-resolution
PTS apparatus described previously.28 A gas mixture was
prepared by flowing helium carrier at a stagnation pressure of
500 mbar over a solid sample of s-triazine cooled to 20 °C. A
pulsed molecular beam of the gas mixture was generated either
by means of a piezoelectrically driven pulsed valve or by a
similar source with an additional heating element.29 To avoid
dimers, the photolysis laser pulse was timed to coincide with
the leading edge of the beam pulse where monomer species
dominate. The velocity distribution of s-triazine was measured
by the hole-burning technique,30 and was found to have a most
probable velocity of 1000 and 1230 m/s and a beam translational
temperature of 20 and 3 K with and without the heating element,
respectively.
Time-of-flight (TOF) distributions of the photofragments were
measured at several scattering angles Θ, where Θ is the angle
between the molecular beam and the laboratory-fixed direction
of the detection axis. The 193, 248, and 308 nm laser pulses
were generated by an ArF, KrF, or XeCl excimer laser (Lambda
Physik EMG 101 MSC), respectively. The 275, 285, and 295
nm laser pulses were obtained by frequency-doubling the output
of a pulsed dye laser (Lambda Physik, SCANMATE 2E), which
was pumped by the frequency-doubled output of a Nd:YAG
laser (Continuum, Powerlite 7020). Typical laser fluences were
400 mJ/cm2 at 193 nm, 500 mJ/cm2 at 248 nm, 80 mJ/cm2 at
275, 285, and 295 nm, and 2.0 J/cm2 at 308 nm. After a flight
path of 34.5 cm the photofragments were ionized by electron
bombardment, discriminated according to their mass-to-charge
ratio (m/e) within a quadrupole mass filter, and detected by an
electron multiplier. A time offset of 3.9(m/e)1/2 ms was
subtracted from all of the TOF spectra before analysis in order
to correct for the ion flight time through the quadrupole mass
spectrometer.
Figure 1. (a) TOF distribution of HCN fragment from photodisso-
ciation of s-triazine at 248 nm. The circles represent the experimental
data, the solid line is the best fit obtained from the translational energy
distribution shown in (b). (b) Center-of-mass translational energy
distribution P(ET) of the HCN fragment.
3. Results
3.1. Excitation at 248 nm. Photofragment signals were
observed with the mass filter set at m/e ) 27 (HCN+) while no
signal was detected for the mass to charge ratio, m/e ) 54
(H2C2N2+). The unpolarized TOF spectra for m/e ) 27 at Θ
) 18° are displayed in Figure 1a. The solid line corresponds
to the calculated best fit distribution, which was obtained by a
forward convolution procedure31 of the center-of-mass (CM)
translational energy distribution P(ET) for the HCN fragment,
shown in Figure 1(b). This P(ET) refers to the translational
energy of each HCN molecule as measured directly from the
time-of-flight spectrum. The overall P(ET) for the dissociation
process also includes the kinetic energy of the partner frag-
ment(s).21,31 Owing to the complex nature of the dissociation
that may include three fragments, it is not possible to unambigu-
ously decide a priori what the overall P(ET) will be, as this
will depend upon the recoil velocity vector distribution for the
other fragments. This point and the model eventually used to
fit the distribution are addressed in the Discussion section. It
should be noted that although the total P(ET) distribution does
depend upon the particular model used, the average P(ET) must
To measure the recoil anisotropy of the photofragment flux,
a 248 nm laser beam was polarized to a degree of 91 ( 1%
using a stack of 10 quartz plates placed at Brewster’s angle to
the laser beam. Polarized TOF spectra were taken at four
polarization angles ꢀ, where the latter defines the angle between
the electric vector of the polarized laser beam and the detection
axis. The polarization angle was rotated by means of a half-
wave plate placed after the quartz plates stack.

Three-Body Photodissociation of 1,3,5-Triazine
J. Phys. Chem., Vol. 100, No. 33, 1996 13943
Figure 3. TOF distributions of the HCN fragments from 248 nm
photodissociation of s-triazine using three different laser fluences as
indicated. The solid line is the forward convolution of P(ET) given in
Figure 1b.
Figure 2. TOF distribution of HCN fragments from photodissociation
of s-triazine at 248 nm. As in Figure 1a but with the molecular beam
generated by the heated source.
be the same for any model and will correspond to the average
value observed in the TOF spectra.
Figure 1a shows the TOF signal which gradually decreases
around 200 µs. To confirm that the effect of clusters in this
region can be ignored, photofragment signals were measured
and the P(ET) distribution was determined through the use of
the heated nozzle (T ∼ 200 °C) pulsed beam source.29 The
result is depicted in Figure 2. Although there is an apparent
difference between the two distributions shown in Figures 1a
and 2, these differences can be entirely attributed to slight
differences in the beam conditions. Once these are accounted
for, the two time-of-flight spectra can be fitted using the same
P(ET), which proves the contribution from cluster photolysis
to the m/e ) 27 signal to be negligible.
The threshold (or upper limit) of the translational energy for
the fastest fragments was determined to be 48 kcal/mol as
indicated by the P(ET) of Figure 1b. To ensure that multiphoton
absorption does not affect the threshold, TOF distributions were
measured with laser fluences ranging between 270 and 1100
mJ/cm2 at Θ ) 45°. The result is shown in Figure 3. Though
there is clearly an increase in signal corresponding to the fastest
HCN molecules at the highest laser fluence used, the shapes of
the photofragment signals remained constant below a laser
fluence of 560 mJ/cm2. We conclude that two-photon processes
are not important below a fluence of 500 mJ/cm2 and hence
that the one-photon process has a threshold translational energy
of ∼48 kcal/mol.
Figure 4. Polarized TOF distribution of HCN fragments of s-triazine
at 248 nm and θ ) 45°. The two set of data were taken with polarization
angles of ꢀ ) -20° and +70°.
basis of the formation enthalpies of s-triazine and HCN, the
cleavage of all three bonds requires a dissociation energy D0 )
39.6 kcal/mol.24,26,27 Consequently, the available energy is 75.7
kcal/mol following excitation at 248 nm (115.3 kcal/mol). The
P(ET) distribution of the detected HCN fragments at 248 nm
(Figure 1b) shows an average value of 17.4 kcal/mol, which
results in an average total translational energy release of 52.2
kcal/mol. This would correspond to 69% of the available energy
and implies that the three HCN fragments are formed with a
remarkably high translational energy and with only 7.8 kcal/
mol of internal energy in each HCN molecule. The observed
P(ET) with the threshold translational energy of 48 kcal/mol
per HCN molecule is unequivocal evidence that the 3-fold
symmetry during dissociation is not preserved for a substantial
number of the dissociating molecules because their sum would
simply exceed the available energy (at least all those with ET
> 25 kcal/mol, see Figure 1b).
Figure 4 shows the polarized TOF spectra for m/e ) 27 taken
at Θ ) 18° with ꢀ ) -20° and 70°. Four independent
measurements (ꢀ ) -20°, 70°, -70°, and 20°) were carried
out, but no significant polarization effect was observed. Thus,
the fragment anisotropy parameter â32 must have a value close
to zero. Because the peak should consist of three types of HCN
fragment (corresponding to the three recoil velocity vectors), it
is not possible to determine a small â value precisely.
Nevertheless, in view of the observation of s-triazine fluores-
cence around 310 nm,33 the measured value of â ≈ 0 is
presumably due to a relatively long lifetime of the initially
excited state of s-triazine.
The available energy is given as EaVl ) hν - D0 + Eint, where
hν is the photon energy, D0 is the bond dissociation energy,
and Eint is the internal energy of the parent molecule.21 The
latter can be neglected for supersonic jet expansion. On the
Finally, it is important to note that the TOF spectrum was
also recorded under scattering angles θ ) 12°, 30°, and 60°. In
all cases only a single peak was obtained. In comparison with

13944 J. Phys. Chem., Vol. 100, No. 33, 1996
Gejo et al.
Figure 5. TOF distribution of HCN fragments from photolysis of
s-triazine at (a) 275 and (b) 285 nm. The circles represent the
experimental data. The solid lines are obtained from the P(ET)
distributions shown in Figure 6.
Figure 7. (a) TOF distribution of HCN fragments from photolysis at
193 nm. The circles represent the experimental data, and the solid line
is the best fit obtained from P(ET) shown in Figure 6. (b) P(ET)
distribution calculated from the TOF spectrum at 193 nm.
available energy. The average translational energy ET obtained
from the P(ET) distributions given in Figure 6 is 15.2 kcal/mol
at 275 nm, 14.7 kcal/mol at 285 nm, and 13.8 kcal/mol at 295
nm. These average translational energies correspond to about
70% of Eavl (Table 1).
Figure 6. Best-fitting translational energy distributions P(ET) of HCN
from from photolysis at 193, 248, 275, 285, and 295 nm. P(ET) is shown
as a function of HCN translational energy in the center-of-mass system
of s-triazine and hence does not include any contribution from the
partner fragments.
3.3. Excitation at 193 nm. Figure 7a shows the photo-
fragment signal at m/e ) 27 from photolysis at 193 nm (148
kcal/mol) using the heated source. Only a very weak signal
was observed when using the source without the heating
element. The room temperature absorption spectrum of s-
triazine shows a broad absorption band with an absorption
coefficient of ∼100 L/(mol‚cm) at 193 nm. The strongly
reduced signal for the unheated source is probably due to a small
absorption coefficient at 193 nm for jet-cooled s-triazine. The
P(ET) distribution calculated from the TOF spectrum at 193 nm
is displayed in Figure 7b and reveals an average translational
energy of the HCN fragments of 19.8 kcal/mol. Despite the
fact that there is only a small difference of 2.4 kcal/mol in the
average translational energy of the HCN photofragments
between photolysis at 193 and 248 nm, the shape of the
distribution is noticeably wider at 193 nm while the threshold
of translational energy remains at ∼48 kcal/mol. The average
the spectrum in Figure 1a, the peak was slightly broader at 60°
and slightly narrower at 12°.
3.2. Excitation at 275, 285, and 295 nm. We measured the
photofragment TOF spectra with mass setting at m/e ) 27
(HCN+) and Θ ) 18° for the photolysis at 275, 285, and 295
nm. Figure 5, a and b, displays the spectra taken at 275 and
285 nm, respectively. The solid lines represent the best fit from
the forward convolution with the P(ET) distributions shown in
Figure 6. It should be noted that the laser light was polarized
with ꢀ ) 0° although the measured anisotropy â = 0 at 248 nm
suggests that this will not affect the TOF spectra. The much
weaker laser power at these excitation wavelengths accounts
for the lower signal-to-noise ratio as compared to the 248 nm
photolysis TOF spectrum. The shape of the P(ET) is not
significantly different from that obtained at 248 nm (Figure 1b),
though the position is slightly changed due to the change in

Three-Body Photodissociation of 1,3,5-Triazine
J. Phys. Chem., Vol. 100, No. 33, 1996 13945
translational energy of 3‚19.8 kcal/mol corresponds to ∼55%
of Eavl
.
Using the heated source, we expected to obtain a better
Franck-Condon overlap for absorption without affecting the
dynamics. In the present case symmetry correlation indicates
that s-triazine dissociates on the lowest potential energy surface.8
This dissociation therefore includes electronic relaxation as a
necessary step, which suggests that absorption involving dif-
ferent Franck-Condon factors should not significantly influence
the dynamics (vide infra).
3.4. Excitation at 308 nm. Figure 8 shows the photofrag-
ment signal for m/e ) 27 detected at Θ ) 45°. In view of the
similarity to the P(ET) measured at 193 nm, the photofragments
were produced not only by a one-photon but also by a two-
photon process. The threshold translational energy of ∼50 kcal/
mol also supports involvement of a two-photon process, because
the high translational energy exceeds the limits imposed by
conservation of momentum for a one-photon process. Although
the energy dependence was measured over the range from 1000
to 2000 mJ/cm2, the integrated TOF signal is not proportional
to the square of the laser power, indicating that the TOF
spectrum contains more than two components. To clarify this
point, we examined the possibility of fitting the TOF signal
with three energy distributions, two of which are the same as
the P(ET) distributions obtained at 295 and 193 nm and the third
one being a low translational energy P(ET). The solid line
through the data points in Figure 8 is the result of the fit using
these three distributions. The first two P(ET) distributions
correspond to one- and two-photon excitations while the third
P(ET), which is used to fit the low-energy region of the TOF
spectrum, is presumably due to the contribution of cluster
photolysis considering the low signal intensity. Such a situation
would arise if the UV absorption is shifted to longer wavelength
upon cluster formation, even though the photolysis laser pulse
was timed to coincide with the leading edge of the beam pulse.
Figure 8. TOF distribution of HCN fragments from photolysis at 308
nm. The over all solid line is the fit consisting of three translational
energy distributions (see text).
4. Discussion
4.1. General Aspect. There are two likely mechanisms for
the photodissociation of s-triazine. The first one involves a
symmetric single step (eq 1) while the second one proceeds
according to a sequential process
Figure 9. Vector diagrams for the three cases of photodissociation
processes considered in this work. (a) a simultaneous symmetric three-
body dissociation, with all three fragments having identical kinetic
energies; (b) three-body dissociation in which the two HCN fragments
generated in the secondary dissociation are ejected at the same angle
to the initial primary dissociation recoil vector; (c) concerted (but
asymmetric) three-body dissociation with one fragment ejected forward
with respect to the recoil direction of the primary H2C2N2 fragment.
H3C3N3 + hν f HCN + H2C2N2
H2C2N2 f 2HCN
(2a)
(2b)
long as the intermediate lifetime is shorter than a rotational
period (vide infra). In the limiting case the H2C2N2 moiety will
not have had time to rotate, so that the two HCN frag-
ments from step 2b will be ejected with the same speed. Since
in this scenario the initial symmetry is essentially preserved,
the decay is referred to as a symmetric sequential dissociation
shown schematically in Figure 9. The symmetric three-body
dissociation is obviously a limiting case of this latter process.
If, however, the lifetime of H2C2N2 is sufficiently long so that
this fragment can undergo rotation, the velocity vectors of the
two HCN fragment will generally have different magnitudes;
this process corresponds to an asymmetric, sequential dissocia-
tion.
4.2. Modeling. We started the modeling of the observed
time-of-flight spectra by considering the existence of a long-
lived H2C2N2 intermediate surviving for several rotational
periods before breaking apart and thus giving rise to a sequential
or stepwise mechanism. To this end we have carried out TOF
simulations based on a forward convolution procedure of the
primary and secondary photofragments.3,34-36 For the sequential
In the case of reaction 1 the three HCN fragments will have
identical kinetic energies in the center-of-mass frame of the
s-triazine molecule, and thus we refer to this pathway as the
symmetric three-body (synchronous) dissociation channel (Fig-
ure 9). The TOF spectrum will consist of a single peak, and
the high-energy threshold of the HCN translational energy
distribution will be at most equal to one-third of the available
energy. The measured TOF spectra and the corresponding P(ET)
(see Figures 1 and 7) show clear evidence that a substantial
number of the fragments are ejected with ET exceeding this
threshold. Therefore, reaction 1 cannot be the dominant
dissociation pathway that is operative following excitation
between 193 and 248 nm. We are forced to consider reaction
2 with the intermediate H2C2N2 which undergoes a secondary
dissociation to form 2 HCN molecules. The second step 2b
may occur on a similar time scale as the initial fragmentation
step 2a or on a much longer time scale.
If the processes 2a and 2b occur on similar time scales, the
dissociation may be considered to be a concerted process as

13946 J. Phys. Chem., Vol. 100, No. 33, 1996
Gejo et al.
dissociation processes the angular HCN distribution is symmetric
with respect to forward-backward scattering about the center-
of-mass frame of the s-triazine parent with the fragment flux
peaking at an angle to the primary recoil direction.34,37,38 The
maximum HCN translational energy threshold corresponding
to all of the available energy being released in the primary
fragmentation step would result in two-thirds of Eavl, which is
∼50 kcal/mol for the photolysis at 248 nm. The very low signal
observed around this energy (Figure 1b) originates from
fragments consistent with this extreme limiting case and clearly
indicates that most dissociation events occur with a smaller
translational energy and hence with a significant amount of
internal excitation in the HCN fragments and probably also in
the H2C2N2 fragment. However, the TOF spectrum based on a
long-lived intermediate should exhibit a bimodal TOF distribu-
tion; the first peak due to the primary dissociation and the
second, slower peak to the two HCN fragments formed in the
secondary dissociation step. In view of the small binding energy
of the HCN dimer of a few kcal/mol,22 the intermediate H2C2N2
may have an internal energy of a few kcal/mol at most, if it is
to live for a significant length of time. The low binding energy
(the shallow potential energy) limits the amount of energy that
may appear as translation in the secondary fragments. Our
simulations showed for all cases involving a long-lived H2C2N2
intermediate with reasonable HCN-HCN binding energies, two
pronounced peaks in the modeled spectra. Thus, the single peak
observed in the experimental HCN TOF spectrum rules out such
a mechanism and indicates that a more concerted dissociation
is operative.
Figure 10. Angular distribution of the secondary photofragments. (a)
angular distribution based on a Lorentzian at a center of 90°; (b) W(δ)
according to eq 3; (c) angular distribution obtained by superimposing
(a) and (b).
lived case above, we satisfied this requirement by using the
expression
Having ruled out the symmetric three-body dissociation and
the sequential or stepwise decay with a long-lived intermediate
as dominant pathways, we attempted to model the single
observed TOF peak by allowing (i) the two HCN fragments
from dissociation of H2C2N2 to have a significant amount of
translational energy and (ii) the H2C2N2 intermediate to have a
finite lifetime for partial rotation. In general, when treating
the sequence of an ABC-type reaction,34-36 one analyzes the
TOF distribution of C in order to determine the recoil distribu-
tion of the primary photofragment pair AB and C. With this
information, one simulates the TOF distributions of the second-
ary fragments A and B with the final recoil velocity vectors. In
the case of s-triazine, however, the dissociation yields three
indistinguishable HCN fragments which contribute to the TOF
spectrum. Therefore, we simulated a sequential dissociation
mechanism in the following manner. As a first step, the P(ET)
of the primary dissociation was determined from the central
region of the TOF spectrum (for example the region of 100-
140 µs in Figure 11a). Secondly, the whole TOF spectrum was
subsequently calculated using a variety of assumed P(ET) and
angular distributions for the secondary dissociation step (see
Figure 10). The secondary P(ET) and the angular distribution
were varied until an acceptable fit had been obtained.
W(δ) ) c sin δ*/sin δ
(3)
for δ ∈ [δ*,180°-δ*] and W(δ) ) c otherwise. This distribu-
tion is shown in Figure 10b for δ* ) 20°. The final angular
probability distribution shown in Figure 10c arises from a
superposition of these two functions depicted in Figures 10a,b.
A reasonable fit for the data at 248 nm was obtained using the
P(ET) distributions for the primary and secondary dissociation
shown in Figure 11b and the angular distribution with a center
of 90° shown in Figure 10c.
Figure 11a depicts the TOF distribution of the HCN fragments
from the 248 nm photodissociation and the partitioning into the
two components. The component of the primary HCN frag-
ments corresponds to the narrower peak and the broader peak
to the HCN molecules from the secondary dissociation. The
primary and secondary P(ET) distributions have similar mean
values and comparable widths (Figure 11b). As is evident from
Figure 11a, the fastest fragments arise from the dissociation of
the H2C2N2 due to scattering from secondary dissociation in
the opposite direction relative to the velocity vector of the
primary HCN fragment. On the other hand, the slowest
fragments are those ejected from the secondary dissociation
along the direction of the velocity vector of the primary HCN.
A dissociation geometry corresponding to a case close to this
situation is drawn in Figure 9.
On the basis of this analysis an interpretation of the decay
mechanism rests on a sequential breaking of bonds. Since
s-triazine is a cyclic compound, three bonds are eventually
broken. Initially two bond fissions are required to release the
primary HCN fragment, and subsequently a third fission is
necessary during the secondary dissociation of the H2C2N2
fragment. The latter decay involves a range of secondary
dissociation recoil vectors with a scattering angle distribution
strongly peaked at δ ) 90° indicating that the majority of
fragmentations proceed with a small amount of rotation,
The analysis of a concerted three-body dissociation developed
by Zhao et al. for s-tetrazine9 used a Gaussian function to
express the spread about the asymptotic velocity vectors of
secondary photofragments. We explored fitting the data with
this function, but the result was unsatisfactory; the distribution
required a function more strongly peaked at 90° to the primary
recoil vector. Therefore, fitting was attempted by a Lorentzian
angular distribution function shown in Figure 10a. The scat-
tering angle δ is the angle between the recoil vector of the
secondary dissociation and the direction of the primary dis-
sociation. Since the final angular distribution should also
include the process with the strong back and forward peak
resulting from symmetry reasons34,37,38 as explained for the long-

Three-Body Photodissociation of 1,3,5-Triazine
J. Phys. Chem., Vol. 100, No. 33, 1996 13947
Figure 12. TOF distributions of the HCN fragments from photolysis
at 193, 275, and 285 nm. The solid lines are calculated TOF
1
2
distributions obtained from P(ET ) and P(ET ) shown in the insets for
a concerted process.
TABLE 1: Partitioning of the Available Energy (Eavl) into
the Translational (ET) and Internal (Eint) Energy of the HCN
Fragments upon s-Triazine Photolysis at Various Excitation
Wavelengthsa
1
b
2 b
λ (nm)
Eavl
ET
Eint
ET /Eavl (%)
ET
ET
193
248
275
285
295
109
76
64
60
57
59
52
46
44
41
50
24
18
16
16
55
69
71
73
72
33
27
24
24
26
25
22
20
Figure 11. (a) TOF distribution of HCN fragments from photolysis
at 248 nm. The circles represent the experimental data, and the solid
lines are calculated TOF distributions obtained from the P(ET)s in (b)
assuming a concerted process with W(δ) shown in Figure 10c. (b)
a All energies are in kcal/mol units. b From P(ET) partitioning.
1
2
Primary and secondary distributions, P(ET ) and P(ET ), of HCN.
primary and secondary dissociation shown in Figures 11 and
12. After photon excitation between 295 and 248 nm ET ,
which coincides approximately with the maximum of the P(ET)
curve in Figure 6, is gradually shifted to slightly higher values
and accompanied by a small broadening of P(ET). In all cases
ET represents about 70% of Eavl. A more pronounced change
is evident if excitation occurs at 193 nm where ET has
decreased to ∼55% of Eavl and P(ET) is considerably broadened
compared to P(ET) at 248 nm. The behavior of the P(ET)
distributions in Figure 6 is an indication that both dissociation
steps are preceded by substantial intramolecular vibrational
redistribution (IVR). As has been pointed out by Lee and co-
workers,39-42 the translational energy release is, under these
conditions, mainly correlated with the exit barrier along the
reaction coordinate and is less influenced by the excess energy,
and hence the excitation energy, when an appreciable exit barrier
is involved. The P(ET) curves for different excitation energies
(but for the same transition state) will therefore merely show a
broader distribution at higher energy and a peak position change
much smaller than the increase in the excess energy. This
parallels our present observation. In recent three-body dis-
sociation studies of azomethane,18,19 CH3-N)N-CH3 f CH3
+ N2 + CH3, and acetone,14 CH3-CO-CH3 f CH3 + CO +
noticeably less than a full rotational period, of the H2C2N2
intermediate. According to the definition given by Strauss and
Houston,10 we may still regard the s-triazine photodissociation
as concerted because it was found necessary to break all three
bonds on similar time scales, i.e., less than a rotational period
of the intermediate. Only under these conditions will a single
TOF peak be observed.
The simulations for the photolysis at 193, 275, and 285 nm
are displayed in Figure 12. In each case the P(ET) curves found
for the primary and the secondary dissociation are similar in
form to those at 248 nm, and also the angular distribution for
the photofragments was comparable to that obtained from the
fit of the 248 nm data.
4.3. Energy of the Photofragments. The average transla-
tional energy ET is defined in terms of the center-of-mass
system of the s-triazine parent molecule and is equal to the sum
of the average translational energies released in the primary and
1
2
secondary dissociation step, ET + ET . The TOF measure-
ments of the HCN fragment yielded the translational energy
distributions P(ET) collected in Figure 6 and the ET values
1
2
listed in Table 1 where we also included ET and ET obtained
1
2
from partitioning of P(ET) into P(ET ) and P(ET ) from the

13948 J. Phys. Chem., Vol. 100, No. 33, 1996
Gejo et al.
CH3, similar P(ET) features were observed. Thus, the P(ET)
distribution of azomethane photofragments measured at 193
nm19 was found to be similar to that at 351 nm, reported
previously,18 and acetone excited at 193 and 248 nm produced
similar photofragment distributions at the two excitation ener-
gies.14
the CH3 photofragment showed two separate peaks due to CH3
emerging from the primary and secondary dissociation step.
5. Conclusions
Following excitation at 308-193 nm s-triazine dissociates
exclusively into three HCN fragments. At the six different
excitation wavelengths used, only one single peak due to the
HCN photofragment has been observed in the TOF spectrum,
the analysis of which revealed the dissociation not to be a
symmetric three-body process with a simultaneous breaking of
the three bonds, but a sequential process with a short-lived
intermediate H2C2N2. The results of our photofragment distri-
bution measurements and our conclusions thus contrast those
of Ondrey and Bersohn8 who reported an anomalous behavior
of P(ET), having a lower ET at 193 nm than at 248 nm, and
symmetric bond breaking. Furthermore, if we consider a
rotational period or less of the intermediate, and in turn a
correlation between the asymptotic velocity vectors of the three
HCN photofragments, as a criterion of concertedness, the present
dissociation process proceeds in two steps but is concerted. The
features of the absorption spectrum in the excitation range 308-
248 nm as well as those of the corresponding photofragment
distributions P(ET) point to electronic predissociation. Hence,
the initially excited state in s-triazine is in a first step
electronically relaxed to a lower electronic state. The finally
dissociative state is, according to symmetry correlation with
three ground state HCN molecules, the electronic ground state
S0 (1A1) of s-triazine.
The energy channeled into the internal degrees of freedom
of the HCN fragment (see Table 1) is only ∼30% of Eavl, except
for 193 nm where it is ∼45%. Under the latter conditions a
substantial amount of Eavl is deposited into vibrational motion
(much less into rotation), as observed by Goates et al.24 who
performed IR emission measurements on the HCN photo-
products. Moreover, they found that the number of quanta in
the bending mode was very much greater than those in the CH
stretching mode, indicating that the time for complete IVR
relaxation to a statistical distribution is not sufficient in the
dissociation process at 193 nm.
4.4. Mechanism. The absorption band of s-triazine in the
excitation range 308-248 nm consists of two transitions:23,43
1
1
1
1
1A1 f E′′(nπ*) and A1 f A2 (ππ*). The excited state E′′
is Jahn-Teller distorted and mixed via pseudo-Jahn-Teller
1
interaction with the higher lying A1 state.43 The absorption
band is highly structured, and below 300 nm fluorescence is
observed.33 It appears, therefore, that these two states are not
directly dissociative; rather, they are deactivated by internal
conversion and/or intersystem crossing33 to a lower electronic
state. Dissociation on the S0 potential energy surface would
then be consistent with the behavior of the fragment distributions
P(ET) discussed above, where prior to dissociation substantial
IVR occurs. Although excitation at 193 nm involves a different
excited singlet state, it still leads to a fragment distribution the
features of which are similar to those of the P(ET) distributions
of the lower excitation energies. It is therefore conceivable that
dissociation following 193 nm excitation occurs on the same
potential energy surface as that using λ g 248 nm.
Since the s-triazine dissociation is clearly not a symmetric
three-body process but a sequential one, the possible concert-
edness of the reaction is now considered within the definition
appropriate to photofragment distribution measurements. Thus,
using the rotational period of the intermediate species as a gauge,
we consider a process uncorrelated or not concerted if the
lifetime of the intermediate exceeds a rotational period.10 Under
these conditions rotational averaging causes a complete loss of
spatial correlation among the asymptotic velocity vectors of the
primary and secondary photofragments.14 For s-triazine the
analysis of the fragment distribution was shown to be successful
only if the angular fragment distribution was assumed to be
narrow, which clearly excludes rotational averaging. Conse-
quently the three fragment velocity vectors are correlated, and
the process can be considered to be concerted. More specifi-
cally, it is a two-step but still a concerted photodissociation,
and there is evidence based on a highly structured absorption
band, an isotropic fragment distribution (â ) 0), and an only
small P(ET) dependence on the excitation energy that dissocia-
tion takes place after electronic relaxation (internal conversion/
intersystem crossing) from the initially excited state to a lower
electronic state (electronic predissociation).
Acknowledgment. Support of this work by the Schweiz-
erischer Nationalfonds zur Fo¨rderung der wissenschaftlichen
Forschung is gratefully acknowledged. J.A.H. acknowledges
support from the Miller Institute for Basic Research in Science
(Berkeley, CA) during the early stages of this research. We
thank PD Dr. Peter Felder for valuable discussions, Dr. Philip
R. Willmott for critically reading the manuscript, and Rolf
Pfister for his assistance with the graphics.
References and Notes
(1) Solomon, J.; Jonah, C.; Chandra, P.; Bersohn, R. J. Chem. Phys.
1971, 55, 1908.
(2) Tamir, M.; Halavee, U.; Levine, R. D. Chem. Phys. Lett. 1974, 25,
38.
(3) Korger, P. M.; Riley, S. J. J. Chem. Phys. 1977, 67, 4483.
(4) Korger, P. M.; Riley, S. J. J. Chem. Phys. 1979, 70, 3863.
(5) Kellman, M. E.; Pechukas, P.; Bersohn, R. Chem. Phys. Lett. 1981,
83, 304.
(6) Baughcum, S. L.; Leone, S. R. Chem. Phys. Lett. 1982, 89, 183.
(7) Dewar, M. J. S. J. Am. Chem. Soc. 1984, 106, 209.
(8) Ondrey, G. S.; Bersohn, R. J. Chem. Phys. 1984, 81, 4517.
(9) Zhao, X.; Miller, W. B.; Hintsa, E. J.; Lee, Y. T. J. Chem. Phys.
1989, 90, 5527.
(10) Strauss, C. E. M.; Houston, P. L. J. Phys. Chem. 1990, 94, 8751.
(11) Gruebele, M.; Roberts, G.; Zewail, A. H. Philos. Trans. R. Soc.
London 1990, A332, 223.
(12) Khundkar, L. R.; Zewail, A. H. J. Chem. Phys. 1990, 92, 231.
(13) Trentelman, K. A.; Kable, S. H.; Moss, D. B.; Houston, P. L. J.
Phys. Chem. 1990, 94, 263.
(14) North, S. W.; Blank, D. A.; Gezelter, J. D.; Longfellow, C. A.;
Lee, Y. T. J. Chem. Phys 1995, 102, 4447.
(15) Wannenmacher, E. A. J.; Felder, P.; Huber, J. R. J. Chem. Phys.
1991, 95, 986.
(16) Baum, G.; Felder, P.; Huber, J. R. J. Chem. Phys. 1993, 98, 1999.
(17) Kim, S. K.; Pederson, S.; Zewail, A. H. J. Chem. Phys. 1995, 103,
477.
(18) North, S. W.; Longfellow, C. A.; Lee, Y. T. J. Chem. Phys 1993,
99, 4423.
(19) Gejo, T.; Felder, P.; Huber, J. R. Chem. Phys. 1995, 195, 423.
(20) Busch, G. E.; Mahoney, R. T.; Morse, R. I.; Wilson, K. R. J. Chem.
Phys. 1969, 51, 449.
(21) Wodtke, A. M.; Lee, Y. T. In Molecular Photodissociation
Dynamics; Ashfold, M. N. R., Baggott, J. E., Eds.; Royal Society of
Chemistry: London, 1987; p 31.
The proposed three-body photodissociation mechanism of
s-triazine is similar to the one suggested for azomethane18,19
but differs from that of acetone. The latter has recently been
reinvestigated by North et al.14 using photofragment translational
spectroscopy and by Kim, Petersen, and Zewail17 by femto-
second observation of the transient intermediate. Both groups
establishedscontrary to some earlier claims 44sthat the mech-
anism is stepwise and not concerted. The TOF spectrum of

Three-Body Photodissociation of 1,3,5-Triazine
J. Phys. Chem., Vol. 100, No. 33, 1996 13949
(22) Pacansky, J. J. Phys. Chem. 1977, 81, 2240.
(23) Bolovinos, A.; Tsekeris, P.; Philis, J.; Pantos, E.; Andritsopoulos,
G. J. Mol. Spectrosc. 1984, 103, 240.
(34) Hintsa, E. J.; Zhao, X.; Lee, Y. T. J. Chem. Phys. 1990, 92, 2280.
(35) Zhao, X.; Nathanson, G. M.; Lee, Y. T. Acta Phys.-Chim. Sin. 1992,
8, 70.
(24) Goates, S. R.; Chu, J. O.; Flynn, G. W. J. Chem. Phys. 1984, 81,
4521.
(25) Osamura, Y.; Uonno, M.; Hashimoto, K. J. Am. Chem. Soc 1987,
109, 1370.
(26) Benson, S. W. Thermochemical Kinetics, 2nd ed.; Wiley: New
York, 1976.
(27) Bystro¨m, K. J. Chem. Thermodyn. 1982, 14, 865.
(28) Felder, P. Chem. Phys. 1990, 143, 141.
(29) Morley, G. P.; Felder, P.; Huber, J. R. Chem. Phys. Lett. 1994,
219, 195.
(30) Minton, T. K.; Felder, P.; Brudzynski, R. J.; Lee, Y. T. J. Chem.
Phys. 1984, 81, 1759.
(31) Sparks, R. K.; Shobatake, K.; Carlson, L. R.; Lee, Y. T. J. Chem.
Phys. 1981, 75, 3838.
(32) Zare, R. N. Mol. Photochem. 1972, 4, 1.
(33) Saigusa, H.; Lim, E. C. J. Chem. Phys. 1983, 78, 91.
(36) Felder, P. Habilitationsschrift Thesis, University of Zu¨rich, 1993.
(37) Miller, W. B.; Safron, S. A.; Herschbach, D. R. J. Chem. Soc.,
Faraday Discuss. 1967, 44, 108.
(38) Smith, D. J.; Grice, R. Mol. Phys. 1991, 73, 1371.
(39) Zhoa, X.; Hintsa, E. J. J. Chem. Phys. 1988, 88, 801.
(40) Krajnovich, D.; Huisken, F.; Zhang, Z.; Shen, Y. R.; Lee, Y. T. J.
Chem. Phys. 1982, 77, 5977.
(41) Butler, L. J.; Buss, R. J.; Brudzynski, R. J.; Lee, Y. T. J. Phys.
Chem. 1983, 87, 5106.
(42) Wodtke, A. M.; Hintsa, E. J.; Lee, Y. T. J. Phys. Chem. 1986, 90,
3549.
(43) Innes, K. K.; Ross, I. G.; Moomaw, W. R. J. Mol. Spectrosc. 1988,
132, 492.
(44) For a discussion see ref 17.
JP9609592