UV photodissociation dynamics of allyl radical by photofragment translational spectroscopy

Article abstract of DOI:10.1063/1.477156

The photodissociation of the allyl radical was studied by photofragment translational spectroscopy following excitation to the C(pver-tilde) (2²B₁) and A(over-tilde)(1²B₁) states. At 248 nm excitation, two different primary channels were detected: H-atom loss and CH₃ elimination. The overall shape of the P(E_T) for the H-atom loss channel suggests a statistical dissociation from the ground potential energy surface of the C₃H₅ system.

Full text of DOI:10.1063/1.477156

UV photodissociation dynamics of allyl radical by photofragment translational
spectroscopy
Domenico Stranges, Martin Stemmler, Xueming Yang, James D. Chesko, Arthur G. Suits, and Yuan T. Lee
Citation: The Journal of Chemical Physics 109, 5372 (1998); doi: 10.1063/1.477156
View online: http://dx.doi.org/10.1063/1.477156
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/109/13?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Photodissociation dynamics of C3H5I in the near-ultraviolet region
J. Chem. Phys. 141, 104316 (2014); 10.1063/1.4894393
Photodissociation dynamics of the phenyl radical via photofragment translational spectroscopy
J. Chem. Phys. 133, 074302 (2010); 10.1063/1.3473743
High resolution photofragment translational spectroscopy studies of the near ultraviolet photolysis of phenol
J. Chem. Phys. 125, 133318 (2006); 10.1063/1.2353818
Dynamics of OH formation in photodissociation of pyruvic acid at 193 nm
J. Chem. Phys. 118, 10093 (2003); 10.1063/1.1572133
Fast beam photodissociation spectroscopy and dynamics of the vinoxy radical
J. Chem. Phys. 106, 3049 (1997); 10.1063/1.473419
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.24.51.181 On: Sat, 29 Nov 2014 09:18:39

JOURNAL OF CHEMICAL PHYSICS
VOLUME 109, NUMBER 13
1 OCTOBER 1998
UV photodissociation dynamics of allyl radical by photofragment
translational spectroscopy
Domenico Stranges,a) Martin Stemmler, Xueming Yang,b) James D. Chesko,
Arthur G. Suits, and Yuan T. Leeb)
Department of Chemistry, University of California, Berkeley, California 94720 and Chemical Science
Division, Lawrence Berkeley National Laboratory, University of California, Berkeley, California 94720
͑Received 26 February 1998; accepted 26 June 1998͒
Photodissociation of the allyl radical, CH2CHCH2, has been studied using the method of molecular
2
2
˜
˜
beam photofragment translational spectroscopy following excitation to the C(2 B1) and A(1 B1)
states by 248 and 351 nm photons. Two different primary channels have been detected following
248 nm excitation: H-atom loss ͑84%͒ and CH3 elimination ͑16%͒. From the product translational
energy distribution and polarization dependence studies, dissociation processes from the
ground-state C3H5 potential energy surface are inferred for both wavelengths. At 248 nm there may
also be a contribution to the H-atom loss channel from predissociation by a higher electronically
excited state. Rice–Ramsperger–Kassel–Marcus ͑RRKM͒ calculations show that the formation of
cyclopropene is not important, while formation of allene and methylacetylene from dissociation of
1- and 2-propenyl radicals are important reaction pathways at both wavelengths. Translational
energy distributions peaking well away from zero provide evidence for CH3 elimination directly
from an allylic structure through a four-member cyclic transition state. © 1998 American Institute
of Physics. ͓S0021-9606͑98͒02137-0͔
tected in the 210–250 nm region.19 Oscillator strengths of
several UV bands have been measured experimentally.20,21
The resonant multiphoton ionization ͑MPI͒ technique was
used to probe the one-photon forbidden transition
I. INTRODUCTION
It is well known that neutral free radicals are important
intermediate species in many important chemical processes.
However, very often only limited information is available on
these ‘‘transient’’ species ͑as compared to stable molecules͒,
mainly owing to the difficulties in producing them under
well defined conditions suitable for definitive experimental
studies.
22,23
2
2
˜
˜
B(3s A1)←X( A2)
with a (2ϩ2) scheme. Three fun-
damental vibrational levels in both the ground and excited
states were identified. Recently Chen and co-workers re-
ported an extensive high resolution spectroscopic study on
this radical.24–26 They have produced a supersonic molecular
beam of allyl radical by flash pyrolysis27 of allyl iodide.
From their partially rotationally resolved MPI spectra they
were able to assign many vibronic bands in the 250–240 nm
region and to deduce the band origins of the strongly inter-
The allyl radical is the simplest conjugated ␲-electron
hydrocarbon radical and represents a benchmark molecule
for both theoretical and experimental studies. It is an inter-
mediate in many thermal,1 combustion,2 photochemical,3 and
surface catalytic4,5 reactions. The allyl radical has been the
subject of many experimental and theoretical investigations.
2
2
2
˜
˜
˜
acting B(3s A1), C(2 B1), and D(1 B2) electronic states.
Several theoretical studies at different levels of calcula-
tions have analyzed the ground-state structure and vibra-
Electron-spin resonance ͑ESR͒ and electron diffraction8
6,7
showed that this radical has a planar structure with C2v sym-
metry in its ground state. Photoelectron spectroscopy9 has
been used to determine the vertical and adiabatic ionization
potentials. Vibrational frequencies and structural information
for the ground electronic state have been obtained from ma-
trix infrared ͑IR͒,10,11 ultraviolet ͑UV͒ resonant Raman
spectroscopy,12–16 and IR diode laser spectroscopy.17 Currie
and Ramsay18 measured the absorption spectrum of the allyl
radical in the 410–370 nm region. A system of diffuse bands
tional frequencies.28–34 The ab initio molecular orbital–
35
configuration interaction ͑MO–CI͒
and spin-coupled
36
valence-bond ͑SCVB͒ methods have also been used to
study the ground and excited electronic states of the allyl
radical. The energies of the excited electronic states were
obtained, which are in good agreement with the experimental
results.18,25
Another interesting aspect of the allyl radical is the pos-
sible isomerization into cyclopropyl radical through an elec-
trocyclic reaction.37–39 From the Woodward–Hoffmann
rules,40 based on MO symmetry considerations, both conro-
tatory and disrotatory pathways are symmetry forbidden on
the ground-state potential energy surface, so that the ground
state of each radical correlates with an electronically excited
state of the other.41 However, the thermal ring opening of
cyclopropyl radical into allyl radical has been observed ex-
2
2
˜
˜
in that region were assigned to the A(1 B1)←X( A2) elec-
tronic transition. Other vibronic bands have also been de-
a͒
`
Present address: Dipartimento di Chimica, Universita ‘‘La Sapienza,’’ and
Centro Studi Termodinamica Chimica alle Alte Temperature–CNR, Dipar-
timento di Chimica, Universita ‘‘La Sapienza,’’ P.le A. Moro 5, Rome I-
`
00185, Italy.
b͒
Present address: Institute of Atomic and Molecular Sciences, Academia
Sinica, P.O. Box 23-166, Taipei, Taiwan, ROC.
0021-9606/98/109(13)/5372/11/$15.00
5372
© 1998 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.24.51.181 On: Sat, 29 Nov 2014 09:18:39

J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
Stranges et al.
5373
FIG. 2. Pyrolytic free radical pulsed molecular beam source. ͑a͒ pulsed
valve, ͑b͒ alumina insulator block, ͑c͒ alumina standoff, ͑d͒ SiC tube, ͑e͒
alumina screws, ͑f͒ electrodes, ͑g͒ water cooled copper block.
FIG. 1. Schematic view of the apparatus. Photolysis laser is perpendicular to
the plane of the figure. PV, pulsed valve; DET, detector; E, direction of the
electric vector of the laser. The polarization angle ␥ is defined with respect
to the detector axis as shown, and the laboratory scattering angle ⌰ is the
angle between the molecular beam and the detector axis.
SiC tube held at 1400–1500 K as determined by beam TOF.
Under these conditions, C3H5I pyrolyses quantitatively to
produce an allyl radical and iodine atom. The mass spectra of
the molecular beam with the nozzle unheated ͑upper trace͒
and heated ͑lower trace͒ are shown in Fig. 3. The quantita-
tive loss of C3H5I (m/eϭ168) and the appearance of Iϩϩ
peak (m/eϭ63.5) in the heated mass spectrum are strong
evidence for the formation of the allyl radical, since doubly
charged ions are much more likely to come from one atom
than from the dissociative ionization of C3H5I. The molecu-
lar beam velocity and its velocity spread were measured by
the time-of-flight technique with a spinning slotted disk. Dif-
ferent parts of the gas pulse can be sampled by changing the
delay between the wheel and the pulsed valve so that opti-
mum experimental conditions are achieved. The delay be-
tween the pulsed valve and the laser trigger can be calculated
from the wheel angular velocity and the delay between the
pulsed valve and the wheel. The beam velocity was mea-
sured to be ϳ3300 m/s with a full width half-maximum
͑FWHM͒ velocity spread of 17%.
perimentally and the activation energy was estimated to be
about 20 kcal/mol.42 This result also agrees with an ab initio
calculation39 in which a highly nonsymmetrical transition
state was found.
Very recently Deyerl et al.43 have investigated the pho-
todissociation of the allyl radical at 248.15 nm. The H-atom
loss channel has been detected, showing an average transla-
tional energy release of 13.9 kcal/mol. Despite all the experi-
mental and theoretical studies described above, a full picture
of the primary photochemical processes of the allyl radical is
still lacking. In the present work, we report a study on the
photodissociation of the allyl radical at 248 and 351 nm ex-
citation by photofragment translational spectroscopy ͑PTS͒.
II. EXPERIMENT
The experimental measurements in this work were car-
ried out on a rotating source photofragment-translational
spectrometer which has been described previously in
detail.44,45 A simple schematic drawing of the experimental
setup is shown in Fig. 1. A pulsed molecular beam of the
allyl radical (C3H5) seeded in helium was skimmed twice
before entering the main chamber, where it was crossed at
90° with a focused pulsed laser beam. Neutral photofrag-
ments scattered from the laser-molecular beam interaction
region and which fly along the detector axis are ionized by
an electron bombardment ionizer at a distance of 36.7 cm
from the interaction region. The ions produced in the ionizer
are then mass selected by a quadrupole mass spectrometer,
and counted with a Daly ion detector. A 486 personal com-
puter interfaced multichannel scaler ͑Turbo-MCS, EG&G
Ortec͒ triggered by the laser pulse is used to collect and store
the time-of-flight ͑TOF͒ spectra. Product laboratory angular
distributions can be obtained by rotating the molecular beam
source around the laser propagation axis, which is perpen-
dicular to the plane defined by the molecular beam direction
and the detector axis. The allyl radical pulsed beam was
generated from the pyrolysis of allyl iodide (C3H5I). The
pyrolytic source used in this experiment ͑Fig. 2͒ is an adap-
tation to our pulsed valve46 of the one designed by Chen
et al.27 A mixture of C3H5I/Heϳ0.2% was expanded at a
stagnation pressure of 760 Torr through a resistively heated
FIG. 3. Mass spectra of the molecular beam of ϳ0.2% allyl iodide in He
taken with the pyrolytic nozzle at ͑a͒ room temperature ͑heat off͒ and ͑b͒
1400 K ͑heat on͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.24.51.181 On: Sat, 29 Nov 2014 09:18:39

5374
J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
Stranges et al.
In all the experiments performed in this work the mo-
lecular beam was crossed at 90° by an output of a Lambda
Physik EMG 103 MSC excimer laser operating at the KrF
͑248͒ and XeF ͑351 nm͒ transitions. A convex lens ͑25 cm
focal length͒ was used to focus the laser beam to a 2.5
ϫ6 mm spot in the interaction region, with fluence of 50–
100 mJ/cm2/pulse for studies at 248 nm excitation and 100–
200 mJ/cm2/pulse for studies at 351 nm excitation. The out-
put of the excimer laser was polarized using a pile of ten
quartz plates at Brewster’s angle with respect to the direction
of laser propagation, yielding Ͼ95% polarized light. The po-
larization direction of the laser beam can be rotated with
respect to the detector axis by adjusting the orientation of the
polarizer.
Time-dependent background, arising from the dissocia-
tive ionization of pulsed C3H5 from the molecular beam, was
present in TOF spectra at masses р41 (C3H5) and at small
angles. The shot-by-shot background subtraction method was
used, which has been described in detail elsewhere.47 Briefly,
the pulsed valve was run at twice the laser repetition rate. A
TOF spectrum with laser on and laser off was recorded, and
the laser-off TOF was then subtracted from the laser-on
TOF.
A forward convolution technique was used to fit the
data. Briefly, a trial center-of-mass translational energy dis-
tribution is convoluted with the instrumental response func-
tions and molecular beam conditions to yield a calculated
TOF spectrum in the laboratory reference frame, which is
then compared to the experimental spectrum. The transla-
tional energy distribution is iteratively adjusted until a good
fit is obtained.
At 248 nm excitation, the intensity of all detected frag-
ments from C3H5 photodissociation varied linearly at lower
laser power, but showed saturation at higher laser power. At
351 nm excitation, the intensity of C3H4 fragments ͑see be-
low͒ varied linearly with the laser power up to the maximum
power of the excimer laser, showing that the absorption cross
section at this wavelength is quite small. The TOF spectra
reported in this work have been recorded under conditions
where saturation effects were absent.
FIG. 4. Schematic view of the modified apparatus for H-atom detection. The
photolysis laser is directed perpendicular to the plane of the figure. PV,
pulsed valve; DET, detector. The molecular beam direction is perpendicular
to the detector axis (⌰ϭ90°).
detector axis to let photofragments going to the detector
pass. In this configuration the molecular beam, the laser
beam, and the detector axis are all perpendicular to each
other. The 1.5 mm diameter laser spot is located at a distance
of 1 cm from the nozzle. While the H-atom detection was not
successful for the allyl radical, owing to a low number den-
sity of the parent molecule, H-atom TOF spectra from the
photodissociation of other systems were recorded.50 More-
over, the same configuration was used to detect H and D
atoms from the photolysis of HI and DI at 248 nm in order to
determine experimentally the instrumental response func-
tions ͑the ionizer width and the neutral flight distance͒ of our
apparatus. In Fig. 5 the H- and D-atom TOF spectra from
photodissociation of HI and DI at 248 nm excitation are
shown. From the peak positions of the two spin-orbit com-
ponents, the neutral flight distance ͑the bond dissociation en-
H and D atoms produced in photodissociation experi-
ments are difficult to detect by mass spectrometry for two
main reasons. The signal is very low ͑kinematics factor͒ and
the inherent background inside the detector at those masses
is very high. The only way to increase the signal-to-noise
ratio ͑S/N͒ is to increase the number density of the parent
molecules by reducing the distance between the nozzle and
the interaction region. H-atom TOF spectra have been mea-
sured with success in our group on several systems48 but
always on crossed molecular beam type machines49 because
of the space available in the fixed source chamber. We have
slightly modified our experimental apparatus ͑also called
‘‘rotating source machine’’͒ in order to detect H-atom time
of flight, the C3H4 momentum-matched photofragment. The
modification was simply the addition of an extension to the
source chamber ͑Fig. 4͒. In the modified setup, the pulsed
valve is located upward inside this box. Two small holes
exist along the laser beam propagation direction to let the
laser beam pass through, and a third hole is present along the
FIG. 5. TOF spectra at a laboratory angle of 90 ° for m/eϭ1(Hϩ) and
m/eϭ2(Dϩ) from photodissociation of HI and DI at 248 nm. The open
circles represent experimental data. The solid line is the calculated TOF
from the convolution on the instrumental response functions of the experi-
mental apparatus ͑neutral flight length, 36.7; ionizer width, 0.9 cm͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.24.51.181 On: Sat, 29 Nov 2014 09:18:39

J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
Stranges et al.
5375
FIG. 7. The center-of-mass translational energy distribution used to fit the
data in Fig. 6.
these TOF spectra is shown in Fig. 7. It is peaked around 0
and has a tail that decreases monotonically up to 59 kcal/
FIG. 6. TOF spectra for m/e 38 (C3H2ϩ) from photodissociation of the allyl
radical at 248 nm at the indicated laboratory angles. The open circles are
experimental points and the solid line is the simulation using the P(ET) in
Fig. 7.
mol. The mean translational energy release, E , for this
͗
͘
T
channel is 13.2 kcal/mol and represents an upper limit for the
reasons mentioned before.
For the CH3ϩC2H2 channel, both momentum-matched
photofragments have been detected. TOF spectra for m/e 26
(⌰ϭ7.5, 15, 20, 25, and 30°͒ and 15 (⌰ϭ30°) have been
collected after 300 000 (m/e 26͒ and 1 500 000 (m/e 15͒
laser shots. In Fig. 8, the TOF spectra collected for m/e 26 at
7.5, 15, 20, and 30° are presented. At ⌰ϭ7.5°, two different
peaks, centered around 80 and 120 ␮s, are clearly resolved.
At larger ⌰ only one single peak is present. The Newton
ergy, D0 , of the two channels is well known51͒ is determined
to be 36.7 cm, and from the widths of the two peaks the
ionizer width can also be determined to be 0.9 cm.
III. RESULTS AND ANALYSIS
In all the TOF spectra presented here, the open circles
represent the experimental data, the dashed lines are single
channel contributions to the forward convolution fit, and the
solid lines are the overall fit to the data. The lowest energy
range of the P(ET) distributions are not reported because
those that are very slow are not scattered beyond 2.5° from
the molecular beam.
A. 248 nm excitation
At 248 nm excitation two primary dissociation channels
are present ͑see below͒. For the H-atom loss channel, TOF
spectra were collected for products at m/eϭ41–38
(C3H5ϩ ,C3Hϩ4 ,C3H3ϩ ,C3H2ϩ) at detector angles, ⌰, of 2.5, 3,
4, 5, 6, and 7.5° from the molecular beam. TOF spectra at
m/eϭ41 were recorded in order to check if the signal ob-
served at lower masses came from the photodissociation of
the allyl radical or from heavier hydrocarbon molecules. The
low signal at this mass ͑ϳ2% of m/eϭ40) is due only to 13
C
natural abundance in C3H4 neutral fragments. TOF spectra at
m/eϭ40–38 show a single peak and they are identical when
recorded at the same laboratory angle, suggesting that they
all originate from the same neutral fragment by electron im-
pact dissociative ionization inside the ionizer. TOF spectra at
m/eϭ39 and 38 have the best S/N. Figure 6 shows the TOF
spectra for m/e 38 at ⌰ϭ3, 5, 6, and 7.5°. The number of
laser shots necessary for each spectrum ranged from 25 000
͑at small ⌰͒ to 100 000 ͑at larger ⌰͒. The center-of-mass
͑c.m.͒ translational energy distribution, P(ET), used to fit
FIG. 8. TOF spectra for m/eϭ26 (C2Hϩ2 ) from photodissociation of allyl
radical at 248 nm at the indicated laboratory angles. The open circles are
experimental points, the dashed line is the contribution from the C2H2ϩCH3
channel, the dotted line is the contribution from the C3H4ϩH channel, and
the solid line is the total fit to the data.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.24.51.181 On: Sat, 29 Nov 2014 09:18:39

5376
J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
Stranges et al.
FIG. 9. Newton diagram for allyl radical photodissociation at 248 nm. The
circles represent the thermodynamic limit for each photofragment.
FIG. 11. TOF spectrum for m/e 15 (CHϩ3 ) from photodissociation of the
allyl radical at 248 nm excitation at ⌰ϭ30°. The open circles are experi-
mental points and the solid line is the simulation using the P(ET) in Fig. 10.
diagram is shown in Fig. 9, with circles representing the
highest possible velocities of the photofragments when all
the available energy is channeled into translation. For the
H-atom loss channel, the C3H4 product can be allene and/or
methylacetylene ͑see Sec. IV͒. It is clear from the compari-
son with TOF of C3H4 that the slow component in the TOF
at m/eϭ26 and 7.5° is due to C3H4 dissociative ionization
and its contribution can be fitted with the P(ET) shown in
Fig. 7. The P(ET) used to fit the acetylene channel is shown
in Fig. 10. It has a totally different shape as compared to the
relative cross sections of fragments a and b monitored at a
given m/e ratio and corrected for kinematics,
o
Rϭ ␴ /␴ ͒ϭ ␴o/␴ ͒ Q /Q ͒ F /F ͒.
͑1͒
͑
͑
͑
͑
a b
a
b
b
a
a
b
Qa and Qb are the ionization cross sections of the products
which result in fragment a and b and were calculated using
the method of Fitch and Sauter.53 Fa and Fb are the prob-
abilities of products in the ionization process to form frag-
ments a and b at the given m/e ratio. We estimate the un-
H-atom loss channel. It is peaked at 16, E ϳ19.1, with a
͗
͘
T
FWHM of 20.5, and extends up to 62.0 kcal/mol. The same
P(ET) has also been used to fit the TOF spectrum at m/e
ϭ15 ͑Fig. 11͒, showing that it is the momentum-matched
counterpart of acetylene.
certainty in the branching ratio obtained in this way to be
o
Ϯ25%. The ratio ␴oa/␴ ͑calculated by our analysis program͒
b
depends on the anisotropy parameter ␤, even if unpolarized
light is used.54 Polarization dependence studies were done
for m/eϭ38, ⌰ϭ4°, and m/eϭ26, ⌰ϭ20°. The polariza-
tion direction of the light can be rotated in the plane defined
by the molecular beam and the detector axis. The anisotropy
parameter is obtained directly from the laboratory angular
distribution55 I(␥),
In order to determine the branching ratio, R, between the
two primary dissociation channels, the TOF spectrum at
m/eϭ26 and 7.5° was used because both contributions were
recorded under the same experimental conditions. R was cal-
o
o
culated according to Eq. ͑1͒,52 where ␴ and ␴ are the
a
b
I ␥͒ϰ1ϩ␤P cos ␥Ϫ␥ ͒͒,
͑
͑
͑
2
0
where ␥ is the laboratory polarization angle, ␥0 is the cor-
rection angle for the c.m. to laboratory transformation, and
P2(␥Ϫ␥0) is the second Legendre polynomial in cos(␥
Ϫ␥0). For the acetylene plus methyl (CH3) channel we did
not observe any dependence of the angular distribution from
the polarization angle, indicating that this dissociation pro-
cess is isotropic (␤ϭ0). For the H-atom loss channel, a very
small negative anisotropy parameter ␤, Ϫ0.05Ϯ.02, was ob-
served. The branching ratio between the H-atom loss channel
and the acetylene plus methyl channel at 248 nm excitation
is determined to be 0.84:0.16.
B. 351 nm excitation
At the 351 nm excitation, only the H-atom loss channel
was observed. TOF spectra were recorded for m/eϭ40–38
at detector angles of 2.5, 3, 4, 5, and 6° from the molecular
beam. Modulated background was also subtracted shot by
shot. At 351 nm the available energy for the same asymptotic
FIG. 10. The center-of-mass translational energy distribution used to fit the
contribution of C2H2ϩCH3 photodissociation channel in Figs. 8 and 11.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.24.51.181 On: Sat, 29 Nov 2014 09:18:39

J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
Stranges et al.
5377
FIG. 14. Dissociation and isomerization reaction schemes. Also indicated
are the unimolecular rate constants calculated by RRKM theory ͑cf. Table
I͒.
FIG. 12. TOF spectra for m/e 39 (C3Hϩ3 ) from photodissociation of the
allyl radical at 351 nm at the indicated laboratory angles. The open circles
are experimental points and the solid line is the simulation using the P(ET)
in Fig. 13.
ϭ4.5 kcal/mol. The anisotropy parameter measured for this
channel was zero. Given the good signal-to-noise ratio for
the TOF’s and their simulation, the uncertainly in E is
estimated to be less than 0.5 kcal/mol at both 351 and 248
nm.
͗
͘
T
products is 33.5 kcal/mol lower as compared to 248 nm. The
Newton circle will be smaller and the neutral C3H4 photo-
fragments can be scattered only up to 6° from the molecular
beam. TOF spectra obtained here show only a single peak,
and they are identical for different m/e values when col-
lected at the same laboratory angle. In Fig. 12, TOF spectra
at m/eϭ39 and ⌰ϭ2.5, 3, 4, and 6° are shown. The number
of laser shots necessary for each spectrum ranges from
200 000 to 400 000. The c.m. translational energy distribu-
tion used to fit these TOF spectra is shown in Fig. 13. It has
a similar shape as compared to the H-atom loss P(ET) at 248
nm. It peaks near 0 and decreases monotonically up to 24.5
kcal/mol. The upper limit for the mean translational energy
C. RRKM calculations
In order to establish the relative importance of possible
isomerization and dissociation reaction channels from the
ground-state potential energy surface ͑PES͒ of the C3H5 sys-
tem ͑see Sec. IV͒ we have performed RRKM calculation to
determine the microcanonical rate constants k(E)’s for all
the reaction pathways shown in Fig. 14. The lack of ab initio
calculations at the same level of theory for all reaction path-
ways on the ground-state C3H5 PES required us to choose
empirical methods to estimate vibrational frequencies and
structures for the different radical species and transition
states ͑TS’s͒. The vibrational frequencies for the allyl radical
have been taken from Refs. 10, 11, 13, and 34, for the cy-
clopropyl radical from Ref. 56. Vibrational frequencies for
the 1- and 2-propenyl radicals were obtained from propylene
removing one C–H stretch and two C–H bending vibrational
modes. For k1 , k5 , k6 , k7 , k8 , and k10 the TS’s have been
defined following the prescriptions given by Pearson and
Rabinovitch.57 TS’s for the isomerization reactions among
the allyl radical and the 1- and 2-propenyl radicals ͑through
1,3 and 1,2 H-atom migration, respectively͒ have been built
in a similar way as Ibuki et al.58 The TS for allyl-
cyclopropyl isomerization (k9 ,kϪ9) was taken halfway. For
the dissociation of allyl radical into C2H2ϩCH3(k2) through
a four-member cyclic TS we have adopted a similar proce-
dure used for k3 and k4 TS’s with an additional lowering of
CH2 wagging, twisting, and rocking vibrations. In Table I,
the calculated rate constants for 248 and 351 nm, together
with energetic information and A-factors, are reported. Some
of the isomerization and dissociation reactions have been
release at 351 nm excitation is determined to be
E
͘
T
͗
FIG. 13. The center-of-mass translational energy distribution used to fit the
data in Fig. 12.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.24.51.181 On: Sat, 29 Nov 2014 09:18:39

5378
J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
Stranges et al.
TABLE I. RRKM microcanonical rate constants, A-factors, excitation energies E, and zero-point energy dif-
ferences E0 between transition state and reactanta for the reactions reported in Fig. 14.
248 nm
k(E)/1010
351 nm
k(E)/109
(sϪ1
E0
E
E
Reaction
(sϪ1
)
)
͑kcal/mol͒
log A
͑kcal/mol͒
͑kcal/mol͒
4.4
2.6
13.0
32.6
3.9
11.6
7.8
19.1
18.3
19.5
1.6
115.3
115.3
115.3
103.2
115.3
100.0
103.2
100.0
103.2
100.0
115.3
87.2
0.38
0.03
5.29
81.5
81.5
81.5
69.7
81.5
66.2
69.7
66.2
69.7
66.2
81.5
53.4
53.4
59.0b
68.1c
48.6d
36.8d
49.8e
34.5e
47.2b
40.5f
45.8b
43.2b
50.0g
21.9g
52.9h
13.7
13.9
13.6
13.1
13.4
13.0
13.2
14.0
13.4
13.4
13.3
13.3
13.7
k1
k2
k3
kϪ3
k4
kϪ4
k5
k6
k7
k8
17.4
1.62
7.44
0.95
7.46
2.70
2.92
0.76
k9
kϪ9
k10
70.3
1.1
115
0.00
87.2
a⌬H0 for reactants taken from Refs. 59 and 61.
bFrom Ref. 59 and references cited therein.
cA barrier of 20 kcal/mol for the reverse reaction is assumed ͑see text͒.
dFrom Ref. 66.
eFrom Ref. 58.
fFrom Ref. 64.
gFrom Ref. 39.
hA barrier of 3 kcal/mol for the reverse reaction is assumed.
investigated previously59 with RRKM theory, and our results
are in agreement with these results. In order to get branching
fractions for the formation of allene, methylacetylene, cyclo-
propene, and acetylene, a set of differential equations for the
reaction rates has been solved with the Laplace transform
method,60 and the results are listed in Table II. For the 248
nm case the initial condition in which only the allyl radical is
present is reported here ͑assuming cyclopropyl radical as the
initial molecular species present, the branching fraction val-
ues change less than one percentage unit, so that the global
picture does not change͒.
The estimates of vibrational frequencies and structures
made by empirical methods, and the use of the harmonic
oscillator model to determine RRKM k(E)’s will certainly
introduce some degree of error into the final branching frac-
tion calculations. However, the qualitative picture given by
these calculations is in agreement with our experimental re-
sults ͑see Sec. IV͒.
IV. DISCUSSION
A. Decay mechanisms at 248 nm
As mentioned above, Chen’s group24–26 reported a de-
tailed spectroscopic study of the allyl radical in the 240–250
nm region by resonant MPI. The band origin for the C
˜
(2 2B1) state was found to be 248.15 nm. The broadband
KrF excimer laser used here certainly covers this transition.
The absorption cross section increases at shorter wave-
lengths, reaching a maximum at ϳ222 nm.21 Several weaker
and broader vibronic bands have also been observed at
shorter wavelengths, implying a shorter lifetime for the ex-
cited state at the shorter wavelengths. The results of a gas-
phase resonant Raman study at 225 nm are consistent with
isomerization to cyclopropyl radicals.12 In fact, the longer
˜
lifetime of the C state at longer wavelengths could be due to
the presence of a barrier for the electrocyclization transfor-
mation by a disrotatory mode.41 In this case the C state de-
˜
cay could be through internal conversion ͑IC͒ to form vibra-
tionally hot allyl radicals either in the ground state or in the
TABLE II. Branching fractions for the different dissociation channels cal-
culated with the Laplace transform method ͑Ref. 60͒.
43
˜
A state. At shorter wavelengths, more rapid decay via
isomerization to cyclopropyl is possible, provided sufficient
energy is available to overcome the aforementioned barrier.
Following IC, allyl radical in the ground electronic state
with 115.3 kcal/mol of vibrational energy can either dissoci-
ate or isomerize. A schematic energetic diagram with pos-
sible dissociation and isomerization processes is shown in
Fig. 15. Thermochemical data have been taken from Refs. 59
and 61, and corrected to 0 K ͑see also Table I͒. There are two
possible dissociation channels directly from an allylic struc-
ture: H-atom loss to form allene, reaction ͑R1͒, and CH3
elimination through a concerted mechanism, reaction ͑R2͒.
As mentioned above for the excited state, the allyl radical
Branching fraction
at 248 nm
Branching fraction
at 351 nm
Reaction products
39%
28% from CH2CHCH2
11% from CH2C•CH3
34%
27%
17% from CH2CHCH2
10% from CH2C•CH3
42%
CH2CCH2ϩH
CHCCH3ϩH
9% from •CHCHCH3
25% from CH2C•CH3
27%
12% from •CHCHCH3
30% from CH2C•CH3
31%
C2H2ϩCH3
c-C3H4ϩH
17% from CH2CHCH2
10% from •CHCHCH3
0.1%
1% from CH2CHCH2
30% from •CHCHCH3
0.0%
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.24.51.181 On: Sat, 29 Nov 2014 09:18:39

J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
Stranges et al.
5379
FIG. 15. Schematic diagram of the energetics of the allyl radical for the possible reaction channels following excitation at 248 and 351 nm. The numbers are
relative energies ͑in kcal/mol͒ for different C3H5 isomers, transition states, and reaction products.
B. The H-atom loss channel at 248 nm
can isomerize by an electrocyclic transformation through a
highly nonsymmetrical TS39 into cyclopropyl radical, reac-
tion ͑R9͒, or into 1- and 2-propenyl radicals through an H-
atom shift, reactions ͑R4͒ and ͑R3͒, respectively. Finally, the
cyclopropyl radical can lose an H-atom to form cyclopro-
pene, reaction ͑R10͒; 1-propenyl radical (•CHCHCH3) can
either lose the H-atom attached to the central carbon atom to
form methylacetylene, reaction ͑R8͒, or can break the C–C
bond to form C2H2 and CH3, reaction ͑R6͒; 2-propenyl radi-
cal (CH2C•CH3) can lose an H-atom to form either methy-
lacetylene, reaction ͑R7͒, or allene, reaction ͑R5͒. There are
five different reaction pathways that produce an H-atom and
allene/methylacetylene/cyclopropene, and two for the forma-
tion of acetylene plus the methyl radical,
The P(ET) for the H-atom loss channel ͑Fig. 7͒ is
peaked near zero and extends up to the thermodynamic limit
for allene/methylacetyleneϩH formation. This observation
suggests that either or both molecular species may be formed
as the final products. From the maximum ET value it is not
possible to establish if the less endothermic dissociation
channel, CH3CCHϩH, is present. This is owing to the small
difference ͑1.5 kcal/mol͒ between the enthalpy of formation
(⌬H0) of allene and that of methylacetylene. Cyclopropene
is less stable than allene/methylacetylene by about 21–22
kcal/mol, therefore the available energy for reaction ͑R10͒
should be reduced by the same amount. This means that if
this channel is present it should contribute at lower ET . We
do not observe any change of the slope of the P(ET) around
the thermodynamic onset for this channel (ET
ϳ37 kcal/mol); moreover, the branching fraction obtained
from RRKM calculations ͑Table II͒ is Ͻ1%. This low value
is mainly due to the lower rate constant k10 for dissociation
as compared to the isomerization reaction kϪ9 ͑Table I͒. The
same RRKM calculations show that allene and methylacety-
lene are produced in almost equal amounts through an exten-
sive isomerization before dissociation. More than 50% of the
dissociation events do not come from allyl radicals directly.
While the shape of the P(ET) is in qualitative agreement
with a statistical dissociation, the mean translational energy
release of 13.2 kcal/mol, in good agreement with what has
been observed previously,43 and a slightly negative anisot-
ropy parameter are in apparent contradiction to the statistical
dissociation picture. An RRKM type P(ET), which was cal-
culated assuming no barrier on the exit channel, predicts an
2
1
2
˜
˜
CH CHCH X A ͒→CH CCH X A ͒ϩH S͒ ͑R1͒
͑
2
͑
2
͑
2
2
2
1
ϩ
g
2
1
2
˜
˜
˜
͑
3
CH CHCH X A ͒→C H X ⌺ ͒ϩCH X AЉ͒
͑
2
͑
2
2
2
2
2
͑R2͒
͑R3͒
͑R4͒
͑R5͒
͑R6͒
2
˜
˜
͑
CH CHCH X A ͒→CH C•CH X͒
͑
2
2
2
2
3
2
˜
˜
͑
CH CHCH X A ͒→•CHCHCH X͒
͑
2
2
2
3
1
2
˜
CH C•CH ϰCH CCH X A ͒ϩH S͒
•CHCHCH →C H X ⌺ ͒ϩCH X AЉ͒
CH C•CH →CHCCH X A ͒ϩH S͒
͑
2
͑
2
3
2
1
ϩ
g
1
2
˜
˜
͑
͑
2
3
2
3
2
1
2
˜
͑R7͒
͑R8͒
͑
3
͑
2
3
1
1
2
˜
•CHCHCH →CHCCH X A ͒ϩH S͒
͑
3
͑
3
1
2
2
˜
˜
͑
CH CHCH X A ͒→c-C3H5 X A ͒
͑R9͒
͑
2
Ј
2
2
of ϳ5 kcal/mol.62 Even if we assume that all the energy
͘
T
2
2
˜
˜
c-C3H5 X A ͒→c-C H X͒ϩH S͒
͑R10͒
E
͑
͑
4
͑
Ј
͗
3
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.24.51.181 On: Sat, 29 Nov 2014 09:18:39

5380
J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
Stranges et al.
from the potential barrier for the reverse reaction is chan-
neled into translation, the predicted E will be only ϳ7–8
changing the initial conditions for the solution of the set of
differential equations, with the cyclopropyl radical as the
only molecular species initially present instead of the allyl
radical, the global picture will not change. The only small
difference is that the branching fraction for reaction ͑10͒
rises to 1%.
͗
͘
T
kcal/mol, still less than the experimental value.
2
2
˜
˜
The transition moment of the C(2 B1)←X( A2) transi-
tion is perpendicular to the C2␯ axis and in the molecular
plane35 (My-type transition͒. Prompt dissociation to form al-
lene plus an H-atom from the ground-state geometry would
yield ␤ϭϪ1. The anisotropy parameter ␤ will tend to zero
for a polyatomic dissociation process with its dissociation
time much longer than the rotational period of the parent
molecule. This kind of behavior is expected for dissociation
processes occurring from the ground electronic state follow-
ing internal conversion from an electronically excited state.
C. The methyl؉acetylene channel at 248 nm
The second primary dissociation channel detected at 248
nm excitation is the formation of acetylene and the methyl
radical. This accounts for 16% of the total dissociation
events. In the studies on the thermal decomposition of the
allyl radical,59 the formation of acetylene has been observed,
and was supposed to come only from 1-propenyl radical dis-
sociation, reaction ͑R6͒. The barrier for the CH3 addition
reaction to C2H2 to form 1-propenyl radical has been found
to be 7.7 kcal/mol.64 This also represents the barrier on the
exit channel for the reverse reaction. The upper limit for the
most probable ET for the dissociation of 1-propenyl into
acetylene and CH3 can be estimated to be about 7.7 kcal/mol,
assuming that all the potential energy barrier is channeled
into translation. However, our P(ET) for this is peaked at 16
kcal/mol ͑Fig. 10͒, much larger than the maximum expected
for decomposition of the 1-propenyl radical. Our P(ET) thus
suggests the existence of a tighter TS for this dissociation
channel. A possible candidate for this transition state is a
four-member cyclic TS, where one CH2 group is rotated by
about 90°. Dissociation of the allyl radical can occur through
reaction ͑R2͒, where C2H2 and CH3 are produced directly
from the allyl radical through this four-member cyclic TS.
The proposed TS structure is similar to the one for H2 elimi-
nation reaction from ethylene where the P(ET) was found to
be peaked at 20 kcal/mol.65 The P(ET) peak position repre-
sents a lower limit to the potential energy barrier height. The
asymmetry of the TS for reaction ͑R2͒ will couple efficiently
the reaction coordinate with the other internal degrees of
freedom, suggesting that the true barrier will be consistently
higher and the products should be rotationally and vibra-
tionally excited. The total branching fraction for this channel
obtained from RRKM calculations ͑Table II͒ is 27%, higher
than the value, 16%, determined experimentally. Moreover,
10% of the 27% fraction is due to reaction ͑6͒ according to
the RRKM calculations. Even though the TOF spectra at
m/eϭ26 ͑Fig. 8͒ and the P(ET) do not show any bimodal-
ity, a contribution from reaction ͑R6͒ can still be present,
especially at low ET . We have simulated TOF spectra with
two different contributions, one with a P(ET) peaked at 16
for reaction ͑R2͒ and the other peaked at ϳ5 kcal/mol for
reaction ͑R6͒. Although this simulation procedure has a cer-
tain degree of arbitrariness, it shows that the contribution of
reaction ͑R6͒ will likely be less than 15%–20% of the total
for this channel, which is still somewhat lower than the cal-
culated value. The contribution of reaction ͑R6͒ from the
RRKM calculations can be lowered by increasing the barrier
for the isomerization between the allyl and 1-propenyl radi-
cals and/or choosing a tighter TS for reaction ͑R6͒. The
branching fractions reported in Table II have been calculated
by considering only statistical dissociation reactions from the
The higher E value found experimentally and the slightly
͗
͘
T
negative ␤ may be due to a competition between two differ-
ent microdissociation mechanisms: internal conversion and
predissociation through an excited electronic state. In this
case, the anisotropy distribution measured should come from
two different contributions, one nearly isotropic (␤ϭ0) and
another with ␤Ͻ0. For the allyl radical, the only rotation that
can change the angle between the initial direction of the tran-
sition moment and the final direction of the C–H bond, prior
to bond scission, is around the axis perpendicular to the mo-
lecular plane. The classical rotational period around this axis,
assuming a rotational temperature of 150 K, is 4.2 ps. From
the partially rotationally resolved spectra we have estimated
˜
a lifetime for the C state at 248 nm of 1–2 ps. This shows
that predissociation should be faster than the rotational pe-
riod showing up in a substantial negative value of ␤ for the
dissociation process. The low measured value of the anisot-
ropy thus signifies the importance of the IC process at 248
nm.
˜
Internal conversion of the C state to the first low lying
2
˜
electronic excited A(1 B1) state will produce allyl radicals
with 11.4 kcal/mol of vibrational energy. Currie and
Ramsay18 observed a number of diffuse vibronic bands in the
2
˜
410–370 nm region, which were assigned to the A(1 B1)
2
˜
←X( A2) transition. The resolution of their spectrophotom-
eter was high enough to resolve rotational structures18 and
they concluded that predissociation, to form allene plus an
H-atom, could be responsible for the diffuseness of those
bands. Irradiation of allyl radicals trapped in an Ar matrix
with 410 nm light was shown to produce new vibrational
bands in the IR spectrum assigned to the cyclopropyl
radical.56 The orbital correlation diagram for the interconver-
sion of allyl and cyclopropyl radicals40,41 shows that the A
˜
state of the allyl radical correlates with the ground state of
the cyclopropyl radical along the conrotatory mode, suggest-
ing that the ring closure process will be fast ͑symmetry al-
lowed͒. This conclusion is also supported by ab initio calcu-
lations at the CASSCF level.63 The main result of this
calculation shows that the conrotatory pathway has no bar-
rier along the reaction coordinate, and the disrotatory path-
way has a barrier less than 1.5 kcal/mol. Their conclusion
was that the cyclization process would be facile. They thus
argued that this cyclization, rather than predissociation, was
responsible for the diffuse character of the vibronic bands of
the first absorption transition. As mentioned above, by
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.24.51.181 On: Sat, 29 Nov 2014 09:18:39

J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
Stranges et al.
5381
V. CONCLUSION
ground potential energy surface, therefore they do not in-
clude a possible contribution to the H-atom loss channel
from predissociation. The higher branching fraction for the
acetylene channel, as compared to the experimental value,
may be due to the fact that the barrier assumed for reaction
͑R2͒ is too low in the RRKM calculations ͑20 kcal/mol͒. The
dissociation products will be internally extremely hot, since
on average about 77% of the available energy for the H-atom
loss channel will be deposited to internal excitation ͑mainly
vibrational excitation͒, and about 72% of the available en-
ergy remains in internal degrees of freedom for the acetylene
plus CH3 channel.
Photodissociation of the allyl radical has been studied by
photofragment translational spectroscopy following excita-
2
2
˜
˜
tion to the C(2 B1) and A(1 B1) states. At 248 nm excita-
tion, two different primary channels have been detected: H-
atom loss ͑84%͒ and CH3 elimination ͑16%͒. The overall
shape of the P(ET) for the H-atom loss channel suggests a
statistical dissociation from the ground potential energy sur-
face of the C3H5 system. The decay mechanism of the ex-
cited state is either through internal conversion into the
˜
ground state or internal conversion into the A state followed
by an electrocyclic transformation into the ground-state cy-
clopropyl radical and, finally, reopening of the ring to form
the allyl radical in the ground state with 115.3 kcal/mol of
vibrational energy. The small negative anisotropy parameter
␤ and the higher mean translational energy release E , as
compared to an RRKM type calculated value, can be ratio-
D. 351 nm dissociation
͗
͘
T
The absorption of a 351 nm photon excites the allyl radi-
2
˜
˜
cal to the A(1 B1) state. The A state decays through an
electrocyclic transformation to the ground state of the cyclo-
propyl radical with 53.4 kcal/mol of vibrational energy. The
absence of angular anisotropy (␤ϭ0), a P(ET) peaked
around zero, and a mean translational energy of 4.5 kcal/mol
are all consistent with this mechanism for the H-atom loss
channel. In fact, a subpicosecond predissociation process to
form allene plus an H-atom will show up with a negative ␤
nalized by the competition between two different decay
mechanisms for the C state: internal conversion and predis-
˜
sociation through a higher excited state. RRKM calculations
show that almost 50% of allyl radicals dissociate to give
allene plus an H-atom and acetylene plus CH3, while the
remaining radicals isomerize into 1- and 2-propenyl radicals
before dissociation to form allene and methylacetylene plus
an H-atom. The P(ET) for the acetylene plus CH3 channel is
peaked away from zero at 16 kcal/mol and is due mainly to
dissociation from an allylic structure. This result, supported
also by RRKM calculations, has been interpreted as disso-
ciation from an allylic structure through a cyclic four-
member transition state. A smaller contribution to this chan-
nel could also come from dissociation of the 1-propenyl
radical.
˜
˜
͑the A state has the same symmetry as the C state͒. An
RRKM type P(ET) for reactions ͑R1͒, ͑R8͒, ͑R7͒, and ͑R5͒,
which are assumed barrierless, predict
a
E
͗ ͘
T
ϭ2.5 kcal/mol. If we add ϳ2 kcal/mol for the repulsive in-
teraction due to the small exit barrier, the agreement with the
experimental value is very good. The threshold for the for-
mation of cyclopropene plus an H-atom, reaction ͑R10͒, is
49.9 kcal/mol. Assuming a barrier of 3 for the reverse reac-
tion, the available energy to overcome this barrier is only 0.5
kcal/mol. The rate constant for reaction ͑10͒, k10 , will be so
low ͑Table I͒ that cyclopropene formation will not compete
with the isomerization into ground-state allyl radical through
a nonsymmetrical TS, kϪ9 . RRKM calculations show that
isomerization of the allyl radical with 81.5 kcal/mol vibra-
tional energy is extensive. Only 18% of the dissociation
products ͑Table II͒ come from an allylic structure. Methy-
lacetylene is predicted to be the most abundant product with
42% yield, allene 27%, and acetylene 31% ͑with 30% from
dissociation of the 1-propenyl radical͒. The inability to detect
signal at m/eϭ26 may be due to different reasons. If the
barrier to the isomerization between the allyl and the 1-
propenyl radicals is higher, the branching fraction for the
acetylene channel will then be reduced. From a comparison
of the signal levels for m/eϭ38–39 at 248 and 351 nm, we
can conclude that the absorption cross section of the allyl
radical at 351 nm is ϳ100 times smaller than 248 nm, which
is in agreement with previous results.35 The kinematics also
plays an important role. For the H-atom loss channel the
signal of the heavy fragment is concentrated in only 6 deg in
the laboratory reference frame, while the Newton circle for
C2H2 extends up to 40 deg. Finally, for the 351 nm excitation
the dissociation products will still be internally hot, since on
average ϳ82% of the available energy will be channeled into
C3H4 internal excitation ͑mainly vibrational excitation͒.
˜
The A state of allyl radical is reached after excitation at
351 nm. The decay mechanism is again an electrocyclic
transformation into the ground-state cyclopropyl radical, fol-
lowed by a reopening of the ring to form the allyl radical in
the ground state with 81.5 kcal/mol vibrational energy.
RRKM calculations also show that isomerization into the 1-
and 2-propenyl radical is extensive, with final production of
methylacetylene and allene plus an H-atom.
ACKNOWLEDGMENTS
D.S. gratefully acknowledges the Consiglio Nazionale
delle Ricerche ͑Italy͒ for a fellowship. This work was sup-
ported by the Director, Office of Energy Research, Office of
Basic Energy Science, Chemical Science Division of the
U.S. Department of Energy under Contract No. DE-AC03-
76SF00098.
1 W. Tsang, Int. J. Chem. Kinet. 10, 1119 ͑1978͒.
2 M. Weissman and S. W. Benson, Prog. Energy Combust. Sci. 15, 273
͑1989͒.
3 R. E. Linder, D. L. Winters, and A. C. Ling, Can. J. Chem. 54, 1405
͑1976͒.
4 R. N. Carter, A. B. Anton, and G. Apai, J. Am. Chem. Soc. 114, 4410
͑1992͒.
5 J. C. Schultz and J. L. Beauchamp, J. Chem. Phys. 87, 3587 ͑1983͒.
6 R. W. Fessenden and R. H. Schuler, J. Chem. Phys. 39, 2147 ͑1963͒.
7 M. J. McManus, R. W. Fessenden, and D. M. Chipman, J. Phys. Chem.
92, 3778 ͑1988͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.24.51.181 On: Sat, 29 Nov 2014 09:18:39

5382
J. Chem. Phys., Vol. 109, No. 13, 1 October 1998
Stranges et al.
8 E. Vajda, J. Tremmel, B. Rozsondai, I. Hargittai, A. K. Maltsev, N. D.
Kagramanov, and O. M. Nefedov, J. Am. Chem. Soc. 108, 4352 ͑1986͒.
9 F. A. Houle and J. L. Beauchamp, J. Am. Chem. Soc. 100, 3290 ͑1978͒.
10 J. W. Huang and W. R. M. Graham, J. Chem. Phys. 93, 1583 ͑1990͒.
11 G. Maier, H. P. Reisenauer, B. Rohde, and K. Dehnicke, Chem. Ber. 116,
732 ͑1983͒.
41 H. C. Longuet-Higgins and E. W. Abrahamson, J. Am. Chem. Soc. 87,
2045 ͑1965͒.
42 G. Greig and J. C. J. Thynne, Trans. Faraday Soc. 62, 3338 ͑1966͒.
43 H.-J. Deyerl, T. Gilbert, I. Fischer, and P. Chen, J. Chem. Phys. 107, 3329
͑1997͒.
44 A. M. Wodtke and Y. T. Lee, J. Phys. Chem. 89, 4744 ͑1985͒.
45 A. M. Wodtke, Ph.D thesis, University of California, Berkeley, 1986.
46 D. Proch and T. Trickl, Rev. Sci. Instrum. 60, 713 ͑1989͒.
47 D. Stranges, X. Yang, J. D. Chesko, and A. G. Suits, J. Chem. Phys. 102,
6067 ͑1995͒.
12 J. D. Getty, M. J. Burmeister, S. G. Westre, and P. B. Kelly, J. Am. Chem.
Soc. 113, 801 ͑1991͒.
13 J. D. Getty and P. B. Kelly, Chem. Phys. 168, 357 ͑1992͒.
14 X. Liu, J. D. Getty, and P. B. Kelly, J. Chem. Phys. 99, 1522 ͑1993͒.
15 J. D. Getty, X. Liu, and P. B. Kelly, J. Phys. Chem. 96, 10,155 ͑1992͒.
16 J. D. Getty, X. Liu, and P. B. Kelly, Chem. Phys. Lett. 201, 236 ͑1993͒.
17 E. Hirota, C. Yamada, and M. Okunishi, J. Chem. Phys. 97, 2963 ͑1992͒.
18 C. L. Currie and D. A. Ramsay, J. Chem. Phys. 45, 488 ͑1966͒.
19 A. B. Callear and H. K. Lee, J. Chem. Soc., Faraday Trans. 64, 308
͑1968͒.
48 R. E. Continetti, B. A. Balko, and Y. T. Lee, J. Chem. Phys. 93, 5719
͑1990͒; Chem. Phys. Lett. 182, 400 ͑1991͒; B. A. Balko, J. Zhang, and Y.
T. Lee, J. Chem. Phys. 94, 7958 ͑1991͒; D. A. Blank, S. W. North, and Y.
T. Lee, Chem. Phys. 187, 35 ͑1994͒.
49 Y. T. Lee, J. D. McDonald, P. R. LeBreton, and D. R. Herschbach, Rev.
Sci. Instrum. 40, 1402 ͑1968͒.
20 H. E. van der Bergh and A. B. Callear, Trans. Faraday Soc. 66, 2681
͑1970͒.
50 D. Stranges, A. G. Suits, and Y. T. Lee ͑unpublished results͒; D. A. Blank,
S. W. North, D. Stranges, A. G. Suits, and Y. T. Lee, J. Chem. Phys. 106,
539 ͑1997͒.
21 N. Nakashima and K. Yoshihara, Laser Chem. 7, 177 ͑1987͒.
22 J. W. Hudgens and C. S. Dulcey, J. Phys. Chem. 89, 1505 ͑1985͒.
23 A. D. Sappey and J. C. Weisshaar, J. Phys. Chem. 91, 3731 ͑1987͒.
24 D. W. Minsek, J. A. Blush, and P. Chen, J. Phys. Chem. 96, 2025 ͑1992͒.
25 J. A. Blush, D. W. Minsek, and P. Chen, J. Phys. Chem. 96, 10,150
͑1992͒.
51 H. Okabe, Photochemistry of Small Molecules ͑Wiley, New York, 1978͒.
52 A. M. Schmoltner, Ph.D. thesis, University of California, Berkeley, 1989.
53 W. L. Fitch and A. D. Sauter, Anal. Chem. 55, 832 ͑1983͒.
54 J. G. Frey and P. Felder, Mol. Phys. 75, 1419 ͑1992͒.
55 G. E. Busch and K. R. Wilson, J. Chem. Phys. 56, 3638 ͑1972͒; R. N.
Zare, Mol. Photochem. 4, 1 ͑1972͒.
26 D. W. Minsek and P. Chen, J. Phys. Chem. 97, 13,375 ͑1993͒.
27 D. W. Kohn, H. Clauberg, and P. Chen, Rev. Sci. Instrum. 63, 4003
͑1992͒; H. Clauberg, D. W. Minsek, and P. Chen, J. Am. Chem. Soc. 114,
99 ͑1992͒; H. Clauberg and P. Chen, ibid. 113, 1445 ͑1991͒.
28 S. D. Peyerimhoff and R. J. Buenker, J. Chem. Phys. 51, 2528 ͑1969͒.
29 T. Takada and M. Dupuis, J. Am. Chem. Soc. 105, 1713 ͑1983͒.
30 C. Cometta-Morini, T.-K. Ha, and J. F. M. Oth, J. Mol. Struct.:
THEOCHEM 188, 79 ͑1989͒.
56 K. Holtzhauer, C. Cometta-Morini, and J. F. M. Oth, J. Phys. Org. Chem.
3, 219 ͑1990͒.
57 M. J. Pearson and B. S. Rabinovitch, J. Chem. Phys. 42, 1624 ͑1965͒.
58 T. Ibuki, T. Murata, and Y. Takezaki, J. Phys. Chem. 78, 2543 ͑1974͒.
59 J. Niedzielski, J. Gawlowski, and T. Gierczak, J. Photochem. 21, 195
͑1983͒; M. Naroznik and J. Niedzielski, ibid. 32, 281 ͑1986͒.
60 J. I. Steinfeld, J. S. Francisco, and W. L. Hase, Chemical Kinetics and
Dynamics ͑Prentice-Hall, Englewood Cliffs, NJ, 1989͒.
61 CRC Handbook of Chemistry and Physics, edited by D. R. Lide ͑CRC,
Boca Raton, FL, 1992͒.
31 E. Fjogstad and M. Ystenes, Spectrochim. Acta A 46, 47 ͑1990͒.
32 P. G. Szalay, A. G. Csaszar, G. Fogarasi, A. Karpfen, and H. Lischka, J.
Chem. Phys. 93, 1246 ͑1990͒.
33 F. Sim, D. R. Salahub, S. Chin, and M. Dupuis, J. Chem. Phys. 95, 4317
͑1991͒.
34 E. Hirota, J. Mol. Struct. 320, 75 ͑1994͒.
62 The RRKM type P(ET) calculated for the reactions ͑3͒, ͑9͒, ͑11͒, and ͑12͒
35 T.-K. Ha, H. Baumann, and J. F. M. Oth, J. Chem. Phys. 85, 1438 ͑1986͒.
36 J. M. Oliva, J. Gerratt, D. L. Cooper, P. B. Karadakov, and M. Raimondi,
J. Chem. Phys. 106, 3663 ͑1997͒.
give the same
E
value.
͘
T
͗
63 M. Yamaguchi, J. Mol. Struct.: THEOCHEM 365, 143 ͑1996͒.
64 J. A. G. Dominguez and A. F. Trotman-Dickenson, J. Chem. Soc. , 940
͑1962͒.
37 L. Farnell and W. G. Richards, J. Chem. Soc. Chem. Commun. , 334
͑1973͒.
65 B. A. Balko, J. Zhang, and Y. T. Lee, J. Chem. Phys. 97, 935 ͑1992͒; E.
F. Cromwell, A. Stolow, M. J. J. Vrakking, and Y. T. Lee, ibid. 97, 4029
͑1992͒; A. Stolow, B. A. Balko, E. F. Cromwell, J. Zhang, and Y. T. Lee,
J. Photochem. Photobiol., A 62, 285 ͑1992͒.
38 P. Merlet, S. D. Peyerimhoff, R. J. Buenker, and S. Shih, J. Am. Chem.
Soc. 96, 959 ͑1974͒.
39
´
S. Olivella, A. Sole, and J. M. Bofill, J. Am. Chem. Soc. 112, 2160
͑1990͒.
66 Von B. Ondruschka, U. Ziegler, and G. Zimmermann, Z. Phys. Chem.
͑Leipzig͒ 267, 1127 ͑1986͒.
40 R. B. Woodward and R. Hoffmann, The Conservation of Orbital Symme-
try ͑Academic, New York, 1970͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.24.51.181 On: Sat, 29 Nov 2014 09:18:39