Infrared laser-induced post-pulse dissociation of CF2HCl and CF2Cl2 under high pressure and fluence conditions

Article abstract of DOI:10.1063/1.472781

The unimolecular decomposition of the halogenated methanes CF₂HCl (one main channel) and CF₂Cl₂ (two main channels) in the focused beam of a pulsed CO₂ laser under high pressure and fluence conditions (p = 100 Pa-2 kPa; Φ=5-200 J/cm²) was studied by a special laser-induced fluorescence (LIF) technique, permitting spatially resolved fragment concentration measurements in the focal region. Considerable amounts of CF₂ product were formed after the end of the laser pulse. In the one-channel-dissociation case of CF₂HCl LIF measurements of the CF₂ yield distribution Y(z,r) can be related to the spatial distribution of the average absorbed energy in the parent molecules. Only part of the absorbed energy is consumed by multiphoton dissociation, while most reactant molecules remain highly vibrationally excited in the focus volume far into the double cone. Using the long-lived CF₂ also as a probe for measuring the rotational, translational, and vibrational temperatures, the redistribution of the internal energy in the molecules and fragments involved is monitored. The post-pulse production of CF₂ is shown to be caused by the energy pooling v-v transfer mechanism, while contributions of pyrolytic and gas dynamic processes are of little importance.

Full text of DOI:10.1063/1.472781

Infrared laserinduced postpulse dissociation of CF2HCl and CF2Cl2 under high
pressure and fluence conditions
W. Strube, J. Wollbrandt, M. Rossberg, and E. Linke
Citation: The Journal of Chemical Physics 105, 9478 (1996); doi: 10.1063/1.472781
View online: http://dx.doi.org/10.1063/1.472781
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/105/21?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Quasiclassical trajectory calculations for the OH(X 2Π) and OD(X 2Π)+HBr reactions: Energy partitioning and
rate constants
J. Chem. Phys. 105, 9897 (1996); 10.1063/1.472855
Coriolisdependent Stark effect of the 2ν3 band of methane observed by saturated absorption spectroscopy
J. Chem. Phys. 105, 9027 (1996); 10.1063/1.472737
Unimolecular decomposition of NO3: The NO+O2 threshold regime
J. Chem. Phys. 105, 6807 (1996); 10.1063/1.472527
The highpressure range of the reaction of CH(2Π) with N2
J. Chem. Phys. 105, 5423 (1996); 10.1063/1.472383
Semiempirical MNDO, AM1, and PM3 direct dynamics trajectory studies of formaldehyde unimolecular
dissociation
J. Chem. Phys. 104, 7882 (1996); 10.1063/1.471504
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:
129.22.67.107 On: Mon, 24 Nov 2014 11:38:36

Infrared laser-induced post-pulse dissociation of CF2HCl and CF2Cl2
under high pressure and fluence conditions
W. Strube, J. Wollbrandt, M. Rossberg, and E. Linke
¨
¨
Humboldt-Universitat Berlin, Institut fur Chemie, Forschungsgruppe Laser-Photochemie,
Rudower Chaussee 5, Haus 4.1, D 12484 Berlin, Germany
͑Received 9 July 1996; accepted 29 August 1996͒
The unimolecular decomposition of the halogenated methanes CF2HCl ͑one main channel͒ and
CF2Cl2 ͑two main channels͒ in the focused beam of a pulsed CO2 laser under high pressure and
fluence conditions ͑pϭ100 Pa–2 kPa; ⌽ϭ5–200 J/cm2͒ was studied by a special laser-induced
fluorescence ͑LIF͒ technique, permitting spatially resolved fragment concentration measurements in
the focal region. Considerable amounts of CF2 product were formed after the end of the laser pulse.
In the one-channel-dissociation case of CF2HCl LIF measurements of the CF2 yield distribution
Y(z,r) can be related to the spatial distribution of the average absorbed energy in the parent
molecules. Only part of the absorbed energy is consumed by multiphoton dissociation, while most
reactant molecules remain highly vibrationally excited in the focus volume far into the double cone.
Using the long-lived CF2 also as a probe for measuring the rotational, translational, and vibrational
temperatures, the redistribution of the internal energy in the molecules and fragments involved is
monitored. The post-pulse production of CF is shown to be caused by the energy pooling
–
v v
2
transfer mechanism, while contributions of pyrolytic and gas dynamic processes are of little
importance. © 1996 American Institute of Physics. ͓S0021-9606͑96͒01345-1͔
INTRODUCTION
EXPERIMENT
The experiments involved real-time, in situ laser-
induced fluorescence ͑LIF͒ detection of the X A1 CF2 pro-
duced in the ir multiphoton dissociation ͑IRMPD͒ of
CF2HCl
For a number of years ir lasers have been used as a
versatile energy source in gas-phase chemistry for selective
reactant activation.1 Pressure and irradiation conditions cho-
sen in practical applications, such as powder production,2
isotope separation,3 or radical generation4 usually present a
compromise between an acceptable yield and improved se-
lectivity of the chemical reactions involved. While most fun-
damental studies are concerned with a detailed understanding
of the primary photophysical and photochemical processes in
infrared multiphoton excitation and decomposition under
collision-free and/or low fluence ͑less than a few J/cm2͒
conditions,5 much less is known about the complex processes
occuring in the high pressure/high fluence regime ͑pressure
pulse length product pϫ␶pϾ2ϫ105 Pa ns, fluence ⌽у5
J/cm2͒, when using standard TEA lasers. This parameter
range is of considerable importance. While the significant
radical yields and subsequent radical reactions in this range
are relevant to practical applications, the high excitation lev-
els reached there can extend the underlying theoretical mod-
els. However, the laser-induced processes and their depen-
dence on fluence and pressure are complex and difficult to
predict in this range.
1
˜
˜
CF2HCl→X CF2ϩHCl ⌬Hfϭ203 kJ/mol,
and of CF2Cl2
˜
CF2Cl2→X CF2ϩCl2
⌬Hfϭ306 kJ/mol,
⌬Hfϭ332 kJ/mol,
CF2Cl2→CF2ClϩCl
and in possible secondary reactions of the primary products
of this latter two-channel unimolecular reaction having a flu-
ence dependent branching ratio, which favors the CF2Cl
channel at high fluences.8
The experimental technique and apparatus were de-
scribed previously.9 Briefly, the halomethanes were photo-
lyzed with tightly focused ir radiation of a homemade pulsed
CO2 TEA laser operating in TEM00 mode. The laser emis-
sion could be grating tuned to the respective absorption fre-
quencies of the halomethanes ͓9R͑32͒ at 1085.8 cmϪ1 for
CF2HCl, 10P͑36͒ at 929.02 cmϪ1 for CF2Cl2͔. The typical
CO2-laser pulse shape consisted of a 350 ns ͑FWHM͒ main
pulse containing approximately 50% of the energy, while the
other 50% were contained in the pulse tail. In order to take
into account the action of this tail, instead of a halfwidth or
an effective pulse length ␶ the total pulse length tpϭ2.9 ␮s is
used throughout this paper. The laser beam was focused with
an fϭ8 cm lens. As a result, the Gaussian fluence field
⌽(z,r) in the focal region was highly inhomogeneous ͓Fig.
5͑b͔͒ with a confocal parameter of 5 mm ͑twice the Rayleigh
range͒ and a waist diameter 2wϭ190 ␮m ͑1/e2 point͒.
One of the phenomena observed under high pressure/
high fluence conditions is the continued dissociation of the
irradiated gas long after the end of the ir laser pulse, an effect
which was found earlier for both substances studied in this
paper. It was assumed to be originating from some collision-
7
induced energy transfer ͑CF2HCl;6 CF2Cl2 ͒. A more de-
tailed study is still missing and will be given in the present
paper.
9478
J. Chem. Phys. 105 (21), 1 December 1996
0021-9606/96/105(21)/9478/8/$10.00
© 1996 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:
129.22.67.107 On: Mon, 24 Nov 2014 11:38:36

Strube et al.: Dissociation of CF2HCl and CF2Cl2
9479
The size of the stainless steel reaction cell was fitted to
the irradiation conditions and the need to exclude wall ef-
fects ͑length 10 cm, diameter 3 cm͒. Gas flow rate ͑1.3 sccm
at 800 Pa͒ and laser repetition rate ͑2 Hz͒ were adjusted to
flush the probed region between laser shots. After checking
the purity of the gases by ir absorption and gas chromatog-
raphy they were then used without further purification. Pres-
sures ranged between 100 and 2000 Pa. In some experiments
argon was used as buffer gas at partial pressures up to 12
kPa.
The LIF technique used for determining the concentra-
tion of the photolysis fragments in real time is based on a
frequency-doubled dye laser as probe ͓pulse duration 5.6 ns
͑FWHM͒, bandwidth 1.5 cmϪ1͔. After exciting with this la-
ser an absorptive transition in the species to be detected, the
measured signal is the fluorescence out of the electronically
excited state, which is a measure of the concentration of that
species in the selected rovibronic state in the electronic
ground state. In order to monitor CF2 it was usually excited
FIG. 1. CF2 product concentration profiles obtained by radial scans at a
fixed axial position ͑zϭ0͒.
The principal ability of our SRF arrangement to collect
in a radial scan at a fixed axial position z all species of a
given sort present at a given time in a region normal to the
photolysis laser axis offers the unique possibility of deter-
mining the total amount M(t) of that species in a volume of
height h. For the given values of focal length f and slit
height h this volume is approximately cylindrical in the fo-
cus core region studied. We thus can write
from the vibronic level Љϭ0 of the ␯ bending vibration in
v
2
1
˜
the electronic ground state X A to the vibronic level Јϭ2
v
1
of the same vibration in its electronically excited state
1
˜
A B1. Individual rotational lines of CF2 were not resolved.
M t͒ϭ2␲h r ϽRC r ,t͒r dr
͑1͒
Ј
˜
Vibrational temperatures of X CF2 were determined by ex-
͑
͑ Ј
Ј
Ј
͵
r ϭ0
Ј
citing it with the dye laser from the low-lying vibronic states
˜
with R being the cell radius. Since all parameters, such as
vibrational temperature, rotational temperature, radiative
lifetime, rates of formation, reaction and relaxation, obtain-
able by conventional LIF can be measured now for their
spatial dependence, new information on processes occurring
in the photolysis zone with its strong gradients is to be
gained.
ϭ0, 2, 3 to a selected vibronic state ͑ Јϭ2͒ in the A state,
vЉ
v
from which the fluorescence to the Љϭ1 state was mea-
v
sured. After corrections of the LIF signals for Franck–
Condon factors and different saturation of the absorption
transitions, the signals yield the relative population of the
vibronic states and thus a vibrational temperature, provided
the population follows a Boltzmann distribution. The satura-
tion correction factors took into account the deviation of the
LIF signal measured for a given dye laser energy Edl from
the linear LIF͑Edl͒ dependence. These deviations were deter-
mined in separate measurements, giving 53%, 40%, and 8%
RESULTS
Post-pulse production of CF2
at the focus centre for Љϭ0, 2, 3, respectively, and for the
standard dye laser energy of 20 ␮J.
v
Delaying the probe laser pulse with respect to the ir pho-
tolysis laser pulse yields the development in time of product
concentration at a fixed location. Figure 2 shows such a
In order to improve spatial resolution and thus to probe
regions of constant ir fluence, a collinear arrangement of the
beams of CO2 laser and dye laser was chosen. Furthermore,
the dye laser was focused ͑fϭ250 mm͒ to a diameter ͑1/e2͒
of 58 ␮m, and the probed cylindrical volume was limited to
5.3ϫ10Ϫ6 cm3 by reducing the height of the parallel mono-
chromator slit ͑2 mm͒, onto which the fluorescence from that
volume was imaged 1:1. Employing the spatially resolved
fluorescence ͑SRF͒ technique,9 spatial concentration profiles
of a given species could be measured in radial and/or axial
scans by changing the relative focus position of the probe
laser with respect to the photolysis laser focus, i.e., to the
processing zone. Figure 1 shows typical radial SRF profiles
taken at a fixed axial position. Timing between the two lasers
was controlled with an analog delay generator at a jitter of
8.5 ns. The delay times td given for the dye laser refer to the
onset of the CO2 laser pulse. The fluorescence was detected
by a photomultiplier tube ͑EMI 9558 QB, integration time
constant 3 ␮s͒, the signal of which was fed to a boxcar
integrator with 400 ns gate.
FIG. 2. Time-resolved fluorescence of CF2 in the focus center ͑zϭ0, rϭ0͒.
Maximum at 2 ␮s by primary ͑photolytic͒ CF2 formation; maximum at 25
␮s by secondary CF2 formation.
J. Chem. Phys., Vol. 105, No. 21, 1 December 1996
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:
129.22.67.107 On: Mon, 24 Nov 2014 11:38:36

9480
Strube et al.: Dissociation of CF2HCl and CF2Cl2
FIG. 3. Time-resolved LIF of CF2 ͑zϭ0, rϭ0͒. Separation of the net LIF
signal in a primary production and a post-pulse production part for the
reactants CF2Cl2 ͑a͒ and CF2HCl ͑b͒.
curve for CF2 in its vibrational ground state in the centre
͑zϭ0, rϭ0͒ of the CO2 laser focus. The main features of this
curve are two consecutive maxima. As will be shown later,
the first is caused by immediate photolytical fragmentation.
The second is treated in the following.
FIG. 4. The number of CF2͑0,0,0͒ radicals 2␲h
͐ ͓CF2(r)͔rdr vs time at
The derivation of kinetic parameters from such measure-
ments, however, is usually not straightforward due to the
interference with reaction, relaxation, and diffusion of the
species out of the probed region. In the present case it was
possible to separate the contribution of diffusion, because ͑i͒
zϭ0 shows retardation of secondary CF2 formation with reactant pressure
͑signals corrected for fluorescence quenching͒. ͑a͒ Reactant CF2Cl2 . ͑b͒ Re-
actant CF2HCl.
˜
a similar curve for X CF2͑0,3,0͒ showed that vibrational re-
laxation is much slower than diffusion, and ͑ii͒ the CF2 den-
sity decay of the photolytically generated CF2 could be fitted
with an expression derived in Ref. 10
nel leading directly to CF2, the yield and primary formation
rate are higher, while the initial rate of secondary CF2 for-
mation is comparable to the CF2Cl2 case at the same pres-
sures. This suggests a common, collision-controlled mecha-
nism.
C tϭ0͒
Ϫr2
␳
͑
C r,t͒ϭ
͑
0r exp
͵
ͩ ͪ
2Dt
4 Dt
0
The effects of diffusion can, however, be completely
eliminated by the integration procedure ͓Eq. ͑1͔͒ given in the
experimental section, which permits to cover all CF2 radicals
at any radial position at a given time. Integrating the radial
concentration profiles measured for different times yields the
curves shown in Fig. 4͑a͒. As can be seen, the amount of CF2
produced after the end of the laser pulse reaches multiples of
the photolytically produced CF2. When the pressure is in-
creased, the position of the curve maximum resulting from
the combined action of CF2͑0,0,0͒ formation and reaction to
other products is shifted towards longer times. Relaxation of
CF2 from higher vibrational states also affects the position of
the maximum.11 Again, the CF2 produced from CF2HCl
shows a qualitatively similar behavior ͓Fig. 4͑b͔͒, but the
effect is less pronounced. Here, the factor for the post-pulse
increase of CF2 is independent of pressure.
2
r
Ϫr
rr
Ј
Ј
ϫ
0r exp
I
dr dr
͑2͒
Ј
Ј
͵
ͩ
ͪ ͩ ͪ
0
4Dt
2Dt
0
for the diffusion out of the probe laser beam, the diffusion
constant of which was found to be only dependent on pres-
sure, not on fluence ͑␳ and r0 are the radii of the probe and
0
photolysis laser beams, respectively; I0 is the modified
Bessel function of the first kind of order zero͒. Hence, the
fraction of the photolytically generated, radially diffusing
CF2 could be subtracted from the measured signal. The re-
sidual signal ͓Fig. 3͑a͔͒ due to essentially post-pulse produc-
tion of CF2 in secondary processes then explains the second
maximum in Fig. 2 at about 25 ␮s for the given pressure.
A qualitatively similar result is obtained for CF2HCl
͓Fig. 3͑b͔͒. Here, since IRMPD proceeds via only one chan-
J. Chem. Phys., Vol. 105, No. 21, 1 December 1996
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:
129.22.67.107 On: Mon, 24 Nov 2014 11:38:36

Strube et al.: Dissociation of CF2HCl and CF2Cl2
9481
As can be seen, the shape of the yield distribution is
essentially determined by the characteristic fluence distribu-
tion ⌽(z,r) shown in the equifluence line plot ͓Fig. 5͑b͔͒.
This distribution was calculated for the case of no absorp-
tion; it would yield a maximum fluence of 100 J/cm2. How-
ever, a numerical calculation starting with a reasonable ␴
value at the entrance window and testing various ␴͑⌽͒ de-
pendences, which would reproduce the average energy ab-
sorbed along the length of the cell, gave no significant
changes of this fluence field ͑Ͻ10%͒. This is due to the
relatively high total transmission value ͑Ͼ50%͒ and thus the
small incremental changes ⌬E(z) of the absorbed energy, so
that ⌽(z,r) is essentially determined by the respective beam
cross section ␲͓w(z)͔2.
In Fig. 5͑a͒ the small axial deviation of the yield distri-
bution from the fluence distribution is caused by absorption
of the counterpropagating CO2 and dye laser beams, respec-
tively, along the length of the cell. The radial perturbation,
however, is due to outward expansion, the velocity of which
is dependent on z. Peak velocities of several hundred m/s
were observed. The decrease of the CF2 concentration to be
seen in the very focus center is assumed to be due to probe-
laser induced photolysis of the hot CF2 precursor/parent
molecule.13 Despite these minor distortion effects the ob-
served decomposition pattern qualitatively reflects the en-
ergy absorption from the ir laser field, which is the source of
all secondary post-pulse processes. Lack of a unique ␴͑⌽͒
relation prevents a quantitative ͗Eabs(z,r)͘ picture14 via
͗Eabs͘ϭ⌽ϫ␴͑⌽͒, where ͗Eabs͘ is the average absorbed en-
ergy per molecule.
Energy transfer processes
As an average value ͗Eabs͘ defines only the ‘‘center of
gravity,’’ not the width and shape of the vibrational energy
distribution in the ensemble of parent molecules. Generally,
the energy distribution is determined by the dynamics of the
excitation processes. Hence, an explanation of the energy
absorption at high halomethane pressures must include not
FIG. 5. ͑a͒ Lines of equal CF2͑0,0,0͒ fragment concentration ͑᭹ measured
values͒ for the reactand CF2HCl. The numbers refer to percent of maximum
value. ͑b͒ Focus of the CO2 laser beam ͑without gas absorption͒. Lines of
equal fluence are given as multiples of the fluence at the focusing lens. The
broken curves ⌽ϭ⌽(z,rϭ0)/e2 give an orientation for comparison with
͑a͒.
only the fluence ⌽(t)ϭ͐
t0I(tЈ)dtЈ, but also processes redis-
tributing the molecular energy. The excitation level ͗Eabs͘ is
given by the equilibrium of the optical pumping rate with the
rates of dissociation and deactivation. Under the given high
pressure conditions deactivation occurs essentially by colli-
sions. At low excitation levels collisional energy transfer is
relatively well understood both experimentally and
theoretically.15 Especially for larger polyatomics, however,
knowledge in the high excitation region is still only
fragmentary.16
In order to gain some insight into the energy transfer
processes involved here, the internal energy of the CF2 prod-
uct will be used as a probe, provided that—at least at some
later time—its rotational and vibrational energies are essen-
tially determined by exchange with its collision partners
rather than by the partitioning of the reactant energy in the
IRMPD process. This situation will have to be checked sepa-
rately for rotational and vibrational energy transfer. Espe-
Inhomogeneous energy absorption
Due to the fluence dependence of the absorption cross
section ␴ in the IRMPD of the parent molecules the inho-
mogenous irradiation of a focused laser results in a spatially
dependent energy absorption, i.e., in a spatially inhomog-
enous dissociation yield. The spatial CF2 yield distribution
was obtained from radial LIF profiles measured at different
axial positions ͑for a given time͒ by connecting points of
equal LIF intensity. These represent lines of equal CF2 con-
centration in a given quantum state. Figure 5͑a͒ shows such
an isoconcentration representation for a delay time tdϭ500
ns after the onset of the CO2-laser pulse. The ‘‘altitude
lines’’ give the fractional dissociation yield in percent of the
maximum value, which corresponds to 20% of the total re-
actant concentration. This value was determined using the
method given in Ref. 12 based on benzene as an emitter of
known density and LIF characteristics.
cially favorable conditions should exist for the
–
v v
transfer
J. Chem. Phys., Vol. 105, No. 21, 1 December 1996
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:
129.22.67.107 On: Mon, 24 Nov 2014 11:38:36

9482
Strube et al.: Dissociation of CF2HCl and CF2Cl2
FIG. 6. Fit of the solution TR(t)ϪT0ϭ(TR(0)ϪT0)exp͑ϪkRmt͒ to the
measured change of rotational temperature TR(t) with time ͑T0Ϫroom tem-
perature͒.
between parent molecule and CF2, since the energy defect
⌬E lowering the transfer rate by a factor f per ⌬EϷ300
cmϪ1 ͑fϭ0.1ϫcollision rate͒17 is less than 10 cmϪ1 even for
FIG. 7. Plot of the vibrational temperature vs time of the CF2 product in the
focus center ͑zϭ0, rϭ0͒.
cold parents and product ͓␯ ͑CF2HCl͒ϭ1116 cmϪ1 vs
8
␯ ͑CF2͒ϭ1112 cmϪ1; ␯ ͑CF2Cl2͒ϭ672 cmϪ1 vs ␯ ͑CF2͒
3
6
2
ϭ672 cmϪ1͔. Furthermore, the
–
v v
transfer rate kvv be-
rium, this extremely low value does not characterize the pro-
cess of rotational relaxation in a cold buffer gas, but the slow
tween the freon molecules ͑number of vibrational degrees of
freedom siϭ9͒ and CF2 should be lower only by a factor of
2 as compared to the transfer between parent molecules,
since
cooling process of the gas heated by –T transfer from the
v
freon molecules, which serve as a heat bath for the CF2 frag-
ments. At the given pressure, room temperature is reached in
about 50 ␮s.
sisj
k
vvϰZ
͑3͒
The vibrational temperatures of the CF2 product, how-
ever, show a different behavior. It was found that even for
the shortest delay time tdϭ500 ns the population follows the
Boltzmann law. Such a distribution was also found in Ref.
20 for nascent CF2 in collision free IRMPD ͑⌽р6 J/cm2͒ of
our reactants CF2HCl ͑TVϭ1160 K͒ and CF2Cl2 ͑Tvϭ1050
K͒. A plot of the vibrational temperatures of the CF2 product
vs time under our high fluence conditions ͑Fig. 7͒ shows not
only that the observed maximum temperatures exceed these
nascent temperatures, it also shows that this maximum exci-
tation is only reached after the CO2 laser pulse for the case
of CF2Cl2. Reducing the pressure by 50% yields this delay
also for CF2HCl. Obviously the measured vibrational energy
distribution in the CF2 product at the given pressure and
fluence conditions must be the result of a fast and intense
siϩsj
with Z the gas kinetic collision rate.18 CF2 is also especially
suited as a probe, since it is long lived and its rotational as
well as vibrational energies can be monitored by LIF.
A rotational temperature of CF2 could only be deter-
mined for times у1 ␮s, ͑у80 collisions at pϭ800 Pa͒, be-
cause only then a model spectrum calculated under the as-
sumption of a Boltzmann distribution with a rotational
temperature TR could be matched to the measured excitation
spectrum of a vibrational band. Extrapolation of our results
͑Fig. 6͒ to shorter times gives maximum temperatures of
1070 K for CF2HCl and 880 K for CF2Cl2, respectively.
From this maximum value the rotational excitation decays at
a rate kRϭ3.3ϫ1011 cm3/͑mol s͒ for CF2HCl and 1.1ϫ1011
cm3/͑mol s͒ for CF2Cl2. This rate was determined from the
solution
–
v v
exchange with the undissociated and highly excited
parent molecules. From the maximum value the vibrational
temperature decays first at a rate of ͑5Ϯ2͒ϫ1011 cm3/͑mol s͒,
which is about the same as the cooling rate of the R/T tem-
perature. While the latter temperature had reached room tem-
perature after about 50 ␮s, continued slowing down of the
TR t͒ϪT ϭ T 0͒ϪT ͒exp Ϫk mt͒
͑4͒
͑
͑
͑
͑
0
R
0
R
of the linear differential equation
dTR
–T rate causes the establishment of thermal equilibrium,
v
ϭϪk m T ϪT ͒,
͑5͒
͑
R
R
0
dt
i.e., TVϭTR/T , only after several hundred microseconds.
The radial Tv profiles in Fig. 8 show that already during
the final part of the ir laser pulse the region of highly excited
parent molecules/products stretches across a volume exceed-
ing the CO2 laser radius by several times. The long-lasting
decay of the vibrational temperature of the CF2 in space and
time reflects the high energy content of the vibrational en-
describing the relaxation process ͑see Fig. 6͒, where T0 is
room temperature and m is concentration in mol/cm3. This
rate is orders of magnitude lower than the collision rate or
the known relaxation rate for rotation/translation kR/Tϭ1.3
ϫ1014 cm3/͑mol s͒ in argon.19 Since at the times and pres-
sures considered the R-T degrees of freedom are in equilib-
J. Chem. Phys., Vol. 105, No. 21, 1 December 1996
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:
129.22.67.107 On: Mon, 24 Nov 2014 11:38:36

Strube et al.: Dissociation of CF2HCl and CF2Cl2
9483
ingly high decomposition rate, which was detected as CF2
formation rate. Large quantities of vibrational energy remain
accumulated in the focal volume stretching far into the
double cone of the focus. The molecular energy is then re-
distributed via different –X channels (Xϭ ,T) during and
v
v
especially after the input of radiative energy, resulting in the
observed smearing out in time of product formation.
It was shown that –T energy transfer, despite causing
v
translational temperatures as high as 1000 K, plays no role in
producing significant amounts of CF2 product. Accordingly,
intermolecular
–
transfer turns out to be the main redis-
v v
tribution route. So long as product concentrations are still
low, transfer involves essentially the parent molecules
–
v v
themselves because of lack of other collision partners. From
‘‘cold’’ molecules in the focus neighborhood the excitation
level is rising towards the focus center. Vibrational energy
transfer of weakly excited molecules is relatively well under-
stood ͑reviewed in Ref. 24͒. The main mechanism
FIG. 8. Decay in space and time of the CF2 vibrational temperatures.
M
͒ϩM
͒→M
ϩ1͒ϩM 2Ϫ1͒ϩ⌬Ekin
͑v1 ͑v
͑v1
͑v2
ergy reservoir with which the CF2 interacts. It is another clue
to explain the delayed dissociation behavior.
is especially efficient ͑typically, 4–30 collisions͒ within the
‘‘ladder’’ of a respective vibration, owing to favorable reso-
nance conditions in this quasiharmonic potential region.
Upon higher excitation, i.e., with increasing anharmonicity
the efficiency of this mechanism was found to drop. Thus it
should not play a significant role at higher excitation levels.
Whether this low-energy mechanism regains its role, how-
ever, when the quasicontinuum of vibrational states is
reached, is still a matter of discussion.
It would be interesting to check now, if the high tem-
peratures observed would explain the measured product
yields on the grounds of a pyrolytic dissociation mechanism.
For a thermal reaction the yield is only a function of tem-
perature, not of the dynamics of reaching that temperature.
We thus solve for the translational temperature of 1000 K the
equation
d CF
͓
͔
2
Few studies exist so far on the
–
v v
energy transfer of
ϭk1 CF HCl Ϫk CF 2ϪkϪ1͓CF2͔͓HCl͔
͑6͒
͓
͔
͓
͔
2
2
2
dt
highly excited freons ͑i.e., in the quasicontinuum͒, particu-
larly of CF2HCl and CF2Cl2. The limited number of experi-
mental observables and the marked dependence of experi-
mental results on the collision partners involved makes it
difficult to establish models of general validity. Some trends,
however, e.g., on the dependence of the average energy
for the thermal CF2 formation rate ͑k1ϭ6.92ϫ1013
exp͑Ϫ233.5 kJ/RT͒ sϪ1 21
kJ/RT͒ cm3 molϪ1 sϪ1 22
,
,
k2ϭ1.3ϫ1011 ͑T/300͒1/2 exp͑Ϫ5.0
kϪ1 ϭ2.14ϫ1011 exp͑Ϫ50.7 kJ/
RT͒ cm3 molϪ1 sϪ1 23͒. Since in the beginning the bimolecu-
lar reaction terms are small with respect to the unimolecular
decomposition term, this is straightforward. As a lower value
one obtains a value of 60 ms to reduce the initial concentra-
tion of CF2HCl to 50%, while the measured rate ͓see Fig.
3͑b͒, 800 Pa͔ gives a time of 1.7 ␮s for reaching half the
maximum concentration ͑net yield YϷ0.5͒. Thus primary
CF2 formation is essentially photolytic ͑IRMPD͒. As was
shown above, the observed decay of the translation tempera-
ture within 50 ␮s also rules out significant contributions
from a thermal dissociation mechanism for times longer than
the CO2 laser pulse length tp .
⌬E transferred per collision on the excitation level, on the
͗
͘
number of atoms in the colliding molecule/collision partner,
or on temperature have been worked out:16,25
͑i͒ ⌬E increases with ͗Eabs͘ of the excited species,
͗
͘
showing a dependence between quadratic and linear; for
more complex molecules ͑C2–R, C3–R͒ saturation was ob-
served.
͑ii͒ ⌬E increases with the number of atoms in the
͗
͘
collision partner from about 5 kJ/mol ͑C2H4Cl2 with He͒ to
about 50 kJ/mol ͑C2H4Cl2 with C4H8͒; for more complex
partners the differences in transfer efficiencies are not pro-
nounced. From data published for partners of comparable
complexity to the freons studied here,26 the transferred ener-
DISCUSSION
The energy source responsible for the observed effects is
the ir radiation contained in the CO2 laser pulse of duration
tp . Upon multiphoton excitation of a resonant vibration of
gies ⌬E should be somewhat greater than 15 kJ/mol, i.e.,
͗ ͘
they should be in the energy range of the ir laser quanta.
͑iii͒ ⌬E is only weakly dependent on temperature. In
͗
͘
the molecule and subsequent fast intramolecular
–
trans-
general, it decreases with rising temperature. Above a limit-
ing temperature of about 1000 K, however, it increases due
to the action of the repulsive part of the potential.
v v
fer high internal excitation levels or even unimolecular de-
composition were achieved. The latter dissociation consumes
part of the absorbed vibrational energy in a photolytically
induced unimolecular IRMPD process with a correspond-
͑iv͒
– transfer is very efficient in establishing a Boltz-
v v
mann distribution out of an originally arbitrary distribution.27
J. Chem. Phys., Vol. 105, No. 21, 1 December 1996
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:
129.22.67.107 On: Mon, 24 Nov 2014 11:38:36

9484
Strube et al.: Dissociation of CF2HCl and CF2Cl2
Using these more or less general trends, notions can be
developed of those processes, which eventually result in the
observed high energy accumulation during the laser pulse
and to the continued dissociation by energy redistribution
after the pulse. Because of its energy dependence ͓cf. ͑i͔͒ the
energy exchange increases during the multiphoton absorp-
tion process, i.e., it is especially strong for energies close to
the dissociation threshold D0. To describe schematically the
effect of the energy transfer on MPA and MPD the following
mechanisms are commonly used:
*
*
**
A ϩM →A ϩM activation,
**
A
→products
ϩM→A ϩM
dissociation,
**
*
*
A
deactivation,
*
where A is the high vibrationally excited parent molecule
**
(EϽD0), A is the overexcited parent molecule (EϾD0),
and M is any collision partner ͑parent molecule, product,
buffer gas͒. All these mechanisms are active both during and
after the laser pulse.
FIG. 9. The desactivating action of the buffer gas enhances absorption, i.e.,
storage of vibrational energy, which eventually causes unimolecular decay
via ‘‘energy pooling.’’
Deactivation should play a major role in the course of
the accumulation of vibrational energy in the irradiated spe-
cies, because on the one hand, it makes less absorbers vanish
by dissociation and on the other hand, it enhances absorption
of desactivated molecules owing to their higher absorption
cross section. The given activation mechanism must be of
importance in the post-pulse vibrational relaxation processes
responsible for collision-induced dissociation. It is called
‘‘energy pooling’’ 28 or ‘‘vibrational up-pumping’’.29 It
starts from two ‘‘hot’’ undissociated parent molecules, the
separate vibrational energies of which are not sufficient for
dissociation. Energy pooling means that the stored energy is
now redistributed such that the sum of the vibrational ener-
gies of the colliding molecules and the relative kinetic en-
ergy along the line interconnecting their centres of gravity is
exceeding the dissociation threshold D0 within one molecule
in the moment of closest approach. As a result of some en-
ergy transfer one of the partners gains a high enough energy
to decompose with the unimolecular reaction rate k(E).
Energy pooling has been shown to occur in a number of
ir laser experiments at higher pressures. In Ref. 30 it was
found responsible for continued spontaneous fluorescence of
a product of CO2-laser irradiated SiF4 at pressures in the
range of a few Torr. In Ref. 31 the fluorescent product of the
CO2-laser-induced dissociation of CDCl3 at 1.3 kPa was
identified by LIF to be CCl2. Post-pulse excitation of the
electronically excited state responsible for this fluorescence
required an energy equivalent to 17 ir laser quanta, i.e., the
vibrational energy transferred during a collision may be
much larger than collision-induced dipole selection rules
would allow. While here contributions from the IRMPD
product C2Cl4, which was vibrationally hot due to absorption
of ir radiation, could not be neglected, experiments with
CH2BrCH2F, where none of the products showed absorption,
yielded comparable results. Even for diatomics such as NO,
reported.32 This would correspond to a total energy trans-
ferred of 23 300 cmϪ1
.
Assuming ͑iv͒ to be fulfilled in our case, the ratio of
activating to deactivating collisions can be given using the
detailed balance of the transition probabilities Pij ,Pji for the
levels Ei ,Ej(EiϾEj)
Pji ␳ E ͒
͑
i
ϭ
exp Ϫ E ϪE ͒/kT .
͑7͒
͓
͑
͔
i
j
Pij ␳ E ͒
͑
j
After calculation of the vibrational state densities ␳(E) ac-
cording to Whitten and Rabinovich,33 for CF2HCl a ratio
␳(D0)/␳(D0Ϫh␯)ϭ1.58 is found, where h␯ is an ir quan-
tum. The normalization condition PijϩPjiϭ1 then yields a
contribution for activating ͑dissociative͒ collisions of 34.9%
for the acceptance of a quantum ⌬E ϭ
͉
EiϪEj
͉ϭ1086
͗
͘
cmϪ1 ͑laser photon͒, if a vibrational temperature of 1500 K
corresponding to an average number n ϭ10 of absorbed
͗ ͘
laser photons is assumed. Since the density of states ratio
becomes more favorable for lower energies, this percentage
value gives a lower limit. According to ͑iii͒ the translational
temperature should have no marked effect.
Hence, excited parent molecules of lower internal en-
ergy, which were shown to make up the majority of the
molecules in the focus neighborhood far into the double
cone, can be involved in unimolecular decay. Since this dis-
sociation mechanism is collision-controlled, the measured
net dissociation rate is orders of magnitude lower than the
IRMPD rate, as can be seen from Fig. 3. The measured decay
rate of the vibrational CF2 temperature, which was much
slower than the translational temperature evolution, reflects
the high energy content of the vibrational energy reservoir
with which it interacts.
collision-induced vibrational population of the Љϭ15 level
The effect of an added inert buffer gas shown in Fig. 9
confirms the picture quite impressively. At a 40-fold buffer
v
from the Љϭ1 level excited by a cw CO laser at 12 kPa were
v
J. Chem. Phys., Vol. 105, No. 21, 1 December 1996
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:
129.22.67.107 On: Mon, 24 Nov 2014 11:38:36

Strube et al.: Dissociation of CF2HCl and CF2Cl2
9485
gas surplus ͑16 000 Pa curve͒ dissociation can be completely
suppressed during the ir laser pulse. Since no absorbers are
lost, the total absorbed energy was increased by a factor of
about 2. Subsequently, this stored vibrational energy is redis-
4 J. Wollbrandt, M. Rossberg, W. Strube, and E. Linke, J. Mol. Spectros. 76
385 ͑1996͒.
5 Ber. Bunsenges, Phys. Chem. 99, 231 ͑1995͒.
6 K. Sugawara, A. Watanabe, Y. Koga, H. Takeo, K. Fukuda, J. Hiraishi,
and Ch. Matsumura J. Phys. Chem. 93, 3647 ͑1989͒.
7 P. K. Chowdhury, Ber. Bunsenges, Phys. Chem. 95, 1626 ͑1991͒.
8 D. Krajnovich, F. Huisken, Z. Zhang, Y. R. Shen, and Y. T. Lee, J. Chem.
Phys. 77, 5977 ͑1982͒.
tributed by
–
v v
energy pooling, yielding the dissociative
action shown. A marked effect of –T processes can be
v
ruled out due to the increased heat capacity ͑by 40 times͒.
The effect of a radiative heating of the CF2 product, as it
may occur at the wavelength chosen for IRMPD of CF2HCl,
could not be clarified. The lack of such a product resonance
at the laser wavelength chosen is not the only reason, why
results quantitatively different from CF2HCl were obtained
for the parent molecule CF2Cl2. Typical of CF2Cl2 is also the
small primary production of CF2, the respective molecular
elimination channel is only of minor importance at high flu-
ences ͑branching ratio Ͻ0.12 for ⌽Ͼ6 J/cm2 8͒. The initial
CF2 concentration being so low, the later increase is that
marked, as was shown in Fig. 4͑a͒. It can only be the result
of secondary reactions.
Still controversial is the contribution of the secondary
dissociation of the major primary product CF2Cl, recombin-
ing to our major stable product 1.2-C2F4Cl2. In Ref. 8 sec-
ondary dissociation of a fraction of 20% was estimated at a
fluence of 6 J/cm2 under molecular beam conditions. Accord-
ing to the results of this paper it is likely that under the
9 W. Strube, J. Wollbrandt, M. Rossberg, and E. Linke, Rev. Sci. Instrum.
65, 129 ͑1994͒.
10
¨
K. Reihs, Ph.D. thesis, University of Gottingen, 1989.
11 W. Strube, J. Wollbrandt, M. Rossberg, and E. Linke, in Time-resolved
Vibrational Spectroscopy, edited by A. Lau, F. Siebert, and W. Werncke
͑Springer-Verlag, Berlin, 1994͒, p. 173.
12 D. S. King, and J. C. Stephenson, J. Am. Chem. Soc. 100, 7151 ͑1978͒.
13 M. Rossberg, W. Strube, J. Wollbrandt, and E. Linke, J. Photochem. Pho-
tobiol. A 80, 61 ͑1994͒.
14 J. L. Lyman, S. D. Rockwood, and S. M. Freund, J. Chem. Phys. 67, 4545
͑1977͒.
15 D. J. Krajnovich, C. Parmenter, and D. L. Catlett, Chem. Rev. 87, 237
͑1987͒.
16 I. Oref and D. C. Tardy, Chem. Rev. 90, 1407 ͑1990͒.
17 J. D. McDonald, Ann. Rev. Phys. Chem. 30, 29 ͑1979͒.
18 V. N. Bagratashvili, V. N. Burimov, M. V. Kuz’min, and A. P. Sviridov,
Preprint 3/1983 ͑Academy of Sciences, USSR, Troitsk, 1983͒.
19 J. C. Stephenson, J. Chem. Phys. 70, 4496 ͑1979͒.
20 D. S. King, Adv. Chem. Phys. 50, 105 ͑1982͒.
21 J. W. Edwards and P. A. Small, Nature ͑London͒ 202, 1329 ͑1964͒.
22
¨
K. P. Schug, Ph.D. thesis, University of Gottingen, 1976.
23 D. S. Y. Hsu, M. E. Umstead, and M. C. Lin, ACS Symp. Series 66, 128
͑1978͒.
present conditions an intense
–
v v
exchange with the hot
parent molecules renders the fragment highly vibrationally
excited, should the nascent CF2Cl not already be hot due to
energy partitioning in the IRMPD process. In any case, this
situation would promote further vibrational heating by ab-
sorption of ir radiation, although the linear absorption band
at 1149 cmϪ1 is off-resonance from our laser wavelength.
Then the situation is comparable to that for CF2HCl, i.e.,
energy pooling explains the qualitatively similar results for
CF2Cl2.
24 G. W. Flynn, Acc. Chem. Res. 14, 334 ͑1981͒.
25 H. Hippler and J. Troe, in Advances in Gas-Phase Photochemistry and
Kinetics, edited by M. N. R. Ashfold and J. E. Baggott ͑Royal Society of
Chemistry, London, 1989͒, Vol. 2, p. 209.
26 M. Quack and J. Troe, in Gas Phase Kinetics and Energy Transfer, edited
by P. G. Ashmore and R. J. Donovan ͑A Specialist Periodical Report͒
͑Chem. Soc., London, 1977͒, Vol. 2, p. 175.
27 C. D. Cantrell, S. M. Freund, and J. L. Lyman, in Laser Handbook, edited
by M. L. Stitch ͑North Holland, Amsterdam, 1979͒, Vol. IIIb, p. 529.
28 I. Oref, J. Chem. Phys. 75, 131 ͑1981͒.
29 G. A. McRae, A. B. Yamashita, and J. W. Goodale, J. Chem. Phys. 93,
5997 ͑1990͒.
1 M. Quack, Infrared Phys. 29, 441 ͑1989͒; Lasers in Chemistry, Special
Issue of Proc. Indian Acad. Sci. ͑Chem. Sci.͒ 103, 201 ͑1991͒; Proc. 2nd
30 V. E. Merchant, Opt. Commun. 25, 259 ͑1977͒.
31 M. Ivanco, G. A. McRae, R. A. Back, J. W. Goodale, and P. E. Lee, J.
Chem. Phys. 96, 5191 ͑1992͒.
Conf. Lasers in Chemistry, Trieste ͑Italy͒, Nov. 1993.
2
¨
J. Forster, M. v. Hoesslin, and J. Uhlenbusch, Appl. Phys. B 62, 609
32 S. Saupe, O. Schulz, W. Urban, I. Adamovich, M. J. Grassi, and J. W.
Rich, Verhandlungen Dt. Phys. Ges. 3, 362 ͑1993͒.
͑1996͒.
3
¨
J. Gothel, M. Ivanenko, P. Hering, W. Fuss, and K. L. Kompa, Appl.
Phys. B 62, 329 ͑1996͒.
33 G. Z. Whitten and B. S. Rabinovitch, J. Chem. Phys. 38, 2466 ͑1963͒.
J. Chem. Phys., Vol. 105, No. 21, 1 December 1996
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:
129.22.67.107 On: Mon, 24 Nov 2014 11:38:36