Infrared frequency-modulation measurements of absolute rate coefficients for Cl+HD→HCl(DCl)+D(H) between 295 and 700 K

Article abstract of DOI:10.1016/S0009-2614(99)00420-0

Absolute rate coefficients and HCl branching fractions for the Cl+HD reaction have been measured. The rate coefficient for Cl+HD is fit well by k₁=(1.28±0.18)×10^-16 T^1.67 e^{-(1646±60)/T} cm³ molecule

Full text of DOI:10.1016/S0009-2614(99)00420-0

4 June 1999
Ž
.
Chemical Physics Letters 306 1999 33–40
Infrared frequency-modulation measurements of absolute rate
ž
/
ž /
coefficients for ClqHD™HCl DCl qD H between 295 and
700 K
)
Craig A. Taatjes
Combustion Research Facility, Mail Stop 9055, Sandia National Laboratories, LiÕermore, CA 94551-0969, USA
Received 12 March 1999; in final form 2 April 1999
Abstract
Absolute rate coefficients and HCl branching fractions for the ClqHD reaction have been measured. The rate coefficient
y
for ClqHD is fit well by k1s 1.28"0.18 =10 16 T1.67 eyŽ1646" 60.r T cm3 moleculey1 sy1, approximately 30% smaller
than the mean of literature rate coefficients for ClqH2 and D2. The branching fraction into the HCl channel is 0.67"0.09
at 295 K and decreases at higher temperature. The present results agree well with literature measurements of relative rates,
confirming different isotope effects in thermal conditions than in molecular beam measurements. q 1999 Elsevier Science
B.V. All rights reserved.
Ž
.
1. Introduction
tion and allows a measurement of the intramolecular
kinetic isotope effect k1ark1b:
The reaction of chlorine atom with hydrogen is a
prototypical three-atom reactive system. This reac-
tion and its isotopic variant ClqD2 have been
k1
a
ClqHD
™
HClqD
DClqH
2
Ž .
k1
w x
b
extensively studied both experimentally 1 and theo-
™
w x
retically 2 . However, the reaction of Cl with the
mixed isotope species HD,
Persky and Klein measured relative rates for ClqH2 ,
ClqD2 , and ClqHD between 235 and 340 K in
k1
ClqHD ™ products
has been less thoroughly investigated. The mixed
isotope reaction provides a further check on the
accuracy of the theoretical understanding of the reac-
1
Ž .
w x
1966 3 , but until now there has been no absolute
measurement of the ClqHD rate coefficient. In
addition, the intramolecular kinetic isotope effect has
w
x
been measured 4,5 . An additional impetus for in-
vestigating the reaction of the mixed isotope is the
recent activity on understanding the dynamics of
these systems. The dynamics of the ClqHD reac-
tion have been studied by Bersohn and coworkers
)
Corresponding author. Fax: q1 925 294 2276; e-mail:
cataatj@ca.sandia.gov
w x
5 , and more recently by Liu in a crossed molecular
0009-2614r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S0009-2614 99 00420-0

(
)
34
C.A. TaatjesrChemical Physics Letters 306 1999 33–40
w x
beam experiment 6 . Interestingly, molecular beam
measurements of intramolecular kinetic isotope ef-
fect tend to differ with thermal kinetic measure-
ments.
beam measurements of several radicalqHD reac-
w
x
tions 6,13–15 . There are systematic differences
between thermal kinetic measurements and predicted
rate coefficients based on cross-sections in molecular
beam scattering experiments. Measurements of the
reactive cross-section as a function of energy for
Reactions of radical species with HD provide an
important test for theoretical understanding of simple
reactions since the primary and secondary kinetic
isotope effects can be separated in the mixed isotope
species. Rate coefficients and intramolecular kinetic
isotope effects of several radical qHD abstraction
reactions have been measured. The reaction of OHq
HD has recently been investigated by Ravishankara
and coworkers, who observed that the rate coeffi-
cient for OHqHD was identical to the mean of the
rate coefficients for OHqH2 and OHqD2 over the
Ž1
.
S D qH2 , D2 , and HD show a trend which is
contrary to kinetic intuition, with sHD )sD )sH
2
2
w x
6 . In contrast, kinetic experiments on reactions of a
number of radicals with hydrogen isotopomers show
reaction rate coefficients for HD which are near the
mean of those for H2 and D2. Differing intramolec-
ular kinetic isotope effects are also expected depend-
ing on whether an insertion or an abstraction mecha-
w
x
nism dominates 16 . Abstractions from HD under
thermal conditions prefer to remove an H atom, i.e.,
XqHD™HXqD, displaying a normal kinetic iso-
tope effect. Reactions which proceed by an insertion
mechanism to form a transient complex show an
inverse kinetic isotope effect under thermal condi-
tions. In a molecular beam, this situation is reversed
w x
temperature range 230–420 K 7 . They also found
that the intramolecular kinetic isotope effect was
well described by assuming the reaction with HD to
be the simple sum of independent H- and D-atom
abstractions, i.e., the ratio of the production of H2O
to the production of HDO in the OHqHD reaction
was the same as the ratio of rate coefficients for
OHqH2 and OHqD2. The abstraction reaction of
O P with HD was measured by Presser and Gordon
using flash photolysis resonance fluorescence meth-
ods 8 . In this case the reaction with HD was again
found to match the arithmetic mean of the rate
coefficients for H2 and D2. Subsequent measure-
ments of the OHrOD branching fractions using a
discharge-flow apparatus found intramolecular ki-
netic isotope effects in disagreement with the sum of
independent abstractions, with the disagreement in-
w x
6 .
The discrepancy between molecular beam mea-
surements and thermal kinetic measurements of the
intramolecular kinetic isotope effect is thought to be
partially a consequence of HD rotation. This phe-
nomenon can occur because of the differing rota-
tional populations in the thermal and molecular beam
experiments. It has been proposed, in connection
with the FqHD reaction, that a kinematic prefer-
ence for D-end attack operates at low JHD because
the center of charge in the molecule is moved away
Ž3
.
w x
w x
w
x
creasing at higher temperatures 9 . Most recently the
reactions of CN with H2 , D2 and HD were measured
from the center of mass 17 . This effect is removed
by HD rotation, with the preference for H atom
abstraction dramatically increasing with increasing
Ž
.
using infrared IR absorption techniques by Mac-
Donald and co-workers, and compared to transition
state theory calculations on an ab initio potential
energy surface 10,11 . In most of these reactions,
the thermal rate coefficient for XqHD is very near
the mean of the rate coefficients for XqH2 and
XqD2 , suggesting that the leaving atom plays little
role in the reactivity.
The dynamics of the reactions of HD with atoms
and small radicals reflect effects of reagent rotation
and vibration on the course of the reaction. The
reaction of F atoms with isotopic hydrogen species
has become a benchmark reaction 12 . The Liu
group has recently undertaken crossed molecular
JHD
.
The reaction of ClqHD is an excellent target for
further investigations of such effects. Dynamical in-
formation on the Clqhydrogen reaction is becom-
ing available from exquisite molecular beam scatter-
ing experiments 2,18 . Beam experiments performed
in the Liu group found HD to be far less reactive
than H2 , and only slightly more reactive than D2
19 , in contrast to the thermal relative rate measure-
ments of Persky and Klein between 235 and 340 K
3 . The H2Cl system has been the object of recent
theoretical attention; a new potential energy surface
is now available, and quantum-mechanical calcula-
w
x
w
x
w
x
w x
w
x

(
)
C.A. TaatjesrChemical Physics Letters 306 1999 33–40
35
tions of rate coefficients and kinetic isotope effects
The cw probe beam is passed through the reactor
multiple times using a cell based on the Herriott
w
x
have been reported 20,21 . The effects of vibration
and rotation on the reaction ha&ve been calculated by
several groups. Aoiz and Banares found that the
scattering of Cl with HD to form HCl and D is
w
x
design 27 . A Herriott cell uses off-axis paths in a
spherical resonator. By propagating the pump beam
along the resonator axis, the pump–probe overlap is
confined to the center of the cell, where the tempera-
Ž
.
vibrationally adiabatic, with HD Õs0 producing
Ž
.
ture is well-controlled "3 K . The multipass cell is
predominantly Õs0 and Õs1 producing mainly
w x
described in detail in a previous publication 27 . In
w
x
Õs1 22 . Song and Gislason have investigated
rotational and vibrational effects in the ClqH2 and
ClqHD reactions, confirming the orientational ef-
fect favoring DCl production at low J and high
Ž
.
addition, two-tone frequency modulation FM is
employed in the present experiments to reduce the
contributions of laser amplitude noise and thermal
w
x
w
x
lensing noise 26,28 .
kinetic energy 23 .
The reactor is a stainless-steel vessel approxi-
mately 1 m in length. The central 60 cm is enclosed
by three resistive heaters under individual micropro-
cessor control using temperature feedback from three
separate thermocouples placed along the length of
In the present study absolute measurements of the
rate coefficient for Cl reacting with HD are pre-
sented. The yield of HCl in the reaction is measured,
which allows the intramolecular kinetic isotope ef-
fect to be estimated. The present results can be
combined with the established literature values for
the reaction with H2 and D2 to extract kinetic iso-
tope effects. The absolute measurements are in rela-
tively good agreement with the relative measure-
Ž
.
the reactor. Pure CF2Cl2 99.7% is used as the
photolyte in all experiments except the measurement
Ž
.
at 700 K, where CCl4 neat was also used. The
concentration of photolyte in these experiments is
typically 2=1014 cmy3, and the initial Cl atom
concentration is 2–3=1012 cmy3. The reaction of H
atoms with CF2Cl2 and with CCl4 should be quite
slow, in analogy with the HqCF3Cl reaction, which
has an activation energy of nearly 9 kcal moly1 and
a rate coefficient of 4=10y14 cm3 moleculey1 sy1
ments of Persky and Klein, showing kHDrkD
dis2crepancy as to the temperature dependence of
)
2
kH rkHD below 400 K, although there is a slight
kH rkHD. Theoretical predictions for the kinetic iso-
2
tope effects are at larger variance with the experi-
mental values, suggesting that additional considera-
tion of possible rotational effects may be necessary
to correctly predict the thermal rate coefficients for
ClqHD.
w
x
at 676 K 29 . Secondary formation of HCl by
reactions of H atoms with the photolyte would ap-
pear as a bi-exponential time behavior of the HCl
signal. No bi-exponential behavior was observed in
these experiments.
Ž
.
Ar 99.9999% serves as the buffer gas in the
2. Experiment
Ž
experiments reported here. The gases CF2Cl2 , Ar,
and HD are flowed through the cell at a rate suffi-
.
The kinetics of the ClqHD reaction have been
cient to replenish the cell between laser shots, but
remain essentially static on the timescale of the
kinetics. The individual gas flows are controlled by
separate calibrated mass flow controllers, and the
pressure in the reactor is actively controlled by a
butterfly valve which changes the pumping speed to
maintain the desired pressure for a given flow.
The reaction of Cl with HD is slightly endother-
investigated using the technique of laser photo-
Ž
.
lysisrcontinuous-wave cw IR frequency-modula-
Ž
.
tion FM spectroscopy. The apparatus used is simi-
w
x
lar to that previously reported 24–26 . Chlorine
atoms are produced by photolysis of CF2Cl2 or CCl4
at 193 nm, and the progress of the reaction is
followed in time by frequency modulation spec-
Ž .
w
x
troscopy using a cw IR probe laser tuned to the R 2
mic 30 , and the contribution of vibrationally ex-
y1
Ž
line of the fundamental vibrational transition in the
H35Cl product, which provides sufficient signal-to-
noise at all temperatures of the present study. Signal
averaging is used to improve signal-to-noise ratios.
cited HCl product which requires ;9 kcal mol
.
of energy is expected to be negligible. Comparison
of HCl absorption traces using Ar with those using
CO2 as a buffer which relaxes HCl Õs1 effi-
Ž
Ž
.

(
)
36
C.A. TaatjesrChemical Physics Letters 306 1999 33–40
. w
x
ciently 31 showed no evidence of vibrational exci-
Ž
tation at the highest temperatures in this study 700
.
K .
The IR light is produced by difference-frequency
Ž
.
generation DFG in LiNbO3, using the output of a
ring dye laser and a single-frequency Arq laser. The
dye laser beam passes through a phase modulator,
which is driven at two closely spaced radio frequen-
Ž
.
cies RF , 499"0.9 MHz. The RF power applied to
Ž
.
the resonant phase modulator 2–3 W is sufficient
to generate a peak phase shift of approximately 1
radian; the power is adjusted to produce the highest
FM signal. The periodic phase shift on the dye laser
beam is carried onto the generated IR beam, and is
equivalent to adding sidebands spaced 498.1 and
499.9 MHz from the central carrier. Differential
absorption of the sidebands relative to the carrier
generates a signal at the frequency corresponding to
the sideband spacing and at the beat frequency be-
tween them, 1.8 MHz. The signal at the intermodula-
tion frequency of 1.8 MHz is demodulated and the
resulting DC signal is digitized, averaged over sev-
eral hundred photolysis laser shots, and stored. This
technique has been described more fully in other
Fig. 1. A typical time-resolved FM measurement of HCl produc-
tion in the ClqHD reaction, taken at 293 K with an HD concen-
tration of 1.03=1015 molecule cmy3
.
where k1 'k1a qk1b, and kx represents all Cl loss
processes other than reaction with HD, including
contributions from Cl reactions with contaminants in
the photolyte or the buffer and diffusion of Cl from
the reaction region. The rate coefficient ky similarly
Ž
represents the removal of HCl principally by diffu-
w
x
publications 26,32 .
.
sion . The HCl time profile is then given by
The time profiles for the production of HCl from
the Cl atomqHD reaction are well described by
simple exponential behavior. Because of the small
rate coefficient for the reaction, the pseudo-first or-
der rate coefficients are relatively low at the pres-
sures of HD used in these experiments, especially at
the lowest temperatures. This is shown in Fig. 1,
where a typical absorption trace at room temperature
is shown. Fitting such a trace requires consideration
of the decay of the HCl concentration, which is
principally due to radial diffusion under the condi-
tions of these experiments. Although radial diffusion
is not expected to be strictly first order, the deviation
of the HCl decay from a single exponential is negli-
gible. The kinetic equations used to describe the
experiment are therefore simple:
w x
w
x
Cl 0 k1a HD
w
x
HCl s
t
w
x
k1 HD qkx yky
w
1
x
HD qk x t
.
= eyk t yey
.
4
k
Ž
y
Ž
.
Ž .
The time traces are therefore fit to the difference of
two exponentials, where the rise time is related to the
production rate of HCl, since in these experiments
w
x
k1 HD )ky. A plot of this time constant versus the
HD concentration gives a straight line whose slope is
k1 and whose intercept reflects the losses of Cl
atoms that do not depend on the HD concentration.
Such a plot is shown in Fig. 2 for the reaction of
ClqHD at 500 K.
The measurements of the HCl yield, fHCl
'
Ž
w
x
w x
k1ark1 s HCl `r Cl when competing reactions
0
.
can be ignored , is performed by comparison with
the Clqpropane or Clqethane reaction, where
fHCl s1. A measurement is made of the time profile
of HCl formation from the reaction of Cl with HD.
Immediately thereafter, ethane or propane is substi-
tuted for the HD, keeping the other conditions identi-
cal. The amplitude of the FM signal is directly
d
w x
w
xw x
w x
Cl syk1 HD Cl ykx Cl ,
dt
d
w
x
w
xw x
w
x
HCl skla HD Cl yky HCl ,
3
Ž .
dt

(
)
C.A. TaatjesrChemical Physics Letters 306 1999 33–40
37
Table 1
Rate coefficients and HCl yields for the reaction of Cl with HDa
Temperature
kClqHD
fHCl
3
cm moleculey1 sy1
Ž .
K
Ž
.
y15
Ž
Ž
Ž
Ž
.
.
.
.
.
.
.
293
350
400
450
500
550
705
6.4"0.6 =10
0.67"0.09
0.63"0.09
y14
y14
y14
y13
y13
y13
2.1"0.2 =10
4.5"0.4 =10
9.3"0.9 =10
1.59"0.14 =10
2.3"0.2 =10
6.9"0.7 =10
0.5"0.1
0.56"0.08
0.5"0.1
Ž
Ž
Ž
a Error bars represent "2s and include estimates of systematic
errors.
played in Fig. 3. The rate coefficients for ClqH2
and ClqD2 , taken from the evaluation of Kumaran
Fig. 2. A plot of the pseudo-first order rate coefficients versus HD
concentration, taken at 500 K. The second-order rate coefficient is
the slope of such a plot. The intercept of ;65 sy1 reflects all
other loss mechanisms for Cl atoms, including diffusion and
reaction with photolyte and buffer gas impurities.
w x
et al. 1 , are also shown for reference. In several
other abstraction reactions with HD, e.g., OH and
Ž1
.
O D , the thermal rate coefficient has been seen to
be very close to the average of the rate coefficients
with H2 and with D2. This would imply that the
departing atom has little effect on the abstraction rate
constant. In Fig. 3, the average of the evaluated rate
coefficients for ClqH2 and ClqD2 is shown as
proportional to the absorption strength, so the size of
the observed HCl signal is proportional to the HCl
produced. The traces are fitted, and the ratio of the
amplitude with HD to that with the alkane reactant
Ž .
gives the HCl yield. Consideration of Eq. 4 shows
that the contributions of competing reactions and the
removal rate of the product must be taken into
account in comparing amplitudes of product time
w x
traces 7 . Ideally, yield measurements would be
w
x
x
made at relatively high values of HD , so that the
x Ž
w
x
x
w
x
w
.
amplitude Cl k1 a HD r k1 HD q kx y k y f
0
w
w
x
w
Cl k1a HD rk1 HD . However, since the rate coef-
0
ficient for Cl reacting with HD is so small, this
condition is only obtained at impracticably high HD
concentrations. Therefore, in the present work the
Ž .
HCl yields are corrected using the model of Eq. 4 .
For most of the present measurements, the correc-
tions to the yield for competing reactions and HCl
removal are much smaller than other sources of
error, chiefly frequency and amplitude instability of
the Arq laser and power fluctuations of the excimer
laser.
Fig. 3. Rate coefficients for the reactions of Cl atom with hydro-
gen molecule isotopomers. The dashed line is the rate coefficient
for ClqH2 , and the dot–dashed line that for ClqD2 , both taken
w x
from the evaluation of Ref. 1 . The mean of the evaluated
ClqH2 and ClqD2 rate coefficients is shown as the thin solid
line. The present measurements of the ClqHD rate coefficient,
shown as the individual points with "2s error bars, fall signifi-
cantly below the mean of the ClqH2 and ClqD2 rate coeffi-
3. Results
cients. The fitted temperature dependence of the ClqHD reaction,
3
k1s 1.28"0.18 =10y16 T1.67 exp y 1646"60 rT cm mole-
The measured rate coefficients for the reaction of
Cl atoms with HD are given in Table 1 and dis-
Ž
.
w
Ž
.
x
culey1 sy1, is shown as the dotted line.

(
)
38
C.A. TaatjesrChemical Physics Letters 306 1999 33–40
isotope effects for Cl reacting with the various iso-
topomers of hydrogen have been measured by
w x
Persky and Klein 3 . Fig. 4 shows the results of
Persky and Klein along with the kinetic isotope
effects calculated from the present rate coefficient
measurements combined with the recommended val-
w x
ues for ClqH2 and ClqD2 from Ref. 1 . The
agreement with the earlier determinations is reason-
able, but there appears to be a larger discrepancy
with the H2rHD isotope effect than with the HDrD2
determination. The kinetic isotope effects decrease
with increasing temperature, and an Arrhenius fit to
the calculated isotope effects yields kH rkHD
s
2 Ž
1.75"0.24 eyŽ80" 30.r T, and kHDrkD s 0.66"
Ž
.
2
0.09 eyŽ490" 60.r T. As can be seen in Fig. 4, the
Fig. 4. Kinetic isotope effects for the Clqhydrogen system.
.
Circles represent kHD rkD , and squares represent kH rkHD
.
2
2
slope of the Arrhenius plot for kH rkHD is signifi-
2
Ž
.
The open data points `, I are the present measurements,
cantly smaller than that from the lower temperature
combined with the evaluated H2 and D2 rate coefficients from
w x
results of Ref. 3 , but the agreement is better for
w x
Ž
.
Ref. 1 . The filled data points v, B are kinetic isotope effect
w x
kHDrkD
.
measurements from Ref. 3 . Calculated kinetic isotope effects
2
w
x
Ž
.
from Ref. 20 are also shown as the dashed line kH rkHD and
the dot–dashed line kHD rkD
The intramolecular kinetic isotope effect of the
2
Ž
.
.
2
ClqHD reaction has also previously been measured
w x
4 . In the present work the relative yields of HCl,
fHCl, have been measured at several temperatures.
the thin solid line. The present results are signifi-
cantly smaller than the average of the H2 and D2
rate coefficients, indicating a secondary kinetic iso-
tope effect of the leaving atom on the rate coeffi-
cient.
This quantity is straightforwardly related to the ki-
Ž
netic isotope effect, hintra 'k1ark1b 'fHClr 1y
.
fHCl . The present results for fHCl are displayed in
Fig. 5 along with HCl yields calculated from avail-
able literature measurements of the thermal intra-
The rate coefficients for ClqHD can be fitted
nearly to within the experimental error estimates
Ž
.
maximum deviation of 20% by a simple Arrhenius
function k1 s1.8=10y11 ey2370r T cm3 moleculey1
sy1. However, significant curvature is seen in an
Arrhenius plot of the corresponding ClqH2 and
ClqD2 reactions, with fitted temperature exponents
of 1.72 and 1.62, respectively, and a similar curva-
ture would be expected in ClqHD. A slightly im-
2
Ž
.
proved fit reduction in x of a factor of 3 is given
by fixing a T1.67 dependence to the pre-exponential
factor for ClqHD, yielding the expression k1 s
1.28"0.18 =10y16 T1.67 eyŽ1646" 60.r T cm3 mole-
Ž
.
culey1 sy1, where the error bars reflect the precision
Ž
.
"2s of the fit. The temperature exponent is cho-
sen as the average of the exponents in the evaluated
Fig. 5. HCl yield in the reaction of Cl with HD, fHCl 'kla rk1.
w x
H2 and D2 rate coefficients 1 , and is fixed to
reduce parameter correlation.
Ž
.
The present measurements are shown as the open circles ` . The
w x
measurements of Bar Yaakov et al. 4 are shown as open squares
While no previous absolute measurements exist
for the rate coefficient of Cl with HD, the kinetic
Ž
.
w x
Ž .
I and the results of Xing et al. 5 as diamonds l . HCl yields
w
x
from the ab initio calculations of Ref. 20 are denoted by an =.

(
)
C.A. TaatjesrChemical Physics Letters 306 1999 33–40
39
molecular kinetic isotope effect. The error estimates
of initial rotational state dependence in the ClqH2
Ž
.
w
x
are "2s and are based on several 3–5 measure-
ments at each temperature and include a considera-
tion of possible systematic errors. While the preci-
sion in the present system is inferior to the relative
reaction address this directly 19 . Quantum mechan-
ical calculations have recently been performed for
w
x
the ClqH2 system 20,34 , and more detailed calcu-
lations of ClqHD may also be desirable. At higher
temperatures the kinetic isotope effects become com-
parable, and at the highest temperatures of the pre-
w x
rate studies of Bar Yaakov et al. 4 , the present HCl
yields are in reasonable agreement with their mea-
surements as well as those of Bersohn and co-workers
sent study, the absolute rate data suggest kH rkHD
2
w x
5 .
slightly greater than kHDrkD . The error bars on the
2
ratio of the absolute rate coefficients reflect the
present experimental uncertainties and the stated pre-
cision of the evaluation of the ClqH2 and ClqD2
4. Discussion
w x
rate coefficients in Ref. 1 . Further experimental
The absolute measurements of the rate coeffi-
cients for the reaction of ClqHD are in reasonably
good agreement with those predicted from relative
rate measurements. The kinetic isotope effect for HD
versus D2 from the absolute measurements agrees
both in magnitude and temperature dependence with
the determinations of Persky and Klein. The ratio of
absolute determinations for H2 versus HD has a
slightly shallower temperature dependence than the
Persky and Klein result, while the magnitude of the
effect near room temperature is in good agreement.
This slight disagreement between the Persky and
Klein results and the absolute measurements also
appears in the kinetic isotope effect kH rkD , which
investigations, including relative rate measurements
at higher temperatures, would help confirm the tem-
perature dependence of the kinetic isotope effects.
Molecular beam measurements of the excitation
function have shown sH rsHD 4sHDrsD , with
the reactive cross-section fo2 r ClqHD similar2 to that
for ClqD2 , and much below ClqH2 19 . Molecu-
lar beam scattering experiments also show a much
higher branching for the DCl product than thermal
energy experiments. The effect of the mass asymme-
try on the reactions of HD is well known, and in a
number of systems, including ClqHD, calculations
show that the D-atom is preferentially abstracted at
low rotational energy, especially at the higher colli-
w
x
2
2
w x
w
x
links the two HD isotope effects 1 . The Persky and
Klein value is also slightly higher than that of Miller
sion energies of molecular beam experiments 23 .
Such a picture is broadly consistent with the differ-
ences in both intramolecular and kinetic isotope ef-
fects in the ClqHD reaction. Whereas the thermal
kinetic isotope effect favors the hydrogen abstraction
by nearly a factor of two, abstraction of a deuterium
atom dominates the beam reaction by a factor of
w
x
and Gordon 33 , and shows a steeper dependence on
temperature.
A new potential energy surface has recently be-
come available for the H2Cl system, and rate coeffi-
cients and kinetic isotope effects have been calcu-
lated using a canonical variational transition state
w x
three 6 . The cross-section for the beam reaction of
w
x
theory method 20 . A comparison of the results of
the present experiments with the predictions of the
theoretical calculations shows reasonable agreement
as to the general magnitude of the kinetic isotope
effects, but significant discrepancies remain. The
HD may be relatively reduced because of the higher
contribution of D-end collisions whose only reactive
pathway is the presumably less favorable D-atom
transfer.
The present results for the ClqHD reaction high-
light some discrepancies of the most sophisticated
calculations with experimental kinetics measure-
ments. The reaction of Cl with hydrogen has gained
the status of a prototypical reaction because of the
extensive theoretical and experimental work which
calculations predict a kinetic isotope effect kH rkHD
2
larger than kHDrkD , whereas the experimental data
2
show that the opposite is the case up to 400 K. The
semi-classical variational transition state theory
method is known to overestimate the barriers for
w
x
w x
high bend states 20 , and perhaps some rotational
effect may underlie the discrepancy between theory
and experiment. Recent experimental examinations
exists on its dynamics and kinetics 2 . Nonetheless,
a consensus on this reaction has not yet been at-
tained. The present absolute rate coefficient measure-

(
)
40
C.A. TaatjesrChemical Physics Letters 306 1999 33–40
w
w
w
x
10 G. He, I. Tokue, L.B. Harding, R.G. Macdonald, J. Phys.
ments for the ClqHD reaction may aid in further
theoretical refinements or in a possible reinvestiga-
tion of the potential energy surface.
Ž
.
Chem. A 102 1998 7653.
x
11 G. He, I. Tokue, R.G. Macdonald, J. Phys. Chem. A 102
Ž
.
1998 4585.
x
Ž
.
12 D.E. Manolopoulos, J. Chem. Soc. Faraday Trans. 93 1997
673.
w
w
w
x
Ž
.
.
13 D.C. Che, K.-P. Liu, J. Chem. Phys. 103 1995 5164.
Acknowledgements
x
Ž
14 Y.-T. Hsu, K.-P. Liu, J. Chem. Phys. 107 1997 1664.
x
Ž
.
15 Y.-T. Hsu, J.-H. Wang, K.-P. Liu, J. Chem. Phys. 107 1997
2351.
The author thanks Leonard Jusinski for expert
technical assistance in performing these experiments
and Prof. Kopin Liu for valuable discussions on the
dynamics of the ClqHD reaction. This work is
supported by the Division of Chemical Sciences, the
Office of Basic Energy Sciences, US Department of
Energy.
w
x
16 K. Tsukiyama, B. Katz, R. Bersohn, J. Chem. Phys. 83
Ž
.
1985 2889.
w
x
17 G.W. Johnston, H. Kornweitz, I. Schechter, A. Persky, B.
Ž
.
Katz, R. Bersohn, R.D. Levine, J. Chem. Phys. 94 1990
2749.
w
x
18 M. Alagia, N. Balucani, P. Casavecchia, D. Stranges, G.G.
Ž
.
Volpi, J. Chem. Soc. Faraday Trans. 91 1995 575.
w
w
x
x
19 K.-P. Liu, personal communication, 1999.
20 T.C. Allison, G.C. Lynch, D.G. Truhlar, M.S. Gordon, J.
Ž
.
Phys. Chem. 100 1996 13575.
w
x
21 S.L. Mielke, T.C. Allison, D.G. Truhlar, D.W. Schwenke, J.
Ž
.
Phys. Chem. 100 1996 13588.
References
&
w
w
w
x
Ž
.
22 F.J. Aoiz, L. Banares, Chem. Phys. Lett. 247 1995 232.
23 J.-B. Song, E.A. Gislason, Chem. Phys. 237 1998 159.
24 J.S. Pilgrim, C.A. Taatjes, J. Phys. Chem. A 101 1997
x
Ž
.
w x
1
S.S. Kumaran, K.P. Lim, J.V. Michael, J. Chem. Phys. 101
x
Ž
.
.
Ž
.
1994 9487.
4172.
w x
2
M. Alagia, N. Balucani, L. Cartechini, P. Casavecchia, E.H.
w
x
Ž
25 J.S. Pilgrim, C.A. Taatjes, J. Phys. Chem. A 101 1997
&
van Kleef, G.G. Volpi, F.J. Aoiz, L. Banares, D.W.
8741.
Schwenke, T.C. Allison, S.L. Mielke, D.G. Truhlar, Science
w
w
x
Ž
.
26 J.T. Farrell, C.A. Taatjes, J. Phys. Chem. A 102 1998 4846.
Ž
.
273 1996 1519.
x
27 J.S. Pilgrim, R.T. Jennings, C.A. Taatjes, Rev. Sci. Instrum.
w x
Ž
.
3
w x
4
A. Persky, F.S. Klein, J. Chem. Phys. 44 1966 3617.
Ž
.
68 1997 1875.
Y. Bar Yaakov, A. Persky, F.S. Klein, J. Chem. Phys. 59
w
w
x
x
Ž
.
28 D.E. Cooper, R.E. Warren, J. Opt. Soc. Am. B 4 1987 470.
29 J.N. Bradley, D.A. Whytock, T.A. Zaleski, J. Chem. Soc.
. Ž
Ž
.
1973 2415.
w x
5
G. Xing, X. Huang, R. Bersohn, K. Tsukiyama, B. Katz, J.
Ž
.
Faraday Trans. 1 72 1976 2284.
Ž
.
Chem. Phys. 102 1995 3169.
w
x
30 J. Berkowitz, G.B. Ellison, D. Gutman, J. Phys. Chem. 98
w x
Ž
.
6
S.-H. Lee, K.-P. Liu, Chem. Phys. Lett. 290 1998 323.
Ž
.
1994 2744.
w x
7
R.K. Talukdar, T. Gierczak, L. Goldfarb, Y. Rudich, B.S.
w
w
x
x
Ž
.
31 S.R. Leone, J. Phys. Chem. Ref. Data 11 1982 953.
32 P.E. Powers, C.A. Taatjes, T.J. Kulp, Appl. Opt. 35 1996
Ž
.
Madhava Rao, A.R. Ravishankara, J. Phys. Chem. 100 1996
Ž
.
3037.
4735.
w x
Ž
.
8
w x
9
N. Presser, R.J. Gordon, J. Chem. Phys. 82 1985 1291.
w
w
x
Ž
.
33 J.C. Miller, R.J. Gordon, J. Chem. Phys. 79 1983 1252.
D.C. Robie, S. Arepalli, N. Presser, T. Kitsopoulos, R.J.
x
34 H.B. Wang, W.H. Thompson, W.H. Miller, J. Chem. Phys.
Ž
.
Gordon, J. Chem. Phys. 92 1990 7382.
Ž
.
107 1997 7194.