Measuring gas-phase chlorine atom concentrations: Rate constants for Cl + HN3, CF3I, and C2F5I

Article abstract of DOI:10.1021/jp972118g

A convenient method is reported for the measurement of gas-phase Cl atom concentrations in the 10¹¹-10¹³ atoms cm^-3 range in a flow reactor. The titration reaction with vinyl bromide is used to measure the absolute Cl atom concentration, and the relative Cl atom concentration is monitored by the HCl(ν) infrared emission intensity produced by the reaction of Cl atoms with hydrogen sulfide. A microwave discharge through dilute flows of Cl₂, CCl₄, CFCl₃, and CF₂Cl₂ in Ar were characterized as sources of Cl atoms in the flow reactor. The elementary rate constants for reaction of Cl atoms with HN₃, CF₃I, and C₂F₅I were measured as (1.1 ± 0.3) × 10^-12, (5.1 ± 1.5) × 10^-13, and (3.9 ± 0.8) × 10^-12 cm3 molecule^-1 s^-1, respectively, at 300 K. These rate constants were obtained after adjustment of the apparent rate constants for secondary reactions.

Full text of DOI:10.1021/jp972118g

J. Phys. Chem. A 1998, 102, 153-159
153
Measuring Gas-Phase Chlorine Atom Concentrations: Rate Constants for Cl + HN3, CF3I,
and C2F5I
G. C. Manke II and D. W. Setser*
Department of Chemistry, Kansas State UniVersity, Manhattan, Kansas 66506
ReceiVed: July 1, 1997; In Final Form: October 20, 1997X
A convenient method is reported for the measurement of gas-phase Cl atom concentrations in the 1011-1013
atoms cm-3 range in a flow reactor. The titration reaction with vinyl bromide is used to measure the absolute
Cl atom concentration, and the relative Cl atom concentration is monitored by the HCl(V) infrared emission
intensity produced by the reaction of Cl atoms with hydrogen sulfide. A microwave discharge through dilute
flows of Cl2, CCl4, CFCl3, and CF2Cl2 in Ar were characterized as sources of Cl atoms in the flow reactor.
The elementary rate constants for reaction of Cl atoms with HN3, CF3I, and C2F5I were measured as (1.1 (
0.3) × 10-12, (5.1 ( 1.5) × 10-13, and (3.9 ( 0.8) × 10-12 cm3 molecule-1 s-1, respectively, at 300 K.
These rate constants were obtained after adjustment of the apparent rate constants for secondary reactions.
I. Introduction
titration with monitoring of the Cl(2PJ) concentration by
resonance fluorescence. However, measurement of the Cl atom
emission intensity, which is in the vacuum-ultraviolet, requires
special equipment. The most common alternative for a [Cl]
monitor is the Cl2-afterglow intensity,8 but this technique has
two problems: the Cl2(B-X) emission intensity in the 600-
700 nm region is weak, and it is overlapped by the nitrogen
first-positive bands, which are commonly observed as a
background emission from many discharge systems. Observa-
tion of the Cl2-afterglow intensity is only reliable for rather high
Cl atom concentrations of 1014.5 molecules cm-3 at room
temperature.9,10
Large NCl(a1∆) and NF(a1∆) concentrations can be chemi-
cally generated in the gas phase using the F/Cl/HN3 reaction
scheme.1 In this system, reactions 1-3 are used to make NCl-
(a1∆,b1Σ+) and NF(a1∆,b1Σ+) at room temperature in a flow
reactor.
F + HN3 f HF + N3
(1)
(2)
(3)
F + N3 f NF(a1∆,b1Σ+,X3Σ-) + N2
Cl + N3 f NCl(a1∆,b1Σ+,X3Σ-) + N2
Two Cl atom titration reactions have been used. The first,
developed initially by Clyne et al.,5 uses reaction 5 to remove
Cl atoms.
The rate constants2 for reactions 1 and 2 are accepted as (1.1 (
0.2) × 10-10 and (5 ( 2) × 10-11 cm3 molecule-1 s-1
,
respectively, at 300 K. The rate constant and product branching
fractions for reaction 3 are still subjects of study.1,3 To
understand the competitive reactions for N3, the concentrations
of F and Cl atoms must be measured.3 Reaction 4,
Cl + ClNO f Cl2 + NO
(5)
Unfortunately, nitrosyl chloride is unstable, and it is often
necessary to synthesize ClNO, since commercial sources usually
contain appreciable amounts of Cl2, NO, and NO2 impurities.
An alternative titrant is vinyl bromide,7 with a room-temperature
Cl + HN3 f HCl + N3
(4)
rate constant of 1.4 × 10-10 cm3 molecule-1 s-1
Cl + C2H3Br f C2H3Cl + Br
.
with a reported rate constant4 of (8.9 ( 1.0) × 10-13 cm3
molecule-1 s-1, also needs to be included in analysis of the
kinetics of a F/Cl/HN3 reaction system. This rate constant needs
to be confirmed, and k3 needs to be established. As part of our
program3a to characterize the generation and removal reactions
of NCl(a1∆), we needed a convenient and reliable way to
measure Cl atom concentrations in the flow reactor. Many other
Cl atom reaction systems also would benefit from the availability
of a convenient method for measurement of the absolute and
relative Cl atom concentrations.
The measurement of Cl atom densities in gas-phase flow
reactors has been accomplished by a number of techniques,
including atomic resonance fluorescence and absorption,5
electron paramagnetic resonance,6 mass spectrometry,7 and the
Cl2-afterglow8 emission intensity. A sensitive method for
measuring absolute concentrations in flow reactors is chemical
(6)
The secondary reaction
Cl + C2H3Cl f C2H3Cl2*
(7)
has a bimolecular rate constant11 of 1.2 × 10-10 cm3 molecule-1
s-1 at atmospheric pressure, because the addition product, C2H3-
Cl2*, can be collisionally stabilized. Experiments examining
the pressure dependence12 of reaction 7 show that C2H3Cl2*
dissociates to Cl + C2H3Cl, and the effective bimolecular rate
constant13 is ∼(2-3) × 10-12 cm3 molecule-1 s-1 at 2 Torr.
This rate constant is sufficiently small that the secondary step
does not remove an appreciable amount of Cl atoms in a flow
reactor at <2 Torr. Thus, the stoichiometry for reaction 6 is
one Cl atom removed for each C2H3Br molecule reacted.
X Abstract published in AdVance ACS Abstracts, December 1, 1997.
S1089-5639(97)02118-X CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/01/1998

154 J. Phys. Chem. A, Vol. 102, No. 1, 1998
Manke and Setser
We used the reaction with vinyl bromide to titrate the Cl
atom concentration with observation of the relative [Cl] by the
HCl infrared chemiluminescence produced by reaction 8.
pseudo-first-order reaction conditions with fixed-point observa-
tion, a plot of ln I(HCl) vs [reagent] has a slope equal to the
product of kR∆t. Dividing the slope by kR gives an experimental
value for ∆t. For the initial test, we selected C2F5I as the
reference reaction. However, the value for the rate constant in
the literature19 gave a reaction time that was ∼3 times longer
than the plug flow prediction. Therefore, we selected the more
thoroughly studied HBr reaction16 as the reference. The
experimental value for ∆t using HBr as the reference reagent
was 20% smaller than the plug flow predictions at 2 Torr
pressure. For the positions of the inlets in this reactor, the
reaction times are 18 and 0.6 ms from the C2H3Br and H2S
inlets to the NaCl observation window, respectively.
Cl + H2S f HCl(V)0,1) + SH
(8)
The rate constant for reaction 814 is 5.7 × 10-11 cm3 molecule-1
s-1, and the reaction is thermochemically restricted to HCl(V
) 0 and 1).15 The HCl(1-0) signal is first-order in [Cl], and
the relative intensity serves as a measure of [Cl] as the [C2H3-
Br] is added during the titration. Linear extrapolation of the
HCl(1-0) emission intensity to zero on a plot of intensity vs
[C2H3Br] gives [Cl]0, since reaction 6 goes to completion and
has 1:1 stoichiometry. The infrared emission intensity from
reaction 8 could be observed with an InSb detector for [Cl] g
3 × 1011 cm-3 in our experiments.
Hydrogen azide is not commercially available, since it is
shock sensitive and decomposes if stored for long periods of
time (3-4 weeks). It was synthesized by the reaction of stearic
acid (Aldrich) with NaN3 (Aldrich) and stored as a 10% Ar
mixture in a 10 L reservoir. The reaction mixture, which usually
consisted of 3 g of NaN3 and 30 g of C17H35COOH, was
evacuated and then heated to ∼100 °C for 3-4 h or until the
HN3 pressure reached ∼40 Torr in a 20 L bulb. The HN3 was
then diluted by Ar. The purity was periodically checked by
mass spectrometry. The major peaks in the mass spectrum were
at m/z ) 40 and 43, corresponding to Ar and HN3, respectively.
Evidence of CO2(m/z ) 44), a frequent impurity from this
synthetic method, was also observed; however, the CO2 impurity
was less than 10% (of the HN3) and, in fact, probably came
mainly from the added Ar. The CF3I and C2F5I were purchased
from PCR.
In this paper the method described above is demonstrated
for measuring Cl atom concentration in our flow reactor. We
then used it to calibrate the reaction time in the flow reactor by
observing the Cl + HBr reaction, which has a known rate
constant.16 Next the rate constants for Cl atom removal by HN3,
CF3I, and C2F5I were measured. We also show that a
microwave discharge through dilute flows of Cl2, CCl4, CFCl3,
and CF2Cl2 in Ar are suitable sources for Cl atoms in a flow
reactor.
II. Experimental Methods
(i) Description of the Flow Reactor. The 0.5 m long, 4.0
cm i.d., Pyrex flow reactor used in this study has been used
extensively for infrared chemiluminescence experiments, and
it has been described elsewhere.17 Although the reactor walls
were not coated, the loss of Cl atoms by recombination at the
wall was not serious. Measurement of the fractional dissociation
in a halocarbon wax coated reactor18 by observation of the
reduction of Cl2 when the discharge was turned on gave
approximately the same results as the titration reaction used
here in the uncoated reactor.18 The reactor was pumped by a
Stokes mechanical pump, and the flow velocity was ∼2750 cm
s-1 at 2.0 Torr. The Ar carrier gas was purified by passage
through no fewer than three molecular sieve filled traps cooled
to 195 K. The Cl atom precursors and all other reagents were
degassed by freeze-pump-thaw cycles before being diluted
with Ar and stored in 12 L bulbs. Dilute mixtures of Cl2
(Matheson), CF2Cl2 (PCR or Matheson), CFCl3 (MG Industries),
and CCl4 (Aldrich) in Ar were passed through a microwave
discharge to generate a flow of Cl atoms. The microwave power
was coupled to the discharge with an Evenson type cavity, and
the power (∼20 W) was adjusted for maximum Cl atom
concentration. The alumina discharge tube was inserted into
the reactor through an aluminum flange. In other applications
we have used a quartz discharge tube. Vinyl bromide (Mathe-
son) was added via a separate tube to the bulk flow; the C2H3-
Br inlet tube terminated just downstream of the point where
the Cl atoms entered the reactor. Controlled flows of H2S
(Matheson) were added 30 cm downstream of the Cl and C2H3-
Br inlets. The HCl(1-0) emission was observed through a NaCl
window placed 3 cm downstream of the H2S inlet. The Ar and
Cl atom precursor flow rates were controlled by needle valves
and monitored continuously by calibrated Hastings mass flow
meters. The flow rates of the dilute C2H3Br and H2S mixtures
were also controlled by needle valves, but their flow rates were
measured by diverting the flows to a vessel of known volume
and measuring the rate of pressure rise.
(ii) Detection Methods. Two different infrared detection
systems were used. In the initial experiments to establish the
titration method, we used a Biorad FTS-60 spectrometer with
an InSb detector, which gave resolved vibrational-rotational
HCl spectra.15,17 The band center of the V ) 1 f 0 emission
is at 2886 cm-1, and the Einstein emission coefficient20 is 40
s-1. For better sensitivity, a band-pass filter for 3125-2500
cm-1 (Perkin-Elmer) was used to block the background thermal
radiation. The emission was collected by a 7.5 cm focal length
lens and focused into the spectrometer.
The FT spectrometer, which was equipped with a liquid
nitrogen cooled InSb detector (Infrared Associates, D* ) 2.165
× 1011 cm Hz0.5 W-1), was operated at 2 cm-1 resolution.
Despite the HCl band-pass filter, the experimental HCl spectra
were superimposed on a weak 300 K thermal background. To
correct for this thermal emission, a background spectrum was
collected with no reagents present and subtracted from the HCl
signal; the only feature remaining after subtraction of the
background spectrum was the HCl(1-0) band. The spectrom-
eter was scanned 256 times (∼3 min), and the I(HCl) signal
was taken as the intensity of the strongest line in the HCl(1-0)
band. After the method was proven, the detection system was
simplified for general use with other flow reactors by using an
isolated InSb detector with the same collecting lens and band-
pass filter. The InSb detector in the stand-alone system was
identical with the one in the spectrometer. A mechanical light
chopper (EG&G Model 196) operated at 50 Hz was placed
between the detector and the 7.5 cm focal length quartz lens.
The current produced by the detector was converted to a voltage
by a homemade preamplifier. The modulated output of the
preamplifier was processed by a lock-in amplifier set for a 1 s
time constant (Princeton Applied Research, Model HR-8),
monitored on a digital electrometer (Keithley Model 614) and
recorded on a personal computer. The thermal background was
subtracted by adjusting the dc offset of the lock-in amplifier,
A calibration of the reaction time was made using a reference
reaction with rate constant kR for removal of Cl atoms. For

Measuring Gas-Phase Cl Atom Concentrations
J. Phys. Chem. A, Vol. 102, No. 1, 1998 155
Figure 1. Plots of I(HCl) as a function of [H2S] and [Cl]. The FTIR
spectrometer was used to measure I(HCl). The data points in panel a
indicated by the circles were obtained by titration. In panel b, CF2Cl2
was the Cl source and the stand-alone InSb detector was used to
measure I(HCl). The Cl atom concentrations were determined by
titration.
Figure 2. Sample Cl atom titration plots: (a) Two titrations on different
days for the same [Cl2], but with [H2S] ) 1.7 × 1013 or 9.0 × 1012
molecules cm-3; the [Cl] was (6.2 ( 0.1) × 1012 atoms cm-3 for both
experiments; I(HCl) was measured with the FTIR spectrometer. (b)
Three titrations for CF2Cl2 as the Cl atom source with [H2S] ) 6.5 ×
1012 molecules cm-3; the I(HCl) was measured by the stand-alone InSb
detector.
and the I(HCl) data were acquired with a 20 s average of the
signal. With optical and electronic filtering, signal-to-noise
ratios of 100:1 or better were routinely attained for [H2S] )
a Cl atom titration or any other relative Cl atom concentration
measurement.
The HCl(1-0) emission intensity increased linearly with the
Cl2 (10% mixture) flow rate through the discharge over the 6.0
× 1011-1.0 × 1013 molecules cm-3 range for [H2S] ) 9.0 ×
1012 (see Figure 1a). The linear fit to the data extrapolates
through the origin, indicating that I(HCl) [Cl2] for the whole
range of [Cl2]; i.e., the fractional dissociation of Cl2 was
constant. This point was confirmed by titration measurements
of [Cl] over this range of [Cl2]. For the other Cl atom
precursors, the linearity of I0(HCl) vs [Cl source] fails for higher
flows, which implies that the effective dissociation efficiency
depends on the concentration of the precursor.
1.0 × 1013 molecules cm-3
.
The most significant advantages of the stand-alone detector
over the FT spectrometer are portability and ease of use. When
using the interferometer to collect the data, at least 3 min was
required to collect a spectrum, and about a half hour was needed
to complete a Cl atom titration. The signal from the InSb
detector can be monitored continuously, and a titration can be
performed in a few minutes. On the other hand, the resolved
spectrum from the FT spectrometer can be very helpful for some
experiments.
The utility of the vinyl bromide titration reaction has been
shown previously7 using mass spectrometry as the measure of
[Cl], and we describe the titration method only briefly here.
The I(HCl) is monitored vs added [C2H3Br] for constant [H2S],
∆t, and [Cl atom precursor]. The [Cl] value is given by the
linearly extrapolated I(HCl) ) 0 intercept on the [C2H3Br] axis.
The reaction time for reaction 6 was 8 ms, so [Cl] will be 80%
III. Experimental Results
(i) Experiments with the FT Spectrometer To Establish
the Titration Method. Because the reaction time for the Cl
+ H2S reaction is short (0.6 ms), the [HCl] follows the
differential rate law, and I(HCl) is first-order in [H2S] and [Cl],
e.g. I(HCl) k8[H2S][Cl]∆t. Figure 1a demonstrates that I(HCl)
removed for [Cl]0 ) [C2H3Br]0 g 3 × 1012 molecules cm-3
.
is proportional to [Cl], for [H2S] ) 9 × 1012 molecules cm-3
.
Two sample titrations using Cl2 as the precursor are shown in
Figure 2a. The titrations were performed on different days with
[H2S] ) 1.7 × 1013 or 0.9 × 1013 molecules cm-3, but for the
same [Cl2] ) 1.0 × 1013 molecules cm-3. The I(HCl) signal,
which was recorded with the FT spectrometer, decreases linearly
with added vinyl bromide, and a smooth extrapolation to zero
I(HCl) gives [Cl]0. The broken line in Figure 2a represents the
calculated Cl atoms remaining at the detector for [Cl]0 ) 6.2
× 1012 atoms cm-3. The agreement between the experimental
The I(HCl) is also proportional to [H2S] for fixed [Cl] as shown
in Figure 1b. For sufficiently large [H2S], the I(HCl) reaches
a plateau because some Cl atoms are removed, and the
differential rate law no longer holds. However, even in this
high [H2S] region, the I(HCl) still scales with [Cl]. Although
nearly any H2S concentration except [H2S] , [Cl] can be used
to generate the I(HCl) signals, the best choice is [H2S] ≈ 5[Cl]0.
Checks should be made to maintain a constant H2S flow during

156 J. Phys. Chem. A, Vol. 102, No. 1, 1998
Manke and Setser
data and the theoretical line shows that the linear extrapolation
from the data for [C2H3Br] e 0.8[Cl]0 is valid. The titrations
in Figure 2a both extrapolate to (6.2 ( 0.1) × 1012 cm-3, and
the same fractional dissociation of Cl2 was obtained in both
experiments. Three additional Cl atom titrations are shown in
Figure 2b for CF2Cl2 as the Cl atom source. For these
experiments, the I(HCl) was recorded with the stand-alone
detector and [H2S] ) 6.5 × 1012 molecules cm-3. The zero
intensity intercepts give [Cl]0 ) 0.95 × 1013, 1.7 × 1013, and
2.9 × 1013 atoms cm-3 for [CF2Cl2] ) 0.52 × 1013, 1.1 × 1013,
and 2.2 × 1013 molecules cm-3, respectively. The titration plots
in Figure 2 show that we can calibrate the I(HCl) signal vs
known [Cl]0 using titrations with vinyl bromide to establish the
absolute Cl atom concentration scale.
Attempts to titrate Cl atoms for [Cl2] < 3 × 1012 molecules
cm-3 gave unreliable results, because reaction 6 does not go to
completion in the fixed 8 ms reaction time for low [Cl]. It
would be possible to increase ∆t by decreasing the pumping
speed and complete the titration. However, this is unnecessary
after preparation of a calibration curve of I(HCl) vs [Cl]. The
lowest [Cl] that can be measured is limited by the sensitivity
of the detector for I(HCl). The smallest reliable signal for the
FT spectrometer was [Cl] ) 3.3 × 1011 atoms cm-3. The
smallest signal detected using the stand-alone detector with the
phase-lock amplifier set for a 1 s time constant was [Cl] ) 6.0
× 1011 atoms cm-3. In principle, the stand-alone detector should
be more sensitive, and even smaller [Cl] could be detected with
better optics and by increasing the time constant and sensitivity
of the lock-in amplifier, but we did not attempt such measure-
ments.
Figure 3. Dissociation efficiency for Cl2, CF2Cl2, CFCl3, and CCl4 vs
precursor concentration. The number of Cl atoms produced per molecule
of Cl2 is 0.6 for the [Cl2] range studied. The apparent dissociation
efficiency for the chlorofluoromethanes decreases linearly, with the
same apparent slope for increasing precursor concentration. The highest
number of Cl atoms per precursor molecule is 1.5, 1.1, and 1.0 for
CF2Cl2, CFCl3, and CCl4, respectively.
(ii) Microwave Discharges of Cl2, CCl4, CCl3F, and CCl2F2
as Cl Atom Sources. In many chemical systems, the presence
of Cl2 is undesirable, and an alternative precursor to molecular
chlorine as a source of Cl atoms is needed. This is the case for
the F/Cl/HN3 reaction scheme, because the reaction rate of NCl-
(a1∆) with Cl2 causes complications. The CF2Cl2, CFCl3, and
CCl4 molecules have all been used in this laboratory as possible
Cl atom sources.3a,21 These molecules, however, are not without
problems of their own. Experiments with the FT spectrometer,
but without the HCl band-pass filter, showed that Cl and F atoms
are both generated when CF2Cl2 and CFCl3 (to a much smaller
degree) are passed through a microwave discharge. The Cl and
F atoms react with H2S to form HCl and HF, respectively. On
the basis of the relative HCl and HF infrared emission intensities,
the nascent vibrational distributions, the HX formation rate
constants, and the Einstein A coefficients,20 we estimate that
the [Cl]:[F] ratio from the discharge through CF2Cl2 was ∼14:
1. Only a trace of HF emission was observed when CFCl3 was
passed through the discharge, so the F atom concentration
generated from CFCl3 was negligible. The HF(V) contribution
to the infrared emission signal detected by the stand-alone InSb
detector is effectively eliminated by the HCl band-pass filter,
and a small concentration of F atoms can be tolerated in the
measurement of the Cl atom concentrations. Another disad-
vantage for CFCl3, and more seriously for CCl4, is the
development of a black residue in the discharge tube after a
few hours of use. This residue suggests the formation of carbon
and possibly Cl2 from the discharge. In any case, the residue
can be removed easily by flowing O2 through the discharge tube
with the discharge operating. One should always be on guard
for other reactive species when passing polyatomic molecules
through microwave discharges. Indeed, we have evidence21 that
CFCl3 gives some species other than Cl atoms when concentra-
tions of g1 × 1013 molecules cm-3 of CFCl3 were used for
NCl(a) and NF(a) quenching experiments.
The Cl atom titration method worked very well with all the
precursor molecules for measurement of the Cl atom concentra-
tion, as shown by the plots in Figure 2b. The average number
of Cl atoms per precursor molecule is plotted vs the precursor
concentration in Figure 3. After calibration of the I(HCl) by
titration, the [Cl] was determined from the relative I(HCl). The
fractional dissociation of Cl2 is low, but constant, at ∼30%;
i.e., each Cl2 molecule gives 0.6 atom up to [Cl2] ) 1 × 1013
cm-3. We have obtained a similar result in a different flow
reactor.18 The concentration range for CF2Cl2, CFCl3, and CCl4
was extended to 3 × 1013 molecules cm-3, and the dissociation
efficiency decreases for higher concentrations. The low con-
centration limit corresponds to 1.5, 1.2, and 1.0 Cl atoms per
CF2Cl2, CFCl3, and CCl4 molecule, respectively. Perhaps
fortuitously, the apparent reduction in dissociation efficiency
with increased precursor concentration was the same for all three
molecules. The number of Cl atoms obtained per CFCl3
molecule seemed less reproducible than for the other precursors;
however, two separate experiments with CFCl3 gave the same
qualitative result.
Using dilute mixtures in Ar (Cl2, 10% or 5%; CF2Cl2, 30%;
CFCl3, 20%; CCl4, 15%), we measured the dependence of the
Cl atom production on microwave power using the titration by
C2H3Br to establish the absolute Cl atom concentration. The
highest fractional dissociation in all cases was for the lowest
microwave power for which the discharge was stable, ∼20 W.
For higher power, the F atom concentrations increased with CF2-

Measuring Gas-Phase Cl Atom Concentrations
J. Phys. Chem. A, Vol. 102, No. 1, 1998 157
Cl2 and CFCl3, and the rate of deposition of the solid residue
increased for CFCl3 and CCl4.
A note of precaution should be reserved for use of the
fluorochloromethanes as Cl atom sources. The cofragment from
dissociation is likely to be halomethyl or halocarbene radicals.
The former have recombination rates with Cl atoms, as well as
with themselves. The Cl + CF3 recombination rate constant at
300 K can be estimated22 as ∼1 × 10-11 cm3 molecule-1 s-1
at 1 Torr. If [Cl] ) [CF3] ) 1 × 1013 cm-3, then one-half of
the Cl atoms could be lost by recombination in 0.01 s. The
recombination rate constants for Cl + CFCl2 and CCl3 will be
similar.22 The decline in the apparent dissociation efficiency
shown in Figure 3 could be a consequence of these recombina-
tion reactions. A microwave discharge in CF4 is commonly
used as a source of F atoms.2,3,21 However, that source gives
2F per CF4 molecule in the 1 × 1013 molecules cm-3 range,
and recombination seems not to be a problem, because the rate
constant for F + CF2 is less than for F + CF3.23
(iii) Measurement of Cl Atom Removal Rate Constants.
The measurement of reliable rate constants for Cl atom removal
in a flow reactor depends on knowing the reaction time.
Experiments with HBr demonstrated that, in fact, the plug-flow
reaction time was approximately correct.
Cl + HBr f HCl(V) +Br
(9)
Reaction 9 has a room-temperature rate constant16 of (1.05 (
0.1) × 10-11 cm3 molecule-1 s-1, and it produces vibrationally
excited HCl (the nascent distribution15 is V ) 0:1:2 ) 33:50:
17) and some vibrationally excited molecules survive to the IR
detection zone. Therefore, the observed IR signal from the Cl
+ H2S reaction must be adjusted for the background emission
from reaction 9. The I(HCl) with the flow of H2S turned off
was subtracted from the total I(HCl) signal obtained from H2S
plus reaction 9.
The rate constants for removal of Cl atoms by HN3, C2F5I,
and CF3I were measured by direct comparison to the HBr
reaction, and, in fact, we did not use the reaction time. Pseudo-
first-order conditions were used for all reagents, so the integrated
rate law simplifies to eq 10 with ∆t fixed for the HBr reaction.
Figure 4. Pseudo-first-order plots of Cl atom concentration vs reagent
concentration. In all cases [H2S] was ∼1 × 1013 cm-3, the reactor
pressure was 2 Torr, and ∆t was 11 ms. (a) The slopes (cm3 molecule-1
)
and Cl atom concentrations (atoms cm-3) for the HBr reaction are as
follows: (b) [Cl] ) 2.3 × 1012, slope ) 1.0 × 10-13; (9) [Cl] ) 8.8
× 1012, slope ) 8.9 × 10-14. (b) For the Cl + HN3 reaction: (b) [Cl]
) 3.1 × 1012, slope ) 1.9 × 10-14; (9) [Cl] ) 6.2 × 1012, slope ) 1.9
× 10-14; ([) [Cl] ) 1.2 × 1013, slope ) 2.0 × 10-14. (c) For the Cl
+ CF3I reaction: (b) [Cl] ) 8.8 × 1012, slope ) 1.1 × 10-14; (9) [Cl]
) 5.2 × 1012, slope ) 9.2 × 10-15; ([) [Cl] ) 2.3 × 1012, slope )
1.4 × 10-14
.
rate constant for Cl + N3 is still uncertain,1,3,18 but it is larger
than 2 × 10-11 cm3 molecule-1 s-1 and each HN3 molecule
will remove two Cl atoms. With this assumption, our rate
constant for HN3 agrees with the previously published value,
although those previous measurements were somewhat indirect.4
The importance of the secondary reaction for CF3I is more
difficult to ascertain. The rate constants for Cl + ICl24 and for
Cl + CF3 (at 1 Torr)22 are 0.8 × 10-11 and ∼1 × 10-11 cm3
molecule-1 s-1, respectively. Numerical integration of the rate
equations suggests that each secondary reaction consumes about
0.5 Cl atom per molecule of CF3I that reacts. Thus, the apparent
rate constant was divided by a factor of 2 to estimate the primary
rate constant.
The Cl + C2F5 recombination reaction probably is close to
its high-pressure limit at 1 Torr, and this reaction will remove
a second Cl atom. We ran a numerical simulation of the Cl +
C2F5I system that included Cl + ICl and Cl + C2F5, with rate
constants of 0.8 × 10-11 and 1.0 × 10-10 cm3 molecule-1 s-1
to estimate the stoichiometry. The results indicate that each
C2F5I molecule that reacts removes about 2.5 Cl atoms, and
the apparent rate constant was divided by 2.5 to estimate the
primary rate constant.
I(HCl)/I0(HCl) [Cl]/[Cl]0 )
exp[-(kHBr[reagent]∆t)] (10)
The slopes of the pseudo-first-order plots for reaction with HN3
and CF3I (see Figure 4 for representative plots), relative to the
slope for HBr, were used to calculate the overall rate constant
(see eq 11 for CF3I).
slope(CF3I)
kCF I
)
kHBr
(11)
3
slope(HBr)
The comparison between HBr and C2F5I is shown in Figure 5.
Several experiments for each reagent were performed, and [Cl]0
and ∆t were both varied. In addition, two different Cl atom
precursors, CF2Cl2 and CFCl3, were used. No systematic
difference in rate constants was noted between the two Cl atom
sources. We did not use Cl2 as the Cl atom source, because
the reaction of CF3 with Cl2 would regenerate Cl atoms. The
data give linear plots, and the slopes were reproducible. The
total Cl atom removal rate constants for HN3, CF3I, and C2F5I
are summarized in Table 1.
The removal of Cl atoms from secondary reactions must be
considered for HN3, CF3I, and C2F5I, before the apparent rate
constants can be converted to elementary rate constants. The
Our rate constants for CF3I and C2F5I are both larger than
the measurements of Ahonkhai and Whittle.19 In their study,
mixtures of Cl2 + C2F5I (or CF3I) + CH4 were photolyzed,

158 J. Phys. Chem. A, Vol. 102, No. 1, 1998
Manke and Setser
reaction 6 is ∼5 times faster than the Cl + ClNO reaction, vinyl
bromide is inherently more sensitive as a titrant than nitrosyl
chloride. In addition, vinyl bromide is stable at room temper-
ature, not corrosive, and commercially available in high purity.
On the basis of these considerations, we recommend C2H3Br
for the titration of gas-phase Cl atom densities of g3 × 1012
atoms cm-3 in a flow reactor. While this laboratory has little
experience in generating or measuring Cl atom concentrations
above 1 × 1014 atoms cm-3, we do not expect any complications
in extending reaction 6 to higher Cl atom concentrations. It is
important to note, however, that the stoichiometry of Cl removal
by vinyl bromide will change if the pressure in the reactor is
too high, due to the increasing importance of the Cl + C2H3Cl
secondary reaction.
In addition to testing the C2H3Br titration reaction, we have
also demonstrated the use of HCl(1-0) emission intensity as a
monitor of the relative [Cl] using a Fourier transform infrared
spectrometer or a stand-alone InSb detector with a band-pass
filter. High-sensitivity IR detectors for the HCl(1-0) spectral
region are commercially available, and in many instances the
observation of IR emission is more easily implemented than
the two-photon laser-induced fluorescence on the vacuum-
ultraviolet atomic resonance fluorescence or absorption tech-
niques. If H2S is not chemically compatible with other reagents
in the flow reactor, other H atom donors with low bond energies,
such as HI or SiH4, could be substituted. After calibration of
the HCl infrared emission intensity by titration to establish the
absolute [Cl], the Cl atom concentrations (g3 × 1011 atoms
cm-3) could be measured from the relative infrared emission
intensities.
Figure 5. Pseudo-first-order plot of Cl atom concentration vs C2F5I
concentration. The reaction pressure was 2 Torr, [H2S] ) 1 × 1013
cm-3, and ∆t ) 11 ms. (a) The slope (cm3 molecule-1) and [Cl]0 (atoms
cm-3) for the HBr experiments are as follows: (b) [Cl]0 ) 6.2 × 1012,
slope ) 8.9 × 10-14; (9) [Cl]0 ) 1.2 × 1013, slope ) 7.5 × 10-14. (b)
For Cl + C2F5I: (b) [Cl] ) 6.2 × 1012, slope ) 6.9 × 10-14; (9) [Cl]0
The Cl atom concentration from a microwave discharge
through four precursor molecules diluted in Ar have been
characterized. Under our conditions, Cl2, CF2Cl2, CFCl3, and
CCl4 are all satisfactory sources of Cl atoms in a flow reactor.
However, CF2Cl2 gives Cl and F atoms in a 14:1 ratio, while
CFCl3 and CCl4 both generate carbon deposits inside the
discharge tube. The F atoms and the deposits are minimized
by operation of the discharge at low power. The possibility
for loss of Cl atoms from recombination reactions with CF2Cl,
CFCl2, and CCl3 radicals should be remembered. Such loss
processes will become important for higher Ar pressures and
longer reaction times.
) 1.2 × 1013, slope ) 8.0 × 10-14
.
TABLE 1: Cl Atom Removal Constants
rate constanta
reagent
apparent value
elementary step
literature
HN3
CF3I
C2F5I
2.2 ( 0.6
1.0 ( 0.3
9.8 ( 2.0
1.1 ( 0.3
0.50 ( 0.15
3.9 ( 0.8
0.89 ( 0.10b
0.17 ( 0.14c
1.5 ( 1.2c
a
In units of 10-12 cm3 molecules-1 s-1; the uncertainties are the
standard deviation from six experiments for each reaction. See text for
justification for corrections for secondary reactions. b Reference 4.
c Reference 19.
The Cl atom removal rate constants for HN3, CF3I, and C2F5I
have been measured at 300 K, relative to a reference reaction,
Cl + HBr. After allowance for secondary reactions, the
elementary rate constants were assigned as (1.1 ( 0.3) × 10-12
,
and the removal of Cl atoms was monitored by end product
analysis using gas chromatography. The ratio of CH3Cl to C2F5-
Cl generated by reaction 12
(0.51 ( 0.15) × 10-12, and (3.9 ( 0.8) × 10-12 cm3 molecule-1
s-1, respectively. The quoted statistical uncertainties do not
include the uncertainties introduced by the corrections for
secondary reactions. In other work21 we have demonstrated that
the major product from the Cl + HN3 reaction is HCl + N3.
Although the rate is relatively slow at room temperature, the
reaction could be a chemical source of N3 at higher temperature.
CH3 (or C2F5) + Cl2 f CH3Cl (or C2F5Cl) + Cl (12)
was related to the competition between Cl + CH4 and Cl +
C2F5I to obtain the rate constant. Although our measurement
is more direct, we have not eliminated the secondary reactions.
We loaded several C2F5I samples taking care to remove any I2
or other impurities, but the same apparent rate always was
obtained. Our rate constant values for CF3I and C2F5I probably
are upper limits to the true elementary rate constant, because
of the uncertainty associated with the treatment of the secondary
reactions.
Acknowledgment. This work was supported by the U.S.
Air Force Office of Scientific Research, Grant F49620-96-1-
0110.
References and Notes
(1) Liu, X.; MacDonald, A.; Coombe, R. D. J. Phys. Chem. 1992, 96,
4907.
IV. Conclusions
(2) Du, K. Y.; Setser, D. W. J. Phys. Chem. 1990, 94, 2425.
(3) (a) Manke, II, G. C.; Setser, D. W. J. Phys. Chem., to be published.
(b) Hewett, K. B.; Setser, D. W. J. Phys. Chem. A 1997, 101, 9125; this
A successful titration reaction requires a large rate constant,
known stoichiometry, and chemical stability of the titrant. Since

Measuring Gas-Phase Cl Atom Concentrations
J. Phys. Chem. A, Vol. 102, No. 1, 1998 159
paper reports the branching ratio for HNF formation in reaction 1 as
(15) Agrawalla, B.; Setser, D. W. In Gas-Phase Chemiluminescence and
Chemiionization; Fontjn, A., Ed.; Elsevier: Amsterdam, 1985.
(16) (a) Persky, A. J. Phys. Chem. 1995, 99, 16192. (b) Rubin, R.;
Persky, A. J. Chem. Phys. 1983, 79, 4310.
(17) (a) Wategaonkar, S. J.; Setser, D. W. J. Chem. Phys. 1989, 90,
251. (b) Rengarajan, R.; Setser, D. W.; Des Marteau, D. D. J. Phys. Chem.
1994, 98, 10568.
(18) Hewett, K.; Setser, D. W. J. Phys. Chem., To be submitted.
(19) (a) Ahonkhai, S. I.; Whittle, E. Int. J. Chem. Kinet. 1984, 16, 543.
(b) Ahonkhai, S. I.; Whittle, E. Int. J. Chem. Kinet. 1981, 13, 427.
(20) Arunan, E.; Setser, D. W.; Oglivie, J. F. J. Chem. Phys. 1992, 97,
1734.
(21) Manke II, G. C. Ph.D. Thesis, Kansas State University, 1997.
(22) (a) Kumaran, S. S.; Su, M.-C.; Lim, K. P.; Michael, J. V.; Wagner,
A. F.; Harding, L. B.; Dixon, D. A. J. Phys. Chem. 1996, 100, 7541. (b)
Michael, J. V. Private communication, 1997.
+0.02
0.03
.
-0.01
(4) Yamasaki, K.; Gueno, T.; Kajimoto, O. Chem. Phys. Lett. 1983,
94, 475.
(5) Clyne, M. A. A.; Cruse, H. W.; Watson, R. T. J. Chem. Soc.,
Faraday Trans. 2 1972, 68, 153.
(6) Westenberg, A. A. J. Chem. Phys. 1968, 43, 1544.
(7) Park, J.-Y.; Slagle, I. R.; Gutman, D. J. Phys. Chem. 1983, 87,
1812.
(8) Bader, L. W.; Ogryzlo, E. A. J. Chem. Phys. 1964, 41, 2926.
(9) Clyne, M. A. A.; Stedman, D. H. J. Chem. Soc., Faraday Trans.
1968, 64, 1816.
(10) Clyne, M. A. A.; Nip, W. S. In Reactive Intermediates in The Gas
Phase; Setser, D. W., Ed.; Academic Press: New York, 1979.
(11) Atkinson, R.; Aschmann, S. M. Int. J. Chem. Kinet. 1987, 19, 1097.
(12) Iyer, R. S.; Rowland, F. S. J. Phys. Chem. 1985, 89, 3730.
(13) The bimolecular Cl atom removal rate constant at ∼2 Torr and
300 K has not been measured. We estimated it using a value for k-1 of 2.7
× 109 s-1; see ref 11.
(23) Plumb, J. C.; Ryan, D. R. Plasma Chem. Plasma Process 1986, 6,
11.
(24) Clyne, M. A. A.; Cruse, H. W. J. Chem. Soc., Faraday Trans. 2
1972, 68, 1377.
(14) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F., Jr.; Kerr,
J. A.; Troe, J. J. Phys. Chem. Ref. Data 1992, 21, 1125.