Kinetics of the reactions of acetonitrile with chlorine and fluorine atoms

Article abstract of DOI:10.1021/jp9521417

The rate coefficients for the reactions of chlorine and fluorine atoms with acetonitrile have been measured using relative and direct methods. In the case of chlorine atoms the rate coefficient k₁ was measured between 274 and 345 K using competitive chlorination and at 296 K using laser flash photolysis with atomic resonance fluorescence. The rate coefficient measured at ambient temperature (296 ± 2 K) is (1.15 ± 0.20) × 10^-14 cm³ molecule^-1 s^-1, independent of pressure between 5 and 700 Torr (uncertainties are 2 standard deviations throughout). This result is a factor of 6 higher than the currently accepted value. The results from the three independent determinations reported here yield the Arrhenius expression k₁ = (1.6 ± 0.2) × 10^-11 exp[-(2140 ± 200)/T] cm³ molecule^-1 s^-1. Product studies show that the reaction of Cl atoms with CH₃CN proceeds predominantly, if not exclusively, by hydrogen abstraction at 296 K. The rate coefficient for the reaction of fluorine atoms with acetonitrile was measured using both the relative rate technique and pulse radiolysis with time-resolved ultraviolet absorption spectroscopy. The rate coefficient for the reaction of F atoms with CH₃-CN was found to be dependent on total pressure. The observed rate data could be fitted using the Troe expression with F_c = 0.6, k₀ = (2.9 ± 2.1) × 10^-28 cm⁶ molecule^-2 s^-1, and k_∞ = (5.8 ± 0.8) × 10^-11 cm³ molecule^-1 s^-1, with a zero pressure intercept of (0.9 ± 0.4) × 10^-11 cm³ molecule^-1 s^-1. The kinetic data suggest that the reaction of F atoms with CH₃CN proceeds via two channels: a pressure-independent H atom abstraction mechanism and a pressure-dependent addition mechanism. Consistent with this hypothesis, two products were observed using pulsed radiolysis with detection by UV absorption spectroscopy. As part of the product studies, relative rate techniques were used to measure k(Cl+CH₂ClCN) = (2.8 ± 0.4) × 10^-14 and k(F+CH₂FCN) = (3.6 ± 0.2) × 10^-11 cm³ molecule^-1 s^-1.

Full text of DOI:10.1021/jp9521417

660
J. Phys. Chem. 1996, 100, 660-668
Kinetics of the Reactions of Acetonitrile with Chlorine and Fluorine Atoms
Geoffrey S. Tyndall* and John J. Orlando
Atmospheric Chemistry DiVision, National Center for Atmospheric Research, P.O. Box 3000,
Boulder, Colorado 80307-3000
Timothy J. Wallington
Ford Research Laboratory, SRL-3083, Ford Motor Company, P.O. Box 2053, Dearborn, Michigan 48121
Jens Sehested
EnVironmental Science and Technology Department, Risø National Laboratory, DK-4000 Roskilde, Denmark
Ole J. Nielsen
Ford Forschungszentrum Aachen, Ford Motor Company, Dennewartstrasse 25, D-52068 Aachen, Germany
ReceiVed: July 26, 1995; In Final Form: October 2, 1995X
The rate coefficients for the reactions of chlorine and fluorine atoms with acetonitrile have been measured
using relative and direct methods. In the case of chlorine atoms the rate coefficient k1 was measured between
274 and 345 K using competitive chlorination and at 296 K using laser flash photolysis with atomic resonance
fluorescence. The rate coefficient measured at ambient temperature (296 ( 2 K) is (1.15 ( 0.20) × 10-14
cm3 molecule-1 s-1, independent of pressure between 5 and 700 Torr (uncertainties are 2 standard deviations
throughout). This result is a factor of 6 higher than the currently accepted value. The results from the three
independent determinations reported here yield the Arrhenius expression k1 ) (1.6 ( 0.2) × 10-11 exp[-
(2140 ( 200)/T] cm3 molecule-1 s-1. Product studies show that the reaction of Cl atoms with CH3CN proceeds
predominantly, if not exclusively, by hydrogen abstraction at 296 K. The rate coefficient for the reaction of
fluorine atoms with acetonitrile was measured using both the relative rate technique and pulse radiolysis with
time-resolved ultraviolet absorption spectroscopy. The rate coefficient for the reaction of F atoms with CH3-
CN was found to be dependent on total pressure. The observed rate data could be fitted using the Troe
expression with Fc ) 0.6, k0 ) (2.9 ( 2.1) × 10-28 cm6 molecule-2 s-1, and k∞ ) (5.8 ( 0.8) × 10-11 cm3
molecule-1 s-1, with a zero pressure intercept of (0.9 ( 0.4) × 10-11 cm3 molecule-1 s-1. The kinetic data
suggest that the reaction of F atoms with CH3CN proceeds via two channels: a pressure-independent H atom
abstraction mechanism and a pressure-dependent addition mechanism. Consistent with this hypothesis, two
products were observed using pulsed radiolysis with detection by UV absorption spectroscopy. As part of
the product studies, relative rate techniques were used to measure k(Cl+CH2ClCN) ) (2.8 ( 0.4) × 10-14
and k(F+CH2FCN) ) (3.6 ( 0.2) × 10-11 cm3 molecule-1 s-1.
Introduction
at 370 and 413 K, using chloroform as the reference gas. Poulet
et al.11 measured absolute values of k1 between 295 and 723 K
using a discharge flow tube, with chlorine atoms in excess, and
mass spectroscopic detection of acetonitrile. The results of
Poulet et al. showed a curved Arrhenius plot, with effective
activation energies of 6 kcal mol-1 at 723 K and 3 kcal mol-1
at 295 K. The value determined at room temperature was 9 ×
10-15 cm3 molecule-1 s-1. Kurylo and Knable10 used flash
photolysis with resonance fluorescence detection of chlorine
atoms to derive an upper limit of 2 × 10-15 cm3 molecule-1
s-1 at 298 K. Clearly, there is a large discrepancy in the
measurements. The high temperature values of Olbregts et al.
and Poulet et al. extrapolate back to values at room temperature
similar to the upper limit of Kurylo, and this has been the
accepted value since 1984.12,13 However, if the results of Poulet
and co-workers11 at 295 K are extrapolated down to stratospheric
temperatures, a significant value of k1 is expected.
Acetonitrile, CH3CN, is emitted into the troposphere from a
number of combustion processes, including biomass burning,
internal combustion engines, and cigarette smoke.1 It is also
used widely as a solvent. Tropospheric concentrations of the
order of 100 ppt have been reported, associated with a total
source strength of 1.1 ( 0.5 Tg/yr.1 Acetonitrile has a lifetime
in the troposphere of approximately 0.5 yr,1 leading to mixing
ratios of 40 ppt in the lower stratosphere and falling to around
0.3 ppt at 45 km altitude.2-6 In the stratosphere, its large dipole
moment and large proton affinity allow it to replace water in
hydrated complexes of ions and to produce the so-called non-
proton hydrates. Vertical profiles of acetonitrile and the
associated ionic chemistry have been modeled by Brasseur and
co-workers.7,8
The reaction between chlorine atoms and acetonitrile (1) has
been studied several times before.9-11
The reaction of OH with acetonitrile appears to proceed by
a complex mechanism, with both addition and abstraction
channels taking place.14 If one assumes that the value of that
rate coefficient at low pressure and in the absence of oxygen
(2 × 10-14 cm3 molecule-1 s-1) corresponds to an abstraction
Cl + CH3CN f HCl + CH2CN
Olbregts et al.9 performed a competitive chlorination experiment
(1)
X Abstract published in AdVance ACS Abstracts, December 1, 1995.
0022-3654/96/20100-0660$12.00/0 © 1996 American Chemical Society

Reactions of Acetonitrile with Cl and F
J. Phys. Chem., Vol. 100, No. 2, 1996 661
reaction, the upper limit for the chlorine reaction seems
anomalously low, since exothermic abstractions by chlorine
atoms are typically a factor of 3-10 faster than the correspond-
ing reactions of OH.12,13 Thus, we have reinvestigated the
reaction of chlorine atoms with acetonitrile using the comple-
mentary approaches of flash photolysis with resonance fluo-
rescence and the relative rate technique. The relative rate
experiments were carried out in two separate chambers, with
very good consistency, and one of the systems was used to
measure the temperature dependence of k1. The results of the
high-temperature relative rate study of Olbregts et al.9 have
been reevaluated using new data on the reference reaction (Cl
+ CHCl3) and are shown to be in excellent agreement with an
extrapolation of the current data to higher temperature.
To complement our study of reaction 1 the rate coefficient
k2 for the reaction of F atoms with acetonitrile was measured
both by use of the relative rate technique and by pulse radiolysis
with detection of transient intermediates using ultraviolet optical
absorption spectroscopy.
then
[CH3CN]
[reference]
t0
k
t0
ln
)
1 ln
(I)
(
)
(
)
k4
[CH3CN]t
[reference]t
where [CH3CN]t and [reference]t are the concentrations of CH3-
0
0
CN and the reference compound at time t0 and [CH3CN]t and
[reference]t are the corresponding concentrations at time t; k1
and k4 are the rate coefficients of reactions 1 and 4, respectively.
An analogous equation holds for fluorine atoms.
The decay of CH3CN and the reference were measured using
their characteristic absorptions in the infrared over the following
wavelength ranges (in cm-1); CH3CN, 1300-1600; CH4, 1200-
1400; CD4, 950-1050; CF3CCl2H, 1100-1200. Initial con-
centrations of the gas mixtures were 15-74 mTorr of CH3CN,
0.2-0.5 Torr of Cl2 or 0.3-0.5 Torr of F2, and 11-90 mTorr
of the reference compound (CH4, CD4, or CF3CCl2H), in 5-700
Torr total pressure of either N2, O2, or air diluent at 296 K. All
reagents were purchased from commercial vendors at purities
of >99% and used without further purification. Experiments
were performed at room temperature, 296 ( 2 K.
b. NCAR Environmental Chamber. Relative rate mea-
surements of k1 were carried out as a function of temperature
in a 48-L stainless steel chamber (S/V ) 0.25 cm-1).18,19 The
chamber is surrounded by a jacket through which ethanol or
water can be circulated to regulate the temperature. Photolysis
was accomplished using light from a xenon arc lamp filtered
by a glass filter (Corning 7-51). Measurements were made
between 274 and 345 K. Infrared spectra were taken using a
Bomem DA3.01 FTIR spectrometer, with a path length of 32.6
m. Spectra were taken by coadding 200 scans at a resolution
of 1 cm-1. Experiments were carried out using both N2 and
O2 as bulk gas. Acetonitrile was degassed before use and was
added to the chamber by sweeping known pressures from a
calibrated bulb using the appropriate carrier gas. Methane was
taken from a cylinder of research grade gas without further
purification.
F + CH3CN f products
(2)
Reaction 2 has been the subject of two previous studies, both
of which were carried out at less than 2 Torr total pressure.15,16
We find k2 to be pressure-dependent, with the values obtained
at low pressure in good agreement with the earlier studies. The
effects of the rate coefficients on calculated CH3CN concentra-
tions in the stratosphere are examined.
Experimental Section
Experiments were carried out in four different systems: (a)
an environmental chamber at ambient temperature with FTIR
detection at Ford Motor Company, (b) a temperature-controlled
chamber with FTIR detection at NCAR, (c) a flash photolysis-
resonance fluorescence experiment at ambient temperature at
NCAR, and (d) a room temperature pulse radiolysis-UV
absorption system at Risø National Laboratory. All have been
described in detail elsewhere, and so only important details will
be summarized here.
c. NCAR Flash Photolysis-Resonance Fluorescence
System. The rate coefficient for chlorine atoms with acetonitrile
was measured using excimer laser flash photolysis of chlorine
gas and resonance fluorescence detection of chlorine atoms.20,21
The reaction vessel consists of a 300-cm3 Pyrex cylinder with
side arms for photolysis and resonance fluorescence detection.
The reactive gases (Cl2, CH3CN) were flowed into the reaction
vessel and photolyzed with an excimer laser at 308 nm (Questek
Model 2440). A chlorine atom resonance lamp was directed
at right angles to the laser beam, and resonance fluorescence
was detected orthogonally to both beams using a solar-blind
photomultiplier (Hamamatsu R1459). Acetonitrile was degassed
by several freeze-pump-thaw cycles at liquid nitrogen tem-
perature and made up into 4-5% mixtures in N2 in a 5-L bulb.
d. Pulse Radiolysis-UV Absorption Experiment at Risø.
The pulse radiolysis transient UV absorption spectrometer and
the experimental procedure used are described in detail else-
where.22,23 Fluorine atoms were generated by irradiation of SF6/
CH3CN gas mixtures in a 1-L stainless steel reaction cell with
a 30-ns pulse of 2-MeV electrons from a Febetron 705B field
emission accelerator. The radiolysis dose could be varied using
steel attenuators. SF6 was always in great excess:
a. Ford FTIR-Smog Chamber System. The apparatus
and experimental techniques employed at Ford have been
described previously.17 The apparatus consists of a Mattson
Instruments Inc. Sirius 100 FTIR spectrometer interfaced to a
140-L, 2-m-long evacuable Pyrex chamber (S/V ) 0.14 cm-1).
The Pyrex chamber was surrounded by UV fluorescent lamps
(GTE F40BLB) which were used to generate halogen atoms
by the photolysis of molecular halogen.
Cl2 (or F2) + hν f 2Cl (or 2F)
(3)
White-type multiple reflection optics were mounted in the
reaction chamber to provide a total path length of 27 m for the
IR analysis beam. The spectrometer was operated at a resolution
of 0.25 cm-1
.
Reaction mixtures consisting of CH3CN, a reference com-
pound (CH4, CD4, or CF3CCl2H), and molecular halogen were
admitted to the reaction chamber in N2, O2, or air diluent. In
the presence of atomic chlorine, CH3CN and the reference
compound decay via reactions 1 and 4.
Cl + CH3CN f products
Cl + reference f products
(1)
(4)
SF6 + e- (2 MeV) f F + products
Transient absorptions were followed by multipassing the
output of a pulsed 150-W xenon arc lamp through the reaction
cell using internal White cell optics. A total path length of 120
Provided that CH3CN and the reference are lost solely by
reactions 1 and 4 and that neither is re-formed in any process,

662 J. Phys. Chem., Vol. 100, No. 2, 1996
Tyndall et al.
cm was used. A McPherson grating spectrometer, Hamamatsu
R 955 photomultiplier, and Biomation 8100 wave form digitizer
were used to isolate, detect, and record the light intensity at the
desired wavelength. Ultraviolet absorption spectra could be
recorded using a Princeton Applied Research OMA-II diode
array equipped with an image intensifier (Type 1420-1024HQ)
installed at the exit slit of the monochromator in place of the
photomultiplier. Gas mixtures were prepared by adding one
component at a time and measuring the corresponding partial
pressure on a MKS Baratron-170 absolute membrane manometer
with a resolution of 10-5 bar. Reagent concentrations used were
as follows: SF6, 1000 mbar; CH3CN, 0-2 mbar. All experi-
ments were performed at 296 K. SF6 (99.9%) was supplied by
Gerling and Holz, and CH3CN (>99.9%) was obtained from
Merck. All reagents were used as received.
Figure 1. Plot of the decay of CH3CN versus those of CH4, CF3CCl2H,
and CD4 in the presence of Cl atoms in the following diluents: 700
Torr of air (filled circles), 700 Torr of N2 (open circles), 700 Torr of
O2 (open squares), 15 Torr of O2 (open diamonds), and 5 Torr of N2
(open inverse triangles) at 296 K.
Results
a. Relative Rate Studies of the Reactions of Cl and F
Atoms with CH3CN at 296 K. Initial experiments in which
Cl2 was photolyzed in the presence of acetonitrile showed
formation of HCl, indicating that a reaction does indeed occur
at room temperature. When N2 was used as the bath gas, only
a short chain decomposition was observed, consistent with the
observations of Olbregts at higher temperature.9 This observa-
tion implies that the reaction of CH2CN with Cl2 (5) is very
slow, as is also the case for other methyl radicals with
electronegative substituent groups (e.g., CH2F, CF3, etc).24
CH2CN + Cl2 f CH2ClCN + Cl
(5)
Other products were observed in nitrogen bath gas, correspond-
ing to further substitution of hydrogen atoms by chlorine (see
later). The products formed in air will be addressed in a separate
publication.
The relative rate technique relies on the assumption that both
CH3CN and the reference compound are removed solely by
reaction with halogen atoms. To verify this assumption,
mixtures of the halogen (either chlorine or fluorine) with CH3-
CN and each of the references used were prepared and allowed
to stand in the dark. In all cases, the reaction of the organic
species with molecular halogen, in the absence of ultraviolet
light, was of negligible importance (<2%) over the typical time
periods used in this work. To test for possible photolysis,
mixtures of CH3CN and the reference compounds were prepared
and irradiated for 2 min in the absence of added halogen. No
photolysis (<2%) of any compound was observed.
The kinetics of reaction 1 were measured relative to reactions
6-8, and reaction 2 was measured relative to reactions 9 and
10.
Figure 2. Plot of the decay of CH3CN versus those of CD4 and CH4
in the presence of F atoms at 700 Torr of air (filled circles) or N2
(open circles) at 296 K.
Cl + CH3CN f products
F + CH3CN f products
Cl + CD4 f CD3 + DCl
Cl + CF3CCl2H f CF3CCl2 + HCl
Cl + CH4 f CH3 + HCl
F + CD4 f CD3 + DF
(1)
(2)
Figure 3. Plots of the decay of CH3CN versus CD4 in the presence of
F atoms at total pressures of 200 Torr (triangles), 30 Torr (circles), 10
Torr (inverse triangle), and 5 Torr (diamonds).
(6)
(7)
As shown in Figure 1, there was no discernible effect of total
pressure over the range 5-700 Torr, or O2 partial pressure over
the range 15-700 Torr, on the kinetics of reaction of Cl atoms
with CH3CN. Linear least-squares analysis of the data in Figure
1 gives k1/k6 ) 1.85 ( 0.10, k1/k7 ) 0.85 ( 0.05, and k1/k8 )
0.11 ( 0.01. The uncertainties quoted are 2 standard deviations.
(8)
(9)
F + CH4 f CH3 + HF
(10)
Using k6 ) 6.1 × 10-15 25
,
k7 ) 1.2 × 10-14 25
and k8 ) 1.0 ×
,
The observed loss of CH3CN versus CD4, CF3CCl2H, and CH4
in the presence of Cl atoms, and that of CH3CN versus CD4
and CH4 in the presence of F atoms, are shown in Figures 1-3.
10-13 12,13
,
gives k1 ) (1.13 ( 0.06) × 10-14, (1.02 ( 0.06) ×
10-14, and (1.10 ( 0.10) × 10-14 cm3 molecule-1 s-1
,
respectively. The results obtained using the different references

Reactions of Acetonitrile with Cl and F
J. Phys. Chem., Vol. 100, No. 2, 1996 663
TABLE 1: Values of k1 Measured in This Study
temperature
(K)
rate coefficientb
methoda
(10-14 cm3 molecule-1 s-1
)
296
274
296
325
336
345
296
R-R
R-R
R-R
R-R
R-R
R-R
FP-RF
1.08 ( 0.12
0.70 ( 0.15
1.22 ( 0.15
2.37 ( 0.30
2.90 ( 0.45
3.49 ( 0.30
1.24 ( 0.20
a R-R, relative rate technique; FP-RF, flash photolysis with resonance
fluorescence. b Uncertainties are 2 standard deviations plus a contribu-
tion to account for uncertainties in the reference reactions, etc.
TABLE 2: Measured Values of k2/k9 at 296 K
press.
(Torr) diluent
press.
(Torr) diluent
Figure 4. Kinetic data for the reaction F + CH3CN from the present
work (filled circles) and Smith et al.15 (open triangle) and Hoyermann
and Seeba16 (open square) versus total pressure. The solid curve is a
fit of the Troe expression to the data (see text for details).
a
a
k2/k9
k2/k9
5
10
30
N2
air
N2
(0.48 ( 0.04)
(0.62 ( 0.04)
(0.94 ( 0.07)
100
200
700
air
N2
air
(1.21 ( 0.10)
(1.27 ( 0.07)
(1.31 ( 0.06)
a Uncertainties are 2 standard deviations.
4 is a three parameter fit of expression II to all the data.
k0k∞[M]
2
-1
0.6{1+[log(k [M]/k )] }
(II)
show good internal consistency. Errors quoted thus far reflect
statistical uncertainty. In addition, we estimate that there is an
additional 10% systematic uncertainty associated with the values
of the reference rate coefficients. We choose to quote a final
value for k1 that is an average of the determinations above with
error limits that encompass the extremes of the determinations
plus an additional uncertainty to account for uncertainties in
the reference rate coefficients. Conventional error analysis
techniques were used to propagate the uncertainties. Hence,
k1 ) (1.08 ( 0.12) × 10-14 cm3 molecule-1 s-1. This average
value is given in the first line of Table 1.
0
∞
k ) kabst
+
k0[M] + k∞
Expression II is the standard Troe expression which has been
modified to include the possibility of a nonzero rate coefficient
at zero total pressure (kabst). Values of k0, k∞, and kabst were
varied simultaneously to provide the best fit values of k0 ) (2.9
( 2.1) × 10-28 cm6 molecule-2 s-1, k∞ ) (5.8 ( 0.8) × 10-11
cm3 molecule-1 s-1, and kabst ) (0.9 ( 0.4) × 10-11 cm3
molecule-1 s-1. Quoted uncertainties are 2 standard deviations.
b. Relative Rate Studies of Reaction 1 as a Function of
Temperature. The stainless steel chamber at NCAR was used
to determine the temperature dependence of reaction 1 between
274 and 345 K. Conditions used were similar to those just
described. Experiments were carried out using both N2 and O2
as bath gases. Methane was used as the reference gas for all
the experiments. Although the rate coefficients for the two
reactions differ by approximately a factor of 8.2 at room
temperature, which is not ideal for a relative rate experiment,
CH4 is the only one of the reference gases for which the
temperature dependence is known with sufficient accuracy to
be used as a reliable reference. The experiments were carried
out at a resolution of 1 cm-1 using 200 coadded scans. Since
the spectra of methane and acetonitrile are both structured,
calibration experiments were performed using standard additions
of the two compounds to ensure that the spectra were linear at
this resolution over the range of concentrations encountered.
Figure 2 shows the data obtained for reaction 2 at 700 Torr
total pressure using both CH4 and CD4 as reference compounds.
Linear least-squares analysis of the data in Figure 2 gives k2/k9
) 1.31 ( 0.06 and k2/k10 ) 0.83 ( 0.04. The uncertainties
quoted are 2 standard deviations. Using k9 ) 4.7 × 10-11 26
and k10 ) 6.8 × 10-11 cm3 molecule-1 s-1 26 gives k2 ) (6.1 (
0.3) × 10-11 and (5.6 ( 0.3) × 10-11 cm3 molecule-1 s-1
,
respectively, at 700 Torr total pressure. The consistency
between results obtained using the different references is
gratifying. A series of experiments performed over the total
pressure range 5-700 Torr of either N2 or air diluent using
CD4 as reference revealed a substantial effect of total pressure
on k2. Representative data are shown in Figure 3, and linear
least squares analysis of the data in Figure 3 and analogous
plots yielded the rate coefficient ratios given in Table 2.
Calibration experiments were performed to ensure that the IR
features of CD4 and CH3CN were linear under the experimental
conditions employed. The data in Table 2 can be placed on an
At room temperature (296 K) the measured ratio was k1/k8
) 0.122 ( 0.010 (2 standard deviations, precision only), leading
to a rate coefficient of (1.22 ( 0.10) × 10-14 cm3 molecule-1
s-1 for reaction 1, using the latest recommendation for reaction
8.12,13 This result is in excellent agreement with that measured
at Ford and described in the previous section. The ratio k1/k8
was found to increase with increasing temperature, indicating
a larger activation energy for reaction 1 than for Cl + CH4 over
this temperature range. Experiments were carried out down to
274 K, but below this temperature the difference in the rate
coefficients became too large to allow accurate determination
of the ratio. Representative data from experiments at 274 and
345 K are shown in Figure 5. The kinetics appeared well-
behaved at all temperatures studied, and the determinations are
summarized in Table 1. The results can be combined to yield
the following Arrhenius expression for k1 (where the uncertain-
absolute basis using k9 ) 4.7 × 10-11 cm3 molecule-1 s-1 26
.
There are two previous measurements of k2, both at low total
pressure. Smith et al.15 measured a value of (1.3 ( 0.2) × 10-11
cm3 molecule-1 s-1 in Ar diluent at a total pressure of around
1 Torr, using HF chemiluminescence. Hoyermann and Seeba16
employed the discharge flow technique using He diluent at a
pressure of 1.2 Torr and found k2 ) (1.25 ( 0.13) × 10-11
cm3 molecule-1 s-1. For the sake of comparison with the data
measured here in N2 diluent, we have assumed that the third
body efficiency of He and Ar is 50% of that of N2. Figure 4
shows a plot of the values for k2 derived in this work as a
function of total pressure together with the corrected values of
Smith et al. and Hoyermann and Seeba. The solid line in Figure

664 J. Phys. Chem., Vol. 100, No. 2, 1996
Tyndall et al.
Figure 7. Second-order plot for Cl + CH3CN measured by resonance
fluorescence at 296 K.
Figure 5. Plots of the decay of CH3CN versus that of CH4 in the
presence of Cl atoms at (a) 273 K and (b) 345 K.
by the reaction of CH2CN with Cl2.
Cl2 + CH2CN f CH2ClCN + Cl
(5)
Some experiments were also carried out in the presence of O2
in an attempt to scavenge CH2CN radicals. However, this was
found to lead to an increase in the apparent rate coefficient. On
closer inspection it was found that the fluorescence decays from
the runs containing O2 were curved downward, indicating that
a reactive product was being formed. The most likely candidate
is the cyanomethylperoxy radical NCCH2O2, which, by analogy
with methylperoxy and ethylperoxy, would be expected to react
rapidly with Cl,28 competing with CH3CN for Cl atoms.
O2 + CH2CN + M f NCCH2O2 + M
(11)
Cl + NCCH2O2 f NCCH2O + ClO
(12)
Figure 6. Arrhenius plot for Cl + CH3CN measured in this study:
open circles, relative rates measured at NCAR; open triangles, relative
rates measured at Ford; filled circle, flash photolysis. Uncertainties are
2 standard deviations, precision only.
Since the relative rate experiments suggested that the reaction
of CH2CN with Cl2 is slow, experiments were carried out in
the absence of O2, but with as low a Cl2 concentration as
possible so as to avoid regeneration of Cl atoms.
ties include possible systematic errors):
The data were fitted where possible using an unweighted,
linear least-squares fit of ln(fluorescence) against time. In the
case of the rapid decays, the necessity to use low [Cl2] meant
that the signal was soon lost in the noise, and so a commercial
nonlinear least-squares fitting routine was used. When this
routine was used to fit the slower decays, comparable first-
order rate coefficients were obtained by both methods (within
10% of each other). The first-order rate coefficients measured
are plotted against acetonitrile in Figure 7. Due to the severe
restrictions imposed by the attenuation of the signal and the
requirement of using low [Cl2], only a small range of acetonitrile
k1 ) (1.70 ( 0.25) ×
10-11 exp[-(2140 ( 200)/T) cm3 molecule-1 s-1
The measurements of k1 made in this study are summarized in
Figure 6. The preexponential factor is reasonable for an
abstraction reaction by Cl, while the activation energy is larger
than that for Cl with either CH4 or CH3Cl, although the C-H
bond is weaker. Knable and Kurylo rationalized the trends
observed for substituted methanes in terms of the electron
affinity or the dipole moment of the reagent. The current results
are consistent with their hypothesis that the high activation
energy for reaction 1 is associated with the extremely large
dipole moment of acetonitrile.
concentrations could be used, (1.0-8.5) × 1015 molecule cm-3
.
Nevertheless, the rate coefficient is well-defined, with a value
of (1.24 ( 0.10) × 10-14 cm3 molecule-1 s-1. In view of the
slowness of the reaction and the experimental difficulties
involved, along with the possibility that some regeneration may
have occurred, we choose to quote a somewhat larger overall
uncertainty, i.e., k1 ) (1.24 ( 0.20) × 10-14 cm3 molecule-1
s-1. This result is in good agreement with the two relative rate
experiments presented earlier in this paper.
The purity of the acetonitrile was checked using FTIR
spectroscopy to look for reactive impurities (specifically HCHO,
CH3OH, and CH3CHO). Upper limits of a few percent were
found for each. Although this impurity level is not low enough
to preclude a contribution to the Cl atom reactivity, the excellent
agreement with the relative rate experiments suggests that no
reactive impurities were present.
c. Absolute Measurement of k1 at 296 K. An absolute
determination of k1 was made at room temperature (296 K) by
monitoring the time resolved resonance fluorescence from
chlorine atoms in the presence of acetonitrile. As discussed
by Kurylo and Knable,10 Cl atom fluorescence was reduced in
the presence of acetonitrile, with the extent of attenuation
consistent with the absorption cross sections near 135 nm given
by Suto and Lee.27 The experiments were thus carried out at
fairly low acetonitrile concentrations, using 80 Torr of nitrogen
as diluent to reduce the rate of loss of Cl atoms by diffusion to
a value less than the loss by reaction with acetonitrile. The
Cl2 concentration was also kept as low as possible (<2 × 1014
molecule cm-3) in order to minimize regeneration of Cl atoms

Reactions of Acetonitrile with Cl and F
J. Phys. Chem., Vol. 100, No. 2, 1996 665
Figure 10. Plot of k1st versus [CH3CN] for experiments conducted at
1000 mbar total pressure using monitoring wavelength of 348 nm.
Figure 8. Spectrum acquired 1 µs after irradiation of a mixture
containing 5 mbar of CH3CN and 995 mbar of SF6. The resolution
was 0.5 nm, and the dwell time was 100 ns.
Figure 11. Results of relative rate experiments involving CH2ClCN
and CH2FCN. Circles are the decay of CH2ClCN versus those of CH3-
CN (in air, filled symbols) and CH4 (in N2, open symbols) in the
presence of Cl atoms. Triangles are the decay of CH2FCN versus that
of CD4 in the presence of F atoms in N2 diluent. All experiments were
performed at 296 K and 700 Torr total pressure.
Figure 9. Absorption at 348 nm observed following the pulsed
radiolysis (using 0.32 of maximum dose) of a mixture of 0.53 mbar of
CH3CN and 999 mbar of SF6. The trace is the result of a single pulse
with no signal averaging. The solid line is a least-squares fit of the
expression Abs(t) ) (A0 - Ainf) exp(-k1stt) + Ainf to the data, where
Abs(t) is the absorbance as a function of time, and A0 and Ainf are the
absorbances at time t ) 0 and infinity, respectively.
tion in Figure 10, where it can be seen that k1st values increased
linearly with [CH3CN]. Linear least-squares analysis of the data
shown in Figure 10 gives k2 ) (5.1 ( 0.5) × 10-11 cm3
molecule-1 s-1 at 296 K and a total pressure of 1000 mbar of
SF6. The y-axis intercept of (1.1 ( 1.7) × 105 s-1 is not
statistically significant. The results are in good agreement with
the high pressure limiting rate coefficient obtained at Ford using
the relative rate technique.
e. Mechanistic Study of the Reactions of Cl and F Atoms
with CH3CN. To provide information on the mechanism of
reactions 1 and 2, experiments were performed at Ford Motor
Co. in which CH3CN/Cl2 and CH3CN/F2 mixtures were subject
to UV irradiation in 700 Torr of N2.
d. Pulse Radiolysis Study of Reaction 2. Following the
pulse radiolysis of SF6/CH3CN mixtures, a rapid increase in
UV absorption at several wavelengths between 280 and 350
nm was observed. Figure 8 shows the spectrum acquired 1 µs
after radiolysis of a mixture of 5 mbar of CH3CN and 995 mbar
of SF6. The absorption consisted of two sets of relatively narrow
vibrational bands, with the strongest being at 348 nm. The
disappearance kinetics of the bands suggested that two distinct
species were absorbing. Control experiments were performed
in which either 1 mbar of CH3CN or 1000 mbar of SF6 was
subjected to pulse radiolysis; no significant transient absorption
at 348 nm was detected (<0.01 absorbance units). Figure 9
shows typical data obtained following radiolysis of a mixture
of 0.53 mbar of CH3CN and 999 mbar of SF6. The trace shown
is a result of a single pulse with no signal averaging. The rise
in absorption is presumably due to a direct product of reaction
2, and so the kinetics can be determined by monitoring the rate
of appearance of this product. The smooth curve in Figure 9
is a first-order fit to the trace which yields a pseudo-first-order
rate coefficient, k1st ) (7.9 ( 1.2) × 105 s-1. The observed
values of k1st are plotted as a function of the CH3CN concentra-
Cl + CH3CN f products
(1)
F + CH3CN f products
(2)
CH2CN radicals formed via the H atom abstraction channel of
reactions 1 and 2 are expected to react with molecular chlorine
or molecular fluorine to give CH2ClCN or CH2FCN, and so
the yields of CH2ClCN and CH2FCN provide insight into the
relative importance of the H atom abstraction channels of
reactions 1 and 2. As a preliminary exercise, experiments were
performed using the relative rate technique to measure the rates
of reactions 13 and 14. Results are shown in Figure 11.
Cl + CH2ClCN f products
(13)

666 J. Phys. Chem., Vol. 100, No. 2, 1996
Tyndall et al.
very short in these experiments, implying that termination
reactions compete with the chain propagation reactions. Thus,
considerable yields of the termination products such as NCCH2-
CH2CN are to be expected.
After subtraction of CH3CN and CH2ClCN from the product
spectra, residual IR features attributable to CHCl2CN and CCl3-
CN were observed. For completeness, a relative rate study of
the reactions of Cl atoms with each of these products was carried
out. Rate coefficient ratios for dichloroacetonitrile relative to
CHD3 and CH4 of 1.24 ( 0.11 and 0.27 ( 0.02, respectively,
were obtained. Using the available literature data for the
reference reactions gives k(Cl+CHCl2CN) ) (2.8 ( 0.4) ×
10-14 cm3 molecule-1 s-1. Quoted uncertainties are 2 standard
deviations. Chlorine atoms were found to react slowly with
trichloroacetonitrile, with a rate coefficient of the order of 10-15
cm3 molecule-1 s-1. However, the observed kinetic behavior
was not straightforward, and a reliable rate coefficient could
not be derived.
Figure 12. Formation of CH2ClCN following the UV irradiation of
CH3CN/Cl2 mixtures in 700 Torr of N2. The filled symbols are observed
data; the open symbols have been corrected for secondary reaction with
Cl atoms.
The yields of the products could all be modeled well using
the rate coefficients derived above; radical-radical termination
rate coefficients were fixed at 5 × 10-11 cm3 molecule-1 s-1
and rate coefficients for the reactions of the radicals with
molecular chlorine between 10-15 and 10-14 cm3 molecule-1
s-1. Finally, after subtraction of all known product features
from the infrared spectra, absorption bands due to one or more
unidentified products were observed at 792, 945, 985, 1024,
1143, 1652, and 1774 cm-1. The results of the product study
presented here show that the major pathway for reaction of Cl
atoms with acetonitrile is hydrogen atom abstraction, although
the occurrence of a slow reaction of Cl with Cl3CCN may
indicate that a minor addition channel may occur. Poulet et
al.11 also found evidence for consecutive chlorination reactions
at low pressures.
Discussion
Figure 13. Formation of CH2FCN following the UV irradiation of
CH3CN/F2 mixtures in 700 Torr of N2. The filled symbols are observed
data; the open symbols have been corrected for secondary reaction with
F atoms.
The results of the experiments presented here demonstrate
conclusively that a reaction between Cl atoms and acetonitrile
occurs at ambient temperatures. The three independent experi-
mental systems described here give indistinguishable results.
We recommend a value for k1 of (1.15 ( 0.20) × 10-14 cm3
molecule-1 s-1 at 296 K. The results can be combined to yield
a temperature-dependent rate coefficient of (1.6 ( 0.2) × 10-11
exp[-(2140 ( 200)/T]. This was derived by taking the
activation energy from the temperature dependence study of k1
and calculating a preexponential factor which reproduced the
mean value at 296 K derived from the independent studies. This
method guaranteed the most systematically consistent repre-
sentation of the temperature dependence of the data. All
measurements of k1 to date are summarized in Figure 14, along
with the line obtained as described above. The product study
shows that reaction 1 proceeds predominantly, if not exclusively,
by a direct abstraction reaction.
F + CH2FCN f products
(14)
Linear least-squares analysis of the data in Figure 11 gives k13/
k1 ) 3.00 ( 0.12, k13/k8 ) 0.34 ( 0.03, and k14/k9 ) 0.77 (
0.04 (where the uncertainties are 2 standard deviations). The
rate coefficient ratios k13/k1 ) 3.00 ( 0.12 and k13/k8 ) 0.34 (
0.03 can be combined to yield a ratio k1/k8 ) 0.11 ( 0.01. This
indirect result is indistinguishable from the value measured
directly as discussed in a previous section. Using the reference
rate coefficients k8 ) 1.0 × 10-13 and k9 ) 4.7 × 10-11 cm3
molecule-1 s-1, we derive k13 ) (3.4 ( 0.3) × 10-14 and k14
)
(3.6 ( 0.2) × 10-11 cm3 molecule-1 s-1
.
The observed yields of CH2ClCN and CH2FCN following
irradiation of CH3CN/Cl2/N2 and CH3CN/F2/N2 mixtures are
shown in Figures 12 and 13, respectively. The initial conditions
used were as follows: [CH3CN] ) 74-120 mTorr, [Cl2] ) 1.4-
5.4 Torr, [F2] ) 1.9-4.3 Torr, in 700 Torr of N2 diluent at 296
K. Corrections were computed for the secondary loss of CH2-
ClCN by Cl atom attack, and CH2FCN by F atom attack, using
the Acuchem chemical kinetic program29 with k13/k1 ) 3.0 and
k14/k2 ) 0.77/1.31 ) 0.59. Observed and corrected data are
given in Figures 12 and 13. There was no discernible effect of
molecular halogen concentration over the ranges used. The solid
lines in Figures 12 and 13 are linear fits to the data which give
yields of 81 ( 6% and 9 ( 1% for CH2ClCN and CH2FCN,
respectively. Errors quoted throughout are 2 standard devia-
tions. It is not surprising that the yields of the monohalogenated
products are not 100%. We know that the chain lengths are
The two previous measurements at ambient temperature were
in substantial disagreement with one another. Poulet et al.11
measured 9 × 10-15 cm3 molecule-1 s-1, in good overall
agreement with this determination, while Kurylo and Knable10
determined an upper limit of 2 × 10-15 cm3 molecule-1 s-1
.
The data obtained by Poulet and co-workers exhibited a curved
Arrhenius plot, and the data obtained at lower temperatures have
been discounted by most evaluations, since heterogeneous loss
of acetonitrile was suspected. However, the data at 296 K are
in agreement with those found here. There is no evidence in
our data for the curvature in the temperature dependence
observed by Poulet et al.11
The competitive chlorination experiments of Olbregts et al.9
yielded values for k1 at 370 and 413 K. They measured

Reactions of Acetonitrile with Cl and F
J. Phys. Chem., Vol. 100, No. 2, 1996 667
matrix and measured infrared and ultraviolet spectra of the
products.39 She observed the bands below 300 nm seen here
but attributed them to CH2CNH. It is unlikely that this molecule
could be formed in the pulse radiolysis experiments, and we
tentatively assign those bands to CH2CN. The bands below
359 nm could be associated with a fluorine atom adduct,
although the good correspondence between the rate coefficients
measured by fluorine atom loss and by the relative rate method
suggest that the ultimate fate of the adduct is reaction.
It is interesting that the reaction of fluorine atoms with
acetonitrile involves an addition mechanism, while that of
chlorine atoms appears to proceed entirely by abstraction. The
reaction of chlorine atoms with alkenes and alkynes is thought
to proceed by electrophilic addition. The strong carbon-
nitrogen triple bond and high ionization potential of acetonitrile
probably make this mechanism unfavorable. However, it is not
clear why the addition of fluorine atoms is so rapid.
Figure 14. Arrhenius plot showing all measurements of k1: filled
triangles, this work; filled circles, Poulet et al.;11 open triangles, Olbregts
et al.;9 open square, upper limit of Kurylo and Knable;10 line, fit to
data described in text.
Although the reaction of F atoms with CH3CN is very fast,
with a rate coefficient of (5.5 ( 1.5) × 10-11 cm3 molecule-1
s-1 at 296 K and ambient pressure, the concentration of F atoms
in the stratosphere is very small, of the order of 1 molecule
production of CH2ClCN versus CCl4 in the competitive pho-
tochlorination of mixtures of CH3CN and CHCl3.
cm-3 40
, and so reaction 2 will not impact the calculated profiles
of CH3CN.
Cl + CHCl3 f HCl + CCl3
(15)
(16)
Acknowledgment. NCAR is sponsored by the National
Science Foundation. Part of the work carried out at NCAR
was funded through NASA’s Upper Atmosphere Research
Program, Grant W-18067. The authors thank Chris Cantrell
and Mark Smith of NCAR and Steve Japar of Ford for their
comments on the manuscript.
CCl3 + Cl2 f CCl4 + Cl
Olbregts et al. evaluated the available data to reach their
reference expression, based largely on the direct data of Clyne
and Walker30 and indirect data of Knox31 and others (discussed
in ref 31). The data produced by Clyne and Walker are usually
regarded as being too high, and current evaluations12,13 are based
on the data of Knox. However, recent relative rate studies by
Wallington32 and Beichert et al.33 at room temperature gave a
rate coefficient k15 ) 1.13 × 10-13 cm3 molecule-1 s-1, which
is roughly 50% higher than the room temperature value derived
from the data of Knox. In order to reevaluate the data of
Olbregts et al., we have taken the room temperature value of
k15 measured by Wallington, with the activation energy recom-
mended in refs 12 and 13 (the mean of two determinations by
Knox). The revised values for k1 measured by Olbregts then
become 4.6 × 10-14 and 9.4 × 10-14 cm3 molecule-1 s-1 at
370 and 413 K, respectively. The revised data of Olbregts et
al. are included in the Arrhenius plot (Figure 14) and are found
to be in excellent agreement with the current data set.
In the stratosphere, the strong temperature dependence of k1
means that it will not compete with the OH reaction, since OH/
Cl is typically about 20, and kOH/kCl is calculated to be around
3-4 for all temperatures encountered in the atmosphere.12
Nevertheless, the reaction will have a larger effect than
previously thought, especially at higher altitudes, where CH3-
CN profiles may be more sensitive to reaction 1.
We have reported the first measurements of the pressure
dependence of the rate coefficient for the reaction of fluorine
atoms with acetonitrile. There is evidence from the FTIR
product study that abstraction occurs at least part of the time,
and the time-resolved UV experiments show the production of
two absorbing species. The species OCN and NHCN, which
are both isoelectronic with CH2CN, are known to have banded
spectra between 300 and 420 nm,34-36 and it would be expected
that CH2CN would also absorb in this region. Although CH2-
CN has been observed in interstellar space through its millimeter
wave spectrum,37,38 its ultraviolet spectrum has not been
positively identified. The spacings of the vibrational progres-
sions in Figure 8 are 400 ( 50 cm-1 for the bands below 350
nm and 980 ( 20 cm-1 for the bands below 300 nm. Jacox
codeposited CH3CN and excited argon atoms in a cryogenic
References and Notes
(1) Hamm, S.; Warneck, P. J. Geophys. Res. 1990, 95, 20 593.
(2) Arnold, F.; Bo¨hringer, H.; Henschen, G. Geophys. Res. Lett. 1978,
5, 653.
(3) Henschen G.; Arnold, F. Geophys. Res. Lett. 1981, 8, 999.
(4) Arnold, F.; Henschen, G.; Ferguson, E. E. Planet. Space Sci. 1981,
29, 185.
(5) Arijs, E.; Ingels, J.; Nevejans, D. Nature 1978, 271, 642.
(6) Ingels, J.; Nevejans, D.; Frederick, P.; Arijs, E. Planet. Space Sci.
1987, 35, 685.
(7) Brasseur, G.; Arijs, E.; De Rudder, A.; Nevejans, D.; Ingels, J.
Geophys. Res. Lett. 1983, 10, 725.
(8) Beig, G.; Walters, S.; Brasseur, G. J. Geophys. Res. 1993, 98,
12 767.
(9) Olbregts, J.; Brasseur, G.; Arijs, E. J. Photochem. 1984, 24, 315.
(10) Kurylo, M. J.; Knable, G. L. J. Phys. Chem. 1984, 88, 3305.
(11) Poulet, G.; Laverdet, G.; Jourdain, J. L.; LeBras, G. J. Phys. Chem.
1984, 88, 6259.
(12) DeMore, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.;
Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.; Molina,
M. J. Jet Propulsion Laboratory Publication 94-26, Pasadena, CA, 1994.
(13) 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.
(14) Hynes, A. J.; Wine, P. H. J. Phys. Chem. 1991, 95, 1232.
(15) Smith, D. J.; Setser, D. W.; Kim, K. C.; Bogan, D. J. J. Phys. Chem.
1977, 81, 898.
(16) Hoyermann, K.; Seeba, J. Z. Phys. Chem. (Munich) 1995, 188, 215.
(17) Wallington, T. J.; Japar, S. M. J. Atmos. Chem. 1989, 10, 399.
(18) Shetter, R. E.; Davidson, J. A.; Cantrell, C. A.; Calvert, J. G. ReV.
Sci. Instrum. 1987, 58, 1427.
(19) Tyndall, G. S.; Orlando, J. J.; Calvert, J. G. EnViron. Sci. Technol.
1995, 28, 202.
(20) Orlando, J. J.; Tyndall, G. S.; Moortgat, G. K.; Calvert, J. G. J.
Phys. Chem. 1993, 97, 10,996.
(21) Tyndall, G. S.; Kegley-Owen, C. S.; Orlando, J. J.; Calvert, J. G.
Manuscript in preparation.
(22) Hansen, K. B.; Willbrandt, R.; Pagsberg, P. ReV. Sci. Instrum. 1979,
50, 1532.
(23) Nielsen, O. J. Risø-R-480, 1984.
(24) Seetula, J. A.; Gutman, D.; Lightfoot, P. D.; Rayez, M. T.; Senkan,
S. M. J. Phys. Chem. 1991, 95, 10 688.
(25) Wallington, T. J.; Hurley, M. D. Chem. Phys. Lett. 1992, 189, 437.
(26) Wallington, T. J.; Hurley, M. D.; Shi, J.; Maricq, M. M.; Sehested,
J.; Nielsen, O. J.; Ellerman, T. Int. J. Chem. Kinet. 1993, 25, 651.

668 J. Phys. Chem., Vol. 100, No. 2, 1996
Tyndall et al.
(27) Suto, M.; Lee, L. C. J. Geophys. Res. 1985, 90, 13 037.
(28) Maricq, M. M.; Szente, J. J.; Kaiser, E. W.; Shi, J. J. Phys. Chem.
1994, 98, 2083.
(29) Braun, W.; Herron, J. T.; Kahaner, D. K. Int. J. Chem. Kinet. 1988,
20, 51.
(30) Clyne, M. A. A.; Walker, R. F. J. Chem. Soc., Faraday Trans. 1
1973, 69, 1547.
(31) Knox, J. H. J. Chem. Soc., Trans. Faraday Soc. 1962, 58, 275.
(32) Wallington, T. J. Unpublished results.
(33) Beichert, P.; Wingen, L.; Lee, J.; Vogt, R.; Ezell, M. J.; Ragains,
M.; Neavyn, R.; Finlayson-Pitts, B. J. J. Phys. Chem. 1995, 99, 13 156.
(34) Perry, R. A. J. Chem. Phys. 1985, 82, 5485.
(35) Herzberg, G.; Warsop, P. A. Can. J. Phys. 1963, 41, 286.
(36) Wu, M.; Hall, G.; Sears, T. J. J. Chem. Soc., Faraday Trans. 1993,
89, 615.
(37) Irvine, W. M.; Friberg, P.; Hjalmarson, A° .; Ishikawa, S.; Kaifu,
N.; Kawaguchi, K.; Madden, S. C.; Matthews, H. E.; Ohishi, M.; Saito, S.;
Suzuki, H.; Thaddeus, P.; Turner, B. E.; Yamamoto, S.; Ziurys, L. M.
Astrophys. J. 1988, 334, L107.
(38) Saito, S.; Yamamoto, S.; Irvine, W. M.; Ziurys, L. M.; Suzuki, H.;
Ohishi, M.; Kaifu, N. Astrophys. J. 1988, 334, L113.
(39) Jacox, M. E. Chem. Phys. 1979, 43, 157.
(40) Wallington, T. J.; Schneider, W. F.; Nielsen, O. J.; Sehested, J.
Faraday Discuss., in press.
JP9521417