Reactions of CH3+ with C2N2, CH2CHCN, and HC3N: A low-pressure/high-pressure study

Article abstract of DOI:10.1021/jp9525954

The association of CH₃⁺ with the three molecules C₂N₂, CH₂CHCN, and HCCCN has been examined using ion cyclotron resonance (ICR) and selected ion flow tube (SIFT) techniques at room temperature. In each reaction, the mean lifetime of the complex (CH₃·N≡C-R⁺)* formed in the association has a major influence on the outcome of the reaction and the product channels that are observed using ICR and SIFT. Termolecular rate coefficients are reported for the association of CH₃⁺ + C₂N₂ for the bath gases M = He, Ar, N₂, and C₂N₂. k₃ = 8.2 × 10^-24 cm⁶ s^-1 (M = C₂N₂). In each system the association product channel occurs in competition with exothermic bimolecular channels. The complex lifetimes in all three systems are in the range 30-70 μs. Very rapid ion-molecule association reactions have been observed in several systems of hydrocarbons and nitriles, and the implications for Titan ion chemistry are discussed briefly.

Full text of DOI:10.1021/jp9525954

4032
J. Phys. Chem. 1996, 100, 4032-4037
+
Reactions of CH3 with C2N2, CH2CHCN, and HC3N: A Low-Pressure/High-Pressure
Study
M. J. McEwan,†,‡ D. A. Fairley,‡ G. B. I. Scott,‡ and V. G. Anicich*
Jet Propulsion Laboratory, 4800 Oak GroVe DriVe, Pasadena, California 91109
ReceiVed: September 5, 1995; In Final Form: December 5, 1995X
+
The association of CH3 with the three molecules C2N2, CH2CHCN, and HCCCN has been examined using
ion cyclotron resonance (ICR) and selected ion flow tube (SIFT) techniques at room temperature. In each
reaction, the mean lifetime of the complex (CH3‚NtCsR+)* formed in the association has a major influence
on the outcome of the reaction and the product channels that are observed using ICR and SIFT. Termolecular
+
rate coefficients are reported for the association of CH3 + C2N2 for the bath gases M ) He, Ar, N2, and
C2N2. k3 ) 8.2 × 10-24 cm6 s-1 (M ) C2N2). In each system the association product channel occurs in
competition with exothermic bimolecular channels. The complex lifetimes in all three systems are in the
range 30-70 µs. Very rapid ion-molecule association reactions have been observed in several systems of
hydrocarbons and nitriles, and the implications for Titan ion chemistry are discussed briefly.
Introduction
they are to planetary ionospheres. This is true for the lower
extents of most gas-enveloped planets and moons.
The unique blend of atmospheric species discovered on Titan
during the Voyager 2 fly-by1,2 has stimulated interest both in
the way in which the observed chemical species are formed
and in the range of molecules and ions that can be anticipated.
One of the most efficient ways of synthesizing molecules is
via the process of ion-molecule association in which an ion
and a neutral molecule combine to form a covalently bound
molecular ion of mass equal to the sum of the masses of the
reactant ion and neutral molecule.
Ion-molecule association (reactions 1a and 1b) followed by
dissociative recombination has been invoked as a mechanism
for synthesizing some of the molecules that have been observed
in interstellar clouds. Under the conditions of very low
temperatures and pressures that exist in the interstellar medium,
stabilization of the association complex is achieved by photon
emission (reaction 1b).3 In the laboratory at higher pressures,
however, the complex generally undergoes collisional stabiliza-
tion (reaction 1c) before it can emit a photon.4
We continue in this work, our studies of the factors that
influence the rates of ion-molecule association processes. We
+
present here the association behavior of CH3 with cyanogen
(C2N2), acrylonitrile (CH2CHCN), and cyanoacetylene (HC3N).
The techniques used for the study include ion cyclotron
resonance (ICR) and selected ion flow tube (SIFT). In the ICR
technique, the relative efficiencies of collisional stabilization
of different molecules can be compared, and this has been done
for the system CH3+/C2N2.
Experimental Section
The pressure dependent measurements (<10-3 Torr) were all
made using an ICR spectrometer at JPL operating at a 1.5 T
magnetic field.9 The ICR instrument was used in two different
operational modes. In the trapped mode, ions were contained
for different periods of time in a known pressure of reactant
gas before being released to drift through the cell and entering
the detection region. In the drift mode of operation the ions
drift through the cell, immediately after their formation, in a
higher pressure of reactant gas than in the trapped mode. The
rate of reaction is found from the rate of disappearance of the
reactant ion in the trapped mode and also from the ratio of the
reactant ion to product ion signals in the drift mode. Perhaps
the most significant difference between the trapped and drift
modes of operation of the ICR cell is that the pressure may be
sufficiently high in the latter mode for collisional stabilization
to compete with other avenues depleting the populations of the
A+ + B h (AB+)*
(AB+)* f AB+ + hν
(AB+)* + M f AB+ + M
(1a)
(1b)
(1c)
One ion that is thought to associate with many different
neutrals in interstellar clouds is CH3
+ 5
since it is relatively
+
unreactive with H2. In earlier studies of CH3 association
reactions with the nitriles HCN,6,7 CH3CN,4 and HC3N,8 we
found the association channel to be remarkably efficient even
when it competes with exothermic bimolecular channels.
Termolecular rate coefficients as large as 1.9 × 10-22 cm6 s-1
were found, for example, in the association between CH3+ and
CH3CN.4 Termolecular rate coefficients of this magnitude
enable three-body processes to compete effectively with bimo-
lecular reactions at pressures as low as 2 × 10-5 Torr. Of
course, such pressures are not relevant to interstellar clouds but
+
excited association complex. The branching ratios for CH3
reacting with CH2CHCN and HC3N were also obtained using a
tandem ICR instrument acquired recently.7
The higher pressure measurements (0.3 Torr) were made
using the Canterbury SIFT.10 In the SIFT technique, rate
coefficients are always determined from the rate of removal of
the reactant ion under pseudo first-order conditions.
Reagent species (except for HC3N) were obtained from
commercial sources and further purified by freeze-pump-thaw
cycles. All measurements were made at room temperature (298
( 5 K). HC3N was prepared by the reaction of methyl
propriolate and NH3 followed by dehydration with P2O5 as
described previously.11
† NRC-NASA Research Associate at JPL, 1993-94.
‡ Department of Chemistry, University of Canterbury, Christchurch, New
Zealand.
X Abstract published in AdVance ACS Abstracts, February 1, 1996.
0022-3654/96/20100-4032$12.00/0 © 1996 American Chemical Society

+
Reactions of CH3
J. Phys. Chem., Vol. 100, No. 10, 1996 4033
Results
+
CH3 + C2N2. The product channels observed by the ICR
and SIFT techniques are different in this reaction, and this
situation often occurs when termolecular association competes
with exothermic bimolecular channels. In the ICR, two channels
are observed (CH2CN+ + HCN production and the association
channel product CH3‚NCCN+) with a combined rate coefficient
k ) 9 × 10-11 cm3 s-1 at a pressure of 3 × 10-5 Torr (of only
C2N2).
CH3+ + C2N2 f CH2CN+ + HCN
f CH3‚NCCN+
(2a)
(2b)
[Note: The chemical structures of the adducts in this manuscript
are not known. We believe that the adducts have covalent
structures and that the potential wells associated with these
structures are greater than ∼150 kJ/mol deep. The chemical
symbols used here for the adducts, e.g., CH3‚NCCN+, are only
meant to indicate the reactant structures and not to contain any
information about the actual adduct structure.] The branching
ratio at this pressure was about 4:1 in favor of CH2CN+ + HCN.
In the SIFT instrument, only the association product was
observed and the pseudobimolecular rate coefficient was 1.7 ×
10-9 cm3 s-1 at a helium pressure of 0.3 Torr. The association
rate coefficient in the SIFT approximates the collision rate
coefficient12 of kcoll ) 1.9 × 10-9 cm3 s-1 and is expected to
be effectively invariant over the pressure range 0.3-0.5 Torr
in the SIFT.
Figure 1. Variation in rate coefficient ko2bs (cm3 s-1) against pressure
of C2N2 (Torr) for the reaction of CH3 with C2N2. The points are
experimental and the curve is the computer-generated line of the best
fit based on the model presented in eqs 3-6.
+
TABLE 1: Termolecular Rate Coefficients ko3bs Measured
+
for the Association Reaction CH3 + C2N2 + M f
CH3+‚NCCN+ + M with the Bath Gas M
a
M
ko3bs
âM/âC N
2
2
He
Ar
N2
C2N2
0.8
1.3
2.4
8.2
0.20
0.33
0.50
1.0
a In units of 10-24 cm6 s-1 b Collision efficiencies âM are relative
.
to âC N . As measured using drift-mode ICR.
2
2
In the ICR, however, the branching ratio changes with
increasing pressure in favor of the association channel CH3‚
NCCN+. By the measurement of the dependence of the rate
coefficient on pressure in the ICR instrument, the termolecular
rate coefficient for formation of CH3‚NCCN+ via collisional
stabilization was found. The rate coefficient at each pressure
was derived from the relative height of the mass peak at m/z )
15 Da (CH3+) to those of the products at m/z ) 40 Da (CH2-
CN+) and m/z ) 67 Da (CH3‚NCCN+) at a measured drift time
and C2N2 partial pressure. The reactions occurring in this
system may be written:
cm3 s-1). The results of this procedure gave the following
values for the rate coefficients: kf ) 2.0 × 10-9 cm3 s-1, k-1
) 276 000 s-1, k-2 ) 13 000 s-1, and k3 ) 8.2 × 10-24 cm6
s-1 for M ) C2N2, where k3 is defined in reaction 8.
k3
9
CH3+ + C2N2 + M
8 CH3‚NCCN+ + M
(8)
The mean lifetime with respect to unimolecular dissociation
of the (CH3‚NCCN+)* complex is τ*â ∼ 4 µs, which is a
consequence of both dissociation to reactants (fast) and dis-
sociation to products (slow).
CH3+ + C2N2 ykf z (CH3‚NCCN+)*
(3)
We also list in Table 1 the relative efficiencies of stabilization
and termolecular rate coefficients for the bath gases He, Ar,
N2, and C2N2. The termolecular rate coefficients for the
different bath gases were found from a least squares minimiza-
tion between the model calculation of reactions 3-6 and the
experimental data. The rate coefficients kf, k-1, and âkcoll for
C2N2 were fixed at the values obtained for M ) C2N2, and the
least squares minimization was carried out allowing â and kr to
vary.
k-1
kr
9
(CH3‚NCCN+)*
8 CH3‚NCCN+ + hν
(4)
(5)
(6)
k-2
(CH3‚NCCN+)* 8 CH2CN+ + HCN
âkcoll
(CH3‚NCCN+)* + M
8 CH3‚NCCN+ + M
CH3 + CH2CHCN. In the ICR instrument, the main
+
product ion of the reaction was at m/z ) 42 Da (CH3CNH+)
with minor product ions at m/z ) 27, 41, and 54 Da. The
product channels assigned to these ions on the basis of their
energetics are
If the observed bimolecular rate coefficient for the reaction is
ko2bs, then it has been shown that ko2bs may be expressed in terms
of the rate coefficient parameters for reactions 3-6 as4
ko2bs ) kf(kr + k-2 + âkcoll[M])/(k-1 + kr + k-2 + âkcoll[M])
CH3+ + CH2CHCN 0.808 CH3CNH+ + C2H2 (-232) (9a)
(7)
The coefficient â is a number between 0 and 1 that gives a
measure of the efficiency of stabilization of the (CH3‚NCCN+)*
complex by the bath gas. A model based on eqs 3-6 was used
to fit the experimental data, and the comparison of the model
calculations to the experimental results shown in Figure 1 for
M ) C2N2. The lower limit of kf was chosen, kf g 1.7 × 10-9
cm3 s-1. This was inferred from the pseudobimolecular rate
coefficient measured in the SIFT experiment, and âkcoll was held
at the collision-limiting rate coefficient12 (âkcoll ) 1.2 × 10-9
0.108 C2H3+ + CH3CN (-91)
0.058 C3H5+ + HCN (-196)
(9b)
(9c)
0.058 CH2CHCNH+ + CH2 (+30)
(9d)
Experimental branching ratios (using the ICR) and exother-

4034 J. Phys. Chem., Vol. 100, No. 10, 1996
McEwan et al.
+
TABLE 2: Comparison of Reactions of CH3 with C2N2, CH2CHCN, and HC3N Using ICR and SIFT Techniques
ICR
SIFT
a
branching
ratio
kICR
branching
ratio
kSIFT
reactant
C2N2
product channel
(10-9 cm3 s-1
)
product channel
(10-9 cm3 s-1
)
∼0.80
CH2CN+ + HCN
0.09 b
∼0.20
CH3‚NCCN+
1.0
CH3‚NCCN+
CH3CNH+ + C2H2
CH3‚NCCHCH2
1.7
5.5
CH2CHCN
∼0.80
∼0.10
∼0.05
∼0.05
>0.95
<0.05
CH3CNH+ + C2H2
C2H3+ + CH3CN
C3H5+ + HCN
5.4
0.35
0.65
+
CH2CHCNH+ + CH2
HCCCN
C3H3+ + HCN
1.5
0.40
0.60
C3H3+ + HCN
CH3‚NCCCH+
4.4c
CH3‚NCCCH+
a Flow tube pressure ) 0.30 Torr of helium. b At a pressure ∼ 3 × 10-5 Torr. The branching ratio is pressure dependent and the termolecular
rate coefficients for different bath gases were measured. k3 (M ) C2N2) ) 8.2 × 10-24 cm6 s-1; see text. c Results from refs 8 and 15.
micities (kJ mol-1) are also listed. In reaction 9a we have
designated CH3CNH+ as the ion product, but it could equally
be an isomer such as CH3NCH+. Channel 9d appears to be
endothermic by 30 kJ mol-1 according to the listed heats of
formation.13 There is, however, sufficient uncertainty in the
listed values to allow the 9d channel to be a few kJ mol-1
endothermic, which is more consistent with our observation of
this channel. No association product was observed in the ICR,
even at bath gas pressures approaching 5 × 10-4 Torr.
in the trapped mode occurred, forming a lower energy complex
that subsequently reacted with HC3N. The product of the
reaction may be a product ion already present in the cell. It is
not clear from our analysis why this decrease occurred. Similar
behavior was observed previously in the CH3+ + HCN system.7
The value of 1.5 × 10-9 cm3 s-1 is most likely the more more
accurate value.
Two studies of this reaction have been reported using the
SIFT technique, and both investigations reported the same
+
A comparison of these results to those of an earlier SIFT
study shows a quite different product distribution from the
reaction. The SIFT study14 found the major product of the
reaction to be the association adduct ion CH3‚NCCHCH2+ with
product channels as observed in the ICR; viz., C3H3 + HCN
and association. At 0.3 Torr of helium, channel 11a was 40%
of the total reaction and channel 11b 60% with a rate coefficient
k ) 4.4 × 10-9 cm3 s-1 8,15
.
This is very close to the collision
two other channels, c-C3H3 + CH2NH and H2C3N+ + CH4.
rate coefficient that is calculated to be 5.5 × 10-9 cm3 s-1
.
+
In view of the apparent discrepancy, we re-examined this
reaction in the SIFT, taking special care to eliminate all traces
of CH2+ contamination from the CH3+ reactant ion swarm. We
found that the product ions C2H3+, C3H5+, and CH2CHCNH+
earlier attributed to CH3+ were in fact due to CH2+ and H3O+,
which were also present in the flow tube. In the present re-
evaluation only two products were observed:
A summary of all the results is given in Table 2.
Discussion
The contrast in results between the low-pressure and high-
pressure techniques is very informative and can be rationalized
in terms of how long the (CH3‚NC-R+)* complex lives.
CH3+ + NC-R h (CH3‚NCR+)*
(12)
CH3+ + CH2CHCN 0.358 CH3CNH+ + C2H2 (10a)
In the CH3+/NCCN system, the low-pressure measurements
found the CH2CN+ + HCN product channel in addition to
association, whereas the SIFT measurements revealed the
association product only, which was produced in the SIFT at a
rate close to the collision rate for the reaction. At flow tube
pressures (0.3 Torr), there are sufficient collisions of the
complex with the helium bath gas to collisionally deactivate
0.658 CH3‚NCCHCH2
(10b)
+
The major channel, as before, was association, but the only other
channel of significance was CH3CNH+ + C2H2. The rate
coefficient measured in the SIFT was k ) 5.5 × 10-9 cm3 s-1
.
This is the same as the collision rate coefficient within the errors
of our measurement.
(CH3‚NCCN+)* before any other exit channel from the C3H3N2
+
surface is found. In the ICR experiment, the much lower
ambient pressures allow unimolecular dissociation of (CH3‚-
NCCN+)* to take place before collisional stabilization can occur.
The smaller rate coefficient in the ICR measurement (k ) 9 ×
10-11 cm3 s-1 at 3 × 10-5 Torr) indicates that unimolecular
dissociation back to reactants is much faster than dissociation
to CH2CN+ + HCN (k-1 . k-2). As the bath gas pressure
increases, the number of complexes stabilized by collision also
increases, and this is seen by the increase in k with increasing
pressure.
CH3+ + HC3N. Two product channels of the reaction were
found in the ICR. The major channel was C3H3+ + HCN, but
small signals corresponding to the association channel CH3‚-
NC3H+ were detected at pressures approaching 5 × 10-4 Torr
of HC3N.
CH3+ + HC3N g0.958 C3H3+ + HCN
e0.058 CH3‚NC3H+
(11a)
(11b)
The measured termolecular rate coefficient for complex
formation allows an estimate of the mean complex lifetime
according to the model presented in reactions 3-6 with τ*â ∼
4 µs. An absolute value for τ cannot be deduced from the
pressure dependence of k alone without knowledge of the
collision stabilization efficiency, â. One piece of information
we learn from the SIFT experiment, where all complexes are
stabilized at 0.3 Torr of helium, is that the mean complex
lifetime is much greater than the time between collisions, viz.,
τ(CH3‚NCCN+)* . 0.25 µs. This is consistent with the
An unusual feature of the ICR study of this reaction was the
difference in rate coefficients measured using the trapped and
drift modes of operation. In the trapped mode, k ) 1.5 × 10-9
cm3 s-1, where k was found from the rate of disappearance of
CH3+. In the drift mode, k ) 7 × 10-10 cm3 s-1, where k was
+
derived from the relative ratio of the CH3 ion signal to the
product ion signals. We have not identified the process
responsible for the decrease in rate coefficient in the drift mode.
It is possible that some radiative stabilization of the complex

+
Reactions of CH3
J. Phys. Chem., Vol. 100, No. 10, 1996 4035
TABLE 3: Solutions from Crossover Fitting
k2a (ICR)
k2a (SIFT)
kf
kcoll
âk-1
âk-2
âτc
k3
19.0
k3a @ 1 × 10-4 Torr
% of k2
a
a
b
b
d
CH3CN
C2N2
HC3N
CH2CHCN
HCN
1.8
0.09
1.5
5.4
0.2
4.0
1.7
4.4
5.5
2.0
5.5
2.0
5.2
5.6
1.68
5.5
52 000
276 000
248 000
10 000
23 000
13 000
100 000
274 000
13.
3.5
2.9
3.5
1.6
1.94
5.33
5.62
4.5
0.82
3.8
0.1-0.5
0.12
0.016
11
4
532 000
1.1
a Units are (×10-9) cm3s-1
.
b Units are s-1 c Units are µs. d Units are (x10-23) cm6 s-1
. .
observation that τ*â ∼ 4 µs, and since â e 1, then τ g 4 µs.
are they due to termolecular processes that have a collision-
induced dissociative component? The only complexes that are
known to dissociate in the flow tube are very weakly bound
complexes, i.e., bound by less than a few tens of kJ. The
collision complexes in the systems that we are discussing here
are believed to have relatively deep potential wells of at least
100 kJ. Can the model and the values of the parameters that
we have deduced support the observation of products other than
the association adduct in the SIFT experiments?
From a past study we have found that â ∼ 0.094 (M ) CH3-
+
CN), and we expect that it will be about the same for CH3
/
C2N2 (M ) C2N2). This is the only experimentally determined
â. If the value of â for (CH3‚NCCN+)* is similar to that for
(CH3‚NCCH3+)*, then τ(CH3‚NCCN+)* ∼ 60 µs. We note that
values of â , 1 occur when collisional stabilization of the
complex occurs in competition with binary channels.
In the CH3+/CH2CHCN system, no association product was
observed in the ICR experiment, even at bath gas pressures as
high as 1 × 10-4 Torr. The absence of association and the
From the experimental data, the SIFT-measured rate gives kf
and the ICR rate gives the ratio of τ-2/τ-1. Four decades of
pressure are required to span the transition from bimolecular
kinetics only, at low pressures, to the region where every
complex formed is stabilized by collision. The crossover
between these two regimes, bimolecular kinetics and termo-
lecular kinetics, occurs in a specific place in each system (see
Figure 2 in the Appendix). This crossover defines the total
lifetime of the collision complex.
In the CH3+/CH2CHCN system, the reverse rate coefficient,
k-1, is much less than the k-1 values in the other two systems.
Nevertheless, it is not possible to make the k-2 value in this
system sufficiently large to account for the extent of dissociative
product formation that is observed in the SIFT experiment. It
would require a larger value of k-2 in the SIFT than was found
in the ICR experiments. The value of the rate coefficient
predicted for k-2 can account for about 5% of the SIFT products
but no more. Alternatively, if the termolecular rate coefficient
is reduced to k3 ) 3 × 10-25cm6 s-1, then the reaction has not
entered the “pressure-saturated” regime in the SIFT and larger
amounts of dissociative products can be expected.
large rate coefficient in the ICR (k ) 5.4 × 10-9 cm3 s-1
=
kcoll) support the conclusion that the exit channel from the (CH3‚-
NCCHCH2+)* complex is found so quickly after formation that
the complex lifetime is too short for collisional stabilization to
be competitive with unimolecular dissociation to products at
the pressures in the ICR instrument. Although the two collision
complexes, (CH3‚NCCN+)* and (CH3‚NCCHCH2+)*, have
about the same lifetimes, their decomposition pathways are quite
different in the ICR cell. In the case of (CH3‚NCCN+)*, the
lifetime is determined by the rapid return to reactants (k-1
.
k-2). In the (CH3‚NCCHCH2+)* case, the lifetime is defined
by the rapid rate toward products CH3CNH+ + C2H2 (k-1
,
k-2). However, since collisional stabilization is the major
+ *
pathway of the reaction in the SIFT, then τ(CH3‚NCCHCH2
)
is longer than the time between collisions, i.e., τ(CH3‚-
NCCHCH2+)* . 0.25 µs.
The CH3+/HC3N system is in between the behavior exhibited
in the previous two cases. The major product channel in the
ICR is C3H3+ + HCN, whereas the major channel in the SIFT
is association occurring in competition with the C3H3+ + HCN
product channel. This system therefore lies somewhere between
the previous two cases (k-1 ∼ 2k-2). Although the two collision
complexes, (CH3‚NC3H+)* and (CH3‚NCCHCH2+)*, have again
about the same lifetimes, their decomposition pathways are
slightly different. In the case of (CH3‚NC3H+)*, the lifetime is
defined by a slower return to reactants. Actually, the forward
and reverse rates are about the same within a factor of 2. The
mean complex lifetime, τ(CH3‚NCCCH+)* is again clearly larger
than the time between collisions in the SIFT (0.25 µs).
The latter two systems can be analyzed in a manner similar
to that for the analysis on the CH3+/NCCN system. Equations
similar to eqs 3-7 can be defined for the CH3+/CH2CHCN and
CH3+/HC3N systems. Although we were not able to measure
the termolecular association directly in the ICR for these two
systems, the nonobservation sets a lower limit on the overall
For the CH3+/HC3N system, the bimolecular process could
contribute up to 40% of the total products observed in the SIFT
but only if the unimolecular lifetime τ g 1 ms. However, a
lifetime this large is also inconsistent with the fact that we were
unable to make a direct experimental determination of the
lifetime in the ICR. With the current parameters deduced for
this system we can account for no more than 5% of the observed
SIFT dissociative products.
We have assumed in these three systems that kf ) kcoll. Under
these conditions in the SIFT, all three reactions should have
entered the pressure-saturated regime, and therefore, the extent
of dissociative channels should be minimal. It is not and this
raises the question of why.
As mentioned earlier, dissociative product channels might
arise from the association adduct through collisional breakup
of the adduct ion. This outcome is unexpected because of the
assumed stability of the association complex in these systems.
Alternatively, it is possible that two independent channels with
two different types of complexes may occur such that one
complex can be stabilized by collision but the other, which has
a very much shorter lifetime, cannot be collisionally stabilized
at the SIFT experimental pressures. Other association systems
have been described before as having two complexes, e.g.,
k3. By use of the known kobs value, the values kf, k-1, k-2, and
2
τ*â are estimated from the fit of the kinetics crossover. The
formulas used in this calculation are given in the Appendix.
The solutions are presented in Table 3 along with the values
from the CH3+/CH3CN and CH3+/HCN systems. One surprise
is that in all three present systems, the τ*â products are very
nearly the same in magnitude. The overall estimate for k3 for
CH3+/HC3N is expected to be about 3.8 × 10-23 cm6 s-1, which
is very fast for a termolecular process.
16
17
i-C3H7+/C2H5NH2 and C2N2+/ C2N2
.
There is one ambiguous point left, i.e., are the dissociative
products observed in the SIFT due to bimolecular processes or
It is not possible with the present experimental data to decide
which mechanism is responsible for the occurrence of dissocia-

4036 J. Phys. Chem., Vol. 100, No. 10, 1996
McEwan et al.
tive channels in the SIFT experiments, either the CH3+/CH2-
CHCN or the CH3+/HC3N system.
not exhibit fast termolecular association in the ICR. The CH3‚-
NCCN+ is thus a covalently bonded molecular ion.
Direct Measurement of Association Complex Lifetime.
We have recently developed a technique for measuring the mean
lifetime of ion-molecule association complexes.18 The tech-
nique requires the presence of a bimolecular product channel
competing with association. The three association reactions in
this study meet that requirement.
Finally, we assess the relevance of association processes to
the atmosphere of Titan. The atmosphere of Titan is predomi-
nately molecular nitrogen, but a substantial pressure of CH4 has
been detected.1,2 Other hydrocarbons that are also expected to
be present are acetylene, ethylene, and ethane. Model calcula-
+
tions predict a peak ion production for CH3 at an altitude of
1100 km21,22 for which the principal loss process is reaction
with CH4.
A+ + B h (AB+)*
(AB+)* f C+ + D
(13)
(14)
(15)
CH3+ + CH4 f C2H5+ + H2
(16)
(AB+)* + M f AB+ + M
It is apparent, however, from this and other studies that rapid
The basis of the technique is to monitor the bimolecular product
ion (C+) while employing double resonance ejection of the
precursor (AB+)* ion. A distribution of lifetimes is found from
the shape of the double resonance ejection curve vs the time of
ejection.18 In our instrument, the shortest ejection times,
corresponding to the largest Vrf amplitudes, are ∼70 µs. This
means complexes with lifetimes shorter than ∼70 µs will show
no double resonance behavior as they are lost from the cell by
unimolecular decomposition before they can respond to ejection
from their power absorption.
All three systems investigated were examined using this
technique. No significant double resonance response was
observed in any of the three systems, which is not inconsistent
with the earlier observation that τ*â (CH3‚NC-R+)* g 4 µs.
In these systems it is expected that â ∼ 0.1. Thus, we also
expect then that the lifetimes are between 30 and 70 µs. If we
make this assumption, then k3 (CH3+/HC3N) is estimated to be
about 3.8 × 10-23 cm6 s-1, and k3 (CH3+/ CH2CHCN) is
estimated from the model (see Appendix) to be between 1 and
5 × 10-24 cm6 s-1. These differences in the k3 values estimated
reflect more than just differences in the lifetimes of the collision
complexes. They also include the effect of changes in dipole
moments of the neutrals and the resultant effect this has on the
collision rate (or kf).
termolecular ion-molecule association reactions occur. The
+
directly measured association reaction in this study is CH3
+
C2N2, which requires an ambient neutral density of 6 × 1014
molecules cm-3 before it is competitive with fast bimolecular
reactions. [The nitrogen molecule has been found to be almost
as effective as the parent neutrals, as a relaxing third body.]
+
Some association reactions of CH3 are much faster than this.
+
For example, the CH3 + CH3CN association with a termo-
lecular association rate coefficient of 1.9 × 10-22 cm6 s-1 4 is
competitive with fast bimolecular reactions at a neutral density
of around 8 × 1012 molecules cm-3
. These densities are
achieved in the lower ionosphere of Titan with molecular
nitrogen as the bath gas. The relatively large mixing ratio of
CH4 in Titan’s atmosphere means that termolecular reactions
+
of CH3 do not constitute significant loss processes for
+
ionospheric CH3 ions. However, the same is not necessarily
+
true of ions that are unreactive with CH4 such as C3H5 and
H2CN+. What is also not clear is the effect of temperature.
The systems studied here are room temperature investigations,
and yet they show association competing efficiently with fast
exothermic channels. For association via collisional stabilization
to be competitive with an exothermic bimolecular reaction, there
must be a barrier in the exothermic exit channel that lengthens
the lifetime of the complex. Few systems have been as closely
investigated as the CH3+/CH3CN system.4,23 Extensive calcula-
Conclusions
+
tions of the C3H6N+ energy surface for the association of CH3
and CH3CN showed a barrier height to the exothermic channel
of only about 15 kJ mol-1 below the entrance energy to the
surface but 85 kJ mol-1 above the energies of the exothermic
products. As the temperature decreases towards the temperature
of Titan’s atmosphere, the existence of this barrier will favor
further the association process over the exothermic channel. We
therefore conclude that ion-molecule association processes
make up an important class of reactions that should be included
in models of Titan up to lower ionospheric altitudes. What is
needed, however, is a much more extensive data base of such
reactions.
The question of the relative efficiency of ion-molecule
association reactions and competing bimolecular channels is an
important one that has application to interstellar synthesis and
planetary atmospheres. The crucial parameter controlling the
relative efficiency of the association mechanism is the complex
lifetime, which is, in turn, influenced by the presence of fast
exit channels from the energy surface of the complex. In this
study, all three ion-molecule systems examined demonstrated
association as the major outcome of a reaction event in the SIFT.
Thus, at a flow tube pressure of 0.3 Torr, the time between
collisions is sufficiently short for collisional stabilization to be
the principal stabilizing mechanism.
Acknowledgment. The work described in this paper was
carried out at the Jet Propulsion Laboratory, California Institute
of Technology, under contract with the National Aeronautics
and Space Administration.
What we have not addressed in the present study is the more
fundamental question of why the lifetimes of the three (CH3‚-
NC-R+)* complexes should or should not differ. The existence
of exothermic reactions, the binding energy of the complex,
and the number of degrees of freedom are the chief factors that
influence the lifetime. Some attempts have been made to assess
these factors since they influence radiative association.19,20 None
of the structures of the association adducts in this study are
known, and therefore, the binding energies for the complexes
are not known. A comparison of the measured termolecular
rate for CH3+ + C2N2 with other R-CN systems that we have
studied4,6,7 suggests a binding energy of at least 150 kJ mol-1
for the CH3‚NCCN+ complex. Weakly bound complexes do
Appendix
The following set of chemical equations can be represented
as a set of partial differential equations. This is a parallel set
of equations to eqs 3-6 in the text.
A+ + B ykf z (AB+)*
(1)
k-1

+
Reactions of CH3
J. Phys. Chem., Vol. 100, No. 10, 1996 4037
termolecular reaction. This solution is used to show the
estimated crossover position for the system CH3+/CH2CHCN,
(Figure 2).
References and Notes
(1) Hanel, R.; Conrath, B.; Flaser, F. M.; Kunde, V.; Maguire, W.;
Pearl, J.; Pirraglia, J.; Samuelson, R.; Herath, L.; Allison, M.; Cruikshank,
D.; Gautier, D.; Gierasch, P.; Horn, L.; Koppany, R.; Ponnamperuma, C.
Science 1981, 212, 192.
(2) Kunde, V. G.; Aikin, A. C.; Hanel, R. A.; Jennings, D. E.; Maguire,
W. C.; Samuelson, R. E. Nature 1981, 292, 686.
(3) Bates, D. R.; Herbst, E. In Reaction Rate Coefficients in Astrophys-
ics; Miller, T. J., Williams, D. A., Eds.; Reidel: Dordrecht, 1988.
(4) McEwan, M. J.; Denison, A. B.; Huntress, W. T. Jr.; Anicich, V.
G.; Snodgrass, J.; Bowers, M. J. J. Phys. Chem. 1989, 93, 4064. Smith, S.
C.; Wilson, P. F.; Sudkeaw, P.; Maclagan, R. G. A. R.; McEwan, M. J.;
Anicich, V. G.; Huntress, W. T. J. Chem. Phys. 1993, 98, 1944. Anicich,
V. G.; Sen, A. D.; McEwan, M. J.; Smith, S. C. J. Chem. Phys. 1994, 100,
5696.
(5) Smith, D. Chem. ReV. 1992, 92, 1473.
(6) McEwan, M. J.; Anicich, V. G.; Huntress, W. T.; Kemper, P. R.;
Bowers, M. T. Chem. Phys. Lett. 1980, 75, 278.
(7) Anicich, V. G.; Sen, A. D.; Huntress, W. T., Jr.; McEwan, M. J. J.
Chem. Phys. 1995, 102, 3256.
(8) Knight, J. S.; Freeman, C. G.; McEwan, M. J.; Adams, N. G.; Smith,
D. Int. J. Mass Spectrom. Ion Processes 1985, 67, 317.
(9) McEwan, M. J.; Denison, A. B.; Anicich, V. G.; Huntress, W. T.
Jr. Int. J. Mass Spectrom. Ion Processes 1987, 81, 246.
(10) McEwan, M. J. In AdVances in Gas Phase Ion Chemistry; Adams,
N. G., Babcock, L. M., Eds.; J. A. I. Press: Greenwich, CT, 1992; Vol 1,
p 1.
(11) Sen, A. D.; Huntress, W. T.; Anicich, V. G.; McEwan, M. J. J.
Chem. Phys. 1991, 94, 5462, and originally in Miller, F. A.; Mellon, D. H.
Spectrochim. Acta, 1967, 23A, 1415.
Figure 2. Expected position of the crossover between bimolecular and
termolecular kinetics for the system CH3+/CH2CHCN. The effective
bimolecular reaction rate coefficient is plotted against the pressure of
both the parent gas and helium as buffers. The parent gas curves are
solid lines while the helium gas curves are dashed lines. The helium
data represent the sum of both channels that were observed in the SIFT.
The data points are taken from Table 2. The closed circles are the
bimolecular kinetics and the open circles are the termolecular kinetics.
kr
9
(AB+)*
8 AB+ + hν
(2)
(3)
(4)
k-2
(AB+)* 8 C+ + D
âkcoll
(AB+)* + M
8 AB+ + M
(12) Collision reaction rate coefficients were calculated using the
Langevin rate for neutrals without permanent dipole moments and using
the parametrized ion-polar collision rate coefficient equations of Su, T.;
Chesnavich, W. J. J. Chem. Phys. 1982, 76, 5183 for neutrals with
permanent dipole moments.
(13) Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin,
R. D.; Mallard, W. G. J. Phys. Chem. Ref. Data, Suppl. 1988, 17.
(14) Petrie, S. A. H.; Freeman, C. G.; McEwan, M. J. Mon. Not. R.
Astron. Soc. 1992, 257, 438.
(15) Bohme, D. K.; Raksit, A. B. Mon. Not. R. Astron. Soc. 1985, 213,
717.
(16) Meot-Ner, M. J. Am. Chem. Soc. 1979, 101, 2389.
(17) McEwan, M. J.; Anicich, V. G. J. Phys. Chem. 1995, 99, 12204.
(18) Anicich, V. G.; Sen, A. D.; Huntress, W. T., Jr.; McEwan, M. J. J.
Chem. Phys. 1991, 94, 4189.
(19) Dunbar, R. C. Int. J. Mass Spectrom. Ion Processes 1990, 100,
423.
(20) Herbst, E.; Dunbar, R. C. Mon. Not. R. Astron. Soc. 1991, 253,
341.
(21) Keller, C. N.; Cravens, T. E.; Gan, L. J. Geophys. Res. 1992, 97,
12117.
(22) Dalgarno, A.; Fox, J. L. In Unimolecular and Bimolecular Ion-
molecule Reaction Dynamics; Ng, C. Y., Baer, T., Powis, I., Eds.; John
Wiley & Sons: Chichester, England, 1994; p 1.
The set of partial differential equations was solved for the
concentration of the chemical species in the equations. The
reaction rate coefficient for the loss of A+ is k2(A+), the
formation reaction rate coefficient for C+ is given by k2(C+),
and the formation reaction rate coefficient for AB+ is given by
k2(AB+). The equations in kf, k-1, kr, k-2, and âkcoll[M] are
given below.
k2(A+) ) kf(k-2 + âkcoll[M])/(k-1 + kr + k-2 + âkcoll[M])
k2(C+) ) kfk-2/(k-1 + kr + k-2 + âkcoll[M]) ) k2(ICR)
k2(AB+) ) kf(kr + âkcoll[M])/(k-1 + kr + k-2
+
âkcoll[M]) ) k2(SIFT) @ 0.3 Torr of He
It is assumed in this formalism that the [C+] are dissociative
products that are seen in the ICR in these experiments. It does
not include any C+ that may result from dissociation associated
with the termolecular reaction. The [AB+] is assumed to include
the radiative association products and all the products of the
(23) Herbst, E.; McEwan, M. J. Astron. Astrophys. 1990, 229, 201.
JP9525954