Gas-phase kinetics of the self reactions of the radicals CH2F and CHF2

Article abstract of DOI:10.1524/zpch.2000.214.5.625

Fluorinated hydrocarbon radical-radical reactions in the gas phase have been studied at low pressure (0.5 ≤ p/mbar ≤ 2) and low temperature (253 ≤ T/K ≤ 333) using the discharge flow reactor molecular beam sampling mass spectrometry (MS) technique. Stable

Full text of DOI:10.1524/zpch.2000.214.5.625

Zeitschrift für Physikalische Chemie, 214, 5, 625Ϫ641 (2000)
by Oldenbourg Wissenschaftsverlag, München

Gas-Phase Kinetics of the Self Reactions
of the Radicals CH F and CHF2
2
1
2,
1
By Thomas Beiderhase , Walter Hack *, Karlheinz Hoyermann
and Matthias Olzmann
3
1
Institut für Physikalische Chemie, Universität Göttingen, Tammannstr. 6,
D-37077 Göttingen, Germany
Max-Planck-Institut für Strömungsforschung, Bunsenstr. 10, D-37073 Göttingen,
Germany
Institut für Physikalische Chemie, Universität Halle, FB Chemie Merseburg,
D-06099 Halle, Germany
2
3
Dedicated to Prof. Dr. Dr. h. c. mult. H. Gg. Wagner
on the occasion of his 70th birthday
(Received February 3, 2000; accepted February 8, 2000)
Reaction Kinetics / Radical Reactions / Mass Spectrometry
Fluorinated hydrocarbon radical-radical reactions in the gas phase have been studied at
low pressure (0.5 Յ p/mbar Յ 2) and low temperature (253 Յ T/K Յ 333) using the
discharge flow reactor molecular beam sampling mass spectrometry (MS) technique.
Stable and labile species have been detected by MS applying low energy electron impact
as well as multiphoton ionisation.
For the combination reaction
CH
2
F ϩ CH
2
F → products
(1)
12
(Ϫ3.9Ϯ1.0)
3
the rate coefficient k ϭ (7.0Ϯ0.8) · 10 (T/298)
low pressure the HF elimination pathway (1b)
cm /mol · s was determined. At
1
CH
2
F ϩ CH
2
F → C
2
H
3
F ϩ HF
(1b)
is the main channel (k(1b)/k
1
ϭ (0.82Ϯ0.015)).
For the CHF radical self reaction
2
CHF ϩ CHF → products
(2)
3
2
2
13
the rate coefficient was measured as k
mol · s. The stabilisation of C * is the main reaction pathway (2a) (k2a/k
HF elimination is of minor importance (k2b/k ϭ 0.2).
No pressure dependence of k and k was observed in the limited pressure range.
2
ϭ (1.7Ϯ0.5) · 10 exp((Ϫ555Ϯ89)/T) cm /
2
H
2
F
4
2
Ն 0.5), the
2
1
2
*
Corresponding author. E-mail: whack@gwdg.de
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

6
26
Th. Beiderhase, W. Hack, K. Hoyermann and M. Olzmann
1. Introduction
Fluorohalocarbons can accumulate in the urban atmosphere due to their low
reactivity. At higher altitudes fluorinated hydrocarbon radicals are formed
via photodissociation.
Data on elementary reactions of fluorinated radicals are essential, too,
for the description and the prediction of the chemistry of halogenated or-
ganic compounds under incomplete combustion as in waste incineration, or
pyrolysis, or under poor operating conditions, leading to products of incom-
plete combustion (PIC’s) [1, 2].
Whereas the self-reactions of CH and CF radicals lead to the combi-
3
3
nation products C H and C F , respectively, the CH F and CHF radicals
2
6
2
6
2
2
can show elimination and disproportionation besides combination; e.g.
CH
CH
2
FϩCH
2
F –→ products
(1)
(1a)
(1b)
(1c)
(1d)
M
2
FϪCH
2
F* –→ 1,2-C
2
H
4 2
F
–
→ C
2
H FϩHF
3
–→ cis/trans C
2
H
2
F
2
ϩH
2
CH
2
FϩCH
2
F –→ CHFϩCH
3
F .
For reaction (1) a room temperature rate constant was determined [3].
Detailed theoretical and experimental kinetic studies in the past have con-
centrated on the thermal decomposition of alkylfluorides (being products
of the combination reaction) and the decomposition of alkylfluorides after
chemical activation via energetically different reactions. For example the
formation of C
H
2 5
F* via the reactions:
C H ϩF
→ C H F*ϩF
2
5
2
2
5
CH
3
2
ϩCH
ϩCH
2
F → C
F → C
2
H
5
F*
CH
3
2
H
5
F*
is followed by the elimination of HF, where the critical energy for this route
is estimated to be in the range of 200Ϫ250 kJ/mol (for a review see for
example [4Ϫ6]).
The rate of the self reactions of the radicals CH
2
F (1) and CHF
2
CHF ϩ CHF
2
2
→ products
(2)
have not yet been studied in detail. The cross combination reaction of the
two radicals CH F ϩ CHF → products will be described elsewhere [7].
2
2
In this paper we report on the radical-radical reactions of CH F and
2
CHF radicals, respectively. The priority is set on the measurements of the
2
rate coefficients complemented with primary product determination for
a comparison with existing data from final product analysis [8, 9]. The
discharge-fast-flow-technique with molecular beam sampling is a suitable
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

Gas-Phase Kinetics of the Self Reactions of the Radicals CH
2
F and CHF
2
627
method to study these reactions. Mass spectrometry using laser induced
multiphoton ionisation (especially for radicals) and electron impact ionisa-
tion is applied for the species detection.
Fig. 1. Experimental arrangement (thermostated discharge flow reactor with time of flight
mass spectrometer).
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

6
28
Th. Beiderhase, W. Hack, K. Hoyermann and M. Olzmann
2. Experimental
The basic experimental set up consisted of an isothermal low pressure dis-
charge flow reactor, a molecular beam sampling system, and a mass spec-
trometer for the detection of stable and labile species. As parts have been
described elsewhere [10, 11] only a summary and the modifications will be
given.
The experimental arrangement is shown in Fig. 1. The reactor was a
conventional double walled flow reactor to be thermostated by a coolant
(
253 Յ T/K Յ 333) with an inner diameter of 20 mm. Additionally at room
temperature a flow reactor with an inner diameter of 36 mm was used. The
precursor molecules CH F and CH , respectively, were added to the main
flow via a movable inlet probe.
3
F
2 2
Samples from the flow reactors were withdrawn continuously into the
ion source of a mass spectrometer. A magnetic deflection type mass spec-
trometer with electron impact ionisation (4.5 Յ EI/eV Յ 29.5, EI ϭ 70 eV)
was combined with single ion counting techniques. Moreover a time-of-
flight mass spectrometer equipped with multiphoton ionisation, was used
for the detection of the radicals: CH
2
F (at λ ϭ 378.0 nm [12]), CHF (at
2
λ ϭ 386.65 nm [13]).
The radicals were produced by the reaction of F atoms (from a micro-
wave discharge in highly diluted F
2
/He mixtures) with the appropriate pre-
cursors, CH F (see [3])
3
F ϩ CH F → CH F ϩ HF
(3)
3
2
and CH F .
2
2
F ϩ CH F → CHF ϩ HF
(4)
2
2
2
The absolute F atom concentration was obtained from titration reactions
see below). Chemicals with the highest commercial grade available were
used (He Ն 99.996%), He/F (He: Ն 99.996; F Ն 99.6) (all Messer-
(
2
2
Griesheim); CH F (Ն 98.0), CH F (Ն 98.0), 1,1,2-C H F (Ն 99.0), C H F
3
2
2
3
3
3
2
3
(Ն 98.0), C
2
H
5
F (Ն 97.0), C
2
HF
3
(Ն 97.0) (all PCR/Heraeus); CH
4
(Ն 99.995) (Merck-Schuchardt).
3. Results
For the reactions (1) and (2) the rates were determined and the products
were measured. As the absolute radical concentrations enter into the evalu-
ation, the radical sources were studied first.
Procedure: radical production, absolute concentration calibration
The experimental procedures of the radical production and the determi-
nation of their absolute concentrations are to be described. The attribution
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

Gas-Phase Kinetics of the Self Reactions of the Radicals CH
2
F and CHF
2
629
Ϫ12
3
Fig. 2. The formation of CH
2
F, the radical source ([F]
0
ϭ 2 · 10 mol/cm , [CH
3
F]
1
0
ϭ
ϭ
ϭ
Ϫ10
3
12
3
2
6
5
· 10 mol/cm ; rate coefficients of the reactions: k
1
ϭ 3 · 10 cm /mol · s (a), k
1
2
3
13
3
13
3
· 10 cm /mol · s (b), k
1
ϭ 1.2 · 10 cm /mol · s (c).) (k
3
ϭ 1.7 · 10 cm /mol · s, k
6
Ϫ1
13
3
0 s , k
5
ϭ 5 ·10 cm /mol · s).
of absolute concentrations to the mass spectrometric signals are attained as
follows. Taking the CH F radical as an example, the reaction scheme to be
2
considered is given as:
CH
CH
CH
CH
3
FϩF
FϩF
Fϩwall → products
FϩCH FϩHF .
F → C
→ CH
2
FϩHF
(3)
(5)
2
→ CHFϩHF
2
(6)
2
2
2
H
3
(1b)
The rate coefficients of the reactions (3) and (5) are known from a previous
study (k ϭ 1.7 · 10 cm /mol · s, and k ϭ 5 · 10 cm /mol · s [3]). Using
mole fractions of CH
mol cm Յ 22 (the actual concentrations are given in the figure captions),
1
3
3
13
3
3
5
Ϫ10
3
F in the order of 1Ϫ3% i.e. 4 Յ [CH F]/10
3
Ϫ3
Ϫ10
Ϫ3
and decreasing amounts of F atoms in the range 0.03 Յ [F] /10 mol cm
0
Յ 3, the F atoms are consumed mainly via reaction (3). For these conditions
Fig. 2 illustrates the following: (i) reaction (5) does not contribute to the
consumption of F atoms when taking k values in a kinetically reasonable
5
1
4
3
range of (0.01Ϫ1) · 10 cm /mol · s around the quoted value of k , (ii) reac-
5
tion (5) does not remove CH
2
F radicals, (iii) the [F] concentration equals
0
the [CH F] concentration at small reaction times (0.2 Յ t/ms Յ 3). Thus
2
the absolute radical concentration [CH F] is known from the relation
2
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

6
30
Th. Beiderhase, W. Hack, K. Hoyermann and M. Olzmann
[
CH
consumption of CH
the difference of the CH
2
F]
0
ϭ [F]
0
ϭ ∆[CH
F, ∆[CH
F mass signals in the absence and presence of F
3
F] in the time range 0.2 Յ t/ms Յ 3. The absolute
3
3
F], is determined by mass spectrometry from
3
atoms and by the known flow of CH F. Moreover the mass spectrometric
3
signal at the mass of the radical is directly linked to an absolute radical
concentration. The ionisation of the radical can be achieved either by elec-
tron impact ionisation at low ionisation energy or multiphoton ionisation at
a specific wavelength. It is noteworthy to state that mixing problems (CH
to the F/He flow) and uncertainties in the reaction time axis do not enter
into the [CH F] concentration evaluation at these early reaction times. At
3
F
2
0
these low CH F radical concentrations, the second order combination reac-
2
tion (1) is suppressed. The first order wall removal is dominant for the
radical depletion and its rate can be determined by the variation of the
Ϫ1
reaction time. The rate coefficients k
6
varies in the range 30 Յ k /s Յ 50,
6
depending only slightly on temperature (10% in the temperature range 250
Ϫ1
1/2
Յ T/K Յ 340 K); a value of k ϭ 50 s (T/333 K) is used in the further
6
evaluations.
The rate coefficient of the combination reaction (1) is studied in a sec-
ond experiment by increasing the F atom concentration such that an appro-
priate decay of the CH
in the range of 5 Յ t/ms Յ 10. The F
controlled by the mass spectrometer. The F
2
F radicals can be measured within a reaction time
dissociation is still complete as
flow is low compared to the
mol/cm ; [F ] Ӷ [He], [CH F] (1.3 mbar)
2
2
Ϫ10
3
total flow ([F ] Ϸ 1.5 · 10
2
0
2 0
3
0
Ϫ8
3
Ϸ 5 ·10 mol/cm ). Thus the conditions in the reactor, the molecular beam
sampling, and the ion source are not changed by increasing the F atom
concentration. Therefore the ion signals of the CH F radicals at the con-
2
dition for studying reaction (1) are well attached to the absolute CH F con-
2
centration via the calibration procedure given above.
The concentration profiles for these higher F atom concentrations, still
fulfilling the relation [F] Ͻ [CH F] , are shown in Fig. 3. The conditions
0
3
0
are similar to those applied experimentally. The F atoms are nearly com-
pletely consumed at reaction times t Յ 1 ms, i.e. the reactions CH F ϩ F
3
→
CH
reaction times larger than 1 ms. The maximum CH
not equal to the [F] concentration. The actual absolute CH
2
F ϩ HF (3) and CH
2
F ϩ F → CHF ϩ HF (5) are unimportant for
F concentration does
F concentration
2
0
2
profile is derived from the calibration procedures. For reaction times greater
than 1 ms well behind the maximum of [CH F](t), for example for a refer-
2
ence time chosen to be tref ϭ 2 ms, the disappearance rate of the CH
radical is governed by the reactions (1) and (6).
2
F
The rate law for the CH F radical in the region where reactions (1) and
2
(
6) are important, is given by:
2
Ϫd[CH
2
F]/dt ϭ 2 k
1
[CH
2
F] ϩ k
6
[CH F] .
2
This differential equation is easily integrated with the boundary condition
CH F](tref) at the reference time t ϭ tref and t′ ϭ t Ϫ tref
[
2
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

Gas-Phase Kinetics of the Self Reactions of the Radicals CH
2
F and CHF
2
631
Fig. 3. The reaction CH
2
F ϩ CH
F] ϭ 2 ·10
ϭ 50 s , k
2
F (1): Simulation of concentration profiles ([F]
0
ϭ
ϭ
Ϫ11
3
Ϫ10
3
12
3
4
· 10
mol/cm , [CH
3
0
mol/cm , k
1
ϭ 6 · 10 cm /mol · s, k
3
1
3
3
Ϫ1
13
3
1.7 · 10 cm /mol · s, k
6
5
ϭ 5 · 10 cm /mol · s) (see text).
[CH
2
F](t) ϭ {[CH
2
F](tref) · k
6
· exp(Ϫk
6
· t′)}/
{
k ϩ 2 k [CH F](r )[1Ϫexp(Ϫk t′)]}
6
1
2
ref
6
Rearrangement leads to:
CH F](tref)/[CH F](t′) ϭ {1 ϩ 2 k
Ϫ 2 k [CH F](t )/k .
[
2
2
1
[CH
2
2
F](tref)/k
6
} exp(ϩk t′)
6
(I)
1
ref
6
Ϫ1
The condition k
6
· t′ Ӷ 1 holds for small reaction times (k
6
ϭ 50 s , t Ͻ 5
ms:k · t′ Ͻ 0.25). Thus series expansion to the first order gives:
6
[CH
2
F](tref)/[CH
2
F](t′) Ϸ 1 ϩ {2 k
1
[CH
2
F](tref) ϩ k
6
} · t′ .
1
(I )
A plot of the normalized inverse [CH
2
F] concentration vs. the reaction time
t′ ϭ t Ϫ tref should yield the slope (2 k [CH F](t ) ϩ k ) which is converted
1
2
ref
6
by the known values of the absolute [CH
2
F] concentration at the starting
to the wanted rate coefficient
point tref, [CH
2
F](tref), and the wall loss rate k
6
for combination, k . Where the first order series expansion of the ex-
1
ponential is a poor approximation, refinements are easily performed. It is
needless to state that the choice of tref and the exact numerical value is not
crucial to the evaluation of k
value of tref can easily be estimated via the rate coefficient of reaction (3)
with the given [CH F] concentration (tref ӷ t ϭ ln([F] /[F])(t)/k [CH F] ;
1
, it is only a reference point. The minimal
3
0
0
3
3
0
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

6
32
Th. Beiderhase, W. Hack, K. Hoyermann and M. Olzmann
i.e. [F]
the CH
It may be emphasized that the evaluation of k
0
/[F] Ͼ 100). This value can be checked experimentally by following
2
F concentration by the mass spectrometer.
1
is not performed by
fitting a CH F radical concentration profile by a complicated reaction mech-
2
anism. The results are based on measurements at three different experimen-
tal conditions: (i) the calibration of absolute CH F concentrations by a
2
titration reaction ([CH
3
F]
0
/[F]
0
ӷ 1, at small reaction times), (ii) the deter-
mination of the wall loss rate ([CH F] /[F] ӷ 1, at longer reaction times),
3
0
0
and (iii) the decay rate of CH
reaction times).
2
F radicals at ([CH
3
F]
0
/[F] Ͼ 1, at longer
0
The absolute F atom concentration was also obtained from the consump-
tion of CH (∆[CH ]) measured at m/z ϭ 16. This signal is not influenced
4
4
by C
CH
2
H
6
which is the final product in the reaction sequence F ϩ CH
4
→
3
ϩ HF, 2 CH . This method of determination of the F atom
3
→ C
2
H
6
concentration takes into account the F atom losses (e.g. the wall losses)
during the transit from the discharge to the reaction zone.
The reaction CH F + CH F → products (1)
2
2
Reaction rates
The rate coefficient k
1
and its temperature dependence (253 Յ T/K Յ 333)
was measured in two different experimental arrangements either by multi-
photon ionisation or by electron impact ionisation. The results are shown in
Fig. 4a, b. The CH F concentration profiles were followed either by elec-
2
tron impact ionisation at EI ϭ 14.0 eV (Fig. 4a) or by multiphoton ionis-
ation (see Fig. 4b) at λ ϭ 378 nm. Using multiphon ionisation the fragment
ion formation of products and the radical precursor CH
3
F at m/z ϭ 33 are
suppressed. From Eq. (I ) the rate constant k was determined. The reference
1
1
time tref (typically 2 ms) was chosen such that the mixing was complete
and a well-defined relation (distance in the flow reactor/reaction time) was
established. The absolute concentration of the CH F radical was determined
2
via a direct calibration of the mass spectrometric sensitivity of the CH2F
radicals (see above). The very low absolute [CH
2
F] concentration resulted
0
in low ion signals with restricted signal-to-noise ratios.
The wall loss rate of CH F radicals was measured in independent
2
experiments. (CH
3
F ϩ F → CH
2
3
F ϩ HF (3) (for example at 300 K: k ϭ
1
3
3
13
3
1
.7 · 10 cm /mol · s [12]); CH F ϩ F → products (5) (k ϭ 5 ·10 cm /
2
5
Ϫ1
mol · s, [3]; CH
2 6
F ϩ wall → products (k
ϭ 50 s ).) The lines represent
the rate coefficient k as shown in Fig. 4a, b.
1
The measurements at different initial conditions by the two experimental
arrangements yielded the results given in Table 1. The initial F atom con-
Ϫ10
Ϫ3
centration varied in the range (1.58 Յ [F] /10 mol · cm Յ 3.4).
0
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

Gas-Phase Kinetics of the Self Reactions of the Radicals CH
2
F and CHF
2
633
A
B
Fig. 4. Reaction CH
mbar; [CH F] ϭ 3.5 · 10
velocity v¯ ϭ 25.3 m/s; determination of [CH
2
F ϩ CH
2
F (1): Determination of the rate coefficient k
1
a) p ϭ 1.7
Ϫ10
3
Ϫ10
Ϫ3
3
0
mol/cm ; 1.6 Յ [F]
0
/10
mol cm Յ 1.8; linear flow
2
F] by electron impact (EI) mass spec-
trometry at an ionisation energy of EI: 14.0 eV. T: ᭡ ϭ 253, ᭿ ϭ 263, ᭢ ϭ 273, and
Ϫ9
3
Ϫ10
Ϫ3
᭹
ϭ 333 K; b) p ϭ 1.2 mbar, [CH
3
F]
0
ϭ 2.2 · 10 mol/cm , 2.1 Յ [F]
0
/10 mol cm
Յ 3.4; determination of CH
2
F by REMPI-MS; λ ϭ 378 nm; E ϭ 6.5 mJ; T: ᭡ ϭ 253 K;
᭿
ϭ 273 K; ᭢ ϭ 298 K; ᭹ ϭ 333 K.
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

6
34
Th. Beiderhase, W. Hack, K. Hoyermann and M. Olzmann
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

Gas-Phase Kinetics of the Self Reactions of the Radicals CH
2
F and CHF
2
635
As no systematic differences between the two sets of measurements are
observed, the values are averaged leading to the temperature dependence
n
using k
1
ϭ k (298 K)(T/298 K)
1
1
2
(Ϫ3.9Ϯ1.0)
3
k
1
(T) ϭ (7.0Ϯ0.8) · 10 (T/298)
cm /mol · s
which corresponds to the usual Arrhenius expression of
1
1
3
k (T) ϭ (1.4Ϯ0.3) · 10 exp((1156Ϯ265)/T) cm /mol · s
1
(error limits: one standard deviation; by regression analysis).
No pressure dependence was observed in the small pressure range.
Reaction products
For reaction (1) the following reaction scheme is considered:
CH F ϩ CH F –→ CH F Ϫ CH F*
2
2
2
2
2
M
CH
2
F Ϫ CH
F* –→ 1,2 Ϫ C
2
H
4
F
2
(1a)
(1b)
(1c)
(1d)
–
→ C
2
H F ϩ HF
3
–→ cis/trans-C
2
H
2
F
2
ϩ H
2
CH
2
F ϩ CH
2
F –→ CHF ϩ CH
3
F .
Mass spectrometric studies at low energy electron impact ionisation (EI ϭ
6 eV) showed signals at m/z with the intensity ratios given in brackets:
m/z ϭ 26 (17.8%), 27 (8.4%), 44 (15.3%), 45 (7.6%), 46 (100%), 64 (69%).
This product spectrum is to be compared with the mass spectrum of C
recorded under identical conditions (m/z ϭ 26 (14.1%), 27 (7.4%), 44
17.0%), 45 (10.4%), 46 (100%)), demonstrating the formation of C H F
1
2
H F
3
(
2
3
along reaction (1b). The calibration of the contribution of the reaction path-
way (1b) to the total CH F radical consumption via the reaction pathways
2
((1a) ϩ (1b) ϩ (1c) ϩ (1d)) was performed by measuring the formation
of C H F with respect to the consumption of CH F. These measurements
2
3
2
lead to the branching ratio (1b)/((1a) ϩ (1b) ϩ (1c) ϩ (1d)) ϭ (82Ϯ15)%.
The small signal at m/z ϭ 64 is attributed to cis/trans-1,2-C H F , since cis-
2
2 2
1
,2-C
H
2 2
F
2
and trans-1,2-C
2
H
2
F show the most intense signal at m/z ϭ 64.
2
Therefore a small contribution of channel (c) to the total reaction is postu-
lated since C H F shows no fragment ion at m/z ϭ 64.
2
4 2
The absence of an ion signal at m/z ϭ 66 (C
2
H
4
F ) down to an ionisation
2
energy of EI ϭ 13.5 eV indicates the absence of a stabilized product C H F
2
4 2
from the reaction pathway (1a) at these low pressures. The formation of
CHF (m/z ϭ 32) was not observed. This experimental finding, however, is
not sufficient to rule out reaction pathway (1d) completely. The dispro-
portionation reaction, pathway (1d), cannot be detected via the CH F prod-
3
uct since this molecule is present in large excesses as the radical precursor.
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

6
36
Th. Beiderhase, W. Hack, K. Hoyermann and M. Olzmann
Fig. 5. Reaction CHF
.2 mbar; linear flow velocity ϭ 17.7 m/s; T ϭ 298 K; [CH
ref ϭ 1.5 ms; determination of [CHF ] by REMPI-TOF mass spectrometry; evaluation
by relation (I ) and by computer simulation of the full mechanism (straight line), see text.
2
ϩ CHF
2
(2): Determination of the rate coefficient k
2
(p ϭ
Ϫ9
3
1
t
2
F
2
] ϭ 1.7 · 10 mol/cm ;
0
2
2
In summary the major product channel at low pressure is the elimination
of HF from the highly excited C H F * via channel (1b).
2
4 2
The reaction CHF
2
+ CHF
2
→ products (2)
Reaction rate
The rate coefficient k is determined via:
2
2
Ϫd[CHF
2
]/dt ϭ 2 k
2
[CHF
2
] ϩ k
7
[CHF ]
2
which leads to Eq. (I
2
) being completely analogous to Eq. (I
](t′) Ϸ 1 ϩ {2 k [CHF } · t′
1
) (see above):
(I )
[CHF
2
](tref)/[CHF
2
2
2
](tref) ϩ k
7
2
with
wall
––→ products .
CHF
2
(7)
ϩ HF (4)) is described
by the rate constant, k ϭ 5.9 · 10 cm /mol · s. The consecutive reaction:
The radical production rate (CH
2
F
2
2
ϩ F → CHF
2
1
3
4
CHF
2
ϩ F → products
(8)
can be neglected for the reason of a high excess ratio [CH
above).
2
F
2
]
0
/[F]
0
(see
The experimental conditions are given in Fig. 5 and in Table 2. The
Ϫ10
initial F atom concentration was varied in the range 2.1 Յ [F] /10
0
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

Gas-Phase Kinetics of the Self Reactions of the Radicals CH
2
F and CHF
2
637
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

6
38
Th. Beiderhase, W. Hack, K. Hoyermann and M. Olzmann
Ϫ3
mol · cm Յ 3.4 and the CH
2
F
2
concentration in the range 1.44 Յ
/10 mol · cm Յ 1.9. The excess ratio was always larger than
/[F] Ն 56. The rate constant for the heterogeneous CHF radical
Ϫ9
Ϫ3
[
[
CH
CH
2
F
2
F
2
]
2
]
0
0
0
2
removal reaction was determined in independent experiments analogous to
those described for the CH F radicals (see above). The results are given in
2
2
Table 2. The small positive temperature dependence of k is given by:
2
1
2
(1.9Ϯ0.3)
3
k (T) ϭ (2.6Ϯ0.6) · 10 (T/298 K)
cm /mol · s
2
or in Arrhenius form as:
1
3
3
k (T) ϭ (1.7Ϯ0.5) · 10 exp((Ϫ555Ϯ89)/T) cm /mol · s .
2
(The errors represent the one standard deviation of a least-squares fit.)
Reaction products
Reaction (2) was studied along the lines presented above. The working
hypothesis for the reaction mechanism is
CHF ϩ CHF –→ CHF Ϫ CHF *
2
2
2
2
M
CHF
2
Ϫ CHF
2
* –→ C
2
H
2
F
4
(2a)
(2b)
(2c)
(2d)
–
→ C HF ϩ HF
2
3
–→ C
2
F
4
ϩ H
2
CHF
2
ϩ CHF
2
–→ CF
2
ϩ CH
2
2
F .
The following product spectra were observed at different ionisation ener-
gies. At the ionisation energy EI ϭ 22.0 eV, the ratios were (i.e. signal
at m/z over the signal at m/z ϭ 82 at EI ϭ 22.0 eV are given in brackets):
m/z ϭ 63 (8%); m/z ϭ 82 (100%); m/z ϭ 83 (48%); m/z ϭ 102 (29%).
For the ionisation energy EI ϭ 19.5 eV the following product spectrum
was obtained: m/z ϭ 63 (Ϫ%); m/z ϭ 82 (94%); m/z ϭ 83 (19%); m/z ϭ
1
02 (18%) and at an ionisation energy of EI ϭ 17.0 eV the ratios: m/z ϭ
6
3 (Ϫ%); m/z ϭ 82 (43%); m/z ϭ 83 (Ϫ%); m/z ϭ 102 (2%).
The ratio of the product signals at m/z ϭ 63, 82, 83, 102 were dependent
on the ionisation energy indicating that they are not to be interpreted as
fragment ions of the same product species. The finding of a peak m/z ϭ
1
02 indicates the formation of C H F , due to channel (2a). The observation
2
2 4
of a peak m/z ϭ 82 has to be interpreted as the formation of C
2b). The ion fragmentation pattern at high ionisation energy (EI ϭ 70 eV)
of the possible products do not interfere at the following m/z values: C H F
2
HF , channel
3
(
2
2 4
(m/z ϭ 83), C
2
F
4
(m/z ϭ 100, 81) and C
2
HF
3
(m/z ϭ 82, 63). The measure-
ments showed only a signal increase at m/z ϭ 63, 82, 83 (pointing to C HF3
2
and C
H
2 2
F
4
) and no signal at m/z ϭ 81, 100 (excluding C
2
F , channel (2c)).
4
The disproportionation reaction (2d) cannot be detected directly via the
product CH F since CH F was present in high excess from the CHF rad-
2
2
2
2
2
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

Gas-Phase Kinetics of the Self Reactions of the Radicals CH
2
F and CHF
2
639
ical source (CH
CF , is also hard to detect directly. CF
indirectly via C since the rate of CF
2
F
2
ϩ F → CHF
2
ϩ HF). The other product of channel (2d),
radicals are also hard to detect
radical combination to give C
2
2
2
F
4
2
2
4
F :
10
3
CF ϩ CF → C F (9) is low (rate coefficient of k ϭ 2.6 · 10 cm /mol · s
2
2
2
4
9
[15]). Thus only high amounts of CF
2
would yield C
2
F under our con-
4
ditions. The absence of an ion signal of CF makes (2d) a minor channel in
2
agreement with the findings of Pritchard et al. [8], quoting Յ17% for chan-
nel (2d).
The calibration of channel (2b) by Ϫ2∆[CHF
2
]/∆[C
2
HF
3
2
] gave k2b/k ϭ
2
0%. The pathways (2c) and (2d) are of minor importance. Therefore the
stabilisation pathway (2a) is deduced to be the major route with an ef-
ficiency of Ͼ50% in the given pressure range.
4. Discussion
The rate coefficient for reaction (1) at room temperature is in agreement
12
3
with the value k (300 K) ϭ 7.5 · 10 cm /mol · s obtained earlier [3]. For
1
the other rate coefficient k (T) no direct measurements are published yet.
2
For either reaction, the kinetic behaviour observed is in accord with a
complex-formation mechanism [16Ϫ19], and we will discuss the results
from a common point of view, employing the general scheme
R ϩ R
→ R Ϫ R*
(A)
(D)
(E)
(S)
R Ϫ R*
R Ϫ R*
→ R ϩ R
→ alkene ϩ HF
R Ϫ R* ϩ M → R Ϫ R ϩ M .
The association reaction (A) leads to a chemically activated complex (*
denotes rovibrational excitation), which is subject to the competition be-
tween redissociation (D), elimination of HF (E), and collisional stabilization
(S) by the bath gas M. From elementary kinetic considerations, with the
steady-state approximation d[R Ϫ R*]/dt ϭ 0, one obtains for the rate of
the disappearance of R:
1
d[R]
dt
k
E
ϩ k
ϩ k ϩ k
and k depend on the average excitation en-
S
[M]
2
2
Ϫ
ϭ kobs[R] ϭ
k
A
[R]
(II)
2
k
D
E
S
[M]
where the rate coefficients k
D
E
ergy of R Ϫ R*, and k [M] ϰ p represents the pseudo-first-order rate coef-
S
ficient for the collisional stabilization. It becomes evident that the observ-
able rate coefficient kobs generally is a complicated function of the
temperature and the pressure and only in the case k
E
ϩ k
S
[M] ӷ k equals
D
the association rate coefficient k ; that is k , which is itself temperature
A
A
dependent, is accessible as the high-pressure limit of kobs
.
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

6
40
Th. Beiderhase, W. Hack, K. Hoyermann and M. Olzmann
Because of the energy dependence of the rate coefficients, it is necessary
to consider the thermochemistry of the reaction systems. For the sake of
consistency, we will do this in terms of BAC-MP4 heats of formation
throughout, which are shown to be in good agreement with experimental
data [20]. From the exothermicity of the association step, it follows an
excitation energy for the nascent intermediate population of 383 kJ/mol in
the case of CH
2
FCH
2
F* and 396 kJ/mol in the case of CHF
2
CHF *. Because
2
these quantities also represent the threshold energies for the redissociation
step, they are to be compared with the threshold energies for the competing
elimination reaction. Here, only little experimental information is available.
For reaction (1b), values of Ϸ220 kJ/mol [6] and ϳ260 kJ/mol [9] were
estimated from chemical activation experiments, and for reaction (2b), a
thermal activation energy of ϳ290 kJ/mol [21] was determined.
These values are clearly below the threshold energies for the redis-
sociation. Therefore, the latter should play only a minor role. Consequently,
according to Eq. (II), the observed rate coefficients for reactions (1) and (2)
are probably close to the corresponding association rate coefficients despite
the low pressures in our experiments. It is interesting, however, to notice
the differences. At room temperature, the value of k is larger than k by
1
2
about a factor of three, and, moreover, the temperature dependences are
opposite. Whether this is due only to differences in the association rate
coefficients k (T) or if the complicated interplay between redissociation,
A
elimination, and stabilization has an additional influence, still has to be
shown. A corresponding master equation analysis with specific rate coef-
ficients from statistical rate theory is in progress [22].
As was estimated above, the excitation energies for CH FCH F* and
2
2
CHF
2
CHF
2
* are comparable. The threshold energy for the elimination step,
however, is lower for CH FCH F than for CHF CHF . Therefore, HF split-
2
2
2
2
off from CH
2
FCH
2
F is faster than from CHF
2
CHF . This is consistent with
2
our experimental finding; the relative branching fractions for the elimi-
nation step are 80% and 20%, respectively. Consequently, the collisional
stabilization dominates in the latter case. This is also in agreement with the
final product analysis for similar conditions reported in the literature [8, 9].
The following ratios for the concentrations of stabilization of elimination
products, [S]/[E], were obtained:
reaction (1): [S]/[E] ϭ 0.042 [9] (this work: ϳ0)
reaction (2): [S]/[E] ϭ 24.2 [8] (this work: Յ0.5) .
As already mentioned, the thermal unimolecular degradation of fluori-
nated hydrocarbons proceeds by HF elimination, too [21, 23Ϫ25]. The re-
ported activation energies of 200Ϫ300 kJ/mol are appreciably lower than
the internal energies of our chemically activated species. Therefore the reac-
tion routes observed in our experiments are directly linked to the unimolecu-
lar decomposition of fluorinated hydrocarbons at high temperatures.
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM

Gas-Phase Kinetics of the Self Reactions of the Radicals CH
2
F and CHF
2
641
Acknowledgements
The continued interest of Prof. Dr. H. Gg. Wagner is gratefully acknowl-
edged. Thanks are due to the Deutsche Forschungsgemeinschaft (Euro-
päisches Graduiertenkolleg (EGK) “Microstructural Control in Free-Radical
Polymerization”) and the Fonds der Chemischen Industrie for financial sup-
port.
References
1
. W. M. R. Seeker, Twenty-Third Symposium (International) on Combustion, The
Combustion Institute, Pittsburgh, PA, p. 867 (1990).
2. C. P. Koshland, Twenty-Sixth Symposium (International) on Combustion, The Com-
bustion Institute, Pittsburgh, PA, p. 2049 (1996).
3
4
. Th. Beiderhase, W. Hack and K. Hoyermann, Z. Phys. Chem. 188 (1995) 227.
. P. J. Robinson and K. A. Holbrook, Unimolecular Reactions, Wiley-Interscience,
London, p. 217 (1972).
5
6
. A. Maccoll, Chem. Rev. 69 (1969) 33.
. J. A. Kerr, A. W. Kirk, B. V. O’Grady, D. C. Phillips and A. F. Trotman-Dickenson,
Discuss. Faraday Soc. 44 (1967) 263.
7
8
9
. Th. Beiderhase, K. Hoyermann and W. Hack, to be published.
. G. O. Pritchard and J. T. Bryant, J. Phys. Chem. 72 (1968) 1603.
. S. W. Benson and G. Haugen, J. Phys. Chem. 69 (1965) 3898.
1
1
1
1
1
0. R. Buth, K. Hoyermann and J. Seeba, Twenty-Fifth Symposium (International) on
Combustion, The Combustion Institute, Pittsburgh, PA, p. 841 (1994).
1. M. Bartels, J. Edelbüttel-Einhaus and K. Hoyermann, Twenty-Third Symposium (In-
ternational) on Combustion, The Combustion Institute, Pittsburgh, PA, p. 131 (1990).
2. J. W. Hudgens, C. S. Dulcey, G. R. Long and D. J. Bogan, J. Chem. Phys. 87 (1987)
4
546.
3. D. V. Dearden, J. W. Hudgens, R. D. Johnson III and B. P. Tsai, J. Phys. Chem. 96
1992) 585.
(
4. O. J. Nielsen, T. Ellermann, T. Bartkiewicz, T. J. Wallington and M. D. Hurley,
Chem. Phys. Lett. 192 (1992) 82.
1
1
1
5. S. Sharpe, B. Harnett and H. S. Sethi, J. Photochem. 38 (1987) 1.
6. J. Troe, J. Chem. Soc. Faraday Trans. 90 (1994) 2303.
7. J. Troe, Twenty-second Symposium (International) on Combustion, The Combustion
Institute, Pittsburgh, PA, p. 843 (1988).
1
8. I. W. M. Smith, J. Chem. Soc., Faraday Trans. 87 (1991) 2271.
9. R. G. Gilbert and S. C. Smith, Theory of Unimolecular and Recombination Reactions,
Blackwell, Oxford (1990).
1
20. C. F. Melius, BAC-MP4 Heats of Formation at 298 K and Error Estimate (Version
8
/28/97), Sandia National Laboratories, Livermore (1997).
2
2
2
2
2
1. G. E. Millward, R. Hartig and E. Tschuikov-Roux, J. Phys. Chem. 75 (1971) 3195.
2. M. Olzmann, W. Hack and K. Hoyermann, to be published.
3. J. A. Kerr and D. M. Timlin, Int. J. Chem. Kinet. 3 (1971) 427.
4. P. Cadman, M. Day and A. F. Trotman-Dickenson, J. Chem. Soc. A, 1356 (1971).
5. P. Cadman, M. Day and A. F. Trotman-Dickenson, J. Chem. Soc. A, 2498 (1970).
Brought to you by | The University of Auckland Library
Authenticated
Download Date | 6/25/15 4:21 AM