Full text of DOI:10.1007/PL00004053

Shock Waves (2001) 10: 445–456
Infrared absorption and thermal decomposition of some CFCs
T. Tsuboi, K. Ishii, K. Itoh
Department of Mechanical Engineering and Materials Science, Faculty of Engineering, Yokohama National University,
7
9–5 Tokiwadai, Hodogaya, Yokohama 240–8501, Japan
Received 26 January 2000 / Accepted 8 August 2000
2 2 2 3
Abstract. The spectra of CCl F , CHClF and CCl F were observed in the wavelength range 7–13 µm.
The decadic molar extinction coefficients were measured and compared with the data calculated in this
work and also with the HITRAN database. Usingthe shock tube technique the emission spectra at high
temperature were measured and usingthese emission spectra the pressure-dependent unimolecular rate
2 2 2
constants of thermal decomposition of CCl F and CHClF were measured in the temperature range
−
5
−4
3
1
200–1900 K and in the density range 10 –10 mol/cm .
Key words: Thermal decomposition, Freon, IR-absorption, IR-emission
1
Introduction
to the results for the VUV region (Chou et al. 1978, Lacher
et al. 1950, Zobel and Duncan 1955). These molecules
Knowledge about molecular absorption coefficients is of have no emission and absorption spectra at around 3–
significant value for the optical sounding of gaseous me- 4 µm. This region is widely used for the measurement of
dia. The experimental spectra were widely used for the hydrocarbon molecules, and, fortunately, there are sev-
determination of local thermodynamic parameters in gas eral suitable photodetectors and window materials which
mixtures. In the laboratory, both low- and high-resolution are easy to use and inexpensive. Freons have spectra at
techniques have been used for various purposes. In the UV around 10 µm (Herzberg 1945(1)), and the number of de-
range, we have measured temperature profiles in propa- tectors and window materials decreases for this spectral
gating flames and detonations (Tsuboi et al. 1975,Tsuboi region.
et al. 1981). An infrared He–Ne laser was used for the
determination of the density of the absorbing molecules the decadic molar extinction coefficients of CO2 and some
Tsuboi et al. 1996). Measurements at a fixed wavelength Freon molecules at CO2 laser wavelengths around 10.6 µm,
The present study is, therefore, an attempt to measure
(
give only one piece of information, but the implementation and to observe the absorption spectra of Freon molecules
generally involves low cost and little experimental effort. around 10 µm, using black body radiation as a light source.
The infrared He–Ne laser (3.39 µm) is a low-cost, easy-to- Furthermore, the thermal decompositions of CCl2F2 and
use tool for the study of media involving hydrocarbons CHClF2 were observed using emissions at around 10 µm.
(Tsuboi et al. 1990,Tsuboi et al. 1985). The infrared CW
CO2 laser (10.6 µm) can also be a tool for the measure-
ment of CO2 concentration and Freon, though it is a bit
difficult to use, because the light intensity (several watts)
is too high to use as an infrared light source.
Adding to these, precise data on the infrared spectra
of CO2 provide for information on Earth-warming calcu-
lation. Knowledge about the infrared spectra of Freon is
also needed for the solution of environmental problems,
and especially the elemental thermodynamic data for the
C–Cl bond and for the related molecules, required for the
development of dioxin disposal techniques. In order to ob-
tain the thermochemical data, we also need to obtain the
IR spectral and kinetic data at high temperatures, adding
2
Experiments
The experiments were performed in a stainless-steel shock
tube with an inner diameter of 100 mm. The test section
was 4.2 m long with a 2 m-long driver section. Four piezo-
electric pressure transducers, set at intervals of 400 mm,
measured incident shock velocities. The test section was
evacuated to 0.5 Pa and the pressure rise due to the leak-
age (and also due to some desorption of the gas adsorbed
onto the tube wall) was 0.02 Pa/min. Before we let the test
gas flow into the test section, we cleaned the tube with
the test gas several times. Thus, we considered that the
residual gas in the tube after this procedure was mainly
Correspondence to: T. Tsuboi (e-mail: tsuboi@ynu.ac.jp)
An abridged version of this paper was presented at the 22nd the test gas mixture itself. In order to know the effect of
Int. Symposium on Shock Waves at London, United Kingdom, impurities in the test section, we compared the emission
from July 18 to 23, 1999
profiles after the first cleaning with those after the fifth

4
46
T. Tsuboi et al.: Thermal decomposition of Freons
cleaning, and we determined that there was no difference. up to four or five figures using a normal monochromator,
We used inert gas (helium) as the driver gas in order to whose wavelength scale was determined using a He–Ne
avoid the influence of hydrogen. Furthermore, the follow- laser (632.82 nm and 3.3922350 µm) in advance. However,
ing calculation was made in order to know the impurity due to the play in the wheel of the monochromator we
effect due to leakage. We usually performed the experi- could not decide the wavelength more precisely. A wave-
ment within 1 min after closing the valve of the vacuum length of more than 6 figures could be estimated by com-
pump. This experimental time was equivalent to the pres- paring the measured density dependence of the molar ex-
sure rise of oxygen by 0.004 Pa due to the inflow of air. tinction coefficients of CO2 with the calculated values. The
3
4
We used 4×10 –4×10 Pa as the initial pressure P1. In wavelengths used here are, therefore, those which could be
the case using 0.04% test gas in argon, the initial test gas determined from the CO2 spectra given in the HITRAN
4
pressure was 4 Pa at P1=1×10 . This meant that the im- database (Rothman et al. 1996). Since the light intensity
purity due to oxygen was 1/1000 and the impurities due of the CO2 laser was too strong, we decreased it by re-
to other molecules were much less significant than this flecting the light twice on ZnSe plates.
value. Therefore, we decided that there was no influence
of impurities on the decay profiles of the test gas.
The light emission at the central axis of the shock
The decadic molar extinction coefficients of pure CO2
molecules were calculated as follows:
3
For the density ρ mol/cm and the temperature T K of
tube and at the point 16 mm from the end wall went the light-absorbing molecule A, the decadic molar extinc-
2
through the ZnSe window of the shock tube, and was fo- tion coefficient εA(ρ, T) cm /mol is given by
cused by the ZnSe object glass onto the entrance of a
I
monochromator. A long bandpass filter was set in front
of the monochromator in order to filter out the radiation
below 7 µm and above 15 µm. A HgCdTe photodetector
−ρ εA(ρ,T ) d
=
10
(1)
I
0
(
Hamamatsu Photonics) was used to detect the IR inten- where d is the optical path length in cm. εA(ρ, T) is related
−1
sity. The light absorption was measured using a similar to the spectroscopic absorption coefficient KA(ρ, T) cm
optical system but without the focusing system. A CO2 by
laser (Nihon Kagaku Engineering) was used as the light
source at 10.6 µm. The density dependence of the decadic
molar extinction coefficients of CO2 molecules was mea-
sured using the CO2 laser light source and compared with
the data calculated using the theoretical model of the light
extinction, in order to determine the precise wavelength
of the CO2 laser. The measurement technique and theo-
retical calculations are given in detail in Sect. 3.1.
Test gases of 0.02, 0.05, 0.2 and 1.0% CCl2F2, CHClF2
and CCl3F diluted with argon were prepared in steel con-
tainers and used after 24–48 h mixing time. The steel con-
tainers set horizontally were rolled and rocked occasion-
ally during the experiments. The purities of the test gases,
which were used without further purification, were: CO2 >
KA(ρ, T) = ρ εA(ρ, T) ln(10)
(2)
The absorption coefficient for the wave number σ is given
by:
ꢀ
KA(σ, ρ, T) = ρ
Sl φl(ρ, T)
(3)
l
2
where Sl(T) in cm /mol is the integrated intensity of line
l and φl is the line shape. The dependence of Sl on the
temperature is
Sl(T)
Sl(Ts)
QRV (Ts) 1 − exp(−hσl/kT)
QRV (T) 1 − exp(−hσl/kTs)
=
ꢁ
ꢂ
ꢃꢄ
ꢀꢀ
−E
1
1
l
×
exp
−
(4)
9
9.9%, CCl2F2 > 99.9% (H2O < 10 ppm, CCl3F > 99.9%,
k
T
Ts
O2 < 0.5 ppm, residuum < 20 ppm), CHClF2 > 99.9%
(
H2O < 10 ppm, O2 < 0.5 ppm, residuum < 20 ppm), O2 > where Ts is a standard temperature (296 K in the HITRAN
9
9.9% (THC < 30 ppm, H2O < 10 ppm), H2 > 99.99% (O2 database), while Eq. (4) involves the line position σl and
ꢀꢀ
<
10 ppm, CO < 10 ppm, CO2 < 10 ppm, THC < 10 ppm, the energy E of the lower level of the transition; QRV is
l
H2O < 10 ppm), Ar > 99.9995% (O2 < 0.5 ppm, CO < the rovibrational partition function for A. The line shape
1
ppm, CO2 < 1 ppm, THC < 0.5 ppm, H2 < 0.5 ppm, is a Voigt profile, given by
H2O < 5 ppm).
ꢅ
ꢆ
+
∞
2
ln 2
π παD(T)
y
exp(−t )
φl(ρ, T) =
with
dt (5)
2
2
y + (z − t)
√
γl(ρ, T)
−
∞
3
Results
√
3
.1 Carbon dioxide spectra
σ − σl
z = ln 2
; y = ln 2
(6)
αD(T)
αD(T)
The infrared CO2 laser has several tens of lines around the
φ depends on the line position σl, the pressure broadening
parameter γl(ρ, T) (Lorentz half-width at half-height) and
the Doppler width αD(T). The value of γl is given by
1
0.6 µm region (Levine 1968). In our laser system a fixed
wavelength could not be obtained, since the wavelength of
the laser changed according to the laser cavity condition,
especially according to the cooling water temperature. Be-
cause of this, we could fortunately use the laser at vari-
ous wavelengths. The wavelength could be determined to
ꢂ
ꢃβl(A)
P
Ts
s
γl(ρ, T) = γ (A)
(7)
l
Ps
T

T. Tsuboi et al.: Thermal decomposition of Freons
447
–1
6
: 942.3834[cm ] = 10.611392[µm]
–
1
6
: 934.8945[cm ] = 10.696394[µm]
5,6
1
5 : 10.611394[µm]
4 : 10.611396[µm]
3
2
1
0
5,6
1
10
10
10
10
5: 10.696391[µm]
4: 10.696389[µm]
3
2
1
0
1
1
1
0
0
0
ε
ε
T=297 K
00 % CO2
1
T=297K.
00 % CO2.
1
wavelength 10.696394 [µm]
wavelength 10.611392µm.
8
7
6
: 10.696387[µm]
: 10.696384[µm]
: 10.696382[µm]
3 : 10.611399[µm]
2 : 10.611401[µm]
1
: 10.611403[µm]
10
–6
–4
–6
–4
3
10
10
10
10 ρ [mol/cm ]
3
ρ [mol/cm ]
Fig. 3. Density dependence of decadic molar extinction coef-
ficients of CO . Solid lines show the calculated dependence at
various wavelengths around 10.6114 µm
Fig. 1. Density dependence of decadic molar extinction co-
2
efficients of CO . Solid lines show the calculated depen-
dence at various wavelengths around 10.6964 µm. Under ρ =
2
−
6
mol/cm3 the density dependence showed different pro-
1
0
εA for the A molecule (here CO2). Figures 1–3 show sev-
eral examples of the density dependence of CO2. Sym-
bols 1–6 in the figures show the decadic molar extinction
coefficient ε of the given wavelength (1/σ). The coeffi-
cients increased with decreasing density in the range be-
files with different wavelengths. Open circles and the error bars
show the experimental values
–1
6
5
: 938.6882[cm ] = 10.653164[µm]
: 10.653166[µm]
−
4
3
low ρ = 10 mol/cm . The calculated data (solid lines in
5
,6
1
3
2
1
0
4 : 10.653169[µm]
the figure) show that the coefficient is not strongly wave-
10
10
10
10
−6
3
length dependent at densities above 10 mol/cm . That
is, the density dependence of the coefficients was almost
the same at different wavelengths, when densities were
ε
−
6
3
above 10 mol/cm . On the contrary, the difference in
the density dependence with different wavelengths became
−6
3
larger when densities were below 10 mol/cm . There-
fore, one could estimate the wavelength by fitting the mea-
sured density dependence of the decadic molar extinction
coefficient to the calculated one.
T=297 K.
1
wavelength 10.653164µm.
00 % CO2.
3
2
1
: 10.653171[µm]
: 10.653173[µm]
: 10.653176[µm]
3
.2 Freon spectra
The decadic molar extinction coefficients of CCl2F2 and
CHClF were measured in the region 7–13 µm at room
2
–6
–4
3
temperature using black body radiation as the light
10
10 ρ [mol/cm ]
source. Since the intensity of the light source was very
Fig. 2. Density dependence of decadic molar extinction coef- low, an average over 400 data points was used. Apart from
ficients of CO . Solid lines show the calculated dependence at this light source, the CO laser was also applied as a light
2
2
various wavelengths around 10.6532 µm
source.
Adding to these experiments the dependencies on
s
wavelength, density, and temperature of the coefficients
of CCl2F2 molecules were calculated by using the follow-
ing relations, obtained from Herzberg 1945(2)
where γ (A) is the standard coefficient at ρs, Ts for the
l
broadening of the molecule A line l by collisions with the
A colliding parameter; βl(A) is the associated temperature
coefficient.
The variables given above beside the wave number σ
can be obtained from the HITRAN database. If the wave
ꢂ
ꢃ
Tv(J, K)hc
kT
Sl(T) ∼ RσlAJ,K gJ,K exp −
(8)
number, i.e. the wavelength for the measurement, is given, where Sl(T) is the integrated intensity of line l at temper-
one can calculate the decadic molar extinction coefficients ature T, R is the IR intensity for each wave number we, σl

4
48
T. Tsuboi et al.: Thermal decomposition of Freons
is the line position, AJ,K is the intensity factor and gJ,K
is the statistical weight for the integers J and K (J ≥ K).
Tv(J, K) is the rotational and vibrational energy and is
described as follows:
6
4
2
10
CHClF2
Meas.
Calc.
ε
Tv(J, K) = G(v) + Fv(J, K)
(9)
ꢀ
G(v) =
(vi + 1/2)wei
(10)
i
Fv(J, K) = (1/2)(Bv + Cv)J(J + 1)
1
0
0
2
+
[Av − (1/2)(Bv + Cv)]K
(11)
ꢀ
A
1
Av = Ae −
α (vi + )
i
2
1
ꢀ
ꢀ
B
Bv = Be −
Cv = Ce −
α (vi + )
i
2
C
1
2
α (vi + )
(12)
Hitran
12
i
where Gv is the vibrational level and Fv(J, K) is the ro-
tational level. Av, Bv, Cv are the effective rotational con-
stants. Ae, Be, Ce are the rotational constants referring to
1
8
10
λ [µm]
the equilibrium position of vibration, and are obtained by Fig. 4. Light extinction spectra (∆λ = 0.16 µm) of CHClF
multiplying the reciprocal equilibrium moment of inertia at 293 K and 0.1 MPa. Mixtures: 0.1, 1.0 and 10% in Ar. Open
2
by h/(8π c). The variables α , α and α iC are correc- circles: measured usingblack body radiation as a li gh t source;
2
A
B
i
i
2
open quadrangles: CO laser light source; solid line: calculated
tion factors for the effective rotational constants, Av, Bv
and Cv. These variables are defined in Herzberg 1945(2).
The expression K in Eq. (11) is for the symmetric top
values; dash–dot line: HITRAN database
2
molecules. CHClF2 and CCl2F2 are asymmetric top
molecules and are unlike the above case. However, we used
this expression for these molecules as slightly asymmetric
molecules.
6
10
CCl2F
2
The wei wavelengths for CHClF2 can be obtained from
the JANAF Thermochemical Data (Chase et al. 1985).
However, some wavelengths had to be changed in order
to fit the calculated absorption spectra to the measured
data. The elemental wave numbers wei were: we1 = 3024;
we2 = 1312; we3 = 920; we4 = 812; we5 = 598; we6 = 417;
we7 = 1320; we8 = 1108; we9 = 340. The wave num-
bers we(i) and the IR intensities R(i) selected for the cal-
culation in the region 7–13 µm were: we(1) = we2 and
R(1) = 85; we(2) = we3 and R(2) = 190; we(3) = we4
and R(3) = 913; we(4) = we7 and R(4) = 34; we(5) = we8
and R(5) = 1080; we(6) = we3 + we6 and R(6) = 1.3;
we(7) = we3 + we9 and R(7) = 2.7; we(8) = we3 + we5
and R(8) = 3.9; we(9) = we4 + we5 and R(9) = 8.5;
we(10) = we4 + we6 and R(10) = 2.8; we(11) = we4 +
we9 and R(11) = 3.0; we(12) = we5 + we6 and R(12) =
Hitran
ε
4
10
Meas.
2
1
0
Calc.
8
10
12
14
λ[µm]
0
.26; we(13) = we5 + we9 and R(13) = 0.55; we(14) =
Fig. 5. Light extinction spectra (∆λ = 0.16 µm) of CCl
93 K and 0.1 MPa. Mixtures: 0.1, 5, 15 and 67% in Ar. Open
circles: measured usingblack body radiation as a li gh t source;
open triangles: CO2 laser light source; solid line: calculated
values; dash–dot line: HITRAN database
2 2
F at
we6 + we9 and R(14) = 0.18; we(15) = we8 + we9 and
2
e
−39
2
e
R(15) = 6.3. Here we used I = 8.3285×10
g cm , I
A
B
−38
2
e
−38
2
=
1.7336×10
g cm and I = 2.4099×10
g cm as
C
A
B
the moments of inertia. The zeros were given for α , α
i
i
C
and α . About 1,000,000 line positions l and integrated in-
i
tensities Sl(T) were calculated from the above-mentioned
energy levels (8)–(11). The decadic molar extinction coeffi- In Fig. 4 one can see how well the decadic molar extinc-
cients between 7 and 13 µm were calculated using Eqs. (2)– tion coefficients can be calculated in the wavelength region
(
7). Because the pressure broadening parameter 7–13 µm using the above-mentioned model and spectral
l
s
γ (CCl2F2) and the temperature coefficient βl(CCl2F2) values.
are unknown, the values 0.1 and 0.7 were roughly esti-
The spectral data of CCl2F2 can be seen in Fig. 5.
mated values selected by considering the values of CO2 in The elemental wave numbers were: we1 = 1099; we2 =
the 10.6 µm region and those of CH4 in the 3.4 µm region. 261.5; we3 = 667.2; we4 = 457.5; we5 = 322; we6 = 918;

T. Tsuboi et al.: Thermal decomposition of Freons
449
1
(,2,5)% CCl F , λ = 10.6114 µm.
2
2
–5
3
i, 1x10 [mol/cm ]
–
5
Meas.
r, 2x10
–
5
5
4
3
ε
i, 2x10
r, 5x10
10
10
10
–
5
–5
i, 3x10
r, 8x10
–5
Calc.
–
5
i, 2x10 , 2%
–
5
r, 5x10 , 2%
–
5
5
x10 , 5%
1
2
3
4
–1
3
Fig. 6. Calculated light extinction (solid lines) and mea-
sured emission spectra (arbitrary unit, dotted lines and
1
0 /T [K ]
open circles) of CCl
2
F
2
around 10 µm. T
2
= 1000 K, ρ
2
=
Fig. 7. Temperature dependence of decadic molar extinction
coefficients of CCl in Ar. Mixtures: 1, 2 and 5% CCl in
Ar. Quadrangles, triangles and circles are the measured values.
−
6
3
5
.0×10 mol/cm , 1.0 % CCl in Ar
2
F
2
2
F
2
2 2
F
Solid lines are the calculated values in this work usingEqs. (2)–
3
(
11). The total densities are given in mol/cm and ‘i’ and ‘r’
we7 = 436; we8 = 1162; we9 = 446. The wave num-
bers we(i) and the IR intensities R(i) used for the cal-
culations in the region 7–13 µm were: we(1) = we1 and
show the incident and reflected shock waves
R(1) = 352; we(2) = we6 and R(2) = 450; we(3) = we8 The coefficients increased with increasing temperature up
and R(3) = 248; we(4) = we1 + we2 and R(4) = 2.0; to about 1000 K. Above 1000 K the coefficients decreased
we(5) = we3 + we4 and R(5) = 1.4; we(6) = we4 + we5 rapidly with increasing temperature. One can see in this
and R(6) = 0.2; we(7) = we5 + we6 and R(7) = 2.5; figure how well the calculated and the measured values
we(8) = we6 + we7 and R(8) = 1.1; we(9) = we7 + we9 agree with each other, though the experimental error was
and R(9) = 1.8; we(10) = we5 + we9 and R(10) = 0.1; large. This indicates that other molar extinction coeffi-
we(11) = we1 + we5 and R(11) = 1.0; we(12) = we4 + we9 cients should be explained using a similar profile, because
and R(12) = 36; we(13) = we4 + we7 and R(13) = 36; the temperature and the density dependencies of the
we(14) = we6 + we9 and R(14) = 1.1; we(15) = we4 + we6 decadic molar extinction coefficients can be described by
and R(15) = 1.1; we(16) = we2+we8 and R(16) = 1.1. The Doppler and pressure broadening. We could not perform
remaining possible wavelengths, i.e. we(17) = we2 + we6, experiments at various wavelengths due to the difficulty
we(18) = we2 + we3, we(19) = we3 + we7, we(20) = we3 + in determining the wavelength of the CO laser system.
2
we9, were regarded as belonging to the strong spectra of
Figures 8 and 9 show the measured wavelength de-
pendence of the decadic molar extinction coefficients of
e
we8, we6, we1 and we1, respectively. Here, we used I
A
−
38
2
e
−38
2
e
=
=
2.1216×10
g cm , I = 3.1929×10
g cm and I
CCl F at 293 K and 0.1 MPa, and the measured temper-
B
C
3
−38
2
3.8148×10
g cm as the moments of inertia. The ature dependence of the coefficients at 10.6531 µm of the
A
B
C
zeros were given for α , α and αi also in this case. CO laser.
i
i
2
The wavelength dependence of the decadic molar extinc-
tion coefficient of CCl2F2 agreed well with the measured
value. The emission intensity at high temperature was
measured behind an incident shock wave with arbitrary
units. This emission spectra of CCl2F2 at about 1000 K
3
.3 Thermal decomposition of CCl2F2
and the decadic molar extinction coefficients are shown in Figure 10 shows the time histories of CCl2F2 at various
Fig. 6. wavelengths. The emission due to the C–F bond was ob-
From the emission spectra one can expect that the served for wavelengths 8.5–9.8 µm. The emission did not
emission of CCl2F2 due to the C–Cl bond can be measured decrease intensely. One example can be seen in the emis-
in the region 10.5–11.5 µm.
sion signal at 9.5 µm in the figure. The emission around
Figure 7 shows the temperature dependence of the 10.4 µm increased just behind the reflected shock wave
decadic molar extinction coefficients of CCl2F2 at and soon reached equilibrium. The emission for the wave-
1
0.6113 µm at various densities. The solid line shows the lengths 10.7–12 µm decreased exponentially, like the ther-
−5
3
calculated extinction coefficients at ρ = 5×10 mol/cm . mal decomposition representative signal behind the re-

4
50
T. Tsuboi et al.: Thermal decomposition of Freons
2
00
00
0
Reflected shock
CCl3F
6
4
2
Measured
Hitran
10
10
10
9
.5 µm
ε
1
1
1 µm
1
0.4 µm
Incident shock
6
8
10
12
0
500
1000
t [µs]
λ [µm]
Fig. 10. Time histories of IR emission of 1.0% CCl
0.4 and 11.0 µm (∆λ = 0.4 µm), T = 1010–1020 K, ρ
= 1960–1980 K, ρ = (1.14–
2 2
F at 9.5,
3
Fig. 8. Light extinction spectra (∆λ = 0.04 µm) of CCl F
1
2
2
=
at 293 K and 0.1 MPa. Mixtures: 0.05, 0.1, 1.0% in Ar. Open
circles: measured usingblack body radiation as a li gh t source.
Dash–dot lines: HITRAN database
−
6
3
(5.05–5.06) × 10 mol/cm ; T
5
5
−
5
3
1
.15)×10 mol/cm . The dotted lines are the calculated emis-
sion profiles at 9.5 and 11 µm usingthe reaction mechanism in
Tables 1 and 2
λ = 10.6532µm
5
1
1
1
1
0
0
0
0
5
% CCl F in Ar
3
–
5
3
CCl2F2
ρ2=1.5x10 [mol/cm ]
ρ2=4.0x10
ρ2=6.0x10
ε
–5
5
–5
–
5
10
ρ5 = 1.5x10
–5
ρ5 = 4.0x10
4
3
2
–5
ρ5 = 9.0x10
Meas.
–5
3
ρ5 = 1.5x10 [mol/cm ]
: 0.02 %, : 0.05 %
4
10
–5
3
ρ5 = 4.0x10 [mol/cm ]
0.02 %, : 0.05 %
:
–5
3
ρ5 = 9.0x10 [mol/cm ]
:
0.02 %,
5
: 0.05 %
1
2
3
3
–1
1
0 /T [K ]
2
6
7
8
Fig. 9. Temperature dependence of decadic molar extinction
coefficient of CCl F at various densities. The light source is a
CO laser. The solid line shows the average of the measured
value
4
–1
1
0 /T [K ]
5
3
2
Fig. 11. Thermal decomposition of CCl
emission for 11 µm
2 2
F , the decay rate of
flected shock wave. The emission signal for 11.0 µm is also
shown in Fig. 10.
decomposition. The thermal decomposition of CCl2F2 was
observed in the mixtures less than 0.05% CCl2F2 in argon,
because it was expected that the effect of secondary reac-
tions could be reduced under 0.05% dilution. Figure 11
shows the rate constant kA (−d ln I/dt at t = 0) of the
decay emission rate induced by the reaction,
Using the above-mentioned comparison of the spec-
tra, the thermal decomposition of CCl2F2 was followed
for a wavelength of 11 µm. The mixtures above 0.2% di-
lution seemed to have somewhat complex reactions. Some
CCl2F2 must have decayed due to bimolecular reactions
induced by radicals which were produced by the thermal
CCl2F2 + (Ar) → CClF2 + Cl + (Ar) [1]

T. Tsuboi et al.: Thermal decomposition of Freons
451
Reflected shock
5
CHClF2
1
1
1
1
0
0
0
0
9.1 µm
–5
2
00
00
0
ρ 5 = 5x10
ρ 5 = 1x10
–4
4
3
2
12.4 µm
1
–5
3
ρ 5 = 2x10 [mol/cm ]
.2 % 0.05 %
1
0.6 µm
0
–5
3
ρ 5 = 5x10 [mol/cm ]
0
.2 %
0.05 %
–4
3
ρ 5 = 1x10 [mol/cm ]
0
.2 %
0.05 %
Incident shock
0
500
1000
7
8
9
t [µs]
4
–1
1
0 /T [K ]
5
Fig. 12. Time histories of IR emission of 1.0% CHClF
2
at
=
9
(
1
.1, 10.6 and 12.4 µm (∆λ = 0.4 µm), T
2
= 850–863 K, ρ
2
2
Fig. 13. Thermal decomposition of CHClF , the decay rate of
= 1570–1600 K, ρ = (1.68– emission for 12.4 µm
5
−
6
3
7.77–7.82) × 10 mol/cm ; T
5
.70) × 10 mol/cm3
−
5
for total densities of ρ = 1.5, 4.0 and 9.0 × 10−5 mol/cm3. not agree with the rate constant of the unimolecular reac-
tion, if one did not highly dilute the mixture. One had to
dilute the test gas to less than 0.05% in argon in order to
decrease the effect of the bimolecular reactions induced by
the intermediate molecules, which were produced during
3
.4 Thermal decomposition of CHClF2
The emission due to the C–F bond was observed at wave- the reactions.
lengths between 8.5 and 9.5 µm. The emission did not
In the case of CCl F the decay signal did not go to
2
2
decrease notably. They were similar to the emission of zero. It seemed to go to a value of a little less than half of
CCl2F2. One example can be seen in the emission signal at the initial value (at t = 0). Then the signal decreased grad-
9
1
.1 µm in Fig. 12. The emission signals of CHClF2 around ually. This decay phenomena depended on the experimen-
2.4 µm were also similar to those of CCl2F2. Figure 12 tal conditions (the mixture temperature behind the shock
shows the signals of CHClF2 decomposition. The thermal wave and the initial CCl F concentration in Ar). With in-
2
2
decomposition of CHClF2 was observed at 12.4 or 12.6 µm. creasing dilution rate the emission signal approached zero.
Under 0.2% dilution it was expected that the decay rate One possible reason for the change in these profiles was
of the emission depended only on the decomposition
that some reaction products had CCl combination and
the emission intensities were of the same order as those of
CCl2F2, and probably CClF2 etc. The second possibility
was that the log Kf value of CCl2F2 was large, or even
larger than the related reaction products, because it is a
very stable species. The variable Kf means the thermo-
dynamic equilibrium constant of formation. The molecule
could not decompose much and was already in a chemical
(probably quasi-chemical) equilibrium in our experiment,
after some of the CCl2F2 had decomposed.
CHClF2 + (Ar) → CF2 + HCl + (Ar) [2]
Figure 13 shows the rate constant obtained from the decay
rate of the emission at 12.4 µm.
4
Discussion
The FTIR monochromator is often used in order to de-
If CClF2 has the same order as the IR intensity of
cide the precise wavelength in the infrared region. How- CCl2F2 at 11 µm, the pure rate constants of CCl2F2 de-
ever, we used a normal monochromator to measure the composition must be two times larger than the measured
molecular light extinction, because of its low cost, and decay rate. The CClF2 has strong IR radiation in the
−
1
obtained the spectra by comparing the measured density 761 cm
region due to CCl stretching. Since the wave-
−1
dependence of decadic molar extinction coefficients with length is far from 11 µm(910 cm ), the IR intensity of
the calculated dependence from Eqs. (2)–(7), using the el- CClF2 must be small. Therefore, our decomposition rates
ementary HITRAN data. The IR emission due to C–Cl include an uncertainty factor of 1. Considering these in-
stretching could be observed at wavelengths in the range fluences we made small computer simulations for mixtures
1
1–12 µm for CCl2F2, CHClF2 and CCl3F. The rate did of 0.04% and 1% CCl2F2 using the reactions in Tables 1

4
52
T. Tsuboi et al.: Thermal decomposition of Freons
a
3
Table 1. Reactions and rate constants for CCl
2
F
2
decomposition. k = 10 exp(−E
A
/RT). Units: mol, cm , kJ. A reference
∗
in parantheses indicates that an estimation was used. shows the estimation from the decay signal at 11 and 9.5 µm emission
values in our experiments
No.
1
Reactions
Rate constants
Forward Backward
Ref.
a
E
A
a
E
A
CCl
2
F
2
+ (Ar)
= CClF
2
+ Cl + (Ar)
(
ρ = 5 × 10− mol/cm3)
6
10.53 212 8.73 −114
11.04 217 9.24 −109
11.56 229 9.76 −97
11.72 229 9.92 −97
∗
This work
−
5
(
ρ = 1.5 × 10 mol/cm3)
5
(
(
ρ = 4 × 10− mol/cm3)
ρ = 9 × 10− mol/cm3)
5
3
4
5
6
7
8
9
CCl
2
F
2
+ Cl
= CClF
= CF
2
+ Cl
+ Cl
+ Cl
+ F
2
2
12.70 138 11.66
13.30 63 12.25
12.70 84 13.83
13.00 167 13.01
13.70 105 13.07
13.70 126 13.72
51 (Sekiguchi et al. 1992)
CClF
2
+ Cl
2
100 (Sekiguchi et al. 1992)
∗
CCl
CCl
CCl
CCl
CCl
2
2
2
2
2
F
2
F
2
F
2
F
2
F
2
+ CClF
+ CClF
2
2
= C
= C
2
2
Cl
Cl
2
F
4
3
121
∗
3
F
26
∗
+ CF
+ CF
+ CF
+ Ar
2
2
2
= CCl
= CClF
= C Cl
2
F
+ CF3
−3
2
∗
2
+ CClF
2
∗
2
2
F
4
11.00
0 13.95
239
1
0 CClF
1 C
2
= CF
2
+ Cl + Ar
+ CF + (Ar)
14.70 209 12.88
7 (Modica and Sillers 1968)
1
2
F
4
+ (Ar)
= CF
2
2
ρ = 6 × 10− mol/cm3)
7
10.64 202 6.46 −69 Schug(1976)
12.04 226 7.86 −45
12.63 235 8.45 −36
(
ρ = 3 × 10 mol/cm3)
ρ = 9 × 10− mol/cm3)
ρ = 5 × 10− mol/cm3)
ρ = 1 × 10− mol/cm3)
ρ = 4 × 10− mol/cm3)
−
6
(
6
(
5
(
13.67 253 9.49 −18
4
(
(
14.20 262 10.01
−9
2
4
14.79 273 10.61
12 Cl
2
+ Ar
= Cl
+ Cl + Ar
13.37 196 12.60 −43 (Baulch et al. 1981)
and 2. The effective digits of the activation energies EA and used log(Kf ) = 11.635 at 1000 K and 6.770 at 2000 K.
were mostly 1 or 2, though the figures in Tables 1 and 2 The rate constants (10) are given in Table 1. The log Kf
are given up to 3 figures. The rate constants were esti- 18.938 (1000 K), 6.234 (2000 K) for CCl2F2, and 12.988
mated in this work by considering the data of Sekiguchi (1000 K), 5.348 (2000 K) for CClF2 were used in this case.
et al. (1992). The effective digits for the pre-exponential
The emission intensities were calculated using decadic
a
factors 10 were also 1 or 2. The backward reaction rate molar extinction coefficients, because we considered that
constants were calculated by using the log Kf , tabulated the emission intensities at our wavelengths must be pro-
in the JANAF table. Several log Kf values which were portional to the absorption intensities. The decadic mo-
not in the JANAF table, were calculated in this work by lar extinction coefficients were roughly estimated and the
3
using wave numbers, moments of inertia, formation en- following values were used: at 11 µm, 5×10 (CCl2F2),
3
3
3
3
3
3
thalpies and other thermochemical values. The profile at 5×10 (CClF2), 5×10 (CCl2F), 5×10 (C2Cl3F3), 5×10
3
3
1
1 µm was the emission mainly due to CCl2F2. The de- (C2Cl4F2), 5×10
(CCl3), 1×10
(C2Cl6), 5×10
3
3
composition rate of the 0.05% mixture was the reaction (C2Cl3F2), 5×10 (C2Cl2F2), 5×10 (C2Cl2F3), 5×10
3
3
[
1] (called R1 in this work). However, though the emission (C2Cl4), 5×10 (C2Cl2F4), 5×10 (CClF3), 5×10
3 3
profiles of the 1% mixture were still due to CCl2F2, the (CCl3F), 5×10 (C2ClF4), 5×10 (C2Cl4F) cm /mol, and
3
3
decomposition rate was not due to the reaction [1] (R1), at 9.5 µm, 5×10 (C2F3), 5×10 (CCl2F2), 4×10 (CClF2),
3
3
3
3
3
3
but it was much smaller than the rate of R1. The CCl2F2 3.5×10 (CF2), 5×10 (CCl2F), 4×10 (CF3), 5×10
3
3
went to quasi-equilibrium with Cl and CClF2 very soon, (C2Cl3F3), 5×10 (C2Cl4F2), 5×10 (C2F6), 5×10
3
3
when one used the log Kf of CF2 from the JANAF ta- (C2Cl3F2), 5×10 (C2Cl2F2), 5×10 (C2Cl2F3), 1.5×10
3
3
ble, or the high thermal decomposition rate constant for (C2F4), 5×10 (C2Cl2F4), 5×10 (CF4), 5×10 (C2ClF4),
3
3
3
CClF2. In this work we changed the log Kf of CF2 a bit 5×10 (CClF3), 5×10 (CCl3F), 5×10 (C2Cl4F), respec-

T. Tsuboi et al.: Thermal decomposition of Freons
453
a
3
Table 2. Additional reactions and rate constants. k = 10 exp(−E
A
/RT). Units: mol, cm , kJ. A reference in parantheses
∗
indicates that an estimation was used. shows the estimation in this work
No.
Reactions
Rate constants
Ref.
Forward Backward
a
E
A
a
E
A
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
3 F
4 CF
5 CCl
2
+ Ar
+ Ar
+ Cl
+ Cl
+ F
= F
= CF
+ F +Ar
+ F +Ar
+ ClF
13.32 141 12.09 −12 Westbrook 1982
2
17.57 443 16.25 −67 Modica and Sillers 1968
∗
2
F
2
= CCl
= CF
= CF
= CF
= ClF
= F
= ClF
= CCl
= CClF
= C
= C
= C
= C
= C
= C
= C
= C
= C
= C
= C
= C
2
F
13.00 259 10.84
13.70 267 13.09
41
5
∗
∗
∗
6 CF
7 CF
8 CF
9 F
0 ClF
1 Cl
2 CCl
3 CCl
4 CCl
5 CCl
6 CCl
7 CCl
8 C
9 C
0 C
1 C
2 C
3 C
4 C
5 C
6 C
7 C
8 C
9 C
0 C
1 C
2 C
3 C
4 C
5 C
6 C
7 C
8 C
9 C
0 C
2
+ ClF
2
2
+ F
2
13.00 351 12.82 −4
12.00 150 12.92 −2
+ CF
2
+ CF
+ Cl
3
+ Cl
+ Ar
2
13.00 41 12.96
14.47 247 13.76
50 (Baulch et al. 1981)
0 (Baulch et al. 1981)
+ Cl +Ar
+ F
+ F
+ F
+ F
2
13.00 41 12.58 135 (Baulch et al. 1981)
∗
2
2
2
2
2
2
F
F
F
F
F
F
2
2
2
2
F
+ F
2
13.00 334 11.26
13.00 125 11.92
23
47
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗
2
+ ClF
+ F
+ Cl
+ F
+ CF
3
2
2
2
2
2
2
2
2
2
2
2
2
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
2
F
F
F
F
F
F
F
F
F
F
F
F
4
3
2
3
2
3
3
2
2
3
2
2
13.00 62 13.69 −55
+ CCl
+ CCl
+ CCl
+ F
2
2
2
F
F
F
2
3
4
3
3
2
2
3
3
2
2
2
12.70 83 13.80
12.70 167 14.85
13.00 251 11.94
81
23
31
2
2
2
+ ClF
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
ClF
ClF
F
F
F
Cl
Cl
Cl
3
3
3
3
3
3
3
2
2
2
2
2
2
2
2
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
3
3
3
3
3
3
2
3
2
2
4
4
4
4
4
+ F
2
13.00 209 13.22 −8
+ F
+ ClF
+ CClF
2
12.00 83 10.88
13.00 125 12.97
80
76
+ CF
+ CF
+ Cl
+ Cl
+ Ar
+ Ar
+ Cl
+ F
2
2
+ CF
+ ClF
+ Cl
3
13.00 125 14.33 111
13.00 125 12.80
13.00 41 11.92
1
30
2
+ Cl +Ar
+ F +Ar
13.00 167 11.93 −18
13.00 293 12.84 −14
= CCl
= CCl
3
+ CF
+ CF
2
13.00 104 10.67
13.00 41 10.89
13.00 125 11.35
3
20
50
2
3
+ Cl
+ Cl
+ F
= C
= C
= C
= C
= C
= C
= C
= C
= CClF
= CF
= C
= C
2
2
2
2
2
2
2
2
ClF
Cl
ClF
Cl
ClF
4
+ Cl
2
+ ClF
+ ClF
2
F
3
13.00 167 10.77 −14
4
13.00 125 11.31
13.00 313 11.20
13.00 125 12.41
58
38
13
+ F
2
F
3
+ F
+ CClF
+ Cl +Ar
+ Cl
2
+ CF
+ Ar
+ Cl
+ Cl
+ Cl
+ F
2
4
2
4
F
F
4
13.00 167 13.03 −13
13.00 41 13.79 99
12.00 41 11.96 222
13.00 83 10.64 14
13.43 41 11.50 127
13.00 125 11.78 135
4
4
2
4
4
4
ClF
4
2
+ CF
+ CF
2
2
3
4
4
6
F
F
2
+ Cl
+ Cl
2
Cl
Cl
3
F
F
2
+ Cl
2
2
2
4
+ ClF
+ CCl
13.00 125 11.55
33
29
+ (Ar)
+ Ar
+ Cl
+ Cl
+ (Ar)
+ Cl
+ Cl
+ F
= CCl
3
3
+(Ar) 12.00 293 8.63
1 CF
2 CF
3 CF
4
4
3
= CF
= CF
= CF
= CF
= CF
= CClF
= CClF
= CCl
3
+ F +Ar
+ ClF
+ ClF
17.08 544 13.90
13.00 293 10.54
13.00 83 11.47 −25
12 (Baulch et al. 1981)
∗
3
2
3
3
8
∗
∗
∗
∗
∗
∗
4 C
2
F
6
+ CF
+ Cl
+ ClF
+ F
+ Cl
3
+(Ar) 17.04 395 13.27
17
5 CClF
6 CClF
7 CClF
3
2
13.00 104 11.72 −2
13.00 251 11.24 −2
13.00 334 11.66 −12
3
3
2
2
2
8 CCl
3
F
+ Cl
2
F
2
13.00 96 11.19
32

4
54
T. Tsuboi et al.: Thermal decomposition of Freons
tively. Practically there was a large influence from CCl2F2, portant. That is, the values XRi of reactions R53 and R55
some influence from CClF2 at 11 µm and some from CClF2 were of the order of only 1/100 of R1 and they were slightly
and CF2 at 9.5 µm. The emission of the other molecules larger than the values of XRi for other reactions. The val-
was small. Therefore, we could obtain the emission pro- ues XRi for other reactions were less than 1/1000 of R1.
file, which is indicated by dotted lines in Fig. 10. This fig- Therefore, we knew that there was no influence by the re-
ure includes also the emission profiles at 9.5 µm, mainly actions of Table 2 on the decay rates at 11 (and 9.5) µm in
due to the vibration of CF stretching. The decay rate at our conditions. If there had been some influence by the re-
1
1 µm was smaller than the thermal decomposition and it actions other than those in Table 1 on the decay rates, the
depended on the reaction [R7,R8,R9] and the log Kf of rate constants in Table 2 would have been quite different,
CF2, as mentioned above. The decay rate of CCl2F2 at or there would have been other reactions.
the first step could mainly describe the reactions in Ta-
The obtained rate constants for unimolecular reaction
ble 1. The rate depended also on the log Kf of CCl2F2, (1) were:
CClF2, CF2 and Cl, because the molecule CCl2F2 and the
related species were chemically very stable.
ꢂ
ꢃ
k1 = 1.1 × 10 exp −2
17 kJ/mol
[1/s]
1
1
RT
We checked the amount of molecules which passed in
ms through the reactions in Table 1. The amount which
for ρ = 1.5 × 10− [mol/cm ]
5
3
1
passed through reaction R1 was about 1.5 times larger
than the amount of the initial CCl2F2 concentration. Ap-
parently this was curious; however, 0.4–0.5 of the CCl2F2
was produced from the reaction R8. Anyway R1 is the
first-step reaction sequence; 0.2–0.4 of the initial CCl2F2
passed through reaction R4 and 0.05–0.3 through R10.
By increasing the dilution rate, the ratio of the amount
of R10 to that of R4 increased. The influence of R3, R5
and R6 was about 1/100, or less than that. The amount
passing through R7 was 0.01–0.05 of the initial CCl2F2.
The amount passing through R11 was the order of 1/100
and that through R12 was −0.01 to 0.05. When the rate
constant of R4 was made 10 times larger, or 1/10, the
separation of the emission profile at 11 µm from the orig-
inal profile was about 30% in the latter part of the reac-
tion. However, the decomposition rate in the former part
of the reaction did not differ from the original profile.
Since the sensitivity was different, as mentioned above,
we could obtain the rate constant for R4 from the emis-
sion at 11 µm and also for R7, R8 from 9.5 µm. Other re-
actions were estimates. We could not decide the reaction
products of CCl2F2 from the heat of the reaction, since
the heat of the reaction CCl2F2 + (Ar) → CF2 + Cl2 +
ꢂ
ꢃ
229 kJ/mol
RT
1
1
k1 = 3.6 × 10 exp −
[1/s]
for ρ = 4.0 × 10− [mol/cm ]
5
3
ꢂ
ꢃ
k1 = 5.3 × 10 exp −2
29 kJ/mol
RT
[1/s]
1
1
for ρ = 9.0 × 10− [mol/cm ]
5
3
The decay signal for CHClF2 went down to almost
zero. The reaction products of the reaction [2] were un-
known. However, we suggest that they were either CF2
HCl or CHF2 + Cl. In this case they must have had
no IR emission around 11 µm; therefore the decay rate at
2.4 µm must have coincided with the rate constant of the
thermal decomposition. Schug measured the time history
of CF2 absorption profiles of CHClF2 thermal decompo-
sition (Schug 1976). Our values agreed well with those of
Schug. We could, therefore, select reaction [2] and the rate
constants were:
+
1
ꢂ
ꢃ
186 kJ/mol
RT
1
0
k2 = 8.8 × 10 exp −
[1/s]
(
Ar) (∆H = −322 kJ/mol) was almost the same as the
reaction CCl2F2 + (Ar) → CClF2 + Cl +(Ar) (∆H =
310 kJ/mol). When the products were CClF2 + Cl, the
for ρ = 2.0 × 10− [mol/cm ]
5
3
ꢂ
ꢃ
−
k2 = 8.8 × 10 exp −1
83 kJ/mol
RT
[1/s]
1
0
decay profiles at 11 µm could fit better to the simulation.
Namely, the rate of decay profile was the same as the
thermal decomposition rate and went quickly to the equi-
librium state in the 1% mixture. We could not identify the
species at 10.4 µm in Fig. 10. The emission increased up to
about 100 µs and then remained almost constant. How-
ever, we could calculate the emission profiles at 10.4 µm
in Fig. 10, when we assumed the following decadic molar
for ρ = 5.0 × 10− [mol/cm ]
5
3
ꢂ
ꢃ
k2 = 5.3 × 10 exp −2
00 kJ/mol
RT
[1/s]
11
for ρ = 1.0 × 10−4 [mol/cm3]
4
2
3
extinction coefficients: 1×10 cm /mol (CCl2F2), 3×10
Since the density region was not large enough in this work,
4
(
CClF2), 1×10 (C2Cl3F3, C2Cl4F2, C2Cl3F2, C2Cl2F2, one could not calculate the precise rate constants for the
3
C2Cl2F3, C2Cl2F4), and 1×10 (CClF3).
∞ 0
high-pressure limit k and the low-pressure limit k . We
Table 2 shows the reactions used in our calculation in calculated them by rough extrapolation (Troe 1974) for
order to know the number of molecules XRi which passed R1:
ꢂ
ꢃ ꢁ cm3
ꢄ
through each reaction. Most rate constants for the reac-
tions were unknown. Since we estimated using hydrocar-
bon reactions, the rate constants were unreliable. How-
ever, we confirmed from XRi that they were not so im-
1
90 ± 20kJ/mol
1
6
k
1
,0
= 2.1 × 10 exp −
RT
mols
ꢂ
ꢃ
240 ± 20kJ/mol
1
2
k1,∞ = 2.4 × 10 exp −
[1/s]
RT

T. Tsuboi et al.: Thermal decomposition of Freons
455
5
4
3
2
–4
–5
10
10
10
10
10
1
400 K
5
4
3
2
1
1
1
0
0
0
10
ρ
1
%CCl F in Ar
2 2
1
300 K
200 K
λ=10.6 µm (∆λ=0.12mm)
1
0% O2
1
% O2
5
% O2
10
–6
–5
–4
–3
5
6
7
10
10
10
10
3
4
–1
ρ [mol/cm ]
10 /T [K ]
5
Fig. 14. Density dependence of thermal decomposition of Fig. 15. Thermal decomposition of CCl
2 2
F , the influence of
CHClF
2
. Solid lines are data in this work and dashed lines oxygen on the decay rate at 10.6 µm. The upper side symbols
are those of Schug(1976). Circles are experimental data in
this work and quadrangles are from Schug
5
show the density ρ behind the reflected shock
our values. We could compare our data with Schug’s data
and confirm the certainty of our data. Since the exper-
iment for CCl2F2 was not performed in a wide pressure
range, the error of the density dependence must have been
large. However, the pressure range in this work was not
far from the high pressure limit. Therefore, our proposed
values k1,∞ should not have a large error. On the contrary,
the second-order rate constants k1,0 must have had large
extrapolation errors, because we could not perform the
experiment in the low-pressure range. One can compare
them with the result for CH3Cl in Kondo et al. (1980).
The effect of additives on CCl F was observed by mix-
here we used SK = 7, 8 and 8, and BK(= E0/kT) = 25,
2
2 and 19, for T = 1400, 1600 and 1800 K. The variable
SK is Seff + 1 = Uvib/RT + 1, and E0 is the threshold
energy at 0 K of the reaction, described in Troe (1974).
The value 294 kJ/mol was used for E0 as an approximate
value. The wave numbers used to calculate the partition
function were used for the values given in Sect. 3.2 for R2.
ꢂ
ꢃ ꢁ cm3
ꢄ
1
50 ± 20 kJ/mol
1
5
k2,0 = 2.5 × 10 exp −
RT
00 ± 20 kJ/mol
RT
mols
ꢂ
ꢃ
k2,∞ = 1.1 × 11 exp −2
[1/s]
1
1
2
2
ing oxygen and hydrogen. Figure 15 shows the influence of
the oxygen concentration on the decay rate of CCl2F2 in
here we used Sk = 6, 6 and 6, and Bk = 29, 27 and 25, for the mixture with 1% CCl2F2 in argon. Figure 16 shows the
T = 1200, 1300 and 1400 K. More detailed calculation for influence of the hydrogen concentration on the decay rate
the high- and the low-pressure limit was not performed in of CCl2F2 in the mixture with 1% CCl2F2 in argon. We
this work, because of the insufficient data.
did not observe the influence of the total density, but we
Figure 14 shows the pressure dependencies of the rate observed only the influence of oxygen and hydrogen. It was
constant k2 at the temperatures 1200, 1300 and 1400 K. difficult to perform experiments by sustaining the total
The circles and the solid lines show data obtained in this density constant behind the reflected shock wave. There-
study, while the data of Schug (1976) are shown with fore, the densities ρ5 together with the rate constants are
quadrangles and the dashed lines. Schug determined the shown in Figs. 15 and 16. As these figures show, the hy-
rate constant by measuring the CF2 absorption at 250 nm. drogen influenced strongly the decay of CCl2F2, while the
In this study the rate constants were obtained from the effect of oxygen was not large. This is because the reac-
IR decay at 12.4 µm due to CHClF2. Both values agreed tions of CCl2F2 with hydrogen atoms or with hydrogen-
relatively well with each other, as shown with open circles containing radicals are active. This comparison was done
and quadrangles in Fig. 14. This shows that the reaction for the 1% CCl2F2 in argon in order to obtain the profiles
products of the thermal decomposition of CHClF2 are CF2 with high S/N ratios. Therefore, this mixture must not
+
HCl. Schug used rate constants 1.5 to 2 times larger at have involved only a unimolecular reaction but also com-
the high pressure limit than our values in the temperature plex bimolecular reactions, because the 1% test gas was
range 1200–1400 K. On the contrary his rate constants at already too diluted. More precise work on these bimolec-
the low-pressure limit were about 1.5 times smaller than ular reactions is needed.

4
56
T. Tsuboi et al.: Thermal decomposition of Freons
–
4
HerzbergG (1945)(1) Molecular spectra and molecular
structure, II infrared and raman spectra of polyatomic
molecules. Van Nostrand Reinhold, New York, pp. 306–320
HerzbergG (1945)(2) Molecular spectra and molecular
structure, II infrared and raman spectra of polyatomic
molecules. Van Nostrand Reinhold, New York, pp. 421, 460,
10
1
%CCl F in Ar
2 2
5
4
3
2
–5
10
10
10
10
λ=10.6 µm (∆λ=0.12mm) 10
ρ
4
81
Kondo O, Saito K, Murakami I (1980) The thermal unimolec-
ular decomposition of methyl chloride behind shock waves.
Bull. Chem. Soc. Jpn. 53: 2133–2140
Lacher JR, Hummel LE, Bohmfalk EF, Park JD (1950) The
near ultraviolet absorption spectra of some fluorinated
derivatives of methane and ethylene. J. Am. Chem. Soc.
0% H2
1% H2
5% H2
7
2: 5486–5489
L ´e vine AK (1968) Lasers, Vol. 2. Marcel Dekker, New York,
p. 159
Modica AP, Sillers SJ (1968) Experimental and theoretical
kinetics of high-temperature fluorocarbon chemistry. J.
Chem. Phys. 48: 3283–3289
Richard Zobel C, Duncan BF (1955) Vacuum ultraviolet ab-
sorption spectra of some halogen derivatives of methane.
Correlation of the spectra. J. Am. Chem. Soc. 77: 2611–
5
6
7
4
–1
1
0 /T [K ]
2
615
Fig. 16. Thermal decomposition of CCl
hydrogen on the decay rate at 10.6 µm
2 2
F , the influence of
Rothman L, Gamache R, Goldman L, Brown L, Toth R, Pick-
ett H, Poynyer R, Flaud J, Camy-Peyret C, Narbe A, Hus-
son N, Rinsland C, Smith M (1992, 1996) HITRAN 92,96
CD-ROM
SchugKP (1976) Untersuchungdes Reaktionsverhaltens von
Fluorkohlenstoffradikalen bei hohen Temperaturen. Zur
Aktivierunsenergie von drei-zentren Eliminierungen. Dis-
sertation 1976, G ¨o ttingen
5
Conclusion
The spectra of CCl2F2, CHClF2 and CCl3F were observed
in the wavelength range 7–13 µm. The pressure-dependent
unimolecular rate constants for thermal decomposition of
CCl2F2 and CHClF2 were measured in the temperature
Sekiguchi H, Kanzawa A, Honda T (1992) Freon decomposition
usingthermal plasma (in Japanese). J. High-Temp. Soc. 18:
216
−5
−4
range 1200–1900 K and in the density range 10 –10
3
mol/cm . The reactions and the rate constants for the Troe J (1974) Fall-off curve of unimolecular reactions. Ber.
CCl2F2 decay were obtained at an early stage of the reac-
tion.
Bunsen. Gesell. Physik. Chem. 78: 478–488
Tsuboi T, Kimura T, Yamamoto K (1975) Temperature mea-
surement of a propa ga tingflame usingUV-absorption of
2
CO . Jpn. J. Appl. Phys. 14: 1483–1488
Tsuboi T, Egawa H, Kobayashi H, Hirose T, Kamei M, Ter-
anaka M (1981) Temperature measurement of detonation
Acknowledgements. The authors thank Prof. K. Sakakibara for
his helpful suggestion to use the Gaussian program, Prof. N.
Arimitsu for her helpful calculation of the CO absorption,
2
and Mr. Y. Torikai, Mr. K. Horie and Mr. T. Ishiyama for
their experimental work.
2
usingUV-absorption of O . Proc. 13th Int. Symp. Shock
Tubes and Waves. New York State University Press, New
York, pp. 141–149
Tsuboi T, Ishii K, Mochizuki Y (1996) Velocity, pressure and
temperature measurement in low-pressure supersonic flow.
Jpn. J. Appl. Phys. 35: 1917–1921
Tsuboi T, Arimitsu N, Du P, Hartmann JM (1990) Tempera-
ture, density, and perturber dependencies of absorption of
the 3.39 µm He–Ne laser by methane. Jpn. J. Appl. Phys.
29: 2147–2151
References
Baulch DL, Duxbury J, Grant SJ, Montague DC (1981) Eval-
uated kinetic data for high temperature reactions. J. Phys.
Chem. Ref. Data 10, Suppl. 1
Chase Jr. MW, Davies CA, DowneyJr. JR, Frurip DJ, McDon- Tsuboi T, Inomata K, Tsunoda Y, Isobe A, Nagaya K (1985)
ald RA, Syverud AN (1985) JANAF thermochemical ta-
bles, 3 Ed., Part I, American Chemical Society and Amer-
ican Institute of Physics, Washington, DC and New York
Chou C, Milstein RJ, Smith WS, Vera Ruiz H, Molina J Mario,
Rowland FS (1978) Stratospheric photodissociation of sev-
eral saturated perhalo chlorofluorocarbon compounds in
current technological use (fluorocarbons-13, -113, -114, and
Light absorption by hydrocarbon molecules at 3.392 µm of
He–Ne laser. Jpn. J. Appl. Phys. 24: 8–13
Westbrook CK (1982) Inhibition of hydrocarbon oxidation
in laminar flames and detonations by halogenated com-
pounds. 19th Symp. (Int.) on Combustion. The Combus-
tion Institute, pp. 127–141
-115). J. Phys. Chem. 82: 1–6