324
HUYBRECHTS ET AL.
are used. The temperature range is 565–843 K and the
pressure of C2HCl5 lies between 0.5 and 760 torr. Ex-
periments in the presence of added He, toluene, and
Cl2 are also considered. Reaction conversions amount
to up to 85%.
optimization or fitting procedure is based on the least-
squares method. The reaction model used is given in
Tables II and III. It involves the homogeneous elemen-
tary reaction steps from Ref. [1] and the heterogeneous
reaction steps suggested in Ref. [4]. Other homoge-
neous and heterogeneous reactions were found to be
negligible for the experimental conditions modeled.
The kinetic Arrhenius parameters E and A from Ta-
ble II are known or estimated for the homogeneous
reactions and estimated or unknown for the heteroge-
neous reactions. The thermochemical parameters from
Table III, which were used to calculate E and A for
back reactions from the “principle of microscopic re-
versibility”, are known, except the enthalpy of forma-
tion and the entropy of ClS, a Cl atom adsorbed on a
surface site S covered with a pyrolytic carbon film. The
unknown parameters were used as adjustment param-
eters. Starting from the initial values, selected kinetic
and thermodynamic parameters were optimized simul-
taneously for the set of experimental points from this
work and from literature [1–5] shown in Figs. 1, 2,
7, 9, and 10 given later (see Table I for an overview
of the experimental conditions). The adjusted and op-
timized parameters and the corrections that were ap-
plied to the known or estimated initial values are also
given in Tables II and III. As can be seen from Ta-
ble II, the corrections for the homogeneous reactions
do not exceed 0.5 kcal/mol for activation energies
and 0.17 Log10 units for A-factors. The optimized
activation energies for the heterogeneous recombina-
tion desorption steps 6 and 7 and the adsorption
step 8 are between 0 and 2 kcal/mol. These low val-
ues are reasonable for such types of exothermic reac-
tions (see differences between E values for reaction
couples from Table II). Their optimized A-factors are
below the initial maximum values calculated from the
wall collision rate of the kinetic gas theory. The small
corrections may be regarded as a steric factor p of 0.02
EXPERIMENTAL
C2HCl5 (Janssen Chimica) was purified by prepara-
tive gas chromatography to a purity level higher than
99.9%. The pyrolyses were carried out in a cylindri-
cal quartz vessel (length = 12 cm, volume = 100 ml,
S/V = 1.04 cm 1). The fresh reactor walls were condi-
tioned by pyrolyzing large amounts of C2HCl5 (ca. 100
torr) at high temperature (ca. 810 K) until reproducible
kinetic results were obtained.
The conventional static system and the procedure
were similar to those described previously [7]. The de-
hydrochlorination of C2HCl5 was followed by measur-
ing the rise in the total pressure 1P(total) with a Pyrex
Bourdon gauge. The reaction products were analyzed
by gas chromatography (GC). The sampling technique
employed [8] evacuates all the products from the re-
action cell by condensation at 196 C. The analyses
were performed on a 2 m column of 30% w/w Silicone
OV101 using 60–80 mesh Chromosorb PAW, operated
between 70 and 170 C at a rate of 2 C/min with a he-
lium flow rate of 3.6 l/h. The analyses show that there
are two parallel reactions, the dehydrochlorination
C2HCl5 → C2Cl4 + HCl
and the dechlorination
(J)
C2HCl5 → C2HCl3 + Cl2
(K)
The latter is about 150 times less important. The
amount of Cl2, about 150 times smaller than HCl, could
not be measured quantitatively by GC. Trace quantities
of C2Cl6 and 1,1,2,2-C2H2Cl4 were detected, whereas
chloromethanes were not observed. 1P(total) is thus
a very good representation of the pressure of C2Cl4 or
HCl produced by reaction (J).
and 0.46 for the recombination desorption steps
6
and 7 and a sticking coefficient of 0.23 for the ad-
sorption step 8, respectively. Note that p is, as ex-
pected, smaller for the C2HCl4 radical compared with
that for the Cl atom. The adjustable steric factor p can
also be considered a catchall for effects not included
in the model, fitting errors, etc. The diffusion of the
species involved in the heterogeneous reactions is not
contributing significantly to the decreased steric fac-
tor p. Simulations accounting for this process indicate
that no important concentration gradients develop. An
analysis of the rates of the reaction steps reveals that
the consumption (or formation) of species at the wall is
determining the rate of the heterogeneous process, and
not the diffusion of the species to (or from) the reactor
wall.
NUMERICAL MODELING OF THE
EXPERIMENTAL RESULTS
The simulated and optimized curves were obtained
using OPTKIN, a user-friendly PC program for
mechanistic modeling of reactions by kinetic and
thermodynamic parameter optimization [9]. This pro-
gram integrates stiff, coupled kinetic differential
rate equations and performs sensitivity analyses. Its