Chen et al.
Article
(Supelco, Bellefonte, PA). The gas mixture was analyzed by the
GC flame ionization detector for quantitative analysis. The gases
were sampled into the mass-selective detector through a fused
silica polyimide capillary (51 μm I.D., 363 μm O.D., 0.75 m
length) after separation in the GC column. After analysis, the
HCl in the gas mixture was neutralized in a solution of sodium
hydroxide prior to exhaust.
2.4. Data Collection Procedure. The reaction was allowed
to stabilize for 0.5-1 h at a particular set of conditions before the
initial data point was taken. During the course of the experiments,
changes in concentration and temperature were chosen in a
random order so that any variation in the system would not
introduce systematic error. At the end of each experiment, we
repeated the first data point in the series. The agreement between
repeat experiments verified there was no significant deactivation
of the catalyst during the experiment, except for CCl4 HDCl.
We determined the kinetic parameters, activation energy,13
reaction order in chlorocarbon,14 hydrogen,15 and hydrogen
chloride,16 and the rate (at different temperatures), while chan-
ging the [HCl]/[H2] ratio for each compound. The reactant
conversion was maintained at <5% in order to assume differ-
ential reactor operation. The reactivity of the individual com-
pounds varied dramatically, requiring measurements of reaction
kinetics at low conversion over a large temperature range. HCl
was included in the feed of all reactions in order to maintain its
concentration constant through the catalyst bed, since HCl
inhibits the hydrodechlorination reaction and limits the con-
version. We compared data from the same experiments carried
out at identical conditions to calculate experimental errors. The
maximum error in reaction orders is 25%, and the error in the
apparent activation energies was 5 kJ mol-1. The reproducibility
of the turnover rate was g90%.
Figure 1. Apparent reaction orders (CH3Cl ([), H2 (2), HCl (9)
and apparent activation energy (O) for the hydrodechlorination of
CH3Cl. The reaction conditions are listed below.13-17
approximation (LDA), with the exchange-correlation functional
proposed by Perdew and Zunger20 and corrected for nonlocality
in the generalized gradient approximations (GGA) using the
Perdew-Wang 91 funtional.21 The interaction between the core
and electrons is described using the ultrasoft pseudopotentials
introduced by Vanderbilt22 and provided by Kresse and Hafner.23
The Pd(110) surface is modeled by a periodic five layer-slab
with the adsorbate placed on one side of the slab. The unit cell size
is 3 ꢂ 3; this size enabled the investigation of different adsorbate
coverage and the presence of different species simultaneously on
the Pd(110) surface. The choice of a limited number of metallic
layers in the model is imposed by the size of the system for which a
tractable calculation may be performed in reasonable time. One
slab is separated from its periodic image in the z-direction by a
Deuterium-exchange experiments were conducted in a well-
mixed batch reactor.17 The reactor was evacuated to 10-5 Torr
with a diffusion pump for a 0.5 h before the introduction of
reactants. The temperature of the oven was raised to the target
temperature at a rate of 5 K min-1. A sample of the gas mixture
was taken from the reactor using a gastight 500 μL syringe
(Hamilton Company, Reno, NV) through a septum fitted to a
union mounted on a sample port on the batch reactor loop. For
each data point, ∼150 μL of the gas mixture was injected into the
GC-MS for analysis.
˚
vacuum space of about 11.4 A. All atoms, except from the two
bottom layers of the slab, were allowed to relax in all optimiza-
˚
tions. The value for the bulk Pd-Pd distance was set at 2.81 A
based on the model of Filhol et al.24 in order to reduce the effect of
the stress in the metallic layer. This optimized value is in good
agreement with the experimentally determined Pd-Pd bulk dis-
2.5. Density Functional Theory Calculations. All calcula-
tions were performed using Vienna Ab-Initio Simulation Package
(VASP) code to perform periodic density functional theory
(DFT) calculations using pseudopotentials and a plane wave
basis set.18,19 The DFT was parametrized in the local-density
25
˚
tance of 2.75 A.
In all calculations performed here, the Brillouin-zone integra-
tions have been done on a 2 ꢂ 3 ꢂ 1 Monkhorst-Pack grid of
k-points for all structures, which yields a correct convergence for
the calculated energy (at a reasonable computational cost). The
search for the transition state is done by following the reaction
coordinate and further optimization of the stationary point along
this reaction path, as described elsewhere.26 These transition
states were further checked by frequency calculations using a
VASP internal routine. The surface was kept frozen during these
calculations. In most of the cases, only one imaginary frequency
was found. In the few cases where two imaginary frequencies were
found, the second one was negligible, as observed previously in a
study of chlorinated molecules.26
(13) The conditions for the determination of the apparent activation energy for
the four chlorinated methane compounds were as follows: 48-92 ꢀC, 10 Torr CCl4,
155 Torr H2, and 58 Torr HCl (CCl4); 92-143 ꢀC, 14 Torr CHCl3, 235 Torr H2,
and 48 Torr HCl (CHCl3); 187-229 ꢀC, 22 Torr CH2Cl2, 196 Torr H2, and 38 Torr
HCl (CH2Cl2); 227-261 ꢀC, 220 Torr CH3Cl, 181 Torr H2, and 74 Torr HCl
(CH3Cl).
(14) The conditions for the determination of the apparent reaction order in
chlorinated methane were as follows: 61 ꢀC, 9-16 Torr CCl4, 173 Torr H2, and 45
Torr HCl (CCl4); 114 ꢀC, 15-40 Torr CHCl3, 237 Torr H2, and 49 Torr HCl
(CHCl3); 201 ꢀC, 21-43 Torr CH2Cl2, 199 Torr H2, and 38 Torr HCl (CH2Cl2);
227 ꢀC, 46-193 Torr CH3Cl, 159 Torr H2, and 64 Torr HCl (CH3Cl).
(15) The conditions for the determination of the apparent reaction order in
hydrogen were as follows: 114 ꢀC, 16 Torr CCl4, 129-256 Torr H2, and 50 Torr
HCl (CCl4); 114 ꢀC, 18 Torr CHCl3, 183-399 Torr H2, and 48 Torr HCl (CHCl3);
198 ꢀC, 25 Torr CH2Cl2, 108-379 Torr H2, and 37 Torr HCl (CH2Cl2); 224 ꢀC, 145
Torr CH3Cl, 82-306 Torr H2, and 48 Torr HCl (CH3Cl).
(16) The conditions for the determination of the apparent reaction order in
hydrogen chloride were as follows: 115 ꢀC, 17 Torr CCl4, 173 Torr H2, 17-50 Torr
HCl (CCl4); 115 ꢀC, 20 Torr CHCl3, 155 Torr H2, 20-64 Torr HCl (CHCl3);
199 ꢀC, 25 Torr CH2Cl2, 127 Torr H2, 22-64 Torr HCl (CH2Cl2); 227 ꢀC, 193 Torr
CH3Cl, 159 Torr H2, 10-64 Torr HCl (CH3Cl).
(17) The reaction conditions for the hydrodechlorination of CH4-xClx (x =
1-3) compounds with deuterium were as follows: 205 ꢀC, 416 Torr CH3Cl, 434
Torr D2 (CH3Cl); 184 ꢀC, 30 Torr CH2Cl2, 261 Torr D2 (CH2Cl2); 95 ꢀC, 18 Torr
CH2Cl2, 328 Torr D2 (CHCl3).
Gas phase C-Cl bond strength values for CH4-xClx (x = 1-4)
were calculated by ab initio DFT calculations using the commer-
cially available Gaussian 98 software package, with the B3PW91/
6-311þG(2d,p)//B3PW91/6-311þG(2d,p) functional and basis set.
For each compound, the energy of the chlorocarbon, energy of
(20) Perdew, J. P.; Zunger, A. Phys. Rev. B 1981, 23, 5048–5079.
(21) Perdew, J. P.; Yue, W. Phys. Rev. B 1986, 33, 8800–8802.
(22) Vanderbilt, D. Phys. Rev. B 1990, 41, 7892–7895.
(23) Kresse, G.; Hafner, J. J. Phys.: Condens. Matter 1994, 6, 8245–8257.
(24) Filhol, J. S.; Simon, D.; Sautet, P. J. Phys. Chem. B 2003, 107, 1604–1615.
(25) Kittel, C. Introduction to Solid State Physics; John Wiley: Chichester, 1996.
(26) Barbosa, L. A. M. M.; Sautet, P. J. Catal. 2002, 207, 127–138.
(18) Kresse, G.; Furthmuller, J. Phys. Rev. B 1996, 54, 11169–11186.
(19) Kresse, G.; Furthmuller, J. Comput. Mater. Sci. 1996, 6, 15–50.
Langmuir 2010, 26(21), 16615–16624
DOI: 10.1021/la1020753 16617