Chemistry Reference
In-Depth Information
We have completed two complementary gas/surface collision studies with
squalane, using incident O(
3
P) with 5 and 1 eV and F(
2
P) with 1 and 0.5 eV initial
translational energy [
28
,
29
]. To understand the viability of using MSINDO for the
QM method, we computed the reaction enthalpies and barriers for the H abstraction
and elimination reactions of O/F with methane and ethane, and compared the results
to those from the advanced ab initio method CCSD(T). These comparisons show
that for the reaction barriers, which are important for the bond making/breaking
events in our simulations, the MSINDO computations are comparable (~0.15 eV)
to the CCSD(T) results. For the methane reaction, the H abstraction reaction (O
þ
CH
4
!
D
E
rxn
¼
CH
3
+ OH) has an
0.342 eV whereas with F the reaction is
D
E
rxn
¼
more exothermic (as expected) with an
1.418 eV, based on MSINDO.
Also, the reaction barriers are completely distinctive for these two reactions, with
the barrier for O + CH
4
being 0.564 eV and that for F + CH
4
only 0.167 eV. The H
elimination reaction (O + CH
4
!
CH
3
O + H) has a quite different thermochemis-
try with a
0.015 eV; the
reaction barriers are 1.869 and 1.769 eV, respectively. Generally, the MSINDO
computed thermochemistry with ethane has slightly lower reaction barriers as
compared to methane for both the H abstraction and elimination channels.
Considering the differences in thermochemistry between F and O reactions with
methane and ethane, it is not surprising that in our F + squalane studies, with
incident energies of 1.0 and 0.5 eV, no H elimination reactions are found; in
contrast, in O + squalane with the incident energy of 5 eV the probability of an
H elimination reaction is as high as 0.14. Additionally, in the O collision studies
double H abstraction and C-C bond cleavage occur with probabilities as high as
0.22 and 0.06, respectively. Another interesting difference in these two reactions is
in the probability that the product does not desorb from the surface by the end of the
simulation time (~10 ps). Upon H absorption, HF has a desorption probability of
0.90-0.95 (dependent on incidence angle), while OH has a bit lower probability of
0.71-0.83. This may be in part due to the average depth of penetration of the
incident atom before the reaction occurs, whereas the O average depth for H
abstraction is ~0.5
˚
deeper than for F. Also, HF is formed with a higher probabi-
lity than OH, with reaction probabilities of 0.78 and 0.41, respectively, which is
understandable given that the MSINDO reaction barriers for H abstraction (of
methane or ethane) are ~3-4 times higher for O than F. However, the copious
formation of HF does not necessarily translate into the efficient transfer of the
incident energy into the product's vibrational, rotational, and translational modes.
To examine these results in more detail, we now look at how the vibrational
states of the HDS component for the H abstraction product are affected by the
choice of F or O as the reactant. We compute these vibrational states based on the
classical histogram method [
48
], wherein the vibrational quantum numbers (calcu-
lated from the vibrational action [
49
]) is rounded to the nearest integer to determine
vibrational state populations. A similar method is used to define rotational quantum
number starting from the classical rotational angular momentum.
The dynamic QM/MM calculations show that the nascent gaseous product
HF leaves the surface with nearly all its vibrational distribution in either the
D
E
rxn
¼
0.081 eV while that for F + CH
4
is
D
E
rxn
¼
Search WWH ::
Custom Search