Chemistry Reference
In-Depth Information
atoms) were computed with semiempirical QM (MSINDO), an efficient approach
to describe bond-breaking for the reactions we are interested in. The coordinates
and forces in our simulations are propagated in time using a rather robust fifth/sixth
order predictor-corrector scheme. One unique aspect of our QM/MM algorithm is
that we allow for a dynamic partitioning of atoms to be treated either with MM or
QM, which we refer to as the “seed atom” method. Though we took a simplistic
approach to the dynamic partition of atoms, we found this to be very useful in
studying a surface like squalane that is very porous, allowing the incident atoms lots
of freedom to move around within the liquid surface. We further discussed in
Sect.
2.3
some of the difficulties involved in reassigning atoms in and out of the
reactive region, as this changes the forces computed from one step to the next with a
different level of theory.
Thus far, we have applied our QM/MM-MD model to study reactions at the
surface of a well known hydrocarbon liquid, squalane, and we have also provided
perhaps the first theoretical study of gaseous atom scattering from an RTIL surface.
With squalane we observed the reactivity of both O(
3
P) and F(
2
P) scattering from
the surface at a few different initial translational energies. The chemistry that
ensued upon bombardment with F(
2
P) with 1.0 and 0.5 eV incident energy was
limited to H abstraction. However, when O(
3
P) hits the surface of squalane at 5 eV,
it is capable not only of H abstraction but also of H elimination and C-C bond
scission. Even though understanding the barriers to these reactions helps determine
whether there is enough initial translation energy to overcome these barriers, it is
only through the dynamics simulations that we gain an insight into the mechanism
of these reactions and the likelihood of these reactions occurring as a function of
position relative to the liquid interface. One major theme that is noted throughout
the discussion in Sect.
4.1
is that the molecular energy (be it vibrational, rotational,
or translational) of the nonthermal nascent products emerging from the surface is
predominantly due to the bond making/breaking event and is largely independent of
the incident atom's energy and collision angle. This indicates that nonthermal
products are incompletely relaxed as they emerge from the liquid surface, even
when they are produced a few angstroms beneath the surface.
Probing the reactivity of our RTIL surface of [emim][NO
3
] was done with both
Ar and O(
3
P) at 5 eV to study both the nonreactive and reactive scattering. The Ar
scattering shows that proton transfers can occur between the cations and anions, and
there exists a small probability that the neutral species emim or HNO
3
can subse-
quently desorb during our simulation time of 7.3 ps. Desorption of neutral emim
and HNO
3
can also occur in the O(
3
P) scattering simulations and other neutrals
such as HNO
2
and O
2
can also be produced. Bombardment of the RTIL with O(
3
P)
can also produce many other chemical species. Consistent with the Minton et al.
experimental studies, we find that the abstraction product OH desorbs from the
RTIL, resulting in an average translational energy that is about half of the input
energy at an angle of incidence of 60
, and even less when the angle of incidence is
more direct.
The QM/MM applications that we have considered have focused on hyperther-
mal chemistry for the most part, and this is a natural direction for this research due
Search WWH ::
Custom Search