Download Heriot-Watt University Scattering Dynamics of Oxygen Atoms on

Heriot-Watt University
Heriot-Watt University
Research Gateway
Scattering Dynamics of Oxygen Atoms on Imidazolium Tetrafluoroborate Ionic Liquid
Surfaces
Marshall, Brooks C.; Smoll, Eric J.; Purcell, Simon; Costen, Matthew Lawrence; McKendrick,
Kenneth George; Minton, Timothy K.
Published in:
Journal of Physical Chemistry C
DOI:
10.1021/acs.jpcc.6b02288
Publication date:
2016
Document Version
Accepted author manuscript
Link to publication in Heriot-Watt University Research Gateway
Citation for published version (APA):
Marshall, B. C., Smoll, E. J., Purcell, S. M., Costen, M. L., McKendrick, K. G., & Minton, T. K. (2016). Scattering
Dynamics of Oxygen Atoms on Imidazolium Tetrafluoroborate Ionic Liquid Surfaces: Dependence on Alkyl Chain
Length. Journal of Physical Chemistry C, 120(23), 12472-12483. DOI: 10.1021/acs.jpcc.6b02288
Scattering Dynamics of Oxygen Atoms on
Imidazolium Tetrafluoroborate Ionic Liquid
Surfaces: Dependence on Alkyl Chain Length
Brooks C. Marshall,1 Eric J. Smoll, Jr.,1 Simon M. Purcell,2 Matthew L. Costen,2 Kenneth G.
McKendrick,*2 Timothy K. Minton*1
1
Department of Chemistry and Biochemistry, Montana State University, Bozeman, Montana
59717, USA
2
Institute of Chemical Sciences, School of Engineering and Physical Sciences, Heriot-Watt
University, Edinburgh EH14 4AS UK
AUTHOR INFORMATION
Corresponding Authors:
*Email: [email protected]; [email protected]
1
ABSTRACT
The dynamics of oxygen-atom scattering from the surfaces of a series of roomtemperature ionic liquids (RTILs) containing the 1-alkyl-3-methylimidazolium cation [Cn mim]
and the tetrafluoroborate anion [BF4] were studied by reactive-atom scattering with mass
spectrometric detection of products (RAS-MS). The length of the alkyl chain was varied (n = 4,
8, 12) in order to investigate the relationship between the scattering dynamics and the density of
alkyl chains on the surface and their ordering.
RAS-MS uses a beam-surface scattering
technique, with a hyperthermal O-atom beam source and a rotatable mass spectrometer detector.
Time-of-flight and angular distributions were collected for inelastically scattered O and
reactively scattered OH and H2O, enabling product flux distributions to be obtained as a function
of final translational energy and scattering angle, P(ET,θf). A new analysis technique was used to
separate these distributions into three distinct components, corresponding to three dynamical
pathways for scattered products: fast and slow impulsive scattering and thermal desorption. The
results of these experiments support previous findings that surface alkyl coverage increases with
increasing alkyl chain length, but these dynamical studies provide the additional insight that the
ionic liquids with alkyl chains of n = 8, 12 have smoother surfaces than the ionic liquid with n =
4 and the pure alkane liquid, squalane (2,6,10,15,19,23-hexamethyltetracosane).
2
I. INTRODUCTION
Room temperature ionic liquids (RTILs) are salts that have melting points below 100 °C.
The ionic liquid-gas interface is particularly important for many applications of interest such as
gas capture and storage,1 gas chromatography,2-3 and supported ionic liquid phase catalysis. 4-5
Knowledge of how gases react with or enter ionic liquids will hopefully contribute to a set of
“design rules” that will ultimately allow ionic liquids to be tailored to maximize efficiency and
minimize waste in these processes.6
Many experimental and theoretical techniques have been used to study the ionic liquidgas interface. A large number of studies have focused on a family of ionic liquids that contain
the 1-alkyl-3-methylimidazolium ([Cnmim]) cation. The wide range of experimental techniques
employed includes sum frequency generation (SFG),7-9 angle-resolved photoelectron
spectroscopy (ARXPS),10-12 low-energy ion scattering (LEIS),13-15 Rutherford backscattering
(RBS),16-18 neutral impact collision ion scattering spectroscopy (NICISS),
19-20
direct recoil
spectroscopy (DRS),21-22 neutron reflectometry,23 and reactive atom scattering (RAS). 24
Theoretical work has employed molecular dynamics (MD) simulations24-26 to study the structure
of the interfacial region and a hybrid approach involving quantum mechanics and molecular
mechanics (QM/MM) to study reactive collisions at the gas-liquid interface.25,
27-28
Both the
experimental results and associated simulations indicate that the surfaces of these liquids become
increasingly enriched with hydrocarbon moieties as the alkyl chain length on the imidazolium
ring increases or as the size of the anion decreases.24
An understanding of surface speciation, which many techniques offer, is not sufficient to
understand gas-liquid interactions.
Impinging gas molecules may follow a multitude of
pathways involving direct scattering, thermal accommodation, and reaction. The organization
3
and structure of interfacial molecules and their ability to absorb energy will dictate the fate of an
incident gas atom or molecule. The outermost molecules of a liquid are often organized in a way
that is different than the bulk. Several groups have conducted extensive studies of the dynamics
of gas-liquid scattering.25,29-42 Imidazolium based ionic liquids have been the focus of a number
of these studies.25, 43-46
We describe here a study which uses the reactive-atom scattering technique with mass
spectrometric detection (RAS-MS) to collect translational energy and angular distributions of
non-reactive as well as reactive products that scatter from liquid surfaces. In particular, we
provide detailed experimental results on the relationship between the reactive O-atom scattering
dynamics and the alkyl chain length for the [Cn mim] cation (n = 2, 8, 12) and the
tetrafluoroborate, [BF4], anion. Such dynamics probe the relative density of alkyl groups at the
ionic liquid surface, and they also reflect the relative smoothness of the surface. We also
propose a new method to distinguish between two different mechanisms for nonthermal
scattering, in which products (either reactive or non-reactive) scatter impulsively from the
surface on a time scale too short for thermal equilibrium to be attained. The new dynamical
information is complementary to our previous study,24 which used LIF and mass spectrometric
detection, as well as molecular dynamics (MD) simulations. Our previous work focused on OH
product yields as a way to infer the relative number of abstractable hydrogen atoms at the
surface, whereas the new work described here uses the dynamical behavior of molecular beam
scattering data to gain insight into the surface structure beyond the identity of species that reside
at the liquid-gas interface.
4
II. EXPERIMENTAL METHODS
The RAS-MS experiments were conducted with the use of a universal crossed-molecularbeams apparatus, configured for beam-surface scattering studies, which employed a rotatable
mass spectrometer detector.24-25,
38
A laser-detonation source47 was used to prepare a pulsed
beam containing hyperthermal O atoms from the breakdown of O2 gas. The production of the
hyperthermal beam begins with a pulse of O2 gas at a stagnation pressure of 550 psig that is
injected through a 1 mm orifice into a conical nozzle from a home-built piezoelectric pulsed
valve. A 7 J pulse of light from a CO2 TEA laser is focused into the throat of the nozzle and
induces a breakdown and heats the gas to >40000 K, whereupon the associated shockwave
dissociates the majority of O2 in the nozzle and accelerates the gas to a nominal velocity of ~7.8
km s-1. For the experiments described here, the beam pulse that exited the nozzle contained
~96% atomic oxygen and ~4% molecular oxygen, both in their ground electronic states, O(3P)48
and O2(3Σg-).49 The pumping speed of the source region limited the repetition rate of the laserdetonation source to 2 Hz.
The hyperthermal beam was characterized by aligning the detector and molecular beam
axes and collecting time-of-flight (TOF) distributions (number density vs. time, N(t)) of massselected products over a known distance from the point near the nozzle orifice where the beam
pulse was initiated and the ionizer of the mass spectrometer detector. The hyperthermal beam
pulse passed through a 2 mm diameter skimmer that was 77 cm from the nozzle orifice and then
through a collimating slit, 1.3 mm wide and 2.8 mm high, which was 3.5 cm downstream from
the skimmer. The center of rotation of the rotatable quadrupole mass spectrometer detector was
18.5 cm from the collimating slit, and the distance from this point to the Brink-type electronimpact ionizer50 of the detector was 33.7 cm. Thus, the total flight path from the nozzle orifice
5
to the detector was 132.7 cm. A synchronized chopper wheel, with three 1.5 mm wide slots and
a rotation speed of 300 Hz, was placed between the collimating slit and the rotation center of the
detector in order to select a narrow range of velocities from the overall beam pulse. Ions,
produced in the ionizer, were filtered by mass-to-charge ratio, m/z, using a quadrupole ion filter
and then detected by a Daly-type ion counter.51
TOF distributions were registered on a
multichannel scaler as ion counts as a function of arrival time. The neutral flight time from the
source to the ionizer was found by subtracting the ion flight time (over the distance from the
ionizer to the ion counter), given by  m / z , where the parameter, α, was empirically
determined to be 3.4 𝜇𝑠√𝑒⁄𝑎𝑚𝑢.
From these TOF distributions, the translational energy
distributions (P(ET), proportional to flux) of O and O2 were derived. Representative translational
energy distributions of the O and O2 components of the chopped beam are shown in Fig. 1. The
experiments were performed on several different days, and the beam conditions varied slightly
each day. The average translational energy of the oxygen atoms ranged from 504 to 511 kJ mol 1
, and the mole fraction of oxygen atoms in the beam varied between 93 percent and 98 percent.
These variations have been explicitly corrected for in the analysis as described below.
The hyperthermal O-atom beam was directed at a continually refreshed liquid surface,
held at 323 K, that was formed by the pick-up of a film on a stainless steel wheel that rotated at
0.25 Hz through a reservoir of liquid (25 ml volume). The liquid surface was placed at the center
of rotation of the mass spectrometer detector such that the surface normal and the molecular
beam axis were contained in the detector rotation plane. After the beam pulse struck the surface,
the resulting scattered species were detected as TOF distributions using the mass spectrometer,
as described above, with a neutral flight path of 33.7 cm. The TOF distributions were collected
for a variety of final angles at specific angles of incidence, θi. Thus, for θi, the product flux as a
6
function of translational energy and final angle, P(ET,θf), could be derived. Note that θi and all
final angles, θf, are polar angles with respect to the surface normal and are in the same plane as
the incident beam and surface normal, with positive θf being defined as angles on the opposite
side of the surface normal from the incident beam.
Three ionic liquids, [C4mim][BF4], [C8mim][BF4 ], and [C12mim][BF4], were used, and
these were acquired commercially from IoLiTec Ionic Liquids Technologies Inc. Beam-surface
scattering experiments were also performed on a pure liquid alkane (squalane) surface, which is
a well-studied liquid that was used as a reference in these experiments. Squalane was purchased
from Sigma Aldrich. The purities of the starting liquids are given in Table S1 of Supporting
Information.
Each liquid sample was degassed under rough vacuum overnight and then
transferred to the liquid reservoir inside the scattering chamber. The sample was then degassed
for >12 hours at 323 K under high vacuum (<10-6 Torr), with the wheel turning in the liquid
reservoir, prior to the collection of any data.
TOF distributions were collected for scattered species with m/z = 16 (O+), 17 (OH+), 18
(H2O+), and 32 (O2+). The TOF distributions for m/z = 16 and 17 were corrected for the known
contributions from dissociative ionization of O2 and H2O. For each mass, TOF distributions
were collected for three incidence angles, θi = 30°, 45°, 60°. The final angles ranged, in 5°
increments, from θf = -10° to 70°, θf = 5° to 70°, and θf = 20° to 70°, for incidence angles of θi =
60°, 45°, and 30°, respectively. The limits to the ranges of final angles were dictated by
experimental geometry.
In order to reduce the effect of long-term drifts in incident beam
intensity and other experimental parameters, the final angle was increased and then decreased
systematically until two TOF distributions had been collected for each m/z at each θf for a given
θi. Then the liquid reservoir was lowered out of the path of the beam, and the incident beam was
7
interrogated by collecting TOF distributions of the O and O2 components of the beam. This
process was then repeated, which resulted in a total of four TOF distributions for each m/z at
each θf, for a given θi. The four TOF distributions were each normalized to the relevant beam
intensity and then summed.
III. RESULTS AND ANALYSIS
Representative TOF distributions of O, OH, and H2O scattering from [C12 mim][BF4] are
shown in Fig. 2. These TOF distributions appear to be bimodal, and, as has been widely invoked
in previous related work, the two components were first assumed to arise from two limiting
dynamical cases for scattering: impulsive, or nonthermal, scattering (IS) and thermal desorption
(TD).33, 52 When a reaction occurs, the analogous cases may also be referred to as Eley-Rideal
and Langmuir-Hinshelwood, respectively.53 The products that arrive at the ionizer after longer
flight times (slower component in the TOF distribution) are assumed to be the result of TD, with
a Maxwell-Boltzmann (MB) speed distribution characterized by the surface temperature. The
TD mechanism implies that the desorbed atoms or nascent reactive products have no memory of
the trajectory of the incident O atoms, whereas the IS mechanism involves inelastic or reactive
scattering on a time scale so brief that some memory of the incident trajectory is retained. The
IS products thus scatter from the surface with a superthermal speed distribution.
From the TOF distributions of scattered species, translational energy distributions, P(ET),
which are flux probability density functions, were determined by using standard analysis
methods that convert from number density to flux (P(ET) ∝ N(t)/t), on the assumption of a
monoenergetic beam.38,
54-55
The justification for this assumption is discussed in Supporting
Information. The first step in the separation of the IS and TD components of a TOF distribution
8
is to transform a Maxwell-Boltzmann distribution of translational energies at the surface
temperature to a TOF distribution and adjust its amplitude to fit the slow (TD) component (red
curves in Fig. 2). Then the IS component of the TOF is determined by subtracting the fitted TD
component from the overall TOF distribution (blue curves in Fig. 2). This TOF distribution is
then inverted to yield a P(ET) distribution for the IS component. The separation of the two
components in this way was is justified by the clear bimodal appearance of the TOF distributions
and the fact that a MB distribution at the surface temperature matches the slower components of
the TOF distributions very well. After the smooth TD simulation is subtracted in the TOF
domain, there is significant noise in the slow part of the remaining IS TOF distribution, which,
when converted to the energy domain results in a high frequency variation in the low-energy
region of the P(ET) distribution. This apparent noise in the low-energy region is not related to
the scattering dynamics, so we have smoothed the P(ET) distributions using a spline interpolation
(see Supporting Information). Figure 3 shows the smoothed P(ET) distributions for impulsively
scattered species that correspond to the TOF distributions shown in Fig. 2.
The P(ET) distributions were also integrated to give relative flux, allowing angular
distributions of scattered flux to be determined. Representative angular distributions for O, OH,
and H2O scattering from the four liquid surfaces for θi = 60° are shown in Fig. 4. Analogous
distributions for θi = 45° and 30° are shown in Figs. S1 and S2, respectively. The scattering of O
and OH is dominated by nonthermal (IS) mechanisms, whereas both thermal (TD) and
nonthermal (IS) mechanisms are significant in the production of H2O from the surface. The
angular distributions of the TD components are described well by the expected cos(f) function
for all incidence angles and for every scattered product. The IS components have maxima far
9
from the surface normal and even beyond the specular direction for the most of the O and OH
products.
The P(ET) distributions and flux angular distributions have been combined into flux
distributions that are a function of both translational energy and final angle, P(E T,f). Because
the scattered products must have lower translational energies than the incident beam, these
distributions were created by truncating all of the smoothed flux-weighted P(ET) distributions at
the incident beam energy (discussed in Supporting Information in the context of the
monoenergetic beam assumption in the analysis). The reactions of hydrocarbons with O(3P) to
form OH and hydrocarbon radicals are only modestly exothermic. The enthalpies of hydrogen
abstraction are -4.5 and -18.5 kJ mol-1 for primary and secondary H atoms, respectively. 56 These
reaction exothermicities are less than 4% of the energy of the incident beam. Because the OH
products may come from abstraction of primary or secondary H atoms and these processes are
nearly thermoneutral, the contributions of these reaction exothermicities to the available energy
have been neglected in the analysis. The P(ET,f) distributions shown in Fig. 5 are interpolated
between measured P(ET) distributions, collected at 5° increments in θf, to create a smooth image
of the flux as a function of translational energy and final angle for an incidence angle of θi = 60°.
The red color in Fig. 5 indicates maximum flux and the dark blue color indicates zero flux. The
scale is different for each column, but it is consistent within each column (corresponding to a
given scattered product).
IV. DISCUSSION
The relationship between surface structure and scattering dynamics is manifested in the
translational energy and angular distributions of the scattered products. We previously proposed
10
a soft-sphere kinematic model,52 similar to the hard-sphere or Baule model,53 that allows the
observed laboratory scattering dynamics to be related to collisions in the center-of-mass (c.m.)
reference frame, by analogy with gas-phase scattering experiments using crossed molecular
beams.57 This kinematic model implies that an impulsive (nonthermal) atom-surface collision is
localized on the surface, with a relatively small number of surface atoms being involved. Such
localized gas-surface collisions have been linked to scattering in the “structural regime” for atom
scattering on metal surfaces.58 If the collisions are localized, then the IS dynamics should
contain information about the nature of the surface on a molecular level. Accordingly, we focus
on the IS, or nonthermal, components of the TOF distributions from which IS translational
energy and angular distributions are derived. Implicit in this discussion is the assumption that
the IS collisions that lead to scattered O and OH involve single collisions on a stationary surface.
Earlier theoretical work has shown that the majority of the collisions of O atoms on a liquid
hydrocarbon surface have one inner turning point and lead to superthermal (IS) products,
whereas a minority of collision trajectories that lead to superthermal products may have multiple
inner turning points. These multiple-bounce collisions tend to yield products with lower
translational energies and broader angular distributions than are observed for single-bounce
collisions.59 We have taken steps to minimize the influence of multiple-bounce collisions in the
analysis (see below). In addition, we note that the velocities of the surface atoms due to thermal
motion are much smaller than the hyperthermal incident velocities used in our experiments and
can effectively be ignored.
The IS translational energy distributions seen in Figs. 3 and 5 appear to be bimodal, with
a broad distribution at higher energies (ET ≈ 100-500 kJ mol-1) and a narrower distribution at
lower energies (<ET> ~ 50 kJ mol-1) that are still roughly an order of magnitude higher than the
11
energies of thermally desorbed species in a MB distribution at the surface temperature (<ET> ~ 5
kJ mol-1). Such bimodal translational energy distributions have been observed in several studies
of hyperthermal atom-surface scattering.25, 52, 54-55 The kinematic analysis in the earlier study of
hyperthermal O-atom scattering on squalane54 suggests that the high energy component in the
translational energy distributions corresponds to scattering in the sideways direction in the c.m.
frame, whereas the low energy component corresponds to backward scattering in the c.m. frame.
The backward-scattered products presumably come from more head-on collisions with surface
moieties that transfer large amounts of energy into the surface collision partner in a single
(nonthermal) collision. Thus, the kinematic factors that lead to backward scattering in the c.m.
frame following single collisions with large energy transfers are largely responsible for the lowtranslational-energy products that are observed in the lab frame. It has been shown in gas-phase
scattering of hyperthermal O atoms with ethane that collisions that lead to backward-scattered
products, which presumably come from low-impact-parameter (“head-on”) collisions,
correspond to more energy transfer into internal modes of ethane, on average, than those that
lead to forward- and sideways-scattered products.60 In that study, the coupling between the
translational energy release and the scattering angle in the c.m. frame motivated an analysis
procedure in which different c.m. translational energy and angular distributions were used to
describe the forward/sideways-scattered products and the backward-scattered products. Given
the likelihood that the low and high energy components in the IS translational energy
distributions shown in Fig. 3 represent qualitatively different dynamics for single-bounce
scattering and the possibility that multiple-bounce collisions play a larger role in trajectories that
lead to low energy products than high energy products, we have considered only the high energy
12
products when attempting to find relationships between scattering dynamics and surface
structure.
The distributions shown in Fig. 3 and Fig. 5 show two distinct components in impulsive
scattering, one with a relatively low energy and broad angular distribution and another with a
much higher energy and super-specular angular distribution.
The high and low energy
components were separated algorithmically. Figure 6 shows a graphical representation of this
separation for at θi = 60°. At this incidence angle, TOF data were collected from -10° to 70° in
5° increments. As shown in Figs. 3 and 5, the P(ET,θf) surface contains a high-energy peak and a
low-energy peak. In general, these peaks are maximally separated at high final angles. Relative
to the high-energy peak, the low-energy peak is radially narrow with relatively small changes in
the extent of energy with angle. Also, as f is reduced from 70° to -10°, spectral weight is
ultimately transferred from the high-energy-peak to the low-energy peak. Because the separation
of these two distributions is most obvious at final angles far from surface normal, the first step in
the algorithm is to separate the two components at the maximum final angle, f = 70°. A region
of the high energy portion of the P(ET, θf = 70°) distribution (blue curve in Fig. 6A) that contains
the peak is fit with a log-normal distribution:
𝐸𝑇 2
𝑙𝑛 ( 𝐸 )
0
Phigh (𝐸𝑇 ) = 𝐴 𝑒𝑥𝑝 (−
)
𝑤
(1)
which provides a good empirical fit to asymmetric shape of the high energy portion of the
translational energy distribution, Phigh(ET). In this analysis, A, x0, and w are fitting parameters.
The log-normal fit (green curve in Fig. 6A) defines Phigh(ET, f = 70°), and its residual (orange
curve in Fig 6A) defines Plow(ET, f = 70°). For all angles below f = 70°, Phigh(ET,f) is defined
by a log-normal fit to P(ET,f) – Plow(ET, f + 5°) (red curve in Fig. 6B), and Plow(ET,f) =
13
P(ET,f) – Phigh(ET,f)
This iterative approach is justified by the fact that Plow(ET,θf) changes
only slightly between final angle increments and in this way the high energy portion is somewhat
isolated before it is fit. This method was tested for robustness by adjusting the initial log-normal
distribution fitting parameters and observing the changes in the resulting separated components
of P(ET, θf).
Plow(ET,f) and Phigh(ET,f) have very different flux angular distributions. Figure 7 shows
the flux angular distributions for both the low and high translational energy components of IS O
and OH, for a representative incidence angle of θi = 60°. The distributions in Fig. 7 have been
normalized to their peaks, in order to compare their shapes.
The low-energy scattering
components have angular distributions that are roughly cosθf, while the angular distributions of
the high-energy components are lobular in shape with maxima beyond the specular angle. The
angular distributions in Figs. 7A and 7B look similar to those for the total IS flux in Fig. 4,
because the majority of the flux is contained in the high-energy component.
The energy transfers to the surface associated with the Plow(ET,f) and Phigh(ET,f) product
distributions are also remarkably different. Figure 8 shows the fractional energy transfers as a
function of deflection angle, χ = 180° - (θi + θf), for the high- and low-energy components of IS
O and OH. The energy transfers for the low-energy components are high, close to 90%, and vary
little with deflection angle.
On the other hand, the energy transfers for the high-energy
components are strongly dependent on deflection angle and cover a broad range from about 3575%. The fractional energy transfers for the high-energy components are all fit well by the softsphere kinematic model that was presented in a previous publication.52
The near-cosine distributions and the large energy transfers that are characteristic of
Plow(ET,f) are superficially reminiscent of thermal desorption. However, these products are
14
clearly nonthermal (compared with average thermal desorption energies of ~ 5 kJ mol-1), with an
average translational energy of ~50 kJ mol-1 and a high-energy tail extending above 100 kJ mol-1
and are likely the result of atom-surface collisions that involve a single bounce or, perhaps, only
a few bounces, but in any case an interaction that occurs on a timescale too short for thermal
equilibrium to be attained.
The reason for the thermal-desorption-like appearance is that
backward scattering in the c.m. frame accompanied by a relatively large transfer of energy to
internal modes of the liquid is reflected by relatively slow product velocities in the laboratory
frame. This interpretation might apply also to Ne scattering on a self-assembled monolayer film,
where it was observed theoretically that single-bounce collisions could lead to a population of
scattered Ne atoms that appeared to be Maxwellian.61
The angular distributions corresponding to the Phigh(ET,f) components in Figs. 7A and
7B suggest a qualitatively different surface structure on the [C8 mim][BF4] and [C12 mim][BF4]
surfaces than on the [C4mim][BF4] and squalane surfaces.
Both the O and OH angular
distributions corresponding to [C8mim][BF4] and [C12mim][BF4] are narrower and more superspecular than the corresponding distributions for [C4 mim][BF4] and squalane. The same trends
are observed in Figs. 4 and S1, which show plots of total IS flux as a function of angle. Note that
the angular distributions for O and OH in Figs. 4 and S1 are dominated by the high-energy
scattering component, but they may be slightly distorted by the low energy component, which
was removed before creating the plots in Fig. 7. The angular distributions may be linked to
relative surface roughness. Rougher surfaces would be expected to correspond to localized
groups that protrude from the surface and promote more “head-on” and multiple-bounce
collisions with an incident O atom or a nascent OH radical, thus deflecting the scattered products
over a broad angular range. In contrast, the higher probability of glancing collisions on a smooth
15
surface would be expected to result in products that scatter into a relatively narrow range of
super-specular (“forward”) angles.
Thus, the angular distributions for [C8 mim][BF4] and
[C12mim][BF4], with their narrower and more super-specular shapes, suggest that the surfaces of
these liquids are slightly smoother than the surfaces of the other two liquids. A relatively smooth
surface could result if the C8 and C12 alkyl chains of the imidazolium cations tended to occupy
the surface in an orientation that is parallel to the liquid-vacuum interface or if these chains
packed tightly, either lying parallel to the interface or standing up vertically.
Our recent
experimental study24 clearly shows that the hydrocarbon density at the surfaces of [C nmim]based ionic liquids grows with n but is not saturated even with n = 12; therefore, the surfaces
cannot have a structure of tightly packed chains. Earlier MD simulations25 of the surfaces of
[C2 mim][NTf2] and [C12mim][NTf2] and our more recent simulations24 of [Cnmim][NTf2] and
[Cnmim][BF4] surfaces indicate that the Cn alkyl chains of these liquids are concentrated at the
interface and do indeed tend to lie more parallel to the interface as the alkyl chain length grows,
but without forming tightly packed layers. Squalane, which is a branched alkane, appears to
project both methyl and methylene groups toward the interface 24, 41, 59 and therefore may have a
more molecularly rough surface than the [C8 mim] or [C12mim] ionic liquids.
Thus, the
conclusion from the angular distributions that the surfaces of [C 8 mim][BF4] and [C12 mim][BF4]
are smoother than those of [C4 mim][BF4] and squalane is direct experimental evidence of what
has been inferred from prior experimental data and MD simulations.
Turning to the fractional energy transfer data, the high-energy components of the IS O
and OH (Fig. 8) should, in principle, reflect the molecular nature of the four surfaces studied. As
the H-atom abstraction reaction to produce OH is almost thermoneutral62 and the mass change on
going from O to OH is small, the observed differences in the scattering behavior of O and OH
16
should reflect the fact that O atoms can scatter from any site on the surface, while OH can only
come from an incident O-atom collision with a hydrocarbon moiety. In the case of non-reactive
scattering of O atoms (Fig. 8A), the energy transfers on [C8 mim][BF4], [C12mim][BF4], and
squalane are essentially the same, within the error of the experiment, whereas the energy transfer
on [C4mim][BF4] is significantly less. The earlier experimental/molecular dynamics study of
alkyl surface coverage as probed by relative yields of OH26 indicates that the alkyl coverage
increases monotonically with alkyl chain length, with squalane by necessity being entirely
covered with alkyl groups. The relatively low O-atom energy transfer on [C4 mim][BF4] may
thus be explained by the small fraction of alkyl groups at the surface and the concomitantly large
fraction of anions. However, the energy transfers for O atoms on the other three liquids do not
follow the expected trend, where one would expect the energy transfer to increase with alkyl
chain length at the surface. We considered the possibility that the grazing incidence angle of 60°
might result in the incident O atoms being obscured by protruding alkyl groups that would block
access to stiffer, anion regions of the [C8 mim][BF4] and [C12 mim][BF4] surfaces, but we found
that these surfaces and squalane also exhibited the same trends in O-atom energy transfer when
the incidence angle was 45° (see Fig. S3). Apparently, the energy transfer of inelastically
scattered O atoms is not sensitive to the surface hydrocarbon density of RTILs with alkyl chain
lengths of at least C8 and higher. A similar observation was reported for the inelastic scattering
of CO2 on RTILs, where the insensitivity of the inelastic energy transfer data to alkyl chain
length on the imidazolium cation beyond 8 carbons led to the conclusion that the surface was
saturated with alkyl chains.44 Although reasonable, this was conclusion was demonstrated not to
be correct through the more-incisive monitoring of the yield of the reactive-product OH yield
from O atom scattering.[ref 24]
17
The fractional energy transfer data for reactively scattered OH (Fig. 8B) are apparently
more sensitive to the surface structure and show slight differences between all four liquids. As
might be expected, the energy transfer is lowest for reactive scattering on [C4 mim][BF4].
Surprisingly, however, energy transfer is greatest on [C8mim][BF4] and it decreases somewhat
on going from [C12mim][BF4] to squalane. The reactive scattering dynamics are thus sensitive to
the nature of the liquid surface, but the origin of this sensitivity is unclear.
A possible
explanation might be the number of primary hydrogen atoms at the surface relative to secondary
or tertiary hydrogen atoms. Our previous molecular dynamics simulations do indeed reveal
relative surface methyl group coverage that decreases in the order, [C8mim][BF4] >
[C12mim][BF4] > squalane > [C4mim][BF4]. If methyl groups present a stiffer collision partner
to the incident O atoms than methylene groups, then the observed trend in fractional energy
transfer should be observed. Other explanations are possible and are under investigation through
a detailed analysis of the molecular dynamics results.
The dynamics of the reactively scattered H2O product are much different than the
scattering dynamics of O and OH. These differences are clearly shown in the P(E T) distributions
in Fig. 3, as well as the total flux angular distributions in Fig. 4. This is not surprising because
water molecules must be formed by a more complex interaction of O atoms with the interface. 54
It is reasonable that these interactions require more energy transfer from the incident atom as two
hydrogen atoms must be abstracted. It is also important to note that the relative yield of H2O is
not the same for each of these liquids, which is to be expected because the evidence suggests that
the surfaces have differing fractions of hydrogen atoms present.
Previous work also
hypothesizes that the nearly monotonic increase in TD H2O products with increasing chain
length is a good representation of the amount of hydrogen present at the interface. 24
18
V. CONCLUSION
The dynamics of inelastic and reactive collisions of hyperthermal O( 3P) on [Cnmim][BF4]
(n = 4, 8, 12) ionic liquids and their comparison with the analogous scattering dynamics on the
pure hydrocarbon liquid, squalane, offer a direct experimental window into the surface structure
at the vacuum-liquid interface. The extent to which the scattering dynamics can serve as a useful
probe of the details of the surface structure depends on the localized nature of the atom-surface
collision. The picture of localized atom-surface collisions is justified in this work and in several
previous studies by the success in describing the impulsive-scattering dynamics with a simple
kinematic model. However, trajectories that involve multiple bounces on the surface or that lead
to very large energy transfers during a single bounce may reduce the quality of the information
on surface structure that is contained in the scattering dynamics. Therefore, we have focused on
the relatively simple inelastic collisions of O and the reactive collisions that produce OH, as
opposed to the more complex reactive collisions that produce H 2O. Furthermore, in an effort to
minimize the influence of trajectories with multiple bounces and large energy transfers on our
analysis, we developed a technique to remove the O and OH products that scattered backward in
the c.m. frame, manifested by their low (but still superthermal) translational energies, and keep
only those with high translational energies that scattered sideways in the c.m. frame. The highenergy products are most likely to arise from single-bounce collisions and retain the information
about the localized region of the surface where the collision occurred. The energy transfers
associated with these collisions (for both O and OH products) were described well by a twoparameter (effective surface mass and internal excitation) kinematic model, but the differences in
the parameters were too small to gain physical insight from their relative values. The key insight
19
derived from this work has thus come from the assumptions behind the novel separation the
impulsive scattering dynamics into two components, the angular distributions of scattered O and
OH and, to some extent, the differences in energy transfer associated with these inelastic and
reactive collisions.
The scattering dynamics presented here go beyond our earlier studies that focused on
surface reactivity. The earlier studies showed clearly that as the alkyl chain on the imidazolium
ring became longer, or if the anion were smaller, then the surface became more hydrocarbonrich. The angular distributions of inelastically scattered O and reactively scattered OH indicate
that the ionic liquid surfaces with n = 8 and 12 are smoother than the ionic liquid surface with n
= 4 and the squalane surface. This increased smoothness is interpreted as a propensity for the
longer alkyl chains on the imidazolium cation to congregate at the surface and lie down more or
less parallel to the interface, and not as a tight packing of the chains at the surface with any
particular order. The information revealed through the energy transfer dynamics is more subtle
than what was evident from the angular distributions. The inelastic energy transfers for squalane
and the ionic liquids with n = 8 and 12 were indistinguishable, indicating (in light of our earlier
work) that inelastic scattering alone is an insensitive probe of the hydrocarbon density at the
surface. However, the energy transfer associated with reactive scattering to form OH, which can
only originate from collisions with a hydrogen-containing moiety at the surface, was sensitive to
the liquid – albeit in an as yet unexplained way. The energy transfers in these collisions
followed the trend, [C4mim][BF4] < squalane < [C12 mim][BF4] < [C8mim][BF4], which cannot
be related simply to the hydrocarbon density at the surface. The energy transfers in reactive
collisions to form OH must therefore probe more detailed aspects of the surface structure, a
20
relationship we expect to be revealed with future theoretical studies of the liquid surfaces
themselves or of gas-liquid collision dynamics.
ACKNOWLEDGMENTS
This work has been supported by the U.S. National Science Foundation (NSF-CHE-1266032)
and by the U.K. Engineering and Physical Sciences Research Council (EP/K032062/1). We are
grateful to Maria A. Tesa-Serrate for many valuable discussions and to Philip J. Woodburn for
assistance in the collection of the data.
21
SUPPORTING INFORMATION
The Supporting Information is available free of charge on the ACS Publications website at DOI:
10.1021/acs.jpcc.6022886. Angular distributions of scattered flux for θi = 45°, 30°. Average
fractional energy transfers for θi = 45°. Comparison between directly inverted and smoothed
translational energy distributions for impulsive scattering. Justification for the use of the
monoenergetic beam approximation. Details of the purities of the four liquids used for the
experiments.
NOTES
The authors declare no competing financial interest.
22
REFERENCES
(1) Keskin, S.; Kayrak-Talay, D.; Akman, U.; Hortaçsu, Ö. A Review of Ionic Liquids Towards
Supercritical Fluid Applications. J. Supercrit. Fluid. 2007, 43, 150-180.
(2) Poole, C. F.; Poole, S. K. Ionic Liquid Stationary Phases for Gas Chromatography. J. Sep.
Sci. 2011, 34, 888-900.
(3) Anderson, J. L.; Armstrong, D. W. Immobilized Ionic Liquids as High-Selectivity/HighTemperature/High-Stability Gas Chromatography Stationary Phases. Anal. Chem. 2005, 77,
6453-6462.
(4) Mehnert, C. P.; Cook, R. A.; Dispenziere, N. C.; Afeworki, M. Supported Ionic Liquid
Catalysis − A New Concept for Homogeneous Hydroformylation Catalysis. J. Am. Chem. Soc.
2002, 124, 12932-12933.
(5) Van Doorslaer, C.; Wahlen, J.; Mertens, P.; Binnemans, K.; De Vos, D. Immobilization of
Molecular Catalysts in Supported Ionic Liquid Phases. Dalton T. 2010, 39, 8377-8390.
(6) Lovelock, K. R. J. Influence of The Ionic Liquid/Gas Surface on Ionic Liquid Chemistry.
Phys. Chem. Chem. Phys. 2012, 14, 5071-5089.
(7) Jeon, Y.; Sung, J.; Bu, W.; Vaknin, D.; Ouchi, Y.; Kim, D. Interfacial Restructuring of Ionic
Liquids Determined by Sum-Frequency Generation Spectroscopy and X-Ray Reflectivity. J.
Phys. Chem. C 2008, 112, 19649-19654.
(8) Iimori, T.; Iwahashi, T.; Ishii, H.; Seki, K.; Ouchi, Y.; Ozawa, R.; Hamaguchi, H.; Kim, D.
Orientational Ordering of Alkyl Chain at the Air/Liquid Interface of Ionic Liquids Studied by
Sum Frequency Vibrational Spectroscopy. Chem. Phys. Lett. 2004, 389, 321-326.
(9) Santos, C. S.; Baldelli, S. Alkyl Chain Interaction at the Surface of Room Temperature Ionic
Liquids: Systematic Variation of Alkyl Chain Length (R = C1-C4, C8) in both Cation and Anion
of [RMIM][R-OSO3] by Sum Frequency Generation and Surface Tension. J. Phys. Chem. B
2009, 113, 923-933.
(10) Lockett, V.; Sedev, R.; Bassell, C.; Ralston, J. Angle-Resolved X-Ray Photoelectron
Spectroscopy of The Surface of Imidazolium Ionic Liquids. Phys. Chem. Chem. Phys. 2008, 10,
1330-1335.
(11) Kolbeck, C.; Cremer, T.; Lovelock, K. R. J.; Paape, N.; Schulz, P. S.; Wasserscheid, P.;
Maier, F.; Steinrueck, H. P. Influence of Different Anions on the Surface Composition of Ionic
Liquids Studied Using ARXPS. J. Phys. Chem. B 2009, 113, 8682-8688.
(12) Lockett, V.; Sedev, R.; Harmer, S.; Ralston, J.; Horne, M.; Rodopoulos, T. Orientation and
Mutual Location of Ions at the Surface of Ionic Liquids. Phys. Chem. Chem. Phys. 2010, 12,
13816-13827.
(13) Caporali, S.; Bardi, U.; Lavacchi, A. X-Ray Photoelectron Spectroscopy and Low Energy
Ion Scattering Studies on 1-butyl-3-methyl-imidazolium Bis(trifluoromethane) Sulfonimide. J.
Electron. Spectrosc. 2006, 151, 4-8.
(14) Villar-Garcia, I. J.; Fearn, S.; De Gregorio, G. F.; Ismail, N. L.; Gschwend, F. J. V.;
McIntosh, A. J. S.; Lovelock, K. R. J. The Ionic Liquid-Vacuum Outer Atomic Surface: A LowEnergy Ion Scattering Study. Chem. Sci. 2014, 5, 4404-4418.
(15) Villar-Garcia, I. J.; Fearn, S.; Ismail, N. L.; McIntosh, A. J. S.; Lovelock, K. R. J. Fine
Tuning the Ionic Liquid-Vacuum Outer Atomic Surface Using Ion Mixtures. Chem. Commun.
2015, 51, 5367-5370.
(16) Nakajima, K.; Ohno, A.; Hashimoto, H.; Suzuki, M.; Kimura, K. Observation of surface
structure of 1-alkyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide using highresolution Rutherford backscattering spectroscopy. J. Chem. Phys. 2010, 133, 044702.
23
(17) Hashimoto, H.; Ohno, A.; Nakajima, K.; Suzuki, M.; Tsuji, H.; Kimura, K. Surface
Characterization of Imidazolium Ionic Liquids by High-Resolution Rutherford Backscattering
Spectroscopy and X-Ray Photoelectron Spectroscopy. Surf. Sci. 2010, 604, 464-469.
(18) Nakajima, K.; Miyashita, M.; Suzuki, M.; Kimura, K. Surface Structures of Binary Mixtures
of Imidazolium-Based Ionic Liquids Using High-Resolution Rutherford Backscattering
Spectroscopy and Time of Flight Secondary Ion Mass Spectroscopy. J.Chem. Phys. 2013, 139,
224701.
(19) Hammer, T.; Reichelt, M.; Morgner, H. Influence of the Aliphatic Chain Length of
Imidazolium Based Ionic Liquids on the Surface Structure. Phys. Chem. Chem. Phys. 2010, 12,
11070-11080.
(20) Ridings, C.; Lockett, V.; Andersson, G. Effect of the Aliphatic Chain Length on Electrical
Double Layer Formation at the Liquid/Vacuum Interface in the [C nmim][BF4] Ionic Liquid
Series. Phys. Chem. Chem. Phys. 2011, 13, 17177-17184.
(21) Gannon, T. J.; Law, G.; Watson, P. R.; Carmichael, A. J.; Seddon, K. R. First Observation
of Molecular Composition and Orientation at the Surface of a Room-Temperature Ionic Liquid.
Langmuir 1999, 15, 8429-8434.
(22) Law, G.; Watson, P. R.; Carmichael, A. J.; Seddon, K. R. Molecular Composition and
Orientation at the Surface of Room-Temperature Ionic Liquids: Effect of Molecular Structure.
Phys. Chem. Chem. Phys. 2001, 3, 2879-2885.
(23) Bowers, J.; Vergara-Gutierrez, M. C.; Webster, J. R. P. Surface Ordering of Amphiphilic
Ionic Liquids. Langmuir 2004, 20, 309-312.
(24) Tesa-Serrate, M. A.; Marshall, B. C.; Smoll, E. J.; Purcell, S. M.; Costen, M. L.; Slattery, J.
M.; Minton, T. K.; McKendrick, K. G. Ionic Liquid–Vacuum Interfaces Probed by Reactive
Atom Scattering: Influence of Alkyl Chain Length and Anion Volume. J. Phys. Chem. C 2015,
119, 5491-5505.
(25) Wu, B.; Zhang, J.; Minton, T. K.; McKendrick, K. G.; Slattery, J. M.; Yockel, S.; Schatz, G.
C. Scattering Dynamics of Hyperthermal Oxygen Atoms on Ionic Liquid Surfaces:
[emim][NTf2] and [C12mim][NTf2]. J. Phys. Chem. C 2010, 114, 4015-4027.
(26) Tesa-Serrate, M. A.; King, K. L.; Paterson, G.; Costen, M. L.; McKendrick, K. G. Site and
Bond-Specific Dynamics of Reactions at the Gas-Liquid Interface. Phys. Chem. Chem. Phys.
2014, 16, 173-183.
(27) Yockel, S.; Schatz, G., Dynamic QM/MM: A Hybrid Approach to Simulating Gas-Liquid
Interactions. In Multiscale Molecular Methods in Applied Chemistry, Kirchner, B.; Vrabec, J.,
Eds. Springer Berlin Heidelberg: 2012; Vol. 307, pp 43-67.
(28) Radak, B. K.; Yockel, S.; Kim, D.; Schatz, G. C. Modeling Reactive Scattering of F( 2P) at a
Liquid Squalane Interface: A Hybrid QM/MM Molecular Dynamics Study. J. Phys. Chem. A
2009, 113, 7218-7226.
(29) Saecker, M. E.; Govoni, S. T.; Kowalski, D. V.; King, M. E.; Nathanson, G. M. Molecular
Beam Scattering from Liquid Surfaces. Science 1991, 252, 1421-1424.
(30) Saecker, M. E.; Nathanson, G. M. Collisions of Protic and Aprotic Gases with Hydrogen
Bonding and Hydrocarbon Liquids. J. Chem. Phys. 1993, 99, 7056-7075.
(31) King, M. E.; Saecker, M. E.; Nathanson, G. M. The Thermal Roughening of Liquid
Surfaces and Its Effect on Gas–Liquid Collisions. J. Chem. Phys. 1994, 101, 2539-2547.
(32) Nathanson, G. M.; Davidovits, P.; Worsnop, D. R.; Kolb, C. E. Dynamics and Kinetics at
the Gas−Liquid Interface. J. Phys. Chem. 1996, 100, 13007-13020.
24
(33) Garton, D. J.; Minton, T. K.; Alagia, M.; Balucani, N.; Casavecchia, P.; Volpi, G. G.
Reactive Scattering of Ground-State and Electronically Excited Oxygen Atoms on a Liquid
Hydrocarbon Surface. Faraday Discuss. 1997, 108, 387-399.
(34) Garton, D. J.; Minton, T. K.; Alagia, M.; Balucani, N.; Casavecchia, P.; Gualberto Volpi, G.
Comparative Dynamics of Cl(2P) and O(3P) Interactions with a Hydrocarbon Surface. J. Chem.
Phys. 2000, 112, 5975-5984.
(35) Kelso, H.; Köhler, S. P. K.; Henderson, D. A.; McKendrick, K. G. Dynamics of the Gas–
Liquid Interfacial Reaction of O(3P) Atoms with Hydrocarbons. The Journal of Chemical
Physics 2003, 119, 9985-9988.
(36) Köhler, S. P. K.; Allan, M.; Kelso, H.; Henderson, D. A.; McKendrick, K. G. The Effects of
Surface Temperature on the Gas-Liquid Interfacial Reaction Dynamics of O(3P) + Squalane. J.
Chem. Phys. 2005, 122, 024712.
(37) Köhler, S. P. K.; Allan, M.; Costen, M. L.; McKendrick, K. G. Direct Gas−Liquid
Interfacial Dynamics: The Reaction Between O( 3P) and a Liquid Hydrocarbon. J. Phys. Chem. B
2006, 110, 2771-2776.
(38) Zhang, J. M.; Upadhyaya, H. P.; Brunsvold, A. L.; Minton, T. K. Hyperthermal Reactions
of O and O2 With a Hydrocarbon Surface: Direct C-C Bond Breakage by O and H-Atom
Abstraction by O2. J. Phys. Chem. B 2006, 110, 12500-12511.
(39) Allan, M.; Bagot, P. A. J.; Costen, M. L.; McKendrick, K. G. Temperature Dependence of
OH Yield, Translational Energy, and Vibrational Branching in the Reaction of O( 3P) with Liquid
Squalane. J. Phys. Chem. C 2007, 111, 14833-14842.
(40) Bagot, P. A. J.; Waring, C.; Costen, M. L.; McKendrick, K. G. Dynamics of Inelastic
Scattering of OH Radicals from Reactive and Inert Liquid Surfaces. J. Phys. Chem. C 2008, 112,
10868-10877.
(41) Allan, M.; Bagot, P. A. J.; Westacott, R. E.; Costen, M. L.; McKendrick, K. G. Influence of
Molecular and Supramolecular Structure on the Gas−Liquid Interfacial Reactivity of
Hydrocarbon Liquids with O(3P) Atoms. J. Phys. Chem. C 2008, 112, 1524-1532.
(42) Perkins, B. G.; Nesbitt, D. J. Correlated Angular and Quantum State-Resolved CO2
Scattering Dynamics at the Gas−Liquid Interface. J. Phys. Chem. A 2008, 112, 9324-9335.
(43) Roscioli, J. R.; Nesbitt, D. J. State-Resolved Scattering at Room-Temperature Ionic LiquidVacuum Interfaces: Anion Dependence and the Role of Dynamic Versus Equilibrium Effects. J.
Phys. Chem. Lett. 2010, 1, 674-678.
(44) Roscioli, J. R.; Nesbitt, D. J. Quantum State Resolved Scattering from Room-Temperature
Ionic Liquids: The Role of Cation Versus Anion Structure at the Interface. J. Phys. Chem. A
2011, 115, 9764-9773.
(45) Ziemkiewich, M. P.; Zutz, A.; Nesbitt, D. J. Inelastic Scattering of Radicals at the Gas-Ionic
Liquid Interface: Probing Surface Dynamics of [BMIM][Cl], [BMIM][BF4], and [BMIM][Tf2N]
by Rovibronic Scattering of NO (21/2). J. Phys. Chem. C 2012, 116, 14284-14294.
(46) Zutz, A.; Nesbitt, D. J. Nonadiabatic Spin–Orbit Excitation Dynamics in Quantum-StateResolved NO(2Π1/2) Scattering at the Gas–Room Temperature Ionic Liquid Interface. J. Phys.
Chem. C 2015, 119, 8596-8607.
(47) Garton, D. J.; Brunsvold, A. L.; Minton, T. K.; Troya, D.; Maiti, B.; Schatz, G. C.
Experimental and Theoretical Investigations of the Inelastic and Reactive Scattering Dynamics
of O(3P) + D2. J. Phys. Chem. A 2006, 110, 1327-41.
25
(48) Garton, D. J.; Minton, T. K.; Maiti, B.; Troya, D.; Schatz, G. C. A Crossed Molecular
Beams Study of the O(3P)+H2 reaction: Comparison of Excitation Function with Accurate
Quantum Reactive Scattering Calculations. J. Chem. Phys. 2003, 118, 1585-1588.
(49) Yeung, L. Y.; Okumura, M.; Paci, J. T.; Schatz, G. C.; Zhang, J. M.; Minton, T. K.
Hyperthermal O-Atom Exchange Reaction O2 + CO2 Through a CO4 Intermediate. J. Am. Chem.
Soc. 2009, 131, 13940-13942.
(50) Brink, G. O. Electron Bombardment Molecular Beam Detector. Rev. Sci. Instrum. 1966, 37,
857-860.
(51) Daly, N. R. Scintillation Type Mass Spectrometer Ion Detector. Rev. Sci. Instrum. 1960, 31,
264-267.
(52) Alexander, W. A.; Zhang, J.; Murray, V. J.; Nathanson, G. M.; Minton, T. K. Kinematics
and Dynamics of Atomic-Beam Scattering on Liquid and Self-Assembled Monolayer Surfaces.
Faraday Discuss. 2012, 157, 355-374.
(53) Harris, J., Mechanical Energy Transfer in Particle-Surface Collisions. In Dynamics of GasSurface Interaction, Rettner, C.; Ashfold, M. N. R., Eds. Royal Soc. Chem: Cambridge, UK,
1991; Vol. 3.
(54) Zhang, J.; Garton, D. J.; Minton, T. K. Reactive and Inelastic Scattering Dynamics of
Hyperthermal Oxygen Atoms on a Saturated Hydrocarbon Surface. J. Chem. Phys. 2002, 117,
6239-6251.
(55) Minton, T. K.; Giapis, K. P.; Moore, T. Inelastic Scattering Dynamics of Hyperthermal
Fluorine Atoms on a Fluorinated Silicon Surface. J. Phys. Chem. A 1997, 101, 6549-6555.
(56) Ausfelder, F.; McKendrick, K. G. The Dynamics of Reactions of O(3P) Atoms with
Saturated Hydrocarbons and Related Compounds. Prog React Kinet Mec 2000, 25, 299-370.
(57) Lee, Y. T., Reactive Scattering I: Nonoptical Methods. In Atomic and Molecular Beam
Methods, Scoles, G., Ed. Oxford University Press: New York, 1988; Vol. 1, pp 553-568.
(58) Rettner, C. T.; Barker, J. A.; Bethune, D. S. Angular and Velocity Distributions
Characteristic of the Transition Between the Thermal and Structure Regimes of Gas-Surface
Scattering. Phys. Rev. Lett. 1991, 67, 2183-2186.
(59) Kim, D.; Schatz, G. C. Theoretical Investigation of Hyperthermal Reactions at the
Gas−Liquid Interface: O(3P) and Squalane. J. Phys. Chem. A 2007, 111, 5019-5031.
(60) Garton, D. J.; Minton, T. K.; Hu, W.; Schatz, G. C. Experimental and Theoretical
Investigations of the Inelastic and Reactive Scattering Dynamics of O(3P) Collisions with
Ethane. J. Phys. Chem. A 2009, 113, 4722-4738.
(61) Yan, T.; Hase, W. L.; Barker, J. R. Identifying Trapping Desorption in Gas–Surface
Scattering. Chem. Phys. Lett. 2000, 329, 84-91.
(62) Pomerantz, A. E.; Ausfelder, F.; Zare, R. N.; Althorpe, S. C.; Aoiz, F. J.; Bañares, L.;
Castillo, J. F. Disagreement Between Theory and Experiment in the Simplest Chemical
Reaction: Collision Energy Dependent Rotational Distributions for H+D 2 → HD (ν′=3,j′) + D. J.
Chem. Phys. 2004, 120, 3244-3254.
26
Figure 1. Representative translational energy distributions of the velocity-selected
atomic and molecular components of the hyperthermal beam (used for the experiment
with [C12mim][BF4]. The O-atom curve has been normalized to 1, and the area under
each curve reflects the relative flux of the respective species.
27
Time of Flight / s
Figure 2. Representative TOF distributions for products scattered from [C12 mim][BF4] with i = 60°.
The yellow circles are the total distribution which is then partitioned empirically into two
components. The red (TD) curves correspond to the TOFs of atoms or molecules that desorb from
the surface with speeds described by a Maxwell-Boltzmann distribution at the surface temperature of
323K. The blue (IS) curves are the differences between the total signal and the TD curves; as such
they correspond to the TOF of molecules that have not reached thermal equilibrium with the surface.
Time zero is the nominal time at which the atomic beam pulse struck the surface. All offsets,
including ion flight time, have been subtracted.
28
Translational Energy / kJmol
Figure 3. Representative flux-weighted translational energy distributions for IS O, OH, and
H2O scattered from a [C12mim][BF4] surface. The complete IS distribution without the TD
component is shown. These curves have been smoothed because the P(ET) distribution that is
derived from the direct inversion of the TOF data has a significant amount of noise in the low
energy region of the distribution (see text). The area under each curve is proportional to the
flux of the scattered species.
29
O
OH
H2O
Figure 4. Flux angular distributions of O, OH, and H2O, corresponding to θi = 60°. The yellow
circles represent the total signal; the blue circles represent the IS component; and the red circles
represent the TD component. The angular distributions of the thermal components are well
described by a cosine distribution about the surface normal, shown here as black lines.
30
Figure 5. Flux as a function of translational energy and final angle, P(ET,f), for i =
60°. Every panel is derived from 17 individual translational energy distributions, each
collected at a given final angle. The overall distribution is then interpolated between the
individual translational energy distributions. Red indicates maximum flux and dark blue
indicates zero flux. The color scale is different between columns, but it is the same
within each column. The left, center, and right columns show the IS behavior of O, OH,
and H2O, respectively. Each H2O distribution has a sharp maximum at low energies and
relatively low amounts of signal at high energies. Because most of the scattered H2O
flux has relatively low translational energies, the Etrans scale for H2O has a maximum of
100 kJ mol-1 to facilitate viewing of its P(ET,f) distributions.
31
Figure 6. Representative decompositions of impulsive scattering P(ET,f)
distributions into low-energy Plow(ET,f) and high-energy Phigh(ET,f) components,
for an incidence angle of 60°. (A) The high-energy peak in P(ET, f = 70°) (blue)
is well-fit with a log-normal distribution (green) over the majority of its width
(gray box). The residual (orange) defines Plow(ET, f = 70°). (B) Analogous
decomposition of P(ET, f = 50°) distribution (blue). P(ET,f) – Plow(ET, f + 5°) is
shown in red. Phigh(ET,f) is shown in green. Plow(ET,f) is shown in orange.
32
A
0
B
15
30
0
15
30
45
O
45
60
O
60
75
C
0
75
D
15
30
45
O
0
15
30
45
60
75
O
60
75
Figure 7. Flux angular distributions of scattered products from squalane (yellow) and
[Cnmim][BF4] ionic liquids, where θi = 60° and n = 4 (blue), 8 (red), and 12 (green). Panels A
and B show the flux angular distributions of inelastically scattered O atoms and reactively
scattered OH radicals, respectively, that correspond to Phigh(ET,θf). Panels C and D show the
flux-angular distributions of O and OH, respectively, that correspond to Plow(ET,θf). All
distributions have been normalized to their maxima in order to facilitate a comparison of their
shapes.
33
Figure 8. Average fractional energy transfer as a function of deflection angle, χ, for (A)
inelastically scattered O and (B) reactively scattered OH from four different liquids:
[C4 mim][BF4] (blue), [C8mim][BF4] (red), [C12mim][BF4] (green), and squalane (black). These
data were collected with θi = 60°. The circles correspond to the Phigh(ET,θf) component and the
squares correspond to the Plow(ET,θf) component. The colored lines represent the best fit
according to the soft sphere kinematic model. 24 The values of the two parameters from the
model (effective surface mass, ms, and the ratio of internal excitation to incidence energy, Eint/Ei)
are shown. The two parameters are coupled, so subtle differences in energy transfer, as observed
here, cannot be interpreted in terms of these parameters.
34
TOC GRAPHIC
35