Uptake on squalane surfaces

From the characterization of the properties of the incident beams discussed in section 4.3, the centre-of-mass collision energies for the Ne and He carrier were estimated to be 6.10 ± 0.05 and 25.0 ± 0.2 kJ mol-1_{, respectively. A comparison of those values with} the estimated E0 for different C-H bonds in table Table 4.1 reveals that the lower Ecoll

only modestly exceeds the estimated E0 values for secondary and tertiary C-H bonds but

is barely above threshold for primary C-H bonds. Molecular dynamics simulations done for squalane [95] show that the primary C-H sites are expected to occupy circa 41% of the surface. Therefore, at the lower collision energy, almost two fifths of the exposed squalane should be unreactive to OD. Further investigation of the excitation function shown in Figure 4.1 shows that these sites, as well as the sites of the secondary and tertiary C-H bonds should become substantially more reactive at the higher Ecoll. A

three C-H bonds suggest that an increase of the collision energy between Ne and He carrier experiments would lead to an expected increase in the uptake of OD (and corresponding decrease in the survival probability) by more than a factor of 2.

Surprisingly, no such decrease is observed in the measured survival probabilities as a function of collision energy in Table 4.6. Instead, the survival coefficients are essentially the same at around 65% for both collision energies, and even for the higher collision energy accessed via the ‘bulk photolysis’ experiments [10-11], within experimental uncertainty. There are undoubtedly some detailed differences in incident angular distributions and possibly in the sampling of the product angular distributions between the current set-up and the previous photolytic experiments that affect the measurement of the survival probability. Some caution should also be exercised when comparing experiments with different incident rotational distributions; these vary subtly between the He and Ne carriers, as characterized in section 4.3, but more significantly in the photolytic experiments. Any rotational-level dependence of the uptake coefficients could therefore, in principle, independently affect the overall uptake. Nevertheless, regardless of the possible influence of these effects, the result in the photolytic experiments continues the essentially flat trend of the lower two collision energies, and so an alternative explanation to account for this trend must be sought.

One interesting possibility is that the expected reduced reactivity in direct, IS-like trajectories at lower <Ek> is compensated by an increase in the proportion that undergo TD-like accommodation at the squalane surface. If these accommodated molecules are able to migrate until they encounter a more reactive secondary or especially tertiary C -H site, then their reaction probability would be enhanced. This explanation would be consistent with the dynamical results from this experiment, suggesting that those molecules that do escape the squalane surface retain some of the characteristics of IS trajectories even at lower collision energies, including a positive correlation between translational and rotational energies. The effect would be even more amplified in reaction at the squalene surface at the higher collision energies, where OH molecules which have been accommodated at the surface have a strongly enhanced probability of undergoing an addition reaction at a double-bond site. Although the differences are modest relative to the experimental uncertainties, it is interesting that the observed OD or OH rotational temperatures (reported in Table 4.4) from scattering from squalene at <Ek> = 29.5 kJ mol-1 _{or previously of 54 kJ mol}-1 _{are consistently higher than those}

from squalane and similar to those from PFPE, despite its known ‘stiffer’ surface. This would be compatible with only the most direct, impulsively scattered OH molecules escaping from squalene, whereas this constraint is weaker in squalane.

These results show that, even for surfaces of a known composition, both the absolute values and the temperature dependence of OH uptake may not simply reflect the behaviour that would be expected based on extrapolation from the corresponding gas-phase reactions. In the case of squalane, the uptake coefficient remains almost constant (at a value of ~0.35) as the collision energy is reduced to more nearly thermal collision energies. Though the explanation for this result may be as yet unknown, it at least provide some reassurance to groups invoking the previous uptake on squalane results in an atmospheric context [3, 17, 26, 54, 80, 211-216] that their conclusions are unlikely to be affected. The uptake coefficients for both squalane and squalene have also demonstrated that the collision-energy dependence is strongly affected by the chemical nature of the surface. This implies that the corresponding temperature dependence of uptake in the atmosphere is almost certainly a non-trivial function of the initial and evolving composition of the aerosol surface during ageing, which has potentially important consequences for the understanding and modelling of such reactions.

4.6 Summary Points

1) OD radicals, generated from the new molecular beam and electric discharge source, were scattered from the surfaces of squalane, squalene and PFPE. The radicals were then detected by LIF using the single-point detection set-up described in section 2.7.1

2) Helium and neon were chosen as the carrier gases for the molecular beam. From a detailed characterization of the radicals in the incident beam, it was determined that

<v> = 1811 ± 8 m s-1_{, <Ek> = 29.5 ± 0.3 kJ mol}-1 _{and that OD radicals were populating} 45% of N = 1 and 28% of N = 2 when using the He carrier, whereas for the Ne carrier it was determined that <v> = 894 ± 8 m s-1_{, <Ek> = 7.20 ± 0.06 kJ mol}-1_{, and that the} relative populations in N = 1 and N = 2 were 48% and 18%, respectively.

3) At the higher collision energy, the OD radicals scattered off the liquid surfaces have clear superthermal translational energies and rotational distributions, indicating the presence of a significant component of IS-like scattering. There is also some evidence of IS-like scattering present from scattering at the lower collision energy, however, its contribution with respect to the TD scattered molecules could not be unambiguously distinguished with the current data.

4) The experimental set-up obtains direct measurements of the survival probability of OD off a liquid surface. This can be related to uptake coefficient values corresponding to single reactions of OD with the surface, provided that no other OD loss mechanisms are present, and that all OD detected scatters off the inert PFPE surface. Both assumptions were found to be valid within the context of this experimental set-up.

5) For squalene, the results show that addition reactions at the vinyl sites on the surface predominate, and that such reactions are negatively activated, with the uptake coefficient decreasing with increasing collision energy.

6) Surprisingly, the results for squalane show that the uptake coefficient is 0.35 regardless of the collision energy used, within statistical uncertainties. A possible explanation for those results suggests that the expected reduction in OD reactivity from IS-type scattered molecules off primary C-H sites at lower <Ek> is compensated by thermally accommodated radicals migrating on the surface until they encounter a more reactive secondary or tertiary C-H bond.