Laboratory and model studies of ice nucleation by secondary organic aerosol and desert dust

In this paper we present two main results. The first is a comparison of nucleation efficiencies derived from cold stages and those derived from dynamic cloud chambers, whereas the second finding is that of a new source of ice nuclei in the atmosphere: secondary organic aerosol (SOA).

Cold stages are frequently used as a method to quantify ice nucleation by mineral dust particles. The technique involves preparing drop samples of mixtures containing different concentrations of mineral particles and observing the temperatures at which they freeze on a cold stage. In this contribution we adopt the methodology of using dynamical cloud chamber experiments to investigate ice nucleation on both mineral dust particles and on SOA.

Our results using kaolinite, NX-illite, and k-feldspar are largely in agreement with previous cold stage studies; however, there are some differences which are explained either by sampling statistics within the cold stage experiments or by the asphericity of the mineral dust samples – these differences can be responsible for up to a factor of five difference in ice nucleation efficiency so it is evident that care should be taken when interpreting results from both techniques. Nevertheless these results largely confirm that the chamber techniques and cold stage techniques produce consistent results for the mineral dust particles investigated. A key message is that ice nucleation parameterisations should not be extrapolated to outside of the range that they were derived.

We also present results of ice nucleation by SOA, which were formed via the photochemical oxidation of SOA precursors in a chamber environment. with a mode size of around 100 nm. These experiments demonstrate that the SOA is able to nucleate ice at -20 to -28 degC; thus revealing that there may be a large, as yet unidentified, source of ice nuclei in the atmosphere. The ice nucleation mode appears to be the immersion / condensation – freezing mode since drops are always observed before ice; hence, it is consistent with the current idea that immersion nucleation is a dominant nucleation pathway at these temperatures. This observation is markedly different to recent observations in the literature that have focussed on heterogeneous nucleation of ice by organic acids at low temperature due to the formation of glasses, leading to inhibition of ice formation and postulated elevation of supersaturation.

At present we are unable to extrapolate this result to warmer temperatures, but it is not inconceivable that different oxidation pathways may provide aerosol that do act as ice nuclei at high temperatures. To put our SOA results into context, we estimate the range of ice nuclei present in the atmosphere as a function of temperature and compare it to previous estimates of IN from other sources. Although the nucleation efficiency of SOA is low the total number of SOA particles in the atmosphere is high enough to suggest from our analysis that the number of potential ice nuclei due to SOA is comparable to, and maybe even exceeds, the number of ice nuclei from other sources.

The conference paper will describe the approach taken in the experiments and the key findings. We will also use a cloud model to elucidate impacts on mixed-phase cloud formation and development.