The ice nucleation behavior of mineral dusts: A comparison of different pure and surface modified dusts

Ice containing clouds cover 40% of the Earth surface at any time and play an important role in both the Earth's radiation budget and the formation of precipitation. In other words, ice containing clouds are important for both climate and weather. The initial step of ice formation in clouds is ice nucleation which takes place either homogenously or heterogeneously, where heterogeneous freezing requires the presence of a foreign ice nucleus (IN), e.g. a mineral dust particle, which induces the ice formation. Heterogeneous ice nucleation occurs at temperatures higher than those observed for homogeneous ice nucleation. Different modes of heterogeneous ice nucleation, i.e., deposition nucleation, condensation freezing, immersion freezing, and contact freezing do exist. Here we will deal specifically with the process of immersion freezing, which is of special interest in mixed phase clouds (Ansmann et al., 2009) influencing both, their radiative properties and precipitation behavior.

In this paper, we will present results from measurements concerning the immersion freezing behavior of different dust particles. Measurements were carried out with LACIS (Leipzig Aerosol Cloud Interaction Simulator, Hartmann et al., 2011), examining size segregated particles consisting of either Arizona Test Dust (ATD), illite (illite-NX, Arginotec), one kind of a Potassium-rich feldspar (containing roughly 80% microcline, from Minas Gerais, Brazil), or kaolinite (where results from a sample from Fluka and a sample from the Clay Mineral Society (CMS, KGa-1b) are available). Results reported here were mostly determined for particles with a mobility diameter of 300 nm. Coatings with sulfuric acid with thicknesses of a new nanometer were applied to the dust particles to examine the respective effect on the particles' ice nucleation ability.

In general, it was found that for uncoated particles the ice nucleation ability of the majority of the dusts was very similar. An outstanding exception was feldspar which induced freezing at much higher temperatures than observed for all other dusts, and CMS kaolinite which was clearly the least ice active. In those cases where different sizes of particles were examined, a dependence of the ice nucleation ability on the particle surface was found.

Coating with sulfuric acid usually reduced the particles ability to act as IN in the immersion mode, although a freezing point depression can be neglected as the particles examined in LACIS are immersed in strongly diluted droplets of a few micrometers in diameter. The most dramatic reduction was observed for the feldspar sample. The other extreme were CMS kaolinite particles for which no reduction was found. Interestingly, the CMS kaolinite particles nucleated ice similar to the other particles when they were coated with sulfuric acid. This is indicative for the other particles containing a particular substance and/or possessing a type of ice nucleating surface feature not present on the CMS kaolinite, which can be destroyed by reaction with sulfuric acid.

In this context the amount of K-feldspar contained in the different dusts could be a possible explanation. It was recently proposed that the ability of mineral dusts to act as IN might originate in K-feldspar (Atkinson et al., 2013), and according to Atkinson et al. (2013), the K-feldspar amounts to 20%, 8% and 5% for ATD, illite-NX and Fluka kaolinite, respectively, while our feldspar sample was mostly composed of it (to about 80%), and the CMS kaolinite does not contain any detectable amount. In other words, according to our results, the amount of K-feldspar present in the different dust samples roughly correlates with the IN activity of the dusts in immersion mode, thereby corroborating the findings of Atkinson et al. (2013). However, we also found clear evidence that it's not K-feldspar in general but rather microcline, one of the end-members of the K-feldspar group, having a triclinic crystal structure, which may be controlling the ice nucleation behavior of mineral dusts.

Additional measurements performed for feldspar particles of different sizes enabled us to quantify the ice nucleation behavior of single ice active sites on the particles. The respective results will be reported in terms of both, ice active site surface densities and contact angle distribution parameters, i.e., parameters/models that can be utilized in atmospheric models.

Acknowledgement: This work is partly funded by the German Research Foundation DFG within the framework of the Ice Nucleation research UnIT FOR 1525 (INUIT), project number WE 4722/1-1.

References: Ansmann, A. et al. (2009), Evolution of the ice phase in tropical altocumulus: SAMUM lidar observations over Cape Verde, J. Geophys. Res., 114 (D17208). Atkinson, J. D. et al. (2013), The importance of feldspar for ice nucleation by mineral dust in mixed-phase clouds, Nature, 498, 355-358. Hartmann, S. et al. (2011), Homogeneous and heterogeneous ice nucleation at LACIS: Operating principle and theoretical studies, Atmos. Chem. Phys., 11, 1753–1767.