Back to overview Show all

Other publication (peer-review)

Publisher ETH , Zurich
DOI 10.3929/ethz-b-000337265

Open Access

Type of Open Access Repository (Green Open Access)


Ice crystals in clouds play an important role in initiating precipitation and act to moderate the Earth’s radiative balance. As such, understanding the mechanisms responsible for ice formation is necessary to quantify the impacts of anthropogenic climate change. Ice crystals are found in both cirrus, composed purely of ice, and mixed-phase clouds, which contain both supercooled cloud droplets and ice crystals. In cirrus clouds, ice crystals can form either homogeneously or heterogeneously, the latter through so called deposition nucleation. In contrast, ice crystals in mixed-phase clouds are primarily observed after the formation of supercooled cloud droplets, indicating that immersion freezing is the responsible ice formation mechanism. Therefore, deposition nucleation and immersion freezing are the two most important heterogeneous freezing pathways in the atmosphere and the focus of this thesis. Deposition nucleation is conventionally described as the formation of ice directly from the vapor phase, without the prior formation of a liquid water phase. However, water is capable of condensing in narrow pores well below water saturation as predicted by the inverse-Kelvin effect. Therefore, it is difficult to rule out the possibility that condensed water in pores is responsible for ice nucleation in conditions below water saturation. In the atmosphere, porous particles are prevalent; they contribute to the largest fraction of airborne dust and are associated with anthropogenic aerosols like soot. As such, understanding the role of pores on ice nucleation is critical for understanding and predicting ice formation globally. To test the role of pores on ice nucleation, we exposed mesoporous silica with well-defined pore diameters ranging from 2.5 to 3.8 nm to varying temperatures and supersaturations with respect to ice in the Zurich Ice Nucleation Chamber. Calculations using Classical Nucleation Theory for deposition nucleation were unable to predict the observed freezing behavior of the mesoporous silica particles. Rather, the freezing behavior was consistent with a pore condensation and freezing mechanism. Furthermore, synthesized particles with an increased concentration of surface hydroxyl groups, as verified by diffuse reflectance infrared Fourier transform spectroscopy, led to a lower onset humidity for ice nucleation. The enhanced ice nucleation ability of silica with the same pore diameter and a higher concentration of surface hydroxyl groups can be reconciled by assuming a lower contact angle and, therefore, a lower relative humidity required for pore condensation and subsequent freezing. Additionally, when the effective contact angle derived from the concentration of hydroxyl groups is low, the ice nucleation ability of the mesoporous silica does not depend on pore size, consistent with the inverse-Kelvin effect for the range of pore sizes tested in this study. However, when the mesoporous silica surface is functionalized with methyl groups, which produces a high contact angle, the dependence on pore size becomes apparent. Meanwhile, at temperatures above the homogeneous nucleation temperature, the addition of both hydroxyl and methyl surface functional groups enhances the ability of the mesoporous silica to nucleate ice heterogeneously depending on the particle type. At conditions above water saturation, immersion freezing occurs when an ice nucleating particle (INP) immersed in a supercooled cloud droplet reaches an activation temperature and triggers ice formation. The INP can become immersed by either acting as a cloud condensation nucleus and initiating the formation of the droplet or through i scavenging. Regardless of how an INP enters a cloud droplet, the characteristics of INP and their concentrations in the ambient atmosphere vary by several orders of magnitude at a given temperature. This variability makes parameterizations for INP concentrations in climate models difficult and can lead to regional biases in cloud lifetime and subsequently in surface temperature. To complement the existing information on INP that act via immersion freezing, a new fully automated drop freezing instrument, DRoplet Ice Nuclei Counter Zürich (DRINCZ) is developed, characterized and validated using a surrogate for atmospherically relevant dust. DRINCZ is shown to be a robust technique for investigating the concentration of INP in both field collected and laboratory samples. Furthermore, the evolution of the INP concentration during a mid- latitude storm system interacting with a high alpine location is examined. Consistent with previous studies, the results indicate that air mass source region and precipitation history can explain the variability in INP concentration observed during the storm.