A burning issue - Soot particles acting as ice cloud seeds

Abstract

The ice phase in tropospheric clouds plays a key role for many atmospheric processes. Ice crystals are important as surfaces for heterogeneous chemical reactions, influence the hydrological cycle through initiation of precipitation formation and determine climate by impacting Earth's energy balance. Ice crystals can form in cirrus clouds at temperatures below 235 K and in mixed-phase clouds (MPCs), where the temperature (T) range is between 235-273 K and supercooled cloud droplets and ice crystals can co-exists in a thermodynamically metastable state. The initial formation of an ice crystal is usually catalyzed by aerosols that can act as ice nucleating particles (INPs), through lowering of the thermodynamic energy barrier that is associated with forming the ice phase. Despite significant advances in understanding atmospheric ice formation over the past decades, representation of the ice phase in global climate models remains associated with major uncertainties in terms of projected radiative forcing. This mainly results from a lack of understanding the aerosol-cloud interactions and their feedback processes, such as impacts on cloud phase, lifetime and albedo effects. While some aerosol species, such as mineral dust, have been identified as important INPs and environmental conditions required for dust particles to nucleate ice seem to be well understood, research on the ice nucleation activity of soot particles remains inconclusive. This despite soot being considered the second most important radiative forcing agent and constituting a major anthropogenic pollutant from fossil fuel and biomass combustion. The contradictory results of the ice formation ability of soot are mainly based on an insufficient understanding of the relationship between the physicochemical particle properties and ice nucleation activity and mechanism, respectively. In addition, the link of the particle properties to how these can change upon atmospheric aging of the soot particles is largely unknown. As such, understanding of these processes is required to quantify the impacts of anthropogenic climate change. The goal of this thesis is to improve the understanding of soot ice nucleation through an experimental investigation of the cloud formation abilities of different combustion particles in a controlled laboratory setup. Laboratory measurements were performed on six different soot types. The ice nucleation activity was studied using a continuous flow diffusion chamber (CFDC) setup over a temperature range from 218-253 K, covering the MPC and cirrus regime, as a function of relative humidity (RH) and soot particle size. Particle physicochemial properties were characterized by a suite of auxiliary measurements, including water vapor sorption and thermogravimetric analysis. A strong dependence on the temperature regime was found, with ice nucleation being absent for T > 235 K, where only supercooled liquid cloud droplets formed on the soot particles. At cirrus temperatures, a dependence of ice nucleation on the particle size was observed, with larger particles showing ice nucleation activity at lower ice supersaturations at a given T. The ice nucleation was strongly linked to the particle properties, in particular the wettability of the soot. This, along with the dependence on the temperature regime revealed a pore condensation and freezing (PCF) mechanism to be the responsible ice nucleation mechanism. The impact of atmospheric aging on the ice nucleation activity of soot was investigated through two independent studies, addressing the effect of cloud processing and aging of particles in acidic aqueous solutions, respectively. To test the role of cloud processing on soot ice nucleation activity, a new experimental platform was developed, where two CFDCs are coupled in series. The setup allows to mimic different cloud processing scenarios in the first CFDC and subsequently test the ice nucleation activity of the cloud processed aerosol particles within the second CFDC. Upon cloud processing, the soot properties changed with processed particles showing enhanced cloud formation potential at T < 233 K. In particular, the cloud processed particles were found to show a change in wettability and particle morphology, with processed soot aggregates forming more compact clusters with increased total pore volume and wettability. An increased ice nucleation activity was found, independent of the cloud processing scenario. Furthermore, the formation of a cloud hydrometeor was identified to be the key factor for enhancing the ice nucleation activity. Overall the better ice nucleation ability was attributed to an enhanced ice formation via PCF, resulting from the changes in particle properties upon cloud processing. Moreover, the effect on soot ice nucleation by aging of soot particles in acidic aqueous solutions, mimicking aging in cloud and haze droplets was investigated. Aging was achieved through exposure of the soots to slightly acidic and aqueous solutions containing sulfuric acid and pure water only. Aged soot samples showed a significantly enhanced ice nucleation activity, which could be linked to distinct chemical functionalities present on the particles by characterization using near edge X-ray fine structure spectroscopy. Finally, to further the understanding on cloud phase fractionation, a new instrument, the high speed particle phase discriminator (PPD-HS) was tested and characterized. PPD-HS is shown to accurately discriminate the shape of spherical cloud droplets and aspherical ice crystals based on symmetry analysis of the spatial intensity distribution of near-forward scattered light. The lower size limit for particle shape discrimination using PPD-HS was experimentally found to be approximately 3 micrometer, with detection rates of a few hundred particles per second, thus showing enhanced capabilities to previous devices. Deployment of PPD-HS within a CFDC setup was successfully tested and provides an alternative to the previous deployment of a particle size threshold for discriminating cloud droplets and ice crystals at MPC conditions. Hence, phase discrimination can be achieved largely independent of particle size when using of PPD-HS.