Abstract
Experiments conducted at the Aerosol Interactions and Dynamics in the Atmosphere (AIDA) chamber located in Karlsruhe, Germany permit investigation of particle properties that affect the nucleation of ice at temperature and water vapor conditions relevant to cloud microphysics and climate issues. Ice clouds were generated by heterogeneous nucleation of Arizona test dust (ATD), illite, and hematite and homogeneous nucleation of sulfuric acid. Ice crystals formed in the chamber were inertially separated from unactivated, or “interstitial” aerosol particles with a pumped counterflow virtual impactor (PCVI), then evaporated. The ice residue (i.e., the aerosol which initiated ice nucleation plus any material which was scavenged from the gas- and/or particle-phase), was chemically characterized at the single particle level using a laser ionization mass spectrometer. In this manner the species that first nucleated ice could be identified out of a mixed aerosol population in the chamber. Bare mineral dust particles were more effective ice nuclei (IN) than similar particles with a coating. Metallic particles from contamination in the chamber initiated ice nucleation before other species but there were few enough that they did not compromise the experiments. Nitrate, sulfate, and organics were often detected on particles and ice residue, evidently from scavenging of trace gas-phase species in the chamber. Hematite was a more effective ice nucleus than illite. Ice residue was frequently larger than unactivated test aerosol due to the formation of aggregates due to scavenging, condensation of contaminant gases, and the predominance of larger aerosol in nucleation.
Acknowledgments
Funding of this work was provided by ACCENT, ETH Zürich, NOAA base funding, PECASE, PNNL LDRD, and the European Union Integrated Project SCOUT-O3. We wish to thank the AIDA team, the IN08 participants, and David Thomson for their assistance in preparing these experiments. We also thank Joseph Ulanowski for providing SID-2 data.
Notes
a OPC concentrations are not corrected for wall loss or size biases.
b Relative to total OPC concentration.