1Technical University of Munich, Institute of Hydrochemistry, Marchioninistr. 17, D-81377 Munich, Germany
2Atmospheric Physics Department, Institute of Physics, St. Petersburg State University, Ulianovskaya 1, 198904 St. Petersburg, Russia
Abstract. The interaction of aerosol particles in the 100–200 nm size range composed of the protein bovine serum albumin (BSA) and the inorganic salts sodium chloride and ammonium nitrate with water vapor at ambient temperature and pressure (25°C, 1 atm) has been investigated by hygroscopicity tandem differential mobility analyzer (H-TDMA) experiments complemented by transmission electron microscopy (TEM) and Köhler theory calculations. BSA was chosen as a well-defined model substance for proteins and other macromolecular compounds, which constitute a large fraction of the water-soluble organic component of air particulate matter.
Pure BSA particles exhibited deliquescence and efflorescence transitions at ~35% relative humidity (RH) and a hygroscopic diameter increase by up to ~10% at 95% RH in good agreement with model calculations based on a simple parameterisation of the osmotic coefficient. Pure NaCl particles were converted from near-cubic to near-spherical or polyhedral shape upon interaction with water vapor at relative humidities below the deliquescence threshold (partial surface dissolution and recrystallisation), and the diameters of pure NH4NO3 particles decreased by up to 10% due to chemical decomposition and evaporation.
Mixed NaCl-BSA and NH4NO3-BSA particles interacting with water vapor exhibited mobility equivalent diameter reductions of up to 20%, depending on particle generation, conditioning, size, and chemical composition (BSA dry mass fraction 10–90%). These observations can be explained by formation of porous agglomerates (envelope void fractions up to 50%) due to ion-protein interactions and electric charge effects on the one hand, and by compaction of the agglomerate structure due to capillary condensation effects on the other. The size of NH4NO3-BSA particles was apparently also influenced by volatilisation of NH4NO3, but not as much as for pure salt particles, i.e. the protein inhibited the decomposition of NH4NO3 or the evaporation of the decomposition products NH3 and HNO3. The efflorescence threshold of NaCl-BSA particles decreased with increasing BSA dry mass fraction, i.e. the protein inhibited the formation of salt crystals and enhanced the stability of supersaturated solution droplets.
The H-TDMA and TEM results indicate that the protein was enriched at the surface of the mixed particles and formed an envelope, which inhibits the access of water vapor to the particle core and leads to kinetic limitations of hygroscopic growth, phase transitions, and microstructural rearrangement processes. Besides these surface and kinetic effects, proteins and comparable organic macromolecules may also influence the thermodynamic properties of the aqueous bulk solution (solubilities, vapor pressures, and chemical equilibria, e.g. for the decomposition and evaporation of NH4NO3.
The observed effects should be taken into account in the analysis of data from laboratory experiments and field measurements and in the modelling of aerosol processes involving water vapor and particles with complex composition. They can strongly influence experimental results, and depending on ambient conditions they may also play a significant role in the atmosphere (deliquescence, efflorescence, and CCN activation of particles). In fact, irregular hygroscopic growth curves similar to the ones observed in this study have recently been reported from H-TDMA experiments with water-soluble organics extracted from real air particulate matter and with humic-like substances.
The Köhler theory calculations performed with different models demonstrate that the hygroscopic growth of particles composed of inorganic salts and proteins can be efficiently described with a simple volume additivity approach, provided that the correct dry solute mass equivalent diameter and composition are known. A simple parameterisation of the osmotic coefficient has been derived from an osmotic pressure virial equation and appears to be well-suited for proteins and comparable substances. It is fully compatible with traditional volume additivity models for salt mixtures, and for its application only the density and molar mass of the substance have to be known or estimated.