Abstract. Aluminosilicates (such as feldspars, clay minerals and micas) and quartz (a crystalline form of silica), constitute the majority of airborne mineral dust. Despite similarities in structures and surfaces they differ greatly in terms of their ice nucleation (IN) efficiency. Here, we show that determining factors for their IN activity include surface ion exchange, NH3 or NH4+ adsorption, as well as surface degradation due to the slow dissolution of the minerals. We performed immersion freezing experiments with the (Na-Ca)-feldspar andesine, the K-feldspar sanidine, the clay mineral kaolinite, the micas muscovite and biotite, and gibbsite (Al(OH)3, a mineral form of aluminum hydroxide) and compare their IN efficiencies with those of the previously characterized K-feldspar microcline and quartz. Samples were suspended in pure water as well as in aqueous solutions of NH3, (NH4)2SO4, NH4Cl and Na2SO4, with solute concentrations corresponding to water activities aw = 0.88–1.0. Using differential scanning calorimetry (DSC) on emulsified micron-sized droplets, we derived onset temperatures of heterogeneous (Thet) and homogeneous (Thom) freezing as well as heterogeneously frozen water volume fractions (Fhet). Suspensions in pure water of andesine, sanidine and kaolinite yield Thet = 242.8 K, 241.2 K and 240.3 K, respectively, while no discernable heterogeneous freezing signal is present in case of the micas or gibbsite (i.e. Thet ≈ Thom ≈ 237.0 K). The presence of NH3 and/or NH4+-salts as solutes has distinct effects on the IN efficiency of most of the investigated minerals. When feldspars and kaolinite are suspended in very dilute solutions of NH3 or NH4+-salts ( ~ 0.99), Thet shifts to higher temperatures (by 2.6–7.0 K compared to the pure water suspension). Even micas and gibbsite develop weak heterogeneous freezing activities in ammonia solutions. Conversely, suspensions containing Na2SO4 cause Thet of feldspars to clearly fall below the water-activity-based immersion freezing description (Δaw = const) even in very dilute Na2SO4 solutions, while Thet of kaolinite follows the Δaw = const curve. The water activity determines how the freezing temperature is affected by solute concentration alone, i.e. if the surface properties of the ice nucleating particles are not affected by the solute. Therefore, the complex behavior of the IN activities can only be explained in terms of solute-surface-specific processes. We suggest that the immediate exchange of the native cations (K+/Na+/Ca2+) with protons, when feldspars are immersed in water, is a prerequisite for their high IN efficiency. On the other hand, excess cations from dissolved alkali salts prevent surface protonation, thus explaining the decreased IN activity in such solutions. In kaolinite, the lack of exchangeable cations in the crystal lattice explains why the IN activity is insensitive to the presence of alkali salts (Δaw = const). We hypothesize that adsorption of NH3 and NH4+ on the feldspar surface rather than ion exchange is the main reason for the anomalous increased Thet in dilute solutions of NH3 or NH4+-salts. This is supported by the response of kaolinite to NH3 or NH4+, despite lacking exchangeable ions. Finally, on longer timescales (hours to days) the dissolution of the feldspars in water or solutions becomes an important process leading to depletion of Al and formation of an amorphous layer enriched in Si within less than an hour. This hampers the IN activity of andesine the most, followed by sanidine, then eventually microcline, the least soluble feldspar.
Contribution of Ion-Exchange and Non-Ion-Exchange Reactions to Sorption of Ammonium Ions by Natural and Activated Aluminosilicate Sorbent
