Potential Sites for Ice Nucleation on Aluminosilicate Clay Minerals and


Potential Sites for Ice Nucleation on Aluminosilicate Clay Minerals and...

1 downloads 156 Views 1MB Size

Perspective pubs.acs.org/JPCL

Potential Sites for Ice Nucleation on Aluminosilicate Clay Minerals and Related Materials Miriam Arak Freedman* Department of Chemistry, The Pennsylvania State University, University Park, Pennsylvania 16802, United States ABSTRACT: Few aerosol particles in clouds nucleate the formation of ice. The surface sites available for nucleus formation, which can include surface defects and functional groups, determine in part the activity of an aerosol particle toward ice formation. Although ice nucleation on particles has been widely studied, exploration of the specific sites at which the initial germ forms has been limited, but is important for predicting the microphysical properties of clouds, which impact climate. This Perspective focuses on what is currently known about surface sites for ice nucleation on aluminosilicate clay minerals, which are commonly found in ice residuals, as well as related materials.

T

he impact of aerosol−cloud interactions on cloud radiative properties is the largest uncertainty in our understanding of the climate system. 1 Clouds in the atmosphere are composed of liquid droplets, ice particles, or a mixture of liquid and ice. It is generally assumed that it is not thermodynamically favorable for ice nucleation to occur by homogeneous nucleation until temperatures are less than approximately −40 °C,2 though the fact that homogeneous nucleation events can occur less efficiently at warmer temperatures affects the microphysical properties of clouds.3 Ice can also form when aerosol particles act as a catalyst for the nucleation of ice. Heterogeneous freezing on aerosol particles can occur below or above homogeneous freezing temperatures. The number of aerosol particles that ultimately form cloud droplets and ice particles affects cloud optical properties, lifetime, and precipitation.2 As with many areas in atmospheric chemistry, understanding aerosol-cloud interactions depends on field, chamber, lab, and theory studies by researchers with backgrounds in chemistry, physics, engineering, and meteorology. Four different hypothetical mechanisms of heterogeneous nucleation have been proposed (Figure 1).4,5 In contact nucleation, the collision of a particle with a supercooled liquid droplet induces the formation of ice. In immersion freezing, a liquid droplet surrounds an insoluble particle, and liquid− surface interactions catalyze the formation of ice. In condensation freezing, a soluble material on a solid particle takes up water prior to the particle freezing. Depending on the system of interest and the experimental method, condensation freezing cannot necessarily be differentiated from immersion freezing or deposition mode nucleation. In deposition mode nucleation, it is experimentally observed that water vapor nucleates ice on a solid particle and bulk liquid water is not present.4 Note that, depending on the method of detection, other processes such as the nucleation of ice in a pore or the freezing of a thin aqueous film may be interpreted as deposition mode nucleation. Deposition mode nucleation can occur at the © XXXX American Chemical Society

Figure 1. Illustration of homogeneous freezing and four heterogeneous freezing mechanisms.

Received: June 22, 2015 Accepted: September 10, 2015

3850

DOI: 10.1021/acs.jpclett.5b01326 J. Phys. Chem. Lett. 2015, 6, 3850−3858

The Journal of Physical Chemistry Letters

Perspective

related to how readily ice can form on a given surface, which depends on the type and density of sites on that surface at which ice can nucleate. Presumably, sites for ice nucleation are composed of functional groups or defect sites on the surface, or a combination of the two. Crystal structure can also aid the formation of ice on a surface. Identifying the sites for ice

lowest supersaturations with respect to ice. Condensation freezing can occur below liquid water saturation and above the relative humidity at which the soluble material deliquesces. Immersion and contact freezing generally require liquid water saturation to maintain the size of the liquid droplet prior to freezing. The types of particles that nucleate ice have a dependence on region, altitude, airmass history, and cloud type. For example, in the tropical upper troposphere, sulfate and sulfate−organic particles dominate.6 At lower altitudes in both the tropics and subtropics, mineral dust is the most common ice residual in cirrus clouds.7 In orographic wave clouds, a variety of residuals are found, including mineral dust, biological particles, playa salt, and biomass burning aerosol.8−12 In arctic stratus clouds, the rare ice particle residuals can include mineral dust, soot from biomass burning, and mineral dust mixed with carbonaceous material or salt/sulfate.13,14 In all of these different cloud types and locations, mineral dust is prevalent with the exception of tropical upper troposphere cirrus. Mineral dust is primarily composed of aluminosilicate clay minerals with additional contributions from quartz, carbonates, oxides, evaporates, mica, and feldspars.15 Aluminosilicate clay minerals compose over half of mineral dust from Africa and Asia.16,17 Many species of aluminosilicate clays have been shown to nucleate ice efficiently by immersion and deposition mode mechanisms.18−20 These minerals are composed of layers of octahedrally coordinated aluminum and tetrahedrally coordinated silicon in different stoichiometric ratios.21 The layers are hydrogen bonded to one another. Hydroxyl groups are located on the aluminum basal plane, interstitial spacings, and edge sites. Isomorphous substitution causes the aluminosilicate layers to be negatively charged, and as a result, cations are adsorbed to the surface and interstitial spacings to balance out the charge.21 One challenge to experimental studies on aluminosilicate clay minerals is that powder samples are rarely pure in composition. Another species of mineral dust of recent interest in the ice nucleation community is feldspars, which weather to form aluminosilicate clay minerals. Although present in smaller quantities in the atmosphere than aluminosilicate clay minerals, certain types of feldspars are efficient at nucleating ice through immersion and deposition mode mechanisms.22,23 Ice nucleation on mineral dust aerosol has been widely studied, and work in this area up to 2012 is reviewed by Hoose and Möhler.20 The species investigated include natural dusts, aluminosilicate clay minerals, and mineral dust mixtures such as Arizona Test Dust. The onset of nucleation spans the whole range of possible supersaturations below −10 °C. Hoose and Möhler note that submicron natural dusts tend to have lower activity toward ice nucleation than other submicron particles.20 Agricultural soils tend to be more active toward ice nucleation due to organic or biological matter.24−26 Inorganic acid coatings on these minerals result in worse nucleation activity, as will be discussed further below. Why are some compounds efficient at nucleating ice and others are not? The concentrations of different types of particulates at relevant altitudes and in different regions are important, but a more efficient heterogeneous material can compensate for a lack of concentration. For example, though lead is not a common naturally occurring component of mineral dust, 67% of ice residuals collected at a mountain-top laboratory in the Swiss Alps contained mineral dust, and 42% of these mineral dust particles contained lead.27 The efficiency is

Identifying the sites for ice nucleation is of fundamental importance for understanding this process. Furthermore, it is an area where physical chemists should be able to contribute greatly to the atmospheric sciences. nucleation is of fundamental importance for understanding this process. Furthermore, it is an area where physical chemists should be able to contribute greatly to the atmospheric sciences. Thus, this Perspective addresses what is known about surface sites for ice nucleation on minerals, focusing primarily on aluminosilicate clay minerals. Surface Sites for Water Adsorption and Ice Nucleation. The molecular-scale mechanism of ice nucleation may depend on the nucleation mode. In deposition mode nucleation, water vapor first adsorbs to a surface, whereas in each of the other nucleation modes, the heterogeneous particle is surrounded by or comes in contact with aqueous solution. Water molecules on the surface of the particle form a nucleus, which then initiates the growth of ice. Pruppacher and Klett have identified five aspects of materials that promote ice nucleation.5 (1) Insolubility: An insoluble heterogeneous material (at the temperature and relative humidity range of interest) is preferred to provide a solid surface on which the ice nucleation can occur. (2) Size: Particles generally need to be larger than 100 nm. (3) Chemical bonding: The ability of a surface to form hydrogen bonds promotes water adsorption and ice nucleation. (4) Crystallography: A surface that has a crystallographic structure close to that of ice will promote ice nucleation. In the case of misfit, dislocations and grain boundaries between regions of crystalline ice and strain in the ice lattice will cause an increase in the interfacial and bulk free energy of the ice. (5) Heterogeneities: Heterogeneities on the surface can provide sites that promote ice nucleation. These heterogeneities can be morphological (steps, cracks, cavities), chemical (hydrophilic contaminant), or electrical. Note that chemical heterogeneities may promote the adsorption of water as in (3), with the distinction that (5) implies that the chemical groups that are promoting the formation of ice are rare contaminants. Which heterogeneities exist on a surface and their role in ice nucleation depends on the composition of the surface and the nucleation mode. As the temperature is lowered, more sites become active toward ice nucleation. These sites may not be distributed evenly among a population of particles, even if they have similar compositions. For simplicity, however, researchers often assume that the number of active sites is proportional to the particle surface area.20,28 The first step of heterogeneous ice nucleation is the interaction of water with the surface. For deposition mode ice nucleation, in particular, water adsorbs onto the surface from the vapor phase. Many studies have been performed on monolayer and multilayer adsorption of water onto single 3851

DOI: 10.1021/acs.jpclett.5b01326 J. Phys. Chem. Lett. 2015, 6, 3850−3858

The Journal of Physical Chemistry Letters

Perspective

crystal mineral surfaces. The use of single crystals allows for precise study of the structure of water on the mineral surface. The one aluminosilicate studied is muscovite, the most abundant dioctahedral mica, and other studies have focused on carbonates and metal oxides. Studies have also been performed on the structure of liquid water at mineral interfaces.29 Muscovite is an aluminosilicate in which there are two layers of tetrahedrally coordinated Si for each layer of octahedrally coordinated Al.21 In the Si layers, Al substitutes for Si to form an Al:Si ratio of 1:3.21 K+ ions balance the negative charge that is caused by these substitutions.21 The hexagonal arrangement of oxygen atoms on the surface of muscovite and its registry with the ice lattice (6% misfit) caused researchers to investigate whether it could be a potentially good material for ice nucleation.30 For studies of water adsorption, muscovite is a convenient surface because it can be cleaved to form large, atomically smooth sheets. The structure of water on muscovite has been studied by a variety of experimental and computational techniques.31−35 Water forms a stable monolayer on muscovite with six water molecules per unit cell, with some of the water molecules adsorbed to the surface and some hydrating the K+ ion.32,34 For thicker water films, K+ can become displaced in the hydration water.34 The RH at which a monolayer is formed is debated, as is the exact nature of the water structure.36 At ambient temperatures and higher relative humidities, water can have a pronounced effect on oxide and carbonate surfaces.37,38 At low exposures, water molecules dissociate upon adsorption to form hydroxyl groups at O vacancies on surfaces such as TiO2, MgO, and CaCO3.39−42 Ambient pressure photoelectron spectroscopy has shown that the first water molecules that do not dissociatively adsorb on TiO2 hydrogen bond strongly to OH and act as the nucleation site for further adsorption.39,43 With long-term exposure to relative humidity, MgO(100) and CaCO3(104) surfaces undergo significant morphological changes due to surface hydrate formation.44−47 MgO forms crystallites of Mg(OH)2.44 Adsorbed water causes the calcite (CaCO3) surface and close subsurface to distort.48 Calcite goes through a series of changes with exposure time from the formation of an amorphous hydrate layer, to the crystallization of vaterite, a polymorph of CaCO3, to crystallization of calcium hexahydrate.44,45 Different regions of the surface have different water contents.44,45 Surface water has a pronounced effect on heterogeneous chemistry of mineral surfaces, as has been recently highlighted in Rubasinghege and Grassian.38 Vibrational sum frequency spectroscopy has been used to sensitively probe the orientation of water molecules at mineral−water interfaces. The structure of water has a strong dependence on pH. In particular, the water dipole flips when the isoelectric point of a surface is crossed, as observed for sapphire (Al2O3), quartz (SiO2), CaF2, and TiO2.49−52 Water molecules are least ordered at the isoelectric point, and most ordered at acidic and basic pH away from the isoelectric point.51 For CaF2, ion exchange reactions result in the formation of Ca−OH at basic pH.50 Variations on this technique can also be used to infer that different surface sites on quartz have different pK values, which affects water organization.53 Geometric constraints are thought to be most important for the origins of water organization on mineral surfaces, rather than water-substrate interactions.29,52

Several density functional theory (DFT) and molecular dynamics (MD) studies have focused on investigating water adsorption and ice nucleation from a molecular perspective on mineral surfaces.54−61 In these studies, the structure of water multilayers and their similarity to the ice structure is investigated. Studies with application to ice nucleation have focused on kaolinite, a common type of aluminosilicate clay mineral. Kaolinite has one layer of Al for every layer of Si (Figure 2). Two different types of basal planes are present: a

Figure 2. Structure of kaolinite showing the Al and Si layers, hydroxylated and nonhydroxylated basal planes, and the edge hydroxyl groups (which are found on the sides of the structure). The kaolinite base structure was used courtesy of C. Ignacio Sainz-Diaz.63

hydroxylated alumina surface and a “hydrophobic” silica surface.21 The edge sites are terminated with OH groups that can become protonated/deprotonated as a function of pH.62 Although early research proposed that the ice-like arrangement of basal plane hydroxyl groups on kaolinite promoted ice nucleation,5 recent computational studies have shown a large degree of misfit between the surface and ice lattices.55−58 DFT studies have shown that water forms a stable layer on the kaolinite surface due to the ability of the surface to accept and donate hydrogen bonds.55 The 2D adsorbed water layer, which corresponds to a water bilayer or equivalently, 2/3 of a monolayer, does not have the structure expected for an initial bilayer of hexagonal ice (Ih) in several respects.55,56 First, the layer forms with H atoms toward the surface or parallel to the surface, providing no dangling OH groups or lone pairs for interlayer H-bonding.56 Second, the 2D layer is flattened relative to what is expected for a bilayer of Ih to compensate for lattice mismatch, which also inhibits interlayer H-bonding.56 In terms of O−O distances, Hu and Michaelides report a mismatch between the kaolinite surface and ice of 3.7−25% from their computations and 0.7−25% from experimental values.56 Water clusters on the surface are not favored, having at best equal stability compared with monolayers, and as a result, multilayers are not expected to form.56 MD simulations have shown that the water adsorption occurs at atmospherically relevant relative humidities (RH) and exhibits collective behavior, meaning that little adsorption is observed prior to monolayer formation and water desorption occurs at a lower RH than water adsorption.57,58 In these simulations, hexagonal rings of water molecules are observed on the Al basal plane, but a misfit strain of 14.0% is calculated compared with Ih.57 Because a strain of 5% is expected to depress nucleation temperatures by 40 °C,64 ice growth is predicted to be inhibited on the hydroxylated basal plane of kaolinite.57 The siloxane basal plane of kaolinite is expected to be inefficient for ice nucleation on kaolinite. Although siloxane has a closer lattice match to Ih, its interactions with water are less 3852

DOI: 10.1021/acs.jpclett.5b01326 J. Phys. Chem. Lett. 2015, 6, 3850−3858

The Journal of Physical Chemistry Letters

Perspective

bellow water saturation and remained full as the RH was lowered to