Amorphous and crystalline aerosol particles interacting with water vaporconceptual framework and experimental evidence for restructuring, phase transitions and kinetic limitations

Picture a particle one hundred nanometers wide — smaller than a wavelength of visible light — floating in the air above a city. You might assume it's a tiny crystal, or maybe a liquid droplet. Mikhailov and colleagues would tell you that assumption is almost certainly wrong. For a large fraction of organic aerosol particles, the truth is stranger: they are glasses. Not the glass in a window, but something responsive to its environment — a material that can slowly absorb water, swell like a sponge, restructure itself, and then collapse back. That distinction, between glassy and crystalline, turns out to matter enormously for clouds, climate, and atmospheric chemistry. The two thousand nine study from Mikhailov, Vlasenko, Martin, Koop, and Pöschl set out to build a conceptual framework for how amorphous particles interact with water vapor, and to test it experimentally. Their instrument of choice was the hygroscopicity tandem differential mobility analyzer, or H-TDMA — a device that selects particles of a known dry size, exposes them to a controlled relative humidity, and measures how much they've grown or shrunk. They ran experiments across relative humidities from about five to ninety-five percent at room temperature, using three carefully chosen substances: crystalline ammonium sulfate as a reference, and two organics that form amorphous particles when spray-dried — oxalic acid and levoglucosan, a sugar derivative produced by burning wood. Start with the crystal, because that's the baseline everyone assumes. Ammonium sulfate behaves exactly as classical thermodynamics predicts. Humidify the air up to about eighty percent relative humidity, and the particle sits there, unchanged — a dry solid. Cross that threshold, and it abruptly dissolves into a saturated droplet. That jump is called deliquescence. Now dry the air back down. The droplet doesn't re-crystallize at eighty percent. It remains liquid all the way down to about thirty percent relative humidity before it snaps back into a solid. This gap between deliquescence at eighty percent and efflorescence at thirty percent is the hysteresis loop, and for crystalline salts, it's well understood and well modeled. The H-TDMA confirmed all of it with sub-percent diameter precision. Below the deliquescence threshold, the team could even resolve about two monomolecular layers of water adsorbed on the crystal surface between fifty and seventy percent relative humidity — a small but measurable signal. This is the world that atmospheric models were built to describe. Then comes oxalic acid, and everything changes. Spray-dry an oxalic acid solution and you don't get a compact crystal. You get a highly porous structure — something the authors describe as a xerogel — with apparent void volume fractions near forty percent. These voids fill with water far earlier than any crystalline threshold would predict. Substantial bulk water absorption began at roughly forty percent relative humidity, even though the crystalline deliquescence for oxalic acid sits near ninety-seven percent. The particle didn't wait for a sharp threshold. It started soaking up water gradually, with stepwise diameter changes — notably a step near forty-three percent relative humidity — consistent with the formation and swelling of a hydrogen-bonded, gel-like network. Full dissolution only came above about eighty percent relative humidity, when absorbed water occupied roughly fifty percent of the particle's volume. The particle had been slowly drinking for forty percentage points of relative humidity before it finally became a liquid droplet. Levoglucosan tells a similar story with its own texture. Unlike the porous oxalic acid xerogel, spray-dried levoglucosan forms compact spheres. But compact doesn't mean inert. Bulk water absorption began at around thirty percent relative humidity — again, well below the crystalline deliquescence at eighty-two percent. Between thirty and sixty percent relative humidity, the growth curve became steep, broadened, and irregular: a signature the paper attributes to an ultra-viscous, rubbery or gel-like matrix slowly swelling as water diffuses in. At about thirty percent relative humidity, the rate of water layer uptake increased by roughly a factor of five — a signal that a moisture-induced glass transition was unlocking the interior. Full deliquescence came above sixty percent relative humidity, when absorbed water exceeded about thirty percent of the dry particle volume, giving a growth factor of around 1.12. For this fully diluted state, the hygroscopicity parameter kappa came out at approximately 0.23, with an effective van't Hoff factor — a measure of how many effective particles a solute molecule produces in solution — of about 1.3. Earlier studies had placed levoglucosan's kappa between 0.17 and 0.21. The higher values reflect the more thoroughly dried starting conditions in these experiments, suggesting that previous studies may have systematically underestimated levoglucosan's hygroscopicity. This is where the classification Mikhailov and colleagues propose becomes useful. They organize amorphous aerosol phases into four types: glassy, rubbery, gel-like, and ultra-viscous liquid. The dividing lines are set by viscosity and molecular diffusivity. In a glassy phase, molecular motion is nearly frozen — diffusivity is so low that water molecules can barely move through the material. Add a little water, and you cross a glass transition into a rubbery state, where the material is flexible but still semi-solid. Add more, and gel-like networks can form — swollen, cross-linked structures that hold water in a supramolecular web. Add still more, and you reach a viscous liquid that eventually behaves like any concentrated aqueous solution. Each of these states has a different relationship with water vapor, and none of them behaves like the crystalline case. The kinetic implications follow directly, and they are stark. If a particle's outer layer is glassy or ultra-viscous, water molecules cannot penetrate or escape quickly enough to track the surrounding humidity. In the laboratory, this showed up as persistent water retention: levoglucosan particles at relative humidities below twenty percent still held water volume fractions above ten percent. They simply couldn't let it go fast enough. In the atmosphere, the consequence is more serious. Köhler theory — the standard framework for predicting at what supersaturation a particle will activate into a cloud droplet — assumes the particle is in thermodynamic equilibrium with the surrounding vapor at every moment. If the particle is glassy, that assumption fails. A particle that theory says should activate into a cloud condensation nucleus at a given supersaturation might not, because water can't get in fast enough. The same logic applies to ice nucleation. Mikhailov and colleagues are explicit: these kinetic limitations may influence the activation of particles as cloud condensation nuclei and ice nuclei in field measurements, laboratory studies, and in the actual atmosphere. The team didn't just measure; they modeled. Using Köhler theory in both its full form and a simplified volume-additive formulation built around the single kappa parameter, they extracted precise hygroscopicity values and cross-checked them against independent thermodynamic databases. For ammonium sulfate, kappa came out at 0.51 with a van't Hoff factor of 2.11 — and the model agreed with measurements to better than one percent across nearly the full humidity range. For oxalic acid dihydrate, kappa was 0.23; for levoglucosan, also 0.23. But for the organics, standard Köhler assumptions had to be extended. The authors applied Flory-Huggins polymer theory — a framework for describing how polymer networks swell in solvent — with fitted interaction parameters of 0.46 for oxalic acid and 0.26 for levoglucosan. These values better captured the gradual, gel-like swelling behavior. They also proposed generalized definitions of deliquescence and efflorescence broad enough to encompass amorphous transitions: not just the sharp crystalline phase change, but the gradual onset of bulk water uptake at ten to twenty percent absorbed water volume, and full liquefaction at thirty to fifty percent. The broader point is this. Levoglucosan is a tracer for biomass burning. Oxalic acid is one of the most abundant organic acids in the atmosphere. If particles like these — and the many other carboxylic acids, carbohydrates, and proteins that tend to form amorphous rather than crystalline phases when dried — are actually glasses and gels rather than crystals or simple liquids, then the models built on crystalline equilibrium assumptions are systematically misrepresenting how those particles grow, how they seed clouds, and how they react with atmospheric oxidants. The semi-solid shells can slow not just water uptake but also the uptake of gaseous photo-oxidants, slowing atmospheric aging in ways that equilibrium chemistry cannot capture. Picture those glassy particles again — drifting through clouds, absorbing and releasing water on their own schedule, indifferent to our equilibrium assumptions. The framework Mikhailov and colleagues built is an invitation to take that image seriously, and to rebuild the models around what the particles actually do. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.