Secondary organic aerosol formation in cloud droplets and aqueous particles (aqSOA)a review of laboratory, field and model studies

If you've followed the story of secondary organic aerosol, or SOA for short, you've probably seen the same plot I have. We take volatile organics in the gas phase, let them oxidize, and then ask how much of that stuff partitions into particles. It works up to a point. But then the real atmosphere pushes back. The oxidation state is off. The size distribution isn't right. The vertical profile doesn't line up. And the light-absorbing parts of the aerosol are underpredicted. Ervens, Turpin, and Weber pulled those threads together in 2011 and said, in essence: we're missing a big piece. That missing piece is water. Not just water vapor, but liquid water—inside cloud droplets and inside the tiny films of water that coat hygroscopic particles. Chemistry in that condensed phase, what we'll call aqueous SOA, or aqSOA, makes products the gas phase doesn't: dicarboxylic acids like oxalate and large, sticky oligomers often grouped under high molecular weight carbon. These compounds don't need to have gas phase parents to matter, and under humid, photochemically active conditions, aqSOA can rival the mass you'd expect from the traditional gas-to-particle route. The logic is simple. Globally, the amount of liquid water associated with particles is on the same order as, and often larger than, the dry particle mass. Put numbers on it and you're looking at aerosol liquid water from a few to a few dozen micrograms per cubic meter, while cloud water is a thousand times more concentrated, roughly a tenth to a gram per cubic meter. And fresh, hydrophobic soot doesn't stay that way—within tens of kilometers it picks up coatings, turns hygroscopic, and carries a little pool of water with it. That's real estate for chemistry. So what actually happens in that water? Two things, and they pull in different directions. First, there's dark chemistry—hydration of dissolved carbonyls like glyoxal and methylglyoxal, followed by reversible oligomer formation. Hydration is easy thermodynamically but only trudges along kinetically; measured rates are slower than a hundredth per second. Glyoxal's oligomerization in dilute solution is similarly unhurried, about a hundred per molar per second. In a cloud droplet, those reactions tend to make products that leave when the droplet evaporates. In a more concentrated particle, where the water is salty and the organics are closer together, those same steps can build larger oligomers that stay behind and add to mass. The second pathway is photochemical, and it's the workhorse. Dissolved organics meet hydroxyl radicals in water, and the reaction is almost diffusion-limited. Think rates around a billion to ten billion per molar per second. Even with hydroxyl concentrations as low as about ten to the minus thirteen molar in cloud water and ten times higher in aerosol water, you end up with effective first-order rates that can turn over a meaningful fraction of the dissolved organics on atmospheric timescales. That means you get small acids like oxalic and glyoxylic acids quickly, and as concentrations climb you see repeating chemical motifs in the products—regular seventy-two dalton jumps that flag oligomers—signaling that high molecular weight material is on the rise. The same pattern shows up when isoprene's water-soluble oxidation products, MACR and MVK, see hydroxyl in water: a blend of small acids and heavier intermediates that push mass into the particle phase. And the inorganic backdrop matters. In sulfate-rich solutions, light and hydroxyl make sulfate radicals, which are plenty reactive themselves. Glyoxal reacts with those radicals with a rate constant on the order of a few hundred thousand per molar per second, and it gets faster as the solution turns more acidic. Ammonium changes the game too. Glyoxal's ammonium-catalyzed pathways follow an empirical law that grows exponentially with ammonium activity and with pH. You don't need the exact exponents to catch the drift—add more ammonium or raise the pH and those dark reactions take off. In evaporating droplets, ammonium and amines help build imidazoles and related aromatic-like structures. Those reactions run slower than hydroxyl chemistry—effective rates around a millionth to a thousandth per second—but they keep working when light is scarce and concentrations are high. What drops out the other side doesn't look like classic gas-derived SOA. Aqueous products are highly oxygenated. Just hydrating glyoxal puts its oxygen-to-carbon ratio around one to two, and oxalate lives at two. Dark, ammonium-catalyzed oligomers tend to be a bit less oxygen-rich, but still high compared to gas phase condensable material. Hygroscopicity—the tendency to pick up water—tracks that oxygenation. AqSOA's hygroscopicity parameter often sits between about zero point one five and zero point six, and you can see growth factors around one point four at ninety percent relative humidity for MACR-derived material. Some of the products surf the surface, literally: oligomeric films can knock a solution's surface tension down by twenty to thirty-five percent, which changes how droplets activate into cloud condensation nuclei. And some aqSOA is yellow-brown. Glyoxal in ammonium solutions starts absorbing in the deep ultraviolet and gradually reddens toward three hundred to four hundred nanometers. That's brown carbon building in a beaker—and, as field data suggest, in the sky. Let's go outdoors. If aqueous processing is important, you should see its fingerprints in time and in composition. In Atlanta, during the Southern Oxidants Study, particulate water-soluble organic carbon rose and fell with particle water and with nitrate, peaking in the humid morning. Mexico City during MILAGRO told a similar story: right after sunrise, as relative humidity climbs and photochemistry wakes up, water-soluble organic carbon and nitrate jumped together, then a second, smaller bump in the afternoon tracked more traditional gas-to-particle SOA. Here's the kicker—only about a third of that fresh water-soluble organic fraction looked volatile. Nitrate was far more volatile. That growing ratio of organics to nitrate as the day wore on points to irreversible mass built in aerosol water. And if you track oxygenation with an aerosol mass spectrometer, you see the oxygen-to-carbon ratio rise with humidity, often peaking around noon when both light and water are abundant. The flip side is just as informative. In Los Angeles, the fraction of organics in the particle phase didn't care much about humidity but did track total organic carbon, a sign that gas phase partitioning dominated. Meanwhile, the humid, biogenic Southeast showed higher oxygenation and stronger signs of aqueous processing, and at Whistler—foggy, cloud-touched air in British Columbia—aqSOA patterns reproduced aged organic aerosol better than gas-only chemistry. Across sites, oxalate shows up as a reliable tracer of water chemistry. In-cloud measurements over Ohio paired oxalate tightly with sulfate—classic aqueous behavior. In marine stratocumulus off Monterey, models that made oxalate in clouds overshot observations by roughly a factor of ten, a hint that post-formation losses like wet scavenging or further chemistry were being missed. And in Mexico City's gas phase, glyoxal had a missing sink—it disappeared faster than it should have—right when particle-phase organics climbed, flagging uptake and aqueous processing into SOA. Models sharpen those inferences. When cloud aqSOA formation was added to the regional model CMAQ, the chronic low bias in water-soluble organic carbon shrank dramatically. The median bias moved from around minus sixty-four percent to about minus fifteen, and the model started to catch the day-to-day swings tied to clouds. That's a big jump in credibility. It also underlines a point from Ervens, Turpin, and Weber: early models paid far more attention to clouds than to the persistent, lower liquid water reality of aerosol particles. That matters, because concentrated aerosol water is where oligomerization locks mass in. How do you represent all of this chemistry without melting your computer? There are two schools of thought. One is to take a detailed multiphase mechanism—think the CAPRAM family—and march radicals and intermediates through every aqueous step in clouds and particles. You gain mechanism, but you carry a huge burden. The other is to step back and approximate. You give the uptake of a soluble precursor an effective coefficient, or you let products live on a volatility ladder—the so-called volatility basis set—and ask how much ends up condensed. Both approaches respect partitioning physics. In the gas-to-particle world, products slide into particles based on their effective saturation concentration and the organic mass waiting for them. In the aqueous world, the dissolved fraction scales with Henry's law, so the particle-phase share rises with solubility and liquid water content. The time-scale framework laid out by Kerminen and Wexler is helpful here: it compares how fast you can move a molecule into water and process it versus how fast it would otherwise create gas-derived SOA. The punchline is intuitive—more water, more solubility, and faster aqueous reactions tip the balance toward aqSOA. Put numbers to that balance and patterns emerge. For isoprene in clouds, parameterized aqueous yields over a few hours can hit three to nine percent when nitrogen oxides are high, but drop below one percent when they are low. In the same air, purely gas-derived yields sit around one to four percent. In a box model that cycles parcels through clouds and clear air, total cloud-made aqSOA can reach about two micrograms per cubic meter, with half of that mass sticking around after three cloud cycles and evaporations. Outside clouds, aqueous chemistry in particle water keeps going, adding roughly another microgram per cubic meter over the same window. That's extended processing, not a one-and-done. Now zoom out to regional scenarios. In humid, anthropogenically influenced places—a Los Angeles-like case around seventy percent humidity or the Po Valley around eighty—aqSOA supplies roughly thirty to fifty percent of the total SOA budget. In a biogenic, very humid Amazon-like case, the aqueous and gas routes come out about even. In a dry Houston-like case, around forty-five percent humidity, aqSOA fades; there's just not enough particle water to matter. Those are model scenarios, but they line up with the field stories you just heard. A last mechanistic contrast helps connect the dots. Gas phase products can find a home in particles fairly readily when there's a microgram or ten of organic mass around—thirty to sixty percent of semivolatile products might be in the particle at a microgram per cubic meter, climbing above sixty percent at ten. For aqSOA, the particle-phase share is limited by how much water you have to dissolve into. The highest values in aerosol water are about thirty percent when particle liquid water sits near ten micrograms per cubic meter, falling to a few percent when water is scant. Clouds, with a thousand times more water, soak up and process much more, but some of those products are too volatile to survive evaporation. The ones that do—oligomers, acids, organosulfates—build the enduring aqSOA mass. Precursor by precursor, the story varies. In concentrated aqueous systems, glyoxal can yield aqSOA at the top end—near complete conversion in some lab setups—while methylglyoxal approaches seventy percent. For aromatic parents like toluene and xylene, the glyoxal and methylglyoxal they produce don't all make it into aqSOA; modeled aqueous yields there sit in the low single digits, around two to three percent, and methylglyoxal's lower solubility holds it back. Isoprene's aqueous yields are on par with its gas-derived ones, while benzene looks unlikely to contribute much through water. And those MACR and MVK numbers from isoprene's family—roughly two to twelve percent and five to ten percent in aqueous labs—fill in the midrange. Step back and you can see why the two thousand eleven synthesis landed so hard. In humid air, especially where biogenic emissions meet anthropogenic nitrogen oxides, aqSOA isn't a footnote. It's a coequal pathway that pushes particles toward higher oxygenation, more water uptake, and, in some cases, stronger light absorption. It also explains why field oxygen-to-carbon ratios climb with humidity, why morning peaks in water-soluble organics travel with nitrate and particle water, and why models that ignore cloud and particle water undercount the organic mass. Where does that leave us? With a to-do list that's refreshingly concrete. Oxalate and glyoxal have proven their worth as tracers, but we need more markers that cleanly tag aqueous processing across different seeds and pH. We need parameterizations—simple, tested ones—that let big models represent both cloud and particle water without pretending they're the same. And we need lab studies that keep pushing into the messy middle of the particle: high ionic strength, phase-separated droplets, interfaces that can be as reactive as the bulk. The reward is practical. Better SOA models mean better climate and air-quality predictions. And as warmer, wetter conditions expand the parts of the world that sit in the aqSOA sweet spot, this chemistry only becomes more relevant. That's the arc Ervens, Turpin, and Weber drew by combining lab kinetics, field fingerprints from Atlanta to Mexico City to Monterey, and models from CAPRAM to CMAQ. Once you let water into the story, a lot of the atmosphere's organic puzzles start to make sense.