Organic aerosol components observed in Northern Hemispheric datasets from Aerosol Mass Spectrometry

If you've ever watched winter sunlight slice through cold air and reveal a haze you didn't know was there, you were looking at a puzzle that atmospheric chemists have been trying to solve for two decades. That haze is filled with organic aerosol—tiny carbon-rich particles small enough to get deep into your lungs and bright enough to matter for climate. It's a wild mix of stuff from tailpipes, trees, kitchen fires, and the chemistry that happens once all of that floats into the sky. The question is simple to say and hard to answer: what is it, and how does it change as it ages? A breakthrough came when researchers started using a tool called the Aerosol Mass Spectrometer, or AMS, to read the chemical fingerprints of these particles in real time. Out of those fingerprints, a pattern emerged. Almost everywhere in the Northern Hemisphere, the organic aerosol can be described as a blend of two big families. One looks like hydrocarbons straight from combustion—the "fresh" primary material from traffic and other sources. That's called hydrocarbon-like organic aerosol, or HOA. The other looks more cooked, more oxygenated, the product of atmospheric chemistry. That's oxygenated organic aerosol, or OOA. And here's the twist: even that oxidized family splits, not cleanly but consistently, into two flavors that reflect how far aging has gone. There's a semivolatile, less oxidized subtype—SV-OOA—and a low-volatility, more oxidized subtype—LV-OOA. How do we tell them apart? Counterintuitively, we listen to just two notes in the mass spectrum. The AMS reports signals at many mass-to-charge ratios, but two are workhorses. One at 44 is largely carbon dioxide ion, a flag for carboxylic acid groups and advanced oxidation. Another at 43 is dominated by a small oxygenated fragment, C2H3O+, which thrives in fresher secondary aerosol. People talk about f44 and f43—the fraction of the total signal that lands at those two masses. In plain terms, more 44 means more oxygen; more 43 often means you're earlier in the oxidation journey. Across dozens of field campaigns, those two fractions plot out a recurring picture. SV-OOA sits lower on the oxygen scale, with a typical f44 near 0.07 and an atomic oxygen-to-carbon ratio—O:C—around 0.35. LV-OOA stakes out a higher ground, with f44 around 0.17 and O:C about 0.73. The entire OOA ensemble averages in between, roughly f44 near 0.14 and O:C around 0.62. Don't treat those as hard boundaries; they're guideposts. At any given site, the "SV" and "LV" labels are relative. The real message is the continuum: a path from fresher, more volatile secondary aerosol to an aged, low-volatility endmember. Plot those f44 and f43 values for components extracted from many sites, and you get a triangle—literally. Everything sits inside a well-defined wedge in that two-dimensional space. SV-OOA tends to fill the lower half of the triangle, LV-OOA the upper half, with total OOA in between. As aerosol ages, points march toward the LV-OOA corner. Describe it another way—put O:C on one axis and the ratio of 44 to 43 on the other—and you see the same arc: rapid oxidation early, then a plateau as chemistry runs out of steam. That leveling-off lands near an O:C of about 1.2. It's like a speed limit for atmospheric oxygenation. Now, this is a podcast, not a manual, but tools matter for trust. So here's how scientists pulled out these components from the chaos. The AMS itself comes in a few flavors: an older quadrupole version, compact and high-resolution time-of-flight versions, all designed and refined over the two thousands by groups led by Jayne, Drewnick, DeCarlo, and reviewed by Canagaratna and colleagues. The core idea is the same: flash-vaporize submicron particles, ionize the vapor, and read the mass spectrum. To make sense of that torrent of data, teams turn to a technique called positive matrix factorization, or PMF. PMF is basically a bookkeeping equation with constraints. Imagine each measured spectrum at each time point—call it x—is a weighted sum of a few underlying source spectra plus some noise. In symbols, x equals G times F, plus an error term. G is the time series of each factor; F is the characteristic spectrum of each factor. The trick is to solve for both at once, without telling the algorithm ahead of time what the factors look like. PMF does that by minimizing the weighted squared errors, while forcing all contributions to be non-negative. In this hemispheric synthesis, researchers used a standard PMF engine in what's called robust mode, and they pressure-tested the solutions—varying starting points, nudging the rotation parameter known as FPEAK, and checking the factors against outside tracers like carbon monoxide, ozone, and nitrate. Where there were strong prior hints—say, biomass burning during a known event—some teams used a cousin of PMF, the Multilinear Engine, to gently steer the solution with constraints. Across 43 AMS datasets, with 27 analyzed in this consistent way, the same families kept showing up. What gives those families away are the fragments in the spectrum. HOA looks like a parade of hydrocarbon ions—peaks at 27, 29, 41, 43, 55, 57, and on up. OOA, by contrast, puts much of its weight at 44. High-resolution measurements confirm that 43, in ambient oxygenated aerosol, is overwhelmingly that C2H3O+ ion, and 44 is tied to oxygenated and acidic groups. That's why f44 and f43 work as a shorthand for oxidation state. And because you often want a single number for "how oxygenated is this," the teams mapped f44 to O:C for unit-mass data using a calibration reported by Aiken and colleagues, while high-resolution data yield O:C directly. If you're worried about stitching together apples and oranges, you're not wrong to ask. Instruments differ, sites differ, seasons differ. But the inter-instrument quirks are generally small. In careful comparisons—down to how diacids look in the AMS—the differences in those key fractions were around a couple of percent. Meanwhile, the factor choices were anchored by reality checks: do the "HOA" time series rise with traffic? Does the "OOA" component peak in the afternoon with photochemistry? Those kinds of diagnostics kept the math honest. All of this triangulation leads to a simple picture of aging you can keep in your head. Fresh emissions and fresh secondary material start lower in the triangle. Sunlight and oxidants go to work. The mass spectrum shifts: 43 becomes less important, 44 swells, O:C climbs. Over time you end up in a regime that looks like LV-OOA, and that region is suspiciously close to spectra of humic-like substances—HULIS—found in atmospheric particles and in natural organic matter. It's a chemical echo: highly oxidized, acidic, low-volatility material that the atmosphere keeps converging toward. Now, here's where the field had to get humble. When scientists tried to recreate this evolution in the lab, the aerosol didn't always climb to the LV-OOA corner. In chamber after chamber, when you plot the highest f44 they saw against how much organic mass was in the air, you see a trend: as the chamber fills with more aerosol, f44 tends to drop. That makes sense if you remember how semivolatile organics behave. At higher particle mass, even relatively unoxidized vapors condense, which dilutes the fraction of highly oxidized material in the particle. The effect is strongest at low loadings—think zero to twenty micrograms per cubic meter—then eases off as you approach a hundred. It's a partitioning story layered on top of chemistry. Chemistry still matters. Chen and colleagues at Caltech ran aromatic systems with higher effective aging—longer residence times and lower wall losses—so even with an average hydroxyl radical concentration on the order of 13 million per cubic centimeter, their aerosol marched farther up the oxidation axis than similar experiments at Carnegie Mellon. Switch to flow tubes and you can push exposures further. In those setups, hydroxyl radical doses reach up to something like eight times ten to the tenth per cubic centimeter-seconds, and the particles get very oxidized, very quickly. That's the lab proving the point: if you give the system time—or dose—it will head toward the LV-OOA regime. But in most chambers, short experiments and sticky walls pull it back. On top of that sit the fingerprints of different precursors. Ozonolysis of alpha-pinene often shows f44 leveling off after the particles grow, suggesting a balance between the formation of oxygenated products and their dilution by additional mass. Photooxidation of the same terpene, or of an aromatic like m-xylene, can tell a different story, with f44 rising as the molecules keep adding oxygen. Same triangle, different paths through it. The power of this synthesis is that it doesn't just catalog spectra; it offers a coordinate system for aging. Zhang and colleagues described that f44–f43 triangle as a standard map, and it has become one. Chamber data sit in the same territory as ambient points but weigh the lower half more heavily; ambient aerosol, with days to weeks of processing, pushes toward the apex. The curve you get by plotting O:C versus the 44-to-43 ratio tracks that same migration, rising fast and then gliding toward that O:C of roughly 1.2. You can visualize the evolution from SV-OOA to LV-OOA without ever seeing a figure. Modelers took that map and built a framework around it. Rather than treating organic aerosol as a single blob that just partitions between gas and particle, Jimenez and colleagues argued for a two-dimensional volatility basis set. Put volatility on one axis and oxidation state—O:C—on the other. Let emissions and chemistry move material around that grid. Suddenly, the triangle isn't just a picture; it's a state space for simulating how a fresh, semivolatile mixture can evolve into something low-volatility and highly oxygenated, consistent with what the AMS sees. There are limits, and they're not shy about them. Calling one component "SV-OOA" and another "LV-OOA" is a site-relative choice, not a declaration that you've measured absolute volatility. PMF solutions can rotate; two different seeds can land you in slightly different basins. That's why they explored those rotations, chose solutions that matched external tracers best, and where the chemistry was obvious, used constrained methods to stabilize the answers. Instruments vary, but intercomparisons suggest those differences are small next to the atmospheric patterns you care about. Why should you care about a triangle of two mass-spectral peaks? Because it gives everyone—from field teams to chamber chemists to climate modelers—a common language. Cities differ. Forests differ. The weather differs from day to day. Yet when you map their aerosol onto f44 and f43, you can line up a traffic-choked Beijing winter with a spruce forest in Finland and see the same aging arc. That lets you compare sites honestly, quantify how much aging has happened, and test if your lab setup is recreating the real world or a caricature of it. There's also a public health and climate angle hiding in those numbers. More oxidized, low-volatility aerosol scatters light differently and acts as a better cloud seed than its fresher cousins. It's also stickier in your lungs and harder to evaporate. Knowing how fast the atmosphere drives material up that oxidation ladder—knowing, for example, that O:C climbs steeply then stalls near 1.2—helps translate emissions into radiative effects and exposure risks more faithfully. If you want to picture the take-home in a single image, imagine a triangle on a sheet of paper. At the bottom edge, fresh secondary aerosol with a decent 43 signal. Up at the top, LV-OOA with a strong 44. Points flow uphill as the sun works on them, stalling near the peak. Chamber experiments dot the lower slopes unless you give them time and tame the walls. And across 43 datasets—urban, rural, ground, aircraft—that same mountain keeps reappearing. Looking ahead, the speculation is modest and focused. Better measurements of true volatility across sites would turn those SV and LV labels from relative tags into physical properties, and chambers with lower wall losses and longer timescales are already helping to close the gap with ambient aerosol. In models, the two-dimensional volatility and oxidation grid is proving to be a flexible scaffold. As more ambient points and more laboratory paths fill in that space, the translation from what we measure to what we predict should get crisper. But even without that future work, the present story is clear. Two fragments in a mass spectrometer—44 and 43—taught us that organic aerosol ages along a common path. It's a simple map drawn from complex chemistry, and for once, the map seems to fit the world.