The research, led by Yifan Zhang and Mingjin Tang of the Guangzhou Institute of Geochemistry, together with colleagues across China, harnessed a powerful tracing tool: the stable isotope ratio of iron, expressed as δ56Fe. Isotopes are atoms of the same element with slightly different masses, and natural processes — evaporation, biological uptake, industrial combustion — can sort them, leaving each source with a subtly different isotopic signature. By measuring this signature in aerosol particles, scientists can in principle work backwards and calculate how much iron each source contributes, much like a forensic accountant tracing money through a labyrinth of accounts.

The trouble has been that the reference values, or endmembers, for anthropogenic sources were poorly known. Field studies routinely assumed that any iron lighter than the crustal value of roughly +0.1 per mil must be human-made, and many studies assigned a single, very low endmember value — sometimes as low as −4.7 per mil — to all pollution sources lumped together. But those assignments rested on a handful of measurements from a handful of sources. Without reliable endmembers, the entire edifice of isotope-based source apportionment wobbled.

To shore it up, the team assembled an unusually broad collection of samples. Seven desert dusts were analyzed, including Saharan dust from the Cape Verde Islands, Arizona Test Dust, and dusts from the Taklimakan Desert, the Gobi, Qinghai, the Tibetan Plateau, and the famous Luochuan loess section of the Chinese Loess Plateau. On the anthropogenic side, they obtained 28 fly ash samples from coal-fired power plants spanning 25 Chinese provinces, 21 fly ash samples from steelworks covering processes from coking and sintering to blast furnaces and casting, and 11 biofuel burning aerosols generated in the laboratory from five crop straws and six wood types burned in a commercial cook stove. They also measured municipal waste incineration fly ash, heavy oil bottom ash, and a certified urban particulate matter reference material.

The analytical work was exacting. Samples were digested in a Class 100 ultra-clean laboratory using acid protocols tailored to each material, with steelwork and biofuel samples requiring microwave-assisted digestion to break down refractory carbonaceous matter. Iron was then chemically purified on cation exchange resin columns, achieving recoveries above 98 percent with negligible procedural blanks, and isotope ratios were measured on a Neptune Plus multi-collector inductively coupled plasma mass spectrometer in pseudo-high-resolution mode to eliminate argon-based interferences. The team verified their precision with replicate measurements agreeing within 0.08 per mil and confirmed accuracy against four international reference standards.

The results painted a strikingly varied picture. Desert dust proved remarkably homogeneous, averaging +0.14 ± 0.10 per mil, closely matching the upper continental crust. Power plant coal fly ash, however, averaged +0.26 ± 0.18 per mil — heavier than dust, not lighter. Steelwork fly ash averaged −0.07 ± 0.41 per mil, only slightly below dust, with the striking exception of coking fly ash, which averaged +0.42 per mil and mirrored coal fly ash, a telltale sign of its shared coal feedstock. Most dramatically, biofuel burning aerosols averaged −0.28 ± 0.39 per mil, with individual samples spanning from −1.23 per mil for pear wood smoke to +0.21 per mil for soybean straw smoke — the first robust endmember determination for this source.

The headline finding is conceptual: not all anthropogenic aerosol iron is isotopically lighter than natural dust. This directly contradicts the conventional assumption underpinning many two-component mixing models, in which lighter-than-crustal iron was automatically attributed to pollution. Coal combustion, one of the largest industrial iron sources, actually tends to enrich the heavier isotope. The authors suggest this could even explain puzzling observations of anomalously heavy iron in some tropospheric aerosols over the Pacific, which previous studies had struggled to attribute. Meanwhile, steelworks and domestic biofuel burning emerge as plausible contributors to the light iron frequently observed in the atmosphere.

The study also carries important caveats that the authors confront candidly. Fly ash particles captured by dust collectors are considerably larger, often tens of micrometers, than the fine aerosols that travel far downwind. Prior work by Kurisu and colleagues has shown that lighter iron isotopes preferentially evaporate in high-temperature combustion and recondense as nanoparticles, potentially making fine emitted particles isotopically lighter than the bulk fly ash measured here. An earlier study near a French steel plant found no such fractionation, leaving the question open — and the authors call for size-resolved measurements of emissions directly at the source. They likewise note that wildfire aerosols, which mix soil and biomass iron, likely carry different signatures than the domestic stove burning simulated here, and that isotope fractionation during dissolution may mean total and soluble iron do not share identical endmembers.

For the field of atmospheric chemistry, the implications are immediate. Simple two-component mixing models that treat all pollution as a single isotopically light reservoir are no longer defensible. The authors point instead to Bayesian frameworks such as MixSIAR, which can accommodate multiple sources and their uncertainties — and which now, for the first time, have measured endmember values to draw upon. Because anthropogenic iron, though a minor fraction of total atmospheric iron, dissolves far more readily than desert dust and thus dominates the soluble iron fertilizing remote oceans, getting these fingerprints right has consequences for modeling marine productivity and the global carbon cycle.

What began as a technical gap — a missing table of isotope values — turns out to reshape how we read the atmosphere itself. Every plume of smoke, every dust storm, every industrial emission now carries a quantifiable isotopic signature, and with this study the dictionary translating those signatures into sources has grown dramatically richer. The next chapter, the authors argue, lies in measuring the particles that actually escape into the air, at the size that matters, from the sources that matter most — a task that will determine whether the atmosphere’s iron budget can finally be balanced with confidence.

Subject of Research: Stable iron isotope endmember values of anthropogenic and natural aerosol sources for atmospheric source apportionment

Article Title: Isotopic composition of aerosol iron from anthropogenic sources: implications for source apportionment of aerosol iron

Article References: Zhang, Y., Zhu, G., Li, R., Liu, M., Yang, Y., Zhang, T., Chen, Y., Ma, J., Wang, X., & Tang, M. (2026). Isotopic composition of aerosol iron from anthropogenic sources: implications for source apportionment of aerosol iron. Atmospheric Chemistry and Physics, 26(19), 14073-14084. https://doi.org/10.5194/acp-26-14073-2026

Image Credits: AI Generated

DOI: 10.5194/acp-26-14073-2026

Keywords: aerosol iron, iron isotopes, source apportionment, desert dust, coal fly ash, steelwork emissions, biofuel burning, atmospheric chemistry, marine biogeochemistry, MC-ICP-MS, soluble iron, air pollution