Where PFAS comes from now: landfills, and the products already in use

Most PFAS regulation is aimed at manufacture. The compounds are controlled at the point they are made or imported, on the reasonable theory that stopping production stops the problem. Two studies published in the past few weeks describe what happens after that theory succeeds: the emissions stop coming from factories and start coming from everything that was made before.

The inventory

The first reconstructs where PFOS and PFOA emissions in mainland China actually came from, year by year, from 1985 to 2023 [s2]. The method is dynamic material flow analysis, combining provincial activity data, trade statistics and sector-specific emission factors to track releases from production, industrial applications, in-use products and end-of-life management into multiple environmental media [s2].

Cumulative emissions came to 2,289.6 tons of PFOS and 846.1 tons of PFOA [s2]. The shape of the curve is an increase-peak-decline pattern, peaking around 2010 and then falling, which the authors attribute to the Stockholm Convention and to domestic regulations [s2]. Early-2010s PFOS emissions turned out to be approximately double previous estimates [s2] — a correction worth noting for anyone reading older exposure models built on the earlier figures. High emissions concentrated in eastern coastal and industrialised regions, with hotspots extending toward central and western China [s2].

The forward-looking conclusion is the one that connects to the second study: with tightening controls on PFOS and PFOA, future emissions will increasingly be driven by legacy products and waste streams [s2].

The waste stream

The second study went to look at one of those waste streams directly, compiling 1,416 leachate samples from 539 facilities in 76 cities across 22 countries, covering 150 legacy and emerging PFAS [s1].

Source strength was high almost everywhere: 93% of sites exceeded 1,000 ng/L of total PFAS and 56% exceeded 10,000 ng/L [s1]. For scale, the US drinking water limits for PFOA and PFOS are set in single-digit nanograms per litre. Leachate is not drinking water and is not consumed directly, but it is a concentrated source term entering wastewater systems and, in some settings, groundwater.

High-burden samples clustered by hotspot region and landfill type rather than by continent [s1]. Industrial landfills carried by far the strongest signature, running 56 to 101 times higher than other landfill types and contributing 62% of the global burden [s1]. Three compounds — FOSA, PFOA and PFBA — accounted for over half of cumulative PFAS [s1].

Country-level fingerprints diverged into long-chain, short-chain and precursor-enriched patterns, which the authors read as recording, respectively, legacy use, replacement inputs and precursor transformation [s1]. Higher concentrations were associated with organic matter, salts and ammonium, most strongly with total organic carbon and chemical oxygen demand [s1]. A regional structural equation model explained only 22.4% of the variance, which the authors take as evidence that what governs leachate PFAS is facility-scale — waste composition, receiving history and operation — rather than regional [s1].

For risk screening, long-chain PFAS dominated the toxicity-weighted pressure, with species sensitivity distribution-derived HC5 values for PFOS of 2.58 to 25.99 ng/L, more than an order of magnitude below the equivalent for PFOA [s1]. Short and ultrashort compounds create a separate pressure through mobility, persistence and resistance to treatment [s1].

A note on the source's numbers

Two figures in the leachate paper's abstract are printed with irregular digit separators: the upper end of the total PFAS range appears as "5542,000 ng/L" and the industrial landfill average as "1141,345 ng/L" [s1]. Read as written, those are 5,542,000 and 1,141,345 ng/L; the separator placement does not match the digit grouping. This appears to be a typesetting artifact rather than a data error, but because the values cannot be verified from the abstract alone, they are reported here as printed rather than silently corrected.

What the two say jointly

The inventory establishes that production-side controls worked, in the sense that emissions peaked and fell after the Stockholm Convention took hold [s2]. The leachate survey establishes that the material did not go away — it went into products, and the products went into landfills, and the landfills leak [s1].

That has a specific policy implication, and both papers reach for it. The leachate authors argue management should shift from end-of-pipe control to identifying fluorinated waste, reducing precursors, controlling mobile short-chain compounds and managing residuals [s1]. The inventory authors point at legacy products and waste streams as the future driver [s2]. Neither is arguing for weaker production controls. Both are arguing that production controls alone answer a question that was largely settled a decade ago.

Limits

Neither study measured human exposure or any health outcome. The leachate compilation is a synthesis of samples collected by many groups under different protocols, and the 22 countries covered are not a random sample of the world's landfills [s1]. The emissions inventory rests on emission factors and activity data, which is modelling rather than measurement — the doubling of the early-2010s PFOS estimate is itself a demonstration of how much such estimates can move [s2].

What to watch is whether leachate becomes a regulated discharge stream in its own right, and whether national inventories elsewhere reproduce the peak-and-decline shape that China's now shows.

Sources

• [s1] Landfill leachate as a global PFAS source reservoir: source-strength heterogeneity, compositional transition, and screening-level risk. Water Research, 6 August 2026. https://doi.org/10.1016/j.watres.2026.126641
• [s2] Highly Resolved Life Cycle Emissions of Per- and Polyfluoroalkyl Substances (PFAS) in China from 1985 to 2023. Environmental Science & Technology, 17 July 2026. https://doi.org/10.1021/acs.est.6c02449