forever chemicals with no return address

Problem statement

Wastewater treatment plants do not make PFAS; they concentrate whatever comes down the sewer into sludge, and about 60% of U.S. sewage sludge is then land-applied as biosolids fertilizer. States are now capping PFOA/PFOS in land-applied biosolids (Michigan prohibits land application above 100 ppb PFOS as of 2026, down from 150 ppb; Virginia's new law sets 50 ppb) or banning the practice outright (Maine 2022, Connecticut 2024). The one intervention with a proven track record is industrial pretreatment: Michigan utilities with an identifiable industrial discharger achieved "up to a 99% reduction in PFOS sewage sludge concentrations." But most sewersheds have no chrome plater or paper mill to point at. Their PFAS arrives diffusely — from household products, laundry, cosmetics, food packaging, commercial laundries, car washes, landfill leachate — and "the actual contribution of households to the load of PFAS in raw wastewater is currently unknown," as is the amount from commercial operations and "the major products or processes leading to high release." A utility facing a biosolids limit it cannot meet has no method to find out which of thousands of upstream sources to act on.

Why this matters

Land application is, in EPA's words, "a crucial and cost-effective management option for our nation's utilities, especially in locations where landfill and incinerator capacity is limited"; when Maine banned it, the state saw "significant issues and cost increases for utilities" with limited landfill space and no in-state incinerators. Every additional state limit turns a source-identification problem into a disposal crisis for utilities that cannot lower their numbers. Farmers who received biosolids for decades face contaminated soil and product; ratepayers face the cost of hauling sludge to distant landfills or incinerators. And because PFAS in leachate from landfilled sludge returns to a treatment plant, moving sludge around does not remove PFAS from the loop — only source control does. Without a way to trace diffuse sources, the only levers left are bans on land application or bans on entire chemical classes in consumer products, both of which are slow, contested and blunt.

What’s been tried and why it hasn’t worked

Michigan's tiered industrial-pretreatment strategy (2018 onward), copied by Wisconsin, Minnesota, Colorado, Maryland and Virginia, uses PFOS as an indicator to identify "the most significant dischargers" — an approach that by design finds point sources and does nothing for a sewershed whose loading is spread across households and small businesses. Monitoring mandates (Washington, Oregon, Florida) generate occurrence data at the plant, not upstream. EPA itself, in its 2026 draft guidance, faults its 2025 draft risk assessment for "failing to conduct a national survey to document occurrence of PFOA and PFOS in sewage sludge," so even the baseline distribution across plants is unknown; state surveys exist for a growing but partial list of states. Sewershed source-tracking studies have been attempted — a 2020 Southern California project with nine water agencies planned ~750 samples across sewer conveyances, influent, treatment units and residential, commercial and industrial sources for 17 PFAS — but such campaigns depend on laboratory LC-MS/MS analysis (EPA Method 1633 covers dozens of PFAS), are expensive and slow, and produce a snapshot of a network whose flows change hourly. Precursor compounds that transform to PFOA/PFOS inside the plant further blur the trail: what is measured in sludge is not what entered the sewer. So a utility can know its sludge is at 80 ppb and still have no defensible way to say which streets, business types or products are responsible.

What would unlock progress

The problem needs a source-apportionment toolkit for diffuse PFAS analogous to what exists for stormwater bacteria (microbial source tracking) or urban air pollution (receptor modeling): fingerprinting by PFAS congener profile and precursor content that can distinguish household laundry from commercial laundries, car washes, landfill leachate or food-service; sampling designs that use the sewer network's tree structure to bisect the loading with far fewer samples; and lower-cost or field screening methods (total organic fluorine, passive samplers) that make repeated upstream sampling affordable. On the data side, a national occurrence survey — the gap EPA names — would let utilities know whether their sludge is unusual before they spend on tracing.

Entry points for student teams

A team with a partner utility could design and pilot a network-bisection sampling plan (passive samplers at manholes chosen by sewer topology, composited over days) and demonstrate whether it localizes loading to sub-catchments at a fraction of the cost of a full campaign. A data team could compile published sewershed and product PFAS congener profiles into an open fingerprint library and test whether receptor-model apportionment is feasible on real influent data. A policy team could model, for one state, the cost trade-off between source-tracking programs and the disposal costs a land-application limit would impose on plants that cannot comply. Skills: environmental chemistry, network sampling design, statistics/chemometrics, wastewater engineering, environmental policy.