the fire hiding in the recycling bin

Problem statement

Lithium-ion and lithium-metal batteries are now embedded in greeting cards, vape pens, earbuds, toys and toothbrushes, and a large share of them end up in household trash and curbside recycling. Once inside the waste system they are compacted in trucks, jostled on conveyor belts, and punctured in shredders — and a damaged lithium cell can ignite hours later, deep in a pile of paper and plastic. The unsolved problem is upstream of firefighting: materials recovery facilities (MRFs), transfer stations and scrap yards have no way to detect and pull batteries — especially the ones glued inside a product — out of a mixed, fast-moving stream before the equipment damages them. Consumer education and drop-off programs have not stopped the inflow, and the batteries are getting smaller, more numerous, and harder to see.

Why this matters

EPA's 2021 review found 64 waste facilities that experienced 245 fires caused or likely caused by lithium batteries between 2013 and 2020, in 28 states, and concluded that the true count is "severely underestimated" because it relied on media reports; every industry representative EPA consulted "agreed that LIB fires pose a threat to their operations that needs to be addressed — most even declared this issue to be the most pressing problem currently facing their industry." A Pacific Northwest landfill that began logging every battery fire in 2017 saw them rise from 21 in 2018 to 47 in 2020. As reported by trade press in 2025 (Resource Recycling, relaying Fire Rover and NWRA figures; neither primary report was checked at intake), Fire Rover's 2024 tally of publicly reported fires at U.S. and Canadian MRFs and transfer stations rose 20% year-on-year to its highest recorded level, and the National Waste & Recycling Association estimates more than 5,000 fires a year at recycling facilities. The consequences are structural: a Plano, Texas MRF destroyed by a likely battery fire in 2016 cost $30 million to replace; MRF operators told EPA that insurance premiums and deductibles were rising and that the number of insurers willing to cover MRFs had fallen from almost 50 to fewer than 10 in three years — one operator called it "the existential threat" to the recycling industry. If recycling facilities become uninsurable, curbside recycling itself is at risk.

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

Facility-level responses have been reactive: tongs and sand buckets for cells spotted smoldering on the belt, water cannons on tipping floors, thermal cameras and automated suppression, employee walk-throughs, and drivers doing "visual audits" of loads. These reduce damage but do not remove batteries before they are damaged. Consumer-facing measures — labeling, drop-off bins, county education campaigns — run into a labeling landscape EPA describes as actively confusing: with no standardized U.S. battery label, cells carry international symbols including the chasing-arrows recycling mark alongside a crossed-out bin, so residents put battery-containing products in the recycling cart. E-cigarettes marketed as "disposable" send whole lithium-metal cells into household trash. On the detection side, the technical bottleneck is that batteries are increasingly embedded: EPA notes that laptops, phones and headphones "often contain embedded LIBs that are encased in plastic or glued into devices, making them very difficult or even impossible to remove intact," and that an iPad with a glued-in cell "may take 40 minutes to disassemble and be worth $1, at most, in scrap." Even dedicated electronics recyclers struggle — one told Resource Recycling in 2025 that "more than 60% of devices his company flags for containing a battery may not actually have one," and X-ray density screening for embedded cells is still in development. Panelists at the same 2025 session could only suggest "compartmentalized" shredding modeled on how mercury lamps were segregated from TVs — a containment strategy, not a detection one. Nothing detects a 1-gram lithium cell inside a plastic toy on a belt moving at MRF throughput (which rose from an average 129 tons/day in 2001 to 214 tons/day in 2014).

What would unlock progress

Two unlocks are visible. First, a pre-shred detection method for embedded cells in mixed streams — the physical signatures are distinctive (dense metal-foil windings, lithium and cobalt/nickel X-ray fluorescence, magnetic and eddy-current responses of steel/aluminum cans, characteristic thermal behavior when perturbed) but no one has fused them into a sensor that works on unsorted material at line speed with an acceptable false-positive rate; the analogy is the metal and X-ray inspection lines used in food manufacturing to catch foreign objects at high throughput. Second, a design and labeling regime that keeps batteries out of the stream in the first place — removable cells, a single unambiguous mark, and extended producer responsibility that makes the party who glues in the battery pay for the fire — which EPA's report frames as requiring "further coordination and action by the wide variety of" waste-system stakeholders. Better incident data would help too: EPA's headline count came from local news because no reporting system exists.

Entry points for student teams

A team could build a bench-scale multi-sensor detector (e.g., X-ray transmission plus eddy-current plus low-cost XRF) and characterize detection rate and false-positive rate on a library of battery-containing consumer products buried in representative MRF material, producing the first open dataset for the problem. A second team could design and test a standardized battery-presence label plus a curbside/transfer-station intercept protocol with a local hauler, measuring battery capture per ton. A policy/design team could map, for one product category (vapes or musical greeting cards), what an EPR and design-for-removal rule would need to specify. Relevant skills: sensor engineering, computer vision, materials/electrochemistry, industrial design, waste-policy analysis.