Battery Recycling
and Secondary Mineral Supply

Battery recycling puts those minerals back into the manufacturing chain. Most of the optimism around this process does not survive contact with the actual chemistry and economics.

Recycling cannot replace primary mining for at least another decade. The timeline math alone rules it out: EV batteries are designed for 8 to 15 years of service, so the 14 million vehicles sold in 2023 will not yield recycling feedstock in volume until the 2030s at the earliest. Many of those packs will pass through a second-life phase in stationary storage before reaching end of life, pushing the recycling date out to the 2040s. In the meantime, what recyclers process is overwhelmingly manufacturing scrap from gigafactories, not retired vehicles. The secondary mineral supply contribution of recycling in 2025 is a fraction of a percent of total lithium and nickel demand. It will remain a fraction for years.

Recovery rates for cobalt and nickel above 95% are commercially demonstrated in hydrometallurgical processing. Lithium runs 80 to 90%. These numbers appear in press releases, policy documents, and investor decks, and they say almost nothing about whether recycled material can go back into a battery.

Cathode precursor manufacturers maintain incoming specifications that most recyclers cannot consistently meet. The 2023 International Battery Seminar included multiple presentations where precursor producers disclosed copper limits below 5 ppm, with some insisting on below 1 ppm. Iron below 5 ppm. Zinc below 5 ppm. Calcium below 10 ppm. Material that misses spec does not enter battery production. It gets sold to the ceramics or steel industry at a steep discount, and the secondary mineral supply chain has gained nothing from the effort.

Copper is the hardest impurity to manage because of how lithium-ion cells are built. The cathode coating sits on aluminum foil. The anode coating sits on copper foil. Shredding a cell co-mingles them at every particle size. Eddy current separators handle copper fragments above ~1 mm. Below 100 microns, fine copper dust persists in the black mass and dissolves into the acid leach solution alongside the target metals.

Removing copper from the pregnant leach solution (the metal-laden sulfuric acid solution produced after leaching black mass) to below 5 ppm requires dedicated chemical steps, and each one introduces complications. Cementation with iron powder works by electrochemical displacement: iron is more electropositive, so adding iron powder precipitates copper as metallic copper. Dissolved iron then contaminates the solution and must be removed in a subsequent hydroxide precipitation step at controlled pH. Overshoot the pH and nickel and cobalt co-precipitate, which is yield loss.

Sulfide precipitation using sodium hydrosulfide is more selective for copper at low pH. Precise dosing is critical. Excess NaHS co-precipitates nickel and cobalt sulfides into the waste stream. The operator is managing a dosing decision on a solution whose composition shifts batch to batch depending on what cell chemistries entered the shredder that week. A batch heavy in consumer electronics cells (LCO chemistry, cobalt-rich, relatively little nickel) behaves differently in sulfide precipitation than a batch dominated by automotive NMC811 cells.

In practice, the copper removal step is where many recyclers discover the difference between a laboratory flowsheet and an industrial operation. At lab scale, you control your input chemistry. You know what went into the leach reactor. You can tune reagent additions precisely because you are working with 500 mL of solution in a beaker. At plant scale, you are processing the output of a shredder that ate a mixed load of prismatic automotive cells, cylindrical consumer electronics cells, and possibly some pouch cells from e-bike battery packs, all in a single campaign. The PLS composition changes from batch to batch and sometimes within a batch. The copper concentration in the incoming solution might be 150 ppm one week and 400 ppm the next, depending on the ratio of copper current collector foil to total cathode mass in the shredder input. Adjusting the cementation or sulfide precipitation parameters for each batch requires either real-time analytical feedback (inline ICP monitoring, which some plants have and many do not) or conservative overdosing that sacrifices yield to guarantee copper removal. Conservative overdosing means losing cobalt and nickel to the precipitate, which is throwing away the most valuable metals in the stream. None of this appears in published flowsheets or corporate presentations.

Redwood Materials, Li-Cycle, Brunp Recycling, SungEel HiTech all face this. When these companies announce "battery-grade" output, it matters whether CNGR, Umicore, or Huayou Cobalt has accepted that material at commercial volume against their incoming cathode precursor specification. Or whether a polishing refining step sits between the recycler's output and the cathode production line. If the latter, what the recycler has produced is an intermediate, not a battery-grade input. Recyclers are not always transparent about which category they occupy.

Every commercial lithium-ion cell contains 2 to 5% fluorine by mass, split between LiPF₆ electrolyte salt and PVDF cathode binder. During the thermal pretreatment that most recyclers run at 300 to 600°C to burn off organics and deactivate residual electrolyte, lithium in the cathode reacts with fluorine to form lithium fluoride. LiF has near-zero solubility in the dilute sulfuric acid used for leaching.

Makuza, Tian, Guo, Chattopadhyay, and Yu published a review in the Journal of Power Sources (vol. 534, 2022) identifying this LiF formation as a primary mechanism behind the gap between theoretical and achieved lithium recovery. A recycler dissolving 98% of cobalt and 96% of nickel from a batch might recover only 82% of lithium from the same material. The missing lithium has been chemically locked into an insoluble compound by the cell's own fluorine content during the pretreatment step that was supposed to make the material safe to handle.

LiPF₆ decomposition also produces HF gas, which requires scrubber infrastructure rated for hydrofluoric acid. Capital cost for HF-rated off-gas systems is substantial. PVDF decomposition above 350°C releases additional HF along with other fluorinated pyrolysis species. The off-gas treatment adds to facility cost and permitting complexity. The solid-phase LiF problem is arguably worse because it represents permanent lithium loss within the process itself, not a safety or environmental cost that can be managed with engineering controls.

PVDF and LiPF₆ production both require hydrofluoric acid sourced from fluorspar mining. The fluorine in spent batteries could be recovered. No commercial recycler does so. It leaves as calcium fluoride sludge for landfill.

The recycling industry splits into two activities that look like one business from outside but have radically different economics.

Mechanical pretreatment covers discharge, shredding, and physical separation. A line handling 10,000 to 20,000 tonnes per year costs $10 to $30 million to build. The product is black mass, copper scrap, aluminum scrap, steel. Black mass margins are thin and volatile. The operator absorbs hazardous waste classification headaches, fire risk during shredding (cells retaining charge, electrolyte vapors), and whatever Basel Convention paperwork is needed to export the black mass if domestic refining capacity does not exist.

Hydrometallurgical refining is where black mass becomes specification-grade nickel sulfate, cobalt sulfate, manganese sulfate, and lithium carbonate or hydroxide. Li-Cycle's Rochester Hub is the most publicly documented Western example. Initial budget: ~$485 million per SEC filings. The project paused in late 2023 after cost overruns and the company restructured. That capital covers acid-resistant reactor vessels, solvent extraction mixer-settler trains, crystallization equipment, wastewater treatment, HF-rated gas handling. The margin in the value chain concentrates here because this is the step where a variable intermediate becomes a product that cathode precursor manufacturers will purchase at battery-grade pricing.

Multiple North American and European companies run preprocessing domestically and ship black mass to Brunp, GEM Co., or Huayou Cobalt in China for refining. The IRA and EU Battery Regulation try to incentivize domestic refining through content requirements and rules of origin. The refining capacity to back those policies does not exist yet at scale in North America or Europe.

Basel Convention hazardous waste classification means cross-border shipment requires prior informed consent, a process taking three to six months per route. Class 9 dangerous goods freight costs three to five times non-hazardous rates. NMC-rich black mass worth $10,000+ per tonne can absorb the freight. LFP-derived material at $2,000 to $4,000 per tonne often cannot.

LFP recycling is going to be structurally unprofitable for the foreseeable future, meaning at least through 2030. Subsidies and gate fees will be necessary to make it happen at all. The circular economy narrative around batteries quietly assumes that recycling pays for itself, which is true for NMC and false for LFP. The industry has been slow to say this because it complicates the story, and the story is what attracts investment.

CATL and BYD shipped more LFP capacity into the Chinese market in 2023 and 2024 than all NMC producers combined. Tesla uses CATL LFP cells in standard-range Model 3 and Model Y. European and North American OEMs are adopting LFP for entry-level BEVs. LFP will constitute a large and growing share of the end-of-life battery stream within a decade.

Recovered metal value per tonne of NMC811 black mass exceeds $10,000 to $15,000 at mid-cycle commodity prices. LFP cathode contains lithium, iron, and phosphate. Iron sulfate trades at $0.10 to $0.30/kg. Lithium carbonate crashed from over $70/kg in late 2022 to below $15/kg by mid-2024 according to Benchmark Mineral Intelligence spot assessments. Total recoverable value per tonne of LFP black mass at those lithium prices: $2,000 to $4,000. Processing cost per tonne does not vary much by chemistry. LFP recycling at current lithium prices loses money.

Add discharge, shredding, thermal pretreatment, leaching, purification, wastewater treatment. Add the insurance. Add the permitting compliance overhead. The fully loaded processing cost per tonne of input is in the $1,500 to $3,000 range for most Western operators. When the recoverable metal value at the other end is $2,000 to $4,000 before refining costs, and the refined product must compete on price with virgin lithium carbonate from Chile and Australia, the margin goes negative.

The EU Battery Regulation mandates 50% lithium recovery by 2027, 80% by 2031, with no chemistry exemption. Recyclers will be legally required to process LFP at a loss. That loss gets absorbed through cross-subsidy from NMC recycling margins, gate fees charged to battery producers (which feed into vehicle pricing), or government subsidies. Probably all three. If lithium prices stay depressed for several years, the cumulative gate fee burden on producers becomes a nontrivial cost line that the automotive industry has been quiet about.

Direct recycling of LFP, which regenerates cathode powder through relithiation without breaking it down to elements, could change the economics by skipping the full leach-separate-resynthesize sequence. Guangdong Brunp has published results. The process works on manufacturing scrap, which is homogeneous and minimally degraded. Whether it works on end-of-life cells with ten years of varied degradation, mixed cycling histories, and unknown storage conditions is undemonstrated at commercial scale. Betting that LFP direct recycling will mature fast enough to close the gap before the 2027 EU compliance deadline is a gamble. If it does not, secondary lithium supply from the fastest-growing battery chemistry becomes hostage to continued willingness to fund an activity that loses money on every tonne.

In 2025, recyclers feed mostly on gigafactory manufacturing scrap. Electrode coating trims, calendering rejects, cells failing formation cycling, moisture-contaminated batches. Scrap rates at mature lines run 5%; ramp-up lines can hit 20% or more. Northvolt publicly acknowledged yield problems at Skellefteå before entering restructuring in late 2024. Panasonic's early ramp at the Tesla Nevada Gigafactory went through an extended yield optimization period reported in trade press (Nikkei Asia, among others).

This scrap is perfect feedstock: single chemistry, well-documented, bulk quantities at one location, no field-use degradation, no safety complications from aged cells. Redwood Materials holds supply agreements with Panasonic and Toyota for it. SungEel HiTech in South Korea processes domestic manufacturer rejects.

End-of-life EV batteries from the first mass-market wave (vehicles produced 2018 to 2025) will not show up at recycling plants in volume until the mid-2030s. Batteries are designed for 8 to 15 years of vehicle service. Many will then enter second-life stationary storage, adding 5 to 15 more years. Nissan and Sumitomo have operated second-life programs with retired Leaf packs in Japan. BMW has used i3 modules in grid storage. A cell made in 2025 might not be recycled until 2048.

Gigafactories will cut scrap rates as they mature, because scrap is expensive. Scrap availability declines. End-of-life volumes have not ramped. The late 2020s and early 2030s could see recycling plants sitting partially idle, having been sized and financed on scrap-era throughput.

Informal dismantlers in receiving countries may extract packs for unregulated second use or crude recycling that recovers copper and aluminum but discards black mass. The EU Battery Regulation's digital battery passport has no legal reach over a Renault Zoe shipped to Dakar. And collection rates for lithium-ion cells in consumer electronics, even in the EU with the most mature framework, hover at ~50% under the Batteries Directive. Devices accumulate in drawers. They get exported inside used laptops.

Graphite is the largest material in a lithium-ion cell by mass, 1.0 to 1.2 kg per kWh for NMC622 versus ~0.15 kg of lithium, and no major recycler recovers it at commercial scale. Pyrometallurgical recycling burns it as fuel. Hydrometallurgical recycling leaves it in the leach residue mixed with carbon black and binder decomposition products. Producing battery-grade anode graphite from that residue has been demonstrated in labs. The 2031 EU recovery targets cover Li, Co, Ni, Mn, Cu. Graphite is absent. China imposed graphite export restrictions in late 2023. Between 50 and 70 kg of graphite per EV pack goes to waste or downcycling.

Manganese has been precipitated out with iron and aluminum as a nuisance impurity in recycling hydromet. Manganese sulfate for batteries trades at $1 to $2/kg versus $4 to $8 for nickel sulfate. CATL and BYD have both announced LMFP cathode production, adding manganese to the LFP structure for higher voltage. NMC development is trending toward high-manganese, low-cobalt formulations. Both shifts increase manganese in future recycling feedstock and tighten battery-grade MnSO₄ supply. Retrofitting a manganese SX circuit (D2EHPA extractant instead of Cyanex 272 used for Co/Ni) takes one to two years and several million dollars.

CATL announced sodium-ion in 2021. HiNa and Faradion (Reliance Industries, 2022) are also commercializing. Sodium-ion cells contain no lithium, no cobalt, no nickel, and use aluminum instead of copper for the anode current collector. Recycling these cells cannot pay for itself at any foreseeable commodity price. Gate fees from producers or regulatory mandates are the only mechanisms that will make it happen.

PVDF binder requires calcination above 500°C or NMP solvent washing to release cathode material from aluminum foil. Water-based binders (CMC/SBR) would simplify this but cell engineers have no reason to adopt them for a recycler's benefit ten years downstream. CTP designs from BYD and CATL make disassembly harder. The EU Battery Regulation mandates battery passports with disassembly info. Financial penalties for poor recyclability do not yet exist.

Thermal events at recycling plants are an insurance problem more than a technology problem. Some insurers have walked away from battery recycling entirely. Others require nitrogen-inerted shredding and large deductibles. Annual insurance: several hundred thousand dollars, omitted from most techno-economic analyses. Hydromet recycling uses 50 to 200 m³ of water per tonne of metal output, and published figures specific to battery recycling are scarce.

Through 2030, secondary supply stays below 5% of lithium demand and 5% of nickel, per IEA and Benchmark Mineral Intelligence projections. Cobalt at maybe 10 to 15% because consumer electronics recycling has operated longer. After 2035 the first mass-market EV batteries reach end of life and volumes grow. How much depends on everything above. Spent NMC811 cathode material contains lithium at 2 to 3% and nickel above 30%, versus ~1.5% Li₂O in spodumene and ~1% Ni in laterite. Energy cost for recovering lithium via hydromet recycling runs 5,000 to 10,000 kWh per tonne of LCE (Dai et al., 2019, Journal of Cleaner Production), versus 15,000 to 20,000 kWh for hard-rock extraction. The grade advantage of spent batteries over geological deposits is large enough that the thermodynamic argument should have settled things years ago. It has not, because collection rates are low, LFP recycling loses money, and most hydromet refining capacity sits in China. Until those change, secondary mineral supply from battery recycling stays marginal.