Lithium iron phosphate

LFP is a lithium-ion cathode built on the olivine structure — the same atomic arrangement as the mineral olivine — rather than on the layered oxide arrangement of the cobalt and nickel cathodes.

That structural difference is the whole of its character. The phosphate group holds its oxygen in strong covalent bonds, so an LFP cathode does not release oxygen when it is abused, and a cell that cannot supply its own oxidiser is very much harder to set on fire. It also cycles for thousands of full charges rather than hundreds, because the lattice barely changes volume as lithium comes and goes.

What it gives up is voltage and capacity. LFP runs at about 3.2 volts against 3.7 for a nickel cathode, and stores less lithium per unit mass, so a pack built with it is heavier and bulkier for the same energy. For a phone that was disqualifying. For a car with a floor to fill and for a grid installation sitting in a field it turned out not to be.

It contains no cobalt and no nickel, which is why it went from a niche chemistry to over half of world cell production in about five years.

Two structures, two failure modes, and everything follows.

A layered oxide cathode holds lithium between sheets of nickel-cobalt-manganese oxide. Take too much lithium out, or get it too hot, and the sheets become unstable and release oxygen from the lattice. That oxygen meets a flammable organic electrolyte inside a sealed can, and the result is thermal runaway — a fire that supplies its own oxidiser and cannot be smothered.

In LFP the oxygen is not held as an oxide ion in a layer but inside a phosphate tetrahedron, covalently bound. Decomposition needs something over 250 °C rather than around 150, and even then the oxygen release is far smaller. An LFP cell can still burn; it is much harder to make it, and that difference has been demonstrated repeatedly in nail-penetration and overcharge testing.

The cycle life comes from the same place. Lithium moves in and out of the olivine framework along one-dimensional channels with a volume change of about 6 per cent and no phase collapse, so the particle is not slowly pulverised by its own breathing. Layered oxides change more and degrade faster.

LFP's original problem was that those one-dimensional channels also make it a poor electronic conductor — orders of magnitude worse than a layered oxide. It was solved by coating each particle in a nanometres-thin carbon layer and making the particles very small, which is why LFP was invented in 1996 and did not become commercially serious until the mid-2000s.

LFP is the clearest recent case of a material winning on supply chain rather than on performance.

Its inputs are iron and phosphate — abundant, cheap, produced everywhere, and not subject to any meaningful concentration risk. NMC's are nickel and cobalt, and cobalt is roughly seventy per cent Democratic Republic of the Congo with the refining overwhelmingly Chinese, carrying a documented artisanal-mining problem that no purchaser has been able to audit away.

So the buyer of an LFP pack is buying out of a commodity risk and a reputational one at the same time, and paying for it in range. When battery-grade nickel and cobalt prices spiked in 2021 and 2022, that trade got dramatically better, and LFP's share of global cell production went from a minority to a majority within a few years — with Chinese manufacturers, who had built the LFP supply chain while the rest of the industry pursued energy density, holding most of the capacity.

The patent history matters to that outcome. The core LFP patents, held through the University of Texas and Hydro-Québec, expired around 2022, and their existence is part of why western adoption lagged.

It is also the harder chemistry to recycle economically, which is the honest cost of having no valuable metal in it: an NMC cathode is worth recovering for the nickel and cobalt alone, and an LFP cathode is iron and phosphate that nobody is short of.

Standard-range electric vehicles, which is now most of them by unit worldwide. The pack is heavier and the range shorter for a given size, and both matter less than the price.

Grid and home storage, where LFP is close to unopposed. A stationary installation does not care what it weighs, it is expected to cycle daily for fifteen or twenty years, and it must not burn down a building — which is LFP's three strengths in the order they matter.

Buses, trucks, forklifts and boats, for the same reasons plus duty cycle.

It has taken almost none of the phone and laptop market, where volume is the binding constraint and a thicker battery is not acceptable.

Better than the nickel-cobalt cathodes on extraction and worse on recycling, and both halves should be said.

On extraction it avoids the two genuinely difficult inputs. Cobalt carries the artisanal mining problem in the DRC, including child labour that has been documented for over a decade and has not been solved by supply-chain auditing. Class 1 nickel is mined and refined in ways that include laterite processing with high energy intensity and, in Indonesia, deep-sea and land tailings disposal that is a live controversy. LFP needs neither.

It still needs lithium, and lithium extraction is its own argument — brine operations in the Atacama drawing on water in one of the driest places on earth, and hard-rock spodumene mining with a higher energy footprint.

On recycling it is the weaker material. Hydrometallurgical recycling is economic because the recovered cathode metals are valuable, and LFP's are not: recovering iron phosphate costs more than buying it. Direct cathode-to-cathode recycling, where the material is refurbished rather than dissolved to elements, is the route that makes LFP recycling work and it is not yet at scale. Regulation is currently doing what economics does not — the EU battery regulation sets recycled-content and collection targets regardless of whether the recovered material pays.