Neodymium Nd · 60

The rare earth at the heart of the strongest magnet ever made — the kind in headphones, hard drives, wind turbines and electric motors.

Why does it matter?

An NdFeB magnet holds roughly a thousand times more magnetic energy than the iron magnets of a century ago, which is why an EV motor can be small enough to sit inside a wheel arch.

Where it is in the Earth

Neodymium belongs to the group of elements called the rare earth elements (REEs) — a set of seventeen metals that, despite the name, are not especially scarce in the Earth's crust. What is scarce is a geological process that concentrates them into one place at grades high enough to mine economically. Neodymium is found in two principal ore minerals: bastnäsite, a fluorocarbonate mineral, and monazite, a phosphate. Both carry a mixture of rare earths together — cerium and lanthanum typically dominate by mass, with neodymium making up a smaller but commercially important fraction of the whole.

The most significant concentrations of these minerals occur in carbonatites, a rare type of igneous rock formed from magma that is unusually rich in carbonate. Carbonatite magmas are thought to originate deep in the mantle and rise through the crust along ancient fault systems, crystallising into bodies that can be extraordinarily enriched in REEs relative to ordinary rock. The deposits at Bayan Obo in China and Mountain Pass in the United States are both carbonatite-hosted, as is the ore body at Mount Weld in Australia, though at Mount Weld the original carbonatite has been deeply weathered over geological time, producing a residual concentration of REE minerals near the surface. This weathering process — where soluble minerals dissolve away and leave behind the more resistant phosphates and carbonates — can actually raise the grade of a deposit compared with the fresh rock beneath.

A second, geologically distinct source is the ion-adsorption clay deposit, found principally in southern China. In these deposits, REEs released by the weathering of granite have been adsorbed — that is, loosely bound — onto the surfaces of clay minerals rather than incorporated into a discrete ore mineral. These clays are particularly enriched in the heavier rare earths, whereas carbonatite deposits like Bayan Obo tend to be skewed toward the lighter REEs, of which neodymium is one. Monazite also appears as a heavy mineral in beach and river placer sands, where wave action or river currents sort and concentrate dense grains. All of these deposit types reflect the same underlying principle: neodymium does not form deposits on its own, but travels with its chemical relatives and must be separated from them later in the processing chain.

Getting it out

All three of the principal neodymium-producing mines listed — Bayan Obo, Mountain Pass, and Mount Weld — are worked as open-pit operations. Open-pit mining suits these deposits because the ore bodies are large and relatively near the surface, meaning the economics of stripping away overlying waste rock (the overburden) are better than trying to follow ore underground. In an open pit, the ground is drilled and blasted in horizontal layers called benches, then loaded onto haul trucks and carried to either a processing plant or a waste dump. The proportion of waste rock moved for every tonne of ore recovered — the strip ratio — varies by deposit and is not given in the available data, but at carbonatite deposits the transition from rich ore to barren rock can be sharp enough to keep strip ratios manageable.

Grade matters enormously in rare earth mining, though it must be read carefully. Published figures are typically expressed as a percentage of total rare earth oxide (TREO) in the ore, but neodymium makes up only a portion of that total — the rest is cerium, lanthanum, praseodymium, and others. A mine with an attractive headline TREO grade might yield relatively little neodymium if its ore is dominated by the less commercially valuable cerium and lanthanum. Bayan Obo is the world's largest source by volume, but it is also an iron ore mine where rare earths are recovered as a co-product of steel production; the economics of that operation are therefore partly insulated from rare earth prices in a way that standalone REE mines are not.

Ion-adsorption clay deposits in southern China are mined very differently. Because the REEs are held loosely on clay surfaces rather than locked inside hard mineral crystals, the ore does not need to be crushed and milled. Instead, a solution — historically ammonium sulfate, increasingly alternatives with lower environmental impact — is injected into the clay in place, the REEs are leached out of the ground, and the pregnant solution is collected at the surface. This in-situ leaching method uses less energy than conventional crushing and flotation, but it has historically caused significant landscape disturbance and water contamination in producing regions.

What pulls on it

Neodymium is pulled out of the ground primarily because it is the key ingredient in neodymium-iron-boron (NdFeB) permanent magnets — the strongest type of permanent magnet made commercially. A permanent magnet holds its magnetism without any ongoing power supply, which makes it far more efficient in a motor or generator than an electromagnet that must be continuously energised. The material-intensity figures in the table illustrate how this plays out: a direct-drive offshore wind turbine, which uses a permanent-magnet generator rather than a gearbox, requires between 100 and 200 kilograms of neodymium per megawatt of generating capacity. That is a large quantity relative to the weight of the turbine's electrical output, and it is why wind power is one of the most discussed end markets.

Electric vehicle traction motors account for the other major growth conversation. A traction motor using NdFeB magnets in its rotor contains between 0.4 and 1.0 kilograms of neodymium per motor. Not all electric vehicles use this architecture — some manufacturers use induction motors or separately excited motors that contain no rare earth magnets at all — but permanent-magnet motors are generally more compact and efficient, which is why they are common in passenger car drivetrains where space and energy density matter. Consumer electronics, robotics and data-centre drives also consume neodymium, though typically in smaller magnets per unit. The USGS notes catalysts and lasers as additional applications; these draw on neodymium's optical and chemical properties rather than its magnetic ones, though magnets dominate by volume.

Demand could shift sharply in either direction under a few conditions. A durable move away from permanent-magnet motors in electric vehicles — driven by cost, supply concern, or engineering preference — would reduce the growth trajectory considerably, though it would not eliminate demand from existing uses. Conversely, faster-than-expected deployment of direct-drive offshore wind, or expansion of industrial robotics, would pull harder on supply. The price history in the table — running from $98 per kilogram in 2021 to a peak of $134 in 2022, then declining to $56 in 2024 before recovering partially to $73 in 2025 — reflects how sensitive the market is to shifts in anticipated demand from the EV sector and to changes in Chinese production policy.

Turning ore into product Level 3

The gap between run-of-mine ore and a neodymium product that a magnet factory can use is wide, and crossing it requires a sequence of distinct industrial steps. At the mine, ore is first crushed and ground — a stage called comminution — to liberate the individual mineral grains from the surrounding rock. The resulting fine material is then concentrated, most commonly by froth flotation for bastnäsite: chemical reagents are added to a water-and-ore slurry so that the target mineral grains attach to air bubbles and float to the surface as a froth, while gangue (waste) minerals sink. Mount Weld ore is concentrated in Australia before being shipped to the Lynas Advanced Materials Plant in Kuantan, Malaysia, for separation; Mountain Pass concentrates and separates on site. The output at this stage is a mixed rare earth concentrate, not yet separated by element.

Separation of the individual rare earths from the concentrate is the technically demanding and capital-intensive heart of the chain. The dominant method is solvent extraction, also called liquid-liquid extraction. The mixed rare earth feed is dissolved in acid, then contacted repeatedly with an organic solvent that preferentially picks up certain elements over others. Because the rare earths are chemically very similar to one another — they sit next to each other in the periodic table and behave almost identically in solution — achieving high purity requires many extraction stages in series, in large banks of mixer-settler units. The output is separated neodymium oxide, typically reported on a neodymium oxide (Nd₂O₃) equivalent basis. China hosts by far the largest separation capacity; Neo Performance Materials in Estonia and Lynas in Malaysia represent the most significant separation capacity outside China, though the data does not break down their individual throughputs.

Beyond separation, the oxide must be reduced to metal — most commonly by molten salt electrolysis or metallothermic reduction — then alloyed with iron and boron to produce NdFeB alloy, cast into strip or powder, pressed and sintered into magnet blanks, and coated to resist corrosion. Each step introduces yield losses and adds cost. The traded form therefore progresses from oxide to metal to alloy to finished sintered magnet, with value added at each stage. China's integrated magnet cluster dominates sintered magnet production, which means that even material separated outside China typically re-enters the Chinese industrial system for final magnet manufacture. Losses through the chain are real but not uniformly reported, and figures in the public literature should be treated with caution unless they are traceable to measured recovery data.

Substitution and recycling Level 3

Within the NdFeB magnet family, praseodymium can substitute for a portion of the neodymium content, and in practice commercial alloys are often specified as NdPr (neodymium-praseodymium) to reflect this interchangeability; the two elements are sometimes sold as a mixed oxide for this reason. Dysprosium or terbium — both heavy rare earths — are added in small quantities to NdFeB magnets intended for high-temperature service, such as in traction motors, because they raise the coercivity (the resistance to demagnetisation) of the magnet. These additions do not replace neodymium but they do create separate dependency chains of their own.

At the level of the magnet itself, the main alternatives to NdFeB are samarium-cobalt (SmCo) magnets, which perform better at very high temperatures but cost considerably more and depend on cobalt, and ferrite (iron oxide) magnets, which are cheap and widely used but carry far less magnetic energy per unit volume. An EV motor redesigned around ferrite magnets would need to be physically larger and heavier to deliver the same output, which is why ferrite is not straightforwardly interchangeable for traction applications. At the motor architecture level, as noted, induction and wound-field synchronous motors avoid rare earths entirely, at some efficiency and packaging cost — a trade-off that different manufacturers have assessed differently.

Recycling of neodymium from end-of-life products is technically possible but currently recovers only a small fraction of annual consumption. The main barriers are practical: NdFeB magnets are often glued, coated, or encapsulated inside assemblies in ways that make disassembly difficult and costly; the volumes of magnets in individual consumer products are small; collection infrastructure for magnet-bearing waste is sparse outside a few specialist streams; and the price of neodymium, while volatile, has not consistently reached levels that make secondary recovery economically attractive compared with primary production. Some direct magnet-to-magnet recycling routes — recovering the alloy without going back to oxide — are in development at pilot scale, but have not yet reached commercial throughput.