Osmond Resources — homeASX: OSM · Orión Project

Minerals Hub / Mining & Production / Why separating rare earths from each other takes so many stages

Mining & Production · Monazite · 5 min read

Why separating rare earths from each other takes so many stages

Some neighbouring rare earths differ so slightly that one solvent-extraction stage parts them only a little, and a small separation factor between a pair means a circuit needs many stages.

Reviewed by Peter Uppal

A long row of glass separating funnels in a metal rack, each holding two liquid layers, the lower drifting from pale green through yellow and peach to rose.
Illustrative artwork: a long row of two-layer separations, each changing the mixture only a little. Not a facility, equipment or material connected to this project. · Illustration · Osmond Hub

The short version

Rare earths are hard to separate from one another because they differ only subtly in ionic radius and coordination number. A separation factor measures how well two of them part in an extraction, and with the extractant P507 some neighbouring pairs, including neodymium and praseodymium, have factors small enough that reaching high purity requires many equilibrium stages. A 2023 paper in Nature says conventional solvent-extraction circuits require dozens or even hundreds of stages to produce high-purity individual rare-earth oxides.

Neighbours that barely differ

The difficulty starts in the atoms. A 2023 paper in Nature states that rare earths are hard to separate because they differ only subtly in ionic radius and coordination number(opens in a new tab). The same paper, citing radii published by Shannon in 1976, puts the drop in trivalent ionic radius from lanthanum to lutetium at only about 0.19 Å(opens in a new tab).

Other authors say the same thing in chemical terms. A 2023 study of extraction columns describes the separation of individual rare earths as hard because their chemical properties are remarkably similar(opens in a new tab). A 2024 study of separating heavy rare earths, as that study uses the term, attributes the challenge to their closely related electronic structures(opens in a new tab).

What a separation factor measures

Two glass vessels on a bench, each with a blue layer over a copper layer: on the left purple and ivory spheres are evenly mixed through both layers; on the right the purple spheres sit almost all in the upper layer and the ivory ones in the lower.
Illustrative diagram: two rare earths shared almost identically between two liquid layers (left), as for Nd/Pr, whose separation factor with the extractant P507 is small, and clearly parted (right), as for Tm/Dy, where a 2024 study using TODGA found a factor of 41; two different extractant systems, not one experiment. Not a facility, equipment or material connected to this project. · Illustration · Osmond Hub

The separation factor (β) between two rare earths is the ratio of their distribution ratios, and characterises how well they separate(opens in a new tab); a factor close to one means two elements distribute almost identically.

The 2023 column study names the pairs where that matters: for some pairs, including Nd/Pr, Gd/Eu and Lu/Yb, separation factors with the extractant P507 are small, which requires many equilibrium extraction stages to reach high purity(opens in a new tab). A 2024 paper sharing an author with that study notes that sufficiently high separation factors between adjacent elements are observed in P507 extraction systems(opens in a new tab); on this publication's reading, the difficulty is pair by pair.

One model makes the cost of a small factor concrete. In a simulated P507 flowsheet in hydrochloric acid, targeting 99.99% praseodymium and neodymium, the Pr/Nd separation required 62 equilibrium stages at a separation factor of 1.70, while simpler bulk splits needed as few as 16 stages(opens in a new tab). In a different set of calculations, using the extractant TODGA, a 2024 study found cascades of 11 to 18 stages would separate dysprosium, erbium, thulium and ytterbium to about 99.2–99.8% purity, with separation factors at 5.0 mol/L HCl ranging from 3.9 (Tm/Yb) to 41 (Tm/Dy)(opens in a new tab).

The 2023 Nature paper gives the scale for whole circuits. It says conventional solvent-extraction circuits require dozens or even hundreds of stages to produce high-purity individual rare-earth oxides(opens in a new tab).

A small separation factor is paid for in stages.

Mixer-settlers in series

Six stirred glass tanks in a row, each with a blue layer over a copper layer; arrows enter top left and leave top right, enter bottom right and leave bottom left; the blue deepens and the copper lightens from left to right.
Illustrative diagram: six counter-current mixer-settler stages of the kind run in a 2015 pilot; industrial circuits can need dozens or hundreds. Not a facility, equipment or material connected to this project. · Illustration · Osmond Hub

Solvent extraction is the most common industrial method for separating rare earths(opens in a new tab), according to the 2023 column study, which names extraction columns and cascades of series-connected mixer-settlers as the industrial equipment, with productivity greater by several orders of magnitude than chromatographic devices(opens in a new tab), a point it makes about extraction apparatus in general. The extracting phase is organic: conventional rare-earth solvent extraction uses organic solvents such as kerosene and phosphonate extractants(opens in a new tab), and P507 is widely used in the Chinese rare-earth industry(opens in a new tab).

A small pilot shows the arrangement. In a 2015 study using the extractant DEHPA, the continuous tests were run counter-current in a sequence of mixer-settler stages, using six extraction stages(opens in a new tab), in a micro continuous pilot plant. The same study found that the Pr/La and Nd/Pr pairs gave small separation factors, which the authors describe as showing how hard these pairs are to separate, in particular for Nd/Pr(opens in a new tab).

Mixtures of extractants are one line of laboratory work. In a 2024 laboratory study using a 0.9 M equimolar mixture of P507 and Cyanex 272 in chloride solution, the dysprosium/terbium separation factor reached 3.42 at one feed concentration, and ranged from about 1.8 to 3.4 across those tested(opens in a new tab).

No single order of cuts

Which elements count as heavy depends on who is counting. USGS traditionally defines the heavy rare earths, on the basis of atomic weight, as terbium through lutetium (atomic numbers 65–71), and includes yttrium with them(opens in a new tab). The simulated praseodymium–neodymium flowsheet draws its light/heavy line elsewhere, between neodymium and samarium(opens in a new tab), so samarium to gadolinium, which fall outside USGS's heavy group, count as heavy in that flowsheet.

Flowsheets also start in different places. That simulated flowsheet, for liquor from ion-adsorption clay at Jeli in Malaysia, is a four-train design that starts with a bulk light/heavy (LREE/HREE) separation and ends with the Pr/Nd separation, which the authors call the challenging step(opens in a new tab). The 2024 P507–Cyanex 272 study proposes a flowsheet for what it calls "medium-heavy" rare earths, calculated from a material balance, whose separation order ran Er/Ho, Ho/Dy, Dy/Tb, Tb/Gd and then Er/Y(opens in a new tab). That proposed design uses five cascades, the first with 25 extraction stages and 15 scrubbing stages, and another with 50 extraction and 30 scrubbing stages(opens in a new tab).

Capacity and alternatives

The IEA's Global Critical Minerals Outlook 2026 puts China's current share of magnet rare-earth separation at 85%(opens in a new tab). On this publication's reading, circuits of many stages are one reason a separation plant is a large undertaking.

The 2023 Nature paper names the inefficiency and "large environmental impact" of rare-earth separations as drivers for research into alternatives(opens in a new tab). In written testimony to the US House Science Committee, an Ames Laboratory scientist said that Ames Laboratory scientists developed the ion-exchange process, which was still used to produce the highest-purity individual elements, while solvent-extraction processes were used for large-scale separation(opens in a new tab). Counter-current chromatography, the 2023 column study notes, has been used mainly for analytical and preparative purposes (its example is herbal pharmaceutical products); its low productivity has prevented use in large-capacity industrial production, though work aims to scale it with conventional extraction equipment(opens in a new tab).

The 2023 Nature paper's own route was at laboratory stage: a lanmodulin protein dimer gave Nd/Dy separation factors of about 8 and 13 (two variants)(opens in a new tab). In a single laboratory column stage in that 2023 study, the unmodified dimer upgraded dysprosium in a model e-waste mixture from 5% to 83% purity, as a share of the Nd/Dy mixture, while recovering neodymium at 99.8% purity(opens in a new tab).

Related

  • Processing Technology — the equipment a separation circuit is built from
  • Ore Processing — the concentrate a separation plant receives
  • Processing Capacity — how separation plants are counted
  • Supply Chain Risk — what concentration at one stage means
  • New Processing Technologies — routes still in development

Sources