Separation (Membrane)

Membrane rare earth separation uses a membrane to pull rare earth ions out of a solution, or to hold them back while other ions pass. The main approaches are supported liquid membranes, membrane solvent extraction, ligand-functionalized membranes and nanofiltration. Independent sources place all of them at the research, pilot or demonstration stage. The International Energy Agency says separation circuits rely predominantly on liquid-liquid solvent extraction. Among ways to cut energy use and emissions in refining, it lists improving separation efficiency through advances in membrane technologies (IEA, May 2026).

How membrane rare earth separation works

Membranes compete for the same job as the mixer-settler trains described under rare earth separation and the resin and contactor columns covered under column-based separation. A 2022 review in Membranes groups membrane methods by mechanism. The membrane can extract rare earths and carry them across, capture them on sites built into the membrane, or reject them while other ions pass through (Bashiri et al., 2022).

Supported liquid membranes and membrane solvent extraction

In a supported liquid membrane, a porous polymer is soaked with an organic liquid that carries the extractant. A 2023 review describes the process as three simultaneous steps: extraction from the feed into the membrane, diffusion across it and stripping into a receiving solution. The review credits liquid membranes with using only small amounts of hazardous reagents (Kaczorowska, October 2023).

Membrane solvent extraction runs this idea in hollow-fiber modules. Researchers at Oak Ridge National Laboratory, part of the DOE's Critical Materials Innovation Hub, dissolved end-of-life products and scrap and passed the solutions through patented modules holding bundles of hollow fibers that let only certain metal ions through, the DOE says (DOE, September 2024).

Stability is the recurring weak point. The 2023 review says supported and emulsion liquid membranes often show low stability in use. Citing the KTH team below, it adds that hollow-fiber supported liquid membranes had not been implemented industrially, mainly because of membrane instability, difficult scale-up and short module life (Kaczorowska). A Swedish team from KTH Royal Institute of Technology and IVL tested three hollow-fiber setups at pilot scale, in a module with 8 square meters of membrane area, on a synthetic feed based on recovering rare earths from apatite concentrate. Performance of the standard supported liquid membrane decayed rapidly. Two other configurations, renewal liquid membrane and emulsion pertraction, were much more stable, and the authors concluded they may be feasible for processing rare earth liquors (Alemrajabi et al., December 2022).

Ligand-functionalized membranes and nanofiltration

Ligand-functionalized membranes build the selectivity into the membrane material. A 2025 viewpoint from the University of Texas at Austin, funded by the DOE Office of Science, describes embedding molecular recognition in polymer membranes for ion separations in general, because conventional membranes struggle to tell chemically similar ions apart in high-ionic-strength solutions (Ganesan and Zofchak, October 2025). In April 2025, the university reported lab tests of artificial membrane channels built from a modified pillararene molecule. They showed a 40-fold preference for europium over lanthanum and a 30-fold preference for europium over ytterbium (UT Austin, April 2025).

Nanofiltration holds back rare earth ions mainly by pore size and surface charge. A 2026 review by researchers in Sichuan, China, funded by Sichuan provincial agencies, found that for commercial nanofiltration membranes, the separation factor between adjacent rare earths generally does not exceed two, too low for high-purity separation. Adding citric acid as a complexing agent under mildly acidic conditions raised separation factors by more than an order of magnitude compared with operation without it (Feng et al., August 2026). REEx concluded that nanofiltration may pre-concentrate rare earths before conventional solvent extraction but is not yet a replacement for China's mature processing systems (REEx, August 2026).

How ready is membrane separation?

The IEA's description of today's separation plants names ion exchange, selective precipitation, distillation and solid-liquid separation as complementary steps to solvent extraction. It does not list membranes among them (IEA).

The 2026 nanofiltration review says long-term operating data on real leachates are extremely scarce. Recent pilot studies showed that commercial membrane elements can be adapted for submerged nanofiltration, but continuous runs of hundreds to thousands of hours on real rare earth leachates are still needed (Feng et al.). REEx reported in October 2026 that no alternative technology, membranes included, has yet displaced solvent extraction as the dominant industrial method (REEx, October 2026).

Who is developing membrane separation (as of October 2026)

Momentum Technologies, Carrollton, Texas. The DOE says Momentum licensed ORNL's membrane solvent extraction process for recovering rare earths and battery materials, and that ORNL and Momentum received follow-on funding under the Bipartisan Infrastructure Law to commercialize it (DOE). In October 2025, Momentum said it had begun processing rare earth waste from magnet swarf and mining tailings at commercial scale at its Carrollton demonstration plant (Momentum, October 2025). In June 2026, the company reported 99.9% pure NdPr oxide from e-waste and magnet production waste, and 99.5% purity for dysprosium and yttrium, all from feedstocks processed at the Carrollton demonstration facility (Momentum, June 29, 2026). In October 2026, Momentum said its first commercial rare earth line at Carrollton, which refines recycled hard-disk-drive feedstock under a tolling agreement, was operational and ramping up, and that it was developing a second commercial-scale plant there ahead of a final investment decision. That release gave different purity figures: 99.5% for NdPr oxide and 99.9% for yttrium (Momentum, October 1, 2026). The commercial-scale descriptions and the purity figures are company-reported.

Research groups. The other named work comes from universities and research institutes: the KTH/IVL pilot, the UT Austin channels and the Sichuan review.

Frequently asked questions