1. What Solvent Extraction Is and What It Does
The rare earth separation problem is one of the most technically demanding challenges in industrial chemistry. Seventeen elements that behave almost identically in solution, with ionic radii that differ by less than 10% across the entire series, must be separated to commercial purity levels measured in percentage points or better. The chemistry that accomplishes this at industrial scale is solvent extraction — and the reagents that make it work are the least-discussed link in the rare earth supply chain.
Solvent extraction — SX — is a liquid-liquid separation process. It works by exploiting the different affinities of individual rare earth ions for two immiscible liquid phases: an aqueous phase (a water-based solution containing the dissolved rare earth elements, typically in sulfuric acid from the leaching steps that preceded it) and an organic phase (an organic solvent containing the extractant reagent dissolved at controlled concentration in a hydrocarbon diluent). When the two phases are mixed in a mixer-settler unit (the individual processing vessel in an SX circuit, which contains a mixing zone where the two phases are vigorously agitated to promote transfer of metal ions, followed by a settling zone where they separate by density back into two distinct layers), the extractant selectively binds to certain metal ions based on their chemical properties — ionic charge, ionic radius, coordination chemistry — and transfers them from the aqueous phase into the organic phase.
The selectivity is the entire mechanism. An extractant that has strong affinity for dysprosium but weaker affinity for neodymium will preferentially load dysprosium into the organic phase in each mixer-settler stage. After enough stages in sequence — the number ranges from dozens to several hundred depending on the separation required — the accumulated selectivity difference produces individual rare earth fractions of commercial purity.
The process requires scrubbing and stripping to complete the circuit. After the organic phase is loaded with the target rare earth, scrubbing removes co-extracted impurities. Stripping — washing the scrubbed organic with acid or water — breaks the extractant-metal bond and releases the purified rare earth into a fresh aqueous solution. The stripped organic is recycled. The purified aqueous stream goes to precipitation and calcination to produce the final oxide.
A complete rare earth separation plant may contain hundreds of individual mixer-settler units, each requiring continuous organic phase circulation, reagent replenishment, pH control, temperature management, and analytical monitoring. The process runs 24 hours a day, 365 days a year.
Plain English
Rare earth separation works by mixing two liquids that don't dissolve in each other — a water-based solution containing dissolved rare earths and an oil-based solution containing the extractant. The extractant grabs certain rare earths preferentially and pulls them into the oil phase. Repeat hundreds of times in sequence and the small preference accumulates into a complete separation. The extractant chemical is what makes it work. Without the right extractant, the separation either doesn't run or produces impure product.
2. The Reagent Family
Four reagent families do most of the work in rare earth solvent extraction. Each has different selectivity, different operating conditions, different producers, and different supply chain characteristics.
D2EHPA (di-2-ethylhexyl phosphoric acid, also known by its Chinese industrial designation P204) is the workhorse of rare earth separation. It has strong affinity for heavy rare earths over light rare earths, making it suitable for the initial separation of heavy from light rare earth fractions in mixed feedstreams. It is produced at industrial scale in China — where it is known as P204 and manufactured by multiple Chinese chemical companies at costs that reflect China's integrated organophosphorus chemical industry. Western sources exist but the Chinese production scale and cost position dominate the market.
PC88A (2-ethylhexyl phosphonic acid mono-2-ethylhexyl ester) is produced primarily by Daihachi Chemical Industry in Japan. Daihachi, headquartered in Osaka, is a mid-sized Japanese specialty chemical company that produces PC88A under that trade name for international markets, while Chinese facilities use a domestically produced equivalent designated P507. PC88A has different selectivity from D2EHPA — it separates certain rare earth pairs, particularly within the light rare earth group, with greater efficiency. Lynas's LAMP facility in Malaysia uses PC88A as a key separation reagent, sourced from Daihachi. The thin supply chain for PC88A outside China runs through a single Japanese specialty chemical company that almost nobody outside the rare earth industry knows exists.
Cyanex 272 (bis(2,4,4-trimethylpentyl) phosphinic acid) is produced by Solvay in Belgium under the Cyanex brand. It has different coordination chemistry from D2EHPA and PC88A, giving it different selectivity useful for specific rare earth separations and for cobalt-nickel separation in battery materials processing. Solvay's Cyanex line represents the most accessible Western rare earth reagent supply chain — a large established European specialty chemical company with global distribution capability.
TBP (tributyl phosphate) is a multi-purpose extractant and diluent modifier used across many solvent extraction applications including rare earth processing, uranium extraction, and nuclear fuel reprocessing. TBP is produced by multiple manufacturers globally and has a less concentrated supply chain than D2EHPA and PC88A.
Plain English
Four main extractants run rare earth separation. D2EHPA is the workhorse — mostly Chinese production. PC88A is made primarily by one Japanese company almost nobody has heard of. Cyanex 272 comes from Solvay in Belgium — the most accessible Western option. TBP is more widely available. The Western rare earth separation supply chain runs through Daihachi for PC88A — a single specialty chemical company in Japan whose supply disruption has no quick substitute.
3. Why China Dominates
China's approximately 85–91% share of global rare earth separation capacity (IEA 2024; USGS 2025) is not solely a function of reagent supply. It is the accumulated result of three decades of deliberate industrial development that combined reagent production, process optimization, infrastructure, and workforce skills in a way that is not easily disaggregated or replicated.
The reagent advantage is real. Chinese rare earth separation facilities have preferential access to domestically produced D2EHPA (P204) and PC88A (P507) equivalents from China's integrated organophosphorus chemical industry, at costs that reflect Chinese domestic production economics. The cost of SX reagents is a meaningful fraction of a separation facility's operating costs — sourcing reagents domestically versus importing from Japanese or Belgian specialty chemical companies is a structural advantage that compounds with every operating hour.
But the deeper advantage is not the reagent. It is the process knowledge.
Rare earth solvent extraction is not a recipe that can be read from a textbook and implemented immediately. The specific extractant concentration, the pH of the aqueous phase at each stage, the organic-to-aqueous phase ratio, the number of extraction stages versus scrubbing stages versus stripping stages, the temperature, the diluent selection, the management of third-phase formation — these parameters interact in ways that require extensive empirical optimization for each specific feedstock chemistry. Chinese separation facilities have been optimizing these parameters for specific ore bodies and specific product specifications for thirty or more years.
In December 2023, China made the nature of this advantage explicit: the Chinese government prohibited the export of rare earth extraction, separation, and magnet-making technologies. Not just the materials — the knowledge. The processing flowsheets, the engineering specifications, the operating parameters that make rare earth separation work at commercial quality are now formally restricted from leaving China. China's legal framework has recognized that the process knowledge is as strategically important as the ore or the separated product.
Plain English
China dominates rare earth separation for two reasons. First: cost — Chinese reagents, Chinese energy, Chinese labor. Second, and harder to quantify: thirty years of running these circuits on this chemistry for these ore bodies has produced process knowledge that lives in the people and the data, not in the reagent bottles. In December 2023, China made the second advantage explicit by restricting technology exports. You can source the same reagents. You cannot source the experience. And now, officially, you cannot buy the knowledge either.
4. The Western Separation Problem
Lynas Rare Earths operates LAMP — the Lynas Advanced Materials Plant — near Kuala Lumpur, Malaysia. It is the largest rare earth separation facility outside China and the primary proof of concept that Western rare earth separation is possible at commercial scale.
LAMP's SX circuits use PC88A sourced from Daihachi Chemical in Japan and other specialty reagents from Western suppliers. It has been operating since 2013 and has accumulated more than a decade of operating experience on its specific Mount Weld ore chemistry. Its neodymium and praseodymium product quality meets the specifications of the NdPr mixed oxide that magnet manufacturers require.
LAMP is the exception that demonstrates the difficulty of the rule. Getting to a functioning commercial-quality rare earth separation facility outside China required more than a decade of operating experience, a high-grade ore body with favorable chemistry, substantial government support, and continuous operational refinement. Lynas's experience demonstrates that Western separation is achievable — it does not suggest it is easy or fast.
The new Western separation facilities — Energy Fuels' White Mesa Mill in Utah, USA Rare Earth's planned refinery, the ReElement Technologies facility in Indiana — are at earlier stages of operational development. They are acquiring the reagent supply chains, the process engineering, and the operating experience that Lynas built over more than a decade. The DFARS January 1, 2027 deadline requires DFARS-compliant rare earth separation. The facilities required to produce it are still accumulating the process knowledge that Chinese facilities built over thirty years.
For these new Western facilities, the PC88A supply chain runs through Daihachi. A facility running PC88A is dependent on Daihachi's supply reliability. The Western separation buildout requires simultaneously developing the reagent supply chain and the process knowledge — two parallel challenges that Chinese facilities resolved decades ago.
Plain English
Lynas is the only proof that Western rare earth separation works at commercial scale outside China. It took more than a decade to get there. The new Western facilities being built for DFARS compliance are earlier in that learning curve. They need the reagents and the process knowledge simultaneously. The PC88A supply chain runs through one Japanese specialty chemical company. The process knowledge takes time the DFARS deadline doesn't allow.
5. The Knowledge Layer
The most important thing about solvent extraction reagents is not the reagents themselves. It is what they reveal about the nature of China's competitive advantage in rare earth processing.
The rare earth supply chain conversation focuses on ore bodies — who controls the mining. It sometimes extends to processing capacity — who has the separation plants. It rarely reaches the level of the reagent chemistry — what specific extractants, at what concentrations, run in what configuration, optimized over how many years of operation for what specific feedstock chemistry.
The reagent layer is where the tacit knowledge lives. D2EHPA does not come with instructions for separating dysprosium from holmium at 99.9% purity from a Jiangxi ionic clay feedstock. PC88A does not come with the optimized pH profile for separating NdPr from the lanthanide mixture in an Australian hard rock ore concentrate. Those parameters were determined empirically, refined over years, and are embedded in process control systems and operator training programs in Chinese facilities.
China's December 2023 technology export restrictions formalized this reality. By prohibiting the export of rare earth extraction and separation technologies — flowsheets, process specifications, operating parameters — China drew a legal boundary around the knowledge layer that its processing facilities have built. The material export controls of April 2025 restricted the physical supply chain. The technology export restrictions of December 2023 restricted the knowledge supply chain. Both are in force simultaneously.
The binding constraints on Western rare earth separation are therefore multiple and parallel: reagent supply, process knowledge, analytical capability, and scale. The reagent supply chain — thin but real, running through Daihachi and Solvay — is necessary but not sufficient. The knowledge layer is the constraint that is hardest to acknowledge because it cannot be solved with money alone and cannot be transferred through a purchase order. It requires time. And the DFARS clock does not wait.
Plain English
You can buy the reagents. You can build the equipment. You cannot buy thirty years of optimization. China made this explicit in December 2023 by restricting technology exports — not just materials, but the knowledge of how to process them. The reagent is available. The process knowledge that makes it work takes years to accumulate. Western facilities are building both simultaneously against a January 2027 deadline. The knowledge layer is the constraint nobody talks about because it is the hardest to solve.
6. Why It Belongs in the Reagent Layer
The Reagent Layer documents the chemistry beneath the mineral. Sulfuric acid dissolves the ore. Hydrofluoric acid cleans the silicon. Sodium cyanide dissolves the gold. Solvent extraction reagents separate the elements.
This page is the deepest layer in the stack. Sulfuric acid, hydrofluoric acid, and sodium cyanide are commodities — they can be priced, sourced from multiple suppliers, and quantified on a standard unit basis. Solvent extraction reagents are specialty chemicals — contract-negotiated, produced by a handful of companies, optimized for specific applications over years of use. There is no public price index. There is no spot market. The supply status indicator replaces the price card because no price card exists. That absence is itself information.
Commodity chemicals are priced by markets because buyers and sellers are numerous and trade is frequent. Specialty chemicals are priced by negotiation because the buyer-seller relationships are long-term, the products are application-specific, and switching suppliers requires process reformulation. The fact that SX reagents are contract-negotiated reflects the depth of the dependency between separation facilities and their specific reagent suppliers.
The DFARS January 1, 2027 deadline requires rare earth magnets whose supply chain is China-free all the way through. That chain runs through ore, through leaching, through solvent extraction, through precipitation, through metal production, through alloying, through sintering. The solvent extraction step sits in the middle. The reagents that enable it are the chemistry beneath the chemistry.
Solvent extraction reagents belong in the Reagent Layer because they are the invisible chemistry inside the visible supply chain. You can see the mine, the separation plant, the magnet factory, and the defense system. You cannot see the organic phase cycling through hundreds of mixer-settler stages at LAMP or at Chinese facilities that run 85–91% of global rare earth separation. But it is happening in every facility that produces the separated oxides that go into every NdFeB magnet whose supply chain is covered on this site.
Plain English
Sulfuric acid, HF, and sodium cyanide can be priced on a market. SX reagents cannot — specialty chemicals priced by negotiation between facilities and suppliers. The absence of a price is information about the market structure. The chemistry that separates neodymium from praseodymium runs on these reagents. China restricted their technology export in December 2023. You can see the mine and the magnet. You cannot see the organic phase cycling through the mixer-settlers between them. This page exists to make that invisible layer visible.
7. Oxalic Acid — The Final Precipitation Reagent
Solvent extraction separates rare earth elements from each other. What it produces is not a finished oxide — it is a loaded organic phase (a liquid mixture where the target rare earth has been chemically transferred from the water-based feed solution into an oil-based solvent) and, after stripping, a concentrated aqueous rare earth solution. That solution still has to be converted into a solid before calcination (high-temperature roasting that drives off remaining chemistry and produces the final oxide powder) can complete the process. The conversion step is precipitation — and the reagent that drives it is oxalic acid.
Oxalic acid works by reacting with dissolved rare earth ions to form rare earth oxalates (insoluble solid compounds that fall out of solution as a fine precipitate). The oxalates are then filtered, washed, and fed into the calcination furnace, where heat converts them to rare earth oxides at greater than 99.9% purity. Without precipitation, there is no solid. Without a solid, there is no oxide. Without an oxide, there is no magnet precursor, no catalyst, no phosphor feedstock.
The dependency has three characteristics that make it structurally significant.
First, oxalic acid is consumed, not recovered. Unlike some reagents that can be regenerated and recirculated through the process, oxalic acid is destroyed in the precipitation reaction. Every tonne of separated rare earth oxide produced requires a fresh purchase of oxalic acid. There is no internal loop. The consumption is continuous and proportional to output.
Second, the industrial-scale supply is predominantly Chinese. China accounts for the majority of global oxalic acid production, and the merchant market for the volumes required by a serious rare earth processing facility — not laboratory quantities, but tonnes per month at sustained output — is effectively a Chinese supply chain. No meaningful Western merchant market exists at processing scale.
Third, it was not on the supply chain risk map. The policy and investment focus on rare earth reagent dependency has concentrated on the solvent extraction extractants — D2EHPA, PC88A, Cyanex 272, TBP — because those are the reagents that perform the separation chemistry. Oxalic acid came after. It was assumed. It was purchased. It was never flagged.
The result is a structural vulnerability that survived every investment in Western rare earth processing infrastructure. A facility that has secured its solvent extraction reagents — that has solved D2EHPA sourcing, that has qualified PC88A alternatives, that has locked in TBP supply — but has not addressed oxalic acid sourcing has not actually solved the reagent problem. It has solved the front end and left the back end exposed.
In June 2026, Nth Cycle (an MIT-derived electro-extraction company) and Ionic Technologies (Belfast) signed a Joint Development and Licensing Agreement integrating Nth Cycle's electrochemical extraction technology into the Ionic MAIL long-loop recycling flowsheet. The specific target: replacing the oxalic acid precipitation step with an electrochemical alternative that requires no reagent addition and generates no chemical waste stream. The integration is targeted for Q4 2026. If it commissions as planned, the Ionic Belfast facility will become the first Western rare earth recycling operation to carry zero Chinese reagent dependency at any processing stage.
The elimination of the oxalic acid dependency is not a marginal process improvement. It closes the last reagent chokepoint in a flowsheet that had solved every other one. The licensing architecture — explicitly targeting US deployment and international replication — means the upgraded flowsheet becomes the template for future facilities, not a one-site fix.
Plain English
Solvent extraction does the separation. Oxalic acid turns the result into a solid. No solid, no oxide. No oxide, no magnet. The acid is consumed every cycle, cannot be recovered, and comes almost entirely from China. A processing facility that solved every other reagent dependency but not this one has not solved the reagent problem. It has just moved the chokepoint one step further downstream — to the step nobody was watching.