Ion-imprinted phosphonate-bis-picolinamide resin for dysprosium and terbium separation from magnet leachate

Heavy rare earths — dysprosium and terbium in particular — sit at the center of the energy transition's most acute supply bottleneck. Both metals are essential dopants in NdFeB permanent magnets used in EV traction motors and wind turbine generators, yet their recovery from end-of-life magnets is nearly non-existent at commercial scale. The primary obstacle is not leaching chemistry but the extraordinarily difficult separations step: once a spent magnet is dissolved, dysprosium and terbium must be partitioned away from a complex soup of neodymium, praseodymium, iron, boron, and trace contaminants. Conventional solvent extraction achieves this only through long, capital-intensive cascades with large organic solvent inventories — a configuration that raises cost, safety, and environmental permitting burdens that few recyclers can bear. This asset addresses that bottleneck directly. The invention is a heavy-rare-earth-selective ion-imprinted polymer resin built around a structurally defined phosphonate-bis-picolinamide ligand (BPDPA-P), where the polymer matrix is formed in the presence of a Dy or Tb template ion to create a cavity that preferentially recalls the heavy-lanthanide ionic radius and coordination geometry. Computational prediction yields a single-pass Dy/Tb separation factor in the range of 4 to 8 and a Dy/Nd separation factor of 30 to 100 — if experimentally confirmed, either figure would represent a step change versus D2EHPA or PC-88A solvent extraction in batch mode. The composition-plus-device-use claim structure is designed to protect the specific ligand architecture and the imprinted device as a unit, not ligand chemistry in the abstract. Timing matters here. Geopolitical supply pressure on heavy rare earths has accelerated recycler investment globally, while domestic-content rules in the United States and Europe are creating regulatory forcing functions that favor closed-loop magnet recycling. A resin that compresses a multi-stage SX cascade into a compact column operation fits exactly the capital profile that emerging magnet recyclers are targeting. The asset is a lead position in the critical-mineral recovery and recycling separations portfolio, not a defensive filing — its claim scope is built around a specific, patentable ligand structure that cannot be designed around simply by adjusting process parameters.

The core material is designated BPDPA-P: a phosphonate-bis-picolinamide ligand in which two picolinamide arms are bridged and functionalized with a phosphonate group. The canonical connectivity of this ligand is recited in the patent claims by SMILES string, and the claims expressly disclaim the broader "BPDPA" topology where connectivity is ambiguous or incorrect — a precision that is both a prosecution choice and a freedom-to-operate move. The ligand design is deliberate: picolinamide arms provide nitrogen-donor coordination to lanthanide ions in a geometry that differentiates the slightly smaller ionic radii of dysprosium (0.912 Å in 8-coordinate geometry) and terbium (0.923 Å) from the larger neodymium (0.983 Å), while the phosphonate group supplies an additional hard-donor oxygen that tightens overall binding and can anchor the ligand to the polymer backbone. The ion-imprinting step — polymerizing around a Dy or Tb template — then enforces a cavity shape that disfavors the slightly larger lanthanides even before ligand coordination selectivity acts, creating a two-layer selectivity mechanism. The claimed family encompasses BPDPA-P as the lead structure and extends to a small set of related architectures: bipyridine-2,2'-dicarboxamide-phosphonate, 1,3-diaminopropane-bis(picolinamide-phosphonate), and a diglycolamide-phosphonate hybrid. These variants are protected as a structural family, not as a broad functional claimed sweep. The phenanthroline-diamide backbone — a known competing ligand topology used in actinide separations — is expressly excluded from claim scope, providing a clean boundary between this IP and prior separations chemistry. Selectivity predictions in the 4–8 range for Dy/Tb and 30–100 for Dy/Nd are grounded in lanthanide-ligand binding calculations from sparse-matrix simulation exercises modeled after the real multicomponent leachate environment, supplemented by phosphate-phase phonon stability computations that confirm the relevant crystallographic reference phases are dynamically accessible under processing conditions. Because the target materials are coordination polymers and molecular ligand systems rather than extended inorganic crystalline solids, the standard machine-learning interatomic potential (MLIP) validation pipeline used elsewhere in the portfolio — multi-potential consensus phonon calculations with MACE, CHGNet, MatterSim, and ORB — is not applicable to the BPDPA-P ligand itself. This is an honest feature of the asset, not a gap: MLIP methods are parameterized for periodic inorganic materials and do not transfer reliably to 4f-metal coordination chemistry with flexible organic backbones. The phonon stability work (WE35A) applies to rare earth phosphate reference phases, confirming that the thermodynamic ground states used to benchmark binding energetics are themselves well-characterized. The binding-selectivity calculations are the operative proof of concept, and they carry a known limitation: the binding matrix was computed using a lanthanide potential that is not fully self-consistent for 4f electrons, and a corrected calculation using 4f-in-core effective core potentials (ECPs) is an explicit open validation gate. This is disclosed candidly — the predicted separation factors should be treated as directionally reliable but requiring experimental confirmation before design-to-spec use. Key target properties, if bench results match prediction: Dy/Tb single-pass separation factor of 4–8 (substantially above the ~1.5–2.5 achievable with conventional extractants in a single contact), Dy/Nd separation factor of 30–100 (enabling near-complete Nd/Dy partitioning in a short column), and meaningful reduction in organic solvent volume relative to a conventional SX train of comparable separation power. The resin format enables column operations with simple aqueous stripping, which eliminates the need for organic phase disengagement equipment and simplifies waste streams.

The addressable market for this technology sits at the intersection of two converging trends: the rapid growth of NdFeB magnet recycling capacity, and tightening regulatory and supply-chain pressure on heavy rare earth sourcing. The global rare earth separations market has historically been dominated by primary mining and hydrometallurgical processing in China, but the recycled magnet stream is growing as EV fleets age, wind turbine retrofits accelerate, and manufacturers face domestic-content obligations. Estimates for the heavy rare earth recycling separations opportunity range from $0.5 billion to $2 billion in addressable revenue as the segment matures — these are estimates, not audited figures, and they are sensitive to how quickly closed-loop magnet recycling scales outside China. The buyers in this market are not commodity chemical companies; they are magnet recyclers building greenfield or retrofit separation facilities. These companies are capitally constrained relative to primary REE processors, which makes the resin's proposition particularly relevant: replacing or shortening a solvent extraction cascade with a column-based solid-phase system reduces both upfront capital (fewer mixer-settlers or pulse columns, less solvent inventory, simpler safety and environmental engineering) and operating cost (lower solvent makeup, simpler phase management). For a recycler processing on the order of hundreds to low thousands of tonnes of magnet per year, this difference can be decisive for project economics. The royalty or licensing logic follows accordingly — a per-kilogram royalty on recovered Dy or Tb, or an upfront license tied to column capacity, would capture value proportionally to the scale at which the technology is deployed. The race window is real but not unlimited. Heavy rare earth supply pressure from geopolitical dynamics has created urgency in the mid-2020s, with significant government-backed investment in domestic recycling capacity in the U.S., Japan, and the EU. Recyclers building new facilities in the next two to four years will make technology selection decisions that could lock in process choices for a decade. A patent-protected, validated resin technology would have substantial leverage in that window. If bench validation is delayed or the predicted separation factors do not fully hold experimentally, the window tightens.

reduces solvent inventory + capital footprint vs long SX cascades

canonical BPDPA-P connectivity lock; phenanthroline-diamide topology excluded

The most direct buyers or licensees are magnet recyclers building or scaling separation capacity: companies in the U.S., Europe, Japan, and South Korea that are processing end-of-life NdFeB magnets from EV motors, hard disk drives, or industrial equipment. At commercial scale, these recyclers face exactly the capital and operational burden that this resin is designed to reduce, and a validated, patent-protected separation material would command a licensing premium relative to commodity SX reagents. Strategic acquirers might also include specialty chemical companies serving the rare earth processing sector — particularly those with existing resin manufacturing capability who would want to lock up the IP before a competitor does — and rare earth refining joint ventures that are being established under government-backed supply-chain programs in the United States and European Union. A secondary buyer class is defense and aerospace supply chain participants. Dysprosium and terbium are critical to the high-temperature performance of NdFeB magnets used in defense applications, and defense contractors or their Tier 1 suppliers have strong incentives to secure domestic supply chains for these materials. A licensing arrangement that gives a defense-adjacent buyer exclusivity in defense applications, while leaving commercial EV/wind markets open for broader licensing, would be a viable deal structure. The asset's clean FTO status and the specificity of the composition claims make it a credible IP anchor for such a transaction.

The central risk is that the predicted separation factors do not fully survive contact with real NdFeB leachate. Magnet leachate contains iron at concentrations that can be orders of magnitude higher than the target lanthanides, and iron interference with nitrogen-donor ligands is well-documented. If the imprinted cavity does not maintain selectivity in the presence of high iron, the resin would require upstream iron removal — which is achievable but adds process steps and reduces the capital-reduction argument. The 4f-in-core ECP correction is also a live uncertainty: if the corrected binding matrix narrows the predicted Dy/Nd separation factor substantially, the case for replacing a full SX cascade weakens, and the technology becomes a polishing step rather than a primary separation tool. These risks are de-risked in sequence: the ECP correction is a computational task that can be completed relatively quickly and cheaply, and bench validation on synthetic leachate followed by real leachate is a standard progression. The phosphonate anchor and picolinamide donor geometry are individually literature-supported for lanthanide coordination, which provides structural confidence that the bench results will be in the right range even if they do not hit the top of the predicted interval. The IP itself is not contingent on the validation results — the composition and device-use claims protect the invention regardless of where the bench separation factors land within the predicted range.