Hafnium nickel tin half-Heusler thermoelectric for mid-temperature waste-heat recovery

HfNiSn and its substituted half-Heusler analogs address one of the most persistent frustrations in practical thermoelectric engineering: the mid-temperature (600–900 K) window where the best commercial materials — lead telluride and silicon-germanium — either require toxic, supply-constrained elements or carry unacceptable fabrication cost. Half-Heuslers in the C1b crystal structure have been studied for decades precisely because they are mechanically robust, thermally stable in air, and composed of earth-abundant transition metals. What has held back commercial adoption is the combination of relatively high lattice thermal conductivity and the difficulty of simultaneously optimizing the power factor without destroying carrier transport. The Lattice Graph portfolio's work on the Half-Heusler thermoelectric family positions undoped HfNiSn and a set of targeted substituted analogs as a credible, patent-protectable path to a lead-free mid-temperature thermoelectric leg with ZT in the 0.8–1.2 range at 800 K. The strategic timing matters because regulatory pressure on lead-containing thermoelectrics is tightening in the European Union and increasingly in Asia, while tellurium's supply chain (dominated as a byproduct of copper smelting) is drawing attention as demand from photovoltaics competes directly with thermoelectric module makers. A material system that delivers competitive ZT at 800 K without either element, from a crystal structure that survives thermal cycling, enters a forced-substitution dynamic that does not depend on a performance breakthrough to gain market traction. The portfolio here is candid about its boundaries: the broad half-Heusler space is partially claimed by existing prior art, and the protectable novelty has been narrowed to specific undoped HfNiSn and Zr-, Ti-, Pd-, and Sb-substituted embodiments. That narrowing is the result of rigorous freedom-to-operate work rather than a weakness — it defines exactly where the white space is.

Each candidate is validated by multiple independent machine-learning interatomic potentials. A material advances only when the engines agree on phonon (dynamic) stability — disagreement is surfaced, not hidden.

HfNiSn adopts the C1b half-Heusler structure (space group F-43m), a filled tetrahedral framework in which Hf occupies the A-site, Ni the B-site, and Sn the X-site. The half-Heusler motif leaves one of the four tetrahedral voids of the full Heusler structure empty, which is directly responsible for its semiconductor character — unlike its fully-filled cousin, C1b HfNiSn is a narrow-gap semiconductor with a band structure amenable to thermoelectric optimization. The crystal symmetry also means the lattice is mechanically stiff and chemically stable at temperatures where many competing materials (including skutterudites) begin to decompose or require inert-atmosphere encapsulation. The key thermoelectric figure of merit, ZT = S²σT/κ, is projected at approximately 0.8–1.2 at 800 K for the lead compound HfNiSn, with a Boltzmann Transport Equation (BTE) power factor of approximately 30 μW/cm/K² at 800 K and a lattice thermal conductivity (κ_lat) of approximately 3 W/m/K. Those numbers reflect the well-known tension in half-Heuslers: the power factor is genuinely high — competitive with or exceeding PbTe in absolute terms — but so is the lattice thermal conductivity compared to the best chalcogenide thermoelectrics. The room for improvement lies in isoelectronic A-site substitution (Hf ↔ Zr, Ti), B-site substitution (Ni ↔ Pd), and X-site substitution (Sn ↔ Sb) to introduce mass and strain disorder that scatters mid-frequency phonons without proportionally damaging the electronic transport. Each of the five compositional embodiments in the protected family — HfNiSn, (Hf,Zr)NiSn, (Hf,Ti)NiSn, Hf(Ni,Pd)Sn, and HfNi(Sn,Sb) — targets a different phonon-scattering mechanism while staying within the same topological crystal framework. Dynamic stability was evaluated using two independent machine-learning interatomic potentials — MACE and MatterSim — which returned minimum phonon frequencies of +0.408 THz (MACE) and +0.179 THz (MatterSim) respectively. Both values are positive across the full Brillouin zone, meaning neither potential identifies any imaginary (soft) mode that would indicate a tendency for the structure to spontaneously distort or decompose. The agreement between two potentials trained on very different datasets, combined with two independent DFT-level single-point validations, constitutes a multi-layer confirmation that the F-43m structure is the correct ground state and is not being artificially stabilized by a single model's biases. Fifteen independent phonon runs were registered across the candidate family, and all returned positive minimum frequencies in the range +0.096 to +0.408 THz, giving statistical confidence in the stability result rather than a single-point calculation. The computational workflow also included Boltzmann Transport Equation calculations for the power factor and lattice thermal conductivity (logged as simulation V-HFNISN-001), which provided the projected ZT range. These calculations were supplemented by cross-engine phonon reseeding (three independent re-initializations) to confirm that the stability result is not sensitive to the initial atomic displacement pattern. What remains computationally open — and is candidly flagged — is the full experimental ZT measurement in the relevant processing window. The BTE power factor and phonon calculations set the target, but the experimental confirmation of doping level, grain boundary microstructure, and thermoelectric module integration will determine where in the 0.8–1.2 ZT range a real device lands.

The addressable market for mid-temperature thermoelectric generation — the 600–900 K range where industrial waste heat, automotive exhaust, and concentrated solar applications operate — is estimated at $1–5 billion globally. That range reflects genuine uncertainty: the market is currently constrained by the cost and toxicity limitations of incumbent materials rather than by lack of demand, so a credible lead-free alternative that reaches module-level cost parity could expand the addressable base beyond what current deployment numbers imply. The primary buyers are waste-heat-recovery module makers who supply into automotive, heavy industry (steel, cement, glass), and distributed power generation channels. Royalty and licensing logic in this space typically runs on a per-module or per-watt basis, with materials licenses commanding meaningful value because the thermoelectric leg is the margin-determining component of any module. A licensor of the composition and device-use rights could command royalties structured around the ZT performance premium over alternatives — roughly, any ZT advantage over the reference material at 800 K translates directly into module efficiency and thus module price. For waste-heat-recovery applications in heavy trucks and industrial furnaces, where a 10% efficiency improvement can represent thousands of dollars per installation per year in fuel or energy savings, the willingness-to-pay per watt of thermoelectric capacity is substantially higher than in consumer electronics thermoelectrics. The supply-chain argument adds a second commercial vector. Tellurium production is approximately 400–500 metric tons per year globally, and photovoltaic CdTe demand alone is projected to consume an increasing fraction of that supply over the next decade. Module makers who are engineering mid-temperature products today face a sourcing risk if they lock in a Te-based chemistry. HfNiSn-family materials use hafnium (a byproduct of zirconium processing with established supply chains), nickel, and tin — none of which face the same concentration-of-supply risks. This is a genuine durability argument that a sophisticated procurement buyer will value independently of the ZT number.

lead-free non-Te mid-T thermoelectric with A/B/X tunability

narrowed to undoped HfNiSn + specific Zr/Ti/Pd/Sb embodiments

The natural acquirers and licensees for this family are waste-heat-recovery module manufacturers who are actively engineering non-lead, non-tellurium product lines for the 600–900 K temperature window. This includes established thermoelectric module companies in Germany, Japan, South Korea, and the United States who supply automotive OEMs and industrial furnace operators. Strategic acquirers from the automotive tier-one supplier base — particularly those already qualified in thermal management components — would also find the composition and device-use rights strategically valuable as regulatory pressure on lead-containing materials intensifies. A bolt-on acquisition of a defined, FTO-cleared half-Heusler claim set is faster and lower-risk than an internal materials development program starting from scratch. A second buyer category is materials companies with existing half-Heusler production capability who could use the substituted composition rights to differentiate their product line without infringing the CN117265309B family. For such a buyer, the license value is partially defensive — securing freedom to practice their own manufacturing program without future dispute — and partially offensive, in enabling them to offer a certified, patent-protected product to module makers who demand IP clarity in their supply agreements. The portfolio's combination of BTE-validated performance projections, multi-potential phonon stability confirmation, and a documented FTO analysis makes the diligence package unusually complete for an early-stage computational materials asset.

The principal technical risk is the gap between the BTE-projected ZT of 0.8–1.2 and experimental confirmation. Half-Heusler thermoelectrics are notoriously sensitive to grain boundary scattering, secondary-phase formation, and doping uniformity — effects that BTE calculations treat only approximately. The spread in the projected ZT range (0.8 to 1.2) reflects this uncertainty, and a real synthesis campaign could land anywhere in that range or, in a worst case, below it if microstructure proves difficult to control. The roadmap to de-risk this is straightforward: a targeted synthesis and characterization program on the five protected compositions, measuring Seebeck coefficient, electrical conductivity, and thermal conductivity as a function of temperature from 300 to 900 K. That program would cost on the order of months and moderate laboratory resources, and its output would either confirm the computational projections or identify which substitution pattern achieves the highest experimental ZT. The secondary risk is patent scope. The narrowing to specific undoped and substituted embodiments, while strategically sound, means that a competitor who identifies a slightly different substitution pattern not covered by the claims can practice in adjacent space without a license. The counter-strategy is to file continuation claims on additional substitution patterns as experimental data accumulates, maintaining a claim set that tracks the validated composition space. The FTO risk from CN117265309B is real but bounded — it determines where you cannot file, not whether the protected embodiments are themselves valid. A buyer with resources to run the experimental confirmation program and file continuation claims based on results is well-positioned to build out this asset from its current early-stage computational foundation into a fully de-risked, experimentally validated licensing position.