Half-Heusler intermetallic thermoelectric module for high-temperature waste-heat recovery

Half-Heusler intermetallics have been theorized as the natural successor to bismuth-telluride and lead-telluride thermoelectrics for high-temperature operation for roughly two decades. The commercial and strategic motivation is clear: Bi2Te3 works well below roughly 500 K; PbTe extends usable range into the mid-temperature regime but carries lead-toxicity burdens; neither material is suited to the sustained 600–900 K exhaust environments found in automotive and industrial waste-heat streams. Half-Heuslers in the M-Ni-Sn and M-Fe-Sb families sit precisely in that gap, with intrinsically lower thermal conductivity than many competing intermetallics, reasonable carrier mobility, and — critically — a composition space built entirely from earth-abundant elements carrying no tellurium or lead supply-chain risk. The headline performance claim here is roughly 2x the temperature-integrated electrical power output versus a conventional Bi2Te3/PbTe segmented couple over the same hot-side/cold-side temperature differential. That figure, if validated in hardware, represents a compelling economic proposition for any automotive integrator or industrial heat-exchanger manufacturer trying to meet fleet fuel-economy or decarbonization targets without exotic feedstocks. The timing dynamic matters. The half-Heusler thermoelectric space is not empty — Bosch, Toshiba, Toyota, and specialist firms like Marlow Industries and ZT Plus have been active filers — but the composition and device claim space remains genuinely fragmented, with established players tending to anchor on specific anion genera (Sb/Bi anion-side) or system-level architectures (liquid-cooling integration) rather than on a clean, independent n-leg module genus. The strategy here is to occupy the M-Ni-Sn n-leg space with a genus-level composition claim while pairing it with a generic p-leg, sidestepping the crowded NbFeSb p-leg pairing territory. For an acquirer or licensee in automotive exhaust harvesting or industrial co-generation, this asset offers a defensible position at the module architecture level — exactly the layer where engineering contracts and supply agreements are written. The asset sits within the broader "dielectric, ferroelectric and wide-bandgap oxides" portfolio as the portfolio's most computationally intensive thermoelectric entry, with the deepest simulation and knowledge-graph density of any asset in that grouping. It is a lead filing, not a defensive backup — the composition and device-use claims are intended to be the primary commercial vehicle, with dependent claims covering specific pairings retained for licensing discussions rather than front-line assertion.

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.

The primary materials system is Hf(1-x)ZrxNiSn, a continuous solid-solution series across the HfNiSn/ZrNiSn binary, crystallizing in the cubic F-43m half-Heusler structure. The 18-electron count that characterizes this family enforces a narrow, semiconducting bandgap — computed here at approximately 0.39 eV — which sits in the ideal range for thermoelectric power factor optimization at elevated temperature. TiNiSn is included in the composition genus as the third member of the M-Ni-Sn triad, offering a lighter-mass A-site that can be alloyed to tune lattice thermal conductivity via mass-variance scattering without disrupting the electronic structure significantly. The p-type leg pairing in the module context involves NbFeSb and VFeSb, which are 18-electron half-Heuslers of the M-Fe-Sb family with complementary transport properties. The computational validation pipeline applied to this family is the most thorough in the portfolio. Two independent machine-learning interatomic potentials — MACE and CHGNet — were used to evaluate phonon band structures for the principal compositions. Both potentials agree that HfNiSn and NbFeSb are dynamically stable: the lowest phonon frequencies are positive (no imaginary modes), with MACE reporting a minimum phonon frequency of +0.365 THz for HfNiSn and +0.422 THz for NbFeSb. VFeSb also clears the stability bar at +0.475 THz. This consensus-based approach is a deliberate guard against false positives: a single potential can produce spurious stability verdicts, particularly for intermetallics with partially filled d-shells, so requiring agreement between two architecturally distinct potentials before advancing a composition provides meaningful confidence. Three independent DFT source calculations further corroborate the electronic structure and phonon results. TaFeSb, a candidate that was evaluated for p-leg inclusion, was rejected on computational grounds: its lowest phonon frequency is strongly negative at −2.495 THz, indicating genuine dynamical instability at the structure examined, and it has been explicitly excluded from the claim genus. Thermal transport is a central design variable for any thermoelectric. The simulation set includes a Cahill-Pohl minimum thermal conductivity sweep across the half-Heusler family, yielding a family mean kappa_min of approximately 1.55 W/m/K. This is the amorphous or glassy lower limit on lattice thermal conductivity, not an achieved bulk value, but it establishes the theoretical floor that grain-boundary engineering and alloying strategies are working toward. Real bulk values in heavily alloyed Hf/Zr/Ti-Ni-Sn ceramics in the literature sit in the 2–4 W/m/K range, meaning there is meaningful headroom between the current state of practice and the computed minimum. Interface molecular dynamics simulations have been run to probe the thermal boundary resistance at leg/electrode contacts, which is a known efficiency bottleneck in module-level performance and an area where simulation-guided contact metallization can improve real device output. An extended genus of structurally related 18-electron half-Heuslers has been screened and several were found computationally stable: CaAgAs, DyNiBi, MgCuP, LiZnAs, and LiZnP. These materials are electronically and structurally analogous to the M-Ni-Sn family but explore different chemistries — alkaline-earth, rare-earth, and light-element variants — that could offer distinct thermal conductivity or carrier-concentration tunability. Their stability verdicts (confirmed by the same two-potential consensus protocol) are the basis for the genus-widening claims added to the family. TiFeTe was also included in the extended genus enumeration. These widening members exist primarily to establish composition-space breadth in the patent claims and to create optionality for licensing; they are not the primary commercial target compositions, which remain HfNiSn/ZrNiSn/TiNiSn for the n-leg.

The addressable market for high-temperature thermoelectric modules is estimated at roughly $1–3 billion annually, spanning automotive exhaust heat recovery, industrial process heat co-generation, and emerging aerospace auxiliary power applications. The buyer profile is well-defined: automotive OEMs and Tier 1 suppliers integrating exhaust gas recirculation and waste-heat recovery systems, industrial furnace and kiln operators seeking parasitic power from process exhaust, and defense/aerospace primes seeking solid-state power conversion without moving parts. Royalty logic in this space typically follows the module unit value — a thermoelectric module for a passenger vehicle application might carry a manufacturer's selling price in the $50–200 range per module, with IP royalty rates in established mechanical and electronic component markets running in the 2–5% range. At moderate penetration of even the automotive exhaust-harvesting segment, the royalty stream on a genus-level composition-plus-device claim is material. The critical commercial variable is the earth-abundant, tellurium-free materials advantage. Tellurium is a byproduct of copper smelting with highly concentrated global supply — roughly 40% comes from a single country — and supply shocks have historically moved Te prices by an order of magnitude within a few years. Any OEM designing a thermoelectric module for production volumes in the millions of units per year faces real procurement risk with Te-bearing materials. HfNiSn and ZrNiSn are composed of hafnium, zirconium, nickel, and tin — all with diversified, established supply chains and no comparable concentration risk. Hafnium carries a price premium as a zirconium co-product, but the module mass fraction is small enough that feedstock cost is not a barrier at automotive production economics, particularly relative to the value of recovered waste heat. The performance claim of approximately 2x temperature-integrated electrical power versus a Bi2Te3/PbTe segmented couple, if supported by coupon-level zT measurements, translates directly into a cost-per-watt-recovered advantage that purchasing engineers can put into an ROI model. This is the sales proposition that converts a composition patent into a module supply agreement or technology license. The window for establishing that first-mover licensing position is near-term, with parallel-filing timing indicating active prosecution rather than a dormant holding.

~2x T-integrated power vs Bi2Te3/PbTe; earth-abundant, no Te supply risk

independent M-Ni-Sn n-leg + generic p-leg module; NbFeSb-p-leg pairing license-flagged dependent

The most natural strategic acquirers and licensees fall into two groups. The first group is Tier 1 automotive suppliers with active waste-heat recovery programs — companies that already have thermoelectric module development programs and are seeking to upgrade from Bi2Te3/PbTe technology before regulatory pressure on fleet fuel economy forces the issue. Toyota is already active in the space as a filer; Bosch, Continental, and BorgWarner are logical targets given their exhaust and powertrain thermal management portfolios. For these buyers, the value proposition is a defensible material and module claim that enables a product differentiated on raw performance and supply-chain resilience, acquired before a competitor does the same. The second group is specialty thermoelectric module manufacturers and industrial energy-harvesting integrators — companies like Marlow Industries (now part of II-VI/Coherent), Alphabet Energy (now absorbed into the industrial thermoelectrics landscape), and emerging players in industrial co-generation. For these buyers, a license rather than an outright acquisition is more likely: the claim provides freedom to practice a high-performance n-leg composition in exchange for a per-module royalty or an upfront technology access fee. An industrial partner integrating thermoelectric recovery into cement kilns, glass furnaces, or petrochemical process heaters would find the earth-abundant, high-temperature composition profile directly aligned with their operating conditions — temperatures above 600 K are routine in those environments and well outside the practical range of Bi2Te3.

The primary technical risk is the gap between computational prediction and measured module performance. The ~2x power output figure is a projection; actual module zT in a production-relevant form factor depends on grain boundary engineering, contact metallization, and thermal cycling stability — none of which are fully captured by phonon stability calculations alone. If coupon-level measurements come in below expectation, the commercial narrative requires adjustment. The mitigation path is straightforward in concept: a focused measurement program on sintered Hf(1-x)ZrxNiSn pellets with candidate contact metals, which is standard thermoelectric characterization work achievable at university or national-lab facilities under a sponsored research agreement before significant capital is committed. The secondary risk is freedom-to-operate, specifically for the NbFeSb p-leg pairing. The FTO picture for the independent M-Ni-Sn n-leg claim is manageable, but the competitive filing density in the half-Heusler space means that any product configuration pairing NbFeSb as the p-leg requires a dedicated FTO opinion before commercialization. Prosecution timeline risk is also real: parallel filing timing means this family is in active prosecution, and claim scope may be narrowed during examination. Buyers or licensees should factor prosecution cost and outcome uncertainty into any valuation, and should engage patent counsel early on the V/Ta A-site prior art read and the Bosch Sb/Bi anion-genus landscape to scope the actual claim footprint they are acquiring.