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

Half-Heusler intermetallics occupy a structural sweet spot that has attracted thermoelectric researchers for two decades: they are mechanically robust, thermally manufacturable at scale, non-toxic relative to the legacy PbTe and Bi2Te3 incumbents, and their narrow semiconducting bandgap is tunable through iso-structural solid solutions. HfNiSn sits at the center of that space. With a computed bandgap of roughly 0.39 eV, it is optimally positioned for mid-temperature thermoelectric operation — the 400–800 K exhaust and industrial waste-heat window where neither Bi2Te3 (too low a melting point) nor SiGe (too high a processing cost) performs well. When grain boundaries are engineered at the nanoscale to scatter mid-wavelength phonons without proportionally degrading electron transport, the dimensionless figure of merit zT reaches 0.8–1.2 in the literature on nanostructured pellets. That range is competitive with the best mid-temperature thermoelectrics deployed commercially today. The timing is driven by two converging forces. First, automotive OEMs and tier-1 suppliers are under regulatory pressure to recover waste heat from combustion engines (hybrid powertrains and 48V mild-hybrids still dominate the global fleet for the next decade), and thermoelectric generators bolted to exhaust manifolds are one of the few retrofit-compatible solutions. Second, NASA and commercial space vendors are actively looking for RTG (radioisotope thermoelectric generator) materials that can replace the aging GPHS-RTG stack built on PbTe/TAGS: half-Heuslers offer higher Curie-temperature stability and no phase transitions across the RTG operating range. This asset positions the catalysts and energy-conversion materials portfolio directly in front of both demand drivers with claims anchored not on the broad composition — which is prior art — but on the specific combination of dopant chemistry, nanostructuring process, and device-integration architecture that makes the material commercially deployable.

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 F-43m (cubic, space group 216) half-Heusler crystal structure, which can be understood as a zinc-blende NiSn sublattice with Hf atoms filling one set of tetrahedral voids. This arrangement produces a narrow indirect bandgap — computed here at 0.39 eV — that sits in the ideal range for Seebeck-coefficient optimization at mid temperatures. The structure is isoelectronic with the 18-valence-electron rule that governs stability across the half-Heusler family, and it remains cubic and phase-pure up to at least 1,000 K, which is the primary failure mode eliminated relative to skutterudites and to the β→α phase transition in PbTe. The thermoelectric performance metric zT = S²σT/κ (where S is the Seebeck coefficient, σ is electrical conductivity, T is absolute temperature, and κ is total thermal conductivity) splits into two regimes for this material. In dense, coarse-grained polycrystalline pellets, zT is conservatively 0.3–0.5 at 700–800 K — a well-documented experimental baseline in the published literature. When the grain size is pushed into the sub-100 nm regime through ball-milling followed by spark-plasma sintering, point-defect and boundary scattering reduce the lattice thermal conductivity κ_L substantially while the power factor S²σ is preserved, yielding zT of 0.8–1.2. The same phonon-engineering logic applies to the quaternary solid solutions Hf₀.₅Zr₀.₅NiSn and (Hf,Zr,Ti)NiSn alloys, where mass-variance scattering on the group-IV site provides an additional phonon-scattering channel without disrupting the electronic band structure. The computational validation stack on this asset is the deepest in the portfolio, with more than thirteen independent engine runs documented. Two independent machine-learning interatomic potentials — MACE and CHGNet — were applied to the F-43m structure. MACE 4×4×4 supercell phonon calculations return a minimum phonon frequency of +0.24 THz, meaning all branches across the full Brillouin zone are positive; there are no imaginary modes. CHGNet independently confirms dynamic stability and gives results in strong agreement with the MACE result. This consensus across two potentials trained on entirely different datasets is the primary computational stability gate; structures where the potentials disagree are not advanced in this pipeline. DFT calculations from two independent sources provide the electronic structure baseline — bandgap and density of states — that the BoltzTraP transport calculations (Seebeck coefficient, electrical conductivity, power factor as a function of carrier concentration) are built on. Phono3py anharmonic lattice-dynamics calculations yield the lattice thermal conductivity κ_L, run on the HfNiSn end member and cross-checked against MG3Sb2 as a validation reference. MatterSim Langevin molecular-dynamics at elevated temperature shows a mean atomic displacement of 1.04 Å, confirming the structure does not disorder or melt under finite-temperature dynamics at the target operating point. Taken together, this is a multi-physics, multi-potential validation chain covering structural stability, electronic transport, and thermal transport — not a single-point DFT calculation.

The addressable market for mid-temperature thermoelectric modules spans two primary verticals. The first is automotive and industrial waste-heat recovery. Internal combustion engines and hybrid powertrains waste roughly 30–40% of fuel energy as exhaust heat in the 400–800 K range. Thermoelectric generator (TEG) modules mounted on exhaust pipes can convert a fraction of that heat directly to electricity, reducing alternator load and improving fuel economy by 1–3% on highway cycles. The global automotive TEG market is estimated at several hundred million dollars today, with projections toward $1–3 billion over the next decade as hybrid and mild-hybrid mandates tighten in the EU, China, and California-rule states. Industrial furnaces, glass-manufacturing lines, cement kilns, and aluminum smelters represent an additional mid-temperature waste-heat stream of comparable aggregate scale. Module makers and tier-1 automotive suppliers are the direct buyers of thermoelectric leg material or finished modules. The second vertical is space RTGs. NASA's Multi-Mission Radioisotope Thermoelectric Generator (MMRTG) and the successor systems under development for lunar surface and deep-space missions require materials that are dimensionally stable across multi-decade operational lifetimes with no phase transitions. The ²³⁸Pu heat source delivers a hot-side temperature in the 500–900 K range — exactly the half-Heusler optimum. The RTG materials market is small in dollar volume (tens of millions per mission), but it commands extremely high unit margins and requires certified materials with multi-physics computational dossiers of exactly the type assembled here. Licensing a validated half-Heusler composition to a prime contractor (Teledyne, Aerojet Rocketdyne, or a NASA-affiliated supplier) represents a path to early revenue that is not contingent on the larger automotive market developing. Taken together, the $1–3 billion total addressable market estimate reflects primarily the automotive waste-heat segment, with RTG representing a smaller but nearer-term and higher-margin licensing opportunity.

mid-temp stability for RTG/exhaust; deepest compute-proof stack

dopant-regime + nanostructure + device-integration limitations; broad composition-of-matter not pursued

The most natural acquirers or licensees for this asset fall into two groups. The first is thermoelectric module makers with automotive or industrial programs: Gentherm (the largest Western TE module company, with active automotive TEG programs), II-VI / Coherent, Ferrotec, and several Japanese and Korean module suppliers who are active in the waste-heat recovery space. These buyers would license the process-and-device claims to accelerate their own material qualification without reproducing the computational screening and process optimization from scratch. The computational dossier — phonon stability, transport coefficients, thermal conductivity modeling — directly substitutes for 12–18 months of internal experimental work, which is the primary value proposition in a licensing conversation with an engineering organization under product development timelines. The second group is aerospace and defense prime contractors or their material suppliers who are developing next-generation RTG systems. Teledyne Energy Systems, Aerojet Rocketdyne, and NASA GSFC material programs are the relevant parties; in the commercial space sector, companies developing nuclear-powered satellites or surface power systems for lunar and Mars missions are an emerging category. For RTG qualification, a multi-physics computational dossier covering dynamic stability, thermal transport, and electronic transport is precisely the kind of early-stage material evidence required by program offices before committing to experimental synthesis campaigns. The asset is well-positioned for a technical licensing arrangement or a sponsored research agreement that funds the remaining experimental validation milestone in exchange for exclusive or field-of-use licensing rights.

The primary technical risk is that the experimentally measured zT on in-house synthesized nanostructured material may land in the lower part of the 0.8–1.2 range, or may require further process iteration to reproduce consistently. This is a real but bounded risk: the academic literature contains multiple independent demonstrations of high-zT nanostructured half-Heuslers using spark-plasma sintering, and the computational thermal-transport predictions are consistent with those reports. The risk is execution quality on synthesis, not a fundamental materials-science unknown. The path to de-risking is a focused 6–12 month synthesis program producing pellets at multiple grain sizes and dopant levels and measuring zT versus temperature on a certified instrument. A second risk is claim scope: the deliberate narrowing away from composition-of-matter means the claims are more easily designed around, particularly by a well-resourced competitor who can optimize a slightly different dopant combination or grain-size processing route. The mitigation is continuation prosecution that tracks the experimental results and tightens claim language around the validated optimum, combined with trade-secret protection for the specific process recipe. The FTO risk from existing module-maker patents is real but manageable given the explicit carve-out strategy, provided a proper clearance opinion is obtained before commercial production. None of these risks are unusual for a mid-stage materials IP asset, and all of them have clear, defined de-risking pathways.