Skutterudite, Zintl, and tin-telluride backup thermoelectric compositions for segmented modules

This asset covers a curated set of backup thermoelectric compositions — skutterudite CoSb3 and its filled variants, Zintl-phase Mg3Sb2 and Mg3Bi2, rhenium-bearing As4Co8Re4, and the IV-VI rocksalt SnTe — that collectively solve a practical problem no single thermoelectric material can address on its own: matching leg chemistry to temperature zone across a segmented or graded module. The fundamental reason segmented modules exist is that no single thermoelectric material maintains competitive conversion efficiency across the full 300–900 K operating range relevant to automotive exhaust heat recovery, industrial furnaces, or aerospace power sources. A high-efficiency module requires p-type and n-type legs, and within each leg the chemistry typically must change at intermediate temperatures. The compositions claimed here are explicitly chosen to interlock: Mg3Sb2 (Zintl, n-type) performs well near room temperature and in the mid-range; CoSb3 and its filled skutterudite derivatives span the mid-to-high range; and SnTe provides a p-type complement in temperature zones where PbTe is the typical incumbent. Together they constitute a coherent "chemistry toolkit" for module builders rather than a single point solution. The filing is honest about its role in the broader catalysts and energy-conversion materials portfolio: it is a backup and defensive holding, not the lead asset. Its practical value lies in several strategic functions. First, it closes off obvious design-arounds for a competitor trying to replicate a segmented module built around the portfolio's primary half-Heusler claims — a competitor would need to negotiate, license, or design around this layer as well. Second, the computational validation already performed means these compositions are shovel-ready for experimental prioritization; a licensee does not start from zero. Third, the deliberate exclusion of broad PbTe and GeTe composition claims (where prior art is dense) keeps the portfolio clean and litigation-resistant, while the specific carve-outs actually claimed — segmented-leg device use and the particular Zintl and skutterudite families — remain defensible whitespace. The timing is appropriate. Waste-heat recovery at scale is being forced by tightening CO2 fleet-average standards in the EU and the United States, by new industrial-energy-audit mandates, and by the economics of hybrid powertrains where even a 2–3% improvement in net system efficiency translates to meaningful fuel and emissions savings. Module makers who currently build around Bi2Te3 and PbTe are actively looking for compositions that extend temperature capability and reduce lead content. This asset is positioned directly in that gap.

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 three primary material families represent distinct crystal-chemical strategies for thermoelectric performance. CoSb3 crystallizes in the Im-3 space group with an open cage structure that is the defining feature of the skutterudite class. The large interstitial voids in the CoSb3 lattice can be partially or fully occupied by rare-earth or alkaline-earth "rattler" atoms (denoted RyCo4Sb12, where R is the filler and y describes the filling fraction). The rattler mechanism is well-documented: the loosely bound filler atoms scatter heat-carrying acoustic phonons strongly while leaving the electron transport relatively unperturbed, reducing lattice thermal conductivity without proportionately degrading electrical conductivity or Seebeck coefficient. The BoltzTraP constant-relaxation-time calculation run on CoSb3 yields a Seebeck coefficient of approximately 272 µV/K at 300 K — a conservative estimate because constant-tau does not capture the full scattering anisotropy present in the doped, filled material. Actual zT in optimized filled skutterudites (e.g., Ce- or Yb-filled CoSb3) measured experimentally in the literature reaches 1.2–1.7 in the 700–900 K range, numbers the constant-tau result does not claim to reproduce, only to directionally validate. Mg3Sb2 belongs to the Zintl phase family, crystallizing in the trigonal P-3m1 space group. Its attractiveness as an n-type thermoelectric has become clear only in the last decade: when doped appropriately (Mn, Te, or Bi substitution), it achieves competitive zT in the 400–700 K range with no toxic heavy metals and earth-abundant constituents. The BoltzTraP calculation on Mg3Sb2 gives a Seebeck coefficient of approximately 362 µV/K at 300 K — the highest in this composition set — again under constant-tau, and again conservative relative to what is achievable with carrier-concentration optimization. The related Mg3Bi2 composition extends the family to a higher-mobility, somewhat lower-Seebeck variant that is useful for tuning the electronic properties of Mg3(Sb,Bi)2 alloys by adjusting the Sb:Bi ratio on the anion sublattice. Phonon transport in Mg3Sb2 has been computed with phono3py, which provides an ab initio estimate of lattice thermal conductivity by evaluating third-order interatomic force constants — a more demanding calculation than harmonic phonon stability alone, and one that gives a physical basis for comparing the intrinsic thermal conductivity of this material against competing legs in the same module. SnTe rounds out the p-type side. It crystallizes in the Fm-3m rocksalt structure, isostructural to PbTe but without the regulatory and supply-chain burden of lead. The inclusion of SnTe is strategic: broad composition-of-matter claims over the PbTe and GeTe rocksalt families were explicitly not pursued given the density of prior art in that space. SnTe, by contrast, offers a narrower and more defensible target. Its valence-band convergence at elevated temperatures can be enhanced by resonant-level doping (e.g., In) or alloying (Mn, Cd), and peak zT in the experimental literature is typically in the 0.6–1.0 range at 700–900 K. The rhenium-bearing As4Co8Re4 composition — an arsenide analog in the skutterudite-adjacent space — adds further coverage at the fringes of the skutterudite family, and two independent machine-learning interatomic potentials (MACE and CHGNet) have been applied to this structure, with both returning positive phonon frequencies across the Brillouin zone, confirming dynamic stability with no imaginary modes. This consensus between independent potentials, calibrated on different datasets and using different architectures, is the first computational gate a candidate must pass before more expensive DFT phonon and transport calculations are warranted. The computational workflow applied to this set exemplifies the Lattice Graph multi-fidelity approach: MLIP screening for dynamic stability first (two potentials, consensus required), BoltzTraP transport calculations for Seebeck and power-factor trends, phono3py for lattice thermal conductivity where warranted, and DFT cross-checking from two independent source calculations. The declared open gate — doped and filled zT confirmation — is honest and specific. The constant-tau BoltzTraP results validate chemical intuition and structure-property trends but do not replace experimental or fully ab initio confirmation of peak zT once the carrier concentration, filling fraction, and microstructure are optimized. A prospective licensee or acquiring party should budget for targeted synthesis and Hall-effect characterization of two to three compositions to close this gate.

The addressable market for thermoelectric modules used in waste-heat recovery is estimated at $0.5–1 billion annually, a figure that reflects current module shipments for automotive, industrial, and aerospace applications rather than the theoretical ceiling if thermoelectrics achieved full penetration of available heat streams. The current market is dominated by Bi2Te3 modules for near-room-temperature applications and PbTe-based modules for mid-to-high temperature ranges. Segmented and graded modules — the target application for this composition set — represent a meaningful and growing subset of this market because they are the practical path to efficiency in applications where the hot-side temperature exceeds 600 K, which includes diesel and gasoline exhaust recovery, steel and glass furnace stacks, and stationary gas-turbine afterburners. The customers for compositions of this type are module makers and, behind them, Tier-1 automotive suppliers and industrial OEMs who qualify thermoelectric modules for integration into heat-recovery systems. The licensing logic follows accordingly: compositions and device-use claims license best as a bundle with manufacturing know-how and module design, because the performance of a segmented module is highly sensitive to interfacial bonding, contact resistance, and thermal expansion matching between adjacent leg segments. A royalty structure based on module shipped value is the most natural enforcement point, given that compositions and device configurations are directly traceable to the finished product. Longer term, the regulatory pressure reducing lead content in industrial and automotive components — paralleling the RoHS directives already in effect in consumer electronics — creates a forced substitution dynamic that specifically favors SnTe over PbTe as a p-type leg option. That substitution is not imminent in all markets, but the direction of travel is clear and the timeline (2030–2035 for many automotive applications) is short enough that module makers are already evaluating alternative compositions. This asset is positioned at the boundary of that transition.

complementary leg chemistries for multi-leg modules

segmented/graded leg use; broad IV-VI rocksalt composition not claimed

The most natural acquirers or licensees for this asset are thermoelectric module manufacturers operating at the mid-to-high temperature segment: companies that build exhaust-heat recovery units for commercial trucks, industrial furnace systems, or aerospace auxiliary power, and who are actively engineering second-generation modules to replace or supplement their existing PbTe and Bi2Te3 leg sets. Marlow Industries (a subsidiary of II-VI / Coherent), Laird Thermal, and Alphabet Energy alumni ventures are representative of the commercial landscape, as are Tier-1 automotive suppliers with dedicated thermoelectric module programs such as Gentherm and relevant business units at Bosch and Eberspächer. For any of these buyers, the value of this asset is primarily defensive and complementary — it fills in the composition and device-use space around the lead half-Heusler claims in the portfolio and removes the risk that a competitor or new entrant could design a segmented module using Zintl or skutterudite legs without engaging the portfolio. A secondary buyer category is material suppliers seeking upstream IP coverage on thermoelectric powders and target materials. Companies supplying pressed CoSb3 or Mg3Sb2 material to module makers would be practicing the composition claims, making them potential licensees or co-development partners. In either the direct-licensing or acquisition scenario, this asset is most valuable bundled with the broader catalysts and energy-conversion materials portfolio — its standalone value as an isolated filing is limited by the narrowness of scope, but as part of a layered strategy covering both the primary half-Heusler compositions and the backup segmented-leg chemistries, it meaningfully strengthens the portfolio's enforceability against any party trying to engineer a thermoelectric module without engaging the portfolio's claims.

The principal technical risk is the open zT gate. The computational results to date are credible and directionally correct, but they do not constitute confirmed performance data on doped or filled compositions. A sophisticated buyer will require either experimental confirmation or a credible de-risking plan before attributing significant value to the performance claims embedded in the filing. The path to closing this gate is clear — targeted synthesis and characterization of two to three compositions, achievable in six to twelve months at a university or national-lab partner — but until it is closed, the asset carries material-performance uncertainty. A second risk is the established prior art in the skutterudite and Zintl literature: while the device-use claims in segmented modules are more defensible than composition-only claims, an aggressive defendant could argue that combining known thermoelectric materials in a known module architecture (segmented legs) is obvious to a practitioner skilled in the field. The countervailing argument is the specific combination and configuration being claimed, which was not previously disclosed in the segmented-module context for these particular compositions. The narrow FTO status is an honest acknowledgment of these constraints rather than a disqualifying factor — the scope actually claimed is chosen to be defensible, not maximal. Buyers should treat this asset as a supporting layer in a module IP strategy rather than as a standalone blocking position.