High-throughput alloy design of superior thermoelectric materials (Allotherm) (completed)

• Grimenes, Øven Andreas; Løvvik, Ole M. & Berland, Kristian (2026). Exceptional thermoelectric properties in Na2TlSb enabled by quasi-one-dimensional band structure. Physical review B (PRB). ISSN 2469-9950. 113(15). doi: 10.1103/jnfj-2n8p. Full text in Research Archive Show summary

  Materials with reduced dimensionality offer beneficial density-of-states (DOS) profiles for thermoelectric energy conversion, but can be impractical in realistic devices. Encouragingly, bulk high-symmetry materials can also exhibit similar quasi-low-dimensional band structures. A striking example is the full-Heusler compound Na2⁢TlSb, whose valence-band energy isosurfaces can form intersecting two-dimensional pockets, i.e., a boxlike structure. The individual energy isosurface sheets resemble those of 1D quantum wires. The combination of high electron velocities (perpendicular to the pockets) and a rapidly increasing DOS with energy in the transport regime (due to the low dimensionality) makes Na2⁢TlSb a representative case where the band structure gives rise to attractive electronic transport properties. However, these beneficial features could be counteracted by high electronic scattering rates due to the large scattering space. In this first-principles study of Na2⁢TlSb, we find that the electronic scattering rates remain modest. This result is linked to the reduced matrix elements of large-momentum (𝐪) scattering across the delocalized energy isosurfaces. The enhanced free-carrier screening due to the large DOS also contributes to reducing scattering. In combination, the low-dimensional features and modest scattering result in excellent electronic transport properties. Combined with an ultralow lattice thermal conductivity of 𝜅ℓ<1W/mK reported in the literature, we predict a thermoelectric figure of merit ranging from 2.4 at 300 K to 4.4 at 600 K. The 𝑛-type properties are also favorable, with 𝑧⁢𝑇 values from 1.5 at 300 K to 3.0 at 600 K.

• Grimenes, Øven Andreas & Berland, Kristian (2025). Common errors in BoltzTraP-based calculations. Computational Materials Science. ISSN 0927-0256. 262, p. 1–8. doi: 10.1016/j.commatsci.2025.114380. Full text in Research Archive Show summary

  Boltzmann transport calculations based on band structures computed from first principles play an important role in modern thermoelectric materials research. Among available codes, the BoltzTraP code is the most widely adopted, but many recent studies contain systematic user errors. We identify three error modes: (1) inserting the electronic thermal conductivity at zero electric field, 𝜅0, in place of the electronic thermal conductivity at zero electric current, 𝜅𝑒, (2) computing the figure of merit 𝑧𝑇 by combining a constant relaxation time of unity while keeping the lattice thermal conductivity 𝜅𝓁 in standard units, and (3) doing both errors at once. We have found many examples of the third error mode, but since the first two are simpler, we suspect they are also present in the literature. For the single parabolic band model, we derive exact analytical limits in the non-degenerate and near-degenerate regimes, and show how mistakes appear in the results for the realistic case study of ZrNiSn. Our results illustrate how faulty calculations can appear reasonable at certain temperatures and Fermi levels, and we provide practical guidance for identifying faulty results and avoiding such pitfalls in thermoelectric transport studies.

• Álvarez, José Luis Garrido; Lopez-Garcia, Javier; Dura, Oscar J.; Schrade, Matthias; Poulia, Anthoula & González, Carlos F. Gutiérrez [Show all 14 contributors for this article] (2025). Synergistic effects of Sb doping and nanostructuring on spark plasma-sintered ZrNiSn thermoelectrics. Materials Advances. p. 1–11. doi: 10.1039/d5ma01100k. Full text in Research Archive Show summary

  Nanostructured ZrNiSn and ZrNiSn1−xSbx (x = 0.02 and 0.05) half-Heusler alloys were successfully synthesized via mechanical milling followed by spark plasma sintering. All samples crystallized in the expected MgAgAs-type crystal structure (space group 𝐹-43m) and retained their nanostructured powder morphology post-sintering. Thermoelectric measurements revealed that the mechanical milling effectively reduced lattice thermal conductivity, while Sb doping significantly enhances the electrical conductivity in the ZrNiSn system. However, extending the milling time beyond 5 hours led to reduced performance due to the increased structural disorder without further thermal conductivity decrease. Hall effect measurements yielded a relatively low effective carrier mass of around 0.7-1.3 me, which accounts for the observed reduction in the Seebeck coefficient. The optimal Sb doping level was found to be x = 0.02. At this doping level, a substantial improvement in thermoelectric performance is detected with a maximum thermoelectric figure of merit (ZT) of approximately 0.6 at 600 K (two folds that of the undoped sample) and a 60% improvement over the unmilled ZrNiSn0.98Sb0.02 alloy These results highlight the synergistic effects of controlled Sb doping and nanostructuring via mechanical milling on optimizing the thermoelectric properties of ZrNiSn-based half-Heuslers.

• Grimnes, Øven Andreas; Snyder, G. Jeffrey; Løvvik, Ole Martin & Berland, Kristian (2025). Thermoelectric transport of strained CsK2 Sb: Role of electron velocities and scattering within extended Fermi surfaces. Physical review B (PRB). ISSN 2469-9950. 111(19), p. 1–12. doi: 10.1103/PhysRevB.111.195205. Full text in Research Archive Show summary

  In this first-principles study, we investigated the thermoelectric properties of the full-Heusler compound CsK2⁢Sb at different compressive strains. This material exhibits a valence band structure with significant effective mass anisotropy, forming tubelike energy isosurfaces below the band edge, akin to that of two-dimensional (2D) systems. Such systems can have a large number of high-mobility charge carriers and a beneficial density of states profile. In the calculations, we predicted a maximum p-type figure of merit (𝑧⁢𝑇) of 2.6 at 800 K, in line with previous predictions of high 𝑧⁢𝑇. This high 𝑧⁢𝑇 arises from the low lattice thermal conductivity of 0.35 Wm−1K−1 and the beneficial electronic band structure. The high density of states significantly increased the electron-scattering space, but this effect was largely compensated by reduced scattering rates of electrons with large momentum 𝐪. We further explored the effect of enhancing the low-dimensionality through compressive strain. This increased the p-type power factor by up to 66%; partly due to more strongly pronounced 2D features of the valence band, but primarily due to increased Fermi velocities. However, compressive strain also increased phonon velocities and hence the lattice thermal conductivity. The maximum p-type 𝑧⁢𝑇 thus only increased slightly, to 2.7 at 1% compressive strain. In the conduction band, strain aligned the Γ- and X-centered valleys, resulting in the optimal n-type 𝑧⁢𝑇 increasing from 0.9 to 2.3 at 2% compressive strain. Thus highly strained CsK2⁢Sb has the potential for both good p- and n-type thermoelectricity.

• Grimnes, Øven Andreas; Snyder, G. Jeffrey; Løvvik, Ole Martin & Berland, Kristian (2024). Thermoelectric transport of strained CsK2Sb: Role of momentum-dependent scattering in low-dimensional Fermi surfaces. arXiv.org. ISSN 2331-8422. doi: 10.48550/arXiv.2412.01681. Full text in Research Archive
• Tranås, Rasmus André; Løvvik, Ole Martin & Berland, Kristian (2022). Attaining Low Lattice Thermal Conductivity in Half-Heusler Sublattice Solid Solutions: Which Substitution Site Is Most Effective? Electronic Materials. 3(1), p. 1–14. doi: 10.3390/electronicmat3010001. Full text in Research Archive Show summary

  Low thermal conductivity is an important materials property for thermoelectricity. The lattice thermal conductivity (LTC) can be reduced by introducing sublattice disorder through partial isovalent substitution. Yet, large-scale screening of materials has seldom taken this opportunity into account. The present study aims to investigate the effect of partial sublattice substitution on the LTC. The study relies on the temperature-dependent effective potential method based on forces obtained from density functional theory. Solid solutions are simulated within a virtual crystal approximation, and the effect of grain-boundary scattering is also included. This is done to systematically probe the effect of sublattice substitution on the LTC of 122 half-Heusler compounds. It is found that substitution on the three different crystallographic sites leads to a reduction of the LTC that varies significantly both between the sites and between the different compounds. Nevertheless, some common criteria are identified as most efficient for reduction of the LTC: The mass contrast should be large within the parent compound, and substitution should be performed on the heaviest atoms. It is also found that the combined effect of sublattice substitution and grain-boundary scattering can lead to a drastic reduction of the LTC. The lowest LTC of the current set of half-Heusler compounds is around 2 W/Km at 300 K for two of the parent compounds. Four additional compounds can reach similarly low LTC with the combined effect of sublattice disorder and grain boundaries. Two of these four compounds have an intrinsic LTC above ∼15 W/Km, underlining that materials with high intrinsic LTC could still be viable for thermoelectric applications.

• Schrade, Matthias; Xing, Wen; Thorshaug, Knut & Belle, Branson (2022). Centimeter-Sized Monolayer CVD Graphene with High Power Factor for Scalable Thermoelectric Applications. ACS Applied Electronic Materials. 4(4), p. 1506–1510. doi: 10.1021/acsaelm.1c00995. Full text in Research Archive Show summary

  Among the many extraordinary qualities of graphene, its thermoelectric properties have attracted significant interest, for example, in active cooling applications. Here, we report on the thermoelectric transport properties of centimeter-sized monolayer CVD graphene, electrostatically controlled by a high-capacity ionic gel. The power factor reaches 7 and 5.4 mW m–1 K–2 for hole and electron conduction, respectively, similar to earlier reports obtained for microdevices despite our devices being over 2 orders of magnitude larger. On the basis of these results, we propose nonvolatile ferroelectric polarization as a scalable technology for graphene-based thermoelectric applications.

• Berland, Kristian; Løvvik, Ole Martin & Tranås, Rasmus André (2021). Discarded gems: Thermoelectric performance of materials with band gap emerging at the hybrid-functional level. Applied Physics Letters. ISSN 0003-6951. 119(8). doi: 10.1063/5.0058685. Full text in Research Archive Show summary

  A finite electronic band gap is a standard filter in high-throughput screening of materials using density functional theory (DFT). However, because of the systematic underestimation of band gaps in standard DFT approximations, a number of compounds may be incorrectly predicted metallic. In a more accurate treatment, such materials may instead appear as low band gap materials and could have good thermoelectric properties if suitable doping is feasible. To explore this possibility, we performed hybrid functional calculations on 1093 cubic materials listed in the MATERIALS PROJECT database with four atoms in the primitive unit cell, spin-neutral ground state, and a formation energy within 0.3 eV of the convex hull. Out of these materials, we identified eight compounds for which a finite band gap emerges. Evaluating electronic and thermal transport properties of these compounds, we found the compositions MgSc2Hg and Li2CaSi to exhibit promising thermoelectric properties. These findings underline the potential of reassessing band gaps and band structures of compounds to identify additional potential thermoelectric materials.

• Tranås, Rasmus André; Løvvik, Ole Martin; Tomic, Oliver & Berland, Kristian (2021). Lattice thermal conductivity of half-Heuslers with density functional theory and machine learning: Enhancing predictivity by active sampling with principal component analysis. Computational Materials Science. ISSN 0927-0256. 202. doi: 10.1016/j.commatsci.2021.110938. Full text in Research Archive Show summary

  Low lattice thermal conductivity is essential for high thermoelectric performance of a material. Lattice thermal conductivity is often computed using density functional theory (DFT), typically at a high computational cost. Training machine learning models to predict lattice thermal conductivity could offer an effective procedure to identify low lattice thermal conductivity compounds. However, in doing so, we must face the fact that such compounds can be quite rare and distinct from those in a typical training set. This distinctness can be problematic as standard machine learning methods are inaccurate when predicting properties of compounds with features differing significantly from those in the training set. By computing the lattice thermal conductivity of 122 half-Heusler compounds, using the temperature-dependent effective potential method, we generate a data set to explore this issue. We first show how random forest regression can fail to identify low lattice thermal conductivity compounds with random selection of training data. Next, we show how active selection of training data using feature and principal component analysis can be used to improve model performance and the ability to identify low lattice thermal conductivity compounds. Lastly, we find that active learning without the use of DFT-based features can be viable as a quicker way of selecting samples.

View all works in NVA

• Løvvik, Ole Martin (2025). Tellur - Grunnstoffene. Full text in Research Archive Show summary

  Tellur er en skikkelig luring. Dette grunnstoffet som ligger helt i skjæringspunktet mellom å være en halvleder og en hel-leder (eller metall som det jo også heter) gir oppgav til mange interessante fysiske fenomener, og er en av grunnene til at det er dagens gjest, Kristian Berland fra NMBU, har det som sitt favorittgrunnstoff. Ikke… Fortsett å lese Tellur

• Løvvik, Ole Martin (2024). Lyddemping i tømmervegger, holdbarhet av hamptau og den høyeste mulige temperaturen. [Radio]. Intervju i Abels tårn, NRK P2. Full text in Research Archive
• Grimenes, Øven Andreas; Løvvik, Ole Martin & Berland, Kristian (2024). The Importance of Electronic Relaxation Times in Complex Fermi Surfaces. Full text in Research Archive
• Løvvik, Ole Martin (2024). Sykling, radongass og oversikt over hva vi vet. [Radio]. Intervju i Abels tårn, NRK P2. Full text in Research Archive
• Grimenes, Øven Andreas; Løvvik, Ole Martin; Snyder, G. Jeffrey & Berland, Kristian (2024). Electron scattering in materials with extended Fermi surfaces. Full text in Research Archive
• Tranås, Rasmus André; Berland, Kristian; Grimenes, Øven Andreas & Løvvik, Ole Martin (2024). New materials with very low lattice conductivity from first principles and machine learning. Full text in Research Archive
• Tranås, Rasmus André; Løvvik, Ole Martin & Berland, Kristian (2023). Screening of low lattice thermal conductivity materials using active learning. Full text in Research Archive
• Grimenes, Øven Andreas; Løvvik, Ole Martin & Berland, Kristian (2023). Complex Fermi surfaces for high ZT. Full text in Research Archive
• Løvvik, Ole Martin (2023). Abels tårns årskavalkade – nye materialer og maskinlæring. [Radio]. Intervju i Abels tårn, NRK P2. Full text in Research Archive
• Løvvik, Ole Martin (2023). Plast som leder strøm og andre nye materialer for det grønne skiftet. [Radio]. Intervju i Ekko, NRK P2. Full text in Research Archive
• Løvvik, Ole Martin (2023). Storsatsing på utvikling av materialer med automatisering og kunstig intelligens. [Radio]. Intervju i Abels tårn. Full text in Research Archive
• Løvvik, Ole Martin (2023). Norwegian efforts towards a materials acceleration platform: combining high-throughput experiments, modelling, and machine learning/artificial intelligence. Full text in Research Archive
• Tranås, Rasmus André; Løvvik, Ole Martin & Berland, Kristian (2022). Finding Low Lattice Thermal Conductivity Compounds in Materials Space: Machine Learning with Active Sampling. Full text in Research Archive
• Løvvik, Ole Martin (2022). Hvordan få strøm fra varme – uten bevegelige deler? Full text in Research Archive
• Tranås, Rasmus André; Løvvik, Ole Martin & Berland, Kristian (2022). Alloying leads to drastic reduction of lattice thermal conductivity of half-Heusler compounds. Full text in Research Archive
• Løvvik, Ole Martin (2022). Bør du lukke døra til komfyren når den er ferdig med å steke? [TV]. Intervju i Praktisk Info med Jon Almaas, TV Norge. Full text in Research Archive
• Løvvik, Ole Martin (2022). Thermoelectrics of alloys and intermetallic materials: potential benefits and challenges of using SPS. Full text in Research Archive
• Tranås, Rasmus André; Løvvik, Ole Martin & Berland, Kristian (2022). Material Informatics: Machine Learning with Active Sampling for Small Training Sets. Full text in Research Archive
• Løvvik, Ole Martin; Schrade, Matthias; Grimenes, Øven Andreas; Tranås, Rasmus André & Berland, Kristian (2022). Predictive screening for thermoelectric properties using atomic-scale simulations and machine learning tools. Full text in Research Archive
• Grimenes, Øven Andreas; Løvvik, Ole Martin & Berland, Kristian (2022). Hybrid Functional Screening Identifies Previously Discarded Thermoelectric Materials. Full text in Research Archive
• Grimenes, Øven Andreas; Løvvik, Ole Martin & Berland, Kristian (2022). More than 130 new low bandgap materials identified in a hybrid functional screening study – some with very promising thermoelectric properties. Full text in Research Archive
• Løvvik, Ole Martin; Rahman, Jamil Ur; Finstad, Terje Gunnar; Gunnæs, Anette Eleonora & Carvalho, Patricia Almeida (2022). Thermoelectric properties of multiple component (high entropy) silicides and antimonides. Full text in Research Archive
• Tranås, Rasmus André; Berland, Kristian; Tomic, Oliver & Løvvik, Ole Martin (2022). Discovering Energy Materials: Low Thermal Conductivity Thermoelectric Compounds and Barocaloric Compounds. Full text in Research Archive
• Løvvik, Ole Martin (2021). Screening of materials for thermoelectric conversion of heat into electricity. Full text in Research Archive
• Løvvik, Ole Martin (2021). Vi skal utvikle materialer for fornybar energi ti ganger raskere enn i dag. [Journal]. Apollon : Forskningsmagasin for Universitetet i Oslo. Full text in Research Archive
• Løvvik, Ole Martin (2021). Casting Materials Acceleration Platform and the accelerated search for thermoelectric materials. Full text in Research Archive
• Tranås, Rasmus André; Løvvik, Ole Martin; Tomic, Oliver & Berland, Kristian (2021). Active sample selection for finding low lattice thermal conductivity materials using machine learning. Full text in Research Archive
• Tranås, Rasmus André; Berland, Kristian; Løvvik, Ole Martin & Tomic, Oliver (2021). Fingerprints of low lattice thermal conductivity compounds: half-Heusler case study and beyond. Full text in Research Archive
• Grimnes, Øven Andreas; Berland, Kristian & Løvvik, Ole Martin (2025). Scattering dependent electronic transport in thermoelectric materials with complex band structures. Norges miljø- og biovitenskapelige universitet (NMBU). ISSN 9788257523039. Full text in Research Archive Show summary

  Thermoelectric (TE) materials, which can convert heat to electricity and vice versa, have great potential for energy harvesting from waste heat and have seen renewed interest in recent years. These materials are also used in many niche applications, such as electrical generators in remote locations, sensors, and cooling for medical and scientific equipment. With the advent of computational methods like density functional theory (DFT), modeling is becoming an increasingly more important tool in the TE field. Providing, among other things, electronic band structures and electronic transport properties, DFT combined with Boltzmann transport theory provides both predictive and explanatory power. A limiting factor, however, has long been the modeling of electron scattering, resulting in the use of coarse approximations like the constant relaxation time approximation or textbook results based on heavily simplified models. More accurate modeling and understanding of electronic scattering in complex materials and band structures could lift the field of TE and result in new and improved TE materials. To find suitable candidate materials for more detailed analysis, a highthroughput (HT) screening of the Materials Project database was performed. Based on criteria such as crystal structure complexity, elements, energy above convex hull, magnetic properties, and valence electron count, 6965 materials were selected. For more accurate band gap estimations, the MBJ functional was employed. This revealed band gaps in 189 materials predicted to be metallic with PBE, and an average band gap increase of 0.43 eV in 331 materials with narrow PBE band gaps. The electronic transport properties and thermoelectric figure of merit (zT) were calculated with a constant relaxation time (τ ) of 10 fs and with a lattice thermal conductivity (κℓ) of 5 W/mK for each material. Using the same, constant τ and κℓ for every material results in both false positives and negatives, but still yields a rough ranking on how applicable the electronic structure is for TE purposes. Based on the high-throughput results, two promising full-Heusler materials were selected for further investigation. CsK2Sb, a known photocatode material, has a complex band structure with features akin to those of two-dimensional (2D) quantum wells. The other material, Na2TlSb, also has a quasi-low-dimensional band structure, but one-dimensional (1D), resulting in a Fermi surface consisting of sheets. Both band structures have a rapidly increasing density of states (DOS) with high-mobility carriers close to the band gap, a favorable trait for thermoelectric materials. Despite high DOS generally being associated with excessive electron scattering, this was not found to be the case in these materials, counterintuitive to conventional wisdom. Both materials were also found to have very low κℓ, resulting in a predicted zT of 2.6 and 4.4 for CsK2Sb and Na2TlSb, respectively. Compressive strain was found to improve the electronic transport properties of CsK2Sb by enhancing the 2D features and reducing electron scattering, but also increased κℓ, resulting in only a marginally increased zT of 2.7 at 1% strain. The changes to materials properties and the resulting electronic scattering rates in strained CsK2Sb prompted further investigation. The materials properties of ZrNiSn were systematically adjusted, and the following changes to scattering rates were observed. While most materials properties changed scattering rates as expected, some gave less intuitive results. Other parameters relevant to scattering, such as carrier concentration (Nc) and temperature (T), were also investigated. In addition to directly affecting different scattering mechanisms, both Nc and T change the location of the Fermi level and the magnitude of free-carrier screening, creating highly interconnected results. The low scattering rates could also, in part, be explained by the exotic geometries of the Fermi surfaces. For momentum-dependent scattering mechanisms, the large separation in k-space prevented scattering between large parts of the surface. Further, the combination of wavefunction overlaps and the geometrical momentum relaxation time factor strongly reduced elastic scattering to large portions of the total scattering space. The results for both materials indicate that Fermi surfaces delocalized in k-space tend to reduce electronic scattering. These insights, combined with knowledge of how materials properties affect scattering, can help build understanding and intuition that can aid the discovery and design of new TE materials.

• Tranås, Rasmus André (2023). Identifying Materials with Low Lattice Thermal Conductivity Using Machine Learning and Computational Modeling. Norges miljø- og biovitenskapelige universitet (NMBU). Full text in Research Archive Show summary

  Lattice thermal conductivity is a key materials property in applications related to thermal functionality, such as thermal barrier coatings, thermal conductors in microelectronics, and solid-state waste-heat recovery devices. The lattice thermal conductivity governs the rate of heat energy transfer in thermoelectric materials, which are materials that can directly convert heat to electricity and vice versa. These materials become interesting in applications that require electricity generation or local cooling. Thermoelectric materials depend on a low lattice thermal conductivity to attain high heat-to-electricity conversion efficiency. The materials used in present thermoelectric generators are often based on toxic or scarce elements. New high-efficiency thermoelectric materials are therefore desired for sustainable and environmentally friendly energy harvesting. Two main research challenges are investigated in this thesis: 1) reducing the lattice thermal conductivity to enhance thermoelectric performance, and 2) identifying new compounds with low lattice thermal conductivity. Addressing these challenges experimentally is a daunting task -- especially for 100s or 1000s of compounds -- as experiments are costly, time-consuming, and require expert domain knowledge. This thesis, therefore, relies on lattice thermal conductivity from theoretical calculations based on quantum mechanical simulations. Addressing challenge 1), the lattice thermal conductivity of 122 half-Heusler compounds is calculated using density functional theory and the temperature-dependent effective potential method. Phonon scattering from partial sublattice substitutions and grain boundaries are included in calculations, in an attempt to reduce the lattice thermal conductivity. We find that isovalent substitutions on the site hosting the heaviest atom should be performed to optimally reduce the lattice thermal conductivity in most half-Heuslers. Compounds with large atomic mass differences can have a large drop in lattice thermal conductivity with substitutions. Examples of such compounds are AlSiLi and TiNiPb, which achieve a $\sim 70$~\% reduction of their lattice thermal conductivity when substituting Si by Ge and Pb by Sn at 10~\% concentration. The reduction from additional scattering mechanisms enables a handful half-Heuslers to attain a lattice thermal conductivity close to $2$~W/Km at 300~K. Calculations for full-Heusler $\rm{AlVFe}_2$ reveal that the introduction of 15~\% Ru substitutions on the Fe-site and 100~nm grain boundaries can reduce the lattice thermal conductivity from 46~W/Km to 7~W/Km. Tackling challenge 2) is done by computational screening for low lattice thermal conductivity compounds. Coupling calculations with machine learning accelerates the screening. When training the machine learning model on calculated lattice thermal conductivities, it learns to recognize descriptor patterns for compounds with low lattice thermal conductivity. The size of the training set is limited by the large computational cost of calculating lattice thermal conductivity. It is therefore challenging to obtain a diverse set of training compounds, especially so because low lattice thermal conductivity compounds tend to be rare. We find that including certain compounds in the training can be crucial for identifying low lattice thermal conductivity compounds. Active sampling enables scouting of the compound space for compounds that should enter the training set. Principal component analysis and Gaussian process regression are used in the active sampling schemes. With Gaussian process regression we screen 1573 cubic compounds, where 34 have predicted lattice thermal conductivity $\leq 1.3$~W/Km at 300 K -- as well as electronic band gaps -- indicating that they could be potential thermoelectric compounds. The findings in this thesis show that certain compounds could have a drastic reduction in the lattice thermal conductivity with sublattice substitutions. Thermoelectric compounds with favorable electronic properties -- but high lattice thermal conductivity -- can be investigated in future studies if there is a potential for a large drop in the lattice thermal conductivity with sublattice substitutions. The machine learning and active sampling schemes are scalable, and future works could expand upon this thesis by including different compound classes in training and screening. This would enlarge the search space for promising thermoelectric compounds, increasing the likelihood of encountering high-efficiency candidates. It is also possible to combine the two challenges faced in this thesis. A machine learning model can be trained to predict the lattice thermal conductivity of compounds with sublattice substitutions. This would further increase the pool of possible compounds where promising thermoelectric compounds could reside.

View all works in NVA