GeTe exhibits excellent p-type medium-temperature thermoelectric properties with low toxicity and good mechanical characteristics, making it highly promising for development in the thermoelectric field. However, GeTe is prone to producing Ge vacancies, leading to high p-type carrier concentration, which results in elevated electronic thermal conductivity and a low Seebeck coefficient. This study systematically analyzes intrinsic and extrinsic defects in GeTe and its alloys, focusing on reducing p-type carrier concentration through first-principles calculations. The results reveal that substituting Ge-sites with Bi (Bi_Ge) yields lower donor defect formation energy, effectively reducing p-type carrier concentration of GeTe and its alloys compared to other elemental doping. Additionally, alloying with certain elements, such as Pb, proves favorable for decreased p-type carrier concentration due to lowered energy levels of valence band maximum (VBM). Inspired by this, screening divalent elements for alloying on Ge-sites reveals that Sr, Ba, Eu, and Yb substantially reduce the VBM of GeTe. Further calculations for Ba and Yb-alloyed GeTe confirm changes in formation energies for donor (favorable) and acceptor (unfavorable) defects. Our work provides a systematic investigation of intrinsic and various extrinsic doping defects in GeTe and its alloys, shedding light on possible strategies of optimizing carrier concentration in these compounds.

Thermoelectric materials are environmental-friendly energy conversion materials. Their performance optimization is a complex issue of multi-parameter coordination, which becomes a challenge. Although the computational simulation and experimental methods for thermoelectric materials have developed rapidly, the efficiency of searching thermoelectric materials still needs to be further improved. Machine learning has some advantages of low computational cost and high prediction speed, which can shorten the search process and accelerate the corresponding studies on the structure and performance optimization of thermoelectric materials. This review introduced the research progress on machine learning for small sample numerical data (data volume is about 10²), large sample numerical data (data volume >10⁴) and image data in thermoelectric materials from the perspective of data types. Moreover, different machine learning algorithm models used for the structure and performance of thermoelectric materials in different data types were discussed. In addition, the future development and application direction were also prospected.

Over the years, the fact that the quaternary diamond-like thermoelectric materials show much lower carrier mobilities than ternary compounds remains mysterious. In this work, by adopting first-principles defect chemistry and electrical transport calculations, the fundamental origin of the difference on carrier mobility between quaternary and ternary diamond-like compounds is addressed, exemplified by Cd₂Cu₃In₃Te₈. The results of defect chemistry show that the main intrinsic defects in quaternary compound Cd₂Cu₃In₃Te₈ are substitutional defects, i.e., Cd_In and Cd_Cu, differing from the copper vacancy defect in ternary Cu-based compound such as CuInTe₂. The low defect formation energies in Cd₂Cu₃In₃Te₈ result in high defect concentrations, which is caused by the similar atomic radii and electronegativities between Cd—In and Cd—Cu. Further calculations show that the low-energy defects are mainly located around the valence band maximum in Cd₂Cu₃In₃Te₈. The electrical transport calculations, considering both the acoustic phonon scattering and ionized impurity scattering, demonstrate that mainly due to the higher concentration of the ionized defects, the mobility of the quaternary Cd₂Cu₃In₃Te₈ is much lower than that of ternary CuInTe₂. Our work sheds light on the intrinsic defects in quaternary diamond-like compounds and their influence on charge transport.

The application of machine learning (ML)-based methods to the study of thermoelectric (TE) materials is promising. Although conventional ML algorithms can achieve high prediction performance, their lack of interpretability severely obstructs researchers from extracting material-oriented insights from ML models. In this work, high ML-based prediction performance was achieved with respect to TE power factors (PFs), and the results were well understood by the SHapley Additive exPlanations (SHAP), a method to identify the correlations between targets and descriptors. We designed a robust PF prediction model for diamond-like compounds via a stacking technique, and the model achieved a coefficient of determination value above 0.95 on the test set. From the SHAP analysis, the PFs were negatively correlated with electronegativity and positively correlated with the descriptor "volume per atom" based on the previously reported dataset. TE domain knowledge was adopted to understand these correlations. This work shows that ML models can achieve high accuracy while exhibiting good interpretability, making them useful for materials scientists.

Electron-phonon coupling (EPC) is a key factor for thermoelectric properties of materials. In this paper, the thermoelectric properties of zinc-blende chalcogenides (p-type) ZnS and ZnSe have been studied through full evaluation of EPC from first-principles, including the influences on both electrical and thermal transport. We find that the polar longitudinal optical phonon scattering is the dominant mechanism for electrical transport. Due to the triple degeneracy near the valence band maximum, the inter-band scattering also has detrimental contributions to the electrical conductivities. For phonon transport, it shows that the lattice thermal conductivity can be reduced by the electron-phonon scattering significantly at high carrier concentrations (e.g., at 300 K with 10²¹ cm⁻³ of hole, the reduction is ~24.9% for ZnS and ~28.4% for ZnSe, respectively). Finally, the p-type thermoelectric figure of merit (ZT) of two systems have been obtained, which are 0.129 for ZnS and 0.141 for ZnSe, at 700 K with their respective optimal hole concentrations. Our work provides a complete and in-depth study of thermoelectric properties in chalcogenides ZnX from the role of EPC. The results suggest EPC plays an important role on the thermoelectric properties and thus full evaluation of EPC is necessary especially for polar materials.

The complex intrinsic defects of four ternary diamond-like ABX₂ (A: Cu/Ag; B: In; X: Te/Se) compounds are studied by first-principles calculations. The results show that Cu-based compounds are likely to generate acceptor V_A defect with p-type, while Ag-based compounds form donor B_A with n-type. The reasons are analyzed by the newly introduced density of energy (DOE).

Superionic conductors, which exhibit liquid-like phonon transport but crystal-like carrier transport, have attracted great attention and broad research interest in the thermoelectric community. Ag₂Te is a superionic conductor; however, its small band gap and large Ag vacancy formation energy impede its application as a prominent p-type thermoelectric material. In this work, synergistic optimization of the thermoelectric performance of Ag₂Te through Cu substitution is realized through a combination of experimental and theoretical efforts. For the electrical transport, Cu substitution systematically increases the band gap of Ag₂Te and reduces the cation vacancy formation energy. These two beneficial effects simultaneously increase the electrical conductivity and suppress the bipolar effect, thereby greatly enhancing the p-type electrical transport properties of Ag₂Te. For the thermal transport, alloying Cu₂Te with Ag₂Te significantly reduces the thermal conductivity through not only point defect scattering but also softening of the interatomic interactions. The latter is attributed to the relatively small Cu atoms vibrating in the oversized 8c sites. This two-fold optimization results in maximum thermoelectric figure of merit zT values of over 1.3 at 773 K for both Ag_1.2Cu_0.8Te and AgCuTe, demonstrating the great potential of Ag_2–xCu_xTe as a promising p-type thermoelectric material system.