Abstract
Layered IV-V-VI semiconductors have immense potential for thermoelectric (TE) applications due to their intrinsically ultralow lattice thermal conductivity. However, it is extremely difficult to assess their TE performance via experimental trial-and-error methods. Here, we present a machine-learning-based approach to accelerate the discovery of promising thermoelectric candidates in this chalcogenide family. Based on a dataset generated from high-throughput ab initio calculations, we develop two highly accurate-and-efficient neural network models to predict the maximum ZT (ZTmax) and corresponding doping type, respectively. The top candidate, n-type Pb2Sb2S5, is successfully identified, with the ZTmax over 1.0 at 650 K, owing to its ultralow thermal conductivity and decent power factor. Besides, we find that n-type Te-based compounds exhibit a combination of high Seebeck coefficient and electrical conductivity, thereby leading to better TE performance under electron doping than hole doping. Whereas p-type TE performance of Se-based semiconductors is superior to n-type, resulting from large Seebeck coefficient induced by high density-of-states near valence band edges.
Highlights
Due to the depletion of fossil fuels and ever-increasing environmental concerns, it is urgent to explore sustainable and clean energies, which has become a global consensus[1]
The optimal doping type (i.e., n-type or p-type) at which the maximum ZT value is achieved for a Firstly, we examined three popular types of activation functions, namely, Sigmod, Tanh[36], and rectified linear unit (Relu)[37]
The results demonstrate that the p-type doped power factor (PF) of Pb2Sb2S5 is moderately larger than that of Pb2Sb2Se5 and Pb2Sb2Te5 in the whole studied concentration range, owing to the higher S of the former
Summary
Due to the depletion of fossil fuels and ever-increasing environmental concerns, it is urgent to explore sustainable and clean energies, which has become a global consensus[1]. It is noted that the chemical composition of the current advanced TE materials is mainly composed of main group IV to VI elements, such as PbTe6, Bi/Sb−Te-based alloys[20,21], and recently reported Sb2Si2Te622, most of which adopt layered crystal structures All these IV-V-VI compounds exhibit ultralow lattice thermal conductivity (0.28−2.02 W m−1 K−1 at 300 K) and suitable electronic bandgaps (>0.1 eV)[19] within the range of good TE materials[23]. Based on extensive examinations of hyperparameters of deep neural networks, two highly accurate ML models are successfully developed and used to predict the maximum ZT (ZTmax) and corresponding doping type at different temperatures for the remaining 30 semiconductors, respectively Among this IV-V-VI family, we identify several compounds that exhibit good TE performance with ZTmax values above 0.8 at 650 K, especially the n-type Pb2Sb2S5 achieving a high ZTmax of 1.2. A set of appropriate hyperparameters is critical for obtaining effective neural network models, which generally includes the activation
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
