Deep learning (DL) is one of the fastest growing topics in materials data science, with rapidly emerging applications spanning atomistic, image-based, spectral, and textual data modalities. DL allows analysis of unstructured data and automated identification of features. Recent development of large materials databases has fueled the application of DL methods in atomistic prediction in particular. In contrast, advances in image and spectral data have largely leveraged synthetic data enabled by high quality forward models as well as by generative unsupervised DL methods. In this article, we present a high-level overview of deep-learning methods followed by a detailed discussion of recent developments of deep learning in atomistic simulation, materials imaging, spectral analysis, and natural language processing. For each modality we discuss applications involving both theoretical and experimental data, typical modeling approaches with their strengths and limitations, and relevant publicly available software and datasets. We conclude the review with a discussion of recent cross-cutting work related to uncertainty quantification in this field and a brief perspective on limitations, challenges, and potential growth areas for DL methods in materials science. The application of DL methods in materials science presents an exciting avenue for future materials discovery and design. 

/pdf/recent-advances-and-applications-of-deep-learning-methods-in-47f5g19w.pdf

Recent advances and applications of deep learning methods in materials science

Thermoelectric coolers as thermal management systems for medical applications: Design, optimization, and advancement

With the fast evolution in greenhouse gas (GHG) emissions (e.g., CO2, N2O) caused by fossil fuel combustion and global warming, climate change has been identified as a critical threat to the sustainable development of human society, public health, and the environment. To reduce GHG emissions, besides minimizing waste heat production at the source, an integrated approach should be adopted for waste heat management, namely, waste heat collection and recycling. One solution to enable waste heat capture and conversion into useful energy forms (e.g., electricity) is employing solid-state energy converters, such as thermoelectric generators (TEGs). The simplicity of thermoelectric generators enables them to be applied in various industries, specifically those that generate heat as the primary waste product at a temperature of several hundred degrees. Nevertheless, thermoelectric generators can be used over a broad range of temperatures for various applications; for example, at low temperatures for human body heat harvesting, at mid-temperature for automobile exhaust recovery systems, and at high temperatures for cement industries, concentrated solar heat exchangers, or NASA exploration rovers. We present the trends in the development of thermoelectric devices used for thermal management and waste heat recovery. In addition, a brief account is presented on the scientific development of TE materials with the various approaches implemented to improve the conversion efficiency of thermoelectric compounds through manipulation of Figure of Merit, a unitless factor indicative of TE conversion efficiency. Finally, as a case study, work on waste heat recovery from rotary cement kiln reactors is evaluated and discussed.

Thermal Management Systems and Waste Heat Recycling by Thermoelectric Generators—An Overview

The anharmonic lattice dynamics of rock-salt thermoelectric compounds SnTe and PbTe are investigated with inelastic neutron scattering (INS) and first-principles calculations. The experiments show that, surprisingly, although SnTe is closer to the ferroelectric instability, phonon spectra in PbTe exhibit a more anharmonic character. This behavior is reproduced in first-principles calculations of the temperature-dependent phonon self-energy. Our simulations reveal how the nesting of phonon dispersions induces prominent features in the self-energy, which account for the measured INS spectra and their temperature dependence. We establish that the phase space for three-phonon scattering processes, combined with the proximity to the lattice instability, is the mechanism determining the complex spectrum of the transverse-optic ferroelectric mode.

Phonon self-energy and origin of anomalous neutron scattering spectra in SnTe and PbTe thermoelectrics

Development of the high performance thermoelectric unicouple based on Bi2Te3 compounds

Solid-state energy conversion through thermoelectric effects remains the technology of choice for space applications for which, their low energy conversion efficiency is largely outweighed by the reliability and technical requirements of the mission. Radioisotope thermoelectric generators (RTGs) enable the direct conversion of the heat released by nuclear fuel into the electrical power required to energize the scientific instruments. The optimization of the conversion efficiency is intimately connected to the performances of the thermoelectric materials integrated which are governed by the transport properties of these materials. Recent advances in the design of highly efficient thermoelectric materials raise interesting prospects to further enhance the performances of RTGs for future exploratory missions in the Solar system. Here, we briefly review the knowledge acquired over the last years on several families of thermoelectric materials, the performances of which are close or even higher than those conventionally used in RTGs to date. Issues that remain to be solved are further discussed.

/pdf/thermoelectric-materials-for-space-applications-1r5qmw9566.pdf

Thermoelectric materials for space applications

The binary copper chalcogenides Cu2−δX (X = S, Se, and Te) have recently gained significant interest due to their high thermoelectric performance at moderate temperatures. In an effort to unveil ne...

Oussama El Hamouli

Papers

Thermoelectric materials for space applications

Phase-Transition-Enhanced Thermoelectric Transport in Rickardite Mineral Cu3‒xTe2