1.

Record Nr.

UNINA9911007485503321

Autore

Amano Tomohito

Titolo

First-Principles and Machine Learning Study of Anharmonic Vibration and Dielectric Properties of Materials / / by Tomohito Amano

Pubbl/distr/stampa

Singapore : , : Springer Nature Singapore : , : Imprint : Springer, , 2025

ISBN

981-9640-24-5

Edizione

[1st ed. 2025.]

Descrizione fisica

1 online resource (XVIII, 219 p. 52 illus., 45 illus. in color.)

Collana

Springer Theses, Recognizing Outstanding Ph.D. Research, , 2190-5061

Disciplina

530.10285

Soggetti

Mathematical physics

Computer simulation

Machine learning

Semiconductors

Condensed matter

Materials science - Data processing

Electronic structure

Quantum chemistry - Computer programs

Computational Physics and Simulations

Machine Learning

Condensed Matter Physics

Condensed Matter

Electronic Structure Calculations

Lingua di pubblicazione

Inglese

Formato

Materiale a stampa

Livello bibliografico

Monografia

Nota di contenuto

Chapter 1 Introduction -- Chapter 2 Density Functional Theory -- Chapter 3 Anharmonic Phonon Theory -- Chapter 4 Modern Theory and Machine Learning of Polarization -- Chapter 5 Dielectric Properties of Strongly Anharmonic TiO2 -- Chapter 6 Dielectric Properties of Liquid Alcohols and Its Polymers -- Chapter 7 Conclusion.

Sommario/riassunto

The book presents the author's development of two first-principles methods to calculate dielectric properties of materials based on anharmonic phonon and machine learning, and demonstrates an in-depth analysis of anharmonic crystals and molecular liquids. The



anharmonic phonon method, combined with Born effective charges, is useful to study dielectric properties of crystals. The recently developed self-consistent phonon theory (SCPH) enables accurate simulations in strongly anharmonic materials. The author reveals that the combination of SCPH with the four-phonon scattering term accurately reproduces experimental spectra, and discusses how anharmonic phonon self-energies affect the dielectric properties. The second method is molecular dynamics with Wannier centers—the mass centers of Wannier functions. The author constructs a machine learning model that learns Wannier centers for each chemical bond from atomic coordinates to accurately predict the dipole moments. The developed method is, in principle, applicable to molecules of arbitrary size. Its effectiveness is demonstrated and the dielectric properties of several alcohols, including dipole moments, dielectric constants, and absorption spectra, are analyzed. This book benefits students and researchers interested in anharmonic phonons, machine learning, and dielectric properties.