From experimentation to prediction: comprehensive study of dielectric properties through experimental research and theoretical modeling

Abstract

This study discusses the experimental findings on the frequency & temperature influences on the dielectric (constant (ε1) and loss (ε2)) of some chalcogenide materials based on Se83Bi17 composition performed in the temperature range 303 K–393 K and frequency range (100–1000000 Hz). As the frequency increases, multiple polarization mechanisms contribute to the reduction of the dielectric constant. The addition of germanium (Ge) to a composition increases ε1 more than tellurium (Te). The dielectric loss decreases with frequency while increasing with temperature and AC conductivity. Understanding these behaviors is important for material characterization and applications in fields like electronics and solar cells. The theoretical section introduces adaptive neuro-fuzzy inference systems (ANFIS), which are utilized in the estimation of the dielectric characteristics of Se83Bi17 (SB), Se83Bi17Te5 (SB-T), and Se83Bi17Ge5 (SB-G). Experimentation-related data are a source of input. ANFIS model of the Takagi–Sugeno type has been trained. With MATLAB, the most effective networks are created. The outcomes of the ANFIS modeling are exceptional. The accuracy of the modeling process is due to the error values. This study demonstrates that the ANFIS technique can accurately anticipate the dielectric properties of the compositions under consideration when they are formed into thin films. The ANFIS can describe the experimental data of the dielectric (constant (ε1) and loss (ε2)) of some chalcogenide materials for all the mentioned temperatures and frequencies. This leads to using the ANFIS model to produce the dielectric (constant (ε1) and loss (ε2)) of some chalcogenide materials for various temperatures and frequencies which there are no experimental data yet to compare with.

Machine learning-assisted prediction of the electronic features of a Schottky diode interlaid with PVP:BaTiO3 composite

“…Improved optical, electronic, and dielectric characterizations of the MS instruments interlaid with a polymer/organic layer without/with dopants of metallic, metallic oxide, and graphene nanoparticles can also be attributed to factors like appropriate stableness of mechanical and dielectric specifications, great storage capacity of charge/energy, simplicity process methods, small weight, and sufficient flexibility [12][13][14][15]. Therefore, in order to improve the efficacy of the MS configuration, potential BH needs to be controlled and manipulated.…”

“…The R s quantity of a diode, the surface/interface states/traps (D it /N ss ) produced amidst the semiconductor and interfacial layer, the non-uniformity of the electric potential barrier and the interlayer deposited at the M/S interface, and other factors are often responsible for these discrepancies. As a result, it makes sense to use other techniques that can cut down on the number of experiments conducted, saving money and effort [12,16,17]. Machine learning (ML) is currently the most popular approach to achieving this goal.…”