Page 137 - Read Online

Chen et al. Energy Mater. 2025, 5, 500120 https://dx.doi.org/10.20517/energymater.2024.311 Page 13 of 21

Table 1. The survey of this review refers to the descriptors, targets, and findings of the study related to ML in material chemistry

Sample Best ML
Publication Samples Descriptors Targets Finding
source algorithms
[51]
Juneja et al., 2019 Material 2,162 binary, ternary, and Maximum phonon frequency, integrated Gruneisen parameter up Analyzed the κ for 120 Low κ materials between 0.13 GPR
l l
project quaternary compounds to 3 THz, average atomic mass, and volume of the unit cell materials in relation to the and 0.98 (CsK Sb, TlI, Ba BiAu,
2
2
parameters SrTePd, Ba SbAu, TlBr, Cs Se,
2
2
PbI , LiFeP, TlCl, Ba AgSb, PbI ,
2 2 2
LaCoTe)
[55]
Zhang et al., 2018 Previous 93 binary Electronegativity, atomic radius, effective nuclear charge, Vander Proposed a method to Increased accuracy about 2.1% Kernel ridge
reports semiconductors Waals radius, covalent radius, row number in the periodic table, increase prediction than before. Used to more regression
block number, enthalpy of formation of gaseous atoms, ionization accuracy by incorporating accurately predict the κ of
l
energy, and valence number property estimation in the materials having more than 0.1
-1 -1
feature space to establish Wm K
ML models
[52]
Juneja et al., 2020 DFT + 2,838 compounds Boiling point, melting point, specific heat, molar specific heat, Found the common Common parameters driving GPR
Boltztrap molar volume, heat of fusion, heat of vaporization, Pauling parameters influencing chemical bonding such as
code electronegativity, first ionization energy, group and period in the electrical and thermal electronegativity, bond
periodic table, elemental thermal conductivity, atomic number, transport properties. strength, bond distance,
atomic mass, covalent radius, van der Waals radius, density, the Tested with 135 volume, and coordination
average bond distance, average bond strength, volume per atom, compounds number
volume of cell, and coordination number
[53]
Juneja et al., 2020 DFT 2,838 nonmetallic Boiling point, melting point, specific heat, molar specific heat, Found the correlation The bond strength obtained GPR
compounds after high molar volume, heat of fusion, heat of vaporization, Pauling between the chemistry of was weak for high coordination
throughput screening 185 electronegativity, first ionization energy, group and period in the bonding and κ numbers, and the bonding
l
compounds were periodic table, elemental thermal conductivity, atomic number, distance was large
analyzed atomic mass, covalent radius, van der Waals radius, density, the
average bond distance, average bond strength, volume per atom,
volume of cell, and coordination number
[50]
Tewari et al., 2020 DFT 315 compounds lattice energy per atom, atom density, band gap, mass density, Segregating and Identified low κ among 315 XGBoost classifier
l
and oxygen ratio by transition metal atoms discovering material based transition metal oxides, values and regression
-1 -1
on κ less than 5 Wm K (cubist, kernel ridge,
l
and Gaussian
process)

potential to be expanded to efficient synthesis of chalcogenide-based TE materials using CVD techniques.

Among chalcogenides, Bi Te -based materials show the best TE properties. Wang et al. have applied ML to optimize hot-extruded Cu Bi Te 2.85+y Se TE
x
2
0.15
3
2
materials for the first time . Principal component analysis (PCA) has been employed to characterize multiple variables. It has been found that the extrusion
[60]
temperature and Cu content are the most important parameters in the Cu Bi Te 2.85+y Se -based TE material processing design. This study developed artificial
2
0.15
x
neural network (ANN), SVR, and RF regression models to analyze the relationships between processing conditions, microstructural information, and TE
properties.