Search

The discovery of new materials is one of the driving forces to promote the development of modern society and technology innovation, the traditional materials research mainly depended on the trial-and-error method, which is time-consuming and laborious. Recently, machine learning (ML) methods have made great progress in the researches of materials science with the arrival of the big-data era, which gives a deep revolution in human society and advance science greatly. However, there exist few systematic generalization and summaries about the applications of ML methods in materials science. In this review, we first provide a brief account of the progress of researches on materials science with ML employed, the main ideas and basic procedures of this method are emphatically introduced. Then the algorithms of ML which were frequently used in the researches of materials science are classified and compared. Finally, the recent meaningful applications of ML in metal materials, battery materials, photovoltaic materials and metallic glass are reviewed.

AlCoCrxFeNi high-entropy alloys (HEAs) coatings were prepared on the surface of 45# steel by laser cladding with different contents of Cr (x = 0.50, 0.75, 1.00, 1.50, 2.00). A model was established to estimate the content of each component in the coatings, based on which three parameters including mixing entropy (△Smix), net driving factor (Ω) (Ω = Tm⋅△Smix/|△Hmix| Ω = T m ⋅ S m i x / | △ H m i x | ), and atomic size difference (δ) were calculated to precisely predict phase constituents in the coatings. The evolution in microstructure and corrosion behavior of the coatings with the change in x was investigated comprehensively. Results indicated that the dilution rate of the coatings was gradually increased with the increase in x. The coatings with x = 1.00 and 1.50 were composed of a single solid solution with the face-centered cubic (FCC) structure. A small number of fine equiaxial FeAl3 particles were also observed in the coatings with x = 0.50, 0.75, 2.00. Corrosion resistance of the coatings was improved first and then decreased with the increase in x in the electrochemical tests carried out in 0.1 mol L−1 HCl and 3.5 wt% NaCl solutions. The similar phenomenon was observed in the immersion tests. The optimum corrosion resistance was obtained in the coating with x = 1.50, in which the content of Cr reached the maximum value. Corrosion resistance of 45# steel could be considerably improved by laser cladding an AlCoCrxFeNi HEA coating on its surface.

Based on the analyses of a large number of investigations related to laser cladding, the modified criterions in terms of the atomic size difference (δ), the mixing enthalpy ( Δ H m i x ), the mixing entropy ( Δ S mix ) and the specific laser energy (K) had been proposed to precisely predict phase constituents of the multi-component laser-clad coatings fabricated in a typical nonequilibrium state. The criterions were followed: 10.8 ≤ ΔSmix ≤ 16.2 J·K−1·mol−1, −17 ≤ ΔHmix ≤ 7 kJ·mol−1, 0 ≤ δ ≤ 14 and 0.04 ≤ K ≤ 0.14 kJ·mm−2, in which the simple solid solution was obtained in a multi-component alloy coating prepared by laser cladding. Among them, δ as the most important criterion was double that proposed in the previous studies, K as a new criterion was first proposed. Moreover, FeCrCoAlNiMox (x = 0.25, 0.75, 1.00, 1.25 and 1.50) high-entropy alloy (HEA) coatings were synthesized on 45# steel substrates by laser cladding, and the investigations into the effect of Mo on dilution rates, macro morphologies, microstructures and mechanical properties (microhardness and wear resistance) of the coatings were carried out. Results showed that the dilution rates of the coatings presented the increasing tendency with the increase in x. The coatings with x = 0.75/1.00 were composed of the single solid solution with a BCC structure. Besides the solid solution, an intermetallic compound rich in Fe and Mo (CrFeNiMo) was also formed in the other three coating with x = 0.25/1.25/1.50. A model was established to estimate the content of each component in the coatings, based on which Δ S mix , Δ H m i x and δ were further calculated. The calculated results of the four parameters ( δ , Δ H m i x , Δ S mix and K) in the five coatings confirmed the validity of the above criterions. With the increasing in x, the average microhardness of the coatings (660, 664, 706, 681 and 642 HV0.2) was first increased and then decreased. The maximum value was obtained in the coating with x = 1.00, which was improved about 3.5 times when compared with that of the substrate (about 200 HV0.2). The average friction coefficient (0.514, 0.462, 0.443, 0.494 and 0.503), wear volumes (0.281, 0.154, 0.119, 0.177 and 0.254 mm3) and wear rates (8.29, 5.64, 4.41, 6.28, and 7.43 mm3·N−1·m−1) of the coatings also confirmed that the coating with x = 1.00 exhibited the optimum wear resistance among the five coatings, which were significantly improved when compared those of the substrate (0.579, 0.34 1 mm3 and 8.42 mm3·N−1·m−1).

Porous Co3O4/reduced graphene oxide (RGO) composites with excellent supercapacitive performance were successfully synthesized on nickel foam (NF) by electrophoretic deposition followed by thermal reduction, and a hydrothermal method. The effects of the introduction of RGO and molar ratio of hexamethylenetetramine (HMT) and Co(NO3)2 on morphology and electrochemical performance of the deposits were investigated. The morphological evolution of Co3O4 prepared on RGO with the increase in molar ratio was followed: dispersed flower-like clusters composed of fine nanosheets (1:2) → coarse flower-like clusters (1:1) → a honeycomb-like structure consisting of fine nanosheets (2:1) → two layers of structures (initial honeycomb-like structure covered with coarse flower-like structure in 3:1 and 4:1). A highest specific surface area was obtained in the sample prepared in the HMT/Co(NO3)2 M ratio of 2:1. The introduction of RGO contributed to the increase in specific surface area of the deposits. For the three-electrode testing system, the specific capacitance of the Co3O4/RGO deposit at the molar ratio of 2:1 reached the highest values of 1138.11 F g−1 at a current density of 1 A g−1, 800.71 F g−1 at a scanning rate of 5 mV s−1, which were higher than those of Co3O4. This Co3O4/RGO deposit also demonstrated the outstanding cycling stability (the specific capacitance of the 80.67% of initial value was retained after 5000 cycles). For the two-electrode testing system, the as-fabricated asymmetric supercapacitor device achieved a high specific capacitance of 108.87 F g−1 at a current density of 1 A g−1 with a stable operational voltage of 1.45 V and a high energy density of 647.05 Wh⋅kg−1.