Laser powder bed fusion processed LaCe(Fe, Mn, Si)₁₃ lattices for magnetic refrigeration: Process optimization, microstructure, and magnetocaloric performance

Abstract

In this study, the optimal laser powder bed fusion (LPBF) processing parameters for fabricating fully dense LaCe(Fe,Mn,Si)13 thin walls have been identified through a machine-learning approach based on the Gaussian process regression (GPR) model. All the specimens and components were fabricated using a continuous laser source. The relationship between the defect fraction of the fabricated thin wall and the line energy density (EL) and hatch (h) is established. The measured defect fraction of specimens fabricated using the validation data sets was very well in agreement with the predicted values of the GPR model, with an error of less than 1%. The microstructure of as-fabricated lattices is contained by α-Fe phases, LaFeSi phases, NaZn13-type phases, and the La/Ce/Si rich phases, which has an amorphous matrix embedded with nanocrystalline. The microstructure of the HTHed lattices presents the α-Fe phases, LaFeSi phases, and NaZn13-type phases. The diamond lattice has high heat exchange efficiency among the four lattices because of its large surface area (1577.1 mm2) and excellent thermal conductivity. Although the LPBF parameters are the same, the maximum isothermal magnetic entropy change (ΔSm) and Tc of the four samples differ. The X-ray powder diffraction test confirms that the HTHed Tube exhibits the highest volume of the NaZn13-type phase. The HTHed Tube saw ΔSm, about 1.63 J kg−1K−1 at 264.5 K. The ΔSm of HTHed Diamond around Tc (1.13 J kg−1K−1 at 219.5 K) is slightly higher than HTHed Gyroid (0.99 J kg−1K−1 at 250.5 K). Essentially, our work accelerates the search for optimal process parameters and guides the direction for lattice design of LaCe(Fe,Mn,Si)13.

Keywords

• Machine learning
• Laser powder bed fusion (LPBF)
• Microstructure
• Heat exchange efficiency
• X-ray powder diffraction
• And magnetocaloric performance