Prediction of density in amorphous and nanocrystalline soft magnetic alloys: A data mining approach

Abstract

The density of amorphous and nanocrystalline soft magnetic alloys is often vital to accurately estimate the saturation magnetic polarization and other properties. However, this information is often missing from the literature of amorphous and nanocrystalline alloys, possibly due to the challenges associated with using the Archimedes buoyancy method with high specific surface area melt-spun ribbons. In this work, a review of the literature has been conducted to determine the density (ρ in 103 kg/m3) of alloys in an amorphous state and, in combination with our experimental observations, is found to be well described by the following regression function: ρ = 7.751 + 0.010CCo + 0.010CNi + 0.027CNb + 0.017CCu – 0.020CB – 0.039CSi – 0.032CP, where Ci is the content of i element in at.% and the fitting error is within approximately 1%. Unlike previous works in this area, this study estimated the change in density upon amorphization and subsequent nanocrystallization for melt-spun Fe-B alloys using a He gas pycnometer. A selection of other common soft magnetic alloy systems including FINEMET, NANOPERM, NANOMET and HiB-NANOPERM are also tested. The change in density upon nanocrystallization is seen to vary between 0.5 and 3%, depending on the alloy composition. This change in density is well described by the volume-weighted average densities of the residual amorphous and crystalline phases. The difference in density between a crystalline precursor ingot and after amorphization by melt spinning was observed to vary between 2 and 3%.