The taxonomy label points at bulk engineering magnet classes (sintered rare-earth permanent magnets, ceramic ferrites, insulated iron-powder soft magnetic composites), but every supplied source is upstream magnetism physics or computational magnetism, so what can actually be assessed here is a fast-improving simulation and discovery toolchain plus a set of cryogenic device demonstrations, not the industrial materials themselves.
Summary
In engineering terms, “magnetic materials” splits into hard magnets that hold a remanent field (the neodymium-iron-boron sintered class, and cheaper ceramic ferrites) and soft magnets that carry alternating flux with low loss (laminated steels, and soft magnetic composites made from insulated iron powder pressed into a core). Those are the materials the concept name implies. None of the 18 supplied sources addresses them. This page therefore assesses what the evidence base does contain, and says explicitly where it cannot support a judgement.
Evidence base
- 5 May 2026: a fully ab-initio spin-lattice dynamics method integrated into VASP recovers the correct magnetic ground state for four materials from random initial conditions and cuts magnetic ML potential energy MAE by up to about one order of magnitude for BiFeO3.
- 28 May 2026: the EPW package is extended to spin-polarised electron-phonon interpolation and validated on ferromagnetic iron and nickel, including finite-temperature carrier resistivity.
- 30 June 2026: a magnetism-aware crystal graph network improves total magnetic moment prediction from a CGCNN test MAE of 2.54 muB to 2.02 muB, raising test R2 from 0.644.
- 2 July 2026: anomalous Hall angle in Fe-doped Co3Sn2S2 nanoflakes is modulated to tan(thetaA) of 0.46, around 25 degrees, versus 0.1 to 3 degrees typical for magnetic materials, giving 7028 microohm-cm/T sensitivity and 23.5 nT/Hz^0.5 detectability at 1 Hz.
- 2 July 2026: an ML-enhanced genetic algorithm predicts the ground-state Pbam structure of the synthesised La4Co4Pb phase, which database-reliant ML predictions had missed, and predicts La5CoPb2 from composition alone.
- 22 July 2026: bulk spin-Seebeck composites made by dynamic powder sputtering and low-temperature sintering of Pt-coated YIG powder show isotropic transverse thermoelectric signals at bulk scale, beyond the diffusion-limited thin-film geometry.
Open questions
- Can the ab-initio spin-lattice dynamics framework be run on a rare-earth transition-metal magnet composition, and at what computational cost per property (anisotropy, Curie temperature) rather than per validation case?
- Does any altermagnetic or PT-symmetric antiferromagnetic effect in this stream survive above room temperature, given that the interfacial signature in RuO2/WSe2 appears only below 55 K?
- Does the Pt-coated YIG powder sintering route generalise to soft magnetic components, and what are its loss, density and yield figures at part scale?
- Is the anomalous Hall angle result reproducible outside single-crystalline nanoflakes, in a deposited film or wafer-level process, at the reported 23.5 nT/Hz^0.5 detectability?
Assessment drafted 2026-08-31 from up to 18 KB sources using the technology-scorecard framework; scores are a draft read pending review.