Opus 5.5 agents discover two room-temperature magnetic semiconductor candidates

A team using Claude Opus 5.5 agents ran density-functional theory searches and designed or identified two Luttinger-compensated antiferromagnetic semiconductors that combine zero net magnetism with spin-sorted electronic states - properties desirable for spintronic memory because they provide large spin windows at band edges without stray magnetic fields. Simulations used a faster PBE+U level and the more accurate hybrid HSE06; key metrics are band gaps, spin-window widths (energy ranges at band edges where states share one spin), and magnetic ordering temperatures relative to thermal energy at room temperature (~26 meV). The core claim is materials that are insulating yet sort electrons by spin over windows far larger than thermal fluctuations, enabling readable, fast, low-interference spintronic bits.

The first candidate is a designed oxide YBaMnFeO5 predicted to be a semiconductor with a 2.35 eV gap, 1.0 eV hole window, 1.4 eV electron window, and simulated magnetic stability near 420 K (≈490 K after calibration), but it requires an ordered Mn/Fe checkerboard that likely disorders during high-temperature synthesis. The second is KVCr(CN)6, synthesized in 1999, predicted to have ~2.1 eV gap with 2.6 eV hole and 1.6 eV electron windows and experimentally stayed ordered up to ~376 K; its crystal chemistry locks metals into inequivalent sites making it more robust, though the only sample contained pore water and showed a small residual moment. All input files, outputs and analysis code are publicly available on GitHub, and simulations assume ideal, dry crystals with some disagreement between methods on solvent effects.