Scientific Motivation: Toward Functional Magnetic Semiconductors Beyond Conventional MIT Paradigms

The metal–insulator transition (MIT) stands as a cornerstone in condensed matter physics, frequently entangled with magnetic ordering. Canonical mechanisms — such as the Mott transition driven by Coulomb repulsion and the Slater transition arising from antiferromagnetic periodicity — typically lead to either ferromagnetic metallicity or antiferromagnetic insulating behavior. While these transitions embody magnetoelectric coupling, their device relevance is often restricted to a single degree of freedom: either spin or charge.

A longstanding challenge has been the realization of ferromagnetic semiconductors that combine large spontaneous magnetization with insulating or semiconducting transport. Such a phase would allow dual-channel control over magnetic and electric states, with transformative implications for low-dissipation logic gates, spintronic memory-in-computation schemes, and neuromorphic platforms.

Our study was driven by the hypothesis that spontaneous spin order, under tailored electronic conditions, could topologically renormalize the Fermi surface — enabling a Lifshitz-type MIT, as shown in Figure 1, distinct from Mott or Slater scenario.

Key Discovery: A Spin-Driven Lifshitz Transition in a 3d–5d Hybridized Quadruple Perovskite

We report the successful synthesis of a novel A- and B-site ordered quadruple perovskite oxide, CaCu₃Ni₂Os₂O₁₂ (CCNOO), via high-pressure and high-temperature techniques. The compound hosts a robust ferrimagnetic ground state, characterized by a Cu²⁺(↑)–Ni²⁺(↑)–Os⁶⁺(↓) spin configuration (shown in Figure 2) and displays a remarkably high Curie temperature of 393 K, maintaining a saturated magnetization of 2.15 µ_B/f.u. at 300 K.

Strikingly, we observe a concurrent metal-to-insulator transition (MIT) at the same temperature (shown in Figure 2). Synchrotron XRD and thermal analyses confirm the absence of structural phase transitions, while infrared spectroscopy reveals a gradually opening direct bandgap below the Curie point. The key insight: this insulating phase emerges purely from electronic reconstruction driven by spin ordering. First-principles calculations reveal that the ferrimagnetic order induces substantial band renormalization, resulting in Fermi surface reconstruction. Assisted by spin–orbit coupling (SOC) and moderate on-site Coulomb repulsion (U), a Lifshitz-type MIT transition is realized — distinct from traditional mechanisms and rarely observed in oxide materials.

Strategic Design: Exploiting Spin–Charge–Orbital Interplay in 3d–5d Systems

Central to our design is the 3d–5d orbital hybridization intrinsic to the AA′₃B₂B′₂O₁₂ quadruple perovskite framework. This crystal structure accommodates magnetic cations at the A′, B, and B′ sites, forming superexchange pathways that are both geometrically flexible and magnetically robust. In CCNOO, the Cu–O–Os angle (~112°) and Ni–O–Os angle (~138°) create strong covalent bridges supporting cooperative magnetism.

We hypothesized that by selecting magnetic ions with moderate moments and hybridization strength, the system could be tuned near the boundary of magnetic and electronic instability. This condition is conducive to spin-driven topological transitions in the electronic structure — specifically, the Lifshitz transition.

Mechanistic Insight: A Correlation-Assisted, Spin-Driven Lifshitz Transition

The MIT observed in CCNOO coincides precisely with the onset of ferrimagnetic order, yet its mechanism deviates from conventional Mott or Slater scenarios. Unlike Slater transitions, which rely on antiferromagnetic symmetry breaking and Brillouin zone folding, CCNOO exhibits a ferrimagnetic ground state without any structural distortion. While a finite Hubbard U is necessary to open a bandgap in theory, the required correlation strength is moderate, suggesting it is not a classic Mott insulator either. Instead, our results point to a spin-driven Lifshitz transition, where magnetic order reconstructs the band structure and eliminates Fermi surface pockets. This topological change, assisted by spin–orbit coupling and electron correlation, leads to the insulating state. We thus identify the transition as a correlation-assisted, spin-topology-driven MIT, beyond the conventional models.

Reflections: From a conceptual standpoint, this work illustrates how targeted crystal chemistry and orbital engineering can lead to emergent collective behaviors. The success of CCNOO underscores the power of interdisciplinary materials design, combining high-pressure synthesis, element-specific spectroscopy, and advanced electronic structure theory. As we look forward, we anticipate this framework will spark broader interest in spin–charge couplings, potentially leading to entirely new classes of functional quantum materials.

We would like to thank our collaborators, funding agencies, and the entire research team for their invaluable contributions to this work. This work is titled as “High-temperature ferrimagnetic order triggered metal-to-insulator transition in CaCu₃Ni₂Os₂O₁₂” and published in Nature Communications.