A New Class of Insulator Can Grow Fresh Electronic Bands on Demand

The electrons inside a Mott insulator are, in a sense, paralysed. Repelling each other so strongly that none can move without displacing a neighbour, they sit locked in place — insulating despite what band theory alone might predict. Crack one of these materials open with a beam of light, though, or nudge its magnetism with a field, and something unexpected happens. New electronic states blink into existence inside what should be an empty gap.

That discovery, formalised in work published in Physical Review B by Masanori Kohno at Japan’s National Institute for Materials Science, could change the way we think about engineering tomorrow’s electronic devices. Not by swapping one material for another. By reaching inside a material and reshaping it.

In conventional semiconductors — silicon, say — the energy bands that electrons can occupy are essentially fixed. You can shift how many electrons sit in those bands by applying a voltage, but the bands themselves stay put. That constraint is fundamental to how transistors, solar cells, and LEDs are designed. You pick a material for its band gap, then engineer around it. The idea that you might dial up a different band structure on the fly, using light or a magnetic field, barely registers as a possibility.

Mott insulators are different. So are Kondo insulators, a related family of materials where localised and mobile electrons tangle up in ways that produce exotic insulating behaviour. In both cases, electrons interact so strongly with each other that you cannot ignore those interactions. And it is precisely those interactions — those strong correlations — that open up a new possibility.

The key is something physicists call spin-charge separation. In an ordinary insulator, spin excitations (perturbations to the magnetic orientation of electrons) and charge excitations (perturbations to electron number) are inseparable; the energy required to create one is the same as the other, roughly equal to the band gap itself. In strongly correlated insulators, that link breaks. Spin excitations can exist at energies far below the charge gap. Spin and charge go their separate ways.

Kohno’s theoretical analysis shows what this separation enables. When you perturb a Mott or Kondo insulator — by doping it with a few extra electrons or holes, by applying a magnetic field large enough to partially magnetise it, or by illuminating it with photons that kick particle-hole pairs across the gap — the perturbation effectively changes the quantum-mechanical ground state from which electronic excitations begin. New electronic modes, with their own dispersion relations, materialise inside the band gap. They inherit their shape partly from the spin or charge excitation that spawned them; their energies are, in a sense, combinations of the conventional electronic dispersion and either the spin or charge dispersion of the unperturbed material.

That is a subtle but quite significant distinction from what happens in conventional band insulators, where no such modes appear. Dope silicon and you shift the Fermi level. Dope a Mott insulator at the right concentration and new bands emerge mid-gap, potentially reshaping whether the material behaves as an insulator, a metal, or something in between.

The work covers a broad range of perturbations and models. Doping — shifting the chemical potential to inject holes or electrons — has been studied before, but Kohno extends the framework to magnetisation-induced modes in spin-gapped Kondo and Mott insulators, and to states driven by spin or charge fluctuations more generally; including, crucially, the nonequilibrium states you get when light shines on the material and excites particle-hole pairs without introducing net charge at all. The dispersion relations of these emergent modes were verified across one- and two-dimensional Hubbard models, periodic Anderson models, and Kondo lattice models, suggesting the physics is robust rather than a quirk of any particular approximation.

There is an important proviso. These new modes only become significant — only develop enough spectral weight to matter for device applications — when a macroscopic number of spins or charges are excited simultaneously. A single photon won’t do it. You need, roughly speaking, a steady illumination that keeps enough particle-hole pairs excited at any given moment to sustain the effect.

Practically speaking, that is demanding but not unreasonable. Kohno describes band-structure engineering based on strong electron correlations as the potential outcome — a route to electronic and optoelectronic devices with tunable functionality that conventional semiconductor physics simply cannot offer. A solar cell, perhaps, whose effective band gap shifts under illumination. Or a switch that flips between insulating and metallic behaviour not by applying a large gate voltage but by tuning a magnetic field through the material’s spin gap.

Whether any of that reaches production is, of course, another matter entirely. The theory is clean; the materials science is considerably messier, and strongly correlated insulators are not the easiest things to work with experimentally. But the conceptual door is open. In conventional semiconductors, the band structure is given. In strongly correlated insulators, according to this work, it might, at least in principle, be chosen.

Study link: https://journals.aps.org/prb/abstract/10.1103/ythd-s2x8