A Smart Material That Routes Heat Like a Computer Chip

They made heat behave like information. In a lab at Osaka Metropolitan University, researchers coaxed thermal radiation into obeying a new set of rules — ones that let engineers steer and store heat the way we route and hold data in electronics.

The trick relies on two very different ingredients: a magneto-optical film that changes how it interacts with light under a magnetic field, and a phase-change alloy known as GST that can flip between states and remember that flip without power. Combined, they form a surface whose thermal emission depends on direction and whose behavior can be switched and retained.

Why does that matter? Because heat normally follows reciprocity: if a surface absorbs infrared radiation from one direction, it will tend to emit the same way. That mirrors a stubborn symmetry in physics that has long limited thermal control. Break that symmetry, and you can make a surface absorb heat from one direction while preferentially emitting it in another. You can, in other words, route thermal energy.

Tests of the device showed something rare. It responded differently to incoming infrared light even when that light hit nearly straight on — a regime where older nonreciprocal designs failed or became inefficient. The result is more practical control over thermal flow at angles where real-world devices operate.

There’s also a pragmatic win: the GST layer preserves whatever state it’s switched into, without needing a continuous power supply. Flip the material into an absorbing state. Leave the power off. The state persists. That memory-like property makes the structure programmable in a literal sense.

Heat can be programmed and stored like data.

Imagine infrared sensors that selectively listen to certain directions, or energy harvesters that funnel waste heat to where it can be put to use. Think of photonic memory that uses light and thermal emission as the information carrier rather than charges in a transistor. The team behind the work — led by Professor Koichi Okamoto and Dr. Shunsuke Murai — points to these possibilities and more.

The approach does have challenges ahead: integrating such layers into compact, manufacturable devices, scaling control of magnetic fields, and pushing performance across broader wavelengths and temperatures. Still, the demonstration marks a conceptual leap. It reframes heat from a passive byproduct to an active signal that can be steered, switched, and stored.

That thought — that warmth might one day be routed like packets across a network of tiny surfaces — changes how you imagine thermal design. It also opens a toolbox for engineers who want to bend infrared light to human purposes, not the other way around.