In the frigid darkness beyond Jupiter, the Voyager spacecraft are still transmitting data back to Earth almost 50 years after launch. There are no solar panels out there — sunlight is 900 times weaker than at Earth. Instead, each probe carries a lump of plutonium-238 that quietly decays, and its heat is converted directly into electricity by a device with no moving parts, no fluids, and no combustion. That device is a thermoelectric generator, and it's essentially a thermocouple scaled up to power a spacecraft.

The physics is disarmingly simple. In 1821, Thomas Seebeck noticed that when two dissimilar metals were joined and one junction was heated, a compass needle nearby deflected. He'd stumbled onto the Seebeck effect: a temperature gradient across a conductor produces a voltage. Modern TEGs use semiconductors (typically bismuth telluride, lead telluride, or silicon-germanium) sandwiched between a hot side and a cold side. Electrons diffuse from hot to cold, and you tap the current off the ends.

What makes TEGs so appealing — and so frustrating — is the tradeoff:

• No moving parts. Voyager's RTG has run continuously since 1977. Try that with a diesel generator.
• Any heat source works. Radioactive decay, a wood stove, a car's exhaust manifold, even body heat.
• Terrible efficiency. Typically 5–8%. A gas turbine crushes that.

That last point is why TEGs haven't taken over the world. But efficiency stops mattering when the alternative is nothing. On the seafloor, in remote pipelines, on Mars rovers, or attached to gas flare stacks in Siberia, a device that reliably makes 100 watts from waste heat for 30 years is worth more than a 40%-efficient engine that needs maintenance you can't provide.

The article also surfaces a wonderful bit of history: in 1909, an inventor named William Coblentz tried to build a thermoelectric solar panel and failed — he realized heat alone didn't do it, only incident light produced power. He'd accidentally reinvented a photovoltaic effect while chasing thermoelectrics. The two technologies are cousins, and the boundary between them was blurry for decades.

There's a domestic angle too. Camping stoves like the BioLite use a small TEG wrapped around the burn chamber to power a fan and charge a phone. Some wood stoves ship with TEG modules that keep circulation fans running during a blackout — the stove powers itself off its own heat. And researchers are chasing "thermoelectric paint" and flexible TEGs that could harvest body heat to power medical sensors indefinitely.

The reason nobody has cracked highly efficient TEGs is deep physics: you need a material that conducts electricity well but conducts heat poorly, and most materials that are good at one are good at both. Metals with high electrical conductivity are usually excellent thermal conductors too. Beating this coupling is a decades-long materials-science quest, and every incremental gain in the "ZT" figure of merit gets published in Nature.