How do thermoelectric generators work, and where are they used?

Why a temperature difference matters

A thermoelectric generator turns a temperature difference directly into electricity. There are no turbines or pistons: the conversion happens inside solid materials. That makes the idea attractive anywhere heat is available and maintenance is hard, from deep space to the human body. For our overview of the technology, see thermoelectric generators.

How it works

In a peer-reviewed 2021 paper in iScience, Ryu, Chung and Park describe a thermoelectric device as one that converts heat into electricity by generating a voltage in a thermoelectric leg placed between a hot region and a cold region, through the Seebeck effect. They treat the device as a heat engine: its efficiency is the power delivered to an external load divided by the heat flowing in.

Because the voltage comes from the temperature difference, output depends on keeping that gap in place. If both sides drift towards the same temperature, the output falls.

Efficiency is usually discussed through a material figure of merit called zT. In the standard textbook formula that Ryu and colleagues set out, maximum efficiency is the temperature difference divided by the hot-side temperature (the limit for any heat engine working between those temperatures), multiplied by a factor that rises as zT rises. They note that the highest reported peak zT values were below 1 until 2000 and have recently reached about 3.

What has been demonstrated

Radioisotope generators are the best-documented use. NASA's overview of radioisotope power systems says a radioisotope thermoelectric generator converts heat from the natural radioactive decay of plutonium oxide fuel into electricity using devices called thermocouples, and has no moving parts. NASA says these generators power New Horizons and the Curiosity and Perseverance Mars rovers, and that its twin Voyager spacecraft, powered by radioisotope power systems, have operated for more than 47 years. It says radioisotope power systems have flown on more than two dozen NASA missions since the 1960s. Each general purpose heat source module, the building block of these generators, is nominally designed to produce 250 watts of heat at the start of a mission. The page gives no electrical output or conversion efficiency, so we do not state one.

Wearable devices are at laboratory stage. Jeong and colleagues (2022), writing in Advanced Science, built a wearable generator that runs on body heat, sunlight, or both. In their tests it produced 2.19 microwatts per square centimetre from body heat alone, at a temperature difference of 5.5 degrees Celsius, and 15.33 microwatts per square centimetre with both sources together, under one-sun illumination. For stable measurement, these figures were taken on a thermally insulating film held at 35 degrees Celsius rather than on skin. The authors report that the device kept 83.1 per cent of its room-temperature output at an ambient temperature of 35 degrees Celsius, and that output from body heat alone almost disappeared at that temperature.

What is only reported

These points are the authors' own judgements or results we have not been able to check independently.

• Jeong and colleagues describe their 15.33 microwatt result as the highest under "actual operating conditions" among wearable thermoelectric generators. That is the authors' judgement, not a verified ranking.
• Ryu and colleagues argue that zT alone does not determine efficiency, and that three parameters do: an averaged figure of merit plus two terms reflecting how material properties change with temperature. This is a theoretical and computational result; we have not seen it tested in a built device.

What is speculative

The following uses are proposals. The sources read for this article do not measure them.

• Clothing-mounted sensors that run on body heat alone, once output rises well above current laboratory levels.
• Modules on vehicle exhaust pipes that recover part of the waste heat to power on-board electronics.
• Battery-free monitors on pipes or machinery that sit across a steady temperature difference for long periods.

We have not verified a peer-reviewed performance figure for industrial or vehicle waste-heat recovery for this article, so its scale remains unknown here. Our tracker on electricity from waste heat follows progress on these uses.

Limits and open problems

Efficiency is a key limit. Even a perfect device cannot beat the heat-engine limit set by the temperature difference, and real devices fall short of it by an amount that depends on their materials. Output also depends on conditions: Jeong and colleagues' results changed with ambient temperature, and their power figures were measured on a film rather than on skin.

This article does not draw on sources about materials cost, toxicity or heat management, so it makes no claims about them.

What to watch next

Watch for independent measurements of wearable devices on live skin, for built devices that test the three-parameter efficiency theory, and for measured output from waste-heat modules rather than material tests alone. In our view, these are likely to be the points where the gap between laboratory figures and practical devices shows most clearly.