Australian researchers are challenging conventional solar energy collection with a novel approach: harvesting energy from heat radiating from the Earth itself. Scientists at the University of New South Wales (UNSW) are developing a “thermoradiative diode” – essentially a reverse solar panel – that converts infrared radiation into electricity, opening possibilities for 24-hour power generation and possibly revolutionizing energy sources for space exploration. While still in early stages, with current output significantly lower than traditional solar panels, the technology holds promise for powering remote devices and supplementing existing systems, particularly in environments where sunlight is limited.
Scientists in Sydney, Australia are exploring a novel approach to solar energy – generating electricity from heat radiating *from* the Earth, rather than relying on direct sunlight. This research could unlock new possibilities for powering devices even after the sun goes down, and potentially offer a unique energy source for space-based applications.
“We’re working on developing devices that generate electricity by emitting light, instead of absorbing it,” explains Jamie Harrison, a postgraduate student at the University of New South Wales (UNSW). “It’s like a reverse solar panel,” he adds.
Harrison is part of a research team at the School of Photovoltaic and Renewable Energy Engineering at UNSW, focused on innovative ways to harness solar energy – including capturing energy after sunset.
The Earth absorbs energy from the sun throughout the day and releases it at night as infrared radiation – a type of light invisible to the human eye, but felt as heat. Researchers at UNSW are developing a semiconductor called a thermoradiative diode that can convert this infrared radiation into electricity. This approach represents a significant departure from traditional solar power generation.
“If you were to look at the Earth at night with an infrared camera, you would see the Earth glowing,” says Professor Ned Ekins-Daukes, who leads the UNSW team. “What’s happening is that the Earth is radiating heat out into the cold universe,” he explains.
While scientists at UNSW weren’t the first to develop a thermoradiative diode, building on work from Harvard and Stanford Universities in the US, their team was the first to directly demonstrate electrical power generation from such a device in 2022.
Currently, the device generates a very small amount of electricity – approximately 100,000 times less than a conventional solar panel. Recent solar activity highlights the potential for both challenges and opportunities in harnessing solar energy.
“It’s enough to power a digital Casio wristwatch from the heat of your body,” says Ekins-Daukes, explaining that the amount of energy the diode can generate is determined by the temperature difference between the heat source and the surrounding environment.
Even operating at peak efficiency, Ekins-Daukes states that on Earth, the diode could generate electricity with a power density of only one watt per square meter.
This is because water vapor and gases like carbon dioxide in the atmosphere also absorb heat from the sun, reducing the temperature difference between the Earth’s surface and the night sky.
However, as Ekins-Daukes sees it, the true potential of this technology lies in space, where the absence of an atmosphere provides a much cooler environment for the diode to operate.
He anticipates the technology being used to power satellites. These systems are typically powered by solar panels, but Ekins-Daukes points out that this has limitations, particularly during periods when the satellite isn’t in direct sunlight.
“Particularly in low Earth orbit… You have 45 minutes of sunlight and then 45 minutes of darkness,” he says. “Obviously, your solar panel only works when the sun is shining. So, the opportunity here is… (to) use other surfaces of the spacecraft, not to fully power it, but to provide some auxiliary power,” he explains.
The diode would generate electricity from the heat absorbed by the satellite while in sunlight, radiating it into the “incredibly cold” space during periods of darkness, says Ekins-Daukes.
Currently, satellites are powered by batteries charged during sunlight hours, but Ekins-Daukes says the diodes present an “opportunity… to extract a little more energy from the satellite surface.”
“There’s a trend in space technology to create smaller satellites that fly in lower orbits, but maintain the same function as the larger ones,” he says. “It’s for this reason that the thermoradiative diode could be useful — it’s lightweight and generates power from unused surfaces.”
The team is planning a balloon flight test this year that will allow them to test the technology in space for the first time.
GALLERY – See Astronomical Discoveries of 2026
Dr. Geoffrey Landis, a scientist working with thermoradiative technologies at NASA’s John Glenn Research Center, says the technology could work for satellites in low Earth orbit, but would only be worthwhile if it could be made “at a very, very low cost.”
“Batteries are cheap,” he says. “You could think about using a thermoradiative diode, but it would probably be more expensive than just using batteries during those 45 minutes,” he adds.
Instead, Landis’s research focuses on using thermoradiative diodes for satellites on deep space missions to the outer planets of the solar system, or on rovers on permanently shadowed regions of the moon.
These missions are currently powered by specialized thermoelectric generators that convert heat – produced by the decay of a radioactive isotope, such as plutonium – into electricity.
“These things are heavy. They weigh about 45 kilograms, have about 200 liters of volume… They’re very expensive, and they’re reserved for major flagship missions because we need to produce plutonium – it’s difficult to produce, expensive to produce, and it’s a scarce resource,” says Dr. Stephen Polly, who works with Landis at NASA.
He says that while plutonium would still be needed to provide a heat source for thermoradiative diodes in deep space, compared to conventional thermoelectric generators, the diodes are much simpler and have fewer moving parts.
“Many smaller diodes would be connected together to create a panel that resembles the arrays of solar cells currently used to power satellites,” says Polly.
“The panel itself is what releases waste heat as light, so they can be much smaller, much more efficient, and a better use of that plutonium resource,” he says.
Thermoradiative diodes are currently made from the same semiconductor materials used in night vision goggles, but Landis says more work is needed to evaluate their viability when exposed to the high temperatures that decaying radioactive isotopes would produce.
Current space systems using these isotopes as heat sources operate at temperatures of around 540 or 1,000 degrees Celsius (1,004 and 1,832 degrees Fahrenheit).
“Nobody has ever thought about operating this type of semiconductor at higher temperatures, so we don’t know much about its durability. And, for a space mission, we want these semiconductors to last 10 years, 20 years, maybe even more,” he adds.
Landis and Polly are investigating new materials for manufacturing and testing a thermoradiative cell, which Polly says should allow the system to operate at temperatures of up to 375 degrees Celsius (707 degrees Fahrenheit).
He says that “if the research results continue to be promising,” then the use of an isotope-heated thermoradiative system “is certainly possible within the next five to ten years.”
At UNSW, Ekins-Daukes’s team has received funding from the U.S. Air Force to refine the diode, so it can operate more efficiently and generate larger amounts of power when used in low Earth orbit satellites, with solar radiation as the sole heat source.
His team is also considering using different materials, similar to those used to manufacture conventional solar cells, which, according to Ekins-Daukes, would allow them to “leverage” the manufacturing processes of solar cells, enabling production to be scaled up more quickly when the diode becomes commercially available – which he expects could happen within the next five years.
