Could it one day be possible to generate electricity from the loss of heat from Earth to outer space? A group of Harvard engineers believe so and have theorized something of a reverse photovoltaic cell to do just this. The key is using the flow of energy away from our planet to generate voltage, rather than using incoming energy as in existing solar technologies.
Federico Capasso, a professor of applied physics, is leading a project that is counter-intuitive and challenges commonly-held physical conventions, yet puts to work findings from almost half a century ago.
"Sunlight has energy, so photovoltaics make sense; you’re just collecting the energy," says collaborator Steven J Byrnes, a postdoctoral fellow at SEAS. "But it’s not really that simple, and capturing energy from emitting infrared light is even less intuitive. It’s not obvious how much power you could generate this way, or whether it’s worthwhile to pursue, until you sit down and do the calculation."
The team has proposed two types of emissive energy harvesters: one that's similar to a solar thermal power generator, and another like a photovoltaic cell, except both would run in reverse to their more conventional counterparts.
The harvester utilizes two plates, one at the temperature of the Earth and another on top made of an ultra-emissive material designed to radiate heat skyward. A case study run in Lamont, Oklahoma found that the heat difference between the plates could generate only a few watts per square meter, day and night. Further difficulties were faced keeping the top cooler plate at a temperature below ambient temperature. The experiments did however show that the principle held in practice – thanks to the thermoelectric effect, differences in temperature do generate work.
"This approach is fairly intuitive because we are combining the familiar principles of heat engines and radiative cooling," says Byrnes.
The second device proposed attempts to generate voltage by harnessing temperatures at a much smaller scale. It relies on temperature differences between nanoscale electronic components, such as diodes and antennas. Using a theory shelved since the late 1960s, the team explored the use of temperature differences to direct electrical noise.
"We found they had been considered before for another application – in 1968 by J.B. Gunn, the inventor of the Gunn diode used in police radars – and been completely buried in the literature and forgotten," Capasso reveals. "But to try to explain them qualitatively took a lot of effort."
Gunn’s diagrams show that if a diode is at a higher temperature than a resistor, a current will be pushed towards the cooler component. Using this idea, Capasso has suggested that microscopic antenna pointed toward the sky could be used to cool areas of a circuit, which would create directional flow as proposed by Gunn.
The result, says Byrnes, is that "you get an electric current directly from the radiation process, without the intermediate step of cooling a macroscopic object."
According to the paper published in Proceedings of the National Academy of Sciences, a single flat device could be coated in tiny circuits, pointed skyward, and used to generate power. Recent advances in small-scale electronics, new materials like graphene, and nanofabrication add to the feasibility of this kind of technology and, while both ideas still have plenty of hurdles to overcome, the Harvard team is open in identifying the remaining challenges.
"People have been working on infrared diodes for at least 50 years without much progress, but recent advances such as nanofabrication are essential to making them better, more scalable, and more reproducible,” says Byrnes. He also identified that existing diodes are limited in their ability to work at low voltages. "The more power that’s flowing through a single circuit, the easier it is to get the components to do what you want. If you’re harvesting energy from infrared emissions, the voltage will be relatively low. That means it’s very difficult to create an infrared diode that will work well."
But with new types of diodes already being designed that can handle lower voltages, such as tunnel diodes and ballistic diodes, this challenge may soon be overcome. The team has suggested that the impedance of the circuit components could be increased to raise the voltage to a more practical level. Byrnes expects a little bit of each will probably be the answer.
Another hurdle is speed.
"Only a select class of diodes can switch on and off 30 trillion times a second, which is what we need for infrared signals," says Byrnes. "We need to deal with the speed requirements at the same time we deal with the voltage and impedance requirements. Now that we understand the constraints and specifications, we are in a good position to work on engineering a solution."