In a conventional TPV system, most of the photons generated in the heated material are reflected back into the material when they reach its surface; it’s the same phenomenon that traps light in fiber-optic cables. When the solar cell and the heated material are brought close together, so that the gap between the two is shorter than the wavelength of the light being emitted, the surface no longer reflects light back. The photons travel from one material to the other as if there were no gap between them. The close spacing also allows electrons on one side of the gap to transfer energy to electrons on the other side. (A vacuum between the heated material and the solar cell maintains a temperature difference between the two that is required to achieve high efficiencies.) Since the heated material emits more photons, the solar cell can generate 10 times as much electricity for a given area, compared with a solar cell in a conventional TPV.

That makes it possible to use one-tenth as much solar-cell material, which cuts costs significantly. Alternatively, it makes it possible to generate more power at lower temperatures, which Peter Peumans, a professor of electrical engineering at Stanford University, says is one of the key advantages of the approach. Conventional thermal photovoltaics can require temperatures of 1,500 °C, he says. The first prototypes from MTPV work well at less than 1,000 °C, and DiMatteo says that, in theory, the technology could economically generate electricity at temperatures as low as 100 °C. This large temperature range could make the technology attractive for generating electricity from heat from a variety of sources, including automobile exhaust, that would otherwise be wasted.

But Peumans says that the technology has a trade-off: because the heated material and solar cell are placed so close together, it’s not possible to put a filter between them to help tune the wavelengths of light that reach the solar cell. This could limit the ultimate efficiencies that the system can reach.

DiMatteo first published work on the MTPV concept in the late 1990s, but it has taken until now to engineer prototypes large enough to be practical. One main challenge has been finding ways to create a gap that’s just one-tenth of a micrometer across and yet can be maintained over the relatively large areas needed for a practical device. DiMatteo says that the company will improve the performance of the devices by making the gap steadily smaller, which computer models suggest will improve efficiency.