A nanophotonic solar thermophotovoltaic device as viewed from the perspective
of the incoming sunlight. Reflective mirrors boost the intensity of the light reaching the
carbon nanotube absorber array (center), enabling the device to reach high temperatures and
record-setting efficiencies. Credit: FELICE FRANKEL
A new approach to harvesting solar energy, developed by MIT researchers, could improve efficiency by using sunlight to heat a high-temperature material whose infrared radiation would then be collected by a conventional photovoltaic cell. This technique could also make it easier to store the energy for later use, the researchers say.
A conventional silicon-based solar cell "doesn't take advantage of all the photons," Wang explains. That's because converting the energy of a photon into electricity requires that the photon's energy level match that of a characteristic of the photovoltaic (PV) material called a bandgap. Silicon's bandgap responds to many wavelengths of light, but misses many others.
To address that limitation, the team inserted a two-layer absorber-emitter device—made of novel materials including carbon nanotubes and photonic crystals—between the sunlight and the PV cell. This intermediate material collects energy from a broad spectrum of sunlight, heating up in the process. When it heats up, as with a piece of iron that glows red hot, it emits light of a particular wavelength, which in this case is tuned to match the bandgap of the PV cell mounted nearby.
This basic concept has been explored for several years, since in theory such solar thermophotovoltaic (STPV) systems could provide a way to circumvent a theoretical limit on the energy-conversion efficiency of semiconductor-based photovoltaic devices.
That limit, called the Shockley-Queisser limit, imposes a cap of 33.7 percent on such efficiency, but Wang says that with TPV systems, "the efficiency would be significantly higher—it could ideally be over 80 percent."
There have been many practical obstacles to realizing that potential; previous experiments have been unable to produce a STPV device with efficiency of greater than 1 percent. But Lenert, Wang, and their team have already produced an initial test device with a measured efficiency of 3.2 percent, and they say with further work they expect to be able to reach 20 percent efficiency—enough, they say, for a commercially viable product.
