New PV Cell Might Realize 70-80% Conversion Efficiency

Aug 1, 2014
Tetsuo Nozawa, Nikkei Electronics

A Japanese research group proposed a photovoltaic (PV) cell that does not use a PN junction and potentially realizes a conversion efficiency of 70-80%.

The group includes Syuuichi Emura, a researcher at the Institute of Scientific and Industrial Research, Osaka University. The cell was proposed at PVJapan2014.

The proposal was to use the polarity inside crystal (the gradient of the internal electric field caused by spontaneous polarization) for the separation of excitors (pairs of an electron and a hole). Though silicon (Si), which is a common material for PV cells, does not have a polarity, many compound crystals used as materials have strong polarities.

When photons are absorbed and excitors are generated due to the gradients of the internal electric fields of such materials, electrons and holes are spontaneously separated in different directions, Emura said. Specifically, he is considering a PV cell having a device structure in which a 300nm-350nm-thick InGaN layer (bandgap: 0.92eV) is sandwiched between an InN layer and electrodes.

In the case of commonly-used PV cells, PN junction plays the role of separating excitors and extracting electrons and holes at different electrodes. There are several advantages in separating excitors by using only the gradient of the internal electric field, but the largest advantage is that it prevents the recombination of electrons and holes as well as thermal relaxation to some extent.

For example, the thickness of the photoactive layers of general Si-based PV cells is several tens of micrometers or more to improve optical absorptance. As a result, short-wavelength, high-energy photons become so-called "hot excitors" away from the PN junction, and they are lost before reaching the PN junction and being separated into electrons and holes due to recombination and thermal relaxation.

(Continue to the next page)

The short-wavelength (shorter than the bandgap) light of conventional unijunction PV cells is lost due to the thermal relaxation while their long-wavelength light is transmitted and cannot be effectively used. This problem causes the performance limit of unijunction PV cells called "Shockley-Queisser limit."

On the other hand, the InGaN photoactive layer of the new PV cell is 300nm-350nm in thickness. Different from Si, InGaN is a direct transition type and has a high optical absorptance.

"With a thickness of about 100nm, it absorbs half of the irradiated light," Emura said.

It means that InGaN absorbs most of the light when its thickness is 300nm. Also, in consideration of the life of the carrier, the distance to the electrodes is short. As a result, phonon scattering hardly occurs, preventing thermal relaxation. In other words, it becomes possible to eliminate one of the two major loss factors that determine the Shockley-Queisser limit, Emura said.

The transmission loss of long-wavelength electromagnetic waves such as infrared light will still remain. But when the bandgap is lowered to 0.92eV by controlling the composition of In in InGaN, the energy ratio of the unusable infrared light can be lowered to 10% of the entire sunlight, he said.

Even with the 10% loss and other losses caused by, for example, reflected light, the overall loss will be 20-30%. In other words, in ideal conditions, it is theoretically possible to realize a PV cell whose conversion efficiency is 70-80%.

However, the PV cell is still in the theoretical stage.

"We have yet to make an actual cell and evaluate it," Emura said.