Unlocking secrets of new solar material

Thursday, 24 April, 2014

Researchers at the US Energy Department’s National Renewable Energy Laboratory (NREL) are analysing a new solar material that has the same crystal structure as a mineral first found in the Ural Mountains in 1839. The material, perovskite, is extremely efficient.

NREL Research Fellow David Ginley said what makes perovskite structures so remarkable is that when processed in a liquid solution, they have unusual abilities to diffuse photons a long distance through the cell. That makes it far less likely that the electrons will recombine with their hole pairs and be lost to useful electricity - and indicates a potential for low-cost, high-efficiency devices. NREL Senior Scientist Daniel Friedman added that the light-absorbing perovskite cells have “a diffusion length 10 times longer than their absorption length”.

The solar cells are made from a relative of the perovskite mineral found in the Ural Mountains. Small changes to the material allow it to absorb sunlight very efficiently. The material is also easy to fabricate using liquids that could be printed on substrates like ink in a printing press, or made from simple evaporation. These properties suggest an easy, affordable route to solar cells.

By playing with the elemental composition, it is also possible to tune the perovskite material to access different parts of the sun’s spectrum. This means the material can be changed by deliberately introducing impurities, and in such a way that it can be used in multijunction solar cells that have ultrahigh efficiencies. Multijunction solar cells are an NREL invention from 1991 but currently have high material costs; cheaper multijunction cells based on perovskites could thus change this.

The band gap of perovskites can be adjusted by changing their compositions to access different parts of the sun's spectrum. Credit: Dennis Schroeder.

In four years, perovskite’s conversion efficiency - the yield at which the photons that hit the material are turned into electrons that can be used to generate electricity - has grown from 3.8% in 2009 to over 16%, with unconfirmed reports of even higher efficiencies arriving regularly. By contrast, efficiencies of single-crystal solar cells grew by less than 50% during their first five years of development, and most other types of solar cells showed similar modest improvements during their first few years.

NREL materials scientists are encouraged by the possibility of further optimising the materials. For example, replacing lead with tin in the cells could improve the efficiency of multijunction cells made from perovskite. Besides switching to a more environmentally friendly material, the change from lead to tin would also allow the finished solar cell to better withstand high humidity.

NREL Senior Scientist Joey Luther said perovskite shows promise to be a whole lot easier to make, compared to most other solar cells. “It doesn’t require high-temperature processing - you can just dip glass into two chemicals and get the material to form on it,” he said.

Luther predicts that researchers will approach perovskite from two different directions. One will be to make the best semiconductor possible without regard to cost, and the other will be to try to make it as cheap as possible. He noted that all the research groups currently working on perovskite are already getting very high efficiencies: “It’s not as if just one person knows the secret.”

Progress in efficiency for solar cells made from perovskite absorbers. Efficiency values were taken from publications and NREL's latest chart on record cell efficiencies.

The theoretical maximum efficiency of a perovskite-based solar cell is about 31%, and multijunction cells based on perovskites could attain higher efficiencies still. According to Luther, “The goal should be to try to get to 28% or higher.”

“This material is so easy to work with,” said NREL Senior Scientist Kai Zhu, who is preparing a precursor solution that converts from a liquid base to an absorber in a device.

“Working on solution processing, we can make a device in one or two days, from beginning to finish.”