真乄科技業的頂尖投資團隊

The researchers

Right-click to download interview with Aditya Mohite and Wanyi Nie(5.24 MB MP3)

"In just two years reported perovskite solar cell efficiencies have reached 20% – other nanomaterials have taken two decades to get to 9%," says Aditya Mohite, a scientist at Los Alamos National Laboratory in New Mexico. "Los Alamos has a reputation for crystal growth so we thought we could come at the field from a different perspective."

Mohite adds that the approach they developed for improved perovskite crystal quality has potential for producing crystals with improved optoelectronic properties for a number of other applications as well. "Typically for all the nanomaterial technologies, the optical properties are good but they tend to have multiple interfaces and this leads to a lot of recombination. So if there is a way we can make better quality crystal it would be good."

Bigger better grain size

Mohite and his postdoc Wanyi Nie – first author on the report of these results – worked together with colleagues at Los Alamos, Purdue University and Rutgers University. They began by trying the usual approach for obtaining high-quality crystal thin films of the perovskite-based PbCH₃NH₃I_3–xCl_x. Traditionally a solution is spin-coated to produce a thin film at low temperature and then placed on a hot plate for "post annealing" to improve the crystal quality.

As luck would have it the material used for these initial experiments was poor. "I was trying the post annealing process and never getting a good efficiency," says Nie. "So I thought maybe the post annealing is too late – maybe I should do the annealing before the film is dry, while the reaction can still occur and the crystal is forming." The result was crystals with millimetre grain sizes and exceptional optoelectronic properties.

Hot-cast perovskite thin films

Assessing optoelectronic improvements

One of the indicators of crystal quality is the hysteresis observed when the voltage sweep is reversed. Defects and grain boundaries lead to trap states that restrict the movement of charge carriers. "So as a result when you sweep your voltage from reverse to forward and from forward to reverse the curves do not lie on top of each other," says Mohite. In contrast the curves for the hot-cast perovskite show no hysteresis and the efficiency was reproducible over 50 devices.

The team also measured the open voltage circuit behaviour as a function of light intensity and identified a relationship between the two that suggests bimolecular recombination processes dominate. This is similar to the behaviour observed in high-quality GaAs devices but these require far more stringent production conditions that are not readily scaled up for industrial production. "It’s like a text-book semiconductor," Mohite adds.

Hot-casting scheme for large-area crystal growth

Next steps

The team are enthusiastic about attempting to push the efficiency further still. "This was a proof of concept device so the architecture was not optimized," says Mohite. "But we know what to do to go about hitting the theoretical efficiency defined by the ‘Shockley–Queisser limit’, which is the ‘Holy Grail’." For the current material this would be 27%.

Nie adds that there is still important work to do to improve the stability. Humidity exposure and oxidation can be avoided by encapsulation but photodegradation presents a more fundamental challenge. "In our recent study we found that if we put the device in high vacuum we still see light-induced degradation. So we are trying to understand the mechanism to find out how to eliminate this by adjusting the working conditions and changing the chemistry to reduce the charging effect."

Full details of the results are reported in Science 347 522-525.