Kiwi postdoc powers Japanese collaboration for computational discovery of new solar cell materials

1 September, 2023

Materials scientists are a little too blessed with all the options they have for making climate-friendly solar cells cheaper and more efficient, so they're turning to Artificial intelligence (AI) to speed up the process. A New Zealand researcher is aiming to be part of that solution.

Solar cell research has come a long way in recent years as scientists race to develop new cost-effective and efficient solar cell materials. But the research has almost been too successful, because now that we know more about what makes a good solar cell, there are literally tens of thousands of new options for solar materials.

“It’s not often in science we have too many solutions,” says Dr Geoffrey Weal, one of only two New Zealanders selected by the Royal Society of New Zealand for a JSPS Postdoctoral Fellowship this year. “That’s the problem facing solar cell chemists around the world as they trawl through the thousands of potential candidates of new materials to find the best one for the next solar cell breakthrough. It’s just not possible for lab chemists to test each and every material.”

So researchers are turning to computing.

The JSPS Fellowship will enable Dr Weal to be part of the solution. He and his colleagues are using computing and AI to help scientists discover which of the tens of thousands of potential new materials will best increase solar cell efficiency.

Professor Daniel Packwood of iCeMS at Kyoto University (where Dr Weal will be based) says that the structure of the material determines how well the solar cell will work.

“So as materials scientists, we’re always on the hunt for new semiconductors that are more efficient. As chemists and physicists we’re working at the level of the molecular structure, experimenting with different combinations to see which absorb more light. Essentially, we’re using chemistry and physics to make the solar cell materials better, finding ways to tune the colour, or how well it conducts electricity, plus thinking about the best options for environmentally friendly manufacturing.”

He said that doing this in a traditional chemistry lab would be a formidable task.

Essentially, we’re using chemistry and physics to make the solar cell materials better, finding ways to tune the colour, or how well it conducts electricity, plus thinking about the best options for environmentally friendly manufacturing.
Professor Daniel Packwood Principal Investigator iCeMS at Kyoto University

“But on a computer, we can design and test hundreds or thousands of options for new solar materials very quickly.”

MacDiarmid Institute Co-Director Professor Justin Hodgkiss agrees.

“By using machine learning, Dr Weal and others are creating computational tools that will allow theorists and experimentalists to rationally design organic solar materials. This will save heaps of time and speed up the process of finding new solar cell materials.”

Professor Hodgkiss says Dr Weal has for the past two years been based in Kyoto University’s on-site integrated data materials lab at Te Herenga Waka Victoria University of Wellington.

“Dr Weal and his colleagues here in Wellington have shown a proof-of-concept of how these machine learning can be used for further OPV materials discovery.

Dr Weal says that that the Fellowship gives him the opportunity to collaborate with world-leading researchers at the Institute for Integrated Cell-Material Sciences (iCeMS) at Kyoto University.

“iCeMS is internationally recognised as a leader in fundamental materials research to advance a range of biological and materials technologies.

“During this Fellowship, I will work with Prof. Daniel Packwood and his group at iCeMS, while collaborating with Prof. Justin Hodgkiss and his group at Victoria University of Wellington (VUW). This collaboration will take advantage of our complementary strengths, allowing us to achieve our project goals, goals which could not be achieved by either the MacDiarmid Institute or iCeMS alone.”

Professor Packwood says Dr Weal’s research sits within the umbrella of an MOU between iCeMS and the MacDiarmid Institute.

“The establishment of the on-site lab in 2021 was the first step, and provided a gateway for collaboration between iCeMS and the Institute. Dr Weal’s Fellowship builds on this and allows for even closer research connectivity.

Professor Hodgkiss says Dr Weal was part of a visit by MacDiarmid Institute researchers to Kyoto University iCeMS in August that will create even more opportunities for collaboration.

“The wide range of research crossovers between members of the MacDiarmid Institute and iCeMS will ensure that this partnership will flourish over the next decade.”

The wide range of research crossovers between members of the MacDiarmid Institute and iCeMS will ensure that this partnership will flourish over the next decade.
Professor Justin Hodgkiss Co-Director The MacDiarmid Institute

JSPS Postdoctoral Fellowships

A maximum of four Japan Society for the Promotion of Science (JSPS) Postdoctoral Fellowships are awarded to New Zealand researchers through the Royal Society each year. The Fellowship programme is fully funded by Japan Society for the Promotion of Science (JSPS). JSPS Postdoctoral Fellowship supports young and excellent New Zealand postdoctoral researchers doing research in Japan for 12 – 24 months.

The Fellowship is not Dr Weal’s first encounter with success. His thesis was placed on the University of Otago Exceptional Doctoral Theses List in 2021 and he was awarded a University of Otago Doctoral Scholarship, University of Otago, 2017 – 2020. He also won first place in the student essay competition, New Zealand Institute of Chemistry 2021, a MESA International Travel Grant, 2019 and the People’s Choice Award Poster Competition at the New Zealand Institute of Chemistry in 2018, among many other awards.

Dr Geoffrey Weal

More background on solar chemistry

We’re all familiar with silicon solar panels. Solar panels need a semiconductor, such as the silicon we're used to, where electrons exist in a band of energies. When the material absorbs light, electrons are moved to another band of energy with a higher potential. The absorbed light therefore creates a voltage, and the electrons flow as a current around a circuit. The output power is just the product of voltage and current.

The exponential growth of installed solar photovoltaics (PV) will continue to be fuelled by new technologies that can have power conversion efficiencies of 30%, over 100x higher than plants, because they have no other responsibilities.

The exciting thing about the next generation PV cells are that they are being developed from cheap organic molecules, and the reason that they may fuel the next PV revolution is that they are lightweight, flexible, and can be printed, and therefore large quantities of solar cells can be churned out at low cost, and for entirely new applications. Their efficiencies are approaching current generation silicon solar cells.

Predicting exciton and charge diffusion in crystalline organic materials

The photophysics of photocurrent generation depends on the material structure. Dr Weal and the iCeMS/MacDiarmid Institute collaboration is focussing on using machine learning and photophysical intuition.

Recently, organic solar materials with high crystallinity have been shown to have exciton diffusion lengths that are comparable to bulk heterojunction systems. Dr Weal is developing tools for calculating photophysical properties for existing, novel, and yet-to-be-synthesised organic materials. The AI that Dr Weal is developing can be used to theoretically extract pairs of molecules within coupling distance and calculate exciton and charge transfer coupling values between molecules.

Dr Weal and his colleagues in Wellington have been creating the tools needed for high-throughput screening of organic photovoltaics (OPV) and have shown a proof-of-concept of how these tools can be used for further OPV materials discovery using machine learning.

Tags: geoffrey weal, justin hodgkiss, icems