Tinkering with Solar Cells To Improve New Material Designs

When your research revolves around material properties, you’ve always got to be on the lookout for interesting behaviour. Jonathan Barnsley knows this well. As a Chemistry PhD candidate at the University of Otago, he studies donor-acceptor (D-A) materials. With help from NeSI supercomputers, he’s uncovering some interesting findings on a molecular level that could have big impacts across a number of industries.

“It is very exciting to be involved because you’re constantly learning,” Barnsley says. “It just takes somebody to notice something that you haven’t noticed in the past and that’s really where things come from. You never know what’s around the corner.”

D-A materials have an ability to exchange charge from an electron rich donor to an electron acceptor, forming a formally positively charged donor and negatively charged acceptor. Barnsley and his supervisor, University of Otago Chemistry Professor Keith Gordon, are exploring ways to apply the most effective qualities of D-A interactions into the design of new and novel materials.

Some of the materials they work with are D-A dyes, which can be used in dye-sensitized solar cells. Improving the efficiency of solar cells has obvious applications in the green energy sector, but findings from their work could also be applied to other technologies, like sensors, and influence the design of materials in other sectors, such as photodynamic therapy or nonlinear optics. Their latest NeSI-supported work has been published in two papers recently, one of which is featured on the cover of ChemPhotoChem.

When investigating D-A structure-property relationships, one of their first steps is to use techniques under the umbrella of spectroscopy. In simple terms, they shine a laser, or bright light source, at a sample of the material and depending on how the light reacts, either passing through or being scattered, the behaviour provides information about the material’s properties. Those experimental results are then compared with data gathered from computational models run on NeSI computing resources and smaller, specialized clusters at the University of Otago.

“That gives us an enriched dataset to really understand how these systems are behaving when they interact with light,” Barnsley says. “Leaps of understanding have been achieved, in part due to some of the computational approaches. We find it is very useful to have that modelling approach to understand our systems.”

This mix of experimental and theoretical approaches is especially useful when experimental data isn’t always obtainable. For example, they can use vibrational spectroscopy to probe a material’s structure and then combine that with density functional theory to model the material’s electronic and geometric parameters.

“There are some things that computational chemistry can probe that we simply can’t observe in the lab, at least at this point in time,” Barnsley says. “It would be difficult for any scientist to remain on the frontiers of dye design and characterisation without computational approaches.”

The predictive capabilities of computational methods also open the door for better observation and understanding of properties like structure, absorption profiles and transition nature — key characteristics that can impact the performance of a material’s design.

“If you’re a synthetic chemist, you can spend a long time making a material and find out eventually it doesn’t have the properties you were hoping for, whereas computational chemistry is becoming better and better at predicting properties ahead of time,” Barnsley says. “That’s a tantalizing thing that may come into play for us in the dye design field.”

Barnsley and Gordon work with a number of local collaborators, including fellow University of Otago researchers Associate Professor James Crowley and Dr Nigel Lucas, as well as some across the Tasman, partnering with the University of Wollongong’s Professor David Officer and Dr Attila Mozer in Australia. Their collaborations even extend overseas, with some projects involving research groups led by Professor Johan Bobacka at Åbo Akademi University in Finland and Dr Grzegorz Lisak at Nanyang Technological University in Singapore.

Currently, a primary source of D-A materials comes from Officer and Mozer, whose Wollongong labs are experimenting with new designs for solar cell systems. Samples are sent to Barnsley and Gordon, who tinker with the systems to identify their properties and explore ways to improve their design.

“We are really lucky to work with people from these groups who are doing all the hard yards making the materials. I basically get to ‘play’ with the materials once somebody’s made them,” Barnsley says. “My work tries to understand how these things interact with light and how we can perhaps make them better at absorbing light or try different combinations to absorb light at different parts of visible region.”

Meanwhile, he also has colleagues developing supramolecular architectures or other complex computational models that he needs to factor into his studies.

“Synthetic chemists are pushing the envelope making wonderfully large, elegant molecules that have all sorts of interesting properties,” he says. “We have to keep up with that and be able to run the modelling and simulations of those big molecules.”

Case in point, they recently worked on their largest simulation yet, a model with 286 atoms and 1,283 electrons. The analysis demanded significant computing power, much more than their group’s computing resources were providing.

“We couldn’t do it on our cluster at all. The jobs were taking six months to complete. Factoring in the number of calculations you had to do to analyse these correctly — it was just too long,” Barnsley says. “So NeSI helped a lot in that situation, where we had these novel, complicated and large architectures.”

While the models Barnsley works with may be getting increasingly complex, the integration with NeSI resources has been fairly simple.

“Some of the proof of how good the experience was with NeSI is that we didn’t really need a lot of help. It was set up in a fairly straightforward manner and it made it really simple to come from our small scale system up to the high performance realm,” Barnsley says. “Help from people like Ben Roberts, Gene Soudlenkov, and Peter Maxwell was outstanding. People who have a deep understanding of NeSI’s systems and computational modelling are an incredible resource and we are glad to have access to the resources that NeSI provides.”

With new project ideas already on the horizon, Barnsley and Gordon’s use of NeSI won’t be slowing down anytime soon.

“The Keith Gordon group will continue to analyse novel and exciting materials, so for us there will always been a need for quality high performance computing,” Barnsley says. “I’m looking forward to a lot of new possibilities and applications of the work we’re doing. It’s as exciting as it’s ever been, so it’s good to be a part of it.”

- - - - - - - - - - - - - - - - - -

Image Credit - J. E. Barnsley et al.: Flicking the Switch on Donor–Acceptor Interactions in Hexaazatrinaphthalene Dyes: A Spectroscopic and Computational Study. ChemPhotoChem. 2017. Volume 1, Issue 10. Page 424. Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.

Recent publications by Jonathan Barnsley and his colleagues, supported by NeSI systems:

• Mitchell, R., Wagner, K., Barnsley, J. E., van der Salm, H., Gordon, K. C., Officer, D. L. and Wagner, P. (2017), Synthesis and Light-Harvesting Potential of Cyanovinyl β-Substituted Porphyrins and Dyads. Eur. J. Org. Chem., 2017: 5750–5762. doi: 10.1002/ejoc.201701003
• J. E. Barnsley, B. A. Lomax, J. R. W. McLay, C. B. Larsen, N. T. Lucas, K. C. Gordon, ChemPhotoChem2017, 1, 432. doi: 10.1002/cptc.201700092
• Synøve Ø. Scottwell, Jonathan E. Barnsley, C. John. McAdam, Keith C. Gordon, and James D. Crowley. “A Ferrocene Based Switchable Molecular Folding Ruler” Chemical Communications (2017) accepted for publication https://dx.doi.org/10.1039/C7CC03358C