Discovering Alternative Solar Panel Materials with Supercomputing – HPCwire

Solar power is quickly growing in the world’s energy mix, but silicon – a crucial material in the construction of photovoltaic solar panels – remains expensive, hindering solar’s expansion and competitiveness with fossil fuels. Now, a team of researchers from the Georgia Institute of Technology and the Hanoi University of Science and Technology are researching ways to bypass silicon entirely – and to conduct the research, they used supercomputers at both the San Diego Supercomputer Center (SDSC) and the Texas Advanced Computing Center (TACC).

Specifically, the Georgia Tech researchers took a closer look at hybrid organic-inorganic perovskites, or “HOIPs,” which could serve as a viable alternative to silicon. Perovskite – a mineral – is naturally found in Russia’s Ural Mountains, but materials researchers have been looking for similarly structured compounds that may be easier to procure or produce. 

Some of the best candidates for HOIPs, unfortunately, are lead-based, leading them to be both more dangerous to humans and more unstable. So Georgia Tech worked with the Hanoi University of Science and Technology to use supercomputer-powered simulations to identify a set of lead-free HOIP candidates.

The simulations were run on SDSC’s Comet, a system with 1,944 Intel Haswell CPU nodes and 72 Nvidia GPU nodes delivering 2.76 peak petaflops, and TACC’s Stampede2, a system with 4,200 Intel Knights Landing CPU nodes and 1,736 Intel Xeon Skylake nodes delivering 18 peak petaflops. The systems were procured through XSEDE, the Extreme Science and Engineering Discovery Environment.

“This XSEDE-supported research relies on large-scale computations — a first step in our overall plan, which begins with showing simulations of this chemical space of HOIPs,” said Huan Tran, a professor of materials science and engineering at Georgia Tech and co-author of the research (which can be found here). “Next, we will use these simulations to collaborate with experimental experts who can synthesize and test the predicted HOIPs – no personal computer can handle this level of computations, hence the XSEDE supercomputers are a critically important aspect of our project.”

In total, thanks to the supercomputer-powered simulations, the researchers discovered a final set of four lead-free HOIPs: methylammonium tin iodide, formamidinium tin iodide, hydrazinium tin iodide and azetidinium tin iodide. Of the four, two have already been synthesized.

The newly identified HOIP candidates. Image courtesy of the authors.

“XSEDE offered us access to leading computational facilities, and this is a very important factor for enabling my research topics and accelerating my projects,” Tran said. “In the coming era of materials informatics, computational materials data is the most important infrastructure and I find Comet, Stampede2, and other XSEDE facilities provide the ideal platform for boosting up the development of these areas.”

To read the article by XSEDE’s Kim Bruch discussing this research, click here.

Source: https://www.hpcwire.com/2020/05/23/discovering-alternative-solar-panel-materials-with-supercomputing/

« »
Malcare WordPress Security