Episode 7: RoboMapper reduces environmental impact of data generation

In this podcast episode, MRS Bulletin’s Laura Leay interviews Aram Amassian from North Carolina State University about his group’s achievements using RoboMapper, a materials acceleration platform. In researchers’ quest to run environmentally-conscious laboratories, Amassian offers a solution that focuses on characterization of materials. Having found that characterization generates a lot of energy, his group developed an automated approach to screening small samples in order to identify ones that warranted more in-depth study. By using their automated approach, the researchers found quantitative structure–property relationships for wide-bandgap perovskites. This work was published in a recent issue of Matter.

Transcript

LAURA LEAY: Welcome to MRS Bulletin’s Materials News Podcast, providing breakthrough news & interviews with researchers on the hot topics in materials research. My name is Laura Leay. Perovskites are a hot topic as they could be used to produce efficient solar cells. Metal halide perovskites are the emerging forerunner as they can be made with low cost and low energy intensity but challenges remain around their stability. There is a wide diversity of materials chemistry that could be probed to find more stable perovskites and doing this using traditional laboratory techniques would be incredibly time consuming. One research group has developed a way of automating this process that uses minimal chemical resources and produces tiny samples that save on analysis time and energy. RoboMapper uses precursor salts in various solutions to print an array of perovskites of different compositions. These printing robots fit on a bench top, taking up a space that’s about a meter square. Hundreds, or potentially even thousands, of samples can be printed in an array. Each sample can be as small as 50 μm in diameter and 600 nm thick, with an option to produce square patches and lines rather than spots. The small size is sufficient for the analytical instruments. Automated analysis stations can then perform microscopy, spectroscopy, diffraction, or electrical characterization using bespoke x–y stages. Professor Aram Amassian from North Carolina State University explains their approach.

ARAM AMASSIAN: If we were to think of ourselves as designing the laboratory not from the perspective of people, but from the perspective of the data we’re trying to generate about materials, how would we go about designing the experimental workflow? If we are going to automate, we no longer need to think about the sample manipulation by hand and so we no longer have to think about the sample size. The other inspiration was all the advances in materials characterization. You can do things down to the nanoscale and to the microscale, so you can collect meaningful data on samples that are very small.

LAURA LEAY: Duplicate arrays of samples can also be shipped to collaborators to perform detailed analysis. Collaborators at the National Synchrotron Light Source II, part of Brookhaven National Laboratory, have a custom-built setup to perform XRD and probe the effects of temperature. The work has also led to a cloud-based data library with collaborators at Iowa State University and UNC Chapel Hill. The aim of finding efficient, stable perovskites and other semiconductors is to help build energy supplies that have a minimum contribution to climate change. This means that sustainability is an important feature of the research. The team worked with an expert in lifecycle analysis to determine what factors of their research were least sustainable.

ARAM AMASSIAN: We did an in-depth analysis of the time, the operational steps, the energy required. We also worked with an environmental economist to do a lifecycle assessment—cradle to grave—evaluated the entire environmental impact of research associated to generating data for perovskites structural data. What we found is that, to our surprise, a lot of the energy generated—carbon footprint, environmental impact—was actually coming from characterization. We thought all of the impact was going to come from reducing the amount of material we’re wasting, the solvent, the single-use plastics. And it’s true that it makes a difference, right? But we were surprised to see how significant it was to reduce the bottlenecks associated to characterization. 

LAURA LEAY: Using smaller samples printed on an array decreased instrument time, which reduced electricity consumption, which means that the automated research is a more sustainable than doing all of the work by hand. Automation also meant that compositions could be screened, allowing the researchers to focus on the most promising candidates.

ARAM AMASSIAN: We used a framework that’s physics-based, that essentially did not require us to actually make the devices, did not require us to make long-term stability tests. All we had to do was just print the materials on a chip and do a test that required an hour in order to predict whether the material was going to be stable and was going to result in good solar cells. Now you have a truly high-throughput screening method that you could at least use to do fast-failing. Fast-failing means that you could at least fail, you know, 99% or 95% of the bad ones and then focus on the 1% or 5% to do more extensive investigations on the ones that seemed more promising.

LAURA LEAY: Using their automated approach, quantitative structure–property relationships for a wide-bandgap perovskite were found. In previous work, the team had looked at this type of perovskite for use in tandem solar cells, where silicon is still used but the perovskite forms an extra layer.

ARAM AMASSIAN: Tandem solar cells are the next revolution in solar cell technology. Think of it as taking solar technologies like silicon and adding an additional layer or sub-cell on them and enhancing their efficiency, reducing the cost in terms of dollars-per-Watt and accelerating de-carbonization of the grid. What our quantitative structure–property relationships revealed is, “What is the right composition at which you have the perfect cubic perovskite, that has the right bandgap and leads to the minimum amount of light-induced halide segregation?” So it was a demonstration of how you would go about doing a pre-screening of an alloy. You would take that and do further optimization. So it was sort of a proof-of-concept. And we translated that to a device and we showed that the concepts of this workflow were sound.

LAURA LEAY: As materials science moves forward, an automated approach could make a big difference in screening materials, developing data sets, and in making lab practice more sustainable.

ARAM AMASSIAN: The Royal Society of Chemistry came up with a report on sustainable laboratory practices. And that was an inspiration to us. That report showed that most scientists are interested in finding ways to make laboratory research more sustainable but they don’t necessarily know how. Hopefully this study show that there are ways of not just incrementally improving sustainability, but there are ways that you could improve sustainability by orders of magnitude, or at least an order of magnitude, which we have shown. And that sustainability improvement comes with very, very significant changes in productivity and efficiency in research.

LAURA LEAY: This work was published in a recent issue of Matter. My name is Laura Leay from the Materials Research Society. For more news, log onto the MRS Bulletin website at mrsbulletin.org and follow us on twitter, @MRSBulletin. Don’t miss the next episode of MRS Bulletin Materials News – subscribe now. Thank you for listening.