SOPHIA CHEN: Welcome to MRS Bulletin’s Materials News Podcast, providing breakthrough news & interviews with researchers on hot topics in materials research. My name is Sophia Chen.
A big part of materials research is, well, discovering new materials. Historically, this process has involved a lot of trial and error, where a human researcher makes an experimental material, runs a whole bunch of tests on it to characterize it, and then iterates upon the recipe to try to make it fit the desired function better.
But this process takes a long time. With the rise of artificial intelligence tools, materials researchers are trying to speed up the discovery process by automating it. The approach has been toward developing an autonomous, self-driving laboratory that can create and characterize materials in a high-throughput way, under the guidance of a researcher.
Vivek Chawla, who currently works as a postdoctoral research associate at Sandia National Laboratories, is working on automating machines that would be necessary for a self-driving lab.
VIVEK CHAWLA: We need multiple instruments that all coordinate with each other in an autonomous fashion, and then we need some form of tools that can make intelligent decisions and navigate the whole process.
SOPHIA CHEN: To that end, Chawla and his colleagues have automated a machine that performs a process known as nanoindentation. Chawla did this work as a postdoc at the University of Tennessee at Knoxville.
In nanoindentation, you’re prodding a material with a sharp tip in a very localized area to test the mechanical properties of a material.
VIVEK CHAWLA: Let’s say you take a finger, you probe this finger on the sand, and now you probe this finger on a cement. You see the difference in the material response when you’re probing. So basically, that's what we are doing, but at a nanoscale with very high resolution force transducers.
SOPHIA CHEN: They automated the machine that does the nanoindentation, also known as the nanoindenter. They were able to automate the machine to operate in a few different modes. In one mode, the nanoindenter could move autonomously to a location that the user specifies.
They could also automate the nanoindenter to work with images of the material. The software identifies features of interest on the material sample, and the nanoindenter can automatically move toward those features.
They demonstrated the automation on a material made of carbon fibers.
VIVEK CHAWLA: The code is given an image, and the code decides what is the ellipticity of the fiber, identifies the pixel locations of interesting features, different orientations, tells the indentor, "Hey, go and do indentation there. Indenter sends back the data, and based on that, it makes the next judgment on where it should go and do the indentation next.
SOPHIA CHEN: Most notably, they did the automation without access to the instrument’s application programming interface, or API. The API is the software interface that allows the computer to directly talk to the indenter.
From a machine safety perspective, it makes sense for the user not to have access to the API.
VIVEK CHAWLA: They have a very valid point in not providing these because in our quest to make an automated indenter, we can very easily break an indenter.
SOPHIA CHEN: Instead, they implemented the automation at the user interface level, essentially telling the computer controlling the nanoindenter where to click.
VIVEK CHAWLA: This work also shows how to build robust autonomous instruments, even if you don't have these APIs available.
SOPHIA CHEN: Without direct access to the manufacturer’s API, they built what Chawla calls a “poor man’s API.”
Chawla says this work is a starting point to show that an autonomous nanoindenter is feasible, and it could convince manufacturers to build future instrument software with native automation support.
VIVEK CHAWLA: How it works is usually you need to show applications, you need to show why it is important, and then slowly things start becoming available. […] Ideally, you want access to API.
SOPHIA CHEN: This work was published in a recent issue of Review of Scientific Instruments. My name is Sophia Chen from the Materials Research Society. For more news, log onto the MRS Bulletin website at mrsbulletin.org and follow us on X, @MRSBulletin. Don’t miss the next episode of MRS Bulletin Materials News – subscribe now. Thank you for listening.
