Featured Product
This Week in Quality Digest Live
Innovation Features
Anton Ovchinnikov
In competitive environments, operational innovation could well be the answer to inventory risk
Brian A. Weiss
Tech transfer can be more efficient
Zach Winn
Optical milk scanner uses materials-sensing technology to measure the health of cows
Tamara Sheldon
EV subsidies are poorly designed, but simple changes could make them more effective and equitable
Rebecca Beyer
Knowing who creates a robot makes it feel more authentic

More Features

Innovation News
MIT course focuses on the impact of increased longevity on systems and markets
Upgraded with blue laser technology
The Ring Dex 2 filling and capping system is designed to simplify production.
The QM certification is awarded for excellence in curriculum design and quality
Recent research finds organizations unprepared to manage more complex workforce
For scenarios that require a rapid response
With block-based editor, anyone can realize projects using AI-based image processing
Product placement lends depth to sustainability
Industrial Scan&Sand solution wins RBR50 Innovation Award

More News

Oak Ridge National Laboratory


Study Harnesses Made-to-Order Design to Pair Properties to Materials

Training based on data from 11,000 quantum chemistry calculations

Published: Monday, May 23, 2022 - 12:02

Astudy led by researchers at the U.S. Dept. of Energy’s Oak Ridge National Laboratory could help make materials design as easy as point-and-click.

The study, published in Nature Partner Journals’ Computational Materials, used an invertible neural network, a type of artificial intelligence that mimics the human brain, to select the most suitable materials for desired properties, such as flexibility or heat resistance, with high chemical accuracy. The team’s findings offer a potential blueprint for customizing scientific design and speeding up the journey from drawing board to production line.

“These results are a really nice first step to expanding capabilities for materials design,” says Victor Fung, a Eugene Wigner fellow at ORNL’s Center for Nanophase Materials Sciences and lead author of the study. “Instead of taking a material and predicting its given properties, we wanted to choose the ideal properties for our purpose and work backward to design for those properties quickly and efficiently with a high degree of confidence. That’s known as inverse design. It’s often discussed, but very few concrete demonstrations of inverse design achieve this kind of high accuracy.”

Neural networks rely on millions of digital neurons and synapses, similar to those in the brain. The neurons can operate independently and don’t necessarily perform calculations in traditional ways. The neurons of an invertible neural network operate in one-to-one pairs like tag teams, a process known as “bijective function approximation.”

The inverse-design approach used in the study leverages advances in the invertible neural architecture to enable forward mapping, or adding up input to produce a result, and backward mapping, or starting with a result and working back to deduce the initial input.

“Traditional algorithms can be very computationally intensive, and they can’t guarantee the best design,” says Jiaxin Zhang, an AI scientist at ORNL and co-author of the study. “We have such a range of possible materials that the search space is huge. But we can generate a smaller batch of samples from existing data and use that to represent the whole search space. We can use those samples to train the neural network, give it a specific desired result, and let it explore all possible candidates. The model learns and delivers more accurate results as it goes.”

The team trained the neural network on data from a total of 11,000 quantum chemistry calculations performed on ORNL’s Compute and Data Environment for Science and the National Energy Research Scientific Computing Center (NERSCC) at Lawrence Berkeley National Laboratory. The team leveraged the neural network to determine the necessary strain on molybdenum disulfide, a 2D material, to result in a specified bandgap, or range of energy, that prevents electrical conductivity. The model was able to specify nearly the exact applied strain needed to tune the bandgap of the material, which allowed further investigation into its metal insulator transition—the change from high electrical conductivity to low electrical conductivity.

The study could lead to more customizable materials design for industry. Credit: Jiaxin Zhang/ORNL, U.S. Department of Energy.

“We reached near chemical accuracy,” Fung says. “In terms of this particular application, it’s definitely a first for this kind of inverse design. This is a general approach with a broad range of applications that’s easy to train and easily scalable, and we’re quite confident it’s the best model for this type of problem currently.”

The team has made the neural network code publicly available and hopes to run additional studies using Summit, ORNL’s 200-petaflop supercomputing system, or Frontier, the forthcoming exascale supercomputing system, to further refine the approach. Summit and Frontier are part of the Oak Ridge Leadership Computing Facility, a DOE Office of Science user facility.

“The more computing resources available, the more samples we can use to train the model,” Zhang says. “We’re excited to leverage this experience to explore other materials and designs.”

This work was performed at the Center for Nanophase Materials Sciences, a DOE Office of Science user facility, and used resources of the NERSCC. Support came from the Center for Understanding and Control of Acid Gas-Induced Evolution of Materials for Energy, an Energy Frontier Research Center funded by the DOE Office of Science, as well as from the Office of Advanced Scientific Computing Research and the ORNL Artificial Intelligence Initiative.

The University of Tennessee-Battelle manages ORNL for DOE’s Office of Science, the single largest supporter of basic research in the physical sciences in the United States. DOE’s Office of Science is working to address some of the most pressing challenges of our time. For more information, visit https://energy.gov/science.

First published April 21, 2022, on Oak Ridge National Laboratory News.


About The Author

Oak Ridge National Laboratory’s picture

Oak Ridge National Laboratory

Oak Ridge National Laboratory is a multiprogram science and technology laboratory managed for the U.S. Department of Energy by UT-Battelle, LLC. Scientists and engineers at ORNL conduct basic and applied research and development to create scientific knowledge and technological solutions that strengthen the nation's leadership in key areas of science; increase the availability of clean, abundant energy; restore and protect the environment; and contribute to national security.