Machine learning automates identification of crystal structures in new materials

January 30, 2020

Providing a method for eliminating some of the guesswork from crystal structure determination, a machine learning-based approach to determining crystal symmetry and structure from unknown samples may greatly improve the speed and accuracy of this process. The new method brings crystallography into the high-throughput world of artificial intelligence (AI). From geology to biology to materials science, identifying crystal structure is critical to understanding its general characteristics and properties. Electron backscatter diffraction (EBSD) is the standard technique for identifying crystal structure. However, while powerful, EBSD requires user input concerning critical elements of structure, such as crystal phase, which can be both time-consuming and prone to error. A fully autonomous approach to more hands-on crystallography would open the door to a high-throughput evaluation of a material's properties, according to the authors. Kevin Kaufmann and colleagues developed an autonomous, machine learning-based method for rapidly determining crystal structure from EBSD data with high accuracy. The authors used a convoluted neural network to identify unique crystal symmetries in EBSD pattern images using the same symmetry features a crystallographer would use. According to the results, the trained algorithm was capable of accurately identifying and classifying various aspects of crystal structure from diffraction patterns of materials that it was not trained on, and with almost no human input. The platform opens the door to high-throughput determination of structures for multiple fields.
-end-


American Association for the Advancement of Science

Related Crystallography Articles from Brightsurf:

Making disorder for an ideal battery
The lithium batteries that power our electronic devices and electric vehicles have a number of drawbacks.

New techniques probe vital and elusive proteins
Researchers at the Biodesign Center for Applied Structural Discovery and ASU's School of Molecular Sciences, along with their colleagues, investigate a critically important class of proteins, which adorn the outer membranes of cells.

Mining molecular data with cryo-EM unveils hidden biological secrets
In the new study, Abhishek Singharoy and his colleagues demonstrate that cryo-EM can be pushed to even greater extremes of clarity, by extracting precious information previously buried in the reams of cryo-EM data.

New substance library to accelerate the search for active compounds
The MX team at HZB and a Group at the University of Marburg have established a new substance library.

Elucidating how asymmetry confers chemical properties
New research by Carnegie's Olivier Gagné and collaborator Frank Hawthorne of the University of Manitoba categorizes the causes of structural asymmetry, some surprising, which underpin useful properties of crystals, including ferroelectricity, photoluminescence, and photovoltaic ability.

It takes a neutron beam to find a proton
Researchers from Osaka University have successfully determined the structure of a bacterial copper amine oxidase using neutron crystallography.

Researchers map protein motion
Cornell structural biologists took a new approach to using a classic method of X-ray analysis to capture something the conventional method had never accounted for: the collective motion of proteins.

Machine learning automates identification of crystal structures in new materials
Providing a method for eliminating some of the guesswork from crystal structure determination, a machine learning-based approach to determining crystal symmetry and structure from unknown samples may greatly improve the speed and accuracy of this process.

Early immune response may improve cancer immunotherapies
Researchers report a new mechanism for detecting foreign material during early immune responses.

Study reveals limitations of method for determining protein structure
A new study by chemists at the University of Arkansas shows that X-ray crystallography, the standard method for determining the structure of proteins, can provide inaccurate information about membrane proteins, which in turn could lead to poor and inefficient drug design.

Read More: Crystallography News and Crystallography Current Events
Brightsurf.com is a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com.