Machine learning enables predictive modeling of 2-D materials
3
Machine learning enables predictive modeling of 2-D materials 7 December 2016, by Joan Koka The badlands of stanene: stanene is softer and consequently much more rippled than its cousins graphene and silicene. Credit: Mathew Cherukara, Badri Narayanan and Subramanian Sankaranarayanan/Argonne National Laboratory Machine learning, a field focused on training computers to recognize patterns in data and make new predictions, is helping doctors more accurately diagnose diseases and stock analysts forecast the rise and fall of financial markets. And now materials scientists have pioneered another important application for machine learning—helping to accelerate the discovery and development of new materials. Researchers at the Center for Nanoscale Materials and the Advanced Photon Source, both U.S. Department of Energy (DOE) Office of Science User Facilities at DOE's Argonne National Laboratory, announced the use of machine learning tools to accurately predict the physical, chemical and mechanical properties of nanomaterials. In a study published in The Journal of Physical Chemistry Letters, a team of researchers led by Argonne computational scientist Subramanian Sankaranarayanan described their use of machine learning tools to create the first atomic-level model that accurately predicts the thermal properties of stanene, a two-dimensional (2-D) material made up of a one-atom-thick sheet of tin. The study reveals for the first time an approach to materials modeling that applies machine learning and is more accurate at predicting material properties compared to past models. "Predictive modeling is particularly important for newly discovered materials, to learn what they're good for, how they respond to different stimuli and also how to effectively grow the material for commercial applications—all before you invest in costly manufacturing," said Argonne postdoctoral researcher Mathew Cherukara, one of the lead authors of the study. Traditionally, atomic-scale materials models have taken years to develop, and researchers have had to rely largely on their own intuition to identify the parameters on which a model would be built. But by using a machine learning approach, Cherukara and fellow researchers were able to reduce the need for human input while shortening the time to craft an accurate model down to a few months. "We input data obtained from experimental or expensive theory-based calculations, and then ask the machine, 'Can you give me a model that describes all of these properties?'" said Badri Narayanan, an Argonne postdoctoral researcher and another lead author of the study. "We can also ask questions like, 'Can we optimize the structure, induce defects or tailor the material to get specific desired properties?'" 1 / 3
Transcript of Machine learning enables predictive modeling of 2-D materials
Machine learning enables predictivemodeling of 2-D materials7 December 2016, by Joan Koka
The badlands of stanene: stanene is softer andconsequently much more rippled than its cousinsgraphene and silicene. Credit: Mathew Cherukara, BadriNarayanan and SubramanianSankaranarayanan/Argonne National Laboratory
Machine learning, a field focused on trainingcomputers to recognize patterns in data and makenew predictions, is helping doctors more accuratelydiagnose diseases and stock analysts forecast therise and fall of financial markets. And nowmaterials scientists have pioneered anotherimportant application for machine learning—helpingto accelerate the discovery and development ofnew materials.
Researchers at the Center for Nanoscale Materialsand the Advanced Photon Source, both U.S.Department of Energy (DOE) Office of ScienceUser Facilities at DOE's Argonne NationalLaboratory, announced the use of machinelearning tools to accurately predict the physical,chemical and mechanical properties ofnanomaterials.
In a study published in The Journal of PhysicalChemistry Letters, a team of researchers led byArgonne computational scientist SubramanianSankaranarayanan described their use of machinelearning tools to create the first atomic-level modelthat accurately predicts the thermal properties ofstanene, a two-dimensional (2-D) material made upof a one-atom-thick sheet of tin.
The study reveals for the first time an approach tomaterials modeling that applies machine learningand is more accurate at predicting materialproperties compared to past models.
"Predictive modeling is particularly important fornewly discovered materials, to learn what they'regood for, how they respond to different stimuli andalso how to effectively grow the material forcommercial applications—all before you invest incostly manufacturing," said Argonne postdoctoralresearcher Mathew Cherukara, one of the leadauthors of the study.
Traditionally, atomic-scale materials models havetaken years to develop, and researchers have hadto rely largely on their own intuition to identify theparameters on which a model would be built. But byusing a machine learning approach, Cherukara andfellow researchers were able to reduce the need forhuman input while shortening the time to craft anaccurate model down to a few months.
"We input data obtained from experimental orexpensive theory-based calculations, and then askthe machine, 'Can you give me a model thatdescribes all of these properties?'" said BadriNarayanan, an Argonne postdoctoral researcherand another lead author of the study. "We can alsoask questions like, 'Can we optimize the structure,induce defects or tailor the material to get specificdesired properties?'"
1 / 3
The Argonne research team that has pioneered the useof machine learning tools in 2-D material modeling.Credit: Wes Agresta/Argonne National Laboratory
Unlike most past models, the machine learningmodel can capture bond formation and breakingevents accurately; this not only yields more reliablepredictions of material properties (e.g. thermalconductivity), but also enables researchers tocapture chemical reactions accurately and betterunderstand how specific materials can besynthesized.
Another advantage of building models usingmachine learning is the process is not material-dependent, meaning researchers can look at manydifferent classes of materials and apply machinelearning to various other elements and theircombinations.
The computational model Cherukara, Narayananand their colleagues have developed describesstanene, a structure made of tin that has caught theeye of researchers in recent years. Interest instanene mirrors a growing interest in 2-D materialsevolving from the 2004 discovery of graphene, asingle-layer arrangement of carbon with attractiveelectronic, thermal and mechanical properties.While stanene remains far from commercialization,researchers find it promising for applications inthermal management (the regulation of heat)across some nanoscale devices.
The study, "Ab Initio-Based Bond Order Potential to
Investigate Low Thermal Conductivity of StaneneNanostructures," appeared in the The Journal ofPhysics Chemistry Letters.
More information: Mathew J. Cherukara et al,-Based Bond Order Potential to Investigate LowThermal Conductivity of Stanene Nanostructures, The Journal of Physical Chemistry Letters (2016). DOI: 10.1021/acs.jpclett.6b01562
Provided by Argonne National Laboratory
2 / 3
APA citation: Machine learning enables predictive modeling of 2-D materials (2016, December 7)retrieved 20 March 2022 from https://phys.org/news/2016-12-machine-enables-d-materials.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, nopart may be reproduced without the written permission. The content is provided for information purposes only.
Powered by TCPDF (www.tcpdf.org)
3 / 3
