Sign up to save your podcasts
Or
Share Machine-Learning in Material Science to find new rare-earth compounds with Dr. Singh
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
Machine-Learning in Material Science to find new rare-earth compounds with Dr. Singh
Join us in this engaging episode with Dr. Singh, who introduces us to an exciting application of machine learning in the realm of material science. He elucidates how chemical alloying can impact the formation enthalpy of rare-earth intermetallics.
The use of machine learning in rare-earth intermetallic design has been minimal, largely due to the limited availability of reliable datasets. To overcome this, Dr. Singh and his team have developed an extensive 'in-house' rare-earth database, containing over 600 compounds. Each entry in this database is enriched with formation enthalpy data and associated atomic features obtained using high-throughput density-functional theory (DFT).
With this resource at their disposal, Dr. Singh's team then applied the SISSO (Sure Independence Screening and Sparsifying Operator) based machine learning method to train and test the formation enthalpies of these rare-earth compounds. This approach enabled them to delve into the effects of transition metal alloying on the energy stability of Ce based cubic Laves phases (MgCu type).
The SISSO predictions, which align well with high-fidelity DFT calculations and X-ray powder diffraction measurements, provide invaluable quantitative guidance for compositional considerations within a machine-learning model. This contributes significantly to the discovery of new metastable materials.
To deepen our understanding, Dr. Singh also analyzes the electronic-structure of a Ce-Fe-Cu based compound, offering insights into the electronic origin of phase stability. This fusion of interpretable analytical models, density-functional theory, and experimental methods presents a quick and reliable design guide for discovering technologically useful materials.
Whether you're a materials scientist, a machine learning enthusiast, or just someone fascinated by the intersection of technology and science, this episode with Dr. Singh is a must-listen!
Keywords: Machine learning, Rare-earth intermetallics, Chemical alloying, Formation enthalpy, High-throughput density-functional theory (DFT), SISSO, Energy stability, Cubic Laves phases, Metastable materials, Electronic-structure, Phase stability, Material discovery.
https://doi.org/10.1016/j.actamat.2022.117759
View all episodes
By Catarina CunhaJoin us in this engaging episode with Dr. Singh, who introduces us to an exciting application of machine learning in the realm of material science. He elucidates how chemical alloying can impact the formation enthalpy of rare-earth intermetallics.
The use of machine learning in rare-earth intermetallic design has been minimal, largely due to the limited availability of reliable datasets. To overcome this, Dr. Singh and his team have developed an extensive 'in-house' rare-earth database, containing over 600 compounds. Each entry in this database is enriched with formation enthalpy data and associated atomic features obtained using high-throughput density-functional theory (DFT).
With this resource at their disposal, Dr. Singh's team then applied the SISSO (Sure Independence Screening and Sparsifying Operator) based machine learning method to train and test the formation enthalpies of these rare-earth compounds. This approach enabled them to delve into the effects of transition metal alloying on the energy stability of Ce based cubic Laves phases (MgCu type).
The SISSO predictions, which align well with high-fidelity DFT calculations and X-ray powder diffraction measurements, provide invaluable quantitative guidance for compositional considerations within a machine-learning model. This contributes significantly to the discovery of new metastable materials.
To deepen our understanding, Dr. Singh also analyzes the electronic-structure of a Ce-Fe-Cu based compound, offering insights into the electronic origin of phase stability. This fusion of interpretable analytical models, density-functional theory, and experimental methods presents a quick and reliable design guide for discovering technologically useful materials.
Whether you're a materials scientist, a machine learning enthusiast, or just someone fascinated by the intersection of technology and science, this episode with Dr. Singh is a must-listen!
Keywords: Machine learning, Rare-earth intermetallics, Chemical alloying, Formation enthalpy, High-throughput density-functional theory (DFT), SISSO, Energy stability, Cubic Laves phases, Metastable materials, Electronic-structure, Phase stability, Material discovery.
https://doi.org/10.1016/j.actamat.2022.117759