| About this Abstract |
| Meeting |
MS&T24: Materials Science & Technology
|
| Symposium
|
Frontiers of Machine Learning on Materials Discovery
|
| Presentation Title |
A Hierarchical Machine Learning Scheme to Identify Promising New Scintillators |
| Author(s) |
Anjana Talapatrra, Ghanshyam Pilania, Christopher Stanek, Blas P. Uberuaga |
| On-Site Speaker (Planned) |
Blas P. Uberuaga |
| Abstract Scope |
When considering the periodic table and all of the elements that can be combined to make a new compound with some functionality, one quickly becomes overwhelmed by the options. Further, several functionalities may be desired, making any search that much harder. In this work, we use a series of machine learning models to down select from the vast number of potential perovskite scintillators to a tractable set that is amenable to experimentation. We first consider the synthesizability of the compound, from two different perspectives: experimental formability and theoretical stability. We then classify the compounds as having either a small or large band gap and, in the case of large band gaps, we train a regression model to predict the band gap. Finally, we determine whether activator states reside in the gap or not, a key feature for scintillation. We end with a list of compounds that have promise as scintillators. |