Abstract
New Nd-Fe-B crystal structures can be formed via the elemental substitution of LA-T-X host structures, including lanthanides (LA), transition metals (T) and light elements, X = B, C, N and O. The 5967 samples of ternary LA-T-X materials that are collected are then used as the host structures. For each host crystal structure, a substituted crystal structure is created by substituting all lanthanide sites with Nd, all transition metal sites with Fe and all light-element sites with B. High-throughput first-principles calculations are applied to evaluate the phase stability of the newly created crystal structures, and 20 of them are found to be potentially formable. A data-driven approach based on supervised and unsupervised learning techniques is applied to estimate the stability and analyze the structure-stability relationship of the newly created Nd-Fe-B crystal structures. For predicting the stability for the newly created Nd-Fe-B structures, three supervised learning models: kernel ridge regression, logistic classification and decision tree model, are learned from the LA-T-X host crystal structures; the models achieved maximum accuracy and recall scores of 70.4 and 68.7%, respectively. On the other hand, our proposed unsupervised learning model based on the integration of descriptor-relevance analysis and a Gaussian mixture model achieved an accuracy and recall score of 72.9 and 82.1%, respectively, which are significantly better than those of the supervised models. While capturing and interpreting the structure-stability relationship of the Nd-Fe-B crystal structures, the unsupervised learning model indicates that the average atomic coordination number and coordination number of the Fe sites are the most important factors in determining the phase stability of the new substituted Nd-Fe-B crystal structures.
Highlights
The major challenge in finding new stable material structures in nature requires high-throughput screening of an enormous number of candidate structures, which are generated from different atomic arrangements in three-dimensional space
For each host crystal structure, a substituted crystal structure is created by substituting all lanthanide sites with Nd, all transition metal sites with Fe and all light-element sites with B
We infer that the large Gauss component corresponds to the distribution of unstable crystal structures and the small Gauss component corresponds to the distribution of potential formable crystal structures. This hypothesis can be verified through comparison with the results of the density functional theory (DFT) calculations, and it can be seen that most of the potential formable crystal structures confirmed by DFT calculation belong to the small Gauss component. This implies that the phase stabilities of the Nd–Fe–B crystals are not significantly related to the coordination number of the Nd sites but are largely determined by the coordination number of the Fe sites, suggesting that, if the Nd sites can be replaced in part by Fe, the crystal structure characteristics of Nd–Fe–B which are directly related to its phase stability can be controlled
Summary
The major challenge in finding new stable material structures in nature requires high-throughput screening of an enormous number of candidate structures, which are generated from different atomic arrangements in three-dimensional space. Given a list of hypothetical structures, ML methods are utilized for recommending the most likely new potential materials using probabilistic models [e.g. Bayesian optimization techniques (Yamashita et al, 2018; Xue et al, 2016b)]. This approach requires a list of potential candidates to be prepared as input, which is primarily based on human intuition. The bottleneck of the current recommendation methods is that a large number of known property materials are required as references for the system to start an effective recommendation process. An ‘understanding’ of the structure–stability relationship can be directly obtained from screening results, which can help in systematically correcting researchers’ suggestions
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
