English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

The recently proposed crystal graph convolutional neural network (CGCNN)offers a highly versatile and accurate machine learning (ML) framework bylearning material properties directly from graph-like representations ofcrystal structures ("crystal graphs"). Here, we develop an improved variant ofthe CGCNN model (iCGCNN) that outperforms the original by incorporatinginformation of the Voronoi tessellated crystal structure, explicit 3-bodycorrelations of neighboring constituent atoms, and an optimized chemicalrepresentation of interatomic bonds in the crystal graphs. We demonstrate theaccuracy of the improved framework in two distinct illustrations: First, whentrained/validated on 180,000/20,000 density functional theory (DFT) calculatedthermodynamic stability entries taken from the Open Quantum Materials Database(OQMD) and evaluated on a separate test set of 230,000 entries, iCGCNN achievesa predictive accuracy that is significantly improved, i.e., 20% higher thanthat of the original CGCNN. Second, when used to assist high-throughput searchfor materials in the ThCr2Si2 structure-type, iCGCNN exhibited a success rateof 31% which is 310 times higher than an undirected high-throughput search and2.4 times higher than that of the original CGCNN. Using both CGCNN and iCGCNN,we screened 132,600 compounds with elemental decorations of the ThCr2Si2prototype crystal structure and identified a total of 97 new unique stablecompounds by performing 757 DFT calculations, accelerating the computationaltime of the high-throughput search by a factor of 130. Our results suggest thatthe iCGCNN can be used to accelerate high-throughput discoveries of newmaterials by quickly and accurately identifying crystalline compounds withproperties of interest.

Developing an improved Crystal Graph Convolutional Neural Network framework for accelerated materials discovery