Prediction of Quantum Materials for Environmental Gas Monitoring

Author : Paul C.H. Li, Sepehr Roonasi, Ka Ho Wong, Shabnam Rehmat

Abstract : The recently discovered topological insulator (TI) exhibits quantum properties of uni-directional edge electrical conduction that favor measurements with high signal/noise ratio. These materials are favorable for highly sensitive and selective environmental monitoring for gases such as NO2 . The materials properties of known TIs, such as Bi2 Se3 , will be used in machine learning (ML) in artificial intelligence (AI) to train the convoluted neural network (CNN) used in deep learning. To optimize sensor performance, new TI materials will be predicted using CNN. This has accelerated material designs by properties prediction with near ab initio accuracy at much faster speed. A trained orbital graph convolutional neural network (OGCNN) has been adapted to accurately predict different crystals with various structure types and compositions. The material properties include formation energy, band gap, Fermi energy, spin-orbital coupling spillage, etc. The material structure is represented by a crystal graph. The atom features include group and period numbers, valence electrons, ionization energy, electron affinity, atomic volume; the bond feature includes atom distance. In addition, the orbital-orbital interactions are also included by the two-dimensional descriptor called the orbital field matrix (OFM). The inclusion of these interactions to encode the local chemical environments of atoms along with embedding of an encoder-decoder network enabled the OGCNN to achieve higher accuracy for prediction of new topological quantum sensing materials.

Keywords : Topological Insulators, Machine Learning, Gas Sensors, OGCNN, Materials Design

Conference Name : International Conference on Environmental Applications of Machine Learning (ICEAML-26)

Conference Place : Lima, Peru

Conference Date : 7th Aug 2026