Quotient complex (QC)-based machine learning for 2D perovskite design

Chuan-Shen Hu,Rishikanta Mayengbam,Kelin Xia,Tze Chien Sum
2024-07-24
Abstract:With remarkable stability and exceptional optoelectronic properties, two-dimensional (2D) halide layered perovskites hold immense promise for revolutionizing photovoltaic technology. Presently, inadequate representations have substantially impeded the design and discovery of 2D perovskites. In this context, we introduce a novel computational topology framework termed the quotient complex (QC), which serves as the foundation for the material representation. Our QC-based features are seamlessly integrated with learning models for the advancement of 2D perovskite design. At the heart of this framework lies the quotient complex descriptors (QCDs), representing a quotient variation of simplicial complexes derived from materials unit cell and periodic boundary conditions. Differing from prior material representations, this approach encodes higher-order interactions and periodicity information simultaneously. Based on the well-established New Materials for Solar Energetics (NMSE) databank, our QC-based machine learning models exhibit superior performance against all existing counterparts. This underscores the paramount role of periodicity information in predicting material functionality, while also showcasing the remarkable efficiency of the QC-based model in characterizing materials structural attributes.
Computational Engineering, Finance, and Science,Algebraic Topology
What problem does this paper attempt to address?
### What problem does this paper attempt to solve? This paper aims to solve the problem of insufficient representation in the design and discovery of two - dimensional (2D) perovskite materials. Specifically, the existing material representation methods cannot fully capture the complex structural features of 2D perovskites, thus hindering their application and development in photovoltaic technology. #### Background and Problem Description 1. **Importance of 2D Perovskites** - 2D halide - layered perovskites show great potential in the field of photovoltaic technology due to their excellent photoelectric properties and stability. - However, the existing material representation methods are insufficient in capturing the high - order interactions and periodic information of 2D perovskites, limiting the efficiency of their design and discovery. 2. **Limitations of Existing Methods** - Traditional methods such as Smooth Overlap of Atomic Positions (SOAP) features, although having achieved certain success, are still difficult to comprehensively describe the complex structure of 2D perovskites. - These methods mainly focus on pairwise atomic interactions, ignoring higher - order or many - body interactions and the periodic information of materials. 3. **Introduction of a New Framework** - To solve the above problems, the authors introduce a novel computational topological framework - Quotient Complex (QC) as the basis for material representation. - The QC framework provides a more comprehensive material representation method by encoding higher - order interactions and periodic information. #### Solutions 1. **Quotient Complex (QC) and Its Characteristics** - QC is a quotient variant of simplicial complexes generated based on the unit cell and periodic boundary conditions of materials. - It can encode high - order interactions and periodic information simultaneously, surpassing previous material representation methods. 2. **Application of Machine Learning Models** - Based on the features generated by QC (called Quotient Complex Descriptors, QCDs), combined with the Gradient Boosting Tree (GBT) model, it is used to predict the band gap of 2D perovskites. - Through verification using the New Materials for Solar Energetics (NMSE) database, the superior performance of the QC - GBT model in predicting the band gap of 2D perovskites is proved. #### Summary This paper solves the deficiencies of existing methods in representing 2D perovskite materials by introducing the Quotient Complex (QC) framework, especially in capturing high - order interactions and periodic information. This new method not only improves the accuracy of material representation but also significantly enhances the ability to predict the material band gap, providing strong support for the design and discovery of 2D perovskites. ### Key Formulas 1. **Definition of Unit Cell** \[ U = U(v_1, v_2, v_3)=\left\{\sum_{i = 1}^{3}a_i v_i\mid a_i\in[0,1]\right\} \] where \(v_1, v_2, v_3\) are basis vectors in three - dimensional space. 2. **Definition of Lattice Group** \[ \Lambda=\Lambda(v_1, v_2, v_3)=\left\{\sum_{i = 1}^{3}a_i v_i\mid a_i\in\mathbb{Z}\right\} \] 3. **Definition of Periodic Set** \[ V = M+\Theta=\bigcup_{w\in\Theta}(M + w) \] where \(M\) is a finite subset within the unit cell and \(\Theta\) is a finite subset of the lattice group \(\Lambda\). 4. **Definition of Quotient Complex** \[ K = K_{\epsilon}(V)=\left\{\text{conv}(X)\mid\text{diam}\right\}