Theoretical Research On Solar Energy Conversion Materials Based On First Principles And Machine Learnin

Solar energy conversion materials include solar cell materials and photo-rechargeable materials（with both photo power generation and energy storage functions）,etc.Solar energy conversion materials,led by solar cell materials,have been a research hotspot in recent decades,resulting in more and more known materials,but it has been difficult to find a winner between high efficiency and low cost.We believe that a large number of excellent solar energy conversion materials are hidden in various literatures.In order to find them,we often need to read a large number of materials science literature manually,find candidates,and verify them through traditional methods such as experiments,which consumes a lot of human and material resources.How to select candidate solar energy conversion materials more quickly and comprehensively and verify them more conveniently has become the key to the problem.Aiming at this problem,this paper seeks to improve the prediction speed of solar energy conversion materials,increase the prediction space and quickly verify or explain the mechanism,mainly including the following aspects:（1）With the help of the ability of computer to deal with high-intensity calculation,the energy band structure,optical properties,ion migration path and potential barrier of photo-rechargeable material（C₆H₉C₂H₄NH₃）₂Pb I₄are calculated and simulated by first principle.The hole polaron is calculated,and the mechanism of lithium ion adsorption and desorption when it is used as an photo-rechargeable battery is explained.（2）Using a database of 20,000 English literature abstracts for unsupervised machine learning,the potential photo-rechargeable materials Na S,Li₂Fe Si O₄and Cd WO₄are successfully predicted,and first principle calculations were performed to verify that they have good photoelectric and lithium ion energy storage properties.（3）Explore the scope of the literature database,use the database containing 200,000Chinese literature abstracts for unsupervised machine learning,predict the potential solar cell materials Ti Nb₂O₇、Bi PO₄and Y₂O₃successfully,and carry out the first principle calculation to verify that they have good photoelectric properties.