Research On Poisson Effect And Machine Learning Of Semiconductor Materials

Materials with negative Poisson's ratio have abnormal mechanical properties,such as transverse contraction(expansion)during uniaxial compression(stretching),and generally have desirable fracture resistance,shear and energy absorption.Due to its excellent mechanical properties,negative Poisson's ratio materials have a wide range of applications in aerospace,medical equipment,national defense engineering and other fields.At present,artificial negative Poisson's ratio materials are widely used in industry.However,the production cost of artificial Poisson materials is high,and negative Poisson's ratio effect is basically realized through structural design,which makes it difficult to endow materials with good electrical properties,and becomes the bottleneck for negative Poisson's ratio materials to move from laboratory to industrialization.Therefore,researchers pay great attention to the research of natural negative Poisson's ratio materials,hoping to find more natural negative Poisson's ratio materials,especially those with excellent electrical properties,to promote the development of functional materials with negative Poisson's ratio.However,at present,there are few natural negative Poisson materials,which are difficult to meet the demand of related industries.Traditional methods of studying Poisson effect of materials are selected by experimental measurement or high-throughput calculation,which will consume a lot of human and time costThe combination of machine learning algorithm and theoretical calculation can greatly accelerate the research speed of Poisson materials.By learning the calculated Poisson's ratio of materials,the model can be constructed to accurately predict the Poisson's ratio of a large number of other materials,so that the materials with negative Poisson's ratio effect can be quickly screened out.In this paper,a regression model for predicting Poisson's ratio of materials is constructed by machine learning based on the Poisson's ratio of existing materials,which can accurately predict the Poisson's ratio of a large number of materials with negative Poisson's ratio.Based on the first-principles calculation of density functional theory,the negative Poisson's ratio effect and semiconductor properties of bulk group-?monochalcogenides are studied in detail,and the generation mechanism of negative Poisson's ratio is discussed.The main work of this paper is as follows:(1)Pymatgen package was used to collect the crystal structure,chemical composition and isotropic Poisson ratio of 5000 Materials in the Materials Project database.After data preprocessing,feature extraction was carried out on the element composition and spatial structure of the material,and 135 feature quantities were obtained.Then,the prediction model of Poisson ratio of the material was constructed by using the random forest algorithm.Finally,we used the 10-fold cross-validation method to verify the robustness of the model,and calculated the evaluation parameters of the model,R2test=0.812 and RMSEtest=0.074,indicating that the model could accurately predict the isotropic Poisson's ratio of materials(2)We focused on the study of bulk group-? monochalcogenides with negative Poisson's ratio effect,and calculated the lattice constants,bond lengths,bond angles and band gaps of bulk group-? monochalcogenides in the ground states of GeS,GeSe,SnS and SnSe by using the first principles of density functional theory.At the same time,we discussed their semiconductor characteristics,and analyzed the variation rule of band gap values when the uniaxial strain in the range of±10%was applied along the armchair and zigzag direction of the crystal structure,respectively(3)After optimizing the structure of the bulk group-? monochalcogenides,uniaxial strain of±10%was applied along the b(zigzag)and a(armchair)direction of the crystal structure,respectively.It was found that when the strain was applied along the direction of b,abnormal mechanical phenomena appeared in the c(vertical)direction of all the four sulfides Among them,GeS has a rare expansion phenomenon on both sides,that is,whether it is stretched or compressed in the b direction,it will expand in the c direction,while the rest three bulk compounds possess negative Poisson's ratios.The negative Poisson's ratio(NPR)of GeSe,SnS and SnSe in the b direction and the c direction is Vbc=-0.015/-0.046/-0.042,respectively.When uniaxial strain was applied along the a direction,we found that the interlayer height of bulk group-? monochalcogenides also possess negative Poisson's ratio effect(4)By analyzing the variation laws of bond length,bond angle,interlayer distance and intralayer height under uniaxial strain,we discussed the mechanism of negative Poisson's ratio effect,and found that NPR effect mainly depends on the interlayer height,which is different from its corresponding two-dimensional material,indicating the van der Waals interaction between layers plays an important role in the mechanical behavior of materials.