Structure Optimization And Configuration Identification Method For Large-size Alloy Clusters

The large-size alloy clusters composed of Au,Ag,Pd and Pt have great potential in the fields of optics,electricity,magnetism and catalysis.The stable structure and properties of alloy clusters have been a hot topic in the field of material chemistry and computational chemistry.The special properties of clusters are closely related to their structure.Therefore,the determination of its stable geometry is the first condition to study the special properties of materials,and the structure is the first to study the properties of materials.Only by understanding the structure of matter can we have a deeper understanding of its nature.In modern science,due to the influence of various external factors,it is still a considerable challenge to determine the structure of matter directly by means of experiments,which will inevitably lead to the low efficiency of research.Through simulation,the structure of the material is not affected by environmental factors,which can provide guidance for the scientific preparation of target materials.In recent years,many methods have been proposed and applied to construct prediction,but there are still some defects in these methods.There are two main difficulties in the prediction of material structure:1.The efficiency and success rate of the existing optimization algorithms are relatively low in large-scale system calculation.2.The method based on density functional theory is accurate,but it takes a long time.In addition,prediction of cluster structure performance relationship has always been a focus area of many calculation and theoretical research.The main contents of this thesis are summarized as follows（1）The structure of Au_nPd_147-n（n=1-147）and Au_nPt_147-n（n=1-147）clusters with 147atoms were optimized by adaptive immune optimization algorithm based on kernel construction.The structural characteristics,stability and distribution of atoms are discussed.The results show that the 147-atom Au-Pd and Au-Pt clusters are complete icosahedral structures.Sequence parameter analysis shows that they all form a nuclear layer structure.（2）Based on the multi-body Gupta potential function,the most stable structure of Ag_nPd_147-n（n=1-147）and Ag_nPt_147-n（n=1-147）clusters with 147 atoms was optimized by using the kernel based adaptive immune optimization algorithm.The structural characteristics,stability and distribution of atoms are discussed.The results show that the 147-atom Ag-Pd and Ag-Pt clusters are all complete icosahedral structures.The sequence parameter analysis shows that they all form nuclear layer structure.（3）Nanoparticles show a variety of structural and morphological characteristics,which are often interrelated,which makes the correlation of structure/property relationship challenging.The optimal and local optimal configurations of univariate and multivariate clusters based on molecular mechanics are collected as the cluster structure database,and the cluster structure feature information is designed.The samples are divided into calibration set and verification set.The cluster structure verification model is designed by using neural network pattern recognition method,and the configuration of the verification set is predicted.The results show that the model has high precision for cluster structure prediction..