【作者】 魏育新；

【导师】 王新强；

【作者基本信息】 重庆大学 ， 凝聚态物理， 2002， 硕士

【摘要】 纳米材料被公认为是21世纪最具有前途的新材料，受到国际上的高度重视，成为目前凝聚态物理学和材料科学的一个前沿研究热点。而纳米团簇作为纳米材料研究的基础，是凝聚态物理研究的最重要的课题之一。团簇可以包含几个到几百个甚至上千个原子。众所周知，随着尺寸的变化，团簇会出现一些新的效应，包括量子尺寸效应、表面效应等。由于具有这些效应，使以团簇作为基本组成单元的纳米材料表现出一些和大块晶体截然不同的独特性质，并为人们所熟知。然而近年来的研究又发现，随着尺寸的进一步减小，纳米团簇中出现了一些令人注目的新现象。比如Na原子团簇（原子数70-200）的熔点随尺寸出现令人惊奇的不规则变化，相差几个原子就可以使熔点改变百分之几十；最新的实验发现由147个Na原子组成的团簇在熔点附近吸收能量温度反而降低，出现了负热容。由10-30个Sn原子组成的团簇的熔点比晶体材料还高出约50K。我们把这些现象称之为反常尺寸效应。研究金纳米团簇的反常尺寸效应不仅具有很高的科学价值，而且对发现和制备具有更为独特性质的纳米材料具有十分重要的指导意义。本文就是利用微正则分子动力学方法和原子嵌入模型来研究金原子纳米团簇反常尺寸效应。原子嵌入法是一种被证明了的能够有效处理贵金属转变的统计性质和动力学性质的多体势能模型。之所以选择金纳米团簇作为研究对象主要是因为一方面是不仅有不少理论结果可以参照，而且还有不少实验结果作检验，使研究更有实际价值；另一方面它是一种很重要的纳米材料，通过对它的研究对其他材料的研究有重要的指导意义。在研究中我们共模拟了从45至767个原子组成的十四种不同尺寸的团簇的升降温过程。首先产生出各种尺寸的面心立方格子球形团簇，然后使团簇温度从室温（300K）一直升高到熔点以上，接着再从熔点以上降至300K以下，在每个温度上都要使团簇达到平衡状态。在模拟中我们得到了各种尺寸的纳米团簇在不同能量下的温度、势能和结构变化图象。研究结果表明在我们模拟的团簇的尺寸范围内：（1）当团簇的原子数目小于90时，团簇的熔点和尺寸不成单调关系，尺寸小的团簇有可能具有更高的熔点，在模拟中Au70的熔点要高于Au80、Au90、Au100、Au120、Au135、Au225的熔点，这说明Au70具有相对稳定的结构，团簇变化少数几个原子就可以使团簇结构发生根本性变化；（2）在尺寸较大（原子数目多于90）的团簇升温过程中观察到负热容，也就是说在熔点附近时，使团簇的能量增加，但是团簇的温度反而降低，这是和大块晶体不同的，小于90个原子的团簇没有观察到负热容现象。存在负热容是因为在熔点<WP=5>附近，当团簇能量增加时，不仅增加的能量全部转化成势能，而且还有一部分动能也转化成势能，动能减少，温度降低；（3）在降温过程中出现结晶温度相对于熔解温度的滞后现象，也就是说结晶温度要小于熔解温度，而在大块晶体中这两者应该是相同的。这种情况说明团簇从有序（固态）转入无序（液态）容易，而从无序到有序较难。并且当团簇的原子数目从120变化到767，尺寸越大，结晶温度和熔解温度的差异也越大。更多还原

【Abstract】 Nanomaterial have become an active topic of material science. Cluster contains few atoms to hundreds of atoms. Its properties are different from a single atom and crystal. With the decreasing number of atom,the cluster may show irregular properties which we call irregular size-effect.Adding or cuting an atom can bring about radical structural change, accordingly lead to the change of properties. We present a detailed molecular-dynamics study of the melting , freezing of gold nanoclusters within the framework of the embedded-atom method . In order to be able to simulate clusters containing more than several hundred particles, it is necessary to resort to an empirical description of the interatomic forces. Here we choose to employ the embedded-atom method, an n-body potential with proven ability to model reliably various static and dynamic properties of transition and noble metals. The model is "semiempirical" in the sense that it approachs the total-energy problem from a local-electron-density view of point. But using a function form with parameters obtained from experiment(equilibrium lattice constant, sublimation energy , bulk modulus, elastic constants. etc)In this study fourteen clusters of different sizes are considered. First we should prepare various face-centered-cubic Au spherical clusters. Of course , the clusters cann’t be perfectly spherical. The atoms follow the Maxwell distribution law. The cluster is heated up to 1000K, then cooled from the highest temperature to 200K. At every temperature the cluster should reach equilibration. Upon approaching the transition from either side, the temperature is changed in steps of 20-30K. In the simulation the change of temperature is achieved by changing energy and use a time step of 0.2fs. In the simulation we get potential energy, total energy and structure varying with time for each cluster. The melting-and-freezing simulations are carried out in the microcanonical ensemble.Our results show (1)the melting point of cluster does not increase monotonically with cluster size. This differs from what we have imagined. The smaller cluster may has higher melting temperature. (2)the microcanonical heat capacity for a greater cluster that contains more than 90 atoms can become negative. Every day experience tells us that if one adds energy to a system it will get warmer. But an increase of energy for nanoclusters may lead to a lower temperature near its solid to liquid transition. This has been observed in experiment for a cluster containing 147 sodium atoms. We try to<WP=7>explain it in term of energy. (3)In the melt-quench cycle we observe the hysteresis . It indicates it is much easier for a cluster to go from an ordered state to a disordered state than the opposite. Namely the freezing temperature is lower than the melting temerature, however both are same for crystal. According to the simulated clusters,the greater the cluster is, the greater the difference is. Finally we discuss the factor that affects the simulation.更多还原