【作者】 安之南；

【导师】 戎咏华；

【作者基本信息】 上海交通大学 ， 材料物理与化学， 2009， 硕士

【摘要】 纳米材料的热稳定性关系到纳米材料能否在较高温度下保持纳米级的晶粒尺寸,而晶粒尺寸的稳定性关系到纳米材料能否在较高温度下保持优异的机械性能和理化性能。因此,纳米材料的热稳定性研究一直是纳米材料研究的一个热点。目前,国内外学者在对纳米材料的热稳定性研究中已经发现了多种抑制晶粒生长的机制,包括:溶质颗粒拖曳,第二相颗粒拖曳,气孔,空位,三叉晶界拖曳,厚度效应等。关于厚度效应抑制晶粒生长,虽然从上世纪40年代末就已被提出来,但国内外对厚度效应对纳米材料晶粒生长抑制作用的研究却鲜有报道,尤其是退火过程中厚度效应抑制晶粒生长的动力学研究,至今仍未有详细的研究。本论文采用各向异性蒙特卡罗方法模拟不同厚度的纳米薄膜退火过程中的晶粒尺寸演化,并采用磁控溅射方法制备纳米多晶薄膜,从理论模拟和实验两方面对纳米薄膜中的厚度效应进行研究。将Read-Shockley各向异性晶界能模型引入蒙特卡罗模拟,并对经典的蒙特卡罗模型进行了改进处理:1.改进通常3D模拟中周期性边界条件的设定,将薄膜厚度方向设置为自由边界,其他两个方向为周期性边界;2.统计晶粒尺寸时只考虑最近邻单元,而不再考虑次近邻和第三近邻;3.减小系统最大取向数为64,在确保取向值不影响晶粒生长的情况下尽可能使取向值的设定满足晶界的一般定义;4.采用加速法,认为选定单元只能从邻近单元中选择取向值,而不再是任意的取向值。模拟结果显示:厚效应并不是在晶粒生长初期便呈现出来,而是当晶粒尺寸达到厚度的0.8-1.2倍的时候才变得明显。通过引入厚度因子来修正Burke提出的动力学模型,可以得到描述纳米薄膜中厚度效应抑制晶粒生长的动力学方程。该方程较前人提出的方程物理含义更为清晰,且与模拟结果更为符合。为从实验上研究纳米薄膜中晶粒生长的厚度效应并检验模拟结果和修正方程的合理性,薄膜体系的设计需要满足研究厚度效应的标准,并尽可能与模拟条件相符,为此本文设计了厚度为纳米尺寸的Co/SiO₂以及Ni/SiO₂多层膜结构,以此研究该薄膜退火时的晶粒尺寸演化过程。该多层膜体系由于每个金属单层均被SiO₂非晶层隔开,可以认为每一个金属单层是一个独立的纳米薄膜结构,并且金属单层满足模拟的边界条件,即薄膜的上下表面处于同一状态。此外,多层膜结构具有很好的统计性,相比单层膜数据可靠性高。薄膜的制备方法采用磁控溅射法,该方法可以沉积表面平整,膜内晶粒尺寸小至10nm左右的金属薄膜,并且通过控制溅射功率和沉积时间以制备不同调制层厚度的纳米级多层膜结构。将多层膜结构在特定温度下退火不同时间。通过XRD,SEM,TEM表征其微观组织,并采用Voigt函数计算平均晶粒尺寸。实验结果显示:1. Co/SiO₂多层膜结构XRD数据显示,金属Co层以大量fcc-β相和少量hcp-α相存在。将晶粒的生长过程用修正后的Burke方程进行拟合,其拟合结果明显好于修正前的Burke方程。两种厚度的Co薄膜中晶粒生长的极限尺寸均小于厚度,与早期单相的晶粒生长研究不完全相符,可能是因为少量hcp-α相的存在,抑制了fcc-β相晶粒生长;2. Ni/SiO₂多层膜结构XRD数据显示,金属Ni层均以fcc-γ相存在。将晶粒的生长过程用修正后的方程拟合可以得到很好的拟合结果,而用Burke方程拟合,结果相对较差。以上的理论模拟和实验结果均表明,Burke方程不能准确描述纳米材料的厚度抑制效应,而本文将厚度因子引入后,建立了新的修正方程,该方程从理论解释和实验数据拟合上均优于Burke方程。更多还原

【Abstract】 The grain size stability of nanocrystalline materials determines whether they can keep their unique mechanic properties and the properties of physics and chemistry in high temperature, while whether the nano-grains can keep their size within nano-scale depends on their thermal stability. Therefore, nanocrystalline materials’thermal stability has extensively attracted researchers through out the world. So far, investigators domestic and overseas have found out several stagnation effects on grain growth, including solute drag, Zener drag, pore, vacancy, triple junction, and thickness effect. Although thickness effect was firstly reported in 1940s, the studies of thickness effect on grain growth in nanocrystalline materials have seldom been done till now, especially on kinetics of thickness effect on grain growth in nanocrystalline materials during annealing process.To investigate grain growth in nanocrystalline materials during annealing process, both 3D anisotropic Monte Carlo simulation and magnetic sputtering experiments were used.Read-Shockley grain boundary energy model together with some modifications were introduced in our Monte Carlo simulation. These modifications were:1. Thin films’thickness direction is set as free boundary condition while the other two directions are periodic boundary conditions;2. Only the nearest units with the same orientation number （Q） are considered as one grain;3. Qmax is set to 64;4. Monte Carlo acceleration is used to the reorientation of each unit available within the nearest neighbors.The simulation results indicate that thickness effect is not exhibited through the whole process of grain growth, but appears only when the average grain size reached 0.8 to 1.2 times of the thickness of the films. By inducing thickness factor, we modified Burke’s grain growth kinetic equation and obtained a new one not only reflecting the mechanism of thickness effect but also according with the experimental data better.In order to verify our Monte Carlo simulation and the modified kinetic equation, experiments of thickness effect on grain growth in nanocrystalline thin films are designed to meet the requirements of thickness effect and satisfy the simulation conditions. We designed two kinds of Co/SiO₂ and Ni/SiO₂ multilayer systems. Each metal layer was separated by its two amorphous SiO₂ neighbors with the two interfaces, which is the same condition as described in our simulation. Additionally, the multilayer system has better statistical estimation than single layer system, which leads to a better reliability of experimental data. The nanocrystalline multilayer films were deposited by means of magnetic sputtering method for its advantages in fabricating films with flat surfaces or/and interfaces and nanoscale grain size. By controlling the sputtering power and deposition time, two group samples with different thicknesses were deposited for each kind of multilayer system, and then were annealed at a certain temperature with different time. Their microstructures were characterized by X-ray diffraction （XRD）, scanning electron microscopy （SEM）, and transmission electron microscopy （TEM）, and the average grain size was calculated by using Voigt function.The research results were concluded as follows:1. The XRD data of Co/SiO₂ multilayer films show that the Co metal layer consists of majority of fcc-βphase and minority of hcp-αphase. The fitted curve of grain growth with our modified equation is better consistent with the experimental data than Burke’s one. However, the maximum average grain size in each group of Co/SiO₂ multilayer film is smaller than the film’s thicknesses. This result disagrees with the early experimental results in grain growth in metal films of single phase, and may be explained as the result of Zener drag, namely the hcp-αgrains as particles inhibits grain growth of fcc-βphase.2. The XRD data of Ni/SiO₂ multilayer films shows that every Ni layer consists of single fcc-γphase. The fitted curve of grain growth experimental data with our modified equation is in better accordance with the data than Burke’s one.The results of simulation and experiments demonstrate that Burke’s equation could not precisely describe thickness effect on grain growth in nanocrystalline thin films. However, by introducing thickness factor into Burke’s equation, we deduced a new modified equation which could better describe thickness effect both in simulation and in experiments.更多还原