Development of combined FIM/STM system and investigation of the mechanism of thin-film growth on the atomic level

虽然表面扩散是薄膜生长中的一个重要过程，但由于很难在不干扰原子运动的情况下直接观察原子的扩散过程，因此使用场离子显微镜（Field ion microscope，简称FIEMS）的工作量有限。这是因为电子显微镜的应用仅限于钨、铂等少数几种金属。我们试图开发一种通过扫描隧道显微镜（STM）观察表面扩散的方法，这种方法适用于观察各种材料的表面。成功地控制了样品温度，在30 ~ 320 K之间获得了良好的Si（111）和Si（100）像，并制作了超高真空蒸发器。我们试图观察原子的运动，而不影响它的运动，只在30 K的样品温度下扫描，原子的位置被冻结，并让原子在更高的温度下自由扩散，而不扫描。但发现由于热漂移的存在，很难一直盯着同一个位置。这意味着某种标记应该是必要的，以便始终找到相同的位置。我们还首次利用Pt和Ir的直接测量数据，通过速率方程和动力学Monte Carlo模拟研究了表面扩散和团簇稳定性对薄膜生长的影响。虽然集群扩散的重要性已经声称一些团体的基础上的定性参数，它是第一次使用实验扩散数据，集群的稳定性是更重要的增长相比，集群扩散。

Although surface diffusion is an important process in thin film growth, because of the difficulty in observing the diffusion process of an atom directly without disturbing its motion, only a limited amount of work using field ion microscope (FIM) have been reported. This is because the application of FIM is restricted to a few metals like tungsten, platinum and so on. We have tried to develop a method to observe surface diffusion by a scanning tunneling microscope (STM) which is useful for observing the surface of a wide range of materials. We succeeded to control the sample temperature and got a nice of image of Si (111) and Si (100) between 30K and 320K.We also fabricated an evaporator for the deposition of silicon in ultra-high vacuum. We tried to observe the motion of an atom without affecting its motion by scanning only at the sample temperature of 30K where the position of the atom is frozen and letting the atom diffuse freely at a higher temperature without scanning. But it is found that because of thermal drift, it is very difficult to look at the same position constantly. This suggests some sort of a marker should be necessary to always find the same position. The sharp top of a tip created by FIM might be a possibility.We also investigated the effect of surface diffusion and cluster stability on thin-film growth by rate equation and kinetic Monte Carlo simulation using the directly measured data for Pt and Ir for the first time. Although the importance of cluster diffusion has been claimed by some groups based on a qualitative argument, it is found for the first time using experimental diffusion data that the cluster stability is much more important for growth compared to cluster diffusion.