Study on Initial Stage of Oxidation of Hydrogen-Terminated Silicon Surface and Structure of Interface between Ultra-thin Metal Film and Ultra-thin Silicon Oxide Film

为了在原子尺度上形成高质量的氧化硅和SiO_2/Si界面，制备超清洁、原子平整的硅表面是非常重要的。也就是说，对于Si(111)表面，可以通过在40%NH_4F溶液中处理来实现，而对于Si(100)表面，可以通过在1100ﾟC的氢气中外延生长Si来实现。在任何一种情况下，Si表面都是以氢为端面的，并且是原子平坦的。由于氢封端的硅表面具有化学稳定性，自然氧化物对硅表面的钝化有望被氢气钝化所取代。本研究用X射线光电子能谱详细研究了原子扁平的端氢Si(111)在300ﾟC的1Torr干氧中氧化的初始阶段。将实验结果与氧化过程的模拟结果进行了比较，发现氧化过程是不均匀的，氧化层局部以层状生长方式进行。然而，如果在600和800ﾟC的1Torr干氧中通过300ﾟC的1Torr干氧中形成的0.5 nm厚的氧化物进行氧化，则随着氧化的进行，界面结构出现周期性的变化。这可以用热氧化物的一层一层生长来解释。这些结果可以理解为，在0.5-0.8 nm的临界厚度以下，氧化不均匀进行，而在这一临界厚度以上，界面处的原子台阶均匀流动，导致热氧化物的层状生长。为了证实这一重要发现，金属/硅界面的形成和价带光谱的测量只需很短的时间。通过对1托干氧在200、300和400ﾟC下氧化初期的研究发现，氧化温度越低，氧化过程越均匀。

In order to form high quality silicon oxide and SiO_2/Si interface in atomic scale, it is important to prepare ultra-clean and atomically flat Si surface. Namely, in the case of Si(111) surface it can be realized by the treatment in 40% NH_4F solution, while in the case of Si(100) surface it can be realized by the epitaxial growth of Si in hydrogen gas at 1100ﾟC.In either case Si surface is terminated with hydrogen and is atomically flat. Because hydrogen-terminated Si surface is chemically stable, the passivation of Si surface with native oxide is expected to be replaced by that with hydrogen. In the present study initial stage of oxidation of atomically flat hydrogen-terminated Si(111) in 1 Torr dry oxygen at 300ﾟC was studied in details by X-ray photoelectron spectroscopy. By comparing this experimtal results with simulated results of oxidation process, it was found that the oxidation proceeds non-uniformly and layr by layr growth of oxide proceeds locally. However, if the oxidation is performed in 1 Torr dry oxygen at 600 and 800ﾟC through 0.5 nm thick oxide formed in 1 Torr dry oxygen at 300ﾟC, the periodic changes in interface structure appeared with the progress of oxidation. This can be understood by the layr by layr growth of thermal oxide. These results can be understood such that below the critical thickness of 0.5-0.8 nm oxidation proceeds non-uniformly, while above this critical thickness the atomic steps at the interface flow uniformly to result in layr by layr growth of thermal oxide. In order to confirm this important discovery, only a little time could be spent for the formation of metal /Si interface and the measurement of valence band spectra of this system. By the way it was found from the study on the initial stage of oxidation in 1 Torr dry oxygen at 200, 300 and 400ﾟC that the oxidation proceeds more uniformly at lower oxidation temperature.