Ab-initio Molecular Dynamis Simulation of Metastalde Defect by Electronic Excitation

用从头算分子动力学方法研究了非晶硅（a-Si）、氢化非晶硅（a-Si：H）和晶体硅（c-Si）中孤立悬挂键（D）的性质。我们发现，在所有情况下，sp^3和sp^2类构型的悬挂键都表现出较大的晶格弛豫。类sp^3构型的中性悬挂键表现出负的有效U性质，具有较大的晶格弛豫。因此，中性悬挂键（D^0）在热平衡时是不稳定的，并且不成比例地变成带负电和带正电的悬挂键; 2D^0 --> D^- + D^+。另一方面，sp^2-like构型的中性悬挂键表现出正的有效U性质，随着化学势的增加，悬挂键中出现了带负电荷的悬挂键，悬挂键中的这种具有正负有效U性质的双稳性与X射线辐照增强非晶硅晶化的微观机制有关，即空位间隙-通过X射线照射形成电子对降低了空位的形成能并增强了结晶。我们还发现类sp^3构型悬挂键的负有效U性质与Staebler-Wronski效应的起源有关，其中类sp^2构型的正有效U的中性悬挂键是由可见光照射下的复合增强缺陷反应产生的。我们的预测可以非常合理地解释所有可用的实验数据。

The nature of the isolated dangling-bonds (D) in amorphous silicon (a-Si), hydrogenated amorphous silicon (a-Si : H) and crystalline silicon (c-Si) are investigated using ab-initio molecular-dynamics simulation. We find that the bistability of dangling-bonds at sp^3- and sp^2-like configuration which exhibit a large lattice-relaxation in all cases. Neutral dangling-bond at sp^3-like configuration shows negative effective-U nature with large lattice-relaxation. Therefore the neutral dangling-bond (D^0) is unstable at the thermal equilibrium and disproportionate into negatively- and positively-charged dangling-bond ; 2D^0 --> D^- + D^+. On the other hand, neutral dangling-bond at sp^2-like configuration shows positive effective-U nature, and positively-, neutrally-, and negatively-charged dangling bonds appear with increasing the chemical potential.These bistabilities in dangling-bonds with the positive and negative effective-U nature are related to the microscopic mechanism of the X-ray-irradiation-enhanced crystallization in amorphous silicon in which the vacancy-interstitial-pair formation by X-ray irradiation reduces the formation energy of the vacancy and enhances the crystallization. We also find that the negative effective-U nature of the dangling-bond at sp^3-like configuration is related to the origin of Staebler-Wronski effect in which the neutral dangling-bonds of positive effective-U at sp^2-like configuration are created by the recombination-enhanced defect-reaction with the exposure to the visible light. Our prediction can explain all available experimental data very reasonably.