Modeling of polysilicon thin-film transistors formed by grain enhancement technology: Metal-induced lateral crystallization.

摘要

Metal-Induced-Lateral-Crystallization (MILC) using nickel (Ni) has been recognized as a promising crystallization technique to form large-grain polysilicon used in many applications such as thin-film-transistor (TFT) on active matrix displays and other system-on-panel applications. In this study, the growth mechanism of MILC was investigated. Based on the mechanism of the crystallization and the time evolution of the metallic impurity in the amorphous silicon film being crystallized, a model to predict the growth rate and the final Ni distribution in the polycrystalline silicon was developed. The accuracy and reliability of the model were extensively validated by experimental results. Thus the model can be used to optimize the grain growth conditions for fabricating high performance TFTs on the recrystallized large-grain polysilicon film.; Variation of performance of the polysilicon TFTs is another crucial consideration to the flat panel display manufacturing and TFT circuit integration, especially when the polysilicon grain size is comparable to the transistor size under grain enhancement technology. As the amount of grain boundaries in the channel is a key factor varying the device performance, a probabilistic model to predict the statistical distribution of the grain boundaries in the channel was created. Device-to-device variation was successfully predicted by cooperating the effects and the statistical distribution of grain boundaries in the channel. The study is not only useful to the polysilicon TFTs formed by MILC, it can also be applied to other polysilicon crystallization technologies to optimize the yield of production and the transistor size before the actual fabrications. For TFT circuit integration using the MILC technology, a technique to form high quality polysilicon in short annealing time and a low thermal budget is always our expectation. To this end, pulsed-rapid thermal annealing was investigated in my research. It was found that it could be an effective method to tackle the limitations of MILC using conventional constant temperature annealing and form large-grain polysilicon in an extremely short annealing period and low thermal budget.