北京大学学报 （ 自然科学版 ） , 第40 卷,第3期.pdf

( ) , 40 , 3 , 2004 5 Acta Scientiarum Naturalium Universitatis Pekinensis, Vol. 40, No. 3 ( May, 2004) A New MD Statistical Noise Improvement 1) Algorithm for Dose Effect Simulation SHI Xiaokang SHI Hao YUMin H UANG Ru ZHANG Xing (Institute of Microelectronics, Peking University, ejiing, 100871; email: shixk@ ime.pku. edu. cn) Abstract A new method for the acceleration of MD (molecular dynamics) simulation of ion implantation into crystalline targets is presented. The READE ( Rare Event Algorithm with Dose Effect) method can well speed up the MD simulation of ion implantation with dose effect. As a result, the time required performing a simula tion with high precision and dose effect is drastically reduced. And many details producing statistical noise in simulation is involved in the READE and help it to be more reasonable and reliable. Key words statistical noise; dose effect; ion implantation; READE; MD simulation Introduction Simulation of low energy ion implantation is very important for ultra shallow junction technology. [ ] Cascade collisions and thus damage buildup notably influence the range profiles and cause dose effect (Fig. ) . The range profiles directly influence the forming of ultra shallow junction. In addition, the [2] damage buildup due to cascade collisions is amply necessary for further annealing simulations . The BED ( Boron Enhanced Diffusion) and TED (Transient Enhanced Diffusion) phenomenon are ba