Efficient kinetic Monte Carlo simulation of annealing in semiconductor materials.

详细信息   

• 作者：Hargrove ; Paul Hamilton.
• 学历：Doctor
• 年：2003
• 导师：Plummer, James D.
• 毕业院校：Stanford University
• 专业：Engineering, Materials Science.;Computer Science.
• CBH：3104236
• Country：USA
• 语种：English
• FileSize：16943474
• Pages：266

文摘

As the semiconductor manufacturing industry advances, the length scales of devices are shrinking rapidly, in accordance with the predictions of Moore's Law. As the device dimensions shrink the importance of predictive process modeling to the development of the production process is growing. Of particular importance are predictive models which can be applied to process conditions not easily accessible via experiment. Therefore the importance of models based on physical understanding are gaining importance versus models based on empirical fits alone. One promising research area in physical-based models is kinetic Monte Carlo (kMC) modeling of atomistic processes.;This thesis explores kMC modeling of annealing and diffusion processes. After providing the necessary background to understand and motivate the research, a detailed review of simulation using this class of models is presented which exposes the motivation for using these models and establishes the state of the field.;The author provides a user's manual for ANISRA ( ANnealIng Simulation libRAry), a computer code for on-lattice kMC simulations. This library is intended as a reusable tool for the development of simulation codes for atomistic models covering a wide variety of problems. Thus care has been taken to separate the core functionality of a simulation from the specification of the model.;This thesis also compares the performance of data structures for the kMC simulation problem and recommends some novel approaches. These recommendations are applicable to a wider class of model than is ANISRA, and thus of potential interest even to researchers who implement their own simulators.;Three example simulations are built from ANISRA and are presented to show the applicability of this class of model to problems of interest in semiconductor process modeling. The differences between the models simulated display the versatility of the code library. The small amount of code written to construct and modify these simulations indicates the value this reusable library may have in easing the development effort associated with kMC simulations.;Finally the author presents a list of open research directions which may lead to more efficient and more widely applicable kMC simulation tools.