First principle determination of dopant ionization levels

Since CMOS (Complementary metal-oxide-semiconductor) circuits were invented in 1963, CMOS technology has made rapid advancements while keeping pace with the so-called Moore's law. With continued scaling, the size of transistors shrinks to the nanometer scale. In order to achieve the desired electronic properties of nanometer transistors, abrupt and high-concentration doping is a critical issue, but extremely challenging because of relatively large surfaces and interfaces as well as complex defect- dopant interactions. Therefore, a deeper understanding of the fundamental behavior and properties of dopants and their interactions with defects in various process conditions is essential in fabrication of nanotechnology-based CMOS devices. In this research, we use first principle quantum mechanical calculations to determine the structure, dynamics and electronic properties of dopants in crystalline silicon (Si). The first part of this thesis focuses on finding methods capable of predicting the ionization levels of dopants, such as P, As, B, and Al, as well as defect-dopant complexes. Since the ionization levels are closely related to doping conditions, it is worthwhile to understand the relationship. We first check carefully the convergence of total energy for the pure and doped systems with respect to calculation conditions, such as k-point mesh size, plane wave cutoff energy, supercell size, and so on. Then, based on the total energy calculations we evaluate the ionization levels of dopants with comparisons to experimental data available in literature. The second part of this thesis examines the structure and dynamics of phosphorus (P) near the Si(100) surface. First, P diffusion in bulk Si is presented as the reference to clarify the surface effect. Then, the relative stability and diffusion of P near the surface are calculated. The results show that surface P is far more energetically favorable than bulk P. Also, our calculations show that with a moderate barrier, bulk P will undergo out-diffusion to the surface, and surface P will migrate along a dimer row. The fundamental findings provide valuable guidance on how to control doping of silicon nanostructures.