Molecular dynamics modeling of solid phase epitaxial regrowth

Abstract

Solid phase epitaxial regrowth (SPER) is of great technological importance in semiconductor device fabrication. A better understanding and accurately modeling of its behavior are vital to the design of fabrication processes and the improvement of the device performance. In this paper, SPER was modeled by molecular dynamics (MD) with Tersoff potential. Extensive MD simulations were conducted to study the dependence of SPER rate on temperature, growth orientation, pressure, and uniaxial stress. The simulation data were fitted to empirical formula, and the results were compared with experimental data. It was concluded that MD with Tersoff potential can qualitatively describe the SPER process. For a more quantitatively accurate model, larger simulation systems and a better interatomic potential are needed.

Solid-Phase Epitaxy

“…Recent MD simulations of Si(001) SPE using the Tersoff 3 potential at 1700 K reproduce qualitatively the enhancement by hydrostatic pressure, the retardation by uniaxial in-plane compression, and the small or nonexistent effect of uniaxial in-plane tension [92]. These simulations overestimate y by about an order of magnitude, but reproduce the experimental activation energy within 0.05 eV; they overestimate the experimental activation volume by about a factor of three and underestimate the retardation effect of in-plane uniaxial compression by about the same factor.…”

“…Although the computation time is limited in an MD simulation, the influence of a wide range of experimental situations has been explored, including the dependence of the SPE rate on temperature [163], interface orientation [92,93], pressure [92,94,171,172], stress [92,94], SPE in confined regions [173,174], and B segregation and precipitation during SPE [175]. Figure 7.23 is an Arrhenius plot of the Si SPE rates extracted from MD simulations using several different potentials and compared with the experimental data characterized by an activation energy of 2.7 eV.…”

Molecular-dynamics simulations of solid phase epitaxy in silicon: Effects of system size, simulation time, and ensemble

“…The system was maintained at 300 K for 100 ps (5 × 10 4 Δt) for relaxing the a=c interface, and then was annealed at various temperatures ranging from 1100 to 1600 K for recrystallization. Lai et al 7) performed MD simulations of SPE growth of amorphous Si using four thermostats: direct velocity scaling, Berendsen, 24) Nosé-Hoover, 25) and Langevin. 26) As a result, all four thermostats were found to give similar results.…”

“…Indeed, the simulations of SPE growth for amorphous Si have been performed so far by several researchers, as summarized in Table I. [1][2][3][4][5][6][7][8][9] Bernstein et al 1) examined the SPE processes of ∼2000 Si atoms using the environment-dependent interatomic potential. From the Arrhenius plot of the growth rates, they found that the activation energy of the SPE growth changes at 950 K: 0.4 ± 0.2 eV below 950 K and 2.0 ± 0.5 eV above 950 K. On the contrary, Motooka et al 2) reported an opposite behavior, based on the MD simulations using 4096 Si atoms interacting via the Tersoff potential: 2.7 eV at the lower temperature region (1450-1600 K) and 1.2 eV at the higher temperature region (1600-2000 K).…”

Impact of supercooled liquid structures on the crystallization processes of amorphous Ge

“…Previous MD simulations were performed under an isochoric (constant volume) condition, [18][19][20][21] but the internal pressure due to volume change associated with a solid-toliquid phase transformation may affect the growth speed. [22][23][24] Therefore, the simulations were performed under an isobaric condition with a time step of 2 fs, and the temperature was controlled by using the Langevin equation. Various empirical interatomic potentials are proposed for group-IV elements.…”