Publication | Closed Access
Stochastic equations of motion for epitaxial growth
192
Citations
67
References
1993
Year
EngineeringAnalytic DerivationStochastic PhenomenonLangevin EquationsNumerical SimulationNanoscale ModelingTransport PhenomenaAnomalous DiffusionMaterials SciencePhysicsStochastic Dynamical SystemBrownian MotionStochastic Differential EquationNatural SciencesSurface ScienceApplied PhysicsDiffusion ProcessStochastic EquationsRegularization SchemeContinuum ModelingMultiscale Modeling
We report an analytic derivation of the Langevin equations of motion for the surface of a solid that evolves under typical epitaxial-growth conditions. Our treatment begins with a master-equation description of the microscopic dynamics of a solid-on-solid model and presumes that all surface processes obey Arrhenius-type rate laws. Our basic model takes account of atomic deposition from a low-density vapor, thermal desorption, and surface diffusion. Refinements to the model include the effects of hot-atom knockout processes and asymmetric energy barriers near step edges. A regularization scheme is described that permits a (nonrigorous) passage to the continuum limit when the surface is rough. The resulting stochastic differential equation for the surface-height profile generically leads to the behavior at long length and time scales first described by Kardar, Parisi, and Zhang [Phys. Rev. Lett. 56, 889 (1986)] (due to desorption). If evaporation is negligible, the asymptotic behavior is characteristic of a linear model introduced by Edwards and Wilkinson [Proc. R. Soc. London, Ser. A 381, 17 (1982)] (due to asymmetric step barriers and/or knockout events). If the latter are absent as well, the surface roughness is determined by an equation independently analyzed by Villain [J. Phys. I 1, 19 (1991)] and Lai and Das Sarma [Phys. Rev. Lett. 66, 2348 (1991)] (which includes only deposition and site-to-site hopping). The consequences of reflection-symmetry breaking in the basic microscopic processes are discussed in connection with step-barrier asymmetry and Metropolis kinetic algorithms.
| Year | Citations | |
|---|---|---|
Page 1
Page 1