J. P. DeVita,L. M. Sander,P. Smereka
Abstract:We present a fast Monte-Carlo algorithm for simulating epitaxial surface growth, based on the continuous-time Monte-Carlo algorithm of Bortz, Kalos and Lebowitz. When simulating realistic growth regimes, much computational time is consumed by the relatively fast dynamics of the adatoms. Continuum and continuum-discrete hybrid methods have been developed to approach this issue; however in many situations, the density of adatoms is too low to efficiently and accurately simulate as a continuum. To solve the problem of fast adatom dynamics, we allow adatoms to take larger steps, effectively reducing the number of transitions required. We achieve nearly a factor of ten speed up, for growth at moderate temperatures and large D/F.
What problem does this paper attempt to address?
This paper attempts to solve the computational efficiency problems encountered in the process of simulating epitaxial growth, especially when the kinetic process of surface adatoms is very fast, resulting in overly long computational time. Although the traditional kinetic Monte Carlo (KMC) method is accurate and flexible, it is relatively slow in some cases, especially at high temperatures or when the adatom diffusion rate is much greater than the island - atom detachment rate. In such cases, most of the computational resources are used to calculate the trajectories of adatoms.
### Core Problems of the Paper
1. **The Problem of Rapid Adatom Kinetics**:
- Under actual growth conditions, the diffusion rate of adatoms is much higher than other processes (such as the detachment of island atoms), which makes most of the time in KMC simulations spent on calculating the movement of adatoms.
- In order to improve computational efficiency, the paper proposes a multi - scale kinetic Monte Carlo (MSKMC) algorithm, which allows adatoms to jump multiple step lengths at once, thereby reducing the total number of jumps and computational time.
2. **The Problem of Simulating Low - Density Adatoms**:
- When the density of adatoms is low, continuum models cannot efficiently and accurately simulate these processes. The MSKMC method proposed in the paper can improve computational efficiency while maintaining the accuracy of macroscopic and mesoscopic kinetics by adjusting the microscopic kinetics.
### Solutions
- **MSKMC Algorithm**: This algorithm significantly improves the computational speed by reducing the total number of jumps through allowing adatoms to make larger - step - length jumps. Specifically, the jump step length of adatoms changes from the original single - nearest - neighbor jump to being able to jump multiple step lengths (for example, \( n_d \) steps), and the jump rate is adjusted accordingly to \( w_0'=\frac{1}{n_d} \).
- **Local Search Strategy**: In order to prevent adatoms from skipping over obstacles (such as steps, attachment sites or other adatoms) during the jumping process, the paper introduces a local search strategy to ensure that adatoms do not skip important structural features.
### Result Verification
The paper verifies the effectiveness and accuracy of the new method by comparing the results of MSKMC and the traditional KMC method. The results show that MSKMC can not only correctly simulate various phenomena in epitaxial growth (such as island shapes, island size distributions in the early growth stage, surface roughening, etc.), but also, under certain physical conditions, its computational speed is 5 to 10 times faster than that of the traditional KMC method.
### Summary
The main contribution of this paper lies in proposing an efficient MSKMC algorithm, which solves the computational efficiency problems encountered by the traditional KMC method in simulating the epitaxial growth process, especially the simulation difficulties in the case of rapid adatom kinetics and low - density conditions.