Inverse processing-structure relation for the nucleation and growth mechanism
- PDF / 461,904 Bytes
- 6 Pages / 612 x 792 pts (letter) Page_size
- 40 Downloads / 263 Views
Inverse processing-structure relation for the nucleation and growth mechanism Mark Jhon1, Yang Hao Lau1, Siu Sin Quek1, and David T. Wu1 1 Institute of High Performance Computing, A*STAR 1 Fusionopolis Way, #16-16 Connexis Singapore, 138632, Singapore
ABSTRACT The formation of realistic, polycrystalline microstructures can be simulated by modeling the kinetics of nucleation and growth. However, it is difficult to perform the inverse simulation, where details of the nucleation and growth process are inferred from geometric properties of the final microstructure. In the present study, we develop a methodology for solving the inverse problem for interface-limited growth in 1D, utilizing a reverse Monte Carlo (RMC) algorithm. The algorithm produces a time dependent nucleation rate that gives a grain size distribution closest to a target distribution. Its results may be used to understand the limitations of manipulating the grain boundary distributions through temperature alone. INTRODUCTION The properties of materials can be optimized through careful engineering of their grain boundary structure. The relationships between property and structure are affected not only by the mean of the grain size, but also by the shape of the grain size distribution. For example, by narrowing the initial particle size distribution of sintered materials, grain coarsening can be inhibited, increasing thermal stability of the structure [1]. Likewise, mechanical properties are sensitive both to the shape as well as the average of the grain size distribution. Refining the average grain size of a metal can cause improvements in strength at the expense of ductility [24]. However, by making the size distribution bimodal [5], or by producing unique ultrafinegrained microstructure from severe plastic deformation [6], it is possible to make a nanocrystalline metal simultaneously strong and ductile. The foregoing examples illustrate the importance of controlling grain size distributions. However, it also highlights the difficulty of tuning grain size distributions. The process required to develop the bimodal grain distributions in [5] is complicated and involves both thermal and mechanical processing. This raises the following question: how much can size distributions be manipulated by controlling temperature alone? In the present study, we consider the simplest version of this problem, assuming a 1D geometry and interface limited growth kinetics. We propose a simple computational methodology to calculate the time-dependent nucleation rate most consistent with a given grain size distribution. Our algorithm is based on the reverse Monte Carlo algorithm (RMC), a technique first developed to generate atomic positions of disordered systems consistent with an experimentally determined structure factor and radial distribution function [7]. RMC has since been extended to include a wide variety of experimental data as inputs to the fitting process (as reviewed in [8]).
The RMC has been applied to several microstructural problems [9-11]: Tong et al. used
Data Loading...
