Analysis of high-throughput materials data by means of computation and simulation
XiaoGang Lu
1School of Materials Science and Engineering, Shanghai University, Shanghai, China
2Materials genome institute, Shanghai University, Shanghai, China
EXTENDED ABSTRACT: High-throughput data refer to high-throughput experimental and computational data. Combinatorial material chip and diffusion multiple are two representative high-throughput experimental techniques to produce a large quantity of materials data. On the other hand, high-throughput computation, based on physical theories and high-performance computer clusters, is able to obtain data that may be difficult from experiments. In order to boost materials design and development, high-throughput data need to be analyzed by means of computation and simulation. A case study is the analysis of diffusion data. Traditionally diffusion coefficients or atomic mobilities are extracted from composition variations measured by diffusion couple experiments, which is considered as one dimensional (1D) method. However, one more step to assemble a diffusion triple produces two-dimensional (2D) composition distributions, which are not adopted to study diffusion data. In the present work, we have developed an efficient numerical inverse approach to directly evaluate composition-dependent atomic mobilities and full matrix of interdiffusivities based on a novel 2D diffusion simulation scheme by analyzing only one or two
diffusion triple over a wide composition range.
Professor Xiao-Gang Lu received his master degree in materials science from Central South University in China in 1995. Since 1999, he studied in Prof. Bo Sundman’s group in the Royal Institute of Technology (KTH), Sweden and received the PhD degree in computational thermodynamics in 2005. He then joined the Thermo-Calc Company in Stockholm and worked for 5 years before he moved to Shanghai, China. He is now a full professor in the school of materials science and engineering in Shanghai University. His research interests are computational thermodynamics and computational kinetics including diffusion controlled phase transformations and precipitation simulation, modeling of thermodynamic and thermo-physical properties at ambient and high pressures. Currently he is involved in several projects aiming to establish thermodynamic, diffusion mobility and physical property databases for steel and Ni-Co-base alloy. He is also dedicated to integrated multi-scale computations and the materials genome projects in China.