NERSC Initiative for Scientific Exploration (NISE) 2011 Awards

Ab initio simulation of mechanical properties of bulk metallic glass and Mo-bases alloys

Wai-Yim Ching, University of Missouri - Kansas City

Associated NERSC Project: Electronic Structures and Properties of Complex Ceramic Crystals and Novel Materials (mp250)

This research is aimed at providing a new way to characterize the mechanical properties of materials (simple or complex). While there are several experimental techniques such as nano-indentation to characterize the strength of materials, it is not precise and requires many parameters which are difficult to obtain. The theoretical construction of the failure envelope based on accurate ab initio simulations will have a large impact in this critical area related to energy science and technology. Petascale supercomputing facility is the only way to obtain the required data. The method and procedures developed and the knowledge generated in this project can be further expanded and applied to other areas of materials design and development.

We propose to develop and implement a precise method to characterize the strength of a material by constructing a three-dimensional failure envelope in the stress space (σxx, σyy, σzz) using data generated from multi-axial tensile experiments on supercomputer. The “experiments” are performed on the supercell of a crystal by applying successive tensile strains in small steps until the stress reaches the maximum which is defined as the failure point (yield point). At each step, the structure is fully relaxed using VASP with high accuracy. The stress data σij corresponding to the applied strain in a given direction at each step are extracted. The “experiment” for each crystal is carried out in as many directions as feasible and the final set of stress (σxx, σyy, σzz) and strain (εxx, εyy, εzz) data at the failure points are collected to construct a failure envelope in the three dimensional stress space. A color code is used to represent the value of |σ| which depicts the specific locations of the weak and strong points in the first quadrant of the stress space. The total area of the envelope or the volume it encloses constitutes a single parameter representing the average strength of a crystal under tensile deformation. The shape and the color of the envelope delineate the variations of the strength in different directions. This technique has been demonstrated by us in a complex bioceramics crystal hydroxyapatite.

We will extend our study to two Mo-based alloys Mo5Si3 and Mo5SiB2 which have some very unique mechanical properties, and to a new class of noncrystalline material, the bulk metallic glasses (BMG). The Mo-based alloys within the Mo-Si-B phases diagram are targeted to replace Ni-based super-alloys for applications in fossil energy technology where the mechanical properties at elevated temperature and in the presence of microstructures play a critical role. BMG is class of noncrystalline materials with some outstanding properties such as high strength, high elastic limit, high corrosion resistance etc. and have wide applications in industry. However, it is also well known that BMG is hindered by its lack of plasticity. It is believed that this is intimately related to the formation of the shear transition zone due to the movements of free volumes in BMG. Understanding the origin of this behavior in BMG required atomistic level information which can only be obtained by accurate large-scale simulations on models of BMG.

In all, we plan to construct the failure envelopes of four supercell models, one each for crystalline phases of Mo5Si3 and Mo5SiB2, one for a BMG model Cu50Zr50 (50% Cu and 50% Zr) and a supercell model for crystalline CuZr to which the results of BMG model Cu50Zr50 will be compared. Each supercell is estimated to be of the size of 256 atoms. Previous simulations on Ti3AlC2 and Cr2AlC indicated that at least 240 data points corresponding to tensile experiments along 240 directions are needed. Once these pilot projects are completed, additional modeling on larger multi-component BMG of different compositions and constituent elements will be planned.