PHY480 Report 2

 

The PHY480 Computational Physics course offered in Spring 2006 provided familiarization to basic UNIX commands and also introduced the students to the Fortran programming language, where we learnt how to employ Fortran to evaluate differential equations and integrals using various numerical techniques. At the later stage of the course, emphasis is placed on molecular dynamics of Lennard-Jones clusters and the determination of vibrational eigen-modes using both matrix approach and fast-fourier transform algorithm.

 

I hope to apply my basic experience in molecular dynamics to perform contact melting simulations between NiTi (intermetallic phase) and pure Nb. The great interest in this subject is attributed to the ability of the above-mentioned 2 materials to initiate melting at 1170°C upon contact. This process can eventually mediate metallurgical bonds between 2 separate NiTi components. It is worthy to highlight that this contact melting temperature is about 200°C and 1300°C below the melting points of NiTi and Nb, respectively. From classical thermodynamic considerations, substantial diffusion across the contact interface has to occur for “impurity effect” to cause interfacial melting. However, classical theory may not be valid at the melting interface due to surface effects.

 

This opens up some scientific questions to be addressed: (1) Whether contact melting can actually occur without mutual migration of species across the interface? (2) Whether the crystal lattices are destroyed on contact before any migration process? (3) Which diffusing species is the rate-limiting factor in the contact melting process? (4) Is there a “latency” time before the melting occurs and what governs it?

 

To address the questions the following concepts and approaches to the molecular dynamics simulation are devised:

 

(1)               A standard Molecular Dynamics method will be modified for contact modelling. The interatomic potentials used may be derived from the Embedding Atom Method (EAM) or from the Finnis-Sinclair Type. These potentials are known to have inner consistency that allows for the treatment of ternary systems.

 

(2)               Division of a calculation cell into 2 halves with a contact border. A simulation is performed to account for any statistical characteristic changes for particles in each half.

 

(3)               Monitoring features of the dynamics of the particles and the stipulated availability of a contact border. Comparing the characteristics of a system when the border initially exists and after its disappearance due to inter-diffusion of particles).

 

(4)               Monitoring any variations of the crystal lattices until different states exists from the contact border.