Surface melting of superheated crystals. Atomistic simulation study

https://doi.org/10.1016/j.cpc.2007.02.073Get rights and content

Abstract

Melting front velocity dependencies on temperature are calculated using the molecular-dynamics method for the EAM models of Al and Fe as well as for the Lennard-Jones system. Different surface orientations are considered. It is shown that the Broughton–Gilmer–Jackson theory of the collision-limited growth can describe the results obtained. The isochoric bulk solid melting and decay under ultrafast heating is simulated for mono- and polycrystalline models.

Introduction

Melting mechanisms and kinetics is a long-standing topic of interest for atomistic simulation. Since the pioneering work of Broughton and Woodcock [1] where the premelting effects on the (100) Lennard-Jones crystal surface were studied, the theory of equilibrium surface properties near the bulk melting point has been well established (e.g., see the review [2]). In a number of works the kinetics of the crystal–liquid transition was studied primarily with regard to the solidification kinetics (e.g., [3], [4], [5] and references therein). Less attention was addressed to the melting kinetics itself: e.g., in [6] the velocity of the melting front propagation was calculated at different temperatures, and the similarity between the open surface melting and the grain boundary melting was shown.

A relatively larger attention to the solidification kinetics is caused by its obvious technological importance for crystal growth applications [7]. At the same time the atomistic-level melting kinetics and the solid superheating effects are usually considered to be negligible for the experiments where melting takes place. Local melting propagates from the open surfaces, grain boundaries and defects of the crystal structure. Taking into account the polycrystalline structure of usual solids it is accepted that the local melting of the substance takes place as soon as the local temperature equals the equilibrium melting temperature.

However under conditions of recent experiments connected with ultrafast high energy deposition at the nanosecond time scale (electrical explosions of wires, laser heating, shock waves) it is expected that the superheated solid phase could be one of the transient states during the solid phase evolution [8] and the melting kinetics is to be taken into account.

In Section 2 of the paper we describe the results on melting front velocity dependencies on temperature for the Lennard-Jones (LJ) solid and Al and Fe modeled by EAM potentials and make some general conclusions on the melting kinetics mechanisms. In Section 3 we compare the results of the atomistic simulation of the isochoric bulk solid melting and decay under ultrafast heating for monocrystalline and polycrystalline solids. We use these results in order to illustrate the possibility of solid superheating under special conditions.

Section snippets

Melting front propagation kinetics

For molecular dynamics simulation of the melting front propagation from the free surface of a superheated crystal we use simulation boxes that are elongated in the z direction and have square cross-section in the xy plane. Initially the simulation box is filled by atoms on the crystal lattice with the specific orientations (100), (110) or (111). At each temperature the lattice constant value is chosen from the zero stress condition. In order to prepare the initial configuration the crystal

Melting of mono- and polycrystalline solids under ultrafast isochoric heating

The molecular-dynamics model for the simulation of the bulk melting uses a cubic simulation box in 3D periodic boundary conditions. The EAM potential model for Cu [13] was used. The initial state corresponds to the equilibrated zero stress single crystal or the polycrystalline structure. The latter was created by dividing the simulation box into a number of Voronoi polyhedra and filling them with random orientations of the f.c.c. lattice (removing the overlapping atoms). In both cases the

Acknowledgements

This work was partially supported by Sandia National Laboratories under the U.S. DOE/NNSA Advanced Simulation and Computing program, the Ministry of Education and Science of RF (project RNP.2.1.1.712) and the RFBR (grant 04-02-17065). A.Yu.K. and V.V.S. gratefully acknowledge the support of the Dynasty Foundation.

References (14)

  • U. Tartaglino et al.

    Phys. Reports

    (2005)
  • J.J. Hoyt et al.

    Mat. Sci. Eng. R

    (2003)
  • S.J. Plimpton

    J. Comp. Phys.

    (1995)
  • V.V. Stegailov

    Comput. Phys. Comm.

    (2005)
  • J.Q. Broughton et al.

    J. Phys. C: Solid State Phys.

    (1978)
  • F. Celestini et al.

    Phys. Rev. E

    (2002)
  • H.L. Tepper et al.

    J. Chem. Phys.

    (2001)
There are more references available in the full text version of this article.

Cited by (0)

View full text