Atomistic study of self-diffusion in Ni, Al and Ni3Al

J. Duan, Yu N. Osetsky, D. J. Bacon

Research output: Contribution to journalArticlepeer-review

6 Scopus citations

Abstract

Atomic-scale computer simulation, both molecular dynamics (MD) and statics (MS), has been applied to study self-diffusion processes in pure Ni, Al and Ni3Al alloy of stoichiometric composition over a temperature range from 0.85 to 0.94 Tm ( Tm is the corresponding melting temperature). A wide set of diffusional characteristics, including the self-diffusion coefficient, jump frequency, correlation factor and activation energy, was obtained by MD for each material Self-diffusion in pure metals occurs via random walk of vacancies and the tracer correlation factor obtained from MD is in a good agreement with that estimated from the theory. In the ordered alloy the diffusion mechanism is more complicated. Thus, Ni diffuses mainly via Ni sites, whereas the diffusion of Al atoms includes formation and annihilation of antisite defects. Once a Ni antisite is formed, Al diffusion occurs via the intrasublattice (ISL) mechanism. At the high temperatures the antistructural bridge (ASB) mechanism is also observed. The conclusions of the MD study are supported by the MS simulation, in which the activation energy was obtained along paths corresponding to possible diffusion mechanisms, namely the six-jumps cycle (SJC), ISL and ASB mechanisms. It was found that the ISL and ASB mechanisms are much more favourable for Al diffusion than the SJC one.

Original languageEnglish
Pages (from-to)423-428
Number of pages6
JournalDefect and Diffusion Forum
Issue number194-199 PART 1
DOIs
StatePublished - 2001
Externally publishedYes

Keywords

  • Antistructural Bridge Mechanism
  • Atomistic Simulation
  • Correlation Factor
  • Intermetallic Compound
  • Intrasublattice Mechanism
  • Ordered Alloy
  • Self-Diffusion
  • Six-Jump-Cycle

Fingerprint

Dive into the research topics of 'Atomistic study of self-diffusion in Ni, Al and Ni3Al'. Together they form a unique fingerprint.

Cite this