Abstract
Nickel-based superalloys are high performance structural materials that exhibitexcellent strength and creep resistance at high temperatures, even in
chemically aggressive environments. This makes them ideal for use in the
construction of ecient turbines for energy generation or aerospace applications.
These superalloys are usually manufactured as single crystals, with
their high strength resulting from dislocation pinning at interfaces between
the fcc matrix an L12-ordered precipitates. Chemical impurities such as
rhenium also aect dislocation mobility, and their inclusion in commercial
materials is standard practice. However, there is currently no detailed understanding
of the atomic-scale mechanisms underlying these processes. This
problem is here addressed at the atomistic level. The typical accuracy level of
rst principle methods is required to describe bond breaking in the distorted
region surrounding a dislocation core and for including impurities in the fcc
matrix, but model systems must be large enough to accommodate the strain
gradients typical of long-range elastic interactions due to the presence of
dislocations. Multiscale methods are therefore required for simulating these
chemo-mechanical processes. The `Learn on the Fly' (LOTF) technique is a
non-uniform precision quantum mechanical/molecular mechanical approach.
It oers a predictor/corrector algorithm for speeding up calculations, and the
possibility of modelling a moving quantum region, useful for fast dislocation
motion due to high simulating temperature or load condition. The scope of
this thesis is to apply this method to metallic systems to conduct, for the rst
time, quantum mechanical accurate simulations of dislocation motion in Nibased
alloys. The QM/MM method corrects the deciencies of the classical
interatomic potential related to inaccurate energetics for the hcp phase, relevant
to the geometry of dislocation cores, and it is capable of reproducing the
correct separation between Shockley partials at high temperature conditions.
| Date of Award | 2016 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Nicola Bonini (Supervisor), James Kermode (Supervisor) & Alessandro De Vita (Supervisor) |