% IMPORTANT: The following is UTF-8 encoded. This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.
@PHDTHESIS{Klppelberg:283574,
author = {Klüppelberg, Daniel Aaron},
title = {{F}irst-principle investigation of displacive response in
complex solids},
volume = {119},
school = {RWTH Aachen},
type = {Dr.},
address = {Jülich},
publisher = {Forschungszentrum Jülich GmbH Zentralbibliothek, Verlag},
reportid = {FZJ-2016-01888},
isbn = {978-3-95806-123-1},
series = {Schriften des Forschungszentrums Jülich. Reihe
Schlüsseltechnologien / Key Technologies},
pages = {XI, 177 S.},
year = {2016},
note = {RWTH Aachen, Diss., 2015},
abstract = {In this work, we discuss two approaches to calculate phonon
spectra of crystals within the all-electron full-potential
linearized augmented-plane-wave (FLAPW) method. This method
is one of the most precise implementations of Kohn-Sham (KS)
density functional theory(DFT) due to the inclusion of all
electrons into the calculation and the use of the full
potential, i.e., no shape approximations are applied to the
potential. The calculation of phonons requires the
force-constant matrix (FCM). The FCM is the second-order
derivative of the KS total energy with respect to two atomic
displacements. Its Fourier transform yields the dynamical
matrix (DM). The eigenvalues of the DM are thesquares of the
phonon frequencies. Its eigenvectors are the polarization
vectors. The first approach to calculate phonons is the
nite-displacement (FD) method. In this method, the FCM is
obtained from displacing one atom at a time from
equilibrium, calculating the forces on all atoms, and
dividing by the displacement amplitude. This is repeated for
each atom and for each spatial direction. In practice, the
number of calculations reduces signicantly by exploiting the
symmetry of the crystal lattice. The FCM is transformed to
the DM and the phonon energies and polarization vectors are
extracted. A drawback of this approach is given by the
necessity to use supercells. The phonon frequencies are only
correct for phonons whose wave vector $\textit{q}$ is
commensurable with the lattice. Hence, to correctly
calculate phonon frequencies at small wave vectors, large
super cells are needed, because the displacement pattern of
such phonons repeats only after many instances of the
primitive unit cell. Since the FD procedure relies on an
analytical derivation of the total energy followed by a
numerical one, precise forces are necessary. Otherwise, the
FCM is not symmetric, forexample. We present a reformulation
of the FLAPW force formalism which includes the whole unit
cell into the calculation of the atomic force contribution
from the core states and which incorporates additional terms
to deal with the slight discontinuity of the LAPW basis
functions and the quantities derived from them. The
improvement of the force precision is demonstrated by the
study of different criteria. We then present phonon spectra
for Al, MgO, GaAs, and EuTiO$_{3}$ obtained by the FD method
from forces calculated in the FLAPW approach using our
reformulation. The second approach to calculate phonon
spectra is density functional perturbation theory(DFPT). In
DFPT, the second-order derivative of the KS total energy is
directly calculated via perturbation theory. DFPT allows the
determination of phonon frequencies at arbitrary wave
vectors $\textit{q}$ from calculations involving the
primitive unit cell, only, by treating a phonon of this wave
vector as the perturbation. The first-order changes of the
basis functions, the electronic density, and the potential
have to be obtained from the Sternheimer equation, which is
the linearized Schrodinger equation. Additionally, the
second-order changes of the external potential, the ion-ion
energy, and the LAPW basis functions are required. We
provide formulas which explicitly include adjustments of the
general DFPT approach when used in conjunction with the
FLAPW method. These adjustments include Pulay terms, the
correct treatment of the core state contribution, and
surface terms which are analogous to those within the force
formalism.},
cin = {PGI-1 / IAS-1},
cid = {I:(DE-Juel1)PGI-1-20110106 / I:(DE-Juel1)IAS-1-20090406},
pnm = {142 - Controlling Spin-Based Phenomena (POF3-142) / 143 -
Controlling Configuration-Based Phenomena (POF3-143)},
pid = {G:(DE-HGF)POF3-142 / G:(DE-HGF)POF3-143},
typ = {PUB:(DE-HGF)11 / PUB:(DE-HGF)3},
url = {https://juser.fz-juelich.de/record/283574},
}