% 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{Schfer:958928,
author = {Schäfer, Tobias},
othercontributors = {Wuttig, Matthias and Siegrist, Theo},
title = {{T}uning charge transport in crystalline phase-change
materials},
school = {RWTH Aachen University},
type = {Dissertation},
address = {Aachen},
publisher = {RWTH Aachen University},
reportid = {RWTH-2023-05483},
pages = {1 Online-Ressource : Illustrationen, Diagramme},
year = {2023},
note = {Veröffentlicht auf dem Publikationsserver der RWTH Aachen
University; Dissertation, RWTH Aachen University, 2023},
abstract = {Today, many objects in our daily lives as well as most of
our society’s critical infrastructure depend on computing
devices—starting from high-performance supercomputers down
to low-performance electronic switches and the upcoming
internet of things. While in past decades miniaturization
allowed to meet the rising demand („Moore’s law“), in
near future confinement and quantum-mechanics will impede
further down-scaling. In the meantime, energy efficiency is
becoming a more and more important topic. As a rising demand
for more complicated calculations can be expected also for
the future, novel concepts are needed to initiate a
revolution in computing chip architecture. It is
foreseeable, that future computing will employ specialized
chips for different computing challenges instead of the
all-purpose chips in use today: Solid state memory can tear
down the separation between memory and computation and
combined units may allow for more efficient evaluation of
large amounts of data. Numerical calculations may be
performed on analog computers and neuromorphic computing
architectures may speed up pattern recognition. Beyond those
electronic computing concepts, spintronic computing may
further reduce the energy consumption and even allow for
quantum computers. The latter might become unparalleled in
performing tasks using the same algorithm on large sets of
input parameters. Those revolutions are fueled by the
development and availability of new materials: Non-volatile
data retention needs to be realized by an intrinsic
material’s parameter like the electrical resistance, but
this parameter needs to be tunable as well on extremely
short time-scales. Analog and neuromorphic calculations,
furthermore, call for gradual changes in resistance.
Spintronic computing might be realized by magnetic doping of
a semiconducting material, while topological insulators
might host states for quantum-computing. Remarkably,
Phase-change materials (PCMs) and related materials may
serve this multifaceted list of requirements. Especially the
fast writing combined with long data retention has already
been commercialized in CD-RWs and DVD-RWs and is thus
technically well-optimized. The electronic properties of
many PCMs are determined by defects, while the ideal
structure of the materials would result in small-bandgap
insulators. Materials like GeSb2Te4 consist of one
sub-lattice of the crystal hosting Te-atoms only, while
every fourth place on the other sub-lattice remains empty.
It has been assumed in previous works that the ordering of
these vacancies determines the mobility of the free charge
carriers, but small grain sizes in the crystal and a phase
transition did not allow to fully rule out competing
explanations. A collaboration within this thesis allowed to
perform similar mobility tuning experiments also in a series
of large-grain fully cubic samples that even feature an
insulator-to-metal transition, providing the missing pieces
of evidence. The transition from metallic to insulating
samples is often described as pure Anderson transition in
literature, while most other materials feature a
Mott-Anderson transition. This work agrees on the general
picture, but the finding of Efros-Shklovskii-hopping and the
direct proof of a Coulomb-gap requires to add small
contributions of correlation to the model of trnasport in
materials like SnSb2Te4. But vacancies do not only determine
the mobility of the charge carriers, but also their
number—only that here it is additional vacancies instead
of the aforementioned “stoichiometric” ones. One would
expect those additional vacancies to be tunable by the
cation-to-anion ratio, but experiments on Ge1-δTe1+δ and
(SnSb2)1-δ(Te4)1+δ show the performance and limitations of
this approach. Likewise, doping with foreign elements might
tune the number of carriers, but the overall performance is
limited due to a drastic reduction in mobility. Replacing
the antimony in SnSb2Te4 by bismuth in SnBi2Te4 allows for
much more efficient ways to change the number of charge
carriers, both by partial isovalent replacement
(Sn(Bi,Sb)2Te4) and by cation-to-anion variation in
(SnBi2)1-δ(Te4)1+δ. Even the type of carriers can be
changed from holes to electrons, while "normal" PCMs are
always hole conductors. The electronic properties of valence
and conduction band are remarkably similar. Furthermore, a
surface state might be present in Sn(Bi,Sb)2Te4, but the
remaining bulk conductivity as well as the chosen
measurement techniques do not allow to produce proof beyond
doubt on th etopological nature of this state. Simulations
show this different behavior of SnBi2Te4 compared to
"normal" PCMs to be caused by the interplay of a multitude
of defects that are possible and allow to tune the carrier
density. “Normal” PCMs, by contrast, feature only one
defect that is energetically possible and thus dominating
the properties of the material. The multitude of defects in
SnBi2Te4 does not only allow to tune the charge carrier
density by changing the composition, but also by heat
treatment, while normal materials mostly increase their
mobility upon annealing. It can be deduced as a general rule
for all here-investigated materials, that an increase in
carrier concentration reduces the mobility, which can be
reasoned by the doping with defects and vacancies as
predominant mechanism. All in all, Phase-change materials
may play an important role to enable the upcoming
revolutions in computing technologies. However, it needs to
be taken into account for the design of new materials, that
they cannot be doped like silicon and other well-known
semiconducting materials. Here doping needs to be performed
via the tuning of defect energetics and thus via changing
the self-doping properties of the material.},
cin = {131110 / 130000},
ddc = {530},
cid = {$I:(DE-82)131110_20140620$ / $I:(DE-82)130000_20140620$},
typ = {PUB:(DE-HGF)11},
doi = {10.18154/RWTH-2023-05483},
url = {https://publications.rwth-aachen.de/record/958928},
}
guest ::