Efterpi Kalesaki

Topological states in multi-orbital HgTe honeycomb lattices

Research on graphene has revealed remarkable phenomena arising in the honeycomb lattice. However, the quantum spin Hall effect predicted at the K point could not be observed in graphene and other honeycomb structures of light elements due to an insufficiently strong spin–orbit coupling. Here we show theoretically that 2D honeycomb lattices of HgTe can combine the effects of the honeycomb geometry and strong spin–orbit coupling. The conduction bands, experimentally accessible via doping, can be described by a tight-binding lattice model as in graphene, but including multi-orbital degrees of freedom and spin–orbit coupling. This results in very large topological gaps (up to 35 meV) and a flattened band detached from the others. Owing to this flat band and the sizable Coulomb interaction, honeycomb structures of HgTe constitute a promising platform for the observation of a fractional Chern insulator or a fractional quantum spin Hall phase.

Morphology and strain of self-assembled semipolar GaN quantum dots in (112¯2) AlN

GaN quantum dots (QDs) grown in semipolar (112¯2) AlN by plasma-assisted molecular-beam epitaxy were studied by transmission electron microscopy (TEM) and scanning transmission electron microscopy techniques. The embedded (112¯2)-grown QDs exhibited pyramidal or truncated-pyramidal morphology consistent with the symmetry of the nucleating plane, and were delimited by nonpolar and semipolar nanofacets. It was also found that, in addition to the (112¯2) surface, QDs nucleated at depressions comprising {101¯1} facets. This was justified by ab initio density functional theory calculations showing that such GaN/AlN facets are of lower energy compared to (112¯2). Based on quantitative high-resolution TEM strain measurements, the three-dimensional QD strain state was analyzed using finite-element simulations. The internal electrostatic field was then estimated, showing small potential drop along the growth direction, and limited localization at most QD interfaces.

Effect of edge threading dislocations on the electronic structure of InN

The open issue of the n-type conductivity and its correlation to threading dislocations (TDs) in InN is addressed through first principles calculations on the electronic properties of a-edge TDs. All possible dislocation core models are considered (4-, 5/7-, and 8-atom cores) and are found to modify the band structure of InN in a distinct manner. In particular, nitrogen and indium low coordinated atoms in the eight-atom core induce states near the valence band maximum and above the conduction band minimum, respectively. The formation of a nitrogen–nitrogen “wrong” bond is observed at the 5/7-atom core resulting in a state inside the band gap. The 4- and 5/7-atom cores induce occupied states resonant in the conduction band due to In–In strain induced interactions and wrong bonds, respectively. These occupied states designate TDs as a source of higher electron concentrations in InN and provide direct evidence that TDs contribute to its inherent n-type conductivity.

Electronic structure of 1/6〈202¯3〉 partial dislocations in wurtzite GaN

The I1 intrinsic basal stacking faults (BSFs) are acknowledged as the principal defects observed on {112¯0} (a-plane) and {11¯00} (m-plane) grown GaN. Their importance is established by recent experimental results, which correlate the partial dislocations (PDs) bounding I1 BSFs to the luminescence characteristics of GaN. PDs are also found to play a critical role in the alleviation of misfit strain in hetero-epitaxially grown nonpolar and semipolar films. In the present study, the energetics and the electronic structure of twelve edge and mixed 1/6〈202¯3〉 PD configurations are investigated by first principles calculations. The specific PD cores of the dislocation loop bounding the I1 BSF are identified for III-rich and N-rich growth conditions. The core structures of PDs induce multiple shallow and deep states, attributed to the low coordinated core atoms, indicating that the cores are electrically active. In contrast to edge type threading dislocations no strain induced states are found.

Screw threading dislocations in AlN: Structural and electronic properties of In and O doped material

Density functional theory calculations were performed on undoped AlN screw threading dislocations (TDs) as well as TDs doped by indium and oxygen, prompted by integrated experiments through transmission electron microscopy and spectroscopic techniques demonstrating enhanced In and O concentrations in screw dislocation cores. It is revealed that screw TDs act as conduction pathways to charge carriers, introducing multiple levels in the bandgap due to overstrained, dangling, and “wrong” bonds formed even in the undoped cores. The presence of impurities and especially metallic In elevates the metal-like electronic structure of the distorted material and promotes the conductivity along the dislocation line. Hence screw dislocations in AlN are established as highly prominent conductive nanowires in semiconducting thin films and prospects for novel, highly functional nano-device materials through exploitation of screw TDs are attested.

Reconstructions and electronic structure of (11-22) and (11-2-2) semipolar AlN surfaces

Τhe energetics, atomic geometry, and electronic structure of semipolar (112¯2) and (112¯2¯) AlN surfaces are investigated employing first principles calculations. For metal-rich growth conditions, metallic reconstructions are favoured on both polarity surfaces. For N rich to moderate Al rich conditions, the (112¯2) planes promote semiconducting reconstructions having 2 × 2 or c(2 × 2) periodicity. In contrast, under the particular range of the Al chemical potential the (112¯2¯) surfaces stabilize reconstructions with excess metal and it is only at the extreme N rich limit that the semiconducting c(2 × 2) N adatom structure prevails. The present study reveals that the reconstructed (112¯2) surfaces do not contain steps in contrast to (112¯2¯) where surface steps are inherent for N rich to moderate metal rich growth conditions and may result in intrinsic step-flow growth and/or growth of parasitic semipolar orientations.

Atomic scale morphology, growth behaviour and electronic properties of semipolar

First-principles calculations relating to the atomic structure and electronic properties of {101[overline]3} GaN surfaces reveal significant differentiations between the two polarity orientations. The (101[overline]3) surface exhibits a remarkable morphological stability, stabilizing a metallic structure (Ga adlayer) over the entire range of the Ga chemical potential. In contrast, the semiconducting, cleaved surface is favoured on (101[overline]3[overline]) under extremely and moderately N-rich conditions, a Ga bilayer is stabilized under corresponding Ga-rich conditions and various transitions between metallic reconstructions take place in intermediate growth stoichiometries. Efficient growth schemes for smooth, two-dimensional GaN layers and the isolation of {101[overline]3} material from parasitic orientations are identified.

\{ 1 0\bar {1}3\}

Observations of easy transition between nonpolar and semipolar orientations during III-Nitride heteroepitaxy identify the 90o 〈1¯21¯0〉 rotation relationship as being very important in defining this coexistence. A rigorous analysis of this relationship using the topological theory of interfaces showed that it leads to a high order of coincident symmetry and makes energetically favorable the appearance of the intergranular boundaries. Principal low-energy boundaries, that could also be technologically exploited, have been identified by high-resolution transmission electron microscopy (HRTEM) observations and have been studied energetically using empirical potential calculations. It is also shown that these boundaries can change their average orientation by incorporating disconnections. The pertinent strain relaxation mechanisms can cause such boundaries to act as sources of threading dislocations and stacking faults. The energetically favorable (101¯0) || (0001) boundary was frequently observed to delimit m...

GaN surfaces

Air annealing of chalcopyrite solar cells at either 200 °C or higher is often known to increase their power conversion efficiency. In this paper, we investigate the nature of this effect for Cu2SnS3 (CTS) solar cells by modeling the experimental external quantum efficiency. We find that the cell efficiency increase stems from increased diffusion length and depletion width and decreased interface recombination at the p-n junction. The increased diffusion length is also reproduced when only the absorber layer is air annealed. When solar cells are annealed in N2, no increase in diffusion length is measured. Hence, we attribute the increase in diffusion length to passivation of the grain boundaries in the bulk by oxygen. The larger depletion width on air and N2 annealing in the devices is independent of the CdS buffer layer thickness and occurs in its absence. We ascribe it to copper diffusion from the absorber layer to the n-type buffer and window layers. Interface recombination positively correlates with increasing buffer layer thickness. Based on our modeling, we conclude that the CTS absorber layer is still too highly doped to obtain large depletion widths and is highly recombination active at the p-n interface.

Interfaces between nonpolar and semipolar III-nitride semiconductor orientations: Structure and defects

The interest in 2-dimensional systems with a honeycomb lattice and related Dirac-type electronic bands has exceeded the prototype graphene1. Currently, 2-dimensional atomic2,3 and nanoscale4-8 systems are extensively investigated in the search for materials with novel electronic properties that can be tailored by geometry. The immediate question that arises is how to fabricate 2-D semiconductors that have a honeycomb nanogeometry, and as a consequence of that, display a Dirac-type band structure? Here, we show that atomically coherent honeycomb superlattices of rocksalt (PbSe, PbTe) and zincblende (CdSe, CdTe) semiconductors can be obtained by nanocrystal self-assembly and facet-to-facet atomic bonding, and subsequent cation exchange. We present a extended structural analysis of atomically coherent 2-D honeycomb structures that were recently obtained with self-assembly and facet-to-facet bonding9. We show that this process may in principle lead to three different types of honeycomb structures, one with a graphene type-, and two others with a silicene-type structure. Using TEM, electron diffraction, STM and GISAXS it is convincingly shown that the structures are from the silicene-type. In the second part of this work, we describe the electronic structure of graphene-type and silicene type honeycomb semiconductors. We present the results of advanced electronic structure calculations using the sp3d5s* atomistic tight-binding method10. For simplicity, we focus on semiconductors with a simple and single conduction band for the native bulk semiconductor. When the 3-D geometry is changed into 2-D honeycomb, a conduction band structure transformation to two types of Dirac cones, one for S- and one for P-orbitals, is observed. The width of the bands depends on the honeycomb period and the coupling between the nanocrystals. Furthermore, there is a dispersionless P-orbital band, which also forms a landmark of the honeycomb structure. The effects of considerable intrinsic spin-orbit coupling are briefly considered. For heavy-element compounds such as CdTe, strong intrinsic spin-‐orbit coupling opens a non-trivial gap at the P-orbital Dirac point, leading to a quantum Spin Hall effect10-12. Our work shows that well known semiconductor crystals, known for centuries, can lead to systems with entirely new electronic properties, by the simple action of nanogeometry. It can be foreseen that such structures will play a key role in future opto-electronic applications, provided that they can be fabricated in a straightforward way.