Session L21: Semiconductors: Atomic Structure and Lattices

Bonding states of Cu atoms in superionic $\alpha$-CuI phase

The fast migration of cations in solids is used for solid state battery. Therefore the mechanism of the migration is of importance for the development new material. The superionic conductor CuI is zinc blend-type structure at low temperature and fluorite-type at high temperature. So the bonding between Cu and I atoms has been considered to be a covalent bonding. The peak positions and the asymmetrical peaks in the pair distribution functions between the Cu-Cu and Cu-I components have been remained to be explained in an experiment and a computational studies. We investigate the electronic states of CuI using the planewave based density functional calculation. We evaluate the charge and bonding states with Bader decomposition method. The stability of the Cu-Cu pairs in the conductor will be discussed using their binding energies.   

Non-equilibrium Phonons in CaWO$_{4}$: Issues for Phonon Mediated Particle Detectors

The CRESST experiment looks for evidence of dark matter particles colliding with nuclei in CaWO$_{4}$, using cryogenic bolometers sensitive to energy deposition $\sim $ 10 keV with a few percent accuracy. Calibration of the energy deposited in the phonon system depends upon the details of the evolution of the non-equilibrium energy in the CaWO$_{4}$ absorber. Our phonon images sensitively measure variations in angular phonon flux, providing key information about the elastic constants and scattering rates that determine the energy evolution. Phonon pulses, created by focused photoexcitation of a 150 nm Cu film, are detected after propagation through 3 mm of CaWO$_{4}$. The 20 ns Ar-ion laser pulse creates a localized (10$^{-3}$ mm$^{2})$ source of 10-20 K blackbody phonons. The sample is at 2 K. Our images show that the elastic constants derived from ultrasonic velocities along high symmetry axes do not accurately predict the total phonon flux along non-symmetry directions. We present new data on the dependence of phonon flux on excitation level and discuss the influence of isotope and anharmonic decay on the shape of phonon pulses in these ultrapure samples. Thanks to J.P. Wolfe and the Frederick Seitz Materials Research Laboratory, Urbana, IL, for partial support of this work.   

Ultrafast anisotropic strain in semiconductors measured by x-ray diffraction

We have used time-resolved x-ray diffraction to probe the non-uniaxial properties of impulsive strains in ultrafast laser-excited III-V semiconductors. Transient shifts of x-ray rocking curves due to the strains are measured from three Bragg reflections whose scattering vectors range from perpendicular to the surface to nearly in plane. Time-dependent strain ellipsoids are then constructed, with a temporal resolution under $\sim$150 ps. We find that the strain consists not only of a longitudinal expansion along the surface normal, but it also includes slight compression along the transverse direction. We compare measurements for GaAs and InSb; their significant differences in electron diffusion rates allow us to distinguish between lattice and electronic effects. Supported by the U.S. Department of Energy.   

Using DFT-based Cluster Expansions to Study Oxygen Adsorption on Platinum and Gold (111) Surfaces

We have studied oxygen adsorption on Platinum and Gold (111) surfaces using Density Functional Theory. We have addressed the limitation on the number of configurations we can consider through DFT through the use of two-dimensional cluster expansions on our DFT data, allowing for rapid energy calculations for any arbitrary surface. We have used the cluster expansion to study adsorption properties and phase behavior on the surface including simulated TPD experiments and atomistic thermodynamics phase diagrams which we compare to experimental behavior.   

Can a re do of field theory numerically speaking help in I - E CDW curve data analysis?

Density wave physics has a plethora of current versus electrical field data from experimental measurements. The author presented in his PhD dissertation a way of giving a false vacuum interpretation of how current can be measured against the magnitude of an applied electric field in laboratory conditions. This article purports to re examine the feasibility of such modeling taking into account Dr. Fred Cooper's work on a time averaging scheme for phi to the fourth power field theory as could be applied to modified wash board potentials.   

Temperature-composition phase diagrams of Gd-doped EuO and EuS

We have computed the temperature-phase diagram of Eu$_{1-x}$Gd$_x$O alloys by combining density functional theory in the generalized-gradient approximation with Hubbard U correction on f-orbitals with the regular cluster expansion and Monte-Carlo approach. The cluster expansion fit has been performed with varying numbers of distinct cluster types until the formal cross-validation score is minimized. Our results indicate that (i) pair interactions are relatively stronger than other cluster types, (ii) the pair terms decay rapidly with distance up to 10 {\AA}, (iii) the pair terms are attractive for direct interactions between cations and repulsive for indirect interactions through anions, and (iv) the calculated convex hull is asymmetric about x=0.5, displaying more deep ground states in Eu-rich regions than in Gd-rich regions. The asymmetry of the convex hull may imply relative instability of Gd-rich compounds, as was shown by previously-reported experimental difficulties to make Gd-rich compounds. A comparison with a similar binary system - sulfur replacing oxygen - is made, showing that both oxides and sulphides are dominated by deformation interaction. The sulphides have a marginal tendency to phase-separate into pure compounds at low temperatures, whereas the oxides tend to order.   

Low energy metastable states and immiscibility in (SiC)$_{1-X}$-(AlN)$_X$

A cluster expansion Hamiltonian was fit to VASP/PAW calculated supercell formation energies, $\Delta E_f$, and first principles based phase diagrams (miscibility gaps) were calculated for the wurtzite-structure pseudobinary system SiC$_{1-X}$AlN$_X$. An unusually wide range of $3 \alt \Delta E_f~ \alt 125$ kJ/mole MX (M= Al, Si; X= N, C) was calculated and all supercells with $\Delta E_f~ \alt 8$ kJ/mole exhibited characteristic (SiC)$_m$(AlN)$n$ crystallography, in which (SiC)$_m$ indicates m SiC-double layers $\bot$ to the hexagonal c-axis, and similarly for (AlN)$n$. The prediction of (SiC)$_m$(AlN)$n$ low-energy metastable states, may explain why one can synthesize SiC$_{1-X}$AlN$_X$ films, or single crystals of arbitrary bulk composition, in spite of the very strong tendency toward immiscibility. Specifically, one expects that metastable films or single crystals will be dominated by a disordered stacking of SiC- and AlN-double layers.   

Alloy Stabilized Wurtzite Ground State Structures of Zinc-Blende Semiconducting Compounds

Although the ground state structures of zinc-blende (ZB) alloys have been extensively studied, the knowledge of the ground state structures of wurtzite (WZ) alloys remains incomplete. Here, the ground state structures of the A$_x$B$_{1-x}$C WZ alloys with $x=$0.25, 0.5, and 0.75 are revealed by a ground state search using the valence-force field model and density-functional theory total energy calculations. It is shown that the ground state WZ alloy always has a lower strain energy and formation enthalpy than the corresponding ZB alloy. Therefore, we propose that the WZ phase can be stabilized through alloying. This novel idea is supported by the fact that the WZ AlP$_{0.5}$Sb$_{0.5}$, AlP$_{0.75}$Sb$_{0.25}$, ZnS$_{0.5}$Te$_{0.5}$, and ZnS$_{0.75}$Te$_{0.25}$ alloys in the lowest energy structures are more stable than the corresponding ZB alloys. To our best knowledge, this is the first example where the alloy adopts a structure distinct from both parent phases.   

First-principles thermodynamic theory of epitaxial alloys: Prediction of spontaneous rotation of epitaxial habits in InGaN and GaAsSb alloys

A general-purpose method for calculating the stablest ground state structures and finite-temperature thermodynamics of AC-BC alloys grown \emph{epitaxially} on a coherent substrate is presented. In addition to the fact that such coherent epitaxy stabilizes certain ordered phases, we discovered that depending on the substrate and the film concentration there is a spontaneous rotation of the stablest film microstructure. This general behavior is revealed for both mixed-cation (In, Ga)N and mix-anion Ga(As, Sb) alloys on a variety of substrates. Such spontaneous rotation of the epitaxial habits can be understood by the tetragonal ratio $\eta\neq1$ of the corresponding bulk structure.   

Atomistic origins of the phase transition mechanism in Ge$_2$Sb$_2$Te$_5$

The fast and reversible phase transition mechanism between crystalline and amorphous phases of Ge$_2$Sb$_2$Te$_5$ has been in debate for several years. Through employing first-principles density functional theory calculations, we identify a direct structural link between the meta-stable crystalline and amorphous phases. The phase transition is driven by the displacement of Ge atoms along the rocksalt [111] direction from stable-octahedron to high-energy-unstable tetrahedron sites close to the intrinsic vacancy regions. Due to the instability of the tetrahedra, the Ge atoms shift away from those sites, giving rise to the formation of local-ordered 4-fold motifs coupled with long-range structural disorder. The high figures of merit of Ge$_2$Sb$_2$Te$_5$ are achieved from the optimal combination of intrinsic vacancies provided by Sb2Te3 and the instability of the tetrahedron sites provided by GeTe.   

First-principles theory of phase stability, solvus boundaries, and coherency strain in LAST (lead-antimony-silver-telluride) and in other doped AgSbTe$_{2}$ thermoelectric alloys

Bulk telluride alloys are promising thermoelectrics [e.g. the figure of merit (\textit{ZT}) of LAST (AgPb$_{m}$SbTe$_{2+m})$ alloys was reported$^{\ast }$ to exceed \textit{ZT}$\sim $2]. Recent theoretical examination$^{+}$ found that precipitation of ordered AgSbTe$_{2}$ phases in rocksalt PbTe likely contributes to the high \textit{ZT} of LAST, and predicted that the isoplethal PbTe-AgSbTe$_{2}$ phase diagram includes highly asymmetric miscibility gap. Here we generalize that analysis by first launching a search for \textit{unknown} (Ag,Pb,Sb)Te non-rocksalt phases (those deviating from the 1:1 cation:anion ratio), and second by presenting an extended analysis of the solubility limits for alloying AgSbTe$_{2}$ with PbTe and other tellurides. In particular, we find that the large asymmetry of the PbTe-AgSbTe$_{2}$ miscibility gap shares a common physical origin with the substitutional site preference for Pb in ordered AgSbTe$_{2}$, and that during coherent precipitation, the coherency strain increases the solubility limits in PbTe-AgSbTe$_{2}$ by a factor of $\sim $2 relative to the predicted$^{+}$ unstrained bulk values. $^{\ast }$K.F. Hsu \textit{et al.}, Science \textbf{303}, 818 (2004). $^{+}$S.V.Barabash \textit{et al.}, Phys.Rev.Lett. \textbf{101}, 155704   

Crystal and electronic structure of quaternary chalcogenide semiconductors I$_{2}$-II-IV-VI$_{4}$ (I=Cu, Ag, II=Zn, Cd, IV=Ge, Sn and VI=S, Se, Te)

Sequential cation mutation in zinc-blende chalcogenide semiconductors, from binary to ternary to quaternary compounds, is systematically studied using first-principles calculations. Several universal trends are found for the crystal and electronic structure of the ternary and two classes of quaternary chalcogenides. We find that (i) most I$_{2}$-II-IV-VI$_{4}$ compounds are more stable in the kesterite structure, rather than the widely-recognized stannite structure; (ii) Cu and Zn layers are easy to be randomized in kesterite Cu$_{2}$ZnSnS$_{4}$ and Cu$_{2}$ZnSnSe$_{4}$; (iii) the band gap decreases during the mutation; (iv) the band gap of Cu$_{2}$ZnSnSe$_{4}$ should be around 1.0 eV, not 1.5 eV as reported in previous absorption measurements.   

The electronic structures and structural properties of the amorphous Ge$_{2}$Sb$_{2}$Te$_{5}$, a phase change memory material

Ge-Sb-Te compound is one of the most promising materials for phase change random access memory. Recently, Ge$_{2}$Sb$_{2}$Te$_{5}$ has been under intensive researches. However, there exists a critical discrepancy between experimental and theoretical observations. In experiment, the ideal glass following 8-$N$ rule has been observed. There are deviations from 8-$N$ rule for melt-quench structures obtained by molecular dynamics calculations. In this presentation, we compare the melt-quench structure with ideal glass. We theoretically obtained the ideal glass using Si-As-Se compounds with a higher covalency The amorphous structure of Si$_{2}$As$_{2}$Se$_{5}$ is obtained by the melt-quench process and the elements are replaced by Ge-Sb-Te. It is found that the resulting Ge$_{2}$Sb$_{2}$Te$_{5}$ structures satisfy the 8-$N$ rule and all Ge atoms are tetrahedrally coordinated. The total energy of the ideal glass is higher than that of the melt-quench structure, explaining why the ideal glass has not been observed in the MD simulations. The electronic structures are also compared between ideal glass, melt-quench structure, and crystalline phase. It is concluded that the electronic character of the melt-quench structure lies in between those of ideal glass and crystalline phase.   

Molecular dynamics study on volume dependence of atomic and electronic structure in amorphous Ge$_2$Sb$_2$Te$_5$

In order to understand the confinement effect of the phase-change memory cell consisting of Ge$_2$Sb$_2$Te$_5$, we carry out first-principles molecular dynamics calculations with the simulation volume equal to or larger than the crystalline volume. The amorphous structures are obtained by rapidly quenching the liquid phase of Ge$_2$Sb$_2$Te$_5$. It is found that the energy gap increases monotonically with the simulation volume. Furthermore, the density of defect levels in the energy gap is reduced in the simulation using the large cell volume. The resistance drift of the phase-change memory cell is explained on the basis of the simulation results.