Session J23: Phase Changes, Metals, PC Memories

Phonon dispersions and structural transitions of CrO$_{2}$ under high pressure

Phonon dispersions of chromium dioxide (CrO$_{2}$) are calculated to investigate the structural phase transitions as a function of pressure. The structural phase transition has been confirmed from the ground state tetragonal CrO$_{2}$ of rutile-type (t-CrO$_{2}$) to the orthorhombic CrO$_{2}$ of CaCl$_{2}$-type (o-CrO$_{2}$). The ferromagnetic and half-metallic property is preserved even in o-CrO$_{2}$. The softening of Raman-active B$_{1g}$ phonon mode, which is relevant to the above structural transition, is also obtained. We will discuss the possible more structural phase transitions from o-CrO$_{2}$ and the related phonon and magnetic properties at much higher pressure.   

Optimizing plasmonic properties of supported silver nanocube monolayers

The refractive index (RI) sensitivity of extinction spectra was compared experimentally for silver nanocubes in solution and in supported monolayers prepared by the Langmuir technique. The size of the nanocubes, RI of supporting dielectric substrate, and the monolayer surface pressure were used as variables in refractive index sensing (RIS) optimization. The dipolar plasmon modes of the colloidal nanocubes were found to have the highest RIS values of 176, 361 and 480 nm/RI units for 40, 80, and 130 nm cubes respectively. The largest figure of merit (FOM) of 4.55 was measured for a quadrupolar mode of 130 nm nanocubes. When compared to suspensions, RIS of supported nanocubes were reduced by $\sim $50{\%} and decreased with increasing monolayer surface pressure. RIS of 40 nm cube monolayers appeared to be sensitive to the substrate RI due to the RI dependent plasmon mode hybridization resulting in dipolar (D) and quadrupolar (Q) modes. RIS sensitivities of D and Q modes for silicon- supported nanocube monolayer were found to be 78 and 170 nm/RI units respectively. The FOM for the Q mode appeared to be 5.5. Intensity of this Q peak was found to increase with the angle of incident light. This work shows the use of high refractive index dielectric substrates, a passive molecular spacer, and large angles of incidence can significantly improve the detection of plasmonic response by supported nanocube monolayers.   

Isotropic Structural Color of Nanostructured Metal Surfaces

Engineering angularly insensitive (i.e., isotropic) structural coloration in metals without employing extrinsic materials have been a challenge due to the strong absorption properties of metals in the visible range. In this study, by combining plasmonic resonances of metallic nanoparticles and incoherent light scattering from deterministic arrays with isotropic and diffuse Fourier space, we demonstrate isotropic structural coloration of metal films with spatial uniformity and angular insensitivity. Specifically, we explore the angular scattering properties of aperiodic gold nanoparticle arrays with Pinwheel geometry and their hyperuniform counterpart (Delaunay-triangulated Pinwheel centroid, DTPC). The structures are designed by electromagnetic simulations based on the Coupled Dipole Method in partnership with three dimensional Finite Difference Time Domain modeling of plasmonic resonant nanoparticles on gold films. The experimental characterization is performed by measuring the far-field scattering spectra of the proposed arrays using angle-resolved reflection spectroscopy under white light illumination. The measured radiation diagrams of the fabricated plasmonic arrays demonstrate controllable and isotropic coloration of gold films. The proposed approach can potentially advance plasmonic applications to display, tagging and colorimetric sensing technologies.   

Ab-initio study of the free liquid Hg surface

The free surface of liquid Hg at two temperatures (T=300 and 450 K) has been studied by using first principles molecular dynamics simulations. The calculated longitudinal ionic density profile shows an oscillatory shape extending several atomic diameters into the bulk and its wavelength is in good agreement with experiment. The associated self-consistent valence electronic density profile shows much weaker oscillations which are somewhat out of phase with the ionic ones. The calculated X-ray reflectivity shows a marked maximum at a wavevector transfer of $q_z$ $\approx$ 2.2 \AA$^{-1}$ whose origin is related to the surface layering. Moreover, it shows a good agreement with the experimental reflectivity data.   

Noncontact technique for measuring the electrical resistivity and magnetic susceptibility of electrostatically levitated melts

Over the last two decades the popularity of levitation methods for studying equilibrium and supercooled melts has increased steadily. Measurements of density, viscosity, surface tension, and atomic structure have become well established. In contrast, measurements of electrical resistivity and magnetic susceptibility of levitated melts have been very limited. To fill this void, we have combined the tunnel diode oscillator (TDO) technique with electrostatic levitation (ESL) to perform inductively coupled measurements on levitated melts. A description of the basic operating principles of the TDO and ESL will be given, as well as a description of the implementation and performance characteristics of this technique. Preliminary measurements of electrical resistivity in the solid and liquid state will be presented for samples of Zr, Si, and Ge, as well as the measurements of ferromagnetic transitions in Fe and Co based alloys.   

Orientational ordering of tetrahedral clusters in $CaCd_6$ alloy

In icosahedral CaCd quasicrystals and related alloy structures, each icosahedral cluster contains an innermost tetrahedral shell that is loosely coupled with its shell. We extract an effective potential for the pair interaction $V(\Omega,\Omega')$ of the orientations $\Omega$ of neighboring clusters, as mediated by the intervening atoms. Using the EAM potentials of Brommer et al.,~\footnote{Brommer, G\"ahler, and Mihalkovi\v{c}, Phil. Mag. 87, 2671 (2007)}we relax all atoms, with each tetrahedron constrained to a chosen (continuously variable) overall orientation but allowing distortions. Using singular value decomposition, the relaxed energies are represented as $V(\Omega, \Omega')=\sum_j A_j f_j(\Omega) f'_j(\Omega')$ where only a few terms are important and the $f_j$'s have simple sinusoidal forms (which can be understood physically). We thus obtain a fit with only a few parameters, in place of the 46-parameter fit of Brommer et al$^1$ based on 12 discrete orientations. By Monte Carlo simulations with the obtained interaction, we determine the pattern of the orientationally ordered state seen experimentally below $\sim 90$K, and check the ordering transition previously simulated only in small system sizes$^1$.   

Electronic structure of Sr$_2$IrO$_4$ and its doped materials measured by angle-resolved photoemission spectroscopy

The electronic structure of iridate Sr$_2$IrO$_4$ and its doped materials Sr$_2$Ir$_{1-x}$Rh$_x$O$_4$ and (Sr$_x$La$_{1-x}$)$_2$IrO$_4$ were investigated by angle-resolved photoemission spectroscopy. For Sr$_2$IrO$_4$, which is a J$_{eff}$=1/2 Mott insulator driven by strong spin-orbit coupling, the dispersion of the predominant Ir 5d bands has been successfully resolved and compared with theoretical calculations. The overall band structure is in line with LDA calculations with strong spin-orbit coupling and moderate electron correlation effects included. The detailed electronic structures of isovalent-doped Sr$_2$Ir$_{1-x}$Rh$_x$O$_4$ and electron-doped (Sr$_x$La$_{1-x}$)$_2$IrO$_4$ are also obtained and compared with the electronic structure of the mother compound. These results help reveal the delicate interplay between charge, spin, orbit, and lattice degrees of freedom in iridates and other correlated electron systems.   

Electronic Structures of Oxygen-deficient PtO$_{2}$

We studied the electronic properties of beta-platinum dioxide ($\beta $-PtO$_{2})$, a catalytic material, based on density functional theory. Using the GGA+U method whose predicted band gap is verified by GW calculations, we found that the creation of an oxygen vacancy will induce local magnetic moment on the neighboring Pt and O atoms. The magnetism originates not only from the unpaired electrons that occupy the vacancy induced gap state, but also from the itinerant valence electrons. Because of antiferromagnetic (AF) coupling and the localized nature of gap states, the total magnetic moment is zero for charge-neutral state (V$_{\mbox{O}}^{\mbox{0}}$) and is $\sim $ 1 $\mu _{B}$ for singly-charged states (V$_{\mbox{O}}^{\pm}$). Calculation of grand potential shows that, the three charge states (V$_{\mbox{O}}^{\mbox{0}}$, V$_{\mbox{O}}^{\pm})$ are of the same stability within a small region, and the negatively charged state (V$_{\mbox{O}}^-$) is energetically favored within a wide range of the band gap. On this basis we discussed the implication on catalytic behavior.   

Theory of Non-Fermi Liquids

The focus of this project is to understand the simplest model needed to explain the physics of the material FeCrAs. The geometry of the FeCrAs crystal is used to perform crystal field splitting calculations, which in turn indicates a model that we will attempt to use to describe two of the main trends in the experimental data collected for the material. The model that we believe captures the physical character of FeCrAs for the resistivity and heat capacity measurements is the double exchange model. Next, we must find a way to describe an antiferromagnetic transition at 125K.   

Evolution between metastable strain glass state and stable martensitic phase in TiNi alloy

Phase transitions occur when the thermodynamic free energy of one phase is not lowest for varying some external conditions, such as temperature, pressure, and others. However, metastable glass states that falls out of equilibrium may be observed on continued cooling. Some frozen metastable states can be further spontaneously transformed into thermodynamic stable ones. The corresponding inverse process, from low temperature stable phase to glass state then to another high temperature stable phase upon heating, seems proving the glass state like the metastable state like in chemical reactions. Here we report the inverse freezing (strain glass state) does not occur for the first time. The lower temperature B19' martensite directly transforms into the higher temperature B2 parent phase and the strain glass also transforms B2 phase in the heating process, although the glass state is in the intermediate between B2 and B19' phases in the cooling process in TiNi single crystal. Completely different transformation sequence reveals the inverse glass transition is not necessary between two stable phases and helps us further deepen the understanding of glass transition. \\[4pt] [1] X. Ren, et. al., Philos. Mag. A 90, 141(2010). \\[0pt] [2] X. Ren, et. al., MRS Bull. 34, 838(2009).   

Ab Initio Study of Phase-Change Materials Doped with magnetic Impurities

Chalcogenide Phase-change materials are of great technological importance due to their ability to undergo reversible and fast transitions between the amorphous and crystalline phases upon heating. Recently, it was shown experimentally that Ge$_{2}$Sb$_{2}$Te$_{5}$ doped with Fe atoms exhibits phase-change behavior for low concentrations of Fe and that both the amorphous and the crystalline phases are ferromagnetic at low enough temperatures. Moreover, the two phases were found to have different saturation magnetization. This finding opens up the possibility of exploiting the phase-change behavior for fast magnetic switching in e.g. spintronic devices. We have investigated the structural, electronic and magnetic properties of Fe-doped Ge$_{2}$Sb$_{2}$Te$_{5}$ by first-principles simulations based on Density Functional Theory. Both amorphous and crystalline (hexagonal and cubic rocksalt) phases of Ge$_{2}$Sb$_{2}$Te$_{5}$ were considered. Our results show that in the amorphous phase, the magnitude of the magnetic moments of the Fe impurities is reduced with respect to the crystalline phases, due to the different local geometries and chemical environments, which explains the experimentally observed magnetic contrast between the two phases.   

Crystallization Times of Ge2Sb2Te5 Nanostructures as a Function of Temperature

Phase-change memory is a promising non-volatile memory technology in which a small volume of a phase change material is reversibly and rapidly switched between amorphous and crystalline phases by a suitable electrical pulse. Amorphization is achieved by fast cooling after a melting pulse while crystallization is achieved through electrical breakdown of the amorphous element that leads to heating above the crystallization temperature for a sufficient period. Significant difference between the crystallization behavior of phase-change materials in bulk, thin films and in nanostructures have been observed[1,2]. We have studied the crystallization times of nanoscale Ge2Sb2Te5 line structures using 50-500 ns voltage pulses with a baseline offset in the 500 K to 625 K range under high-vacuum. The baseline voltage allows measurement of resistance before and after the pulse and crystallization time. Current and voltage were recorded with high time-resolution for amorphizing pulse and very long periods of time to observe the crystallization time. The crystallization time decreases for increasing temperature but it remains on the order of seconds even at elevated temperatures. [1] H. Wong et al., Proc IEEE 98, 2201, 2010. [2] T. Zhang et al., Scr. Mater. 58, 977, 2008.   

Electrical Resistivity of Liquid Ge2Sb2Te5 in Patterned Nanostructures

Phase change memory devices are based on the electrical resistivity contrast between the amorphous and the crystalline phases of chalcogenide materials [1]. Since melting is required to amorphize the material, knowledge of the liquid state properties is critical for device design. Ge2Sb2Te5 (GST) is the most studied phase change material. However, the two experimental values reported to date for electrical resistivity of liquid GST (one on a thin film [2], the other in bulk [3]) differ by an order of magnitude. We have extracted the electrical resistivity of liquid GST from single pulse measurements on a large number of encapsulated GST line structures with varying lengths, widths and thicknesses. Each structure is self-heated to melt by the microsecond voltage pulse while voltage and current are monitored using an oscilloscope. The liquid state resistivity is calculated from slopes of liquid state resistance versus 1/width, fitted as a function of length. The results we obtained for the liquid resistivity of GST in nanostructures are close to those obtained from measurements on bulk GST [3]. 1. H. Wong et al., Proc IEEE 98, 2201, 2010. 2. T. Kato and K. Tanaka, Jap. J. Appl. Phys. 44, 7340, 2005. 3. R. Endo et al., Jap. J. Appl. Phys. 49, 5802, 2010.   

3-D Numerical Study of Switching Dynamics in Nanoscale Phase Change Memory Devices

Phase change memory (PCM) is currently regarded as a strong candidate technology to replace Flash memory in the market. In this work we report a detailed numerical study of the switching process in a nanoscale PCM cell, namely its switching dynamics during SET (turn on) and RESET (turn off) operations. A comprehensive picture of the electrical, thermal and phase change dynamics is obtained using a multiphysics approach with coupled differential equations in the framework of a three dimensional finite element model. The complexity of the problem was handled by constructing separate submodels; an electrical model which involves a temperature and phase dependent electrical conductivity, a thermal model where the joule heating from the electrical current serves as the heat source and involves temperature and phase dependent thermal conductivity and a phase change model. In this presentation we will concentrate on the electrical and thermal submodels in detail. The results of the phase change model taking into account homogeneous and heterogeneous nucleation kinetics will be discussed in another presentation. We will compare the numerical results with experimental data on GST based nanoscale phase change devices with various contact sizes and shapes.   

3-D Simulation Model of Phase Change and Percolation in Phase Change Memory

Even though phase change memory (PCM) appears as a promising nonvolatile solid state memory with its high signal to noise ratio and superior scalability compared to other memory technologies, the complex nature of the phase change process necessitates advanced numerical modeling to optimize the performance of nanoscale memory cells. The phase change and the percolation processes of a nanoscale PCM cell during SET (turn on) and RESET (turn off) operations have been simulated based on a three dimensional finite element model. A multiphysics approach with coupled differential equations is used to observe and understand the phase change and percolation dynamics. The model to represent the PCM is divided into submodels consisting of an electrical, a thermal and a phase change model that affects nucleation kinetics of crystallites. Coupling the submodels in the framework of the multiphysics approach, this model allows us to predict threshold voltage and recrystallization temperature for switching by detecting the critical conditions for the formation of a conductive percolation path in phase change process. These results will be compared to the experimental results to be carried on. The subject of electrical and thermal model will be mentioned in another presentation.