Prof. John M. Wills (1), and Ann E Mattsson (2) (1) Head of Theoretical Division, Los Alamos National Laboratiry, Los Alamos, USA,

Transcription

1 Abstracts for the WORKSHOP in Electronic Structure Theory organized by Linköping Linnaeus Initiative for Novel Functional Materials, Theory and Modeling Platform Prof. Ann E. Mattson, Sandia National Laboratories, Albuquerque, USA. Enabling, and Enhancing the Predictive Power of, Science Based Engineering. Computational Engineering is ultimately limited by available Materials Models. However accurately the conservation equations (mass, momentum, energy, etc.) underlying continuum Engineering codes are implemented, and however many different physical phenomena (fluid dynamics, electromagnetism, etc.) these codes can describe, the predictive power is set by how well the material response is described. Traditionally, empirical materials models have been developed based on experimental data, often taken at large scale. However, computational investigations have recently arisen as a powerful complement to experimental studies in all areas of Materials Science. This paves the way for obtaining data for a larger part of the parameter space of materials models, where experimental data is unavailable because these experiments are dangerous, expensive or otherwise impossible to perform. It also allows us to investigate fundamental materials response mechanisms in a systematic way. In this way both the data set used for informing a model and the functional form of the model itself can be investigated and improved. In 1929, Dirac pointed out that with the Dirac/Schrödinger equation as a newly found fundamental physical law of nature, it was only a matter of solving this equation in order to know all the properties of a system. However, to this day no method exists capable of directly solving the equation for the complex and large-scale problems encountered in Engineering, including designing new materials for increased efficiency of solar cells or batteries. Instead, methods bridging the length scales between the Ångström scale of the Dirac/Schrödinger equation and the millimeter and larger engineering scales are essential. A vital step in this multi-scale chain is the Nobel Prize winning [1] Density Functional Theory (DFT), formulated and made into a practical scheme in two foundational papers by Hohenberg and Kohn [2], and Kohn and Sham[3] in 1964 and In this presentation I will describe various efforts to develop Science Based Materials Models for Engineering use but I will also discuss the very first step in the multi-scale chain, the exchange-correlation (xc) functional, the part of DFT that connects to the Dirac/Schrödinger equation. The fidelity of the approximations to the unknown, exact or divine [4], xc functional determines the achievable accuracy with DFT. Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy's National Nuclear Security Administration under contract DE-AC04-94AL [1] Walter Kohn, 1998, Walter Kohn's Nobel lecture: Electronic structure of matter--wave functions and density functionals, Rev. Mod. Phys. 71, 1253 (1999). [2] Inhomogeneous Electron Gas, P. Hohenberg and W. Kohn, Phys. Rev. 136, B864 (1964). [3] Self-Consistent Equations including Exchange and Correlation Effects, W. Kohn and L. J. Sham, Phys. Rev. 140, A1133 (1965). [4] In Pursuit of the "Divine" Functional, Ann E. Mattsson, Science 298, 759 (25 October 2002). Prof. John M. Wills (1), and Ann E Mattsson (2) (1) Head of Theoretical Division, Los Alamos National Laboratiry, Los Alamos, USA, (2) Sandia National Laboratories, Albuquerque, NM 87185, USA Accurate Total-Energy Electronic Structure Calculations with RSPt Several methodologies exist for calculating the electronic structure and energy of materials within the context of Density Functional Theory (DFT). Often several versions of a given methodology are available, distinguished by differing computational implementations. Even when these methodologies are implemented to solve the same set of equations, it is often the case that reported results differ between methodologies. Sometimes this is the result of deliberate approximation, but not infrequently results

2 purported to be converged differ between methodologies. In order to make definitive the predictions of necessarily approximate density functionals, it is important to establish that different methodologies, taken to completion, can provide identical results. This is particularly important when establishing the impact of new density functionals built to address phenomena such as relativity and confinement which are not well described by available functionals. In this talk, I will discuss the RSPt implementation of the Full-Potential Linear Muffin-Tin Orbital (FPLMTO) method. The FPLMTO method, which solves for the total energy of a material using a formally exact potential and density obtained from an optimized, site-centered basis, is often regarded as less accurate than methods such as FLAPW that use a large basis of augmented plane waves. I will illustrate how the predictions of RSPt are taken to convergence, and compare results with calculations obtained using the Vienna Ab-initio Simulation Package (VASP), a method that uses a large plane-wave basis. I will further describe the implementation of the Dirac equation in RSPt, which uses four-component basis functions obtained from the Dirac equation on a muffin-tin geometry, to solve the full-potential Dirac problem. This method is particularly suited to address the properties of rare-earths and actinides without uncertainty due to the treatment of relativity and to provide for the development and testing of new functionals able to deal with both electron confinement and relativity. Calculations of the low pressure energy-volume relation in thorium are compared with results obtained with other methods, illustrating the effect of full relativity in this rather simple actinide, and demonstrating the convergence of the Dirac-RSPt method. Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy's National Nuclear Security Administration under contract DE-AC04-94AL Prof. H. Ebert, Ludwig-Maximilians-Universitaet Muenchen, Department Chemie, Muenchen. Ab-initio studies on spin-orbit induced properties of magnetic solids. Spin-orbit coupling gives rise in magnetic solids to a large number of effects that are at the same time of great scientific as well as technological interest. As examples for this, results of recent investigations based on a fully relativistic Dirac formalism on the magneto-crystalline anisotropy, the Gilbert damping parameter and the anomalous Hall effect will be presented. To achieve a coherent treatment of the magnetic anisotropy the Breit interaction was recently included in the underlying Hamiltonian. In this way the spin-orbit induced part of the anisotropy energy as well as the dipolar shape anisotropy are treated on the same quantum mechanical footing. As is demonstrated for various layered magnetic systems, this approach gives results very similar to the common classical approximation to the shape anisotropy. The magnetization dynamics in magnetic materials is conventionally described in a phenomenological way on the basis of the Landau- Lifshitz-Gilbert (LLG) equation, with the Gilbert damping parameter representing damping processes. Results of ab-initio calculations of the Gilbert damping parameter and its temperature dependence will be presented and discussed for a number of disordered transition metal alloys. The Kubo's linear response formalism used for these calculations also provides a sound and powerful basis for investigations of transport properties of ferromagnetic alloys giving access in particular to ab-initio investigations on the spontaneous Hall effect. Corresponding results including a decomposition into intrinsic and extrinsic contributions will be presented and discussed for a number of transition alloys. Prof. Sergey Simak, Linköping University, Sweden. New theoretical tools for studying temperature dependent properties of materials. We will discuss our in-house method for thorough and accurate determination of thermodynamic and physical properties of strongly ahnarmonic systems, the temperature dependent effective potential technique (TDEP). It is based on ab initio molecular dynamics followed by a mapping onto a model Hamiltonian that describes the complex lattice dynamics. The method, in particular, allows one to study the behavior of force constants and elastic response upon the temperature-driven dynamical stabilization of crystal structures that are unstable at low temperatures and to predict pressure-temperature phase diagrams of different materials, including alloys. The corresponding set of codes and the necessary expertise are available to all groups participating in the LiLi-NFM environment.

3 Assoc. Prof. Weine Olovsson. Linköping University, Sweden. Theoretical spectroscopy from firstprinciples calculations. By performing first-principles calculations for obtaining spectroscopic properties it is possible to investigate the electronic structure and bonding in materials. In addition, the results can be used as a "fingerprint" in order to characterize different structures. In this brief talk I will demonstrate some of my recent results for different material studies utilizing software packages such as WIEN2k (FPLAPW), FEFF (multiple scattering), VASP (PAW) and the exciting-code (FPLAPW). The results include: X-ray absorption near-edge structure (XANES) spectra, using density functional theory (DFT) supercell technique with the core-hole approximation as compared with many-body perturbation theory through solving the Bethe-Salpeter equation (BSE). Core-level binding energy shifts (CLS) as a function of concentration and pressure in substitutionally disordered metallic alloys, as well as disorder broadening effects. An investigation of XANES and extended X-ray absorption fine structure (EXAFS) spectra for amorphous materials. Assoc. Prof. Rickard Armiento, Linköping University, Sweden. Challenges and Opportunities for Materials Design with High-Throughput Computing. I will introduce our various efforts towards high-throughput computational materials design. The talk includes an overview of the methodology of database-driven high-throughput computations and ongoing efforts to improve the accuracy of the computational methods to treat systems with localized states, e.g., transition metal compounds, as well as systems with van der Waals forces. I will cover our recent work in density-functional theory (DFT) on a semi-local (GGA-type) functional with a derivative discontinuity that addresses the description of orbital localization, charge transfer, and band gaps in DFT without incurring extra computational overhead compared to usual semi-local DFT functionals (e.g., LDA, PBE). A few applications of high-throughput materials design that has recently been performed, and which are planned for the future, will also be discussed, e.g., engineering materials for high-performance piezoelectrics and materials for use in photovoltaics. PhD Davide G. Sangiovanni, Linköping University. Toughness enhancement in transition metal nitrides. Enhanced toughness in brittle ceramic materials, such as transition metal nitrides (TMN), is achieved by optimizing the occupancy of shear-sensitive metallic electronic-states. This theoretical research aims to solve an inherent long-standing problem for hard ceramic protective coatings: brittleness. High ductility/toughness (reduced brittleness), in combination with high hardness, is thus one of the most desired mechanical/physical properties in modern coatings. Density functional theory (DFT) calculations are carried out to understand the electronic origins of ductility, and to predict novel TMN alloys with optimal hardness/toughness ratios. Importantly, one of the TMN alloys identified in this theoretical work has subsequently been synthesized in the laboratory and exhibits the predicted properties. Assoc. Prof. Gueorgui Gueorguiev, Linköping University, Sweden. Synthetic growth of nanostructured materials. Inherently nanostructured materials exhibit a rapidly expanding role in materials science, nano-opto-electromechanical systems, solid state electronics, optoelectronics and chemistry of catalysis. Modeling of nanomaterials provides reliable structural knowledge, cheaper way to find best candidates for desirable applications, and practical recipes how to synthesize them. We have developed an original approach - the Synthetic Growth Concept (SGC) for predictive simulations of inherently nanostructured materials and lowdimensional phases. SGC achieves reliable simulation of structural formation and growth of a compound, identification of its bonding features and structural patterns, relating them to its properties, and assessing its synthesis feasibility by evaluating deposition techniques, precursors and their concentrations. I'll mostly discuss 3 original and technologically prospective classes of materials: Si-based cluster assembled materials, carbon-based nano-units and graphene-like templates for nano-device applications, and III-Nitrides including low-dimensional phases and gas-phase chemistry relevant to synthesis of III-Nitrides.

4 PhD Cecilia Goyenola, Linköping University, Sweden. Designing carbon-based thin films from graphene-like nanostructures. In the same way as graphite can be modeled and understood as a stack of graphene layers, some nanostructured carbon-based thin films, such as fullerene-like thin films, can be modeled as assemblies of doped graphene-like low-dimensional templates. In addition, when atoms of an element different from carbon, such as fluorine or sulphur, substitute carbon atoms in a graphene-like network, important bonding and structural changes occur. With the aim of predicting the structural patterns and their impact on the properties of such compounds, computational methods can give invaluable insight, allowing models of graphene-like nanostructures and derived thin films to be studied. In this context, we have employed the synthetic growth concept, an original theoretical approach based on the density functional theory. It has been a powerful simulation tool which helped to define a whole new class of nanostructured compounds: carbon based thin films with fullerene-like and graphene-like structural features. Viktor Ivády a,b,*, R. Armiento a, I. A. Abrikosov a, E. Janzén a, and A. Gali b,c. a Linköping University, b Wigner Research Centre for Physics, Budapest, Hungary, c Department of Atomic Physics, Budapest University of Technology and Economics, Budapest, Hungary. The role of the screening in the density functional applied on correlated orbitals in an sp 3 electron bath. First principles characterization of structural, electrical and optical properties of the mixture of correlated and uncorrelated states is still a quite challenging task in computational physics. As it is well known, the density function theory in the Kohn-Sham scheme predicts qualitatively wrong band structure for strongly correlated systems. To overcome this shortcoming, the orbital dependent LDA+U method was proposed which introduces a correction to the LDA on the subset of correlated orbitals. However, for sp 3 systems probably the most widespread methods are the hybrid functionals in the generalized Kohn-Sham scheme, which introduce a homogeneously screened Coulomb interaction and treat the orbitals uniquely in the core and interatomic region. Recently, it was also shown that by choosing a proper screening parameter for the hybrid functional, correlated systems can be described accurately within this scheme. In our work we investigated the mixed system of correlated and uncorrelated states as transitional metal impurities in semiconductor host which built up from built up from d-orbital like defect states in the bath of sp 3 hybridized delocalized orbitals. Our results clearly show that the widespread HSE06 hybrid functional cannot describe such complex system correctly. The source of the discrepancies is the homogeneous screening of the bare Coulomb interaction. By introducing the point defect to the host semiconductor, this approximation becomes invalid and causes significant errors for transition metal impurities. We developed a scheme, HSE06+V w, which can correct the HSE06 functional by introducing an additional screening potential to break the homogeneous screening. The corrected charge transition levels and optical excitation energies are remarkably close to the results of experimental measurements and many-body perturbation theory calculations. Assoc. Prof. Jonas Björk, Linköping University, Sweden. Bottom-up assembly of covalent architectures perspectives from theory. Several studies have illustrated how large macromolecular structures can be assembled from molecular building blocks on metal surfaces. This experimental success has mainly been obtained via trial-and-error coupled to insight from organic chemistry, and to date there is little understanding of how these surfaceassisted reactions proceeds on a submolecular level. Here, we provide such insight from density functional theory based transition state calculations. First, we discuss the mechanisms of halogen-based covalent assembly, in which molecular building blocks with specific hydrogen atoms replaced by halogens are used [1]. The halogens are split-off more easily than their hydrogen counterparts, thereby generating unsaturated carbon atoms that concomitantly couple. Here, we focus on the formation of biphenyl from bromobenzene and iodobenzene, as model reactions for this type assembly, on the (111)-facets of the three coinage metals. Reaction barriers of dehalogenation, diffusion processes and phenyl-phenyl recombination, are provided [2]. The approach allows for a systematic investigation of how the different surfaces affect the elementary processes involved in halogen-based covalent assembly.

5 Recently, we introduced a new class of surface-assisted reactions: the homo-coupling between terminal alkynes on Ag(111) [3]. Our transition states calculations show that the reaction process is most likely initiated by the covalent coupling, followed by the dehydrogenation. This gives a reaction mechanism that is fundamentally different from the halogen-based approach, where the opposite it observed. [1] L. Grill, et al., Nature Nanotech. 2, 687 (2007) [2] J. Björk et al., J. Am. Chem. Soc. 135, 5768 (2013) [3] Y.-Q. Zhang, et al., Nat. Commun. 3, 1285 (2012) PhD Nina Bondarenko, Uppsala University, Natalia Skorodumova, KTH Royal Institute of Technology, Stockholm, Sweden. Polarons and bipolarons in alkaline earth oxides. Hole polarons and bipolarons bound to cation vacancies are the most probable causes of local magnetization in nonmagnetic semiconductors. Recently this phenomenon has drawn much attention described in the terms of d0 magnetism. Large number of experimental studies has shown that alkaline metal oxides are exhibiting localised hole polarons and bipolarons. However, even in their relatively simple case many fundamental questions of polaron physics are not yet fully understood. The difficulties underlying the problem originate both from the experimental uncertainties and the drawbacks of current theoretical models. We report the results of our study of polaron and bipolaron electron topology at the (100) MgO/CaO interfaces within density-functional theory using Hubbard-U corrections for the Op states. We show the preferable space topology of bi- and mono-polarons and analyze how polaron properties depend on degree of anysotropy in the system, type of cation defect and local strain. We have shown possibility of bi mono-polaronic transitions lying in the plan of interface (quasi two dimensional) or perpendicular to the plan of interface (quasi three dimensional). Assoc. Prof. Martin Magnuson, Linköping University, Sweden. Science and Opportunities at the new Synchrotron Radiation Facility MAX IV. MAX IV is a new high-brilliance 3 rd generation synchrotron radiation x-ray source that is presently being built in Lund. The facility will be state-of-the-art with nano-sized x-ray beams when it starts in 2015 and will open up new opportunities in many interdisciplinary fields such as materials science, life-science, earth science and nano-science. The advantages of the anticipated performance with exceptionally low emittance, high-brilliance and large coherence properties will be discussed in experiments such as spectroscopy, diffraction and x-ray imaging. An overview of the x-ray science possibilities and experiments at the initial program at the first seven beamlines will be given. As MAX IV will open new opportunities in many fields, it is of high interest to Swedish and international research user groups.