Research activities

Research in the ERMES group is mainly focused on the application of modern techniques of computational condensed matter physics/materials science to theoretical investigations of important aspects of the physics of semiconductors, insulators and nanostructures. In the following, you will find brief summaries of present and past research areas in the ERMES group.

Active areas:

1. Ab Initio Infrastructure for High-Throughput Materials Innovation and Discovery. Molecular biologists can nowadays design complex biosystems that are able to perform novel specialized biological functions. This tremendous progress is based on a revolution in sequencing techniques that was driven (1) by the high-throughput strategies to identify the genome of living organisms and (2) by methodologies that enable to map the genetic information onto specific functionalities. This revolution in biological techniques is paralleled in materials science by (1) the advent of high-throughput data generation frameworks that exploit innovative computer architectures (accelerated hardware) combined with (2) the progress toward computational data-processing methodologies that empower scientists to distill the complex interactions at the origin of specific materials properties. Our goals are to assemble data sets of materials properties by building on the synergy between high-throughput (HT) first principles calculations, hardware enhancements, and experimental literature; and to develop effective tools for analyzing the data and discover materials property descriptors. The proposed research is firmly grounded on the massively parallel infrastructure of the QUANTUM ESPRESSO density functional theory codes, and on the HT framework AFLOW, used to create AFLOWLIB.ORG, a library of materials properties that is continuously expanded thanks to the synergistic effort of the PIs. Our scientific goals are to map the materials genome to enable accelerated materials discovery, innovation and technological transfer by exploiting and controlling the interplay between interacting degrees of freedom, especially charge-phonon interplay in mixed-valence compounds, the coupling between electronic, mechanical, thermal effects in materials for energy conversion, and topological decompositions and spectral sampling for element substitution in critical technologies. These steps imply a computational effort of such scope that can be achieved only through extensive access to the leadership computing facilities and dedicated algorithmic developments. The realization of these objectives not only will provide the scientific community and the society at large with effective tools for the discovery and design of advanced materials, but will also give back to our nation a leadership role in the development and distribution of innovative scientific software for high-performance materials simulations.

2. Electonic, spintronic and thermal transport in nanostructures. WanT is an open-source, GNU General Public License suite of codes that provides an integrated approach for the study of coherent electronic transport in nanostructures. The core methodology combines state-of-the-art Density Functional Theory (DFT), plane-wave, norm-conserving pseudopotentials calculations with a Green's functions method based on the Landauer formalism to describe quantum conductance. The essential connection between the two, and a crucial step in the calculation, is the use of the maximally-localized Wannier function representation to introduce naturally the ground-state electronic structure into the lattice Green's function approach at the basis of the evaluation of the quantum conductance. Moreover, the knowledge of the Wannier functions of the system allows for the direct link between the electronic transport properties of the device with the nature of the chemical bonds, providing insight onto the mechanisms that govern electron flow at the nanoscale. The WanT package operates, in principles, as a simple post-processing of any standard electronic structure code. The WanT code is currently interfaced to the codes in the Quantum-ESPRESSO distribution (http://www.quantum-espresso.org, http://www.pwscf.org).
WanT calculations will provide the user with:
- Quantum conductance spectrum for a lead-conductor-lead geometry;
- Density of states spectrum in the conductor region;
- Centers and spreads of the maximally-localized Wannier functions of the system.
The development and maintenance of the WanT code is promoted by the National Research Center on nanoStructures and bioSystems at Surfaces (S3) of the Italian INFM-CNR (http://www.s3.infm.it) and the University of North Texas. The present release of the WanT package has been realized by Andrea Ferretti (S3), Benedetta Bonferroni (S3), Arrigo Calzolari (S3) and, Marco Buongiorno Nardelli (UNT).

3. Methodology development for atomistic materials with strong Coulomb interaction and exchange effects. Novel materials with stronger electron localization and correlation are vigorously sought for their rich physical and chemical properties. We are working on developing theoretical methodologies and computational techniques that can simultaneously address the localized (molecular) and delocalized (solid-state) electronic structure of these materials, in a seamlessly integrated computational framework.
For instance, we have shown an efficient algorithm that accurately interconverts the electronic structure between atomic-orbital (AO) and plane-wave Hilbert spaces. AO representations are not only desirable for computational accuracy and finiteness of the basis set, but also to gain a better chemical interpretation of the quantum-mechanical wavefunction. They are essential in a gamut of applications such as: construction of model Hamiltonians for correlated-electrons and magnetic systems; evaluation of quantum transport properties design of semiempirical potentials for solids and biomolecules; calculation of exact-exchange integrals; and applications within linear scaling density-functional theory, coupled cluster, quantum Monte Carlo and the GW methods. Moreover, the computation of local Hamiltonians allows the calculation of the electronic states of materials on ultra dense k-space grids for accurate Brillouin zone (BZ) integrations, an essential requirement for the high-throughput computational materials applications central to the mission of the Materials Genome Initiative.

4. Chemical reactions on low dimensional nanostructures and in nanoconfinement. Most chemical reactions of practical interest are carried out in micro- and nanoporous materials, which can enhance or reduce reaction yields through various different effects, including an increase in the surface area per unit volume, selective adsorption of reactants and/or products, and shape-catalytic effects, among others. A fundamental understanding of the role of each one of these different effects could lead to a systematic procedure for the design of improved catalytic materials that take advantage of all of them simultaneously. Despite the multitude of factors that can affect reactions in confinement, we can classify these effects into three main groups: (1) shape-catalytic effects, i.e. the effect of the shape of the confining material and/or the reduced dimensionality of the porous space, (2) physical (or “soft”) effects, including the influence of dispersion and electrostatic interactions with the confining material, and (3) chemical (or “hard”) effects, interactions that involve significant electron rearrangement, including the formation and breaking of chemical bonds with the confining material. The latter is usually considered to be the actual catalytic effect, and it is the one that has the most obvious influence on the reaction rates, as it alters the reaction mechanism. However, the first and second type of effects can also have a strong influence on both the rates and equilibrium yields, as has been shown in several recent theoretical calculations and experimental studies.

5. Geometrical and electronic properties of surfaces and interfaces. Recent advances in our ability to grow epitaxial, crystalline oxides on semiconductors have opened the full spectrum of electronic, optical and magnetic behavior in oxide dielectrics to field effect phenomena in semiconductors. At issue is an understanding of the role of nanostructure in the materials properties as the physical size of the system decreases to the point that classical electrodynamics is no longer viable. The barrier height for electron exchange at a dielectric/semiconductor interface has long been interpreted in terms of Schottky’s theory with a modification from gap states induced in the semiconductor by the bulk termination. It has recently been shown, with the structure specifics of heteroepitaxy, that the electrostatic boundary conditions are set by dipole formation in a distinct interface phase that acts as a “Coulomb Buffer”. This Coulomb buffer is tunable, offsets the relative electrostatic potential on either side of the interface, and functionalized the barrier height concept itself. When Schottky and Mott formulated the barrier height theory for a metal/semiconductor junction, and later when Anderson formulated the band-edge offset problem for semiconductor/semiconductor junctions, there was no consideration given to interface states as contributions to the electrostatic boundary conditions. The charge distribution at the interface was treated simply as a superposition of the bulk-terminated junction. These theories are insightful but consistently misrepresent the barrier height or band-edge offsets because real interfacial structure variations modify the intrinsic band lineup. While the bulk-termination view of the problem has been enhanced over the years these approaches are fundamentally flawed because they leave out the interface physics and its chemical bonding-induced charge transfer.

To put it succinctly, the classical electrodynamic interpretation traditionally used in these nanoscale systems is no longer capable for describing the essentially physics that give rise to the dramatic materials properties observed in these systems. Instead, it is the nanoscale features of these systems that give rise to the materials properties. To gain the necessary understanding of these systems requires a quantum mechanical and mesoscopic description. Our goal will be to investigate the most relevant scientific questions in order to completely understand the nanoscale physical properties of COS systems. We will focus on the determination of the interface structure amid the semiconductor and the alkaline earth oxide. The corresponding electronic structure and the role of the interfacial phase in tuning the properties of the interface and how this phase can be manipulated to obtain the desired behavior of the MOSFET structure. The band offsets, and the fundamental role played by the Coulomb buffer in determining the barrier height between the crystalline dielectric and the silicon. Interfacial strain effects, point and extended defects, stochiometry and interfacial doping. Moreover, we will be able to evaluate physical properties, such as local density of states and phonon spectra, that, linked with experimental measurement, will provide invaluable characterization tools of the different samples.

Past projects:

Electronic and transport properties of nanostructures. Given that the technological momentum is drastically shrinking the dimensions of devices, it is of crucial importance to ask if our physical understanding of electronic and transport properties of materials can be extrapolated to the unprecedented small spatial and temporal scales of nanostructured components. In particular, electronic transport takes a completely new meaning in molecular-size systems where macroscopic transport theory does not hold anymore.

In this respect, I investigated the mechanism of quantum conductance in carbon nanotubes. These systems are becoming prototypical for studies of transport in nanostructures and hold substantial promise for applications as active components in nano-devices. I have discovered several structure-dependent classes of electrical behavior in nanotubes at equilibrium and under strain, both under DC and AC conditions. this analysis, based on conductance calculations using realistic tight-binding and {\it ab initio} Hamiltonians, unveiled the crucial role played by structural deformations, such as bending, defects and tube-tube contacts, as primary factors in the control of the electrical behavior of carbon nanotubes. These results provide a clear interpretation of recent experimental findings and suggest avenues for the use of nanotubes as components in electronic devices. This work was recognized through a number of invited talks at international conferences and workshops, including the APS March 2000 and MRS Spring 2000 meetings.

Strength and plasticity of carbon nanotubes. Carbon nanotubes are extremely strong and resilient and thereby excellent candidates for use as light weight super-strong fibers and in novel high-strength mechanical devices. With the goal of investigating the ultimate strength of carbon nanotubes and their mechanism of failure, I have identified the first stages of the mechanical yield of these systems and unraveled their peculiar mechanical properties: plastic or brittle behaviors can occur depending on the external conditions and tube geometry. While brittle behavior determines the ultimate strength of nanotubes, plasticity opens the way of producing nanodevices from mechanical relaxations that modify the tube and its local electronic structure. Massively parallel first-principles calculations and semi-empirical methods have been employed in a quantitative analysis of the mechanical strength of nanotubes. These simulations have shown that carbon nanotubes have the highest strength of any known material. A complete map of the mechanical response of these systems has also been obtained. This is probably the first such map for any material obtained from theoretical considerations and computer simulation. It provides a clear connection between atomistic behavior and inherently mesoscopic phenomena such as crystal plasticity and fracture. This work has attracted wide attention in discipline-specific journals and popular reviews (PRL ``Focus'', Nature ``News & Views'') as well as invited talks. The predictions of the microscopic strength of carbon nanotubes have recently been confirmed by experiments in Rice, Harvard and Washington Universities.

Electronic, structural and dynamical properties of semiconductor surfaces and interfaces. We have used first-principles electronic structure calculations to study the influence of strain and composition on interface properties of nitride-based heterostructures. Central to our investigation was the issue of polarization fields induced by the low symmetry of nitride systems and the strain induced by the lattice-mismatch between the different layers. The results have demonstrated that strain and composition play an important role in tuning the band offsets of nitride interfaces, and thus provide a natural way of controlling the properties of heterojunctions.
Furthermore, we have shown that polarization fields in nitrides can be strong enough to induce a spatial separation of the electron and hole wavefunctions in a quantum well and thus reduce the interband recombination rate. This effect intrinsically limits the performance of quantum-well-based lasers and should be taken into account in the engineering of high performance devices. In a broader perspective, I have investigated the structural and electronic properties of a number of semiconductor surfaces, with particular emphasis on the interplay between microscopic structure and measurable quantities. For example, we have performed a combined experimental-theoretical study of the structure and dynamics of the InSb(110) surface. Inelastic He scattering experiments were interpreted and explained using {\it ab initio} DFT-LDA techniques. We obtained a thorough description of the structural and dynamical properties of this surface, including a geometrical description of the low energy surface phonon modes and an interpretation of energy loss peaks. I have also examined the initial stages of H-induced passivation of dangling bonds, using GaAs(110) and Si(111) surfaces as paradigmatic examples. The results show that H adsorption plays a critical role in removing the relaxation and reconstructions present on the clean surfaces.

Growth dynamics of carbon nanotubes. In order to enable technological applications of carbon nanotubes based on their strength or their electronic and transport properties, they have to be grown in sufficient quantity and quality, which is very difficult at present. Fundamental issues in controlling growth have thus to be resolved. We have investigated the growth of multiwalled carbon nanotubes using extensive classical molecular dynamics simulations. Our primary interest was to understand the role of the ``lip-lip'' interaction between concentric tube tips in the growth dynamics. We have shown that the lip-lip interaction provides a natural mechanism for the observed closing of nanotube shells, which accounted for the experimentally observed growth behavior. This allowed us to provide a theoretical estimate of the mean time for closure as a function of both the temperature and nanotube diameter. In particular, we predicted that lowering the growth temperature should lead to longer tubes.

Structural phase transition and dynamical properties of crystals under high pressure. Linear response theory within the DFT-LDA formalism and Landau theory of phase transitions were used together, for the first time, to predict and qualitatively interpret a structural phase transition. I have studied the relative stability of different high-pressure phases in various cesium halides. My analysis revealed the essential role played by a soft phonon mode as the primary order parameter of the phase transition, and the relevance of the phonon-strain coupling in stabilizing the high-pressure phases. I was able to reinterpret the experimental X-ray diffraction data for these compounds and unambiguously determine their high-pressure phases. This is particularly interesting in two respects: these systems are abundant in the earth mantle and the knowledge of their low-symmetry phases is relevant for predicting its dynamical behavior. Moreover, they are of fundamental interest for investigation of band-overlap metalization processes at high pressure.