Search

The relationships between orbital magnetization and some magnetic properties derived from dichroic spectroscopy are further clarified.

Ferroelectric (FE) materials display very peculiar structural, electrical, and electromechanical features. Their properties have been addressed, by means of ab initio simulations, from 1990. Since then, a great number of different contributions has greatly expanded our understanding of this important class of materials; quite often, these contributions were based on novel theories, methods, and algorithms. Nowadays, many properties can be accurately computed and predicted from first principles for a given FE material. Even more important, the work of the last decade has revolutionized our qualitative view of the phenomenon of ferroelectricity at large. We review the major advances concerning structural and lattice-dynamical properties, spontaneous and induced polarization, dielectric and piezoelectric properties, and structural phase transitions at finite temperature. Special attention is devoted to illustrating the novel methods which made such advances possible.

The bonding pattern of a covalent semiconductor is disrupted when a surface is cut while keeping a rigid (truncated bulk) geometry. The covalent bonds are partly reformed (with a sizeable energy gain) when reconstruction is allowed. We show that the ``electron localization function'' (ELF)---applied within a first--principles pseudopotential framework---provides un unprecedented insight into the bonding mechanisms. In the unreconstructed surface one detects a partly metallic character, which disappears upon reconstruction. In the surface reformed bonds, the ELF sharply visualizes strongly paired electrons, similar in character to those of the bulk bonds., 6 pages, 8 included figures; revtex file. To appear on Solid State Communications

We give evidence that some nonmagnetic (i.e., spin-integrated) properties of the antiferromagnetic late transition-metal monoxides (MnO, FeO, CoO, NiO) are substantially noncubic below the N\'eel temperature even when assuming the ideal rocksalt structure for the ions. Our findings, which are at variance with the currently accepted picture, are based on ab initio and model calculations for the case study of MnO. The calculated zone-center optic phonon frequencies and Born effective charge tensor of this material show a significant magnetic-induced anisotropy.

The dynamical charges measure the macroscopic polarization induced by a relative sublattice displacement in a polar insulator. When the material has a mixed ionic/covalent character such charges present nontrivial features. Apparently similar oxides have strikingly different behaviors: some (like the ferroelectric perovskites) have giant dynamical charges, while others (like ZnO) are quite normal and display dynamical charges close to the nominal static ones. The rationale for such differences is found, upon closely scrutinizing the covalence mechanism. Furthermore, the effects of electron-electron interaction on the polarization of a model highly correlated dielectric are discussed.

In a mixed ionic/covalent insulator the dynamical ionic charges are much larger than the static ones. The phenomenon is particularly relevant in ferroelectric perovskites, and is well understood at the mean-field level. Here we use a model Hamiltonian to investigate polarization in a strongly correlated, and partially ionic, insulator. Charge transport is evaluated as a Berry phase. At a critical value of the electron-electron interaction, the model has a transition from a band insulator to a Mott insulator: the static ionic charge is continuous across the transition, whereas the polarization is discontinuous. Above the transition, the anion transports a positive charge.

Recently, the Berry's phase approach has made the macroscopic polarization of ferroelectric perovskites accessible to full quantum-mechanical treatment. First-principle calculations for KNbO3 have demonstrated—amongst other things—the overwhelming importance of asymmetric Nb-O bond length variations: such variations drag a giant macroscopic current across the sample during the ferroelectric distortion. Here we investigate the microscopic mechanism mostly responsible for such giant currents. It is shown that only the highest (so-called O2p) electronic bands significantly contribute to the polarization; furthermore evidence is given that a sizeable hydridization between the O2p oxygen orbitals and the Nb 4d orbitals occurs, this hybridization being the sole responsible for the giant currents.

In ferroelectric perovskites, the strong spontaneous polarization is due to giant values of the Born effective charges, which are much larger than predicted on the basis of the nominal ionicity: we investigate the physical origin of such charges, which by definition measure the current traversing the sample during polarization reversal. Our first-principles calculations for ${\mathrm{KNbO}}_{3}$ identify unambiguously the dominant mechanism with the presence of covalent bonding. The giant charges are reduced to their nominal values in a computational experiment performed on a fake material, where covalence has been artificially suppressed.

Nonlinear piezoelectricity has recently been detected in strained-layer superlattices where CdTe is pseudomorphically grown along (111) on different materials. We investigate this problem from first principles using density-functional theory. As a byproduct, we provide the ab initio calculation of several properties of this material, including the equilibrium lattice constant, Born effective charges, internal strain parameter, zone-center optical-phonon frequencies, dielectric constant, elastic constants, and the «clamped-ions» piezoelectric constant. Nonlinear effects on all of them are considered. Our major result is that piezoelectricity is accurately linear over a wide range of volume-conserving strains, while it displays strong nonlinearity whenever the strain is not volume conserving

The vibrational properties of ZnSe, ZnTe, CdSe, and CdTe are determined by density-functional perturbation theory. To this end we have generalized this method so as to explicitly account for the nonlinear core correction to the exchange and correlation energy of systems treated with pseudopotentials. Furthermore, we have implemented a method to enhance the transferability of pseudopotentials of group-II atoms with shallow d electrons frozen in the core. The accuracy obtained in this way is similar to that previously achieved for elemental and III-V semiconductors.

Contrary to common textbook statements, the macroscopic electric polarization of a crystal cannot be unambiguously defined as the dipole of a unit cell. The only measurable quantities are either derivatives of the polarization (dielectric susceptibility, Born effective charges, piezoelectricity, pyroelectricity) or finite differences (ferroelectricity). The differential concept is a basic one in theoretical work too. I sketch here the basic theory and the state of the art in its implementation to real materials, both for induced and spontaneous polarization; for the latter case, I also propose a novel approach.

I provide here a comprehensive presentation of the so called 'modern theory of polarization', based on a Berry phase. This theory is implemented as a standard option in most electronic-structure codes for both crystalline and noncrystalline materials, and is routinely used whenever features ot the macroscopic electrical polarization of materials are addressed. Several reviews of the modern theory exist in the literature; with respect to them, the present work has important elements of novelty. In this work I present a formal derivation, starting from the many-body wavefunction at the most general level, and eventually specialising to the case of single-determinant wavefunctions for either disordered or crystalline materials.

The absolute deformation potentials (ADP's) in dielectrics depend in general on shape and surface effects, owing to the long-range nature of Coulomb interactions. We prove that, for a given direction of uniaxial strain, he ADP's of crystalline solids are well defined in terms of bulk properties in Bravais and diamond lattices, while they are not in polar lattices. We give an explicit expression for an arbitrary strain orientation and calculate them in Si and Ge using ab initio methods.

We introduce an alternative approach to the first-principles calculation of NMR shielding tensors. These are obtained from the derivative of the orbital magnetization with respect to the application of a microscopic, localized magnetic dipole. The approach is simple, general, and can be applied to either isolated or periodic systems. Calculated results for simple hydrocarbons, crystalline diamond, and liquid water show very good agreement with established methods and experimental results., submitted to Phys. Rev. Lett; second half completely rewritten to address referee's comments

Our present understanding of the dielectric response in semiconductors is based on recent theoretical work. We review here some of the existing calculations and we discuss the picture of the basic mechanisms and trends which emerges from the results.

The static dielectric properties of liquid and solid water are investigated within linear response theory in the context of ab initio molecular dynamics. Using maximally localized Wannier functions to treat the macroscopic polarization we formulate a first-principles, parameter-free, generalization of Kirkwood's phenomenological theory. Our calculated static permittivity is in good agreement with experiment. Two effects of the hydrogen bonds, i.e., a significant increase of the average local moment and a local alignment of the molecular dipoles, contribute in almost equal measure to the unusually large dielectric constant of water.

First-principles-based methods are used to determine the external dielectric susceptibility (i.e., the polarization response to the external electric field) and the internal susceptibility (i.e., the polarization response to the average internal field) in ferroelectric dots, wires, and films, as a function of the electrical boundary conditions. While the external susceptibility is obviously positive, we find that the internal one is negative over a wide range of boundary conditions for all kinds of nanostructures. A Landau-type phenomenological model provides a rationale for all of our findings.

Short duration cadmium diffusion in sealed ampoules was studied at a temperature of 550/spl deg/C and a Cd/P molar ratio of 0.28. The starting sources were physical mixtures of either elemental Cd metal and red P or Cd/sub 3/P/sub 2/ and red P. The Cd profiles were determined by SIMS. In relatively short diffusion anneals up to two hours, the surface concentration, or solubility of Cd with either diffusion source is the same. However, Cd penetration was shallower in samples diffused with elemental (Cd,P) mixture than in samples diffused with (Cd/sub 3/P/sub 2/,P) mixture as the starting source. The slower diffusion of Cd with elemental starting source (Cd,P) appears to be related to the kinetics involved in reaching the thermal equilibrium between CdP/sub 2/ and CdP/sub 4/. The formation of Cd/sub 3/P/sub 2/ en route to the eventual formation of CdP/sub 2/ and CdP/sub 4/ or surface related kinetics are proposed as plausible mechanisms. The results appear to be consistent with the establishment of thermal equilibrium between CdP/sub 2/ and CdP/sub 4/ at a Cd/P ratio of 0.28 per the available phase equilibria data. The diffusion kinetics and defect structure are discussed on the basis of quasi-equilibrium among the three condensed phases InP, CdP/sub 2/ and CdP/sub 4/.

The concept of macroscopic polarization is the basic one in the electrostatics of dielectric materials: but for many years this concept has evaded even a precise microscopic definition, and has severely challenged quantum-mechanical calculations. This concept has undergone a genuine revolution in recent years (1992 onwards). It is now pretty clear that—contrary to a widespread incorrect belief—macroscopic polarization has nothing to do with the periodic charge distribution of the polarized crystal: the former is essentially a property of the phase of the electronic wavefunction, while the latter is a property of its modulus. An outline of the modern viewpoint is presented. Experiments invariably address polarization derivatives (permittivity, piezoelectricity, pyroelectricity,…) or polarization differences (ferroelectricity), and these differences are measured as an integrated electrical current. The modern theory addresses this same current, which is dominated by the phase of the electronic wavefunctions. First-principle calculations based on this theory are in spectacular agreement with experiments, and provide thorough understanding of the behavior of dielectric materials.

We investigate some surfaces of a paradigmatic sp bonded metal--namely, Al(110), Al(100), and Al(111)--by means of the "electron localization function" (ELF), implemented in a first-principle pseudopotential framework. ELF is a ground-state property which discriminates in a very sharp, quantitative, way between different kinds of bonding. ELF shows that in the bulk of Al the electron distribution is essentially jelliumlike, while what happens at the surface strongly depends on packing. At the least packed surface, Al(110), ELF indicates a free-atom nature of the electron distribution in the outer region. The most packed surface, Al(111), is instead at the opposite end, and can be regarded as a jellium surface weakly perturbed by the presence of the ionic cores., Comment: 5 pages with 6 figures. Surface Science (in press)

The insulating state of matter is characterized by the excitation spectrum, but also by qualitative features of the electronic ground state. The insulating ground wavefunction in fact: (i) displays vanishing dc conductivity; (ii) sustains macroscopic polarization; and (iii) is localized. The idea that the insulating state of matter is a consequence of electron localization was first proposed in 1964 by W. Kohn. I discuss here a novel definition of electron localization, rather different from Kohn's, and deeply rooted in the modern theory of polarization. In fact the present approach links the two features (ii) and (iii) above, by means of essentially the same formalism. In the special case of an uncorrelated crystalline solid, the localization of the many-body insulating wavefunction is measured - according to our definition - by the spread of the Wannier orbitals; this spread diverges in the metallic limit. In the correlated case, the novel approach to localization is demonstrated by means of a two-band Hubbard model in one dimension, undergoing a transition from band insulator to Mott insulator., Comment: 12 pages with 3 figures. Presented at the workshop "Fundamental Physics of Ferroelectrics", Aspen Center for Physics, February 2000

Within an ab-initio HF scheme, we use both Berry-phase calculations and supercell calculations in order to compute the dynamical charges for lattice dynamics and the electronic dielectric constant for KNbO_3 and BaTiO_3. Comparison with experimental data indicates that HF provides a description of the electronic properties of this material whose accuracy is of the same order as the LDA one. There are however significant differences between the two sets of results, whose origin is scrutinized. Motivated by the study of surface and domain-boundary properties, we also present some results for BaTiO_3 slabs, including both genuinely isolated and periodically repeated slabs with different terminations. The capability of dealing with a genuinely isolated slab is a virtue of the localized-basis implementation adopted here. We demonstrate, amongst other things, the nontrivial dynamical-charge neutrality of BaTiO_3 [001] surfaces., 11 pages, 6 figures. Presented at the Fifth Williamsburg Workshop on First-Principles Calculations for Ferroelectrics

The problem of Fermi-level pinning at semiconductor-metal contacts is readdressed starting from first-principles calculations for Al/GaAs. We give quantitative evidence that the Schottky barrier height is very little affected by any structural distortions on the metal side---including elongations of the metal-semiconductor bond (i.e. interface strain)---whereas it strongly depends on the interface structure on the semiconductor side. A rationale for these findings is given in terms of the interface dipole generated by the ionic effective charges., 5 pages, latex file, 2 postscript figures automatically included

We have investigated the macroscopic polarization and dynamical charges of some crystalline dielectrics presenting a mixed ionic/covalent character. First principles investigations have been done within the Hartree-Fock, LDA, and model GW approaches. All calculations have been performed on the same footing, using the all-electron FLAPW scheme. Apparently similar oxides have strikingly different behaviors: some (like the ferroelectric perovskites) have giant dynamical charges, while others (like ZnO) are quite normal and display dynamical charges close to the nominal static ones. We find the rationale for such differences.

The local Zn/Se relative concentration at the interface in ZnSe-GaAs(001) heterostructures synthesized by molecular beam epitaxy was found to be controlled by the Zn/Se flux ratio employed during the early growth stage of ZnSe on GaAs. Correspondingly, the valence band discontinuity varies from 1.20 eV (Zn-rich interface) to 0.58 eV (Se-rich interface). Comparison with the results of first-principles calculations suggests that the observed trend in band offsets is related to the establishment of neutral interfaces with different atomic configurations.

In the present paper we discuss the electronic properties of semiconductor heterojunctions, focusing on the band offset problem. We address interface-specific phenomena, where the conditions of growth -- including controlled contamination and strain effects -- may significantly alter the properties of the junction. We study the effects of ultrathin intralayers (i.e., heterovalent implantation) both at homojunctions (such as GaAs/Ge/GaAs) and heterojunctions (such as GaAs/Si/AlAs). In both cases, our theory demonstrates how the intralayers control the band offset: this result is confirmed by recent experimental observation in several systems. We then consider the band-offset engineering at lattice-mismatched heterojunctions, studying the paradigmatic case GaAs/Si. We evaluate the band-offset for several fully relaxed inequivalent interfaces. Both macroscopic strain and microscopic morphology effects are considered.

Band offsets at lattice-mismatched heterojunctions can be tuned owing to their dependence on macroscopic strain, and hence on the substrate composition. The system studied here, GaAs/Si(001), is lattice mismatched and heterovalent, offering thus an additional flexibility, due to the intrinsic nonbulk character of the band offset at heterovalent junctions. Starting with a study of macroscopic and microscopic elasticity, we evaluate the band offset for several fully relaxed inequivalent interfaces. Both macroscopic strain and microscopic morphology strongly affect the offset between the topmost Si and GaAs valence bands, which, consequently, is tunable, in principle, by as much as 1.1 eV.

We review in this paper the electronic properties of semiconductor heterojunctions. We focus on interface-specific phenomena, where the conditions of growth—including controlled contamination—may significantly alter the physical properties of the junction. We start outlining the basic concepts of our first-principles theory of the band offsets. We then discuss in some detail the case of Ge/GaAs, where the band offset significantly depends on interface features, owing to the difference in chemical valence which induces electrostatic effects. For this system, sound experimental evidence of noncommutativity of the band offsets has been reached. We then discuss the effects of ultrathin intralayers at heterojunctions and homojunctions, starting with the case of heterovalent implantations, such as GaAs/Ge/GaAs. In these cases, the intralayers control the band offset: the effect has been predicted by our theory and subsequently experimentally observed in several systems. We review some recent measurements and their current interpretation, particularly in relationship with interdiffusion and atomic ordering at the interface. Finally, we discuss the electronic and optical properties of isovalent intralayers, i.e ultrathin quantum wells, such as AlAs/GaAs/AlAs and Si/Ge/Si. We show how the spatial confinement of carriers in an isovalent intralayer strongly enhances interband optical properties and can be used for the optical characterization of semiconductor interfaces at the atomic scale

We examined band discontinuities in AlAs-Ge-GaAs(001) and GaAs-Ge-AlAs(001) single-quantum-well structures, as well as individual isolated Ge-GaAs(001), Ge-AlAs(001), GaAs-Ge(001), and AlAs-Ge(001) heterojunctions. We found that well-defined in- equivalent neutral interfaces are established in III-V/IV/III-V structures for Ge coverages as low as 1--2 monolayers. Deviations from the transitivity and commutativity rules of heterojunction behavior reflect inequivalent local interface environments rather than charged interfaces.

We have created 0.35–0.45 eV band offsets at Ge homojunctions using Ga–As dipole intralayers, with the Ge valence band edge on the As side of the junction at lower energy. This is, to our knowledge, the first time that intralayer control of band discontinuities is extended to homojunctions, thereby expanding the potential domain of band gap engineering. Because these offsets occur over just a few atomic spacings, they rival heterojunction band edge discontinuities in breadth. The offsets were measured with synchrotron‐radiation photoemission spectroscopy. Similar band offset magnitudes occur for both ‘‘Ga‐first’’ and ‘‘As‐first’’ growth sequences, consistent with a truly dipolar effect. Both cleaved Ge(111) and thick ≊50 A Ge films deposited on cleaved Ge(111) were used as substrates, obtaining consistent results. The sign and magnitude of the effect is in agreement with a ‘‘theoretical alchemy’’ model. Experimental evidence of microdiffusion is discussed.

Band offsets can be modified at semiconductor heterojunctions or created at homojunctions by depositing thin intralayers of group-IV atoms at III-V/III-V polar interfaces. We present here a theoretical study of Si and Ge intralayers deposited along (001) at GaAs and AlAs homojunctions, and at GaAs/AlAs heterojunctions. Our results show that the offset is very sensitive to the coverage and abruptness of the intralayer. A comparison with recent experiments for Si in GaAs/AlAs suggests that Si atoms are confined over two atomic planes for coverages lower than about 0.5 monolayers, whereas for higher coverages Si diffusion occurs.

The problem of whether band offsets at semiconductor interfaces are determined by bulk properties of the constituents or substantially affected by interface phenomena is critically readdressed. In particular, the conditions under which band offsets depend or not on the interface orientation are examined. State-of-the-art pseudopotential calculations are performed in a supercell geometry for GaAs/AlAs and GaAs/Ge grown in different directions. A new approach to the calculation of band offsets at semiconductor heterojunctions is proposed, based on ab-initio linear-response theory. The offset is shown to be the sum of two terms, the first of which depends only on bulk properties and is therefore independent on interface orientation and abruptness. The second, which vanishes for isovalent and non polar heterojunctions, does depend on the details of the interface geometry, but can be calculated -once the geometry is known- from such simple quantities as the lattice parameters and dielectric constants of the two constituents. The linear-response results are compared to the supercell ones and to the experimental data.

We developed recently a simple theory for the dielectric response of a doped semiconductor, and calculated the effect of the valence dispersive screening on the ionized-impurity-limited (IIL) mobility. The effect was found to result in a small reduction of the IIL mobility. In this paper we apply the same approach to a semimetal. In this case the effect results in a strong dielectric enhancement of the IIL mobility.

Investigates the possibility of using phase shifts calculated for electron-atom scattering in performing a band-structure calculation in the solid. No explicit determination of the atomic potential is needed in the augmented plane wave method, because matrix elements can be directly related to the phase shifts. Conduction bands have been calculated for solid argon, and a comparison with other existing calculations is made.

The band structure of crystalline methane has been calculated in a linear-combination-of-molecular-orbitals approximation. In this molecular tight-binding method the Bloch sums are built up of Hartree-Fock orbitals of the free molecules. Interactions between the states of different molecules are calculated directly from molecular integrals without reducing the exchange interaction to an exchange potential. The approximations involved are shown to be well justified for a crystal made of closed-shell units. The band structure is obtained by solving secular equation in 400 points along the relevant symmetry lines. The calculated bands are relatively narrow, less than 2.5 eV wide, around the molecular levels. The band-to-band gap in solid methane is 24.5 eV. This value has the same order of magnitude as in solid rare gases.

The Hartree-Fock conduction band structure of solid neon is computed by the APW method. The logarithmic derivatives used have been obtained directly from the static-exchange electron-atom scattering phase shifts. A detailed comparison is made with the existing literature, where the HF conduction bands are obtained with the use of different schemes.

The long-standing problem of determining which interface-specific properties affect the band offset at semiconductor heterojunctions is readdressed using a newly developed theoretical approach. The actual interface is considered as a perturbation with respect to a reference periodic system (virtual crystal). By comparison with state-of-the-art self-consistent calculations, we show that linear-response theory provides a very accurate description of the electronic structure of the actual interface in a variety of cases, and sheds light on the mechanisms responsible for the band offset. Results are presented for a number of lattice-matched junctions, both isovalent and heterovalent. It is shown that—within linear response theory—band offsets are genuine bulk properties for isovalent interfaces, whereas they do depend on the atomic structure of the junction for polar interfaces between heterovalent semiconductors. In the latter case, however, the interface-dependent contribution to the offset can be calculated—once the microscopic geometry of the junction is known—from such simple quantities as the lattice parameters and dielectric constants of the constituents. Perspectives for extending the theory to non-lattice-matched systems are also briefly discussed.

The problem of whether band offsets at semiconductor interfaces are determined by bulk properties of the constituents or substantially affected by interface phenomena is critically readdressed. In particular, the conditions under which band offsets do not depend on the interface orientation are examined. State-of-the-art ab-initio pseudopotential calculations are performed for (GaAs)3(AlAs)3 grown in the (001), (110), and (111) directions. Our results are analysed through a novel definition of the interface charge distribution which does not snake any use of ideal reference interfaces: the dipole corresponding to such a distribution directly yields the potential drop across the interface. Our calculations give for the (001), (110), and (111) interfaces a band offset of 0.49, 0.51, and 0.49 eV respectively, thus indicating that orientation independence holds in this case. However, in the case of the (111) orientation, two inequivalent interfaces exist whose offsets slightly differ (0.07 eV); associated with this difference we also found a net interfacial charge accumulation at the two inequivalent interfaces (±2.8 ×10−4 electrons per unit surface cell). Our results are finally interpreted through a new model based on crystal symmetry and whose only ingredients are the bulk charge densities of the the two constituents. The model — though not reproducing the fine details of the (111) superlattice — is in excellent agreement with our first-principles results and with available experimental data.