Search

This paper discusses in what sense an over-represented multi-word pattern in a corpus of translated texts can be said to be typical of translated text. The purpose of the discussion is to assess, from a quantitative as well as from a qualitative perspective, the status of translational collocation data retrieved through data-driven methods from a comparable and parallel aligned cor¬pus of English and Swedish original and translated texts. The study focuses on the explicitation of clausal relations in translations from English into Swedish. In some cases, lexical and grammatical contrast lead to explicita¬tion, but in others explicitation is due to different factors.

This article evaluates a method of frequency-based retrieval of translation-specific linguistic surface patterning on the basis of two pilot corpus studies of fiction translated from English into Swedish. The focus of the article is on the relevance of different methods for corpus research on source language transfer, necessary grammatical adaptation, and translation strategies conditioned by non-imperative cross-linguistic factors. The two studies, which started from overrepresentation of grammatical words, indicated that the overall distribution differences are related to register differences, source language transfer due to cross-linguistic compatibility, and grammatical adaptation due to target language syntactic restrictions. It is hypothesized that the method used will be suitable primarily for cross-linguistic textual and register-oriented study as well as for the study of necessary adaptations to target language grammatical restrictions. On the basis of this hypothesis, it is suggested that the search for translation corpus patterning indicative of other features than these will benefit from a more theory-driven orientation, although semi-automatic corpus markup is likely to play an important part in the development of new theories and methods.

The local electronic structure of N atoms in Ga ( In ) AsN diluted alloys ( N concentration of 3%) is determined by soft-X-ray emission and absorption spectroscopies as element specific probes. The experimental spectra reflect the local 2p orbital-projected density-of-states of N impurities, the main recombination centers in Ga ( In ) AsN , which appears to deviate dramatically from crystalline GaN in both valence and conduction bands. In particular, we observe a N local charge transfer from the valence band maximum to deeper valence states, which fundamentally limits the optical efficiency of Ga ( In ) AsN , unless different N local environments are formed. The incorporation of In in large concentrations forms In -rich N local environments such as In 4 N , which become the main recombination centers in Ga ( In ) AsN due to a local decrease of the band gap. A k-conserving process of resonant inelastic X-ray scattering is discovered, which allows probing of the k-character of valence and conduction states despite the random alloy nature of Ga ( In ) AsN .

Resolving the 3D wave vector k in photoemission mapping of the band structure E( k ) requires knowledge of the unoccupied final states. Both dispersions and lifetimes of these states can be achieved by very-low-energy electron diffraction (VLEED). By incorporating the non-free-electron and excited-state effects in the final states, combining VLEED with photoemission provides accurate mapping of the valence E( k ) resolved in the 3D wave vector and under control of the intrinsic accuracy. We here concentrate on the most accurate combined method which uses angle-dependent VLEED and photoemission measurements. It provides access to many Brillouin-zone lines using one crystal surface, and benefits from an intensity gain and a better intrinsic accuracy near the Fermi level.

We report on self-energy effects in the electronic structure of Cu as a prototype weakly correlated system containing electron states of different localization. The unoccupied and occupied excited states were mapped fully resolved in the three-dimensional k using very-low-energy electron diffraction and photoemission, respectively. The self-energy corrections to the density-functional theory show distinct band- and k-dependence. These results are well described by quasiparticle GW calculations, especially for less localized states. We find correlation of the self-energy effects with spatial overlap of the one-electron wave function with the electron density, and elucidate the essential physics of this effect based on the electron-gas exchange correlation behavior.

Using ab initio calculations we have been able to show that the traditionally assumed charge redistribution on III-V semiconductor (110) surfaces has to be modified. The use of ab initio local ionicity calculations confirms that the relative polarity of the cation and anion increases but that the valence charge density locally decreases on both types of site at the surface. The decrease in charge is continuous as a function of atom radius and the overlap with the core levels diminishes. For small radii the cation loses less charge than the anion, which gives a positive electrostatic shift of the surface core level relative to the anion. Hence the traditional explanation for part of the surface core-level shift in terms of initial-state electrostatic effects is not satisfactory. Instead we believe the observed core-level shifts to be related to screening effects stemming from the dangling bonds in the final-state picture.

Alternative general definitions of ionicity and ionic radii based on the charge density null-flux concept are introduced. This method is local in nature and possible to adapt to any atom in any physical system including surfaces, interfaces, defects and alloys. These and other explicit advantages, as compared to the standard approaches, are discussed in detail. The current approach is especially suitable for density functional methods, and is applied to a number of III-V semiconductors, alkali halides and a few alkaline earth compounds. Good agreement with traditional results is found.

Ultrathin AlAs(100) layers of 1-, 2-, and 5-ML thickness buried in GaAs are investigated by nb initio calculations. Unique experimental soft-x-ray emission spectra are explained in terms of interfa ...

The mechanisms of the stability of bulk terminated semiconductor polar (111) surfaces are investigated for the first time with ab initio methods. In the cases of InAs(111), InP(111) and GaP(111), charge is found to be removed from the surfaces, which avoids an electrostatic catastrophe and stabilizes the surfaces. These new findings properly account for the experimental existence of these unreconstructed surfaces. The GaAs(111) surface is, however, predicted to be unstable in accordance with experiment.

Arsenic deposited on Ge(100) has been investigated by means of ab initio LDA calculations. The results support a previous interpretation of photoemission data in terms of symmetric As dimers. Additional new surface states have been found in an energy range not investigated experimentally. Upon adsorption of hydrogen the As dimer bonds become much weaker, resulting in an essentially increased As–As separation. The modified surface state band structure is predicted to contain some features which should be clearly visible in photoemission spectra.

Excited-state unoccupied band structure E ( k ) above the vacuum level, which forms the final states in photoemission (PE), may be directly accessed by Very-Low-Energy Electron Diffraction (VLEED) at the energies below ∼40 eV. Basically, VLEED reveals the critical points in the bands with dominant surface transmission. Recently developed methods of VLEED data analysis exploit this fact to determine the surface-parallel E ( k ‖ ) directly, and surface-perpendicular E ( k ⊥ ) by a fitting technique. For many materials the unoccupied bands, contrary to common belief, prove to be not free-electron-like, and significantly influenced by self-energy corrections Δ Σ . As the VLEED experimental E ( k ) contains all these effects, its use in PE band mapping in the valence band provides absolute, accurate and approximation-free, determination of the electronic structure.

Experimental very-low-energy electron diffraction (VLEED) data on the H-terminated Si(111) surface are analyzed using ab initio pseudopotential band calculations performed for bulk Si and an H-Si slab. The VLEED spectral structure reflects the unoccupied $E(\mathbf{k})$ of bulk Si, with the effect from the H layer being insignificant. The experimental information about the unoccupied $E(\mathbf{k})$ is used to analyze the final-state effects in photoemission.

Very-low-energy electron diffraction (VLEED) is a direct probe for the unoccupied electronic band structureE(k) corresponding to the upper bands of photoemission (PE). These bands can contain significant non-free-electron and self-energy effects. Use of the experimental VLEED bands in PE and inverse PE allows to determine the valence and conduction bandE(k) absolutely, i.e. with complete and approximation-free control of the 3D wavevector k.

The layer-perpendicular dispersions of the typical layered TMDCs and were studied by combining determination of the upper unoccupied bands by very-low-energy electron diffraction (VLEED) with mapping of the lower occupied bands by photoemission (PE). We found that the upper bands of these materials are very complicated, and are compatible neither with the free-electron, nor with the ground-state approximation. Knowledge of the upper bands allowed us to carry out a PE experiment optimized for the -resolved mapping of the lower bands. The PE data were consistently rationalized, using a map of the PE intensity as a function of the binding energy and the photon energy . We found that the PE intensity is well described by direct, -conserving, transitions, with minor shifts of PE peaks being basically a consequence of their broadening due to finite electron and hole lifetime. Finally the lower bands were mapped explicitly, using the PE peaks with minimal shifts and the experimental upper bands. The obtained is very consistent, and shows overall agreement with full-potential LAPW calculations.

Calculations of the soft x-ray emission spectrum have been carried out on a buried Si(100) layer in GaAs and compared with experimental data. We find that Si occupies Ga- as well as As-sites and that the local density of states is different for these two cases.

Quantum well states in very thin epitaxial ErAs layers on GaAs(100) have been found in angle-resolved photoelectron spectra. From the dispersive properties of the quantum well states effective masses are obtained representing electron motion parallel to the surface layers and orthogonal to the layers.

As adsorption and interaction with InP(110) is investigated by means of total-energy minimization calculations. We find that the As-P exchanged configuration has higher energy than the As/InP(110) epitaxially continued layer structure (ECLS), for all types of As reservoirs considered. In the presence of an additional As layer in ECLS (i.e., 1.5 ML adsorbed As), however, the exchanged configuration is only slightly higher in energy than the nonexchanged one. We conclude that the As-P exchange process is energetically unfavorable, and should in any case not be complete at room temperature, as suggested in a recent report. Our conclusion is supported by results of a photoemission study, including As adsorption, desorption, and redeposition, according to which the process is nonreversible.

We have combined photoelectron spectroscopy (PES) and very-low-energy electron diffraction (VLEED) to study the electronic band structure $E(\mathbf{k})$ of a material with complicated unoccupied upper bands, which are incompatible with the free-electron approximation. Using VLEED, we have experimentally determined these bands, and accordingly optimized the PES experiment. PES band mapping using the VLEED upper bands enabled the first consistent resolution of the lower occupied bands, in particular, the perpendicular dispersion $E({k}_{\ensuremath{\perp}})$, for a layered material. The combination of PES and VLEED is a powerful method to resolve even very complicated $E(\mathbf{k})$ absolutely.

We derive the excited-state bands of Cu above the vacuum level along the $\ensuremath{\Gamma}X$ and $\ensuremath{\Gamma}L$ lines, utilizing very-low-energy electron diffraction (VLEED) experimental data from the (100) and (111) surfaces, respectively. This is done by a band-fitting technique, based on empirical pseudopotential calculations of the excited-state bands and the corresponding reflectivity $R(E).$ The calculations are iterated until a good fit is obtained between calculated and experimental energies of $dR/dE$ extrema, each of which corresponds to a band irregularity point. We analyze how the excited-state bands are connected with VLEED and photoemission, and introduce a partial transmission (PT) to characterize the coupling properties of each band: If the absorption ${V}_{i}$ is small, the current adsorbed in the crystal from an incident-beam-excited Bloch wave in VLEED, or the photocurrent from a photoexcited final-state Bloch wave, are both proportional to the PT of the Bloch wave. The excited-state bands, obtained from VLEED band fitting, are found to deviate from free-electron bands notably enough to influence photoemission spectra. In particular, along $\ensuremath{\Gamma}X$ we identify two different bands, which constitute final states in the photoemission from the (100) surface. As similar effects are far more prominent for nonmetals, we conclude that determination of the upper bands by VLEED is essential for accurate photoemission band mapping.

We present a comparative core level photoemission study of the interaction between As and the (110) surfaces of GaAs and InP. In both cases it is found that As forms well ordered overlayers with the periodicity of the substrate surface. The interactions in the two cases are, nonetheless, very different. The bonding at GaAs is very weak, and the adsorbed species desorbs below 100 °C. In contrast, the interaction with InP (110) is significantly stronger, and at temperatures around 330 °C an As–P exchange reaction is indicated and is seen the core level spectra. The latter data are interpreted in terms of a recently proposed structure model involving an As-covered InAs layer.

The InAs(001) 2 × 4 and 4 × 2 surfaces have been investigated by angle-resolved photoemission. The X 3 and X 5 points were found to be located 6.0 and 2.7 eV below the valence band maximum, respectively, and the dispersion of bulk bands along the Г-X direction in the bulk Brillouin zone were well described by a theoretical calculation. From angle-resolved valence band spectra measured along the high symmetry directions [110] and [110], three surface induced states were identified on both the InAs(001)4 × 2 and the InAs(001)2 × 4 surface.

Experimental photoelectron spectra are usually interpreted using rather crude approximations for the upper states into which the electrons are excited. Better knowledge about these excited states could substantially improve the accuracy of valence band mapping by photoelectron spectroscopy. We here demonstrate that VLEED measurements are ideally suited for accurate determination of the desired upper states. This is illustrated by model calculations including absorption and self-energy corrections. The close correspondence between so-called irregularity points of the excited-state bands and the total electron reflectivity is established, which opens up the possibility for direct mapping of irregularity points by comparison with experimental VLEED spectra, and for fitting of the whole excited-state bands between these points. The proposed scheme is finally used to determine the excited-state bands of Cu along from measurements on Cu(111).

We report VLEED/TCS measurements on the layered transition metal dichalcogenide . By combining the experimental results with simple VLEED calculations, we have been able to determine the unoccupied bands constituting the upper bands in photoemission. Our analysis reveals accurate positions of so-called coupling bands, which predominantly contribute to both electron transmission and photoemission. Better knowledge about these bands may substantially improve the accuracy of valence band mapping, particularly in studies of layered materials like , as the upper bands of these are too complicated to be well approximated by free-electron-like bands. We discuss in some detail the connection between VLEED and photoelectron spectroscopy, and we obtain for the first time a detailed recipe for -resolved band mapping, with utilization of VLEED results. Some previous photoemission results are re-examined in the light of our present findings, and the power of combined VLEED/photoemission studies is stressed.

The valence and core electronic structure of the sputtered and annealed 1 × 1 periodic GaAs(311)A surface has been studied by angle-resolved photoelectron spectroscopy. Five surface bands are identified and their dispersions along high symmetry lines in the surface Brillouin zone are mapped out. While electron-counting indicates a metallic surface, the experiment show no evidence for partly occupied surface bands. Analysis of the core level spectra reveals one As and two Ga surface shifted components. The results are discussed in terms of surface geometry models.

An angle-resolved photoemission study of MBE grown InAs/GaAs(110) hetero-structures was carried out to investigate the establishment of valence bands as a function of overlayer thickness. The valence band spectra were found to change gradually up to thicknesses well above 10 nm. The data are interpreted in terms of excitations within the overlayer from a combination of substrate and overlayer initial states, the former tailing into the overlayer.

The electronic structure of molecular-beam-epitaxy-grown InAs(1\ifmmode\bar\else\textasciimacron\fi{} 1\ifmmode\bar\else\textasciimacron\fi{} 1\ifmmode\bar\else\textasciimacron\fi{})2\ifmmode\times\else\texttimes\fi{}2 and InSb(1\ifmmode\bar\else\textasciimacron\fi{} 1\ifmmode\bar\else\textasciimacron\fi{} 1\ifmmode\bar\else\textasciimacron\fi{})2\ifmmode\times\else\texttimes\fi{}2 surfaces is investigated by angle-resolved photoelectron spectroscopy. Valence-band spectra, and dispersions of five surface-related structures, are presented. The qualitative similarities of data from the two surfaces indicate that they are very similar, with respect to atomic and electronic structure. Comparisons with other (111) surfaces support the identification of the surface-related structures. \textcopyright{} 1996 The American Physical Society.

Surface sensitive high resolution core level spectroscopy has been applied to the molecular beam epitaxy grown InAs ( 1 1 1 )2 × 2 and InSb ( 1 1 1 )2 × 2 surfaces. For both systems the In 4d core level consists of one dominating component while the Group V core levels are deconvoluted into four components. This analysis is consistent with a surface model where the topmost layer consists entirely of arsenic or antimony. In this model, Group V atoms form trimers bound to Group V atoms in the first double layer, leaving a single Group V rest atom per unit cell.

Systematic experimental data on resonant inelastic X-ray scattering (RIXS) in GaN near the N K-edge are presented for the first time. Excitation energy dependence of the spectral structures manifests the band structure effects originating from momentum selectivity of the RIXS process. This finding allows obtaining k-resolved band structure information for GaN crystals and nanostructures.

The surface electronic structure of the In-rich InSb(100)4 X 1 surface has been studied by angle-resolved photoelectron spectroscopy. To determine the origin of different contributions in spectra, direct bulk interband transitions were first identified using a semi-empirical band structure calculation and assuming direct transitions into free electron-like final bands. Three surface-induced features were identified and their dispersions have been mapped along symmetry directions of the surface Brillouin zone.

Surface science goes back to the beginning of the 19th century. It was, however, only with the development of ultra high vacuum (UHV) technology in the 1960,s that reproducible experimental results from surfaces on the atomic scale could be obtained. At about the same time high speed computers appeared which made it possible to apply realistic computations of surface properties. The last decade ha8 seen a large progress in surface 8cience and the field is under rapid development. This is largely due to the continuous development of the experimental and the theoretical methods. Many basic properties of surfaces are now in principle understood. Much more research is however necessary to achieve detailed descriptions, in particular of complex systems such as interfaces, overlayers and multilayers. A coordinated interplay between theory and experiment is desirable to achieve an efficient progress. In this article we will illustrate the present understanding of semiconductor surfaces as achieved mainly by photoelectron 8pectroscopy using synchrotron radiation. We will choose three examples, namely results from (i) core level studies on a clean surface, (ii) valence band studies of a clean surface, and (iii) core level and valence band studies on an adsorbate system. Before presenting these results we will summarise the present understanding of the atomic and the electronic structure of crystalline semiconductor surfaces.

As part of an extensive project on the surface electronic structure of semiconductor compounds we report here on an experimental and theoretical study of the InP(110) surface. By comparison of angle-resolved photoemission data with results from empirical tight-binding calculations it has been possible to identify electronic surface related features and specify their origin.

We report photoemission studies of the valence band structure of VSe 2 and of VSe2 intercalated with Cs. Pure VSe 2 showed significant band dispersion both perpendicular and parallel to the layers, i.e. the valence band of VSe 2 is 3D in character, confirming self-consistent LAPW band structure calculations. After Cs intercalation the perpendicular band dispersion vanished, while that parallel to the layers remained, showing that the valence band structure had become 2D. The observed changes go far beyond the rigid band model, but are largely understandable in terms of intercalation-induced decoupling of the VSe 2 layers, and charge transfer from the Cs.

The electronic structure of the InAs( 111 )1 × 1 surface has been investigated by angle resolved photoelectron spectroscopy along the symmetry lines Γ K , Γ M , and Γ M ′ of the surface Brillouin zone. The bulk valence band structure was calculated using a combination of the linear augmented plane-wave method and the relativistic augmented plane-wave method. We have projected the theoretical bulk band structure onto the surface Brillouin zone to separate surface states from surface resonances. Two surface related structures, S1 and S2, have been observed and their Ei(k‖) dispersions are established. Both S1 and S2 show the symmetry of the 1 × 1 surface Brillouin zone, which is consistent with the observed 1 × 1 LEED pattern. We identify S1 as the As-derived dangling bond state, and S2 is associated with the backbonds connecting the As atoms in the surface layer with the underlying In layer.

We have used photoemission spectroscopy to study the valence band structure of clean and in situ Cs intercalated VSe2. The results show that the valence band structure of VSe2 is transformed, from initially being of 3D character, to become essential 2D, as Cs is intercalated. The changes in the electronic structure are to a large extent understandable as caused by intercalation-induced de-coupling of the VSe2 layers, and electronic charge transfer from Cs to the host material. Comparison with band structure calculations supports this picture. Our findings indicate that the Cs/VSe2 system may be of great value in studies of “reduced dimensionality” phenomena.