SiO2 surface defect centers studied by AES

A theoretical model is proposed on how a Si dangling bond associated with an oxygen vacancy on a SiO₂ surface (E_s′ center) should be observed by Auger electron spectroscopy (AES). The Auger electron distribution N_A(E) for the L₂₃VV transition band is calculated for a stoichiometric SiO₂ surface, and for a SiO_x surface containing Si-(e^?O₃) coordinations. The latter is characterized by an additional L₂₃VD transition band, where D is the energy level of the unpaired electron e^?. The theoretical N_A(E) spectra are compared with experimental N(E) spectra for a pristine, and for an electron radiation damaged quartz surface. Agreement with the theoretical model is obtained if D is assumed to lie ≈2 eV below the conduction band edge. This result shows that AES is uniquely useful in revealing the absolute energy level of localized, occupied surface defect states. As the L₂₃VD transition band (main peak at 86 eV) cannot unambiguously be distinguished from a SiSi₄ coordination L₂₃VV spectrum (main peak at 88 eV), supporting evidence is presented as to why we exclude a SiSi₄ coordination for our particular experimental example. Application of the Si-(e^?O₃) model to the interpretation of SiO₂Si interface Auger spectra is also discussed.     

Hydrogen chemisorption on silicon (100) 2 × 1

Changes in the valence band density of states (DOS) of a (100) silicon surface that accompany he chemisorption of atomic hydrogen onto that surface are deduced from a study of the changes in the L_2,3VV Auger lineshape. Complementary changes in the conduction band DOS are inferred from changes in L_2,3VV-core-level characteristic loss spectra (CLS). The chemisorbed hydrogen layer is identified as the dihydride phase from low energy electron diffraction measurements. Upon hydrogen adsorption the DOS at the top of the valence band decreases and new energy levels associated with the Si-H bonds appear lower in the band. Assuming that the Auger signal from the hydrogen covered sample consists of a superposition of a signal from silicon atoms bonded to hydrogen in the dihydride layer and an elemental-Si signal from the substrate, a N(E) difference spectrum with features due only to the dihydride is obtained by subtracting the background corrected, loss deconvoluted L_2,3VV signal for a clean (100)Si surface rom the corresponding signal for the hydrogen covered surface. Comparisons of the energy position of the major peak in this difference spectrum with that of the main peak in a gas phase silane Si-L_2,3VV spectrum, and of the corresponding Auger energy calculated empirically, indicate a hole—hole interaction energy of ~8 eV for the two-hole final state in the gaseous system and zero for the dihydride surface system. Hydrogen induced changes in the conduction band DOS are less apparent than those of the valence band DOS with only the possibility of a decrease in the DOS at the bottom of the conduction band being inferred from the CLS measurements. Electron stimulated desorption of hydrogen from the dihydride layer is adduced from changes in the Auger lineshape under electron beam irradiation of the surface. Hydrogen induced changes in the near-elastic electron energy loss spectra (ELS) are also reported and compared with previously published ELS results.     

The valence band of free K clusters studied by photoelectron and Auger spectroscopies

The valence states of free neutral potassium clusters produced by a gas aggregation source were probed by synchrotron radiation based photoelectron spectroscopy. The first ionization energy (IE) of the clusters was determined to be 10% larger than the work function of bulk potassium. Using electrostatic concepts and the IE, the mean size of the clusters was estimated to be ≈2000 atoms. Further information about the valence band was provided by investigation of the Auger process initiated by the ionization of the 3p level with a subsequent emission of an Auger electron from the valence band (M_2,3VV). Plasmon satellites were observed in Auger spectra of free metal clusters.     

Chemisorption of Si on Al(111) surfaces: A local-chemical-bond analysis from Auger transition density of states

Auger and electron loss spectroscopies have been used to study the local chemical bond between Si and Al, in the first stages of growth of Si deposited at room temperature on Al(111) surfaces. Si follows a layer-by-layer mechanism up to 2 monolayers with the formation of an Al(111)-3 × 3-Si structure at about 0.44 monolayers. A detailed analysis of the L_2,3VV Auger spectra for this structure allows to interpret the Si and Al Auger transition density of states (TDOS) in terms of the actual p-like partial DOS centered on the Si and Al sites. The experimental results indicate a strong SiAl interaction with the formation of a p-type local covalent bond between the Si and Al surface atoms.     

The electronic structure and the band gap of nano-sized Si particles: competition between quantum confinement and surface reconstruction

The electronic structure and especially the band gap of Si_n clusters (n=3–45 atoms) is studied by photoelectron spectroscopy. Contrary to expectations of quantum confinement, almost all clusters studied here have a band gap smaller than that of crystalline Si or even display a continuous (metallic) density of states. We attribute this to covalent bond formation analogous to the reconstructions observed on single-crystal surfaces. Additionally, for Si₃₀ and Si₃₃ a gap size of 0.6 eV (0.4 eV) is observed, supporting the prediction of stable, spherically symmetric structures of these particular clusters. Received: 18 November 1999 / Accepted: 24 November 1999 / Published online: 5 April 2000     

The temperature dependence of the photoluminescence from overgrown GaAs/Al0.3Ga0.7As quantum dots and wires

Photoluminescence measurements on GaAs/Al_0.3Ga_0.7As quantum dots and wires fabricated using electron bears lithography and reactive ion etching are reported both before and after regrowth with a layer of Al_0.4Ga_0.6As. Dots exhibit little change in luminescence efficiency from the bulk with a reduction in diameter either before or after regrowth. Surface recombination therefore appears to be suppressed. In wires, however, luminescence intensity is very sensitive to wire width, decreasing rapidly with this parameter, but recovers and becomes independent of size after overgrowth. The temperature dependence of the photoluminescence from the dots and wires showed that dots and wires less than 150nm in width luminesced to higher temperatures than the larger diameter structures and dots liminesced to higher temperatures than wires of comparable width. This suggests that there is a finite coherence area effect which increases the radiative lifetimes of excitons in the quantum structures due to the geometric constraint, in the lateral direction in the wires and in all three directions in the dots. Below 20K bound exciton luminescence dominates in the dots but not in the wires. In wires it is still possible for the excitons to diffuse to nonradiative sites within the exciton lifetime. Regrowth at 750°C causes migration of aluminium into the quantum well and causes the shape of the well to become parabolic resulting shifts in the exciton emission to shorter wavelengths, making it difficult to separate the effect of processing from those due to quantum confinement.     

Electroluminescence from Si quantum dots/SiO2 multilayers with ultrathin oxide layers due to bipolar injection

Si quantum dots/SiO₂ multilayers with ultrathin oxide layers (2.4 nm) were fabricated on a p-type Si substrate in order to enhance the hole injection. Besides the luminescence band at 900 nm which was also shown in photoluminescence spectra, another strong luminescence band near the infrared region (1200 nm) can be observed in electroluminescence spectra. It can be assigned to the band-edge emission from the quasi 2-dimensional potential well in the Si substrate. Moreover, it is interesting to find the reduction of photoluminescence intensity under biased conditions which can be attributed to the occurrence of non-radiative Auger recombination process in charged Si quantum dots.     

Mechanisms for oxygen adsorption on the Si(110) surface studied by Auger electron spectroscopy

Oxygen adsorption on the Si(110) surface has been studied by Auger electron spectroscopy. For a clean annealed surface chemisorption occurs, with an initial sticking probability of ～6 × 10^?3. In this case the oxygen o_kll signal saturates and no formation of SiO₂ can be detected from an analysis of the Si L_2,3VV lineshape. With electron impact on the surface during oxygen exposure much larger quantities are adsorbed with the formation of an SiO₂ surface layer. This increased reactivity towards oxygen is due to either a direct effect of the electron beam or to a combined action of the beam with residual CO during oxygen inlet, which creates reactive carbon centers on the surface. Thus in the presence of an electron beam on the surface separate exosures to CO showed adsorption of C and O. For this surface subsequent exposure in the absence of the electron beam resulted in additional oxygen adsorption and formation of SiO₂. No adsorption of CO could be detected without electron impact. The changes in surface chemistry with adsorption are detectable from the Si L_2,3VV Auger spectrum. Assignments can be made of two main features in the spectra, relating to surface and bulk contributions to the density of states in the valence band.     

Transition density of states for Si(100) from L1L23V and L23VV Auger spectra

The L₁L₂₃V and L₂₃VV Auger spectra of sputtered and annealed Si(100) have been measured and the transition density of states extracted. The line shapes for the two transitions differ, indicating the importance of matrix element effects. Whereas the L₁L₂₃V line shape closely resembles results of other measurements of the Si density of states, the L₂₃VV line shows a strong emphasis on p-like states.     

Room temperature visible light emission from Si/SiO2 multilayers: Roles of interface electronic states and silicon phase

We have studied luminescence properties and microstructure of 20 patterns Si/SiO₂ multilayers. The photoluminescence spectra consist of two gaussian bands in the visible-infrared spectral region. It has been demonstrated that the strong PL band is caused by the radiative recombination in the Si/SiO₂ interfaces states, whereas the weaker band originates from radiative recombination in the nanosized Si layers. The peak shift of this latter band shows a discontinuity that corresponds to a crystalline-to-amorphous phase change when the Si layers are thinner than 30 Å. The peak energy as a function of the layer thickness is interpreted using a quantum confinement model in the case of amorphous Si layers.     

Core-hole screening effects in the initial state of Auger emission across the transition metal series

Supercell method is used to study the relaxation and screening effects on the initial state of the Auger transition in metals. Our consideration is based on the assumption that when a core-hole exists long enough before the Auger transition occurs, the occupied valence states relax to screen the core-hole which results in a redistribution of the valence electrons, in particular within the atom that contains the core-hole. In order to make the interaction between the core-holes sites at different atoms negligible, the real metal is simulated by supercells repeated periodically. In each supercell one atom is considered to have a core-hole and many others not to have one. The electronic states concerned by the Auger transition are calculated by the self-consistent full-potential linearized augmented plane wave (FLAPW) method. Different responses of the local valence band on the site of the core-hole have been shown depending on whether the d-bands are partially or completely filled. According to the final state rule, the screening to the two holes in the local valence band after the Auger transition has also been considered, as examples, for Ni and Cu metals. The result shows that, with the existence of two holes in them, the states of the local valence band of Cu relax to atomic-like impurity states, while the local valence band of Ni changes to a much narrow band at the bottom of the original band. As examples, L₃VV and M₁VV Auger spectral profiles of Cu have been calculated in reasonably good agreement with the experiment.     

The structure of theL 2,3 VV-auger line of silicon and silicon compounds

TheL _2,3 VV Auger transitions of Si, SiO₂, and SiC have been measured and compared with the self-fold electron density of states. The data indicate that Auger matrix effects must be included to explain the structure of the Auger lines. A comparison with soft X-ray measurements of Wiech shows, that the measured Auger line shape is nearly identical with the self-foldK _β emission band. The selection rules for X-ray emission lead then to the conclusion that mostlyp-like valence electrons are involved in the Auger transition. This result indicates the relative importance ofs andp states in Auger transitions which is in accordance with theoretical calculations of Feibelman et al.     

On the Auger-photoelectron coincidence spectroscopy spectra of Cu(100), Cu metal and CuOx/Cu surface

The M₃–VV Auger-photoelectron coincidence spectroscopy (APECS) spectrum of Cu(100) and the L₃–VV APECS spectra of Cu metal and CuOx/Cu surface are analyzed in detail. The narrowing and energy shift of the photoelectron line in the M₃–VV APECS spectrum is well predicted by the present theory. The spectrum shows the presence of the M₂–M₃(V)–VV(V) decay in which a hole in the 4s band hops away prior to the decay of M₃ hole. The analysis of the L₃ photoelectron spectra of Cu metal measured in coincidence with the ³F or ¹G Auger line raises a question concerning the presence of two different core–hole states upon the L₃ level ionization recently proposed by Thurgate and Jiang [Surf. Sci. 466 (2000) L807]. The analysis of the L₃–VV APECS spectrum of CuOx/Cu shows that the final-state charge–transfer interaction plays an important role in CuO.     

Synchrotron investigations of the specific features in the electron energy spectrum of silicon nanostructures

The Si L _{2, 3} x-ray absorption near-edge structure (XANES) spectra of porous silicon nanomaterials and nanostructures with epitaxial silicon layers doped by erbium or containing germanium quantum dots are measured using synchrotron radiation for the first time. A model of photoluminescence in porous silicon is proposed on the basis of the results obtained. According to this model, the photoluminescence is caused by interband transitions between the energy levels of the crystalline phase and oxide phases covering silicon nanocrystals. The stresses generated in surface silicon nanolayers by Ge quantum dots or clusters with incorporated Er atoms are responsible for the fine structure of the spectra in the energy range of the conduction band edge and can stimulate luminescence in these nanostructures.     

Intrinsic luminescence of amorphous and disordered crystalline systems

A unified method is developed for describing the steady-state luminescence of exciton fluctuation states for weak excitation in different disordered systems. The phononless luminescence band is found to be formed by “radiative” states of the fluctuation tail in the density of states, i.e., by states for which nonradiative states are either nonexistent or have a low probability. The shape of the emission spectra calculated including the phonon interaction is in good agreement with experimental luminescence spectra of α Si:H and of solid solutions of ZnSe_(1−c)Te_c and CdS_(1−c)Se_c. Fiz. Tverd. Tela (St. Petersburg) 40, 890–891 (May 1998)     

Microscopic theory of Auger recombination in quantum wires

An analysis is made of mechanisms for Auger recombination of nonequilibrium carriers in cylindrical quantum wires. It is shown that two different Auger recombination mechanisms take place in these wires: a quasi-threshold and a nonthreshold mechanism. Both mechanisms are associated with the presence of heterobarriers but are of a different nature. The quasi-threshold mechanism is attributed to the spatial confinement of the carrier wave functions to the region of the quantum wire and in this case the quasi-momentum conservation law is violated and the Auger recombination process is intensified. As the radius of the wire increases, the quasi-threshold Auger recombination process goes over to a threshold process. The nonthreshold mechanism is caused by the scattering of an electron (hole) at the heterojunction; the rate of this nonthreshold Auger recombination tends to zero in the limit of an infinite-radius wire.     

Band-structure engineering of gold atomic wires on silicon by controlled doping

We report on the systematic tuning of the electronic band structure of atomic wires by controlling the density of impurity atoms. The atomic wires are self-assembled on Si(111) by substitutional gold adsorbates and extra silicon atoms are deposited as the impurity dopants. The one-dimensional electronic band of gold atomic wires, measured by angle-resolved photoemission, changes from a fully metallic to semiconducting one with its band gap increasing above 0.3 eV along with an energy shift as a linear function of the Si dopant density. The gap opening mechanism is suggested to be related to the ordering of the impurities.     

Luminescence of CdSe/Al<Subscript>2</Subscript>O<Subscript>3</Subscript> quantum wires at high photoexcitation levels

Nonlinear dependence of the CdSe/Al₂O₃ quantum wire luminescence intensity on the photoexcitation level and an asymmetric broadening of luminescence spectra accompanied by a high-frequency shift of the maximum are observed and explained in terms of the dominant filling of exciton phase space in wires with different cross sections.     

Manifestation of outgoing resonance in stokes and anti-stokes spectra of ZnTe and ZnMgTe quantum wires

The spectra of resonant light scattering by ZnTe quantum wires have been measured at excitation energies of 2.18–2.72 eV. The quantum wires have been grown on Si(100) and GaAs(100) substrates by molecular beam epitaxy. The effect of outgoing resonance with the electron transition energy E ₀ on the intensity of phonon lines of the Stokes spectrum and on the intensity ratio of the Stokes and anti-Stokes spectral lines has been studied. The energy E ₀ has been determined in ZnTe and ZnMgTe quantum wires from the edge luminescence spectra.     

Potential of Si1−xGex alloys for Auger generation in highly efficient solar cells

Recent investigation on Si solar cells demonstrated the utility of Auger generation for the creation of more than merely one electron/hole pair per absorbed photon. The semiconductor Si requires a minimum photon energy of about 3.4 eV for this internal carrier multiplication. The current of a Si cell is therefore not significantly increased by Auger generation when the cell is illuminated by an air mass 1.5 spectrum, which contains only few photons with energies above 3.4 eV. Use of Si_1–x Ge _x alloys promises a lower onset energy. Unfortunately, incomplete data on band structures ofrandom Si_1–x Ge _x alloys preclude a detailed quantitative discussion of the full potential for these materials. Nevertheless, (i) analogies to our own quantum efficiency data from pure Si, (ii) the calculated band structure of the hypothetical,ordered zincblende type Si_0.5Ge_0.5 crystal, and (iii) optical data for Si_1–x Ge _x alloys indicate an optimum Ge content ofx=0.6 tox=0.7.