Search

An analysis of surface Pb core‐level shifts from rock salt structure PbS(100) suggests large top layer contractions for group IVA element chalcogenides, in contradiction to expectations from other binary cubic semiconductors like MgO. To test this hypothesis, we have performed a structure analysis on PbTe(100) using eight beams of diffracted low‐energy electrons from PbTe(100) at 50 K. An intensity analysis, based on relativistic potentials shown to be accurate for fifth row elements, leads to the conclusion that the Pb sublattice in the top layer is contracted by 0.23 A (7%), whereas the Te species is unrelaxed, and the spacing between the second and third layers is 0.07 A (2%) larger than that of the bulk. Our analysis gives results analogous to those found for the (100) surfaces of other cubic materials rather than the large uniform top‐layer contraction suggested by the analysis of the core‐level shifts.

Previously measured low‐energy electron diffraction (LEED) intensity data for the (110) surfaces of the isoelectronic series GaP, ZnS, and CuCl, are reanalyzed using dynamical LEED calculations, a new selection of independent structural variables, and an automated optimization procedure. The bond lengths of the atoms in the top layer and perpendicular displacements of the cation‐anion sublattices in the top two layers are selected as the independent structural variables. All other bond lengths are taken to be equal to their bulk values. An automated gradient method search procedure is utilized to obtain these structural variables by optimizing the fit of the measured intensities by the calculated ones as determined by the x‐ray R factor. This procedure permits a precise comparison of the bond lengths of species in the top layer from one material to another. Our major finding is that the bond lengths, d, between the anion and cation in the top layer depend on the spectroscopic ionicity, f, via [(d−dB)/dB]=...

The chemistry, overlayer morphology, and band bending occurring upon room temperature deposition of Ga and Pb on GaAs(100) are studied via high resolution photoemission spectroscopy, low energy electron diffraction and contact potential difference measurement. Both interfaces are abrupt and unreacted. At submonolayer coverage, Ga and Pb grow two dimensionally on GaAs(100). An ordered (1×2) superstructure is obtained for Pb(1/2 ML)/GaAs(100) by low temperature annealing of the interface (∼300 °C). Clustering is dominant at high coverage. With Ga, the Fermi level (EF) converges toward a position 0.68 eV above the valence band maximum (VBM), in agreement with the position obtained for the GaAs(110) surface. The room temperature deposition of Pb leads to EF stabilization at 0.57 eV above VBM, independent of the initial surface stoichiometry (Ga or As rich). The evolution of the band bending as a function of metal coverage emphasizes in both cases the importance of adatom‐induced states at low coverages and th...

Epitaxial layers of the ionic zinc blende compound CuCl are grown on the (110) surface of GaP. The growth is performed by congruent evaporation from a single CuCl source. Surface and interface properties of CuCl/GaP are studied with Auger electron spectroscopy, ultraviolet and x‐ray photoemission spectroscopy, electron energy‐loss spectroscopy, and low‐energy electron diffraction. The interface formed at 100 °C is abrupt, but undergoes a massive chemical reaction which leads to the formation of a copper phosphide layer at high temperature (≥300 °C). The valence band discontinuity is 0.85 eV ±0.15 at the abrupt CuCl/GaP interface. The CuCl(110) surface is atomically ordered and exhibits a (1×1) unit cell. Its atomic geometry is determined by multiple scattering analysis of low‐energy electron diffraction intensities. The surface is found to be relaxed in a way which is entirely compatible with the ‘‘universal’’ structure of cleavage surfaces of tetrahedrally coordinated III–V and II–VI compound semiconductors.

The electron beam patterning of crystalline CaF2 and Ca0.5Sr0.5F2 thin films grown by molecular‐beam epitaxy is demonstrated. The mechanism leading to patterning is the removal of fluorine through high energy electron beam exposure and dissolution of the resulting metal oxide in a water‐based solution. A scanning Auger microprobe and a scanning electron microscope are used as exposing tools. The development of the films is performed either in deionized water or in a weak HCl solution. The chemical analysis of the patterned features confirms the development of small windows. The feasibility of the fabrication of micrometers features is verified.

Interfaces formed between metals (In, Ag) and cleaved (110) surfaces of III–V compound semiconductors (GaSb, GaAs) held at low temperature are studied with low‐energy electron diffraction and electron energy‐loss spectroscopy. We consider the unrelaxation of the semiconductor surface structure by the metal adatoms, the adsorption sites on the surface and their implications for semiconductor band bending. The results are discussed in light of recent scanning tunneling microscopy measurements on metal/semiconductor interfaces and tight binding calculations of preferential adsorption sites.Interfaces formed between metals (In, Ag) and cleaved (110) surfaces of III–V compound semiconductors (GaSb, GaAs) held at low temperature are studied with low‐energy electron diffraction and electron energy‐loss spectroscopy. We consider the unrelaxation of the semiconductor surface structure by the metal adatoms, the adsorption sites on the surface and their implications for semiconductor band bending. The results are discussed in light of recent scanning tunneling microscopy measurements on metal/semiconductor interfaces and tight binding calculations of preferential adsorption sites.

In this article we investigate the growth of GaAs on two different vicinal surfaces of Ge (100), cut 6° off the (100) plane toward the (110) plane or toward the (111) plane. Both substrates exhibit evidence of a regular array of double steps. Ge substrates and GaAs films are characterized with low energy electron diffraction, low temperature photoluminescence, scanning tunneling microscopy, electron channeling, and scanning electron microscopy. GaAs grown on wafers cut toward the (111) plane exhibits high quality as compared to reference GaAs samples, whereas GaAs grown on wafers cut toward the (110) plane displays clear evidence of three-dimensional growth and low crystallinity.

We present a systematic investigation of the formation of Schottky barriers between n- and p-GaN(0001)−(1×1) grown by metalorganic chemical vapor deposition and a series of high and low work function metals (Mg, Al, Ti, Au, and Pt). Al, Ti, and Mg react at room temperature with nitrogen, whereas Au and Pt form abrupt, unreacted interfaces. We find that the Fermi level movement on both n- and p-GaN is consistent with variations in metal work functions, but limited by surface or interface states. Upon annealing, the incorporation of Mg increases the density of acceptors as seen on both n- and p-GaN. In spite of similar work functions and chemical reaction with nitrogen, Ti and Al show drastic differences in Schottky barrier formation due to differences in the nature of the products of reaction. AlN is a wide band gap semiconductor whereas TiN is a metallic compound.

A study of the formation of an interface between the archetype organic semiconductor, 3,4,9,10 perylenetetracarboxylic dianhydride (PTCDA), and GaAs(100) is presented. We examine the growth of thin (few monolayers) and thick (up to 3000 A) organic layers are examined at room and low temperature, and conclude that bonding across the PTCDA/GaAs interface is stronger than the PTCDA van der Waals intermolecular bond. Although the molecular stacking distance of 3.21 A is clearly seen in x‐ray diffraction, the films are polycrystalline, and no interface molecular ordering is detected. The interface electronic structure and the energy band lineup are studied with photoelectron spectroscopy. The highest occupied molecular orbital in PTCDA is found at 0.7 eV below the top of the GaAs valence band, in agreement with a vacuum level alignment a this interface.

The work presented here shows a correlation between the position of Pb‐induced gap states, responsible for the initial movement of the Fermi level (EF) at the Pb–GaAs(100) interface, and the position of the Pb 5d core level measured by photoemission spectroscopy. The difference in Pb‐induced gap states is obtained by depositing submonolayer amounts of Pb on the As‐rich (2×4)–C(2×8) or the Ga‐rich (4×2)–C(8×2) GaAs(100) surfaces. These reconstructions are known to produce a large dipole‐induced difference in the work function of the clean surface. The correlation suggests that the reconstruction‐induced dipole is mostly preserved during the initial stage of Pb deposition and shifts the energy levels of the adsorbate.

The composition and structure of the CuBr(100) surface and CuBr/GaAs(100) interface are studied with low‐energy electron diffraction, ultraviolet photoemission spectroscopy, x‐ray photoemission spectroscopy, and Auger electron spectroscopy. The CuBr(100) surface exhibits a c‐(2×2) superstructure. Assuming that the surface is semiconducting and that all anion and cation surface dangling bonds are, respectively, full and empty, we show with electron counting that the c‐(2×2) structure is due to a half‐monolayer of anion or cation vacancies instead of the usual dimerization of (100) surface atoms. The interface composition, however, does not appear to follow the electron matching requirement and seems to be dominated by chemistry.

Work function and band bending measurements are performed on n‐ and p‐type GaAs(100) surfaces prepared by molecular‐beam epitaxy and As decapping. The band bending is almost independent of the decapping temperature and surface reconstruction [(1×1)→C(4×4)→C(2×8)→(4×2)]. The work function and electron affinity undergo large and systematic variations of up to 400 meV which consistently establish maxima and minima of these quantities for the (2×4)‐C(2×8) and (4×2)‐C(8×2) structures, respectively. These variations are due to surface dipoles induced by charge exchanges between the cation and anion dangling bonds in the top two layers of the crystals. These charge exchanges are fully explained by applying a simple electron counting model to each reconstruction.

Epitaxial GaAs/Ca0.5Sr0.5F2(1800 A)/GaAs heterostructures have been grown by molecular‐beam epitaxy on (100), (111)A, (511)A, (511)B, (711)A, and (711)B GaAs substrates. We used reflective high‐energy electron diffraction, Auger electron spectroscopy, low‐energy electron diffraction, Rutherford backscattering (RBS), scanning electron microscopy, and Raman scattering to study the surface morphology, crystallinity, and chemical composition of these films. On the (100) orientation, the top GaAs had a flat microstructure; the surface crystallinity improved with increased thickness from 500 A to 1 μm, but the size of grains due to antiphase disorder also increased with thickness. On the (111)A orientation, the GaAs formed (100)‐facets to lower the surface free energy. The surface morphology and near‐surface crystallinity improved with increased thickness. The top GaAs films from (511)A and (511)B substrates showed the best near‐surface crystallinities from RBS, electron channeling pattern, and Raman scattering...

The surface structure, morphology, and crystallinity of high Miller index (511) and (711) GaAs epitaxial layers prepared by molecular‐beam epitaxy are studied with electron diffraction, scanning electron micorscopy, Raman scattering, and Rutherford backscattering. All A‐ and B‐(511) and (711) surfaces exhibit an ordered array of steps with (100) terraces and (111) risers. The (511)‐B layers exhibit the smoothest surface morphology, and the best surface and bulk crystallinity among these layers.

In order to extend the comparison of the initial stages of Schottky barrier formation to different surfaces of III–V semiconductors, we present here an investigation with low‐energy electron diffraction, Auger electron spectroscopy, and ultraviolet photoemission spectroscopy of In– and Ag–GaSb(100) interfaces at room temperature and after annealing. The initial GaSb surfaces exhibit a (3×2) structure. The low‐coverage growth of In and Ag is essentially two dimensional. It is followed by substantial islanding (In) especially after annealing. Band bending measurements show no significant movement of the Fermi level from its initial position near the valence band maximum.

Contact potential difference (CPD) measurements using a Kelvin probe coupled with synchrotron radiation are used to investigate various aspects of the problem of surface photovoltage (SPV) induced by the synchrotron radiation at (110) and (100) GaAs surfaces. A large and quasipermanent SPV is found at surfaces of low doped and low temperature (110) samples. SPV discharge mechanisms are investigated. Finally, the CPD technique is used to define conditions which minimize SPV and allow accurate measurements of band bending at low temperature. Band bending measurements are reported for interfaces between metals and (110) and (100) GaAs surfaces.

The atomic geometries of the (110) surfaces of GaSb and ZnTe are determined by comparing calculated elastic low energy electron diffraction (ELEED) intensities with those measured at T=125 K. The quality of the model description of the observed intensities is measured by the x ray R factor Rx. Using the minimization of Rx as the structure determination criterion, the atomic geometries of these two surfaces are shown to be very similar, in fact, nearly identical to within the uncertainties inherent in the analysis. GaSb(110) is reconstructed via a bond‐length conserving rotation in its uppermost atomic layer characterized by ω1=30±2°, and the Sb relaxed outward from the substrate. ZnTe(110) exhibits an analogous bond‐length conserving rotation of ω1=28±2° in its uppermost layer. In addition, however, this layer is relaxed by 0.05±0.05 A toward the substrate, and the second layer exhibits a counter relaxation of the Zn outward by 0.025±0.05 A and the Te inward by 0.025±0.05 A. The spectroscopic ionicity of ...

The Al/GaP(110) interface, formed by evaporating Al on a clean, well‐ordered, sputter‐annealed GaP(110) surface was studied with ELEED and AES. Evidence for island formation, when Al is evaporated on the room‐temperature substrate, comes from the slow decrease of the Ga and P AES peaks and the persistence of the substrate ELEED pattern for up to 75 ML (monolayer) of Al. Limited substrate decomposition was also observed for higher rates of Al evaporation. Upon annealing at moderate temperature, Al undergoes a metal–cation replacement reaction whereby Al–P bonds are formed and metallic Ga appears at the surface of the system. With an initial Al coverage of more than 8 ML and a 690 °C annealing cycle, a few layers of AlP are formed.

LEED intensity profiles recorded from AlxGa1−xAs(110) MBE surfaces are compared with data obtained from GaAs(110)‐Al(ϑ) annealed interfaces. This comparison confirms the hypothesis of a heat induced Al–Ga replacement reaction and the formation of AlGaAs when Al is annealed on GaAs(110). AlAs(110) produces I–V profiles very different from those of GaAs(110) and preliminary modeling computations indicate small surface atomic reconstructions. The effect of the elemental As layer used to protect AlxGa1−xAs(110) surfaces during ambient transfer is also investigated.

The initial formation of several metal–GaAs(110) interfaces has recently been reexamined by evaporating the metal on the semiconductor surface cooled to 70–100 K. Overlayer homogeneity and metal‐induced unrelaxation are increased at low temperature (LT), whereas clustering, interdiffusion, and chemical reactions are substantially inhibited. Ag and Au interfaces with n‐ and p‐GaAs were studied with soft x‐ray and ultraviolet photoemission spectroscopies (SXPS, UPS) and low‐energy electron diffraction. Both interfaces displayed the very asymmetric LT pinning of the Fermi level (EF) reported earlier: (1) fast EF movement at ultralow coverage (

The polar (311) surfaces of GaAs exhibit four inequivalent bulk (1×1) terminations each of which exhibits a metallic electronic character associated with partly occupied surface states. Experimentally prepared surfaces also display a (1×1) symmetry, but the tendency for surfaces to reconstruct towards an insulating ground state suggests adatom surface structure models which yield the saturation of the surface dangling bonds. We examine two Ga–adatom models on an otherwise clean As‐terminated (311) surface. Empirical total‐energy minimization calculations reveal that the As‐terminated surfaces exhibit two possible minimum‐energy adatom structures with the top‐layer (adatom) Ga in a bridge or hollow site. In the bridge‐site configuration, the Ga species occupy twofold sites on top of a relaxed As‐terminated surface. In the hollow‐site configuration, the Ga occupy threefold sites slightly above a nearly unrelaxed surface. Both correspond to simple saturated‐bond models of the surface chemical bonding but onl...

Low‐energy electron diffraction (LEED) intensity–voltage data for a complete set of first‐order diffraction beams of the GaAs (110) surface is averaged at constant momentum transfer and compared with kinematical and dynamical model calculations for several surface atomic geometries. The atomic displacements in the surface and first two subsurface layers are obtained via this analysis. The influence of multiple scattering is examined by computing the single and multiple scattering contributions to the intensity–voltage profiles associated with atomic distortions within the first three lattice layers. This procedure allows the identification of the major features of the structure in the LEED–CMTA (constant momentum transfer average) profiles either with multiple scattering effects or with single scattering resonances derived from the lattice geometry.

Oscillations are found in the integrated intensity of elastic electrons scattered from amorphous Si. These oscillations are attributed to the multiple scattering of elastic waves between a central atom and its neighbors, an effect which is similar to but more general than that found in extended x‐ray absorption fine structure and related phenomena. The Fourier analysis of the signal obtained with amorphous Si produces a dominant peak which closely corresponds to twice the nearest‐neighbor distance. The result of a first‐principle multiple‐scattering calculation reproduces remarkably well in the experiment and clearly demonstrates the physical processes responsible for these oscillations.

In this study, Sn/GaAs interfaces are studied at both room temperature (RT) and low temperature (LT) (70–100 K) using soft x‐ray and ultraviolet photoemission spectroscopies and low‐energy electron diffraction. The movement of the Fermi level EF on the RT surface as a function of Sn coverage is initially slow on n‐GaAs and fast on p‐GaAs. This asymmetry is typically found only at LT for other metal/GaAs interfaces, where it is related to the delayed onset of metallic behavior of the overlayer. In contrast to these other systems, however, Sn/GaAs gives a classic example of Stranski–Krastanov growth at RT. The first two‐dimensional layer does not bring EF close to its final position on n‐GaAs. At the onset of three‐dimensional nucleation (3–4 A), EF drops abruptly to its final position on n‐GaAs. This drop is delayed to ∼5 A at LT. As for Ag and Au/GaAs, this relatively sharp change in the position of EF appears to be related to the formation of clusters and the appearance of the metallic behavior of the ov...

The atomic geometries of GaP(110) and ZnS(110) are reexamined using the R-factor minimization procedure, developed for GaAs(110) and previously applied to GaSb(110), ZnTe(110), InAs(110), and AlP(110), to analyze experimental elastic low-energy electron diffraction intensities. Unlike most of the earlier cases, both GaP(110) and ZnS(110) exhibit two distinct minimum-Rx structures which cannot be distinguished by analysis of the shapes of the intensity profiles alone. One region of best-fit structures exhibits top-layer displacements normal to the surface characterized by a small bond-length-conserving, top-layer rotation (omega aproximately 2-3 deg), a small relaxation of the top layer away from the surface, and a 10 percent expansion of the top-layer bond length. The other region of best-fit structures is the conventional one: nearly bond-length-conserving rotations of omega = 26-28 deg in the top layer and a small (approximately 0.1 A) contraction of the uppermost layer spacing. This ambiguity may be removed, however, by consideration of the integrated beam intensities. The conventional region of structural parameters provides a decisively better description of the relative magnitudes of the integrated beam intensities and hence is the preferred structure.

The atomic structure of Al on GaAs(110) is studied by soft x‐ray photoemission spectroscopy (SXPS) and elastic low energy electron diffraction (ELEED) for coverages ranging from 0.5 to 8.5 monolayers (ML). The core level measurements show a heat‐induced chemical shift of Al–2p to higher binding energy consistent with an Al–Ga replacement reaction and the formation of AlAs, and a decrease in emission from Al–2p and Ga–3d, indicating a penetration of Al in the lattice and thermal desorption or formation of islands of Ga freed by the replacement reaction. The ELEED multiple scattering analysis yields a correspondence between the initial experimental coverage and the structure of the interface. The low coverage (0.5–1.0 ML) data are best described by a replacement of Ga in the second layer, the medium coverage (1.5–2.5 ML) data by a replacement in the second and third layers and the high coverage (5.5–8.5 ML) data by replacing Ga in all layers (pure AlAs).

Preliminary results for the atomic geometry of the (110) surface of ZnS are reported. The surface structure is determined by comparing dynamical calculations of elastic low‐energy electron diffraction (ELEED) intensities with those measured at T = 300 K using a pulsed floodgun technique to suppress the charging of the insulating ZnS sample. In contrast to other tetrahedrally‐coordinated, zincblende‐structure compound semiconductors for which the (110) surfaces are known to be reconstructed, i.e., GaAs, InSb, InP, and ZnTe, the measured ELEED intensities from ZnS(110) are described acceptably by calculations based on the truncated bulk atomic geometry and minor variants thereof. Measures of the quality of this description of the ELEED intensity data are the x‐ray reliability (R) factor, Rx = 0.222, and the Zanazzi–Jona R‐factor, RZJ = 0.229. The minimum x‐ray R‐factor which we have achieved thus far, i.e., Rx = 0.219, is obtained for a puckered surface geometry in which the sulfur species relaxes verticall...

In this paper, a study is conducted of the composition and structure of epitaxial ZnSe films grown by congruent evaporation on GaAs(110) at a rate of 2 A/min. It is found that the films grown on 300 C GaAs are nearly stoichiometric and form an abrupt interface with the substrate. Films grown at higher temperature (T greater than 350-400 C) are Se rich. The crystallinity of films grown at 300 C is good and their surface atomic geometry is identical to that of a ZnSe crystal. The GaAs-ZnSe interface geometry seems to be dominated by the Se-substrate bonds. The adsorption of Se, during the formation of very thin ZnSe films (2-3 A), produces a (1 x 2) LEED pattern and modifications of the LEED I-V profiles, which probably indicate a change in the substrate atomic relaxation.

Because of the expense of dynamical low‐energy electron diffraction (LEED) computational analyses, a more economical approach to the comparison between measured and computed LEED intensity profiles of nonpolar compound semiconductor surfaces consists of determining an initial estimate of surface atomic position coordinates via comparison of experimental data with kinematical calculations. Once an acceptable region of position parameter space has been identified, further refinements can be obtained by dynamical LEED analyses. This paper presents the results of the kinematical portion of such an analysis applied to GaAs(110) and ZnSe(110) surfaces. It is found that both materials exhibit similar surface atomic rearrangements characterized by bond length distortions and vertical displacements of the two sublattices in the uppermost atomic layer. These rearrangements differ from the ’’rotated’’ rearrangement reported earlier for GaAs(110) principally in that the structures identified herein involve small chan...

LEED IS (intensity–momentum transfer) data has been used to verify atomic displacements in the first subsurface layer of GaAs(110). IS profiles averaged over a range of incidence (ϑ) and azimuthal (φ) angles, subject to constant S, data for selected values of ϑ and φ, and similar IS profiles computed from kinematical calculations are intercompared to assess the relative importance of inadequate data averaging and variations of geometric parameters. Preferred atomic position parameters in the terminating region of the lattice are given.

The atomic structure of 1/2 monolayer of Al on GaAs(110) is examined by low energy electron diffraction (LEED) intensity analysis. Room temperature deposition produces a disordered overlayer that does not modify the substrate structure. However, an ordered 1×1 structure results when the system is annealed at 450°C. The AES and LEED measurements suggest a reaction of Al on GaAs to form AlAs, in agreement with some of the photoemission data. As a preliminary result, the multiple scattering analysis indicates that a structure for which Al replaces the second layer Ga provides the most satisfactory description of the measured LEED intensities.

The first determination of the atomic geometry of the (110) surface of InSb is reported. The structure is determined by comparing dynamical calculations of elastic low‐energy electron diffraction (ELEED) intensities with those measured at T=150 K. Initial candidate structures were obtained from an energy‐minimization calculation and from a kinematic analysis of room‐temperature ELEED intensities. Neither candidate proved fully compatible with the measured intensities so another search for the correct structure was performed using dynamical calculations. This procedure yields as the most probable surface structure for InSb(110) one in which the top layer undergoes both a rigid rotation of 28.8° and a 0.05 A contraction toward the substrate. The Sb atoms move outward and the In atoms inward, giving a relative vertical shear of 0.78 A in this uppermost layer. In the second layer, the In moves outward 0.09 A and the Sb inward 0.09 A. The resulting structure is analogous to that of covalently bonded GaAs(110) but different from that of ZnTe(110) which exhibits a more ionic bonding. We infer that highly covalent zincblende‐structure compound semiconductors exhibit reconstructions of their (110) surfaces characterized by (1) large bond rotations in the uppermost layer, with the anion moving outward and cation inward, (2) small bond‐length alterations (i.e., a few percent or less), and (3) distortions from the bulk geometry which penetrate at least two atomic layers in from the surface.

A survey is presented of determinations of the atomic geometries of the (110) surfaces of zincblende‐structure III–V and II–VI compounds via analyses of elastic low‐energy electron diffraction intensities, of both integrated and angle‐resolved ultraviolet photoemission spectra, and of electron paramagnetic resonance signals associated with adsorbed O−2 complexes. Comparison of the proposed model geometries for GaAs(110) reveals that they are compatible both with each other to within the accuracy of the analytical methodologies and with energy‐minimization calculated structures. Results for other materials are fragmentary, but suggest incompatibilities between spectroscopically‐determined and energy‐minimized structures for ZnTe(110) and InSb(110). All of the model surface structures embody large movements of species in the uppermost atomic layer, with the cation moving inward by Δd⊥?0.4 A and the anion outward either less or not at all. The reconstructions penetrate to the second and perhaps third atomic layer for covalently bonded materials like GaAs and InSb. Spectroscopic evidence for the penetration of the reconstructions beneath the top layer is weak, however, for more ionic compounds like ZnTe.

We have studied via LEED, AES, and EELS the (110) and (100) interfaces between GaAs and ZnSe or Se deposited in UHV on sputter‐annealed substrates. These interfaces are reactives. Although nearly stoichiometric (ϑ>20 A) crystalline ZnSe films can be grown on GaAs by congruent evaporation, the ZnSe–GaAs interfaces are nonstoichiometric and include a thin intermediate (Ga, Se) compound layer due to a Se–As exchange taking place during the early stages of growth. Various phases of the (Ga, Se) compound can be obtained, depending on the availability of Se at the interface. The structure and chemical properties of the Se–GaAs interfaces are also investigated and related to the chemical reaction observed during the formation of the ZnSe–GaAs interfaces.

The recent determination of the atomic structure of CdTe(110) and GaSb(110), and of GaAs(110)‐In, GaAs(110)‐Sb, and GaP(110)‐Al has triggered a reevaluation of the atomic structures of compound semiconductor surfaces and interfaces. The atomic geometries of the (110) surfaces are dominated by a top bilayer bond rotation and a small contraction of the bilayer toward the substrate. The magnitude of the rotation angle approximately correlates with the lattice constant of the crystal. No simple correlation can be found between the surface relaxation and the ionicity of the crystal. The long range atomic structure of column III and V elements adsorbed on GaAs(110) and GaP(110) is evaluated, before and after annealing, in terms of adatom electronic configuration and interface reactivity. Column III elements form disordered overlayers whereas column V elements form ordered structures and modify the substrate relaxation. Annealing reactive interfaces (Al on GaAs and GaP) leads to a metal–cation replacement reacti...

The morphology, chemistry, and band bending at the Ca/GaAs(110) interface are studied with Auger electron spectroscopy, electron energy‐loss spectroscopy, low‐energy electron diffraction, x‐ray photoemission spectroscopy, and ultraviolet photoemission spectroscopy. Ca is found to grow nearly uniformly on GaAs at RT. There is a Ca–Ga exchange reaction at the interface. An intermediate position of the Fermi level (EF) resulting from the formation of adsorbate induced states and native defects is found at low coverage at 0.75–0.9 eV above the top of the valence‐band maximum (VBM). An additional shift of EF leading to a final position 0.55 eV above VBM takes place when metallicity develops in the overlayer. The results for this interface support Schottky barrier models based on gap states induced or modified by the metal.

Auger electron spectroscopy and low‐energy electron diffraction are used to study the interface formed by the evaporation of Sb on room temperature GaAs(110). In contrast to Al, the Sb overlayer is ordered and produces a (1×1) diffraction pattern with intensity profiles very different from those measured from the clean substrate surface. The interface is sharp and stable under heat treatment, the Sb film is continuous and thermal desorption experiments reveal the strong bonding that exists between the first Sb monolayer and the substrate.The atomic structure of the GaAs(110)‐p(1×1)‐Sb(1 ML) system is analyzed with multiple scattering computations. Three types of structures have been examined: chains of Sb atoms parallel and antiparallel to the top layer Ga–As chains, Sb2 dimers attached to the surface Ga species, and a ’’jellium’’ type structure in which one Sb is bonded to the surface Ga atom and the other Sb is randomly distributed above the surface. Only single scattering computation, however, has been used for the later model. Qualitative description of the measured intensities are achieved by structures of the first (antiparallel chains) and third models.

A model for the atomic structure of the annealed Ge(111) 2×8 surface is proposed on the basis of intensity measurements in which the fractional order beams are attributed to a ripplelike distortion of (at least) the top two double layers in a manner consistent with charge density wave (Peierls) instabilities or compresive stress deformations derived from distributed rehybridization. The distortion is similar to that suggested for the annealed Si(111) 7×7 surface from a similar analysis. The superlattice is actually an 8×8 unit cell; the observed 2×8 reconstruction follows from selected structure factor cancellations owing to the specific topography of the distortion.

Note: Univ wisconsin madison,dept phys,stoughton,wi 53589. princeton univ,dept elect engn,princeton,nj 08544. Kelly, mk, univ wisconsin madison,ctr synchrotron radiat,stoughton,wi 53589.ISI Document Delivery No.: C8191Part 1 Reference LSE-ARTICLE-1986-011 Record created on 2006-10-03, modified on 2017-05-12

Recent experiments on the formation of Schottky barriers reveal a temperature-dependent pinning of the Fermi level (E F ) which cannot be explained in terms of the current published models. Al, Au, Ag, In, and Sn are evaporated on room temperature (RT) and low temperature (LT= 80 K) cleaved n- and p-GaAs(110). We find a substantial decrease in surface metal clustering at LT, leading to more homogeneous films, especially with Al, In, and Sn. Interface mixing and chemical reaction are also partly inhibited. Cancellation of the semiconductor surface relaxation is observed for Al and Sn. From the electronic point of view, low temperature has opposite effects on the pinning rate of E F on n- andp-GaAs. Pinning on n-GaAs as a function of coverage is dramatically retarded. Pinning on p-GaAs remains faster than on n-GaAs, as fast as at RT, faster for Al and In. These asymmetric rates make it impossible to explain the initial stages of Schottky barrier formation for all substrates with a single mechanism. They suggest that several independent but concomitant mechanisms might be at work in the pinning process.

The Al–GaP(110) and Al–InP(110) interfaces are characterized with low energy electron diffraction and Auger electron spectroscopy. Both interfaces are reactive, i.e., the replacement of the cation by Al is thermodynamically favorable. However, Al evaporated on the room temperature substrates has a high surface mobility and forms metallic islands which leave the atomic structure of the uncovered semiconductor areas unperturbed. Limited Al–substrate reaction can be detected at room temperature. Moderate annealing triggers a large scale Al‐cation replacement reaction and leads to the formation of AlP layers. The crystalline quality of the AlP layers is good for Al on GaP (0.01% lattice mismatch between AlP and InP) and poor for Al on InP (7% lattice mismatch between AlP and InP). The atomic geometry of AlP(110) is analyzed with multiple scattering LEED computation. A best fit structure is obtained which is consistent with the correlation recently obtained between atomic structure and atomic size on (110) semiconductor compound surfaces.

The surface structure of monocrystalline ZnSe(110) and of 4-5-nm-thick ZnSe(110) layers epitaxially grown on GaAs(110) is investigated by means of elastic LEED and AES; the results are analyzed using the computer programs and R-factor methods of Duke et al. (1981 and 1983), presented in graphs and tables, and compared to those for GaAs(110). Significant differences are attributed to bond-length-conserving outward rotation of Se and inward rotation of Zn in the top layer, with an angle of 4 deg between the actual plane of the cation-anion chain and the truncated bulk surface. The R intensities measured for ZnSe(110) and GaAs(110) are given as Rx = 0.22 and RI = 0.21 and Rx = 0.24 and RI = 0.16, respectively.

We use soft x‐ray photoemission spectroscopy to study the Ga/GaAs and Mn/GaAs interfaces formed at room temperature (RT) and at low temperature (LT=80−100 K). The LT Ga/GaAs interface exhibits the strongly inhibited clustering observed at other nonreactive metal/GaAs interfaces. The Mn–GaAs reaction is slowed somewhat. At both interfaces, the effect of temperature on EF movement is in accord with the trends previously set: initial band bending on n‐GaAs is slowed at LT, whereas band bending on p‐GaAs is not affected significantly. For both Ga and Mn/GaAs, EF movement at low coverages is consistent with the explanation proposed for the sp and noble metals: introduction of donor states from, e.g., the specific adsorbate levels. Studies of the noble metal interfaces with GaAs have showed that the establishment of the final EF position is related to the appearance of metallicity at the interface; Mn/GaAs provides direct evidence for that interpretation.