Low-energy electron diffraction study of the reconstruction and orientational stability of Si(331)

Using low-energy electron diffraction, we have studied the reconstruction of Si(331). We find a previously unreported (12 × 1) reconstruction is characteristic of the clean surface; we present evident that a previously observed (13 × 1) reconstruction is not the equilibrium structure. The (12 × 1) reconstruction disorders via a strongly first-order phase transition at approximately 810°C. Heating carbon contaminated surfaces to about 950°C causes faceting to (111) and {17151} orientations.     

Surface structure evolution during Sb adsorption on Si(1 1 1)–In(4×1) reconstruction

The Sb adsorption process on the Si(1 1 1)–In(4×1) surface phase was studied in the temperature range 200–400 °C. The formation of a Si(1 1 1)–InSb (2×2) structure was observed between 0.5 and 0.7 ML of Sb. This reconstruction decomposes when the Sb coverage approaches 1 ML and Sb atoms rearrange to and (2×1) reconstructions; released In atoms agglomerate into islands of irregular shapes. During the phase transition process from InSb(2×2) to Sb (θ_Sb>0.7 ML), we observed the formation of a metastable (4×2) structure. Possible atomic arrangements of the InSb(2×2) and metastable (4×2) phases were discussed.     

Observation of true c(8 × 2) symmetry in scanning tunnelling microscopy images of the clean InSb(001) surface

Filled and empty state scanning tunnelling microscopy images of the sputtered and annealed InSb(001) surface are presented. The sputter-anneal preparation generates a surface with two distinct phases. The dominant phase possesses a unit cell with true c(8 × 2) symmetry, whereas the other phase is attributed to an asymmetric 1 × 3 reconstruction. The presence of a c(8 × 2) unit cell in filled state images is in contrast to previous reports, which identified only a 4 × 1 unit cell. The true c(8 × 2) symmetry further indicates, the available structural model is used as a guide, that the current interpretation of features in filled state images is incorrect. This result may necessitate a reevaluation of the structural model for the InSb(001)-c(8 × 2) surface.     

Changes of phase and intensity of optical SHG with Ag deposition on Si(111)-7×7 surfaces

The phase and intensity of optical second harmonic generation (SHG) were studied in Ag deposition on the Si(1 1 1) surfaces. The 7×7 reconstruction of the surface changed to the 1×1 structure in room temperature deposition. The SHG intensity reduced and the phase shifted to +180° with this structural change. Meanwhile, the 7×7 reconstruction changed to the reconstruction in 400 °C deposition. With this structural change, the SHG intensity increased and the phase shifted to −250°. An analysis with anharmonic oscillator model showed that the changes in room temperature deposition were due to the quenching of the surface dangling bond states of the 7×7 reconstruction by the Ag deposition, whereas those in 400 °C deposition were due to the appearance of new surface states of the reconstruction.     

The adsorbate-induced removal of the Pt{100} surface reconstruction Part I: NO

The molecular adsorption of NO on both the reconstructed (hex) and unreconstructed (1 × 1) surfaces of Pt{100} has been studied using a combination of infrared reflection-absorption spectroscopy (IRAS) and low energy electron diffraction (LEED) at temperatures between 90 and 300 K. On the (1 × 1) surface at 300 K adsorbed NO gives rise to an N-O stretching band at initially 1596 cm⁻¹ shifting to 1641 cm⁻¹ at a coverage of θ = 0.5. The LEED pattern at this coverage is interpreted in terms of a c(4 × 2) structure in which all the molecules occupy a single type of adsorption site between the on-top and bridge positions. At temperatures below 300 K, a higher coverage disordered phase is observed, giving rise to an N-O stretching band at 1680 cm⁻¹ associated with an on-top NO species. On the (hex) phase surface above 210 K, NO adsorption gives rise to bands characteristic of adsorption on the (1 × 1) phase indicating that the reconstruction is immediately lifted. Below 200 K initial adsorption actually occurs directly on the (hex) phase, resulting in a band at 1680 cm⁻¹, which is assigned to on-top NO. This band increases in intensity until, at a critical coverage dependent on temperature, the (hex) → (1 × 1) surface phase transition is induced. This is indicated by the disappearance of the band at 1680 cm⁻¹ and its replacement by bands characteristic of adsorption on islands of the (1 × 1) structure.     

ESDIAD studies of the structure of TiO2(110)(1 × 1) and (1 × 2) surfaces and interfaces in conjunction with LEED and STM

The TiO₂(110)-(1 × 1) surface and its reconstruction as a (1 × 2) form have been studied with low energy electron diffraction (LEED), electron stimulated desorption ion angular distribution (ESDIAD) and scanning tunnelling microscopy (STM). Oxygen ion desorption occurs within a lobe perpendicular to the (1 × 1) surface, changing to two off-normal lobes for the (1 × 2) reconstruction. This transformation in the ESDIAD pattern is consistent with the added Ti₂O₃ row model of the (1 × 2) reconstruction proposed by Onishi and Iwasawa. STM studies of the stoichiometric and electron irradiated surfaces reinforce the association of the O⁺ ESD contribution with majority sites at the surface. Adsorption of acetic acid on the (1 × 1) surface produces a (2 × 1) overlayed and induces a reconstruction of the underlying substrate. ESDIAD reveals H⁺ ions emitted off-normally from dissociatively adsorbed acetate, and along the surface normal from surface hydroxyls. Adsorption of acetic acid on the (1 × 2) surface does not modify the LEED pattern, but ESDIAD reveals H⁺ desorption with a weaker off-normal contribution consistent with the Ti₂O₃ model of the reconstruction.     

Metastable (1 × 2) and (1 × 3) reconstructions of Pd(110)

By adsorption and subsequent reduction of oxygen on Pd(110), metastable (1 × 2) and (1 × 3) reconstructed surfaces have been produced. Oxygen was not present after the reduction but a small amount of residual hydrogen (< 0.15 monolayers) remained. However this is not the origin of the reconstruction as adsorption of this amount of hydrogen on the clean surface did not cause reconstruction. The structures were stable up to ˜ 370 K, and at higher temperatures they reverted to (1 × 1). These results are compared with Rh(110) where similar reconstructions have been found.     

An STM investigation of the Cu(110)−c(6 × 2)O system

The formation of the c(6 × 2)−O phase on a Cu(110) surface, after completion of the (2 × 1)−O structure, was observed by scanning tunneling microscopy (STM). The phase is composed of isotropic structural elements on fourfold hollow sites of the substrate lattice, which form a quasi-hexagonal array and manifest themselves as large protrusions in the STM images. Individual units of this type are mobile and also represent stable nuclei within a (2 × 1) surrounding. Nucleation is activated and occurs preferentially at steps, in contrast to previous findings with the (2 × 1) phase. Structural implications of additional weak features in high resolution images and of the observed change in two-dimensional density of Cu atoms are discussed.     

Surface structures of S/Ni(111) in (√3 × √3 )R30° and (5√3 × 2) phases studied by surface extended X-ray absorption fine structure spectroscopy

The local surface structures of S/Ni(111) in the ( √3 × √3) R30° and (5√3 × 2) phases have been investigated by means of polarization-dependent sulfur K-edge surface EXAFS. In the (√3 × √3 ) R30° phase, sulfur adatoms are found to occupy threefold hollow sites with a S---Ni distance of 2.13 Å and an inclination angle ω of the Sz.sbnd;Ni bonds at 44° from the surface plane. In contrast, in the (5√3 × 2) phase, it is determined that the Sz.sbnd;Ni bond is longer, 2.18 Å, more inclined, ω = 31°, and that the coordination number is not 3 but 4. These results strongly support a picture involving reconstruction of the top nickel layer to form a rectangular structure. Consideration of several models proposed for the (5√3 × 2) phase leads to one which is compatible with both the present results and results recently reported using STM.     

Well-ordered (1 0 0) InAs surfaces using wet chemical treatments

Atomic ordering of HCl-isopropanol (HCl-iPA) treated and vacuum annealed (1 0 0) InAs surfaces was studied by scanning tunneling microscopy (STM), low-energy electron diffraction (LEED), and reflectance anisotropy spectroscopy (RAS). On the as-treated surface, a diffused (1 × 1) pattern is observed, which successively evolves to the β₂(2 × 4)/c(2 × 8) and (4 × 2)/c(8 × 2) ones after annealing to 330 °C and 410 °C, respectively. At the intermediate temperature of 370 °C, an ₂(2 × 4)/(4 × 2) mixed reconstruction is observed. Reflectance anisotropy spectra are compared with those of the corresponding reconstructions observed after As-decapping and found to be quite similar. Therefore we conclude that high-quality (1 0 0) InAs surfaces can be obtained by wet chemical treatment in an easy, inexpensive and practical way.     

Structure investigation of the nitrogen induced Cu(110)−(2 × 3) phase with 180° low energy impact collision ion scattering spectroscopy

The 180° low energy impact collision ion scattering spectroscopy with detection of noble gas neutrals (180°-NICISS) has been used to investigate the nitrogen saturated Cu(110) surface, which is known to exhibit a (2 × 3) diffraction pattern. The nitrogen induced (2 × 3) phase appears to be the result of a surface reconstruction of a new missing row type, in which every third 100 row of Cu atoms of the first layer is missing. The 180° NICISS patterns further indicate within an accuracy of 0.1–0.2 Å, that the double periodicity in the [1 0] direction is not due to the reconstruction of the Cu surface. Its origin has to be found in the arrangement of the N atoms.     

Spatially resolved dynamics of the TiO2(1 1 0) surface reconstruction

We use low-energy electron microscopy to image the reversible transformation of the TiO₂(1 1 0) surface between a high-temperature 1 × 1 structure and a low-temperature 1 × 2 structure. The reconstruction dynamics are novel: 1 × 2 bands nucleated during cooling at the steps of the starting 1 × 1 surface and then grew laterally from the steps. The transformation kinetics are dominated by mass flow from the surface to the bulk, a process that facilitates converting the high-density 1 × 1 phase to the lower-density 1 × 2 phase. We have also imaged how the 1 × 1 surface reconstructs to 1 × 2 phase after sufficient oxygen is removed from the crystal’s bulk during vacuum annealing. 1 × 2 bands also nucleated and grew laterally from the initial 1 × 1-surface’s steps. However, because this isothermal 1 × 1-to-1 × 2 transition occurs largely by mass redistribution on the surface, the steps of the initial 1 × 1 surface and final 1 × 2 surface are offset. We propose models of mass redistribution during the 1 × 1/1 × 2 phase transition to explain this effect. We conclude that the phase transition is first-order because it always occurred by the nucleation and growth of discrete phases. Finally, we show that quenching can roughen TiO₂’s surface by forming pits and that changing temperature causes step motion on 1 × 2 surfaces.     

Surface reconstructions of the Si(100)–Ge system

The surface reconstruction of epitaxial Ge layer on Si(100) was studied with ultrahigh vacuum scanning tunneling microscopy. The surface with 0.8 ML Ge grown in the presence of a hydrogen surfactant reveals the same structures as found in chemical-vapor-deposited Ge on Si(100): (i) defective (2×1) structure at 290°C, (ii) irregular (2×N) in Ge layer and defective (2×1) in bare Si regions at 420°C, and (iii) (2×N) in Ge-covered regions and c(4×4) in bare Si regions at 570°C. The morphology of step edges does not change with temperature, implying that the c(4×4) reconstruction is anisotropic in nature.     

The Si(001) c(4×4) surface reconstruction: a comprehensive experimental study

We have carried out a comprehensive experimental study of the Si(001) c(4×4) surface reconstruction by scanning tunneling microscopy (STM) (at room temperature and elevated temperatures), Auger electron spectroscopy (AES), reflection high-energy electron diffraction (RHEED) and low-energy electron diffraction (LEED). Si(001) samples were kept under ultra-high vacuum (UHV) at around 550°C until the c(4×4) reconstruction appeared. STM contrast of the c(4×4) reconstruction is strongly influenced by electronic effects and changes considerably over a range of bias voltages.

The c(4×4) surface reconstruction is a result of stress which is caused by incorporation of impurities or adsorbates in sub-surface locations. The resulting c(4×4) reconstruction in the top layer is a pure silicon structure. The main structural element is a one-dimer vacancy (1-DV). At this vacancy, second layer Si-atoms rebond and cause the adjacent top Si-dimers to brighten up in the STM image at low bias voltages. At higher bias voltage the contrast is similar to Si-dimers on the (2×1) reconstructed Si(001). Therefore, besides the 1-DV and the two adjacent Si-dimers, another Si-dimer under tensile stress may complete the 4× unit cell. This is a refinement of the missing dimer model.     

STM investigations of CO coadsorption with O and with S on Ni(110)

The interaction between CO coadsorbed with oxygen and sulfur on Ni (110) has been studied with room temperature STM and LEED. In the case of CO/O/Ni(110)−(θ_o0.3 ML), it is found that due to a large local repulsion between the differing species, the coadsorbed species phase separate into large domains of O−(3 × 1) and CO−p2mg(2 × 1) structure. Similarly in the case of CO/S/Ni(110)−(θ_s0.4 ML), at low local coverages of coadsorbed CO, island segregation of CO and S-c(2 × 2) is observed. At locally saturated CO coverage, the S-c(2 × 2) structure transforms into long -S-S- chains running predominantly along the [ ] direction and separated by a local p2mg(2 × 1)-CO structure; this transformation is attributed to the large CO-CO repulsion in the condensed overlayer structure.     

Hydrogen adsorption on Mo1−xRex(110) (x=0–0.25) surfaces

The effects of adsorbed H on the Mo_1−xRe_x(110), x=0, 0.05, 0.15, and 0.25, surfaces have been investigated using low-energy electron diffraction (LEED) and high-resolution electron energy loss spectroscopy (HREELS). For the x=0.15 alloy only, a c(2×2) LEED pattern is observed at a coverage Θ0.25 ML. A (2×2) pattern is observed for H coverages around Θ0.5 ML from surfaces with x=0, 0.05, and 0.15. Both c(2×2) and (2×2) patterns are attributed to reconstruction of the substrate. At higher coverages, a (1×1) pattern is observed. For the alloy surface with x=0.25, only a (1×1) pattern is obtained for all H coverages. Two H vibrations are observed in HREELS spectra for all Re concentrations, which shift to higher energies at intermediate coverages. Both peaks exhibit an isotopic shift, confirming their assignment to hydrogen. For Re concentrations of x=0.15 and higher, a third HREELS peak appears at 50 meV as H (D) coverage approaches saturation. This peak does not shift in energy with isotopic substitution, yet cannot be explained by contamination. The intrinsic width of the loss peaks depends on the Re concentration in the surface region and becomes broader with increasing x. This broadening can be attributed to surface inhomogeneity, but may also reflect increased delocalization of the adsorbed hydrogen atom.     

Structure and stability of the anatase TiO2 (101) and (001) surfaces

Well-defined (101) and (001) surfaces of anatase TiO₂ were studied for the first time by secondary-electron imaging and low-energy electron diffraction. Initially, both surfaces show a crystalline structure corresponding to the bulk anatase phase, buried under an atomically thin contamination layer. After mild sputtering with 500 eV Ne⁺ ions, we have observed a surface phase transition from tetragonal anatase to face-centered cubic titanium monoxide TiO. Subsequent annealing at 900 K restores the (1×1) anatase structure at the (101) surface. On the (001) surface, however, a (1×4) reconstruction appears. The unreconstructed structure can be recovered by exposing the surface to oxygen. These observations demonstrate the stability of the anatase surfaces and illustrate the feasibility of preparing and investigating clean surfaces of this technological important form of TiO₂.     

Hydrogen induced structure changes of GaAs(1 0 0) surfaces

The interaction of hydrogen with GaAs(1 0 0) surfaces has been studied at room temperature by means of high-resolution electron energy-loss spectroscopy (HREELS), low-energy electron-diffraction (LEED), and Auger electron spectroscopy (AES). Sample cleaning by Ne⁺ ion bombardment (500 eV) and annealing resulted (with increasing temperatures) in “1 × 1”, 1 × 6, 4 × 1, 4 × 6 and c(8 × 2) LEED structures. As a function of hydrogen exposure, the intensity ratio of the stretching vibrations (As-H/Ga-H) is shown to be characteristic for the specific reconstructed surface. In particular the arsenic hydride concentration gradually decreased in all cases. In addition, an initial weakening of the fractional order LEED spots occurred with increasing hydrogen exposures. Finally, a strong 1 × 2 structure was observed independent of the reconstruction we started with. Simultaneously, a shift of the energetic position of the Ga-H stretching vibration occurred.     

The mesoscopic and microscopic structural consequences from decomposition and desorption of ultrathin oxide layers on Si(100) studied by scanning tunneling microscopy

The spatially inhomogeneous decomposition and desorption reaction of oxide layers with coverage 1-0.3 monolayers (ML) from a silicon (100) surface has been studied using scanning tunneling microscopy (STM). After desorption, microscopic changes to the (2 × 1) reconstruction produce two variations on the dimer row reconstruction with decreased surface atom density. A (2 × n) vacancy chain reconstruction and a c(4 × 4) incomplete row reconstruction were observed; a structure for the latter is proposed. Both reconstructions are metastable, reforming the (2 × 1) reconstruction upon heating. At greater length scales during desorption from an initial 1.0 ML coverage, the mesoscopic changes to the surface structure include pitting and roughening, with up to a measured 20 fold increase in the edge density as compared to the clean Si(100) surface.

These structural changes suggest a reaction mechanism involving a substantial rearrangement of the substrate silicon. From an initial 1.0 ML oxygen coverage, using measured void size distributions at total desorption levels of 13% and below — before voids have begun to coalesce — the evolution of void sizes during initial desorption can be followed. A mechanism for desorption is proposed in which silicon atoms must diffuse from adjacent clean surface area to the oxide boundary, producing a reactive complex from which SiO is desorbed. Void growth rates derived from two rate limiting cases for this desorption reaction mechanism can be compared to measured results. We show that the measured void area evolution is consistent with a reaction mechanism where the rate limiting step for monolayer desorption is the promotion of a silicon atom in a lattice site to a mobile monomer within the void.     

Surface reconstructions and phase transitions on the GaAs(111)B surface

The (111)B surface of GaAs has been investigated using scanning tunneling microscopy (STM) and a number of different reconstructions have been found at different surface stoichiometries. In accordance with electron diffraction studies, we find the series (2 × 2), (1 × 1)_LT, ( ) and (1 × 1)_HT with increasing annealing temperature, corresponding to decreasing surface As concentration. The (1 × 1)_LT is of particular interest, since it only occurs in a narrow temperature window between the two more established reconstructions, the (2 × 2) and the ( ). We find the (1 × 1)_LT to take the form of a mixture of the local structures of both the (2 × 2) and ( ) phases, rather than having a distinct structure. This is behaviour consistent with a kinetically limited system, dominated by the supply of As adatoms to the surface, and may be an example of a continuous phase transition. Above the (1 × 1)_LT transition, atomic resolution images of the ( ) surface reveal only a three-fold symmetry of the hexagonal structural units, brought about by inequivalent surface bonding due to the 23.4° rotation of the surface unit cell relative to the substrate. This is responsible for the disorder found in the ( ) reconstruction, since the structure may form in one of two domains. At lower surface As concentration, the (1 × 1)_HT surface adopts a structure combining small domains of a 19.1° structure and random disorder. There is no apparent similarity between the (1 × 1)_LT and (1 × 1)_HT structures, which may be due to our measurements being conducted at room temperature and without an As flux to control the surface As concentration.