LEED,Auger and plasmon studies of negative electron affinity on Si produced by the adsorption of Cs and O

We have made LEED, Auger, and Plasmon measurements to study how Cs and O adsorb onto the (100) surface of p-type degenerate Si to produce negative electron affinity (NEA). A key factor to producing NEA was found to be a highly ordered Si surface as reflected by very high quality 2×2 LEED patterns. When NEA is produced, both the adsorbed Cs and O give the same LEED pattern as the original Si surface, but with a general enhancement of the half-order spot intensity. The adsorption of both Cs and O is strongly self-limiting, apparently controlled by the number of available appropriate sites on the surface. If Cs and O adsorb amorphously, NEA is not achieved. The thermal desorption of Cs occurs over a fairly broad temperature range centered at about 550°C. After Cs desorbs, the remaining O reverts spontaneously from an ordered layer to an amorphous layer, and then desorbs at about 800°C with an activation energy of 3.3 eV.Measurements of backscattered electron energy losses due to plasmons have shown that the Si surface plasmon is reduced in energy from 12.5 V to 7.0 V by the Cs-O layer. From this, an effective dielectric constant ε = 5.3 for this layer can be deduced which, in turn, enables us to characterize completely the Cs-O dipole layer.The geometrical model described by Levine for the NEA surface is consistent with our experimental results.     

LEED-Auger characterization of GaAs during activation to negative electron affinity by the adsorption of Cs and O

A combination of low energy electron diffraction (LEED) and Auger electron spectroscopy (AES) has been used to study the formation of the negative electron affinity (NEA) condition on surfaces of p-type, degenerate, (100) and (111) GaAs. Activation to NEA is achieved by adsorbing Cs and O onto atomically clean GaAs in repetitive cycles of first Cs and then O. Before activation, the clean GaAs surfaces exhibit their characteristic LEED patterns. However, once obtained, there is no significant correlation between the quality of these LEED patterns and the final activation. The adsorption of both Cs and O during activation to NEA is amorphous. Auger measurements have shown that the first photoemission maximum occurs after the adsorption of about a half monolayer of Cs. The initial O adsorption occurs on the GaAs surface between the Cs atoms. The adsorbed O interacts strongly with Cs at any stage during the activation. Peak photosensitivities, after completion of the Cs and O adsorptions, were in the range 400 to 1100 $\text{μA}\text{lumen}$. The final activation does not correlate with the quantity of Cs and O on the surface. The temperature dependence of the photosensitivity of NEA GaAs (100) activated at ?170°C has a broad maximum at about ?50°C and a subsidiary maximum at about 160°C. In addition, the photoemission at ?170°C can be either increased or decreased by having heated the sample up to 200°C, even though no Cs or O desorption has taken place. These results can be traced to changes in work function rather than to changes in bulk properties. While the LEED patterns from clean GaAs show no structural changes with temperature, such changes are observed when Cs is on the surface. It is suggested that changes both in photoemission and in LEED patterns are due to the temperature-induced mobility of Cs on GaAs. An atomic model for the NEA surface is discussed in terms of a layer of Cs and O atoms about 10 Å thick on the GaAs.     

Plasmons in monolayer Na,K and Rb films adsorbed on Ni(100)

Electronic excitations in denser monolayer Na, K and Rb films and Na duolayer films adsorbed on a Ni(100) surface have been investigated using Electron Energy Loss Spectroscopy (EELS). Lateral adatom distributions were monitored by LEED. Angular integrated EEL spectra from the ordered c(2 × 2)Na, coverage θ = 0.5, and the ordered hexagonal structures of K and Rb, θ = 0.29, show prominent losses at 3.1, 1.9 and 1.7 eV, of presumably collective nature. The loss energies shift with coverage as ∝ θ^0.4 and as ∝ θ^0.8 for the Na and K, Rb respectively. Angular resolved EEL spectra indicate an only weak dependence of the loss energies on the momentum transfer, Q. In particular the K and Rb losses seem to pass through shallow energy minima, which is predicted by the “box model”. Low energy losses observed at ?1.3 and ?1.0 eV for the c(2 × 2)Na and the hexagonal K and Rb, respectively, are tentatively identified with interband excitations. The observed interband energies yield, when introduced in the “box model”, 3.1., 2.3 and 2.4 eV for the Na, K and Rb, Q = 0 plasmon energies, which is in fair agreement with the observed plasmon loss energies.     

梯度掺杂与均匀掺杂GaN光电阴极的对比研究

为了提高负电子亲和势(NEA)GaN光电阴极的量子效率,利用金属有机化合物化学气相淀积(MOCVD)外延生长了梯度掺杂反射式GaN光电阴极,其掺杂浓度由体内到表面依次为1×10¹⁸ cm^-3,4×10¹⁷ cm^-3,2×10¹⁷ cm^-3和6×10¹⁶ cm^-3,每个掺杂浓度区域的厚度约为45 nm,总的厚度为180 nm.在超高真 关键词： NEA GaN光电阴极 梯度掺杂 量子效率 能带结构     

Growth of extra-thin ordered aluminum films on Si(110) surfaces

The formation of surface structures upon Al deposition onto a Si(110) surface was studied by LEED and AES. The “4 × 6”, “1 × 9”, 2 × 1, 1 × 1 ordered Si(110)-Al surface phases and epitaxial Al(110) domains were observed depending on Al coverage and substrate temperature. The formation phase diagram was drawn for the Al/Si(110) system.     

Electronic properties of alkali metal/silicon interfaces: A new picture

Adsorption of Na and Cs on the Si(100)2 × 1 surface in the monolayer range is investigated by core level and valence band photoemission spectroscopy using synchrotron radiation. The alkali metals are found to induce an electronic interface state near the Fermi level while hybridization between alkali adsorbate “s” and silicon substrate “3p” valence electrons occurs. These results provide evidence that the alkali metal/silicon bonding is covalent. This covalent bond is weak and polarized while plasmon at the alkali metal core level indicates adsorbate rather than substrate metallization.     

A LEED study of the Si{100}(1 × 1)H surface structure

Extensive LEED intensity-energy data have been collected for a Si{100}(1 × 1)H surface and dynamical theory LEED calculations have been performed using an ideal unreconstructed Si(100) surface to model this structure. Agreement between experiment and theory is good indicating that the probable structure for this surface does involve (weakly scattering) H atoms on the “dangling bonds” of an unreconstructed Si(100) surface, and that difficulties in achieving good agreement between experiment and theory for the clean Si{100}(2 × 1) surface is more probably due to deficiencies in the model structure than to deficiencies in the non-structural aspects of the LEED theory.     

The GaP(001) surface and the adsorption of Cs

GaP(001) cleaned by argon-ion bombardment and annealed at 500°C showed the Ga-stabilized GaP(001)(4 × 2) structure. Only treatment in 10^?5 Torr PH₃ at 500°C gave the P-stabilized GaP(001)(1 × 2) structure. The AES peak ratio $\text{P}\text{Ga}$ is 2 for the (4 × 2) and 3.5 for the (1 × 2) structure. Cs adsorbs with a sticking probability of unity up to 5 × 10¹⁴ Cs atoms cm^?2 and a lower one at higher coverages. The photoemission measured with uv light of 3660 Å showed a maximum at the coverage of 5 × 10¹⁴ atoms cm^?2. Cs adsorbs amorphously at room temperature, but heat treatment gives ordered structures, which are thought to be reconstructed GaP(001) structures induced by Cs. The LEED patterns showed the GaP(001)(1 × 2) Cs structure formed at 180°C for 10 h with a Cs coverage of 5 × 10¹⁴ atoms cm^?2, the GaP(001)(1 × 4) Cs formed at 210°C for 10 hours with a Cs coverage of 2.7 × 10¹⁴ atoms cm^?2, the GaP(001)(7 × 1) and the high temperature GaP(001)(1 × 4), the latter two with very low Cs content. Desorption measurements show three stability regions: (a) between 25–150°C for coverages greater than 5 × 10¹⁴ atoms cm^?2, and an activation energy of 1.2 eV; (b) between 180–200°C with a coverage of 5 × 10¹⁴ atoms cm^?2, and an activation energy of 1.8 eV; (c) between 210–400°C with a coverage of 2.7 × 10¹⁴ atoms cm^?2, and an activation energy of 2.5 eV.     

A LEED crystallographic analysis for the Rh(111)-(3 × 3)30°-S surface structure

An intensity analysis with low-energy electron diffraction is reported for the $\text{(}\text{3}\text{×}\text{3}\text{)30°}$ surface structure obtained by the adsorption and presumed dissociation of H₂S on the (111) surface of rhodium. Intensity-versus-energy curves were measured with a video LEED analyser for nine diffracted beams at normal incidence, and comparisons made with the renormalised forward scattering method for four different types of structural models in which the metal atoms remain in their regular bulk positions. The best correspondence between the experimental and calculated intensities occurs with sulphur atoms adsorbed in the “expected” 3-coordinate adsorption sites. The reliability index proposed by Pendry is minimised with S atoms 1.53 Å above the topmost metal layer; this corresponds to nearest-neighbour RhS bond distances equal to 2.18 Å. Comparisons are made with structural data available for related systems, and with the predictions of a model analysis of surface bond lengths given recently by one of the authors.     

Optical properties of dangling-bond states at cleaved silicon surfaces

The spectral dependence of surface photovoltage and surface photoconductance both under continuous illumination as well as LEED I/V spectra were studied with cleaved Si(111)-2 × 1 surfaces at 130 K. Between 0.23 and 0.5 eV a doubly peaked absorption band was found with opposite sign compared to the SPV and SPC signals at higher photon energies. This band is due to electronic transitions from occupied to empty dangling-bond states located at the raised and the lowered rows of atoms in the 2 × 1 reconstruction, respectively. This absorption shows a pronounced dependence on the polarization of the incident light which correlates with the spatial symmetry of the dangling-bond states. Anneals at up to 500 K remove the low-energy absorption peak and equalize the 2 × 1 reconstruction: The homogeneous Si(111)-2 × 1 structure exhibits a buckling of 0.3 Å and a dangling-bond absorption with a threshold at 0.42 eV and a maximum at 0.47 eV. An anneal at 750 K, forming the 7 × 7 structure, destroys the peak of opposite sign in SPV and SPC and only leaves a broad tail with a threshold of 0.32 eV.     

Study of Ag/Si(111) submonolayer interface: I. Electronic structure by angle-resolved UPS

Angle-resolved ultraviolet photoelectron spectra have been measured for well defined Ag/Si(111) submonolayer interfaces of (1) $\text{Si(111)(}\text{3}\text{×}\text{3}\text{)R30°-Ag}$, (2) “Si(111)(6 × 1)-Ag”, and (3) Ag/Si(111) as deposited at room temperature. Non-dispersive and very narrow (FWHM ～ 0.4–0.5 eV) Ag 4d derived peaks are found at 5.6 and 6.5 eV below the Fermi level for surface (1) and at 5.3 and 6.0 eV for surface (2). Dispersions of sp “binding” states in the energy range between E_F and Ag 4d states have been precisely determined for surface (1). Electronic structures similar to those of the Ag(111) surface, including the surface state near E_F, have been observed for surface (3).     

Further dynamical and experimental LEED results for a clean W<“001”-(1 × 1) surface structure determination

Full dynamical layer-doubling calculations have been made for comparison with precision LEED spectra for the clean W “001”-(l × 1) surface at approximately 470 K. Using 45 beams and 10 phase shifts, multi-layer spacing calculated spectra are critically compared with 12 experimental curves involving 5 different beams and 5 incidence angles. Both visual judgements and a semi-quantitative peak deviation/penalty evaluation yield the same result. A surface-bulk layer spacing of 1.51 ± 0.05 Å is concluded, a 4.4% contraction, in contrast to the most recent other determination of 1.40 ± 0.03 Å. This analysis re-emphasizes the need for a reliable and objective criterion for comparing observed and calculated LEED spectra, and corrects a potentially important input to the analysis of more complex systems. For example, at low temperatures (<370 K) the W “001” clean surface rearranges to a c(2 × 2) structure.     

Electronic states of (2 × 1) and (1 × 1) (111) surfaces of Ge,Si, diamond,GaAs and Ge on Si

The “dangling-bond” surface state dispersion curves, E(k), have been calculated for the (2 × 1) and (1× 1) (111) surfaces of Ge, Si, and diamond, for (1 × 1) GaAs, and for (2 × 1) Ge on Si. The calculations employ the sp³s1 empirical tight-binding model of Vogl et al. and the atomic relaxation of Feder et al. The surface state band gaps are in good agreement with optical-absorption and electron-energy-loss measurements for Ge and Si. For the assumed epitaxial geometry, Ge on Si is predicted to shift the dangling-bond states downward by ≈0.1 to 0.4 eV.     

Leed,aes and photoemission measurements of epitaxially grown GaAs(001), (111)A and (1̄1̄1̄)B surfaces and their behaviour upon cs adsorption

Epitaxially grown GaAs(001), (111) and (1?1?1?) surfaces and their behaviour on Cs adsorption are studied by LEED, AES and photoemission. Upon heat treatment the clean GaAs(001) surface shows all the structures of the As-stabilized to the Ga-stabilized surface. By careful annealing it is also possible to obtain the As-stabilized surface from the Ga-stabilized surface, which must be due to the diffusion of As from the bulk to the surface. The As-stabilized surface can be recovered from the Ga-stabilized surface by treating the surface at 400°C in an AsH₃ atmosphere. The Cs coverage of all these surfaces is linear with the dosage and shows a sharp breakpoint at 5.3 × 10¹⁴ atoms cm^?2. The photoemission reaches a maximum precisely at the dosage of this break point for the GaAs(001) and GaAs(1?1?1?) surface, whereas for the GaAs(111) surface the maximum in the photoemission is reached at a higher dosage of 6.5 × 10¹⁴ atoms cm^?2. The maximum photoemission from all surfaces is in the order of 50μA Im^?1 for white light (T = 2850 K). LEED measurements show that Cs adsorbs as an amorphous layer on these surfaces at room temperature. Heat treatment of the Cs-activated GaAs (001) surface shows a stability region of 4.7 × 10¹⁴ atoms cm^?2 at 260dgC and one of 2.7 × 10¹⁴ atoms cm^?2 at 340°C without any ordering of the Cs atoms. Heat treatment of the Cs-activated GaAs(111) crystal shows a gradual desorption of Cs up to a coverage of 1 × 10¹⁴ atoms cm^?2, which is stable at 360°C and where LEED shows the formation of the GaAs(111) (√7 × √7)Cs structure. Heat treatment of the Cs-activated GaAs(1?1?1?) crystal shows a stability region at 260°C with a coverage of 3.8 × 10¹⁴ atoms cm^?2 with ordering of the Cs atoms in a GaAs(1?1?1?) (4 × 4)Cs structure and at 340°C a further stability region with a coverage of 1 × 10¹⁴ at cm^?2 with the formation of a GaAs(1?1?1?) (√21 × √21)Cs structure. Possible models of the GaAs(1?1?1?) (4 × 4)Cs, GaAs(1?1?1?)(√21 × √21)Cs and GaAs(111) (√7 × √7)Cs structures are given.     

Adsorption of CO on clean and oxygen covered Ni(111) surfaces

Adsorption of CO on Ni(111) surfaces was studied by means of LEED, UPS and thermal desorption spectroscopy. On an initially clean surface adsorbed CO forms a √3 × $\text{√3}\text{R30°}$ structure at θ = 0.33 whose unit cell is continuously compressed with increasing coverage leading to a c4 × 2-structure at θ = 0.5. Beyond this coverage a more weakly bound phase characterized by a $\text{√7}\text{2}$ × $\text{√7}\text{2R19°}$ LEED pattern is formed which is interpreted with a hexagonal close-packed arrangement (θ = 0.57) where all CO molecules are either in “bridge” or in single-site positions with a mutual distance of 3.3 Å. If CO is adsorbed on a surface precovered by oxygen (exhibiting an O 2 × 2 structure) a partially disordered coadsorbate 2 × 2 structure with θ_o = θ_co = 0.25 is formed where the CO adsorption energy is lowered by about 4 kcal/mole due to repulsive interactions. In this case the photoemission spectrum exhibits not a simple superposition of the features arising from the single-component adsorbates (i.e. maxima at 5.5 eV below the Fermi level with O_ad, and at 7.8 (5σ + 1π) and 10.6 eV (4σ) with CO_ad, respectively), but the peak derived from the CO 4σ level is shifted by about 0.3 eV towards higher ionization energies.     

Interaction of O2 with Pt(100): I. Equilibrium measurements

Two newly discovered phases on the Pt(100) surface produced by the adsorption of oxygen have been investigated using Rutherford baekscattering (RBS), nuclear microanalysis (NMA), work function changes (Δφ) and LEED. One phase is associated with the oxygensaturated surface (0.63 ± 0.03 monolayers0.81 × 10¹⁵ O atoms cm^?2), where a very complex LEED pattern is observed; the other is observed at an average coverage of 0.44 ± 0.05 monolayers and gives rise to a (3 × 1) LEED pattern (when observed at room temperature). For both surfaces, RBS measurements indicate large (? 0.025 nm) Pt atom displacements. Also discussed is a new method for preparing the “clean” (1 × 1)-Pt(100) surface without the need for NO adsorption/decomposition.     

The lineshape of the L2,3 VV Auger spectrum of silicon

An attempt is presented to understand the details of the lineshape of the Si L_2,3 VV Auger spectrum from the (111) surface in the 7 × 7 superstructure. In the experiments we have followed the variation of the lineshape induced by adsorption of O₂, H₂O, CO and by bombardment with 3 keV Ar⁺ ions, over a range from a small perturbation of the surface to major changes in surface structure. For small perturbations from the clean surface we were able to resolve changes in the local density of states at surface silicon atoms. By unfolding the experimental spectra, effective transition densities of states result, which compare quite closely with calculated densities of states, apart from a certain enhancement of surface features in the experiments. All peaks in the experimental spectra can be explained, based on densities of states at the surface of pure Si(111) (7 × 7) (91.8 and 84.8 eV), Si(111) + adsorbed oxygen (70.6 eV), SiO₂, (78.9 and 64.5 eV) and plasmon losses, at 71.0 and 57.5 eV for the clean surface.     

The surface structure of Si(100) surfaces using averaged LEED: II. The (2×1) clean surface structure

Constant momentum transfer averaged LEED data from a Si(100) (2 × 1) clean surface structure have been produced for the $\text{(00), (10), (11), (}\text{1}\text{2}\text{0)}$ and $\text{(1}\text{1}\text{2}\text{)}$ beams. All averages show strong features in addition to those attributable to single scattering from the bulk. If this structure is assumed to also originate from single scattering, the surface reconstruction must be deep (>^～4 layers). Alternatively this structure can be ascribed to multiple scattering. As data from the Si(100) (1 × 1)H surface structure produced “good” averaging over an identical range of data, this latter conclusion has considerable bearing on the future usefulness of this averaging approach to surface structure analysis by LEED.     

Structural analogy between GeSi(111)-5 × 5 and Si(111)-7 × 7 surfaces

Low energy electron diffraction (LEED) patterns for the GeSi(111)-5 × 5 surface are reported and compared to those for the Si(111)-7 × 7 surface. Parallels between the observed LEED patterns are explained by a structural analogy between GeSi(111)-5 × 5 and Si(111)-7 × 7 surfaces. Both the (5 × 5) and (7 × 7) patterns are shown to be consistent with structural models of the triangle-dimer type previously proposed for Si(111)-7 × 7 surface.     

Temperature dependence of surface electronic structure of Si(111) surface

Temperature dependence of angle-resolved ultraviolet photoelectron spectra has been obtained for Si(111) surfaces starting with a thermally quenched “1 × 1” surface and ending with a high temperature “1 × 1” surface. it has been found that the surface state at 0.8 eV below the Fermi level exhibits degradation with the increase in temperature, which explains the difference of surface electronic structures between a quenched “1 × 1” surface and a high temperature “1 × 1” surface. Electron correlation effect in a dangling-bond derived surface state is postulated as a cause for the phenomenon.