Unreconstructed Si Research Articles

Ultrathin (In, Mg) films on Si(111): A nearly freestanding double-layer metal

We report on the formation of triple-atomic-layer metal films on Si(111) by deposition of Mg onto In double-layer films. The deposited Mg atoms are intercalated between the In layers and Si substrate and form a buffer layer with $(\sqrt{3}\ifmmode\times\else\texttimes\fi{}\sqrt{3})$ periodicity on an unreconstructed Si(111) surface. Two metal layers with nearly hexagonal close-packed arrangement are stacked in an ABC sequence on the buffer layer. The (In, Mg) triple-layer phase shows free-electron-like electronic structure with circular Fermi surface. Unlike the single Fermi circle of the In double layers, the Fermi surface of the (In, Mg) triple layers is composed of two circles with different radii. The larger and smaller Fermi circles are found to come from bonding and antibonding states between the top and middle layers of the three layers. The bottom layer acts as a buffer layer to saturate the Si dangling bonds and realize a nearly freestanding double-layer metal.

In Situ Observation of Epitaxial Li-Si-Nanostructure Formation on Si(111)

Silicon is considered as anode material for Li-ion batteries due to its high theoretical capacity with a value of 4200 mAh/g for Li22Si5. The associated volume change exceeding 300 % during de-/lithiation of silicon is considered as the major problem for application with a possible solution in using nanostructured material. Various details of the formation of Li-Si phases are presently under discussion like the formation of amorphous vs. crystalline structures or the Li-Si phase order during lithiation.To shed light on the relevant processes during the Li-Si phase formation a clean Si substrate was exposed under ultra-high condition in a molecular beam epitaxy (MBE) environment to an atomic lithium flux. During the experiment the Si substrate temperature was gradually reduced from initial 700°C down to room temperature. Simultaneously, the onset of the nucleation and the formation of Li-Si phases were observed by in situ synchrotron x-ray diffraction (XRD), azimuthal reflection high-energy electron diffraction (ARHEED), and line-of-sight quadrupole mass spectrometry (QMS) of the desorbing Li. Initial only Si surface reconstructions are observed involving only the top most atomic layer. At lower temperatures, Li-Si nanostructures are formed. First, crystalline Li12Si7 appears, followed by the formation of Li21Si5 phases. Both phases grow epitaxially. The epitaxial relationship is Si[111]||Li12Si7[010] and Si[11-2]||Li12Si7[100] as well as Si[111]||Li21Si5[111] and Si[01-1] ||Li21Si5[01-1]. Further confirmation besides the XRD data is gained by investigating the Li-Si surfaces. The Li21Si5(111) surface exhibits an approximately nine times larger (1x1) surface unit cell than the unreconstructed Si(111) surface. Ex situ investigations of various intermediate stages of the Li-Si phase formation by atomic force microscopy show nanostructures which formed on the substrate surface with a height of about 10 nm and 100 nm diameter much smaller than usually reported in Si anode material. It demonstrates that the formation of the Li-Si phases depends on the Li kinetics as well as the Si release mechanism from the crystalline substrate and its diffusion. Acknowledgements: This work was supported by the Leibniz-Gemeinschaft under the project No. SAW-2011-PDI-230. The experiments are carried out at the synchrotron radiation facility Bessy II (Helmholtz-Zentrum Berlin, Germany).

Theoretical studies of Li incorporation into Si(111)

The incorporation of Li atoms into Si(111) is studied with the aid of density functional theory (DFT). We find that Li adsorption removes the (1 × 2) reconstruction on this surface. For the whole range of chemical potential values of Li at which Li adsorption occurs, only a high coverage structure (1 ML) is stabilized on unreconstructed Si(111). We find for this surface that the simultaneous penetration of two or more Li atoms into the subsurface is very unfavorable. However, the penetration barrier for a single Li atom into the subsurface on 1 ML-Li/Si(111) is very similar to that on clean Si(111). The present work can help us to better understand the process of Li insertion into Si anodes of Li-ion batteries.

First principles analysis of the initial oxidation of Si(001) and Si(111) surfaces terminated with H and CH3

Transition state analyses have been carried out within a density functional theory setting to explain and quantify the distinctly different ways in which hydrogen and methyl terminations serve to protect silicon surfaces from the earliest onset of oxidation. We find that oxidation occurs via direct dissociative adsorption, without any energy barrier, on Si(111) and reconstructed Si(001) that have been hydrogen terminated; oxidation initiates with a barrier of only 0.05 eV on unreconstructed Si(001). The commonly measured protection afforded by hydrogen is shown to derive from a coverage-dependent dissociation rate combined with barriers to the hopping of adsorbed oxygen atoms. Methyl termination, in contrast, offers an additional level of protection because oxygen must first undergo interactions with these ligands in a three-step process with significant energy barriers: adsorption of O(2) into a C-H bond to form a C-O-O-H intermediate; decomposition of C-O-O-H into C-O-H and C=O intermediates; and, finally, hopping of oxygen atoms from ligands to the substrate.

Preparation, IR spectroscopy, and time-of-flight mass spectrometry of halogenated and methylated Si(111)

The preparation of chlorine-, bromine-, and iodine-terminated silicon surfaces (Si(111):Cl, Br, and I) using atomically flat Si(111)-(1×1):H is described. The halogenated surfaces were obtained by photochemically induced radical substitution reactions with the corresponding dihalogen in a Schlenk tube by conventional inert gas chemistry. The nucleophilic substitution of the Si-Cl functionality with the Grignard reagent (CH3MgCl) resulted in the unreconstructed methylated Si(111)-(1×1):CH3 surface. The halogenated and methylated silicon surfaces were characterized by Fourier transform infrared (FTIR) spectroscopy and laser-induced desorption of monolayers (LIDOM). Calibration of the desorption temperature via analysis of time-of-flight (TOF) distributions as a function of laser fluence allowed the determination of the originally emitted neutral fragments by TOF mass spectrometry using electron-impact ionization. The halogens were desorbed atomically and as SiXn (X = Cl, Br) clusters. The methyl groups mainly desorbed as methyl and ethyl fragments and a small amount of +SiCH3.

Investigation of fivefold symmetry at the liquid Pb/Si(001) interface: Anab initiomolecular dynamics study

The structure of liquid Pb on an ideal Si(001) surface was studied experimentally a decade ago by means of x-ray diffraction and the results were interpreted in terms of the presence of fivefold symmetry Pb structures captured transiently by the potential created by the unreconstructed Si(001) surface. We critically analyze this interpretation in the light of the results obtained in an extensive ab initio molecular dynamics study of a system comprising 314 Pb atoms and 175 Si atoms setup in 7 (001) ideal layers (a total number of 1956 valence electrons) in a slab geometry. The structure found for the first Pb layer is very different from that of bulk Pb, mostly consisting in one-dimensional lines. However, we do observe the possibility of forming transient structures, in particular icosahedral caps. We discuss if this result can explain the modulation found experimentally in the diffracted intensity.

Nonresonant formation ofH−near unreconstructedSi(100)surfaces

We calculate ab initio the fraction of outgoing negative hydrogen ions that are normally incident on an unreconstructed Si(100) surface with kinetic energies between 50 and 150 eV. The ground-state electronic structure of the surface is derived from a self-consistent screened Thomas-Fermi-von Weizsaecker pseudopotential including Wang-Teter shell structure corrections. Orbitals and energies of the electronic states in this potential are obtained by solving Kohn-Sham equations. The dynamics of the transfer of a single electron during the ion-surface collision is represented within the Newns-Anderson model, including image-charge interactions and electron translation factor. We show that the outgoing H{sup -} fraction evolves at large distances from the surface due to nonresonant transitions from the valence band levels of the substrate into the affinity level of H{sup -}. In particular, we show that electron capture from dangling-bond surface-state resonances determines the final negative-ion fraction. We find good qualitative agreement with the experimental results of Maazouz et al. [Surf. Sci. 398, 49 (1998)] for the scattering of hydrogen atoms and ions on silicon surfaces, even though our calculations do not include the effects of reconstruction and projectile motion parallel to the surface.

STM observation and first-principles determination of Ge nanoscale structures on Si(111)

Scanning tunneling microscopy observations and first-principles quantum mechanical calculations were employed to investigate the nanoscale structures formed on Si(111) surfaces upon germanium deposition at a coverage of $\ensuremath{\sim}0.3$ monolayer. At room temperature, Ge atoms form nanoclusters with sizes of $1.5--6\phantom{\rule{0.3em}{0ex}}\text{nanometers}$ in width. After annealing, the nanoclusters become two-dimensional islands with typical size of $\ensuremath{\sim}10$ nanometers in width. We propose that the annealing or high-$T$ deposition results in a partial transformation of $(7\ifmmode\times\else\texttimes\fi{}7)$ reconstructed unit cells to unreconstructed Si(111) configurations on which the Ge adatoms reside at the ${T}_{4}$ sites and form a $(\sqrt{3}\ifmmode\times\else\texttimes\fi{}\sqrt{3})R30\ifmmode^\circ\else\textdegree\fi{}$ reconstruction.

Scanning Tunneling Microscopy of Ethylated Si(111) Surfaces Prepared by a Chlorination/Alkylation Process

Scanning tunneling microscopy (STM) and computational modeling have been used to study the structure of ethyl-terminated Si(111) surfaces. The ethyl-terminated surface was prepared by treating the H-terminated Si(111) surface with PCl5 to form a Cl-terminated Si(111) surface with subsequent exposure to C(2)H(5)MgCl in tetrahydrofuran to produce an alkylated Si(111) surface. The STM data at 77 K revealed local, close-packed, and relatively ordered regions with a nearest-neighbor spacing of 0.38 nm as well as disordered regions. The average spot density corresponded to approximately 85% of the density of Si atop sites on an unreconstructed Si(111) surface. Molecular dynamics simulations of a Si(111) surface randomly populated with ethyl groups to a total coverage of approximately 80% confirmed that the ethyl-terminated Si(111) surface, in theory, can assume reasonable packing arrangements to accommodate such a high surface coverage, which could be produced by an exoergic surface functionalization route such as the two-step chlorination/alkylation process. Hence, it is possible to consistently interpret the STM data within a model suggested by recent X-ray photoelectron spectroscopic data and infrared absorption data, which indicate that the two-step halogenation/alkylation method can provide a relatively high coverage of ethyl groups on Si(111) surfaces.

Covalent Attachment of Acetylene and Methylacetylene Functionality to Si(111) Surfaces: Scaffolds for Organic Surface Functionalization while Retaining Si−C Passivation of Si(111) Surface Sites

Si(111) surfaces have been functionalized with Si-CC-R species, where R = H or -CH3, using a two-step reaction sequence involving chlorination of H-Si(111) followed by treatment with Na-CC-H or CH3-CC-Na reagents. The resulting surfaces showed no detectable oxidation as evidenced by X-ray photoelectron spectroscopic (XPS) data in the Si 2p region, electrochemical measurements of Si-H oxidation, or infrared spectroscopy. The Si-CC-R-terminated surfaces exhibited a characteristic CC stretch in the infrared at 2179 cm-1, which was strongly polarized perpendicular to the Si(111) surface plane. XPS measurements in the C 1s region showed a low binding energy peak indicative of Si-C bonding, with a coverage that was, within experimental error, identical to that of the CH3-terminated Si(111) surface, which has been shown to fully terminate the Si atop sites on an unreconstructed Si(111) surface. The Si-CC-H-terminated surfaces were further functionalized by exposure to n-C4H9Li followed by exposure to para Br-C6H5-CF3, allowing for introduction of para -C6H5CF3 groups while maintaining the desirable chemical and electrical properties that accompany complete Si-C termination of the atop sites on the Si(111) surface.

3 keV Ar scattering from unreconstructed Si(1 0 0) at grazing incidence: Molecular dynamics simulation

In this paper, grazing scattering of 3keV Ar along the [100] direction of an ideal (unreconstructed) Si(100) surface is simulated using the molecular dynamics (MD) method. A detailed analysis of the scattering trajectories is performed. The scattering process in many cases involves a sequence of simultaneous collisions between projectile and target atoms. The scattered intensity is presented as a function of the polar scattering angle, the azimuthal scattering angle and the energy loss. Various prominent features of the spectra are assigned to specific scattering trajectories.

Atomistic-to-continuum description of vacancy cluster properties in crystalline silicon

A synergistic combination of molecular dynamics and statics calculations based on the empirical Environment-Dependent Interatomic Potential (EDIP) is used to compute the thermodynamic properties of vacancy clusters (voids) in silicon. All cluster formation properties are found to follow a simple size scaling law, leading to a compact expression for void free energies. An estimate for the free energy of the unreconstructed Si (111) surface is found to compare well with experimental measurements. The results should be useful for the development of accurate process simulators for void formation during crystal growth and wafer thermal annealing.

X-ray photoelectron spectroscopy studies on hexadecafluoro-copper-phthalocyanine ultrathin films deposited onto Si(100) 2×1

We have deposited in ultrahigh vacuum conditions very thin films of hexadecafluoro copper phthalocyanine onto both a clean, reconstructed Si(100) 2×1 substrate and a clean but unreconstructed Si(100) surface. We have studied the organic–inorganic interface by means of monochromatic X-ray photoemission spectroscopy. The measurements show an effective interaction between the molecule and the silicon substrate, in particular for a well-reconstructed silicon surface. The results suggests the important role of the negatively charged sites on the Si (100) 2×1 surfaces in the silicon–molecule interaction.

Oxidation and Stabilization of Unreconstructed Hydrogen- and Fluorine-Terminated Si(100) Surface: A Periodic Density Functional Study

The understanding and control of silicon surfaces is of great importance in the production of silicon-based electronic devices made from semiconductor materials and constructed on silicon single-crystal substrates. This is the first study to see the effect of surface etching by fluorine on the stability of the unreconstructed Si(100) surface using periodic density functional calculations. All the possible fluorine substitution sites are considered, and the results are compared with the existing experimental observation in terms of suitability of fluorine substitution on the surface. The results are compared with H-terminated surfaces to prove the efficiency of fluorine in stabilizing the unreconstructed Si(100) surface. Oxidation of the Si(100) surface for both H- and F-terminated surfaces were also studied to propose the plausible mechanism of oxidation. Two kinds of experimental situations were mimicked for oxidation; (1) oxidation on the surface, i.e., generation of the Si−O−H bond, and (2) oxidation on the bridging bond between two silicons, i.e., generation of Si−O−Si, were compared to show the feasibility of Si−O−Si bond formation. Oxidation through atomic oxygen was followed throughout the calculation. The calculations were performed with smaller clusters using density functional methodology to validate and rationalize the current understanding. The results were further supported by molecular electrostatic potential maps generated from periodic density functional calculation results.

Dynamic repulsion of surface steps during step flow etching: Controlling surface roughness with chemistry

The development of surface roughness during step flow etching of vicinal one- and two-dimensional surfaces was investigated using kinetic Monte Carlo simulations. The two-dimensional simulations were performed on an unreconstructed Si(111) lattice and were designed to model aqueous silicon etching. In one dimension, the etching of a surface with noninteracting steps leads to a progressive roughening of the surface. In contrast, steps on a two-dimensional surface dynamically repel one another during etching, even when the steps are noninteracting. Dynamic step–step repulsion is the consequence of a feedback mechanism that is induced by step collisions and mediated by the morphology of the etching step and the presence of etch pits. The stability of this feedback mechanism to morphological perturbations is analyzed. During step flow etching, feedback can be enhanced either by increasing the roughness of the steps or the density of pits on the etched surface. Under certain conditions, etching can produce step distributions that are more regular than the corresponding equilibrium distribution. The implications of this observation on the production of very flat surfaces by etching are discussed.

Substitutional geometry and strain effects in overlayers of phosphorus on Si(111)

The structure and bonding topology of phosphorus adatoms on Si(111) surfaces have been investigated by scanning tunneling microscopy (STM) and scanning tunneling spectroscopy, in conjunction with low-energy electron diffraction and Auger electron spectroscopy. At room temperature P adatoms substitute for the Si adatoms of the $\mathrm{Si}(111)7\ifmmode\times\else\texttimes\fi{}7$ surface at low coverages as revealed by the chemical contrast between P and Si adatoms in the filled-state STM images. A statistical evaluation of the STM images within the framework of a simple reaction model suggests that the corner adatoms of the faulted half of the $(7\ifmmode\times\else\texttimes\fi{}7)$ unit cell act as primary reaction centers for the ${\mathrm{P}}_{2}$ molecules from the gas phase, in agreement with theoretical expectation. At elevated temperature (500--650 \ifmmode^\circ\else\textdegree\fi{}C) a $(6\sqrt{3}\ifmmode\times\else\texttimes\fi{}6\sqrt{3})R30\ifmmode^\circ\else\textdegree\fi{}$ structure is the prevailing P-induced surface reconstruction. Atomically resolved STM images show that this structure is a domain-wall structure containing hexagonal domains that tesselate the entire surface. The 6% contracted $(1\ifmmode\times\else\texttimes\fi{}1)$ phosphorus domains are separated by straight, 4% expanded light domain walls of irregular lengths. The structure displays a complex in-phase---antiphase---stacking fault relationship between adjacent domains, which has been modeled successfully with a P adlayer in a substitutional bonding geometry on an unreconstructed Si(111) surface. The large tensile surface stress introduced by the P-Si bonding is responsible for the domain-wall formation and precludes the formation of a global $(1\ifmmode\times\else\texttimes\fi{}1)$-P structure on Si(111) surfaces.

Reflectance anisotropies of silicon surfaces: analysis of spectra in terms of surface excess susceptibilities

Reflectance anisotropy (RA) spectra of semiconductor surfaces provide a measure of the contribution of the outermost atomic layers of the solid to the optical reflectivity. Experimental RA spectra are not yet well understood and a suitable means of analysis is needed. Interpretation of features of RA spectra in terms of the surface excess susceptibility is argued to be preferable to interpretation in terms of transitions between occupied and unoccupied states at the surface. RA spectra of the unreconstructed Si(001) and Si(110) surfaces are calculated using the discrete dipole method and these are analysed in terms of surface excess susceptibilities and surface polarization.

Reflectivity and reflectance anisotropy of Si(100): a polarisable bond model

The reflectivity and reflectance anisotropy (RA) of the unreconstructed Si(100) surface are calculated using a model in which the semiconductor is a slab of polarisable bonds interacting by induced dipolar fields. Parameters for the calculations are obtained from ab initio calculations. The dielectric response of the slab is analysed in terms of dipole-moment normal modes which are called dipole waves. At the \\ ̄ gG point of the surface Brillouin zone, dipole waves are polarised either parallel or perpendicular to the surface. We find that the normal-incidence reflectivity of the dipole lattice reproduces the magnitude of the experimental reflectivity of silicon in the energy range 1–6 eV but lacks spectral features. Local fields near the surface of the slab differ from the local field in the bulk, and consequently the surface of the slab responds to optical radiation at a slightly higher frequency than the bulk, and the slab has a maximum RA of the order of 0.7%. A comparison is made between the theoretical RA spectrum and the experimental RA spectra of the Si(100)-(2 × 1) and Si(100)-(2 × 1)-As surfaces.

Visualization of Tight‐Binding Calculations–The Electronic Structure and Electron Localization of the Si(100) Surface

AbstractThe advantage of computer graphics in the visualization of tight‐binding calculations is highlighted in a model study of the reconstruction of the Si(100) surface. Three different surface models—the unreconstructed surface Si(100)‐(1 × 1), and symmetric and asymmetric pairing of surface atoms Si(100)‐(2 × 1)—are investigated on the basis of density of states (DOS), local (projected) density of states (LDOS) and crystal orbital‐overlap population (COOP) analysis. For the visualization of the real‐space properties of tight‐binding calculations, two‐ and three‐dimensional images of the total (TED) and partial electron densities (PED) are shown. The PED calculated near the Fermi level are compared to densities of HOMOs and LUMOs in molecular systems and used to analyse constant current mode STM images, obtained by applying bias voltages of different sign. They show excellent agreement with STM experiments. The electron‐localization function (ELF) has been shown to describe chemical bonds in molecules and solids surprisingly well. Here, the ELF is calculated for surfaces. In order to visualize the shape of the “dangling” surface bonds and bonds connecting surface atoms, two‐ and three‐dimensional representations of the ELF are discussed. Using the reconstruction of the Si(100) surface as an example, we show that combining methods for extracting information from quantum mechanical calculations, such as PED, TED and ELF, leads to a more comprehensive description of the electronic surface structure. With the help of computer graphics, chemical concepts routinely used for describing local properties of molecules can be transferred very effectively to extended systems.

Tensor LEED for the approximation of nondashspherical atomic scattering

We present a new approach to consider nondashspherical atomic charge densities for the computation of intensities in low energy electron diffraction (LEED). This might be important for covalently bonded atoms for example in adsorbed molecules, semiconductors or compound materials. In a first step the nondashspherical charge distribution of a surface atom is approximated by Hartree-Fock calculations applied to a small atomic cluster surrounding the atom under consideration. Then, starting with a LEED reference calculation for spherical scatterers as usual, the deviation from the spherical charge distribution is taken into account by means of the corresponding change of the atomic scattering matrix which becomes nondashdiagonal. Eventually, this is made to enter the perturbation scheme tensor LEED. This procedure allows easy access to trial structures with realistic atomic scattering. We applied the method to an example of practical importance, i.e. to the unreconstructed Si(100) surface. A rough estimation neglecting multiple scattering between the nondashspherical and spherical part of the atomic potential proves, however, that the influence of nondashspherical scattering on intensities in a typical energy range (20–150 eV) is practically negligible.