Polarization dependence of angle-resolved photoemission spectroscopy of graphite

We have used variable polarization synchrotron radiation to map the valence band electronic structure of graphite by angle-resolved photoemission spectroscopy (ARPES). The experimental results with two orthogonal linear polarization of light signifies the contribution of either even or odd symmetry with respect to the crystal mirror plane towards the photoemission intensity. The σ₁ and σ₂ valence bands show odd reflection symmetry while the π valence band shows even symmetry with respect to the mirror plane. The measured ARPES spectrum using left and right circular polarized lights shows asymmetry in intensity around M point of the Brillouin zone, which ultimately mimicking different partial wave character of σ₁ and σ₃ bands.     

Thin graphite overlayers: Graphene and alkali metal intercalation

Using LEED and angle resolved photoemission for characterisation we have prepared graphite overlayers with down to monolayer thickness by heating SiC crystals and monitored alkali metal intercalation for the multilayer films. The valence band structure of the monolayer is similar to that calculated for graphene though downshifted by around 0.8 eV and with a small gap at the zone corner. The shift suggests that the transport properties, which are of much present interest, are similar to that of a biased graphene sample. Upon alkali metal deposition the 3D character of the π states is lost and the resulting band structure becomes graphene like. A comparison with data obtained for ex situ prepared intercalation compounds indicates that the graphite film has converted to the stage 1 compounds C₈K or C₈Rb. Advantages with the present preparation method is that the graphite film can be recovered by desorbing small amounts of alkali metal and that the progress of compound formation can be monitored. The energy shifts measured after different deposits indicate that saturation is reached in three steps. Our interpretation is that in the first the alkali atoms are dispersed while the final steps are characterized by the formation of first one and then a second (2 × 2) ordered alkali metal layer adjacent to the uppermost carbon layer.     

Cluster growth of Cu on graphite: XPS,Auger and electron energy loss studies

Auger, XPS and EELS techniques have been used to investigate the core levels, the d-valence band and the electronic transitions of different UHV deposited Cu clusters on graphite. The decreasing of the Cu particle size produces core levels and valence band shifts towards higher binding energies. Lower extra-atomic screening of the conduction electrons near the excited atom and shift of the d-band towards the isolated atom levels are claimed to explain these effects. The EELS results suggest that, for smallest clusters, no structural change but only a lattice parameter contraction of the f.c.c. cage occur.     

The nature of the adsorption bond between graphite islands and iridium surface

To reveal the nature of adsorption bonds between two-dimensional graphite islands and iridium (111) and (100) faces, a study has been made of the adsorption of potassium and cesium atoms on the surface of these systems, using thermal desorption and Auger electron spectroscopy, as well as surface ionization and thermionic emission techniques. The graphite islands are shown to be weakly bound to the iridium substrate by Van der Waals forces. The unsaturated valence bonds at the periphery of the graphite islands are “lowered down” on to the metal. The recess between the graphite layer and the metal is filled by adsorbing particles through defects in the graphite layer. The atoms can penetrate into the recess in two ways: at T > 1000 K directly from the flux incident on the surface, and at T < 1000 K also by migration from the graphite island surface. The adsorption capacity of this state is ～ (2?3) × 10¹⁴cm^-2. Thermal destruction of the islands at T > 1900 K liberates the potassium and cesium atoms from under the graphite islands. Our study suggests that the reason for the “raised” position of the islands lies in the valence bonds of the graphite layer being saturated, the valence bonds of the metal and its crystallographic orientation being less significant. Therefore one may expect the graphite layer to be raised also above other metals as well. The filling by cesium of the recess between the graphite layer and iridium and of the adsorption phase on the graphite surface, does not change the general “graphitic” shape of the carbon Auger peak. This cesium results, however, in a pronounced splitting of the negative spike on the carbon peak (which provides information on its location relative to the graphite layer) indicating the appearance in the valence band of graphite near the Fermi level of two narrow (～ 2?3 eV) regions with an enhanced density of states originating from the presence of the alkali metal.     

Evidence for an alkali-like conduction band in alkali graphite intercalation compounds

The valence bands of pure graphite and several alkali graphite intercalation compounds (AGIC's) were studied by UPS (hv = 21.2 eV). The most significant observation is an intensity peak at the Fermi energy E_F in the intecalation compounds. This peak is mainly due to alkali-like s-states. The density of states at E_F is enhanced by a factor of 30 compared to pure graphite. The alkali-like conduction bands in the first stage AGIC's are similar to those of pure alkali metals.     

Valence and core state modifications in Pt3Ti

XPS data for the valence band, the Pt 4? states, and the Ti 2p states are presented for the intermetallic Pt₃Ti. Relative to the Pt valence band, the Pt₃Ti band shows a decrease in the density of states just below the Fermi level and a shift of the centroid to higher (binding) energy. Ti 2p and Pt 4? binding energies showed relatively large shifts with respect to the pure metals. These changes in the valence band density of states and core level binding energies are interpreted as arising from hybridization of the d-orbitals in both metals to form strong intermetallic bonds.     

A comparison of electronic properties of various modifications of graphite

We have investigated various electronic band structures for different graphite modifications using the extended tight binding method. The specific crystal structures studied are the two naturally occurring graphite modifications (Bernal and rhombohedral), and a hypothetical configuration where the carbon atoms in consecutive layers are directly above each other. The latter structure has not been observed in pure graphite, but it is the backbone structure for different stages of intercalated graphite, e.g. Li-graphite compounds LiC_n. On comparing band structures for various graphite modifications we find important differences in the π-bands close to the Fermi energy, the region dominating transport and low energy excitation properties.     

Evolution of the band structure of quasiparticles with doping in copper oxides on the basis of a generalized tight-binding method

Two methods for stabilizing the two-hole ³ B _1g state as the ground state instead of the Zhang-Rice singlet are determined on the basis of an orthogonal cellular basis for a realistic multiband pd model of a CuO₂ layer and the dispersion relations for the valence band top in undoped and doped cases are calculated. In the undoped case, aside from the valence band, qualitatively corresponding to the experimental ARPES data for Sr₂CuO₂Cl₂ and the results obtained on the basis of the t-t′-J model, the calculations give a zero-dispersion virtual level at the valence band top itself. Because of the zero amplitude of transitions forming the virtual level the response corresponding to it is absent in the spectral density function. In consequence, the experimental ARPES data do not reproduce its presence in this antiferromagnetic undoped dielectric. A calculation of the doped case showed that the virtual level transforms into an impurity-type band and acquires dispersion on account of the nonzero occupation number of the two-hole states and therefore should be detected in ARPES experiments as a high-energy peak in the spectral density. The computed dispersion dependence for the valence band top is identical to the dispersion obtained by the Monte Carlo method, and the ARPES data for optimally doped Bi₂Sr₂CaCu₂O_8+δ samples. The data obtained also make it possible to explain the presence of an energy pseudogap at the symmetric X point of the Brillouin band of HTSC compounds.     

Photoemission study of Pd Zr and Cu Zr alloys

The XPS valence bands and core levels of the alloys Pd_1?xZr_x (0<x<1) and Cu_1?xZr_x (0<x<1) have been measured. The alloys prepared by coevaporation are crystalline — but their valence band spectra are close to those of the metallic glasses of the same compositions. The large valence band and core level shifts observed for Pd can be explained by a simple theory, not necessitating the postulation of a new type of bonding in these systems.     

X-ray C K-emission band spectra of graphite fluoride

X-ray C K-emission band spectra of graphite fluoride were measured for the first time by the use of the electron beam excitation. The problem of first importance in obtaining X-ray spectra was the decomposition of graphite fluoride induced by electron beams. The decomposition of the samples during the measurements was controlled by moving the samples at a constant rate. The spectra thus measured showed strong high energy satellite bands. Peak positions of these satellites were 5 eV and 8 eV higher relative to that of the main band. Each satellite was considered to be attributable to the CF and CF₂ bonding, respectively. A similar spectrum was obtained from graphite fluoride, (C₂F)_n. Owing to the zig-zag form of the C-C bond, the polarization of the C K-emission bands of well-ordered flaky graphite fluorides was not observed distinctly.     

Electronic structure of lithium graphite

X-ray photoemission spectra of vacuum cleaved LiC₆, prepared from highly oriented pyrolytic graphite, provide a direct measure of the filling of the graphite π bands by electrons from Li. The resulting Fermi energy shift is in agreement with recent band structure calculations and indicates near unity charge transfer from Li. Core level spectra exhibit shifts compatible with the expected charge transger and line shapes showing strong asymmetries resulting from the metallic character of this compound.     

Low energy photoelectron spectroscopy on alkali graphite intercalation compounds

The conduction band of various stages of alkali graphite intercalation compounds has been studied by low energy photoelectron spectroscopy (hv ? 6.55 eV). The dissimilar behaviour of the width β of the conduction band peak as a function of photon energy for C₆Li and C₈M (M = K, Rb, and Cs) is discussed in terms of different band types in the vicinity of the Fermi level. The stage dependence of β is measured and interpreted for the system C_xK (for stages 1, 2, 4, and 5).     

Studies on optimization for HgCdTe heterojunction photodiodes

Band-edge profiles are calculated for a graded heterojunction diode, i.e. a wide band gap n-HgCdTe layer on a 0.1 eV gap p-HgCdTe substrate, when the energy level for the valence band top (measured from the vacuum level) varies spatially in the same way as the alloy composition does (non-zero valence band offset). The conditions for no barrier-formation in the profiles are derived from the calculations as functions of profiles of both dopant densities and composition. The following characteristics were obtained: (1) Potential barriers do not appear when the relation N_A-N_D⩽10¹⁵cm⁻³ holds. (2) The positive valence band offset value, which corresponds to a practical case, alleviates a limitation to the alloy composition of the n-side crystal. (3) As the valence band offset increases, the region free from barrier formation shifts toward a region with smaller dopant densities, as well as toward a region with steeper distribution of alloy composition. Using the last result on the alloy composition, conditions for crystal growth are briefly discussed.     

Two hole spectra and valence states in chromium and chromium silicides

Auger spectra for L₃M₂₃V and L₃V V transitions involving, respectively, one and two valence holes in the final state, have been measured for Cr and CrSi₂ using both X-ray photons and electrons as ionization source. Careful subtraction of the energy losses from the raw data permits determination of the lineshape of the Auger spectra. The valence hole spectral functions derived from the L₃M₂₃V transitions are compared with valence band spectra obtained by X-ray photoemission. The comparison provides direct evidence of the importance of multiplet coupling between the 3p and 3d holes in the final state. Results for the spectral function of two valence holes are consistent with the outcome of band structure calculations, although some correlation effects seem to be present.     

On the interpretation of valence band photoemission spectra of NiO

The valence band photoemission spectrum and the L₃VV Auger spectrum of NiO are compared. The satellite found in the valence band of NiO and other Ni compounds is interpreted as an unscreened 3d⁷ final state, whereas the main d-emission is a 3d⁷ final state screened by a d electron in an exitonic state.     

ESCA studies of core level shifts and valence band structure in nonstoichiometric single crystals of titanium carbide

Core level shifts and valence band structure from six single crystals of TiC_x have been studied by X-ray photoelectron spectroscopy (ESCA). The     

Photostimulated desorption of metal adatoms: potassium on graphite

Photodesorption is observed of single K atoms from a graphite surface covered with less than 1 monolayer of potassium. The desorption cross section has a threshold at ħω ≈ 3 eV and a maximum at ħω_max ≈ 4.9 eV. Polarization measurements indicate a substrate-mediated mechanism. The coverage dependence suggests that only the ionic 2D, K-phase is photoactive. The proposed mechanism includes attachment of hot electrons, photoexcited in the bulk, to the K 4s adsorbate resonance of energy E_res. The band structure of graphite causes a narrow energy distribution of hot electrons, which yields ħω_max ≈ 2E_res.     

Effects of defects on the electronic structure of ion-irradiated graphite

The effects of N⁺ and Ne⁺ ions impinging on a graphite target are studied by ultraviolet photoelectron spectroscopy. Changes in the valence band of the N⁺-irradiated graphite surface are found to be inherently different from the Ne⁺-ion-induced structural modification. They reveal a build-up of additional -defect states at the top of the band, and confirm what appears to be a distinct character of the influence of nitrogen on an amorphous carbon matrix.     

Galvanomagnetic effects in graphite—II: The influence of trigonal warping of the constant energy surfaces on σxx(Bz)

The dispersion relationship for the electrons near the region of band overlap in graphite corresponds to the case where the constant energy surfaces are strongly warped in the k_x?k_y plane with three-fold symmetry (trigonal warping). The effects of this warping on the galvanomagnetic tensor component σ_xx(B_z) are examined. In the present calculation the trigonal warping is treated using a simplified mathematical model. Implications with respect to the calculation of the density of free carriers in graphite are discussed.