Photoemission Final State Research Articles

Electronic structure of MnO.

We have studied the electronic structure of MnO by photoemission spectroscopy. Mn 3d--derived emission is found to be confined within \ensuremath{\sim}10 eV of the top of the valence band, and we find no evidence for an intrinsic satellite at higher binding energies. The photoemission intensity of the whole valence band is enhanced for photon energies in the Mn 3p\ensuremath{\rightarrow}3d core absorption region. These observations suggest that ${\mathit{d}}^{4}$ and ${\mathit{d}}^{5}$L photoemission final states (L denotes a ligand hole) are close in energy and are strongly hybridized with each other. The spectra are analyzed in terms of configuration-interaction theory using a ${\mathrm{MnO}}_{6}$ cluster model. We thus find that the ligand-to-Mn d charge-transfer energy \ensuremath{\Delta} is comparable to the intra-atomic Coulomb energy U\ensuremath{\simeq}7.5 eV. Namely, MnO is indeed close to the boundary between the Mott-Hubbard (U) and the charge-transfer (U>\ensuremath{\Delta}) regimes in the Zaanen-Sawatzky-Allen phase diagram, and the ${\mathit{d}}^{4}$ and ${\mathit{d}}^{5}$L states are nearly degenerate. The large \ensuremath{\Delta} value leads to a highly ionic character in the Mn-O bonding.

X-ray photoelectron diffraction from a free-electron-metal valence band: Evidence for hole-state localization.

X-ray-photoelectron diffraction has been measured from an Al(001) single crystal for the free-electron-like valence band and the well-localized 2s core level, using Mg K\ensuremath{\alpha} radiation. Both show the same high anisotropy of 65% in the azimuthal distribution of the photoemitted electrons at 45\ifmmode^\circ\else\textdegree\fi{} away from the surface normal. Moreover, the measured intensity patterns are essentially identical, indicating a strong similarity in the photoemission final states. This similarity provides evidence for a strong degree of hole localization in high-energy valence-band photoemission.

Photoemission study of the valence-band electronic structure in FexO, Fe3O4, and alpha -Fe2O3 single crystals.

Valence-band photoemission spectra have been measured from cleaved stoichiometric iron oxide single crystals using synchrotron radiation. The total photoelectron yield is observed to increase dramatically at photon energies above the Fe 3p\ensuremath{\rightarrow}3d excitation threshold, and the resonant onset correlates directly with the oxidation state of the Fe cations. Features in the total yield profiles are interpreted in terms of atomic multiplets. The phenomenon of resonant photoemission is used to distinguish the Fe 3d-derived valence states from the overlapping unhybridized O 2p states in each of the oxides. By use of cleaved single crystals, the subtle differences in the valence-band photoemission features associated with ferrous (${\mathrm{Fe}}^{2+}$) and ferric (${\mathrm{Fe}}^{3+}$) cations have been determined. The measured energy distribution of the photoemission final states in single crystal \ensuremath{\alpha}-${\mathrm{Fe}}_{2}$${\mathrm{O}}_{3}$ are in good agreement with recent measurements and configuration-interaction cluster calculations by Fujimori et al., which take into account ligand-to-metal charge transfer. Constant-initial-state spectra measured across the 3p\ensuremath{\rightarrow}3d threshold in each oxide indicate that most of the photoemission intensity near the top of the valence band is derived from hybridized cation-3d--ligand-2p states, whereas the emission at higher binding energies results from primarily 3${d}^{n\mathrm{\ensuremath{-}}1}$ final states. On this basis, each of the iron oxides is classified as a charge-transfer rather than a Mott-Hubbard insulator.

Determination of valence states in Fe3O4 by resonant photoemission

We report resonant photoemission measurements across the Fe 3p→3d photoabsorption threshold from the (110) surface of cleaved single-crystal Fe3O4 using synchrotron radiation. The resonant enhancement effect is used to distinguish the Fe 3d-derived states from the overlapping O 2p states in the valence band. The Fe 3d-derived states are found to extend about 18 eV below the Fermi level. Constant-initial-state spectroscopy across the Fe 3p excitation threshold reveals that this large valence bandwidth is due to a significant amount of cation–ligand hybridization; this result is in accord with recent configuration interaction cluster calculations of the photoemission final states in iron oxides reported by Fujimori et al. The electronic levels that are primarily associated with the different Fe2+ and Fe3+ cation environments in Fe3O4 are identified based on a comparison with photoemission spectra measured from cleaved FexO (x≂0.945) and α-Fe2O3 single crystals. Upon O2 adsorption, the Fe3O4(110) surface is rapidly oxidized, as evidenced by a decrease in the photoemission features associated with the Fe2+ valence state.

Valence bands, oxygen in planes and chains, and surface changes for single crystals of M2CuO4 and MBa2Cu3Ox (M=Pr,Nd,Eu,Gd).

X-ray photoemission results for single crystals of ${M}_{2}$${\mathrm{CuO}}_{4}$ (M=Pr,Eu,Gd), M${\mathrm{Ba}}_{2}$${\mathrm{Cu}}_{3}$${\mathrm{O}}_{\mathrm{x}}$ (M=Nd,Gd), and CuO are sintered ${\mathrm{La}}_{1.85}$${\mathrm{Sr}}_{0.15}$${\mathrm{CuO}}_{4}$ and ${\mathrm{YBa}}_{2}$${\mathrm{Cu}}_{3}$${\mathrm{O}}_{6.9}$ show valence-band spectra within 10 eV of the Fermi energy that are remarkably similar in appearance, with contributions that reflect Cu-O hybrid states and the rare-earth 4f states. For ${\mathrm{Pr}}_{2}$${\mathrm{CuO}}_{4}$ and ${\mathrm{NdBa}}_{2}$${\mathrm{Cu}}_{3}$${\mathrm{O}}_{\mathrm{x}}$, there are two distinct 4f features due to ligand screening in the photoemission final state. The rare-earth 5p core-level emission overlaps the O 2s emission and reveals complex 5p-4f multiplet interactions. All O 1s spectra show a dominant peak at \ensuremath{\sim}528 eV that can be resolved into features separated by \ensuremath{\sim}0.7 eV. These reflect inequivalent oxygen bonding configurations in the lattice and are associated with the planes and chains for the 1:2:3 compounds and the planes and off-planes for the 2:1:4 compounds. The lower-binding-energy feature is associated with the Cu-O chains of the 1:2:3 compounds and the Cu-O planes of the 2:1:4 compounds. In addition to the O 1s main line for the Cu-O planes there is also a weak satellite. Time dependent studies of the Cu 2p and O 1s emission indicate surface modification, dependent upon the quality of the cleave. The effects of surface changes and the presence of imperfections are discussed in the context of surface studies and surface superconductivity.

Work-function changes and photoemission final-state relaxation of Ne, Ar, Kr, Xe, H 2 and N 2 on gallium

UP spectra were recorded for several gases physisorbed on a polycrystalline Ga film well below 10 K. Layer-dependent shifts of electronic binding energies are observed for the first time for Ne and H 2 . In all cases studied here the layer-dependent shifts of binding energy are larger than the work-function changes. Therefore, it is concluded that these shifts are caused by a change of relaxation of the photoemission final state.

Calculated local density X-ray and photoemission spectra for superconducting La 2−xM xCuO 4: Localization of Cu-3d

Comparison of UPS (XPS) and XES spectra for the high T c superconductor La 2− x M x CuO 4 - calculated using highly precise local density results - with available experimental data reveals a large shift of about 2 eV to lower energies for the Cu-d related features. This shift is interpreted qualitatively in terms of a simple two-level model taking into account the hole-hole repulsion at the Cu sites in the photoemission final state. This simple model explains nicely the magnitude of the shift, the occurence of a Cu related satellite below the valence-bands, and the charge-transfer screening of the Cu-d photo-holes by O-p electrons. The O-states, especially the bonding states at the bottom of the valence band, are less affected and agree reasonably well with the experimental findings.

Coexistence of biholes and electron-bihole complexes in the photoemission final state in cuprous halides.

Coexistence of biholes and electron-bihole complexes in the photoemission final state in cuprous halides.

Noble- and transition-metal clusters: The d bands of silver and palladium.

A comparison of x-ray photoemission from Ag and Pd clusters grown on amorphous carbon substrates highlights the importance of the unfilled 4d band in Pd clusters. In both cases, the valence-band spectra show the d-band narrowing with decreasing cluster size, as expected. In both cases, also, there is a positive shift of the binding energies of the d-band centroids and of the core levels, primarily due to the unit positive charge that remains on the cluster in the photoemission final state, as occurs for other metal clusters on amorphous carbon. In Ag clusters the core-level shift is smaller than the valence-band shift because in small clusters the Coulomb energy of the charged cluster suppresses the conduction electron screening of the core hole. By contrast, in Pd clusters the increased localization causes a reduction in the d-electron density of states at ${E}_{F}$, resulting in a transition to s-electron screening and hence a core-level shift that is larger than the valence-band shift.

4fphotoemission from Ce clusters and disordered reaction products at Ce/Si and Ce/GaAs interfaces

Synchrotron radiation photoemission has been used to study the 4f emission from 4f electrons in Ce clusters and disordered reaction products at Ce/GaAs(110) and Ce/Si(111) interfaces. For low Ce coverages on GaAs(110), the reacted interface forms Ce-As local configurations with two 4f-derived features similar to those of bulk CeAs. The shallower 4f peak is located \ensuremath{\sim}1 eV below ${E}_{F}$, indicating that it is not due to fully relaxed 4f holes, consistent with the As p\ensuremath{\rightarrow}Ce 4f charge-transfer screening mechanism. For the Ce/Si interface, small Ce clusters formed at low coverage have only a single 4f peak at -3.3 eV, with no evidence of a second peak nearer ${E}_{F}$. This is attributed to the reduced number of neighboring Ce 5d orbitals for the clusters, apparent stability of the \ensuremath{\gamma} phase relative to the \ensuremath{\alpha} phase (in the ground state), and reducted 5d\ensuremath{\rightarrow}4f charge-transfer screening probability in the photoemission final state. At high coverages, Ce overlayers cover the reacted interfaces, but our results show that the 4f emission near ${E}_{F}$ is not established until \ensuremath{\theta}\ensuremath{\simeq}20 monolayers, suggesting that the 4f-5d hybridization in metallic Ce is long ranged compared with the short-range Ce-4f/As-p hybridization.

Unit Charge on Supported Gold Clusters in Photoemission Final State

Unit Charge on Supported Gold Clusters in Photoemission Final State

Resonant photoemission at the 3p-threshold of Ge and GeO 2

Resonant photoemission final states are reported for the elemental semiconductor Ge and its oxide GeO 2 at the Ge3p photothresholds. The line positions and the multiplet structure of the resonating peaks identifies them as an atomic-like 3d 8 configuration. Differences in the intensity profiles and the exact location of the resonance reveal the importance of a change in initial state electron configuration for this basically atomic effect. The effective Coulomb interaction between the two 3d-holes is U eff ( 1G) = 12.8, 13.2 ± .2 eV for Ge and GeO 2 respectively. A review of all elemental results then allows the restriction of this resonance to elements between Fe and Rb. The line positions of the resonant final states of the remaining elements As, Se, Br, Kr can then approximately be predicted.

X-ray photoemission study of Ce-pnictides

All the Ce pnictides crys allize in the NaCl structure with lattice constant increasing systematically from CeN to CeBi. Since Ce has the largest $4f$-orbital radius in the entire lanthanide series, it is quite interesting to follow the evolution of his level as the distance between the Ce atoms is varied. The core levels and the valence-band region of CeN, CeP, CeAs, and CeSb have been studied by high-resolution x-ray photoemission spectroscopy. The $4d$ and $3d$ core-level spectra of Ce show clearly that the $4f$ state remains strictly localized in all compounds. In CeN a superposition of lines corresponding to two valences is found which reveals the intermediate valence character of this compound. In the valence-band spectrum, the $4f$ level is found to be pinned at the Fermi level and superimposed on the $5d$ states. The extended valence states originating from the anion $p$ states are located at higher binding energies. In the compounds of heavier rare-earth elements the $4f$ state moves away from the Fermi level and gradually overlaps in energy with $p$ states, so that the probability for interatomic Auger processes in the photoemission final state increases rapidly. The natural width of the $4f$ peak is already 0.8 eV in CeP, and becomes so large in CeAs and CeSb that this line can no longer be unambiguously identified in the other valence-band structure. Finally, it can be concluded from the similarity of the $4d$ spectrum of metallic Ce and CeN that the $4f$ electron remains localized in the $\ensuremath{\alpha}$ phase of this metal.