The resonant behavior of anatase TiO2 (101) and (001) surfaces has been investigated using synchrotron photoemission spectroscopy. The data are compared with earlier photoemission work from rutile TiO2(110) and calculations for bulk anatase in order to elucidate the degree of Ti-O hybridization in the valence band. The results for the (101) surface show good general agreement with bulk band-structure calculations. Deviations from the bulk band structure in the case of the (001) surface are attributed to the reconstruction of this surface. A small peak is observed at around 1 eV binding energy (relative to the Fermi energy) for both surfaces following the creation of surface defects (O vacancies), which is thought to arise from surface Ti3+. The attenuation of this peak by gentle heating in O-2 is attributed to healing of the surface O vacancies.

The interaction of NH3 with perfect and defective TiO2(110) surfaces at 300 and 185 K has been studied using synchrotron radiation photoemission. To increase the sensitivity to Ti3+ states, the Ti p-d resonance at an incident photon energy of 47 eV has been used. Ammonia adsorbs mainly molecularly at low coverages at 300 K, with some amounts of NH2 and OH radicals. Upon annealing to high T both desorption and dissociation occurs. A maximum work function variation of −0.57 and −0.38 eV has been found for the perfect and defective surfaces, respectively. No significant differences in the electronic structure between adsorption at the two temperatures have been found, except for a higher coverage on the perfect surface.

Synchrotron radiation photoemission has been used to study the interaction of H2O with defective and nearly-perfect TiO2(110) surfaces at temperatures between 160 and 300 K. Ti3+ 3d defect sites are implicated in the adsorption process, and by tuning the photon energy to 47 eV we find that a resonant photoernission process gives an enhanced photoermission sensitivity to the 3d defect states. Defects are produced on TiO2(110) by annealing to 1000 K in UHV; subsequent exposure to 104 L O2 produces nearly perfect surfaces, based on the suppressed Ti3d emission. Both nearly perfect and defective surfaces give rise to dissociative adsorption of H2O at 300 K. The saturation coverages are near 0.1 ML, independent of the initial defect concentration; however, the rate of dissociative adsorption (sticking probability) is higher on defective surfaces. The enhanced sensitivity to the Ti3+ defect states has allowed the observation of a surprising effect; the dissociative adsorption of H2O results in increased defect state intensity on the nearly perfect surfaces. This apparent charge-transfer to the substrate implies that a new model for the dissociation process on oxide surfaces is needed. At 160 K H2O adsorbs molecularly on both the nearly-perfect and the defective surfaces. Subsequent annealing experiments allow estimates of the interaction energies involved in the dissociation process.

Recent efforts in our laboratory to understand the mechanisms involved in stimulated desorption have forced a re-evaluation of the factors that influence surface stability under ionizing radiation. Our experimental results in desorption from different surfaces of TiO2 [1] have shown that the local surface atomic arrangements can be as important in determining ion desorption probability as the electronic excitations. Likewise, our studies of desorption from MgO [2] have shown that in this highly ionic compound with a large Madelung potential at an O2− site, desorption of O+ does not readily occur when the Mg 1s core level is excited. The Knotek-Feibelman model [3] predicts a large ion signal in this system. It appears that the geometry, lattice dynamics, and excited state lifetime have large effects on the ion desorption probability [2, 4]. To expand our efforts in futher elucidating the role of geometry and lattice dynamics in ion desorption, it is important for us to be able to measure the angular distribution of ions leaving the surface under photon bombardment, and at the same time determine their kinetic energy and mass. In addition, since ion signals often are weak, we also require an analyzer that collects over a large solid angle.

Metal oxides, in particular those that are highly ionic, are ideal systems for the study of mechanisms of stimulated desorption [1]. Many of these ionic materials are also maximal-valency compounds and are useful for examining the Knotek-Feibelman mechanism of stimulated desorption [2]. MgO is one of the most ionic of these materials and, with the Mg2+ cation in a 1s 22s 22p 6 configuration, it is also maximal-valent. Calculations indicate that the (100) surface Madelung energy for a positive charge at a lattice O- site is quite large (>23 eV) [3]. From this perspective, one would expect an interatomic Auger decay process to be operative and, with the strong Madelung repulsion, intense ion desorption signals. Although this would appear to imply that MgO should give intense ion desorption signals, previous studies have shown it to be stable under electron irradiation [4]. Studies of desorption from this material therefore should be instrumental in assessing the validity of charge-transfer and lattice dynamical concepts.

The performance of the “high-flux” toroidal grating monochromator (HFTGM) at the NBS SURF-II synchrotron storage ring is assessed. Two gratings are studied: one with a ruled profile and the other having a laminar profile. The laminar profile is shown to reduce substantially the intensity of higher-order diffracted light with only a small decrease in the intensity of the first order light. The dependence of the energy resolution as a function of the area of the grating illuminated is also discussed.

There have been many studies of the mechanisms of stimulated desorption and of the influence of surface structure, however few investigations have considered their simultaneous interplay. Our recent studies of stimulated desorption from TiO2 surfaces have shown that there is a significant interaction between both of these aspects.

The interaction of O2, CO2, and H2O with bulk BaO and BaO adlayers adsorbed on W(001) has been examined using ultraviolet photoelectron spectroscopy. H2O reacts with bulk BaO to form Ba(OH)2, while CO2 forms a surface layer of BaCO3. Water and carbon dioxide also react with a (√2×2)R45–BaO monolayer adsorbed on W(001) to produce adsorbed OH and CO3 species bound to the tungsten substrate. The interaction of O2 with W(001) is enhanced by the presence of a BaO monolayer on the substrate. The observations are compared with the results of previous studies.

Core-exciton-induced desorption of ${\mathrm{O}}^{+}$ and ${\mathrm{H}}^{+}$ from MgO(100) and MgO(111) has been observed using photon excitation energies spanning the O K edge. Electron-yield data from partially oxidized Mg implies that these states are localized in the near-surface region. ${\mathrm{O}}^{+}$ and ${\mathrm{H}}^{+}$ desorption results from the decay of different O core-exciton states as well as the states produced by interband transitions. The O excitonic levels are interpreted in terms of their related atomic origin.

Electron attenuation lengths in condensed molecular solids have been measured utilizing monochromatized synchrotron radiation from the NBS SURF-II storage ring. The attenuation lengths are approximately 13, 10 and 9 Å, respectively, for water, methanol and cyclohexane and show only a slight energy dependence over the electron kinetic energy range covered (18–68eV). The experiment consisted of monitoring the attenuation of Cu(100) substrate photoelectrons as solid H2O, CH3OH and C6H12 were condensed at 90 K by dosing from a microcapillary array. Accurate measurement of adsorbate layer thickness was accomplished by calibration of the doser; this procedure is described in detail. Verification of this calibration was accomplished by comparison with thermal desorption and UV photoelectron spectroscopic measurements which are able to distinguish monolayers from multilayers.

We describe a method for using synchrotron radiation to measure accurately electron attenuation lengths in condensed molecular solids as a function of electron energy. It consists of measuring the attenuation of photoelectrons from a well characterized, relatively inert, cooled surface as a condensable overlayer is deposited. As photoelectrons from the substrate escape they pass through and are scattered in the overlayer. This scattering appears as a decrease in the intensity of the substrate photoelectron peak. The measurement of this decrease as a function of layer thickness gives the electron attenuation lengths directly. By using monochromatized synchrotron radiation for the photoemission excitation source, one can tune the photon energy and, hence, obtain the attenuation lengths as a function of electron kinetic energy.

The dependence of the electron- and photon-stimulated desorption (ESD, PSD) O+-ion yield on surface preparation from TiO2 (110) and (001) surfaces has been studied. Angle-integrated electron-stimulated desorption yields have been measured versus annealing temperature from room temperature sputtered surfaces to 900°C annealed surfaces. Both the surface cation valence state and the surface geometry change as a function of annealing temperature, giving rise to a rich variety of ESD ion angular distribution (ESDIAD) patterns. These patterns are discussed in terms of possible models of local surface structure.

The resonant enhancement in photoelectron spectra at the 4d edges of rare earth atoms and metals is also found in yield spectra of desorbed ions from the surfaces of the oxides of Sm, Eu and Yb following the photon excitation. The analysis of the 4d → 4f resonance leads to a picture of an indirect mechanism of ion desorption which is mainly caused by the flux of energetic 4f photoelectrons from the bulk. In this case the dominant desorption through secondary processes limits the use of the photon-stimulated desorption (PSD) to determine to which type of atom the desorbing species was attached.

The 3p excitation cross section in the 3d transition metals shows a resonant maximum in photoabsorption and electron energy loss spectroscopy. The resonant 3p excitation is shown to decay into various decay channels with Auger decay and direct recombination being most prominent. Electron and photon stimulated ion desorption from the 3d transition metal oxide surfaces is initiated by Auger induced 2 hole and 2 hole. 1 electron final states. In Cr(110)O2 the O+ yield differs significantly from the total secondary electron yield. This rules out electron stimulated ion desorption induced by secondary electrons. It is the first instance of a pronounced core hole state sensitivity observed in PSD.

The pulse counting and delay circuitry for a threshold photoelectron-photoion coincidence mass spectrometer is presented along with the automation system using a CEC LSI 11/23 computer and CAMAC instrumentation. A switching input register and logic circuits are used to measure alternately a time of flight coincidence mass spectrum and a background accidental coincidence spectrum, at selected photon energies. The high voltage, fast rise time pulse used for time of flight analysis creates special signal isolation problems which are handled with unique pulse blanking techniques.

An interpretation and comparison of photon-stimulated desorption yields of ${\mathrm{H}}^{+}$ ions from OH on Ti and Cr and from bulk solid ${\mathrm{H}}_{2}$O indicate that desorption occurs through two entirely different mechanisms. The first involves in intramolecular excitation of the OH adsorbate producing a ${H}^{+}$ yield similar to that in bulk ${\mathrm{H}}_{2}$O. The second involves metal core-level excitation followed by Auger decay and is an example of molecular adsorbate dissociation arising from a metal-substrate Auger decay. This is a further generalization of the Knotek-Feibelman model applicable for desorption in ionic systems.