C.R. Brundle

Interpretation of the x-ray photoemission spectra of cobalt oxides and cobalt oxide surfaces

Abstract CoO and Co3O4 have been studied by high-resolution X-ray photoemission. The characteristic binding energies in the Co 2p 3 2 , 2p 1 2 , and 3s regions, their band shapes and widths, the associated shake-up structure, the O(1s) and O(2s) BEs, and the valence band spectra have been examined. The two oxides are readily distinguished from their spectra though it is shown that the O(1s) BEs are identical at 529.50 ± 0.14 eV. Argon and oxygen ion sputtered surfaces were examined to establish the integrity of the oxides. A higher BE O(1s) component (530.7–531.6 eV), the intensity and BE of which vary with the treatments mentioned above, corresponds to non-stoichiometric surface oxygen. The results are discussed with respect to the electronic structures of the oxides and the often conflicting earlier studies of these oxides.

Core and valence level photoemission studies of iron oxide surfaces and the oxidation of iron

The core and valence level XPS spectra of FexO (x ~ 0.90–0.95); Fe2O3 (α and γ); Fe3O4; and FeOOH have been studied under a variety of sample surface conditions. The oxides may be characterized by a combination of valence level differences and core-level effects (chemical shifts, multiplet splittings, and shake-up structure). FeII and FeIII states are distinguishable, but octahedral and tetrahedral sites are not. The O 1 s BE cannot be used to distinghuish between the oxides since it has a nearly constant value. Fe 3d valence level structure spreads some 10 eV below EF, much broader than suggested by previous UPS and photoelectron-spin-polarization (ESP) measurements for FexO and Fe3O4. Fe surfaces (films, foils, (100) face) yield predominantly FeIII species when exposed to high exposures of oxygen or air, though there is evidence for some FeII also. At low exposures the FeII/FeIII ratio increases.

CO oxidation on Pd(100): a study of the coadsorption of oxygen and carbon monoxide

Oxygen adsorption, and coadsorption and reaction with CO were studied with temperature programmed reaction spectroscopy (TPRS), low energy electron diffraction (LEED), and high resolution electron energy loss spectroscopy (EELS). Oxygen adsorption at 300 K was studied for coverages up to 0.5, while coadsorption with CO was studied in the temperature range of 80 to 450 K, and O and CO coverages of 0 to 0.25 and 0 to 0.8, respectively. Oxygen adsorbed into p(2 × 2) islands for coverages in excess of 0.05 and yielded a fully developed p(2 × 2) structure for 14 monolayer coverage at 300 K. The p(2 × 2)O pattern was gradually replaced by a c(2 × 2)O structure as the oxygen coverage was increased to 0.5. Oxygen desorption occurred in three temperature programmed desorption states at 840, 730, and 695 K. The highest temperature state was populated at all coverages, whereas the 730 and 695 K states appeared as the oxygen coverage exceeded 0.25 and 0.36, respectively. The 695 K state was unusually narrow suggesting attractive interactions in the oxygen adlayer for coverages of 0.36 to 0.5. For the surface precovered with 0.25 monolayer of oxygen, CO adsorption at 80 K caused a disordering of the p(2 × 2)O structure. At lower oxygen coverages CO initially adsorbed at sites apart from the oxygen domains, but also adsorbed within oxygen islands upon filling of the exterior sites. CO adsorption in the interior of the islands produced a directly interacting COPdO complex characterized by a CO stretching frequency of 2125 cm−1. Reaction-limited CO2 evolution occurred in three states; viz. reaction of CO with (1) isolated oxygen atoms or p(2 × 2)O islands at 420 K, (2) disordered oxygen islands at 360 K, and (3) by a low temperature reaction between 100 and 310 K due to the COPdO complexes. No evidence of molecularly adsorbed CO2 or carbonate (CO3) reaction intermediates was observed by EELS at 80 K.

X- ray photoemission study of the interaction of oxygen and air with clean cobalt surfaces

Abstract The interaction of oxygen with polycrystalline cobalt surfaces has been studied at 300 K (1 × 10−6 to 1 × 10−5 Torr) using high-resolution (monochromatized) X-ray photoemission. At high exposures (> 100 L nominal) CoO is identified as the product from the nature of the Co 2 p 3 2 , 2 p 1 2 , 3 s , and valence band spectra. There is no evidence for measurable amounts of Co3O4 or Co2O3. Two O 1s features are observed at both high and low (10L) exposures. The dominant O 1s feature at 529.5 ± 0.2 eV corresponds to the oxide and a minor feature at 531.3 ± 0.2 eV is attributed to non-stoichiometric surface oxygen. Exposure to air produces quite different results, with a dominant O 1s feature at 531.5 ± 0.2 eV and dominant Co 2 p 3 2 and 2 p 1 2 features centered at 781.3 ± 0.2 eV and 797.1 ± 0.2 eV. These three values are very close to those reported here for bulk Co(OH)2. Ion etching of the air-exposed surface removes this dominant surface product rapidly revealing some oxide and finally metal.

Oxygen adsorption and thin oxide formation at iron surfaces: An XPS/UPS study

Abstract The interaction of oxygen with iron films at 80 and 300 K has been followed by XPS, and, in part by HeI/HeII UPS. At 80 K adsorption ceases near monolayer coverage (θ = 1), when the sticking probability, Sexp, which has remained constant, drops sharply to zero. At 300 K, Sexp decreases with increase in θ but the fast adsorption stage has no abrupt end. There is a period between θ ~ 0.5 and 1 (θ = 1 corresponds to one O atom per surface Fe atom) where Sexp remains approximately constant. Exposure to 0.5 atm oxygen at 300 K produces an oxide layer estimated to be about 20 A thick. Monitoring the O(1s) and Fe(2p 3 2 ) BEs and the UPS spectrum as a function of θ leads to the following conclusions. (1) For θ ⩾ 0.2 an O(1s) BE of 530.2 ± 0.2 eV is dominant; (2) below θ = 0.2 a higher BE O(1s) is observable in addition to the 530.2 eV peak; (3) despite the very different kinetics of 80 and 300 K, no bonding differences are observable in the XPS up to θ = 1 ; (4) below θ=1 neither the present UPS or XPS data can completely characterize the electronic nature of the reaction products; (5) at high coverage (θ ⩾ 1.5) the dominant species is FeIII; (6) the adsorption kinetics are explainable only in terms of the presence of a mobile precursor state, and it is suggested that the difference between 80 and 300 K may be accounted for by the different degrees of order established in the chemisorbed layer at each temperature.

An X-ray photoelectron spectroscopy study of the chemical changes in oxide and hydroxide surfaces induced by Ar+ ion bombardment☆

In the course of X-ray photoemission studies on the oxidation of metals and alloys and on bulk oxides, we found that in addition to physical sputtering Ar+ ion bombardment can in many cases reduce an oxide to a mixture of the original oxide, lower oxides and metal. The effect is even more pronounced in systems containing hydroxyl groups which are readily destroyed by the ion beam. Specific examples for oxidized cobalt, nickel and iron surfaces and their bulk oxides and hydroxides are given. The relative reduction rates of CoII and FeIII in CoFe2O4 are also examined. From these observations, it is clear that any depth compositional profiling using ion sputtering in conjunction with Auger or X-ray photoelectron spectroscopy should be treated with extreme caution. The mechanism for the chemical changes induced by ion bombardment is briefly discussed.

Adsorption and decomposition of ammonia on a W(110) surface: Photoemission fingerprinting and interpretation of the core level binding energies using the equivalent core approximation

We report the first XPS data for ammonia adsorption, condensation and decomposition on a W(110) surface. Monolayer, “second layer” and multilayer NH3 as well as NH2, NH and N species can be characterized by a specific N(1s) electron binding energy. We discuss the observed binding energies within a thermodynamic framework, using the “equivalent core approximation”. This model has been previously successfully applied to core level binding energies of gaseous molecules and solids. The agreement between calculated and experimental N(1s) binding energies for some species is excellent, and we conclude that for the considered adsorbates the variation of the N(1s) binding energies is primarily determined by the ground state properties rather than by different relaxation energies in the final state. We also briefly discuss the activity of the W(110) face towards NH3 decomposition and also present some data for NH3 dissociation on an oxygen predosed surface.

The interaction of oxygen with gadolinium: UPS and XPS studies

The interaction of oxygen with evaporated Gd films at 300 K has been studied for the first time using AlK α XPS and Hel and Hell UPS. The characteristic changes in the Gd(6s5d) and O(2p) valence bands, Gd(4f), Gd(5p) and Gd(4d) core levels, and O(2s) and O(1s) core levels were studied. Evidence is presented for the initial formation of an intermediate oxidation state at low exposure (characterized by a new Gd valence band with a maximum in the DOS at ∼ 2.5 ev below EF and an ∼ 0.6 eV shift in Gd(4f)) prior to formation of Gd2O3 where the Gd(6s5d) valence band disappears completely, as expected for Gd3+. In the higher exposure range there is little further increase in the oxide thickness, which is estimated as ⩾ 20 A, but there is a slow replacement of O by OH, as characterized by a second O(1s) feature at 532.3 eV and OH 1π and 3σ orbitals in UPS at ∼ 6.7 and 11.5 eV. The interpretations are supported by parallel studies on bulk Gd2O3 and by Ar+ sputtering experiments. Comparisons are made to other rare earth oxidation studies.

A search for precursor states to molecular nitrogen chemisorption on Ni(100), Re(0001) and W(100) surfaces at ~20 K

Abstract We have studied adsorption and condensation of molecular nitrogen on Ni(100), Re(0001) and W(100) surfaces by XPS at surface temperatures ⩾ 20 K in order to search for a precursor state to the linearly bonded chemisorbed state (γ-state). Our data however show XPS spectra only from γ-N 2 at low exposures and from physisorbed and condensed N 2 at higher exposures. No intrinsic precursor is found, allowing an upper bound of approximately 3 kJ/mol to be placed on the activation barrier between any such state and the chemisorbed state. The physisorbed N 2 observed on top of γ-N 2 covered patches acts as an extrinsic precursor to chemisorption.

Cluster model calculation and photoemission studies of molecularly adsorbed NO on Ni

Self-consistent scattered wave cluster model calculations for the NONi system are presented with NO bonded N-down in a fourfold coordination site. The results suggest that the charge is transferred from metal into the single occupied 2π∗ molecular orbital of NO which is thus significantly perturbed. The interaction involves the metal s, p, and d-derived states. We calculate total and local “density of states” defined within the framework of the cluster model and compare the results with Ultraviolet Photoemission experiments recorded at 80 K. At 300 K partial dissociation occurs as judged from UPS and XPS. Increased photoemission observed at about 2 eV below the Fermi level in the molecularly adsorbed state is attributed to the alteration of density of states due to the interaction of the 2π∗ molecular orbital of NO with the substrate. A detailed analysis of the nature of the relevant molecular orbitals is also given.