P.R. Norton

A photoemission study of the interaction of Ni(100), (110) and (111) surfaces with oxygen

Abstract Photoelectron spectroscopic studies of the oxidation of Ni(111), Ni(100) and Ni(110) surfaces show that the oxidation process proceeds at 295 and 485 K in two distinct steps: a fast dissociative chemisorption of oxygen followed by oxide nucleation and lateral oxide growth to a limiting coverage of 3 NiO layers. The oxygen concentration in the 295 K saturated oxygen layer on Ni(111) was confirmed by 16 O(d,p) 17 O nuclear microanalysis. At 295 and 485 K the oxide growth rates are in the order Ni(110) > Ni(111) > Ni(100). At 77 K the oxygen uptake proceeds at the same rate on all three surfaces and shows a continually decreasing sticking coefficient to saturation at ~2.1 layers (based upon NiO). An O 1 s b . e . = 529.7 eV is associated with NiO, and O ls b.e.s of ~531.5 and 531.3 eV can be associated, respectively, with defect oxide (Ni 2 O 3 ) or (in the presence of H 2 O) with an NiO(H) species. The binding energies (Ni 2p, O 1s) of this NiO(H) species are similar to those for Ni(OH) 2 . Defect oxides are produced by oxidation at 485 K, or by oxidation of damaged films (e.g. from Ar + sputtering) and evaporated films. Wet oxidation (or exposure to air) of clean nickel surfaces and oxides, and exposure of thick oxide to hydrogen at high temperature results in an O 1s b.e. ~531.3 eV species. Nuclear microanalysis 2 H( 3 He,p) 4 He indicates the presence of protonated species in the latter samples. Oxidation at 77 K yields O 1s b.e.s of 529.7 and ~531 eV; the nature of the high b.e. species is not known. Both clean and oxidised nickel surfaces show a low reactivity towards H 2 O; clean nickel surfaces are ~10 3 times less reactive to H 2 O than to oxygen.

Adsorption of co on Pt(111) studied by photoemission, thermal desorption spectroscopy and high resolution dynamic measurements of work function

Abstract The adsorption of CO on Pt(111) has been studied by XPS, UPS, thermal desorption mass spectroscopy and by dynamic and static work function measurements at 95, 298, 377,403 and 453 K. At all temperatures and coverages ≳0.1 monolayers two states are populated, the ratio of which is controlled by kinetic considerations at T ≲ 320 K and thermodynamics at T ≳ 400 K. The state with the lowest heat of adsorption is bridge bonded CO and has a dipole moment ≲ 0.004 D with the negative end of the dipole pointing outwards. The more strongly bound state is linearly bonded CO which has a dipole moment of ∼0.04 D, positive end outwards. The difference in the heat of adsorption is ∼3800 J mol −1 , half the value derived from an analysis of the thermal desorption spectra. It is suggested that the frequency factor for desorption of the bridge species is ∼5 times that of the linearly bonded CO, perhaps because of the greater entropy change in desorption of the former. The change in work function at θ = 0.33 between 80 and 300 K cannot be explained solely by a change in the ratio of bridge to linear bonded CO but must reflect changes in the order of the adsorbed layer which then effects the dipole moment by through-metal interactions. The ratio of bridge to linear states determined from thermal desorption measurements is different from that measured by XPS at 298 or 100 K, because of interconversion at T ≳ 400 K. XPS and UPS measurements indicate that the sequence of CO levels in the adsorbed state is 1π, 5σ, 4σ in order of increasing binding energy and that the 1σ-4σ splitting is smaller (by ∼0.5 eV) in bridge bonded than in linearly bonded CO. A very weakly chemisorbed state is detected at 77 K which desorbs below 100 K. The intensity ratio of the observed UPS bands are compared to those strongly in chemisorbed CO, and it is suggested that either (a) the bands result from a different weakly adsorbed state than can tumble on the surface or (b) that the extra peaks are due to shake-up processes which are enhanced because of the weakness of the interaction between CO and Pt(111) at high coverages.

The pt(100) (5 × 20) ⇋ (1 × 1) phase transition: A study by Rutherford backscattering, nuclear microanalysis, LEED and thermal desorption spectroscopy

Abstract The reconstruction exhibited by clean Pt(100) surfaces [(5 × 20) LEED pattern] is removed by the adsorption of CO. Rutherford backscattering (RBS) indicates that 1.65 ± 0.05 × 10 15 Pt atoms cm −2 move into registry with the bulk upon adsorption of 6.4 ± 0.4 × 10 14 CO molecules cm −2 ( θ = 0.50 ± 0.03 monolayers). The data indicate that some atoms in the second and perhaps even subsequent layers must be displaced by ≳0.01 nm in the reconstructed surface. By contrast, only 1.3 ± 0.1 × 10 15 Pt atoms cm −2 move back into registry upon adsorption of H 2 or D 2 , and the LEED pattern also indicates that residual reconstruction remains. The stability of the CO-covered, H-covered and “almost clean” (1 × 1) surfaces (the latter prepared by NO and H 2 treatments with a residual H-coverage of ∼1 × 10 14 H atoms cm −2 ) was investigated by RBS. The CO-covered surface starts to reconstruct only when the CO coverage drops below 0.5 monolayers ( T ≳ 450 K) while the H-covered surface (produced by adsorption on the (5 × 20) surface) reconstructs rapidly at T ≳ 350 K, by which temperature the adsorbed hydrogen coverage drops below ∼0.2 monolayers. The “almost clean” surface reconstructs at T ≳ 390 K and the data indicate that the process exhibits an activation energy of 88 ± 17 kJ mol −1 . The absolute coverages of CO and D were determined by nuclear microanalysis (NMA) and excellent agreement was achieved between the LEED and NMA data. The saturation CO coverage was found to be 0.77 ± 0.03 monolayers, consistent with the observed c(4 × 2) LEED pattern. Deuterium (and hence hydrogen) coverages of 1.54 ± 0.1 × 10 15 D (H) atoms cm −2 ( θ = 1.20 ± 0.08) were found at saturation at ∼150 K and the hydrogen adsorbed on the (1 × 1) surface was more strongly bound than that resulting from adsorption on the (5 × 20) surface.

Absolute coverage and isostemc heat of adsorption of deuterium on Pt(111) studied by nuclear microanalysis

Abstract The absolute coverage (θ) of deuterium adsorbed on Pt(111) in the ranges 180 T −6 P −2 Pa D 2 has been determined by nuclear microanalysis using the D( 3 He, p) 4 He reaction. From these data, the isosteric heat of adsorption ( E a ) has been determined to be 67 ± 7 kJ mol −1 at θ ≲ 0.3. This heat of adsorption yields values of the pre-exponential for desorption (10 −5 to 10 −2 cm 2 atom −1 s −1 ) that lie much closer to the normal range for a second order process than those determined from previous isosteric heat measurements. The E a versus θ relationship indicates that the adsorbed D atoms are mobile and that there is a repulsive interaction of 6–8 kJ mol −1 at nearest neighbour distances. At 300 K the coverage decreases to ≲ 0.05 monolayer (≲ 8 × 10 13 D atoms cm −2 ) as P → 0, apparently invalidating a recent model of site exchange in the adsorbed layer.

Interaction of O2 with Pt(100) II. Kinetics and energetics

Abstract The kinetics and energetics of the interaction of O 2 with Pt(100)-hex and (1 × 1) surfaces were studied by thermal desorption spectroscopy, work function techniques and X-ray photoemission. Three states of adsorbed oxygen are formed in roughly equal amounts at saturation (saturation coverage = 0.81 ± 0.04 × 10 15 O atoms cm −2 ). The state desorbing at the lowest temperature (β 1 ) which exhibits a very narrow desorption peak is associated with a phase transition (complex to (3 × 1)) involving Pt atom displacements. The next state to desorb (β 2 ), is best modeled by second order kinetics, a constant activation energy for desorption E d 0 (≈ 10 −3 cm 2 atom −1 s −1 , E d ≈ 160 kJ mol −1 ) and is associated with a further phase transition, (3 × 1) → hex. The high temperature state (β 3 ) which populates first with an initial sticking coefficient of ≈ 4 × 10 −3 at 573 K is believed to be associated with surface defects in the hex-overlayer. The rate of adsorption of the β 2 state at T ≳ 573 K increases with increasing coverage at 0.1 ≲θ ≲ 0.3, probably because of nucleation and growth of (3 × 1) islands which act as traps for further adsorption. The maximum sticking coefficient into either the β 1 or β 2 states is ≈ 10 −3 . The initial sticking coefficient of O 2 on Pt(100)−(1 × 1) is ≈ 0.1.

Interaction of O2 with Pt(100). I: Equilibrium measurements

Abstract Two newly discovered phases on the Pt(100) surface produced by the adsorption of oxygen have been investigated using Rutherford baekscattering (RBS), nuclear microanalysis (NMA), work function changes (Δφ) and LEED. One phase is associated with the oxygensaturated surface (0.63 ± 0.03 monolayers0.81 × 10 15 O atoms cm −2 ), where a very complex LEED pattern is observed; the other is observed at an average coverage of 0.44 ± 0.05 monolayers and gives rise to a (3 × 1) LEED pattern (when observed at room temperature). For both surfaces, RBS measurements indicate large (⩾ 0.025 nm) Pt atom displacements. Also discussed is a new method for preparing the “clean” (1 × 1)-Pt(100) surface without the need for NO adsorption/decomposition.

An investigation of the adsorption of oxygen and oxygen containing species on platinum by photoelectron spectroscopy

It is shown that XPS can detect 0.01 monolayers of adsorbed carbon or oxygen and can identify the chemical state of the adsorbed atom(s). Two states of adsorbed oxygen were resolved by thermal desorption spectroscopy and by XPS. The O 1s binding energies (FEB) were 530.2 and 533 eV below the platinum Fermi level for the strongly and weakly adsorbed states respectively. (FEB) did not vary with coverage. The resulting apparent variation of (VEB), the vacuum level referenced value, is discussed in terms of a simple model for the work function Φ which was measured in situ. UPS indicated that the weakly adsorbed state is probably molecular, with levels at 6.1, 9.3, 10.4 and l2.4 eV below the Fermi level. The main change in the UPS spectra produced by the strongly adsorbed state was a reduction of a peak close to the Fermi level.

Absolute coverages and hysteresis phenomena associated with the CO‐induced Pt(100) hex⇄(1×1) phase transition

Temperature hysteresis in the hex⇄(1×1) phase transition in the Pt(100)–CO system has been investigated using Rutherford backscattering spectroscopy, nuclear microanalysis, work function measurements, LEED, and thermal desorption spectroscopy. The onset of the hex→(1×1) transition occurs at an average surface coverage of 0.08±0.05 monolayers while the onset of the (1×1)→hex transition occurs at 0.25±0.05 monolayers. The results are consistent with a recently proposed model for the phase transition.

A photoelectron spectroscopic study of the adsorption of CO and CO2 on platinum

Abstract The adsorption of CO and CO2 on platinum has been studied by UV photoelectron spectroscopy using both He I and He II radiation. The modulation of the intensity of the spectral features observed for adsorbed CO as the photon energy is changed is used to assign the observed levels. The results are in reasonable agreement with recent theoretical and experimental work. The levels are observed to shift by different amounts compared to gas phase CO because of chemical binding effects. The adsorption of CO2 produces spectral features that are shifted by the same amount compared to gas phase CO2. This, together with the absence of any localized attenuation of the platinum valence band and the low heat of adsorption, indicates that CO2 is physisorbed on platinum.

The heat of adsorption of hydrogen on platinum

Abstract The heat of adsorption ( q ) of hydrogen on clean polycrystalline platinum filaments has been determined as a function of the coverage (θ), temperature and method of preadsorption using UHV adiabatic calorimetric techniques. Adsorption at 273 °K results in a stepped shape q -θ curve with q ( θ =0)= 24.7 kcal mole −1 . At 77 °K, the form of the curve depends on the temperature of preadsorption. The relative roles of a priori and induced heterogeneity are discussed and while both must play a part, the temperature dependence of the form of the q -θ curves suggest that the latter is the more important of the two in these experiments.