Laser induced thermal desorption of carbon monoxide from Fe(110) surfaces

Thermal desorption of CO is induced by bombarding an Fe(110) surface with pulses of a neodymium glass laser. The maximum amplitude of the desorption signal is recorded by a mass spectrometer as a function of the laser pulse intensity and of the CO coverage for both single pulses and sequences of pulses. Since the half width of the laser pulses is only 30 ns the shape of the desorption signal is mainly determined by the time-of-flight of the desorbed particles. There is strong evidence that the latter obey a Maxwell-Boltzmann distribution of temperature T_d, identical in the low temperature range with the maximum surface temperature T_s. Above T_s = 600 K, however, T_d is smaller than T_s. The experimental observations are analyzed successfully with the first order rate equation for desorption.     

Electron stimulated desorption of carbon monoxide from a (111) niobium surface

Electron stimulated desorption of CO from the (111) face of a Nb single crystal produced both CO⁺ and O⁺ ions after adsorption at 150°K on a clean surface. When the surface was heated to above 250 °K only O⁺ ions were observed, and this current disappeared as the temperature was increased to 700 °K. Readsorption (at 150 °K) was inhibited following the 700 °K heating. These data indicate the formation on heating of a tightly bound surface phase with very low ionic desorption cross section. Threshold energies for CO⁺ and O⁺ ion production were 10.0 ± 0.5 eV and 19.0 ± 0.5 eV, respectively. The cross section for electron stimulated depopulation of the O⁺ producing phase was (4 ± 1) × 10^?18 cm² for 100 eV electrons.     

A method for assessing the coverage dependence of kinetic parameters: Application to carbon monoxide desorption from iridium (110)

A method for analyzing thermal desorption mass spectra has been developed for determining the coverage dependence of the pre-exponential factor and the desorption energy in the Arrhenius equation for a one-step desorption process. The method, which involves variable heating schedules, is applicable to spectra in which several features appear. However, if the desorption process involves multiple steps, or if substantial desorption from multiple sites occurs for any one coverage, this method cannot be used. The method is applied to CO desorption from the (110) surface of Ir. Three features can be resolved in the desorption spectrum. Both the preexponential factor and the desorption energy vary strongly with coverage, and a compensation effect occurs.     

Decomposition of carbon monoxide on a (110) nickel surface

New investigations of the (110) nickel/carbon monoxide system have been made using low energy electron diffraction (LEED), Auger electron spectroscopy (AES), mass spectroscopy and work function measurements. Room temperature adsorption of CO on the surface was reversible with the CO easily removable by heating in vacuum to 450°K. The CO formed a double-spaced structure on the surface which, however, was unstable at room temperature for CO pressures less than 1×10^?7 torr. Work function changes greater than + 1.3 eV accompany this reversible CO adsorption. Irreversible processes leading to the build-up of carbon, and under certain circumstances oxygen, on the surface were the primary concern of the measurements reported here. These processes could be stimulated by the electron beams used in LEED and AES, or by heating the clean surface in CO. The results of AES investigations of this carbon (and oxygen) build-up, together with CO desorption results could be explained on the basis of two surface reactions. The primary reaction was the dissociation of chemisorbed CO leaving carbon and oxygen atomically dispersed on the surface. The second reaction was the reduction of the surface oxygen by CO from the gas phase. The significance of the dissociation reaction to COdesorption studies is discussed.     

Thermal desorption of molecular oxygen from SnO2 (110) surface: Insights from first-principles calculations

First-principles density functional theory calculations in the generalized gradient approximation, with plane wave basis set and pseudopotentials, have been used to investigate the desorption pathways of molecular oxygen species adsorbed on the SnO₂ (110) surface. Energetics of the thermodynamically favored precursors is studied in dependence on the surface charge provided either by surface defects or by donor type impurities from the near-surface region. The resonant desorption modes of O₂ molecules are examined in the framework of ab initio atomic thermodynamics and relationship of these results to experimental observations is discussed.     

Mechanisms of the catalytic carbon monoxide oxidation on Pt (110)

The kinetics of the carbon monoxide oxidation on a clean Pt (110) crystal were investigated in an ultra-high vacuum system by utilizing Auger electron spectroscopy, low-energy electron diffraction and residual gas analysis. Two different catalytic reaction mechanisms were found to prevail for the experimental conditions chosen. In the temperature range, 100 < T ? 220°C, where essentially CO was preadsorbed on the Pt surface the subsequent adsorption of O₂ was competitive and the reaction exhibited the characteristics of a Langmuir-Hinshelwood mechanism. In this case the onset of the CO₂ formation was delayed by a characteristic time which depended strongly on temperature (“induction period”). A simple model for the Langmuir-Hinshelwood reaction was developed which permitted a more detailed evaluation of the kinetic curves yielding an activation energy for the catalytic reaction of 2.9 kcal/mole. On the other hand, when oxygen was preadsorbed on the Pt surface (T > 90°C) the subsequent reaction with CO occurred immediately and was temperature independent. This behavior was interpreted in terms of an Eley-Rideal mechanism. Both reactions were used for titration of the adsorbed species. From area measurements under the titration curves it was concluded that the saturation coverage for CO and oxygen on Pt(110) is approximately the same.     

Formic acid desorption from graphitized Ni(110)

The adsorption/desorption behavior of formic acid from a monolayer of graphite carbon on Ni(110) was studied using AES, LEED and flash desorption spectroscopy. Formic acid adsorbed at 165 K did not form multilayers of adsorbate. Instead, due to strong hydrogen-bonding interactions the formic acid formed a two-dimensional condensed phase on the surface and exhibited zero-order desorption kinetics initially for a 30-fold change in initial coverage. The zero-order desorption rate constant was k_d = 10¹⁸ exp[?68.2 kJ mol^?1/RT]s^?1, suggesting a desorption transition state with nearly full translational and rotational freedom on the surface. The desorption kinetics and the coverage limit were consistent with the formation of a surface polymer-monomer equilibrium.     

Thermal desorption spectroscopy of Ni,Cu, Ag and Au from W(110)

Ni, Cu, Ag and Au on W(110) in the submonolayer range are studied by thermal desorption spectroscopy with the goal of obtaining information on lateral interactions and on the phase state of the adsorption layer in the temperature range in which desorption occurs. Ni, Cu and Ag are found to desorb over a wide coverage range from the two-phase (vapor-condensate) region while Au desorbs only from the single-phase vapor region. Segments of the coexistence curve are determined. The desorption energies have the following limiting values: 4.35–4.95. 3.2–3.85, 2.8–3.55 and 3.3–4.1 eV for Ni, Cu, Ag and Au, respectively.     

Preferential ion-induced desorption of adsorbed carbon monoxide from an iron (111) single crystal surface

Adsorption-desorption of CO on an iron (111) single crystal surface has been studied by a combined technique of ion-induced desorption (IID) and thermal desorption spectroscopy (TDS). Only a non-dissociative adsorbate (α-state) is preferentially sputtered by ion bombardment of energy 3 keV, while a dissociative adsorbate (β-state) is much less sputtered at this energy.     

Stimulated desorption from CO chemisorbed on Cr(110)

Electron stimulated desorption (ESD) experiments using a time-of-flight pulse counting method are reported for molecular CO chemisorbed on the Cr(110) surface at 90 K. Consistent with previous qualitative observations, negligible CO⁺ and O⁺ desorption signals were measured from the ₁CO overlayer which saturates at 1/4 monolayer. For θ_CO > 0.25, a terminally-bonded (₂CO) binding mode is populated in addition to the existing ∝₁CO binding mode and the ion yield increases sharply. For ₂CO, both O⁺ and CO⁺ ions are observed; the CO⁺ ions desorb with characteristically lower kinetic energies than O⁺ ions. Near saturation coverages of CO(ads), an observed decrease in the O⁺ yield is attributed to adsorbate-adsorbate interactions which reduce the ion desorption probability, as seen in ESD studies of terminally-bonded CO on other metals. These results are considered in the context of two possible models proposed for the ₁CO binding state and related ESD observations for CO chemisorbed on Fe(001) and potassium-promoted Ru(001).     

Thermal desorption of cyanogen from Pt(100)

Thermal desorption of cyanogen adsorbed on Pt(100) was studied by flash desorption mass spectrometry. By investigating the parent ion and all possible fragmentation products in the mass spectrometer during desorption it was concluded, that desorption takes place exclusively as molecular C₂N₂. Three desorption peaks were observed at 140, 410 and 480°C denoted as α, β₁ and β₂. The respective surface coverages at saturation were determined by quantitative evaluation of the flash desorption curves to be 2.0 ± 0.2 × 10¹⁴ and 5.5 ± 1.0 × 10¹⁴$\text{molecules}\text{cm}{}^2$ for the α and the β states, respectively. First order desorption kinetics was suggested by the coverage dependencé of the desorption spectra for both α and β states with desorption energies of 12 and 38–42 $\text{kcal}\text{mole}$, respectively. A large difference in the sticking probabilities of α and β states was observed with initial values of 0.06 (α) and 0.9 (β). Adsorption experiments at elevated temperatures led to the assumption, that α and β states coexist on the surface with no or very little interactions between them. The results are discussed in terms of different models for the adsorption states.     

Fractional-order desorption of D2 from a Pd(110) surface

The desorption of deuterium from the low temperature (α₂) state on a Pd(110) surface was studied by TDS, LEED and Δφ. Simultaneous measurements of the isothermal desorption of the α₂ state, intensity of a half-order beam of the (1 × 2) phase and work function change show clearly that the desorption of this low temperature state is associated with the phase transition (1 × 2) to (2 × 1). This transition correlation with a work function change of ∼ 70 mV. The desorption follows nearly zero-order kinetics with an activation energy of ∼ 71 kJ mol⁻¹. The desorption kinetics are thus closely analogous to those on Ni(110).     

Electron impact desorption of Xe from the tungsten (110) plane

The electron stimulated desorption of Xe adsorbed on the clean and on oxygen and CO covered tungsten (110) surfaces has been investigated. Only neutral Xe desorption was observed; for Xe on clean W a very small initial regime with cross section 10^?17cm² is followed by a slow decay with cross section 3×10^?19cm². The Xe yield varies nonlinearly with coverage, suggesting desorption from edges of islands or from sites with less than their full complement of nearest neighbor Xe atoms. Desorption from oxygen or CO covered surfaces results in an apparent desorption cross section identical to that of the underlying adsorbate. This results from a kicking off of Xe by electron desorbed O or CO. The true cross sections for these processes are ～10^?14cm² for Xe-0 and ～10^?15 cm² for Xe-CO. Some speculations about the mechanism, particularly the absence of ions are presented.     

Interaction of oxygen, carbon monoxide and nitrogen with (001) and (110) faces of molybdenum

A combination of low-energy electron diffraction and retarding potential measurements was employed to study gaseous adsorption on atomically clean (001) and (110) Mo single crystal surfaces. Adsorption of oxygen on the (001) surface at room temperature occurred with a sticking coefficient close to unity and produced a large increase in work function and appreciable changes in the intensity distributions of the integral order diffraction beams, without the appearance of any new diffraction beams. These results indicate that a surface monolayer of oxygen was formed with a unit mesh having the same dimensions as that of the underlying molybdenum surface. Exposures above 6 × 10⁻³ Torr-sec produced a uniform decrease in intensities, thus indicating a second monolayer with amorphous structure. On heating, two additional surface structures were observed, characterized by one-half and one-third order beams, respectively. A clean surface was obtained by heating above 1100 °C. An exposure of 1 to 7 × 10⁻⁷ Torr-sec of oxygen for the (110) face resulted in two types of patterns characteristic of lattices with one-quarter and one-half the surface density of the (110) Mo face, with an increased work function accompanying the latter pattern. Exposure of the clean surface at 400 to 800 °C produced similar patterns of enhanced intensities with no increase in work function. Possible models are discussed. It is concluded that place exchange models account for these results, as well as the one-half and one-third order structures on the (001) face, in a more satisfactory manner than adsorption above the surface. An exposure to 10⁻⁵ Torr-sec produced a monolayer coverage with a unit mesh similar to that of the molybdenum substrate. Additional exposure resulted in further amorphous adsorption. Adsorption of CO produced changes in the intensity distributions, with the appearance of no new maxima, for both (001) and (110) Mo surfaces. Nitrogen, at an exposure of 3 × 10⁻³ Torr-min did not adsorb on either the (001) or (110) Mo surface, but when dissociated by electron impact it adsorbed on both Mo surfaces with the same dimensions of unit mesh as those of the Mo substrates and with an increase in work function of 1.05 eV for the (001) and 0.05 eV for the (110) surface.     

Adsorption of O on Mo(110) surface from first-principles calculation

First-principles calculations based on density functional theory (DFT) have been performed to study O adsorption in on-surface and subsurface sites. For different coverages, hollow site is found to be the most stable on-surface adsorption site. For the subsurface adsorption at the bare Mo surface, the adsorption energies are found to be higher than those at the on-surface sites, suggesting that these sites are less stable. However, the presence of preadsorbed O overlayer enhances the binding energy of subsurface adsorption, particularly for the adsorption of O at octahedral site. Further, vibrational frequencies, work-function and density of states are presented for O adsorption in on-surface sites.     

The interaction of hydrogen and carbon monoxide with oxygen on Cu(110)-Fe surfaces

The interaction of hydrogen and carbon monoxide with oxygen adsorbed on Cu(110)-Fe surfaces has been studied with ellipsometry, Auger electron spectroscopy and low energy electron diffraction. With carbon monoxide copper can be reduced completely and Fe_0.95O partially. With a model which is only an extension of the scheme for the reduction of pure Cu(110) by CO, the reduction of Cu(110)-Fe can be simulated. The lateral orientation of Fe_0.95O with respect to the copper matrix changes during repetitive oxidation-reduction cycles. At 725 K oxygen deficient iron oxide segregates to the surface. With hydrogen all oxygen can be removed.     

The adsorption of carbon monoxide on TiO2(110) supported palladium

Palladium overlayers deposited on TiO₂(110) by metal vapour deposition have been investigated using LEED, XPS and FT-RAIRS of adsorbed CO. Low coverages of palladium (<3 ML) deposited at 300 K adsorb CO exclusively in a bridged configuration with a band (B₁ at 1990 cm⁻¹) characteristic of CO adsorption on Pd(110) and Pd(100) surfaces. When annealed to 500 K, XPS and LEED indicate the nucleation of Pd particles on which CO adsorbs predominantly as a strongly bound linear species which we associate with edge sites on the Pd particles (L* band at 2085 cm⁻¹). Both bridged and linear CO bands are exhibited as increases in reflectivity at the resonant frequency, indicating the retention of small particle size during the annealing process. Palladium overlayers of intermediate coverages (10–20 ML) deposited at 300 K undergo some nucleation during growth, and adsorbed CO exhibits both absorption and transmission bands in the B₁ (1990 cm⁻¹) and B₂ (1940 cm⁻¹) regions. The latter is associated with the formation of Pd(111) facets. Highly dispersed Pd particles are produced on annealing at 500 K. This is evidenced by the dominance of transmission bands for adsorbed CO and a significant concentration of edge sites, which accommodate the strongly bound linear species at 300 K. Adsorption of CO at low temperature also allows the identification of the constituent faces of Pd and the conversion of Pd(110)/(100) facets to Pd(111) facets during the annealing process. High coverages of palladium (100 ML) produce only absorption bands in FT-RAIRS of adsorbed CO associated with the Pd facets, but annealing these surfaces also shows a conversion to Pd(111) facets. LEED indicates that at coverages above 10 ML, the palladium particles exhibit (111) facets parallel to the substrate and aligned with the TiO₂(110) unit cell, and that this ordering in the particles is enhanced by annealing.