Search

We have developed a system for rapid data acquisition from low‐energy electron diffraction (LEED) patterns. The essential components of the system are a sensitive silicon intensified target video camera, a commercial video interface, and a PDP 11/23 microcomputer. The system is designed to acquire up to ten diffraction spot intensities and profiles (total) in a short time relative to the time scales of typical surface reactions. The data acquisition system is used in conjunction with standard Varian■ four‐grid LEED optics in an ultrahigh vacuum chamber. The abilities of the system with respect to versatility, speed, linearity, and signal‐to‐noise are demonstrated, and data from a study of oxygen chemisorption on Pd(100) are presented.

We monitor Bragg intensity oscillations, during the growth of Pt films on Pd(100), using a conventional LEED apparatus. Distinct oscillations are observed even at 80 K, where thermally activated surface diffusion is negligibly slow. We have shown previously that the oscillations are consistent with the requirement of a four-fold hollow adsorption site. The experimental data suggest that surface diffusion switches on around T = 150 K. In an effort to extend our model to the temperature range where diffusion begins, we present have a preliminary analysis of master equations which incorporate diffusion and the appropriate adsorption-site requirement. The model predicts that typically growth becomes more layer-by-layer-like with the onset of diffusion, mimicking the experimental data for Pt Pd (100) . The role of diffusion is complex, since lateral (intralayer) diffusion leads to clustering, creation of more adsorption sites, and a diminution of layer-by-layer-like growth. On the other hand, downward or “lemming” (interlayer) diffusion makes growth more layer-by-layer-like.

Electron energy loss spectroscopy has been used to study the vibrational energies of NO and CO coadsorbed on the Ru(001) surface. Adsorption of CO competes directly with adsorption of one of the two molecular states of NO. This is interpreted in terms of the geometry of the adsorption sites for CO and NO and is consistent with models proposed previously for adsorption of CO and NO separately on the Ru(001) surface.

Water can displace cyclohexane from chemisorption sites on a Ru(001) surface. In so doing, thermal desorption spectroscopy indicates that the water forms three-dimensional, hydrogenbonded clusters which are two layers deep, much as it would on the clean surface. Water forces the cyclohexane into weakly bound states. This is apparently driven by a combination of effects: first, a single water molecule is more strongly bound to the ruthenium surface than a cyclohexane molecule and second, the packing density of water in the hydrogen-bonded bilayer is much greater than the packing density of the chemisorbed cyclohexane. Taking both effects into consideration, replacement of a chemisorbed cyclohexane layer with a water bilayer is energetically favored by 5 to 8 kcal per mole of surface metal atoms. The data also show that the preferential chemisorption of water on ruthenium is not limited by the kinetics of diffusion or displacement in these experiments, but rather reflects a true equilibrium state.

It is known that clean Pt(100) normally exists in a reconstructed (‘‘hex’’) phase, that a metastable clean unreconstructed phase [the (1×1)] can be prepared, and that during adsorption of CO the hex→(1×1) transition occurs. In this and the following paper, we describe an investigation of the CO/Pt(100) system which clarifies the mechanism of this adsorbate‐induced Pt phase transition. The experimental techniques included fast video‐LEED techniques, thermal desorption spectroscopy, electron energy loss spectroscopy, and work function measurements. Adsorption of CO at low temperature (T≲400 K) is discussed in this paper. CO on (1×1)‐Pt forms a c(2×2) overlayer near and at the ideal coverage of θ=0.5, in which the CO molecules occupy on‐top adsorption sites. Repulsive CO–CO interactions cause this structure to form. CO adsorption on hex‐Pt proceeds via formation of small areas with the same local structure, a c(2×2) layer of CO on (1×1)‐Pt, even when the total coverage of CO is low (0.05

We have investigated the temperature and coverage‐dependent transformations of the ordered structures of oxygen on Pd(100). The four ordered structures are the chemisorbed c(2×2) and p(2×2) lattices, and the reconstructed p(5×5) and ((5)1/2×(5)1/2)R27° lattices. We present evidence that the p(5×5) reconstruction forms in an activated stepfrom c(2×2) regions. The onset of p(5×5) formation is associated with an increase in oxygen sticking coefficient. In thermal desorption, there are three states which can be correlated directly with the structure of the adsorbed phase during desorption: At lowest coverage, the α state shows the traits of second‐order kinetics and is due to desorption from a disordered adlayer. At higher coverage, the β state appears and is due to desorption from a layer with c(2×2) order. There is a lower barrier to desorption in the β state than in the α state because of the repulsive second‐nearest‐neighbor interactions in the c(2×2). At highest coverage, the sharp and narrow γ state eme...

The adsorption rate for O2 on Rh(111) at 335 K exhibits Langmuir (1 − θ) kinetics rather than the expected (1 − θ)2 dependence. Adsorption occurs with a sticking coefficient near unity on empty sites, producing a mixture of ordered and disordered domains. At θ ⩽ 0.5, the intensity of the(1, 12)overlayer beam increases in proportion to θ2 as expected for ordered oxygen island formation. Extensive ordering occurs only in the very last stages of monolayer formation. The reaction of the ordered-O layer with H2(g) to produce H2O takes place rapidly above 275 K in the 10−8−10−6 Torr range and can be followed quantitatively by observation of the(1, 12) LEED beam intensity. The reaction accurately exhibits kinetics which are first order in H2 pressure, implying an efficient trapping mechanism for adsorbed hydrogen on the ordered-O layer. The activation energy for reaction is5.3 ± 0.3 kcalmole; each reaction event leads to a loss of order from a ∼-300 A2 domain. This suggests that near saturation coverage, extensive disordering occurs for small decreases in ordered oxygen coverage during reaction with Hz(g) and is consistent with the massive ordering behavior seen in the final stages of oxygen adsorption. The reaction of H2(g) with disordered oxygen species follows approximate half-order kinetics in PH2 as expected for a mechanism which does not involve hydrogen atom trapping.

The influence of chemisorbed nitrogen and oxygen atoms on the subsequent adsorption of NO on Ru(001) at 150 K has been investigated by high resolution electron energy loss spectroscopy. The presence of the NO dissociation products can block adsorption of one of the two states of molecular NO which are populated on the clean Ru surface.

The adsorption of H2O at low temperature on clean and oxygen dosed nickel (111) has been studied using thermal desorption spectrometry (TDS) and electron-stimulated desorption (ESD). Both the single and multilayer H2O regimes have been investigated. For first layer chemisorbed water, the thermal desorption data can be modeled well assuming first order kinetics with a desorption energy of about 9.8 kcalmol. For ice multilayers, a leading edge analysis gives a desorption energy of 11.5 ± 0.5 kcalmol. ESD measurements detected both H+ and H2O+ ions, and several types of ESD measurements were made: determination of ion yield versus electron beam energy, determination of ion kinetic energy distribution curves (lEDC's), and determination of the temperature dependence of the ion yield (TDIY) at fixed excitation energy. The latter measurements show a correlation with the thermal desorption data: TDIY of H2O+ during desorption of multilayer H2O films gives 11.0 ± 0.5 kcalmol for the desorption energy, in excellent agreement with the TDS results. Additional features are resolved in the temperature dependence of the ESD signal which are absent in the TDS spectra. Characteristic ion yield and ion kinetic energy distribution curves are given for adsorbed layers on the clean surface in both the single and ice multilayer regime. For the oxygen predosed surface, TDS was used to isolate a H2O desorption state at higher temperature which is due to dissociative recombination, and ESD was used to characterize the ion yields and IEDC's of H+ from hydroxyl groups adsorbed on this type of surface. Comparison of the data for H+ ions from adsorbed water and hydroxyls indicates that ESD in this case is sensitive primarily to the local parent O-H bond, and insensitive to secondary chemical bonds.

A commercial, conventional low‐energy electron diffraction apparatus is used to monitor Bragg intensity oscillations during the growth of Pt on Pd(100). The effect of substrate temperature between 80 and 400 K is investigated. Between 80 and 300 K, two to three Bragg oscillations are observed. The oscillation amplitude damps out quickly as film coverage increases at fixed temperature, but damp out less quickly at the higher substrate temperatures. Above ∼350 K, reconstruction of the Pt overlayer interferes with the oscillations. These data indicate that a kinetic barrier, most probably the barrier to surface diffusion, inhibits the system from achieving macroscopic equilibrium, and that the true equilibrium growth mode for this system is layer‐by‐layer. A new, analytical procedure is used to determine the coverage distribution within the layers from the Bragg intensities during growth. Bragg oscillations are predicted to occur at low substrate temperatures where surface diffusion is minimal and deposition...

Utilisation d'une nouvelle methode d'emission IR pour realiser les premieres mesures de la largeur des bandes du mode vibrationnel support-molecule sur une surface de monocristal bien caracterisee. Pour un recouvrement de saturation, la largeur de bande observee pour le mode de C−Ni, pour CO sur Ni(100) est de 15 cm −1

The mechanism of the (5×20) → (1times;1) transition of Pt(100) during adsorption of CO has been investigated using a fast Video-LEED technique. Analysis of the coverage-dependence of the intensities of several LEED spots on different surfaces leads to a straightforward model. The phase transition occurs by a nucleation-trapping mechanism of the adsorbed CO. The driving force is the difference in stabilities of CO adsorbed on the reconstructed and unreconstructed Pt surfaces.

As part of a larger study of small fluorinated molecules on metal and semiconductor surfaces we investigated the adsorption of NF3 on Ru(0001) using measurements of electron‐stimulated desorption ion angular distribution (ESDIAD), low‐energy electron diffraction (LEED), thermal desorption spectroscopy, and Auger electron spectroscopy (AES). At 100 K, NF3 adsorbs with a high sticking probability. Some molecular NF3 desorbs on heating, with a peak maximum at 200 K. All molecular desorption is complete by 400 K, but AES shows that nitrogen and fluorine remain on the surface up to 1100 K. The LEED pattern of the majority of the surface remains (1×1) at all coverages and annealing conditions. At most locations on the Ru(0001) surface (planar regions), ESDIAD is dominated by F+ emission normal to the surface over the temperature range 100–1100 K. On stepped regions of the surface (near edges and defects) highly directional and intense hexagonal patterns of F+ emission are observed (200–1100 K). The off‐normal beams desorb at 67°±3° away from the normal. It appears that most ESDIAD patterns in this system are derived from F or NFx (x

We describe thermal desorption studies of both fluorinated and hydrogenated diethers on the Ru(001) surface. We find that 1,1′-[methylenebis(oxy)]bisethane (CH3CH2OCH2OCH2CH3) chemisorbs on Ru(001) with a bond strength of 53–69 kJ mol . About 0.05 monolayer decomposes. For comparison, 0.04 monolayer of 1,2-diethoxyethane (CH3CH2OCH2CH2OCH2CH3) decomposes. The chemisorbed diethoxyethane molecules desorb in three states. The desorption energy of the majority state is 58–62 kJ mol . 1,2-Perfluorodiethoxyethane (CF3CF2OCF2CF 2OCF2CF3) bonds less strongly (44–47 kJ mol ) to Ru(001) than its hydrogenated analog, diethoxyethane. No perfluorodiethoxyethane decomposes. The diethers bond between 12% and 35% more strongly than the monoether analogs.

We have investigated the mechanism and driving force of the CO‐induced phase transition of the Pt(100) surface. In a preceding paper, we have concentrated on the mechanism by which CO removes the surface reconstruction. As discussed, the clean reconstructed (hex) surface of Pt(100) is more stable than the unreconstructed (1×1) phase, but if the CO coverage on hex exceeds a small critical value (θ≳0.05) nucleation of (1×1) patches occurs and proceeds until, at θ=0.5, the entire surface has been converted to the (1×1) phase. In this paper we present data which quantitatively describe the energetics of this system. Values for the heats of adsorption of CO on both the reconstructed and unreconstructed phases, as well as values for the preexponential factors for the desorption rates, are determined from quasiequilibrium LEED measurements at different coverages. The difference in the low coverage heats of adsorption of CO on the hex and (1×1) phases (27.5 vs 37.5 kcal/mol) is the driving force in the Pt phase transformation during adsorption. On the other hand, during desorption, the hex phase does not return until the coverage has decreased below 0.3. This measurement allows an estimation of the difference in stability between the clean (1×1) and hex phases: about 9–13 kcal per mole of Pt atoms. The transition between the two phases which are in equilibrium [COgas+hex‐Pt(100) vs COads+(1×1)‐Pt(100)] shows a hysteresis due to kinetic limitations. These kinetic effects are characteristic of a nucleation process involving a critical coverage.

The adsorption of CO on Rh(111) has been studied by thermal desorption mass spectrometry and low-energy electron diffraction (LEED). At temperatures below 180 K, CO adsorbs via a mobile precursor mechanism with sticking coefficient near unity. The activation energy for first-order CO desorption is 31.6 kcal/mole (νd = 1013.6s−1) in the limit of zero coverage. As CO coverage increases, a (√3 ×√3)R30u overlayer is produced and then destroyed with subsequent formation of an overlayer yielding a (2 × 2) LEED pattern in the full coverage limit. These LEED observations allow the absolute assignment of the full CO coverage as 0.75 CO molecules per surface Rh atom. The limiting LEED behavior suggests that at full CO coverage two CO binding states are present together.

A metastable c(2 x 2) structure is formed by atomic oxygen on Pd(100) under conditions of slow diffusion combined with rapid adsorption. We propose that formation of the c(2 x 2) under these conditions must be controlled by characteristics of the dissociative adsorption event itself.

Influence de la variation de la vitesse de chauffage sur la distribution des molecules dans des etats d'adsorption separes. Etude sur deux exemples: eau et benzene se desorbant de Ru (001). Identification des especes metastables adsorbees

A simple manipulator is described for use in ultrahigh vacuum. Unhindered rotation within vacuum is provided by a commercial differentially pumped adapter to which the entire XYZ stage is attached. The combination of movements provided by the XYZ stage, the rotating vacuum flange, and a gimbal permits alignment of the sample with any of the peripheral view ports. The resistively heated sample support, mounted through a cold finger, permits rapid variation of sample temperature from several hundred degrees Kelvin to near cryogenic. It is anticipated that this design could be easily used with many existing types of commercial vacuum systems, with the consequent advantage of increased ease and simplicity of both mechanical motion and cryogenic cooling.

We present the results of an investigation of Rh films on Ag(100). The films are studied using Auger-electron spectroscopy, low-energy electron diffraction, x-ray photoelectron spectroscopy, ultraviolet photoemission spectroscopy, ion-scattering spectroscopy, and scanning Auger microscopy. Overlayer characteristics are examined at substrate temperatures of 300 and 600 K. We find that the equilibrium configuration is not predicted by any of the three traditional growth modes (Frank--Van der Merwe, Stranski-Krastanov, or Volmer-Weber). Rather, the equilibrium film structure is that of a Ag-Rh-Ag sandwich, most probably flat. Formation of the sandwich is thermodynamically driven by the difference in surface free energies between Ag and Rh, and is kinetically accessible because of the high mobility of the Ag atoms.

Atomic oxygen forms a metastable c(2×2) phase on Pd(100) under conditions of rapid adsorption (high pressure) and slow diffusion (low sample temperature). One possible explanation is that oxygen molecules require an 8-fold ensemble of empty sites for dissociative chemisorption, and that subsequent adatom motion is limited and creates no neighboring pairs of filled sites. We describe the properties of the adlayer predicted by such a model.

The co‐adsorption of CO and H2 on Rh(111) at low temperature (∼ 100 K) has been studied using thermal desorption mass spectrometry (TDS) and Low‐Energy Electron Diffraction (LEED). The probability of adsorption of CO on rhodium pretreated with hydrogen has been found to decrease slowly with increasing amounts of hydrogen on the surface. In addition, the effect of surface hydrogen on the CO LEED patterns indicates segregation of hydrogen and CO. These results can be explained in terms of a strong repulsive CO–H interaction and a mobile precursor model of CO adsorption.

The purpose of this review is to compare and discuss recent experimental and theoretical results in the field of H2O-solid interactions. We emphasize studies of low (submonolayer) coverages of water on well-characterized, single-crystal surfaces of metals, semiconductors and oxides. We discuss the factors which influence dissociative versus associative adsorption pathways. When H2O adsorbs molecularly, it tends to form three-dimensional hydrogen-bonded clusters, even at fractional monolayer coverages, because the strength of the attractive interaction between two molecules is comparable to that of the substrate-H2O bond. The template effect of the substrate is important in determining both the local orientation and long-range order of H2O molecules in these clusters. The influence of surface additive atoms (e.g., O, Br, Na, K) and also surface imperfections (e.g. steps and defects) on the surface structure and chemistry of H2O is examined in detail. Some results on single-crystal substrates are compared with earlier measurements of H2O adsorption on high-area materials.

The adsorption of H2O on Ru(001) at temperatures of 95 K and 165 K has been investigated using electron energy loss spectroscopy, low-energy electron diffraction, and ultraviolet photoelectron spectroscopy. The vibrational data, together with the structural information available, supports a model in which aggregates of water molecules form via population of a first (chemisorbed) layer and subsequent hydrogen bonded layers. At the higher temperature of adsorption, 165 K, major differences in the vibrational spectra are apparent. Possible reasons for these differences are discussed.

We report a surface science study of fluorinated and hydrogenated ethers adsorbed at a metal surface. Fluorinated ethers bond less strongly to ruthenium surfaces than the hydrogenated analogs, both for atomically-smooth and atomically-rough surfaces. For the hydrogenated ethers, 0.03 to 0.17 monolayers undergo decomposition, while fluorinated ethers do not decompose significantly.

The authors report studies of the adsorption bond strengths and decomposition of perfluorodiethyl ether and hexafluoroacetone on clean and oxygen-dosed Ru(001). Perfluorodiethyl ether bonds weakly to Ru(001) (39 kJ/mol) without decomposing. This binding energy is typical for a metal-adsorbate interaction in which the adsorbate donates electrons to the metal from the oxygen lone pair. Hexafluoroacetone bonds to the surface in both a weakly bonded configuration (38 kJ/mol) and a strongly bonded configuration (66-87 kJ/mol), and there is some decomposition. They compare the chemistry of these oxygenated fluorocarbons to analogous oxygenated hydrocarbons. They propose that the net effect of fluorination is to weaken those chemisorption bonds which rely upon electron donation from the oxygen lone pairs to the metal but to strengthen the chemisorption bonds which rely primarily upon electron donation from the metal into antibonding C-O orbitals. Both phenomena can be explained by electronic effects resulting from fluorination.

International audience; We present a study of the surface of the-Al-Pd-Mn approximant phase based upon scanning tunneling microscopy and low-energy electron diffraction. Several structures are observed on two different samples grown either by the Bridgman technique or by a self-flux method, and which contain various degrees of disorder. We also describe some other complex crystalline phases that are sometimes observed on the fivefold surface of Al-Pd-Mn quasicrystalline samples after the sputter-annealing cleaning process under ultrahigh vacuum conditions. This includes the T approximant phase resulting from surface decomposition after a high-temperature annealing.

It is well-established that three features are resolved in the thermal desorption spectra of H 2 O from clean Ru(001): one at low temperature due to desorption from ice multilayers, and two at higher temperatures due to desorption from a chemisorbed bilayer. However, we find that for D 2 O the highest-temperature state is strongly suppressed. By varying the heating rate during H 2 O desorption, we obtain evidence that the highest-temperature state can be formed by conversion from the intermediate state. Conversion is much slower for D 2 O than for H 2 O, so that direct desorption is kinetically favored over conversion and the high-temperature state is suppressed for D 2 O. The rate of conversion is about 3 to 8 times slower for D 2 O than for H 2 O, which is consistent with a model in which the rate-determining step for conversion is rotation of one or more water molecules within a hydrogen-bonded cluster. This work provides the first evidence for an isotope effect in thermal desorption of water. An important ramification is that adsorbed D 2 O is not always interchangeable with H 2 O for chemical studies of isotopic mixing or for spectroscopic studies of vibrational frequency shifts, as is commonly assumed.

In the present work, we have investigated the adsorption of cyclohexane on Ru(001), using EELS (both in the specular and off-specular directions), angle-resolved photoelectron spectroscopy, and TDMS (combined in one UHV system) to investigate the orientation and symmetry of the adsorbed molecule and compare the results with the above-mentioned structural model [3].

We describe a simple, inexpensive electrical circuit which protects a turbomolecular pumped chamber against vacuum loss in the event of a power failure. The circuit is timed to ensure proper sequencing of events, and also to distinguish between a short and long power failure.

The adsorption of oxygen on Rh(111) at approximately 100 K has been studied by thermal desorption mass spectrometry (TDS), Auger electron spectroscopy (AES), and low energy electron diffraction (LEED). Oxygen adsorbs in a disordered state at 100 K and orders irreversibly into an apparent (2 x 2) surface structure upon heating to T ≥ 150 K. The kinetics of this ordering process have been measured by monitoring the intensity of the oxygen (1,1/2bar) LEED beam as a function of time with a Faraday cup collector. The kinetic data fit a model in which the rate of ordering of oxygen atoms is proportional to the square of the concentration of disordered species due to the nature of adparticle interactions in building up an island structure. The kinetic data are shown in Fig 1 for five temperatures at which the ordering was measured, where f(N0) is the integrated form of the kinetic rate law. On the basis of these data, the activation energy for ordering is 13.5±0.5 kcal/mol. This is consistent with values reported for oxygen ordering on Ir(111), 16-19 kcal/mol [1], and for oxygen migration on Ru, 14 kcal/mol [2].

We have determined the number of atomic layers necessary to support the bulk-like (5×1) reconstruction in a thin film of platinum on a Pd(100) substrate. The Pt film grows layer-by-layer, epitaxially, up to at least three atomic layers at room temperature. It reconstructs at ∼420 K, similar to surfaces of bulk Pt. Reconstruction takes place at Pt coverages above 1/3 monolayer, i.e. only one atomic layer of Pt is necessary to support the reconstruction in this system.

We have investigated adsorption of water on Ru(100) using electron energy loss spectroscopy and thermal desorption spectroscopy. On this row-and-trough substrate, there is a single desorption state associated with the chemisorbed layer, which shifts down from 240 to 180 K as coverage increases. There is no isotope effect, i.e., desorption spectra of H 2 O are the same as those of D 2 O, within experimental error. Less than 0.1 monolayers of H 2 O dissociate and exchange between chemisorbed hydrogen and D 2 O is not measurable. Pre-adsorbed oxygen and hydrogen cause distinctive changes in the shapes of the water desorption peaks on Ru(100). Electron energy loss spectroscopy indicates that the extent of hydrogen bonding increases as water coverage increases at 85 K. Upon annealing and desorption, most changes in the vibrational spectra can be attributed to decreasing coverage, with the exception of some changes in the low-frequency modes. Comparison with previous studies of the hexagonally close-packed Ru(001) surface reveals that surface morphology plays a major role in determining the properties of water adlayers on ruthenium surfaces.

Carifornia 91 I25, USA Received 31 January 1983; accepted for publication 6 April 1983 The adsorption of cyclohexane on Ru(001) at 90 K has been investigated by thermal desorption mass spectrometry, EELS, UV photoemission and LEED. Thermal desorption indicates the adsorption of the undissociated molecule first in a chemisorbed monolayer (Td = 200 K) with subsequent formation of multilayers (Td = 165 K) at higher exposures. The vibrational spectrum obtained by EELS is characterized by a frequency shift of the C-H stretching mode from 2920 cm-’ (multilayer) to 2560 em-’ for the chemisorbed monolayer. Off-specular EELS data indicate two different electron scattering mechanisms for the C-H stretching mode. Whereas for the C-H stretching mode of the multilayer, large angle electron impact scattering is observed, the C-H soft-mode of the monolayer is largely due to small angle dipolar scattering. The He I photoelectron spectra of cyclohexane multilayers are characteristic of the undissociated molecule. A new assignment of C(2s) and the lowest C(2p) level, based on a comparison with benzene, shows that the chemisorbed monolayer is characterized by the absence of emission or broadening of the 2a,, level. This is attributed to C,, symmetry of the chernisorbed layer and to a possible interaction of the 2a,, orbital with the metal surface.