Dissipative dynamics within the electronic friction approach: the femtosecond laser desorption of H2/D2 from Ru(0001)

An electronic friction approach based on Langevin dynamics is used to describe the multidimensional (six-dimensional) dynamics of femtosecond laser induced desorption of H(2) and D(2) from a H(D)-covered Ru(0001) surface. The paper extends previous reduced-dimensional models, using a similar approach. In the present treatment forces and frictional coefficients are calculated from periodic density functional theory (DFT) and essentially parameter-free, while the action of femtosecond laser pulses on the metal surface is treated by using the two-temperature model. Our calculations shed light on the performance and validity of various adiabatic, non-adiabatic, and Arrhenius/Kramers type kinetic models to describe hot-electron mediated photoreactions at metal surfaces. The multidimensional frictional dynamics are able to reproduce and explain known experimental facts, such as strong isotope effects, scaling of properties with laser fluence, and non-equipartitioning of vibrational, rotational, and translational energies of desorbing species. Further, detailed predictions regarding translations are made, and the question for the controllability of photoreactions at surfaces with the help of vibrational preexcitation is addressed.     

Six-dimensional potential energy surface for H2 at Ru(0001)

The six-dimensional (6D) potential energy surface (PES) for the H(2) molecule interacting with a clean Ru(0001) surface has been computed accurately for the first time. Density functional theory (DFT) and a pseudopotential based periodic plane-wave approach have been used to calculate the electronic interactions between the molecule and the surface. Two different generalized gradient approximation (GGA) exchange-correlation functionals, PW91 and RPBE, have been adopted. Based on the DFT/GGA calculated potential energies, an analytical 6D PES has been constructed using the corrugation reducing procedure. A very accurate representation of the DFT/GGA data has been achieved, with an average error in the interpolation of about 3 meV and a maximum error not larger than about 30 meV. The top site is found to be the most reactive site for both functionals used, but PW91 predicts a higher reactivity than RPBE, with lower-energy and earlier-located dissociation barriers. The energetic corrugation displayed by the RPBE PES is larger than the PW91 PES while the geometric corrugation is smaller. The differences between the two PESs increase as the distance of the molecular center of mass to the surface decreases. A direct comparison with experimental investigations on H(2)/Ru(0001) could shed light on the suitability of these XC potentials often used in DFT calculations.     

Collision-induced intramolecular vibration—vibration energy transfer in CO2: CO2(0001) + H2/D2

Intramolecular vibration—vibration energy transfer cross sections have been calculated for CO₂(00⁰1) + H₂/D₂ → CO₂(11¹0) + H₂/D₂, → CO₂(10⁰0) + H₂/D₂, and → CO₂(20⁰0) + H₂/D₂ based on the mechanism that the energy mismatch is transferred to the translational motion. For CO₂ + H₂, the calculated cross section for CO₂(00⁰1) + H₂ → CO₂(10⁰0) + H₂ is in good agreement with experimental data. Cross sections for the processes (00⁰1 → 11¹O) and (00⁰1 → 20⁰0) are found to be too small compared with experimental data. For CO₂ + D₂, (00⁰1 → 10⁰0) is also the most important process and appears to represent experimental data at room temperature. The sum of three cross sections of CO₂ + H₂ is always greater than that of CO₂ + D₂ over the temperature range of 100–2500 K.     

Manipulating the activation barrier for H2(D2) desorption from potassium-modified palladium surfaces

In this work the permeation and desorption of hydrogen (deuterium) from potassium-modified Pd(111) and polycrystalline palladium surfaces have been studied in the temperature range from 350 to 523 K. Time-of-flight spectroscopy has been used to determine the translational energy distributions of associatively desorbing H(2)(D(2)) molecules as a function of the potassium coverage and additional isotropic O(2) and CO background pressures. It turned out that the energy distribution of the hydrogen desorption flux is thermalized for the clean Pd surfaces but hyperthermal for the potassium-covered surfaces. The activation barrier for adsorption was found to increase with the potassium coverage but to decrease again in the presence of coadsorbates such as O(2) or CO. Especially by choosing different isotropic CO pressures, the effective desorption barrier for hydrogen could be reversibly decreased and increased, which resulted in the equivalent changes of the mean kinetic energies of the desorbing H(2) molecules.     

XUV free-electron laser desorption of NO from graphite (0001)

We report results of femtosecond laser induced desorption of NO from highly oriented pyrolytic graphite using XUV photon energies of hν = 38 eV and hν = 57 eV. Femtosecond pulses with a pulse energy of up to 40 μJ and about 30 fs duration generated at FLASH are applied. The desorbed molecules are detected with rovibrational state selectivity by (1 + 1) REMPI in the A(2)Σ(+) ← X(2)Π γ-bands around λ = 226 nm. A nonlinear desorption yield of neutral NO is observed with an exponent of m = 1.4 ± 0.2. At a fluence of about 4 mJ/cm(2) a desorption cross section of σ(1) = (1.1 ± 0.4) × 10(-17) cm(2) is observed, accompanied with a lower one of σ(2) = (2.6 ± 0.3) × 10(-19) cm(2) observable at higher total fluence. A nonthermal rovibrational population distribution is observed with an average rotational energy of = 38.6 meV (311 cm(-1)), a vibrational energy of = 136 meV (1097 cm(-1)) and an electronic energy of = 3.9 meV (31 cm(-1)).     

Potential energy surface intersections in the C(1D)H2 reactive system

Potential energy surface (PES) intersection seams of two or more electronic states from the 1 1A', 2 1A', 3 1A', 1 1A", and 2 1A" states in the C(1D)H2 reactive system are investigated using the internally contracted multireference configuration interaction method and the aug-cc-pVQZ basis set. Intersection seams with energies less than 20 kcal/mol relative to the C(1D) + H2 asymptote are searched systematically, and finally several seam lines (at the linear H-C-H, linear C-H-H, and C(2v), geometries, respectively) and a seam surface (at Cs geometries) are discovered and determined. The minimum energy crossing points on these seams are reported and the influences of the PES intersections, in particular, conical intersections, on the CH2 spectroscopy and the C(1D) + H2 reaction dynamics are discussed. In addition, geometries and energies of the 1 1A2 and 1 1B2 states of methylene biradical CH2 are reported in detail for the first time.     

The scattering of low energy C60 on graphite (0001) surfaces

The interaction between C₆₀ molecules with a graphite (0001) surface has been investigated by means of molecular dynamics simulations. The initial energies of the C60 molecules are 90 and 270 eV, respectively. An empirical model potential suggested by Takai et al. is used to describe the interaction between carbon atoms in the C₆₀ molecule and between the atoms forming the graphite substrate. The interaction between the C₆₀ atoms and the graphite atoms is modeled by a suitable Lennard-Jones potential. The resilience of scattered C₆₀ molecules is observed and its energy distribution is in reasonable agreement with available experimental data, showing no significant dependence of the rebounding translational energy on the incident kinetic energy. The energy partition in the collision has been analyzed in detail and a two-step collision model speculated in the experiments has been discussed based on the simulation results.     

Distinguishing the formation of C2D+4 ions from C2D+6 (by D2 loss) and from C2D5H+ (by HD loss) in a Reflectron spectrometer

The D₂ loss from C₂D⁺₆ ions and the HD loss from C₂D₅H⁺ ions has been investigated in a photoelectron photoion coincidence experiment employing a reflecting ion time of a flight mass spectrometer (Reflectron). The experiment is able to distinguish the metastable formation of C₂D⁺₄ ions (m/z = 32) from C₂D⁺₆ ions by D₂ loss and from C₂D₅H⁺ ions by HD loss simultaneously in a mixture of deuterated ethanes. The breakdown curves of the title reactions are presented and compared to the H₂ loss from C₂H⁺₆ ions. The HD loss from C₂D₅H⁺ is shifted by 67 meV and the D₂ loss from C₂D⁺₆ is shifted by 108 meV with respect to the H₂ loss from C₂H⁺₆. This shift reflects a strong kinetic isotope effect which is most likely due to tunneling of H/D atoms through a barrier.     

Adsorption of CO and O2 on W2C(0001)

CO, O(2), and H(2) adsorption on a clean W(2)C(0001)√13×√13 R ± 13.9° reconstructed surface at room temperature (RT) were investigated using high-resolution electron energy loss spectroscopy (HREELS). The W(2)C(0001) adsorbs CO molecularly and adsorbs O(2) dissociatively, but does not adsorb H(2) at RT. In the CO adsorption system, two C-O stretching (antisymmetric CCO stretching) modes were found at 242.3 meV (1954 cm(-1)) and at 253.0 meV (2041 cm(-1)). The low-frequency site is occupied at first with subsequent conversion to the high-frequency site with increasing coverage. Additionally, a small peak was apparent at 104.5 meV (843 cm(-1)), and a middle peak at 50-51 meV (400-410 cm(-1)), which are assignable to a symmetric stretching mode and a hindered translational mode, respectively, of a CCO (ketenylidene) species. These observations are consistent with the CO adsorption model on top of the surface carbon. For oxygen adsorption, two adsorption states were found at 65.2-68.1 meV (526-549 cm(-1)) and 73.6 meV (594 cm(-1)): typical frequencies to oxygen adsorption on metal surfaces. Results suggest that atomic oxygen adsorption occurred on a threefold hollow site of the second W layer.     

Vibrational energy relaxation at surfaces: O2 chemisorbed on Pt(111)

It is shown that the O-O stretch for O₂ chemisorbed on Pt(111) is likely to be damped via excitation of electron-hole pairs. Vibrational spectroscopic data for this system are analyzed using a model which assumes an adsorbate-induced resonance in the vicinity of the Fermi energy. In accordance with theory, the vibrational line profile is asymmetric and from the deduced linewidth γ, asymmetry parameter τ and vibrational polarizability α_v, we estimate the magnitude of the electron-phonon coupling parameter and derive information as to the nature of the adsorbate-induced resonance state.     

CO Blocking of D2 Dissociative Adsorption on Ru(0001)

The influence of pre‐adsorbed CO on the dissociative adsorption of D₂ on Ru(0001) is studied by molecular‐beam techniques. We determine the initial dissociation probability of D₂ as a function of its kinetic energy for various CO pre‐coverages between 0.00 and 0.67 monolayers (ML) at a surface temperature of 180 K. The results indicate that CO blocks D₂ dissociation and perturbs the local surface reactivity up to the nearest‐neighbour Ru atoms. Non‐activated sticking and dissociation become less important with increasing CO coverage, and vanish at θ_CO≈0.33 ML. In addition, at high D₂ kinetic energy (>35 kJ mol^?1) the site‐blocking capability of CO decreases rapidly. These observations are attributed to a CO‐induced activation barrier for D₂ dissociation in the vicinity of CO molecules.     

n-alkanes on Pt(111) and on C(0001)Pt(111): chain length dependence of kinetic desorption parameters

We have measured the desorption of seven small n-alkanes (C(N)H(2N+2), N=1-4,6,8,10) from the Pt(111) and C(0001) surfaces by temperature programed desorption. We compare these results to our recent study of the desorption kinetics of these molecules on MgO(100) [J. Chem. Phys. 122, 164708 (2005)]. There we showed an increase in the desorption preexponential factor by several orders of magnitude with increasing n-alkane chain length and a linear desorption energy scaling with a small y-intercept value. We suggest that the significant increase in desorption prefactor with chain length is not particular to the MgO(100) surface, but is a general effect for desorption of the small n-alkanes. This argument is supported by statistical mechanical arguments for the increase in the entropy gain of the molecules upon desorption. In this work, we demonstrate that this hypothesis holds true on both a metal surface and a graphite surface. We observe an increase in prefactor by five orders of magnitude over the range of n-alkane chain lengths studied here. On each surface, the desorption energies of the n-alkanes are found to increase linearly with the molecule chain length and have a small y-intercept value. Prior results of other groups have yielded a linear desorption energy scaling with chain length that has unphysically large y-intercept values. We demonstrate that by allowing the prefactor to increase according to our model, a reanalysis of their data resolves this y-intercept problem to some degree.     

The sticking of H and D atoms on a graphite (0001) surface: the effects of coverage and energy dissipation

Classical trajectory methods are used to examine the trapping and sticking of H and D atoms on the graphite (0001) surface. Total energy calculations based on density functional theory are used to construct the model potential energy surface, and graphite clusters of up to 121 atoms are considered. For hydrogen to chemisorb, the bonding carbon must pucker out of the surface plane by roughly 0.4 A. For incident energies above the 0.2 eV barrier, any trapped H atoms must rapidly dissipate their excess energy into the surrounding lattice within a few vibrations of the C-H stretch in order to remain bound. For sufficiently large clusters, the C-H bond stabilizes within about 0.1 ps. The sticking probability for D at 150 K is in the range of 5%-10%, more-or-less consistent with the most recent measurements in the limit of zero coverge. Variation with isotope and substrate temperature is weak. We estimate that the sticking cross section for adsorption at the para site, directly across the sixfold carbon ring from an already adsorbed H atom, can be four or more times larger that the zero coverage sticking cross section.     

Scanning tunneling microscopy of sulfur and benzenethiol chemisorbed on Ru(0001) in 0.1 M HClO4

In situ scanning tunneling microscopy (STM) combined with linear sweep voltammetry was used to examine spatial structures of sulfur adatoms (SA) and benzenethiol (BT) molecules adsorbed on an ordered Ru(0001) electrode in 0.1 M HClO4. The Ru(0001) surface, prepared by mechanical polishing and electrochemical reduction at -1.5 V (vs RHE) in 0.1 M HClO4, contained atomically flat terraces with an average width of 20 nm. Cyclic voltammograms obtained with an as-prepared Ru(0001) electrode in 0.1 M HClO4 showed characteristics nearly identical to those of Ru(0001) treated in high vacuum. High-quality STM images were obtained for SA and BT to determine their spatial structures as a function of potential. The structure of the SA adlayer changed from (2 x mean square root of 3)rect to domain walls to (mean square root of 7 x mean square root of 7)R19.1 degrees and then to disordered as the potential was scanned from 0.3 to 0.6 V. In contrast, molecules of BT were arranged in (2 x mean square root of 3)rect between 0.1 and 0.4 V, while they were disordered at all other potentials. Adsorption of BT molecules was predominantly through the sulfur headgroup. Sulfur adatoms and adsorbed BT molecules were stable against anodic polarization up to 1.0 V (vs RHE). These two species were adsorbed so strongly that their desorption did not occur even at the onset potential for the reduction of water in 0.1 M KOH.     

Enhanced selectivity towards O(2) and H(2) dissociation on ultrathin Cu films on Ru(0001)

The reactivity of Cu monolayer (ML) and bilayer films grown on Ru(0001) towards O(2) and H(2) has been investigated. O(2) initial sticking coefficients were determined using the King and Wells method in the incident energy range 40-450 meV, and compared to the corresponding values measured on clean Ru(0001) and Cu(111) surfaces. A relative large O(2) sticking coefficient (～0.5-0.8) was measured for 1 ML Cu and even 2 ML Cu/Ru(0001). At low incident energies, this is one order of magnitude larger than the value observed on Cu(111). In contrast, the corresponding reactivity to H(2) was near zero on both Cu monolayer and bilayer films, for incident energies up to 175 meV. Water adsorption on 2 ML Cu/Ru(0001) was found to behave quite differently than on the Ru(0001) and Cu(111) surfaces. Our study shows that Cu/Ru(0001) is a highly selective system, which presents a quite different chemical reactivity towards different species in the same range of collision energies.     

Hyperthermal Ar atom scattering from a C(0001) surface

Experiments and simulations on the scattering of hyperthermal Ar from a C(0001) surface have been conducted. Measurements of the energy and angular distributions of the scattered Ar flux were made over a range of incident angles, incident energies (2.8-14.1 eV), and surface temperatures (150-700 K). In all cases, the scattering is concentrated in a narrow superspecular peak, with significant energy exchange with the surface. The simulations closely reproduce the experimental observations. Unlike recent experiments on hyperthermal Xe scattering from graphite [Watanabe et al., Eur. Phys. J. D 38, 103 (2006)], the angular dependence of the energy loss is not approximated by the hard cubes model. The simulations are used to investigate why parallel momentum conservation describes Xe scattering, but not Ar scattering, from the surface of graphite. These studies extend our knowledge of gas-surface collisional energy transfer in the hyperthermal regime, and also demonstrate the importance of performing realistic numerical simulations for modeling such encounters.     

Isotope effect in Ni + C2H4 (C2D4) association reaction

The association reactions of atomic nickel with ethene and fully deuterated ethene in carbon dioxide buffer gas at 295 K have been investigated in the pressure range 5–100 torr, using a laser photolysis-laser fluorescence technique. By comparison with results of ab initio quantum chemistry calculations for the complex Ni[C₂H₄], the data are shown to be consistent with reaction on both ground and excited state potential energy surfaces. A simple rate equations treatment is described which shows the form of the pressure dependence of the second-order recombination rate coefficient in this case. Under conditions which are expected to hold for the Ni + C₂H₄ (C₂D₄) reaction, the pressure dependence has the standard Lindemann-Hinshelwood form, with the limiting high pressure rate constant given by an apparent value which reflects the degree to which the participating electronic states are coupled by nonadiabatic transitions. The limiting high pressure behavior of the recombination rate coefficient for Ni + C₂H₄ is not strongly affected by deuterium isotope substitution. However, the effect on the low pressure rate constant is large and consistent with RRKM unimolecular reaction theory. This validates the use of RRKM calculations for estimating the binding energy of the complex from kinetic data. The binding energy of Ni[C₂H₄] is estimated to be 35.2 ± 4.2 kcal mol^?1. © 1994 John Wiley & Sons, Inc.     

Six-dimensional quantum dynamics of dissociative chemisorption of H2 on Ru(0001)

Six-dimensional quantum dynamics calculations on dissociative chemisorption of H(2) on Ru(0001) are performed. The six-dimensional potential energy surface is generated using density functional theory. Two different generalized gradient approximations are used, i.e., RPBE and PW91, to allow the results to be compared. The dissociation probability for normally incident H(2) on a clean Ru(0001) surface is calculated. Large differences between the reaction probabilities calculated using the RPBE and PW91 are seen, with the PW91 results showing a much narrower reaction probability curve and a much higher reactivity. Using the reaction probabilities and assuming normal energy scaling reaction rates are generated for temperatures between 300 and 800 K. The rate generated using the PW91 results is higher by about a factor 5 than the rate based on the RPBE results in the range of temperatures relevant to ammonia production.     

Negative ion formation in electron-stimulated desorption of CF2Cl2 coadsorbed with polar NH3 on Ru(0001)

Photon-induced dissociation of CF2Cl2 (freon-12) in the stratosphere contributes substantially to atmospheric ozone depletion. We report recent results on dissociation and negative ion formation in electron-stimulated desorption (ESD) of CF2Cl2 on Ru(0001), when CF2Cl2 is coadsorbed with a polar molecule (NH3), for electron energies ranging from 50 to 300 eV. Two different time-of-flight methods are used in this investigation: (a) an ESD ion angular distribution detector with wide collection angle and (b) a quadrupole mass spectrometer with narrow collection angle and high mass resolution. Many negative ESD fragments are seen (F-, Cl-, FCl-, CF-, F2-, and Cl2-), whose intensities depend on the surface preparation. Using both detectors we observe a giant enhancement of Cl- and F- yields for ESD of CF2Cl2 coadsorbed with approximately 1 ML of NH3; this enhancement (>10(3) for Cl-) is specific to certain ions, and is attributed to an increased probability of dissociative electron attachment due to "trapped" low-energy secondary electrons, i.e., precursor states of the solvated electron in NH3. In further studies, the influence of polar NH3 spacer layers (1-10 ML) on ESD of top-layer CF2Cl2 is determined, and compared with thick films of condensed CF2Cl2. The magnitudes and energy dependences of the Cl- yields are different in these cases, due to several contributing factors.