The most common mechanism of catalytic surface chemistry is that of Langmuir and Hinshelwood (LH). In the LH mechanism, reactants adsorb, become thermalized with the surface, and subsequently react. The measured vibrational (relaxation) lifetimes of molecules adsorbed at metal surfaces are in the range of a few picoseconds. As a consequence, vibrational promotion of LH chemistry is rarely observed, with the exception of LH reactions occurring via a molecular physisorbed intermediate. Here, we directly detect adsorption and subsequent desorption of vibrationally excited CO molecules from a Au(111) surface. Our results show that CO (v = 1) survives on a Au(111) surface for ~1 × 10 s. Such long vibrational lifetimes for adsorbates on metal surfaces are unexpected and pose an interesting challenge to the current understanding of vibrational energy dissipation on metal surfaces. They also suggest that vibrational promotion of surface chemistry might be more common than is generally believed.

We investigated the translational incidence energy (Ei) and surface temperature (Ts) dependence of CO vibrational excitation upon scattering from a clean Au(111) surface. We report absolute v = 0 → 1 excitation probabilities for Ei between 0.16 and 0.84 eV and Ts between 473 and 973 K. This is now only the second collision system where such comprehensive measurements are available – the first is NO on Au(111). For CO on Au(111), vibrational excitation occurs via direct inelastic scattering through electron hole pair mediated energy transfer – it is enhanced by incidence translation and the electronically non-adiabatic coupling is about 5 times weaker than in NO scattering from Au(111). Vibrational excitation via the trapping desorption channel dominates at Ei = 0.16 eV and quickly disappears at higher Ei.

Molecular beam surface scattering is used to compare vibrational excitation of N 2 molecules in collisions with clean Pt(111) and Au(111) surfaces under UHV conditions. Direct single-bounce collisions are dominant under all conditions of this work, as evidenced by narrow specular angular scattering distributions and translational incidence energy dependent rotational temperatures. N 2 (v = 0 → 1) vibrational excitation is observed for Pt(111), but not Au(111). The excitation probabilities, ranging from ∼10 −4 to ∼10 −3 , follow an Arrhenius surface temperature dependence and increase with translational incidence energy with zero threshold. The observations are the typical fingerprint of nonadiabatic vibrational excitation due to an electron mediated excitation mechanism, identified in previous work on NO and CO scattering from metal surfaces. The unfavorable electron affinity of the N 2 molecule (EA = −2.3 eV) makes this observation surprising and we discuss possible excitation mechanisms. ■ INTRODUCTIONThe ability to predict whether or not a reaction will happen when a molecule collides with a surface is essential if catalysts are to be designed from first principles. Real catalytic systems are often extremely complex and results from simple model systems, which allow a detailed picture of the surface−molecule interaction and energy transfer to be built, are essential to test theoretical predictions and develop deeper understanding.Molecular beam scattering from well characterized singlecrystal surfaces has provided valuable information about the dynamic processes occurring during the collisions, demonstrating that while some are electronically adiabatic (mechanical translational to vibrational (T−V) excitation, for example, vibrational excitation of ammonia on Au(111) 1 ), other scattering events can only be adequately explained by electronically nonadiabatic mechanisms, where vibrational excitation occurs with the relaxation of a hot electron−hole pair (EHP) of the metal.Electronically nonadiabatic vibrational energy transfer exhibits dynamical fingerprints. Early work on vibrational excitation of NO directly scattered from Ag(111) 2,3 found a strong surface temperature dependence and the lack of a translational energy threshold for vibrational excitation, providing important evidence that the necessary energy comes from the surface, and not from the kinetic energy of the molecule. Energy transfer by a mechanical phonon-tovibration excitation mechanism is unlikely to be efficient given the mismatch in the energy of the high frequency molecular vibration and the low frequency phonons, leaving coupling to EHPs as the most likely energy source for vibrational (EHP-V) excitation. More recently, striking examples have shown that many quanta of vibrational energy can be transferred to or from the colliding molecule due to electronically nonadiabatic mechanisms. These effects can be seen in multiquantum vibrational relaxation 4 or excitation 5,6 and in the emission of electrons from low work function surfaces. 7 Meas...

No abstract