We quantitatively evaluate the contribution of electron-hole pair excitations to the reactive dynamics of H 2 on Cu(110) and N 2 on W(110), including the six dimensionality of the process in the entire calculation. The interaction energy between molecule and surface is represented by an ab initio six-dimensional potential energy surface. Electron friction coefficients are calculated with density functional theory in a local density approximation. Contrary to previous claims, only minor differences between the adiabatic and nonadiabatic results for dissociative adsorption are found. Our calculations demonstrate the validity of the adiabatic approximation to analyze adsorption dynamics in these two representative systems. DOI: 10.1103/PhysRevLett.100.116102 PACS numbers: 82.65.+r, 34.35.+a, 68.49.Df, 82.20.Kh The adiabatic Born-Oppenheimer approximation is ubiquitous in the theoretical study of elementary reactive processes at surfaces. Still, there is ample experimental evidence of electronic excitations associated to gas or surface reactions, that can potentially break down the applicability of the adiabatic approach. Electron-hole pairs appear, for instance, in the detection of chemicurrents during the chemisorption of gas-phase species on thin metal films [1,2], as well as in the measurement of electron emission following the scattering of molecules in highlyexcited vibrational states on metal surfaces [3,4]. Although the existence of energy dissipation through electron-hole (e-h) pair excitations is widely accepted, there is not a definite quantitative answer on the role of electronic excitations in the adsorption and reaction rates of diatomic molecules at metal surfaces [5].Laursen et al. [6] predicted that e-h pair excitations should alter substantially the adsorption dynamics of H 2 in Cu (110). Contravening this prediction, full sixdimensional (6D) adiabatic calculations of the dynamics based on an ab initio potential energy surface have been shown to provide a good description of the experimental results on this system [7]. Measurements of the dissociative adsorption and diffractive scattering of H 2 on Pt(111) are reasonably well described within the adiabatic approximation as well [8]. For heavier molecules, it was claimed that strong energy dissipation effects due to the excitation of e-h pairs are responsible for the strong disagreement between (adiabatic) theoretical and experimental sticking coefficients for N 2 on Ru(0001) [9,10]. Díaz et al. [11] showed afterwards that when the full dimensionality of the process is taken into account, adiabatic calculations are much closer to experiments. This result suggests that nonadiabatic effects might be smaller than previously predicted. Still, the lack of a theoretical calculation, based on state-of-the-art interaction potentials, that explicitly includes the e-h pair excitation channel and the full dimensionality of the process keeps this controversy open.In this Letter, we evaluate the contribution of e-h pair excitations to the dissociative adsorp...

An efficient method is proposed to construct the six-dimensional Potential Energy Surface (PES) for diatomic molecule-surface interactions from low dimensional cuts obtained in ab initio calculations. The efficiency of our method results from a corrugation-reducing procedure based on the observation that most of the corrugation in a molecule-surface PES is already embedded in the atom-surface interactions. Hence, substraction of the latter leads to a much smoother function which makes accurate interpolations possible. The proposed method is a general one and can be implemented in a systematic way for any system. Its efficiency is illustrated for the case of H2/Pd(111) by using recent ab initio data. We report also the results of very stringent checks against ab initio calculations not used in the interpolation. These checks show the high accuracy of our method.

Juaristi et al. Reply: In the preceding Comment [1] on our recent Letter [2], Luntz et al. claim that their results show that the local density approximation for electronic friction (LDAF) is in general not reliable for estimating the electronic friction for molecules on metal surfaces. This is an unfounded statement based on very limited information.Our LDAF calculation for the friction coefficient involves at most two further approximations as compared to those involved in the model used by Luntz et al.. (i) First, a local approximation for the electronic density is used. This is acknowledged to be a reasonable approximation. The LDAF was shown to be appropriate for calculations of the friction coefficient for atoms in metal surfaces, incidentally by some of the authors of the comment [3]. (ii) An additional approximation is that two-center effects are disregarded. Two-center effects were shown not to change the order of magnitude in the case of H 2 [4], but no studies in other systems have been performed to our knowledge.

Following earlier three-dimensional ͑3D͒ calculations, we present results of four-dimensional ͑4D͒ calculations on dissociative and diffractive scattering of H 2 from Pt͑111͒ by extending the 3D model with a second degree of freedom parallel to the surface. A 4D potential energy surface ͑PES͒ is constructed by interpolating four 2D PESs obtained from density-functional theory calculations using the generalized gradient approximation and a slab representation of the metal surface. The 4D calculations show that out-of-plane diffraction is much more efficient than in-plane diffraction, providing a partial explanation for the paradox that diffraction experiments measure little in-plane diffraction, whereas experiments on reaction suggest the surface to be corrugated. Calculations for off-normal incidence of vϭ0 H 2 show that, in agreement with experiment, initial parallel momentum inhibits dissociation at low normal translational energies, and enhances reaction for higher energies. Our 4D calculations also show that the reaction of initial vϭ1 H 2 is vibrationally enhanced with respect to vϭ0 H 2 , as was found in the 3D model, even though H 2 ϩPt(111) is an early barrier system.