Representing high-dimensional potential-energy surfaces for reactions at surfaces by neural networks
نویسندگان
چکیده
The determination of dissociative adsorption probabilities based on first-principles total-energy calculations requires a numerically efficient and accurate interpolation scheme in order to be able to run a sufficient number of trajectories. Here we present a neural network scheme for the construction of a continuous potential energy surface (PES). We illustrate the accuracy and efficiency of our method for H2 interacting with the (2 · 2) potassium covered Pd(100) surface. The sticking probability of H2/K(2 · 2)/ Pd(100) is determined by molecular dynamics simulations on the neural network PES and compared to results using an independent analytical interpolation. 2004 Elsevier B.V. All rights reserved.
منابع مشابه
Estimating and modeling monthly mean daily global solar radiation on horizontal surfaces using artificial neural networks
In this study, an artificial neural network based model for prediction of solar energy potential in Kerman province in Iran has been developed. Meteorological data of 12 cities for period of 17 years (1997–2013) and solar radiation for five cities around and inside Kerman province from the Iranian Meteorological Office data center were used for the training and testing the network. Meteorologic...
متن کاملComputational study of three dimensional potential energy surfaces in intermolecular hydrogen bonding of cis-urocanic acid
متن کامل
Ready, Set and no Action: A Static Perspective on Potential Energy Surfaces commonly used in Gas-Surface Dynamics
In honoring the seminal contribution of Henry Eyring and Michael Polanyi who first introduced the concept of potential energy surfaces (PESs) to describe chemical reactions in gas-phase [Z. Phys. Chem. 12, 279–311, (1931)], this work comes to review and assess state-of-the-art approaches towards first-principle based modeling in the field of gas-surface dynamics. Within the BornOppenheimer and ...
متن کاملDescriptions of surface chemical reactions using a neural network representation of the potential-energy surface
A neural network NN approach is proposed for the representation of six-dimensional ab initio potentialenergy surfaces PES for the dissociation of a diatomic molecule at surfaces. We report tests of NN representations that are fitted to six-dimensional analytical PESs for H2 dissociation on the clean and the sulfur covered Pd 100 surfaces. For the present study we use high-dimensional analytical...
متن کاملCalculation for Energy of (111) Surfaces of Palladium in Tight Binding Model
In this work calculation of energetics of transition metal surfaces is presented. The tight-binding model is employed in order to calculate the energetics. The tight-binding basis set is limited to d orbitals which are valid for elements at the end of transition metals series. In our analysis we concentrated on electronic effects at temperature T=0 K, this means that no entropic term will be pr...
متن کامل