Chebyshev high-dimensional model representation (Chebyshev-HDMR) potentials: application to reactive scattering of H2 from Pt(111) and Cu(111) surfaces

Thomas, Phillip S.; Somers, Mark F.; Hoekstra, Anne W.; Kroes, Geert–Jan

doi:10.1039/c2cp40173h

Cited by 17 publications

(12 citation statements)

References 98 publications

(121 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…11,13,48,78,116,129,133,[141][142][143]145,161,163, The RPBE DF, which was designed to correct for overbinding of atoms and molecules on metal surfaces, 21 has also found much use in dynamics studies of H 2 -metal surface scattering. 53,114,[190][191][192][193][194][195][196][197][198][199] An indication of the level of accuracy that may be obtained with these functionals can be obtained by inspecting Table 1, where we have collected averages of mean unsigned errors (MUEs) in reaction barrier heights computed for two databases of gas phase chemical reactions (HTBH38/08 and NHTBH38/08) by Peverati and Truhlar 167 for a few selected functionals. 53,114,[190][191][192][193][194][195][196][197][198][199] An indication of the level of accuracy that may be obtained with these functionals can be obtained by inspecting Table 1, where we have collected averages of mean unsigned errors (MUEs) in reaction barrier heights computed for two databases of gas phase chemical reactions (HTBH38/08 and NHTBH38/08) by Peverati and Truhlar 167 for a few selected functionals.…”

Section: A Density Functional Theorymentioning

confidence: 99%

Quantum and classical dynamics of reactive scattering of H₂ from metal surfaces

2016

Self Cite

View full text Add to dashboard Cite

We review the state-of-the art in dynamics calculations on the reactive scattering of H2 from metal surfaces, which is an important model system of an elementary reaction that is relevant to heterogeneous catalysis. In many applications, quantum dynamics and classical trajectory calculations are performed within the Born-Oppenheimer static surface model. However, ab initio molecular dynamics (AIMD) is finding increased use in applications aimed at modeling the effect of surface phonons on the dynamics. Molecular dynamics with electronic friction has been used to model the effect of electron-hole pair excitation. Most applications are still based on potential energy surfaces (PESs) or forces computed with density functional theory (DFT), using a density functional within the generalized gradient approximation to the exchange-correlation energy. A new development is the use of a semi-empirical version of DFT (the specific reaction parameter (SRP) approach to DFT). We also discuss the accurate methods that have become available to represent electronic structure data for the molecule-surface interaction in global PESs. It has now become possible to describe highly activated H2 + metal surface reactions with chemical accuracy using the SRP-DFT approach, as has been shown for H2 + Cu(111) and Cu(100). However, chemical accuracy with SRP-DFT has yet to be demonstrated for weakly activated systems like H2 + Ru(0001) and non-activated systems like H2 + Pd(111), for which SRP DFs are not yet available. There is now considerable evidence that electron-hole pair (ehp) excitation does not need to be modeled to achieve the (chemically) accurate calculation of dissociative chemisorption and scattering probabilities. Dynamics calculations show that phonons can be safely neglected in the chemically accurate calculation of sticking probabilities on cold metal surfaces for activated systems, and in the calculation of a number of other observables. However, there is now sufficient evidence to suggest that the decision on whether or not to neglect phonons should be taken with care, with appropriate consideration of the observable to be computed and of the relevant surface temperature. AIMD calculations have provided valuable insights into the mechanisms that are operative in the dissociative adsorption and absorption of hydrogen on/in precovered metal surfaces. Classical and quantum dynamics calculations have shown that the reaction probability of H2 on Pt surfaces consisting of (100) steps and (111) terraces can to a very good approximation be computed as a weighted average of the reactivities on the steps and terraces. Progress obtained with dynamics calculations on the scattering of H2 from alloys and from simple low index metal surfaces is also reported. Insights that may be obtained on the reactivity of a metal surface from the prominent presence of out-of-plane diffraction or, conversely, the complete absence of diffraction, are discussed. A new field has been opened up by experiments on H2 scattering from surfaces at fast graz...

show abstract

Section: A Density Functional Theorymentioning

confidence: 99%

Quantum and classical dynamics of reactive scattering of H₂ from metal surfaces

2016

Self Cite

View full text Add to dashboard Cite

show abstract

“…The calculation of the integrals has been subject of many works aiming to make RS-HDMRs feasible. [79,80,82,87,97] Lower order (<k) terms are also unnecessary but may still be used for the same reasons as in the nested NN scheme. Without them, the fit becomes simply…”

Section: Multibody and Multimode-like Representations By The Choice Omentioning

confidence: 99%

“…[92][93][94] When truncated at k < d, the representation is approximate, and it has been shown that vibrational levels computed with k 4 have errors of a few cm 21 . [95][96][97][98][99][100][101] The terms in each of the sums are called component functions. Instead of using an NN to directly fit a multidimensional PES, it is possible to use NNs to fit the component functions.…”

Section: Multibody and Multimode-like Representations By The Choice Omentioning

confidence: 99%

See 1 more Smart Citation

Neural network‐based approaches for building high dimensional and quantum dynamics‐friendly potential energy surfaces

Manzhos

Dawes

Carrington

2014

Int J of Quantum Chemistry

192

148

View full text Add to dashboard Cite

Development and applications of neural network (NN)-based approaches for representing potential energy surfaces (PES) of bound and reactive molecular systems are reviewed. Specifically, it is shown that when the density of ab initio points is low, NNs-based potentials with multibody or multimode structure are advantageous for representing high-dimensional PESs. Importantly, with an appropriate choice of the neuron activation function, PESs in the sum-of-products form are naturally obtained, thus addressing a bottleneck problem in quantum dynamics. The use of NN committees is also analyzed and it is shown that while they are able to reduce the fitting error, the reduction is limited by the nonrandom nature of the fitting error. The approaches described here are expected to be directly applicable in other areas of science and engineering where a functional form needs to be constructed in an unbiased way from sparse data.

show abstract

“…For example, the energies calculated at particular geometries can be fit to an analytic functional form. 28,[31][32][33][34][35][36][37] The PES can be expressed in a neural network 38,39 or cluster-type [40][41][42] representation. Alternatively, there are several approaches based on local expansions around geometries where the energy and other properties have been calculated explicitly.…”

Section: Introductionmentioning

confidence: 99%

Modified Shepard interpolation of gas-surface potential energy surfaces with strict plane group symmetry and translational periodicity

Frankcombe

Collins

Zhang

2012

The Journal of Chemical Physics

View full text Add to dashboard Cite

A new formulation of modified Shepard interpolation of potential energy surface data for gas-surface reactions has been developed. The approach has been formulated for monoatomic or polyatomic adsorbates interacting with crystalline solid surfaces of any plane group symmetry. The interpolation obeys the two dimensional translational periodicity and plane group symmetry of the solid surface by construction. The interpolation remains continuous and smooth everywhere. The interpolation developed here is suitable for constructing potential energy surfaces by sampling classical trajectories using the Grow procedure. A model function has been used to demonstrate the method, showing the convergence of the classical gas-surface reaction probability.

show abstract

Chebyshev high-dimensional model representation (Chebyshev-HDMR) potentials: application to reactive scattering of H2 from Pt(111) and Cu(111) surfaces

Cited by 17 publications

References 98 publications

Quantum and classical dynamics of reactive scattering of H₂ from metal surfaces

Quantum and classical dynamics of reactive scattering of H₂ from metal surfaces

Neural network‐based approaches for building high dimensional and quantum dynamics‐friendly potential energy surfaces

Modified Shepard interpolation of gas-surface potential energy surfaces with strict plane group symmetry and translational periodicity

Contact Info

Product

Resources

About

Chebyshev high-dimensional model representation (Chebyshev-HDMR) potentials: application to reactive scattering of H2 from Pt(111) and Cu(111) surfaces

Cited by 17 publications

References 98 publications

Quantum and classical dynamics of reactive scattering of H2 from metal surfaces

Quantum and classical dynamics of reactive scattering of H2 from metal surfaces

Neural network‐based approaches for building high dimensional and quantum dynamics‐friendly potential energy surfaces

Modified Shepard interpolation of gas-surface potential energy surfaces with strict plane group symmetry and translational periodicity

Contact Info

Product

Resources

About

Quantum and classical dynamics of reactive scattering of H₂ from metal surfaces

Quantum and classical dynamics of reactive scattering of H₂ from metal surfaces