Study On The Scaling Relationship Between Adsorption Energies Of Atoms And Small Molecules On The Surfaces Of Metals And Alloys

Catalysts are the touchstone of chemical production,since they can improve the rate of chemical reactions,activate the inert reaction materials,lower the production cost,increase the output of target product and decrease the output of by-product.Therefore,to develop highly-efficient catalysts has a very strong practical application value,and is of great significance to the sustainable development.The traditional catalyst research method is mostly trial and error,which not only needs high research cost,but also has low efficiency.With the advances of computer technology and the continuous improvement of computational chemistry-related theory,theoretical calculation method is playing an increasingly important role in the field of searching and selecting target catalysts.At present,the scheme proposed by N?rskov and Bligaard is mostly used in the screening of catalytic materials.In this scheme,the adsorption energy is often used as the descriptor to characterize performance of a catalyst.Currently,the estimation and prediction of adsorption energy usually uses the parameters calculated by computationally expensive electronic structure methods,which will undoubtedly increase the research cost.This paper takes the adsorption energy as the research object,aiming at building the scaling relationship with non-quantum chemical calculation parameters,and providing key information for the design and screening of new catalyst materials.The main research contents of this paper are as follows:Firstly,the adsorption energies of 12 atoms on 38 metal surfaces were calculated by density functional theory and the properties of metal substrates and adsorbed atoms were correlated with the adsorption energies.The results show that when adsorption is dominated by ionic bonding,there is a good linear relationship between the adsorption energy and the electronegativity difference between the adsorbate and the substrate.With the help of machine learning methods,universal adsorption energy prediction models were established with the minimum average absolute error between the predicted and DFT calculated values being 0.17 e V.Given the increasing importance of alloys in catalysis,we calculated the adsorption energies of 12 atoms on the surfaces of 12 binary alloys.It is found that the adsorption energy on the surfaces of binary alloys depend mainly on the interaction between the adsorbed atoms and the active component metals in the alloys.And the monatomic adsorption energy on the surfaces of binary alloys was successfully predicted with the machine learning models built on the basis of the adsorption data of atoms on pure metals,and the average absolute error between the predicted values and the DFT calculated values was less than 0.27 e V.Chapter 5 deals with the adsorption energy of 18 species involved in carbon dioxide hydrogenation.The adsorption energies of 18 species on 7 kinds of pure metal surfaces and 12 kinds of binary alloy surfaces were calculated by density functional theory.The results show that the adsorption energy of 18 species on the surfaces of pure metals can be expressed only by the adsorption energies of C and OH.For the adsorption energy of 18 species on the surfaces of binary alloys,at least 4 parameters（adsorption energy of C,O,H₂O,CO₂）are needed in order to better describe the adsorption energies to map the whole reaction network.Further analysis shows that the adsorption energy on binary alloy surfaces can be effectively represented by the adsorption energies on the corresponding pure metal surfaces.