Theoretical Research On Hydrogen Energy And Lithium Metal Batteries

In recent years,the problems of the energy crisis and environmental pollution have become increasingly prominent.To deal with this problem,reducing the utilization of fossil energy,actively developing new energy resources and establishing a reasonable green energy structural system are very important strategies.Among them,the development of high-efficiency electrocatalysts for hydrogen energy and the development of high-energy-density batteries(such as lithium metal batteries)are considered to be an important part of establishing a sustainable green energy system,so they have received extensive attention and triggered a research upsurge.However,the understanding of the electrochemical catalytic mechanism of catalysts is still obviously insufficient,and the structure-activity relationship in the electrochemical system is still not clear enough,and there is no reliable theory to guide the rational material design and synthesis of experiments.In order to solve the above problems,this paper mainly carried out two research works:(1)Based on theoretical calculations,the electrocatalytic reaction mechanism for the efficient production and utilization of hydrogen energy was clarified;(2)Based on multi-scale simulations,the accurate prediction of Ion conduction behavior at the electrolyte and solid electrolyte interface(Solid Electrolyte Interphase,SEI)for lithium metal batteries.The specific research results are briefly described as follows:(1)Mechanism of compressive stress on the hydrogen evolution reaction(HER)of pure metal Ir,and predicted a new strategy to increase the catalytic performance of metal Ir by increasing the compressive stress.(2)Mechanism of the effect of metal Rh surface compressive stress on the performance of Hydrazine Oxidation Reaction(HzOR).The calculation results show that appropriately increasing the compressive stress can reduce the reaction energy of the potential determining step of HzOR,and also reduce the risk of catalyst poisoning.(3)The effect of metal oxide IrO2 solid phase structure on the catalytic performance of oxygen evolution reaction(OER).The calculation results show that different stacking methods and layer spacing will cause IrO2 to have a huge impact on the OER reaction mechanism and pathway,which reasonably explains the significant difference in activity observed in the experiment.(4)The effect of bulk phase doping and surface engineering on the activity of alloyed RuNi catalysts for hydrogen oxidation reaction(HOR).Through theoretical calculations,the adsorption energy of H,OH,H2O and the formation energy of H2O on different engineering RuNi alloys are predicted.These calculations prove that the dual-engineering catalyst has obvious advantages in catalytic activity compared with the single-engineering catalyst,which not only explains the experimental phenomena,but also provides a theoretical basis for the development of new materials.(5)Using the self-developed XPS prediction technology to characterize the solid electrolyte interface.This method can predict the XPS peak position and intensity at the same time,so as to achieve direct comparison with the experimental XPS,which not only can be used to verify the accuracy of the simulation,but also can realize the reverse analysis of the experimental structures.(6)In order to further improve the XPS prediction efficiency,a machine learning-based XPS for rapid prediction method was developed,the calculation results show that the machine learning model can predict the error of binding energy within 0.05 eV,the predicted XPS spectrum is almost completely consistent with the simulation results and experimental results.More importantly,the machine learning method has increased the prediction speed by hundreds of times.(7)Using the machine learning method to quickly screen high-diffusion coefficient under low-temperature electrolytes,and accurately predict the anion and cation diffusion coefficients in 1441 electrolyte systems through molecular dynamics simulation.Using structural features and property features as descriptors,the tame machine learning model can accurately predict the ion diffusion coefficient of the electrolyte.The machine learning model can accurately predict the ion diffusion coefficient,and the root mean square error is within 0.3825 e-10 m2/s.Based on this model,through interpretable machine learning,it is concluded that electrolytes with low density and weak van der Waals interactions usually have higher diffusion coefficients.