仿生資料庫

Database hydrogen catalytic property of metal nanoparticles-adsorbed oxides(金屬奈米粒子吸附氧化物產氫特性資料庫)

Introduction

From the first-principle structural optimization, it shows that the adsorption energy of hydrogen atoms on the surface of metal nanoparticle-adsorbed oxides will determine its catalytic activity and hydrogen evolution reaction efficiency. The adsorption energy will be affected by (1) the type of nanoparticles (2) the type of oxide structures (3) the adsorption position. Therefore, the database selects these three important factors that affect the adsorption energy of hydrogen atoms as the setting parameters, and combines the neural network algorithm and the genetic algorithm to provide the adsorption energy of hydrogen under different parameters. The learning process is up to 1,000,000 times, the root mean square (RMS) error is less than 0.0057, and the false rate is less than 0.01.

由第一原理之結構優化計算可以得知，氫原子於金屬奈米粒子吸附氧化物表面之吸附能會決定其催化活性及產氫效率，而吸附能會受到(1)奈米粒子種類(2)氧化物種類(3)吸附位置的影響，因此本資料庫挑選這三種影響氫原子吸附能的重要參數作為設定條件，結合類神經網路演算法和基因演算法，提供在不同參數條件下金屬奈米粒子吸附氧化物之吸附能。學習程序為1,000,000次，均方根（RMS）誤差小於0.0057，誤率則小於0.01。