仿生資料庫

Database of Surface Plasmon Resonance Characteristics from Urchin-Like Au NPs Heterostructure
(海膽狀金奈米粒子之表面電漿共振特性資料庫)

Introduction

The effect of surface plasmon resonance (SPR) from noble metal nanostructures such as gold nanoparticles (Au NPs) has been proposed to promote the generation of energetic hot electrons as well as boosting resonant energy transfer (RET), thereby resulting in significantly enhancing of the solar-light harvesting and energy-conversion efficiency. Herein Urchin-like Au NPs with ansymmetric structures have been synthesized through seed-mediate method. In this database, position of absorption peak and FWHM of Urchin-like Au NPs are evaluated with experimental parameters including time of Gold colloid (50-300 ul) , concentration of HAuCl₄ (10^-4-5*10^-4 M) and Hydroquinone (1000-3000 ul). The database is established by data mining and machine learning through the strategy of combining artificial neural networks with genetic algorithm. The learning process is up to 1,000,000 times, the root mean square (RMS) error is less than 0.012, and the false rate is less than 0.001.

貴金屬奈米結構如金奈米顆粒（Au NPs）的表面電漿共振（SPR）的作用已被提出，透過改變奈米顆粒之型貌，結構及尺寸，達到選擇性吸收特定之可見光波長。本文透過晶種成核法合成了具有表面電漿共振之海膽狀金奈米粒子(Urchin-like Au NPs)。在資料庫中，使用包括晶種添加量 (50-300 ul) 和HAuCl₄的濃度（10^-4-5*10^-4 M）和Hydroquinone添加量 (1000-3000 ul) 的實驗參數評估海膽狀金奈米粒子的吸收峰位置、FWHM。資料庫中的數據是以資料探勘和機器學習方式，經結合基因演算與人工神經網絡的深度學習所建立。學習程序為1,000,000次，均方根（RMS）誤差小於0.012，誤率則小於0.001。