運用機器學習於抗反射光學薄膜之優化與應用

Abstract

綜觀顯示技術的發展，微型發光二極體(Micro Light Emitting Diode Display, Micro LED)顯示器已是繼液晶顯示器(Liquid-Crystal Display, LCD)、有機發光二極體(Organic Light-Emitting Diode, OLED)顯示器之後另一個新興且蓬勃發展的顯示技術，憑藉著其高對比度、長壽命、低功耗等優點更被視為重要次世代的顯示技術。得益於微型發光二極體極微小的晶粒尺寸，透明顯示器已成為了其重要的應用場景。然而，目前微型發光二極體透明顯示器仍有一些工程問題尚待解決，如：製程良率、發光效率以及顯示器背側漏光影響使用者視覺體驗等，影響應用推廣。本研究將針對透明顯示器背側的漏光探討利用抗反射光學薄膜改善的可能性。 首先，第一章將簡介薄膜光學、抗反射光學薄膜及造成透明顯示器背側漏光的成因及其現有解決方案，並概述後續用於設計、優化抗反射光學薄膜的機器學習演算法及人工神經網絡。第二章則聚焦於如何建構一光學薄膜優化系統，當中包含了以 Maxwell’s Equations 以及 Fresnel Equation 作為基礎的轉移矩陣及人工神經網絡。第三章則會利用第二章所建構的光學薄膜優化系統，依據微型發光二極體透明顯示器的應用場景擬定對應的設計方案，並進行實作及驗證。最後，經由實驗驗證，本研究所設計的抗反射光學薄膜根據視角的不同至多有效減少約 50%顯示器背側漏光的產生。

With the continuous evolution of display technologies, the Micro Light-Emitting Diode (Micro LED) display has emerged as a promising next-generation technology following Liquid Crystal Displays (LCDs) and Organic Light-Emitting Diode (OLED) displays. Owing to its high contrast ratio, long operational lifetime, and low power consumption, Micro LED has been widely regarded as an important display technology. Benefiting from the ultra-small chip size of Micro LEDs, transparent displays have become one of their key application areas. However, Micro LED transparent displays still face some engineering challenges, including process yield, luminous efficiency, and the light leakage on the backside of the display, which can negatively impact the viewer's visual experience. These issues are among the main reasons why this technology has yet to reach widespread commercialization. This study aims to explore the potential of using anti-reflection optical thin films to mitigate light leakage issue on the backside of transparent displays. Chapter 1 introduces the fundamentals of thin-film optics and anti-reflection optical films along with the origin of the light leakage observed on the backside of transparent displays and the existing solutions. It also provides an overview of the machine learning algorithms and artificial neural networks applied in the subsequent design and optimization processes. Chapter 2 focuses on the construction of an optical thin-film optimization system based on Maxwell’s equations and Fresnel equations, employing the transfer matrix method and neural network models. In Chapter 3, the proposed system is applied to develop an optimized anti-reflection thin-film design tailored to Micro LED transparent display applications, followed by fabrication and experimental validation. Experimental results demonstrate that the designed anti-reflective thin film can reduce the visibility of the backside light leakage by up to approximately 50%, depending on the viewing angle.