CMOS玻璃蓋片自動光學檢測機台之設計及開發

Abstract

玻璃蓋片為CMOS影像晶片封裝時，裝置於CMOS上方之玻璃材質，有保護CMOS之功用，由於光線必須穿透這層玻璃才能到達CMOS晶片，故這片玻璃之透明度與均質度必須非常高，才不致影響成像效果。玻璃蓋片主要會發生之瑕疵包括，刮傷、裂痕、氣泡、可移動之灰塵、不可移動之灰塵、油酯、有機物污染與水紋…等，這些瑕疵會影響玻璃之均質度與透光度，導致封裝後的CMOS成像出現缺陷。

目前業界多是以人工目視對玻璃蓋片進行抽檢，受限於人眼的最小可辨識能力約20μm，較難配合5μm之檢測精度需求。且人工目視容易失誤且耗時過久，勢必無法滿足市場顧客對於品質一致性的要求。本研究設計出適當的光源與取像機構，配合所開發之AOI(Automatic optical inspection)軟體演算法，架構出完整的自動視覺檢測系統。

本研究的目的，在發展出一套適用於CMOS玻璃蓋片瑕疵自動視覺檢測的機台，使能檢測出玻璃蓋片所可能會發生的瑕疵，以取代目前之人工目視檢測作業，改善人工檢測所不足的檢測精度，並提昇檢測速度與減少誤判率，達到100%全檢之理想目標。

The optical glass lid is the covering material of CMOS chip in the packaging process. The function of optical glass lid is to protect the CMOS chip. The lid must be highly penetrable and uniform, and will not affect the imaging result of CMOS. The major defects of optical glass lid include scratches, cracks, bubbles, removable particles, non-removable particles, greases, organic contaminations, and watermarks. These defects reduce the penetrability and uniformity of optical glass lid.

Up to now, most CMOS package factories use human operators to inspect the optical glass lid by sample. But human inspection is too unstable and slow to satisfy the costumers. In this paper, we will devise a light source system, an image collecting mechanism, and a set of AOI software algorithms to construct an automatic optical system for optical glass lid inspection.

The purpose of this paper is to develop a novel AOI system for optical glass lid. This system can find out possible defects of optical glass lid and can improve the accuracy and speed of inspection. In CMOS packaging factories, this system can replace human inspection and work synchronously with the CMOS packaging line.