影像感測元件暗電流與晶圓製程及原材晶圓製程間關係之研究

Abstract

關於近來CMOS影像感測器（Image Sensor）的應用，已是到了無處不在的地步，從最初商務為主的傳真機、影印機、掃描器，到今日娛樂為主的Internet視訊電眼，倒車用的影像輔助系統等都是。而相較於CCD，CMOS擁有較低的驅動電壓，較完整的整合封裝特性，也因此在手持(Portable)功能的表現上較為優異，再加上成本低使得CMOS比CCD有著量產的優勢。但CMOS並非毫無缺點，如雜訊過高、亮度低時感應度不佳及偏高的漏電流等，針對這些缺點如果能在CMOS的製程中加以改善，那麼CMOS完全取代CCD便指日可待了。本論文以改善CMOS 影像感測器中的暗電流(Dark Current)為主，從Wafer製程的角度出發，配合FA(Failure Analysis) test vehicle的失效分析測試平台來找出漏電源(Leak source)及缺陷種類(Defects Type)。再根據這些分析，將相關的製程逐一層別出來，進而 執行製程改善與效果驗證。實驗中，發現光二極體(Photo Diode)及傳輸閘(Transfer Gate)旁的Floating Node(浮節點)的表面破壞(Surface Damage)與STI附近缺陷(Defects) 有著較高的暗電流(Dark Current)，這暗電流在CMOS Image Sensor表現出來的就是白像素(White Pixel)。而相關製程主要以電漿電荷所導致缺陷的製程(Plasma charging-induced defect process)，如RPO film、Spacer film etch、Ion implant等。而缺陷(Defects)的來源除來自MOS本身製程外，另一部分則來自原晶圓製程(Raw Wafer Process)的製程缺點(Process Weakness)，原晶圓製程的過程中會產生本質上的缺陷(Intrinsic defects)，如因過多氧而析出的BMD(Bulk Micro Defects)，這缺陷所產生的差排凹洞(Dislocation Pits)與白像素似乎有直接的關係。本論文則主要是專注於分析Ion Implant Process 對於影像感測元件產生Plasma charging-induced defect的成因與改善的方法以及對原晶圓製程因氧析出而產生缺陷，進而生成暗電流的原因之研究。

With the flourishing development of CMOS image sensor technology recently, the applications of this technology are broadly. At first, the application for the commercial affairs includes fax-machine, printer and scanner, as today for amusement affairs, like inter-net video eye and car image assist system. Comparing with the CCD, CMOS image sensor has lower driving voltage and better integral package character. For production superiority, CMOS image sensor performs excellent portable ability and low cost performance than CCD. But CMOS is not flawless. For example, the noise is higher, poor sensitivity as low illumination and high leak current. In the light of these shortcomings, maybe we can improve them in CMOS process flow by continue improvement actions. Then, it can be expected soon that CMOS image sensor will take the place of CCD completely. I take the “Dark current” research as the principle in my thesis. We look for the leak source and defect type from the point of view of wafer process and the result of failure analysis by proper test vehicle. Base on these results of analysis, we can break down to more detail process, and then carry out the process improvements and verification. In these experiments, there are more dark current exist the surface of floating node and STI defects by the side of photo diode and floating node. The dark current in CMOS image sensor represent white pixel. The major related process of “Plasma charging-induced defect” are RPO film、Spacer film etch and Ion implant. …Etc. Defect source are from not only process weakness, but the process weakness of raw material also. The process of raw material wafer will generate intrinsic defects, which is precipitated by oxygen and to be a defect BMD (Bulk Micro Defect). These defects are caused by dislocation pits, which have strong correlation with white pixel counts. The thesis focus on the improvement methodology of ion implant process which has the plasma charging-induced defect and the research of the correlation between the defects generated by raw wafer process and the source of dark current.