電阻式記憶體串、並聯之量測與浮動端之影響與多晶矽薄膜電晶體及非晶相銦鎵鋅氧化物薄膜電晶體之低頻雜訊分析

Abstract

在本論文中，第一部分為電阻式記憶體 (RRAM)的物理機制模型。RRAM是一種非揮發性記憶體，被視為最可能取代目前傳統快閃(flash)記憶體的候選者之一。RRAM的操作模式為利用外加電壓，可將其在低電阻態(LRS)與高電阻態之(HRS)間轉換，由此記錄邏輯之0或1。其優點在於具有低功率消耗與低操作電壓(寫入電壓 < 3V,讀取電壓~0.1V)、結構簡單(在過鍍金屬氧化物上下各夾金屬電極)、可多階操作 (可大幅提升記憶體密度)、讀取與寫入速度非常快 (< 10 ns)、耐用度高 (> 10 年)等優勢。元件的製備由工研院(ITRI)提供，量測一部分為工研院所提供，一部分則在台灣大學量測。我們對RRAM做了許多不同的實驗，例如改變電流限流(current compliance), 最大負電壓VSTOP,以及分析RRAM 元件之串、並聯之操作。我們發現串聯SET時會發生所謂的浮動端影響，進而提出物理模型來解釋，並提出方法來避免此現象之發生。 第二部份為薄膜電晶體低頻雜訊的理論分析。元件的製備為友達光電(AUO)所提供，並在國家奈米中心(NDL)進行低頻雜訊量測。一開始先分析McWhorter Model之物理，並與其他在文獻上其他兩組模型做比較。接下來從量測結果進一步探討物理機制並加以解釋。 第三部分為IGZO材料之低頻雜訊量測，元件的製備為奇美光電(CMO)所提供，並在國家奈米中心(NDL)進行低頻雜訊量測。因這部分目前比較少人再研究，希望透過低頻雜訊的量測了解其物理特性。

In this thesis, the first part is the model and physical mechanism of Resistive random access memory (RRAM). RRAM is one of the many types nonvolatile memory, which is the most promising candidate to replace traditional flash memory. External voltage is added on RRAM to switch the device between low resistance state (LRS) and high resistance state (HRS), and recorded to Logic 0 or 1. RRAM has advantages of low power consumption and low operation voltage (write voltage <3V, read voltage ~0.1V), a simple structure (a transition metal oxide layer between top and bottom layer electrode), multilevel operation (can dramatically increase memory density), very fast read and write speed (<10ns), high durability (>10yrs). Our device is prepared by Industrial Technology Research Institute (ITRI), some measurements are done in ITRI, and others and done in NTU. We have did different experiment test on RRAM, such as different current compliance, maximum negative voltage Vstop, and analysis of series operation of two RRAM cell. We discover for series SET test, there will be so-called floating terminal effect; we will propose a model for thi phenomenon, and a way to prevent this. The second part is the physical analysis on Thin-Film Transistor (TFT). The device is prepared by AU Optronics Corp.(AUO), and measured at National Nano Device Laboratories (NDL). We will first start from the analysis of physics of McWhorter Model compare with other two models in the literature. Next, we will further discuss and explain the physical mechanism from the measurement results. The third part is the low frequency analysis of α-IGZO, where the device is prepared by Chimei Innolux Corporation (CMO), and measured at National Nano Device Laboratories (NDL). Since there are little research on this topic, we hope with the analysis of low frequency noise, we can learn more about the physical mechanism of α-IGZO.