快閃記憶體元件抹寫週期可靠度中續航力與操作電壓及氧化製程之關係

Abstract

隨著半導體元件製造科技的進步，元件的規格也越縮越小。因此，快閃記憶體的耐用性議題變得更加重要。在這篇論文中，因為氧化層的劣化是影響快閃記憶體續航力的主要原因，我們透過比較使用不同操作電壓的記憶體元件來研究氧化層劣化的機制。此外，我們將分別使用自由基氧化製程和熱氧化製程來形成氧化層的快閃記憶體元件來實驗，比較其耐用性的表現，檢測自由基氧化製程是否能夠改善氧化層的品質。 在經過一萬次寫入抹除之後，透過新的研究方法，我們能夠將浮動閘極電荷變化獨立出來，單獨分析在抹寫中產生的氧化層捕獲電荷量。透過實驗數據的分析和可能模型的假設，我們能夠計算出氧化層中正電荷和負電荷的數量及分布，並和實驗數據做對比，從而得知氧化層在多次抹寫之後劣化的原因，並可以得知自由基氧化製程和熱氧化製程相比之下的優缺點。

Semiconductor devices is continuously scaling down as device process technology being improved. Thus, endurance issue becomes more important especially in NAND flash memory. In this thesis, since tunneling oxide degradation is the main factor in endurance characteristic, we operated NAND flash memory cells with different biases to investigate the mechanism of oxide degradation. In addition, to see if Radical Oxidation process can improve the oxide quality, we compared the performance of cells whose tunneling oxide are grown by Thermal-Dry Oxidation (Dry) and Radical Oxidation (RO) respectively. After 10K cycles, by using new method we can investigate the oxide trapped charge (QOX) generated by program and erase (P/E) cycles without the influence of floating gate (FG) charge (QFG). By examining the measurement data and assumed possible models, the profile of positive and negative charges in the tunneling oxide can be estimated. Thus we can find out what happened in the tunneling oxide during P/E cycles and compare the advantages and disadvantages of radical oxidation process and thermal-dry oxidation process.