Title: 探討金屬雜質對閘極氧化層崩潰電壓的影響與改善
Investigation into Influence of Metallic Impurity on Gate Oxide Breakdown Voltage and Improvement
Authors: 俞文光
陳家富
工學院半導體材料與製程設備學程
Keywords: 閘級氧化層崩潰電壓;閘極氧化層預洗;gate oxide breakdown voltage;pre-gate clean
Issue Date: 2006
Abstract: 本論文針對生產線中0.15μm製程記憶體的閘極氧化層(gate oxide)崩潰電壓(breakdown Voltage)的下降變化與閘極氧化層預洗(pre-gate clean)過程之間的關係作探討。 經由感應耦合電漿質譜分析儀(inductively coupled plasma-mass spectrometer, ICP-MS)與二次離子質譜儀(secondary ion mass spectroscopy, SIMS)的實驗數據證明了此產品的閘極氧化層受到金屬雜質污染,而穿透式電子顯微鏡( transmission electron microscopy, TEM)的影像顯示晶片的矽底材有缺陷,間接證明了製程發生了異狀。從確定問題製程站別的實驗結果推論此產品應是在閘極氧化層生長前的閘極氧化層預洗(pre-gate clean)步驟有異常,導致閘極氧化層(gate oxide)崩潰電壓(breakdown voltage)突然變低。在尋找污染路徑的實驗中得知此金屬雜質污染是來自於廠務中央雙氧水管路。 本論文提出兩個解決方法,結果顯示(1) 廠務中央供酸管路必須定期維護,並提高更換過濾器的頻率從一年一次提高為半年一次,以避免污染源的再現,(2) 將預洗(pre-gate clean)配方中RCA clean 之SC2+SC1 clean順序改為SC1+SC2,即使在廠務中央供酸管路污染尚未被移除的狀況下,此新的方法能使崩潰電壓維持在規格內。方法(1)可以降低管路受到污染的機率,方法(2)可確保產品不會再受到管路污染的影響。同時實施這兩個解決方法更可確保產品的崩潰電壓以後不會再受到類似事件的影響。另外,為了加強監控產品,已經建立一個能快速監控電性的機制來及時反映問題。由於量產的產品從投片到電性測試有好幾百道製程,一旦產品在電性測試的站別才發現異常時,出問題的站別可能已經跑了數週,此法就是用最少的製程做出主動區的絕緣(isolation)及簡單的金屬氧化半導體(metal-oxide- semiconductor, MOS) 的結構後就把晶圓作電性量測。此法可在兩天之內就可得到數據,有別於舊法必須等待數週。 本論文證實確實執行以上三個方法,可以改善閘及氧化層崩潰電壓突然異常下降之故障。
In this paper we focus on the relationship between the breakdown voltage of gate oxide and pre-gate cleaning because the breakdown voltage of gate oxide in 0.15μm product dropped in the production line suddenly. And then solutions are proposed and carried out in the production line. We broke down processes to verify the problem and found that the pre-gate oxide cleaning process could be the main root cause. Before the contaminations were found, we switched SC1+SC2 of RCA Clean in the Pre-Gate Clean. Switching SC1 and SC2 is able to have breakdown voltage stay in spec. Data from ICP-MS (Inductively Coupled Plasma-Mass Spectrometer) and SIMS (Secondary Ion Mass Spectroscopy) also show that the gate oxide was contaminated by the metallic impurities. Also we used different chemical suppliers and chemical delivery paths to verify where the contaminations came from and experimental data show that the contaminations came from H2O2 in the facility central supply system. Therefore two solutions were proposed and carried out : 1.) the facility central supply system should be maintained regularly and the filters should be replaced more frequently ; from every 12 months to every 6 months 2.) the pre-gate oxide cleaning should be changed to SC1->SC2->HF from SC2->SC1->HF ,so that the breakdown voltage can stay in spec before the contaminations are removed. The above solutions have been accomplished and proven effectiveness. In order to monitor gate oxide quality, a rapid monitor system has been set up so that such a case won’t kill so many wafers in the future. By this rapid monitor, the process engineer can have data within 2 days. To carry out the above solutions can prevent such an excursion in the future.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT009375525
http://hdl.handle.net/11536/80300
Appears in Collections:Thesis


Files in This Item:

  1. 552501.pdf

If it is a zip file, please download the file and unzip it, then open index.html in a browser to view the full text content.