電子束微影之低溫短顯影研究

Abstract

微影技術是積體電路IC製程的關鍵技術，也是讓摩爾定律能夠繼續前進的重要推手，半導體工業之所以能夠快速發展，晶片能夠越做越小且價格越來越便宜都與微影技術的發展息息相關，一直以來光學微影是半導體製程的主流，其優點為可大量生產且速度快成本低廉，相較於其他微影技術有很大的優勢，然而隨著晶片越來越小傳統的光學微影技術已經面臨了極限，需要新的微影技術研究，目前電子束微影(E-beam, Electron Beam Lithography)及極紫外光微影(EUV, Extreme Ultraviolet Lithography)是未來顯影技術研究的主流。 本論文主要研究如何改善電子束微影中的鄰近效應。為了改善鄰近效應我們使用了三種方法，改變電子阻劑、縮短顯影時間、降低顯影溫度，在變換電子阻劑的實驗中我們分析了使用不同電子阻劑的優缺點，並選擇了適合進行後續實驗的電子阻劑，接者我們利用電子束系統設計出單點實驗，透過曝光劑量與顯影條件的改變來觀察高斯單點的直徑變化，從實驗結果我們發現透過結合短顯影時間與低溫顯影二種條件，能夠大幅降低鄰近效應，繪製出較小的圖案。之後我們藉由實驗的結果以及傳統電子束散射的理論基礎一步步推導了適用於短顯影時間的速率模型並進行了擬合，發現擬合結果非常好，且各種參數都能夠有很好的物理解釋，接著我們利用低溫短顯影的技術以及模型理論，進行了不同週期奈米線的實驗，成功製作出了9nm的小線寬，並更進一步做出週期30nm的小線寬圖形，同時也證實模型的預測正確而且由高斯單點得到的低溫短顯影實驗結果確實可以運用於畫線上而且有相同的趨勢，最後我們透過反應式離子蝕刻和蒸鍍的方式成功的將我們的圖案轉移至基板上，製作出了小週期奈米線的結構。

Lithography is the key technology in integrated circuits manufacturing process, and the improvement of it is the main reason that Moore's law can keep going. The rapid development of semiconductor industry and chips can become smaller and cheaper are closely related to the progress of lithography.For a long time, optical lithography is the mainstream in semiconductor industry, it is superior to other lithography method because of its mass production with high speed and low cost. However, with chips size become smaller and smaller optical lithography has reached its limit, it is necessary to investigate a new method for lithography, E-beam (Electron Beam Lithography) and EUV (Extreme Ultraviolet Lithography) are the main research direction lithography method in the future. In this thesis, we focus on how to reduce the proximity effect in electron beam lithography. We propose three method to solve proximity effect, including changing the resist, shortening the developing time, and reducing the developing temperature. After the experiment of changing resist, we compare the pros and cons in different resist, and we choose the most suitable resist for our subsequent experiment. Then we use the electron beam system to design single spot experiment and observing the diameter broadening of spots by changing exposure dose and developing condition. Through the experiment result, we found that by combining the short time and low temperature development we can reduce the proximity effect substantially. We further use the experiment result and traditional electron scattering research to derive our model for short interval development, then we fitted our model with experiment data and found that it fitted very well and every coefficient in our model have a good physical meaning. We next use our model to write different pitch lines pattern, and succeed in making 9nm small line width and 30nm small pitch line patterns, and also proved our single spot experiment result can apply to write line patterns and have the same trend. Finally, we use reactive ion etch and evaporation to transfer our pattern to silicon substrate successfully.