延遲烘烤效應與新開發聚焦圖像在深紫外光微影技術之應用

Abstract

當製程技術進展到次微米或深次微米階段時，微影製程技術面臨了相當大的挑戰。主要的原因是曝光的波長必須再降低，意即必須使用波長較短的雷射光源取代原先所一直使用的汞燈光源。在曝光光源轉換到雷射的同時，首先面臨到的問題就是雷射光源的輸出功率問題。由於雷射光源的輸出功率太低，原先所使用的光阻必須更換成化學放大型的深紫外光光阻。對於化學放大型的光阻，曝光後烘烤時的熱能，可使曝光時產生的光酸形成連鎖反應，而讓曝光區域的光阻可以被顯影液所溶解。在深次微米的範圍內，使用深紫外光曝光系統是目前所知技術中最成熟且穩定的技術。因此，全球均使用此技術來執行 0.18 微米或更小線寬的生產與開發。但是化學放大型的深紫外光光阻本身有一個很大的缺點就是對於環境與製程過程非常敏感。例如當環境中的氨氣濃度過高時會使得光阻形成一個 T 型的頂端。若是在光阻曝光後無法立即執行烘烤，此時間延遲即會造成線寬的變異。所幸在光阻中加入鹼性物質可以改善上述兩個缺點。本論文的研究重點包含兩個部份：第一部份即為研究深紫外光光阻在曝光與曝光後烘烤兩者之間有延遲時，光酸濃度變化對於線寬的影響。我們建立了一個模型來準確的描述線寬變化與時間延遲兩者間的關係。在研究過程中，我們發現線寬的變化在一定時間後會達飽和值而不再變動。若利用故意的長時間延遲，不僅可以得到穩定的線寬，同時可以得到較小的光學繞射效應以及較大的製程空間。假使受到機台限制而使曝光後的光阻必須立刻烘烤時，我們所推導出的公式，可以準確的預測線寬隨時間變化的大小值。透過自動化控制，可以將任何異常的晶片挑出而不致產生後續的困擾，因而達到即時製程監控的目的。 眾所皆知的，降低曝光波長與增大曝光機台的數值孔徑可以曝出較小的線寬。而在縮小線寬的同時，曝光的景深也會因此而變小。通常來說，大於 0.8 微米的景深是量產的最小要求。但是在下一代曝光機台愈來愈難製作且成本愈來愈高的情況下，既使景深已被壓縮到 0.6 微米或者甚至 0.4 微米，也會被拿來量產。此時聚焦的準確性也就愈加顯得重要。本論文的第二部份為利用新開發出的一種圖形來準確的量測出曝光機台的最佳焦點。此種圖形運用四種不同大小的圓洞，在加大能量的情況下，使圓洞與圓洞之間的光阻可被繞射的光線所曝開而被顯影液所移除。利用不同大小圓洞在失焦情況下縮小的程度不同，可使這四種圓洞所組成的條狀圖形產生位移。此時，即可利用原先量測層與層疊對關係的機台快速且準確的決定曝光機台的焦點位置。除此之外，此種圖形更具備了廣泛的應用性。從量測機台特性的角度來看，此圖形可以決定機台的最佳焦點、聚焦面傾斜度、聚焦面曲度、或甚至像差等。而由製程方面來看，此圖形可用來補償晶片表面高低差不同所造成的失焦或傾斜度誤判等。除此之外，更可廣泛的利用在其他與最佳焦點相關的應用上，例如：晶片邊緣晶粒的最佳焦點與傾斜度微調、透鏡加熱膨脹所造成最佳焦點偏移的修正、即時產品的最佳焦點與傾斜度修正等等，均可使用此種新開發出的特殊圖形來解決。 總體來說，本論文的主要研究範圍是要解決深次微米技術中所面臨的光阻與最佳焦點兩個大問題。當然，在邁入深紫外光階段的過程中，還面臨著許多問題，例如雷射光源的製作與能量提昇、透鏡材料的製作與研磨、光阻的開發與製作、光罩的製作與修正、缺陷的檢查與降低…..等等，這些都是非常重要且急待解決的問題。本文僅儘個人所能，提供相關的實驗與推導結果，期能對下一世代的製程技術開發略盡棉薄之力。

The lithographic technology faces a challenge when the semiconductor moves into the submicron or deep-submicron era. The principal issue is the exposure wavelength should be lower to quarter-micron or even smaller, i.e., the conventional halogen lamp should be replaced by the laser light source. However, the drawback of the laser light source is its relative lower output power. To conquer the lower power output, the original ultra-violet (UV) photoresist should be replaced by the chemically amplified deep UV photoresist. For a positive DUV resist, a radiation sensitive acid generator is decomposed during exposure, and the subsequent acid-catalyzed thermal reaction at an elevated temperature, i.e. post exposure bake (PEB), makes the resist soluble. As well known, chemically amplified (CA) resist based on acid catalysis for DUV lithography is a promising technology for patterns of 0.18 μm or less. Previously, the main problems for DUV resists were airborne contamination and linewidth change with different delay times. For the positive DUV resist, the generation of “T-top” at the resist-air interface is attributed to neutralization of the photogenerated acid by airborne organic bases, such as ammonia, during post exposure delay (PED). Fortunately, a resist system comprising of a CA resist and an organic base not only prevents a T-top formation, but also suppresses acid diffusion reaction within resist film (the linewidth change is mainly induced by the acid diffusion). In this work, a model was established to describe the linewidth according to the PED time, based on the mechanism of neutralizing organic base and photogenerated acid. It has been found that the linewidth broadened immediately after exposure and eventually became saturated. The resist pattern undergo a long term PED can not only obtain a reliable linewidth control, but also receive a larger process window and smaller optical proximity effect (OPE). If the stepper and track should be running under an in-line configuration, the linewidth variation induced by PED can be prevented by employing our model and automatic control system. Printing smaller linewidth is achievable by either reducing the exposure wavelength or increasing the numerical aperture (NA). However, the Rayleigh depth of focus, DOF = □ □/(2NA2), has become smaller with technology advances. Although 0.8μm or larger DOF is normally a minimum requirement for mass production, the DOF for a specific pattern is generally smaller than 0.6μm. This is owing to that the pattern smaller than the machine's capability is always drawn on the reticle to lower the user’s cost of ownership (COO). In light of the decreasing DOF of modern small wavelength and high NA lithographic tools, the position of best focus must be determined accurately and efficiently. This work presents a novel bar-in-bar (BIB) pattern to monitor the focus and tilting of production wafers. The inner and outer bars contain various hole sizes. When defocused, the shrinkage of the smaller patterns is more significant than that of the larger ones, thus causing the center of gravity to shift. Through the organization of the bar patterns, the centers of inner and outer bars shift in opposite directions when defocused. An overlay measurement tool can be used to easily measure the shift between the centers of inner and outer bars. Therefore, a second order polynomial equation can precisely fit the measured BIB shift. In addition, an accurate and reliable focus value can be obtained with a maximum error less than 0.05μm by simply differentiating the fitting equation. By adding the unique BIB to the scribe lanes of the production wafers, the best focus and tilting of the lithographic tools can be acquired when measuring a layer-to-layer overlay shift and, then, can be fed back to the exposure tool as a valuable reference for following processing wafers. This BIB can also be extended to other useful information, such as lens heating correction, edge die leveling adjustment, and wafer chuck flatness. In summary, this work focused on two of the main problems, PED effect and best focus determination, in the submicron and deep-submicron era. Of course, there are still many urgent issues need to be solved, such as, laser power output improvement, lens manufacturing, photoresist development, reticle manufacturing, defect reduction, etc. This work is just hoping to give some advises and helps to the lithographic field.