減光型相移圖罩材質研究與二孔偏軸發光成像模擬

Abstract

嵌附式減光型相移圖罩與偏軸發光均為常用之解像度增進技術。因此本論文針對這兩部份做研究，第一部份重點在研究目前主流微影波長193 nm與248 nm嵌附層材料之光學性質，第二部份為減光型相移圖罩搭配偏軸發光應用之模擬。 嵌附層材料薄膜不論使用何種方式沉積，均無法完全如鏡面般光滑，且薄膜表面粗糙度會引起散射光的形成，導致紫外光/可見光光譜儀量測之反射率與透射率較真實值低。粗糙度越大時，誤差相對越大，是故，以粗糙度為修正因子，對反射率與透射率之量測值做適當修正是必要的。修正後之反射率與透射率再應用於反射率-透射率法，所得之折射率與吸收係數不但接近商業儀器橢圓儀之量測值，更優於n&k分析儀之量測結果，所需花費亦較低廉，足證此修正法之價值。 在電腦模擬實驗部分，使用美商KLA-Tencor之微影模擬軟體ProLith v. 8.0，調整各項製程參數進行模擬計算，以探討二孔發光所產生的側葉效應、空間影像與阻劑輪廓等關係。模擬結果發現”隙寬”與”隙與側葉中心點距離”兩者之關係，可計算導出指數公式，加上聚光當量(數值孔徑)與發光孔距為修正項，計算結果將更貼近真實值。模擬顯示二孔垂直位向發光之側葉效應遠小於平行位向，對微影製程影響不大。二孔垂直位向為偏軸發光而平行位向類似沿軸發光，因此垂直位向之聚焦深度較大，在相同顯影條件下，阻劑輪廓可發現二孔垂直位向曝光所得之隙深度較平行位向深；如調整顯影條件使兩種位向之發光方式均曝出相同尺寸之隙時，則平行位向之側葉可能顯影於阻劑上。此外，模擬實驗亦驗證s偏振光較佳之干涉現象，可搭配二孔垂直位向獲得較好之成像品質，未偏振光次之，p偏振光最差。模擬探討了s偏振光、p偏振光與未偏振光在聚光當量(數值孔徑)、孔距、孔徑、隙寬與線寬等變因下，空間影像對比度(像比)之變化。 未來使用濕浸式微影或機台之聚光能力不斷提升之時，s偏振光源將漸受重視。

Embedded attenuated phase-shifting mask (EAPSM) and off-axis illumination (OAI) are most popular resolution enhancement techniques (RETs). Hence, we do some research to these two techniques in this paper. The points of the first part are optical properties of materials of embedded layer for 193 nm (ArF) & 248 nm (KrF) lithography. And the second part is the simulation of application of attenuated phase-shifting mask with off-axis illumination. No matter what kind of method we use to deposit thin film, it can’t be completely smooth. And the rough surface of thin film will cause the scattering light. Hence, when we use UV/VIS spectrometer to do metrology of reflectance (R) and transmittance (T) will get smaller results than the true values. The error will get larger when roughness gets larger. So, it is necessary to modify the metrology values of R and T by roughness as a modification factor. When we use the modified results to the R-T method, the resulting refraction (n) and extinction coefficient (k) are close to the metrology results of Variable Angle Spectroscopic Ellipsometer (VASE), and better than n&k analyzer. Also, the cost is cheaper, so that the modified R-T method is very useful in getting n & k. In the simulation part, the software we use is KLA-Tencor’s Prolith v. 8.0. By choosing proper parameters, we study the side lobe, aerial image and resist profile in dipole illumination. By different space width, we can get different distance between space and side lobe. And we try to derive a formula. By adding two modification factors of numerical aperture (NA) and distance between the two illumination holes, the calculation results are very close to the true values, According to the simulation results, the side lobe in vertical orientation is much smaller than parallel orientation in dipole illumination, so it has a smaller effect to lithography process. In dipole illumination, the vertical orientation is off-axis illumination and the parallel orientation is like in-axis illumination. So we can get larger depth of focus (DOF) in vertical orientation. We also can see deeper resist profile in vertical orientation than parallel orientation with the same develop condition. If we adjust the develop condition to make the same space width, we can see significant side lobe in resist profile with parallel orientation. Otherwise, we found the s-polarized light has better interference. With vertical dipole illumination, we can get better image quality. And the unpolarized light has a middle quality, the p-polarized is worst. We also discuss the variation of aerial image contrast for s-polarized, unpolarized and p-polarized light with different NA, pole-distance (σcenter), pole radius (σradius), space width and line width. For future immersion lithography and higher NA lens, the s-polarized light will become more important.