可接地基板於電子束微影之應用

Abstract

在電子束微影系統中，如果阻劑的導電度不佳時，容易引發多餘電子堆積的現象。本實驗在塗佈阻劑前先將金屬沉積於基板上或深入基板內，並將金屬接地，以形成接地層、接地釘以及接地牆等結構，透過接地層、接地釘以及接地牆的導電性可讓累積在基板中的電子排出。此外，金屬存在的另一項功用在於金屬與電子束之間所存在的影像力(Image force)。影像力為影像電荷所衍生之電場，由於電場線與接面垂直，可幫助電子於穿透阻劑時以垂直於表面的方向入射，之後再透過金屬將多餘電子接地，降低電子束鄰近效應之發生。而降低電子束鄰近效應將會使實驗上微影出來的圖案與所定義圖案真確性較為接近，以增加電子束微影系統的效能。 本篇論文中將探討可接地基板於低電壓電子束微影之應用，而電子束的加速電壓設定為5kV。本文所定義的可接地基板是將金屬沉積於基板上或深入基板內，且將金屬接地，所形成的接地層、接地釘以及接地牆等結構。本文中的可接地基板中依照金屬的位置分為三種，第一種是在基板上方沉積整片金屬層，第二種是在基板上方曝光區域的周圍沉積金屬層，以上兩種都是屬於接地層的結構，而第三種是除了在基板上方曝光區域的周圍沉積金屬層以外，金屬並深入基板內形成樁(piles)以製成接地釘的結構，或者金屬深入基板內形成溝(trench)以製成接地牆的結構。除此之外，本實驗中的可接地基板依照基板的材料又分為兩種，分別為矽與二氧化矽，其分別具有不同的金屬沉積位置與厚度，總共分為多種樣品。我們將不同阻劑實作在這些樣品上，並將電子束微影後的實驗結果以掃描式電子顯微鏡觀察。最後，透過分析實驗結果中的直線的雙邊邊緣的粗糙度，討論可接地基板對實驗結果的影響。 我們從實驗結果得到的結論是可接地基板在加速電壓為5kV的低能量電子束微影中的確可以改善圖案的品質。在三種不同阻劑於可接地矽基板上所表現的微影結果中，直線品質最差的樣品，即矽基板上未鍍鈦的樣品，它的LWR除以線寬的值是直線品質最佳的樣品，即矽基板上方沉積厚度40nm的鈦的樣品的2.5倍左右。而在zep520A於可接地二氧化矽上所表現的微影結果中，直線品質最差的樣品，即二氧化矽上未鍍鈦的樣品，它的LWR除以線寬的值是直線品質最佳的樣品的2倍左右。除此之外，在不同的基板材料上利用低電壓電子束微影技術製作圖案，也會得到不同的微影結果。從實驗結果可發現，分別在未鍍任何金屬的矽基板與二氧化矽上製作圖案，在二氧化矽上所得到的微影結果較差，但是若利用鈦在二氧化矽上製作結構，則可改善在二氧化矽上製作樣品時所表現出來的微影結果，且其微影結果甚至能夠比在矽基板上未鍍鈦的樣品所表現出來的微影結果更好。

In electron beam lithography, it is easily to have many electrons to be collected if the conductivity of the photo-resist is not good enough. In this experiment, we evaporated a thin metal layer on or in the substrate to form groundable layer, groundable nail and groundable wall before spin-coating the photo-resist on the substrate, and the metal is grounded so that the electrons collected in the substrate can be discharged by the conductivity of the metal. Besides, the other function of the metal is that there exists image force between the metal and the electron beam. Image force is an electric field produced by image electric charges. Because the electric field is perpendicular to the surface, it can help the electrons to pass through the photo-resist in the direction of vertical to the surface. Then the electrons can be grounded through the metal and we can reduce the proximity effect. By reducing the proximity effect, the patterns we get after the lithography can be closer to the patterns we have defined before the experiment. And then we can enhance the performance of the electron beam lithography system. In this thesis, we are going to study the applications of groundable substrates in low voltage electron beam lithography. And the acceleration voltage is 5kV. The groundable substrate in this thesis is defined as the substrate with metal on the substrate or inside the substrate, and the metal is grounded in order to form groundable layer, groundable nail and groundable wall. There are three kinds of locations of the metal. The first kind of locations is where the metal is evaporated on the whole surface of the substrate. The second kind of locations is where the metal is evaporated around the exposure area on the substrate. These two kinds of samples belong to groundable layer. The third kind of locations is where the metal is evaporated around the exposure area on the substrate and is filled in the piles in the substrate to form groundable nail or the metal is filled in the trenches in the substrate to form groundable wall. Besides, we design many samples which are different from each other in the location and the thickness of the metal based on two different substrate materials, silicon and silicon dioxide. Then we try different kinds of photo-resist on these samples. And we observe the result by using scanning electron microscope. At last, we analyze the line-width roughness of the patterns and discuss the effect of groundable substrate in experimental results. We can reach a conclusion from the experimental results that using groundable substrates can improve the quality of patterns in low voltage electron beam lithography. In the results of three different resists on groundable silicon substrate, the value of line-width roughness divide by line-width of the sample with the worst quality is two point five times larger than the sample with the best quality. And in the results of zep520A on groundable silicon dioxide substrate, the value of line-width roughness divide by line-width of the sample with the worst quality is two times larger than the sample with the best quality. Besides, we can have different results by using different substrate materials in low voltage electron beam lithography. From the experimental results, in the case of there is no metal evaporated on the substrate, we find out that the quality of patterns on silicon dioxide substrate is worse than the quality of patterns on silicon substrate. But if we have titanium evaporated on the silicon dioxide, we can improve the quality of the patterns in experimental results. And with proper titanium structure on or in the silicon dioxide substrate, the quality of the results can be even better than the results of samples with no metal evaporated on the silicon substrate.