製備程序的變因對於低介電孔洞型二氧化矽薄膜性質的影響

Abstract

孔洞型二氧化矽薄膜作為低介電材料具有介電常數低、熱穩定性佳、製程簡單的優點，不過也因其具有孔洞存在，造成機械強度不佳的問題。為了降低介電常數並提升機械強度，達到半導體工業應用的標準，本研究嘗試利用以下方法來改善製程，以期達到上述之目標。 本研究第一部份嘗試改變回流加鹽類實驗的表面修飾步驟以及鍛燒完畢取出前之設定末溫並考慮加入浸潤程序，期望使樣品吸水機率降至最低，同時納入安全上之考量。實驗以70℃酸性條件下加入鹽類TPAOH回流攪拌合成實心的二氧化矽顆粒膠體溶液，而後加入界面活性劑參與反應，旋轉塗佈於矽晶片上，經過軟烤、鍛燒、表面修飾得到低介電薄膜，而後以CV、SEM、nanoindentor量測薄膜性質；第二部份是藉由一連串的測試來決定如何利用水熱法，在鹼性條件下合成具備沸石結構顆粒之膠體溶液，而後添加適量界面活性劑參與反應，以旋轉塗佈的方式鍍於矽晶片上製備低介電常數薄膜，而後應用CV、IV、nanoindentor、SEM、XRD及氮氣吸附脫附系統分析薄膜性質。 第一部份研究結果顯示，有無加入浸潤程序對於薄膜電性並無明顯影響，而鍛燒末溫設定愈高，介電常數可以降愈低，機械強度也相對較好，但仍不能達到工業應用標準。第二部份研究結果顯示，aging時間、水熱溫度及時間會直接影響薄膜的介電特性。Aging時間愈長，產生的晶種數目愈多，後續形成的薄膜結構愈完整，對於電性的量測也有幫助；此外水熱溫度愈高或水熱時間愈長可以使沸石顆粒的結晶性愈完整，配合少量的界面活性劑即可使k值下降至2.30，反觀水熱溫度愈低或反應時間愈短則結晶性愈差，須要配合較多的界面活性劑，才可使k值降至2.16。漏電流及機械強度方面皆可達到工業的需求。 目前本實驗室所製備之最佳薄膜，介電常數=2.16，漏電流=8.37*10-8 A/cm2，楊氏係數=18.0GPa，硬度=2.0GPa。

Mesoporous silica has been widely used for low-k materials because of it’s low dielectric constant, good thermal stability, and easily-preparation procedures. However, the porosity within the film makes the mechanical strength bad. To solve the problem and reach the industrial standard, we made use of some methods as follows which can make the process improved. First, we tried to change the procedures of surface modification (under reflux process), the end set point of calcination temperature and add soaking step in order to lower the opportunity of absorbing water and safety concern. The reflux temperature was 70℃ under acid conditions and adding TPAOH to synthesize the colloidal solution with SiO2 solid particle in it. Then we added tween80 and spanned on P-Si wafer. After baking, calcination and surface modification, we got the low-k films and analyzed by CV、SEM、nanoindentor. Second, we made a series of tests and decided how to synthesize the colloidal solution with zeolite particle in it by hydrothermal process under base conditions. Then we also added tween80, got the low-k films and analyzed by CV、IV、nanoindentor、SEM、XRD、N2 adsorption/desorption system. From the results of first part, soaking step didn’t affect the electric properties of the films. The k value became lower and the mechanical strength became stronger while the end set point of calcination temperature became higher. But the mechanical strength still couldn’t be up to the industrial standard. From the second part, we found that aging time, hydrothermal temperature and time would directly affect the dielectric properties of the films. When the aging time became longer, the films had better structures. The electric properties would be measured more easily. In addition, raising the hydrothermal temperature or prolonging the time would make the particles more crystalline. Adding a little tween80 could lower the k value to 2.30. On the contrary, lowering the hydrothermal temperature or shorting the time would make the particles less crystalline. Adding more tween80 could lower the k value to 2.16. Of all the mechanical strength and the leakage current could meet the industrial standard. The best films our lab can prepared had k value=2.16, leakage current=8.37*10-8 A/cm2, elastic modulus=18.0GPa, hardness=2.0GPa.