新穎接枝型起孔洞劑之多孔性低介電材料探討(I)

Abstract

32 奈米元件製程預計需引進超低介電材料(k≤ 2.5)，其關鍵是將起孔洞劑加入含矽 基的介電質中，然後將低温起孔洞劑於矽基交連後燒除或將高温起孔洞劑延後至金屬層 完成後方燒除。但是旋塗製備的薄膜，其孔洞在相同的起孔洞劑量下會比PECVD 製備 的薄膜大許多而且分佈不均。這源自於起孔洞劑在溶液中或薄膜中因後續熱處理引起之 聚集(aggregation)現象。然而實際應用在後段的製程上，低介電材料之孔洞需要小(<5-10 nm)又分佈均勻。迄今已有研究以快速升溫、加入分散劑、或改變酸鹼值(pH value)等方 式降低起孔洞劑的聚集，進而改善孔洞大小及分佈，然而其成效仍相當有限。 為突破以上缺點， 本計劃將以ATRP 合成具有矽氧烷官能基的起孔洞劑 (PMMA-siloxane 及 PS-siloxane)，並使用接枝方式與基材鍵結而形成新式的低介電材 料。本計劃第一年研究目標可分為三項：(1) 合成具有矽氧烷結構的高温起孔洞劑及其 接枝反應與驗證，(2) 定量顯示孔洞大小及分佈之優勢，(3) 探討分子量與溶劑對孔洞 大小的影響。第二年將持續完成 (1) 探討最大孔隙率下影響孔洞聚集或塌陷之因素及改 良；(2) 探討及鑑量新型低介電薄膜的介電常數，機械性質，及化學等特性。並延伸ATRP 合成PMMA-siloxane 低温起孔洞劑， 探討起孔洞劑之燒除溫度對上述各項之影響與差 異性。

For ultra-low-k dielectrics (k≤ 2.5) used as ILD for 32 nm node and beyond, large porosity was introduced into low k materials matrix using porogens which were burned out immediately after dielectrics deposition or after completion of a Cu/low-k layer in a late-porogen removal scheme. For spin-on dielectrics with large porosity, the pore size is relatively large with wide distribution compared to the porous PECVD dielectric using small molecules as porogen. Such pore morphology may arise from the aggregation of porogen in the solution, and even in the spin-casted film during the curing step. However, for a mechanically robust low-k film with better reliability, it is highly desirable to have small and well-dispersed pore sizes (<5-10 nm) with tight distribution. Yet, approaches such as fast curing rate or adding surfactants have been undertaken to retard the porogen segregation in the solution and casted-films to improve the pore size in porous low-k films with limited success. In this proposal, a novel functionalized porogen is grafted onto the backbone of low-k precursor, which is further crosslinked to low-k matrix with well-dispersed and discrete porogens. This proposal plans to use a spin-on organosilicate such as methylsilsesquioxane (MSQ) as the matrix. In particular, α-siloxane-polystyrene (PS-siloxane) and α-siloxane-poly(methylmethacrylate) (PMMA-siloxane), in which PS and PMMA are high-temperature and low-temperature porogen, respectively, will be synthesized by atom-transfer-radical-polymerization (ATRP) reaction to achieve extremely low polydispersity, desirable for tight pore distribution. Subsequently, PS-siloxane will be grafted onto low-k matrix, MSQ synthesized from methyltrimethoxy silane (MTMS) by sol-gel process, to form a novel low-k MSQ with grafted PS porogen, designated as MSQ-g-PS. We will first validate PS-siloxane structure synthesized by ATRP using 1H-NMR and the grafting of PS-siloxane onto MSQ by comparing their thermal stability using thermal gravimetric analysis. In addition, the pore morphology, pore size and distribution of porous low-k MSQ with grafted-PS porogen (~20% porosity; molecular weight: 5000 g/mole) and their advantages over conventional MSQ/porogen hybrid will be quantitatively illustrated by scanning electron microscope and grazing-incidence small-angle-x-ray-scattering measurement. Upon the confirmation of our proposed grafting porogen scheme, we will investigate the effect of molecular weights of PS on the final pore size/distribution by adjusting the initiator contents and reaction time during ATRP step. Moreover, the impact of solvent compatibility on the final pore size and distribution will be examined. In the 2nd year, we will focus on extending the dielectric constant to the lowest limit by exploring maximum porosity without interconnected pores or film collapse. The pore morphology and pore/distribution for films with large porosity will be also investigated to examine any aggregation issue. The materials properties and processes characteristics of MSQ-g-PS films such as dielectric property, thermo-mechanical properties, and chemical properties will be characterized as function of porosity, MW, and solvents. Finally, we plan to study the synthesis, pore morphology, materials properties, and processes characteristics of low-temperature porogen using PMMA-siloxane based on the same methodology established in the 1st year for MSQ-g-PS films.