本論文建立了兩個三元平衡相圖，分別為927℃靠近矽鍺端之Nb-Si-Ge三元平衡相圖與600℃ Cu-Si-Ge三元平衡相圖。並且也對Cu/Ge及Cu/Si之固態擴散反應行為作初步的探討。本論文之主要目的在提供產學界在設計Si1-xGex/Si或Si元件接觸材料時之基本熱力學參考資料。 在三元平衡相圖的研究中，我們結合X光繞射(XRD)、電子微探儀(EPMA)及金相觀察等分析結果，建立927℃靠近矽鍺端之Nb-Si-Ge三元平衡相圖與600℃ Cu-Si-Ge三元平衡相圖。在927℃靠近矽鍺端之Nb-Si-Ge三元平衡相圖研究方面，證實在927℃下靠近矽鍺端之Si-Ge-NbSi2-NbGe2 梯形區域中僅有Nb(Si1-xGex)2與 Si1-yGey兩相存在，也就是說NbSi2會和NbGe2形成C40結構之連續固溶液Nb(Si1-xGex)2。此固溶液相之不同組成x值與其晶格常數a、c值的關係也被本實驗測得。此外，我們發現在Nb(Si1-xGex)2-Si1-yGey兩相間之tie-lines有偏向NbSi2和Ge兩端的趨勢，推測此乃因NbSi2的生成焓(enthalpy of formation)較負於NbGe2之故。再者tie-lines 亦顯示NbSi2 和NbGe2皆無法單獨與Si1-yGey固溶液達到熱力學穩定狀態，唯有Nb(Si1-xGex)2固溶液方能與Si1-yGey固溶液達到熱力學穩定狀態。 在600℃ Cu-Si-Ge三元平衡相圖研究方面，確認Cu7Si與Cu5Ge會形成不等組成範圍A3結構之(Cu7Si, Cu5Ge)連續固溶液相。另外發現Cu3Si至少可溶進19.1 at﹪Ge，而Cu3Ge則最多只溶進2 at﹪Si，Cu3Si與Cu3Ge並不會形成連續之固溶液。由本實驗之結果推測Pearson’s Handbook of Crystallographic Data for Intermetallic Phases上之ε1-Cu3Ge相晶格結構資料可能有誤，ε-Cu3Ge和ε1-Cu3Ge相之XRD繞射圖譜可能相同。若此兩相繞射圖譜不相同則可能是Cu-Ge二元相圖上有關ε-Cu3Ge和ε1-Cu3Ge之穩定存在溫度有誤。再者本實驗發現Cu3Ge只會溶進極少量之Si，形成接近Cu3Ge電阻係數值之Cu3(Si1-xGex)，其能與Si1-yGey相達熱力學穩定狀態，並且仍保有Cu3Ge極低電阻係數值之優點。因此我們可利用此特性，在Si1-yGey上設計出穩定存在之Cu3(Si1-xGex)接觸材料。 由擴散實驗顯示銅鍺接觸時Cu3Ge相生成速率遠大於Cu5Ge。且Cu3Ge相在600℃下只有單一結構之反應層產生。推測Cu-Ge二元相圖對於兩不同結構之Cu3Ge相的穩定存在溫度可能有誤。或者兩不同結構之Cu3Ge相的交互擴散係數相差較大。另外本實驗也發現在600℃下以夾具的方式進行銅矽擴散反應時，接觸之界面處並無任何擴散的跡象。 本研究之最終目的是在提供探討NbSi2與Si1-yGey/Si接觸時之熱穩定性以及Cu3Ge與Si1-yGey/Si及Si接觸時所需之基礎數據。這些資料當對微電子業界在開發新的接觸材料之金屬化製程有相當助益。 In this thesis , the semiconductor-rich region of the Nb-Si-Ge ternary isotherm at 927℃ and the Cu-Si-Ge ternary isotherm at 600℃ were determined by using X-ray diffraction、electron-probe microanalysis and metallography. The main objective is to provide the necessary thermodynamic information for designing contact materials for applications in SiGe or Si devices. In the second part of this thesis, the diffusion behaviors of Cu-Si-Ge system were preliminarily studied. The objective is to provide the kinetic information for understanding the reaction between Cu and SiGe. It was confirmed that at 927℃ NbSi2 and NbGe2 form a continuous solid solution Nb(Si1-xGex) with the C40 crystal structure. It was also shown that, other than Nb(Si1-xGex) and Si1-yGey , there is no known binary or ternary phase within the Si-Ge- NbGe2- NbSi2 trapezoid. The lattice parameters of Nb(Si1-xGex) were determined. The tie-lines for the Nb(Si1-xGex)- Si1-yGey two phase region tilt slightly toward the NbSi2 and Ge corners presumably because the enthalpy of formation for NbSi2 is more negative than that of NbGe2. The tie-lines also show that the NbSi2 and NbGe2 are not stable when they are in contact with SiGe solid solution alone. In the Cu-Si-Ge ternary isotherm study at 600℃ study, we confirmed that Cu7Si and Cu5Ge formed a continuous solid solution with the A3 crystal structure ,but Cu3Si and Cu3Ge didn’t. We found that the crystal structure information of e1-Cu3Ge in the ”Pearson’s Handbook of Crystallographic Data for Intermetallic Phases” could have some errors. The diffraction pattern of ε-Cu3Ge and ε1-Cu3Ge could be the same. Otherwise the stably existing temperature of ε-Cu3Ge and ε1-Cu3Ge in the Cu-Ge binary phase diagram would have wrongs. The Si would slightly dissolve into Cu3Ge to form the Cu3(Si1-xGex) with low electrical resistivity close to Cu3Ge and the Cu3(Si1-xGex) would be stable on the SiGe substrate. Therefore, we can use this property of Cu3(Si1-xGex) to design the stable contact material on the SiGe substrate. In the second part of this thesis, Cu/Ge binary diffusion couple experiment at 600℃ was performed. The growing rate of the Cu3Ge phase is greater than that of the Cu5Ge phase. The one and only crystal structure of the Cu3Ge phase was detected. We assumed that the stably existing temperature of the Cu3Ge phase with different crystal structures in the Cu-Ge binary phase diagram had some errors, or the variation in interdiffusion coefficients between the two different crystal structures of the Cu3Ge phase was large. We found that no detectable reaction in the Cu/Si binary diffusion couple occurred at 600℃. The main objective is to provide the necessary thermodynamic information for NbSi2 and Cu3Ge when they contact with SiGe or Si substrate. The information would be a great benefit to the development of metallization for the new contact materials in the microelectronic industrial.
