449829 案號 87116160 如年4月 修正1 五、發明說明（1) 本發明是有關於一種金屬化製程（Metallization process)，且特別是有關一種在金屬化製程中降低鋁銅合 金層凹凸現象（Metal void and hillock)的方法。 在有回蝕或有化學機械研磨法的鎢插塞製程（W-p lUg process with etch back or with CMP)中，用於内連線 之金屬層通常由鈦層/氮化鈦層/鋁銅合金層/氮化鈦層 (Ti/TiN/AlCu/TiN-ARC)、氮化鈦層/鋁銅合金層/氮化鈦 層（TiN/AlCu/TiN-ARC)及氮化鈦層/鋁銅合金層/鈦層/氮 第1圖即習知有回蝕或有化學機械研磨法的鎢插塞製 程中形成鋁銅合金堆疊（AlCu stack)的示意圖。 在第1圖中’半導體基底10表面具有一介電層12(如二 氧化矽層），用以形成鎢插塞。鋁銅合金堆疊則依序在介 電層1 2上沈積一鈦層14(Ti)、氮化鈦層16(TiN)、鋁銅合 金層18(AlCu)、氮化鈦層20(TiN-ARC)以形成 Ti/TiN/AlCu/TiN-ARC結構。鈦層14及氮化鈦層16的主要 目的是增進鋁銅合金層18與介電層丨2間的黏著能力。而氮 化鈦層20的主要目的則是反反射，亦即，避免因鋁銅合金 層18反光而造成微影製程的錯誤。 在這種金屬化製程中，整個銘銅合金堆疊在完成後必 須在表面旋塗一層非有機SOG(Non-〇rganic S0G)、低電容 率值HSQ (Low-k HSQ)或有機Fla]：e(Organic Flare)，並 進行固化製程（Curing)以增加旋塗介電層的穩定性。 不過’無論是採用何種旋塗材料，這種紹銅合金堆疊449829 Case No. 87116160 Amended as of April 1 V. Description of the invention (1) The present invention relates to a metallization process, and in particular to a method for reducing the unevenness of aluminum-copper alloy layers during metallization (Metal void and hillock). In the tungsten plug process (Wp lUg process with etch back or with CMP) with etch back or chemical mechanical polishing, the metal layer used for the interconnect is usually composed of a titanium layer / titanium nitride layer / aluminum-copper alloy layer. / Titanium nitride layer (Ti / TiN / AlCu / TiN-ARC), titanium nitride layer / aluminum-copper alloy layer / titanium nitride layer (TiN / AlCu / TiN-ARC) and titanium nitride layer / aluminum-copper alloy layer / Titanium layer / Nitrogen Figure 1 is a schematic diagram of forming an aluminum-copper alloy stack (AlCu stack) in a conventional tungsten plug process with etchback or chemical mechanical polishing. In the first figure, a surface of the semiconductor substrate 10 has a dielectric layer 12 (such as a silicon dioxide layer) for forming a tungsten plug. For the aluminum-copper alloy stack, a titanium layer 14 (Ti), a titanium nitride layer 16 (TiN), an aluminum-copper alloy layer 18 (AlCu), and a titanium nitride layer 20 (TiN-ARC) are sequentially deposited on the dielectric layer 12. ) To form a Ti / TiN / AlCu / TiN-ARC structure. The main purpose of the titanium layer 14 and the titanium nitride layer 16 is to improve the adhesion between the aluminum-copper alloy layer 18 and the dielectric layer 2. The main purpose of the titanium nitride layer 20 is anti-reflection, that is, to avoid errors in the lithography process caused by the reflection of the aluminum-copper alloy layer 18. In this metallization process, the entire Ming Cu alloy stack must be spin-coated with a layer of non-organic SOG (Non-〇rganic S0G), low-permittivity value HSQ (Low-k HSQ), or organic Fla] after completion: e (Organic Flare), and a curing process is performed to increase the stability of the spin-on dielectric layer. But ’no matter what kind of spin coating material is used, this copper alloy stack
0503-3890-EFl.ptc 第4頁 449829 五、發明說明（2) 在固化製程中都會產生極大的熱應力（Thermal stress)， 使銘銅合金堆疊表面出現明顯的凹凸現象（Metal void and hillock)。並且，固化製程的高溫，鋁銅合金層亦可 能與底下的氮化鈦層反應，影響銘銅合金堆疊的導電特 性。 有鑑於此，本發明的主要目的便是提供一種在金屬化 製程中形成鋁銅合金堆疊的方法，其可以減輕鋁銅合金堆 疊所承受的熱應力，進而大幅消除鋁銅合金堆疊表面的凹 凸現象。 本發明的另一目的便是提供一種在金屬化製程中形成 鋁銅合金堆疊的方法，其利用破真空製程預先在氮化鈦層 表面形成阻障用的薄氧氮化鈦層，因此鋁銅合金層與底下 氮化鈦層間的不當反應亦可以充分避免。 為達上述及其他目的，本發明乃提供一種在金屬化製 程中降低鋁銅合金堆疊凹凸現象的方法。這種方法的步驟 係’首先，提供一基底（如二氧化矽層）、並在基底表面依 序形成有離子化金屬電紫鈦層（Ionized metal plasma Ti，ΪΜΡ Ti)及氮化鈦層。然後，再進行破真空製程，藉 以在氮化鈦層表面形成有氧氮化鈦層，用以避免氮化鈦層 與隨後形成的鋁銅合金層反應。接著，利用高溫低功率在 氮化鈦層表面形成鋁銅合金層，藉以減輕鋁銅合金層的應 力，此時的鋁銅合金由於利用IMP-Ti當底層，因此可有效 地產生（lll)AlCu優選方向，即（lll)-textured ;以及， 在鋁銅合金層表面形成反反射層，藉以避免鋁銅合金層反0503-3890-EFl.ptc Page 4 449829 V. Description of the invention (2) Extreme thermal stress will be generated during the curing process, which will cause obvious concave and convex phenomena on the surface of the copper alloy stack (Metal void and hillock) . In addition, due to the high temperature of the curing process, the aluminum-copper alloy layer may also react with the underlying titanium nitride layer, which affects the conductive properties of the copper alloy stack. In view of this, the main object of the present invention is to provide a method for forming an aluminum-copper alloy stack in a metallization process, which can reduce the thermal stress experienced by the aluminum-copper alloy stack, thereby greatly eliminating the unevenness on the surface of the aluminum-copper alloy stack. . Another object of the present invention is to provide a method for forming an aluminum-copper alloy stack in a metallization process. A vacuum breaking process is used to form a thin titanium oxynitride layer on the surface of the titanium nitride layer in advance. Improper reactions between the alloy layer and the underlying titanium nitride layer can also be fully avoided. To achieve the above and other objectives, the present invention provides a method for reducing the unevenness of aluminum-copper alloy stacks during metallization. The steps of this method are as follows. First, a substrate (such as a silicon dioxide layer) is provided, and an ionized metal plasma Ti (UMP Ti) layer and a titanium nitride layer are sequentially formed on the surface of the substrate. Then, a vacuum breaking process is performed to form a titanium oxynitride layer on the surface of the titanium nitride layer to prevent the titanium nitride layer from reacting with the aluminum-copper alloy layer formed later. Next, an aluminum-copper alloy layer is formed on the surface of the titanium nitride layer by using high temperature and low power, thereby reducing the stress of the aluminum-copper alloy layer. At this time, the aluminum-copper alloy can effectively generate (lll) AlCu because it uses IMP-Ti as the bottom layer. The preferred direction is (lll) -textured; and, an anti-reflection layer is formed on the surface of the aluminum-copper alloy layer to prevent the aluminum-copper alloy layer from reflecting.
C:\ProgramFiles\Patent\0503-3890-E.ptd第 5 頁 .98^3_________ 五、發明說明（3) ~ '一--: 光，成微影製程的錯誤。在這種方法中，破真空製程可以 將氮化鈦層曝露在空氣中或利用氧化氮電漿處理（NO plasma treatment)氮化鈦層，反反射層可以是氮&矽 層’而基底則可以是二氧化鈦層或其他材料。 為讓本發明之上述和其他目的、特徵、和優點能更明 顯易懂，下文特舉一較佳實施例，並配合所附圖 細說明如下： ’ 圖式說明 第1圖係習知有回钱或有化學機械研磨法的鎢插塞製 程中形成鋁銅合金堆疊的示意圖；以及 第2A〜2C圖係本發明在金屬化製程中降低鋁銅合金堆 疊凹凸現象的方法的示意圖。 實施例 為避免鋁銅合金層與底下的氮化鈦層反應，並降低銘 銅合金堆疊所承受的熱應力，進而減少其表面產生的凹凸 現象’本實施例的主要特徵係，在形成鋁銅合金層前，預 先對氮化鈦層進行破真空製程（將氮化鈦層曝露於空氣或 以氧化氮電漿處理氮化鈦層）’藉以使氮化鈦層表面與空 氣或氡化氮電漿（N2〇 plasma)的氡反應形成一薄氧氮化鈦 層。薄氧氮化欽層是用來避免氮化鈦層與隨後形成的鋁鋼 合金層因固化製程的高溫而不當反應。本發明的另一個特 徵係’以IMP欽層及氮化鈦層改變紹銅合金層形成表面的 (1 U )晶格排列，並利用高溫低功率形成低應力鋁銅合金 層。由於（111)結構的鋁銅合金層亦可以降低金屬厚子遷C: \ ProgramFiles \ Patent \ 0503-3890-E.ptd page 5 .98 ^ 3 _________ V. Description of the invention (3) ~ 'a-: Light, lithography process error. In this method, the vacuum breaking process can expose the titanium nitride layer to the air or use a nitrogen oxide plasma treatment (NO plasma treatment) on the titanium nitride layer. The anti-reflective layer can be a nitrogen & silicon layer and the substrate can be It may be a titanium dioxide layer or other material. In order to make the above and other objects, features, and advantages of the present invention more comprehensible, a preferred embodiment is exemplified below and described in detail in conjunction with the accompanying drawings as follows: Schematic diagram of forming an aluminum-copper alloy stack in a tungsten plug process with a chemical or mechanical polishing method; and FIGS. 2A to 2C are schematic diagrams of a method for reducing unevenness of an aluminum-copper alloy stack in a metallization process according to the present invention. Example In order to prevent the aluminum-copper alloy layer from reacting with the underlying titanium nitride layer, and reduce the thermal stress on the copper-copper alloy stack, thereby reducing the unevenness on its surface, the main feature of this embodiment is to form aluminum-copper Before the alloy layer, a vacuum breaking process is performed on the titanium nitride layer in advance (exposing the titanium nitride layer to air or treating the titanium nitride layer with a nitrogen oxide plasma) so as to electrically contact the surface of the titanium nitride layer with air or tritium nitride. The osmium reaction of N2plasma forms a thin titanium oxynitride layer. The thin oxynitride layer is used to prevent the titanium nitride layer and the subsequently formed aluminum steel alloy layer from reacting inappropriately due to the high temperature of the curing process. Another feature of the present invention is that the IMP layer and the titanium nitride layer are used to change the (1 U) lattice arrangement on the surface of the copper alloy layer, and a low-stress aluminum-copper alloy layer is formed by using high temperature and low power. Due to the (111) aluminum-copper alloy layer, the metal thickness can also be reduced.
C:\Prograra Files\patent\0503-3890-E.ptd第 6 頁 449829 ^ 87116160 五、發明說明（4) 修正 移能力（Stress migration)，因此亦有助於減輕紹銅合金 堆疊表面的凹凸現象。 請參考第2A〜2C圖，此為本發明在金屬化製程中降低 鋁銅合金堆疊凹凸現象的方法的示意圖。 首先，如第2A圖所示，在半導體基底30(如矽基底）表 面形成一介電層32(如二氧化矽層）’用以形成金屬化製程 中的鋁銅合金堆疊。介電層32是形成鋁銅合金堆疊用的基 底，其亦可以其他材料取代。隨後，在介電層32表面依序 形成内金屬製程鈦層34(IMP-Ti)及氮化鈦層36。在這個步 驟中，内金屬製程鈦層34及氮化鈦層36是隨後形成鋁銅合 金層的黏著層。並且，由於内金屬製程鈦層34及氮化鈦層 36可以改變鋁銅合金層形成表面的晶格排列方式，隨後形 成的鋁鋼合金層亦可以保證為（11 1 )結構。 接著’如第2B圖所示’對氮化鈦層36進行破真空製程 (Vacuum break)。破真空製程可以將半導體基底3〇表面的 氮化鈦層36曝露在空氣中，或利用氧化氮電漿處理（n2〇 plasma treatment)氮化鈦層36表面以達到。在這個實施 例中，破真空製程的主要目的是使半導體基底3〇表面的氮 化敛層36與空氣中或氧化鈦電漿中的氧反應，藉以形成一 薄氧氮化鈦層38(TiN0)。薄氧氮化鈦層38可以用作氮化鈦 層36與隨後形成的鋁銅合金層間的阻障層，藉以避免這兩 層薄膜因固化製程的高溫而產生不當反應。 接著’如第2C圖所示，利用高溫低功率形成的（丨u ) 結構鋁銅合金層40、並在鋁鋼合金層4〇表面形成反反射的C: \ Prograra Files \ patent \ 0503-3890-E.ptd page 6 449829 ^ 87116160 V. Description of the invention (4) Correction of migration (Stress migration), so it also helps to reduce the unevenness of the surface of the copper alloy stack . Please refer to FIGS. 2A to 2C, which are schematic diagrams of a method for reducing unevenness of an aluminum-copper alloy stack in a metallization process according to the present invention. First, as shown in FIG. 2A, a dielectric layer 32 (such as a silicon dioxide layer) 'is formed on the surface of a semiconductor substrate 30 (such as a silicon substrate) to form an aluminum-copper alloy stack in a metallization process. The dielectric layer 32 is a base for forming an aluminum-copper alloy stack, and it may be replaced by other materials. Subsequently, an internal metal process titanium layer 34 (IMP-Ti) and a titanium nitride layer 36 are sequentially formed on the surface of the dielectric layer 32. In this step, the titanium layer 34 and the titanium nitride layer 36 in the inner metal process are adhesion layers for subsequently forming an aluminum-copper alloy layer. In addition, since the titanium layer 34 and the titanium nitride layer 36 in the inner metal process can change the lattice arrangement of the aluminum-copper alloy layer forming surface, the subsequent aluminum-steel alloy layer can be guaranteed to have a (11 1) structure. Next, as shown in FIG. 2B, a vacuum break is performed on the titanium nitride layer 36. The vacuum breaking process can be performed by exposing the titanium nitride layer 36 on the surface of the semiconductor substrate 30 to the air, or by using a nitrogen oxide plasma treatment on the surface of the titanium nitride layer 36. In this embodiment, the main purpose of the vacuum breaking process is to cause the nitrided layer 36 on the surface of the semiconductor substrate 30 to react with oxygen in the air or titanium oxide plasma to form a thin titanium oxynitride layer 38 (TiN0 ). The thin titanium oxynitride layer 38 can be used as a barrier layer between the titanium nitride layer 36 and the subsequently formed aluminum-copper alloy layer, so as to prevent the two layers of films from generating an inappropriate reaction due to the high temperature of the curing process. Next, as shown in FIG. 2C, an aluminum-copper alloy layer 40 having a (丨 u) structure formed by using high temperature and low power is formed, and an anti-reflection layer is formed on the surface of the aluminum-steel alloy layer 40.
4 49 829 五、發明說明（5) 氮化鈦層。如第2A圖的步騍所述，由於内金屬製程鈦層34 輿氮化鈦層3 6可改變鋁銅合金層形成表面的晶格排列方 式’因此’在這個步驟中形成的鋁銅合金層40可確保為 (111)結構。（11 1)結構的鋁銅合金層4 0具有減輕鋁銅合金 堆疊應力的功能，因此可降低鋁銅合金堆疊表面產生的凹 凸現象。另外’為進一步降低鋁銅合金堆疊遷移的熱應力 及凹凸現象，鋁鋼合金層40係以高溫低功率形成。一般而 言’鋁銅合金層40可在400 t溫度及4KW功率下沈積形成。 另外’反反射的氮化鈦層42則可避免因鋁銅合金層40反光 而造成微影製程的錯誤。 在這個實施例中’由於鋁銅合金層4 〇係以高溫低功率 形成’已適度減輕了 IS鋼合金堆疊的熱應力，故隨後進行 的固化製程南溫便不會再對銘銅合金堆疊產生嚴重的熱應 力、並造成嚴重的凹凸現象。 綜上所述’本發明在鎢插塞製程中形成金屬堆疊的方 法中，由於破真空以形成薄氧氮化鈦層，故銘銅合金層與 氮化鈦層的反應可以避免。另外’由於利用内金屬製程鈦 層及氮化鈦層為鋁銅合金層的黏著層，所形成的鋁銅合金 層可具有（111)結構，其可以降低金屬堆疊所承受的熱應 力遷移’進而減少金屬堆疊因隨後固化製程所帶來的強大 熱應力而引起凹凸現象。 雖然本發明已以較佳實施例揭露如上，然其並非用以 限定本發明，任何熟習此技藝者，在不脫離本發明之精神 和範圍内’當可做更動與潤飾，因此本發明之保護範圍當4 49 829 5. Description of the invention (5) Titanium nitride layer. As described in step 2A of FIG. 2, since the titanium layer 34 and the titanium nitride layer 36 in the inner metal process can change the lattice arrangement of the aluminum-copper alloy layer formation surface, “therefore” the aluminum-copper alloy layer formed in this step. 40 can be guaranteed to be (111) structure. (11 1) The structure of the aluminum-copper alloy layer 40 has the function of reducing the stacking stress of the aluminum-copper alloy. Therefore, the concave-convex phenomenon generated on the surface of the aluminum-copper alloy stack can be reduced. In addition, in order to further reduce the thermal stress and unevenness of the aluminum-copper alloy stack migration, the aluminum-steel alloy layer 40 is formed with high temperature and low power. In general, the 'aluminum-copper alloy layer 40 can be formed at a temperature of 400 t and a power of 4 KW. In addition, the 'reflective titanium nitride layer 42 can avoid the lithography process error caused by the light reflection of the aluminum-copper alloy layer 40. In this embodiment, 'as the aluminum-copper alloy layer 40 is formed at a high temperature and low power', the thermal stress of the IS steel alloy stack has been moderately reduced, so the subsequent curing process at South temperature will no longer produce a copper alloy stack. Severe thermal stress and cause severe unevenness. In summary, in the method of forming a metal stack in the tungsten plug process of the present invention, since the vacuum is broken to form a thin titanium oxynitride layer, the reaction between the copper alloy layer and the titanium nitride layer can be avoided. In addition, 'Because the titanium layer and the titanium nitride layer in the inner metal process are used as the adhesion layer of the aluminum-copper alloy layer, the aluminum-copper alloy layer formed can have a (111) structure, which can reduce the thermal stress migration experienced by the metal stack' and further Reduces unevenness caused by metal stacks due to the strong thermal stress caused by subsequent curing processes. Although the present invention has been disclosed in the preferred embodiment as above, it is not intended to limit the present invention. Any person skilled in the art can change and retouch without departing from the spirit and scope of the present invention. Therefore, the protection of the present invention Range when
C:\Program Files\Patent\0503-3890-E.ptd第 8 頁 449829 案號 87116160 月, 條正 曰 修正 五、發明說明（6) 視後附之申請專利範圍所界定者為準 符號說明： 10, 30〜半導體基底 12, 32〜介電層 1 4，3 4〜鈦層 16, 20, 36〜氮化鈦層 18, 40〜鋁銅合金層 3 8〜薄氧氣化鈦層C: \ Program Files \ Patent \ 0503-3890-E.ptd Page 8 449829 Case No. 87116160 Month, Article No. Amendment V. Description of Invention (6) According to the definition of the scope of the attached patent application, the symbol description shall prevail: 10, 30 ~ semiconductor substrate 12, 32 ~ dielectric layer 1 4, 3 4 ~ titanium layer 16, 20, 36 ~ titanium nitride layer 18, 40 ~ aluminum-copper alloy layer 3 8 ~ thin oxygenated titanium layer
0503-3890-EFl.ptc 第9頁0503-3890-EFl.ptc Page 9
II