475217 五、發明說明（l) 發明之領域 本發明係提供一種決定化學機械研磨（Chemical , mechanical p〇i ishing，CMP)製程終點的方法，尤指一種 利用紅外光光譜（infrared spectroscopy)之光學終點決 定方法。 背景說明 在半導體製程中，化學機械研磨（chemical mechanical p〇iishing，CMP)技術是最廣為使用的半導體 晶片平坦化技術，例如CMP技術可用來去除雙層（dual layer)結構中之第一薄膜層。目前評多方法均可用來決定 C Μ P製程終點（e n d p 〇 i n t )，一般最常用的是利用光學原理 進行目標薄膜層（t a r g e t 1 a y e r )的監測。然而利用此方法 偵測CMP終點時，欲去除的目標層需要有足夠的厚度，以 使光偵檢器在CMP過程中所偵測到的反射光強度資料可製 作一軌跡曲線（trace curve)，進而可用來偵測CMP終點。 一般來5兄’目標層的厚度需要超過300 〇埃以上，如此由反 射光強度資料才能產生一執跡曲線。 請參閱圖一，圖一為習知偵測CMP終點之方法示意 圖。如圖一所示，一未經研磨之半導體晶片1 1被置放在晶 座（wafer head)15之支撐部位（holder)13中。研磨塾475217 V. Description of the Invention (l) Field of the Invention The present invention provides a method for determining the end point of a chemical mechanical polishing (CMP) process, especially an optical end point using infrared spectroscopy. Decide the method. BACKGROUND In the semiconductor process, chemical mechanical polishing (CMP) technology is the most widely used semiconductor wafer planarization technology. For example, CMP technology can be used to remove the first thin film in a dual layer structure. Floor. At present, many methods can be used to determine the end point of the CMP process (e n d p o int). Generally, the most commonly used method is to monitor the target thin film layer (t a r g e t 1 a y e r) using optical principles. However, when using this method to detect the CMP end point, the target layer to be removed needs to have sufficient thickness, so that the reflected light intensity data detected by the photodetector during the CMP process can create a trace curve. It can then be used to detect CMP endpoints. Generally, the thickness of the target layer needs to be more than 300 angstroms, so that a reflection curve can be generated from the reflected light intensity data. Please refer to Figure 1. Figure 1 is a schematic diagram of a conventional method for detecting the CMP endpoint. As shown in Fig. 1, an unpolished semiconductor wafer 11 is placed in a holder 13 of a wafer head 15. Grinding
475217 五、發明說明（2) --- (polishing pad) 12,位於晶片u下方，由一研磨平台 (p 1 aten) 16所支撐。研磨平台16具有一窗口（未顯示^可穿 過研磨墊12與研磨平台16至半導體晶片丨丨之目標層表面。 一馬達（motor) 1 9用來驅動晶座1 5與研磨平台丨6而一控制 器（control ler) 18控制兩者之旋轉速度。一垂直馬達 (verticalhmotor)20置入以使晶座Η與研磨塾Μ可做垂直 接觸。.此外，CMP製程設備尚包含一研磨液供應管（slurry s u p p 1 i e r t u b e ) 1 4 ’用來供應晶片1 1與研磨墊1 2間的研磨 液（slurry)0 根據X Μ P製程’晶座1 5與研磨平台1 6均可以一特定速 岸各自旋轉，以使研磨液均勻分佈於研磨墊丨2之上。經由 调整係數設定’晶片1 1之目標層可藉由與研磨液間之化學 反應以及與研磨墊1 2間之機械研磨而被磨成平坦。習知 CMP之終點偵測系統係利用電腦2 1處理目標層之反射光 線，製作出執跡曲線2 2以決定研磨終點。細言之，習知之 CMP設備包含了一光學偵測裝置1 7，可產生固α定波數的光 線。此光線經由研磨墊，1 2上的洞口以一預定傾斜角度照射 晶片1 1上之目標層’其反射光強度可藉由光偵測裝置1 7持 續偵測。接著，資料被傳回控制器1 8以及電腦2 1，最後並 在電腦螢幕上顯示成軌跡曲線2 2。在執跡曲線2 2上之反射 光強度I的明顯改變的時段内，利用一預先設定的視窗邏 輯（w i n d 〇 w 1 〇 g丨c ) 5 1與5 2以決定一 c Μ Ρ製程之研磨終點。475217 V. Description of the invention (2) --- (polishing pad) 12, is located under the wafer u, and is supported by a polishing platform (p 1 aten) 16. The polishing table 16 has a window (not shown) which can pass through the polishing pad 12 and the polishing table 16 to the surface of the target layer of the semiconductor wafer. A motor 1 9 is used to drive the wafer 15 and the polishing table 6 and A controller (controller) 18 controls the rotation speed of the two. A vertical motor (verticalhmotor) 20 is placed so that the wafer holder Η and the grinding 塾 can make vertical contact. In addition, the CMP process equipment also includes a grinding fluid supply The tube (slurry supp 1 iertube) 1 4 'is used to supply the polishing liquid (slurry) between the wafer 1 1 and the polishing pad 12 2 According to the XP process, the crystal base 1 5 and the polishing platform 16 can have a specific speed. Rotate each to uniformly distribute the polishing liquid on the polishing pad. 2. The target layer of the wafer 1 1 can be set by the adjustment coefficient through chemical reaction with the polishing liquid and mechanical polishing with the polishing pad 12. Grinding is flat. The conventional CMP end-point detection system uses a computer 21 to process the reflected light from the target layer to make a tracking curve 22 to determine the end point of polishing. In detail, the conventional CMP equipment includes an optical detection Device 1 7 can produce Light with a fixed alpha wave number. This light passes through the polishing pad, and the hole on 12 illuminates the target layer on the wafer 11 at a predetermined tilt angle. The reflected light intensity can be continuously detected by the light detection device 17. Then, the data is transmitted back to the controller 18 and the computer 21, and finally displayed on the computer screen as a trajectory curve 22. During a period of significant change in the reflected light intensity I on the trajectory curve 22, a pre- The set window logic (wind 〇w 1 〇g 丨 c) 5 1 and 5 2 to determine the grinding end point of a c MP process.
第5頁 475217 五、發明說明（3) 然而，習知用來偵測CMP製程終點的方法需要有厚度 大於3 0 0 0埃之目標薄膜層，如此才能使利用電腦製作的執 跡曲線可適合偵測CMP製程終點。 發明概述 ， 本發明之主要目的在提供一種新穎的CMP製程終點偵 測方法，毋需限定所需目標薄膜層之厚度，以解決上述問 題。 在本發明之最佳實施例中，本發明係提供一種利用紅 外光光譜決定化學機械研磨製程終點的方法，用來去除半 導體晶片上之第一薄膜層而不致造成下方之第二薄膜層過 度研磨。該方法首先將一紅外光光源導向半導體晶片，收 集每一薄膜層之紅外光吸收度之資料以產生半導體晶片上 每一薄膜層的標準吸收曲線。由於每一薄膜層吸收之紅外 光的波數（wavenumber)不同，藉由CMP製程進行時，紅外 光吸收度改變而可偵測到每一層之吸收曲線而可觀察到兩 定義（d e f i n e )的紅外光吸收曲線。經過一段時間後，該第 一薄膜層的紅外光吸收度漸漸的降低，直到吸收曲線偵測 不出有明顯變化。此時CMP製程的終點即在於所偵測到的 第一薄膜層之紅外光吸收有不再有顯著變化而第二薄膜層 的紅外光吸收開始變化時。Page 5 475217 V. Description of the invention (3) However, the conventional method for detecting the end point of the CMP process requires a target thin film layer with a thickness greater than 300 angstroms, so that the computer-made track curve can be adapted. Detect end of CMP process. SUMMARY OF THE INVENTION The main object of the present invention is to provide a novel method for detecting the end point of a CMP process without the need to limit the thickness of a desired target film layer in order to solve the above problems. In a preferred embodiment of the present invention, the present invention provides a method for determining the end point of a chemical mechanical polishing process by using infrared light spectrum to remove a first thin film layer on a semiconductor wafer without causing excessive grinding of the second thin film layer below. . In this method, an infrared light source is first directed to a semiconductor wafer, and data of infrared light absorption of each thin film layer is collected to generate a standard absorption curve of each thin film layer on the semiconductor wafer. Because the wave number of the infrared light absorbed by each thin film layer is different, when the CMP process is performed, the infrared light absorbance is changed, the absorption curve of each layer can be detected, and two defined infrared rays can be observed. Light absorption curve. After a period of time, the infrared light absorption of the first thin film layer gradually decreases until the absorption curve does not detect a significant change. At this point, the CMP process ends when the detected infrared light absorption of the first thin film layer no longer changes significantly and the infrared light absorption of the second thin film layer starts to change.
第6頁 475217 五、發明說明（4) 本發明係利用簡易精確之CMP製程終點之紅外光光譜 偵測方法，以去除第一薄膜層並暴露出位於下方第二薄膜 層，且不會造成過度研磨。 ，、 發明之詳細說明 ， 請參照圖二，圖二為本發明偵測CMP製程終點之方法 示意圖。如圖二所示，一半導體晶片30被置放在支撐物23 之中，且由晶圓載座2 5固定位置。半導體晶片3 0位於研磨 平台2 6上方之研磨墊24上，且包含有一第一薄膜層31與第 二薄膜層2 9。第一薄膜層3 1與第二薄膜層2 9係分別為一矽 氧層和一氮石夕層。兩者均可穿透過紅外光且吸收不同波數 之紅外光。本發明首先將一紅外光米源2 8朝向一窗口 2 7而 使紅外光光源2 8發出之紅外光通過研磨平台2 6、研磨墊 2 4、半導體晶片3 0、晶座25，最後到達偵檢器3 2。在CMP 進行的過程中，偵檢器3 2利用所收集的資料以製作成一圖 表，該圖表可顯示位於研磨墊2 4上的半導體晶片3 0之矽氧 層與氮矽層在每一時段紅外光吸收度。因此，用來顯示第 一薄膜層3 1與第二薄膜層2 9之吸收度變化的圖表之是固定 以毫秒（m i 1 1 i s e c 〇 n d )作為單位時間。 請參閱圖三，圖三為本發明在CMP製程進行中之矽氧 層與氮矽層之紅外光吸收度與波長之關係圖。如圖三所 示，吸收曲線A代表C Μ P製程在剛開始進行時的吸收情形，Page 6 475217 V. Description of the invention (4) The present invention uses a simple and accurate infrared light spectrum detection method at the end of the CMP process to remove the first thin film layer and expose the second thin film layer below, without causing excessive Grinding. For a detailed description of the invention, please refer to FIG. 2. FIG. 2 is a schematic diagram of a method for detecting an end point of a CMP process according to the present invention. As shown in FIG. 2, a semiconductor wafer 30 is placed in a support 23 and fixed by a wafer carrier 25. The semiconductor wafer 30 is located on the polishing pad 24 above the polishing table 26, and includes a first thin film layer 31 and a second thin film layer 29. The first thin film layer 31 and the second thin film layer 29 are a silicon oxide layer and a nitrogen oxide layer, respectively. Both can penetrate infrared light and absorb infrared light of different wave numbers. In the present invention, an infrared light source 28 is directed toward a window 27, and the infrared light emitted by the infrared light source 28 is passed through the polishing platform 26, the polishing pad 24, the semiconductor wafer 30, and the crystal seat 25, and finally reaches the detection测 器 3 2. During the CMP, the detector 32 uses the collected data to make a chart, which can show the silicon oxide layer and nitrogen silicon layer of the semiconductor wafer 30 on the polishing pad 24 at each time period. Light absorption. Therefore, the graph used to show the change in the absorbance of the first thin film layer 31 and the second thin film layer 29 is fixed with milliseconds (m i 1 1 s e c ○ n d) as the unit time. Please refer to FIG. 3. FIG. 3 is a graph showing the relationship between the infrared light absorbance and the wavelength of the silicon-oxygen layer and the silicon-nitrogen layer during the CMP process of the present invention. As shown in Figure 3, the absorption curve A represents the absorption of the CMP process at the beginning.
第7頁 475217 五、發明說明（5) 其包含有兩個尖峰（peak)l、2。每一尖峰代表個別薄膜層 的吸收度。舉例來說，由矽氧層構成的第一薄膜層，其吸 收尖峰2可在波數1 1 0 0 - 1 0 0 0 cm-1間偵測出來，其吸收值約 1. 5。而由氮矽層所構成之第二層之吸收尖峰1則可在 8 5 0 - 7 5 0 cm-1間偵測出來，其吸收值為0. 3。 然而隨著CMP製程的進行，如圖三所示，吸收曲線開 始下降至曲線B、C。在某一時間點時，吸收曲線C的尖峰2 明顯不再有任何下降情形產生。當代表矽氧層吸收之尖峰 2偵測不出明顯變化且代表氮矽層吸收之尖峰1剛開始下降 時即為CMP製程終點。矽氧層之吸收尖峰2缺乏顯著改變與 氮矽層之吸收尖峰1剛開始下降顯示CMP製程終點。由於尖 峰1由曲線B到C有一明顯下降情形，因此可能發生過度研 磨，故CMP終點通常被決定在曲線B與C之間。 其中，第一薄膜層3 1為矽氧層而第二薄膜層2 9為氮矽 層，亦使用第一薄膜層3 1為氮矽層而第二薄膜層2 9為矽氧 層取代，此時之C Μ P製程終點決定於當氮矽層之紅外光吸 收尖峰不再有明顯變化且矽氧層之紅外光吸收尖峰開始下 降時。 同時，本發明之方法亦可應用於淺溝隔離（shal low trench isolation, STI)之CMP製程中，用來去除由石夕氧 層構成之介電層，而可使位於介電層下方且由氮矽層構成Page 7 475217 V. Description of the invention (5) It contains two peaks 1 and 2. Each spike represents the absorbance of an individual film layer. For example, the absorption peak 2 of the first thin-film layer composed of a silicon oxide layer can be detected at a wave number 1 1 0-1 0 0 0 cm-1, and its absorption value is about 1.5. The absorption peak 1 of the second layer formed by the nitrogen-silicon layer can be detected between 850-750 cm-1, and the absorption value is 0.3. However, as the CMP process progresses, as shown in Figure 3, the absorption curve begins to fall to curves B and C. At a certain point in time, it is clear that the peak 2 of the absorption curve C no longer has any decline. When the peak 2 representing the absorption of the silicon-oxygen layer cannot detect a significant change, and the peak 1 representing the absorption of the silicon-silicon layer has just started to fall, it is the end of the CMP process. The lack of significant changes in the absorption peak 2 of the silicon oxide layer and the absorption peak 1 of the nitrogen silicon layer just began to decline, indicating the end of the CMP process. Because peak 1 has a significant decrease from curve B to C, over-grinding may occur, so the end point of CMP is usually determined between curves B and C. The first thin film layer 31 is a silicon oxide layer and the second thin film layer 29 is a silicon nitrogen layer. The first thin film layer 31 is a silicon nitrogen layer and the second thin film layer 29 is a silicon oxide layer. The end of the CMP process at that time is determined when the infrared light absorption peaks of the nitrogen-silicon layer no longer change significantly and the infrared light absorption peaks of the silicon-oxygen layer begin to decrease. At the same time, the method of the present invention can also be applied to the CMP process of shallow low trench isolation (STI) to remove a dielectric layer composed of a silicon oxide layer, and can be located below the dielectric layer and composed of Nitrogen silicon layer composition
第8頁 475217 五、發明說明（6) 的終止層（s t ο p 1 a y e r )暴露出來。淺溝隔離之C Μ P製程終 點即決定於介電層之紅外光吸收尖峰不再有明顯變化，而 終止層之紅外光吸收尖峰開始下降時。 1 與習知技術相較，本發明提供，一種有效且簡單的CMP 製程終點偵、測方法。藉由使用紅外光吸收度的偵測以去除 第一薄膜層，而不致過度研磨到位於下方的第二薄膜層。 同時，習知用來決定CMP製程終點之技術需有一需超過 3 0 0 0埃的第一薄膜層，因為厚度小於3 0 0 0埃之薄膜層無法 產生用來進行終點偵測的執跡曲線。然而，本發明之方法 不需限定第一薄膜層之厚度。 以上所述僅為本發明之較佳實施例，凡依本發明申請 專利範圍所做之均等變化與修飾，皆應屬本發明專利之涵 蓋範圍。Page 8 475217 5. The termination layer (s t ο p 1 a y e r) of the description of the invention (6) is exposed. The end of the CMP process for shallow trench isolation is determined by the fact that the infrared light absorption peaks of the dielectric layer no longer change significantly, and the infrared light absorption peaks of the termination layer begin to decrease. 1 Compared with the conventional technology, the present invention provides an effective and simple method for detecting and measuring the endpoint of a CMP process. The detection of infrared light absorption is used to remove the first thin film layer without excessive grinding to the second thin film layer located below. At the same time, the conventional technique used to determine the end point of the CMP process requires a first thin film layer that needs to exceed 300 angstroms, because a thin film layer with a thickness of less than 300 angstroms cannot generate a track curve for endpoint detection. . However, the method of the present invention does not need to limit the thickness of the first thin film layer. The above description is only a preferred embodiment of the present invention, and all equivalent changes and modifications made in accordance with the scope of the patent application for the present invention shall fall within the scope of the invention patent.
第9頁 475217 圖式簡單說明 圖示之簡單說明 終點之方法示意圖。 程終點之方法示意圖。 程進行時矽氧層與氮矽層之 圖一為習知偵測CMP製程 圖二為本發明偵測CMP製 圖三為本發明經由CMP製 I R吸收度幾波長之關係圖。 圖示之符號說明 10 CMP機台 12' 24 研磨塾 14 研磨液供應管 16> 26 研磨平台 18 控制器 20 垂直馬達 22 軌跡曲線 28 紅外光光源 31 第一薄膜層_ 5卜 52 視窗邏輯 1 1 > 30 半 導 體 晶 片 13、 23 支 撐 物 15' 25 晶 座 17 光 偵 測 裝 置 19 馬 達 21 電 腦 27 窗 α 29 第 二 薄 膜 層 32 偵 檢 器Page 9 475217 Simple illustration of the diagram Simple illustration of the diagram The method of the end point. Schematic illustration of the method for the end of the process. The silicon oxide layer and silicon silicon layer during the process are shown in Figure 1. Figure 1 shows the conventional detection CMP process. Figure 2 shows the detection CMP process of the present invention. Figure 3 shows the relationship between IR absorption and wavelength of the CMP process according to the present invention. Explanation of symbols in the figure 10 CMP machine 12 '24 Grinding 塾 14 Grinding liquid supply tube 16> 26 Grinding platform 18 Controller 20 Vertical motor 22 Trajectory curve 28 Infrared light source 31 First thin film layer 5 5 52 Window logic 1 1 > 30 semiconductor wafer 13, 23 support 15 '25 wafer base 17 light detection device 19 motor 21 computer 27 window α 29 second thin film layer 32 detector
第10頁Page 10