200400098 玖、發明說明 (發明說明應敘明：發明所屬之技術領域、先前技術、内容、實施方式及圖式簡單說明） 【發明戶斤屬之技術領域3 相關申請案之交叉參考 本申請案係為於2002年12月17曰提出申請之美國專利 5 申請案序號為第10/321，150號@丁-280-1；3)，於2002年3月 22曰提出申請之美國專利申請案序號為第10/105,016號 (NT-250-US)，於2002年7月15日提出申請之美國專利申請 案序號為第10/197,090號（NT-248-US)，以及於2002年1月 17曰提出申請之美國專利申請案序號為第10/052,475號 10 (NT_238-US)之部分接續申請案，於此該所有申請案併入 本案以為參考資料。 本申請案係請求對於在2002年12月27曰提出申請之美 國臨時申請案第60/436,706號（NT-278-P4)，在2002年12月 23曰提出申請之美國臨時申請案第60/436，108號（NT-278-15 P3)，在2002年10月10曰提出申請之美國臨時申請案第 60/417,544號（NT-278-P2)，在2002年9月27日提出申請之 美國臨時申請案第60/415,579號（NT-278-P)，在2002年7月 19曰提出申請之美國臨時申請案第60/397,110號（NT-273-P)，在2002年3月12曰提出申請之美國臨時申請案第 20 60/365,016號（NT-249-P)之優先權，於此該所有申請案併 入本案以為參考資料。 技術領域 本發明係有關於半導體積體電路之製造，更特定言之 ，係有關於一種傳導及絕緣層之化學機械拋光的方法。 200400098 玖、發明說明 【先前技術2 發明背景 傳統式半導體元件一般係包括一半導體基板，通常係 為一矽基板，以及複數之相繼構成的介電層間（dielectric 5 interlayers)，諸如二氧化矽，以及以傳導性材料所製成的 傳導性路控或互連（interconnections)。近來由於銅與銅合 金其之優越的電移（electromigration)以及低電阻率的特性 ’已受到相當的關注使用作為互連材料。互連通常係藉由 將餘刻的特徵或腔室中的銅，藉由一金屬化製程填注進入 10介電層間所構成。銅金屬化製程之較佳的方法係為電鍍。 於一積體電路中，多層的互連網路係相關於基板表面橫向 地L伸。構成在連續層中的互連能夠使用通孔（vias)或接 點電氣地連接。於一典型的製程中，首先在半導體基板上 構成一絕緣層。執行圖案化及蝕刻製程用以在絕緣層中構 成諸如溝或通孔的特徵。在以一阻障層以及之後 的一種晶層（seed layer)塗佈位在表面上的特徵之後，將銅 電鍍填注特徵。然而，電鍍製程，除了填注該等特徵外， 同時在基板之頂部表面上導致一銅層。此過量的銅係為所 月的超載（overburden)，並應在接續的加工步驟之前去除。 2〇 第1八圖係顯示該經電鍍基板9，例如一矽晶圓，之一 範的邛为8。應注意的是基板9可包括元件或是其他的金 屬與半導體部分，為了說明於第1A圖中並未顯示。如第 1A圖中两f , ’、=如一通孔10以及一溝13的特徵係構成在一 絕緣層14中，士妻l —产 堵如二氧化矽層，其係構成在基板9上。通 200400098 玖、發明說明 孔與溝13以及絕緣層14之頂部表面15係經由一電鍍加工以 一沉積的銅層16加以覆蓋並填注。傳統地，在圖案化與餘 刻之後，絕緣層14首先以一阻障層18加以塗佈，典型地， 係為一组（Ta)或是紐/氮化组（TaN)合成層。以阻障層丨8塗 5佈通孔與溝以及絕緣層之表面15，用以確保良好的黏著性 並使用作為一阻障材料，防止銅擴散進入半導體元件以及 絕緣層中。再者，一種晶層（未顯示），通常係為一銅層， 係沉積在阻障層上。在接續之銅沉積期間，種晶層構成一 傳導性材料基底用於生成銅薄膜。當電鍍銅薄膜時，沉積 10之銅層16快速地填注通孔10，但以保角（c〇nf〇rmal)的方式 塗佈寬溝13以及頂部表面15。當持續沉積加工用以確保溝 同時受到填注時’在基板9上構成一銅層或超載層 (overburden)。傳統地，在銅電鍍之後，能夠使用不同的 材料去除加工，例如，化學機械拋光（CMp)、蝕刻或電蝕 15 刻（electroetching)，將非所欲的超載層去除。 化學機械拋光（CMP)加工傳統地包括將一半導體晶圓 或其他的該基板，壓按靠著一具有拋光漿液之濕的移動拋 光表面。該漿液可為鹼性、中性或是酸性，並且一般係含 有氧化鋁、氧化鈽（ceria)、氧化矽或其他之硬質研磨陶瓷 20顆粒。拋光的表面典型地係為一平面墊，並以在化學機械 拋光（CMP)技藝中所廣為熟知的聚合材料所製成。一些拋 光墊包含研磨顆粒（固定的研磨墊）。該等墊係可與化學機 械拋光（CMP)溶液結合使用，該溶液可不包含任何研磨顆 粒。假若墊係為多孔的，則拋光漿液或溶液可輸送至墊之 200400098 玖、發明說明 表面，或可流經該墊至其之表面。於化學機械拋光（CMP) 加工期間，一晶圓載具支撐著一待加工之晶圓並將晶圓表 面放置在一化學機械拋光（CMP)墊上，並以一受控制之壓 力將晶圓壓按靠在墊上同時轉動該墊。該墊同時係可構形 5為一線性拋光帶，其能夠如一線性帶橫向地移動。藉由靠 著該墊移動晶圓，靠著晶圓而移動該墊或是以上該二移動 方式，當將拋光漿液供應至墊與晶圓表面間之界面時執行 加工。 如第1B圖中所示，首先應用化學機械拋光（CMP)降低 10覆蓋在絕緣層丨4之頂部表面15上之銅層厚度至阻障層18。 接著’去除位在頂部表面上的阻障層18用以將銅與阻障層 限定在通孔10、12以及溝13中。然而，於該等加工期間， 轉疋拋光之終點，無論是將銅層向下拋光至阻障層或是將 阻障層向下拋光至絕緣層，在工業上係為一重要的問題。 15 美國專利第5,6〇5,760號中說明一種拋光墊其係以一固 體均勻的聚合物薄片所製成。該聚合物薄片在一特定的波 長範圍下係為透光的。聚合物薄片的表面不包含任一研磨 材料，並不具任何吸收或是運送漿液顆粒之固有的能力。 近來，終點檢測系統已利用具有一視窗或複數視窗之 20轉動墊或是線性帶系統施行。於該等狀況下，當墊或是帶 移動時’其係在就地（in_situ)之監視器上方通過進行自晶 圓表面測量反射係數（reflectance)。反射上的變化顯示出 拋光加工的終點。然而，在拋光墊中開啟的視窗能夠使拋 光加工變得複雜，並且擾亂了墊或帶之均質性。此外，該 10 200400098 玖、發明說明 等視窗可致使聚積拋光加工之副產品與漿液。 因此’當基板係使用化學機械拋光（CMP)加工拋光時 ’持、’’貝地為求對於能夠精確並有效地檢測在基板上的一終 點的方法與裝置。 5 如第1B圖中所示，首先應用化學機械拋光（CMP)降低 覆蓋絕緣層14之頂部表面15之銅層厚度至阻障層18。接著 ’去除位在頂部表面上的阻障層丨8用以將銅與其餘之阻障 層限定在通孔10以及溝13中。然而，於該等加工期間，均 勻地降低經拋光銅層之厚度在工業上係為一重要的問題。 10在將其加工時必需維持金屬層厚度的均勻一致性，因此將 在銅終點之後的過度拋光降至最低且基板不致過度拋光， 因為過度拋光會導致過度地碟形下陷（dishing)、凹陷 (erosion)以及其他的缺陷。再者，銅層與阻障層之拋光不 足會致使電氣短路或是其他的缺陷。於拋光加工期間的不 15 一致性’可因一不均句一致的拋光加工抑或是位在基板上 金屬層之一不均勻一致的厚度或是以上該二狀況所造成。 將一基板之絕緣層拋光係為另一化學機械拋光（CMp) 的應用。淺溝隔離（STI)係為一加工方式，其係藉由在基板 之表面中構成絕緣溝用以防止在相鄰電路間的電移。該等 20溝典型地係以氮化矽（SigN4)與二氧化矽（Si〇2)填注。為填 注該等溝，首先將一氮化矽層沉積在基板之表面上，接著 覆以二氧化矽覆蓋層。必需自基板之表面去除過多的二氧 化矽與氮化矽，留下一平滑的氮化矽層覆蓋著大部分之基 板表面以及填注溝區域之氧化矽與氮化矽層。典型地藉由 200400098 玖、發明說明 化學機械拋光（CMP)執行將過多的二氧化矽與氮化矽去除。 第1C圖係顯示一基板52，例如一矽晶圓，之一示範的 部分51的橫截面視圖，其係覆以二絕緣材料層。一適於淺 溝隔離（STI)的溝53，係構成在基板52之表面中。一底部絕 5緣層54以及一頂部絕緣層55覆蓋著包括溝53之基板52之表 面。底部絕緣層54以及頂部絕緣層55，例如，係可分別為 氮化矽與二氧化矽。應注意的是絕緣層54及55覆蓋基板％ 的正個表面。為凡成淺溝隔離（STI)加工，必需去除過多的 絕緣材料。 10 15 20 第1D圖係為在絕緣層54及55已拋光至一所欲程度之後 基板52之示範的部分51的橫截面視圖，亦即已將過多的 絕緣材料去除之後。例如，係可藉由化學機械拋光（cMp) 執行絕緣層之拋光。應注意的是絕緣層54之一平滑層，亦 即見化梦覆蓋基板52之表面，以及絕緣層54及55(亦即氮 化矽與二氧化矽）填注溝53。 目别淺溝隔離（STI)技術有關的問題包括藉由光學干涉 计執仃測量二氧化矽厚度的困難性，因為厚度測量信號本 身係週期性地隨著二氧切厚度的增加或降低而重複。此 外、’厚度測量信號對於環境因素，諸如水分（水薄膜）以及 檢測角，係為敏感的。 /兩，於目則技術的一附加的問題在於傳統式度量衡工具 係而自其之載具頭將基板移開用以執行終點檢測。 一均句一致的拋光加 (CMP)成本，同時增加加 工係可顯著地降低化學機械拋光 工的生產率。當晶圓尺寸變得較 12 200400098 玖、發明說明 大時，例如，300公釐及更大，由於晶圓的表面積較大所 以以一均勻一致的方式將厚度平面地降低變得更為困難。 因此，需要一經改良的方法與裝置用於在利用化學機 械拋光（CMP)加工拋光基板時監測並維持經拋光層的一致 5 性。 【發明内容】 發明概要 本發明有利地提供一種在材料去除加工中，諸如化學 機械拋光（CMP)，用於控制平面性的拋光方法與裝置。本 10發明之一具體實施例包括在該一材料去除加工中執行終點 檢測的此力。另一具體實施例提供一種連同一壓力控制技 術的智慧型終點檢測，其係可選擇地將拋光壓力施加在一 工作件上的特定區域。 根據本發明之一觀點提出一種化學機械拋光（CMP)裝 15置一用於拋光一工作件之一表面以及用於檢測-化學機械 拋光（CMP)終點。該化學機械拋光（^^巧裝置包括一透光 的拋光構件、一工作件支架、一支撐板、以及一光學檢測 系統。抛光構件，例如，可為一抛光帶、一拋光塾、或是 聖式的拋光構件。拋光構件，較佳地包括研磨顆粒， 2〇拋光工作件的表面並係可於一或更多的方向上（較佳地為 線性方向，但同時亦能夠於其他的方向，例如，環形）移 作件支木支撐著工作件並係構形靠著拋光構件壓按 乍件工作件支架，例如，係可為-晶圓載具頭或是其 他的結構用於支撐晶圓。支撐板係設計用以當靠著拋光構 13 200400098 玖、發明說明 件壓按工作件時，支撐拋光構件。支撐板，例如，係可為 -平臺或是其他的支撐結構。光學檢測系統檢測化學機械 拋光（CMP)終點，並係配置在拋光構件的下方。光學檢測 系統包括-光源與-檢測器。光源傳送輸出信號通過支撐 5板以及拋光構件至工作件的表面。檢測器接收來自工作件 的表面經由拋光構件與支撐板進入的反射信號。 根據本發明之另一觀點提出一種將一工作件之一表面 拋光以及檢測一化學機械拋光（CMp)終點的方法。根據本 方法，將工作件壓按靠著一透光的抛光構件。㈣構件係 10藉由一支撐板而受到支撐。工作件之表面係以拋光構件加 以拋光。拋光構件係可在一或更多的線性方向上移動。輸 出之光學信號係從一《源輸送通較撐板以及抛光構件至 工作件之表面。光源係配置在拋光構件之下方，因此拋光 構件係介於光源與工作件之表面間。進入的反射光學信號 15係自工作件之表面通過拋光構件以及支撐板而由檢測器接 收。檢測器係配置在拋光構件的下方。 根據本發明之一進一步的觀點提出一種將一或更多工 作件拋光以及提供化學機械拋光（CMP)終點檢測的方法。 根據本方法，-透光的拋光構件係配置在一供應區域與一 0接又區域之間。拋光構件具有一第一端部與一第二端部， 以及-拋光側邊和一背部。第一端部係最初脫離該供應區 域並與接受區域連接，以及第二端部維持與接受區域連接 。一第一卫作件係藉由在-拋光區域中於-或更多的線性 方向上移動抛光構件的一部分而受抛光。使用一光學檢測 14 200400098 玖、發明說明 系統檢測第一工作件之一第一化學機械拋光（CMP)終點。 光學檢測系統輸送輸出信號至第一工作件，並經由拋光構 件接收來自第一工作件的進入反射信號。拋光構件係位在 光學檢測系統與第一工作件之間。 5 根據本發明之另一觀點提出一種將一工作件之一表面 拋光以及用於檢測一化學機械拋光（CMP)終點的化學機械 拋光（CMP)裝置。化學機械拋光（CMP)裝置包括一供應捲 筒以及一接受捲筒、一透光的拋光構件、一加工區域、一 構件用於在一或更多的線性方向上移動拋光構件的一部分 10 、以及一構件用於檢測一化學機械拋光（CMP)終點。拋光 構件具有二端部。一端部係附裝至供應捲筒以及另一端部 係附裝至接受捲筒。加工區域在二端部之間具有拋光構件 的一部分。用於檢測化學機械拋光（CMP)終點的構件輸送 光學信號至工作件的表面，並經由拋光構件接收來自工作 15 件表面的反射光學信號。拋光構件係位在用於檢測的構件 與工作件之間。 根據本發明之一進一步的觀點提出一種將一工作件的 一表面拋光以及檢測一化學機械拋光（CMP)終點的方法。 根據該方法，在一加工區域中該工作件係受支撐致使工作 20 件的表面係暴露至一透光的拋光構件的一部分。藉由線性 雙方向地移動拋光構件之部分，而拋光晶圓的表面。針對 工作件藉由輸送輸出之光學信號通過拋光構件至工作件， 並持續地檢查從工作件反射並經由拋光構件接收的進入光 學信號的相對強度，而確定一化學機械拋光（CMP)終點。 15 200400098 玖、發明說明 僅藉由導論提供先前對於本發明之觀點的說明。於此部分 中並未對以下定出本發明之範料申料鄉圍造成限制。刀 本發明之-第二示範的具體實施例包括一抛光站係具 有-工作件支架，以及-可撓曲的抛光構件。抛光構件係 5靠著藉由-平臺而受支撐之工作件，對著抛光構件之背部 供應-流體。該平臺包括複數之孔用於供應流體，並同時 包括複數之感應器能夠檢測工作件加工之終點。將該等孔 聚集在-起用以產生壓力區域並且典型地每一區域係與一 感應器結合，但可與更多或較少的感應器結合。一電腦接 10收感應器信號並控制流體流動而使抛光最佳化。例如，假 若工作件上的一特定位置抵達終點，則電腦降低流體流動 至4位置同時維持流體流動至其他區域。 於本發明之另一示範的具體實施例中，用於檢測一多 層半導體晶圓之一加工終點的一種感測裝置，其係包括一 15光源對著半導體晶圓之一表面發射光線、一色彩感應器用 以在反應入射光線後自半導體晶圓之表面感測反射的色彩 ，並用以產生一感應器信號、以及一判定電路（decision circuit)與色彩感應器結合並構形用以至少部分地基於感應 器信號判定是否已抵達晶圓加工終點。 20 於本發明之另一示範的具體實施例中，一種用於檢測 一半導體晶圓之一加工終點的終點檢測系統，其包括一感 測裝置係構形用以感測與半導體晶圓之一表面相關的一尺 度（metric)並基於該尺度產生一感應器信號。終點檢測系 統同時包括一判定電路係與感測裝置結合，並構形用以至 16 200400098 玖、發明說明 少部分地基於感應器信號判定是否已抵達晶圓加工終點、 以及一可移動的結構係與該感測裝置結合用以將感測裝置 定位以感測該尺度。 於本發明之另一示範的具體實施例中，一種用於檢測 5 一多層半導體晶圓之一加工終點的方法，其係包括對著半 導體晶圓之-表面發射光線、在反應入射光線後感測自半 導體晶圓之表面反射的色彩，並基於所感測之反射色彩產 生一感應器信號、以及至少部分地基於感應器信號確定是 否已抵達晶圓 力口工終點〇 10 於本發明之一觀點中，流體控制器獨自地控制流體流 動至壓力區域。此觀點之一特性在於本發明同時能夠選擇 地從位在平臺中的特定孔排放流體，用以降低，並且甚至 消極地影響該壓力區域。 於本發明之另一觀點中，在加工期間轉動工作件以及 15該等平臺孔係同中心地配置，並且每一同中心環件係代表 一壓力區域。 於本發明之另一觀點中，流體控制器獨自地控制流體 流動至位在平台上的該等同中心環件。 於本發明之另一觀點中，該拋光構件係為透光的。 20 於本發明之另一觀點中，該拋光構件係包括視窗。 於本發明之另一觀點中，該感應器係為光感應器。 於本發明之另一觀點中，該感應器係為聲學的厚度感 應器（acoustic thickness sensors) 〇 於本發明之另一觀點中，該感應器係為色彩感應器。 17 200400098 玖、發明說明 於本發明之另一觀點中， 的結構。 該感應器係附裝至一可移動 於本發明之另一觀點中， (fiber optic threads)。 於本發明之另一觀點中， 该感應器係使用光纖線 該工作件實質上係保持不動 但於拋光加οι期間係可轉動並平移地移動。於本發明之 較佳的觀點中，該平移的移動係小於_壓力區域面積。 本务明之優點包括理想地抛光工作件的能力，從而節 10 15 省時間與金錢。 圖式簡單說明 本發明的前述與其他之特性、觀點與優點當結合下列 圖式閱讀由以下的詳細說明將變得更為顯而易見的，其中： 第1A圖係為一示範的基板在將材料沉積在基板表面上 之後的橫截面視圖； 第1B圖係為第1 a圖之示範的基板在接受一傳統式的 化學機械拋光（CMP)加工之後的橫截面視圖； 第1C圖係為一示範的基板在將絕緣材料沉積在基板表 面之上後的橫截面視圖； 第1D圖係為第1C圖之示範的基板在接受一傳統式的 化學機械拋光（CMP)加工之後的橫截面視圖； 第2圖係為一示範的CNU系統的橫截面側視圖，該系 統包括一本較佳具體實施例之一用於加工諸如晶圓的工作 件的示範終點檢測系統； 第3圖係為第4圖之示範的化學機械拋光（CMP)系統以 18 200400098 坎、發明說明 及根據本發明之觀點供終點檢測系統所用的示範控制系 統的橫截面俯視圖； 第4圖係為包括第2圖之示範終點檢測系統之示範的化 學機械拋光（CMP)系統的橫截面側視圖； 第5A-C圖係為一工作件表面的視圖； 第6 A圖係圖示本發明之一具體實施例的一種工作件加 工糸統； 第6B圖係圖示本發明之另一具體實施例的一種工作件 加工系統； 〇 第6C圖係圖示本發明之另一具體實施例的一種工作件 加工系統； 第7A-B圖係圖示於第6A_6B圖中本發明之一具體實施 例之平臺； 第8圖係為本發明之一具體實施例之一感應器的分解 15視圖； 第9A-BB1係@示利用本發明之加頌獲得的壓力量變 曲線（pressure profiles); 第10A-C圖係圖示本發明之一具體實施例將一工作件 抛光； 20 帛11圖係圖示本發明之-具體實施例將-I作件拋光 ，其中顯示視工作的外形而有不同的力向量； 第12圖係為本發明之一具體實施例的一平臺其係具有 一減震緩衝層；以及 第13 A-B圖係圖示一呈㈣音#点丨 田士人外 畀體實鉍例，用於藉由從一工作 19 200400098 玖、發明說明 件之背面施以壓力而變化壓力量變曲線； 第Η圖係為一色彩感測裝置的一具體實施例，用於檢 測一多層的半導體晶圓的一加工終點，其中色彩感測裝置 包括一光源、一色彩感應器以及一判定電路； 第15圖係為一種用於檢測一多層的半導體晶圓的一加 工終點的方法，其之一具體實施例的流程圖； 第16 Α圖係圖示一用於就地終點檢測之終點檢測裝置 的一具體實施例的俯視圖，其係包括一可移動的結構與一 感測裝置； 第16B圖係圖示一用於就地終點檢測之終點檢測裝置 的一具體實施例的側視圖，其係包括可移動的結構與感測 裝置； 第17A圖係圖示安置在一示範化學機械拋光（CMP)裝 置中的一終點檢測裝置的一具體實施例，其中該化學機械 15抛光（CMP)裝置包括一載具頭、一抛光構件、終點檢測裝 置以及一執道，並且該化學機械拋裝置係處在一 拋光模式； 第17B圖係圖示安置在一示範化學機械拋光（CMp)裝 置中的、終點檢測裝置的一具體實施例，其中該化學機械 20 拋光（CMP)裝置包括—載具頭、—拋光構件、終點檢測裝 置以及執道，並且該化學機械拋光（CMp)裝置係處在一 非拋光模式； 第18圖係為一種在一 檢測一多層半導體晶圓的 化學機械拋光（CMP)裝置中用於 一加工終點的方法的一具體實施 20 200400098 玖、發明說明 例的流程圖，該裝置具有一載具頭以及一拋光構件，並且 其中該半導體晶圓係附裝至載具頭。 【實施方式3 較佳實施例之詳細說明 5 如以下的說明，本發明提供用於針對材料去除加工， 諸如化學機械拋光（CMP)，的就地終點檢測的一種方法與 一種系統。於整個參考圖式中相同的代表符號係代表相同 的元件。 A.終點檢測系統 10 第2圖係顯示一示範的化學機械拋光（CMP)裝置10〇， 其係包括一拋光構件102以及一載具頭104。拋光構件，例 如，係可為一拋光帶、一拋光墊或是另一型式的拋光構件 。拋光構件102包括一上或加工表面ι〇6以及一下表面1〇8 。拋光構件之下表面108係配置在一支撐板109上，諸如平 15臺’並繃緊。拋光構件與載具頭係經配置因此工作件面係 接近拋光構件，其係可接近或是接觸拋光構件。於此具體 實施例中，拋光構件102係為一透光的拋光構件。一拋光 溶液110係在拋光構件1〇2之加工表面106上流動，並且拋 光構件係在一組滾輪112上方藉由一移動機構（未顯示）在單 20 一方向抑或是雙方向上移動。於此具體實施例中，拋光構 件係在雙方向上移動。拋光溶液110係可為一銅拋光溶液 或是一研磨拋光漿液。溶液110係可由晶圓之一側或雙側 進給在拋光構件上，或是其可同時經由拋光構件進給在晶 圓表面上，或疋上述二方式皆可。一待加工之晶圓114係 21 200400098 玖、發明說明 由載具頭104所支撐，因此晶圓之一前表面116，之後將視 為表面’係完全地暴露。載具頭1〇4係可將晶圓垂直地上 下移動，同時經由一軸118轉動晶圓114。晶圓114之表面 116係可具有於第1A圖中所示的結構具有一銅層16(包括種 5晶層與沉積銅），係能夠向下拋光至阻障層18(如第1B圖中 所示），同時就地利用本發明執行終點檢測。於此實例中 ，起載層係為銅（Cu)，阻障層is係為艇（Ta)。絕緣層14可 為一氧化石夕（Si〇2)或一低介電值或是極低介電值材料所製 成。於此具體實施例中，一終點監測裝置12〇，較佳地包 10括一光學發射器與檢測器，係配置在拋光構件1〇2的下方 。當將銅層向下拋光至位在絕緣層之頂部表面15上的阻障 層18時（見第1A-1B圖），終點監測裝置12〇檢測拋光終點。 阻障層18 —經暴露並由裝置12〇所檢測，則停止加工。於 一可任擇的步驟中，如為所欲，可持續加工直至將阻障層 15向下抛光至下伏的乳化物層（underlying oxide layer)。如以 下所《兒明，裝置120係可配置於位在平臺1〇9中的腔室中。 本發明之裝置120能夠為任一的光學監測裝置係用於監測 反射率的變化。儘管於此係使用銅作為一實例材料，但是 本發明同時可用在去除其他的材料，例如導體，諸如鎳 20 (Nl)、鈀（Pd)、麵（Pt)、金（Au)、斜（Pb)、錫（Sn)、銀（Ag) 以及其之合金，鈕（Ta)、氮化鈕（TaN)、鈦（Ti)及氮化鈦 (ΤιΝ)，以及絕緣體和半導體。於加工期間，轉動晶圓丨14 並藉拋光構件102之加工表面106與表面116接觸，並在拋 光溶液11〇係在加工表面106上流動並弄濕晶圓之表面116 22 200400098 玖、發明說明 時移動。 如第3圖中所示係為-平面圖，同時第4圖係為橫截面 視圖，監測裝置120係配置於構成在平臺ι〇9中的一腔室 122中。如第4圖中所示，腔室122之頂部能夠藉由一透明 5的視窗加以密封。於此具體實施例中，按適當的尺寸製作 腔室122並構形用以容納監測裂置之伸長的主體沿著腔室 122的移動。腔室122的位置係與位在拋光構件與下伏平臺 上的晶圓之相對位置互相關聯。於加工期間，可藉由一移 動機構（未顯示）沿著腔室移動監測裝置用以掃描晶圓之半 1〇徑。由於掃描的動作，所以監測到介於晶圓之邊緣與晶圓 之中心間的不同位置。該腔室係可擴大超過晶圓之中心， 因此田日日圓轉動時藉由在腔室中滑動該監測裝置俾便產生 掃描的動作’例如，能夠沿著晶圓之直徑完成讀取寬的 光譜。能夠執行此掃描程序作為一持續的加工，或是於加 15 工步驟中進行。 於此具體實施例中，一附裝至監測裝置的反射鏡126 使輸出的光子k號128投射在晶圓表面上。反射鏡126因而 谷許進入的經反射光學信號丨3 〇或是反射光學信號抵達監 測裝置120。於可交替的具體實施财，以不同的形式使 2〇用監測裝置，諸如可撓曲的微光纖，可去除使用一反射鏡 ，並且k號可直接地由裝置輸送至銅表面。該裝置確定終 點’亦即’當經反射光學信號13〇的強度改變時阻障層18 係立即暴露（見第1B圖）。假若化學機械拋光（CMp)加工係 持續進行用以去除阻障層，則當絕緣層14之頂部表面15暴 23 200400098 玖、發明說明 路日寸經反射的信號的強度係再次改變（見第1B圖）。藉由監 測裝置所產生或由其所引導的光學信號，其之波長範圍為 600-900奈米。輸出之光學信號可藉由裝置12〇之一發射器 所產生，諸如一具有一截斷器（ch〇pper)或一發光二極體或 5疋雷射的白光發射器。根據目前的一較佳具體實施例，藉 由裝置120的一檢測器接收經反射的光學信號。一示範的 檢、]器係了為熱電檢測器（pyroelectric detector)。進入 的光學信號首先可通過一帶通濾光片（bandpass mter)其係 °又置用以實貝上消除所有的波長而留下由檢測器所檢測到 1〇的該一波長。於此具體實施例中，輸出的與經反射的信號 有利地行進通過係為透光的拋光構件。另一可交替的具體 實施例係配置一多重監測裝置陣列，該裝置係固定在自平 板（星狀）之中心延伸之徑向構成的腔室中，其係與晶圓之 中“致’用以監測在晶圓表面上的信號變化。再者，可 15交替地，可沿著一單一腔室分配複數之監測裝置。以此方 式，監測裝置可自轉動的晶圓表面之中心、中間以及邊緣 區域收集資料。 根據本發明之一觀點，整個拋光構件係以透明的材料 製成，並且不因終點檢測而需額外的視窗。於此具體實施 2〇例中，拋光構件包括一合成的結構係具有一頂部透明研磨 層構成在一透明的襯底（backing)材料上。在加工期間一研 磨層係與工作件接觸，並包括精細的研磨顆粒分佈在一透 明的黏結劑基材（binder matrix)中。與本發明一同使用的 一示範線性拋光構件結構，係可包括一薄的透明研磨塗層 24 200400098 玖、發明說明 ’例如厚度為5微米至⑽微米’堆疊在一透明的密拉襯底 材料师打上，該材料係由位在加州，Hayward 的Mipox公司所販售。研磨層之厚度係可為㈣米至⑽微 米，同時襯底層之厚度係可為〇5至2公髮。研磨層中研磨 5顆粒的尺寸係位在約為〇.2至〇5微米的範圍内。顆粒所用 之示範的材料係可為氧化石夕、氧化銘或氧化鈽。-較不透 明的抛光構件，但仍能於本發明使用，同時係由位在明尼 蘇達州的3M公司所販售。儘管於一些具體實施例中拋光 構件可包括研磨顆粒，但是抛光構件同時能夠以不具研磨 10顆粒的透明聚合材料所製成。 如以上所說明，當研磨拋光構件自晶圓表面去除材料 並且當阻障層或氧化層暴露時，經反射的光線強度受到改 炎於實例中，一透明的拋光構件具有使用約為1〇微米 厚度的研磨層以及厚度為〇.5至丨〇公釐的透明密拉層 (Mylar layer)。於此實例中，該研磨層具有〇·2至〇·5微米經 燻過的氧化石夕顆粒。波長為675奈米的光束（輸出）係輸送通 過此拋光構件，並監測於整個化學機械拋*(cMp)加工中 強度的變化。利用此拋光構件觀測到整個銅錄加工中， 、’、i反射之光線的強度係維持為2的任意（合標準的）強度值。 20然而，阻障層（鈕層）一經露出則強度降至為}。再者，當自 氧化層之頂σρ去除阻卩早層並露出氧化層時，經反射之光線 的強度係降至為〇. 5。 於第3圖中所示，於較佳的具體實施例中，該監測裝 置120係與一電腦132連接，該電腦同時可電氣地與一載具 25 200400098 玖、發明說明 頭控制态(未顯不)連接，儘管應瞭解的是可以複數種方式 、τ /汁算但並非必需具有一處理器的一電腦，而能取 代地使用分離的或積體邏輯電路，包括但非限定在AS· $ X及可式化閘極陣列上。當在銅層上作業其下方具有阻 p早層時，一經露出阻障層則由於反射率的改變而改變了出 自I測裝置的輪出信號，並終止機械拋光加工。 般地’根據本發明之觀點的終點檢測裝置與方法係 化用在i更多的工作件上，用以檢測在每一工作件上的 一或更多的終點。例如，根據本發明之一觀點，一化學機 10械拋光（CMP)的終點檢測加工針對一單一的工作件，諸如 一晶圓，可檢測到具有複數之化學機械拋光（CMP)終點。 化學機械拋光（CMP)終點能夠具有個別的拋光順序以及相 配合的個別的加工狀K物，自晶圓的表面去除超載之 金屬’係可代表一第一化學機械拋光（CMP)終點，去除位 15在晶圓之特徵外的阻障層可代表一第二化學機械拋光 (CMP)、、、ς點。k號強度的第一低限（thresh〇ld)或程度係可 用於檢測第一化學機械拋光（CMp)終點，因此當藉由檢測 系統所觀測到的信號強度下降至或低於該第一低限或程度 時’確定已抵達第一化學機械拋光（CMp)終點。可使用信 20號強度的其他低限（threshold)或程度用以檢測其他的化學 機械拋光（CMP)終點。例如，用於檢測一第二化學機械拋 光（CMP)終點，當藉由檢測系統所觀測到的信號強度下降 至或低於一第二低限或程度（其係低於第一低限或程度）時 ，確定已抵達第二化學機械拋光（CMP)終點。 26 200400098 玖、發明說明 應瞭解的是於前述之說明與附加的申請專利範圍中， 名判工作件表面及’工作件的表面”係包括，但不限定， 加工之前的工作件之表面以及構成在工作件上之任一層的 表面，包括導體、氧化金屬、氧化物、旋轉塗佈的玻璃 5 (spin-on glass)、陶甍等。 B·智慧型終點檢測系統 如以下所說明，本發明提供一種就地用於材料去除加 工，諸如化學機械拋光（CMP)，之厚度一致性控制與一終 點檢測的方法。於此系統中，拋光構件使用諸如視窗或是 1〇透明部分之元件可為透光的，或是部分透光的。 第5A-C圖係圖示一工作件表面的視圖。第5八圖係圖 不在將一薄膜16，例如銅，已沉積覆蓋之後的一晶圓9。 曰曰圓C括複數之電路構成在晶圓基板5 1 中，其係 針對說明而顯示，其中n係為任意值。該等電路中之每一 15電路包括大量的特徵係以沉積的傳導性薄膜填注，通常覆 蓋一阻障層。化學機械拋光（CMP)加工去除超載部分並將 傳導性薄膜留在該等特徵中。然而，應注意的是所具有的 一總體的表面厚度變化在利用一諸如化學機械拋光（cMp) 的加工去除超載部分時係需加以推平。由於表面變化，一 2〇加工係簡單地將薄膜U拋光至一預定的厚度，可能會造成 特定區域過度拋光而其他區域會拋光不足。 第5B圖係圖示在晶圓114上局部的表面變化，其係針 對說明而作了稍許的放大。就以上所述，由於表面變化， 一加工係簡單地將薄膜16拋光至一預定的厚度，可能會造 27 200400098 玖、發明說明 成特定區域過度拋光而其他區域會拋光不足。 第5C圖係圖示具有所欲之拋光終點的晶圓，其中傳導 性層係位在特徵中並去除了超載部分。 於一具體實施例中，本發明之厚度一致性的檢測與控 5制系統利用其之即時厚度測量能力以及其之涵蓋加工參數 的控制，維持經加工表面之厚度的一致性。根據得自於經 加工之晶圓的表面的即時厚度資料，在一化學機械拋光 (CMP)加工期間，厚度一致性的控制系統變化拋光參數用 以均勻一致地拋光一層。因此，抵達經拋光層之終點總體 10地涵蓋晶圓表面而不致造成對目標層過度拋光及拋光不足 。藉由局部地變化拋光構件下方的壓力而改變拋光參數， 因此特定位置之拋光係快於其他的位置。 於本發明之一觀點中，本發明藉使用所檢測之即時終 點資料維持經加工表面的一致性。根據自經加工晶圓之表 15面所得之即時資料，在化學機械拋光（CMP)均勻一致地拋 光一層期間厚度一致性的控制系統變化拋光參數。 儘管於此係使用銅作為一實例材料，但是本發明同時 可用在去除其他的材料，例如導體，諸如鎳（Ni)、鈀（pd) 、鉑（Pt)、金（Au)、鉛（Pb)、錫（Sn)、銀（Ag)以及其之合金 20 ’钽（Ta)、氮化组（TaN)、鈦（Ti)及氮化鈦（TiN)，以及絕緣 體和半導體。 第6A圖係顯示具有一厚度一致性的控制單元560的一 種示範化學機械拋光（CMP)裝置550。化學機械拋光（CMP) 裝置可進一步地包括一研磨性拋光構件102以及一載具頭 28 200400098 玖、發明說明 刚。拋光構件1()2包括一上或加工表面ig6以及_下表面 108。拋光構件之下表面1〇8係配置在一支推板6⑼上諸 如一平3：，並繃緊。拋光構件較佳地包括一合成的結構係 -有頂邛透明研磨層構成在一透明的襯底材料上。在加 5工期間-研磨層係與卫作件接觸，並包括精細的研磨顆粒 分佈在-透明的黏結劑基材中。與本發明一同使用的一示 範線性拋光構件結構，係可包括—薄的透明研磨塗^ 1 如厚度為5微米至⑽微米，堆疊在一透明的密拉襯底材料 (Mylar backing)上’該材料係由位在加州，㈣评如的 1〇 MiPox公司所販售。研磨層之厚度係可為5微米至1〇〇微米 ’同時襯底層之厚度係可為〇.5至2公董。研磨層中研磨顆 粒的尺寸係位在約為〇·2至〇.5微米的範圍内。 平室包括複數之孔620a-620n，於第6B圖（同時見於第 7A-7B圖）中詳細地顯示，於加工期間用於在拋光構件的下 15方產生-流體壓力。假若使用一包括研磨料的化學機械抛 光（CMP)漿液或拋光溶液，則拋光構件1〇2係可以非研磨性 的拋光構件所取代。孔62〇心62〇11係可與藉由流體供應單 兀562所供應之一流體連接。於此具體實施例中，拋光構 件102係可為一透光的拋光構件，但同時可為一拋光構件 2〇其中具有視窗或是以透光的部分所組成。於本發明之一觀 點中，流體供應單元562包括轉動的流量計係控制流體流 動至平$。例如，可以將流至每一平臺區域的流體控制在 0至5立方呎/分鐘（cfm)。可交替地，可以藉由於市場上所 販售的電子式質量流量控制器控制流體流動並測量。該等 29 200400098 玖、發明說明 電子式質量流量控制器係可以軟體控制並使其自動化。示 範的質量流量控制器係由SMC及Celerity所販售。 選定具有足夠可撓性的拋光構件符合所施加的壓力， 並與一對著晶圓表面相關的局部壓力連通。示範的具體實 5 施例使用一可撓曲的聚合體拋光構件，充分地將壓力傳送 至局部區域。假若拋光構件的撓曲性不足，例如以一鋼帶 加強’則壓力會連通涵蓋一較大的區域，並且系統可持續 拋光晶圓之非所欲的區域。 一抛光溶液112係在拋光構件102之加工表面106上流 10動’並且拋光構件係在一組滾輪113上方藉由一移動機構（ 未顯不）在單一方向抑或是雙方向上移動。於此具體實施 例中’拋光構件較佳地係在雙方向上移動。拋光溶液U2 係可為一銅拋光溶液或是一研磨拋光漿液。溶液丨丨2係可 由晶圓之一側或雙側進給至拋光構件上，或是其可同時經 15由拋光構件進給在晶圓表面上，或是上述二方式皆可。一 待加工之晶圓114係由載具頭1〇4所支撐，因此晶圓之一前 表面116，之後將視為表面，係完全地暴露。載具頭1〇4係 可將晶圓垂直地上下移動，同時經由一軸118轉動晶圓ιΐ4 。晶圓114之表面116初始係可具有於第5八圖中所示的結構 〇 ，有一銅層16(包括種晶層與沉積銅），能夠向下拋光至一 終點（如第5C圖所示），同時就地執行以下本發明之厚度一 致性的檢測與控制加工。於此，同時利用阻障層去除步驟 持續加工，因此將絕緣層之頂部表面15上的阻障層拋光直 至絕緣層14露出或是抵達阻障層終點。於此實例中，超載 30 200400098 玖、發明說明 層係為銅（Cu)，阻障層18係為鈕（Ta)以及絕緣層14係為二 氧化矽（Si02)。 一致性控制單元包括一流體供應單元562，用於將流 體（例如’空氣）輸送至平臺600。一致性控制單元同時包括 5 具有一中央處理單元（CPU)的一電腦控制器564、記憶體、 顯示器、鍵盤以及其他的共同元件。電腦564係與一系列 之示範的感應器630a-630n結合，其中n係為通過一感應器 控制器566的一任意的感應器識別符號（63〇卜63〇(1同時顯示 於第6Β及7Α-7Β圖中）。感應器630a_630n係配置在平台中 10與平台中的流體孔620a-620n相鄰。於此具體實施例中， 平台的孔較佳地係以特定的方式分組，例如以一環形方式 刀佈母一組之孔（見第6B，7A_7B圖）。示範的感應器可包 括厚度感應器與終點檢測感應器。如以下所說明，每一組 之孔（即所熟知的壓力區域）係與流體供應單元連接，輸送 15由電腦控制器564所控制的流體壓力。流體供應單元能夠 針對彼此獨立的每一壓力區域變化流體壓力（如同流體流 動）。 於本發明之一觀點中，感應器63〇α·63〇η係為終點感 應器包括一光學發射器及檢測器配置在拋光構件的下方。 2〇例如，當將銅層向下拋光至位在絕緣層之頂部表面15上的 阻障層18時，終點感應器檢測拋光終點（見第1Α-1Β圖）。 如上述的解釋，本發明能夠控制來自平臺之不同的區 域的局部壓力，用以增加或降低在晶圓上的局部抛光率。 因此，本發明之一關鍵的觀點係為藉由使用位在平台上的 31 200400098 玖、發明說明 不同Μ力區域提供不同抛光率的能力。藉由充分地控制在 每一個別的壓力區域上流體或空氣壓力程度，改良此系統 之拋光靈敏性。針對壓力區域建立精確地控制壓力程度， 依次’導致較佳地控制晶圓上的局部拋光率。 5 如第6]3及6<:圖中所示，於較佳的具體實施例中，同 時可藉由自平板之頂部將過多之空氣去除而達到該等具有 預定壓力程度的分開壓力區域。如以下更為詳盡的說明， 藉由容許經控制地洩漏至大氣或一真空源，本發明調節流 動涵蓋鄰近壓力區域之所吹送的過多空氣，亦即調節鄰近 10區域間的串音干擾（cross_talk)，並致使在鄰近區域中空氣 壓力程度上的改變。如第6B圖中所示，於一具體實施例中 ，示範的系統1000具有空氣洩放閥。於此具體實施例中， 電腦控制器與感應器單元為了清楚起見並未顯示。系統主 要係由平臺600、晶圓載具1〇4用以支撐待加工之晶圓114 15 、以及拋光帶102或拋光墊所組成。如以上說明，帶1〇2具 有頂部表面106或是一加工表面，以及背面1〇8。晶圓114 之前表面116面向拋光帶1〇2之頂部表面。上述拋光帶與拋 光溶液的細節係作為例子，因此為了清楚起見，於此不在 贅述。 2〇 與第6A圖比較，第6B圖係更為詳細地顯示平臺6〇()。 如第6B圖中所示，平臺6〇〇可具有一上表面61〇將一基底塊 件612圍住。上表面係劃分成同中心的壓力區域，即第一 區域zl、第二區域以、第三區域幻以及第四區域z4。該等 同中心的區域同時例示於第7A_7B圖中。區域zl_z4包括孔 32 200400098 玖、發明說明 62〇a-62〇n。於第6B圖中，每一區域可包含二或更多的孔 。例如，第一區域Zl包括孔6施諸如此類。感應器6術_ 630η同時係配置在每一區域中。為了清楚起見，第佔圖中 並未包括電腦控制器與感應器單元以及與此單元連接的連 5接裝置（見第6A圖）。再者，於表面010中的每一區域係以 如第6B圖中所示之方式與一空氣室614a_614d相對應。例 如，位在第一區域zl中的孔620a係藉由將空氣流經室61乜 而進氣，位在第二區域z2中的孔620b係藉由將空氣流經室 614b而進氣諸如此類。室614a_614d係構成為環形同中心 ίο的溝槽，其係分別地經由空氣管路616a_616d連接至一空 氣供應單元562。每一空氣管路6l6a-616d係經由一或更多 的空氣口 618a_618d與對應的室連接。再者，藉使用連接 器，例如T型連接器，每一空氣管路6i6a-616d係分別地與 壓力控制裝置622a-622d連接。於此具體實施例中，壓力 15控制裝置係為與空氣管路616a-616d連接的空氣閥622a_ 622d。於此方面，每一閥係與其中之一壓力區域結合，例 如’第一閥622a係用於第一區域；21，以及第二閥622b係用 於第—區域z 2諸如此類。 閥 622a-622d 包括排氣口 624a-624d。排氣口 624a_624d 20係可連接至外界大氣環境或真空（未顯示），用於自系統 1000將經排放的空氣去除。於此具體實施例中，經由該等 閥件，能夠調節可自排氣口 624a-624d排放的空氣量，從 而調節在一壓力區域上的正壓力。當開啟閥622a-622d時 ’其將流經管路616a-616d之某一百分比的空氣排出。於 33 200400098 玖、發明說明 此方面’在該等區域中能夠使用閥622a-622d產生一正壓 力或一負壓力或是零壓力。利用一真空連接裝置，能夠在 壓力區域中產生一負壓力或是零壓力。 然而’與該區域結合的閥之最重要的功能在於當來自 5鄰近區域的過多空氣流動涵蓋該區域並致使在該區域的空 氣壓力增加時，排放空氣用以調整一壓力區域中的壓力程 度。於此具體實施例中，空氣供應單元能夠以相同的空氣 流動率供應空氣至每一壓力區域，以及變化至個別壓力區 域的流動率用以在拋光帶1 〇2下方建立一具有一預定空氣 10 壓力量變曲線的空氣區域。 如第6C圖中所示的另一具體實施例中，平臺6〇〇包括 流體放出孔1400，其係較佳地，配置在該等區域之間用以 自平板之頂部去除過多的流體。藉由容許流體經由流體放 出孔而洩漏至大氣或一真空源，本發明消除了過多流體會 15流動覆蓋该等鄰近壓力區域的問題，並因此大體上將該等 鄰近區域間的串音干擾降至最低。如此，依次，在平臺上 大體上產生獨立的壓力區域，有利地容許在每一壓力區域 中使用不同的壓力程度。如第6(：圖中所示，於一實例中， 流體放出孔1400a· 1400d係配置在同中心的壓力區域η、z2 20 、z3及z4之間，該等區域具有流體孔62〇a-62〇d以及感應器 630a-630d。介於每一區域之間，複數之放出孔係構成在 一單一的或是多於一的環形路徑上。每一環形路徑可具有 至少一管路其係具有複數之放出孔1400a-1400b。例如， 複數之位在區域zl與z2之間的放出孔14〇〇a係可沿著一單 34 200400098 玖、發明說明 一的環形路徑或是包括複數之放出孔的二同中心的環形路 徑而構成。儘管，於此具體實施例中放出孔係沿著環形路 徑構成並介於該等區域之間，其可以任一方式分佈，諸如 徑向，並涵蓋在本發明之範疇内。於此具體實施例中，放 5出孔係按適當的形狀製成圓形或環形；然而，其係可具有 矩形或是其他的幾何形狀或是其可按適當的形狀製成為一 環形狹縫。於化學機械拋光（CMP)加工期間，一諸如空氣 的流體在拋光帶102的下方經由位在每一區域中的流體孔 620a-620n噴射，同時支撐著晶圓114的載具頭1〇4係下降 1〇至拋光f上。當拋光帶102係在平臺600上方移動時，流體 通過孔620 a-620η在抛光帶102的下方施以壓力。介於壓力 區域間的放出孔將自壓力區域zl-z4流出之過多的流體放 出’並防止該等區域間的串音干擾。於加工期間，晶圓 114係可至少平移約為放出孔之直徑的兩倍，用以平衡放 15出孔可能局部化的效應。可開啟每一放出孔至大氣壓力， 或可連接至一真空系統（未顯示）。如第6(：圖中所示，於此 具體實施例中，每一流體放出孔1400a-1400d係個別連接 至大氣壓力。母一放出孔1400a-1400d係獨立地開啟與外 部壓力連通，並個別地將過多之流體放出至大氣。然而， 20 放出孔之最重要的功能係獨立地調整每一壓力區域中的壓 力程度。例如，於第一區域中的壓力大小係可藉由進給較 高的流量至zl並將第一區域ζι流出的過多之流體經由放出 孔1400a放出，而高於相鄰之壓力區域z2的壓力，因此其 不致影響z2中的壓力。於此具體實施例中，空氣供應單元 35 200400098 玖、發明說明 同時能夠以相同的空氣流動率供應空氣至每一壓力區域， 以及變化至個別壓力區域的流動率用以於每一流體中建立 一特定的壓力程度。如此在拋光帶1〇2下方產生一具有一 預定空氣壓力量變曲線。 5 第7A_7B圖係顯示具有區域zl_z4之平面610的平面圖 ，包括孔62〇a_62〇n及感應器630a_63〇n。於此具體實施例 中，不範的感應器630a_630n可為光學的終點感應器，較 佳地包括一光學發射器與檢測器，並係從工作件配置在位 於拋光構件下方的平臺中。例如，感應器63〇a_63〇n可坐 10落在或接近區域zl_z4，其係代表一壓力區域其中流體壓 力係藉由流體供應單元562選擇地加以控制。儘管於此具 體實施例中使用坐落在平臺中的示範光學感應器，但是能 夠於系統中使用任一型式的感應器坐落在任一適合的位置 ，而涵蓋在本發明之範疇内。如第7B圖中所示，每一區域 15 可包括複數之同心圓’並且進一步地期待在一些狀況下一 區域可不具有一感應器。感應器單元566接收未經處理的 感應器信號（例如，反射的光線）並產生電氣的感應器信號 ，其係被輸送至電腦564(見第6A圖），其係以上述的方式 控制流體供應單元562。 20 本發明之終點感應器能夠為任一光學監測裝置，其係 用以監測經拋光層之反射率的變化。參考第8圖，每一感 應器63Ox包括一輸送光纖632x係提供一反射離開工作件 114的光線（見代表符號710)，以及一接收光纖634x係接收 經反射的光線。當例如銅層係經向下拋光至位在絕緣層之 36 200400098 玖、發明說明 頂部表面15上的阻障層18時（見第1A-1B圖），終點感應器 藉由經反射的光線的變化檢測該拋光的終點。於此方面， 輸出與進入的信號係行進通過透光的拋光構件102。在化 學機械拋光（CMP)終點檢測中使用該等感應器，係揭露纟 5 於2002年1月17曰提出申請之美國專利申請案序號為第 10/052,475號中。 化學機械拋光（CMP)係為一加工製程，大體上根據以 下方程式拋光一表面： 拋光率=常數X速度X壓力 10 本發明利用控制局部壓力的能力，用以增加或降低局 部的拋光率。因此，本發明之一主要的觀點係為在不同的 壓力區域中使用不同的拋光率的能力。 利用壓力區域zl及z2例示一操作順序，用以建立於第 9A圖中所示的壓力量變曲線。應瞭解的是為了範例的目的 15使用二區域。與在第9A圖中的一壓力量變曲線相似的曲線 係能夠利用壓力區域zl、z2、z3及z4而構成。於第9A圖中 所示的壓力量變曲線能夠藉由在第一壓力區域zl中具有一 高空氣壓力P1而建立，但在附近的第二區域22中係為一較 低的空氣壓力。於操作當中，例如此係可首先在第一區域 20 zl中建立壓力P1而執行，該壓力係由空氣供應單元以一第 一預定的空氣量流動至第一區域zl所建立。於壓力？1建立 期間，第-閥622铸經調整用以排放來自第一管路6心的 第-空氣流動的-部分。在第二區域z2中建立壓力p2，例 如’係可藉由將第-敎的空氣量流經第二空氣管路6说 37 200400098 玖、發明說明 而完成，同時藉由將第一預定的空氣流經排氣口 624b而將 壓力降低至P2。於此由第一區域至第二區域的任一空氣流 動可將在第二區域中的壓力增加至壓力p3。根據本發明， 在第二區域z2中的壓力程度增加藉經由第二閥而自第一預 5定的流動排放更多的空氣而徹底改變。由於通氣，而發生 降低了被引導至第一區域的第一流動量，並且於第二區域 z2中的壓力程度恢復至P2的壓力程度。針對每一區域使用 不同的空氣流動執行相同的加工製程。於此狀況下，壓力 程度再次藉由排放預定的空氣流動量而加以調整。 1〇 同時使用區域zl&z2例示另一操作順序，用以建立於 第9B圖中所示之壓力量變曲線。與在第叩圖中的壓力量 變曲線相似的曲線係能夠利用壓力區域zl、z2、以及以而 構成。於第9B圖中所示的壓力量變曲線能夠藉由在第一壓 力區域zl中具有一低空氣壓力P1而建立，但在附近的第二 15區域Z2中係為一較高的空氣壓力。於操作#中，例如此係 可首先在第二區域z2中建立壓力P2而執行，該壓力係由空 氣供應單元562以一第一預定的空氣量流動至第二區域 所建立。於建立壓力p2期間，第二閥622b亦經開啟或閉合 用以排放第-空氣流動的一部分。在第一區域心建立壓 力P1，例如，係可藉由將第一預定的空氣量流經第一空氣 管路616a而完成，同時藉由將流經排氣口 62知的第一預定 空氣的一預定部分排放而將壓力降低至ρι。於此由第二區 域z2至第-區域zl的任一空氣流動可將在第一區域^中的 壓力增加至壓力P3。如先前的狀況，在第一區域Η中的壓 38 200400098 玖、發明說明 力程度增加藉經由第一閥622a而自第一預定的流動排放更 多的空氣而徹底改變。由於通氣，而發生降低了被引導至 第一區域zl的第一流動量，並且於第一區域中的壓力程度 恢復至P1的壓力程度。針對每一區域使用不同的空氣流動 5執行相同的製程。於此狀況下，壓力程度再次藉由排放預 定的空氣流動量而加以調整。該等結合第9A-9b圖所說明 之加工製程同時可動態地加以控制。例如，可以由如第6B 圖中所不配置在每一壓力區域21_24中的壓力感應器之輸 入駟料控制或調節該等閥件。當由於來自相鄰區域的空氣 10流動造成在一區域中的壓力增加，則閥件將預定的空氣量 排放用以調整在該區域的空氣壓力。藉由一控制器接收來 自感應器的壓力輸入資料，能夠控制經由閥件的通氣。 當在具有一位在下方的阻障層的一銅層上作業時，阻 障層經路出，則由於反射率的改變而使來自終點感應器 I5的信號變化。參考第10八_10(：：圖，於示範的加工中，晶圓 中的一區域係可較另一區域需要更多的拋光，或是一區域 較另-區域較快地變薄，因此針對一區域該銅終點係較另 -區域更加快速地抵達。銅終點—經藉由終點感應器所檢 測，則於麼力區域中的空氣壓力係降低用以使在該區域中 2〇的進-步拋光放慢速度或是消除進一步抛光的需求。可交 替地，在尚未抵達終點的其他區域中可增加空氣壓力。利 用不同的去除率，在完成的區域中大體上不再去除銅，而 另-區域能夠繼續拋光。於此本發明之觀點係基於其之有 闕於終點的狀態而龍力區域施以不同的空氣塵力。 39 200400098 玖、發明說明 第10B-10C圖係圖示智慧型終點檢測功能的一實例。 如第10B圖中所示，工作件表面係以代表符號92〇a所定義 。在若干拋光時刻之後，該表面係降低至代表符號92〇b， 並且接近位在感應器630c附近的區域該層係極薄。在更多 5的拋光時刻之後，當表面向下拋光至代表符號920c(920c-1 及920c-2)時’感應器630c將檢測表面中的變化以及控制器 560將降低至該區域的壓力（流體流率）。因此，該區域將經 歷較少的拋光，同時其他的區域在原始的速率下持續地拋 光。當然’假若係如此要求，同時期望的是能夠增加流體 10 動至特疋的未完成區域。所有的區域一經拋光（所有的 感應器皆回報已抵達終點），則完成加工。 儘管以上已詳述不同的較佳具體實施例，但熟知此技 藝之人士應可立即察知的是可對示範的具體實施例作複數 種修改，實質上不致背離本發明之新穎的講授内容與優點。 15 c ·具體實施例的變化形式 於本發明之一觀點中，能夠使用音響感應器取代上述 的光學感應器。於此觀點中，感應器63如_63011即時檢測 經拋光層的厚度，同時加工晶圓，並經由感應器單元566 供應此資料至電腦。電腦564因而鑑定所供應的厚度資料 2〇，假若檢測在經去除層中的不平坦性，並藉由變化在晶圓 上-或更多的拋光參數，諸如在拋光構件或浆液合成物下 方的工氣壓力，選擇性地再次調整材料去除率而獲得整個 晶圓表面的厚度一致性。 於本發明之另一觀點中，第11圖係圖示拋光-工作件 40 200400098 玖、發明說明 視工作件之特徵顯示不同的壓力向量910a_910d。箭頭越 長代表力量越大。假若一工作件區域需要更多的拋光，因 而電腦控制器指示流體供應單元對該區域提供增加的壓力 。同樣地，當一區域不需額外的拋光時，因而電腦控制器 5指不流體供應單元對該區域提供較小的壓力。 於本發明之另一觀點中，一熱交換器係與流體供應裝 置直列地（in-line)與平臺結合，因此控制輸送至平臺之流 體的溫度，並能夠維持在一預先設定的溫度。平臺能夠進 一步地包括一溫度感應器，為了提供反饋至熱交換器，為 1〇 了維持拋光構件之一預定的溫度。 D·具有緩衝層的平臺 於利用上述的一拋光構件的一化學機械拋光（CMP)加 工期間，複數種因素可損害拋光構件、晶圓表面抑或是以 上一者。就晶圓表面而言，當介於待拋光的工作件表面與 15拋光構件表面間接觸時任何的不平行，會對工作件表面造 成損害。在任何的化學機械拋光（CMp)加工之前，平臺表 面與待拋光的工作件表面應對正，因此其大體上係為平行 的任何顯著地偏離此平行會致使工作件的一部分更為接 近平臺表面，同時會將工作件表面的另一部分配置遠離平 2〇臺表面。該表面部分，或是所謂的位在工作件上的高點， 即更為接近平臺表面可過度拋光或是碰撞平臺表面，造成 工作件表面的損害並同時對拋光構件造成損害。介於平臺 與工作件表面間之該未對正，亦即不平行，在抛光包含基 板的低介電值材料期間係特別地造成損害。由於低介電值 41 200400098 玖、發明說明 材料之易相結構，在拋光齡電值基板_與平臺發生 任何的碰撞可完全地損害低介電值材料結構。 π 3，洛入固 5 W π黑抛尤稱件與平臺間 的任一大顆粒會刮傷或損害薄的固定的研磨拋光構件。再 者’終點視窗應平順地與平臺表面對正。任何顯著地未對 正之視心部會在平臺的表面上構成—凸塊，並會刮傷抛 光構件或損害工作件。 該等問題此夠利用與於此所述之平臺結合的減震介質 而加以避免。於一實例中，該減震介質係為-介於拋光構 1〇件與平臺表面間的減震緩衝層。於此所說明之具體實施例 能夠包括平臺、拋光構件（具有或不具研磨料）以及拋光溶 液（具有或不具漿液）的任一結合方式。 第12圖係顯示平臺600其具有一減震緩衝層1300係附 裝在平臺表面610之頂部±。緩衝層测係可以軟性的聚 。材料所製成’諸如聚氨酯或是任何該可禁得起化學機械 抛光（CMP)加工之化學環境的材料。緩衝層13〇〇可具有與 平臺流體孔620a_620n之相同型態的第一孔132〇a_132〇n , 以及與感應器630a-630n同型態的第二孔133〇a_133〇n。於 此具體實施例中，孔1320a]32〇n之尺寸係可大於流體孔 2〇 620a_620n之尺寸。於化學機械拋光（CMP)加工期間，孔 1320a-1320n容許一流體，諸如空氣，在拋光構件1〇2的下 方喷射，同時支撐晶圓114的載具頭1〇4係降低至拋光構件 上。拋光構件因而較佳地係在平台上方，包括緩衝層，以 雙方向線性運動地移動。當然，拋光構件可在其他的方向 42 200400098 玖、發明說明 上移動，例如環形方向。 當拋光構件102係在緩衝層1300上方移動時，在抛光 構件102下方經由孔1320施加流體壓力。緩衝層容許流體 分佈遍及並涵蓋平臺，但額外地提供安全性用以避免平臺 5硬質表面、拋光構件以及晶圓表面間的意外接觸。本發明 致使針對易碎的低介電值以及極低介電值材料的化學機械 拋光（CMP)加工提供特定的優點。軟質的緩衝層吸收對於 晶圓的任何瞬時震波，並將對低介電值材料造成的損害降 至最低。 10 15 20 除了前述的具體實施例之外，本具體實施例針對低介 電值材料基板提供一種經改良的化學機械拋光（CMp)加工 。儘管與傳統式拋㈣件純使用固定式研磨拋光構件可 提供較低的過度碟形下陷與凹陷（dishing and簡i〇n) ,但 是位在固定式研磨拋光構件上的硬質表面當使用在具低介 電值材料的基板上時，可產生較高的缺陷或是局部的分層 (de—)。如先前所提及，在鋼金屬化製程中使用低 介電值材料-般極為易碎並具有不良的黏著性。控制介於 基板與拋光構件間的摩㈣、數，對於防止在不同的化學機 械拋光（⑽)步_間發生低介電值㈣分層係為重要的 。相關於在銅/低介電值材料的整合中低介電值材料的整 體強度以及導致損害的化學機械拋光(CMP)之技術挑戰， 皆可利用本發明之加卫而降低或甚至消除。 的拋光溶液。然而，於本發明之一加工中 利用固^研磨拋光材料的傳統式技術可使用不具襞液 一示範基板的 43 200400098 玖、發明說明 一銅層係可利用一固定研磨拋光構件而去除，同時一包含 一預定漿液量的拋光溶液係被輸送至該固定研磨拋光構件 上。該等增加的顆粒潤滑拋光構件表面，並降低位在經拋 光的基板表面上的橫向力。示範的顆粒包括，但不限定， 5氧化鋁、氧化鈽、氧化矽或是任何其他的金屬氧化物或聚 合物樹脂珠。在拋光溶液中顆粒的濃度係可為重量比從 0.1至40% ，更佳地為從〇_5至5% 。一示範的拋光溶液可藉 由增加氧化銘或氧化矽顆粒至一銅拋光溶液中，諸如由 3M公司所販售的CPS-11溶液，製備而得。 10 E.多層拋光 於另一具體實施例中，可於一整體的化學機械拋光 (CMP)工具中，在個別的化學機械拋光（CMP)站中使用的 個別之拋光構件上執行去除銅與阻障層。於第一化學機械 拋光（CMP)站中，於第一加工順序，基板之銅層係利用固 15定研磨拋光以及一包含顆粒的拋光溶液加以去除。拋光加 工係可利用減震缓衝層1300加以執行，其係於前述的具體 實施例中結合第12圖加以說明。於加工期間，使用一與第 12圖中所示相似的一系統，將一晶圓降低在固定研磨拋光 構件上，以及將一包含潤滑顆粒的拋光溶液配送至拋光構 20 件上。如以上所述，將固定研磨拋光構件在緩衝層丨3〇〇上 方移動，同時在拋光構件下方施以一流體壓力。銅層一經 向下去除至位在低介電值材料之表面上的阻障層時（見第 1B圖），在一第二化學機械拋光（CMP)站中執行一阻障層去 除加工。於此步驟中，於第12圖中所示的一化學機械拋光 44 200400098 玖、發明說明 (CMP)站係可使用一聚合/非固定研磨拋光構件。拋光構件 係可以諸如聚氨酯的一種軟性聚合材料所製成。於此實例 中，於阻障層去除期間，將一具選擇性的拋光溶液配送至 適於阻障層材料去除的聚合拋光構件上，同時移動拋光構 5件並如上所述在拋光構件的下方施以一流體壓力。此加工 步驟的順序將在低介電值材料上的應力以及將所導致之分 層降至最小，以及將碟形下陷與凹陷降至最低。 於另一具體實施例中，可於相同的化學機械拋光 (CMP)站中執行去除銅及阻障層。在將阻障層去除之前針 10對去除銅而執行第一步驟。根據此加工順序，於第一步驟 中，大量的銅係可在固定研磨拋光構件上向下去除至位在 固定研磨拋光構件上的阻障層。於此步驟，拋光溶液可或 未包含顆粒。於一第二步驟中，結合固定研磨拋光構件與 具有顆粒的拋光溶液，用於自阻障層的表面去除剩餘的銅 層同時在工作件上施以一向下力，例如，可為一相對低 的向下力。接續戎等步驟，於另一化學機械拋光（CMp)加 工站中，在一軟性聚合拋光構件上執行一阻障層去除步驟 ，同時將一鈕選擇性拋光溶液配送至拋光構件上，並同時 在工作件上施以一低的向下力。 20 F•載具頭的壓力變化 第13A-B圖係圖示一針對變化壓力量變曲線的具體實 施例，其係藉由自晶圓114後方施以壓力。於此具體實施 例中，利用頭104對晶圓114施以壓力梯度，同時將晶圓支 撐在適當的位置。一可撓曲或膨脹的薄膜121〇與載具頭的 45 200400098 玖、發明說明 形狀一致，典型地係為環形的形狀，並係附裝與一升高的 表面區域之内周圍相鄰。膨脹的薄膜121〇在加工期間提供 一依從的晶圓支撐。膨脹的薄膜121〇係以薄的依從材料所 建構而成，諸如彈性體，較佳地係為vit〇n@。該薄膜較佳 5地使用結合黏著劑與扣件或夾緊機構而附裝至載具頭1〇4 。此附裝的結構當膨脹時，將薄膜121〇支撐並密封。 儘管不範的具體實施例係描述為一膨脹的薄膜，但是 薄膜係可交替地以一可撓曲，但不需為膨脹的，依從的材 料所建構而成。假若薄膜係非膨脹的，則可使用一海綿型 10 (spongy type)材料迫使晶圓靠著拋光構件。 參考第13A圖，薄膜121〇係劃分成複數之區域121〇心 1210e，其中可具有任意數目之區域。為了施以一壓力梯 度至工作件，所以供應流體進入該等區域並同時可自該等 區域排出。如以下所說明，來自流體管路1224心1224己的 15流體係用以將可膨脹薄膜1210膨脹，並經由發生的加工維 持該膨脹。於加工期間，藉由薄膜施加的壓力較佳地係位 在〇·1至10 psi的範圍内。 在加工期間可以複數種方式的其中之一方式可將晶圓 支撐在適當的位置。如第13B圖中所示，其中之一方式係 20藉使用一扣件1212a-1212b。該一扣件1212較佳地將晶圓 支撐在一固定的位置上，同時不致妨礙到表面加工。另一 用於將晶圓支撐在適當位置的技術，係為在晶圓與薄膜之 間使用真空，與在美國專利申請案序號為第1〇/〇43,656號 中所說明的方式相似，於此併入本文以為參考資料。在操 46 200400098 玖、發明說明 作中，在將晶圓114配置在薄膜1210上之後，襯底構件係 膨脹直至下層與薄膜1210接觸。接著排空載具頭腔室施以 真空吸住晶圓114。當對腔室施以真空時，介於袋狀部分 間的連接區域或是低凹部分提供低壓空間，並從而致使鄰 5接的薄膜部分塌陷落入低凹部分。如此，依次，在晶圓 114之背面上產生複數之低壓空間。該低壓空間的作用如 同吸盤（suction cups)並提供足夠的抽吸動力用以在加工期 間固持晶圓。 在拋光時藉由個別的壓力管路1224a-1224e與一壓力 1〇控制器1220連接。該等管路容許壓力控制器在晶圓之背面 處產生一可變化的壓力梯度，因此已經存在於晶圓之前表 面上薄膜去除率的一致性，於加工期間係可藉由在晶圓背 面的不同壓力加以控制。例如，對晶圓之中心施以較高的 壓力而對晶圓之周圍施以較小的壓力，與位在晶圓之周圍 15處的機械元件相較位在晶圓中心處加工之機械元件係顯著 地增加，增加了來自中心區域的材料去除率。 第13B圖同時係顯示平臺16〇〇，其係與上述的平臺6〇〇 相似，或可為一具有拋光構件固定於其上的平坦表面。於 本發明之該一觀點中，晶圓與拋光構件間的相對運動係藉 2〇由移動拋光構件、載具頭或是該二者而達成。於任一狀況 下’基板表面監測感應器63〇a-630n係安裝在平臺中，並 經由拋光構件抑或通過位在拋光構件中的一開口監測晶圓 。位在平臺1600中的感應器與第6Λ圖中所示者相似，與一 感應器單元566與電腦控制器564連接。電腦控制器控制壓 47 200400098 玖、發明說明 力控制器mo並提供反饋至加工系統，為了控制施加位在 工作件上的每一區域的壓力並最理想地加工工作件。相關 於第HK：圖之流程圖說明上述㈣，可在不同的時刻使用 此方法在工作件的不同區域處選擇終點。 5 G·具有色彩感應器的感測裝置 於-具體實施例中，用於一多層晶圓之終點檢測的一 感應器係為一色彩感應器。於本文中，名詞，，色彩，，意謂自 表面反射或發射光線不同的特性中的至少一特性。經反射 光線具有色彩變化的特性，例如複數之波長。第14圖係圖 10不一色彩感測裝置1405的一示範具體實施例，其係用於檢 測-多層半導體晶圓的m終點，其中該色彩感測裝置 包括一光源1410、一色彩感應器142〇以及一判定電路143〇 。名詞’’感測結構”係可與名詞，，感測裝置，，可替換地使用 。如以下進一步地說明，色彩感應器係可為一單一波長感 15應器或是一複合波長感應器（多波長感應器）。色彩感測裝 置，例如，係可結合一淺溝隔離（STI)的化學機械拋光 (CMP)程序使用。相關於以上的第1(：與1〇圖提供一示範的 淺溝隔離（STI)的化學機械拋光（CMP)程序的說明。 於示範的具體實施例中，光源對著半導體晶圓之一表 20面發射入射光。色彩感應器係光學地與光源結合並反應入 射光線自半導體晶圓之表面感測經反射光線，係為一所謂 的反射色彩。於一觀點中，色彩感應器係可為一單一波長 感應器。色彩感應器係構形反應反射色彩而產生一感應器 k说。判定電路係與色彩感應器結合，並係構形用以基於 48 200400098 玖、發明說明 至少部分的感應器信號判定是否已抵達晶圓加工終點。 於本發明之一觀點中，光源與色彩感應器係坐落極接 近晶圓。於另一觀點中，光源係與一光纖結合。於此觀點 中，光源包括光纖之輸出端。同樣地，色彩感應器係可與 5 一光纖結合用以感測反射色彩。於此觀點中，色彩感應器 包括光纖。 如上所述，可以一多波長感應器取代一單一波長感應 器。光源可發射多光譜入射光，並且色彩感應器可感測一 多光譜反射。多光譜係意謂具有至少二波長。於本發明之 10 一觀點中，色彩感應器係構形感測波長範圍自400至800奈 米延伸的光線。於另一方面，光源發射白色入射光並且色 彩感應器感測一紅-綠-藍（RGB)反射。 判定電路係構形用以至少部分地基於感應器信號判定 疋否已抵達晶圓加工終點。判定電路可包括一比較器用以 15 對照一低限反射色彩（threshold reflection color)比較來自 半導體晶圓表面的反射色彩。低限反射色彩，例如，係可 為來自一已抵達其之加工終點的樣品半導體晶圓的反射色 彩。於此方面’基於來自比較器的反射色彩比較資料判定 是否已抵達加工終點。反射色彩比較資料，例如，係可比 20 較反射波長。於本發明之另一觀點中，判定電路係利用演 算法確定是否已抵達晶圓加工終點。 低限反射色彩係可藉由感測一已知材料的反射色彩而 初始化。於一觀點中，低限反射色彩係基於來自一樣品半 導體晶圓之一氧化石夕（Si〇2)層的一反射。於另一觀點中， 49 200400098 玖、發明說明 低限反射色彩係基於來自一樣品半導體晶圓之氮化石夕 (Si3N4)層的-反射。於另—觀點中，晶圓之_上層係為銅 (Cu)以及一下層係為一阻障層，諸如鈕（Ta)或是氮化鈕 (TaN)或鈕/氮化鈕（Ta/TaN)。於此觀點中，低限反射色彩 5係可基於來自一已拋光至阻障層之樣品半導體晶圓的一反 射。可交替地，低限反射色彩係可基於來自樣品半導體晶 圓的一銅層的一反射。再者於可交替的方式中，低限反射 色彩係可基於來自樣品半導體晶圓的一絕緣層的一反射。 於進一步的觀點中，半導體晶圓的一層係為親水的 10 (hydroPhllie)以及另一層係為疏水的（hydr〇ph〇bic)。（親水 的係意謂易於保持水分，同時疏水的係意謂不易於保持水 分）。例如，晶圓之一上層係可由親水的二氧化矽所組成 Π夺aa圓之一下層係為疏水的氮化石夕。由於二氧化碎層 係為親水的，所以一薄水膜典型地係構成在其之表面上。 15 J而s 一淺溝隔離化學機械拋光加工將晶圓向下抛光至 氮化矽層時，典型地在氮化物表面上具有少量或是無水分 。在氮化矽表面上未有水分係考慮到加工終點之前後一致 的測量。 如以上相關於第14圖之說明，感測裝置係可結合淺溝 2〇隔離化學機械拋光使用。當一接受淺溝隔離化學機械拋光 的半導體晶圓係自二氧化矽層55拋光至氮化矽/二氧化矽 界面時（參考第1C及1D圖），反射色彩係由微綠的（通常為 4-5 kA)變化至黃或紫。於此實例中，氮化矽/二氧化矽界 面係代表加工終點。因此，再次參考第14圖，色彩感測裝 50 200400098 玖、發明說明 置藉由監測反射色彩何時由微綠變化至黃或紫，而能夠檢 測一淺溝隔離化學機械拋光加工何時已成功地抵達加工終 點。先前的淺溝隔離化學機械拋光技術係為示範的，並且 其他的技術係為預料中的。 5 色彩感測器係可容許在感測角及感測距離，亦即由色 彩感應器至半導體晶圓之表面的距離，上作變化。於一觀 點中’色彩感應器係配置在一段考慮到能感測到最理想光 學信號的感測距離處。例如，感測距離係可為2_ 1〇公釐。 可操作感測裝置用以在一預定的頻率下在半導體晶圓 10上執行終點檢測。例如，感測裝置係可每隔第50個晶圓加 以測試，用以確定一晶圓拋光製程的精確度。 第15圖係為用於檢測一多層半導體晶圓之一加工終點 ’例如使用色彩感測裝置1405，的一方法的一具體實施例 之流程圖。於步驟1510中，對著一半導體晶圓之一表面發 15射入射光。於步驟1520中，在反應入射光自半導體晶圓之 表面感測一反射色彩。於步驟153〇中，基於所感測的反射 色彩產生一感應器信號。於步驟丨54〇中，至少部分地基於 感應器信號確定是否已抵達晶圓加工終點。 使用色彩感應器可降低或去除結合其他型式的光電感 2〇應器所產生的問題，諸如受限制的微分能力 (differentlati〇n capacity)以及無法補償在目標距離上的變 動。可與本發明一同使用的一示範色彩感應器係由位在新 澤西州，WoodcliffLake，的Keyence&司所販售。 H.用於就地終點檢測的可移動結構 51 200400098 玖、發明說明 針對就地終點檢測而言，一感測裝置係可與一可移動 結構結合。由於感測裝置與可移動結構結合，所以不需將 半導體晶圓自其之加工承座，亦即載具頭丨〇4，移開即可 在一半導體晶圓上執行終點檢測（相關於第2圖）。於一具體 5實施例中，一終點檢測系統包括一感測裝置係構形用以感 測一與一半導體晶圓之表面相關的度量，用以基於該度量 產生一感應器信號。該系統同時包括一判定電路係與該感 測裝置結合，並構形用以至少部分地基於感應器信號判定 是否已抵達晶圓加工終點。該系統進一步地包括一可移動 10結構係與該感測裝置結合，定位感測裝置用以感測該度量。 感測裝置可包括，例如，上述相關於第14圖所說明的 光源1410以及色彩感應器1420。於此觀點，光源以及色彩 感應器係與可移動結構結合，用以感測來自半導體晶圓之 表面的反射色彩。於另一觀點中，光源以及色彩感應器係 15與可移動結構結合，用以掃描半導體晶圓之表面。於另一 觀點中，可移動結構將色彩感應器定位用以感測反射色彩 。感測裝置同時包括判定電路1430。可交替地，感測裝置 可為一與上述相關於第14圖所說明的感測裝置14〇5不同的 感測裝置。 2〇 弟16A圖係圖示用於就地終點檢測的一終點檢測裝置 1610的一具體實施例之俯視圖，其係包括一可移動結構 1620以及一感測裝置1630。可移動結構係與感測裝置結合 並使感測裝置能夠定位在不同的位置。例如，可移動結構 可將感測裝置定位在一主動位置（感測位置），或是一不活 52 200400098 玖、發明說明 動位置（非感測位置）。感測裝置檢測使用上述技術，諸如 反射色彩感測’檢測一晶圓加工終點。同時係可使用兑他 的終點檢測技術。感測裝置可包括一光電感應器，諸如上 述相關於第14圖所說明的色彩感應器。再次相關於第16A 5 圖，感測裝置係可結合一判定電路用以至少部分地基於由 感測裝置所產生的資料判定是否已抵達晶圓加工終點。第 16B圖係為用於就地終點檢測的一終點檢測裝置161 〇的一 具體實施例之側視圖，其係包括可移動結構1620以及感測 裝置1630。 10 第17 A圖係圖示終點檢測裝置1710，其係位在一示範 的化學機械拋光裝置1700中，其中化學機械拋光裝置包括 載具頭104、拋光構件102、終點檢測裝置1610以及一軌道 1730，並且該化學機械拋光裝置係處在拋光模式下。該執 道提供一路徑供終點檢測裝置在其上行進，用以執行就地 15 終點檢測。如上所述，第ΠΑ圖顯示處在拋光模式下的化 學機械拋光裝置，將處在向下位置的載具頭以及晶圓114 之底部表面116與拋光構件102之拋光表面1〇6接觸。儘管 於第17A圖中係圖示化學機械拋光裝置處在拋光模式下， 但是終點檢測裝置係處在未作動的位置，意謂終點檢測裝 20 置並非位在感測裝置在晶圓之底部表面上執行終點檢測的 一位置上。 第17B圖係圖示終點檢測裝置1610，其係位在一示範 的化學機械拋光裝置1700，其中化學機械拋光裝置包括載 具頭104、拋光構件102、終點檢測裝置1610以及執道1730 53 200400098 玖、發明說明 ，並且該化學機械拋光裝置1700係處在非抛光模式下。如 上所述，第17B圖顯示處在非拋光模式下的化學機械抛光 裝置，將載具頭處在升高的位置以及晶圓之底部表面與抛 光構件之撤光表面接觸。將載具頭處在一升高的位置，終 5點檢測裝置在載具頭下方沿著軌道移動並將感測裳置定位 在晶圓之底部表面下方，從而將終點檢測裝置定位在一作 動的位置。儘管定位在晶圓之底部表面下方，但是感測裝 置在半導體晶圓上執行終點檢測。例如，感測裝置可自晶 圓表面感測反射色彩。應注意的是不需為了執行終點檢測 10而將晶圓自載具頭卸载。 假若感測裝置確定已抵達終點，接著可將晶圓自載具 員卸載並被載送至—接續的加卫站。於―觀點中，可移動 結構可將半導體晶圓移動（載送）至接續的加工站。 15 可移動結構可為任-型式適於將感測裝置定位用於就 地終點檢測的構件，諸如一搬運梭（shuttle)、臂件或β 1 他型式的構件。於-觀點中，可移動結構係為-清潔：運 ㈣之功能係用以將晶圓在抵達加工終點之後㈣= 潔室（未顯示）。於此觀點巾，清潔搬運梭係設計作為可= 動結構用於定位感測裝置。假若感測裝置_1 繼係位在-作動位置時’已抵達終點，接著：晶圓二 載在清潔搬運梭上（亦即可移動結構）並載送至清潔室接為 清潔作業。應瞭解的是軌道並非為本發明所必需。例如又 假若可移動結構係為_臂件則可不需軌道。 假若感測裝置辞定並未抵達終點，則接著終點檢測裳 54 200400098 玖、發明說明 置係自載具頭下方移開（恢復至一未作動位置），並且載具 頭下降用以將晶圓之背面配置與拋光構件之拋光表面接觸 作額外的拋光。拋光晶圓並移動終點檢測裝置進入適當位 置用以檢測晶圓加工終點這樣的循環，係可持續直至抵達 5 終點為止。 於本發明之另一觀點中，轴118以及載具頭使晶圓旋 轉，如第17A及17B圖中位在軸上方以環形箭頭所顯示。 於此觀點中，由於晶圓係受旋轉，所以終點檢測裝置能夠 藉由在-涵蓋晶圓半徑的平直路徑上移動而掃描晶圓的整 1〇個表面。可交替地，假若未旋轉晶圓，終點檢測裝置可配 置馬達用以旋轉終點檢測裝置，因此能夠掃描整個晶圓 表面。終點檢測裝置可替代地具有多感測裝置用以掃描整 個晶圓表面。 第18圖係為用於檢測在一具有一載具頭以及一拋光構 B件的化學機械拋光裝置中，諸如示範的化學機械抛光裝置 1600，一多層半導體晶圓之一加工終點的一方法的一具體 實施例之流程圖。於步驟1810中，停止拋光半導體晶圓。 於步驟1820中，藉由升高載具頭移開與拋光構件接觸的半 導體晶圓。於步驟1830中，在半導體晶圓之一底部表面的 2〇下方移動一感測裝置。於步驟1 840中，對著半導體晶圓之 底部表面自感測裝置發射入射光。於步驟丨85〇中，在反應 入射光利用感測裝置自半導體晶圓之底部表面感測一反射 色彩。於步驟1860中，至少部分地基於反射色彩確定是否 繼續拋光半導體晶圓。於一觀點中，該方法進一步地包括 55 200400098 玖、發明說明 中斷拋光半導體晶圓，以及假若感測到一所欲的反射色彩 即移動半導體晶圓至另一加工站。 I.結論 本發明之優點包括提供拋光至一選定的終點之最理想 5 工作件的能力。於本發明之一觀點中，可使用於此所說明 之技術用以拋光尺寸變化的晶圓。例如，該等技術可用以 拋光直徑為200公釐、300公釐、400公釐、500公釐或其他 直徑的晶圓。於本發明之一觀點中，使用相同平臺拋光不 同尺寸的晶圓。 10 應瞭解的是於前述的說明與附加的申請專利範圍中， 名詞晶圓表面”與，’晶圓的表面’’係包括，但非限定，在加 工之前晶圓的表面以及構成在晶圓上任一層的表面，包括 導體、氧化金屬、氧化物、旋轉塗佈的玻璃（spin_〇n gkss) 、陶瓷等。名詞’’晶圓’’、，，半導體晶圓，，以及，，基板，，係可替 15 換地使用。 應瞭解的是於此所說明的具體實施例與觀點係可結合 以任-適合的方式_同操作。例如，感測裝置^彻及/ 或可移動結構1620係可與上述㈣g型終點m统及/ 或載具頭壓力變化系統結合，用於提供涵蓋半導體晶圓的 2〇厚度-致性。前述的結合係僅為實例。同時可考量其他的 結合與具體實施例。 同時應瞭解的是儘管特定的晶圓加工，諸如化學機械 光已加以次明，但可結合任何其他型式的晶圓加工實 行本發明，諸如電化學機械沉積（ECMD)。 56 200400098 玖、發明說明 已揭鉻示範的具體實施例與最佳模式，可對揭露的具 體實施例作修改以及變化，同時涵蓋在由以下申請專利範 圍所限定之本發明的主題與精神内。 【圖式簡單說明】 5 第1A圖係為一示範的基板在將材料沉積在基板表面上 之後的橫截面視圖； 第1B圖係為第1A圖之示範的基板在接受一傳統式的 化學機械拋光（CMP)加工之後的橫截面視圖； 第1C圖係為一示範的基板在將絕緣材料沉積在基板表 10 面之上後的橫截面視圖； 第1D圖係為第1C圖之示範的基板在接受一傳統式的 化學機械拋光（CMP)加工之後的橫截面視圖； 弟2圖係為一示範的CNU系統的橫截面侧視圖，該系 統包括一本較佳具體實施例之一用於加工諸如晶圓的工作 15 件的示範終點檢測系統； 第3圖係為第4圖之示範的化學機械撤光（CMP)系統以 及一根據本發明之觀點供終點檢測系統所用的示範控制系 統的橫截面俯視圖； 弟4圖係為包括第2圖之不範終點檢測系統之示範的化 20 學機械拋光（CMP)系統的橫截面侧視圖； 第5A-C圖係為一工作件表面的視圖； 第6 A圖係圖示本發明之一具體實施例的一種工作件加 工糸統； 第6B圖係圖示本發明之另一具體實施例的一種工作件 57 200400098 玖、發明說明 加工系統； 第6C圖係圖示本發明之另一具體實施例的一種工作件 加工系統； 第7A-B圖係圖示於第6A-6B圖中本發明之一具體實施 5 例之平臺； 第8圖係為本發明之一具體實施例之一感應器的分解 視圖， 第9 A-B圖係圖示利用本發明之加工所獲得的壓力量變 曲、線（pressure profiles); 10 第10A_C圖係圖示本發明之一具體實施例將一工作件 拋光； 第11圖係圖示本發明之一具體實施例將一工作件拋光 ，其中顯示視工作的外形而有不同的力向量； 第12圖係為本發明之一具體實施例的一平臺其係具有 15 —減震缓衝層；以及 第13A-B圖係圖示一具體實施例，用於藉由從一工作 件之背面施以壓力而變化壓力量變曲線； 第14圖係為一色彩感測裝置的一具體實施例，用於檢 測一多層的半導體晶圓的一加工終點，其中色彩感測裝置 20包括一光源、一色彩感應器以及一判定電路； 第15圖係為一種用於檢測一多層的半導體晶圓的一加 工終點的方法，其之一具體實施例的流程圖； 第圖係圖示-用於就地終點檢測之終點檢測裝置 的一具體實施例的俯視圖，其係包括一可移動的結構與一 58 200400098 玖、發明說明 感測裝置； 第16B圖係圖示一用於就地終點檢測之終點檢測裝置 的一具體實施例的側視圖，其係包括可移動的結構與感測 裝置； 第17A圖係圖示安置在一示範化學機械拋光（CMp)裝 置中的一終點檢測裝置的一具體實施例，其中該化學機械 拋光（CMP)裝置包括一載具頭、一拋光構件、終點檢測裝 置以及一執道，並且該化學機械拋光……巧裝置係處在一 拋光模式； 第17B圖係圖示安置在一示範化學機械拋光（CMp)裝 置中的一終點檢測裝置的一具體實施例，其中該化學機械 拋光（CMP)裝置包括一載具頭、一拋光構件、終點檢測裝 置以及一執道，並且該化學機械拋光^乂^裝置係處在一 非拋光模式； 15 第18圖係為一種在一化學機械拋光裝置中用於 檢測一多層半導體晶圓的一加工終點的方法的一具體實施 例的流程圖，該裝置具有一載具頭以及一拋光構件，並且 其中該半導體晶圓係附裝至載具頭。 【圖式之主要元件代表符號表】 8.·.不範的部分 14…絕緣層 9···經電鍍基板 15…頂部表面 10··.通孔 12…通孔 13...溝 16.. .銅層 18.. .阻障層 51.. .示範的部分 59 200400098 玖、發明說明 5 2 · · ·基板 53···溝 54··.底部絕緣層 5 5…頂部絕緣層 1〇〇···化學機械拋光裝置 102.·.抛光構件 104·.·載具頭 10 6 · · ·加工表面 108·..下表面 109···支撐板 110…拋光溶液 112···滾輪/拋光溶液 113…滾輪 114 · · ·晶圓 116.··前表面 118 · · ·轴 120···終點監測裝置 122…腔室 124.··透明的視窗 126·.·反射鏡 128…輸出的光學信號 130…進入的經反射光學信號 132…電腦 510a-510n.··晶圓基板 550···化學機械拋光裝置 560…厚度一致性的控制單元 562···流體供應單元 564···電腦控制器 566···感應器控制器 600···支撐板/平臺 610···上表面 612···基底塊件 614a-614d···空氣室 616a-616d.··空氣管路 618a-618d···空氣口 620a-620n.··孔 622a-622d...壓力控制裝置 624a-624d···排氣口 630a-630n·.·感應器 630x…感應器 632x.··輸送光纖 634x···接收光纖 710.. .光線 910a-910d···壓力向量 920a...工作件表面 920b·"表面 920c-l，920c-2·..表面 1000.. .系統 60 200400098 玖、發明說明 1210.. .薄膜 1210a_1210e...區域 1212a-1212b...扣件 1220.. .壓力控制器 1224a-1224e…流體管路 1300…減震緩衝層 1320a-1320n···第一孔 1330a_1330n···第二孔 1400.. .流體放出孔 1420…色彩感應器 1430.. .判定電路 1400a-1400d…流體放出孔 1405.. .色彩感測裝置 1410.. .光源 1600…平臺/化學機械拋光 裝置 1610.. .終點檢測裝置 1620.. .可移動結構 1630.. .感測裝置 1700.. .化學機械拋光裝置 1710.. .終點檢測裝置 1730.. .軌道 Z1-Z4···第一至第四區域 61200400098 发明 Description of the invention (The description of the invention should state: the technical field to which the invention belongs, the prior art, the content, the embodiments, and the drawings are briefly explained. The serial number of the US patent 5 filed on December 17, 2002 is 10 / 321,150 @ 丁 -280-1; 3), the serial number of the US patent filed on March 22, 2002 No. 10 / 05,016 (NT-250-US), US Patent Application No. 10 / 197,090 (NT-248-US) filed on July 15, 2002, and January 17, 2002 The serial number of the U.S. patent application serial number 10 / 052,475 10 (NT_238-US) filed for application is a continuation application, and all applications are incorporated herein for reference. This application is for US Provisional Application No. 60 / 436,706 (NT-278-P4) filed on December 27, 2002, and US Provisional Application No. 60 / filed on December 23, 2002 No. 436, 108 (NT-278-15 P3), US Provisional Application No. 60 / 417,544 (NT-278-P2) filed on October 10, 2002, and filed on September 27, 2002 U.S. Provisional Application No. 60 / 415,579 (NT-278-P), U.S. Provisional Application No. 60 / 397,110 (NT-273-P), filed on July 19, 2002, March 12, 2002 The priority of U.S. Provisional Application No. 20 60 / 365,016 (NT-249-P) is filed, and all applications are hereby incorporated by reference. TECHNICAL FIELD The present invention relates to the manufacture of semiconductor integrated circuits, and more particularly, to a method for chemical mechanical polishing of conductive and insulating layers. 200400098 发明 Description of the invention [Prior Art 2 Background of the Invention Traditional semiconductor devices generally include a semiconductor substrate, usually a silicon substrate, and a plurality of successive dielectric interlayers (such as silicon dioxide), and Conductive road control or interconnections made of conductive materials. Recently, due to its superior electromigration and low resistivity characteristics, copper and copper alloys have received considerable attention for use as interconnect materials. Interconnects are usually formed by filling the remaining features or copper in the cavity into a 10 dielectric layer by a metallization process. The preferred method for copper metallization is electroplating. In an integrated circuit, the multilayer interconnection network is related to the substrate surface extending L laterally. The interconnects formed in continuous layers can be electrically connected using vias or contacts. In a typical process, an insulating layer is first formed on a semiconductor substrate. Patterning and etching processes are performed to form features such as trenches or vias in the insulating layer. After coating the features on the surface with a barrier layer and a seed layer thereafter, the features are electroplated with copper. However, the electroplating process, in addition to filling these features, also results in a copper layer on the top surface of the substrate. This excess copper is overburdened by the month and should be removed before subsequent processing steps. 20 Figure 18 shows the plated substrate 9, such as a silicon wafer, with a typical 邛 of 8. It should be noted that the substrate 9 may include components or other metal and semiconductor parts, which are not shown in FIG. 1A for explanation. As shown in FIG. 1A, the characteristics of two f, ′, = such as a through hole 10 and a groove 13 are formed in an insulating layer 14, and the wife 1 produces a plug such as a silicon dioxide layer, which is formed on the substrate 9. In 200400098, the description of the invention, the top surface 15 of the hole and trench 13 and the insulating layer 14 is covered and filled with a deposited copper layer 16 by an electroplating process. Traditionally, after patterning and etching, the insulating layer 14 is first coated with a barrier layer 18, which is typically a group (Ta) or a button / nitride group (TaN) composite layer. Barrier layer 丨 8 is applied to cover the surface 15 of the through-holes and trenches and the insulating layer to ensure good adhesion and is used as a barrier material to prevent copper from diffusing into the semiconductor element and the insulating layer. Furthermore, a crystalline layer (not shown), usually a copper layer, is deposited on the barrier layer. During subsequent copper deposition, the seed layer forms a substrate of conductive material for forming a copper thin film. When the copper film is electroplated, the copper layer 16 deposited 10 quickly fills the through hole 10, but coats the wide groove 13 and the top surface 15 in a conformal manner. When the continuous deposition process is used to ensure that the trenches are simultaneously filled, a copper layer or an overburden is formed on the substrate 9. Traditionally, after copper electroplating, different materials can be used for removal processing, such as chemical mechanical polishing (CMp), etching or electroetching, to remove undesired overload layers. Chemical mechanical polishing (CMP) processing traditionally involves pressing a semiconductor wafer or other substrate against a wet moving polishing surface having a polishing slurry. The slurry can be alkaline, neutral, or acidic, and generally contains hard abrasive ceramic 20 particles of alumina, ceria, silica, or other. The polished surface is typically a flat pad and is made of a polymeric material that is well known in chemical mechanical polishing (CMP) technology. Some polishing pads contain abrasive particles (fixed abrasive pads). These pads can be used in combination with a chemical mechanical polishing (CMP) solution, which may not contain any abrasive particles. If the pad is porous, the polishing slurry or solution can be transported to the surface of the pad 200400098, description of the invention, or can flow through the pad to its surface. During chemical mechanical polishing (CMP) processing, a wafer carrier supports a wafer to be processed and places the wafer surface on a chemical mechanical polishing (CMP) pad, and presses the wafer under a controlled pressure. Lean on the pad while turning it. The pad is also configurable 5 as a linear polishing belt that can move laterally like a linear belt. By moving the wafer against the pad, moving the pad against the wafer, or the above two moving methods, processing is performed when the polishing slurry is supplied to the interface between the pad and the wafer surface. As shown in FIG. 1B, first, chemical mechanical polishing (CMP) is applied to reduce the thickness of the copper layer covering the top surface 15 of the insulating layer 4 to the barrier layer 18. Next, the barrier layer 18 on the top surface is removed to confine the copper and the barrier layer in the through holes 10, 12 and the trenches 13. However, the end of transition polishing during these processes, whether polishing the copper layer down to the barrier layer or polishing the barrier layer down to the insulating layer, is an important issue in the industry. 15 U.S. Patent No. 5,605,760 describes a polishing pad made of a thin, uniform polymer sheet. The polymer sheet is transparent in a specific wavelength range. The surface of the polymer sheet does not contain any abrasive material and does not have any inherent ability to absorb or transport slurry particles. Recently, the endpoint detection system has been implemented using a 20-rotation pad with a window or multiple windows or a linear belt system. Under these conditions, when the pad or belt is moved, it is measured over the in-situ monitor by measuring the reflection coefficient from the surface of the wafer. The change in reflection indicates the end of the polishing process. However, the windows opened in the polishing pad can complicate the polishing process and disturb the homogeneity of the pad or belt. In addition, the 10 200400098 玖, invention description and other windows can cause the accumulation of by-products and slurry of polishing processing. Therefore, when the substrate is polished by chemical mechanical polishing (CMP), the method and apparatus are to be able to accurately and effectively detect an end point on the substrate. 5 As shown in FIG. 1B, first, chemical mechanical polishing (CMP) is applied to reduce the thickness of the copper layer covering the top surface 15 of the insulating layer 14 to the barrier layer 18. Then, the barrier layer 8 on the top surface is removed to define the copper and the remaining barrier layers in the through hole 10 and the trench 13. However, uniformly reducing the thickness of the polished copper layer during these processes is an important industrial issue. 10 When processing it, it is necessary to maintain the uniformity of the thickness of the metal layer, so that the excessive polishing after the copper end point is minimized and the substrate is not over polished, because over polishing will cause excessive dishing and depression erosion) and other defects. Furthermore, insufficient polishing of the copper layer and the barrier layer may cause electrical shorts or other defects. The non-uniformity during the polishing process may be caused by a uniform polishing process or an uneven thickness of one of the metal layers located on the substrate or the above two conditions. Polishing the insulation layer of a substrate is another application of chemical mechanical polishing (CMp). Shallow trench isolation (STI) is a processing method in which an insulating trench is formed in the surface of a substrate to prevent electrical migration between adjacent circuits. These 20 trenches are typically filled with silicon nitride (SigN4) and silicon dioxide (SiO2). To fill the trenches, a silicon nitride layer is first deposited on the surface of the substrate, followed by a silicon dioxide coating. It is necessary to remove too much silicon dioxide and silicon nitride from the surface of the substrate, leaving a smooth silicon nitride layer covering most of the substrate surface and the silicon oxide and silicon nitride layers filling the trench area. Removal of excess silicon dioxide and silicon nitride is typically performed by 200400098 玖, invention description Chemical mechanical polishing (CMP). Figure 1C shows a cross-sectional view of a substrate 52, such as a silicon wafer, one exemplary portion 51, which is covered with two layers of insulating material. A trench 53 suitable for shallow trench isolation (STI) is formed in the surface of the substrate 52. A bottom insulation layer 54 and a top insulation layer 55 cover the surface of the substrate 52 including the grooves 53. The bottom insulating layer 54 and the top insulating layer 55 may be, for example, silicon nitride and silicon dioxide, respectively. It should be noted that the insulating layers 54 and 55 cover the positive surface of the substrate. For shallow trench isolation (STI) processing, excess insulating material must be removed. 10 15 20 Figure 1D is a cross-sectional view of an exemplary portion 51 of the substrate 52 after the insulating layers 54 and 55 have been polished to a desired level, that is, after excess insulating material has been removed. For example, the polishing of the insulating layer can be performed by chemical mechanical polishing (cMp). It should be noted that one of the smoothing layers of the insulating layer 54, that is, the surface of the substrate 52 covered by the dream, and the filling layers 53 filled with the insulating layers 54 and 55 (i.e., silicon nitride and silicon dioxide). Issues related to shallow trench isolation (STI) technology include the difficulty of measuring the thickness of silicon dioxide with an optical interferometer, because the thickness measurement signal itself is repeated periodically as the thickness of the oxygen cut increases or decreases. . In addition, the 'thickness measurement signal is sensitive to environmental factors such as moisture (water film) and detection angle. / Two, an additional problem with the objective technology is that traditional weighing and weighing tools remove the substrate from its carrier head to perform endpoint detection. Consistent polishing and polishing (CMP) costs, while adding processing systems can significantly reduce the productivity of chemical mechanical polishing. When the wafer size becomes larger than 12 200400098 mm, and the invention description is large, for example, 300 mm and larger, it becomes more difficult to reduce the thickness in a uniform and flat manner due to the large surface area of the wafer. Therefore, there is a need for an improved method and apparatus for monitoring and maintaining the consistency of a polished layer when a polished substrate is processed by chemical mechanical polishing (CMP). SUMMARY OF THE INVENTION The present invention advantageously provides a polishing method and apparatus for controlling planarity in material removal processing, such as chemical mechanical polishing (CMP). A specific embodiment of the present invention 10 includes this force for performing endpoint detection in the material removal process. Another embodiment provides a smart endpoint detection with the same pressure control technology, which selectively applies polishing pressure to a specific area on a work piece. According to an aspect of the present invention, a chemical mechanical polishing (CMP) device is provided for polishing a surface of a work piece and for detecting a chemical mechanical polishing (CMP) end point. The chemical mechanical polishing device includes a light-transmissive polishing member, a work piece holder, a support plate, and an optical detection system. The polishing member can be, for example, a polishing tape, a polishing pad, or a saint. Polishing members. Polishing members preferably include abrasive particles. The surface of the work piece may be polished in one or more directions (preferably in a linear direction, but also in other directions. For example, a ring-shaped work piece supports the work piece and is configured to press the work piece holder against the polishing member, for example, it can be a wafer carrier head or other structure for supporting the wafer. The support plate is designed to support the polishing member when it is pressed against the work piece against the polishing structure 13 200400098, the invention description piece. The support plate, for example, can be a platform or other support structure. The optical detection system detects chemistry The end of mechanical polishing (CMP), and it is arranged under the polishing member. The optical detection system includes-light source and-detector. The light source transmits output signals through the support of 5 plates and the polishing member to the work The detector receives the reflected signal from the surface of the work piece through the polishing member and the support plate. According to another aspect of the present invention, a surface of a work piece is polished and a chemical mechanical polishing (CMp) endpoint is detected. Method. According to this method, the work piece is pressed against a light-transmissive polishing member. The concrete member system 10 is supported by a support plate. The surface of the work piece is polished with a polishing member. The polishing member can be Or more linear direction. The output optical signal is from a source to the support plate and the polishing member to the surface of the work piece. The light source is arranged below the polishing member, so the polishing member is between the light source and the work Between the surfaces of the pieces. The incoming reflected optical signal 15 is received by the detector from the surface of the work piece through the polishing member and the support plate. The detector is arranged below the polishing member. According to a further aspect of the present invention, a method is provided. Method for polishing one or more work pieces and providing chemical mechanical polishing (CMP) endpoint detection. According to the method,- The light polishing member is disposed between a supply area and a 0-then area. The polishing member has a first end portion and a second end portion, and-a polishing side and a back portion. The first end portion is initially disengaged. The supply area is connected to the receiving area, and the second end portion remains connected to the receiving area. A first work piece is received by moving a part of the polishing member in the -polishing area in-or more linear directions. Polishing. Using an optical inspection 14 200400098 玖, invention description system detects the first chemical mechanical polishing (CMP) end point of one of the first work pieces. The optical detection system sends an output signal to the first work piece, and receives the first signal from the first through the polishing member Incoming reflection signal of the work piece. The polishing member is located between the optical detection system and the first work piece. 5 According to another aspect of the present invention, a surface of a work piece is polished and used to detect a chemical mechanical polishing ( CMP) chemical mechanical polishing (CMP) device. A chemical mechanical polishing (CMP) apparatus includes a supply roll and a receiving roll, a light-transmitting polishing member, a processing area, a member for moving a portion of the polishing member in one or more linear directions 10, and A component is used to detect a chemical mechanical polishing (CMP) endpoint. The polishing member has two ends. One end is attached to the supply roll and the other end is attached to the receiving roll. The processing area has a part of the polishing member between the two ends. The component used to detect the end of chemical mechanical polishing (CMP) sends optical signals to the surface of the work piece, and receives reflected optical signals from the surface of the work piece via the polishing component. The polished component is positioned between the component for inspection and the work piece. According to a further aspect of the present invention, a method for polishing a surface of a work piece and detecting a chemical mechanical polishing (CMP) endpoint is proposed. According to the method, the work piece is supported in a processing area such that the surface of the work piece is exposed to a part of a light-transmissive polishing member. The surface of the wafer is polished by moving a portion of the polishing member linearly and bidirectionally. A chemical mechanical polishing (CMP) end point is determined for the work piece by transmitting and outputting the optical signal through the polishing member to the work piece, and continuously checking the relative intensity of the incoming optical signal reflected from the work piece and received through the polishing member. 15 200400098 (ii) Description of the invention The previous description of the point of view of the present invention is provided by way of introduction only. In this section, no limitation is imposed on the following application materials for determining the present invention. Knife-A second exemplary embodiment of the present invention includes a polishing station having a work piece holder and a flexible polishing member. The polishing member system 5 supplies a fluid to the back of the polishing member against a work piece supported by a -platform. The platform includes a plurality of holes for supplying fluid, and also includes a plurality of sensors capable of detecting the end point of the work piece processing. The holes are clustered together to create pressure zones and each zone is typically combined with a sensor, but can be combined with more or fewer sensors. A computer receives 10 sensor signals and controls fluid flow to optimize polishing. For example, if a specific position on the work piece reaches the end point, the computer reduces the fluid flow to the 4 position while maintaining fluid flow to other areas. In another exemplary embodiment of the present invention, a sensing device for detecting a processing end point of a multilayer semiconductor wafer includes a 15 light source emitting light toward a surface of a semiconductor wafer, a The color sensor is used to sense the reflected color from the surface of the semiconductor wafer after responding to incident light, and is used to generate a sensor signal, and a decision circuit is combined with the color sensor and configured to at least partially Based on the sensor signal, it is determined whether the wafer processing end point has been reached. 20 In another exemplary embodiment of the present invention, an end point detection system for detecting a processing end point of a semiconductor wafer includes a sensing device configured to sense one of the semiconductor wafers. A surface-related metric generates a sensor signal based on the metric. The endpoint detection system also includes a determination circuit combined with the sensing device and is configured to 16 200400098. The invention description is based in part on the sensor signal to determine whether the end of the wafer processing has been reached, and a movable structure system and The sensing device is combined to position the sensing device to sense the scale. In another exemplary embodiment of the present invention, a method for detecting a processing end point of a 5-layer semiconductor wafer includes emitting light toward the surface of the semiconductor wafer, and reflecting the incident light. Sensing the color reflected from the surface of the semiconductor wafer, generating a sensor signal based on the sensed reflected color, and determining whether the wafer gate end point has been reached based at least in part on the sensor signal From a standpoint, the fluid controller alone controls the flow of fluid to the pressure region. One of the characteristics of this viewpoint is that the present invention can also selectively discharge fluid from a specific hole in the platform to reduce and even negatively affect the pressure region. In another aspect of the present invention, the rotating work piece and the 15 platform holes are concentrically arranged during processing, and each concentric ring member represents a pressure region. In another aspect of the invention, the fluid controller alone controls the flow of fluid to the equivalent central ring on the platform. In another aspect of the present invention, the polishing member is transparent. 20 In another aspect of the invention, the polishing member includes a window. In another aspect of the present invention, the sensor is a light sensor. In another aspect of the present invention, the sensor is an acoustic thickness sensor. In another aspect of the present invention, the sensor is a color sensor. 17 200400098 (ii) Description of the invention In another aspect of the present invention, the structure of. The sensor is attached to a fiber optic thread which is movable in another aspect of the present invention. In another aspect of the present invention, the sensor is an optical fiber cable, and the work piece is substantially stationary but can be rotated and moved in translation during polishing. In a preferred aspect of the present invention, the translational movement is smaller than the area of the pressure region. Benefits of this book include the ability to ideally polish work pieces, saving you time and money. The drawings briefly explain the foregoing and other characteristics, viewpoints, and advantages of the present invention. The following detailed description will become more apparent when reading the following drawings, in which: Figure 1A is an exemplary substrate in which materials are deposited. A cross-sectional view after the substrate surface; FIG. 1B is a cross-sectional view of the exemplary substrate of FIG. 1 a after being subjected to a conventional chemical mechanical polishing (CMP) process; FIG. 1C is an exemplary Cross-sectional view of the substrate after the insulating material is deposited on the substrate surface; FIG. 1D is a cross-sectional view of the exemplary substrate of FIG. 1C after undergoing a conventional chemical mechanical polishing (CMP) process; The figure is a cross-sectional side view of an exemplary CNU system including an exemplary endpoint detection system for processing a work piece such as a wafer according to one of the preferred embodiments; FIG. 3 is a diagram of FIG. 4 A cross-sectional top view of an exemplary chemical mechanical polishing (CMP) system at 18 200400098 Angstroms, the description of the invention, and an exemplary control system for an endpoint detection system according to the viewpoint of the present invention; FIG. 4 A cross-sectional side view of an exemplary chemical mechanical polishing (CMP) system including the exemplary endpoint detection system of FIG. 2; FIGS. 5A-C are views of a work piece surface; and FIG. 6A is a diagram illustrating the present invention. A work piece processing system of a specific embodiment; FIG. 6B is a work piece processing system illustrating another specific embodiment of the present invention; and FIG. 6C is a view illustrating a specific embodiment of the present invention. Work piece processing system; Figures 7A-B are diagrams showing a platform of a specific embodiment of the present invention in Figures 6A_6B; Figure 8 is an exploded 15 view of an inductor of a specific embodiment of the present invention; 9A-BB1 is a series of pressure profiles obtained by using the present invention; Figures 10A-C are illustrations of a working example of a particular embodiment of the present invention; a work piece is polished; The specific embodiment of the present invention polishes the I-piece, which shows different force vectors depending on the shape of the work; FIG. 12 is a platform according to a specific embodiment of the present invention, which has a shock-absorbing buffer layer ; And Figure 13 AB is a diagram showing # 点 丨 Tian Shiren's external carcass is an example of bismuth, which is used to change the amount of pressure curve by applying pressure from the back of a work 19 200400098 玖, the description of the invention; Figure VII is a specific implementation of a color sensing device For example, it is used to detect a processing end point of a multilayer semiconductor wafer, wherein the color sensing device includes a light source, a color sensor, and a judgment circuit; FIG. 15 is a diagram for detecting a multilayer semiconductor crystal A method for processing an end point of a circle, a flowchart of one specific embodiment thereof; FIG. 16A is a top view illustrating a specific embodiment of an end point detection device for local end point detection, which includes a movable FIG. 16B is a side view illustrating a specific embodiment of an endpoint detection device for local endpoint detection, which includes a movable structure and a sensing device; FIG. 17A A specific embodiment of an endpoint detection device disposed in an exemplary chemical mechanical polishing (CMP) device is shown, wherein the chemical mechanical 15 polishing (CMP) device includes a carrier head, a polishing structure , An end point detection device, and a method, and the chemical mechanical polishing device is in a polishing mode; FIG. 17B illustrates a specific embodiment of the end point detection device disposed in a model chemical mechanical polishing (CMp) device. The chemical mechanical polishing (CMP) device includes a carrier head, a polishing member, an end-point detection device, and a road, and the chemical mechanical polishing (CMp) device is in a non-polishing mode; FIG. 18 is A specific implementation of a method for a processing end point in a chemical mechanical polishing (CMP) device for detecting a multilayer semiconductor wafer 20 200400098 玖, a flowchart of an illustrative example of the invention, the device has a carrier head and a The member is polished, and wherein the semiconductor wafer is attached to a carrier head. [Embodiment 3 Detailed Description of Preferred Embodiments 5] As described below, the present invention provides a method and a system for in-situ endpoint detection for material removal processing, such as chemical mechanical polishing (CMP). The same reference numerals represent the same elements throughout the reference drawings. A. End-point detection system 10 FIG. 2 shows an exemplary chemical mechanical polishing (CMP) apparatus 100 including a polishing member 102 and a carrier head 104. The polishing member, for example, may be a polishing tape, a polishing pad, or another type of polishing member. The polishing member 102 includes an upper or processed surface 106 and a lower surface 108. The lower surface 108 of the polishing member is arranged on a support plate 109, such as flat 15 ', and is taut. The polishing member and the carrier head are configured so that the work piece surface is close to the polishing member, which can approach or contact the polishing member. In this embodiment, the polishing member 102 is a light-transmitting polishing member. A polishing solution 110 flows on the processing surface 106 of the polishing member 102, and the polishing member is moved over a group of rollers 112 in a single direction or both directions by a moving mechanism (not shown). In this embodiment, the polishing member is moved in both directions. The polishing solution 110 can be a copper polishing solution or an abrasive polishing slurry. The solution 110 may be fed onto the polishing member from one or both sides of the wafer, or it may be fed onto the wafer surface via the polishing member at the same time, or both. A wafer 114 to be processed 21 200400098 玖, description of the invention is supported by the carrier head 104, so one of the wafers 'front surface 116, after which it will be regarded as the surface', is completely exposed. The carrier head 104 can move the wafer vertically up and down, and simultaneously rotate the wafer 114 through a shaft 118. The surface 116 of the wafer 114 may have the structure shown in FIG. 1A with a copper layer 16 (including a seed layer and a deposited copper), which can be polished down to the barrier layer 18 (as shown in FIG. 1B). (Shown) while using the present invention to perform endpoint detection in situ. In this example, the lifting layer is copper (Cu) and the barrier layer is is a boat (Ta). The insulating layer 14 may be made of silicon oxide (SiO2) or a low-dielectric value or a very low-dielectric value material. In this embodiment, an endpoint monitoring device 120, preferably including an optical transmitter and a detector, is disposed below the polishing member 102. When the copper layer is polished down to the barrier layer 18 on the top surface 15 of the insulating layer (see Figs. 1A-1B), the end point monitoring device 120 detects the end point of the polishing. Barrier layer 18-After exposure and detection by device 120, processing is stopped. In an optional step, if desired, the process can be continued until the barrier layer 15 is polished down to the underlying oxide layer. As described below, Erming, the device 120 can be arranged in a chamber located in the platform 109. The device 120 of the present invention can be any optical monitoring device for monitoring changes in reflectance. Although copper is used as an example material, the present invention can also be used to remove other materials, such as conductors such as nickel 20 (Nl), palladium (Pd), surface (Pt), gold (Au), and oblique (Pb). ), Tin (Sn), silver (Ag) and alloys thereof, buttons (Ta), nitride buttons (TaN), titanium (Ti) and titanium nitride (TiN), and insulators and semiconductors. During processing, the wafer is turned 14 and brought into contact with the surface 116 by the processing surface 106 of the polishing member 102, and the polishing solution 110 flows on the processing surface 106 and wets the surface of the wafer 116 22 200400098 发明, description of the invention When moving. As shown in FIG. 3 is a plan view, and FIG. 4 is a cross-sectional view. The monitoring device 120 is arranged in a chamber 122 formed in the platform ι09. As shown in Figure 4, the top of the chamber 122 can be sealed by a transparent window. In this embodiment, the cavity 122 is made in an appropriate size and configured to accommodate the movement of the cracked elongated body along the cavity 122. The position of the chamber 122 is related to the relative positions of the wafers on the polishing member and the underlying platform. During the processing, a moving mechanism (not shown) can be used to move the monitoring device along the chamber to scan the half-diameter of the wafer. Due to the scanning action, different positions between the edge of the wafer and the center of the wafer are detected. The chamber can be enlarged beyond the center of the wafer, so when Tian Yen Yen rotates by sliding the monitoring device in the chamber, a scanning action is generated. For example, a wide spectrum can be read along the diameter of the wafer . This scanning procedure can be performed as a continuous process or in 15 processing steps. In this embodiment, a reflecting mirror 126 attached to the monitoring device causes the output photon number 128 to be projected on the wafer surface. The reflecting mirror 126 thus arrives at the monitoring device 120 with the reflected optical signal or the reflected optical signal. For alternative implementations, different types of monitoring devices, such as flexible microfibers, can be removed using a reflector, and the k number can be directly transported from the device to the copper surface. The device determines the end point, i.e., the barrier layer 18 is immediately exposed when the intensity of the reflected optical signal 13o changes (see Figure 1B). If the chemical mechanical polishing (CMp) processing is continued to remove the barrier layer, when the top surface of the insulating layer 14 is exposed to the storm, the intensity of the reflected signal will change again (see section 1B). Figure). The optical signals generated by or guided by the monitoring device have a wavelength range of 600-900 nm. The output optical signal can be generated by a transmitter of the device 120, such as a white light transmitter with a chopper or a light emitting diode or a 5 疋 laser. According to a presently preferred embodiment, a reflected optical signal is received by a detector of the device 120. An exemplary detector is a pyroelectric detector. The incoming optical signal can be passed through a bandpass filter (bandpass mter), which is used to eliminate all the wavelengths on the shell, leaving the wavelength of 10 detected by the detector. In this embodiment, the output and reflected signals advantageously travel through a polishing member that is transparent. Another alternate embodiment is configured with an array of multiple monitoring devices. The device is fixed in a radially formed chamber extending from the center of a flat plate (star shape). It is used to monitor the change of the signal on the surface of the wafer. Furthermore, it is possible to alternately distribute a plurality of monitoring devices along a single chamber. In this way, the monitoring device can rotate the center and center of the wafer surface According to an aspect of the present invention, the entire polishing member is made of a transparent material and does not require an additional window for endpoint detection. In this specific implementation, the polishing member includes a synthetic The structure has a top transparent abrasive layer formed on a transparent backing material. During the processing, an abrasive layer is in contact with the work piece and includes fine abrasive particles distributed on a transparent binder substrate. matrix). An exemplary linear polishing member structure used with the present invention may include a thin transparent abrasive coating 24 200400098 玖, description of the invention 'for example thick 5 μm to ⑽ μm is stacked on a transparent dense substrate material division, which is sold by Mipox Company in Hayward, California. The thickness of the abrasive layer can be ㈣m to ㈣m, and The thickness of the backing layer can be from 0 to 5 hairs. The size of the ground 5 particles in the grinding layer is about 0. In the range of 2 to 0.05 microns. Exemplary materials used for the granules can be oxidized stone, oxidized oxide or thorium oxide. -A less transparent polished member that can still be used in the present invention and is sold by the 3M company in Minnesota. Although the polishing member may include abrasive particles in some embodiments, the polishing member can also be made of a transparent polymeric material without abrasive particles. As explained above, when the abrasive polishing member removes material from the wafer surface and when the barrier or oxide layer is exposed, the reflected light intensity is modified. In the example, a transparent polishing member has a use of about 10 microns Thickness of the abrasive layer and thickness of 0. 5 to 0 mm transparent Mylar layer. In this example, the abrasive layer has fumed oxide particles of 0.2 to 0.5 micron. A light beam (output) with a wavelength of 675 nm is passed through the polished member and the intensity change is monitored throughout the chemical mechanical polishing (cMp) process. Using this polished member, it was observed that the intensity of the light reflected by, ', and i was maintained at an arbitrary (standard) intensity value during the entire copper recording process. 20 However, once the barrier layer (button layer) is exposed, the strength decreases to}. Furthermore, when the early barrier layer is removed from the top σρ of the oxide layer and the oxide layer is exposed, the intensity of the reflected light is reduced to 0. 5. As shown in FIG. 3, in a preferred embodiment, the monitoring device 120 is connected to a computer 132, and the computer can be electrically connected to a vehicle 25 200400098 at the same time. No) connection, although it should be understood that there can be multiple ways, τ / Jue, but it is not necessary to have a computer with a processor, but can use discrete or integrated logic circuits instead, including but not limited to AS X and programmable gate array. When working on a copper layer with an early p-blocking layer underneath, once the barrier layer is exposed, the wheel-out signal from the I test device is changed due to the change in reflectance, and the mechanical polishing process is terminated. Generally, the end point detection device and method according to the viewpoint of the present invention are used on more work pieces to detect one or more end points on each work piece. For example, according to an aspect of the present invention, a chemical mechanical polishing (CMP) endpoint detection process is performed on a single work piece, such as a wafer, and a plurality of chemical mechanical polishing (CMP) endpoints can be detected. The end point of the chemical mechanical polishing (CMP) can have an individual polishing sequence and a matching individual processed K-shaped object. The removal of overloaded metals from the surface of the wafer can represent a first end point of the chemical mechanical polishing (CMP). 15 The barrier layer outside the features of the wafer may represent a second chemical mechanical polishing (CMP), ..., dot. The first threshold or degree of intensity k can be used to detect the end point of the first chemical mechanical polishing (CMp), so when the signal intensity observed by the detection system drops to or below the first low The time limit or extent is determined to have reached the end of the first chemical mechanical polishing (CMp). Other thresholds or degrees of intensity 20 can be used to detect other chemical mechanical polishing (CMP) endpoints. For example, for detecting a second chemical mechanical polishing (CMP) end point, when the signal intensity observed by the detection system drops to or below a second lower limit or degree (which is lower than the first lower limit or degree) ), It is determined that the second chemical mechanical polishing (CMP) end point has been reached. 26 200400098 发明. Description of the invention It should be understood that in the scope of the foregoing description and additional patent applications, the surface of the work piece and the 'surface of the work piece' include, but are not limited to, the surface and composition of the work piece before processing. The surface of any layer on the work piece includes conductors, oxidized metals, oxides, spin-on glass, ceramics, etc. B. Intelligent endpoint detection system is described below, the present invention Provides a method for in-situ material removal processing, such as chemical mechanical polishing (CMP), thickness uniformity control and an end point detection. In this system, the polishing member using a component such as a window or a transparent part can be Transparent, or partially transparent. Figures 5A-C are views showing the surface of a work piece. Figures 5 and 8 are diagrams showing a wafer 9 after a thin film 16, such as copper, has been deposited and covered. A circuit including a plurality of circles C in the wafer substrate 5 1 is shown for explanation, where n is an arbitrary value. Each of these 15 circuits includes a large number of features to deposit Conductive film filling, usually covered by a barrier layer. Chemical mechanical polishing (CMP) processing removes the overloaded part and leaves the conductive film in these features. However, it should be noted that there is a general variation in surface thickness A process such as chemical mechanical polishing (cMp) is used to remove the overloaded part when it is removed. Due to surface changes, a 20 process simply polishes the film U to a predetermined thickness, which may cause over-polishing of a specific area The other areas will be insufficiently polished. Figure 5B shows the local surface change on the wafer 114, which is slightly enlarged for illustration. As mentioned above, due to the surface change, a processing is simply a thin film 16 Polishing to a predetermined thickness may result in 27 200400098 玖, the invention shows that specific areas are over-polished and other areas are under-polished. Figure 5C shows a wafer with the desired polishing end point, where the conductive layer is It is located in the feature and the overload part is removed. In a specific embodiment, the thickness consistency detection and control system of the present invention uses it Thickness measurement capability and control of its covered processing parameters to maintain consistent thickness of the processed surface. Based on real-time thickness data from the surface of the processed wafer, the thickness during a chemical mechanical polishing (CMP) process Consistent control system changes polishing parameters to uniformly and uniformly polish a layer. Therefore, reaching the end point of the polished layer generally covers the wafer surface without causing excessive polishing and insufficient polishing of the target layer. By locally changing the polishing member The pressure underneath changes the polishing parameters, so the polishing at a specific position is faster than other positions. In one aspect of the present invention, the present invention maintains the consistency of the processed surface by using the detected immediate endpoint data. According to the self-processing The real-time data obtained on the 15th surface of the wafer, the polishing system changes the polishing parameters during the uniform polishing of the chemical mechanical polishing (CMP) layer. Although copper is used as an example material, the present invention can also be used to remove other materials, such as conductors, such as nickel (Ni), palladium (pd), platinum (Pt), gold (Au), lead (Pb) , Tin (Sn), silver (Ag) and its alloys 20 'tantalum (Ta), nitride group (TaN), titanium (Ti) and titanium nitride (TiN), as well as insulators and semiconductors. FIG. 6A shows an exemplary chemical mechanical polishing (CMP) apparatus 550 having a control unit 560 having a uniform thickness. The chemical mechanical polishing (CMP) apparatus may further include an abrasive polishing member 102 and a carrier head 28 200400098. Description of the invention. The polishing member 1 () 2 includes an upper or processed surface ig6 and a lower surface 108. The lower surface 108 of the polishing member is arranged on a push plate 6⑼ like a flat 3: and is taut. The polishing member preferably comprises a synthetic structural system-a transparent topped abrasive layer is formed on a transparent substrate material. During processing-the abrasive layer is in contact with the work piece and includes fine abrasive particles distributed in a transparent adhesive substrate. An exemplary linear polishing member structure for use with the present invention may include-a thin transparent abrasive coating ^ 1 stacked on a transparent Mylar backing material such as a thickness of 5 microns to ⑽ microns The materials are sold by 10MiPox, a California-based company. The thickness of the abrasive layer may be 5 μm to 100 μm ′ and the thickness of the substrate layer may be 0.1 μm. 5 to 2 directors. The size of the abrasive particles in the abrasive layer ranges from about 0.2 to 0. 5 micron range. The flat chamber includes a plurality of holes 620a-620n, which are shown in detail in Fig. 6B (also seen in Figs. 7A-7B), and are used to generate a -fluid pressure below the polished member during processing. If a chemical mechanical polishing (CMP) slurry or polishing solution including an abrasive is used, the polishing member 102 may be replaced with a non-abrasive polishing member. The holes 62o and 62o11 are fluidly connectable to one of those supplied through the fluid supply unit 562. In this embodiment, the polishing member 102 may be a light-transmitting polishing member, but at the same time, it may be a polishing member 20 having a window or a light-transmitting portion. In one aspect of the invention, the fluid supply unit 562 includes a rotating flow meter system to control fluid flow to a flat level. For example, fluid flow to each platform area can be controlled at 0 to 5 cubic feet per minute (cfm). Alternately, fluid flow can be controlled and measured by electronic mass flow controllers sold on the market. The 29 200400098 玖, description of the invention The electronic mass flow controller can be controlled by software and automated. Exemplary mass flow controllers are sold by SMC and Celerity. The polishing member is selected to have sufficient flexibility to comply with the applied pressure and communicate with a local pressure associated with a pair of wafer surfaces. The exemplary embodiment 5 uses a flexible polymer polishing member to fully transmit pressure to a local area. If the flexibility of the polishing member is insufficient, for example, reinforced with a steel strip, the pressure will cover a large area, and the system can continue polishing the undesired area of the wafer. A polishing solution 112 is moved on the processing surface 106 of the polishing member 102 and the polishing member is moved above a set of rollers 113 in a single direction or both directions by a moving mechanism (not shown). In this embodiment, the 'polishing member' is preferably moved in both directions. The polishing solution U2 can be a copper polishing solution or an abrasive polishing slurry. The solution 2 can be fed to the polishing member from one or both sides of the wafer, or it can be fed to the surface of the wafer through the polishing member at the same time, or both. A wafer 114 to be processed is supported by the carrier head 104, so one front surface 116 of the wafer will be regarded as the surface and is completely exposed. The carrier head 104 can move the wafer vertically up and down, and at the same time rotate the wafer through a shaft 118. The surface 116 of the wafer 114 may initially have the structure shown in Figure 5-8. There is a copper layer 16 (including the seed layer and the deposited copper), which can be polished down to an end point (as shown in Figure 5C). ), While performing the following thickness consistency detection and control processing of the present invention. Here, the barrier layer removal step is used for continuous processing, so the barrier layer on the top surface 15 of the insulating layer is polished until the insulating layer 14 is exposed or reaches the end of the barrier layer. In this example, overload 30 200400098 98, description of the invention The layer system is copper (Cu), the barrier layer 18 is a button (Ta), and the insulating layer 14 is a silicon dioxide (Si02). The consistency control unit includes a fluid supply unit 562 for delivering fluid (e.g., 'air') to the platform 600. The consistency control unit also includes a computer controller 564 with a central processing unit (CPU), memory, display, keyboard, and other common components. The computer 564 is combined with a series of exemplary sensors 630a-630n. Among them, n is an arbitrary sensor identification symbol (63b, 63b (1 simultaneously displayed on 6B and 7A) through a sensor controller 566. -7B). The sensors 630a_630n are arranged in the platform 10 and adjacent to the fluid holes 620a-620n in the platform. In this specific embodiment, the holes of the platform are preferably grouped in a specific way, for example, in a ring Shape the hole of the female knife group (see Figure 6B, 7A_7B). Demonstration sensors can include thickness sensor and end point sensor. As explained below, each group of holes (known as the pressure area) ) Is connected to the fluid supply unit, and conveys 15 the fluid pressure controlled by the computer controller 564. The fluid supply unit can change the fluid pressure (such as fluid flow) for each pressure region independent of each other. In one aspect of the present invention, The sensor 63〇α · 63〇η is an end sensor. The sensor includes an optical emitter and a detector arranged under the polishing member. 2 For example, when the copper layer is polished down to the top of the insulating layer When the barrier layer on 15 is 18, the end sensor detects the polishing end point (see Figures 1A-1B). As explained above, the present invention can control the local pressure from different areas of the platform to increase or decrease the Local polishing rate on a circle. Therefore, one of the key aspects of the present invention is the ability to provide different polishing rates for different M-force regions by using 31 200400098 玖 on the platform. The invention is fully controlled at each The degree of fluid or air pressure on another pressure region improves the polishing sensitivity of the system. Establishing an accurate control of the pressure level for the pressure region, in turn, leads to better control of the local polishing rate on the wafer. 5 as described in [6] 3 and 6 <: As shown in the figure, in a preferred embodiment, at the same time, the separated pressure regions having a predetermined pressure level can be achieved by removing excess air from the top of the plate. As explained in more detail below, by allowing controlled leakage to the atmosphere or a vacuum source, the present invention regulates the flow of excess air that covers adjacent pressure zones, that is, adjusts crosstalk interference between adjacent 10 zones (cross_talk ) And cause a change in the degree of air pressure in the adjacent area. As shown in Figure 6B, in a specific embodiment, the exemplary system 1000 has an air relief valve. In this embodiment, the computer controller and the sensor unit are not shown for clarity. The system is mainly composed of a platform 600, a wafer carrier 104 for supporting the wafer 114 15 to be processed, and a polishing tape 102 or a polishing pad. As explained above, the tape 102 has a top surface 106 or a machined surface, and a back surface 108. The front surface 116 of the wafer 114 faces the top surface of the polishing tape 102. The details of the polishing tape and the polishing solution mentioned above are taken as examples, so for the sake of clarity, they will not be repeated here. 20 Compared with FIG. 6A, FIG. 6B shows the platform 60 () in more detail. As shown in Figure 6B, the platform 600 may have an upper surface 61o surrounding a base block 612. The upper surface is divided into concentric pressure regions, that is, a first region z1, a second region, a third region, and a fourth region z4. The isocentric areas are exemplified in Figures 7A-7B. The area zl_z4 includes holes 32 200400098 玖, invention description 62〇a-62〇n. In FIG. 6B, each region may include two or more holes. For example, the first region Z1 includes a hole 6 and the like. 6 sensors _ 630η are arranged in each area at the same time. For the sake of clarity, the computer controller and sensor unit and the connection device connected to this unit are not included in the figure (see Figure 6A). Furthermore, each area in the surface 010 corresponds to an air chamber 614a-614d as shown in Fig. 6B. For example, the hole 620a in the first region z1 is fed by passing air through the chamber 61 乜, the hole 620b in the second region z2 is fed by passing air through the chamber 614b, and so on. The chambers 614a_614d are formed as annular concentric grooves, which are connected to an air supply unit 562 via air lines 616a_616d, respectively. Each air line 6116-616d is connected to a corresponding chamber via one or more air ports 618a-618d. Furthermore, by using a connector, such as a T-connector, each air line 6i6a-616d is separately connected to the pressure control device 622a-622d. In this embodiment, the pressure 15 control device is an air valve 622a-622d connected to the air lines 616a-616d. In this regard, each valve system is combined with one of the pressure regions, for example, 'the first valve 622a is used for the first region; 21, and the second valve 622b is used for the first region z 2 and the like. Valves 622a-622d include exhaust ports 624a-624d. Exhaust ports 624a_624d 20 can be connected to the outside atmosphere or vacuum (not shown) for removing exhausted air from the system 1000. In this specific embodiment, the amount of air that can be discharged from the exhaust ports 624a-624d can be adjusted via these valves, thereby adjusting the positive pressure in a pressure region. When the valves 622a-622d are opened, it exhausts a certain percentage of the air flowing through the pipes 616a-616d. On 33 200400098, the description of the invention In this regard, the valves 622a-622d can be used in these areas to generate a positive pressure or a negative pressure or zero pressure. With a vacuum connection, a negative or zero pressure can be generated in the pressure region. However, the most important function of the valve combined with this area is that when excessive air flow from 5 adjacent areas covers the area and causes the air pressure in the area to increase, the exhaust air is used to adjust the pressure level in a pressure area. In this specific embodiment, the air supply unit can supply air to each pressure region with the same air flow rate, and the flow rate is changed to individual pressure regions to establish a predetermined air 10 below the polishing belt 10 The air area of the pressure curve. In another specific embodiment shown in FIG. 6C, the platform 600 includes a fluid discharge hole 1400, which is preferably disposed between these areas to remove excess fluid from the top of the plate. By allowing the fluid to leak to the atmosphere or a vacuum source through the fluid discharge hole, the present invention eliminates the problem that excessive fluid will flow to cover these adjacent pressure regions, and thus substantially reduces crosstalk interference between these adjacent regions. To the lowest. In this way, in turn, substantially independent pressure zones are generated on the platform, which advantageously allows different pressure levels to be used in each pressure zone. As shown in FIG. 6 (:), in one example, the fluid discharge holes 1400a · 1400d are arranged between the concentric pressure regions η, z2 20, z3, and z4, and these regions have fluid holes 62〇a- 62〇d and sensors 630a-630d. Between each area, a plurality of discharge holes are formed on a single or more than one circular path. Each circular path may have at least one pipeline and its system There are plural release holes 1400a-1400b. For example, the plural release holes 1400a located between the areas z1 and z2 may follow a circular path of a single 34 200400098 玖, invention description 1, or include multiple releases. The holes are formed by two concentric circular paths. Although, in this embodiment, the discharge holes are formed along the circular path and interposed between these areas, they can be distributed in any manner, such as radial, and are covered in Within the scope of the present invention. In this specific embodiment, the 5 hole is made into a circle or a ring in an appropriate shape; however, it can have a rectangular or other geometric shape or it can have an appropriate shape. Made into a circular slit. During a mechanical polishing (CMP) process, a fluid such as air is sprayed under the polishing tape 102 through the fluid holes 620a-620n located in each area, while the carrier head 104 supporting the wafer 114 is lowered 1 〇 to polishing f. When the polishing belt 102 is moved above the platform 600, the fluid passes pressures under the polishing belt 102 through the holes 620a-620η. The release holes between the pressure areas will be from the pressure areas zl-z4 Excessive fluid flow out 'and prevent crosstalk interference between these areas. During processing, the wafer 114 can be translated at least about twice the diameter of the discharge hole to balance the localization of the 15 discharge holes. Effect. Each vent hole can be opened to atmospheric pressure, or can be connected to a vacuum system (not shown). As shown in Figure 6 (: shown in the figure, in this embodiment, each fluid vent hole 1400a-1400d is Individually connected to atmospheric pressure. The female-releasing holes 1400a-1400d are independently opened to communicate with external pressure, and individually release excess fluid to the atmosphere. However, the most important function of the 20-releasing hole is to independently adjust each pressure Area The degree of pressure. For example, the magnitude of the pressure in the first region can be higher than the adjacent pressure region by feeding a higher flow rate to zl and discharging too much fluid out of the first region ζι through the discharge hole 1400a. The pressure of z2, so it will not affect the pressure in z2. In this specific embodiment, the air supply unit 35 200400098 发明, invention description can simultaneously supply air to each pressure region with the same air flow rate, and change to individual pressure The area flow rate is used to establish a specific pressure level in each fluid. Thus, a curve with a predetermined amount of air pressure is generated under the polishing belt 102. 5 Figure 7A_7B is a plan view showing a plane 610 having a region zl_z4, including holes 62a_62〇n and sensors 630a_63〇n. In this specific embodiment, the non-standard sensors 630a-630n may be optical end sensors, preferably including an optical transmitter and a detector, and arranged from the work piece in a platform located under the polishing member. For example, the sensors 63a-63n may sit at or near the area zl_z4, which represents a pressure area in which the fluid pressure is selectively controlled by the fluid supply unit 562. Although the exemplary optical sensor located in the platform is used in this specific embodiment, any type of sensor can be used in the system to be located at any suitable location, and is included within the scope of the present invention. As shown in Fig. 7B, each region 15 may include a plurality of concentric circles' and it is further expected that the region may not have a sensor under some conditions. The sensor unit 566 receives unprocessed sensor signals (eg, reflected light) and generates electrical sensor signals, which are sent to a computer 564 (see Figure 6A), which controls the fluid supply in the manner described above. Unit 562. 20 The end sensor of the present invention can be any optical monitoring device for monitoring the change in the reflectance of the polished layer. Referring to FIG. 8, each sensor 63Ox includes a transmission fiber 632x system for providing a light reflecting off the work piece 114 (see reference numeral 710), and a receiving fiber 634x system for receiving the reflected light. When, for example, the copper layer is polished down to the position of the insulating layer 36 200400098 玖, the barrier layer 18 on the top surface 15 of the invention description (see Figures 1A-1B), the end sensor uses the reflected light. The change detects the end of the polishing. In this regard, the output and incoming signals travel through the light-transmitting polishing member 102. The use of these sensors in chemical mechanical polishing (CMP) endpoint detection is disclosed in US Patent Application No. 10 / 052,475, filed on January 17, 2002. Chemical mechanical polishing (CMP) is a processing process that generally polishes a surface according to the following formula: Polishing rate = constant X speed X pressure 10 The present invention uses the ability to control local pressure to increase or decrease the local polishing rate. Therefore, one of the main aspects of the present invention is the ability to use different polishing rates in different pressure regions. An operation sequence is exemplified by using the pressure regions z1 and z2 to establish the pressure amount curve shown in FIG. 9A. It should be understood that the two regions are used for example purposes. A curve system similar to a pressure amount change curve in Fig. 9A can be constructed using the pressure regions zl, z2, z3, and z4. The pressure change curve shown in FIG. 9A can be established by having a high air pressure P1 in the first pressure region z1, but a relatively low air pressure in the second region 22 nearby. In operation, for example, this system may be performed by first establishing a pressure P1 in the first region 20 zl, which is established by the air supply unit flowing to the first region zl with a first predetermined amount of air. Under pressure? During the establishment, the -valve 622 is adjusted to discharge the -air-flowing portion from the center of the first pipe 6. Establishing a pressure p2 in the second zone z2, for example, 'can be accomplished by flowing the -th volume of air through the second air line 37 200400098 玖, the invention description, and at the same time by the first predetermined air Flowing through the exhaust port 624b reduces the pressure to P2. Any air flow from the first region to the second region here can increase the pressure in the second region to a pressure p3. According to the present invention, the degree of pressure increase in the second zone z2 is completely changed by discharging more air from the first predetermined flow through the second valve. As a result of the ventilation, the first flow volume guided to the first region is reduced, and the pressure level in the second region z2 is restored to the pressure level of P2. Perform the same process using different air flows for each area. In this case, the pressure level is adjusted again by discharging a predetermined amount of air flow. 10 Simultaneously using the area zl & z2 exemplifies another sequence of operations for establishing the pressure amount curve shown in Fig. 9B. A curve similar to the pressure amount curve in the second figure can be constructed by using the pressure regions z1, z2, and. The pressure change curve shown in Fig. 9B can be established by having a low air pressure P1 in the first pressure zone zl, but it is a higher air pressure in the nearby second 15 zone Z2. In operation #, for example, this is performed by first establishing a pressure P2 in the second zone z2, and the pressure is established by the air supply unit 562 flowing to the second zone with a first predetermined amount of air. During the build-up pressure p2, the second valve 622b is also opened or closed to discharge a part of the first air flow. Establishing the pressure P1 in the center of the first region can be accomplished, for example, by flowing a first predetermined amount of air through the first air line 616a, and by A predetermined portion is discharged to reduce the pressure to ρm. Here, any air flow from the second region z2 to the -th region zl can increase the pressure in the first region ^ to the pressure P3. As in the previous situation, the pressure in the first area Η 38 200400098 玖, description of the invention The increased degree of force is completely changed by discharging more air from the first predetermined flow through the first valve 622a. As a result of the ventilation, the first flow amount guided to the first region z1 is reduced, and the pressure level in the first region is restored to the pressure level of P1. The same process is performed using different air flows 5 for each area. Under this condition, the pressure level is adjusted again by discharging the predetermined air flow. The processes described in conjunction with Figures 9A-9b can be controlled dynamically at the same time. For example, these valves can be controlled or adjusted by the input of pressure sensors which are not arranged in each pressure zone 21_24 as shown in Fig. 6B. When the pressure in one area increases due to the flow of air 10 from the adjacent area, the valve member discharges a predetermined amount of air to adjust the air pressure in the area. By a controller receiving pressure input from the sensor, ventilation through the valve can be controlled. When working on a copper layer with a barrier layer below it, the barrier layer exits, and the signal from the endpoint sensor I5 changes due to the change in reflectance. With reference to the eighteenth _10 (:: diagram, in the exemplary processing, one area in the wafer may require more polishing than another area, or one area becomes thinner faster than the other-area, so For one area, the copper end point is reached more quickly than the other area. The copper end point—as detected by the end point sensor, the air pressure in the force area is reduced to make the 20% increase in the area. -Step polishing slows down or eliminates the need for further polishing. Alternately, air pressure can be increased in other areas that have not yet reached the end point. With different removal rates, copper is generally no longer removed in the completed area, and The other area can continue to be polished. Here, the viewpoint of the present invention is based on its state of being at the end point, and the Longli area applies different air dust force. An example of the end-of-line detection function. As shown in Figure 10B, the surface of the work piece is defined by the representative symbol 92〇a. After several polishing times, the surface is reduced to the representative symbol 92〇b, and is close to the sense This layer is extremely thin in the area near the sensor 630c. After more 5 polishing times, when the surface is polished down to the symbol 920c (920c-1 and 920c-2), the sensor 630c will detect changes in the surface and The controller 560 will reduce the pressure (fluid flow rate) to this area. As a result, this area will experience less polishing while the other areas continue to be polished at the original rate. Of course, if this is required, and the desired It is able to increase the flow of fluid 10 to the special unfinished area. Once all the areas are polished (all the sensors report that they have reached the end point), the processing is completed. Although the different preferred embodiments have been detailed above, it is well known It should be immediately apparent to those skilled in the art that multiple modifications can be made to the illustrated specific embodiments without substantially departing from the novel teaching content and advantages of the present invention. 15 c · Variations of the specific embodiments are one of the present invention In the viewpoint, an acoustic sensor can be used instead of the above-mentioned optical sensor. In this viewpoint, the sensor 63 such as _63011 detects the thickness of the polished layer in real time, and simultaneously adds Process the wafer and supply this data to the computer via the sensor unit 566. The computer 564 thus identifies the supplied thickness data 20, if unevenness in the removed layer is detected, and changes are made on the wafer-or More polishing parameters, such as the pressure of the working gas under the polishing member or slurry composition, selectively adjust the material removal rate again to obtain the thickness uniformity of the entire wafer surface. In another aspect of the present invention, the 11th The picture shows the polishing-work piece 40 200400098 玖, description of the invention shows different pressure vectors 910a_910d depending on the characteristics of the work piece. The longer the arrow, the greater the power. If a work piece area requires more polishing, the computer controller indicates The fluid supply unit provides increased pressure to the area. Similarly, when an area does not require additional polishing, the computer controller 5 means that the fluid supply unit provides less pressure to the area. In another aspect of the present invention, a heat exchanger is combined with the fluid supply device in-line with the platform, so the temperature of the fluid delivered to the platform is controlled and can be maintained at a preset temperature. The platform can further include a temperature sensor to maintain a predetermined temperature of one of the polished components for providing feedback to the heat exchanger. D. Platform with buffer layer During a chemical mechanical polishing (CMP) process using a polishing member as described above, a plurality of factors may damage the polishing member, the wafer surface, or the former. As far as the wafer surface is concerned, any non-parallelism between the surface of the work piece to be polished and the surface of the 15 polishing member will cause damage to the surface of the work piece. Before any chemical mechanical polishing (CMp) processing, the platform surface should be aligned with the surface of the work piece to be polished, so it is generally parallel. Any significant deviation from this parallel will cause a part of the work piece to be closer to the platform surface. At the same time, another part of the surface of the work piece will be arranged away from the flat 20 surface. This surface part, or the so-called high point on the work piece, that is, closer to the surface of the platform, may be excessively polished or hit the surface of the platform, causing damage to the surface of the work piece and causing damage to the polishing member at the same time. This misalignment, i.e., non-parallel, between the platform and the surface of the work piece, causes particular damage during polishing of the low dielectric material containing the substrate. Due to the low dielectric value 41 200400098 玖, the easy phase structure of the material, any collision between the polished substrate and the platform can completely damage the structure of the low dielectric material. π 3, Luo Ruogu 5 W π Black polishing, any large particles between the scale and the platform will scratch or damage the thin fixed abrasive polishing member. Furthermore, the end window should be smoothly aligned with the surface of the platform. Any significantly misaligned epicenter will form a bump on the surface of the platform and will scratch the polishing members or damage the work piece. These problems can be avoided by using a shock absorbing medium in combination with the platform described herein. In one example, the shock-absorbing medium is a shock-absorbing buffer layer between 10 polished structures and the surface of the platform. The specific embodiments described herein can include any combination of a platform, a polishing member (with or without abrasive), and a polishing solution (with or without slurry). Figure 12 shows a platform 600 having a shock absorbing layer 1300 attached to the top of the platform surface 610. The buffer layer test system can be softly aggregated. Made of materials such as polyurethane or any material that can withstand the chemical environment of chemical mechanical polishing (CMP) processing. The buffer layer 1300 may have a first hole 1320a-132n of the same type as the platform fluid holes 620a-620n, and a second hole 1330a-133n of the same type as the sensors 630a-630n. In this embodiment, the size of the holes 1320a] 320n may be larger than the size of the fluid holes 20620a-620n. During chemical mechanical polishing (CMP) processing, the holes 1320a-1320n allow a fluid, such as air, to be sprayed below the polishing member 102, while the carrier head 104 supporting the wafer 114 is lowered onto the polishing member. The polishing member is therefore preferably tied above the platform and includes a buffer layer that moves linearly in two directions. Of course, the polishing member can be moved in other directions, such as a circular direction. When the polishing member 102 is moved above the buffer layer 1300, a fluid pressure is applied below the polishing member 102 via the hole 1320. The buffer layer allows fluid to be distributed throughout and covers the platform, but additionally provides safety to avoid accidental contact between the hard surface of the platform 5, the polished components, and the wafer surface. The present invention has led to specific advantages in chemical mechanical polishing (CMP) processing of fragile, low dielectric and very low dielectric materials. The soft buffer layer absorbs any transient shocks to the wafer and minimizes damage to low dielectric materials. 10 15 20 In addition to the foregoing specific embodiments, this specific embodiment provides an improved chemical mechanical polishing (CMp) process for a substrate of a low dielectric material. Although the use of fixed abrasive polishing members with traditional casters can provide lower excessive dishing and depression, the hard surface on the fixed abrasive polishing members is used when On substrates of low dielectric materials, higher defects or local delamination can occur. As mentioned earlier, the use of low dielectric materials in steel metallization processes-is extremely fragile and has poor adhesion. Controlling the friction and number between the substrate and the polishing member is important to prevent low-dielectric ㈣ stratification from occurring between different chemical mechanical polishing (⑽) steps. Technical challenges related to the overall strength of low dielectric materials and the damage caused by mechanical mechanical polishing (CMP) in the integration of copper / low dielectric materials can be reduced or even eliminated using the protection of the present invention. Polishing solution. However, the conventional technique of using a solid polishing material in one of the processes of the present invention can use 43 200400098 without a liquid substrate. An exemplary description of the invention is that a copper layer system can be removed by a fixed polishing component, A polishing solution containing a predetermined amount of slurry is delivered to the fixed abrasive polishing member. These increased particles lubricate the surface of the polished member and reduce the lateral force on the surface of the polished substrate. Exemplary particles include, but are not limited to, alumina, hafnium oxide, silica, or any other metal oxide or polymer resin beads. The concentration of particles in the polishing solution can be from a weight ratio from 0. 1 to 40%, more preferably from 0 to 5 to 5%. An exemplary polishing solution can be prepared by adding oxide particles or silicon oxide particles to a copper polishing solution, such as the CPS-11 solution sold by 3M Company. 10 E. Multi-Layer Polishing In another embodiment, the removal of copper and barrier layers can be performed on individual polishing members used in individual chemical mechanical polishing (CMP) stations in an integrated chemical mechanical polishing (CMP) tool. In the first chemical mechanical polishing (CMP) station, in the first processing sequence, the copper layer of the substrate is removed by solid grinding and polishing and a polishing solution containing particles. The polishing process can be performed by using the shock absorbing buffer layer 1300, which is described with reference to FIG. 12 in the foregoing specific embodiment. During processing, a system similar to that shown in Figure 12 was used to lower a wafer onto a fixed abrasive polishing member and distribute a polishing solution containing lubricant particles to 20 polishing members. As described above, the fixed abrasive polishing member is moved above the buffer layer 300, and a fluid pressure is applied under the polishing member. Once the copper layer is removed down to the barrier layer on the surface of the low dielectric material (see FIG. 1B), a barrier layer removal process is performed in a second chemical mechanical polishing (CMP) station. In this step, a chemical mechanical polishing 44 200400098 shown in Figure 12 is used. (CMP) The system can use a polymerized / non-fixed abrasive polishing member. The polishing member may be made of a soft polymer material such as polyurethane. In this example, during the removal of the barrier layer, a selective polishing solution is dispensed onto a polymeric polishing member suitable for the removal of the barrier layer material, while moving the 5 polishing members and under the polishing member as described above. Apply a fluid pressure. This sequence of processing steps minimizes stress on low dielectric materials and minimizes the resulting delamination, and minimizes dish-like depressions and depressions. In another embodiment, the removal of copper and the barrier layer can be performed in the same chemical mechanical polishing (CMP) station. The first step is performed by removing the copper before removing the barrier layer. According to this processing sequence, in the first step, a large amount of copper can be removed downward on the fixed abrasive polishing member until the barrier layer on the fixed abrasive polishing member is removed. At this step, the polishing solution may or may not contain particles. In a second step, a fixed abrasive polishing member and a polishing solution with particles are used to remove the remaining copper layer from the surface of the barrier layer and apply a downward force on the work piece, for example, it can be a relatively low Downward force. Following the steps such as Rong, in another chemical mechanical polishing (CMp) processing station, a barrier layer removal step is performed on a soft polymerized polishing member, and a button selective polishing solution is distributed to the polishing member, and at the same time, Apply a low downward force on the work piece. 20 F • Pressure change of the carrier head Figures 13A-B show a specific embodiment of the change pressure curve, which is applied by applying pressure from behind the wafer 114. In this specific embodiment, the head 104 is used to apply a pressure gradient to the wafer 114 while supporting the wafer at an appropriate position. A flexible or swellable film 121〇 is in conformity with the shape of the carrier head 45 200400098 玖, description of the invention, typically a ring shape, and is attached adjacent to the inside of a raised surface area. The expanded film 121 provides a compliant wafer support during processing. The expanded film 121o is constructed of a thin, compliant material, such as an elastomer, preferably viton @. The film is preferably attached to the carrier head 104 using a combination of an adhesive and a fastener or clamping mechanism. This attached structure, when expanded, supports and seals the film 121. Although a non-limiting embodiment is described as an expanded film, the film may alternatively be constructed of a flexible, but not necessarily, expanded, compliant material. If the film is non-intumescent, a sponge type 10 material can be used to force the wafer against the polishing member. Referring to FIG. 13A, the thin film 1210 is divided into a plurality of regions 12110 and 1210e, which may have any number of regions. In order to apply a pressure gradient to the work piece, the supply fluid enters these areas and can be discharged from these areas at the same time. As explained below, a 15-flow system from the fluid line 1224 core 1224 itself is used to expand the expandable film 1210 and maintain the expansion through the processing that occurs. During processing, the pressure applied by the film is preferably in the range of 0.1 to 10 psi. One of a number of methods can be used to support the wafer in place during processing. As shown in Figure 13B, one of the methods is to use a fastener 1212a-1212b. The fastener 1212 preferably supports the wafer in a fixed position, and at the same time does not hinder the surface processing. Another technique for supporting a wafer in place is to use a vacuum between the wafer and the film, similar to the way described in U.S. Patent Application Serial No. 10 / 〇43,656, here This article is incorporated as reference material. In the operation of 2004200498 and the description of the invention, after the wafer 114 is arranged on the film 1210, the substrate member is expanded until the lower layer contacts the film 1210. The carrier head chamber is then evacuated and a vacuum is applied to the wafer 114. When a vacuum is applied to the chamber, the connection area between the bag-like portions or the low-concave portion provides a low-pressure space, and thereby the adjacent film portion collapses into the low-concave portion. In this way, a plurality of low-pressure spaces are sequentially generated on the back surface of the wafer 114. This low pressure space acts like suction cups and provides sufficient suction power to hold the wafer during processing. During polishing, it is connected to a pressure 10 controller 1220 through individual pressure lines 1224a-1224e. These pipes allow the pressure controller to generate a variable pressure gradient at the back of the wafer, so the consistency of the film removal rate that already exists on the surface before the wafer can be determined by using the Different pressures are controlled. For example, a higher pressure is applied to the center of the wafer and a smaller pressure is applied to the periphery of the wafer. Compared with the mechanical elements located at the periphery of the wafer, the mechanical elements are processed at the center of the wafer. The system is significantly increased, increasing the material removal rate from the central area. FIG. 13B also shows a platform 160, which is similar to the platform 600 described above, or may be a flat surface with a polishing member fixed on it. In this aspect of the invention, the relative movement between the wafer and the polishing member is achieved by moving the polishing member, the carrier head, or both. In any case, the substrate surface monitoring sensors 63a-630n are installed in the platform, and the wafer is monitored through the polishing member or through an opening in the polishing member. The sensor located in the platform 1600 is similar to that shown in FIG. 6A, and is connected to a sensor unit 566 and a computer controller 564. The computer controller controls the pressure 47 200400098 玖, description of the invention The force controller mo provides feedback to the processing system, in order to control the pressure applied to each area on the work piece and to process the work piece optimally. The flow chart related to the HK: diagram illustrates the above. You can use this method to select the end point at different areas of the work piece at different times. 5 G · Sensing device with color sensor In a specific embodiment, a sensor used for endpoint detection of a multilayer wafer is a color sensor. In this context, the term, color, means at least one of the different characteristics of reflected or emitted light from the surface. Reflected light has characteristics of color change, such as multiple wavelengths. FIG. 14 is an exemplary embodiment of a color sensing device 1405 shown in FIG. 10, which is used to detect the m-terminus of a multilayer semiconductor wafer. The color sensing device includes a light source 1410 and a color sensor 142. 〇 and a decision circuit 143〇. The term "sensing structure" can be used interchangeably with nouns, sensing devices, and, as further explained below, the color sensor can be a single-wavelength sensor or a composite wavelength sensor ( Multi-wavelength sensor). Color sensing devices, for example, can be used in conjunction with a shallow trench isolation (STI) chemical mechanical polishing (CMP) process. Related to the above 1 (: and 10) provides an exemplary light Description of the trench mechanical isolation (STI) chemical mechanical polishing (CMP) procedure. In the exemplary embodiment shown, the light source emits incident light on one of the semiconductor wafers. The color sensor is optically combined with the light source and reacts. The incident light senses the reflected light from the surface of the semiconductor wafer and is a so-called reflected color. In one point of view, the color sensor can be a single wavelength sensor. The color sensor is configured to reflect the reflected color and is generated. A sensor k said that the judging circuit is combined with a color sensor and is configured to determine whether or not it has reached the wafer processing based on at least part of the sensor signal of 48 200400098 (invention description). In one aspect of the invention, the light source and the color sensor are located very close to the wafer. In another aspect, the light source is combined with an optical fiber. In this perspective, the light source includes the output end of the optical fiber. Similarly, The color sensor can be combined with a 5-optical fiber to sense the reflected color. In this view, the color sensor includes an optical fiber. As described above, a single-wavelength sensor can be replaced by a multi-wavelength sensor. The light source can emit multiple spectra Incident light, and the color sensor can sense a multi-spectral reflection. Multi-spectrum means having at least two wavelengths. In the tenth aspect of the present invention, the color sensor is configured to sense the wavelength range from 400 to 800 nm The extended light. On the other hand, the light source emits white incident light and the color sensor senses a red-green-blue (RGB) reflection. The decision circuit is configured to determine whether or not 疋 has arrived based at least in part on the sensor signal. Wafer processing end point. The judging circuit may include a comparator 15 to compare the reflection color from the surface of the semiconductor wafer against a threshold reflection color. Low-reflection color, for example, can be the reflection color from a sample semiconductor wafer that has reached its processing end point. In this regard, 'based on the comparison of the reflection color from the comparator to determine whether the processing end point has been reached. Reflection The color comparison data, for example, is more reflective than 20. In another aspect of the present invention, the determination circuit uses an algorithm to determine whether the end of the wafer processing has been reached. The low-reflection color can be determined by sensing a known The material's reflective color is initialized. In one view, the low-reflection color is based on a reflection from a SiO2 layer of a sample semiconductor wafer. In another view, 49 200400098, invention It is explained that the low-reflection color is based on the -reflection of a nitride nitride (Si3N4) layer from a sample semiconductor wafer. In another view, the upper layer of the wafer is copper (Cu) and the lower layer is a barrier layer, such as a button (Ta) or a nitride button (TaN) or a button / nitride button (Ta / TaN). ). In this view, the low reflection color 5 may be based on a reflection from a sample semiconductor wafer that has been polished to a barrier layer. Alternatively, the low-reflection color may be based on a reflection from a copper layer of a sample semiconductor wafer. Furthermore, in an alternate manner, the low-reflection color can be based on a reflection from an insulating layer of a sample semiconductor wafer. In a further perspective, one layer of the semiconductor wafer is hydrophilic 10 (hydroPhllie) and the other layer is hydrophobic (hydrophobic). (Hydrophilic means that it is easy to hold water, while hydrophobic means that it is not easy to hold water). For example, one of the upper layers of the wafer may be composed of hydrophilic silicon dioxide. One of the lower layers of the aa circle is a hydrophobic nitride. Since the fragmented dioxide layer is hydrophilic, a thin water film is typically formed on its surface. 15 J and s A shallow trench isolation chemical mechanical polishing process, when polishing a wafer down to a silicon nitride layer, typically has little or no moisture on the surface of the nitride. The absence of moisture on the surface of silicon nitride is a consistent measurement considering the end of the process. As described above with respect to FIG. 14, the sensing device can be used in combination with shallow trench 20 CMP. When a semiconductor wafer subjected to shallow trench isolation chemical mechanical polishing is polished from the silicon dioxide layer 55 to the silicon nitride / silicon dioxide interface (refer to Figures 1C and 1D), the reflection color is slightly green (usually 4-5 kA) to yellow or purple. In this example, the silicon nitride / silicon dioxide interface represents the end of processing. Therefore, referring to FIG. 14 again, the color sensing device 50 200400098 玖, description of the invention By monitoring when the reflected color changes from slightly green to yellow or purple, it is possible to detect when a shallow trench isolation chemical mechanical polishing process has successfully arrived Processing end point. Previous shallow trench isolation chemical mechanical polishing techniques were demonstrated, and other techniques were expected. 5 The color sensor allows changes in the sensing angle and sensing distance, that is, the distance from the color sensor to the surface of the semiconductor wafer. At one point of view, the 'color sensor is arranged at a sensing distance that takes into account the most ideal optical signal. For example, the sensing distance may be 2-10 mm. The sensing device is operable to perform endpoint detection on the semiconductor wafer 10 at a predetermined frequency. For example, the sensing device may be tested every 50th wafer to determine the accuracy of a wafer polishing process. Fig. 15 is a flowchart of a specific embodiment of a method for detecting a processing end point of a multilayer semiconductor wafer, for example, using a color sensing device 1405. In step 1510, incident light is emitted toward a surface of a semiconductor wafer. In step 1520, a reflected color is sensed from the surface of the semiconductor wafer in response to incident light. In step 1530, a sensor signal is generated based on the sensed reflected color. In step 540, it is determined whether the end of the wafer processing has been reached based at least in part on the sensor signal. The use of a color sensor can reduce or eliminate problems associated with other types of photoresistors, such as limited differential capacity and inability to compensate for changes in target distance. An exemplary color sensor that can be used with the present invention is sold by Keyence & Division of Woodcliff Lake, New Jersey. H. Movable structure for local endpoint detection 51 200400098 000, description of the invention For local endpoint detection, a sensing device can be combined with a movable structure. Due to the combination of the sensing device and the movable structure, it is not necessary to remove the semiconductor wafer from its processing support, that is, the carrier head, and remove the end point detection on a semiconductor wafer. 2)). In a specific embodiment, an endpoint detection system includes a sensing device configured to sense a metric related to the surface of a semiconductor wafer, and to generate a sensor signal based on the metric. The system also includes a determination circuit combined with the sensing device and configured to determine whether the end of the wafer processing has been reached based at least in part on the sensor signal. The system further includes a movable structure coupled with the sensing device, and the positioning sensing device is used to sense the metric. The sensing device may include, for example, the light source 1410 and the color sensor 1420 described above in relation to FIG. 14. In this view, the light source and color sensor are combined with a movable structure to sense the reflected color from the surface of the semiconductor wafer. In another aspect, the light source and color sensor 15 are combined with a movable structure to scan the surface of the semiconductor wafer. In another aspect, the movable structure positions the color sensor to sense the reflected color. The sensing device also includes a decision circuit 1430. Alternatively, the sensing device may be a sensing device different from the sensing device 1405 described above in relation to FIG. 14. 20A is a top view illustrating a specific embodiment of an endpoint detection device 1610 for local endpoint detection, which includes a movable structure 1620 and a sensing device 1630. The movable structure is combined with the sensing device and enables the sensing device to be positioned at different positions. For example, the movable structure can position the sensing device in an active position (sensing position), or it can be inactive. 52 200400098 玖, description of the movable position (non-sensing position). Sensing device detection uses the techniques described above, such as reflective color sensing 'to detect the end of a wafer process. At the same time, the endpoint detection technology can be used. The sensing device may include a photo sensor, such as the color sensor described above in relation to FIG. Related to FIG. 16A 5 again, the sensing device can be combined with a determination circuit to determine whether the end of the wafer processing has been reached based at least in part on the data generated by the sensing device. FIG. 16B is a side view of a specific embodiment of an end point detection device 161 0 for local end point detection, which includes a movable structure 1620 and a sensing device 1630. 10 Figure 17A illustrates an end point detection device 1710, which is located in an exemplary chemical mechanical polishing device 1700. The chemical mechanical polishing device includes a carrier head 104, a polishing member 102, an end point detection device 1610, and a track 1730. And the chemical mechanical polishing device is in a polishing mode. The lane provides a path for the end point detection device to travel on to perform the end point detection in place. As described above, FIG. 11A shows the chemical mechanical polishing device in the polishing mode, which contacts the carrier head and the bottom surface 116 of the wafer 114 in the downward position with the polishing surface 106 of the polishing member 102. Although FIG. 17A illustrates that the chemical mechanical polishing device is in the polishing mode, the endpoint detection device is in an inactive position, which means that the endpoint detection device 20 is not located on the bottom surface of the sensing device on the wafer. On a position where the end point detection is performed. FIG. 17B illustrates the end point detection device 1610, which is located in an exemplary chemical mechanical polishing device 1700. The chemical mechanical polishing device includes a carrier head 104, a polishing member 102, an end point detection device 1610, and a road 1730 53 200400098. The invention is explained, and the chemical mechanical polishing device 1700 is in a non-polishing mode. As described above, Fig. 17B shows the chemical mechanical polishing device in the non-polishing mode, which places the carrier head in an elevated position and the bottom surface of the wafer in contact with the light-removing surface of the polishing member. Place the carrier head in a raised position, the final 5-point detection device moves along the track below the carrier head and positions the sensing skirt below the bottom surface of the wafer, thereby positioning the end point detection device in an action s position. Although positioned below the bottom surface of the wafer, the sensing device performs endpoint detection on the semiconductor wafer. For example, the sensing device may sense the reflected color from the wafer surface. It should be noted that there is no need to unload the wafer from the carrier head in order to perform the end point detection 10. If the sensing device determines that the end point has been reached, the wafer can then be unloaded from the carrier and carried to the next station. From the point of view, the movable structure can move (carry) the semiconductor wafer to a subsequent processing station. 15 The movable structure may be any type of component suitable for positioning the sensing device for local end point detection, such as a shuttle, arm or β 1 other type of component. From the point of view, the movable structure is -Clean: The function of the transport is to move the wafer after the end of the process has been reached = clean room (not shown). In this regard, the cleaning and handling shuttle system is designed as a movable structure for positioning the sensing device. If the sensing device_1 is in the -actuated position ', it has reached the end point, and then: the wafer 2 is loaded on the cleaning shuttle (that is, the structure can be moved) and sent to the clean room for cleaning operations. It should be understood that tracks are not necessary for the present invention. For example, if the movable structure is an _arm member, no track is required. If the sensing device determines that the end point is not reached, then the end point detection is performed. 2004 200498 98, the description of the invention is to move away from the carrier head (return to an inactive position), and the carrier head is lowered to move the wafer The rear surface is configured to contact the polishing surface of the polishing member for additional polishing. The cycle of polishing the wafer and moving the end point detection device into the appropriate position to detect the end point of the wafer processing is continued until the end point 5 is reached. In another aspect of the present invention, the shaft 118 and the carrier head rotate the wafer, as shown in Figures 17A and 17B by a circular arrow above the shaft. In this point of view, since the wafer is subject to rotation, the endpoint detection device can scan the entire 10 surface of the wafer by moving on a straight path that encompasses the radius of the wafer. Alternatively, if the wafer is not rotated, the end point detection device can be equipped with a motor to rotate the end point detection device, so that the entire wafer surface can be scanned. The endpoint detection device may alternatively have multiple sensing devices to scan the entire wafer surface. FIG. 18 is a method for detecting a processing end point of a multi-layer semiconductor wafer in a chemical mechanical polishing apparatus having a carrier head and a polishing structure B, such as the exemplary chemical mechanical polishing apparatus 1600. A flowchart of a specific embodiment. In step 1810, polishing of the semiconductor wafer is stopped. In step 1820, the semiconductor wafer in contact with the polishing member is removed by raising the carrier head. In step 1830, a sensing device is moved below 20 on a bottom surface of one of the semiconductor wafers. In step 1840, incident light is emitted from the sensing device against the bottom surface of the semiconductor wafer. In step 850, a reflected color is sensed from the bottom surface of the semiconductor wafer using a sensing device in the reaction incident light. In step 1860, it is determined whether to continue polishing the semiconductor wafer based at least in part on the reflected color. In one aspect, the method further includes 55 200400098, a description of the invention, interrupting polishing the semiconductor wafer, and moving a semiconductor wafer to another processing station if a desired reflection color is sensed. I. Conclusion The advantages of the present invention include the ability to provide the most ideal work piece for polishing to a selected endpoint. In one aspect of the invention, the techniques described herein can be used to polish wafers of varying dimensions. For example, these techniques can be used to polish wafers with a diameter of 200 mm, 300 mm, 400 mm, 500 mm, or other diameters. In one aspect of the invention, wafers of different sizes are polished using the same platform. 10 It should be understood that in the scope of the foregoing description and additional patent applications, the terms "wafer surface" and "wafer surface" include, but are not limited to, the surface of the wafer before processing and the structure of the wafer. The surface of any layer, including conductors, oxidized metals, oxides, spin-on glass (spin_on gkss), ceramics, etc. The term "wafer" ,,, semiconductor wafer, and, substrate, It can be used in place of 15. It should be understood that the specific embodiments and viewpoints described herein can be combined in any suitable manner and operated in the same way. For example, the sensing device and / or the movable structure 1620 It can be combined with the above-mentioned 终点 g-type end point system and / or carrier head pressure change system to provide 20-thickness consistency of semiconductor wafers. The aforementioned combinations are only examples. At the same time, other combinations and specifics can be considered. It should also be understood that although specific wafer processing, such as chemical mechanical light has been sub-clarified, the present invention can be practiced in conjunction with any other type of wafer processing, such as electrochemical mechanical deposition (ECMD) 56 200400098 发明 Description of the Invention The specific embodiments and best modes of the chrome demonstration have been disclosed. Modifications and changes can be made to the specific embodiments disclosed, and they are also covered by the subject matter and spirit of the invention as defined by the scope of the following patent applications. [Schematic description] 5 FIG. 1A is a cross-sectional view of an exemplary substrate after depositing materials on the substrate surface; FIG. 1B is an exemplary substrate of FIG. 1A after receiving a conventional chemical machinery Cross-sectional view after polishing (CMP) processing; FIG. 1C is a cross-sectional view of an exemplary substrate after depositing an insulating material on the substrate surface 10; FIG. 1D is an exemplary substrate of FIG. 1C Cross-sectional view after undergoing a conventional chemical mechanical polishing (CMP) process; Figure 2 is a cross-sectional side view of an exemplary CNU system including one of the preferred embodiments for processing An exemplary endpoint detection system for 15 pieces of work such as wafers; FIG. 3 is an exemplary chemical mechanical withdrawal (CMP) system of FIG. 4 and an endpoint detection system according to the viewpoint of the present invention A cross-sectional top view of the exemplary control system used in the system; Figure 4 is a cross-sectional side view of the exemplary chemical mechanical polishing (CMP) system including the exemplary endpoint detection system of Figure 2; Figures 5A-C Figure 6A is a view of the surface of a work piece; Figure 6A is a work piece processing system illustrating a specific embodiment of the present invention; Figure 6B is a work piece 57 200400098 illustrating another specific embodiment of the present invention发明 Invention processing system; Figure 6C is a work piece processing system illustrating another embodiment of the present invention; Figures 7A-B are one embodiment of the present invention illustrated in Figures 6A-6B. 5 Example platform; FIG. 8 is an exploded view of a sensor according to a specific embodiment of the present invention, and FIG. 9 AB is a diagram illustrating pressure curves and pressure profiles obtained by using the processing of the present invention; 10 Fig. 10A_C illustrates polishing a work piece according to a specific embodiment of the present invention; Fig. 11 illustrates polishing a work piece according to a specific embodiment of the present invention, which shows different force vectors depending on the shape of the work ; Figure 12 is based on A platform of a specific embodiment of the invention has 15-damping buffer layer; and FIGS. 13A-B illustrate a specific embodiment for changing pressure by applying pressure from the back of a work piece Fig. 14 is a specific embodiment of a color sensing device for detecting a processing end point of a multilayer semiconductor wafer, wherein the color sensing device 20 includes a light source, a color sensor, and a Judgment circuit; FIG. 15 is a flowchart of a method for detecting a processing end point of a multi-layered semiconductor wafer, which is a specific embodiment; FIG. 15 is a diagram illustrating an end point for in-situ end point detection A top view of a specific embodiment of the detection device, which includes a movable structure and a sensing device of the 2004200498; invention description; FIG. 16B illustrates a specific example of an end point detection device for local end point detection A side view of the embodiment, which includes a movable structure and a sensing device; FIG. 17A illustrates a specific embodiment of an endpoint detection device disposed in a model chemical mechanical polishing (CMp) device, Wherein, the chemical mechanical polishing (CMP) device includes a carrier head, a polishing member, an end point detection device, and a road, and the chemical mechanical polishing device is in a polishing mode; FIG. 17B illustrates the arrangement A specific embodiment of an endpoint detection device in an exemplary chemical mechanical polishing (CMp) device, wherein the chemical mechanical polishing (CMP) device includes a carrier head, a polishing member, an endpoint detection device, and a road, and The chemical mechanical polishing apparatus is in a non-polishing mode; FIG. 18 is a specific embodiment of a method for detecting a processing end point of a multilayer semiconductor wafer in a chemical mechanical polishing apparatus. The apparatus has a carrier head and a polishing member, and wherein the semiconductor wafer is attached to the carrier head. [Representation of the main components of the diagram] 8. ·. Irregular parts 14 ... Insulating layer 9 ... Plated substrate 15 ... Top surface 10 ... Through hole 12 ... through hole 13. . . Ditch 16. . . Copper layer 18. . . Barrier layer 51. . . Demonstration part 59 200400098 玖, Description of the invention 5 2 ··· Substrate 53 ··· Ditch 54 ·. Bottom insulating layer 5 5 ... Top insulating layer 100 .. Chemical mechanical polishing device 102. ·. Polished member 104 .. Carrier head 10 6 . Lower surface 109 ... Support plate 110 ... Polishing solution 112 ... Roller / polishing solution 113 ... Roller 114 ... wafer 116. Front surface 118. Shaft 120. End point monitoring device 122. Chamber 124. Transparent window 126 .. · Reflector 128… output optical signal 130… received reflected optical signal 132… computer 510a-510n. Wafer substrate 550 ... Chemical mechanical polishing device 560 ... Control unit 562 for thickness consistency ... Fluid supply unit 564 ... Computer controller 566 ... Sensor controller 600 ... Support plate / Platform 610 ... upper surface 612 ... base block 614a-614d ... air chambers 616a-616d. Air lines 618a-618d ... Air ports 620a-620n. Holes 622a-622d. . . Pressure control device 624a-624d ... exhaust port 630a-630n ... · Sensor 630x ... Sensor 632x. Transmission fiber 634x ... Receiving fiber 710. . . Ray 910a-910d ... pressure vector 920a. . . Workpiece surface 920b " surface 920c-l, 920c-2 ·. . Surface 1000. . . System 60 200400098 发明, invention description 1210. . . Film 1210a_1210e. . . Area 1212a-1212b. . . Fastener 1220. . . Pressure controller 1224a-1224e ... fluid line 1300 ... shock absorption buffer layer 1320a-1320n ... first hole 1330a_1330n ... second hole 1400. . . Fluid discharge hole 1420 ... color sensor 1430. . . Judgment circuit 1400a-1400d ... fluid discharge hole 1405. . . Color sensing device 1410. . . Light source 1600 ... platform / chemical mechanical polishing device 1610. . . Endpoint detection device 1620. . . Movable structure 1630. . . Sensing Device 1700. . . Chemical mechanical polishing device 1710. . . End-point detection device 1730. . . Track Z1-Z4 ... First to fourth zones 61