200427977 玫、發明說明： 【發明所屬之技術領域】 本發明係關於一種在設備製造過程中，供谓測半導體 (諸如矽）中表面層金屬污染、偵測由熱處理製程且尤其 是快速熱處理（RTP)製程所引起之表面層金屬=毕用^ =破壞性方法及裝置，該方法及裝置既作爲—品質控制度 量亦作爲一診斷方法以鑒別在設備製造過程中之户理= 題。 处。 【先前技術】 在過去的50年中，矽技術中所取得之進展已在晶片效能 方面產生了顯著的進步。快速收縮設備幾何尺寸與對高效 能電路之技術需求已將吾人帶領到了一點，其中一單=晶 片所能包含之電晶體的數量已升至5百萬。此意即：若$ 一單個200毫米晶圓之上可有2〇〇個晶片，則每個晶圓可總 共包含1 0億電晶體。200427977 Rose, description of the invention: [Technical field to which the invention belongs] The present invention relates to a method for measuring metal surface contamination in a semiconductor (such as silicon) in a device manufacturing process, detecting a heat treatment process and especially rapid heat treatment (RTP) ) Surface layer metal caused by the manufacturing process = complete use ^ = destructive method and device, which serves as both a quality control measure and a diagnostic method to identify household management issues in the equipment manufacturing process. Office. [Previous Technology] Over the past 50 years, advances in silicon technology have produced significant improvements in chip performance. Fast shrinking device geometry and technical requirements for high-efficiency circuits have led us to a point where the number of transistors that a single wafer can contain has risen to 5 million. This means that if there are 200 wafers on a single 200mm wafer, each wafer can contain a total of 1 billion transistors.
Ik著Β曰片尺寸接近01微米，設備良率對缺陷及雜質已 變得愈來愈敏感。該等缺陷及雜質可能是在無意間帶入生 產線中。監測該等雜質，尤其是由處理過程中之金屬污染 引起之該等雜質的存在之能力，及對該等雜質對ic效能與 良率之不良影響的理解是關鍵的，其可作爲品質控制與製 程控制之方法並可用於開發方法以得到改良且更乾淨之 處理技術。 高溫下金屬雜質之高擴散性及溶解度（此可導致該等金 屬雜質及其錯合物之活化）導致對在熔爐區域控制缺陷及 86614 200427977 污染士重要需求。該等雜質一旦被·散4舌，即可導致接合設^ 備效能之降級（degradat i on)、增加之洩漏及’設備可靠性 之降級。同樣在溶爐區域，高溫處理可導致熱滑動 (thermal si ip)，吾人已熟知熱滑動可導致設備效能之降 級。 現在’伴隨於減少生産成本及改善生産力快速反饋方面 正在進行的改良而來的是啓動緊密製程控制之需求；於製 程中關於缺陷的不確定性可導致時間及金錢的損失。已有 許多詳細研究表明在一 M0S (金屬氧化物半導體）設備中， Fe如何可以導致閘氧化物之降解（degradati〇n)。國際標 準（參見如國際半導體材料之國際半導體路標2〇〇1， SEMATECH ， 3101 Industrial Terrace Suite 106 ， Austin TX 78758)建議總Fe濃度必須低於lx l〇1Qcnf3以避免良率 損失。 用以監測污染之習知方法係基於表體壽命（bulk lifetime)及擴散量測。該等方法較緩慢，且需依散佈於 大部分晶圓之雜質而定。對於某些特定熱處理步驟而言， 並非必須使甩藉由使用短暫退火時間及高溫而提供高晶 圓產出之快速熱處理（RTP)。此意即於RTp過程中形成之任 何雜質或缺陷將沒有時間擴散入表體，但是將會達到高至 足以對良率造成顯著害處之表面濃度。爲使用表體壽命分 析之污染測試，應進一步將該等晶圓退火以將該等金屬驅 起入該表體。其延緩了該製程，且於任何狀況下，吾人都 不欲晶圓受到進一步的熱處理。 86614 200427977 【發明内容】 本發明之-目的係提供—種用以偵測半導體（諸如石夕） 中表面層金屬污染之方法及裳置，該方法及裝置減少了— 二或全部的上述缺點。 :發明之-特定目的係提供一種方法及裝置，其提供比 先W技術更爲快速的對由熱處理製程造成之污染之評估。 本發明之一特定目的係提供一種可適用於一改良之品 質控制度量的方法及裝置。 σσ 本發明之另一目的係提供一種用以偵測半導體（諸如 石夕）中表面層金屬污染之方法及裝置’該方法及裝置提供 作爲-診斷方法之潛在額外功能性以在設備冑造過程中 鑒別處理問題。 因此’根據本發明之第一態樣，一種尤其在作爲設備製 造之一部分的熱處理過程中彳貞測半導體結構内之雜質的 方法包括如下步驟： 、 將半導體結構之表面曝露於來自一適當光源之至少一 個高強度光束（較佳爲雷射’尤其爲高強度雷射），且收 集因及光束激勵該半導體結構而産生之光致發光； 番對所收集之光致發光訊號做出分析，且將該分析用作半 ‘體適用‘I·生之品質分級的基礎以供4 一步設備製造處理 之用。 該品質分級步驟包括：執行對所收集之光致發光訊號的 數值分析；將該數值分析之結果與—預定之可接受光致發 光規格’諸如吾人已知關於良好品質的光致發光預定範圍 86614 200427977 進行比較；且基於該比較做出對該半導體結構之品質分 級0 在簡單的替代方法中，該方法包括以下步驟：測定平 均光致發光強度；將該平均值與光致發光之預定的可接受 規格範圍進行比較；且如上所述基於該比較做出對該半導 體結構之品質分級。 該平均值可爲基於整個結構區域所發射之平均光致發 光強度的全區平均值，或可爲局部區域平均值，其中將該 結構之區域分成二維陣列之子區域，爲各子區域測定一平 均光致發光強度，將各子區域之平均值與一預定的可接受 光致發光規格進行比較，且如上所述之品質分級係以該比 較爲基礎。此方式是有利的，因爲在一全區平均值中，對 一單獨缺陷之響應可能被埋沒，即使該缺陷已嚴重到足以 確定其為一不良品。於一適當子區域尺寸下，則可確保仍 能4貞測到該響應。 基於該平均值的一預定光致發光規格之使用僅爲一實 例。於該替代方法中，尤其當使用子區域方法時，可將其 他數值參數施用至光致發光訊號之分析中，諸如標準差、 區域性極大值及/或極小值、自一預定之基線的偏差或其 他數值刀析方法，以在局部或全區基礎上測定該光致發光 響應與吾人已知與具有良好品質之半導體結構有關的預 定參數之間的偏差。雖然下文做出對基於平均發光之數值 刀析的參考應瞭解其僅作爲示例，且爲比較所觀察之塑 應與預定之可接受響應而選擇的精確數值參數對本發^ 86614 10 200427977 並非關鍵。 在由南強度光束之特性所決定之解析度下，光致發光技 術産生-空間解析PL映射。可藉由本方法之另外較佳特徵 (feature)利用該特點，但是對於在處理過程中作爲對整 個晶圓之簡單且快速品質測試的本發明之基礎目的= 言，只需獲得在整個晶圓區域之上的平均凡強度結果。可 將该平均PL強度結果與結合使用較慢分析方法（如表體載 體壽命方法或電子良率測試方法）之研究所開發之一預定 可接受規格範圍相聯繫。吾人驚奇地發現，如下詳述，在 自本發明之基於近表面之PL技術與傳統上所使用之表體 載體壽命方法獲得之雜質資料之間可顯示一密切且明顯 的線性相互關係。 光束（尤其光束功率及/或波長及/或光斑尺寸）經控制 為便於在該半導體結構中於一選擇深度下鑒別缺陷，以自 適當近表面深度（如自該半導體結構之上部12微米處） 收集PL資訊。對於某些材料與設備而言，較小之深度，如 下至5微米或1微米可能比較適當。 因此本發明係可用以監測表面污染及邊緣滑動的一種 "ί貞測-監測工具。由於該技術量測該表面區域，其將彳貞測 使用習知方法偵測不到（譬如由於近表面缺陷及污染未擴 散入表體）但是在對設備品質及效能之影響中非常具有決 定性作用的近表面缺陷及.污染。因此該方法尤其適用於現 代快速熱處理技術，在該等快速熱處理技術中，於處理過 程中引入之雜質可能未完全擴散入表體係表體方法之一 86614 11 200427977 公認的缺陷。爲使表體方法有效，可能需使晶圓經 外退火步驟。相反地’本發明 月之方法可於雜質量特定存在 於表面，存在於待製造設備 a才便用而無需在分析 之刖退火以將該等雜質趕入表體。 吓 根據本發明’首先衫平均光致發光之-狀可接受規 格範圍，且然後爲達到品質控 又規 认—s 貝控制之目的，使用其作爲任何 、、、口疋晶圓之結果的參者。兮箱— 疋之規格範圍將包含—極小 及/或極大光致發光值。特定古之，五λ 口 1 —人 吾人已知視雜質中所 =特定化學物質而定，光致發光訊號可按不同方式受 到衫響。因此，該規格範圍較佳包含-極小及一極大光致 發光值。 徑人九致 視所量測之結果是否扁^Γ ^ “預疋之規格範圍之内做出σ 質控制判定。舉例而言，對在 口口 t在5亥乾圍之内可判定爲“合 格，，，而對在該調之外可判定爲“不合格,，。“不人 2產'可能被丢棄，使其經受矯正措施諸如額外清： 將有所#化获： 所涉及之特定材料及製程 :有所、化。糟由使於-晶圓受到本發明之方法處理時之 PL響應與根據現存先前枯 寻之 質控制規格範圍相關聯，測 見存。° 下給出一實例。 丨……接纽範圍。以 -旦在現存製程之現存製程規格與該製程之規 圍之間建立聯繫，本發明相 & 丁於无刖技術方法提供非當古 之産出。舉例而言，對〜 捉伢非书问 而t不、、 （ 0 〇毫米）晶圓均等物 而β，五刀鐘左右即可獲取 木向ί木用現存表體處理方 86614 -12 - 200427977 法則約需一小時。因此，本發明之技術，除了特別地適合 經快速熱處理之晶圓，其中超過處理時間度量之雜質的緩 f艾擴散速率將導致無退火步驟之先前技術表體方法失 效’不論晶圓所使用之特定熱處理技術是何技術，本發明 之方法亦提供超過表體方法之産出速度上的優點。 光致發光（PL )光譜技術係一供觀測半導體中雜質與缺 陷上之内部及外部電子轉移用之非常靈敏的技術。當矽在 低溫受到高於材料帶隙之雷射輻射激勵時，産生電子空穴 對。該等載體可按各種不同方式重組，其中一些重組方式 導致發光。可將於低溫下形成之電子空穴對截留於矽中的 雜質上，且該等電子空穴對發射具有該相互作用之特徵的 光子’藉此在光致發光光譜中給出雜質特定資訊。現有大 量應用於矽的PL光譜技術，其包括於不同處理步驟之後矽 之特徵描述、設備製造（譬如，植入、氧化、電聚蝕刻、 點缺陷錯合物之偵測及錯位之存在）之特徵。最重要的應 用之一包括對淺施體及受體（諸如砷、硼及磷）之非破壞 性量測。特別地，該技術能夠對該等淺施體及受體之濃度 進行量測。但是，在習知應用中爲獲取光學中心之該光譜 資訊及明確化學鐾別，需於液氦溫度下執行量測。在整個 製造業中，吾人已知於室溫下，PL訊號明顯變弱且可獲 取極少有用之光譜資訊。 因此較佳採用室溫技術，尤其諸如由世界專利申請案卯 第98/1 1425號描述之一非破壞性技術，該技術使基於室溫 之P L半導體結構中電活性缺陷之偵測成爲現實。該專利申 86614 200427977 請案揭示了 一具有‘工業應用之PL技術，於其應用中，該技 術能夠於數分鐘之内産生影像，且該技術在産生尤其是靠 近晶圓表面（設備於其中被製造）之較小單個缺陷的微成 像方面具有另一額外優點。 該技術以一適合工業應用之速率提供關於半導體或矽 結構中之缺陷的資訊，且尤其使吾人能夠目測半導體或矽 結構之上層區域中且尤其是在該結構表面附近之缺陷。該 技術利用在一半導體或矽結構中於缺陷處電子空穴對之 增強的非輻射重組以增強該半導體或矽結構之一凡影像 中的對比度以便增強對該影像巾缺陷之觀測。因此將^界 專利第98/ 1 1 425號中之較佳PL技術以引用的方式併入本 文0 1…吧歸功於藉 較小、空間解析度較佳為n20微米、較理 米且具有1 04至i 0 9瓦特/平方公分之間的峰值或平均功 密度之雷射探測的探測容量，因此區域性缺陷對 P L強度具有更大影響，且吾人亦相 、 丑〇人亦相^，該方法之 是由於自聚焦激勵後，所注入之載體密度較古 加了於缺陷處非輕射重組之可能性且因此二 ^ 物理定位。在某此較佳f f ψ 9 了缺陷： 一1乂仫貝施例中，本發明藉由製借且古, 表性PL響應的缺陷之一空間映射，且 利用該方法。 灵佳爲—空間影像1 此處對高強度雷射之論述意味包含 密产雷射，立p — 限制）高功^ 在度田射，思即不管雷射功率在 ^ 4輪射被聚焦。 86614 200427977 在本發明之一較佳實施例中，使用一脈衝雷射激勵源， 且較理想地，發光資料被量測及/或收集發光影像作爲時 間函數。此意味著深度與空間解析度皆被改善，且可被用 以獲取在缺陷之載體俘獲截面上之資訊。亦可使用時間解 析量測以量測有效載體壽命且獲取壽命映射。 本發明之PL技術於整個晶圓區域生成一空間解析凡映 射。在本發明之主要方法中，其後處理該資料映射以在整 個晶圓提供一平均PL水平，將該平均pL水平與參考標準相 比較以作出品質控制判定。若將該方法用於該合格^不合 格品質控制判定，則僅有該平均PL水平有t彡，而藉由該 方法産生之映射的解析度則不重要。大約7毫米之解析度 足矣。於該水平之解析度下，處理時間被縮短，且測試産 出速率被最大化。舉例而言，自一12英寸（3〇〇毫米°) 2 晶圓獲取一結果可只需五分鐘。 然而，本發明之技術的一特定優點係其可另外被用以生 成貝穿在測試中之半導體表面的PL訊號之空間解析映 、且尤其可被用以生成該等訊號之空間解析影像。因 此，在-較佳實施例中，該方法進一步包括生成該映射及 /或該影像的步驟。於該等環境中，以G· 5毫米或以下之解 析度進行映射/成像作業可能較爲適當。 a車t佳地，該方法進一步包括將空間解析PL映射儲存於適 斗儲存構件上，及/或藉由適當處理構件傳送自空間 :析映射獲得之數位化資料以供前向處理之用。特定言 可按此種方式儲存及/或處理與熱處理之連續階段所 36614 200427977 産生之結果有關的連續數位化映射，以鑒別問題可能出現 於處理過程中之哪一特定階段。 較佳地，該方法進一步包括將任何生成之PL影像顯示於 適當顯示構件上的步驟。 在該較佳實施例中，充分利用該技術在一在測試中之半 導體結構的PL響應上産生空間解析資料的能力，其適於 (如所指示）在較高解析度下作業。因此，産出將比當該 技術僅被用作基礎（如“合格/不合格”）品質控制二= 之基礎時慢。一較佳且更爲發達之品質控制機制可因此使 用該更快速、基礎之技術以處理各單元，且使用由空間解 析貝料之集合提供之額外功能性以取樣製程批量。 於本發明之另一態樣中，進一步利用本發明之特徵，包 括v斷方法以鑒別在設備製造過程中之處理問題，該診 斷方法包括根據本發明之第一態樣在設備製造過程之複 數個處理步驟之末端’且尤其是在複數個熱處理步驟之末 端於至—個半導體結構上陸續執行該债測方法； 另外在該半導體結構之整個表面上，收集及儲存與PL 響應之一空間解析映射有關之資料； ，當完成該等處理階段時，譬如藉由産生一良率映射的方 式電測試該最後産品； 使來自在最德奈α 士 次 中運行表現不理想區域之電測試的 負料與來自各種_ _此 壬〇又之空間解析PL映射的資料相關 聯以繁別哪製籍卩比f 一 I程卩白段可能是造成有問題之區 的問題的原因。 \ w κ π 86614 -16- 200427977 舉例而言，熱處理中一特定階段之PL映射可說明—特定 區域於該處理階段之後可能具有一定量之雜質。該結果電 測試可能表明於該特定區域中之晶片正執行次規格。根據 :發明之該態樣的方法因此能夠蓉別有問題的處理階 段’使其作爲待進—步調查之可能的㈣者。因此，本$ 明之此態樣可充當整個製造製程之一言爹斷方〉去。 ' 於極端狀況中’頃發現來自—單個熱處理階段之 :析PL映射’或自該映射生成之-影像可展示熔爐結構二 産生之-污染㈣。其可呈現清楚證 題與該特定熔爐有關。 今疋可木問 ^方法適用於任何基礎半導體結構，於該等結構之上， =處理以上述方式製造設備。特定言之，該方法適用 於基於矽晶圓及矽合金之鈐 ^ 層晶圓製造該等設備，孽如 間早早層晶圓或自多 中形成該等設備。…基礎梦晶圓上之-Α晶層 二=:另—態'樣，1尤其於作爲設備製造之- 彳伤的熱處理過程中偵測 又 包括-高強度光源，較佳爲：體'、、°構中之雜質的裝置’其 將來自該錢之高強度光=’且尤其爲高強度雷射； 的-表面之上的構件；收隹♦焦至測試中之半導體結構 體結構産生的貫穿測試本構件，其收集因光束激勵半導 發光資料；分析構件，1 /半導體結構的該表面之光致 -比較器’其將分析數值分析所收集之資料; 較。 /、預疋之可接受規格參數進行比 86614 -17- y /7 執行該基礎方、、表 基礎之資料，疑如 置生成以該PL響應之數值分析爲 民立、 S如所述的貫穿整個區域之一平均 局部區域資料之集人m虎或 進行比較。 〇將该―貝料與預定之可接受規格參數 但是，爲執行上；+，十^ 需另外包括將所“改進的替代方法，該裝置較佳 、 集之凡資料解析爲一貫穿半導體區域 之空間解析PL映射 ^ 轉換爲一 PL影像之^ 要包括將所解析之資料 冓件，及/或儲存該映射/影像、尤其是 件連、，映射/影像供將來比較用之影像/資料儲存構 杜及/或將該映射/f彡像料至—適當遠程資料處理器之 :件:及/或將一影像及/或相關資料顯示給一使用者之影 顯不構件（諸如一視覺顯示幕）。 、本^月之另—悲樣中，提供—電腦程式及/或-適當 -式化之電腦以執行本文前述方法之一些或所有步驟，且 尤：在所收集之PL資料上執行資料處理步騾，舉例而言， 決定整個晶圓區域之平均P L，及/或自所收集之P L資料空 間解析-PL映射’及/或將該平均值與—預定之可接受規 格範圍進行比較’及/或比較連續空間解析凡映射以便利 作爲製程診斷一部份的問題處理階段之鑒別。 【實施方式】 圖1令所展示之裝置主要包括一 PL成像顯微鏡，其中： 朝右手面’包括一組雷射(3_8);朝底部，包括一樣品臺， ^ X Y至或卜室，朝左手面，包括一微處理器（40)及 ~顯示幕(39) ’且圖中心部位展示了供導引光線通過該 86614 200427977 系統的各種光學組件。 在圖1所展示之實施例中，接也丄太 、彳j T钕供六束雷射以探測樣品中 :不同深度。但是，僅使用一個雷射，或確實使用更多數 量之雷射都在本發明之範嘴之内。無論如何，該等雷射中 的至少一束爲高強度雷射’且較理想地具有0· 1毫米至0. 5 微：之間的光斑尺寸’及10‘至1〇9瓦特/平方公分之間的功 率密度。提供一與該組雷射相耦接之雷射選擇器（16)以 選擇-束或多纟雷射以供使用，且另外亦選擇該等雷射之 頻率與波長。 使用習知光學裝置’諸如光學纖維（9)以指引光朝向 準直儀（10)及雷射束擴展器（11)。將—切趾板（12) 置於雷射束擴展器（11)與光束分離器（31)之間。光束 分離器（3U #由物鏡（34)指引來自前述雷射之一部份 光朝向樣品（2 )。 提供-自動聚焦控制H (30)，且將其耦接至一壓電驅 動聚焦臺（33)。顯微鏡裝備有一習知轉臺（36)，該轉臺 分別具有供微觀檢查用之至少一個高數值孔徑物鏡及一 供在觀檢查用之低數值孔徑物鏡（34，35 )。此外，亦提 供一耦接於轉臺（36)的光學位移量測系統（38)。 提供電纜敷設以將自動聚焦控制器（3〇)連接至微處理 器（40)’且亦將一顯微鏡物鏡分度裝置（32)連接至微 處理器（40 )。 於光束分離器（31)之下游具有雷射缺口濾光器之濾光 輪（1 3 )，在濾光輪（1 3 )之下游具有一向旁邊搖擺之折 86614 19 200427977 豐鏡（1 4) ’將於下文描述其功能。與該鏡（丨4)排列成行的 係一供波長選擇用之濾光輪（27)，且在濾光輪（27)之後係 一與適當CCD 2-D陣列偵測器（29)相連之變焦透鏡。 在冷鏡（1 7)最前面之光學路徑中具有無限系統補償透 鏡（37)，其將光反射向供波長選擇用之另一濾光輪（23) 及水焦透鏡（2 4 )，該聚焦透鏡在一 U V及可視光之偵測器 (25)的最前面。將偵測器（25)耦接至鎖定放大器（26)。其 用以獲取該等表面之一反射影像。 在冷鏡（17)的最後面具有仍供波長選擇用之另一濾光春 輪（18)，且在濾光輪（18)之最後面具有一聚焦透鏡 及供針孔選擇用之另一孔徑輪（19)，孔徑輪（19)在供偵測 發光用之偵測器（21)的最前面。 uv及可視區域偵測器（25)及紅外線伯測器（21)皆被 接至鎖定放大器（26)。 下文說明該系統之運行。 藉由若干束雷射（3_8)提供波長之範圍以探測樣品中之 =同平面。可藉由頻率產生器（16)調變該等雷射以使得可 藉由谓測器自背景輻射分離自樣品⑵發射之訊號，藉由 鎖定放大器(26)使該等偵測器與雷射調變頻率同步。在另 一實施例巾’可藉由使用—束可調諸雷射及/或―光學表 數振蕩器產生波長之該範圍。將各束雷射連接至—多 光學纖維⑻且與其㈣成行以使縣何或所有該等 皆可㈣樣品⑵。該多分支光學纖維之共同末端終止於 一準直所出現之光的光學系統（10)中。該光學系統虚一光 86614 -20- 200427977 束擴展器（11)排列成行，光束擴展器（丨1)使雷射束直徑與 樣品（2 )之上的顯微鏡物鏡（34，35 )所要求之直徑相匹 配。擴展之光束然後穿過一將光學能量均勾分配於光束區 域之上的切趾板（1 2 )。 藉由一光束分離器（31)反射該擴展切趾光束，且該光束 傳到顯微鏡物鏡（34及35)。藉由一顯微鏡物鏡（34或35) 將该光束聚焦於樣品上。於微觀模式中，選擇該物鏡以將 該光束聚焦至一繞射有限光斑尺寸。藉由一分度機構（32) 操作之轉臺（36)允許該物鏡可被轉變爲宏觀模式，其中可 照射到樣品之更廣區域。在另一實施例中可移除切趾板 (1 2 )’如此可將宏觀模式之光斑變小以允許更高之注入 量 ° 一光學位移感應器（3 8 )量測到樣品之距離，且藉由一通 過抗聚焦控制器（3 0)之反饋迴路，維持藉由壓電致動聚焦 臺（33)産生之正確間隔。 藉由顯微鏡物鏡（34)(其處於微觀模式）收集來自樣品 之光致發光訊號，且藉由光束分離器（3丨）及濾光輪（丨3) 中之缺口濾光器傳輸回該訊號，濾光輪（13)含有與雷射波 長範圍相匹配之缺口濾光器。缺口濾光器移除任何所反射 之雷射光，僅讓光致發光訊號通過。 將折疊鏡（1 4)轉到光束外以允許訊號傳到鏡筒透鏡 (37)且傳到冷鏡（丨7)上，其中可將鏡筒透鏡（37)倂入以補 償可使用之任何無限顯微鏡物鏡。冷鏡（丨7)將在所選截止 點（約700奈米）之下的該等波長反射至將訊號聚焦到偵測 200427977 器（25)中之聚焦透鏡（24)。在谓測哭平隹沒 貝列為水焦透鏡（24)前面之 一濾光輪（2 3)包含濾光器以分離所選之波+ $ 位於在截止點之上的波長範圍中 ㈤τ < 4仞先致發光訊號 穿過冷鏡（17)，且藉由透鏡（22)#相如以& a 兄以〇被相似地聚焦到偵測器 (21)中。該訊號亦穿過一含有濟来考以八紐 π恩尤為以分離所選之波長帶 的濾、光輪（1 8)。 於一被置於偵測器（21 )之前的孔徑輪（丨9 )中包含一系 列不同直徑之針孔。可藉由壓電致動器（2〇)轴向移動該孔 徑輪，如此可放置該等針孔以使其與所要之影像平面共φ 焦。藉由該種方式，可將樣品⑴中位於不同深度之平面 成像以提供精確深度資訊。 將來自偵測器（21，25)之電訊號供給鎖定放大器（26)， 在鎖定放大器（26)中藉由來自頻率産生器（15)之一參考 訊號使該電訊號與雷射（3-8)之調變頻率同步。然後將該 電訊號供給中央處理器（40)以供分析用。藉由光栅掃描樣 品臺以獲取PL影像。或者亦可運用使用檢流計鏡之光學掃 描。 _ 在另一運行微觀模式中，將折疊鏡（14)轉到光致發光訊 號之光束内。該轉向訊號穿過一濾光輪（27)，且進入變焦 透鏡（28)中，其中濾光輪（27)包括濾光器以分離所選之波 長帶。變焦透鏡允許使用不同放大率將樣品（2)上之受照 射光斑成像於CCD二維陣列（29)上。其允許了可於不同解 析度下使樣品（2 )之受照射區域成像。將來自CCD陣列之電 訊號供給中央處理器（4〇)以供分析之用。 86614 -22- 200427977 圖2中圖解歧明了資料之處理。藉由圖工之收集裝置(以 :圖形態展示作設備1〇2)自樣品(m)收集pL資料。於一 弟處理路彼中，根據本發明之主要態樣，將PL映射資料 傳到處理态（103)，其處理該資料以測定貫穿樣品(1〇1) 之整個區域的平均PL強度。 將所得之平均值傳至比較器⑴4)，其使該平均PL強度 Γ料與在資料儲存器（1〇5)之内的—預定之所儲存的㈣ 祀圍相關％ ’且基於該比較將—品質控制判定傳到控制單 元（1〇6)上。視運行之較佳模式，控制單元（106)可爲，舉 0而^冑#作者給出指示之―顯示構件，或—自動處 理系統之控制單元’其在被發送至控制單元（iQ6)之判定 的基礎上可接著按照一自動處理順序行動，譬如作出一接 受/排除選擇，將在測試中之樣品(1〇1)轉移至矯正處理 等0 虛線展示一第二處理路徑，其反映了本發明之可選擇的 第二態樣。於此項可選擇之態樣中，同樣將與貫穿樣品 (1 0 1)表面之PL強度映射相對應之資料傳至—處理單元 (11 0 )，處理單元（u 〇)可將該資料解析爲一貫穿樣。 (101)表面之數位化空間解析強度映射。將所得之映射傳 至一資料儲存器（11丨）及一視覺顯示幕（112)。資料儲存器 (111)能夠儲存及比較來自連續結果之資料，邀 言如自複數 個樣品及/或在複數個處理步驟後自相同樣品産生之次 料。可將該資料，譬如結合設備測試（如電 ^ 电叫式，如最終 産出設備之良率測試）使用，以測定哪裡 j犯出現生産中 86614 -23 - 200427977 之缺陷。 一爲確疋疋否可使用平均PL訊號量測熔爐處理過程中之 π杂，使用表體壽命標準方法量測所選擇的含有不同量之 金屬污染的300毫米氧化晶圓。記錄各晶圓之映射，且然 後於圖1所示之系統上量測該映射。藉由直接比較來自各 糸統之晶圓映射，可創造污染量對平均几訊號之校準圖。 圖3展示一典型實例。當觀察該圖時，清晰可見在平均 PL訊號與金屬污染濃度之間存在直接線性相互關係。因此 y將自一晶圓映射獲取之平均凡訊號用作晶圓品質之度修 量。記錄該PL映射且然後將其與一預定值進行比較。使用 該規定範圍以測定晶圓品質，且因此創造了 一合格/不合 才°氣1控製程序。可於圖4參見該程序之實例。 圖5及6展示空間解析映射之實例，可使用如此之映射 (譬如）結合本發明之較佳的第二態樣鑒別缺陷之位置，在 。亥第—態樣中本發明不僅被用作一種品質控制方法，亦被 用作一種製程控制構件。 在建立結合圖3及4描述之程序中的規定範圍後，可接著鲁 將該範圍與任何色彩或灰階系統聯繫，且生成一說明晶圓 扣質之影像。圖5係一實例，其中較亮之區域在規格之内， 而較暗之區域在規格之外。可使用該簡單色彩編碼或灰色 、扁碼以k測整個樣品中之晶圓品質的變化，其中以晶圓映 射中之較暗區域展示在規定範圍之外的區域。 可使用該方法在其他熔爐製程（諸如RTP)中及可使用熱 處理的任何地方評定晶圓品質。 86614 -24 - 200427977 在使用平均PL訊號水平檢測晶圓品質後，可以更高之解 析度記錄進一步量測。亦可將晶圓映射記錄於晶圓背面。 其有助於鑒別污染源，且尤其有助於鑒別污染源是否來自 晶圓之後表面。圖6展示在熔爐處理後自晶圓獲取之晶圓 映射。於該實例中，晶圓之背面受到來自晶圓支架之污 染。在熱處理過程中，污染擴散至晶圓正面。藉由圖6b(晶 圓正面）中所明顯展示之與圖6a(晶圓背面）中熔爐結構相 對應之雜質的清晰描繪可見污染之擴散。 圖中之實例表明可使用更高解析度之晶圓映射以測定 工具受到污染時之“指紋，，。在對完全處理晶片進行電測 試之最後部份時，若有一區域顯示明顯良率問題，則可將 晶圓映射用作診斷目的以鑒別可能引入導致該問題之污 染的可能的肇事熔爐或處理階段。 可將微分干涉對比成像（D 1C)模塊加入圖i所示之工具 以使能夠债測滑動線。因此該系統可使用PL晶圓映射债測 金屬污染，且使用一工具上的DIC系統而非使用兩個單獨 的機器偵測滑動（一般同樣生成於熔爐區域）。 因此根據本發明’將光致發光工具用作一快速製程#制 工具以測定晶圓品質。獲取晶圓映射，且數值分析所量測 之PL響應以提供一品質度量，其中光致發光響應與吾人已 知和具有良好品質之半導體結構相關之預定參數之間的 偏差被用作品質控制判定之基礎。 在所給出之基礎實例中，其可僅爲在整個區域 a <上的平 均響應之比較。但是一般而言，污染可非常具有巧，眭 86614 -25 - 200427977 且平均PL訊號(在韻映射申（崎縣暴之像素上所求 之=),能未必爲該污染之可靠指示，。若使用一爪網 格細分晶圓映射，則佳用jp«始pi八 j使用+ APL刀析可偵測到非常具有區 域性之訊號。 於記錄-晶圓映射之後，可施用—虛擬網格，且使用適 當軟體將其顯示於晶圓映射之頂部。可使用該虛擬網格執 行分析。同時局部區域分析亦允許結構中問題位置之更優 定位。舉例而言’可指示出不合格之網格元件，·當網格元 件允許吾人更詳細地檢測所關心<區域時，可於相同位置 啓動微觀掃描；且可用適當形式輸出晶圓映射分析中不合 格元件之記錄。 局部網格方法並不局限於基於平均值之數值分析。可使 用任何適當之預先規定的參數（包括平均強度、PL極小 值極大值、標準偏差及基線），以測定污染區域。由於 在貫穿晶圓之訊號中的變化不均勻，pL訊號基線方法可爲 一更爲有用之參數。將在下文介紹此技術。 圖7a展示一典型晶圓映射，而圖7b展示相關的PL強度之鲁 直方圖。 在晶圓映射中藉由與基線值之偏差偵測到污染，且可設 定界限。但是，藉由污染修正該晶圓映射中之^平均值。 而所用之值應代表一未被污染的晶圓之訊號水平。圖8所 不之峰值代表一未被污染晶圓之真實PL值。修正基線函數 以具有一峰值將允許用戶準確追蹤污染且具有更高靈敏 度。 86614 -26- 200427977 一適當算法包括以下步驟： 1·界定峰值爲直方圖中之極大值，且將其用於基線。 2 · 搜索峰值之資料。 3·然後計异±70 %之極大值（其由使用者規定），接著重 新界定峰值極大值爲該等點之中心位置。 4·然後計算峰值極大值之精確值，且接著界定pL水平。 5·同樣允許使用者輸入來自未被污染晶圓之典型基線 值，若存在相等強度之兩個峰值，其將有所幫助。 晶圓之基線係對應於點之最大數量的pL值。 藉由下列關係式界定基線偏差： 基線偏差=PL值-基線 PL值係網格各元件之AVG pL。必須對各元件量測極限偏 差。 【圖式簡單說明】Ik wrote that the chip size is close to 01 micron, and the device yield has become more and more sensitive to defects and impurities. Such defects and impurities may be inadvertently brought into the production line. The ability to monitor the presence of these impurities, especially those caused by metal contamination during processing, and to understand the adverse effects of these impurities on IC performance and yield are critical, and can be used as quality control and Process control methods can also be used to develop methods for improved and cleaner processing techniques. The high diffusivity and solubility of metal impurities at high temperatures (which can lead to the activation of these metal impurities and their complexes) has led to important requirements for controlling defects in the furnace area and 86614 200427977 polluters. Once these impurities are dispersed, it can lead to degradat on the efficiency of joint equipment, increased leakage, and degradation of equipment reliability. Also in the melting furnace area, high temperature treatment can cause thermal si ip, and we are well aware that thermal sliding can cause degradation of equipment efficiency. Now, along with ongoing improvements in reducing production costs and improving rapid feedback on productivity, there is a need to initiate tight process control; uncertainty about defects in the process can lead to loss of time and money. Many detailed studies have shown how Fe can cause the degradation of gate oxides in a MOS (metal oxide semiconductor) device. International standards (see, for example, International Semiconductor Roadmap 2001 for International Semiconductor Materials, SEMATECH, 3101 Industrial Terrace Suite 106, Austin TX 78758) recommend that the total Fe concentration must be lower than 1x10Qcnf3 to avoid yield loss. Conventional methods for monitoring pollution are based on bulk lifetime and diffusion measurements. These methods are slower and depend on the impurities that are scattered on most wafers. For some specific heat treatment steps, rapid heat treatment (RTP) that provides high wafer yield by using short annealing times and high temperatures is not necessary. This means that any impurities or defects formed during the RTp process will not have time to diffuse into the surface, but will reach a surface concentration high enough to cause significant harm to yield. To use the contamination test for the life analysis of the watch body, the wafers should be further annealed to drive the metals into the watch body. It delays the process, and under no circumstances do we want the wafer to undergo further heat treatment. 86614 200427977 [Summary of the invention] The purpose of the present invention is to provide a method and a device for detecting metal pollution on a surface layer in a semiconductor (such as Shixi), the method and the device reduce the two or all of the above disadvantages. : Invention-Specific Purpose is to provide a method and apparatus that provide a faster assessment of contamination caused by heat treatment processes than prior W technologies. It is a specific object of the present invention to provide a method and apparatus applicable to an improved quality control metric. σσ Another object of the present invention is to provide a method and device for detecting surface metal contamination in semiconductors (such as Shixi). The method and device provide potential additional functionality as a diagnostic method for manufacturing equipment. Identify and handle problems. Therefore, according to the first aspect of the present invention, a method for detecting impurities in a semiconductor structure, particularly during a heat treatment process as part of the manufacture of a device, includes the steps of: exposing the surface of the semiconductor structure to At least one high-intensity light beam (preferably a laser, especially a high-intensity laser), and collecting photoluminescence generated by the semiconductor structure with the light beam; and analyzing the collected photoluminescence signal, and This analysis was used as the basis for the semi-'body-applicable 'I. quality classification for 4 step equipment manufacturing processes. The quality grading step includes: performing a numerical analysis of the collected photoluminescence signals; and a result of the numerical analysis with—a predetermined acceptable photoluminescence specification, such as a predetermined range of photoluminescence that we know about good quality 86614 200427977 to make a comparison; and based on the comparison to make a grade for the quality of the semiconductor structure. 0 In a simple alternative method, the method includes the steps of: determining the average photoluminescence intensity; Accept the specification range for comparison; and based on the comparison as described above, grade the quality of the semiconductor structure. The average value may be a whole-area average value based on the average photoluminescence intensity emitted by the entire structure area, or may be a local area average value, in which the structure area is divided into sub-areas of a two-dimensional array, and a sub-area is determined for each sub-area. The average photoluminescence intensity is compared with the average value of each sub-region with a predetermined acceptable photoluminescence specification, and the quality classification as described above is based on the comparison. This approach is advantageous because in a global average, the response to a single defect may be buried, even if the defect is severe enough to determine it as a defective product. With an appropriate sub-region size, it is ensured that the response can still be measured. The use of a predetermined photoluminescence specification based on the average value is only an example. In this alternative method, especially when using the sub-region method, other numerical parameters may be applied to the analysis of the photoluminescence signal, such as standard deviation, regional maximum and / or minimum, deviation from a predetermined baseline Or other numerical analysis methods to determine the deviation between the photoluminescence response and predetermined parameters that we know are related to semiconductor structures of good quality on a local or full-area basis. Although a reference to numerical analysis based on average luminescence is given below, it should be understood that it is only an example, and the precise numerical parameters selected to compare the observed plastic with a predetermined acceptable response are not critical to the present invention ^ 86614 10 200427977. At a resolution determined by the characteristics of the south-intensity beam, the photoluminescence technique produces a spatially resolved PL map. This feature can be utilized with another preferred feature of the method, but for the basic purpose of the invention as a simple and fast quality test of the entire wafer during processing, it is only necessary to obtain the entire wafer area Above the average Van intensity results. This average PL intensity result can be linked to a predetermined acceptable specification range developed by a research institute that uses a slower analysis method, such as a surface carrier life method or an electronic yield test method. I was surprised to find that, as detailed below, a close and obvious linear correlation can be shown between the impurity data obtained from the near surface-based PL technology of the present invention and the traditionally used surface carrier carrier life method. The beam (especially the beam power and / or wavelength and / or spot size) is controlled to facilitate identification of defects in the semiconductor structure at a selected depth to a suitable near-surface depth (eg, 12 microns above the semiconductor structure) Collect PL information. For some materials and equipment, smaller depths, down to 5 microns or 1 micron, may be appropriate. Therefore, the present invention is a " verity detection-monitoring tool " which can be used to monitor surface contamination and edge slip. Because this technology measures the surface area, it can't be detected by conventional methods (such as near surface defects and pollution has not spread into the surface), but it is very decisive in affecting the quality and performance of the device. Near surface defects and contamination. Therefore, this method is especially suitable for modern rapid heat treatment technologies. In these rapid heat treatment technologies, impurities introduced during the process may not fully diffuse into the surface body of the surface system. 86614 11 200427977 A recognized defect. For the surface body method to be effective, the wafer may need to undergo an annealing step. Conversely, the method of the present invention can exist on the surface only when the impurity is present, and it can be used only when it is present in the equipment to be manufactured. According to the present invention, 'firstly, the average photoluminescence of the shirt is within the acceptable specification range, and then for the purpose of quality control and certification, it is used as a reference for the results of any wafer. By. Xi Box—The range of specifications for tritium will include—minimal and / or maximal photoluminescence values. In particular ancient times, five lambda mouths 1—people We know that depending on the specific chemical substance in the impurities, the photoluminescence signal can be heard in different ways. Therefore, the specification range preferably includes -minimum and a maximum photoluminescence value. Whether the measured result of Jiren Jizhi is flat ^ Γ ^ "The quality control judgment is made within the range of the pre-specified specifications. For example, if the mouth t is within 5 Haiganwei, it can be judged as" Qualified, and for those outside of this tune that can be judged as "unqualified,". "Inhuman 2 properties" may be discarded and subjected to corrective measures such as additional clearance: will be obtained #: Specific materials and processes: some. This is due to the fact that the PL response when a wafer is processed by the method of the present invention correlates with the range of quality control specifications based on existing prior search, as measured by the present. An example is given below °.丨 ... range of joints. In order to establish a connection between the existing process specifications of the existing process and the scope of the process, the present invention & Ding Wu Wu technology method provides non-dangerous output. For example, if you want to capture the non-book, but not t (, 0 mm) and the wafer equivalent, and β, you can obtain the wood surface using the existing surface treatment method in about five minutes. 86614 -12-200427977 The rule takes about an hour. Therefore, the technology of the present invention is particularly suitable for wafers subjected to rapid thermal processing, in which the slow diffusion rate of impurities exceeding the processing time metric will result in the failure of the prior art surface method without the annealing step. What is the specific heat treatment technology, the method of the present invention also provides advantages over the output speed of the surface body method. Photoluminescence (PL) spectroscopy is a very sensitive technique for observing internal and external electron transfer on impurities and defects in semiconductors. Electron-hole pairs are generated when silicon is excited by laser radiation above the band gap of the material at low temperatures. These vectors can be recombined in a variety of different ways, some of which recombine to cause luminescence. Electron-hole pairs formed at low temperatures can be trapped on impurities in silicon, and the electron-hole pairs emit photons having the characteristics of the interaction, thereby giving impurity-specific information in the photoluminescence spectrum. There are a large number of PL spectroscopy technologies applied to silicon, including the characterization of silicon after different processing steps, and equipment manufacturing (such as implantation, oxidation, polyetching, detection of point defect complexes and the existence of misalignment) feature. One of the most important applications includes non-destructive measurement of shallow donors and acceptors such as arsenic, boron, and phosphorus. In particular, this technique enables the measurement of the concentrations of such shallow donors and acceptors. However, in the conventional application, in order to obtain the spectral information of the optical center and clarify the chemical difference, the measurement needs to be performed at the temperature of liquid helium. Throughout the manufacturing industry, we know that at room temperature, the PL signal is significantly weakened and very little useful spectral information can be obtained. Therefore, it is preferred to use room temperature technology, especially a non-destructive technology such as described in World Patent Application No. 98/1 1425, which enables the detection of electroactive defects in a P L semiconductor structure based on room temperature. The patent application 86614 200427977 application discloses a PL technology with 'industrial applications'. In its application, this technology can produce images within minutes, and the technology is particularly close to the surface of the wafer where the device is Manufacturing) has the additional advantage of microimaging smaller individual defects. This technology provides information about defects in semiconductor or silicon structures at a rate suitable for industrial applications, and in particular enables us to visually inspect defects in the upper layers of semiconductor or silicon structures, and especially near the surface of the structure. This technology utilizes enhanced non-radiative recombination of electron-hole pairs at the defects in a semiconductor or silicon structure to enhance the contrast in one image of the semiconductor or silicon structure in order to enhance the observation of defects in the image towel. Therefore, the better PL technology in Patent No. 98/1 1 425 is incorporated herein by reference. 0 1 ... thanks to the smaller size, the spatial resolution is preferably n20 micrometers, the better than the meter, and has 1 The detection capacity of laser detection for peak or average work density between 04 and i 0 9 watts / cm 2, so regional defects have a greater impact on PL intensity, and we are also ugly and ugly. ^ The method is that after self-focusing excitation, the density of the injected carrier is higher than the possibility of non-light recombination at the defect and therefore physical positioning. In some preferred f f ψ 9 the defects are as follows:-In a 1 乂 仫 case, the present invention uses the method to map a space of one of the defects of the ancient and expressive PL response. Lingjiawei—Space Image 1 The discussion of high-intensity lasers here includes dense-produced lasers, and p-limitation) high power ^ shots in Dutian, that is, regardless of the laser power is focused in ^ 4 rounds. 86614 200427977 In a preferred embodiment of the present invention, a pulsed laser excitation source is used, and ideally, the luminescence data is measured and / or the luminescence image is collected as a function of time. This means that both depth and spatial resolution are improved, and can be used to obtain information on the defect-captured cross-section of the carrier. Time analysis measurements can also be used to measure the effective carrier life and obtain a life map. The PL technology of the present invention generates a spatially resolved map over the entire wafer area. In the main method of the present invention, the data map is subsequently processed to provide an average PL level throughout the wafer, and the average pL level is compared with a reference standard to make a quality control decision. If this method is used for the qualified quality control decision, only the average PL level has t, and the resolution of the mapping generated by this method is not important. A resolution of about 7 mm is sufficient. At this level of resolution, the processing time is shortened and the test output rate is maximized. For example, obtaining a result from a 12 inch (300 mm °) 2 wafer can take only five minutes. However, a particular advantage of the technology of the present invention is that it can additionally be used to generate spatially resolved maps of PL signals that are pierced through semiconductor surfaces under test, and in particular can be used to generate spatially resolved images of such signals. Therefore, in a preferred embodiment, the method further includes the step of generating the mapping and / or the image. In these environments, mapping / imaging with a resolution of G · 5 mm or less may be appropriate. Preferably, the method further includes storing the spatially resolved PL map on a bucket storage component, and / or transmitting the digitized data obtained from the space: analysis map by an appropriate processing component for forward processing. SPECIFIC SPECIFICATIONS Continuous digital maps related to the results produced by successive phases of heat treatment 36614 200427977 can be stored and / or processed in this way to identify which specific stage of the process the problem may occur. Preferably, the method further comprises the step of displaying any generated PL image on an appropriate display member. In the preferred embodiment, the technology is fully utilized in its ability to generate spatially resolved data on the PL response of the semiconductor structure under test, which is suitable (as indicated) for operation at higher resolutions. As a result, the output will be slower than when the technology is used only as a basis (eg “pass / fail”). A better and more developed quality control mechanism can therefore use this faster, more basic technology to process each unit, and use the additional functionality provided by the collection of spatially resolved shellfish to sample the process batch. In another aspect of the present invention, the features of the present invention are further utilized, including a v-cut method to identify processing problems in the device manufacturing process. The diagnostic method includes a plurality of numbers in the device manufacturing process according to the first aspect of the present invention. The end of each processing step 'and especially at the end of a plurality of heat treatment steps, the debt measurement method is successively performed on one semiconductor structure; in addition, a spatial analysis of collection and storage and PL response on the entire surface of the semiconductor structure Map the relevant information; when the processing stages are completed, such as by electrically testing the final product by generating a yield map; make the negative of the electrical test from areas where the performance is not ideal in the most Denai α-times It is expected that the data from various __ this non-zero spatially resolved PL maps will be associated with each other, which may be the cause of the problematic area. \ w κ π 86614 -16- 200427977 For example, the PL mapping of a specific stage in heat treatment can indicate that a specific area may have a certain amount of impurities after the treatment stage. The resulting electrical test may indicate that the wafers in that particular area are performing sub-specs. According to: the method of this aspect of the invention can therefore be used to identify problematic processing stages, making it a possible candidate for further investigation. Therefore, this aspect of Ben Mingzhi can be used as one word in the entire manufacturing process. 'In extreme conditions' were found to come from-a single heat treatment stage: Analytical PL mapping "or the images generated from the mapping-can show the contamination of the furnace structure II. It can present clear evidence that it is related to that particular furnace. This method can be applied to any basic semiconductor structure, and on top of these structures, = processing equipment manufactured in the manner described above. In particular, this method is applicable to the manufacture of such devices based on silicon wafers and silicon alloy wafers, such as forming wafers early or early. … -A crystal layer 2 on the basic dream wafer =: another-state 'like, 1 especially as a device manufacturing-the detection of the heat treatment process of the wound also includes-a high intensity light source, preferably: the body', The device of impurities in the structure 'It will come from the high-intensity light of the money =' and especially a high-intensity laser;-a component on the surface; collected by focusing on the structure of the semiconductor structure under test Throughout testing this component, it collects semiconducting luminescence data due to beam excitation; the analysis component, 1 / photo-comparator of the surface of the semiconductor structure, which will analyze the data collected by numerical analysis; comparison. / 、 The ratio of acceptable specifications in advance is 86614 -17- y / 7 Execute the basic information of the basic party, table, if you set the numerical analysis of the response of the PL to be private, S as described in the The average local area data of one of the entire area is collected or compared. 〇 The “shell material” and the predetermined acceptable specification parameters, but, for implementation; +, 10 ^ In addition, the “improved alternative method, the device ’s better and more comprehensive information should be parsed into a through the semiconductor area Spatially parsed PL map ^ Converted to a PL image ^ includes an image / data storage structure that converts the parsed data file and / or stores the map / image, especially a piece of data, and maps / image for future comparison And / or the mapping to the appropriate remote data processor: pieces of: and / or a video display component (such as a visual display screen) that displays an image and / or related data to a user ). In addition to this month—in sadness, provide—computer programs and / or—appropriate-styled computers to perform some or all of the steps in the method described above, and in particular: execute the data on the collected PL data Process steps, for example, determine the average PL for the entire wafer area, and / or interpret the PL data from the collected PL data space-PL mapping 'and / or compare the average with a-predetermined acceptable specification range' And / or more continuously empty Analyze all mappings to facilitate identification of problem processing stages as part of process diagnostics. [Embodiment] The apparatus shown in Fig. 1 mainly includes a PL imaging microscope, where: 'the right-hand side' includes a set of lasers (3_8) Towards the bottom, including a sample stage, ^ XY to or Bu room, to the left-hand side, including a microprocessor (40) and ~ display screen (39) 'and the center of the figure shows for guiding light through the 86614 200427977 Various optical components of the system. In the embodiment shown in FIG. 1, six laser beams are also used to detect the sample: different depths. However, only one laser is used, or indeed more A large number of lasers are within the scope of the present invention. In any case, at least one of the lasers is a high-intensity laser 'and ideally has a range of 0.1 mm to 0.5 micro: Spot size 'and power density between 10' and 109 watts / cm2. A laser selector (16) coupled to the set of lasers is provided to select a beam or multiple lasers for selection Use, and additionally select the frequency and wavelength of these lasers. Use conventional optical devices such as optical fibers (9) to direct light towards the collimator (10) and laser beam expander (11). Place the apodization plate (12) on the laser beam expander (11) And the beam splitter (31). The beam splitter (3U # is directed by the objective lens (34) toward a sample (2) from a portion of the aforementioned laser. Provide-auto focus control H (30), and Coupled to a piezo-driven focusing stage (33). The microscope is equipped with a conventional turntable (36), each of which has at least one high numerical aperture objective lens for microscopic inspection and a low value for observation inspection Aperture objectives (34, 35). In addition, an optical displacement measurement system (38) coupled to the turntable (36) is also provided. A cable is provided to connect the autofocus controller (30) to the microprocessor (40) 'and also to connect a microscope objective indexing device (32) to the microprocessor (40). A filter wheel (1 3) with a laser notch filter downstream of the beam splitter (31), and a fold that swings sideways downstream of the filter wheel (1 3). 86614 19 200427977 Feng mirror (1 4) 'will Its function is described below. A filter wheel (27) for wavelength selection is arranged in a row with the mirror (丨 4), and a zoom lens connected to an appropriate CCD 2-D array detector (29) is arranged after the filter wheel (27). . In the front optical path of the cold mirror (1 7), there is an infinite system compensation lens (37), which reflects the light to another filter wheel (23) and a water-focus lens (2 4) for wavelength selection. The lens is in front of a UV and visible light detector (25). A detector (25) is coupled to the lock-in amplifier (26). It is used to obtain a reflection image of one of these surfaces. On the rear surface of the cold mirror (17), there is another filter spring wheel (18) for wavelength selection, and at the end of the filter wheel (18), there is a focusing lens and another aperture wheel for pinhole selection. (19), the aperture wheel (19) is at the front of the detector (21) for detecting light emission. UV and visible area detectors (25) and infrared primary detectors (21) are connected to the lock-in amplifier (26). The operation of the system is explained below. A range of wavelengths is provided by several lasers (3_8) to detect = coplanar in the sample. The laser can be modulated by the frequency generator (16) so that the signal emitted from the sample plutonium can be separated from the background radiation by the detector, and the detector and laser can be caused by the lock-in amplifier (26). Modulation frequency synchronization. In another embodiment, this range can be generated by using a beam-tunable laser and / or an optical table oscillator. Each beam of laser is connected to a multi-fiber optical fiber and is aligned with it so that all or all of these can be used as samples. The common ends of the multi-branched optical fibers terminate in an optical system (10) that collimates the emerging light. The optical system imaginary light 86614 -20- 200427977 beam expander (11) is arranged in a row, and the beam expander (丨 1) makes the laser beam diameter and the microscope objective lens (34, 35) above the sample (2) required Match the diameter. The expanded beam then passes through an apodization plate (1 2) that distributes optical energy evenly over the beam area. The extended apodized beam is reflected by a beam splitter (31), and the beam is passed to microscope objectives (34 and 35). The beam is focused on the sample by a microscope objective (34 or 35). In the micro mode, the objective lens is selected to focus the beam to a diffraction-limited spot size. The turntable (36), which is operated by an indexing mechanism (32), allows the objective lens to be transformed into a macro mode in which a wider area of the sample can be illuminated. In another embodiment, the toe plate (1 2) can be removed, so that the spot size of the macro mode can be reduced to allow a higher injection amount. An optical displacement sensor (3 8) measures the distance to the sample. And through a feedback loop through the anti-focus controller (30), the correct interval generated by the piezoelectrically actuated focusing stage (33) is maintained. The photoluminescence signal from the sample is collected by the microscope objective (34) (which is in the micro mode), and transmitted back through the notch filter in the beam splitter (3 丨) and the filter wheel (丨 3), The filter wheel (13) contains a notch filter that matches the laser wavelength range. The notch filter removes any reflected laser light and allows only the photoluminescence signal to pass. Turn the folding lens (1 4) out of the light beam to allow the signal to pass to the barrel lens (37) and to the cold mirror (丨 7), where the barrel lens (37) can be inserted to compensate for any useable Infinity microscope objective. The cold mirror (丨 7) reflects these wavelengths below the selected cut-off point (about 700 nm) to a focusing lens (24) that focuses the signal to the detection 200427977 (25). A filter wheel (2 3) is listed in front of the hydrofocal lens (24) at the measurement level, and contains a filter to separate the selected wave + $ in the wavelength range above the cutoff point ㈤τ < The 4 仞 proluminescence signal passes through the cold mirror (17), and is similarly focused into the detector (21) by the lens (22) # 相 如 以 & a brother 〇. The signal also passes through a filter and a light wheel (18) that contain the Zirconium and New Zealand π En especially to separate the selected wavelength band. An aperture wheel (丨 9) placed in front of the detector (21) contains a series of pinholes of different diameters. The aperture wheel can be moved axially by a piezoelectric actuator (20), so that the pinholes can be placed so that they are in φ focus with the desired image plane. In this way, planes at different depths in the sample stack can be imaged to provide accurate depth information. The electric signal from the detector (21, 25) is supplied to the lock-in amplifier (26). In the lock-in amplifier (26), the electric signal and the laser (3-) are made by a reference signal from a frequency generator (15). 8) The modulation frequency is synchronized. This electrical signal is then supplied to a central processing unit (40) for analysis. A raster-scanned sample stage was used to obtain the PL image. Alternatively, an optical scan using a galvanometer mirror can be used. _ In another operating micro mode, turn the folding mirror (14) into the light beam of the photoluminescence signal. The steering signal passes through a filter wheel (27) and enters the zoom lens (28), wherein the filter wheel (27) includes a filter to separate the selected wavelength band. The zoom lens allows the illuminated spot on the sample (2) to be imaged on a CCD two-dimensional array (29) with different magnifications. It allows imaging the irradiated area of the sample (2) at different resolutions. The signal from the CCD array is supplied to the central processing unit (40) for analysis. 86614 -22- 200427977 The processing of the data is clearly illustrated in Figure 2. The pL data was collected from the sample (m) by means of a collection device (shown as a device 102 in the form of a figure) of a drawing worker. In the processing of Lupi, according to the main aspect of the present invention, the PL mapping data is transmitted to the processing state (103), which processes the data to determine the average PL intensity across the entire area of the sample (101). The obtained average value is transmitted to the comparator ⑴4), which makes the average PL intensity Γ be related to the predetermined stored 围 within the data storage (105), and based on the comparison, -The quality control decision is passed to the control unit (106). Depending on the preferred mode of operation, the control unit (106) may be, for example, 0 and ^ 胄 # The author gives instructions-display components, or-the control unit of the automatic processing system, which is being sent to the control unit (iQ6). Based on the judgment, you can then follow an automatic processing sequence, such as making an acceptance / rejection selection, transferring the sample (1101) in the test to the correction processing, etc. 0 The dashed line shows a second processing path, which reflects this An alternative second aspect of the invention. In this optional aspect, the data corresponding to the PL intensity mapping across the surface of the sample (1 0 1) is also transmitted to the processing unit (11 0), which can be analyzed by the processing unit (u 〇). For a run through. (101) Digitized spatially resolved intensity map of the surface. The obtained mapping is transmitted to a data storage (11 丨) and a visual display screen (112). A data storage (111) is capable of storing and comparing data from consecutive results, such as invitations from multiple samples and / or secondary materials generated from the same sample after multiple processing steps. This information can be used, for example, in conjunction with equipment testing (such as electric ^ calling, such as yield test of the final output equipment) to determine where the 86 86 -23-200427977 defects in production occurred. The first is to determine if the average PL signal can be used to measure π impurities in the furnace treatment process. The standard life of the body is used to measure the selected 300 mm oxidized wafers containing different amounts of metal contamination. Record the mapping of each wafer, and then measure the mapping on the system shown in Figure 1. By directly comparing wafer maps from each system, a calibration map of contamination versus average signals can be created. Figure 3 shows a typical example. When looking at the graph, it is clear that there is a direct linear correlation between the average PL signal and the metal pollution concentration. Therefore, y uses the average signal obtained from a wafer map as a measure of wafer quality. The PL mapping is recorded and then compared to a predetermined value. This specified range was used to determine wafer quality, and therefore a pass / fail 1 control procedure was created. See Figure 4 for an example of this procedure. Figures 5 and 6 show examples of spatially resolved mappings. Such mappings (for example) can be used in conjunction with the preferred second aspect of the invention to identify the location of a defect at. In the first aspect, the present invention is used not only as a quality control method, but also as a process control component. After establishing the specified range in the procedure described in conjunction with Figures 3 and 4, the range can then be linked to any color or grayscale system and an image illustrating the quality of the wafers can be generated. Figure 5 is an example in which the lighter areas are within the specifications and the darker areas are outside the specifications. The simple color coding or gray, flat code can be used to measure the change in wafer quality in the entire sample with k, where the darker areas in the wafer map are displayed outside the specified range. This method can be used to evaluate wafer quality in other furnace processes, such as RTP, and where thermal processing can be used. 86614 -24-200427977 After the average PL signal level is used to check the wafer quality, it can be further measured with a higher resolution record. The wafer map can also be recorded on the back of the wafer. It helps to identify the source of contamination and, in particular, whether the source of contamination is from the back surface of the wafer. Figure 6 shows the wafer map obtained from the wafer after the furnace process. In this example, the backside of the wafer is contaminated from the wafer holder. During the heat treatment, the contamination spreads to the front side of the wafer. The diffusion of contamination is clearly visible by the clear depiction of the impurities corresponding to the furnace structure in Fig. 6a (back side of the wafer), which is clearly shown in Fig. 6b (front side of the wafer). The example in the figure shows that higher resolution wafer mapping can be used to determine the "fingerprint" of a tool when it is contaminated. If the last part of the electrical test of a fully processed wafer is performed, if an area shows a significant yield problem, Wafer mapping can then be used for diagnostic purposes to identify possible furnaces or processing stages that could introduce contamination that could cause the problem. A differential interference contrast imaging (D 1C) module can be added to the tool shown in Figure i to enable debt The sliding line is measured. Therefore, the system can use PL wafer mapping to detect metal contamination, and use a DIC system on one tool instead of two separate machines to detect sliding (generally also generated in the furnace area). Therefore according to the present invention 'Using the photoluminescence tool as a rapid process # manufacturing tool to determine wafer quality. Obtain wafer mapping and numerically analyze the measured PL response to provide a quality metric, where the photoluminescence response is consistent with what we know and Deviations between predetermined parameters related to semiconductor structures of good quality are used as the basis for quality control decisions. In the basic examples given, It may only be in the entire area a Comparison of average responses on <. But in general, pollution can be very clever, 眭 86614 -25-200427977 and the average PL signal (required on the rhyme map application (required on the pixel of Qixian County) =) may not be a reliable indicator of the pollution, if A one-jaw grid is used to subdivide the wafer mapping, then the best use of jp «beginning and + APL analysis can detect very regional signals. After recording-wafer mapping, you can apply-virtual grid, And use appropriate software to display it on top of the wafer map. The virtual grid can be used to perform the analysis. At the same time, the local area analysis also allows better positioning of problem locations in the structure. For example, 'can indicate an unqualified grid Components, when the grid component allows us to detect the concerns in more detail < Micro area scan can be started at the same location when the area is; and records of defective components in wafer mapping analysis can be output in an appropriate form. The local grid method is not limited to numerical analysis based on average values. Any appropriate pre-defined parameters (including average intensity, PL minimum maximum, standard deviation, and baseline) can be used to determine the contaminated area. Due to the uneven variation in the signal throughout the wafer, the pL signal baseline method can be a more useful parameter. This technique is described below. Figure 7a shows a typical wafer map, and Figure 7b shows the relative histogram of PL intensity. Contamination is detected in wafer mapping by deviation from the baseline value, and limits can be set. However, the mean value in the wafer map is corrected by contamination. The value used should represent the signal level of an uncontaminated wafer. The peaks shown in Figure 8 represent the true PL values of an uncontaminated wafer. Modifying the baseline function to have a peak value will allow the user to accurately track pollution with higher sensitivity. 86614 -26- 200427977 An appropriate algorithm includes the following steps: 1. Define the peak as the maximum value in the histogram and use it for the baseline. 2 · Search peak data. 3. Then calculate the maximum value of ± 70% (which is specified by the user), and then redefine the peak maximum value to the center position of these points. 4. The exact value of the peak maximum is then calculated and then the pL level is defined. 5. It also allows the user to enter typical baseline values from uncontaminated wafers. It will be helpful if there are two peaks of equal intensity. The baseline of the wafer is the pL value corresponding to the maximum number of points. Baseline deviation is defined by the following relationship: Baseline deviation = PL value-Baseline PL value is the AVG pL of each element of the grid. The limit deviation must be measured for each component. [Schematic description]
以上參照所附圖式圖丨—8，以舉例說明之方式描述本 明，其中： X 圖1係對用以獲取PL資料之適當裝置的圖解說明； 圖2係一展示如何處理資料之示意圖； 圖3展示一 pl映射與藉由習知方法測定之雜質量的相互 關係； 圖4展不作爲材料規袼之決定因素的PL量測之使用； 圖5展不在一晶圓上規格區域内外之一空間解析映射； 圖6展示一晶圓前後之PL影像； 圖7及8展不根據本發明之一種可能的數值分析技術。 86614 200427977 【圖式代表符號說明】 1 樣品臺 2 樣品 3-8雷射 9 光學纖維 10準直儀 11雷射束擴展器 12 切趾板 1 3, 1 8, 23, 27 濾光輪 _ 14折疊鏡 1 5頻率車生器 16雷射選擇器 17冷鏡 19孔徑輪 20壓電致動器 2 1紅外線偵測器 22,24聚焦透鏡 β 25 UV及可視區域偵測器 2 6鎖定放大器 28變焦透鏡 29 CCD二維陣列 30自動聚焦控制器 3 1光束分離器 32物鏡分度裝置 86614 -28 - 200427977 3 3壓電驅動聚焦臺 34, 35顯微鏡物鏡 36轉臺 37鏡筒透鏡 3 8 光學位移量測系統 3 9顯示幕 40微處理器 1 0 1 樣品The above description describes the present invention by way of example with reference to the attached drawings 丨 -8, where: X FIG. 1 is a diagrammatic illustration of a suitable device for obtaining PL data; FIG. 2 is a schematic diagram showing how to process the data; Figure 3 shows the correlation between a pl mapping and the amount of impurities measured by conventional methods; Figure 4 shows the use of PL measurement which is not a determining factor for material specifications; Figure 5 shows the inside and outside of the specification area on a wafer A spatially resolved map; Figure 6 shows the PL images before and after a wafer; Figures 7 and 8 show a possible numerical analysis technique according to the present invention. 86614 200427977 [Illustration of Symbols in the Drawings] 1 Sample Stage 2 Sample 3-8 Laser 9 Optical Fiber 10 Collimator 11 Laser Beam Extender 12 Apodization Plate 1 3, 1 8, 23, 27 Filter Wheel_ 14 Folding Mirror 1 5 frequency car generator 16 laser selector 17 cold mirror 19 aperture wheel 20 piezoelectric actuator 2 1 infrared detector 22, 24 focus lens β 25 UV and visible area detector 2 6 lock amplifier 28 zoom Lens 29 CCD two-dimensional array 30 Auto focus controller 3 1 Beam splitter 32 Objective indexing device 86614 -28-200427977 3 3 Piezo driven focusing stage 34, 35 Microscope objective 36 Turntable 37 Tube lens 3 8 Optical displacement Test system 3 9 display screen 40 microprocessor 1 0 1 sample
102收集裝置 1 0 3處理器 1 0 4 比較器 1 0 5資料儲存器 I 0 6控制單元 110處理單元 111資料儲存器 II 2視覺顯示幕102 Collection device 1 0 3 Processor 1 0 4 Comparator 1 0 5 Data storage I 0 6 Control unit 110 Processing unit 111 Data storage II 2 Visual display
86614 29-86614 29-