201207579 六、發明說明: 【發明所屬之技術領域】 本發明是關於檢測出形成於罩幕的罩幕對準標記,及 形成於工件的工件對準標記,使兩者以成爲事先設定的位 置關係的方式進行對位,經由罩幕將光線照射於工件的曝 光裝置中’檢測出工件對準標記所用的檢測方法,尤其是 有關於作爲工件對準標記,使用形成於印刷基板等工件的 圓形凹部的情形的工件對準標記的檢測方法者。 【先前技術】 在藉由微影成像術來製造半導體元件、印刷基板、液 晶基板等的圖案的工序中,使用著曝光裝置。曝光裝置是 將形成罩幕圖案的罩幕、及轉印該圖案的工件對位(對準) 於所定的位置關係,之後,於塗布光阻的工件經由罩幕進 行照射曝光光。藉此,使得罩幕圖案被轉印(曝光)至工 件。 曝光裝置的罩幕與工件的對位,一般爲如下地進行。 (1) 藉由對準顯微鏡檢測出形成於罩幕的對準標記(以 下稱爲罩幕標記)、及形成於工件的對準標記(以下稱爲工 件標記)。 (2) 在裝置的控制部中進行畫像處理藉由對準顯微鏡所 檢測的罩幕標記與工件標記,求出各個的(對準顯微鏡的 視野的)位置座標。 (3) 使得兩者的位置成爲事先所設定的位置關係的方 201207579 式,進行移動罩幕或工件。 又,罩幕與工件,是必須進行平面內的2方 向與Y方向)及旋轉方向(8方向)的對位。因此,奪 與工件標記是分別形成於兩處以上。 在第6圖表示檢測出工件標記的對準顯微鏡 略構成。又’如上述地’罩幕標記及工件標記是另 兩處以上’因此對準顯微鏡10也因應於其而設灑 上,惟在同圖中,僅表示1個(1處)》 對準顯微鏡10是由:半反射鏡l〇a、透鏡L1 CCD攝影機10b所構成。11是進行畫像處理等纪 置,12是監測器,W是形成有工件標記WAM的工 作爲檢測出工件標記WAM的方法眾知者,笔 匹配所致的檢測,及依邊緣檢測所致的檢測。又, 謂檢測不僅取入對準標記畫像,也包括檢測其位s 者。 針對於圖案匹配,例如在日本專利文獻1的段 〜0009有簡單的說明。 此方法,例如,利用對準顯微鏡1 〇進行擴大 的表面,再利用CCD攝影機進行攝影。控制部11 攝影畫像內,檢測出與經註冊的工件標記的畫像-位(圖案),進行求出其位置的畫像處理》 邊緣檢測是在攝影畫像內,將亮度變化大的音丨 爲圖案的邊緣進行檢測出圖案的方法。使用第7 1 說明。 向(X方 〖幕標記 10的槪 、別形成 i兩處以 、L2及 丨演算裝 件。 「依圖案 在此所 I (座標) 落 0004 工件 W 是掃描 •致的部 !分考量 1簡單地 -6- 201207579 第7(a)圖是利用對準顯微鏡的CCD所攝影的工件表 面的畫像的例子。當圖案形成於工件的表面,則其部分是 照明光的反射率會變化’與其他部分在明亮(亮度)不相 同。圖中附有陰影線的橢圓形的部分爲圖案。 如第7(b)圖所示地’在上述所攝影的畫像中,以—定 間隔朝一方向(圖中箭號的方向)進行微分畫像的亮度分 布,並將微分値(亮度變化)的最大部分作爲圖案的邊緣(圖 中的黑圓形·)。 如第7(c)圖所示地,將此在攝影畫像全體重複地進 行,並將所求得的邊緣彼此間連結所得到的圖形作爲圖 案。又,演算此圖案的例如重心,作爲圖案的位置。 (先前技術文獻) (專利文獻) 專利文獻1 :日本特開200 1 - 1 1 0697號公報 【發明內容】 (發明所欲解決之課題) 工件爲印刷基板時,作爲工件標記,有利用藉由雷射 照射形成於基板的孔[凹部,以下稱爲通孔(via hole)]的情 形。其理由是如下所述。 在製造印刷基板中,爲了取得多層形成的配線層間的 耦合,藉由雷射照射形成多數Φ ΙΟΟμηι左右的通孔(貫通 層間的孔)。所以,形成層間耦合用的通孔,順便將工件 標記例如形成於工件的周邊部,就不必另外形成對準標記 201207579 所用的工序。 但是,印刷基板1 0 0是形成通孔1 〇 1之後’如第8 (a) 圖所示地,在表面施以銅102等的鍍金屬’之後,爲了曝 光,如第8(b)圖所示地,貼著所謂乾膜光阻的厚度數十 μηι薄膜狀的光阻(光阻薄膜)1 〇 3。所以,形成作爲工件標 記的通孔101是成爲藉由光阻薄膜103所覆蓋。 作爲工件標記的通孔,是隔著如此地被黏貼的光阻薄 膜1 03,利用對準顯微鏡被觀察。這時候’以通孔1 〇 1的 開口部分的光阻薄膜1 03的鬆弛情況’利用對準顯微鏡的 攝影手段(C C D)的通孔1 0 1的觀看態樣(形狀或明暗或色調) 有種種變化。 第9(a)圖至第9(f)圖是攝影被光阻薄膜所覆蓋的通孔 的畫像例。如同圖所示地,通孔是看似白色,或看似淺灰 色至深灰色,或在白色中看似黑色之點等觀看態樣爲各式 各樣。 如此地,若通孔亦即工件標記的觀看態樣有變化,則 在檢索與所註冊的畫像一致的圖案的圖案匹配中,有無法 檢測出工件標記的情形。 所以,在此種工件標記的檢測中,發明人等考慮了不 是進行圖案匹配,而是進行了依邊緣檢測所致的圖案檢 測。若爲邊緣檢測,則即使工件標記的觀看態樣不相同’ 亦由於在通孔的邊緣部分產生亮度變化,因而考慮到可檢 測工件標記(通孔)。 如上述地,在邊緣檢測中,將畫像中的亮度變化最大 -8- 201207579 部分認作爲邊緣’而連續其邊緣來檢測出圖案。圖案爲如 通孔似爲圓形時’則將所檢測到的各邊緣連結成近似圓 形’而演算該圓的中心位置作爲圖案的位置。 然而’若直接適用習知的邊緣檢測方法,則知道會有 無法檢測出工件標記的正確位置的情形。 本發明是如被光阻薄膜所覆蓋的通孔等般,將觀看態 樣依狀況而異的圖案使用作爲工件標記時,明瞭在習知的 邊緣檢測中無法檢測出工件標記等圖案的正確位置的原 因’可據此來檢測出圖案的正確位置,本發明的目的是在 於提供一種一面使用邊緣檢測方法,一面可正確地檢測出 觀看態樣爲各式各樣的工件標記等圖案的位置的對準標記 的檢測方法。 (用以解決課題的手段) 針對於在習知的邊緣檢測中有無法檢測出工件標記的 正確位置的情形,經精心檢討之結果,在其一原因發現到 如以下的理由。 如上述地,在習知中依邊緣檢測所致的圖案位置檢測 是採取將亮度的變化爲最大的部分檢測作爲邊緣,連結所 檢測到的邊緣成爲圓形近似,演算該圓的中心位置的順 序。 如第9圖所示地,通孔是依黏貼於其上面的光阻薄膜 而有各種不相同的觀看態樣,惟大致上孔中是白色或淺灰 色,工件(基板)的部分是深灰色,通孔的邊緣部分爲黑 -9- 201207579 色。 在此,如第10(a)圖所示地’若覆蓋通孔 薄膜1 03,以左右均等的狀態下垂在孔中’則 部分的中心與通孔的中心大約一致。 但是,如第10(b)、(c)所示地,若光阻薄1 孔101中偏向一方而下垂,則看似白色的部分 孔的中心並不一致。此種現象是光阻薄膜103 101中起皺紋之狀態下也會產生。 如上述地,在邊緣檢測中,將亮度變化的 測出作爲邊緣。在如第1 0圖所看到的通孔(工 中,亮度變化的最大部分,是看似白色的部分 部分的界限部分,此爲,亦即將看似白色的部 爲工件標記。 但是,原來的工件標記的位置是通孔1 0 1 第10(a)圖所示地通孔101的中心與看似白色 —致時較佳,惟如第9(b),(c)圖所示地看似白 心與通孔1 0 1的中心未一致時,是無法檢測 (通孔)1 01的正確位置。 依據上述發現,爲了解決此問題,發明人 討,而想到如下。通孔的邊緣是在畫像中爲最 周邊的基板的底子的深灰色部分之間,有稍 化。 因此,例如從通孔的中心朝著周邊進行觀 化時,若將從此黑色變化成深灰色的部分作爲 1 〇 1的光阻 看似白色的 漠103在通 的中心與通 即使在通孔 最大部分檢 件標記)1 0 1 與看似黑色 分檢測成作 的位置,如 部分的中心 色部分的中 出工件標記 等再加以檢 黑,而與其 微亮度的變 察顏色的變 邊緣,而求 -10- 201207579 出近似於連結此複數邊緣位置所形成的封閉曲線的圓的中 心,就可正確檢測出通孔(工件標記)的位置。 爲了實現此,以如以下的順序進行邊緣檢測。 與習知同樣地,求出利用對準顯微鏡進行攝影的工件 的畫像的亮度分布,有關於所定方向微分其亮度而求出亮 度的變化。 但是,並不是如習知地將亮度變化(微分値)的最大部 分作爲邊緣,針對圖案的畫像,朝著複數的放射方向求出 對於距離的亮度分布,並將亮度變化的極大値或極小値的 位置,針對各方向分別抽出每一個而加以組合。 又,求出近似於通過上述極大値或極小値之位置的封 閉曲線的複數圓,該複數圓之內,選擇半徑(直徑也可以) 爲與工件對準標記的半徑(直徑也可以)最近的圓,求出其 中心位置。 例如,從圖案的畫像的中心附近朝著周邊來觀察顏色 的變化’將亮度變化的微分値爲事先所設定的値以上而上 面有凸出部分(在黑-> 白的變化時微分値爲正的情形)的微 分値的位置作爲邊緣,將連結如此地所求出的邊緣彼此間 的封閉曲線作爲圖案。 接著’若如此地所得到的圖案大小,爲與工件標記 (通孔)的設計値的大小大約一致,則將此作爲工件標記, 演算其中心位置而求出。 亦即’在本發明中,如下所述地來解決上述課題。 (1)取得形成於工件的圖案的畫像。 -11 - 201207579 (2) 針對於從上述取得的圖案的畫像中心附近,或是朝 著中心方向的複數放射方向,求出對於距離的亮度的分 布。 (3) 微分上述求出的亮度分布,求出對於距離的亮度的 變化的微分値,針對於上述各放射方向求出該微分値的極 大値或極小値的位置。 (4) 求出複數近似於通過分別抽出每一個針對於上述各 放射方向所求出的上述極大値或極小値的位置而加以組合 之位置之封閉曲線的圓。 (5) 比較上述複數的圓的半徑(直徑)及圓形的工件對準 標記的半徑(直徑),由上述複數的圓中,選擇與工件對準 標記的半徑(直徑)最近的圓。 (6) 演算上述選擇的圓的中心位置,並將該中心位置作 爲工件對準標記的位置。 (發明之效果) 在本發明中,求出近似於通過亮度變化的微分値或極 小値的位置之封閉曲線的複數圓,選擇最接近於工件對準 標記的大小的圓而求出工件對準標記的位置之故,因而即 使有光阻薄膜黏貼於工件,也可檢測出工件標記的通孔的 正確位置。 【實施方式】 第1圖是表示本發明的一適用對象的投影曝光裝置的 -12- 201207579 構造的圖式。在同圖中,MS是罩幕平台。在罩幕平台 MS,置放保持著形成有罩幕標記MAM與罩幕圖案MP的 罩幕Μ。 從光照射裝置1出射曝光光。所出射的曝光光是經由 罩幕Μ,投影透鏡2,被照射在被載置於工件平台W S上 的塗布光阻的工件W上’罩幕圖案ΜΡ被投影在工件W 上而被曝光。 在投影透鏡2與工件W之間,有可朝同圖的箭號方 向移動的對準顯微鏡10設置於兩處。在將罩幕圖案MP 曝光於工件W上之前’將對準顯微鏡10插入在圖示之位 置,檢測出罩幕標記 mam與形成於工件的工件標記 WAM,進行罩幕Μ與工件w的對位。對位之後,對準顯 微鏡1 〇是從工件W上避開。又’在第1圖中’僅表示設 置兩處內的一方的對準顯微鏡。 如上述地,對準顯微鏡1 〇是由:半反射鏡1 〇a ’透鏡 LI,L2及CCD攝影機l〇b所構成。在對準顯微鏡1〇,設 有照明進行攝影的工件w的表面所用的照明手段1 0C。 利用對準顯微鏡10的CCD攝影機l〇b進行受像的罩 幕標記MAM像,工件標記WAM像等,是被傳送至控制 部11。 控制部1 1是具備記憶著處理以上述CCD攝影機1 〇b 進行受像的畫像的畫像處理部1 1 a,工件標記的大小,罩 幕標記的位置座標資訊,邊緣檢測所用的參數等各種參數 的記憶部1 1 b。 -13- 201207579 又,控制部1 1是具備:從以CCD攝影機1 Ob進行受 像而在畫像處理部1 1 a經畫像處理的畫像進行邊緣檢測, 將利用邊緣檢測所得的圖案形狀,與事先註冊的工件標記 形狀比較評價,來判定此圖案是否爲作爲工件標記所檢測 的圖案,而檢測出中心位置座標的工件標記中心位置檢測 部1 1 c,及檢測出作爲工件標記的圖案的位置座標,爲一 致於被記憶在記憶部1 1 b的罩幕標記像的位置座標的方 式,移動工件平台WS或是罩幕平台MS(或是其雙方)的對 位控制部〗1 d,及用以利用作業人員的指示將工件標記的 大小’或邊緣檢測的條件(下述的微分臨界値)等註冊至記 憶部1 1 b的註冊部1 1 e。 工件平台WS或罩幕平台MS,是利用藉由上述對位 控制部1 1 d被控制的工件平台驅動機構4,罩幕平台驅動 機構3被驅動,而進行XY方向(X,γ:平行於罩幕平台 MS’工件平台WS面而互相地正交的2方向)移動,而且 以垂直於XY平面的軸爲中心進行旋轉。 上述控制部1 1中,連接監測器1 2,以上述畫像處理 部1 1 a畫像處理之畫像則於監測器1 2之畫面,例如第4 圖地加以顯示。 使用上述第4圖及第1圖,第2圖,第3圖,針對於 依在控制部所進行的邊緣檢測所致的工件標記檢測的方法 進行說明。 又,第2圖是表示所檢測出的畫像(與模式圖時之畫 像不相同),及此畫像的放射方向的亮度分布及其微分 • 14- 201207579 値’第3圖是表示近似於連結所檢測出的畫像的亮度的微 分値的極大値或是極小値的位置的封閉曲線的複數圓。 在第1圖中,控制部1 1是從如第4圖所顯示的畫像 中’進行檢索工件標記。在此檢索例如使用被稱爲探針檢 測的方法。探針檢測是欲檢測在2値(黑白)畫像內具有同 一濃度的像素的集合體(例如圓形或四角形等形狀的部 分)’ 一般所使用的方法。又,欲檢索工件標記,也可使 用探針檢測以外的檢測方法。 亦即,控制部1 1是由利用探針檢測,利用對準顯微 鏡1 0被檢測出的如第4圖的畫像,選擇表示工件標記形 狀的圓形形狀的部分。接著,表示其圓形的形狀的部分成 爲對準顯微鏡1 0的視野中心的方式移動工件平台w S, 又’轉換對準顯微鏡1 0的倍率,經擴大而受像於CCD攝 影機1 〇b。 第2(a)圖是作成如上述地利用對準顯微鏡1〇的CCD 攝影機1 〇b進行受像的工件標記WAM的通孔的畫像的模 式圖。在工件黏貼著光阻薄膜,而覆蓋通孔的上面。 同圖的白色部分爲通孔之孔的部分,而其周圍的顏色 深的部分爲通孔的邊緣部分,又其周圍的顏色稍淺的部分 爲工件(基板)的部分。 首先’控制部1〗的工件標記中心位置檢測部U c,是 由如第2 (a)圖地受像的圖案的畫像之中心附近,如同圖的 虛線A所示地針對於複數放射方向,求出對於距離的畫像 的亮度分布。又’圖案畫像的基本上的中心位置,是利用 -15- 201207579 上述探針檢測予以求出。 在第2(b)圖,表示有關於某一方向的亮度分布。橫軸 爲放射(掃描)方向的距離’縱軸爲亮度。如同圖所示地, 通孔的內側(孔的部分)是映出白色而亮度高。由此朝著外 側前進而達到通孔的邊緣會變黑色而亮度急速地降低。當 從通孔之邊緣移到基板的部分,則亮度是稍上昇。 接著,工件標記中心位置檢測部1 1 C是微分如第2(b) 圖地所得到的亮度分布的曲線。其結果爲第2(c)圖。橫軸 爲放射(掃描)方向的距離’縱軸是微分値,亦即爲亮度變 化。 如同圖所示地,亮度的變化是亮度從明亮變化成黑暗 (從白色變成黑色)的部分(圖中白邊緣1:微分値的極小値) 爲最大。習知是把此亮度變化最大的白邊緣1的部分檢測 出作爲工件標記(通孔)的邊緣。 但是,在本發明中,不僅該白邊緣1,也檢測出亮度 由黑暗變化成明亮(由黑色至白色)的黑邊緣。爲了此採用 如下的手法。爲了忽略白邊緣1,事先在控制部設定微分 値的臨界値,某一値以下(負値)的微分値是作成忽略。由 白色變化成黑色的部分是微分値成爲負。因此,白邊緣是 被忽略。 接著,將微分値爲臨界値以上而在上面有凸部的極大 値部分,亦即將亮度由黑色變化成白色的部分(黑邊緣), 檢測出作爲邊緣的候補點。在第2(c)圖中,黑邊緣(極大 値)是以黒邊緣1,2,3所示的3處。在第2(d)圖中,表 -16- 201207579 示畫像的白邊緣與黑邊緣的部位。 如此地’針對於各放射方向求出黑邊緣的位置。又針 對於所檢測出的各黑邊緣,如第3圖所示地,連結最近的 黑邊緣彼此間而近似作成圓形。亦即,針對於各放射方向 分別每一個地抽出亮度變化的極大値位置經組合,求出近 似於通過上述極大値位置的封閉曲線的複數圓。 接著,由所得到的各圓計算出各個的半徑r (或是直 徑)。在控制部,輸入有事先工件標記的通孔的半徑(或是 直徑)(設計値),將各圓的半徑r(或是直徑)與通孔的半徑 (或是直徑)作比較,將具有與通孔的半徑(或是直徑)最近 數値的圓作爲通孔的邊緣,亦即工件標記的圖案。接著, 求出其圓的中心位置,將此作爲工件標記的位置。 在第3圖中,表示圓是連結黑邊緣1的圓C1,連結 黑邊緣2的圓C 2,連結黑邊緣3的圓C 3的3種類,惟不 僅爲該3種類的圓,例如同圖所示地,也有描繪圓C 1〜 C3以外的圓C4,圓C5等的情形。 此爲,如同圖所示地,例如未能檢測出黑邊緣1〜3 之內的一部分,或是將畫像上的雜訊檢測出作爲黑邊緣 等,在各放射方向中,並不一定求出正確的黑邊緣1〜3 的位置的情形之故。 因此,針對於各放射方向分別抽出各一個亮度變化的 極大値的位置經組合,而如上述地近似複數之圓,針對各 圓如上述地求出如以下U)所述的半徑記分推定表示邊緣位 置的圓。 -17- 201207579 又’因應於需要,求出如以下的(b)〜(d)的記分予以 合計,將記分最高的圓,推定爲表示工件標記(通孔)的邊 緣位置的圓。 (a) 半徑記分 如上述地以圓的半徑r接近多少設計値作爲指標而求 出半徑記分。亦即,若半徑r爲與設計値相同,則評價爲 一分,而若未成爲所定値以上則評價爲0分。具體上,例 如利用以下式求出半徑記分。 半徑記分=1-丨近似圓半徑-設計半徑|/a (b) 邊緣數記分 使用幾個邊緣而是否圓近似作爲指標以求出邊緣數記 分。亦即,所使用的邊緣數愈多而記分變高。具體上,例 如利用以下式求出邊緣數記分。 邊緣數記分=近似的邊緣數/b (c) 邊緣位範圍記分 以邊緣距近似圓的圓周多少距離作爲指標求出邊緣位 置範圍記分。例如第5(a)圖所示地,全邊緣在圓周上時, 則記分成爲滿分,而如第5(b)圖所示地,邊緣未在圓周上 時’則記分是變低。具體上,例如利用以下式求出邊緣位 置範圍記分。 邊緣位置範圍記分=1 -(距圓周的距離的最大最小的差)/c -18- 201207579 (d)邊緣對稱性記分 以相對的邊緣被檢測多少作爲指標求出邊緣對稱性記 分。例如放射方向爲6方向時’如第5 (c)圖所示地,若三 個成對(pair)全存在則記分變高’而如第5(d)圖所示地, 若未存在成對時,則記分變低。具體上,例如利用以下式 求出邊緣對稱性記分。 邊緣對稱性記分=(α +雙方檢測成對數Μ +單方檢測成對數/r )/d 又,在上述第2(c)圖中,表示由白色變化成黑色的極 小値,由黑色變成白色的極大値的例子,惟變更亮度變化 的採用方法,也有由白色變成黑色的極小値,由黑色變成 白色的極小値的情形。其情形爲將極大値的位置檢測作爲 白邊緣的位置,而將極小値的位置檢測作爲黑邊緣的位 置。 在第1圖的裝置中,在工件標記中心位置檢測部1 1 C 如上述地求出工件標記的位置,依據此進行罩幕Μ與工件 W的對位’罩幕Μ與工件W的對位,如下地進行。 (1 )將照射光從光照射裝置1或未圖示的對準光源照射 至罩幕Μ,經由對準顯微鏡1 〇的CCD攝影機1 〇b進行受 像罩幕標記M A Μ像’而傳送至控制部1 1。控制部1 1的 畫像處理部11a是將上述罩幕標記ΜΑΜ像變換成位置座 標而被記憶在記憶部1 1 b。 -19- 201207579 又,罩幕標記的檢測方法是各種方法被提案,視需要 例如請參考專利文獻1等。 (2) 接著,將照明光從對準顯微鏡10的照明手段10c 照射至工件 W,如上述地進行邊緣檢測,檢測出工件 W 上的工件標記WAM,而控制部1 1是求出其位置座標。 (3) 控制部11是所記憶的罩幕標記MAM的位置座 標,及所檢測的工件標記WAM的位置座標成爲所定的位 置關係的方式,移動工件平台WS(或是罩幕平台MS,或 是其雙方),來進行罩幕Μ與工件W的對位。 【圖式簡單說明】 第1圖是表示本發明的一適用對象的投影曝光裝置的 構成圖。 第2(a)圖至第2(d)圖是表示所檢測的畫像(模式圖), 及此畫像的放射方向的亮度分布與其微分値的圖式。 第3圖是表示近似於連結畫像的亮度的微分値的極大 値或極小値的位置的封閉曲線的複數圓的圖式。 第4圖是表示顯示於監測器上的工件標記像的圖式。 第5(a)圖至第5(d)圖是說明近似的圓的記分的圖式。 第6圖是表示檢測工件標記的對準顯微鏡1 〇的槪略 構成的圖式。 第7(a)圖至第7(c)圖是說明依邊緣檢測所致的圖案檢 測方法的圖式。 第8(a)圖至第8(b)圖是表示使用作爲工件標記的通孔 -20 - 201207579 的形狀例的圖式。 第9(a)圖至第9(f)圖是表示將被覆蓋於光阻薄膜的通 孔予以攝影的畫像例的圖式。 第10(a)圖至第10(c)圖是說明被覆蓋於光阻薄膜的通 孔的觀看態樣的圖式。 【主要元件符號說明】 1 :光照射裝置 2 :投影透鏡 3 =罩幕平台驅動機構 4 ·工件平台驅動機構 1 〇 :對準顯微鏡 1 0 a :半反射鏡 10b : CCD攝影機 l〇c :照明手段 1 1 :控制部 1 1 a :畫像處理部 1 1 b :記憶部 1 1 c :工件標記中心位置檢測部 1 1 d :對位控制部 1 1 e :註冊部 1 2 :監測器 L1,L 2 :透鏡 Μ :罩幕 -21 - 201207579 MS :罩幕平台 MAM :罩幕標記(罩幕對準標記) MP :罩幕圖案 W :工件 W S :工件平台 WAM :工件標記(工件對準標記) -22201207579 VI. Description of the Invention: [Technical Field] The present invention relates to detecting a mask alignment mark formed on a mask and a workpiece alignment mark formed on the workpiece so that the two are set to a predetermined positional relationship The method of detecting the position of the workpiece alignment mark by illuminating the light through the mask to the workpiece, in particular, as a workpiece alignment mark, using a circle formed on a workpiece such as a printed substrate. The method of detecting the workpiece alignment mark in the case of the recess. [Prior Art] An exposure apparatus is used in the process of manufacturing a pattern of a semiconductor element, a printed circuit board, a liquid crystal substrate or the like by lithography. In the exposure apparatus, the mask forming the mask pattern and the workpiece to which the pattern is transferred are aligned (aligned) in a predetermined positional relationship, and then the workpiece coated with the photoresist is irradiated with exposure light through the mask. Thereby, the mask pattern is transferred (exposed) to the workpiece. The alignment of the mask of the exposure apparatus with the workpiece is generally performed as follows. (1) An alignment mark (hereinafter referred to as a mask mark) formed on a mask and an alignment mark (hereinafter referred to as a workpiece mark) formed on a workpiece are detected by an alignment microscope. (2) Performing image processing in the control unit of the device The position marks of each (aligned with the field of view of the microscope) are obtained by aligning the mask mark and the workpiece mark detected by the alignment microscope. (3) Move the mask or workpiece by making the position of the two positions the previously set positional relationship 201207579. Further, the mask and the workpiece must be aligned in the two directions and the direction of rotation (eight directions) in the plane. Therefore, the workpiece marks are formed in two or more places. Fig. 6 shows an outline of an alignment microscope that detects a workpiece mark. In addition, 'the above-mentioned 'mask mark and workpiece mark are two or more', so the alignment microscope 10 is also sprinkled according to it, but in the same figure, only one (1) is shown. 10 is composed of a half mirror l〇a and a lens L1 CCD camera 10b. 11 is the image processing and the like, 12 is the monitor, W is the work of forming the workpiece mark WAM, the method is known to detect the workpiece mark WAM, the detection by the pen matching, and the detection by the edge detection . In addition, the detection includes not only the alignment mark image but also the position s. For the pattern matching, for example, a paragraph 0009 of Japanese Patent Document 1 has a brief description. This method, for example, uses an alignment microscope 1 to perform an enlarged surface, and then photographs using a CCD camera. The control unit 11 detects an image-bit (pattern) of the registered workpiece mark in the photographed image, and performs image processing for obtaining the position. Edge detection is a pattern in which a brightness change is large in the photograph image. The edge performs a method of detecting a pattern. Use the 7th 1 description. To the X side, the 标记 幕 幕 , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , -6- 201207579 Figure 7(a) is an example of a portrait of the surface of the workpiece photographed by the CCD of the alignment microscope. When the pattern is formed on the surface of the workpiece, the reflectance of the illumination light is changed in part with the other parts. The brightness (brightness) is different. The elliptical part with the hatching in the figure is a pattern. As shown in Fig. 7(b), 'in the above-mentioned photographed image, the direction is in a certain direction (in the figure) The direction of the arrow is used to perform the luminance distribution of the differential image, and the largest part of the differential 値 (luminance change) is used as the edge of the pattern (black circle in the figure). As shown in Fig. 7(c), A pattern obtained by connecting the obtained edges to each other is used as a pattern. Further, for example, the center of gravity of the pattern is calculated as the position of the pattern. (Prior Art Document) (Patent Literature) Patent Literature 1 : Japan Special 2 00 1 - 1 1 0 697 OBJECTIVE OF THE INVENTION PROBLEM TO BE SOLVED BY THE INVENTION When a workpiece is a printed circuit board, a hole formed in the substrate by laser irradiation is used as a workpiece mark [recessed portion, hereinafter referred to as a through hole) In the case of manufacturing a printed circuit board, in order to obtain coupling between wiring layers formed by a plurality of layers, a plurality of through holes (through holes between the layers) of a plurality of Φ ΙΟΟμηι are formed by laser irradiation. Therefore, the through hole for interlayer coupling is formed, and the workpiece mark is formed, for example, on the peripheral portion of the workpiece, so that it is not necessary to separately form the process used for the alignment mark 201207579. However, the printed substrate 100 is formed after the through hole 1 〇1 'As shown in Fig. 8(a), after applying a metal plating such as copper 102 to the surface, for exposure, as shown in Fig. 8(b), the thickness of the so-called dry film photoresist is tens薄膜ηι Film-shaped photoresist (photoresist film) 1 〇 3. Therefore, the through hole 101 which is formed as a workpiece mark is covered by the photoresist film 103. The through hole as a workpiece mark is pasted so as to be adhered thereto of The resist film 101 was observed using an alignment microscope. At this time, the 'relaxation of the photoresist film 103 with the opening portion of the through hole 1 〇1' utilizes the through hole of the photographic imaging means (CCD) of the alignment microscope 1 0 1 There are various changes in the viewing state (shape or shading or color tone). Figures 9(a) to 9(f) are examples of the through holes in which the photoresist is covered by the photoresist film. The hole looks like white, or looks like light gray to dark gray, or looks like black in white. It looks like a variety of views. So, if the through hole, that is, the view of the workpiece mark, changes. In the pattern matching of the pattern matching the registered portrait, there is a case where the workpiece mark cannot be detected. Therefore, in the detection of such a workpiece mark, the inventors have considered that pattern matching by edge detection is performed instead of pattern matching. In the case of edge detection, even if the viewing state of the workpiece mark is different, the brightness of the workpiece is detected in the edge portion of the through hole, so that the workpiece mark (through hole) can be detected. As described above, in the edge detection, the portion of the image in which the brightness change is maximum -8 - 201207579 is regarded as the edge ' while the edge is continuous to detect the pattern. When the pattern is such that the through hole is circular, the detected edges are connected to an approximately circular shape, and the center position of the circle is calculated as the position of the pattern. However, if the conventional edge detection method is directly applied, it is known that there is a case where the correct position of the workpiece mark cannot be detected. In the present invention, when a pattern in which a viewing condition varies depending on a state of view is used as a through hole covered by a photoresist film, it is understood that the correct position of a pattern such as a workpiece mark cannot be detected in the conventional edge detection. The reason for this is that the correct position of the pattern can be detected accordingly. It is an object of the present invention to provide a method for detecting the position of a pattern such as a variety of workpiece marks on the one side using the edge detection method. Alignment mark detection method. (Means for Solving the Problem) In the case where the correct position of the workpiece mark cannot be detected in the conventional edge detection, the results of careful examination revealed the following reasons. As described above, in the conventional method, the position detection by the edge detection is performed by taking a partial detection that maximizes the change in luminance as an edge, and connecting the detected edge to a circular approximation, and calculating the order of the center position of the circle. . As shown in Fig. 9, the through holes are in various different viewing states depending on the photoresist film adhered thereto, but the holes are generally white or light gray, and the portion of the workpiece (substrate) is dark gray. The edge portion of the through hole is black-9-201207579 color. Here, as shown in Fig. 10(a), if the through hole film 103 is covered and sagged in the hole in a state where the right and left are equal, the center of the portion substantially coincides with the center of the through hole. However, as shown in the tenth (b) and (c), when the photoresist thin film 1 is tilted toward one side and sags, the center of the partial holes which are white is not uniform. This phenomenon is also caused in a state where the photoresist film 103 101 is wrinkled. As described above, in edge detection, the measurement of the change in luminance is taken as an edge. In the through hole as seen in Fig. 10 (the largest part of the brightness change in the work, it is the boundary portion of the portion that looks white, and this is the part that appears to be white as the workpiece mark. However, the original The position of the workpiece mark is the through hole 1 0 1 . The center of the through hole 101 shown in Fig. 10(a) is preferably white, as shown in Fig. 9(b), (c). When the center of the hole does not coincide with the center of the through hole 110, the correct position of the (through hole) 101 cannot be detected. According to the above findings, in order to solve this problem, the inventors have thought of the following. In the case of the image, the dark gray portion of the base of the most peripheral substrate is slightly reduced. Therefore, for example, when viewing from the center of the through hole toward the periphery, the portion that changes from black to dark gray is taken as 1 The photoresist of 〇1 looks like the white desert 103 in the center of the pass and the pass even in the largest part of the through hole check mark) 1 0 1 and the position where the black component is detected, such as the central color portion of the part The workpiece mark and the like are checked for black, and the brightness of the workpiece is changed. It becomes an edge, for the sake -10-201207579 an approximate circle connecting the center of this closed curve formed by a plurality of edge positions, the position can be accurately detected through hole (workpiece mark). In order to achieve this, edge detection is performed in the following order. Similarly to the conventional method, the luminance distribution of the image of the workpiece photographed by the alignment microscope is obtained, and the luminance is differentiated in the predetermined direction to determine the change in luminance. However, it is not conventionally known that the largest portion of the luminance change (differential 値) is used as an edge, and for the image of the pattern, the luminance distribution for the distance is obtained toward the radiation direction of the plurality, and the luminance is greatly changed or minimized. The position is extracted for each direction and combined. Further, a complex circle approximated to a closed curve passing through the position of the maximum 极 or the minimum 値 is obtained, and within the complex circle, the selected radius (diameter may also be) is the closest to the radius (diameter of the alignment mark) of the workpiece. Circle and find the center position. For example, the change in color is observed from the vicinity of the center of the image of the pattern toward the periphery. 'The differential of the change in brightness is 値 above the 値 set in advance and the bulge is on the upper part (in the case of black-> white change, the differential is In the positive case, the position of the differential 値 is used as the edge, and the closed curve between the edges thus obtained is connected as a pattern. Then, if the size of the pattern obtained in this manner is approximately the same as the size of the design flaw of the workpiece mark (through hole), the center mark is calculated as the workpiece mark. That is, in the present invention, the above problems are solved as described below. (1) An image of a pattern formed on a workpiece. -11 - 201207579 (2) The distribution of the brightness with respect to the distance is obtained in the vicinity of the center of the image of the pattern obtained as described above or in the direction of the plurality of rays in the center direction. (3) The luminance distribution obtained as described above is differentiated, and the differential 値 of the change in the luminance of the distance is obtained, and the position of the maximum 极 or the minimum 値 of the differential 値 is obtained for each of the radial directions. (4) A circle is obtained which approximates a closed curve of a position where each of the maximum enthalpy or minimum enthalpy obtained for each of the radial directions is extracted and approximated. (5) Compare the radius (diameter) of the above-mentioned plural circle and the radius (diameter) of the circular workpiece alignment mark, and select the circle closest to the radius (diameter) of the workpiece alignment mark from the above-mentioned plural circle. (6) Calculate the center position of the circle selected above and use the center position as the position of the workpiece alignment mark. (Effects of the Invention) In the present invention, a complex circle approximated to a closed curve of a position of a differential enthalpy or a minimum enthalpy by a change in luminance is obtained, and a circle closest to the size of the workpiece alignment mark is selected to obtain a workpiece alignment. The position of the mark, so that even if a photoresist film is adhered to the workpiece, the correct position of the through hole of the workpiece mark can be detected. [Embodiment] FIG. 1 is a view showing a structure of a projection exposure apparatus to which the present invention is applied, -12-201207579. In the same figure, the MS is a mask platform. In the mask stage MS, the mask 形成 which is formed with the mask mark MAM and the mask pattern MP is placed. The exposure light is emitted from the light irradiation device 1. The exposure light that is emitted is projected onto the workpiece W coated with the photoresist on the workpiece stage W S via a mask ’, and the mask pattern 投影 is projected onto the workpiece W to be exposed. Between the projection lens 2 and the workpiece W, there are two alignment microscopes 10 which are movable in the direction of the arrow of the same figure. Before the mask pattern MP is exposed on the workpiece W, 'the alignment microscope 10 is inserted at the position shown, and the mask mark mam and the workpiece mark WAM formed on the workpiece are detected, and the mask Μ is aligned with the workpiece w. . After the alignment, alignment of the microscope 1 is avoided from the workpiece W. Further, in the first drawing, only the alignment microscope in which one of the two places is provided is shown. As described above, the alignment microscope 1 is composed of a half mirror 1 〇a ' lens LI, L2 and a CCD camera 10b. The illumination means 1 0C for illuminating the surface of the workpiece w to be photographed is provided in the alignment microscope. The mask mark MAM image of the image received by the CCD camera 10b of the alignment microscope 10, the workpiece mark WAM image, and the like are transmitted to the control unit 11. The control unit 1 1 is provided with an image processing unit 11 a that memorizes the image to be imaged by the CCD camera 1 〇 b, the size of the workpiece mark, the position coordinate information of the mask mark, and parameters such as parameters for edge detection. Memory unit 1 1 b. -13-201207579 In addition, the control unit 1 1 is configured to perform edge detection on a portrait image processed by the image processing unit 11 a from the CCD camera 1 Ob, and to use a pattern shape obtained by edge detection and register in advance. The workpiece mark shape comparison evaluation determines whether the pattern is the pattern detected as the workpiece mark, and detects the workpiece mark center position detecting portion 1 1 c of the center position coordinate, and detects the position coordinate of the pattern as the workpiece mark. In order to conform to the position coordinates of the mask mark image memorized in the memory portion 1 1 b, the alignment control portion 1d of the workpiece platform WS or the mask platform MS (or both of them) is moved, and The size of the workpiece mark or the condition of the edge detection (derivative threshold 下述 below) or the like is registered to the registration unit 1 1 e of the storage unit 1 1 b by an instruction of the worker. The workpiece stage WS or the mask stage MS is driven by the workpiece stage driving mechanism 4 controlled by the above-described registration control unit 1 1 d, and the mask stage driving mechanism 3 is driven to perform the XY direction (X, γ: parallel to The masking platform MS' is moved in two directions orthogonal to each other in the WS plane of the workpiece platform, and is rotated about an axis perpendicular to the XY plane. The control unit 1 1 is connected to the monitor 1 2, and the image processed by the image processing unit 1 1 a is displayed on the screen of the monitor 12, for example, the fourth figure. The method of detecting the workpiece mark by the edge detection by the control unit will be described with reference to Fig. 4 and Fig. 1, Fig. 2, and Fig. 3 described above. In addition, Fig. 2 is a view showing the detected image (not the same as the image in the pattern diagram), and the luminance distribution and the differential of the radial direction of the image. 14-201207579 第'Fig. 3 shows the approximate connection The maximum value of the differential 値 of the brightness of the detected image or the complex circle of the closed curve of the position of the extremely small 。. In Fig. 1, the control unit 1 1 searches for the workpiece mark from the image "as shown in Fig. 4". Here, for example, a method called probe detection is used. The probe detection is a method generally used to detect an aggregate (for example, a portion such as a circle or a quadrangle) having pixels of the same concentration in a 2 値 (black and white) image. Also, to search for the workpiece mark, a detection method other than probe detection can be used. In other words, the control unit 11 is a portion of the circular shape which indicates the shape of the workpiece mark by the image of the fourth figure which is detected by the probe and detected by the alignment microscope 10. Next, the portion showing the circular shape moves the workpiece stage w S so as to align the center of the field of view of the microscope 10, and the magnification of the alignment microscope 10 is converted to the CCD camera 1 〇b. Fig. 2(a) is a schematic view showing an image of a through hole in which the workpiece mark WAM of the image is received by the CCD camera 1b of the alignment microscope as described above. The photoresist film is adhered to the workpiece to cover the upper surface of the through hole. The white portion of the same figure is the portion of the hole of the through hole, and the dark portion around it is the edge portion of the through hole, and the lighter portion around it is the portion of the workpiece (substrate). First, the workpiece mark center position detecting unit Uc of the 'control unit 1' is the vicinity of the center of the image of the image received by the second image (a), and is directed to the plural radiation direction as indicated by the broken line A in the figure. The brightness distribution of the portrait for the distance. Further, the basic center position of the pattern image is obtained by the probe detection described in -15-201207579. In the second (b) diagram, there is shown a luminance distribution in a certain direction. The horizontal axis is the distance in the radial (scanning) direction, and the vertical axis is the brightness. As shown in the figure, the inside of the through hole (the portion of the hole) reflects white and has high brightness. Thereby, the outer edge is advanced to reach the edge of the through hole, which becomes black and the brightness is rapidly lowered. When moving from the edge of the through hole to the portion of the substrate, the brightness is slightly increased. Next, the workpiece mark center position detecting unit 1 1 C is a curve which differentiates the brightness distribution obtained as shown in the second (b) figure. The result is the 2nd (c) figure. The horizontal axis is the distance in the radial (scanning) direction. The vertical axis is the differential 値, which is the change in brightness. As shown in the figure, the change in brightness is the portion where the brightness changes from bright to dark (from white to black) (white edge 1: small 値 of the differential 图 in the figure) is maximum. It is conventional to detect the portion of the white edge 1 where the change in luminance is the largest as the edge of the workpiece mark (through hole). However, in the present invention, not only the white edge 1, but also the black edge whose brightness changes from dark to bright (from black to white) is detected. In order to do this, the following methods are used. In order to ignore the white edge 1, the threshold 微 of the differential 値 is set in advance in the control unit, and the differential 値 below one (negative 値) is made negligible. The part that changes from white to black is that the differential 値 becomes negative. Therefore, the white edges are ignored. Next, the differential 値 is set to be a critical 値 or more, and a maximum 値 portion of the convex portion is formed thereon, that is, a portion (black edge) whose luminance changes from black to white, and a candidate point as an edge is detected. In the 2nd (c) diagram, the black edges (maximum 値) are 3 places indicated by the edges 1, 2, and 3. In the second (d) diagram, Table -16 - 201207579 shows the white and black edges of the portrait. Thus, the position of the black edge is obtained for each radial direction. Further, for each of the detected black edges, as shown in Fig. 3, the nearest black edges are connected to each other to be approximately circular. That is, the maximum 値 positions for which the luminance changes are extracted for each of the radiation directions are combined to obtain a complex circle which is close to the closed curve passing through the maximum 値 position. Next, the respective radiuses r (or diameters) are calculated from the obtained circles. In the control unit, the radius (or diameter) of the through hole with the workpiece mark is input (design 値), and the radius r (or diameter) of each circle is compared with the radius (or diameter) of the through hole, and The circle closest to the radius (or diameter) of the through hole is used as the edge of the through hole, that is, the pattern of the workpiece mark. Next, the center position of the circle is obtained, and this is taken as the position of the workpiece mark. In Fig. 3, the circle is the circle C1 connecting the black edges 1, the circle C 2 connecting the black edges 2, and the three types of the circles C 3 connecting the black edges 3, but not only the circles of the three types, for example, the same figure Also shown is a case where a circle C4 other than the circles C1 to C3, a circle C5, and the like are drawn. Therefore, as shown in the figure, for example, a part of the black edges 1 to 3 cannot be detected, or the noise on the image is detected as a black edge, etc., and it is not necessarily obtained in each radial direction. The correct black edge 1 to 3 position of the situation. Therefore, the positions of the maximum enthalpy of each of the luminance changes are combined for each radial direction, and the circle of the complex number is approximated as described above, and the radius score estimation expression as described in the following U) is obtained for each circle as described above. The circle at the edge position. -17- 201207579 In addition, the scores of (b) to (d) below are obtained as needed, and the circle with the highest score is estimated as a circle indicating the edge position of the workpiece mark (through hole). (a) Radius score As described above, the radius score is obtained by using the design radius 接近 as the index of the radius r of the circle. That is, if the radius r is the same as the design ,, it is evaluated as one point, and if it is not equal to or greater than the predetermined value, it is evaluated as 0 point. Specifically, for example, the radius score is obtained by the following equation. Radius score = 1 - 丨 approximate circle radius - design radius | / a (b) Edge number score Use several edges and round approximation as an index to find the edge number score. That is, the more the number of edges used, the higher the score. Specifically, for example, the edge number score is obtained by the following equation. Edge number score = approximate number of edges / b (c) Edge bit range score The edge position range score is obtained by taking the distance from the circumference of the approximate circle as an index. For example, as shown in Fig. 5(a), when the full edge is on the circumference, the score becomes a perfect score, and as shown in Fig. 5(b), when the edge is not on the circumference, the score is lower. Specifically, for example, the edge position range score is obtained by the following equation. Edge position range score = 1 - (the difference between the maximum and minimum distances from the circumference) /c -18- 201207579 (d) Edge symmetry score The edge symmetry score is obtained by using the relative edge detected as an index. For example, when the radiation direction is 6 directions, as shown in the fifth (c) diagram, if the three pairs are all present, the score becomes higher, and as shown in the fifth (d), if it does not exist When paired, the score becomes lower. Specifically, for example, the edge symmetry score is obtained by the following equation. Edge symmetry score = (α + both detected as logarithm Μ + unilateral detected logarithm / r ) / d Also, in the above 2 (c), it shows a very small flaw from white to black, from black to white The example of the great ambiguity, but the method of changing the brightness change also has a very small flaw from white to black, and a very small flaw from black to white. The situation is that the position detection of the maximum flaw is taken as the position of the white edge, and the position of the extremely small flaw is detected as the position of the black edge. In the apparatus of Fig. 1, the workpiece mark center position detecting unit 1 1 C obtains the position of the workpiece mark as described above, and accordingly, the alignment of the cover Μ and the workpiece W is performed, and the alignment of the cover Μ and the workpiece W is performed. , proceed as follows. (1) Irradiation light is irradiated from the light irradiation device 1 or an alignment light source (not shown) to the mask curtain, and is transmitted to the control by the image pickup mask mark MA 'b via the alignment microscope 1 〇 CCD camera 1 〇b Department 1 1. The image processing unit 11a of the control unit 1 1 converts the mask mark image into a position coordinate and stores it in the memory unit 1 1 b. -19- 201207579 In addition, various methods are proposed for the method of detecting the mask mark. For example, please refer to Patent Document 1. (2) Next, the illumination light is irradiated from the illumination means 10c of the alignment microscope 10 to the workpiece W, edge detection is performed as described above, and the workpiece mark WAM on the workpiece W is detected, and the control unit 1 1 determines the position coordinates thereof. . (3) The control unit 11 is a position coordinate of the memorized mask mark MAM, and the position coordinate of the detected workpiece mark WAM is in a predetermined positional relationship, and the workpiece platform WS is moved (or the mask platform MS, or On both sides, the alignment of the mask and the workpiece W is performed. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a view showing the configuration of a projection exposure apparatus to which the present invention is applied. Figs. 2(a) to 2(d) are diagrams showing the detected image (schematic diagram), and the luminance distribution in the radial direction of the image and its differential 値. Fig. 3 is a view showing a complex circle of a closed curve which approximates the position of the maximum 値 or the minimum 値 of the differential 値 of the brightness of the connected image. Figure 4 is a diagram showing the image of the workpiece displayed on the monitor. Figures 5(a) to 5(d) are diagrams illustrating the scores of approximate circles. Fig. 6 is a view showing a schematic configuration of an alignment microscope 1 检测 for detecting a workpiece mark. Figures 7(a) to 7(c) are diagrams illustrating a pattern detecting method by edge detection. Figs. 8(a) to 8(b) are diagrams showing an example of the shape of the through hole -20 - 201207579 which is used as a workpiece mark. Figs. 9(a) to 9(f) are diagrams showing an example of an image to be photographed by a through hole covered with a photoresist film. Figs. 10(a) to 10(c) are diagrams for explaining a view of a through hole covered with a photoresist film. [Main component symbol description] 1 : Light irradiation device 2 : Projection lens 3 = Mask platform drive mechanism 4 · Workpiece platform drive mechanism 1 对准: Alignment microscope 1 0 a : Half mirror 10b : CCD camera l〇c : Illumination Means 1 1 : Control unit 1 1 a : Image processing unit 1 1 b : Memory unit 1 1 c : Work mark center position detecting unit 1 1 d : Registration control unit 1 1 e : Registration unit 1 2 : Monitor L1, L 2 : Lens Μ : Mask 21 - 201207579 MS : Mask platform MAM : Mask mark (mask alignment mark ) MP : Mask pattern W : Workpiece WS : Workpiece platform WAM : Work mark (workpiece alignment mark ) -twenty two