200844649 九、發明說明： 【發明所屬之技術領域】 本發明係關於包含用於製造液晶顯示裝置（Liquid Crystal Display :以下，稱爲LCD)等之灰階遮罩、微細圖 案的光罩之缺陷檢查方法以及缺陷檢查裝置等。此外，還 關於一種形成於遮罩之圖案的轉印方法。 【先前技術】 現在，在LCD的領域中，薄膜電晶體液晶顯示裝置 (Thin Film Transistor Liquid Crystal Display:以下稱爲 T F T- L C D )相較於C RT (陰極射線管），因爲有能輕易薄型化 且消耗電力低的優點，所以現在正迅速地商品化。TFT-LCD 具有以矩陣狀排列的各畫素上排列有TFT之構造的TFT基 板、及對應各畫素而排列有紅色、綠色以及藍色之畫素圖 案的彩色濾色片而在液晶層的介入下所重合的槪略構造。 在TFT-LCD中，製造步驟數很多，光是製造TFT基板就要 使用 5〜6片光罩。在這種狀況下，在「月刊 FPD Intelligence」、1 999年5月、P.31〜35(非專利文獻1)中， 提出了使用4片光罩來進行TFT基板之製造的方法。 此方法係藉由使用具有遮光部、透光部和半透光部的 光罩（以下稱爲灰階遮罩），來減低使用的遮罩片數。在此， 半透光部就是指在使用遮罩並將圖案轉印至被轉印體上 時，使透過之曝光光線的透過量減低既定量，且控制被轉 印體上之光阻膜的顯影後之殘膜量的部分，這種同時具備 半透光部、遮光部、透光部的光罩則被稱爲灰階遮罩。 200844649 作爲能在此使用的灰階遮罩’已知有半透光部以微細 圖案所形成之構造者。例如，如第4(a)圖所示，具有與源 極/汲極對應之遮光部1、透光部2及與通道部對應之半透 光部3，半透光部3係在使用灰階遮罩之LCD用曝光機的 曝光條件下，形成由解析極限以下之微細圖案所組成的遮 光圖案3a的區域。3b係半透光部3在曝光機之曝光條件 下的解析極限以下的微細透過部。遮光部1和遮光圖案3 a 通常是由一起以鉻和鉻化合物等的相同材料組成之相同厚 度的膜所形成。透過部2與微細透過部3 b係一起在透明基 板上未形成遮光膜等之透明基板的部分。使用灰階遮罩之 LCD用曝光機的解析極限在大部分的情況下，步進式之曝 光機則是大約3 μιη，鏡面投影式之曝光機則是大約4μιη左 右。因此，例如，第4(a)圖中，可將半透光部3之微細透 過部3b的空間寬度設爲未滿3μηι，將遮光圖案3a之線寬 設爲未滿 3 μιη而作爲曝光機之解析極限以下的圖案。或 者，可藉由調整曝光機的解析條件，來設定上述微細圖案 爲實質地非解析的條件。 上述微細圖案型的半透光部在設計時能考慮到，在半 透光部的設計，具體而言，會有把用於維持遮光部和透光 部之中間半色§周效果的微細圖案做成線與間隙（1 i n e a n d space)型或做成點狀（網點）型，或者做成其他圖案的選擇， 此外，在線與間隙型的情況下，線寬要做到什麼程度，要 怎麼處理光透過部分和遮光部分的比率，要將全體的透過 率設計到什麼程度等。 200844649 在以上述大型LCD用曝光機進行曝光的時候，因爲通 過半透光部3的曝光光線就全體而言，曝光量變得比透光 部還要少，所以在通過透光部2的曝光光線而使被轉印體 之正型光阻在顯影後消失之條件的時候，介由此半透光部 3而曝光之被轉印體上的正型光阻方面，在顯影後只有膜 厚變薄而殘留於基板上。此時，光阻會因爲曝光量的差異， 與通常的遮光部1對應之部分及與半透光部3對應之部分 方面，對顯影液的溶解度會產生差異，所以顯影後的光阻 形狀係如第4(b)圖所示，與通常的遮光部1對應之部分4 1 爲例如大約1 . 3 μιη，與半透光部.3對應之部分4 3爲例如大 約0.3 μηι，與透光部2對應之部分就是無光阻的部分42， 來使膜厚選擇性地變化。然後，在無光阻的部分4 2進行被 加工基板的第1蝕刻，由灰化等來除去與半透光部3對應 之薄的部分4 3之光阻，在此部分進行第2蝕刻，藉以利用 1片遮罩來進行以往2片份量遮罩的步驟，可減少遮罩片 數。 作爲如同上述灰階遮罩之由微細圖案所組成的半透光 部之檢查方法，已知道有特開2 0 0 3 - 3 0 7 5 0 1號公報（專利文 獻1 )所揭露的檢查方法。亦即，專利文獻1中記載了一種 灰階遮罩之缺陷檢查方法’該灰階遮罩具有：遮光部；透 光部；以及半透光部，其以減低透過已調整透過量之區域 的此區域之光之透過量而選擇性地改變光阻的膜厚爲目 標’且前述半透光部於灰階遮罩上具有形成在使用灰階遮 罩的曝光機之解析極限以下之遮光圖案的區域所組成，該 200844649 缺陷檢查方法包含：掃描前述半透光部而獲得透過率信號 的步驟；對前述透過率信號實施使用前述灰階遮罩時近似 於半透光部之透過率特性的修正處理的步驟；以及前述修 正處理後之透過率信號當超過預設之半透光部的透過率缺 陷之臨界値時，判斷爲在半透光部中發生缺陷的步驟。 【發明內容】 在上述專利文獻1揭露的先前技術中，有鑑於檢查灰 階遮罩之半透光部時，在對僅由以往之遮光部和透明部所 組成之遮罩所進行的比較檢查中不充足的觀點，而提出了 預先設定透過率之臨界値，而檢測出缺陷的方法。亦即， 在掃描遮罩內之圖案而獲得的透過率信號方面，使用預定 之透過率缺陷的臨界値而檢測出缺陷。另外，在檢查由微 細圖案所組成之半透光部時，掃描該部分而獲得透過率信 號時，因爲在此透過率信號中有時會發生週期性變動，所 以會施行用於使此信號平坦化的模糊處理。而且，藉由進 行缺陷檢查，能避免檢測出疑似缺陷的情形。作爲此時的 模糊處理，能採用被以往畫像處理所使用的模糊功能。 不過，本發明人等發現了會產生即使藉由上述檢查方 法也無法排除的疑似缺陷（並非缺陷，而是被判定爲疑似缺 陷者），或者無法以足夠的精度來檢測出缺陷等在遮罩之檢 查精度方面不佳的情形。 例如，第5圖係表示包含遮光部與由透光部之線與間 隙的微細圖案所構成之半透光部3的遮罩圖案。中央之遮 光部（遮光圖案3 a)和此兩側透光部（微細透過部3 b )的線寬 200844649 胃爲Ιμιη。這就是藉由上述先前技術的方法來掃描圖案並 獲得透過率信號。換言之，使具備光源以及照明光學系統、 物鏡的光學系統以及攝影手段（能適當地選擇採用線 CCD(Charge Coupled Device)或者 TDI(Time Delay Integration)的CCD等）平行於遮罩主平面地對遮罩進行掃 描（或者’對上述光學系統以及攝影手段，使遮罩移動並掃 描）’藉以獲得透過率信號。因爲攝影手段（在此爲線CCD) 的畫素尺寸在此爲1 μπι，所以掃描時，攝影手段之各畫素 (在此表示畫素Α〜F)和圖案的相對關係係如第5圖的1〜 1 〇 (圖中附加圓形之數字而表示）所示，會有各式各樣的時 候。 若爲「1」之位置關係時，作爲.CCD的攝影畫像，理 論上考慮到如同第6圖之6 1的梳型畫像。這是因爲第5圖 之「1」的位置關係的情況下，畫素C、D、E會分別與透 光部、遮光部、透光部的位置一致，作爲此時的CCD之輸 出，C之透光部是1〇〇%(在此以ι〇0%來表示透光部的透過 率），D爲零，E爲100%，所以理論上成爲第6圖中以虛線 表示的透過率分佈曲線6 1。但是，由於光的干涉，透過率 分佈會變寬至某個程度，實際上會取得第6圖之實線所示 的透過率分佈曲線62之狀態。 接著，在CCD的畫素和微細圖案處於如第5圖之「5」 的位置關係時，取得第6圖中以一點鏈線所示之透過率分 佈曲線63的狀態。這時候，能獲得畫素B爲40%左右，C 爲60%左右，D爲40%左右，E爲60%左右的透過率信號。 -10- 200844649 然後，按照專利文獻1的方法，分別對第6圖之透過 率分佈曲線62和63施行畫像處理技術所使用的模糊處 理，接著比較施行模糊處理後的透過率和預設之臨界値。 不過，在此’從完全相同之圖案所獲得的透過率分佈之曲 線，便有62(「1」之位置關係的情況）和63(「5」之位置關 係的情況）不同。因此，對該等進行相同模糊處理而獲得的 透過率分佈也如第7圖中之曲線7 1 (對曲線62進行模糊處 理）和7 2 (對曲線63進行模糊處理）般之不同。因此，按照 半透光部之臨界値的設定方法，上述「1」之位置關係時與 「5」之位置關係時，會發生缺陷判定之結論不同的情況。 此缺陷判定雖取決於半透光部的透過率容許範圍，但在例 如5 1 0%的時候，即使依據例如「5」之位置關係時所得 的模糊處理後之透過率分佈曲線 72而判定該遮罩爲良 品，在^ 1」之位置關係時，模糊處理後之透過率分佈曲線 7 1也會發生超過透過率容許範圍的部分，且該遮罩被判斷 爲不良的情況。在此，雖可暫時將臨界値設定爲即使在「1」 之位置關係時也不會被判定爲不良的程度（換言之，使透過 率容許範圍的寬度ΔΤ擴大），但這麼做的話，會發生檢查 精度下降的問題。 此外，爲了進一步發展上述專利文獻1的手法，則考 慮了反映遮罩使用時之曝光機的解析條件，以正確地把握 無法解析之微細圖案部的透過率’所以取得更微細的透過 率信號，使其趨近於曝光條件的模糊狀態下進行平坦化即 可。例如，如第8(a)圖、第8(b)圖所圖示，相對於圖案的 200844649 尺寸，使用相對較小之畫素的攝影手段即可。第8 (a)圖係 表示1/2畫素尺寸的情況，第8(b)圖係表示1/5畫素尺寸 的情況。不過，具備較小畫素的攝影手段會超過用來作爲 一般液晶裝置製造用之缺陷檢查裝置的規格，不但裝置的 成本增加，.且檢查時間會變得非常長。當然，因爲取得的 透過率信號之資料量也會膨脹，所以資料處理變得困難。 最初，採用微細圖案的半透光部方面，該部分之曝光光線 透過率能滿足所需規格是首要任務，雖然逐一檢查微細圖 # 案的形狀的意義不大，但如果要同上述般準備微細的畫素 尺寸之裝置方面，則不論從成本的角度看來、或從檢查時 間的角度看來都不具效率。 如同上述，在藉由微細圖案而形成半透光部的灰階遮 罩中，該微細圖案的尺寸與對其進行檢查的檢查機之攝影 畫素尺寸彼此接近時，檢查時的彼等相互位置關係會使透 過率分佈的輸出發生變動，有無缺陷的判斷也不一致。這 是因爲拍攝半透光部時由於攝影元件之畫素尺寸而導致漏 • 失的畫像資訊會影響缺陷判定。近幾年，半透光部（例如， 薄膜電晶體的通道部）之尺寸也有變得更加小的趨勢，另 外，遮罩使用者所期望之半透光部的轉印性之需求也會多 樣化，因此’半透光部的微細圖案尺寸也必須要小，會發 生無法忽視與畫素尺寸之抗衡的狀況。 特別是’用於製造液晶表示裝置的灰階遮罩爲大型， 一邊爲30cm以上，根據情況有時會成爲100em以上的尺 寸，所以不但檢查效率極爲重要，還因爲單價高，所以缺 -12- 200844649 陷判定方面要求要有高可靠度。 本發明係有鑑於上述以往的問題點而完成者，第1目 的在於提供灰階遮罩的缺陷檢查方法以及缺陷檢查裝置 等，能夠正確地判定半透光部具有形成在使用灰階遮罩時 曝光之曝光條件下的解析極限以下之微細圖案之區域的灰 階遮罩在半透光部有無缺陷，並提升缺陷檢查的可靠度。 另外，本發明之第2目的爲提供具有採用這種缺陷檢查方 法之缺陷檢查步驟的灰階遮罩之製造方法。此外，第3目 的爲提供一種圖案的轉印方法，該圖案係形成於實施上述 缺陷檢查步驟而獲得的灰階遮罩。 爲了解決上述課題，本發明有以下的構成。 (構成1) 一種灰階遮罩之缺陷檢查方法，該灰階遮罩 具有：遮光部；透光部；以及半透光部，其減低透過已調 整透過量之區域的此區域光之透過量而選擇性地改變被轉 印體上之光阻的膜厚爲目標，且前述半透光部具有形成在 使用灰階遮罩時曝光之曝光條件下的解析極限以下之微細 遮光圖案的區域，該缺陷檢查方法之特徵爲具有：掃描前 述半透光部而獲得透過率信號的步驟；以及比較前述透過 率信號和預設之半透光部的透過率容許値，來判定前述半 透光部有無缺陷的判定步驟，在獲得前述透過率信號的步 驟中，以既定光源來照射前述半透光部並透過前述半透光 部之透過光束從正焦位置以既定量錯焦的像，會被攝影手 段所拍攝，且由該拍攝畫像獲得透過率信號。 (構成2)如構成1的灰階遮罩之缺陷檢查方法，其中， 200844649 在前述攝影時，根據前述半透光部之圖案形狀或者尺寸、 或者根據檢查機的能力、或者根據遮罩使用時的曝光條 件、或者根據遮罩曝光後之處理步驟條件，來決定前述錯 焦量。 (構成3)如構成1或構成2的灰階遮罩之缺陷檢查方 法’其中，以在拍攝前述半透光部內之正常前述微細遮光 圖案形成區域時，排列之攝影手段的畫素當中，相鄰之畫 素的光強度差爲5 %以下，來作爲決定前述錯焦量的條件。 (構成4)如構成1至構成3中任一構成的灰階遮罩之 缺陷檢查方法，其中，前述攝影手段之畫素尺寸和前述微 細遮光圖案形成區域之遮光圖案或者透光圖案的線寬比範 圍係1 / 2〜2。 ’ (構成5)如構成1至構成4中任一構成的灰階遮罩之 缺陷檢查方法，其中，前述半透光部的透過率容許値係被 設定在使用以下兩種透過率信號時之判定結果相異的範圍 中：藉由前述攝影手段，不錯焦而拍攝灰階遮罩之半透光 部的微細遮光圖案形成區域，並對該攝影畫像進行用以信 號平坦化的畫像處理而得的透過率信號；以及進行前述錯 焦而拍攝獲得的透過率信號。 (構成6) —種光罩的缺陷檢查方法，其係具有微細圖 案部之光罩的微細圖案部的缺陷檢查方法’該方法之特徵 爲具有：掃描前述微細圖案部而獲得透過率信號的步驟； 比較前述透過率信號和預設之微細圖案部的透過率容許 値，來判定前述半透光部有無缺陷的判定步驟’在獲得前 -14- 200844649 述透過率信號的步驟中，以既定光源來照射前述微細 部並透過前述微細圖案部之透過光束從正焦位置以既 錯焦的像，會被攝影手段所拍攝’且由該拍攝畫像獲 過率信號。 (構成7) —^種灰階遮罩之缺陷檢查裝置，該灰階 具有：遮光部；透光部；以及半透光部，其減低透過 整透過量之區域的此區域光之透過量而選擇性地改變 印體上之光阻的膜厚爲目標，該裝置之特徵爲具有： 系統，其藉由平行光源以及受光透鏡來掃描形成於前 罩內的圖案，並接收透過光束；攝影手段，其拍攝所 的透過光；以及判定手段，其使用從該攝影手段之攝 像中獲得的透過率信號，來和預設之半透光部的透過 許値做比較，來判定前述半透光部有無缺陷，而前述 系統以及/或者前述攝影手段係具有錯焦手段，使得使 透過光束的像以既定量從正焦位置錯焦而得的像會被 手段所拍攝。 (構成8)如構成7的灰階遮罩之缺陷檢查裝置，， 具有控制手段，其根據前述半透光部之圖案形狀或尺 或者檢查機之能力、或者遮罩使用時的曝光條件、或 遮罩曝光後之處理步驟條件相關的錯焦量之決定要件 的輸入’來將前述光學系統或者前述攝影手段驅動並 於滿足該錯焦量的位置。 (構成9)如構成7或構成8的灰階遮罩之缺陷檢 置’其中’前述判定手段中係預先記憶有前述半透光 圖案 定量 得透 遮罩 已調 被轉 光學 述遮 接受 影畫 率容 光學 前述 攝影 ί中， 寸、 者1與 資料 維持 查裝 :部之 200844649 透過率容許値範圍、前述透光部之透過率容許範圍、前述 遮光部之透過率容許範圍，並進行和前述透過率信號的比 較。 (構成1 〇) —種灰階遮罩的製造方法，其特徵爲具有缺 陷檢查步驟，其使用構成1至構成5中任一構成所記載之 缺陷檢查方法來進行檢查缺陷。 (構成1 1 ) 一種圖案轉印方法，其特徵爲：將曝光光線 照射於構成1 〇記載之製造方法所獲得的灰階遮罩，將形成 Φ 於前述灰階遮罩上的圖案轉印至被轉印體上。 藉由本發明之灰階遮罩的缺陷檢查方法，具有：掃描 半透光部而獲得透過率信號的步驟，而該半透光部具有形 成在使用灰階遮罩時曝光之曝光條件下的解析極限以下之 微細遮光圖案的區域；以及比較前述透過率信號和預設之 半透光部的透過率容許値，來判定前述半透光部有無缺陷 的判定步驟，在獲得前述透過率信號的步驟中，以既定光 源來照射前述半透光部並透過前述半透光部之透過光束從 ® 正焦位置以既定量錯焦的像，會被攝影手段所拍攝，且由 該拍攝畫像獲得透過率信號，並使用以此方式獲得之透過 率信號來進行前述判定步驟。藉此，即使是相同的被檢查 圖案，因.爲也能迴避由於攝影手段之畫素和圖案的相對位 置而使缺陷檢查判定變得各不相同的問題，能高精度地判 定半透光部具有形成在使用灰階遮罩時曝光之曝光條件下 的解析極限以下之微細圖案之區域的灰階遮罩在半透光部 有無缺陷，所以可以提升缺陷檢查的可靠度。另外，該缺 -16- 200844649 陷檢查方法並不侷限於所謂的灰階遮罩，也可以適當地應 用於包含微細圖案之光罩的微細圖案部的缺陷檢查。 另外，藉由本發明的缺陷檢查裝置，能適當地實施本 發明的缺陷檢查方法。 另外，藉由本發明的灰階遮罩的製造方法，由於具有 採用這種本發明之缺陷檢查方法的缺陷檢查步驟，所以能 獲得已實施高可靠度之缺陷檢查的灰階遮罩。 進一步，藉由將曝光光線照射於實施上述缺陷檢查步 • 驟而獲得之灰階遮罩，並將形成於該遮罩的圖案轉印至被 轉印體上的圖案轉印方法，因爲使用已實施高可靠度之缺 陷檢查的灰階遮罩，所以特別能夠防止半透光部的轉印不 良。 【實施方式】 以下，參照圖式來說明用於實施本發明之最佳形態。 本發明係一種灰階遮罩之缺陷檢查方法，該灰階遮罩具 有：遮光部；透光部；以及半透光部’其減低透過已調整 ® 透過量之區域的此區域光之透過量而選擇性地改變被轉印 體上之光阻的膜厚爲目標’且前述半透光部具有形成在使 用灰階遮罩時曝光之曝光條件下的解析極限以下之微細遮 光圖案的區域，該缺陷檢查方法具有：掃描前述半透光部 而獲得透過率信號的步驟；以及比較前述透過率信號和預 設之半透光部的透過率容許値’來判定前述半透光部有無 缺陷的判定步驟。然後’在獲得前述透過率信號的步驟中， 以既定光源來照射前述半透光部並透過前述半透光部之透 -17- 200844649 過光束從正焦位置以既定量錯焦的像，會被攝影手段所拍 攝，且由該拍攝晝像獲得透過率信號。使用這樣獲得的透 過率信號來進行前述判定步驟。 在本發明中，如同上述地掃描構成灰階遮罩之半透光 部的微細圖案並進行檢查的時候，具體而言，攝影時適用 的錯焦條件係必須從正焦位置來決定錯焦的既定量’亦 即，錯焦量。具體而言，根據例如前述半透光部的圖案形 狀或者尺寸、或者根據檢查裝置的能力、或者根據遮罩使 φ 用時之曝光條件、或因應遮罩曝光後之處理步驟條件來決 定此錯焦量。在此，檢查裝置的能力方面，則舉出例如檢 查裝置的光學系統或者攝影手段（例如C CD)的光軸方向之 動作精度。另外，遮罩使用時的曝光條件方面，則舉出曝 光機的解析度等。此外，在曝光後的遮罩之處理步驟條件 方面，會有遮罩使用者所期望之光阻膜厚量容許範圍、膜 厚變動量容許範圍或者光阻殘膜的形狀等。最好是採用這 些當中的至少一個來決定錯焦量。特別是半透光部的圖案 ® 形狀或者圖案尺寸是重要的決定要件。 另外’被錯焦的攝影畫像係在拍攝半透光部之正常該 微細遮光圖案形成區域的時候，該部分之透過率便成略平 坦所引起之使攝影的光強度變成大致平坦的情況是典型 的，決定上述錯焦量的時候’例如，在拍攝前述半透光部 內的時候，能選擇在排列的攝影手段之畫素中相鄰的畫素 彼此的光強度差爲既定量以下（例如5 %以下）的條件，但當 然不侷限於此。 -18- 200844649 另外，本發明在作爲檢查對象的半透光部之圖案尺寸 方面未被限定，也可應用於例如微細遮光圖案之線寬爲〇 · 5 〜2·5μπι左右者。同樣地，微細遮光圖案中的透光圖案之 線寬也能有效地應用於0.5〜2.5 μιη左右的情況。此外，本 發明在微細圖案形狀方面也無特別限制，可以有效地應用 在例如1〜2 μιη線寬左右的線與間隙之情形，或上述線寬 之矩形圖案所排列的形狀等。 另外，CCD等的攝影手段之畫素尺寸並未被特別限 定，在具有例如〇 . 2 μιη以上之畫素尺寸的時候，能有效地 採用本發明。 此外，在本發明中，作爲檢查對象的半透光部之微細 圖案之線寬和攝影手段之畫素尺寸的相對關係方面，並不 需要特別設限，但例如，微細圖案之線寬相對於攝影手段 之畫素尺寸爲過大的情況下，因爲在該攝影手段中能充分 拍攝檢查對象的圖案，所以能採用和一般形狀缺陷之檢查 相同的檢查（例如，die-to-die和die-to-data的比較檢查）。 另一方面，在與上述相反的情況下，即使使用本發明之錯 焦的光學模糊，也很難獲得充分的效果。因此，攝影手段 之畫素尺寸和微細遮光圖案形成區域的線寬比範圍爲 1 /2〜2的時候，能顯著地獲得本發明的效果。在此所謂的線 寬也包含形成於半透光部的遮光圖案之線寬、或透光圖案 的任一種情況。 決定上述錯焦條件（錯焦量），使用以在那種條件下獲 得之攝影畫像爲基礎的透過率信號，與預設之半透光部之 200844649 透過率容許値之比較的缺陷判定步驟中，例如，將透光部 之透過率設爲100%的時候，將半透光部之透過率的中央値 設爲50%，將50士10%的透過率分佈設爲容許値。此外，藉 由遮罩使用者的曝光機或者遮罩使用者之光阻處理等之條 件來適當地決定此透過率容許値。此外，半透光部之透過 率的中央値雖在上述舉例表示5 0 %，但也能夠藉由遮罩使 用者的處理條件等，而適當地在20〜60%的範圍中進行選 擇。 • 此外，前述半透光部的透過率容許値係被設定在使用 以下兩種透過率信號時之判定結果相異的範圍中：如同前 述習知技術，藉由前述攝影手段，不錯焦而拍攝灰階遮罩 之半透光部的微細遮光圖案形成區域，並對該攝影畫像進 行用以信號平坦化的畫像處理而得的透過率信號；以及如 同本發明進行前述錯焦而拍攝獲得的透過率信號。在上述 之情況下，藉由本發明，因爲能避免前述以前習知技術造 成缺陷判定結果變得各不相同的問題，能進行可靠度高的 ® 檢查缺陷，所以本發明的效果顯著。特別是藉由本發明的 檢查方法，能有效地檢測出半透光部之微細圖案的線寬（CD) 錯誤所引起的缺陷。 作爲適合實施本發明之缺陷檢查方法的缺陷檢查裝 置，根據本發明，具有：光學系統，其藉由平行光源以及 受光透鏡來掃描形成於遮罩內的圖案，並接收透過光束； 攝影手段，其拍攝所接受的透過光；以及判定手段，其使 用從該攝影手段之攝影畫像中獲得的透過率信號，來和預 -20-[Technical Field] The present invention relates to defect inspection of a photomask including a gray scale mask and a fine pattern for manufacturing a liquid crystal display device (hereinafter referred to as LCD) Method and defect inspection device, etc. Further, it relates to a transfer method of a pattern formed on a mask. [Prior Art] Now, in the field of LCD, Thin Film Transistor Liquid Crystal Display (hereinafter referred to as TF T-LCD) is easier to thin than C RT (Cathode Ray Tube). Since it has the advantage of low power consumption, it is rapidly becoming commercialized. The TFT-LCD has a TFT substrate in which TFTs are arranged on each pixel arranged in a matrix, and a color filter in which red, green, and blue pixel patterns are arranged corresponding to the respective pixels, and is in the liquid crystal layer. The strategic structure that coincides with the intervention. In the TFT-LCD, the number of manufacturing steps is large, and it is necessary to use 5 to 6 reticle for manufacturing the TFT substrate. In this case, a method of manufacturing a TFT substrate using four masks is proposed in "Flip Intelligence", May, 999, and P.31 to 35 (Non-Patent Document 1). This method reduces the number of masks used by using a photomask having a light shielding portion, a light transmitting portion, and a semi-light transmitting portion (hereinafter referred to as a gray scale mask). Here, the semi-transmissive portion means that when the mask is used and the pattern is transferred onto the transfer target, the amount of transmitted light transmitted through is reduced by a predetermined amount, and the photoresist film on the transfer target is controlled. A portion of the residual film amount after development, which has a semi-transmissive portion, a light-shielding portion, and a light-transmitting portion, is called a gray-scale mask. 200844649 As a grayscale mask that can be used here, a construct in which a semi-transmissive portion is formed in a fine pattern is known. For example, as shown in Fig. 4(a), the light-shielding portion 1 corresponding to the source/drain electrodes, the light-transmitting portion 2, and the semi-transmissive portion 3 corresponding to the channel portion are provided, and the semi-transmissive portion 3 is used in gray. Under the exposure conditions of the exposure machine for LCD of the step mask, a region of the light-shielding pattern 3a composed of the fine pattern below the analysis limit is formed. 3b is a fine transmissive portion in which the semi-transmissive portion 3 is equal to or lower than the analysis limit under the exposure conditions of the exposure machine. The light shielding portion 1 and the light shielding pattern 3a are usually formed of a film of the same thickness composed of the same material such as chromium and a chromium compound. The transmissive portion 2 and the fine transmissive portion 3b are not formed with a portion of the transparent substrate such as a light shielding film on the transparent substrate. The resolution limit of an LCD exposure machine using a gray scale mask is, in most cases, about 3 μm for a stepper exposure machine and about 4 μm for a mirror projection type exposure machine. Therefore, for example, in the fourth (a) diagram, the spatial width of the fine transmissive portion 3b of the semi-transmissive portion 3 can be set to less than 3 μm, and the line width of the light-shielding pattern 3a can be set to less than 3 μm as an exposure machine. The pattern below the resolution limit. Alternatively, the fine pattern can be set to be substantially non-analyzed by adjusting the analysis conditions of the exposure machine. The semi-transmissive portion of the fine pattern type can be designed in consideration of the design of the semi-transmissive portion, specifically, a fine pattern for maintaining the effect of the middle half color of the light-shielding portion and the light-transmitting portion. Make line and gap (1 ineand space) type or dot (mesh) type, or make other patterns. In addition, in the case of online and gap type, what is the line width to be done, how to deal with it The ratio of the light transmitting portion to the light blocking portion is to what extent the overall transmittance is designed. 200844649 When the exposure is performed by the above-mentioned large-sized LCD exposure machine, since the exposure light passing through the semi-transmissive portion 3 is less than the light-transmitting portion as a whole, the exposure light passing through the light-transmitting portion 2 is exposed. When the positive photoresist of the transfer target disappears after development, the positive photoresist on the transfer target exposed by the semi-transmissive portion 3 has only a film thickness after development. It is thin and remains on the substrate. At this time, since the photoresist has a difference in the solubility of the developer in the portion corresponding to the normal light-shielding portion 1 and the portion corresponding to the semi-light-transmitting portion 3 due to the difference in the amount of exposure, the photoresist pattern after development is different. As shown in Fig. 4(b), the portion 4 1 corresponding to the normal light shielding portion 1 is, for example, about 1.3 μm, and the portion 4 3 corresponding to the semi-light transmitting portion .3 is, for example, about 0.3 μm, and is transparent. The portion corresponding to the portion 2 is the portion 42 having no photoresist to selectively change the film thickness. Then, the first etching of the substrate to be processed is performed in the portion 4 2 where the photoresist is not formed, and the photoresist of the thin portion 43 corresponding to the semi-transmissive portion 3 is removed by ashing or the like, and the second etching is performed in this portion. By using one mask to perform the conventional two-piece masking step, the number of masks can be reduced. As a method of inspecting the semi-transmissive portion which is composed of a fine pattern as in the above-described gray-scale mask, the inspection method disclosed in Japanese Laid-Open Patent Publication No. H03-303 (Patent Document 1) is known. . That is, Patent Document 1 describes a method for inspecting a defect of a gray scale mask, which has a light shielding portion, a light transmitting portion, and a semi-light transmitting portion for reducing an area through which the transmitted amount is adjusted. The film of the region selectively changes the film thickness of the photoresist as the target' and the semi-transmissive portion has a light-shielding pattern formed on the gray-scale mask below the resolution limit of the exposure machine using the gray-scale mask. The 200844649 defect inspection method includes the steps of: scanning the semi-transmissive portion to obtain a transmittance signal; and applying the gray-scale mask to the transmittance signal to approximate a transmittance characteristic of the semi-transmissive portion. The step of correcting the processing; and the step of determining that the transmittance signal in the semi-transmissive portion is defective when the transmittance signal after the correction processing exceeds the threshold of the transmittance defect of the predetermined semi-transmissive portion. SUMMARY OF THE INVENTION In the prior art disclosed in the above Patent Document 1, in view of the inspection of the semi-transmissive portion of the gray scale mask, a comparative inspection is performed on the mask composed only of the conventional shading portion and the transparent portion. In the case of insufficient views, a method of detecting a defect by setting a threshold of transmittance is proposed. That is, in terms of the transmittance signal obtained by scanning the pattern in the mask, the defect is detected using the critical threshold of the predetermined transmittance defect. Further, when the semi-transmissive portion composed of the fine pattern is inspected, when the portion is scanned to obtain a transmittance signal, since the periodicity fluctuation sometimes occurs in the transmittance signal, the signal is flattened. Fuzzy processing. Moreover, by performing defect inspection, it is possible to avoid detection of a suspected defect. As the blurring processing at this time, the blurring function used by the conventional image processing can be employed. However, the present inventors have found that a suspected defect (not a defect but a judged defect) which cannot be excluded by the above-described inspection method is generated, or a defect or the like cannot be detected with sufficient accuracy in the mask. The situation in which the inspection accuracy is poor. For example, Fig. 5 is a view showing a mask pattern including a light-shielding portion and a semi-transmissive portion 3 composed of a fine pattern of lines and spaces of the light-transmitting portion. The line width of the central light-shielding portion (light-shielding pattern 3 a) and the light-transmitting portions (fine-transmissive portion 3 b ) on the both sides is 200844649. The stomach is Ιμιη. This is by scanning the pattern and obtaining the transmittance signal by the method of the prior art described above. In other words, an optical system including a light source, an illumination optical system, an objective lens, and an imaging means (a CCD or the like which can appropriately select a line CCD (Charge Coupled Device) or a TDI (Time Delay Integration)) is placed in parallel with the main plane of the mask. The cover is scanned (or 'to the above optical system and photographic means to move the mask and scan) to obtain a transmittance signal. Since the pixel size of the photographing means (here, the line CCD) is here 1 μm, the relative relationship between each pixel of the photographing means (here, the pixel Α~F) and the pattern is as shown in Fig. 5 during scanning. The 1 to 1 〇 (indicated by the number of the circle attached to the figure) shows a variety of times. In the case of the positional relationship of "1", a comb-shaped portrait as in Fig. 6 is theoretically considered as a photographing image of .CCD. This is because, in the case of the positional relationship of "1" in Fig. 5, the pixels C, D, and E are respectively aligned with the positions of the light transmitting portion, the light blocking portion, and the light transmitting portion, and the output of the CCD at this time is C. The light transmitting portion is 1% (here, the transmittance of the light transmitting portion is represented by ι〇0%), D is zero, and E is 100%, so theoretically, the transmittance is indicated by a broken line in FIG. Distribution curve 6 1. However, due to the interference of light, the transmittance distribution is broadened to some extent, and the state of the transmittance distribution curve 62 shown by the solid line in Fig. 6 is actually obtained. Then, when the pixel and the fine pattern of the CCD are in the positional relationship of "5" in Fig. 5, the state of the transmittance distribution curve 63 indicated by the one-dot chain line in Fig. 6 is obtained. At this time, it is possible to obtain a pixel B of about 40%, C of about 60%, D of about 40%, and E of about 60%. -10- 200844649 Then, according to the method of Patent Document 1, the transparency processing used in the image processing technique is performed on the transmittance distribution curves 62 and 63 of Fig. 6, respectively, and then the transmittance after the blur processing and the threshold of the preset are compared. value. However, the curve of the transmittance distribution obtained from the completely identical pattern differs between 62 (the case where the positional relationship of "1" is) and 63 (when the positional relationship of "5" is concerned). Therefore, the transmittance distribution obtained by performing the same blurring processing is also different as the curve 7 1 (blurring the curve 62) and 7 2 (blurring the curve 63) in Fig. 7 . Therefore, according to the method of setting the threshold 半 of the semi-transmissive portion, when the positional relationship of the above "1" is in the positional relationship with "5", the conclusion of the defect determination may be different. The defect determination depends on the transmittance allowable range of the semi-transmissive portion. However, for example, at 510%, the transmittance analysis curve 72 after the blurring process is obtained based on the positional relationship of, for example, "5". When the mask is a good product, in the positional relationship of ^1", the transmittance distribution curve 71 after the blurring also occurs in a portion exceeding the allowable range of the transmittance, and the mask is judged to be defective. Here, the threshold 値 can be temporarily set to such an extent that it is not judged to be defective even in the positional relationship of “1” (in other words, the width ΔΤ of the transmittance allowable range is expanded), but this may occur. Check the problem of reduced accuracy. In addition, in order to further develop the technique of the above-mentioned Patent Document 1, it is considered that the analysis conditions of the exposure machine when the mask is used are used, and the transmittance of the fine pattern portion that cannot be analyzed is accurately grasped, so that a finer transmittance signal is obtained. It is sufficient to flatten it in a blurred state close to the exposure conditions. For example, as shown in Fig. 8(a) and Fig. 8(b), it is sufficient to use a relatively small pixel photographing means with respect to the size of the pattern of 200844649. Fig. 8(a) shows the case of 1/2 pixel size, and Fig. 8(b) shows the case of 1/5 pixel size. However, the photographic means with a smaller pixel will exceed the specifications used as a defect inspection device for general liquid crystal device manufacturing, and the cost of the device will increase, and the inspection time will become very long. Of course, data processing becomes difficult because the amount of data of the obtained transmittance signal also expands. Initially, in terms of the semi-transmissive portion of the fine pattern, the exposure light transmittance of the portion can satisfy the required specifications is a primary task, although it is not significant to examine the shape of the fine pattern one by one, but if it is prepared as described above, it is fine. The device aspect of the pixel size is not efficient from the point of view of cost or inspection time. As described above, in the gray scale mask in which the semi-transmissive portion is formed by the fine pattern, when the size of the fine pattern and the photographic pixel size of the inspection machine for inspecting the same are close to each other, the mutual position at the time of inspection The relationship changes the output of the transmittance distribution, and the judgment of whether or not there is a defect is also inconsistent. This is because the image information that is lost due to the pixel size of the photographic element when the semi-transparent portion is photographed affects the defect determination. In recent years, the size of the semi-transmissive portion (for example, the channel portion of the thin film transistor) has become smaller, and the demand for the transferability of the semi-transparent portion desired by the user is also varied. Therefore, the size of the fine pattern of the semi-transmissive portion must also be small, and a situation in which the size of the pixel cannot be ignored can be ignored. In particular, the gray scale mask used for manufacturing the liquid crystal display device is large, and the size is 30 cm or more on one side, and may be 100 cm or more depending on the case. Therefore, not only the inspection efficiency is extremely important, but also because the unit price is high, the shortage is -12- 200844649 Requirement requires high reliability. The present invention has been made in view of the above conventional problems, and a first object of the present invention is to provide a defect inspection method and a defect inspection device for a gray scale mask, and can accurately determine that the semi-transmissive portion is formed when a gray scale mask is used. The gray-scale mask in the region of the fine pattern below the resolution limit under the exposure condition under exposure has defects in the semi-transmissive portion, and improves the reliability of the defect inspection. Further, a second object of the present invention is to provide a method of manufacturing a gray scale mask having a defect inspection step using such a defect inspection method. Further, the third object is to provide a pattern transfer method which is formed in a gray scale mask obtained by performing the above defect inspection step. In order to solve the above problems, the present invention has the following configuration. (Configuration 1) A method for inspecting a defect of a gray scale mask, the gray scale mask having: a light shielding portion; a light transmitting portion; and a semi-light transmitting portion that reduces a light transmission amount in the region that transmits the adjusted transmittance And selectively changing the film thickness of the photoresist on the transfer target, and the semi-transmissive portion has a region of the fine light-shielding pattern formed below the resolution limit under exposure conditions when the gray-scale mask is used, The defect inspection method is characterized in that: the step of scanning the semi-transmissive portion to obtain a transmittance signal; and comparing the transmittance signal with a transmittance permitting of a predetermined semi-transmissive portion to determine the semi-transmissive portion In the step of determining whether or not there is a defect, in the step of obtaining the transmittance signal, the semi-transmissive portion is irradiated with a predetermined light source, and the transmitted light beam transmitted through the semi-transmissive portion is imaged at a positive focal position with a predetermined amount of misfocus The photographing means captures a transmittance signal from the photographed image. (Configuration 2) The defect inspection method of the gray scale mask of the configuration 1, wherein 200844649 is used according to the pattern shape or size of the semi-transmissive portion, or according to the capability of the inspection machine, or when used according to the mask. The amount of misfocus is determined by the exposure conditions or the processing step conditions after the exposure of the mask. (Configuration 3) The defect inspection method of the gray scale mask of the configuration 1 or the configuration 2, wherein the pixels of the photographing means arranged in the normal light shading pattern forming region in the semi-transmissive portion are photographed The difference in light intensity between adjacent pixels is 5% or less, which is a condition for determining the amount of misfocus. (Configuration 4) The defect inspection method of the gray scale mask of any one of the configurations 1 to 3, wherein the pixel size of the photographing means and the line width of the light-shielding pattern or the light-transmitting pattern of the fine light-shielding pattern forming region The ratio range is 1 / 2 to 2. (Configuration 5) The defect inspection method of the gray scale mask of any one of the configurations 1 to 4, wherein the transmittance of the semi-transmissive portion is allowed to be set when the following two kinds of transmittance signals are used. In the range in which the determination results are different: the fine light-shielding pattern forming region of the semi-transmissive portion of the gray-scale mask is photographed by the above-mentioned photographing means, and the photographed image is subjected to image processing for signal flattening. The transmittance signal; and the transmittance signal obtained by performing the aforementioned misfocusing. (Configuration 6) A defect inspection method for a mask, which is a defect inspection method for a fine pattern portion of a mask having a fine pattern portion. The method is characterized in that the method includes scanning the fine pattern portion to obtain a transmittance signal. Comparing the transmittance signal and the transmittance permitting of the predetermined fine pattern portion to determine whether the semi-transmissive portion has a defect or not, in the step of obtaining the transmittance signal before the pre--14-200844649, the predetermined light source The transmitted light beam that has passed through the fine portion and transmitted through the fine pattern portion is imaged by the photographing means from the positive focus position and is captured by the photographing means, and an overshoot signal is obtained from the photographed image. (Structure 7) - A defect inspection device for a gray scale mask, the gray scale having: a light shielding portion; a light transmitting portion; and a semi-light transmitting portion that reduces a light transmission amount in the region through the entire transmission amount Selectively changing the film thickness of the photoresist on the printing body, the device is characterized by: a system for scanning a pattern formed in the front cover by a parallel light source and a light receiving lens, and receiving the transmitted light beam; And the means for determining the transmitted light; and the determining means for determining the semi-transmissive portion by comparing the transmittance signal obtained from the imaging of the photographing means with the transmission of the predetermined semi-transmissive portion Whether or not there is a defect, and the above-described system and/or the above-described photographing means have a misfocusing means such that an image of the transmitted light beam is imaged by means of a plurality of images which are offset from the positive focus position. (Configuration 8) The defect inspection device of the gray scale mask of the configuration 7 has a control means according to the pattern shape of the semi-transmissive portion or the ability of the ruler or the inspection machine, or the exposure condition when the mask is used, or The input of the determination element of the amount of misfocus associated with the condition of the processing step after the exposure of the mask is used to drive the optical system or the aforementioned imaging means to a position that satisfies the amount of misfocus. (Configuration 9) The defect detection of the gray scale mask of the configuration 7 or the configuration 8 wherein the above-mentioned determination means pre-stores the semi-transmissive pattern, and the transparent mask is adjusted to be rotated. Rate optics, the above-mentioned photography, the inch, the 1 and the data maintenance check: Department 200844649, the transmission allowable range, the transmittance of the light-transmitting portion, and the allowable range of the transmittance of the light-shielding portion, and Comparison of transmittance signals. (Configuration 1) A method of manufacturing a gray scale mask, which is characterized in that it has a defect inspection step for performing inspection defects using the defect inspection method described in any of the configurations 1 to 5. (Configuration 1 1) A pattern transfer method characterized in that an exposure light is applied to a gray scale mask obtained by the manufacturing method described in the above, and a pattern formed on the gray scale mask is transferred to On the body to be transferred. The defect inspection method of the gray scale mask of the present invention has a step of: scanning a semi-transmissive portion to obtain a transmittance signal, and the semi-transmissive portion has an analysis under exposure conditions exposed when a gray scale mask is used a region of the fine light-shielding pattern below the limit; and a step of determining the presence or absence of the defect in the semi-transmissive portion by comparing the transmittance signal and the transmittance permitting of the predetermined semi-transmissive portion, and obtaining the transmittance signal In the case where the semi-transmissive portion is irradiated with a predetermined light source and the transmitted light beam transmitted through the semi-transmissive portion is imaged by the photographing means from the positive focus position, the image is captured by the photographing means, and the transmittance is obtained from the photographed image. The signal is used, and the above-described decision step is performed using the transmittance signal obtained in this way. In this way, even if the same pattern to be inspected is used, it is possible to avoid the problem that the defect inspection determination is different due to the relative position of the pixels and the pattern of the photographing means, and it is possible to accurately determine the semi-transmissive portion. The gray-scale mask having the region of the fine pattern formed below the resolution limit under the exposure conditions when the gray-scale mask is exposed using the gray-scale mask has defects in the semi-transmissive portion, so that the reliability of the defect inspection can be improved. Further, the missing inspection method is not limited to the so-called gray scale mask, and may be suitably applied to the defect inspection of the fine pattern portion of the photomask including the fine pattern. Further, the defect inspection method of the present invention can be suitably carried out by the defect inspection device of the present invention. Further, with the manufacturing method of the gray scale mask of the present invention, since the defect inspection step using the defect inspection method of the present invention is provided, it is possible to obtain a gray scale mask which has been subjected to defect inspection with high reliability. Further, the pattern transfer method of transferring the pattern formed on the mask to the transfer target by irradiating the exposure light to the gray scale mask obtained by performing the above defect inspection step is used Since the gray-scale mask for performing the defect inspection with high reliability is implemented, it is particularly possible to prevent the transfer failure of the semi-transmissive portion. [Embodiment] Hereinafter, the best mode for carrying out the invention will be described with reference to the drawings. The present invention relates to a method for inspecting a defect of a gray scale mask, the gray scale mask having: a light shielding portion; a light transmitting portion; and a semi-light transmitting portion that reduces the light transmission amount in the region that passes through the adjusted amount of transmission And selectively changing the film thickness of the photoresist on the transfer target as the target', and the semi-transmissive portion has a region of the fine light-shielding pattern formed below the resolution limit under the exposure condition when the gray-scale mask is exposed, The defect inspection method includes a step of scanning the semi-transmissive portion to obtain a transmittance signal, and comparing the transmittance signal with a transmittance of the predetermined semi-transmissive portion to determine whether the semi-transmissive portion is defective or not. Decision step. Then, in the step of obtaining the transmittance signal, the semi-transmissive portion is irradiated with a predetermined light source and transmitted through the semi-transmissive portion through the -17-200844649 over-beam from the positive focal position to a quantitatively misfocused image. It is photographed by the photographing means, and a transmittance signal is obtained from the photographed image. The above-described determination step is performed using the transmittance signal thus obtained. In the present invention, when the fine pattern constituting the semi-transmissive portion of the gray scale mask is scanned and inspected as described above, specifically, the misfocus condition applicable at the time of photographing is determined from the positive focal position to determine the misfocusing. Both quantitative 'that is, the amount of misfocus. Specifically, the error is determined according to, for example, the pattern shape or size of the semi-transmissive portion, or according to the capability of the inspection device, or the exposure condition when φ is used according to the mask, or the processing step condition after exposure of the mask. The amount of coke. Here, in terms of the capabilities of the inspection apparatus, for example, the optical axis of the inspection apparatus or the accuracy of the optical axis direction of the imaging means (e.g., C CD) can be cited. Further, in terms of exposure conditions at the time of use of the mask, the resolution of the exposure machine and the like are given. Further, in terms of the processing steps of the mask after the exposure, the allowable range of the thickness of the photoresist film desired by the user, the allowable range of the film thickness variation, or the shape of the photoresist residual film may be masked. It is preferable to use at least one of these to determine the amount of misfocus. In particular, the pattern of the semi-transparent portion ® shape or pattern size is an important decision. In addition, when the photographic image that is misfocused is photographed in the normal light-shielding pattern forming region of the semi-transmissive portion, the transmittance of the portion is slightly flat, and the light intensity of the photographing becomes substantially flat. When the amount of the above-mentioned misfocus is determined, for example, when the inside of the semi-transmissive portion is imaged, the difference in light intensity between adjacent pixels in the pixels of the arranged imaging means can be selected to be less than or equal to a predetermined amount (for example, 5) % below), but of course not limited to this. In addition, the present invention is not limited to the pattern size of the semi-transmissive portion to be inspected, and may be applied to, for example, a line width of the fine light-shielding pattern of about 5 5 to 2·5 μπι. Similarly, the line width of the light-transmitting pattern in the fine light-shielding pattern can be effectively applied to the case of about 0.5 to 2.5 μm. Further, the present invention is not particularly limited in terms of the shape of the fine pattern, and can be effectively applied to, for example, a line and a gap of about 1 to 2 μm line width, or a shape in which the rectangular pattern of the line width is arranged. Further, the pixel size of the photographing means such as a CCD is not particularly limited, and the present invention can be effectively employed when it has a pixel size of, for example, 2 μm or more. Further, in the present invention, the relative relationship between the line width of the fine pattern of the semi-transmissive portion to be inspected and the pixel size of the photographing means does not need to be particularly limited, but for example, the line width of the fine pattern is relative to In the case where the size of the photographing means is too large, since the pattern of the inspection target can be sufficiently captured by the photographing means, the same inspection as the inspection of the general shape defect can be employed (for example, die-to-die and die-to). -data comparison check). On the other hand, in the case of the opposite of the above, it is difficult to obtain a sufficient effect even if the optical blur of the focus of the present invention is used. Therefore, when the pixel size ratio of the photographing means and the line width ratio of the fine light-shielding pattern forming region are in the range of 1 /2 to 2, the effects of the present invention can be remarkably obtained. Here, the line width also includes any of the line width of the light-shielding pattern formed in the semi-transmissive portion or the light-transmitting pattern. Determining the above-mentioned misfocus condition (wrong focus amount), using the transmittance signal based on the photographic image obtained under that condition, and the defect determination step in comparison with the transmittance of the predetermined semi-transmissive portion of 200844649 For example, when the transmittance of the light transmitting portion is 100%, the center 値 of the transmittance of the semi-transmissive portion is set to 50%, and the transmittance distribution of 50% and 10% is set as the allowable enthalpy. Further, the transmittance permitting is appropriately determined by masking the user's exposure machine or masking the user's photoresist processing. Further, although the center 値 of the transmittance of the semi-transmissive portion is 50% as exemplified above, it can be appropriately selected in the range of 20 to 60% by masking the processing conditions of the user or the like. • In addition, the transmittance of the semi-transmissive portion allows the enthalpy to be set in a range in which the determination results when the following two kinds of transmittance signals are used are different: as in the aforementioned conventional technique, the photographing means is used to shoot well a fine light-shielding pattern forming region of the semi-transmissive portion of the grayscale mask, and a transmittance signal obtained by performing image processing for signal flattening on the photographed image; and transmission obtained by photographing the above-mentioned misfocus as in the present invention Rate signal. In the above case, according to the present invention, since the problem that the defect determination results are different from each other in the prior art can be avoided, the highly reliable ® inspection defect can be performed, and the effect of the present invention is remarkable. In particular, by the inspection method of the present invention, it is possible to effectively detect a defect caused by a line width (CD) error of the fine pattern of the semi-transmissive portion. According to the present invention, a defect inspection apparatus suitable for carrying out the defect inspection method of the present invention has an optical system that scans a pattern formed in a mask by a parallel light source and a light receiving lens, and receives a transmitted light beam; Shooting the received transmitted light; and determining means using the transmittance signal obtained from the photographic image of the photographing means, and pre--20-
200844649 設之半透光部的透過率容許値做比較，來判定前述 部有無缺陷，而前述光學系統以及/或者前述攝影手 有錯焦手段，使得使前述透過光束的像從正焦位置 量錯焦而得的像會被攝影手段所拍攝。 使用第1圖來具體說明上述缺陷檢查裝置。第 表示本發明之缺陷檢查裝置的構成主要部分之立體 如第1圖所示，缺陷檢查裝置2 0具備：光源 於射出平行光的照明光學系統2 2 ;接收透過光的物 以及檢測透過率信號的攝影手段24。在此，光源2〗 光學系統2 2、物鏡2 3以及攝影手段（例如c C D ) 2 4 各個光軸一致的狀態下，相對於作爲在照明光學系糸 物鏡2 3之間配置之被檢查對象的灰階遮罩1 〇之主 平行地相對移動。或者相反地，亦可藉由使光學系 固定位置，而遮罩1〇在其主平面方向上移動，結果 與光學系統相對移動，進行遮罩主平面上之任意點& 如同這般，缺陷檢查裝置20係具有攝影手段 藉由光源2 1、照明光學系統22以及物鏡23而相對 在遮罩1 0內形成的圖案，並檢測出透過率信號。更 說明，具有使遮罩1 〇與上述光學系統相對移動並掃 10主表面之整個區域的手段（通常是遮罩台座移動 沿著掃描方向以物鏡2 3而接收來自遮罩1 〇的透過 一步藉由具備CCD等之受光元件的攝影手段24來 透過率信號。此外，此缺陷檢查裝置2 0具有能在從 置而移動既定量的錯焦狀態下進行攝影的錯焦手段 半透光 段係具 以既定 1圖係 圖。 21 ;用 鏡23 ; .、照明 係在使 充22和 平面而 統處於 使遮罩 勺攝影。 24，其 地掃描 具體地 描遮罩 手段）， 光，進 檢測出 正焦位 ，因此 -21 - 200844649 在第1圖例示之實施形態中具備控制手段，使上述物鏡23 以及/或者攝影手段24相對於遮罩10主表面而在光軸方向 (第1圖中的箭頭2 5)上相對地移動。此控制手段係根據前 述半透光部的圖案形狀或者尺寸、或者根據檢查裝置的能 力、或者根據遮罩使用時之曝光條件、或輸入與遮罩曝光 後之處理步驟條件相關的錯焦量決定要件，根據這些決定 要件來決定錯焦量時，則以滿足該錯焦量的方式，來將上 述光學系統（在此爲物鏡23)以及/或者攝影手段24從正焦 # 驅動至以既定量偏移的錯焦位置，且檢查期間維持那種狀 態。 此外，第1圖雖未圖示，但上述缺陷檢查裝置20係具 備處理手段（CPU等），其能夠進行裝置全體之控制機構、 上述光學系統的移動、攝影畫像的演算處理（信號化）、或 與判定手段之臨界値（透過率容許値）的比較、缺陷判定 等。另外，前述判定手段中預先記憶有前述半透光部之透 過率容許値範圍、前述透光部的透過率容許範圍、前述遮 ® 光部的透過率容許範圍，並進行和前述透過率信號的比較。 接著，說明採用本發明的缺陷檢查方法來進行檢查缺 陷的具體事例。 以下，進行具備具有如同第4(a)圖之微細圖案區域的 半透光部（在此透過率是5 0%)之灰階遮罩的缺陷檢查的範 例。在此使用的缺陷檢查裝置係上述第1圖所示的構成。 使用線CCD作爲攝影手段，其畫素尺寸是Ιμιη。作爲檢查 對象的灰階遮罩圖案係與第5圖相同，半透光部的微細圖 -22-200844649 The transmittance of the semi-transmissive portion is allowed to be compared to determine whether there is a defect in the portion, and the optical system and/or the photographer has a misfocusing means such that the image of the transmitted beam is misaligned from the positive focus position. The captured image will be taken by photography. The above defect inspection apparatus will be specifically described using FIG. First, the main part of the configuration of the defect inspection device of the present invention is as shown in Fig. 1. The defect inspection device 20 includes a light source for illuminating optical system 2 2 that emits parallel light, and a device that receives transmitted light and detects a transmittance signal. Photography means 24. Here, the light source 2, the optical system 2, the objective lens 2, and the photographing means (for example, cCD) 2 4 are inspected with respect to each other as the object to be inspected between the objective lens 23 and the illumination optical system. The grayscale mask 1 〇 is moved in parallel relative to each other. Or conversely, by fixing the position of the optical system, the mask 1 移动 moves in the direction of its principal plane, and as a result, moves relative to the optical system to perform any point on the main plane of the mask & The inspection device 20 has a pattern formed by the light source 21, the illumination optical system 22, and the objective lens 23 with respect to the mask 10, and detects a transmittance signal. More specifically, there is a means for moving the mask 1 相对 relative to the optical system and sweeping the entire area of the main surface 10 (usually the mask pedestal moves along the scanning direction to receive the passage from the mask 1 by the objective lens 2). The transmittance signal is transmitted by the imaging means 24 including the light receiving element such as a CCD. The defect inspection apparatus 20 has a misfocusing means semi-transmissive section capable of performing photography while moving from a predetermined amount of misfocusing state. With a set of 1 maps. 21; with a mirror 23; ., the lighting system is used to make the charge 22 and the plane in the mask scoop photography. 24, its ground scan specifically mask means), light, into the detection In the embodiment shown in Fig. 1, a control means is provided to cause the objective lens 23 and/or the imaging means 24 to be in the optical axis direction with respect to the main surface of the mask 10 (Fig. 1). The arrow 2 5) moves relatively. The control means is determined according to the pattern shape or size of the semi-transmissive portion, or according to the capability of the inspection device, or according to the exposure condition when the mask is used, or the amount of misfocus associated with the processing step condition after the exposure of the mask. When the amount of the wrong focus is determined according to the determination requirements, the optical system (here, the objective lens 23) and/or the imaging means 24 are driven from the positive focus # to the predetermined amount in such a manner as to satisfy the amount of the wrong focus. The offset of the offset is maintained and that state is maintained during the inspection. In addition, although the first aspect is not shown, the defect inspection apparatus 20 includes a processing means (such as a CPU) that can perform a control mechanism of the entire apparatus, a movement of the optical system, and an arithmetic processing (signalization) of the photographed image. Or comparison with the critical value of the determination means (transmission tolerance), defect determination, and the like. Further, in the determination means, the transmittance permitting range of the semi-transmissive portion, the transmittance permitting range of the light transmitting portion, and the transmittance permitting range of the light shielding portion are stored in advance, and the transmittance signal is performed. Comparison. Next, a specific example in which the defect inspection method of the present invention is used to perform the inspection defect will be described. Hereinafter, an example of the defect inspection of the gray scale mask having the semi-transmissive portion (the transmittance of which is 50%) as in the fine pattern region of Fig. 4(a) is performed. The defect inspection device used here is the configuration shown in Fig. 1 described above. A line CCD is used as a photographing means, and its pixel size is Ιμιη. The gray scale mask pattern to be inspected is the same as in Fig. 5, and a fine pattern of the semi-transmissive portion -22-
200844649 案係線與間隙之遮光部被夾在透光部間的形狀， 光部和其兩側透光部的線寬皆爲1 μιη。 將作爲被檢查對象的上述灰階遮罩安裝於第 陷檢查裝置，於第2圖模式地表示本發明之光學 的攝影畫像。在第2圖中，掃描時之CCD的各晝 和攝影畫像的相對位置關係也會如1〜1 〇 (圖中用 字記述）所示地呈現出各式各樣。此外，第2圖的| 雖不見得正確地表現實際的攝影畫像，但實際上 焦的光學模糊，所以遮罩之與遮光部1對應的畫 及遮罩之與半透光部3對應的畫像區域13的交 確。 第3圖係表示從光學錯焦之狀態的攝影畫 過率分佈。在此，不論CCD的畫素和攝影圖案 關係爲「1」時之透過率分佈曲線3 1、或相對 「5」時之透過率分佈曲線3 2的任一情況下’ 透過率都會變得大致平坦，且兩者的透過率分 差異，所以將半透光部之透過率容許値 △ Τ設 5 0± 10 %的時候，在第3圖的範例中，任一種情 同的缺陷檢查判定，能夠迴避以往即使是相同 案，缺陷檢查判定也會變得各不相同的不佳狀 藉由本發明，能防止缺陷檢查判定會由於攝影 和圖案的相對位置而受到影響。 如同以上所說明，藉由本發明的灰階遮罩 方法，即使是相同的被檢查圖案，也能迴避由 中央之遮 1圖的缺 錯焦狀態 素（Α〜F) 帶圓形數 I式圖中， ，由於錯 像區域1 1 界並不明 取得之透 .相對位置 置關係爲 i透光部的 ;幾乎沒有 定爲例如 i中都是相 丨被檢查圖 I。亦即， ^段之畫素 b缺陷檢查 >攝影手段 -23- 200844649 之畫素和圖案的相對位置造成缺陷檢查判定變得各不相同 的問題，能高精度地判定半透光部具有形成在使用灰階遮 罩時曝光之曝光條件下的解析極限以下之微細圖案之區域 的灰階遮罩在半透光部有無缺陷，所以可以提升缺陷檢查 的可靠度。 在大基板尺寸的LCD裝置製造用灰階遮罩中，採用以 前的缺陷檢查方法時，常會有檢測爲透過率信號之疑似缺 陷的情況，檢查的可靠度不太高，所以本發明的缺陷檢查 # _方法在實用化LCD製造用灰階遮罩方面有極大的效果。這 種情況在LCD製造用遮罩以外的其他顯示裝置也相同。此 外，在LCD製造用遮罩方面，包含對製造LCD時必要的所 有遮罩，例如包含用於形成TFT(薄膜半導體）、低溫多晶 矽的TFT、濾色片等的遮罩。在其他顯示裝置製造用遮覃 方面，則包含製造有機EL(電致發光）顯示器、電漿顯示器 等所必需的所有遮罩。 因爲本發明係使用從錯焦之攝影畫像取得的透過率輸 ® 出信號，而不進行圖案辨識的檢查方法，所以能迴避微細 圖案檢查時發生特有圖案形狀所引起之疑似缺陷的問題 (無法降低臨界値的問題），因此，可降低臨界値，能滿足 包含灰階遮罩和微細圖案之光罩要求精度（規格）。 另外，藉由本發明的灰階遮罩的製造方法，由於具有 採用這種本發明之缺陷檢查方法的缺陷檢查步驟，所以能 獲得已實施可靠度高之檢查缺陷的灰階遮罩。此外，將曝 光光線照射於實施上述缺陷檢查步驟而獲得的灰階遮罩’ -24- 200844649 透過將形成於該遮罩的圖案轉印至被轉印體的圖案轉印方 法，因爲使用已實施可靠度高的檢查缺陷的灰階遮罩，所 以特別能防止半透光部的轉印不良。 另外’藉由本發明的缺陷檢查方法以及缺陷檢查裝 置，除了灰階遮罩以外，也可以應用在如同光罩之線寬爲 3 μιη未滿之線與間隙的微細且高精度的圖案之缺陷檢查， 而獲得掃描圖案部並施行錯焦之光學模糊的均勻透過率信 號，並根據此信號，進行形狀及尺寸等之檢查缺陷，因而 可檢測出微細及高精度的缺陷。作爲包含這種微細圖案的 光罩，可列舉出LCD製造用光罩和有機EL顯示器、電漿 顯示器等的顯示裝置製造用光罩，亦即具有用於形成TFT 通道部和接觸孔部等之微細圖案的光罩等。 此外，本申請案係以在2007年3月2日申請之日本專 利申請案第2007-05 3 273號之優先權爲基礎而主張其利 益，其所有內容在此作爲參照文獻。 【圖式簡單說明】 第1圖係表示本發明之缺陷檢查裝置的構成主要部分 之立體圖。 第2圖係模式地表示在本發明之缺陷檢查方法中光學 錯焦之狀態的攝影畫像之平面圖。 第3圖係表示從光學錯焦之狀態的攝影畫像取得之透 過率分佈的圖。 第4 (a)圖係表示微細圖案型之灰階遮罩的一例之平面 圖。 -25- 200844649 第4(b)圖係使用灰階遮罩而獲得之光阻圖案的截面 圖。 第5圖係表示藉由攝影手段來掃描微細圖案時之攝影 手段的畫素和微細圖案之相對位置關係的平面圖。 第6圖係表示藉由攝影手段來掃描微細圖案時的攝影 手段之輸出透過率信號的圖。 第7圖係表示對攝影手段之輸出透過率信號施行畫像 處理之既定模糊處理而得的透過率分佈圖。 第8圖係表示藉由攝影手段來掃描微細圖案時之攝影 手段的畫素和微細圖案之相對位置關係的平面圖，相對於 第5圖所示的畫素尺寸，第8(a)圖係表示1/2畫素尺寸’ 第8(b)圖係表示〗/5畫素尺寸的情況。 【主要元件符號說明】 1 遮 光 部 2 透 過 部 3 半 透 光 部 3a 遮 光 圖 案 3b 微 細 透 過 部 10 灰 階 遮 罩 11 畫 像 區 域 13 畫 像 區 域 20 缺 陷 檢 查 裝 2 1 光 源 2 2 照 明 光 學 系 統 -26- 200844649 23 物 鏡 24 攝 影 手 段 25 目U 頭 3 1 透 過 率 分 佈 曲 線 32 透 過 率 分 佈 曲 線 41 與 通 常 的 遮 光 部 1 對 應 之部分 42 與 透 光 部 2 對 應 之 部 分 43 與 半 透 光 部 3 對 應 之 部 分 6 1 透 過 率 分 佈 曲 線 6 2 透 過 率 分 佈 曲 線 6 3 透 過 率 分 佈 曲 線 7 1 曲 線 72 曲 線 -27 -200844649 The light-shielding portion of the case line and the gap is sandwiched between the light-transmitting portions, and the line widths of the light portion and the light-transmitting portions on both sides are 1 μm. The gray scale mask as the object to be inspected is attached to the trap inspection device, and the optical photographing image of the present invention is schematically shown in Fig. 2 . In Fig. 2, the relative positional relationship between each CCD and the photographic image at the time of scanning is also expressed as shown in 1 to 1 〇 (indicated by words in the figure). Further, although the image of Fig. 2 does not necessarily represent the actual photographic image correctly, the image of the mask corresponding to the light-shielding portion 1 and the image corresponding to the semi-transmissive portion 3 of the mask are masked. The intersection of area 13 is true. Fig. 3 is a graph showing the distribution of the photographic magnification from the state of optical misfocus. Here, in any case where the transmittance of the CCD and the photographic pattern relationship is "1", the transmittance distribution curve 3 1 or the transmittance distribution curve 3 2 at the time of "5", the transmittance will become substantially It is flat, and the transmittance of the two is different. Therefore, when the transmittance of the semi-transmissive portion is allowed to be 値 Δ Τ set to 50 ± 10%, in the example of Fig. 3, any of the same defect inspection determinations, It is possible to avoid the disadvantages in which the defect inspection determinations are different even in the case of the conventional case. According to the present invention, it is possible to prevent the defect inspection determination from being affected by the relative positions of the photographing and the pattern. As described above, with the gray scale mask method of the present invention, even the same pattern to be inspected can avoid the false focus state element (Α~F) of the central mask 1 with a circular number I pattern. In the case, the position of the erroneous image region 1 1 is not clearly obtained. The relative position relationship is the i-transmission portion; it is hardly determined that, for example, i is the same as the image I is inspected. That is, the segmental pixel b-defect inspection> photography means -23- 200844649 The relative position of the pixels and the pattern causes the defect inspection determination to be different, and the semi-transmissive portion can be formed with high precision. The gray-scale mask in the region of the fine pattern below the resolution limit under the exposure conditions under exposure using the gray scale mask has defects in the semi-transmissive portion, so that the reliability of the defect inspection can be improved. In the gray scale mask for manufacturing a large substrate size LCD device, when the previous defect inspection method is used, there is often a case where a suspected defect of the transmittance signal is detected, and the reliability of the inspection is not so high, so the defect inspection of the present invention is performed. The # _ method has great effects in the practical use of grayscale masks for LCD manufacturing. This is the same in other display devices than the mask for LCD manufacturing. Further, in the mask for LCD manufacturing, all the masks necessary for manufacturing an LCD, for example, a mask for forming a TFT (Thin Film Semiconductor), a TFT of a low temperature polysilicon, a color filter, or the like are included. In the case of other concealers for display device manufacturing, all the masks necessary for manufacturing an organic EL (electroluminescence) display, a plasma display, and the like are included. Since the present invention uses the transmittance transmission signal obtained from the photographic image of the wrong focus without performing the pattern recognition inspection method, it is possible to avoid the problem of the occurrence of the suspected defect caused by the characteristic pattern shape at the time of the fine pattern inspection (cannot be lowered) The problem of critical enthalpy), therefore, can reduce the critical enthalpy, and can meet the required accuracy (specification) of the reticle including the grayscale mask and the fine pattern. Further, according to the method for manufacturing a gray scale mask of the present invention, since the defect inspection step using the defect inspection method of the present invention is provided, it is possible to obtain a gray scale mask which has been subjected to inspection defects having high reliability. Further, the exposure light is irradiated to the gray scale mask obtained by performing the above-described defect inspection step '-24- 200844649 by transferring the pattern formed on the mask to the transfer method of the transfer target, since the use has been carried out The gray-scale mask for inspecting defects is highly reliable, so that the transfer failure of the semi-transmissive portion can be particularly prevented. In addition, by the defect inspection method and the defect inspection device of the present invention, in addition to the gray scale mask, it is also possible to apply a defect inspection of a fine and high-precision pattern such as a line and a gap of a line width of 3 μm. Further, a uniform transmittance signal obtained by scanning the pattern portion and performing optical blurring of the misfocus is obtained, and inspection defects such as shape and size are performed based on the signal, so that fine and high-precision defects can be detected. The mask for manufacturing such a fine pattern includes a photomask for manufacturing an LCD, a photomask for manufacturing a display device such as an organic EL display or a plasma display, and the like, that is, a TFT for forming a TFT channel portion, a contact hole portion, or the like. A mask with a fine pattern, etc. Further, the present application claims the benefit based on the priority of Japanese Patent Application No. 2007-05 3 273, filed on March 2, 2007, the entire disclosure of which is hereby incorporated by reference. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a perspective view showing a main part of a configuration of a defect inspection device of the present invention. Fig. 2 is a plan view schematically showing a photographing portrait in a state of optical misfocus in the defect inspection method of the present invention. Fig. 3 is a view showing a transmittance distribution obtained from a photographic image in a state of optical misfocus. Fig. 4(a) is a plan view showing an example of a gray pattern mask of a fine pattern type. -25- 200844649 Figure 4(b) is a cross-sectional view of a photoresist pattern obtained using a grayscale mask. Fig. 5 is a plan view showing the relative positional relationship between the pixels and the fine patterns of the photographing means when the fine pattern is scanned by the photographing means. Fig. 6 is a view showing an output transmittance signal of a photographing means when a fine pattern is scanned by a photographing means. Fig. 7 is a view showing a transmittance distribution map obtained by performing a predetermined blurring process on the output transmittance signal of the photographing means. Fig. 8 is a plan view showing the relative positional relationship between the pixels and the fine patterns of the photographing means when the fine pattern is scanned by the photographing means, and the figure 8(a) is expressed with respect to the pixel size shown in Fig. 5. 1/2 pixel size 'Fig. 8(b) shows the case of /5 pixel size. [Description of main component symbols] 1 Light-shielding portion 2 Transmissive portion 3 Semi-transmissive portion 3a Light-shielding pattern 3b Fine-transmissive portion 10 Gray-scale mask 11 Image region 13 Image region 20 Defect inspection device 2 1 Light source 2 2 Illumination optical system-26- 200844649 23 Objective lens 24 Photographing means 25 Head U Head 3 1 Transmittance distribution curve 32 Transmittance distribution curve 41 A portion corresponding to the normal light-shielding portion 1 and a portion corresponding to the light-transmitting portion 2 and a portion corresponding to the semi-transmissive portion 3 6 1 Transmittance distribution curve 6 2 Transmittance distribution curve 6 3 Transmittance distribution curve 7 1 Curve 72 Curve -27 -