201241872 六、發明說明： 【發明所屬之技術領域】 【0001】 本發明係關於一種處理裝置之異常診斷方法及其異常診斷系 統。 【先前技術】 [0002] 一般而言，半導體晶圓等之製程中，為於半導體晶圓或LCD 基板等之被處理基板的表面形成光阻圖案，使用光微影技術。此 一光微影技術，依序施行：於被處理基板之表面塗布光阻液的光 阻塗布步驟、於卿成之光_將圖練光㈣光處理步驟、以 及將顯影液供給予曝光處理後之基板的顯影處理步驟等一系列之 處理’可於被處理基扳之表面形成既定的光阻圖案。此等之步驟， 係於搭載有處理被處理基板之各種處理單元、搬運祕理基板之 搬運單元等的光阻塗布·顯影處理裝置施行。 【0003】 連續⑽ 生常異原常因的情^ 【0004】 此一情況，作為推定異常原因的裝置 之一 已知有：依據作201241872 VI. Description of the Invention: [Technical Field of the Invention] [0001] The present invention relates to an abnormality diagnosis method of a processing apparatus and an abnormality diagnosis system therefor. [Prior Art] [0002] Generally, in a process of a semiconductor wafer or the like, a photoresist pattern is formed on a surface of a substrate to be processed such as a semiconductor wafer or an LCD substrate, and a photolithography technique is used. The photolithography technique is sequentially performed: a photoresist coating step of applying a photoresist to the surface of the substrate to be processed, a light of the film of the crystal, a light treatment step of the light, and a solution for applying the developer to the exposure process. A series of processes, such as a development processing step of the substrate, can form a predetermined photoresist pattern on the surface of the substrate to be processed. These steps are carried out by a photoresist coating and developing treatment apparatus equipped with various processing units for processing a substrate to be processed, a transport unit for transporting a secret substrate, and the like. [0003] Continuous (10) Frequently different causes of common cause ^0004] In this case, as one of the devices for estimating the cause of the abnormality, it is known that:
Sir狀因子的監視對象之監視結果:推定裝 【=】 製造裝置（例如’參考專利文獻1)。 對象载的ΐ導體製造裝置’將依據組合複數監視 應異㈣別㈣進行比較，對於異常資料，檢索關於對應之 201241872 異常判別資料的模式中是否有 it日1，讀取與此—赋對應之料相 之Monitoring result of the monitoring target of the Sir-like factor: estimated installation [=] Manufacturing apparatus (for example, 'Reference Patent Document 1). The target-made tantalum conductor manufacturing device 'will be compared according to the combination complex number monitoring difference (4) (4), and for the abnormal data, whether there is an it day 1 in the pattern of the corresponding 201241872 abnormality discriminating data, and the reading corresponds to this Material phase
[白1知技術文獻] M[白一知技术文章] M
[專利文獻] 【0006] [專利文獻1]日本特開2006_216920號公報 【發明内容】 [發明所欲解決的問題] 【0007】 ，則，專利文獻i記載之半導體製造 =至 >、’監視縣之檢職超翻 =’在某-時點中 操作者，對操作者選出之複數監視對象，為異常並通知 組合，與異常判別資料加以比較。此 ^測時的檢測値 定閾値的其他各時點中，未進行昱常月]檢測値未超過固 【0008】 ~ ° _上^:置生一異二監= 【0009】 μΪϊ上述問題，本發明之目的在於提供—種，可ϋ祕切 方法及其異常診斷系統。权〜精度的處理裝置之異常診斷 [解決問題之技術手段] 【0010】 徵為=3;題判置ί異常診斷方法，其特 到;處理製程之感測器所輸出的訊號收集 满間時間序列龍，是否與複數判紐件- 201241872 該術:齡叹診斷步驟， 【0011】 該判ίί料位之時間序列資料； 為複數監視區間’在各列個時間序列分割 -致，並製作由此等判定: =疋否與複數判定條件 -個處縣置巾⑽理铸。處理置=的=步驟、及 入至各處理袭置後，至將苴搬 例如自將晶圓W搬 【0014】 f 【οοίΓ 偵測異f之時間序列資料。 此外，5亥複數判定條件，宜包含以判 ;==:;rr 幅度是否 至設定錄 201241872 資料之 ^ _ 千句^ 才示準差所計算出之變動閾値内、最大 ;Γ:Γ閾値内、最小値是否位於既定閾値内。 式資料斷步驟宜具備排序步驟’計算該判定資料與該模 ^排i —致度高的順序進行異常原因選擇對象之順 【0017】 —致兄拉第1再診斷步驟，在該診斷步驟之對照不 :=更?模式資料並進行對照，推定異常原因； ί變動傾向的關連性以;常;口：巧資= 内為止之狀條J更為下位的順位 除了指與全部的處，對照不—致的情況， 常原無法確認包含由於與複數異 式資Μ更;驟之對照不-致的情況，宜將該模 件的對照，重“ssr為止之判定條 與該S資備:序步驟，計算該判定資料 象之順位排序。 蚁度向的順序進行異常原因選擇對 【0020】 ::步驟步Ϊ更=:製:?驟’在該 料；以及第2再診斷步驟，將合而成的再判定資 將騎匈疋資料、與依擄各個 201241872 因發生之時間序列資料其 判定資料之模式#料進行對/㈣、!?而預先設定並對應於該再 5變更，除了指1個判定條件變更為m原因。此處，判定條 =了1 複數=】因—致而無法確認為1個ίίϊ因兄^:包含由於與 第2再診斷步驟之對照不再^資，製作步驟，於該 再度製作該再判定資料；以 '月一步變更該判定條件以 該再判定資料、與依據各個原驟，將再度製作的 變動傾向而預先設定並對於;間序列資料其特有 料進行·，_定^^再度齡_射仪:賴之模式資 【0022】 與該^資=驟，計算該判定資料 象之順位排序。 又冋勺頁序進行異常原因選擇對 【0023】 的常，之異 於處理被處理體之處理!^ =作σΡ心自藉由設置 號收集到之與時間；之出的訊 1 牛一致，並製作由此等判找果組合而成^仪^魏判定條 照，以推定異常原因。鳩於糾疋糾之核式資料進行對 【0024】 該判= 序列㈣’縣各碱理單位之邮序列資料. 複數皿視《，在各個監視區間中，狀是否與複數判定^^ 201241872 致 你，^衣作由此等判定結果組合而成的判定㈣μ- ♦ * 立係包含，複數個處理I置之各處處’處理單 個處理裝置中的處理步驟。衷置知仃的處理步驟、及一 【0025】 否該判定資 '雜作由此㈣歧果 ίίιΐί 5 出的異常之_序师料，：碰、—I叙以該異常侧部镇測 此等判定結果以二4判=咖定條件-致，並製作由 【0027】 【0028】 比外，該 =日，序賴k上限値與下限値的巾二… ::=Γ1ΐ閣値内、時間積分値是否位於既定= 竺，値為止之到達時間是否位於既定閾値内、至處^以止 數欺條件，宜包含以下判定條件中之至少i個： 呒定閾 无定閾 各個時間序列的平均值盘並# m i二二r r 値是::定二 一勤，外’該診斷部，宜進—步計算該判定資料與該模式資料之 t’i—μ高__行料賴麵之順位排序。 法、此4丨月況更具備第1再診斷部’於診斷部之對·昭π—絲 情況，藉由變更該模式資料並進行對 疮= 的 仏件’絲如與各個異常原因發生之_相資料其特有^ 201241872 傾向的關連性高低之排序， ·該變 起至既定順位内為止 ^讀，健自最高位 式資料，宜與自最高位起照；該變更後之模 之判ρ》的判定結果對照。…為下位的順位内為止 變更=^料，宜進一步 ”此外’该第1再診斷部，宜 + 貧，之-致度，以— 判定資料與該模式 排序。 負序進仃異书原因選擇對象之順位 【0033] 不-二& : 定f料製作部，於診斷部之對昭 一致，作由6 與變更後之判定條^ 斷部，將該再判定^果=2的再判定資料；以及第2再診 料其特有變動傾向預；： =常原因發生之時間序列資 進行對照，轉因认並對應於該再狀資料之模式^ 【0034】 不-d兄宜處理：於該第2再診斷部之對日召 作的該再判定2再診斷部再度製 特有變動傾向而二異㊉原因發生之時間序列資料1 式資=】對,¾異 式資=;^ f^5」宜進—步計算該再縱資料舆該模 位排序。又 致度同的順序進行異常原因選擇對象之順 [本發明之效果] 201241872 【0036】 依本發明處理裝置之里火_ 判定與時間1㈣的時法及其異常简^統，藉由 致，並製作由顺縣 ^料’是否與複數狀條件一 設定並對應於判定資料其财變動傾向而預先 而可沿著_相=貞料進行龍’以推定異常原因， 度。此外，由於將自^里裝置之狀態，可提高異常診斷之精 與預先設定之模列資料製作之判定資料请 行異常診斷。 、進對知，推定異常原因，故可輕易地進 【實施方式】 【0038】 係就將本發明明之實施形態加以詳細說明。此處， 光阻以;=:=於被處理體為晶㈣之 光阻塗布·顯影處理裝置，如圖1及圖2戶斤干，士 φ山 =，?成：用於將收納晶圓w之晶盒=搬=由= ίϋ自此一載具區塊1其晶陳盒cb内取出之H 二先阻塗布·顯影處理的處理區塊2、以及於此，之日0= 介面區塊3連設的曝光區塊4。 <理£塊2介由 【_】[Patent Document 1] [Patent Document 1] JP-A-2006-216920 SUMMARY OF INVENTION [Problems to be Solved by the Invention] [0007] Patent Document i describes semiconductor manufacturing = to > The prefecture's overcrowding = 'in the time-point, the operator selects the multi-monitoring object selected by the operator, and gives an exception and notifies the combination, and compares it with the abnormality discriminating data. In the other time points of the detection threshold 値, the 未 月 ] ] ] ] ] 0008 [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ The object of the invention is to provide a method, a secret cutting method and an abnormality diagnosis system thereof. Abnormal diagnosis of the processing device of the weight-accuracy [Technical means for solving the problem] [0010] The levy is =3; the problem is judged ί abnormal diagnosis method, which is special; the signal output by the sensor for processing the process is collected. Sequence dragon, whether or not with the plural number of pieces - 201241872 The technique: the age sigh diagnosis step, [0011] the time series data of the judgment ίί material level; for the complex monitoring interval 'in the time series of each column segmentation - and made by These judgments: = 疋 No and plural judgment conditions - a county county towel (10). After the processing = the step of the processing, and after the processing is performed, the processing is performed, for example, by moving the wafer W. [0014] f [οοίΓ Detecting the time series data of the different f. In addition, the 5 haidu plural judgment condition should include the judgment; ==:; rr amplitude to the set record 201241872 data ^ _ thousand sentences ^ to show the change threshold calculated within the threshold, the maximum; Γ: Γ threshold 値, Whether the minimum 値 is within the established threshold. The data breaking step should have a sorting step 'calculating the judgment data and the order of the module is high, and the order of the abnormal reason is selected. [0017] - the brother pulls the first re-diagnosis step, in the diagnosis step Control not: = more? Mode data and comparison, presumed the cause of the abnormality; ί correlation of change propensity; often; mouth: Qiaozi = the figure of the inside J is lower than the position of the whole In the case of non-conformity, Changyuan cannot confirm that it is included in the case of the comparison with the plural-type foreign-investigation; if the comparison is not correct, the comparison of the module should be repeated, and the determination clause of the ssr and the S-supplied: In the sequence step, the ranking of the judgment data is calculated. The order of the ant direction is selected according to the order of the abnormality [0020]::step step ==: system: step 'in the material; and the second re-diagnosis step, The re-judgment of the combination will be based on the data of the Hungarians, and the pattern of the data of each of the 201241872 time series data is determined to be / (4), !? and is preset and corresponds to the change 5, except Refers to the reason why one judgment condition is changed to m. Here , the judgment bar = 1 complex number =] because of the - can not be confirmed as 1 ίίϊ because the brother ^: contains the comparison with the second re-diagnosis step no longer, the production steps, the re-judgment data is re-created; The one-month change of the determination condition is based on the re-judgment data, and the change tendency to be re-created according to each original step is set in advance, and the specific information of the inter-sequence data is performed, and the ___ : Lai's model capital [0022] and the ^ capital = step, calculate the ranking of the judgment data image. The order of the exception is selected for the [0023], which is different from the processing of the processed object! ^ = σ Ρ 自 自 自 自 自 自 自 自 自 自 自 自 自 自 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 收集 Ρ 收集 收集Correction and correction of the nuclear data to the [0024] The judgment = sequence (four) 'the county's various basic units of the mail sequence data. Multiple dishes view, in each monitoring interval, whether the shape and the plural number ^^ 201241872 To you, ^The judgment of the combination of the judgment results of the clothing (4) -- ♦ * The system contains, the processing of the processing in a single processing device in a plurality of processing I. The processing steps of the processing unit, and a [0025] whether the judgment is miscellaneous (4) Fruit ίίιΐί 5 Out of the abnormal _ sequence teacher,: touch, -I Syria with the abnormal side of the town test such judgment results by two 4 = coffee conditions - and made by [0027] [0028] In addition, the = day, the order of the upper limit of k and the lower limit of the towel 2... ::=Γ1ΐ阁値, time integral 値 is located at the established = 竺, 到达, the arrival time is within the established threshold, everywhere ^ The number of fraud conditions should include at least i of the following determination conditions: 呒定 threshold has no threshold, and the average value of each time series is #米二二rr 値 is:: two, one diligent, the other 'the diagnosis department, It is advisable to calculate the order of the judgment data and the t'i-μ high__ line of the pattern data. In the case of the 4th month, the first re-diagnosis department has the same information in the diagnosis department, and the situation of the sputum is changed by the change of the model data. _ phase data its unique ^ 201241872 tendency of the relevance of the ranking, · The change to within the established order to ^ read, from the highest position data, should be from the highest position; the change of the model ρ The judgment result is compared. ...changes to the lower position of the lower position = ^ material, it is advisable to further "external" the first re-diagnosis department, should be + poor, to the degree, to - determine the data and the pattern sorted. Negative order into the book reason selection The order of the object [0033] No-two & : The f-material creation unit, which is consistent with the diagnosis in the diagnosis unit, is determined by 6 and the changed determination section, and the re-judgment = 2 is re-determined. The data; and the second re-survey material has a special tendency to change;: = the time series of the frequent causes are compared, and the transfer reason corresponds to the pattern of the re-form data ^ [0034] Not-d brother should handle: In the second re-diagnosis unit, the re-diagnosis of the re-diagnosis unit of the second re-diagnosis unit re-establishes the characteristic change tendency and the time-series data of the occurrence of the dimorphism is 1 type of capital =] right, 3⁄4 alien type capital =; ^ f^ 5" should be advanced - step to calculate the re-segment data 舆 the module order. In the same order, the object of the abnormality is selected. [Effect of the present invention] 201241872 [0036] According to the present invention, the time method of the fire_determination and time 1 (4) of the processing device and its abnormality are obtained by And it is determined whether or not the condition of the Shun County material is set in accordance with the condition of the plural number and corresponds to the tendency of the financial change of the judgment data, and the cause of the abnormality can be estimated along the _ phase = the material. In addition, due to the state of the device, it is possible to improve the accuracy of the abnormal diagnosis and the determination data of the pre-set template data. The present invention is described in detail with reference to the embodiments of the present invention. Here, the photoresist is a photoresist coating/developing treatment device with the object being treated as a crystal (4), as shown in Fig. 1 and Fig. 2, the φ山=, 成: for accommodating the wafer w crystal box = moving = by = ί ϋ from this carrier block 1 taken out of its crystal box cb H 2 first resistance coating development processing block 2, and here, the day 0 = interface area Block 3 is connected to the exposure block 4. <Leading block 2 through [_]
收納1 ^有：晶圓E盒裝卸站6，具備可载置複數個密閉 收納有杈數枚’例如25枚之晶圓w的晶圓匣盒CB 開閉=7’設置於此―晶随盒裝㈣6間所配置的壁面；以及傳 遞手段A1，用於介由此一開閉部7自晶圓ϋ盒CB取出曰圓w 【0041】 曰曰uw。 一於上述處理區塊2 ’將加熱•冷卻系統之單元多段化的棚架單 兀m、U2、U3，以及於包含後述塗布•顯影單元之各處理^元 201241872 間進行晶圓W之傳遞的主搬運機構A2、A3，自前方側起依序交 互配列地設置。亦即，自載具區塊i側觀察棚架單元口丨、 U3及主搬運機構A2、A3為配列成前後一列，且各自之連接部位 形成未圖示之晶圓搬運用的開口部，晶圓W可於處理區塊2内自 由地自一端側之棚架單元U1移動往另一端側之棚架單元U3為 止。此外主搬運機構A2、A3 ’係置於區隔壁8所包圍之空間内”、'， 該區隔壁8，由自載具區塊丨觀察時配置為前後方向之棚架單元 Ul、U2、U3側其一面部、後述之例如右側的液處理單元说'、仍 巧，一面部、以及成為左側之一面的背面部所構成。另，圖中， 符號12、13為溫濕度調節單元，具備各單元使用之處理液 調節裝置與溫濕度調節用的導管等。 皿又 【0042】 液處理單元U4、U5 ’例如如圖】所示，在成為塗布液（光阻 液）或顯影液等藥液供給用之空間的收納部M之上，將塗布單元 cot、具備顯影裝置之顯影單元DEV及反射防止膜形 BARC等堆疊複數段，例如5段地構成。此外，上述之 二1；2、，於施行液處理料w、U5所進行的處理其前 ^及後處理之各種單元堆疊複數段，例如1G段地構成，具有加 …（烘烤）晶圓W之加熱單元、冷卻晶圓冒之冷卻單 【0043】 寸 區地2中的棚架單元U3之後側，介由介面區塊3與曝光 .連接。此一介面區塊3，由前後地設置於處理區塊2與曝 區塊4間，各自以例如筐體包圍之搬運室15及搬運室16所 t=搬運室15之中央部設置搬運機構A4，可於x軸、Y轴及 二i Z軸）方向自由移動，且可將臂部圍繞鉛直軸地旋轉；此 機，構成為可接觸傳遞單元⑽）17、高精度溫度 (未圖示）、周邊曝光單元19、緩衝晶随盒（未圖示) 區塊2所具備之棚架單元U3，與此等之各單元進行晶圓 【0044】 12 201241872 動同上述地構成之光阻塗布·顯影處理裝置中的曰圓之, 盒邙載置於載具區塊！之载置部5省開^^圓 ί 之蓋體而藉由傳遞手段A1取出晶圓W。之後，ί ίί : 之前處之;·個棚架，施行作為塗布處理 朵阳ΐ ί 處冷卻處理後，於塗布單元COT冷允 熱皁元加熱（烘烤處理），牛 Υ3之—個棚架的加 遞單元往介面‘父4二=二,辦⑽之傳 ^處理。接受過周邊曝光處理之晶圓t/藉由周邊曝 至向精度溫度調節單元，於此一高精产1 f構A4被搬運 W表面之溫度高精度地溫度調節至對内，將晶圓 ，度。搬運機構M將此一曝内之溫度的 ^日日圓W，稭由搬運機構A 至16。所搬運 ί元r;:反，路徑搬運 以成為棚架之!二要的光阻。將晶圓= 2 ’接著《冷卻單元冷卻後，使JL0S，丨载(後供烤處 原本的晶圓匣盒CB。 八1载/、£塊1之载置台上 【0045】 20，明之異常診斷系統2〇加以說明。里常岭斷备 常原;影處理裝置中偵測發生之異常以； ，數製程處理裝置的複合處理3 ’為施行既定處理 诊斷系統20，藉送收菸邱％义 Θ 、圖3及圖4所示，異當 理部、處理單 》邛36 ’舁为別設置於複數製程處理F= Π又動的日爾列資料41之訊號輸出至異常診 201241872 20，設置例如壓力咸 ^ 器、位置感測哭、二二:度感測器、流量感測器、液面感測 可於3°。另， 2卜異常铺二22異$ = ^ 2〇之理論構成，係由資料收集部 部25、再判定資料製二診斷部曰24、第！再診斷 時間序列資料41 ；資料保持部29 記憶有： 判1;、及判Ϊ定資料43、模式資料44、再 資料收集部21，自藉由分別# 、 的訊號，收集與時間—同變動^岸二王 =之感測器30所輸出 2㈣隐時間序列資料4 於資料保持部 變動的壓力、溫戶、、、ά曰，^斤幻貝枓41，有例如：與時間一同 【_】 ㈣、液位、位置、扭力、速度等。 料41 ’與後述』f變收二部21收集到之時間序列資 時間序列資料41較:而偵測作為異常之 單位：9?數個處理裝置 判疋資料製作部23 ,鸦：士 g A & 42，依據判定條件資^夂個部22 _出的異常資料 在各個監視區間中，二、n間序列分割為複數監視區間， 定結果組合而成二致，並製作由判 f。另，判定資料製作部23，可在處理二記憶判定資料 兄，將各個感測器30收集到之里^測為30為複數 間序列分割為複數監視區間^^=2’分別於各個時 風現£間+，判定是否與複 14 201241872 數判定條件一致，並掣你士 43。如此藉由製作判定^斗由幻此荨可判，果組合而成的判定資料 之變動傾向。 、 了孑、，、田且明確地標示異常資料42 【0050】 分値是否位於既定閾値内、於既定閾倍内、時間積 定閾値内、至處理結束為止之間是否位於既 於依據複數時間序列資斜々疋^於既疋閾値内、是否位 所計算出之變_仙〃、序觸平均値與其標準差 ，定條件之間値、件是否一致。各 29記憶判定條件資料46.。σ個瓜視&間決定，於資料保持部 【0051】 診斷部24，藉由將判定資料製 兵對應於判定資料43之模式::衣作m料43， 此外，診斷部24，可計皙判定次粗進仃對照以推定異常原因。 =1】的順序進行異常原因選擇對象之順位“。& 模式資料44,依據各個里常原 =向而預先設定，並與：^^料41其特 $ J原因，有例如處理液供給部5()4 成:作為既定 吕内產生的氣泡、未喷吐光 x生之.犯入處理液供給 :間點不一等。若於處理裝置發管滲漏、閥開閉 ，’將此等在各個異常原因f 而模式資料44 4，以製作判定資料43B夺之判傾向之時間序列資料 料44 ’預先記億於資料果。所 資科^麵常相===：： 【0053】 201241872 此外 資料41 生之時間序列 44 ’可對1個判定資料幻 白_杈式。亦即，模式資料 料41其特有變動傾向與複數判發生之時間序列資 位順序製作由複數階層構成之模式^連^之=序，以此-順 位的順位内為止之判定條件的判定自至較既定順位更為下 其後；序下位的判定條件“更下二 第1再診斷部25，於診斷部24之對 更模式資料44並進行對照 ^=情，，藉由變 Τ’可計算判定資料43與模式W再， ====位r此處 複數致而無料認為丨;^^包含由於與 判定條件的判定結果對押之上声模自最至既定順位内為止之 【0056】 更判不—朗情況，變 後之判定絲岭喊致’並製作由變更 除了指將丨_定料變更為其他狀條 16 201241872 更1個判条件其既定閾値的情況。此外 除了指與全部的異常原因不一致的情况以卜對的情況， 常原因-致而無法確認為i個異常原因的 於與複數異 【0057】 月〆〇 定次^ 2 f珍斷部27 ’藉由將以再判定資料製作部26势作之爯刺 疋貝枓45、與於各個既定異常原因決 衣作之再判 $式_44進行触，以推定異常仙。此V，第2^f 45 27 ’可計算再判定資料45與模式資料44之 $再杉斷部 的順序進行異常原因選擇對象之順位排序。度，以—致度高 【0058】 甘—顯不處理部28，將例如異常偵測結果、里 一 热員示裝置37。此外，顯示處理部28，可=9、，°果顯示於 44等-同顯示。_出異常的情況及;^二貝=、模式資料 於顯示裳置37使警報訊息明滅、或鳴| “原因咐況，可 【0059】 流排=====，，介由内部匯 ^ Mi〇m^35 ^ 37 ^w 36 ^33' 部記憶 =貞測。P 22、判定資料製作部23、診斷部％、第]^ 2卜異 再判定資料製作部26、第第1再矽斷部25、 控制部3!與實行於其上1程—處理部28，係以 【0061】 主記憶部32由等構 ；;29 j 貼 δ己憶保持。 為°己丨思區域之構造 【0062] 201241872 己憶部S、由1更碟、快閃記憶體等構成，預先記憶用於 ^工制。卩31施侧述處理之程式與處理條件資料等的配方，此 外，=控制部31之指示，將此—程式之資料供給予控制部31， 亚記憶自控制部31供給之資料。 【0063】 記憶媒體34》電腦可讀取之記憶媒體 之程式。用於在控制部31施行前述處理之程=㈡ 触記憶。作為記醜體34，.可使用 【0064】 操作部35，為使作業者對異常診斷系統2〇給予指令， 設備等、以及將其等與内部匯流排38連接之介面。 2細作4 35，輸人該配方的選擇、異常_ 指令’供給予控制部3卜此外，作章者"辦训始寺之 式資料44、判定條件資料4/等^^^ _如模 【0065】 叉' ^裝置37 ’由CRT或LCD等之顯示器 二令’顯示圖表化的異常資料42、異常 果等。此外，顯示裝置37可具備，依來自控制果 警報之揚聲器等的輸出設備。 、。-P々而鳴 【0066】 收發部36具備··可將與設置於各處理部之 的配線進行連接之介面機能。控制部31 接 測㈣接收時間序列資料4卜時間序職料4^^6：；自各感 有收納於未圖示之其碗j服料的情形。此 ^貧料44， 由收發部36 ’自伺服器等接收其等資料。控制部31介 【0067】 '' 接著，對異常價測部22之異常偵測方法加 測揚監視自藉由監視處理製程之感測器3〇所輪出的訊號^ 18 201241872 ° 6 (a) 73進扞# f將取于之守間序列資料41與上下變動閾値資料 貞測作為異f之日_列資料41的異常資料幻 平均ί 3 構成，如圖5所示’由資料選擇部22a、 情況，、資iri ’設定部22c及摘測部22d所構成。此— 選擇次^71、夺^ Μ ’除了記憶時間序列資料41等以外，亦記恃 及均資料72、上下變動閣値資料73、係數資料； 【0069】 4卜ίίί擇部瓜’自資料收集部21收集到之時間序列資料 料7广出二為各個有用處理單位其時間序列資料41的複數選擇資 持部29記憶選擇㈣71。另，資料選^3 阻涂Ϊ由上述光_布•顯影處理裝置處理晶®…後，依據以杏 '4顯影處理裝置中評價晶圓之處理的檢查裝置3 ί ’廷出選擇資料71 (參考圖5)。亦即，藉檢查裳置311結 理狀態異常之晶圓w的時間序列資料41，不被選為選g 【0070】 料’自以資料選擇部瓜選出之複數選擇資 均貧料72，於資料保持部29記憶平均資料72。貪 閑値設定部22c，自平均資料計算部娜計算 次 的、與預先輸入之係數資料74及修正値資料75，以下料 昇與時間一同變動的上下變動閾値資料73。 χ ）計 M=N±kx〇±NxH...... (1) 此處各記號表示： Μ:上下變動閾値資料値 Ν:各個時間序列的平均資料値 k :係數 ' 19 【0072】 201241872 間序糊點中的複數模式資料値之標準差 值;過操作部35設定之任意數 73之容許範圍，設定為可於至少、，y將上下變動閾値資料 處，上下變動閾値㈣73之容許範列區間變更。此 象的時間序列資料41與 ^，侧部22d將監視對 定為異常的範圍。 下又動間値貪料73進行比較時，未判 【0073】 如圖6 (a)所示，偵測部22d 視對象的時間序列資料41，與如同上址―一 21收集到之監 料73進行比較，於監視對象的時 ^疋^^變動閾値資 外時，偵測異常資料42。 j貝枓41位於该閾値範圍 【0074】 的時=料㈣資料保持部29作為異“二=; 〈弟1貫施形態> 其次’對以上述異常診斷祕20診斷作為構成 衫處理裝置其複數處理部之一的處理液供仏土布頦 此—處理液供給部5G，對施行過反射防止5 2，以二兑明。 之晶圓W，塗布光阻液。 、成处里、冷卻處理 【0076】 處理液供給部50,如圖1及圖7所示，载 __ 加_ 53，介由加壓管路53a加壓此—光阻容器51 m夜壓 201241872 达往儲藏槽52 ;隔膜型泵54，介由過漉器&自儲藏槽 =液’並介由處理液供給管路55a往噴嘴55喷吐；加壓源％，用 =將隔膜型泵54介由加壓管路56a進行加壓；減壓源57，、用於將 泵54介由減壓管路57a進行減壓；開 * 吸闕 ^於處理Γ給管路^ ;電子氣動調節器EV，插設於力口 屋官路56a以調整泵内之壓力；可變孔％ ς ; Ζ制氣體之流通·，以及旋盤⑼，將㈣嘴 曰曰圓W水平地保持，配置於塗布單元c〇T 、 /{ 【0077】 於加壓管路53a，在加壓源53與光阻容器51 力管路*，在加壓源％與電子氣動;周；器：ί 曰%置'於減壓管路％，在隔膜型泵54與可變 ，58之間故置壓力感測器观。此外，讀理 , 泵3Γ與59a之間設置流量感測器^。以此等减 達至異常診斷系統2Q。亦即，處 ^ ^之傳 3〇b、撕、雇、30e監視異常之發f。、、^ 5〇 ’错由4個感測器 【0078】 【0079】 流量=。=;資常異顺 常資料42a進行相同處理：：二圖d 異 仙的圖表。此-情況，如圖9(a/r=t:之 21 201241872 42為處理褒置之各個處理單位 ^為，自停止處理液之供給至⑵ [0080] ^二人’將以步驟Μί偵測出的里 /刀割為複數監視區严曰 1，在各個巧/ϋ4 2，於各個時間序列 條件-致，並製作由判間中’敏是否與複數判定 【0081】 、，〇而成的判定資料43a(步驟Μ2> 此一情況，如圖9 (a)、圖Q r _ 預先設定之判定條件資料46，m戶;處理單位τ，依據 中，判定是顯概判個監視㈣D1、D2，， 【0082】 此情況，判定條件為：平3 感測器：al〜a4、壓力感測哭=値=否位麵定閾値内（流量 既定間値内（流量感測： ^〜二 ^，分値是否位於 準差所計算出之變動閾値内間^的平均値與其標 〜⑷，於個別的各個監視區了4、麗力感測器： 判定。 ]D1 D2、D3、D4以既定閾値 【0083】 區間of、D〇2所二刀對:’f常資料必、42b ’在各個監視 製作由此等般結果組合而成定條件—致’ =❿，將滿足判定條件= °另，_圖^之判定 定條件之判定結果以「而」表^'、、。果以0Κ」表不，未滿足判 [0084] ^ 43a # 44 ^4： ( ；^f(^ 22 201241872 (b)及圖12所不之具體的模式資料4如、她、 數^構成之模式資料44。此處，異常原因為上^ 賀吐光阻液；Q:光阻液自供仏其、&、p V汉為P.未 模，料二、rr依：以Storage 1 ^ Yes: Wafer E-box loading and unloading station 6 is provided with a plurality of wafer cassettes CB that can be placed in a plurality of sealed containers, for example, 25 wafers. Boxes (4) 6 disposed wall surfaces; and a transfer means A1 for taking out the circle w from the wafer cassette CB by means of an opening and closing portion 7 [0041] 曰曰uw. In the processing block 2', the scaffolding units m, U2, and U3 which are multi-staged in the unit of the heating and cooling system, and the transfer of the wafer W between the processing units 201241872 including the coating and developing unit described later. The main transport mechanisms A2 and A3 are arranged in an orderly arrangement from the front side. In other words, the scaffolding unit port, the U3, and the main transporting mechanisms A2 and A3 are arranged in a row before and after the self-supporting block i side, and the respective connecting portions form an opening for wafer transport (not shown). The circle W can be freely moved from the scaffolding unit U1 on one end side to the scaffolding unit U3 on the other end side in the processing block 2. In addition, the main transport mechanisms A2, A3' are placed in the space surrounded by the partition wall 8, ', and the partition wall 8 of the area is configured as a scaffolding unit U1, U2, U3 in the front-rear direction when viewed from the carrier block 丨The liquid processing unit on one side and the right side, for example, to be described later, is said to be ', and it is a case of a face and a back surface which is one side of the left side. In the figure, the symbols 12 and 13 are temperature and humidity adjusting means, and each of them is provided. The treatment liquid adjusting device used for the unit, the catheter for temperature and humidity adjustment, etc. The liquid processing unit U4, U5 ', for example, as shown in the figure, is a liquid medicine such as a coating liquid (photoresist liquid) or a developing solution. In the accommodating portion M for the space for supply, the coating unit cot, the developing unit DEV including the developing device, and the anti-reflection film-shaped BARC are stacked in a plurality of stages, for example, five stages. Further, the above two 1; The processing of the liquid processing materials w and U5 is performed by stacking a plurality of units of the front and the rear processing, for example, a 1G segment, having a heating unit for baking (baking) the wafer W, and cooling the wafer. Cooling list [0043] inch area 2 The rear side of the scaffolding unit U3 is connected to the exposure via the interface block 3. The interface block 3 is disposed between the processing block 2 and the exposed block 4 from the front and the rear, and is respectively carried by the casing. The chamber 15 and the transfer chamber 16 are t=the central portion of the transfer chamber 15 is provided with a transport mechanism A4, which is freely movable in the x-axis, the Y-axis, and the two-i-Z axis), and can rotate the arm portion around the vertical axis; The contact unit (10) 17 and the high-precision temperature (not shown), the peripheral exposure unit 19, and the buffer unit (not shown) include the scaffolding unit U3 of the block 2, and the like. The unit performs the wafer [0044] 12 201241872 In the same manner as the above-mentioned photoresist coating and development processing device, the cassette is placed in the carrier block! The mounting portion 5 is omitted. After the cover body, the wafer W is taken out by the transfer means A1. Thereafter, ί ίί : before; a scaffold is applied as a coating treatment, and after cooling treatment, the coating unit COT heats the heat to heat the soap element. (baking treatment), the burdock 3 - a scaffolding delivery unit to the interface 'father 4 two = two, do (10) pass ^ The wafer t received by the peripheral exposure process is exposed to the precision temperature adjustment unit by the periphery, and the temperature of the surface of the W surface is highly accurately adjusted to the inside by a high-precision 1 f-structure A4. Round, degree. The handling mechanism M will take the temperature of the exposed day yen W, the straw from the handling mechanism A to 16. The carrying ί元 r;: reverse, the path is carried to become the scaffolding! Will wafer = 2 'then" after the cooling unit is cooled, make JL0S, load (the original wafer cassette CB for the grilling area. 8 1 load /, block 1 on the stage [0045] 20, Mingzhi The abnormality diagnosis system is described in the following section: Li Changling is shut down in Changyuan; the abnormality is detected in the shadow processing device; and the composite processing of the digital processing device 3' is used to execute the predetermined treatment and diagnosis system 20, and the loan is collected. %Θ, Figure 3 and Figure 4, the different management department, processing list "邛36 '舁 is not set in the complex process processing F = Π Π 的 的 日 41 41 41 41 41 41 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 , for example, pressure salt device, position sensing crying, 22: degree sensor, flow sensor, liquid level It can be measured at 3°. In addition, 2 Bu exceptions are different from the theoretical structure of 22 = $ 2 ^ ^ 2〇, which is from the data collection department 25, and then the second data diagnosis department 曰 24, the first! The time series data 41 is re-diagnosed; the data holding unit 29 memorizes: the judgment 1; the judgment data 43, the pattern data 44, and the re-collection unit 21, which are collected and time-changed by the signals of #, respectively ^ shore two king = sensor 30 output 2 (four) hidden time series data 4 in the data retention department changes the pressure, the temperature, the household, ά曰, ^ 幻 幻 枓 枓 41, for example: with time [_] (4) Liquid level, position, torque, speed, etc. The material 41' is compared with the time series resource time series data 41 collected by the f-receiving two-part 21, and is detected as a unit of abnormality: 9? Several processing device judgment data creation units 23, Crow: g g A & 42, according to the determination condition, the abnormal data of the unit 22 _ is in each monitoring section, the second and n sequences are divided into a plurality of monitoring sections, and the result is combined to form a second. In addition, the determination data creation unit 23 can process the two memory determination data brothers, and collect the respective sensors 30 to measure 30 as the complex sequence division into the complex monitoring interval ^^=2' respectively. It is now +, and it is judged whether it is consistent with the judgment condition of the 201214872 number, and you are 43. In this way, by making a judgment, it is judged that the judgment data of the combination is a tendency to change. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , The sequence is slanted and 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 于 _ _ _ _ _ Each 29 memory judgment condition data 46. In the data holding unit [0051], the diagnosing unit 24 determines the pattern of the data to correspond to the mode of the determination data 43: the clothing is made of m material 43, and the diagnosis unit 24 can calculate皙 Determine the next coarse 仃 control to estimate the cause of the abnormality. =1] The order of the abnormal cause selection object ". & mode data 44, which is set in advance according to each of the constants = direction, and with: ^^ material 41 its special $ J reason, for example, the processing liquid supply part 5 () 4 into: as a bubble generated in the established Lv, no spit light x raw. Into the treatment liquid supply: the point is not the same. If the treatment device leaks, the valve opens and closes, 'this is in For each abnormal cause f, the pattern data 44 4 is used to produce the time series data 44 of the judgment data 43B. The data is pre-recorded in the data. The subject is always ===:: [0053] 201241872 The data 41 life time series 44 ' can be used to determine the data illusion _ 。 。 。 。 。 。 。 。 。 。 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式 模式The order of the ^ is the order of the judgment condition from the order of the order to the next, and the judgment condition of the lower order is the second lower first diagnostic unit 25, and the diagnosis unit 24 For the more mode data 44 and compare ^^, by, change 'Computable judgment data 43 and pattern W again, ==== position r here is plural and is not considered to be 丨; ^^ contains due to the judgment result of the judgment condition, the upper acoustic mode is from the highest to the predetermined position. [0056] It is judged that the situation is not--language, and the change is judged by Silken's slogan's and the production is changed by the change, except that the 丨 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ In addition to the case where the situation is inconsistent with all the abnormal causes, the common cause - can not be confirmed as the cause of i abnormalities and the plural number [0057] month 〆〇 ^ ^ 2 f 珍 27 27 ' The abnormality is estimated by touching the 爯 疋 枓 枓 、 、 、 、 、 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 In this V, the second ^f 45 27 ' can calculate the order of the object of the abnormal cause selection by the order of the re-judgment data 45 and the pattern data 44. The degree is high, and the processing unit 28, for example, an abnormality detecting result and a heat indicating device 37. Further, the display processing unit 28 can display the same as -, and display the same. _ Out of the abnormal situation and; ^ two shell =, mode data in the display set 37 to make the alarm message clear, or sound | "cause the situation, can [0059] stream =====, through the internal sink ^ Mi〇m^35 ^ 37 ^w 36 ^33' Part memory = 贞 。 P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P P The broken portion 25, the control portion 3! and the upper portion of the processing portion 28 are executed by [0061] the main memory portion 32 is isomorphic; [0062] 201241872 The memory unit S is composed of a single disc, a flash memory, etc., and is memorized in advance for use in the system. The recipe of the processing program and the processing condition data are further described, and the control unit is further controlled. Instructed by 31, the data of the program is supplied to the control unit 31, and the data supplied from the control unit 31 is sub-memorized. [0063] The memory medium 34 is a computer-readable program of the memory medium, which is used in the control unit 31. The process of the foregoing processing = (2) touch memory. As the ugly body 34, the operation unit 35 can be used [0064] for the operator to give the abnormality diagnosis system 2 Instructions, devices, etc., and interfaces for connecting them to the internal bus bar 38. 2Details 4 35, the selection of the formula, the error _ the instruction 'for the control unit 3, in addition, the chapter maker " The information of the beginning temple type 44, the judgment condition data 4/etc. ^^^ _如模【0065】 The fork '^device 37' is displayed by the CRT or LCD display, etc. 'Displaying the abnormal data 42 and the abnormal fruit. Further, the display device 37 may be provided with an output device such as a speaker that controls an alarm, etc., and the transmission/reception unit 36 is provided with a wiring that is provided in each processing unit. The interface function is controlled by the control unit 31. (4) Receiving the time series data 4 time sequence material 4^^6: The situation is that the bowl j material is not included in the illustration. The portion 36' receives its data from the server or the like. The control unit 31 [0067] '' Next, the abnormality detecting method of the abnormality measuring unit 22 is added to the monitoring sensor 3 by the monitoring processing process. The signal that is turned out ^ 18 201241872 ° 6 (a) 73入捍# f will be taken from the order The data 41 and the up-and-down fluctuation threshold 値 data are measured as the abnormal data imaginary average ί 3 of the date _ column data 41 of the different f, as shown in FIG. 5 'by the data selection unit 22a, the case, the capital iri' setting unit 22c and the extract The measurement unit 22d is configured. This is the selection of the second and the second, and the data is as follows: in addition to the memory time series data 41, the data is also recorded 72, the upper and lower changes of the information 73, the coefficient data; [0069] 4 mb. The time series data collected from the data collection unit 21 is the second selection 71 of the time series data 41 of each useful processing unit. In addition, the data selection ^3 is coated by the above-mentioned light-cloth development processing device, and then the inspection device 3 ί 'Ting out the processing device for evaluating the wafer in the apricot '4 development processing device Refer to Figure 5). That is to say, by checking the time series data 41 of the wafer w with the abnormal state of the 311, it is not selected as the selection g [0070] The material selected by the data selection department is selected as the average poor material 72, in the data The holding unit 29 memorizes the average data 72. The stagnation setting unit 22c calculates the up-and-down fluctuation threshold data 73 which is calculated from the average data calculation unit and the coefficient data 74 and the correction data 75 which are input in advance. χ) Meter M=N±kx〇±NxH...... (1) The symbols here indicate: Μ: up and down variation threshold 値 data 値Ν: average data of each time series 値k: coefficient ' 19 [0072] 201241872 The standard deviation value of the complex pattern data in the order paste; the allowable range of the arbitrary number 73 set by the operation unit 35 is set to be at least, y will be up and down the threshold value data, and the up and down variation threshold 四 (four) 73 is allowed The variation of the range is changed. The time series data 41 and ^ of the image, the side 22d will monitor the range of the abnormality. When the comparison is made between the next and the next 73, it is not judged [0073] As shown in Fig. 6 (a), the detecting unit 22d views the time series data 41 of the object, and the monitoring information collected as the above address - 21 When the object is monitored, the abnormal data 42 is detected when the object is monitored. When the 枓 枓 枓 位于 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 007 The treatment liquid of one of the plurality of processing units is supplied to the processing liquid supply unit 5G, and the photoresist W is applied to the wafer W, which is subjected to the over-reflection prevention, and the photoresist is applied. [0076] The processing liquid supply unit 50, as shown in FIG. 1 and FIG. 7, is loaded with the pressure line 53a via the pressure line 53a, and the photoresist container 51 m is pressed to the storage tank 52 at night pressure 201241872; The diaphragm type pump 54 is spouted through the sump & self-storage tank = liquid ' and through the treatment liquid supply line 55a to the nozzle 55; the pressure source is used, and the diaphragm type pump 54 is passed through the pressure line. 56a is pressurized; a pressure reducing source 57 is used to depressurize the pump 54 via the pressure reducing line 57a; open * suction 阙 is applied to the Γ supply line ^; an electropneumatic regulator EV is inserted in the force Mouth official road 56a to adjust the pressure inside the pump; variable orifice % ς; circulation of the tanning gas, and the reel (9), the (four) mouth 曰曰 round W horizontally, Placed in the coating unit c〇T, /{0077] in the pressurizing line 53a, in the pressurizing source 53 and the resistive container 51 force line*, in the pressurized source % and electro-pneumatic; week; % is set to % of the pressure reducing line, and the pressure sensor is placed between the diaphragm type pump 54 and the variable 58. In addition, the flow sensor is disposed between the pumps 3A and 59a. The reduction is up to the abnormal diagnosis system 2Q. That is, the transmission of ^^3, b, tear, hire, 30e monitoring abnormality f.,, ^ 5〇 'wrong by 4 sensors [0078] [0079] 】 Flow = .=; The same information is used for the same processing: 2: Figure d (Figure 2 (a/r=t: 21 201241872 42 for each processing device) The processing unit ^ is, from the supply of the stop treatment liquid to (2) [two people' will be detected by the step Μί, the cut/cut is cut into a plurality of surveillance areas Strictly, at each time / ϋ 4 2, at each time The sequence condition-induced, and the determination data 43a formed by the determination of whether or not the sensitivity is determined by the complex number [0081], and is formed (step & 2 > this case, as shown in Fig. 9 (a), Fig. Q r _ Judgment Piece of data 46, m household; processing unit τ, according to the judgment, it is judged to be a monitoring (four) D1, D2, [0082] In this case, the judgment condition is: flat 3 sensor: al ~ a4, pressure sensing crying =値=No bit surface is set within the threshold ( (the flow rate is within 値 (flow sensing: ^~2^, whether the 値 is located in the variation threshold calculated by the quasi-differential 値 的 値 値 値 値 値 値 ( ( ( ( ( Each monitoring area is 4, Lili sensor: judgment. ]D1 D2, D3, D4 with a predetermined threshold 値 [0083] The interval of, D 〇 2 two pairs of pairs: 'f often data must, 42b' in each monitoring production and other results combined to form a condition - to ' = ❿, the judgment result of satisfying the judgment condition = °, and the judgment condition of the _ diagram ^ is expressed by "and". If the result is 0Κ, it does not satisfy the judgment [0084] ^ 43a # 44 ^4: ( ;^f(^ 22 201241872 (b) and the specific model data of Figure 12, such as her, number ^ Mode data 44. Here, the cause of the abnormality is the upper ^ He Xie photoresist liquid; Q: the photoresist liquid self-supply, &, p V Han is P. unmodulated, material 2, rr according to:
，日才間序列貧料41的特有變動傾向，即示、$， 發生時之各個異常原因P、Q : 原因p、Q、R 41，與判定資料j /、特有交動傾向的時間序列資料 D卜D2、D3、、D4，在^^於f個時間序列分割為複數監視區間 樣的判定條件U1〜二=4=中’4以，判定資料咖時同 ⑷進行判定之情況的4。M C1〜C4、dl〜d4、61〜e4、η〜 【0085】 若對’顯示為「◎、。、△」之格子， 發生異常原因p、〇、 m、、 〇〇、△」之格子係為，若 ，為異常值發生之判定條=固 -」的顯示，由於旅从％ rt干另扠式貝枓44a、44b、44(:之 其特有變_^之=^與各個異常原因發生之時間序列資料4i 的判定條Γ1連性低關定條件，故料成柄照之對象 【0086】 fl 〜⑸〜的、bl 〜M、Cl 〜c4、dl〜d4、el〜e4、 4ϊ其特有變動傾常^因P、Q、R發生之時間序列資料 高位的判定條件為°連性向低的順位排序，以關連性最高之最 定條件為「〇:，更3」，其次以自最高位起下移1個順位的判 件為「△」，加以順^^固即自最高位起下移2個順位的判定條 【_7】 、W。 診斷部24 ψ. 的判定結果（僅式貝料44a ’僅對照最高位之判定條件 料顿為1昭3 再診斷部25情照的變更後之模式f …、自^位起至下移〗個順位内（◎、〇)為止之判 23 201241872 ，條件的判定結果。此外，第丨再診斷部μ巾賴的模 ，，對照自最高位起至下移2個順位内（◎、。、△)之 條件的判定結果。 々正<刊疋 【0088】 假，，判定資料43a與模式資料构之對 卿，亦即，模式資料_之:= ’、中個異_原因與判定資料43a—致的情況，可推定 m 驟、、Κ，驟）。另一方面，若被判斷為對照不一致 Ξμ6)。則判定資料咖，與變更之模式資料顿對照（步 【0089】 進行3 /上示之狀資料他與模式資料44a之對㈣果 的格子u2、b2、e2、f4f=&4：^巧p其顯示「◎」 由 被 於與複數異對照—致。此一結果 判斷為對财-致（步驟M3 j '，、、料料1個異常原因 【0090】 之後，於步驟M3判斷為對 措由將判定資料43a、盘變更 致（步驟M3 : NO)，則 式貧料44a變更為模式資料44b，鱼 (a)、（b)所示，將模 上述’模式歸44a與變更之模式資^貝，咖進行對照。如同 之模式貧料44的關係，變更後之次b為，由複數階層構成 至下移1個順位内（◎與。）為止之’對照自最高位起 變更後之模式資料44b，包含變旻兮疋餘件的判定結果。因此， 「◎」，更對照其他的「。」广則之模式資料44a其全部的 【0091】 24 201241872 圖11 (b)所示之判定資料43a與模 異常原因P，關於判定條件a3，對於「貝144b的對照結果， 其判定條件a3之判定結果為「〇κ ^ ’後得的判定資料43a 判斷不為異常原因。此外，显 %此，由於對照不一致， 常原照不〜 另一方面，異常原因Q，如圖11(b)所-「 記之全部的格子中，由於判定資料 「◎及〇」標 不-致之判定條件不存在。此一結果，為「NG」，故 原因Q對照—致（步驟祕：YES)，故^推貝由於與異常 T M4)。此一情況，於顯示裝置37顯示里，為Q (步 之狀況的異常診斷結果(步驟工)阻液 另一方面，假定於步驟Μ6被判 驟施：_，藉由將判定資料恤、亡情況（步 上述地，模式資料咏，為_自最。如同 之判定條件的判定結果。如此，藉順位内為止 式資料44e而重複與判定資位之 原因（步驟M7 : YES—)。 ’可推定異常 【0094】 依上述處理裳置之異常診斷方法及其異 判定從與時間—同變動的時間序·料41 藉由 否與複數判定條件一致，並製作由:’果二二！資料 間序列育料41其特有變動傾向而触奴並對庫生之時 之模式資料44進行對照，以減異常原因，而可 ^個處料財_置之祕，可辟異常間序列監 由於將自監視對象的異常資料42a、42b製作之^定此外’ 25 【0095】 201241872 料44a、44b進行對照，推定異常原因，故可輕 個置於處理裝置的感測器篇、3〇c、3〇d、施為複數 ，各個感測器偵測出的異常資料仏、伽，是否分 資=条件一致，並製作由此等判定結果組合而成的模式 44c ’可將詳細且明確地標示異常資料似、伽 各瓣常原因發 Λ w八^歹料1其特有受動傾向的模式資料44a'44卜44c， 處口^:3=亍=，故可更輕易地賴作為關於處理裝置之 IS:7吊原因的故障原因，並提高其精度。 # 41Λ^T # =、D2 密之異常診斷·早位王體如丁判疋的情況相較可進行更精 【0097】 ” r料41與上下變動閾値資料 伽，故可更輕易地確認作為 的故障原因，並提高其精度。 4置之處理製私其異吊原因 【0098】 此外’判定條件係藉由：平灼信 感測器:al〜a4、屋力感測器:d二 既定閾値内（流量感測器：bl〜M ;巧積刀値疋否位於 否位於依據複數時間序列資料之夂個| =貝裔· el〜e4)、是 差所計—_ ( 〜⑷而加關定，由於可更明確地標示判定=== 26 201241872 44b、44c之各自的變動傾向，故可更輕易地確認作為關於處 5 :里製程其異常原因的故障原因，並提高其精度。 ㈣宙診斷部24之對照不一致的情況，可將模式資料4如 ^為柄式貧料44b以進行對照，可推定更正確之異常原因，並 枓44疋、f對知之模式資料输，重複實行判定資料43與模式資 奴衫朋，並稍财魏傾向^ 製程其異常賴“於處理裝置之處理 【0101】 <弟2實施形態> 判斷ϋ: i貝料43a之模式資料44a進行對照（步驟Μ3)，若 變依據各個異常賴發致時’列㈣4 Ρ 1特有 對照，、推定異===判定資料45的模式資料44d進行 【_】 更為^判;ί=Ϊ變更(步驟M7)，係指將1個判定條件變 如缺的情況，例如將「異常資料42之平均値是= 27 201241872 於既疋間f直内」之掌丨丨^^夂^^ 位於既定間値内」之判$ 為「異常資料42之最小値是否 既定閾値的情況，例如對或指變更1個判定條件其 間値内，縮小既定 致，ίΪ作謂概判定條件— 將再判定資料45 資料45 (步驟⑽）， 則可推定異常原因異常原因―致（步驟磨..YES)， 異常辑結果（步^；^) ) °此—情況’於顯示裝置37顯示 【0104】 二巧ί^^(=照不-致(步驟 定資料45 (牛驟更判^件（步驟Μ7)，再度製作再判 常:因此:重複― 【0105】 如同上述，於診斷部24之對照致 將再判定資料-、與對; ^ 5之模式育料44d進行對照以推定異常原因，可藉變 而製作的再判定資料45推定異常原因，故 確乍為關處理震置之處理製程其異常原因的故障原因 【0106】 此外’於第2再診斷部之對照不一致的情況 卉 作再判定資料45 ’重複實行將“製作5 =疋貝^5、與對應於再判定鑛45之模式㈣44d ， 17精進一；/ k更判定條件或異常値之檢測性能而製作的再判定資 28 201241872 【0107】 <第3實施形態> 於判Ϊ 式判定資，與對應 資料44b進行對昭 j則，判疋-貝料《a，與變更之模式 ^ M3 ^ M3 模式資料44b之一致产（牛驟M 可计异判疋貧料43a與 常原因it擇對象之順= 一 $高的順序進行異 M4) 〇 财1 4M11) ’推定異常原因（步驟 【0108】 模式之-致度，係指對於例如 ί此上i果:=原因選擇對象分別計算(步驟剩。依 【〇口1〇9广 順序進行順位排序（步驟剌。 象，致度之異常原因選擇對 37顯示異常診斷結果（步驟‘。1 置 對象的情況及其他=== 【0110】 -致定資料43a與模式資⑽之 =關於處理装置之處理製程其以 【0111】 29 201241872 <其他實施形態> 另’上述第1實施形態中，於牛 的情泥（步驟M6 : NO)，將判定資乂料 不巧斷為對照不—致 資料44進行對照（轉廳）變更之模式 原因，但在即便重複實行步驟Μ6 κ^以推疋異常 一致的情況（步驟應：N0)，與對照仍不 判定條件（步驟Μ?)，判定是錢Km地’可變更 作由判定結果組合而成的再判㈣件―，並製 判定資料45，與對應於再判定資料45 ^ 藉由將再 驟M9) ’推定異常原因（參考圖3、圖4進仃對照（步 【0112】 ° 驟τντί述第2實施形態中’若於步驟Μ9被判斷為對照不-致（+ =9/NO) ’則進-步變更判定條件（步驟 ;^ 45 (步驟M8)，將再度製作的再判 作的再判定資料45之模式賴44進行 自步驟M7至步驟M9為止之步驟以推定異常原因丄 27夕m複貝订自步驟術至步驟M9為止之步驟而第2再診斷部 仍不一致的情況（步驟膽：N〇)，與第1實施形態同 ί:，由將判定資料43，與變更之模式資料44進行對照J 驟Μ6)，推定異常原因（參考圖3、圖8)。 【0113】 此外，即便重複實行第1實施形態中的步驟Μ6而第i再診 ，部土5之對照仍不一致的情況（步驟M6 : N〇)，即便重複實‘ ，2實施形態中的自步驟M7至步驟M9為止之步驟而第2再診斷 部J7之對照仍不一致的情況（步驟M9 : N〇)，第1再診斷部乃 及=2再診斷部27，與第3實施形態同樣地，可計算判定資料43 ，模式資料44之一致度（步驟刖:川），以—致度高的順序進行異 常原因選擇對象之順位排序（步驟M11)，推定異常原因（.參考 圖3、圖14)。 〆 30 【0114】 201241872 此等資料製麵定㈣f為騎餅42並依據 檢测之時間序列資料41 亚不限疋於此’可街於 f乍由此等判找果組合而1的條/1，並 中至少1個項目為NG的叙’f此一判定資料43 =進行對照，施行異常原因二^式資 以資料4卜無遺漏地進行異常診斷;可成 【0115】 不限it，本發明之實施形態的—例進行說明，但本發明並 ，此一形悲而可採用各種態 ^並 ϊ^ίΜ!：° ?f； 數處::先=布本=;;ir=。，不僅應用於心 之處縣置。 裝置，亦可應用於僅施行1個處理 【0116】 &晶_外之被處理體，例如，:二s: 【圖式簡單說明】 【0037】 農置本發批料診__光時布•顯影處理 圖2顯示上述光阻塗布•顯影處理裝置之概略平面圖。 圖3顯示本發明之異常診斷系、_理論構成之方塊圖。 圖4顯示本發0狀異常診嶋統雜略彳滅之方塊圖。 圖5顯示本發明巾的異f/f貞測部之理論構成的概略方塊圖。 31 201241872 _圖6顯示本發明中之時間序列資料的一例之圖表（a)及顯 示異常資料的一例之圖表（b)。 &圖7顯示構成光阻塗布·顯影處理裝置之處理液供給部與異 书診斷系統其構成的概略剖面圖。 圖8顯*本發明之第丨實施職㈣異常 程序 的流程圖。 圖9 .拉賴王里〉夜供給部中的流量感測器之異常資料的Ϊ 表（a) ’顯示壓力感測ϋ之異料的圖表⑻。、 圖1〇顯示本發明中的判定資料之—例的說明圖。 ，11顯示對照本發明中的欺#料與 ，顯示對照判定資料與變更後的模式資料之—例白^ 圖。圖顯示本發明中進—步被變更之模式資料的一例之說明 流程齡本㈣之第2 料辑綠之程序的 流程L14顯林發明之第3實施賴㈣常賴方法之程序的 【主要元件符號說明】 【0117】 1載具區塊 12、13溫濕度調節單元 收納部 15、16搬運室 17傳遞單元 19周邊曝光單元 2處理區塊 20異常診斷系統 21資料收集部 32 201241872 22異常偵測部 22a資料選擇部 22b平均資料計算部 22c 閾値設定部 22d偵測部 23判定資料製作部 24 診斷部 25 第1再診斷部 26再判定資料製作部 27 第2再診斷部 28 顯示處理部 29資料保持部 3 介面區塊 30感測器 - 30b、30c、30d壓力感測器 3 0e流量感測器 31控制部 ’ 32 主記憶部 33 外部記憶部 34記憶媒體 35操作部 36收發部 37顯示裝置 38内部匯流排 39檢查裝置 4曝光區塊 41時間序列資料 42、 42a、42b異常資料（異常之時間序列資料) 43、 43a判定資料 44、 44a、44b、44c 模式資料 33 201241872 45再判定資料 46判定條件資料 5載置部 50處理液供給部（處理裝置） 51 光阻容器 52儲藏槽 53、56加壓源 53a、56a加壓管路 54 隔膜型泵 55喷嘴 55a處理液供給管路 57減壓源 57a減壓管路 58 可變孔 59a開閉閥 59b 回吸閥 6晶圓匣盒裝卸站 60旋轉夾盤 61過濾器 7 開閉部 71記憶選擇資料 72平均資料 73上下變動閾値資料 74係數資料 75修正値資料 8區隔壁 al〜a4、bl〜b4、cl〜c4、dl〜d4、el〜e4、fl〜f4 判定條 A1傳遞手段 A2、A3 主搬運機構 34 201241872 A4、〜A5 搬運機構 BARC 反射防止膜形成單元 CB 晶圓匣盒 COT 塗布單元 DEV顯影單元 D1〜D4 監視區間 Ml〜Mil 步驟 P、Q、R異常原因 T處理單位 U1〜U3棚架單元 U4、U5液處理單元 W 晶圓（被處理體） 35The tendency of the special change of the day-to-day sequence of the poor material 41, that is, the indication, $, the occurrence of each abnormal cause P, Q: the reason p, Q, R 41, and the judgment data j /, the time series data of the unique interaction tendency Db D2, D3, and D4 are divided into four cases in which the f-time series is divided into the plurality of monitoring intervals, and the determination condition U1 to two=4=medium 4, and the determination is made when the data is determined to be the same as (4). M C1 to C4, dl to d4, 61 to e4, and η to [0085] If the grid of 'displayed as ◎, ., △' is displayed, the grid of abnormal causes p, 〇, m, 〇〇, △ If it is, the judgment bar = solid-" is displayed for the abnormal value, because the brigade is from the % rt, the other forks 44a, 44b, 44 (: its characteristic change _^ = ^ and each abnormal cause The occurrence of the time series data 4i determines that the condition 1 is low and the condition is low, so the object of the handle is [0086] fl ~ (5) ~, bl ~ M, Cl ~ c4, dl ~ d4, el ~ e4, 4 The unique condition is due to the high condition of the time series data of P, Q, and R. The condition of the high degree is the order of the low degree. The most stable condition is "〇:, more 3", followed by the highest The judgment piece that moves down one position is "△", and the decision piece [_7] and W that are moved down from the highest position by two ranks is determined. W. The judgment result of the diagnosis unit 24 ψ. The feed material 44a' is only in accordance with the determination condition of the highest position, and the pattern is changed to 1 after the change of the diagnosing unit 25, and the mode is changed from the position to the lower position ( 〇 之 23 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 Judgment result. 々正<刊疋[0088] False, the judgment data 43a and the pattern data are constructed, that is, the pattern data _:: ', the middle difference _ the reason and the judgment data 43a It can be estimated that m, Κ, )). On the other hand, if it is judged that the control is inconsistent Ξμ6). Then determine the data coffee, and the changed mode data comparison (step [0089] to perform 3 / above the data and he and the pattern data 44a (four) fruit grid u2, b2, e2, f4f = &4; p shows that "◎" is determined by the comparison with the plural. This result is judged as the correctness (step M3 j ', , 1 material cause of abnormality [0090], and then judged as correct in step M3. When the determination data 43a and the disk are changed (step M3: NO), the lean material 44a is changed to the pattern data 44b, and the fish (a) and (b) are shown, and the mode of the above mode is changed to 44a and the mode of change. According to the relationship between the mode and the poor material 44, the second b after the change is the change from the plural level to the next one (◎ and .) The mode data 44b includes the determination result of the remaining pieces. Therefore, "◎" is more in contrast to the other "." wide mode data 44a, all of which are [0091] 24 201241872, as shown in Figure 11 (b) The determination data 43a and the mode abnormality cause P, regarding the determination condition a3, the comparison result of the "Bei 144b" The determination result 43a of the determination condition a3 is "〇κ^', and the judgment data 43a is judged not to be the cause of the abnormality. Further, since the comparison is inconsistent, the normal original is not ~, on the other hand, the abnormal cause Q, as shown in the figure 11(b) - "In all the grids, the judgment condition "◎ and 〇" is not marked - the judgment condition does not exist. The result is "NG", so the cause Q is contrast--(step secret : YES), therefore, ^ is due to the abnormality T M4). In this case, in the display device 37 display, it is Q (the abnormal diagnosis result of the condition of the step (step work) is blocked, on the other hand, it is assumed in step Μ6 It is judged that the stipulations are: _, by judging the information shirt, the death situation (steps above, the pattern data 咏, is the _ from the most. As the judgment result of the judgment condition. Thus, by repeating the information within the order 44e Reason for determining the position (step M7: YES-). 'Probable abnormality' (0094) The abnormality diagnosis method according to the above processing and the different judgment from the time-synchronization time sequence material 41 by the no and plural The judgment conditions are the same and are produced by: '果二二The sequence between the data breeding material 41 has its own tendency to change and touches the slaves and compares the pattern data 44 at the time of the Kushen period to reduce the cause of the abnormality, and can be used to calculate the abnormality of the sequence. The abnormality data 42a and 42b of the monitoring target are compared with the '25 [0095] 201241872 materials 44a and 44b, and the cause of the abnormality is estimated. Therefore, the sensor can be lightly placed in the processing device, 3〇c, 3〇d, applied as a complex number, the abnormal data detected by each sensor, gamma, whether the distribution of funds = the same conditions, and the combination of the results of the determination of the results of the pattern 44c ' can be detailed and clearly marked Abnormal data, gamma flaps often cause Λ w8^歹料1 its unique mode of propensity 44a'4444 44c, mouth ^:3=亍=, so it can be more easily used as a processing device IS: 7 causes the cause of the failure and improves its accuracy. # 41Λ^T # =, D2 The diagnosis of the abnormality of the denseness is better than the case of the early king's body, such as Ding Jue [0097] ” r material 41 and the upper and lower variation threshold 値 data gamma, so it is easier to confirm the fault Reason, and improve its accuracy. 4 The treatment of the system is different from the cause of the suspension [0098] In addition, the 'judgment condition is: Ping burning letter sensor: al ~ a4, house force sensor: d two established threshold (Flow sensor: bl~M; Qiaoji knife 位于 No is located in the plural according to the time series data | = Bei Shi · el ~ e4), is the difference meter - _ (~ (4) plus definite Since the change tendency of the judgment === 26 201241872 44b, 44c can be more clearly indicated, it is easier to confirm the cause of the malfunction as to the cause of the abnormality in the process 5: and improve the accuracy. If the control of the department 24 is inconsistent, the mode data 4 can be compared with the stalk-type material 44b for comparison, and the more accurate cause of the abnormality can be estimated, and the data of the mode is transmitted, and the determination data is repeated. With the mode of slavery, and a little bit of Wei tend to ^ process Processing by the processing device [0101] <Different 2 embodiment> Judging ϋ: i pattern material 44a of the material 43a is compared (step Μ3), and if it is changed according to each abnormality, the column (four) 4 Ρ 1 unique control , and the model data 44d of the determination data 45 is judged to be [_] more judgments; ί=Ϊ change (step M7) refers to the case where one judgment condition is changed to be absent, for example, "abnormal data" The average value of 42 is = 27 201241872 In the case of the f f 直 」 」 丨丨 ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ Or, if one of the determination conditions is changed, the predetermined condition is reduced, and the condition is determined. The data 45 will be re-determined (step (10)), and the cause of the abnormality may be estimated to be abnormal (step grinding: YES) , the result of the exception (step ^; ^)) ° This - the situation 'displayed on the display device 37 [0104] Di Qiao ί ^ ^ (= not to - (step to determine the information 45 (the cow is more judged (step Μ 7 ), re-production and re-judgment: therefore: repeat - [0105] As above, in the diagnosis department 24 Judging the data--and the pair; ^5 mode breeding 44d to compare the cause of the abnormality, and the re-judgment data 45 which can be made by the change is used to estimate the cause of the abnormality, so it is determined that the abnormality cause of the treatment process of the shock treatment is The cause of the fault [0106] In addition, the situation in which the second re-diagnosis unit is inconsistent is the re-judgment data 45' repeated execution will be "produce 5 = mussels ^5, and the mode corresponding to the re-determination of the mine 45 (four) 44d, 17 refined (1) / k is a re-determination resource that is determined by the detection condition or the detection performance of the abnormality 2012 28 201241872 [3107] <Third embodiment>疋-Bay material "a, with the change mode ^ M3 ^ M3 mode data 44b consistent production (the cow M can count the difference between the poor material 43a and the usual reason it selects the object = a high order of the same M4 ) 〇财1 4M11) 'The reason for the presumption of the abnormality (step [0108] mode-to-degree, refers to the calculation of the object for example: ??? According to [〇口1〇9广序 order sorting (step 剌. Image, cause abnormality selection to display abnormal diagnosis results for 37 (step '. 1 set the object and other === [0110] - set The information of the data 43a and the model (10) = the processing procedure of the processing device is [0111] 29 201241872 <Other Embodiments> In the first embodiment, in the case of the cow's mud (step M6: NO), It is judged that the information is unreasonably broken as the mode of the control (transfer hall) change, but in the case of repeated implementation of the step Μ6 κ^ to push the abnormality (the step should be: N0), and the control If the condition is not determined (step Μ?), it is judged that the money Km' can be changed as a re-judgment (four) of the combination of the judgment results, and the judgment data 45 is determined, and the re-judgment data 45^ M9) 'Determining the cause of the abnormality (refer to Fig. 3 and Fig. 4 for the comparison control (step [0112] ° τντ ί in the second embodiment, if it is judged as the control is not - (+ = 9/NO) in step Μ9 Then, the step-by-step change determination condition (step; ^ 45 (step M8), The re-judgment data 45 of the re-determined re-determination method 45 performs the steps from step M7 to step M9 to estimate the cause of the abnormality, and the second re-diagnosis is performed from the step to the step M9. In the case where the parts are still inconsistent (step: N〇), in the same manner as in the first embodiment, the determination data 43 is compared with the changed pattern data 44 (J) (6), and the cause of the abnormality is estimated (refer to FIG. 3 and FIG. 8) [0113] In addition, even if the step Μ6 in the first embodiment is repeated and the i-th diagnosis is repeated, the control of the part soil 5 is still inconsistent (step M6: N〇), even if the actual ', 2 embodiment is repeated In the case where the steps from step M7 to step M9 are still inconsistent with the second re-diagnosis unit J7 (step M9: N〇), the first re-diagnosis unit and the=2 re-diagnosis unit 27, and the third embodiment Similarly, the degree of coincidence of the determination data 43 and the pattern data 44 can be calculated (step: 川), and the order of the abnormal cause selection objects can be sorted in the order of high degree (step M11), and the cause of the abnormality is estimated (. 3. Figure 14). 〆30 [0114] 201241872 This The data is fixed (4) f is the riding cake 42 and according to the time series data of the test, 41 Asia is not limited to this, and the street can be judged by f乍, and then at least one item is determined. For the NG's narration, this judgment data 43 = for comparison, the reason for the abnormality is used, and the abnormality diagnosis is performed without any omission; the data can be [0115] not limited to the example of the embodiment of the present invention. To illustrate, but the present invention, this form of sorrow can adopt various states ^ ϊ ^ Μ Μ : : : : : : : : : : : : : : : : : : : : : : : : : : : : : Not only applied to the county where the heart is located. The device can also be applied to only one processed object [0116] & crystallized outside, for example, two s: [simple description of the drawing] [0037] Cloth Development Process FIG. 2 shows a schematic plan view of the above-described photoresist coating and development processing apparatus. Fig. 3 is a block diagram showing the abnormality diagnosis system of the present invention. Fig. 4 is a block diagram showing the abnormality of the abnormality of the present invention. Fig. 5 is a schematic block diagram showing the theoretical configuration of an exclusive f/f detecting portion of the towel of the present invention. 31 201241872 - Fig. 6 shows a graph (a) of an example of time series data in the present invention and a graph (b) showing an example of abnormal data. <Fig. 7 is a schematic cross-sectional view showing the configuration of a processing liquid supply unit and a heterogeneous diagnostic system which constitute a photoresist coating and developing treatment apparatus. Fig. 8 is a flow chart showing the procedure of the fourth (4) abnormality procedure of the present invention. Fig. 9. 异常 王 王 〉 异常 异常 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量 流量Fig. 1A is an explanatory view showing an example of the determination data in the present invention. 11 shows the comparison between the comparison judgment data and the changed pattern data in the comparison with the present invention. The figure shows an example of the pattern data of the step change in the present invention. The flow of the program of the second program of the second edition of the process of the fourth edition of the process of the green program of the invention is based on the third component of the program of the invention. Explanation of Symbols [0117] 1 Vehicle Block 12, 13 Temperature and Humidity Adjustment Unit Storage Unit 15, 16 Transfer Room 17 Transfer Unit 19 Peripheral Unit 2 Processing Block 20 Abnormal Diagnosis System 21 Data Collection Unit 32 201241872 22 Abnormal Detection Part 22a data selection unit 22b average data calculation unit 22c threshold setting unit 22d detection unit 23 determination data creation unit 24 diagnosis unit 25 first re-diagnosis unit 26 re-determination data creation unit 27 second re-diagnosis unit 28 display processing unit 29 data Holding unit 3 interface block 30 sensor - 30b, 30c, 30d pressure sensor 30 e flow sensor 31 control unit 32 main memory unit 33 external memory unit 34 memory medium 35 operation unit 36 transceiver unit 37 display device 38 internal bus 39 inspection device 4 exposure block 41 time series data 42, 42a, 42b abnormal data (abnormal time series data) 43, 43a determination data 44, 44a, 44b, 44c mode 33 201241872 45 Re-determination data 46 Judgment condition data 5 Mounting unit 50 Process liquid supply unit (processing device) 51 Photoresist container 52 Storage tanks 53, 56 Pressurizing sources 53a, 56a Pressurizing line 54 Membrane type pump 55 Nozzle 55a Process liquid supply line 57 pressure reducing source 57a pressure reducing line 58 variable hole 59a opening and closing valve 59b sucking valve 6 wafer cassette loading and unloading station 60 rotating chuck 61 filter 7 opening and closing unit 71 memory selection data 72 average data 73 up and down variation threshold 値 data 74 coefficient data 75 correction 値 data 8 area partitions a1 to a4, bl~b4, cl~c4, dl~d4, el~e4, fl~f4 decision strip A1 transfer means A2, A3 main transport mechanism 34 201241872 A4, ~A5 Transport mechanism BARC Anti-reflection film forming unit CB Wafer cassette COT coating unit DEV developing unit D1 to D4 Monitoring section M1~Mil Step P, Q, R Abnormal cause T processing unit U1~U3 Scaffolding unit U4 , U5 liquid processing unit W wafer (subject to be processed) 35