201115288 六、發明說明： 本發明主張美國臨時專利申請案第61/222，102號之優先權， 其名稱為「Methods and Systems for Advance Equipment Control/Advance Process Control for Plasma Processing Tools」，申請 於2〇〇9年6月30日，發明人為Venugopal等人，其全文係以參考 文獻之方式合併於此。 本案為美國專利申請案第12/555,674號之部份連續案且主張· 其優先權’其名稱為「Arrangement for Identifying Uncontrolled Events at the Process Module Level and Methods Thereof」’發明人為 Huang等人’申請於2009年9月8曰，此美國專利申請案更主張 美國臨時專利申請案第61/222,024號之優先權，其名稱為 Arrangement for Identifying Uncontrolled Events at the Process Module Level and Methods Thereof」，發明人為 Huang 等人，申請 於2009年6月30日，前述兩案係以參考文獻之方式合併於此。 【發明所屬之技術領域】 本發明係關於電漿處理，特別是關於在基板處理期測 理室内的快速暫態事件之方法與裝置。 ]彳貝叫處 【先前技術】 y處理的進步.已為半導體卫業帶來成長。為了更星 I丄須能夠處理基板成為高品f的半導體製置。+ =格控概程錄，以在基板歧_達齡人滿意的結^ §衣程麥數(例如；RP功率、壓力、值厭兩 、、'口果 度等等)超出預定區間(window)時，可能二成^電襞密 (例如粗劣_刻輪廓、低選擇性;dm愤理結果 定區“識別出當製程參數超出預 在基:反處理期間，可能損傷基板和域賴 某二不可㈣事件可能會發生。為了翻出這些不可控^^之 201115288 等等)有關之⑽。;電壓、反射功率、壓力 組件之狀況和/或訊號電裂處理 表不處理室中之原子或多重組件。，、k、且牛」。司將用以 析由感應器收集之資料類型和㈣量已增加。藉由分 料)相關'Γ資料料和處_容資料(處理室事件資 正動作⑽2方2=超卜過預定區間之參數。因此，可提供改 板和/或處理室組=2^。不可控鄉件，藉此進一步防止對基 【發明内容】 資料組盥一組許、，隹/ =円此方法包含一組感應器，用以將一 位快速暫能搴徠 、督心爭件右弟一- 貝料組包含潛在原 件發生期^的+早^法亦包含儲存發生於潛在原位快速暫態事 弧光簽章相比^二早。此方法更包含將電子簽章與-組儲存之 類為第-原^迷符合，^方法更加包含將電子簽章分 定此第-原位快^以心預定之閾值範圍來判 【實施方式] 、 圖式所展示之實施例加以詳 細節。然而，孰朵ιίΐϊ對本發明之完整了解，提出許多特定 部此等特定細::、=本技，者應當了解’本發明可在沒有部份或全 將不再詳述羽=加以貫施。在其它例子中，為了不模糊本發明， 以白處理步驟和/或結構。 亦可涵蓋包種貝鈀例，包括有方法以及技術。應記住本發明 電腦可讀媒體(在其上儲存有用以執行本發明技 述 在以下敘述中 ^毛明將參照其一些如隨附之 201115288 術之實施例的電腦可讀指令)制 如，半導體、磁性、光磁、光讀媒體可包含’例 讀碼之電腦可讀媒體。再者，太鉻Ba二匕形式之用以儲存電腦可 施例之設備。此等設備可包含用二執用以實施本發明實 業的專用和/或可程式化電路。此等言^^本發明實施例相關之作 或經適當程式化之專用型計管梦#又、列子包含通用型電腦和/ :相關之各種作業的專用/VV式化電 解決期==二能夠有效且有效率地 析搬器之互連工具魏的先前技術總邏翻有域層次分 數個f程模組，其恃—個;程模複 5 可^個群組工具控制器(沉），例如crc^tM、 以上^製程模_制器(PMC)，例如PMC m、個 器來收蝴·呈參數之資料謙身料）。==間 理期間，硬數個感應器(例如感應器118、12〇、12 、二处 128、130、132、134、136、138 和 14〇)可與製程模、目 之資料。可“器種 .電壓資料。在^集 用二從製程模組收集資料之感應器可為“3、 用 ’—感應器與另—感應器可具有極小或沒有 通常，配置感應器以收集有關一個以上特定參數之量測資 6 201115288 ί握部份感應器並非用以執行處理，每—_库哭可料 理類比資料，计11Γ者界寻)連接。通常配置計算模組來處 二:]將原始類比資料轉換成數位格式。. mtMti 〇 ^18 144^PMC 110 ^ 資料。將减庫哭二隹、、’ ' 处理感應器118所接收之類比電壓 ϊ =集之細專送到主機層次分析伺服器(例如資 中，在將貞料從舰格式#換成數位格式。在一例子 路徑146傳送到資料盒142之前，計管模 ?二==:收集，_轉換成數位格式。… 收,、處理和ι-if 數個來源(包含感應器和製程模組） 析貢料之中央式分析飼服器。通常，—個資料」 U於處理在基板處理期單—製造商之所有群心具收集= 之資:Ϊ送$^料二^42 L料量可顯著少於感應器所收集 -之貝枓里。通常，感應器可從集大量資料。 吓4 : 收集資料之速率可高達每秒丨百葸： j子中感應 應器收集之資料會被傳送然而，只有—部份由感 不將感應器所收集之整個資料流傳 因是由於當使用符合成本效應之芮田：12的個原 制。到達資料盒142之網路線協物的網路頻寬限 ^ Π8 - 12〇 - 122 - 124 , Ϊ二送1 單:如資料盒_大量資料。換言 感應置之大量資料時， A)與貝枓盒142之間的網路路徑可 Γΐΐίϊ/由上述可知，若資料盒142不能處理 傳入^比直’傳运中之負料包可能會被中斷或需要重新傳送，因 此將額外的負擔加諸於已經嚴重餘之網路沒。 斗斗此料ί—14^可能無*處理來自多個來源之大量傳入資 枓，亚同其匕重要功能，例如處理和分析資料。如上所述， 201115288 =古Γ 2不只用以接收傳人f料包，並且时處理和分 =有傳人賢料流。由於資料盒142是對於收集中 器，資料盒142需要足夠的處理能力以對過量it 由於資料盒142之處理資源有限，只有部份從各 盒，在—例子中，在可由單—感 γ皮發送到資料盒142。在-例子中，只有由感應器^ 料的摘要可被傳送到資料盒142。 莱之貪 除了從複數個感應器接收資料，資料盒142亦可 2器接收資料。在-個例子中，可由每—個製程模組“ 1ί，料和歧内容f料(處理室事件簡)，並發送到資料 理室資料。例如，可由PMC11Q收集製程模組資料3 内谷貧料，並經由路徑148傳送到CTC 104。CTC 104不口总 理來自PMC 110之資料，並且處理來自群組工具内之 = 組控制器(例如PMC 112、PMC 114和PMC 116)的資料、。沾板 诵用工具㈣11妓之資雛著經由半導體設備通訊標準/ j叹備&組(SECS/GEM)界面被傳送到晶圓廠主機1〇2 2二cjrc m將從歷11G、112、114和/或116 _之資』 曰=150透過SECS/GEM156而傳送到晶圓薇主機102。例如， 主機102不只可從CTC104接收資料，並可從其它申且工 如CTC⑽㈣8)接收資料。由晶圓廠主機 ，之貝料接著經由路徑158被發送到資料盒142。由於收隼中 數量，並非所有傳送往晶圓廠主機撤之資料都會被發 ί /4Γ钭盒142。在許多例子中，只有資料的摘要可被傳送到資料 心盒⑷可處理、分析和/或關聯化由感應器和製程模組控 木之育料。例如’若識別出—個異常，資料盒12G可接著 刦疋問題之來源，例如與？％^ 11〇中執行之配方步驟不—致之參 201115288 數。一旦已識別此問題之來源，資料盒l42 ，收此訊息之後，晶圓廠主機102可透過secs/g舰 著將訊息轉達到預定:: 受影響基板之後或 ίίϊ,=損傷，並且可能已使—個以上之處理室組件負 衫曰，因此增加廢棄物與增加所有權成本。 、 =;;ΐ之;r42可能仍需時間來處理、⑽ 組延遲接收封鎖訊息的原因是由於經由資料 至資料盒142的實際資料會顯著少於所^^ 了傳送由感應器118所收集之1千兆赫兹資料 人⑷1上，、有份貧料(約1-5赫茲)被傳送。因此，即使資料 :的。因冒f其Γ有來源接枚大量資料，所接收之資料通常是不完 整資料鑑於資料盒142可能無法存取來自所有來源之完 枓、判疋不可控制事件可能會耗費時間。 此外’㈣傳送到資料盒142所經之路徑可能不同。在 感應哭和ϊίί資料轉換成數位資料之後，直接從感應器裂置(即 i妹Hi5十异換、组)傳送出資料。相反的，將由製程模純集之 億ς ^ k長網路路徑(至少經由群組工具控制器和晶圓庭主 因此，資料盒142直到已接收所有相關.資料流才能完成其 此路142之間的網路路徑較長’並且透過 得I之貝枓w通吊會面臨至少兩個瓶頸。第一個瓶頸是在 201115288 群制器。由於群組工具内之製程模組所收集之資料被傳 ，第—個瓶頸會因為必須經由單-_ "二控:4處縣自各健程模組之資料流而產生。有鑑 =ί=ί：ίίίί^，频响瀬之網路路 1〇2之泣旦口個群組工具控制器接收資料，進入晶圓薇主機 亦由於轉接收中之資料而經職塞。 件，要來自不同來源之資料以判定不可控制事 ㈣的======料 徑傳送響正動作之前’透過相同的冗長路 翻。f —Ϊ造成延遲的因素是_化來自各個㈣源之資料的挑 程模料盒142接收巾之資料流通常是從每_絲和/或製 ίίί *料的摘要’因為此等㈣流可能在不同時間間隔 為間絲、118傳送到資料盒⑷的資料流可以i秒 來自PMC 110之資料流可以2秒為間隔。因此，‘ 取rJ疋不可控制事件之前，可能需要時間來關聯化資料流。 經過的傳送到資料盒142所 正確識财猶卿之純行更辨件，可“要在可 群』目1提供之解決方案的另—個缺點是所有權成本.。除了唯持 〜糸統的成本’額外成本與感應器裝置有關。由於每個感 10 201115288 為不同的品牌/型式/樣式’每個感應器裝置通常包含一 ，牙個計算模組。通常需要實體空間來置放每—個感應4 ，置域應11裝置的成本可能會變得很高，特別是對i 地產價格报高的地區來說。 疋了居 為了減少製程模組内之不可控制事件之實 伺服态。圖2展不一互連工具環境的簡單方塊圖，豆呈有用以 =化感應器與製程模組控制器之間的資料的群組工^次解決^ 木0 1，群組工具可包含複數個製程模組控制器（例如 二212、214和216)。為了收集資料以做分析，每個製程 模、、且控制器可與複數個感應器(例如感應器218、22〇、 226、228、230、232、234、236、238 和 240)相連接。每個减库哭 魏(例如絲魏線244)與其對應之製程模 二。%麥數資料。感應器收集之資料可為類比 計鼻模、組218b)可在將資料經由路經246發 运到群組層:V刀析伺服器(例如遠程控制器、242)之前，將資料處理 和轉換成触料。 心概理 卜每個製雜組控亦可傳送#料(例如製程模 處理内容資料)到群組工具控制器（例如CT(： 2Q4和 例子中，可將pMC 21〇收集之經 綱。除了從PMC⑽接收資料之外，CTC2〇4亦可= 匕衣紅巧組控制态(例如PMC 212、214和216)接收資料。工 具控，之$料接著經由路徑250被發送到晶圓廠主機202。 政广0^曰主機2〇2和CTC 2〇4之間，可將序列分接頭與網路 Ϊ、,’，序列Ϊ接頭2〇8可截取由CTC 204發送到晶圓廠主機202之. 資Ϊ :此貝?冒被複製，且資料流的副本會經由路徑254被傳送 二2。右晶圓廄主機與一個以上之群組工具控制器相 連接，就母個群組工具控制器而言，專用遠程控制器會與群組工 11 201115288 具控制器聯結。在一個彳φ+ s . CTC 206 252 ==路徑258被傳送到遠程控制：0二貝 聯結之遠程控制器(242)不同。 ws^ 資料因利用/個遠程控制器’來處理來自各個群組工具之 個遠程控制器將處理來自較少、數聯結。由於每 轉一群相工且二文目貝枓源(例如製程模:1a控制器和 ϊΓ每諸，每個遠程控制器能夠處理 木源之較大1的貧料。在—例 若識 模組控制器(在此例子中，其為PMC2l^/鎖i到預期之製程 個群群組工具而非來自複數 更好與更快的分析=提二因此’遠程控制器可執行 在下-個基板批次中發生(例 挽救至少—部份預定之之封鎖訊息能夠使製程工程邮 雖然遠程控制器解決方荦比資 器解決方案仍取決於摘要好，遠程控制 =能發生的問題可能維持不明。再者了因此，，處理期 間輸仍不是—直接健，，電 12 201115288 異’其可能使得遠程控制器要關聯化來 自感應β之貝枓與來自製程模組之資料變得困難。 因，’即使逐程控制器解決方案 遠程控制器解決方案仍是不適當的。充其量，此封鎖訊 土=，:_之問題在下個基板處理期間發生。 成之停:可：=== 用乂識別不了控制事件之即時解決方案是必須的。 依，本發明之實施例，提供—種製程層 ^rocess.eve troubleshooting architecture，PLTA)，* t 應器之間平衡負載與容錯_t例更包含在感 在-實L;中感應器交流》 ϊ飼服器可從製程模組和感應器連續1 ;器可從製程模組接收資料，且製程=二=收; 可收:個基板。在基板處理期間， 之資料。若此處理花費—小日士 ^，母隔1〇0毫秒收集關於壓力 之資料項。然而，除了^資料36,〇〇〇筆關於壓力參數 資料(例如電麼偏壓、溫度等等)#=夕卜複數個其它製程 會收集到相當可觀的資料量。）日此，到基板處理已完成的時候， 集之在=析用4=，數個製程模组收 到可用以處理從複數個群組工I收Ufa 242)，或是被傳送 !之資料盒哨。由於資料流^秦之貝料的分析伺服器(例如圖 和/或關聯化此資料。再者自複數個來源，需要時間來分析 冉者由於先前技術之分析伺服器可能無法 13 201115288 處理和分析所有收集 會被傳送到分析伺鮮。=’只有—部份從各個來源收集之資料 料流之複雜任務需要^ ’調:處理' Μ化和/或分析資 在本發明之一熊揭由疋此輕易取得之時間。 料(granular data)^分批’發明人在此了解到如果有更多精細資 分析來自單—來源之可執行更準確與更快的分析。為了 源之資料。在一實施例d 服器必須分析來自較少來 分析資料。換古之，接在衣私模組層次提供裝置以處理和/或 模組及其，應之感應J執#上，組層次分析飼服器以對每個製程 上處理器憶鋈分析伺服器包含可具有—個以 上之感 以彼此ff其處理能力，以 器收集之資料。由於每個^ 异模組以處理感應 處理咨1經歷資料超載，同時處 例子中，若 理器2可被用來協助處理器i以^來自接^^或無資，時，處 再者，在先前技術中，由料瞀上1之育料。 樣式，若-計算模組故障，並它 1^^=^不同的品牌/型式/ 之間的工作量可重新分配。例如，有#要，處理器 =作量可被錄分_奸處理 ^ 其功能， ^内容可了解到，本發明之處理器排益^皮修好。由上 因此亦減少置放計算模組所需之實體空間。^计异杈組之需要， 在本發明之一實施例中，處理器可被分為％## 處理器和次級處理器。配置她和次級處:兩兩來^級 14 201115288 = 聯結，則 處感應器2和3聯結 这兩個感應器(2和刃之資料。 ^逋處理來自 在=施例中，共享記憶體主幹可包含一個以上 =初、、及处理$組不只可用來處 = 處理來自製程模相夕咨极lL aL曰4應°。之貧枓並且可用來 個來源(例如哭外，配置初級處理器組以關聯化各 級處=組可用來傳送封鎖訊息到製程模組控制器鎖心初 ί 當可，以下圖式及朗而更加明白。 ，間早偷要。雖然製造商可有-個以上之非!::2 、、且工具被用來說明本發明之—實;早， 組工具。 口所不之例子包含具有四個製程模組之單-群201115288 VI. INSTRUCTIONS: The present invention claims the priority of US Provisional Patent Application No. 61/222,102, entitled "Methods and Systems for Advance Equipment Control/Advance Process Control for Plasma Processing Tools", applied at 2〇 On June 30, 1989, the inventor is Venugopal et al., the entire disclosure of which is hereby incorporated by reference. This is a continuation of the U.S. Patent Application Serial No. 12/555,674 and claims that its priority is "Arrangement for Identifying Uncontrolled Events at the Process Module Level and Methods Thereof" and the inventor applied for Huang et al. September 8, 2009, this U.S. patent application claims the priority of U.S. Provisional Patent Application No. 61/222,024, entitled "Arrangement for Identifying Uncontrolled Events at the Process Module Level and Methods Thereof", inventor Huang et al. The application was filed on June 30, 2009, and the aforementioned two cases are hereby incorporated by reference. BACKGROUND OF THE INVENTION 1. Field of the Invention This invention relates to plasma processing, and more particularly to methods and apparatus for rapid transient events in a processing chamber during substrate processing. ] 彳贝叫 [Previous technology] The progress of y processing has brought growth to the semiconductor industry. In order to be more star-shaped, it is necessary to be able to process the substrate to be a high-quality semiconductor device. + = grid control overview, in order to meet the satisfaction of the substrate _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ When it is possible, it may be 20% electric (such as poor _ outline, low selectivity; dm anger results in the fixed area) to identify when the process parameters exceed the pre-base: during the reverse processing, may damage the substrate and the domain Lai No (4) events may occur. In order to turn over these uncontrollable ^15 201115288, etc.) (10); voltage, reflected power, pressure component status and / or signal sclerotherapy processing table does not process the atom or multiple in the chamber Components., k, and cattle. The Division will use the type of data collected by the sensor and (4) the amount has increased. By means of the sub-materials, the relevant information and the information (the processing room event is positively operated (10) 2 square 2 = super-adjusted the parameters of the predetermined interval. Therefore, the modification and/or processing room group = 2^ can be provided. Uncontrollable rural parts, thereby further preventing the base [invention] data group 许 、,, 隹 / = 円 This method contains a set of sensors for a quick temporary 搴徕, supervision of the pieces The right brother one - the shell material group contains the potential original occurrence period ^ + early ^ method also contains the storage occurring in the potential in situ fast transient event arc signature compared to ^ two early. This method also includes the electronic signature with the - group The storage method is the first-original match, and the method further includes assigning the electronic signature to the first-in-situ fast-to-center threshold range to judge [embodiment], and the embodiment shown in the figure is detailed. However, in the complete understanding of the present invention, the specific details of the present invention are proposed. In other examples, in order not to obscure the invention, the steps and/or structures are white-processed. It is also possible to cover the case of the palladium, including methods and techniques. It should be borne in mind that the computer readable medium of the present invention (on which is stored useful to carry out the teachings of the present invention in the following description) The computer readable instructions of the embodiment of 201115288 are, for example, semiconductor, magnetic, magneto-optical, optical reading media may include a computer readable medium of the example reading code. Furthermore, the chrome Ba binary form is used for storage. Computer-executable devices. These devices may include dedicated and/or programmable circuits that are used by the present invention to implement the present invention. These are related to or appropriately programmed for use in the embodiments of the present invention. Type meter management dream #又,列子 contains general-purpose computer and /: related to various operations of the dedicated /VV-type power-saving solution period == two can effectively and efficiently resolve the interconnector tool Wei's prior technology total logic There are domain level scores of f-range modules, which are one-by-one; Cheng model complex 5 can be a group tool controller (sink), such as crc^tM, above ^process mode controller (PMC), such as PMC m, To collect the butterfly and present the parameters of the data. == During the interval, a number of sensors (such as sensors 118, 12〇, 12, two 128, 130, 132, 134, 136, 138, and 14〇) can be used with the process module and the data. "Device type. Voltage data. The sensor used to collect data from the process module can be "3. Use" - the sensor and the other sensor can have little or no normal, configure the sensor to collect relevant The measurement of one or more specific parameters 6 201115288 ί grip part of the sensor is not used to perform processing, each - _ library cry can be used to compare analog data, counted. Usually the computational module is configured in two places:] Convert the original analog data into a digital format. . mtMti 〇 ^18 144^PMC 110 ^ Information. The library will be sent to the host analytic server (for example, in the capital, and the data will be changed from the ship format# to the digital format by the subtraction of the analog voltage ϊ = set received by the processing sensor 118. Before an example path 146 is transferred to the data box 142, the meter module 2 ==: collection, _ is converted to a digital format.... Receive, process, and ι-if several sources (including sensors and process modules) Centralized analysis of the feeding device of the tribute. Usually, the data is processed in the processing period of the substrate - all the collections of the heart of the manufacturer = the capital: Ϊ send $ ^ material 2 ^ 42 L material can be significant Less than the sensor collected - in the shell. Usually, the sensor can collect a large amount of data. Scared 4: The rate of data collection can be as high as one hundred per second: the data collected by the sensor in the j sub-sent will be transmitted Only the part of the data collected by the sensor is not due to the use of the cost-effective 芮田:12 original system. The network bandwidth limit of the network route agreement reaching the data box 142 ^ Π8 - 12〇- 122 - 124 , Ϊ二送1 Single: as information box _ a lot of information. In other words When a large amount of data should be placed, the network path between A) and the cassette 142 can be Γΐΐίϊ/ from the above, if the data box 142 cannot process the incoming packet, the negative packet in the transport may be interrupted. Or need to re-transmit, so the extra burden is added to the already severe network. Fighting this material ί—14^ may not deal with a large amount of incoming assets from multiple sources, and its important functions, such as processing and analyzing data. As mentioned above, 201115288 = GuΓ 2 is not only used to receive the passer-by, but also to process and divide = there is a flow of people. Since the data box 142 is for the collection, the data box 142 needs sufficient processing power to handle the excess. Since the processing resources of the data box 142 are limited, only part of the box is used, in the example, the single-sensing gamma can be used. Send to data box 142. In the example, only the summary by the sensor can be transferred to the data box 142. In addition to receiving data from a plurality of sensors, the data box 142 can also receive data. In an example, each process module can be “1 ,, material and ambiguous content f (process room event summary), and sent to the data room data. For example, the process module data can be collected by PMC11Q 3 And transmitted to the CTC 104 via path 148. The CTC 104 does not have the information from the PMC 110 and processes the data from the group controllers (eg, PMC 112, PMC 114, and PMC 116) within the group tool. The board tool (4) 11妓 is passed to the fab host via the semiconductor device communication standard / j sigh & group (SECS/GEM) interface. 1 2 2 2 cjrc m will be 11G, 112, 114 And/or 116 _ 』 150 = 150 is transmitted to the wafer host 102 through the SECS/GEM 156. For example, the host 102 can receive data not only from the CTC 104, but also from other applications such as CTC (10) (4) 8). The round factory host, which is then sent to the data box 142 via path 158. Due to the number of receipts, not all of the data transferred to the fab host will be sent 254. In many instances, Only a summary of the data can be transferred to the data box (4) for processing , analyzing and/or associating the nurturing of the wood by the sensor and the process module. For example, if an abnormality is identified, the data box 12G can then be the source of the robbery problem, for example, executed in ?%^11〇 The recipe step is not—the number of 201115288. Once the source of the problem has been identified, the data box l42, after receiving the message, the fab host 102 can transfer the message to the reservation through the secs/g ship:: After the affected substrate Or ίίϊ, = damage, and may have caused more than one of the process chamber components to be negative, thus increasing waste and increasing the cost of ownership. , =;; ΐ; r42 may still take time to process, (10) group delay receiving blockade The reason for the message is that the actual data passing through the data to the data box 142 will be significantly less than the 1 gigahertz data collected by the sensor 118, and there is a poor material (about 1-5 Hz). ) is transmitted. Therefore, even if the data is: because of the fact that it has a large amount of data, the received data is usually incomplete. In view of the fact that the data box 142 may not be accessible from all sources, the judgment may not be Control event It can take time. In addition, '(4) The path to the data box 142 may be different. After the sensor is crying and the data is converted into digital data, it is directly transmitted from the sensor (ie, the i-Hi5 is changed, the group) The data will be reversed. The process will be purely set by the process network. (At least via the group tool controller and the wafer master. Therefore, the data box 142 can only complete all relevant data streams.) The network path between the roads 142 is long and there are at least two bottlenecks through the I. The first bottleneck is the 201115288 grouper. Since the data collected by the process module in the group tool is transmitted, the first bottleneck will be generated by the data flow of each of the four health-sense modules through the single-_ "second control: four counties. Having =ί=ί: ίίίί^, the frequency response to the network road 1 〇 2 泣 旦 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个 个Pieces, to obtain information from different sources to determine the uncontrollable matter (4) ====== Before the transmission is positive, the same lengthy road is turned over. f - The factor causing the delay is that the data from the respective (four) source of the pick-up module 142 receives the data stream of the towel is usually from the summary of each _ silk and / or _ _ _ _ _ because this (four) stream may The data stream transferred to the data box (4) at different time intervals for the inter-wire, 118 can be i seconds. The data stream from the PMC 110 can be spaced for 2 seconds. Therefore, it may take time to correlate the data stream before taking the rJ疋 uncontrollable event. After passing through the data box 142, the correct line of knowledge is more accurate. The other shortcoming of the solution that can be provided in the group can be the cost of ownership. In addition to only the system The cost 'extra cost' is related to the sensor device. Since each sense 10 201115288 is a different brand / type / style 'each sensor device usually contains one, one calculation module. Usually need physical space to place each one Induction 4, the cost of setting up the device should be very high, especially for the area where the real estate price is high. In order to reduce the actual servo state of the uncontrollable event in the process module. A simple block diagram of the interconnected tool environment, the bean is used to solve the problem between the sensor and the process module controller. ^木0 1, the group tool can include multiple processes Module controllers (eg, two 212, 214, and 216). To collect data for analysis, each process module, and controller can be coupled to a plurality of sensors (eg, sensors 218, 22, 226, 228, 230) , 232, 234, 236, 238, and 240) Each subtraction of the library crying Wei (such as the silk Wei line 244) and its corresponding process model two.% wheat number data. The information collected by the sensor can be analogy meter nose, group 218b) can be used to pass data through the way 246 Shipped to the group layer: Before the V-knife server (such as the remote controller, 242), the data is processed and converted into a touch. The mind can also be used to transfer the material (for example, the process mode). Processing content data) to the group tool controller (for example, CT (: 2Q4 and examples, pMC 21〇 can be collected. In addition to receiving data from PMC (10), CTC2〇4 can also be controlled by 匕衣红巧组The state (eg, PMCs 212, 214, and 216) receives the data. The tool control is then sent to the fab host 202 via path 250. 政广0^曰Host 2〇2 and CTC 2〇4, The sequence tap and the network Ϊ, ', the sequence Ϊ connector 2 〇 8 can be intercepted by the CTC 204 to the fab host 202. Ϊ : This is copied, and the copy of the data stream will pass the path 254 is transmitted two 2. The right wafer controller is connected to more than one group tool controller, and is controlled by the parent group tool. In this case, the dedicated remote controller will be associated with the grouper 11 201115288 controller. At a 彳φ+ s . CTC 206 252 == path 258 is transmitted to the remote control: 0 two-bay coupled remote controller (242) The difference between the ws^ data and the remote controllers from each group tool will be processed from fewer, numbered connections due to the use of a remote controller. Since each group of commutators and two catalogues (eg process) Mode: 1a controller and ϊΓ each, each remote controller can handle the larger one of the wood source. In the case of the module controller (in this case, it is PMC2l^/lock i to the expected process group group tool instead of from the complex number and faster analysis = mention two therefore 'remote controller Executable in the next-substrate batch (for example, at least some of the predetermined blockade messages can make the process engineering post. Although the remote controller solves the solution, the solution still depends on the summary, remote control = can occur The problem may remain unclear. Moreover, during the processing period, the transmission is still not - direct health, and it may cause the remote controller to correlate the data from the induction beta and the data from the process module. Difficulty. Because, 'Even the remote controller solution for the per-process controller solution is still not appropriate. At best, this blockade message =, : _ problem occurs during the next substrate processing. == It is necessary to identify an immediate solution for controlling events. According to an embodiment of the present invention, a process layer ^rocess.eve troubleshooting architecture, PLTA), * t Inter-balanced load and fault tolerance _t is more included in the sense - real L; sensor communication" ϊ feeding device can be continuous from the process module and sensor 1 ; device can receive data from the process module, and process = two = receivable; receivable: a substrate. During the processing of the substrate, the data. If the processing cost - Xiao Ri Shi ^, the mother separated the data item about the pressure of 1 〇 0 milliseconds. However, in addition to ^ data 36, 〇〇〇 Pen on the pressure parameter data (such as electric bias, temperature, etc.) #= 夕 复 a number of other processes will collect a considerable amount of data.) On this day, when the substrate processing has been completed, the set is in the analysis With 4=, several process modules receive the data box whistle that can be used to process Ufa 242) from multiple group workers. Because of the data flow ^Qin's analysis server (such as map and / or associating this information. Again from multiple sources, it takes time to analyze the latter due to prior analysis of the server may not be able to process and analyze 13 201115288 All collections will be transmitted to the analysis store. = 'Only - some of the complex tasks of data streams collected from various sources need to be 'tune: process' deuterated and / or analyzed in one of the inventions The time of easy access. Granular data^ batches 'Inventors here know that if there are more detailed analysis of the single-source implementation, more accurate and faster analysis. For the source of information. In an implementation Example d The server must analyze the data from less analysis. In the past, the device is provided at the level of the private module to process and/or the module and its sensory J. For each processor on the process, the analysis server contains data that can have more than one sense of each other to process the data, and the data collected by the device is processed by each processing module. At the same time, in the example, if the processor 2 can be used to assist the processor i to receive the ^^ or no capital, then, in the prior art, the material of the material 1 is used. If the -computation module fails, and the workload between its 1^^=^ different brands/types/ can be redistributed. For example, there are #要, processor=quantity can be recorded. It can be understood that the processor of the present invention is improved, thereby reducing the physical space required for placing the computing module. In the embodiment of the present invention, The processor can be divided into %## processor and secondary processor. Configure her and secondary: two or two to level 14 201115288 = junction, then sensor 2 and 3 are connected to these two sensors (2 And the data of the blade. ^逋Processing from the example, the shared memory trunk can contain more than one = initial, and processing $ group can not only be used = processing from the process mode 夕 极 pole lL aL 曰 4 should ° It is barren and can be used as a source (for example, crying, configure the primary processor group to associate the levels = group can be used to transmit the blockade message The information to the process module controller locks the heart at the beginning of the ί, if the following figure and Lang and understand more., early stealing. Although the manufacturer can have more than one!!: 2, and the tool is used Illustrating the present invention - the actual; early, group tool. The example of the mouth contains a single-group with four process modules
306 ϋΐ程模組收集之資料由其對應之製程模¥制哭(PMC ==:=:r312)所收集，並經由== 可由雨結34G而傳送到晶_主機3, 組資料和触心製程模 =機=資料並不依賴於專= 反而疋，士貧料可被歸槽且可用於未來的分析。難排解 Μ在Γ實施例中’提供製轉組層次分油鮮(APECS 3⑷ 資料f f 感應器316、318和32Q從_ 308收隼 ;從=:二^=，3〇8收集電麵； 和332來收隹1 Y處 分別經由感應器纜線330 口 感ί Γ集之資料可接著經由路經您、324 谈其中之被傳达到八舰讀以進行處理和/或分析。 15 201115288 不同於先前技術，感應器收集之資料在被傳送到分析 (APECS 314)之前，不需要被預處理_口被摘要化）。在—實施^ 中，取代用計算模組來處理資料，每個感應器可包含—簡|資 轉換器，可用以在發送資料到APECS314之前將類比資^換成 數位資料。或者，在一實施例中，f料轉換器，例如場效可程式 閘極陣列(field-pr0grammabie gate 町吵，FpGA)，可内建於 314。在一例子中'，每個處理器可包含一資料轉換器演算法 將資料巧換成數位格式成為其處理功能的—部份。由上述内= 了解，藉由去除對計算模組的需求，則會需要較少的 置放群組工具及其硬體。因此，可減少所有權成本。、 a „CS 314被專时處理僅來自—製鋪組及其對腐之 土應态的育料，使APECS 314能夠處理來自單一來源之較高g 量。換言之’取代必須減少從每個感應器傳送之資 回署 APECS 314來處理大部份(若非全部)由每個感應器收集料 一例子^，取代只傳送10到15筆資料項進行分析，現來十^ 個感應1§之兩千筆以上之資料項可由APECS 314進行分因 ^組用於APECS 314以進行處理和分析之資料流為較完整之資 料。中，APECS 314亦絲處理來自製程模組之資 巾龍流在由分油服11 伽資料盒或遠 2触之，：會經由—通過各個伺服器(例如群‘工具控 二合〒等)之冗長資料路徑而傳送)，製程模組收集二資 科日破直接傳送到APECS 314，而不需經過j：它飼服残、、 PMC 308 334 apecs 送到息經由路㈣直接傳 展亍二於製程模組層次分析編之進-步的細節。 功"ini—實施例中，製程模組層次分析飼服器之丨 製二。欠= W為又向伺服益，且用以處理傳入資料】 16 201115288 在識別出不可控制事件時傳送封鎖訊息。 資，源可來自兩個主要來源：感應器收集之資料和製程模組 ，集之貢料。在-實酬巾’处⑽㈣從複數個感應器 J器 410、412、414、416、420、422'424 和 426)接收傳入資料〜。 ^鑑於-些雜J1具擁有者可能已紐資可觀的錢在傳統感麻 $裝置(具有計异模組之感應||)中’ ApECS 4⑻可用以從傳統H 器裝置與改良式感應器(不需要計算模組之感應器)兩者接收資ϋ' 在一實施例中，APECS 400可包含一界面，例如乙太網路交 換态418 ’用以與傳統感應器裝置(例如感應器410、412、414和 =6)互相作用。在一例子中’在將數位資料傳送到ApEcs働(經 哭H^m、434或436)之前，先藉由計算模組41〇b將感應 二。1 浦從類比格式轉換成數位格式。配置乙太網路交 像統感應器裝置互相作用，以接收資料流。接著將資 經由路控446、448、彻或松)APECS伽内之處理 〇。（ 、、4、406和顿)的其中之一，以進行處理。 應哭^置_#製鱗數，可制改良式感 輪、、且。由於收集之資料不需要被摘要化，故不再 來進行處理。反而是’在—實施例中，改良式感應 換器(圖中未示），例如低價的FPGA，用以將“ :内，；式。或者，取代設置資料轉換器在感應 ===於継⑽外部或内部，去除計算 置放和維持計算’實f上去除了用_、 在本發明之一實施例中，APECS 4〇〇包含一組 虛0人資料。此處理器組可為實體處理單元、 二之“ iit每個處理器負責處理來自與此處理器聯 幻2經由路徑44〇流t=、r，由處理盗404來處理從感應器 之音祖、、* 之貝料/;,L。在另一例子中’感應器424收集 、八机！由路控442被傳送到處理器4〇6以進行處理。 、 17 201115288 -例二及巧料於錢麵配置。在 仍可存在其它關係。在—例子t=對-關係， 以上來源之資料，在另_例 ^^處理糾處理來自-個 理來自-個感應器之資料流。了配置—個以上之處理器以處 在一貫施例中，每—個虛採哭处古 > 此，當一個以上之處理哭 。^子/、旱記憶體主幹428。因 若卿哭426 ίΐΐ载呤可執行負载平衡。在-例子中， 言之體主幹;提^境。換 =ί理會被重新分_其它 =來處理來自感應器424之資料流。‘貝Ϊ ；；分 例中，継⑶—内可存在兩種處理ϋ。第-種虚 严ϋί處理器(例如處理器404、406或顿）。配置每個Ϊ級 二，從其對應之感應器接收之資料流。此外，在-實施 二用以分析資料以及識別其對應之感應器所存在 ，一種處理态被稱為初級處理器(402)。雖然圖4只展示一 ^及，理H，初級處雜之數目可取決於使用者的崎。在一杏 =中’可配置—個初級處理ϋ來處理來自-似上感應器之 料k。在一例子中，將感應器收集之資料流經由路徑^僖 送到初級處理器402以進行處理。 έ ^級^理态之另一個資料來源是製程模組。換言之，製程模 城集之製轉崎料和處動容資料會由擁處理1來處理。' =一例子、中，將製程模組收集之資料透過製程控制匯流排經由路 徑454傳送到APECS 400。此資料在經由路徑446流入初級處理 18 201115288 器402之前，先穿過乙太網路交換器418。 l22 ^424 铀斿办ώ , 外你力例子中’亦由初級處理器402 間以關^感應器之資料流與來自-製程模組之資料流之 料要师料源之資料路#現在為約相同長度，關聯化資 月liif經，挑戰簡單許多。在—例子中，由於資料 相關資料流_糊\ ‘ 之所有 來源之資料的關聯化。 〜才胃大巾田間化來自不同 較大資===之;，度的 在先前技射，因為可_分析 於現在每個分析伺服器只負責分析來，° 則技術，由 从此製程模組聯結之感應器）的資料有源:(製程模組 量。由於已經大幅降低了資海之數s =低了*料源之數 ，單一來源之較大量資料量Γ有^提能力處理 達成，各個來源之資概之間的更佳_化、。了更精細的細節，可 ,若識別出問題(例如不可控鮮件），初 鎖訊息到製織组。在—實施例巾，刺 以傳送封 封鎖訊息從継cs 400傳送到製程模組。藉由在兩 19 201115288 數，輸出線’封鎖訊息不需要在可被傳送之前 ° m此’實f上免去了正確祕式化封鎖訊自铁 鍾 近即時封鎖訊㈣賴組件軸訊息或接 在另-例子中可:==f二=二器：广 疑難排解結構。藉提供了製程層次 料精細度以進行分析，而產生更; 貫質減少對處理室組件的損傷。、〃數里的基板，亚可 行及ί本確:態ίΐ發明，解到，藉由能夠執 理快速暫能辜術你U 之衣転層次疑難排解結構，可識別及管 事件、尖； 事ί理㈣可快速發生且通常維持短時間之 3，如微發弧事件、釋放事件、尖突事件等等）。由“件 板整個基 已經識別出發生在基板上之微發孤事件的時候^此== 20 201115288 ，損，且剩餘之基板批:欠也可能已經受損。此外，亦可 處理室内之硬體組件的損傷。 產生對 子欠發胁岐鶴祕11，㈣敝職速暫態電 結果）。然而，大部份快速暫態感應^ 快，態感應器通常沒有將此資料“Ιίϊί 之電子簽早(可用以·潛在傷錄事件)的能力。 。我 考慮到其中的狀況，例如，在蝕刻處理期 成微發狐產生。如本文所述，微雜意指當 宰1^ 板"^圖案造成損傷(例如破壞材料層、破裏圖 生之事件。藉由使用”探針，可收集 與識前時已發錄速鶴事件(例如微發弧事件ί 反而是’快速暫態感'應器收集之資料 由使用者或錄體_來分析。在—例讨: 濟出严於其綱^ 紅悲事件。讀貧料的工作可能不是耗費數星期，就是 分析數：計程可能需要時間來 對，，_或==:的^ 由於被發弧事件通常不是可預期之現象，偵測快速暫 是困/之工作。換言之，例如，微發^非 代表 母個微發弧事件可獨特簽章所 圖5展示-微發弧事件(曲線5〇2)的簡單 圓事件發生時，電麵電流訊號會 歷陡降(5Q4)。財當和輸罐逐漸上❹]高频卿i、r 21 201115288 依據本發明之實闕，提供方法和裝置來處理在錄處理系 統之處理室内的快速暫態事件(例如微發孤事件)。本發明之實施例 包含用以伽快速暫態事件⑽如微魏)的方法 ，包，由執行將-簽章與已知快速暫態簽章(例如 ^歹 二 =ΐί ΐ電子簽章的方法。本發明刺 么重性的方法。本發明之實施例另包含用 2理时暫“相在㈣製造環__鶴減至最小的方 在本木巾可使用祕發弧作為例子來說明 微發弧，且可包含可能在基板處理二發、生之 任何快速暫悲事件。反而是，本案之 t 明並不限定於所提出之實施例。..W例丁用且本發 f數Λ的概眼'岐暫態感應_如vui 速取樣暫㈣測演算法可财 、、且，以提供對感應器(例如VI探針)…接之计昇模 探針)接收資料。在另一徐浐仞由/面及攸感應器(例如VI 針#制哭、石Γ於5另Λ例中，由與感應器控制器(例如vii冗 D互相作用之計算模組來執行快速取樣暫態演匕： 22 201115288 如VI探針)直㈣咖之分析模組 相作VI探針)或者是由與感應器(例如VI_^ .，:在之=來:=在，然後在㈣二 章)可被儲存並發送到分析模袓^ ^電^唬波形(例如電子簽 綱咖4))以進行上;析;;= 只有關於潛在快速暫能帝子欠土日 感應态層次執行偵測， 送到分析模組以做進：：步之^的資料會被向前發 資料路徑而傳送之資料、、& & 可執行過濾以減少沿著 行分以進 ^潛在互:册r 反而疋，作為製程層次疑難排解 不而要過濾貧料。 =cs 314)可具有能夠處理大量資°料的快=\分(例如 iii的製程層次疑難排解結構，可實質上去除二發明 結構十可能產生的—般資料流量壅塞。'^去t其匕類型分析 能夠，速且有效率地分析數以萬計心資料樣品。；明之分析模組 ^ 組電，簽 發弧的各種已知波形，可被儲存在^^事件(例如微 :快逮暫的;電料庫中 件的嚴重性。在一例子中 y接者判疋此快速暫態事 报小的影響或沒有影響此，對處理中之基板有 程度之事件。在事件可被分類為具.有低嚴重性 處理中之基板。因此，此二能已經損傷了目前 程度。 ⑽速暫悲事件可被分類為具有高嚴重性 23 201115288 藉由識別快速暫態事件之嚴重性，可 此快速暫態事件。在本發明之一實施例十，地處理 件之嚴重性來提供預定動作程序。在一 ‘取快速暫態事 度的快速暫態事件可引起警告，而具有高 重性程 事件則可造成蝕刻處理(例如)被終止。 性私度的快逮暫態 為幫助說明’圖6A展示，在本發明之—實 之簡單方塊圖。處理系統600可包含於 中’^里環境 室6=。在基板處理期間，氣體(圖令未示)可mm之處理 hox)m 生，ΐΐίίΓ二 而造成快速暫態事件發 組616識別。在一實二中^ 樣暫態_演算法模 可包含用以定義快速暫態事件演算法模組你 板處，期間來執行快速取樣暫態模二施例中，可在基 探丁組路徑-被發侧 在-實施例中，％探針控制大‘，官理VI探針610。 算法模組616。 °σ亦可包含快速取樣暫態偵測演 在另—實施例中，快速取樣暫離偵測、.當管 $ VI探針控制器612交 二616可為能 610收集之資料可經由VI探針㈣^ 換吕之，VI探針 偵測演算法模組616。葬由°。612被傳送到快速取樣暫態 成為獨立式模組，若VI ^針^態偵測演算法模組616 不需要修正VI探針控制器61=。°π 12無法操作額外的處理，則 傳送經由路議直接 制器612，分析模組618 代將貧料傳送到VI探針控 616。藉由將資料直接傳逆 ^逮取樣暫態偵測演算法模組 料不需要被預處理。& ，VI探針⑽收集之資 玄除叶异模組(例如VI探針控制器 24 201115288 °反岭’ 分析模組⑽來識別潛 由分析模組態電子簽章，可 由將潛在快速暫態電子簽章與組⑽可藉 章(例如一組弧光簽章)相比較來執行j二之:組快速暫態簽 果’則認定已發生快速暫態事件。π早1^ 1。若識別出符合結 •重性#熟者判定快速暫態事件之嚴 (例如強度)程度。因此，提伶m，事件可具有不同的嚴重性 重性。在-實施^中定每個快速暫態事件的嚴 ，圍。如-例子，可將在電流程度/間 f於15微秒之持續時間(定義為從下降到的I降與 债測晶圓上之損傷的適當閾值。 T日 1)，視為用以 序。夬f暫態事件的嚴重性程度，可採取-動作程 範圍聯結：二者施例Πΐ被預定，且可與嚴重性程度/閾值 =為=的; 右古。。去〔、 々電左和琶流下降之快速暫熊電子欠童可·、士 '目或 巧i程度之事件1可能服基之二被視為 内之即時快=暫；說明用以偵測在製造環境 _演算法不是分析模組的1部=仏程圖，其中快逮取樣暫態 在處择=基=板處理。考慮到其一，例如， 快逮暫’彳貞f處理室内之基板處理。在步驟704a， 電壓和電〉:::。;歹其:νι、針)可傾測電性參數(例如在不同階段之 ⑽Μ、基波和_。於大約相同之時間，在步tot 25 201115288 可執行快速取樣暫態偵測演算法。 在下-步驟706，判斷^皆在 乾例如’快速取樣暫態侧演算法可件i否存在。換言 _例如弧)之縣值。若潛在快速暫 之The data collected by the 306 process module is collected by its corresponding process module cry (PMC ==:=:r312), and can be transmitted to the crystal_host 3 via the rain knot 34G via ==, group data and touch Process mode = machine = data does not depend on the specific = instead, the poor and the poor can be returned to the pool and can be used for future analysis. Difficult to solve the problem in the Γ embodiment, 'provide the system to transfer the oil layer fresh (APECS 3 (4) data ff sensors 316, 318 and 32Q from _ 308; from =: two ^ =, 3 〇 8 collection of electricity; and 332 to receive 1 Y through the sensor cable 330 respectively 口 之 之 之 之 之 可 可 可 可 可 可 可 、 、 、 324 324 324 324 324 324 324 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 The technology, the data collected by the sensor, does not need to be pre-processed before being transmitted to the analysis (APECS 314). In the implementation, instead of using the computing module to process the data, each sensor can include a simple converter that can be used to exchange the analog data into digital data before sending the data to the APECS 314. Alternatively, in one embodiment, a f-material converter, such as a field-effect programmable gate array (field-pr0grammabie gate, FpGA), may be built in 314. In one example, each processor can include a data converter algorithm that replaces the data into a digital format for its processing functions. From the above = know that by removing the need for computing modules, fewer placement group tools and their hardware are required. Therefore, the cost of ownership can be reduced. , a „CS 314 is treated exclusively by the time-making group and its nurturing of the rot, so that APECS 314 can handle higher g quantities from a single source. In other words, the replacement must be reduced from each induction. The device transmits the APECS 314 to process most (if not all) samples collected by each sensor, instead of transmitting only 10 to 15 data items for analysis, and now there are 10^ sensors 1 § More than one thousand items of data can be processed by APECS 314. The data flow for APECS 314 for processing and analysis is more complete. Among them, APECS 314 also handles the flow of materials from the process module. The oil distribution service 11 gamma data box or far 2 touch:, will be transmitted through the lengthy data path of each server (such as group 'tool control two-in-one 〒, etc.), the process module collects the second-party branch directly Transfer to APECS 314, without going through j: it feeds the disability, PMC 308 334 apecs sent to the information via the road (4) directly spread the second step in the process module analysis of the step-by-step details. - In the embodiment, the process module hierarchical analysis of the feeding device System 2. Under = W is for servo and is used to process incoming data. 16 201115288 Sends a block message when an uncontrollable event is identified. The source can come from two main sources: information collected by the sensor and process The module, the collection of the tribute. At the - the real reward towel '10 (4) from the plurality of sensors J 410, 412, 414, 416, 420, 422 '424 and 426) to receive the incoming data ~. ^ Given - some miscellaneous J1 owners may have considerable money in the traditional sense of the device (with the sensing module | |) ' ApECS 4 (8) can be used from the traditional H device and improved sensor (no calculation module required The sensor receives both of the funds'. In an embodiment, the APECS 400 can include an interface, such as an Ethernet switching state 418' for use with conventional sensor devices (eg, sensors 410, 412, 414, and = 6) Interaction. In an example, before transferring the digital data to ApEcs働 (by crying H^m, 434 or 436), the sensing module 41〇b is first converted from the analog format by the calculation module 41〇b. Digital format. Configure the Ethernet sensor device to interact with each other to receive funds. Flow. Then pass the traffic control 446, 448, Tear or loose) APECS gamma processing 〇. (, , 4, 406 and Dun) one of them for processing. Should cry ^ set _# scaly number The improved sense wheel can be made, and since the collected data does not need to be abstracted, it is no longer processed. Instead, in the embodiment, the improved induction converter (not shown), for example A low-cost FPGA is used to put ":in,; Alternatively, instead of setting the data converter in the sense of === outside or inside the 継(10), the computational placement and maintenance calculations are removed. The real f is removed. In one embodiment of the invention, the APECS 4 includes a group. Virtual 0 person data. This processor group can be an entity processing unit, and the two of the "iit" are responsible for processing each processor from the processor. The stream is traversed by the path 44, t=, r, and the processing thief 404 is used to process the ancestors from the sensor. In the other example, 'sensor 424 is collected, eight machines! It is transmitted by the road control 442 to the processor 4〇6 for processing., 17 201115288 - Example 2 and the material In the face configuration, there may still be other relationships. In the example t=peer-relationship, the data from the above sources, in the other case ^^^^^^^^^^^^^^^^^^^^^^ More than one processor is in a consistent practice, each one is a virtual crying ancient > This, when more than one handles crying. ^子/, drought memory trunk 428. Because Ruoqing cry 426 ΐΐ呤Executable load balancing. In the example, the body of the word; the context is changed. The change will be re-divided _ other = to process the data stream from the sensor 424. 'Bei Ϊ;; (3) - There may be two kinds of processing ϋ. The first type of 虚 ϋ processor (for example, processor 404, 406 or ton). A data stream received from its corresponding sensor. In addition, in the implementation - two to analyze the data and identify the presence of its corresponding sensor, a processing state is called the primary processor (402). Figure 4 shows only one and the same, the number of primary impurities may depend on the user's singularity. In apricot = 'configurable' - a primary processing ϋ to process the material k from the -like sensor. In one example, the data stream collected by the sensor is sent to the primary processor 402 via the path for processing. Another source of data for the ^^^ state is the process module. In other words, the process of the process module set The raw materials and the moving data will be processed by the processing 1. In an example, the data collected by the processing module is transmitted to the APECS 400 via the process control bus via path 454. This data flows into the primary processing via path 446. 18 201115288 Before the 402, first through the Ethernet switch 418. l22 ^ 424 uranium 斿 ώ 外 外 外 外 外 外 外 外 外 外 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级 初级The material flow of the module must be materialized The source of the data road # is now about the same length, associated with the monthly capital liif, the challenge is much simpler. In the example, the data related to the data flow _ paste \ 'all sources of the association. The fieldization comes from different large capitals ===; the degree is in the previous technique, because it can be analyzed. Now each analysis server is only responsible for the analysis, ° technology, the sensor connected from the process module) The information is active: (the number of process modules. Since the number of capitals has been greatly reduced s = low * the number of sources, the larger amount of data from a single source has reached the ability to handle the achievement, the source of each source Better between _,. For more detailed details, if you identify a problem (such as uncontrollable items), first lock the message to the weaving group. In the embodiment, the thorn is transmitted from the 継cs 400 to the process module. By the number of 2011 19288, the output line 'blocking message does not need to be removed before the transmission can be transmitted. The correct secret blocking is blocked from the iron clock. (4) The component axis message is connected or received. In another example, you can: ==f two = two devices: a wide-ranging problem-solving structure. By providing a process level of material fineness for analysis, the production is more; the quality reduces damage to the process chamber components. The substrate in the number, the sub-feasible and the ambiguous: the state ΐ ΐ ΐ , , ΐ ΐ ΐ ΐ ΐ ΐ ΐ ΐ ΐ ΐ ΐ ΐ ΐ ΐ ΐ ΐ ΐ ΐ 藉 藉 藉 藉 藉 藉 藉 藉 藉 藉 藉 藉 藉 藉 藉 藉 藉 藉 藉 藉 藉The reason (4) can occur quickly and usually maintains a short time of 3, such as micro-arc events, release events, spike events, etc.). By the time when the entire base of the board has identified the micro-orphan event that occurred on the substrate, this == 201115288, the damage, and the remaining substrate batch: the owe may have been damaged. In addition, it can also handle the hard indoors. The damage of the body component. The generation of the pair is under threatening, and the (4) 敝 速 暂 暂 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 Ability to sign early (available for potential incidents). . I considered the situation, for example, the micro-hair fox produced during the etching process. As described in this article, micro-hybrid refers to the damage caused by the pattern of the board (such as the destruction of the material layer, the destruction of the material. By using the probe), the speed of the record can be collected and recorded. Crane events (such as micro-arc events ί, but the 'quick transients' information collected by the user or the recording _ to analyze. In the case of: the economy is stricter than its outline ^ Red sorrow event. Reading poverty The work may not be a few weeks, or the number of analysis: the meter may take time to correct, _ or ==: ^ Since the arcing event is usually not a predictable phenomenon, the detection is fast and sleepy / work In other words, for example, the micro-hairs non-representative mother micro-arc events can be uniquely signed. Figure 5 shows that the simple circular event of the micro-arc event (curve 5〇2) occurs, the electrical surface current signal will drop sharply. (5Q4). The money and the cans are gradually smashed] HF, i 21, 15, 288, according to the invention, providing methods and apparatus for processing fast transient events (such as micro-orphan events) in the processing chamber of the recording processing system Embodiments of the present invention include gamma fast transient events (10) such as Wei Wei) A method, a package, a method of performing a signature-signature with a known fast transient signature (for example, a method of electronic signature). The method of the invention is sturdy. The embodiment of the present invention further comprises 2 Time to temporarily "phase (4) manufacturing ring __ crane to minimize the square in the wood towel can use the secret hair arc as an example to illustrate the micro-arc, and may contain any rapid processing of the second process, the birth of any slow However, the case of this case is not limited to the proposed embodiment. The case of W is used in this case and the f-number of the current eye is '岐 transient induction _ such as vui speed sampling temporary (four) measurement algorithm Finance, and, in order to provide a sensor (such as VI probe) ... connected to the lift model probe) to receive data. In another Xu Wei / face and 攸 sensor (such as VI needle # system cry, sarcophagus 5 In another example, a quick-sampling transient interpretation is performed by a sensor controller (such as a vii-duplex computing module: 22 201115288, such as a VI probe). Needle) or by means of sensors (eg VI_^.,: in ==: in, then in (4) 2) can be stored and sent to the analysis module ^ ^ ^ 唬 waveform (such as electronic signing coffee 4)) to carry out; analysis;; = only on the potential fast temporary power emperor owing to the soil day sensing state level detection, sent to the analysis module to do:: step ^ The data will be forwarded to the data path and transmitted by the data, && executable filtering to reduce the number of points along the line to the potential mutual: book r instead, as a process level troubleshooting, not to filter poor materials . =cs 314) can have a fast =\ points that can handle a large amount of materials (for example, the process level troubleshooting structure of iii, which can substantially remove the data flow congestion that may be generated by the second invention structure. '^ Go to t Type analysis can quickly and efficiently analyze tens of thousands of heart data samples. The analysis module ^ group of electricity, the various known waveforms of the arc, can be stored in ^ ^ events (such as micro: fast catch temporary The severity of the components in the electrical library. In an example, the y picker determines whether the fast transient event has little or no effect on the substrate, and the event has a degree of event. The event can be classified as a tool. There is a substrate with low severity of processing. Therefore, this two can already damage the current level. (10) Suspended sorrow events can be classified as having high severity 23 201115288 By identifying the severity of fast transient events, this can be quickly State event. In one embodiment of the present invention, the severity of the processing component provides a predetermined operational procedure. A fast transient event in a fast transient event can cause a warning, while a high severity event can be a warning event. Can cause etching The processing is terminated, for example. The fast-moving transient of the sexuality is to help illustrate 'Figure 6A shows a simple block diagram of the present invention. The processing system 600 can be included in the environmental environment 6=. During the processing of the substrate, the gas (not shown) can be processed by hm), which causes the fast transient event to be identified by the group 616. In a real two, the transient_algorithm model can be used to Defining the fast transient event algorithm module at your board, during the execution of the fast sampling transient mode, in the case of the base probe group-issued side - in the embodiment, the % probe control is large, The official VI probe 610. The algorithm module 616. °σ can also include fast sampling transient detection in another embodiment, fast sampling temporary detection, when the tube $ VI probe controller 612 cross two 616 can be collected by 610. The VI probe (4) is replaced by the VI probe detection algorithm module 616. The funnel is transmitted to the fast sampling transient to become a stand-alone module, if VI ^ Needle detection algorithm module 616 does not need to modify the VI probe controller 61 =. °π 12 can not operate extra Then, the transmission is passed through the direct controller 612, and the analysis module 618 transmits the poor material to the VI probe 616. By directly transmitting the data, the sampling transient detection algorithm module does not need to be pre-processed. Processing. & VI probe (10) collected by Xuan Xuan leaf module (such as VI probe controller 24 201115288 ° anti-ridge 'analysis module (10) to identify the latent analysis module configuration electronic signature, can be potentially fast The electronic signature and the group (10) can be compared with the chapter (for example, a set of arc signatures) to perform the second: the group fast transient signings to determine that a fast transient event has occurred. π early 1^ 1. If identified The degree of strictness (such as intensity) of a fast transient event is determined by the acquaintance. Therefore, to raise m, events can have different severity. In the implementation - the strictness of each fast transient event is determined. For example, the duration of the current/interval f is 15 microseconds (defined as the appropriate threshold from the drop to the drop of I to the damage on the die wafer. T day 1), as an order . The severity of the 夬f transient event can be taken in conjunction with the scope of action: both instances are scheduled, and can be associated with severity/threshold = =; . Go to [, 々 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左In the manufacturing environment, the algorithm is not a part of the analysis module = the process map, in which the fast-acquisition sampling transient is at the option = base = board processing. Considering one of them, for example, it is quick to catch the substrate processing in the processing room. At step 704a, voltage and power are >:::.歹 :: νι, pin) can tilt electrical parameters (such as (10) Μ, fundamental wave and _ at different stages. At about the same time, in step tot 25 201115288 can perform fast sampling transient detection algorithm. - Step 706, judging that both are in the process of, for example, the 'fast sampling transient side-effect algorithm can be present. In other words, for example, an arc, the county value. If it is potentially fast
At士 〜取·^悲偵測演算法可句八_ ，彳甘。換1 恶事件(例如微發弧)之標準值 用以疋義潛在快速^ =樣暫態_演算法所定義之料符合㈣ 件’且VI楝針會繼續偵測基 “生潛在快迷暫態事 然而，若識別出潛在快速暫能畜^，。 存在潛在快速暫態事件發生時^丨步驟708，可倍 在下-步驟710，將儲存的波开電流波形。 中，只有與潛在快逮暫態事件模組。在-實施命 藉由只傳絲在快逮暫態電子簽章 =會被儲存與傳送。 外，由於感應器控制器(例如VI探針枓流失減至最小。此 析模組可不需要包含用以分析#料並執行預處理，分 速暫態事件之動作程序的快速處理！t。Μ和判定對於晋在快 八4 ^下—步驟712，由分析模組執行簽章比斜。产-刀析板組可將潛在快速暫態電子簽盘在一貫施例中， 較。在-實施例中，彳將此組快速暫能龙立ϋ速暫態簽章相比 7實施例中’此資料庫亦可包含非快資料庫中。在 夠被執行。 . 簽早以使得關聯化能 在下一步驟714,判斷出潛在快速暫能㊉ 早比對的結果是沒有識別出符合者，曰;子簽羊之類別。若簽 ^會被分類域興趣之快速暫態電子簽⑯，速暫,¾、電子簽章 中，可摒棄此潛在快速暫態電子簽章。=V716)。在—實施例 潛在快速暫態電子簽章加人資料庫中，實施例中，可將此 章(步驟718)。 马新的快速暫態電子簽 然而，若簽章比對的結果是識別出快速新…干 了—步驟720，判定此快速暫態事件之3^日子簽早’則在 性可分佈為從低到高。在-實施例中在—例子中’嚴重 閾值範圍。在-實施例中，可將快速暫態電_^基於一組預定的 驟718)。步驟718為選擇性步驟^ :早加人資料庫中(步 小而用於偵測即時快速暫態 26 201115288 事件。 在下—步驟722,判定動作炉皮 可執行動作程序。在-實施例中^百：^已經判定嚴重性程度， 具有低嚴重性程度之快 ^動作程序。在-例子中， 在另-例子中，具有中子簽早可引發對操作者之通知。 警告。在又—例;=;;;之快速暫態電子簽章可弓丨發 引發基板處理之終止。重性程度之快速暫態電子簽章可 度以及與嚴重性程度騎之動作程序可由制者來配置嚴重性程 之方法的-實施造環境内之即時快速暫態事件 快速暫態事件，其中快球 1亦可使用此方法以偵測即時 份(在-實施例中)。在此類产产:悲、偵測演算法是分析模組的—部 取代VI探針控制哭，來勃中’可由分析模組(例如APECS314) 财，。在一實施 實施例中，分析模組與感應哭怏f處理計算模組。在-集並直接傳送到分析模組。接連接。所以，貢料由感應器收 之裝置和^ i。在了用，測原位即時快速暫態事件 板處理之後才執行快速暫態 :J件::在U完成-基板批次的基 測工具以判定快速㈣ 2。再者’可能需要複雜的量 與否是不可預期的，ΐ能必;，其ff快速暫態事件的存在 定可能已發生之潛在損傷。貝里測—基板批次中的每個基板以判 時的快測實施例提供了在基板處理期間即 傷減至最低。此外，、柯於先餘基板批:妹/祕理室之損 沒有人為干擾的自動化處理Iff ’此_處理為需要較少或 配置之條件轉铜值^而疋已較義可由使用# 有鏗於可在製造環境中用以自動偵測快速暫態事件。 件），可減少實際發生事件(例如微發孤事 遲在先别触中，此延遲可能耗費數小時或甚至數星期。 201115288 f而’，本鑛述之方法和/或裝置，可將此延遲減少為 笔秒，猎此減少所有權之總成本。 ㈣本伽已缝她佳實闕純_，其仍存在屬於本 化、變更和鱗物。雜在此提供了數種例子， &些例子係作為_之用，而_·定本發明。 解釋fi申ft，本案之名稱和摘要，其不應用以 方Γ提供於此’因此不應用以解釋或限制本 本”陳述於申請專利範财。若本文用到「組」 ί-此寻詞語意指其通常理解的數學意義，包含零、一、 發明之方法和設備。因砼，，νπΓΒΛ[ί t山 匕J力式木员現本 於本發明之直實附之申請專利範圍應以包含屬 下加神和料中的所有變化、變更和均等物的原則 【圖式簡單說明】 在隨附圖式中’本發明係以例示之 以說明，在圖式，同參照符號表示相同I元件，其式加 境的先讀術中具有主機層次分析祠服器之互連工具環 方炒Ξ2ίΓΐί群組工具層次解決方案之互連工具環境的η單 簡單i輯總@;本1月之員_中’製程層次疑難排解結構的 的簡^用展圖示;，在本發明之實施例中，製賴組層次分析词服器 圖5展示微發弧事件之簡單曲線圖； 圖6A和6B展示，在本發明夕每 塊圖； 科月之心例中’處理環境之簡單方 28 201115288 圖7展示，在本發明之實施例中，用以偵測在製造環境内之 即時快速暫態事件之方法的簡單流程圖，其中快速取樣暫態偵測 演算法不是分析模組的一部份。 • 【主要元件符號說明】 . 102晶圓廠主機 104群組工具控制器(CTC) 106群組工具控制器(CTC) 108群組工具控制器(CTC) 110製程模組控制器(PMC) 112製程模組控制器(PMC) 114製程模組控制器(PMC) 116製程模組控制器(PMC) .118感應器 118b計算模組 120感應器 122感應器 124感應器 126感應器 128感應器 130感應器 132感應器 134感應器 136感應器 138感應器 • 140感應器 142資料盒 144感應器纜線 146 路徑 148 路徑 29 201115288 150 路徑 156半導體設備通訊標準/通用設備模組(SECS/GEM) 158 路徑 202晶圓廠主機 204群組工具控制器(CTC) 206群組工具控制器(CTC) 208序列分接頭 210製程模組控制器(PMC) 212製程模組控制器(PMC) 214製程模組控制器(PMC) 216製程模組控制器(PMC) 218感應器 218b計算模組 220感應器 222感應器 224感應器 226感應器 228感應器 230感應器 532感應器 234感應器 236感應器 238感應器 -240感應器 242遠程控制器 -244感應器纜線 -246 路徑 248 路徑 250 路徑 252 路徑 30 201115288 254 路徑 256序列分接頭 258 路徑 260遠程控制器 — 302 晶圓廠主機 304群組工具控制器(CTC) 306製程模組控制器(PMC) 308 製程模組控制器(PMC) 310製程模組控制器(PMC) 312製程模組控制器(PMC) 314製程模組層次分析伺服器(APECS) 316感應器 318感應器 320感應器 322路徑 324 路徑 326 路徑 328感應器纜線 330感應器纜線 332感應器纜線 334 路徑 336 路徑 338 連結組 340 連結 • 400製程模組層次分析伺服器(APECS) - 402 處理器 404 處理器 406 處理器 408 處理器 410感應器 31 201115288 410b 計算模組 412 感應器 414 感應器 416 感應器 418 乙太網路交換器 420 感應器 422 感應器 424 感應器 426 感應器 428 共享記憶體主幹 430 路徑 432 路徑 434 路徑 436 路徑 438 路徑 440 路徑 442 路徑 444 路徑 446 路徑 448 路徑 450 路徑 452 路徑 454 路徑 456 數位輸出線 458 路徑 502 曲線 504 陡降 506 向原區 600 處理系統 602 處理室 32 201115288 604基板 606 —組RP產生器 608 —組匹配器 610 VI探針 * 612 VI探針控制器 . 614 路徑 616快速取樣暫態偵測演算法模組 618分析模組 650路徑 702開始基板處理 704偵測基板處理 704a感應器偵測電性參數 704b執行快速取樣暫態偵測演算法 706潛在快速暫態事件是否存在？ 708儲存電壓和電流波形之快照 710將波形傳送到分析模組 712執行簽章比對 714是否符合？ 716識別為不是感興趣之快速暫態電子簽章 718加入資料庫中 720判定嚴重性程度 722判定及執行動作程序 33At Shi ~ take · ^ sad detection algorithm can be sentenced _, 彳 。. The standard value of the 1 evil event (such as micro-arc) is used to deny the potential fast ^ = transient _ algorithm defined material meets (4) ' and the VI 楝 pin will continue to detect the base However, if a potential fast transient animal is identified, there is a potential fast transient event. Step 708 can be repeated. In the next step 710, the stored wave current waveform is stored. The transient event module. In the implementation of the life-by-pass only the wire in the fast-moving transient electronic signature = will be stored and transmitted. In addition, due to the sensor controller (such as the VI probe 枓 loss is minimized. The module may not need to include a fast processing of the action program for analyzing the #material and performing the pre-processing, the speed-dividing transient event! t. 判定 and the decision for the promotion in the fast 8 4 - step 712, the execution of the analysis module The chapter-tool skewing group can be used to apply the potential fast transient electronic sign in the consistent application. In the example, the group is compared with the fast-acting Longli idle transient signature. In the embodiment, 'this database can also be included in the non-fast database. It is enough to be executed. In the next step 714, it can be determined that the result of the potential fast temporary energy comparison is that the match is not identified, and the child is signed. If the sign is used, the fast transient electronic sign of the domain interest is classified. 16, in the temporary, 3⁄4, electronic signature, can abandon this potential fast transient electronic signature. = V716). In the embodiment of the potential fast transient electronic signature plus database, in the embodiment, This chapter (step 718). Ma Xin's fast transient electronic signature, however, if the result of the signature comparison is to identify a fast new ... done - step 720, determine the fast transient event 3 ^ days sign early ' The nature may be distributed from low to high. In the embodiment - in the example - a critical threshold range. In an embodiment, the fast transients may be based on a predetermined set of steps 718. Step 718 is Selective step ^: early in the database (small step is used to detect the instant fast transient 26 201115288 event. In the next step 722, determine the action furnace can perform the action program. In the embodiment - ^ hundred: ^ The degree of severity has been determined, and the procedure has a low degree of severity. In the example, in another example, having a neutron signature can trigger a notification to the operator. Warning. In the case of the fast transient electronic signature, the fast transient electronic signature can trigger the substrate processing. Termination. The degree of severity of the fast transient electronic signature and the severity of the riding procedure can be configured by the system to implement the severity process - the implementation of real-time fast transient events in the environment, rapid transient events, The fastball 1 can also use this method to detect the instant share (in the embodiment). In this kind of production: the sorrow, the detection algorithm is the analysis module - the part replaces the VI probe to control the crying The 'can be analyzed by the analysis module (for example, APECS 314). In an embodiment, the analysis module and the induction crying processing module are processed. In-set and directly transmitted to the analysis module. Connected. Therefore, the tribute is received by the sensor and ^ i. In the case of the in-situ real-time fast transient event board processing, the fast transient is performed: J:: Finishing the sub-substrate batch test tool in U to determine the fast (4) 2 . Furthermore, it may be that the amount of complexity is unpredictable, and that the existence of the ff fast transient event may be a potential damage that may have occurred. Each of the substrates in the Berry-to-substrate batch was provided with a fast-measurement example of the time to minimize damage during substrate processing. In addition, Ke Yuxian's remaining substrate batch: the sister/secret room damage has no human intervention for the automatic processing of the Iff 'this _ processing for the need for less or configured conditions to transfer the copper value ^ and 疋 has been more than can be used #有铿It can be used in the manufacturing environment to automatically detect fast transient events. ()), can reduce the actual occurrence of events (such as micro-orphans, delays may be missed, this delay may take hours or even weeks. 201115288 f and ', the method and / or device described in this mine, you can The delay is reduced to pen seconds, and this reduces the total cost of ownership. (4) Benga has sewed her good 阙 pure _, which still exists in the localization, change and scale. Miscellaneous provides several examples, & The examples are used as _, and _· 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 If this article uses "group" ί - this vocabulary means the mathematical meaning of its usual understanding, including zero, one, invention methods and equipment. Because 砼,, νπΓΒΛ[ί t山匕J力式木员 is present The scope of the patent application of the present invention is intended to cover all changes, modifications and equivalents of the subordinates and materials. [Simplified description of the drawings] In the accompanying drawings, the invention is exemplified Description, in the schema, the same reference symbol I component, its pre-reading has a host-level analysis, and the interconnect tool of the device is fired. 2ίΓΐί Group tool level solution interconnection tool environment η单简单i-i total@@本本In the embodiment of the present invention, the hierarchical analysis of the word processor is shown in Figure 5, which shows a simple graph of the micro-arc event; Figure 6A and 6B shows, in the present invention, each block diagram; in the heart of the month, the simple process of the processing environment 28 201115288 Figure 7 shows, in the embodiment of the invention, used to detect the instant fast transient in the manufacturing environment A simple flow chart of the method of event, in which the fast sampling transient detection algorithm is not part of the analysis module. • [Key component symbol description] . 102 fab host 104 group tool controller (CTC) 106 group Group Tool Controller (CTC) 108 Group Tool Controller (CTC) 110 Process Module Controller (PMC) 112 Process Module Controller (PMC) 114 Process Module Controller (PMC) 116 Process Module Controller ( PMC) .118 sensor 118b computing module 120 sensor 122 sensing 124 sensor 126 sensor 128 sensor 130 sensor 132 sensor 134 sensor 136 sensor 138 sensor • 140 sensor 142 data box 144 sensor cable 146 path 148 path 29 201115288 150 path 156 semiconductor device communication standard / General Equipment Module (SECS/GEM) 158 Path 202 Fab Host 204 Group Tool Controller (CTC) 206 Group Tool Controller (CTC) 208 Sequence Tap 210 Process Module Controller (PMC) 212 Process Mode Group Controller (PMC) 214 Process Module Controller (PMC) 216 Process Module Controller (PMC) 218 Sensor 218b Calculation Module 220 Sensor 222 Sensor 224 Sensor 226 Sensor 228 Sensor 230 Sensor 532 Sensor 234 Sensor 236 Sensor 238 Sensor - 240 Sensor 242 Remote Controller - 244 Sensor Cable - 246 Path 248 Path 250 Path 252 Path 30 201115288 254 Path 256 Sequence Tap 258 Path 260 Remote Controller - 302 Fab Host 304 Group Tool Controller (CTC) 306 Process Module Controller (PMC) 308 Process Module Controller (PMC) 310 Process Module Controller (PMC) 312 Process Module Controller (PMC) 314 Process Module Analytic Server (APECS) 316 Sensor 318 Sensor 320 Sensor 322 Path 324 Path 326 Path 328 Sensor Cable 330 Sensor Cable 332 Sensor Cable 334 Path 336 Path 338 Link Group 340 Link • 400 Process Module Analytic Server (APECS) - 402 Processor 404 Processor 406 Processor 408 Processor 410 Sensor 31 201115288 410b Computing Module 412 Sensor 414 Sensor 416 Sensor 418 Ethernet Switch 420 Sensor 422 Sensor 424 Sensor 426 Sensor 428 Shared Memory Trunk 430 Path 432 Path 434 Path 436 Path 438 Path 440 Path 442 Path 444 Path 446 Path 448 Path 450 Path 452 Path 454 Path 456 Digital Output Line 458 Path 502 Curve 504 steep drop 506 to the original zone 600 processing system 602 processing room 32 201115288 604 substrate 606 - group RP generator 608 - group matcher 610 VI probe * 612 VI probe controller. 614 path 616 fast sampling transient detection algorithm Module 618 analysis module 650 path 702 begins substrate processing 704 detection The substrate processing 704a sensors to detect electrical parameters 704b perform rapid sampling transient detection algorithm 706 potential fast transient event exists? 708 Storing a snapshot of the voltage and current waveforms 710 Transferring the waveform to the analysis module 712 Is the signature comparison 714 compliant? 716 identifies a fast transient electronic signature that is not of interest 718 joins the database 720 determines the severity level 722 determines and executes the action program 33