201009970 九、發明說明： 【發明所屬之技術領域】 本發明是關於一種洩漏偵測方法’且特別是關於一種 製程反應室的洩漏偵測方法，其能夠在使用光學發射光譜 儀（optical emission spectrometry，OfiS)方法或智能 HMS(smart HMS)來偵測專用製程反應室的洩漏時準確並 迅速地偵測是否發生洩漏。 【先前技術】 使用電漿之半導體製程可大致分爲乾式蝕刻製程（dry etching processes)、化學氣相沈積製程（chemical vapor deposition processes )以及濺鍍製程（sputtering processes ) 〇 使用電漿之半導體製造裝置分爲：藉由依據產生電漿 之電極的構造在彼此面對的平行板之間施加射頻（radi〇 frequency，RF)功率來産生電漿的電容輕合電裝 (capacitively coupled plasma，CCP)型裝置、藉由施加 RF功率至反應管外部之線圈來産生電製的電感^合電聚 ❹ （inductively coupledPlasma，ICP)型裴置 功率和磁場來產生電衆的磁強化反應 =: (magneucally enhanced reactive ion etching，施·)型 裝置以及藉由輕合微波和磁場來產生電聚的電子迴旋共振 (electron cyclotron resonance，ECR)型震置。 '、 在使用職之乾絲職財，反 室二 ==_之封閉空= 力羊至上電極和下電極，上電極和下電極安裝成 反應 施加 201009970 在製程反應室内彼此間隔預定距離，藉此形成電場。 反應氣體被電場激活以變成電漿。處於電漿態的離子 與位於下電極之晶圓上的薄膜反應，使得晶圓上的薄膜被 蝕刻成所希望的形狀。 在這種情況下，使用電漿之半導體製造裝置需要實時 監測並控制電漿製程，使得能夠以所希望的方式進行電漿 製程。 也就是，在電漿蝕刻製程及電漿CVD製程中會産生 大量的反躺錄，賊後與反應氣體或絲（phQtc)aresist) 反應’藉此産生聚合體（p〇lymer)。由於聚合體附著於玻 墻表面或製程反應f内壁’可能使製程參數發生變化並產 生微粒。因此’在半導體製程朗，這些情況成為玻璃的 缺限因素（defective factor)，從而導致産量減少。 因此，爲了減少這些缺限因素，要以預定時間重複地 進行製程反應t的預防性轉（pfeventive maintenanee， PM)。 此時，在製程反應室内沈積預定量的薄膜後，進行類 似於侧餘之乾式清洗製程1餘製歡應室内出現 的聚合體。在乾式清洗製程中，需藉由終點_器（end point detector)來债測餘刻終點（etcMng - p〇int)，使 得，程反應室贼㈣部件如損（寧捕）狀態、反 應氣體的消耗以及生產力得以_ # 、然而、’當在製程反應室中進行pM前後用 0ES分析電 漿並且透過分析·_時，_的精度下降，並且難以 201009970 偵測清洗製程的蝕刻終點。 發生ίΐ:由程反應室進行PM，部件狀態 即使未發伽，電漿也與進行m 的^誤。…在補情況下，可能出現谓測到存錢漏 程，娜地進行乾式清洗製 Γ=?圍(specrange)應大於此分佈。因:ί 由OES偵測不到細微洩漏ό 曰 室之H殘it運行停止（mn d〇wn)時，製程反應 在運餘讀量相較於進行連續運行時發生變化。 情、兄進行處理之玻璃的電聚分佈相較於一般 漏偵測的規格範圍。 月，吧徑制洩 【發明内容】 ❹ 麵3 =於一種製程反應室的細測方法，其考 ΐ，23 部件狀態在進行™前後發生明顯變 ，藉由在進行PM後以預定周期或時 影響的茂漏_之頻格銘图“㈣置對Λ漏有 在無錯誤的产、况下嘴放目i電衆穩&爲止’而能夠 生洩漏。、隋况下準確且迅速地偵測製程反應室内是否發 的一方面’提供了 一種製程反應室的魏 以使车ί雜ϊ在以預定時間重複地進行預防性維護 +¥體製程期間的製程反應室的缺陷因素降低時，部 7 201009970 件狀態發生明顯變化的情況下，即使未發生洩漏也偵測到 洩漏之錯誤。此洩漏偵測方法包括：第一步驟，設置警報 規格、故障規格及預防性維護規格，該些警報規格、故障 規格及預防性維護規格，在進行預防性維護之前的半導體 製程期間，基雜意設置之雖錢基礎上擴大至預定範 圍，第二步驟’基於第一步驟中設置的警報規格、故障規 格及預防性維護規格的範圍來判斷是否發生洩漏；第三步 驟，在進行預防性維護後的半導體製程期間，在自動重設 的特徵值的基礎上，將警報規格及故障規格重設至預定範 圍；以及第四步驟，基於第三步射設置之警報規格及故 障規格的範圍判斷是否發生浪漏。 第步驟中任思设置之特徵值是藉由以下方式獲得. 在半導體製程之系嶋作的初_段，狀之若干取樣玻 璃的電聚光發射中搜尋對〶漏有影響的有效波段，並且夢 由積分波段中的强度來計算面積或計算峰值强度。 曰201009970 IX. Description of the Invention: [Technical Field] The present invention relates to a leak detection method, and more particularly to a method for leak detection of a process chamber, which is capable of using an optical emission spectrometer (OfiS) A method or smart HMS (smart HMS) to detect leaks in a dedicated process chamber accurately and quickly. [Prior Art] A semiconductor process using a plasma can be roughly classified into a dry etching process, a chemical vapor deposition process, and a sputtering process. A semiconductor manufacturing apparatus using a plasma Divided into: Capacitively coupled plasma (CCP) type which generates plasma by applying radio frequency (RF) power between parallel plates facing each other according to the structure of the electrode for generating plasma. The device generates an inductively coupled Plasma (ICP) type power and a magnetic field by applying RF power to a coil external to the reaction tube to generate a magnetic strengthening reaction of the electric group = (magneucally enhanced reactive Ion etching, a type of device, and an electron cyclotron resonance (ECR) type shock generated by lightly combining microwaves and magnetic fields to generate electropolymerization. ', in the use of dry wire business, anti-room two == _ closed air = force sheep to the upper and lower electrodes, the upper electrode and the lower electrode are installed in response to the application of 201009970 in the process chamber within a predetermined distance from each other, thereby An electric field is formed. The reaction gas is activated by an electric field to become a plasma. The ions in the plasma state react with the film on the wafer on the lower electrode, so that the film on the wafer is etched into a desired shape. In this case, a semiconductor manufacturing apparatus using plasma requires real-time monitoring and control of the plasma process so that the plasma process can be performed in a desired manner. That is, a large amount of anti-rear is generated in the plasma etching process and the plasma CVD process, and the thief reacts with the reaction gas or the phQtc aresist to generate a polymer (p〇lymer). Since the polymer adheres to the surface of the glass wall or the inner wall of the process reaction f may change the process parameters and produce particles. Therefore, in the semiconductor manufacturing process, these conditions become the defective factor of the glass, resulting in a decrease in production. Therefore, in order to reduce these disadvantages, the preventive transfer of the process reaction t (pfeventive maintenanee, PM) is repeated at predetermined times. At this time, after depositing a predetermined amount of the film in the process chamber, a polymer similar to that in the dry cleaning process of the side is performed. In the dry cleaning process, the end point detector is used to measure the residual end point (etcMng - p〇int), so that the process chamber thief (4) parts are damaged (Ning catch) state, reactive gas Consumption and productivity can be _#, however, 'When the plasma is analyzed by 0ES before and after pM in the process chamber, the accuracy of _ decreases, and it is difficult to detect the etching end of the cleaning process at 201009970. Occurs: PM is carried out from the reaction chamber, and the state of the component is even if the gamma is not emitted. ... In the case of supplements, there may be a measure of the deposit and loss, and the dry cleaning system should be larger than this distribution. Because: ί When OES does not detect a small leak ό H When the H-residue of the chamber stops running (mn d〇wn), the process response changes when the remaining reading is compared to the continuous operation. The distribution of the electro-polymerization of the glass treated by the love and the brother is compared to the specification range of the general leak detection. Month, the diameter of the pipe venting [invention] ❹ face 3 = in a process of the process of the fine measurement method, the test, 23 component state before and after the TM change significantly, by the PM after a predetermined period or time The impact of the leak _ the frequency of the map "(4) set the right leak in the production of no error, the mouth is placed in the eyes of the electricity and stable; and can be a leak., accurately and quickly detect In the process of measuring whether the reaction chamber is in the process of providing a process chamber, the Wei process is used to make the preventive maintenance of the process chamber during the predetermined period of time. 7 201009970 In the case of a significant change in the state of the device, a leak error is detected even if no leak occurs. This leak detection method includes the first step of setting alarm specifications, fault specifications and preventive maintenance specifications. , fault specifications and preventive maintenance specifications, during the semiconductor manufacturing process before preventive maintenance, the base of the miscellaneous settings is expanded to the predetermined range, the second step 'based on the first step The range of alarm specifications, fault specifications, and preventive maintenance specifications set to determine whether a leak occurs. The third step is to set the alarm specification based on the characteristic value of the automatic reset during the semiconductor process after preventive maintenance. And the fault specification is reset to a predetermined range; and the fourth step is to determine whether a leak occurs based on the range of the alarm specification and the fault specification set by the third step. The characteristic value of the setting in the first step is obtained by the following method. In the initial stage of the semiconductor process, the electric concentrating emission of several sampled glasses is searched for an effective band that affects the leakage, and the dream is calculated from the intensity in the integrated band to calculate the area or calculate the peak intensity.
第步驟中《X置的警報規格在第一步驟中設置之 值的基礎上設置於±5-7%的範圍内。 、 第-步驟中設置的故障規格在第一步驟中設置 值的基礎上設置於12-15%的範圍内。 魏絲餘獅轉護的起點 L點並持續叙週期或時間，該預定週期或時間從 程之半導體製造系統的操作持續直至㈣穩定。， 預防性維護規格在第一步驟中設置之特徵值 設置於±15-20%的範圍内。 勺基礎上 201009970 、第…步驟中自動重叹的特徵值，可藉由在由穩定電衆 來進订半導體製程之系統操料，引人之若干取樣玻璃的 電漿的光發射進行自動分析及平均獲得。 第二步驟中重設的警報規格在重設特徵值的基礎上設 置於±5-7%的範圍内。 第三步驟中重設的故障規格在重設特徵值的基礎上設 置於±12-15%的範圍内。 第二和第四步驟包括：當在半導體製程正在進行的狀 態下，於預定時間内未輸入資料時，認為運行停止狀態， 當在認為運行停止的狀態下輸人資料時，忽略警報規格， 且僅當變化偏離故障規格的範圍時，判斷發生洩漏並產 警報。 為讓本發明之上述和其他目的，特徵和優點能更明顯 易懂，下文特舉較佳實施例，並配合所附圖式詳細說明如 下。 【實施方式】 下文中’將參照附圖來描述本發明的範例性實施例。 圖1是顯示根據本發明實施例之製程反應室内的浪漏 偵測型態的示意圖。 參照圖1’當以預定時間重複地進行預防性維護 (preventive maintenance ’ PM)使得半導體製程期間的製 程反應室的缺陷因素減少時，部件狀態發生明顯變化。在 這種情況下’即使未發生洩漏，也可能偵測到洩漏。進行 根據本發明實施例之製程反應室的洩漏偵測方法可避免這 201009970 種偵測錯誤，此方法包括第一至第四步驟。 第-步驟包括在系統操作的初始階段所引入之若干取 樣玻璃之電漿光發射巾，搜尋對、朗有影響的有效波段 (wavelength band) ’且藉由積分波段中的强度來計算面積 或》十^*峰值强度，並藉此操取及設置特徵值(啦如vahe) 以及在進行PM之前的半導體製程中，設置馨報規格 gaming spec)、故障規格(fauit spec)&pM規格的過程， φ 这些疋在设置之特徵值的基礎上擴大至預定範圍。 當僅擷取一個特徵值時，在確定洩漏時可能出現大量 錯誤，並且準確性及可靠性降低。因此，引入若干取樣玻 璃，並且從，樣玻璃中搜尋對洩漏有影響的若干有效波 段。然後，藉由透過積分波段中的强度來計算面積或計算 峰值强度而擷取若干特徵值。在此若干特徵值中，'任一特 徵值是任意没置的。可依據製程反應室來改變設置標準。 此時，警報規格可在設置之特徵值的基礎上設置於 ±5_7%的範_ ’而輯規格可在設置之特徵i的基礎上 β 設置於±12-15%的範圍内。 ΡΜ規格應用於ΡΜ的起點及終點並應用預定週期（例 如’當引入第300 #玻璃時的時間點），此預定週期從新 製程之半導體製造系統操作開始直到電漿穩定或 的時間週期（例如，兩天）。ΡΜ規格可在設置之特徵值 的基礎上設置於±15-20%的範圍内。 ' 也就是，當假定在ΡΜ期間部件狀態發生變化，使得 電漿的變化接近±15%，並且如果洩漏導致±5%的變化時， 201009970 總變化爲±20%。因此，當在一般情況下將pM規格設置在 ±5%的範圍内來仙漏時’設置之±5%的規格範圍包括 在總改變的±2〇%範_。目此，#進行製程反應室的 時’出現偵測錯誤，其中即使未發生茂漏，但由於超&規 袼而偵測到發生洩漏。 因此，考慮到進行PM後部件狀態發生變化，由於冬 進行PM時未應料報規格及故障規格，pM規格範圍^ _ 特徵值的基礎上擴大至±15-20%。 第二步驟包括基於在第一步驟中設置之警報規格、故 障規格及PM規格的範圍來判斷是否發生线漏的過程。 也就是，在進行PM之前對製程反應室的電漿進行先 發射刀析。僅备分析之電漿光發射的變化在設置之特徵值 的基礎上偏離警報規格的範圍（土5_7%)及故障規格的範 圍（土12-15%)時，由於超出規格，洩漏偵測器判斷製程 反應室内發生洩漏。 進步’當分析之電漿光發射的變化在設置之特徵值 =基礎上未偏離警報規格的範圍（±5-7°/❶）及故障規格的 範圍（±12-15%)時，洩漏偵測器判斷製程反應室内 生洩漏。 此時，在第二步驟中，當在半導體製程正在進行的狀 恕下=預定時間未輸入資料時，泡漏偵測器認為其運行停 止。當在認為運行停止的狀態下輸入資料時，洩漏偵測器 忽略±5-7%的警報規格’並且僅當變化偏離輯規格的範 圍（±12-15%)時產生警報。 11 201009970 铪产疋’备發生運行停止時，製程反應室的溫度及殘 、氣’相較於進行連續運行時發生變化，並且在運行 停^後立即進行處理之玻璃的電漿分佈相較於一般情況略 f增加。因此，考慮到出現由分料致的朗侧錯誤， ΐ忽ί±5_7%的警報規格的狀態下’僅當變化偏離故障規 ，耗圍J±12_15%)時’⑽仙n才判斷發生茂漏。 接著’半導體、統控鮮元在誠制器判 已發生洩漏後産生警報。 f二步驟包括在進行PM之後的半導體製程期間，自 動重設之特徵㈣基礎上重設警報規格及輯規格的過 程0 、也就是，當在進行PM之後，經過電聚的穩定後進行 半導體製程系狀操作時，自動地分析若干狀取樣玻璃 之電樂的光魏，且紐進行平均賴取並設置特徵值。 進-步’將警報規格及故料格在設置之紐值的基 擴大至預定範圍。 此時，警報規格可在重設之特徵值的基礎上設置於 ±5-7%的範_ ’並且故障規格可在重設之特徵值的基礎 上設置於±12-15%的範圍内。 第四步驟包括基於在第三步驟中設置的警報規格及故 障規格的範圍判斷是否發生洩漏的過程。 也就是，當進行PM之後在電漿穩定的狀態下進行 製程時，對製程反應室之電衆進行光發射分析。當分析之 電襞光發射中的變化在重設之特徵值的基礎上偏離警報規 12 201009970 格的範圍（±5·7%)及故障規格的範圍（土 12_15%)時，由 於超出規格，洩漏偵測器判斷製程反應室内發生洩漏。 進一步’當分析之電漿光發射的變化在重設之特徵值 的基礎上沒有偏離警報規格範圍（土5_7%)及故障規格範 圍（±12-15%)時，洩漏偵測器判斷製程反應室内未 浪漏。 此時，在第四步驟中，當進行PM之後在半導體製程 _ 正在進行的狀態下，於預定的時間内未輸入資料時，洩漏 ，測=認為其運行停止。當在認為運行停止的狀態下輸入 資料時，洩漏偵測器忽略±5_7%的警報規格，並僅當變化 偏離故障規格的範圍（±12-15%)時産生警報。 严也就是，當發生運行停止時，製程反應室的溫度及殘 餘氣體相較於連續地運行時發生變化，並且在運行停止後 立即進行處理之玻璃的電黎分佈相較於一般情況略微地增 加。因此慮到出現由分佈導致的$麵測錯誤，在忽 =±5_7%的警報規格的狀態下，僅當變化偏離故障規格的 範圍（±12-15%)時’浪漏偵測器才判斷發生沒漏。接著， 半V體製造系統的系統控制單元，在茂漏债測器判斷發生 洩漏後產生警報。 在本示範性實施例中’於PM前後設置任意特徵值之 後，在設置之特徵值的基礎上擴大用以浪漏偵測之警報規 格、故障規格及PM規格的範圍，並且考慮pM之後的部 件變化來避免即使未發生茂漏也偵測_漏的錯誤。 雖然已顯示和描述本發明的幾個範例性實施例，但本 13 201009970 領域熟知此項技藝者顯而易見在d 圍及其等同物定義之本發明的範=背離由所附申請專利範 這些實施例進行各種修改。因^與精神的情況下，可對 附之申請專利範®所界定者為準。發明之保護範圍當視後 根據本發明，考慮到進扞ΡλΛ & 改#，仃1^^碉後部件狀態發生顯著 ’ ^仃PM後將對茂漏有影響之茂漏制的警報規In the first step, the alarm specification of X is set in the range of ±5-7% based on the value set in the first step. The fault specifications set in the first step are set in the range of 12-15% based on the set values in the first step. The starting point of Weiss Yushi transfer is L point and continues the cycle or time, which continues from the operation of the semiconductor manufacturing system until (4) stability. The preventive maintenance specifications are set in the range of ±15-20% in the first step. Based on the spoon, the feature values that are automatically re-sighed in the steps of 201009970, can be automatically analyzed by the light emission of the plasma of some sampled glass by the system of ordering the semiconductor process by the stable electricity. Obtained on average. The alarm specifications reset in the second step are set to within ±5-7% based on the reset feature values. The fault specification reset in the third step is set within the range of ±12-15% based on the reset eigenvalue. The second and fourth steps include: when the semiconductor process is in progress, when the data is not input within the predetermined time, the operation is stopped, and when the data is input while the operation is stopped, the alarm specification is ignored, and Only when the change deviates from the scope of the fault specification, it is judged that a leak has occurred and an alarm is generated. The above and other objects, features and advantages of the present invention will become more <RTIgt; [Embodiment] Hereinafter, exemplary embodiments of the present invention will be described with reference to the drawings. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a schematic view showing a leak detection pattern in a process chamber according to an embodiment of the present invention. Referring to Fig. 1', when the preventive maintenance (PM) is repeatedly performed for a predetermined time so that the defect factor of the process chamber during the semiconductor process is reduced, the state of the component changes significantly. In this case, a leak may be detected even if no leak has occurred. The leak detection method of the process chamber according to the embodiment of the present invention can avoid the 201009970 detection error, and the method includes the first to fourth steps. The first step consists of a number of sampled glass plasma light-emitting wipers introduced during the initial phase of system operation, searching for a pair of influential effective wavelength bands 'and calculating the area by the intensity in the integrated band or Ten ^ * peak intensity, and by this operation and set feature values (such as vahe) and in the semiconductor process before the PM, set the gamification spec), fault specification (fauit spec) & pM specification process , φ These 扩大 are expanded to a predetermined range based on the set feature values. When only one feature value is taken, a large number of errors may occur in determining the leak, and accuracy and reliability are lowered. Therefore, a number of sampled glass are introduced and a number of effective bands that have an effect on the leak are searched from the sample glass. Then, several eigenvalues are extracted by calculating the area by calculating the intensity in the integrated band or calculating the peak intensity. Of the several feature values, 'any trait value is arbitrarily set. The setting criteria can be changed according to the process chamber. At this time, the alarm specification can be set to ±5_7% of the range _' based on the set characteristic value, and the size specification can be set within the range of ±12-15% based on the set feature i. The ΡΜ specification is applied to the start and end points of the crucible and applies a predetermined period (for example, 'the point in time when the 300th glass is introduced') from the operation of the semiconductor manufacturing system of the new process until the plasma is stable or the time period (for example, Two days). The ΡΜ specification can be set within ±15-20% based on the set eigenvalues. That is, when it is assumed that the state of the component changes during the enthalpy, the change in the plasma is close to ±15%, and if the leakage causes a variation of ±5%, the total change of 201009970 is ±20%. Therefore, when the pM specification is set in the range of ±5% under normal conditions, the specification range of ±5% of the setting is included in the ±2〇% of the total change. Therefore, a detection error occurred when #processing the reaction chamber, and even if no leak occurred, a leak was detected due to the super & Therefore, considering the change in the state of the component after the PM is performed, the specification and the fault specification of the PM are not reported when the PM is performed in winter, and the pM specification range is expanded to ±15-20% based on the eigenvalue. The second step includes determining whether or not a line leak occurs based on the range of the alarm specification, the fault specification, and the PM specification set in the first step. That is, the plasma of the process chamber is first subjected to a pre-emission analysis before the PM is performed. The analysis of the plasma light emission of the analysis only deviates from the range of the alarm specification (5_7% of the soil) and the range of the fault specification (12-15% of the soil) based on the set characteristic value, and the leak detector is out of specification. Determine the leak in the process chamber. Progressive' When the analysis of the plasma light emission changes is based on the set characteristic value = not deviating from the alarm specification range (±5-7°/❶) and the fault specification range (±12-15%), the leak detection The detector judges that the process reaction is leaking indoors. At this time, in the second step, when no data is input for a predetermined time under the progress of the semiconductor process, the bubble detector considers that its operation is stopped. When data is input while the operation is stopped, the leak detector ignores the alarm specification of ±5-7% and generates an alarm only when the variation deviates from the range of the specification (±12-15%). 11 201009970 铪 疋 ' When the operation stops, the temperature and residual gas in the process chamber change compared with the continuous operation, and the plasma distribution of the glass treated immediately after the stop is compared with The general situation is slightly increased by f. Therefore, in consideration of the occurrence of a remote side error caused by the material distribution, the state of the alarm specification of ±±5_7% is 'only when the deviation deviates from the fault gauge, and the cost is around J±12_15%.' leak. Then, 'Semiconductor and Regulatory Controls produced an alarm after the leak was made. f The second step includes the process of resetting the alarm specification and the series specification based on the feature of the automatic reset during the semiconductor process after the PM is performed, that is, after the PM is performed, the semiconductor process is stabilized after the electropolymerization is stabilized. In the case of the system operation, the light of the electric music of several sampling glasses is automatically analyzed, and the nucleus is averaged and the characteristic value is set. Step-by-step expands the alarm specification and the base of the set value to the predetermined range. At this time, the alarm specification can be set to ±5-7% of the range _ on the basis of the reset characteristic value and the failure specification can be set within the range of ±12-15% based on the reset characteristic value. The fourth step includes a process of judging whether or not a leak has occurred based on the range of the alarm specification and the fault specification set in the third step. That is, when the process is performed in the state where the plasma is stable after the PM is performed, the light emission analysis is performed on the electric power of the process chamber. When the change in the analyzed electro-optical emission is deviated from the range of the alarm gauge 12 201009970 (±5·7%) and the range of the fault specification (12_15% of the soil) based on the reset characteristic value, due to exceeding the specification, The leak detector determines that a leak has occurred in the process chamber. Further, when the change of the plasma light emission of the analysis does not deviate from the alarm specification range (soil 5_7%) and the fault specification range (±12-15%) based on the reset characteristic value, the leak detector judges the process response. There is no leakage in the room. At this time, in the fourth step, when the semiconductor process is in progress in the semiconductor process _ in progress, the data is not input for a predetermined time, and the measurement is considered to be stopped. When inputting data while the operation is stopped, the leak detector ignores the alarm specification of ±5_7% and generates an alarm only when the variation deviates from the range of the fault specification (±12-15%). Strictly, when the operation stops, the temperature of the process chamber and the residual gas change compared to the continuous operation, and the distribution of the glass of the glass processed immediately after the stop of the operation is slightly increased compared with the general case. . Therefore, considering the occurrence of the $face measurement error caused by the distribution, in the state of the alarm specification of ±5_7%, the leakage detector only judges when the variation deviates from the range of the fault specification (±12-15%). No leakage occurred. Next, the system control unit of the semi-V body manufacturing system generates an alarm after the leak detector determines that a leak has occurred. In the present exemplary embodiment, after setting any characteristic value before and after the PM, the range of the alarm specification, the fault specification, and the PM specification for the leak detection is expanded based on the set feature value, and the components after the pM are considered. Change to avoid detecting _ leak errors even if no leaks occur. While a few exemplary embodiments of the present invention have been shown and described, it is apparent that those skilled in the art in the field of the present invention will be apparent to those skilled in the art. Make various modifications. In the case of the spirit and the spirit, it may be subject to the definition of the patent application. According to the present invention, in consideration of the invention, the state of the component after the 捍Ρ ^ ^ amp 仃 仃 仃 ^ ^ ^ ^ ^ ^ ^ ’ ’ ’ ’ ’ 仃 仃 仃 仃 仃 仃 仃 仃 仃 警报 警报 警报 警报 警报 警报 警报 警报 警报 警报 警报
。、故障規格以及⑽規格之的範B擴大預定週期或時間 直至電漿穩定爲止。目此’可轉確並迅速地判斷製程反 應室内是否發生洩漏並偵測細微洩漏。 【圖式簡單說明】 圖1顯示根據本發明實施例的製程反應室内洩漏偵測 型態的示意圖。 【主要元件符號說明】 無. The fault specification and the specification of the (10) specification expand the predetermined period or time until the plasma is stable. This can be used to accurately and quickly determine if there is a leak in the process chamber and detect a small leak. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic view showing a leak detection pattern in a process reaction chamber according to an embodiment of the present invention. [Main component symbol description] None
1414