200421411 (1) 玖、發明說明 【發明所屬之技術領域】 本發明是關於’蝕刻處理裝置等所使用的電漿處理方 法’及陳化處理終了檢測方法以及電漿處理裝置。 【先前技術】 舉例言之’蝕刻處理裝置等處理裝置具備有：氣密構 造的處理容器；及配設在此容器內，用以保持被處理體的 保持體’在處理容器內產生電漿，對被處理體施加規定的 處理。而繼續進行被處理體的處理時，處理容器內會被副 生成物等污染，或使內部零件消耗。因此，需要暫停處理 裝置’進行處理容器內的清潔或更換消耗品等之保養工作 。而在結束保養工作後再度起動處理裝置。 例如蝕刻處理裝置，在再起動時，需向處理容器內供 應規定片數的仿真晶圓（d u m m y w a f e r)而重複蝕刻循環， 將處理容器內調整到生產時被要求的狀態，所謂陳化處理 。完成陳化處理後，要檢查蝕刻率或晶圓面內的蝕刻均一 性等。在陳化處理時，使用由多枚仿真晶圓獲得的測量資 料’例如從終點檢測器獲得的發光光譜的測量資料，進行 資料解析。而觀察此解析資料的變化，判斷陳化處理是否 終了。 然而，傳統的解析資料，很難看出作爲陳化處理終了 的判斷基準的變化，究竟是陳化處理引起的變化，亦即依 據處理容器內的狀態變化的變化，或依據各仿真晶圓間的 -6 - (2) (2)200421411 溫度變化的變化。因此，存在有很難判斷陳化處理是否終 了的課題。 亦即，本發明人等曾進行’例如使用多變量解析的一 種的主成分分析，如後述對測量資料作資料解析，在此解 析結果看出有表示變化的兩個大峰値，很難判斷陳化處理 是否已結束。以下說明此主成分分析。這時是藉由與傳統 同樣的手法採取測量資料。例如，第1天供應1 3 0枚仿真 晶圓進行蝕刻，第2天進入生產製程而供應3 0枚仿真晶 圓進行蝕刻。使用第1天的第5 1〜60枚的仿真晶圓及第 1 2 1〜1 3 0枚的仿真晶圓分別獲得的發光光譜的測量資料 進行主成分分析，獲得第8A圖、第8B圖、及第9A圖、 第9B圖所示的解析結果。在主成分分析，是使用發光光 譜中，在1 93 nm〜4 1 9 nm的短波長領域的297種波長對 一片仿真晶圓在一分鐘內按每3秒測量各波長強度1 8次 ，而依據此等測量資料進行主成分分析。而分別求出各測 量時的主成分得分及殘差，描繪HOTELLINGS TSQUARE( 主成分的平方和）的結果是第8A圖、第9A圖，描繪殘差 的平方和（殘差得分）的結果是第8B圖、第9B圖。從此等 解析結果也可以明白，任一曲線圖均在第1天及第2天的 資料有很大的峰値，要判斷陳化處理終了很難。再者，各 曲線圖的橫軸表示測量次數。 本發明是爲了解決上述課題而完成者，其目的在提供 ，能夠明確判斷陳化處理終了的電漿處理方法，及陳化處 理終了的檢測方法以及電漿處理裝置。 -7- (3) (3)200421411 本發明人等對辨認出有兩個峰値的原因作種種檢討的 結果，發現原因是在資料解析用的測量資料的採取方法。 因此獲得，藉由在採取資料時對處理容器做特定的處置， 排除仿真晶圓間的溫度變化的影響，確實掌握陳化處理所 形成的變化，便能夠確實判斷陳化處理終了的結論。 【發明內容】 本發明是根據上述結論而完成者，本發明的申請專利 範圍第1項所述之電漿處理方法是，將試驗用被處理體供 給處理裝置的處理容器內，以進行陳化處理時，用以檢測 上述陳化處理終了的方法，其特徵爲具備有：向上述處理 容器內供應上述試驗用被處理體，冷卻上述處理容器內後 ，再度向上述處理容器內供應複數個上述試驗用被處理體 時，使用所獲得的複數個測量資料進行多變量解析，作成 用以預測上述陳化處理終了的預測式的製程；及依據上述 預測式，檢測進行上述陳化處理時的陳化處理終了的製程 〇 本發明的申請專利範圍第2項所述之電漿處理方法是 ，如申請專利範圍第1項所述之發明，其中上述多變量解 析使用主成分分析。 本發明的申請專利範圍第3項所述之電漿處理方法是 ，如申請專利範圍第1項或第2項所述之電漿處理方法， 其中上述測量資料使用電漿的發光光譜。 本發明的申請專利範圍第4項所述之電獎處理方法是 -8- (4) (4)200421411 ，如申請專利範圍第3項所述之發明，其中使用上述發光 光譜的波長中，對殘差的貢獻率高的波長。 本發明的申請專利範圍第5項所述之電漿處理方法是 ，如申請專利範圍第1項或第2項所述之發明，其中上述 測量資料使用電氣計量裝置所獲得的高頻電壓。 本發明的申請專利範圍第6項所述之電漿處理方法是 ，如申請專利範圍第1項或第2項所述之發明，其中上述 測量資料使用電氣計量裝置所獲得的高頻電流。 本發明的申請專利範圍第7項所述之電漿處理方法是 ，如申請專利範圍第1項或第2項所述之發明，其中上述 測量資料使用電氣計量裝置所獲得的高頻電壓與高頻電流 的相位差。 本發明的申請專利範圍第8項所述之陳化處理終了檢 測方法是，將試驗甩被處理體供給處理裝置的處理容器內 ，以進行陳化處理時，用以檢測上述陳化處理終了的方法 ，其特徵爲具備有：向上述處理容器內供應上述試驗用被 處理體，冷卻上述處理容器內後，再度向上述處理容器內 供應複數個上述試驗用被處理體時，使用所獲得的複數個 測量資料進行多變量解析，作成用以預測上述陳化處理終 了的預測式的製程；及依據上述預測式，檢測進行上述陳 化處理時的陳化處理終了的製程。 本發明的申請專利範圍第9項所述之電漿處理裝置， 其特徵爲具備有：用以收容被處理體的處理容器；測量此 處理容器內的電漿的發光光譜的檢測器；以及連接在此檢 -9- (5) (5)200421411 測器，輸入來自此檢測器的測量資料之控制裝置，在將試 驗用被處理體供給上述處理容器內，以進行陳化處理時， 向上述處理容器內供應上述試驗用被處理體，冷卻上述處 理容器內後，再度向上述處理容器內供應複數個上述試驗 用被處理體時，依據藉由上述檢測器測量的複數個測量資 料，使用多變量解析程式進行多變量解析，作成用以預測 上述陳化處理終了的預測式，依據此預測式，檢測進行上 述陳化處理時的陳化處理終了的控制裝置。 本發明的申請專利範圍第1 0項所述之電漿處理裝置 ’其特徵爲具備有：用以收容被處理體的處理容器；設在 此處理容器的電氣計量裝置；以及連接在此電氣計量裝置 ，輸入來自此電氣計量裝置的測量資料之控制裝置，在將 試驗用被處理體供給上述處理容器內，以進行陳化處理時 ’向上述處理容器內供應上述試驗用被處理體，冷卻上述 處埋容器內後，再度向上述處理容器內供應複數個上述試 驗用被處理體時，依據藉由上述檢測器測量的複數個測量 資料，使用多變量解析程式進行多變量解析，作成用以預 測上述陳化處理終了的預測式，依據此預測式，檢測進行 上述陳化處理時的陳化處理終了的控制裝置。 【實施方式】 茲參照第1圖至第7圖所示實施形態說明本發明如下 〇 本實施形態的電漿處理裝置1是例如第1圖所示，備 -10- (6) (6)200421411 有：可以保持所希望的高真空度，表面經過氧化鋁加工， 且成電氣方式接地的處理容器2 ;配設在此處理容器2的 底面中央，且載置被處理體（例如晶圓）W的下部電極3 ; 由下方支持此下部電極3且經絕緣構件2 A配設在處理容 器2底面的支持體4 ;經由空隙配設在下部電極3，且形 成爲空心狀的上部電極5。下部電極3是經由匹配器6A 連接在例如2 MHz的高頻電源6，上部電極5是經由匹配 器7A連接在較下部電極3的頻率高的例如60 MHz的高 頻電源7。下部電極3連接有高通濾波器8，上部電極5 連接有低通濾波器9。同時，處理容器2底面的排氣口 2B 經由排氣管1 1 A連接在排氣裝置11，此排氣裝置1 1可真 空排出處理容器2內的空氣，維持所希望的真空度。再者 ，以下將按需要，將下部電極3與支持體4合倂稱作載置 台1〇。 上部電極5的上面中央形成有氣體導入管5A，此氣 體導入管5 A是經由絕緣構件2 C貫穿處理容器2的上面 中央。而經由氣體供應管1 3將處理氣體供應源1 2連接在 此氣體導入管5 A，從此處理氣體供應源1 2供應蝕刻氣體 。亦即，處理氣體供應源12具有：C 5 F 8氣體供應源 12A、0 2氣體供應源12B及At*氣體供應源12C，此等氣 體供應源12A、KB、12C分別連接在氣體供應管13的支 管1 3 A、1 3 B、1 3 C。各支管1 3 A、1 3 B、1 3 C從上游側向 下游側順序設有對應C 5 F 8氣體供應源1 2 A、Ο 2氣體供 應源12B及Ar氣體供應源12C的流量控制裝置12D、 -11 - (7〉 (7〉200421411 12E、12F及閥12G、12H、12 I，經由此等流量控制裝置 12D、12E、12F及閥12G、12H、12 I將供給處理容器2 內的鈾刻用氣體控制成規定流量。 在上部電極5下面均等分散形成多數孔5B，從各孔 5 B向處理容器2內均等分散供應處理用氣體。因此，在 藉由排氣裝置1 1將處理容器2抽真空，同時從處理氣體 供應源1 2以一定流量供應規定的蝕刻用氣體之狀態下， 對下部電極3及上部電極5分別施加高頻電力，使其在處 理容器2內產生蝕刻用氣體的電漿，對下部電極3上的晶 圓W施加規定的蝕刻。此下部電極3裝設有溫度感測器（ 未圖示），經由溫度感測器恆常監視下部電極3的晶圓W 的溫度。 載置台1 〇形成有通過規定的冷媒（例如，傳統上習知 的氟系流體、水等）的冷媒流路1 0 A，冷媒在冷媒流路 1 ο A內流通期間，下部電極3受到冷卻，經由下部電極3 冷卻晶圓W，將晶圓W控制在所希望的溫度。同時，在 下部電極3上配置有絕緣材料構成的靜電夾鉗1 4，靜電 夾鉗1 4內的電極板1 4 A連接有高壓直流電源1 5。靜電夾 鉗14是藉由高壓直流電源15施加於電極板14A的高電 壓而在表面產生的靜電，靜電吸著晶圓W。下部電極3的 外周緣配置有圍繞靜電夾鉗1 4的聚焦環1 6，經由聚焦環 1 6使電漿聚集於晶圓W。 載置台1 0形成有將He氣等熱傳導性氣體當作後側 (backside)氣體供給的氣體流路10B，氣體流路10B在載 -12- (8) (8)200421411 置台10上面的複數個地方開口。此等開口部與形成在載 置台1 0上的靜電夾鉗1 4的貫穿孔一致。因此，當向載置 台1 〇的氣體流路1 0B供應背側氣體時’背側氣體將經由 氣體流路1 〇 B從靜電夾鉗1 4的貫穿孔流出’均勻擴散到 靜電夾鉗1 4與晶圓W間的整個空隙，提高空隙間的熱傳 導性。再者，在第1圖，1 7是形成在處理容器2的開關 晶圓W的運進運出口用的閘口閥。 在電漿處理裝置1安裝有例如終點檢測器1 8，使用 此終點檢測器1 8測量處理容器2內的電漿的發光光譜， 將此測量値取進控制裝置1 9內。此控制裝置1 9儲存有多 變量解析程式，例如主成分分析用的程式，而經由此程式 進行主成分分析。此成分分析用的程式是在處理容器2進 行陳化處理時，用以解析陳化處理用資料。資料解析用的 資料可使用終點檢測器1 8的發光光譜的測量資料。測量 資料是使用例如在193 nm〜95 0 nm範圍內的1 024種波 長。 以下說明，本實施形態的陳化處理資料解析方法，製 作預測所謂陳化處理終了的預測式的測量資料的採取方法 。亦即，更換處理容器2內的淸潔或聚焦環（未圖示）等消 耗品後’爲了使處理容器2穩定，以下述程序進行陳化處 理。首先，在第1天起動電漿處理裝置1後，向處理容器 2內供應仿真晶圓（裸矽）W。然後，從氣體供應管1 3向處 理容器2內供應蝕刻用氣體，以保持一定真空度的狀態， 從局頻電源6、7施加例如60 MHz及2 MHz的高頻電力 -13- (9) (9)200421411 ，進行蝕刻。對例如1 3 0枚仿真晶圓W重複進行此項處 理。以處理1 3 〇枚的仿真晶圓W結束第1天的處理。 然後，暫停鈾刻處理，以接在電源的狀態，亦即，以 可以立刻再起動的狀態將處理容器2放置幾個小時以上。 而將因蝕刻處理變成高溫的處理容器2本身及下部電極3 等內部零件冷卻至設定溫度。 接著’在第2天再度以生產過程之條件，例如將3 〇 枚的仿真晶圓W —次一枚供給處理容器2內，對各仿真 晶圓W重複蝕刻循環。剛開始蝕刻循環時，因爲處理容 器2被冷卻，從第1枚仿真晶圓w進行至第3 0枚期間， 處理容器2本身及處理容器2內的下部電極3、聚焦環等 零件的溫度慢慢上昇。本實施形態是對3 0枚中之最初2 0 枚的有溫度變化的時點對仿真晶圓W在1分鐘內測量1 8 次發光光譜，將上述297種波長的發光強度用作主成分分 析的測量資料。因此，此等測量資料反映有處理容器2內 的溫度變化。 而使用上述測里資料進丫丁主成分分析。例如，對2 0 枚仿真晶圓進行m次（本實施形態是1 8 X 2 0 = 3 6 0次）的 測量’若各次測量存在有n個（本實施形態是2 9 7個波長 的發光強度）測量資料，則包含測量資料的行列可以用數 式1表示之。此行列的行是一次測量所獲得的測量波長的 測量資料成爲其成分，其列則是各波長因時間而變化的測 量資料爲其成分。而在控制裝置1 9依據各該測量資料求 出平均値、分散値、標準偏差後，以平均値與標準偏差値 -14- (10) (10)200421411 規格化。依據此等規格化的値，使用相關行列進行複數個 測量資料的主成分分析，求出固有値及其固有向量。固有 値表示各測量資料的分散的大小，按固有値之大小順序， 將其定義爲第1主成分、第2主成分、......第η主成分。 同時，各固有値分別有所屬的固有向量（權重）。通常，主 成分之次方數愈高對資料的評價的貢獻率愈低，其利用價 値愈輕。 數式1」200421411 (1) 发明 Description of the invention [Technical field to which the invention belongs] The present invention relates to a 'plasma processing method used for an' etching processing apparatus' and the like, and an aging treatment end detection method and a plasma processing apparatus. [Prior Art] For example, a processing device such as an etching processing device includes: a processing container having an airtight structure; and a holding body disposed in the container to hold the object to be processed 'generates plasma in the processing container, A predetermined process is applied to the object. If the processing of the object is continued, the processing container may be contaminated with by-products or the like, or internal parts may be consumed. Therefore, it is necessary to suspend the processing device 'to perform maintenance work such as cleaning of the processing container or replacement of consumables. After finishing the maintenance work, the processing device is started again. For example, when the etching processing device is restarted, it is necessary to supply a predetermined number of dummy wafers (dumm y w a f e r) to the processing container and repeat the etching cycle to adjust the processing container to a state required during production, so-called aging treatment. After the aging process is completed, check the etching rate or the uniformity of the etching in the wafer surface. In the aging process, the measurement data obtained from a plurality of simulated wafers is used, for example, the measurement data of the luminescence spectrum obtained from the endpoint detector, to perform data analysis. Observe the changes in this analytical data to determine whether the aging process is over. However, with the traditional analytical data, it is difficult to see whether the change of the judgment criterion as the end of the aging process is the change caused by the aging process, that is, the change in the state of the processing container, or the change between the simulated wafers. -6-(2) (2) 200421411 Changes in temperature. Therefore, there is a problem that it is difficult to judge whether the aging treatment has ended. That is, the present inventors have performed, for example, a kind of principal component analysis using multivariate analysis. As described later, data analysis is performed on the measurement data. The analysis results show that there are two large peaks that indicate changes, and it is difficult to judge Chen. Whether the processing has ended. The principal component analysis will be described below. At this time, the measurement data is taken in the same way as the traditional method. For example, on the first day, 130 simulated wafers are supplied for etching, and on the second day, the production process is entered and 30 simulated wafers are supplied for etching. The principal component analysis was performed using the measurement data of the luminescence spectrum obtained from the 5 1 to 60 simulated wafers and the 121 to 130 simulated wafers on the first day, and the 8A and 8B images were obtained. And the analysis results shown in Figs. 9A and 9B. In principal component analysis, 297 kinds of wavelengths in the short wavelength range of 1 93 nm to 4 1 9 nm are used in a luminescence spectrum to measure the intensity of each wavelength 18 times per 3 seconds on a simulated wafer in one minute, and Principal component analysis was performed based on these measurement data. The principal component scores and residuals at each measurement were obtained, and the results of drawing HOTELLINGS TSQUARE (the sum of squares of principal components) are shown in Figures 8A and 9A. The results of plotting the sum of squares of residuals (residual scores) are 8B and 9B. From these analysis results, it can be understood that the data of any graphs on the first day and the second day have large peaks, and it is difficult to judge that the aging treatment is over. The horizontal axis of each graph indicates the number of measurements. The present invention has been made in order to solve the above-mentioned problems, and an object thereof is to provide a plasma processing method capable of clearly judging the end of aging treatment, and a detection method and a plasma processing device that are aged out. -7- (3) (3) 200421411 The inventors conducted various reviews on the causes of the two peaks identified, and found that the reason was the method of measuring data used for data analysis. Therefore, it is obtained that, by taking specific treatment of the processing container when taking the data, excluding the influence of the temperature change between the simulated wafers, and truly grasping the changes formed by the aging treatment, it is possible to determine the conclusion of the aging treatment. [Summary of the Invention] The present invention has been completed based on the above conclusion. The plasma processing method described in the first patent application scope of the present invention is to supply a test object to a processing container of a processing device for aging. The method for detecting the end of the aging treatment during processing is characterized by supplying the test object to be processed into the processing container, cooling the processing container, and supplying a plurality of the above to the processing container again. In the case of a test object, a multivariate analysis is performed using the obtained plurality of measurement data to prepare a predictive process for predicting the end of the aging process; and the aging process when the aging process is performed is detected based on the predictive equation. The end of the chemical treatment process. The plasma processing method described in the second patent application scope of the present invention is the invention described in the first patent application scope, wherein the multivariate analysis uses principal component analysis. The plasma treatment method described in item 3 of the scope of patent application of the present invention is the plasma treatment method described in item 1 or 2 of the scope of patent application, wherein the above measurement data uses the emission spectrum of the plasma. The electric award processing method described in item 4 of the scope of patent application of the present invention is -8- (4) (4) 200421411. The invention described in item 3 of the scope of patent application, in which the wavelength of the above emission spectrum is used, The wavelength at which the residual contribution rate is high. The plasma processing method described in item 5 of the scope of patent application of the present invention is the invention described in item 1 or 2 of the scope of patent application, wherein the above-mentioned measurement data uses a high-frequency voltage obtained by an electrical metering device. The plasma processing method described in the sixth aspect of the patent application scope of the present invention is the invention described in the first or second aspect of the patent application scope, wherein the measurement data uses a high-frequency current obtained by an electrical metering device. The plasma processing method described in item 7 of the scope of patent application of the present invention is the invention described in item 1 or 2 of the scope of patent application, wherein the measurement data uses the high-frequency voltage and high voltage obtained by the electrical metering device. Frequency current phase difference. The method for detecting the end of aging treatment described in item 8 of the scope of patent application of the present invention is to supply the test object to the processing container of the processing device for the aging treatment to detect the end of the aging treatment. The method is characterized by comprising: supplying the test object to be processed into the processing container, cooling the processing container, and then supplying the test object to the processing container again, using the obtained plural number A multivariate analysis is performed on each measurement data to create a predictive process for predicting the end of the aging process; and a process that ends the aging process when the aging process is performed is detected based on the predictive equation. The plasma processing apparatus according to item 9 of the scope of patent application of the present invention is characterized by comprising: a processing container for accommodating an object to be processed; a detector for measuring a light emission spectrum of the plasma in the processing container; and a connection Here is a control device that checks the 9- (5) (5) 200421411 sensor, and inputs the measurement data from this sensor. When the test object is supplied to the processing container for aging treatment, When the test object to be processed is supplied in the processing container, and after cooling the inside of the processing container, a plurality of the test objects to be supplied to the processing container are again used based on the plurality of measurement data measured by the detector. The variable analysis program performs multivariate analysis to create a prediction formula for predicting the end of the aging process. Based on the prediction formula, a control device that detects the end of the aging process when the aging process is performed is detected. The plasma processing apparatus described in item 10 of the scope of patent application of the present invention is characterized by being provided with: a processing container for accommodating an object to be processed; an electrical measuring device provided in the processing container; and an electrical measuring device connected to the processing container. Device, a control device that inputs measurement data from this electrical metering device, and when the test object is supplied to the processing container for aging treatment, the test object is supplied to the processing container and cooled. After being buried in the container, when the plurality of test objects to be processed are supplied to the processing container again, a multivariate analysis program is used to perform multivariate analysis based on the plurality of measurement data measured by the detector, and used for prediction. The prediction formula for ending the aging process, and based on the prediction formula, a control device that detects the end of the aging process when the aging process is performed is detected. [Embodiment] The present invention will be described with reference to the embodiments shown in Figs. 1 to 7 as follows. The plasma processing apparatus 1 of this embodiment is shown in Fig. 1, for example. (6) (6) 200421411 There are: a processing container 2 which can maintain a desired high degree of vacuum, whose surface is processed with alumina, and which is electrically grounded; is disposed in the center of the bottom surface of the processing container 2 and mounts a processing object (such as a wafer) W The lower electrode 3 supports the lower electrode 3 from below and is arranged on the bottom surface of the processing container 2 via the insulating member 2 A through the insulating member 2 A. The lower electrode 3 is disposed through the gap and is formed as a hollow upper electrode 5. The lower electrode 3 is a high-frequency power source 6 connected to, for example, 2 MHz via a matcher 6A, and the upper electrode 5 is a high-frequency power supply 7 connected to the lower electrode 3 via a matcher 7A, such as 60 MHz. A high-pass filter 8 is connected to the lower electrode 3, and a low-pass filter 9 is connected to the upper electrode 5. At the same time, the exhaust port 2B on the bottom surface of the processing container 2 is connected to the exhaust device 11 via an exhaust pipe 1 1 A. This exhaust device 11 can exhaust the air in the processing container 2 in a vacuum and maintain a desired degree of vacuum. It should be noted that the combination of the lower electrode 3 and the support 4 will be hereinafter referred to as a mounting table 10 as necessary. A gas introduction pipe 5A is formed in the center of the upper surface of the upper electrode 5, and this gas introduction pipe 5A penetrates the center of the upper surface of the processing container 2 via an insulating member 2C. A process gas supply source 12 is connected to the gas introduction pipe 5 A via a gas supply pipe 13, and an etching gas is supplied from the process gas supply source 12. That is, the process gas supply source 12 includes: a C 5 F 8 gas supply source 12A, a 02 gas supply source 12B, and an At * gas supply source 12C. These gas supply sources 12A, KB, and 12C are connected to the gas supply pipe 13 respectively. The branch pipes 1 3 A, 1 3 B, 1 3 C. Each branch pipe 1 3 A, 1 3 B, 1 3 C is provided with a flow control device corresponding to C 5 F 8 gas supply source 1 2 A, 0 2 gas supply source 12B, and Ar gas supply source 12C in order from the upstream side to the downstream side. 12D, -11-(7> (7> 200421411 12E, 12F and valves 12G, 12H, 12 I), through these flow control devices 12D, 12E, 12F and valves 12G, 12H, 12 I will be supplied to the processing container 2 The uranium engraving gas is controlled to a predetermined flow rate. A plurality of holes 5B are uniformly dispersed and formed under the upper electrode 5, and the processing gas is evenly distributed from the holes 5B into the processing container 2. Therefore, the processing is performed by the exhaust device 11 When the container 2 is evacuated and a predetermined etching gas is supplied from the processing gas supply source 12 at a constant flow rate, high-frequency power is applied to the lower electrode 3 and the upper electrode 5, respectively, so that etching is generated in the processing container 2. A gas plasma applies a predetermined etching to the wafer W on the lower electrode 3. The lower electrode 3 is provided with a temperature sensor (not shown), and the wafer of the lower electrode 3 is constantly monitored by the temperature sensor. Temperature of W. A predetermined refrigerant is formed on the mounting table 1 〇 For example, conventionally, the refrigerant flow path 10 A of the conventional fluorine-based fluid, water, etc., and the refrigerant flows in the refrigerant flow path 1 ο A, the lower electrode 3 is cooled, and the wafer W is cooled through the lower electrode 3, The wafer W is controlled at a desired temperature. At the same time, an electrostatic clamp 14 made of an insulating material is arranged on the lower electrode 3, and a high-voltage DC power source 15 is connected to the electrode plate 1 4 A in the electrostatic clamp 14. The clamp 14 is static electricity generated on the surface by the high voltage applied to the electrode plate 14A by the high-voltage DC power source 15 and the wafer W is electrostatically attracted. A focus ring 1 surrounding the electrostatic clamp 14 is arranged on the outer periphery of the lower electrode 3. 6. The plasma is collected on the wafer W through the focusing ring 16. The mounting table 10 is formed with a gas flow path 10B for supplying a thermally conductive gas such as He gas as a backside gas, and the gas flow path 10B is loaded. -12- (8) (8) 200421411 The openings are located in a plurality of places on the mounting table 10. These openings are consistent with the through-holes of the electrostatic clamps 14 formed on the mounting table 10. Therefore, when the mounting table 1 〇 When the back side gas is supplied to the gas channel 1 0B, the back side gas will pass through the gas The circuit 10B flows out through the through-holes of the electrostatic clamp 14 and spreads uniformly to the entire gap between the electrostatic clamp 14 and the wafer W, thereby improving the thermal conductivity between the gaps. Furthermore, in Figure 1, 17 is A gate valve for the loading and unloading of the switching wafer W formed in the processing container 2. The plasma processing apparatus 1 is equipped with, for example, an end point detector 18. The end point detector 18 is used to measure the plasma in the processing container 2. Take the measurement spectrum into the control device 19. This control device 19 stores a multivariate analysis program, such as a program for principal component analysis, and performs principal component analysis through this program. This component analysis program is used to analyze the aging data when the processing container 2 is aging. For the data analysis, measurement data of the emission spectrum of the end point detector 18 can be used. The measurement data uses, for example, 1,024 wavelengths in a range of 193 nm to 95 0 nm. The aging processing data analysis method according to this embodiment will be described below. It is a method for generating predictive measurement data that predicts the end of the aging processing. In other words, after replacing consumables such as a cleaning tool or a focus ring (not shown) in the processing container 2 ', in order to stabilize the processing container 2, the aging treatment is performed in the following procedure. First, after the plasma processing apparatus 1 is started on the first day, a dummy wafer (bare silicon) W is supplied into the processing container 2. Then, the etching gas is supplied into the processing container 2 from the gas supply pipe 13 to maintain a certain degree of vacuum, and high-frequency power such as 60 MHz and 2 MHz is applied from the local frequency power supply 6, 7-13 (9) (9) 200421411, etching. This process is repeated for, for example, 130 simulated wafers W. The processing of the first day ends with the processing of 130 simulated wafers W. Then, the uranium engraving process is suspended to be connected to the power source, that is, the processing container 2 is left in a state where it can be immediately restarted for several hours or more. In addition, the internal parts such as the processing container 2 itself and the lower electrode 3 that have become high temperature due to the etching process are cooled to a set temperature. Next, on the second day, the conditions of the production process are again used, for example, 30 simulated wafers W are supplied to the processing container 2 one at a time, and the etching cycle is repeated for each simulated wafer W. At the beginning of the etching cycle, since the processing container 2 is cooled, the temperature of the processing container 2 itself and the lower electrode 3 and the focus ring in the processing container 2 is slow during the period from the first simulation wafer w to the 30th. Slow rise. In this embodiment, the light emission spectrum of the simulated wafer W is measured 18 times within 1 minute at the time when the temperature of the first 20 of the 30 is changed, and the light emission intensity of the aforementioned 297 wavelengths is used as the principal component analysis. Measurement data. Therefore, these measurement data reflect the temperature change in the processing container 2. The above-mentioned data were used for YAD principal component analysis. For example, perform m measurements on 20 simulated wafers (18 X 2 0 = 360 times in this embodiment). 'If there are n measurements (in this embodiment, there are 297 wavelengths) Luminous intensity) measurement data, the ranks containing the measurement data can be expressed by Equation 1. The row of this row and column is the measurement data of the measurement wavelength obtained by one measurement, and its column is the measurement data of each wavelength that changes with time as its component. After the control device 19 obtains the average 値, the dispersed 値, and the standard deviation based on each of the measurement data, the control unit 19 standardizes the average 値 and the standard deviation -14- (10) (10) 200421411. Based on these normalized radon, the principal component analysis of a plurality of measurement data is performed using related rows and columns to obtain the intrinsic radon and its eigenvector. The intrinsic 値 represents the size of the dispersion of each measurement data, and it is defined as the first principal component, the second principal component, and the η-th principal component in the order of the magnitude of the intrinsic 値. At the same time, each eigenvector has its own eigenvector (weight). In general, the higher the power of the main component, the lower the contribution rate to the evaluation of the data, and the lighter its utilization price. Equation 1 "
A H 本實施形態是對20枚的仿真晶圓進行m次的測量， 按各測量分別取η個測量資料，對應第i次測量的第j個 固有値的第j主成分，以數式2表示之。而在此第j主成 分代入由具體的第i次測量所獲得的測量資料（Xll、 X i 2、......X i η )所獲得的値，成爲對第i次測量的第j主成 分的得分。因此第j主成分的得分t j由數式3定義，第j 主成分的固有向量Pj由數式4定義。使用行列X與固有 向量P j時，第j主成分的得分t ^由數式5表示之。若使 用主成分的得分與各固有向量，行列X便由數式6表示 -15- (11)200421411 數式2」 hj 二 Xη?j\ + Xi:Pj2 ^ V XinPjn 「數式3」 乂.AH In this embodiment, 20 measurements are performed on 20 simulated wafers, η measurement data is taken for each measurement, and the j-th principal component corresponding to the j-th intrinsic chirp of the i-th measurement is expressed by Equation 2. . Here, the j-th principal component is substituted into the 値 obtained from the measurement data (Xll, X i 2,... X i η) obtained from the specific i-th measurement, and becomes the i-th measurement of the i-th measurement. The score of the j principal component. Therefore, the score t j of the j-th principal component is defined by Equation 3, and the eigenvector Pj of the j-th principal component is defined by Equation 4. When the rank X and the eigenvector P j are used, the score t ^ of the j-th principal component is expressed by Equation 5. If the score of the principal component and each eigenvector are used, the rank X is represented by Equation 6 -15- (11) 200421411 Equation 2 "hj two Xη? J \ + Xi: Pj2 ^ V XinPjn" Expression 3 "乂.
數式4 Λ Ρ, Λ 參 J» 尸， 數式5」Equation 4 Λ Ρ, Λ see J »Corpse, Equation 5"
數式6」 X - ίλΡχ + t2P^ Η----+ 其中，p n T是p n的轉置行列。 因此，在主成分分析’縱使有多種測量資料，可以綜 合例如第1主成分及第2主 數統計資料，調查少數統計 ，判斷陳化處理終了。例如 成分’最多到第3主成分的少 胃#即可掌握陳化處理的狀況 ’―般是如果第1、第2主成 -16- (12) (12)200421411 分的固有値的累計貢獻率超過9〇 %時，依據第1、第2主 成分的評價的可靠性便很高。第1主成分是如上述表示測 量資料分散最廣的方向，適合於掌握處理容器2的陳化處 理隨時間的變化，判斷陳化處理終了。第1、第2主成分 無法掌握的變化可由殘差得分來掌握。本實施形態是求出 弟 1主成分。 因此，本實施形態是以下述條件對仿真晶圓 W施加 蝕刻’從這時的測量資料的主成分分析，固有値可以使用 測量資料的相關行列求出，最大的固有値成爲第1主成分 得分的分散。第1主成分的固有向量可以使用固有値及相 關行列求出。算出各測量資料的主成分得分，而描繪各個 主成分得分的平方和（HOTELLINGS TSQUARE)的結果便 是第2圖所示的曲線圖。從此曲線圖可以明白，僅在第1 天的測量資料看到很大的峰値，第2天的測量資料就看不 到有很大峰値，可以確實判斷陳化處理終了。同時，描繪 各測量資料的殘差的平方和的結果便是第2B圖所示的曲 線圖。在此圖也是僅在第1天的測量資料看到很大的峰値 ，可以確實判斷陳化處理終了。再者，各曲線圖的橫軸表 示測量次數。本實施形態是對一枚仿真晶圓W進行1 8次 的測量，而且在兩天內處理1 6 0枚的仿真晶圓W，因此， 刻度是到1 8 X 1 6 0 = 1 8 8 0。 〔處理條件〕 處理裝置：電容耦合型平行平板電漿處理裝置 -17- (13)200421411 仿真晶圓（裸矽）：3 00 mm 下部電極的電源高頻頻率及電力：2 MHz， 上部電極的電源高頻頻率及電力：60 MHz， 處理壓力：25mTorr 蝕刻用氣體：C 5 F 8 = 29 s c c mEquation 6 "X-ίλΡχ + t2P ^ Η ---- + where p n T is the transposed rank of p n. Therefore, even if there is a variety of measurement data in the principal component analysis, it is possible to integrate statistics such as the first principal component and the second principal number, investigate a few statistics, and judge that the aging process is over. For example, the component 'up to the third main component of Shaowei # can grasp the state of aging treatment'-generally, if the first and second main components are -16- (12) (12) 200421411 points, the cumulative contribution rate of the intrinsic 値When it exceeds 90%, the reliability of the evaluation based on the first and second principal components is high. The first principal component indicates the direction in which the measurement data is most widely dispersed as described above, and is suitable for grasping the change of the aging treatment of the processing container 2 with time and judging that the aging treatment is finished. Changes that cannot be grasped by the first and second principal components can be grasped by the residual score. In this embodiment, the main component of Brother 1 is obtained. Therefore, in the present embodiment, etching is performed on the simulated wafer W under the following conditions. From the principal component analysis of the measurement data at this time, the intrinsic volume can be obtained using the correlation rank of the measurement data. The largest intrinsic volume becomes the dispersion of the first principal component score. The eigenvector of the first principal component can be obtained by using the eigenvalue and the related ranks. The principal component score of each measurement data is calculated, and the result of drawing the sum of squares of each principal component score (HOTELLINGS TSQUARE) is the graph shown in Figure 2. From this graph, it can be understood that only the measurement data on the first day saw a large peak, and the measurement data on the second day did not see a large peak, it can be sure that the aging treatment is over. At the same time, the result of plotting the sum of squared residuals of each measurement data is the curve shown in Figure 2B. In this picture, only a large peak is seen in the measurement data on the first day, and it can be surely judged that the aging treatment is over. The horizontal axis of each graph indicates the number of measurements. In this embodiment mode, one simulation wafer W is measured 18 times, and 160 simulation wafers W are processed in two days. Therefore, the scale is 1 8 X 1 6 0 = 1 8 8 0 . [Processing Conditions] Processing Device: Capacitively Coupled Parallel Flat Plasma Processing Device-17- (13) 200421411 Simulation wafer (bare silicon): 3 00 mm High-frequency frequency and power of the lower electrode: 2 MHz, of the upper electrode Power frequency and power: 60 MHz, processing pressure: 25mTorr Etching gas: C 5 F 8 = 29 sccm
Ar=750sccm， Ο 2 = 47 背側氣體：H e = 1 5 T 〇 r r (電極中央部）Ar = 750sccm, Ο 2 = 47 Backside gas: He = 1 5 T 〇 r r (central part of the electrode)
40 T o r r (電極邊緣部） 處理溫度：上部電極 =60。(：，側壁 =60 =2 0 °C 如以上所說明，依據本實施形態時，亦 陳化處理終了的預測式的製作方法時，因爲 晶圓W進入生產過程的階段中途暫行停止電 後，將處理容器2放置數小時，令處理容器 部的下部電極3等零件冷卻後，再度進入 2 0枚仿真晶圓W之間，採取判斷陳化處理 資料，因此可以獲得反映處理容器2本身及 零件的溫度變化的測量資料，可以從其解析 溫度變化的峰値。同時，由於將此解析結果 時’得確實檢測陳化處理終了，並加以判斷 檢出陳化處理終了後進行電漿處理，則可對 的鈾刻處理。 第3A圖、第3B圖是表示本發明的其 資料解析方法之圖。本實施形態與上述實施40 T o r r (electrode edge portion) Processing temperature: upper electrode = 60. (:, Side wall = 60 = 2 0 ° C As explained above, according to this embodiment, when the predictive production method is also aged, the wafer W enters into the stage of the production process after the power is temporarily stopped in the middle. After the processing container 2 is left for several hours, the parts such as the lower electrode 3 of the processing container section are cooled, and then they are again entered between 20 simulated wafers W to judge and age the processing data. Therefore, the processing container 2 and its parts can be reflected. The temperature change measurement data can be used to analyze the peak value of the temperature change. At the same time, since the analysis result must be sure to detect the end of the aging treatment, and judge to detect the end of the aging treatment and perform the plasma treatment, then The uranium can be processed. Figures 3A and 3B are diagrams showing the data analysis method of the present invention. This embodiment is similar to the above embodiment.
3 8 0 0 W 3 3 00 W ° C，下部電極 即，依據預測 是在使用仿真 〖漿處理裝置1 2本身及其內 生產過程處理 終了用的測量 下部電極3等 結果排除依據 用在陳化處理 。因此，確實 晶圓施以穩定 他實施形態的 形態不同，是 -18- (14) (14)200421411 在求出就各仿真晶圓W獲得的1 8個測量資料（297種波 長）的平均値後，使用此等平均値進行主成分分析，求出 固有値及固有向量。而描繪各仿真晶圓 W的主成分得分 的平方和及殘差的平方和，其結果便是第3 A圖、第3B 圖所示的曲線圖。從第3A圖、第3B圖可以淸楚看出， 與第2A圖、第2B圖所示的上述實施形態的曲線一樣， 可以判斷陳化處理的終了。再者，橫軸的數値是仿真晶圓 的數目。 而，第4圖、第5A圖、第5B圖是表示本發明的另 一其他實施形態的資料解析方法之圖。本實施形態是如第 4圖所示，選擇例如對測量資料的殘差的貢獻大的波長（ 例如第4圖中以Ο所圍的波長）1 〇種，就此等1 〇種波長 ，與上述實施形態同樣，按每一仿真晶圓 W求其平均値 後，使用此等平均値進行主成本分析，求出固有値及固有 向量。而描繪各仿真晶圓W的第1主成分得分及殘差得 分，其結果便是第5 A圖、第5B圖所示的曲線圖。從第 5A圖、第5B圖可以淸楚看出，與第2A圖、第2B圖所 示的實施形態比較，曲線的曲折減少，成爲順暢圓滑的曲 線，可以更簡單判斷陳化處理的終了。 而第6A圖、第6B圖是表示本發明的再一其他實施 形態的資料解析方法之圖。本實施形態是與第5 A圖、第 5B圖所示實施形態同樣，選擇對殘差的貢獻率高的1 0種 波長。而第5A圖、第5B圖所示實施形態是採用仿真晶 圓 W的各波長的測量資料的時間平均値，但本實施形態 -19- (15) (15)200421411 是採用測量資料本身，這一點與第2 A圖、第2 B圖所示 情形相同。然而，第2A圖、第2B圖所示者的行列的行 ’其成分是一次測量所獲得的測量資料’其列的成分是各 波長因時間變化之測量資料，但本實施形態是按各仿真晶 圓W與各波長將行與列轉置。一行是就1 〇波長測量1 6 次，因此，具有10 X 1 6 = 1 60成分。一列是因爲將晶圓 2〇枚放進訓練集合，因此，具有20成分的主成分分析的 訓練集合是2 0行、1 6 0列的行列。依據此行列，與上述 實施形態同樣進行主成分分析，描繪主成分得分的平方和 與殘差的平方和，其結果便是第6A圖及第6B圖所示的 曲線圖。從第6A圖、第6B圖可以淸楚看出’與第5A圖 、第5 B圖所示的曲線圖比較，曲線的曲折減少，成爲順 暢圓滑的曲線，可以更簡單判斷陳化處理的終了。 對第7圖所示的電漿處理裝置20也可以與上述電漿 處理裝置1同樣應用本發明，並可期望收到跟上述電漿處 理裝置1同樣的作用效果。此電漿處理裝置2 0是如第7 圖所示，備有：由鋁等導電性材料構成的處理容器2 1 ; 配設在此處理容器2 1內的底面，且兼用以載置晶圓W的 載置台的下部電極2 2 ;分開規定間隔配設在此下部電極 22上方，且兼用作蝕刻氣體的供應部的空心狀接地的上 部電極23 ;及可賦予旋轉磁場的磁場形成手段24，而在 控制裝置25的控制下，使磁場形成手段24形成的旋轉磁 場B作用於產生在處理容器2 1的上下兩電極間的磁場， 以高密度電漿對晶圓W進行均一的電漿處理。 -20- (16) (16)200421411 處理容器2 1上面連接有連通上部電極23的氣體供應 管26，經由氣體供應管26及上部電極23，從氣體供應源 (未圖示）向處理容器21內供應蝕刻氣體。處理容器21的 側面連接有連結在未圖示的真空排氣裝置的氣體排氣管 2 7，經由真空排氣裝置及氣體排氣管2 7將處理容器2 1內 減壓，保持規定真空度。下部電極22連接有高頻電源28 ，從高頻電源28向下部電極22施加高頻電力，在電極 22、23間產生蝕刻氣體的電漿，對下部電極22上的半導 體晶圓W施加例如規定的蝕刻處理。 電漿處理裝置20安裝有例如終點檢測器29，使用此 終點檢測器2 9測量處理容器2 1內電漿的發光光譜，再將 此測量結果取進控制裝置25內。此控制裝置25儲存有解 析程式之例如主成分分析用的程式，可經由此程式進行主 成分分析。此主成分分析用的程式是在陳化處理容器2 1 時，用以解析陳化處理用的資料。解析資料用的資料使用 終點檢測器2 9的發光光譜的測量資料。測量資料使用例 如在193 nm〜950 nm範圍的1024種波長。 再者，上述各實施形態是以主成分分析爲例子，說明 判斷陳化處理終了的資料的解析手法，但也可以使用其他 的多變量解析。同時，上述各實施形態是就使用電漿的發 光光譜時進行說明，但是也可以使用其他測量資料，例如 由設在電漿處理裝置內的電氣量測裝置（VI探針）所檢出的 高頻電壓、高頻電流、高頻電壓與高頻電流的向位差等容 易受到處理容器內的溫度變化影響的測量資料。同時，上 -21 - (17) (17)200421411 述各實施形態是以蝕刻處理裝置爲例子進行說明，但其他 電漿處理裝置也可以應用本發明。 依據本發明的申請專利範圍第1項至第7項記載的說 明’可以提供，能夠明確判斷[陳化處理結果的電漿處方 法及陳化處理終了檢測方法以及電漿處理裝置。 【圖式簡單說明】 第1圖是表示應用本發明的陳化處理資料的解析方法 及陳化處理終了檢測方法的電漿處理裝置的一個例子的架 構圖。 第2A圖、第2B圖是表示有關由本發明一實施形態 獲得的第1圖所示電漿處理裝置的測量資料的解析結果之 Η ’第2Α圖是表示測量資料的主成分點數的平方和的變 動的曲線圖、第2Β圖是表示測量資料的殘差點數的變動 的曲線圖。 第3 Α圖、第3 β圖是表示由本發明的其他實施形態 獲得的解析結果之圖，是分別對應第2 A圖、第2 B圖的 曲線圖。 第4圖是表示本發明的另一其他實施形態所用的發光 光譜的測量資料對殘差的貢獻率的曲線圖。 第5A圖、第5B圖是表示使用第4圖所示波長的平 均値所獲得的解析結果之圖，是分別對應第2A圖、第2B 圖的曲線圖。 第6A圖 '第6B圖是表示使用第4圖所示波長所獲 -22- (18) (18)200421411 得的解析結果之圖，是分別對應第2A圖、第2B圖的曲 線圖。 第7圖是表示應用本發明的陳化處理資料的解析方法 及陳化處理終了檢測方法的電漿處理裝置的另一個例子的 架構圖。 第8 A圖、第8 B圖是表示以傳統的解析手法所獲得 的解析結果之圖，是使用陳化處理第1天的第5 1〜6〇片 的仿真晶圓時的分別對應第2A圖、第2B圖的曲線圖。 第9A圖、第9B圖是表示以傳統的解析手法所獲得 的另一解析結果之圖，是使用陳化處理第1天的第1 2 1〜 1 3 0片的仿真晶圓時的分別對應第2 A圖、第2B圖的曲線 圖0 [圖號說明】 1、 2 0 :電漿處理裝置 2、 2 1 :處理容器 擊 2 A :絕緣構件 2 B :排氣口 2 C :絕緣構件 3、 22 :下部電極 4 :支持體 5、2 3 :上部電極 5 A :氣體導入管 5B ··孔 23- (19) (19)200421411 6、7、2 8 :高頻電源 6 A、7 A :匹配器 8 :高通濾波器 9 :低通濾波器 - 1 0 :載置台 1 0 A :冷媒流路 10B :氣體流路 I 1 :排氣裝置 _ II A :排氣管 1 2 :處理氣體供應源 1 2 A、1 2 B、1 2 C :氣體供應源 1 2 D、1 2 E、1 2 F :流量控制裝置 1 2G、1 2H、1 21 :閥 13、26 :氣體供應管 13 A、13B、13C :支管 Μ :靜電夾鉗 ¥ 1 4 A :電極板 . 1 5 :高壓直流電源 1 6 :聚焦環 1 7 :閘口閥 1 8 :終點檢測器 1 9 :控制裝置 24 :磁場形成手段 2 5 :控制裝置 -24 - (20)200421411 27 :氣體排氣管 W :晶圓3 8 0 0 W 3 3 00 W ° C, the lower electrode is based on the prediction that the simulation is used. [Pulp processing device 1 2 itself and the measurement of the end of the production process are used to measure the lower electrode 3. The results are excluded from aging. deal with. Therefore, it is true that the form in which the wafer is stabilized is different from -18- (14) (14) 200421411. The average value of 18 measurement data (297 wavelengths) obtained for each simulated wafer W is obtained. Then, a principal component analysis is performed using these average 値, and the 値 and the eigenvector are obtained. The sum of the sum of the squares of the principal component scores and the sum of the squares of the residuals of each of the simulation wafers W are the graphs shown in Figs. 3A and 3B. It can be clearly seen from FIG. 3A and FIG. 3B that, similar to the curve of the above embodiment shown in FIG. 2A and FIG. 2B, it can be judged that the aging process is finished. Furthermore, the number on the horizontal axis is the number of simulated wafers. Figs. 4A, 5A, and 5B are diagrams showing a data analysis method according to another embodiment of the present invention. In this embodiment, as shown in FIG. 4, for example, 10 kinds of wavelengths (for example, wavelengths surrounded by 0 in FIG. 4) that have a large contribution to the residuals of the measurement data are selected. These 10 kinds of wavelengths are the same as those described above. In the same manner as in the embodiment, after calculating the average value for each simulation wafer W, the main cost analysis is performed using these average values to obtain the intrinsic value and the intrinsic vector. The first principal component score and residual score of each simulation wafer W are plotted, and the results are graphs shown in Figs. 5A and 5B. As can be clearly seen from Figures 5A and 5B, compared with the embodiment shown in Figures 2A and 2B, the tortuosity of the curve is reduced, and it becomes a smooth and smooth curve. It is easier to judge the end of the aging process. 6A and 6B are diagrams showing a data analysis method according to still another embodiment of the present invention. In this embodiment, similar to the embodiment shown in Figs. 5A and 5B, 10 kinds of wavelengths having a high contribution rate to the residual are selected. In the embodiment shown in FIGS. 5A and 5B, the time average of the measurement data of each wavelength of the simulation wafer W is used. However, in the embodiment -19- (15) (15) 200421411, the measurement data itself is used. One point is the same as that shown in Figs. 2A and 2B. However, in the rows and columns shown in FIGS. 2A and 2B, “the components are measurement data obtained by one measurement”, and the components in the columns are measurement data of each wavelength changing with time. However, this embodiment is based on each simulation. The wafer W and each wavelength transpose the rows and columns. One line is measured 16 times for 10 wavelengths, so it has 10 X 1 6 = 1 60 components. One column is because 20 wafers are put into the training set. Therefore, the training set with 20 component principal component analysis is 20 rows and 160 columns. Based on this rank, the principal component analysis was performed in the same manner as in the above embodiment, and the sum of the square of the principal component score and the sum of the squared residuals were plotted. The results are the graphs shown in Figures 6A and 6B. From Figures 6A and 6B, you can clearly see that compared with the graphs shown in Figures 5A and 5B, the curve's tortuosity is reduced, and it becomes a smooth and smooth curve. It is easier to judge the end of the aging process. . The present invention can also be applied to the plasma processing apparatus 20 shown in FIG. 7 in the same manner as the plasma processing apparatus 1 described above, and it is expected that the same effects as those of the plasma processing apparatus 1 can be obtained. As shown in FIG. 7, this plasma processing apparatus 20 includes a processing container 21 made of a conductive material such as aluminum, and a bottom surface disposed in the processing container 21 and also used for wafer placement. W lower stage electrode 2 2 of the stage; a hollow grounded upper electrode 23 which is arranged above the lower electrode 22 at predetermined intervals and also serves as a supply part of the etching gas; and a magnetic field forming means 24 which can impart a rotating magnetic field, Under the control of the control device 25, the rotating magnetic field B formed by the magnetic field forming means 24 acts on the magnetic field generated between the upper and lower electrodes of the processing container 21, and uniformly plasma-processes the wafer W with a high-density plasma. . -20- (16) (16) 200421411 The processing vessel 2 1 is connected to a gas supply pipe 26 communicating with the upper electrode 23, and passes from the gas supply source (not shown) to the processing vessel 21 via the gas supply pipe 26 and the upper electrode 23. An etching gas is supplied inside. A gas exhaust pipe 27 connected to a vacuum exhaust device (not shown) is connected to the side of the processing container 21, and the inside of the processing container 21 is depressurized through the vacuum exhaust device and the gas exhaust pipe 27 to maintain a predetermined vacuum level. . A high-frequency power source 28 is connected to the lower electrode 22, and high-frequency power is applied from the high-frequency power source 28 to the lower electrode 22 to generate a plasma of an etching gas between the electrodes 22 and 23. The semiconductor wafer W on the lower electrode 22 is subjected to, for example, regulations. Etching process. The plasma processing device 20 is equipped with, for example, an end point detector 29, and the end point detector 29 is used to measure the light emission spectrum of the plasma in the processing container 21, and the measurement result is taken into the control device 25. The control device 25 stores a program such as a program for principal component analysis, and can perform principal component analysis through the program. This principal component analysis program is used to analyze the aging data when the aging container 2 1 is aged. As the data for analyzing the data, the measurement data of the emission spectrum of the end detector 29 was used. Measurement data is used, for example, 1024 wavelengths in the range of 193 nm to 950 nm. In addition, each of the embodiments described above uses the principal component analysis as an example to explain the analysis method of the data to judge the aging process, but other multivariate analysis may be used. In the meantime, the above-mentioned embodiments are described in the case of using the plasma emission spectrum, but other measurement data may also be used, such as the height detected by the electrical measuring device (VI probe) installed in the plasma processing device. Measurement data such as high-frequency voltage, high-frequency current, the high-frequency voltage and the high-frequency current misalignment are easily affected by temperature changes in the processing vessel. In the meantime, each embodiment described in the above -21-(17) (17) 200421411 is described by taking an etching processing device as an example, but other plasma processing devices can also apply the present invention. According to the descriptions in items 1 to 7 of the scope of patent application of the present invention, it is possible to provide a clear judgment of [plasma prescription method of aging treatment result, aging treatment end detection method, and plasma treatment device. [Brief Description of the Drawings] Fig. 1 is a block diagram showing an example of a plasma processing apparatus to which the aging processing data analysis method and aging processing end detection method of the present invention are applied. Figures 2A and 2B show analysis results of measurement data about the plasma processing apparatus shown in Figure 1 according to an embodiment of the present invention. 'Figure 2A shows the sum of the squares of the main component points of the measurement data. FIG. 2B is a graph showing changes in the number of residual points of the measurement data. Figures 3A and 3β are graphs showing analysis results obtained by other embodiments of the present invention, and are graphs corresponding to Figures 2A and 2B, respectively. Fig. 4 is a graph showing the contribution rate of the measurement data of the emission spectrum used in another embodiment of the present invention to the residual error. Figures 5A and 5B are graphs showing the analysis results obtained by using the average chirp of the wavelength shown in Figure 4, and are graphs corresponding to Figures 2A and 2B, respectively. Fig. 6A 'and Fig. 6B are graphs showing the analysis results obtained using the wavelength shown in Fig. -22- (18) (18) 200421411, which are graphs corresponding to Figs. 2A and 2B, respectively. Fig. 7 is a block diagram showing another example of a plasma processing apparatus to which the aging processing data analysis method and aging processing end detection method of the present invention are applied. Figures 8A and 8B are diagrams showing the analysis results obtained by the conventional analysis method. They correspond to the 2A when the 51 to 60th artificial wafers on the first day of aging treatment are used. Fig. 2B is a graph of Fig. 2B. Figures 9A and 9B are diagrams showing another analysis result obtained by the conventional analysis method, and respectively correspond to the 1 21 to 130 simulated wafers on the first day of aging treatment. Graphs 2A and 2B 0 [Description of drawing numbers] 1, 2 0: Plasma processing device 2, 2 1: Treatment container 2 A: Insulating member 2 B: Exhaust port 2 C: Insulating member 3, 22: lower electrode 4: support 5, 2 3: upper electrode 5 A: gas introduction tube 5B · hole 23- (19) (19) 200421411 6, 7, 2 8: high-frequency power supply 6 A, 7 A: Matcher 8: High-pass filter 9: Low-pass filter-1 0: Mounting table 1 0 A: Refrigerant flow path 10B: Gas flow path I 1: Exhaust device_ II A: Exhaust pipe 1 2: Treatment Gas supply source 1 2 A, 1 2 B, 1 2 C: Gas supply source 1 2 D, 1 2 E, 1 2 F: Flow control device 1 2G, 1 2H, 1 21: Valve 13, 26: Gas supply pipe 13 A, 13B, 13C: Branch tube M: Electrostatic clamp ¥ 1 4 A: Electrode plate. 1 5: High-voltage DC power supply 16: Focusing ring 17: Gate valve 1 8: End point detector 1 9: Control device 24: Magnetic field forming means 2 5: Control device-24-(20) 200421411 27 : Gas exhaust pipe W: Wafer
-25--25-