TWI306269B - - Google Patents

Description

1306269 A7 B7 V. INSTRUCTION DESCRIPTION OF THE INVENTION Q) 1 [Field of the Invention] (Please read the back of the page and then fill in the page.) The present invention relates to a processed object or device such as a wafer processed by a semiconductor manufacturing apparatus. BACKGROUND OF THE INVENTION [Description of the Prior Art] Various processing devices in a semiconductor process are used. For example, a film forming process of a target object such as a semiconductor wafer or a glass substrate, or a processing device such as a plasma processing apparatus in an etching process is widely used. Each processing device has its own process characteristics inherent to the object to be processed. Therefore, when the processing object (for example, a semiconductor wafer) is subjected to, for example, an etching process using each processing device, for example, a test wafer is prepared in advance, and the test wafer is periodically subjected to an etching process according to the processing result (for example, The amount of chipping of the test wafer, etc.), and the state of the processing device at that time is determined. Printed by the Ministry of Economic Affairs, the Intellectual Property Office, and the Consumer Cooperative. However, the method of judging the state of the processing device at that time based on the test wafer requires a lot of test wafers to be produced. Further, since a plurality of S-type wafers are processed by a processing device, it is necessary to measure the respective processing results each time, so that the measurement of the fabrication of the test wafer and the measurement of the processing result require a lot of work and time. Further, a process monitoring method of a plasma processing apparatus is proposed in Japanese Laid-Open Patent Publication No. Hei 10 - 1 2 5 6 6 0. This method uses the applicable wafer before processing, and creates an electrical signal reflecting the state of the plasma and a mode associated with the processing characteristics of the electric paddle. The electrical paper size obtained when processing the actual wafer is applied to the Chinese national standard ( CNS ) A4 size (210X 297 mm) 1306269 A7 B7 V. INSTRUCTIONS (2) The detection of the signal is substituted into the mode to predict the plasma processing characteristics. This method is excellent in the point of predicting the plasma processing characteristics, and includes actual wafer processing including a change in the batch (L 〇t ) with time and a sudden change in the applied state of the high-frequency power. The forecast is not sufficient and it is necessary to improve one. The present invention has been made in view of the above problems of the prior art, and its object is to provide a predictive formula for process characteristics by providing only a small amount of operational data and process characteristic data obtained by collecting and processing a small number of samples. (Mode) 'The prediction method and the processing device for predicting the processing result of the process characteristics can be simply and accurately predicted by applying the operation data when the object to be processed is processed to the prediction formula. SUMMARY OF THE INVENTION In order to solve the above problems, according to a first aspect of the present invention, a method for predicting a processing result is provided, which is processed one by one (for example, an etching process) in a processing chamber of a processing device such as a plasma processing device. The process of the object to be processed (for example, a semiconductor wafer) predicts the processing result based on the operation data of the processing device and the processing result data, and the feature includes: a step of collecting the operation data and the processing result data; and according to the collected operation data and a step of performing multivariate analysis on the processing result data (data group); a step of obtaining a correlation relationship between the operation data and the processing result data via the multivariate analysis; and obtaining the correlation relationship by using the correlation relationship The size of the paper to be processed is applicable to the Chinese National Standard (CNS) Α4 specification (210X297 mm) (please read the notes on the back and fill out this page), τ Ministry of Economic Affairs Intellectual Property Bureau employee consumption cooperative printing-5- Ministry of Economic Affairs Intellectual Property Bureau S Workers Consumption Cooperative Printed 1306269 Λ7 B7 V. DESCRIPTION OF EMBODIMENT () The operation data at the time of the object to be processed other than the body, and the steps of predicting the processing result. In order to solve the above problems, according to a second aspect of the present invention, a processing apparatus for processing (for example, uranium engraving) a plurality of processed objects (for example, semiconductor wafers) one by one in a processing chamber is provided according to the operation data and processing. The processing device for predicting the processing result, such as a plasma processing device, includes: a mechanism for storing the operational data; a mechanism for storing the processing result data; and performing multivariate according to the stored operational data and the processing result data a mechanism for analyzing; a mechanism for obtaining a correlation between the operation data and the processing result data by the multivariate analysis; a mechanism for storing the obtained correlation relationship; and obtaining the correlation relationship by using the correlation according to the stored correlation relationship The operation data at the time of the object to be processed other than the object to be processed, and the mechanism for predicting the processing result. According to the first aspect and the second aspect of the present invention, the correlation between the operational data and the processing result data can be obtained by multivariate analysis only by collecting a small amount of operational data and processing result data obtained by processing, for example, a small number of samples. (for example, a predictive formula such as regression). Then, only the operation data when the object to be processed is processed is placed in the correlation, and the processing result of the object to be processed can be predicted simply and accurately. Moreover, the above-mentioned multivariate analysis is subject to the Chinese National Standard (CNS) A4 specification (210x297 mm) if it is subjected to heavy regression analysis (please read the note on the back and fill in this page). 1306269 Λ7 B7 V. Inventive Note (,) 4, by means of re-regression analysis, even if it is a large number of variable data, it is easy to find the regression equation of the correlation between the operational data and the processing result data. (Please read the precautions on the back and fill out this page.) If the PLS method is used for the above multivariate analysis, even if it is a large number of variable data, it is easy to find the operation data and the processing result data. The relationship of the relationship. Further, the operation data may include temperature data of the mounting table on which the object to be processed is placed, and may include data of the back gas pressure. The operation data is easy to affect the processing result data (related). Since the temperature data of the stage and the back gas pressure are included, the prediction accuracy of the processing result can be improved. Further, the operation data may include a standard deviation of the back gas pressure (for example, a back side gas pressure such as He gas), and may include an in-plane pressure difference of the object to be treated including the back gas pressure (for example, the back gas is centered, Information on the pressure difference at the center and the edge of the three systems can also be used. Since the standard deviation of the back gas pressure and the like is indicative of the stability of the back gas pressure, in particular, the processing result data of the object to be processed is useful for predicting the in-plane uniformity of the amount of wafer W to be removed, and the prediction accuracy is also improved. . The Ministry of Economic Affairs, the Intellectual Property Office, the employee consumption cooperative, and the operation data includes at least the voltage data of the high-frequency power source applied when the object to be processed is processed, and at least includes the cumulative operation time of the high-frequency power source. The information is also available. Further, it is also possible to include both the voltage data of the high-frequency power source and the data of the accumulated operation time of the high-frequency power source. The voltage data of the high-frequency power source and the accumulated operation time of the high-frequency power source are particularly applicable to the paper size of the processed object by the processing result data for predicting the amount of chipping (for example, etching rate). China National Standard (CNS) A4 Specification (210X29*7 mm) 1306269 A7 B7 V. Invention Description () 5 High usability and improved prediction accuracy. (Please read the precautions on the back and fill out this page.) The cumulative operation time of the above-mentioned high-frequency power supply can be reset to zero every time the maintenance of the processing room is performed. For the cumulative time of application of the high-frequency power of the trace data, for example, maintenance by a wet cleaning (Wet cleaning), for example, the cumulative application time is zero, so that each wet cleaning cycle can be obtained. Apply accumulated time data. Therefore, when the accumulated time of application of the high-frequency power is used as the operation data, the prediction result can be accurately predicted even if the processing result data is changed by the wet cleaning. In the case of the uranium engraved object of the information on the in-plane uniformity of the data of the above-mentioned object to be processed, or the amount of the in-plane uniformity of the amount of the object to be processed, The processing result data may be a processing result of the object to be processed which is etched as information on the in-plane uniformity of the data of the amount of the object to be processed or the amount of the chipping amount of the object to be processed. According to this, only a small amount of the operation data and the processing result data obtained by the processing of, for example, a small number of samples are etched, and the information on the in-plane uniformity of the data of the object to be processed or the amount of the in-plane uniformity of the amount of the object to be processed is etched. The processing result of the processed body can also be predicted simply and accurately. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a cross-sectional view showing a processing apparatus relating to a first embodiment of a prediction method to which the present invention is applied. Fig. 2 is a view showing a multi-variable analysis device according to the embodiment, a paper scale applicable to the Chinese national standard (€奶) 八4 specification (210 parent 297 mm)~~~~1306269 at Β75, invention description () The Ministry of Economic Affairs, the Intellectual Property Office, and the employee-consumer cooperatives have printed on the display of the amount of disassembly and analysis. The display is more than the case where the block 3 square chart is obtained. Figure. The change of the indicator map 2 The time-varying temperature of the time-varying graph is the surface temperature of the wall-wall diagram of the 7VJ 3 2 咅 的 得 得 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 The change of the graph of Fig. 2 shows that the temperature of the graph 5 is the result of the RH 咅 Η 0. When the amount of resolution is changed, 4^-^^, '-^ shows the difference. 2 The deviation is marked. The apparent force is that the pressure of the 6-body diagram is higher, and the amount of the solution is much larger. When the graph is changed, the oscillating pressure shown by the graph is the source of the electric crystal of the electric source 7 . The 2-quantity map of the inner surface of the quantitative variation sentence which is analyzed by the disintegration of the figure is used to cut the visible film. The process and the materials produced are obtained and transferred to the equipment. ~ Quantitative graph 3 The graph 値 値 用以 用以 用以 用以 用以 用以 用以 用以 用以 用以 用以 用以 用以 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 The mapping is the 关1 of the 1 关 关 关 关 1 1 1 关 关 关 关 关 关 关 关 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 The surface of the map is shown in the form of a section with a large amount of analysis and a large amount of analysis. It is the example of the block diagram of the block 1 of the block diagram. The set of the sample is analyzed. The volume of the circle is changed. The rate of the crystal is more than the rate of the etched 2. The film is represented by oxygen. The D is V C. 1 W Figure round paper size is applicable to China National Standard (CNS) Α4 specification (210X297 mm) (please read the back note first and then fill out this page) -9- 1306269 at B7 V. Invention description b) Figure. Figure 1 4 ( a ) is the use of optical data to display the operating data, not carried out (please read the back of the note before filling this page) Pre-processing etch rate prediction 値 and the actual measurement 値, Figure 14 4 (b) It is a graph showing the correlation between the predicted 値 and the measured 値. Figure 1 5 (a) is a graph showing the prediction of uranium engraving rate and the measured enthalpy of the optical data and tracking data for the operational data. Figure 15 (b) shows the correlation between the predicted enthalpy and the measured enthalpy. relation chart. Fig. 16 (a) is a graph showing the prediction of the etching rate and the measured enthalpy when the tracking data of the operating data is not used, and Fig. 16 (b) is a correlation diagram showing the predicted enthalpy and the measured enthalpy. Fig. 1 7 ( a ) is a graph showing the prediction of the etching rate and the measured enthalpy when the VI probe data is used for the operation data, and the pre-processing is not performed. Figure 17 (b) shows the correlation between the predicted 値 and the measured 値. . Fig. 1 8 ( a ) is a graph showing the predicted 値 and measured 値 of the etch rate when the optical data is used for the operation data, and Fig. 18 (b) shows the correlation between the predicted 値 and the measured 値. relation chart. Figure 1 9 (a) is a graph showing the prediction of the etching rate and the measured enthalpy when using the optical data of the operating data and the printed materials of the employee's consumer cooperatives of the Intellectual Property Office of the Ministry of Economic Affairs. 9 (b) is a graph showing the correlation between predicted 实 and measured 値. Fig. 20 ( a ) is a graph showing the prediction of the etching rate and the measured 値 when using the tracking data of the operating data, and Fig. 20 (b) shows the correlation between the predicted 値 and the measured 値. Figure. Figure 2 1 (a) shows the VI probe data for the operation data. The Chinese National Standard (CNS) A4 specification (210X297 Gongdong) is applied to the paper scale. -10- 1306269 A7 B7 Printed by the Intellectual Property Office of the Ministry of Economic Affairs V. INSTRUCTION OF THE INVENTION (8) The prediction of the etch rate when using the pre-processing of 0 CS and the 1 1 of the actual measurement 图 Figure 2 1 (b) shows the correlation between the predicted 値 and the measured 〇1 1 Fig. 2 2 ( a ) is a graph showing the prediction of the etch rate and the measurement of the etch rate when using the optical data for the operation data using the optical data, Μ % Back 1 I Figure 2 2 ( b) is the correlation between the predicted 値 and the measured 图. Fig. 0 1 1 Fig. 2 3 ( a ) is the etch rate when the SNV treatment is performed using the optical data and the tracking information 1 1 I material. Figure 2 3 (b) of the prediction 値 and the tube re-measurement Related to the actual measurement 本页 Nw^ 嗓1 〇 1 I Figure 2 4 ( 3 ) is the etch rate when the tracking data of the operating data is used to perform the processing of the 1 刖 V Predicted 値 and measured 値 map » 1 1 Figure 2 4 ( b ) is the correlation between the predicted 値 and the measured 〇 1 Figure 2 5 ( a ) shows the use of the VI data in the operational data into the 1 1 row The prediction of the etch rate at the pre-processing of SNV and the 1 图 of the measured 値 Figure 2 5 ( b ) is a correlation diagram showing the predicted 値 and the measured 〇Μ 1 令 Figure 2 6 ( a ) shows the operational data Using the optical data to predict the etch rate and the tube 値 when pre-processing for Μ SC, 1 1 Figure 2 6 (b) shows the correlation between predicted 实 and measured 〇1 1 2 7 ( a ) is showing operation The data using optical data and tracking data for the prediction of uranium engraving rate using Μ SC 値 and 1 I 値 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图Fig. 1 1 Fig. 2 8 ( a ) is a graph showing the prediction of the etching rate and the measured enthalpy when the tracking data of the operating data is used for the pre-processing of the 1 SC 1 1 1 The paper scale is applicable to the Chinese country. Standard (CNS) A4 specification (210X297 mm) -11 - 1306269 A7 __ B7 V. Description of invention () 9 Figure 2 8 (b) is a correlation diagram showing the predicted 値 and the measured 値. Figure 29 (a) shows the use of the VI probe data for the operation data, and enters (please read the precautions on the back side and then fill out this page). The prediction of the engraving rate when using the pre-processing of the MSC and the graph of the actual measurement ' 2 9 ( b ) is a graph showing the correlation between predicted 实 and measured 値. Fig. 30 is a table summarizing the prediction error P E from the experimental results in each of the graphs (a) of Figs. 14 to 29. Fig. 31 is a table summarizing the correlation coefficient R from the experimental results in each of Figs. 14 to 29 (b). Fig. 3 2 is a table summarizing the influence variables V I P of the various types of data in the tracking data on the predicted results. Fig. 3 3 ( a ) is a graph showing the prediction of uranium engraving rate and the measured enthalpy when using the data of the high frequency voltage VP p removed by the tracking data, and Fig. 3 3 (b) shows the correlation between the predicted enthalpy and the measured enthalpy. relation chart. Fig. 3 4 (a) is a graph showing the predicted 値 of the etch rate and the measured 値 when the data of the applied cumulative time of the high frequency power is removed by the tracking data, and Fig. 34 (b) shows the predicted 値 and the measured 値. Correlation diagram. Figure 3 5 (a) shows the prediction of the etch rate when using the tracking data to remove only the data of the cumulative time of the application of the v PP and the high-frequency power printed by the Intellectual Property Office of the Intellectual Property Bureau. Fig. 3(b) is a graph showing the correlation between predicted 实 and measured 値. [Description of symbols] 1 : Processing room 1 A : Upper room paper size applicable to China National Standard (CNS) A4 specification (210X297 mm) -12- 1306269 Λ7 B7 Ministry of Economic Affairs Intellectual Property Bureau employee consumption cooperative printing 5, invention Explanation (Μ ) 10 1 B : Lower chamber ic: Exhaust pipe 1 D : APC valve 2 : Lower electrode 2 A : Insulating material 3 : Support 3 A : Cold coal flow path 3 B : Gas flow path 4 '· Upper Electrode 4 B : Hole 5 : Dipole ring magnet 6 : Smell valve 7 : High frequency power supply 7 A : Integrator 7a, 7B, 9a: Power meter 7b: Accumulator 8 : Electrostatic chuck 8 A : Electrode plate 9 : DC power supply 10: Processing device 10a: Polymagnetic ring 12: Ball screw mechanism 13: Wind box 14: Cold coal piping (please read the notes on the back and fill out this page)

, 1T This paper scale applies to China National Standard (CNS) Α4 specification (210Χ297 mm) -13- 1306269 V. Invention L) A7 B7 1 5 Gas introduction mechanism 1 6 Bellows cover 1 7 Piping 1 8 , 18 ; Process gas Supply system 1 8 A , 1 8 A : C 4 F ί 3 gas supply ί source 1 8 D , 1 8 D - 0 2 gas supply source 1 8 G , 1 8 GA r gas supply source 1 8 B , 1 8 E , 1 8 Η , 1 8 B /, 1 8 Ε ^ Η - Valves 1 8 C , 1 8 F , 1 8 I , 1 8 C / , 1 8 F ^ (Please read the notes on the back and fill out this page ) 1 8 I > : Mass Flow Controller

1T 20: Optical measuring device 5 0: Multivariable analysis device 5 1: Operation data storage unit 5 2: Processing result data storage unit 5 3: Multivariate analysis program storage unit 5 4: Multivariate analysis processing unit Economic Department Intellectual Property Bureau employee consumption cooperative printing 5 5: Multivariate analysis result storage unit 60: Input/output device 100: Processing device [Detailed description of preferred embodiment] Hereinafter, a plasma uranium engraving device to which the present invention is applied will be described in detail with reference to the accompanying drawings. The first embodiment of the case of the prediction method of the processing result. This paper scale applies to the Chinese National Standard (CNS) Α4 specification (210Χ297 mm) -14- 1306269 Λ7 B7 V. Inventive Note (12) First, the plasma etching apparatus of the first embodiment will be described as a magnetron reaction. A saturating processing device (hereinafter referred to as [processing device 10]). As shown in FIG. 1, the processing apparatus 1 includes, for example, a processing chamber 1 made of aluminum, and a support body 3 made of aluminum which can be lifted and lowered by the insulating material 2A to support the lower electrode 2 disposed in the processing chamber 1. a jet head that is above the support 3 and supplies process gas and has an upper electrode (

Shower head) (hereinafter also referred to as [upper electrode] as needed) 4 〇 The upper portion of the processing chamber 1 is formed by a small-diameter upper chamber 1 A, and the lower portion is formed by a large-diameter lower chamber 1 B. The upper chamber 1 A is surrounded by a dipole ring magnet 5. In the dipole ring magnet 5, a plurality of anisotropic segment magnets are housed in a casing made of a ring-shaped magnetic body, and the entire upper chamber 1A is formed in the same direction in one direction. Horizontal magnetic field. An inlet and outlet for transferring the incoming wafer W is formed in the upper portion of the lower chamber 1 B, and a gate valve (Gate valve) 6 is attached to the inlet and outlet, and a high-frequency power source 7 is connected to the lower electrode 2 via the integrator 7 The frequency power source 7 applies a high frequency power of 13.56 Μ Η z to the lower electrode 2 to form an electric field in the vertical direction between the upper chamber 1 and the upper electrode 4. The integrator 7A includes a measuring device (not shown) for measuring the high-frequency (RF) voltage V pp on the lower electrode 2 side (the output side of the high-frequency voltage) 〇 on the integrator 7 A and the lower electrode 2 side (high The output side of the frequency power is connected to the power meter 7 B. The power meter 7 B is measured from the high-frequency power source. The paper size applies to the Chinese National Standard (CNS) VIII 4 specifications (210X297 mm) (please read the back note first and then fill in this page) %? Ministry of Economics intellectual property Bureau employee consumption cooperative printing -15-

1306269 AV B7 V. Inventive Note (13) 7 high frequency power P. In the upper chamber 1A, a magnetic field is generated by the electric field generated by the high-frequency power source 7 and the horizontal magnetic field generated by the dipole ring magnet 5 in the upper chamber 1A, and a power award is generated for the process gas supplied into the upper chamber 1A. A static chuck 8 is disposed on the top surface of the lower electrode 2, and a DC power source 9 is connected to the electrode plate 8 A of the electrostatic chuck 8. Therefore, the wafer W is electrostatically adsorbed by the electrostatic chuck 8 by applying a high voltage to the electrode plate 8 A by the DC power source 9 under a high vacuum. On the outer circumference of the lower electrode 2, a collecting ring 10a is disposed, and the plasma generated in the upper chamber 1A is concentrated on the crystal pattern w, and is disposed on the lower side of the collecting ring 10 a. The exhaust ring 11 of the upper portion of the support body 3. A plurality of holes are formed at equal intervals in the circumference of the exhaust ring 11 over the entire circumference, and the gas in the upper chamber 1 A is discharged to the lower chamber 1 b through the holes. The support 3 can be raised between the upper chamber 1 A and the lower chamber 1 B via a ball screw mechanism 1 2 and a bellows 13 . Therefore, in the case where the wafer W is supplied onto the lower electrode 2, the lower electrode 2 is lowered to the lower chamber 1b via the support 3, and the open gate valve 6 is supplied to the lower electrode 2 via a transfer mechanism (not shown). . A cold coal flow path 3 A ' connected to the cold coal pipe 14 is formed inside the support 3 to circulate the cold coal in the cold coal flow path 3 A via the cold coal pipe 14 to adjust the wafer W to a predetermined temperature. . In the support 3, the insulating material 2 A, the lower electrode 2, and the electrostatic chuck 8, a gas flow path 3 B ' is formed in the gap between the electrostatic chuck 8 and the wafer w by the gas introduction mechanism 15 via the gas pipe 15 A. The predetermined paper size is applicable to the China National Standard (CNS) A4 specification (21〇χ297 public) (please read the note on the back and fill out this page). Printed by the Ministry of Economic Affairs, Intellectual Property Bureau, Staff Consumer Cooperative Printed-16- 1306269 at B7 V. INSTRUCTION DESCRIPTION (,) 14 Force supply of He gas as back side gas 'The heat conductivity between the electrostatic chuck 8 and the wafer W is improved by the He gas. The pressure detecting pressure sensor (not shown) of the back side gas displays the detection 値 on the pressure gauge 15B. In addition, 16 is a bellows cover 16. A gas introduction portion 4A' is formed on the top surface of the air jet head 4. Here, the gas introduction portion 4A is connected to the process gas supply system 18 via a pipe 17. The process gas supply system 18 has a C 4 F 8 gas supply source 1 8 A, a 0 2 gas supply source 1 8 D, and an Ar gas supply source 1 8 G. These gas supply sources 1 8 A, 1 8 D, and 1 8 G are via valves 1 8 B, 1 8 E, 1 8 Η, and Mass flow controllers 1 8 C, 1 8 F, 1 8 I, respectively. Each gas is supplied to the air jet head 4 at a predetermined flow rate, and is internally adjusted with a mixed gas having a predetermined mixing ratio. The plurality of holes 4 B are disposed uniformly over the bottom surface of the air jet head 4, and the mixed gas is supplied from the air jet head 4 to the upper chamber 1 A as a process gas via the holes 4 B. Further, in Fig. 1, 1 C is an exhaust pipe, and 19 is an exhaust system constituted by a vacuum pump or the like connected to the exhaust pipe 1 C. For example, as shown in Fig. 2, the processing device 10 includes a multivariate analysis device 50 that statistically processes the operation data and the processing result data, and an input/output device 60 that inputs the processing result data and outputs information such as the analysis result. The processing device 1 performs multivariate analysis of the operation data and the processing result data via the multivariate analysis device 50, and obtains the correlation between the two, and outputs information such as the analysis result to the input/output device 60 as needed. As shown in Figure 2, the analysis device 50 has the operating data storage paper scale applicable to the Chinese National Standard (CNS) Α 4 specification (210X297 mm) (please read the back note and fill out this page), ?? Intellectual Property Office Staff Consumer Cooperative Printed -17-1306269 at B7 V. Invention Description (π) 15 Storage Unit 5 1. Processing Result Data Storage Unit 5 2. Multivariate Analysis Program Storage Unit 5 3. Multivariate Analysis Processing Unit 5 4 and a multivariate analysis result storage unit 5 5 . The operation data storage unit 51 constitutes a mechanism for storing operation data, and the processing result data storage unit 52 constitutes a mechanism for storing processing result data. The multivariate analysis processing unit 54 constitutes a mechanism for predicting the processing result based on the correlation with the mechanism for obtaining the correlation between the operation data and the processing result data (for example, the prediction formula and the regression equation). The multivariate analysis result storage unit 5 5 constitutes a mechanism for storing the correlation obtained by the multivariate analysis processing unit 54. The multivariate analysis device 50 may be configured by, for example, a microprocessor that operates in accordance with a program from the multivariate analysis program storage unit 53. The operation data storage unit 5 1 , the processing result data storage unit 5 2 , and the multivariate analysis result storage unit 5 5 may be configured by a recording unit such as a memory, and each memory area may be disposed in a recording mechanism such as a hard disk. The composition is also ok. The multivariate analysis device 50 stores the data in the operation data storage unit 51 and the processing result data storage unit 52 by inputting the operation data and the process characteristic data, and then stores the data and the multivariate analysis program storage unit 53. The program is taken out to the multivariate analysis processing unit 504, and the multivariate analysis processing unit 54 performs multivariate analysis of the operation data and the process characteristic data, and stores the processing result in the multivariate analysis result storage unit 55. Here, the operation data means the detection data obtained by the plurality of measuring devices attached to the processing device 10 when the wafer W is processed, and the processing paper size is applicable to the Chinese National Standard (CNS) A4 specification (210×297 mm). (Please read the notes on the back and fill out this page.) Ordered by the Ministry of Economic Affairs, Intellectual Property Office, Staff and Consumer Cooperatives -18- 1306269 Λ7 B7 V. Invention Description () 16 The data means the result of processing the wafer W. The process characteristic data of the obtained wafer W and the device state data regarding the state in the processing chamber 1. The operation data is intermittently measured between the processing wafers W, and the processing result data is measured as needed after the processing of the wafer. These measurement results are stored in the respective storage sections 5 1 and 5 2 . In the first embodiment, the relationship between the operational data and the processing result data is obtained, and the information on the operational data that is likely to affect the processing result is preferable. In the first embodiment, the temperature of the plurality of positions in the processing chamber 1 and the pressure of the back side gas and the electrical data of the processing apparatus 10 are used. In the process characteristic data in the processing result data of the first embodiment, for example, the etching amount of the yttrium oxide film containing the wafer W having the yttrium oxide film on the surface or the etching of the in-plane uniformity of the amount of the etched amount is used. . The device state data in the processing result data can be used to display information on the device state including the deposition film thickness of the by-product in the processing chamber 1 and the consumption amount of the member such as the magnetism collecting ring 10 a. In the first embodiment, process characteristic data among the processing result data is used, in which the in-plane uniformity of the amount of removal of the wafer W is also used. The temperature in the processing chamber 1 uses the temperature of the gas jet head 4 of the upper electrode, the temperature of the inner wall surface of the processing chamber 1, and the temperature of the lower electrode 2 in the first embodiment. In particular, the influence of the temperature of the lower electrode 2 is large. These temperatures can be measured by a known temperature sensor (not shown) such as a thermocouple disposed in each part. More specifically, the temperature in the processing chamber 1 is used to treat the average temperature between one of the above-mentioned crystal grains in each of the above-mentioned portions. The pressure in the processing chamber 1 can be used, for example, the process gas in the processing chamber 1. The paper size is applicable to the Chinese National Standard (CNS) A4 specification (210×297 mm) (please read the notes on the back and fill in the page) Property Bureau Staff Consumer Cooperative Printed -19- 1306269 A7 _ B7 V. Inventive Note () 17 Pressure or Hee gas pressure of the back side gas. In the first embodiment, the pressure in the processing chamber 1 uses the pressure of the back side gas. The electric data of the processing device 10 can be, for example, a fundamental wave of a high-frequency power applied by the high-frequency power source 7, a voltage of a harmonic, a current, a phase 'impedance, or the like. In the first embodiment, the high-frequency voltage (R F voltage) V p p on the output side of the integrator 7 A measured by a measuring device (not shown) in the integrator 7 A is used. The high-frequency voltage vp p is, for example, as shown in Fig. 7, and can be reflected in the prediction 値 even if there is a momentary large fluctuation in the vicinity of 60 hours. The in-plane uniformity of the amount of the cerium oxide film to be removed from the wafer W used in the process characteristic data is measured by, for example, the film thickness of the yttrium oxide film in the surface of the wafer w before and after the treatment. The in-plane uniformity of the error between the film thicknesses is obtained. The in-plane uniformity is determined by (the minimum 値 of the maximum 値 measured 値) / (the average 値 of 2 X measured 値). In the first embodiment, the multi-variable analysis device 50 uses a multivariate analysis program to obtain a plurality of types of operation data as explanatory variables (description variables), and process characteristic data. The explanatory variables (destination variables and destination variables) are the following (1) relational expressions (predictive equations such as regression equations, and models). In the regression equation of (1) below, X means a matrix describing variables, and Y means a matrix of variables to be explained. Further, B is a regression matrix composed of coefficients (weights) of explanatory variables, and E is a residual matrix (Residual matrix). Y - BX + E ··· ( 1 ) This paper scale applies to the Chinese National Standard (CNS) A4 specification (210X297 mm) -20- 1306269 A7 B7 V. Invention Description (18) In the first embodiment, For the above (1), use for example (please read the note on the back and then fill out this page). PLS (partial least squares) shown in JOURNAL OF CHEMOMETRICS, V0L.2 (p P 2 1 1 - 228) (1998) ,

Partial Least Squares). This P L S method has a plurality of explanatory variables and explanatory variables for each of the matrices X and Y, and a relationship between X and Y can be obtained if there are a few actual measured enthalpy. Moreover, even with a relationship obtained with a small number of measured flaws, stability and high reliability are characteristics of the P L S method. The multi-variable analysis program storage unit 53 stores the program for the PLS method, and the multi-variable analysis processing unit 54 processes the operation data and the process characteristic data in the order of the program, and obtains the above (1) formula for multivariate analysis. The result storage unit 55 stores the result. Therefore, when the above formula (1) is obtained in the first embodiment, the process characteristics can be predicted by inserting the operation data into the matrix X as an explanatory variable. Moreover, this prediction is a reliable one. The Ministry of Economic Affairs Intellectual Property Office employee consumption cooperative prints, for example, the ith principal component corresponding to the i-th inherent 値 of the X τ Y matrix ′ is represented by t i . When the matrix X uses the score t 1 and the vector P i of the i-th principal component, the following equation (2) can be expressed. If the matrix Y uses the score ti and the vector c ^ of the i-th principal component, the following can be used ( 3) Expression. In addition, in the following formulas (2) and (3), + i and ¥1 + 1 are the remainder matrices of Y and Y, and are further transposed matrices. The following index T means the transposed matrix. X= tiPi+ t2P2+ t3p3+··· tiPi + Xi+1 ··· (2) This paper scale applies to Chinese national standards (CNS > eight 4 traces (210X297 public) -21 - 1306269 Λ7 Α7 Β7 V. Invention description ( 19 Y= tlCl+ t2c2+ t3c3 + ... tiCi + Yi + i (3) Further, the PLS method used in the first embodiment calculates the above equations (2) and (3) with a small amount of calculation. The complex inherent 値 and the eigenvectors of the related cases. The PLS method is implemented in the following order: First, the centering and scaling operations of the matrix X, Y are performed in the first stage. Set i = 1 and let Χι = Χ, Yi = Y. And ' U i sets the first column of matrix Y 1. In addition, centering refers to the operation of deducting the average 値 of each column from each column of each column, scaling It refers to the operation (processing) of dividing each column of each column by the standard deviation of each column. After finding Wi = XiTUi / ( Ιΐ / ΐΐί) in the second stage, normalize the determinant of w ί and find ti = X Iwi. Also, the same processing is performed for the matrix Y. After finding C i = YiT ti / ( ti T ti ), Of c: the determinant obtained u i = Y i c i / C c i τ c i). In the third stage, the X load (Loading) pi = XiTti / C t * T t i ) and the Y load Q.sYiTui/CUiTui are obtained. Further, b i = u i T t , / ( t ' T t ,) which returns u to t is obtained. Next, the residual matrix Xi = Xi - tlPlT ' residual matrix Yi = Y1 - bitlClT is found. And 'increasing the i setting i = i + 1 , repeating the processing from the second stage. Repeat these processes until the program according to the PLS method satisfies the predetermined stop condition or converge the residual matrix Χ1 + ι to zero to find the maximum inherent paper size of the residual matrix. The Chinese National Standard (CNS) Α4 specification is applied. (210X297 mm) (Please read the note on the back and fill out this page) Ordered by the Ministry of Economic Affairs, Intellectual Property Office, Staff Consumer Cooperatives -22- 1306269 A7 B7 V. Invention Description (2Q) and its inherent vector. The P L S method converges to the stop condition of the residual matrix X i + i or zero. The repetition of the calculation of only about 10 times, the residual matrix converges to the stop condition or zero. The remainder of the matrix is typically converge to a stop condition or zero by a repetition of 4 to 5 calculations. Using the maximum intrinsic 求出 obtained by this calculation process and its eigenvector, the first principal component of the χτγ matrix is obtained, and the maximum correlation between the X matrix and the γ matrix can be known. Next, an embodiment of the operation of the processing apparatus 10 and the method of the present invention will be described. In the first embodiment, the predetermined wafer W is processed by first obtaining the above formula (1) for predicting the process characteristics by multivariate analysis. At the processing stage of the wafer W, the process characteristics at that time can be predicted by inserting the operation data at any time point into the equation (1). When the operation of the processing apparatus 10 is started, the support 3 is lowered to the lower chamber 1 of the processing chamber 1 via the ball screw mechanism 12, and the gate valve 6 is placed on the lower electrode 2 by the open inlet and outlet wafer W. After the wafer W is transferred, the gate valve 6 is closed and the exhaust system 19 is operated to maintain the predetermined degree of vacuum in the processing chamber 1. At this time, the gas introduction mechanism 15 supplies He gas as the back gas, and the wafer W and the lower electrode 2 are specifically made to have thermal conductivity between the electrostatic chuck 8 and the wafer W, thereby improving the cooling efficiency of the wafer W. On the other hand, the process gas supply system 18 supplies C 4 F 8 gas at a flow rate of 16 sccm, and supplies 〇2 gas at a flow rate of 300 S ccm, and also supplies A r at a flow rate of 40 sc cm. gas. The pressure in the processing chamber 1 at this time is, for example, 5 3 m T 〇 r r . In this state, if the high-frequency power supply 7 applies high-frequency power at 190 W, the Chinese National Standard (CNS) A4 specification (210 X 297 mm) is applied to the dipole paper scale (please read the back of the cabinet first) Fill in this page) Order Ministry of Economic Affairs Intellectual Property Bureau Staff Consumer Cooperative Printed -23- 1306269 A7 B7 V. Invention Description L) The function of the ring magnet 5 combines to cause the magnetron discharge to occur, and the plasma of the process gas is generated to etch the crystal. An oxide film of a circle W. After the end of the etching, the processed wafer W is transferred from the processing chamber 1 by the operation opposite to the incoming, and the same processing is repeated for the subsequent wafer W. The predetermined number of processing is processed, and the processing is completed. In one embodiment, before the actual wafer W is processed, the actual wafer W is mixed with the same 6 wafers W and 19 virtual wafers (

Dummy wafer ) 2 5 wafers are batch (L 〇t ), treated with 3 [min/min] processing time 'repeated 1 1 batch every 10 hours or 5 hours to get Multi-variable analysis is performed on the operation data and process characteristics of six wafers W. In the first embodiment, the PLS method can be completed by using a small number of data. Therefore, only the operation data and the process characteristic data of the wafer W in the second and the first batches are used, and the above-mentioned data is obtained from the data using the PLS method. (1) Formula. Further, six wafers W are inserted into the first, third, fifth, first and second sheets of each batch. During the processing of the wafer W, the temperature of each of the wall surface of the upper chamber 1 a and the lower electrode 2 of the jet head (upper electrode) 4 and the processing chamber 1 is intermittently detected as operation data, and these detection signals τ : T 2 and T 3 are sequentially input to the multivariate analysis device 5 via the A/D converter, and stored in the lotus transfer data storage unit 51. Further, the pressure of the He gas is intermittently detected as another operation data, and the detection signal P is sequentially input to the multivariate analysis device 50 via the A/D converter, and based on these inputs, the multivariate analysis processing unit 5 calculates This paper scale applies to China National Standard (CNS) A4 specification (210X297 public Chu) (please read the note on the back to fill out this page) 丨0' Good Ministry of Economic Affairs Intellectual Property Bureau Staff Consumer Cooperative Printed -24- 1306269 at B7 V. Description of Invention (22) The standard deviation is stored in the operation data storage unit 51. (Please read the precautions on the back and fill out this page.) In addition, the voltage of the high-frequency power supply 7 is intermittently detected as other operation data. 'This detection signal V is sequentially input to the multivariable analysis device via the A/D converter. It is stored in the operation data storage unit 51. Then, the average value of each wafer w of each operation data other than the He gas pressure is obtained by the multivariate analysis processing unit 5 to determine the standard deviation of each wafer W of the operation data with respect to the He gas pressure. Next, the operation data storage unit 51 stores the average 値 and standard deviation ′ of each of the operation data of each wafer W or prepares it to the next process as it is. Here, the time-lapse change of the detection signal T i of the upper electrode temperature of all the wafers W, the detection signal T of the wall surface temperature, and the detection signal T 3 of the lower electrode temperature are shown in FIG. 3 . Figure 5. The time-dependent change of the standard deviation of the detection signal P indicating the He gas is shown in Fig. 6, and the time-lapse change of the average 値 of the detection signal V showing the high-frequency power is shown in Fig. 7. Printed by the Ministry of Economic Affairs, the Intellectual Property Office, and the employee's consumer cooperative, the processed wafer W is taken out, and the amount of clipping of 13 points in the plane of the yttrium oxide film of the wafer W is input to the multivariate analysis by the input/output device 6 Based on the input, the device 50 calculates the in-plane uniformity via the multivariate analysis processing unit 5, and calculates the data as the process characteristic data, and stores it in the processing result data storage unit 52. The time-dependent change showing such process characteristic data (in-plane uniformity) is shown in Fig. 8. Among the operating data and process characteristics data shown in Figures 3 to 8, the operating data and process characteristics of the second and first batches are applied to the Chinese National Standard (CNS) A4 specification (210X297 mm) by this paper scale. -25- 1306269 A7 B7 V. Inventive Note () 23 The PLS method is used to obtain the regression matrix B and the residual matrix E of the above formula (1). Further, using this formula, the process characteristic data of the wafer W in the batch and the batch other than the above batch is predicted, and a graph in which the X mark is drawn is shown in Fig. 9. Further, the graph shown by the □ mark of Fig. 9 is the actual measurement of the process characteristic data. In Fig. 9, the predictions of the second batch and the first batch of batches are the same as the actual measurement, because the wafer W at this time is used when the equation (1) is obtained. It is known that the prediction of the process characteristic data of other wafers W is also very close to the measured 値 of each batch (every 10 hours). In particular, it is possible to confirm the deterioration prediction of the uniformity in the vicinity of 60 hours. This is a reflection of the sudden decrease in the high frequency (R F ) voltage confirmed in Fig. 7 . That is, as shown in FIG. 3 to FIG. 6 , the data of the state in the processing chamber 1 such as the upper electrode temperature, the wall temperature, the lower electrode temperature, and the He gas pressure which can detect the batch fluctuation over time are reflected as shown in FIG. It is effective to detect both of the data that reflect the application state of the high-frequency power. Fig. 10 shows the correlation between the predicted 値 and the measured 値 shown in Fig. 9. This correlation is clearly shown in Fig. 1 若 If the correlation coefficient R = 0 · 9 0 5 3 is high, the correlation is high, and the prediction 一致 is consistent with the measured 値. Further, in the first embodiment, the comparison of the predicted enthalpy and the actual measurement 値, the actual measurement of the operation data and the process characteristic information relating to all the wafers W are shown in Figs. 3 to 9 . Further, when the in-plane uniformity of the amount of removal of the wafer W in the process characteristic data of the first embodiment is predicted by the results of the experiment, the average 値 of each wafer W at the lower electrode temperature is obtained. The standard deviation of each wafer w of the H e gas pressure is important for the use of operational data to improve prediction accuracy. This paper scale is applicable to China National Standard (CNS) Α4 specification (210X297 mm) (please read the note on the back and fill out this page), τ Ministry of Economic Affairs Intellectual Property Bureau employee consumption cooperative print -26- 1306269 at B7 V. OBJECT OF THE INVENTION (24) (Please read the precautions on the back side and fill in the page again.) In this embodiment, before processing the actual wafer w, a wafer is used in the same wafer W (in the first embodiment). In the first and second batches of the first and second batches, the operating data and the process characteristics data were obtained as described above. Using these few operating data and process characteristic data, the regression equation (1) is obtained by the P L S method, and the operation data of an arbitrary wafer W is detected when the actual wafer W is processed. Moreover, the in-plane uniformity of the actual wafer W can be predicted as process characteristic data only by inserting each operation data into the regression equation (1). Moreover, process prediction with extremely high precision can be performed. According to the above description, the Ministry of Economic Affairs Intellectual Property Office employee consumption cooperative prints the operation data and processing result data (for example, process characteristic data) when processing a small number of test wafers like a predetermined batch of wafers according to the first embodiment, according to The collected data group (operation data and processing result data) is subjected to multivariate analysis, and the correlation between the operation data and the processing result data is obtained through the multivariate analysis, and the correlation is used to predict the amount of wafer shaving. The processing result of the wafer W such as internal uniformity (for example, process characteristics), when the wafer W is actually processed, the in-plane uniformity of the wafer W can be accurately predicted only by obtaining the operation data of the wafer W. As a process feature. Further, when the correlation between the operation data and the processing result data is obtained by multivariate analysis, the regression equation (1) can be efficiently obtained in a short time by using the P L S method. Therefore, according to the first embodiment, it is not necessary to fabricate a plurality of test wafers as in the prior art, or to process a plurality of test wafers using the processing apparatus 1 to measure the respective processing results, and to test the fabrication of the test wafers and the determination of the processing results. A lot of work and time. Moreover, the processing result can be predicted with high precision than the conventional prediction method. This paper scale is applicable to China National Standard (CNS) A4 specification (210X297 mm) -27- 1306269 A7 B7___ V. Invention description () 25 (Please read the note on the back and fill in this page) Again, if you follow the first In the embodiment, the data that easily affects the process characteristic data (in-plane uniformity of the wafer W), that is, the temperature at the plural position in the processing chamber 1 (the upper electrode temperature 'the wall surface temperature of the processing chamber 1 and the lower electrode) are used. Temperature), pressure in the processing chamber (back gas pressure such as He gas), and electrical data (voltage of high frequency power), so the correlation between the operating data and the process characteristic data is strong, and the process characteristics can be predicted with high precision. Further, since the process characteristic data uses the in-plane uniformity of the wafer W, it is possible to accurately predict the uniformity of the in-plane of the wafer W to be etched. Further, in the first embodiment, the test wafers of the second batch and the first batch are used, and the correlation between the measured enthalpy and the predicted enthalpy is obtained, but in the case where the correlation is obtained, the actual process is processed between the wafers W. The test wafer may be processed periodically to determine the correlation, or the test wafer may be processed irregularly to determine the correlation. After finding a correlation, the prediction can be further improved by using the appropriate test wafer to update the correlation, {/ by [-rV? ^ degrees. In the first embodiment, the operation data of the first embodiment is used, and the temperature of the upper electrode, the wall temperature of the processing chamber, and the temperature of the lower electrode are used, but the temperature of other parts that easily affect the process characteristics may be used. It is better if you use the temperature of at least one position. In particular, when the process characteristic data predicts the in-plane uniformity of the amount of chipping of the wafer w, the lower electrode temperature is preferable. Further, although the pressure in the treatment chamber is used, the pressure of the process gas is used. In particular, the process characteristic data. When predicting the in-plane uniformity of the amount of wafer W to be removed, the paper scale using the pressure indicating the He gas is applicable to the Chinese National Standard (CNS) A4 specification (210×297 mm) -28- 1306269 A7 B7 V. INSTRUCTION DESCRIPTION k) The standard deviation of stability is better, and the in-plane pressure difference of the wafer W using He gas (for example, the pressure difference when the back gas is centered, centered, and edged) Etc.) Better. Further, in the electric data of the operation data of the first embodiment, the voltage of the high-frequency power source is used, but the fundamental wave of the high-frequency power source, the current of the harmonic wave, the phase, and the impedance may be used. Further, in the first embodiment, the processing result data uses the process characteristic data, and although the process characteristic data uses the in-plane uniformity of the amount of the wafer W to be removed, other process characteristic data except the amount of wafer W is removed. It is also possible to use an etching characteristic such as a line width or a taper angle which shows an etching pattern. Next, a second embodiment in the case where the prediction method of the processing result of the plasma etching apparatus of the present invention is applied will be described in detail with reference to the accompanying drawings. In the second embodiment, the same portions as those in the first embodiment are denoted by the same reference numerals, and detailed description thereof will be omitted. In the second embodiment, the parameter as the operation data is changed or added, and the processing result data is subjected to multivariate analysis using the amount of removal of the wafer W (e.g., etching rate) in the process characteristic data, and the etching rate of the wafer W is predicted. First, the plasma etching apparatus according to the second embodiment will be described with reference to Fig. 11 for a magnetron reactive etching processing apparatus (hereinafter referred to as "processing apparatus 100"). Further, the same reference numerals are given to the same portions in Fig. 11. The detailed description is omitted. In the jet head 4 of the processing apparatus 100 shown in Fig. 11, a spectrometer (hereinafter referred to as "optical meter") for detecting plasma light in the processing chamber 1 is disposed. The specific wave paper size obtained by this optical meter 20 is applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm) (please read the note on the back and then fill out this page) Employee Consumption Cooperative Printed -29- 1306269 at B7 V. INSTRUCTION DESCRIPTION g) The luminescence intensity of a long range (eg, 2 0 0 to 95 nm) is optical. The process gas supply system 1 8 / is connected to the gas introduction unit 4 A via a pipe 17 . The process gas supply system 1 8 > has a C5F8 gas supply source 1 8 A , a 0 2 gas supply source 1 8 D , and an Ar gas supply source 1 8 G /. These gas supply sources 18A >, 18D>, 18G> are supplied at predetermined flow rates via valves 1 8 B / , 1 8 E >, 18 Η - and mass flow controllers 18C -, 18F > 18 1 / respectively. Each of the gases is supplied to the air jet head 4, and is internally adjusted with a mixed gas having a predetermined mixing ratio. Each gas flow rate can be detected by each of the mass flow controllers 1 8 C >, 1 8 F /, 1 8 I. In the second embodiment, the gas flow rate of the C5F8 gas and the gas flow rate of the Ar gas are detected among the respective gas flows. The data of these gas flows being detected is made into tracking data. An APC (Auto Pressure Controller) valve 1 D is disposed in the exhaust pipe 1C, and the opening of the A P C valve is automatically adjusted with respect to the pressure in the process chamber 1. In the second embodiment, the A P C opening degree by the A P C valve 1 D is detected. The detected A P C opening is made into a tracking data. A power meter 9a for detecting an applied current of the electrostatic chuck 8 and applying a voltage is connected between the electrode plate 8A of the electrostatic chuck 8 and the DC power source 9. The data of the applied current and the applied voltage of the electrostatic chuck 8 detected by the power meter 9a is made into tracking data. The gas introduction mechanism for introducing the back side gas (for example, He gas) is 5 The paper size is applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm) (please read the notes on the back and fill out this page) Intellectual Property Bureau employee consumption cooperative printing -30- 1306269 Λ7 B7 V. Invention description ς. 28 is provided with, for example, a mass flow controller (not shown) by which the gas flow rate of the back side gas is detected by the mass flow controller. The gas flow rate of the back side gas is made into a tracking data together with the gas pressure of the back side gas detected by the pressure gauge 15B. The above integrator 7A includes, for example, two variable capacitors C1, C2, a capacitor C, and a coil. L, impedance integration is obtained via variable capacitors C 1 , C 2 . The position of the variable capacitors C 1 and C 2 in the integrated state is made into tracking data. The integrator 7 A is provided with a power meter 7 a by which the voltage V d c between the supply line (wire) of the high-frequency power and the ground of the processing device 1 计 is measured. The voltage V d c between the high-frequency power supply line (wire) and the ground is made into tracking data. An electric measuring device (for example, a VI probe) 7 C is attached to the lower electrode 2 side of the integrator 7 A (the output side of the high-frequency voltage), via which the electric measuring device 7 C is applied to the lower electrode 2 The frequency power P is based on the fundamental wave of the plasma generated in the upper chamber 1 A (the wave of the high-frequency power and the reflected wave) and the high-frequency voltage V of the harmonic, the high-frequency current I, the high-frequency phase P, Impedance Z is used as electrical data detection. Among them, the wave of the high-frequency power and the reflected wave are made into tracking data. Further, the high frequency voltage V, the frequency current I, the local frequency phase P, and the impedance Z of the higher harmonics are made into V I probe data. An integrating portion 7b that accumulates the application time of the high-frequency power is connected between the high-frequency power source 7 and the power meter 7B. The cumulative time of application of the high frequency power detected by the accumulating portion 7b is also created as tracking data. The so-called application of this paper scale applies the Chinese National Standard (CNS) A4 specification (210X297 mm) (please read the back note on the front page and then fill out this page), 11 Ministry of Economic Affairs Intellectual Property Bureau employee consumption cooperative print -31 - Ministry of Economic Affairs, Intellectual Property Bureau, Staff Consumer Cooperatives, Printing 1306269 Λ7 Β7 V. Invention Description L) The cumulative time is the cumulative time for applying high frequency power to the wafer W each time. The accumulating unit 7b resets the application time for resetting the high-frequency power by zero for each maintenance of the processing apparatus 1 。. Therefore, the cumulative application time of the high-frequency power referred to here is the cumulative application time until the next maintenance. The above-described maintenance can be used, for example, to remove wet cleaning, consumables, or measuring instrument exchanges such as by-products (e.g., particles) in the processing apparatus 1000 caused by etching. In the second embodiment, each time the wet cleaning is performed for resetting, the applied cumulative time is zero. Next, a block diagram of the multivariate analysis device 5 in the second embodiment is shown in Fig. 12. In the second embodiment, the operation data detected by each measuring device is divided into optical data, tracking data, and V I probe data. The optical data uses the luminescence spectral intensity of a wavelength range of, for example, 2 0 0 to 950 nm which is detected by the optical measuring device 20 described above. The tracking data is in addition to the temperature (upper electrode temperature T i , wall surface temperature Τ 2, lower electrode temperature T 3 ) in the processing chamber 1 described in the first embodiment, the pressure of the back side gas, and the output of the integrator 7 A. In addition to the information on the side of the frequency (RF voltage) V pp, the following information is added. Gas flow rate of C 5 F 8 gas and Ar gas to be treated gas, flow rate of back side gas, AP C opening degree by AP C valve 1 D, applied current of electrostatic chuck 8 and applied voltage, integrator 7 A The position of the variable capacitors C 1 and C 2 in the middle, the voltage V dc between the high-frequency power supply line and the ground in the integrator 7 A, the wave of the high-frequency power, and the reflected wave paper are only used in the Chinese national standard ( CNS ) A4 size (210X297 mm) ~~一-32- (Please read the note on the back and fill out this page again.) Book 1306269 at B7 V. Invention description) (Please read the notes on the back and fill out this page. ) The data and the cumulative time of application of high-frequency power are added to the tracking data. Further, the pressure and flow rate of the back side gas use, for example, the flow rate in the center and the edge of the wafer w. The V I probe data uses a high-frequency voltage V of a harmonic, a high-frequency current I, a high-frequency phase P, and an impedance. The process characteristic data uses the amount of wafer W removed. Specifically, the amount of dicing of the wafer W of the wafer W is etched (A/ mi η ) at the time of uranium etching of the CVD oxide film formed by, for example, CVD (Chemical Vapor Growth) on the surface of the wafer W. data. In the multivariate analysis device 50 according to the second embodiment, for example, optical data or the like is used as the explanatory variable, and the etching rate of the wafer W in which the processing result data is one of the process characteristic data is described as a variable (destination variable) The regression equation ((1) relational expression) described in the first embodiment described above is obtained using, for example, a multivariate analysis program for the PLS method. Then, the operation data is input to the regression equation obtained, and the etching rate of the wafer W is predicted. In the multivariate analysis processing unit 5 of the second embodiment, the multivariate analysis processing unit 54 of the second embodiment performs the multivariate analysis such as the calculation of the relational expression (regressive equation) of (1), and is used for the operational data. And processing the result data for pre-processing. The pre-processing is used, for example, to selectively perform any of 0 S C (Orthogonal Signal Correction), MSC (Multiplicative Signal Correction), and SNV (Standard Normal Variate Transformation). The pre-processing using the above 0 S C is generally a pre-processing of removing the unrelated component (component perpendicular to γ) from the objective variable Y by the explanatory variable X. In addition, the details of the pre-processing using the above 0 S C are disclosed, for example, in the paper scale applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm) -33-1306269 at B7 V. Invention description k)

Wold, et al.. (1 99 8a), Orthogonal Signal Correction of Near-Infrared Spectra, Chemometrics and Intellegent Laboratory Systems, 44, 1 75- 1 8 5 ° Pre-treatment with the above SNV is generally used to calibrate the sample (Sample) Here, the influence of the error of the operation data and the processing result data of each wafer W is performed before each sample is subjected to the reference in the data direction. Specifically, the pre-processing using the above S NV is to correct each sample by standard deviation, for example. Further, details of the pre-processing using the above SNV are disclosed, for example, in Barnes, et al... (1989), Standard Normal Variate Transformation and De-trending on Near-Infrared Diffuse Reflectance Spectra, Applied Spectroscopy, 43, 772-777 ° The pretreatment using the above MSC is generally a pretreatment for correcting the sample by obtaining a desired spectrum and making the dispersion between the samples smaller. Specifically, the pretreatment using the above M S C is, for example, that each sample is averaged in the wavelength direction (ideal spectrum), and a linear regression line with the ideal spectrum is calculated for each sample. The slope and slice obtained from the linear regression line are used to correct the data for each sample. In addition, the details of the pre-processing using the above S Ν V are not disclosed, for example, in Gelad, et al. (1985), Linearization and Scatter-infrared Reflactance Spectra of Meat, Applied Spectroscopy 3, 491-500 °. The operation of the processing device 100. When the operation of the processing device 100 is started, the detection data intermittently detected by each measuring device such as the optical measuring device 2 is sequentially input to the multivariable analyzing device 5. The paper size is applied to the Chinese National Standard (CNS) Α 4 specification. (210Χ297 mm) (Please read the notes on the back and fill out this page again)

1T Ministry of Economic Affairs Intellectual Property Bureau employee consumption cooperative printing -34- 1306269 A7 B7 V. Invention description (32). Here, the conditions at the time of the etching treatment are such that the pressure in the processing chamber is 50 m τ, the high-frequency power applied to the lower electrode is 150 〇w, and the processing gas is a mixed gas of C 5 F 8 and 〇 2 and A r . The back side gas is He gas. Next, the average value of each wafer W of each operation data is obtained by the multivariate analysis processing unit 54. Next, the operation data storage unit 5 1 stores the average value of each operation data of each wafer W, or prepares for the next process as it is. Then, the processed wafer W is taken out, and the etching rate of the CVD oxide film of the wafer W is input to the multivariate analysis device 50 by the input/output device 60, and the input is stored in the processing result data storage unit 52 as a process. Characteristic data. Further, after the pre-processing or the pre-processing, the regression equation using the PLS method (the relational expression of (1)) is obtained. Here, the relationship between the number of processed wafer W and the measured etching rate is shown in Fig. 13. In Fig. 13, WC (wet cleaning cycle) 1 is a section to the wet cleaning of the first processing apparatus 100, and WC 2 is performed by the first wet cleaning to the second time. The section 'w C 3 from the wet cleaning is the section from the second wet cleaning to the third wet cleaning, and the WC 4 is the third wet cleaning to the fourth In the section from the wet cleaning, the regression data ((1) relationship is obtained by the PLS method from the data of the wet cleaning cycle WC1 (the first to the 16th) in the operation data and the processing result data. The regression matrix B and the residual matrix E. Furthermore, using this formula, the above WC 1 and the above | (WC2 other than 3 1 (17th to 36th), WC3 (the 37th paper size applies to the Chinese National Standard (CNS) A4 specification (210X297 mm)) ' _ -- (Please read the notes on the back and fill out this page)

, ST. New Economy Ministry Intellectual Property Bureau employee consumption cooperative printing -35-1306269 A7 B7 V. Invention description () 33 ~ 47), WC4 (48th to 52nd) processing results data crystal Information on the etching rate of the circle W. (Please read the precautions on the back side and fill out this page.) A graph in which the prediction result of the etching rate of the wafer W is marked with a mark is shown in each of Figs. 14 to 29 (a). Among the figures (a), the graph indicated by the ◊ mark is the actual measurement of the data of the etching rate of the wafer W. The prediction error (PE: Prediction Error) was calculated for each of the experimental results of Fig. 14 to Fig. 29 (a). The prediction error P E is obtained by subtracting the square of the prediction 値 from the actual measurement of the data of each wafer, and dividing the number of wafers for processing the integration to obtain the square root. The prediction error P E is optimal, and the smaller the 値 is, the less the error between the measured 値 and the predicted 値

Q The relationship between the predicted 値 and the measured 値 shown in each of the graphs (a) of Figs. 14 to 2 is shown in Fig. 14 to Fig. 29 (b). The correlation coefficient R was obtained for the experimental results of each of Fig. 14 to Fig. 29 (b). The correlation coefficient R is optimal at 1, indicating that the larger the 値 is, the more relevant it is. Therefore, the closer the above prediction error P E is to ◦ and the closer the correlation coefficient R is to 1, the better the prediction accuracy. In addition, for the above experimental results, since WC 1 to WC 4 are etched under the same etching conditions, the prediction error PE and correlation coefficient are used, and w C 1 to WC are used. 4 information. However, the experimental results of the etch rate using the VI probe data as explanatory variables (Fig. 17.7, Fig. 2, Fig. 25, Fig. 29) are for experimental reasons, since only WC4 and other WC 1 to WC 3 Since the etching rate conditions are different, the prediction error PE in the experimental results of Fig. 17, Fig. 21, Fig. 25, Fig. 29 is obtained, and the paper size is applied to the Chinese National Standard (CNS) A4 specification (210X297 mm) -36- 1306269 A7 B7 V. Description of invention ς,) 34 For the correlation coefficient, use WC 1 to WC 3 data other than WC 4 data. (Please read the precautions on the back and fill out this page.) Figure 1 4 to Figure 1 7 shows the experimental results in the case of multivariate analysis using the P L S method without pre-processing. Fig. 14 is a case where the above optical material is an explanatory variable, and Fig. 15 is a case where the above optical data and the above-mentioned tracking data are explanatory variables. Figure 16 is the case where the above tracking data is used as an explanatory variable. Fig. 17 is a view showing the above V I probe data as a description of the variables. Fig. 18 to Fig. 2 1 show experimental results of the case where the pre-processing performs the multivariate analysis by the P L S method after the above 0 S C is performed. Fig. 18 is a case where the above optical data is used as an explanatory variable, and Fig. 19 is a case where the above optical data and the above-mentioned tracking data are explanatory variables. Fig. 20 is a case where the above tracking data is an explanatory variable, and Fig. 21 is a case where the above V I probe data is an explanatory variable. Printed by the Intellectual Property Office of the Ministry of Economic Affairs, the Consumers' Cooperatives. Figure 2 2 to Figure 2 shows the experimental results of the multivariate analysis using the P L S method after the above SNV is performed. Fig. 2 2 shows the case where the above optical data is used as an explanatory variable, and Fig. 23 is a case where the above optical data and the above-mentioned tracking data are explanatory variables. Fig. 24 shows the case where the above tracking data is an explanatory variable, and Fig. 25 shows a case where the above V I probe data is an explanatory variable. Fig. 26 to Fig. 2 show experimental results of the case where the pre-processing performs the multivariate analysis by the P L S method after performing the above MS C. Fig. 26 is a case where the above optical data is used as an explanatory variable, and Fig. 27 is a case where the above optical data and the above-mentioned tracking data are explanatory variables. Figure 2 8 is to make the paper scale applicable to China National Standard (CNS) A4 specification (210X297 mm) -37-1306269 B7 V. Invention description ς5) The above tracking data is for the case of explanatory variables, and Figure 29 is for the above V The probe data is the case of the variable. (Please read the precautions on the back and fill out this page.) From the experimental results in each of the above figures (a) of Figure 14 to Figure 2, the prediction error PE is found in the table shown in Figure 3, from the above figure. The experimental results obtained in the respective ffl (b) of 丄4 to Fig. 2 9 are obtained by the correlation coefficient R and are shown in Fig. 31. From the point of view of the data used in multivariate analysis, if the prediction error p E is the largest in the case of using optical data according to Fig. 30, the use of optical data and tracking data, using the v probe information The situation is small in the order in which the tracking data is used, and the situation in which the tracking material is used is small. Furthermore, 'according to Fig. 3, the correlation coefficient R is the smallest in the case of using optical data, and becomes larger depending on the case of using optical data and tracking data, the case of using V丨 probe data, and the order of using tracking data. The most common use of tracking data. The Ministry of Economic Affairs, the Intellectual Property Office, and the Employees' Cooperatives are printed. Therefore, the viewpoints of the data used in the multivariate analysis are based on the overall situation. 'The prediction accuracy depends on the use of optical data, the use of optical data and the tracking of data, and the use. V. The case of the probe data 'increased the order of the S tracking data. The prediction using the tracking data = the best, and is valid for the prediction. θ This paper scale applies to the Chinese National Standard (CNS) Α4 specification (210X297 mm). For the case of using the tracking data with the best prediction accuracy, it is more important from the perspective of the overall situation. 3 0 'Predictive error Ρ Ε If 〇 s C is removed, the case of pre-processing is smaller than the case where pre-processing is not performed. Moreover, the prediction error Ρ Ε for the case of '隹广前” is smaller in the order of 0 SC, S Ν V, M sc, pre-38 A7 B7 1306269 V. Invention description (^) (Please read the back of the note first) Matters refilling this page) The situation is minimal when dealing with MSCs. Further, if the correlation coefficient R is removed in the case of 0 S C according to Fig. 31, the case where the pre-processing is performed is larger than the case where the pre-processing is not performed. Further, the correlation coefficient R becomes larger in the order of 0 S C, S Ν V, and M S C for the pre-processing, and the pre-processing is the largest in the case of performing M S C . Therefore, in the case of using the tracking data having the best prediction accuracy, the viewpoint of the presence or absence of the pre-processing or the type of the pre-processing, if the pre-processing is removed and the 0 SC is performed, it is better than the case where the prediction accuracy is pre-processed. The pre-processing situation is good and effective. Furthermore, in the case of pre-processing, the prediction accuracy is better in the order of 0 S C, S Ν V, and M S C , and the pre-processing is optimal for the prediction accuracy of the M S C case. From the above, for the etch rate of the predicted wafer W, it is explained that the variable uses the tracking data for multivariate analysis, and then before the multivariate analysis, the pre-processing is optimal for the prediction of the M S C and is most effective. Printed by the Intellectual Property Office of the Ministry of Economic Affairs, the Consumer Cooperatives. Here, which of the above-mentioned tracking data is reviewed most affects the forecast results. For the various types of data in the tracking data, the influence variable V I P (Variable Influenceon Projection) for the prediction result is shown in Fig. 3 . The influence variable V I P is the magnitude of the influence of each explanatory variable X when the expected purpose variable Y is displayed. For example, let a be the component (Component), R be the load vector, W be the trade-off vector (Weight vector), and R2 y be the correlation coefficient of Υ, to normalize the quadratic of the influence variable VIP (W[ a] The sum of the components of X (R 2 y [ a ]) is expressed. According to Figure 3 2, the paper size of the integrator 7A affecting the variable VIP is applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm) -39-1306269 at B7 V. Invention Description (q7) The voltage (RF voltage) vpp is maximum 'second' and the application time of the high frequency power is large. Therefore, it is known that the application time of the high-frequency voltage v p P and the high-frequency power greatly affects the prediction result. Therefore, if the cumulative frequency of the tube frequency voltage V P P and the application frequency of the remaining frequency is removed from the tracking data, and the etching rate of the wafer w is multivariate-analyzed, the experimental results shown in Figs. 33 to 35 can be obtained. The results of the prediction of the etching rate of the wafer W are shown by the □ mark in each of the figures (a) of Figs. 3 to 35. The figure indicated by the ◊ mark in each of the figures (a) is the actual measurement of the uranium engraving rate of the wafer w. The correlations between the predicted enthalpy and the actual measured enthalpy shown in each of the graphs (a) of Figs. 3 3 to 3 are shown in Fig. 33 to Fig. 35 (b). FIG. 3 is a case where only the data of the high-frequency voltage Vp p is removed by the tracking data, and FIG. 34 is a case where only the data of the applied cumulative time of the high-frequency power is removed by the tracking data. FIG. 35 is the use of the tracking data. Only the case of the data of the cumulative time of application of the high-frequency voltage VPP and the high-frequency power is removed. For the experimental results of each of the graphs (a) of FIG. 3 to FIG. 3, if the prediction error PE is calculated separately, it is 4 9 · 7 A/ m 1 η, 55 ' 1 A/mi η, 66 · 3 A/mi. η. Here, if the prediction error of 4 3 · 7 A/mi η is used for comparing all the data of the above-mentioned tracking data (Fig. 16 ( a )), it is known that the prediction error is in Fig. 3 3 to Fig. 5 (a). Any situation is also greater than in the case of Figure 16 (a). This paper scale applies to China National Standard (CNS) VIII 4 specifications (21〇X29? 公楚) (please read the note on the back and fill out this page). The Ministry of Economic Affairs, Intellectual Property Bureau, Staff Consumer Cooperative Printed -40- 1306269 A7 B7 Ministry of Economic Affairs Intellectual Property Bureau employee consumption cooperative printing 5, invention description () 〇〇 Next, for the experimental results of each figure (b) of Figure 3 3 to Figure 3, if the correlation coefficient R is calculated separately, it is 0 · 82, 〇. 83, 〇. 5 7. Here, if the correlation coefficient of the case of all the materials of the above-mentioned tracking data (Fig. 16 (b)) is compared with 0.9 〇, the correlation coefficient is found in Fig. 3 3 ~ Figure 3 5 (b) is also smaller than the case of Figure 16 (b). Therefore, the use of only the removal of the tube frequency voltage vp P of the tracking data is more open (Figure 3 3); use only The case where the tracking data of the application time of the high-frequency power is removed (Fig. 34): Any case where the tracking data of the high-frequency voltage vpp and the application time of the high-frequency power is removed (Fig. 36) can be used. It is confirmed that the prediction accuracy is lower than the case where all the data of the tracking data is used. It is also confirmed that the prediction accuracy of the tube frequency V pp and the high frequency power is the worst. Therefore, when predicting the uranium engraving rate of the wafer W, the tracking data has at least the high frequency voltage V pp effective. Further, it is preferable to apply the accumulated time of the high-frequency power. As described above, according to the second embodiment, the operation data and processing when processing a small number of test wafers like a wet-cycle (WC) wafer are collected. The result data (for example, the process characteristic data) is subjected to multivariate analysis based on the collected data group (operation data and processing result data). The correlation between the operation data and the processing result data is obtained by the multivariate analysis. Predicting the processing result (for example, process characteristics) of the wafer W such as the amount of wafer removal (for example, the etch rate), when the wafer W is actually processed, the operation data of the wafer w can be obtained only by high precision. Ground (please read the notes on the back and fill out this page) — Book the paper size for the Chinese National Standard (CNS) A4 specification (210X297 mm) -41 - 13062 69 Λ7 B7 V. Inventive Note ς.) Predict the amount of chipping of the wafer W (for example, etching rate) as the process characteristic. Further, when the correlation between the operation data and the processing result data is obtained by multivariate analysis, the regression equation (1) can be efficiently obtained in a short time by using the PLS method. Therefore, according to the second embodiment, it is not necessary to It is known that many test wafers are produced, or a plurality of test wafers are processed by the processing device 10 to determine the respective processing results, and it is not necessary to make a lot of work and time for the fabrication of the test wafer and the measurement of the processing results. Moreover, the processing result can be predicted with high precision than the conventional prediction method. According to the second embodiment, the high-frequency voltage V p P and the tracking data including the application time of the high-frequency power, the optical data, the VI probe data, and the like are added to the data used in the first embodiment. As the operational data, the data that is easy to influence the process characteristic data can improve the prediction accuracy of the process characteristic data. In particular, the wafer can be further improved by using a trace data of a high-frequency voltage V pp and an application time of application of high-frequency power which are easily affected by the amount of chip removal (for example, etching rate) of the wafer W as a process characteristic data. The prediction accuracy of the amount of W cut. Further, before performing multivariate analysis, the prediction accuracy of the process characteristic data can be further improved by performing predetermined pre-processing. Further, since the process characteristic data uses the etching rate of the wafer W, it is possible to accurately predict the uranium engraving rate of the wafer W to be etched. As described above, even if optical data or V I probe data is used as the operation data, the prediction accuracy can be improved. However, for example, the Chinese National Standard (CNS) A4 specification (210X297 mm) is applied to the paper scale as shown in Figure 13. (Please read the note on the back and fill out this page.) Co-operative printing -42- 1306269 A7 _ B7 V. Invention description () 40 (Please read the precautions on the back and fill out this page.) Before and after maintenance such as wet cycle, process characteristics data (such as saturation rate) In the case where the tendency of the change greatly changes, there is a case where the prediction accuracy is lowered. For example, in FIG. 14 and FIG. 17, a wet cleaning cycle (w C 3 ) other than the wet cleaning cycle (WC 1 ) of the regression equation (model) using multivariate analysis is performed, and the prediction accuracy is lowered. At this point, if the tracking data including the high-frequency voltage v PP and the application cumulative time of the high-frequency power is used as the operation data, the prediction can be improved in all the wet cleaning cycles (~WC4) as shown in Fig. 16. Precision. In particular, the cumulative time of application of the high-frequency power of the tracking data is obtained by, for example, performing maintenance such as wet cleaning, so that the cumulative application time is zero, so that the cumulative time of application of each wet cleaning cycle can be obtained. Therefore, if the cumulative time of application of the high-frequency power is used as the operation data, even if the wet cleaning is performed, the processing result information of the tendency change can be predicted with high precision. According to the present invention, a small amount of operational data and process characteristic data obtained by collecting only a small number of samples can be provided, and a predictive formula of process characteristics can be obtained, and then only The operation data when the object to be processed is processed is inserted into the prediction formula, and the prediction method of the processing result of the process characteristic can be predicted simply and accurately. The invention made by the inventors of the present invention is specifically described above, but the invention is not limited to the embodiments of the invention, and various modifications may be made without departing from the spirit and scope of the invention. These modifications and modifications are understood to fall within the technical scope of the present invention. This paper scale applies to China National Standard (CNS) A4 specification (210X297 mm) -43- 1306269 at B7 V. Invention description (41) (Please read the note on the back and fill in this page) For example, in the second implementation above The morphological processing data is used as the process characteristic data. Although the process characteristic data uses the amount of dicing of the wafer w (for example, the etch rate), other process characteristic data uses etching characteristics such as line width or taper angle of the etch pattern. Information is also available. In addition, the device state data relating to the device state such as the film thickness of the by-product in the processing chamber and the consumption amount of the component such as the magnetism ring 1 〇a is also used in the processing result data in the above-described first and second embodiments. can. By using the film thickness of the by-product, the amount of consumption of the component such as the magnetism ring 1 〇a as the device state data, it is possible to predict the cleaning period of the processing device 1 或 or the exchange of parts such as the magnetism ring 10 a. period. Further, in the first and second embodiments, the case where the wafer W is processed by etching is described, but a processing apparatus such as a film forming process other than the etching process can be applied to the present invention. Further, it is not limited to the wafer 〇 economic department of the object to be processed, and is printed by the Intellectual Property Office of the Intellectual Property Office. In the first and second embodiments, the multi-variable analysis is performed using the PLS method to obtain the regression equation ( 1) However, it is also possible to obtain the intrinsic enthalpy and its eigenvector by using a conventionally known number calculation method other than the PLS method (for example, a power multiplication method or the like). [Industrial Applicability] The present invention is applicable to, for example, a method and a processing apparatus for predicting a processing result of a processed object or a device state such as a wafer processed by a semiconductor manufacturing apparatus, and is particularly applicable to plasma etching. A plasma processing apparatus and a method of predicting the processing results in such a device. This paper scale applies to the Chinese National Standard (CNS) A4 specification (210X297 mm) -44 -

Claims (1)

1306269 A8 B8 C8 D8 VI. Patent Application Range 1 1. A method for predicting the processing result is a process of treating a treated object by plasma in a processing chamber of a plasma processing apparatus, according to the operation data and processing of the plasma processing apparatus a method for predicting a result of processing a data, comprising: a step of collecting the operation data and the processing result data; and performing a multivariate analysis step according to the collected operation data and the processing result data; a step of determining a correlation between the operation data and the processing result data; and a step of predicting a processing result by using an operation data when the processing target other than the object to be processed is obtained by the processing according to the correlation relationship; The operation data includes temperature data that is disposed in the processing chamber and that is used to mount the workpiece of the object to be processed. 2. A method of predicting a result of processing as described in claim 1 wherein the operational data further comprises information on a back gas pressure supplied between the object to be processed and the stage. Printed by the Intellectual Property Office of the Ministry of Economic Affairs, Employees' Consumption Cooperatives 3. The method for predicting the processing results as described in item 1 of the patent application scope, wherein the operational data further includes back gas pressure supplied between the object to be processed and the stage Information on the standard deviation. 4. The method of predicting a processing result according to claim 1, wherein the operation data further comprises an in-plane pressure difference of the object to be treated supplied to the back gas pressure between the object to be processed and the stage. data. 5. A method for predicting the processing results is applicable to the Chinese National Standard (CNS) A4 specification (210X297 mm) in the paper scale of the plasma processing device. 1306269 A8 B8 C8 D8 VI. Patent application scope 2 Processing chamber with plasma treatment The process of the object to be processed, the method for predicting the processing result based on the operation data of the plasma processing device and the processing result data, characterized in that: the step of collecting the operation data and the processing result data; according to the collected operation data and the a step of processing the result data to perform multivariate analysis; a step of obtaining a correlation between the operation data and the processing result data by the multivariate analysis; and, according to the correlation, using the processing to obtain the correlation relationship The operation data of the object to be processed and the step of predicting the processing result; the operation data includes at least the accumulated operation time data of the high frequency power source, and the plasma of the processing gas is generated in the processing chamber. 6. A method of predicting the processing result as set forth in claim 5, wherein the cumulative operating time of the high frequency power source is reset to zero each time the maintenance of the processing room is performed. Printed by the Ministry of Economic Affairs, the Intellectual Property Bureau, and the Consumer Cooperatives. 7. A processing device, which is a process for processing a processed object by plasma in a processing chamber, and a processing device for predicting the processing result based on the operating data and the processing result data, and is characterized by: a mechanism for storing the processing data; a mechanism for storing the processing result data; performing a multivariate analysis mechanism based on the stored operation data and the processing result data; and obtaining the operation data and the processing result capital paper by the multivariate analysis The scale applies to the Chinese National Standard (CNS) A4 specification (210X297 mm) -46- 1306269 A8 B8 C8 D8 6. The organization that applies for the patent related scope 3; the organization that stores the relevant relationship; and according to the storage The correlation relationship 'a mechanism for predicting a processing result using an operation data when the object to be processed other than the object to be processed is obtained, and the operation data includes: being disposed in the processing chamber for placing the processed Temperature data of the mounting table of the body. 8. The processing apparatus of claim 7, wherein the operational data further comprises: information on a back gas pressure supplied between the object to be processed and the stage. 9. The processing apparatus of claim 7, wherein the operational data further comprises: information on a standard deviation of back gas pressure supplied between the object to be processed and the stage. The processing device according to claim 7, wherein the operation data further includes: a data of an in-plane pressure difference of the object to be treated supplied to the back gas pressure between the object to be processed and the stage . Ministry of Economic Affairs, Intellectual Property Office, Staff and Consumer Cooperatives Printing 1 1. A processing device is a processing device that processes a processed object by plasma in a processing chamber, and predicts a processing result based on the operating data and the processing result data, and is characterized by: a mechanism for storing the operation data; a mechanism for storing the processing result data; a mechanism for performing multivariate analysis based on the stored operation data and the processing result data; and obtaining the operation data and the processing result capital by the multivariate analysis The paper scale applies to the Chinese National Standard (CNS) A4 specification (210X297 mm) -47- 1306269 Ministry of Economic Affairs Intellectual Property Bureau employee consumption cooperative printed A8 B8 C8 D8 6. The organization that applies for the patent relationship 4; And a mechanism for predicting the processing result according to the stored relationship data of the processed object other than the object to be processed that obtains the correlation relationship; the operation data includes at least: The accumulated operating time data of the frequency power supply is in the processing room Health plasma processing gases. 1. The processing apparatus according to claim 1, wherein the cumulative operation time of the high-frequency power source is reset to zero each time the maintenance of the processing chamber is performed. This paper scale applies to the Chinese National Standard (CNS) A4 specification (210X297 mm) -48 -