JP2004039952A - Plasma treatment apparatus and monitoring method thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a monitoring method of a plasma treatment apparatus which can operate the plasma treatment apparatus always in a normal state by measuring optical data accurately even after the operation for a fixed time, and to provide the plasma treatment apparatus. <P>SOLUTION: In the method, the operation condition of a plasma treatment apparatus 100 is monitored by using a plurality of detectors including an optical detector 20 for detecting optical data attached to the plasma treatment apparatus 100 by utilizing detection values of each detector detected for each workpiece. In the optical detector 20 for detecting optical data, the variation of detection values of the plasma treatment device 100 is recorded. Emission intensity from an optical source 30 provided in opposition to the optical detector 20 is detected immediately before each treatment and compared with emission intensity from the optical source 30 in the initial state of the plasma treatment device 100, and optical data are corrected. When emission intensity is attenuated to a specified value or lower, an alarm is emitted. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a method for monitoring a plasma processing apparatus and a plasma processing apparatus, and more particularly, to improving the reliability of optical data detected for monitoring the state of the plasma processing apparatus, thereby enabling more accurate status of the plasma processing apparatus. TECHNICAL FIELD The present invention relates to a plasma processing apparatus monitoring method and a plasma processing apparatus capable of grasping the condition.
[0002]
[Prior art]
In a semiconductor manufacturing process, various types of manufacturing apparatuses and inspection apparatuses are used. These processing devices have many parameters for controlling or monitoring the operation status thereof, and control or monitor them so that various processes can be performed under optimum conditions.
[0003]
The above parameters include, for example, the flow rate of a processing gas introduced into a processing chamber and the processing gas for converting the processing gas into plasma in a plasma processing apparatus that performs film formation and etching processing on an object to be processed such as a semiconductor wafer or a glass substrate. There are controllable parameters (hereinafter referred to as control parameters) such as a distance between electrodes of the electrodes installed in the room and a high-frequency power applied to at least one of the electrodes.
[0004]
Also, parameters such as optical data such as plasma spectroscopic analysis for grasping the state of the plasma excited in the processing chamber, and electrical data such as high-frequency voltage and high-frequency current of fundamental and harmonics based on the plasma (hereinafter referred to as plasma) (Referred to as reflection parameters).
[0005]
In addition, when a high-frequency power is applied to the electrodes in the processing chamber, parameters such as the capacity of a variable capacitor in a matching state of a matching unit provided for impedance matching and a high-frequency voltage measured by a measurement area in the matching unit ( (Hereinafter referred to as device state parameter).
[0006]
When performing processing, control parameters are set to values that are considered optimal, and the plasma processing apparatus is controlled so that optimum processing can always be performed while monitoring the plasma reflection parameters and apparatus state parameters with the respective detectors. However, since these parameters are tens of types, it is very difficult to determine the cause when an abnormality is found in the operation state.
[0007]
Thus, for example, Japanese Patent Application Laid-Open No. 11-87323 discloses a monitoring of a processing apparatus that analyzes a plurality of process parameters of a semiconductor wafer processing system and statistically correlates these parameters to detect a change in process characteristics or system characteristics. Methods and processing devices have been proposed.
[0008]
As another method, one of the multivariate analysis is principal component analysis, which is one of the multivariate analyses, so that the plurality of parameters can be combined into a small number of parameters and the operation status of the processing device can be monitored based on a small number of statistical data. There is a method of evaluating the driving situation by using the method of (1).
[0009]
Furthermore, the above-mentioned plasma reflection parameters are detected, and if an abnormality is found, any of the above-mentioned control parameters or equipment state parameters is abnormally generated using a partial least squares method, which is one of the multivariate analyses. A method of predicting and specifying whether or not it has been proposed.
[0010]
[Problems to be solved by the invention]
Incidentally, in any of the above methods, it is important to accurately detect each parameter in order to grasp the processing state. However, there is a problem that a detection value may fluctuate due to a change over time of the plasma processing apparatus depending on a parameter, so that accurate detection data cannot be obtained.
For example, as the optical data included in the plasma reflection parameter, a value detected by an optical detector provided outside the processing chamber through a detection window provided on a side wall of the processing chamber is used. For this reason, the detection window is contaminated by operating the plasma processing apparatus for a long time, and the contaminants cause reflection and absorption of light, thereby attenuating the transmittance of plasma light emission. There was a problem that measurement could not be performed.
[0011]
Therefore, the present invention has been made in view of the above-described problems of the conventional plasma processing apparatus monitoring method and the plasma processing apparatus, and an object of the present invention is to provide an accurate optical processing method even after the plasma processing apparatus has been operated for a long time. New and improved plasma processing equipment can be obtained by improving the accuracy of predicting the state of the equipment and processing results, and accurately monitoring the plasma processing equipment. To provide a method for monitoring a plasma processing apparatus and a plasma processing apparatus.
[0012]
[Means for Solving the Problems]
In order to solve the above problems, according to a first aspect of the present invention, detection is performed for each object to be processed using a plurality of detectors including at least an optical detector for detecting light (for example, light intensity) in a processing chamber. A method for monitoring a plasma processing apparatus, wherein a multivariate analysis is performed using detection values of the respective detectors as analysis data such as optical data, and information on plasma processing is monitored based on the results. A variation detection step of detecting a variation of a detection value of the plasma processing apparatus in an optical detector; and a correction step of correcting the detection value based on the variation to obtain the analysis data. The monitoring method of the plasma processing apparatus described above is provided.
[0013]
In the fluctuation detecting step, when the plasma processing apparatus is in an initial state, light is radiated from the light source provided for fluctuation detection into the processing chamber, and the light intensity (e.g., light emission intensity) of the light from the light source is measured. Detecting and storing the detected value as a reference value, and detecting the light intensity of the light from the light source before performing plasma processing on each of the objects to be processed, and calculating a ratio of the detected value to the reference value. Determining and storing the variation as a variation, wherein the correcting step comprises detecting the light in the processing chamber with the detector during the plasma processing of each object to be processed, and determining a ratio determined as the variation from the detected value. The detection value may be corrected by dividing the value of 1/2 to the data for analysis.
[0014]
According to such a method according to the present invention, when the plasma processing apparatus is operated for a certain period of time or more, for example, the detection accuracy of optical data is reduced due to contamination of the detection window and the optical detector, and the plasma processing apparatus changes over time. Even if the optical data fluctuates, it can be corrected so that the optical data detection accuracy is the same as in the initial state, such as at the time of delivery or immediately after maintenance, which was good, so the plasma was always kept in a normal state The processing device can be operated, and a decrease in yield can be prevented.
[0015]
In particular, by performing multivariate analysis using analysis data including optical data, control of the state of the plasma in the above apparatus, the state of the apparatus such as a high-frequency voltage applied to the above apparatus, the flow rate of the processing gas, and the like. When predicting various data such as the processing results such as the state and the amount of abrasion of an object to be processed such as a wafer, even if the detection accuracy of optical data is reduced due to the operation of the plasma processing apparatus for a certain period of time or more, the prediction accuracy is reduced. Therefore, high prediction accuracy can be always maintained, and the plasma processing apparatus can be monitored more accurately.
[0016]
Further, it is determined whether or not the ratio obtained as the fluctuation amount is equal to or less than a predetermined amount, and when it is determined that the ratio is not equal to or less than the predetermined amount, the processing in the correction step is performed. The method may include a step of issuing a warning by an alarm without performing the processing of the correction step. According to this, when the detection value from the light source is attenuated below a certain amount, for example, an alarm for urging maintenance is issued, so that the light detection accuracy is reduced to such an extent that normal monitoring cannot be performed anymore by correction. Even in such a case, the maintenance can be performed so that the plasma processing apparatus can always be operated in a normal state, and a decrease in the yield can be prevented.
[0017]
Further, if the amount of fluctuation is detected for each wavelength of light and the correction is performed for each wavelength of light, optical data can be corrected more accurately. Further, if it is configured such that it is determined whether or not the ratio of the wavelength with the largest variation (for example, the wavelength with the largest attenuation) among the ratios of the respective wavelengths obtained as the variation is equal to or smaller than a predetermined fixed amount, the optical detection Since the warning can be issued even if all the wavelengths of the light detected from the container are not attenuated, the warning can be issued at a more appropriate maintenance time. Further, if it is configured such that it is determined whether or not the ratio of the wavelength having the greatest influence on the result of the multivariate analysis among the ratios of the respective wavelengths obtained as the fluctuation amount is equal to or smaller than a predetermined amount, the optical detector A warning can be issued at an appropriate maintenance time so as to prevent the fluctuation of the detected value at the time from affecting the result of the multivariate analysis.
[0018]
Further, a principal component analysis may be performed as a multivariate analysis using the analysis data, and, for example, operation information of the plasma processing apparatus may be monitored as information on the plasma processing. Further, a partial least squares method is performed as a multivariate analysis using the analysis data, and for monitoring information on the plasma processing, for example, an apparatus state of the plasma processing apparatus or a processing result of an object to be processed is predicted. Thus, the plasma processing apparatus may be monitored.
[0019]
In order to solve the above problems, according to a second aspect of the present invention, there is provided a plasma processing apparatus for generating plasma by applying high-frequency power to perform plasma processing on an object to be processed in a processing chamber. An optical detector for detecting light in the processing chamber via the detection window provided on the wall, and a light source window provided on a wall of the processing chamber at a position opposed to the detection window for detecting fluctuation. A light source for irradiating light into the processing chamber; and a reference value detecting means for detecting light from the light source when the plasma processing apparatus is in an initial state with the optical detector and storing the detected value as a reference value. Ratio calculating means for detecting light from the light source by the optical detector before performing plasma processing on each of the objects, calculating and storing the ratio of the detected value to the reference value, and performing plasma processing on each of the objects. Sometimes the light in the processing chamber is Correction means for detecting by a chemical detector, correcting by dividing the value of the above-mentioned power of 1/2 to the detected value, and using the corrected value as optical data; There is provided a plasma processing apparatus characterized by comprising monitoring means for performing a variable analysis and monitoring information relating to plasma processing based on the result.
[0020]
Further, the light source emits a plurality of wavelengths, the reference value detecting means detects the light intensity of each wavelength of the light from the light source, stores the detected value as a reference value, and the ratio calculating means Detecting the light intensity for each wavelength of light before the plasma processing of each processing object, storing the ratio of the detected value to the reference value, and correcting the light during the plasma processing of each processing object. The detected value of the light intensity for each wavelength may be corrected to be optical data.
[0021]
According to such an apparatus of the present invention, even if the detection accuracy of optical data is reduced by operating the plasma processing apparatus for a certain period of time or longer, the optical data is initialized before processing each object. Since the correction can be made to be the same as in the case of the above, the plasma processing apparatus can always be operated in a normal state, and a decrease in yield and the like can be prevented.
[0022]
In particular, by performing multivariate analysis using analysis data including optical data, control of the state of the plasma in the above apparatus, the state of the apparatus such as a high-frequency voltage applied to the above apparatus, the flow rate of the processing gas, and the like. When predicting various data such as the processing results such as the state and the amount of abrasion of an object to be processed such as a wafer, even if the detection accuracy of optical data is reduced due to the operation of the plasma processing apparatus for a certain period of time or more, the prediction accuracy is reduced. Therefore, high prediction accuracy can be always maintained.
[0023]
Further, it is determined whether the ratio calculated by the ratio calculating means is equal to or less than a predetermined fixed amount. If it is determined that the ratio is not equal to or less than the predetermined amount, the detection value is corrected by the correcting means, and it is determined that the value is equal to or less than the certain amount. At times, a warning may be issued by an alarm device. According to this, even if the light detection accuracy is reduced to such an extent that normal monitoring can no longer be performed by correction, maintenance can be performed so that the plasma processing apparatus can always be operated in a normal state, and the yield can be reduced. A decrease or the like can be prevented.
[0024]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, preferred embodiments of a method for monitoring a plasma processing apparatus and a plasma processing apparatus according to the present invention will be described in detail with reference to the accompanying drawings. In this specification and the drawings, components having substantially the same function and configuration are denoted by the same reference numerals, and redundant description is omitted.
[0025]
FIG. 1 is a schematic configuration diagram showing a plasma processing apparatus 100 according to an embodiment of the present invention, FIG. 2 is a block diagram showing an example of a multivariate analysis means of the plasma processing apparatus 100, and FIG. FIG. 4 is a diagram showing the wavelength dependence of the emission intensity of the light source 30 in the initial state of the device 100, FIG. 4 is a diagram showing the wavelength dependence of the emission intensity of the light source 30 through a dirty detection window, and FIG. 5 is a flowchart showing a correction method of the embodiment.
[0026]
As shown in FIG. 1, a plasma processing apparatus 100 includes an aluminum processing chamber and a vertically movable aluminum support 3 that supports a lower electrode 2 disposed in the processing chamber 1 via an insulating material 2A. And a shower head (hereinafter also referred to as an upper electrode) 4 which is arranged above the support 3 and supplies a process gas and also serves as an upper electrode.
[0027]
The processing chamber 1 has an upper part formed as a small-diameter upper chamber 1A and a lower part formed as a large-diameter lower chamber 1B. An entrance for loading / unloading the wafer W is formed in the upper part of the lower chamber 1B, and a gate valve 6 is attached to the entrance. A high-frequency power source 7 is connected to the lower electrode 2 via a matching unit 7A, and a high-frequency power P of 13.56 MHz is applied from the high-frequency power source 7 to the lower electrode 2 so that the upper electrode 1 in the upper chamber 1A. 4 and an electric field in the vertical direction. This high-frequency power P is detected via a power meter 7B connected between the high-frequency power supply 7 and the matching device 7A. The high-frequency power P is a controllable parameter and constitutes the above-described control parameter.
[0028]
An electric measuring device (for example, a VI probe) 7C is attached to the lower electrode 2 side (high-frequency voltage output side) of the matching device 7A, and the high-frequency power applied to the lower electrode 2 via the electric measuring device 7C. P detects a high frequency voltage V and a high frequency current I of a fundamental wave and a harmonic based on the plasma generated in the upper chamber 1A as electric data. These electric data constitute the above-mentioned plasma reflection parameters.
[0029]
Further, the matching device 7A includes, for example, two variable capacitors C1 and C2 and a coil L and performs impedance matching via the variable capacitors C1 and C2. The capacitances of the variable capacitors C1 and C2 in the matching state and the high-frequency voltage Vpp measured by a measuring instrument (not shown) in the matching unit 7A are determined by the above-described apparatus together with the APC (Auto Pressure Controller) opening degree described later. Configure state parameters.
[0030]
An electrostatic chuck 8 is disposed on the upper surface of the lower electrode 2, and a DC power supply 9 is connected to an electrode plate 8A of the electrostatic chuck 8. Therefore, the wafer W is electrostatically attracted to the electrostatic chuck 8 by applying a high voltage from the DC power supply 9 to the electrode plate 8A under a high vacuum. A focus ring 10 is arranged on the outer periphery of the lower electrode 2, and collects plasma generated in the upper chamber 1A on the wafer W. An exhaust ring 11 attached to an upper portion of the support 3 is provided below the focus ring 10. A plurality of holes are formed in the exhaust ring 11 at equal intervals in the circumferential direction over the entire circumference, and the gas in the upper chamber 1A is exhausted to the lower chamber 1B through these holes.
[0031]
The support 3 can be moved up and down between the upper chamber 1A and the lower chamber 1B via the ball screw mechanism 12 and the bellows 13. Therefore, when the wafer W is supplied onto the lower electrode 2, the lower electrode 2 descends to the lower chamber 1B via the support 3, releases the gate valve 6, and transfers the wafer W via a transfer mechanism (not shown). It is supplied to the lower electrode 2. The inter-electrode distance between the lower electrode 2 and the upper electrode 4 is a parameter that can be set to a predetermined value and is configured as a control parameter.
[0032]
A coolant channel 3A connected to the coolant pipe 14 is formed inside the lower electrode 2, and the coolant is circulated in the coolant channel 3A via the coolant pipe 14 to adjust the temperature of the wafer W to a predetermined temperature. Further, a gas flow path 3B is formed in each of the support 3, the insulating material 2A, the lower electrode 2, and the electrostatic chuck 8, and a fine passage between the electrostatic chuck 8 and the wafer W from the gas introduction mechanism 15 via the gas pipe 15A. He gas is supplied to the gap as a backside gas at a predetermined pressure to increase the thermal conductivity between the electrostatic chuck 8 and the wafer W via the He gas. A bellows cover 16 is provided below the exhaust ring 11.
[0033]
A gas introduction section 4A is formed on the upper surface of the shower head 4, and a process gas supply system 18 is connected to the gas introduction section 4A via a pipe 17. The process gas supply system 18 includes an Ar gas supply source 18A, a CO gas supply source 18B,₄F₈Gas supply 18C and O₂It has a gas supply source 18D. These gas supply sources 18A, 18B, 18C, 18D supply the respective gases to the shower head 4 at a predetermined set flow rate via valves 18E, 18F, 18G, 18H and mass flow controllers 18I, 18J, 18K, 18L. , And is adjusted as a mixed gas having a predetermined compounding ratio inside. Each gas flow rate is a parameter that can be detected and controlled by each of the mass flow controllers 18I, 18J, 18K, and 18L, and is configured as a control parameter.
[0034]
A plurality of holes 4B are uniformly arranged on the lower surface of the shower head 4 over the entire surface, and a mixed gas is supplied as a process gas from the shower head 4 into the upper chamber 1A through these holes 4B. An exhaust pipe 1C is connected to an exhaust hole at a lower portion of the lower chamber 1B, and the inside of the processing chamber 1 is exhausted through an exhaust system 19 including a vacuum pump and the like connected to the exhaust pipe 1C so that a predetermined gas pressure is obtained. Holding. The exhaust pipe 1C is provided with an APC valve 1D, and the opening is automatically adjusted in accordance with the gas pressure in the processing chamber 1. This opening is a device state parameter that indicates the device state and cannot be controlled.
[0035]
A detection window 28 for detecting light emission in the processing chamber 1 is provided on a side wall of the upper chamber 1A, and a spectroscope (hereinafter referred to as an optical detector) for detecting plasma emission in the processing chamber 1 is provided immediately outside the detection window 28. 20) are provided. The plasma state is monitored based on the optical data on the specific wavelength obtained by the optical detector 20, and the end point of the plasma processing is detected. The optical data constitutes a plasma reflection parameter that reflects the plasma state together with electrical data based on the plasma generated by the high frequency power P.
[0036]
Further, a detection window 32 is provided opposite to the detection window 28 of the upper chamber 1A, and a light source 30 for detecting fluctuation of optical data which is a feature of the present embodiment is provided immediately outside the detection window 32 together with the detection window 32. Provided. The light source 30 is a light source that emits light to check the initial state of the plasma processing apparatus and the degree of contamination of the detection windows 28 and 32 immediately before each processing.
[0037]
As the light source 30, at least one or more light sources (e.g., emitting one wavelength) that emit light of a plurality of wavelengths, for example, 1024 in the range of 200 to 950 nm, for example, are detected as optical data of the plasma reflection parameters. A plurality of light sources or one light source emitting light of all wavelengths) is preferable. For example, a deuterium lamp as a continuous spectrum light source, a xenon flash lamp, a xenon lamp, a mercury xenon lamp, a low-pressure mercury lamp as a bright-spectrum light source, and a hollow cathode lamp can be used.
[0038]
The output of the continuous spectrum light source is generally high in the order of xenon flash lamp, deuterium lamp, xenon lamp, and mercury xenon lamp in the order from the lowest. For example, a xenon flash lamp can be compared in several microseconds by pulse lighting. If this is the case, a high output that is about 1000 times higher than that of the continuous lighting type can be obtained.
[0039]
Regarding the stability of light output, the xenon flash lamp, the mercury xenon lamp, the xenon lamp, and the deuterium lamp are similarly stable in the order from the lowest, but use the central part of light emission using a convex mirror or slit. Therefore, a highly stable light output can be used even with a light source having low stability such as a xenon flash lamp.
[0040]
As shown in FIG. 2, the plasma processing apparatus includes multivariate analysis means 200. The multivariate analysis unit 200 includes, for example, a multivariate analysis program storage unit 201 that stores a multivariate analysis program, and an electrical unit that intermittently samples signals from the electrical measuring device 7C, the optical detector 20, and the parameter measuring device 21. A signal sampling unit 202, an optical signal sampling unit 203, a parameter signal sampling unit 204, and a model formula for predicting a plurality of control parameters and apparatus state parameters based on a plurality of plasma reflection parameters (electrical data and optical data) are stored. A model formula storage means 205 for performing calculation, a calculating means 206 for calculating a plurality of control parameters and / or device state parameters through the model formula, and a control parameter and / or device state parameter based on a calculation signal from the calculating means 206. Prediction, diagnosis, And a prediction, diagnosis and control means 207 to perform the control. The calculation means 206 is provided with a storage means 206a for storing data from each sampling means such as the optical signal sampling means 203. Each of the above parameters constitutes analysis data.
[0041]
Further, a control device 22 for controlling the plasma processing apparatus, an alarm device 23, and a display device 24 are connected to the multivariate analysis means 200, respectively. The control device 22 continues or interrupts the processing of the wafer W based on a signal from the prediction / diagnosis / control unit 207, for example. The alarm 23 and the display device 24 notify the abnormality of the control parameter and / or the device state parameter based on the signal from the prediction / diagnosis / control means 207 as described later. The parameter measuring device 21 collectively shows measurement of a plurality of control parameters such as flow rate detection.
[0042]
The calculation means 206 and the prediction / diagnosis / control means may be constituted by, for example, a microprocessor which operates based on a program from the multivariate analysis program storage means 201. The multivariate analysis program storage unit 201, the model expression storage unit 205, and the storage unit 206a may be each configured by a storage unit such as a memory, or may be configured by providing each storage region in a storage unit such as a hard disk. .
[0043]
The storage means 206a may be provided in the arithmetic means 206 as shown in FIG. 2, or may be provided separately from the arithmetic means 206 so as to be accessible from the arithmetic means 206. The storage unit 206a may function as an operation data storage unit that constitutes a unit that stores operation data such as plasma reflection parameters, control parameters, and apparatus state parameters from the respective sampling units. When predicting the processing result data, the storage unit 206a may function as a processing result data storage unit that configures a unit that stores the processing result data such as the wafer scraping amount.
[0044]
In the present embodiment, as the multivariate analysis, for example, JOURNAL OF OF CHEMOMETRICS, VOL. 2 (PP211-228) (1998), a partial least squares method (hereinafter, referred to as a PLS (Partial Least Squares) method) can be used. In the PLS method, a plurality of plasma reflection parameters are used as explanatory variables, and a target variable is a plurality of control parameters and apparatus state parameters. Equation) is used as a method for creating the equation. In the following model formula (1), X means a matrix of explanatory variables, and Y means a matrix of explanatory variables. B is a regression matrix including coefficients (weights) of explanatory variables, and E is a residual matrix.
Y = BX + E (1)
[0045]
In the PLS method, even if there are many explanatory variables and explained variables in each of the matrices X and Y, the relational expression between X and Y can be obtained if there are a small number of actual measured values. In addition, it is a feature of the PLS method that even a relational expression obtained with a small number of measured values has high stability and reliability.
[0046]
In the present embodiment, the program for the PLS method is stored in the multivariate analysis program storage means 201, and the calculation means 206 processes data as explanatory variables and objective variables in accordance with the procedure of the program. Then, the result is stored in the model formula storage means 205. Therefore, in the present embodiment, if the above equation (1) is obtained, the objective variable can be predicted by applying the data as the explanatory variable to the matrix X. Moreover, the predicted value becomes highly reliable.
[0047]
For example, X^TThe i-th principal component corresponding to the i-th eigenvalue for the Y matrix is t_iIs represented by The matrix X is the score t of the i-th principal component._iAnd the vector pi, the following equation (2) is used, and the matrix Y is the score t of the i-th principal component._iAnd the vector c_iIs expressed by the following equation (3). In the following equations (2) and (3), X_{i + 1}, Y_{i + 1}Is a residual matrix of X and Y, and X^TIs the transpose of matrix X. In the following, the index T means transposed matrix.
X = t₁p₁+ T₂p₂+ T₃p₃+ ・ ・ + T_ip_i+ X_{i + 1}... ▲ 2 ▼
Y = t₁c₁+ T₂c₂+ T₃c₃+ ・ ・ + T_ic_i+ Y_{i + 1}・ ・ ・ ▲ 3 ▼
[0048]
Thus, the PLS method used in the first embodiment is a method of calculating a plurality of eigenvalues and respective eigenvectors when the above equations (2) and (3) are correlated with a small amount of calculation.
[0049]
The PLS method is performed according to the following procedure. First, in the first stage, the operations of centering and scaling of the matrices X and Y are performed. Then, i = 1 is set, and X₁= X, Y₁= Y. U₁Matrix Y₁Is set in the first column. The centering is an operation of subtracting the average value of each row from the individual value of each row, and the scaling is an operation (processing) of dividing the individual value of each row by the standard deviation of each row.
[0050]
In the second stage, w_i= X_i ^Tu_i/ (U_i ^Tu_i), Then w_iIs normalized, t_i= X_iw_iAsk for. The same processing is performed for matrix Y to obtain c_i= Y_i ^Tt_i/ (T_i ^Tt_i), Then c_iNormalize the determinant of_i= Y_ic_i/ (C_i ^Tc_i).
[0051]
In the third stage, X loading (load amount) p_i= X_i ^Tt_i/ (T_i ^Tt_i), Y load q_i= Y_i ^Tu_i/ (U_i ^Tu_i). Then, u is returned to t_i= U_i ^Tt_i/ (T_i ^Tt_i). Then, the residual matrix X_i= X_i-T_ip_i ^T, Residual matrix Y_i= Y_i-B_it_ic_i ^TAsk for. Then, i is incremented to set i = i + 1, and the processing from the second stage is repeated. These series of processes are performed until a predetermined stop condition is satisfied according to the program of the PLS method, or the residual matrix X_{i + 1}Is repeated until converges to zero, and the maximum eigenvalue of the residual matrix and its eigenvector are obtained.
[0052]
The PLS method uses the residual matrix X_{i + 1}Quickly converges to the stop condition or zero, and the residual matrix converges to the stop condition or zero only by repeating about 10 calculations. Generally, the residual matrix converges to a stop condition or zero by repeating the calculation four to five times. X is calculated using the maximum eigenvalue and its eigenvector obtained by this calculation process.^TThe first principal component of the Y matrix is obtained, and the maximum correlation between the X matrix and the Y matrix can be known.
[0053]
In this embodiment, in order to specifically determine the model formula (1) by the PLS method, a predetermined number (for example, 18) of wafers are subjected to plasma processing (for example, etching processing) as a training set in advance, and a plurality of explanatory variables (for example, The plasma-reflecting parameters consisting of the above-mentioned electrical data and optical data) and target variables (the above-described control parameters such as the high-frequency power P and the gas flow rate, and the device state parameters such as the capacitances of the variable capacitors C1 and C2) are measured. Using these average values, a model formula (1), which is a relational expression between the explanatory variable and the objective variable, is obtained by the PLS method.
[0054]
When the processing of the wafer W is performed by the plasma processing apparatus 100, the high frequency voltage V and the high frequency current I of the fundamental wave and the harmonic based on the plasma are measured by the electric measuring device 7C, and the specific wavelength of the plasma is measured by the optical detector 20. The emission intensity of each wavelength in the range is measured, the average value thereof is obtained and substituted into the model formula, so that the predicted values of a plurality of control parameters and apparatus state parameters for each wafer W can be calculated. This makes it possible to constantly monitor the state of the plasma processing apparatus 100. If any of the control parameters and the apparatus state parameters is abnormal, an alarm is issued from the alarm device 23 to notify the abnormality. It is possible to control the plasma processing apparatus 100 to always operate in a normal state.
[0055]
Here, a method for obtaining accurate optical data in order to perform the above monitoring and control accurately will be described. FIG. 3 is a diagram in which a deuterium lamp is used as the light source 30 and the light emission intensity at each wavelength is measured by the optical detector 20 through the detection windows 28 and 32 in a clean state. FIG. FIG. 4 is a diagram in which the light emission intensity at each wavelength is measured by the optical detector 20 through the windows 28 and 32 and normalized by each light emission intensity detected through the clean window. The measurement is performed every 1 nm.
[0056]
The horizontal axis in FIG. 3 represents the wavelength (nm), and the vertical axis represents the emission intensity. As shown in FIG. 3, the emission intensity of the deuterium lamp through the clean detection windows 28 and 32 fluctuates in a range of about 230 to 450 nm. On the other hand, after operating for about 80 hours, the emission intensities of the respective wavelengths are measured in the same manner through the detection windows 28 and 32 that are considered to be in a considerably dirty state due to the adhesion of the foreign matter, and the detection window in a clean state is obtained. Assuming that the emission intensity at each wavelength at 28 and 32 is 100%, and expressed as a relative emission intensity with respect to that, as shown in FIG. 4, the attenuation is large on the low wavelength side and a gentle upward curve is drawn. The relative emission intensity is, for example, 50% at 226 nm, 67% at 260 nm, and 96% at 400 nm.
[0057]
As can be seen from FIGS. 3 and 4, the emission intensity at each wavelength detected by the optical detector 20 varies when the detection windows 28 and 23 are clean. In the range from to 400 nm, the attenuation is larger on the lower wavelength side, and the following procedure is taken to apply accurate optical data to the above model formula.
[0058]
As shown in FIG. 5, first, in step 501, the light source 30 emits light through the clean detection windows 28 and 32 immediately after cleaning or the like, and in step 502, the light emission intensity at each wavelength is measured by the optical detector 20. This measured value is called “initial value”. Next, in step 503, the initial value is stored in the storage means 206a of the arithmetic means 206. Subsequently, in step 504, the first plasma processing (for example, etching processing) is performed on the first wafer W by the plasma processing apparatus 100. The optical data serving as the initial value is used as a reference when correcting the optical data thereafter.
[0059]
In the second plasma processing by the plasma processing apparatus 100, the light source 30 is caused to emit light without generating plasma immediately before performing the plasma processing on each wafer W (step 505), and the light emission intensity at each wavelength is measured by an optical detector. Measure at 20. The measured value at this time is referred to as “pre-processed measured value”. Next, in step 506, the calculating means 206 compares the measured value before processing with the initial value for each wavelength to obtain a relative emission intensity which is a ratio (variation amount) of the measured value before processing to the initial value. It is stored in the storage means 206a.
[0060]
Next, in step 507, it is determined whether or not the relative light emission intensity is equal to or less than a predetermined fixed amount (hereinafter, referred to as “reference value”). If it is determined in step 507 that the relative light emission intensity is not lower than the reference value, the plasma processing is actually performed on the wafer W in step 508. At the time of such actual plasma processing, the light source 30 is turned off to emit no light, and only the light emission intensity emitted from the plasma is measured by the optical detector 20 in step 509. In this case, the optical detector 20 measures the emission intensity at a predetermined wavelength (for example, 200 to 950 nm).
[0061]
Subsequently, at step 510, the measured value from the optical detector 20 is divided by the value of the 発 光 power of the relative emission intensity for each wavelength calculated above to make correction for each wavelength, and the corrected value is obtained. (Correction value) is used as optical data. The reason for dividing by the value of the 強度 power of the relative emission intensity is that the emission intensity is attenuated twice by passing through the two detection windows 28 and 32 in the correction, whereas the actual plasma processing is performed. In this case, the light emission intensity is attenuated by only one detection window 28, so that the correction is made in consideration of this.
Then, in step 511, the correction value of the optical data is substituted into a model expression as a plasma reflection parameter together with other electrical data and the like, and a multivariate analysis for predicting the above-mentioned control parameters, apparatus state parameters, and the like is performed, and the plasma processing is performed. It monitors and controls the device 100. Thereafter, the process returns to step 505 to continue the processing.
[0062]
If the relative emission intensity, which is the ratio of the measured value before the processing to the initial value, falls below the reference value, the detection windows 28 and 32 determine that the correction cannot correct the multivariate analysis, and the alarm is turned off. It is preferable that an alarm be issued from 23 (step 512) to notify that it is time for maintenance. If an alarm is issued, the process is immediately stopped, and appropriate measures such as cleaning are performed (step 513). This can prevent the yield from lowering. As shown in FIG. 4, it is known that the reference value is determined so that the attenuation is large on the low wavelength side.
[0063]
Further, in the present embodiment, a case has been described in which the PLS method, which is one of the multivariate analyses, is used to create a model formula to monitor and control the plasma processing apparatus 100. The method is also useful for monitoring a plasma processing apparatus using principal component analysis, which is another type of multivariate analysis.
[0064]
Principal component analysis conceptually summarizes the parameters of the plasma processing apparatus by compiling them into a small number of parameters such as principal components and residuals, and monitoring the fluctuations of those parameters with respect to predetermined values. It is assumed that. As described above, when there are a large number of plasma reflection parameters, control parameters, apparatus state parameters, and the like as parameters indicating the state of the plasma processing apparatus, the parameters are collected into a small number, so that the situation can be easily grasped. The details of the principal component analysis are described in, for example, Japanese Patent Application No. 2000-201729.
[0065]
In the above case, a method of multiplying a parameter indicating a state of each plasma processing apparatus by a weighting coefficient (loading vector) to obtain a main component is used. When, for example, the first principal component is obtained by performing principal component analysis on the parameter values indicating the state of the plasma processing apparatus in a normal state or in a normal state and an abnormal state, the weight coefficient of the optical data is determined by the processing used. It changes depending on the type of gas and the cause of the abnormality. A reference value is determined for the relative emission intensity at the wavelength at which the value of the weighting coefficient (the component of the loading vector of the first principal component) is the largest, and an alarm is issued when the value falls below the reference value, so that the influence on the principal component analysis is minimized. It may be possible to accurately monitor and control the plasma processing apparatus by taking into account that the optical data having a large value does not fluctuate beyond a certain level.
In addition to the weighting factors of the principal components, the relative emission intensity at the wavelength where the value of the component of the residual vector in which the higher-order principal components having a low contribution rate are combined into one, or simply at the wavelength where the value variation is the largest. The reference value may be determined for.
[0066]
As described above, according to the present embodiment, by operating the plasma processing apparatus for a certain period of time or more, the detection accuracy of the optical data decreases due to the contamination of the detection window or the optical detector. Even if the optical data fluctuates due to the change, it can be corrected so that the detection accuracy of the optical data before processing each wafer is the same as that in the initial state such as after maintenance where the detection accuracy was good. The plasma processing apparatus can be operated in a normal state, and a decrease in yield can be prevented.
[0067]
In particular, by performing multivariate analysis using analysis data including optical data, control of the state of the plasma in the above apparatus, the state of the apparatus such as a high-frequency voltage applied to the above apparatus, the flow rate of the processing gas, and the like. When predicting various data such as processing results such as the state and the amount of wafer scraping, it is possible to prevent a decrease in the prediction accuracy even if the detection accuracy of optical data is reduced by operating the plasma processing apparatus for a certain period of time or more. Therefore, high prediction accuracy can be always maintained, and the plasma processing apparatus can be monitored more accurately.
[0068]
For example, a multivariate analysis method is performed using data for analysis including optical data, confirming that the plasma processing apparatus has stabilized after the plasma processing apparatus has started operating, and that the plasma processing apparatus has deviated from a stable state during operation. Corrects optical data fluctuations due to changes over time in the plasma processing equipment when performing confirmation of processing, prediction of control parameters and apparatus state parameters, uniformity of etching rate, process prediction of pattern dimensions, etching shape, damage, etc. Therefore, control and prediction can be performed more accurately, and the effect of applying this embodiment is great.
[0069]
Further, it is determined whether the ratio of each wafer to the reference value detected by the optical detector before the plasma processing is equal to or less than a predetermined amount (whether or not the correction width is equal to or more than a certain amount). When the judgment is made, the optical data is corrected, and when it is judged that it is less than a certain amount, a warning is issued by an alarm without making correction, so if the detected value from the light source is attenuated below a certain amount, For example, a warning that prompts maintenance is issued. Therefore, even if the light detection accuracy is reduced to the extent that normal monitoring can no longer be performed with correction, the plasma processing apparatus is always maintained in a normal state by performing maintenance. It can be operated and a decrease in yield can be prevented.
[0070]
In addition, the amount of change in optical data due to the aging of the plasma processing apparatus is detected for each wavelength of light, and the correction of optical data is corrected for each wavelength of light. Can be.
[0071]
In addition, as a criterion for determining whether or not to issue a warning such as a warning by a warning device, for example, the wavelength of the largest variation (for example, the wavelength of the largest attenuation) of the ratio of each wavelength obtained as the variation of the optical data is used. It may be based on whether the ratio is equal to or less than a predetermined fixed amount. By doing so, a warning can be issued even if all the wavelengths of the light detected from the optical detector are not attenuated, so that a warning can be issued at a more appropriate maintenance time.
[0072]
In addition, as a criterion for determining whether or not to issue the above-mentioned warning, the ratio of the wavelength that has the largest influence on the result of the multivariate analysis among the ratios of the respective wavelengths obtained as the variation is equal to or less than a predetermined amount. It may be based on whether or not. In this manner, a warning can be issued at an appropriate maintenance time that can prevent a change in the detection value of the optical detector from affecting the result of the multivariate analysis.
[0073]
As described above, the preferred embodiments of the monitoring method and the plasma processing apparatus according to the present invention have been described with reference to the accompanying drawings, but the present invention is not limited to such examples. It is obvious that a person skilled in the art can conceive various changes or modifications within the scope of the technical idea described in the claims, and those changes naturally fall within the technical scope of the present invention. It is understood to belong.
[0074]
For example, the configuration of the plasma processing apparatus is not limited to the above example, and the present invention can be applied to an apparatus that performs processing using plasma and uses optical data such as spectral analysis. .
[0075]
【The invention's effect】
As described above, according to the present invention, after the plasma processing apparatus has been operated for a long time, the correction is made to be the same as in the initial state, such as at the time of delivery or immediately after maintenance, in which the detection accuracy of optical data was good. As a result, accurate optical data can be obtained, and when maintenance is required, the fact is notified, so that the accuracy of the apparatus status and processing results can be improved, and the plasma processing apparatus can be monitored accurately. it can.
[Brief description of the drawings]
FIG. 1 is a schematic configuration diagram showing a plasma processing apparatus 100 according to one embodiment of the present invention.
FIG. 2 is a block diagram showing an example of a multivariate analysis unit 200 of the plasma processing apparatus 100.
FIG. 3 is a diagram showing the wavelength dependence of the emission intensity of the light source 30 in an initial state of the plasma processing apparatus 100.
FIG. 4 is a diagram showing the wavelength dependence of the emission intensity of the light source 30 through a dirty detection window.
FIG. 5 is a flowchart illustrating a method for correcting optical data.
[Explanation of symbols]
1 Processing room
1A @ Upper room
1B lower room
1C exhaust pipe
1D valve
2 lower electrode
2A insulation
3 Support
3A refrigerant channel
3B gas flow path
4 shower head (upper electrode)
4A Gas introduction section
4B hole
6 Gate valve
7 High frequency power supply
7A matching device
7B power meter
7C @ Electric measuring instrument
8 mm electrostatic chuck
8A electrode plate
9 DC power supply
10 focus ring
11 exhaust ring
12mm ball screw mechanism
13 bellows
14 refrigerant pipe
15 Gas introduction mechanism
15A gas piping
16 bellows cover
17 piping
18 Process gas supply system
18A gas supply source
18B gas supply source
18C gas supply source
18D gas supply source
18E valve
18I Mass Flow Controller
19mm exhaust system
20 ° optical detector
21 Parameter measuring instrument
22 control device
23 alarm
24 display device
28 ° detection window
30 ° light source
32 detection window
100 plasma processing equipment
200 multivariate analysis means
201 Multivariate analysis program storage means
202 electrical signal sampling means
203 Optical signal sampling means
204 parameter signal sampling means
205 model expression storage means
206 arithmetic means
206a storage means
207 Prediction, diagnosis and control means
W wafer

Claims (11)

A multivariate analysis is performed by using, as analysis data, a detection value of each of the detectors detected for each object to be processed using a plurality of detectors including an optical detector for detecting light in the processing chamber. A monitoring method of a plasma processing apparatus for monitoring information on a plasma processing based on a result,
A fluctuation amount detecting step of detecting a fluctuation of a detection value of the plasma processing apparatus in the optical detector;
A correction step of correcting the detection value based on the variation amount to obtain the analysis data;
A method for monitoring a plasma processing apparatus, comprising: In the fluctuation detecting step, when the plasma processing apparatus is in an initial state, light is emitted from the light source provided for fluctuation detection into the processing chamber, and the light intensity of the light from the light source is detected. Storing a value as a reference value, and detecting the light intensity of the light from the light source before performing plasma processing on each of the objects to be processed, and calculating a ratio of the detected value to the reference value as the variation. Remembering,
In the correcting step, the light in the processing chamber is detected by the detector at the time of plasma processing of each object to be processed, and a value of a 1/2 power of a ratio obtained as the fluctuation amount is divided from the detected value. Correcting the detection value to obtain the analysis data,
The method for monitoring a plasma processing apparatus according to claim 1, wherein: It is determined whether or not the ratio obtained as the fluctuation amount is equal to or less than a predetermined amount. When it is determined that the ratio is not equal to or less than the predetermined amount, the processing in the correction step is performed. The monitoring method for a plasma processing apparatus according to claim 1, wherein a warning is issued by an alarm device without performing the step processing. 4. The method according to claim 1, wherein the amount of change is detected for each wavelength of light, and the correction is performed for each wavelength of light. 5. The monitoring of the plasma processing apparatus according to claim 4, wherein it is determined whether or not the ratio of the wavelength having the largest variation among the ratios of the wavelengths obtained as the variation is equal to or less than a predetermined amount. Method. 5. The method according to claim 4, wherein, among the ratios of the respective wavelengths obtained as the fluctuation amounts, a ratio of wavelengths having the greatest influence on the result of the multivariate analysis is determined to be equal to or less than a predetermined amount. The monitoring method of the plasma processing apparatus described in the above. The method according to any one of claims 1 to 4, wherein principal component analysis is performed as multivariate analysis using the analysis data, and operation information of the plasma processing apparatus is monitored. . A partial least squares method is performed as a multivariate analysis using the analysis data, an apparatus state of the plasma processing apparatus or a processing result of an object to be processed is predicted, and the plasma processing apparatus is monitored based on the predicted value. The method for monitoring a plasma processing apparatus according to claim 1. A plasma processing apparatus for generating plasma by applying high-frequency power to perform plasma processing on an object to be processed in a processing chamber,
An optical detector that detects light in the processing chamber through the detection window provided in a wall of the processing chamber;
A light source for irradiating the processing chamber with light for fluctuation detection through a light source window provided on a wall of the processing chamber at a position facing the detection window;
Reference value detection means for detecting light from the light source when the plasma processing apparatus is in an initial state with the optical detector and storing the detected value as a reference value;
Ratio calculating means for detecting light from the light source by the optical detector before the plasma processing of each object to be processed, and calculating and storing a ratio of the detected value to the reference value;
At the time of plasma processing of each object to be processed, light in the processing chamber is detected by the optical detector, and corrected by dividing a value of a 1/2 power of the ratio from the detected value. Correction means,
Monitoring means for performing multivariate analysis using at least the optical data and monitoring information relating to plasma processing based on the result;
A plasma processing apparatus comprising: The light source emits a plurality of wavelengths;
The reference value detection means detects light intensity of each wavelength of light from the light source, stores the detected value as a reference value,
The ratio calculating means detects light intensity for each wavelength of light before plasma processing of each object to be processed, and stores a ratio of the detected value to the reference value,
10. The plasma processing apparatus according to claim 9, wherein the correction unit corrects the detected value of the light intensity for each wavelength of the light at the time of the plasma processing of each object to be processed into optical data. It is determined whether or not the ratio calculated by the ratio calculating means is equal to or less than a predetermined amount. If it is determined that the ratio is not equal to or less than the predetermined amount, the detection value is corrected by the correcting means. 11. The plasma processing apparatus according to claim 9, wherein a warning is issued by a vessel.

Images