Substrate processing apparatus and substrate processing method

Abstract

According to one embodiment of the present invention, in a substrate processing method in which a substrate is processed by a heater that heats the substrate to perform a semiconductor process, the substrate processing method includes: measuring a correlation formula of a temperature of the heater in one or more independent variables and dependent variables; inputting a measured value corresponding to the independent variable into the correlation formula, and calculating a predicted temperature of the heater; and calculating an estimated temperature by applying a Kalman filter to the predicted temperature.

Description

Technical Field

[0001] The present invention relates to a substrate processing device and a substrate processing method, and more particularly to a substrate processing device and a substrate processing method capable of confirming the temperature of each area of ​​a heater. Background Art

[0002] Generally speaking, heaters used in semiconductor equipment include multiple zones, that is, multiple zones with independent heating wires on each zone, but it is difficult to install a number of temperature measurement sensors (TC) corresponding to the number of zones. Specifically, the heater has a circular structure, and the zones where the temperature can be measured / controlled are limited to 1 or 2 (for example, only the temperature of a part of the heater, i.e., the center, can be measured). When the heater is divided into more than that number of zones, there is a problem that a temperature sensor that can measure / control the temperature of each zone cannot be installed.

[0003] Therefore, the output of the controllable area is used as a reference to proportionally control the output of other areas. However, this may cause problems in temperature confirmation or process results, and a lot of time is required for temperature tuning.

[0004] In particular, since semiconductor processes generally depend on heater temperature, it is important to measure / control the temperature of each area of ​​the heater, and the above problem has a negative impact on the process uniformity of the substrate. In the manufacture of semiconductor devices, with the recent miniaturization and high functionality of semiconductor devices, processing accuracy of several nanometers is required. In order to achieve such processing accuracy, it is necessary to maintain the heater temperature uniformly to achieve a stable semiconductor process. Summary of the invention Technical issues

[0005] An object of the present invention is to provide a substrate processing apparatus and a substrate processing method capable of confirming the temperature of each zone of a heater.

[0006] Another object of the present invention is to provide a substrate processing apparatus and a substrate processing method that can reduce the time required for temperature tuning and improve equipment utilization.

[0007] Other objects of the present invention will become more apparent from the following detailed description and accompanying drawings. Problem Solution

[0008] According to one embodiment of the present invention, in a substrate processing method for processing a substrate by a heater, the heater heats the substrate to implement a semiconductor process, and the substrate processing method includes: for a correlation equation of one or more independent variables and dependent variables, inputting a measured value corresponding to the above-mentioned independent variable into the above-mentioned correlation equation, and calculating the predicted temperature of the above-mentioned heater, the independent variable is the above-mentioned heater-related parameter, and the dependent variable includes the above-mentioned heater measured temperature; and applying a Kalman filter to the above-mentioned predicted temperature to calculate the estimated temperature.

[0009] The method may further include: a step of associating the independent variable and the dependent variable and saving them in a database; and a step of calculating the coefficients contained in the correlation equation by regression analysis of the independent variable and the dependent variable.

[0010] The above-mentioned heater includes the first area to the nth area (n=2, 3, ..., m, m is an integer), and the above-mentioned independent variables may include: the temperature of one of the above-mentioned first area to the nth area, the power ratio of the first area to the nth area, the current value supplied to the heating wire set in the area where the above-mentioned predicted temperature is to be calculated, the resistance value of another one of the above-mentioned first area to the nth area, and the resistance value of the area where the above-mentioned predicted temperature is to be calculated.

[0011] The above parameters may be voltage value, current value, temperature value, resistance value, power value, power ratio, output, etc.

[0012] Any one of the first to n-th regions may be a circular region located at the center of the heater.

[0013] Another one of the first to n-th regions may be an annular region surrounding one of the first to n-th regions.

[0014] The correlation equation for calculating the predicted temperature (Tn) of the n-th region can be expressed as the following “Formula 1”.

[0015] <Mathematical formula 1>

[0016] Predicted temperature (Tn) = Y intercept + (temperature of one of the first to nth regions * a) + (power ratio of the first region * b1) + (power ratio of the second region * b2) + ... + (power ratio of the nth region * bn) + (current value supplied to the heating wire provided in the nth region * c) + (resistance value of another of the first to nth regions * d) + (resistance value of the nth region * e)

[0017] In the above , Y intercept, a, b, c, d, and e are the above coefficients.

[0018] The method may further include the step of placing a temperature measuring sensor (eg, a chip TC) capable of measuring the temperature of the first to n-th regions on the heater to obtain the measured temperature.

[0019] According to one embodiment of the present invention, a substrate processing device processes a substrate by a heater, wherein the heater heats the substrate to implement a semiconductor process, and the substrate processing device comprises: a database that associates and stores the measured temperature of the heater among one or more independent variables and dependent variables; a storage unit that stores the correlation between the independent variable and the dependent variable; a calculation unit that calculates the coefficients included in the correlation by regression analysis of the independent variable and the dependent variable; a temperature control mechanism that has a Kalman filter that calculates an estimated temperature based on the predicted temperature; and a measurement mechanism that obtains a measured value corresponding to the independent variable. The calculation unit inputs the measured value into the correlation to calculate the predicted temperature of the heater.

[0020] Effects of the Invention

[0021] According to an embodiment of the present invention, the temperature of each region of the heater can be confirmed, thereby achieving feedback control of each region and high-precision semiconductor processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 FIG. 1 is a diagram schematically showing a substrate processing apparatus according to an embodiment of the present invention.

[0023] Figure 2 It is shown Figure 1 Diagram of the heater shown.

[0024] Figure 3 It is shown Figure 1 Block diagram of the temperature control mechanism shown.

[0025] Figure 4 It is shown by Figure 3 The flowchart shows the process of calculating the estimated temperature by the temperature control mechanism and the measuring structure.

[0026] Figure 5 It is shown in the application Figure 3 A graph showing the change in measured temperature of the first region before and after the Kalman filter is shown.

[0027] Figure 6 It is shown in the application Figure 3 The graph shows the change in RMSE value of the first region before and after the Kalman filter.

[0028] Figure 7 It is shown in the application Figure 3 A diagram showing changes in the measured temperature and estimated temperature of the first region after the Kalman filter.

[0029] Figure 8 It is a representation of each zone of the heater Figure 4 A graph showing estimated and actual temperatures. DETAILED DESCRIPTION

[0030] Below, refer to the attached Figures 1 to 8 , the preferred embodiments of the present invention are described in more detail. The embodiments of the present invention can be transformed into various forms, and the scope of the present invention should not be interpreted as being limited to the embodiments described below. Each embodiment is provided to further explain the present invention in detail to a person skilled in the art. Therefore, in order to emphasize a clearer description, the shapes of the elements shown in the figures may be exaggerated.

[0031] Figure 1 FIG. 2 is a diagram schematically showing a substrate processing apparatus according to an embodiment of the present invention. Figure 1 As shown, the substrate processing apparatus includes a processing chamber 1 for performing a processing process on a processed object, and a gas supply mechanism 2 supplies gas into the processing chamber 1. The gas in the processing chamber 1 is exhausted by a scheduling mechanism 4, and a valve 3 adjusts the exhaust volume to control the pressure in the processing chamber 1. A pressure gauge 5 measures the pressure in the processing chamber 1.

[0032] Figure 2 It is shown Figure 1 The heater 11 is provided in the processing chamber 1 for supporting the object to be processed. The heater 11 can be divided into six areas, namely, the first area to the sixth area (① to ⑥). Each of the divided areas is provided with a heating wire (not shown), which is independently controlled by a power supply mechanism (not shown). At this time, the power supply mechanism can supply power to each heating wire in accordance with a certain proportion (or power ratio, such as %) of the power supplied to the heater 11 as a whole.

[0033] The first region (①) may be a circular region located in the center of the heater 11, and the second region (②) may be an annular region disposed around the first region. The third to sixth regions (③ to ⑥) may be regions obtained by dividing the annular region disposed around the second region at equal angles (e.g., 90 degrees). However, unlike the present embodiment, the heater 11 may be divided into less than 5 or more than 7 regions, or may be divided into 6 regions in a manner different from the above.

[0034] The temperature control mechanism 14 and the measuring mechanism 15 are used to control the temperature of the heater 11. The measuring mechanism 15 includes a TC (21) which is provided in the first region ① of the heater 11 and measures the temperature of the first region (①) of the heater 11. In addition, as described later, the measuring mechanism 15 includes a temperature measuring sensor (for example, a chip TC (w)) corresponding to a temperature measuring portion. The temperature measuring sensor (for example, a chip TC (w)) is provided on the upper part of the heater 11, and the temperature measuring sensor has a thermocouple (TC) which corresponds to a location located in each region of the heater 11. The thermocouple wire (TCwire) connected to the thermocouple (TC) is connected to the temperature control mechanism 14, and the temperature of each region can be measured.

[0035] The measuring mechanism 15 further includes a power measuring unit, a current measuring unit, and a resistance measuring unit, and the power measuring unit, the current measuring unit, and the resistance measuring unit measure the current power, the current, and the resistance of each area of ​​the heater 11 .

[0036] Figure 3 It is shown Figure 1 The block diagram of the temperature control mechanism is shown. The values ​​of temperature / power / current / resistance measured by the measuring mechanism 15 for each area of ​​the heater 11 are transmitted to the temperature control mechanism 14, and the temperature control mechanism 14 estimates the temperature of each area of ​​the heater 11 based on the values. The temperature control mechanism 14 includes: a calculation unit 24, a display unit 22, a data input unit 24, a database 25, a correlation formula storage unit 26 and a Kalman filter 21.

[0037] Figure 4 It is shown by Figure 3 The flow chart of the process of calculating the estimated temperature by the temperature control mechanism and the measuring structure shown in the figure. Figure 4 , a temperature control method based on a temperature control mechanism and a measuring mechanism is explained.

[0038] First, a temperature measuring sensor (e.g., chip TC(w)) is provided on the heater 11. The temperature measuring sensor (e.g., chip TC(w)) has a chip shape and can measure the temperature of each area of ​​the heater 11 (S1). After heating the heater 11 for a sufficiently long time under a variety of learning conditions (e.g., heating conditions, stabilization, interference, etc.), the temperature of each area of ​​the heater 11 is measured. At the same time, the learning conditions (e.g., heating conditions, stabilization, interference, etc.) such as power / current / resistance are also measured for each area of ​​the heater 11 (e.g., power measurement value / current measurement value / resistance measurement value) (S2). The measured temperature and learning conditions (e.g., heating conditions, stabilization, interference, etc.) at this time are input / saved in the database 25 through the data input mechanism 24.

[0039] The data input mechanism 24 can be a device that an operator can manually input data, such as a keyboard, mouse, stylus, etc., preferably a medium drive such as a floppy disk drive that can read electromagnetic information, and most preferably a network interface that automatically reads the measured temperature and learning conditions (for example, heating conditions, stabilization, interference, etc.).

[0040] After these data are fully stored in the database 25, the calculation unit 24 performs calculations for associating the measured temperature with the learning conditions (e.g., heating conditions, stabilization, disturbance, etc.) to obtain a correlation. Here, in order to compile the correlation, for example, a polynomial of the predicted temperature with the learning conditions (e.g., heating conditions, stabilization, disturbance, etc.) as variables is assumed, and then regression analysis is performed to obtain each proportionality coefficient.

[0041] Specifically, in x experiments, the temperature of each area of ​​the heater 11 is measured based on the learning conditions (for example, heating conditions, stabilization, interference, etc.), and the learning conditions such as power / current / resistance (for example, heating conditions, stabilization, interference, etc.) are directly output from the measuring mechanism 15 to the database 25, and are simultaneously associated with the temperature of each area of ​​the heater 11 and stored.

[0042] As shown above, it can be seen that when a data set of measured temperature and learning conditions (e.g., heating conditions, stabilization, interference, etc.) for each region is obtained, the following correlation formula (S3) can be obtained from the data set, and the predicted temperature can be obtained from the learning conditions (e.g., heating conditions, stabilization, interference, etc.) through the correlation formula. That is, the correlation formula for calculating the predicted temperature (Tn) of the nth region is expressed as the following <Mathematical Formula 1>.

[0043] <Mathematical formula 1>

[0044] Predicted temperature (Tn) = Y intercept + (temperature of one of the first to nth regions * a) + (power ratio of the first region * b1) + (power ratio of the second region * b2) + ... + (power ratio of the nth region * bn) + (current value supplied to the heating wire provided in the nth region * c) + (resistance value of another of the first to nth regions * d) + (resistance value of the nth region * e)

[0045] In the above <Mathematical Formula 1>, Y intercept, a, b, c, d, and e are the above coefficients.

[0046] Specifically, the regression analysis can use the learning conditions (e.g., heating conditions, stabilization, interference, etc.) of the heater 11 (specifically, the temperature of the first area ①, the power ratio of the first to sixth areas ① to ⑥, the current value of the area, the resistance value of the second area ②, the resistance value of the area) as independent variables, and the measured temperature of the area as a dependent variable. The regression analysis can be SVM regression analysis, linear regression analysis, Gaussian regression analysis, etc. In addition, if the above correlation is observed, it can be seen that the predicted temperature is a proportional coefficient relative to the learning conditions (e.g., heating conditions, stabilization, interference, etc.), and the calculation unit 24 calculates the proportional coefficient based on the data set.

[0047] The correlation equation for calculating the predicted temperature can be determined by the method described above. Unlike the present embodiment, the correlation equation may be different according to the structure of the heater 11 (e.g., the configuration of the divided areas), etc. In addition, regarding the correlation equation, the accuracy of the correlation equation can be improved by using a polynomial including high-order terms such as second-order and third-order.

[0048] When performing the above calculation, the calculation unit 24 may be a central processing unit used in a general personal computer, or an integrated circuit specially made for such calculation. After the correlation equation is compiled as shown above, the correlation equation is recorded in the correlation equation storage unit 26.

[0049] After completing these series of operations, the temperature measuring sensor (for example, wafer TC(w)) provided on the heater 11 is removed. After that, although no temperature measuring instrument is provided on the substrate processing apparatus, the predicted temperature of each area of ​​the heater can be calculated based on the correlation formula recorded in the correlation formula storage unit 26 by measuring the learning conditions (for example, temperature rise conditions, stabilization, interference, etc.). The correlation formula compiled as shown above is stored in the correlation formula storage unit 26, and is read from the correlation formula storage unit 26 as needed, thereby calculating the predicted temperature of each area of ​​the heater (S4).

[0050] On the other hand, the measured value read by the measuring mechanism 15 not only affects the accuracy but also has noise that generates spikes, and the Kalman filter 21 can intelligently infer the actual state of the system. This inference can be performed by the Kalman gain (generally represented by 'K').

[0051] Figure 5 It is shown in the application Figure 3 The graph of the measured temperature change of the first area before and after the Kalman filter is shown. Figure 5As shown, the measured temperature (WaferTemp) measured by the temperature measuring sensor (for example, chip TC(w)) for the first area (①) contains noise. It can be seen that the value (Kalmanfilter) of the measured temperature (WaferTemp) applied with the Kalman filter has a tendency similar to the measured temperature (HeaterTemp) measured by TC(21) and has a certain error.

[0052] In short, the predicted temperature calculated by the temperature control mechanism 14 is input into the Kalman filter, and the Kalman filter can calculate the estimated temperature (S5) based on the predicted temperature.

[0053] (1) System state equation

[0054] x_Predict=(A*Xk)+B_uk

[0055] *A is the time for measuring temperature change, which is the same each time and is 1.

[0056] Xk is the initial value (changed by subsequent calculations).

[0057] B_uk is an input value that is independent of the system and is 0 at temperature.

[0058] (2) Error covariance

[0059] p_Predict=(A*Pk)+Q

[0060] Pk is the value of the associated uncertainty, which is taken to be 1 (ie, the uncertainty is assumed to be almost 0).

[0061] Q is an element for assigning a weighted value. The larger the Q is, the greater the weighted value assigned to the measured value relative to the estimated value.

[0062] (3) Kalman gain

[0063] K_gain=p_Predict / (H*p_Predict+R)

[0064] H is the transformation coefficient and R is the variance of the observation.

[0065] (4) Calculation of presumed value

[0066] Xk=x_Predict+K_gain*(z_Din-x_Predict)

[0067] z_Din is the current measured value.

[0068] (5) Error covariance update

[0069] Pk=(1-K_gain)*p_Predict

[0070] Indicates the degree of correlation between the prediction error and the Kalman gain.

[0071] Figure 6 It is shown in the application Figure 3 The figure shows the change of RMSE value in the first area before and after the Kalman filter. Figure 6 In the figure, the left side is a graph showing the change of the measured temperature and predicted temperature of the first area measured by the temperature measuring sensor (e.g., chip TC(w)), and the right side is a graph showing the change of the measured temperature and predicted temperature when the Kalman filter is applied. By observing the RMSE (Root Mean Square Error) of each, it can be seen that the prediction accuracy (RMSE) is improved when the Kalman filter is used.

[0072] Figure 7 For application Figure 3 The figure shows the changes in the measured temperature and estimated temperature of the first area after the Kalman filter. Figure 7 As shown, it can be seen that the measured temperature (TC Wafer temperature) has a tendency similar to the estimated temperature (predicted temperature) and has a certain error, which is within 0.5 degrees.

[0073] 8 is the representation of each area of ​​the heater Figure 4 The figure shows the estimated temperature and the actual temperature. Figure 8 As shown, it can be seen that the change in the estimated temperature (predicted temperature) and the change in the process result (film thickness) have a certain linearity.

[0074] The above describes a method for calculating the estimated temperature of each area of ​​the heater 11 using the learning conditions of the heater 11 (e.g., heating conditions, stabilization, interference, etc.), but the estimated temperature can also be calculated using other learning conditions that are not the learning conditions described above (e.g., heating conditions, stabilization, interference, etc.).

[0075] As described above, according to the embodiment of the present invention, even if a temperature measuring instrument is not provided for each area of ​​the heater 11 , the temperature can be checked, thereby optimizing the temperature control of the heater 11 .

[0076] Although the present invention has been described in detail by means of the preferred embodiments, other embodiments are possible. Therefore, the technical concept and scope of the following claims are not limited to the preferred embodiments.

[0077] Industrial Applicability

[0078] The present invention can be applied to various types of semiconductor manufacturing equipment and manufacturing methods.

Claims

1. A substrate processing method, wherein the substrate is processed by a heater, wherein the heater heats the substrate to implement a semiconductor process, wherein: include: A step of calculating a predicted temperature of the heater by inputting a measured value corresponding to the independent variable into a correlation equation of one or more independent variables and dependent variables, wherein the independent variable is the heater-related parameter and the dependent variable includes the measured temperature of the heater; and A step of applying a Kalman filter to the predicted temperature to calculate an estimated temperature.

2. The substrate processing method according to claim 1, wherein: Also includes: The step of associating the independent variable with the dependent variable and saving them in a database; and The step of calculating the coefficients contained in the above correlation equation by regression analysis of the above independent variables and the above dependent variables.

3. The substrate processing method according to claim 1, wherein: The heater includes a first region to an nth region, where n=2, 3, ..., m, and m is an integer. The above-mentioned independent variables include: the temperature of one of the above-mentioned first region to nth region, the power ratio of the first region to nth region, the current value supplied to the heating wire set in the region where the above-mentioned predicted temperature is to be calculated, the resistance value of another one of the above-mentioned first region to nth region, and the resistance value of the region where the above-mentioned predicted temperature is to be calculated.

4. The substrate processing method according to claim 3, characterized in that: One of the first to n-th regions is a circular region located at the center of the heater.

5. The substrate processing method according to claim 4, characterized in that: Another one of the first to n-th regions is an annular region surrounding one of the first to n-th regions.

6. The substrate processing method according to any one of claims 3 to 5, characterized in that: The correlation equation for calculating the predicted temperature (Tn) of the nth region is expressed as the following <Math. 1>: <Mathematical formula 1> Predicted temperature (Tn) = Y intercept + (temperature of one of the first to nth regions * a) + (power ratio of the first region * b1) + (power ratio of the second region * b2) + ... + (power ratio of the nth region * bn) + (current value supplied to the heating wire provided in the nth region * c) + (resistance value of another of the first to nth regions * d) + (resistance value of the nth region * e), In the above <Mathematical Formula 1>, Y intercept, a, b, c, d, and e are the above coefficients.

7. The substrate processing method according to claim 1, wherein: Also includes: A step of placing a wafer TC capable of measuring the temperatures of the first to n-th regions on the heater to obtain the measured temperatures.

8. A substrate processing device, which processes a substrate by means of a heater, wherein the heater heats the substrate to implement a semiconductor process, characterized in that: include: A database storing the above-mentioned heater measured temperature in association with one or more independent variables and dependent variables; A storage unit for storing a correlation expression between the independent variable and the dependent variable; A calculation unit, which calculates the coefficients contained in the above correlation equation by performing regression analysis on the above independent variables and the above dependent variables; a temperature control mechanism including a Kalman filter for calculating an estimated temperature based on the predicted temperature; and The measuring means obtains a measured value corresponding to the above-mentioned independent variable, The calculation unit inputs the measured value into the correlation equation to calculate a predicted temperature of the heater.

9. The substrate processing method according to claim 1, wherein: The heater includes a first region to an nth region, where n=2, 3, ..., m, and m is an integer. The above-mentioned independent variables include: the temperature of one of the above-mentioned first region to nth region, the power ratio of the first region to nth region, the current value supplied to the heating wire set in the region where the above-mentioned predicted temperature is to be calculated, the resistance value of another one of the above-mentioned first region to nth region, and the resistance value of the region where the above-mentioned predicted temperature is to be calculated.

10. The substrate processing apparatus according to claim 9, wherein: The correlation equation for calculating the predicted temperature (Tn) of the nth region is expressed as the following <Math. 1>: <Mathematical formula 1> Predicted temperature (Tn) = Y intercept + (temperature of one of the first to nth regions * a) + (power ratio of the first region * b1) + (power ratio of the second region * b2) + ... + (power ratio of the nth region * bn) + (current value supplied to the heating wire provided in the nth region * c) + (resistance value of another of the first to nth regions * d) + (resistance value of the nth region * e), In the above <Mathematical Formula 1>, Y intercept, a, b, c, d, and e are the above coefficients.

Images