Substrate processing apparatus, temperature control program, method of manufacturing semiconductor device, and temperature control method

Abstract

The cooling control section is configured to acquire at least one of the temperature of the heating section and the temperature of the processing chamber, the opening degree of the cooling valve, and information of the exhaust fan, so as to adjust the opening degree of the cooling valve in such a manner that the error between the predicted temperature series calculated according to the prediction model and the target temperature series calculated according to the ratio of the change from the current target temperature to the final target temperature becomes minimum.

Description

Technical Field

[0001] This disclosure relates to a substrate processing apparatus, a temperature control program, a method for manufacturing a semiconductor device, and a temperature control method. Background Technology

[0002] A semiconductor manufacturing apparatus is known as an example of a substrate processing apparatus, and a vertical forming apparatus is another example of a semiconductor manufacturing apparatus. In a vertical forming apparatus, a boat that holds multiple substrates (hereinafter also referred to as wafers) in multiple layers is moved into a processing chamber inside a reaction tube while holding the substrates. The substrates are processed at a predetermined temperature in multiple zones while temperature control is performed. Up to this point, in conventional heater temperature control, the heater is turned off during cooling, but in recent years, efforts have been made to improve the cooling characteristics after substrate processing.

[0003] For example, Patent Document 1 discloses a semiconductor manufacturing apparatus in which heating based on a heater unit and cooling based on a cooling unit are performed in parallel, following a predetermined heating rate and a predetermined cooling rate. Patent Document 2 discloses a semiconductor manufacturing apparatus in which temperature characteristics are automatically acquired beforehand, and temperature control is performed using these characteristics, thereby preventing deviations in the controller's performance.

[0004] In the control of cooling gas flow rate in the aforementioned cooling unit configuration, during rapid cooling, the varying cooling rates in each zone can sometimes lead to discrepancies in the historical temperature records of those zones. Furthermore, in feedback control based on PID calculations, appropriate parameters need to be predefined; however, the optimization of these PID parameters must proceed through a trial-and-error process to explore optimal values, and the results largely depend on the operator's intuition and experience.

[0005] Existing technical documents

[0006] Patent documents

[0007] Patent Document 1: International Publication No. 2018 / 100826

[0008] Patent Document 2: Japanese Patent Application Publication No. 2019-145730 Summary of the Invention

[0009] The purpose of this disclosure is to provide a technique that can improve temperature deviation within a range using optimal parameters.

[0010] According to one aspect of this disclosure, a technology is provided that comprises:

[0011] The reaction tube forms the processing chamber for processing the substrate inside;

[0012] A heater unit is disposed outside the reaction tube and has a heating section for heating the substrate;

[0013] A cooling unit having a cooling valve that supplies cooling medium to the space between the heater unit and the reaction tube;

[0014] An exhaust fan that supplies the cooling medium to the cooling unit; and

[0015] A cooling control unit acquires a prediction model that infers a predicted temperature for at least one of the predicted temperatures of the heating element and the processing chamber. This prediction model includes information about the exhaust fan, a final target temperature that will become the future target, and the opening degree of the cooling valve. The cooling control unit acquires information about at least one of the temperatures of the heating element and the processing chamber, the opening degree of the cooling valve, and the exhaust fan, and adjusts the opening degree of the cooling valve in a manner that minimizes the error between the predicted temperature series calculated according to the prediction model and the target temperature series calculated using the rate of change when the current target temperature changes to the final target temperature.

[0016] Invention Effects

[0017] According to this disclosure, it is possible to improve the temperature deviation within a range using optimal parameters. Attached Figure Description

[0018] Figure 1 This is a partial truncated front view illustrating a substrate processing apparatus according to an embodiment of the present disclosure.

[0019] Figure 2 This is a front cross-sectional view of a substrate processing apparatus according to one embodiment of the present disclosure.

[0020] Figure 3 This is a diagram illustrating the temperature control unit of one embodiment of the present disclosure.

[0021] Figure 4 This is a diagram illustrating the hardware configuration of the controller in a substrate processing apparatus according to an embodiment of the present disclosure.

[0022] Figure 5 This is a control block diagram of the internal cooling control unit according to one embodiment of the present disclosure.

[0023] Figure 6 This is a flowchart illustrating the first effective set method used in this disclosure.

[0024] Figure 7 This is a flowchart illustrating the second effective set method used in this disclosure.

[0025] Figure 8This is a control block diagram of the internal cooling control unit of another embodiment of this disclosure.

[0026] Figure 9 This is a control block diagram of the internal cooling control unit during the generation of the rapid cooling prediction model disclosed herein.

[0027] Figure 10 This is a control block diagram illustrating an example of the automatic acquisition process of the rapid cooling prediction model of this disclosure.

[0028] Figure 11 This is a flowchart illustrating an example of a temperature-related process in the film-forming process of an embodiment of the present disclosure.

[0029] Figure 12 It is shown Figure 11 The flowchart shown illustrates the temperature changes inside the furnace.

[0030] Figure 13 It is used for explanation Figure 11 The flowchart shown illustrates the operation of the control unit 200, temperature control unit 64, and cooling control unit 300.

[0031] Figure 14 (A) is a graph showing the furnace temperature and temperature deviation of each zone when temperature control is performed using the cooling control unit of the comparative example. Figure 14 (B) is a graph showing the furnace temperature in each zone and the temperature deviation between zones when temperature control is performed using the cooling control unit of this embodiment.

[0032] Figure 15 (A) is a graph showing the measured value of the furnace temperature, the predicted temperature, and their errors when temperature control is performed in the cooling control unit of this embodiment without using information from the exhaust fan. Figure 15 (B) is a graph showing the measured temperature, predicted temperature, and their errors when temperature control is performed using the cooling control unit of this embodiment. Detailed Implementation

[0033] <One embodiment of the present disclosure>

[0034] Hereinafter, an embodiment of the present disclosure will be described with reference to the accompanying drawings. Furthermore, the drawings used in the following description are schematic, and the dimensional relationships and ratios of the elements shown in the drawings do not necessarily correspond to reality. Additionally, the dimensional relationships and ratios of the elements in the various drawings may not be consistent with each other.

[0035] In this embodiment, such as Figures 1-3As shown, the substrate processing apparatus 10 of this disclosure is configured as a processing apparatus 10 for performing processing steps in a semiconductor device manufacturing method.

[0036] Figure 1 The substrate processing apparatus 10 shown includes a process tube 11, which serves as a supported longitudinal reaction tube. The process tube 11 is composed of an outer tube 12 and an inner tube 13 arranged concentrically. The outer tube 12 is made of quartz (SiO2) and is integrally formed into a cylindrical shape that is closed at the top and open at the bottom. The inner tube 13 is formed into a cylindrical shape that is open at both the top and bottom. The interior of the hollow part of the inner tube 13 forms a processing chamber 14 for loading a boat (described later), and the lower opening of the inner tube 13 forms a furnace opening 15 for loading and unloading the boat. As described later, the boat 31 is configured to hold multiple wafers, which serve as substrates, in a long arrangement. Therefore, the inner diameter of the inner tube 13 is set to be larger than the maximum outer diameter of the wafer 1 being processed (e.g., a diameter of 300 mm).

[0037] The lower end between the outer tube 12 and the inner tube 13 is hermetically sealed using a manifold 16, which is constructed in a generally cylindrical shape. The manifold 16 can be easily installed and removed from the outer tube 12 and the inner tube 13 for replacement. By supporting the manifold 16 on the frame 2 of the CVD apparatus, the process tube 11 is mounted vertically. In the figures thereafter, there are also instances where only the outer tube 12 is shown as the process tube 11.

[0038] By utilizing the gap between the outer pipe 12 and the inner pipe 13, an exhaust passage 17 with a circular annular cross-section of a certain width is formed. For example... Figure 1 As shown, one end of an exhaust pipe 18 is connected to the upper part of the side wall of the manifold 16, and the exhaust pipe 18 extends to the lowermost end of the exhaust passage 17. An exhaust device 19 controlled by a pressure control unit 21 is connected to the other end of the exhaust pipe 18, and a pressure sensor 20 is connected midway through the exhaust pipe 18. The pressure control unit 21 is configured to perform feedback control on the exhaust device 19 based on the measurement results from the pressure sensor 20.

[0039] A gas inlet pipe 22 is provided below the manifold 16, leading to the furnace opening 15 of the inner tube 13. A raw material gas supply device and an inactive gas supply device (hereinafter referred to as the gas supply device) 23 are connected to the gas inlet pipe 22. The gas supply device 23 is configured to be controlled by a gas flow control unit 24. The gas introduced into the furnace opening 15 from the gas inlet pipe 22 flows within the processing chamber 14 of the inner tube 13, passes through the exhaust passage 17, and is discharged through the exhaust pipe 18.

[0040] A sealing cap 25, with its lower opening closed, is connected to the manifold 16 from the lower vertical direction. The sealing cap 25 is constructed in a disc shape with an outer diameter approximately the same as that of the manifold 16, and is configured to move vertically via a boat lift 26 installed in the waiting room 3 of the frame 2. The boat lift 26 consists of a motor-driven feed screw shaft assembly and a bellows, etc., and the motor 27 of the boat lift 26 is controlled by a drive control unit 28. A rotating shaft 30 is rotatably supported on the centerline of the sealing cap 25, and the rotating shaft 30 is driven to rotate by a rotating mechanism 29, which is controlled by the drive control unit 28. A boat 31 is vertically supported at the upper end of the rotating shaft 30.

[0041] The boat 31 has a pair of end plates 32 and 33 at the top and bottom, and three holding members 34 vertically mounted between them. Multiple holding grooves 35 are equally spaced along the long side of each of the three holding members 34. The holding grooves 35 in the same segment of the three holding members 34 are configured to open opposite each other. The boat 31 holds multiple wafers 1 horizontally and with their centers aligned by inserting wafers 1 between the holding grooves 35 in the same segment of the three holding members 34. A heat insulation cover 36 is disposed between the boat 31 and the rotation shaft 30. The rotation shaft 30 is configured to support the boat 31 in a state where it is lifted from the upper surface of the sealing cover 25, thus separating the lower end of the boat 31 from the furnace opening 15 by an appropriate distance. The heat insulation cover 36 insulates the area near the furnace opening 15.

[0042] Outside the process tube 11, heater units 40, which are longitudinally placed heating devices, are arranged concentrically and supported by the frame 2. Each heater unit 40 includes a housing 41. The housing 41 is made of stainless steel (SUS) and is formed into a cylindrical shape, closed at the top and open at the bottom, preferably cylindrical. The inner diameter and total length of the housing 41 are set to be greater than the outer diameter and total length of the outer tube 12. Furthermore, in this embodiment, the heater unit 40 is divided into seven control zones from the upper end to the lower end: U1, U2, CU, C, CL, L1, and L2.

[0043] A heat-insulating structure 42 is provided inside the housing 41. In this embodiment, the heat-insulating structure 42 is formed in a cylindrical shape, preferably in a cylindrical shape, and the side wall portion 43 of the cylindrical body is formed in a multi-layer structure. In addition, it includes a partition portion 105 that divides the side wall portion 43 into multiple areas in the vertical direction, and a heating element 56 that is provided on the inner side of the side wall portion 43 and heats the wafer 1 of the processing chamber 14.

[0044] Heater unit 40 Figure 3The configuration shown is controlled by a temperature control unit 64. In addition, in the heater unit 40, thermocouples 65 and 66 are provided in pairs for each control zone, corresponding to control zones U1, U2, CU, C, CL, L1, L2.

[0045] Thermocouple 65 is a heater thermocouple that detects the temperature between the outer tube 12 and the heater unit 40 in each control zone. Thermocouple 65 is configured to measure the ambient temperature near the heating element 56 in each control zone. The temperature detected by thermocouple 65 will be used as the heater temperature below. Alternatively, the temperature of the heating element 56 can also be used as the heater temperature.

[0046] Thermocouple 66 is a cascaded thermocouple used to detect the temperature between the outer tube 12 and the inner tube 13 in each control zone. Thermocouple 66 is configured to measure the furnace temperature, which is the temperature of the processing chamber 14 in each control zone. The temperature detected by thermocouple 66 will be referred to as the furnace temperature below.

[0047] The temperature control unit 64 is configured to adjust the energization of the heating element 56 in each control zone based on the temperature information detected by the thermocouples 65 and 66 in each control zone, so as to control the temperature of the processing chamber 14 at the desired timing so that the temperature is set by the control unit 200.

[0048] Additionally, unidirectional dampers 104, serving as anti-backward diffusion sections, are provided in each area within the housing 41. These unidirectional dampers 104 are configured to supply cooling gas 90, acting as a cooling medium, to the internal space 75 via the gas flow path 107 by opening and closing. When no cooling gas 90 is supplied from a gas source (not shown), the unidirectional dampers 104 are closed to prevent backflow of ambient gas in the internal space 75. Alternatively, the opening pressure of the unidirectional dampers 104 can be varied according to the area. Furthermore, a heat-insulating cloth, acting as a blanket, is provided between the outer peripheral surface of the sidewall portion 43 and the inner peripheral surface of the housing 41 to absorb the thermal expansion of the metal.

[0049] like Figure 1 As shown, the top wall portion 80, serving as the top, covers the upper end of the side wall portion 43 of the heat-insulating structure 42 in a manner that encloses the internal space 75. An exhaust port 81, which forms part of the exhaust path for discharging ambient gas from the internal space 75, is formed in the top wall portion 80 in a ring shape. The lower end of the upstream side of the exhaust port 81 extends into the internal space 75. The downstream side of the exhaust port 81 is connected to an exhaust pipe 82. The exhaust pipe 82 is connected to an exhaust fan 84. The exhaust fan 84 is configured to supply cooling gas 90, which serves as a cooling medium, to a cooling unit, which is described later, and discharge it via the exhaust pipe 82.

[0050] The pressure control unit 21, gas flow control unit 24, drive control unit 28, temperature control unit 64, and cooling control unit 300 are configured to be electrically connected to the control unit 200 for communication. The pressure control unit 21, gas flow control unit 24, drive control unit 28, temperature control unit 64, and cooling control unit 300 (described later) are configured to perform control according to the instructions of the control unit 200.

[0051] [Composition of cooling unit 301]

[0052] Next, use Figure 2 The cooling unit 301 in this embodiment will be described in detail.

[0053] In this embodiment, the cooling unit 301 is divided into multiple cooling zones (U1, U2, CU, C, CL, L1, L2) corresponding to multiple control zones. Each cooling zone is configured with an intake pipe 101 for supplying cooling gas 90, a cooling valve 102 located on the intake pipe 101 and acting as a flow guide valve to regulate the gas flow rate, and multiple openings (quench holes) 110 for ejecting cooling gas toward the process pipe 11. The cooling valve 102 supplies cooling gas 90, which serves as the cooling medium, to the internal space 75 between the heater unit 40 and the process pipe 11.

[0054] The cooling valve 102 is configured to adjust the flow rate of the cooling gas 90 introduced into the intake pipe 101 by opening and closing the cooling valve 102, corresponding to the ratio of the length of each cooling zone. This adjusts the flow rate and velocity of the gas ejected from the opening 110 toward the process pipe 11. In other words, by adjusting the opening degree of the cooling valve 102 in accordance with the components within the intake pipe 101 using the cooling control unit 300, the flow rate and velocity of the cooling gas 90 introduced into each cooling zone can be changed. Thus, the cooling valve 102 is configured to have a different opening degree controlled for each cooling zone. The cooling valve 102 is configured to be controllable by the cooling control unit 300.

[0055] Additionally, a one-way damper 104 is provided downstream of the cooling valve 102 on the intake pipe 101 to prevent back diffusion of ambient gas from the processing chamber 14. Cooling gas 90 is discharged from the exhaust port 81 located on the upper side of the internal space 75. Therefore, the one-way damper 104 is provided in each zone to effectively store cooling gas 90 and prevent convection between the intake pipe 101 and the thermal insulation structure 42 when quenching is not used.

[0056] Additionally, cooling zones (e.g., in the area where the wafer 1 is held on the boat 31) extend from approximately the same height as the uppermost layer of the area where the wafer 1 is held to the lowermost layer of the area where the wafer 1 is held. Figure 2In the cooling zone, the cooling gas 90 blown out by U2, CU, C, CL, L1) is provided with openings 110 so that the flow rate and velocity are equal. Specifically, the openings 110 are arranged at equal intervals in the circumferential and vertical directions in the cooling zone, and are configured to blow out into the internal space 75 through the gas flow path 107.

[0057] The heat insulation structure 42 used in the heater unit 40 is also used as a cooling unit 301. The heat insulation structure 42, as described above, has a cylindrical sidewall portion 43, which is formed in a multi-layered structure. This sidewall portion 43 is configured to divide the space into multiple cooling zones (U1, U2, CU, C, CL, L1, L2) along the vertical direction. For example, the partition portion can be configured to separate the sidewall portion 43 into multiple cooling zones (U1, U2, CU, C, CL, L1, L2) in the vertical direction, or it can be configured to provide a space between the partition portion 105 and the sidewall portion 43. The gas flow path 107 is configured to connect the suction pipe 101 to the internal space 75, and to blow cooling gas 90 into the internal space 75 through the opening 110 for each cooling zone.

[0058] In addition, the opening 110 is configured such that the blown cooling gas 90 avoids the heating element 56.

[0059] Furthermore, in this embodiment, the partition 105 is configured such that the number of control zones and the number of cooling zones are the same. The number of control zones and the number of cooling zones can be set in any configuration, regardless of the specific configuration.

[0060] The exhaust pipe 82 is configured to be connected to the exhaust fan 84, and the cooling gas 90 is discharged through the exhaust function of the exhaust fan 84.

[0061] Furthermore, the cooling control unit 300 is configured to be electrically connected to the cooling valve 102 and to indicate the opening degree of the cooling valve 102. Additionally, the cooling control unit 300 is configured to be electrically connected to the exhaust fan 84 and to act as a switch to indicate the operation of the exhaust fan 84.

[0062] In this embodiment, the cooling unit 301 adjusts the opening of the cooling valve 102 for each cooling zone via the cooling control unit 300, and then simultaneously starts the exhaust fan 84. As a result, the flow rate of the supplied cooling gas for each cooling zone can be adjusted, and the cooling capacity for each cooling zone can be adjusted.

[0063] [Structure of the Control Unit]

[0064] Next, the configuration of the control unit 200 will be shown as an example.

[0065] like Figure 4As shown, the control unit 200 includes: a controller body 203 comprising a CPU (Central Processing Unit) 201 and a memory 202, etc.; a communication interface 204 as a communication unit; a storage device 205 as a storage unit; and a display / input device 206 as an operation unit. In other words, the control unit 200 includes components that are typical of a controller.

[0066] CPU 201 forms the central part of the operation unit, executes the control program stored in storage device 205, and executes the process (e.g., process for manufacturing) recorded in storage device 205 according to the instructions from display / input device 206.

[0067] Furthermore, the recording medium 207 used to store the operating program of the CPU 201 may include ROM (Read Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), flash memory, hard disk, etc. Here, RAM (Random Access Memory) functions as the CPU's working area.

[0068] The communication unit 204 is electrically connected to the pressure control unit 21, the gas flow control unit 24, the drive control unit 28, the temperature control unit 64, and the cooling control unit 300 (sometimes collectively referred to as sub-controllers), and is able to send and receive data related to the operation of each component.

[0069] The storage unit 205 has a program storage area for storing documents such as the aforementioned process procedures. This program storage area stores programs that cause the substrate processing apparatus to execute a sequence of controlling the power supply to the heater in a manner that makes the predicted temperature series calculated according to a pre-preserved furnace temperature prediction model approach the future target temperature series, and / or programs that cause the substrate processing apparatus to execute a sequence in which the opening of the cooling valve 102 is adjusted according to a quench prediction model to minimize the error between the predicted temperature series and the target temperature series calculated using the ratio of the change from the current target temperature to the final target temperature that will become the future target. The prediction model includes information about the exhaust fan 84 (described later in this embodiment), the final target temperature, and the opening of the cooling valve 102, and predicts at least one of the following temperatures: the heater temperature and the furnace temperature. Additionally, a parameter storage area (not shown) stores at least various parameters for implementing the aforementioned prediction models. Furthermore, at least one prediction model within a specified temperature range is stored.

[0070] In one embodiment of this disclosure, a control unit 200 has been described as an example, but the implementation is not limited to this; a conventional controller system can be used. For example, the above-described process can also be executed by installing the program onto a general-purpose controller from a recording medium 207 such as a CD-ROM or USB, which stores a program for performing the above-described process. Alternatively, a communication unit 204, such as a communication line, communication network, or communication system, can be used. In this case, for example, the program can be published on a bulletin board of the communication network, and the program can be provided via the network and overlay. Then, the program provided in this way can be started and executed in the same way as other applications under the control of the OS (Operating System), thereby enabling the execution of the above-described process.

[0071] [Composition of the Cooling Control Unit]

[0072] Next, use Figure 5 Explain the control configuration of the cooling control unit 300.

[0073] The cooling control unit 300 consists of a furnace temperature acquisition unit 351, a temperature history record storage unit 353, an exhaust history record storage unit 355, a valve opening history record storage unit 357, an individual characteristic creation unit 359, a target temperature series creation unit 361, a comprehensive characteristic creation unit 363, a constraint optimization calculation unit 365, and an opening signal supply unit 367.

[0074] The target temperature from the control unit 200 is input to input terminal S. The furnace temperature from thermocouple 66 is input to input terminal F. The on / off signal information of the exhaust fan 84 from the control unit 200 is input to input terminal L. The target temperature, input terminals S, and input terminals F correspond to the number of thermocouples 66, but... Figure 4 Since they are all of the same composition, only one is shown in the diagram.

[0075] As described above, each cooling zone is equipped with an intake pipe 101 and a cooling valve 102, but... Figure 5 For illustrative purposes, only one illustration is shown. That is, the cooling valve 102 can be set to different opening degrees in each zone, supplying cooling gas to the intake pipe 101 in each zone.

[0076] Thermocouples 66 are arranged in the same number as each zone in the cooling zone of the inner tube 13, and convert the temperature near the wafer 1 into a small voltage and output it.

[0077] The cooling control unit 300 is configured to acquire input signals from input terminals S, F, and L at minute intervals according to a pre-set control cycle, and update the output signal at minute intervals.

[0078] The furnace temperature acquisition unit 351 acquires the minute electrical current from the thermocouple 66, smooths it to remove noise, and converts it into a detected temperature according to this physical characteristic. In other words, the furnace temperature acquisition unit 351 acquires the furnace temperature detected by the thermocouple 66. The furnace temperature acquisition unit 351 has a number corresponding to the number of thermocouples 66.

[0079] The temperature history storage unit 353 inputs the furnace temperature or heater temperature of the entire area from the furnace temperature acquisition unit 351 and stores this data for a certain period in the temperature history storage area. The temperature history storage unit 353 writes data sequentially at predetermined intervals relative to the temperature history storage area, starting from the initially acquired temperature. When the temperature history storage area is full, the oldest data is discarded, and new data is written to that location. In this way, the temperature history storage unit 353 is configured to store past temperatures from the present to a certain period.

[0080] To standardize the time-related expression, the temperature written at a specific time t is considered as the temperature one step earlier (e.g., expressed as y(t-1) shown in Equation 1). The input temperature is the average temperature calculated from the electromotive force of thermocouple 66 up to the writing time.

[0081] The exhaust history storage unit 355 receives the switch signal of the exhaust fan 84 from the control unit 200 and stores the data related to the input switch signal of the exhaust fan 84 in the exhaust history storage area for a certain period of time.

[0082] The valve opening history storage unit 357 receives opening information from the cooling valves 102 that are supplied to the entire area, and stores this data in the valve opening history storage area for a certain period of time. The valve opening history storage unit 357 writes data sequentially at predetermined intervals, starting from the initially acquired opening. Once the valve opening history storage area is full, the oldest data is discarded, and new data is written to that location. In this way, the valve opening history storage unit 357 is configured to store only the openings from the current time up to a certain period of time prior.

[0083] To standardize time-related expressions, the opening value written at a specific time t is considered to be the temperature before the first operation (e.g., V as shown in Equation 1). a (Expressed as in (t-1)). The input opening is the opening calculated in the previous processing and will continue to be output up to this time.

[0084] The individual characteristic creation unit 359 obtains a rapid cooling prediction model for a specific cooling zone (described in detail later) from the storage unit 205, obtains specified past temperature data of the furnace temperature or heater temperature from the temperature history storage unit 353, obtains data related to specified past on / off states of the exhaust fan 84 from the exhaust history storage unit 355, and obtains specified past opening data of the cooling valve 102 from the valve opening history storage unit 357. Hereinafter, the individual input response characteristic matrix S, as described in Equations 2 and 3, is calculated. sr and individual zero-response characteristic vectors S zr Individual input response characteristic matrix S sr and individual zero-response characteristic vectors S zr Only the number of furnace temperatures set as the controlled objects (= the number of zones) is calculated. As described above, the quench prediction model is obtained from the control unit 200, but a quench prediction model storage unit may also be provided within the cooling control unit 300, for example. The above is merely one example.

[0085] [Quick Cooling Prediction Model]

[0086] A quench prediction model is a mathematical formula used to calculate the predicted temperature for at least one of the heater temperature and the furnace temperature, using the following formula 1.

[0087] [Mathematical Expression 1]

[0088]

[0089] Here, Let y(t-1) be the predicted temperature at time t, and y(t-1) be the temperature one step earlier (the previous time).

[0090] V a (t-1), V a (t-2), ..., V a (tn) represents the opening degree of the cooling zone before one iteration, two iterations, ..., n iterations.

[0091] V b (t-1), V b (t-2), ..., V b (tn) represents the opening degree of the zone adjacent to one side of the cooling zone before the first, second, ..., nth time iterations.

[0092] V c (t-1), V c (t-2), ..., V c (tn) represents the opening degree of the zone adjacent to the other side of the cooling zone before the first, second, ..., nth time iterations.

[0093] f(t-1), f(t-2), ..., f(tm) are the data for the exhaust fan being turned on (=1) and off (=0) one time before, two times before, ..., m times before.

[0094] y0 is the reference temperature, for example, assumed to be near room temperature. The reference temperature y0 is the temperature within the range of 20°C to 30°C. a1, ..., a n b1, ..., b n c1, ..., c n d and d are specified coefficients. n and m are arbitrarily set values, representing the necessary amount of past data. Predictive models are stored for each cooling zone and can be used for control calculations. In other words, the quench prediction model corresponds to various temperature ranges.

[0095] According to Equation 1, when the temperature before the first step is the reference temperature y0, "y(t-1)-y0" is zero, and the result is as follows:

[0096] [Mathematical Expression 2]

[0097]

[0098] The predicted temperature equals the temperature before the first test, which equals the reference temperature y0. If the furnace temperature is assumed to be room temperature, and the cooling valve 102 is set to fully open, the furnace temperature will remain unchanged even with the maximum flow of cooling gas (room temperature). This indicates that Equation 1 is appropriate at the reference temperature.

[0099] Furthermore, according to Equation 1, it is shown that, for example, if a quench prediction model is set to be related to the temperature of zone C, then depending on each coefficient, not only the opening degree of the cooling valve 102 in zone C will be affected by the predicted temperature, but also the opening degree of the cooling valve 102 in adjacent zones CU or CL. Thus, the mutual interference between zones can be expressed depending on each coefficient. Additionally, in the quench prediction model of Equation 1, the data of the opening (=1) and closing (=0) of the exhaust fan 84 are used as constant terms, thereby improving the temperature error of the zone at the start of quench. Here, when the exhaust fan 84 is started, all cooling valves 102 are closed until just before quench begins; therefore, it can be assumed that there is no impact on the furnace temperature. However, by considering the effect of the exhaust fan 84's operation on the exhaust gas in the furnace environment during the small time at the start of quench, the temperature error of the zone at the start of quench is improved, and temperature controllability is enhanced.

[0100] Here, adjacent zones are pre-defined considering cooling characteristics. For example, sometimes two adjacent zones are needed due to interference. In addition, depending on the characteristics of the cooling unit, the cooling gas flows upward within the internal space 75, so there are cases where only two adjacent zones on the vertically lower side are defined.

[0101] Furthermore, according to Equation 1, even when the opening degree of the cooling valve 102 in this zone or adjacent zones is all zero (= fully closed), the predicted temperature change will still occur due to the term related to d. Thus, it is possible to express cooling other than cooling from the cooling gas 90 from the intake pipe 101, for example, cooling performed by natural cooling or unintentional intermittent airflow.

[0102] Equation 1 above is expressed by the state-space model shown in Equation 2 below.

[0103] [Mathematical Expression 3]

[0104]

[0105] Here, matrices A, B, and C are shown below. Furthermore, for simplicity, n = 4 and m = 3.

[0106] [Mathematical Expression 4]

[0107]

[0108] [Mathematical Expression 5]

[0109]

[0110] C = [1 0 0 0 0 0 0 0 0 0]

[0111] In addition, the vectors x(t), u(t), and the output y(t) are shown below.

[0112] [Mathematical Expression 6]

[0113]

[0114] [Mathematical Expression 7]

[0115]

[0116] Here, Y = {y(t) - y0}.

[0117] In Equation 2, if u(t) is input at time t and u(t) is input thereafter, the predicted temperature after t+1 becomes Equation 3.

[0118] [Mathematical Expression 8]

[0119]

[0120] [Mathematical Expression 9]

[0121] Here,

[0122]

[0123] This is the predicted temperature vector.

[0124] In Equation 2, for simplicity, n = 4 and m = 3 are given as examples, but Equation 3 is not limited to these. Furthermore, in Equation 3, S... zr S is an individual zero-response characteristic vector. sr This is the individual input response characteristic matrix.

[0125] The number of rows for each row is calculated based on the number allowed, which depends on the control cycle and the processing power of the CPU used by the controller.

[0126] Individual zero response characteristic vector S zr This represents the amount of change in the predicted temperature vector due to past temperatures and the past opening degree of cooling valve 102. Additionally, the individual input response characteristic matrix S... sr This represents the amount of change in the predicted temperature vector due to the influence of the calculated opening degree of the cooling valve 102.

[0127] The following section uses the corresponding region to distinguish the individual input response characteristic matrix S. sr Individual zero-response characteristic vector S zr And when predicting the temperature vector, the individual input response characteristic matrix corresponding to region a is represented by S. sr -a represents the individual zero-response characteristic vector corresponding to region b, denoted by S. zr -b and other expressions.

[0128] [Mathematical Expression 10]

[0129] In addition, the predicted temperature vector corresponding to region e is used Expressions such as...

[0130] When the set temperature is updated, the target temperature column creation unit 361 inputs the target temperature, the current target temperature, and the heating / cooling rate from the control unit 200, and calculates the individual target temperature column vector S. tg The heating / cooling rate indicates the rate of change from the current target temperature to the final target temperature that will become the future target. For example, based on a setting of 1°C / min, it shows an indication of a change at a rate of 1°C within 1 minute. Information input from the control unit 200, for example, if the current target temperature is 100°C, the updated set temperature is 200°C, and the heating / cooling rate is 10°C / min, then the target temperature = 200°C, the current target temperature = 100°C, and the heating / cooling rate = 10°C / min. Subsequently, before the current target temperature reaches 200°C, for example, if it is 150°C, and the set temperature is 300°C with the heating / cooling rate updated to 1°C / min, then the target temperature = 300°C, the current target temperature = 150°C, and the heating / cooling rate = 1°C / min.

[0131] Then, the target temperature column creation unit 361 creates individual target temperature column vectors S when the heating and cooling rates are zero and when they are other than zero. tg Switch to another device.

[0132] First, when the heating and cooling rates are zero, the target temperature column creation unit 361 calculates the individual target temperature column vector S according to the following reference settings. tg .

[0133] (1) Temperature rise / fall deviation = target temperature - current target temperature

[0134] (2) Heating and cooling time = Absolute value (temperature deviation) ÷ Reference heating and cooling rate

[0135] (3) Reference setpoint = Current target temperature + Temperature rise / fall deviation × (1 - exp(Elapsed time ÷ (Heating / fall time ÷ Time constant))

[0136] Set the time constant, for example, to 1.0.

[0137] Next, when the heating and cooling rates are other than zero, the target temperature column creation unit 361 calculates the individual target temperature column vector S according to the following reference settings. tg .

[0138] (1) Temperature rise / fall deviation = target temperature - current target temperature

[0139] (2) Heating and cooling time = absolute value (temperature deviation) ÷ heating and cooling rate

[0140] (3) Reference setpoint = Current target temperature + Temperature rise / fall deviation × (Elapsed time ÷ Temperature rise / fall time)

[0141] Individual target temperature column vector S tg For the purposes of the following explanation, it will be expressed as in Equation 4.

[0142] [Mathematical Expression 11]

[0143]

[0144] The timing and row number in Equation 4 correspond to those in Equation 3, etc. The number of target temperature column creation units 361 is the same as the temperature that is the controlled object, that is, there are only the same number as thermocouples 66.

[0145] The following section uses corresponding regions to distinguish individual target temperature column vectors S. tg When corresponding to region a, use S tg-a The expression, in the case corresponding to region e, is expressed as S. tg-e Expressions such as...

[0146] The comprehensive feature creation unit 363 inputs the individual input response feature matrix S from multiple individual feature creation units 359. sr and individual zero-response characteristic vectors S zr And from multiple target temperature column creation units 361, individual target temperature column vectors Stg are input to create a comprehensive characteristic equation.

[0147] First, make the individual input response characteristic matrix S sr Deformation. Individual input response characteristic matrix S sr This expresses the change in predicted temperature when u(t) is input at time t and thereafter continues to be input. If different values ​​u(t) ~ u(t+Np-1) are input at all control timings instead of keeping u(t), then the second term on the right side of Equation 3 is as follows. Furthermore, let Np be the row number of Equation 3.

[0148] [Mathematical Expression 12]

[0149]

[0150] In known model predictive control, it is assumed that different values ​​u(t) to u(t+Np-1) are input to the timing of all operations, and these are calculated to obtain the result. However, the processing performance of the CPU of the cooling control unit 300 is insufficient. Therefore, in this disclosure, the second term on the right side of expression 3 is set as follows by setting the input mode to fixed.

[0151] [Mathematical Expression 13]

[0152]

[0153] [Mathematical Expression 14]

[0154]

[0155] As described above, the individual input response characteristic matrix S is... sr By transformation, we obtain Equation 5 from Equation 3.

[0156] [Mathematical Expression 15]

[0157]

[0158] [Mathematical Expression 16]

[0159]

[0160] In Equation 5, change S dsr This is used as the individual input response characteristic matrix. When using corresponding regions for differentiation, the individual input response matrix corresponding to region a is expressed as S. dsr-a wait.

[0161] Next, regarding Equations 5 and 4 above, we will arrange them in all cooling zones that are the objects of control.

[0162] [Mathematical Expression 17]

[0163]

[0164] [Mathematical Expression 18]

[0165]

[0166] As described above, the comprehensive characteristic creation unit 363 calculates the comprehensive input response characteristic matrix U expressed by equations 6 and 7. dsr Comprehensive zero-response characteristic vector U zr Comprehensive target temperature vector U tg And output it.

[0167] The constraint optimization calculation unit 365 inputs the comprehensive input response characteristic matrix U from the comprehensive characteristic creation unit 363. dsr Comprehensive zero-response characteristic vector U zr Comprehensive target temperature vector U tg The optimal opening degree is then calculated using a method known as the effective set method, which will be described later.

[0168] The opening signal supply unit 367 has only the number of cooling zone divisions, that is, only the same number as the connected cooling valves 102. It acquires the opening degree corresponding to the constraint optimization calculation unit 365 in a specified control cycle and updates the opening degree indication relative to the cooling valves 102.

[0169] [First Efficient Set Method]

[0170] This describes the first effective set method used in this disclosure.

[0171] The effective set method, based on the constraints of Equation 9 below, finds the solution vector x that maximizes the evaluation function f(x) given by Equation 8 below.

[0172] [Mathematical Expression 19]

[0173]

[0174] b≥A·x…(Equation 9)

[0175] In equations 8 and 9, c, Q, b, and A are assigned constant matrices or vectors. Additionally, the symbol T denotes inversion. At this point, by implementing the effective set method, it is possible to... Figure 6 The process shown yields the solution vector x.

[0176] In S201, the solution x of equation 9 is not within a valid range.k Then, let A be the set of rows in Equation 9 where the equality sign is valid. e b e In S201, A e b e All are empty sets. Furthermore, let A be the set of rows in Equation 9 where the equality sign is not valid. d b d In S201, A d =A, b d =b.

[0177] In S203, solve the next simultaneous equation, and let the solution be x and λ. If x = x k Then proceed to S205. If x ≠ x k Then proceed to S207.

[0178] [Mathematical Expression 20]

[0179]

[0180] In S205, determine whether all elements of λ are 0 or higher. If they are 0 or higher, proceed to S213. If not all elements of λ are 0 or higher, proceed to S211.

[0181] In S207, α is calculated according to Equation 10 below. i a i From A respectively d b d Extract the value of row 1. If α = 1, proceed to S205. If α < 1, proceed to S209.

[0182] [Mathematical Expression 21]

[0183]

[0184] In S209, from A d b d Remove the constraints used when calculating α (<1) according to Equation 10 [b] i a i}, and to A e b e Add to it, then proceed to S203.

[0185] In S211, the element with the smallest negative value λ is selected from A. e b e Delete the contents contained in A e b e The corresponding constraint in the inner constraint [b i a i}, and to Ad b d Add to it, then proceed to S203.

[0186] In S213, the solution x obtained in S203 is taken as the optimal solution and the process ends.

[0187] Figure 6 The effective set method shown explores the combinations of rows in Equation 9 where the equality sign is valid by using an additional multiplier λ, and can find the solution that satisfies Equation 9 and sets Equation 8 to the maximum.

[0188] [Control applied to the effective set method]

[0189] Next, the application method of the effective set method in this disclosure will be explained.

[0190] In the comprehensive characteristic creation unit 363, the predicted temperature column (predicted temperature vector) of the furnace temperature in Equation 6 and the target temperature column (comprehensive target temperature vector) in Equation 7 can be obtained. Therefore, in the constraint optimization calculation unit 365, the square of the error between the target temperature column and the predicted temperature column is used as the evaluation function. The evaluation function V(u(t)) becomes as shown in Equation 11 below.

[0191] [Mathematical Expression 22]

[0192]

[0193] If we compare the inside of the parentheses outside the second term of Equation 11 with Equation 8, then c and Q in Equation 8 can be replaced by the following expressions respectively.

[0194] [Mathematical Expression 23]

[0195] c = U dsr T ·[U tg -U zr ], Q = U dsr T ·U dsr

[0196] Therefore, using the aforementioned effective set method, the solution within the parentheses surrounding the second term of Equation 11 can be maximized. Thus, the solution that minimizes the evaluation function V(u(t)) can be found, and an evaluation function V(u(t)) that minimizes the square of the error between the target temperature series and the predicted temperature series can be created. The simultaneous equations are calculated in such a way that this evaluation function V(u(t)) is minimized. Then, by solving these simultaneous equations, the opening degree of the cooling valve 102, which is included in the solution of the predicted temperature series, can be obtained, and the cooling control unit 300 adjusts the opening degree of the cooling valve 102.

[0197] Next, regarding Equation 9 related to the constraints, for the sake of simplicity, only the portion concerning the opening degree of zones a to c is shown in Equation 12 below, where the power supply value k for each zone... a k b k c By assigning upper and lower limits to the left side of the arrow, respectively, and establishing inequalities as shown on the right side of the arrow, we can fit to Equation 9. In Equation 12 below, LL a UL a LL represents the upper and lower limits of the electricity supply value relative to zone a, respectively. b UL b LL c UL c Similarly, these represent the upper and lower limits of the power supply value relative to zone b and zone c, respectively. For example, set as LL. a =0%, UL a =100%.

[0198] [Mathematical Expression 24]

[0199]

[0200] [Second Effective Set Method]

[0201] Next, the second effective set method that can be used in this disclosure will be explained. In the above... Figure 6 In the effective set method shown, when the CPU's processing power is insufficient, there are cases where the calculation does not finish within the specified control cycle. Therefore, it replaces... Figure 6 The process is set up in Figure 7 The solution vector x can be obtained in the process.

[0202] and Figure 6 The difference between the first effective set method and the previous one is that S215 is added immediately after the start, changing the processing from S201 to S217, and proceeding from S209 and S211 to the added S219, advancing to S203 or S213 through the decision in S219. The following only explains the differences from the first effective set method.

[0203] In S215, the loop count is initialized.

[0204] Then, in S217, select the solution x for which the equality sign of equation 9 is not in a valid range. k To prepare for the possibility of the optimization calculation ending midway in S219 described later, the selected solution is specifically set to the upper limit of the range where the equality sign in Equation 9 is not valid. For example, in region a, the opening degree is 0 ≤ V. a When (t)≤100, the choice solution is set to V. a(t) = 99.9, etc. By making choices in this way, the constraints added in S209 become the upper limit of the priority constraints, so that even if the optimization calculation ends midway, a safe calculation result can be obtained.

[0205] In S219, the number of iterations is counted. If it is within the specified number, proceed to S203. If it exceeds the specified number, proceed to S213, take the solution x obtained in S203 as the optimal solution, and end the process.

[0206] By setting Figure 7 This process allows for the calculation of the optimal solution with minimal processing, thus enabling the calculation to be completed within a specified control period.

[0207] <Second Embodiment of the Present Disclosure>

[0208] Next, a second embodiment of this disclosure will be described. In the cooling control unit 300 of the second embodiment of this disclosure, the temperature detected by the thermocouple 65 is input to the furnace temperature acquisition unit 351 instead of the thermocouple 66. That is, the furnace temperature acquisition unit 351 acquires the heater temperature detected by the thermocouple 65 and controls it according to the target temperature. Therefore, even without the thermocouple 66, the same effect as that achieved by the above embodiment can be obtained by utilizing the temperature detected by the thermocouple 65.

[0209] <Third Embodiment of the Present Disclosure>

[0210] Next, the third embodiment of this disclosure will be described.

[0211] Figure 8 This is a control block diagram of the internal structure of the cooling control unit 300 according to the third embodiment of this disclosure. In the third embodiment, instead of Figure 5 The control block diagram shown uses a comprehensive characteristic creation unit 369 instead of a comprehensive characteristic creation unit 363, and an optimization calculation unit 371 is used instead of a constraint optimization calculation unit 365. Hereinafter, only those related to the above will be described. Figure 5 The control boxes shown are different parts; detailed descriptions of the same parts are omitted.

[0212] The comprehensive characteristic creation unit 369 inputs the individual input response characteristic matrix S from the individual characteristic creation unit 359, which has a number of divisions. sr and individual zero-response characteristic vectors S zr And from the target temperature column creation unit 361, which has a number of partition numbers, input the target temperature column vector S. tg Create a comprehensive characteristic equation.

[0213] The comprehensive characteristic equations are created using the methods shown in Equations 13 and 14, instead of Equations 6 and 7.

[0214] [Mathematical Expression 25]

[0215]

[0216] [Mathematical Expression 26]

[0217]

[0218] In Equations 13 and 14, the difference with region a is calculated after the second paragraph. Region a can be set as the basis for calculating the difference in advance using parameters, etc. Here, region a is used as the basis for calculating the difference, but it can also be a region other than region a. In addition, the time and row number of Equations 13 and 14 correspond to those of Equations 6 and 7.

[0219] Then, the comprehensive characteristic creation unit 369 calculates the comprehensive input response characteristic matrix U expressed by Equations 13 and 14. dsr Comprehensive zero-response characteristic vector U zr Comprehensive target temperature vector U tg And output it.

[0220] The constraint optimization calculation unit 371 inputs the comprehensive input response characteristic matrix U from the comprehensive characteristic creation unit 369. dsr Comprehensive zero-response characteristic vector U zr Comprehensive target temperature vector U tg The optimal opening degree is calculated using the effective set method described above.

[0221] In the optimization calculation unit 371, the evaluation function is used as follows: for the reference region, a function is used that sums the squares of the errors between the target temperature column and the predicted temperature column; for other regions, a function is used that sums the squares of the differences between the predicted temperature column of that region and the predicted temperature column of the reference region. Here, the function that sums the squares of the differences between the predicted temperature column of that region and the predicted temperature column of the reference region is considered in terms of the weight matrix Z. The evaluation function V(u(t)) becomes Equation 15.

[0222] [Mathematical Expression 27]

[0223]

[0224] [Mathematical Expression 28]

[0225] That is, if c = U dsr T ·Z·[U tg -U zr ], Q = U dsr T ·Z·U dsrBy making a replacement, the above-mentioned effective set method can be applied.

[0226] Here, the weight matrix Z is a diagonal matrix in which the weight for the evaluation relative to the deviation from the reference area is set to 1, and the weight for the evaluation relative to the difference between the reference area and other areas is set to Z. Z uses, for example, values ​​from 1 to 10.

[0227] [Mathematical Expression 29]

[0228]

[0229] According to Figure 8 The control method of the cooling control unit 300 shown in the figure can take into account the temperature deviation of the temperature range of each zone when controlling the temperature, and can control the cooling of the temperature set in each zone by roughly using the same temperature history record.

[0230] [The composition of the update process for the quench prediction model]

[0231] This illustrates the automatic acquisition order of the quench prediction model illustrated in Equation 1. Based on this order, the coefficients of the quench prediction model (a1, ..., a1 in Equation 1) are determined. n b1, ..., b n c1, ..., c n , d).

[0232] Figure 9 The diagram illustrates the processing block performed by the cooling control unit 300 when generating the rapid cooling prediction model.

[0233] The random opening signal supply unit 373, based on a command from the control unit 200, instructs the corresponding cooling valve 102 to randomly select an opening degree (hereinafter referred to as random opening degree) from a discrete three-value range. The random opening degree signal supply unit 373 only has the same number of cooling zones as the cooling valves 102. The possible values ​​of the random opening degree and the duration until the change can be input from the control unit 200 or set in advance using parameters, etc.

[0234] The quench prediction model update unit 375, based on commands from the control unit 200, retrieves the quench prediction model from the storage unit 205, necessary past temperature data from the temperature history storage unit 353, necessary past exhaust fan 84 on / off information from the exhaust history storage unit 355, and necessary past valve opening information from the valve opening history storage unit 357. It then calculates the latest quench prediction model obtained at that point in time, updates it, and records it again. The quench prediction model update, after initiation, is repeated at a pre-defined cycle, with the action only repeated for a pre-defined time before ending.

[0235] The quench prediction model update unit 375 only contains the number of cooling zones, that is, only the same number as the cooling valves 102. The number of terms in the quench prediction model (the values ​​of n and m in Equation 1) or the mutual interference situation (Equation 1) can be input from the control unit 200 or can be set in advance using parameters.

[0236] [Update methods for quench prediction models]

[0237] Next, the update method for the quench prediction model using the quench prediction model update unit 375 will be described. The update method in this disclosure uses a method called sequential least squares. Equation 17 below is expressed using a matrix / vector.

[0238] [Mathematical Expression 30]

[0239] The predicted value of the difference between the predicted temperature and the temperature a day ago. As shown in Equation 17 below.

[0240]

[0241] Here, let Y = y(t-1) - y0.

[0242] Here, time t represents this processing step, and the latest data among the elements of x(t) becomes V. a (t-1) is because, as mentioned above, the time when the opening value obtained from this process is set to t-1.

[0243] The latest coefficients θ(t) of the prediction model are calculated in the manner described in 18 below.

[0244] [Mathematical Expression 31]

[0245] θ(t)=θ(t-1)+η(t)·k(t)…(Formula 18)

[0246]

[0247] η(t)=Δy(t-1)-x T (t-1)·θ(t-1)

[0248]

[0249] Here, Δy(t-1) is the difference between the temperature obtained this time and the temperature obtained last time (=y(t-1)-y(t-2)). P is a parameter called the forgetting coefficient, which is preset as a parameter. P(t) is the coefficient error correlation matrix, which is recorded along with the rapid cooling prediction model each time it is updated. The initial values ​​are set, for example, 100~1000 as the unit matrix of the features.

[0250] The coefficients θ(t) of the quench prediction model are recorded in a designated storage area within the cooling control unit 300 after a pre-set time has elapsed.

[0251] [Automatic acquisition order of the quench prediction model]

[0252] Next, use Figure 10 This explains the automatic acquisition sequence of the rapid cooling prediction model performed in the cooling control unit 300.

[0253] In S300, the furnace temperature is controlled to the target temperature T1 by means of the control unit 200. At this time, control is performed by the feedback loop of the heater unit 40, the temperature control unit 64, and the thermocouple 66.

[0254] In S304, the drive of the exhaust fan 84 is started (turned on) according to the instruction from the control unit 200, and at the same time, the exhaust fan 84 is driven by the instruction from the control unit 200. Figure 9 The configuration shown begins with updating the rapid cooling prediction model using the cooling control unit 300. (As in...) Figure 9 As explained, the cooling control unit 300 independently instructs the cooling valves 102 of each cooling zone to a random opening degree, and updates the rapid cooling prediction model (Equation 18). If a preset time has elapsed since the start of this step, the random opening degree instruction is stopped, the rapid cooling prediction model is determined, and recorded in a designated storage area within the cooling control unit 300. Furthermore, at the end of this step, the control unit 200 stops (turns off) the drive of the exhaust fan 84.

[0255] In S306, it is determined whether the quench prediction model determined in S304 is appropriate. The determination is based on the number of times S304 has been executed, the convergence status of the prediction model during the execution of S304, or a combination of these.

[0256] The convergence of the prediction model refers to whether the change in the coefficients of the rapid cooling prediction model (=||θ(t)-θ(t-1)||; the criterion for coefficient change) is greater than or less than the threshold.

[0257] If the judgment result is inappropriate (=No), then return to S300 again. If the judgment result is appropriate (=Yes), then the automatic acquisition sequence of the quench prediction model ends. When implementing the temperature control of this disclosure in the cooling step S5 described later, the appropriate quench prediction model determined by the individual characteristic creation unit 359 is read out and used.

[0258] [The second quench prediction model disclosed herein]

[0259] Next, the second quench prediction model of this disclosure will be explained.

[0260] In the quench prediction model of Equation 1 above, in order to achieve sufficiently high accuracy in predicting the temperature, it is sometimes necessary to set the pre-set value of n in Equation 1 to a sufficiently large value. However, the processing power of the CPU of the cooling control unit 300 is insufficient, so sometimes if the value of n is increased, the control operation cannot be completed within the specified control cycle. Therefore, the inventors of this disclosure have discovered a quench prediction model of Equation 19 below, which can replace the quench prediction model of Equation 1 above.

[0261] [Mathematical Expression 32]

[0262]

[0263] Here, for example, natural numbers such as k=2 or k=10 are used.

[0264] y(tk) represents the deviation from the reference temperature relative to the temperature k times ago.

[0265] V a (tk), V a (t-2k), ..., V a (t-nk) represents the opening degree of the cooling zone before k, 2k, ..., nk times.

[0266] V b (tk), V b (t-2k), ..., V b (t-nk) represents the opening of the regions adjacent to one side of the region before k, 2k, ..., nk.

[0267] V c (tk), V c (t-2k), ..., V c (t-nk) represents the opening of the regions adjacent to the other side of the region before k, 2k, ..., nk.

[0268] f(tk), f(t-2k), ..., f(t-mk) are information about the exhaust fan 84 before k, 2k, ..., mk times, related to its on (=1) and off (=0) states. Other elements are the same as in Equation 1 above.

[0269] Then, if k = 1, Equation 19 is consistent with Equation 1 above. That is, if Equation 19 is used as a material for estimating the predicted temperature, the data up to the nth time are used, but in order to reduce the computational load, only the data for each k samples are used.

[0270] Furthermore, if only data per k samples is used, outliers may sometimes be used due to noise, etc. Therefore, the following formula 20 can also be used with data per k samples after low-pass filtering such as simple moving average has been applied.

[0271] [Mathematical Expression 33]

[0272]

[0273] [Mathematical Expression 34]

[0274] Here, for example,

[0275] Or it can be set as follows:

[0276]

[0277] By using Equation 19 or Equation 20 above as a rapid cooling prediction model, the predicted temperature can be obtained with high accuracy even when the frequency component contained in the characteristics is relatively small, and the amount of control calculation can be reduced.

[0278] Next, use Figures 11-13 The temperature timing performed in the substrate processing apparatus 10 is illustrated with an example. Figure 12 The reference numerals S1 to S6 shown in the figures indicate the production process. Figure 11 Each step S1 to S6.

[0279] In step S1, the furnace temperature is maintained at a target temperature T0, which is lower than the processing temperature T1. The control unit 200 inputs the target temperature to the temperature control unit 64. The temperature control unit 64 provides feedback on the temperature detected by thermocouple 66 or thermocouple 65, and controls the power supply value to the power supply circuit 63 based on the target temperature input from the control unit 200, thereby controlling the furnace temperature to maintain the target temperature T0. At this time, the wafer 1 is not moved into the processing chamber 14. From step S1 to step S4 (described later), the control unit 200 shuts off the drive of the exhaust fan 84 and notifies the exhaust history storage unit 355 of the cooling control unit 300 of information related to the shutdown signal of the exhaust fan 84. In addition, from step S1 to step S4 (described later), temperature control is not performed by the cooling control unit 300, and the cooling valve 102 is in a closed state.

[0280] In step S2, if a predetermined number of wafers 1 are loaded into the boat 31, the boat 31 holding one set of wafers is lifted by the boat lift 26 via the sealing cap 25 and moved into the processing chamber 14 of the inner tube 13 (boat loading). The sealing cap 25, having reached its upper limit, is pressed against the manifold 16, thereby sealing the interior of the process tube 11. The boat 31 is stored in the processing chamber 14 while supported by the sealing cap 25. At this time, the temperature of the boat 31 and the wafers 1 is lower than the furnace temperature T0, and the insertion of the wafers 1 held by the boat 31 into the furnace results in the introduction of ambient gas (room temperature) into the furnace. Therefore, the furnace temperature temporarily becomes lower than T0, but under the control of the temperature control unit 64, the furnace temperature stabilizes again at T0 after a short period of time.

[0281] In step S3, exhaust is vented from the interior of the process tube 11 using the exhaust pipe 18. Furthermore, the furnace temperature is gradually increased from temperature T0 to the target temperature T1 used for the prescribed processing of the wafer 1 through timing control by the temperature control unit 64. The error between the actual temperature rise inside the process tube 11 and the target temperature controlled by the temperature control unit 64 is corrected using feedback control based on the measurement results of thermocouples 65 and 66. Additionally, the boat 31 is rotated by the motor 29.

[0282] In step S4, if the internal pressure and temperature of the process tube 11, as well as the rotation of the boat 31, are set to a stable state, the raw material gas is introduced from the gas inlet pipe 22 into the processing chamber 14 of the process tube 11 using the gas supply device 23. That is, the temperature control unit 64 acquires the heater temperature or furnace temperature and the power supply value within a predetermined control cycle, and controls the power supply value output to the heating element 56 by adjusting the power supply value, thereby maintaining and stabilizing the furnace temperature at the target temperature T1. The raw material gas introduced through the gas inlet pipe 22 flows within the processing chamber 14 of the inner tube 13 and is discharged through the exhaust pipe 18 via the exhaust passage 17. While flowing through the processing chamber 14, a predetermined film is formed on the wafer 1 through a thermal CVD reaction where the raw material gas contacts the wafer 1 heated to a predetermined processing temperature.

[0283] If the prescribed processing time has elapsed, in step S5, after the introduction of the processing gas is stopped, a purging gas such as nitrogen is introduced into the interior of the process pipe 11 from the gas inlet pipe 22. Simultaneously, cooling gas 90 is supplied from the suction pipe 101 to the gas flow path 107 via the one-way damper 104. Then, cooling gas 90 is blown into the interior space 75 from multiple openings 110 that serve as cooling gas supply ports. Finally, the cooling gas 90 blown into the interior space 75 from the openings 110 is discharged through the exhaust port 81, the exhaust pipe 82, and the exhaust fan 84.

[0284] In step S5, after the processing of the substrate is completed, the furnace temperature is rapidly reduced (dropped) from temperature T1 to a lower temperature T0. At this time, if the control unit 200 starts (turns on) the exhaust fan 84, it notifies the exhaust history storage unit 355 of the cooling control unit 300 of information related to the start signal of the exhaust fan 84. Then, control is performed based on the control of the cooling control unit 300 to adjust the opening of the cooling valve 102 to obtain the desired temperature trajectory. At this time, temperature control based on the temperature control unit 64 is not performed, and the power supplied to the heater unit 40 is set to zero. That is, the temperature control unit 64 is configured to set the power supply value output to the heating element 56 in each control zone to zero.

[0285] Through the flow of the cooling gas 90, the heat insulation structure 42 and the process pipe 11 are rapidly cooled at a high rate in order to forcibly cool the entire heater unit 40. Furthermore, the cooling gas 90 can be used to isolate the internal space 75 from the processing chamber 14. However, to further improve the cooling effect or prevent corrosion of the heating element 56 at high temperatures due to impurities in the gas, inactive gases such as nitrogen can also be used as the cooling gas.

[0286] If the temperature of the processing chamber 14 drops to the target temperature T0, in step S6, the boat 31 supported by the sealing cover 25 is lowered by the boat lift 26, thereby being removed from the processing chamber 14 (boat unloading). At this time, the control unit 200 shuts off the drive of the exhaust fan 84 and notifies the exhaust history storage unit 355 of the cooling control unit 300 of the information related to the shutdown signal of the exhaust fan 84. In addition, at this time, temperature control based on the cooling control unit 300 is not performed, and the cooling valve 102 is closed.

[0287] Then, if there are remaining wafers 1 that should be processed but have not yet been processed, the processed wafers 1 on the boat 31 are replaced with unprocessed wafers 1, and the series of processes of these steps S1 to S6 are repeated.

[0288] Each of the above steps S1 to S6 proceeds to the next step only after a stable state has been achieved, where the furnace temperature is within a pre-defined small temperature range relative to the target temperature and this state has been maintained for a pre-defined time or longer. Therefore, in the heating step, for example, step S3, rapidly converging the furnace temperature to the target temperature T1 is an important control performance indicator.

[0289] Furthermore, since the same processing is performed on the multiple wafers 1 held by the boat 31, it is desirable to ensure that the furnace temperatures of the multiple control zones follow approximately the same temperature trajectory during these steps. Therefore, reducing the difference between the maximum and minimum values ​​of the furnace temperatures in the multiple control zones (hereinafter referred to as the interval temperature deviation) is an important control performance indicator.

[0290] According to the cooling control unit 300 based on this disclosure, range deviation can be reduced. Furthermore, even when the deviations in the various temperature characteristics of the heater are large, or when the engineer does not have sufficient time, thermal characteristics can be automatically acquired, enabling optimized control methods with minimal or no parameter adjustments. Therefore, the desired device performance can be easily obtained. Additionally, according to the cooling control unit 300 based on this disclosure, by considering the influence of furnace ambient gas exhaust based on the operation of the exhaust fan 84 within a small timeframe at the start of rapid cooling, the temperature error in the range at the start of rapid cooling is improved, thereby enhancing temperature controllability.

[0291] [Example 1]

[0292] Next, use Figure 14 (A) and (B) illustrate Example 1 in which the cooling control unit 300 of this disclosure is applied to the above-described cooling step (step S5).

[0293] Figure 14 (A) shows the application of the comparative example's cooling control unit 300 to the above-mentioned... Figure 11 A graph showing the furnace temperature trajectory in each zone under step S5. Figure 14 (B) illustrates the application of the cooling control unit 300 of this embodiment to the above-described... Figure 11 A graph showing the furnace temperature trajectory in each zone under step S5. The cooling control unit in the comparative example controls the opening of the cooling valve 102 in such a way that the deviation between the temperature detected by the thermocouple outside the reference zone and the temperature detected by the thermocouple in the reference zone is zero.

[0294] right Figure 14 The temperature deviation of the temperature control range shown in (A) is compared with... Figure 14 By comparing the temperature deviation within the temperature control range shown in (B) of this embodiment, it was confirmed that the temperature deviation within the range can be reduced by performing the temperature control of this embodiment. Furthermore, in Figure 14 In fact, a comparison was made. Figure 3 The furnace temperatures in zones L2 to U1 are shown.

[0295] [Example 2]

[0296] Next, use Figure 15(A) and (B) describe the application of the cooling control unit 300 of this disclosure to the above-mentioned Figure 11 Example 2, which describes the cooling step in step S5 of the above.

[0297] Figure 15 (A) is a graph showing the measured value of the furnace temperature, the predicted temperature, and the prediction model error as an error of the temperature control performed in the cooling control unit 300 of this embodiment without using the information of the exhaust fan 84 and using the rapid cooling prediction model. Figure 15 (B) is a graph showing the measured value of the furnace temperature, the predicted temperature, and the prediction model error as an error of the temperature when the cooling control unit of this embodiment is used for temperature control.

[0298] like Figure 15 (A) and Figure 15 As shown in (B), it was confirmed that by using a quench prediction model that utilizes information from exhaust fan 84, the error between the measured quench temperature and the predicted temperature is smaller compared to the case where temperature control is performed using a quench prediction model without utilizing information from exhaust fan 84. In particular, it was confirmed that the prediction model error at the start of quenching can be reduced, thereby improving temperature controllability.

[0299] The above describes in detail the implementation methods of this disclosure. This disclosure is not limited to the above implementation methods and embodiments, and various modifications can be made without departing from the spirit of this disclosure.

[0300] Furthermore, in the above embodiment, temperature control using a rapid cooling prediction model is described in step S5, but temperature control using a prediction model can also be performed in other steps. For example, in step S4, the temperature control unit 64 can be configured to perform control as follows: acquire the heater temperature or furnace temperature and power supply value at a predetermined control cycle; use the prediction model stored in the storage unit 205 to adjust the power supply value output to the heating element 56 in a way that minimizes the deviation between the final target temperature and the predicted temperature; thereby maintaining and stabilizing the furnace temperature at the target temperature T1.

[0301] Furthermore, in the above embodiments, an example of providing a temperature control unit 64 and a cooling control unit 300 separately has been described, but this disclosure is not limited thereto, and the temperature control unit 64 and the cooling control unit 300 may also be provided as a single control unit.

[0302] Furthermore, while the above embodiments illustrate an example of forming a specific film on wafer 200, the types of films disclosed herein are not specifically limited. For example, it can also be appropriately applied to various types of films such as nitride films (SiN films) or metal oxide films formed on wafer 200.

[0303] Furthermore, semiconductor manufacturing apparatuses that process semiconductor wafers, such as the substrate processing apparatus described in the above embodiments, can also be applied to LCD (Liquid Crystal Display) manufacturing apparatuses that process glass substrates.

[0304] Explanation of reference numerals in the attached figures

[0305] 1. Wafer (substrate)

[0306] 40 heater units

[0307] 84 exhaust fan

[0308] 300 Cooling Control Unit

[0309] 301 Cooling Unit.

Claims

1. A substrate processing apparatus, configured to include: The reaction tube forms the processing chamber for processing the substrate inside; A heater unit is disposed outside the reaction tube and has a heating section for heating the substrate; A cooling unit having a cooling valve that supplies cooling medium to the space between the heater unit and the reaction tube; An exhaust fan that supplies the cooling medium to the cooling unit; and A cooling control unit acquires a prediction model that infers a predicted temperature for at least one of the predicted temperatures of the heating element and the processing chamber. This prediction model includes information about the exhaust fan, a final target temperature that will become the future target, and the opening degree of the cooling valve. The cooling control unit acquires information about at least one of the temperatures of the heating element and the processing chamber, the opening degree of the cooling valve, and the exhaust fan, and adjusts the opening degree of the cooling valve in a manner that minimizes the error between the predicted temperature series calculated according to the prediction model and the target temperature series calculated using the ratio of the change from the current target temperature to the final target temperature.

2. The substrate processing apparatus according to claim 1, wherein, The cooling control unit is configured to include a temperature history storage unit that stores at least one of the temperatures of the heating unit and the processing chamber, an exhaust history storage unit that stores the on / off signals of the exhaust fan, and a valve opening history storage unit that stores the opening information output to the cooling valve. The temperature history storage unit, the exhaust history storage unit, and the valve opening history storage unit store each data for a specified period.

3. The substrate processing apparatus according to claim 1, wherein, The cooling control unit also includes a creation unit, which acquires the prediction model and acquires past temperature data of at least one of the temperatures of the heating unit and the processing chamber, past on / off data of the exhaust fan, and past opening data of the cooling valve, and calculates individual input response characteristic matrices and individual zero response characteristic vectors.

4. The substrate processing apparatus according to claim 1, wherein, The cooling unit is divided into multiple cooling zones. The prediction model is a mathematical formula for calculating the predicted temperature, expressed by the following Equation 1. [Mathematical Expression 1] Here, Let y(t-1) be the predicted temperature at time t, and y(t-1) be the temperature one step ago. V a (t-1), V a (t-2), ..., V a (tn) represents the opening degree of a cooling zone one time before, two times before, ..., n times before. V b (t-1), V b (t-2), ..., V b (tn) represents the opening degree of the zone adjacent to one side of the cooling zone before the first, second, ..., nth time iterations. V c (t-1), V c (t-2), ..., V c (tn) represents the opening degree of the zone adjacent to the other side of the cooling zone before the first, second, ..., nth time iterations. f(t-1), f(t-2), ..., f(tm) are the switching data of the exhaust fan before one, two, ..., m times. y0 is the reference temperature, n and m are arbitrarily set values, and a1, ..., a n b1, ..., b n c1, ..., c n d and d are the specified coefficients.

5. The substrate processing apparatus according to claim 4, wherein, The reference temperature y0 is a temperature within the range of 20℃ to 30℃. The values ​​of n and m represent the necessary amount of past data.

6. The substrate processing apparatus according to claim 3, wherein, The individual zero-response characteristic vector represents the amount in the predicted temperature vector that has changed due to the influence of past temperatures and past opening degrees, and the individual input response characteristic matrix represents the amount in the predicted temperature vector that has changed due to the influence of the opening degree calculated this time.

7. The substrate processing apparatus according to claim 3, wherein, The cooling control unit also has the capability to calculate an individual target temperature column vector S, expressed by the following mathematical formula 2. tg The target temperature column creation section, The target temperature column creation unit is configured to calculate the individual target temperature column vector S based on the target temperature, the current target temperature, and the ratio of the change from the current target temperature to the final target temperature. tg , [Mathematical Expression 2] 。 8. The substrate processing apparatus according to claim 7, wherein, The target temperature column creation unit calculates the temperature rise / fall deviation between the target temperature and the current target temperature, and divides the absolute value of the temperature rise / fall deviation by the rate of change. When the rate of change is zero, the reference setting value is calculated using the following formula: Reference setpoint = Current target temperature + Temperature rise / fall deviation × (1 - exp(Elapsed time ÷ (Heating / fall time ÷ Time constant)) When the rate of change is not zero, the reference setting value is calculated using the following formula: Reference setpoint = Current target temperature + Temperature rise / fall deviation × (1 - exp(Elapsed time ÷ (Heating / fall time)) The individual target temperature column vector S is calculated based on the aforementioned benchmark setting value. tg .

9. The substrate processing apparatus according to claim 7, wherein, The cooling control unit also includes a comprehensive characteristic creation unit, which creates a prescribed equation based on the individual input response characteristic matrix, the individual zero response characteristic vector, and the individual target temperature column vector. The cooling unit is divided into multiple cooling zones. The integrated characteristic creation unit is configured to arrange the individual zero-response characteristic vector, the individual input response characteristic matrix, and the individual target temperature column vector in all cooling zones that are controlled objects, and to create a predicted temperature column including the integrated input response characteristic matrix and the integrated zero-response characteristic vector, and a target temperature column including the integrated target temperature vector.

10. The substrate processing apparatus according to claim 9, wherein, The cooling control unit is further configured to include a calculation unit that creates an evaluation function representing the square of the error between the target temperature series and the predicted temperature series, and calculates a specified set of simultaneous equations in a manner that minimizes the evaluation function. The calculation unit obtains the opening degree of the cooling valve by solving the prescribed simultaneous equations, which are included in the solution of the predicted temperature series.

11. The substrate processing apparatus according to claim 10, wherein, The cooling control unit is configured to include an opening signal supply unit that updates the opening degree of the cooling valve obtained from the calculation unit within a predetermined control cycle.

12. The substrate processing apparatus according to claim 1, wherein, The heater unit is divided into multiple control zones and equipped with temperature sensors to detect the temperature of each control zone. The cooling unit is divided into multiple cooling zones, and each cooling zone is equipped with a cooling valve.

13. The substrate processing apparatus according to claim 12, wherein, The prediction model that predicts the predicted temperature of at least one of the temperatures of the heating section and the processing chamber in each cooling zone corresponds to their respective temperature ranges.

14. A program product comprising a temperature control program executed in a substrate processing apparatus, the substrate processing apparatus comprising: The reaction tube forms the processing chamber for processing the substrate inside; A heater unit is disposed outside the reaction tube and has a heating section for heating the substrate; A cooling unit having a cooling valve that supplies cooling medium to the space between the heater unit and the reaction tube; and An exhaust fan that supplies the cooling medium to the cooling unit. The temperature control program causes the substrate processing apparatus to perform the following steps: A prediction model is obtained to infer the predicted temperature of at least one of the predicted temperatures of the heating element and the processing chamber, the prediction model including information of the exhaust fan, the final target temperature to be the future target, and the opening degree of the cooling valve. Acquire at least one of the temperatures of the heating element and the processing chamber, the opening degree of the cooling valve, and information about the exhaust fan; as well as The opening of the cooling valve is adjusted in such a way that the error between the predicted temperature series calculated according to the prediction model and the target temperature series calculated using the ratio of the change from the current target temperature to the final target temperature is minimized.

15. A method for manufacturing a semiconductor device, comprising a step of raising the temperature of a processing chamber for processing a substrate from a predetermined temperature to a processing temperature, a step of processing the substrate while maintaining the processing temperature, and a step of cooling the temperature of the processing chamber from the processing temperature after the processing step. The process of cooling the temperature of the processing chamber includes the following steps: Acquire at least one of the temperatures of the heating element and the processing chamber, the opening degree of the cooling valve, and information about the exhaust fan; and The opening of the cooling valve is adjusted in such a manner that the error between a predicted temperature series calculated by a prediction model that infers at least one of the predicted temperatures of the heating element and the processing chamber and a target temperature series calculated using the ratio of the change in the current target temperature to the final target temperature that will become the future target is minimized. The prediction model includes information about the exhaust fan, the final target temperature, and the opening of the cooling valve.

16. The method for manufacturing a semiconductor device according to claim 15, wherein, In the process of lowering the temperature of the processing chamber, the power supply value output from the heating unit is set to zero.

17. A temperature control method, comprising the following steps: Acquire at least one of the following: the temperature of the heating element and the temperature of the processing chamber; the opening degree of the cooling valve; and information about the exhaust fan. The opening of the cooling valve is adjusted in such a manner that the error between a predicted temperature series calculated by a prediction model that infers at least one of the predicted temperatures of the heating element and the processing chamber and a target temperature series calculated using the ratio of the change in the current target temperature to the final target temperature that will become the future target is minimized. The prediction model includes information about the exhaust fan, the final target temperature, and the opening of the cooling valve.

Images