Substrate processing apparatus, substrate processing apparatus control method, and control program

Abstract

Provided are a substrate processing apparatus, a substrate processing apparatus control method, and a control program.　The substrate processing apparatus performs substrate processing through cooperative control performed by a plurality of control systems. The substrate processing apparatus comprises an estimator that estimates a disturbance in a first control system, and a controller that controls a control object in a second control system different from the first control system by reflecting a result of estimation by the estimator.

Description

SUBSTRATE PROCESSING APPARATUS, CONTROL METHOD FOR SUBSTRATE PROCESSING APPARATUS, AND CONTROL PROGRAM

[0001] The present disclosure relates to a substrate processing apparatus, a control method for a substrate processing apparatus, and a control program.

[0002] In substrate processing equipment such as semiconductor manufacturing equipment, control is sometimes performed for each functional unit using estimated quantities obtained using soft sensing technology rather than physical sensors. Here, a functional unit refers to a system for controlling one or a small number of specific physical quantities, such as a pressure control system, a temperature control system, an RF (Radio Frequency) control system, or a gas flow rate control system.

[0003] International Publication No. 2021 / 014724

[0004] The present disclosure provides a substrate processing apparatus, a control method for the substrate processing apparatus, and a control program that can improve performance in cooperative control by a plurality of control systems.

[0005] The substrate processing apparatus of the present disclosure is a substrate processing apparatus that performs substrate processing through cooperative control by multiple control systems, and is equipped with an estimator that estimates a disturbance in a first control system, and a controller that controls a control object in a second control system different from the first control system by reflecting the estimation result by the estimator.

[0006] According to the present disclosure, it is possible to improve performance in cooperative control by a plurality of control systems.

[0007] FIG. 1 is a schematic diagram showing an example of the configuration of a substrate processing apparatus according to an embodiment. FIG. 2 is a block diagram showing the internal configuration of a control device. FIG. 3 is an explanatory diagram explaining the configuration of a control system in a comparative example. FIG. 4 is an explanatory diagram explaining an example of the configuration of a control system in embodiment 1. FIG. 5 is an explanatory diagram explaining a first application example in a substrate processing apparatus. FIG. 6 is an explanatory diagram explaining a second application example in a substrate processing apparatus. FIG. 7 is an explanatory diagram explaining a third application example in a substrate processing apparatus. FIG. 8 is a flowchart explaining the procedure of processing executed by a control unit. FIG. 9 is a schematic diagram showing an example of displaying disturbances. FIG. 10 is an explanatory diagram explaining a method of generating a learning model. FIG. 11 is an explanatory diagram explaining a method of using a learning model.

[0008] An embodiment will be described below with reference to the drawings. In the description, the same elements or elements having the same functions are designated by the same reference numerals, and redundant description will be omitted.

[0009] (Embodiment 1) Figure 1 is a schematic diagram showing an example of the configuration of a substrate processing apparatus according to an embodiment. The substrate processing apparatus 1 according to the embodiment includes an apparatus main body 10 in which substrate processing is performed, and a control device 50 that executes control related to the substrate processing. The substrate processing apparatus 1 is an apparatus for performing processes such as etching, film formation, CMP (Chemical Mechanical Polishing), ashing, and cleaning on a substrate (wafer W) that is a processing target. The control device 50 controls the operation of each hardware component included in the apparatus main body 10, acquires various data obtained during processing, and monitors the status of processing in the apparatus main body 10.

[0010] The apparatus main body 10 has a cylindrical chamber 11 whose interior can be sealed. The chamber 11 is made of, for example, aluminum and is connected to a ground potential. A mounting table 12 made of a conductive material such as aluminum is provided inside the chamber 11. The mounting table 12 is a cylindrical table on which a substrate (wafer W) to be processed is placed. The mounting table 12 is configured to also function as a lower electrode.

[0011] An exhaust path 13 is formed between the side wall of the chamber 11 and the side surface of the mounting table 12, and serves as a path for discharging gas above the mounting table 12 to the outside of the chamber 11. An exhaust plate 14 is disposed midway along the exhaust path 13. The exhaust plate 14 is a plate-like member having a large number of holes, and functions as a partition plate that divides the space within the chamber 11 into an upper space and a lower space.

[0012] The upper space of the chamber 11 is a reaction chamber 17 where plasma etching is performed, and the lower space is an exhaust chamber (manifold) 18. An exhaust pipe 15 that exhausts gas from the chamber 11 is connected to the exhaust chamber 18. An exhaust plate 14 captures or reflects plasma generated in the reaction chamber 17 to prevent leakage into the exhaust chamber 18. The exhaust pipe 15 is connected to an exhaust device via an APC (Adaptive Pressure Control) valve 16. The exhaust device reduces the pressure inside the chamber 11 and maintains it at a desired vacuum state.

[0013] A first high frequency power supply 19 is connected to the mounting table 12 via a matching box 20. The first high frequency power supply 19 supplies high frequency bias power of, for example, 400 kHz to 13.56 MHz to the mounting table 12. The matching box 20 suppresses reflection of the high frequency power from the mounting table 12, maximizing the efficiency of supplying the high frequency bias power to the mounting table 12.

[0014] An electrostatic chuck (ESC) 22 having an electrostatic electrode plate 21 therein is disposed on the upper surface of the mounting table 12. The electrostatic chuck 22 has a shape in which an upper disk-shaped member having a smaller diameter than the lower disk-shaped member is stacked on top of a lower disk-shaped member. The electrostatic chuck 22 is made of aluminum, and its upper surface is thermally sprayed with a ceramic or the like. When a wafer W is placed on the mounting table 12, the wafer W is placed on the upper disk-shaped member of the electrostatic chuck 22. A first DC power supply 23 is connected to the electrostatic electrode plate 21. The electrostatic chuck 22 attracts and holds the wafer W by generating an electrostatic force such as Coulomb force due to a voltage applied to the electrostatic electrode plate 21 from the first DC power supply 23.

[0015] An annular edge ring 24 is placed on the electrostatic chuck 22 so as to surround the peripheral edge of the wafer W. The edge ring 24 is made of a conductive material (e.g., silicon) and focuses plasma in the reaction chamber 17 toward the surface of the wafer W, thereby improving the efficiency of the etching process.

[0016] An electromagnet may be provided above the chamber 11. The control device 50 can control the sheath thickness by controlling the characteristics of the magnetic field inside the chamber 11 using the electromagnet provided above the chamber 11. The control device 50 controls the strength of the magnetic field generated by the electromagnet in accordance with the phase of the high-frequency power, the potential of the electrode to which the high-frequency power is applied, the reflected wave power, etc., thereby controlling the characteristics of the magnetic field inside the chamber 11.

[0017] The mounting table 12 includes an annular coolant chamber 25 extending in the circumferential direction, for example. A low-temperature coolant is circulated and supplied from a chiller unit to the coolant chamber 25 through coolant piping 26. The low-temperature coolant may be cooling water or Galden (registered trademark). The mounting table 12, cooled by the low-temperature coolant, cools the wafer W and the edge ring 24 via the electrostatic chuck 22.

[0018] The electrostatic chuck 22 has a plurality of heat transfer gas supply holes 27. A heat transfer gas such as helium (He) gas is supplied to the plurality of heat transfer gas supply holes 27 via a heat transfer gas supply line 28. The heat transfer gas is supplied to a gap between the attracting surface of the electrostatic chuck 22 and the back surface of the wafer W via the heat transfer gas supply holes 27. The heat transfer gas supplied to the gap functions to transfer heat from the wafer W to the electrostatic chuck 22.

[0019] A heater may be provided inside the electrostatic chuck 22, and a temperature control module may be connected to the heater. The temperature control module adjusts the heater power output to the heater in response to a temperature command from the control device 50, thereby adjusting at least one of the electrostatic chuck 22 and the wafer W to a target temperature.

[0020] A shower head 29 is provided on the ceiling of the chamber 11, facing the mounting table 12. The shower head 29 includes an upper electrode 33 having a large number of gas holes 32, a cooling plate 34 to which the upper electrode 33 is detachably attached, and a lid 35 that covers the cooling plate 34. A buffer chamber 36 is provided inside the cooling plate 34. A gas inlet pipe 37 is connected to the buffer chamber 36. The shower head 29 diffuses gas introduced through the gas inlet pipe 37 in the buffer chamber 36 and supplies the gas into the reaction chamber 17 via the large number of gas holes 32.

[0021] A second high frequency power supply 31 is connected to the upper electrode 33 via a matching box 30. The second high frequency power supply 31 supplies high frequency power for plasma excitation, for example, of about 40 MHz, to the upper electrode 33. The matching box 30 suppresses reflection of the high frequency power from the upper electrode 33, maximizing the efficiency with which the high frequency power for plasma excitation is supplied to the upper electrode 33. In this embodiment, the high frequency power for plasma excitation is applied to the upper electrode 33, but it may also be applied to the mounting table 12.

[0022] The cooling plate 34 has a cooling mechanism and cools the upper electrode 33. The cooling mechanism has a spiral or annular coolant chamber 38 extending in the circumferential direction and coolant piping 38a. A low-temperature coolant is circulated and supplied from a chiller unit to the coolant chamber 38 via the coolant piping 38a. The low-temperature coolant is cooling water, Galden (registered trademark), or the like. The upper electrode 33 becomes hot due to heat input from the plasma. In this embodiment, the upper electrode 33 and the cooling plate 34 are closely attached to each other, and heat from the upper electrode 33 is dissipated by the cooling plate 34, thereby dissipating heat from the upper electrode 33 and cooling the upper electrode 33.

[0023] In the embodiment, the substrate processing apparatus 1 may be a semiconductor device such as a film formation apparatus, an etching apparatus, an ion implantation apparatus, an ashing apparatus, a sputtering apparatus, or an exposure apparatus. The substrate processing apparatus 1 may also be a semiconductor manufacturing system including one or more semiconductor manufacturing apparatuses and a transfer apparatus for transferring wafers W. Furthermore, the substrate processing apparatus 1 may also be a display manufacturing apparatus for manufacturing flat display panels (FDPs) such as liquid crystal display panels and organic electroluminescence (EL) panels.

[0024] 2 is a block diagram showing the internal configuration of the control device 50. The control device 50 is, for example, a dedicated or general-purpose computer including a control unit 51, a storage unit 52, a communication unit 53, an operation unit 54, and a display unit 55.

[0025] The control unit 51 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), etc. The ROM included in the control unit 51 stores control programs and the like that control the operation of each hardware unit included in the control device 50 and the operation of each hardware unit included in the device main body 10. The CPU in the control unit 51 reads and executes the control programs stored in the ROM and the computer programs described below that are stored in the storage unit 52, and controls the operation of each hardware unit, thereby causing the entire device to function as the control device 50 of the present disclosure. The RAM included in the control unit 51 temporarily stores data used during execution of calculations.

[0026] In the embodiment, the control unit 51 is configured to include a CPU, a ROM, and a RAM, but the configuration of the control unit 51 is not limited to the above. The control unit 51 may be, for example, one or more control circuits or processing circuits including a GPU (Graphics Processing Unit), an FPGA (Field Programmable Gate Array), a DSP (Digital Signal Processor), a quantum processor, volatile or non-volatile memory, etc. The control unit 51 may also have functions such as a clock that outputs date and time information, a timer that measures the elapsed time from when a measurement start instruction is given until when a measurement end instruction is given, and a counter that counts numbers.

[0027] The storage unit 52 includes a storage device such as a hard disk drive (HDD), a solid state drive (SSD), an electronically erasable programmable read-only memory (EEPROM), etc. The storage unit 52 stores various computer programs executed by the control unit 51 and various data used by the control unit 51.

[0028] The computer program (program product) stored in the storage unit 52 includes a control program PG for causing a computer to execute a process of estimating a disturbance in a first control system for a plurality of control systems in the substrate processing apparatus 1 and controlling a control target in a second control system different from the first control system by reflecting the estimated disturbance result. Specific examples of the control systems and control methods will be described in detail later.

[0029] The control program PG may be a single computer program or a group of programs consisting of multiple computer programs. The control program PG may be executed by multiple computers working together. The control program PG may partially use an existing library.

[0030] A computer program including the control program PG is provided, for example, by a non-transitory recording medium RM on which the computer program is readably recorded. The recording medium RM is a portable memory such as a CD-ROM, USB memory, a Secure Digital (SD) card, a microSD card, or a CompactFlash (registered trademark). The control unit 51 reads various computer programs from the recording medium RM using a reading device (not shown) and stores the read computer programs in the memory unit 52. The computer programs stored in the memory unit 52 may also be provided via communication. In this case, the control unit 51 acquires the computer programs via communication via the communication unit 53 and stores the acquired computer programs in the memory unit 52.

[0031] The communication unit 53 includes a communication interface for transmitting and receiving various data to and from an external device. A wired or wireless communication interface conforming to a communication standard such as a local area network (LAN) can be used as the communication interface of the communication unit 53. The external device may be the device main body 10 or a user terminal (not shown). When data to be transmitted is input from the control unit 51, the communication unit 53 transmits the data to the destination external device, and when data transmitted from the external device is received, the communication unit 53 outputs the received data to the control unit 51.

[0032] The operation unit 54 includes operation devices such as a touch panel, a keyboard, and switches, and receives various inputs and operations from a user, etc. The control unit 51 acquires information input through the operation unit 54 and performs appropriate control based on various operation information provided by the operation unit 54.

[0033] The display unit 55 includes a display device such as a liquid crystal monitor or an organic EL (Electro-Luminescence) monitor, and displays information to be notified to the user or the like in response to an instruction from the control unit 51 .

[0034] The control device 50 may be a single computer or a computer system configured with multiple computers and peripheral devices. The control device 50 may also be a virtual machine whose entity is virtualized, or may be a cloud. Furthermore, the control device 50 may be provided separately from the device main body 10, or may be provided inside the device main body 10 as an element of the device main body 10.

[0035] 3 is an explanatory diagram illustrating the configuration of a control system in a comparative example. The control system in the comparative example includes two control systems: a first control system S1 and a second control system S2. The first control system S1 includes a controller C1 and a controlled object O1, and the second control system S2 includes a controller C2 and a controlled object O2.

[0036] The controller C1 of the first control system S1 generates a control input u1 for the control object O1 based on a control command given from the outside and outputs it to the control object O1. The control object O1 is controlled by the input of the control input u1, and its state changes. The state change of the control object O1 (or the environmental change accompanying the state change of the control object O1) is measured by a physical sensor. The sensor output y1 of the physical sensor is fed back to the controller C1. The controller C1 performs feedback control based on the sensor output y1.

[0037] Similarly, the controller C2 of the second control system S2 generates a control input u2 for the control object O2 based on a control command given from the outside and outputs it to the control object O2. The control object O2 is controlled by the control input u2, and its state changes. The state change of the control object O2 (or the environmental change accompanying the state change of the control object O2) is measured by a physical sensor. The sensor output y2 of the physical sensor is fed back to the controller C2. The controller C2 performs feedback control based on the sensor output y2.

[0038] It is assumed that multiple control systems are intricately intertwined and interfere with each other in the substrate processing apparatus 1. For example, it is assumed that a phenomenon occurring in the control object O2 of the second control system S2 becomes a disturbance to the control object O1 of the first control system S1, and affects the state change (sensor output y1) of the control object O1.

[0039] Even under such circumstances, if the phenomenon occurring in the controlled object O2 of the second control system S2 can be observed and the controller C2 can be feedback controlled based on the physical quantity yα obtained as the observation result, it is possible to improve the performance of the controller C2.

[0040] However, in reality, there may be cases where a physical sensor capable of observing the physical quantity yα does not exist, and it may be difficult to create a model of the controlled object O2 that includes the dynamics of the physical quantity yα. In such cases, the performance of the controller C2 cannot be improved.

[0041] Therefore, in this embodiment, we propose constructing an estimator E1 (see Figure 4) that estimates the physical quantity yα as a disturbance in the control object O1, and reflecting the physical quantity yα estimated by the estimator E1 in the control by the controller C2.

[0042] FIG. 4 is an explanatory diagram illustrating an example configuration of a control system according to the first embodiment. The control system according to the first embodiment includes two control systems: a first control system S1 and a second control system S2. The first control system S1 includes an estimator E1 in addition to the controller C1 and the controlled object O1 described above. The second control system S2 includes a controller C2 and the controlled object O2. In the example of FIG. 4, for simplicity, the first control system S1 and the second control system S2 are each provided as one system, but at least one of the first control system S1 and the second control system S2 may be provided in plurality. Furthermore, the controllers C1 and C2 may each be dedicated hardware, or may be realized as a software function executed by the control unit 51.

[0043] 3 , the controller C1 of the first control system S1 generates a control input u1 in response to a control command from the outside (e.g., the control device 50) and outputs the generated control input u1 to the control object O1, thereby controlling the control object O1. A change in the state of the control object O1 (or an environmental change accompanying the change in the state of the control object O1) is measured by a physical sensor and obtained as a physical quantity y1. The controller C1 acquires the physical quantity y1 obtained from the physical sensor and performs feedback control of the control object O1.

[0044] Similarly, the controller C2 of the second control system S2 generates a control input u2 in response to a control command from the outside (for example, the control device 50) and outputs the generated control input u2 to the control object O2, thereby controlling the control object O2. A change in the state of the control object O2 (or an environmental change accompanying the change in the state of the control object O2) is measured by a physical sensor and obtained as a physical quantity y2.

[0045] Here, it is assumed that a physical quantity yα that cannot be observed by a physical sensor is input as a disturbance to the control target O1. A specific example will be described in detail later, but for example, in a temperature control system that controls the temperature of the electrostatic chuck 22, heat input from plasma is a physical quantity that cannot be measured by a physical sensor and can become a disturbance in the temperature control system.

[0046] In this embodiment, an estimator E1 is provided to estimate a physical quantity yα that cannot be observed by a physical sensor. The estimator E1 is a software sensor that uses a control input u1 to the control object O1 and a physical quantity y1 that is the output of the control object O1 as inputs to estimate a disturbance to the control object O1. The estimator E1 is constructed using an observer or a Kalman filter that models the control object O1. An existing method is used to construct the model. In general, the model can be constructed by utilizing information on the physical quantity y1 (actual value) observed for the control object O1 when the control input u1 is input.

[0047] The control input u1 and the physical quantity y1 are time-series data. The estimator E1 receives the control input u1 and the physical quantity y1, which are time-series data, and estimates the disturbance (estimated quantity yα_hat) in time series. The estimator E1 outputs the estimated estimated quantity yα_hat to the controller C2 of the second control system S2.

[0048] When a control command is input from outside, the controller C2 controls the controlled object O2 using a control law that reflects the estimation result by the estimator E1. The control law is set in advance to include the control command and an estimated quantity yα_hat estimated by the estimator E1 as parameters. Furthermore, when a physical quantity y2 measured by a physical sensor is used for feedback control, the control law may be set to further include the physical quantity y2 as a parameter. The controller C2 generates a control input u2 using the control law and outputs the generated control input u2 to the controlled object O2, thereby controlling the controlled object O2.

[0049] In embodiment 1, the estimated quantity yα_hat related to the phenomenon (state described by the physical quantity yα) that occurs when the controlled object O2 is controlled is reflected in the control law, and the controlled object O2 can be controlled, thereby improving the performance of the controller C2.

[0050] Specific application examples will be described below. Fig. 5 is an explanatory diagram illustrating a first application example in the substrate processing apparatus 1. The control system of the substrate processing apparatus 1 includes, for example, an ESC temperature control system 110, an RF control system 120, and a magnetic field control system 130. These control systems are a series of control systems in which the control device 50 performs cooperative control during plasma processing.

[0051] The ESC temperature control system 110 is a control system for controlling the temperature of the electrostatic chuck 22. The ESC temperature control system 110 includes a controller and a controlled object. The controller in the ESC temperature control system 110 is a temperature adjustment module 111, and the controlled object is the ESC 112 (electrostatic chuck 22). The temperature adjustment module 111 adjusts heater power based on a control command given by the control device 50, and controls the amount of heat generated by a heater provided inside the ESC 112. The temperature of the ESC 112 increases due to the heat generated by the heater. The temperature y1 of the ESC 112 is measured over time by a temperature sensor (physical sensor), and the measurement results are fed back to the temperature adjustment module 111.

[0052] The ESC temperature control system 110 may further include a controller (chiller unit) for cooling the ESC 112, a controlled object (refrigerant), and the like.

[0053] The RF control system 120 is a control system for controlling the plasma in the chamber 11. The RF control system 120 includes a controller and a controlled object. The controller in the RF control system 120 is a high-frequency power supply 121 (first high-frequency power supply 19 and second high-frequency power supply 31), and the controlled object is plasma 122. The high-frequency power supply 121 adjusts the frequency and other parameters of the high-frequency power applied to the mounting table 12 and the upper electrode 33 in accordance with a control command given by the control device 50, thereby controlling the plasma generated in the chamber 11. The RF control system 120 may measure a measurable physical quantity (e.g., a reflection coefficient) of the plasma generated in the chamber 11 using a physical sensor and feed it back to the high-frequency power supply 121.

[0054] The magnetic field control system 130 is a control system for controlling the magnetic field within the chamber 11. The magnetic field control system 130 includes a controller and a controlled object. The controller in the magnetic field control system 130 is a magnetic field regulator 131, and the controlled object is an electromagnet 132 provided above the chamber 11 to control the magnetic field within the chamber 11. The magnetic field regulator 131 controls the magnetic field within the chamber 11 by adjusting the magnitude and phase of the current output to the electromagnet 132 in accordance with a control command given by the control device 50. The magnetic field control system 130 can control the sheath thickness by controlling the magnetic field within the chamber 11. The magnetic field control system 130 may measure the magnetic field within the chamber 11 with a physical sensor and feed the result back to the magnetic field regulator 131.

[0055] When plasma is generated in the chamber 11 by the control of the RF control system 120 and the magnetic field control system 130, heat is input from the plasma to the ESC 112. The heat input from the plasma becomes a disturbance in the ESC temperature control system 110. Therefore, in this embodiment, a model (ESC model 113) for estimating the heat input from the plasma is prepared, and the heat input to the ESC 112 is estimated by the ESC model 113. The ESC model 113 may estimate a physical quantity related to the heat input, such as a heat flux, instead of the heat input.

[0056] The ESC model 113 is an estimator (soft sensor) using an observer or a Kalman filter, and is described using parameters such as the density, specific heat, heat flux cross-sectional area, thickness, temperature, heat input, and heat extraction of the ESC 112. The observer may be a state observer, a disturbance observer, or the like. The Kalman filter may be an extended Kalman filter, an unscented Kalman filter, or the like. Alternatively, the ESC model 113 may be a particle filter or an estimator (soft sensor) using the recursive least squares method, or any suitable method may be selected as long as it can perform online sequential estimation. The ESC model 113 is constructed to estimate the heat input from the plasma to the ESC 112 in a time series manner when data such as the temperature y1 of the ESC 112 measured in a time series by a temperature sensor is input. The ESC model 113 outputs the estimated heat input value yα_hat to the RF control system 120 and the magnetic field control system 130.

[0057] In the RF control system 120 and the magnetic field control system 130, the heat input value can be reflected in the control law as a parameter related to the plasma generated in the chamber 11 in each control, thereby improving the performance of the controller.

[0058] 6 is an explanatory diagram illustrating a second application example in the substrate processing apparatus 1. The control system of the substrate processing apparatus 1 includes, for example, an ESC temperature control system 110, a gas flow rate control system 140, and a pressure control system 150. These control systems are a series of control systems that are cooperatively controlled by a control device 50 during plasma processing.

[0059] 5, and is a control system for controlling the temperature of the ESC 112. The ESC temperature control system 110 includes a controller and a controlled object.

[0060] The gas flow control system 140 is a control system for controlling the flow rate of gas supplied to the chamber 11. The gas flow control system 140 includes a controller and a controlled object. The gas flow control system 140 includes a controller and a controlled object. The controller in the gas flow control system 140 is a flow controller 141, and the controlled object is, for example, a valve 142 provided in the gas inlet pipe 37. The flow controller 141 adjusts the opening of the valve 142 in response to a control command given by the control device 50, thereby controlling the flow rate of gas supplied into the chamber 11. The gas flow control system 140 may measure the gas flow rate with a physical sensor and feed the measured value back to the flow controller 141.

[0061] The pressure control system 150 is a control system for controlling the pressure inside the chamber 11. The pressure control system 150 includes a controller and a controlled object. The pressure control system 150 includes a controller and a controlled object. The controller in the pressure control system 150 is a pressure controller 151, and the controlled object is, for example, a valve 152 (APC valve 16) provided in the exhaust pipe 15. The pressure controller 151 adjusts the opening of the valve 152 in accordance with a control command given by the control device 50, thereby controlling the pressure inside the chamber 11. The pressure control system 150 may measure the pressure inside the chamber 11 with a physical sensor and feed the measured pressure back to the pressure controller 151.

[0062] When plasma is generated by controlling the flow rate and pressure of the gas supplied into the chamber 11 using the gas flow rate control system 140 and the pressure control system 150, heat is input from the plasma to the ESC 112. The heat input from the plasma becomes a disturbance in the ESC temperature control system 110. Therefore, a model (ESC model 113) for estimating the heat input from the plasma to the ESC 112 is constructed, and the heat input is estimated using the constructed ESC model 113. The ESC model 113 may estimate a physical quantity related to the heat input, such as a heat flux, instead of the heat input.

[0063] The ESC model 113 is an estimator (software sensor) using an observer or a Kalman filter, and is described using parameters such as the density, specific heat, heat flux cross-sectional area, thickness, temperature, heat input, and heat extraction of the ESC 112. The ESC model 113 is constructed so that, when data such as the temperature y1 of the ESC 112 measured over time by a temperature sensor is input, the ESC model 113 outputs the heat input from the plasma to the ESC 112 over time. The ESC model 113 outputs the estimated heat input value yα_hat to the gas flow control system 140 and the pressure control system 150.

[0064] In the gas flow rate control system 140 and the pressure control system 150, the heat input value can be reflected in the control law as a parameter related to the plasma generated in the chamber 11 in each control, thereby improving the performance of the controller.

[0065] 7 is an explanatory diagram illustrating a third application example in the substrate processing apparatus 1. The control system of the substrate processing apparatus 1 includes, for example, an ESC temperature control system 110, an RF control system 120, a magnetic field control system 130, and an in-wafer uniformity control system 100. These control systems are a series of control systems that are cooperatively controlled by the control device 50 during plasma processing.

[0066] The ESC temperature control system 110, RF control system 120, and magnetic field control system 130 are similar to those shown in Fig. 5 and each include a controller and a controlled object. In addition to the controller and controlled object, the ESC temperature control system 110 includes an ESC model 113 for estimating heat input to the ESC 112.

[0067] The in-plane uniformity control system 100 is a control system positioned above the ESC temperature control system 110, the RF control system 120, and the magnetic field control system 130. The in-plane uniformity control system 100 controls the operation of each of the control systems 110 to 130, thereby controlling the in-plane uniformity of the substrate (wafer W) processed by the substrate processing apparatus 1.

[0068] The in-plane uniformity control system 100 includes a controller 101. The controller 101 receives a control command from the outside (e.g., the control device 50) and an estimation result (estimated quantity yα_hat) estimated by the ESC model 113 of the ESC temperature control system 110. When a control command from the outside is input, the controller 101 generates control commands (temperature command, reflection command, magnetic field command) for the ESC temperature control system 110, the RF control system 120, and the magnetic field control system 130, respectively, using a control law that reflects the estimation result by the ESC model 113. The control law used by the controller 101 is provided for each of the control systems 110 to 130, and is set in advance to include the external control command and the estimated quantity yα_hat estimated by the ESC model 113 as parameters.

[0069] The controllers of the control systems 110 to 130 generate control inputs for the respective control objects in response to control commands from the in-plane uniformity control system 100, and control the respective control objects by outputting the control inputs to the respective control objects. Each controller may perform feedback control using a physical quantity measured by a physical sensor in each of the control systems 110 to 130.

[0070] As described above, in embodiment 1, even if there is a physical quantity that cannot be measured by a physical sensor, it is estimated as a disturbance in the first control system, and the estimation result is reflected in controlling the second control system, which is different from the first control system, thereby improving the performance of the controller in the second control system.

[0071] Second Embodiment In a second embodiment, a configuration will be described in which the function of the estimator E1 is realized by a control unit 51 of a control device 50. The substrate processing apparatus 1 includes the first control system S1 and the second control system S2 described in the first embodiment.

[0072] 8 is a flowchart illustrating the procedure of the process executed by the control unit 51. The control unit 51 sequentially executes the following process, for example, at predetermined time steps. The control unit 51 outputs a control command to the first control system S1 and controls the first control system S1 (step S101). In the first control system S1, the controller C1 receives the control command from the control unit 51 and generates a control input u1, and the control object O1 is controlled based on the control input u1. The state of the control object O1 changes as a result of the control of the control object O1, and the state of the control object O1 is measured by a physical sensor provided in the first control system S1.

[0073] The control unit 51 acquires the control input u1 generated by the controller C1 and the physical quantity y1 measured by the physical sensor from the first control system S1 via the communication unit 53 (step S102).

[0074] The control unit 51 estimates a disturbance to the control object O1 using the acquired control input u1 and physical quantity y1 (step S103). A model of the control object O1 is used to estimate the disturbance. The model of the control object O1 is constructed using an observer or a Kalman filter. An existing method is used to construct the model. In general, the model can be constructed by utilizing information on the physical quantity y1 (actual value) observed for the control object O1 when the control input u1 is input.

[0075] The control unit 51 controls the second control system S2 using a control law that reflects the estimated result of the disturbance (step S104). The control law is set in advance to include the estimated result of the disturbance (estimated amount yα_hat) as a parameter. The control unit 51 generates a control command for the second control system S2 using the control law that reflects the estimated result of the disturbance, and controls the second control system S2 by outputting the generated control command to the second control system S2. In the second control system S2, a controller C2 generates a control input u2 in response to the control command from the control unit 51, and the controlled object O2 is controlled based on the control input u2.

[0076] Because the state of the control object O2 changes when the control object O2 is controlled, the state of the control object O2 may be measured by a physical sensor provided in the second control system S2. The second control system S2 may also output the measurement results measured by the physical sensor to the control unit 51. When controlling the second control system S2 in step S104, the control unit 51 may perform feedback control taking into account the measurement results measured by the physical sensor.

[0077] In the present embodiment, a physical quantity that is not observed by a physical sensor (for example, heat input from plasma to the electrostatic chuck 22) is estimated in a time series manner by the control unit 51. The control unit 51 may cause the display unit 55 to display the estimated physical quantity (disturbance).

[0078] FIG. 9 is a schematic diagram showing an example of a display of a disturbance. FIG. 9 shows an example in which the heat input to the electrostatic chuck 22 is displayed on the display unit 55 as a distribution within the substrate surface. A color contour map, for example, is used to display the heat input distribution. Since the control unit 51 estimates the heat input value in a time series manner, the heat input distribution within the substrate surface may be displayed as a video. In the example of FIG. 9, a video playback button BT is provided on the display screen. When the video playback button BT is pressed by the operation unit 54, the control unit 51 plays a video relating to the heat input distribution within the substrate surface.

[0079] As described above, in the second embodiment, the function of the estimator E1 can be realized in the control unit 51. Furthermore, in the second embodiment, it is possible to visualize time-series data of disturbances that cannot be measured by physical sensors.

[0080] (Embodiment 3) In embodiment 3, a configuration will be described in which a virtual experiment is performed using a learning model that has learned the relationship between control commands, etc. and disturbances, and the behavior of the device is optimized by repeating the virtual experiment.

[0081] FIG. 10 is an explanatory diagram illustrating a method for generating a learning model. In the third embodiment, various data corresponding to the various conditions θ are obtained by operating the substrate processing apparatus 1 under various conditions θ. The conditions θ may be preset values ​​such as a process recipe, or may be parameters representing the state of the substrate processing apparatus 1 during operation. The conditions θ are expressed as a vector including, for example, the type of gas introduced into the chamber 11, the gas pressure in the chamber 11, the voltage applied to the plasma, their time-series changes, the continuous operating time of the substrate processing apparatus 1, the time elapsed since the last maintenance, the number of operations since the last maintenance, and machine-difference features of the substrate processing apparatus 1. Alternatively, the conditions θ may be expressed as a vector obtained by encoding the above parameters using principal component analysis, an autoencoder, or the like.

[0082] The control system of the substrate processing apparatus 1 according to the third embodiment includes a first control system S1 and a second control system. A controller C1 in the first control system S1 receives a control command y from an external device (for example, a control device 50). ref1 A control input u1 is generated in accordance with the control input u1, and the controlled object O1 is controlled based on the generated control input u1. A change in the state of the controlled object O1 (or an environmental change accompanying the change in the state of the controlled object O1) is measured by a physical sensor. The sensor value (physical quantity) of the physical sensor is denoted as y1.

[0083] The controller C2 in the second control system S2 receives a control command y from an external device (for example, the control device 50). ref2 A control input u2 is generated in accordance with the control input u2, and the controlled object O2 is controlled based on the generated control input u2. A change in the state of the controlled object O2 (or an environmental change accompanying the change in the state of the controlled object O2) is measured by a physical sensor. The sensor value (physical quantity) of the physical sensor is denoted by y2.

[0084] Here, when a physical quantity d2 that cannot be observed by a physical sensor exists in the second control system S2 and the physical quantity d2 becomes a disturbance in the control object O1, the physical quantity d2 (disturbance) can be estimated by using an estimator E1 (an observer or Kalman filter that models the control object O1) as described in embodiment 1. Similarly, when a physical quantity d1 that cannot be observed by a physical sensor exists in the first control system S1 and the physical quantity d1 becomes a disturbance in the control object O2, the physical quantity d1 (disturbance) can be estimated if the control object O2 can be modeled. Hereinafter, the estimated amounts of disturbance will be denoted as d1_hat and d2_hat, respectively.

[0085] The control device 50 generates a data set (θ, y ref1 , y ref2 , y1, y2, d1_hat, d2_hat) are acquired and stored in the database DB. The database DB may be provided in the storage unit 52 of the control device 50 or may be provided outside the control device 50.

[0086] The control unit 51 of the control device 50 generates a learning model MD by learning with an existing algorithm such as machine learning using a data set stored in the database DB. The learning model MD is generated based on the condition θ, the control command y ref1 , y ref2 , and the estimated quantities y1_hat and y2_hat for the physical quantities y1 and y2 are input, the system is trained to output the estimated quantities d1_hat and d2_hat of the disturbances.

[0087] Here, the estimated quantities y1_hat and y2_hat may be values ​​(estimated quantities) obtained after the controllers C1 and C2 are adjusted by the disturbance feedback mechanism described in the first embodiment. Furthermore, when performing a simulation using the learning model MD, measurements from physical sensors cannot be obtained, so virtual estimated quantities y1_hat and y2_hat for the physical quantities y1 and y2 may be used. For example, the estimated quantities y1_hat and y2_hat may be virtual values ​​obtained when the control systems S1 and S2 are completely independent (when the disturbance is zero). Furthermore, the estimated quantities y1_hat and y2_hat may be virtual values ​​obtained when virtual disturbances d1 and d2 determined by a preset function or virtual disturbances d1(θ) and d2(θ) determined according to the condition θ are input to the controllers C1 and C2.

[0088] The learning model MD is a function of the condition θ and the control command y ref1 , y ref2 In this case, the data set acquired by the control device 50 during learning includes the condition θ, the control command y ref1 , y ref2 , the estimated amounts d1_hat and d2_hat of the disturbances are sufficient, and the physical quantities y1 and y2 measured by the physical sensors are not essential. Also, the learning model MD may be a model that does not require the input of the condition θ. In this case, the learning model MD is ref1 , y ref2 The sensor is trained to output estimated disturbance quantities d1_hat and d2_hat in response to the inputs of the sensor (and physical quantities y1 and y2).

[0089] While the example of FIG. 10 shows a configuration including two control systems S1 and S2, similar modeling is possible for configurations including one or more control systems. Furthermore, the interaction between a set of functions having multiple control systems as elements may also be modeled in a similar manner. For example, if the ESC temperature adjustment function controls the heater and the cooler individually, the heat input from the plasma is not a disturbance to the heater or the cooler alone, but rather a disturbance to their combined system. Therefore, it is reasonable to model the disturbance between the energy supply system to the plasma (including frequency control and power control of the high-frequency power supply) and the ESC temperature adjustment function (including heater control and chiller control).

[0090] Furthermore, a similar modeling may be performed for a higher-level function set having a plurality of control systems as elements, or a higher-level function set having a function set as an element.

[0091] 11 is an explanatory diagram illustrating a method for using the learning model MD. The learning model MD generated as described above can be considered as a digital replica or digital twin of the interactions between control systems in the apparatus (substrate processing apparatus 1). The control device 50 can use the learning model MD when performing a virtual experiment of the apparatus on a simulator.

[0092] Furthermore, the control device 50 executes a simulation using the learning model MD multiple times and evaluates the results to determine an optimal control command y ref1 , y ref2 For example, the control device 50 can generate an optimal heater output command value y that achieves both reduction in operation costs and the ability to follow an ideal wafer temperature change profile. ref1 and the optimal chiller output command value y ref2 can be found by repeating virtual experiments for various control command values ​​using the above method.

[0093] In the third embodiment, the control device 50 generates the learning model MD and executes a simulation using the generated learning model MD, but an external computer (computer) may generate the learning model MD and execute a simulation using the generated learning model MD. The computer that executes the above calculations may be any computer, such as a classical computer, an analog computer, a quantum annealer, an optical computer, or a gate-type quantum computer.

[0094] The embodiments disclosed herein are to be considered in all respects as illustrative and not restrictive. The scope of the present invention is defined by the claims, not by the above meaning, and is intended to include all modifications within the meaning and scope of the claims.

[0095] Furthermore, the features described in each embodiment can be combined with each other. Independent claims and dependent claims described in the claims can be combined with each other in any combination, regardless of the reference format. Furthermore, claims may be written in a multiple claim format that references two or more other claims, or in a multiple claim format that references at least one other multiple claim (multi-multi claim format).

[0096] REFERENCE SIGNS LIST 1 Substrate processing apparatus 50 Control device 51 Control unit 52 Storage unit 53 Communication unit 54 Operation unit 55 Display unit PG Control program RM Recording medium

Claims

1. A substrate processing apparatus that processes substrates through cooperative control of multiple control systems, comprising: an estimator that estimates disturbances in a first control system; and a controller that controls a control target in a second control system different from the first control system by reflecting the estimation result by the estimator.

2. The substrate processing apparatus according to claim 1, wherein the estimator is a soft sensor.

3. The substrate processing apparatus according to claim 1, wherein the estimator is an observer, a Kalman filter, a particle filter, or an estimator based on a recursive least squares method.

4. The substrate processing apparatus according to claim 1, wherein the estimator estimates the disturbance in a time series manner.

5. The substrate processing apparatus according to claim 4, further comprising a display that displays a time series change in the disturbance estimated by said estimator.

6. The substrate processing apparatus according to claim 1, wherein the controller controls the controlled object using a control law that reflects the estimation result by the estimator.

7. The substrate processing apparatus according to claim 1, further comprising a control device that generates a learning model that learns the relationship between the control commands for each of the first control system and the second control system and the disturbance, and generates control commands to be given to each of the first control system and the second control system based on a virtual experiment using the learning model.

8. A substrate processing apparatus as described in claim 1, further comprising a third control system different from the first control system and the second control system, wherein the third control system: acquires an estimation result by the estimator; generates control commands for each of the first control system and the second control system using a control law that reflects the acquired estimation result; and controls the control target to be controlled by the third control system by giving the generated control commands to each of the first control system and the second control system.

9. The substrate processing apparatus according to any one of claims 1 to 7, wherein the substrate processing is a plasma processing in which plasma is generated in a chamber that houses a substrate mounting table, and a substrate to be processed that is placed on the substrate mounting table is processed using the plasma generated in the chamber, the first control system is a control system for controlling the temperature of the substrate mounting table, the control object of the second control system is at least one of a plasma generation source, a sheath thickness, a gas flow rate in the chamber, and a pressure in the chamber, the estimator estimates heat input from the plasma as the disturbance, and the controller controls the control object using a control law that reflects parameters related to heat input estimated by the estimator.

10. A control method for a substrate processing apparatus in which substrates are processed through cooperative control by multiple control systems, the control method comprising the steps of: estimating a disturbance in a first control system; and controlling a control target in a second control system different from the first control system by reflecting the estimated disturbance.

11. A control program for a substrate processing apparatus in which multiple control systems perform cooperative control to process substrates, the control program causing a computer to execute the following process: estimating a disturbance in a first control system; and controlling a control target in a second control system different from the first control system by reflecting the estimated disturbance.

Images