Processing apparatus, management apparatus, lithographic apparatus and article manufacturing method

Abstract

A processing device, a management device, a lithographic apparatus, and an article manufacturing method are disclosed. A processing device includes a driver configured to drive a control object, and a controller configured to control the driver by generating a command value for the driver based on a control error. The controller includes a first compensator configured to generate a first command value based on the control error, a second compensator configured to generate a second command value based on the control error, and an adder configured to obtain the command value by adding the first command value and the second command value. The second compensator includes a neural network whose parameter value is decided through learning, and input parameters input to the neural network include at least one of a driving condition of the driver and an environmental condition of a surrounding of the control object in addition to the control error.

Description

Technical Field

[0001] This invention relates to a processing apparatus, a management apparatus, a photolithography apparatus, and a method for manufacturing articles. Background Technology

[0002] Various techniques are provided for improving the control accuracy of a controlled object. Japanese Patent Publication No. 2006-128685 describes updating the parameters of a feedforward controller through iterative learning in a control system obtained by combining a feedback controller and a feedforward controller.

[0003] In recent years, techniques have been proposed to use neural networks to control objects and improve accuracy. The neural network is optimized for each controlled object. The optimized neural network is called the learned network. The learned neural network is then used to control the controlled object.

[0004] Control devices using neural networks can determine the parameter values ​​of the neural network by performing reinforcement learning. However, since the state of the controlled object changes over time, even a neural network optimized at a given time will no longer be optimal because the state of the controlled object has since changed. Therefore, the control accuracy of the control device deteriorates due to changes in the state of the controlled object. Summary of the Invention

[0005] This invention provides a technique that helps to suppress the degradation of control accuracy caused by changes in the state of the controlled object.

[0006] In one aspect, the present invention provides a processing apparatus comprising a driver configured to drive a controlled object, and a controller configured to control the driver by generating a command value for the driver based on a control error, wherein the controller includes a first compensator configured to generate a first command value based on the control error, a second compensator configured to generate a second command value based on the control error, and an adder configured to obtain a command value by adding the first command value and the second command value, the second compensator including a neural network for which parameter values ​​are determined by learning, and the input parameters to the neural network include, in addition to the control error, at least one of the driving conditions of the driver and the surrounding environmental conditions of the controlled object.

[0007] In a second aspect, the present invention provides a management device for managing a processing apparatus as defined in the first aspect, the management device comprising a learning device configured to re-determine parameter values ​​based on the control results of a controller on a controlled object, wherein the learning device repeatedly learns while changing at least one of driving conditions and environmental conditions.

[0008] In a third aspect, the present invention provides a photolithography apparatus for performing a process of transferring an original pattern onto a substrate, comprising a driver configured to drive a controlled object to perform the process, and a controller configured to control the driver by generating a command value for the driver based on a control error, wherein the controller includes a first compensator configured to generate a first command value based on the control error, a second compensator configured to generate a second command value based on the control error, and an adder configured to obtain the command value by adding the first command value and the second command value, the second compensator comprising a neural network for learning to determine parameter values, and configured to use the neural network to control the driver, wherein the input parameters input to the neural network include, in addition to the control error, at least one of the driving conditions of the driver and the ambient conditions surrounding the controlled object.

[0009] The present invention provides, in its fourth aspect, a method for manufacturing an article, comprising transferring a pattern of a master copy onto a substrate using a photolithography apparatus as defined in the third aspect, and processing the substrate that has undergone the transfer, wherein the article is obtained from the processed substrate.

[0010] Further features of the invention will become clear from the following description of exemplary embodiments (with reference to the accompanying drawings). Attached Figure Description

[0011] Figure 1 It is a block diagram showing the structure of the manufacturing system;

[0012] Figure 2 It is a block diagram showing the arrangement of the processing apparatus;

[0013] Figure 3 This is an example Figure 2 A block diagram showing the arrangement of the processing apparatus;

[0014] Figure 4 This is a flowchart illustrating the operation of the management device in the learning sequence;

[0015] Figure 5 This is a view illustrating the arrangement of the scanning exposure apparatus;

[0016] Figure 6 This is a flowchart illustrating the operation in an actual sequence of scanning exposure apparatus; and

[0017] Figure 7A and Figure 7B This is a graph illustrating the effect of reducing the control error of the controlled object according to the embodiment. Detailed Implementation

[0018] In the following description, embodiments will be detailed with reference to the accompanying drawings. It should be noted that the following embodiments are not intended to limit the scope of the claimed invention. Multiple features are described in the embodiments, but the invention is not limited to claiming all of these features, and multiple such features may be suitably combined. Furthermore, in the drawings, the same reference numerals are given the same or similar constructions, and repeated descriptions thereof are omitted.

[0019] <First Embodiment>

[0020] Figure 1 The structure of a manufacturing system MS according to this embodiment is shown. The manufacturing system MS may include, for example, a processing unit 1, a control unit 2 that controls the processing unit 1, and a management unit (learning unit) 3 that manages the processing unit 1 and the control unit 2. The processing unit 1 is a device that performs processing on a target object, such as a manufacturing device, an inspection device, or a monitoring device. The concept of processing may include the processing, inspection, monitoring, and observation of the target object.

[0021] Processing device 1 may include a controlled object and control the controlled object using a neural network that determines parameter values ​​for it through reinforcement learning. Control device 2 may be configured to send drive commands to processing device 1 and receive drive results or control results from processing device 1. Management device 3 may perform reinforcement learning to determine multiple parameter values ​​of the neural network of processing device 1. More specifically, management device 3 may determine multiple parameter values ​​of the neural network by repeatedly sending drive commands to processing device 1 and receiving drive results from processing device 1 while changing all or some of the multiple parameter values. Management device 3 can be understood as a learning device.

[0022] All or some of the functions of control device 2 can be incorporated into management device 3. All or some of the functions of control device 2 can be incorporated into processing device 1. Processing device 1, control device 2, and management device 3 can be physically integrated or formed separately. Processing device 1 can be controlled by control device 2 as a whole, or it can include components controlled by control device 2 and components not controlled by control device 2.

[0023] Figure 2An example of the arrangement of processing device 1 is shown. Processing device 1 may include a stage mechanism 5, which includes a stage (holder) ST as the controlled object. Processing device 1 may also include a sensor 6 for detecting the position or state of the stage ST, a driver 7 for driving the stage mechanism 5, and a controller 8 for receiving the output from the sensor 6 and providing a command value to the driver 7. The stage ST can hold a positioning target object. The stage ST can be guided by a guide (not shown). The stage mechanism 5 may include an actuator AC for moving the stage ST. The driver 7 drives the actuator AC. More specifically, for example, the driver 7 can supply current (electrical energy) to the actuator AC corresponding to the command value given from the controller 8. The actuator AC can move the stage ST by force (mechanical energy) corresponding to the current given from the driver 7. The controller 8 can control the position or state of the stage ST as the controlled object using a neural network that determines parameter values ​​for it through reinforcement learning. The drive command unit 9 gives the drive target position of the stage ST as the controlled object and the drive conditions of the driver 7. The environmental sensor 10 detects the environmental conditions around the stage ST and provides the detected environmental conditions to the controller 8.

[0024] Figure 3 This is an example. Figure 2 The block diagram shows the arrangement of the processing device 1. The controller 8 may include a subtractor 81, a first compensator 82, a second compensator (neural network) 83, and an adder 84. The subtractor 81 can calculate the control error as the difference between the drive command (e.g., target position) given from the control device 2 and the detection result (e.g., position of the stage ST) output from the sensor 6. The first compensator 82 can generate a first command value by performing a compensation calculation on the control error provided by the subtractor 81. The second compensator 83 includes a neural network. In addition to the control error provided by the subtractor 81, the neural network can also take the drive conditions input from the drive command unit 9 and the environmental conditions measured by the environmental sensor 10 as input parameters, and generate a second command value by performing a compensation calculation. The drive conditions and environmental conditions are information that is not input to the first compensator 82. The adder 84 can generate a command value by adding the first command value and the second command value. The controller 8, the driver 7, the stage mechanism 5, and the sensor 6 form a feedback control system that controls the stage ST, which is the controlled object, based on the control error.

[0025] The first compensator 82 can be, for example, a PID compensator, but it can be other compensators. For example, when L represents the number of inputs, M represents the number of intermediate layers, and N represents the number of outputs (L, M, and N are all positive integers), the second compensator 83 can be, for example, a neural network defined by the product of an L×M matrix and an M×N matrix. Multiple parameter values ​​of the neural network can be determined or updated through reinforcement learning performed by the management device 3. The first compensator 82 is not always necessary; the command values ​​provided to the driver 7 can be generated solely by the second compensator 83.

[0026] The drive conditions input from the drive command unit 9 may include, for example, at least one of the following: the current position of the stage, the target position, the drive direction, the drive stroke, the speed, the acceleration, the jerk, and the snap, but may be other drive conditions. Furthermore, the conditions may be values ​​such as the maximum value, the average value, or the variance of a series of drive operations. Alternatively, the conditions may be current values, past history at a specific time, or future target values ​​after a given time.

[0027] The environmental conditions input from the environmental sensor 10 may include, for example, at least one of the following: pressure, temperature, humidity, vibration, wind speed, and flow rate around the stage ST, but may be other conditions as long as they can be measured by the sensor. This value may be a current value, a past value prior to a given time, or a future value predicted based on past changes. Values ​​obtained by filtering the values ​​measured by the sensor may be used.

[0028] When the reward obtained from the control result of the controller 8 of the processing device 1 on the stage ST does not meet the predetermined criteria, the management device 3 can be used as a learning device or a relearning device to execute the learning sequence. In the learning sequence, the set of parameter values ​​consisting of multiple parameter values ​​of the second compensator (neural network) 83 can be determined or re-determined through reinforcement learning.

[0029] Figure 4 The operation of the management device 3 in the learning sequence is illustrated. In step S101, the management device 3 can initialize multiple parameter values ​​(a set of parameter values) of the second compensator (neural network) 83. In step S102, the management device 3 can send a command to the processing device 1 to drive the stage ST, which is the controlled object. More specifically, in step S102, the management device 3 can send a drive command to the controller 8 of the processing device 1 via the control device 2. In response, the controller 8 of the processing device 1 can cause the driver 7 to drive the stage ST according to the drive command, thereby controlling the position of the stage ST.

[0030] In step S103, the management device 3 can obtain drive data indicating the drive state of the stage ST, which is the controlled object in step S102, from the controller 8 of the processing device 1 via the control device 2. The drive data may include, for example, at least one of the output from the sensor 6 and the output from the subtractor 81. In step S104, the management device 3 can calculate a reward based on the drive data obtained in step S103. The reward can be calculated based on a predefined formula. For example, if the reward is calculated based on the control error, it can be calculated using a formula that gives the reciprocal of the control error, a formula that gives the reciprocal of the logarithm of the control error, a formula that gives the reciprocal of a quadratic function of the control error, etc., but it can be calculated using other formulas. In one example, the second compensator (neural network) 83 is more superior as the reward value increases. Conversely, the second compensator (neural network) 83 may be more superior as the reward value decreases.

[0031] In step S105, the management device 3 generates a new set of parameter values ​​by changing at least one of the multiple parameter values ​​of the second compensator (neural network) 83, and sets the new parameter values ​​in the second compensator (neural network) 83. Steps S106, S107, and S108 can be the same as steps S102, S103, and S104, respectively. In step S106, the management device 3 can send a command to the processing device 1 to drive the stage ST. More specifically, in step S106, the management device 3 can send a drive command to the controller 8 of the processing device 1 via the control device 2. In response, the controller 8 of the processing device 1 can cause the driver 7 to drive the stage ST according to the drive command, thereby controlling the position of the stage ST. In step S107, the management device 3 can obtain drive data indicating the drive state of the stage ST in step S106 from the controller 8 of the processing device 1 via the control device 2. In step S108, the management device 3 can calculate a reward based on the drive data obtained in step S107.

[0032] In step S109, the management device 3 determines whether the reward calculated in step S108 is higher than the reward calculated in step S104. Then, if the reward calculated in step S108 is higher than the reward calculated in step S104, in step S110, the management device 3 uses the set of parameter values ​​obtained after performing the change operation in step S105 as the latest parameter values. On the other hand, if the reward calculated in step S108 is not higher than the reward calculated in step S104, in step S111, the management device 3 does not use the set of parameter values ​​obtained after performing the change operation in step S105 and returns to step S105. In this case, in step S105, a new set of parameter values ​​is set in the second compensator (neural network) 83.

[0033] If step S110 is executed, then in step S112, the management device 3 determines whether the reward calculated in the immediately preceding step S108 meets the predetermined criteria. If the reward meets the predetermined criteria, then... Figure 4 The processing shown here ends. This means that the set of parameter values ​​generated in the immediately preceding step S105 is determined to be the set of parameter values ​​after reinforcement learning. The neural network with parameter values ​​set after reinforcement learning can be referred to as the learned model. On the other hand, if it is determined in step S112 that the reward calculated in the immediately preceding step S108 does not meet a predetermined criterion, then the management device 3 repeats the processing that started from step S105.

[0034] The driving conditions and environmental conditions during step S102 are not constant, and some or all of the possible conditions may change. That is, the management device 3 (learning device) repeats reinforcement learning while changing at least one of the driving conditions and environmental conditions. Furthermore, during the learning step, learning is repeatedly performed while changing the combination pattern among a first number of first combination patterns of driving conditions and environmental conditions. If the reward exceeds a predetermined value while learning is repeatedly performed, then the number of combination patterns of driving conditions and environmental conditions can be increased. That is, in this case, learning can be repeatedly performed while changing the combination pattern among a second number of second combination patterns, where the second number is greater than the first number.

[0035] The inventors have discovered that even if the history of control errors is the same, the future behavior of the controlled object may change due to differences in the driving conditions of the controlled object or the surrounding environmental conditions. In this embodiment, to address this situation, the input parameters to the neural network may include at least one of the driving conditions and environmental conditions in addition to the control error. Therefore, the neural network is learned to output command values ​​that can suppress the control error to the driver 7.

[0036] The processing device 1 can operate as a means of including the learned model (second compensator 83) obtained in the learning sequence within a sequence of processing the target object (hereinafter referred to as the actual sequence). In one example, the processing device 1 can execute the actual sequence under the management of the management device 3. However, in another example, the processing device 1 can execute the actual sequence independently of the management device 3.

[0037] <Second Embodiment>

[0038] The following will refer to Figure 6 The description applies the above manufacturing system MS to Figure 5The scanning exposure apparatus 500 shown is an example of a scanning exposure apparatus 500. The scanning exposure apparatus 500 is a step-scanning exposure apparatus that performs scanning exposure on a substrate 14 using slit light formed by a slit member. The scanning exposure apparatus 500 may include: an illumination optics system 23, a master substrate stage mechanism 12, a projection optics system 13, a substrate stage mechanism 15, a first position measuring device 17, a second position measuring device 18, a substrate marking measuring device 21, a substrate transport unit 22, and a controller 25.

[0039] The controller 25 controls the illumination optics system 23, the original plate mounting stage mechanism 12, the projection optics system 13, the substrate mounting stage mechanism 15, the first position measuring device 17, the second position measuring device 18, the substrate marking measuring device 21, and the substrate transport unit 22. The controller 25 controls the process of transferring the pattern of the original plate 11 onto the substrate 14. Furthermore, the controller 25 may include the functions of the controller 8 according to the first embodiment. The controller 25 may, for example, be formed by a programmable logic device (PLD) such as a field-programmable gate array (FPGA), an application-specific integrated circuit (ASIC), a general-purpose computer with a program installed, or a combination of all or some of these components.

[0040] The original mounting stage mechanism 12 may include an original mounting stage RST for holding the original 11 and a first actuator RAC for driving the original mounting stage RST. The substrate mounting stage mechanism 15 may include a substrate mounting stage WST for holding the substrate 14 and a second actuator WAC for driving the substrate mounting stage WST. The illumination optics system 23 illuminates the original 11. The illumination optics system 23 shapes the light emitted from the light source (not shown) into a long, strip-shaped or arc-shaped slit light in the X direction, for example, by a light-shielding member such as a shielding blade, and illuminates a portion of the original 11 with this slit light. The original 11 and the substrate 14 are held by the original mounting stage RST and the substrate mounting stage WST, respectively, and are arranged in nearly optically conjugate positions (on the object plane and image plane of the projection optics system 13) via the projection optics system 13.

[0041] The projection optical system 13 has a predetermined projection magnification (e.g., 1, 1 / 2, or 1 / 4) and projects the pattern of the master 11 onto the substrate 14 via slit light. The area on the substrate 14 where the pattern of the master 11 is projected (the area illuminated by the slit light) can be referred to as the illumination area. The master stage RST and the substrate stage WST are configured to move in a direction (Y direction) orthogonal to the optical axis (Z direction) of the projection optical system 13. The master stage RST and the substrate stage WST are scanned relative to each other synchronously at a speed ratio corresponding to the projection magnification of the projection optical system 13. This scanning of the substrate 14 in the Y direction relative to the illumination area transfers the pattern formed on the master 11 to the injection area of ​​the substrate 14. Then, by sequentially performing this scanning exposure on multiple injection areas of the substrate 14 while moving the substrate stage WST, the exposure process of one substrate 14 is completed.

[0042] The first position measuring device 17 includes, for example, a laser interferometer and measures the position of the master stage RST. For example, the laser interferometer illuminates a reflector (not shown) provided in the master stage RST with a laser beam and detects the displacement (displacement relative to a reference position) of the master stage RST by the interference between the laser beam reflected by the reflector and the laser beam reflected by a reference surface. The first position measuring device 17 can obtain the current position of the master stage RST based on this displacement. In this example, the first position measuring device 17 can measure the position of the master stage RST using a position measuring device (e.g., an encoder) instead of a laser interferometer. The substrate marking measuring device 21 includes, for example, an optical system and an image sensor, and can detect the position of a mark provided on the substrate 14.

[0043] The second position measuring device 18 includes, for example, a laser interferometer and measures the position of the substrate stage WST. For instance, the laser interferometer illuminates a reflector (not shown) provided in the substrate stage WST with a laser beam and detects the displacement of the substrate stage WST (displacement relative to a reference position) by the interference between the laser beam reflected by the reflector and the laser beam reflected by a reference surface. The second position measuring device 18 can then determine the current position of the substrate stage WST based on this displacement. In this example, the second position measuring device 18 can measure the position of the substrate stage WST using a position measuring device (e.g., an encoder) instead of a laser interferometer.

[0044] Sensors 30, 31, and 32 are positioned near the controlled object and can detect environmental conditions surrounding the controlled object, such as pressure, temperature, humidity, vibration, wind speed, and flow rate. Figure 5In the example shown, sensor 30 is arranged near the substrate stage WST, sensor 31 is arranged near the original stage RST, and sensor 32 is arranged near the projection optics system 13.

[0045] The scanning exposure apparatus 500 is required to accurately transfer the pattern of the original plate 11 to the target position on the substrate 14. To achieve this, it is important to accurately control the relative position of the original plate 11 on the original plate stage RST relative to the substrate 14 on the substrate stage WST during scanning exposure. Therefore, as a bonus, a value for evaluating the relative position error (synchronization error) between the original plate stage RST and the substrate stage WST can be used. To improve the detection accuracy of the markings on the substrate 14, it is important to accurately position the substrate stage WST below the substrate marking measurement device 21. Therefore, as a bonus, a value for evaluating the control error of the substrate stage WST while imaging the markings can be used. To increase throughput, it is important to increase the substrate transport speed. Furthermore, it is important that the control errors of the substrate transport unit 22 and the substrate stage WST converge to below a predetermined value within a short time after the drive is completed during substrate loading and unloading. Therefore, as a bonus, a value for evaluating the convergence time of the substrate transport unit 22 and the substrate stage WST can be used. Each of the substrate mounting stage mechanism 15, the original plate mounting stage mechanism 12, and the substrate transport unit 22 is an example of an operation unit that performs the operation of transferring the pattern of the original plate 11 to the substrate 14.

[0046] Figure 6 An example of the actual sequence of the scanning exposure apparatus 500 is shown. This actual sequence begins when the management device 3 instructs the controller 25 of the scanning exposure apparatus 500 to execute the actual sequence (substrate processing sequence). The substrate processing sequence may, for example, include steps S301, S302, and S303 as multiple sub-sequences.

[0047] In step S301, the controller 25 controls the substrate transport unit 22 to load (transport) the substrate 14 onto the substrate mounting stage WST. In step S302, the controller 25 can control the substrate mounting stage mechanism 15 so that the mark on the substrate 14 falls within the field of view of the substrate mark measuring device 21, and control the substrate mark measuring device 21 to detect the position of the mark on the substrate 14. This operation can be performed for each of the multiple marks on the substrate 14. In step S303, the controller 25 controls the substrate mounting stage mechanism 15, the original plate mounting stage mechanism 12, the illumination optical system 23, etc., so that the pattern of the original plate 11 is transferred to each of the multiple injection areas of the substrate 14 (exposure step). In step S304, the controller 25 controls the substrate transport unit 22 to unload (transport) the substrate 14 from the substrate mounting stage WST.

[0048] In step S301, in order for the substrate transport unit 22 to accurately place the substrate 14 on the substrate mounting stage WST, a high positioning accuracy of the substrate transport unit 22 is required. In this case, the speed, acceleration, and jerk of the substrate transport unit 22 can be obtained as driving conditions input to the second compensator 83. As environmental conditions, the pressure when the substrate transport unit 22 adsorbs the substrate 14 can be obtained, or the output from the acceleration sensor provided on the substrate transport unit 22 when the substrate transport unit 22 is driven can be obtained.

[0049] In step S302, it is required that the error of the substrate mounting stage WST be converged as quickly as possible by driving the substrate mounting stage WST so that the mark on the substrate 14 is located directly below the substrate mark measuring device 21. In this case, the driving conditions input to the second compensator 83 may be at least one of the velocity, acceleration, and jerk of the substrate mounting stage WST. Alternatively, the driving conditions may be at least one of the direction and distance when driving the substrate mounting stage WST so that the mark to be measured next is located directly below the substrate mark measuring device 21 with a given mark located directly below the substrate mark measuring device 21. In addition, the environmental conditions may be at least one of the change in pressure in the space measured by the pressure sensor when the substrate mounting stage WST is driven and the output from the accelerometer provided near the substrate mark measuring device.

[0050] In step S303, the driving conditions input to the second compensator 83 can be at least one of the following information.

[0051] • Coordinates used to designate the injection area as the exposure target on the substrate

[0052] • Dimensions of the injection zone in the X and / or Y directions

[0053] • The speed, acceleration, jerk, and / or drive direction when moving the substrate stage or master stage during exposure.

[0054] The environmental conditions input to the second compensator 83 may be the intensity of the exposure light illuminating the substrate and at least one of pressure, temperature, humidity, vibration, wind speed and flow rate detected by sensors 30, 31 and 32.

[0055] Examples of control objects for which a neural network is formed are the substrate stage mechanism 15, the original substrate stage mechanism 12, and the substrate transport unit 22; however, the neural network can be incorporated into other components. For example, multiple components such as the substrate stage mechanism 15, the original substrate stage mechanism 12, and the substrate transport unit 22 can be controlled by a single neural network, or multiple components can be controlled by different neural networks respectively. Furthermore, the same or different learned models can be used for the transport sequence, measurement sequence, and exposure sequence as the learned models. In the calculation of the reward, the same or different formulas can be used for the transport sequence, measurement sequence, and exposure sequence.

[0056] Figure 7A and Figure 7B This is a graph showing the reduction effect of the control error of the controlled object when this embodiment is applied. Figure 7A The time profile of acceleration as a driving condition is shown. When comparing graph 600, indicated by the solid line, with graph 601, indicated by the dashed line, the change in acceleration up to the midpoint is the same, but the timing of the acceleration returning to 0 differs. (Reference) Figure 7B Curve 603 represents the control error when the controlled object is driven with the acceleration profile of curve 600, while curve 604 represents the control error when the controlled object is driven with the acceleration profile of curve 601. Within interval 602 where acceleration conditions are the same, curves 603 and 604 indicate identical error waveforms, but the error transitions become different in the middle. By inputting acceleration conditions into the neural network, a neural network capable of suppressing errors can be provided, even if the error history is the same but differences arise in the error due to variations in future acceleration conditions.

[0057] The example of applying the manufacturing system MS to the scanning exposure apparatus 500 has been explained above. However, the manufacturing system MS can be applied to other types of exposure apparatus (e.g., stepper apparatus) or other types of photolithography apparatus such as imprint apparatus. In this case, the photolithography apparatus is an apparatus for forming patterns on a substrate, and this concept includes exposure apparatus, imprint apparatus, and electron beam tracing apparatus.

[0058] The following describes a method for manufacturing articles (e.g., semiconductor IC components, liquid crystal display components, or MEMS) using the aforementioned photolithography apparatus. The article manufacturing method may include a transfer step of transferring a pattern from a master copy onto a substrate using a photolithography apparatus, and a processing step of processing the substrate that has undergone the transfer step, thereby obtaining an article from the processed substrate.

[0059] When the photolithography apparatus is an exposure apparatus, the article manufacturing method may include: a step of exposing a substrate (substrate, glass substrate, etc.) coated with a photosensitive agent, a step of developing the substrate (photosensitive agent), and a step of processing the developed substrate in other known steps. Other known steps include etching, resist removal, dicing, bonding, and encapsulation. According to this article manufacturing method, articles of higher quality can be manufactured compared to conventional articles. When the photolithography apparatus is an imprinting apparatus, the article manufacturing method may include: a step of forming a pattern made of a cured product of an imprinting material by using a mold to mold an imprinting material onto a substrate, and a step of processing the substrate using the pattern.

[0060] While the invention has been described with reference to exemplary embodiments, it should be understood that the invention is not limited to the disclosed exemplary embodiments. The scope of the following claims should be given the broadest interpretation to cover all such modifications and equivalent structures and functions.

Claims

1. A processing apparatus, characterized in that, include: The driver is configured to drive the controlled object; as well as The controller is configured to control the driver by generating command values ​​for the driver based on control errors. The controller includes: The first compensator is configured to generate a first command value based on the control error. The second compensator is configured to generate a second command value based on the control error, and The adder is configured to obtain the command value by adding the first command value and the second command value. The second compensator includes a neural network whose parameter values ​​are determined by learning. The input parameters fed into the neural network include, in addition to the control error, at least one of the driving conditions of the driver and the environmental conditions surrounding the controlled object. Among them, the driving conditions and environmental conditions, which are input parameters, are information that is not input to the first compensator.

2. The apparatus according to claim 1, wherein, The driving conditions include at least one of the following: the position of the controlled object, the driving direction, the driving stroke, the speed, the acceleration, the jerk, and the latch.

3. The apparatus according to claim 1, wherein, Environmental conditions include at least one of the following: pressure, temperature, and vibration around the controlled object.

4. A management device for managing a processing apparatus according to any one of claims 1 to 3, the management device comprising: The learning device is configured to learn and re-determine parameter values ​​based on the controller's control results on the controlled object. The learning device repeatedly learns while changing at least one of the driving conditions and environmental conditions.

5. The apparatus according to claim 4, wherein, If the reward obtained from the controller's control of the controlled object does not meet the standard, the learning device repeats the learning process while changing at least one of the driving conditions and environmental conditions.

6. The apparatus according to claim 5, wherein, If the reward obtained from the controller's control result on the controlled object does not meet the standard, the learning device repeats the learning process while changing the combination mode among the first combination modes of the first quantity of driving conditions and environmental conditions. When the reward exceeds a predetermined value while the learning is repeated, the learning device repeats the learning process while changing the combination pattern in a second combination pattern that is greater than the first number.

7. A photolithography apparatus for performing a process of transferring a pattern from an original image onto a substrate, characterized in that, include: The driver is configured to drive the control object to perform the aforementioned processing; as well as The controller is configured to control the driver by generating command values ​​for the driver based on control errors. The controller includes: The first compensator is configured to generate a first command value based on the control error. The second compensator is configured to generate a second command value based on the control error, and The adder is configured to obtain the command value by adding the first command value and the second command value. The second compensator includes a neural network that learns to determine parameter values, and is configured to use the neural network to control the driver. The input parameters fed into the neural network include, in addition to the control error, at least one of the driving conditions of the driver and the environmental conditions surrounding the controlled object. Among them, the driving conditions and environmental conditions, which are input parameters, are information that is not input to the first compensator.

8. A method for manufacturing an article, characterized in that, include: The original pattern is transferred to the substrate using the photolithography apparatus according to claim 7; as well as Processing substrates that have already undergone transfer printing, The items are obtained from processed substrates.

Images