Temperature control method and related device
By using a target model to dynamically adjust control parameters in semiconductor process equipment, the problem of unstable temperature control in the process chamber was solved, achieving rapid temperature stabilization and efficient control, thus improving process performance.
Patent Information
- Application Number
- CN202411125191.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-15
- Publication Date
- 2026-03-03
Smart Images
Figure CN121604754A_ABST
Abstract
Description
Technical Field
[0001] This specification relates to the field of semiconductor technology, specifically to automatic control technology within the semiconductor technology field, and more specifically to a temperature control method and related apparatus. Background Technology
[0002] For semiconductor process equipment such as Physical Vapor Deposition (PVD), it is usually necessary to control the temperature inside the process chamber during the process to meet the process requirements, that is, to ensure that the temperature inside the process chamber is stable at the set target temperature value, so as to ensure good process results for wafers and other processed objects.
[0003] Currently, most methods for controlling the temperature of process chambers in semiconductor manufacturing equipment employ PID (Proportional-Integral-Derivative) methods to control the operation of heating components (such as electrostatic chucks, ESCs) to achieve closed-loop control of the process chamber temperature. However, these technologies often suffer from poor temperature control of the process chamber, failing to ensure that the temperature remains stable at the set target value, resulting in poor wafer processing performance. Summary of the Invention
[0004] This specification provides a temperature control method and related apparatus to improve the temperature control effect.
[0005] To achieve the above technical objectives, the embodiments of this specification provide the following technical solutions:
[0006] In a first aspect, one embodiment of this specification provides a temperature control method applied to a semiconductor process apparatus, the semiconductor process apparatus including a process chamber, the process chamber including a heating element, a gate valve, and a process power supply, the process power supply being used to provide power during the process in the process chamber, the semiconductor process apparatus also carrying a target model, the temperature control method comprising:
[0007] During the heating phase of the process chamber, the heating element is controlled to operate based on initial control parameters; these initial control parameters are determined by the target model based on target parameters.
[0008] In response to the change information of the target parameter, the heating component is controlled to work based on the optimized control parameters, which are obtained by the target model by adjusting the initial control parameters according to the changed target parameter;
[0009] The target parameter includes at least one of the operating state of the process power supply, the on / off state of the valve, and the opening time of the valve. The change information of the target parameter includes at least one of the following: the operating state of the process power supply changes from off to on, the on / off state of the valve changes from off to on, and the opening time of the valve changes from 0 to a non-zero value.
[0010] Secondly, one embodiment of this specification provides a semiconductor process apparatus, including: a controller, a process chamber, heating components, a gate valve, and a process power supply, wherein the controller is equipped with a target model; wherein...
[0011] The controller is configured to control the heating element to operate based on initial control parameters during the heating phase of the process chamber; the initial control parameters are determined by the target model based on target parameters.
[0012] In response to the change information of the target parameter, the heating component is controlled to work based on the optimized control parameters, which are obtained by the target model by adjusting the initial control parameters according to the changed target parameter;
[0013] The target parameter includes at least one of the operating state of the process power supply, the on / off state of the valve, and the opening time of the valve. The change information of the target parameter includes at least one of the following: the operating state of the process power supply changes from off to on, the on / off state of the valve changes from off to on, and the opening time of the valve changes from 0 to a non-zero value.
[0014] Thirdly, one embodiment of this specification also provides a computing device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the temperature control method described above.
[0015] Fourthly, one embodiment of this specification also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the temperature control method described above.
[0016] Fifthly, embodiments of this specification provide a computer program product or computer program, the computer program product including a computer program stored in a computer-readable storage medium; the processor of the computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program to implement the steps of the temperature control method described above.
[0017] As can be seen from the above technical solution, the temperature control method provided in this specification, during the heating stage of the process chamber, controls the heating component to work based on the initial control parameters determined according to the target parameters using a target model, thereby meeting the temperature control requirements of the process chamber during the heating stage. If the target parameters change, the optimized control parameters obtained by adjusting the initial control parameters based on the changed target parameters using the target model are used to control the heating component to work. In this way, the influence of the working state of the process power supply, the opening and closing state of the valves, and the changes in the opening time of the valves on the temperature of the process chamber can be considered during the process. Optimized control parameters that are more in line with the actual working environment can be formulated to control the heating component, achieving the purpose of dynamically adjusting the control parameters according to the changes in the target parameters. This meets the requirement of stabilizing the temperature of the process chamber near the set target temperature value, thereby achieving the purpose of optimizing the temperature control effect.
[0018] Furthermore, in the temperature control method, both the initial control parameters and the optimized control parameters are determined based on the target model at different stages, eliminating the need for manual parameter setting and adjustment by maintenance personnel, thus offering the advantage of high efficiency. Attached Figure Description
[0019] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this specification. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0020] Figure 1 A schematic flowchart of a temperature control method provided for one embodiment of this specification;
[0021] Figure 2 A schematic diagram of a semiconductor process apparatus provided for one embodiment of this specification;
[0022] Figure 3 A schematic diagram of the control framework provided for one embodiment of this specification;
[0023] Figure 4 A schematic diagram of the structure of a target model provided for one embodiment of this specification;
[0024] Figure 5 This is a schematic diagram of the structure of a computing device provided for one embodiment of this specification. Detailed Implementation
[0025] Unless otherwise defined, the technical or scientific terms used in the embodiments of this specification shall have the ordinary meaning understood by one of ordinary skill in the art to which this specification pertains. The terms "first," "second," and similar terms used in the embodiments of this specification do not indicate any order, quantity, or importance, but are merely used to avoid confusion of constituent elements.
[0026] Unless the context otherwise requires, throughout this specification, "a plurality of" means "at least two," and "including" is interpreted as open-ended or encompassing, that is, "including, but not limited to." In the description of this specification, terms such as "one embodiment," "some embodiments," "exemplary embodiment," "example," "specific example," or "some examples" are intended to indicate that a particular feature, structure, material, or characteristic associated with that embodiment or example is included in at least one embodiment or example of this specification. The illustrative representations of the above terms do not necessarily refer to the same embodiment or example.
[0027] The technical solutions in the embodiments of this specification will now be described with reference to the accompanying drawings. All other embodiments obtained by those skilled in the art based on the embodiments in this specification without inventive effort are within the scope of protection of this specification.
[0028] Overview
[0029] Taking PVD (Polymer Deposition and Conversion) equipment as an example, the process chambers of PVD equipment are typically in a vacuum state. During the process, different types of process gases, gas flow rates, and target temperature values need to be introduced according to different process requirements. These factors all affect the quality of the deposited thin film, among which temperature control is the most complex and time-consuming part of the process. Currently, related technologies use PID control methods to regulate the ESC (Electronic Stress Controller) in PVD equipment. The inner and outer heating wires of the ESC act as the actuators of the PID control system, and the thermocouple corresponding to the inner heating wire of the ESC acts as the feedback mechanism of the PID control system, providing feedback signals based on the actual temperature of the ESC. During the operation of the ESC, the temperature controller controls the ESC based on pre-tuned PID parameters. After each process chamber is assembled and adjusted, the historical PID parameters obtained from years of experience need to be fine-tuned. This fine-tuning process requires manual adjustments by maintenance personnel based on experience, which is inefficient and time-consuming. In addition, the inventors discovered that temperature control of the process chamber based on fixed PID parameters may result in a mismatch between the PID parameters and the environmental parameters of the process chamber during some operating stages. This can lead to a situation where the temperature control based on the original PID parameters cannot stabilize the temperature of the process chamber near the set target temperature value, resulting in poor temperature control performance.
[0030] To address this issue, the inventors further discovered that during the heating phase of the process chamber, the heating component can be controlled based on initial control parameters determined by the target model according to the target parameters, thus meeting the temperature control requirements of the process chamber during the heating phase. If the target parameters change, in response to the change information, the optimized control parameters obtained by adjusting the initial control parameters based on the target model according to the changed target parameters are used to control the heating component. In this way, the influence of the working state of the process power supply, the opening and closing state of the valves, and the changes in the opening time of the valves on the temperature of the process chamber can be considered during the process. Optimized control parameters that are more in line with the actual working environment can be formulated to control the heating component, achieving the goal of dynamically adjusting the control parameters according to changes in the target parameters. This satisfies the requirement of stabilizing the temperature of the process chamber near the set target temperature value, thereby achieving the goal of optimizing the temperature control effect.
[0031] Furthermore, in the temperature control method, both the initial control parameters and the optimized control parameters are determined based on the target model at different stages, eliminating the need for manual parameter setting and adjustment by maintenance personnel, thus offering the advantage of high efficiency.
[0032] Based on the above concept, this specification provides a temperature control method. The temperature control method provided by this specification will be described exemplarily below with reference to the accompanying drawings.
[0033] Exemplary methods
[0034] To be applied Figure 1 Taking a controller of a semiconductor process equipment as an example, some embodiments of this specification exemplify the temperature control method. The controller can be a lower-level machine and / or a higher-level machine of the semiconductor process equipment, or it can be a temperature controller independent of the lower-level machine and / or the higher-level machine, or it can be a temperature controller integrated into the lower-level machine and / or the higher-level machine. The controller can carry a target model. The semiconductor process equipment also includes a process chamber, which includes heating components, a gate valve, and a process power supply. The process power supply provides power during the process in the process chamber. The semiconductor process equipment also carries a target model, but this specification does not limit this. The temperature control method includes:
[0035] S101: During the heating stage of the process chamber, the heating component is controlled to operate based on initial control parameters; the initial control parameters are determined by the target model based on target parameters.
[0036] S102: In response to the change information of the target parameter, the heating component is controlled to work based on the optimized control parameters, wherein the optimized control parameters are obtained by the target model by adjusting the initial control parameters according to the changed target parameter;
[0037] The target parameter includes at least one of the operating state of the process power supply, the on / off state of the valve, and the opening time of the valve. The change information of the target parameter includes at least one of the following: the operating state of the process power supply changes from off to on, the on / off state of the valve changes from off to on, and the opening time of the valve changes from 0 to a non-zero value.
[0038] refer to Figure 2 , Figure 2 A schematic diagram of a feasible semiconductor process apparatus is shown. This apparatus may include a process chamber, a host computer, a temperature controller, a transfer platform, a gate valve, a gate valve status sensor, a mass flow meter, and a vacuum pump. The process chamber contains a temperature sensor (e.g., a first sensor and a second sensor) and heating elements. The heating elements may include heating wires. The temperature controller can implement the temperature control method provided in the embodiments of this specification to control the operation of the heating wires. The vacuum pump is used to evacuate the process chamber. The mass flow meter is connected to the process chamber via a gas pipe and is used to detect the flow rate of the process gas entering the process chamber. The transfer platform is used to transfer wafers into the process chamber. The gate valve can control the sealing state of the process chamber. The gate valve status sensor can detect the opening and closing state of the gate valve. In some embodiments, the opening and closing state of the gate valve detected by the gate valve status sensor is transmitted to the temperature controller as part of the target parameters, enabling the temperature controller to execute the temperature control method provided in the embodiments of this specification based on target parameters such as the operating state of the process power supply. Taking PVD equipment as an example, the process power source may include a plasma power source. In this case, the process power source can be used to excite the process chamber to generate plasma so that the process chamber can operate based on plasma.
[0039] The warm-up phase refers to the process in which the process chamber heats up from room temperature to a certain temperature during the execution of the process. During this phase, the process chamber typically does not perform any specific process (which may include etching, film deposition, etc.). Only when the temperature in the process chamber reaches or approaches the set target value will the process enter the process phase and perform the target process on the wafer.
[0040] The initial control parameters used in the heating stage can be determined based on the target parameters during the heating stage using a pre-trained target model. This eliminates the need for maintenance personnel to manually set or adjust the parameters, which is highly efficient. Furthermore, with sufficient data for training the target model, the generalization ability of the target model can make the determined initial control parameters more consistent with the actual conditions of various types of process chambers, which is beneficial for optimizing the temperature control effect during the heating stage.
[0041] Once the chamber temperature reaches a certain value, the specific process can begin. During the process, parameters such as the opening and closing status and opening time of the valves between the transmission platform and the process chamber may affect the temperature of the process chamber. For example, after the valve is opened, the temperature inside the process chamber decreases, accelerating the heat dissipation of the heating components. At this time, the transfer function of the entire system changes significantly. If the initial control parameters are continued to be used for temperature control, the temperature of the process chamber cannot be quickly stabilized at the set target temperature value, resulting in poor temperature control performance.
[0042] Therefore, upon detecting changes in the target parameters, the heating component is controlled based on optimized control parameters in response to these changes in the objective function. This makes the control parameters (i.e., optimized control parameters) more consistent with the actual state of the process chamber. During the process, the influence of the working state of the process power supply, the opening and closing state of the valves, and the changes in the valve opening time on the process chamber temperature can be considered. Optimized control parameters that better suit the actual working environment can be formulated to control the heating component, achieving the goal of dynamically adjusting the control parameters according to changes in the target parameters. This meets the requirement of quickly stabilizing the temperature of the process chamber near the set target temperature value, thus achieving the goal of optimizing temperature control.
[0043] In determining the initial control parameters and optimizing the control parameters, in order to comprehensively consider the influence of the state of the process chamber on temperature control, in one embodiment, the target parameters also include at least one of the following: the output power of the process power supply, the real-time temperature inside the process chamber, the process gas flow rate, the derivative of the set target temperature value and the deviation value; the deviation value includes the difference between the set target temperature value and the real-time temperature inside the process chamber.
[0044] In this embodiment, analysis of the complete working process of the process chamber reveals that the aforementioned parameters affect the temperature within the process chamber from different perspectives. Incorporating these parameters into the target parameters allows for a comprehensive consideration of the influence of the process chamber's state on the chamber temperature when determining the control parameters, thus improving the accuracy of the control parameters determined by the target model.
[0045] Furthermore, the target parameters are not simply composed of the deviation value; rather, the derivative of the deviation value is included as part of the target parameters. This derivative can be the first derivative, which monitors the changing trend (e.g., increasing or decreasing) of the output value (i.e., the real-time temperature within the process chamber). This trend can be used to predict the output value at the next moment. Thus, the target model can predict the output value based on the derivative of the deviation value and determine the control parameters based on the predicted value. This ensures that the determined control parameters better match the actual state of the process chamber at the next moment, improving the temperature control effect.
[0046] In one embodiment, the real-time temperature within the process chamber can be characterized by multiple temperature values, allowing the real-time temperature within the process chamber to accurately reflect the actual temperature, improving the precision of the determined control parameters and enhancing the temperature control effect. Specifically, the real-time temperature within the process chamber includes: the real-time temperature at the valve position, and / or, the real-time temperature of the heating element;
[0047] The real-time temperature at the valve location is obtained based on multiple temperature sensors installed at the valve location.
[0048] Combination Figure 2 The first sensor can be used to detect the real-time temperature of the heating element, and the second sensor can be used to detect the real-time temperature at the valve position. In some embodiments, there can be multiple second sensors, and the detection values of multiple second sensors can be averaged to obtain the real-time temperature characterizing the valve position. In other embodiments, due to the relatively small internal space of the heating element, there can be only one first sensor. However, this specification does not limit the specific number of each temperature sensor (first sensor and second sensor), and it depends on the actual situation.
[0049] refer to Figure 3 , Figure 3 A schematic diagram of the control framework of the temperature control method provided in this specification is shown. The target model receives target parameters and outputs control parameters (initial control parameters or optimized control parameters). The temperature controller outputs current to the heating wire based on the control parameters to control the operation of the heating wire. Figure 3 In the middle, T G (t) represents the real-time temperature at the valve position (e.g., valve orifice) at time t, S G (t) represents the opening and closing state of the valve at time t, where T is the valve's opening and closing state. n(Δt) is the valve opening time, r(t) is the target temperature value set at time t, y(t) is the output value at time t (i.e., the real-time temperature in the process chamber (this temperature may include the real-time temperature of the heating element)), e(t) is the error at time t, M(t) is the process gas flow rate at time t, I(t) is the current output to the heating wire at time t, and S P P(t) represents the output state of the process power supply at time t, and P(t) represents the output power of the process power supply at time t.
[0050] The following implementation methods provide a feasible training process for a target model. Specifically, in one implementation method, the target model is trained based on a first dataset and a second dataset.
[0051] The first dataset includes pre-calibrated first sample data, which includes: first sample control parameters and first sample parameters. The type of the first sample control parameters is the same as the type of the initial control parameters, and the type of the first sample parameters is the same as the type of the target parameters. Among the first sample parameters, the operating state of the process power supply includes a closed state, the opening and closing state of the valve includes a closed state, and the opening time of the valve includes 0.
[0052] The second dataset includes pre-calibrated second sample data, which includes: second sample control parameters and second sample parameters. The type of the second sample control parameters is the same as the type of the optimized control parameters, and the type of the second sample parameters is the same as the type of the target parameters. In the second sample parameters, the working state of the process power supply includes an on state, the switching state of the valve includes an on state, and the opening time of the valve is greater than 0.
[0053] In this embodiment, by designing the first and second sample data to have the same data format but different values for some parameters, the goal of training the same neural network model sequentially using the first and second datasets is achieved. This allows the final trained target model to determine initial control parameters based on target parameters during the heating phase, and to determine optimized control parameters based on changes in the target parameters. Thus, the same target model can be used to meet the needs of rapid self-tuning of initial control parameters and dynamic adjustment of optimized control parameters, improving temperature control performance while reducing the number of required neural network models and lowering the cost of implementing this method.
[0054] Accordingly, based on the first and second datasets mentioned above, one embodiment of this specification provides a feasible training method. Specifically, the training process of the target model includes:
[0055] Construct an initial network model; the number of input layer nodes in the initial network model is the same as the number of parameter types in the first sample parameters and the second sample parameters;
[0056] The first dataset is divided into a first training set and a first validation set. The initial network model is trained using the first training set. During the training process, the weights and activation functions of the initial network model are updated based on the backpropagation algorithm. When the loss function value of the initial network model is less than a set loss value, the trained initial network model is validated using the first validation set. During the validation process, the hyperparameters of the initial network model are adjusted to obtain an intermediate network model.
[0057] The second dataset is divided into a second training set and a second validation set. The intermediate network model is trained using the second training set. During the training process, the weights and activation functions of the intermediate network model are updated based on the backpropagation algorithm. When the loss function value of the intermediate network model is less than a set loss value, the trained intermediate network model is validated using the second validation set. During the validation process, the hyperparameters of the intermediate network model are adjusted to obtain the target model.
[0058] refer to Figure 4 The model structure of the initial network model, intermediate network model, and target model can include an input layer, hidden layers, and an output layer, with the number of nodes in the input layer corresponding to the number of parameters in the target parameters. Figure 4 In the input layer, there are n nodes (e.g., ...). Figure 4 In the X1~X8 array, each node receives T. G (t), S G (t), M(t), S P (t), T n One of the parameters (Δt), P(t), de / dt, and r(t), and the hidden layer has q nodes (e.g., Figure 4 The output layer has m nodes (e.g., Z1 to Z7). Figure 4 In the diagram, Y1 to Y3), the weights between the input layer and the hidden layer are V, the nodes between the hidden layer and the output layer are W, the input variable is X, and the activation function is ReLU.
[0059] Hidden layers can be represented as
[0060] The output layer can be represented as
[0061] The error function can be expressed as
[0062] The activation function can be expressed as f(x) = max(0,x);
[0063] In this embodiment, during the training process, the model's weights and activation functions are updated based on the backpropagation algorithm. This allows for efficient calculation of the gradient of each weight, avoiding the need to calculate the derivative of each weight individually, which greatly speeds up the calculation and improves the training efficiency of the model.
[0064] In another embodiment, a feasible method for acquiring a dataset is provided. Specifically, the process of acquiring the first dataset includes:
[0065] Set a first preset parameter, which includes a first preset control parameter and a set target temperature value;
[0066] The semiconductor process equipment is controlled to operate according to a set strategy using a first preset parameter. During operation, other parameters in the first sample parameters, excluding the first preset parameter, are collected at a set frequency to obtain a set of first sample parameters. Some parameters in the first preset control parameters are changed, and the process returns to the step of controlling the semiconductor process equipment to operate according to the set strategy using the first preset parameter, until multiple sets of first sample parameters are obtained. The set strategy includes: heating to a set target temperature value, and after the heating component stabilizes at the set target temperature value, cooling the heating component to room temperature. The obtained first sample parameters are labeled to obtain the first dataset.
[0067] The process of obtaining the second dataset includes:
[0068] Set a second preset parameter, which includes a second preset control parameter and a set target temperature value;
[0069] The semiconductor process equipment is controlled to operate according to the set strategy using the second preset parameters. During operation, other parameters in the second sample parameters, excluding the second preset parameters, are collected at a set frequency to obtain a set of second sample parameters. Some parameters in the second preset control parameters are changed, and the process returns to the step of controlling the semiconductor process equipment to operate according to the set strategy using the second preset parameters, until multiple sets of second sample parameters are obtained. The obtained second sample parameters are labeled to obtain the second dataset.
[0070] After parameter acquisition and before data labeling, the acquired data can be preprocessed to remove obviously abnormal data. Specifically, during preprocessing, an allowable error value can be selected to filter the acquired data and remove data that does not meet the specifications. For example, when creating a dataset, the mean and standard deviation can be used to remove non-compliant data to ensure the accuracy of model training. The resulting set after filtering is {|y(k)―y(k―1)|<ε}, where k represents the k-th sampling point, ε is the minimum error value of the data that meets the requirements, the mean represents the center position of the heating component temperature during the measurement process, and the standard deviation reflects the ESC temperature variability during the measurement process.
[0071] In this embodiment, the first dataset can be used to pre-train the initial network model to obtain an intermediate network model that can generate initial control parameters based on the target parameters; the second dataset can be used to fine-tune the intermediate network model so that the trained target model can adjust the initial control parameters based on the changed target parameters when the target parameter changes are detected. The second dataset only needs to collect the change information of the target parameters at the process point temperature to meet the training requirements, and it does not need to be heated from room temperature to the actual process temperature, which helps to reduce the production cost of the second dataset.
[0072] In one specific implementation, taking initial control parameters and optimized control parameters including PID parameters as an example, a feasible process for obtaining a first dataset and a second dataset is provided, the process including:
[0073] For the first dataset, determine a set of initial PID parameters, set a target temperature value, and collect the real-time temperature inside the process chamber at a certain sampling frequency (e.g., the real-time temperature inside the process chamber can be characterized by the center temperature of the heating element) until the real-time temperature inside the process chamber stabilizes. Cool the real-time temperature inside the process chamber to room temperature, modify P while keeping I and D constant, set the target temperature value, and repeat the process of collecting the real-time temperature inside the process chamber at a certain sampling frequency (e.g., the real-time temperature inside the process chamber can be characterized by the center temperature of the heating element) until the real-time temperature inside the process chamber stabilizes. Repeat the above steps n times. Then, keep P and D constant, modify I, set the target temperature value, and repeat the above steps n times. Finally, keep P and I constant, modify D, set the target temperature value, and repeat the above steps n times to obtain multiple first sample data.
[0074] For example, suppose the first sample data includes: f, P, I, D, r(t), y(t), T n (Δt),S G (t),T n (Δt),e(t),M(t),SP P(t), where the meanings of each parameter are described above, the first sample data obtained can be represented as the following matrix:
[0075]
[0076] In this matrix, each row of data can be used as a first sample set.
[0077] Similarly, for the second dataset, the process of collecting the second sample data is similar to that of collecting the first sample data. The collected second sample data is assumed to include: f, P, I, D, r(t), y(t), T n (Δt),S G (t),T n (Δt),e(t),M(t),S P P(t), where the meanings of each parameter are described above, the corresponding second sample data can be represented as the following matrix:
[0078]
[0079] After the target model is trained on the first and second datasets, the model can be tested and evaluated. The real-time temperature inside the process chamber is collected and compared with the set target temperature value to evaluate whether the control action of the target model is within the specified range.
[0080] Finally, the target model can be deployed. During the temperature control process of the process chamber, the real-time temperature at the valve position, the valve's opening and closing status and opening time, the output status of the process power supply, the process gas flow rate, and the temperature setpoint are input into the target model. The real-time temperature of the heating element after heating is obtained through the temperature controller. The deviation between the set target temperature value and the real-time temperature is compared to see if it is within the process requirements. If not, the real-time temperature value is re-acquired and the output value of the target model (i.e., the initial control parameters or optimized control parameters) is updated. If so, the control parameters output by the target model are used to control the operation of the temperature controller.
[0081] One feasible embodiment of this specification provides an implementation of a feasible temperature control method, specifically, this implementation method can be based on, for example... Figure 2 The semiconductor process equipment shown includes a process chamber equipped with structures such as temperature sensors (e.g., a first sensor and a second sensor) and heating components. The heating components may include heating wires, and the temperature controller can implement the temperature control method provided in the embodiments of this specification to control the operation of the heating wires of the heating components.
[0082] This temperature control method may specifically include the following steps:
[0083] S1. The real-time temperature at the valve position and the real-time temperature inside the process chamber are collected in real time by a temperature sensor. The process gas flow rate is collected in real time by a mass flow meter in the process chamber. The valve status and opening and closing time between the transmission platform and the process chamber, as well as the process power output status, are monitored by a temperature controller. The data is stored in the memory of the temperature controller.
[0084] S2, Constructing with T G (t), S G (t), M(t), S P (t), T n Using (Δt), P(t), de / dt, and r(t) as input nodes, and the actions executed by the temperature controller as output, the steps for constructing the neural network model are as follows:
[0085] S2.1. Collect data from the first sample data according to the method described above;
[0086] S2.2 Select an acceptable error value, filter the collected data, remove outlier data that does not meet the specifications, and finally construct a data matrix from the data that meets the specifications;
[0087] S2.3. Label the dataset. Manually label the constructed data matrix. Use the labeled data and the corresponding timestamp set as the final first and second sample data. (During data filtering, some data will inevitably be discarded. It is necessary to match data at the same time, that is, label the physical quantities at the same time. For example, at time t1, set the value as r(t1), and output the value y(t1) and T at time t1.) G (t1) represents the real-time temperature at the valve position (e.g., valve orifice) at time t1, S G (t1) represents the opening and closing state of the valve at time t1, e(t1) is the error at time t1, M(t1) is the process gas flow rate at time t1, I(t1) is the current supplied to the heating wire at time t1, and S... P (t1) represents the output state of the process power supply at time t1, and P(t1) represents the output power of the process power supply at time t1. These physical quantities are corresponding, and the next set of data is also one-to-one. For example, in the above example, (0.1,100,50,0,300,50,30,0,0,0.2,0,0,10) is a set of data.
[0088] S2.4 Determine the training set and validation set. Divide the first sample data into a training set and a validation set. The ratio of the training set to the validation set can be, for example, 8:2. The first dataset is now complete.
[0089] S2.5 Training the Model: Determine the number of input layer nodes, hidden layer layers, activation functions, and output layer nodes to select the neural network model. Train the model using the training set from S2.4. The number of channels for the first sample data in the first dataset corresponds to the number of nodes in the network model's input layer. Input the first sample data into the model's input layer. The activation functions between layers add non-linearity to the network model, enabling it to learn complex tasks. The hidden layers extract features between the input and output, providing accurate signals to the output layer. The output layer automatically adjusts the controller's PID based on the trend of the first-order error derivative, outputting the PID parameters required for the current chamber type. During training, backpropagation updates the model weights and activation functions, iterating repeatedly until a satisfactory loss function value is obtained.
[0090] S2.6 Model Validation: The trained network model is validated using the validation set in S2.4 to further determine the model's hyperparameters.
[0091] S2.7 Model testing and evaluation: Compare the real-time temperature values with the set values to evaluate whether the control actions of the neural network model are within the specified range.
[0092] S2.8 Model Deployment: Deploy the neural network model to the process chamber controller, close the valve between the transmission platform and the process chamber, shut off the process power supply, shut off the process gas flow, set the required process temperature value and input it into the neural network model. The model will provide a PID parameter, which is then used by the PID controller to obtain the temperature value after ESC heating. Compare the set temperature value with the real-time temperature value to see if they are within the process requirements. If not, re-acquire the real-time temperature value and update the output value of the neural network model. If they are, use the parameters output by the neural network model to control the PID controller.
[0093] S2.9, Creating the second dataset:
[0094] S2.9.1. Using the neural network model trained on the first dataset, determine the PID parameters during the heating process. Heat the ESC to the target temperature. With the process gas flow rate determined, open and close the valve at a certain frequency, and collect the current ESC center temperature and valve port temperature n times. Change the process gas flow rate, open and close the valve at the same frequency, and collect the current ESC center temperature and valve port temperature n times. Fine-tune the PID parameters and repeat the above steps. With the process power supply output state, change the PID parameters n times and collect the ESC center temperature and valve port temperature at a certain sampling frequency.
[0095] The resulting data matrix is as described above;
[0096] S2.9.2 After obtaining the data matrix, the preprocessing, labeling, and partitioning of the training and validation sets are similar to the process of creating the first dataset, and will not be elaborated here.
[0097] S3.0: The model is trained using the training set in the second dataset. The output layer dynamically adjusts the PID controller based on the network model's prediction of future ESC temperature data by the process power supply output status, valve opening and closing time, and process gas flow rate, so that the temperature controller responds in a timely manner and keeps the ESC surface temperature stable. During the training process, backpropagation updates the model weights and activation function, and iterates repeatedly until a satisfactory loss function value is obtained.
[0098] S3.1 Model Validation: Validate the model using the validation set in the second dataset to further determine the model's hyperparameters;
[0099] S3.2 Model testing and evaluation: The temperature value collected in real time is compared with the set value to evaluate whether the control action of the neural network model is within the specified range. If the evaluation is within the specified range, the final target model is obtained.
[0100] S3.3 Model Deployment: Deploy the target model to the controller of the process chamber, collect initial control parameters and optimize control parameters and input them into the target model. The temperature value after ESC heating is obtained through the PID controller. Compare the temperature setpoint with the real-time temperature value to see if they are within the process requirements. If not, re-acquire the initial control parameters, optimize the control parameters and update the target model output value. If so, use the parameters output by the target model to control the PID controller.
[0101] Exemplary related devices
[0102] In one exemplary embodiment of this specification, a semiconductor process apparatus, such as... Figure 2 As shown, it includes: a controller, a process chamber, heating components, a gate valve, and a process power supply. The controller is equipped with a target model; wherein,
[0103] The controller is configured to control the heating element to operate based on initial control parameters during the heating phase of the process chamber; the initial control parameters are determined by the target model based on target parameters.
[0104] In response to the change information of the target parameter, the heating component is controlled to work based on the optimized control parameters, which are obtained by the target model by adjusting the initial control parameters according to the changed target parameter;
[0105] The target parameter includes at least one of the operating state of the process power supply, the on / off state of the valve, and the opening time of the valve. The change information of the target parameter includes at least one of the following: the operating state of the process power supply changes from off to on, the on / off state of the valve changes from off to on, and the opening time of the valve changes from 0 to a non-zero value.
[0106] Optionally, the target parameter further includes at least one of the following: the output power of the process power supply, the real-time temperature inside the process chamber, the process gas flow rate, the derivative of the set target temperature value and the deviation value; the deviation value includes the difference between the set target temperature value and the real-time temperature inside the process chamber; the real-time temperature inside the process chamber includes: the real-time temperature at the valve position, and / or, the real-time temperature of the heating element;
[0107] The semiconductor process equipment further includes: a first sensor and a second sensor, wherein the first sensor is disposed in the heating component and is used to detect the real-time temperature of the heating component;
[0108] The second sensor is located on the side of the valve facing the process chamber and is used to detect the real-time temperature at the valve location.
[0109] The controller can be used to execute the temperature control method described in any of the above embodiments. For specific limitations of the temperature control method, please refer to the relevant description above, which will not be repeated here.
[0110] In one exemplary embodiment of this specification, a temperature control device is also provided, applied to a semiconductor process apparatus. The semiconductor process apparatus includes a process chamber, the process chamber including a heating element, a gate valve, and a process power supply. The process power supply is used to provide power during the process in the process chamber. The semiconductor process apparatus also carries a target model. The temperature control device includes:
[0111] A temperature control module is used to control the operation of the heating component based on initial control parameters during the temperature rise phase of the process chamber; the initial control parameters are determined by the target model based on target parameters.
[0112] The process control module is used to control the operation of the heating component in response to changes in the target parameters, based on optimized control parameters. The optimized control parameters are obtained by the target model by adjusting the initial control parameters according to the changed target parameters.
[0113] The target parameter includes at least one of the operating state of the process power supply, the on / off state of the valve, and the opening time of the valve. The change information of the target parameter includes at least one of the following: the operating state of the process power supply changes from off to on, the on / off state of the valve changes from off to on, and the opening time of the valve changes from 0 to a non-zero value.
[0114] Specific limitations regarding the temperature control device can be found in the limitations regarding the temperature control method above, and will not be repeated here. Each module in the aforementioned temperature control device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.
[0115] Exemplary computing device
[0116] Another embodiment of this specification also proposes a computing device, see [link to documentation]. Figure 5 As shown, an exemplary embodiment of this specification also provides a computing device, including: a memory and a processor, the memory storing a computer program, the processor executing the computer program to perform the steps of the temperature control method according to various embodiments of this specification described above.
[0117] The internal structure of the computing device can be as follows: Figure 5 As shown, the computing device includes a processor, memory, network interface, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it follows the steps of the temperature control method according to various embodiments of this specification as described in the above embodiments.
[0118] The processor may include the main processor, as well as baseband chips, modems, etc.
[0119] The memory stores a program that executes the technical solution of this invention, and may also store an operating system and other critical business functions. Specifically, the program may include program code, which includes computer operation instructions. More specifically, the memory may include read-only memory (ROM), other types of static storage devices capable of storing static information and instructions, random access memory (RAM), other types of dynamic storage devices capable of storing information and instructions, disk storage, flash memory, etc.
[0120] The processor can be a general-purpose processor, such as a general-purpose central processing unit (CPU), a microprocessor, etc., or an application-specific integrated circuit (ASIC), or one or more integrated circuits used to control the execution of the program of the present invention. It can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.
[0121] Input devices may include devices that receive data and information input by the user, such as keyboards, mice, cameras, scanners, light pens, voice input devices, touch screens, pedometers, or gravity sensors.
[0122] Output devices may include devices that allow information to be output to the user, such as displays, printers, speakers, etc.
[0123] The communication interface may include any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.
[0124] The processor executes the program stored in the memory and calls other devices, which can be used to implement the various steps of any of the temperature control methods provided in the above embodiments of this specification.
[0125] The computing device may also include a display component and a voice component. The display component may be a liquid crystal display screen or an e-ink display screen. The input device of the computing device may be a touch layer covering the display component, or a button, trackball or touchpad set on the casing of the computing device, or an external keyboard, touchpad or mouse, etc.
[0126] Those skilled in the art will understand that Figure 5The structures shown are merely block diagrams of some structures related to the solutions in this specification and do not constitute a limitation on the computing devices on which the solutions in this specification are applied. Specific computing devices may include more or fewer components than those shown in the figures, or combine certain components, or have different component arrangements.
[0127] Exemplary computer program products and storage media
[0128] In addition to the methods and devices described above, the temperature control methods provided in the embodiments of this specification can also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the temperature control methods according to various embodiments of this specification as described in the "Exemplary Methods" section above.
[0129] The aforementioned computer program product can be implemented through hardware, software, or a combination thereof. In one optional embodiment, the computer program product is specifically embodied in a computer storage medium; in another optional embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.
[0130] The computer program product described herein can be written in any combination of one or more programming languages to perform the operations of the embodiments described herein. These programming languages include object-oriented programming languages such as Java and C++, as well as conventional procedural programming languages such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.
[0131] Furthermore, embodiments of this specification also provide a computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor of the steps in the temperature control methods according to various embodiments of this specification as described in the "Exemplary Methods" section above.
[0132] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this specification can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.
[0133] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0134] The embodiments described above are merely illustrative of several implementation methods outlined in this specification. While the descriptions are specific and detailed, they should not be construed as limiting the scope of the solutions provided in this specification. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this specification, and these all fall within the scope of protection of this specification. Therefore, the scope of protection for this patent should be determined by the appended claims.
Claims
1. A temperature control method, characterized in that, An application is made in semiconductor process equipment, the semiconductor process equipment including a process chamber, the process chamber including heating components, a gate valve, and a process power supply, the process power supply being used to provide power during the process in the process chamber, the semiconductor process equipment also carrying a target model, the temperature control method including: During the heating phase of the process chamber, the heating element is controlled to operate based on initial control parameters; these initial control parameters are determined by the target model based on target parameters. In response to changes in the target parameters, the heating component is controlled to operate based on optimized control parameters, which are obtained by adjusting the initial control parameters according to the changed target parameters from the target model. The target parameter includes at least one of the operating state of the process power supply, the on / off state of the valve, and the opening time of the valve. The change information of the target parameter includes at least one of the following: the operating state of the process power supply changes from off to on, the on / off state of the valve changes from off to on, and the opening time of the valve changes from 0 to a non-zero value.
2. The method according to claim 1, characterized in that, The target parameters also include at least one of the following: the output power of the process power supply, the real-time temperature inside the process chamber, the process gas flow rate, the derivative of the set target temperature value and the deviation value; the deviation value includes the difference between the set target temperature value and the real-time temperature inside the process chamber.
3. The method according to claim 2, characterized in that, The real-time temperature inside the process chamber includes: the real-time temperature at the valve location, and / or the real-time temperature of the heating element; The real-time temperature at the valve location is obtained based on multiple temperature sensors installed at the valve location.
4. The method according to claim 1, characterized in that, The target model was trained based on the first dataset and the second dataset; The first dataset includes pre-calibrated first sample data, which includes: first sample control parameters and first sample parameters. The type of the first sample control parameters is the same as the type of the initial control parameters, and the type of the first sample parameters is the same as the type of the target parameters. Among the first sample parameters, the operating state of the process power supply includes a closed state, the opening and closing state of the valve includes a closed state, and the opening time of the valve includes 0. The second dataset includes pre-calibrated second sample data, which includes: second sample control parameters and second sample parameters. The type of the second sample control parameters is the same as the type of the optimized control parameters, and the type of the second sample parameters is the same as the type of the target parameters. In the second sample parameters, the working state of the process power supply includes an on state, the switching state of the valve includes an on state, and the opening time of the valve is greater than 0.
5. The method according to claim 4, characterized in that, The training process of the target model includes: Construct an initial network model; the number of input layer nodes in the initial network model is the same as the number of parameter types in the first sample parameters and the second sample parameters; The first dataset is divided into a first training set and a first validation set. The initial network model is trained using the first training set. During the training process, the weights and activation functions of the initial network model are updated based on the backpropagation algorithm. When the loss function value of the initial network model is less than a set loss value, the trained initial network model is validated using the first validation set. During the validation process, the hyperparameters of the initial network model are adjusted to obtain an intermediate network model. The second dataset is divided into a second training set and a second validation set. The intermediate network model is trained using the second training set. During the training process, the weights and activation functions of the intermediate network model are updated based on the backpropagation algorithm. When the loss function value of the intermediate network model is less than a set loss value, the trained intermediate network model is validated using the second validation set. During the validation process, the hyperparameters of the intermediate network model are adjusted to obtain the target model.
6. The method according to claim 5, characterized in that, The process of obtaining the first dataset includes: Set a first preset parameter, which includes a first preset control parameter and a set target temperature value; The semiconductor process equipment is controlled to operate according to a set strategy using a first preset parameter. During operation, other parameters in the first sample parameters, excluding the first preset parameter, are collected at a set frequency to obtain a set of first sample parameters. Some parameters in the first preset control parameters are changed, and the process returns to the step of controlling the semiconductor process equipment to operate according to the set strategy using the first preset parameter, until multiple sets of first sample parameters are obtained. The obtained first sample parameters are labeled to obtain the first dataset. The set strategy includes: heating to the set target temperature value, and after the heating component stabilizes at the set target temperature value, cooling the heating component to room temperature. The process of obtaining the second dataset includes: Set a second preset parameter, which includes a second preset control parameter and a set target temperature value; The semiconductor process equipment is controlled to operate according to the set strategy using the second preset parameters. During operation, other parameters in the second sample parameters, excluding the second preset parameters, are collected at a set frequency to obtain a set of second sample parameters. Some parameters in the second preset control parameters are changed, and the process returns to the step of controlling the semiconductor process equipment to operate according to the set strategy using the second preset parameters, until multiple sets of second sample parameters are obtained. The obtained second sample parameters are labeled to obtain the second dataset.
7. A semiconductor process apparatus, characterized in that, include: The system includes a controller, a process chamber, heating components, valves, and a process power supply; the controller is equipped with a target model. The controller is configured to control the heating element to operate based on initial control parameters during the heating phase of the process chamber; the initial control parameters are determined by the target model based on target parameters. In response to changes in the target parameters, the heating component is controlled to operate based on optimized control parameters, which are obtained by adjusting the initial control parameters according to the changed target parameters from the target model. The target parameter includes at least one of the operating state of the process power supply, the on / off state of the valve, and the opening time of the valve. The change information of the target parameter includes at least one of the following: the operating state of the process power supply changes from off to on, the on / off state of the valve changes from off to on, and the opening time of the valve changes from 0 to a non-zero value.
8. The device according to claim 7, characterized in that, The target parameters also include at least one of the following: the output power of the process power supply, the real-time temperature inside the process chamber, the process gas flow rate, the derivative of the set target temperature value and the deviation value; the deviation value includes the difference between the set target temperature value and the real-time temperature inside the process chamber. The real-time temperature inside the process chamber includes: the real-time temperature at the valve location, and / or the real-time temperature of the heating element; The semiconductor process equipment further includes: a first sensor and a second sensor, wherein the first sensor is disposed in the heating component and is used to detect the real-time temperature of the heating component; The second sensor is located on the side of the valve facing the process chamber and is used to detect the real-time temperature at the valve location.
9. A computing device, characterized in that, The device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the temperature control method according to any one of claims 1 to 6.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the temperature control method according to any one of claims 1 to 6.