Control method and device for automatically adjusting wafer thinning thickness

Abstract

The invention provides a control method and device for automatically adjusting the wafer thinning thickness, and relates to the technical field of semiconductor manufacturing. The method comprises the following steps: acquiring an initial TTV value measured by an online TTV measurement module in a wafer thinning process; based on the initial TTV value, utilizing a TTV compensation model to obtain a corrected TTV value; if the corrected TTV value is not in the set specification, determining an optimal TTV angle parameter through a pre-trained prediction model; and the angle of a grinding disc in the wafer thinning machining equipment is controlled to be adjusted to the TTV optimal angle parameter so as to act on next wafer thinning thickness machining. According to the method, real-time accurate control and continuous optimization of the wafer thinning process can be realized, and the production efficiency and the product quality are improved.

Description

Technical Field

[0001] The present application relates to the field of semiconductor manufacturing technology, and in particular to a control method and device for automatically adjusting the thickness of a wafer thinning. Background Art

[0002] In the field of semiconductor manufacturing, especially in the wafer thinning process, precise control of wafer thickness variation (TTV) is crucial to ensure wafer performance and quality. Although the wafer thinning technology in related technologies is relatively mature, it still has some shortcomings.

[0003] At present, the common practice is to conduct sample tests before each shift, and adjust the parameters such as the inclination angle of the carrier through manual judgment to control TTV. However, this method has obvious defects: on the one hand, the subjective factors of human judgment have a greater impact, resulting in the inability to guarantee the accuracy of the results; on the other hand, the single adjustment method cannot adapt to the dynamic changes in the subsequent production process, and it is difficult to meet the requirements of higher-precision TTV control. In addition, the determination of TTV inclination adjustment in related technologies often relies on empirical formulas, which further increases the uncertainty of the results. These problems not only affect the accuracy of wafer thinning, but may also lead to reduced production efficiency and increased production costs. Summary of the invention

[0004] Based on this, it is necessary to provide a control method for automatically adjusting the wafer thinning thickness to address the above technical problems, which method can achieve real-time and precise control of the wafer thinning process.

[0005] In a first aspect, the present application provides a control method for automatically adjusting the wafer thinning thickness, the method being applied to a control module in a wafer thinning system, the wafer thinning system further comprising an online TTV measurement module and a wafer thinning processing device, the method comprising: Obtaining an initial TTV value measured by an online TTV measurement module during wafer thinning; wherein the measurement points collected by the online TTV measurement module are smaller than the measurement points collected by the FRT measurement device, and the sampling resolution is smaller than the sampling resolution of the FRT measurement device, and the FRT measurement device measures the TTV value of the same wafer after thinning is completed; Based on the initial TTV value, a TTV compensation model is used to obtain a corrected TTV value; wherein the TTV compensation model is obtained by training at least TTV values ​​measured by an online TTV measurement module and an FRT measurement device on the same wafer; If the corrected TTV value is not within the set specification, the optimal TTV angle parameter is determined by a pre-trained prediction model; wherein the environmental characteristic data during wafer thinning is used as a constant of the prediction model, the TTV angle parameter is used as an optimization target of the prediction model, and the expected TTV value is used as an output target of the prediction model to perform a traversal of the feasible set of TTV angles to obtain the optimal angle parameter; wherein the environmental characteristic data includes process parameter characteristics; The grinding disc angle in the wafer thinning processing equipment is controlled to be adjusted to the TTV optimal angle parameter to act on the next wafer thinning thickness processing.

[0006] In one embodiment, obtaining a corrected TTV value based on the initial TTV value using a TTV compensation model includes: Inputting the initial TTV value into the high- and low-resolution mapping model to obtain a virtual TTV value measured by a virtual FRT measurement device; the high- and low-resolution mapping model is obtained by mapping TTV values ​​measured on the same wafer using at least an online TTV measurement module and an FRT measurement device; The initial TTV value and the virtual TTV value are spliced ​​and input into the TTV compensation model to obtain the corrected TTV value.

[0007] In one embodiment, training the TTV compensation model includes: Establish a TTV compensation model based on neural network; Splicing the first resolution data and the second resolution data of the same wafer to obtain spliced ​​data; wherein, for the same wafer, the first resolution data at least includes the TTV value measured by the online TTV measurement module, and the second resolution data at least includes the TTV value measured by the FRT measurement device; The spliced ​​data is used as the input of the initial TTV compensation model, and the TTV value in the second resolution data used for splicing is used as the output to train the TTV compensation model.

[0008] In one embodiment, determining a high-resolution and low-resolution mapping model includes: Establishing an encoder f for converting the second resolution data to the first resolution data, and establishing a decoder g for converting the first resolution data to the second resolution data; The encoder f and the decoder g are trained based on the first resolution data and the second resolution data of the same wafer, and a high-low resolution mapping model is obtained based on the trained encoder f and the decoder g.

[0009] In one embodiment, training the prediction model includes: Acquire sample characteristics of the sample wafer, wherein for any sample wafer, the sample characteristics include environmental characteristic data, TTV angle parameters and TTV values; Build a prediction model based on sample characteristics; Introduce the L1 regularization coefficient to shrink the coefficients of non-correlated feature data in the prediction model; Introduce L2 regularization term to limit the weight of feature data; The prediction model is trained using the gradient descent method to gradually adjust the parameters of the prediction model.

[0010] In one embodiment, optimizing the TTV angle parameter includes: Determine the value range of TTV angle parameters; Traversing the parameter grid within the value range, for each parameter combination in the parameter grid, using the prediction model to calculate the TTV value; the TTV angle parameter includes a first angle parameter of the first support leg and a second angle parameter of the second support angle, and the parameter combination consists of the first angle parameter and the second angle parameter; In the traversal parameter grid, the parameter combination that minimizes the TTV value is selected as the optimal TTV angle parameter.

[0011] In one of the embodiments, the environmental characteristic data includes time domain characteristics and frequency domain characteristics in addition to process parameter characteristics; Among them, obtaining environmental characteristic data includes: Build a data acquisition device that can collect process parameters, process parameters and TTV values ​​in real time during wafer thinning. The data acquisition device includes a built-in PLC data real-time acquisition module of the wafer thinning processing equipment and an external acquisition sensor module; The collected process parameters and procedure parameters are subjected to feature extraction to obtain time domain features, frequency domain features and process parameter features.

[0012] In one of the embodiments, a PLC data real-time acquisition module built into the wafer thinning processing equipment is used to collect process parameters and process parameters inside the wafer thinning processing equipment through the TCP / IP protocol; The external acquisition sensor module is used to realize A / D conversion of analog signals through the data acquisition card and encapsulate digital signals through the Modbus protocol.

[0013] In a second aspect, the present application also provides a control device for automatically adjusting the wafer thinning thickness, the device comprising: An acquisition module is used to acquire an initial TTV value measured by an online TTV measurement module during wafer thinning; wherein the measurement points collected by the online TTV measurement module are smaller than the measurement points collected by the FRT measurement device, and the sampling resolution is smaller than the sampling resolution of the FRT measurement device, and the FRT measurement device measures the TTV value of the same wafer after thinning is completed; A compensation module, for obtaining a corrected TTV value based on the initial TTV value using a TTV compensation model; wherein the TTV compensation model is obtained by at least training the TTV value measured by the online TTV measurement module and the FRT measurement device on the same wafer; An optimization module is used to determine the optimal TTV angle parameter through a pre-trained prediction model if the corrected TTV value is not within the set specification, use the environmental characteristic data in the wafer thinning process as a constant of the prediction model, use the TTV angle parameter as the optimization target of the prediction model, and use the expected TTV value as the output target of the prediction model to perform a traversal of the feasible set of TTV angles to obtain the optimal angle parameter; wherein the environmental characteristic data includes process parameter characteristics; The adjustment module is used to control the grinding disc angle in the wafer thinning processing equipment to adjust to the TTV optimal angle parameter to act on the next wafer thinning thickness processing.

[0014] In a third aspect, the present application further provides a wafer thinning system, the wafer thinning system comprising: Wafer thinning processing equipment, used for thinning wafers; Online TTV measurement module collects wafer thickness data in real time during wafer thinning process of wafer thinning equipment; The control module includes a memory and a processor. The memory stores a computer program. When the processor executes the computer program, the control method for automatically adjusting the wafer thinning thickness of the first aspect is implemented.

[0015] The above-mentioned control method for automatically adjusting the wafer thinning thickness uses an online TTV measurement module to obtain the initial TTV value in real time during the wafer thinning process; uses a TTV compensation model trained based on the measurement data of the online TTV measurement module and the FRT measurement equipment to correct the initial TTV value to obtain a corrected TTV value that is closer to the FRT measurement result; if the corrected TTV value exceeds the set specification, the optimal TTV angle parameter is determined by a pre-trained prediction model, and the prediction model uses environmental feature data (including process parameter features) as a constant and TTV angle parameters as the optimization target, and obtains the optimal angle parameter by traversing the feasible set; the grinding disc angle in the wafer thinning processing equipment is adjusted to the optimal TTV angle parameter to optimize the next wafer thinning process. This method can achieve real-time precise control and continuous optimization of the wafer thinning process, and improve production efficiency and product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A flow chart of a control method for automatically adjusting wafer thinning thickness in one embodiment; Figure 2A flowchart of obtaining a corrected TTV value using a TTV compensation model based on an initial TTV value in one embodiment; Figure 3 is a structural diagram of a TTV correction value obtained by using a TTV compensation model in one embodiment; Figure 4 A flowchart of training a TTV compensation model in one embodiment; Figure 5 A flowchart for determining a high-resolution and low-resolution mapping model in one embodiment; Figure 6 is a structural diagram of a high-low resolution mapping model in one embodiment; Figure 7 A flowchart of training a prediction model in one embodiment; Figure 8 A flowchart of optimizing TTV angle parameters in one embodiment; Fig. 9 A flow chart of wafer thinning run-to-run control in one embodiment; Fig.10 A device diagram of a control device for automatically adjusting wafer thinning thickness in one embodiment; Fig.11 FIG. 4 is a structural diagram of a wafer thinning system in one embodiment. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0018] In one embodiment, Figure 1 As shown, a control method for automatically adjusting the wafer thinning thickness is provided, the method is applied to a control module in a wafer thinning system, the wafer thinning system also includes an online TTV measurement module and a wafer thinning processing device, the method includes the following steps: Step 101: obtaining an initial TTV value measured by an online TTV measurement module during wafer thinning; wherein the measurement points collected by the online TTV measurement module are smaller than the measurement points collected by the FRT measurement device, and the sampling resolution is smaller than the sampling resolution of the FRT measurement device, and the FRT measurement device measures the TTV value of the same wafer after thinning is completed; Wafers are thin sheets of high-purity semiconductor materials used in semiconductor manufacturing. They are the basic materials for manufacturing semiconductor devices such as integrated circuits and chips. Common wafer materials include silicon wafers, silicon carbide wafers, and sapphire wafers.

[0019] The online TTV measurement module can obtain the initial TTV value of the wafer in real time during the thinning process. The number of measurement points and sampling resolution are relatively low. When measuring in real time on the production line, it is necessary to quickly obtain data to find obvious problems in time. If there are too many measurement points or the resolution is too high, it will greatly increase the measurement time and calculation amount, affecting production efficiency. Therefore, the online TTV measurement module improves the measurement speed by reducing the measurement points and lowering the sampling resolution.

[0020] The FRT measuring equipment performs more detailed and accurate TTV measurement on the same wafer after wafer thinning. Therefore, the accuracy and comprehensiveness of wafer measurement are required to be higher. The FRT measuring equipment can collect more measurement points and has a higher sampling resolution, thus providing more accurate TTV value information. Among them, the TTV value refers to the change in wafer thickness, which is used to evaluate the quality of the wafer after thinning.

[0021] Step 102: obtaining a corrected TTV value based on the initial TTV value using a TTV compensation model; wherein the TTV compensation model is obtained by training TTV values ​​measured on the same wafer using at least an online TTV measurement module and an FRT measurement device; In the process of training the TTV compensation model, the online TTV measurement module and the FRT measurement equipment can be used to measure multiple wafers and collect TTV data of multiple wafers. Furthermore, the TTV data of these wafers include the TTV values ​​of low-resolution online measurements and a small number of measurement points, as well as the TTV values ​​of high-resolution FRT measurements and a large number of measurement points. By comparing these two measurement results, the TTV compensation model can learn the deviation between the online measurement value and the actual high-precision FRT measurement value.

[0022] Furthermore, for the training of the TTV compensation model, paired online and FRT measurement values ​​can be used as training data and input into a neural network. By adjusting the parameters of the TTV compensation model, the output of the TTV compensation model is made as close as possible to the high-precision TTV value measured by FRT, thereby minimizing the error between the predicted value and the true value. During the training process, the TTV compensation model can learn how to map the low-resolution online measured TTV value to the high-resolution FRT measured TTV value. After the TTV compensation model is trained, the initial TTV value obtained by the online measurement can be input into the TTV compensation model, and the TTV compensation model can output a corrected TTV value to make it closer to the true high-precision FRT measurement result.

[0023] Step 103: If the corrected TTV value is not within the set specification, determine the optimal TTV angle parameter through the pre-trained prediction model; wherein the environmental characteristic data during the wafer thinning process is used as a constant of the prediction model, the TTV angle parameter is used as the optimization target of the prediction model, and the expected TTV value is used as the output target of the prediction model to perform a traversal of the feasible set of TTV angles to obtain the optimal angle parameter; wherein the environmental characteristic data includes process parameter characteristics; The prediction model can be trained based on a large amount of historical data and actual measurement results. The prediction model can predict the corresponding TTV value based on the input environmental characteristic data and TTV angle parameters. Among them, the environmental characteristic data can be used as a constant of the prediction model. The environmental characteristic data can include process parameter characteristics, such as temperature, pressure, and rotation speed during the thinning process. The environmental characteristic data is relatively stable during the production process and has an important impact on the TTV value.

[0024] The TTV angle parameter can be the optimization target of the prediction model. It is necessary to adjust the TTV angle parameter so that the TTV value output by the prediction model reaches the expected target. The specific operation is to perform a traversal of the feasible set of TTV angles, that is, to conduct a comprehensive search and attempt within the possible value range of the TTV angle parameter, calculate the predicted TTV value corresponding to each possible angle parameter combination, and then select the optimal angle parameter combination that makes the predicted TTV value closest to the expected value from each possible angle parameter combination.

[0025] Step 104: Control the grinding disc angle in the wafer thinning processing equipment to adjust to the TTV optimal angle parameter to act on the next wafer thinning thickness processing.

[0026] During the wafer thinning process, the angle of the grinding disc can affect the TTV of the wafer. After the optimal TTV angle parameters are determined through the prediction model, the angle of the grinding disc in the wafer thinning equipment needs to be adjusted accordingly. Furthermore, the angle of the grinding disc is adjusted to the optimal angle parameter to ensure that the expected TTV value can be achieved during the next wafer thinning process to meet the quality requirements of wafer thinning.

[0027] In this embodiment, during the wafer thinning process, the method uses a TTV compensation model trained based on the measurement data of an online TTV measurement module and an FRT measurement device to correct the initial TTV value to obtain a corrected TTV value that is closer to the FRT measurement result; if the corrected TTV value exceeds the set specification, the optimal TTV angle parameter is determined by a pre-trained prediction model, the prediction model uses environmental feature data (including process parameter features) as a constant, and the TTV angle parameter as the optimization target, and obtains the optimal angle parameter by traversing the feasible set; the grinding disc angle in the wafer thinning processing equipment is adjusted to the optimal TTV angle parameter to optimize the next wafer thinning process. This method can achieve real-time precise control and continuous optimization of the wafer thinning process, and improve production efficiency and product quality.

[0028] In one embodiment, Figure 2 As shown, based on the initial TTV value, a TTV compensation model is used to obtain a corrected TTV value, including the following steps: Step 201: inputting the initial TTV value into the high-low resolution mapping model to obtain the virtual TTV value measured by the virtual FRT measurement device; the high-low resolution mapping model is obtained by mapping the TTV values ​​measured on the same wafer using at least the online TTV measurement module and the FRT measurement device; The high- and low-resolution mapping model can be obtained by mapping the TTV values ​​measured by the online TTV measurement module and the FRT measurement device on the same wafer. It should be noted that the online TTV measurement module can obtain the initial TTV value of the wafer in real time during the thinning process, but due to the small number of measurement points and low sampling resolution, the accuracy of the initial TTV value is limited. The FRT measurement device can perform more detailed and precise measurements on the same wafer after wafer thinning is completed. The number of measurement points collected is large and the sampling resolution is high, which can provide more accurate TTV values.

[0029] Furthermore, the high-low resolution mapping model can be trained using the relationship between the two measurement results to learn how to map the low-resolution TTV value measured online to the high-resolution FRT measurement value. For example, the initial TTV value can be input into the high-low resolution mapping model, and the high-low resolution mapping model can output a virtual high-precision and high-resolution TTV value based on the previously learned mapping relationship. The virtual TTV value can be closer to the real high-precision measurement result, which helps to more accurately evaluate and control the flatness of the wafer during the production process.

[0030] Step 202: The initial TTV value and the virtual TTV value are concatenated and input into a TTV compensation model to obtain a corrected TTV value.

[0031] The initial TTV value is low-resolution data obtained by the online TTV measurement module, and the virtual TTV value is high-resolution data obtained by mapping the initial TTV value through the high-low resolution mapping model. The two TTV values ​​are spliced ​​together to form a more comprehensive input vector.

[0032] It should be noted that the TTV compensation model can adopt a neural network architecture, and optimize the model parameters by using a loss function such as mean square error through forward propagation and back propagation algorithms. During the training process, the TTV compensation model can continuously adjust the weights and biases to minimize the difference between the predicted value and the actual FRT measurement value.

[0033] The spliced ​​initial TTV value and the virtual TTV value are input into the trained TTV compensation model, which can output a corrected TTV value. The corrected TTV value can be closer to the real high-precision FRT measurement result, which helps to more accurately evaluate and control the flatness of the wafer during the production process.

[0034] In this embodiment, the method can make the online measurement result closer to the actual high-precision FRT measurement result.

[0035] In one embodiment, Figure 3 As shown, the TTV value measured online can be mapped through the high-resolution and low-resolution mapping models to obtain a high-resolution virtual TTV value. The high-resolution virtual TTV value and the low-resolution TTV value are spliced, and the initial TTV value and the virtual TTV value are spliced ​​and input into the TTV compensation model. The TTV compensation model combines these data and outputs a corrected TTV value.

[0036] In one embodiment, Figure 4 As shown, training the TTV compensation model includes the following steps: Step 401: Establishing a TTV compensation model based on a neural network; Collect online TTV measurement values ​​and FRT measurement values ​​of the same wafer. These measured TTV value data can be used to train the TTV compensation model. Furthermore, to build the TTV compensation model, the data needs to be cleaned and normalized, and divided into training sets and test sets. Select a suitable neural network architecture, such as a multilayer perceptron or a convolutional neural network, define the number of neurons in the input layer, hidden layer, and output layer, select the activation function and loss function, and use the optimization algorithm to adjust the model parameters so that the corrected TTV value output by the TTV compensation model is as close to the actual FRT measurement value as possible. After training, the test set can be used to evaluate the performance of the TTV compensation model, and the TTV compensation model can be adjusted and optimized if necessary.

[0037] Step 402: splicing the first resolution data and the second resolution data of the same wafer to obtain spliced ​​data; wherein, for the same wafer, the first resolution data at least includes the TTV value measured by the online TTV measurement module, and the second resolution data at least includes the TTV value measured by the FRT measurement device; The first resolution data includes at least the TTV value measured by the online TTV measurement module. These data are acquired in real time during the thinning process. The number of measurement points is small and the sampling resolution is low. They are mainly used to quickly detect the flatness of the wafer. The second resolution data includes at least the TTV value measured by the FRT measurement equipment. These data are obtained by more detailed and precise measurement of the same wafer after the wafer thinning is completed. The number of measurement points is large and the sampling resolution is high. It can provide more accurate and detailed wafer flatness information.

[0038] By stitching the data of these two resolutions, we can combine the advantages of both, taking advantage of the real-time nature of online measurement and the high precision of FRT measurement, providing a richer and more comprehensive data basis for subsequent analysis and modeling. For example, the low-resolution TTV value measured online and the high-resolution TTV value measured by FRT can be matched and stitched according to the unique identification (wafer ID) of the wafer to form a complete data set for training the TTV compensation model, thereby improving the accuracy and reliability of the TTV compensation model.

[0039] Step 403: Using the spliced ​​data as input of an initial TTV compensation model, using the TTV value in the second resolution data used for splicing as output, and training the TTV compensation model.

[0040] The spliced ​​data is input into the TTV compensation model, and the high-precision second resolution data, that is, the TTV value measured by FRT, is used as the target output. Furthermore, by adjusting the parameters of the TTV compensation model, the TTV compensation model can learn how to relate the low-resolution online measurement value to the high-resolution FRT measurement value, and then output a more accurate corrected TTV value based on the initial TTV value and the virtual TTV value measured online, so that it is closer to the real high-precision FRT measurement result.

[0041] In this embodiment, the method establishes a TTV compensation model so that the initial TTV value measured online can be supplemented by the model to generate a high-precision FRT measurement result close to the actual value, which helps to more accurately evaluate and control the flatness of the wafer.

[0042] In one embodiment, Figure 5 As shown, determining the high and low resolution mapping models includes the following steps: Step 501: Establish an encoder f for converting second resolution data to first resolution data, and establish a decoder g for converting first resolution data to second resolution data; The second resolution data may refer to the TTV values ​​of high resolution and multiple measurement points collected by the FRT measurement equipment. The first resolution data refers to the TTV values ​​of low resolution and a small number of measurement points collected by the online TTV measurement module.

[0043] The role of the encoder f can be to convert the high-resolution second resolution data into low-resolution first resolution data. This can be achieved by downsampling, feature extraction, or dimensionality reduction of the high-resolution data. Exemplarily, the convolution layer or pooling layer in the neural network can be used to extract the main features of the high-resolution data and map it to the low-resolution space. Further, the encoder f can learn how to extract key information from the high-resolution data and represent it in the form of low-resolution data.

[0044] Decoder g is to convert the low-resolution first-resolution data back to high-resolution second-resolution data. This process is relatively complex and requires upsampling, deconvolution or other generative methods to restore the high-resolution details of the data. For example, the deconvolution layer or pixel shuffle operation in the neural network can be used to gradually enlarge the low-resolution data and fill in the details to make it close to the original high-resolution data. Among them, the goal of decoder g is to reconstruct the high-resolution data as accurately as possible while retaining the main features in the low-resolution data.

[0045] Step 502: Train an encoder f and a decoder g based on the first resolution data and the second resolution data of the same wafer, and obtain a high-low resolution mapping model based on the trained encoder f and decoder g.

[0046] After sufficient training, the encoder f can learn how to extract key information from high-resolution data and map it to low-resolution space, while the decoder g can recover the details of high-resolution data from low-resolution data. Combining the trained encoder f and decoder g can form a high-low resolution mapping model. That is, after obtaining the low-resolution TTV value measured online, the high-low resolution mapping model can be used to convert the low-resolution TTV value into a virtual high-resolution TTV value.

[0047] For example, for the first resolution data and the second resolution data of the same wafer, a data mapping model between high and low resolutions can be constructed. Assume that the first resolution data has The data is .

[0048] The second resolution data has indivual, ,exist For this wafer, the first resolution data has few points and belongs to low-dimensional samples; the second resolution data has many points and belongs to high-dimensional samples. For the mapping management between high- and low-dimensional samples, establish arrive Encoder f , and reverse training arrive Decoder g , and obtain high and low resolution mapping models.

[0049] Furthermore, the encoder g and TTV compensation model in the trained high- and low-resolution mapping model are extracted to form an online low-resolution TTV correction model. As encoder input, obtain virtual high-dimensional input ,by and Input into the compensation model to obtain the corrected TTV value.

[0050] In this embodiment, the trained encoder and decoder can accurately convert low-resolution online measurement data into high-resolution virtual FRT measurement data, thereby improving the accuracy and reliability of online measurement and making it closer to the actual high-precision FRT measurement results.

[0051] In one embodiment, Figure 6 As shown, it mainly includes encoder f, decoder g and TTV compensation model. High-resolution data is converted into low-resolution data by encoder f, and decoder g restores low-resolution data to high-resolution data. Encoder f and decoder g are trained based on high-resolution data and low-resolution data of the same wafer, and a high-low resolution mapping model is obtained based on the trained encoder f and decoder g.

[0052] In one embodiment, Figure 7 As shown, training the prediction model includes the following steps: Step 701: obtaining sample characteristics of a sample wafer, wherein for any sample wafer, the sample characteristics include environmental characteristic data, TTV angle parameters and TTV values; For any sample wafer, its sample characteristics include environmental characteristic data, TTV angle parameters and TTV values. Among them, the environmental characteristic data may include various process parameters in the wafer thinning process, such as temperature, pressure, rotation speed, etc. The process parameters have an important influence on the thinning effect of the wafer. The TTV angle parameter can affect the thinning thickness and surface flatness of the wafer. The TTV value may refer to the initial TTV value and the final high-precision TTV value measured by the online TTV measurement module and the FRT measurement equipment respectively, and the more accurate wafer flatness data obtained after correction by the TTV compensation model. These data together constitute the sample characteristics of the sample wafer. By analyzing and modeling the sample characteristics of the sample wafer, the TTV of the wafer can be more accurately predicted and controlled, the thinning process can be optimized, and the product quality can be improved.

[0053] Step 702: construct a prediction model based on sample characteristics; Collect historical sample features of multiple sample wafers, organize and preprocess these sample feature data, including data cleaning, normalization, feature selection and other operations to ensure data quality and consistency. Furthermore, a suitable algorithm can be selected to build a prediction model, and the environmental feature data and TTV angle parameters in the sample features are used as input variables, and the corrected TTV value is used as the target output variable. By adjusting the parameters of the prediction model, the prediction model can learn the mapping relationship between the input variables and the output variables. Furthermore, after the prediction model training is completed, a part of the sample data that did not participate in the training can be used to verify and test the prediction model to evaluate the accuracy and generalization ability of the prediction model, such as by calculating the root mean square error, mean absolute error and other indicators to measure the prediction performance of the prediction model.

[0054] For example, based on the sample characteristics of the sample wafer and the corrected TTV value, a wafer thinning TTV prediction model is established. The constructed prediction model can be a model .in, is the eigenvector, is the target variable, is the model parameter.

[0055] Step 703: Introduce the L1 regularization coefficient to shrink the coefficients of non-correlated feature data in the prediction model; The training process of the prediction model is based on a large amount of sample feature data. Among the sample feature data, some sample feature data are closely related to the TTV value and are the key factors affecting the TTV value; while some sample feature data have little relationship with the TTV value, or even have no direct correlation, that is, non-correlated feature data.

[0056] The L1 regularization coefficient can refer to adding a penalty term to the loss function so that the prediction model tends to shrink the coefficients of these non-correlated feature data to zero during the training process. Features that do not contribute significantly to the prediction of the TTV value will be automatically excluded from the prediction model, thereby achieving the purpose of feature selection. For example, during the training of the prediction model, if a certain environmental feature data (such as temperature) has a negligible effect on the TTV value, after the introduction of L1 regularization, the coefficient corresponding to the temperature feature will be shrunk to zero (without affecting the effect of the prediction model), and the temperature feature will no longer participate in the prediction calculation of the prediction model.

[0057] Step 704: introducing an L2 regularization term to limit the weight of feature data; The L2 regularization term adds a penalty term proportional to the sum of the squares of the feature data weights to the loss function, which makes the prediction model tend to reduce the absolute value of the weight during the training process. The weights of features that have a smaller impact on the prediction of the TTV value will be further reduced, thereby reducing interference with the output of the prediction model. For example, during the training of the prediction model, if a certain environmental feature data (such as pressure) has a relatively small impact on the TTV value, after the L2 regularization term is introduced, the weight corresponding to the pressure feature will be limited to a smaller range to avoid excessive impact on the output of the prediction model.

[0058] Step 505: Use the gradient descent method to train the prediction model to gradually adjust the parameters of the prediction model.

[0059] Gradient descent is an iterative optimization algorithm. Its core idea is to calculate the gradient of the loss function with respect to the parameters of the prediction model and update the parameters in the opposite direction of the gradient, thereby gradually reducing the value of the loss function and making the prediction results of the prediction model closer to the actual value. During the training process, a suitable loss function, such as mean square error (MSE), is defined to measure the difference between the TTV value predicted by the prediction model and the actual TTV value. Further, the parameters of the prediction model are initialized, usually using random initialization or zero initialization. The sample features in the training data set are input into the prediction model, the prediction output of the prediction model is calculated, and the loss value under the current parameters is calculated according to the loss function. The gradient of the loss function for each parameter is calculated by the back propagation algorithm, and the parameters of the prediction model are updated according to the learning rate and the gradient value. This process is repeated many times until the value of the loss function converges to a smaller range or the preset maximum number of iterations is reached.

[0060] Using the gradient descent method to train the prediction model can effectively optimize the model parameters and improve the prediction accuracy and stability of the prediction model. For example, in the prediction model of wafer thinning, adjusting the parameters by the gradient descent method can enable the prediction model to more accurately capture the complex relationship between environmental feature data and TTV angle parameters and TTV values.

[0061] Exemplarily, a training data set is collected, which contains an input feature vector x and a target variable y. A loss function J(β) is defined, which is usually composed of a data fitting error term and a regularization term, such as L1 and L2 regularization terms, to prevent overfitting and improve the stability of the prediction model. The form of the loss function is: .

[0062] Where m is the number of samples, y i is the true value of the i-th sample, x i is the feature vector of the i-th sample, β is the model parameter, λ 1 and λ 2 is the regularization coefficient.

[0063] The training process is to initialize the model coefficient β, usually using a zero vector. Iterate over all coefficients β and convert each coefficient β j As variables, lock other coefficients to the results of the previous calculation and treat them as constants to find β under the current conditions j The optimal solution. At the kth iteration, the weight coefficient is updated Need to meet: .

[0064] In each iteration, the loss function is calculated with respect to each parameter β j The gradient of involves calculating the derivative of the data fitting error and the derivative of the regularization term: .

[0065] Update each parameter β according to the gradient value j , so that it gradually approaches the optimal value. Repeat the above process until the value of the loss function converges to a stable state or reaches the preset maximum number of iterations.

[0066] In this embodiment, the method can effectively remove the interference of non-correlated features, simplify the prediction model structure, limit feature weights, avoid overfitting, and improve the stability and generalization ability of the prediction model through the combination of L1 regularization and L2 regularization. The trained prediction model can more accurately predict the TTV value of the wafer based on the input environmental feature data and TTV angle parameters.

[0067] In one embodiment, Figure 8 As shown, the optimization of TTV angle parameters includes the following steps: Step 801: Determine the value range of the TTV angle parameter; According to the physical limitations and process requirements of the wafer thinning equipment, the maximum and minimum values ​​of the TTV angle parameters in actual operation can be clearly defined. The physical limitations can be provided by the manufacturer to ensure the safe operation of the physical limitations and the stability of the process. For example, the angle adjustment mechanism of the grinding disc may only be physically adjusted within a certain range. Exceeding this range may cause damage to the equipment or malfunction.

[0068] Different wafer materials and thinning targets may require different angle settings, and process requirements are derived from actual production experience and a large amount of experimental data. You can also refer to historical data and actual production experience to further refine and adjust the theoretical value range to ensure that in actual applications, the value of the TTV angle parameter can meet both the physical limitations of the equipment and the optimization requirements of the process, thereby providing a reliable basis for subsequent model optimization and process parameter adjustment.

[0069] Step 802: traverse the parameter grid within the value range, and for each parameter combination in the parameter grid, use the prediction model to calculate the TTV value; the TTV angle parameter includes a first angle parameter of the first support leg and a second angle parameter of the second support leg, and the parameter combination consists of the first angle parameter and the second angle parameter; The TTV angle parameter includes a first angle parameter of the first support leg (the first angle parameter may be ) and a second angle parameter of the second support foot (the second angle parameter may be ). The parameter combination may consist of a first angle parameter and a second angle parameter.

[0070] Determine the physical limitations of the equipment and process requirements and The value range of The range can be [0°,10°], The range of can be [5°,15°]. The scope and The range of is divided into multiple equally spaced values ​​to generate a parameter grid. For example, The range is divided into 0°, 2°, 4°, 6°, 8°, 10°. The range is divided into 5°, 8°, 11°, and 14°. The parameter grid will contain 6×4=24 groups of different parameter combinations.

[0071] Step 803: In the traversal parameter grid, select the parameter combination that minimizes the TTV value as the optimal TTV angle parameter.

[0072] For example, suppose that at a certain grid point, the parameter combination ( =5°, =10°) corresponds to a TTV value of 2.0. At another grid point, the parameter combination ( =6°, =12°) corresponds to a TTV value of 1.8, the latter may be a better parameter combination.

[0073] In this embodiment, by comparing the TTV values ​​of all combinations, the parameter combination with the smallest TTV value is selected as the optimal solution. This method can fully cover possible parameter combinations, avoid missing the optimal solution, and improve process accuracy and product quality.

[0074] In one embodiment, the environmental characteristic data includes time domain characteristics and frequency domain characteristics in addition to process parameter characteristics; wherein, obtaining the environmental characteristic data includes: building a data acquisition device, the data acquisition device can collect process parameters, process parameters and TTV values ​​in the wafer thinning process in real time, the data acquisition device includes a PLC data real-time acquisition module built into the wafer thinning processing equipment, and an external acquisition sensor module; extracting features of the collected process parameters and process parameters to obtain time domain characteristics, frequency domain characteristics and process parameter characteristics.

[0075] During the wafer thinning process, the collection of environmental characteristic data is achieved by building a special data acquisition device. The data acquisition device can collect various key information in the wafer thinning process in real time, providing a basis for subsequent data analysis and model building.

[0076] The data acquisition device includes a built-in PLC data real-time acquisition module and an external acquisition sensor module. Among them, the PLC data real-time acquisition module is integrated inside the wafer thinning processing equipment and is responsible for obtaining the internal process parameters of the equipment during operation, such as temperature, pressure, speed, etc. The internal process parameters are usually directly measured by sensors inside the equipment, which can reflect the operating status and process conditions of the equipment. The external acquisition sensor module is installed outside the equipment to collect other process parameters, such as current value, vacuum pressure, vibration frequency, temperature of the grinding area, etc. Process parameters can be measured by additional sensors, such as accelerometers used to detect the vibration of the equipment, and ammeters used to collect current signals during the processing process.

[0077] The collected raw data needs to be preprocessed to ensure the quality and consistency of the data, including data cleaning to remove outliers and noise; data normalization to scale data of different ranges to a unified range; and data alignment to ensure the consistency of data from different sources in time.

[0078] The preprocessed data will be subjected to feature extraction to obtain time domain features, frequency domain features and process parameter features. Time domain features include mean, variance, peak value, etc., which can reflect the distribution and change trend of data in time; frequency domain features can be obtained through Fourier transform and other methods, such as power spectrum density, frequency peak, etc., which reveal the distribution and periodic changes of data in frequency. The extracted time domain features, frequency domain features, process parameter features and surface shape results TTV form the sample features of each wafer. , .

[0079] in and They are the time domain characteristics and frequency domain characteristics of the timing parameters respectively; is the process parameter characteristic; and is the angle adjustment value of TTV.

[0080] In this embodiment, by comprehensively collecting and processing environmental feature data, rich and accurate information can be provided for subsequent model training and prediction. The extraction of time domain features and frequency domain features can capture the changing patterns of data in time and frequency, which helps to discover potential process problems and optimization points.

[0081] In one embodiment, a PLC data real-time acquisition module built into the wafer thinning processing equipment is used to collect process parameters and process parameters inside the wafer thinning processing equipment through the TCP / IP protocol; The external acquisition sensor module is used to realize A / D conversion of analog signals through the data acquisition card and encapsulate digital signals through the Modbus protocol.

[0082] The PLC data real-time acquisition module built into the wafer thinning processing equipment collects the process parameters and process parameters inside the equipment through the TCP / IP protocol. As the core equipment of industrial automation, PLC can monitor and control the operating status of the equipment in real time. During the wafer thinning process, the PLC module is responsible for collecting various process parameters inside the equipment, such as temperature, pressure, speed, feed rate, etc. The process parameters can be measured by the sensors inside the equipment. The PLC module converts these analog signals into digital signals and transmits them to the data storage server or control system through the TCP / IP protocol to realize real-time monitoring and data recording of the equipment's operating status.

[0083] The external acquisition sensor module is used to realize the A / D conversion of analog signals through the data acquisition card and encapsulate the digital signal through the Modbus protocol. The external acquisition sensor module usually includes various sensors, such as vibration sensors, sound sensors, current sensors, etc. These sensors are used to monitor the environmental parameters outside the equipment and the physical quantities of the equipment during operation. The data acquisition card performs A / D conversion on these analog signals, that is, converts the continuous analog signals into discrete digital signals for computer processing and analysis. The converted digital signals are encapsulated through the Modbus protocol. Modbus is a protocol widely used in industrial communications that allows data exchange between devices from different manufacturers. The encapsulated digital signals can be transmitted to the host computer or other control system for further processing and analysis.

[0084] In this embodiment, through the coordinated work of these two modules, comprehensive monitoring and data collection of the wafer thinning process can be achieved, providing a basis for subsequent data analysis, model building and process optimization.

[0085] In one embodiment, Fig. 9 The figure shows the run-to-run control process of wafer thinning. The historical data of wafers of the same thinning equipment are collected through the data acquisition module, including time domain features, frequency domain features, process parameter features, and surface result TTV data. Time series feature extraction, time series noise reduction, and feature extraction after frequency domain transformation are performed on multimodal data to convert multi-source data into structured data. A TTV correction model is established using online measurement and FRT high-resolution measurement data in historical data. According to the surface result of the wafer, the corrected TTV prediction value is used as the TTV value of the surface result. When the processing TTV of the wafer is within the specification, the current process is maintained to continue production when processing the next wafer. When the processing TTV of the wafer exceeds the specification, the TTV angle adjustment value is obtained by optimization control.

[0086] Using the collected wafer thinning data , construct a high-dimensional, nonlinear regression prediction model for TTV, use RMS as the loss function and introduce L1 and L2 regularization terms. Optimize the prediction model parameters and obtain the current optimal prediction model when the training conditions are met. In actual operation, as the wafer data increases and accumulates, the wafer data is dynamically updated and a new prediction model is established. .

[0087] Using predictive models When predicting the TTV of the next wafer processing, the prediction model structure remains unchanged. , keep the process parameters using the set values ​​without adjustment; for the time domain and frequency domain characteristics of the timing parameters, the next wafer to be predicted is not processed yet and the process data cannot be obtained temporarily, so the average value of the process data of the current most recent n wafers is used as the wafer to be predicted. and . Use grid search to optimize the angle adjustment parameters and calculate the and As the recommended TTV angle adjustment value.

[0088] In one embodiment, Fig. 9 As shown in the figure, a closed-loop control process is formed through real-time measurement, online data acquisition, model prediction and parameter optimization, aiming to continuously improve the accuracy and quality of wafer thinning. The high- and low-resolution TTV correction models map the low-resolution online TTV measurement data to the high-resolution FRT data to improve the accuracy of online measurement. The IPG thickness measurement and online TTV measurement modules obtain the thickness and TTV value of the wafer in real time during the thinning process, and these data are used together with the process parameters to dynamically adjust the thinning process. Among them, IPG thickness measurement refers to the real-time measurement of the thickness of the wafer using integrated process thickness measurement technology during the wafer thinning process. FRT measurement is performed after the wafer thinning is completed, providing high-precision flatness data for final quality assessment and model training. According to the measurement results and process parameters of the current wafer, the optimal angle parameters are calculated, and the angle of the grinding disc is adjusted to optimize the subsequent wafer thinning process. The data acquisition system runs through the entire process, collecting various data to provide support for model training and process control.

[0089] Based on the same concept, Fig.10 As shown, the present application also provides a control device for automatically adjusting the wafer thinning thickness, the device comprising: The acquisition module 1001 is used to acquire the initial TTV value measured by the online TTV measurement module during the wafer thinning process; wherein the measurement points collected by the online TTV measurement module are smaller than the measurement points collected by the FRT measurement device, and the sampling resolution is smaller than the sampling resolution of the FRT measurement device, and the FRT measurement device measures the TTV value of the same wafer after the thinning is completed; A compensation module 1002 is used to obtain a corrected TTV value based on the initial TTV value using a TTV compensation model; wherein the TTV compensation model is obtained by training the TTV value measured by the same wafer using at least an online TTV measurement module and a FRT measurement device; The optimization module 1003 is used to determine the optimal TTV angle parameter through a pre-trained prediction model if the corrected TTV value is not within the set specification, use the environmental characteristic data in the wafer thinning process as a constant of the prediction model, use the TTV angle parameter as an optimization target of the prediction model, and use the expected TTV value as an output target of the prediction model to perform a traversal of the feasible set of TTV angles to obtain the optimal angle parameter; wherein the environmental characteristic data includes process parameter characteristics; The adjustment module 1004 is used to control the grinding disc angle in the wafer thinning processing equipment to be adjusted to the TTV optimal angle parameter to act on the next wafer thinning thickness processing.

[0090] In one embodiment, the compensation module 1002 uses a TTV compensation model based on the initial TTV value to obtain a corrected TTV value, specifically for: inputting the initial TTV value into a high- and low-resolution mapping model to obtain a virtual TTV value measured by a virtual FRT measurement device; the high- and low-resolution mapping model is obtained by mapping the TTV values ​​measured on the same wafer using at least an online TTV measurement module and a FRT measurement device; the initial TTV value and the virtual TTV value are spliced ​​and input into the TTV compensation model to obtain a corrected TTV value.

[0091] In one embodiment, the compensation module 1002 trains the TTV compensation model including: establishing a TTV compensation model based on a neural network; splicing the first resolution data and the second resolution data of the same wafer to obtain spliced ​​data; wherein, for the same wafer, the first resolution data at least includes the TTV value measured by the online TTV measurement module, and the second resolution data at least includes the TTV value measured by the FRT measurement device; using the spliced ​​data as the input of the initial TTV compensation model, and using the TTV value in the second resolution data used for splicing as the output, to train the TTV compensation model.

[0092] In one embodiment, the compensation module 1002 determines a high- and low-resolution mapping model, which is specifically used to: establish an encoder f from second resolution data to first resolution data, and establish a decoder g from first resolution data to second resolution data; train the encoder f and decoder g based on the first resolution data and the second resolution data of the same wafer, and obtain a high- and low-resolution mapping model based on the trained encoder f and decoder g.

[0093] In one embodiment, the optimization module 1003 trains the prediction model to obtain sample features of the sample wafers, wherein, for any sample wafer, the sample features include environmental feature data, TTV angle parameters and TTV values; based on the sample features, a prediction model is constructed; an L1 regularization coefficient is introduced to shrink the coefficients of non-correlated feature data in the prediction model; an L2 regularization term is introduced to limit the weight of the feature data; and the prediction model is trained using the gradient descent method to gradually adjust the parameters of the prediction model.

[0094] In one embodiment, the optimization module 1003 is specifically used to determine the value range of the TTV angle parameter; traverse the parameter grid within the value range, and use the prediction model to calculate the TTV value for each group of parameter combinations in the parameter grid; the TTV angle parameter includes a first angle parameter of the first support leg and a second angle parameter of the second support leg, and the parameter combination consists of the first angle parameter and the second angle parameter; in the traversed parameter grid, select the parameter combination that minimizes the TTV value as the optimal TTV angle parameter.

[0095] In one embodiment, the environmental characteristic data includes not only process parameter characteristics but also time domain characteristics and frequency domain characteristics; wherein, the environmental characteristic data is obtained. The optimization module 1003 is specifically used to build a data acquisition device, which can collect process parameters, process parameters and TTV values ​​in the wafer thinning process in real time. The data acquisition device includes a built-in PLC data real-time acquisition module and an external acquisition sensor module of the wafer thinning processing equipment; the collected process parameters and process parameters are feature extracted to obtain time domain characteristics, frequency domain characteristics and process parameter characteristics.

[0096] In one embodiment, the optimization module 1002 is a PLC data real-time acquisition module built into the wafer thinning processing equipment, which is used to collect the process parameters and process parameters inside the wafer thinning processing equipment through the TCP / IP protocol; the external acquisition sensor module is used to realize the A / D conversion of analog signals through the data acquisition card, and encapsulate the digital signal through the Modbus protocol.

[0097] Based on the same concept, Fig.11 As shown, the present application also provides a wafer thinning system 1100, which includes: a wafer thinning processing equipment 1110, which is used to thin the wafer; an online TTV measurement module 1120, which collects the thickness data of the wafer in real time during the wafer thinning process of the wafer by the wafer thinning processing equipment; a control module 1130, including a memory 1131 and a processor 1132, the memory 1131 stores a computer program, and the processor 1132 implements a control method for automatically adjusting the wafer thinning thickness when executing the computer program.

[0098] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A control method for automatically adjusting the wafer thinning thickness, characterized in that: The method is applied to a control module in a wafer thinning system, wherein the wafer thinning system further comprises an online TTV measurement module and a wafer thinning processing device, and the method comprises: Obtaining an initial TTV value measured by the online TTV measurement module during wafer thinning; wherein the measurement points collected by the online TTV measurement module are smaller than the measurement points collected by the FRT measurement device, and the sampling resolution is smaller than the sampling resolution of the FRT measurement device, and the FRT measurement device is a TTV value measurement performed on the same wafer after thinning is completed; Based on the initial TTV value, a TTV compensation model is used to obtain a corrected TTV value; wherein the TTV compensation model is obtained by training the TTV value measured on the same wafer using at least the online TTV measurement module and the FRT measurement device; If the corrected TTV value is not within the set specification, the optimal TTV angle parameter is determined by a pre-trained prediction model; wherein the environmental characteristic data during wafer thinning is used as a constant of the prediction model, the TTV angle parameter is used as an optimization target of the prediction model, and the expected TTV value is used as an output target of the prediction model to perform a traversal of the feasible set of the TTV angle to obtain the optimal angle parameter; wherein the environmental characteristic data includes process parameter characteristics; The grinding disc angle in the wafer thinning processing equipment is controlled to be adjusted to the TTV optimal angle parameter to act on the next wafer thinning thickness processing.

2. The control method for automatically adjusting wafer thinning thickness according to claim 1, characterized in that: Based on the initial TTV value, a TTV compensation model is used to obtain a corrected TTV value, including: Inputting the initial TTV value into a high-low resolution mapping model to obtain a virtual TTV value measured by a virtual FRT measurement device; the high-low resolution mapping model is obtained by mapping TTV values ​​measured on the same wafer using at least the online TTV measurement module and the FRT measurement device; The initial TTV value and the virtual TTV value are spliced ​​and input into a TTV compensation model to obtain a corrected TTV value.

3. The control method for automatically adjusting wafer thinning thickness according to claim 2, characterized in that: Training the TTV compensation model includes: Establishing the TTV compensation model based on a neural network; Splicing the first resolution data and the second resolution data of the same wafer to obtain spliced ​​data; wherein, for the same wafer, the first resolution data at least includes the TTV value measured by the online TTV measurement module, and the second resolution data at least includes the TTV value measured by the FRT measurement device; The spliced ​​data is used as input of the initial TTV compensation model, and the TTV value in the second resolution data used for splicing is used as output to train the TTV compensation model.

4. The control method for automatically adjusting wafer thinning thickness according to claim 2, characterized in that: Determining the high and low resolution mapping models includes: Establishing an encoder f from the second resolution data to the first resolution data, and establishing a decoder g from the first resolution data to the second resolution data; The encoder f and the decoder g are trained based on the first resolution data and the second resolution data of the same wafer, and a high-low resolution mapping model is obtained based on the trained encoder f and decoder g.

5. The control method for automatically adjusting wafer thinning thickness according to claim 1, characterized in that: Training the prediction model includes: Acquire sample characteristics of the sample wafer, wherein for any sample wafer, the sample characteristics include environmental characteristic data, TTV angle parameters and TTV values; Based on the sample characteristics, construct a prediction model; Introducing an L1 regularization coefficient to shrink the coefficients of non-correlated feature data in the prediction model; Introduce L2 regularization term to limit the weight of feature data; The prediction model is trained using a gradient descent method to gradually adjust the parameters of the prediction model.

6. The control method for automatically adjusting wafer thinning thickness according to claim 1, characterized in that: The optimization of the TTV angle parameters includes: Determine the value range of TTV angle parameters; The parameter grid is traversed within the value range, and for each parameter combination in the parameter grid, the TTV value is calculated using the prediction model; the TTV angle parameter includes a first angle parameter of the first support foot and a second angle parameter of the second support angle, and the parameter combination consists of the first angle parameter and the second angle parameter In the traversal parameter grid, the parameter combination that minimizes the TTV value is selected as the optimal TTV angle parameter.

7. The control method for automatically adjusting wafer thinning thickness according to any one of claims 1 to 6, characterized in that: The environmental characteristic data includes not only the process parameter characteristics but also time domain characteristics and frequency domain characteristics; Wherein, obtaining the environmental characteristic data includes: Building a data acquisition device, which can collect process parameters, process parameters and TTV values ​​in real time during wafer thinning, and the data acquisition device includes a built-in PLC data real-time acquisition module of the wafer thinning processing equipment and an external acquisition sensor module; The collected process parameters and procedure parameters are subjected to feature extraction to obtain time domain features, frequency domain features and process parameter features.

8. The control method for automatically adjusting wafer thinning thickness according to claim 7, characterized in that: The PLC data real-time acquisition module built into the wafer thinning processing equipment is used to collect the process parameters and process parameters inside the wafer thinning processing equipment through the TCP / IP protocol; The external data acquisition sensor module is used to realize A / D conversion of analog signals through a data acquisition card and encapsulate digital signals through the Modbus protocol.

9. A control device for automatically adjusting the wafer thinning thickness, characterized in that: The device comprises: An acquisition module, used for acquiring an initial TTV value measured by the online TTV measurement module during the wafer thinning process; wherein the measurement points collected by the online TTV measurement module are smaller than the measurement points collected by the FRT measurement device, and the sampling resolution is smaller than the sampling resolution of the FRT measurement device, and the FRT measurement device is a TTV value measurement performed on the same wafer after the thinning is completed; A compensation module, configured to obtain a corrected TTV value based on the initial TTV value using a TTV compensation model; wherein the TTV compensation model is obtained by training the TTV values ​​measured on the same wafer using at least the online TTV measurement module and the FRT measurement device; An optimization module, for determining the optimal TTV angle parameter through a pre-trained prediction model if the corrected TTV value is not within the set specification, using the environmental characteristic data in the wafer thinning process as a constant of the prediction model, using the TTV angle parameter as an optimization target of the prediction model, and using the expected TTV value as an output target of the prediction model to perform a traversal of the feasible set of the TTV angle to obtain the optimal angle parameter; wherein the environmental characteristic data includes process parameter characteristics; The adjustment module is used to control the grinding disc angle in the wafer thinning processing equipment to be adjusted to the TTV optimal angle parameter to act on the next wafer thinning thickness processing.

10. A wafer thinning system, characterized in that: The wafer thinning system comprises: Wafer thinning processing equipment, used for performing thinning processing on the wafer; An online TTV measurement module collects thickness data of the wafer in real time during the wafer thinning process of the wafer thinning equipment; A control module comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the control method for automatically adjusting the wafer thinning thickness as described in any one of claims 1 to 8 when executing the computer program.

Images